Ar

chi
vedWebCont
ent Cont
enuWebar
chi
vé
Saving the Twentieth Century: Sauvegarder le XXe siicle :
The Conservation of Modern la conservation des mat6riaux
Materials modernes

Proceedings ofa Conference Les actes de la conf6rence

Symposium '91 - Saving the Twentieth Century Symposium 9l - Sauvegarder le XXe sidcle
Ottawa, Canada Ottawa, Canada
15 to 20 September 1991 du 15 au 20 septembre l99l

Organized by the Canadian Conservation Organis6 par I'Institut canadien de

Institute, Communications Canada conservation, Communications Canada

Edited by Sous la direction de

David W. Grattan David W. Grattan

Canadian Conservation Institute Institut canadien de conservation

Ottawa, Canada Ottawa, Canada

1993 t993
Canadian Cataloguing in Publication Data Donn6es de catalogage avant publication
(Canada)

Symposium '91 - Saving the Twentieth Century Symposium 91 - Sauvegarder le XXe sidcle
(1991 : Ottawa, Ont.) (1991 : Ottawa, Ont.)

Saving the twentieth century : the conservation Saving the twentieth century : the conservation
of modern materials : proceedings of a confer- of modern materials : proceedings of a confer-
ence, Otlawa,Canada, 15 to 20 September ence, Ottawa,Canada, l5 to 20 September
1991 :Sauvegarder le XX" sidcle : la conser- l99l : Sauvegarder le XX" sidcle : la conser-
vation des matdriaux modernes : les actes de vation des mat6riaux modemes : les actes de
la conf6rence, Ottawa (Canada) du 15 au la conf6rence, Ottawa (Canada) du 15 au
20 septembre 1991 20 septembre 1991

Prefatory material and abstracts in English Texte pr6liminaire et r6sum6s en anglais et

and French. en frangais.
Includes bibliographical references. Comprend des r6f6rences bibliographiques.
rsBN 0-660-57854-9 ISBN 0-660-57854-9
DSS cat. no. NM95-5812-1992 No de cat. MAS NM95-5812-1992

l. Museum conservation methods Canada l. Mus6es M6thodes de conservation

Congresses. 2. Museum techniques
- Canada- Canada- -Congrds. 2. Mus6ologie -
Canada
Congresses. 3. Materials
-
Conservation Congrds. 3. Mat6riaux
-
Conservation et
-and restoration Congresses. - I. Grattan, -restauration Congrds. I.- Grattan, David W.
-
David W. II. Canadian Conservation Institute.
-
IL lnstitut canadien de conservation. III. Tife.
lll. Title. IV. Title: Sauvesarder le XXe siecle. IV. Titre : Sauvegarder le XXe sidcle.

AMl4l.S28 1992 069'.53'0971 AM141.S28 1992 069',.53'0971

c92-099400-0F. c93-099400-0F
Proceedings published subsequent to Ce document a 6t6 publi6 d la suite du
Symposium '91 - Saving the Twentieth Symposium 91 - Sauvegarder le XXe sidcle,
Century, organized by the Canadian Conserva- organis6 par I'Institut canadien de conserva-
tion Institute, Communications Canada, and tion, Communications Canada, et pr6sent6
held at the Skyline Hotel, Ottawa, e l'h6tel Skyline, Ottawa, du 15 au
I 5 to 20 September I 991 . 20 septembre 1991.

o Communications Canada, Ottawa, 1993.

o Communications Canada. Ottawa. 1993.

All rights reserved. No part of this publication Tous droits r6serv6s. La reproduction d'un
may be reproduced or transmitted, in any form extrait quelconque de ce livre, par quelque
or by any means, electronic or mechanical, proc6d6 que ce soit, tant 6lectronique que
including photocopying, recording, entering m6canique, ou par photocopie, microfilm,
in an information storage and retrieval system, bande magn6tique, disque ou autre, sans
or otherwise, without prior written permission I'autorisation 6crite de l'6diteur est
ofthe publisher. interdite.

Cover Design: Sophie Georgiev. Graphisme de la couverture : Sophie Georgiev,

Canadian Conservation Institute lnstirut canadien de conservation

PRINTED IN CANADA IMPRIME AU CANADA

Available ftom: Vendu par :

Extension Services Services de diffusion externe

Canadian Conservation Institute Instinrt canadien de conservation
Communications Canada Communications Canada
Ottawa, Ontario Ottawa (Ontario)
KIA OC8 KIA OC8
Acknowledgements Remerciements

Organizing Committee Comitd organisateur

J. Cliff McCawley (chair) J. Cliff McCawley (pr6sident)

Henri Benoit Henri Benoit

Malcolm Bilz Malcolm Bilz
Marie-Claude Corbeil Marie-Claude Corbeil
Valerie Dorge Valerie Dorge
David Grattan David Grattan
Maureen MacDonald Maureen MacDonald
Janet Mason Janet Mason
Wanda McWilliams Wanda McWilliams
Charlotte Newton Charlotte Newton
Jean T6treault Jean T6treault
David Tremain David Tremain
Season Tse Season Tse

The Organizing Committee thanks the speak- Le comit6 organisateur tient d remercier les
ers, session chairs, presenters ofposters, and conf6renciers, les pr6sidents de session et tous
participants in the panel discussion for their les participants d la table ronde pour leur
conhibution to the Symposium. contribution au symposium.

Editorial Committee Comite de rtdaction

David Grattan (chair) David Grattan (pr6sident)

Valerie Dorge Valerie Dorge

Charlotte Newton Charlotte Newton
Jean T6treault Jean T6treault

Production Committee Comite de production

Deborah Robichaud (chair) Deborah Robichaud (pr6sidente)

A.P. Doming A.P. Doming

Sophie Georgiev Sophie Georgiev
David Grattan David Grattan
Sandra LaFortune Sandra LaFortune
Linda Leclerc Linda Leclerc
Publication Publication

The editor thanks the authors for their manu- Le r6dacteur en chef tient ir remercier les au-
scripts and for their kind cooperation during the teurs de leur contribution au projet ainsi que de
editorial process. The editor also acknowledges leur coop6ration. Il d6sire 6galement exprimer
the help of the following people in the prepara- sa gratitude d toutes les personnes qui lui ont
tion of this publication: pr6t6 main-forte :

Caroline Shaughnessy: Caroline Shaughnessy :

English editing R6vision des textes anglais
Andr6 La Rose: Andr6 La Rose :
French editing and translation R6vision des textes frangais et traduction
Sophie Georgiev: Sophie Georgiev :
Design and layout Graphisme et production
Sandra LaFortune and Linda Leclerc: Sandra LaForhure et Linda Leclerc :
proofreading
Editorial assistance and Aide d la r6daction et correction d'6preuves
Valerie Dorge, Charlotte Newton, and Valerie Dorge, Charlotte Newton et
Jean T6treault: Jean T6treault :
Technical review R6vision des textes techniques
Translation Bureau, Departrnent of the Bureau des traductions du
Secretary of State: Secr6tariat d'Etat:
Translation of abstracts Traduction des r6sum6s

Consultation Consultation

We acknowledge the advice and assistance Nous tenons ir remercier certaines personnes
of the following people in plaruring the qui, par leurs conseils et leur aide, ont con-
Symposium: tribu6 d la planification du symposium :

Dr. David J. Carlsson: David J. Carlsson, Ph.D. :

National Research Council Canada, Conseil national de recherches Canada,
Ottawa Ottawa
Karen Graham: Karen Graham :
Canadian War Museum, Ottawa Mus6e canadien de la guerre, Ottawa
Robson Senior: Robson Senior:
National Museum of Science and Mus6e national des sciences et de la
Technology, Ottawa technologie, Ottawa
Dr. David M. Wiles: David M. Wiles, Ph.D. :
Plastichem Consulting, Victoria, B.C. Plastichem Consulting, Victoria (C.-B.)
Introduction Introduction

The conference "Symposium'91 - Saving the La conference < Symposium 9l - Sauvegarder

Twentieth Century" took place in Ottawa in le XX" sidcle : la conservation des mat6riaux
the autumn of I 991 . Organized and hosted modernes > a eu lieu ir Ottawa au cours de
by the Canadian Conservation Institute (Com- l'automne de 1991. Symposium 91, mis sur
munications Canada), Symposium'9 I formed pied par I'Institut canadien de conservation
one of a series of intemational meetings de- (Communications Canada) qui en 6tait 6gale-
voted to specific conservation problems. ment I'hdte, s'inscrivait dans le cadre d'une
s6rie de rencontres internationales consacr6es
Each Symposium addresses topics of major d l'6tude de probldmes de conservation
importance to museums particularly to particuliers.
-
museums in Canada. Topics are selected that
have been generally overlooked and that pose Chacune de ces rencontres porte sur un
new and demanding challenges. In this way, we sujet d'importance pour tous les mus6es
hope to create a better awareness of problems,
-
surtout pour les mus6es du Canada. Et chacun
to stimulate thought, and to encourage research. de ces sujets correspond ir un domaine qui,
The theme for Symposium '91, modem materi- bien que g6n6ralement n6glig6, n'en consti-
als, meets all these criteria. tue pas moins un nouveau d6fi de taille. Les
conf6rences visent donc d sensibiliser davan-
Museums and galleries reflect society, and are tage les gens d certains probldmes, tout en
influenced by opinion and fashion. As material favorisant la r6flexion et en stimulant la recher-
culture becomes more sophisticated, museums che. Symposium 91, qui portait sur les mat6-
collect a much wider range of objects. Museum riaux modernes, respectait tous ces critdres.
curators are, therefore, assembling collections
that include new slmthetic materials, complex Les mus6es et les galeries d'art sont le reflet
electronic circuitry, sophisticated alloys, and de la soci6t6, et ils ne sont dds lors pas d I'abri
many new types of coatings. Consider the de I'opinion et des modes. Au fur et d mesure
following list: compact disc, radio, computer, qu'6volue la culture mat6rielle, les conser-
spacesuit, aircraft, plywood, and plastic doll. vateurs de ces 6tablissements en viennent ir
All are among the rapidly degrading artifacts collectionner une garnme toujours plus vari6e
that were discussed at Symposium '91. d'objets, faits d partir de nouvelles matidres syn-
th6tiques, de c ircuits 6lecfr onique s complexes
"Complex" is perhaps the word that best et d'alliages de haute technicit6, et dont le re-
describes modern objects. They are becoming couvrement comporte lui aussi de nouveaux
increasingly complex in structure, in materials mat6riaux. Les disques compacts, les radios, les
used, and especially in function. Conservators ordinateurs, les combinaisons spatiales, les
are expected to stabilize and conserve artifacts a6ronefs, les contreplaqu6s et les poup6es en
from all periods in history. However, perhaps plastique ne sont que quelques-uns des objets
by inclination and certainly by training, they qui, se d6gradant rapidement, figuraient au pro-
have concentrated on studying past technolo- grarnme de Symposium 91.
gies rather than current ones. Although modern
materials are generally perceived as being Le mot < complexe ) est sans doute celui qui
more stable than traditional substances, permet le mieux de d6crire les objets modemes,
nothing could be further from the truth. Many qui deviennent en effet toujours plus complexes
twentieth-century artifacts are very unstable, quant ir leur structure, aux mat6riaux qui les
and some have problems that are new to conser- composent et, surtout, ir leur fonction. On
vators, such as radioactivity, polymer degrada- s'attend donc d ce que les restaurateurs arrivent
tion, and aluminum corrosion. d stabiliser et d conserver des objets de toutes
les 6poques de I'histoire. Or, peut-€tre par
Conservators are having to cope with these penchant naturel et sans doute en raison de leur
objects using their limited training in materials formation, les restaurateurs se sont consacr6s
science and an ethical code conceived with davantage d l'6tude des techniques du pass6
more traditional artifacts in mind. plut6t que celles du monde contemporain. Et
m6me si I'on a g6n6ralement tendance d croire
The degradation ofrecently acquired artifacts is que les mat6riaux modemes sont plus stables
forcing curators to review collecting policies. que les anciens, rien n'est moins s0r. Nombre
Should museums acquire examples of modem d'objets du XXe sidcle sont en effet fort ins-
artifacts after extensive use has accelerated the tables, et posent aux restaurateurs des probld-
degradation process? Should more emphasis be mes in6dits, qui tiennent entre autres d la
placed on documentation in order to preserve radioactivit6, d la d6gradation des polymdres et
information about the artifact if the artifact ir la corrosion de I'aluminium. Lorsqu'il ont d
itself cannot be preserved? How are we to r6soudre de tels probldmes, les restaurateurs
preserve and maintain this rapidly developing ne disposent donc que d'une formation rudi-
material culture? mentaire dans le domaine de la science des
mat6riaux, et se voient contraints d'appliquer
The aim of Symposium '91 was to approach des rdgles d'6thique qui ont 6t6 congues pour
these issues over a broad fiont. It involved des objets beaucoup plus anciens.
experts in science and technology, both from
within the museum field and outside, and in- La ddgradation des objets r6cemment acquis
cluded many museum professionals. We hoped force les conseryateurs de mus6es d revoir les
to outline the cunent state of knowledge and politiques qui pr6sident d l'6tablissement des
to set future directions for research. As the collections. Les mus6es devraient-ils acqu6rir
following pages show, these goals were accom- des moddles d'objets contemporains qui ont 6t6
plished. As for the conservation of modem grandement utilis6s et dont le processus de
materials, we have only just begun.... d6gradation est d'autant plus avanc6? Vaut-il
mieux privil6gier la documentation, la conser-
Editorial Committee vation de renseignements sur les objets, si
ces derniers ne peuvent pas 6tre conserv6s?
Comment r6ussira-t-on ir conserver les moddles
de cette culture mat6rielle qui 6volue si rapide-
ment, et d maintenir ces collections d'objets?

Ces questions devaient 6tre envisag6es, lors du

Symposium 91, dans le cadre d'une perspective
plus large. Aussi a-t-on fait appel tant d des
sp6cialistes des sciences et de la technique wu-
vrant d I'int6rieur ou d I'ext6rieur des mus6es
qu'aux professionnels de ces 6tablissements.
La conference visait d foumir un apergu des
connaissances acfuelles dans ce domaine, et d
d6cider des orientations que prendra la recher-
che dans le secteur. Les pages suivantes mon-
trent, de fagon 6loquente, que ces objectifs ont
6t6 atteints. Et, pour ce qui est de la conserva-
tion des mat6riaux modemes proprement dite,
force est de constater qu'elle n'en est qu'ir ses
d6buts...

Comite de r6daction
Table of Contents / Table des matiires

1. Modern Materials in collections / utilisation des mat6riaux modernes

au sein des collections

Les mat6riaux modemes au Mus6e de la civilisation d Qu6bec : Sylvie Marcil

un d6fi pass6, pr6sent et futur

Modern Materials in the Collection of the Musie t0

de la civilisation d Qutbec (abstract)

The condition Survey of Sound Recordings at the National Jan Michaels l3

Library of Canada: Implications for Conservation

Les enregistrements sonores d la Bibliothdque nationale du 21

Canada : une dtude d'etut et son incidence sur la conservation htsumd)

Plastics in the Science Museum, London: A Curator's View Susan Mossman 25

Les plastiques au National Museum of Science and Industry 34

le point de vue d'une conservatrice (resumi) -
Membership and Aims of the Plastics Historical Society John Morsan 5I

La composition de la Plastics Histoical Society et ses objectifs 40

(rdsumd)

2. Conservation Policies and Plans / Politiques et projets en matiire de conservation

A Joint Project on the Conservation ofPlastics by John Morgan 43

The Conservation Unit and the Plastics Historical Society

La consentation des plastiques : un projet mixte du Sentice de 50

consentation de la Museums & Galleries Commission et
de la Plastics Historical Society (rtsumt)

Conserving the Science Museum Collections Roger Price and 5l

Anne Moncrieff

La conservation des collections du National Museum 54

ofScience and Industry (rdsumd)
3. History of Technology / Histoire de la technologie

Rubber: Its History, Composition and Prospects for Conservation M.J.R. Loadman 59

Le caoutchouc : son histoire, sa composition et ses IJ

perspectives de conservation (r,!sumi)

Ardil: The Disappearing Fibre? MaryM.Brooks 8l

L'Ardil, lafibre qui disparait? (risumd) 9l

Plastics Found in Archives Alan Calmes 95

Les plastiques prdsents dans les archives (rdsumt) 102

4. Processes of Deterioration / Processus de d6gradation

Changes in Polymeric Materials with Time David M. Wiles 105

La transformation des matdriatn polymdriques avec le temps (resumt) lll

The Physical Aging of Polymeric Materials Christopher W. ll3
McGlinchey

Le vieillissement des matdriaux polymtriques (rdsumi) t20

La pr6vision du comportement d long terme de mat6riaux Jacques Lemaire 123

polymdres synth6tiques d'aprds des exp€riences de
vieillissement artifi ciel

Predicti on of the Long-Term B ehaviour of Synthetic Polyn eri c 133

Materials from Artificial Ageing Experiments (abstract)

Composition Implications of Plastic Artifacts: A Survey of R. Scott Williams 135

Additives and Their Effects on the Longevity of Plastics

La composition des objets en plastique et leur longdvitd (resume) 152

The Nature and Origin of Surface Veneer Checking in Plywood Mark D. Minor 155

Une dtude prdliminaire de I'origine et de la nature des t&

craquelures qui se produisent dans le placage des contreplaques (rdsumd)
5. Case Studies and Specific Problems with Materials/
Etudes de cas et problimes particuliers pos6s par les mat6riaux

Deterioration of Cellulose Nitrate Sculptures Made by Gabo Michele Derrick, 169

and Pevsner Dusan Stulik and
Eugena Ordonez

La dbgradation de sculptures en nitrate de cellulose exdcuttes 181

par Gabo et Pevsner (resumb)

A.W. McCurdy's Developing Tank: Degradation of an Early Plastic Robert D. Stevenson 183

La cuve de dbveloppement d'A.W. McCurdy : un exemple 187

de la ddgradation des premiers plastiques (resumt)

Degradation of Polyurethanes in 2Oth-Century Museum Textiles Nancy Kerr and 189

Jane Batcheller

La ddgradation des polyarithannes prdsents dans les textiles du 204

Xf sidcle conservds dans des musrles lrtsumi)

Treafing Early Regenerated Cellulose Textiles: Two Case Histories Emma Telford 207

Le traitement de textiles anciens en cellulose rdgindrte : 2ll

deux dtudes de cas (resumd)

Treatnent of 2fth-Century Rubberized Multimedia Costume: Clare Stoughton- 213

Conservation of a Mary Quant Raincoat(ca.1967) Harris

Le traitement d'un costume multimedia caoutchouti du Xf siicle : 220

la conservation d'un impermeable Mary Quant datant de
1967 environ ftisumd)

Spacesuits: NASA's Dream Conservator's Nightmare Mary T. Baker and 223

- Ed McManus

Les combinaisons spatiales : un r€ve pour la NASA, mais un 229

cauchemar pour les spdcialistes de la restauration (rtsumd)

Conservation of Paintings on Delaminated Plywood Supports Donald C. Williams 231

and Ann Creager

La conservation de peintures sur supports de contreplaqui 241

ddcollAs (rdsumb)

Chicken Bones and Cardboard: The Conservation of a Anton Rajer and 243
Collection of Eugene Von Bruenchenhein's Art Emil L. Donoval

Les os de poulet et le carton ou la conservation d'une collection 247

d' oeuvres d' Eugene Von B ruenchenhein (resume)
Ensuring a Future for Our Present High-Tech Past: Lessons from Jon Ecklund 249
the ENIAC for the Conservation of Major Electronic Technology and Beth Richwine

Garantir I'avenir des vestiges de notre haute technologie : 259

les leqons d tirer de I'expbrience de I'ENIAC pour assurer
la conseruation des grandes techniques de I'electronique (rdsumi)

Modern Metals in Museum Collections Lorna R. Green and 261

David Thickett

Les mdtaux modernes dans les collections de mus,le (rtsumil 270

Finishes on Aluminium A Conservation Perspective Chris Adams and 273

- David Hallam

L'application de rev€tements sur I'aluminium : un point de vue 284

axi sur la consewation (rdsumd)

Radiation Hazards in Museum Aircraft John Ashton 287

Les risques d'exposition aux radiations provenant d'aeronefs 295

exposds dans les mtndes (rdsumd)

Conservation of One of Alexander Calder's Largest Mobiles Albert Marshall and 301
Shelley Sturman

La restauration d'un des plus grands mobiles d'Alexander Calder (rdsumd) 305

Conservation of Weathering Steel Sculpture John Scott 307

La conseryation de sculptures en acier (rdsumd) 318

6. Testing and Development of Conservation Processes /

Elaboration et mise i I'essai des m6thodes de conservation

An Evaluation of Eleven Adhesives for Repairing Don Sale 325

Poly(methyl methacrylate) Obj ects and Sculpture

L'|valuation de I I adhtsifs en vue de leur utilisation pour la 337

rtparation d'obje* et de sculptures en poly(mtthacrylate
de mLthyle) (rtsumi)

Labelling Plastic Artefacts Julia Fenn 341

La pose des numdros d'enregistrement sur les objets en plastique (rtsumt) 349
Degradation Rates for Some Historic Polymers and the Potential David W. Grattan 351
of Various Conservation Measures for Minimizins Oxidative
Degradation

Le rythme de degradation de certains polymdres d'importance historique 360

et les diverses mesures de consenation qui permettraient de rdduire au
minimum la ddgradation oxydative (risume)

A Field Trial for the Use of Ageless in the Preservation of Rubber Yvonne Shashoua 363
in Museum Collections and Scott Thomsen

L'essai pratique de I'Ageless pour la conservation du caoutchouc dans 371

les collections de musde (resume)

La mise au point d'un traitement cathodique de stabilisation Christian Degrigny 373

de vestiges a6ronautiques immerg6s en alliages d'aluminium

An Electrolytic Treatment for Stabilizing Submerged Aluminium 379

Al loy Aircraft Fragments (abstract)

The Discoloration of Acrylic Dispersion Media James Hamm, 381

Ben Gavett, Mark
Golden, Jim Hayes,
Charles Kelly, John
Messinger, Margaret
Contompasis and
Bruce Suffield

La dbcoloration des medias de dispersion acrvliques (rtsum6) 392

7. Methods of Analysis and Identification / M6thodes d'analyse et d'identification

Practical Pitfalls in the ldentification of Plastics Helen C. Coxon 395

Les dfficultes pratiques liees d I'identification des plastiques (rdsume) 407

The Identification and Characterization of Acrvlic Emulsion Carol Stringari 411

Paint Media and Ellen Pratt

L' identification et la caractdrisation des peintures-imulsions 438

acryliques (rtsumd)
Modern Materials in Collections
Utilisation des mat6riaux modernes au sein des collections
Les mat6riaux modernes au Mus6e de la civilisation ir Qu6bec :
un d6fi pass6o pr6sent et futur

Sylvie Marcil

Muste de la civilisation
Qudbec (Quebec)
Canada

R6sum6 Le Muste ne dispose que de ressources limitdes,

si I'on prend en compte son calendrier d'exposi-
Les mattriaux modernes prennent de plus en tions chargt. Aussi son equipe de conseryation-
plus d'importance dans les collections du Muste restauration doit-elle s'en tenir d la mise en
de la civilisation d Qubbec, qui presente, sous dif- application de mesures prtfoentives et d des inter-
firents thimes, un patrimoine souvent assez ventions minimales. Et il lui tarde de mettre sur
rdcent. Lavaridtd de ces matdriata ne cesse de pied un programme prdventif global pour
croltre caoutchoucs, plastiques de tous genres, ces matdriaux.
-
laques et vernis, metuux et leur composition
-,
nous est souvent inconnue. Ils se retrouvent dans
presque toutes les collections alimentation, Introduction
-
accessoires de mode et de beautd, ameublement,
jeur etjouets, outils et dquipement, transports et En reprenant, au terme du pr6sent symposium,
communications dans des proportions vari6es.
- le sujet de la communication d'introduction,
ir savoir << Les mat6riaux modernes dans le
Qu'ils entrent dans la composition d'objets ex- mus6e >, nous voulons en quelque sorte boucler
posts ou mis en rdserve, ces matdriaux complexes
la boucle. Nous n'aurons pas n6cessairement
posent de nouveaux defis d I'equipe de conserva-
fait le tour de la question mais, en effectuant
tion-restauration, compte tenu du mandat qui a
ainsi un retour au point de d6part, nous serons
ttd confie au Musde par le ministdre des Affaires
culturelles du Qudbec. La prdsente communica- mieux en mesure de d6terminer si cette rencon-
tion Jitit prtcisiment ttat de ces ddfis. tre nous a effectivement permis d'avancer
un peu.
Dans le cas d'expositions, ces d6fis tiennent sur-
tout d des questions de responsabilitd Sans offrir de solutions ni tirer de conclusions
partagee entre deux seruices - qui est
et d'auitude, de des communications pr6c6dentes, le texte qui
-
m€me qu'au statut des objets eux-memes et aux suit pr6sente tout simplement une r6alit6, celle,
conditions ambiantes paniculidres qu' ils exigent. trds commwre, d'une conservatrice-restaura-
Dans le cas des collections, ils se posent, pour trice d'un mus6e qui collectionne l'histoire
ainsi dire, d toutes les 2tapes : depuis I'identifica- en devenir, dans divers secteurs, et qui utilise
tion et la numdrotation des objets ou l'tvaluation diff6rentes th6matiques pour f illustrer. Le
de leur ttat, jusqu'd leur nettoyage, leur traite- contenu t6moigne des d6frs que pr6sente la
ment et leur entreposage, et en passant par les conservation des mat6riaux modemes, pr6-
divers contr6les eux-m€mes tout au-
- des objets
tant que des conditions ambiantes des rtserttes
sents dans un nombre croissant d'obiets de

q ui do iv ent s' effec t u er.

- la collection.
Au fil du texte sont ins6r6es plusieurs photo- services du Mus6e et plus particulidrement
graphies : les figures I d 5 offrent quelques
-
celui des expositions ainsi que quelques
exemples de mat6riaux modernes en exposi-
-,
agences ext6rieures et, d I'occasion, le grand
tion; les figures 6 d 10 sont des exemples de public. Les interventions se font surtout ir la
la d6t6rioration de mat6riaux modemes et les r6serve, dans un atelier 6quip6 ir cette fin.
figures 1l e 13 offrent quelques exemples d'in-
terventions. Il est d noter que si certains objets Les mat6riaux modernes
ont le statut d'objets fabriquds ou proviennent au Mus6e de la civilisation
de collections priv6es, il s'en trouve aussi d'au-
tres qui ne constituent encore que des acces- D6chet hier, tr6sor aujourd'hui, d6chet au-
soires, mais qui pourraient bien 6ventuellement jourd'hui, tr6sor demain. I1 ne fait aucun
devenir des objets de collection. Tous ces doute que les objets modernes, faits de mat6-
mat6riaux se c6toient dans les expositions. riaux modernes, se retrouvent d6jd dans nos
expositions, et qu'ils entrent en grand nombre
Le Mus6e de la civilisation dans nos collections. Leur vari6t6 ne cesse de
croitre : caoutchoucs et plastiques de tous gen-
Le Mus6e de la civilisation est situ6 d Qu6bec res, laques, peintures et vernis, m6taux, cartons,
m€me, au bord du majestueux fleuve Saint- contreplaqu6s, textiles, etc. Ils se retrouvent
Laurent. Il loge dans un immeuble trds r6cent, dans presque tous les secteurs I'alimenta-
qui a ouvert ses portes en 1988. L'6quipe du tion, les accessoires de mode et- de beaut6,
Mus6e est jeune, dynamique et ambitieuse. Elle l'ameublement, I'art populaire, les jeux et les
r6ussit ir attirer quelque 800 000 visiteurs par jouets, les outils et l'6quipement, les transports
ann6e, en faisant en sorte que le Mus6e soit et les communications, etc.
ouvert et accessible d tous.
Ces mat6riaux sont trds complexes, h6t6rogdnes
La r6serve du Mus6e, oi sont entrepos6es les et en constante 6volution, et nous devons donc
collections, est situ6e dans la ville de Vanier, d d6sormais apprendre ir les conserver et 2r les res-
environ 10 km du Mus6e. Une importante par- taurer, au m6me titre que les autres. Il s'agit ld
tie du Service des collections y travaille, dont d'un d6fi de taille pour tous les conservateurs-
les 6quipes charg6es du registrariat, du cata- restaurateurs, mais plus particulidrement pour
logage, de la recherche et des prCts d'objets, ceux qui sont isol6s dans des mus6es oi les
ainsi que celle de la conservation-restauration. collections sont assez diverses et nombreuses.

La conservation-restauration I. Les expositions

L'6quipe de conservation-restauration se Outre les expositions itin6rantes, le Mus6e

compose d'une technicienne en conservation pr6sente une vingtaine d'expositions par ann6e
pr6ventive, Elisabeth Forest, et, depuis mai dans les l0 salles du bdtiment principql, qui
I 989, d'une conservatrice-restauratrice. Des occupent approximativement 5 700 m", et
techniciens en restauration viennent. au besoin, dans deux autres bdtiments, ir savoir la maison
se greffer d cette 6quipe. Le mandat que le Chevalier et I'entrepdt Thibodeau*. Une propor-
ministdre des Affaires culturelles a confi6 au tion marqu6e des objets expos6s se compose de
Mus6e n'est pas un mandat de recherche. Le mat6riaux modemes. Les deux tiers environ des
Mus6e est plutdt charg6 des fonctions pr6ven- objets de chaque exposition sont emprunt6s ou
tives, tandis que le Cenhe de conservation du acquis ir cette fin.
Qu6bec (CCQ) continue d'assumer en grande
partie la fonction curative. Les expositions ont une influence consid6rable
sur le travail de conservation-restauration et
Compte tenu de ce mandat et du roulement sur le d6veloppement des collections, qui com-
des expositions au Mus6e, le travail de conser- portent de plus en plus d'objets cr66s d partir
vation-restauration s'effectue plutdt ir un niveau de mat6riaux modernes. Ces mat6riaux sont
de consultation, de pr6vention ou d'interven- en effet trds pr6sents, car les thdmes qui sont
tion minimale. L'6quipe dessert tous les

4
Figure I : Autopsie d'un sacvert (Photo : Pierre
Soulard, Archives du Mus,le de la civilisation, 348-
PH-6). Figure 3 : Jeux (Photo : Piene Soulard, Archives du
Muste de la civilisation, 341-PH-3).

t:.' @
t;ht::';'
e;;.,*.:|3

Figure 2 : Ephemdre (Photo : Piene Soulard,

Archives du Musde de la civilisation. 330-PH-21).
Figure 4 : Quebec sur glace (Photo : Pierre Soulard,
Archives du Muste de la civilisation, 3 l9-PH-20).
abord6s lors de telles expositions I'alimenta-
tion, la beaut6 et I'environnement,-pour n'en Tous ces objets n'aboutissent pas n6cessaire-
nommer que quelques-uns 6tablissent une ment dans nos collections, mais il faut quand
comparaison entre le pass6 -et le pr6sent, tout en m€me assurer leur conservation pendant leur
explorant les tendances futures. s6jour au Mus6e. Et c'est de ld que proviennent
les diffrcult6s, qui tiennent surtout d des ques-
Cela met donc en valeur un patrimoine sou- tions de responsabilit6 et d'attitude, ou au statut
vent frds r6cent : depuis les ensembles de cos- des objets eux-m6mes. Le Service des expo-
m6tique ou un canot de fibre de verre d'une des sitions est, avec celui des collections, respon-
demidres courses sur glace jusqu'aux systdmes sable des objets. Vu la nature consultative de sa
audiovisuels qui ne sont pas encore sur le fonction, la conservation-restauration n' inter-
march6 chez nous. vient qu'au besoin, d diff6rents moments de la
planification. Un objet r6cent quel qu'en soit
le propri6taire ou la personne - responsable ne
-
 abim6 pendant I'installation s'il 6tait coll6,
- La d6marca-
6pingl6 ou peintur6, par exemple.
tion entre I'artefact et I'accessoire n'est d'ail-
leurs pas toujours 6vidente. Il est difficile pour
les conservateurs d'6tablir la valeur d'une ac-
quisition en comparant les co0ts de gestion
et d'enheposage de I'artefact ir ceux de I'objet
neufet en prenant en compte les axes de
d6veloppement de la collection. Mais le d6fi
ne s'arr6te pas ld. Lorsque les objets ont int6gr6
les collections, ce sont d'autres pr6occupations
qui entrent en jeu.

II. Les collections

Actuellement, prds de 65 000 objets sont enhe-
pos6s d la r6serve du Mus6e, dans un espace
d'un peu plus de 3 000 m'. Il est difficile
Figure 5 : Souffrir pour €tre belle (Photo : Piene
d'6tablir le nombre d'objets qui comportent
Soulard, Archives du Musee de la civilisation,
des mat6riaux modernes, car leurs fiches ne
303-PH-21).
font pas toujours 6tat des mat6riaux qui les
commande pas autant de respect qu'un bronze composent, mais ils constituent, de toute 6vi-
de I'Antiquit6, par exemple. De plus, certaines dence, un pourcentage non n6gligeable de cet
pidces de m6me genre ont tantdt le statut d'ob- ensemble. On sait d6jd que les plastiques et les
jets de mus6e, ou d'artefacts, tantOt celui d'ac- caoutchoucs font partie, ir eux seuls, de plus
cessoires, et ne regoivent donc pas toujours de 2 000 de ces objets. Ce nombre augmente
la m6me attention, ce qui porte parfois d rapidement au fur et d mesure que progresse
confusion. la validation des fiches d'objets aux fins d'int6-
gration d la base de donn6es sur vid6odisque,
En ce qui conceme les conditions d'exposi- un systdme dont I'efficacit6 d6pend de la pr6-
tion (la lumidre, la temperature et I'humidit6 cision et de I'exhaustivit6 de I'information qui
relative), on opte plut6t pour la prudence. Le y est vers6e.
Mus6e possdde d'ailleurs un systdme 6volu6 de
contrdle, mais pour les salles d'exposition du Quel que soit le mat6riau, il faut donc I'identi-
bdtiment principal seulement. Les facteurs qui fier, num6roter I'objet, en d6terminer l'6tat, le
posent la plus grande difficult6 sont les pol- nettoyer, le traiter au besoin, I'entreposer dans
luants atrnosph6riques et la compatibilit6 entre les meilleures conditions possibles et, finale-
les mat6riaux entre ceux qui composent l'ob- ment, le contrdler. Beaucoup plus facile d dire
jet lui-m6me, -de m6me qu'entre ces derniers et qu'd faire, surtout lorsqu'il s'agit de mat6riaux
les mat6riaux de construction environnants. De modemes. Pour y arriver, il faut avoir une
plus, il arrive souvent que I'on ne connaisse ni connaissance approfondie de la nature des
la composition des objets ni celle des objets mat6riaux et disposer de bons outils de travail.
environnants, et on ne peut donc pr6voir les La tdche est 6norme et les priorit6s, difliciles
r6actions qu'ils auront dans un montage of ils d 6tablir.
se c6toient.
1. L'identification
Il est par ailleurs possible que, au terme d'une Dans un tel contexte, I'identification est le
exposition, I'objet passe du statut d'accessoire premier probldme qui se pose. L'absence
d celui d'artefact. On essaie toutefois de d6ter- d'identification des mat6riaux rend difficile
miner son statut futur avant le montage, pour le catalogage et les autres 6tapes de la mise en
s'assurer qu'il sera, au besoin, ad6quatement r6serve, de m6me que la r6daction des vignettes
restaur6 et pour 6viter qu'il ne soit inutilement accompagnant les objets.

6
L'identification exige une formation en science
des mat6riaux et un bon sens de l'observation.
MOme si on reconnait la pr6sence d'un
mat6riau moderne, il faudra faire appel ir un
conservateur-restaurateur, qui est souvent le
seul d poss6der la formation qui permettra de
I'identifier. Ce sp6cialiste ne dispose n6an-
moins pas toujours de tous les outils dont il a
besoin pour trouver les r6ponses, et il dewa
donc trouver des moyens simples et rapides
pour y arriver : tests. 6quipements. ouvrages
de r6f6rence, etc. Si on veut que le catalogage '#w
suive le rythme des acquisitions, on ne peut, en
effet, se permettre d'effecfuer une recherche Figure 6 : Ornement de Nodl en forme de
pouss6e pour chaque objet. Aussi une forma- violon (Musde de la civilisation, 89-1679). La laque
tion plus approfondie et continue facilite-t-elle prdsente de grandes fissures, qui laissent paraitre le
verre soffid sous-jacent. (Photo : Elisabeth Forest)
le travail.

2. La num6rotation
Lorsque le mat6riau a 6t6 identifi6, il se peut
que I'on ne puisse pas num6roter I'objet en em-
ployant la m6thode traditionnelle de I'encre et
du vernis qui risquerait de faire fondre le
mat6riau. -C'est le cas de certains plastiques
les nitrates cellulosiques, par exemple -
caoutchoucs, surfaces vernies et m6taux -, laqu6s.
Il est donc bon de faire part de cette mise en
garde aux personnes qui sont charg6es de
cette 6tape, s'ils n'ont pas d6jd regu la
formation pertinente.
Figure 7 : Poupde (Musie de la civilisation, 79-138)
Dans ces cas, nous utilisons parfois des constitude d'un mattriau composite, possiblement
6tiquettes de polyester dactylographi6es, qui une pdte de bois et recouverte d'une couche de
servent habituellement d la num6rotation des plastique non identifie. Ce recouvrement craque
textiles. Nous les fixons ir I'objet avec un et se souldve. (Photo : Elisabeth Forest)
adh6sifsoluble d I'eau en I'occurrence, du
Cellofas - plutdt mal d cer-
mais il adhdre
-, Nous n'avons pas encore, de
taines surfaces.
toute 6vidence, trouv6 la m6thode id6ale.

3.L'6tat de I'objet
Contrairement d ce qu'on aurait pu imaginer, la
plupart des mat6riaux modemes mis en r6serve
au Mus6e paraissent stables et ne montrent pas
de signes de d6t6rioration avanc6e. Une 6va-
luatign trds sommaire a permis d'6tablir que la
majorit6 des mat6riaux serait en trds bon 6tat.
Ceci r6sulte probablement du fait que plusieurs
acquisitions d'objets modernes sont r6centes, et Figure 8 : Poupie (Musde de la civilisation,
qu'on a choisi, dans la mesure du possible, des 88-6495) de caoutchouc rembourr,le. Le caoutchouc
obiets en bon 6tat. sous tension se fissure et laisse paraitre la bounure,
probablenent de laine. (Photo : Elisabeth Forest)
 4. Le nettoyage
Le nettoyage demeure I'op6ration la plus
courante d laquelle sera soumise I'objet. Le
temps, les ressources financidres et les donn6es
sont souvent trop limit6s pour pousser plus loin
I'intervention. On s'en tient donc d de simples
tests, ir I'aide de m6thodes et produits usuels.
Il est certain que cette 6tape est plus facile si le
materiau d 6t6 identifie. En cas d'effet negatif,
on se limite au d6poussi6rage.

Lorsque la m6thode de nettoyage a 6t6 d6fi-

Figure 9 : Jeu de parchesi (Musee de la civilisation, nie diff6rentes questions se posent. Jusqu'or)
89-1655). Le papier vemi est ddchird, ainsi que les doit-on aller? Faut-il conserver tous les
bandes de tissu du centre et des c6tds. (Photo : 6l6ments? Doit-on disposer syst6matiquement
Llsabeth I. ore.st )
des contenus d caractdre pharmaceutique ou
alimentaire, par exemple? Et comment doit-on
s'y prendre?

.\rf

Figure l0 : Parapluie (Collection pivde). Les cou-

leurs du tissu ont pdli et le bouton de plastique est ,l/
compbtement Jis surd. (P hoto : Elis abeth Forest) ,,,
't1,"
'\

Les manifestations de d6t6rioration, lorsqu'il

s'en trouve, sont caract6ristiques : les plas-
tiques et les caoutchoucs craquent, durcissent,
jaunissent, suintent; les laques, les peinfures et Figure I 1 : Rasoir (Collection privee). La plupaft
des interventions se limitent d un nettoyage de
les vernis craquent, se souldvent, s'effiitent; les
sudace lorsque le matdriau rdagit bien.
m6taux se cassent, se corrodent, et leurs pla-
cages tombent; les cartons se d6chirent, plient,
se d6forment etjaunisssent; les textiles se
5. Le traitement
tachent et se d6chirent; les contreplaqu6s fen- Un simple nettoyage n'est parfois pas
dent et se d6laminent, etc. Certains mat6riaux suffisant pour assurer la stabilit6 d'un objet.
sont plus touch6s les caoutchoucs et les Si on connait la cause de la d6t6rioration, on
-
plastiques mous, certiaines laques et les cafions
peut parfois intervenir. Il est toutefois difficile
de formuler un traitement et d'en 6valuer les
notamment. L'utilisation qui a 6t6 faite de
l'objet ir I'origine, combin6e d la nature du cons6quences sans avoir de solides connais-
mat6riau, est responsable en grande partie de sances du mat6riau. Les ouwages de r6f6rence,
cette d6terioration. rares, 6pars et souvent trop th6oriques, ne sont
pas d'un grand secours. On se r6fdre plutdt d
des colldgues plus exp6riment6s. On doit
aussi affronter des questions d'6th!que, car il
est difficile d'etablir un principe directeur
d'intervention qui tienne compte de l'intention
du fabricant et des autres consid6rations pour
la conservation.

Comme pour le nettoyage, les quelques inter-

ventions curatives qui s'effectuent au Mus6e
sont limit6es, compte tenu du manque de temps
et de connaissances approfondies. Les r6sultats
sont de ce fait in6gaux. La diversit6 des mat6-
riaux ne nous permet pas de passer en rerv'ue
tous les geffes de traitements qui ont 6t6 mis d
I'essai, mais nous en retiendrons tout de m€me
quelques-uns.

Figure 13 : Le traitement, fait au Centre de

conseryation du Qutbec (CCQ), a permis de
stabiliser la capote de cuirette < American Leather >
avec Ltn entoilage de mousseline de soie colle avec
du BEVA et de reiquilibrer les roues en adaptant
des rubes de ndoprine. (Photo ; Michel Elie. CCQt

Les caoutchoucs wlcanis6s, mousse ou

latex demeurent- non trait6s pour I'instant.
On les- protdge avec du Armor-All, quand
ils sont en bon 6tat, mais on remplace aussi cer-
taines parties manquantes avec des mat6riaux
6quivalents modemes, ce qui est parfois plus
rapide et plus 6conomique, et qui permet mieux
de respecter I'original et son utilisation. Il
faudra voir comment de telles interventions
resisteront aux ann6es.

6. L'entreposage
Determiner les meilleures conditions d'en-
treposage possibles n'est pas toujours facile
Figure 12 : Carrosse (Musee de la civilisation,
89-1655). La capote Atuitfendue et trois des lorsqu'il s'agit de mat6riaux modernes. Bien
cerceaux de caoutchouc manquaient aux jantes. que les rdgles g6n6rales d'entreposage s'ap-
pliquent; il faut d6terminer les besoins par-
ticuliers des mat6riaux en ce qui d trait d
Les meubles, les valises ou les jouets, par exem-
ple, ont souvent un recouvrement qui est fait de I'humidit6, d la temp6rature et ir la lumidre, et d
leur protection contre les polluants atmosph6ri-
textile enduit, de plastique (du vinyle, notam-
ques. L'identification des mat6riaux est 6gale-
ment) ou de carton lamin6 et parfois vemi. On
ment primordiale pour 6tablir leur compatibilit6
r6ussira d refermer les fentes dans les textiles d
avec les autres mat6riaux environnants.
I'aide d'une doublure de soie de polyester col-
l6e avec du BEVA, d faire adh6rer le vinyle au
On a d6jd beaucoup am6lior6 I'entreposage
bois avec du Rhoplex AC-33 et i consolider
g6n6ral au Mus6e, et on continue de le faire,
des cartons d6lamin6s avec du Cellofas.
mais on n'a toujours pas de systdme de contr6le
environnemental adapt6 aux collections, et on de I'ICOM sur les mat6riaux modernes est
ne s'est pas encore pench6 sur les besoins par- une excellente initiative et un grand pas dans la
ticuliers des mat6riaux. Le fait que les objets bonne direction. Il regroupe une foule de ren-
soient souvent compos6s de plusieurs mat6- seignements pertinents. Je suis certaine que le
riaux vient compliquer les choses. L'espace symposium aura 6galement 6t6 profitable. Il
diminue d6jd d rue d'ail avec les acquisitions, s'agit < simplement > de continuer dans cette
surlout quand on doit entreposer s6par6ment direction. Maintenant que les probldmes et
chacun des el6ments de I'objet (un jeu en boite, leurs causes sont mieux connus, les conser-
par exemple, ori l'on retrouverait des pions en vateurs-restaurateurs isoles misent davantage
m6tal et en plastique dans un contenant de sur leurs collegues pour pousser plus loin leur
papier et de carton). Dewait-on enfreposer recherche, et pour publier les r6sultats de leurs
ces 6l6ments par mat6riau, si les mat6riaux travaux, autant les bons que les mauvais.
constituants sont incompatibles? La question
demeure sans r6ponse pour I'instant. Il y a beau- Remerciements
coup d'arguments contre cette option, d'ordre
logistique notamment. J'adresse mes trds sincdres remerciements d
Elisabeth Forest, qui participe activement au
7. Le contr6le travail dont j'ai fait 6tat dans la pr6sente com-
Puisque la plupart des mat6riaux modernes de munication, et qui a ex6cut6 bon nombre des
la collection du Mus6e semblent, pour l'instant, photographies qui ont 6t6 pr6sent6es. Les autres
en bon 6tat, il importe de nous concentrer sur diapositives ont 6t6 tir6es des archives du
I'aspect pr6ventif. Dans un avenir rapproch6, il Mus6e. Je tiens 6galement ir remercier les
faudra d6finir un plan d'6valuation global pour manutentionnaires Ga6tan Gigudre et Lise
ces mat6riaux et en assurer un certain contr6le. Dionne, responsables de la num6rotation et de
Cette 6tape sera facilit6e par le vid6odisque, la mise en r6serve, qui ont emball6 les objets
mais compliqu6e par le fait que ces mat6riaux qui n'ont malheureusement pas pu 6tre pr6sen-
sont dispers6s dans I'ensemble de la collection. t6s en d6monstration.
L'ail averti des conservateurs responsables
de secteurs sera d'une grande utilit6 pour les Note
rep6rer et pour 6tablir les priorit6s de traite-
ment. Encore faudra-t-il trouver des traitements xPour obtenir de plus amples renseignements
qui soient efficaces! techniques. pridre de communiquer avec :

Conclusion Sylvie Marcil

334, rue de la Tourelle
Ilest 6vident qu'on ne peut ignorer I'arriv6e Qu6bec (Qu6bec) clR lC8
continuelle de nouveaux mat6riaux au Mus6e. T6l6phone : (418) 648-1590
Pour un 6tablissement poss6dant une telle quan- T6l6copieur : (418) 529-4195
tit6 et une telle diversit6 d'objets, la conserva-
tion-restauration constitue un d6fi 6norme, et Abstract
les besoins sont proportionnels. Pour bien s'ac-
quitter de cette t6che, il faut s'attaquer globale- Modern Mqterials in the Collection of the Musie
ment au probldme, mOme si cela ne semble pas de la civilisation
6vident de prime abord.
Modern materials have become increasingly im-
Le conservateur-restaurateur ne peut 6tre un portant in the collections of the Musde de la civili-
sp6cialiste de chaque materiau qu'il rencontre. sation, which presents a thematic via,u of recent
Il history. The composition of these artifacts is often
faut qu'il puisse compl6ter sa formation par
unlmown, is becoming increasingly diverse, and
des ateliers ou des s6minaires sp6cialis6s. Il
includes such materials as rubber, plastics, lac-
faut aussi qu'il ait facilement accds ir une infor- quers and vanishes, and metals. These materials
mation regroup6e d I'int6rieur de bibliographies can be found in varying propotlions in almost
choisies, etc. Le bulletin du groupe de travail every area ofthe collection: food, fashion and

10
beauty accessories, fumiture, games and toys, attihtde as much a,s to the status of the artifacts
tools and equipment, transportation, and contmu- themselves and to the particular imbient iondi-
nications. tions thqt require. In the case ofcollections, these
challenges exist at all stages,for example, identi-
Whether one considers the composition of dis-
fiing and numbering the objects; evaluating their
played or of stored artifocrs. th'ese ,o*pi", n,ot"- state; cleaning, treating, and storing them; and
rials present new challenges to the conservation monitoing
team. This is particularlt' true in light of the man- - of the objects themselves as much
as of the ambient storage conditions.
date given to the Museum bv the Quebec Ministry
of C u lt ura I Affa i rs. Th i s p re s e nt a t i o n ou tIin es The Muste de la civilisation has limited re-
these challenges. sources, considering its heavy exhibition sched-
ule. Hence, its team of consentators must aclhere
In the case ofexhibitions, these challenges mainly to preventive measures and use minimal interyen-
relate to issues of responsibility, which is sharecl tion. A global program of prcvention.for these
bv two departments within the museum, and of materials has yet to be established.

il
 The Condition Survey of Sound Recordings at the National Library of Canada:
Implications for Conservation

Jan Michaels

The National Library of Canada

Ottawa, Ontario
Canada

Abstract Introduction
The National Library of Canada is the deposit li- The phonograph was invented in 1877 by
braryfor, among other things, published sound Thomas Edison. Like many inventors of
recordings in Canada. In the summer of 1990, as genius, he claimed that it was by "the merest
part ofa three-year program, a condition survey accident":
of sound recordings was conducted of National
Library collections. Separate suryeys of reel-to- I was singing to the mouthpiece of a tele-
reel tapes, cylinders, LPs, 45s and 78s were done. phone, when the vibrations of the wire sent
the fine steel point into my finger. That set
Information was obtained, which will assist in me to thinking. If I could record the actions of
collection management as well as in preservation the point, and then send the point over the
and conservation planning. Shelving methods and same surface afterwards. I saw no reason whv
containers were examined for suitabilin and con- the thing would not talk.
dition. Lignin and pH testi were conducted on
containers made of paper-based materials. As for I tried the experiment, frst on a strip of tele-
the sound recordings themselves, base materials
graph paper and found that the point made an
and oxide lavers were identified. For open reel
alphabet. I shouted the word 'Halloo! Hal-
tape, rub and smell tests were canied out, as well
loo ! ' into the mouthpiece, ran the paper back
as physical examinationfor over 20 categories of
over the steel point, and heard a faint 'Halloo!
damage, including creasing, s tretching, flaking, Halloo!' in retum. I determined to make a ma-
blocking and plasticizer migration. Similar tests chine that would work accurately, and gave
for the other mediawere conducted as appropri- my assistants instructions, telling them what I
ate. For discs, groove wear also was examined. had discovered.

This paper discusses the suruqv's results and They laughed at me. I bet fifteen cigars with
implications. In the short term, our new-found in- one of my assistants, Mr. Adams, that the
formation will help in planning preservation ac- thing would work the first time without a
tivities including re-housing and copying. The break, and won them. That's the whole story.
survey also directs attention to unresolved conser- The discovery came through the pricking of a
vation issues, both technical and ethical. An unre- finger.'
solved ethical issue remains: is the aim to repair
in order to play an item one more time, or to con- From that "merest accident" was born a
serveforposterity? The paper concludes with a technology that has transformed cultural life.
discussion of the National Library's attempts to
come to gips with the problem of establishing a
discipline of sound recording conservation.

13
This transformation is despribed by an anony- major media in the collections and to do some
mous Canadian in Moogk': sound tests, too. In the end, within the allotted
time, only the cylinders, 78s, 45s, LPs and reel-
Records have done more to spread music and to-reel tapes were surveyed. No sound tests
culture to the far-flung comers of Canada, were conducted. A decision on whether or not
and indeed the world, in the first fifty years to complete the survey as originally planned
of its [sic] invention, than personal perform- has not been made. Costs will have to be
ances did in all the centuries that passed be- weighed against the potential value of any
fore. Until the advent of radio, and even for
additional information.
many years after that, the gramophone was
the main source of entefiainment for the
The surveys of the discs and magnetic tapes
settler, the hunter and happer, the isolated
were conducted using a stratified random sam-
farmer, and the little hamlets and communi-
ties in the Canadian north and far west.
ple with approximately 400 items per stratum.
This provides a95oh confidence level. (Further
The National Library, as keeper of the Cana- information on thg sampling method can be
dian published heritage, has been receiving pub- forurd in Bullock.')
lished sound recordings through legal deposit
since 1969. In order to fulfill its mandate, it has In the questionnaire we recorded the physical
purchased or received as gifts or donations tens condition ofthe sound recordings and their con-
of thousands more recordings that were made tainers. Base materials and oxide layers were
prior to legal deposit. The National Library also identified. Over 20 categories of damage to the
has thousands oftapes recorded by or for Cana- recording media were surveyed. pH of the con-
dian composers and performers. As well, there tainers was assessed using Phydrion Insta-
are talking recordings, children's tapes, Cana- check surface pH pencils and lignin content
dian National Institute for the Blind cassettes using the phloroglucinol test.
and educational kit recordings. The Library has
piano rolls, cylinder recordings, 78 rpm records Results and Implications
(78s),45 rpm records (45s),337: rpm records
This paper briefly discusses some results of the
(also known as long-playing records or LPs),
disc and reel-to-reel tape surveys. The cylinder
reel-to-reel magnetic tapes, cassettes, 8-track
survey is not discussed. Detailed analysis of all
cassette tapes, compact discs (CDs), and music
the sound surveys will be available at a later
videos. All told, the National Library of
date.
Canada's Recorded Sound Collection now
holds over 120,000 recordings of Canadiana.
Discs
In the summer of 1989, as the fust step of a
78s: The phonograph was invented in 1877.
new initiative in preservation planning, the
Twenty years later, in 1897, shellac was intro-
National Library undertook the examination of
duced as the major component in the disc. The
the nature and condition ofall its collections.
78 rpm disc was the principal commercial re-
Starting with eight categories of printed items,
cording medium of the 1930s and 1940s."
the surveys continued, evaluating microforms
McWilliams describes these shellac recordinss
and sound recordings. In 1992 the last ofthe
in the following way:
suryeys, manuscripts, will be completed. Once
all surveys are concluded and the analyses fin- The term shellac, as used in record manufac-
ished, the National Library will have a compre- ture. did not mean at that time or sub-
hensive picture of the state of its holdings and sequently
-
a disc made entirely of shellac.
the areas of greatest threat. This will greatly - a convenient way of referring
It was, rather,
facilitate preservation management. to a compound material. Shellac contained
fillers, such as limestone or slate, pigment
In some senses, our sound recording survey (usually carbon black), lubricants, such as
was the least successful of the surveys carried zinc stearate, and binders and modifiers, such
out. Ambitious plans were made to survey all as Congo gum and vinsol.'

t4
Forhrnately, most shellac discs are really Of the inner sleeves, 92oh are originals. Sleeves
very stable. McWilliams goes on to say, composed of paper constitute 55% while those
"Properly formulated shellac cures as it ages of paper with plastic liners make up 260/o. Of
a cross-polymerizatron occurs, which guaran-- the paper sleeves, 97o/o are acid paper and 93Yo
tees good long-term life."o have a surface pH of3 or 4.

However, McWilliams has reported that the Unfornrnately data indicates that all of these re-
composition of shellac compounds deteriorated cordings also are damaged: 74%o are scratched
during the World Wars when shellac supplies and99o/o are dirty. Figure 2 depicts the major
were intemrpted. As a result war-year discs forms of damage to these 3373 rpm recordings.
may be less stable.
45s: The 45 rpm record with its large-diameter
Despite their long-term stability, shellac discs spindle was f-rst introduced by RCA in the
are nonetheless quite brittle and relatively easy 1950s. Currently, these 45s are stored in their
to damage. Results indicate that all of the Na- original outer jackets. Results indicate that99%o
tional Library's collection of 78 rpm recordings of the National Library's outer jackets are acid
are damaged because of dust and dirt. Figure I paper. All of the 45s show some kind of dam-
depicts the major forms of damage to these 78s. age. Figure 3 depicts the major forms of dam-
age to these 45s.

Overall Comments
Records are published with jackets on the
DIRT
outside and sleeves as innerprotectors. Data
SCRATCHES
indicates that99o/o of all types of discs in the
collection, as a group, are in acidic recordjack-
FINGERPRINTS ets. Of the inner liners, 92o/o are originals. Of

STAINS

CHIPS

BTJBBLES
DIRT

GOUGES
SCRATCHES
EDGEFLAKE
FINGERPRINTS

BUBBLES

STAINS

Figure I Damage to 78 rpm recordings. CHIPS

WARPAGf,
LPs: The long-playing disc or LP was first
manufactured in the early 1950s and by the 20 & 60 80 100 120
1960s was ubiquitous. The National Library
collection is not stored on proper shelving or I Pecent with damage

with adequate support. Results indicate that

99%o are still stored in their original outerjack-
ets. Of these outer jackets, 98o/o are composed
of acid paper, and 8lo/o are of ligneous paper. Figure 2 Damage to 331/z rpmrecordings

l5
 drag generates enough heat that the plastic
partially melts (though not enough to de-
form), causing a microscopic flow around the
stylus into which dust can be embedded
DIRT
permanently.d

SCRATCIIES
Record player needles cause additional damage
to recordings because of the friction between
FINGERPRINTS
the record grooves and the stylus. The friction
contact is needed to reproduce the signal, but it
STAINS
is also harmful because it causes gradual defor-
B[MBLES
mation or bending of the grooves. The result is
a progressive distortion that builds up from one
playback to the next.'This is why it is recom-
clilPs
mended that a record only be played once in
WARPAGE
any 24-hour period.

20 40 60 80 100 r20 The single greatest problem in the,Breservation

of disc recordings is groove wear.'" This dam-
I Percent with damge
age is not visible except under a microscope.
Another major form of damage to discs, and the
second most common found in the survey, is
scratching. Most scratches occur as a result of
Figure 3 Damage to 45 rpm recordings. use. As playing is one of the major causes of
damage to recordings, a priority must be the
the inner sleeves, 58Yo are paper,97o of which minimizing of damage to originals. Clearly the
are acidic. In general, paper inner sleeves best solution to the potential damage caused by
should be avoided because paper breaks down dust is not to play the recordings. A second-
over time, contaminating the surface and best solution is not to use a stylus to play the re-
grooves of the recording with dust-like paper cordings. This is no longer impossible. Within
debris. Obviously acidity in the paper acceler- the last year a record player has become avail-
ates this breakdown. As well, there is a possibil- able commercially that uses a laser beam rather
ity that ligneous paper can damage recordings.' than a stylus to play the disc. This revolution-
ary piece of equipment has taken almost a dec-
Clearly, we are facing a massive re-sleeving ade to develop. The National Library recently
program. As well, the original jackets will acquired the first such laser turntable outside
require de-acidification and storage space. Japan.

Results indicate that99o/o of the discs are dirty. The ELP laser turntable uses five laser beams
As dust easily can be imbedded permanently to track the record. It can compensate for warp-
into the plastic, it is considered damage for age, discs ofvarious sizes and speeds, certain
purposes ofthe survey. This high level ofdust types ofgroove damage or variable groove
must be related to the inadequate air filtering widths. Even discs that are flakins can be
system in the Library as well as the large num- l
played successfully with this systim.l Use of
ber ofacidic sleeves andjackets. According to this turntable will assist in minimizing damage
St.-Laurent, when a record is played: to the collection.

... only a small point of the stylus is actually Fortunately the long-term prognosis for discs is
making contact with the groove walls. One quite good. Their predicted life expectancy is
and a halfgrams ofstylus pressure on such a
more than a century when stored in ordinary
minute surface translates to several tons
library environments. The Pickett and Lemcoe
ofpressure per square inch. The resulting
report states:

t6
 The actual potential storage life with respect Many tape recordings from the 1950s will be
to chemical degradation of an individual disc found on acetate-base tape. Cellulose acetate
is dependent on its exact formulation (includ- is unstable. It will eventually crumble, de-
ing both kinds and amoult of stabilizer and stroying the sound recording, or cause
extender used) and its thermal history prior to patches of the magnetic coating to fall off, de-
acquisition (including processing and mold- stroying areas ofthe recording. Acetate-base
ing). Apparently, small changes in these pa- tapes may hold up well for.years but eventu-
rameters can change the potential storage life ally they will self-destruct.' '
with respect to chemical degradation by sev-
eral decades.'' Polyester or pre-stretched poly(ethylene
terephthalate) (PET) was reserved initially for
Unfortunately evidence does exist that 45s may products exposed to severe conditions, such as
not be so permanent. Very early 45s seem to be those used in the military. Its use spread very
quite stable, but later, polystyrene was com- gradually over a period of 10 years beginning
monly used for 45s due to greater economies in around 1960. Until about 1970. both cellulosg^
manufacture over vinyl. These polystyrene triacetate and polyester tapes were produced.'o
discs are inherently unstable. They fiacture rela-
tively easily." The outer layer of the record sw- Experts project a life expectancy (i.e., the ac-
face can peel offat the area ofcontact like ceptable maintenance of mechanical properties)
old house paint.'' Thoush the National- Li- for cellulose triacetate at approximately 300
brary's coilection of 45s"does not appear to be years, and for PET at several millennia. Under
in worse condition than the other discs, this adverse storage conditions, the polyester base
may be a result of benign environmental condi- appears to be much more stable than the acetate
tions and relatively short life rather than any in- base. Degradation ofcellulose acetate at very
herent longevity in their makeup. Clearly it is high humidity (RH 80%) is extremely rapid.
important to conduct further research before the Cellulose acetate tapes stretch with dampness
45s begin to self-destruct. to produce waviness and contract with dry-
ing.'' This is corroborated in our results: the
Reel-to-Reel Tapes cellulose acetate tapes were reported to be
curled much more frequently than the polyes-
Magnetic recordings have existed since the end ter. Conversely, as the production of acetate
of the l9th century. However, the first success- tape requires the use ofplasticizers, overly dry
ful tape recordings were not demonstrated until storage conditions contribute to the loss of
1935 atthe Berlin Radio Exhibition. Harold these aggnts and the film becomes extremely
Lindsay built the first Ampex machine in 1947, brittle."'As well, cellulose acetate is decom-
successfully introducins the medium to the posed by acids and alkalis.2l This is of great
I5
United States. Low-piiced-urits became concem as results indicated that 93% ofthe
available in the mid-1950s.16 tape storage boxes are acidic. The thin base
layer has a front coating, namely the binder,
There are usually three layers in a magnetic which contains a magnetic recording material,
tape: the base or substrate onto which the which is commonly ferric oxide. Brown er al.
binder and recording material is coated; the decribe these binders:
binder, which bonds the recording material to
the base; and the recording material, which is The magnetic material is embedded in a poly-
capable of being magnetized and contains the meric binder. Common binders are based on
information recorded. polyester polyurethanes. Sometimes there is a
carbon coating on the back, which is also em-
The survey indicated that in the National bedded in a polymeric binder. The carbon
Library's collection 78o/o of the tape bases are coating dissipates electrostatic charge.
polyester, 20Yo are cellulose acetate and only
The polyester pollurethane binders used on
l% is paper. McWilliams states:
tapes are highly cross-linked materials with
complex structures. They can be expected

l7
 to have 5-10 times as many ester groups as
urethane groups. Aliphatic esters are used;
these are more susceptible to trydrolytic degra-
dation than the PET substrate." CREASES

OXIDE RUBS OTF

And in a later paper Smith et al. state:
LOSS OF OXIDE

Other materials such as lubricants, adhesives, SCRATCHES

or stabilizers are frequently added to this
TEARS
layer. Polyester polyurethane is subject to
auto catallic hydrolysis, which was expected BRITTLE
to limit the lifetime of the tape. Some types
CTJRL
of magnetic tapes contain chromium oxide
particles, which do affect the chemical degra- STICKINESS
dation mechanisms. This has recently been re-
EDGE WAVINESS
ported by researchers at IBM."
VINEGAR SMELL

In order for the magnetic tape to work properly,

the binder must adhere so frrmly to the base ma-
terial and to the iron oxide that it will resist the I Percent with daroge

stresses ofplayback and storage without craz-

ing, flaking, or peeling. It must maintain these
properties despite chemical degradation and
loss of residual solvent or plasticizer.'- Brown Figure 4 Damage to reel-to-reel tape recordings.
et al. indicate:
Sound recording experts have recommended
Thus the degradation of interest in magnetic that tape should not be stored on plastic rppls.
tapes is primarily that of the binder. On degra- Metal, unslotted reels are recofirmended.'"
dation this softens; adjacent layers oftape These solid, unslotted reels prevent uneven
may stick together or the binder may stick to exposure to the environment. Waites
the recording heads. Ultimately the tape be- explains:
comes unreadable."
Because the hub is the strongest and most sta-
The binder is more sensitive to hvdrolvsis than ble part of the reel, it is the best means of reel
PET.26 The^short estimated useful tapi lifetime support during storage. When the reel is sup-
of20 years'' is directly related to the failure of ported by the hub, there is little if any weight
the binder layer. In some cases, binder break- resting upon its flanges. This protects the
down has been discovered in items only five flanges from problems such as bending and
years old.28 nicks.

The results of the reel-to-reel tape survey were Under no circumstances should a reel be
very worrisome: for 30% of the tapes, the oxide stored resting on its flanges. Paper notes
rubs off; 23o/o cannot be played. Figure 4 indi- about stored data or other sources of
cates damage to the tape recordings. contamination should not be put in the
storage container.3l

The reels onto which tapes are wound can be

another source of damage. Results indicate that The Association for Recorded Sound Collec-
690/o of the National Library tape reels are plas-
tions (ARSC) report also recommends that
tic and 3lo/o are metal. Waites indicates that the liner notes pg stored separately from the sound
reel's "...basic function is to protect the tape recordings." The effect of acid migration on
from damage and contamination. It is often the tapes is reflected in the condition survey re-
reel itself. damaged througl mistreatment. that sults. The pH of paper notes for all items
in turn damages the tape.'Zg where the oxide rubbed off was reported to be
5 or less.

l8
Clearly the potential loss of information that is Pickett and Lemcoe recommend a surveillance
indicated by the results is distressing. Informa- procedure of inspection and test based on stabi-
tion is quickly disappearing in large amounts. lizer exhaustion as the most feasible means of
Unforfunately very little research is being car- determining the need for re-recording these
ried out to allay this worldwide problem. Gerry discs. They say:
Gibson of the Library of Congress said in 1989:
...there is evidence that detectable changes in
With the exception of work now being carried the chemical composition of the record can be
out on magnetic tape for the National Archives used to indicate incipient failure due to chemi-
[in the United States] by the National Bureau of cal deterioration, although more information
Standards (United States et al. 1986), reported is needed to develop analytical techniques
evaluation of Sony's optical Century Media data and surveillance procedures. This aspect of
by the NBS, and of the effects of flre upon the problem might be made the subject ofad-
sound and audiovisual recording supports by the ditional study.rl
French Ministdre de la culhre et de la communi-
cation for the Bibliothdque nationale (Paris) Thirty years later, analysis ofresidual products
(Fontaine 1987), virtually no independent work is being pursued in France by Fontaine with the
is going on at this time on topics directly related intention that these can be used as indicators of
to audio preservation. Further, relatively little is degradation.36
known about the presewation, conservation, or
aging problems or properties of sound rec^ord-
A remedy for another area of concem for tape
ings from directly related scientific strldy."
recordings, binder breakdown, is feasible.
In fact, the 1959 Pickett and Lemcoe study Commercial firms can, through a carefully
Preservation and Storage ofSound Record- controlled time-temperafure cycling process,
reverse the binder breakdown. Kent describes
ings,long out of print, remains the basis of
the technique:
most^o^f the conservation knowledge in the
field."
Heat allows the binder system to rebond
temporarily. The treatment is not permanent,
Conservation Issues but gives at least a 30-day time window in
which to work with the tape. Provided the
Pickett and Lemcoe point out that shellac and time-temperature cycling is done correctly,
poly(vinyl chloride) (PVC) discs deteriorate there is no measurable high-frequency loss,
in opposite ways. Whereas shellac discs will noise increase or increase in print-through.
slowly, progressively become embrittled even Treated tapes appear to play with no diffi-
in a good storage environment, PVC discs suf- culty and no apparent damage. It also appears
fer from an increasingly rapid embrittlement at that the tape couldbe re-treated later for an-
the end of their storage life. They recommend other use period.''
skilled judgment to determine when a disc has
become so embrittled that it should be re- Similar environmentally based binder rejuvena-
recorded. They indicate that: tion programs that reverse tape hydrolysis are
discussed in papeqs^by Sidney Geller'o and
Such embrittlement is often noticed by the Edward Cuddihv."
decrease in flexibility of a disc or by playback
(with good equipment) resulting in disc wear There is a critical need for basic research into
so serious that the powder will dirty a soft both the conservation science and treatment
white cloth wiped across the surface.'" ofsound recordings. The problem is that the
profession of sound recording conservator does
This is reminiscent of a short poem I used to not exist. Even the question ofproper cleaning
know, written by the Danish poet and inventor ofrecordings has not been examined by a
Piet Hein, about making perfect toast: one was conservator or scientist. Gerry Gibson laments,
supposed to put it in the toaster until it began to "What is known is based upon trial and error,
smoke and then toast it for fwo minutes less. not upo.q controlled, objective. scientific
sfudv."au

t9
The National Library is in the initial stages of select items for retention and the remainder
plans for a conservation program for its sound could be dubbed onto an archival medium.
recordings. The intention is to create two posi-
tions for sound recording conservators: one to At this time, however, there are two major prob-
concentrate on disc technologies and the other, lems with this solution of selective retention
magnetic tape technologies. Such positions will and global coplng. There is no proven archival
be challenging. Clearly a strong background in medium for sound carriers.'' In the case of
chemistry will be needed to ensure that non- discs, records may well last longer than any me-
destructive. reversible. ethical conservation dium onto which they are dubbed. Copies them-
techniques are developed. Much scientific re- selves must, over a period of years, be copied.
search will be required before treatment deci- Though tape is most often used for dubbing,
sions can be made. How exactly can we repair Smith et al. report:
a broken 78 so that it can be played again?
How can we simulate grooves so the needle can It has been found that in many cases old tapes
track the record at least one more time? Are we can be read only once. Therefore a tape test-
talking about repair for only one more play or a ing program involving random sampling
more perrnanent repair? If a one-play repair is might eventually destroy a substantial portion
of a tape library without finding bad tapes."
adopted do we immediately reverse the repair
once completed in order to preserve the
The second problem is related to the increased
original?
sophistication of sound systems. A copy made
today is usually dubbed using a computerized
As we have seen, tapes are even more problem-
system that easily filters out the original's
atic. Can anything be done when there is cata-
clicks and pops, adjusts the pitch and removes
strophic failure, that is, when the binder has
background noise. But this adjustrnent in sound
separated from the backing? Can something be
is not reversible on a copy. We still need to
done when the lubricant or plasticizer has
have the original to get as close to the per-
dried, making the tape very brittle?
former's or composer's original intent.
Perhaps the most stimulating and challenging
A major study of preservation needs for sound
debate of this Symposium is that of retention:
recordings was undertaken in 1987 by the
can we, should we, at what price? There are
intemational Association for Recorded Sound
many analogies that can be drawn between
Collections. The results and recommendations,
sound recordings and books: both are mechani-
compiling more than 400 pages, call for action:
cal, they function, they move, they are used.
Both can be damaged through use. Both carry There is a clear and urgent need to preserve
information and are usually mass produced. our surviving heritage of sound recordings.
Both are relatively easily copied: microfilm, Sound recordings have been created and dis-
photocopy, dubbing. And both speak ofa cul- seminated for nearly a century, for the most
ture and its time. That trapper and his cylinder part with little thought for their lasting signifi-
player, the isolated homesteader listening to cance to society. For the most part, recogni-
gramophone recordings in retreat from the sur- tion of their scholarly value has come about
rounding wildemess. a teenager pouring over only recently along with the creation of archi-
his or her favourite 45s. Just as bibliographic val facilities, collections, and preservation
integrity and artifactual value are important projects. ... In the meantime, very large num-
considerations when deciding on whether to re- bers of sound recordings including unique
tain a book after microfilming, similar consid- material -
are rapidly deteriorating.
-
erations are important for sound recordings.
Preservation of archival collections of sound
recordings, both in theory and practice, has
Unquestionably at least a proportion of original only recently begun to receive widespread,
published sound recordings should be kept. serious attention. As with any field of study
Criteria based on rarity, value, examples of dif- which is only now moving through the early
ferent media, major changes to carrier or sound stages of development, the field of audio
recording technologies could be used to help

20
 archiving is characterized by widely diver- si le rangement sur 4tagires et les contenants
gent practice, doubt. confusion and a myriad convenaient, et d tvaluer I'etat des obiets ainsi
of questions.*' conservtls. Des tests portant sur la lignine et le
pH des contenants d base de papier ont itt
Nearly all of the conclusions reached by the menis, et I'on a identifi| les matdriaux qui en'
planning study group reflected the lack ofco- trent dans la composition du support des enregis-
trements sonores eux-mAmes, de mAme que ses
ordinated, carefully planned research into audio
couches d'oxyde. Dans le cas des rubans de
preservation and conservation problems.44
magndtophone d bobines, on a effectut des essais
Gerry Gibson has expressed it best:
defrottement et d'odeur, et I'on a tentd d'identi'
Clearly, the solution can not be to endlessly fier plus de 20 genres de dommages, dont le plis-
sement, I'ttirement, I'effitement, le blocage et la
rerecord holdings. We must search to find a
migration du plastifiant. Les autres supports d'en-
more permanent storage media [sic] and to ac-
registrement ont dtd soumis, dans la mesure du
cept an archival format good for 50 or more
possible, d des essais analogues. Dans le cas des
years. Further, we must actively and aggres-
disques, I'usure des sillons a en outre 6td 6valute.
sively work together, since the job is far too
large for any one or two collections to under-
Au cours de la prtsente communication, nous
take. We must carry out coordinated research
traiterons des rdsultats et de I'incidence de cette
into the various factors that affect the long
etude. A court terme, les renseignements nou-
term storage and retrieval of the data and ma-
veaux que nous fournissons faciliteront la planifi-
terials in our collections. We must work to-
gether to build the shared pool ofknowledge
cation des mesures qui seront prises pour assurer
la prdsertation de ces enregistrements leur
which is necessary to prevent premature fail- -
relogement et leur copie notamment. L'etude
ure of the items in our care, and, thus, loss of
the knowledge of our civilization. Only in fait par ailleurs ressortir certains problimes de
conserl)ation qui demeurent non rdsolus, et qui
that manner can we assure that the informa-
sont aussi bien d'ordre technique qu'dthique.
tion that thev carrv will be transmitted to
Ainsi, sur le plan de l'6thique, la questions de-
fut r.e g"noutions.a5
meure de savoir si la rdparation du mtdium doit
d'abord viser d obtenir une audition supplhmen-
It is time that the conservation community took taire de I'enregistrement ou d assurer sa conser'
up the challenge! vation pour la posttrit4. Nous terminerons en
traitant des efforts que ddploie la Bibliothique
R6sum6 nationale en',ae de rtsoudre la dfficulti que
pose la crdation d'une discipline distincte pour
Les enregistrements sonores d Ia Bibliothique la conservation des enregistrements sonores.
nationale du Canada : une dtude d'titat et son
incidence sur la conservation References
C'est d la Bibliothique nationale du Canada que l.Quoted in Moogk, Edward B., Roll Back the
sont archivds, entre autres, les enregistrements
Years: History ofCanadian Recorded Sound
sonores produits au Canada. Au cours de l'6td de
and its Legacy. Genesis to I930 (Ottawa:
1990, les enregistrements sonores conservbs d cet
Atubnssement ontfait I'objet d'une etude d'6tat,
National Library of Canada, 1975) p. 5.
mende dans le cadre d'un programme de trois
Original source, J.B. McClure, ed., Edison and
ans. Les rubans de magnetophone d bobines, His Inventions, (Chicago: Rhodes and McClure
les cylindres, les disques longue durbe, les Publishing Co., 1895).
45 tours et les 78 tours ont tous 6tt soumis d
une dvaluation distincte. 2. Mooglq Roll Back the Years,p.8.

L'information ainsi recueillie facilitera la gestion 3. Bullock, Allison, National Library of

des collections, de m€me que la planification des Canada 1989 Survey ofthe Non-rare Printed
mesures d prendre pour assurer leur prdsertation Collections (Ottawa: National Library of
et leur conservation. On a ainsi cherche d savoir Canada, 1989, in press).

2l
4. McWilliams, Jerry, "Sound Recordings in 15. Gibson. Gerald D.. "Preservation and
Swartzburg," in: Conservation in the Library: Conservation of Sound Recordings."
A Handbook of Use and Care of Traditional Paper presented to the [U.S.] National
and Non-traditional Mqterials, ed. Susan Archives, February 28, 1989,p.4.
Garretson, (Westport: Greenwood Press, 1983)
p.164. 16. McWilliams, "Sound Recordings in
Swartzburg," p.164.
5. McWilliams, Jerry, The Preservation qnd
Restoration of Sound Recordings (Nashville: 17. McWilliams, "Sound Recordings in
American Association for State and Local Swartzburg," p. 169.
History, 1979)p.6.
18. Fontaine, Jean-Marc, Degradation de
6. McWilliams, "Sound Recordings in I' enregistrement magnetique audio (Paris:
Swartzburg," p. 165. Ministdre de la culture et de la communication,
1987) p. 10.
7. David Grattan and Helen Burgess, personal
communication, Canadian Conservation I 9. Fontaine, Degradation de I' enregis trement,
Institute, 1991. p. ll.
8. St.-Laurent, Gilles, "The Care and Handling 20. Fontaine, Degradation de l'enregistrement,
of Recorded Sound Materials," Commission on p.13.
Preservation and Access Report, (Washington,
D.C.: Commission on Preservation and Access, 21. Blank. Sharon. "An Introduction to Plastics
l99l) p. 8. and Rubbers in Collections," Studies in Conser-
vation.35, 1990, p.53.
9. Heckmann, Harold, "Storage and Handling
of Audio and Magnetic Materials," in: 22.Brown, D.W., R.E. Lowry and L.E. Smith,
Preservation of Library Materials, ed. Merrily Prediction of the Longterm Stability of
A. Smith (Paris: Saur, 1987) p. 68. P olyes ter- B as ed Re cordin g Me d ia, NB SIR
82-2530 (Washington, D.C. : National Bureau
10. Gibson, Gerald, "Preseryation of Non-paper ofStandards, June 1982) p. 9.
Materials," in: Conserving and Preserving
Library Materials, Kathryn Luther Henderson 23. Smith, L.E., D.W. Brown and R.E. Lowry,
and William T. Henderson, eds. (Urbana-Cham- Prediction of the Longlerm Stabiliry of
paign: University of Illinois, 1983) p. 103. P olye s t er- B as ed Re cord in g M edia,
(Washington, D.C.: National Bureau of
11. Gilles St.-Laurent, personal communica- Standards, June 1986) pp.2-3.
tion, 1991.
24.Pickett and Lemcoe, Preservation and
12. Pickett, A.G. and M.M. Lemcoe, Storage, p. 56.
Preservation and Storage ofSound Recordings
(Washington: Library of Congress, 1959) 25. Brown, Lowry and Smith, Prediction of
p.30. Long-term Stability, 1982, p. 9.

1 3. McWilliams, Preserttation and Restoration. 26. Brown, Lowry and Smith, Prediction of
p.42. Long-term Stabilin, 1982, p. 2.

14. Alexandrovich, George, "Phono Cartridges 27. Smith, Brown and Lowry, Prediction of
and Communications," Broadcast Engineering, Long-term Stability,l986, p. 1.
August 1982,p.26.
28. Kent. Scott. "Binder Breakdown in
Back-Coated Tapes," Recording Engineer/
Producer, July 1988, p. 80.

22
29. Waites, J.B., "Care, Handling, and 38. Geller, Sidney 8., Care and Handling of
Management of Magnetic Tape," in: Magnetic Computer Magneti c Storage Media, NBS
Tape Recordingfor the Eighties, ed. Ford Special Publication 500-10 I (Washington,
Kalil, NASA Reference Publication 1075 D.C.:U.S. Dept. of Commerce, June 1983)
(NASA, April 1982) p. 50. pp. 94 and I 15.

30. Ad Hoc Subcommittee on the Preservation 39. Cuddihy, Edward F., "Stability and Preser-
of Sound Recordings of the National Archives vation of Magaetic Tape," in: Proceedings of
and Records Administration (NARA) Advisory Conservation in Archives: International
Committee on Preservation. Minutes of meet- Symposium, Ottawa, Canada, May 10-12,
ing July 29-30,1987, Washington, D.C., p. 13. 1988, (Paris: International Council on
Archives, 1989)p.204.
31. Waites, "Care, Handling, and Manage-
ment," p.54. 40. Gibson, "Preservation and Conservation,"
p.7.
32. Association for Recorded Sound Collec-
tions (ARSC), Associated Audio Archives 4l . Association for Recorded Sound Collec-
Committee, Final Pedormance Report. Audio tions (ARSC), Final Pedorrnance Report, p.5.
Preservation: A Planning Study (Washington,
D.C.: ARSC, December 31,1987) p. 54. 42. Smith, Brown, and Lowry,Prediction of
Long-term Stability, 1986, p. 20.
33. Gibson, "Preservation and Conservation,"
p. l. 43. Association for Recorded Sound Collec-
tions (ARSC), Final Performance Report,p. ll.
34. Pickett and Lemcoe, Preseryation and
Storage,p.26. 44. Association for Recorded Sound Collec-
tions (ARSC), Final Perforrnance Report, p.4.
35. Pickett and Lemcoe, Preserttation and
Storage, p. 49. 45. Gibson, "Preservation and Conservation,"
p. 17.
36. Fontaine, Degradation de l' enregistrement,
p.18.

37. Kent, "Binder Breakdown in Back-Coated

Tapes," p. 81.

z)
Plastics in the Science Museum, London:
A Curator's View

Susan Mossman

Collections Division
Science Museum
London, U.K.

Abstract stored separately, and, in cases ofsevere degrada-

tion, are isolatedfrom thosewhich, sofar, ap-
It is only in the fairly
recent past that plastics pear to be "healthy."
have become afeature of many museum collec-
tions. During the lastfve years, curators have Current research into the degradation ofthe earli-
begun to notice that objects made ofplastics de- est cellulosic plastics appears to suggest that in
grade with time, and sometimes very rapidly, 50 years' timefew of these will remain. The cura-
indeed. These provide problemsfor the curator, tor has to consider very carefully whether to dis-
panicularly in a technological museum such as play the most vulnerable types of plastics, which
the Science Museum, where we collect artefacts will then be subjected to quickened rates ofpho-
that may be wholly or partly made of plastics. As todegradation. Should the curator acquire two of
time goes on, our plastics artefacts, with their every plastics object: onefor display and hence
attendant problems of degradation, will only disposable and another to keep in optimum stor-
increase in number. age conditions to give it the longest lift possible?

The Science Museum's collection of plastics Introduction

contains about 1,500 objects, and is ofgreat im-
portance because it contains a ich selection of How long will our plastics collections survive?
historic plastics, ranging from Parkesine, the first This is the question that curators are now hav-
semi-s.vnthetic plastic, to a sample of the earliest ing to ask themselves. It has become evident
polyethvlene. We continue to collect actively in over the last few years that collections ofplas-
thefield of modern plastics materials, in particu- tics might not last forever. They degrade, and
lar composites and biodegradable plastics. In some faster than others. Sadly, it also appears
addition, plastics appear in many of our other
that the most historic and earliest plastics are
collections where they play an increasingly impor-
the ones that are most lulnerable.
tant role, especially in telecommunications, elec-
trical engineering, medicine and transport.
John Morgan explains elsewhere in this publica-
Particular problems we have noted occur in the tion what a polymer is and briefly introduces
earliest semi-synthetic plastics, that is, those con- some of the historic plastics. This paper refers
taining cel lulose nitrate. to many of the same plastics, but from the
point of view of a curator of plastics of varying
lle are in the process of improving the storage of kinds, from the oldest to the most modem,
our plastics, giving due attention to the effects of referring in particular to objects in the Science
photodegradation and high temperatures as well Museum's collections. The paper's flrst part
as ensuring that the various types ofplastics are maintains a roughly chronological framework.

25
The Science Museum's collection of plastics Vulcanite is well represented in the collection
(about 1,500 in number) is rich in the high- that includes many pieces made by Thomas
lights of plastics' history, and many pieces are Hancock, including a plaque of himself, dated
(as far as we can tell) in quite good condition. ca. 1843, and made of rulcanized rubber. Re-
We have obtained specimens from a variety of markable are some ornate vulcanite plaques
sources, ranging from industry to relatives of made by Hancock and sealed up in glass-
the inventors of certain plastics. We are also fronted passe-partout frames, which were dis-
fortunate in owning two large groups of plas- played in the Rubber Exhibition at the Science
tics (together numbering over 500 objects) that Museum n 1929. These have survived in excel-
are not only useful because they cover the full lent condition; possibly linked with the fact that
range of plastics materials from the late l9th their sealed environment has set up a stable mi-
century until ca. 1970, but also because they cro-climate, with no access to oxygen to encour-
contain objects of aesthetic importance. Some age the degradation process (Figure l). Only
of the more important items are addressed in one plaque has begun to fade slightly.
the following discussion.

The collection includes the rubber-based prod-

ucts that predate what is regarded as the first
plastic, Parkesine. Some of the rubber-based
compounds were used in their unmodified
state, such as gutta-percha. These fall into the
category defined as natural plastics.

Gutta-percha is a dark brown substance ob-

tained from the palaquium tree, which is found
in Malaysia. It became a very popular material
in the early Victorian period and was used for a
multiplicity of purposes. The Science Museum
possesses a number of examples of this mate- Figure I Vulcanite plaques, ca. 1843, qre
rial, including an inkstand dated to about 185 I elaborately decorated and in excellent condition.
(moulded to commemorate the use of gutta- This is probably linked to the lack of oxygen
available in their sealed environment.
percha to insulate the first submarine telegraph
cable from England to France in 1850). Gutta-
Perhaps the greatest treasures in the plastics col-
percha possesses excellent insulating proper-
lection are the 87 examples of Parkesine dated
ties, and so was ideal for this purpose. The
from I 855 to I 88 1 . Parkesine, invented by
main problem we have with gutta-percha is that
Alexander Parkes, is generally accepted to be
with age the surface cracks, and the object be-
the first plastic, although at this early date, of
comes brittle and fragile.
course, it was a modified natural product rather
than a fully synthetic material. Parkesine is
Other natural plastics in the collection are hom,
made of cellulose nitrate, which is made from
tortoiseshell and shellac, as well as a very rare
nitric acid, sulphuric acid and cellulose (ob-
material called Bois Durci (literally translated
tained from such sources as cotton flock). It
as "hardened wood") blended from sawdust and
is then mixed with vegetable oils and small
albumen.'
amounts of organic solvents. This mixture
forms a pliable dough that can be moulded into
Certain other rubber-based substances, such as
a variety ofshapes. A collection ofParkesine
vulcanite, are themselves sometimes regarded
objects presented to the Science Museum in the
as semi-synthetic plastics, since they are modifi-
1930s by the family of the inventor contains a
cations ofa nafural substance. In the case of
wide range of items, from crude samples to or-
r.ulcanite, sulphur is added to natural rubber to
nate hair-slides inlaid with silveq brass and
produce a harder material.
mother-of-pearl (Figure 2). Other pieces of
Parkesine also convey religious themes. A

26
remarkable piece is a tiny carved head gf celluloid, combining innovation with sharp mar-
Christ, measuring 2.7 cm in height. A signifi- keting skills. He realized there would be a keen
cant number of these pieces show Alexander market for celluloid combs and wipe-clean
Parkes'skills as a carver. collars and cuffs.

A fibre based on cellulose nitrate, and de-

scribed as artificial silk, was patented by Count
Hilaire de Chardonnet in 1884, and samples
were shown at the Paris Exhibition in 1889.
We have an advertisement for Chardonnet silk
showing dyed and undyed samples that date to
1896 (Figure 3).

7XR-rrFrerTf,b
",:'
srLr,{..
' '-''
The Artificiai Silk Company. Iirnited.
+
Figure 2 Parkesine hair-slides with brass, silver and
mother-of-pearl inlay, ca. 1862. Metal inlay is ltfting
A ,Vr-'rr Btttt'rsn /,rlr''srRr',
due to shrinkage of Parkesine.

Parkes set up the Parkesine Company Ltd.

to manufacture Parkesine. After the company
failed in 1868, his one-time Works Manager,
Daniel Spill, tried to make money out of
Parkes' invention, which Spill renamed Xylo-
nite. The Xylonite Company (set up by Spill in
1869) also produced material made of imitation
ivory (called Ivoride) and coral. The Science
Museum possesses objects made of Ivoride
and imitation coral, which were manufactured
by Spill's company in the 1870s. Of particular
interest is the Ivoride Death's Head walking-
stick handle, which was Spill's own. Samples

Neither Parkesine, Xylonite nor Ivoride were

very satisfactory products due to problems with
flammability. John Wesley Hyatt, an American, Figure 3 Chardonnet silk, 1896. Yellowing;
experimented to develop a more stable product. Iiquidizing; based on cellulose nitrate (artifcial silk
Hyatt made the breakthrough in 1870 when he patented by Count Hilaire de Chardonnet in 1884, of
found that camphor made an excellent solvent which samples were shown at the Paris Exhibiion
and plasticizer for cellulose nitrate. flyatt pat- in 1889).
ented his invention on l2 July 1870' and
called his product Celluloid, the name by Our collection includes samples of other semi-
which most products based on cellulose synthetic plastics, such as the milk-based plas-
nitrate are known today. tic. casein. and the less flammable cellulosic
plastic, cellulose acetate.
Spill then entered painful years of litigation
with Hyatt, who by l872had set up the Cellu- In 1909 came the next important development
loid Manufacturing Company. Spill finally lost in the history of plastics, when a brilliant Bel-
the battle, and Hyatt made a great success of gian chemist, Leo Baekeland, patented the
frst phenolic polymer. When combined with a

27
filler, such as wood flour or cotton flock, phe- Material. This consists of a number of cosmetic
nol-formaldehyde forms an excellent material boxes ofcellulose acetate and later, urea-for-
that can be moulded into a durable plastic maldehyde, made in Paris by the workshops of
with excellent electrical insulating properties. Editions Paris E. Fornells between the years
Baekeland's invention was quickly adapted to 1913 to 1940 (Figures 4 and 5). Eduard For-
many uses, primarily electrical fittings, but was nells Marco, an Andorran by birth, was a mas-
also commonly used for telephones and radio ter carver recognized for his skills and taken on
casings. A point to note is that phenolic plastics as a craftsman by Ren6 Lalique. Fornells was
are dark in colour, due to their fillers. A remark- responsible for the famous cherry box design
able object in our collection is a coffur, in- marked Lalique, made between l9l I and 1913,
vented by James Doleman and made in 1938 of which the Museum also has an example in
by the Ultralite Casket Co. Ltd. The coffin is its collection (Figures 6 andT).
claimed to be the largest phenolic moulding
ever made. It was manufactured from imitation In addition to a wide range of products illushat-
walnut phenolic with a wood flour hller (a spe- ing the applications and particular advantages
cial type of moulding material produced by Ba- of individual plastics materials, the Science
kelite Ltd. of London). Various stories are told Museum collects machinerv used for the
as to why the coffin did not go into large scale
production. According to the inventor's son
(personal communication) it was due to his
father's death in 1944, durngWorld War II.

By 1928, a method had been perfected of cast-

ing phenol-formaldehyde without a filler. The
result was the ability to produce cast phenolics
in bright colours. Cast phenolic was used to imi-
tate natural materials, such as amber and jade.
The Science Museum collection includes a
FADA Radio, dated to the 1940s. This is a
collector's item and made of imitation amber
catalin (cast phenolic).

The mid to late 1920s saw the advent of the Figure 4 Fomells Material: cosmetic boxes of
colourlhl as well as durable and waterproof c e ll ulo s e ac e tat e a nd late r, ur e a-fo rm a I d e hy d e,
plastics made of thiourea-formaldehyde and made in Paris by the workshops o/ Editions
urea-formaldehyde. Identified by such trade Paris E. Fomells, ca. 1913 to 1940.
names as Beetleware and Bandalasta Ware,
these plastics were popular for domestic ware
and picnic sets. They are well represented in
our collection.

With the 1930s, came polyethylene, polysty-

rene, acrylics, and polyamides. We have a sam-
ple of the first polyamide knitted tubing, made
by du Pont in July 1935. Nylon was to be its
successor and would become a runaway suc-
cess as an excellent thread for stockings, and
later in solid form as a material for gears and
bearings.

As well as very practical objects, the plastics

collection includes items important for their
artistic merit. Among the finest is the Fornells
Figure 5 Close-up of Fornells' trade mark.

28
 not one of chemistry.a There was a need to de-
velop a reactor that could produce greater quan-
tities without a mnaway reaction leading to an
explosion in the reaction vessel. By 1939, ICI
had developed the first fuIl-scale polyethyl-
ene plant, which was capable of producing
100 tons (based on a 50-litre reactor). ICI
continued to expand their production capacity,
which was necessary, as polyethylene played a
vital role during the Second World War, for ex-
ample, in radar. The nine-litre reaction vessel
was an important step in this development.

The Science Museum is also active in collect-

Figure 6 Cherry box made by Eduard Fornells
ing archive material, for example, trade cata-
Marco (between l9ll and 19I3) shows Lalique
trade mark.
logues, associated technical literature and
sample books. Particularly significant is a gutta-
percha catalogue from the Gutta-Percha Com-
* pany, ca. 1851, which shows a range ofobjects

€{!t
from cherubs to coats of arms moulded in this
material. In addition, the Museum holds a num-
. *-,8 ber of casein button sample cards (part archive

{tt'-
; lf . t*il'
tt- t1
and part object) dated to the first halfofthe
20th century.

Such archive material is useful because it helps

us assess the range ofobjects that were avail-
able at the time the catalogue was produced and
it also helps us identify objects. Work diaries
and private papers often give us insight into the
development of the plastic, working practices,
and problems encountered.
Figure 7 Close-up of trade mark on Lalisue cherrv
box. It is essential when acquiring an object, particu-
larly modem one such as a composite ski
a
production of plastics. The machinery has sepa- boot, to acquire the relevant technical literature
ratg conservation problems (i.e., metal, wood at the same time. A modern Salomon ski boot
and rubber), which are dealt with elsewhere in
(1985) in the collection is made of four injec-
this publication. tion-moulded parts in nylon with polycarbonate
clips; the boot is lined with a polyurethane
An important example of such machinery is a foam sock.
pilot-plant, nine-litre reaction vessel for poly-
ethylene, which was used at Imperial Chemical In the absence of technical literature, it would
Industries (ICI), Winnington, in 1938 to pro- be almost impossible to find out what plastics a
duce the first ton of polyethylene (of which we complicated object is made of, without much
are fortunate to have a sample). ICI had begun
time and expense. Howeveq sometimes manu-
the chemical work that led to polyethylene as facturers are reluctant to reveal the secrets of
early as 19323 andwere finalfy successful in re- their latest invention. Occasionally, we have to
peating the discovery somewhat fornritously in wait for over 50 years before the full story be-
1935. Due to the very high pressures required hind a particular plastic is released.
to produce polyethylene, ICI regarded produc-
ing larger amounts as an engineering problem, Another area to explore and collect is oral his-
tory. Although we have no formal programme

29
in this area at the Science Museum, it is often problem we have with other Parkesine objects
invaluable to record interviews with pioneers of is shrinkage. This is particularly evident with
plastics and later have the text transcribed. decorative combs that have metal inlays. In
some cases, the object has shrunk so much that
On a broader level, plastics materials are not the inlay has popped ou1.
only confrred to the plastics collection; many
other Science Museum collections contain ob-
jects of plastics in combination with other mate-
rials. Some that deserve mention are collections
of electronics, telecommunications, medicine
and aeronautics.

Are plastics worth preserving? We of course

are biased and hope that the brief resum6 of
some of the treasures in our collections are
convincing proof of their value and importance.
Many, in particular the Parkesine and polyethyl-
ene objects, hold great significance for the his-
tory ofthe science and technology ofplastics.

The curator has to make decisions as to what

to collect, both on the grounds of the object's Figure 8 Samples ofdeteriorated Parkesine (thin
significance and longevity. This does not mean films and pen nib), ca. 1862.

that the more l'ulnerable plastics should not be

In the case ofa Parkesine denture, degradation
collected; however, it is important to be aware
was rapid. The denture had been on display for
of their probable lifetime as well as ensure that
all important information is recorded about two years. We noticed that yellowish powder
them upon acquisition. Other ways of recording was deposited around it; when we examined the
denture, it was so brittle, it immediately fell
the shape ofa plastics object should be consid-
into pieces.
ered. Various suggestions of doing this are
explored later.
Other pieces on display appear to be fairly
stable, but these are monitored regularly.
Which Plastics Are Most at Risk?
When a material has been stressed, for example
Cellulose Nitrate-Based Plastics
a cellulose nitrate comb made of imitation tor-
Our immediate concern is with those plastics
toiseshell, it crazes where it has been bent into
based on cellulose nitrate: Parkesine, Ivoride,
shape.
Xylonite and Celluloid, of which we have ex-
amples in our collections. Certain examples
When metal is in contact with cellulose nitrate
have shown characteristic signs of degradation,
objects, in the form ofinlays or hinges for ex-
as also discussed in detail by John Morgan.
ample, it appears to speed up the degradation of
These include sweating, crazing, discolouration
the cellulose nitrate, providing a focus around
(yellowing in the case of imitation ivory cellu-
which degradation begins and then spreads
lose nitrate objects) and the giving off of acidic
throughout the object.
fumes.

Curiously enough, the majority of our Parkes- Other Plastics

ine objects are in good, and sometimes excel- At present our cellulose acetate objects do not
lent, condition. The items that have suffered appear to be exhibiting signs ofdegradation.
most are the thin films and smaller pieces, such However, the Tate Gallery in London has expe-
as a pen nib and a tiny carved head that are rienced problems with sculphrres.made from
both now in fragments (Figure 8). The main cellulose acetate by Naum Gabo.' In addi-
tion, the problems of deteriorating cellulose

30
triacetate filrn are well known.o So we can fore- o The objects in store are kept in the dark. This
see problems with our cellulose acetate objects. is essential as plastics are very susceptible to
photodegradation, particularly when exposed
Currently our casein (also known by such trade to ultraviolet light.
names as Erinoid and Galalith) objects appear
to be in remarkably good condition; however, o Deteriorating plastics, especially cellulosics,
we are well aware of their moisture-retaining
are separated from "healthy" objects.
qualities as well as the danger of warping and
cracking if the humidity levels change too
r Gloves are always worn when handling
radically.
plastics, as fingerprints can indelibly mark
Phenolic plastics (generally referred to as certain types.
Bakelite) appear to be fairly stable unless bro-
ken (when they become susceptible to biologi- o Sticky labels are not applied to objects as we
cal attack due to their fillers made of wood have found that the glue can "eat into" the
flour and cotton flock). The main problem surface.
appears to be with fading (green to brown).
We have to deal with the practicalities of stor-
Cast phenolic plastics (often known as catalin) ing a large number of different plastics as well
are also susceptible to colour change, having a as our other collections (which are themselves
tendency to turn yellow. made of many and varied materials).

Solutions: What Are We Doing If we keep cellulose nitrate-based plastics, such

at the Science Museum? as Parkesine and Celluloid, on open shelving
for ventilation, as is recommended practice,
Thorough Records any gases released may attack other plastics.
The first priority is to record the object fully, Currently most plastics, apart from the largest
both in writing and visually. We are doing this pieces, are wrapped in acid-free tissue and
for the Science Museum's plastics objects. We packed in acid-free boxes. We are well aware
write a detailed description of the object on a this may not be ideal for the cellulosics in
record card, including date and manufacturer's particular.
name. Other space on the card is available for
dimensions, material, and most important of all, A few pieces have been laid out on open
condition. We date the card so that we have a shelves on plasiazote, which appears to be a
report of the item's condition at the time it is relatively stable base. We are monitoring this
recorded. This information will be important arrangement.
as more plastics begin to deteriorate, either by
discolouring, shrinking or by showing various Of course, a separate, well-ventilated room for
other symptoms, such as crazing and cracking. each type of plastic, with a separate room for
those that have begun to degrade, would be
It is important to photograph the object with a ideal. However, every museum has constraints
scale, preferably using colour. We are undertak- on space, and we are no different. Cost is an-
ing a programme that incorporates this practice. other limiting factor.

Stores We might consider using another room at our

We have improved the storage conditions for store in which we might install a ventilated
our plastics, thus: case for those cellulose nitrate objects that are
deteriorating. Even the air for ventilation is
r The plastics are separated by type of material. not without attendant problems: the store is at
Olympia, Kensington, in a busy part of London
and so suffers from air pollution due to heavy
. They are stored in stable conditions that are
traffrc. We cannot allow air to come into the
kept as cool as possible.
case without filtering it or cleansing it in parts replaced as they wear out? Or should the
some way. machinery be preserved in a suspended state,
but in good condition?
Display Material
Whilst we try to protect our plastics in store Moreover, many of the early plastics were
from many of the dangers that could initiate or produced using somewhat rough and ready
accelerate their degradation, what about those methods, which could on occasion be quite
plastics on display? We have to accept that, at dangerous. For example, producing cellulose
present, displaying plastics will shorten nitrate-based plastics, such as celluloid, is po-
their lives. tentially afirehazard. Within the plastics
manufacturing centre at Oyonnax (in La Val
Light levels can be reduced and ultraviolet Plastiques, Jura Mountains, France) is located
light can be excluded by the use ofoptical fil- La Grande Vapeur, a factory that was used to
ters. We have excluded natural light from our produce objects, mainly combs and spectacle
plastics gallery at South Kensington. frames, initially of cellulose nitrate, and later
ofcellulose acetate fromca. 1908. The factory
We also look forward to the results of the work had an in-built sprinkler system as well as a sys-
being carried out by Julia Fenn at the Royal tem of cells with steel doors for each workshop,
Ontario Museum, in setting up a sealed case for so that if fire broke out in one workshop it
plastics, and using a system ofcleansing the at- could be contained.
mosphere using scavengers and absorbents as
well as a method of ensurins the circulation of To be realistic, the factory inspector and the
air within the case.7 Science Museum's safety officer would not be
happy if we tried to reproduce these types of
Using traditional display methods will mean fireproof "cabines" in the Science Museum. In
that in some cases, for example with the cellu- fact, we were not allowed to display a block of
losics, complete breakdown may come very celluloid together with its pilot machinery in
rapidly. Indeed, it is vital for the curator and the 1970s due to fears about the flammability
conservator who are mounting the display of of the plastic.
plastics, to be fully aware of the implications of
their actions. The recipes used to make early plastics were
not standard. Rather like cooking, they varied
It would be most sensible to display only those according to the person making the plastic. We
plastics objects that are duplicates or easily re- have recipe books for casein in the Science
placeable. This revision in curatorial attitudes Museum. Other recipe books exist for cellu-
would have a direct inlluence on collecting poli- loid.* However, records show that the plastics
cies. The most satisfactory solution would be to pioneers spent a fair amount of time unjam-
collect two of everything: one to be kept in ming the machinery when it became clogged
ideal collections in store, and the other to be due to variations in the mixture. Of course to-
kept for display purposes. day, things are very different. Mixtures and
temperatures are precisely measured, and pro-
Solutions to the Problem of Display: vided one has access to the amounts used, it
Pros and Cons would be possible to replicate these. However,
modem machinery is usually on an enornous
Replicas scale, and this might well prove impractical.
A mould could be taken of the object, in order
to make a replica. This might be suitable where
original machinery has survived. However,
the use of the original machinery provides an * Ipswich Public Records Office holds recipe books
ethical dilemma for the curator of technology, for the manufacture of Cellulloid, from BXL Plastics,
as well as practical diffrculties. Should the ma- Wardles Story, Brantham in Essex, the site ofthe old
chinery be maintained in working order and Xvlonite Factorv.
Three-Dimensional Recording visitor would then be able to look at objects not
on display. This would also satisff the visitor
The object could be copied in a more techno- who wants to look at items made of a particular
logical fashion. Three-dimensional photo- plastic, without having to put the objects on
graphy is an option; the object could be display and thus contribute to their degradation.
photographed stereoscopically, so that the Moreover, it would be beneficial to those visi-
picture could be viewed in three dimensions tors with a particular interest. They could select
through special spectacles. This would give only the objects they wished to view for study
a clear idea of shape. Three-dimensional pho- purposes. Again, this practice would cut
tographs could be taken from all importrant down the unnecessary handling ofobjects
angles. Ifthe object degraded, a three- in store.
dimensional record of its form would at least
exist. Such a system is already in use in the However, such a system would have to be used
archaeological world, and in the Victoria and with caution. Visitors will always want to see
Albert Museu;n department of sculpture objects. Though this system will allow visitors
conservation.d to see objects not on display due to lack of
space or fragility of the objects, it will never
A more advanced technique would be to record totally replace the display ofreal objects.
the three-dimensional image electronically.
Such a system is available, for example, the Conclusion
company Tektronix produces a suitable pack-
age. However, all such systems cost money. Today the Science Museum is active in collect-
The Science Museum is currently undertaking ing new plastics materials, specifically those
a programme of photographing objects with engineered for a precise purpose, or those that
the long-term view of transferring the images to have found a new use. For example, recycled
computer; our outstation, the National Museum and biodegradable plastics are an increasingly
of Film, Photography and Television at Brad- important area in which to collect. We have
ford, is considering doing the same for its pho- even collected a flowerpot made of recycled
tographic archives. Another outstation, the plastic. Another recent acquisition is a
National Railway Museum at York, is explor- Biopol bottle, a degradable plastic produced
ing digital storage for its photographic material. by ICI.
Three-dimensional computer imaging is some
way down the ling (on current judgement. As for the rich collections that we have inher-
about five years).' However. if the obiects are ited, we are now becoming increasingly aware
already photographed in this format. the images that we have to put into practice what we leam
could be transferred to computer when suffi- about their stability. We have had to assess
cient funding is available. Currently such a them for condition and probable lifetime. We
proposal is in preparation for the Science have improved our storage; in particular cutting
Museum's collection of plastics. out exposure to light wherever possible, and
maintaining stable storage conditions, specifi -
In the long term, the computerized three- cally of temperature and humidity. We are
dimensional recording of the object would have keeping a careful eye on the more lulnerable
many benefits. The image would be readily plastics, in particular those based on cellulose
available at the touch of a button. It would no nitrate and cellulose acetate. As yet, we have no
longer be always necessary to unpack the ob- ideal solution to the problems of their degrada-
ject from store, which would reduce both unnec- tion but we are exploring new ways of prolong-
essary handling as well as superfluous exposure ing their lives and watching with interest work
to light. Equally, computer terminals could be being done at other institutions in Britain and
put in the plastics gallery and at information abroad, notably in Ontario, Canada.
banks throushout the museuln. The museum

JJ
Acknowledgement accordons ainsi aux ffits de la photodigradation
et des hautes tempdratures toute I'attention qu'ils
The author thanks Sue Cackett and Dr. Derek mtritent, et nous veillons d ce que les divers gen-
Robinson for their helpful comments on this res de plastiques soient mis en nlserve sipard-
paper, and Ben Booth for the information he ment et d ce que les plastiques qui se sont grave-
supplied. ment dtgradds soient isol4s de ceux qui semblent
toujours < intacts D.

All photographs are provided courtesy ofthe

S'ilfaut en croire les rdsultats de la recherche ac-
Trustees of the Science Museum, London.
tuelle sur la dtgradation des premiers plastiques
cellulosiques, il semble qu'il n'en restera quefort
R6sum6 peu dans 50 ans. Aussi les spicialistes de la res-
tauration doivent-ils ne ddcider d'exposer les
Les plastiques au National Museum of plastiques les pltts yulntrables qu'aprds avoir
Science and Industry le point de vae d'ane soigneusement analvst la question, puisque ces
consewatrice - matiires seront alors sujettes d une photodegra-
dation accildrde. Et il convient sans doute de se
Ce n'est qu'assez rbcemment que les plastiques demander s'il nefaudrait pas toujours acqudrir
ontfait leur entrde dans les collections de muste. deux exemplaires du m€me objet en plastique : un
Au cours des cinq dernidres anntes, les specia- premier qui serait exposd on pourrait
listes de la restauration ont par ailleurs com- 4ventuellement disposer
- etetundont
second qui serait
mencd d constater que les objets en plastique se mis en riserve
-
ddgradaient avec le temps, et parfois m€me trds - et qui serait conseryi dans des
conditions optimales pour durer le plus
rapidement. Une telle dtgradation n'est pas sans longtemps possible.
poser des probldmes d ces spdcialistes, surtout
dans un musde technologique comme le National
Museum of Science and Industry, oil les collec-
References
tions comportent des objets qui sont partiellement
ou entidrement constituts de plastique. Le nom-
l. Williamson, C., "Bois Durci," Plastiquarian,
bre des objets de musde en plastique continuera
vol. l, Winter, 1988, p.8.
par ailleurs de s'accroitre, multipliant d'autant
les problimes de ddgradation de ce matdriau. 2.Hyatt, J.W.. "Improvement in Treating and
Moulding Pyroxyline," U.S. Patent no. 105338.
La collection de plastiques du muste regroupe
quelque I 500 objets. Elle est trds importante car 3. Kennedy, C., rcl: The Company That
elle renferme une riche sdlection de plastiques Changed Our Lives (London: Hutchinson,
qui ont une valeur histoique t986) p.64.
- depuis le
P arkes i ne, I e pre mi er p I asti que s emi -synthefi que,
jusqu'd un ichantillon du premier polydthyl?ne. 4. Wilson, Gordon, personal communication,
Et des matidres plastiques modetnes continuent ICI Petrochemicals and Plastics Division
de s'y ajouter
- des plastiques mixtes et
bioddgradables notamment. Les plastiques sont
Licensing Manager, U.K.,9 August 1991.

en outre prtsents dans nombre d'objets des au- 5. Pullen, D. and J. Heumann, "Cellulose
tres collections, puisqu'ils jotrcnt un rdle de plus
Acetate Deterioration in the Sculptures of
en plus grand dans des domaines comme les
Naum Gabo," Modern Organic Materials,The
tilicommunications, le gtnie 4lectrique, la
Scottish Society for Conservation and Restora-
midecine et le transpor"t, pour n'en citer que
quelques-uns. tion, Edinburgh, 1988, pp.57-66.

Les probldmes particuliers que nous avons 6. Edge, M. et al., "Cellulose Acetate: An
relevts se posent dans le cas des premiers plas- Archival Polymer Falls Apart," Modern
tiques semi-svnthttiques, c'est-d-dire ceux qui Organic Materials, The Scottish Society for
renferment du nitrate de cellulose. Conservation and Restoration, Edinburgh,
t988,pp.67-79.
Nous cherchons actuellement d amdliorer la
mise en rdserve de nos plastiques. Nous

.A
)+
7. Fenn, J., "Scavengers for Controlling EUROPE
Combinations of Emissions from Exhibitions Wilhelm Koepp Zellkautschuk GmbH & Co.,
of Mixed Polymers," Polymers in Conserva- D5100 Aachen, Postfach 848,
tion, Congress, Manchester Polyechnic and Hegelsbendenstrasse 20, Germany.
Manchester Museum, July 17-19, 1991.
Companie lnternationale de Plastique
8. Larson, John, "The Three Dimensional Re-
Biod6gradable, 24 Boulevard Princesse
cording ofSculpture," Lecture in the series Tfte
Charlotte, Immeuble Est-Ouest, MC98000
State of the Arl, Royal College of Art/ Victoria
Monte Carlo, Monaco.
and Albert Museum, London, March 15, 1990.

EASTERN EUROPE
9. Booth, B., personal communication, July
BP Chemicals GmbH, Zaunergasse 4,
1991.
A-1030 Vienna, Austria.

NORTH AMERICA
Suppliers
United Foam Plastics Corporation, 172 East
Plastazote: Main Street, Georgetown, Massachusetts
01833-2107, U.S.
U.K.
Polyformes Ltd., Cherrycourt Way, AUSTRALIA
Stanbridge Road, Leight on Buzzard, Dunlop Foam Products Group, P.O. Box l,
LU7 8UH. Mordialloc, Victoria 3 I 95, Australia.

35
 Membership and Aims of the
Plastics Historical Societv

John Morgan

P lasticsHistorical Society
Plastics & Rubber Institute
London, U.K.

Abstract and The Conservation Unit (of the Museums

& Galleries Commission) have started a co-
The centenary ofplastics was celebrated in 1962, operative program ofwork into the degradation
yet it is only in comparatively recent years that problems of plastics materials. The international
an interest in collecting plastic mateials has membership of the PHS encompasses a variety of
emerged. In 1968, a History Discussion Circle people, including some from i ndustrv, education
was formed in Britain by the Plastics Institute. and research as well as designers, collectors and
The idea of recording the reminiscences o.l some stafffrom museums and auction houses. Many
of the plastics pioneers, first mooted at the time members have their ovn plastics collections,
of the centenary celebrations, was re-ffirmed. covering a diversitlt of th-emes. The society is thus
However, the Circle was short lived, and it was able to call upon a wide range ofexpertise and ex-
not until 1986 that sfficient general interest in p er i e n ce, and i t s j o urnal Plastiquaria nr c o n t ai n s
plastics had developedfor the Plastics Historical news and articles ofinterest to collectors and
Society (PHS) to become a reality. plastics historians.

The PHS is an independent society, ffiliated to What Is Plastic?

the Plastics & Rubber Institute with which it
shares headquarters. Its aims are "to promote In general terms we all know what plastics are,
the study, preservation and sharing of informa- but a simple definition that includes only plas-
tion on all historical aspects ofplastics, and to tics materials and excludes everything else is
encourage the recording ofany current develop- not easy to find. One dqfurition that has pre-
ment adjudged to be of value to future genera-
viously been suggested' includes the following
tions." A longletm aim is the establishment of a
characteristics:
National Plastics Museum, but in the meantime ef-
fort is being devoted to saving historically impor- o Solid materials
tant records and objects from being consigned to
the rubbish tip. This is especially important when
old, establishedfirms are taken over by someone . Organic, or organic/inorganic polymers
who has little svmpathy for the histoical
signifcance of the company. o Natural, chemically modified natural or
wholly synthetic polymers (or mixtures of
The PHS forms links with other organizations these)
having associated interests,for example, the
Kunsts toffe - Museums -Verein (KMV), a German
o Capable of being compounded with
plastics museum society. In addition, the PHS
colourants, plasticizers, etc.

JI
. Capable of being moulded at some stage in It was not until the 1920s that the word plastics
production was first used as a generic name for the new
materials. During this same period the poly-
There are some arbitrary exclusions, such as meric nature of plastics was elucidated and this
adhesives, fabrics, paints and rubbers, but opin- scientific understanding enabled new polymers
ions on this differ and the Plastics Historical to be researched and to be established in new
Society (PHS) takes a fairly broad view in its ways. Even so, new polymers were still discov-
definition of plastics. ered by accident during other investigations,
but their polymeric nature was immediately rec-
How Old Are Plastics? ognized unlike generations earlier when
polymers- were sometimes produced by experi-
It is generally considered that plastics materials ment, but the sticky residues were discarded as
were first exhibited in England at the Intema- failures. Injection moulding, one of the key
tional Exhibition of 1862 by Alexander Parkes. processes in plastics fabrication, was also devel-
The material was called Parkesine and was oped during the 1920s. This and the various fac-
made from cellulose nitrate, a chemically tors already mentioned are reflected in the
modifi ed natural polymer. sharp rise in the output of plastics, starting in
about 1930 and continuing to the present day,
The roots of the industry, however, lie in the the progression being hindered only slightly by
earlier manipulation of naturally occurring plas- the oil crisis in the 1970s. Figures I and2 re-
tics. Detailed mouldings were being made from spectively show the growth in world plastics
hom as long ago as the 17th century; the vul- production during 1900 to 1940 and 1940 to
canization of rubber was discovered in 1839 1990, from 20 thousand to 100 million tons
and in 1850 gutta-percha was used to protect per year, a five-thousand-fold increase.
and insulate the first submarine telegraph
cables. Also in the 1850s, shellac was com- The fact that plastics were initially regarded as
pounded with wood flour to mould Union cases substitutes no doubt contributed to their poor
to protect and display early photographic im- popular image an image that was reinforced
ages. Albumen, principally from blood, was
- years by the inferior quality
in those and later
also compounded with wood flour to make oftheir design and application a criticism
-
that is sometimes levied even today.
Bois Durci, which was widely used to make
decorative inkwells, plaques and other items.
The Formation of the Plastics
At fnst, the new plastics were mainly regarded Historical Society
as substitutes for naturally occurring materials,
such as shellac, tortoiseshell and ivory. Such A History Discussion Circle was formed in
material was in short supply and was becoming Britain in 1968. but it was short lived and it
technically inadequate for the demands of in-
dustry. At the tum of the 20th century, casein,
a plastics material based on the protein from
milk, was introduced in Germany, and for
many years it was known as artificial horn. At
about the same time the search for a synthetic
substitute for shellac produced the material that
many consider the real starting point of the plas-
tics industry Bakelite. This material was pro-
-
duced by compounding phenol-formaldehyde,
a new and entirely synthetic resin, with wood
flour, the same type of filler that was com-
pounded 60 years earlier with natural shellac
to mould such items as Union cases and mirror Figure I Worldplastics productionfrom 1900 to
and brush backs. 1940.

38
 which they have been involved, whereas many
others are motivated by the design possibilities
of plastics compared to other materials. The
warm touch and handling qualities of plastics
are also important to some collectors.

How to best care for a plastics collection is one

question that remains largely unresolved, but an
equally important consideration is the extent to
which plastics should be cleaned, polished or re-
stored. An example is the brown discolouration
that occurs to cast phenolic resins, such as
Catalin and Bakelite, the effect being particu-
Figure 2 World plastics production
larly apparent with blue or white colours. Their
from 1940 to
1990. ready discolouration in natural or artificial light
is one reason why warm colours, especially am-
ber, were more popular. M*y, if not most, cast
was not until alrnost 20 years later that suffi-
phenolic objects now appear much darker in
cient general interest in the historical aspects of
colour, and in many cases are of an entirely
plastics had developed for the Plastics Histori-
different hue to what they were at the time of
cal Society to become a reality.
manufacture. The colour and polish may be re-
stored by an abrasive cleaning process, but at
The PHS, founded in 1986, has established the
the expense of the removal of surface material.
following aims: The same is true of the more familiar moulded
Bakelite, but because of discolouration during
. To promote the study, preservation and moulding, original colours were usually drab.
sharing of information on all historical Nevertheless, colour change does occur the
aspects ofplastics -
original colour can usually be seen by inspecf
ing an unexposed part of the moulding, and one
o To encourage the recording ofany current may be tempted to remove the discolouration.
developments adjudged to be of value to Opinions on the advisability of such treatment
future generations differ widely and debates on these and related
topics are lively and sometimes a little
o To establish a National Plastics Museum controversial.

o To co-operate with other organizations With the aim of addressing these concems, and
interested in the historical aspects ofplastics the more urgent questions about conservation,
the PHS and The Conservation Unit (of the Mu-
The intemational membership of the PHS en- seums & Galleries Commission) have started a
compasses a variety of people, including some co-operative program of work into the degrada-
from all sectors ofthe plastics industry, staff tion problems of plastics materials in museum
from museums and educators, as well as and private collections. (This is discussed in
designers, historians and many others. more detail in a second paper by Morgan in
this publication).
Approximately half of the members are collec-
tors of plastics whose themes range from "any- Activities of the Plastics Historical
thing so long as it's plastic" to specialized Society
collections ofone particular era, one particular
material or a particular type of object. Reasons The PHS is an enthusiastic body of individuals
for collecting are as varied as the objects col- engaged in widely differing aspects of plastics
lected. Those from the industry itself are often history. Some of the topics listed below are
influenced by materials and processes with covered in their discussions.

39
o Recording reminiscences of the "old-timers" ainsi que le paftage d'informations sur l'histoire
who worked on many of the processes and de ces matidres, et defavoriser la consignation de
toute nouvelle ddcouverte qui pounait dtre jugee
materials that have been developed this
importante pour les gdndrations.futures. A long
century
terme, elle envisage de cnler un musie national
des plastiques. Dans I'entre-temps toutefois, elle
o Rescuing records from factories that are met tout en @uvre pour |viter que des dossiers ou
closing des objets d'impoftance du point de vae histori-
que ne soient ditruits. Et ce genre d'action prend
o Researching particular manufacturers, items tout son sens lorsque, par exemple, une entreprise
or materials 4tablie de longue date passe aux mains de gens
qui se prioccupent peu de ce qu'elle reprdsente
sur le plan histoique.
. Ananging exhibitions of plastics
La PHS entretient des relations avec d'autres
o Organizing meetings or visits of interest to organismes qui partagent les m€mes intdr€ts,
plastics historians d o nt I e Kun st st offe - Mus e um s - Ver ei n, une s o ci 6 t d
musdale allemande qui s'inttresse aussi aux
These topics, and many more, are reported in plastiques. Elle s'est de plus unie au Service de
the PHS magazine Plastiquarian together with conservation de la Museums & Galleries
news items and articles of general interest. Commission pour crder un programme qui per-
mettra d'approfondir les probldmes de dtgrada-
R6sum6 tion des matiires plastiques.

La composition de la Plastics Historical Society Parmi les membres diversifi4s de la PHS qui
a ses objectifs proiennent de la communautd internationale,
certains reprdsentent les secteurs de I'industrie,
Si 1962 a marqut le centenaire du plastique, de l'4ducation et de la recherche, tandis que
on n'a pourtant commenci qu'assez ricemment d'autres se recrutent parmi les concepteurs, les
d collectionner des objetsfaits d partir de ce collectionneurs ou le personnel des musdes ou
matiriau. En 1968, le Plastics Institute a crtb des entreprises de vente aux enchires. Par
un cercle de discussion historique en Grande- ailleurs, nombre des membres de la PHS possi-
Bretagne, et on a alors repris I'idde, lancte pour dent leurs propres collections de plastiques, qui
la premiirefois lors des ctltbrations du cen- couvrent toute une gamme de thdmes. C'est
tenaire, de noter les rtminiscences de certains donc dire que la PHS peut faire appel d des com-
pionniers du domaine des plastiques. Ce cercle pdtences et d des exptriences vaides. Son bulle-
n'a toutefois pas fait longfeu, et il faudra atten- rin Plastiquariu contient enfin des nouvelles et
drejusqu'en 1986 pour que I'intdr€t gdntral des articles qui sauront inttresser toute personne
pour les plastiques deienne assez grand pour qui se spdcialise dans la collection des plastiques
amener la criation de la Plastics Histoical ou dans I'histoire de ce matdriau.
Societ"v (PHS).
Reference
Lo PHS est une socitti indtpmdante, ffili4e au
Plastics & Rubber Institute, dont elle partage les 1. Redfarn. C.A.. "What is Plastic?" Plastics
quartiers gtndraux. Ses objectifs sont de promou- and Rubber International, vol. 2, no. 3,1917,
voir I'ttude et la conservation des plastiques, p.134.

40
Conservation Policies and Plans
Politiques et projets en matiire de conservation
A Joint Project on the Conservation of Plastics by
The Conservation Unit and the Plastics Historical Society

John Morgan

Plastics Historical Society

Plastics & Rubber Institute
London, U.K.

Abstract applied, whilst anticipation offuture degradation

problems malt enable the development of
Mounting concern about the deterioration of appr op n a t e prev en ti ve m eas ure s.
early plastics artefacts has led to co-operarion be-
tween The Conservation lJnit of the liuseums & Dffirent plastics degrade in diferent ways, and
Galleries Commission and the Plastics Historical it is therefore necessary to classifi objects accord-
Society ofLondon to devise a program ofwork ing to material. It is particularly important that
aimed at combating the problem. plastics susceptible to auto-catalvtic degradation
mechanisms-the most damagingform of deter, ,-
A step-by-step approach has been adopted, the ration-be identified and kept under obsenation.
first part of which is the publishing ofpreliminary
guidelinesfor the storage and displav ofplastics The progress and preliminary results of this w,ork
objects. These guidelines are based upon our will be presented and discussed.
current understanding of the ageing of these
materials. Introduction
The second part ofthe program is the determina- Museums and private individuals collect in-
tion ofthe nature and extent ofpresent degrada- creasing amounts of plastic, both as artefacts
tion problems so that appropiate priorities for in their own right and as component parts of a
research can be established. A survev encompass- wide range of items. This is an inevitable conse-
ing a w,ide range of histoical collections hai quence of the increasing importance of these
been started, the aims being to: provide details on
materials in our daily lives, starting with the
the o,pe of mateial collected; ascertain the pre-
exhibition of the fust semi-synthetic plastics in
sent condition of objects; determine Ihe atmos-
1862 and culminating with the arrival of "The
pheric conditions under which artefacts are
stored and displayed; and give information about Plastics Age" in 1979, when the volume of plas-
preventive and curative measures alreadlt being tics produced world-wide exceeded that of steel.
/bllowed by conservators and collectors.
Polymer degradation was observed in the l6th
It is already clearfrom the work to date that if cenfury by explorers who were returning with
plastics are Io suruive they will require more in samples of a newly discovered material. Later
the way of preventive consentation than has hith- known as rubber, the material was said to have
erto been given to nrost otlter materials. Proce- "perished" during the long sea voyage. Thomas
dures for arresting detefioration of plastics Hancock studied the degradation of rubber by
alrcady affected need to be established so that measuring the water loss from a sealed rubber
curative techniques may be researched and bag over 30 years, from 1820 to 1850. The

43
result, shown in Figure l, illustrates a tlpical o publish preliminary guidelines for the storage
course of degradation for many polymers; a and display ofplastics
relatively long induction period showing little
change followed by a period of accelerating . carry out a survey of plastics held in U.K.
degradation. In the 1850s problems due to collections
degradation of gutta-percha were encountered
when this material was first used as insulation o collate information relevant to the
on telegraph cables. degradation of plastics in collections

. provide training workshops for conservators

and collectors

o research stabilizing treatments and curative

techniques

The project is still in the early stages, but th9

preliminary guidelines have been published'
and the first of the training seminars took place
in May of 1992. The first stage of the survey is
in the form of a questionnaire, which has been
sent to a wide range of British museums, and
Figure I These data areJrom an experiment to also to several private collectors ofplastics.
determine water loss fi'om a sealed rubber bag. The
experiment was conducted by Thomas Hancock over
a period of 30 years, from I 820 to I 850. The data
Preliminary Findings from the Survey
illustrate a typtcal course of degradation for many
polymen: a relatively long induction period showing First, many private collections are devoted
little change follow'ed by a period of accelerating almost exclusively to plastics while others, for
degradation. example, button or toy collections, contain a
significant proportion of these materials. Quite
Degradation has serious economic implications often the objects are made entirely of plastic. In
to plastics manufacturers who have devoted museum collections, plastics are widely distrib-
much effort to limiting the extent of degrada- uted but often in comparatively small numbers
tion during the service life of their products. Im- and usually form only a part ofthe objects.
proved quality control and the use of stabilizing Very few museums have displays devoted to
additives has enabled plastics to be used in ap- these modern materials. Figure 2 shows the
plications from which they would otherwise be distribution of plastics in collections by period
excluded. However, little attention has so far of manufacture.
been paid to limiting the slow deterioration of
historical plastics, some of which were manu-
facfured before their true nature had been eluci-
dated and when the science of additives was
little more than a "black art."

In recognition of mounting concem about

the deterioration ofplastics in historical collec-
tions, The Conservation Unit (of the Museums
& Galleries Commission of Great Britain) and
the Plastics Historical Society have started a pre l88o- 1900- 19m- 194G 1960- post
18a0 1900 1920 1940 1960 1980 1980
co-operative program of work, the aims of
which are to:
Figure 2 Popularity ofvarious periods in plastics
collections.

44
 The survey reveals that approximately half essentially linear or they can form a three-di-
of the museums replying to the questionnaire mensional, crosslinked network of relatively
possess more than 100 objects wholly or partly short chains. The former group comprise the
made from plastics. A slightly smaller propor- thermoplastic materials, that is, they soften and
tion report between 10 and 100, with only a flow when heated; the crosslinked polymers are
few (less than l0%) recording less than 10 or thermosetting, and once formed do not soften
more than I,000 such objects in their care. when heated and will normally char before
melting. Rubbers are a special class of thermo-
The most popular kinds of objects are those for setting polymer having long, flexible polymer
the home, followed closely by personal items chains with few crosslinks.
and toys. Mostly, they date from 1920 to 1960,
but surprisingly, more than one-third of the mu- Some linear polymers, such as polyethylene
seums report holding items manufactured since and nylon are partially crystalline, the degree of
1980. crystallinity varying with purity, processing
conditions and grade ofpolymer. Crystalline re-
Approximately twothirds of the museums gions are less permeable (e.g., to oxygen) than
report some deterioration of plastics, the most amorphous regions and as a consequence, deg-
common being crazing, discolouration or fad- radation may be less pronounced in the more
ing. Figure 3 shows the extent and kinds of crystalline material.
deterioration in plastics collections.
Polymers are also divided into two types,
addition polymers that are formed by "chain
growth" reactions, and condensation polymers
100 that are formed by "step growth" mechanisms.
Thermoplastics may be of either type whereas
o/ thermosetting plastics are almost always
formed from condensation polymers. Exam-
850 ples of addition polymers are polyethylene,
polypropylene, polystyrene, vinyl polymers
and acrylic polymers. Examples of condensa-
tion polymers are polyesters, polycarbonate,
0 nylon and phenol formaldehyde. The two types
Fadrng Bl@m Crang differ structurally in that addition polymers con-
tain only carbon atoms along the backbone of
the polymer chain whereas condensation poly-
Figure 3 Proportion ofcollections reporting various
mers also contain atoms other than carbon (e.g.,
forms of degradation in plastics objects.
oxygen) at regular intervals along the chain.
The two bpes are generally susceptible to dif-
The next stage of the survey will be to visit ferent degradation processes. The hetero-chain
a wide range of collections in order to obtain polymers, for example, may be susceptible to
more detailed information about the composi- hydrolysis chemical reaction with water
tion of objects and materials, particularly those resulting in-decomposition, or scission of
giving cause for concern. The results will be chemical bonds.
used to determine priorities in the final part of
the program, which is the setting of standards Physical Effects
for the care of collections, and the development These effects are associated with loss or
of stabilizing treatments and curative tech- migration of additives, absorption of liquids
niques for plastics materials. or vapours, crazing due to stress or fatigue,
mechanical damage, or excessive heat or cold.
Degradation of Plastics
Physical causes may be responsible for the
Plastics are made from long-chain molecules, following changes:
or polymers. The polymer chains can be

45
o distortion or dimensional change which may eventually lead to cracking, but the
. crazing or cracking maintenance of stable conditions will limit the
o surface deposit (often tacky) rate and, possibly, the extent of such changes.
o changes in flexibility
Plasticizer migration and loss may also occur
(Figure 4 shows typical degradation for cellu- due to changes in miscibility of the poly-
lose nitrate.) Cellulosic materials, casein and mer/plasticizer system brought about by chemi-
nylon are among the plastics that contain sig- cal degradation. In such cases, attempts to
nificant quantities of moisture, and their mois- reverse the process and replace lost plasticizer
ture content can vary with changes in humidity. by re-absorption would be ineffective.
Moisture absorption results in dimensional
changes and consequently the formation of Chemical Effects
stress. Varying humidity levels therefore pre- As with all materials, polymers can react chemi-
sent a low frequency fatigue stress and this is cally by coming into contact with other sub-
partly responsible for the crazing of casein stances. This includes those substances carried
plastics. in the atmosphere, such as oxygen, ozone,
moisture and pollutants. Chemical reactions
that involve scission of the polymer chain are
potentially the most serious. Because of their
network structure, thermosetting polymers can
withstand more chain scission than thermoplas-
tics and are generally found to be more stable.

Chemical effects are nearly always progressive

and irreversible, leading eventually to complete
Figure4 Cellulose nitrate shoehom showing disintegration. They are, therefore, usually
degradation 4,pical of this polymer. more serious than physical effects.

The combination of stress and certain liquids Evidence for chemical effects may be:
or vapours, which would otherwise have little
or no effect, can produce environmental stress- o colour change
cracking. The stress may be an applied stress or o chalkiness or surface bloom
it may be residual from moulding or machining c crazing
operations, or it may have developed with age. o embrittlement with loss of strength
This phenomenon became widely known from o evolution ofdegradation products (often
the premature failure of early polyethylene bot- acidic)
tles containing detergents. Rigid thermoplas- . softening or tackiness
tics, such as polysflrene and poly(methyl
methacrylate), are subject to stress-cracking in Major factors involved in bringing about
the presence of lower alcohols, paraffins, white chemical changes are:
spirit or oils. Acrylic materials intended for
outdoor exposure need to be annealed at a . light, especially ultraviolet
temperature of 80oC in order to minimize risk r heat
from this effect. . stress
o oxygen
Many physical effects can be controlled by the r moisture, including humidif
maintenance of stable conditions of tempera- . ozone and other atmospheric contaminants
ture and humidity and by avoiding mechanical (including those from nearby degrading
stress or contact with liquids and vapours that
objects)
might be absorbed. Migration of plasticizers in . contact with other materials (intentionally
such materials as cellulose acetate and poly(vi-
or by accident)
nyl chloride) result in these plastics becoming . some forms of biological attack
more rigid. This rigidity may cause distortion,

46
Not all of the above are damaging to any par- or synthetic rubbers are susceptible to this type
ticular material, but combinations of two or of degradation, which is potentially very dam-
more may be synergistic (i.e., the combined ef- aging as it can propagate at an accelerating rate
fect is much greater than the sum of the effects throughout the material.
considered individually). Antagonistic effects,
more simply known as "bad neighbours," can Light fades many colourants and discolours
occur between two materials. In this case, the many materials. Colour change can therefore
close proximity of one material adversely af- be a useful and responsive indicator of exces-
fects the stability of another. For example, the sive exposure to light. However, colouring mat-
copper core ofelectrical cables can accelerate ter may also mask early signs of deterioration.
the degradation of plastics insulation. A similar
effect was observed in the early 19th century Sometimes, degradation is conhned to the sur-
by Thomas Hancock the rubber coatings of face, as may be the case with some reactions in-
his waterproof textiles- were prematurely de- itiated by light on opaque materials. This type
composed by colouring matter present on ofdeterioration should be distinguished from
some dyed cotton fabrics. that where the whole of the material is affected,
as the consequences are much less serious.
Hydrolysis by atmospheric moisture is the main
degradation mechanism for the cellulose esters, Conditions to which a material has been
such as cellulose nitrate and cellulose acetate. exposed in the past often play a part in deter-
Hydrolysis plays a significant part in the degra- mining the nature and extent of future deteriora-
dation of polyesters (thermoplastic and thermo- tion. Removal of the object from adverse
setting) and nylon, and can also conhibute to conditions does not necessarily halt the degrada-
the degradation of polymer additives, such as tion processes. Conditions during polymer
plasticizers. The process usually results in the manufacture, compounding or moulding may
formation of acidic by-products, which acceler- also influence future degradation. For example,
ate hydrolysis so that decomposition proceeds during the manufacture of cellulose nitrate and
at an accelerating rate and the build-up ofacid- cellulose acetate polymers, acidic treatments
ity itself can initiate other degradation proc- are used and acidity that is not completely re-
esses. Depending on the polymer, the site of moved by subsequent processing is known to
hydrolysis may be in the polymer backbone (as affect the stability of these polyrners. Degrada-
is the case with many condensation polymers) tion may also be initiated by overheating the
or in side groups. The former is more serious material during moulding.
because it causes chain scission.
Control of Degradation
Hydrolysis is accelerated by acids or alkalies
and it is therefore important to ensure that Early detection of degradation and its likely
traces of such materials are not present in the cause is most important this can only be
vicinity of plastics prone to hydrolysis. -
achieved ifobjects are examined regularly and
their condition noted. Sometimes it may be nec-
Chemical bonds are also broken by other forms essary to conduct physical or chemical tests in
of energy, such as heat, light and mechanical an attempt to monitor the extent of any change.
forces. The broken bonds generally form highly
reactive species, namely free radicals, and the Plastics respond in different ways to the vari-
fate of the polyrner is controlled largely by ous environmental factors, and identification
chemical reactions associated with these. Light, will be necessary in order to appreciate how the
especially from the more energetic ultraviolet material might degrade. The minimum require-
region, causes bond scission in many polymers. ment will be the recognition of those materials
In particular, it may initiate an auto-catalytic that need to be kept under special environ-
form ofdegradation, known as autoxidation, mental conditions. The following tests and ob-
to which hydrocarbon polymers are especially servations may be sufficient for this purpose:
vulnerable. Polyethylene, polypropylene, ny- colour and appearance; handling qualities; hard-
lon, gutta-percha and plastics based on natural ness; density; end use; method of manufacture;

47
trademark; chronology; smell; signs of typical constant. Most plastics absorb moisture to
degradation; and specific tests. some degree and stress may develop as a result
of changing moisture content - stress is a
Confirmatory tests using various analytical contributory factor in many kinds of
procedures may still be necessary, especially degradation.
with objects giving cause for concem. The
large range of plastics now in production makes Ventilation
the task of identification more difficult for Ventilation is important for those materials that
recent material. emit gaseous degradation products. For exam-
ple, the cellulosic materials and poly(vinyl chlo-
Illumination ride) emit acidic products, which as well as
Light has long been recognized as a major fac- accelerating degradation may affect other parts
tor in the deterioration of polymers. It is damag- of the object or other materials and artefacts in
ing to all plastics and its exclusion is the single the vicinity. Metals may be corroded by the
most effective step that can be taken to mini- acidic vapours.
mize degradation. Since ultraviolet (UV) is the
most damaging part of the spectrum of light, it Cleaning
is important not to expose objects to it. Even Plastics materials often become soiled during
with ultraviolet removed, the light levels storage, handling and use. Many are electro-
should be maintained as low as circumstances static in nature and attract dust whilst others
permit. slowly bleed plasticizer to the surface and be-
come tacky. They do not seem to benefit from
Temperature Control the polishing action ofrepeated handling as do
Temperature influences the rate at which some other materials (e.g., horn and bronze).
physico-chemical reactions proceed the Perhaps they lack the right combination of hard-
higher the temperature the faster the -rate. It af- ness and wear characteristics. Only occasion-
fects the rate at which reactants, such as oxygen ally does a plastics object develop an attractive
and moisture, permeate into a material as well appearance or patina from the combined effects
as the rate at which reaction products permeate of ageing, wear and polishing. The most likely
out of the material. The optimum storage tem- plastics to do so are those for which colour
peratures for longevity of the various materials plays only a subtle part of their attraction, for
have yet to be determined. Long-term mainte- example, vulcanite and moulded Bakelite. In
nance of temperatures much below ambient such cases cleaning with a soft, dry brush or
would be expensive and may prove to be of cloth is probably all that is necessary.
little benefit.
It is recommended that plastics (with certain
For all plastics, the provision of a stable tem- exceptions) should be cleaned periodically to
perature is advised especially with objects con- remove surface contamination that may have
taining other materials that may have different built up over a period of time and, particularly,
thermal expansion coefficients. to remove any degradation products that might
accelerate deterioration. The frequency at
Humidity Control which this should be carried out will depend
The required level of humidity will depend upon circumstances, but with cellulosic materi-
upon the particular plastics material. For the als and casein it is recommended that this
cellulosic plastics, relative humidity should, should be done not less than once every five
ideally, be maintained below about 40%o. Ca- years.
sein and nylon become brittle and subject to
stress if their moisture content is allowed to be- Objects should be washed with tepid water con-
come too low, and for these materials a relative taining a small quantity of liquid detergent; the
humidity of 60oh is better. The moisture level use of a soft brush may be necessary for tex-
is less important with most other plastics but tured surfaces. Afterwards, the material should
for all groups it is important that it remains be rinsed with clean water and immediately

48
dried using an absorbent material. Do not soak half immersed in ESBO and left at room tem-
for long periods of time it is much better to perature for about two years. As shown in Fig-
rewash if soiling remains.- Only in rare circum- ure 5, the increase in size ofthe crazedarea
stances and after careful consideration should exposed to the atmosphere indicates continuing
solvents be used. Even if the material is not degradation. The areas immersed in the ESBO
directly attacked by the solvent, stress-crazing have not increased in size. Other objects show-
may result and this might not appear until later. ing initial signs ofdegradation have been
smeared with ESBO and the initial results
Care should be exercised with objects contain- are also encouraging.
ing metals that might corrode, or with hollow
objects that may be difficult to dry on the inner Conclusion
surfaces. Disassembly, if possible, is advised in
such circumstances. Some plastics objects have already undergone
serious changes to their properties and appear-
There may be occasions when objects, perhaps ance. These changes are unlikely to be revers-
soiled with tar or the residue from self-adhesive ible and all that can be done is to attempt to
labels, are difficult to clean with aqueous solu- retain those objects in a stable condition for
tions and in such cases one of the following the future.
solvents should be considered: white spirit, pe-
troleum spirit fuel for cigarette lighters and iso-
propyl alcohol (propan-2-ol). These solvents
possess optimum combinations of cleaning
power and inertness towards most plastics. Use
a cloth moistened with the minimum amount
of solvent. Do not immerse. rub too harshlv or
subject to stress. and do not use on rigid th-er-
moplastics, such as "untoughened" polystyrene
or acrylic because of the risk of stress-crazing.
Even for other materials. a wise precaution is to
test a small area on a hidden part of the mould-
ing. Remember that these solvents are highly
inflammable and toxic.

Degrading cellulose nitrate and cellulose ace- Figure 5 Cellulose nitrate mirror back showing pro-
tate that show crazing or porous areas should nounced degradation in four areos of the perimeter
not be washed. Stabilizing treatments are being at one-quarter and three-quarters ofthe height.
sought and the work on ESBO by Williamson' (Photo: cowtesy of C.J. & M.L. Ililliamson, The
described below shows promising results. Mansion House, Ford, Shrewsbury, U.K.
0743 850267.) Fig. 5a shows the mirror at an
early stage of the degradation. Fig. 5b shows the
ESBO and Cellulose Nitrate mircor at a later stage. The degraded areas in the
Cellulose nitrate is susceptible to hydrolysis by top half of the mirror that were not treated with
atmospheric moisture. Nitric acid is released ESBO continued to grow. The degraded areas in
and this promotes further hydrolysis so that, the lower half, which had been immersed in ESBO,
once started, degradation proceeds at an appeared to remain stable.
accelerating rate. It is possible that the applica-
tion of an acid acceptor, in a form that could mi- Plastics do slowly degrade in the atmosphere
grate into the material, might prove effective in
and their life will be considerably shortened if
retarding degradation by neutralizing the acidic kept under the wrong conditions. Since it is not
by-products. Epoxidized soya bean oil (ESBO) generally possible to eliminate completely the
has been tried. A mirror back, probably from
various environmental factors responsible for
the 1920s, exhibited typical degradation in the the deterioration of plastics, it becomes neces-
thick rim section in four distinct areas. It was sary to control the degree of exposure to them.

49
This will require classification of objects prioitaire. lJne ttude, couvrant wt grand nombre
into groups of materials, an assessment of the de collections historiques, a ainsi ttd amorcie
various risks followed by the provision of ap- dans le but de : founrir des d,ltails quant aux gen-
propriate storage and display conditions. Light, res de mattriaux qui se rett'ouvent dans les collec-
tions; dvaluer l'6tat actuel des objets: ddtenniner
humidity, temperature, ventilation and cleaning
les conditions ambiantes dans lesquelles les ob-
procedures all need to be controlled and
jets sont mis en rtserve et exposts: ffiir de l'in-
monitored.
.fotmation sur les mesures prtventives et curatives
qu'utilisent dtjd les specialistes de la restaura-
It is probably not feasible to provide the best tion et les responsables de collections.
environmental conditions for storing everything
collected and so it will be necessary to optimize Les travaux effectuds rdvilent clairement que,
the use of those facilities that can be made pour durer, les plastiques doivent, comparative'
available. This will require an assessment of ment d la plupaft des autres mattfiaux, Jhire
which materials and objects are at most risk I'objet de plus de mesurcs de conseruation
from deterioration. Such considerations should prtventive. Aus si faut-i I ilaborer des mtthode s
not be excluded when determining the collec- qui permettront d'arrAter la ditdrioration des
plastiques qui sont ddjd touchis
tion policy (i.e., what artefacts are collected).
- et donc effec-
tuer de la recherche dans le domaine des tech-
R6sum6 niques curatives et les mettre en application
tenter de prdvoir les problimes de ddgradation
-, et
La conservationdes plasfiques : un proiet mirte que risquent iventuellement de poser ces matd-
du Sewice de consewalion de la Museums & fiaux et donc ddfnir les mesures pr'4ventives
les
- appropri4es pour les combattre
plus
Galleries Commission et de la Plustics
Histoical Society
Des plastiques dffirents se digradent de manidre
Source grandissante d'inq uietudes, I a dit,lriora- dffirente, et il devient dds lors ndcessaire de clas-
tion des premiers objets de muste en plastique a sifier au prialable les objets en plastique suivant
amend le Service de conseruation de la Museums leur composition. Il importe tout particulidrement
& Galleries Commission et la Plastics Historical de reconnaitre les plastiques qui sont sujets d
Society de Londres d unir leurs elforts pour met- l'action des mdcanismes de dtgradation autocata-
de ddtdrioration la plus
tre sur pied un programme qui ise d enrayer lytique
- soit laetforme
destructrice de les garder sous obseruation.
ce probl'ime. -
Pour v arriver, on a adoptti une approche Nous dtcrirons, dans la prtsente communication,
comportant divers es ttapes. I'dvolution des travaux, tont en fournissant un
apergu de leurs rtsultats prtliminaires.
Ainsi, dans un premier temps, des directives
prdliminaires sur la mise en rtsewe des objets en References
plastique et sur letu" exposition seront publi4es.
Elles seront etubhes enfonction des connats' l. Morgan, J., Conservation of Plastics,
sances que I'on a dijd sur le vieillissentent de An Introductlon (London: Museums &
ce matdriau. Galleries Commission, I 99 I ).

Dans un autre temps, on cherchera d ddterminer 2. Williamson, C.J., Polymers in Conserva-

la nature et I'ampleur des problintes actuels de t n, Intemational Conference, Manchester,
io
dtgradation, afin de dLfinir les secteurs de la U.K., July 1991. (Proceedings to be published).
recherche qui devront €tre considirds defaqon

50
Conserving the Science Museum Collections

Roger Price and Anne Moncrieff

Science Museum
The National Museum of Science and Industry
London, U.K.

Abstract members of conservation staff and are given

basic training.
During 1989 to 1990 a nev,conservation
section was established at the Science Museum Like other museums that collect modent artefocts,
in London. Because itsfacilities are, at present, one oJ'the greatest challenges facing the Science
relativelv modest compared with some other na- Mttseum is how,to deal with the problem of the
tional museurns, the conservation section aims to degradation of modem materials. Although im-
protect the Museum's collections b1'rigorous proved quality ofstorage and displav might
managenent of its resources. help, research into new methods for tteatment
is urgentfu required. In the absence ofits own
In recognition of some damage that might have researchfacilities, the Museum has begun spon-
occurred becattse of the lack of clear guidelines, soring research at properly equipped institutions.
a neu'Code ofConservation Practice is being Bv such collaboration, u,hich need not be prohibi-
drawn up. This will include standardsfor storage tivebt expensive, some progress can be made in
and displat, (draw'ing on the Storer Report, spon- this dfficultfield.
sored jointll' b1' the Science Museum and the Mu-
seums & Galleries Commission), obserttations on In viev, of the diversity of materials collected,
at ailable h'eatntents for various materials, and all of u,hich have their own special problems, it
an explanation ofthe new conservation recording u,ould be sensible to broaden the research. Ide-
svstem. This document will serve as the basis for all1t, 17it should be organized on a world-w,ide
actions that might affect the presentation ofthe basis, with like-minded museums agreeing upon
Museum's collections. some wat' of atoiding duplication of work so that
resources can be more elfectivelv directed to-
In most technical museums, conservation is w'ards solving the problems that beset us.
undertaken bv skilled craft technicians. However,
few ofthem operate in the mainstream ofconser- Introduction
vation practice as it is understood, for example,
in fne arts or archaeology. Largefi, for that rea- Many museums specialize in aspects of indus-
son, some objects have been over restored, which trial and technological history and numerous
ffictively destroys evidence of historical signifi- non-specialist museums now acquire such arte-
cance. To remedy the situation, the Science facts as part of their wider collecting activities.
Museum runs a three-year Conservation oflndus- Among long-established museums of industrial
trial Collections Training Course specifically history, the size and scope of the collection is
designed to meet the needs oftechnicians.
greater than ever. New materials are replacing
Similarly, object-cleaning practices have been
reviewed. Obiect cleaners are nort included as
traditional metals and wood, making the

5l
conservation of industrial artefacts a complex lacking the rigour and discipline of modern
and demanding task. conservation practice. The Forum meets twice
yearly with the Director of the National Mu-
The National Museum of Science and Indus- seum of Science and Industry to discuss ways
try, which includes the Science Museum in in which the conservation of industrial material
London, the National Railway Museum and can be raised to a standard more appropriate for
the National Museum of Photography Film the needs of today and tomorrow.
and Television, is widely acknowledged as the
world's pre-eminent museum devoted to the An early achievement of the Forum is the 1989
history ofscience, technology, industry and publication by J.D. Storer, The Conseryation
medicine. Its collections are the largest, the of Industrial Collections: a Surve-v, sponsored
most comprehensive and the most significant jointly by the Science Museum and the Muse-
anywhere. To ensure that the preservation of ums & Galleries Commission. In this important
these outstanding collections fully meets the re- work, Storer surveys the facilities for conserva-
quirements of modern museum practice, a new tion and storage in 43 museums throughout the
conservation section was established durins country holding industrial and related material,
1989 to 1990. and makes a number of recommendations on
standards that should be met.
Maximizing Our Resources
Another direct result of the Forum's work is
The conservation section's most urgent task the 1989 inauguration of the Conservation of
was to identifu and quantify the major prob- Industrial Collections Training Course, run by
lems that threatened the well-being of the the Science Museum. As its most urgent prior-
Museum's objects so that, by rigorous manage- ity, the Forum advocates staff training in two
ment, the available resources could be deployed distinct areas: frst, the need to recover tradi-
most effectively. Quickly realizing that clear tional engineering craft skills (a need that must
guidelines or procedures were needed, work be and to some extent is being met by appren-
commenced on the compilation of a detailed ticeship to experienced craftspersons in muse-
Code of Conservation Practice. ums and private workshops); and second, the
need for a conservation approach based on the
The Code contains instructions to all staff on methodical study of an object, its history, dete-
how to use the conservation facility. This in- rioration and treahnent, of which a written re-
cludes everyone who is in any way involved in cord must be made.
taking actions that might affect the preservation
of the collections, including curators, designers, This new training course is specially designed
stores officers and building managers. It also to meet the needs of engineering technicians in
contains an explanation ofthe new conserva- the full-time employ of a museum. Over three
tion recording system, environmental standards years they receive 70 days of theoretical instruc-
for storage and display, and guidelines for con- tion during which they gain experience in the
servation staff on materials and techniques. wider issues of conservation. They must also
compile a written portfolio based on work they
Central to this initiative is the work of the have undertaken to meet the requirements of a
Conservation of Industrial Collections Forum. practical syllabus. Their progress is assessed
Founded in 1987, this group consists ofconser- regularly and there are examinations at the end
vators, curators and engineers from national of each year. Successful graduates are awarded
and local authority museums and related institu- a Certificate in the Conservation of Industrial
tions, as well as restorers in private practice. Collections.
For industrial collections in general, the ap-
proach to conservation is not as developed as The Science Museum has always employed
it is in, for example, fine arts or archaeology; staffto clean the objects displayed in the galler-
rather, the care ofindustrial objects is based ies. Staff cleaners have considerab le practical
upon traditional engineering craft skills, often experience but, until now, have had no formal
ofa very high order, but all too frequently training. In recognition of the continued need

5?
for this work, 12 cleaners were assigned to the Objects made from cellulose nitrate are also
conservation section in 1990. Three of them showing signs of deterioration. Because the
have since been promoted to act as supervisors Centre for Archival Polymeric Materials at
for each of three cleaning teams. In the last Manchester Polytechnic was already working
year, they have all had some training, which on the degradation of cellulosic film materials,
was based on standards used by the National the Science Museum provided funding for addi-
Trust. The large number of objects on display tional work. The project seeks to characterize
means that these cleaners will be busy dusting the polymer in both original and artificially
for most of the time, however, they will be aged samples, with the aim of finding methods
integrated more fully into the section and to assess the state ofdeterioration and to slow
given more haining so that they can work down the rate of decay.
with conservators on more specialized
cleaning tasks. The Museum has a few l9th-century cars
whose original upholstery is in poor condition.
Through this endeavor the Science Museum en- To conserve the upholstery rather than restore
sures that the objects on display are presented it with new materials, a joint project was set up
attractively and are also treated with appropri- with the Leather Conservation Centre in North-
ate care. Another important gain from such a ampton. New materials and methods were in-
regular cleaning programme is that the objects vestigated to effect the repair and reinforcement
displayed are subject to frequent detailed obser- of the old leather. So faq certain experimental
vation and any change in their condition can be methods have been successful and have been
recorded and treated where necessary. used on two of the cars.

Our Approach to Degrading Condition surveys on objects in the stores have

Modern Materials confirmed early impressions that some commer-
cial protective coatings are effective in retard-
Like other museums that collect modern arte- ing metal corrosion. The Museum will continue
facts, one ofthe greatest challenges facing the to use these, drawing on the experience of other
Science Museum is how to deal with the prob- users. The Museum is also discussing the
lem of the degradation of modern materials. choice ofoils for lubrication and hopes to spon-
Although improved quality of storage and dis- sor research into the degradation ofoils and
play might help, research into new methods for metal surfaces in collaboration with a univer-
treatnent is needed. With no research facilities sity research department outside London.
of its own, the Museum has begun to sponsor
research at institutions that have the staff and While working with existing research groups
equipment for such work. has the advantage ofhaving access to expertise,
it also has the slight disadvantage ofnot having
The Science Museum has some aluminium complete control over the project and the direc-
alloys, especially castings, that are corroding. tion that it might take. It is not easy to find re-
In collaboration with the Corrosion and Protec- search groups to work on the materials that
tion Centre at the University of Manchester interest the Science Museum; nor are most stu-
Institute of Science and Technology (UMIST), dents prepared to work on old materials. Fur-
the Museum has obtained a Science and Engi- thermore, the problems do not always fit neatly
neering Research Council Award for a Ph.D. into the time available to Ph.D. students. Never-
student to carry out research into this problem. theless, this is an effective way to use limited
The aim is to develop a technique that will resources and with the right partner such co-
reduce the rate of decay to a minimum. The operation can be successful and rewarding.
frst two years ofthis three-year project have
brought some success on simulated objects, There is a very real problem with complex
but the methods have yet to be tested in a objects, such as older computers. An example
real situation. of one approach used by the Science Museum
is the formation of the Computer Conservation

53
Society. Collaborating with the British Com- duplication ofwork so that resources can be
puter Society, the Museum has allocated more effectively directed towards solving the
a curator post and provided facilities for the problems that beset us.
enthusiastic and expert volunteers; the techni-
cians, engineers and computer operators who R6sum6
once actually worked with these same early
computers. La conservation des collections du National
Museum of Science and Industry
The Society first focused on restoring to work-
ing order a Fewanti Pegasus vacuum-tube En 1989-1990, une nouvelle section de conserva-
computer dating from 1958 and an Elliott 803 tion a ttt crtte att National Museunt of Science
germanium-transistor machine dating from the and Industrl, de Londres. Contme ses installations
sont,pour le moment, relativement modestes
early 1960s. This work has been fully docu-
comparativement d celles de certains autres
mented and is regarded as a continuation ofthe
mustes nationaux, cette section a pour mandat
working life of these objects. In the process of de prtsener les collectiotts du musie en assurant
this work young engineers and technicians are une gestiort rigoureuse des ressources.
trained to care for these machines. Though the
computers will only work for a few more years, Il a ttt ,"econnu que l'absence de lignes direc-
the experience enables the Museum to capture tfices claires explique certains des domntages
and record the expertise ofthe designers, build- qui se seraient produits, et l'on travaille donc ac-
ers, maintainers, programmers and users of tuellenrent d la ridaction d'un nouveau rtpeftoire
early computers. des rtigles et usages en matidre de consenation.
Ce document renfennera des normes pour la
The partnership between the Museum and these mise en rdserve et l'exposition des objets qui
computer professionals has also created a pro- s'inspireront du rapport qu'a prdpar6, sous - le
gramme of emulation whereby software is de- patronage mixte du musde et de la Museums &
signed to make a modern computer behave just Galleries Commission, la personne responsable
like the old machines, including their idiosyn- des rtserves
-, des observations sur les traite-
ments clui peuvent €tre appliquds aux divers
crasies. This enables anyone who is interested
mattriaux et des explications au sujet du nouveau
to experience something of the work of the
systdme d'enregistrement auxfins de conserva-
early pioneers. A publication on the work of
tion. Et il servira de document d'orientation
the Computer Conservation Society is planned
lorsque viendra le moment de prendre des me-
for 1992. sures qui pourraient influer sur la conservation
des collections du National Museum of Science
Conclusion and Industry'.

Conservation staff at our Museum keep in Dans la plupart des musies de la technique, les
touch with work going on in other museums by travaux de conservation sont confits d des tech-
attending conferences and through many infor- niciens et techniciennes qualifi,!s et expirimentds.
mal contacts, for example, the ICOM Modem Il s'en trouve ntanmoins peu patmi eux qui tra-
Materials Working Group and The Conserva- vaillent suivant les rigles de consenation qui
tion Unit of the Museums & Galleries Commis- ont cours, par exemple, dans un musde des beaux-
sion (whose Conservation Research Policy arts ou de I'archdologie. Et c'est sans doute ce
Group is keeping a record of conservation re- facteur qui explique le mieux que I'on retrouve
search in the United Kingdom). parfois, dans ces mttsdes, des objets qui ont 6ti
beaucoup tt'op restauris, jusqu'd en perdre pra-
tiquement toute signification historique. Soucieux
In view of the diversity of materials now col-
de faire sa part pour remtdier d cette situation,
lected by museums, all of which have their
le muste a mis sur pied un cours de formation
own special problems, it would be sensible to
de trois ans qui, portant sur la consentation des
broaden research into their deterioration and re- collections indwtielles, est spdcialement conQu
medial measures. Ideally, this should be organ- pour rdpondre aux besoins de ces techniciens et
ized on a world-wide basis, with like-minded techniciennes. Les usages entourant le nettoyage
museums agreeing upon some way of avoiding

54
des objets ont 4galement itd passts en retae, si qu'ffictuent, dans ce domaine, des organismes
bien que des personnes sptcialement affectdes ad,lquatement equipds. Il devient d'is lors possi-
d cette tdche ont 6tt inttgries au personnel de ble, grdce d une telle forme de collaboration, de
conservation, et qu'un cours deformation de rdaliser des progrds dans ce secteur plut6t dffi-
base a tti mis sur pied d leut'intention. cile, et ce, sans que le muste ait d effectuer de
gros ddbours,ls.
Le probldme de la dtgradation des mattriaux
modernes constitue sans doute l'un des plus Compte tenu de la diversitt des matidres qui se
grands d6fis auxquels le musde ait dfaire face d retrouvent dans les collections, et qui prisentent
I'instar des autres musdes qui collectionnent des toutes des probldmes particuliers, il serait
objets modernes. Et s'il est sans doute utile de rai so n n abl e d' 4largir l' ho izon des rec herc he s.
chercher d amdliorer la qualite des methodes de Une telle entrepise devrait iddalement €tre in-
mise en rtserve et d'exposition, il n'en demeure tegrde d un projet de portte mondiale, dans le
pas moins urgent de poursuivre la recherche de cadre duquel les mustes ayant les m€mes intdr€ts
nouvelles m'lthodes de traitement. Ne disposant se seraient mis d'accord sur unefaqon d'iviter
pas d'installations de recherche qui lui permet- les dedoublements, etferaient ainsi en sorte que
traient de poursuivre lui-m€me de tels trava*r, les ressources serent effectivement d la rdsolu-
le musde a commencd d parrainer des recherches tion des probl?mes.

))
History of Technology
Histoire de la technologie
 Rubber: Its History, Composition and Prospects
for Conservation

M.J.R. Loadman

Malaysian Rubber Producers' Research Association

Tun Abdul Razak Laboratory
B r i c ke n donbury, H ertford
United Kingdom

Abstract material. Possible causes and effects of surface

changes will be discussed and ways of distinguish-
Everyone living in the "modern" world is familiar ing between them presented.
v'ith rubber and its properties. Perhaps ii is that
ven'familiarin that has bred. if not iontempt. at Knowing v'hat the material is, and from the state
least an unthinking acceptance ofthe material of its surface what has happened to it so far in its
and its position in our civilization. The term ,,ntb- life, consider now what can be done to optimize
ber" is used to descibe any polyner that has or its display life. Wilst there is no one answer for
appears to have elastic properties, and there are all elastomers and their products, many suJfer
many such polymers of both natural and sltnthetic degradation through a series of common phenom-
oigin on the market today. In a "rubber,,prod- ena: oxygen, ozone, heat, light, mechanical work,
uct. the elastomer may well be less than 509o 61 pro-oxidant metals, bacteria and even that
its total mass since many additional chemicali dreaded word "chemistry"; combinations of these
ma1'have been added to produce a seniceable phenomena often have a synergistic effect. An ar-
article, and these in turn may have been selected tempt will be made to weove a path through this
from h un dre ds of p o ss i bi li tie s. minejield so that the best conservation procedure
'of
is chosen for each of the various tvpes elas-
It is not intended, in this paper, to delve deeply tomeric products on display.
into the scientific details of rubber product formu-
lation, but it is essential that, before starting any
conservation process, the conservator is aware of Introduction
the material with which he or she is dealing and
how,inappropiate treatment can do more harm There can be no-one living in the "modem"
than good. The history of elastomeric materials world who is not familiar with rubber and its
will be reviewed, with particular attention to de- properties, but perhaps it is that very familiarity
velopments in the 20th century, which might as-
that has bred, ifnot contempt, at least an un-
sist the conserttator in classtfying and possibly
thinking acceptance of the material and its
dating any elastomeric product. To further these
position in our civilization.
aims, a brief description of some simple methods
of analysing these materials will be included.
The natural material has been used for at least
Visually displeasing changes to the surface ofan 2,000 years. It may, even today, be used "raw"
elastomeric product are a common event and can for crdpe soles of high quality shoes, or mixed
sometimes pres age surface degradation, al though with chemicals in the latex state, prior to hav-
this is not alwal,s the case: some surface changis ing formers dipped into it to produce such
are intended and are beneficial to the life of the articles as baby bottle teats, condoms or

59
surgeons' gloves. The mixed (or compounded) ozone and even that dreaded word "chemistry";
latex may also be treated to produce latex combinations of these phenomena often have a
thread suitable for the finest underwear whilst, synergistic effect. An attempt will be made to
at the other extreme, the dried rubber can be weave a path through this minefield so that the
mixed with more chemicals, often including best conservation procedures may be selected
carbon black, to manufacture the strongest of for the various elastomeric products on display.
engineering products, such as base isolation
units for buildings in earthquake zones, con- The history of natural rubber, or NR as it is usu-
veyor belts and, accounting for by far the great- ally known, from a plaything over 2,000 years
est area ofusage ofelastomers, aircraft and car ago to the founding member of one of the most
tyres. If you doubt the remarkable properties of important classes of materials of modern day is
this material, remember the faith you put in the a fascinating story, unequalled, I believe, for
four 'handprints' ofthe tyres on the road be- any material. The more recent development of
neath your car. synthetic elastomers is equally interesting, aris-
ing as it did from a mixture of the chemist's ba-
In the last 150 years or so, hundreds of chemi- sic desire to mimic and improve on nature and
cals have been added to rubber to modify the the absolute need for a synthetic replacement
polymer itself, or its properties. And numerous for NR by both Germany and the United States
synthetic elastomers have been invented and during the Second World War.
manufactured, put through the same process of
mixing with all the chemicals used with natural Highlighted in an appendix are those events in
rubber, and more. Yet still, for general purpose the history of rubber that seem to me to be im-
applications, and, indeed, many specialized portant, or interesting, in the development of
ones, the sap from the weeping tree, which the rubber industry of today. The list, although
originated in Brazil, gives rubber, a material extensive, is still superficial and very personal,
unequalled in its many useful properties. but even more brief is the summary included
here.
With such a wide variety of end products and
constituent chemicals it is impossible to do Although the oldest rubber known is fossilized
other than generalize, but there are some useful and some 60 million years old, the detailed
observations that can be made. modern history began some 2,000 years ago
with the New World involvement in "rubber"
Materials on display must look good, be they in produced from a "weeping tree." It is probable,
a shop window to encourage sales, or on dis- however, that the earliest rubber came from
play in a museum. Surface deterioration is a Castilloa elastica and not the Hevea brazilien-
common event in both environments and can sls, which produces essentially all the world's
sometimes presage surface degradation, al- natural rubber today.
though this is not always the case; some surface
changes are intended and beneficial to the life Skipping past references from Columbus and
of the material. The possible causes and effects Torquemada, the first real European involve-
of surface changes will be discussed, and ways ment was by the French during the middle of
of distinguishing between them presented. the l8th century, when de la Condamine and
Fresneau attempted, without success, to start
Knowing what the material is, and from the the first manufacturing plants for rubber goods
state of its surface what has happened to it so in Europe. Their main interest was in latex dip-
far in its life, consider now what can be done to ping but they could not ship the latex to Europe
optimize its display life. Many ideas have been without it coagulating a visual and olfactory
put forward, some possible and some definitely -
effect similar to milk curdling!
counter-productive. Whilst there is no one an-
swer for all elastomers and their products, For 50 years little happened, except that
many suffer degradation through a series of Priestly coined the name "rubber" for the mate-
common phenomena: heat, light, mechanical rial that rubbed out pencil marks. Then, within
work, pro-oxidant metals, bacteria, oxygen, a very short period from 1820 to 1839, there

60
was a tremendous growth of interest, when in none succeeded in the long term except
the United Kingdom, Hancock invented his ma- Hevea braziliensis and this has not been bet-
chine to convert lumps of solid rubber into a tered, although the American govemment re-
useable homogeneous gum (a process he called search continued into Guyaule right up until
"pickling" to confuse his competitors), and last year, and during the 1930s the Russians
Macintosh developed his three-layer water- used dandelions as a source material, having
proof fabric. In North America, at the same access to nothing else.
time, a substantial market developed for
dipped rubber shoes, whilst Chaffee invented In 1873 Henry Wickham appeared, and the
his mill and calender, the designs of which story of his shipping 70,000 seeds of Hevea
are basically the same as those in use today. braziliensis to Kew Gardens in London has
Chaffee also founded the first American entered rubber folklore, mostly due to Wick-
rubber company. ham's and others' fireside embellishments of
the diaries and notes that Wickham kept on his
By 1839 the American rubber bubble had burst, trip. What rs known rsthat2,397 seeds germi-
mainly due to the stink of crude unstabilized nated at Kew. and 1.919 of these were sent to
rubber as it decomposed in the heat of summer, Sri Lanka, l8 to Java and "a few" to Singapore.
but also accelerated by the economic collapse Yet little is made of the fact that the British gov-
of the country. In the U.K. Hancock and Macin- ernment forgot to pay freight charges and most
tosh survived due to their company's "Macin- of the seedlings died! Here precise details be-
tosh's Waterproof Double Textures." However, come cloudy, but a month later another 100
the product, although better than anything else, seedlings were sent to Sri Lanka from Kew
was still liable to putrefu and, whilst "state of and yet another 100 were sent in the following
the art," it could notbe consideredperfect. year (1877).

In that same year of 1839, Goodyear discov- In 1878, 22 seedlings were sent from Sri Lanka
ered, by accident, that heating a mix of rubber, to Singapore. Ridley, the Director of the Bo-
white lead and sulphur resulted in a highly tanic Gardens of Singapore and the man who
elastic material that was rubber "cured" of its could fairly claim to be the father of the rubber
problems. It no longer went brittle in the cold plantations in Malaysia, believed that these
and soft in the heat, nor did it seem to putrefr were seedlings from a batch of Brazilian seeds
so easily. Thus the process of heating rubber sent to Kew by another collector, Robert Cross.
with sulphur became known as the "curing" These were the seedlings from which most of
process as well as vulcanization, a name the world's rubber trees have developed. Thus
probably coined somewhat later by an other- it is possible that Cross, not Wickham, should
wise unremembered friend of Hancock's. be thought ofas the true father ofthe natural
Mr. Brockedon. rubber industry. In 1899 the first "plantation
rubber" was shipped from Sri Lanka and the
In 1857 Thomas Hancock published his classic rubber production industry was bom. For an-
guide to the U.K. rubber industry and his illus- other 50 years "wild rubber" from trees, not
trations give some idea of the breadth of uses necessarily Hevea braziliensis, continued to
to which rubber was being put. Not many are come out of Brazil. This was often called "Para
missing from a list of today, since these include rubber," after the port of Para from whence it
air-proofproducts, hoses and tyres, nautical, do- was shipped.
mestic and travel equipment as well as a range
of seals and washers. To the following people, therefore, lies the
credit for founding the rubber industry of to-
In the second halfofthe l9th century explorers day: Goodyear, the idealist and fanatic who
around the world were searching for plants that sacrificed his family, income and health in his
could produce "rubber." These were being attempts to do anything and everything with
shipped to politically and climatically accept- rubber; Hancock, the pragmatic businessman
able locations in the hope of discovering one first, and engineer second, who came to grips
that could be "harvested" profitably. In fact, with the mastication of rubber: Chaffee.

6l
unlmown to most today, but the inventor of the general-purpose elastomers, including NR,
still standard two-speed, two-roll mill and its account for some 80% of the current elastomer
extension, the calender. On the rubber produc- market.
tion side is the famous Wickham (or perhaps
less known, Cross), Hooker and Markham at To my mind, the most important features of
Kew Gardens and Ridley at the Botanic Gar- this summary in the area of conservation are
dens of Singapore. Perhaps, furally, the motor the following:
car should be mentioned, hardly conceived dur-
ing the period covered by those listed above, r Any artefact manufactured before 1840 is not
but the undoubted catalyst to launch the rubber lulcanized.
industry into the 20th cenflrry on an increas-
ingly large scale. e Identification of curatives/additives will help
in dating products manufactured between
In this century a few critical dates can still be 1840 and 1920.
picked out from the great mass of scientific
advancement: o Identification of the polymer is essential
after 1930.
o 1921to 1923 - The development of most
of the chemicals used to "accelerate" and r Non-black filled products were not protected
clean up the chemistry of sulphur with antidegradents before the mid 1950s.
lulcanization
Obviously, more detailed datings may be possi-
o 1925 - Can be considered the start of the ble by considering all ingredients found, and
synthetic rubber industry thus it is worth considering what can be done in
this field. I have not gone into scientific detail
o 1930s - The introduction of amine here, as knowing what can be done, not how to
derivatives as antioxidants do it, is the important thing and analytical de-
tails are not needed to ask the experts to look
o 1933 to 1945 - The development of familiar for something.
synthetics, such as polybutadiene, styrene-
butadiene, nitrile, chloroprene (neoprene), Aspects of Analysis
polyurethane and silicone
Excluding surface effects, which are considered
o 1950s - The introduction ofnon-stainins later, useful information can be obtained
phenolic antioxidants through these three questions:

o 1958 - The introduction of fluoroelastomers o What is the polymeric system?

o 1962 - World production of natural and o What can be said about the cure and
synthetic rubbers each reaches two million protective systems?
tons per annum
r What inorganic materials are present?
o 1990 - Worldoutputof rubberaround
l5 million tons; 2:l synthetic:natural The first is most easily answered by pyrolysis
followed by either infra-red spectroscopy
It is worth noting that, whilst many special (P-IR) or gas chromatography (P-GC). Charac-
elastomers have been developed that can oper- teristic traces that are obtained for each poly-
ate in environments where NR would not func- mer using both techniques and polymer blends
tion satis factorily, none of the general-purpose can be quantified ifthe analytical conditions
elastomers can offer any improvement over are right. Typical patterns for NR are illus-
NR for run-of-the-mill applications. These trated in Figures I and2. These techniques
suffer from one disadvantage in that they

62
cannot be considered non-destructive. Neverthe-
less, P-IR can use only l0 mg whilst P-GC
requires, say, 100 pg and this can usually be
spared from any material. Figure 3 shows a
sample ready for P-GC analysis; the baby
feeder teat gives an idea ofthe scale. The
equipment need not be highly sophisticated I i r'
l

and costs would be 10,000 to 15,000 dollars. .il,

iii, i. l

Examination of the cure and protective systems

is potentially the cheapest of all analytical
techniques, if a gram or two of material can be
spared, since thin layer chromatography (TLC) Figure 2 Pyrolysis gas chromatogram of natural
can be used to separate and visualize virtually rubber. -
all of the candidate chemicals. The sample is ex-
tracted with a suitable solvent, and this solution specific for oxidative degradation of NR, giv-
is then concentrated and spotted onto the ing an orange spot and streak, whilst another en-
bottom of a TLC plate. This is then placed in a ables a distinction to be made between NR and
special tank containing a low volume of a liq- the material that is its nominal synthetic equiva-
uid. which elutes the various components up lent, generical ly call ed " PI" for polyisoprene.
the plate at differing rates. Treatment of the
"developed" plate with particular chemicals pro- If the sample size is limited to milligrams, high
duces variously coloured spots from the compo- performance liquid chromatography (HPLC) is
nents present, which enables identification of the analytical method of choice. This method
the materials with a high degree of specificity, can give as much data, and quantiff it, but costs
provided reference materials or data are avail- more 30,000 dollars upwards!
able. Thus, for a few cents, much useful -
information can be obtained. It may also be It should
be noted here that I have talked only
ofinterest to note that one spray is relatively ofsulphur vulcanization and not ofany other
t1pes. Since the discovery of modem accelera-
tors in 1920, sulphur vulcanization has domi-
nated the methods of crosslinking polyolefins
and probably accounts for over 95o/o of the
market. Peroxide, urethane, and radiation

ll
tl

Figure I Infra-red spectra of a natural rubber Figure 3 Sample being inserted into a pyrolyser coil
pyrolysate (top) and thinfilm (bottom). for P-GC identific ation.

63
crosslinking of polyolefins must be considered be taken to examine a piece Iiee of surface con-
"niche" areas for very specific properties. tamination or, more safely, a section through
Considering the whole elastomer field and the bulk of the material. These instruments are
pseudo-elastomer field, there are two signifi- not cheap but my experience in the U.K. is that
cant exceptions to sulphur wlcanization, the most museums know of their potential and are
pollurethane and silicone rubbers. These have able to persuade either a local university or in-
multi-functional groups in a few molecules of dustrial organization, such as my own, to have
low molar mass that enable three dimensional tu*:,""*u-tn"o''0oe@e
networks to build up during pollmerization.
Here the word "l'ulcanization" is not appropri-
3200
ate whereas "curing" might be. In fact, this OX€s A-4r Pa&b *m!ed aom
Conb'Mi€d L&lDr €atch
2aaa
latter word is used by the resins and plastics in-
dustry quite happily, and reasonably, since their 24

crosslinking processes also cure the sticky and

I

flow problems of the starling materials.

1

12gO I

Finally, any inorganic elements present must aoo I

be considered and this includes any added sul-

phur that would indicate a cured product. For
400
sl ct K
1r
1.1\
simplicigz, and even the chance of fiue non- ot23456789t0
destructive analysis, the scanning electron Figure 5 X-ray spectrum of contaminant with, inset,
microscopy (SEM) with a built-in electron mi- a micrograph of the particle.
croprobe X-ray analyser must be considered.
The X-ray spectrum will immediately identify Such a technique was of instant importance
the elements present and allow a "guesstimate" a few years ago when I was asked to identify
of their levels. Figure 4 shows the spectrum of the "elastic" material found inside a sealed am-
a wlcanizate with zinc and sulphur together phora during a Middle-Eastem excavation (Fig-
with aluminium and silicon, the last two indicat- ure 6). lt was assumed to be a seal of some soft
ing a clay filler. The characteristic peaks ofbar- and, whilst having a brittle skin, was obviously
ium suggest barium sulphate, used to increase "elastic" or rubbery undemeath. A minute
the density of cerlain materials, and there is sliver was taken from within a whorl and identi-
also a trace of iron. In Figure 5 an impuriqz is fied by P-GC as NR. This was exciting as there
identified as rust; it must be emphasized that was no record of rubber being used outside
this is a surface technique with a sample pene- the New World prior to 1500, although with
tration of about 10 microns. This feature is ob- dandelions and the many hundreds of other
viously relevant to the next section; in the rubber-bearing plants aiound. it was not impos-
context of bulk element composition care must sible. SEM X-ray analysis threw the "cat
amongst the pigeons" as this showed both zinc
"t,
and sulphur at "modem" levels of addition.
:"" , , HPLC examination of a further sliver showed
mercaptobenzothiazole (MBT). This placed the
artefact after 1920 and thus raised doubts about
the other artefacts found in the vicinity of the
amphora. It certainly caused local red faces, but
at least prevented a much larger scale of embar-
rassment. There are many different sources of
analytical techniques that can be used to exam-
ine elastomers, and one of which I am co-
author, listed in the bibliography, discusses
these in some detail. Many useful tricks are
Figure 4 Scanning electron microscope documented there and perhaps there is one that
- energ,,
dispersive X-ray spectrum of afilled rubber deserves mention here, as it is cheap! All that is
vulcanizate. required is a press with platens that can be

o+
 chemical that is intended to bloom is wax, the
sole reason for its adding being to bloom and
produce an ozone-impelious film on the sur-
face of the rubber product.

Modified blooms result from protective agents

on or near the surface of the rubber reacting
with the environment and undergoing a chemi-
cal change. The non-reacted chemical in the
bulk of the rubber then migrates to the surface
to decrease the concentration gradient, where it
is then further consumed, eventually building
up a protective skin. At this stage the migration
stops until the skin is removed or broken and
"repairs" are needed. Classic examples of this
heated to about 175'C. Hot pressing non-'urrl- are the paraphenylenediamine antidegradents.
canized rubber will result in a smooth film, but
if the rubber is rulcanized it will stretch and Pseudo blooms are what I call dull surfaces,
tear before shrinking to something like its origi- which appear to be suffering from a bloom but
nal size. One important application is in distin- which on closer examination, often under an
guishing between vulcanizates and such electron microscope, are actually suffering
materials as plasticized PVC (arguably called from surface degradation of the rubber. In very
"rubber") and the thermoplastic elastomers, extreme cases, sometimes called "chalking" or
which are taking an increasing share of the
"frosting", the rubber is completely eaten away
current elastomer market. leaving a powdery surface of the inorganic
filler. This has led some people to claim that
Surface Effects fillers can bloom, but this is not possible as
they are insoluble in rubber. They only "come
Examination of the surface of a rubber product to the surface" because the surface is eaten
or artefact immediately reveals whether the sur- away. Also included in this category is surface
face looks good and "respectable" or whether
crazing,to which I will refer later.
there is something wrong with it. If the latter is
true, there could be several reasons and these Surface contamination, staining and
need to be identified before carrying out any discolouration are always problems, particu-
treafinent. The following categories of surface larly with the sort of materials conservators
effects can be defined. But before discussing might be examining. These have to be identi-
each in tum, I must emphasize that it is possi-
fied, possibly by SEM, before taking action to
ble for more than one of these effects to be restore the material, remembering that any
present at the same time.
"clean-up" procedure must take into account
surface effects or defects if fuither damase is
A true bloom is a thin layer on the surface of a
to be avoided.
rubber article of one of the chemicals added
before, or produced during, lulcanization. This The final category, hazing, only applies to
chemical actually migrates from the bulk to the transparent or translucent products and is
surface by a mechanism that will not be dis- caused by either insoluble crystals, particles or
cussed here beyond saying that the chemical immiscible liquid droplets suspended in bulk
has to be soluble to a ceftain extent in the rub-
rubber. These may or may not migrate as well,
ber and be present at a level above that solubil- but little can be done about them. Although not
ity. Chemicals that can bloom are sulphur, zinc a surface effect, it is not always possible to dis-
dialkyldithiocarbamates, mercaptoben zthiazole tinguish between the two with a cursory exami-
and zinc mercaptobenzimidazole. With the pos- nation. Tables I and II illustrate some simple
sible exception ofthe last, they are not intended tests that can be used to provide information
to bloom in a properly formulated product. One

65
on possible surface effects. It is usually possi- area, once again I recommend IhebookThe
ble to find some part of an article that can be Analysis of Rubber and Rubber-like Materials,
experimented on, but as I mentioned earlier, as listed in the bibliography.
more than one effect may be present at any one
time and therefore a degree of care is required Ageing
in drawing any conclusions. The tests them-
selves are perfectly straightforward and do not In any discussion about ageing it is enlight-
require any further elaboration. ening to begin by quoting from Hancock's
classic book The Origin and Progress of the
A more sophisticated examination should be CAOUTCHOUC, or India-rubber Manufacture
carried out ifpossible and it is obvious that the in England, which was published in 1857. The
SEM with X-ray analytical microprobe must be following observation seems to have been
the technique of choice, although some informa- made sometime during 1825 to 1826:
tion is available from chemical spot tests and
multiple internal reflection infra-red spectros- The injurious effects olthe srm's rays upon thin
copy as well as IR microscopy. Once again, films of rubber we discovered and prolrded agarnst
before much damage accrued. All these things are
there are many fricks that can be used to facili-
now cheaply known to those who have followed us
tate examination, but these are beyond the by men leaving our employ and the specificatrons
scope of this paper. For more details in this of our patents; but they had all to be undergone in

Table I
Simple Tests to ldentify "Bloom" Types

1) Did bloom develop in storage?

I

yes
I

I
Surface contamination or inorganic haze
v
2) Does it disappear on heating?

yes Oxidized antidegradant,

Surface degradation,
Basic zinc stearate
I

it
3) Can it be removed with solvent wipe? Basic zinc stearate
ri
vlno
r

I
yes
Try other.o,u"nrl
yes no
I
yes
''l
Try complexing agent-t
I

I I

J lro
V
True bloom <- Haze due to organics

66
 our early progress at an enorTnous cost, as well as not prepared (quite reasonably) to compromise
trouble and vexation: and none but those who have
passed through the ordeal can concelve the mortrfi-
this performance so that the product might be
catron experienced during those years. displayed for a few hundred years in a museum.
These effects may also interact so, with the
It would appear that there really is nothing new exception of the protective agents themselves,
under the sm! Unfortunately there is a great dif- none will be discussed in any detail. Their
ference between knowing about something and significance will become apparent as the
circumventing it particularly for the time discussion continues.
-
scale that interests museums. The factors that
can influence the life of an elastomeric product Ageing processes involve three distinct mecha-
fall into fwo categories. which I will cafl "prod- nistic routes that can be identified as follows:
uct characteristics," about which nothing can be
done, and "ageing processes." about which o continuing sulphur .vulcanization chemistry
something can be done.
. shelfageing
Product characteristics describe the product as
it was made and include such variables as: . atmospheric ageing
which elastomer was used; whether the product
is filled or not and if so, whether it is black, col- Sulphur chemistry is particularly important in
oured or white; the vulcanizing system; the pro- lulcanized products that are resisting a distort-
tective system; whether it is properly designed ing force. Chemistry does not stop when the
and manufacfured or has built-in stresses; and vulcanizing temperature is reduced from some
the size and shape. 150'C to ambient it just proceeds more
slowly. One of the-features of a r.ulcanizate is
All of these have an effect on the product's known as its "set" and this is due to the fact that
service and storage lives but unfortunately the C-Sx-C bonds, which form the chemical
there is a conflict in that the manufacturer is crosslinks between polymer chains, can break
concemed with optimizing the product's per- between sulphur atoms and then the loose ends
formance during an acceptable service life with re-attach to either their original partner or an-
some pre-use storage, but the manufacturer is other loose end. Ifthe vulcanizate is under a
distorting force, then, as the bond breaks, the
elasticity of the polymer chain causes it to relax
against the force so that the bond will re-form
Table II further down the chain. The result is a slow
Colour Changes and Their Significance in "slippage" of the lulcanizate, which cannot be
"Bloom" Analysis reversed. This can be minimized by the appro-
1) ls the colour change light-rnduced2 (Mask some
priate choice ofcrosslinking agents, since the
of sample for reference) so-called "effi cient" vulcanizing systems.
Grey/Brown amrne a,/o which are low in sulphur and high in accelera-
Prnk phenolic a,/o
tor, have a much higher proportion of monosul-
2) ls the colour change heat-rnduced? phidic crosslinks than "conventional" cure
lf yes, wtth
systems. As the C-S bond is stronger than the
strckrnesyembnttlement oxdatNedeoradatton
lf yes, wtthout S-S one, it will not break so easily, but, once
strckrnesyembnttlement amrnea,/o there is a wlcanizate, its propensity to set can-
3) ls the colour change patchy?
not be changed. It is essential, therefore, that
Reactron of external contamtnatton wtth vulcanized (and of course non-vulcan ized) arte-
rubber chemtcals facts are stored and displayed with as little dis-
4) Does colour change appear after laundering or torting forces as possible. Vehicles, for
contact wth metals? instance, should not be left on inflated tyres.
Copper/lron reactton wtth dtthocarbamates but should be blocked up and the ryres ieflated
5) Does colour change accompany poor agerng? to a few pounds per square inch (psi). Although
Further evidence for Copper/lron contamt natton the chemical changes will continue in an un-
stressed object and are significant in terms of

67
product usage, they will not significantly affect In real life rulcanizates, particularly when
the appearance ofa display object. black-filled, it seems that the ingress of oxygen
is limited and, apart from the surface 0.5 mm or
Shelf ageing is basically oxidative ageing and, so, the bulk rubber remains in excellent condi-
apart from the obvious influence ofoxygen, the tion. It is this experimental fact that enables
catalytic effects ofheat, light and pro-oxidant conservation scientists to predict with confi-
metals have to be considered. If the chemistry dence service lives well in excess of 100 years
of oxidative degradation was considered in de- for engineering products, such as base isolation
tail it would fill in a lifetime or two and still units for buildings, or bridge bearings that take
leave unanswered questions, but in a simplistic up the expansion and contraction of most mod-
form, it is fair to say that oxidation of a sulphur em bridges.
lrrlcanizate proceeds via at least one chain reac-
tion sequence. This, fundamentally, introduces Light-catalysed oxidation is particularly no-
C-C and C-O-O-C crosslinks between polymer ticed in unfilled or light-coloured objects and
chains as well as C-O-O-C rings within the can result in an inelastic skin which, as it
same polymer chain, whilst another set of thickens, cracks in random directions and pro-
reactions between the sulphur atoms of the duces a pattem known as "crazing" and this is
crosslinks and oxygen breaks the crosslinks. illustrated in Figure 7. In thin sheets the effect
The mechanism of chain scission is still not has been called "light stiffening", whilst in
completely understood but theories have been highly filled articles the degradation can result
proposed that involve both the "in-chain" in complete loss of the resinified elastomer,
C-O-O-C groups and the C-O-O-C crosslinks. ending up with the "chalking" effect that I men-
The reactions between sulphur and oxygen are tioned earlier. Because of the lighrabsorbing
those that eventually lead to sulphuric acid for- properties of black-frlled materials, this effect
mation, which is a particular problem with is not normally seen, although a "bronzing" ef-
ebonites. fect has been claimed to result from light-in-
duced ageing of some very smooth-surfaced,
Therefore, there is alarge range ofboth sequen- black-filled products. I must emphasize that
tial (or chain) reactions and competing reac- light itself does not cause degradation of rub-
tions, and the ones that predominate depend on bers; oxygen must be present and the effect is
factors such as the formulation or composition light-catalysed oxidation. The effect varies with
of the rulcanizate as well as the influences of wavelength, UV being the most dangerous and
heat, light and metal catalysis. Heat ageing bal- red/oranse the least so.
ances the rate of reaction of oxygen with the
elastomer and the rate of diffusion of the oxy-
gen into it. If the temperature is relatively low,
ithas been postulated, for an unprotected elas-
tomer, that diffusion through the elastomer pre-
dominates and therefore there is slow oxidation
throughout the product, but at higher tempera-
fures (e.g., >80'C) the rate of oxidation be-
comes much more important than the rate of
diffusion, so substantial oxidation occurs on
the surface and an oxidized (hard) surface skin
forms. As oxidation continues, the chain break-
down may become more significant and the
hard surface can then soften and tum sticky. To
complicate matters further, under certain condi-
tions this order can be reversed and an initially
sticky degraded surface can harden with further
Figure 7 Crazing pattern induced on a thin white
oxidation.
disc ofrubber.

68
The effects of trace levels of pro-oxidant met- cracks perpendicular to the stress in the side
als, such as copper and manganese, have been wall, whilst the white strip in the side wall,
known for years and were used by Goodyear in which must have had an identical life, shows
I 837 to improve his sticky raw rubber by paint- the random crazing ofoxidative degradation.
ing the mouldings with a concentrated solution
of copper nihate. The "improvement" was actu-
ally a hard oxidized skinl These metals acceler-
ate the oxidative degradation, again forming
brittle skins on both raw and l.ulcanized rub-
bers as well as evenfually degrading the elas-
tomer to a sticky mess, as illushated in the
baby feeder teats in Figure 8. There has been
much dispute as to how much copper an un- \\}
filled r,ulcanizate can reasonably tolerate with-
out specific protection. There is no doubt that it
depends upon the "state" ofthe copper and
"state" probably relates to its chemical form
and degree of solubility. Certainly 10 parls
per million copper is sufficient to destroy a
product, as Figure 8 shows. Figure 9 An aircrafi white wall tyre showing crazing
in the white region and ozone/flex cracking in the
black region.

I have not discussed ageing by work, other than

flex cracking, as in most exhibits this will not
be valid. However, once movement is intro-
duced into an elastomeric product, complica-
tions can occur none of which is beneficial
to the life of the-product. The two dangers for
any dynamic exhibit are first that any protective
layer of oxidized rubber, antidegradent or wax
is cracked, in which case oxidation will pro-
ceed very rapidly from the location of the
crack, and second that a cycle ofwork on a
thick piece of rubber can result in heat build-up
within the rubber, which is eventually dissi-
Figure B Babyfeeder teats, one degraded through pated by catastrophic failure ffom within the
high copper contamination. bulk, as shown in Figure 10. Obviously this
should be considered and investigated prior to
Atmospheric ageing is different from shelf initiating any "dynamic" exhibit. Indeed, in any
ageing in that it is characterized by the attack such exhibit it may well be appropriate to re-
of ozone on the rubber. It is often called flex place the moving elastomeric components with
cracking or ozone cracking as detailed examina-
tions of the types of cracking have shown the

il"'ss
same pattem. It is interesting to note that this
type of cracking occurs more readily at night
than in the day as light-catalysed oxidation
tends to obscure the ozone cracking. Following
from this, ozone cracking will occur in
black-fi lled materials, where light-catalysed
oxidation is minimal. A classic picture in Fig- Figure I0 Heat build-up test pieces showing
ure 9 illustrates an aircraft fzre where the top c a tast rop hic i nte rna I fa i lure.
(black) area shows the typical ozone-induced

69
modem replicas. For a much more exhaustive can have half-a-dozen trade names. The ones I
account of the ageing of rubber, please refer to shall use are those familiar to me in the U.K.,
the Malaysian Rubber Producers' Research but in books, such as the BRMA toxicity guide
Association publication Natural Rubber, to rubber chemicals listed in the bibliography,
Science and Technologt, as detailed in the they can be related to their equivalents from
bibliography. other manufacturing sources.

Conservation Options The main difference between the amines and

phenols is that the former stain to a lesser or
After that brief glimpse of what is going on greater extent and the latter do not again
during the life or storage of an elastomeric within limits (see Table III). -
product, consider now what can be done to pro-
long its display life. In truth, options are re-
stricted to adding protective agents, coating the
product with an impermeable membrane or re-
Table III
Amine and Pherol Antidegradents for Rubber
moving the cause of the trouble oxygen and
ozone. Assuming that the light is- adjusted as in- The amines consist of four basic categories:
dicated, the temperature kept low, and all un-
necessary sources ofstress removed, these keto-amrne good antioxrdants ELEZS
actions still will not stop some of the sulphur condensates
chemistry liom continuing. But, as pointed out qurnoil nes good antroxrdants Flectol H
earlier, the sulphur chemistry will not, by itself. poor ozone resrsrance

harm the exhibit for several centuries. substrtuted good antroxrdants DDA
drphenylamrnes some ozone resrstance Octamrne
substttuted good antroxrdants IPPD
There have been well over 1,000 patents issued p-phenylenediamrnes good antrozonants Wrngstay 100
conceming protective agents for rubber vulcani-
zates antioxidants and antiozonants but The phenols can be divided into four categories:
there - - situ-
are relatively few on the market. The mono-ohenols moderate antroxrdants lonol/BHT
ation is complicated by the many trade names least starnrng
used, but for chemically different materials, the polyhydroxyphenols moderateantroxrdants Wrngstay L
relatrvely non-statntng
number is probably no more than 100. Because
of the cost of introducing new chemicals and bls phenols good antioxidants 2246
possrble starnrng
testing against various consumer regulations, it phenols good antroxidants Santowhtte
throbrs
is not likely that any new wonder material will weak antrozonants crystals
be produced. Current efforts are aimed at meet- slrght stainrng
ing market requirements through blends of
chemicals, but remember, "market" require-
ments are not the same as a museum's! If the intention is to add one of these materials
to a vulcanizate then there is little option other
Essentially the antidegradents in use today can than to dissolve the chemical in a suitable sol-
be divided into two types, those derived from vent, immerse the vulcanizate and wait for dif-
aromatic amines and those from hindered phe- fusion to take place. As long ago as 193 I this
nols. There are a few others such as phosphites, was proposed as a method of giving protection
used with phenols in the plastics area as there is to objects made without added antidegradents,
a synergistic effect, and mercaptobenzimida- but it never found commercial favour and,
zole (MBI), which is effective against metal although I have seen no experimental ageing re-
catalysis. Again, this is used in conjunction sults based on this method, I am not surprised.
with a conventional antidegradent. The usual idea is to use a solvent for the antide-
gradent that will swell the rubber surface and al-
For simplicity, trade names will be used in this low diffusion of the antidegradent into the bulk
section. These are easier than many of the of the rubber. This may be acceptable for a new
chemical names that take a line or two of print, wlcanizate, but if the object is not vulcanized,
but have the disadvantage that one chemical or is degraded so that the surface is structurally

70
weak, it will either dissolve or swell to such an be built up by topical application, provided the
extent that the surface will be severely dam- limitations of cracking, repair and surface
aged. The rates of diffusion of antidegradents finish are appreciated.
are quite slow this is essential to their opera-
tion so they-will also take a long time to The phenolic antidegradents are not considered
- into the rubber and. on this time scale.
diffuse antiozonants, only antioxidants. When oxygen
much damage could be done to the network, attacks a rubber molecule various chain reac-
particularly since the solvent will be extracting tions occur, which result in, among other
whatever it can from the lulcanizate whilst the things, polymer chain breakage as well as the
antidegradent is diffi.rsing in. If this method is insertion of further crosslinks. The phenols of-
used for long-term protection, then a non-sol- fer an altemate path in the chain reaction se-
vent for the rubber should be used, methanol quence, which stops the chain breakage fiom
perhaps, as the antidegradent will still diffirse progressing. They do not stop chain reactions
into the rubber, but there will be less extraction from starting, so their effect is, at best, to slow
from the bulk of existing chemicals. There down the oxidative breakdown by perhaps five
will, of course, still be dissolution of any sol- times. They do not form a protective skin if oxi-
uble layer and, ifthere is a stable surface skin dized on the rubber surface and they need to be
already, either of oxidized rubber or insoluble intimately dispersed/dissolved in the rubber to
oxidized antidegradent residues, this may function.
be impervious to the antidegradent being
added! Any physical barrier will stop or reduce oxy-
gen/ozone diffusion into the rubber and might
If the choice is not to diffirse protective chemi- or might not crack depending on its brittleness.
cals into the unprotected rubber, there is still Again, its capability to self-heal will depend on
the option ofadding a protective surface coat- whether it can migrate from the bulk or
ing, but before considering the various treat- whether it is an added surface layer.
ments that are available, one point must be
made. Antidegradents can be considered to There is also the problem of light coloured arti-
operate by physical (barrier) means, chemical cles and UV catalysed oxidation. As with plas-
reactions, or both. It is crucially important to tics, UV stabilizers can be added to the bulk
understand that the different types ofantide- product or in a surface furish. As mentioned,
gradents work in different ways and thus give light by itself does not degrade rubber, it
visually different surface finishes. catalyses oxidative degradation.

The p-phenylenediamines (PPDs) firnction as Finally, there is the "tart-up" factor. Treatment
both antiozonants and antioxidants and at least with a wax, polish or silicone oil can make a
one of their modes of operation is to migrate or rough degraded surface look beautiful just by
bloom to the surface of a product where they re- reducing the light scattering. The coating mate-
act with the oxygen and/or ozone to form a brit- rial will also have some barrier effect and so
tle skin, which increases in thickness until it might prolong the life a little, but there will be
presents an impermeable barrier to the gas. Any no great long-lasting effect.
damage to the skin, such as by cracking, is re-
paired by further blooming. There are many dif- There is one word of warning about any topical
ferent PPDs and one of the main reasons for application; as mentioned earlier, any surface
selecting a particular one is its solubility in the disruption by solvent swelling/drying will po-
polymer system being protected. This affects tentially be disruptive and could do more harm
the rate of blooming, which is crucial to long- than good. Ifthe product is black and protected
term protection. PPDs oxidize to blue/pur- with PPDs (i.e., post early 1930s) or non-black
ple/black materials and readily stain; this and made after the mid 1950s, thus probably
causes no problems with a black-filled material containing phenolic antioxidants, any protec-
that is not in contact with a light coloured one, tive layer could be washed away by a solvent
but obviously would be catastrophic for, say, a and this would not be a good idea. If the inten-
white thread in underwear. A surface skin can tion is to apply any such application, it should

7l
be the mildest water-based emulsion or a and still perfect, its projected life in the same
relatively dry spray of a solvent-based system. environment must run to many hundreds
ofyears.
Another well-established barrier technique is
the coating of the product with an oil lacquer It is obvious that none of these methods meets
that is then lulcanized. The procedure is de- the conservator's ideal of doing nothing irrecov-
scribed in detail by Gottlob, and Figure I I erable to the samples, but in the chemistry area
shows rubber shoes over 70 years old that were at least, there are few options. lndeed, ifthe
given this finish and that are immaculate today; sample is light-coloured and thin, the additional
they were stored underwater, which may be strengthening supplied by a rulcanized coating
relevant. The electron micrograph illustrated in might be a considerable advantage. After all,
Figure 12 clearly shows the layer oflacquer on old Master works are revarnished. whv not
the surface (running vertically slightly left of old elastomers?
centre) and, as this layer is over 70 years old
Since the degradation to be prevented is caused
by oxygen, and to a lesser extent by ozone, the
obvious answer must be to prevent these gases
from reaching the elastomer. In the absence of
a barrier layer the exhibit must be in an en-
closed vessel that can either be evacuated or
filled with an inert gas. Evacuation is not to be
encouraged because ofboth the potential
danger should the case implode and the possi-
bility of chemicals being volatilized from the
Figure I 1 Rubber overshoes recoveredfrom a ship rubber and condensing on the glass. Nitrogen is
sunk some 70 years ago. cheap and easy to acquire, and a low pressure
gauge. showing that the pressure is one or
fwo psi above atrnospheric pressure, will
give an easy check on leaks.

The shoes discussed earlier are by no means the

only case of rubber being recovered from under-
water. The fyre and inner tube of a Wellington
bomber lost during the Second World War in a
Scottish loch were recovered and the tube in-
flated with no problems. Similarly, a seal
from a hatch of the first British submarine 11ol-
land I , lost during trials in l90l , was found to
be in perfect condition under a layer ofsurface
oxidation. The oldest industrial sample I have
ever examined was a sewer seal, installed in
l86l and removed during the replacement of
rusted iron pipework in 1963. The sewer is il-
lustrated in Figure 13, with the seal just visible
under the rubble as a gentle arc in the fore-
ground. In spite of the distinctly hostile envi-
ronment it had been in, and the surface loss
through abrasion, it was still a good seal and
would have lasted many more years if only
Figure I2 Scanning electron micrograph ofcross- the pipework had lasted. Storage under cold
section through the rubber layer ofthe overshoe water might be a viable preservation option
showing the lacquer as a smooth /ilm about although I have never seen it suggested as a
I 5 pm thick running venically just left of centre. deliberate ploy.

72
The one essential point
is that whatever mode
of preservation is attempt-
ed, the object should be
mounted as it is intended
to be displayed, as soon as
possible after the treafinent
or even before, and then not
touched, flexed or otherwise
distorted. Obviously any
mounting should be as
stress-free as possible.

Conclusion
In summary, therefore,
the only real danger to elas-
tomeric artefacts displayed
in museums is oxygen, al-
though light, heat and ozone
might have a part to play in
Figure I 3 An I86l sewer withJunctional rubber seal arcing across the
some environments.
foreground.
There is no wonder chemi-
Acknowledgement
cal that will prevent oxidation nor is there
- recom-
even one that could be universally The author thanks the Board of the Malaysian
mended for all elastomers is there ever Rubber Producers' Research Association for
likely to be. -nor
permission to attend Svmposium '91 andto
publish this paper.
The best that phenolic antioxidants can do is
slow down oxidative damage by a factor of Figure 7 first appeared in Revue Gtntrale du
about five. What this means depends on the in- Caoutchouc, 3 l, 1954:. 419 and is reproduced
dividual article and its initial life expectancy. with permission of that joumal.
Black products ofa reasonable thickness Figure 9 first appeared tn Transactions Institu-
(i.e., >5 mm) should cause no problems over
tion ofthe Rubber Industry,2l,1945:49 and is
many decades and probably centuries, although reproduced with permission of the Plastics and
the surface may appear to deteriorate margin- Rubber Institute.
ally. There might be a benefit from coating
these products with a mix of wax and an
R6sum6
amine antioxidant.
Le caoutchouc : son histoire, sa composition
Light-coloured or non-vulcanized articles are at et ses perspectives de conservution
great risk and should be barrier protected. This
must be the best option for long-term protec-
Quiconque vit dans ce monde dit < moderne >
tion, and the best barrier is a sealed case and an connaitfort bien le caoutchouc et ses propri'ltds.
inert atmosphere. If the storage temperature can Et c'est sans doute justement parce que cette ma-
be reduced to2"C to 5oC. this could be an tidre nous est sifamilidre que nous I'abordons,
added advantage. sinon avec un certain mtpris, du moins avec une
certaine indifflrence, malgrt la place qu'elle

t)
 Appendix
Natural Rubber: History of lts Industrial Development to the 20th Century

lst Millenium Mexico Ball courts/fi gurines holding balls.

B.C.

Aztecs 6th Cent. South Balls, dipped feet -> shoes, coated fabrics.
Mayas & possibly America Pictures copied in National Museum, Mexico.
earlier.

l0th Cent. SW US Ball courts and ball games.

->
Objects in Peabody Museum, Harvard
University.

Columbus 1493 Haiti First European recorded to have seen

rubber balls and retumed some to Europe.

Torquemada t6l5 Mexico Taught Indians how to waterproof cloth

and make dipped goods.

Charles de 1735 to 1740 South In the Andes, described how Indians "milked"
la Condamine America trees for liquid to waterproof fabrics. The
lndians called the tree "HEVA" and the sum
from the liquid "CAHUTSCHU." de la
Condamine christened the "milk" "LATEX."

Fresneau 1743 to 1746 French Realized the potential of the material and
Guiana infected France with enthusiasm for rubber
research. The problem was that latex could
not be shipped to Europe without 'going bad'
and solidiffing. Tree was Hevea braziliensis.

HerrisanV 1761 France Found dry rubber would dissolve in turpentine

Macquer but the resulting dried film was sticky and
soft.

Macquer 1768 France Replaced turps with ether and cast strong
films that were not sticky.

1769 France Made riding boots for Frederick the Great by

multiple dipping process.

Priestly t770 UK Noted that an artist's shop in London sold a

(ofoxygen half inch cube of material for erasing pencil
fame) marks for three shillings. He called it 'INDIA
RUBBER" having found from whence it came.

Roberts/ t790 UK First patent referring to rubber rubber

Dight - oil painting.
solution for treating canvas before

Hummel t 8l3 US Gum elastic vamish first US patent that

mentions rubber. -

75
Thomas 1819 UK First saw rubber.
Hancock

1820 UK First patent for dry rubber; cut strips for

elasticating clothes, braces, etc.

1820 UK Invented his "pickling" machine, which enabled

dry rubber to be worked into a "dough."

1820 US Dipped shoes appeared in the US, made in

South America, exported to Paris gilded,
-
"fashioned" and exported back to America.

by 1823 US Direct imports from Brazil.

Charles 1820 US Born.

Goodyear

Macintosh 1823 UK Realized that if fabric coated with rubber

solution then had another layer of fabric
applied to rubber, the three-layer sandwich was
waterproof and not sticky
-"MACINTOSH."
1825 UK Pitch/rubber solution
->
sheets for coating
ship bottoms, etc.

Faraday 1826 UK Established empirical formula as CsHs.

1825 to 1830 UK Dozens ofuses.

by 1830 US Over 500,000 pairs of rubber overshoes

had been imported.

Chaffee l83l US Rubber/turps/lampblack paint to waterproof

leather.

1832 US Roxbury India Rubber Co. founded (first US

rubber company).

1834 to 1836 US Invented the two-roll, trvo-speed mill

and then enabled it to be heated. Still the
standard mill of today.

Hancock/ 1834 UK Formed Chas. Macintosh and Co.

Macintosh

Chaffee 1835 US Invented the three-roll mill for

"calendering," which is still the basic
procedure today.

Goodyear I 834 to I 83 5 US Became intrigued by rubber

-
some say
obsessed.

76
Hancock 1837 UK Invented the spreader, the standard
coating machine of today.

by 1837 US Economic crisis. Rubber "bubble" burst.

I 839 US Existing rubber industry in the US

finished but 500,000 non-vulcanized shoes
per annum still coming from Brazil.

But in the same vear:

Goodyear I 839 US Left a mix of rubber, sulphur and white lead

on a hot stove and the resulting material was
"CURED" of all its defects. No longer
softened on heatinglhardened on cooling
and lost its stickiness.

l 840 US First commercial vulcanized material rubber

thread for "shirred" cloth. -
Goodyear I 843 US First application for US vulcanization patent.

Hancock 1842 to 1843 UK Identified sulphur in a piece of Goodyear's

cured rubber. Could not duplicate "CURE" as
did not know about white lead, but effectuated
cure with rubber/molten sulphur.

I 843 UK Produced "Hard rubber" (Ebonite) with long

treatment of rubber with molten sulphur.

l 843 UK In November Hancock obtained UK prov.

patent.

Goodyear 1844 UK In February UK patent application refused.

Parkes 1846 UK "Cold cure" process discovered

(sulphur chloride).

Thompson 1 845 UK Patented the pneumatic tyre but no vehicles

suitable to make it a commercial successl

I 851 US Vulcanized rubber shoes being manufactured at

a rate of over 5 million pairs per annurn.

But in the UK there was more interest in vulcanized "Macintosh', material:

Hancock 1857 UK Published his "Personal Narrative. "

Williams 1860 UK Decomposed natural rubber and isolated

isoprene (CsHs).

77
Thompson 1 861 UK First commercial solid vulcanized tyres.

I 861 ? "Hard rubber" frst called Ebonite.

Murphy I 870 US Recognized "oxidation" as cause of

deterioration in rubber.

Collins 1872 UK Commissioned to report on Rubber in Brazil.

Wickham 1813 Brazll Commissioned by Kew to collect seeds of

Hevea braziliensis.

1 875 Brazll Dispatched 70,000 seeds to Kew:2,397

germinated.

1876 Singapore 50 seedlings arrived. Died due to neglect.

Cross? 1877 Singapore 22 seedlings arrived and survived. Basis of

virtually all Asian trees today.

Bouchardat 1879 France Re-polymerized isoprene to "rubber. "

Daimleri l 885 Germany Invented the motor car.

Benz

Ridley 1 888 Singapore Took over Botanic Gardens. Began a one-

man crusade to develop plantations.

Dunlop 1888 UK "Reinvented" the first pneumatic tyre but now

bicycles and vehicles available to use it.

Bartlett/ I 890 UK Developed tyre rim designs essentially

Welch similar to those in use today.

I 899 First plantation rubber shipped from Sri Lanka.

1900 World production of NR approaches 50,000

tons.

Kronstein 1902 Germany Polymerized styrene.

Oenslager 1906 US First chemical to "accelerate" vulcanization

(Aniline) then diphenylthiourea (DPU).

Lebedev l9l0 USSR Polymerized 1,3-butadiene to give a

rubbery material.

l9l0 Wild rubber peaks at about 90,000 tons per

annum. Plantation rubber (Malaysia/Sri Lanka)
reaches 10,000 tons.

78
Ostromislenski 1915 USSR First organic wlcanization systems without
sulphur: nitrobenzene and peroxides.

About this time the importance of zinc oxide

in accelerated cures was appreciated.

Kratz 1920 Diphenylguanidine (DPG).

1920 World production of NR 350,000 tons.

Bruni 1921 Italy Mercaptobenzothiazole (MBT)

Bedford 1921 US independently.

Lorentz 1922 US Thiurams.

Caldwell 1922 US Xanthates.

Bruni 1923 Italy Zinc dithiocarbamates.

Russell 1923 US Patented organic fatty acidlzinc oxide

use in vulcanization.

1924 Germany Fossilized rubber 60 million years old found.

1925 Germany Serious work began on the synthesis of

Synthetic polybutadiene (Buna).

1925 First commercial antioxidants (amines)

introduced. They were staining.

Rosenbaum 1926 US "S;mthetic rubber is dead."

Patrick 1930 US "Thiokol" first commercial synthetic rubber.

- is not a sulphur cure just metal
Note that this
oxides and possibly quinones. -
Semon 1930 US Suggested diffusion after vulcanization of
protective agents into rubbers.

1930 World production of NR 850,000 tons.

Tschunker 1930 Germany Buna N and Buna NN discovered (nitrile

rubbers of 25o/o ACN and 35% ACN).

1930s Introduction of amine derivative antioxidants

(staining).

du Pont 193 I US Duprene. Became polychloroprene (Neoprene).

IG Farben 1932 Germany Sulphenamides adducts of MBT with amines.

-
1932 USSR Russia manufactured SKA followed by SKB
(polybutadiene rubbers). -

79
Tschunker 1933 to 1934 Germany Buna S patented and produced (Styrene
butadiene copolymer).

l 93s USSR Sovprene; equivalent to Neoprene.

Thomas t937 US Butyl rubber.

Union Carbide/ t939 US First use of plasticized PVC as cable sheath.

Goodrich

1939 to 1941 US Nitrile rubber produced.

1940 World production of NR 1.5 million tons.

World production of Synthetic Rubber (SR)
100,000 tons.

1940 USSR SKI polymerized isoprene (synthetic "NR").

-
t943 US GR-S production started (now known as SBR)
Government rubber-styrene styrene
butadiene rubber (similar to ->
Buna S).

Pintin 1943 Germany First patent relating to wethane rubbers.

Dow Corning/ t945 US Silicone rubber. Again not normally a

General Electric sulphur cure, peroxides used.

1950s Inhoduction of Phenolic antioxidants

(non staining).

Last wild rubber exported from Brazil.

Goodrich 1954 US Announced high cis synthetic "NR."

1954 US Introduction of sub stituted

para-phenylenediamine antiozonants.

Firestone I 955 US Low cis synthetic "NR."

du Pont l 958 US Fluoroelastomers.

1960 World production crossover point.

Both NR & SR about 2 million tons.

MRPRA 1970 UK Urethane crosslinking of NR.

1990 World production of NR 5 million tons.

World production of SR l0 million tons.

1990 Perhaps 5 billion Hevea braziliensis

trees worldwide.

80
Ardil: The Disappearing Fibre?

Mary M. Brooks

York Castle Museum

York, Yorkshire
United Kingdom

Abstract new regenerated and synthetic fibres in the

early 20th century. The acetates using cellulose
Ardil is a regenerated proteinjibre produced pulps were well established by the late 1920s.
commercially for clothing and domestic.furnish- Truly synthetic fibres came a little later. Du
ings by Imperial Chemical Industries between Pont developed nylon in the U.S. and in 1939
l95l and 1957. This paper examines the develop- granted ICI a production licence. There was a
ment, characteristics, manufacturing methods real need for suitable fibres and textiles to sat-
and marketing of Ardil. Fev'examples of Ardil
isS both a growing world market and specific
are known, so this paper represents current work
military requirements. *
toward identifying textiles containing Ardil and
makes recommendations for their consewation.
Ardil reprcsents a particular challenge to both However, the cellulose-based fibres did not
the conservator and the curator and highlights yield fabrics that felt warm to wear and there
the importance of researching, identi,fying and was much competition to create an artificial
re co rd i n g manufac tured .fi bre s. wool-like fibre using a protein source. In order
to form a fibre, a polymer must have large
Introduction molecules that are capable of crystallizing,
achieving a reasonable degree oforientation
Ardil is a relatively little-known regenerated and having a high degree ofpolarity to give
protein fibre, which had a short but intriguing intermolecular cohesion. Possible alternative
life as a "wonder fabric," manufactured by Im- fibre sources included casein, a milk-rerived
perial Chemical Industries (ICI) after the Sec- product, waste leather and fish protein. Casein-
ond World War. So far, few named examples based fibres were developed under trade names,
have been identified in British collections. This such as Lanitol (Italy), Fibrolane (Britain) and
may indicate the nature of the long-term stabil-
ity of the fibre or problems in distinguishing
* ln 1942, the Ministry of Economic Warfare in
Ardil flom other fibres and blends unless it is
trade-marked. London was urgently questioning British legations in
Beme and Stockholm regarding the Axis powers tex-
tile supplies: "What is the output by Solanum GmbH
Development of Ardil ofcellulose pulp liom potato tops and other new
forms of cellulose? ... What is the extent of fibre
The development and eventual failure of Ardil production by Germany and Italy respectively from
as a technologically and economically viable fi- Broom plant (Ginstra)?... Your answers to all the
bre involved a complex mix of interacting fac- above may have an important bearing in enemy
tors. There was great interest in developing supply position." '

8l
Aralac (U.S.) and companies in both Europe It is worth noting here that Courtaulds were
and America were experimenting with vegeta- also working on regenerated protein fibres.
ble protein sources. Interest and development They developed Fibrolane A, BC and BX,
was such that in 1946 the American Society for made from case_in, and Fibrolane C, made from
/
Testing Materials proposed that all regenerated peanut protein. The relationship between the
protein-based fibres derived from sources, such two industrial giants was a complex mix of
as casein, peanuts, soya beans or other vegeta- competition and integration. Courtaulds had
ble proteins, should be described using the ge- vast experience in spinning and weaving silk
neric term "azlon." and regenerated viscose mourning crepes, but
relied on chemicals produced by ICI. Con-
It was against this background that W.S. Ast- versely, ICI had little knowledge of fabric
bury of Leeds University and A.C. Chibnall manufacture or consumer-led marketing, but
and K. Bailey of Imperial College, London, dominated the chemicals field. This relation-
approached ICI with the results of their experi- ship was formalized in a 1928 agreement by
ments in producing fibres from vegetable which Courtaulds agreed to abstain from chemi-
protein. Their initia"l research used protein ex- cal manufacture, taking all their supplies from
tracted from hemp,'but by 1936 they had ICI, whilst ICI promised to refrain from artifi-
switched to peanuts, the nut of a sub-tropical cial "silk" fibre production. The respective tex-
plant Arachis hy'pogae 2., as the protein source. tile technology skills of these two companies
Peanuts contain the proteins arachin and con- was to be crucial in decidins the fortunes
arachin. Hydrolysis of the amino acids in these of Ardil.
proteins has shown them to be very similar to
those in wool. Peanuts or groundnuts were Under wartime production pressure, ICI and
grown widely throughout the British Empire Courtaulds joined together in 1940 to form a
as well as in America, China and Borneo. How- company named British Nylon Spinners specifi-
ever. ICI was careful to stress that Ardil had no cally set up to produce nylon for parachutes.
connection with the great groundnut scheme in Despite this, Courtaulds persisted in regarding
Africa, which collapsed in scandal in the early nylon as a competitor to its traditional interests
1950s. Nevertheless, peanuts were an important in artificial "silks," the viscose rayons based on
crop, with a yield of around eight million tons cellulose. In fact, they campaigned to have ny-
per annum during the 1930s. lon classified as a new type ofrayon. In keep-
ing with this conservative attitude, Courtaulds
Astbury, Bailey and Chibnall refined their re- refused to join ICI in a prpposed post-war joint
search and proposed a process for producing venture to develop Ardi1.6 In 1943, the earlier
a fibre from de^natured vegetable protein dis- agreement between ICI and Courtaulds lapsed.
solved in urea.3 Their patent described it as ICI felt free to move into the field of fibre pro-
"the production of artificial filaments, threads, duction, but, significantly, they lacked Cour-
films, and the like from solutions of denafured taulds's expertise in fibre technology and
or degenerate or coagulated protein material by consumer marketing.
extruding such solutions into a diluent or other
regenerating medium."' They approached ICI ICI felt sufficiently confrdent to announce the
and reached an agreement whereby the com- creation of Ardil just before the war. An Ardil
pany took over development.' The project was production plan was launched in 1938, but
based at the ICI research plant in Ardeer, Scot- suspended due to the outbreak ofhostilities.
land hence the name Ardil was registered as Sufficient fibre was produced to make a few
- mark. Of the ICI divisions, the Scot-
the trade Ardil/wool blend suits: some of these were
tish plant had the most experience of fibre spin- apparently still being worn in ,1951 when fuIl-
ning and weaving. ICI researchers developed siale production coimenced.g Post-war, ICI
practical processes for production that moved urgently needed to diversify. In particular, they
away from the original theogetical propositions needed to find an altemative production activ-
and filed their own patents.o ity for their Scottish Explosives Division. Ardil

82
was selected as an ideal new product. The pilot sales never rose above 2.6 million pounds. In
plant was established in 1946 although produc- 1955, producing Ardil cost 47 old pence a
tion was initially limited to half a ton a week. pound (approx. 20 new pence, or 40p Can.) as
In July 1941, ICI proposed a capital expendi- opposed to 24 old pence a pound (approx. 10
ture of 2.1 million pounds for Ardil with an ex- new pence, or 20( Can.) for viscose staple fibre
pected return of 70% when the main plant was despite the fact that the raw material was a com-
operating at half capacity.'" Building of the parable price. " A drop in wool prices further
manufacturing plant at Dumfries, Scotland, reduced the competitive advantage of Ardil as a
commenced in 1948 with an initial capacity of bulking fibre.
22 million pounds per year although production
was well below this amount. Caught in this vicious circle, Ardil never
gained a firm foothold in the market. However,
Simultaneously with this development, synthetic fibres overall had made a significant
Unilever, a company with various interests impact. ICI set up Group F to develop their
including margarines, was building a plant at manufactured fibres including Terylene and
Bromborough, Merseyside, to handle the raw Ardil in 1956 and were still expressing cautious
peanut meal. Unfortunately, disputes with optimism regarding the future of Ardil: "Sales
Unilever over processing charges caused de- of Ardil were disappointing ... but work is in
lays. The major problem however was an unex- hand to evaluate an Ardil fibre with imnroved
12
pected shortfall in world peanut production properties." This improvement never material-
associated with the failure of the East African ized, and in 1957 ICI took the decision to cut
groundnut scheme. Ardil manufacture com- their losses, which had been running at 3.7 mll-
menced in 195 1 again, most unfortunately, lion pounds and Ardil production ceased.
coinciding with a- depression in the textile It had an effective commercial life of only
trade. Of the 316 firms who had indicated inter- six years.
est in 1947, only 76 placed small trial orders in
195 l. The auguries were not good. Manufacturing Methods: Successes
and Problems
Technical difficulties, which will be discussed
more fully later, meant that Ardil was best used The basic principle of manufacturing was to
as a blend with wool, cotton or other manufac- extract the proteins, arachin and conarachin,
tured fibres. This resulted in marketing prob- from the peanut and treat them so that they
lems as such blends could not be marketed as formed a solution from which a fibre could
100% pure single fibre products and therefore be spun.
required special promotion to consumers.
Despite the support of the Bradford Dyer's As- The peanuts, whose fruits grow underground,
sociation and other wool and cotton manufactur- were hand pulled, shelled and the reddish skin
ers who were prepared to invest in Ardil, some removed to prevent discolouration. For maxi-
Divisions of ICI were always skeptical. The mum protein yield, the nuts needed to be ma-
Dyesfuffs Division, which had the most experi- ture but ungerminated. The decorticated nut
ence of the textile market, was never prepared consists of 43o/o to 48o/o oil,24Yo to 26Yo pro-
to declare confidence in the fibre. In the end, in- tein and 26oh to 29o/o carboltydrate. After
temal dissension and technical problems meant blanching, the nuts were ground to a meal and
that economic realities conquered. solvents used to extract the oil, which was
a useful by-product used for salad oil or
If fudil was to be attractive as abulking agent marganne.
for the more expensive natural frbres, it had to
be cheap. Such cheapness could only be Care was needed to select suitable solvents and
achieved when production levels were high to control temperature during this process. The
enough to realize economies of scale and while oil-free meal contained about 40%o to 49o/o pro-
costs were high it was hard to build up sales tein. It was dissolved in dilute alkali and acidi-
volume. Over the five years of production, fied with sulphur di-oxide to pH 4.5 to extract

83
the protein. This process was carefully control- light colours without bleaching. It was nearly
led to keep the colour as light as possible.* neutral with not more than 0.3Yo acetic acid and
a slightly higher moisture regain rating. Ardil F
The solid residue, mainly carbohydrate, was was a pale fawn colour and was used when dy-
used as a base for cattle fodder. The precipi- ing darker colours. It contained 4% sulphuric
tated protein was then washed and dried, giving acid and some formaldehyde. Ardil K was
a creamy powder known as "Ardein," which made in heavier deniers, but was otherwise
was almost pure arachin and conarachin. This similar to Ardil F. The range of deniers avail-
was then dissolved in dilute caustic soda giving able corresponded with those of wool: rang-
a viscose solution that was allowed to mature ing from fine grades to thicker deniers for
for 24 hours. This denaturation process allowed carpeting.
the folded long-chain molecules present in the
soluble protein to open out into an extended Dyeing Ardil
form so giving better mechanical properties.
Ardil dyed well both on its own and in blends.
Once this process was completed, the solution Mass-dyed Ardil was produced by introducing
was pumped through spinnerets into a coagulat- pigment into the fibre during manufacture, giv-
ing bath of 2% sulphuric acid and' l5o/o sodium ing highly light and water fast colours. This
sulphate to form the fibre. Urea or caustic soda was a cheaper dyeing process and 12 standard
solutions had been used in the earlier develop- colours were available, chiefly for use in carpet
ment stages. The tow fibre was then hardened yarns. Most wool dyes, such as acid and
by a formaldehyde treatment under acid condi- chrome dyes, as well as some direct and vat
tions to improve insolubility. The chemistry of dyes, could be used successfully. Dyes sensi-
this process of forming molecular crosslinks- . tive to formaldehyde could not be used. The dif-
was complex and not iltogether understood.14 ferential reactions of fibres to dyes meant that
Altematively, improved wet strength could be cross-dyeing of blends of Ardil with other fi-
achieved by acetylation or treatment with 0.3% bres produced interesting ingrained speckled
glyoxal polymer. A natural crimp developed if effects. Ardil could also be printed using
the fibre was wetted, stretched and released. traditional methods.
The fibre could be extruded in any diameter re-
quired. Although it could be produced as a con- Fibre Properties
tinuous filament, it was often cut into staple
lenglhs of between 1.3 cm to 20 cm depending Physical Properties: Ardil had many of the
on the application. It could then be spun using a characteristics associated with the natural pro-
woollen, worsted or cotton system. tein fibres warmth, softness, resilience and
the ability -
to absorb moisture and generate heat
Three types of Ardil were produced. Ardil B when wet. It had a remarkable wool-like tex-
was pale cream and so could be easily dyed in ture despite being less strong than wool. Al-
though not thermoplastic, its good crease
recovery and draping qualities were valuable in
*Moncrieff details one example of this process as
improving the performance of cellulose fabrics.
follows: "200 parts by weight of fat extracted peanut
Blended with synthetics, it improved wearing
meal are stirred with 3500 parts water at 20"C for
qualities through its ability to absorb moisture.
10 minutes. Then I 50 parts lo/o to 2To caustic soda
are added over a period of 20 minutes bringing the
liquid to a pH of 8.0 to 8.5. Stining is continued for Key Characteristics
one hour and the resulting solution is clarified. After
clarification the solution is pH 8.3 and is claimed to Colour: Probably as a result of some peanut
have practically no colour. Next sulphur di-oxide gas skins remaining despite skinning and blanch-
is passed through the solution until the pH value is ing, Ardil had a creamy yellow colour. Colour
4 to 5. This acidification results in a copious white varied with Ardil F and Ardil K being darker
precipitate, which allowed to settle, is centrifuged, than Ardil B. At one point, attempts were made
washed a1d dried. The yield is 84 parts air-dry to develop a peanut with a grey skin to reduce
protein."'' such colouration, but this was a long-term

84
project and such peanuts were not used for those of scoured wool. Comparative informa-
production. tion is given in Table I.

Elongation and Abrasion Resistance: Ardil

had good elongation at 50Yo with good elastic Table I
recovery so it was hard to break the fibre. How- Percentage Moisture Regain of Ardil in
ever, it had poor abrasion resistance unless Comparison to Other Fibres at 6594 RH
blended with wool. A 50/50 ArdiVwool blend and 20"C
had improved resistance possibly due to lubrica-
tion of the wool by the Ardil fibre. Fibre Moisture Regain (o/d

Wool 15.0
Behaviour with High Temperatures: Like Ardrl B, neutralrzed/bleached 14 0 to 15.0
wool, Ardil had low flammability and resisted Casernfrbre 13 0
mass-dyed,
Ardrl F, Ardil K, 12 0 to 13.0
high temperatures without softening or melting. unneutralized
It began to char at250'C. Excessive dryness Vrscose rayon 1 1.0
if the fibre was re- Vrcara protern fibre 10.0
caused embrittlement, but
Cotton LC
tumed to standard humidity and temperature, it Cellulose acelate 60
regained flexibility. Contemporary technical in- Nylon 4.5
Terylene polyester o.4
formation puts great stress on the low fire risk
of Ardil, presumably reflecting problems en-
countered when working other highly flamma- Source: lCl Technical Bulletin GL. 1
ble regenerated fibres. Extra fre cover was not "The Properties of Ardil Protein Fibre".
required by insurance companies for factories
processing Ardil. Felting and Shrinkage Behaviour Patterns:
Due to the lack of any surface scales, Ardil did
Electric Properties: Ardil had a low dielectric not felt, but felting could be induced. When
rating and gave few static electric problems blended with high grade wool, Ardil could ac-
when processed at normal humidities. celerate felting shrinkage, possibly as a result
of lubricating the wool fibres.
Tensile Strength: This remained the great
weakness of the fibre. The tensile strength of Behaviour under Ultraviolet Light: Ardil had
Ardil was ^calculated at about 8 kg/mm' to high UV resistance. An ArdiVwool blended
l0 kg/mml, which is lowler than that of wool at yam exposed to UV light for 150 hours in a
l2kg/mm' to2}kglmm". When wetted the fi- carbon arc fadeometer showed no significant
bre extends about l5% in length and about 5% strength reduction.l6
in diameter, but on drying, retums to its origi-
nal dimensions. Tensile strength was greatly Tolerance of Pests/Fungi: Unlike wool, Ardil
reduced in the wet state, dropping ffom I g/den- was mothproof (a point that was made much of
ier when dry to 0.3 g/denier when wet. Pro- in advertisements) and was more resistant to
posals for chemical treatrnents to increase mildew.
crosslinking and overcome this significant
strength loss included stretching in a bath of Comparison with Wool: Technical details,
mercuric acid and acetic acid, which had obvi- comparing the qualities of Ardil with wool, are
ous safety problems.l4 Traill and Simpson given in Table IL
showed that an increase of 50% in wet strength
and25%o in dry strength was possible if the fi- Chemical Properties
bre was treated with basic chromium sulphate
and formaldehyde after dyeing with acid or Ardil fibre is essentially made up of two pro-
chrome dves. '' teins, arachin and conarachin. Chemical reac-
tion with formaldehyde causes these to become
Moisture Absorption: The moisture content insoluble in, and resistant to, the aqueous liq-
was l4o/o. The moisture absorption and regain uors used in textile processing. Ardil F and
rates at standard conditions were similar to Ardil K and the mass-dyed fibres are acidic in

85
 Behaviour with Solvents: Ardil is insoluble in
Table II the standard organic solvents. It may therefore
Technical Properti es of Ardi I be solvent cleaned, solvent selection depending
on the behaviour ofany blended fibre.
Properties Ardil Wool

Morsture regarn at Ardil and Other Fibres

65% RH & 20'C
Types F & K 12% 16%
Type B 14"k 16./" As aresult of the poor tensile and wet strength,
Ardil was almost always used in blends either
Tensrle strength (kg/mm2) 8to 10 '12lo 20
as an economic bulking agent with such natural
Tenacrty (g/denrer) fibres as wool, or to improve the characteristics
condfioned 0.7to09 ofcellulose and regenerated cellulose fibres.
wet 0.4 to 0.6
ICI recommended that it should not be used
Elongaton at break (%) alone. The blended fabrics were designed to be
dry 40 to 60 30 suitable for a wide range of end uses from cloth-
weI 50 to 70
ing to domestic fumishing fabrics. A ratio of
Heat ot wettrng (Cal./9.) 26 6 269 50% ludil/sUolo merino wool was used for
sweaters while 40Yo Ardlll60% viscose was
Torsonal ngrdrty (dynes/cm2) 1.2 x 1o1o 1.3 x 1o1o
used in carpeting. Ardil, viscose and nylon
Specrfrc gravrty 13r 'l 33 blends were used for hard-wearing, lightweight
Young's modulus
and cheaper suiting cloths. The Whitworth Art
(g/denrer % measured al 014 0.20 Gallery, Manchester holds a group of fabric
100 extenson per mrnute) samples from ICI in a wide range of blends and
fabric types. These include mass-dyed grey
Source: Anonymous untitled typescript, presumably Ardil/viscose/woolsuiting; Ardil/wool tweed;
lOl document, Whitworth Art Gallery, Manchester. Ardil/viscose/nylon shirting; printed Ardil,/Peru-
vian cotton dress fabric; ArdiVwool jersey
knit; Ardiliviscose velour coating and
reaction and contain about49/o sulphuric acid. Ardil/wooVviscose blanketing. Technical de-
Ardil B differs in being nearly neutral with ap- tails of the yams and weaves are given on the
proximately 0.3% acetic acid. swatch cards.

Acid and Alkali Resistance: Ardil was highly An Ardil blended yam may be used either as
resistant to acids and acid tendering. This en- warp or weft. Amongst the Whitworth samples,
abled it to withstand many of the standard shirting fabric Pattern 31 uses 30% Ardil B,
processes for textile production, which often in-
600Z viscose and, lUYo nylon in both warp and
volved acids. However, it had low resistance to weft. Other examples combine an Ardil blend
alkalis. The fibre would swell making it sensi- with a pure fibre as the other yam. Pattem 341,
tive to mechanical damage. Unlike wool, how- a lighfweight printed fabric, combines a warp
ever, Ardil does not require sulphur crosslinks of 100% cotton with a weft composed of 33%
for its structural stabilizing and therefore can Ardil and 670/oPeru Tanguis cotton. A few
resist chemical damage by alkalis. Normal vat swatches have differing blended yams in the
dyeing could be carried out on Ardil/cellulose warp and weft. The velour coating sample, pat-
blends. This was an important factor as many tem No. l8l, has a warp of 25% Ardil B with
of these fabrics were destined for use in 757o viscose and a weft of 50Yo Ardil B with
women's and children's clothins and were 500/o viscose.
often dyed.
Due to the differential strengths of the various
Behaviour with Bleaches: Like wool, sodium fibres, there were technical problems in prepar-
hypochlorite and sodium chlorite could cause ing the fibres. It was important that a uniform
degradation in Ardil and hydrogen peroxide blending was achieved to avoid problems and
bleaching was preferred. special techniques were required. In general,

86
Ardil and the complementary fibre(s) were Fabric Care
blended together before carding. ICI published
a range of technical information for the trade ICI recommended that Ardil could be washed
(see Appendix l). Conventional processing ma- using normal domestic or commercial methods,
chinery could be used avoiding the need to in- the governing factor being the needs of the com-
vest in an expensive new plant. plementary fibre. They rather cautiously stated
that, although ArdiVcotton blends could be
When blending with wool, Ardil needed washed as normal cotton. it was preferable th,qt
particular care to prevent overstretching and such blends should be "washed as for wool." ''
breaking. Relaxation shrinkage was possible in This presumably relates to possible increased
Ardil/wool blends when wet. Altering the per- loss of wet strength in a hotter washing proc-
centage of the blend allowed for the creation ess. Ironing temperatures were again govemed
of fabrics with different qualities. Alone, Ardil by the requirements of the other fibres used
would not permanent-pleat. However, when in the blend.
blended with a sufficiently high proportion of
a natural fibre, the resulting fabric could be in- As Ardil was stable in most organic solvents it
duced to carry a pleat. A 50/50 ArdiUwool could safely be commercially dry-cleaned.
blend gave a fabric that took a sharp crease
but hung out well whlle 25o/o to 35W65o/o to Marketing Ardil
75oh Ardlllwool was softer and did not carry
a crease well. Such blends were ideal in knits. ICI launched Ardil with extensive marketing.
Ardil was also used in hat felts. The British Hat The trade magazine International Textiles, The
and Allied Feltrnakers Research Association British Export Journal ofTextiles and Fash-
recommended blends of Ardil and rabbit fur or ions gave it wide coverage, reflecting the gen-
Ardil and wool. eral sense of amazement at this new fibre:
"Starting metaphorically with a bag of monkey
Lancashire cotton weavers were particularly nuts and an idea, British chemists of ICI have
interested in exploring blends to improve evolved a new synthetic fibre which, they
cotton's handle, crease resistance and drape. claim. is wool-like. does not shrink and is not
When working with cotton, it was recom- attacked by moths.;'20
mendet' that Ardil, the weaker fibre. should be
oiled.'* Methods of dealing with ArdiVcottg4 ln 1944, post-war shortages and technical prob-
mixes were extensively discussed by Dyke.'' lems were still holding up production and ICI
Ardil/cellulosic warps were sized either with was only able to promise that samples would
sago or Cellofas B. ln cotton blends, fudil fi- shortly be ready for the trade. However, ICI
bres showed a tendency to migrate to the sur- held a major exhibition in London in 1946.
face during washing. This could be controlled As shown in Figure l, the showcase factory
by using a tightly spun yanl gnd problems were in Dumfries was featured in an advertisement
reduced with plain weaves.'o These fabrics in the 195 I Scottish Festival of Britain
were used for nightwear, shirts, lightweight publication.2l
suits and dress fabrics.
Manufacturers were courted. ICI staff lectured
Tertiary blends exploited the wool-like handle and published articles on the new fibre. For ex-
of Ardil whilst adding strength from the other ample, n 1952 F.M. Dyke from ICI addressed
fibres. Lightweight blends of ArdiVviscose and the Oldham branch of the Textile Institute and
wool were used for velours, suitings and dress his lectures were subsequently published." ICI
fabrics. Nylon was used to improve strength went to considerable lengths to provide techni-
and abrasion resistance. Blankets were pro- cal information to textile spinners and weavers.
duced using a cotton or cottor/viscose warp (Titles are listed in Appendix 1.) ICI proposed
with an Ardil/wooVviscose weft while carpet- a wide range of end uses for Ardil, shessing the
ing could be produced using blends of Ardil, economic advantages ofcheapness and price
wool and viscose. stability in comparison with pure natural fibres.

87
 contained a useful section on thgpew synthetic

rt| l}t[ PI|0TI]ll l'llfllli fibres that aimed to do just this." The new fi-
bres are described as "Among the wonders of
the post-war world." Ardil is included and
praised as being wool-like whilst being cheaper
and for improving the behaviour of cotton and
viscose. The need for blending due to lack of
strength is clearly stressed. Care advice in-
cludes a warning against shrinkage when wash-
ing wooVArdil blends.

Examples of Ardil in U.K. Museum

Collections
This new ICI factory at Durnfries is now So far, few examples of Ardil have been
producing 'Ardil' protein fibre, and full producrion located in U.K. museum collections.
- at the rate of 22 million pounds per year--will
ruchicrcd b_r the cnd of 1951.
The York Castle Museum Wallis archive con-
sists of a donation of costume, spanning four
}It,CRI,{I" CHEMICAI, INDUSTRIOS LI\,TITID generations of one family. It includes a night-
\OBEL DIVISION dress belonging to Amy Wallis dating from
the 1950s. Old-fashioned in style and similar
to Viyella, a wooVcotton blend in handle and
appearance, it was made by Potters (Museum
number 431 .78). As shown in Figure 2,thela-
Figure I ICI advertisement for Ardil featuring the bel indicates that the nightdress contains Ardil
showcase factory in Dumfries, Scotland. and that it is styled by Unique from Potter's
Ardingle. The nightdress is in reasonable condi-
Ardil was also promoted directly to the public. tion and has clearly undergone regular washing.
In the 1950s, advertisements appeared in Microscopic examination of the fibres showed
women's magtvines, such as Good Housekeep- cotton plus Ardil.
ing. Ardll is described as 'rthe man-made pro-

s-
tein fibre soft as cashmere. smooth as silk
warm and-absorbent as wool. (It's moth resis-
tant, too.) Blended with other fibres, it gives
clothes the unmistakable touch of luxury at
prices you can afford." A housewife of the
1950s is shown selecting clothes for her family
pyjamas for her son, dresses for her daugh-
-ter and shirts for her husband. The copy encour-
ages her purchases; "Happy families wear
clothes that contain Ardil." Interestingly, she
never seems to wear it herself." A scarf made
with Ardil, which is now in the Nottingham
Museum of Costume and Textiles, may have
been a promotional gift from ICI to troops in Figure 2 Labelfrom nightdress of the 1950s, Wallis
the Korean War.
family archive, York Castle Museum (Museum num-
ber 43 1.7 8).
There was considerable concern to make sure
that consumers understood the new fibres and As previously mentioned, the Nottingham
knew how to care for them. In 1956 the News Museum of Costume and Textiles holds a scarf
of the World, Household Guide and Almanac from the 1950s that is labelled "Nobel Division.

88
An Ardil Blend, ICI Ltd." (Museum number
1979,609). The other fibre in the blend is not
known. The scarf is woven with a crimson and
green cross check and blue stripes with fringed
ends. According to the donor, the scarf was sup-
plied directly from ICI, but was not hard-wear-
ing as it had an unfortunate tendency to
dissolve in saliva.*

The largest collection of fibre samples, woven

fabric swatches and technical literature from
ICI, together with a bag of peanuts, is held by
the Whitworth Art Gallery, Manchester. There
are also tow samples, apparently of all three
Figure 3 Ardil mounted in XAM, magnified 400 times
types of Ardil. With the exception of one
swatch made by T. & J. Tinker Ltd. of Holm- Chemical and Physical Tests: When analys-
firth. Huddersfield. all the fabrics are from ing blended fibres, it is necessary to use a
ICI's Ardil Fibre Factory in Dumfries. The Sci- scheme. such as that outlined by the Textile
ence Museum, London, also holds some Ardil Institute to identifli the various elements.'- In
samples. this programme, the sample is tested using
normal sodium hypochlorite and 0.5 normal
Identification of Ardil sodium hydroxide to identify the presence ofre-
generated protein, wool or tussah silk. The pres-
Ardil is not easy to identify visually, particu- ence ofother fibres can also be established and
larly asit usually occurs as a blended yam and suitable tests carried out. Other techniques are
may vary widely in appearance. This may help required to identify which protein fibre is
to explain why so few examples have been iden- presented.
tified. It is important to sample both warp and
weft in blends thought to contain Ardil as they Burn Test: This test will establish the presence
may contain different fibres in different propor- of a protein frbre without specific identifica-
tions. It is also necessarv to identifu the other tion. Ardil bums without forming a bead with a
fibres present. smell of buming feathers as do wool and other
regenerated protein fibres.
Microscopic Analysis: On a microscopic level,
the longitudinal appearance ofArdil is straight, Wet Strength: This test can be used to estab-
uniform along the length and almost structure- lish the presence ofa regenerated protein fibre.
less although with some longitudinal striations. A marked loss of shength when the fibre is wet-
The edges are smooth. Liquid paraff,rn has been ted and then tested using a Strength Test ma-
suggested as a suitable mounting medium. The chine indicates the presence of Ardil or another
cross-section is circular with some slight pit- regenerated protein fibre.
ting, as shown in Figure 3.
Measurement of Fibre Density: This test will
* identify Ardil specifically. Moncrieff describes
Scarf mid 1950s, 1105 mmx216 mm. Label woven the method of establishing the specific gravif
yellow on black, "Nobel Division, An Ardil Blend,
of textile fibres. A glass tube is filled with two
ICI Ltd." A note from the donor, R.C.G. Williams,
liquids of contrasting density, which are al-
reads "supplied direct from ICI Ltd. who developed
the fibre type Ardil made from the protein part of pea-
lowed to diffuse into each other. The tube is
nuts. The fibre was made to augment the supply of then calibrated using fibres of known density
wool and was of importance during the Korean War before testing unknown fibres. Using this
( 1950 to 1953) when wool was in short supply. The method, it is possible to distinguish Ardil with
fibre has a "wool-ile" (sic) handle but is not very hard a specific gravity of 1.30 fi'om wool (specific
wearins and so failed when wool became plentiful." gravity 1.32) and casein (specific gravity

89
 1.2r.24 An alternative method using a Tecam Unfortunately, at the moment, there is insuffi-
Density Gradient Column apparatus is set out cient evidence to establish a degradation pat-
by the Textile Institute. tem for Ardil. Once more textiles containins
Ardil have been identified it will be possiblJ to
Infra-red Spectroscopy: Once a known stand- assess the behaviour of naturally aged exam-
ard has been established, Ardil can be recog- ples. Tentative conservation recommendations
nized by the chemical constitution shown on can be developed based on documentary evi-
the infra-red spectrum in comparison with the dence. Experience in dealing with degraded
spectra of other known fibre samples. At pre- wool fibres is also relevant. Full testing for fi-
sent no such standard has been established. In- bre stability and dye fastness should precede
fra-red spectroscopy showed a protein fibre, any interventive treatment. The conservation re-
similar but not identical to silk. quirements of other fibres and materials in the
garment should also be considered.
Solubitity Tests: Solubilify tests can be used
to distinguish Ardil from other protein fibres. Wet-cleaning should be approached with cau-
In 80% sulphuric acid, Ardil is unchanged tion as Ardil never achieved good wet strength.
whereas wool and casein dissolve. Caustic soda It is possible that degradation could intensify
5% at boiling point can be used to distinguish this weakness. Relaxation shrinkage could oc-
between protein and regenerated protein fibres. cur, particularly in ArdiVwool blends. Little is
Ardil will stay stable whereas wool dissolves known about the possible reaction ofdegraded
and casein shrivels up. Ardil will dissolve when Ardil with detergents or other wash bath addi-
heated at97"C in 18% solution of sodium hy- tives. High alkalinity would be undesirable as
droxide whereas ot]r_er regenerated protein fi- the fibre becomes lulnerable to mechanical
bres remain stable.2s The ability of Ardil to damage at a higher pH, and a non-ionic deter-
dissolve in a cold aqueous solutio I of sodium gent with sodium carboxy methyl cellulose as
hypochlorite (3.5 g available chlorineilihe) is an anti-soil redeposition agent would be prefer-
used to analyse.percenjgges ofcomponent fi- able. The evidence from the donor of the Not-
bres present in blends.' tingham scarf that it dissolved in saliva would
suggest that the fibre is not stable when ex-
Stain Tests: Moncrieff reports that the regener- posed to certrain enzymes.
ated protein fibres casein, Ardil, soybean and
zen all give a yellorv-orange colour with cold Bleaching protein fibres is highly problematic
Shirlastain A stain.' Shirlastain A is used for and not recommended. Ardil was never stable
the identification of non-thermoplastic fibres, in sodium hypochlorite or sodium chlorite.
such as cotton, wool and regenerated fibres.
The presence of formaldehyde may affect the When new, Ardil was stable in most orsanic
results of staining tests. solvents. Following testing, solvent cleining
may be acceptable. Suitable extraction facilities
Tests for Formaldehyde: Ardil is sfrengthened will be required to meet health and safety stand-
with a crosslinking treatment using formalde- ards. It is not clear how the use of formalde-
hyde or glyoxal to modifu the fibres so as to hyde as an after-treatment will affect the
improve stability and to allow such standard long-term stability of Ardil. Neither is it clear
textile processes as dyeing to be carried out. whether textiles containing Ardil might be a
The presence of formaldehyde canbe identified potential risk to other vulnerable objects in a
using the Chromotropic Acid Test.'" mixed media collection, particularly metals,
pigments and paper, on account oforganic acid
Implications for Conservation vapoq; resulting from the formaldehyde treat-
ment."' Isolating identified textiles and careful
Once the presence of Ardil has been estab- monitoring is recommended.
lished, the conservator should be able to make
better informed decisions regarding the treat-
ment of the textile.

90
Conclusion GL. I The Properties of Ardil Protein Fibre
GL.2 - The Construction and Finishing of
Until more textiles containing Ardil are discov- -
Dimensionally Stable Ardil Fibre/Cellulosic
ered, it is difficult to assess the long-term stabil- Fabrics
ity and behaviour of the fibre. The author WD. 1, 2,3 & 4 The Processing of
would appreciate receiving any information 3l/zDenier and 5 Denier Ardil Protein Fibre
about such items. The speed with which both on the Worsted System
known examples and contemporary informa- WN. I The Processing of Ardil Fibre on the
tion about a recent commercially produced re- Woolen- System
generated fibre have disappeared should alert CN. l. 2,3 &4
both curators and conservators to the impor- - The Processing of 2,3t/z
and 5 Denier Ardil Protein Fibre on the Cotton
tance of acquiring suitable examples and record- Spinning System
ing information on modern fibres as fully as WG. I Sizing of Ardil Fibre/Cellulosic
possible. Warps -
FF. tudil Fibre in Fur Felts
Acknowledgement FW.- I Ardil Fibre in Wool Hat Felts
FG. I - The Scouring and Bleaching of Yams
The author thanks The Conservation Unit, -
and Fabrics Containing Ardil Protein Fibre
Museums & Galleries Commission for gener- FG.2 The Application of Crease Resistant
ous support and grant aid; Fiona Strodder, Cas- Resins-to Blends Containing Ardil Fibre and
tle Museum, Norwich; Michael Robertson, ICI Viscose
Chemicals and Polymers Ltd.;Jeremy Farrell, FG. 3 The Finishing of ArdiVCotton
Museum of Costume and Textiles, Nottingham; - Fabrics
Blended
David Howell and Alain Colombini, Textile AN. I The Quantitative Analysis of Ardil
Conservation Sfudios, Hampton Court Palace - and
FibreArlylon Ardil Fibre/Cotton Blends
for infra-red spectroscopy analysis; Jennifer AN. 2 The Quantitative Analysis of Ardil
Harris and Ann Tullo, Whitworth Art Gallery, - and Ardil FibreA/iscoseA.,lylon
FibreA{ylon
Manchester; Sonia O'Connor, York Archae- Blends
ological Trust; Helen Durrant and Josie Shep- AN. 3 The Analysis of Ardil
pard, York Castle Museum; and Mark Suggitt, - andQuantitative
Fibre/Wool Ardil Fibre/Viscose/Wool
Yorkshire and Humberside Museums Council. Blends
Photographs by Richard Stansfield, York Cas- Forthcoming
tle Museum, copyright City of York Leisure The Dyeing of Ardil Protein Fibre "B",
Services. - "F" and "K" and Ardil Fibre Unions
The Printing of Ardil Fibre Unions
Source of Materials - The Application of Caledon Dyestuffs
- to Ardil Fibre "B"/Cellulosic Fibre
Shirlastain A Stain is available from Shirley Blends
Developments Ltd., PO Box 61, 856 Wilmslow
Road, Didsbury, Manchester M20 8SA Eng-
R6sum6
land.
L'Ardil, la Jibre qui disparatt?
Appendix 1 Fibre obtenue par rdgindration de prottines,
I'Ardil, qui a ,!tt produit commercialement par
Technical Bulletins I'Impeial Chemical Industies (ICI) entre 195 I
et 1957. a servi d confectionner des v€tements et
ICI published the following list of available dfabriquer des pidces d'ameublement. La prd-
and forthcoming Technical Bulletins on sente communication traite du dtveloppement de
"Ardil A New Staple Fibre for the Textile I'Ardil et des caracttristiques de ce produit, de
- (ICI Fibres Division, Harrogate, no
Industry," m€me que de ses techniques defabrication et de
date, possibly 1949): sa commercialisation. Et puisqu'il n'existe que
peu d'exemples connus de cettefibre, elle rend

9l
compte des travaux qui s'effectuent actuellement 7. Moncrieff, R.W., Man-made Fibres,6th edn.
pour retrocer les tissus contenant de l'Ardil, et (London: Newnes-Butterworths, 1 975).
foit un certain nombre de recommandations au
sujet de la conservation de ce mattriau. L'Ardil 8. Lutyens, W.F., "Icvcourtauld' s Relations, "
constitue donc un ddfi pour les spdcialistes tant Central Registry 2 lCl, 2 F ebruary 1945,
1 00 I 21
de la conservation que de la restauration, et il in: Reader, Imperial Chemica l Industries, 197 0.
illustre toute I'importance que rev€tent I'etude,
I' identification et I' enregistrement des
9. "From Ground Nuts to Ardil," ICI Magazine.
fbres synthttiques. April 1951.

References 10. "Ardil Supporting Memo," Central Registry

3l I I l, 6146, 8 July 1947,in: Reader, Imperial
l. Ministry of Economic Warfare, London, Let- Chemical Industries, 1970.
ter to NBM Legation in Berne and Stockholm,
2l October 1942. Public Record Offrce I l. "The Cost of Production of Ardil Fibre,"
L40-9n.
Central Registry 3115,9 August 1955, in:
Reader, Imperial Chemical Industries, 1970.
2. Chibnall, A.C., K. Bailey and W.T. Astbury,
"Improvements Relating to the Production of 12. ICI Review', 1956, p.21.
Artifi cial Filaments, " British P atent 467,7 04,
Provisional Specification No. 29161, 1935.
13. Moncrieff, R.W ., Man-made Fibres
(London: National Trade Press Ltd., 1957).
3. "A Perspective on the Preparation ofProtein
Fibres," Central Registry Papers (CR), 29 June
14. Traill, D., "Some Trials by Ingenious
1939. ICI Head Offrces Records Centre in:
Inquisitive Persons: Regenerated Protein
W.J. Reader, Imperial Chemical Industries. A
Fibres," Journal Society of Dyers and Color-
History (Oxford: Oxford University Press,
rsrs, July 1951, vol. 67 ,pp.257-270.
1970).
15. Traill, D., G.K. Simpson and ICI, "Improve-
4. Chibnall, A.C., K. Bailey and W.T. Astbury,
ments in or Relating to a Method for Improving
"Improvements in or Relating to the Production
the Strength of Artificial Insolubilized Protein
of Artificial Filaments, Threads, Films, and the
Filaments," British Patent 639,342 (1947).
Like," British Patent 467 ,704, Provisional
Specification No. 20927, 1936.
16. Untitled, undated typescript, presumably
from ICI, now held at the Whitworth Art
5. "Ardil Development," Notes of a Meeting
Gallery, Manchester.
on l4 February 1939. ICI Development Ex.
Committee Papers in: Reader, Imperial
17. Dyke, F.M., "Ardil Protein Fibre Blends,"
Chemical Industries, 1970.
The Textile Weekly, February 1952,p.491.

6. Various ICI Patents:

18. "The Production of Shirting Cloths," 1C1
"Hardening of Protein Filaments," British Pat-
Technical Bulletin FG.l, undated.
ent 492,67 7 (1932) and 492,89 5 ( 193 8).
"Insolubilizing Protein Fibres," British Patent 19. Ardil Protein Fibre. A New Staple Fibre
766,360 (r9s4).
"Insolubilizing Protein Fibres," British Patent for the Textile Industry (Harrogate, U.K.: ICI
Fibre Division, undated).
787,s88 (less).
"Protein Filaments," British Patent 758,445
20. "ICI Launch Monkey Nut Fibre,"
(19s6) and 7 57 ,2r5 ( 1953).
I nternation al Textil es, v ol. 12, 1944.
"Method of Insolubilizing Protein Filaments, "
British Patent 763,501 (1954).
"Insolubilizing Protein Fibres," British Patent
758,560 (1954).

92
21. Scott-Moncrieff, G., Living Traditions of 25. Canoll-Porczynski, C.2., Manual of
Scotland (London: HMSO for The Council of Man-made Fibres (Guildford, U.K.: Astex,
Industrial Design Scottish Committee on the 1960).
occasion of the Festival of Britain, l95l).
26. "Qualitative Analysis of Ardil Fibre,Atrylon
22. Advertisements: Good Housekeeping, and Ardil Fibre/ViscoseA.{ylon Blends," 1C1
August 1955, p. 16, November 1955,p.7, Technical Bulletin AN.2.lCI. 1956.
December 1958, p. 5:Woman's Journal,
April,May 1956, p.4. 27.Hatchfield, p.R. and J. CarpenteE
"Formaldehyde. How Great is the Danger to
23. News of the World, Household Guide and Museum Collections?" Centre for ConJervation
Almanac (London: News of the World, 1956). and Technical Studies. Harvard Universitv Art
Museums, 1987.
24. The Textile Institute ldentification of
Textile Materials,Tth edn. (Manchester: The
Textile Institute, 1975).

93
 Plastics Found in Archives

Alan Calmes

National Archives and

Records Adminis trat io n
l[/ashington, D.C.
U.S.A.

Abstract example, have gone through a number of

formats with different plastic materials, from
Ifith the rise ofplastics in the 1950s and 1960s, physically embossed grooves on dictaphone
there emerged novel ways ofrecording informa- belts, to magnetically charged particles on
tion. Dictaphone belts, Thermofax copies, and tape and, more recently, to laminated laser-pro-
all sorts of magnetic tapes joined vin1.,l discs and duced, digitally encoded disc recordings. Each
c el luloid film as base materials for permanentll'
involves a complex combination of plastics.
holding valuable information, such as speeches,
pictures, music, and data. Such mateials t.vpi-
cally reach an archive two or three decades after The desire to have quick copies ofpaper docu-
their creation, presenting a technically compli- ments led to the development of a variety of
cated preservation challenge. The cycle ofdevel- wet and dry and sometimes heat-processed
opment, usage, support, and obsolescence of coated papers during the 1940s and 1950s.
polymers in recording media is examined. Aging Some of these processes involved plastics. Be-
c haracteri s ti cs, c o ns erv a tion m easu re s, an d con- ginning in the early 1960s, the plain paper elec-
version options are examined in the context of trostatic photocopier produced an image using
arc hiv a I pres ervation adm i ni stratio n. copolymers mixed with carbon black and fused
to the paper surface. Early conservation meth-
Overview ods used plastics for the lamination of fragile
documents. The document, tissue paper, and
Throughout the l9th century, paper was practi- sheets of cellulose acetate were heat-pressed
cally the only material used for record-keeping, into a melt. Plastic enclosures and boxes have
except for some special applications ofparch- been used in archives, and plastic parts are
ment. Often, papers were bound together and found in information recording equipment. Ad-
stored in leather volumes, or tri-folded together ditionally, paints containing plastics have been
and pressed into wooden file boxes. Cellulose used on surfaces ofshelves and on containers,
nitrate, as a filler for pyroxylin bindings and as and plastic adhesives have been used to hold
a transparent substrate for photographic filrn, boxes and envelopes together.
marked the advent of manufactured plastics in
archives. During the first half of the 20th cen- Records are usually 20 to 30 years old by the
fury, vinyl, in the form of phonograph records, time they reach the care of an archivist. Before
replaced wax cylinders and joined film as a records are transferred to an archive, they may
plastic material for holding information. Since have been stored in harsh environmental condi-
1950, more and more plastics have been used in tions and subjected to rough handling. Archival
record-keeping practices. Sound recordings, for materials made of plastic often arrive in need of

95
special attention. For example, cellulose nitrate Many product names are used by non-special-
film may arrive in an advanced state of deterio- ists as generic names of plastics, such as Du
ration and be highly flammable and in need of Pont's Mylar for poly(ethylene terephthalate),
special handling, packaging and storage; dicta- Rohm & Haas's Plexiglas for poly(methyl
phone belts may arrive cracked and broken. methacrylate), and General Electric's Lexan for
An archivist needs to know the aging charac- polycarbonate.
teristics of plastics, conservation measures suit-
able for plastics, and copying techniques. Additives
Especially vulnerable are non-paper records, Plastics are seldom used in a pure state. Polyes-
such as motion pictures, video recordings, pho- ter used for encapsulation is generally de-
tographs, sound recordings, and computer data. scribed as a simple polyester, but even it
Some of the most important information of the contains some by-products left over from the
second half of the 20th century will require spe- manufacturing process, such as lubricants and
cial conservation and duplication efforts to pre- some silica compounds to prevent blocking.
serve the history of nations. A partial list of Polypropylene and polyethylene have antioxi-
important events recorded on plastics would in- dants added so that they can be melted and
clude: political debates, presidential addresses formed into sheets or poured into molds with-
and news conferences on radio and television; out undergoing oxidation. Additives add to the
satellite mapping and environmental observa- complex nature of plastic materials. Lubricants
tions; and motion pictures of historic events. in magnetic tape, for example, can ooze out of
An example of valuable information on a plas- the tape onto its surface. The reading-head ofa
tic medium is the sound recording of the tape-drive will collect the oozed-out lubricants
Nuremberg Trials, which was recorded on a and this can cause the reading-head to either
long loop of cellulose acetate film. The em- fail to read the data or gouge the surface of
bossed grooves and bumps have nearly disap- the magnetic tape, destroying the recorded
peared as the plastic material has gradually information.
retumed to its pre-embossed smooth state. A
special machine had to be built with a special Since plastic products may melt, burn and
tracking stylus to play the loop. spread a fre rapidly, flame retardants are often
added. Some local fire codes require the addi-
The majority of plastics found in archives tion of flame retardants in plastic materials
are thermoplastic materials, such as cellulose found in household and office fumiture. and.
nitrate; cellulose acetate; polycarbonate; therefore, flame retardants should be included
poly(methyl methacrylate) (PMMA or acrylic); in plastics found in abundance inside any build-
nylon; poly(vinyl chloride) (PVC); polystyrene; ing. Some flame retardants, however, can
poly(ethylene terephthalate) (PETP or polyes- evaporate in small quantities and affect materi-
ter); polyurethane; and polyolefins, such as als in contact with them. Some halogens, for
polypropylene. Some thermosetting resins are example, may produce oxidants that can react
also found, such as polyurethane, epoxy resin, with the silver halides of photographic film.
melamine formaldehyde resin, and phenolic
resins, such as phenol-formaldehyde resin, Changes in the Composition of Plastics
and Bakelite.
Plastics have evolved rapidly since 1950.
The following are examples of plastics found Chemical formulations have been replaced or
in archives: cellulose nitrate film; various cellu- modified to achieve desired results. New appli-
lose acetate films; vinyl (PVC) phonograph cations or improved materials brought about
discs; polyester encapsulation; polyolefin the obsolescence ofone form ofplastic in favor
shrink wrapping; polyethylene boxes and enve- of another. As a result, it is difficult to identifu
lopes; polyurethane binders on magnetic tape; old plastics found in archives without conduct-
acrylic sheets and blocks used in exhibits; ep- ing laboratory tests.
oxy resin adhesives and coatings; and nylon
gears.

96
Manufacturing processes have changed. Most plastics adhere to the ink and toner ofpaper
plastics originally served more immediate documents or to the imaging layer of photo-
needs and were not designed for long life. graphic prints. PVC plastic sheets have this
There was almost an assumption in our society quality because ofthe plasticizer, such as di-
that plastic products were disposable, and if octyl phthalate, which acts like a solvent in
continued use of the product was desired, it dissolving and athacting the copolymer ingredi-
would have to be replaced by a new and better ents oftoners and dyes used in electrostatic
product. This philosophy is changing, as manu- copies and photographic images. Furthermore,
facturers are beginning to produce engineering PVC should not be used for albums or filed
plastics with substantial durability and environ- with papers because it is an unstable plastic
mental resistance. Paradoxically, there are now that decomposes to produce hydrogen chloride
environmental concems that plastics are here (HCl) and, with a little water, this forms hydro-
forever; this fear has led to the development of chloric acid.
biodegradable plastic s for throw-away produc ts.
Plasticizers are solutes placed in hard plastics
The need to carry out periodic duplication on to reduce the glass transition temperature, and
plastic media extends to all types and formats: consequently to improve the flexibility of the
sound recordings, video recordings, motion pic- material. Poly(vinyl chloride) is an example of
ture film, and computer tapes. How often this is such a material that is normally glass-like. Plas-
done depends on how the medium is stored and ticizers are not chemically bound to the poly-
handled, and on the characteristics ofthe mer, and, over time, may come out of solution
plastics used. and be found on the surface ofthe plastic from
which they may evaporate if heated, or rub off.
Physical Characteristics Eventually, plastics that are plasticized will
"dry out," shrink, and crack. Other additives,
Plastics can be rigid or flexible, soft or hard; such as oils, lubricants, antioxidants, and cy-
they can be molded to almost any shape. Dur- anamides (also known as carbodiimides) may
ing manufacturing processes, plastics are malle- ooze out onto the surface of the plastic material
able and can be stretched as well as molded. as it ages. With the subsequent reduction in the
These attributes, convenient for making any solubility of the plasticizer in the plastic, a
shaped item, can cause problems later. With white powder is often found on the surfaces of
sufficient heat, plastics can retum to the malle- old films and tapes consisting of a number of
able state and change shape. When stretched additives that have come out of solution in the
during a forming process, plastics can be made plastic support and,/or recording layer.
into thin films; however, the material will con-
tinue to have a memory of an earlier state, Aging Characteristics
and try to revert back to a previous condition.
Many plastic films will relax back to their pre- Cellulose Nitrate
stretched dimensions if they are heated even Originally, almost all black-and-white 35-mm
briefly above a transition temperature that var- motion picture film was on cellulose nitrate
ies from one plastic to another. film. It was used exclusively for studio work
from the 1920s to the 1950s. Cellulose nitrate
Plastics can be made with such a smooth sur- was not used for color film, or for l6-mm and
face that when placed in contact with another 8-mm home movie film.
smooth surface, such as a photograph, pressed,
and then removed, the photograph will be left When ignited, cellulose nitrate bums very rap-
with a shiny surface. This process is sometimes idly. Nitrate film decomposes with the emis-
called "ferrotyping" (a term borrowed from pho- sion ofoxides ofnihogen. The reactions are
tography, referring to the process oftransfer- highly exothermic and are responsible for the
ring an image directly onto the smooth surface spontaneous ignition of cellulose nitrate. De-
ofa specially prepared iron plate). Such plastic spite the hazards, commercial film makers
sheets should not be used in albums. Another preferred cellulose nitrate film over cellulose
deleterious process, "offsetting," is when

91
 acetate safety film because it was easy to han- Cellulose Acetate
dle and produced a very clear, sharp image The same general mechanism of deterioration
when projected. By 1950, however, after many of cellulose nitrate occurs in a similar way in
theater fires, fire codes mandated the use of other cellulose esters. The same acid hydrolysis
safety film. During the past 10 years, after sev- occurs in all cellulose materials, but the other
eral devastating and dangerous cellulose nitrate esters produce only an acid that catalyzes fur-
film vault explosions and fires, archives and li- ther acid hydrolysis, not an oxidizer like nitro-
braries have copied most cellulose nitrate film gen dioxide in the case of cellulose nitrate. The
images onto safety film (a cellulose acetate or degradation process ofcellulose acetate film
polyester film) and disposed of the cellulose takes longer than that of cellulose nitrate film,
nitrate film. There remains, however, some but once started, the autocatalytic chemical
spliced-in cellulose nitrate film within reels reaction cannot be stopped. The result ofthe
of safety film. Nitrate-based film stock can be self-destruction of cellulose acetate film is
identified by feel; it is softer and more supple somewhat different from cellulose nitrate film,
than cellulose acetate or polyester film. When in that an intermediate stage of deterioration
degrading, its appearance may be deceiving. It between a good condition and a powdery condi-
is therefore safer to have a laboratory test film tion can be seen. Acetic acid, a product of cellu-
to confirm its nitrate content. lose acetate degradation, can be detected by its
vinegar odor. The presence ofacidic gases and
The aging of cellulose nitrate is characterized particles found in polluted air will initiate the
by rapid change once the deterioration process degradation process of cellulose acetate film.
begins. Prior to the onset of deterioration, there Unlike nitrate film, however, the image layer is
is no serious shrinkage of the material, image not chemically affected by the by-products of
quality is good, and the images can be copied the decomposition of the acetate substrate.
easily. The kinetics ofthe reaction are such that
there appears to be virtually no intermediate The first safety-based films were cellulose
stage between the time when the film is in good mono-acetate and cellulose diacetate. The term
condition and the time when it is obviously de- "safety-base" was used because acetate films do
teriorated. Archivists have seen reels of cellu- not bum easily. By the 1970s, triacetate began
lose nitrate film change from excellent to to replace diacetate as the favored substrate
extremely poor condition in two months. When for film.
it deteriorates, cellulose nitrate film produces
sticky-brownish, powdery-fibrous globs. Manufacturing experience found that the
diacetate substrate took diazo salts more readily
Gases emanating from deteriorating cellulose than the triacetate base and thus the diacetate
nitrate film can initiate the process of deteriora- base was used in diazo films until polyester
tion in neighboring films of all types. Chemi- began to be used in the 1980s.
cally, cellulose nitrate, upon decomposition,
produces its own oxidizer, and, therefore, once Researchers at the Image Permanence Institute
the chemical bonds begin to break down, a in Rochester, N.Y., are frnding little difference
rapid autocatalytic reaction sets in. The reaction between the degradation processes of cellulose
produces its own heat, which accelerates the diacetate film and cellulose triacetate film. One
initial breakdown. The process can be fast is not necessarily more stable than the other.
enough to produce fire and, if film is tightly The process, however, might take longer with a
compacted, an explosion can result. Once a de- triacetate than with a diacetate. Within one cate-
grading cellulose nitrate film has been identi- gory there are likely to be variations from one
fied, the archivist must move quickly to copy batch to another, depending on formulas. and
the images onto safety film and to dispose of additives, and manufacturing processes.'
the cellulose nitrate film. The continued useful-
ness of a reel of cellulose nitrate film depends The long, intermediate stage of deterioration of
upon good environmental conditions, such as cellulose acetate film is characterized by shrink-
clean, cool, and dry air. ase. When the cellulose acetate film base

98
shrinks, the emulsion layer on top, which does microfilm in the 1980s, but it co-existed with
not shrink, is deformed into a mass of wrinkles. cellulose triacetate-based microfilm.
Since the image is within the emulsion layer,
the image becomes illegible, unless there is a With the right combination of conditions, acid
way to copy or transfer the emulsion layer be- hydrolysis can break the bonds between mono-
fore the wrinkling obscures the image. The mers of polyester. During this process acids are
choices are to copy the image before this oc- created that in tum break more bonds. Once
curs or to laboriously remove and reapply the started, degradation is autocatalytic. Hydrolysis
image layer after it occurs. The greater the can be initiated by acids present in polluted air
shrinkage, the more difficult it is to recover the or left over from the manufacturing process.
information. When motion picture film shrinks, Oxides of nitrogen from automobile exhaust
the sprocket holes are no longer in the right form acids in the atmosphere that can acceler-
place, making the copying process, necessary ate the degradation process ofplastics. Exclu-
to save the images, difficult and expensive. sion of water from the air, that is, maintaining
The shrinkage will be uneven; consequently, en- a low relative humidity, is an effective strategy
gineered sprockets with a different spacing may to prevent hydrolysis. Using scrubbing systems
not provide a solution. or treated charcoal filters to remove pollutant
gases from the air are other strategies. The
Until the late 1950s, cellulose acetate films lower the temperature the slower the rate of
were used as a very thin base material for chemical reactions. To slow down degradation,
sound recording tape and for some early video carbodiimides are added to react with acids and
and computer tape. With age, and accelemted antioxidants are added to prevent reaction with
by elevated temperatures and high relative hu- oxygen; however, the additives will eventually
midities, thin cellulose acetate tape becomes be cons^umed, ooze out or evaporate from the
brittle and will break easily. Brittleness is an in- plastic.'''
evitable condition. Plasticizers used during the
manufacfuring process will ooze out onto the Polyester film is bi-axially oriented or "bal-
surface of the tape in the form of white droplets anced." Cellulose acetate is often uni-axially
that look like powder. To complicate matters, a stretched. Due to its limberness, polyester tape
recording layer, such as one composed of iron is more difficult to handle than cellulose acetate
oxides dispersed in polyurethane, will have tape. Polyester tends to respond to tension. For
characteri sti cs of degradation different from example, it will curl under tension. These char-
that of the base material. acteristics vary with the thickness of the tape.
The much thicker photographic film will dem-
A break in cellulose acetate magnetic tape is onstrate different properties, such as springi-
usually clean. It can be spliced back together ness rather than limbemess. Polyester has what
again with little loss of information for an ana- is called "plastic memory," and tries to go back
log sound or video recording. This is different to its pre-stretched state.
from polyester-based magnetic tapes that
stretch considerably before breaking. In order to reduce volume, polyester-based
magnetic tape for computer use has become ex-
Polyester tremely thin, but there is, concomitantly, a
Poly(ethylene terephthalate) entered the scene higher risk of loss of information, as a small
in the late 1950s. Because of its strength it was amount of stretching or some other dimensional
used, even in a very thin film, as the substrate change can cause the loss ofdata.
for all computer and video tapes. Soon thereaf-
ter, polyester replaced cellulose acetate as the Engineering Plastics
base material for sound recording tape. The Engineering plastics, such as nylon, polycarbon-
switch from cellulose acetate to polyester-based ate, and phenolic resins, have replaced metal
photographic film has been very slow. The de- for machine parts in modem information record-
mand for a strong, long lasting microfilm ing and retrieving devices. For example, video
brought about the use ofpolyester for players now have more plastic machine parts

99
than ever before. Plastic machine parts have adhesives can damage not only the surfaces
the advantage of being lightweight, tough, and where they are applied but also can initiate the
wear resistant, they do not need lubrication, degradation of nearby areas of papers, films,
and their gear trains operate quietly. Contrary and tapes.
to popular opinion that plastics are cheap substi-
tutes for metal components, plastic machine Polyolefi ns and Polystyrene
parts are often more expensive than the metal Plastic containers have significant advantages
parts they replace. Quality is a factor with plas- over metal cans and cardboard boxes. They do
tic parts as well as with metal parts. Quality of not corrode, are lightweight, and are unaffected
the product depends upon quality of the manu- by high humidity or water. Polypropylene mo-
facturing process for the plastic part and the tion picture containers are being used in some
quality of assembly into a machine. archives. There are no solvents or plasticizers
in polyolefins. However, colorants may come
There is a need for guidelines for the long-term to the surface and cause a problem and flame
maintenance of plastic machine parts. They retardants required by fne codes may slowly,
should not be lubricated. There is a problem even at normal temperatures, emit small quanti-
when part of a gear train is plastic and another ties of reactive materials, which might affect
part metal, because the lubricant needed for the the contents ofthe container.
metal part may cause the plastic part to deterio-
rate. Spare parts should be obtained before the Plastic cartridges can warp and cause magnetic
machine becomes obsolete. tape to mis-track when read and to mis-align
when rewound. In the latter case the edges of
Acrylics, Polycarbonates and Epoxies the tape may rub against the sides of the con-
Acrylic resins. polycarbonates. epoxy resins. tainer. Plastic pressure pads in magnetic tape
and various mixtures of these are used as cartridges have been known to crumble after a
shields, supports, adhesives, coatings, and few years. The pieces ofpad can damage the
toners for information recording systems. We reading machine and get between the layers of
have, however, very little experience on the use tape, causing the tape to be wound unevenly.
of these plastics in archives. Acrylics are better
known by their product niunes, such as Plexi- Because magnetic tape is thin and will sag
glas and Lucite, but acrylic products come in a against the flange when stored horizontally,
variety of forms from solid materials to liquids. magnetic tape should be stored vertically,
Some acrylic products turn yellow and become suspended on a hub; otherwise, an unevenly
brittle upon exposure to ultraviolet light or un- wound tape will result in having its outlying
frltered sunlight. Recent materials use ultravio- tape-edges folded under by the weight of the
let blockers to reduce the damage caused rest of the tape. Motion picture film is thicker
by light. and stiffer than magnetic tape and is stored hori-
zontally without flanges with the film resting
Poly(methyl methacrylate) (PMMA) has been on the surface of the container. The fihn must
used for compact discs (CDs) and CD-ROM be wound evenly and snugly. Instead of metal,
discs. Often, such discs were also coated with hubs and flanges are sometimes made of poly-
epoxy. Consumers began to note failures in styrene. In special cases glass has been used
these products in the late 1980s. Beginning in for flanges. Plastic is substantially cheaper and
1990, CDs and CD-ROMs were beginning to lighter than metal or glass; thus, we can expect
be made from polycarbonate, which is tougher to see more plastic.
and more resistant to change than PMMA.
Combinations of Plastic Materials
Much work remains to be done on plastics used
in adhesives. For example, pressrue-sensitive la- Combinations of materials abound in archives.
bels may fall offarchival boxes, tape splices on Some of these are borlnd together into lami-
motion picture film or magnetic tape eventually nated "sandwiches." For example, bound vol-
may fail and need to be replaced. Plastic umes consist of many layers of materials glued

100
together. An old phonograph record was con- produce the recording, or research will be
structed of a layer of shellac painted onto glass necessary to determine the appropriate size of
or metal. Magnetic tape base is made up of a stylus and use an appropriate transducer and
layer of pollurethane laminated to a substrate amplifier and play the recording at the
ofpolyester. Each layer has a different coeffi- correct speed.
cient of expansion. Changes in environmental
conditions can cause the separation oflayers. The pace oftechnological change has
The chemical products of decomposition of one quickened during the 20th century. Vinyl
layer can affect the other layer. One layer may records have been in use for over 50 years,
become brittle while the other remains flexible. analog sound recordings on magnetic tape for
50 years, digital recording for 30 years, and for-
Static Electricity mat changes are now occurring nearly every
decade. Plastics have played a role in the quick-
Plastics can hold a static electric charge and can ening of change by providing an infinite variety
release sparks and electromagnetic radiation. of new materials for adaptation. For example,
Static electricity is a problem for the microelec- computer information systems are constantly
tronics of some machines. Some manufacturers being upgraded with increased information den-
try to mitigate the problem of static electricity sity and reading speed, such as dye-polymers
by adding an anti-static layer to the back of used in some optical-magnetic recording sys-
tape or placing an anti-static additive directly tems to provide very high density storage and
into the plastic. Tapes coming out of long-term fast random access.
storage should be equilibrated at greater than
30% relative humidity to di-sipate the electrical The preservation of machine-readable informa-
charge and metal reels should be grounded. tion depends on periodic copying. Ifcopying is
Static electricity also attracts dust and grit that carried out properly and in a timely manner,
must be kept away from magnetic tape or disc there is little loss of information. However,
surfaces to prevent a reading-head from bump- there must be an assumption of continued cost
ing into them. Note that polystyrene, the plastic to pay for periodic duplication and new hard-
material used for tape flanges, can also collect a ware/software updates.
static charge. Dust and grit are also undesirable
on film since they scatter light, cast shadows Conclusion
and scratch the emulsion. Also, it is undesirable
to attract dust and grit at the time of polyester People in our society have ambivalent ideas
encapsulation of documents. about plastics. The memory of how plastic toys
break and cannot be fixed gives the impression
Obsolescence of plastics as cheap, ephemeral, and disposable.
Certainly, most people would not give a plastic
Rapid changes in technology during the object for a keepsake. From another point of
20th century have compounded the problems view, seeing plastics floating around in the mid-
conceming the maintenance of information in Atlantic Ocean stimulates environmental con-
other than readable form. For example, from cems that plastics pollute and will never go
the 1930s through the 1950s, sound-recorded away. The replacement of metal parts with plas-
dictation was kept on a cellulose acetate belt, tic parts is accepted by some as progress and is
sometimes referred to as a dictabelt from a dic- seen as an improvement; others would argue
taphone. At first, the sound frequencies were that the switch represents a cheapening of the
embossed mechanically into the surface of the product, even though it may look better.
cellulose acetate belt; later, the sound was re-
corded magnetically onto the surface of the Since plastics can be manufactured and molded
belt. A simple stylus can translate embossed into almost any shape at a fraction of the cost
vibrations into sounds, like that used for vinyl of other materials, there are economic forces
disc players. To play the magnetically recorded driving the use of plastics. With an acceptable
dictation on a belt, it will be necessary to obtain short life expectancy ofmost products today,
the same tlpe of machine as that used to perhaps attributable to rapid obsolescence,

l0l
frequent changes in style, and the desire to re- R6sum6
duce manufacturing costs. manufacturers are
using the less expensive plastics in products Les plastiques prisents dans les archives
that are virrually disposable after a few years.
Some plastics, however, such as polyester and L'utilisation de plus en plus rtpandue des
melamine are expected to remain durable and plastiques dan.s les annies 50 et 60 a foit naitre
to last for cenfuries. The various plastics have de n o ttve I I es faq on s d' e n regi strer I' i nfo mt ati o n.
their various environmental requirements, and Les bandes de Dictaphone, les copies Thermofax
as long as they are met, the plastic materials et toutes sortes de rubans nngnttiques ont ainsi
rejoint les disques en vinvle et les.films de cellu-
will serve their intended uses well. Maintaining
loid au rang des matiires de base qui seruent de
a benign environmental storage condition is
support petmanent aux enregistrements de dis-
key to extending the life of any material. Con- cours, de photographies, de pidces de musique,
stant low relative humidity between 30o/o and de donndes. Or, en ghtiral, ces documents ne
50% should benefit all plastics. Temperatures sont placds dans les archiyes que deux ou trois
should be kept as low as practical, between dic'ennies aprds leur creation, de sorte que leur
5oC and 20oC. consenation n'est pas sans poser d'tnormes
dfficultds techniques.
Plastics are manufactured materials subject to
endless changes in proprietary formulations. Nous examinerons, dans le cadre de la
After a formula has been used and the plastic prdsente comnunication, le cycle de developpe-
product is replaced by a new model, no one ment des poll'mires qui entrent dans la composi-
will know what formula was used or what addi- tion de tels supports et leur utilisation, tout en
tives were placed in the old plastic product; in traitant des mesures de soutien qui peuvent leur
which case, it is almost impossible to predict €tre appliqudes et de leur obsolescence. Nous
the life expectancy of the medium. Only when aborderons dgalement, du point de vue de I'ad-
we know the history of the material and its ministration d'un service d'archives, leurs carac-
chemical composition can we reasonably ex-
tiistiques de vierllissenrent, de m€me que les
diverses mestffes et options de conversion aux-
pect a certain performance and life expectancy
quelles on peut at,oir recours pour assurer la
of the plastic medium. We need the cooperation consentation de ces supports et des informations
of manufacturers to reveal the complete for- qu'ils contiennent.
mula of each plastic. With advance warning in
hand, archivists, therefore, can program replace-
References
ment costs to allow for a periodic migration of
information from one plastic medium to l. Adelstein, P.2., J.M. Reilly, D.W.
the next. Nishimura, and C.J. Erbland, "Stability of
Cellulose-Ester Based Photographic Film,"
Acknowledgement SMPTE Journal. May 1992, pp. 336-353.

The foregoing information on plastics found in 2. Smith, L.E. et al., "Prediction of the Long-
archives was derived primarily from interviews Term Stability of Polyester-Based Recording
with archivists and technical staff in the Na- Media," (Gaithersburg, Md.: National Bureau
tional Archives and Records Administration of Standards, 1986) Report No. NBSIR-
(NARA) and with polymer chemists and photo- 86t3474.
graphic film and magnetic tape standards ex-
perts at the National Institute for Standards 3. Smith, L.E., "Factors Goveming the
and Technology (NIST). Susan Lee-Bechtold, Long-term Stability of Polyester-Based
Chief Chemist, and Charles W. Mayn, sound Recording Media," Restaurator, The Interna-
and video recording engineer, both of NARA, tional Journal for the Preservation ofLibrary
and Leslie E. Smith, polymer chemist, and and Archival Materials, vol. 12, 1991,
Thomas Bagg, both at NIST, were particularly pp.20l-218.
helptul.

r02
occupe dtsormais dans notre civilisation. Le les bact,lries et m€me, la < chimie D, ce mot re-
terme ( caoutchouc > designe tout pob,a)vs q11i doutable), qui ont souvent un elfet qtnergique.
possdde ou semble poss'lder de l'ilasticit2, et il Aussi tenterons-nous, dans le cadre de la pr6-
existe aujourd'hui nombre de polym'ires d'ori- sente communication, de nous frat'er un chemin d
gine naturelle ou synthdtique qui r,lpondent d travers ce vtritable < champ de mines >, et
cette d|finition. Or, l'2lastom'ire peut tr'is bien d'tlaborer une mdthode qui permettra de choisir,
constituer moins de 50 2(' de la masse totale d'un pour chacun des divers genres d'dlastom,ires ex-
produit dit < de caoutchouc >, puisque plusieurs posis, la techniEte de restauration qui convient
substances chimiques, choisies parmi des cen- le mieux.
taines, peuvent avoir ttt ajoutdes pour lui donner
une forme utilisable. Bibliography
La prdsente communication ne vise nullement History of the Rubber Industty, eds. P.
d fournir une analyse scientifique detuiilee de la Schidrowitz and T.R. Dawson (Cambridge,
prdparation des produits de caoutchouc. Il n'en U.K.: W. Heffer and sons, 1952).
demeure toutefois pas moins que, avant d'amor-
cer tout travail de restauration d'un objet de Hancock, T., The Origin and Progress of the
muste, les sptcialistes de ce domaine devront
CAOUTCHOUC, or India-rubber Manufacture
absolument connaitre les matiriaux qui le com-
in England, eds. Longman, Brown, Green
posent et €tre bien conscients dufait qu'ils ris-
(London: Longmans and Roberts, 1857).
quent, s'ils n'appliquent pas le traitetnent
approprii, de luifaire plus de mal que de bien.
Aussi tenterons-nous de pr,!senter un historique Gottlob's Technolog, of Rubber I 92 5, Eng.
des dlastomdt'es qui met tout particuli'irement Trans., J.L. Rosenbaum (London: Maclaren and
l'accent sur certaines ddcouvertes du Xf si?cle sons,1927).
qui permettront aux spicialistes de la conserva-
tion de mieux classifier, voire de dater, les pro- Woll H. and R. Wolf, Rubber, a Story of Glory
duits dlastomdres. Au-deld de cet objectif, nous and Greed (New York: Covici and Friede,
dicrirons bi,ivement certaines des m,lthodes 1936).
simples qui peuvent ,?tre utilisdes pour analvser
ces produits. Chemistry and Technologv of Rubber, ed. C.C.
Davis (New York: Reinhold Publishing Corp.,
Il arive souvent que la sutface d'un produit t937).
ilastomdre en vienne d prtsenter des change-
ments qui, quoique ddplaisants sur le plan visuel, Buist, J.M., Ageing and Weathering of Rubber
ne sont pas toujours le prtsage d'une ddgrada-
(Cambridge, U.K.; W. Heffer and sons, 1956).
tion de la sudace; certains de ces changements
sont m€me voulus, pour permettre au produit de
Cook, J.G., Rubber (London: Frederick Muller,
durer plus longtemps. Nous passerons en revae
l 963).
les causes et les effets iventuels de tels change-
ments de surface, tout enfournissant des tech-
niques qui serviront d les distinguer entre eux. Wake, W.C., B.K. Tidd and M.J.R. Loadman,
Analysis of Rubber and Rubber-like Materials
Si I'on connait bien les ntatdiaux qui composent (London: Applied Science, 1983).
un objet de musde et que I'on peut, d'apr,is I'dtat
de sa surface, ddterminer ce qui lui est ariv,!, il Natural Rubber Science and Technologlt, ed.
sera d'autant plw Jacile de dejinir les mesures d A.D. Roberts (Oxford: Oxford University
prendre pour optimaliser sa vie utile d'exposi- Press,1988).
tion. Il n'est pas de solution qui puisse s'ap-
pliquer d tous les |lastomdres et d leurs produits, Toxicity and Safe Handling of Rubber
mais on retiendra qu'ils se digradent le plus sou- Chemicals (Birmingham: British Rubber
vent sous I'action de toute une sbrie de facteurs Manufacturers Association. I 990).
courants (l'oxvgdne, I'ozone, la chaleur, la lu-
midre , le travail ntbcanique, les mitaux oxydants,

74
Processes of Deterioration
Processus de d6gradation
Changes in Polymeric Materials with Time

David M. Wiles

Plastichem Consulting
Victoria, B.C.
Canada

Abstract Less well known is the phenomenon of physical

aging, whereby polymeric materials continue to
Virtually all plastics, fibres, rubbers, paints and alter for w,eeks, months or years after they have
protective coatings, as well crs paper, wood, skin solidified or otherwise attained the physical form
and hides, owe thetr useful characteristics to the in which they are used. Originally obsented in
relatively high molecular weights of their mole- thermoplastics belov, their glass transition tem-
cules. Values in molecular weights range from a peratures (Tg), physical aging arises because all
temperature-dependent propefties, which change
feu' t ho u s a n d t o s ev er al m i I I i on. Unfo nu n at e ly,
fabrication, handling, use or misuse of these abruptly at Tg, continue to change below that
materials result in deleteious changes (usually temperature, albeit verv slowly. Physical agtng is
decreases) in the molecular weights of the con- observed in amorphous, glassy polymers, in the
stituent molecules with concomitant reducti on amorphous phase of semi-crystalline polymers
in the materials' useful properties. The detailed and in the rubbery matrix offilled rubbers. As a
science ofthe changes at the molecular level is result, over time, many mateials become stifer
highllt complex and, frequently, not well under- and more brittle as "rates of relaxation" de-
stood. Nevefiheless, somefeatures common to the crease. Because physical aging affects polymer
degradation of many polymeric materials with segmental mobility, it should also affect chemical
time have been elucidated and can indicate the degradation, photo-oxidation, swe lling and de-
wa-v to improved preservation practices. swelling, and crosslinking reactions in a w'ide
v ari ety of macromo I ecul ar sys tems.

C hemical aging includes oxidative deterioration

as a result ofexposure in air to light, heat or ion- Introduction
izing radiation, and also includes hydrolysis and
auack fui acids or bases. Biodegradation can be The degradation of materials is identified by
considered a special case ofchemical breakdown the user/observer as an unacceptable change in
caused by microbial enzymes. The reactions in- characteristics, be they mechanical, chemical,
volved in chemical aging are numerous and some- optical, or electrical. In the case of polymeric
what material-specific, but the more important materials, that is, those comprising very large
ones canfrequently be related to the oxidation of molecules, it is usually an alteration in the mo-
liquid hltdrocarbons. The discolouration, embit- lecular weight of these molecules that results in
tlement and reduction in various phltsical proper-
an undesirable change in properties leading,
ties that accompany oxidative deterioration can
for example, to mechanical failure or discol-
be minimized by the use of appropriate combina-
tions of stabilizing additives. Microbial suscepti-
ouration. Occasionally, with some materials,
bilit)) is best dealt wilh bv "good housekeeping" crosslinking occurs (the formation of chemical
practices. bonds between molecules) with a resultant

105
increase in molecular weight, but more com- Hydrolysis, as well as deterioration from expo-
monly chemical reactions cause a reduction in sure to acids or bases, can be a problem in hu-
the molecular weight of the large molecules, mid or "hostile" environments for polymers
which collectively are the origin of the useful having specific structural features that are inher-
properties of polymers. Not infrequently, how- ently susceptible. Ester linkages are hydrolyz-
ever, unacceptable property changes can occur able, for example, and molecules that can react
even though very little change in molecular with acids or bases will usrrally be degraded by
weight can be measured. In short, polymer deg- them. Microbial attack may be a problem for
radation is widespread, complex and sometimes natural polymers since, in warm, moist air, fun-
difficult to evaluate; it can also be very difficult gal enzymes can oxidatively degrade many
to prevent. such macromolecules. Biodegradation is fortu-
nately not a factor in the deterioration of many
Much of the progress in elucidating the degra- synthetic polymers.
dation of macromolecules is summarized in
the articles and monographs listed in the refer- It is worth remembering that the molecular
ences.'-'- It is iustihable to look for simolifica- weights of polymeric materials are really very
tions and geneialities that assist in under- high and that their desirable properties will
standing and dealing with polymer degradation. be lost when only a very small fraction of the
Tuming first to chemical aging, there are nu- bonds per molecule are broken. The molecular
merous features common to the oxidative dete- weight of the molecules in cotton averages over
rioration of many kinds of macromolecules as a 2 million, in wood cellulose over 1.5 million,
result ofexposure in air to light, heat, ionizing and in paper several hundred thousand. Among
radiation or mechanical action. Some of these the synthetic polymers, nylons and polyesters
overall features are summarized in the figure having superior physical properlies may have
below. It has proven useful to relate these kinds molecular weights of only 30 thousand or so,
of chemistry to the oxidation of liquid hydrocar- but polyolefins are characterized by values
bons. Such model compound studies have iden- from a few hundred thousand to more than a
tified mechanisms that apply (at least in part) to million. The breaking of one bond in, say,
the degradation of solid polymers; kinetics, the l0 thousand in polymer chains will have a dev-
rates of the critical reactions, are rather more astating effect on the properties of the material
composition- and state-specifi c. Nevertheless, (unless the scission occurs near the ends ofthe
considerable progress has been made in devis- molecules). Such sensitivity is rarely observed
ing methods (largely chemical) of postponing in small-molecule science.
for prolonged periods the inevitable
degradation. Physical aging is an entirely different phenome-
non in that it is thermodynamically driven and
cannot be stopped although it is readily revers-
ible. Moreover, it is by no means invariably un-
desirable. All temperature-dependent properties
of materials that change abruptly at the glass-
transition temperature continue to change be-
low Tg. In effect, materials become stiffer and
more brittle; rates of relaxation decrease so the
response of a material to deformation or impact
changes, usually for the worse. Physical aging
in amorphous polymers and related materials
persists for a very long time and affects a vari-
Ro. +.OH ety of properties so that repair and replacement
,ro"r,F*t is tricky and, of course, longevity may be com-
prom ised. Additionql^information can be found
in the references.''-"'

106
Oxidative Degradation The identification ofthose processes that cause
the loss of the essential properties of polymers
A very high proportion of the common poly- has consumed a very great deal of research time
mers, both natural and synthetic, have mole- during the past 40 years and it is by no means
cules that consist of chains of carbon atoms complete. One singularly useful approach has
bonded together or chains of carbon atoms in- been to identify the more important products of
terspersed with other atoms, such as oxygen polymer degradation and, using principles es-
(e.g., in polyesters, polycarbonates, polyethers) tablished with low molecular weight analogues
or nitrogen (e.g., in polyamides, polyure- for the polymers, "reconstruct" the chemistry
thanes). Invariably, hydrogen is bonded to that gave rise to the products. Thus, under ther-
most of the backbone carbon atoms but, occa- mal stress, the weakest bonds in the system are
sionally, so is oxygen or chlorine, for example. most likely to break; these could be tertiary
Aromatic rings may be part of the polymer carbon-hydrogen bonds, for instance, or peroxi-
chains (e.g., PET, aramids) or pendant from dic linkages. In the case of light-induced dam-
them (e.g., polystyrene). Energy in one form or age, it is the shortest wavelengths (highest
another is introduced into macromolecules energy) in the incident radiation that do most of
either during fabrication (or other handling) or the damage and there must be appropriate chro-
during use and, inevitably, chemical bonds of mophores (light-absorbing groups) present for
the types referred to above will be broken ho- any damage to occur. For example, the more
molytically, that is, to produce highly reactive cotton or other cellulosic is purified, the whiter
free radicals, in pairs. Sometimes, it is these pri- it becomes and the less susceptible it will be to
mary reactions by which molecular weights are actinic deterioration because the relevant chro-
reduced below acceptable levels, but it is fre- mophores are present in extractable impurities
quently the case that the damage is done in sec- rather than in the polysaccharide chains them-
ondary or even tertiary reactions. The varied selves. Another example is that the purer a poly-
response of many common polyners to degra- olefin, the less rapidly it will photodegrade,
dative environments is illustrated in Tables I since it is the chromophoric impurities in these
and Il,.which are taken from Carlsson and thermoplastic molecules that give rise to their
Wiles.' photosusceptibility.

Table I
Susceptibility of Unstabilized Polymers to Degradation
Resistance to degradative process'

Thermal Photeoxidation
Polymer Oxidation Weathering Ozone Hydrolysis Oxidation
polyethylene p e e f
polypropylene vp vp e p
polystyrcne I Y
poly(m ethyl methacrylate) s I
poly(tetraf luoroethylene) e e vp
polyamrde (Nylon-6 and -6,6) f f f t
polyacMonrtrile p e f
s
poly(vinyl chloride) vp p p
poly(ethylene terepthalate) s g
s
polyorymethylene p p I f vp
polycarlconate I p
poly(phenylene oxrde) vp t g s
poly(ester urethane) f t I
s s
poly(ether urethane) p p f t
s
poly(/.n-phenylene
rsophthalamtde) e
poly(p-phenylene
terephthalamtde) e

' Key: e = excellent, g = good, t = farr, p = poor, and W = very poor

t07
Given that some carbon-carbon bonds are bro- apply whether bonds were cleaved initially
ken in due course, a general series ofreactions by heat, light, ionizing radiation or mechanical
may be written: stress. The behaviour of a reactive chemical
species, such as a polymer radical, is deter-
Polymer heat, lrght Carbon{entred mined by environment (both macro and micro)
Molecules RH 2R' radicals, rn
more than by origin; the radical has no memory
ronrztng parrs
radration (see Figure 1). This feature has been helpful in
the elucidation of polymer degradation mecha-
RH
R'+02-t -> ROz' TROOH*R' nisms as well as in the development of many
highly effective stabilization systems.
heat RH
ROOH 'OH + RO' ROH + R'
lrght / \\ There are many synthetic polymers that have
-> PRH -> chromophores as part ofthe repeat units, that
HzO + B' R'+ ketones
is, they are built into the polymer chains. Car-
2RO2' Stable Products bonyl groups in polyesters, nylons (including
-> aramids), polycarbonates and polyurethanes are
examples of chemical structures that absorb spe-
Two features ofthis sequence should be obvi- cific wavelengths in terrestrial sunlight. Nor-
ous. Since each initial carbon-centred radical rish-type photodegradation reactions ensue,
(R') gives rise to several other radicals, this is a leading to the formation of radicals that un-
branching chain reaction overall and, ifnothing dergo the same kinds of chemistry illustrated in
intemrpts the sequence, oxidative degradation the scheme shown earlier. There are other kinds
ofa polymer can proceed rapidly. A second fea- ofphotochemistry and other types ofdegrada-
ture is that there is likely to be some difficulty tion reaction in the case ofcer|ain specific poly-
in sorting out all the degradation products and mers, but the scheme is valid for a wide variety
where they came from, especially since some of macromolecules.
are themselves heat- and light-sensitive. A
third, less obvious feature ofthis reaction In addition to reducing the tensile or flexural
scheme is that it should apply (at least in part) performance of polymeric materials, oxidative
to a wide variety of polymer types and it should degradation llequently results in surface

Table II
Thermal Deterioration of Polymers Maximum lJse Temperaturea

Polymer Generic Type Film Thickness, mm Maximum Use Temperature,'C

poly(vrnyl chloride) 2.O 50

polyethylene f,U
polyoxymethylene 0.7 50
poly(methyl methacrylate) 1.5 50
polyslyrene 31 50
poly(phenylene oxrde) 15 50
polyamrde o.7 65
polycarbonate" o.7
epoxy resrns 3.0 90
$ilcones 0.7 105
poly(ethylene terephthalate) 0.2 105
alkyd resn 1.5 130
polytetraf I uoroethylene 0.9 150
phenohc resin 2.0 150
polyrmrde 0.1 200
aramtd" o2 220

a Recommended use temperature at whrch 50% of the ongrnal drelectric strength, tensle, and rmpact properttes are
retatned for 1 1 ,000 h under low contrnuous stress In some cases, stablhzers may be present. Underwnters
Laboratones data.
b Bisphenol A
c Aromatrc polyamrde.

r08
cracking, discolouration and enhanced surface fungi that operate by excreting water-soluble
wettability, for example. In all cases, the trick enzymes onto a biosusceptible substrate.
is to try to identify those chemical reactions Subsequent breakdown of the molecules of the
that cause the loss of useful properties since substrate material into water-soluble fragments,
these are the reactions that need to be orevented for example, two carbon-atom chunks, is fol-
as much as possible. Commonly. materials are lowed by transfer back inside the mycelial
subjected to heat and light stress simultane- cells. Fungi like to be warm and moist and they
ously. Conventional wisdom has it that photo- require oxygen because, like mammals, they
chemical reactions have no activation energy. derive energy from the oxidation ofcarbon to
that is, there is no significant effect on rate of carbon dioxide. Over tens of millions of years,
increases or decreases in temperature. Even fungal systems have evolved that can degrade a
though this is true for most primary photoproc- wide variety of materials, some of them rather
esses, a number ofthe subsequent reactions toxic. There is, however, as yet no fungus that
will, in fact, be govemed by Arrhenius princi- has the en4rmes to break down most synthetic
ples. No one has yet determined how to de- polymers, which, by and large, do not occur in
scribe this situation quantitatively, for example, nafure and have been invented within the past
in terms of lifetime predictions, but it should be 60 years. Nevertheless, microbial growth can
kept in mind that the effects of light and heat to- readily occur on surface dirt of bioinert materi-
gether on polymers are more severe, sometimes als like plastics, causing, at the very least, unac-
synergistically so, than either separately. ceptable aesthetic consequences.

Moisture-Based Deterioration Fungi proliferate by sporulation and firngal

spores are literally everywhere, waiting for con-
It is self-evident that polymers that are highly ditions favorable for germination. Commercial
hydrophobic are unlikely to be susceptible to fungicides are available, but these are relatively
hydrolysis; there are no hydrolyzable groups toxic compounds unsuitable for many materials
present. Using polyolefins as an example, these and situations. Whether there is a need to pre-
plastics cannot hydrolyze, are highly resistant vent growth on a susceptible substrate or on the
to acids and bases, and are microbially inert. contaminated surface of an inert material, the
These characteristics give rise to applications in best way is to maintain "good housekeeping"
geotextiles, food packaging and automobile practices. In other words, it is usually sufficient
parts, for example. Some polyesters are more to keep things cool, clean and dry, in air condi-
hydrolyzable than others. PET, being relatively tioned premises if possible, in order to mini-
resistant, is highly useful as textile fibres, soft mize the undesirable effects of funsi and other
drink bottles and prosthetic devices, whereas a micro-organisms.
simpler polyester, such as poly(glycolic acid),
is an effective "biodegradable" suture material. Physical Aging
Chemical structure at the molecular level is the
critical factor and a very reasonable basis for As the temperature of a solid, amorphous
material selection. It is not surprising, there- polymer is raised, the kinetic energy of the
fore, that many naturally occurring polymers molecules will increase, but the resultant
are not soluble in water (or many other sol- vibrations and rotations will be significantly
vents) at ambient temperatures. Likewise, sus- restricted as long as the material retains its
ceptibility to reaction with (and destruction by) glass-like structure. At a specific temperature,
acids and bases is predictable on the basis of characteristic of each polymer type, there is a
polymer structure compared to that of analo- measurable change in the system where glass-
gous small molecules. like properties give way to rubber-like behav-
iour, owing to the onset of greater rotational
To simplify somewhat, biodegradation is com- freedom and more segmental motion (20 to
monly molecular fragmentation resulting from 50 chain atoms) of the polymer chains. This
the chemical effects of microbial enzymes. A temperature is called the glass transition tem-
very cornmon manifestation of this would in- perature or Tg. Since the properties of a rigid
volve one of the more than 80.000 kinds of glass and a rubbery plastic are very different,

109
the Tg of a polymer is one of its most important almost impossible to quantify on the basis of
characteristics. Values range from -102"C for any free-volume models that may be applied to
cis- 1 ,4-polybutadiene, -67oC for natural rubber, the aging ofplastics.
-20"C for polyethylene, 57"C for nylon 6,6,
69oC for PET, 8l'C for PVC, 100'C for poly- Polymer Stabilization
styrene, and up to 149"C for bisphenol-A poly-
carbonate. Thus, some common polymers are A very high proportion ofthe reactions that de-
below their Tgs at room temperature. More- grade polymers involve free radicals: reactive,
over, this is the case also for the amorphous neutral species that tend to be chemically self-
phase of semi-crystalline polymers and the rub- perpetuating in a hydrocarbon matrix. Two
bery matrix of filled rubbers and other compos- ways of approaching the requirement to stabi-
ites. Amorphous glassy solids are not in lize polymers are (a) reducing the rates of for-
thermodynamic equilibrium after solidification mation of these radicals, and (b) preventing
and thus, over a wide temperature range and their destructive reactions by deactivating them
for a very long time, will undergo characteristic first. It is self-evident that in selecting a mate-
property changes. This is called physical aging rial, recognition of the need for longevity in the
and is quite different than, and distinct from, intended use environment will be factored in
chemical aging, such as thermal degradation with mechanical, aesthetic, and cost criteria. In
and photo-oxidation. connection with approach (a), it is advisable to
avoid the formation of relatively labile chemi-
The temperature range over which physical cal groups or relatively weak bonds in a poly-
aging occurs can be quite broad and frequently mer during storage, handling, application or
includes the use-temperatures of common plas- fabrication. If the material is heat sensitive, for
tics. Physical aging can be explained qualita- example, the time that it is exposed to air at
tively on the basis of the free-volume concept high temperatures should be minimized; like-
so that, for example, the time dependence of wise, light-sensitive materials should be pro-
mechanical properties is found to be inde- tected from exposure to sunlight, light from arc
pendent of chemical structure. lndeed, creep lamps or from some fluorescent lamps. Longer
(stress relaxation) curves of numerous types of wavelength light can be a problem for coloured
polymers measured at various temperatures and materials. In the case of polymers that are to be
aging times can all be superimposed to form a exposed to near ultraviolet (UV) wavelengths
single master curve. (the erythemal, or sunburn region, 290 nm to
315 nm), UV-absorbing stabilizers are com-
The consequences ofphysical aging for the ma- monly incorporated as lowlevel additives.
terials specialist are twofold. On the one hand it Some pigments are IJV protective although
is a major phenomenon that determines the be- the anatase form of TiOz can act as a
haviour of a material to a large extent by chang- photosensitizer.
ing its relaxation times. The ability of rigid
plastics to withstand prolonged stresses, that is, Hydroperoxide groups fastened to polymer
the fact that they can be used as load-bearing chains represent a particularly insidious type of
materials for long periods, is not an inherent thermal and photochemical instability. Present
property but one that is developed by physical just at or below detection limits, hydroperox-
aging during the deformation period. On the ides can initiate the degradation of polymers
other hand, because it persists for a very long containing aliphatic and even aromatic carbon-
time, physical aging affects segmental mobility hydrogen bonds. Sulphur- and phosphorus-
of polymer molecules so it probably also af- containing compounds, as well as hindered
fects chemical degradation, photo-oxidation, amines and nickel chelates, are among the
swelling and de-swelling, in addition to me- highly efficient stabilizers that are added to
chanical relaxation phenomena. A knowledge polymers to protect them by decomposing ad-
of the aging behaviour of a plastic is indispensa- ventitious hydroperoxides to form more stable
ble in the prediction of its long-term properties compounds, before those hydroperoxides can
from short-term tests. Effects on crazing and en- initiate polymer degradation.
vironmental stress-cracking are important but

ll0
It is inevitable that covalent bonds will be bro- R6sum6
ken and radicals will be formed during the fabri-
cation, application and use of macromolecular La transformation des matiriaux polymiriques
materials. With regard to approach (b) men- avec le temps
tioned earlier, stabilization can be achieved by
trapping radicals. The most corrmon kinds of ) I'instar du papier, clu bois et des peaux, pra-
radicals that can be trapped (deactivated) are tiquement tous les plastiques, tolttes les.fibres,
alkyl and alkylperoxide, with the former being tous les caoutchoucs, toutes les peintures et
much more reactive and, in the presence of air, tous les rev€tements protecteurs doivent leurs
caractiristiques utiles d la masse, relativement
converting in what is usually a very fast reac-
tlevde (variant de quelques milliers d plusieurs
tion into the latter. It is very common to use sta-
mi I I i ons ), de I eur s mol t cu I e s. M a I heu reus emen t,
bilizing additives that deactivate alkyl radicals; la fabrication, la manutention et la bonne ou mau-
hindered phenols can protect thermoplastics vaise utilisation de ces matdriaux peuvent modi-
during fabrication (in the melt) as well as dur-
.fi", - voire habituellement abatsser - la masse
ing long-term exposwe to warm temperatures. de leurs moldcules constituantes, et attdnuer
Hindered amines (HALS) react with alkyl and de fagon concomitante leurs propri ttds utiles.
peroxide radicals catalytically, as well as de- L'ttude dans le ddtail des modifications qui se
composing hydroperoxides, and impart un- produisent d l'hchelle moltculaire demeure un
usually good stability to many polymers, secteur scientifique trds complexe, qui est souvent
including paints. dfficile d saisir. Il est toutefois certains aspects,
communs au vieillissement de plusieurs mat6-
A variety of other stabilizers is included in riaux poltmtiques, qui ont ttd dlucidds. et qui
polymer formulations to cope with specific peuvent ainsi out,ir la voie d une amdlioration
instability problems, for example, metal soaps des techniques utilisees pour leur conservation.
to neutralize the HCI generated in the dehy-
drochlorination of poly(vinyl chloride); anti- La digradation chimique de ces mattriaw s'ex-
plique notamment par la ddtbrioration o4tdative
ozonants to protect polymers containing
qui se produit lorsque, exposbs d I'air, ils entrent
carbon-carbon unsaturation, such as diene rub-
en contact avec la lumiire, la chaleur ou des
bers, ffom attack by ambient ozone. lndeed, it
rayonnements ionisants, mais aussi par leur
is noteworthy that synthetic polymers are never ddcomposition sous I'elfet de I'hydrolyse ou de
pure and may include additives of many kinds I'action d'acides ou de bases. La bioddgradation
for the modification of mechanical, swface, peut, par ailleurs, €tre considirbe comme un cos
aesthetic and chemical properties. Few of these particulier de ddcomposition chinique causte
will be significant in contributing to stability. par des enzymes microbiens. Les reactions qui
interviennent dans la dtgradation chimique sont
nombreuses et, dans une certaine mesure, particu-
Conclusion lidres d chaque matidre; ndanmoins, les plus im-
portantes peuvent Ji'2quemmeft ete rattachies
Energy, oxygen and time combine to change d l'oxydation d'hydrocarbures liquides. Pour
the characteristics of polymers. The rates of un- rdduire au minimum les phdnomdnes de decolo-
desirable changes can be reduced by selecting ration, defragilisation et de perte de propietes
the most durable materials initially, combining physiques qui accompagnent la ddtdrioration
o4vdative de ces matdiaux, il sffit d'utiliser un
these where possible with stabilizing com-
mdlange appropri4 d'additifs stabilisateurs. Et la
pounds, reducing the exposure to degradative
meilleure fagon d'attdnuer leur sensibilitd aux mi-
influences, and ensuring the least harmful use
crobes demeure la mise en pratique de bonnes
conditions and service environments. The effec- mithodes d'entretien.
tiveness of such approaches is maximized by
developing a comprehensive understanding of Le vieillissement de ces nrateriaux pollmrbriques
macromolecules and the chemistrv of them. qui, solidifids ou avant atteint autrement leur
fome d2finitive d'utilisation, continueront d se
dbgrader durant des semaines, des mois, voire
des annies, demeure ndanmoins un phdnomine
beaucoup moins bien connu. D'abord observd

lll
dans les themtoplastiques, au-dessotrs de la tem- 8. DurabiIi\, o.f Mao'omolecular Materials,
ptrature de tt'ansition vitreuse (Tv), le vieillisse- American Chemical Society Symposium Series
ment de ces mattriaur se produit parce que No.95. 1979.
toutes les proptibtds lides d la temperatw'e, qui
changent bntsquement d Tv, continuent ndan- 9. Long-term Properties of Polymers and
moins d changer au-dessous de cette temptra- Polym eric Materials, Applied Polymer
ture, mCme si ce n'est que trds lentement. Le Symposium No. 35, 1979.
vieillissement s'observe che: les poll'nr'ires amor-
phes, vitreux, chez les polvmdres semi-cristallins 10. McKellar, J.F. and N.S. Allen, Photochem-
en phase amorphe, de m,?me que dans la matrice istry of Man-made Polymers (London: Applied
caoutchouteuse des caoutchoucs chargtls. Avec le Science Publishers Ltd., 197 9).
tentps, nombre de matdriaux detiennent ainsi plus
rigides et plusfragiles, aufur et d mesure que di- I l. Allara, D.L. and W.L. Hawkins, eds.
ntinue le < taux de relaxation >. Comme le vieillis- Stabilization and Degradation of Polymers,
sement influe sur la mobilite segntentale des Advances in Chemistry Series No. 169,1976.
polymires, son action devrait aussi sefairc sentir
sur la dtgradation chimique, la photo-o4vdation, 12. Ultraviolet Light-induced Reactions in
le gonJlement et le d'lgonflement et les rtactions
ers, American Chemical Society
P ol ym

de rdticulation de toute une gamme de q,s727n"t Symposium Series No. 25,1976.

ntacromoltctilaires.
13. Wiles, D.M. "The Photodegradation of
Fibre-Forming Polymers," in Degradation
References and Stabilization of Polyvners, Chapter 7 ,
G. Geuskens, ed. (London: Applied Science
l. Carlsson D.J. and D.M. Wiles, "Degrada- Publishers, 1975).
tion," in: Encyclopedia of Polymer Science
and Engineering, Volume 4, 2nd edition (New 14. Rinby, B. and J.P. Rabek, Photodegrada-
York: John Wiley & Sons Inc., 1986) tion, Pho to-oxi dation and Ph otostabi lization of
pp.630-696. Polymers; Principles and Applications (New
York: Wiley-Interscience, I 975).
2. Grassie, N., ed. Developments in Polymer
Degradation, volume series, nos. I to 5 15. Road,8.E., P.E. Tomlins and G.D. Dean,
(London: Applied Science Publishers Ltd., "Physical Ageing and Short-term Creep in
1917, 1979, 1981, 1982, 1984). Amorphous and Semi-crystalline Polymers,"
Polymer, vol.7, July 1990, pp. 1204-1215.
3. Jellinek, H.H.G., ed. Degradation and
Stabiliz ation of P olymers (Amsterdam: 16. Bouda, V., "The Nature of Glassy State In-
Elsevier, 1983). stability," Po lymer Degradation and Stability,
vol.24,1989, p. 3 19.
4. Allen, N.S., ed. Degradation and Stqbiliza-
tion of Polltolefns (London: Applied Science 17.ChaL C.X. and N.G. McCrum, "Mechanism
Publishers Ltd., 1983). of Physical Aging in Crystalline Polymers,"
Polymer, vol.21, 1980, p.706.
5. Davis, A. and D. Sims, IAeafuering of
Polwers (London: Applied Science 18. Aklonis, J.J. and W.J. MacKnight,lntroduc-
Publishers, Ltd., 1983). tion to Polymer Viscoelastrciry (New York:
John Wiley and Sons, 1983).
6. Moiseev, V.V. and G.E. Zakov, Chemical
Resistance of Polymers in Aggressive Media 19. Ferry, J.D.,Viscoelastic Properties of
(New York: Plenum Press, 1982). Polymers, 3rd edition (New York: John Wiley
and Sons, 1980).
7. Pappas, S.P. and F.H. Winslow, eds.
P ho t od egr ada ti o n and Ph oto s t abi liz ati on of 20. Stuik, L.C.E. Physical Aging in Amorphous
C o at ings, American Chemical Society Polymers and Other Materials (New York:
Symposium Series No. I5 l, 198 I .
Elsevier.1978).

tt2
The Physical Aging of Polymeric Materials

Chris topher W. Mc Gl inc hey

P aintings Conservation
TheMetropolitan Museum of Art
New York, N.Y.
U.S.A.

Abstract this phenomenon, and make superior products to

the earlier ones comprised of identical materials.
Ph1'sical aging of polymers can be defined as the Distinguishing between dffirences in chemical
changes in spatial arrangement ofmacromole- and physical aging w,ill enable conservators to
cules and side chains w,ith respect to one another. think more clearllt about theirfundamental ap-
This molecular reorganization can cause several proach to the complex aging process of
observable changes. Thermal properties, such as polltmeric materials.
melting point, glass transition and crvstallization,
can become modified, possibly transforming the Introduction
solubilin and optical characteristics in the proc-
ess. More tactile qualities, such asflexibilit-v, em- The elusive property that made macromole-
brittlernent and drape, are also changes brought cules a controversial subject fiom the time of
on bv the process ofphysical aging. their discovery until as recently as the 1920s is
their "polymeric" nature. It was precisely this
Polymer morpholog,,, the study of polymer order, property that prevented macromolecules from
depends upon the chemistry of the polymer, its
being identified using existing methods of frac-
structure and the thermal history to n'hich it has
tionation, purifi cation and crystallization. Re-
been exposed. Though the tetms amorphous and
sults flom these methods appeared to be in
crvstalline are often applied to polvnters, it is
more accurate to think of these terms as extremes direct conflict with results based on diffusion
on a spectrum and to consider that polvmers, de- and viscosity measurements that gave inordi
pending upon their stntcture, have the potential nately high molecular weights. The methods
of being crystallized (achieving a higher state of that failed to properly detect polymers were the
order) or quenched (achieving greater disorder traditional methods that had been successfully
or becoming more amorphous). Physical aging applied for centuries in the characterization of
can have signifcant effects on the aforesaid prop- inorganic and lower molecular weight organic
erties, but unlike chemical degradation, phvsical materials. Early distillations of macromolecular
aging would be completely reversible were it not materials resulted in the decomposition of the
fbr the complication that occurs when chemical parent material into fractions of low molecular
degradation alters the relationship between tem- weight. Neglecting the possibility of degrada-
peratures of degradation and melting point. tion, one attempt to solve this paradigm was
through the concept ofcolloidal forces. These
In the early stages of commercial polymers, colloidal forces, though undetectable, were
physical agingwas poorlv understood and not
thought to reside in the residue of the distilla-
kn ow i ng l1t .fac t o re d i nt o
for mu I a ti o ns. M o re r e - tion process. Two of the prominent scientists
cent polvmer scientists now carefully consider

113
that fi nally settled the "polymer""controversy dependent upon the primary bond forces. Al-
were Staudinger' and Carothers.' They showed though secondary bonds are relatively weak
through slmthesis that macromolecules consist compared to covalent bonds, they can have a
of many (poly) smaller repeat groups (mers) significant cumulative effect on the macro-
joined by conventional covalent bonds. scopic properties of molecules.

These and other historic proofs that led to It is possible that changes brought on by the
the acceptance of polymers soon gave way to altered physical properties of the polymer can
their full scale development and specialization. be reversed in-situ, while those derived from
Polymer materials rapidly began to be reliably chemical degradation cannot. Therefore it can
processed into products and finishes with a be advantageous to determine the source ofob-
wide variety of uses (e.g., automobile tires, servable changes. For example, the yellowing
plexiglass, traffic stripes and vinyl upholstery). of organic materials is obviously due to the
This was largely due to the homogeneity of cer- "irreversible" chemical production of chromo-
tain properties, which assure reproducible acti- phores. But graying, embrittlement and de-
vation conditions, melting point and solubility creased solubility are due to alterations in a
for different batches of the same material. The material's physical properties. Modified physi-
most significant properties of the polymer in- cal properties that result fiom chemical degrada-
clude chemical composition, molecular weight, tion cannot be erased; they merely become the
molecular weight distribution and branching. In new physical properties ofthe particular sys-
addition, slmthetic polymers are more likely to tem. On the other hand, the components of
be free of unstable impurities comlnon to many physical changes not initiated by chemical deg-
nafural polymers, which can initiate degrada- radation are theoretically reversible. It is evi-
tive chemical reactions. However, while these dent that the glass transition temperature (Tg)
new products were likely to have extended life- increases as oxidative degradation proceeds due
times, improperly selected processing condi- to an increase in polar secondary forces. How-
tions may have occasionally shortened their ever, it is also possible for a material to remain
life spans. Polymers do not necessarily possess chemically unchanged while showing an in-
intrinsic properties whose qualities are immedi- crease in Tg through physical aging.
ately apparent. Selection of processing condi-
tions can either bring out desired properties The purpose of this paper is to discuss and illus-
-
transparency, strength and dimensional stability trate the reversible and non-reversible changes
or result in a product that exhibits hastened that occur for thermoplastic polymers after
-embrittlement, warpage or cloudiness. processing and the passage of time. It is hoped
that this information will enlighten conserva-
In order to study the changes in synthetic poly- tors' understanding of these materials, whether
mers upon aging it is necessary to define the they are for their own use or as the medium of
differences between chemical and physical the artist or a previous conservator. This will
properties. Chemical degradation processes are enable conservators to not only choose their
related to changes in the covalent bond struc- treatments more judiciously but also assess the
ture. A covalent bond is defined as the strong aging mechanisms occurring within arfwork
attractive force derived from the sharing of containing such polymers.
two elecffons between two atoms. They are
graphically represented as the lines between Polymer Physical Properties
monomeric units, as well as between the atoms
within each monomer. Physical properties, in Thermoplastic polymers are by definition
addition to being dependent on molecular materials that can be cycled from the solid state
weight, molecular weight distribution and to the fluid state a number of times without a
branching of polymers, also depend upon secon- change in their molecular composition; this
dary bond forces. Secondary forces are attrac- trait is characteristic oftheir physical proper-
tive and repulsive forces that arise from the ties. The term thermoplastic is intended to dis-
local electron densities that surround the tinguish these materials from those that require
atoms in molecules, and are therefore in part chemical reactivity or radiation to cure them

tt4
into an ineversible polymerized state. Thermo- scanning calorimeter (DSC) measurement for
plastic polymers, while stable, are not necessar- the Tg, corresponds to the differences betw.een
ily exempt from thermo- and photo-chemically heat capacity of the liquid and glass states.'Of
induced degradation. Given this finite chemical crystalline polyrners, those that are quenched
stability, the physical changes derived from will show a more intense Tg than those an-
these chemical changes are therefore non- nealed because the phenomenon is derived
reversible. from non-cry.stalline conversely amorphous
domains. ' Below-the Tg the polymer ma-
Thermoplastic polymers have to somehow be -trix is sufficiently restricted in movement to
transferred from their raw material form into prevent further crystallization. The physical
a final shape or finish. This transformation re- processes that take place beneath the Tg in
quires energy that is supplied in the form of amorphous domains are termed physical aging.
heat, pressure and/or solvent action. Whether Physical aging occurs between the glass transi-
polymers in this transient state are in the form tion, Tg, and the next lower secondary transi-
of polymer solutions or are heated above their tion defined as Tg .o Secondary transitions are
melting point (Tm) they are by definition amor- denoted T", Tp, Ty, etc.; T6, corresponds to the
phous. This is because polymers are not locked Tg, while lower transitions maintain their
within an ordered mafrix; neighboring mole- Greek symbols and are due to weaker molecu-
cules randomly exchange their position. Rheol- lar motions at lower temperatures.
ogy, the study of the flow and deformation of
matter, is frequently used to study the response Thermal Processing
of polymer melts and solutions to stresses im- It is important for thermally processed poly-
posed by processing forces. Instrumentation for mers to be stable at the necessary processing
these measurements include viscometers, dy- temperature. If thermo-oxidative degradation is
namic mechanical analyzers, calorimeters, melt expected to occur, antioxidants in small quanti-
rheometers and melt indexers. ties can be introduced to extend the effective
processing time. Common methods of process-
The crystallization of ther^moplastics occurs ing include injection molding, film and fiber
only between Tg and Tm.r The conformation extrusion, calendaring and casting. These opera-
changes required for the transition to a crystal- tions are discussed in greater detail elsewhere.'
line state do not occur instantaneously. Time is Solvents are sometimes added to aid process-
required for polymer chains, side chains and ing. Afterwards, they are usually intended to
functional groups to organize into a three-di- evaporate out but sometimes small amounts re-
mensional order. Annealing (heating a material main as plasticizers (e.g., water in certain nylon
between its Tg and Tm) combined with a slow processes). In processing, polymer choice de-
rate of cooling will optimize the amount of or- pends upon both the intended use ofthe prod-
ganization for a given polymer. Conversely, uct and the actual processing method. There is
quenching (rapidly cooling llom above the no polymer that is ideally suited for all process-
Tm to below the Tg) of a polymer will mini- ing methods. Some polymers (e.g., polyethyl-
mize the crystallinity of its molecular structure. ene) can be easily adapted for different
Thus, it is possible to achieve a range of confor- processing methods, but usually at some com-
mations or morphologies for a polymer by promise, for example, mechanical properties at
controlling the dwation of the cooling period. the expense of optical properties or vice versa.
Annealing can also be govemed by processing
variables, such as pressure, temperature and In thermal processing, crystalline polymers
solvent evaporation. tend to be tumed into fibers and films while
amorphous materials lend themselves well as
Below the Tg, molecular motion is drastically molded objects that need good clarity and iso-
reduced. (Note: Tg is not as precise as melting tropic strength. Synthetic fibers are processed
point; the transition occurs over a temperature from crystalline polymers so that when they are
range.) This reduced molecular motion corre- drawn, their crystalline domains are oriented
sponds to a reduction in heat capacity the inten- along the hber axis (machine direction)
sity, or shift in baseline of the differential

ll5
resulting in their higher strength in that direc- cracking and crazingwithin coatings' derived
tion. For example, nylon fiber after drawing stresses imparted from drying processes, a laser
(extending its length by pulling at a rate faster interferometry technique has been dev,e^loped to
than that at which it comes out of the die exit) make these stress patlems observable.''
has a greater strength in the machine direction.
Additionally, polyethylene pellets are opaque, The preparation of polymer thermoplastic coat-
but after processing, films and fibers can be- ings can include a combination of heat, solvent
come more transparent due to orientation of and pressure. Usually, the less soluble these ma-
crystalline domains previously randomly dis- terials are, the greater their crystallinity. Sol-
tributed. Stress whitening is a result of orient- vent cast films high in crystalline content tend
ing in one direction to such an extent that the to be dwable and impervious to attack by or-
polymer develops hne voids dging processing ganic solvents when dry. (This is the benefit
that cause extensive scattering." gained from materials that require additional en-
ergy to formulate.) Thermoplastic amorphous
Processing forces, in conjunction with the resins with a low Tg are often used as compo-
cooling effect of the mold, transfer significant nents to pressure-sensitive adhesive coatings
stresses into the polymer. Stresses that are because their soft, amorphous domains reduce
formed in pattems are obseryable by the tech- the surface tension of adhesive formulations to
nique of flow bireftingence." If these pattems improve their flow, or tackiness. Amorphous
are not allowed to relax sufficiently they will coatings of higher Tg are commonly found in
become frozen within the material. These fro- over-the-counter applications because of their
zen stress patterns may relieve themselves shelf stability, solubility in mild solvents and
through the process ofphysical aging, or be ease of application. Hard amorphous materials
coaxed along by heating. A significant effort are also those most often used in conservation
has been spent in developing processing opera- practices because they are most likely to remain
tions that ease these stresses prior to being soluble in solvents similar to the ones in which
frozen into materials. they were originally dissolved.

Physical Considerations of Amorphous polymers are easily dissolved be-

Solvent-Borne Polymers cause there is no three-dimensional order that
Those formulating coatings have developed a must be broken up prior to entering the amor-
wide range of solvents and additives to accom- phous liquid state. The lack of order in amor-
modate the many specific end-use requirements phous substances permits more voids into
in the coatings industry. A host of additives which solvent molecules can penetrate, leading
have been developed to prevent almost every to the switching of polymer-polymer interac-
defect known to coatings; some examples are tions to solvent-polymer interactions. These
additives to avert floating, flooding, flattening, voids, or unoccupied spaces between atoms and
fouling, skinning and settling. Rheological molecules, are often referred to as the free vol-
modifiers are also added to improve a coating's ume of the system. Crystalline materials, on the
handling properties during application. While a other hand, require additional energy to break
solvent must have the appropriate solubility up the ordered structure that results from rela-
characteristics to make a proper paint, the rate tively strong polymer-polymer interactions and
at which it evaporates plays a significant role in generally a lower free volume compared to
how the surface sets up to dry. The evaporation amorphous substances.
t'
rate ofa solvent from a paint is different from
the evaporation rate ofthe pure solvent and var- Free volume permits solvent molecules to
ies from polymer to polymer. The surface ten- penetrate and to replace polymer-polymer
sion of the solution (which depends partlqlly interactions with polymer-solvent interactions.
upon the solvent) controls film leveling.'" Un- Since free volume increases with temperature,
suitable surface tension is responsible for such solubility of both amorphous and crystalline
flaws in pppearance as "orange peel" and "cra- materials is increased by heating.
tering."" In order to help study the checking,

l16
While solvent evaporation cannot be strictly lik-
ened to the cooling of polymer melt, some par-
allels do exist; these depend upon the rate of 116
LDPE
\
solvent release and the ability of the polymer
tE \
resin to crystallize. It has been shown that mo- 7a
I

lecular relaxation processes have occurred for I

annealed organic glasses that are the result of =

solvent removal; fu rthermore, these relaxation 375

processes are solvent dependent.'' For coatings

with crystalloid resins the rate of solvent loss
may have a great bearing on the final properties 0
60@ 70@ 80@
TemFatu6 (c)
of the surface coating. This is because the time
required for crystallinity to develop will be
Figure 1 DSC cooling curves for LDPE and
based upon the free volume available per rurit
h-PDCPD
of time; a fast evaporator will cause the free 3

volume to decrease at a fast rate, thereby re-

stricting motions that would be allowable in a
l

slow er evaporating solution. The quickly dried

sample, with its lower crystallinity, will be sub- E
I
I

sequently more soluble. If the Tg of the amor- a

E I
I

h.PDCPD
phous component is low enough, physical
aging processes mentioned above may also re- ;
sult. Latent crystallization can only occur for
materials exposed to temperafures between
their Tg and Tm for significant periods of time.
80@ r@@ 1&00
Results and Discussion
Degree of Crystallinity for Hydrocarbons Figure 2 DSCheating #)i', roru rro
h-PDCPD.
The range of crystallinity can be illustrated
by comparing the DSC cooling curves of two
The effect of processing forces on the thermal
aliphatic hydrocarbon solids, such as low den-
properties of LDPE is determined by analyzing
sity polyethylene (LDPE) and hydrogenated
the DSC curve from the first heating cycle of
poly(dicyclo pentadiene) (h-PDCPD), as shown
the injection molded resin. In this instance,
in Figure l LDPE is a long linear polymer
frozen stresses are derived fiom the imposed
with some branching and h-PDCPD is a cyclic
quench cooling ofthe high-pressure injecting
bridged structure. LDPE crystallizes at 89'C
of molten polymer into a mold much cooler
while the h-PDCPD does not crystallize. This
than the plastic. As a result it does not have the
is to be expected because the highly branched
opportunity to achieve the optimum degree of
structure of the h-PDCPD prevents crystal-
crystallinity. The DSC scan of the injection-
lization. Secondly, DSC heating curves show
molded sample shows a Tg of 4l'C and a Tm
that DCPD has a Tg at 49"C while LDPE has a
of l05oC, which suggests the presence ofboth
Tm of 106'C (Figure 2). Notice that the pro-
amorphous and crystalline components respec-
gression for thermal transition for LDPE is
tively @igure 3). After the sample is annealed
Tg<Tc <Tm, where Tc is the crystallization
at 80oC, prior to the DSC heat, the diminished
temperature. It is clear that the morphology of
Tg is the result of enhanced crystallization.
LDPE may depend upon thermal history where-
as h-PDCPD remains amorphous under the
Thin sections viewed in polarized light
same conditions. The crystalline nature of
show a radial orientation orthogonal to the
LDPE makes it more difficult to dissolve com-
picture plane (Figure 4). Annealed radial and
pared to h-PDCPD, which is more soluble and
tangential samples cut from this circular cross-
hence appropriate as a solvent cast vamish.
section show the stress-release mechanism:

Itl
 cooling results (not shown) are similar to the
h-PDCPD sample. However, first heat DSCs of
a 34-year-old PVAC film, which was mechani-
11
1/l
cally removed fiom a panel painting, indicate
tg il either the release of mechanical stresses in
'il
ltl
amorphous domains or the melting of crystal-
9
I
line domains both in the range of 37'C to 44oC
I
1L (Figure 6). Normally the first heat is discarded
2
L\
I
because aging history makes it complicated and
l
rarely reproducible. This is certainly true here
l because the peak in Figure 6 has taken as long
as 34 years to develop. The fact that this is not
reproducible does not make this result any
Figure 3 DSC heating curves for injection-molded less valuable. It is simply demonstrative of the
and annealed LDPE. aging history of this particular sample. This
phenomenon must take a long time to form be-
cause, as can be seen fiom the annealed sample,
it was not possible to regain this apparent Tm
within the time frame of the DSC experiment.

After heating this sample to 45oC for five min-

utes and then reducing the temperature to 25oC,
visible light transmitklnce. measured spectro-
scopically, increased permanently by 7o/o.lt is
therefore most likely that the opacification after

Figure 4 Thin section cut from injection-molded

LDPE rod, viewed in crossed polars. 350C
unsymmetrical stress distributions in the tangen-
tial sample cause out-of-plane warpage while
the radial cut sample does not (Figure 5). This
warpage is similar to planks of wood cut with 450C
the same orientation, although the derivation
of these stresses is radically different. For com-
plex shapes of molded plastics frozen stresses
can be equally complicated. For early plastics 55nC
(whose color was a mixture of two different
pigmented resins, such as red and white poly-
styrene) the incomplete mixing of these colors
leads to a grain pattern in the product. It is pos-
sible that frozen stresses can follow this srain.
650C
This pattem, while visible. is not n"."ttuiily d.-
monstrative of fiozen stresses; the molding con-
ditions may have eased these stresses by using
a mild cooling temperature for the mold and,/or
750C
a lower injection pressure.

Effect of Thermal History on Varnishes

The amorphicity of poly(vinyl acetate) (PVAC) Figure 5 Heat-treated sample tangentially cutfrom
is illustrated by DSC cooling curves. The DSC LDPE cross-section.

118
 resin and in the remaining two pairs the appear-
ance in effective solubility was about the same
for both samples. These results suggest the
main change in solubility is that the rate at
6
€ which solvent is taken up varies significantly.
B
-5 ..'-4sxeat Given enough time, however, eventual solubil-
I: ity is approximately the same.

This is not to suggest that all the changes in

this film are derived from changes in purely
physical properties. By comparing the IR spec-
tra of this 34-year-old film to fresh PVAC, it is
evident that some degree of chemical degrada-
Figure 6 DSC traces 34-year-old PVACfilm.
tion has taken place (Figure 7). The increase in
polarity from oxidation products chemically
34 years is from the formation of crystalline do-
formed during aging is likely to require a sol-
mains that are large enough to scatter light and
vent ofenhanced polarity to re-dissolve the
that the increase in transparency upon heating
is due to the conversion of crystalline domains
film. It is therefore apparent thatthese materials
change as a result of both chemical and physi-
back into amorphous ones. PVACs, though not
cal properties. Additionally, the change in
normally crystalloid, apparently are to a small
physical properties, the result from the chemi-
degree when given enough time, in this case
cal changes illustrated in the IR spectra, impart
34 years.
their own effect on the physical properties that
are not reversible. This is opposed to the por-
While long-term induced crystallization is cer-
tion ofphysical changes represented by the
tainly responsible for a portion ofthe increased
DSC results that are reversible.
opacification or grayness of these films it is
likely not to be the only reason. The denser
structure associated with the increase in crystal-
linity has (by definition) contracted during the
aging process. Ifsufficient, this contraction
may contribute to the lack of adhesion at the
paint/varnish interface and furthermore may
permit diffirsing air to accumulate in this re-
gion, thereby increasing scattering even more.
The fact that these films are so easily peeled off
of paintings lends credence in some instances
to this hypothesis. It is also possible that soft
resins allow more airbome particles to stick to
their surfaces. Figure 7 Infrared spectra ofPVACs cast on salt
plate.
Once the proportion of crystallinity has been
reduced within the material through heating, Conclusion
solubility properties change. After a five min-
ute exposure to xylene, the post treatment sam- The methods by which thermoplastic polymers
ples imbibe more solvent and coalesce more are made into their final product is as varied as
extensively. After measwing the solvent uptake the class of thermoplastics themselves. While
for five pairs of heated and unheated samples, polymers have intrinsic properties that cannot
on average the preheated sample imbibed be hidden, such observable properties as
0.51 pl and the unffeated film took 0.1I pl in strength, clarity and solubility are contingent
a five minute period. After two hours of expo- upon processing forces. It is understood that
sure the results were less reproducible: in three solvents, temperature and pressure can affect
cases the heated sample cleared away more these observable properties. Stress-release

t19
mechanisms result from frozen stresses of im- stick with wax, were suspended perpendicular
properly molded objects. The process under to the microscope's stage. This assembly was
which these stresses are relieved depend upon attached to a plate of adjustable height on either
polymer structure and ambient exposure condi- side of the stage. The vamish films were placed
tions. After sufficient aging PVAC develops beneath these pipettes and weighed down with
crystalline domains that are large enough to small washers. Prior to filling with solvent, the
scatter light and modify solubility. Addition- pipettes were adjusted to ensr.ue simultaneous
ally, the risks involved in the removal of old contact with each surface. Samples were then
synthetic restorations can possibly be induced rinsed with an aliphatic solvent and allowed to
by warming them to moderate temperatures dry ovemight and then photographed.
(e.g., 40'C) in order to reduce the time required
for their solubility. Visible light transmittance measurements were
calculated at 550 nm using a Perkin-Elmer
Finally, with the ushering in of a class of lambda anay UV/VIS spectrophotometer.
chemically stable synthetic products, yellowing
and other phenomena associated with oxidative Acknowledgement
degradation are no longer the only degradation
processes that divert artistic works from their The author thanks John M. Brealey and Hubert
original meaning. Opacifi cation, scattering and von Sonnenbwg for creating and maintaining
(when appropriate) warpage are changes that the interdisciplinary environment of the Sher-
may or may not be easily reversible. It is our re- man Fairchild Department of Paintings Conser-
sponsibility to leam more about these changes, vation; Jeflrey Jennings and Marie L. O'Shea
if not for the conservator, for the sake of future for providing valuable comments regarding the
generations so they don't romanticize about the subject matter of this paper; and Marycolette
"gorgeous grays" of the 20th century. Hruskocy for providing assistance during the
investigative stages of this paper.
Experimental
R6sum6
The Tg, Tc and Tm of polymers were deter-
mined using a liquid nitrogen cooled differen- Le vieillis sement des matiriaux polyvniriques
tial scanning calorimeter, model DSC 7
(Perkin-Elmer). Temperatures and specific Le vieillissement d'un polymdre peut se ddjinir
heats of transition were calibrated usins indium comme le rdarrangement spatial des macro-
and zinc standards. moldcules et chaines lattrales qui le composent.
Il s'accompagne pafois de plusieurs change-
ments observables. Ses propidtds thermiques
point defusion, sa transitionvitreuse et sa-
Low density polyethylene (Rexene 143,
son
Dupont) was injection molded (at 400'C under
a pressure of 9.65 MPa using a water cooled
cristallinild, notamment
- pourront ainsi se
modifier, tandis que sa solubilitd et ses cqract4ris-
Van Dorn injection molding machine) into the tiques optiques varieront d'autant. Le vieillisse-
form of a rod with a diameter of 5 mm. Plates ment suscite en outre une transformation des
I mm thick were cut from the rod, tangential proprittds plus tactiles du polymdre, telles sa
and radial samples were cut from this plate (Fig-
Jlexibilit4, safriabiliti et son drapt.
ure 1). These samples were placed on a heat-
able stage, Metratherm model 1200d (BBC), La morphologie du polymdre l'6tude de la dis-
and observed through a stereoscope (Zeiss). position des macromolicules -les unes par rapport
Approximate temperatures were determined by aux autres est lide d sa chimie, d sa structure
correcting them against the observed boiling
-
et aux tempdratures auxquelles il a dte expost. Si
point of water. Stress-release mechanisms were I'on qualifie soavent les polymires de cristallins
photographed using a Polaroid c€rmera attach- ou d'amorphes, il serait pourtant plusjuste de
ment on the stereoscope. considirer qu'il ne s'agit ld que de deux ex-
ffAmes, que les polym'ires peuvent, selon leur
Solubility tests were as follows. A pair of structure, devenir plus cristallins lorsque leurs
-
macromoldcules prdsentent, les unes par rapport
2 microliter pipettes, mormted on a wooden

120
aux autres, une disposition rdgulidre ou plus 5. Ferry, J.D., Viscoelastic Properties of Pollt-
amorphes lorsqu'elles offrent une disposition mers,3rd edn. (New York Wiley, 1980) p. 280.
-
moins rtguliire. Le ieillissement peut avoir des
effets importants sur les propi2tds hnumdrdes 6. Struik, L.C.E., "Volume Relaxation in
plus haut mais contrairement d ce qui se pro-
-
duit dans le cas d'une degradation chimique
Polymers," Rheologica Acta, vol. 5, 1966,
-
c es effets demeureraient entidrement reversibles
pp. 303-3 1 1.
si ce n'6tait de cette complication qui survient
quand une degradation chimique altdre le rap-
7.Han,C.D., Rheology in Polwer Processing
(New York: Academic Press, 1976).
port entre les temptratures de dtgradation et le
point de finion.
8. Woodward, A.E., Atlas of Polyner Morphol-
A t'Apoque des premiers polymdres commercianx, ogr (Munich: Hanser Pub., 1989) p.363.
on comprmait mal le vieillissement, et on n'en
tenait guire compte dans I'ilaboration des prdpa- 9. Meeten, G.H., ed., Optical Properties of
rations. Aujourd'hui, les sptcialistes des po- Polwers (London: Elsevier, 1986) p. 101.
lymires prennent soigneusement en compte ce
facteur, si bien que I'on obtient, d partir des 10. Keunigs, R. and D.W. Bousfield, "Analysis
m€mes matidres, des produits suptieurs aux pre- of Surface Tension Driven Leveling in
miers. Forts d'une telle distinction entre la ddgra- Viscoelastic Films," J. Non-Newtonian Fluid
dation chimique et le vieillissement, les Mechanics, vol. 22, I 987, pp. 219-233.
spdcialistes de la restauration seront en mesure
de dtfnir plus clairement leur approche fonda- I l. Quach, A., "Polymer Coatings: Physics
mentaleface d ce processus complexe que consti-
and Mechanics of Leveling," Industrial and
tue le vieillissement des matdriaux polymtiques.
Engineering Chemistry Product Research and
Development,vol. 12,no.2,1973, pp. I l0-l16.
References
12. O'Brien, R.N. and W. Michalik, "Laser
l.Staudinger, H., "The Constitution of
Interferometry for Internal Stress Measurement
Polyoxymethylenes and Other High Molecular
in Coatings," Journal of Coatings Technologt,
Compounds," Helvetia Chimica Acta, 1925,
vol. 58, no. 735, 1986, pp. 25-30.
pp.67-70.
I 3. Morawetz, H., Macromolecules in Solu-
2. Carothers, W.H., "An Introduction to the
tions,2nd edn. (New York: Wiley, 1974).
General Theory of Condensation Polymers,"
Journal of the American Chemical Society,
14. Petrie. S.E.B.. "Thermal Behavior of
vol. 5, no. 1, 1929, pp. 2548-2559.
Annealed Organic Glasses," Journal of
Polwer Science Part A-2, vol. 10,1972,
3. Hall, C., Polymer Materials,2nd edn. (New
pp.1255-1272.
York: Wiley Interscience, 1989) p. 47 .

4. Twi, 8., Thermal Characterization of

P olymericMaterials (Orlando: Academic
Press, 1981) p.167.

t2r
La pr6vision du comportement ir long terme de mat6riaux polymires
synth6tiques doapris des exp6riences de vieillissement artificiel

Jacques Lemaire

Laboratoire de photochimie
Centre national d'|valuation de photoprotection
Univers it6 B lais e-Pas cal
Clermont-Ferrand
France

R6sum6 macroscopiques (propridtds mtcaniques, opacitd,

aspect en surface, etc.). Le vieillissement acctldr!
Nous ddcrirons le photovieillissement des se confond ainsi, enfait, avec une oxydation ac-
p o \,rn'ir e s sy th dti qu es s e re tro uvan t coura m- cdl6rde, et lefacteur d'accilhration peut €tre
ment dans des sculptures modernes qui sont ex- etubh d partir de la vitesse de laformation, dans
posdes dans les musbes extirieurs. Les polyesters des conditions artificielles et naturelles, des pro-
insaturds rdticults ainsi que divers polyacrylates duits d'oxydation critiques. C'est donc dire qu'il
et polymtthacrylates seryiront d'exemples de ma- est possible de prdvoir la duree de vie d'un po-
trices non stabilisdes dont les propi2tds peuvent lymire en sefondant sur les mesures de durde de
€tre modifiees au cours de la polymdrisation, par vie obtenues dans des conditions artifcielles et
sdchage d partir de la solution ou de I'tmulsion en prenant alors en compte lefacteur d'accdl6ra-
et au moyen d'addilifs. Nous ddcrirons dgalement tion. Cette approche mdcanistique se compare
la photo-ox.vdation et I'oxydation thermique bien aux techniques empiriques fonddes sur la
d' dlas tomdres diiniques (butadidne, styrdne- simulation des conditions physiques et chimiques
butadidne et c aoutchouc de ni trile-butadidne), rdelles et sur les variations de proprietes phy-
car on ne saltrait Ahborer une strattgie de siques macroscopiques qui en rtsultent.
stabilisation sans d'abord comprendre le
mdcanis me d' oxvdation. Pour la plupart des polymdres synthttiques, la
degradation oxydative, qui tient soit d la struc-
On peut avoir recours, pour tenter de prd,,toir ture normale du monomdre polymdrise, soit d des
la vitesse du vieillissement des polymdres dtfauts chimiques, s'obserue d un tris foible de-
synthdtiques qui ont 6tt utilists, d I'approche grd d'oxydation. Ilfaut d'is lors avoir recours d
mdcanistique. Aux termes d'une telle approche, des techniques spectrophotomdtriques in situ trds
les polymires solides sont considirds comme des sensibles et informatives pour ddterminer la na-
< rdacteurs > chimiques, c'est-d-dire comme des ture des divers intermediaires et produits finals.
milieux qui sont le site de certaines riactions
chimiques. Elle suppose l'identification des Introduction
mdcanismes qui se produisent, d I'echelle mol6-
culaire, dans des conditions de vieillissement Les mat6riaux polymdres, naturels ou syn-
acc d I drt, artifi ci e I les, et d' uti I isation normale, th6tiques, que I'on utilise soit pour la conserya-
naturelles. Les effets les plus ndfastes proviennent tion et la restauration des cuwes d'art, soit
de I'oxydation causde par une tropforte exposi- pour leur 6laboration, pr6sentent I'inconv6nient
tion d la lumidre, d la chaleur ou d d'autres
majeur d'6voluer dans le temps, avec une perte
contraintes statiques et dvnamiques. Et une
progressive de leurs propri6t6s physiques et des
analyse chimique poussde permettra d' expl iquer
modifications d'aspect. Les ph6nomdnes de
les modifications de propri6tds physiques

123
vieillissement, qui sont reconnus comme com- chimiques (oxyd6s, par exemple) et cette
plexes depuis fort longtemps, ne font en fait 6volution chimique est responsable de la
I'objet d'analyses que depuis vingt ans, leur d6gradation des propri6t6s physiques. Le
6tude ayant 6t6 largement diff6r6e d la suite des vieillissement physique des polymdres n'in-
premiers succds obtenus gr6ce ir des m6thodes tervient que dans le cas des mat6riaux qui
trds empiriques. sont soumis d des contraintes physiques trds
importantes, et il n'a donc aucune importance
Les premiers probldmes de fiabilit6 des dans le cas des auwes d'art.
mat6riaux polymdres en cours d'usage se
sont en effet pos6s dds les ann6es 1945-1950 o L'6volution chimique ne d6pend pas
aux m6caniciens qui utilisaient des pidces en
des contraintes m6caniques qui sont appli-
plastique. Ces m6caniciens n'ont alors con-
qu6es; ces contraintes ne modifient que
sid6r6 que les aspects macroscopiques du vieil-
les cons6quences physiques de l'6volu-
lissement des systdmes, et ils n'ont 6tudi6 que
tion chimique.
les variations des propri6t6s physiques utiles en
appliquant d ces systdmes, en laboratoire, des
contraintes physiques et chimiques repr6senta-
. L'acc6l6ration de l'6volution chimique ne
tives de celles auxquelles ils sont soumis en d6pend donc pas des contraintes m6caniques
cours d'usage. Par exemple, le vieillissement externes et intemes. L'analyse de l'6volution
en conditions climatiques 6tait examin6 dans chimique permet de convertir la dur6e d'ex-
des enceintes de < simulation > qui reprodui- p6riences en laboratoire en dur6e d'usage
saient les contraintes naturelles de I'environne- dans des conditions naturelles (l'6volution
ment (la lumidre, la chalew, 1'Oz, l'HzO, etc.). chimique constituant dds lors la base de tout
De telles m6thodes ne permettaient de pr6voir transfert de donn6es).
le comportement d long terme des polymdres
qu'en 6tablissant une < corr6lation > avec les o L'6volution chimique est une caract6ristique
situations r6elles de vieillissement. Ces de chaque m6canisme d'6volution d'un
m6thodes de contr6le plutdt que de pr6vi- mat6riau donn6. Une formulation pr6ci-
sion - gudre 6volu6 depuis 1950; leur
n'ont se (polymdre + charges + pigments ou colo-
- s'est par contre beaucoup d6velopp6, et
emploi rants + additifs) doit 6tre caract6ris6e par un
elles ont foumi de nombreux resultats que I'on facteur d'acc6l6ration qui lui est propre. On
n'arrive pas d ( rationaliser >. ne peut transferer directement un classement
de divers mat6riaux ou de diverses formula-
A partir des ann6es 70, une approche plus tions obtenu en laboratoire aux conditions du
cognitive du comportement ?r long terme des
terrain sans tenir compte de facteurs d'ac-
mat6riaux polymdres s'est d6velopp6e, qui se
c6l6ration n6cessairement diff6rents.
fondait sur des analyses, au niveau mol6culaire,
des modifications chimiques des chaines macro-
o L'acc6l6ration des 6volutions chimiques est
mol6culaires qui apparaissent en cours d'usage.
Cette approche a surtout 6t6 utilis6e pour les non seulement autoris6e, mais elle est fonda-
polymdres synth6tiques qui, conhairement mentalement indispensable car :

aux mat6riaux organiques naturels, pr6sentent

un ordre au niveau mol6culaire, avec une a) on ne sait pas extrapoler les donn6es
r6p6tition plus ou moins r6gulidre de I'unit6 recueillies dans les phases pr6coces d'6vo-
monomdre. Acfuellement, on peut consid6rer lution des mat6riaux (les m6thodes de la
cornme acquis les principes suivants : cin6tique homogdne ne sont pas
acceptables);
o Un mat6riau polymdre 6volue ir la fagon d'un
r6acteur chimique ou photochimique; sa b) on doit amener le mat6riau d un niveau
d6gradation suppose I'apparition de concen- d'6volution chimique qui enhaine une
trations g6n6ralement faibles de groupements d6gradation m6canique.

124
o Il est, par contre, indispensable de provoquer - La description bas6e sur les produits obser-
une acc6l6ration des 6v6nements chimi- vables en spectroscopie vibrationnelle (in-
ques en maintenant leur repr6sentativit6. frarouge ou Raman) doit Otre associ6e aux
Cette repr6sentativit6 doit d'abord 6tre la variations de propri6t6s m6caniques. Ces pro-
cons6quence des contraintes physiques et duits correspondent en effet aux voies princi
chimiques qui sont appliqu6es, lesquelles pales d'oxydation: ils sont en concentrations
doivent elles-m6mes €tre repr6sentatives faibles mais leur apparition correspond ir une
des contraintes en cours d'usage. La repr6- vraie d6t6rioration de la matrice.
sentativit6 doit 6tre surtout v6rifi6e en com-
parant les m6canismes r6actionnels en - La description bas6e sur les produits
conditions acc6l6r6es, d'une part, et en condi- observables en spectroscopie 6lec-
tions d'usage, d'autre part. Dans ce domaine, tronique (ultraviolet, lumidre visible, colo-
un m6canisme r6actionnel peut Cfre d6crit d rim6trie ou 6mission de fluorescence) doit
l'aide de s6quences r6actionnelles (et non de Otre associ6e aux variations d'aspect. Ces pro-
processus 6l6mentaires), et chaque s6quence duits les produits de jaunissement, par
r6actionnelle doit 6tre reconnue dans les 6vo- exemple- sont d6tect6s ir des concentrations
lutions artificielles et naturelles. Il s'agit ld de
-
trds inf6rieures d celles des produits princi-
la base de I'approche << m6canistique > qui, de paux d'oxydation observ6s en infrarouge ou
I 975 d nos jours, a pu 6tre d6velopp6e dans en Raman, et ils ne correspondent g6n6rale-
notre laboratoire pour la plupart des po- ment pas d une perte des propri6t6s
lymdres couramment utilisds. m6caniques de la matrice.

o Tout vieillissement acc6l6r6 correspond Il est donc possible de pr6voir la dur6e de vie
en fait d une acc6l6ration de l'6volution d'un mat6riau polymdre soumis ir des contrain-
chimique. Quand un m6canisme chimique tes lumineuses ou thermiques en pr6sence d'Oz
suppose plusieurs chemins r6actionnels d'im- et d'eau :
portance 6quivalente, on ne peut esp6rer ac-
c6l6rer tous ces chemins r6actionnels avec les
r en d6terminant, en conditions acc6l6r6es de
m€mes facteurs d'acc6l6ration. L'exp6rience laboratoire, la cin6tique de I'apparition d'un
en laboratoire d6forme alors la r6alit6 des < produit critique >>, c'est-d-dire un produit
conditions d'usage. De m€me, si, d l'6volu- qui s'accumule, selon une loi lin6aire, avec la
tion chimique, se supelposent des ph6no- dur6e d'exposition et dont I'apparition traduit
mdnes de transfert physique (oxygdne ou une coupure de chaine macromol6culaire;
stabilisant, par exemple), la pr6sence de
o en d6terminant, toujours en conditions
plusieurs processus dynamiques interdit le
transfert des conditions acc6l6r6es cr66es en acc6l6r6es de laboratoire, la corr6lation exis-
laboratoire aux conditions d'usage. On ne tant entre les variations de propri6t6s macro-
peut op6rer ce transfert que pour les systdmes scopiques d'usage (propri6t6s m6caniques ou
dont l'6volution chimique est contr6l6e par aspect) et les variations de la concentration de
un m6canisme chimique pr6pond6rant. ce produit critique; on pr6cise ainsi le seuil
Cette situation se pr6sente assez souvent tol6rable d'6volution chimique;
dans le cas du photovieillissement, of le
m6canisme photo-oxydatif est alors contr6- . en comparant la cindtique de I'apparition
lant et les ph6nomdnes de diffusion de l'oxy- du produit critique en conditions acc6l6r6es,
gdne, suffisamment rapides pour ne pas d'une part, et dans une phase pr6coce de
limiter I'oxydation. vieillissement en cours d'usage, d'autre part,
on obtient le facteur d'acc6l6ration qui permet
o La description de l'6volution chimique doit, de convertir la dur6e d'exposition en labora-
enfin, Otre associ6e au critdre de d6gradation toire en duree d'usage.
choisi :

t25
Dans le pr6sent expos6, trois classes exem- Ces polymdres se retrouvent souvent dans des
plaires de polymdres seront 6voqu6es, d savoir : sculptures contemporaines (de T. Grand, de
Niki de Saint Phale ou de J. P. Raynaud, par
l. la classe des polyesters insatur6s, comme exemple). L'insolubilit6 de ces polymdres
exemples de mat6riaux polymdres dont le r6ticul6s rend l'analyse de l'6volution chimique
comportement d long terme peut etre particulidrement difficile. Il a 6t6 possible nean-
modul6 par la structure m6me du polymdre moins de d6montrer les faits suivants :
et par f introduction d'additifs stabilisants;
l. Les groupements responsables de I'ab-
2. polyacrylates et des
la classe des sorption de la lumidre naturelle ext6rieure
polym6thacrylates, comme exemples de (1, > 300 nm) ou int6rieure (1, > 340 nm)
polymdres relativement stables dont le sont les groupements de mal6ates-fumarates
comportement d long terme dEpend pourtant r6siduels et les groupements de phtalates, les
dans une large mesure du compos6 6tranger oligomdres de styrdne n'6tant pas suscep-
qui a 6t6 introduit; tibles d'absorber cette lumidre. Ces
groupements absorbants provoquent, par
3. la classe des 6lastomdres di6niques, comme dissociation en radicaux, I'amorgage de la
exemples de polymdres dont le comporte- photo-oxydation.
ment d long terme a 6t6 pr6l'u, d tort, d partir
d' essais non repr6sentatifs. 2. Ces mat6riaux sont le sidge de ph6nomdnes
de < jaunissement )) photolyique (lapho-
La photolyse et la photo-oxydation tolyse d6signant des 6v6nements photo-
des polyesters insatur6s chimiques non modifiables par l'oxygdne).
Ces produits de la photolyse qui absorbent
Les polyesters insatw6s sont des polymdres de les rayons de la lumidre visible (),:: 400 nm)
condensation r6sultant de l'action de diacides peuvent €ne d6truits par photo-oxydation.
sur des dialcools (glycols). L'un au moins des
diacides est insatur6, ce qui permet la r6ticula- 3. Ces mat6riaux sont simultan6ment le sidge
tion par oligom6risation in situ d'un agent de de ph6nomdnes photo-oxydatifs. Cette
r6ticulation. On obtient alors un r6seau tridi photo-oxydation qui d6pend de la nature
mensionnel donnant au produit final son carac- -
des diacides et du glycol utilis6s, du r6ticu-
tdre d'irr6versibilit6 thermique le classant dans lant (styrdne ou acrylates) et des modes de
les thermodurcissables. Si I'on utilise le styrdne provoque 6galement des
r6ticulation
comme agent de r6ticulation, on obtient des -
jaunissements, mais elle est surtout respon-
polyesters insatur6s de structure :
sable de I'apparition de groupements oxy-
d6s (groupements hydroxyl6s et carbonyl6s),
qui s'accumulent et qui sont associ6s aux
variations de propri6t6s physiques (des
o o microfi ssurations superficielles,
rr I IcH-c-
rl
o-R-o- par exemple).
-c-cH- I
T" 4. Les jaunissements qui apparaissent de
H-
+o fagon plus importante en I'absence ou en
d6ficit d'oxygdne qu'en pr6sence d'oxygdne
sont dus soit aux seuls groupements de
styrdne (dans les polyesters insatur6s
6labor6s uniquement d partir d'anhydride
oi R peut Otre un noyau aromatique di- mal6ique), soit ir l'action conjugu6e de grou-
substitu6 en ortho ou en m6ta selon que le pements de phtalates et de styrdne (dans les
diacide utilis6 est un acide orthophtalique polyesters 6labor6s d partir de m6langes
ou isophtalique. d'anhydride mal6ique et d'acide phtalique).

126
5. L'oxydabilit6 des polyesters insatur6s photo-amorceurs, les unit6s r6actives et la na-
est surtout attribuable aux insaturations de ture des produits observ6s. Le tableau I porte
mal6ates ou de fumarates qui n'ont pas r6agi sur les polyesters obtenus d partir d'un melange
dans la r6ticulation. Une r6ticulation obte- d'acide phtalique et d'anhydride maldique et le
nue avec du styrdne r6duira ainsi I'oxydabi- tableau II, sur les polyesters obtenus d partir
1it6, mais elle accentuera le pouvoir jaunis- d'anhydride mal6ique.
sant. Une r6ticulation obtenue avec un acry-
late et un m6thacrylate diminuera le pouvoir Puisque la structure mol6culaire d'un polyester
jaunissant, mais le taux d'insaturation r6si- insatur6 r6ticul6 peut varier trds largement
selon la nature des monomdres, des r6ticulants
duel sera plus important et l'oxydabilit6,
et du mode de r6ticulation, la photor6activit6
plus grande. Dans ce dernier cas, les cou-
sera tout aussi variable. Elle pourra ainsi notam-
ches superficielles oxyddes donneront lieu
ment Otre adapt6e pour obtenir, par exemple,
ir des microfissurations, et il en r6sultera
une microfissuration r6duite, si I'on accepte
un blanchissement.
qu'un jaunissement important puisse se pro-
duire. En fait, la durabilit6 de ce mat6riau peut
6. La nature du dialcool est 6galement impor-
6tre largement am6lior6e grdce d l'emploi de
tante I'usage de n6opentylglycol permet
- I'oxydabilit6, par exemple.
de reduire
photostabilisants. Lorsqu'un polyester est si-
multan6ment le sidge d'une photolyse et d'une
photo-oxydation, ou d'un autre double proces-
Les mecanismes d'6volution chimique des sus photochimique du genre, il convient d'utili-
polyesters insatur6s r6ticul6s d l'aide de styrdne ser un m6lange form6 des substances suivantes :
sont sch6matis6s dans les tableaux I et II,
qui indiquent les groupes chromophores r un < anti-U. Vx. )) qui inhibera partiellement
la photolyse, ainsi que la photo-oxidation

Tableau I
Les micanismes d'bvolution chimique des polyesters obtenus d partir d'un mtlange
d'acide phtalique et d'anhydride mal4ique

Groupes chrornophores
photo-amorceurs Unlt6s r6actlves Prodults

-OCOCH=CHC@-
prodults
\ absorbant
\ [-n .l R. *o -ococH-cHcoo- en IR
I
l--- R:'
2
I
li (hydroxyles
I
et
" '-
n"./'lR' :I -cH- CH
2
carbonyl6s )
I
ge
cage hors gHt
----l
lMal6ates- l' '/ I caSe
Funarates
I | | n
I
-cH 2 -c-
t-
f
^-l
\7

t27
 Tableau II
Les mtcanismes d'4tolution chimique des polyesters obtenus d partir d'anh1'dride mal2ique

Groupes chronophores Unlt6s r6actlves Prodults

photo-anorceurs
-OCOCH=CHCOO-
produl ts
\ absorbant
Fnl R.*o -ococH-cH@o- en lR
luat6ates-l
l--____-;
hu l'l+ I_-_) + __2
l_---)l I (hydroxyl6s
lrunaratesl I I

In'l
L". J R'
-cH-cH
et
carbonyl6s )
cag€
l2
hora CH
cage

,0
H
I

tant de la matrice que des produits de la circonstance se rencontre, en fait, de fagon clas-
photolyse (par comp6tition d' absorption); sique, dans tous les mat6riaux polymdres qui
n'absorbent pas la lumiere lorsque )" > 300 nm;
. un ( antioxydant redox > qui n'inhibera que l'amorgage photochimique de leur oxydation
la photo-oxydation de la matrice (sans inhi- est 6galement dt d des compos6s 6trangers
bition de la photo-oxydation des produits chromophores et photo-inducteurs. Mais, dans
de la photolyse). la plupart des cas, cet amorEage ne correspond
qu'd une phase initiale, les groupements oxyd6s
L'emploi d'un seul anti-U. V. risque d'accen- form6s sur la chaine principale du polymere
tuer le photojaunissement. Il faut l'assister d'un jouant trds vite le r6le de chromophores
antioxydant redox qui ne modifiera que les photo-amorceurs. Les poly(m6thacrylates
oxydations radicalaires sans perturber les de m6thyle), ou PMMA, se caract6risent, au
photo-oxydations directes. contraire, par le fait que leur durde de vie sem-
ble compldtement tributaire des contaminants.
La photo-oxydation
La photo-oxydation du PMMA se traduit par
des d6riv6s acryliques
I'apparition de produits hydroxyl6s observa-
Les polyacrylates et les polym6thacrylates bles en spectroscopie infrarouge ii transform6e
de Fourier (IRTF) e 3 580 et3 320 cm-'. Ces
sont des polymdres satur6s qui ne pr6sentent
pas de sites intrinsdques de fragilit6 photo- produits, qui atteignent une concentration
lytique et photo-oxydative. Les processus de constante au terme d'une phase pr6coce de for-
Nonish de type II qui interviennent dans les mation, ont 6te identifi6s d des groupements
esters satur6s ont une importance r6duite. Par alcooliques associ6s, par liaisons < hydro-
gdne >. soit aux groupements esters (ir .
contre, ils peuvent 6tre le sidge d'une oxydation
photo-induite par tout compos6 6tranger suscep- 3 580 cm-') soit entre eux (d 3 320 cm-'). Les
tible d'absorber la lumidre et de donner nais- produits hydroxyl6s sont les seuls photo-pro-
sance d des radicaux. Ces polymdres sont donc duits ais6ment observables; il se forme 6gale-
trds sensibles d toute contamination (introduite, ment des groupements responsables de l'6lar-
au moment de l'6laboration ou avec le temps, gissement de la bande d'absorption des groupe-
par migration dans les systdmes solides). Cette ments esters, mais il est impossible d'attribuer
des caractdres quantitatifs ir cette observation.

128
A grandes longueurs d'onde (1, > 300 nm), la absorption au-deld de 300 nm dans les spectres
photo-oxydation est essentiellement le fait de ultraviolets de plaques 6paisses (de 3 d 6 mm).
compos6s extrinsdques chromophores et photo-
inducteurs. Aucun photoproduit chromophore Le m6canisme de I'oxydation photo-induite
n'est form6 au cours de la photo-oxydation du PMMA peut 0tre repr6sent6 ainsi :
du PMMA et n'assure le relais des photo-
inducteurs initiaux. La photor6action s'arr€te
donc quand le photo-inducteur est entidrement
X (photo-lnducteur)
consomm6; ceci intervient 6galement ir trds
faible taux d'oxydation de la matrice de I r',
PMMA expos6e sous forme de film.
'' '*'o"' '

Quand, par conhe, une plaque 6paisse de

PMMA est expos6e en conditions accel6r6es
ou en conditions naturelles, la photo-oxydation cnz-c
l*'
est localis6e dans les quelque I 000 pm de la
- -
couche superfrcielle; elle r6sulte tant de la diffu- *flt'
sion d'oxygdne dans la matrice que d'une diffu-
I
sion qui se produit en direction oppos6e aux
photo-inducteurs. Cette deuxidme diffusion a
6t6 mise en 6vidence lors d'une exp6rience,
. -c
cH
i",
au cours de laquelle deux photo-oxydations - -
sont intervenues successivement, et of les
400 premiers pm les plus photo-oxyd6s ont 6t6
pr6lev6s au terme de la premidre photo-oxyda-
'f" I

tion. La deuxidme photo-oxydation s'est pro-

l02'PH
J
duite d des vitesses nettement inferieures ir
celles qui avaient 6t6 observ6es lors de la pre-
midre. Cette remarque justifie d'ailleurs les
techniques de r6novation du PMMA qui se fon-
dent sur une 6limination des couches oxyd6es
les plus superfi cielles.

Les insaturations r6siduelles sont des sites r6ac-

tifs qui disparaissent au cours de l'oxydation de
la mafrice de PMMA, bien que, n'absorbant pas _C_C_ +.OH
les photons de l" >300 nm, elles ne soient pas
responsables de I'amorgage de la photo-
oxydation du PMMA.

La fonction d'un photo-inducteur

-
c'est-2r- flr
dire d'un compos6 susceptible d'absorber la 9Hc
I

lumidre et de donner naissance d des radicaux I -cH-c-

tl oc
olll'
libres r6actifs, qui ne se recombineront donc -C-C-+H-O
pas entre estpeu sp6cifique. La nature coocH3 \HO'ffir
l\
des contaminants chromophores du PMMA
peut etre trds vari6e et d6pendre des conditions
qui ont pr6sid6 d l'6laboration, ir la mise en (non obeervable) groupcnenta hydroxyles
cuvre et d la formulation du PMMA qui fait assocl6e cntre eux
I'objet de I'analyse, et il n'est pas utile de
pr6ciser la nature mol6culaire du photo-induc-
teur, dont la pr6sence se manifestera par une PH: Hydrogdne du polltmdre

t29
Les polyacrylates d'alkyle pr6sentent des Depuis 1985, nous analysons le m6canisme
m6canismes de photo-oxydation analogues ?r de photo-oxydation de diff6rents types de poly-
ceux du PMMA. Un ph6nomdne nouveau peut butadidnes (BR), de polyisoprdnes (IR) et de
apparaitre, par contre, quand on polym6rise copolymdres styrdne-butadidne (SBR) ou acry-
in situ des monomdres ou des oligomdres lonitrile-butadidne (NBR). Nous avons proc6d6
d'acrylates pour constituer des couches adh6- d une etude d6taill6e des evolutions photo-
sives d'assez forte 6paisseur. Cette polym6risa- chimiques en utilisant diverses techniques
tion initiale est souvent inhib6e par I'oxygdne spectroscopiques (infrarouge, IRTF, micro-
et elle ne s'effectue que trds partiellement; les IRTF, ultraviolet ou Raman). En outre, les per-
insaturations r6siduelles restent donc en concen- turbations apport6es par des noirs de carbone
trations importantes au sein de la matrice au et des oxydes photo-actifs ont 6t6 analys6es.
terme de la polym6risation. Ces insaturations L'emploi de nouvelles techniques analyiques
r6siduelles sont susceptibles de donner lieu d bien adapt6es aux milieux trds opaques la
une polym6risation d long terme et il se cr6e -
microspectrophotom6trie IRTF et la spectro-
alors un r6seau polym6rique dans la matrice photom6trie IRTF avec d6tecteur opto-
initiale, avec une r6duction du volume occup6. acoustique, par exemple ont permis de
Il apparait alors des zones d'h6t6rog6n6it6 visi- reconnaitre, lors de l'6volution de mat6riaux fi-
bles d I'ail nu. Ce ph6nomdne est courant dans nis (rulcanis6s et pigment6s), les m6canismes
le cas d'cuvres oi de fortes 6paisseurs d'ad- qui interviennent dans le cas d'6lastomdres
h6sifs acryliques ont 6t6 utilis6es. non transform6s.

La photo-oxydation Comme pour tout pollmdre < non

d'6lastomires di6niques absorbant >, la lumidre est en fait absorb6e par
des chromophores ext6rieurs aux 6lastomdres
Jusqu'2r ces dernidres ann6es, les 6lastomdres examin6s (probablement des impuret6s ou des
di6niques n'ont pu, essentiellement d cause de d6fauts que l'on ne peut contr6ler). En outre,
facteurs d'ordre analytique, 6tre examin6s dans la vitesse initiale de photo-oxydation, g6n6rale-
Ie cadre de I'approche < m6canistique >. Dans ment trds faible, est une donn6e qui d6pend es-
leur forme finie, ces materiarx sont fort com- sentiellement des antioxydants r6siduels, et la
plexes et profond6ment modifi6s par la lulcani- p6riode d'induction ne peut 6tre consid6r6e que
sation ou la r6ticulation, et ils contiennent comme une caract6ristique extrinsdque. Cette
g6n6ralement des pourcentages 6lev6s de p6riode d'induction n'excede pas une heure
charges absorbantes en ultraviolet, en lumidre dans les conditions de photo-oxydation ex-
visible et en infrarouge. Les m6thodes spectro- ploit6es dans la cadre de la pr6sente 6tude.
scopiques qui sont habituellement utilis6es Au terme de cette p6riode, les contaminants
pour observer les quelques modifications inhibiteurs sont consomm6s et des chromopho-
chimiques des chaines polym6riques (d un de- res intrinsdques se forment (hydroperoxydes,
oh) ne c6tones a-B insatur6es, c6tones satur6es). La
916 d'avancement souvent inferieur d 1
peuvent plus 6tre mises en auwe ais6ment. photo-oxydation vraie de 1'6lastomdre inter-
La durabilit6 des 6lastomdres synth6tiques vient alors, selon le sch6ma de la page suivante.
6labor6s d partir de didnes n'avait donc 6t6
examin6e que sur le plan macroscopique, le On sait que I'oxydation radicalaire des alcdnes
vieillissement de ces mat6riaux expos6s en suppose g6n6ralement une extraction d'hy-
conditions naturelles ou simul6es n'6tant carac- drogdne sur le carbone situ6 en a de la double
t6ris6 que par des variations de propri6t6s liaison. Ce m6canisme d'amorgage par extrac-
mecaniques ou d'aspect (des microfissurations tion entre en comp6tition avec I'addition du
superficielles, par exemple). Une telle approche radical avec la double liaison. Une r6action
n'a pu foumir, d 1'6vidence, aucune indication en chaine d'hydroperoxydation permet d'ex-
sur la nature exacte de l'6volution chimique, les pliquer la formation d'hydroperoxydes o-B
seules connaissances de cet ordre n'ayant pu insatur6s. Aux hydroperoxydes associes qui
6tre acquises que sur des compos6s moddles de apparaissent d des concentrations maximales
faible poids moleculaire. de 60 mmol.ks-' doit 6tre attribuee la bande

130
d'absorption d 3 400 cm-l
environ, qui n'est pas
observ6e en oxydation I
o,
thermique d 60"C et qui J*'
disparait par photolyse r. (rO^.
a\ )
sous vide d 35"C.
+m \-
Il faut admettre que I'hy- | \ addition g.r
J aonule liaiaona
droperoxydation en cr de
{llr{11:O{ll-
la double liaison intervient
+l rH(ro2tl)
avant toute saturation de
cette double liaison pour
expliquer la formation de
l.'
-{Hfcrbcr+-cx- -oEcr- ) -cx.-qf,}FcH2-
c6tones a-p insatur6es par
photolyse (ou thermolyse) o
de ces hydroperoxydes. Si
t.
rO
les c6tones cr-B insatur6es OH
ttl Jtt
absorbant d 1 696 cm-' se -cH2-cH-cH-cH- -cHfcH{t-cH-
forment dans une r6action
en cage, habituellement I o
I
rencontr6e dans les po- lr6rctio da
OH
satuation
lymdres, les groupements I
nrl
alcooliques r6sultent de I
t
radicaux alcenoxydes qui -CH"-CH:CI|-CH- q4 -q1^-611:CK-C1t- +
ne sont pas recombin6s ' I t l.
OH'

dans la cage. oH
1o.u*.i*
cag€
La bande d'absorption i Jen
| 726 cm'' a ete attribuee -CH2-CH:CH-C- + H^O
aux groupements c6toni- o
ll'
ques saturds que I'on peut
Fatofi-l \--,- r\
I gaturdes I -l r6rctidl ds I l. \r
observer ais6ment lors ,rl ,"i*"u"" I lhu'oz \ i"*'etiEation cis-trans
d'oxydations thermiques. l-;F;6
I Eatu€E I
I r--.-\ +
acides cr, p inaatur6g
En photo-oxydation, les
c6tones ne peuvent ap-
paraitre qu'en faibles con- PH: Hydrogdne du polvnftre
centrations stationnaires car les processus
photochimiques de Norrish de types I et II pro- montrent que la photo-oxydation des polybu-
voquent leur conversion. En particulier, la for- . tadidnes ne depend guere du genre de micro-
mation d'acides satur6s (absorbant d | 717 cm-' structure, puisque toute insaturation, quelle
environ) intervient dans toute matrice poly- qu'elle soit, entraine une forte photo-oxyda-
m6rique oi sont form6es interm6diairement bilit6 de ces 6lastomdres ind6pendamment
des c6tones satur6es; cette conversion de c6to- de la structure.
nes satur6es en acides suppose un processus
de Norrish de type I. La photo-oxydation d'un 6lastomdre se traduit
6galement par une modification importante du
La disparition des insaturations r6siduelles est r6seau. Aux scissions de chaines intervenant
directement observ6e d partir des spectres infra- lors de la formation des groupements oxyd6s,
rouges et elle est i l'origine de la formation de s'opposent des r6ticulations. Les photopassiva-
tous les produits d'oxydation. En outre, cette tions tres remarquables, observ6es m6me en
disparition permet d' interpr6ter I a r6ticulation films minces (de I 50 d 2 l0 pm), s'interprdtent
observ6e en R.M.N. Les donn6es cin6tiques par une augmentation de l'imperm6abilit6 d

l3l
I'oxygdne des couches superficielles oxyd6es 2. Les radicaux alc6noxydes correspondants
et r6ticul6es. Il convient de noter que, en donnent naissance i des processus plus
photovieillissement artificiel d'6lastomdres vari6s :
non transform6s, les films deviennent, sous CH:
exposition, des photordacteurs trds h6t6rogdnes I

qui ne s'oxydent que superficiellement. Le

caur du film n'est alors pratiquement
pas modifie. PH
/I
-cH-cH-c-cH2-

/u'
CH: / -
CH"-CO-GH,- +'cH-cH-
Les copolymdres SBR et NBR 6tudi6s se sont
V (i | 722 cn-r)
comport6s comme les homopolymdres de poly- -cH:cH-c-cH"-
butadidne; les unit6s de styrdne et d'acryloni-
l'
I
\
trile sont demeur6es pratiquement inertes au --ctt-cn-co-cH2-
cours de la photo-oxydation, ce qui signifie \ +'cH3
(a 1 6e3 cm-l)
que tous les radicaux produits dans la matrice
\
r6agissent en fait sur les sites insatur6s.
cH-cH-co-cH3 + 'cHr-
Les polyisoprdnes pr6sentent des photo- (A I 693 crn-1)
oxydations qui interviennent selon le m6me
m6canisme g6n6ral. Mais la structure m6me Dans le cas du polyisoprdne, les processus
du polyisoprdne entraine les differences de B-scission des radicaux alc6noxydes provo-
suivantes : quent donc des coupures de chaines. Il s'agit ld
d'un facteur qui diff6rencie nettement le poly-
l. La dissym6trie de I'unit6 isopr6nique laisse isoprdne des autres 6lastomdres pour lesquels
pr6voir la formation de deux radicaux diff6- les coupures de chaines interviennent par d'au-
rents par r6action d'extraction d'hydrogdne : tres mecanismes moins primaires et apparem-
cHs cHs
ment moins fr6quents.
I
Il est maintenant possible de d6crire l'6volu-
-CH-CH-C-CH2- ou -CH-C-CH-CH2-
tion chimique d'6lastomdres di6niques non
(a) (b) transform6s d I'aide des diff6rents produits
interm6diaires et frnals qui se forment au cours
Les hydroperoxydes apparaissent en plus
d'oxydations photothermiques (d basse tem-
p6rature) ou thermiques. Il est donc possible
fortes concenfrations dans les polyisoprdnes
de comparer cette 6volution en photovieillisse-
que dans les autres 6lastomdres, ce qui est com-
ment ou en thermovieillissement artificiel et
patible avec une structwe tertiaire de ces hy-
celle qui se produira en vieillissement clima-
droperoxydes. Les radicaux (a) semblent donc
tique. Il apparait alors clairement que le vieillis-
6tre essentiellement form6s et leurs formes
sement climatique d'6lastomdres di6niques non
m6somdres :
transform6s se r6duit essentiellement d une oxy-
CH. cHe dation photochimique et qu'il ne correspond, en
t-
-cH-cH:c-cH2 <-> -cH:cH-9-.",
I
aucun cas, ir une ozonisation (ce demier genre
d' oxydation pr6sentant une stcchiom6trie tout
d fait differente de celle d'une oxydation photo-
chi mique ou thermique).
permettent de rendre compte de cette structure
tertiaire : Par ailleurs, il convient de signaler que la
pr6sente 6tude a 6t6 prolong6e pour permettre
CHs
I'analyse de m6langes form6s d'6lastomdres
I
-cH:cH-c-cH2- vulcanis6s et contenant de forts pourcentages
I
de noir de carbone et de pigments photo-
ooH actifs (ZnO et TiOz). Les r6sultats obtenus

r32
jusqu'ir maintenant montrent, d l'6vidence et convertir les dur6es d'essai en temps r6el
-
malgr6 les difficult6s d'ordre analytique rencon- d'usage dans des conditions moyennes
tr6es que le vieillissement climatique de ces d'utilisation.
-, expos6s d la lumidre s'explique
polymdres
essentiellement par une oxydation photo- Note
chimique, et non par une ozonisation.
*Un anti-U. V. est un compos6 mol6culaire
Conclusions qui absorbe fortement les rayons ulfraviolets
et dont les 6tats excit6s sont susceptibles de se
La pr6vision du comportement d long terme d6sactiver, de fagon non radiative, avec une trds
des mat6riaux polymdres synthdtiques en condi- grande efficacit6.
tions d'usage reste diffrcile. Mais I'avancement
des connaissances sur les m6canismes d'6volu- Abstract
tion rend cette pr6vision de moins en moins
al6atoire. En fait, les progrds r6alis6s r6cem- Prediction ofthe Long-Term Behaviour of
ment r6sultent de la conionction de deux Synthetic Po lymeric Mate rials from Artilicial
facteurs, d savoir : Ageing Experiments

o La mise au point de dispositifs exp6rimentaux The photo-ageing of synthetic polymers com-

d'6tude en laboratoire qui permettent d'exa- monly used in modem sculptures found in out-
miner les seuls ph6nomdnes de photovieillis- door museums will be described. CrossJinked
unsaturated po lyesters and various polyacrylates
sement en conditions anhydres (avec applica-
and polymethacrylates will be presented as exam-
tion de contraintes de lumidr:, de chaleur ples of unstabilized matrices whose properties
et d'Oz) ou en conditions de concentration can be modified during the polvmeisation, on
d'eau maintenue (avec application de dryingfrom solution or emulsion and by introduc-
contraintes de lumidre, de chaleur, d'Oz tion of additives. Photo-oxidation and thermal
et d'eau). Il devient alors possible de oxidation of dienic elastomers (Butadiene, 54,-
hi6rarchiser I' importance des diff6rents rene-butadiene and Nitrile-butadiene rubber)
m6canismes de photovieillissement, de and oxidation ofpolyacetals will be described.
thermovieillissement et de vieillissement A basic understanding ofthe oxidation mecha-
nism is pre-requisite to any stabilisation strateg/.
hydrolytique, et de n'acc6l6rer en condi-
tions artificielles que le m6canisme le
The prediction ofthe rate ofageingfor synthetic
plus important. polymers in use can be based on the "mechanis-
tic" approach. This approach entails the recogni-
o Le d6veloppement de m6thodes microana- tion ofsolid po$rners as chemical "reactors",
lytiques qui permettent de suiwe in situ l'6vo- that is, environments where certain chemical re-
lution chimique de microzones des systdmes actions are facilitated. The mechanistic approach
polym6riques. Ces techniques de microspec- involves the identification of mechanisms at the
trophotom6trie vibrationnelle la microspec- molecular scale, in artificial accelerated condi-
-
trophotom6trie IRTF, par exemple rendent
tions and in use. The most detimental effects on
-
possible I'analyse des couches 6l6mentaires
pofi,rners are caused by oxidation which is in-
duced by light, heat, static and dynamic stress.
de 5 pm et l'6laboration des profils des pro- Vaiations in macroscopic physical properties
duits d'oxydation et additifs qui se trouvent (mechanical propenies, opacity, surface aspect,
dans les parties les plus superficielles du etc.) can be explained by analyzing the chemistry
mat6riau. Elles peuvent de plus €tre utilis6es, fully. Thus, accelerated ageing is actually acceler-
grdce ir des d6tecteurs opto-acoustiques, pour ated oxidation and the accelerationfactor can be
analyser des milieux trds opaques. Enfin, ces determined from the rates at which critical oxida-
tion products form in artificial and natural condi-
reconnaissances analytiques des m6canismes
tions. Lifetimes of polymers can therefore be
d'6volution permettent, tout en garantissant la
predicted from the measured lifetimes in artifcial
repr6sentativit6 des essais de laboratoire, de

IJJ
c ondi t i otrs, t aking ac ce I erati on fact o rs int o ac - In ntost $,nthetic pobmers, the oxidative degrada-
cottnt. This mechanistic approach contpares fa' tion, u,hich is controlled either by'the nonnal
vourablv with entpiical techniEtes based on structure of the polltnerized monomer or b1'
simulation of actual physical and chemical condi' chemical dekcts, is obsented at a very lou' degree
tions and the resulting vanations in macroscopic of oxidation. Ven' sensitive and infornative in-
physical properties. situ spectrophotometric techniques are required
to detetmine the nature of the various intermedi-
ate and final products.

t34
Composition Implications of Plastic Artifacts:
A Survey of Additives and Their Effects on the Longevity of Plastics

R. Scott Williams

an C ons erv ation In s titute

Can adi
Communications Canada
Ottawa. Canada

Abstract Introduction
Plastic objects are usually described by A plastic is not a pure chemical product but
reference to their main polymer component, rather a formulation or composition made by
i.e., polyetfuilene sheet, vinyl or poly(vinyl chlo- proper mixture of a base polymer with a combi-
ride) upholstery, acrylic sculpture, polystyrene nationof additlves. Some cornmon base poly-
or styrene box, etc. Infact, all manufactured mers are listed in Table I. Polymers can be
plastic items are complexformulations of poly- produced as pure materials but these are never
mers with the additives that are required to adequate for serviceable products. To produce a
give the base pol"vmer suitable end-use proper- useful plastic, the inherent chemical, physical,
ties. Additives such as plasticizers, stabilizers, mechanical, optical, electrical, and other proper-
colorants, processing aids, etc. are as impor- ties of the polymer must be modified by the in-
tant as the base polymer in determining how corporation of additives, such as those listed in
long an object serves its intended purpose, or Table II. Plastics also are subjected to a variety
how long it survives in a museum. of physical treatrnents to modify such proper-
ties as appearance, printability, and shape.
The main polymer groups that comprise (Table III lists some of these.) The nature of
plastics found in typical museum objects are the base polyrner, additives, and fabrication all
brieflv introduced, and the additives in the plas- affect the service life and museum-longevity
tics madefrom these polymers are discussed in ofobjects. This paper discusses the mode of
detail. Results of recent chemical analyses of action of some additives and the deterioration
deteriorated plastic objects are used to illus- ofplastic objects due to their change or loss.
trate the effect of additives and base polymers
on the longevity of the plastics. Deterioration General Characteristics of Additives
and damage such as accretiow, blooming,
cracking, crazing, discoloration, embrittle- Additives should be efficient. that is. effective
ment, oozing, softening, etc. are related to at low concentration (most are used at concen-
changes in, or loss of, the additives. These trations of less than lo/o,with the notable excep-
changes are Wuenced by the museum environ- tions of plasticizers, pigments, fillers, and
ment. Generql guidelines for storage of plastic reinforcing fibres typically used at concentra-
objects, both to increase their longevity and to tions of l0o/o to 35olo), low-cost, convenient,
prevent damage to neighbouring objects, safe to use and handle, and should not impart
are given. undesirable characteristics such as colour. taste.

135
 Compatibility and Permanence
Table I Polymers and additives are compatible if they
Common Base Polvmers in Plastics can be intimately blended with each other to
form a homogeneous composition. Additives
(Note: The abbreviaions are those grven In ASTM that are highly compatible with a given resin do
Standard D 1600-83: Standard Abbrevratrons of
Terms Relatrng to Plastrcs by the American Society for not exude to form droplets or liquid surface
Testino and Materials.) films, nor do they bloom as a crystalline sur-
face crust. Incompatibility is usually indicated
by migration and exudation of substances on
Acrylonrtnle-butadrene-styrene ABS the plastic surface (also called spewing or
Cellulose acetate CA
Cellulose acetate butyrate CAB blooming), or by poor physical properties. This
Cellulose acelate propronate CAP is not always evident after short storage periods
Cellulose plastrcs, general CE but may take months or years to appear. Perma-
Cresol formaldehyde UT
Cellulose nitrate CN nence rcfers to the ability of an additive to re-
Cellulose propionale CP main unchanged within various environments
Casern UJ
so that the plastic retains its desired properties
Cellulose tnacetate CTA
Epoxy, epoxide EP during use, as opposed to simple incompatibil-
Ethyl cellulose EC ity,which becomes evident by time alone and
Ethylene-vinyl acetate EVA
lmpact polystyrene IPS
is not related to the exposure environment. The
MelamrneJormaldehyde MF permanence of an additive is usually deter-
Polyamrde (nylon) PA mined by its volatility, migration, extractabil-
PolycaIbonate rw
Polyethylene PE ity, and stability to heat and light.
Poly(ethylene terephthalate) PET
Phenol{ormaldehyde PF
Compatibility can be thought of as mutual
Polyisobutylene PIB
Poly(methyl methacrylate) PMMA solubility, although phenomena other than solu-
Polypropylene PP bility can create compatibility (e.g., molecular
Polystyrene t'b
Polytehaf luoroethylene PTFE
entanglement). Additives are often relatively
Polyurethane PUR small molecules that can diffuse through the
Poly(vrnyl acetate) PVAC polymer matrix. The larger the additive mole-
Poly(vrnyl alcohol) PVAL
Poly(vnyl butyral) PVB cule, or the more crosslinked or crystalline the
Poly(vnyl chlonde) PVC polymer, the slower is the diffusion. As a result
Poly(vnyhdene chlonde) PVDC
Styrene-acrylonitnle SAN
Styrene-butadrene SB
Srlcone plastrcs
Urea{ormaldehyde
sl
UF
Table II
Unsaturated polyester UP Additives in Plastics

Plasticrzers (50% rn some PVC)

Stabrlizers
or odour. They should also maintain their Antioxdants
properties over time, be resistant to extraction Heat stabllizers
UV absorbers
during service, be thermally stable at processing
temperatures, be stable to light, and be unaf- Processng Ards
fected by pollutants. Additives should be Intemal lubncants
Mould release agents
compatible with the polymer and other addi- Slip agents
tives, and be capable of blending mutually. Blowrng agents
They are intended to remain effective through-
End-use Modifiers
out the serttice lifetime of the plastic. Note that Colourants
the designed service lifetime of a plastic object - Organrc dyes and prgments
- Inorganlc prgments
is always much shorter than the desired mu-
Reinforcing frbres
seum lifetime. It would be nice if additives Fillers and extenders
remained effective throughout their museum Anltstattc agents
Antrblock agents
lifetimes, but this is seldom the case. Barrier coatngs
Lamrnatrng process

136
of diffusion driven by concenhation gradients, Hedley 1980, or Horie 1987 for discussions
additives can migrate to surfaces where, if they of solubility parameters). Table IV lists the
are not volatile, they collect as a discrete ex- Hildebrand solubility parameters of some
uded layer (this property is sometimes used common polymers and compounds used as
intentionally as in the case of slip agents and plasticizers. Compounds most commonly used
antistats). This migration continues until the as plasticizers for poly(vinyl chloride) (PVC)
concentration gradient disappears, at which have Hildpbrand solubility parameters between
time the rate of diftrsion from the interior to 17 .2 MPai'z and 23.3 MPa''2 bracketing
the surface equals the rate of reverse diffirsion 19.6MPa'2, the solubility parameter of PVC.
from the surface to the interior. Ifthe exudate The solubility parameters of other polymers
is removed, a new concentration gradient is set (e.g., cellulose nitrate and polyethylene) are
up and more additive migrates to the surface to outside the compatibility range of plasticizers
replace the lost material. Volatile additives that are suitable for PVC. Thus different types
evaporate at the surface, to be replaced by more ofplasticizers are required for different poly-
additive from the bulk of the plastic. In this situ- mers, as is verified by formulation practice.
ation there is a constant loss ofadditive to the
atmosphere and a constant diffusion of additive Compatibility and Dielectric Constant
from the interior to the surface. Eventually, all Solubility parameter is only one of many
the additive will be lost (and possibly redepo- properties that can be evaluated to predict plas-
sited elsewhere to cause harm). ticizerlpolymer compatibility. Dielectric con-
stant is also an important properly of a solvent.
Although plasticizers will be discussed in detail The dielectric constant of a compound is a
later, additive/polymer compatibility using plas- measure of its polarity and its polarizability.
ticizers as an example is discussed here to Polarizability is a measure of the response of
clarify discussions of all additives. a molecule to an electric field and thus of the
intermolecular forces between a solvent and
Compatibility and the a solute, that is, between a plasticizer and a
Hildebrand Sotubitity Parameter polymer. The interaction of these forces affects
Plasticization is similar to dissolution of the compatibility. Observation of compatible plasti-
polymer by the plasticizer compound. Thus the cizers shows that those with dielectric constants
Hildebrand solubility parameter, which is a between 4 and 8 are compatible with PVC,
measure of solubility or solvent power, may be which has a dielectric constant of 3.2
a predictor of compatibility (see Barton 1983, (Table IV).

Exudation of Additives due to

Table III Incompatibility
Pl as ti c F abric ation Pro c ess es Figure 1 shows the relationship between solu-
bility parameter and dielectric constant for
compounds commonly used as plasticizers (for
Shaping Processes all plastics, not just PVC). The zone of compati-
Molding, e}trusron, casting bility for PVC is indicated by the dotted lines.
Machining (cuttrng, gnndrng, dnlIng,
stamping, etc.) As can be seen there is a restricted group of
Foamrng compatible plasticizers for PVC. Similar zones
of compatibility could be plotted for all other
Surface Treatments
Surfa@ textunzrng, embosstng plastics and would include a different group
Surface activation (to permit printrng) of plasticizers.
- Flame
- Corona drscharge
Coating This Figure helps to explain the occurrence
- Painting, pnntrng of exudates and blooms. Some additives have
- Metallrzrng, electroplatrng
- Lamrnaton solubility parameters and dielectric constants
Joining outside the zone of compatibility. We should
Adhesive bonding expect these to exude because they are incom-
Welding
Mechanrcal fastening (screws, nvets, etc.) patible. Exudations may also occur if the

t37
 Table IV
Pl asti cizers : Solubilit"v Parameters
and Dielectric Constants

Solubility
Paralpeter Dielectric
Plasticizers (Mpa") Constant

Phthalate esters:
Di (2-ethylhexyl phthal ate), 14.9-18.0 5.2
(dioctyl phthalate), DOP
Dibutylphthalate, DBP 17.0-19.2 6.4
Butyl benzylphthalate 15.5-18.2 6.4

Phosphate esters:
Triphenyl phosphate, TPP 1e.6 (20.7)
Tricresyl phosphate, TCP 17.2-20.2 7.2

Diacid esters:
Di (2-ethyl hexyl) adipate,
(dioctyladipate), DOA ('t7.4) 4.1
Di(2-ethylhexyl) azelate,
(dioctylazelate), DOZ 19.0 4.0
Diethylene glycol dibenzoate 7.1

Miscellaneous:
Epoxidized soya oil, ESO (18.2) 5.5
Adipic acid polyester 19.0 6.0
Tri(2-ethylhexyl)
trimellitate, (trioctyl
trimellitate). TOTM 18.4 4.7
Camphor 18.2 11.4
Castor oil 8.9

Polymers

Polytetrafl uoroethylene 12.7 2.0

Polyethylene 16.2-16.8 2.25-2.35
Polypropylene 18.8 2.2
Polystyrene 17.6-19.8 2.4-2.6
Poly(vinyl acetate) 18.0-22.5 3.5
Poly(methyl methacrylate) 18.4-19.4 3.3-3.6
Poly(vinyl chloride) 19.0-20.2 3.3-3.5
Polycarbonate 19.0-20.2 3.0-3.2
Poly(ethylene terephthalate) 21.9 3.0-3.6
Cellulose nitrate 21.5-23.5 7.0-7.5
Cellulose diacetate 22.3-23.3
Cellulose acetate 27.2-27.8 3.5-7.5
Nylon 66 27.8 4.O

r38
 solubility parameters and dielectric constants of mutual solubilities and may lead to separation
polymers and additives shift out of the compati- of components (incompatibility). As aging pro-
bility zone as chemical reactions take place in ceeds, these changes make the plastic composi-
the plastic as it ages. For instance, it is possible tion more and more different from the initial
that dibutyl phthalate, a very compatible PVC design formulation. Since the initial formula-
plasticizer, may hydrolyze to form butyl alco- tion was designed by delicately balancing many
hol and phthalic acid. The solubility parameter properties of different polymers and additives,
and dielec,tric constant of Qutyl alcohol are the likelihood of incompatibilities increases.
23.3 MPa''2 and 36.4 MP a''2 respectively, val- The plastic should be designed so that these
ues well outside the zone of compatibility. We developing incompatibilities do not affect the
would expect this hydrolyic degradation prod- properties of the plastic during its design life-
uct (of an additive, not the polymer) to exude. time. Unforfunately, over the extended duration
This could happen for any additive that changes in a museum, problems of incompatibility start
upon aging. Similarly, if the polymer changes to show up.
enough to shift its solubility parameter and di-
electric constant, then the changed polymer Deterioration of Plastics
would have a new compatibility zone. Un-
changed additives might now lie outside the Plastics deteriorate by l) chemical degra-
new compatibility zone of the aged (changed) dation of the base polymer and/or the additives,
plastic. They might now be incompatible, 2) physical processes like bending and breaking
and therefore exude. or exudation of components (perhaps resulting
from chemical changes), and 3) biological
Incompatibility due to Aging of Base agents (molds, fimgi, rodents, people, etc.),
Polymer and Addifives just as all other organic and biological museum
There is a particular relationship befween the objects deteriorate (Table V). Physical proc-
base polymer and its additives in an unaged esses include diffusion and molecular redistri-
plastic. lnitially, everything is designed to be bution where there is no change in the chemical
mutually compatible. Upon aging, compatibili- bonds in the polymers or additives. These
ties change. Light-, heat-,
and oxygen-induced reac-
tions of the various com-
20 tldninal
ponents occur. Stabilizers
cornpa6lity
prevent reactions of the lmrts basod
on€
polymer macromolecules,
but in doing so the stabi-
lizers are changed into r- 14
@
new products, with differ- 6
ent compatibilities with !12
the base polymers. And" Eto
t Nminal
@mpa&a.ty
in spite of the stabilizers, o lmil8 baaocl
mA
changes do occur in the 88
polymer. Slow macro-
molecular rearangements
increase crystallinity. A
few crosslinks pull poly-
mer molecules together
and decrease molecular
o 5 r0 15 20 25 Xt 35 40 45 50 55 60
free volume, squeezing
U6locticdr6tan, t
out additives. Oxidation
of macromolecules and Figure I Compatibility zone defned by solubility parameter and dielectric
additives tend to increase constants ofplasticizers. Curyes: l, carboxtlic acids; 2, hltdrocarbons:
their polarities. All this 3, chlorinated hydrocarbons; 4, esters: 5, ethers; 6, alcohols; 7, all;yl nitiles;
leads to changes in 8, aldehydes;9, ketones; 10, nitro compounds. (from Darbv, Touchette and
Sears 1967.)

t39
 Degradation Profile
Table V To help visualize degradation processes,
Agents of Degradation degradation profiles can be created by plotting
changes in such properties as colour, brittle-
Heat during Processing and use
ness, tack, surface gloss or crazing or chalking,
Light during use
Atmosphere oxygen and Pollutlon release of volatile substances, and many other
Morsture hydrolysts chemical, mechanical, and physical properties.
Solvents dtssolutton and solvolysls
Feller (1977) described four stages ofdegrada-
Brologrcal Agents bactena, fungi, insects,
rodents, PeoPle tion inception, induction, increase to maxi-
Physrcal Changes crystallizalon, migratlon mum- rate or steady state, and decreasing rate
Mechanical Wo|* fatigue, creep
stages and introduced a generalized profile
- degradation as measured by oxy-
of oxidative
are usually slow and can be accommodated by gen uptake. Examples of degradation profiles
compatibility considerations during design of are shown in Figure 2 for yellowing, for weight
the plastic (See also McGlinchey, this publica- gain due to oxygen uptake during oxidation, for
tion). This sort of degradation takes a long time solubility, and for oxidation (measured by IR
to show. Often it is not detected, or not signifi- absorbance due to hydroperoxide formation).
cant, during the service lifetime of a plastic, but These degradation profiles show the state of
becomes a problem after long periods in a mu- degradation ofa plastic object at any stage in
seum. Chemical degradation alters the arrange- its lifetime and permit some prediction of the
ment and type of chemical bonds present in the course of its degradation. This is valuable in the
plastic. The primary chemical changes are due museum/conservation context. Chemical degra-
to oxidation, although hydrolysis is a factor in dation follows the same general pattem for all
some polymers. Oxidation can be initiated or plastics and organic materials although the
promoted by heat (thermal oxidation) and light
breadth and height of the curve may change.
(photo oxidation). Light in the absence of It is worth examining degradation profiles in
oxygen can also cause chemical degradation. more detail to see how plastic objects degrade,
to identify the stage ofdegradation achieved by
The inherent properties of plastics affect their a particular object, to predict what future can be
degradation. For example, amorphous polymers expected for a particular object, and to deter-
allow easier diffusion of oxygen and water than mine how the course ofdegradation ofobjects
do crystalline polymers, therefore oxidation can be altered so as to prolong their useful life.
and hydrolysis are likely to be more pro-
nounced in the former. Plastics with glass tran- Consider a plastic with a degradation profile
sition temperatures below room temperature are shown in curve A in Figure 3. If the duration of
in a rubbery state, which allows greater molecu- the inceptiorVinduction stage can be increased,
lar mobility and greater oppornrnity for reactive the amount of degradation that will have oc-
degradative species or degradation products to curred at a future observation date will be
move through the plastic thereby increasing the
decreased, even ifthe shape (slope) ofthe later
rate ofdegradation reactions. The presence or portions of the curve is unaltered. This situation
absence ofcertain functional groups in the poly-
is shown on curve B. Note that the amount of
mer macromolecules affects the susceptibility
degradation at the future observation date on
to certain degradation reactions. For instance, curve B (labelled Ds) is less than that on the
esters and amides are susceptible to hydrolysis,
original curve (Da). If the inception/induction
aromatic rings and carbonyl groups absorb ul- period cannot be increased, but the rate of deg-
traviolet (UV) light and are sensitive to pho- radation can be reduced (i.e., reduced slope),
todegradation, and pure hydrocarbons like then the total amount of degradation at the fu-
polyethylene, polypropylene, and rubber con- ture observation date can be reduced. This is
tain no chemically bound oxygen and so will shown in curve C. Note that Dc is less than De.
oxidize only if exposed to oxygen. Unforhr- As another example of possible intervention,
nately these latter hydrocarbons contain tertiary
consider that the durations ofthe various stages
hydrogens, which makes them very susceptible
cannot be changed but the magnitude of each
to degradation by free-radical processes. can be decreased. This is shown in curve D.

140
 O Sample I
aSampte 2

al Slmpl6 on Gls
oA Samptes in Alumlnrum Panr

Hours in Own * IOOoC

*,1'J.',,1:';:u^11"T.0r!i,ij0,tror: res,r
leo
loo ."f
3ao
dza ."I
d cxi rol I
4?O 'ool
801
{ro €oi
Eeo .ol
ito 201
100 ol

Figure 2 Degradation profiles produced by measuring diferent properties of plastics Top teft yellowing
o/'
ru-bber cement (from Feller and Encke 1982); Top right: ireghigain of rubier cement (from
Feller and'Encke
I 982 ): Botrom lefl:, Hydloperoxide concentration in photo-oxldiid poiyprop,tene
as measured bv IR absor-
bance at J400 cm ' with no inhibitor (A), with nickel chelate inhibitor (B) and with piperidine
H)LS inhibitor
(C) (from Grattan 1978); Bottom right: Change in solubility of varnish r'esins (from Felter
1975).

Again' the total amount of degradation at the grofile. They function either by preventing the
future observation date @o) is less than in the initiation reactions caused by light and h# ex-
case without intervention. posure, or by interfering witir th"e free-radical
chain reactions, thereby preventing propagation
In general, to decrease the degradation ofaplas- reactions. They also scavenge non-free-ra"dicat
tic at some future date the objectives should be products of degradation that-cause or catalyze
to 1) increase the duration ofeach ofthe early further degradition. Ifstabilizers are effeciive
stages ofthe deterioration_profile (i.e., delay they alterihe degradation profile by lengthen-
the onset of later stages), 2) decrease the rate of ingthe time scai6 and reducing the exteit of
change from each earlier stage to the next, and degradation. When stabiliz"r.iuil (by loss or
3) decrease the magnitude of degradation for change) the degradation profile becomes much
each stage. steepir and the plastic degrades more rapidly.

Stabilizers The stabilizers to be discussed here include

Stabilizers (inhibiton) are additives that are antioxidants, heat stabil izers, and ultraviolet
used to make changes in the degradation

t4l
 cleavage to produce two new oxygenated radi-
De4radation Trofile cals (hydroxyl and alkoxyl). Thus the number
of radicals multiplies and oxidative degradation
o of the polymer via free-radical reactions is
E accelerated.
\E
o In the initial stages of this chain reaction, radi-
o cals react with non-radical polymer macromole-
cules. Eventually, as the radicals multiply, their
concentration becomes large enough that they
react with each other as well as with the non-
'6
radical macromolecules. The reaction of one
free-radical with another produces a non-radi-
cal. Thus radical-radical reactions terminate
t'(r)
Jl
6'a) free-radical chain reactions. Eventually the
lncreaainq Aqe of }bJect, I
rate of termination of radicals equals the rate
ofproduction ofradicals and a steady state is
Figure 3 Altering the degradation profile of a plas- reached. This corresponds to the steady state
tic. Curtes: A, unaltered profile; B, induction period
region ofthe degradation profile degrada-
increased but shape ofprofile unchanged (B is "par- -
tion continues, but at a constant, non-accelerat-
allel"to A); C, induction period unchanged but rate
ofdegradation decreased (slope decreased); D, in- ing rate.
duction period and rate ofdegradanon unchanged
but magnitude ofdegradation decreased. D* Da, Prevention of Oxidation
Dg, and DL, are amounts of degradation at some ar- Key stages in the oxidation process are the
bitrary future observation time, t, -for each scenario. initial formation of alkyl free-radicals by ener-
The end of the induction period is indicated by getic processes and the subsequent production
ttme, tt. ofhydroperoxides by reaction ofthe alkyl
free-radicals with oxygen. Oxidation can be
light stabilizers (including ultraviolet light
prevented if formation of either of these species
absorbers).
is prevented or if these are converted to some-
thing innocuous before they have a chance to
Antioxidants react with polymer macromolecules. Thus to
prevent oxidation the aims should be the
Oxidation Reactions
following:
To understand how antioxidants work, the
chemical reactions that take place in the base
e prevent access ofoxygen to the plastic by
polymer during oxidation must first be consid-
removing oxygen from its surroundings
ered. This subject is vast and has been dis-
(anaerobic conditions) or by using oxygen
cussed in the conservation context by Grattan
(1978) and de la Rie (1988) and in much barrier layers on the plastic
greater technical detail by Carlsson and Wiles
(1986). For oxidation to occur, oxygen must be r prevent initial formation of alkyl radicals,
absorbed and dissolved in the polymer. It then or scavenge those radicals formed before they
diffuses to reactive sites on the polymer macro- have a chance to react with oxygen in the
molecule where oxidation reactions can take plastic (radical scavenger antioxidants)
place. Such reactive sites are usually free-radi-
cals produced when energy in the form ofheat, o destroy peroxides that do form before they
radiation (commonly UV and visible light but cleave to produce new radicals (peroxide
also gamma- and X-rays, or electron beams), decomposer antioxidants)
and mechanical work, etc. is absorbed by the
plastic and breaks chemical bonds in polymer r prevent the development of, or remove,
macromolecules. These free-radicals react with active sites from the polymer macromolecules
oxygen to produce hydroperoxides. Hydroper-
oxides readily undergo thermal and photo

142
Anaerobic (oxygen-free) atmospheres can be greater detail, with emphasis on conservation
produced using bell jars with inert gas purging applications. Free-radical scavengers are typi-
or by enveloping objects in oxygen barrier cally hindered phenols (e.9., butylated hydroxy
membranes (including bell jars) with enclosed toluene, BHT) or secondary aryl amines (e.g.,
oxygen scavenger chemicals such as Ageless diphenylamine derivatives). These react with
(Grattan l99l). This method may be suitable peroxy and alkoxy free-radicals to convert
for a few special objects in a museum, but it is them to non-radicals. Organophosphites and
useless for plastics in normal use. various organic sulfur compounds, such as
thiodipropionic acid, are effective peroxide
Oxygen-impermeable layers can be applied, decomposers. Organophosphites react with hy-
or allowed to develop by controlled exudation droperoxides to form alcohols and phosphates.
from the body of the plastic. Some wax "anti- Organic sulfides react with hydroperoxide to
oxidants" function in this manner during fab- form alcohols and organic sulfoxides. These
-
rication of the plastic a wax is incorporated that antioxidants are usually used in combinations
migrates to the plastic surface where it forms a with each other and with other stabilizers. A
thin oxygen-impermeable barrier layer that pre- typical stabilizer package may consist of pheno-
vents oxygen dissolution in the plastic. This is lic or amine antioxidant, a thiodipropionate
an example of an intentional exudation of an ad- ester. a phosphite, a metallic stearate (for lubri-
ditive. It is commonly used for rubber. Disrup- cation), and a metal deactivator (for chelating
tion ofthis barrier layer gives oxygen access to any traces of detrimental metal ions, usually
the plastic and oxidative degradation can occur. from polymerizatron catalyst residues).
Thus conservation treatments and handling pro-
cedures should be designed so that this surface
oxygen-barrier layer is not disrupted.
Table VI
Protection by Antioxidants Antioxidants
Antioxidants are compounds that inhibit or re-
Charn Propagatton Radtcal Termtnators
tard oxidative degradation of polymers by at-
mospheric oxygen during fabrication, storage, Alkylated phenols and polyphenols
and use. They are added during formulation - Butylated hydroxy toluene (BHT)
Secondary aryl amrnes
and become intimately mixed with the polymer.
They function at the molecular level by react- Hydroperoxrde Decomposers
ing with a particular functional group of a
Organophosphttes
nearby polymer macromolecule that is reacting, Thtoesters
or has just reacted, with oxygen dissolved in
Metal Deactrvators
the polymer. There are two classes of antioxi-
Hydrazdes
dants. Free-radical scavengers (radical or Tnazoles
chain terminators) react with chain-propagating - Benzotnazole

free-radicals before these free-radicals have a

chance to react with oxygen or polymer macro-
molecules, thereby preventing the formation of Polymers with hydrogens on tertiary carbon
more radicals and propagation of the chain reac- atoms, such as polypropylene and its copoly-
tions of oxidative degradation. Peroxide decom- mers, are particularly susceptible to oxidation
posers (sometimes referred to as secondary initiated by free-radicals produced by hydrogen
antioxidants or synergists) convert peroxides abstraction. During manufacture and process-
and hydroperoxides into non-radical and stable ing, unstabilized polypropylene yellows and
products before they become reactive free- changes in melt viscosity (a function of molecu-
radicals that propagate degradation reactions. lar weight). These polymers require antioxidant
stabilizers. Since they are exposed to high proc-
es sing temperatures, I ow volati lity c ompounds
Many compounds are used as antioxidants and
only the more common can be described (Table should be used. Typical antioxidants for
VI). A review by de la Rie (1988) describes polypropylene are high molecular weight,
polymer stabilizers and their reactions in low volatility alkylated phenols and polyphe-
nols with esters of thiodipropionic acid

143
(a sulfur-containing compound) and phosphites in polyrner chain breakage or chain-scission
at total concentration of 0.25o/o to l%o. and the formation of various oxygen-containing
species, such as peroxides, hydroperoxides,
Metal ions from catalyst residues, such as alcohols, aldehydes, and ketones. Physical
titanium and metal containing pigments (zinc, changes accompany these chemical changes,
iron, chromium, cadmium, etc.), induce hydro- leading to increased stiffness, insolubility, and
peroxide decomposition and cause undesirable exudation ofplasticizers (in the case ofplasti-
discolouration. Such metal deactivators as cized products). The degradation and stabiliza-
hydrazides and triazoles often are added in tion of PVC is described in great detail by Nass
conjunction with antioxidants to protect against /1976).
adverse effects of metals.
The main degradation reactions involve certain
Most hydrocarbon polymers (e.g., natural and labile chlorine atoms that are easily abstracted
synthetic rubbers based on polyisoprene, poly- from PVC molecules. The aims of stabilization
butadiene and its copolymers or blends like are to prevent the abstraction of chlorine atoms,
acrylonitrile-butadiene-styrene, polyisobuty- and/or prevent the reaction between chlorine
lene, etc., and such synthetic plastics as atoms and the polymer. This can be done by the
high-impact polystyrene that have a rubber following methods:
component) have unsaturated carbon-carbon
double bonds that are subject to oxidation that . exchanging chlorine atoms with stable
is accelerated by heat and UV light. This oxida- substituents that become more firmly bonded
tive aging is often characterized by yellowing to the PVC molecule
and embrittlement. These are commonly pro-
tected by 0.lo/o to 2.5%o of an antioxidant sys- r preferentially reacting the abstracted chlorine
tem generally consisting of organic phosphites with additives instead of PVC
and low-volatility hindered alkylidene bisphe-
nols often in conjunction with UV stabilizers. . scavenglng hydrogen chloride to prevent it
from catalyzing further reactions
Heat Stabilizers
o using antioxidants to decrease oxidation reac-
Heat stabilizers are added to polymers to retard tions in the PVC or the polyene sequences to
their decomposition when exposed to heat en- prevent chain-scission, crosslinking, and
ergy and mechanical stress during mixing, pro- other free-radical reactions
cessing, reworking of scrap, outdoor exposure,
and/or other storage or use conditions through- Table VII lists some of the more common types
out their service lives. Since they are most of heat stabilizers for PVC. Predominant in use
commonly used in PVC plastics, heat stabi-
lizers for this plastic will be discussed in detail.
Table VII
PVC degrades by a mechanism referred to as Heqt Stabilizers for PI/C
dehydrochlorination, in which hydrogen chlo-
ride is produced when hydrogen and chlorine Primary Stabrhzerc
atoms are eliminated from adjacent carbons on Banum/cadmium soaps
a PVC molecule. The acidic hydrogen chloride Calcium/zrnc soaps
catalyzes further degradation of the PVC and Lead soaps
Antrmony compounds
corrodes production equipment. The hydrogen Organotins
chloride elimination creates polyene sequences - Trn caboxylates
- Tin mercaptrdes and mercaptoesters
consisting of conjugated carbon-carbon double
bonds. This causes yellowing, which deepens Secondary (auxrlary) Stabr lrzers
through red and brown to black as degradation
Epoxy
increases. Oxidation reactions of non-reacted - Epoxrdrzed soya orl (ESO)
polymer molecules and of the polyene se- Organophosphrtes
quences of degraded polymer molecules result

144
among these is the group with heavy metal salts Thus, PVC stabilized as above, has a period
(zinc, cadmium, stannous tin, etc.) paired with with no deterioration while the barium (or
alkaline or alkaline earth metal salts (sodium, calcium) soap is present, followed by very
potassium, magnesium, calcium, barium, etc.). rapid deterioration after the barium (or calcium)
The synergism between the two types of salt soap is consumed leaving the cadmium (or
make the combinations much more cost-effec- zinc) chloride to catalyze degradation. This sort
tive than either single salt. of behaviour is exemplified by the deterioration
of vinyl rooftops on cars, which takes place
Primaty stabilizers are a group of compounds after about 5 to 10 years. On the degradation
that can be used as additives, by themselves, to profile this manifests itself as a lengthened in-
impart significant heat stabilization to a PVC duction period followed by a very rapid rate of
polymer. Secondary or auxiliary stabilizers deterioration in the second stage (e.g., Figure 3,
cannot by themselves impart stability, but can Curve B). It is the duration of the induction pe-
dramatically increase the efficiency of other riod that is of critical imoortance in the mu-
stabilizers by synergistic effects. seum context.

Primary Heat Stabilizers Among the earliest effective heat stabilizers

Barium/cadmium carboxylate soap systems for PVC were inorganic lead compounds (e.g.,
are the most widely used PVC heat stabilizers. basic lead sulfates, dibasic lead phosphite, ba-
Cadmium exchanges its carboxylate anion for sic lead carbonates, and complexes like basic
labile chlorine atoms of PVC creating a car- lead silicosulfate). Now there is a larger vari-
boxylate-containing PVC molecule that is ety ofboth inorganic and organic lead salts
stabilized against loss of other chlorine atoms (mainly soaps like basic lead phthalates, basic
in the form ofhydrogen chloride. This ex- lead maleates, and dibasic and normal lead
change forms cadmium chloride, which, stearates, which are also used for their lubricat-
unfortunately, catalyzes PVC degradation. Deg- ing properties). Lead functions in the same way
radation is prevented by reacting the cadmium as cadmium andzinc, stabilizing the PVC by
chloride with the barium soap also present in exchange of labile chlorine atoms with other
the heat stabilizer to make barium chloride and anions (e.g., stearates). Lead chloride is pro-
regenerate the cadmium carboxylate soap. Bar- duced, but, unlike cadmium andzinc chlorides,
ium chloride does not cause PVC to degrade. this does not catalyze PVC degradation. Eventu-
The regenerated cadmium soap is available for ally these stabilizers are consumed and become
more stabilization by labile chlorine exchange. ineffective. Lead stabilizers tended to be used
Eventually the barium soap is consumed by in opaque, rigid, and flexible PVC extrusion
reaction with the cadmium chloride. Then, no and injection moldings such as phonograph re-
barium soap remains to scavenge cadmium cords, conduit, pipe, and wire and cable. Lead
chloride, which then catalyzes PVC degrada- stabilizers are cornmon in wire and cable
tion resulting in catastrophic deterioration. insulation because of their electrical properties
Because these barium/cadmium stabilizers are (low volume resistivity) and because lead
toxic they are not likely to be found in objects chloride is water-insoluble and non-
that contact food and drug products. conducting.

Calcium/zinc soap stabilizers function in the Antimony stabilizers have been used for
same way as barium/cadmium soaps, but they about 25 years, especially in phonograph rec-
are non-toxic and have been sanctioned for ords. These have drawbacks because they are
food-packaging film, blister packaging, blow- unstable in UV light and, like lead stabilizers,
molded bottles, beverage tubing, blood bags, also suffer from sulfide staining (producing an
etc. Like the barium/cadmium system, the cal- orange stain) when in contact with any sulfides.
cium/zrnc system is effective as long as there is They are cost-effective because they have a syn-
calcium soap available to scavenge zinc chlo- ergistic response with metal soaps and improve
rides and regenerate zinc soaps. Eventually performance at lower concentrations.
the calcium soap is exhausted and degradation
proceeds.

t45
Recently commercialized organotin stabilizers Advantages and Disadvantages
are most efficient for rigid PVC, because, al- of Different Stabilizers lvlussurn
though more expensive, they excel in the most Consequences -
demanding high temperafure and high pressure Lead stabilizers are relatively inexpensive but
(high shear) fabricating processes that must be they have several drawbacks. They are opaque
used for non-plasticized PVC. These are based salts and cannot be used in clear plastics. Lead
on tetravalent tin (they are usually dialkyl tin compounds are toxic and do not have FDA ap-
esters) and are true organometallic compounds proval. They are also susceptible to sulfur stain-
where the metal is covalently bonded to carbon. ing, which may be of some consequence in
Organotin carboxylates like dibutyltin dilaurate museums. Early "plastic" objects are often
and dibutyltin maleate were among the earliest made of sulfur-vulcanized rubber. Storage of
organotin stabilizers. These have low toxicity, lead-stabilized PVC with sulfur-vulcanized rub-
low volatility, and good UV stability. Organo- ber may lead to sulfur staining of the PVC, that
tin maleate provides UV stability and sparkling is, the production ofbrown or black lead sul-
clarity to flexible and rigid PVC. Some alkyl fide by reaction of lead from the stabilizer with
tins have low toxicity and are sanctioned by the volatile sulfur-lulcanizing agents from the rub-
U.S. Food and Drug Administration (FDA) ber. Gaseous sulfur pollution may also cause
(e.g., di-n-octyltinbis-(iso-octylthioglycolate) lead staining. Additional sources of sulfur are
and di-n-octyltin maleate). Organotin mercap- other (more modem) PVC objects that contain
tides and mercaptoacid esters are liquids that organotin sulfide based heat stabilizers. Cad-
are used in the most difficult-to-process rigid mium and antimony stabilizers undergo similar
PVC applications, such as pipe and profile ex- reactions with sulfur to produce yellow and or-
trusion and injection and blow molding. These ange stains. Thus heavy metal stabilizers and
have poor UV stability relative to organotin sulfur-containing stabilizers cannot be used in
maleates and strong characteristic skunky the same product because of the formation of
odours. lead sulfide stains. Similarly PVC with heary
metal stabilizers cannot be stored with
Secondary Heat Stabilizers sulfur-containing plastics and rubbers, direct
Epoxy compounds with oxirane and ester contact especially should be avoided.
functional groups improve the performance of
heavy metal carboxylate soaps. These are pre- Ultraviolet Stabilizers
pared by epoxidizing natural unsaturated vege-
table oils like soya bean and linseed oil, or by When exposed to light, especially UV, some
epoxidation of synthetic esters of unsaturated plastics are degraded. To have an effect, the
animal or vegetable fatty acids (e.g., butyl-9,10- light must be absorbed by the plastic and must
epoxy-stearate and octyl-epoxy-tallate). Their have sufficient energy to rupture chemical
ability to accept chloride ions preferentially bonds. In general, UV light has sufftcient en-
(thereby preventing the formation of heary ergy and, for cerlain bonds, so does blue light.
metal chlorides that catalyze degradation) is the Certain functional groups, called chromo-
source of their enhanced performance. Thus the phores, absorb light and transform into
more the oxirane content, the more effective is energetically excited and highly reactive states.
the epoxy. Epoxidized soya and linseed oils are Because the functional groups have absorbed
also plasticizers that add low temperature flexi- UV light energy they contain more energy than
bility. These are also FDA sanctioned. in their unexcited normal or "ground" states.
This excess energy must be dissipated and the
Organophosphites (di- and trialkyl, di- and extent to which the plastic will, or will not, be
triaryl, and mixed alkyl aryl phosphites) are degraded depends on how this energy
most often used in conjunction with barium/ dissipation occurs.
cadmium and calcium/zinc stabilizers or some-
times with organotin or antimony stabilizers. Absorbed energy in the excited chromophore
They provide heat and UV light stability and can be dissipated by 1) non-damaging conver-
improve early colour and long-term UV light sion to heat (i.e., vibration of molecules),
stability. 2) non-damaging energy transfer to nearby

r46
molecules which in turn dissipate the energy o be compatible with the polymer
as non-damaging heat, or 3) degradative rup-
ture of weak chemical bonds in the chromo- o be stable at processing temperatures
phore molecule or in nearby molecules to
which the energy is transferred. . have low volatility so that it is not lost during
processlng
Different plastics vary greatly in their resis-
tance to light damage (Carlsson and Wiles o not contribute colour
1986). Some plastics, such as polycarbonates,
polyesters, and aromatic polyurethanes, have Choice also is influenced by other additives pre-
strongly absorbing chromophores as parts of sent, such as antioxidants, heat stabilizers, fill-
their molecular structures. Other plastics, such ers, and pigments.
as polyethylene and polypropylene, have no
chromophores in the pure polymer, but unfortu- Three types of uV stabilizers are used UV
nately they usually contain small amounts of absorbers (UV screens), energl transfer - agents
catalyst residues from polymerization, or ke- (quenchers), and fr e e -r' a d i c a I s c ave n gers. Each
tones and hydroperoxides from oxidation, of these goups contains many compounds that
which are chromophores and absorb energy, are effective UV stabilizers, some of which are
leading indirectly to bond cleavage and forma- listed in Table VIII. Synergistic enhancement
tion offfee-radicals in the plastic. In air, of stabilizing activity is often achieved by
free-radicals higger chain reactions leading to simultaneous use of stabilizers that act by dif-
more bond cleavage and destruction. Once initi- ferent mechanisms. Some compounds act by
ated by light, these oxidative chain reactions several different mechanisms. In his review of
require no light to continue, so it is important polymer stabilizers de la Rie (1988) describes
to prevent or inhibit the initial photochemical UV stabilizers and their reaction mechanisms
events that produce these radicals. Since photo- in greater detail.
chemical production of radicals leads to sub-
sequent oxidation reactions, antioxidants
(preferably UV-stable ones) are usually added
to retard thermal oxidative degradation by
Table VIII
Ultravi olet St abi lizers
scavenging free-radicals and chemically
inactivating peroxides. UV Absorbers

Carbon black
Thus there are two main mechanisms for
Zrnc oxrde
UV degradation: I ) direct rupture of chemical Substrtuted benzophenones
bonds by absorbed UV energy followed by (yellow drscolouratron)
Substrtuted benzotnazoles
rapid oxidation of the radical fragments, and Substrtuted acrylonrtnles
2) rupture in nearby molecules caused by en- Aryl esters
ergy transferred from UV-excited impurities, - Salicylates and benzoates

such as catalyst residues or oxidation products Energy Transter Agents

in the plastic. (Quenchers)

Cobalt and nrckel chelates

Choice of UV absorber stabilizers is governed (green/blue drscolouratron)
by the spectral region in which light absorption
Free-radical Scavengers
leads to degradation. This spectral region,
namely the "activation spectrum," is that por- Hrndered amrnes (HALS)
tion of the UV (and/or visible) spectrum, pri-
marily responsible for the degradation of a
particular polymer. A UV stabilizer should thus IIV Absorbers
do the following:
UV absorbers compete with the polymer for
absorption of the incident UV light. The UV ab-
. absorb strongly at the wavelengths sorber absorbs the UV light before or more effi-
of maxi-
mum sensitivity of the polymer ciently than the polymer molecule. The excess
energy in the UV absorber is converted into

t47
harmless heat (vibrational energy of the mole- Nickel chelate complexes are commonly used
cule). By absorbing the UV light at the surface, quenchers that function by absorbing LfV light
less UV light reaches the interior of the plastic. energy into their highly conjugated structures
It is important to note that this is essentially a and dissipating the energy as harmless IR radia-
surface phenomenon. Although the entire bulk tion by a resonance stabilization mechanism.
of the plastic may contain UV absorber, only They also have some hydroperoxide decompos-
that absorber in a surface layer measuring a few ing abilities. Unfornrnately they are inherently
tens of microns thick is the effective stabilizer. greenish or tan coloured. Some benzophenone
derivatives function both as direct IJV absorb-
Some inorganic pigments are efficient UV ab- ers and as energy-transfer agents, absorbing in
sorbers. Carbon black is most common and ef- the range of 230 nm to 350 nm. As noted, these
fective, but titanium dioxide andzinc oxide are have a tendency to yellow under processing or
also used. When high levels of inorganic pig- exposure to light.
ments can be used, additional UV stabilizer is
not required. The main disadvantage with these Free-radical Scavengers
pigments is that they produce opaque plastics. Free-radical scavengers do not absorb or block
UV light. They scavenge the free-radicals pro-
Colourless and transparent UV absorbers, duced by photolytic reactions. They prevent
such as benzophenone derivatives (e.g., ortho- degradation by reacting so quickly that the free-
hydroxybenzophenone) absorbing at 230 nm to radicals have no time to undergo subsequent
350 nm, andbenzotriazole derivatives absorb- photo-oxidative degradation reactions with the
ing at 280 nm to 390 nm, can be used in clear polymer.
plastics. Although these UV absorbers are not
consumed during their interaction with the UV Compounds typically used for this include
light, some are consumed by reaction with free- substituted tetramethyl piperidines, a class of
radicals (this is especially true for o-hydroxy- compounds referred to as hindered amine light
benzophenone, which is as commonly added stabilizers (HALS). These HALS have excep-
for its free-radical scavenger antioxidant tional activity that is attributed to mechanisms
properties as for its UV absorber properties). in which the active species are regenerated and
An unfortunate property ofbenzophenones and recycled. They are not consumed by their stabi-
benzotriazoles is that the products oftheir reac- lizing reactions and therefore have long life-
tions with free-radicals are usually highly col- times in the plastic. HALS function as energy
oured, causing yellowing of the plastic. Aryl quenchers, peroxide decomposers, and/or alkyl
esters including aryl salicylate, resorcinol radical terminators.
monobenzoate, and aryl esters of terephthalic
and isophthalic acid undergo light-induced Plasticizers
reaffangement to give photochemically stable
products, which are derivatives of2-hydroxy- A polymer is said to be plasticized when it
benzophenone which absorb in the range of is made more flexible by the addition of com-
230 nm to 350 nm and which are effective pounds to the base polymer. Typical plasti-
IJV absorbers. Unfortunately these also cause cizers are high-boiling organic liquids or
discolouration by formation of quinoid by- low-melting solids that are added to hard or
products, like the benzophenone derivatives. tough resins to impart softness or flexibility, re-
duce stiffness, increase impact resistance, and
Quenchers (Energy Transfer Agents) ease processing by lowering melt viscosity and
Quenchers interact with energetic excited-state fabricating temperatures. Camphor, the first
chromophores. Excess energy is transferred plasticizer, inhoduced about I 870, was used at
from the excited chromophore to the quenching concentrations of about 50o/o in cellulose nitrate
stabilizer, allowing the excited chromophore to to make celluloid. This formulation was the
return to its stable ground state. The excess en- principal thermoplastic until camphor supplies
ergy in the excited quencher is dissipated by from Japan were threatened during WW II.
non-damaging conversion to heat (vibrational Increasing scarcity combined with camphor's
energy) in the quencher molecule. objectionable odour led to its gradual

148
replacement with phosphate e sters, parti cularly softening action (plasticization) usually is attrib-
triphenyl phosphate. Phthalates were developed uted to their ability to reduce the intermolecular
in the 1920s. Di-2-ethylhexyl phthalate, com- attractive forces between the polymer macro-
monly referred to as dioctyl phthalate or DOP, molecule chains. The plasticizer molecules
currently the predominant plasticizer for all insinuate themselves between those of the
plastics, was patented in 1933. Modem plasti- polymer, thereby preventing the polymer-to-
cizers are used at concentrations of 10% to polymer interactions that create rigidity (i.e.,
25o/obut sometimes as high as 50o/o (Figure 4). increased tensile strength and decreased elonga-
Eighty-five per cent of plasticizers are used in tion). The plasticizer acts as a lubricant to facili-
PVC and these are most commonly esters of tate movement of the polymer macromolecules
carboxylic acids or phosphoric acid. Plasticiza- over each other and provides internal lubricity.
tion of PVC is reviewed in detail by Darby and Plasticizers have varying degrees of solvating
Sears (1976). action on resins. A plasticizer is a solvent that
is involatile.
Plasticizers are not chemically bound to the
polymer but merely dissolved in it. Plasticizer Different plasticizers have different efhcien-
cies different concentrations are required to
- the same level of plasticization. In gen-
achieve
eral, monomeric plasticizers are the most effi-
cient. Plasticizer efficiency may be measured
by a number ofphysical properties, such as ten-
E
-t''.\'.\ sile strength, modulus of elasticity, elongation
"'.i".
t'F at break, and hardness. Figure 4 shows the vari-
o
z ation of the mechanical properties of tensile
strength and elongation at break with plasti-
cizer concentration and plasticizer type. Hori-
J
6 zontal lines indicate the onset of plasticization.
zU
At higher concentrations plasticization is evi-
dent, whereas at lower concentrations the plas-
tic remains stiff and relatively unaffected by the
*

.-t;;ffi
2? -'
i plasticizer content. Figure 4 also shows that
; 200 different plastic izers have different effi ciencies,
0
z that is, at a particular concentration different
ro0
ot
,,'/4.4i2( compounds produce different levels of
flexibility.
lo 20 30
o/o
PLASTTCTZER CONC.
Plasticizers that exhibit good compatibility
are designated primary plasticizers. Those that
Figure 4 Efect of concentration ofvarious plasticiz- exhibit partial compatibility or that exude or
ers on tensile strength and elongation at break of bloom on standing are usually called secondary
PVC. Plasticizers: I, dicycloheryl phthalate; 2, plasticizers and cannot be used alone. The
phthalate polyester; 3, adipic polltester; 4, ticresyl
distinction between primary and secondary
phosphate ; 5, I iquid butadiene-acrylonitrile copoly-
plasticizers is vague, depending partly on poly-
mer; 6, di-2-ethylhexyl phthalate; 7, di-2-ethylhexyl
adipate (from Darby and Sears 1976). The concen- mer, concentration desired, and environment
tration at v'hich the curves cross the line marking the and conditions ofend use (see previous discus-
initial value of tensile strength or elongation indi- sion of plasticizers in section on additive com-
cates the onset ofplasticization. At any plasticizer patibility). These are external plasticizers,
concentration greater than this, the plastic will be compounds added to the base polymer. Internal
sofier and more flexible than in the non-plasticized plasticization can be accomplished by copoly-
state. For these plasticizers the concentration at on- merization of a flexible polymer with a rigid
set ofplasticization ranges from about 694 to 22% polymer. Vinyl chloride-vinyl acetate copoly-
plasticizer content. This is also a measure of the
mer is an example where the vinyl acetate im-
effciency of the plasticizer-the lower the onset
parts flexibility to the PVC. In this type of
concentrafion, the more efticient the plasticizer.

149
plasticization, no "external" compounds are of acids and alcohols. The acids may be cyclic
added to the base polymer. Instead a new poly- (e.g., phthalic, benzoic, trimellitic, terephthalic)
mer with greater flexibility is chemically pro- or linear (e.g., adipic, azelaic, sebacic, ethylhex-
duced by modifying the polymer backbone anoic, isodecanoic). The alcohols may be
with flexible units. Such "intemally plasticized" monohydric (e.9., 2-ethylhexanol, isodecanol,
plastics avoid plasticizer migration problems. butanol) or polyhydric (e.g., propylene glycol,
Internally plasticized plastics will not be dis- ethylene glycol, pentaerythritol). Furthermore
cussed fi.uther. the alcohol may be linear or branched. Aryl (cy-
clic) and alkyl (linear) esters ofphosphoric acid
Pressure exerted on plastics can cause are also cofiImon.
migration of plasticizers. This can lead to exu-
dation and blooming of plasticizers or to loss of Secondary plasticizers for PVC include various
plasticizer from specific areas in compression aromatic and mixed aromatic-aliphatic oils,
(Figure 5). This might cause excessive exuda- chlorinated paraffrns, some po ly(alpha -methyl
tion at the weight-bearing points of objects rest- sfyrene) derivatives, and inexpensive esters of
ing in storage. Also where there are snap high molecular weight alcohols and organic
fittings under compression, plasticizer canbe acids that have marginal compatibility with
lost. This leads to embrittlement. Examples of PVC.
this are the location where a doll's arm snaps
over a protruding socket at the shoulder, or The most common general purpose plasticizers
where a hose fitting over a tube is often held by are cyclic dialkyl phthalates. As the alkyl side
a hose clamp that adds more pressure and in- chain of the plasticizer increases in molecular
creases migration and subsequent embrittle- weight volatility decreases, water extractability
ment due to plasticizer loss. Plasticizer loss is decreases, oil extractability increases, effi -
usually accompanied by shrinkage resulting ciency decreases, processing becomes more
from loss of volume (i.e., that normally occu- difficult, and specific gravity decreases. Trimel-
pied by the plasticizer in the plastic). This leads litate esters (i.e., derived from trimellitic acid)
to cracking, again exacerbated in areas often- with branched or linear alkyl side chains having
sion or compression. 7 to 10 carbons are cyclic plasticizers ofhigher
molecular weight, which have low volatility
Composition of PVC Plasticizers and low water extractability. Major applica-
All primary plasticizers for PVC are esters, tions are in high temperature wire and non-fog-
functional groups that are reaction products ging automotive compounds for upholstery and
panels in car interiors. (Note: Fogging plasticiz-
ers are those that evaporate from plastic compo-
nents inside automobiles then condense as the
bluish oily film on the inside of the car
a<
windows.)
U;
6
o
) Linear plasticizers have improved low tempera-
N
ture performance. Their linear structure (no
9 rings) make them less compatible with PVC
F
and they tend to exude. Substitution ofalkyl
groups by aryl groups (rings) improves per-
formance. Alkyl benzyl phthalates are high
solvating plasticizers with increased
PRESSURE, kg,/cm2 compatibility with PVC. For example, butyl
benzyl phthalate is used in vinyl asbestos floor
Figure 5 Loss ofplasticizer due to pressqre on
PVC. Note that the pressure of 0.1 kg/cm' (which n tiles because it has less tendency to migrate
equivalent to that exerted bv a stack of 100 dimes or from the tile into the flooring adhesive, thus
quarters) can be easily achieved on objects in stor- preventing adhesive softening caused by the
age, d the entire weight of the objects rests on small plasticizer. Citrate esters are approved for food
area bearing points. (from Darb)t and Sears I 976) products.

150
Phosphate plasticizers impart flame resistance. phthalate, for plasticization. Triacetin may also
Triaryl (tricresyl, cresyl diphenyl, triphenyl) be used. Such materials have greater volatility
phosphates dissolve easily and are low in vola- than dioctyl phthalate (which is typically used
tility. As aryl groups are substituted by alkyl in PVC) and are more subject to loss by evapo-
groups, the properties ofthe plasticized plastic ration leading to embrittlement and cracking
tend to become less useful. For example, tri-2- due to shrinkage.
ethylhexyl phosphate is only good at low tem-
peratures, has high oil extractability, high Plasticizers with high solvating power for
volatility, and poor heat stability (needs PVC are suitable for poly(vinyl acetate),
stabilizer). PVAC. Dibutyl and butyl benzyl phthalates,
and glycol dibenzoates are widely used in
Polymeric plasticizers (polyesters) are usually PVAC. Phthalates and dibenzoates have been
based on linear dibasic acids plus polyols (espe- detected in several PVAC adhesives (white
cially glycols). Their peffnanence improves glues) analyzed in the Adhesive Testing Project
with increased molecular weight. They gener- at CCI (Down and Williams 1988).
ally have low volatility, good to very good ex-
traction resistance to aqueous and hydrocarbon Toluene ethyl sulfonamides are used to soften
solvents, but have poor low temperature per- some grades of nylon. Poly(vinyl butyrals) are
formance and are inefficient and difficult to most often plasticized with glycol derivatives,
process. Epoxy plasticizers are generally epoxi some adipates, and some phosphates. Alkyl
dized oils (linseed and soya bean) and alkyl ep- benzyl phthalates and phosphates are used in
oxy esters (tallates, oleates, linseed oil acids) acrylics. Polprethanes can be plasticized with
with performance properties like polyesters. many of the same compounds that are used
The presence of the oxirane goup allows these for PVC.
to also serve as auxiliary heat and light stabi-
lizers. Some soya and linseed oil epoxies are Extraction of Plasticizers
also FDA approved. Extraction of plasticizers from PVC by coat-
ings, such as nitrocellulose, can mar finishes
Aliphatic and aromatic hydrocarbons and and cause discolouration and tackiness. This
chlorinated paraffins are generally not compat- problem can be greatly reduced by the use of
ible with PVC. These are used as minor compo- plasticizers with little or no affinity for the
nents of a plasticizer system as extenders to coating and by the selection of plasticizers with
reduce cost in those few applications where resistance to migration. Isophthalates (DIOP)
they provide advantageous end-use properties. and terephthalates (DOTP) have very low
migration to nitrocellulose coatings and offer
Plasficizers for Plastics Other Than PVC a marked improvement over all orthophthalate
The amorphous character of PVC allows esters, but are very expensive.
pasticization over a wide range of plasticizer
concentrations, whereas the higher crystallinity As a consequence of these mutual compatibili-
of other polymers means that the region in ties, it is essential that objects made of different
which plasticizing is possible is smaller. This polymers (such as these) be kept out ofcontact
means that plasticizer compatibility must be with each other. Plastic2er migration and plasti-
critically evaluated. Plasticizers must be chosen cizer transfer will occur, usually with softening
with care. and increased tackiness ofobject surfaces. This
also applies to the choice of storage materials.
Generally the plasticizers used for poly(vinyl Storage materials must be carefully chosen to
chloride) work well for cellulose nitrate, resist damage by plastic objects and their
ce llulose acetate propionate, cellulose acetate additives.
butyrate, and ethyl cellulose. Cellulosics are
best plasticized with dibutyl phthalate, alkyl- Conclusion
oxy phthalates, triaryl phosphates, and sul-
fonamides. Cellulose acetate requires lower When collecting, displaying, and storing plastic
alkyl phthalates, such as dimethyl and diethyl objects it must be remembered that plastics are

l5l
complex mixtures of polymers and additives. Nous decrirons donc brievernent, dans le cadre
Changes in either the polymer or any of its addi- de la prdsente communication, les principaux
tives will result in changes in the plastic, usu- groupes de polymires auxquels appartiennent
ally in a way that is described as degradation. les plastiques qui entrent dans la composition
This paper describes the wide variety of addi- des objets de musie les plus typiques, pour en-
tives that are used and the delicate balance of suite traiter dans le detail des additifs qui sont
properties that are required to ensure that all prdsents dans les plastiquesfabriques d partir
the ingredients are, and remain, compatible in a de ces polymdres. Les rtsultats d'analyses
well formulated plastic. The effects of antioxi- chimiques ricentes, portant sur des objets en
dants, heat and light stabilizers, and plasticizers plastique dtgradis, seront utilisbs pour mon-
on the longevity ofplastics has been discussed trer I'incidence des additifs et des polymdres
in some detail. Other additives not discussed de base sur la longivite des plastiques. Leur
here, such as colourants (pigments and dyes), digradation, de m€me que d'autres dommages
mineral additives like fillers, extenders and rein- qui leur sont causis accrition, cristallisa-
forcing agents, and processing aids like anti- -
tion, craquellement, fssuration, ddcoloration,
static agents, intemal lubricants, mold release fragilisation, suintement, ramolliss ement,
agents, and slip agents, also affect the longevity etc. est lide d une modification ou d une
of a plastic, although probably not to the same perte-,d'additifs. Et les conditions ambiantes ne
degree as those discussed. Degradation profiles sont en outre pas ttrangdres d de tels change-
were introduced as an aid to understanding the ments. Nous fournirons enfin des lignes direc-
states and stages ofdegradation ofplastics as trices gmtrales sur la mise en rdserve des
determined by the rates and extents ofdegrada- objets en plastique, qui permettront, d'une
tion processes occurring in aging plastics. The part, d'accroitre leur longtvite et, d'autre part,
ways that additives and changes in additives d'empdcher qu'ils ne causent des dommages
can influence the shape of the degradation pro- aux m ati eres envi ro n na n tes
file and thus the longevity of the plastic object
were discussed. References

R6sum6 Barton, A.F.M. CRC Handbook of Solubility

Parameters and Other Cohesion Parameters
La composition des objas en plastique (Boca Raton, Florida: CRC Press, Inc., 1983).
et leur longevitd
Carlsson, D.J. and D.M. Wiles, "Degradation,"
Les objets en plastique sont gendralement in: Encyclopedia of Polymer Science and Engi-
ddcrits d'apris leur principal constituant po- neering, Volume 4 ed. H.F. Mark et al. (New
lymdrique unefeuille de polyithyldne, un York John Wiley & Sons, 1986) pp. 630-696.
-
rev€tement en vinyle ou en polychlorure de
vinyle, une sculpture en acrylique ou une bofte Darby, J.R. and J.K. Sears, "Theory of Solva-
en polystyrdte ou en styrdne, par exemple. Or, tion and Plasticization," in: Encyclopedia of
tous les articlesfabriquis en plastique sont PVC, Volume l, Chapter I0 ed.LJ. Nass (New
constituts de prtparations complexes de York Marcel Dekker, lnc.,1976) pp. 385-503.
polym,ires, auxquels viennent s'ajouter les ad-
ditifs necessaires pour conferer au polymdre de Darby, J.R., N.W. Touchette and K. Sears,
base les propridtds qu'exige son utilisation fi- "Dielectric Constants of Plasticizers as Predic-
nale. Et les additifs notqmment les plasti- tors of Compatibility with Polyvinyl Chloride,"
fiants, les stabilisants, les colorants, les agents Polymer Engineering and Science, October
favorisant le traitement, etc. sont tout aussi 1967 , pp. 295-309.
importants que le polyrndre de - base quand il
s'agit de determiner la durte de vie utile de de la Rie, R., "Polymer Stabilizers. A Survey
tels objets ou la ptriode durant laquelle ils with Reference to Possible Applications in the
pourront €tre conservds dans un muste. Conservation Field," Studies in Conservation,
vol. 33, no. 1, 1988, pp.9-22.

t52
Down, J.L. and R.S. Williams, "Report on Ad- Grattan, D.W., "The Oxidative Degradation of
hesive Testing at the Canadian Conservation In- Organic Materials and Its Importance in the
stitute," Symposium '88 Proceedings (Ottawa: Deterioration of Artifacts," Journal of the Inter'
Canadian Conservation Institute, in press). national Institute for Conservation - Canadian
Group,vol. 4, no. 1, 1978, pp. 17-26.
Feller, R.L., "Studies on the Photochemical Sta-
bility of Thermoplastic Resins," Preprints of Grattan, D.W., "Degradation Rates for Some
the 4th Triennial Meeting of the International Historic Polymers and the Potential of Various
Council of Museums Committeefor Conserva' Conservation Measures for Minimizing Oxida-
tion Committee. Venice. 7 5 12214, 197 5, tive Degradation," this volume.
pp. l-11.
Hedley, G., "solubility Parameters and Varnish
Felleg R.L., "Stages in the Deterioration of Or- Removal: A Survey," The Conservator,vol.4,
ganic Materials," in Preservation of Paper and 1980, pp. l2-18.
Textiles of Historic and Artistic Value. Ad-
vences in Chemistry Series, No. 164 ed. J.C. Horie, C.V., Materials for Conservation :
Williams (Washington, D.C.: American Chemi- Organic Consolidants, Adhesives and Coatings
cal Society, 1977) pp. 314-335. (London: Butterworths, 1987).

Feller, R.L. and D.B. Encke, "Stages in Deterio- Nass, L.I., "Theory of Degradation and Stabili-
ration: The Examples of Rubber Cement and zation Mechanisms," in: Encyclopedia of PVC.
Transparent Mending Tape," tn'. Science and Volume I, Chapter 8 ed. L.I. Nass (New York:
Technologlt in the Sertice of Conservation. Pre- Marcel Dekker, Inc., 1976) pp.27l-293.
prints of the Contributions to the Washington
Congress, 3-9 September 1982 ed. N.S. Brom- Nass, L.I., "Actions and Characteristics of Sta-
melle and G. Thomson (London: International bilizers," in: Encyclopedia of PVC. Volume I ,
Institute for Conservation of Historic and Artis- Chapter 9 ed.Ll. Nass (New York: Marcel
tic Works, 1982) pp. 19-23. Dekker, Inc., 1976) pp.295-384.

153
The Nature and Origin of Surface Veneer
Checking in Plywood

Mark D. Minor

Department of Obj ects

C onserv ation
TheMetropolitain Museum of Art
New York, N.Y.
U.S.A.

Abstract By the third week, checks extendingfrom the

original lathe checl<s had extended to approxi-
Pllnuood is a material occarring with greater matelv 7596 to 8096 of the veneer thickness. By
frequency in both decorative andfine arts collec- the ninth week, several cracks had reached the
t i o n s. LThi I e fund am ent a I ly w o o d, i t s s t ru ct ure sudace of the panels. At week 20, a majority of
and characteristics fall into a completelv different the cracks had extended to the surface, regardless
realm than solid timber. One degradation prob- of the severity of the moisture change.
lem evident in many pieces offurniture and paint-
ing supports are longitudinal surface veneer A modelforformation of checks is suggested,
checks. These cracks are disfiguing to the dis- based on compressive damage to the wood cells
play surface, they break up the continuity ofthe caused by moisture driven expansion, followed
surface coatings, and they expose the underlying by tensile stress in subsequent shrinkage. The con-
layers and adhesives to air, moisture and grime. cept that plywood is not an inherently stable mate-
As a precursor to responsible treatment of these rial but afragile one is stressed. Implicationsfor
objects, the nature and mechanism of these conservation are discussed, in particular the po-
checks need to be understood by conservators. tential dangers ofsurface stabilization and
reintegration.
The study to be discussed attempted to show a
tangible connection between an inherent vice in Introduction and Objective
rotarv-peeled plvwood veneer (lathe checks) and
the formation of surface checking in pfi,wood Plywood is a material showing up with greater
panels. It consists oftwo sections: a survev ofre- frequency in decorative arts collections. While
search literature on the mechanics ofthe veneer plywood is fundamentally wood, its structure
cutting process combined with a microscopical
and characteristics make it completely different
examination of plywoods qtcled to equilibrium
from solid timber. One problem evident in
at different moisture contents. A sample group of
many pieces of l9th- and 2Oth-century objects
p[u:oods was cycled through two difering mois-
ture extremes, representing an acceptable and an that incorporate plywood is longitudinal sur-
extreme climatic change, for 20 weeks. Periodic face veneer checking. These checks are disfigur-
samples were taken and examined microscopi- ing to the surface, they break up the continuity
callt in cross-section to assess any changes to ofthe surface coatings, and they expose under-
the wood structure. lying layers and adhesives to air, moisture
and grime.
Damage to the wood structure occurred much
more quickly than expected, even in the accept- To devise optimal treatment methods and pre-
able 3594 to 5994 relative humidity (RH) cycle. ventive conservation measures, the nature and

r55
origin of veneer checks need to be fully under- severe situation can be seen in a close-up ofthe
stood. Specifically, it needs to be ascertained chair by Charles Eames (Figures 3 and 4)
whether (and to what extent) the checks are a which he termed "DCW" @ining Chair Wood).
result of relative humidity (RH) fluctuation. Also, the effects of surface checking can be
(RH fluctuation could cause mechanical dam- seen on many painted plywood panels, both
age on the cellular level, which would be ob- new and old. Obviously, where ground and
served as checks and these checks could be paint layers are not flexible enough to remain
augmented by processing methods, such as intact over a developing check, they will crack,
rotary lathe checking.) in some cases much to the detriment of the
pictorial effect of the painting.
The purpose of this study is to provide empiri-
cal information on the nature and origin of sur-
face checks in plywood. The study will seek a
connection between rotary-peel lathe checks in
tangentially cutveneer, and surface check for-
mation in plywood panels composed of such ve-
neer. This will be shown by exposrng plywood
to numerous high/low extremes of RH, and sub-
sequently examining it in cross-section. The
study will underscore the misconception that
plywood is an inherently stable material. It
will also help in the construction of a plausible
model that explains the propagation of surface
checks in plywood, in relation to the veneer cut- Figure 2 Detail ofsurface veneer checking, Breuer
tabl e. Note the accompanying finish film cleavage.
ting process, and the hygromechanical proper-
ties of wood. It is hoped that the study will
create interest in plywood degradation among
other conservators, and provide some prelimi-
nary recommendations on approaches to the
conservation of ply.wood.

Consider some examples of surface checks in

plywood objects (Figures I and2). The frst is
a 1936 table by Marcel Breuer. The surface
checks are evident on this piece, and have exac-
erbated a condition of brittle delamination of
the cellulose nitrate hnish. A similar but less
iiil:t,ii:uii:iJlt::iiJJ\i]jjJ:ii\

Figure I NestingTable, Mqrcel Breuer, 1936. The

Metropolitan Museum of Art, Gtft of Daniel and Figure 3 DCW Chair, Charles Eames, 1946.
Suzanne Geller in memory of Bert and Phyllis Geller Collection of the Fine Arts Museum of
(1e8e.s2). San Francisco (I 987.62. I ).

156
 grain direction would render such designs
useless.

The third advantage, ofgreatest interest here, is

the dimensional stability of plywood. Plywood
is widely considered by non-specialists to be ul-
timately stable, that is, absolutely immune from
dimensional movement, since its alternating
layers effectively lock each other into place. ln
truth, it is not dimensionally stable. In going
from the green state, or with moisture content
at fiber saturation point, to the oven-dry condi-
Figure 4 Detail ofsurface veneer checks, tion, plyr,vood experiences a dimensional
DCW Chair. change that is but a fraction ofthat ofsolid
wood.^The shrinkage is 0.2o/o compared to
Plywood as an Art Material 8.0o ,2 smaller by a factor of 40. Because of its
significantly reduced potential for hygroscopic
Plyrvood has a diverse history of usage, which dimensional changes, plywood panels can be ef-
can be seen as a response to its unique charac- fectively glued to perpendicular or cross-grain
teristics. Plywood embodies at least three key frame elements. Such construction is either dan-
advantages. The flrst advantage lies in its gerous or impossible in solid wood, as can be
strength properties, which show remarkable seen in numerous fumiture case constructions,
advantages over solid wood of equivalent di- where drawer rails and leg posts have been fas-
mension.' Since plywood, by definition, con- tened tight across the width of the sides. Inevita-
sists ofan odd number ofveneer plies bound bly, such constructions lead to compression set
at 90o angles to one another, the weakness of the panel as relative swelling, caused by
inherent in a thin piece of wood that of stiff- moisture absorption, exceeds the proportional
ness or modulus of elasticity perpendicular to limit in compression of the wood. (Concise de-
the grain direction is mitigated. Plywood has scriptions of the nature of compression set are
virtually isotropic (i.e., directionless) strength given by Hoadley'and Stamm.") Since longitu-
properties in its "sheet" or planar dimension. dinal compressive and tensile strengths exceed
Obviously, a thin sheet of plywood, when incor- those perpendicular to the grain direction, sub-
porated into a seat, will support a much greater sequent shrinking results in tensile cracking
load than a piece of solid wood of similar size; across the grain.
a distinct advantage, elegantly exploited by
such designers as Alvaar Aalto in the 1930s. Plywood gains in strength and stability relative
to solid wood through its cross-banded struc-
A second advantage is the ability of pll.wood ture. Forced stability in plywood is not without
to be formed. Again, it is the fundamental pre- its drawbacks, however. The restricted dimen-
condition of being laminated that allowed users sional change (i.e., compared to solid wood) in
to modify plywood from flat sheets to fantastic- relation to moisture content change is the cause
ally curved forms like the Arabesque table of of many of the problems apparent in art and
Carlo Mollino. artifacts incorporating pl1.wood.

Such curved forms rely on the adhesive bond Lathe Checking and the Veneer
between laminae to "set" the shape upon cur- Cutting Process
ing. Solid wood could be processed to equiva-
lent shapes, but at great effoft, and again, Surface checking can be found on weathered
would maintain its unfortunate perpendicular- solid wood ut *E[ as plywood.s The mecha-
to-grain weakness. In addition, solid wood nisms involved in either case are not obvious.
shaped into tight curves would incorporate For plywood, it is also not obvious from which
areas of "short grain," wherein the far weaker direction (i.e., surface downward, or flom the
shear strength of solid wood petpendicular to

r57
undemeath working upwards) the checks upward and forward as a result of the cutting
evolve. The obvious culprits are lathe checks. wedge entering the tissue. Since the resulting
chip is effectively levered upward, there is a
Since the I 890s, rotary lathe cutting has been noteworthy compression on the upper surface
the preferred manner to obtain the large sec- of the wood blank.
tions of venee,r needed to form sheets for mak-
ing plywood.o A log or "blank" of wood, first The combined stresses form a shear force,
boiled or steamed for softening, is spun against resulting from the very complex mixture of ten-
alarge fixed knife, cutting a carefully control- sion and compression. The shear plane is ori-
led thickness ofcontinuous veneer. Such ve- ented at approximately 60o to the direction of
neer will obviously be sliced as a tangential cut. At some point in the cutting process, the
spiral from the blank. Unforfunate by-products shear stress will exceed the tensile strength of
of this cutting action are lathe checks, which the wood cells, and failure (i.e., a lathe check)
form on the downward or knife side of the ve- will occur. At this moment, the stresses are re-
neer being formed. Lathe checks are periodic, lieved by the rapid increase in strain (deflection
radially oriented cracks in veneer, resulting upward of the veneer). The stresses will be cy-
from a complex interaction of compressive _ clic, since both the tension and compression
and tensile stresses during rotary lathe cutting/ will increase as a chip is forced up the knife
(Figure 5). From early on, lathe operators face, and be released somewhat after failure,
learned to optimize the settings of the lathe and only to increase again as the cut continues.
the pre-conditioning of the wood to minimize
lathe checking. Refinements such as the nose- The result is a sheet of veneer with periodic
bar, a pre-compression apparatus, fupher re- checking on its under (or loose) side. Even with
duced the formation of lathe checks.d modem refinements, production of rotary-cut
veneer without knife checks is a theoretical pos-
sibility, and in practice, all such veneer will
contain checks to some extent.

The Hygromechanical Properties

of Plywood
ri
How might the hygromechanical properties of
plywood affect its degradation? As can be seen
trlI iri r" in the following data, there is a notable differ-
rtr'
Ir-t+\ ence in the equilibrium moisture content
(EMC) of plywood in comparison to that of,
solid wood up to the fiber saturation point.'"

Figure 5 Stress distribution duringveneer Plywood EMC RHo/o Solid Wood EMC
cutting. "C" denotes compression, "T" tension,
and "M" moment. (After Lenev 1960). 6% 40 -8%
10% 70 -13%
28k 100 -28"k
Theoretical work begun in the 1940s continues
In simple terms, plywood is less hygroscopic
to shed a great deal oflight on the origin of
than solid wood. It is prudent to consider what
lathe checks. There are a number of studies on
might cause this phenomenon. A ready explana-
this topic that are noteworthy.g-l3 Thir research
tion is the volumetric presence of a great deal
proved the phenomenon to be extremely com-
of adhesive. For a given volume of plywood,
plex. Based on these findings, a very idealized
compared to an equal volume of solid wood,
and simplified veneer cut can be summarized
more water-impervious material, in the form
here. As a knife enters a block of wood (Figure
ofadhesives, is present. But this is only part of
5), various compressive and tensile stresses
the answer.
become apparent. Wood cells are forced both

158
A more interesting explanation is that of demonstration was devised to enable a cross-
reduced hygroscopicity: wood under high com- sectional viewing of wood cells in plywood
pressive stress experiences a small but signifi- through numerous cycles ofRH changes. Con-
cant reduction in hygroscopicity, known as the ditions of extreme RH that an artifact might
Barkas effect.'' Obviously. individual plies, be- actually be exposed to were established. (The
ing crossJaminated, are restrained ffom swel- soaking, freezing and boiling associated with
ling in the tangential direction by the adhesive industrial standards were avoided. For an inter-
bond to the adjacent ply (or plies). esting critique ofindustrial test standards in
plyw6od. see Canoll.l8) Two relative humidity
We can deduce a situation in which hyg.o- regimes were chosen: a moderate one of 33%
scopicity reduces as RH and, therefore, mois- to 59%io RH, and a more drastic yet still realistic
fure content (MC) increases. It follows from one of l2o/o to 85% RH. Small chambers were
general principles that as long as there is any constructed, using saturated salt solutions to
degree of hygroscopicity in wood, any rise in create the two regimes. (The following satu-
RH must be accompanied by a certain degree rated salt solufions were used: lithium chloride
of swelling. It is easy to see the somewhat can- l2Yo,magnesium chloride 33%, sodium bro-
celing effects that further swelling would have mid,e 59o/o and ammonium sulfate 85%.)
on further hygroscopicity. But the reduction in
hygroscopicity will not be enough to account By monitoring the equilibrium moisture con-
for the reduction in swelling.seen in plywood tents (EMCs) of the samples, it was decided
as compared to solid wood.'o that a 3.5 day changeover time (i.e., a complete
low-high cycle took one week) was adequate.
Where does the swelling go? This is not a naive Weights were recorded at both the low and
question. On a cellular level, one theory hy- high EMCs to determine whether a significant
pothesizes "inward swelling, " where restrained gain or loss in weight occurred during the test.
wood is forced to expand "inside" into the cell At the end of the exposure period to low humid-
lumen, or the empty space. gt the center of ves- ity (EMC at minimum, shrinkage at maximum)
sels, fibers, and fracheids.' 'Theoretically. on a several circular samples (6 mm in diameter)
molecular level, enough compression might in- were removed from each test piece.
duce a physical/chemical change in the wood
tissue, with a redistribution of bonds between Two types of plywood samples were used in
polymeric cellulose chains within the lignin this demonstration. The first, representing a
matrix. Both models imply permanent set (this "new" piece of plywood, was commercial
phenomenon is known in the world of wood Baltic birch plywood. This consisted of nine
technology as compression set). As compressed plies of rotary-cut birch veneer, glued with a
wood shrinks in a decreasing RH, it will do so hot-press urea-formaldehyde adhesive. The sec-
from its compressed dimension, effectively try- ond, representing an "old" or historic plywood,
ing to further reduce its size. But plywood is was taken from an early 2Oth-cenflry object. It
bound, one layer to another. Eventually, the consisted ofthree plies oftangentially cut birch
compression of the swelling surface veneer veneer, with a casein or soya adhesive.
will become a tension that exceeds the tensile
strength of the shrinking veneer. In this man- All the samples were embedded, and examined
ner, through numerous MC cycles, we can hy- microscopically, in both reflected visible light
pothesize that tensile cellular failure might and long-wave ultraviolet illumination (UV ex-
occur in the surface veneer. This failure may citer wavelength 365 nm, splitter wavelength
perhaps eventually become a full-fledged check. 395 nm, barrier wavelength 420 nm). The con-
dition(s) of the surface veneers were noted and
An Empirical Demonstration photographed. Long-wave UV has been pre-
of Check Formation viously used in wood research to examine the
wood-adhesive interface and the onset of fun-
As a means of visually gauging the onset and gal degradation.l9'20 It has the advantage
propagation ofsurface checks in plywood, a of showing the lignin-rich areas of wood

r59
structure, and in this work aided the detection
ofcheck proliferation and cell wall breakage.

Results of Demonstration
l. The direction/evolution/resolution of lathe
checks.

The lathe checks in the new plywood, while

quite variable in nature, extended approxi-
mately one-third of the veneer thickness
(Figure 6). The dark lumens seen in the photo-
micrographs are those completely infused Figure 7 Typical lathe check in the uncycled "old"
sample. Note the nonlangential orientation of the
with embedding medium, and tend to indicate
upper veneer, as well as the compression visible in
increased cellular permeability due to damage. vessel elements.
Lathe checks occur with a spacing that is
roughly equal to the thickness ofthe veneer,
and extend upward at an angle ofapproxi- the samples was the extension of existing lathe
mately 45o. There appears to be a great ten- checks (Figure 8). After this first cycle, the
dency for lathe checks to resolve or end at lathe checks had extended a great distance into
some barrier point, such as a ray, a row of late- the veneer, in some cases to within 10 cells of
wood cells, or a vessel element. Figure 7 shows the surface. But most noteworthy was that the
a lathe check zone in the old plywood sample, degree of damage was largely independent of
the lower check resolvins at a ravlvessel the amplitude of RH variation. Note the dam-
intersection. age to the sample in the "extreme" cycle seen
in Figure 8, compared to the sample from the
"moderate" or regime more likely to be seen in
a museum, in Figure 9. There is an alarmingly
similar degree of damage. This very pointedly
underscores the sensitivity of plywood to
swings in RH in the early stages of its life. The
sample shown in Figure 10 again demonstrates
the possibility of extreme damage in one cycle.
Note the crushed layer of cells at the top of this
sample they are a result of compression and
-

Figure 6 Typical lathe check in the uncycled "new"

plywood, extending to the left at an angle, then
changing direction towards vertical. Note extent oJ'
adhesive layer penetration. (N.B.: all photomicro-
graphs are at 150 x.)

2. Checking evolves from lathe checks, much

more quickly than would be expected.

An original goal of this study was to visually

determine from which direction, up or down, Figure 8 Cycle #1, extreme RH range ("new" ply).
veneer checks originated and evolved. By the Note extension of crack system, and middle lamella
completion of the first humidity cycle, it was (inter-cellular) failure in vertical section at lower ar-
evident that the primary change occurring with row. The check terminates at a rqv. near the suiace.

160
 shown in Figure 14, vtrtually no damage that
can be directly ascribed to this humidity cy-
cling occurred, since extant surface coating ma-
terial can be seen penetrating the check. This
indicates that it is an old check. The great de-
gree of variation (Figure l5) in damage to the
older samples makes it difficult to connect any
damages noted with this demonstration with
particular causes. Since lathe/surface checks of
many different lengths and widths were noted
at the start, it became impossible to attribute
any widening or lengthening to further RH cy-
Figure 9 Cycle#1, moderate RH range ("new" ply) cling. It was apparent, however, that the fre-
Photo shot4)s relative damage after one cycle in the quency of checking did not increase, that is, it
moderate RH range. Top margin oJ photo cone- did not appear that any new cracks developed
sponds to surface of the veneer. during the demonstration in either type of sam-
ple. This strongly suggests that checks are only
associated with original lathe checks, since
no new checks developed, either from above
or below.

Figure l0 Cycle #1, extreme RH range ("new" ply).

Extension o;f crack system to within I0 cells o/
surface. Nctte crushed cells at surface infused with
embedding medium.
Figure I I Cycle #2, moderate kH range ("new"
ply). This sample shows no more damage than that
machine surfacing during manufacturing, and ofthe first cvcle.
are not directly related to MC cycling.

3. Very little change occurs after the initial

damage.

Aside from a slight "widening" of the area of

damage within a crack system, and a gradual in-
crease in the length and definition ofthe cracks,
very little change was noted in 18 humidity cy-
cles for both plywoods. Figures I I to 16 show
a series of cross-sections of plywood samples
exposed to an increasing number of RH cyclic
changes. In cycle 9 (Figure 12) the crack has
reached the surface, while another sample (Fig- Figure l2 Cycle #9, moderate RH range ("new"
ure 13) exposed to 18 cycles shows a check re- ply). Typical damage after nine cycles. Check is
solved at aray somewhat further down, albeit effectively to surJace. Note wider "footprint" oJ'
near the surface. In the aged pl1'wood sample crack system.

161
Figure I 3 Cycle #18, extreme RH range ("new" Figure l5 Cycle #18, moderate RH range ("old"
ply). Note long, vertical, middle lamella failure, as ply). No extended lathe checks are visible in this
well as severe and abrupt changes of direction. sample; only surface damage in conjunction with
vessel elements can be seen.

Figure 14 Cycle#6, moderate RH range ("old" ply).

Note the sur/hce coating moterial penetrating the Figure l6 Cycle #18, extreme RH range ("new"
check at the su4face. ply). Full check extension.

Two or three moderate humidity cycles pro- plausible explanation of the decline in crack
vided enough compressive stress to extend a propagation with increasing RH cycles is that
lathe check nearly to the surface, yet another l6 the lengthened crack becomes its own expan-
cycles caused, at most, the crack to be barely sion "release valve." Cracks provide a space to
complete and visible only under high magnifica- absorb expansion before it causes compressive
tion (Figure l6). Since the percentage of weight stress leading to fuither failure. It follows that
gain did not change noticeably with ftrther cy- as the crack lengthens or widens, the relative
cling, the amount of expansion due to adsorp- degree ofexpansion needed to induce compres-
tion must remain virnrally constant as shown in sive damage would increase. Hence the slow
the data below. progress ofthe crack after its initial extension.

Weiqht Gain Due to Moisture UDtake bv Plvwood 4. Checks do not typically evolve at ray
Samole Group Weight Gain % (Mean) interfaces on the surface and work downward.
new plv, extreme cvcle 2.56
old plv, extreme cvcle 3.58 One theory ofcheck propagation predicts that
new ply, mgderate cycle o.74 differing cellular orientation would make the
old ply, moderate cycle 0.84 interface of rays with vessels and fibers the
primary location of check formation.'' The
(Note the difference in moisture uptake be- reasoning is that any tangential or radial shrink-
tween the "new" and "old" sample groups.) One age in vessel or fiber tissue would be restrained

162
by the radially orientated rays, causing a local- degree ofRH change will be enough to propa-
ized stress zone. No downward extending gate checks that extend throughout the thick-
checks were ever noted at a raylfi,ber or raylves- ness ofplywood veneers, and in a short period.
sel interface. After l8 "extreme" cycles, the On the bright side, it also has been shown that
new plywood showed no surface damage (i.e., while damage may develop quickly, it occurs
damage other than the crushed outer layer of on a relatively minor scale, and the propagation
cells, an occurence that is independent ofhu- of damage is slow. While checks were gener-
midity cycling). In the old sample only, how- ated during the test, at no time were any
ever, a small number of surface checks of gaping checks produced, such as those on the
varying sizes were noted from the start. These Breuer table.
were always associated with a vessel element
(Figure 15). The fact that these checks often It is hoped that the importance of paintings in
extended fairly deep implied that they were not most collections will help to underscore this
merely a phenomenon caused by mechanical point, for we only have to look at a small sam-
surfacing. The frequency ofthese checks did pling of the many paintings executed on ply-
not change throughout the demonstration; in no wood grounds to see the effects ofsurface
instance was a surface check seen to connect veneer checking. For paintings on plywood, it
with a lathe check to become a full-thickness is not only the wood surface that is disfigured,
check. While surface checkingproper onply- but the decorative paint layers as well, quite
wood does exist, its frequency ofoccurrence is possibly to the great disfigurement of the pic-
far less than that of surface damage resulting ture. Just as furniture finishes tend to give way
from the extension ofa lathe check. This latter with the evolving check, so do the ground and
damage seems more in line with the explana- color layers in paintings.
tion of weathering offered by Feist and Hon for
solid wood.' No other explanation of the origin This leads to an important point with respect
ofthese checks can be offered. to recommendations. What do we do when we
are "too late"? As long as the RH variability
Conclusion and Conservation that causes checks continues, any filling of
Recommendations these checks will exacerbate the crack. The
fill provides a wedge for further compression,
It should be evident that plywood is far from which negates the "release valve" potential. It is
a perfectly stable material. In view of its poten- unknown at this time how great an effect filling
tial for checking, we must conclude that ply- will have. It is also unknown whether or not the
wood is not as stable as solid wood {iom a filling ofan applied decorative layer (such as
conservation point of view. Plywood was re- paint) with differing mechanical properties will
garded as a panacea on its inception, and as have a subsequent effect on the underlying ply-
a material that allowed artistic freedom for wood veneer. The object's environment, and
designers in the 1920s and 1930s. We as conser- the mechanical properties of the fill material,
vators, however, must now consider plywood will play an imporiant role in determining the
as a product fraught with inherent problems. extent of further damage. These are important
This insight demands a deparhre from the intui- areas for further study of plywood
tive understanding held by most in favor of a conservation.
new understanding: plywood is not stable, not
inert, and is not able to survive the most de- Checks also provide a receptacle into which
manding environmental conditions. dust and airbome grime may settle. Covering
plywood objects, whenever possible, is there-
Ifplywood surface checks are considered to be fore a more important preventive conservation
disfiguring to objects, then pieces in decorative measure than it is for solid wood objects. Not
arts collections composed of plywood should only do years of accumulated grime begin to
and must be considered to be as fragile or pose the same, inadvertent potential harm as a
perhaps more fragile than those constructed of fill, but dirt-filled cracks have the added effect
solid wood. It has been shown that even a mild of becoming "beautiful black lines" across

163
blonde wood, a problem known and loved by Les dommages touchant la structure du bois
all who have tried to deal with it. sont appants beaucoup pltts rapidement que
privu, m€me dans le cas du cycle reprdsentant
Further investigations will perhaps yield en- un changenent de conditions ambiantes accep-
table (de 35 d 59'% d'humiditd relative). D'is la
couraging treatment possibilities. For now there
troisiime semaine, les craquelures qui partaient
are no easy answers. Surface veneer checking is
de la craquelure de ddroulage originale attei-
in large part a classic "inherent vice" problem.
gnaient un,e profondeur de quelque 75 d 80 % du
Although latent, it is a difficulty we must unfor- placage. A la neuviime semaine, nombre d'entre
tunately live with. elles avaient atteint la surface du panneau. Et, d
la vingtidme semaine, la maioritd s'dtaient dten-
R6sum6 dues d la sudace, inddpendamment de I'amplett
du changement d' humiditt.
Une etude priliminaire de l'origine et de Ia
nature des craquelures qui se produisent dans Ie Un moddle est proposi pour expliquer la.forma-
placage des contreplaquis tion des craquelures. Il sefonde sur les dom-
mages risultant de la compression que subissent
Le contreplaqut est un matiriau qui se retrouve les cellules de bois lorsque I'humiditd, qui afa-
de plus en plus souvent dans les collections d'art vorist wte expansion du mattriau, finit par se
ddcoratifet de beaux-arts. Bien que ce produit retirer, lui intposant ainsi un effurt de tension.
soitfondatnentalementfait de bois, sa structure et On insiste sur lefait que le contreplaqud est un
ses caracftristiques le placent dans une cattgorie matiriau non pas stable maisfragile. Et I'on
totalement dilftrente de celle des pi'ices de bois traite de l'incidence de ces dtcouvertes sur la
ntassif. Les craquelures longitudinales qui se pro- consen,ation de ce matdriau, notamment en
duisent dans le placage de nombreux meubles ou ce qui a trait aux risques tventuels que peuvent
supports de peinture constituent un exemple des prdsenter Ia stabilisation et la r2paration
probldmes de dbgradation que peut prdsenter ce de la sudace.
matdriau. De telles craquelures d,!figurent la sur-
face exposie, elles brisent la continuitd des rev€te- References
ments de surface et ellesfont que les couches
sous-jacentes et les adhdsifs entrent en contact
l. Forest Products Laboratory, Forest Service,
avec I'air, I'humidtrd et Ia saletd. Pour qu'un
U.S. Dept. of Agriculture, Handbook of lilood
traitement tclaird puisse €tre appliqut d ces
and Wood-Bas ed Materia ls.for Engineers,
objets, ilfaut d'abord que les specialistes de
la restauration comprennent la nature de ces Architects, and Builders (New York: Hemi-
craquelures et les mdcanismes qui prdsident d sphere Publishing Corp., 1989) pp. 1l-17.
leur formation.
2. Schniewind, A.P., ed., Concise Encyclopedia
La prtsente communicationfail ttat d'une ttude of llood and llood-Based Materials (Oxford:
qui visait d etablir I'existence d'un lien riel entre Pergamon Press, 1989) p.224.
un vice prdsent dans le placage (des craquelures
de ddroulage) et la formation de craquelures d la 3. Hoadley, Bruce, Understanding Wood
surface du contreplaqui. Elle comportait deux vo- (Newtown, CT: The Taunton Press, 1980).
lets : I'etubhssement d'une liste de documents de
recherche portant sur la mtcanique du dtroulage 4. Stamm, A.J.. Wood and Cellulose Science
des placages el un examen au nticroscope de (New York: Ronald Press, 1964).
contreplaquds qui avaient traversd un cycle, en
iquilibre stable, d diverses teneurs en humidit2. 5. Feist, W.C. and D. Hon, "Chemistry of
Un 1channllon de contreplaqud a ainsi ttd sou'
Weathering and Protection," in: The Chemistry
mis, pendant vingt semaines, d deux cltcles de
of Solid l(ood, ACS, Advances in Chemistry
t a ux d' h umi di t t diffh'e n t s, rep rt s e n t a n t re sp e c -
Series, Roger Rowell, ed. (Washington D.C.:
tivement un changentent de conditions ambiantes
acceptable et extr€me. Des prdl'ivements oil eft American Chemical Society, 1984).
faits pdriodiquement, qui ont ttt examints, en
coupe, au microscope pour dvaluer toute
modification de la structure du bois.

164
 6. Susan Klim, "Composite Wood Materials in 1 4. Schniewind. Concise Encl,clopedia, 1989,
Twentieth Century Fumiture," Preprints of the p.224.
Wooden Artifucts Group, American Institute
for Conservation Annuql Meeting, 1990. 15. Barkas, W.W., "The Swelling of Wood
Under Stress," Forest Products Research
7 . Leney, L., Mechanism of Veneer Formation Bullerin (G.8.),1946.
at the Cellular Level, Research Bulletin
No. 744, University of Missouri Agricultural 16. Kollman, F.P., E.W. Kuenzi and A.J.
Experiment Station, 1960. Stamm, Principles of lTood Science and
Technologt Vol. II: l4/ood Based Materiqls
8. McMillan, C.W., "The Relation of (New York: Springer Verlag, 1975) p. 116.
Mechanical Properties of Wood and Nosebar
Pressure in Production of Veneer," Forest 17. Skaar, C., "Wood-Water Relationships," in:
Products Journal, vol. 8, no. 1, 1958, The Chemistry of Solid Wood, ACS, Advances
pp.23-32. in Chemistry Series, Roger Rowell, ed.
(Washington, D.C.: American Chemical
9. Fleischer, H.O., "Experiments in Rotary Society, 1984) p. 143.
Veneer Cutting," Forest Products Research
SocieQ Proceedings, vol. 3, no. 20,1949. 18. Canoll. M.N.. "We Still Don't Boil
Houses: Examination of the Boil-Dry-Boil
10. Kivimaa, E., "Investigating Rotary Veneer Test in Plywood Standards," Forest Products
Cutting with the Aid of a Tension Test," Journal, vol. 28, no. 5, 1978, pp.23-27.
Forest Products Journal, vol. 6, no. 7, 1957,
pp.251-255. 19. Murmanis, L. and B. River, "Microscopy of
Abrasive-Planed and Knife-Planed Surfaces in
1 l. Feihl, O., "Peeling Defects in Veneer: Wood-Adhesive Bonds," Wood and Fiber
Their Causes and Control," Canadian Forest Science, vol. 15, no. 2, pp. 102-l 15.
Products Research Bureau Technical Note
No.25. 20. Krahmer, R., "Detecting Incipient Brown
Rot with Fluorescence Microscopy," llood
12. Cumming, J.D. and B.M. Collett, Science, vol. 15, no. 2, pp. 78-80.
"Determining Lathe Settings for Optimum
Veneer Quality," Forest Products Journal, 21.Schniewind. A.P.. "Mechanism of Check
vol.20, no. ll, 1970,pp.20-27. Formation," Forest Products Journal, vol. 13,
no. I 1, 1963, pp. 47 5-480.
13. Chih, L.C., Fundamental Analysis of the
Ven eer C u tti ng Pro ces s, unpublished
dissertation (University of Califomia at
Berkeley, 1970).

165
Case Studies and Specific Problems with Materials
Etudes de cas et problimes particuliers pos6s par les mat6riaux
Deterioration of Cellulose Nitrate Sculptures
Made by Gabo and Pevsner

Michele Derrick, Dusan Stulik Eugena Ordonez

Getty C onserv ation Institute The Museum of Modern Art

Marina del Rey, California New York, New York
U.S.A. U.S.A.

Abstract together to create open structures in which the

interior space of the sculptures actively inter-
Three sculptures by Naum Gabo and Antoine acted with the exterior space. That is, Gabo
Pevsner in The Museum of Modern Art's collec- started constructing sculpture in terms of space
tion exhibit varying degrees of deteioration instead of volume or mass. A few years later,
and rangefrom very good to poor in overall Gabo's older brother, Antoine Pevsner, a
condition. The sculptures are made of thermally painter, also began to create three-dimensional
shaped pieces ofcellulose ester (cellulose nitrate) constructions. Actually, Gabo described their
attached together with an adhesive also based on works as four-dimensional because time as ex-
cellulose ester. The translucent cellulose nitrate
pressed through rhythm was also an element
pieces, which originally may have been uncol-
(Lassaw 1957, 160). The works of these two
ored, are now shades ofbrown. The opaci/ied
brothers highlight the formal vocabulary of the
pieces, which were colored white or black, have
most likel, also changed in tone. Some of the constructivist tradition for which conventional
pieces exhi bit finefi, spaced cracks whi le others methods and materials were no longer applica-
have small drops of a clear to light-brown liquid ble. Although Gabo and Pevsner did not ascribe
present on the surface. Conosion products arc to the utilitarian, propagandistic agenda that the
frequently found where metallic components are Russian constructivists, such as Vladimir Tatlin
near the deteriorated cellulose nitrate. The proc- and Alexander Rodchenko, pursued in their art,
ess of deteioration of these works as well as there are formal similarities in these artists'
works from other collections were studied using work, including their unconventional use of
F o uri e r t ra n sfo rm i nfr ar ed nt i c ro s c opy (F TI R ), modern materials. (See Lodder 1983 for an
X-ray dffiaction (XRD), scanning electron mi- in-depth study of Russian constructivism.)
croscopy coupled with X-ray spectrometry and
backscatter electron imaging (SEM-EDS) and The manufacture of new materials, such as plas-
light microscopy. Information obtained about tics, allowed artists to explore new territories
the possible deterioration mechanisms and the
technically. The modem plastics of the early
role thatfllers and plasticizers may plalt is pre-
1900s provided sculptors with a medium that
sented and proposed conservation treatments
was lightweight, easily cut and molded into a
are discussed.
variety of shape-maintaining forms. The plastic
could be made either transparent or opaque and
Introduction in many colors. The versatility of this new ma-
Naum Gabo moved to Oslo in l9l4 where in terial of the 20th century appears to have been
perfect for their sculptures of the future except
1915 he created his first constructions. These
consisted of flat and curved planes joined
for one problem, the plastics in the early works

r69
of Gabo and Pevsner were cellulose esters. The which is usually responsible for the fast deterio-
early cellulose-derived plastics are inherently ration ofa work, is dependent on the presence
unstable and undergo slow, spontaneous degra- ofoxygen and water because the nitrogen ox-
dation under normal room conditions. ides themselves do not react with the cellulose
nitrate polymer.
This study incorporates the results of the exami-
nation and analysis ofthree sculptures in The Nearly every step in the production of cellulose
Museum of Modem Art's (MoMA) collection nitrate has an effect on the final stability of the
in New York as well as the general findings on material. Cellulose nitrate is prepared by treat-
several works by Gabo and Pevsner in other ing purified cellulose with nitric acid in the
museums. In all six works sfudied, covering a presence of sulfuric acid. The starting cellulose
period ofabout l0 years (1917 to 1926), cellu- needs to be lignin-free since lignin is unstable
lose-based plastics were an integral parl of the and can photodegrade to produce a complex
construction. The range and degree ofdeteriora- mixture of acidic materials that will promote
tion that these plastics have undergone reflect degradation of the cellulose nitrate (Yarsley
numerous factors, including the composition 1964). However, because pure cellulose was ex-
of the starting material for the plastic and the pensive at the tum of the century, many compa-
environmental history of the work of art. nies bought cheaper material, then cleaned and
bleached it. This processing usually left resid-
Cellulose Nitrate Preparation ual acids and chlorine in the cellulose, which in
and Deterioration turn could speed up the deterioration ofthe cel-
lulose nihate. Thus, to regain some stability, ad-
Pullen and Heuman (1988) found that a number ditional materials, such as sodium bisulfite and
of Gabo's sculptures from the 1920s in the Tate triphenyl phosphate, were put in as scavengers
Gallery were made fiom cellulose nitrate and ofthe acids and chlorine.
that "most were in good condition." Their study
showed that the problematic plastic sculptures In the nitration step, sulfuric acid was incorpo-
were those made from cellulose acetate. Inter- rated to safely control the rate and extent of
estingly though, all four works by Pevsner in the reaction. However, this resulted in small
the present study were found to be made from amounts of sulfate esters in the polymer, which
cellulose nitrate and these works showed seri- later produced sulfuric acid and initiated rapid
ous overall or local deterioration. Prior to dis- deterioration. The early remedy for this was to
cussing six works by Gabo and Pevsner, a brief add alkaline additives to the polymer. But when
background on cellulose nitrate, its manufac- added in small amounts, the alkalis were used
ture and degradation processes, will be given. up rapidly while large amounts of added alkali
attacked the cellulose nitrate itself.
Thermal, hydrolytic and photochemical proc-
esses are the primary decomposition mecha- The nitrated cellulose (pyroxylin) was com-
nisms for the degradation of cellulose nitrate. bined with necessary solvents and plasticizers
(For a thorough description ofthese processes to form a clear, plastic material. A variety of
see Selwitz 1988.) In addition, secondary plasticizers were tried for the production of cel-
mechanisms due to reactions of additives and lulose nitrate, such as phosphates, phthalates,
impurities in the cellulose nitrate also influence oils, resins, gums and camphor. Camphor was
the degradation rates. Even when prepared un- considered the best and was the most com-
der the best conditions, cellulose nitrate under- monly used since it not only acts as a plasti-
goes slow, spontaneous degradation in which cizer but also as a stabilizer. It forms hydrogen
the NOz groups split off to form nitrogen oxide bonds with the unsubstituted hydroxyl group
gases. This thermal degradation is accelerated producing a non-crystalline, solid solution of
by ultraviolet light. Upon exposure to water cellulose nitrate that is strong and flexible. Cel-
and oxygen, the nitrogen oxides form nitric and lulose nitrate with camphor as its plasticizer is
nitrous acids, which then catalyze hydrolyic de- referred to as celluloid.
composition reactions. The autocatalyic cycle,

170
Although camphor was the best plasticizer were used to identify and quantify additives in
available it caused some problems. Impurities the polymer as well as identify the principal
in the camphor are one of the causes of the yel- component of the plastic.
lowing of cellulose nitrate objects. Also, be-
cause it forms hydrogen bonds with the Sculptures
cellulose nitrate, the proportion of camphor in
the final product is important. Too little cam- The earliest sculpture included in the present
phor results in a brittle material subject to craz- study is Head of a Woman by Naum Gabo,
ing. On the other hand, excess camphor slowly shown in Figure 1 and in the collection of The
sublimes from the object, leaving a porous or Museum of Modem Art. It is generally be-
cracked material that is more susceptible to oxi- lieved to have been made in Russia after the art-
dation and hydrolysis and thus to accelerated ist's retum there in 1917, although the exact
deterioration. Even in materials that have the date of execution is not known. Nash bnd Merk-
optimum proportion of camphor (20%to 40%), ert (1985) give arguments for possible dates of
the camphor will slowly sublime from the ex- 1917 to 1920 as well as circa 1923.
posed surfaces promoting deterioration, and
when the object degrades, more surface is usu-
ally exposed and camphor is lost more rapidly.
(For further information on the use of camphor
in celluloid see Reilly 1991.)

Stabilizers were often added to scavenge free

radicals and unwanted acids. Urea and diphenyl
amine were found to be effective initially, but
later crystallized out and became more harmful
than beneficial. Zinc oxide was commonly used
as a filler and opacifier. It also acts as a stabi-
lizer by rapidly consuming nitric acid to form
zinc nitrates.

Sample Analyses

The decomposition of cellulose esters has been

documented and studied extensively and the in-
stability of these materials is generally acknow-
ledged (Sirkis 1982). What is intriguing in the
case of these sculptures, however, is the type
and extent ofdegradation that occurs from
work to work and flom location to location
Figure I Naum Gabo, Head of a Woman (ca. l9l7
within a sculpture. To obtain a better under-
to 1920, after awork oJ 1916). Construction in
standing of the relation between various celluloid and metal, 62.2 cm x 48.9 cm x 35.4 cm.
deterioration mechanisms and the material Collection, The Museum oJ Modem Art, New York,
composition, chemical analyses were per- Purchase.397. 38.
formed on samples from several sculpfures
and from several sites within each sculpture
The work was formed ffom thermally shaped,
whenever possible.
cream-colored pieces of opacified cellulose ni-
trate. Frequently Gabo would make models of a
The samples, acquired as broken fragments or
work in materials, such as cardboard or wood,
removed as scrapings. cross-sections and
prior to executing the sculpture in metal. In a
exudates, were analyzed by FTIR microscopy
letter from 1948, Gabo described MoMA's
at the Getty Conservation Institute and by light
Head of a lloman as being "finished and com-
microscopy, SEM-EDS and X-ray diffraction at
plete" but "in fact, the original model which I
The Museum of Modem Ar1. These methods

t71
always make before executing a work in a In Alexi Pevsner's biography of his two broth-
proper and solid material. In the case of this ers (1964), Naum and Antoine, he mentions
head...the material is celluloid and therefore un- that Gabo taught Pevsner "how to handle metal
stable..." (Nash and Merkert 1985,202).In and plastic materials." Head of a Woman (Fig-
1951 and 1964 Gabo borrowed the head to use ure 2) made by Antoine Pevsner is dated 1923
as a model for a bronze version now in the col- (inscribed under the artist's signature in the
lection of Gabo's family. lower left corner) and is in the collection of the
Hirshhom Museum, Washington, D.C. The
Since its acquisition in 1938, the sculpture has work is in extremely poor condition and is no
been frequently exhibited at the Museum. longer exhibited. The sculpture's pieces were
When on display, it has been hung above eye made from translucent cellulose nitrate except
level in a comer with no vitrine in a similar for the creamy-white base that was opacified
fashion to the way it was presented in the Van with zinc oxide. As shown in Figure 3 the trans-
Diemen Gallery in 1922. (See Nash and Merk- lucent pieces have yellowed and cracked se-
ert 1985, 2l for aphotograph ofthe 1922 exhi- verely, joined pieces have separated and forms
bition showing this work. It should be noted have fallen off. The pieces have also distorted
that Nash believes the object shown in this pho- out-of-plane by forming bubbled areas.
tograph is actually an earlier metallic version
and not this celluloid version as other scholars
believe.) In contrast to some other works done
about this time by Gabo and Pevsner, to be dis-
cussed, this work is in very good condition.
Nonetheless, by 1947, Gabo objected to the
work being exhibited in its "present condition"
(Richards 1947).lt appears that the damage in-
cluded at least two pieces of plastic that had
fallen off. The work was not repaired by the afi-
ist in time for the Gabo-Pevsner Exhibition in
1948 and therefore was not included. It may a ::,1:'aa.:a/:l

have been this sculpture that Gabo is indirectly Figure 2Antoine Pevsner, Head of a Woman
quoted as describing as "an early celluloid (1923) Hirshhorn Museum and Sculpture Garden
(Photo taken in l99l).
model, now in the collection of Alffed Barr,
[which] has cracked and discolored severely,
but was not intended to be permanent" (Siegl When the work was acquired in the early
and d'Hamoncourt 1968). 1970s, it was placed in a wall-mounted vitrine
case. By 1982 the sculpture had darkened,
Currently, there are long, fine cracks especially cracks had formed along the join lines and
in the forehead area, some broken or missing "beads of clear oily liquid" were present on
pieces and separation ofa few pieces from the the surface of the sculpfure (Aks 1991). These
main body, for example, the right proper shoul- drops were analyzed in 1983 and found to be
der has separated along the join line. Most of an aqueous basic zinc nitrate solution (Erhardt
the damage appears to have resulted from me- et al. 1983). The 1983 analytical report fuither
chanically induced stresses caused by handling states that "zinc nitrate is a result ofattack of
the work or environmental factors. The pieces nitric acid on the zinc oxide pigment of the cel-
are a creamy-white except for three that are dis- luloid on which the drops form....It was only
tinctly yellow, namely the proper right shoulder when the object was sealed up that the gener-
and two pieces near the chin area. FTIR analy- ated acid was confined so that reaction with
sis of the discolored pieces showed them to be the available alkali (zinc oxide) became inevita-
cellulose nitrate (as were the other pieces). X- ble." After cleaning the sculpture, it was placed
ray diffraction analysis indicated that zinc ox- in storage in its vitrine and an efforl was made
ide was the principal opacifying pigment both to minimize exposure to light by covering the
in the yellow and cream-colored pieces. The case with brown paper. By 1988 the sculpture
cause ofdiscoloration has not been determined. was in a "severe state of deterioration....A clear

112
 Figure 4 Antoine Pevsner, Bust (1923 to
1924). Construction in metal and celluloid,
53 cm x 59.4 cm. Collection, The Museum of
Modern Art, New York. Purchase. 396. 38.

color, as shown in Figure 5), as well as the

formation of light-brown drops of liquid over
several ofthe pieces. There is also localized cor-
rosion of the metal in areas where the deterio-
Figure 3 Antoine Pevsner, Head of a Woman, rated cellulose nitrate contacts metallic pieces,
detail of bottom portion. such as the copper background or iron bolts.

residue has reformed over the entire sculpfural

surface mostly in the form of oily beads" (Aks
1988). It was also noted that the cracking had
worsened significantly and that "several plastic
panels had collapsed and separated from the
sculpture" (Aks 1988). For a while the work
was kept in a fume hood where it was well ven-
tilated and currently it is kept in an uncovered
ply.wood box in storage. When the work was
examined recently for this study, only a small
amount of the oily drops was observed.

The presence ofa liquid exudate has been no- Figure 5 Antoine Pevsner, Bust, detail ofupper
ticed on other works such as Pevsner's -Basl portion showing extensive cracking of a translucent
(shown in Figure 4) completed between 1923 piece and spotting on an opaque piece (Photo taken
199r).
and 1924. This relief sculpture, acquired by
MoMA in 1938, is in relatively good condition
even though it also has problems. It is made out Until 1980 this work was frequently exhibited
of translucent and opacified cellulose nitrate at MoMA. Exhibition photographs show that it
pieces, the latter primarily opacified with zinc was hrurg without a vitrine on a wall. In 1989,
oxide. The cellulose nitrate has deteriorated in a for as yet undetermined reasons, the sculpture
varief of ways, including cracking, discolora- started producing a liquid exudate especially
tion of translucent pieces, spotting (the pres- on the opacified cellulose nitrate pieces,
ence of small, roundish areas slightly whiter in such as the topmost curved piece. A small
portion of the liquid was analyzed with FTIR

t73
 works at MoMA showed that potassium, chlo-
rine, silicon, calcium and sulfur are usually pre-
sent in both the opacified and the translucent
cellulose nitrates.

Figure 6 Calcium sulfate dihydrate crystals

precipitated out of liquid degradation product,
5 00 times magniJication.

spectroscopy and determined to be primarily Figure 7 Higher atomic weight elements appear
composed of inorganic nitrate(s). Over several brighter. Top surface ofthe cross-section is diago-
months, while the liquid was stored in a cov- nally across the lower left corner, 500 times
ered well-slide, crystal formation took place. magniJication.
With slight heat more crystals formed. Several
chemical species appear to be present in the A third piece at MoMA, Zorso (illustrated in
exudate and have only been partly identified at Figure 8) dated 1924 to 1926, was also made
this time. X-ray diffraction of a large cluster of by Pevsner. This work consists primarily of
crystals showed these to be calcium sulfate di- translucent cellulose nitrate pieces that are
hydrate. Figure 6 is a secondary electron image often attached with metal hardware to similarly
of these crystals magnified approximately 500 shaped copper pieces. (Figure 9 is a detail of
times. Energy-dispersive spectrometric analysis the proper right hand showing the copper and
(EDS) of the same crystals showed that along cellulose nitrate held together with a screw.)
with calcium and sulfur, zinc was also present Torso was exhibited for many years on a pedes-
(though apparently not as a crystalline com- tal and without a vitrine. In December 1954,
pound since it was not distinguished by XRD). the sculpture was accidently knocked over
SEM-EDS examination of a cross-section from while on view, resulting in several broken
an area where excessive liquid had accumulated pieces. The treatment report states that the
indicates that there is about a 20 micron deple- physical effects of the fall were exacerbated by
tion of higher atomic weight elements {iom the the deteriorated condition of the cellulose ni-
surface ofthe piece as indicated in Figure 7. trate (which was incorrectly identified as cellu-
(Because of the sample size required for cross-
lose acetate in the report). The treatment
sections and the delicate nature ofthe sculp- consisted of replacing a number of the pieces,
ture, confirmation of these preliminary results mostly around the lower head area. After being
will have to be carried out on cellulose nitrate repaired, the work was heavily exhibited until
artifacts exhibiting similar symptoms.) Macro- 1968 when it was removed after the black cellu-
scopically, this effect was observed by an in- lose nitrate base (which had been progressively
creased translucency at the surface, albeit deteriorating since at least 1964) was too buck-
exhibiting a slightly browner tonality. It thus led and fragmented. For most of the time be-
appears that zinc salts as well as other com- tween 1968 and 1978 the sculpture was kept in
pounds (such as calcium salts) are leached out storage. Part ofthe reason for storing this ob-
as part ofthe degradation process. EDS analy- ject, as well as other cellulose nitrate objects,
sis on a number of small scrapings of cellulose was the fear that they would spontaneously
nitrate from this work as well as the two other

174
 exhibition and to lightly drape them with paper
to keep dust off if they are not on view.

Torso was once described as being able to

"exploit light through the various densities
and directions of the planes" (Read 1948). The
transparency ofthe plastic was also used to set
up a "rhyhmic dialogue" with the curved cop-
perbacking (Rankin 1988). The photograph of
the sculpture presented in this paper was taken
in 1933 and indeed shows how light could have
penetrated the cellulose nitrate forms and re-
flected offthe copper pieces. It no longer is
able to do this as effectively because the cellu-
lose nitrate pieces are all now discolored and
some have cracked badly enough to form essen-
tially opaque layers. A detail ofthe head area
(illustrated in Figure l0) shows what is left of a
cellulose nitrate piece and how corroded some
ofthe copper pieces have become.

Figure 8 Antoine Pevsner, Torso (1924 to

1926). Construction in plastic and copper,
74.9 cm x 29.4 cm x 38.7 cm. Collection, The
Museum of Modern Art, New York.
Ktttherine S. Dreier Beauest. ,/85. 53.

Figure I0 Pevsner,Torso, detailJi'om head area

(Photo taken in 1991).

At present there is only a slight amount of

liquid formation on the cellulose nitrate sur-
face. FTIR analysis of the liquid, which in this
case crystallized out instantly on contact with a
Figure 9 Pevsner, Torso, detail ofhand (Photo microscope slide, showed it to be inorganic ni-
taken in I99I). trates (Figure I I is an FTIR spectrum of these
crystals). The XRD pattem indicated the pres-
combust. Although it was unlikely that the ob- ence of ammonium nitrate as one of the compo-
jects would combust since the percent nitrate in nents (JCPDS 8452). The copper pieces have
polymers (10%) is lower than that in explosives also corroded to various degrees forming a blue-
(13% to l4o/o) or frlm (l2o/o), suggestions were green, very fine-grained powder. Figure l2 is a
made to put them in airtight vitrines filled with backscatter electron image of the powder show-
nitrogen or to store them in explosion-proof ing the general size distribution ofthe particles.
steel cabinets. It appears instead that Torso was Analyses so far indicate that it is a form of
kept in storage with a plastic film draped over copper nitrate, possibly gerhardite. The same
it. The present policy at MoMA is to keep air corrosion product was found on the copper
freely circulating around these objects while on backins in Pevsner's Busf.

115
Figure I I FTIR spectrum of crystallized liquid
exudate from Pevsner Torso.
very yellow and brittle

3998 3332 2666 2000 1741 182 1223 9U 705

wavenumber (CM_r)

Figure I 3 Infrared spectrafor cellulose nitrate at

va ri ous s ta ges of degrada t ion.

Figure I 2 Conosion product from the Pevsner

Torso, 2,000 times magnifcation.

FTIR analysis ofcondensed fractions collected

from heating cellulose nitrate pieces show that
camphor was used as a plasticizing agent along
with an additional oil. There is a direct correla-
tion between the absence of camphor and se-
vere degradation of the sample as shown in
Figure 13. The top infrared spectrum is ofcellu-
lose nitrate in good condition and the bottom
one is ofthe nitrate salt degradation product.
Two stages of the deterioration are shown in
between. Noticeable in the spectra is the in-
crease in the intensity of a nitrate salt band
at 1340 cm-'. There is a corresponding de-
crease in the intensity of the carbonyl band at
1735 cm'' , which is due to decreasing amounts
of camphor, the plasticizer. (See Figure 14 for
reference spectra ofcellulose nitrate and
wavenumber (cM_l)
camphor.)
Figure 14 Infrared spectrafor cellulose nitrate,
Cross-sections taken from broken fragments camphor and liquid obtained from the Pevsner
of the sculpture were examined to assess the Torso.

r76
extent of deterioration into the pieces of cellu-
lose nitrate (i.e., depth-wise). Figure l5 shows
an inffared spectrum produced by analyzingthe
surface ofthe cellulose nitrate piece. The other
spectra were collected further into the sample,
specifically at I mm and 2 mm. The spectra
show that while the surface of the sculpture con-
tains cellulose nitrate with a diminished amount
of plasticizer, the celluloid 2 mm into the sam-
ple is in good condition.

2 mm inside

1 mm inside

Outside surlace

F i gu re t 5 ;:;":;; ";:;i:o * *
"
co r e c te d a t
dffirent ", ",,in a cross-section sample
depths from the
PevsnerTorso. The spectra obtainedfrom the outer
Figure I6 Antoine Pevsner, Portralt of Marcel
area ofthe cross-section show that there is a
Duchamp, Yale University Art Gallery, Gift of
depletion of camphor near the suryface.
Collection Sociiti Anonyme.

Another work of this period by Pevsner is the The sculpture is undergoing degradation proc-
Portrait of Marcel Duchamp in the Yale Uni- esses, such as excessive liquid production on
versity Art Gallery collection, which was com- the surface (illustrated in Figure l7), cracking
missioned in the spring of 1926 and completed and buckling of the cellulose nitrate pieces and
in the summer of the same year. This work, like extensive corrosion of metallic parts. These me-
the Bust mentioned above, is a relief attached tallic parts include the copper backing, which is
to a metallic backing. The arlist used both clear now covered with a blue-green corrosion prod-
and opaque plastic for this sculpture including uct, an iron-based strip of metal, which is now
pieces colored purple and black. By 1941, a pile of rust and the copper- and iron-based
when the Yale University Art Gallery acquired hardware, which have also corroded. The Yale
the work, conservation records indicated that University Art Gallery files indicate that at
pieces were already warped and discolored. It least since 1984 until now the work has been
was restored in 1957 by the artist after it was kept in storage in an essentially airtight vitrine.
smashed en route to an exhibition. At that time Andrew Petryn, former conseryator at Yale,
the artist replaced broken pieces and changed thought that the work might have been kept
the backing (originally a zinc plate) to copper in a vitrine for a much longer period of time,
because the latter harmonized better with the although an exact date was not known
discolored pieces (Herbert 1984). Figure l6 (personal communication, I 99 l).
shows the work after the 1957 repair. Within a
few years the copper backing started corroding A detailed photograph from late 1984 shows
as the plastic pieces kept deteriorating. By an active corrosion problem but for the most
7979, a replica in plexiglass had been made for part the pieces ofcellulose nitrate do not appear
the gallery by a former assistant of Pevsner badly cracked or out of plane. As mentioned
(Herbert 1984). earlier though, the pieces are now badly

177
Figure 1 7 Pevsner, Portrait of Marcel Duchamp,
detail ofbottom edge oJ'sculpture (Photo taken
in 199 I ).

cracked, many of them exhibiting widely

spaced cracks with the plastic in between the
cracks badly buckled, as shown in Figure 18.
It thus appears that a significant amount of the
cracking and buckling has occurred in the last
few years. As will be discussed fuither in the
following paragraphs, this type of damage ap-
pears to occur when the objects are kept in a
sealed vitrine and then suddenly exposed to a
new environment when the vitrine is opened.
This change in environment allows the immedi-
ate loss ofvolatile degradation byproducts.
This in furn appears to create intemal stresses Figure I9 Antoine Pevsner, Portrait of Marcel
subsequently released by crack formation. It Duchamp, Yale University Art Gallery, Gilt of
C ol le c tion S oc i et e Anorytme.
is known that the vitrine of the Portrait of

Marcel Duchampwas opened in 1986 to re-

move some fragments that had fallen offf.
Rapid deterioration appears to have followed as
shown by comparing the photograph in Figure
19 (taken around the mid-1980s) with the re-
cently photographed details pictured in
Figures 17 and 18.

When Gabo was asked how the invention of

plastics affected the course of his works, he re-
plied, "I did a great deal of work in plastics for
only one reason: to accentuate the transparent
character ofspace" (Kuh 1962, 102). The cones
in Construction in Space: Two Cones (1921)
in the Philadelphia Museum of Art exemplify
Gabo's use of plastic for its transparent quali-
ties. The artist also used opaque plastics
though, in white, black and red for some of
the pieces in this work, thus emphasizing their

*These fragments were analyzed for this study and

Figure 18 Pevsner, Portrait of Marcel Duchamp, found to be cellulose nitrate. It is not known what
detail ofplastic in upper right portion oJ'sculpture piece(s) they are from or if they are fiom the original
(Photo taken in I99I). 1926 pieces or the last 1957 restoration by the artist.

178
planes and curves. In its present condition, as
the detail shown in Figure 20 indicates, both
transparency and form have been seriously
compromised with the deterioration of the plas-
tics. (For an earlier photograph, see Nash and
Merkert 1985.215.)

Figure 2l Fragments from Naum Gabo's

*ff
Construction in Space: Two Cones (Photo taken
in 1991).

the 1927 work), which he had in his studio and

which were still in good condition (Siegl and
d'Hamoncourt 1968). These were analyzed and
all were found to be cellulose acetate (Miklas
Figure 20 Naum Gabo, Construction in Space: 1968). As described by Pullen and Heuman
Two Cones, detail (Photo tqken in 1991). (1988) though, when Gabo used cellulose ace-
tate in 1968 to make a replica of this work for
For the most part, since the work was acquired the Tate Gallery, rapid deterioration again oc-
in 1952 it has been kept in a sealed vitrine. Seri- curred. Unfortr-rnately, as had been the case
ous physical damage occurred after the sculp- for the Philadelphia Museum of Art work, the
ture was removed for study from its airtight Tate Gallery version had also been stored and
vitrine in 1960 when condensation was noticed exhibited in a closed virrine.
on the glass (Siegl 1966). Upon opening the
case "a strong odor escaped from the case, then During an interview in 1953, Gabo was asked
cracks developed in some parts ofthe construc- if his constructions, which seemed so delicate,
tion, and on the following day the plastic base were permanent like traditional sculpture. Gabo
and the thick center stem were fractured to bits" replied, "There is nothing really permanent in
(Siegl 1966). Figure 21 shows the fragments of the world. To base a work of art on perrna-
the center stem. These have been identified as a nency is to base it on something which does
cellulosic material possibly a regenerated cellu- not exist" (Chanin 1953). Although some of
lose*. (Based on the appearance of the work, it the works described above attest to the consid-
seems highly likely that other types of plastics erably less permanent nature of the early plastic
were also used, but no further analysis has been sculptures, other works have survived surpris-
done yet.) After replacing some of the broken ingly well. It is clear that albeit the inherent
pieces the sculpture was put back in its vitrine instabiliry of the materials used, the environ-
because it seemed that there was little that ment plays a significant role in extending or
could be done. It continued to crack and sag drastically reducing the life of the works.
and in 1968 it was taken to Gabo's home in
Connecticut for possible repair. The artist sug-
gested that it was how the object had been kept *The fragments were sampled by Dr. Robert Feller
rather than the materials themselves that had in 1968 who found them to be "a generated cellulose
enhanced the deterioration, since he had early or methylcellulose." Dr. Feller expressed surprise at
works, including a model for this work, that these findings because "I had no idea that [these ma-
were in much better condition (Gabo 1968). terials] could be fabricated into sheets" (Feller 1968).
Gabo produced several pieces of old plastic The samples were reanalyzed in 1992 at the Philadel-
(believed by the artist to be as old as those in phia Museum of Art by Beth Price who agreed with
the 1968 findines.

119
Conservation Implications object with some structural integrity. Methods
for reinfroducing the lost plasticizers and stabi
The primary goal of conservation is to halt the lizers have also been investigated by Green and
deterioration without altering the appearance or Bradley. They attempted to reintroduce a plasti-
structural integrity of the object. However, sev- cizerby soaking the object in dibutyl phthalate.
eral studies have shown that with movie films The procedure was unsuccessful though be-
the only way to stop degradation in cellulose cause the cellulose nitrate was found not to
nitrate is to freeze the object (Allen et al. 1988). be porous enough to permit penetration of the
Since this is not practical, a solution must be liquid into the bulk of the object. While further
devised that at least minimizes the damage and work is being done to determine future heat-
extends the lifetime of the work. ment procedures, it is imperative that measures
be taken now to minimize further degradation
Environmental control is critical to the stability of the works. For example, our study has
of the work. Since temperature and humidity shown that the degradation is directly related to
fluctuations may cause stresses, particularly at the loss of camphor, therefore it is important to
the join lines, it is important to keep the fluctua- minimize its sublimation. One possible method
tions to a minimum. The deterioration proc- to do so is given below.
esses can be both thermally and photolytically
induced, so lowering the temperature of the The degradation processes become autocata-
storage or display room is advisable and pro- lytic in the presence of oxygen and water va-
tecting the piece flom ultraviolet light is por. Thus, attempts have been made to create a
crucial. closed environment for cellulose nitrate objects
by placing them in sealed containers. Unfortu-
One inherent cause ofdegradation is the acid nately, this has frequently resulted in serious
content ofthe object. Studies have looked into degradation of the sculptures. Considerable
the effects of removing the accumulated acids work needs to be carried out in designing dis-
play cases that can maintain a low relative hu-
by washing the object along with introducing
midity, provide protection fiom ultraviolet light
acid neutralizers into the polymer (Green and
and minimize the degradation processes. For ex-
Bradley 1988). Water, however, can have an
ample, such a case might contain vapor-phase
extremely deleterious effect on the structural in-
camphor to prevent further camphor sublima-
tegrity ofdegraded cellulose nitrate and its use tion and consequent depletion of this stabilizing
should be avoided on such sculptures. (Various plasticizer. The case might also include an air
other solvents, e.g., acetone, may also affect the circulation system that incorporates acid scav-
object's appearance as well as increase the deg- engers and/or nitrogen oxide specific absorbers.
radation rate.) Although the results have not yet Each case should be designed specifically to
been promising, one suggested method worth prevent or counteract the sculpture's deteriora-
further sfudy has been the vapor phase addition tion problems and to this extent one needs to
of ammonia to the object (Sirkis 1982). But take into account the work's material composi-
caution must be taken not to induce any base- tion and condition. Until such display cases can
cataly zed degradation of the cellulose nitrate. be constructed, it has been recommended that it
is best to place the sculpture in an area with ade-
Studies have also been undertaken to devise quate ventilation to prevent the nitrogen oxides
methods for dealing with the effects of the dete- from accumulating. Low relative humidities are
rioration process. Preliminary attempts have also recommended, but studies need to be done
been made to consolidate cracked or powdery to determine the optimum value, since a rela-
cellulose nifrate with adhesives likewise based tive humidity that is too low may also cause
on cellulose nitrate, for example, HMG (Green problems. Further work is being carried out
and Bradley 1988). This type of treatment prob-
by the authors to try to determine general
environmental parameters. This work involves
ably would not be reversible, might change the
the detailed material identification of various
appearance of the object and would interfere
cellulose ester artifacts exhibiting serious degra-
with future analytical work, but it might be one
dation and the exposure ofthese artifacts to
of the few options available for providing the various environmental conditions.

180
Conclusion de trds bon d mtdiocre. Ces sculptures sont
constitudes de pidces d'ester de cellulose (nitrate
The cellulose nitrate sculptures examined in de cellulose) qui, thenniquement formt es, sont
this study exhibit various types and degrees of coll\es ensemble avec un adhesif 'lgalement d
degradation. These range from surface crazing base d'ester de cellulose. Les pidces translucides
to cracking, discoloration, liquid production on en nitrate de cellulose, qui dtaient probablement
the surface, and major structural damage. The incolores d I'origine, ont maintenant pris une co-
process of deterioration was studied using loration qui tire d divers degr,ls sur le brun, tan-
FTIR spectroscopy, light microscopy, X-ray dis que les pidces opaques, blanches ou noires d
diffraction and SEM-EDS. The authors found I'oigine, n'ont vraisemblablement plus le m€me
that the various states of desradation relate to ton. De fines craquelures se sont formdes sur cer-
three factors: taines pidces, alors que d'autres prisentent, sur
leur sudace, des gouttelettes de liquidevariant
. past storage and display conditions entre le clair et le brun clair. Bien souvent, des
o composition of additives and impurities produits de conosion se sont d,!pos,!s aux points
within the piece oil le nitrate de cellulose se trouve d proximitd
o physical construction of the work d'tltments mdtalliques. Le processus de dbgrada-
tion de ces @uvres et d'objets provenant d'autres
At present there are no established treatments collections a ete Atudie en ayant recours d la
for reversing some ofthe degradation processes spectrom<itrie infrarouge d transform,le de
that these objects have undergone. Conserva- Fourier (IRTF), d la dffiaction des rayons X, d
tion treatment of the objects will try to mini- la microscopie 2lectronique d balayage associde
mize further degradation by direct control of d la spectrometrie de rayons X et d I'imagerie
the environment. It is imperative that the tem- d lectronique de retrodffision (microscopie 6lec-

perature, relative humidity and ultraviolet light tronique d balayage-dffiaction t lectronique de s

exposure be controlled. This may eventually rayons X) et d la microscopie photonique. Il est
require building specifically designed display fait itat, dans la prtsente communication, des
cases. mtcanismes qui ont pu contribuer d la dtgrada-
tion de ces objets et du r6le que les charges et
plastifiants auraient jout d cet 6gard. On y
Acknowledgement
ffictue enfin une analyse des traitements de
conservation qui sont proposds.
The authors thank the following individuals
for their assistance: Kristen Hoerman of Yale
University Art Gallery, New Haven; Lee Aks References
of the Hirshhom Museum and Sculpture Gar-
den, Washington, D.C.;Andrew Lins, Melissa Aks, Lee, Conservation Report, Hirshhorn
Meighan and Beth Price of the Philadelphia Museum and Sculpture Garden, Smithsonian
Museum of Art; and Denick Pullen of the Tate Institution, 1988.
Gallery, London. Special thanks are given to
Aks, Lee, Reply to questionnaire sent by the
Patricia Houlihan of The Museum of Modem
authors to Mr. Aks, 1991.
Art. New York. for her numerous contributions
in studying MoMA's sculptures. The authors Allen, N.S., M. Edge, C.V. Horie, T.S. Jewitt
also thank Keith Lauer of Massachusetts for his and J.H. Appleyard, "The Nature of the Degra-
generous contribution of cellulose ester objects
dation of Archival Cellulose-ester Base Motion
to be used for future experimentation. Picture Frkn," Journal of Photographic
Science, 36, 1988, pp. 34-39.
R6sum6
Chanin, A.L., "Gabo Makes a Construction,"
La digradation de sculptures en nitrate de Artnews, vol. 52,November 1953, pp. 3446.
cellulose exicuties par Gabo et Pevsner
Erhardt. David et al.. Smithsonian CAL
Trois sculptures de la collection du Museum of Analyical Report # 4lll, Conservation
Modern Art de New York extcuties par Naum Analytical Laboratories, Smithsonian
Gabo et par Antoine Pevsner presentent dif- Institution, 1983.
flrents degris de degradation, leur ttat variant

181
Feller, Robert to Theodor Siegl, April 5,1968, Pullen, Derek and Jackie Heuman, "Cellulose
Philadelphia Museum of Art Conservation Acetate Deterioration in the Sculptures of
Files, fletter]. Naum Gabo," Preprints of Contributions to the
Modern Organic Materia ls Meeting. Edin-
Gabo, Naum to Ted Siegl, April 7,1968, burgh, Scottish Society for Conservation and
Philadelohia Museum of Art Conservation Restoration, 1988, pp. 57-66.
Files, fletter].
Rankin, Elizabeth, "A Betrayal of Material:
Green, Loma and Susan Bradley, "An Problems of Conservation in the Constructivist
Investigation into the Deterioration and Sculpture of Naum Gabo and Antoine
Stabilization of Nitrocellulose in Museum Pevsner," Leonardo, vol. 3 1, no. 3, 1988,
Collections," Preprints of Contributions to pp.285-290.
the Modern Organic Mqterials Meeting,
Edinburgh, Scottish Society for Conservation Read, Herbert (introduction), Ruth Olson and
and Restoration, 1988, pp. 8l-95. Abraham Chanin (text)" Constructivism: the
Art of Naum Gabo and Antoine Pevsner Q{ew
Herbert, Robert et al., eds., The Soci4tt York: The Museum of Modem Art, 1948).
Anonyme and the Dreier Bequest at Yale
Universily: A Catalogue Raisonne Qiew Reilly, J.A., "Celluloid Objects: Their Chemis-
Haven: Yale University Press, 1984). try and Preservation," Journal of the American
Institute for Conserttation, vol. 30, no.2,1991,
JCPDS Powder Dffiaction FiIe,JCPDS pp. 145-162.
Intemational Center for Diffraction Data,
Swarthmore, Pennsylvania, I 980. Richards, Monawee to Dorothy Dudley,
December 16,1947, The Museum of Modern
Kuh, Katherine, The Artist's Voice: Tall<s with Art, Painting and Sculpture Files, letterl.
I7 Artists (New York: Harper and Row, 1962).
Selwitz. Charles. Cellulose Nitrate in
Lassaw, Abram and Ilya Bolotowsky, "Russia Conservation (Los Angeles: The J. Paul Getty
and Constructivism," in: Gabo : Consttactions, Trust,l988).
Sculpture, Paintings, Drawings, Engravin gs,
trans. Lund Humphries (Massachusetts: Siegl, Theodore, Bulletin of the Philadelphia
Harvard University Press, 1957). Museum of Art, vol. 62, no. 291, 1966,p. 1 53.

Lodder Christina, Russian Constructivism Siegl, Theodore, Memo, April 7, 1968, Phila-
(New Haven and London: Yale University delphia Museum of Art Conservation Files.
Press, 1983).
Siegl, T. and A. d'Hamoncourt, Memo, January
Miklas, Helen, Analyical Report from Rohm 25,1968, Philadelphia Museum of Art Conser-
and Haas, Bristol, Pennsylvania, 1968. vation Files.

Nash, Steven and Jom Merkert, Naum Gabo: Sirkis, Linda, "The History, Deterioration and
Sixty Years of Constructivlsz (New York: Conservation of Cellulose Nitrate and Other
Prestel-Verlag, I 985). Early Plastic Objects," Master's Thesis,
Institute of Archaeology, London, 1982.
Petryn, Andrew, personal communication,
1991. Yarsley, V.E., W. Flavell and N.G. Perkins,
Cellulose Plastics, Plastics Institute, London,
Pevsner, Alexi, I Biographical Sketch of Mv 1964.
Two Brothers, Naum Gabo and Antoine
Pevsner (Amsterdam: Augustin and Zimmer, F itz, Nifro cel lu lo s e E s t er Lacquer s :
Schoonman. 1964). Their Composition, Application and Uses
(London: Chapman & Hall Ltd., 1934).

t82
A.W. McCurdy's Developing Tank:
Degradation of an Early Plastic

Robert D. Stevenson

Historic Resource Conservation Laboratory

Canadian Parks Service
Dartmouth, Nova Scotia
- Atlantic Region
Canada

Abstract The tank was designed in the late 1800s by

A.W. McCurdy. Until I saw this object I knew
A protoQpe of a daylightfilm developing tank, of A.W. McCurdy only as the father of J.A.D.
of the collection of the Alexander Graham Bell McCurdy, who piloted the first heavier-than-air
National Historic Park in Baddeck, Nova Sco- flight in Canada. But A.W. McCurdy was for
tia, is presented as an example of the problems many years secretary to Alexander Graham
associated with early plastics. Bell. Mabel Bell, Alexander's wife, was a great
supporter of A.W. McCurdy and invested in his
The tankwas signed and dated by its maker (E. company while he was perfecting the tank,
Schneider, Nov. 1899), andwas presented to even borrowing to do so. No doubt this is why
Mrs. Mabel Bell by A.W. McCurdy, its de- the prototype tank was presented to her on No-
signer and Alexander Graham Bell's secretary vember 25,1899. (Details of the presentation,
for many years (and the father of J.A.D. including the date, are inscribed on the back of
McCurdy, the pilot of thefirst heavier-than-air the tank, as shown in Fizure l.)
flight in Canada).
The tank was discovered in its display case
covered wilh droplets of a liquid, which, upon . :
analysis, turned out to be sulphuric acid. The
-.":
dfficulties encountered during conservation
are presented os a c(tse study. Reference is
made to techniques used in the conservation of
two Marconi artifacts that contain similar
plastics.

Introduction and History

This paper is about the plastic-bodied prototype
of a daylight film developing tank. This tank
is of interest because of its association with
famous people, its precise dating and its Figure 1 The developingtank as it appears today.
spectacular deterioration. Photo: Canadian Parks Semice.

183
The tank is also signed and dated "E. Schneider, tank to George Eastman (of Kodak fame), who
Nov. 1899" on the inner bevel of the window declared it not practical; made further modifica-
area of the removable top. The signature is that tions; and eventually sold the patent rights to
of the tank's maker I presume. This dating Eastman for a small fortune.
gives the object unique value as a record ofthe
state of the art of plastics work during that Observation and Discovery
period.
The material of the tank bears a striking resem-
At the Alexander Graham Bell National His- blance to the materials found in two Marconi ar-
toric Park in Baddeck, Nova Scotia, there is a tifacts at our laboratory, parts of which were
large and wonderful collection of papers relat- identified as hard rubber (ebonite). Testing by
ing to Bell's work and life, amongst which are Fourier Transform Infra-red Spectroscopy
letters to him from his wife. In her letters from (FTIR) has determined that the body and lid of
the spring, summer and fall of 1899, Mabel the tank are also ebonite.
Bell often mentions A.W. McCurdy's work on
the developing tank. To my intense frustration This developing tank is a beautifully made ex-
the last letter in the files to mention this work is ample of the plastic technology of the day, but
dated five days before McCurdy presented the it was in poor condition when it frst came to
prototype tank to her. And I cannot find men- my attention in the summer of 1988. There
tion of such a presentation elsewhere in the were droplets of a liquid distributed on all ex-
frles. posed surfaces ofthe artifact(Figure 2). The
artifact was inside an unsealed display case con-
The only mention of the tank material in Mabel structed from wood, plywood and glass. As
Bell's letters would lead one to believe it is soon as I opened the case and smelled the acrid
made of hard rubber (ebonite). She refers sev- odour of acid I knew there was a problem! The
eral times to Mr. McCurdy's search for hard liquid on the artifact's surface was brown and
rubber for producing the tank: sticky, did not feel greasy and, strangely, had
no apparent odour when isolated from the
from a letter dated May 20, 1899: artifact.
Mr. McCurdy found hard rubber boxes
the size he will require for sale at a dollar
and a quarter, the roll ofcelluloid [pre-
sumably for the film type he was work-
ing on] would cost only a few cents, ...

from a letter dated May 29,1899:

Meantime Mr. McCurdy is investigating
the best and cheapest material for the
box. Hard rubber will do, but it is expen-
sive, and he thinks something else might
be better.

from a letter dated June 2, 1899: Figure 2 The top ofthe tank showing the droplets of
He thinks that he must frst have the box liquid on its surface as found on site.
made of hard rubber and have the whole
thing complete as a practical working ma- I wrapped the artifact in acid-free tissue and
chine though rough and capable of fur- transported it to our regional conservation
-
ther improvement before offering it to laboratory. The liquid left a brownish stain on
capitalists. the tissue. This started me thinking that a layer
of colour on the artifact was being washed
During this same period A.W. McCurdy pat off by the liquid. To ascertain what was in the
ented various aspects of the design; showed the liquid, I briefly put the artifact which was
-
184
dryr.tg out in a sealed chamber to raise the l) wiping with an alcohol-saturated, soft
RH. I then -removed a few drops of the liquid cotton pad
by pipette and sent them offfor analysis. I result streaks of whitish-yellow colour
dipped a pH test strip in a droplet and got a 2) wiping with a water-saturated, soft
rather startling reading ofpH 0.5- the artifact cotton pad
should have been smoking! The result was result: less streaking, pale whitish colour
confirmed when the liquid was identified as overall
sulphuric acid. 3) wiping with an acetone-saturated, soft
cotton pad
I removed the developing tank from the cham- resull faint streaks, surface dull
ber and let it acclimate to the ambient RH. I 4) wiping with a small amount of alcohol on
made one brief and regrettable attempt to wipe
a soft cotton pad
off the residue of liquid. Having tested the
result: no more streaks, surface dull
bottom of the artifact and noticed no problem,
5) wiping with a dry, soft cotton pad
I then wiped one side of it. This left streaks,
result: surface less dull (A dragging sensa-
and I stopped the attempt (Figure 3).
tion was noted with this step, perhaps due
to the slightly rough nature of the tank's
surface.)

The cotton pads used in the initial alcohol swab-

bing picked up quite a lot of yellowish material,
which is presumably composed of the plastic
and its breakdown products. Lesser amounts
were collected by the acetone pad and water
pad, though perhaps this is due to the sequence
of operations rather than to the differing solu-
bilities of the various solvents. It would not be
wise to prolong this process as obviously it re-
moves material from the surface.
Figure 3 Streaks caused by the initial wiping oJ
the tank.
One point that should be noted about this proce-
dure is that it was not possible to predict the sol-
A Conservation Dilemma vents required or the order in which they were
used. The method I used on the tank was the
Why did I get streaking on the side but not on
same as that used on the Marconi artifacts, but
the bottom? This remained a mystery until the
the sequence I followed was different. It was
following year when a conservator at our labo-
noticed (while working on the Marconi arti-
ratory worked on two artifacts for the Marconi
facts) that different pieces of the same artifact
exhibit at Glace Bay, Nova Scotia. These arti-
required different treatments. Some parts re-
facts contained hard rubber (ebonite) parts that
quired only alcohol, for example, while others
had similar dull brown surfaces and showed
seemed to be unaffected by any ofthe solvents.
shiny black where they were not exposed to
Working on such artifacts makes for an
light, just like the developing tank. To remove
exciting day!
dirt and stains from these ebonite parts a
method of swabbing the surface was developed
When I first saw the artifact covered in droplets
using solvents such as acetone, ethanol, water
I was very concerned about the metal parts.
and Varsol.
However, they showed no apparent corrosion at
the time, even though they were surrounded by
I used a modified version of this method to re-
sulphuric acid! I am sorry to report that in the
move streaks I had caused on the developing
three years the artifact has been at the labora-
tank, following the sequence below (which
tory, corrosion has started (Figure 4). The metal
only took a few minutes):

185
 As well, there is a roll of clear plastic (also
probablj cellulose nitrate) inside the tank. In
sey.eral of the historical papers, both in the labo-
ratory notes and in Mabel Bell's letters, refer-
ence is made to A.W. McCurdy trying to invent
a new way to hold a roll of film so that it could
be processed without the adjacent layers touch-
ing each other. In one laboratory note he de-
scribes and draws a thick edging (probably
leather) sewn onto the film. This is exactly the
appearance ofthe roll in the tank at present, as
pictured in Figure 6. McCurdy considered this
Figure 4 Corrosion beginning on the metal pilIs of idea to be more valuable than the developing
the tank. All metal parts were similarly afected. tank, but it did not tum out that way.

parts have not been treated yet, but I plan to

remove and treat them and insert Mylar wash-
ers or some other protective barrier under them
before they are reinstalled. Yet this is not a
wholly satisfactory solution because the metal
parts all screw into the plastic, making direct
contact unavoidable.

Another problem area involves a sandwich of at

least two layers of red-coloured plastic (prob-
ably cellulose nitrate) in the centre of the lid.
This sandwich is covered by two hinged panels
of ebonite and may have been for viewing the
film development process without exposing the
film to damaging levels of light (Figure 5).
There is a yellow powdery accretion growing
between the layers.
Figure 6 A film separator invented by McCurdy,
which has remained in the tank throughout its life, so
./itr as is known. Photo: Canadian Parks Service.

Although these parts are deteriorating, I have

made no attempt to treat them. Stabilization of
the storage/display environment as noted below
should be sufficient for the roll of film. How-
ever, it is not possible to treat the plastic sand-
wich because it cannot be removed from the
tank's lid without causing further damage.

Plans are being made to create a refrigerated

storageidisplay environment for the tank that
will maintain cool, dry conditions using silica
Figure 5 A front view of the tank shows the wording gel to buffer the RH and activated charcoal to
EBEDEC. This is McCurdy's brand name, which is protect against the various vaporous breakdown
Greekfor "out ofBaddeck." Photo: Canadian Parks products. It may be necessary to create an oxy-
Semice. gen-free atmosphere as well.

186
For a more detailed discussion of the deteriora- Bacon. Louise. "The Deterioration of Four
tion of ebonite see the summary of the paper by Giorgi Flutes Made of Ebonite and a Possible
Louise Bacon listed in the Bibliography. Method for Their Conservation," ed. Victoria
Todd, Preprints for the UKIC 30th Anniver-
Acknowledgement sary Conference, 1988,pp. 96-1 00.

The author thanks scientists at the Historic Describes how the surface ofebonite un-
Resource Conservation Laboratory, Canadian dergoes severe decomposition through
Parks Service Headquarters, Ottawa, for their the action of light, air and moisture. The
analysis of the swface liquid and for the FTIR surface acid formed by this process can
work that identified the plastics. be wiped away, but in extreme cases neu-
tralization may be necessary. This can be
R6sum6 done with a dilute solution of sodium bi-
carbonate, rinsing and quick drying in in-
Lu cuve de developpement d'A.*11. McCurdy : dushial methylated sprit. For storage
un exemple de la digradation des premiers and display, light levels should be re-
plasfiques
duced to a minimum and UV eliminated.
RH must be kept low. If possible, oxygen
Un prototype de cuve de diveloppement pour
pellicule exposde d la lumidre du jour, qui fait should also be eliminated.
partie de la collection du parc historique national
Alexander Graham Bell, d Baddeck (Nouvelle- Brydson, J.A., Plastics Materials,
Ecosse), est prdsentt pour illustrer les probldmes ISBN 0 408 00142 9, (London/Boston:
que posent les premiers plastiques. Newnes-Butterworths, I 975).

Signde et datte par lefabricant (E. Schneider, Green, H. Gordon, The Silver Darl (Frederic-
Nov. 1899), la cuve a 6td prdsentte d It(" Mabel ton, New Brunswick: Brunswick Press, 1959).
Bell par son concepteur, A.IU. McCurdy. Se-
crttaire d'Alexander Graham Bell pendant de A book about the first heavier-than-air
nombreuses anndes, ce dernier est en outre le flight in Canada, and the people involved
pire de John A. D. McCurdv, qui sera le premier in the event. These people include the in-
d piloter un < plus lourd que l'air > au Canada.
ventor, Alexander Graham Bell, the pilot,
J.A.D. McCurdy, and his father, A.W.
On a decouvert la cuve, dans sa vitrine, couverte
de gouttelettes des gouttelettes d'acide sulfu- McCurdy.
-
rique, comme I'ont r6,tlt des analyses ultdrieu-
res. Nous prisentons les probldmes rencontrds Sparke, Penny, ed.,The Plastics Age From
sous la forme d'une ,ltude de cas et faisons 6tat Modernity to Post Moden?iql, ISBN -
des techniques qui ont 6tt utilisdes au cours de la | 85177066 6, (London: Victoria and Albert
restauration de deta objets de Marconi contenant Museum, 1990).
des plastiques semblables.
Essays that examine the history of plastic
Bibliography products from the late 19th century up to
the present within the changing cultural
Adams, John V., Plastics Arts Crafts,ISBN climate of the period.
08492a1 5, (TorontoAJew York/London: Van
Nostrand, 1948).

A briefhistory ofplastics technology and

usage.

187
Degradation of Polyurethanes in
20th-Century Museum Textiles

Nancy Kerr and Jane Batcheller

Department of Clothing and Textiles

University of Alberta
Edmonton, Alberta
Canada

Abstract was commercially dry-cleaned. Saturation with

dry-cleaning solvent (perchloroethylene) and the
The term "polyurethane" applies to a diverse mechanical agitation ofthe cleaning cycle caused
group ofmaterials used in 20th-century textiles. the oxidized polyurethane component to break
They may be fibres (spandex), adhesives or con- down. Suggestions will be givenfor the safe
solidants found in synthetic suedes, foams in cleaning of polyurethane-containing garments.
bonded/laminated fabrics, surface foams of
imitation suedes and water-repellant coatings or Introduction
leather-look surfaces on somefabics. There is
muchfor conservators to learn about polyure- Polyurethanes are a very diverse group ofpoly-
thanes, including the variety ofways they are mers with many textile applications. The name
used in textiles, how to assess the condition of polyrethane comes from the periodic urethane
the material and what form degradation will take. links (NHCOO-) in the polymer "backbone,"
This presentation will address these factors. The however, this name may be misleading because
mode of degradation of polyarethanes will be dis- polymers are prepared from polyester, poly-
cussed, based on a reviqu ofthe extensive litera-
ether or other short-chain polymers, and con-
ture on this topic.
tain a variety of other links as well. Because
polyurethanes do not always form the exposed
Photo-oxidation is parricularly damaging to
surface ofa fabric or garment, curators and con-
polyurethanes and is the most common cause of
damage in the textiles we have examined. Poly- servators may not realize that many 2Oth-cen-
urethanes based on aliphatic diisocyanates and hry textile collections contain polyurethanes.
polyester soft segments are more resistant to Polyurethanes appear in textiles primarily as
p h ot o - oxidati o n t h an thos e fo rm e d fro m aro mati c elastomeric hbres and yarns (spandex), and as
diisocyanates and polyether soft segments. Micro- films, coatings and foams supported by tradi-
organisms can grow onpolyurethanes in such tional fabric structures. These materials are sen-
forms as crackedfilms. As growth proceeds, sitive to a variety of degrading influences and
enzwes are secreted which can cause random require care in storage, display and cleaning if
chain cleavage at the polyurethane linkages or they are to be preserved.
scission at chain ends.
Polyurethanes have been produced for commer-
A number of garments containing polyurethane cial use since the early 1950s. Research on
as fibers, foams, films or adhesives will be shown, these polymers began before the Second World
and the damage they exhibit will be discussed. War. While Wallace Carothers was developing
In many of these garments, no evidence of photo-
polyamides in the United States in the 1930s,
oxidative damage was visible until the garment

189
 Otto Bayer and his associates at Farbenfabriken is possible to make many different polyure-
were experimenting with the polymerization of thanes with a variety of physical and mechani-
isocyanates and produced linear polyurethanes cal properties (Figure l).
in 1938 (Hepbum 1982). The commercial pro-
duction of polyester-based polyurethane foam
began in the 1950s and was followed by the Rigid

worldwide production of rigid and flexible

foams i
I

foams, coatings, moldings and adhesives Semr-rrgrd foams

I
Flexible Surface-coatings
(Backus 197 7, 642-643). Elastomeric fibres, foams I
intended to replace rubber, were commercially I

available as spandex fibres around 1960 Cast Increasng

elastomers degree of
(Joseph 1986,142). Museum textiles collected branchrng or
since the early 1960s may include polyurethane Poromencs crosslinkrng
in the form of spandex elastomers in founda- textile coatrngs

tion garments, foam layers in foam-bonded and Spandex Ftlms

laminated fabrics, foam paddings in tailored frbres
garments, water-repellent or waterproof coat-
ings on textiles, coatings on synthetic leathers,
the foam layer in artifrcial suedes, and as Millable
poromerics in shoe uppers. The most recent use elastomers
of polywethane is as a microporous breathable Increasing chain stiftness, Interchain
coating on rainwear. attraction, crystalInity

The objectives ofthis paper are (l) to describe I

Figure Structure/property relationships in
the formation and textile-related uses of poly- polyurethanes (Hepburn 1982). Used with permis-
urethanes, (2) to review the literature on the sion of Elsevier Applied Science Publishers.
degradation of polyurethanes and present exam-
ples of degradation in polyurethane-containing The usual way to prepare polyurethanes is in
garments and (3) to describe the degradation in two steps (Figure 2). The starting material is
four new polyurethane fabrics, which have most often a short polymer of polyester or
been exposed to sunlight through glass for 63 polyether with hydroxyl groups at each end
days. The outdoor exposures were conducted in (polyol), although other polymers, such as poly-
order to determine whether sunlight (through caprolactone, may be used (Carter 197 l, 17 4).
glass) would cause damage similar to that seen The polyester or polyether polyol is reacted
often in polyurethane-containing garments ana- with an excess of a diisocyanate to form a pre-
lyzedby the Textile Analysis Service.* Four polymer (molecular weight approx. 15,000 to
new types of polywethane fabric were chosen 20,000). Two of the most frequently used diiso-
for testing because little information is avail- cyanates (Figure 3) are 4,4'-diphenylmethane
able on their structure, physical or mechanical diisocyanate (MDD and a mixture of 2,4- and,
properties. These fabrics included a very realis- 2,6-toluene diisocyanate (TDI) (Backus 1977,
tic synthetic leather and suede and two polyure- 649). The prepolymer is a low melting point
thane-coated fabrics. Suggestions for the solid or liquid unsuitable for most applications,
conservation of textiles containing polyure- thus, the second step involves linking the prepo-
thanes will also be given. lymers into long-chain polymers by reaction
with a chain extender, usually a diol or diamine
Formation of Polyurethanes (Lenz 1967,191). When the chain extender is a

A number of authors have discussed in detail

the formation of polyurethane polymers *The Textile Analysis Service was set up in the
(Hepburn I 982; Backus 1977, 642; Lenz 1967, Department of Clothing and Textiles in 1970. Its
180; Rosthauser andNachtkamp 1987, 121). purpose is to analyze and resolve problems encoun-
Perhaps no other polymer offers the versatility tered during the use and care ofgarments, household
of polyurethane. Through molecular tailoring it and industrial textiles, furs and leathers.

190
diol, the polyurethane polymer contains we- POLYOLS

thane links; if it is an amine, urea links are polyelhyl.n. orldo (polyelh.r) polypropylene oxlde (polyether)

formed (Figures 2 and3).If carbon dioxide is

uo-[cH.-cu,-o]-oH Ho-[3]:"". o]-o"
released during the reaction, a polyurethane nn
foam is made. Polyfunctional starting materials poly6lhylen. adlpale (polyest€r)
will produce polymers with branching or oo
no - cr.r" - cr.r, -[o
crosslinking. These polymers are more rigid - d
-cr.cHrcHrcH. -b-o-cH"-cH.]-ox
than rmcrosslinked polymers and are well
suited for rigid foams and surface coatings (Fig- I SOCYANATES

dllsocyanato (MDl)
ure l). At this point, the polymers are either ex-
truded as fibres, a continuous film or molded. o=c=N-lgtscHr19FN=c=o
Elastomeric textile fibres are formed by wet
spinning into a coagulation bath, dry spinning 2,4-loluene dllsocyanale and 2,6-tolu6n€ dllsocyanate (TDl)

after dissolving in a solvent, or melt spinning CHr CHr

rA.r="-o * o="-*tA-r-"-o
(Carter 1971,177).
YvN=C,O

CHAIN EXTENDERS
Ho.A A/WVVWV\ /1 o H

Poly..t.r or
Poly.lhrr Polydar
cl cl
't
ocN:.Nco
dlsoqanb
I
I
ocN--Nco
diisq.n.r.
"."@-"r,-@"n,
I

Y
H;:f::' Figure 3 Starting materials frequently used
**:-X-E*E-[-^* Pr.polymor
in the production ofpolyurethanes (Hepburn 1982).
Used with permission of Elsevier Applied Science
I
| + HrN-R-Nrh * I
Publishers.
I d.anoe chain "O"O",",O
diolor wal., I

@a hb b,{s u€rhans rin\s

i i
segments or rigid domains in an elastomer
"lEX-- [-E-","u'u*"-3 [ X-E-."-* consist ofurethane, urea and other links
'l
Polyurethane Polymor
- formed from the diisocyanate and chain exten-
der. They are held to neighbouring molecules
Figure 2 Two-stepformation of polyarethane by H-bonding, dipole-dipole interactions or
polymers. A polyester or polyether polyol reacts first Van der Waal's forces. In effect, the hard seg-
with a diisocyanate, then with a chain extender ments act as tie points or reinforcing "fillers" in
(Hepburn 1982). Used with permission of Elsevier
the soft matrix. They affect properties, such as
Applied Science Publishers.
response to high temperatures and strength
(Hepburn 1982,69; Carter 1977, 174).
When polyurethane is produced for elastomeric
fibres it has a unique molecular structure that
confers high elongation under stress and excel- In recent years, water-bome polyurethanes or
lent elastic recovery. Hepbum (1982) gives aqueous dispersions ofpolyurethanes have be-
structural formulae for a variety of polyesters come increasingly important because they can
be made into adhesives and coatings with little
and polyethen, isocyanates and chain extenders
used in making elastomers. The polyester or
or no co-solvent (Rosthauser and Nachtkamp
polyether oligomer provides a long flexible 1987,121). This has reduced solvent emissions
amorphous "soft" segment that stretches easily and lowered production costs. The films do not
under stress. The nature of the soft segments af- harm plastic surfaces and adhere well to many
fects response to low temperatures, hardness, surfaces. Water serves as the chain extender,
tear strength and modulus. Hepbum (1982,73) reacting with isocyanate groups on the prepo-
states, "Polyethers generally give elastomers
lymer and forming urethane links. Hydrophilic
having a lower level of physical properties than groups, ionic groups (SOr-Na ) or water-
the polyester-based materials due to the weaker soluble segments may be incorporated into
inter-chain attractive forces." The "hard" the main chains.

191
Textile Uses of Polyurethanes
When the word pollurethane is mentioned,
many people imagine foam materials or
injection-molded plastics, however, museum
collections have a great variety ofother arti-
facts containing polyurethanes. Collections of
20th-century textiles include garments contain-
ing spandex or elastane fibres (generic names
in North America and Europe respectively)
(Textile Institute 1975). These spandex or
elastane f,rbres made of either polyether- or
polyester-based polyurethanes appeared as a re-
placement for rubber yarns in 1962 n women's
foundation garments and bathing suits. Figure 4 Imitation suedefabric (Supersuede) con-
sists of/ine (4 pm) polyesterfibres embedded in a
In another form, polyurethane was used in the polyurethane foam base.
1960s for foam-bonded and laminated fabrics
(Mansfield 1985). In bonded fabrics, a thin
layer (l mm to 3 mm) of polyurethane foam
was used to hold together a face fabric and
backing fabric, usually a tricot knit. The foam
was flame-heated on both sides to melt it and
bond it to the two exterior fabrics; altematively,
an adhesive was used with the foam as a bond-
ing agent. Fabrics laminated with foam were
similar to foam-bonded fabrics. however. no
backing fabric was used. The face fabric was
simply adhered by flame or adhesive to a thick
Iayer (3 mm to 5 mm) of foam. A very light
weight, sometimes inferior fabric, was thus
given body and stability so that it would be suit-
able for dress-wear, coatings and upholstery.
Figure 5 Polyurethane foam, whichforms the base
In the late 1970s and early 1980s polyurethane of Supersuede.
foams were used in the production of imitation
suede fabrics. In the simplest fabric structures, Belleseime, do not have a woven or knit fabric
a warp knit fabric is coated on one surface with base. They are composed ofa random web of
a fine polyurethane foam to give a soft suede- extremely fine polyester or nylon bi-constituent
like fabric (e.g., Suede 2l by Kanebo, Japan). If fibres (60%), within a polyurethane foam
the foam is extremely fine grained, it produces matrix (40%) (Figures 4 and 5).
a "handle" or "feel" very much like suede. Very
fme polyester fibres may be embedded in the One of the most important uses for polyure-
foam, or the fibers from the base fabric may thanes is for coating textiles to produce novelty
be brushed or raised and incorporated into the fabrics, such as synthetic leather, or to produce
foam to give a fibrous surface texture. The fabrics with improved water or wind resistance.
foam/fibre mixture may cover both surfaces of The coating may be applied to the face or the
a woven or knitted fabric base, producing a re- reverse side of the fabric. It may be very thin
versible material with the appearance of suede (Figure 6) or a rather thick (l mm) foamed
on each surface. (e.g., Cavsuede by Collins and layer with an embossed surface resembling
Aikman or Ultrana). The most expensive and leather (Figure 7). In the transfer coating proc-
possibly the most realistic synthetic suede ess, which is used to produce many coated
fabrics, Ultrasuede, Supersuede, Clarino and fabrics and most leather-like materials. the

192
 Polyurethanes are also used as a frnish on real
leather. A split leather, for example, can be
improved greatly by using an aqueous polyure-
thane-sizing treatment in the tanning process
(Hepburn 1982,l5l). A completely artificial
leatheq Corfam, was developed by Du Pont de
Nemours and called a poromeric. This material
was used for shoe and boot uppers. It consisted
of polymer-impregnated non-woven nylon fab-
ric, a layer of cellular structured polyrethane
foam and a tough outer skin layer ofpolyure-
thane. The polyurethane is either a polyether-
or polyester-based material (Webster 1985).
Figure 6 Arctic Armour is an example o/ a woven Polyurethane uppers on shoes or boots were
fabic undercoated with a 15 pm to 60 pm layer o/ not as breathable as manufacturers had adver-

\a
po lye s te r- b as e dp olyurethane. tised. Du Pont ceased production ofCorfam
poromerics, but they are still produced in
Poland under the trade name Polfam
(Backus 1977).
rxi;df,r.l *r
!t-So$ Degradation of Polyurethanes
Polyurethanes are degraded by exposure to
light, heat, and chemicals that cause hydrolysis,
and micro-organisms. Each of these agents will
be discussed and referenced to recent literature.
The weak points on a polyurethane molecule
are the various reactive bonds in the main poly-
mer chains, such as urethane, ester, ether, am-
ide, urea, and biuret. These bonds vary in their
Figure 7 A realistic synthetic leather (Ultraleather)
resistance to hydrolysis and oxidation.
is shown in cross-section. The I5 p.m polyurethane
skin with leather-like grain is denser than the
300 pm polyurethanefoam layer beneath it. A
Light: Some authors suggest that polyure-
ticot knit stabilizes thefabric. thanes have good resistance to photo-oxidation
caused by visible and ultraviolet (UV) radiation
polyurethane is first spread in a thin layer over (Iyer 1989; Hepburn 1982) and others indicate
an embossed or smooth silicone-release paper. that light resistance is a problem (Du Pont
This "skin" layer will eventually form the film 1976; Gadette and Lemaire 1984; 1+bu-Zeid
coating or leather-look surface ofthe fabric. A et al. 1984).
tie coat is applied, which laminates a backing
fabric to the skin. After drying and curing, the Photodegradation by UV and visible light pro-
silicone-release paper is removed to reveal the duces a variety ofbreakdown products (Rek et
leather-like swface (Durst 1985). al. 1988; Rinby and Rabek 1975). Photodegra-
dation is accelerated by the presence of mois-
The newest polyurethane-coated fabrics are ture and ammonia (R6nby and Rabek 1975).
advertised as waterproof or water-repellent and Residual catalysts, such as tertiary amines or di-
breathable (Mansfield 1985; Reagan 1990). valent tin, left from the polymerization of rigid
The "breathability" is achieved in a solid poly- polyurethane foams appear to accelerate degra-
urethane film by incorporating hydrophilic dation of the foam by UV light and heat (Abu-
groups in the polymer. The films may also be Zeid and Nofal 1986). Ions from metals, such
breathable because of a microporous strucfure as Cu, Al, andZn, form chelates with some
that allows water vapour but not liquid water to polyurethanes and sensitize the polymers to
diftuse through (Table I). UV light degradation (Ranby and Rabek 1975).

193
 vary in their sensitivity to heat. For many poly-
urethane elastomers, 80oC represents a maxi-
mum use temperature (Hepburn 1982, 351). At
70"C to 80"C, the tear strength of a polyure-
quinone-imide
thane elastomer may only be about 50% of its
room temperature value. Du Pont Bulletin L-86
(1976) includes the results of a test involving
H'. rH". 50 cycles of machine washing (60'C) and trin-

(
i?
Y c'.',
ble drying (82'C) on fabrics containing Lycra
spandex or rubber thread. After the wash/dry
v"* test, the Lycra spandex fabrics withstood
\,/
10,000 cycles ofrepetitive elongation to 50%
cHz
on a Zwick Flex Tester, whereas half the rubber
l specimens failed at 5,000 cycles. The Interna-
photo-Fries rearrangement tional Fabricare Institute recommends that poly-
urethane-coated fabrics, imitation leathers or
Figure 8 Yellowing of polyurethanes is related to a suedes and foam laminates be dried at a tem-
photo-Fies rean'angement and to theformation of a perature not exceeding 50'C in order to mini-
quinone-imide group .ftom the diis ocyanate, MDI.
mize heat damage (Oehlke 1980; Scalco 1989;
Rising and Scalco l99l). The temperature at
most sensitive to photo-oxidation. Light sensi- which appreciable degradation occurs varies
tivity depends on the structure ofthe isocyanate from about 180'C to 220"C (lyer 1989;
and the polyol used to form the polyurethane. Abu-Zeid et al. 1983; Grassie et al. 1980).
The most unstable confrguration is a polyure-
thane made llom an aromatic diisocyanate Barendregt and Van den Berg (1980) suggest
(especially MDI) and a polyether polyol that two successive reactions play a role in the
(Rosthauser and Nachtkamp 1987). Polyester aging ofpolyurethanes. Both reactions have
polyols used to form the soft segments of a approximately the same activation energies
polyurethane molecule and aliphatic diisocy- (102 kJ/mole and 106 kJ/mole), but the frst
anates are most resistant to UV light (Rek et al. reaction is the most important below 100.C.
1988). Rinby and Rabek (1915) report that They found that heat-induced aging was accel-
photo-oxidation depends on the diisocyanate erated by oxygen and moisture. Hepbum
used in forming the polymers. The stability (1982, 35 l) indicates that heat causes oxidative
of three common diisocyanates to UV light cleavage of ether and ester linkages in polyure-
follows the order: diphenylmethane diisocy- thanes, although the ester linkages are consid-
anate < toluene diisocyanate < hexamethylene ered more stable. Grassie and co-workers
diisocyanate. (1980) used thermal volatilization analysis
(TVA) and thermo-gravimetry (TG) to study
It is possible to stabilize polyurethanes to heat degradation mechanisms in a series of
light degradation by incorporating antioxidants. polyurethanes made from polyester soft seg-
Increased resistance of polyurethanes to weath- ments and methylene bis(4-phenyl isocyanate)
ering and radiation can be achieved by incorpo- (MBPD. They demonstrated that the threshold
rating benzotriazole UV absorbers and hindered degradation temperature increases as the con-
amine light stabilizers (HALS) (Gabriele et al. tent ofthe polyester soft segment increases. A
1983; Stohler and Berger 1988). Ifa light-sensi- proposed reaction sequence for the polyester-
tive aromatic diisocyanate must be used in for- based polyurethanes is presented. Degradation
mulating a polyurethane, it is preferable for the in rigid polyurethane foams has been noted at
isocyanate group (-NCO) to be separated from temperatures as low as 60oC (Abu-Zeid et al.
the benzene rings by one or more methylene 1983). Specimens were heated for l5 minutes
goups (Iyer 1989). at 60oC, 80"C, l00oC, 150'C, 200'C and
250'C and their photo-acoustic spectra were
Thermal: Because polyurethanes vary consid- analyzed. At 200'C and thereafter a major peak
erably in structure and composition, they also was noted and assigned to the quinone-imide

195
group. This is one group considered responsi- solvents, such as acetone, methanol and diethyl
ble for the yellowing or darkening of polyure- ether. The strength of polyurethane elastomers
thanes on aging. is lowered as solvents become more polar. Sol-
vents of very high polarity, such as dimethyl
Chemical: One of the limitations of certain formamide and metacresol, will generally dis-
polyurethanes is their ease of hydrolysis solve polyurethanes.
(Hepburn 1982, 345). In order to determine
whether a particular polyurethane is sensitive to Microbial: Although most synthetic polymeric
hydrolysis, it is necessary to know its chemical materials are not readily attacked by micro-
makeup. It has been well established, for exam- organisms, polyurethanes are unusual in their
ple, that polyester-based polyurethanes are susceptibility to microbial degradation, espe-
more easily hydrolyzed than polyether-based cially by fungi (Wales and Sagar 1991). As
pollurethanes (Du Pont 1976; Pavlova el al. referred to in Webster ( I 985) and Shuttleworth
1985; Rosthauser and Nachtkamp 1987; and Seal (1986), when a polyurethane film or
Chapman et al. 1990). The three main polyols foam has been damaged by microbial growth,
used in preparing elastomers are generally evidence of the damage will include the
ranked for hydrolytic stability as follows: following:
polyether > polycaprolactone > polyester
(Hepbum 1982,345). . loss of mass
. surface peeling or flaking
Pavlova et al. (1985) state that polyester-based . surface crazingor deep cracking
polyurethanes are hydrolyzedby alkalies that o staining of the cracks from the growth of
cause saponification of the ester group. When fungal hyphae
an ester link in the polymer backbone is bro- o matting of the interstices of foam
ken, new carboxyl and alcohol groups are o surface discolouration (bloom) from fungal
formed; the carboxyl group catalyzes further hy- growth beneath
drolysis. Polyether-based pollurethanes, on the o the formation of holes in films
other hand, are subject to acid hydrolysis ac-
cording to the accepted mechanism, and have Darby and Kaplan (1968) determined that
very good stability in alkaline solutions. The polyester-based polyurethanes are more easily
rate ofhydrolysis depends not only on the degraded by micro-organisms than polyether-
chemical composition of a polymer, but also based polyurethanes. In their study, 100
on the diffusion properties of the polymeric polyether- and polyester-based polyrrethanes
matrix. ln their search for polyurethanes with of known formulation were challenged with a
improved resistance to hydrolysis and thermo- mixed fungal inoculum (Aspergillus niger,
oxidation, Chapman et al. (1990) have devel- A. flavus, A. versicolor, Penicillium funiculo-
oped experimental polyurethanes that lack both sum, Pullalaria pullulans, Trichoderms sp. and
ester and ether links in the soft segments. They Chaetomium globosum). The polyester polyure-
contain non-self-associating polyamide blocks thanes supported excellent growth after three
and yield polyurethane films that are more hy- weeks whereas the polyether polyurethanes
drolytically stable than polyester polyurethanes showed only light to moderate attack. Enzymes
and less readily oxidized than polyether were able to attack only if the polyether or poly-
polyurethanes. ester segment had at least three -CH2 groups in
a straight carbon chain (Kaplan et al. 1968).
The resistance of polyurethanes to a wide They emphasized the need to know the exact
variety of reagents can be found in a number of formulation of a polyurethane in order to make
references (Hepburn 1982,355-356; Du Pont meaningfu I evaluations. Other researchers have
1976).ln general, when polyurethanes are new confirmed that polyester-based polyurethanes
and in good condition, they have excellent are most susceptible to fungal damage (Filip
resistance to solvents used for solvent scouring, 1978: Webster 1985: Shuttleworth and Seal
finishing and dry-cleaning, such as aliphatic, 1986). Shuttleworth and Seal (1986) found
aromatic and chlorinated hydrocarbons. They that the most extensive deterioration of
also have good resistance to mildly polar

196
polyester-based polyurethanes was caused by
Gliocladium roseum, although fungi belonging
to other genera (Alternaria, Fusarium, Penicil-
lium and Ulocladium) were able to produce
similar results over a longer period of time. A
major study of the mechanism of polyurethane
biodeterioration is being conducted by re-
searchers at the British Textile Technology
Group (Wales and Sagar, 1985, 1988,
and 1991).

Degradation in Garments Analyzed

by the Textile Analysis Service
Many garments containing polyurethanes
have been examined by the University of Al-
berta, Textile Analysis Service during the past
20 years. These include coated fabrics used in
rainwear, imitation suedes, imitation leathers
and foam-bonded or laminated fabrics. The
garments belong to individuals and are not
museum artifacts, but they exhibit typical signs
ofdegradation that may be present in polyure- SW
' 'r,l:'::
thane-containing garments within museum
collections. The garments examined in the labo-
)t*
ratory have developed problems within the first
few years of manufacture. As with many materi-
als, the degraded condition of the polyurethane Figure 9 Wen this imitation suede gatment was
may not be readily apparent or discemible by dry-cleaned, loss oJ' polyurethane foam occuned on
casual examination. the collar, shoulders, and cffi, areas subject to light
exposure and abrasion during use.
The most fiequent cause of deterioration in
these garments is light damage. It is most evi- In luxurious imitation suedes like Ultrasuede,
dent in imitation suede garments in which the Belleseime, and Supersuede, the pollurethane
surface is a layer of fine polyurethane foam. foam forms the base of the fabric in which the
These fabrics often exhibit loss of foam in ar- suede-like polyester fibres are embedded (Fig-
eas of the garment that received the most light ures 4 and 5). When the foam deteriorates fiom
exposure: the shoulders, outer sides ofthe exposure to light or oxides of nitrogen, the
sleeves, and top of the collar and lapels. When fabric shreds easily in areas ofstress, such as
dry-cleaned, the light-damaged foam crumbles across the shoulders, at the "armseye" seams,
and is lost from the fabric base, since it is not or buttonholes. These fabrics also deteriorate
able to withstand immersion in the cleaning at collar and cuffedges ffom abrasion during
solvent and abrasion from commercial dry- wear. In addition, the dry-cleaning process
cleaning cycles (Figure 9). causes fabrics to soften and lose body
(Figure 10).
On one occasion, the most severe loss of foam
occurred under the collar and lapels where light In the deterioration of imitation leather or poly-
had not degraded the polyurethane foam. It is urethane-coated fabrics used in rainwear, light
possible that residual catalyst or a breakdown damage is sometimes, but not always, evident.
product from the foam was trapped under the During dry-cleaning, the polyurethane film
collar and lapels and accelerated the degrada- may separate from the base fabric to which it
tion of the foam in this area. is bonded because of swellins of the film in

t97
perchloroethylene or softening of the adhesive
or tie layer holding the film to the base fabric.
The film balloons away from the backing fabric
and often tears in its weakened, swollen condi-
tion. During solvent extraction and drying, the
separated film may stick to itself or re-adhere
to the base fabric in a wrinkled and distorled
manner (Figure 1l). Often film separation
on/y occurs where light has damaged the
polyurethane.

$
i$

i
i
\
\
\
i
Figure I I The thin polyurethanef lm coating the
surface of this imitation leather garment has sepa-
Figure I0 Ultrasuede skirt, turned inside ouL rated from the backingfabric during dry-cleaning.
exhibits severe pilling and a hole where the In its weakened condition, the /ilm has tom and
polyester Jibres have shredded from stress. adhered to itse(.

198
Bonded and laminated fabrics having alayer
offoam beneath a surface fabric created many
problems for manufacturers during the late
1960s and early 1970s. Ifa face fabric was
loosely woven or knitted, it often appeared to
change colour as the foam layer beneath it yel-
lowed from light and atmospheric fumes. An-
other common problem dwing washing or
dry-cleaning was the separation of the foam
layer from the face or backing fabric. This local-
ized separation oflayers gave the garment a
bubbled appearance, as shown in Figure 12.
The separation often occurred in light-damaged
areas. There are several causes for the separa-
tion: insufficient softening of the foam before
adhering it to the face and back layer, or weak-
ening ofbond due to deterioration ofthe foam
as it ages.

The Textile Analysis Service has also exam-

ined bathing suits reinforced with spandex that
has stretched and lost its elasticity after several
months of continuous use in a chlorinated pool.
Polyurethanes are damaged by chlorine in pool
water or hypochlorite bleaches.

Experimental
Four new polyurethane fabrics were chosen
for testing because they are tlpical of synthetic
suede, synthetic leather and coated water-
repellent fabrics. The fabrics are described in
Table II. They were subjected to aging in sun-
light outdoors in order to determine whether
sunlight would produce degradation similar to
that seen in polyurethane garments analyzedby
the Textile Analysis Service. The polyurethane Figure 12 The dry cleaner was unaware of a water-
fabrics and a set of AATCC Blue Wool Light repellent polyurethane film bonded to the underside
Fastness Standards were exposed to daylight of this rain jacket. During dry-cleaning, thefilm in
one panel partially separated and shrank causing
through glass as described in AATCC Test
puckering of the face fabic.
Method l6C-1988 (AATCC 1989). The tem-
peratwe in the ventilated exposure cabinets was
not monitored continuously but was known to determined. Colour difference was determined
reach at least 45oC when the air temperature on a Hunter Lab D25M-9 Tristimulus colorime-
was 30oC. After exposure for 63 days during ter following AATCC Test Method 153-1985.
June and July 1991, Blue Wool Standard L-8 The CIE 1976L*a*b* system was used to de-
showed a colour change equivalent to Step 4 termine the difference in colour between the un-
on the Gray Scale for colour change. This exposed fabric and the exposed fabric. Changes
amount of daylight exposure is equivalent to in flexibility were determined according to
320 AATCC Fading Units (AFU) (AATCC ASTM Test Method D1388-64 - Cantilever
1989, 35). Following exposure to daylight, Test (ASTM 1990). Fabric mass was measured
changes in fabric colouq stiffness (flexural as specified in CAN/CGSB4.2 No. 5.1-M90
rigidity), mass and bursting strength were (CGSB 1990). Changes in bursting strength

t99
 Table II
Fabrics Used in Experimental Study

Fabric Structure Fibre Content Cost (1991)

(Sample Number)

Supersuede 60% polyestertibres $85/m

Springs Industnes 40% polyurethane Joam
Extremely fine fibres (4 pm) (polyether based)
embedded rn polyurethane foam
(1 and 2)

Ultraleather Plus face-polyurethane

Spnngs Industnes (polyester based)
Polyurethane skin-foam back-7 Oo/" cupram monlum
coated on a tricot knrt rayon/3o% nylon
(3 and 4)

Arctrc Armour face-nylon $8/m

Plarn weave face fabnc back-polyurethane
undercoated with white (polyester based)
polyurethane layer
(15 pm to 60 pm)
(6, 7, and 8)

Thintech garment hner knrt-nylon $19/m

3M Corporatron coatrng-75% polyurethane
Tncot knrt coated wrth (polyether basedy
transparent polyurethane 257o olefrn matnx
layer on a microporous
oletin matnx (9, 10, and 1 1)

following aging were determined according to the Thintech fabric became opaque during light
CAN/CGSB-4.2 Method No. I 1.2-M89 (CGSB exposure and the fabric shrank from heat. The
1990). Bursting strength was measured on an olefin matrix that makes up 25o/o of the coating
Instron Universal tester. Changes in surface to- layer is very heat sensitive. During the June and
pography and fabric structural details were ex- July exposure period, the outdoor temperature
amined with the aid of a Cambridge Stereoscan reached a maximum of about 30oC and the tem-
250 scanning electron microscope after speci- peraftre in the cabinet reached 45'C to 50'C
mens were sputter-coated with gold. in fulIsun.

Results and Discussion Colour Change: Ultraleather, Thintech, and

Arctic Armour all showed visible yellowing of
The four experimental polyurethane-containing the polyurethane layer after light exposure for
fabrics were examined frequently during the 63 days (Table III). The greater change showed
63-day exposure period. The turquoise blue by Thintech (LE:12.4) than by Ultraleather
Supersuede faded rapidly, as expected. It was (A E:5.3) or Arctic Armour (A E:3.3) is to be
assigned an AATCC lightfastness rating of expected because the polyurethane component
L-4, which is the minimal lightfastness rating in Thintech is ether-based whereas the other
recommended for most clottring. The pink two fabrics contain ester-based polyurethane
Arctic Armour nylon fabric coated with poly- coatings (Rek et al. I 988). The colour change
urethane also faded rapidly, even though the shown by Supersuede (A E:29.1) does not
polyurethane layer rather than the nylon was represent a change in the polyurethane foam be-
exposed to the light. The fading may have been cause it is hidden by polyester fibres compris-
caused by radiation that penetrated the opaque ing the fabric surface; rather, it represents
polyrethane film or atmospheric fumes in the fading ofthe turquoise-dyed polyester fibres on
exposure cabinet. The polyurethane coating on the surface ofthe fabric

200
 Table III
Changes in Phltsical Properties of Polyurethane-containing Fabrics after Exposure through
Glass to 63 Days of Sunlight
Fabric Structure and Flexural Rigidity Color Difference AE Masso Bursting
Composition o/o
changea (crE1976 (gm') o/o
change Strength
L*a*b* units) 7o chanoe
Suoersuede nonwoven
60% polyesterfibers +24 29 1" 246 -094 - tJ
40ol. PUR foam
Ultrateather Plus PUR
foam{oating tncot +28 254 -053 -10
knit backinq
Arctrc Armour
Woven nylon + tJ 3.3c 140 + 5.10 -75
PUR coatinq
Thrntech 3M
tncot knrt -27 12.4 89 -265 .78
PUR coatrno

a + sttfe[ - more flexble

b = =
represents fadrng of turquoise fabnc
' represents colour change ot polyurethane coating and fadrng of the ptnk fabric beneath
" mass before outdoor exoosure

Bursting Strength: Two fabrics, Arctic Ar- predominates (Rek et al. 1986). The Thintech
mour and Thintech, were extremely weak after fabric was more flexible after light exposure.
exposure to light. Their loss of bursting During outdoor exposure the fabric shrank
strength after a63-day exposure period,T5Yo slightly and curled at the edges probably be-
and 7 8o/o respectively, demonstrates how sens i- cause of the olefin matrix. The specimens
tive some polyurethanes are to light. The fail- curled and twisted during testing, making it dif-
ure of Arctic Armour, which has an ester-based ficult to measure flexural rigidity accurately.
polyurethane film, is due, in part, to the poor
light resistance of the nylon face fabric. During Scanning Electron Microscopy: Polyurethane
the bursting test, the nylon fabric shred under films are known to become crazed, cracked and
minimal presswe. The ester-based urethane pitted during outdoor exposure to light, oxygen,
film on this fabric, although expected to have air pollutants and micro-organisms (Gabriele
good light resistance, was not sufficiently et al. 1983). The polyurethane surfaces ofthe
strong to hold the degraded nylon fabric to- Ultraleather, Thintech and Arctic Armour fab-
gether. The formation of crosslinks during irra- rics were examined with the aid of an electron
diation could contribute to the observed loss in microscope. Figure 13 shows the original ap-
bursting strength. Bursting strength is control- pearance (left) and appeamnce after 63 days of
led by the least extensible fibres or yams in a outdoor exposure (right). There is little visual
fabric; they reach their maximum extension and difference in the Ultraleather (Figure 13a), or
break before the more extensible fibres or yarns the Arctic Armour (Figure 13b). This finding
have been stretched to their limit. is not surprising since both have ester-based
urethane coatings and Ultraleather showed little
Flexural Rigidity: The flexural rigidity or stiff- loss of mass (-0.53%) after exposure. Since the
ness ofboth the Supersuede and Ultraleather exposure cabinet was covered with glass and
fabrics increased 24o/o and 28% respectively. only filtered air entered the cabinet, fabrics did
The increase in stiffness is a recognized change not become soiled with particulate. The Thin-
in polyurethanes exposed to light and oxygen tech fabric (ether-based urethane) (Figure l3c)
and is caused by crosslinking ofthe polyure- shows more surface change than the other two
thane molecules (Rek et al. 1986). Both chain specimens. There appears to be some pitting of
scission and crosslinking are induced by light the surface; these pits are consistent with the
exposure, but crosslinking eventually loss of mass measure G2.65%\. The surface of

201
I3a Ultraleather - Unaged l3a Ultraleather - Aged

l3b Arctic Armour - Unaged l3b Arctic Armour - Aged

l3c Thintech - Unaged l3c Thintech - Aged

Figure l3 After outdoor exposure through glassJbr 63 days (420 AATCCfadingunits) Ultralearher (a) and
Arctic Armour (b) show no sur/itce change, whereas Thintech (c) shows evidence of pitting. Left: unexposed.
Right: exposed.

202
the Thintech specimen was analyzed with en- Dry-cleaning solvents do not damage polyure-
ergy dispersive X-rays. No change in elements thane and may be used for hand cleaning;
on the surface was observed after exposure. however, adhesives used to bond polyurethane
films and foams to fabrics are sometimes sol-
In conclusion , after a brief 63-day exposure pe- uble in dry-cleaning solvents (Figure 11). Com-
riod behind glass, the four specimens contain- mercial dry-cleaning is not recommended for
ing polyurethane showed physical changes that cleaning garments containing polyurethane as
are typical for ester- and ether-based urethanes. the process involves a considerable amount of
Yellowing, increased brittleness and loss of mechanical action and cannot be considered
bursting strength were expected and occurred safe for aged polyurethane films, coatings or
in most specimens. The woven nylon compo- foams (Figures 9, 10, and I l).
nent of Arctic Armour degraded much faster
than the polyurethane coating. In addition, the Polyurethane can be identified and distin-
heat-sensitive olefin matrix in Thintech added a guished from other similar materials by buming
complicating variable that made it difficult to tests and chemical solubility tests (Textile Insti-
assess the changes in the polyurethane film in tute 1975). Elastomeric polyurethane fibres
this fabric. (spandex or elastane) are easily distinguished
from rubber fibres by their microscopic appear-
Conservation of Polyurethanes ance (round cross-section) or by buming. The
characteristic odor of buming rubber will iden-
In general, the conservation considerations for tiff rubber. Synthetic suedes and leathers used
textiles containing polyurethane are the same in textile applications since the 1960s will be
as for other textiles. During storage and dis- made of either polyurethane or poly(vinyl chlo-
play, light exposure must be minimal, I.fV hght ride). Imitation leathers and suedes composed
excluded, relative humidity in the range of of polyurethane can sometimes be quickly dis-
50Yoto 55Yo, and high temperatures avoided. tinguished from products made of poly(vinyl
Storage and display areas should also have ade- chloride) by a Beilstein copper wire test (Tex-
quate ventilation to dissipate any volatile prod- tile Institute 197 5, 227). Poly(vinyl chloride)
ucts of degradation. Polyurethane-containing produces a green flame in this test while poly-
artifacts should be stored separately from other urethane does not. Solvents to dissolve or
textiles. Layers of foam, films or coatings disintegrate polyurethane coatings and foams
should not be left in direct contact in storage as include boiling dimethyl formamide and
they may stick together. For example, sleeves boiling NaOH (10%) (Textile Institute
should not rest directly on the sides ofa gar- 197s, 160).
ment and collars and lapels should be lifted
slightly or separated from the body ofthe gar- Acknowledgement
ment with inserts of fabric or acid-free tissue.
The authors thank several people who provided
Standard wet-cleaning procedures used in assistance in the preparation ofthis paper. Cath-
textile conservation are suitable for cleaning erine Jacobsen, a Clothing and Textiles student,
polyurethane. The wash water should remain gathered material for the literature review and
in the neutral range of pH 5 to 8. It is important helped perform physical tests on the experimen-
to minimize the agitation, flexing and mani- tal fabrics. George Braybrook, Dept. of Ento-
pulation of the garment during the cleaning mology, took the scanning electron micro-
process, particularly when the textile contains graphs, Linda Tumer took the photographs of
polyurethane foam. Aged foams are frequently Textile Analysis Service garments, and Diana
brittle and readily crumbled with too much han- Parsons typed the manuscript.
dling. Textiles containing polyurethane should
not be bleached, and they should not be heated
to dry, or pressed.

203
R6sum6 enfin des suggestions sur lafagon de nettoyer, en
toute sdcuitd, les v€tements renfermant du
La ddgradation des polyurethannes prisen8 polyurtthanne.
dans les t&iles du XX sidcle consewds dans
des musies References

Le terme < polyurdthanne > s'applique d un ltbu-Zeid, M.E., E.E. Nofal, F.A. Abdul-
groupe diversifi4 de matidres qui entrent daru la Rasoul. M.A. Marati. G.S. Mahmoud and A.
fabrication des textiles du x,f si,icle. Au nombre Ledwith, "Photoacoustic Study of Thermal
de ces matiires,figurent lesfibres (spandex), les Degradation of Polyurethane," Journal of
adhtsifs et les agents de consolidation des su'ides Applied Polyner Science, vol. 28,
synthdtiques, les mousses des tissus contre-coll4s, no. 7, 1983, pp. 2317 -2324.
les mousses de surface des simili-suides ainsi que
les rev€tements hydrofuges ou les surfaces imi-
Abu-Zeid, M.E, E.E. Nofal and L.A. Tahseen,
tant le cuir de certains tissus. Il incombe aux
sptcialistes de la restauration de se familiariser "Photoacoustic Study of tfv, UV-thermal and
Weathering Degradation of Rigid Foam
avec les diverses.facettes de ces mati,ires
depuis leurs multiples modes d'utilisation dans
- Polyurethane," Journal of Applied Polymer
les textiles jusqu'd I'tvaluation de I'dtat d'un Scimce, vol. 29, no. 8, I 984, pp. 2443-2452.
tissu ou de la forme sous laquelle se manifestera
leur ddgradation. Tous ces aspects seront traitis Abu-Zeid, M.E. and E.E. Nofal, "Effect of Cata-
dans le cadre de la prdsente communication, au lyst Residues on the Degradation of Rigid
cours de laquelle nous dtudierons en outre le Foam Polyurethane," Journal of Applied
mode de d'lgradation des polyurdthannes, en Polymer Science, vol. 31, no. 8, 1986,
passont en rewe les nombreux documents qui pp.2407-2415.
existent sur le sujet.
Allen, N.S. and J.F. McKellar, "The Role of
La photo-oxydation est particulidrement dom- Luminescent Species in the Photo-oxidation of
mageable pour les polyartthannes, et elle de- Commercial Polymers," tn: Developments in
meure la premidre cause des dommages subis Polymer Degradation-2, N. Grassie, ed.
par les textiles que nous avons examinis. Les (London: Applied Science Publishers, 1979)
polyartthannes d base de diisocvanates alipha- pp. 145-148.
tiques et de segments souples en polyester rtsis-
tent miew d la photo-oxydation que ceux qui sont
American Association of Textile Chemists
formts d paftir de diisocyanates aromatiques et
de segments mous de polydther. Des micro- and Colorists , Technical Manual of the
organismes peuvent crottre sur les polyurdthan- AATCC, Vol. 64 (Research Triangle Park:
nes, sous forme de pellicule craquelie. Au fur AATCC, 1991).
et d mesure de la croissance, des enzvmes sont
sdcrdtdes, qui peuvent couper les chaines au American Society for Testing and Materials,
hasard d la liaison urtthanne ou prds de I 990 Annual Book of Standards, Y ol. 7 .01
leurs extrdmitts. (Philadelphia, P.A.: ASTM, 1990) pp.36l-364.

Un certain nombre de v€temmts renfermant du B ackus, J.K., " Polyurethanes," in:. P o ly mer iz a -
polyur,lthanne sous forme de fibres, de mousses, tion Processes, C.E. Schildknecht and I. Skeist,
de pellicules ou d'adhtsifs seront prdsentds, avec eds. (New York J. Wiley & Sons, 1977)
une description des dommages qu'ils ont subis. pp.642-680.
Dans nombre de cas, les v€tements n'ont prisentd
aucun signe de dommage photo-oxydatifavant
Barendregt, R.B. and P.G Van den Berg, "The
d'€tre nettoyis d sec dans un dtablissement com-
Degradation of Polyurethan e," Thermo chim a
mercial. Leur saturation avec le solvant de net-
toyage d sec (perchlorodthyline) et I'agitation
Acta,vol.38, 1980, pp. 181-195.
mdcanique d laquelle ils ont 6td soumis pendant
le cycle de nettoyage ont fait que le polwr6- Canadian General Standards Board, Textile
thanne oxvdd s'est ddcompost. Nous fournirons Te s t Me th o ds, Can /C G S B - 4. 2 (Ottaw a:
CGSB.l990).

204
Carter, M.E., Essential Fiber Chemistry Gardette, J.L. and J. Lemaire, "Photo-Thermal
(New York: Marcel Dekker, l97l). Oxidation of Thermoplastic Polyurethane
Elastomers: Part 3 - Influence of the Excitation
Chapman, T.M., D.M. Rakiewicz-Nemeth, J. Wavelengths on the Oxidative Evolution of
Swestock and R. Benrashid, "Polyurethane Polyurethanes in the Solid State," Polymer
Elastomers with Hydrolytic and Thermo- Degradation and Stability, vol. 6, 1984,
oxidative Stability. l. Polyurethanes with pp. 135-148.
N-alkylated Polyamide Soft Blocks," Journal
of Polyner Science: Part A: Polymer Grassie, N., M. Zulfiqar and M.I. Guy,
Chernistry, vol.28, no. 6, 1990, pp.1473-1482. "Thermal Degradation of a Series of Polyure-
thanes," Journal of Polymer Science: Polwer
Darby, R.T. and A.M. Kaplan, "Fungal Chemistry Edition, vol. 18, no. l, 1980,
Susceptibility of Polyurethan es," Appli ed pp.265-274.
Microbiology, vol. 16, no. 6, 1968,
pp.900-905. Hepbum, C., Polyurethane Elas tomers
(London: Applied Science Publishers, 1982).
Du Pont de Nemours & Co. (Inc.), "Lycra
Spandex Fiber," Du Pont Technical Holker, J.R., "Weatheing," Textiles, vol. 17,
Information Bull etin Z-86 (Wilmington: no. 3, 1988, pp.64-71.
Du Pont, 1976).
Iyer, S.B., "Recent Developments in
Durst, P., "PU Transfer Coating of Fabrics for Polyurethanes - XXI,U Man-made Textiles in
Leather-like Fashion Products." Journal of India,vol.32, no. 4,7989, pp. 158-160.
Coated Fabrics, vol. 14. 1985. pp. 227-240.
Joseph, M.L., Introductory Textile Science
Fllip,2., "Decomposition of Polyurethane in (NewYork: Holt, Rinehart& Winston, 1986).
Garbage Landfill Leakage Water and by Soil
Micro-organi sms," European Journal of Kaplan, A.M., R.T. Darby, M. Greenberger and
Applied Microbiology and Biotechnology, M.R. Rogers, "Microbial Deterioration of Poly-
vol. 5, 1978, pp.225-231. urethane Systems," Deve lopments in Industri a l
Microbiologl,,, vol. 9, 1968, pp. 201-217.
Gabriele, P.D., J.R. Geib, R.M. Iannucci and
W.J. Reid, "Photochemical Degradation and Lenz, R.W., Organic Chemistry of Synthetic
Biological Defacement of Polymers," in: The High Polymers (New York: Interscience
Effects of Hostile Environments on Coatings Publishers, 1967).
and Plastics, D.P. Garner and G.A. Stahl, eds.
ACS Symposium Series 229 (Washington, Mansfield, R.G., "High Performance Coatings
D.C.: American Chemical Society, 1983) and Finishes Expand Uses," Textile World,
pp. 3 l7-330. May, 1985, pp. 58-60.

Gardette, J.L. and J. Lemaire, "Oxydation Oehlke, N., "Polyurethane Coated Fabrics,"
photothermique d' 6lastomdres de polyur6than- IFI Bulletin FF288 (Silver Spring, M.D.:
nes thermoplastiques, Partie 2: Influence de la International Fabricare Institute, I 980).
longueur d'onde et des conditions d'irradiation
sur le jaunissement des polyur6thannes d Pavlova, M., M. Oraganova and P. Novakov,
structure aromatique," Die Makromolekulare "Hydrolytic Stability and Protective Properties
Chemie, vol. I 83, no. 10, 1982, pp. 2415-2425. of Polyurethane Oligomers Based on Polyes-
ter/Ether/Polyols," Polymer, vol. 26,no. 1 l,
1985, pp. 1901-1905.

205
R6nby, B. and F.F. Rabek, Photodegradation, Stohler, F.R. and K. Berger, "New Results
Photo-oxidation and Photostabilization of Upon the Stabilization of Polyurethane
Polymers (London: John Wiley, 1975) Systems Against Photo-oxidative lnfluences,"
pp.241-243. Die Angewandte Makromolekulare Chemie,
uol' 158/159' 1988' pp' 233-246'
Reagan, E., "Examining the Moistur"
Managers," Outside Business, April, 1990, Textile Institute, Identification of Textile
pp.3746. Materials, Tth edn. (Manchester: Textile
Institute, 1 975).
Rek, V., H.J. Mencer and M. Bravar, "GPC and
Structural Analysis of Polyurethane Degrada- Wales, D.S. and B.F. Sagar, "The Mechanism
tion," Polymer Photochemistry,vol. T ,1986, of polyurethane Biodeterioration," in: Biodete-
pp.273-283. rioraion of Plastics and Polymers, K.G. Seal,
ed. (Kew, Surrey: Biodeterioration Society,
Rek, V., M. Bravar, T. JociC and E. Govorcin, 1985) pp. 56-69.
"A Contribution to the UV Degradation of Poly-
urethanes," Die Angewandte Makromolekulare Wales, D.S. and B.F. Sagar, "Mechanistic
Chemie, vol. 158/159, 1988, pp. 247-263. Aspects of Polyurethane Biodeterioration," in:
Biodeteriorafion 7,D.R. Houghton, R.N. Smith
Rek, V., M. Bravar, E. Govorcin and M. and H.O.W. Eggins, eds. (London: Elsevier
Suceska, "Mechanical and Structural Studies of Applied Science, 1988) pp. 351-358.
Photo-degraded Polyuretha ne," P o lymer D e gr a-
dation and Stabilit"v,vol.24,1989,pp.399-411. Wales, D.S. and B.F. Sagar, "Effect of

Rising, J. and M. Scalco,

,,Ultraleather@,', iJti:::tJft:;.,1,"Pw,::t:::H:;'#;"t
IFI Bulletin FF386 (Silver Spring, M.D.: Biodegradation 8,H. Rossmore, ed. (London:
International Fabricare Institute, l99l). Elsevier Applied Science, 1991) pp. 241-248.

Rosthauser, J.W. and K. Nachtkamp, Webster, R.M., "A Case Study of the
"Waterborne Polyurethanes," in:. Advances in Biodeterioration of Some Boot Uppers Made
Urethane Science and Technologt, Vol. 10, From a Polyurethane Poromeric," Journal of
K.C. Frisch and D. Klempner, eds. (Basel: the American Leather Chemists Associafion,
Technomic, 1987) pp. 12l-161. vol. 80, no. 6, 1985, pp.172-183.

Scalco, M., "Foam Laminates," IFI Bulletin

FF364 (Silver Spring, M.D.: lnternational
Fabricare Institute, I 989).

Shuttleworth, W.A. and K.J. Seal, "A Rapid

Technique for Evaluating the Biodeterioration
Potential of Polyurethane Elastomers," Appli ed
M icro b io loglt and Biotechnolo gt, v ol. 23,
1986, pp. 407-409.

206
Treating Early Regenerated Cellulose Textiles: Two Case Histories

Emma Telford

Textile Conservation Studio

The National Trust
Aylsham, U.K.

Abstract combined treatment of ammonia vapourfollowed

by wet-cleaning is done to see whether conver-
This paperfocuses on the positive and negative sion to salts renders acids more easilv removable.
aspects of some conventional and less conven-
nonal textile conservation treatments to regener- Introduction
ated cellulose textiles.
In the final year of training at the Textile
The effects of wet-cleaning and dffirent.forms Conservation Centre, Hampton Court Palace,
of de-acidification treatments are discussed in students are required to select and treat a textile
relation to two case histoies. The aim is to make object. This paper describes the conservation of
practical judgements about potential treatments, a piece of an early 20th-century costume that in-
drav,ingfrom both scientific eidence and per- corporated a manufactured fibre element.
sonal experience.
Twentieth-century material and manufactured
Thefirst case history, essentially a non-treatment,
frbres in particular had been entirely over-
concerns a piece ofa costume ca. 1912. The cos-
tume is composed of silk and cuprammonium
looked in the teaching programme at the time,
rayon fabics. Discussion focuses on the need to and little had been researched or documented
de-acidify or remove acid soilingfrom this textile, on this topic from the point of view of
bearing in mind the effect of water on regener- conservation.
ated cellulose fibres and their consequent re-
duced wet strength. The decision not to wet-clean The artefact was privately owned and the client
is argued. gave an almost entirely open brief, with the re-
quest that the artefact be made safe for storage
The second case history concerns part ofa cos- and occasional display for study purposes.
tume from the 1920s. An attempt to quantify the
effect ofwater on degraded regenerated cellulose Description and Attribution
is made by conducting thefollowing analyses
both before and afterwet-cleaning: l) surfoce pH The artefact was a dress composed of a pink
to establish the extent of removal of acid soiling; silk underdress and a black gauze overdress
and 2) tests to identify any reduction in tensile
made from cuprammonium rayon (or cupro as
strength.
it is also known). This latter fibre was identi-
fied by microscopic examination and solubility
In addition, experiments are conducted with
tests. This makes the dress an early example of
ammonia vapour to establish its efticiencv as an
alternative in reducing acidity. Specifically, a this fibre, which only came into general use

207
ca. 1910.t The gauze and silk layers ofthe lengthwise striations, smooth, featureless.
dress were stitched together and were not In comparison with ordinary viscose rayon,
separable. which has a much more lobal cross-section
giving rise to numerous lengthwise striations,
It was dated to ca. l9l2 by costume experts it is easily distinguishable.
from the Victoria and Alberl Museum, the
Museum of London and the National Museums o The frbres appear moderately brittle and some
of Scotland. The rough and ready construction evidently suffer the type of break known as
ofthe dress originally suggested that it had
brittle break. This is identified by the angular
been manufactured later than 1912, but in the
nature of the broken end, which has a step-
style of that period perhaps for drama or fancy
like feature. Other types of failure occurring
dress purposes. However, apparently such poor
in the fibres are tensile breaks, which occur
workmanship was in fact quite typical of even
when the fibre is still strong and which is rec-
haute couture clothing of the period. Other fea-
tures fypical of the period that helped to date ognizable by the granular nature ofthe bro-
the dress were the following: the very popular ken end; and axial failure, where the fibre is
colour combinations of the time, pink and fractured along its axis probably as a result of
black; a short filmy overdress over a longer the wear process.
underdress; heavy surface decoration; and
large clumsy fastenings. o Dark spots that sometimes show up on the
fibres may be attributable to static charging
However, the dress does have a history of under the scanning electron micrograph in
theatrical use, having been part ofa uriversity areas that have insufficient gold covering.
drama department wardrobe for some years. However, the dark, pitted features evident on
This would account for some of the damage some fibres may be due to a number of other
(e.g., heavy sweat-staining), which is more things. Bacterial, fungal, or acid attack (dur-
consistent with it being wom as a theatrical ing spinning or subsequently) may be respon-
costume rather than as a dress. sible, and depending on the dyestuffused, it
may be a dyestuffattack. Early examples of
Condition cuprammonium rayon (such as this) are likely
to have been left acid after manufacture, as
The poor condition ofthe dress probably re-
early washing techniques were ineffective
sulted from its use in the theatre. It was gener-
and any acidity would not have been
ally soiled, badly stained wrderarm and ripped
neutralized.
at the shoulders and around the hem. Both
gauze and silk fabrics were also badly creased
and worn in places. The cuprammonium rayon Treatment Options
appeared to be fairly degraded and brittle, being
The most effective treatment method from the
particularly weak in the weft. In order to estab-
point of view of stabilization would have been
lish in greater detail the condition, and also
wet-cleaning. This would have removed much
study and gain familiarity with the fibre charac-
soiling, creasing and also acidity, and was be-
teristics under very high magnification, fibre
lieved to be beneficial to the cuprammonium
samples from the hem area of the gauze and the
rayon, which is particularly susceptible to acid
braid were examined by scanning electron mi-
degradation. In investigating this possibility a
croscope. On examination of the photomicro-
grapls and with the help of the useful Hearle variety oftests, including one for dye fastness,
were conducted. It transpired that the pink silk
et al.'reference it was possible to determine
dye and also the black braid dye were not wet-
the following characteristics:
fast, and this eliminated the possibility of wet-
cleaning. However, the likely effect of wet-
r The fibres exhibit typical cuprammonium
cleaning on the degraded cuprammonium rayon
rayon cross-sectional and longitudinal fea-
also needed to be investisated.
tures, that is, kidney-like cross-section, no

208
The major drawback of cuprammonium rayon Underdress: Work commenced on the silk
is its poor tensile strength. This is due largely underdress, starting with the hem area, and in-
to the highly amorphous polymer system that volved relaxation by means of a cold poultice,
prevents sufficient hydrogen bondlng between followed by stitching to a dyed silk fabric sup-
polymer chains to impart strength.' This amor- port. The underarm areas were treated in the
phous structure (60% to 65% of the polymer) same way, although because of access prob-
also accounts for its reduced wet strength, lems to the underarm areas, caused by the
which is roughly half that of the dry. Water can gauze, it was unfortunately necessary to undo
penetrate easily and force the polymer chains original seaming in the gauze in order to ade-
apart, thus reducing further the inter-polymer quately support and stitch the silk. Wom,
hydrogen bonding. These hydrogen bonds will fragile areas of silk, such as the pleated cuffs,
rarely re-form in their original positions, which buckle and sash, were covered with a protective
is why the fibre is so susceptible to deformation layer of silk crepeline.
during and following wet treatrnents, and also
why distortion and creasing is likely to remain Gauze Overdress: Structural damage to the
even after strains have been removed in the bodice part or the gauze was also given a full
dry state. support with silk crepeline. The open leno
weave structure of the gauze allowed stitching
Cuprammonium rayon consists almost entirely without creating even tiny holes, or any dam-
of pure cellulose. However, the polymer chains age. A pink support was chosen to minimize
are short, on average having a degree ofpoly- reduction in the transparency ofthe gauze. The
merization of 250 to 300 low compared to gauze was not relaxed with humidity prior to
other cellulosics, such as -cotton (5,000) and supporting, because it may be inadvisable to
flax (18,000). Thus, the polymers contain a introduce moisture to an acidic textile without
high number of reactive end groups, making first rinsing away acidic products. The short pe-
the fibre more susceptible to attack by acids. riod during which the textile is damp may be
Hence the need to remove acid soiling. In addi- sufftcient for acids to start attacking the textile.
tion, the cuprammonium rayon contains some Damage caused during this time could be
aldehyde, ketone and carboxylic acid groups, extensive.
which are created during manufacfure, making
it prematurely slightly degraded. Cupram- In wet-cleaning, the acids are released, but are
monium rayon in a degraded condition would diluted immediately and washed away quickly.
naturally have an increased susceptibility to Due to the acid-sensitive nature of cupram-
acid attack. It is also reasonable to argue that monium rayon a treatment involving humidity
any wet treatrnents applied to cuprammonium was not used.
rayon in that state may damage the textile sub-
stantially by breaking the inter-polymer bonds Issues Arising from Treatment
that may not re-form.
The project raised some issues that were worth
Threads removed for testing to water showed investigating further.
no apparent deterioration. However, a sample
large enough to test the strength before and af- . What were the possibilities of neuhalizing
ter wetting could not be removed. In the event, acidity in regenerated cellulosic materials
fugitive dyes precluded any wet treatment as without wet-cleaning?
previously mentioned, but it was still important
to know whether the benefits of removing acid- r What likely effect would wet and other
ity would outweigh the possible disadvantages treatrnents have on the tensile strensth of
of wet-cleaning.
such textiles?

Treatment Experiments: Some simple experiments were

conducted on a kimono dated to the 1920s that
The treatment of the textile was essentially non-
could be sacrificed for this purpose. The textile
interventionist.

209
had a plain weave structure and a printed floral difficult due to the very fragile nature of the
design. The fibre was identified by microscopy weft. Once dry, the textile was visibly cleaner
and solubility tests as viscose rayon. and more lustrous.

The condition of the textile appeared to be Two other samples were placed in a desiccator
generally good and fairly typical ofthis age. in the presence of a small amount of aqueous
However, on close examination the weft ammonia solution (15%), and the humidity was
threads seemed extremely weak and the textile gradually raised with an ultrasonic humidifier
had to be handled with care. to 95Yo over a period of almost four hours. The
relative humidity was monitored constantly in
The aim of the experiment was to try an the chamber, and the pH of the textile, once
ammonia vapour treatment as a means of every hour. At the end ofthe process the pH
de-acidifying the textile without wetting it. A readings ofthe textiles were around pH 8.
subsidiary aim was to investigate whether con-
version of the acids to ammonium salts would One piece was given a minimal wet treatment
render the acids more easily removable in wet- that involved rinsing in deionized water.
cleaning, which would allow a shorter wet proc-
ess to be used. Thus a sample was treated with Results: Following these three treatments the
a rapid wet process (i.e., wash) after exposure samples were tested for tensile strength and the
to the ammonia vapour. results were as follows:

Tensile Strength Measurement: A sample of Untreated 6.7 kc/denrer rn weft 30.0 kq/denrer In warp
untreated textile was tested to determine its ten- Wet{leanrng 7 75 kgy'denrer In weft 30.3 kg/denier In warp
sile strength. The test was carried out according onlv
NH3 vapour 7.8 kg/denrer rn weft 27 7 kg/denrer In warp
to British Standard B52576:1986 on an Instron ontv
apparatus. The weft in particular was found
NH3 vapour 6.17 kg/denier rn weft 27 5 kg/denrer in warp
to be extremely weak, with a mean strength of olus wet nse
6.7 kgldenier. The warp was stronger with a
mean strength of 30 kg/denier.
Discussion
pH Measurements: pH tests on the untreated
textile showed it to be acidic, pH 4.5. For the These data suggest that the textile has been
purposes of these tests this was considered slightly strengthened by the wet-cleaning and
sufficiently acidic to show clearly any change also by the ammonia vapour treatment in the
resulting from treatment. weft direction, although the change is minimal.
The only loss of strength, again minimal, was
Test Treatments: One section of textile was caused by the treatment with NH3 vapour fol-
given standard wet-cleaning with a solution of lowed by a wet rinse. In the warp direction also
0. l% Synperonic'N' (non-ionic detergent) and the results seem favourable, with slight decline
0.05% sodium carboxy methylcellulose in sof- in strength for the two NH: vapour-treated sam-
tened water at a temperature of 19oC. The sam- ples and no real change for the wet-cleaned
ple was subjected to a series ofthree baths, sample.
which involved soaking initially and sponging.
This was followed by a series of rinses in sof- Thus it would appear that the wet treatrnent
tened water, with a final rinse in deionized on this textile certainly has no harmful effects.
water. The pH of the rinsing was monitored However, the NH: vapour treatments seem to
throughout, and this gradually increased have been as effective at reducing acidity. The
through repeated changes ofwater to pH 6. pH readings on all the samples remain neutral.

It was noted that once wetted out, the textile Conceming the sample treated only with ammo-
contracted quickly in both warp and weft direc- nia vapour, the question remains as to whether
tions. It thus required some manipulation to the salts in the textile will remain stable over
reshape it when laying it out to dry. This was time, or whether they will convert back to the

2r0
acid state. This depends on the sort ofenviron- R6sum6
ment to which the textile is exposed. It has
been suggested that the de-acidification will Le traitement de tefiiles anciens en
last only for a matter of mpnths before conver- cellulose riginirie : deux itudes de cas
sion back to acids occurs." Ifso, then there is
La prtsente communication porte sur les aspects
an argument for rinsing out the salts before
positifs et nigatifs de certains traitements de
they have a chance to convert back. The use of
conservation, classiques ou plus inhabituels, qui
buffers, such as sodium bicarbonate, however,
sont appliquts aw textiles de cellulose rtgtnerte.
may render these salts more stable. This possi-
bility has not been investigated. Les ffits du lavage et de diffdrentes formes
de traitement de ddsacidification sont abordds
Another disadvantage of ammonia vapour dans le cadre de deux 2tudes de cas. Et, notre
treatments is that the textile is humidified in its dbmarche vise d porter des jugements pratiques
soiled state, which as mentioned previously is sur divers traitements dventuels, en nousfondant,
certainly a cause for anxiety. d'une part, sur des faits scientifiques et, d'autre
part, sur des exptriences personnelles.
Itis worth remembering that wet-cleaning is
not used solely as a means of de-acidifying a La premidre Atude de cas, qui est essentiellement
textile. It also allows the relaxation and realign- un exemple de non-traitement, porte sur un cos'
ment of the threads and fibres, and often im- tume datant de I9l2 environ. Ce costume est
proves the appearance and handle ofa textile, constitui de tissus en soie et en rayonne au
generally restoring some of the physical fibre cuproammonium. Notre exposd poftera surtout
properties. Altemative treatments cannot sur l'|ventualitt de ddsacidif er ce textile ou d 'y
enlever les taches d'acide, en tenant compte de
do this.
facteurs comme I'effet de I'eau sur la cellulose
rdgdntrde et la diminution subsdquente de sa
Conclusion rdsistance au mouilli. Nous dbbattrons en outre
de la question de savoir s'il faut ou non laver
Although each textile must be considered ce textile.
according to its structure and condition when
devising treatrnents, the data obtained indicate La deuxiime 6tude de cas porte sur un costume
that wet processes can be applied to early regen- des anndes 20. Nous tenterons de fournir des
erated cellulosic fibres. Extra care needs to be donndes quantitatives sur l'ffit
de I'eau sur la
taken when handling such textiles in the wet cellulose r,lgdntree d,lgradbe d partir des analy-
state. Ammonia vapour has also been shown ses suivantes, qui ont dti ffictudes avant et aprAs
to work as a de-acidifying treatment on regen- le lavage :
erated cellulosic textiles. From these simple
experiments, however, it is not possible to es- l. une dvaluation du pH d la sudace, afin de
tablish whether the effects are too ephemeral to d'lterminer dans quelle mesure les taches
make it worthwhile. In any case, the aesthetic d'acide ont 6tt elimindes;
effects of wet-cleaning are far superior.
2. divers essais, afin de dtterminer la risistance
Acknowledgement d la tension.

Instron tensile test measurements were con- Des exptiences sont en outre effectudes avec
ducted at the Scottish Textile and Technology de la vapeur d'ammoniac, afin d'ttablir si cette
Centre, and photomicrography was conducted substance permet aussi de reduire l'acidit6. Et, ce
at the English Heritage Laboratories in genre de traitement, suivi d'un lavage, est
Fortress House. notamment employt pour ddterminer si, une fois
transformts en sels, les acides s'dliminent
plus facilement.

2tl
References 3. Cockett. S.R.. lr Introduction to Man-made
Fibres (London: Pitman, 1966).
I. Gordon-Cook, J ., Handbook of Textile
Fibres, VoL 2, Man-made Fibres,4th edn. 4. Sykas, Philip, personal communication,
flondon: Merrow Publishing Co. Ltd., 1968). Manchester City Art Galleries, February, 1991.

2.Hearle, J.W.S., B. Lomas, W.D. Cooke and

I.J. Duerdon, Fibre Fracture and lMear of
Materials An Atlas of Fracture, Fatigue and
Durability,- Ellis Horwood Series in Polymer
Science and Technology (Chichester:
Ellis Horwood Ltd., 1989).

212
Treatment of 20th-C entury Rubberized Multimedia Costume :
Conservation of a Mary Quant Raincoat (ca.1967)

Clare Stoughton- Harris

The Textile Conservation Centre

Hampton Court Pslace
East Molesey, Suney
U.K.

Abstract Problems and solutions are evaluated in relation

to treatment of other rubberized objects from the
This paper is based on a 20-week project that same museum collection. The technological and
was submitted in partfulfllment of a postgradu- industrial backgrounds ofsuch objects are investi-
ate diploma in textile conservation awarded gated to allow a better understanding of the mate-
by the Courtauld Institute of Art, University of rials and degradation processes, and to support
London. The project investigated the care of the conservation treatment that seeks to clean, re-
rubberized costume at The Gallery of English form, stabilize and support the object for storage,
Costume, Platt Hall, Manchester, England. study and display.

The paper demonstrates the problems of conserv- Introduction

ing 2}th-century rubberized costume by looking
in depth at a Mary Quant raincoat (ca. 1967). In 1985 a rubberized raincoat, designed by
This is a composite garment compising wool, Mary Quant, was donated by its owner to The
cotton, acrylic, viscose rayon, cellulose acetate Gallery of English Costume, at Platt Hall in
rayon, two rubberized finishes, plastic buttons, Manchester. By 1990, the raincoat's condition
metal fastenings and adhesives.
and stability was giving some cause for con-
cem. This initiated an investigation into the
The raincoat is ofimportance as an excellent
care of rubberized objects collected at Platt
example of Mary Quant's work in the late 1960s,
Hall, which was carried out as a postgraduate
and because it is handled frequently for study pur-
research project. The Mary Quant raincoat was
poses. It had become soiled and showed signs of
deterioration, and thus required attention to re- treated and thus formed the basis of a case
main an important study piece. The project is of study.
importance on a wider scale because of its multi-
media nature. It illustrates many of the inherent Platt Hall primarily houses collections of cos-
problems associated with rubberized composite tume made or worn in England or by English
costume. people. It also has a special objective to collect
examples of costume from the Manchester re-
The paper sets out full characterization of all the gion, which has been the centre of the rubber-
materials concerned and assesses the deteiora- ized rainwear industry since its beginnings.' It
tion mechanisms caused by: 1) inherent instabil- is therefore natural that the Platt Hall collection
ity and the interaction ofthe component parts; 2) has come to include a number of rubberized
use, bothfrom wear and study; and 3) mviron- sarments and accessories.
mental conditions.

213
Objects range from a late-l9th century rubber-
ized cotton bathing cap (from the Cunnington
collection) to a rubberized cotton raincoat de-
signed by BIBA in the early 1970s.

The Problem
Four main areas of concem over the waterproof
rubberized items of costume were identified:

o Many of the rubberized objects were of

unknown materials; the waterproof coatings
had not been identified.

o Some of the rubberized coatings were

deteriorating: they were starting to crumble,
dry out, or go sticky. Many rubberized ob-
jects had distorted and partially set, making
them liable to further damage.

o Storage of the objects was a problem in that,

as with all museums, space is limited and the
potential harmful interaction of rubberized
and other materials stored in close proximity Figure I
Front view of the Mary Quant raincoat
is not well understood. before conservation.

o The collections are all available to researchers through a frst order red plate. Solubility tesJs
on request and may be required for temporary were used to substantiate the identification.'
display from time to time.
The white (warp and weft) was identified as
The Mary Quant raincoat served as a good illus- a blend of cotton, acrylic and viscose rayon,
tration of these points and its treatment forms a and the black (warp and weft) as a chrome dyed
case study. The objective of the treatment was blend of wool, acrylic and cotton. The dye was
to allow the raincoat to remain accessible to re- identified by examination of a black wool fibre
searchers and to restore as far as possible the using a scanning electron microscope with en-
original intended appearance ofthe coat. This ergy dispersive X-ray analysis. The rubber
required that the stability of the raincoat be en- layer on the inside ofthe houndstooth textile
sured and the rate of deterioration be reduced was identified as neoprene (polychloroprene,
where possible. one of the earliest synthetic rubbers) using pyro-
lysis-gas chromatography (P-GC). Both natural
Description and Analysis and synthetic rubbers are usually blended with
other ingredients to give specific properties. Us-
The raincoat (Figure l) is of simple cut and ing a scanning electron microscope with X-ray
made of black and white houndstooth fabric analysis, the main inorganic components of the
that has a rubberized waterproofing layer on the neoprene were identified as calcium, probably
inside.Jextile fibres were identified by micro- in the form ofcalcium carbonate, and silicon or
scopy.'Natural fibres were recognized by their magnesium oxide, which were probably used
morphology in longitudinal section, and the as curing agents. Titanium was also identified,
synthetic components by their interference col- and was probably added as an oxide for bright-
ours when viewed between crossed polars at ening' even though, in this case, the rubberized
45" (i.e., between the extinction positions) coating was not intended to be seen.

214
The collar is faced with a black twill weave infra-red spectrum. Analysis was carried out
cotton that also has a rubberized coating on one using an infra-red transmitting microscope
side. This was identified as natural rubber using attached to an FTIR spectrometer.
P-GC. As the rubber was black it was assumed
to be filled with carbon black. The raincoat was constructed usins a combina-
tion of machine stitching and sticklng with
Both rubberized fabrics used in the raincoat's rubber solution.
construction are known as single textured fab-
rics because the waterproofing is applied to one This complex mixture of materials is typical
side of a single layer of cloth. (A double texture of 20th-century material culture. The raincoat is
is the altemative type of waterproofing, and in not unique; it is a mass-produced object and
this, two layers of fabric have a layer of rubber like many fashion items from the 1960s, not
between them.) Both types, double and single particularly well made. There are many faults
textured rubberized fabrics, were found in ob- in construction. The back panel, for example, is
jects from the collection at Platt Hall. not cut straight, which causes one shoulder to
be higher than the other. Tucks are taken in the
The whole raincoat is lined with green/brown facings to prevent them from hanging below
tabby weave fabric (Figure 2), which was iden- the hem, and the collar is permanently distorted
tified with solubility tests as cellulose acetate. by its crooked attachment.
In-situ electrolytic spot tests, as described by
Townsend,' identified the metallic non-textile Condition
parts ofthe raincoat. These included coated
brass underarm eyelets, a brass press-stud fas- Extemally, the instability of the coat was not
tener, and a zip fastening that is composed of immediately apparent. However, it was dis-
such metals as zinc alloy, brass, iron and nickel torted into heavy (i.e.,large, deep) vertical
plate. There were also three plastic buttons. The folds and the whole coat was discoloured. On
button material was identified as casein by its the inside of the coat where the neoprene layer
was visible, the fuIl extent of the degradation
was evident.

The neoprene coating on the houndstooth fabric

appeared to be in a worse state than the black
rubber used on the collar. ln unexposed areas,
such as inside the facings, the neoprene coating
was still supple and had not cracked. However,
in exposed areas and areas ofwear, such as
around the shoulder region, it had disintegrated
to leave behind a much weakened textile
structure.

The difference in the state ofthe neoprene and

the black natural rubber coatings was not just
due to the fact that one was synthetic and the
other was natural rubber; much was due to the
compounding. The natural rubber filled with
carbon black is more resistant to deterioration,
whereas the white neoprene is much more vul-
nerable as it contains white or pale pigments
and fillers. The polymer matrix is eaten away
leaving behind the chalk filler. Light is unable
to penetrate carbon black-filled rubbers and
Figure 2 Inside upper back: cellulose acetate lining, thus they are less r,ulnerable to the effects of
rubberized cotton collar and labels. light, particularly photo-oxidation.

215
No trace of antidegradents was found in the o The presence ofgeneral soiling deposited
neoprene. Ifantidegradents had been added during wear; it was established that the rain-
during the original compounding it is thought coat had never been wet- or dry-cleaned.
that they would have been of the phenolic type,
which are non-toxic and often non-staining. o Further analysis indicated that the cellulose
They are thus suitable for rubber used in cloth- fibres were considerably degraded and yel-
ing. However, they are volatile, can be leached low/brown cellulose degradation products,
out by water, and are not particularly stable to resulting from acid hydrolysis and the produc-
light. They are therefore lost easily. This loss
tion of aldehydes, were present. (The pres-
would be exacerbated in a coat exnosed to
ence ofaldehyde groups resulting fiom acid
heavy rain.
hydrolysis was established by a Fehlings re-
agent test on the cellulose fibres.) The aver-
Because the neoprene coating was a strucfural
age pH of the textile surface was in the region
stiffening element as well as a waterproof
layer, its deterioration was the predominant ofpH 5.
reason for the coat's poor condition.
r Although no antidegradents were found in the
Distortions in the coat, caused by wear and neoprene, it was possible that some other
hanging in storage (and which had become set), compounding residues had been absorbed
were clearly visible in the side panels; here the into the textile fabric, accounting for some of
textile weave hangs on the bias. The textile the discolouration.
weave had distorted and the neoprene had de-
formed by slowly flowing to the new position As mentioned before, the lining had a number
and then setting. This is possible at room tem- ofneoprene particles stuck to it, and it had also
perature as the glass transition temperature of developed a rippled effect (Figure 2). It was as
neoprene is -50'C. At the centre-back hem a if the acetate had been pressed against the rub-
heavy fold had also developed. berized houndstooth fabric so that the impres-
sion of the weave was hansferred onto the
Neoprene is known to be less resistant than lining material. It then transpired, in talking to
natural rubber to water' and there is evidence the original owner, that the coat had been dried
to suggest that when wet, neoprene becomes in fiont of a central heating hot air vent. This
fairly tacky. This was apparent in the form of would have forced the lining material against
deposits ofneoprene particles on the coat's cel- the outer fabric, thus forming the impression
lulose acetate lining. The lining had also ad- and in some areas making the lining stick to
hered lightly to the neoprene in some areas. the neoprene. Furthermore, the cellulose acetate
had discoloured as a result of gas-fume or acid
There was evidence of "metal poisoning" where fading, possibly induced by the acidity created
copper pight have acted as a catalyst for rubber
ageing.o This can be seen in the underaffn area
around the brass eyelets (Figure 3) and behind
the press-stud fastener in the centre ffont. How-
:; ;;
ever, heat and the build-up of moisture in the
underarm areaare thought to have been the
predominant degrading factors.

The houndstooth fabric was abraded in areas of

wear, such as around the cuff and centre-front
opening. The discolouration of the white yams
was thought to have occurred as a result of the
following three factors, which may have been Figure 3 Underatm eyelets punched through the
linked and/or self-perpetuating : neoprene-coated houndstooth fabric and showing
resultant degradation of the rubberized layer.

2r6
by the degrading rubber components and Also, there was a minor concem that any re-
general soiling. There were also rips in the maining antidegradents might be lost in the wet-
lining in both underarm areas. cleaning process; however, if any were still
present, they were probably non-functioning.
Investigations into Possible Treatment
Treatment
Preventive conservation was not considered
adequate and thus a procedure involving the fol- After surface cleaning with a low-powered vac-
lowing actions was investigated: uum suction unit the coat was stuffed with ny-
lon net until it became quite rigid. The purpose
o Reduce acidity and therefore limit the poten- of the stuffrng was to help maintain the three-
tial for acid hydrolysis of the textile compo- dimensional quality and to prevent the neo-
nents and reduce the possibility for further prene and textile from moving apart when
fading of the cellulose acetate wet.

o Remove degradation products In addition, a layer of non-woven spun-bonded

polyester Reemay was inserted between the cel-
o Remove body dirt, grease and general soiling
lulose acetate lining and the surface of the neo-
prene on the main body of the coat. This was to
act as a barrier layer and prevent the neoprene
o Enhance the aesthetic quality of the object by
from sticking to the lining fabric while wet.
removing the folds and distortions
During the wet-cleaning process, the coat was
Following the treatment, passive conservation supported on nylon mesh screens. These were
would be implemented to improve storage and also used to tum the coat.
reduce damage through handling.
The coat was wet-cleaned in soft water with a
A cleaning treatment was investigated. It was 0.1olo non-ionic Synperonic 'N' detergent and
realized that one treatment might be capable of
0.05 g per litre of sodium carboxymethyl cellu-
removing the damaging factors and also reduc-
lose as a suspending agent. Final rinsing was
ing the distortion. A wet-cleaning treatment,
carried out using deionized water.
known to be safe for the textiles, was proposed
(the dyes were tested for wet fasbress and
After about 30 minutes in the wash bath the
found to be non-fugitive) with the proviso that
neoprene started to plasticize. The coat was gen-
it would only be employed if tests indicated tly sponged on the front and back surfaces. A
that the natural rubber and neoprene would not
fair amount of soiling was removed in the wash
be unduly affected. While investigating the re-
water. The heary folds in the coat relaxed with
action of the natural rubber and neoprene to a no additional manipulation. However, the
wet treatment it was realized that, like the tex- weave ofthe back and side panels needed
tile elements, they might be plasticized and al- easing back into place.
low for relaxation of the folds and realignment
of the raincoat to its ideal shape. It was thought To d.y the coat, muslin was draped over it to
that the neoprene especially would plasticize to
absorb as much dirt and degradation products
some degree.
as possible. This proved very efficient although
the constituents of the brown stain that resulted
It became possible, therefore, to devise atreat- on the muslin were not identified.
ment whereby the coat could be cleaned and
reshaped in the same process. There were, how-
Having improved the appearance and removed
ever, risks involved in the wet-cleaning freat- some possible degrading elements, further stabi-
ment. Whilst it was expected that the textile
lization of the houndstooth fabric was required
elements of the coat would swell in water and to prevent firrther loss ofthe neoprene, particu-
the neoprene would soften, the manageability larly in the shoulder region (Figure 4) and to
of a wet raincoat was an unknown quantity.

2r7
 ",q.ff

Figure 4 Inside shoulder region illustrating

considerable loss ofthe neoprene surface and Figure 5 Underatm eyelet area after application of
remaining weakened textile sttucture. the spun-bonded polyester Reemay and Klucel G.

maintain the shape established in the wet A hanging system was made to provide maxi-
treatment. mum support; to allow the object to be stored,
studied and displayed; and to limit the degree
The reapplication of antioxidants, which of handling. The hanging mechanism was
has been"suggested as a ffeatrnent for degrading based on a system used at Platt Hall, but with
rubber.''o was considered for the neoprene. some modification. It consists of an armature
but abandoned at an early stage due to the un- for the shoulders, covered with polyester wad-
known response of the textiles. Coatings and ding and cotton lawn, and a second armature to
consolidants were considered, but a less drastic which the main body support is attached. Both
treafinent was adopted. A water-soluble adhe- armatures slide onto a vertical pole that has a
sive was used to attach a support to the hook at the top. This can be hung in the gallery
textile. cupboard, but when required for study, it can
be unhooked and placed on a stand. This re-
The non-woven spun-bonded polyester, Ree- duces the handling and means the researcher
may 2250 (of weight per unit area 17 {m', i.e., can view the object in the round. Detachable
the finest grade), was used as a support and ad- sleeve supports, which are secured with velcro
hered with a 50% solution of Klucel G (hydroxy at the shoulders, limit the manipulation neces-
propylcellulose) in water. Patches of the Ree- sary to remove the raincoat from its support. A
may were adhered over the degraded areas of case cover made with ties at the side keeps the
neoprene and a 5Yo Klucel G solution was raincoat in position while hanging. Figure 6
brushed over them. The Klucel G penetrated be- shows the raincoat afler conservation.
tween the patch and the object where it formed
an adhesive bond. Support was given to areas The Collection of 2Oth-Centurv
where the textile was most vulnerable, which
Costume
generally corresponded with the areas from
which the coat would be hung. In effect the The heatment of the Mary Quant raincoat has
Reemay patches acted as barriers between brought to light some issues that need to be ad-
the neoprene and the cellulose acetate lining dressed when collecting not only rubberized
(Figure 5). costume but all costume from the 20th centurv.

The ripped lining in the underarm areas was Identification

supported with suitably dyed medium-weight When objects are accessioned by the museum
Habutai silk held in place with stitching. A their composition is not usually known. Whilst
couching technique and fine polyester thread the identification of natural fibres and most syn-
were used. thetics is comparatively straightforward, the

218
 Waterproof Company, which produced the
coat, was dissolved in 1984.

Inherent Instability
The majority of modern costume acquisitions
are mass produced. In the 1960s, it seems that
the original purpose of such garments was not
longevity. The decade saw the development of
a fast-growing fashion business relying on built-
in obsolescence. Thus garments were not made
to last a lifetime, but perhaps one season only.

Some general conclusions can be drawn with

specific reference to rubberized costume as a
result ofthe survey ofsuch objects at Platt Hall
and from the treatment of the Mary Quant rain-
coat. Any object containing rubber is prone to
fairly rapid deterioration through shelf ageing
or normal oxidation of the rubber components.
Oxidation is initiated by thermal, photochemi-
cal or mechanical stimuli. The behaviour of
such materials is not well understood by most
conservators. It may seem that the most desir-
able treafinent is to eliminate oxygen by the use
ofoxygen absorbers and scavengers. However,
Figure 6 Three-quarterview ofthe raincoat after this could be slightly impractical for costume
consentation.
where the object is fairly large and requires
support.
non-textile synthetic components are not so eas-
ily characterized, thus conservation needs may The use ofeither double or single textured
not be obvious. S^imple methods can be used to rubberized fabrics has an effect on the vulner-
identify plastics." but full and reliable identifi- ability of the garment to degradation. It was
cation can only be carried out with instrumental evident that the double textured fabrics were in
methods, which are usually not within the better condition probably because the rubber
scope of a museum conservation department. layer was protected on both sides.

Without the help of The Malaysian Rubber With regard to the use of dark and light rub-
Producers' Research Association, it would bers, it was noted that the Mary Quant raincoat
not have been possible to identify and analyse had suffered as a result ofthe light coloured
the rubber compounds. It seems that staff of neoprene. This can be compared with a cotton
museums with smaller collections will have windcheater from 1956, which has a similar
to continue to rely on the goodwill of such rubber proofing and is degrading. However, a
establishments. utilitycape, also single textured but proofed
with a black rubber, is in very good condition.
With contemporary objects, information on
the materials originating from the manufacturer The cut of the textile in a rubberized garment
may be available when the object is acces- affects the stresses and hence the subsequent
sioned. In retrospect, it is unfortunate that such mechanical degradation of the rubberized coat-
information was not recorded at the time of ings. Components, such as copper (in the fasten-
the raincoat's acquisition in 1985. This is par- ings), have a local degradative effect on the
ticularly unfortunate because the London rubber. Copper acts as a catalyst for oxidation
of rubber.

2r9
Environmental Observations Acknowledgement
The behaviour ofthe objects in the cooler
stores and the warm office, where objects are The author thanks the Malaysian Rubber
studied, was particularly apparent. When re- Producers' Research Association (Hertford,
moved from the stores, most of the objects England) for analysis of the rubber; The An-
seemed fairly inllexible. However, once they cient Monuments Laboratory for help; English
had acclimatized to the warner environment, Heritage, London, for dye identification; and
they could be handled with greater ease. There- The Rubber and Plastics Research Association
fore, it was proposed that the objects should (RAPRA), Shropshire, for analysis of casein
have a conditioning period and not be handled plastic.
immediately when removed from the stores.
R6sum6
Conclusion
Le traitement d'un costume multimidia
The heatment carried out on the Mary Quant caoutchoutd du XX siicle : la conservation d'un
raincoat has gone some way towards seeking a impermdable Mary Quant datant de 1967 environ
solution to the problems identified at the outset
of the project. However, identification of the La prtsente communication sefonde sur un projet
varying materials required extensive use of d'une durde de vingt semaines prdsentt en we de
extemal resources and an assessment of the de- I'obtention d'un dipl6me d'etudes supdrieures en
conservation des textiles offert par le Courtauld
terioration mechanisms indicated the complexi-
Institute of Art de I'University of London. Des
ties involved in a multimedia object. The
travaux de recherche, portant sur I'entretien des
treatrnent initially appears successful and any costumes caoutchouttis conservts d la Gallery
immediate concern over the coat's storage is of English Costume de Platt Hall, d Manches-
now limited. However, the long-term effects of ter (Angletene), ont ttt effectuds dans le cadre
the cleaning, reshaping and stabilization tech- de ce projet.
nique will require monitoring, and an assess-
ment should be made of the storage, display Nous chercherons d montrer les probldmes que
and handling to which the object will be subjec- pose la conservation de costumes caoutchouttis
ted. The object can be studied by researchers du XX si'dcle en examinant en profondeur un im-
with minimal handling and manipulation by its permdable Mary Quant datant de 1967 environ.
new hung-storage method. V€tement crdt d partir de plusianrs mattiaux, il
est constitui de laine, de coton, d'acrvlique, de
With regard to the other rubberized garments in rayonne de viscose, de rayonne d'acdtate de cellu-
the collections at Platt Hall. a more detailed re- lose et de deux rev€tements caoutchoutts, et il
port on their condition has 6een documentedl0 comporte des boutons en plastique, des attaches
en mdtal et des adhdsifs.
and areas of concern pin-pointed with the possi-
bility of conducting some minimal treatrnents
Cet impermdable, qui constitue un excellent
based on those carried out on the Mary Quant
exemple du travail de Mary Quant d lafin des
raincoat. Howeveg it is clear that preventive
annies 60, sert souvent d'objet d'6tude, et il est
rather than risky active conservation methods donc souvent manipuli. Il s'est sali et il a com-
be implemented to give adequate support and menci d se dttdriorer avec le temps, si bien que
limit manipulation, so as to prevent further dis- des mesures ont &i €tre pises pour lui consenter
tortions and subsequent stress damage. It is im- son importantefonction d'objet d'Atude. Le projet
possible to prevent the deterioration ofthe a par ailleurs une portee plus vaste du simplefait
rubber coatings unless oxygen is eliminated. de la nature mulfimedia de cet objet, qui permet
On a more positive note, it is hoped that by giv- ainsi d'illustrer nombre de probldmes inherents
ing some attention to the rubberized costume, aux costumes caoutchouttis constituds de plu-
further consideration will be given to collection sieurs matdriaux.
policy, material identification and special
handling procedures. Nous procdderons, dans le cours de notre
communication, d une caractdrisation compldte

220
des matidres qui constituent cet objet, et nous 4. Townsend, J.H., "The Identification of
tvaluerons les micanismes de dtttrioration Metals: Chemical Spot Tests," in: Modern
attibuables aux facteurs suivants : Metals in Museums, eds. R.E. Child and
J.H. Townsend (London: lnstitute of
l" son instabilitd inhdrente et les interactions
Archaeology Publications, I 987).
enfre ses diverses parties;
2" son utilisation, rdsultant tout aussi bien d'une
5. Blow, C.M. and C. Hepburn, Rubber
usure normale que de safonction d'objet
Technologt and Manufacture, 2nd edn.
d'itude; (London: Butterworths, 1982).
3" les conditions ambiantes.

Nous proctderons enfin d une 4valuation des

6. Buist, J.M., Ageing and Weathering of
Rubber (Cambridge: Heffer, 1956).
probldmes soulevds et des solutions qui auront 6tt
trouvies pour y remidier, en tenant compte des
autres objets caoutchoutis quifont partie de la 7. Grattan. D.W.. "Rubber Deterioration: Can
m€me collection. Les anticddents technologiques Antioxidants Save Artifacts?" International
et industriels de ces objets seront en outre ana- Institutefor Conservation - Canadian Group
lysds pour acquirir une meilleure connaissance Newsletter, vol. 12, no. 4, 1987.
des matdriaux qui les composent et de leurs pro-
cessus de ddgradation, ainsi que pour Atubhr les 8. Murray, W.A.E., The Deterioration and
bases des traitements de conservation qui seront Conservation of Natural Rubber: A Prelimi-
appliquds d ces objets pour les nettoyer, leur re- nary Report, unpublished diploma report,
donner une forme, les stabiliser et les soutenir en Durham University, 1988.
vue de leur mise en rdserve, de leur Atude et de
leur exposition. 9. Mossman, S., "Simple Methods of Identi$-
ing Plastics," in: Preprints of The Modern
References Materials Meeting, Edinburgh, Scottish
Society for Conservation and Restoration,
l. Levitt,S., "Manchester Mackintoshes: A 1988, pp. 41-45.
History of the Rubberized Garment Trade in
Manchester," Textile History, vol. 17, no. 1, 10. Stoughton-Harris, C., "The Conservation of
1986, pp. 5l-70.
a ca. 1967 Raincoat Designed by Mary Quant,"
unpublished diploma report, The Textile Con-
2. The ldentification of Textile Materials servation Centre, Hampton Court Palace, in
(Manchester: Textile Institute, 1 985).
conjunction with the Courtauld Institute of Art,
Universitv of London. 1991.
3. Loadman, J., personal communication,
Malaysian Rubber Producers' Research
Association, 1991.

221
Spacesuits: NASA's Dream Conservator's Nightmare
-

Mary T. Baker Ed McManus

C on s erv ation Analyti c al L ab oratory National Air and Space Museum

Smiths onian Ins titution Smiths onian I nstitution
Washington, D.C. Llashington, D.C.
U.S.A. U.S.A.

Abstract the suits will be reviewed. In addition, treatment

considerotions, with afocus on the treatments
Human spaceflight was a pivotal achievement currently wed on the suits, will be discussed.
in U.S. history, and the spacesuits used are testi-
mony to the technological advances involved in Introduction
reaching this goal. Unfortunately, the materials
in these suits are deteriorating rapidly, endanger- Modern materials show up in virnrally every
ing the only records of some stages of the space muselrm. However, air and space museums, un-
race technology. Some materials, such as certain like traditional museums, primarily collect arti
neoprene/natural rubber blends, are degrading facts of relatively recent manufacture. Despite
due to poor manufacturing techniques necessi- the advances made in materials engineering
tated by the accelerated pace ofthe space pro-
and processing since the dawn ofthe space age,
gram. Other mateials, such as plasticized PVC
there is much to be leamed, especially with re-
tubing, pose a threat to the rest of the artifact.
gard to how such materials will age. Spacesuits
Still others are deteiorating due to conditions to
are an example ofa product ofthe space age,
which they were exposed by the National Aero-
nautics and Space Administration (NASA) testing made of materials that held up to extreme condi-
and use. Present methods ofdisplay and storage tions, that are now endangered by very ordinary
ofthese artifacts are not sufficient to delay the environments.
degradation process and, in many cases, are
accelerating it. It cannot be assumed that new objects and new
materials require less care than the more tradi-
Treatment and storage decisionsfor these objects tional museum objects. At this time, unfortu-
are not simple, as the suits contain avariety of nately, it is suspected that the Smithsonian's
mateials, including metals, synthetic and natural National Air and Space Museum (NASM) may
textiles, synthetic and natural plastics, rubber be losing some historically significant space ar-
and adhesives, as well as paints andvarnishes; tifacts because ofa casual approach to modern
clearly, an interdisciplinary approach to their materials and objects of recent manufacture.
conservation is needed.
Although, for example, Apollo spacesuits worn
during the first lunar mission compare in sig-
This paper will discuss the results of testing some
nificance to the Wright Flyer, they received far
proposed storage conditions for these materials,
less attention and storage care. It became appar-
which were designed with large collections and
low budgets in mind. Recent research on the ac- ent that storage, display and treatment guide-
celeration of agtng in some materials by the con- lines were needed if the suits were going to
ditions encountered in NASA's use and testing of survive for future senerations.

223
Spacesuit Acquisitions store mannequins. Spacesuits were removed
from exhibit when they began to "look bad."
On March 3,7967, the National Aeronautics There were no investigations into why the
and Space Administration INASA) and the changes in the suits were made. Some space-
National Air and Space Museum created an suits were used in demonstrations and were
agreement. Called "Agreement between the considered to be expendable.
NASA and the Smithsonian Institution concem-
ing custody and management of NASA histori- In the absence of a collections rationale , and a
cal artifacts," it is registered under NASA collections maintenance program, a rather inter-
Management Instruction 1052.85. The National esting consensus developed regarding the sig-
Air and Space Museum was essentially given nificance of the various types of suits. Flown
first refusal on space hardware and equipment suits (suits that were actually worn by astro-
that was retired from active service. nauts on space missions) were considered to be
historically significant and therefore worthy of
In 1968, the spacesuit worn by Project Mercury preservation. Unflown suits (suits worn in train-
astronaut Alan Shepard was acquired by ing or intended as back-up suits) were much
NASM; it was the first spacesuit of the collec- less significant, and were considered to be ex-
tion. Between 1968 and 1976,many more suits pendable. The perceived categories of signifi-
were acquired due to increased activity in hu- cance provided a convenient excuse for
man spaceflight programs and the race to the tolerating poor storage conditions and con-
moon. There was no collections rationale or sumptive use, and scavenging for spare parts.
preservation planning for the new space age ac-
quisitions. Many of the suits that were collected Spacesuit Development**
by NASM were placed on loan to other muse-
ums, and in some instances back to NASA for Long before the space program was a reality,
exhibit in their visitor centers. suits were being designed to keep humans alive
in low pressure environments. Early achieve-
Prior to 1975, spacesuits not on exhibit or loan ments were made in the early 1930s, when alti-
were stored in a warehouse in the Georgetown tude records were being made and broken by
section of Washington, D.C. They were packed balloonists and pilots. During the 1930s and
in boxes or suitcases. In 1975 the spacesuits 1940s, pressure suits were developed for short-
were relocated to the Garber Facility, where term emergency use by jet pilots. None were
they were hung on hangers and placed in tall, very mobile when used in low pressure, nor
cedar-lined cases. It is not known whether the were they comfortable. In the early 1950s the
cedar-lined cases were used because they were U.S. Air Force set out to fulfill a requirement
available, or because the cedar lining was con- that a pressure suit be developed for pilots of
sidered to be a preservation measure. At that the B-52 plane. The suit had to be comfortable
time, there were no storage buildings with mu- enough to be used for many hours at a time and
seum quality environmental control at the Gar- flexible enough to allow the pilot sufficient
ber Facility. Many buildings were unheated. movement for flying. At the same time, the
U.S. Nary was developing its own high altitude
One of the primary reasons for accepting suits suit with similar requirements. Thus, in the
was for exhibit; the spacesuits were treated as early 1960s, when John F. Kennedy proclaimed
exhibit props. Back-up suits and training suits the need for a space program that would eventu-
were collected for spare parts. No thought was ally put a man on the moon, decades of re-
given to exhibit conditions, such as light levels, search on spacesuit development had already
temperature and relative humidity, or the length been completed.
of time suits could be safely exhibited. The
suits were stuffed with ill-fitting departrnent
** There are few histories ofspacesuit development;
* A more complete discussion of the acquisition and the most complete is that of Mallan.' Another helpful
treatment philosophies has been given by Baker and resource comes in the person of Lillian Kozloski at
McManus elsewhere.' the Smithsonian Institution.r

224
The suits for the first missions (Mercury) were a suit that was inflated, but none of the de-
modified Navy high-altitude suits with a venti- signed systems were satisfactory. The inflated
lated undergarment developed by the Air Force. suits, like balloons, were very difiicult to bend,
They consisted of a rubberized cloth suit that causing near immobility of the astronaut. The
could be pressurized, an outer garment and a solution for the Gemini suits was to use a simi-
close-fitting helmet. They were very difficult lar rubber-coated fabric bladder to that used in
to move in, especially when in use in a vacuum, the Mercury missions, which covered most of
but this was considered a minor problem, since the astronaut like a loose wet suit. This bladder
the Mercury astronauts would complete their was restrained from expanding under pressure
missions sitting in their capsules. The suit was by a net (originally Dacron, later Teflon),
mainly an emergency protection from acciden- which decreased the ballooning effect, but was
tal loss of pressure during the mission. The still flexible, allowing bending of the joints.
outer layer of the suits was an aluminized cloth, The aluminized outer garment used in Mercury
which, it was hoped, would protect the asho- was used only in the protofzpe Gemini suits; it
naut ffom radiation and extremes of heat during was replaced by Nylon and Nomex outer layers
orbit and especially during atmosphere re-entry and aluminized Mylar inner layers. The Gemini
(Figure 1). suits had increased abrasion resistance and mo-
bility, and could be used for the first space-
walks (Figure 2).

Figure 2 Gemini spacesuit, once worn by Edwin

I|thite, now on display in NASM. The outer layer
is aJitcsimile, made by David Clark Company, the
original was damaged by exhibit and has been
retired. Photo by Virginia Pledger.

The Apollo missions were the culmination of

the race to the moon; the suits needed not only
to protect the astronaut on the surface of the
moon, but also to allow travel far from the cap-
sule or lunar lander. Modifications on the suits
Figure IMercury spacesuit, wom by L. Gordon included improved helmet design and a liquid
Cooper. Notice the aluminized cloth exterior and coolant garment (LCG) for temperature control.
lbam padded helmet. Photo courtesy oJ'NASA. The LCG was a union suit with tubing running
throughout; water flowed through the tubes,
The next set of missions (Gemini) put a new carrying away excess heat from the astronaut's
demand on the suits. The astronaut needed to body. The LCG was a great improvement over
be able to perform work in the spacesuit inside the Gemini method, which used air circulation
and outside the capsule. Previously, attempts for cooling. The improved helmet design elimi-
had been made to solve the problem of bending nated the close-fittins foam cushion around the

225
head that had been used in Mercury and Gem- collection of only the suits used on missions
ini helmets; this allowed for more mobility does not give a complete story of the inventive-
inside the helmet (Figure 3). ness that went into suit design.

Research is required at this time to gather

historic information about the design, construc-
tion and testing of spacesuits. Time is of the
essence because some of the companies that
were involved with spacesuit research and
development no longer exist, and researchers
who were involved are growing older. Research
is also important to the understanding of the de-
terioration process. That information will en-
able us to develop better exhibit and storage
techniques.

Research on the NASM Suits

The materials in the spacesuits that were show-

ing the most degradation are the soft rubber
pieces, such as gloves, boots, linings and gas-
kets, and the adhesive used to laminate the
aluminized fabrics. The rubber parts were sof-
tening and flowing, then hardening irreversibly
into distorted shapes. The adhesive used in
laminating the aluminum to the fabric was be-
coming very brittle and flaking off, along with
the aluminum. Other problems included the
poly(vinyl chloride) tubing on the LCGs, which
was starting to weep plasticizer; the foam in-
Figure 3 Apollo spacesuit.
side the Mercury and Gemini helmets was hard-
ening, and the textiles showed wear spots and
Shuttle suits had added requirements, which
discoloration.
included greater ease ofdonning and the capac-
ity for reuse. The previous suits were made
The initial work on this project had centered
for one-time use only, and for a particular
around following the free radical population, in
astronaut. Now, the suits were to be made in
an attempt to determine if there were still reac-
generic sizes, with, it was hoped, a long shelf
tive sites in the materials. This was a matter of
lile. The traditional zipper construction was
concem since, in the presence ofoxygen, free
abandoned for teflon seals, which were more
radicals cause oxidation and chain scission, but
flexible and allowed the astronauts to dress
in the absence of oxygen, fiee radicals cause
themselves and to dress in less time.
crosslinking, as llas been discussed in more de-
tail by Schnabel.'This information could be
The above evolution of suits is a mere outline
important in choosing an oxygen-free storage
ofall the changes that took place in spacesuits atmosphere. Electron Spin Resonance (ESR)
over the course of years. While the changes can
had been used to show very high populations
be followed by examination of the flown suits,
of free radicals in the glove materials, which
this does not tell the whole story. Many ideas
were determined to be very stable; most likely
and improvements were tested on protob/pe
they were the result of the antioxidant in the
suits; some of these changes were adopted for
gloves having scavenged the radicals. Since
the final suits, others were not. Some suits were
these stable radicals are not likely to be a threat
of a completely different design, having been
to the suit materials, other causes for their
part ofa contract bid from a new company. A
breakdown were investigated.

226
It was noted in a survey of the suits that many others were dipped in plain distilled water. The
of the unflown suits were in much worse condi- oxidation of the samples was monitored hourly
tion than the flown suits. Some of this differ- during artificial aging at 60"C, 70oC, 80oC and
ence was attributed to the respectful handling 90oC for various periods. The FTIR spectra
of the flown suits by NASA and NASM, yet were collected, and the absorbance ofthe car-
some of the problems could not be explained bonyl band was measured. The spectra showed
by poor handling alone. Some background the greatest difference between pool and dis-
research showed that extensive artificial aging tilled water at 90oC and no sisnificant differ-
testing had been conducted on the natural rub- ence at 60'C (Figure 4).
ber/neoprene blends as well as on the aluminiz-
ing adhesive, the results of which predicted that
at01520
the materials should have fared better than they
did, even after storage at high temperatures.

More background research revealed that most

of the unflown suits were used in training the
astronauts for space; some of this training in-
cluded wearing the suits in a swimming pool
to simulate maneuvering at low gravity. These
pools were kept chlorinated; some sources re-
called using calcium hypochlorite. (Such pools
are still being used for the same purpose by the
present manufacturer of space shuttle suits.)) 4ooo 35oo 1@o 5@
'@ 'ffr"nrrotffnr-rt*
Since hypochlorite is a strong oxidizer, it
seemed likely that residues from the swimming Figure 4 FTlR-microspectra of polvisoprene: (l)
pool tests could be accelerating the degradation dipped in distilledwater, unaged; (II) dipped in
of the spacesuit materials. Since natural rubber distilled water, aged at 80"C for 3.5 hours; (Lil)
dipped in pool water, unaged; (IV) dipped in pool
seemed to be the suit material that was suffer-
water, aged at 80'Cfor 3.5 hours Note the peak at
ing the most from aging, it was a logical choice t
I 7 2 5 cni is equal for the two unaged sampies. but
for the first material tested. The natural rubber is much greaterfor the pool-dipped aged sample
was cast on aluminum foil sheets from solution than the distilled-water aged sample, indicating
and allowed to dry. These films could then be greater oxidation ot that point, due to the hypo-
monitored by Fourier Transform Infrared Spec- chlorite in the pool water. Also note, the growth of
troscopy (FTIR), using a microscope attach- the broad peak at around 3400 cm-' , u'hich suggests
ment in reflectance mode. The dried films were fo rm at i o n of p e ro x id e s.
very irregular; they were examined under the
microscope and the smoothest areas were cho- While this is an interesting phenomenon, if the
sen. These areas were marked with the tip of a increased oxidation rate of the rubber chain is
disposable pipette, which coincidentally is the not causing a like increase in chain scission,
same size as the analysis area of the FTIR-mi- then the mechanical properties of the rubber
croscope. (In-depth information on the FTIR- are rurlikely to have been changed by the hypo-
microscope theory and methods as they relate chlorite residue. In other words, the additional
to conservation has been given by Baker, von oxidation seen by the FTIR could be in the
Endt, Hopwood and Erhardt.o) With this mark form of stable carbonyls on the rubber chain,
as a guide, the microscope could be set to ana- and no extra chain scission would actually be
lyze the exact same spot each time; the spectra happening. Molecular weight determinations
taken from such a spot on a given day would would be necessary to monitor the chain scis-
overlay perfectly with spectra taken from the sion, as well as any other molecular weight
same spot one week later. changes, such as ciosslinking. These mo'iecular
weight changes were monitored by the use of a
Several natural rubber/aluminum foil sheets Size Exclusion Chromatograph (SEC), made up
were dipped in simulated pool water, while of a multi-angle laser light scattering detector

227
(MALLS) and ultrastyragel columns. The col- reappraisal. While this work may have found
umns included a linear, mixed bed column, able an answer to why the non-mission suits are de-
to separate molecules by molecular size for a teriorating faster, the need to look for possible
range of molecular weights from 500 to ways to stop this deterioration still exists. In ad-
20,000,000. An additional column, useful for dition, work will continue to test storage meth-
low molecular weight molecules, was added to ods for all the suits not on exhibit; at present,
increase the sensitivity on the low end of they are housed at low temperature, in the dark,
this range. with internal support and an air filtering sys-
tem. Questions of interest that have come about
The pool-treated and distilled water-treated include the following: Will it be cost-effective
rubber samples were analyzed by the SEC- to contain the suits in bags with an oxygen-ffee
MALLS set-up and the results indicated a atmosphere? Can NASM demand proper
difference in the molecular weight distribution mannequins (Figure 5) for the entire collection?
between the distilled water- and pool-treated What restoration, if any, should be allowed of
rubber samples. Not all of the samples have the suits?
been analyzed to date, so there are no comela-
tions that can be made with time or tempera- These questions apply to the suits on exhibit, as
ture, as yet. well. Would it be possible to acquire air-tight,
oxygen-free. UV-screening, temperature-con-
Conclusion trolled display cases? While this goal may be
unrealistic, NASM is reviewing its loan policy
These experiments show that exposure to cal-
cium hypochlorite will accelerate the oxidation
ofrubber, apparently both by decreasing the in-
duction time and increasing the rate of reaction.
Ramifications of these findings pose the follow-
ing questions: Will storage in nitrogen slow or
stop this process (or will it favor another reac-
tion, such as crosslinking)? Is there another
way to stop it (such as cold storage or chemical
treatrnent)? These areas will be studied in the
future. Other future plans include studying
some of the other materials that are in trouble,
such as neoprene, the nitrile adhesive used in
making the aluminized fabric, and the foam
inside the helmets.

In summary, NASA has a collection of histori-

cally important spacesuits, most of which have
not received the best handling in the past. In
this collection, a subgroup exists ofsuits that
have not been considered important because
they were not used on missions. However,
many of them contain intermediate stages of
development, which also appeared in the mis-
sion suits; considering the amount of records
that have been destroyed by NASA and the suit
manufacturers, these non-mission suits may be Figure 5 Virginia Pledger (ight) and Lillian
the only record ofthe course ofthese develop- Kozloski installing mannequin head in Gemini suit
ments. Most of these suits have probably been helmet. Virginia Pledger designed and constructed
mannequins (under contractfor NASM) thatfit
exposed to a strong oxidizer in the form of cal-
the new criteria for dilplal of spacesuits as given by
cium hypochlorite in addition to the careless Baker and McManus.'
handling they received before the recent

228
 and recalling some of its suits for re-evaluation. mieux, car ces objets contiennent des matdriaux
The idea is to display the suits for public educa- tris varits, dont des m,ltaux, des textiles svn-
tion of the history they represent without fur- thdtiques et naturels, des plastiques synthttiques
ther endangering the only surviving artifacts et naturels, du caoutchouc, des adhesifs, des pein-
of the human space race. tures et des vernis; et il coniendra donc, de toute
dviden c e, d' adopt er u ne approc he multi d i s cipli -
naire pour assurer leur conseryation. Nous trai-
An interesting development, as a result of
terons donc, dans la prdsente communication,
this project, has been the contacts that have
des rdsultats de I'essai de diverses propositions
been established with companies currently
qui, visant de grandes collections tout en tenant
manufacturing spacesuits. We are concemed
compte d'un budget restreint, difinissaient les
with the preservation of historic spacesuits conditions de mise en riserve de ces matidres.
while the manufacturers have a desire to extend Nous passerons ensuite en rewe les plus recents
the useful working life of spacesuits currently travaux de recherche sur le vieillissement ac-
in service. Both of us are, for once, very much ctl2rd de cerlaines mati?res qui serait attribuable
interested in the same goals; the next genera- aux conditions dans lesquelles elles sont mises d
tion may, as a result, have more comprehensive l'essai ou utilisdes d la NASA. Enfin. nous trai-
and stable documentation of the advances being lerons de certaines considdratiois relatives aux
made in human spaceflight today. traitements de conservation, en mettant tout par-
ticuli,irement I'accent sur cew qui sont actuelle-
R6sum6 ment appliquts aux combinaisons spatiales.

Les combinaisons spatiales : un rAve poar la References

NASA, mais un cauchemar poar les sicialistes
de Ia restauration l. Baker, M.T. and E. McManus, "History,
Care, and Handling of America's Spacesuits:
Le lancement d'engins spatiaw habites a 6td un Problems in Modern Materials," Journal of the
dvdnement marquant de l'histoire des Etats-Unis, American Institute for Conservation, vol. 3l
et les combinaisons spatiales que portaient alors (t992) pp. 77 -85 .
les astronautes tdmoignent des progrds techno-
logiques sttr lesquels reposait cette relussite. 2. Mallan, Lloyd, Suiting Upfor Space: The
Ntanmoins, les matdriaux qui entrent dans la
Evolution of the Spacesuit (New York: The
composition de ces combinaisons se ddtdriorent
John Day Company, l97l).
tris rapidemerrt, menaQant ainsi l'existence m€me
des settls objets qui tdmoignent encore de cer-
taines 2tapes de la course technologique de 3. Kozloski, L. Suited Upfor Living and
I'espace. La ddgradation de ces matdriaux I(orking in Space: U.S. Spacesuits of the
amalgames de caoutchouc naturel et de ndo- - des National Air and Space Museum,NASM
prine, par exemple s'explique tant6t par la Artifact Series, Smithsonian Pres s, unpubli shed
-
mauvaise qualiti des techniques de fabication manuscript.
auxquelles on a dA, dans la course ffi4nde du
programme spatial, avoir recours. Tant6t, comme 4. Schnabel, W., Polymer Degradation,
dans le cas des tubes de polvchlorure de vinyle Principles and Practical Applications
plastifid, ils constituent une menace pour le reste (Munich: Hasner International, l98l) pp. 86-89.
de I'objet. Et dans d'autres cas encore, ils se
ddttiorent d cause des mauvaises conditions 5. Gomes, Cheryl, personal communication,
auxquelles ils ont tte soumis, lors de leur mise d ILC Dover, March, 1991.
I'essai ou de leur utilisation, d la National
Aeronautics and Space Administration (NASA).
6. Baker, M., D. von Endt, W. Hopwood and
Les mdthodes qui sont actuellement mises en
application pour I'exposition et la mise en
D. Erhardt, "FTIR Microspecfrometry: A
rdserve de tels objets ne sont pas sffisantes pour
Powerfirl Conservation Analysis Tool,"
retarder la degradation et, dans nombre de cas, Preprints of the I6th Annual AIC Meeting,
elles l'accdldrent. Or, il n'est par ailleurs pas June l-5, 1988, New Orleans, Louisiana
(Washington, D.C.: American Institute of
facile de dbterminer le genre de traitement ou le
mode de mise en rbserve sui leur conviendra le Conservation, 1988) pp. l-13.

229
 Conservation of Paintings on
Delaminated Plywood Supports

Donald C. Williams Ann Creager

Cons erv ati on A naly tic al Labor atory National Museum of American Art
Smithsonian Institution Smithsonian Institution
Washington, D.C. Washington, D.C
U.S.A. U.S.A.

Abstract A planned exhibition of Johnson's workforced

conservators at the National Museum of Amei-
As industrial production of modified wood materi- can Art and the Conservation Analytical Labora-
als developed in the early 20th century, plywood tory of the Smithsonian Institution to explore
panels and other composites were used by artists treatment optionsfor the paintings. Several proce-
as substitutes for more traditional art materials. dures were conceived and designed spectfically
Nowhere was this more evident than in painting, for the treatment of the painting supports, but
where artists employed plyuood as a cheaper, could be applied to any wood laminar structure.
more uniform, and versatile replacement for solid The tvvo primary optionsfor the treatments were
lumber substrates in panel paintings. Now, a half to either reintegrate the existing artifact as a
century or more later, as these modern substrates whole, or to remove the lamina bearing the paint
begin to deteriorate, those caringfor these layer.
artifacts arefaced with an entirellt new set of
problems. This paper discruses the ethical and technical is-
sues surrounding the artifacts and their treat-
These problems became evident in a group of ments, reviews the modelsfor each of the major
paintings by the American artist William H. options and subsequent variants considered,
Johnson (1901 to 1970) in the collections ofthe and illustrates the treatments as they were finally
National Museum of American Art. Johnson, who executed.
was active in both the U.S. and Europe bettveen
1926 and 1946, was academically trained but Introduction
chose to paint in a style approaching the vernacu-
lar. Probablyfor reasons ofeconomy, Johnson In late 1989, the final checklist of paintings for
executed several paintings on plywood sheets be- the major show of William H. Johnson's later
tween 1939 and 1944. Near the end of his life, ill
works was received in the National Museum of
healthforced his hospitalization. At that time and American Art laboratory. Of the 39 paintings
after his death, a dispute over the disposition of
from the collection that were requested for the
his belongings led to the abandonment ofseveral
ofthese paintingsfor approximately l0 years.
show, l0 were on plywood supports. They
For some of this peiod the paintings were prob- were in various states of delamination and
ably on the docks of New York City. Extended ex- structural deterioration. Deletion of any of
posure to the unregulated outdoor environment these works from the show was not a considera-
functioned as an accelerated aging test on the tion, since they all were crucial images, and in-
paintings and their substrates, resulting in gross cluded several of the artist's masterworks llom
delamination, buckling, and fracturing of the the period. We knew that this show would
plyvood laminae. travel, probably extensively, and that although

231
venue selection might be subject to our opin- returned to New York. There he began his
ion, selection would not rest on conservation series of interpretations of black life in Harlem
standards alone. Therefore, for the most se- and the rural South, depicted in a simple flat
verely damaged paintings, major treatment style. The paintings on plywood are from this
was the only option. period between 1939 and 1944.

Before discussing the specific problems, materi- In the early 1940s, Johnson's life took on tragic
als and treatments, it is worth taking a brief proportions. A fire burned his studio apartment
look at the life ofthe artist. It is not easy to un- and, in 1943, Holcha died ofcancer. Johnson,
derstand how such a powerfully original painter who had already started to show signs oferratic
has achieved little recognition to date unless it behavior, now began to clearly manifest the be-
is due to the fact that he was black and he was ginnings of a debilitating mental illness. He
not a part of any movement, school or group. made at least one return trip to Denmark with
However, it is easy to understand, from the se- all his paintings and possessions in tow but was
quence ofevents in his life, how a large body eventually sent back to New York in 1947, and
of his innovative and spiritually lyrical works was immediately placed in a mental hospital
came to fall into such profound disrepair. on Long Island. He never painted again. He
remained in the hospital for 23 years until his
William H. Johnson was born in the small death in 1970.
town of Florence, South Carolina in 1901. His
schooling was intermittent and although his tal- It can only be imagined how those paintings
ent was recognized and encouraged by several were transported back and forth between conti-
teachers, he was unable Io realize his ambitions nents or how they were stored in a dockside
in the segregated South. At 17, he moved to warehouse following Johnson's committal to
New York and eventually entered the five-year the institution. It is this more than nine-year pe-
program at the National Academy of Design. riod ofextended exposure to an unregulated en-
The artist Charles Hawthome was his teacher, vironment that has functioned as an accelerated
mentor, and friend, and upon his graduation aging test on the paintings and their substrates,
in 1926, Hawthome personally raised resulting in loss, delamination, buckling and
enough money to send him to Europe for fracturing of the plywood laminae and the
three years. generally sad condition of all of these works.

Johnson spent one year in Paris, where he was In 1956, all of Johnson's works were tumed
influenced by Soutine, the European Expres- over to the Harmon Foundation, anorganua-
sionist, and had several one-man shows. He tion that had long sought to make Johnson a rec-
then settled in the south ofFrance and painted ognized artist. When the Foundation closed in
while traveling extensively through Europe. He 1967 , over I ,100 paintings, drawings and prints
returned to the United States and set up a studio were given to the National Collection of Fine
in Harlem, but went back to Denmark in 1930 Arts, now known as the National Museum of
and married Holcha Krake, a Danish weaver American Ar| which is part of the Smithsonian
and ceramicist. It was during this time in Den- Institution.
mark that he painted many of his most boldly
beautiful works, depicting eruptive force scenes Since the collection was acquired, many of the
of harbors, f,ords, mountains and people, with Johnson paintings have been treated. This most
a wildly rich palette. Because poverty was al- recent show consists of works from his last ac-
ways with him, most of these paintings were tive years. Ten of the paintings on plywood sup-
executed on a coarse burlap feed bag material, port required treatrnent. Of these, four required
which is now in a very fragile and damaged removal of the back plies and major structural
condition. work with attachment to new supports. Four
others received localized structural treatment
After eight years of traveling and painting without removal of the original plies. The re-
through Europe and North America and with maining two paintings had been remounted and
World War II looming, Johnson and his wife treated 2l years ago with subsequent structural

232
failure. These required reversal ofthe old warping were ones with a total thickness ofjust
treatment and employment of the current one. over 3.5 mm: these thinner panels were also of
much poorer grade and fabrication quality, and
Some generalizations of the artist's materials contained knots and surface imperfections.
and techniques can be made. All l0 paintings
are in oil medium. Handling and application of For purposes of brevity, only one painting from
the medium vary from those where it is simple each of the three categories of treatment is
and direct to those where paint is thickly ap- discussed.
plied with brushed or palette knife impasto.
This variance ofapplication bears no relation to Local Structural Treatment:
subsequent structural condition. There is some Folk Family (1939 to 1944)
cracking, cleavage and paint loss on both thinly
and thickly applied surfaces, although generally This painting was on a 5 mm-thick plywood
the paint layers remain in good condition. support. Both sides of the support were primed
with a brown gritty paint layer. The paint was
By the time these paintings were done, Johnson generously applied with considerable impasto.
seemed to have abandoned the use of his more There was no varnish or surface coatins. The
traditional lead white and calcite srounds. condition of the support was poor, witlibroad
These panels are all primed with i pinkish- convex warps creating planar distortion. All
brown or dark-brown paint layer on both sides corners had delaminated and there were losses
of the support. This preparation layer is gritty in one or more of the ply layers at all corners.
on some paintings, and on others it is thinly The top horizontal edge was completely delami-
applied and does not obscure the texture ofthe nated and buckled and it was broken up by sev-
wood. eral vertical cracks and splits in the uppermost
half. The splits varied from 1.3 cm to l5 cm.
While most of the paintings had no vamish There were losses ofsupport and design layers
layer, several had a dull streaked, slightly dis- along the top edge as well. Despite the poor
colored coating that had been randomly ap- condition ofthe support, the paint layers re-
plied. On all, there was a substantial layer of mained in good condition. Although some verti-
embedded dirt and grime as well as fibers and cal hairline cracks were present in the upper
chips ofpaint stuck to the surfaces from other half of the painting, most of the paint was sta-
paintings that had leaned and pressed against ble and secure. The surface texture was rough,
them in storage. due possibly to the gritty priming layer, and the
paint was unevenly matte and glossy. There
It is in the choice of plywood supports for each were small damages and gouges randomly
of these paintings where a more direct correla- throughout, some surface abrasion, and numer-
tion can be made between the quality and type ous scattered edge losses. The surface had paint
of material and the resulting vulnerability and spatters and a substantial layer of heavily em-
ultimate damage sustained. The construction of bedded dirt and grime.
each of the plywood panels was balanced, in
that they consisted of an odd number of plies Treatment
(this is almost always true for plywood). The The surface dirt was removed with a 50% solu-
two-face plies were oriented in the same direc- tion of di-ammonium citrate in distilled water
tion while that of the core ply was at a right an- followed with a distilled water rinse. Overall
gle to them. (Note: There were only three-ply planar distortions were reduced by placing the
panels). The flanking-face plies were of the painting face up over moistened blotters, then
same kind and quality of wood, and presum- covering it with a sheet of glassine, a thick felt
ably had the same moisture content. However, blanket and furally a sheet of masonite and
those panels that sustained less distortion and heavy weights. The delamination along the up-
separation between plies were of a uniformly per edge was consolidated with hot liquid hide
thick three-ply construction. Their thickness glue (3l5 g) injected berween the pliei. and
was a total of 5 mm. Those panels with the then the edge placed between padded clamps
most severe delamination, buckling and

z))
for flattening. A filling compound of sawdust direction while that of the core ply was at a
mixed with CM Bond M3 [poly(vinyl acetate) right angle to them. All three plies were of the
emulsion] was used to replace the missing cor- same thickness and kind of wood. The total
ners and other sections. On the front ofthe combined thickness of plies was less than
painting these fills were subsequently covered 3.5 mm. The quality or grade of wood was very
with a layer of gesso. A thin brush coat of Solu- poor, each ply having numerous inherent de-
var Matte Vamish was applied as an isolating fects, such as knots and voids. There was over-
layer and the inpainting carried out with dry all warping of the panel as a unit and inter-ply
pigments and AYAC fpoly(vinyl acetate) resin] warping and buckling. (The condition of the
in ethanol over a base of watercolor. This painr panels is shown in Figures 2 and 3.) These
ing and the three others treated in the same way warps and buckles were running with the grain
will continue to be monitored for any further of each ply. The paint-layer face ply was com-
planar distortion or delamination, but it is our pletely split down the middle, the edges of the
hope that more drastic heatment will never split forming a cupped ridge at the apex. Four
be necessary. vertical splits, 2.5 cm to 5.1 cm in length, ran
from the bottom upward on the right side and
Major Structural Treatment: each individual split was severely warped in a
Going to Market (1939 to 1944) convex manner. There was complete loss of
face and back ply along the right side, up to
This panel painting, shown in Figure l, repre- 10.2 cm of image, exposing the buckled core
sents those paintings that were most severely ply. In several areas, the core ply had split and
warped, buckled, split and delaminated. There pushed up on the painted top ply, causing it to
was no question of a partial or remedial ap- split and sustain splintered losses.
proach to treatment. The fragility and thinness
of the combined plies made these paintings dif-
ficult even to tum and handle. Typical exam-
ples of damages common to each are described
in the following paragraphs.

The support was of three-ply construction, the

grains of the face plies extending in the same

Figure I Going to Market by William H. Johnson.

Raking light shows central split and complete loss oJ Figure 2 Detail of Going to Market sftowing wash-
face plies exposing buckled and split core plywood. board-like warping oJ vertical splits and ply loss.

zJ+
 cushioned surface and a large spatula used to
remove the back and core plies one layer at a
time.

Treatment of Remounted Paintings

The next two paintings to be heated were those
Figure 3 Cross-sectional view of typical inter-ply in which treatments carried out more than20
delaminationfound on all the paintings in varying years ago had failed (as a result of a number of
degrees. factors combined). Like those described above,
these older treatments had also involved
The thinly applied priming layer was a pinkish- removal of two plies, but unlike the cunent ap-
brown paint with a somewhat gritry texture. proach no flattening ofwarped or buckled areas
The paint was of moderate thickness with some was carried out. In the original heatment each
areas of brushed and palette knife impasto. painting had been mounted with a wax-resin ad-
There was random abrasion and small paint hesive to different supports. One painting was
losses, and a thin, streaked surface coating that adhered to a strainer-supported masonite board,
was gritty and uneven in its application. Dirt and the other to an aluminum-honeycomb
and grime were embedded in this coating as panel. The wax adhesive also served to fill the
well as on the paint surface, giving an overall voids where the warped original support did
gray appearance. not make contact with the new support. In some
places the wax was almost 5.5 mm thick. Fail-
Treatment ure of the bond between materials on these
The painting was cleaned and areas of insecure paintings had been accelerated due to an active
and semi-detached paint were infused with lending program that allowed the paintings to
Beva 3 7 I [poly(ethylene-vinylacetate)] and set travel to venues having inadequate temperature
down with heat. We decided that the plies of and humidity controls.
this painting support were too thin to attempt
delamination first, so the entire panel was pre- Treatment
flattened as a unit. The painting was placed These paintings were removed from their sup-
face down on a l0 mm{hick sheet of closed- ports and the wax-resin scraped offand leached
cell neoprene. A sheet ofglassine separated the from the plyr,vood with solvent until clean. Im-
neoprene and the painting surface. The back of perfections were sanded, and warps and buck-
the panel was misted generously with water and led areas were flattened with moisture and
blotters were laid on top. Plexiglass sheets were pressure.
placed over this and weighted. This process
was repeated over many days with misting, It was at this point in heatment that the
blotter changes, and gradually increasing Conservation Analytical Laboratory of the
weights. [n some areas of closely-spaced Smithsonian lnstitution was called upon to help
warps, large padded clamps were used. work out a new support system with careful
attention and debate given to the choices of
When the painting was as flat as seemed possi- adhesives.
ble and completely dry, it was placed face up.
The surface was then faced with Tosa Tenjugo General Treatment Considerations
Japanese tissue and wheat flour paste and when for Plywood Artifacts
this was dry, another facing with cheese cloth
was applied with an Acryloid 8-67 emulsion Degraded artifacts constructed of plywood pose
composed of l}ohwater,lo/o aerosol OT (a a vexing conservation problem that requires
clear liquid containing 10% diacef,l sodium careful consideration of several factors. If the
sulfosuccinate, water and mutual solvent), and problem with the artifact is not defined accu-
89% Acryloid 8-67 stock solution. The paint- rately, the proposed or executed treatments
ing was again placed face down on the same can range from those that are inspired and

235
successful to those that are based on ignorance can be considered expendable and the impor-
and therefore doomed to fail. tant lamina (i.e., that containing the painted
image), can be removed from the degraded
The major considerations are the history, tech- structure and remounted onto another substrate.
nology and character of the original raw mate- Altematively, the plywood support may be
rial and its subsequent utilization in the artifact. considered integral to the artifact and be reinte-
Klim's recent article on composite wood materi- grated with the important lamina into a single
als includes a review of plywood history.' An unified element. Both general approaches are
understanding of the nature of plywood is vital justifiable based on the specihc circumstances
to responding appropriately to its deterioration. of the treatment.

Next to consider is the artist's motive or intent In the case of the Johnson paintings, there were
in using the material in question. Why was this degrees ofdamage that varied from virhral de-
particular material selected rather than another? lamination and complete loss of structural integ-
There are several probable "correct answers," rity to minor, localized separations. In keeping
which take account of cost, convenience, avail- with this range of deterioration, a number of
ability of materials, uniformity, and stability of specific treatment options were formulated
materials. Usually, all of these factors play a within the two major directions already men-
role in the artist's selection of material, and ply- tioned. These treatment options were tempered
wood paneling is considered like all other art- by cost and time limitations and by the possibil-
ists' materials. Plywood was (and is) a cheap, ity that the paintings conservators conducting
easily worked, planar altemative to solid or the treatments would be more familiar, more
joined-wood panel supports for painters and comfortable and more skilled with manip-
also serves the more widespread needs of ulating paint and canvas than wood and the
cabinet-making and related crafts. adhesive systems best suited to wood. This pos-
sibility fostered a desire for a "low-tech" ap-
Another critical step for paintings on plywood proach requiring little woodworking expertise
supports is careful evaluation of the object: yet allowing for control over the process. By
how delaminated is the structure, how stable "low-tech" we mean that the process should be
is that which remains, and how did the damage straightforward with simple steps and with little
occur? chance for error in preparation and over the
course of the treahnent. In addition, there were
Finally, how does the presence or lack of ply- the usual concerns over "retreatability." Finally,
wood as a raw material affect the interpretation as would be expected in treating painted wood,
ofthe object? For three-dimensional objects, there were requirements for minimizing en-
such as fumiture or sculpfure, the answer is croachment by water and organic solvents into
pretty clear the plywood is an integral part the obj ect. Introduc ing water-based adhesives
of the object- and without it there is no artifact. into a hygroscopic material raised the risk of
For paintings on plywood, however, the answer extreme stress tlrough swelling and shrinking,
may be very different. Traditionally, art histori- and solvent-based adhesive systems have
ans and restorers have considered that the negative effects on the paint film and on the
painted image rs the artifact and the plywood is conseryators.
simply the support for that image. Therefore,
while the plywood substrate for a painting may Testing Treatment Methods with
remain important for ancillary reasons, it may Sample Panels
not necessarily be inviolable, as would be the
case with most three-dimensional objects In order to explore our ideas for treatment, a
constructed of plywood. group of mock painting-on-plywood panels
were created for sacrificial testing. A sheet of
Removal Versus Reintegration plywood similar to that used in the paintings
to be treated was purchased, cut into 60 cm
There are two main directions for treatment of squares and painted. These squares were then
delaminated plywood. The plywood support

236
subjected to extremely harsh environmental We narrowed our scope of options for new
conditions in order to induce the delamination supports to include only materials with the
observed in the Johlson paintings. Unfortu- same properties as plyvood. This was not only
nately, the quality of the plywood purchased because the original artifacts were made of
for the mock-up panels clearly surpassed plywood, but also because plywood has many
that used by Johnson. It is a credit to modern inherent benefits. These benefits, which have
adhesive and plywood technology that after been mentioned previously, include cost, avail-
I 5 cycles of 72 hours of submersion in water ability, ease of manipulation, the ability to
followed by 72 hours in the oven at 60"C, the work quickly or slowly and dimensional stabil-
panels were still in generally good condition. ity. To use these properties to their fullest, the
There was widespread cracking and cupping new supports were designed around high-
on the surface and some localized delamina- quality plywood.
tion, but nothing approaching the damage ex-
pected. The panels were nevertheless extremely Torsion-box Support Panels
useful in evaluating the physical manifestations
of our ideas. A number of treatment processes One of the earliest proposals we considered
using these panels are described below. was to remount the lamina onto a solid ply-
wood sheet. However, the weight of plywood
Remounting the Decorated Lamina thick enough to maintain planarity proved to be
excessive for safe handling and exhibition. We
In the instances where the painting support was then turned to the idea of improving the proper-
badly damaged or mutilated, remounting the ties of plywood by utilizing "torsion-box pan-
painted surface on a new support was a viable els." These have been used for some time in the
option. The main questions were: l) onto what fabrication of custom fumiture. Torsion boxes
material will the painting be mounted? and 2) are constructed of ribs and thin-facing laminae
how will the painting be adhered to the new that act together as a panel comprised ofa se-
support? ries of tiny l-beams. The weight of the support
panel is reduced greatly but the panel is still
Several supports were considered for remount- nearly as strong as a solid plywood substrate.
ing the painting once it was removed from the Our hollow-core torsion boxes were designed
remaining fragments of original substrate. and fabricated using marine/aircraft grade ma-
Much of the decision making for this option hogany and poplar plywood faces over wooden
was influenced by the earlier failure of a re- ribs (Figure 4). The structure ofthese support
mounting treatment onto aluminum honeycomb panels conceptually is remarkably similar to
that was used for one of the Johnson paintings. that of aluminum-honeycomb panels, that is,
Having little experience with aluminum honey- their strength, stiffrress and lightness are out-
comb and a dislike of superimposing materials standing and they can be built to any size.
of greatly dissimilar properties, we decided to (They were made as 60 cm squares for the
limit the choice of new sub-
strate materials to wood or
wood-like materials, which
were sympathetic to the origi- - /-a
nal fabrication. nature. and in- .' I

tent ofthe artifacts. The

paintings were on wood, albeit
a modified form of wood, and
in our opinion, the best solu-
..- id
tion to the problem was to ap-
1 s'ply 4 smm piywmd face
proach it as a wood problem 2 3mm x 1 ?cm mahogany srd ribs
3 1cn x 1 ?cm mahogany grd pe.rmet.r
we wanted to remain with a 4 3-ply 3 omm plywood face
system that was based on woodi
glued-to-wood. This led to.our Figure 4 A cutaway diagram of the torsion-box support panel
process and material selections.

zJl
mock-paintings.) The torsion-box panels had
the additional advantages ofbeing easy and
cheap to build and made from material very
similar to that of the original plywood supports.
In overall construction, however, the torsion-
box panel was quite different.

Construction of the Remounting Panels

Probably the most critical factor for assuring a
high quality panel is to make sure that the verti-
cal dimension of all ribs (i.e., the rib dimension Figure 5 Torsion-box panel, side view. Topface ply
determrning the thickness of the panel) is identi- is thinner than bottom ply, so that later adhesion of
cal. The easiest way to achieve this is by surfac- painted lamina to the top will bring the two face-ply
thichresses almost exactly equal.
ing a board using a power planer and cutting
thin stnps from this board for all rib-stock.
Since mahogany, Swietenia spp., has long been extremely quickly, precluding the need for
highly regarded for its dimensional stability, it clamping the ribs while they set. Interior ribs
was selected for the ribs and prepared to a thick- were cut from the 3 mm-wide strips and at-
ness of 12 mm. Interior ribs of 3 mm and pe- tached in the same manner to form a grid of
rimeter ribs 25 mm wide were cut from the ribs approximately l0 cm apart. (This measure-
12 mm{hick prepared stock using a table saw ment is not critical.) The top edges of the ribs
with a fine-tooth blade. and the underside of the 3 mm face were then
coated with the hot hide glue and the face was
In most small-scale wood construction the no- placed on the panel and weighted or clamped.
tion of "balanced lamination" is an important
one for the long-term stability ofthe object. As mentioned, these panels were remarkably
In other words, what is done to one side or strong and light. Before mounting the painted
component must be done to the other side. lamina, the 60 cm x 60 cm x 19.5 mm panels
With that in mind, one of the pl1'rvood faces weighed 3 kg. Strength was tested by securing
selected for the panel was 3 mm thick and the opposite edges ofthe panel and then loading it
other 4.5 mm. Later adhesion of the painted with approximately 80 kg at the center there
lamina from the mock-painting to the 3 mm- was no discemible deflection. -
thick face brought the total face thickness to
approximately 4.2 mm. The difference in thick- Similar support panels that substitute a paper
ness of the two faces is practically undetect- honeycomb for the internal rib grid are also an
able. (In Figure 5 a side view ofthe torsion box excellent option and provide nearly the same
is shown. Note that the top face ply is thinner strength characteristics without the need to actu-
than the bottom ply so that later adhesion of ally fabricate the interior of the panel. The pe-
the painted lamina will make both top and rimeter ribs are attached and the paper grid is
bottom facing plies approximately equal cut from a stock sheet to fit the opening formed
in thickness.) by the perimeter ribs, then glued as described
above. The perimeter ribs must be exactly the
For convenience, the panel was fabricated same thickness as the paper core.
using hot glue (315 g). The 4.5 mm face was
placed on a very flat work table and the perime- The Adhesive System for Remounting
ter ribs were rub-joined to it using full-bodied Paintings on Plywood
hot hide glue. (It is extremely important to
choose the assembly work surface carefully From an adhesion-strength perspective, the best
-
the panel will adopt the contour of that surface adhesive for wood is a protein adhesive, or a
during fabrication and maintain it afterwards. hot, animal-derived glue. It is easily prepared in
The new panel will not be planar if the assem- water, it is quite stable and is relatively easily
bly surface is not planar.) The hot glue gels reversed. One of the mock-paintings was

238
remounted successfully onto a thin plywood mock-painting lamina was placed immediately
support using hot hide glue. Nevertheless, we on the tissue and a vacuum was pulled, clamp-
were concemed about introducing substantial ing the entire unit in place. The result was ex-
quantities of moisture into the painted lamina. actly what we wanted a very strong bond
This, combined with the desire for minimal with the amount of water- reduced by about
use oforganic solvents, restricted us to heat- 75oh. This method was selected as the treat-
activated adhesives for the remounting. AYAA ment procedure for the Johnson paintings that
and Beva 371 were flocked onto separate sheets needed complete remounting (Figure 6).
of Japanese tissue paper, which were then used
as heat-activated "dry mount" adhesives. The One altemative to this process would be using
painted lamina was glued by the tissue interleaf hot hide glue with gel suppressants, such as salt
to the new support on a heated table with a or urea, to give the glue an indefinite working
mylar vacuum membrane. Some success was time. In this manner, the pressure of time limita-
achieved but the results were not satisfactory tions for assembly would be removed.
enough to select either ofthese adhesives
without fu rther ex perimentation. Reintegrating the Separated Laminae
We kept retuming to the notion that hide Three reintegration schemes were defined, at-
glue was the best adhesive if we could reduce tempted on models and found to be acceptable
the amount of moisture in the system. The solu- for this circumstance. (This does not imply that
tion was to manipulate the hide glue in a man- there were only three methods. These were the
ner similar to that used for the AYAA and ones tried that satisfied our requirements.)
Beva 371. However, since hide glue cannot be
flocked it was introduced to the tissue as a liq- The frrst method was the most straightforward.
uid and later used as a solid. Sheets oftissue Adhesive was injected or poured into the void
paper were laid flat on a sheet of mylar and pre- left by the delamination and the object was
pared hot hide glue was brushed onto the tissue clamped between release films and two flat
until it was thoroughly impregnated, and al- cauls until the adhesive had set thoroughly.
lowed to dry ovemight. Hot hide glue films can
be cast onto mylar, but the tissue functions as a The second method also involved the introduc-
means of handling the hide glue (and other ad- tion of liquid adhesive, although to our knowl-
hesives) more easily. It also acts as a release edge, the technique is specific to hot hide glue.
agent should the interface need to be removed This process is most useful on a cracked and
in a future treatrnent. cupped surface that has not significantly de-
tached. The hot hide glue is brushed generously
Using the tissue impregnated with hide glue onto the surface, and a transparent membrane
as a heat-only activated adhesive did not prove (e.g., thin polyethylene) is placed over the liq-
successful, but heat in combination with a light uid glue. A warm lamp shining on the surface
water misting of the tissue yielded a remark- keeps the glue fluid under the membrane. By
ably strong adhesive material. The support massaging the polyethylene film with the fin-
panel was placed on the heated work surface gertips, the glue is forced into the delamination
with the 3 mm-thick face exposed. This face voids (the film acts as a hydraulic membrane).
was sprayed lightly with water, and the impreg- Once sufficient glue has been squeezed into
nated tissue was then placed on the support the voids, the membrane is removed, the excess
face and treated in the same way. The glue cleaned offthe surface, and the object

1. Suction membrane
2 Parnted lamtna
3. Glue-rmpregnated trssue
4. 3-ply, 3.0 mm plywood face
5. 1 cm x 1 7 cm mahogany grid penmeter
6. 3 mm x 1.7 cm mahogany gnd nbs
7. s-ply, 4.5 mm plywood face

Figure 6 A cross-section diagram ofthe remounted paintings.

239
clamped as described above. The primary fac- treahnents on a heated vacuum table, on which
tor in choosing this technique is that it should a membrane envelope was partially prepared by
be attempted only on surfaces that are very adhering two edges to the table. The support
stable. panel was placed on the heated table, and once
the panel was warmed, the mounting surface
The third method involved a modified version was misted lightly with water from a spray
of the glue-impregnated tissue described above. bottle. The tissue was immediately put in place
The stiff, dry impregnated tissue was cut to fit on the panel and misted with water to reacti-
an individual void. It was then sprayed lightly vate the glue, and the moistened tissue was
with water and placed in a deep-freeze chest at smoothed out by hand. The painted lamina was
about -30oC. Once the tissue had frozen solid it then placed on the panel and the vacuum enve-
could be inserted into even deep voids like a lope taped shut. Once the air was evacuated,
stiff blade. The fiozen adhesive tissue sweated felt blankets were placed on top to retain heat
naturally as it warmed, and the condensed in the unit and to allow fuIl adhesive flow to
moisture combined with the thawed water to occur, the heat was turned offand the re-
reactivate the glue. The addition of gentle heat mounted painting allowed to cool slowly. The
followed by clamping pressure yielded an vacuum was left on until the glue had set. The
entirely satisfactory bond between the elapsed time from first misting to membrane
delaminated strata. evacuation was aboul four minutes.

Remounting the Johnson Paintings The remounted painting was removed from the
table and placed under gentle weight or clamp-
Once the treatment of the painting was under- ing ovemight (usually with "sausage" weights
way sufficiently so that the painted lamina was around the perimeter). This final step was prob-
completely stable, the support panel was con- ably overkill to a certain degree, but it did re-
structed to fit. The specifications were given to sult in the paintings being well-adhered to the
the museum cabinet shop, and the craftsmen new supports. The only problem we encoun-
did an excellentjob ofexecuting our design tered was that the combination of heat and vac-
(Figure 7). Ifthere is no access to in-house uum led to a too-rapid drying of the glue at the
fabrication, any competent commercial cabinet very edges of the painting. This was alleviated
shop can do the same. While the panels were simply by wicking small amounts of water at
being built, the hide glue tissue was prepared in the edge of the remounted painting. This imme-
the manner described previously. diately reactivated the glue and was followed
by gentle weighting or clamping.

w
w,wY

M&'w
Figure 7 During construction o/ interior gid oJ
torsion-box panel. (Panels were constructed in the
workshop at the National Museum of American Ar|
Smiths onian Institution. )

The treatment in remounting a painting is a

very rapid process that needs extensive and Figure 8 Going to Market after delamination of core
accurate preparations. We conducted all our and back plies and attachment to torsion-box panel.

240
Following the implementation of the new Museum of American Aft. Johnson, qui a peint
mounting system, the final aspect of treatment aux Etats-Unis et en Europe entre 1926 et 1946,
was the compensation of the losses (Figure 8). a, en ddpit de saformation acadtmique, choisi
de peindre suivant un style quasi-vernaculaire.

Conclusion Probablement pour des raisons d'tcononie, entre

1939 et 1942, il a peint nombre de ses tableaux
Plywood is certainly a modern material by defi- sur des panneaux de contreplaqu2. Il a dil €tre
hospitalisi vers lafin de sa vie, si bien que ses
nition, but forhrnately for us, its technology is
euvres sont tombtes dans I'oubli pendant une
well known. In addition, the technology of as-
dizaine d'anndes, soit durant son hospitalisation
sembling wood veneers is one familiar to most
puis, d cause d'un litige relatif d la disposition de
wood conservators. Ironically, this modern ses biens, pendant la ptriode qui a suivi sa mort.
material may be best manipulated using ancient Ses tableaux ont probablement passt certaines de
craft techniques, even when treating degraded ces anndes sur les quais de laville de New York.
artifacts. By combining traditional craft tech- Leur exposition prolongde d un environnement
nique with the knowledge of this modem mate- extiieur non contr6ld a eu, sur les peintures et
rial, conservation treatments can be undertaken sur leurs supports, des effets semblables d ceux
with considerable confidence. d'un essai de vieillissement acctldrd, de sorte que
les diverses couches du contreplaqud ontfini par
Acknowledgement se ddcoller, par se boursoufler et par se casser.

Photographs by the authors, courtesy ofthe Or, la mise sur pied prochaine d'une exposition
National Museum of American Art and the des rpuwes de Johnson a ament les spicialistes
Conservation Analytical Laboratory, de la restauration du National Museum of Ameri-
Smithsonian Institution, Washington, D.C. can Art et du Conservation Analytical Laboratorlt
de la Smithsonian Institution d btudier les divers
traitements qui pounaient €tre appliquts d ces
R6sum6 peintures. Plusieurs methodes sptcialement
adaptdes au traitement du support des peintures
Lq conservution de peintures sur supports de ont ainsi dtb conques, qui peuvent aussi €tre
contreplaqud ddcollis appliquees d toute sttuchre de bois constitude
de plusieurs couches. Deux options principales
Attfur et d mesure que s'est dtveloppde la pro- s'offraient alors, d savoir remettre ces objets de
duction industrielle de matdriata de bois modilit musde entidrement en etat ou enlever la couche
au ddbut du X,f si?cle. les panneaux de contri- portant la peinture.
plaqui et d'autres matiires mixtes en sont venus
d se substituer au matiiel d'artiste classique. Et L'article porte sur les diverses questions d'ordre
ce pincipe vaut tout particuli,irement pow'le do- tthique et technique qui peuvent se poserface
maine de la peinture, oi les artistes prdftraient le aux objets de musde et d leur traitement. Il fait en
contreplaqui, moins cher et plus uniforme et ver- outre 6tat de chacune des pincipales options qui
satile, aux suppot"ts de bois massif pour exicuter ofi eft envisagtes et de ses variantes subsdquen-
leurs peintures sur pann eau. Auj ourd' hui, soit tes, tout enfournissant une illustration des
environ un demi-siicle plus tard, ces supports traitements qui ontfinalement ttd appliques.
ont commenct d se dtttiorer, ce qui n'est pas
sans poser toute une sdrie de problimes inddits
aux responsables des euwes peintes sur de
Reference
tels matdriaux.
l. Klim, S. "Composite Wood Materials in
Twentieth-Century Furniture," AI C Wooden
Des probldntes de cet ordre se sont posts pour la
premidre fois dans le cas d'une sirie de peintures Artifocts Group, Richmond meeting 1990
de I'artiste amtricain William H. Johnson (1901-
(Washington, D.C.: American Institute for
1970) faisant partie de la collection du National Conservation. 1990).

241
Chicken Bones and Cardboard: The Conservation
of a Collection of Eugene Von Bruenchenhein's Art

Anton Rajer Emil L. Donoval

Madison, Wisconsin John Michael Kohler Arts Center

U,S.A. Sheboygan, Wisconsin
U.S.A.

Abstract Solvent cleaning tests on most of the paintings

indicated that the thick, discolored resin varnish
Eugene Von Bruenchenhein (1910 applied by the artist had a solubility parameter
-1983), a self-
trained artist, created artfromfound objects and close to that of the paintfilm. Varnish removal
howehold materials. Living in Milwaukee, l(is- w,as not attempted because there was no viable
consin, he supported himself and his wife, Marie, system to remove the varnish without possible
on the modest salary ofa baker and created his damage to the paint and because ofthe tight
works at night and on weekends. sc heduling of exhibitions. Instead, the varnished
paintings were coated with a thin laver of Acry-
His materials ranged from the typical oils, loid F- 1 0 and Acryloid 8-67 to retard further yel-
enamel, watercolors -
to the unconventional lowing of the varnish. Those paintings that had
-
chicken and turkey bones, oven-baked clalts, card- - not been varnished originally by the artist were
board and plvwood, house paint, and concrete. left unvarnished. All paintings were cleaned with
Many other mateials that he used, such as PVAC a mild soap solution.
glue, are unique to the 20th century.
Treatment of the sculptures rangedfrom light
After Von Bruenchenhein's death, more than cleaning to repairs and inpainting. Repairs were
300 of his 1,500 total works were acquired by the made with PVAC emulsion. Inpaintingwas done
John Michael Kohler Arts Center (JMKAC) in with acrylics and watercolors. Some items re-
Sheboltgan, Ll'isconsin. This paper discusses the quired only cleaning.
Center's conseryation efforts on a collection of
Von Bruenchenhein's works. The photographs needed the least attention. The
anist hadfrequently photographed his wife, devel-
Von Bruenchenhein's artfalls into three catego- oping black-and-white montage prints in the bath-
ies: paintings, sculptures, and photographs. Be- room sink of their home. Hundreds of these pho-
fore treatment, all objects were inventoied and a tographs from the I 950s and I 960s were dusted,
conservation survey was prepared. The objects inventoried, and placed in storage envelopes.
were cleaned superficially by dusting, as the art-
ist's home had been heated by a coal stove and The conservation challenges posed by this collec-
all objects were covered with a thick laver of tion were overcome with curatorial assistance, a
soot. Then, conservation priorities were estab- minimalist approach to intervention, and respect
lishedfor the Center's collection with two major for the artist's intent. The treatment of this collec-
goals in mind: to stabilize the collection and to tion underscored the realization that today's con-
prepare most of itfor the JMKAC and Milwaukee servation methods do not have all the answers
Art Museum exhibitions. Severely damaged ob- in caingfor such a unique collection bv a
jects were set aside for treatment at afuture date. 20th-centurv anist.

z+J
Introduction Milwaukee Art Museum in 1988. This paper
discusses the conservation of Von Bruenchen-
Twentieth-century art historians have hein's works in the JMKAC collection.
documented and described a type of isolated
creative individual known as the "naive" or Materials and Techniques
"outsider," whose artistic style is highly indi-
vidualistic and divorced from major artistic Working outside the traditional aesthetic circles
movements. One such individual, Eugene Von usually associated with art schools and the artis-
Bruenchenhein (1910 - 1983), a self-trained art- tic establishment of his time, Von Bruenchen-
ist, created his works from found objects and hein supported his art both with his small salary
household materials. He was self-taught and and his personal resourcefulness.
relied on his family and friends for supplies and
moral support. He received minor recognition Von Bruenchenhein fashioned artist's tools
for his relentless creative efforts during his life- from a variety of items, including combs,
time. Despite his poverty, a lack of recognition, leaves, sticks, and other found objects. He
and many other obstacles, Von Bruenchenhein made his own paintbrushes from tufu of his
created more than 1,500 artworks over a wife's hair, bound with sfring and glued into
40-year period. He also wrote poetry and es- the end of plastic soda fountain straws, which
says on the nature of art, life and the universe he then painted with aluminum paint to give
and developed theories and inventions he them the appearance of commercially made
hoped would aid humanity. brushes.

Von Bruenchenhein lived in Milwaukee, His choice of materials likewise reflected his re-
Wisconsin, and supported himself and his wife, sourcefulness. These ranged from typical ones,
Marie, on a modest salary, creating his works such as oils, enamel, and watercolors, to uncon-
in the evenings and on weekends. He worked ventional ones. Many of the materials he used,
at a local bakery for many years, and according such as PVAC solvent-based model airplane
to rumor took an early retirement because of glue and PVAC white glue, are unique to the
health difficulties. His income had been modest 20th century. Other materials included chicken
to begin with; his retirement left him and Marie and turkey bones, from which he constructed
nearly penniless. miniature thrones and towers (Figure l); oven-
baked clay, from which he fashioned crowns
He had no studio, pursuing his artistic vision and pots in the shape of animals; cardboard and
in the living or common area of his tiny home. plywood, from which he made painting sup-
He never exhibited or sold a work in his life- ports and non-functioning musical instruments;
time. He was regarded in his neighborhood as and house paint, with which he created paint-
a strange man and some of his artworks, such ings and painted some of his sculptures and
as the large concrete masks he installed outside concrete pieces, which he fashioned into
the house, were periodically vandalized. Never- bas-relief sculptures.
theless, he continued to create his works with a
persistence and dedication that might be best A few crowns were fabricated from glass
characterized as obsessive. His unique vision Christmas ornaments (one of the few instances
and arrangement of materials qualify him as in which Von Bruenchenhein used ready-made
an artist. objects in his artworks) and twisted aluminum
salvaged from beer cans. Miscellaneous items,
After the artist's death, and following the settle- such as a painted and incised fiberglass plaque
ment of his estate, more than 300 of his works and Indian arrowheads, chipped out ofbottle-
passed into the collection of the John Michael glass, were also among the oeuwe.
Kohler Arts Center (JMKAC) in Sheboygan,
Wisconsin. The Center sponsored the conserva- Inventory Phase
tion of Von Bruenchenhein's collection, a book
about the artist, and an exhibition that was After acquiring the collection, the
shown at the Center in 1984 and at the JMKAC removed it from the artist's home

244
 Treatment of the collection took about
l8 months. It was conducted by a team compris-
ing a conservator and three technicians, who
carried out the work in phases, which corre-
sponded with the three categories of works.

First, the surfaces of all the works were cleaned

by dusting with a soft brush and canned air.

All paintings were cleaned with mild soap and

water. The artist had chosen to protect some
paintings by placing a layer of newspaper be-
tween them. Where newsprint stuck to the paint-
ings, the conservation team dampened it with
water before attempting to remove it. Flaking
paint was stabilized with BEVA 371. Inpaint-
ing, where necessary, was accomplished with
acrylic emulsion paints and watercolors.

Solvent cleaning tests on most of the paintings

indicated that a thick resin vamish had been
Figure I Chair by Eugene Von Bruenchenhein.
applied by the artist. The varnish was also dis-
Chicken and turkey bones, glued and painted.
colored. Unfortunately, this vamish had a solu-
6 3/4 in. x 4 I/2 in. x 4 in. (17.1 cm x I1.4 cm x
10.2 cm) No. 84.V8.8.68. Photo counesy of the bility characteristic close to that of the paint
John Michael Kohler Ans Center. film. Varnish removal was not attempted be-
cause there was no viable system to remove the
vamish without possible damage to the paint
for inventory and conservation. Because the and because ofthe tight scheduling ofexhibi-
Von Bruenchenhein home was heated with tions. Instead, the vamished paintings were
coal. the arfworks. like the interior of the house coated with a thin layer of Acryloid F-10 and
itself, were covered with a fine layer of soot. Acryloid B-67 n the hope that the synthetic re-
All objects were inventoried and assigned a sin would retard further yellowing of the vamish.
number according to the category they
represented. Out of respect for the artist's intention, the
paintings that had not been varnished originally
During the inventory phase, an informal conser- were left unvarnished.
vation survey was prepared for the entire collec-
tion. Many objects, such as the bone towers, The conservation team attempted to duplicate
were extremely fragile and showed evidence of the artist's original surface effects using the
having been repaired by the artist himself. As Acryloid F-l 0/Acryloi d 8-67 varnish mixture.
mentioned earlier, many of the paintings were
covered with soot, as well as a thick, discolored A five-degree scale of surface texture/reflectiv-
resin. Fortunately, the artist's photographs ap- ity was developed, ranging from gloss to matte.
peared stable; he had boxed these himself.
No gloss meter was used; the five categories
Conservation and curatorial priorities were es- were subjectively constructed by determining
tablished to stabilize the works, to prepare the five grades of gloss/matte on Von Bruenchen-
objects for two exhibits, and to publish a book hein's paintings. The categories ranged from
on the collection. high gloss (like a plate of glass) to none at all.
Vamish formulas were designed to reproduce
Treatment the degrees. The surface ofeach original paint-
ing was rated according to this scale, and the
Von Bruenchenhein's art falls into three catego- conservation team duplicated the appropriate
ries: paintings, sculptures, and photographs. effect.

245
Research on the suitability of the new cleaning Several details of Von Bruenchenhein's paint-
methods, such as the Wolbers' techniques (i.e., ings are noteworthy. First, he serialized his
utilizing surfactants or indushial enzlmes in paintings, often incising a number and his sig-
gel systems), is still in progress. Cross-sec- nature into the painting while it was still wet.
tional microscopy has revealed that, for most This is unusual because it breaks the estab-
paintings analyzed thus far, the artist applied lished practice of signing a painting when fin-
thick, white grounds and followed these with ished, not while in progress. And, many of his
applications of very thin layers of paint. Thus paintings from the 1950s deal with the subject
the paint layer was applied in a fashion much of nuclear holocaust. Swirling, kinetic mush-
lke a glaze. Von Bruenchenhein frequently in- room clouds, some with demonic faces, as
cised this top layer, in this way achieving a shown in Figure 2, ascend into glowing
greater depth and luminosity. atomic skies.

Previously unframed paintings in the collection Treatment of the sculptures ranged from light
were framed, with curatorial input as to the cleaning to repairs and inpainting. Broken clay
style and aesthetic of the frames Von crowns, animal pots, and objects fabricated
Bruenchenhein himself had made. Color from chicken or turkey bones were repaired
schemes were selected to harmonize with with PVAC or cyanoacrylate adhesives.
those in the paintings.
Of all items in the collection, the photographs
required the least attention. The artist had fre-
quently photographed his wife (Figure 3), de-
veloping the black-and-white montage prints in
,
the bathroom sink of their home. Hundreds of
these photographs liom the 1950s and 1960s
-i*a:
a*
rd' &'-*
.#'
,dr
f'
nr
da

*:

Figure 2 Untitled # 673 by Eugene Von Bruenchen- Figure 3 "Portrait of artist's wife Marie",
hein, 1958. Oil on cardboard. 28 1/4 in. x 22 l/4 in. Eugene Von B ruenchenhein. Gelatin-silver print,
(71.8 cm x 56.5 cm) No. 84.V8.p.401. Photo (21.9 cm x 13.2 cm) No. 84.V8.PH.2696. Photo
courtesy of the John Michael Kohler Arts Center. courtesy of the John Michael Kohler Arts Center

246
were dusted, inventoried, and placed in storage
envelopes. Among the photographic collection
employait
- telle la colle de polyac,ltate de
vinvle (PAV) pour la confection d'avions mi-
are some prints that were hand tinted by the niatures sont des produits particuliers au
artist. s:f siicle.-

Conclusion ) la mort de l'artiste, plus de 300 de ses

I500 euwes ont 6td acquises par le John
The art of Eugene Von Bruenchenhein has
Michael Kohler Arts Center OMKAO de She-
boyan (Wisconsin), et la prtsente communication
posed significant curatorial and conservation
challenges over the past nine years. Conserva- fait dtat des efforts qu'a deployes le JMKAC pour
conserver ces euvres.
tion efforts have been grounded in respect for
the artist's intent and have been marked by Les rpuwes de Von Bruenchenhein se classent
minimal intervention. In short, it is clear that to- parmi trois cattgories : la peinture, la sculpture
day's conservation methods cannot solve all the et la photographie. Avant d'amorcer le traiternmt
problems raised by this unique collection. On- de conservation, on a dressd un inventaire de tous
going conservation research has raised possibili- ces objets, et prdpard une 4valuation des traite-
ties, such as the use of the Wolbers' techniques ments qui leur seraient appliquis. La surface des
for cleaning paintings, which may prove useful objets a ensuite tti ddpoussidrde car, comme la
in the future conservation of the collection. maison de I'artiste avait dti chauffee au bois, ils
itaient couverts d'une dpaisse couche de suie. En-
Von Bruenchenhein's unique contribution to fin, un ordre de piorite a 6te 2tabli pour le traite-
20th-century art is finally being recognized and ment des objets de la collection du JMKAC,
his visionary legacy preserved for future stu- traitement qui aurait deux grands objectifs, d sa-
dents ofthe "naive" or "outsider" aesthetic. voir stabiliser, dans un premier temps, la collec-
tion et prdparer, dans un second temps, le plus
grand nombre possible de ces @uvres en wte de
Acknowledgement leur exposition auJMMC et au Milwaukee
Art Museum.
The authors thank the John Michael Kohler
Arts Center for helping them prepare this paper Les essais de nettoyage avec solvant effectuds sur
and for providing the photographs ofEugene les tableaux ont montrd que, dans la plupart des
Von Bruenchenhein's work. cas, I'tpais vernis de resine ddcolord qu'avait
appliqud I'artiste prdsentait une solubilitd appro-
R6sum6 chante de celle de la couche picturale. On n'a
donc pas tentd de ddbarrasser ces @uwes de leur
Les os de poula a Ie carton ou Ia vernis car on ne disposait alors d'aucune tech-
consewation d'une collection d'tpuvres nique qui aurait permis de le faire sans risque
d'Eugene Von Bruenchenhein pour la couche picturale, et d I'interieur de
l'4chtancier prdvu pour leur exposition. On a
Eugene Von Bruenchenhein ( I 9 I 0- I 98 3 ), donc plut6t optt pour I'application, sur les ta-
artiste amelricain autodidacte, crdait des (Euvres bleaux ainsi vernis, d'une mince couche d'Acry-
d'art d partir d'objets trouvtis et de produits loid F-10 et 8-67 pour retarder leurjaunissement
domestiques. E tabli d Milwaukee (llisconsin), ultdrieur, tandis que les autres, non vernis d
il assurait sa subsistance et celle de safemme I'oigine par l'artiste, ont 6td laissts tels quels.
Maie grdce d un modeste salaire de boulanger, Tous ont nianmoins dtd nettoyds avec un
et il s'adonnait d son art les soirs et les fins savon ldger.
de semaine.
Les sculptures ont, pour leur part, fait I'objet
Pour exicuter ses(Euwes, il avait recours tant6t de niveaux variables de traitement depuis le
d des matdriaux usuels peinture d I'huile, -
simple nettoyage jusqu'aux rdparations ou re-
6mail et aquarelle
-
tant6t d des matiires plut6t touches. Les riparations ont 6td effectudes avec
inusittes
-
os de poulet ou de dinde, argile cuite, une tmulsion de PAV, et les retouches, avec de
-
carton, contreplaqui, peinture domestique et I'acrylique et de I'aquarelle. Certaines@uwes
bdton. Nombre de ces autres matidres su'il n'ont exigi qu'un liger nettoyage.

247
 le
Ce sont les photographies qui ont exigt I'aide de conservateurs et en adoptant une appro-
pho-
moins de soins particuliers. L'artiste, qui che qui, fondde sur le pincipe d'un minimum
ses
tographiait souvent safemme, dtveloppait d'interyentions, respecte I'intention de I'artiste.
montage
dpreuves en noir et blanc destindes au Le traitement de ces rpuwes a par ailleurs permis
de
dans l'6tier de sa salle de bain. Des centaines de constater que les techniques de conservation
itd
ces photographies, des annies 50 et 60, ont actuelles ne permettent pas toujours de risoudre
enveloppes
dtpoussidrdes et glisstes dans des toutes les dfficultts que pose la preservation
avant d'€tre mises en rdserve. d'une collection d'euvres du Xf sidcle
ausst untque.
Les probldmes de conservation qu'a posds cette
collection ont dtd rdsolus enfaisant appel d

248
Ensuring a Future for Our Present High-Tech Past: Lessons
from the ENIAC for the Conservation of Major Electronic Technology

Jon Eklund and Beth Richwine

Division of Conservation
Smithsonian Institution
I(ashington, D.C.
U.S.A.

Abstract consewation and restoration aspects of the

ENIAC finishes will be presented in this paper, as
In March of 1990, the National Museum of well as the historical implications of the work.
American History opened the largest technology
exhibition in the Smithsonian's history: The Introduction
Information Age. The largest and one of the most
histoically significant of the objects in the exhibi- In March of 1990, the National Museum of
tion was a partial re-creation of the first success- American History NMAH) opened the largest
ful, electronic, general-purpose computer : the and most expensive technology exhibition in
Electronic Numerical Integrator And Computer the Smithsonian's history: The Information
(ENIAC). Occupying several hundred cubicfeet
Age.Behindthe 1,300 square metres (14,000
and weighing several tons, the I 5 exhibited units
square feet) of communications and computer
varied widely in their conservation problems.
history developed by the curatorial team was a
Choices among classical inpainting, complete
massive effort by the NMAH Conservation
repainting and component replacement had to
be made according to such competingfactors
Department to preserve and prepare some 750
as exhibit schedule constraints, consewation historic artifacts and a similar number of graph-
standards and experiments on the behavior of ics from America's high-technology past. Of
available modern versions offinishes. Before these, the largest and one of the most signifi-
conservation was completed some l5 people, cant was a partial re-creation of the first suc-
including the object's curator, had put in over cessful, electronic, general-purpose computer:
180 hours oftreatment under the coordination of the Electronic Numerical Integrator And Com-
a conservator who is a coauthor of this paper. puter (ENIAC). Occupying several hundred cu-
Such close collaboration between curators and bic feet and weighing several tons, the units
conservators is not as common as it should be. Be- varied widely in conservation problems. Before
cause of the joint effort here, however, informa- conservation was completed, some l5 people
tion about this artifuct that came to light during had worked on the project under the supervi-
the c ons ervati on effort stimulated furt her carato-
sion of Conservator Beth Richwine of the
rial research on the wrinklefinishes used on the
Smithsonian Division of Conservation. The va-
ENIAC and research on other electronic instru-
riety of conservation solutions chosen were a
ments ofthe 1940s and 1950s. Thesefinishes
produced a particular "look" and'feel" to these function of exhibit schedule constraints as well
as conservation standards and research into the
instruments, which were noticeably different from
both their predecessors ofthe 1920s and succes- paint technologies used for electronics during
sors ofthe post-Sputnik era. The analysis, the era of World War II.

249
Conservation always has an intrinsic historical tables for World War II artillery pieces. Thus
component; even if the objects to be treated the ENIAC, like so many other technological
have little historical significance, the conserva- achievements over the centuries, was encour-
tor must know something of the object's spe- aged and funded by the govemment, and
cific history as an artifact (what it was made of, specifically the war department.
how it was made and what its present condition
is) in order to properly treat it. In view ofthis, The problem of aiming artillery is complex.
it is surprising that there have been so few di- There is a specific path called a trajectory for
rect collaborations between curators and conser- each projectile or firing. The trajectory varies
vators in muselrm projects that study or exhibit for each type of loading, shell, gun, and each
artifacts. ln the case of the conservation of the angle that the gun is tilted (e.g., in hard or soft
ENIAC for The Information Age, however, the ground in abreeze of "x" miles per hour and a
strong time pressures and the importance and barometric pressure of "y" mm of mercury).
complexity of the artifact made such a collabo- During World War II each of these trajectories
ration particularly useful, and a member of the required a complex set of calculations (Figure
Smithsonian curatorial team, Jon Eklund, Cura- l). With aptitude and intensive training, and the
tor of the History of Information Technology use of an electrically driven mechanical calcula-
and of Chemistry, elected to devote much of his tor, a "computer" (i.e., a person who computes)
efforts to working with the Division of Conser- could calculate one ofthese paths or trajecto-
vation in this task. In turn, information about ries in about 20 hours. With the many develop-
this artifact that came to light during the conser- ments in new guns and shells in World War II,
vation effort stimulated further curatorial re- the Ballistic Research Laboratory of the U.S.
search on the finishes used on the ENIAC and Army was falling further and further behind in
other electronic instruments of the period. meeting the demands for tables. Faster methods
These finishes produced a particular "look" and were needed.
"feel" to these instruments that were noticeably
different from both their predecessors ofthe
1920s and successors ofthe post-Sputnik era.
The analysis and the treatrnent of the ENIAC 5ilNfftI FIINIEI TRI|JECIITE
finishes will be presented in this paper, as well
as the historical implications of the work. [|lunE rilRmilt IlrE iln
First we will outline the historical background
of this importantartlfact, then say something
(n'-u)3+-uY (it.i'J
about its history and situation as an object from
a curatorial and conservation standpoint. We *(03-Y)i-il+t=0,
will follow this with a discussion of the treat-
ment strategies and particularly about the con-
servation decisions that had to be made singly 3+ - 3-t= o
andjointly by the conservator and the curator. tN THEsf tqJATtoNS, Ar ANy PotNT (x,v)
IN SPACT
History u= CCI{POi,|ENT 0F VELOCITY 0F AIR lN x DIRECTION

v = COi{POffNT ff VELOCITY 0i AIR lN I DIRECTION

Surprisingly, perhaps, most early computa-
tional devices were not intended for daily use. d, THE VELOCITY tr SCUrc, lS GIVEN BY

The motivation and end product for most such \T dP

^2 _E-J
machines were tables of numbers to be used in
P= PRESSURE
hand calculations. The best known examples
are the tables of logarithms and the trigonome- = DENSITY
'
try tables long remembered by the pre-elec-
tronic calculator generation. Table production Figure I An example of a dilferential equation used
was the motivation for the ENIAC. The tables by the ENIAC to calculate a trajectory'.
for which people lusted in this case were firing

250
In the late thirties and early forties, an electro- electron tubes were unstable and changed their
mechanical calculating device called a dffiren- working characteristics over their lifetimes; it
tial analyzer was developed and built at the was a challenge to get a dozen to work together
Massachusetts Instifute of Technology in Cam- over a long period of time. For the ENIAC, this
bridge, Massachusetts, and a similar one at the meant great care had to be taken to match tubes
Moore School of Electrical Engineering, a part wherever necessary and to watch carefully for
of the University of Pennsylvania in Philadel- signs of change or malfunction. In the end, the
phia. It could solve a trajectory using about original ENIAC had 40 rack units plus three
20 minutes of calculating time and perhaps rolling "function tables" on wheels. It was ar-
half of that to do a simple change or two for the ranged in a "IJ" shape in a large room (Figure
next path. Though an impressive achievement 2). In November 1946, ENIAC was disassem-
providing a considerable savings in time, it still bled and moved to Aberdeen Proving Ground.
wasn't fast enough. Strictly speaking, World War II was over be-
fore ENIAC produced many ballistic tables.
By 1942, a physicist named John W. Mauchley It was used on post-war calculations for the
was on the staff of the Moore School and had H-bomb for the war department. On the civilian
become friends with a brilliant graduate stu- side, it produced calculations for weather fore-
dent, J. Presper Eckert. They had many discus- casting models, number theory, and fluid flow
sions about the possibilities of building in hydrodynamics. After ENIAC was moved to
computing machinery out of ulha-fast elec- Aberdeen it was put to work on hring tables
tronic components. Mauchley even wrote a and produced them for the Korean War. By
short paper entitled "The Use of High-Speed that time, it could calculate a path in 20 sec-
Vacuum Tube Devices for Calculating." He onds. ENIAC was used until October 2,1955
theorized that such a machine could calculate when it was shut down and gradually disman-
paths every 100 seconds. tled. All told, nearly 30 of the rack units sur-
vived and most were donated to the Smith-
In early 1943 the U.S. Army sent out a mathe- sonian in the late 1950s and early 1960s.
matician, Lt. Herman H. Goldstine, to the
Moore School and other research installations Physical Description
associated with the Army's Ballistic Research
Laboratory program. His task was to see if Fourteen of the fifteen ENIAC electronic
there were machines or ideas for speeding up rack units in the exhibition are vertically ori-
the process of making firing tables. Hearing ented and generally similar in appearance
of Mauchley and Eckert's work, Goldstine re- (Figure 3). Each of the vertical units is
viewed their ideas and arranged
for a presentation at the Aber-
deen Proving Ground in Mary-
land. Their project was
approved, and construction of
an electronic computer started
May 31, 1943 at the Moore
School with Mauchley in
charge ofthe general or concep-
fual design and Eckert in
charge ofthe designs for the
specific circuits. The ENIAC
design grew fiom a device that
used about 500 electron tubes
to a completed machine that, as-
toundingly, used close to
18,000 electron tubes. Even the
lower number of 500 was phe-
Fisure 2 ENIAC as it looked in the 1950s at Aberdeen Provins Ground.
nomenal for that time, since

2s1
approximately 213 cmxJ6 cmx 61 cm deep 6l cm to 183 cm (2 ftto 6 ft). Called "buses,"
(7 ftx2.5 ft x 2 ft). Several of these have hun- these were used to make connections among
dreds of antique vacuum tubes attached in the the rack units during operation of the ENIAC
back. The fifteenth unit, a "function table," and were stacked across the bases with cables
is horizontally oriented and measures and plugs leading to the units.
183 cm x 30 cm x 152 cm (6 ftx I ft x 5 ft).
It contains over 900 knobs. The units vary in The Kyoto Exposition
weight from about 68 kg (150 lb) to 136 kg
(300 lb), depending on the types of fittings The first conservation treatment on the ENIAC
contained in the rack. was not done for The Information Age exhibi-
tion. In the fall of 1987, five ENIAC units were
sent to Kyoto, Japan, for a cultural exchange
festival, the World Expo of Historical Cities.

One of the most noted features of the ENIAC

was the large number of tubes (Figure 4). For
the original exhibit in the NMAH, the curato-
rial decision was to show the vacuum tubes.
However, for the Kyoto trip, the tubes were too
fragile and rare to travel. Every tube had to be
taken out by hand and packed separately and
then reinstalled when the pieces were retumed
from Japan. In 1988 these pieces were moved
'l'-'..x:*""j
fr -o-r.- {$
again and put into storage when the Math Hall
was taken down for buildins renovations.

ll)
*:
L:

Figure j Intedace unitforfunction table#3, before

consentation. The non-original "feet" were added to
allow it to stand by itself.

The units are made of mild steel, but there are

also knobs, sockets and other accessories on the
units that are made of smooth plastics. The ma-
jority ofthe knobs and sockets appear to be a
phenol formaldehyde (Bakelite) plastic. Behind
the knobs, small plates of an unknown soft,
white metal are painted flat-black or dull-gold
with white or black letterins. These indicate the Figure 4 Looking over afew hundred ofthe
name of the unit. ENIAC's 18,000 elecnon tubes in the late 1940s is
Army liaison Lt. Herman H. Goldstine, who origi-
In addition to the large units there are numer- nally brought the govemment's attention to the
ous horizontal flat units in three leneths from Eckert-Mauchley proj ect.

252
The conservation on the five units sent to Japan vertical rack units no longer had their original
taught much about the ENIAC to both curato- bases new ones had to be made so that the units
rial and conservation staff. The decisions made could stand alone with stability.
at that time meant that several of the units were
in rather good shape and all but ready to install B) Gross Damage
in a new exhibit. Thus this early project served Dents, Bends, and Torquing
as a model for the later and larger effiort to There were scattered dents and bends in the
re-create a portion of the ENIAC in Zle metals on some pieces. In the past, when many
Inform atio n Age exhibition. of the units had been moved, the frames had
become torqued*. Because of this many of the
The Washington Exposition screw holes in the panels no longer lined up
exactly with screw holes in the frames.
In May 1990, The Information lge exhibit was
to open at NMAH. To give visitors a clear Missing Parts snd Picces
sense of the size of the early mainframe com- There were many missing screws in the frames.
puters, the design called for a walk-through re- There were also more than l0 missing upper
creation of about 40 percent of the ENIAC panels and three missing grills from the frames'
using l5 of the surviving units. In January of lower halves that needed to be replaced. The
that year, the five ENIAC units that had gone to missing panels were primarily flat in the center
Japan along with the l0 more that were taken with side flanges that were bent over on two
from one of the Smithsonian off-site storage ar- sides to fit between the two major vertical
eas were gradually brought down to the Objects supports. The original grills measured about
Conservation Laboratory for treahnent. The 6l cm x 76 cm(2 ftx2.5 ft).
large size and weight of the units made their
moving and handling diffrcult, but the prob- C) Surface Damage
lems were too complex and the storage site too Paint Losses and Instability
confining to try to conserye them in the field. The painted surfaces of the major elements
from the original ENIAC were covered with a
Problems black wrinkle paint. Samples of this paint, ana-
lyzed using infrared spectrometry and gas chro-
The units varied widely in the conservation matography, were found to contain an alkyd
problems they presented. Often the solutions resin (Figure 5). Alkyd resin is made by chemi-
had to be worked out between the department cally combining a natural ester-based oil with
of conservation staffand the curatorial staff a resin. The wrinkle finish was first developed
Problems included the following: during the 1920s but did not actually start to
become popular until the early 1940s. The two
A) Overall structural instability of the major active ingredients for the wrinkling proc-
free-standing units ess were generally a natural ester-based oil
such as Tung oil and a natural ester resin such
B) Gross damage, including missing panels as Congo copal. Two fairly thick coats of wrin-
and parts kle paint needed to be applied in fairly quick
succession. The basic mechanism is one of
C) Surface damage, including instability of the
paint surfaces and the need to integrate paint
losses and treat other visual defects * As normally understood and as scientifically de-
fined, torque is the moment of a system of forces caus-
There was also a need to give the units long- ing rotation rather than the rotation of twisting itself.
term protection from dust and from the public. However, we have used the term here because "twist-
ing" is not adequate to describe what is happening to
A) Structural Instability the frames. We use torquing to mean that moments of
Originally, the ENIAC units were bolted to- force were applied that caused the cabinets or frames
gether on special bases made for the site. Since to become out-of-square.

2s3
differential shrinkage of the two layers with
rapid evaporation of the solvents. Other addi-
tives affected the pattem and size of the wrin-
kle. The wrinkle paint was advertised for use to
increase grip, lessen eye strain because ofde-
creased reflection and reduce finishins ofthe
metal before painting.

Figure 6 Accumulator unit #19 illustrates the bind

and extent of paint damage and loss in many of the
ENIAC sections.

Defects Subsequent to the ENIAC's

l{orking Lifetime
Many of the panels had rough identification la-
bels scrawled on the front. Most were marked
with orange grease-pencil but some were
marked with white paint. We do not know for
certain if these markings were made during the
working lifetime of the machine. Some of the
units also had thick adhesive residues in areas
where a fiber-reinforced tane label had been
applied.

Treatment
Specific aspects of treatment are organized
under similar subheadines to those cited in the
discussion of problems.

Figure 5Gas chromatographs show comparisons of In general the units were lrst vacuumed to re-
the oiginal ENIAC wrinkle paint with the modern move a thick accumulation of dust and dirt
version used in the conseruation effort. -
all were very dusty. Plastic knobs were cleaned
using soft cloths or cotton swabs and distilled
water.
The conditions of the wrinkle paint ranged
from good to fair depending on each unit.
Stabilizing the Units
Many of the units had varying degrees of past Feet, composed of 7 .6 cm x 7.6 cm
restoration and conservation. Two of the units (3 in. x 3 in.) angle-rolled steel were made
had severe loss of the wrinkle paint due to rust
to allow units to stand alone. These new bases
and improper surface preparation. Other units
allowed the units to be moved and installed in-
had losses ofpaint on single parts, scratches,
dividually in the exhibit area as their conserva-
chips and abrasions. There were numerous tion treatments were completed. Three units
small areas of flaking finish and loss primarily that were bolted together had a special base
due to rust. The areas of rust on the panels were
made, allowing them to stand as a triple unit.
very scattered and appeared as groupings of
Bases were also designed so that the units
small dots or large irregular areas of fine rust could be bolted together in groups as they
(Figure 6).
were in the orisinal installation.

254
Removal of Dents and Bends of modern plastics. Even though they were not
The dents and distortions in the panels were painted and did not have quite the same texture
fairly minor. One or two panels had to be as the originals, they blended in nicely.
straightened somewhat to make them fit, but
the dents and other distortions were left Surface Instability and Consolidation
untouched. In order to try to save the remaining finish on
the panels, the rusted surfaces were coated with
Torquing Problems a resin to act as a consolidant by, in essence,
Over the years, the panel frames had torqued binding the frnish and the corrosion products to
slightly due to moving and incautious storage. the surface of the metal. The consolidant was
As a result, the screw holes in many of the old either painted on with a small brush around the
panels and almost all of the new ones had to be areas ofloss or sprayed on, depending on the se-
custom drilled or enlarged to allow them to fit. verity of the flaking.

Replacement of Missing Parts and Pieces Inpainting with Acrylics

We decided that the missing panels should be On the panels where the paint was in relatively
replaced so that the units appeared as complete good condition the losses were isolated with an
as they were during the ENIAC's working life- acrylic resin and inpainted with acrylic paints.
time. New panels for the units were constructed The isolating resin is easily soluble so it en-
out of26-gauge galvanized steel sheet. They ables removal of any inpainting in those areas
were then measured for the screw positions, in the future. The initial treatments provided a
drilled, and the smaller replacement pieces solid base so that the units could travel to the
were painted. The two large bases that held exhibit site needing at most a few minor aes-
the ENIAC's connecting buses were painted in thetic touch-ups later on.
black wrinkle vamish thinned to work in
spray guns. A major problem for surface reconstruction
arose because some of the ENIAC units also
The new panels were fastened to the rmits using had thick adhesive residues in areas where a fi-
new screws painted with black acrylic paint to ber-reinforced tape label had been applied. We
better integrate them with the existing screws. were unable to find a solvent mild enough to
Further inpainting and integration with the remove these residues without damaging the
resin-varnish mixture was done as needed after paint. Where possible the residues were scraped
installation of the units. Because of the diffi- off and overpainted with acrylic paints. The
culty in matching the texture and the surface problem with this treatment was that the adhe-
sheen, the inpainted losses are still somewhat sive residues along with the acrylic overpaint
evident. Thus it is fortunate that the most in- produced a smooth surface with a reflectance
painted areas are not highly lit. different from that of the surrounding surface.
The only way to blend in these areas was to
In addition to the l5 basic units there were the painstakingly dot them in with a glossy resin.
numerous horizontal flat units or "buses" that
were stacked across the bases of the basic units Orange Grease-pencil Markings
once the ENIAC was assembled. At frst, the orange grease-pencil letters on one
panel were removed with solvent. It was very
Equally important, many of the original cables difficult to prevent the solvent from carrying
and plugs were missing. Even though the the orange residues ofthe grease-pencil out-
ENIAC units were not being restored back to ward into the surrounding wrinkle finish and
functional use, both the conservation staff and creating ugly orange halos. Because of the
the curator felt it imporNant to restore these as finish it was impossible to remove all of the
closely as possible to their in-use appearance. grease-pencil in the interstices without damag-
New plugs and cable components were ob- ing the paint undemeath. This required stronger
tained at an electronic supply store. They were solvents or poulticing, which meant that a sol-
similar in appearance to the originals but made vent had to remain on the paint surface for a

255
long time. At the same time some controversy
arose over whether or not the markings were
put on while in original use, in storage or even
later by the Smithsonian. We decided to only
remove or overpaint those areas that would be
most visible and visually disturbing while on
exhibit.

White Paint Markings

The writing in insoluble white paint on some of
the panels was carefully picked out using a scal-
pel, and any remaining white paint was covered
over with a black acrylic paint. Unforhrnately,
there was still a considerable amount of paint
that remained in the interstices of the wrinkled
Figure 7 An example of the heat lamp anangements
areas. This smoothed out the finish but made used to induce the wrinkle effect in the Jinish. Note
the areas more obvious. In order to visually the winkle boundaries spreading out from two areas
break up the smooth surfaces caused by acrylic at the top ofthe panel.
inpainting and remaining white paint, a glossy
resin was dotted on. This process was satisfac- while. We do not know exactly what caused
tory for many of the small scattered areas of this, but there seems to be many possibilities,
loss, but the smoothness of the finish or lack of including dirt, mixing of paint, film thickness,
texture, such as in the original paint, had made amount of time between heating and spraying,
the larger areas of inpainted losses obvious. We and temperature of the heat lamps.
decided to try to alleviate this problem by using
wrinkle paint for inpainting. Close examination showed that generally
our wrinldes were not as fine as those of the
Modern Wrinkle Finish original f,rnish. Perhaps a commercial estab-
Initially we were able to purchase black lishment's equipment could control the vari-
Ruf-Koat Wrinkle Vamish (made by the ables more closely and produce better results
Illinois Bronze Powder and Paint Co.) at an on average. However, on further examination,
electronics supply store. This type offinish is we did find that some areas of the original wrin-
still used by the electronics industry and comes kle finish on the ENIAC and other instruments
in a small selection of colors. The wrinkle paint of the period seemed to have some of the same
would wrinkle on its own if merely left to dry problems that we experienced. Thus overall the
overnight, but the wrinkles were large and repainting process went remarkably well and
often areas would not wrinkle at all. Normally could be done in most facilities.
in a commercial establishment the panels
would have been baked in an oven after paint- We found the manufacturer, Illinois Bronze
ing. Our small oven in the laboratory was not Powder and Paint Co., very cooperative when
nearly large enough so we decided to try to get we explained our project and problems. Local
the same effect using infiared heat lamps. supplies of this product were often limited, but
we were able to get the product directly. We
Often, after applying the wrinkle paint, we were also able to get advice on usage of the
would position the heat lamps long enough at paint.
one end of the piece to start the wrinkling proc-
ess. Then we would gradually move them down Inpainting with Wrinkle Finish
the length of the piece to heat the metal and the In order to inpaint the wrinkle paint in small
paint surface as the wrinkling progressed (Fig- areas with a fine brush, the paint was sprayed
ure 7). Occasionally we would still end up with from aerosol cans into a small metal tin and
small areas that would not wrinkle even after used quickly and in small amounts before the
leaving them exposed to the heat lamps for a rapid evaporation of the solvent made the paint

256
too thick to paint easily. The texture and sheen
of the applied wrinkle paint were slightly differ-
ent than the original wrinkle frnish. Care had to
be taken not to paint over the edges to avoid
thick build up; the effect was much more suc-
cessful than acrylic or any other substitute
paints tried. Even the bulk wrinkle paint in the
can had different surface effects. No matter
how much it was thinned, the wrinkle effect
was not as fine grained. Ultimately only the
paint from the aerosol spray can gave really
good inpainting results.

Repainting and Overpainting

The original literature on wrinkle finishes Figure 8 Accumulator unit #19 at a later stage
claimed that the paint adhered very well to bare shows blocking arrangements just pior to painting.
metal. However, as the condition of the ENIAC The original paint was too badly damaged in many
proved, this did not always hold true. While a areas ofthis object to be reliably consolidated.
few ofthe units appeared to be in very good
condition, most had mild to severe paint loss. then with two coats of the black wrinkle paint.
Much of this could have been due to the im- The presence of primer indicated which units
proper storage conditions that promoted rust. In- were scraped and repainted since the originals
itially we tried to inpaint only in the losses on were never primed. We then painted the panels
some of the badly rusted panels, but the appear- with the wrinkle paint and positioned the heat
ance was not satisfactory. We thus decided to lamps to cause fine wrinkles.
repaint them entirely with the black wrinkle
paint. Thereafter, wherever flaking paint was se- Finally, in order to better integrate the fresh and
vere and wherever the losses would be visible glossy appearance of the repainted panels with
while on exhibit, we brushed or scraped all of the original panels, which were blanched or
the remaining paint offand repainted entirely. appeared lightened from dirt, both were coated
with a wax-resin mixture. This acted like a
The knobs and other areas not to be painted varnish to produce a satin-gloss appearance.
were blocked out or covered. For large areas,
painters' masking tape and brown paper were Protection from Dust and the Public
used (Figure 8). At first the many small set-in
plastic sockets were covered with tape and then Because the plugs could easily be pulled out
the edges were further sealed by painting on art- and knobs twisted until broken off, we felt it
ists' liquid latex masking fluid. In tests this necessary to protect the ENIAC units from pry-
seemed to work well but after painting one of ing hands and souvenir hunters. The design con-
the large panels we decided that it was too diffr- cept of the exhibit was to be as open as possible
cult to paint the latex on evenly around the so a plexiglass barrier was placed in front of the
tape. The tape seemed to work well if it was units. While generally protecting the units, this
carefully cut, applied, and sealed by pressing type of banier unfortunately still allows the
with our fingemails around the edges. This was units to accumulate dust and dirt and at the
also a lot less time-consuming than painting same time makes cleanins access difficult.
around each ofthe hundred or so sockets. The
tape was, of course, removed soon after Conclusion
painting.
The 15 or so people who worked on the conser-
In hopes that the newly applied paint would ad- vation of the ENIAC together contributed over
here better, the panels were painted frst with 520 hours of time. This project was a challenge
grey Rust-oleum Auto Primer spray paint and because of the size of the artifact, the time

2s7
constraints and also the nature of the primary Even a partial restoration of the elechonic appa-
problem: the unusual finish. Over a year's ratus that so pervade 20th-centwy science and
worth of time could have easily been taken to technology will pose continuing challenges as
treat the ENIAC considering the size and num- such artifacts are collected and exhibited. For
ber of units, but one has to be realistic when example, preliminary research shows that the
working with large, multi-unit objects. Also, pervasive black wrinkle finishes were chal-
visitors to history museums, as opposed to fine lenged, giving way to other color variants and
or decorative art museums, expect the func- metallic glazes. There were defrrite preferences
tional objects on display to look used. Thus it is for color and appearance offinishes at various
often not only unrealistic but undesirable to re- periods that might even be characterized as
turn an historic object to its original condition lashions or at least fads.
(Figures 9 and l0).
Manufacturers of laboratory fumiture and appa-
ratus in the late 1950s and early 1960s moved
quickly to take advantage of increased funding
during the immediate post-Spuhrik period.
Brightly colored, "modem" metallic finishes
graced power supplies as well as automobiles.
In the late 1960s and 1970s, flat-beige and blue-
matte finishes were fashionable for a while and
even the dull-black of military gear was used
a bit, though it caught on much more with con-
sumer electronics. Similarly, some manufactur-
ers experimented with brushed aluminum,
though, it too, was more popular for high-
fidelity gear.

In future cases, conservators will face finishes

composed primarily of one type of polymer,
Figure 9 Accumulator unil #19 after consemation
but there may be formulations with several
polymers, as this was common practice in the
past. Each of these trends in finishes will pre-
sent specific problems for the conservator, de-
pendihg on the degree to which the curators
and designers want the surface to be restored
or simply stabilized and consolidated.

The close collaboration between curator and

conseryator made this project run much more
smoothly when it came to making decisions on
the different treatment options. It is crucial to
have the curatorial staff help with treafinents of
major artifact systems like the ENIAC not only
Figure I0 A bank ofeight ofthe original ENIAC because it leads to faster decision makins and
units on display rn The Information Age exhibition easier solutions to problems. but because-it
at the National Museum of American History. At
leads to leaming more about the objects. And
right is a video clip selectorJbaturing the ENIAC's
chief engineer, Presper Eckert, in a I 988 interview
isn't that what a museum is for?
explaining key aspects ofthe computer.

258
R6sum6 restauration et de la conservation ndriteft d'Atre
rdptttes. Les donntes recueillies dans le cours
Garantir l'avenir des vesfiges de notre haute des travaux de restauration ont mend d des re-
technologie : les legons d tirer de I'expthience cherches en conservation plus poussies sur le
de I'ENIAC pour sssurer la conservation des rev€tement vermicult de I'ENIAC, ainsi qu'd
grandes techniques de I'ilectronique des dtudes portant sur d'autres appareils 6lec-
troniques des anndes 40 et 50. Un tel rev€tement
Au mois de mars 1990, le National Museum donne d ces appareils un aspect particulier, d
of American History prtsentait, sous le titre de la tnte comme au toucher, qui est sensiblement
The Information Age (L'dre de I'information), dffirent de celui de leurs predecesseurs, des
la plus grande exposition technologique de I'his- anndes 20, et de leurs successeurs, de la peiode
toire du Smithsonian. Le plus volumineux des postdrieure aa Spoutnik La prdsente communica-
objets exposts, et le plus important au point de tion portera sur I'analyse, sur la conservation et
vue historique, etait une reproduction partielle sur la restauration du rev€tement de I'ENIAC,
du premier ordinateur ilectronique tout usage, tout en traitant des rtpercussions historiques des
I' E lectronic Numerical Integrator And Computer travaux qui ont 6tt effectuds.
(ENIAC). Occupant un volume total de quelques
centaines de mitres cubes et pesant plusieurs Bibliography
tonnes, les I 5 unites exposdes posaient des
probl'imes de conseryation tris diversifi,ls. On a Hopwood, Walter, "Paint Samples from the
ainsi eu d choisir, pour les composants, parmi ENIAC," Smithsonian CAL Analvsis Report:
la retouche classique, la repeinture compldte et February, 1990.
la substitution de piices, en tenqnt compte de
facteurs concurrents comme les contraintes du Stern, Nancy. From ENIAC to UNIVAC
calendrier d'exposition, les normes de conserva-
(Bedford, Massachusetts : Digital Press, 198 I ).
tion et les donntes d'experiences effectudes pour
dtudier la rdaction de versions modernes de
rev€tements. Avant lafin des travaux de restau- Waldie, W.A., The Science of Wrinkle Finish-
ration, environ I5 membres du personnel, dont ing (Dayton, Ohio: Research Press, Inc., 1949).
le conservateur responsable de l'obiet, avaient
consacrd 180 heures aux optrationi de traite- Williams, M.R.,l History of Computing
ment, sous la coordination de I'un des auteurs de Technolog,, (Englewood Cliffs, New Jersey:
la presente communication. Des expdriences de Prentice Hall. 1985).
collaboration aussi 4troite entre sptcialistes de la

259
Modern Metals in Museum Collections

Lorna R. Green and David Thickett

Department of Cons ervation

TheBritish Museum
London. U.K.

Abstract to identify the cause of the conosion; to develop

the Oddy test for materials to be used in the stor-
It is well known that materials usedfor the stor- age and display ofthese metals; and to produce
age and display of antiquities can emit harmful guidelines for their storage and display.
gases that initiate corrosion and cause other dete-
rioration processes. Since 1972 consentation sci- The corrosion processes involved have been inves-
entists at the British Museum have been using an tigated by identification of the corrosion products
accelerated ageing test known as the Oddy test to and a survey of the conosion literature. Using
evaluate the potential of materials to cause corro- this information as a basis, a series of conosion
sion ofmetals. The test has been used to evaluate experiments was carried out to investigate the ef-
the corrosive effect on silver, copper and lead, fects of the known common pollutants in the mu-
the metals that have shown the greatest potential seum environment on these metals. Guidelines
to corrode in the Museum. The introduction of the for the storage and display of these metals were
testing procedure dramatically reduced the num- prepared.
ber of instances of objects corroding in exhibition
cases. It also reduced the rate at which silver Introduction
tarnished.
Itis well known that materials used for the stor-
In common with many museums, the British age and display of antiquities can emit harmful
Museum has continued to collect into the 20th gases that initiate corrosion and cause other de-
century, bringing modern metals into the collec- terioration processes.' Since I 972, the British
tion. Artefacts, coins and medals made of alu- Museum has been using an accelerated ageing
minium, zinc and magnesium and their alloys
test, the Oddy test, to evaluate the potential of
form a small but significant part of the collection. materials to cause coqosion of metals in the
Such metals are commonlyfound in galleries of
modern art. These metals are no less susceptible Museum's collection.'This test has been used
to corrosion than the metals of antiquity. Alu- to evaluate the corrosive effect of materials on
minium artefocts have exhibited both localized pit- the common metals of antiquity, silver, copper
ting corrosion and theformation of a thin layer of and lead, which are most readily corroded in a
white corrosion. Zinc money from the First and museum environment. The introduction of the
Second World Wars has beenfound to corrode testing procedure resulted in a dramatic reduc-
completely when stored in wooden coin cabinets. tion in the number of objects corroding in exhi-
Magnesium has also beenfound to corrodeform- bition cases. In common with many museums,
ing awhite surface corrosion layer. These obser- the British Museum has continued to collect
vations have caused concern to carators and into the 20th century. This has brought modern
consenators. As a result. a studv was undertaken metals into the collection. Artefacts, coins and

261
medals made of metals, such as aluminium, a protective film of aluminium oxide, AlzO:.
magnesium andztnc, and their alloys, now Formic acid, however, peptizes this oxide film,
form a small but significant part of the producing a gelatinous layer that is no longer
collection. protective. Organic chlorides and hydrogen
chloride also disrupt the oxide layer. Particu-
Aim larly in outdoor environments, aluminium is
sensitive to sulphur dioxide. Surface analysis
The aim of this study was twofold: first, to as- of aluminium liom an indoor environment has
sess the condition of modem metals in museum shown that it accumulates adherent particles
collections, and second, to identifu materials of chloride and sulphate. Pitting corrosion is
often used for storage and display purposes in characterislic of chloride and carboxylic acid
muserurs, which are potentially corrosive to corrosion.l0'11 In general, addition of alloying
zinc, aluminium and magnesium. metals, such as copper, reduces the corrosion
resistance of aluminium.'' Previous tests have
Modern Metals and Their shown that aluminium corrodes in an enclosed
Corrosion Processes environment containing wea-formaldehyde ad-
hesive, in tropical conditions. D.ying oils and
Zinc painted plywood had no effect.'
Although zinc was known as a distinct metal
in lndia in the l4th century and probably earlier Magnesium
in China, it was not until 1740 that the frrst Magnesium was isolated in 1808. Industrial pro-
commer-cial zinc extraction plant was built in duction commenced on a modest scale towards
Europe.3 Hence, when compared with metals the close of the 19th century.
of antiquity, silver, lead and gold, zinc may be
considered a "modern" metal. Indoors, the primary reaction of magnesium is
to form a hydroide, Mg(OH)2. This film be-
In damp air, in the absence of carbon dioxide, comes protective by secondary reactions, result-
zinc hydroxid e Zn(OH)z forms. However, ing in the formation of hydrated carbonates and
where carbon dioxide is present, a thin grey- sulphates. The most severely corroding environ-
white covering of carbonate possibly ZnCOt ments for magnesium and its alloys are dilute
and ZnCOtZn(OH)z forms, which protects chloride solutions and dilute acids, particularly
against further corrosion.- In the presence of- . mineral acids and.qrganic acids, such as acetic
rilphrr dioxide. zinc sulphate Zn'SO+ forms.5'6 and tartaric acids." Alloying metals, such as
Zinc corrodes in the presence of acetic acid, manganese and zinc, reduce the elecfrode poten-
formic acid and formaldehyde.' Gaseous and tial from that of pure magnesium, and the corro-
aqueous hydrogen halides also affect zinc.S As sion rate increases accordingly. It has been
with most metals, the corrosion rate of zinc in- recommended that red lead or white lead paints
creases with increasing relative humidity. It has should not be used with magnesium and its al-
been reported that zinc-tin alloys have shown loys, as traces oflead are precipitated electro-
greater corrosion resistance than zinc. Previous lytically aTfl cause pronounced local cell
tests undertaken in differing climates have corToslon.
shown that zinc corrodes in the presence of dry-
ing oils, such as tung oil in plylvood packing Magnesium ribbon, stored in the laboratory,
cases, and with formaldehyde adhesives.' acquired a dense black covering where it had
been exposed to the air. Tests with materials
Aluminium used to make packing cases showed that magne-
Aluminium was discovered in 1820. Marked sium alloys corroded in wooden boxes of pine,
gowth of the aluminium industry occurred in beech and birch plywood. As a result recom-
the vears 1914 to 1918.1 mendations were made that wood should
be avoided when packing magnesium
In pollution-free conditions, aluminium is alloy parts.'
resistant to corrosion due to the formation of

262
A Survey of Modern Metals Aircraft Parts
in Museum Collections Aircraft parts are generally made of aluminium
and magnesium alloys for their low density and
The first stage was to assess the condition of high strength. A number were examined and
modern metals in ambient museum environ- sampled at RAF Cardington, which holds part
ments. Objects in the collections of the Tate of the reserve collections of the Royal Air
Gallery, Victoria and Albert Museum, Royal Force Museum at Hendon, London. Corrosion
Air Force Museum and the British Museum of some artefacts had occurred prior to collec-
were examined for signs of conosion. Corro- tion, as they were often retrieved from wreck-
sion products were found on some objects age after many years in the sea or outdoor
containing zinc, aluminium and magnesium. environments. In one instance, a gearbox cas-
However, many objects made from these ing was found to be composed of a zinc-copper
metals exhibited no signs of corrosion. alloy, with corrosion products containing
chlorides.
Sculpture
A sculpture in magnesium consisting of The majority of the artefacts are stored on open
144 squares was found to have a white pow- shelving in heat-sealed polyethylene bags. The
dery deposit around the edges ofthe squares. "bagging-up" is carried out during periods of
This corrosion was identified as a mixture of low humidities. This appears to be very success-
magnesium oxide, MgOz, and magnesium bo- fu| and no signs of further degradation of the
rate hydrate, Mg5(BOl)O(OH)s.2HzO. It is metal pieces are visible.
thought that this may have formed as a resqlt
ofboric acid being used as a fire retardant,t) or Coins
from cleaning agents around the sculpture. The largest concentration of modern metals
was found to be in the Department of Coins
In a sculpture by Beuys, comprising a zinc and Medals at the British Museum.
box containing fat resting in a painted wooden
showcase, extensive zinc corrosion was ob- All coins are stored in nonJacquered mahog-
served. The zinc was in good condition in any coin trays in traditional wooden cabinets
1989. However, the showcase is part of the with tight-fitting wooden doors, thus the
composition and the artist had instructed that storage environment has negligible air
the paint be maintained in good order. Follow- exchange.
ing repainting in August 1990, the zinc devel-
oped a thick, white corrosion crust over its German coins issued during the years of 1914
entire exposed surface. Analysis by X-ray to 19l8 and 1939 to 1945 were found to be cor-
diffraction (XRD) has shown this to be zinc roding, in most cases, producing a bulky white
formate hydrate, CzHzO qZn.2HzO. corrosion product.

A screen-printed aluminium alloy sheet, by More recent currency was observed to have a
Longo, approximately five years old had devel- thin surface layer of white corrosion. Several
oped elongated patches ofwhite corrosion, cor- of these coins had been retrieved from a small
responding to the direction of printing. This fountain in one of the Museum galleries, where
piece had been on open display and has been visitors throw their loose change. The chloride
stored in wooden packing cases. Because of level (58 ppm), conductivity (0.7 S cm-'
difficulty in obtaining a sample, analysis of the [where S represents Siemens]) and pH (8.3)
corrosion was not undertaken in this instance. of the water from the fountain were similar
to levels found in tap water.
Bidri Ware
Bidri ware, made from a zinc alloy, was found Ten coins in varying states of preservation were
to be corroding. The corrosiop_was identified as selected for fi.rther examination (Figure 1).
I6
simonkolleite, Zns (OH)sClz.

263
 coins, it was possible to analyse a non-corroded
and corroded area. The corrosion was further
analysed by powder XRD. Results are shown
in Table I.

Discussion of Coin Analysis

Most of the coins analysed were identified as
zinc. In most cases the corrosion appeared to be
active. Zinc formate hydrate was identified on
one coin, and its formation was apparently due
to exposure to formaldehyde, or formic acid,
both of which can be emitted by wood and
Figure 1 Corrosion on zinc coins stored in wooden wood composites.
cabinets. Photo: Department of Conservation,
British Museum.
Although chlorides were present on the surface
of the other zinc coins, only one chloride-con-
Analysis of Selected Coins taining corrosion product was identified, si-
monkolleite, Zns(OH)gClz. The most likely
The coins were analysed using the scanning
source ofthese chlorides is from handlins the
electron microscope with energy dispersive
coins.
X-ray analysis SEM (EDXA). On most of the
Table I
Condition of Coins and Analysis of Corrosion Products
NA = not analyzed
tr = trace

Coin Condition Elements Detected Corrosion

sEM (EDXA) Product (XRD)
Metal Corrosion
German, General white surface Zn, (tr: Cl, Ca) Zn, Cl, (tr: Fe, Ca) Amorphous
5 Gelder Stadtmark, thin corosion laver
circa 1 920
German, General white thin Zn, (tr: Pb, Cl) Zn, (tr: Fe, Cl, Si, K ) zinc formate hydrate
10 Mark, corrosion layer, patctres CzHzOqZn.2HzO
1919 of thicker white conosion
German, Localized patches, Zn, (tr: Cl, K, Ca) Zn, Cl, (tr: Fe, Pb) Simonkolleite
50 Pfenniq, 1920 white conosion Zn. (OH)aCl,
German, Small patches, white Zn, (tr: Pb, S, Cl, Si, Al) NA NA
50 Piennig, corrosion, generally
1917 qood condition
Czech, General white surface NA Zn, Cl, (tr: S, Fe) Zinc carbonate
2 Haleru, thin conosion laver hydroxide
circa 1920 7n.1CO"\"lOHl"
lcelandic, Thick white surface Zn, Cu, (tr: Cl, Si, Al, Zn, Cl, (tr: Si, S, Al, Co) Zinc carbonate
5 Qrc, corrosion K, Ca) hydroxide
1 954

French, General white surface Al, (tr: Ag, Cl, Si, Ca, Al, (tr: S, Cl, Ag, Si, Unidentified
1 Franc, thin conosion layer, K, Sn, Cu) Fe, Cu)
1 949 with thicker oatches
German, (Dresden), Few white patches Ni, Zn, Cu, (tr: Mg, Zn, Ni, Cl, (tr: Cu, Zinc carbonate
20 Pfennig, of corrosion Ca, Fe, Cl) S, FE)
circa 1920
German, Some small patches Mg, (tr: Si, Al, Ca, Mg, (tr: Fe, S, Ca, Cl, Magnesium carbonate
10 Mark, of corrosion Mn, S) si,K) hydroxide hydrate
1943 Mn./Cf).\,/OFl\. Fl"O

Japanese, Thick surface Al, (tr: Ag, Si, Ca, K) Al (tr: Si, Ag, Fe) Unidentified
1 Yen, corrosion layer
1 986

264
The corrosion product on the aluminium As pure magnesium conodes on exposure to
coin contained chloride, but was not positively moist air, it is unlikely to be encountered in
identified as it gave a poor XRD pattern. Again pure form in museum objects. Only one magne-
the source of the chloride is probably handling. sium alloy coin and one sculpture were identi-
The magnesium corrosion, a carbonate, is prob- fred in the survey. Magnesium alloys are more
ably due to reaction with atrnospheric carbon stable. but not fully resistant to pollutantsJoten-
dioxide. This reaction may be catalysed by the tially present in a museum environment.' '''o
presence oforganic acid vapours.
Materials Corrosive to
Ethnographic Artefacts Modern Metals Identified
by the Oddy Test
In the Deparhnent ofEthnography, a group of
approximately 40 objects was examined. Most In the British Museum, the corrosive effects of
have been in the Museum since the beginning all materials to be used in the storage and dis-
of this century. All are loosely wrapped in acid- play ofsilver, copper, lead an$ their alloys are
free tissue and stored in boxes made from resin- assessed using the Oddy test.' This involves en-
bonded fibre. Several zinc objects were found closing the material under investigation in a
to have patches of corrosion, though many boiling tube with a clean, degreased coupon of
pieces were in good condition. Zinc corrosion metal, chosen to represent the composition of
was also noted on some Indian anklets stored in the artefacts to be stored or displayed, in close
wooden drawers. In general the aluminium was proximity to the material. To avoid contact
in good condition, even where it was in direct with the test material, the metal coupon is sus-
contact with wood. Nevertheless, two instances pended in the tube using nylon monofilament.
of corrosion of aluminium were observed and To accelerate corrosion reactions, a small vol-
samples were taken for analysis. ume of water is added to create a high humid-
ity. Solid carbon dioxide is added when lead is
Analysis of Ethnographic Artefacts used to accelerate corrosion, as carbonates are
formed as corrosion products. The boiling tube
Simonkolleite, Zns(OH)sClz. was found on the is sealed and placed in an oven at 60oC (see
edge of a zinc frame holding a museum label. Figure 2).
The corrosion on the anklets was identified as
basic zinc carbonate. Both these corrosion prod- The condition of the metal coupon is examined
ucts, which are similar to those found on the after 28 days and used to evaluate the suitabil-
coins, may have formed due to handling and ity of the test material. The method of conduct-
the presence of volatile organic acids in the ing the tests is under review and will be the
atmosphere. Two of the aluminium corrosion subject ofa later publication.l9
products were amorphous. Although the others
remain unidentified, they were found, using the
SEM (EDXA), to contain chlorides, presum-
ably from handling.

Survey Conclusion
From this limited survey it appears that zinc
presents the greatest problem. This was ex-
pected as zinc is known to corrode in the pres- cottonwool metal
ence of common museum pollutants, such as with coupon
organic acids. drstilled water

Aluminium objects showed fewer instances

of corrosion, but in general, they may be sus-
ceptible to a build-up of gases in an enclosed Figure 2 Diagrammatic representation of the
showcase or storage area. Oddy test.

265
This test procedure was used to assess the tests and the analysis are shown in Tables II,
corrosive effects of common materials used in III, IV, V and VI.
storage and display in museums on zinc, alu-
minium and magnesium and their alloys. Tests Discussion of Oddy Tests
were carried out with and without solid carbon
dioxide added, to evaluate its effect. Zinc
Zinc showed slight corrosion in the presence of
The metals used in the tests were zinc, from a water, and water and carbon dioxide. However,
sheet already in the Museum; aluminium foil; the wood and wood composites caused much
Duralumin, an aluminium alloy containing a more severe corrosion. The zinc coupon with
few percent ofcopper; magnesium sheet; and oak was so extensively corroded that it was be-
AZ3l,* a magnesium alloy with small concen- ginning to laminate. Addition of carbon dioxide
trations of aluminium, manganese andztnc. did cause increased corrosion with oak and ply-
wood, but it is not considered to have had an
Common materials used in storage and display effect on the outcome of the test, that is, the
in museums and known to be potentially corro- classification of the test material as unsuitable
sive to metals of antiquity were selected for the for use, or suitable for temporary or
tests. These included wood, wood composites, permanent use.
paint, plastic and wool. The last was selected
because wool-based fabrics, such as felt, are Zinc oxide was identified as the major corro-
often used to dress display cases or in storage sion product with the oak and plywood tests,
even though wool is known to evolve sulphur- even when additional carbon dioxide had been
containing gases that are corrosive to silver and added. Later analysis showed that z;nc carbon-
other materials. ates were also present. This suggests that there
had been reaction with atmospheric carbon di-
Tests with oak and plywood were set up in du- oxide after removal of the zinc pieces from the
plicate, one set having carbon dioxide (COz($) tests. Similar results have been reported by
added, in order to investigate whether its pres- other workers.4
ence is necessary to promote corrosion reac-
tions of the metals and alloys under test. Formic acid attacked the zinc, producing a
Controls were set up in ambient conditions, thick corrosion crust. Paints had some deleteri-
and, as in an Oddy test, with only water or ous effect on zinc. Corrosion occurred in the
water with COz(s). Because of limited time and presence ofwool, but the resultant corrosion
availability of some sample materials all of the did not appear to contain sulphur.
samples were tested against zinc and alu-
minium, but only some were tested against Aluminium
magnesium, AZ3l magnesium alloy and Aluminium conoded only in the presence of
Duralumin. formic acid. Some wood and wood composites
did cause a slight dullness to the surface ofthe
The metal coupons were examined after 28 test pieces, but no corrosion product was
days. Several samples of corrosion were taken formed. Similarly, wool and woollen felt pro-
and analysed using XRD. The results of the duced a dull surface, which was yellow/brown
in places. The latter is possibly due to deposits
from the wool condensing on the metal.

'Speofrcatron, Magnesrum Elektron Ltd. Report

Duralumin showed more evidence of surface
Senal No ML11561
dulling and discolouration in the tests. Corro-
Zn Al Sr Cu Mn Fe sion formed on Duralumin in the tests with ply-
075 2 7 <001 <0 01 0.23 0.003
wood and with formic acid, the latter producing
Nl Ca Zt Others Mg both aluminium and copper corrosion. The cor-
<0 005 <0.01 <0.01 0.3 Rem
rosion products produced weak but similar
(units In percent) Rem = Remarnder XRD patterns. Although they were compared

266
with known XRD pattems of common alu- acquired a thin corrosion layer in many of the
minium corrosion products, they were not iden- tests, the samples were not as extensively at-
tified. They appeared to be different l}om the tacked as the pure magnesium samples. Carbon
unidentified pattems obtained from samples of dioxide did not affect the outcome of tests with
aluminium corrosion from objects in the ethno- magnesium or A231.
graphic collection in the British Museum. Addi-
tion of COz to the tests had little effect on the Conclusion from Oddy Tests
aluminium or Duralumin.
The results suggest that the Oddy test can be
Magnesium used to evaluate the corrosive effects ofstorase
Magnesium was particularly susceptible to cor- and display materials on zinc, aluminium, malg-
rosion under the test conditions. The "control" nesium and their alloys. The addition of carbon
conditions, that is, with water alone, and water dioxide is not necessary as it does not affect the
with carbon dioxide, caused slight corrosion of results of the test. A control, with the metal and
the test pieces with corrosion products being water only, should always be carried out and
white in all cases. However, some of the materi- used for comparison with other tests.
als under test caused considerably more
corrosion. Only the degrading PVC and the It should be borne in mind that wood mav varv
paints did not cause corrosion ofthe pure according to its age, storage and seasoning.
magnesium test pieces. This in turn may affect its corrosive nature.
Wood composites, such as chipboard and ply-
The magnesium alloy AZ3l was also suscepti- wood, can also have variable effects because of
ble to corrosion. However, although it had variability in composition. The drying period

Table II
Condition of Zinc and Maior Conosion Products after the Oddy Test with a Range of Materials

Test Condition after Test Product

Control (window sill) NC
Material in Oddy Test
Darkening of surface ZnO
COz Darkening of surface, light thin corrosion ZnO
layer on low edge
Oak Extensive white corrosion ZnO
Oak, COz Thick white corroslon, lamtnation of metal NA
Plywood Extensive white corrosion NA
Plywood, COe Extensive white corrosion ZnO + Zn+COs(OH)o.HzO
Formic acid (1 drop) Thick white corrosion zinc formate hydrate
Zn(COOH)z.2HzQ
Cardboard Shght white surface corrosion NA
Gloss paint Yellow/whrte thin corroston laver NA
on surface
Vinyl matt paint Slight surface corrosron layer NA
Scoured, unspun wool Heavy yellow/white corrosion NA
Degrading PVC Slight corrosion layer NA
Mahogany Light surface corrosion layer ZnO
Pine Thick whrte corrosion, lamination ZnO
of metal
Medium-density Thick white corrosion ZnO
fibreboard (MDF)
Chipboard Entire surface covered with possibly Zns(COg)z(OH)o
white corrosion

267
 Table III
Condition of Aluminium and Major Corrosion Products after the Oddy Test
with a Range of Materials
NC = no change

Control (window sill) NC

Material in Oddy Test
NC
COz NC
Oak Dulling of surface
Oak, COz Dulling of surface
Plywood Dulling of surface
Plywood, COz Dulling of surface
Formic acrd (1 drop) Otf-white surface, metal disintegrattng
Cardboard NC
Gloss paint NC
Vinyl matt paint NC
Scoured, unspun wool Some yellow/brown patches but no corrosion
Degrading PVC NC
Mahogany NC
Pine NC
Medium-density NC (resinous deposit)
fibreboard (MDF)
Chipboard NC (resinous deposit)
Woollen felt Dulling of surface, yellow/brown in places

Table IV
Condition of Duralumin, an Aluminium Alloy, and Major Corrosion
Products afier the Oddy Test with a Range of Materials

NC=no NA = not analvsed

Test Conditions Condition after Test Corrosion Product
Control (window sill) Slight dulling
Material in Oddy Test
Dulling of surface
COz Small patch of whrte corrosion on lower edge NA
Oak Small patch of white corrosion on lower edge NA
Oak, COz Small patch of white corrosion on lower edge NA
Plywood Few soots of corrosron NA
Plywood, COz Few soots of corrosion Unidentified
Formic acid (1 drop) Patches of white corrosion Unidentified
sometimes "watery," and some green corrosion
Cardboard NC NA
Gloss paint Some dulling NA
Vinyl matt paint Some dulling NA
Scoured, unspun wool layer
Yellowish thin corrosion NA
Degrading PVC Grey, powdery corrosion NA
on lower edge

268
 Table V
Condition of Magnesium and Major Corrosion Products after the Oddy Test
with a Range of Materials

NC = no chanqe NA = not
Test Conditions Condition after Test Corrosion Product (XRD)

Control (window sill) NC

Material in Oddy Test
- Darkened surface, white corrosion lower edge NA
COz Darkened surface, white corrosion lower edge NA
Oak Heavy corrosion Magnesium carbonate hydroxide
hydrate Mg(COa)+.5HeO
Oak, COz Thick voluminous corrosion Mg(COs)+.5HzO
Plywood Extensive corrosion Mg(COs)a.5HeO
Plywood, COz Several spots of corrosion NA
Formic acid (1 drop) Metal eaten away, white corrosion Mg(COe)a.5HzO
Cardboard Heavy outbreaks, white corrosion NA
Gloss paint Slight corrosion, lower edge NA
Vinyl matt paint Some corrosion on lower half NA
Scoured, unspun wool Heavy corrosion over entire surface Mgs(COe)+(OH)z.4HzO
Degrading PVC Some corrosion on lower half amorphous

Table VI
Condition of AZ3 I, a Magnesium Alloy, and Major Corrosion Products
after the Oddy Test with a Range of Materials

Test Conditions Condition after Test Corrosion Product

Control (window sill) NC

Material in Oddy Test
Black surface, white corrosion, lower edge NA
COz Black surface, white corrosion, lower edge NA
Oak Spots of white corrosion over entire surface NA
Oak, COz Spots of white corrosion over entire surface Magnesium carbonate
hydroxide hydrate
Mg(COg)+(OH)z.5HzO
Plywood White thin corrosion layer over enttre surface NA
Plywood, COz White thin corrosion layer over entire surface Mg(COs)c(OH)e.5HzO
Formic acid (1 drop) Metal eaten away, white corrosion Magnesium formate hydrate
Mg(COOH)z.2HzO
Cardboard White thin corrosion layer and spots NA
of corrosion
Gloss paint White "watery" corrosion NA
Vinyl matt paint White thin corrosion layer over entire surface NA

Scoured, unspun wool Heavy off-white thin corrosion layer Mgs(COs)c(OH)e. 5HeO
Degrading PVC Black surface, white corrosion lower edge NA

269
should be considered when dealing with paints, for help with ethnography collections;
as the rateof emission of corrosive gases usu- J. Heumann of the Tate Gallery and D. Heath
ally decreases with increased drying period. and G. Martin of the Victoria and Albert
Museum for permission to examine parts of
Recommendations for Storage their collections and undertake analysis on
and Display of Modern Metals corrosion products; P. Whelan of Magnesium
Elektron Ltd. for generous donation of samples
Aluminium and Duralumin appear to be stable of magnesium and magnesium alloy and for
towards the gases of common materials used in experimental work; staff members at the Royal
storage and display in museums, with the ex- Air Force Museum at Hendon; R. Funnell,
ception of high concentrations of formic acid. It curator at the Royal Air Force Museum's
is therefore recommended that since formalde- Cardington outstation, for information regard-
hyde may oxidize to formic acid at the metal ing the use of modern metals in aircraft and
surface, high formaldehyde emitters and materi- help and permission to examine parts of the col-
als that can give off formic acid should not be lection and to take samples; V. Tremain and B.
used in conjunction with aluminium and its al- Burr for typing; T. Springett and K. Lovelock
loys. Such materials may include urea-formal- for photography; and the Conservation Unit for
dehyde and phenol-formaldehyde adhesives helping finance the trip to Ottawa to present
and composite wood products whose wood and this paper.
adhesive content may vary from batch to batch.
Although PVC did not cause corrosion in these Suppliers
tests it is not recommended for use.
Magnesium Ribbon
Zinc and magnesium corrode in the presence of Merck Ltd., Broom Road, Poole, Dorset,
organic acids. Thus woods, wood composites, BHI2 4NN, U.K.
paints and adhesives that evolve organic acids
should not be used in the storage and display of Magnesium Alloy AZ3l
these metals and their alloys. Since high humid- Magnesium Elektron Ltd., P.O. Box 6,
ity will promote the corrosion reactions, these Swinton, Manchester M27 2LS, U.K.
metals should be kept in a relatively dry envi-
ronment, for example, below 50% RH. Storing Nylon Monofilament
artefacts made from these metals in stove- Available as fishing line from fishing tackle
enamelled or epoxy-coated metal cabinets will shops or as nylon thread from haberdashers.
provide a stable environment.
R6sum6
Conclusion
Les mdtaux modernes duns les collections
This study has shown that modem metals, par- de musie
ticularly zinc and magnesium, are susceptible
to corrosion by common atrnospheric pollutants Il est bien connu que les matidres utilistes pour la
in museums. However, if care is taken when se- mise en rdserve et l'exposition d'objets anciens
lecting materials for use in storage and display, peuvent libtrer des gaz nocifs qui sont d la source
these metals should not present more of a prob- de divers processus de degradation, dont la corro-
lem than the metals of antiquity. sion. Aussi, pour ivaluerjusqu'd quel point ces
matiires peuvent causer de la cotosion chez les
mttaux, les sptcialistes de la restauration du
Acknowledgement Brilish Museum se servent-ils, depuis 1972, d'un
essai devieillissement accilird qui, connu sous le
The authors thank W.A. Oddy, Keeper of the nom de < test Oddy >, a ttd utilisd pour mesurer
Department of Conservation, for editing and leur effet corrosdsur I'argent, sur le caivre et
permission to publish; S. Bradley for guidance sur le plomb, soit les trois mttaux qui ont le plus
and advice; C. Enderly and P. Attwood for help tendance d se conoder. De telles mesures ont
with coins and medals collections; J. Osborn d'ailleurs permis de rdduire de fagon marqude

270
le nombre des objets qui se corrodent dans les 2. Oddy, W.A., "An Unsuspected Danger in
vitrines d'exposition, et de ralentir le ternisse- Display," Museums Journal, vol. 73, no. I,
ment de l'argent. 1973,pp.27-28.

) I'irsta, des autres mustes, le British Museum 3. Smith, E., ed., Monographs on Industrial
continue, en ce Xf siicle, de collectionner des Chemistry (London: Longman Green and Co.,
objets, de sorte que divers objets des pidces de
monnaie ou des mtdailles, entre autres -faits
1918).

d partir de mttaux modernes I'aluminium, -

-
le zinc, le magnisium ou des alliages de ces
4. Haynie, F.H., "Theoretical Air Pollution and
mitaux &lsormais une partie impor- Climate Effects on Materials Confrmed by
-forment
tante de ses collections. Ces mdtaux se retrouvent Zinc Corrosion Data," in: Durability of
d'ailleurs souvent dans les galeries d'art mo- Building Materials and Components,Proc. of
derne, et ils ne risquent pas moins de se corroder the lst International Conference, Ottawa, 1978,
que les mitaux anciens. On a ainsi constatd que pp.157-175.
des objets en aluminium presentaient une coffo-
sion localisde par piqfires ou une mince couche 5. Skerry, B.S. et al., "Corrosion in
blanche de corrosion. Que des pidces de monnaie Smoke, Hydrocarbon and SOz Polluted
en zinc datant de la premidre et de la seconde Atmospheres - III The General Behaviour of
guerre mondiale, qui avaient ete mises en riserve Zinc," Corrosion Science,vol. 28, no. 7, 1988,
dans des armoires en bois, s'etuient complite- pp.72l-740.
ment corrodtes. Et qu'une couche blanche
superficielle de corrosion s'ttaitformte sur du
6. Rance, V.E. and H.G. Cole, Corrosion of
magnAsium. Ces constatations ne sont pas sans
Metals by Vapoursfrom Organic Materials:
prioccuper les spdcialistes de la restauration,
A Survey (London: Her Majesty's Stationary
qui ont donc entrepis une ttude en we de ddter-
miner les causes de la corrosion, mettre au Office, 1958).
point des tests Oddy pour les matidres que I'on
voudrait emplover pour la mise en riserve et I'ex- 7. Mellor, J.W., A Comprehensive Treatise on
position d'objets faits d partir de ces mdtaux et Inorganic and Theoretical Chemistry (London:
4noncer des directives pour la mise en rdserve et Longmans Green and Co., 1923),vol. [V,
I'exposition d'objetsfaits d partir de ces mdtaw. p.476.

Pour mieux comprendre ces divers processus de 8. Duncan, S., "Aluminium: Its Alloys, Coat-
la corrosion, on a d'abord commencd par identi- ings and Corrosion," in'. Modern Metals in Mu-
fier les matiires qui en sont d la source, et par seums (London: Institute of Archaeology
passer en rewe les documents qui esistent sur Publications, 1988) pp. 27 -32.
le sujet. Une sdie d'expbriences a ensuite 6ta
mende en sefondant sur ces donntes, afn d'etu- 9. Graedel, T.E., "Corrosion Mechanisms for
dier I'incidence qu'ont, sur ces mitaux, les Aluminiums Exposed to the Atmosphere,"
substances polluantes connues qui se
J. Electrochem. Soc., vol. 136, no. 4, 1989, pp.
retrouvent dans les mustes.
204C-212C.
Les directives pour la mise en rdserye et
10. Gorson, M.G., Aluminium and lts Allolts
I'exposition d'objetsfaits d partir de ces mitaux
ont ttd prtpardes. (New York: Van Nostrand Reinhold, 1926).

I 1. Farmer, H. and C. Porter, "Corrosion of

References Metals in Association with Wood. Part 3.
Corrosion of Aluminium in Contact with
1. Blackshaw, S.M. and V.D. Daniels, "The Wood," Wood, vol. 28, 1963,pp. 505-507.
Testing of Materials for Use in the Storage and
Display in Museums," The Conservator,vol.3, 12. Roberts, Sheldon, Magnesium and lts
1979,pp.16-19. l//oys (New York: John Wiley and Sons Inc.,
l 960).

271
13. Beck, A.,The Technology of Magnesiunt 17. Brimblecombe, P., "The Chemistry of
and lts Alloys,3rd edn., A translation from Museum Att," Environmental Monitoring and
German by the technical staffs of F.A. Hughes Control, SSCR Dundee, pp. 56-64.
& Co. Ltd. and Magnesium Elektron Limited
of Magnesium und seine Legierungen (London: 18. Brimblecombe, P., "The Composition
F.A. Hughes & Co. Ltd., 1943)p.298. of Museum Atmospheres," Atmospheric
Environment,vol.24B, no. 1, 1990, pp. l-18.
14. Beck, A., The Technologt of Magnesium,
p. 308. 19. Thickett, D. and L.R. Green, "Re-evalu-
ation of the Testing Procedures for Materials
15. Bradley, S.M., personal communication, for Storage and Display in The British Mu-
1991. seum," The British Museum Conservation
Research Section Internal Report 1991.22.
16. La Niece, S. and G. Martin, "The Technical
Examination of Bidri Ware," Studies in
Conservation, vol. 32, I 987, pp. 97 -101.

272
Finishes on Aluminium A Conservation Perspective
-

Chris Adams and David Hallam

Conservation Section
Australian ll'ar Memorial
Canberra. Australia

Abstract Methods of categorization and analysis developed

and used at the AWM are discussed, including the
In terms of volume of production, aluminium is reproduction of various coatings from the origi-
the world's leading non-ferrous metal. Because nal formulatio ns. Some c ons ervation strategies
ofits lightness and unique physical and chemical have been developed, some are proposed. The
properties, its suitability for the construction of implications of this methodology andfuture pros-
aircraft was recognized early this century. pects conclude the discussion.

Concurrent with the development of aluminium This paper attempts to initiate debate on the
and ils alloys as a construction mateial arose above through a discussion ofthe relevance of
the need for stable and resistant surface finishes. surface coalings studies to tie museum
Su rfa ce fi ni s hin g tec hni q ue s deve I ope d rapi dly i n
both Europe and the United States. Today several Introduction
important classes offinishes are used.
Most conservation projects have a logical begin-
The Australian War Memorial (AWM) has an ning. In July 1982 we decided to restore our
extensive collection ofaluminium aircraft. In this Messerschmitt Me 262A-l aircraft. The Conser-
paper these aircraft are used to illustrate the de- vation Section of the Australian War Memorial
velopment of aluminium alloys as structural mate- (AWM) was given the task of writing the speci-
rials and ofthe associated coating technologies.
fication for the project. We found this task was
Particular emphasis is placed on oxide conver-
sion coatings.
not easy, but it did provide us with a good intro-
duction to aircraft technology. Our attention
In museums, it has been our observation that was soon drawn to the grey-green "anodizing"
thesefinishes have, to a large extent, been ig- on the aluminium surfaces of unpainted por-
tions of the aircraft.
nored
- to the detiment of objects, their history
and cowervation in general. These coatings can
offer important historical and technical informa- Before treahnent commenced, we wanted to
tion. As intinsic parts of the object, every effort know more about the materials on which we
should be made to devise conservation strategies were to work. Communication with other muse-
to retain them. Further, it is only through an un- ums with similar aircraft uncovered a large gap
derstanding of the construction of an object and in their knowledge of aluminium surface coat-
its deteioration that appropriate treatments can ings used in the construction of aircraft. Some
be developed. had stripped the coatings dwing restoration,
others had no idea that coatinss were indeed

273
present. The lack of understanding within the Eventually data on oxide coating types, com-
museum community as to the importance of bined with information on the alloy composi-
metal surface modification was obvious. The tion and paint types should assist in studies of
need to conserve the metal surfaces as well as the authenticity and provenance ofobjects.
the body ofthe object was not appreciated by
those working with technology collections. So far, we have carried out literature and histori-
cal searches on aluminium finishes, history and
We decided to use aluminium oxide coatings technology. We have also developed analytical
as an example of a modified metal surface and approaches to characterize and identiff the coat-
undertake investigations from historical and ings used on aircraft in the AWM collection.
conservation points of view. (The same sort of
study is needed for ferrous metals, magnesium, Historical Information
zinc, copper, and all other modem metals and
their alloys.) In order to understand finishes on aluminium
we must first gain an appreciation of alu-
Fortnnately the Me 262A-l restoration project minium as a structural material. It has been ar-
was aborted and the aircraft was returned to gued that it did not supersede wood as the main
storage for further study and a more sympa- aircraft construction material until its corrosion
thetic treatment was undertaken 10 years later. behaviour was understood.and control proc-
Our approach to aircraft is now quite different esses had been developed.' This was accom-
in that it is based firmly in the conservation plished independently in several countries. In
ethic. the Unlted States. Alclad was developed by
Alcoa.' The English developed anodizing.'
Aims The Europeans, particularly the Germans, de-,
veloped the use ofoxide conversion coatings.-
In this study we aimed to confirm historical The Japanese developed the oxalic acid
sources of information about the treatment of anodizing process.'
aluminium and the manufacturing of aircraft.
We also wanted to confirm or add to the infor- Aluminium finishes can be divided into the fol-
mation on the technology transfer between the lowing classes: mechanical finishes; chemical
Germans and Japanese during the 1930s and furishes; metallic coatings; organic and inor-
1940s. ganic coatings; anodizing; and chemical conver-
sion coatings. In our study we are primarily
As conservators we wanted to understand interested in the chemical conversion coatings,
more about the history and technology of fin- but we will also briefly consider the other
ishes on aluminium; the chemistry of the alu- classes.
minium surface and its interactions with the
environment; and the effects of any possible Aluminium Surface Finishes
treatrnent on the aluminium surface and its
long-term stability. The oxide film on aluminium forms immedi-
ately on exposure to the atmosphere. (Oxide
This work is by no means complete, but we film thicknesses are shown in Table L) It has a
hope that it will go some way towards a better molecular volume of 1.5 times that of the met-
understanding of the behaviour of aluminium al. It is thus under compressive stress, and cov-
so that we may develop conservation ap- ers the metal continuously. Furthermore, it can
proaches that maintain an object's integrity. In cope with substantial deformation without rup-
museums the general approach to aluminium ture. As the film grows naturally, it breaks
is to assume that it is unstable. Thus any treat- from the metal surface to form its own struc-
ment cornmences with the stripping of the un- ture, and this reduces the stress on the film. Im-
stable natural oxides along with the synthetic perfections develop in this transition, and these
conversion coating, followed by the application accelerate the breakdown of the film. Break-
of inhibited or uninhibited organic coatings. down is also accelerated by the solubilization

114
 Table I
Thiclmess of Oxide Coatings on Aluminium

Goatings Thickness in nm

Natural oxide on Al or Al-Mg alloy

(formed below 300'C) 1to3
Naturaloxide on Al
(formed above 300'C) 30

Naturaloxide on Al-Mg alloy

(formed above 300'C) 3,000

Normal chemical oxide coating

(e.9., MBV, Alrok, etc.) 2,500 to 5,000

Normal barrier layer anodizing 250 to 750

Normal protective anodic coating

(e.9., Sulphuric acid anodizing) 5,000 to 30,000

Hard anodic coatings

(e.9., for engineering purposes) 25,000 to 150,000

ofthe film by aggressive anions, such as chlo- normally undertaken as part of a rough fitting
rides and sulphates. or frnishing process.

By growing the films in a controlled manner Abrasive Blasting: Normally this is done to
during the manufacture, thicker, compact produce an even finish on castings or forgings
films with a higher corrosion resistance can be and to remove the thick oxide laye; formed
formed. Althoush this was realized in the during the manufacturing process.o
lgth centuryf iiwas not developed until l9l6
in Germany." Scratch Brushing: Wire scratch brushes are
used to produce a matt surface. The bristles
Mechanical Finishes move at 90 m to 120 m per min (300 ft to
400 ft per min). The surface may also be pick-
Mechanical finishesT are created by cutting and led to enhance the effect. Parts so treated are
milling. These mechanical processes work the normally decorative and are lacquered. Unlac-
upper layers of the metal and impose stress. quered components can be cleaned with an
Mechanical finishes may either increase or de- eraser.
crease susceptibility to corrosion. The more
stressed the surface, the more likely corrosion Burnishing and Polishing: The same type of
is to occur. Hence a polished surface should be methods are used for aluminium as for other
more stable than a rough milled surface. The metals.
basic mechanical finishes are as follows:
Ball Burnishing: Rotating barrels filled with
Grinding and Dressing: This is done with metal balls and a silicate detergent are used to
emery wheels or abrasive belts with an open produce a bumished surface on small parts.
abrasive structure to prevent clogging. It is

275
Rough Polishing: This is used as a preliminary Acid Treatments of Aluminium: Early treat-
step to other finishing techniques to remove the ments of aluminium seem to have used non-
scratches from dressing and remnants of the etching chromic acid, since traces of chromate
casting skin. Normally the abrasive (silicon have been found on the aluminium engine cowl
carbide or alumina) is belt- or wheel-mounted, skin from the SPAD aircraft at the National Air
and runs at speeds of 1,800 m to 2,400 m per and Space Museum in Washington, D.C. dur-
min (6,000 ft to 8,000 ft per min). ing restoration treatment. A typical chromic
acid solution is composed of: chromic acid 5o/o,
Polishing or Buffing: Mops with polishing sulphuric acid l5o/o and water 80%. The earliest
compound of alumina, tripoli, and chrome ox- reference we have found is Bengston.g
ides are used. The compound is applied to the
mop as a wax mix, for example, one part mon-
I
Hydrofluoricl0 and phosphoric acid basedl so-
tane wax and three parts alumina. The part is lutions are used widely to remove oxides and
held against the mop until the glazing scratches etch the alloy surface. Typical formulations are
are removed. Final polish is done on a loose the following:
mop after removing the previous polish resi-
due. Tints can be produced by using a final hydrofluoric acid (40%) r0%
polish that is coloured. nitric acid (specific gnvity 1.42) r0%
water 80%
or:
Chemical Finishes
phosphoric acid r0%
Frosting: Frosted finishes are normally pro- butyl alcohol 40%
isopropanol 30%
duced in either a 20%ohot (50'C to 80 'C)
water 20%
caustic soda solution or a mix of nitric and hy-
wetting agent (i.e.,
drofluoric acid. Both processes involve control- naphthalenesulphonicacid) traceo/o
led corrosion ofthe object and produce a
surface full of micro-pits.6
Metallic Coatings
Acid Cleaning: This practice primarily in Chemical: Baths have been used to cement
mineral acid by immersion or application is a thin coatings of copper to aluminium before _ -
common pre-treatnent for aluminium in the electroplating, without the use of electricity.l2
fabrication of a component. Acid cleaning is
used in the following ways: cleaning with little Electroplate: Aluminium can be electroplated.
or no etch prior to organic coating or electro- In the AWM we have examples of nickel,
plating; removing smut and neutralizing after al- chrome and copper plating. Most of the electro-
kaline etching; removing oxides; matt or bright plated objects are ofJapanese origin, and none
etching for decorative fmishes; deep etching of have survived well.
pattern or relief; very deep etching for chemical
milling or contouring; undercut etching for ap- Applied Metal Coatings
plication to teflon (PTFE); and electrolytic etch-
Spray: Spray coating of pure aluminium onto
ing for lithographic printing processes.
castingq a rare type of treatment that predates
^is
Alclad."
Etching: Aluminium can be etched using a bi-
tuminous relief and an etchant composed of
Roll-on: Work by Dixla at Alcoa to create
100 mL of lUYo copper sulphate, 20 mL of con-
Alclad changed the use of aluminium in aircraft
centrated hydrochloric acid, and glycerine as a
and resulted in the demise of wooden struc-
thickening agent.
tures. As a rule, the outer pwe aluminium coat-
ing is 5% to l0%o ofthe total cross-sectional
Chemical Milling: Developed in the U.S. in area. Duralplat, Alplat, Albondur, Avionalplat,
1949, this is also a form of controlled etching
and Chitonal are but a few of the trade names
that is now used to machine material without
used in the late 1930s by other frms through-
stress.
out the world producing clad aluminium.

276
Anodizing: At present 5%.9f all aluminium Boiling: Thisprocess is used in the production
manufactured is anodized.ls In the 1930s and of caoacitors." but was used earlv this cenrurv
1940s it was used most extensively on naval as a general finishing method on
aircraft and in particular on seaplanes. Anodiz- aluminium.
ing processes can be identified visually, al-
though at times process colouring can make Acid Treatment: The finishes produced by
this more difficult. Anodized surfaces are stable phosphoric acid and chromic acid mixtures
except when exposed to industrial and saline used during manufacfure can be considered
environments for periods longer than 40 years. to be conversion coatings, as they leave alu-
Most of the processes are based on chromic minium chromate and phosphate residues on
acid, sulphuric acid or oxalic acid. the surface. These residues have a slight grey
colour and some iridescence. Often these treat-
Chromic Acid: The process of anodizing was ments were done as part of the foundry or fabri-
flrst introduced as the Bengough-Stuart'o chro- cation shop process to clean the metal and were
mic acid process in 1927 for the protection of not applied specifically for corrosion resis-
Duraluminium seaplanes. The process is still tance. Thus they differ significantly from the
widely used in Britain under specification DEF later (post 1945) conversion coatings.
l5l. In the U.S. this process y;as modifred by
the Battle Memorial lnstitute' ' using more con- Oxide Films
centrated solutions, a constant voltage and addi- These tend to be derived from alkaline based
tions of sulphuric acid. treatrnents containing chromates. The main
types are as follows:
Sulphuric Acid: The use of sulphuric acid for
anodizing was developed in the U.S. and pat- Bauer Vogel @V): The aluminium is im-
ented in Britain in 1927 bv C.H.R. Gower and mersed in a boiling solution of alkal,ine carbon-
S.O. O'Brien. ate and chromate for about an hour.'

Oxalic Acid: The oxalic acid process was first Modified Bauer Vogel (MHV): This modifica-
developed in Japan. The Germans adopted it tion on the original process'' reduces the treat-
under the name Eloxal during the 1930s and ment time to five minutes and produces a light-
1940s (note: German authors also use the name to dark-grey fikn. The intensity of the colour
Eloxal GS to denote a direct-current process depends on the time of immersion. The coating
employing sulphuric acid). A second oxalic can be sealed with sodium silicate. Colour can
acid based process called Ematal was devel- vary with immersion time, surface state and al-
oped in 1935 in Swizerland. loy composition, as shown in Table II.

Electro-brightening: Bright anodizing is used Erftwek @W): This22 was based on the MBV
to produce such items as reflectors on lights process but contained silicate and could be used
and automotive trim (e.g., Volkswagen logos on alloys of high copper content, giving a much
and Volvo bumper bars). Two processes were lighter colour than the MBV process. The solu-
patented in 1936 in England and the U.S. The tion contains flu1ryide, chromate and carbonate.
English Bytal processrd patented by the British The related LW" process produces a clear
Aluminium Company is based on sodium car- coating.
bonate and tri-so,{ium phosphate solutions. The
American Alzak't process uses a fluoroborate Alrok: This American version of the MBV
solution. The finishes resulting from these proc- process was developed by Alcoa and used po-
esses have a high reflectance but do not have tassium carbonate and sodium dichromate. The
outstanding corrosion resistance. protective film was sealed with a potassium di-
chromate solution.
Chemical Conversion Coatings
Pylumin: This process23 is based on di-sodium
These can be classified and described as phosphate and was used by the British.
follows:

277
 Table II
Effects of Surface State and Alloy Composition on Appearance of MBV Coatings

Alloy State of Appearance with Casting Skin

o/o
Composition Surface Intact Removed

Al 99.8 Fine Green/grey Green/grey

si 0.08
Fe 0.11 Coarse Spotty Green, metallic,
spotty
AI 99.2 Fine Green/grey Dark grey
Si 0.36
Fe 0.41 Coarse Spotty Green/grey,
local light
spotting
AI 99.5 Fine Milky grey Light grey
Si 0.16
Fe 0.22 Coarse Spotty Light grey

Acid Processes Conversion Coatings reproduce known coatings using original formu-
The principal acid-based processes were devel- lations and procedures obtained from original
oped in the 1940s in the U.S. lmmersion times are sources. We considered three processes, Bauer
typically 30 seconds to five minutes at room tem- Vogel, Modified Bauer Vogel and Erftwerk, to
perature. The main types are as follows: be appropriate. These would provide a good
representation of the processes likely to be
Phosphate Conversion Coatings: The solu- forlnd on the aircraft to be studied.
tions contain phosphate, chromate and fluoride
ions. Films produced are greenish in colour and In all the original formulations we recognized
consist of chromium and aluminium phos- that immersion time was the major variable in
phates. Common trade names are Alodine producing coatings of appropriate thicknesses
and Alocrom. that possess sufficient adherent properties.
These times were varied significantly by the
Chromate Conversion Coatings: More re- original formulators either inten-ti_onally to ob-
cently, solutions containing chromate fluoride taiicertain properties or effects25 or acciden-
and nitrate have been used to produce a golden tally as a result ofpoor quality control.
brown coating containing chromate as a major
constituent (e.g., Alocrom 1200 and its Therefore, a certain amount of experimentation
derivatives). was necessary initially to replicate the homoge-
neous coating that could occur on the objects
As environmental legislation restricts the use of to be investigated. Figures I and2 show scan-
chromium, manufacturers are now avoiding its ning electron micrographs of MBV coatings
use and tending towards the use of phosphates produced by immersion for 25 minutes and
and an {ustralian-developed cerium-based 5 minutes respectively. The 25-minute sample
coating.24 shows that the adherent properties of the layer
have been lost. The S-minute sample adheres
Analytical Techniques to and is, apparently, quite homogeneous in nature.
Characterize and Identify Coatings
on the AWM Aircraft Samples were then taken from three aircraft
from the Second World War period, namely,
Before any analytical techniques were a Messerschmitt Me 109G-6 (1945), a
investigated, we considered it appropriate to Messerschmilt Me 2624-l (1944), and, a

278
 experience, it is sometimes possible to differen-
tiate between chromate coatings, anodization
and chemical conversion coatings by noting
interference fringes and colour differences.
Although visual examination is useful, we
require far more specific information than it
can provide.

Optical Microscopy: lnformation obtained by

cross-sectioning, polishing, etching, and visual
examination under a high-powered optical mi-
croscope can provide useful inf^o-rmation for the
identification of the base alloy.'' The nature of
the base metal is of significance, as historically
Figure I MBV coating produced by immersionfor it is known that only certain classes of goatings
25 minutes. were used with certain types of alloys.'

Structure samples taken from the three aircraft

studied are shown in Figures 3, 4 and 5. Al-
though they have similar chemical composi-
tions, the grain structure of these samples are
somewhat dissimilar and display the fact that
their manufacfuring processes were not the
same. For example, it is apparent that although
all the samples have been hot- or cold-rolled,
the more rounded nature of the grains in the
Me l09G-6 sample indicate that an extensive
process of annealing has occurred. Conversely,
the very elongated grains exhibited in the other
two samples indicate that no annealing was un-
dertaken. It was further noted that the only non-
Figure 2 MBV coating produced by immersionfor
clad alloy was that of the Zero fabrication sheet.
5 minutes.

Optical microscopy can also be used to reveal

Mitsubishi A6M2 Zero (1942\. These were cho-
the presence ofa surface coating, although
sen because they had a high treatment priority
no information as to its composition can be
and because they represented typical examples
gained. In some instances microchemical
of the mass-produced fighting aircraft of two
nations. Also, these aircraft were considered to
techniques have revealed the presence of
be in their original condition, as their histories
were known.

Our primary aim was to investigate and, if r.::': .

necessary, develop techniques by which we t..'.

could categorne and compare surface coatings :i
: ,i,.. iK
I

in detail. 'ir
.
;:..e
Examination Techniques ..-..,..'.i
:.:, ,
Visual Examination: This type of examination
ofan alloy surface can quite often reveal the .*M|:
presence of some sort of inorganic surface Figure 3 Me l09G-6 fabrication sheet, base metal
coatins or other surface finish.'o With structure (1,000x).

279
 equipment used for these studies was a
Cambridge Stereoscan 360 fitted with a Tracor
Northem EDXA analyser using a standardless
semi-quantitative analysis program. A sum-
mary of base alloy analysis fiom the three
study aircraft is presented in Table III.

These results indicate that all were alu-

minium/copper alloys, which would place them
in the modern categorization of the 2,000 series
of copper-containing aluminium alloys. It
should be noted that aluminium and masne-
Figure 4 Me 2624-l /itbrication sheet, tail section sium concentrations are combined becaJse
base metal structure (1,000x). the aluminium Ka peak and the magnesium Ka
peak are close in energy distribution at the con-
centrations encountered, and differentiation is
impossible. To identiff positively a particular
alloy, the magnesium content should be known.
;-: Problems arise however, when the composition
of the surface coatings are to be studied. Typi-
cally the film thickness of a chemical conver-
sion coating will range from 0.5 microns to
5 microns. Figure 6 shows a SEM image of
the conversion coating on the sample taken
from the Zero drop tank. Operatingat a
20 KeV energy level, that which is required
Figure 5 Mi A6M2 Zerofabrication sheet, drop for an analysis of this tlpe, the electron beam
tank base metal structure (1,000x). has a penetration into aluminium of around
1.0 micron to 3.0 microns, depending on cir-
chromium and other metals associated with cumstances and a minimgm diameter of ap-
oxide coatings.28 proximately 1.0 micron.29 Thus, the abilii
to detect elements that are perhaps present in
Scanning Electron Microscopy and Energy small proportion from a thin layer is severely
Dispersive X-ray Analysis (EDXA): When hampered. The analysis obtained in most cases
combined with information obtained from the is that of the underlying base metal rather than
optical microscope regarding grain structure, that of the surface coating. Some success has
qualitative and quantitative elemental analysis been achieved, however. Figure 7 indicates an
of the base metal will aid in classification. The EDXA trace of the analysis of the drop tank

Table III
Standardless Quantitative EDXA Analysis of Base Metal Alloys

% by weight composition
Al&Mg Mn Cu Fe

tank
Zero drop 96.93 0.45 2.39 0.22
sheet
Zero fabrication 96.54 0.40 2.98 0.08
Me109G-6 fabrication sheet 96.90 0.51 2.44 0.14
Me262A-1 tail 96.83 0.41 2.43 0.34
Me262A-1main plane 96.71 0.49 2.58 0.22

280
 X-ray Photoelectron Spectroscopy (XPS):
XPS analysis uses low-energy X-rays (from
aluminium or magnesium targets) to produce
a spectrum of photoelectrons and Auger elec-
trons. This non-destructive technique yields
quantitative data regarding chemical state, oxi-
dation state and elemental composition of the
surface. Argon-Ion-Depth-Profiling is a tech-
nique that uses a focussed beam ofargon ions
to etch material at controlled rates. XPS is then
used to analyse the exposed surfaces. Data ob-
tained include percentages of atomic concentra-
tions. The equipment used for these studies was
a Kratos IXSAM 800 from the University of
Figure 6 Scanning electron micrograph showing
thickness of typical chromate conversion coating New South Wales.
(Zero drop tank).
Table IV is a summary of the analysis of a
sample taken from the Me l09G-6. Data were
obtained with surface contamination removed
and depth profiling at2nrn intervals down to
a maximum depth of 100 nm. The high carbon
and silicon content might suggest surface con-
tamination. However, as the silicon is present
as the oxide, it is unclear_i_f this arises from a
water-glass over-coatingr) or due to plasticizer
contamination from polymeric storage contain-
Figure 7 EDXA analysis of Zero drop tank ers. It was only after surface etching that the
conversion coating. true composition of the conversion coating was
revealed. Since chromium was found in signifi-
from the Zero where chromium has been cant amounts (shown in the 100 nm analysis) it
detected. is obvious that a chromate conversion coatins
is present.
Obviously the thicker the coating, the easier it
is to detect chromium. Although it was deter- The presence of chromium and aluminium in
mined that coatings were present in some in- the oxide form was demonstrated, together with
stances they could not be specifically identified. sodium and chlorine in significant amounts in
the bulk material of the surface coating. The for-
To date, experiments of mounting the sample mer confirms the existence of a chromate-based
with different orientations to increase the re- coating.
sponse from the surface, both in cross-section
and plane-section, have proved to be oflittle As yet we have not progressed to the stage of
advantage. being able to differentiate between surface coat-
ings produced by similar processes (e.g., BV,
X-ray Diffraction (XRD): XRD of swfaces EW and MBV). A far larger sample base will
treated with the MBV process indicated the be needed for this. Indications are that a suffi-
presence of an aluminium oxide polymorph. ciently accurate procedure can be developed.
No chromium species were detected, probably
because if there were any, they would be in an Summary
amorphous form. If any crystalline species
were present, they would have a peak probably All the alloys examined so far have been
masked by the large aluminium peak. This found to be copper-containing, equivalent to
technique therefore does not appear to be the 2,000 series. All samples, except that flom
very useful. the Zero drop tank, were of the clad type,

281
 Table IV
XPS Analysis of Surface of Me 109G-6 Fabrication Sheet

ELEMENT CONCENTRATION (% by atomic weight)

As Presented With Removal of 100 nm

Surface Contamination into Bulk
Aluminium 10.83 30.04 50.07
Chromium 0.11 0.77 1.09
Silicon 1.57 0.00 0.00
Sodium 0.16 0.10 0.22
Oxygen 22.12 33.82 34.23
Carbon 63.62 34.41 13.57
Chlorine 0.48 0.85 0.81
Nitrogen 1.10 0.00 0.00

where the copper-containing core has been can enable identification of the chemical ele-
covered with a more corrosion-resistant layer ments at known depths within the bulk material.
of almost pure aluminium. Grain structurei
vary, indicating somewhat different methods of The results so far fall somewhat short of our
manufacfu re and/or variations in post-manufac- goal of obtaining specific chemical composi-
ture freatment procasses. tions and differentiating between similar proc-
esses. Notwithstanding, enough encouragement
The Zero drop tank and the Me l09G-6 fabrica- and information has been gained to suggest the
tion sheet both revealed the presence of a chro- detail of information that we might expect in
mate chemical conversion coating. We could the future.
not fingerprint the type of oxide coating nor
could we attribute it to a particular process. On the down side, XPS analysis requires so-
phisticated and expensive equipment and is
The aim of this preliminary investigation, to de- time consuming in its implerqentation. The
vise a practical and effective strategy to differ- minimum sample size Q mm') needed for this
entiate and identify coatings, has not proved to technique is also of concem, posing both practi-
be a simple task. We have demonstrated that no cal and ethical problems.
single technique provides all the information
necessary to be able to reach positive conclu- This exercise demonstrates that both technical
sions as to the treatments used on the alu- and historical information must be combined
minium. with accurate material analysis in order to ob-
tain a balanced perspective when studying the
Optical Microscopy combined with EDXA deterioration ofthe object and developing sub-
analysis provides information enabling the iden- sequent conservation strategies.
tification of the base alloy and manufacturing
processes, but falls down when examinins the Conservation of Aluminium
added surface coating (EDXA is not a trui sur- with Oxide Films
face analysis technique). The preferred method
of approach for examining these surface coat- Stability of Aluminium: Any conservation
ings is to use X-ray Photoelectron Spectros- strategy depends on the stability of the material
copy (XPS). This technique can provide the and the environment to which the material was
composition analysis of a swface and when exposed in the past. Aluminium objects that
combined with Argon-Ion-Depth-Profiling, it have not been exposed to an aggressive environ-
ment are quite stable. Oxide-coated obiects are

282
even more stable. Generally speaking, alu- Electrolysis: This technique is described by
minium exposed to clean environments should C. Degrigny elsewhere rr this publication.
need little treatment. Dust, sulphates, chlorides
and high relative humidity (greater than 80%) Use of Cqrium-based Coatings: Cerium-based
will break down the stable natural oxide film. coatings35 have been used by il{cl-eod to treat
The porous oxide coating that results is non- the hull of the l2-metreyachtAustralia 2, and
protective and fraps pollutants, hence acceler- they appear to work well.
ated corrosion begins.
Conclusion
Envilonmental Corrosivity: This can be meas-
ured3o and quantified. Ifobjects are stored in- The sand-blasting approach to the conservation
doors in dry climates (e.g., Canberra or Arizona of large technological items should be made
with an average relative humidity (RH) below extinct. Armed with the knowledge obtained
55%) corrosion will be minimal, unless the fiom historical sources and the dataprovided
object has been previously rendered unstable. by modem analytical procedures (that are now
within the reach of every major conservation es-
Stabilization of Aluminium Oxide Films: tablishment) there is no longer any real excuse
Merely placing an object indoors qe.duces its for the destruction ofobject integrity during
corrosion rate by a factor of I,000" and the conservation treaftnent. An integrated approach
use of a cloth cover will reduce this another that involves the development of the following
20 times. A further simple step in reducing cor- four main strategies must be adopted:
rosion is the use of dehumidification to reduce
RH. It has^b^een a standard military practice for o A sound understanding of the nature of the
some time" and could easily be adopted by mu- surfaces of modem metal objects, which
seums. It should be noted that the use of dehu- includes base metal chemical coatings and
midifiers will extend the lifetime of anv overlying organic coatings.
corrosion inhibitor.
o An understanding of the interactions of
Use of Inhibited Wax Coatings: Although modern metals with their environment. For
inhibited wax coatings (based on petroleum example, we know that clean aluminium will
sulphonate and tertiary amines) are used on op- not corrode in a clean environment at any-
erational aircraft, their use in a museum needs thing less than 80% RH. At the other extreme
thorough investigation. (At the AWM, Dinol-
we know that aluminium removed from sa-
inhibited wax coatings are used. The Duxford
line environments will continue to corrode
branch of the Imperial War Museum uses the
actively even at low RH values. We need to
same compounds in an annual maintenance pro-
study other types of environments in order to
gram on outside display items.) The concern is
develop strategies for protection.
that these products are being used without
enough basic information or knowledge of their
long-term effectiveness. " r A study ofthe nature and effectiveness
of post-treatment organic coatings, such as
Stabilization by Altering the Aluminium Water Displacing Corrosion Preventatives
Surface : Traditional abrasive surface cleanins (WDCP). Electrochemical methods of evalu-
methods, such as glass beads at 3 . I x lOs Pa
- ation also appear to hold promise.
(45 psi) to 6.9 x l0' Pa ( 100 psi). remove the
oxide coatings. By using plastic abrasives'" o The development of techniques for the
and low pressures (below 1.0 x l0) Pa differential removal of non-original organic
[15 psi]), porous oxides can be removed and coatings. Working from the premise that the
the conversion coating left intact. Care must be integrity of the original surface should be
exercised in the choice ofthe abrasive grade maintained, it is imperative that less aggres-
and delivery pressure. The treated surface sive techniques are developed for the removal
should be coated with an inhibited wax. of overlying layers. Recent developments by

283
 Richard Wolbers for the treatment of works rev€tenent. Cette lacune a pu avoir des rdpercus-
of art and fumiture using gel and emulsion sions nbgatives tant sur le plan de I'historique et
technology should be investigated further. des objets eux-m€mes que sur celui de la conser-
The use of low velocity, plastic-bead blasting vation en gdn,lral. Nous demeuronsfermement
convaincus que ces revefuments peuventJburnir
for the removal of aluminium oxide corrosion
d'importants ren seignements d' ordre tant histo-
products should also be studied in detail.
rique que technique. Et, comme ils font partie
integrante de I'objet, il devient impdrieux
Acknowledgement d'dlaborer des strattgies qui permettront d'as-
surer leur conservation. Qui plus est, la mise au
The authors thank the Aushalian War Memo- point de techniques de traitement appropri6es
rial; the National Air and Space Museum of the passe ntcessairement par une connaissance du
Smithsonian lnstitution; and the Canadian Con- mode de construction des objets et des facteurs
servation Institute. Scanning Ele ctron Micros- qui tendent d les d,ltdriorer. Nous ddcrirons les
copy and Energy Dispersive X-Ray Analysis mtthodes de catdgorisation et d'analltse qui y ont
(EDXA) were performed at the Electron Mi- 6ti mises au point et utilisbes, dont la reproduc-
croscopy Unit of the Australian National tion de divers rev€tements d partir des prtpara-
University, Canberra, Australia. X-ray Photo- tions originales. Apris avoir tlabort ou propost
certaines stratigies de conservation, nous
electron Spectroscopy was performed at the
conclurons en traitant des rdpercussions d'une
Surface Analysis Facility, University of New
telle mdthodologie et enfournissant un apergu
South Wales.
d es pe rspect ives d' aven i r.

R6sum6 References
L'application de revAtuments sur l'uluminium :
un point de vue axi sur la consewation
l. Schatzberg, E., Dissertation; Excerpts
Related to Dural Corrosion (Philadelphia:
L'aluminium demeure le mdtal nonfeneux le University of Philadelphia, June 30, 1989).
plus produit dans le monde. Dds le dtbut du
siicle, on a constatd que, dufait de sa l4gdreti et 2.Dix, E.H., "Aluminium Clad Products," in:
de ses proprittts physiques et chimiques uniques, Engineering Laminates ed. A. Dietz (New
il convenait bien d la construction des adronefs. York John Wiley and Sons, lnc., 1949) pp.
382-402.
Or, avec la mise au point de I'aluminium et d'al-
liages d' aluminium utilisables comme mattriau 3. Bengough, G.D. and H. Sutton, "The
de construction, est nt le besoin de crder des Protection of Aluminium and Its Alloys
rev€tements de sudace qui soient d lafois stables against Corrosion by Anodic Oxidation,"
et rtsistants. Ces rev€tements de sudace se sont Engineering no. 122, August 27, 1926,
diveloppds rapidement en Europe et aux Etals- pp.274-277.
Unis, si bien qu'aujourd'hui plusieurs grandes
catigories de rev€tements de ce genre sont 4. Bauer, O. and O. Vogel, International Zeit-
utilistes.
schrift fur Metallographie, v ol. 8, I 9 I 6, p. I 0 I .
L'Australian l(ar Memoial possdde une collec-
5. Setoh, S. and A. Miyata, "Electrolytic
tion importante d'adronefs en aluminium. Et ce
sont ces a,lronefs qui sont utilisds pour illustrer, Oxidation of Aluminium and Its Industrial
dans la prtsente communication, I'ivolution des Applications," in P ro c e edin gs l(or I d
alliages d'aluminium servant dans la consttuc- Engineering Congress Tolqto I 9 29, vol. 22,
tion, ainsi que les techniques connexes de rev€te- 1931, pp. 73-100.
ment, dont, d'unefagon toute particuliire, les
rev€tements constituds d'une couche d'oxyde. 6. Buff, H., Liebigs Annalen der Chemie,
vol. 102, no. 3, 1857, p.265.
Aux musdes, nous avons constatd que, dans
une large mesure, I'on n'a pas tenu compte du

284
7. Samuels. L.E.. "The Effects of Mechanical 19. Mason, R.8., U.S. Patent 2108603, 1938, as
Finishing Processes on Metal Surfaces," in: The cited in Wemick et al., 1987.
Finishing of Aluminium ed. G.H. Kissin (New
York Reinhold Publishing Corporation, 1963) 20. Altenpohl, D.G., "Use of Boehmite Films
pp. 192-214. for Corrosion Protection of Aluminium," Cono-
sion. A Journal ofScience and Engineering,
8. Von Zeerleder, A., The Technologt of vol. 18, no. 4, April 1962,pp.143-153.
Aluminium and lts Light Alloys (Amsterdam:
Nordemann Publishing Company, I 936). 21. Eckert, G.,Aluminium,vol. 19, 1937,
p. 608, as cited in Kissin, G.H., 1963.
9. Bengston, H., "Methods of Preparation of
Aluminium for Electroplati rg," Tr ens act io n s of 22. Y er einigte Aluminium Werke A.G.,
the Electrochemical Society, vol. 88, 1945, German Patent 6781 19,1937 , as cited in
p.307. Wemick et al., 1987.

10. Bullough, W. and G.E. Gradam,Journal 23. "Pylumin Process," British Patent 441088.
of the Electrodeposition Technical Society,
vol.22,1947,p. 169. 24. Amott, D.R., B.R.W. Hinton andN.E.
Ryan, "Cationic-film-forming Inhibitors for the
I 1. Bowen, 8.W., Aluminium and Non-Ferrous Protection of AA 7075 Aluminium Alloy
Review, vol. 3, 1938, p. 16, as cited in Wernick against Corrosion in Aqueous Chloride Solu-
et al., 1987. tion," Corosion,vol.45, no. 1,January 1989,
pp.12-18.
12. Wemick, S., R. Piner and P.G. Sheasby,
The Surface Treatment and Finishing of 25. Kissin, G.H., The Finishing of Aluminium
Aluminium and lts Alloys (Teddington, (New York: Reinhold Publishing Corporation,
England: Finishing Publications Ltd., 1987). 1963).

13. Ministry of Defence, "Protection of Alu- 26. Galler, J.W. and E.J. Vaughan, Protective
minium Alloys by Sprayed Metal Coatings." Coatings for Metals (London: Charles Griffin
Defence Standard 03-3(l), December 28,1970. & Company Limited, 1950).

14. Dix. E.H.. "ALCLAD": A Nqw Corrosion 27. Greaves, R.H. and H. Wrighton, Practical
Resistant A luminium Product (Washington Microscopical Metallograplry (London:
D.C.: Metallurgist, Research Bureau, Chapman Hall, 1960).
August 1927).
28. Biestek, T. and J. Weber, Electrolytic and
15. Gabe, D.R., Principles of Metal Surface Chemical Coatings, A Concise Survqt of Their
Treatment and Protectio,n (Oxford: Pergamon Production, Properties and Testing (Redhill:
Press.1978). Portcullis Press, 1978).

16. Bengough, G.D. and J.M. Stuart, British 29. Goldstein, J.I. and H. Yakowitz, Practicql
Patent 223994, 1923. Scanning Electron Microscopy (New York:
Plenum Press, 1 977).
17. Slunder, C.J. and H.A.H. Pray,Metal
Finishing, vol. 44, 1946, pp. 419-483, as cited 30. Bemdt, H., "Measwing the Rate of
in Wernick et al., 1987. Atmospheric Corrosion in Microclimates,"
Journal of the American Institutefor
18. Pullen. N.D.. "Protection from Corrosion Conservation, vol. 29, 1990, pp. 207 -220.
of Aluminium and Its Alloys," Metal Industries
(London), vol. 44,1934, pp. 133-136, I 87-l 88.

28s
31. Rice, D.W., R.J. Cappell, W. Kinsolving 34. Non-Chemical Depainting Compounds,
and J. Laskowski, "Indoor Corrosion of Equipment Technologt. Suggestion for Dry
Metals," Journal of the Electrochemical Blast Paint Removal on Aircraft (Notre Dame,
Society,vol. 127, no.4, April 1980, Indiana: Aero-Blast Products Inc., 1984).
pp.89l-901.
35. Hinton, B.R.W. and R. Arnott David, "The
32. Cargocaire Inc., The Dehumidification Characteristics of Corrosion Inhibiting Films
Handbo o k (Amesbury, Massachusetts: Formed in the Presence of Rare Earth Cations,"
Cargocaire Engineering Corporation, I 988). Microstructural Science, vol. 17, 1986,
pp. 3l l-319.
33. Tromans, D., "Effect of Organic
Adsorbents on the Aqueous Shess Corrosion
Cracking of AA 7075-T651 Aluminium
Alloy," Corrosion. The Journal of Science
and Engineering, vol. 42, no. I 0, October
1986, pp. 601-608.

286
Radiation Hazards in Museum Aircraft

John Ashton

Conseryation Section
Australian Whr Memorial
Canberra. Australia

Abstract Museums Program. This document mentioned

U.S. Federal and Air Force directives that dealt
A survey of the possible radioactivity hazards as- with the possession, use and precautions of ra-
sociated with museum aircraft and their compo- dioactive material and it also described work-
nents was carried out on the Australian War able radiation hazardpolicies. One of the most
Memorial collection. The survey was performed telling paragraphs in that document warns "we
using a portable contamination monitor to detect
must all recognize our obligation to minimize
the presence ofradioactive material and an ioni-
the potential for inadvertent radiaJion exposure
zation chamber suruey meter to show the dose
rate measurements.
to the public and the workforce."r

The survey results show a large number ofradio-

Although radium-painted dial faces are the best
active sources both in aircraft and in individual known sources of radiation in aircraft, several
instruments in storage. Some of the instruments types of radioactive materials have been used
indicate an exposure rate ofover 300 mR/h, in aircraft, including tritium, promethium- I 4 5,
and one switch in a German aircraft indicates radium-226, krypton-85, magnesium-thorium,
60 mUh. The Npes of instruments. parent air- and depleted uranium. For example, the Das-
craft, and exposure rate readingsfor surface and sault Mirage III-0 jet aircraft (built under li-
100-mm distances are listed. cense in Australia) is equipped with an Atar
engine that has the compound magnesium-
The different types of radioactive material used in thorium in the compressor halves and centre
aircraft, the terminolog,, used in radioactivity castings. A special fatigue and creep inhibitor
measurement, and the safe handling, storage and in light alloys of the Mirage, known as ZTI and
labelling of radioactive aircraft components are also containing magnesium-thorium, w€ts used
discus sed. Suggestions for methods of protecting in the 1970s and later. This material was consid-
the public and museum stafffrom exposure to
ered a hazard and thus the castings, foundry
radiation are offered.
pieces and grindings were stored in dispersed
areas, although the material stock had very low
Introduction radiation readings. The air intake castings of
the Mirage have rare earth alloys in their com-
In 1990 I was alerted to the possible danger of
position.' In March 1991, the Australian Nu-
radiation hazards in the Australian War Memo-
clear Safeguards Authority was in the process
rial (AWM) aircraft collection. Upon return to
of identifiing the concentrations of these mate-
Australia from the United States, David Hallam
rials (magnesium-thorium, ZTI andrare earth
brought a copy ofa document published by
alloys) around the country.
the U.S. Department of Defence Air Force

287
Radioactive materials have been used in igniter while that emitted by the luminous and phos-
plugs, counterweights, aircraft structural com- phorescent materials is pale green.
ponents near engine exhaust areas, gearboxes,
windshield drains, circuit breakers, switches, In section AN 05-10-1, the Pioneer Instrument
exit signs, flight-instrument dials and engine Company air speed indicators and one cabin air
fluid-level gauges, compasses, camera lenses, flow gauge are described by type numbers and
and aircrew wrist-watches. Nucleonic oil the dials'graduations and divisions are listed
gauges containing krypton-85 have been cited according to their type of material. Radium
as needing specific storage requirements.' paint was used on seven out of l0 ofthese in-
struments and some fluorescent lacquer was
Certain luminescent safety devices on aircraft also used. A suction gauge also made by Pio-
may contain up to l0 curies of tritium or up to neer is shown as having a "radium-painted dial
300 millicuries of promethium-145 as sealed and pointer." Technical order 05-20-27 in the
sources.' (These amounts seem higher than same publication deals with the Kollsman In-
would be expected; perhaps the readings should strument Division Type C-2 rate-of-climb indi-
be millicuries and microcuries respectively.) In- cator, which has the pointer hand, major
strument radiation surveys cannot detect tritium graduations, numerals (except 0.5) and the arc
and may not detect some promethium-l45 between the 6,000 ft graduations painted with
sources because of their low energy beta emis- luminous material. The remainder of the letter-
sions. Items exhibiting luminescence that do ing and numerals are treated with fluorescent
not give a reading on the survey meter should lacquer.
be suspected ofcontaining one ofthese radionu-
clides and managed accordingly. These materi- Thermocouple thermometers made by Lewis
als are dangerous in the long term. The half-life Engineering are covered in Technical order
ofradium for radioactive decay is 1,620 years 0540D-2.In this section the use of luminous
whereas the half-life for elimination from the material is only mentioned as an "as required"
body is 45 years.3 entry indicating that some of these instruments
are likely to contain radioactive material and
Literature Survey should be treated accordingly. The overhaul of
fuel contents gauges manufactured by the Liqui-
Quite often the technical publications issued by dometer Corporation is covered in section AN
the various authorities that use aircraft do not 05-65A-6, in which there is no mention of the
mention radiation hazards. They only mention dial-painting medium.
aircraft instruments in the context of their use
by aviators and their basic fitting to the aircraft A "three element nucleonic fuel quantity sys-
by technicians. tem" referred to in an engineering paper de-
scribes its radiation source as consisting of
The U.S. Army Air Corps Technical Ordersa coiled stainless steel or aluminium tubins
has several relevant paragraphs in different sec- containing krypton-S5, a radioactive gasf
tions describing the use of radioactive material.
Of the 18 General Electric tachometers tabled The Kollsman Instrument Handbook6 mentions
in section AN 05-5-4, three are listed as having the use of radium paint on pointers as a cause
luminous scale markings, while the others have of balancing problems if the paint is newly ap-
phosphorescent or fluorescent markings and plied and still wet. Page 16 of the same docu-
pointers. The luminous material is radium ment deals with the application of radium paint
paint, which emits light without any external to dials. The information is of historical interest
source of excitation. The fluorescent material only, and the methods of application and deni-
emits light when exposed to ultraviolet (UV) als ofdanger are not supported by the present
radiation. The phosphorescent material emits understanding of radiation hazards.
light when exposed to UV and in addition con-
tinues to emit light for a period of time after the Most old luminous paint contains radium-226
removal of the UV sowce. The light emitted by at much highgr concentrations than is now
the fluorescent material is pale yellow in colour usually used.' Another source of radioactive

288
material, which may be overlooked, is the fluid The ingxpensive monitors have some inherent
inside a compass, which is in contact with the faults.' The readings in mRA are not accurate
luminous paint on the dial. Although the fluid (the accuracy factor stated by the manufacturers
(normally an alcohol solution) may not dissolve is usually plus or minus 10% of full scale for
the radium oxides, it may carry loose particles the calibrating source, commonly cesium 137)
that are hazardous to health. and ifthe count rate is excessively high the
monitor may be prone to "runaway" or "paraly-
The loss of tritium from luminous aircraft sis." A much more expensive type of monitor
exit signs has been examined.' Although only has an ionization chamber. This type of instru-
about2o/o of the tritium was present in the ox- ment can provide a reliable measurement over
ide form, (with the remainder as tritium gas) a wide range of incident energies.
tritium oxide was measured because the biologi-
cal effect is greater than that of the elemental The AWM uses an inexpensive monitor9 to
tritium. ltbout99o/o of the tritium released do an initial check for radiation sources. Ifany
from a broken exit sigrr escapes in the frst readings above background are observed, a reli-
1.75 hours. able measurement is then sought from the local
Radiation Safety Section of the Departrnent of
According to Harley, "a radiation protection Health (government experts who offer an accu-
program must consist of two parts evalu- rate measuring service for $A60/hr the most
- of the
ation and control. Evaluation consists -
effective and prudent method of surveying for
measurements required to estimate possible the presence of radioactive materials).
radiation exposures and the necessary interpre-
tation of these measurements. Control consists The AWM .rr*.y10 was conducted using a
of taking agtions required to reduce radiation portable contamination monitor (Nuclear Enter-
exposure."d prises PCM5/1 with beta/gamma probe BPI/4
Geiger-Muller Counter) to detect radioactive
Surveying the Collection material. According to the manufacturer the
accuracy rate is plus or minus 25olo. Dose rate
Monitoring Equipment measurements were made using the low energy
Determining if an aircraft component is a radia- mylar window (1.7 mglcm'density) of a Vic-
tion hazard may be achieved by using an inex- toreen 450 Ionization Chamber Survey Meter.
pensive radiation monitor capable of measuring The accuracy stated for this instrument is plus
ionizing radiation by means of a Geiger-Muller or minus l0%o, exclusive of energy response.
(G-M) tube with a thin mica end window.
When a ray or particle of ionizing radiation The presence of radon gas was determined by
strikes the fully enclosed G-M tube it is sensed using a Nuclear Enterprises Radioactive Gas
electronically and displayed by a pointer on an Monitor RGMl/1. Accuracy is stated at plus or
analogue dial face. The instrument may be set minus 20%. The check for radon was thought
to respond to each ionizing event with a flash- necessary because ofthe accumulation of
ing light and an audible beep. Alpha and beta stored instruments loose in cardboard boxes on
radiation are usually measured in counts per a mobile shelving system. The shelves were left
minute (cpm) or in mSv,/h in SI units while closed for four days after which the activity
mR/h is commonly used to measure gamma ra- was read as the shelves were opened. Radon
diation. About 5 to 25 cpm at random intervals was detected but the amount was insignificant.
(depending on the monitoring location) can be
expected from naturally occurring background Although no mention was found in the U.S.
radiation. The instrument should have a range military literature of the use of radioactive mate-
switch to change the scale of the readings, rial in temperature and fuel gauges, the AWM
for example,0 to 0.5 mR/h (0 to 500 cpm), survey indicated the use of radium in all gauges
0 to 5 mR& (0 to 5,000 cpm) or 0 to 50 mR/h monitored.
(0 to 50,000 cpm).
The results of the survey follow, listing com-
plete aircraft first (Table I), then individual

289
instruments in storage (Table II). The instru- Messerschmitt Me BF 109G-6
ments are listed by their common names along This aircraft from the same period as the
with the radioactivity readings, which are taken Me 1638-1 would be expected to have the
at the surface of the instrument glass (or just same levels of radioactivity. This is not re-
above the surface ofpaint ifthere is no glass) flected in the recorded readings, as the aircraft
and also at a 100-mm distance from the surface. is in a dismantled state and has no instrument
Figure I shows the relationship between the panel fitted. The instruments are stored and are
two readings of a sample of American instru- listed amongst the German ones. The one read-
ments. A general background reading taken in ing obtained from two radio frequency selector
the cockpit area ofeach aircraft and in the stor- buttons was very high. This was a case where
age areas proved to be between 0. I mRA and radium oxide paint had been applied quite
0.2 mR/h in most cases, and is listed as BG in thickly in a groove around the perimeter of
the aircraft tables and as a figure of 0.15 mRlh each button with a central accumulation of
in the stored instrument tables. paint about 4 mm in diameter. These buttons
did not appear to have any lacquer or other
coatings on them.
4S

Mitsubishi A6M2Zero
400
This fully restored aircraft shows only two
instruments as having radioactivity readings.
3S

3@
Commonwealth CA-5 Wirraway
This aircraft was built as a training aircraft and
2g has about 80% of its complement of instru-
ments, none of which appear to be hazardous at
200 reasonable distances.

1S Link Flight Simulator Trainer

The Link simulator exhibited a single radiation
1@ reading from one instrument which, because of
its construction, did not allow a surface meas-
0s urement. The background reading in the seat
area was almost three times the norm.
0@
3=A 9q
699 ?+ Commonwealth CA-12 Boomerang
\R!
:< This early World War II Australian aircraft
showed fairly normal readings except for one
engine manifold pressure gauge.
Figure I The relationship between surface and
100-mm radiation activity readings of individual
English Electric Canberra B20
Am eric an aircraft fl ight ins trument s.
(built by Government Aircraft Factory)
This Australian built Canberra from the 1950s
Discussion of Survey Results was from a period when the hazards of radioac-
tive emissions from instruments were becom-
Messerschmitt Me 1638-1
ing unacceptable and the use ofnon-radioactive
This aircraft has an almost complete instrument
materials was becoming more prevalent. Only
panel and cockpit fittings and it appean to be
one item on the pilot's panel indicated any ra-
quite representative of the type of German air-
dioactivity (a low reading) and the navigator's
craft for the period. The background radioactiv-
area had only one switch on one side panel that
ity reading is a little on the high side and some
registered quite high on the monitor.
of the surface readings are 20 times the back-
ground readings.

290
Avro Lancaster MK I Stored Instruments
The Lancaster is a large aircraft with multiple The individual readings of stored instruments
instrument positions. Most were found to be taken at the surface and at 100-mm distances
fairly average in radioactive readings with a revealed some alarming statistics. One evident
few notable exceptions, mainly the engine per- trend is that all instruments with broken,
formance instruments. The 100-mm readings cracked or no glass are likely to be very hazard-
were found to be distorted by the mass of instru- ous to health. In addition, the British aircraft
ments side by side, and were considered not instruments seem to have more, and slightly
worth recording. higher, readings than aircraft liom other
countries.
North American B25 J Mitchell
This North American aircraft, although built Reducing Exposure to Radiation
later than the others, still showed a few instru-
ments with high readings. The most obvious All types of radiation share the property of en-
hazard in this aircraft was the multitude of ergy loss by absorption in passing through mat-
switches with heavy deposits of radioactive ter. The degree ofabsorption depends on the
paint, all very prominent and easy to touch. type of radiation and the absorbing material.
This absorption process results in ionization,
Supermarine Spitfire MK IIa which is the removal of an extra nuclear
The Spitfire readings are included, but only as a electron from an atom of the absorbing material.
guide. The Turn and Bank indicator, which has
no protective glass front, probably interfered The intensity of radiation decreases with the
with all the other readings from this aircraft's inverse square of the distance (d) on moving
relatively small instrument panel. Further read- away from the radiation source (Figure 2).
ings will be taken once the glass has been re- Thus the intensiff l0 m away is only lo/o of the
placed, but in the meantime the aircraft cockpit intensity at I m.* Therefore the further away
is out-of-bounds. one is from the source ofradiation, the better.

EXWUE€fO PADIATION
tlAY E @aC6O .tY- t. /ncrryv* 9lllEU/hG
2.
,.
rc e
hrm alffitZ'ef
AN n
\/ n
ilo\] )-\
(r )l /tl ()
d'7 dzZ t\ il dzt
EffilE,AE HH A(rffiE,@e iu fil
,x ' )(/4
F,Tffi/EE47E
, x/g l0i
\q [01 tl
tlll
\t \{ ll

\ttt /)
IJ z-)

Figure 2 Reduction of radiation exposure (d : distance; x : exposure rate).

291
Since radiation dose to a worker is directly pro- Disposal of Radioactive Materials
portional to exposure time, it is obvious that re-
ducing the exposure time by half (by working In the Australian Capital Tenitory where the
quickly) is not as effective a safety measure as AWM is situated, there is no legal disposal
increasing the distance from the radioactive method of alpha-emitting radiation material.
source (e.g., by using tongs or similar tools). This policy means that any instruments or
Notwithstanding, nothing should be done to switches with radium paint that are considered
an aircraft component by non-radiation-trained too hazardous to display must be stored by the
museum staff that would release or disnrrb the AWM. Other states and territories in Australia
radium-based paint. may have different regulations regarding this
type of material disposal. A national repository
Storage of Radioactive Materials for radioactive waste disposal may be built in
the future. Until then, these materials remain
Individual radioactive instruments should be our responsibility.
stored in two or more strong plastic bags and
then in boxes clearly labelled to warn th4t radio- Conclusion
active luminous compounds are present.- In the
case of instruments with missing or broken Museum collections containing aircraft most
glass faces, an aluminium or lead lined box likely have some radioactive sources. The most
may be advisable as an extra precaution. Seek prominent sources of ionizing radiation are ra-
expert advice from local radiation safety organi- dium-painted instruments and switches. Ap-
zations to clariff such situations. Some proximately 30% of the AWM collection of
switches and selector buttons in aircraft have aircraft instruments and other components indi-
had radioactive luminous paint applied. These cated radioactivity above normal background
items can be removed if they are in an area levels.
where physical damage to the paint is likely to
occur, or they can be sealed with a clear acrylic Exposure to radiation may be reduced by the
lacquer. follow ing three barriers :

Marking Radioactive Components o Increasing the shielding between the source

and the subject
The U.S. Air Force Museum Program Hand-
book states that red adhesive dot labels are to r Reducing the time of exposure of the source
be used to identify radioactive commodities in to the subject
aircraft or in display cases this being to alert
maintenance personnel ofthe - dangers present. o Increasing the distance between the source
A green dot is to be used to identify non-radio- and the subject
active instruments. This system ensures that all
aircraft instruments or suspect materials are The most economic barrier that can be raised
surveyed and appropriately identifred. The against overexposure is the third one.
AWM is not using this method of marking
instruments, as the aircraft in the collection are Specific advice on radiation safety is limited in
neither open for public inspection nor main- this paper both because of my lack of expertise
tained by aircraft technicians. Conservation in this area and the various legislations by coun-
maintenance of the collection does include cil, state, and country on the management of
cleaning inside aircraft, but this is only per- ionizing radiation. I can only stress that collec-
formed by staffaware of the hazards of disturb- tion managers, be they curators or conservators,
ing radioactive materials. The instruments in seek the assistance and advice oftheir local
storage are identified with standard radioactive radiation safety experts.
material warnins labels.

292
Appendix A. Gtossary Gamma emission is a secondary process follow-
ing certain alpha or beta decays. The nucleus
Radiation is a form of energy and may be ex- may be left in an unstable state as a result of the
plained as the emission and propagation of en- alpha and beta emissions and the excess energy
ergy through any material in the form of waves. is released in the form of gamma radiation.
Gamma radiation can be very penetrating and
Radi oactivity is the spontaneous disintegration absorbing material only reduces the intensity.
of atomic nuclei with emission of corpuscular Gamma radiation liom a radium source is re-
or electromagnetic radiation. All elements of duced to 50% of its incident value bv 25.4 mm
atomic number greater than 83 exhibit radioac- of lead.
tive decay.
Bremsstrahlureg is secondary radiation
Nuclear radiation emissions are, mosl com- arising from collision or deviation between fast-
monly, alpha (cr), beta (B) and gamma (y) moving electrons and atoms.
radiations.
Appendix B. Radioactive Materials
Gamma radiation is electromagnetic similar
- but
properties to visible light or ultraviolet, Radon: heaviest noble gas, radioactive, half-life
more penetrating, and may traverse several 3.82 days. Symbol for radium emanation or ra-
centimetres of lead. don is RaEm.

Beta and alpha radiations areparticulate Emanations: heavy inert gases resulting
-
extremely small particles moving at very high from decay ofnatural radioactive elements;
velocity. e.g., radium radon, thorium thoron,
actinium -> -> lived and
actinon. These are short
Alpha particles are emitted by many of the -> radioactive elements.
decay to other
heavy artificial and naturally radioactive ele-
ments. The particle is the same as the nucleus Tritium: radioactive isotope of hydrogen with
of the helium atom (2 protons and 2 neutrons). a half-life of 12.5 years. Very rare as naturally
The particle is heary and slow moving and occurring material. It is used as a radioactive
dissipates energy easily most are completely tracer or label and is very important in radiobi-
-
absorbed by a few centimetres of aiq or less ology. It has also been used as a substitute for
than 0.005 mm of aluminium, or a sheet of pa- radium oxides on instruments to provide lumi-
per. They are not considered a very serious nescence. It produces low energy beta radiation
health hazard extemal to the body, although that cannot penetrate beyond the glass lens of
internally they are most serious owing to their the instrument.
very short range and great ionizing power.
They are among the most radiotoxic nuclides. Promethium-145: rare earth, no known
The decay of the alpha particles in radium re- stable isotopes in nature. Half-life greater than
sults in the transformation to radon (gas). 20 years. Emits medium energy beta particles.

Beta particles are composed of electrons (posi- Krypton-85: rare gas, colourless, odourless,
trons and negatrons) emitted by both hear,y and forms one millionth of the atmosphere.
light radioactive elements. These particles pos-
sess kinetic energy because of their high emis- Radium-226: one of the alkaline-earth metals,
sion speed. The more energetic beta particles emits alpha, high-energy beta, gamma and
approach the speed of light. The range in air Bremsstrahlung.
may be more than a metre and they can pene-
trate several millimetres of aluminium or plas- Radioactive Luminous Compounds: These con-
tic, however the particles emitted from tritium sist ofa phosphor, such as zinc sulphide, inti-
or carbon-14 may be stopped by a few centime- mately mixed with a radioactive material, such
tres ofair. as radium-226, the radiation from which acti-
vates the phosphor to emit light. The alpha

293
particles emitted by a material, such as radium, Effective Half-Li.fe (Ts): always less than both
destroy the phosphor more quickly than the physical and biological half-lives. Time during
beta particles emitted by promethium-145 or which the activity in a particular biological
tritium. When the phosphor is destroyed the system is reduced to halfby both radioactive
luminous compound becomes ineffective, but decay and biological excretion. The physical
most of the radioactivity remains. and biological halflives of a radioactive iso-
tope can be combined to find the effective
Appendix C. Measurement of Radiation half-life in the body by:
l/Teffective : l/Tphysical + llTbiological
There are many different but related terms
used to provide measurement data on radiation. Dose: the quantity of ionizing radiation to
Some of the most commonly used terms are the which a sample has been exposed.
following:
Dose Equivalent: the biological effect conected
Electron volrs (eV): The measurement of the for certain factors measured as sieverts (Sv) in
energy ofany radiation (and hence its ability to SI units and rem (rem) in conventional units
penetrate biological tissue) is expressed in elec- (1.0 rem: 0.01 Sv).
tron volts. This is the energy gained through
the acceleration ofan electron by an electrical Dose Equivalent Rate: shown as sievert per
potential of one volt. This is somglimes ex- second (Sv,s-') or rem per second (rem s-')
pressed in SI units as 1.602 x 10-19 Joules; (1.0 rem s-':0.01 Sv s-').
because of the small size of this unit the ener-
gies are usually given in thousands or millions Expos4re'. measured in coulomb per kilogram
of electron volts (keV or MeV). (C kg ') in SI units. Roentgen (R) was the old
(and probably betJer known) unit of exposure
The usual range of energies for single particles (l R:2.38 x 10-- C/kg).
or rays of common radiations are the following:
Alpha particles 4 MeV to 8 MeV Absorbed Dose: meastued in Joule per kilo-
Beta particles 0 MeV to 4 MeV gram (J kg-') known as gray (Gy) in SI units.
Gamma rays 0 MeV to 2 MeV The rad was the former measurement of the
amount of energy deposited by radiation in the
Activity'. Radioactive isotope activity was for- human or animal body (l Gy: 100 rads). A
merly expressed in terms of the curie (Ci), the more realistic understanding of the gray may be
millicurie (pCi) or the microcurie (pCi) (after seen in the two following observations of the
the famous Polish-bom pioneer of radioactiv- rad, which has only one hundredth of the
ity). One curie is the activity associated with gray's value:
any radi-opctive isotope when it is undergoing
3.7 x l0''nuclear disintegrations per second, l. The normal background level ofradiation
historically equivalent to one gram of radium. (from radioactive rocks, cosmic rays.and
aOK;
The current SI unit is the becquerel (Bq), which internal deposits, such as potassium i.
equals one nuclear transformation per second. about 0.2 rad per year

Physical Half-Lfe (Tp): the time during which 2. A dose of about 700 rad delivered in a short
an aggregate of radioactive atoms decays to time to the whole body has a 95% probability
half its number. of causing death. The same dose delivered to
only a small part of the body is less serious.
Biological Half-Life (Ta): the time during
which the number of atoms of a stable nuclide Quality Factor: The concept of the absorbed
ingested into the body is reduced to one half dose has to be applied with caution as it makes
(this depends on the radioactive element and no allowance for the different types of radia-
the excretion rate). tion. Equal doses measured in grays of 5 MeV
alpha particles and220 keV X-rays would give

294
very different biological effects. An approxi- mis en rtserve individuellement, ont 6td ddcou-
mate weighting factor called the quality factor vertes. Certains instruments atteignent ainsi un
has been ascribed to each type ofradiation taux d'exposition supdrieur d 300 mMh, tandis
without specifuing the biological effect. qu'un interntpteur, plact d I'intdrieur d'un
adronefallemand, indique, d lui seul, un tata de
Rern: the acronym for Roentgen equivalent 60 mRh. Le taux d'exposition a ttd mesurt en
man. It is the amount of ionizing radiation that sutface et d une distance de 100 mitres de hau-
teur. Les taux sont indiques selon le Qpe d'instru-
produces the same damage to a human as one
ment et d'atronef.
roentgen ofabout 200 keV X-rays.
A ces donntes s'ajoutent une description des dif-
The still commonly used value of the millirem
(m rem) appea$ to be interchangeable with
flrents genres de matidres radioactives qui sont
utilisdes dans les adronefs, un lexique des termes
milliRoentgen (mR).The dose equivalent in en usage dans le domaine de l'|valuation de la ra-
rem is the product of the absorbed dose in rad dioactiviti, ainsi qu'une difinition des methodes
and the quality factor. Equal doses ofdifferent sfires de manutention, de mise m rdserve et
types of radiation measured in rem produce d'ttiquetage des composants radioactifs. On
approximately the same biological effect. suggire enfin certains moyens qui permettront
de protdger tant le personnel que le public des
Reliable measuring instruments indicate the musies d'une exposition aux radiations.
exposure in milliRoentgen per hour (mR/h) or
in counts per minute (cpm). Radiation types References
that have a quality factor of one may be ap-
proximately equated for biological material s : l. United States Air Force Museum Program
Exposure measured in Roentgen: absorbed Handbook (Ohio: USAFM, Wright-Patterson
dose measured in rad: dose equivalent AFB, Ohio, 1988).
measured in rem.
2. Baker, Dennis, personal communication, His-
Radioresistanl: able to withstand considerable torical Projects Group, Hawker de Havilland,
radioactive doses without injury. 1991.

Ra.diosensitive: this normally refers to organs 3. Code of Practicefor the Safe Use of Radio-
quickly injured or changed by irradiation. The active Luminous Compounds (Canberra: Aus-
gonads, the blood-forming organs and the cor- tralian Departrnent of Health, 197l).
neas are biologically most radiosensitive in
humans. 4. Chamberlain, K.8., ed., Aus tralian National
Universily lonizing Radiation Safety (Canberra:
R6sum6 Australian National University, Canberra,
1989).
Les risques d'ryosition aux radiations
provenant d'aironefs eryosds dans les musies 5. Walker, W.P. and F.J. Steinbrenner, "Design
and Testing of Aircraft Nucleonic Fuel Quan-
Une dtude, portant sur les isques tttentuels d'ex- tity System," SocieQ of American Engineers pa-
position aux radiations provenant des aeronefs per 690673,1969.
exposds dans les musbes et de leurs composants
a 6td mente d l'Australian War Memorial. Pour 6. Handbook of Airplane Instruments (New
deceler la presence de matidres radioactives, York Kollsman Instrument Division of Souare
on s'est alors servi d'un dttecteur portatifde ra-
D Co., 1940).
diation, tandis qu'un appareil de mesure d cham-
bre d'ionisation a dtt utilisd pour |valuer la
7. Niemeyer, R.G., "Tritium Loss from Tritium
dose absorb,le.
Self-luminous Aircraft Exit Signs," Isotopes
De nombreuses sources radioactives, tant d and Radiation Technolog1,,, vol.7 , no. 3, 1970.
I'intdieur des adronefs que parmi les composants

295
8. Harley, J .H, Dangerous Properties of Indus- 10. Evans, Y ., Report to Australian ll'ar Mu-
triql Materials, 3rd edn. (New York: Reinhold, seum on Radiation Survey of Aircrafi Collec-
1968). tion,AWMFileg0/0534(14), ll Dec. 1990.

9. Operation Manualfor the Radiation Alert

Monitor 4,Extech Equipment Pty Ltd., Victo-
ria, Australia.

296
 Table I
Radiation Survey ofAircraft (ca. 19l7 to 1960s)

InstrumenULocation Surface 100 mm I Instrumenulocation Surface 100 mm

Messerschmitt Me 1 638-1 (1945) Avro 504K (1 918) no Instruments

frtted
Arr Speed Indrcator 06 BG
Altrmeter 40 o.4
Chronometer 4.O BG Airco DHs (1918) BG
Oxygen Supply 1.0 BG
Arr Temperature 1.0 BG Bell 47G-38 Sioux (1 960s) displayed out
VHF Radro Buttons 3.0 v,4J of reach

Cockpit area background reading 0.2 Pfalz D Xll (1918) no reaorngs

Messerschmitt Me/BF 109G-6 (1945) Royal Aircraft Factory SEsA (1917) no readrngs

VHF Radio Buttons (Stbd) 60.0 1.0 Albatros D Va (1917) no readrngs

Cockpit area background reading 0.1 Messerschmitt Me 262A-1 Jabo (1944) no Instru-
ments
Mitsubishi A6M2 Zero (1942) frtted

Vanometer 2.0 BG Gloster Meteor F8 (early 1950s) BG

Arr Temperature 0.34 BG
Bell UH-l8 lroquois (1960s) BG
Commonwealth CA-s Wirraway (1942)
Avro kncaster Mk l(1942)
Rate of Clmb 0.8 EU
Turn and Bank 0.3 BG Prlot's Panel
Air Speed Indtcator 0.3 BG (Stbd side readrng at 300 mm 1 0)
Manifold Pressure 0.34 BG (Port side readrng at 600 mm 0.6)
Orl Temperature 0.9 DU Inm
Fuel Oil Temperature 0.6 BG P4 Compass 1.1
Arr Temperature 1.1 BG Undercarnage Indtcator
Flaps Posttion I.J
Cockprt area background readtng 0.1 Beam Approach 120
Altimeter 50
Link Flight Simulator Trainer (1950s) Arr Speed Indtcator 6.5
Artrficral Honzon 83
(Readrngs 20 mm from surface 2.5) AFN Drrection u.5
(Readrngs at seat dlstance 0.6) Turn and Slip 4.5
Rate of Chmb 17.0 4.0
Commonrvealth CA-12 Boomerang (1942) Engrne Boost: 1 10.0
2 10.0
Manifold Pressure 20.0 1.0 3.0
Air Temperature 3.5 ou 4 3.0
OilTemperature 2.0 ou Engrne RPM. 1 20.0 2.5
Voltage (at 20 mm) 25 EU z 20.0 25
Others (untdentrfred) 1.0 2 20.0 25
4 20.0 1.C
English Electric Canberra 820 (1950s) Flaps Posrtion 8.0
Air Temperature 6.0
Pilot's panel (Stbd) DF Indtcator 1.0
Fuel Transfer Swttch 10 bu DR Compass Repeater 0.8
Oxygen Regulator t.o
Navigator's area Oxygen Contents 1.5
Stbd of seat, lower of Brakes Pressure 1.8
two rear components 23.0 t.z

Note: All the readtngs are gtven In milliroentgens per hour (mFi/h) The surtace and 100-mm readtngs used In the survey are gtven
In mR/h as they are more manageable In that format. To convert the readings to counts per minute multiply by 1,000. lf ihe reaorng
rs undetectable from the background radration level it is denoted by BG (background).

297
 Table I (continued)
Radiation Survey ofAircraft (ca. I9l7 to 1960s)

InstrumenVLocation Surface 100 mm InstrumenULocalion Surface 100 mm

Engrneer's Panel Brakes Pressure 0.45 BG

(Panel at 300 mm 1.6) Hydrau|tc Pressure 0,4 BG
Radiator Temperature 1 8.0 BG Orygen Flow Indrcator 4.0 BG
12.0 BG
10.0 BG Centre Swrtch Panel
; 140 BG Lrghts, Fuel Transfer,
Fuel Contents: No.3 Stbd 14 BG Pnme Pumps, etc. 4.0 1.0
No.2 Stbd 30 BG Armament Panel Swttches 2.2 1.0
No.2 Port 12.0 BG Electncs Panel Swrtches
No.3 Port 10.0 BG Navrgator Lights, Master Right,
No.1 Stbd 1.2 BG Gear Warnrng Horn, Prlot,
No. 1 Port 1.4 BG Flatra^, | 6fi at. 4.5 1.0
Orl Pressure: Rear 8.0 BG
Front 4.5 BG Emergency Compass BG BG
Orl Temperature (x 4) 2.5 BG
Supermarine Spitfire Mk lla (1941)
Navrgator's Table/Panel
Altimeter 10 BG (Background radiation level (BG) 0.15)
Arr Speed Indrcator 7.0 'I 8 (300 mm from panel 0.7)

North American 825 J Mitchell (1945) Brakes Pressure/Supply 32 BG

Oxygen Regulator/Gauges 3.0 BG
Background In cockprt 0.02 Navrgatron Lrght Swrtch 2.0
Arr Speed Indtcator 10.0 2.0
Oxygen Flow Indrcator 4.0 EU Altimeter 25.0 3.0
Air Speed Indicator BG BG Compass Drrectron Indtcator 10.0
DR Compass Repeater 1.0 BG Turn and Bank (No glass) 50.0
Arttftcial Horizon 2.6 BG Rate of Clrmb 15 0
Rate of Chmb BG BG Artrfrcral Horizon 140
Turn and Bank BG BG Engrne RPM 100
Altrmeter BG BG Fuel Pressure 100
Fuel Suctron 200
No 2 Engrne Orl Pressure 100
Manfold Pressure 0.1 BG Radrator Temperature 8.0
Speed In RPM 0.1 BG Arr Temperature 7.0
Orl Pressure 0.1 bu Fuel Contents 7.0
Fuel Pressure 0.1 EU Gun Srght Drmmer Swttch 4.5
Undercarnage Indrcator 0.2 bu Generator Supply and Swrtch 2.0
OrlTemperature 0.1 BG
Cyhnder Temperature 0.1 BG
Carburetoa Arr Temperature 0.15 BG All these readings may be suspect due to the influence
Suctron 0.4 BG of the unglazed Turn and Bank lndtcator; further
Air Temperature 0.1 BG readrngs should be made when the otfendrng
Instrument rs reglazed.
Fuel Contents: Front o2 DU
Rear 035 BG
Auxhary 50.0 10

Note:All thereadtngsaregtvenrnmilliroentgensperhour(mF/h) Thesurfaceandl00-mmreadtngsusedtnthesurveyaregiven

In mFl/h as they are more manageable in that format. To convert the readings to counts per minute mulflply by 1,OOO. lf the reaorng
rs undetectable from the background radiatlon level tt ts denoted by BG (background).

298
 Table II
Stored Instruments
(Background (BG) reading 0.15 mR/hD

Readings in mR/hr
Accn. No, lnstruments Country Condition .|00
Surface mm
n6015 lArr Pressure (0{ OO kg/cm2) I Japan OK 10.15 0.15
/6350.001 lAir Speed Indicator lcermany OK lnzn
l- '- 020
/11504 lAttrmeter OK 0.15
6 001
/1 601
lentarn 10.80
| lAltrmeter Japan OK 4.50 0.60
I

ln 6016.002 lAltrmeter Japan OK 3.00 0.40

lAWM34872 lAltrmeter Germany OK 0.50 0.15
lAWM34822 Altimeter (0-5 km) Germany OK 0.15
i0.80
Artrfrcral Honzon Japan OK 0.15 u. tc
] |
n 1009 Arr Speed Indrcator Amenca OK 1.40 u. tc
I |
t11017 ]Arr Speed Indicator OK 0.15
lBntarn 12.50
t12177 lAir Speed Indrcator I Japan Broken glass 200 00 '11 00
/1 6002 002
I lArr Speed Indicator leritain OK 4.50 150
/1 601 6.002
| lArr Speed Indrcator Japan OK 4.00 1.10
|
/6350.002 Arr Speed Indtcator Germany OK 0.20
10.70
1n152 Arr Speed Indtcator Britarn OK | 11.00 1.00
ln31s lArr Speed Indrcator Germany Broken glass lso.oo 1.00
t8144 lArr Speed Indrcator Bntarn Broken glass 050
AWM34863
lz.so
lnir Speed Indrcator Germany OK
l0 70 015
)t2o194 lAutoprlot Altitude America OK 0.15 0.15
|
/8096 lchronometer Gemany OK 3.00 0.50
l/8482 Chronometer Japan OK 1.60 0.15
1t11012 Compass America OK 0.40 u tc
l/8o74 Compass Japan OK 2.00 050
/8464 I Compass Germany OK 0.60 015
)t8424 I
Compass E2A Bntain OK 2.00 n tE
l/8376 lCompass P8 Bntarn OK 6.50 1.00
In6413 Compass Type 96 Japan OK 0.50 0.30
lnu6 Drift Indrcator Germany No glass 100.00 8.00
/16937 Engrne Boost Pressure Japan OK 4.20 100
n321 Engine Boost Pressure Germany OK 0.70 0.15
/8139 Engrne Boost Pressure Britatn OK 80.00 200
AwM34884(x2) Engine Boost Pressure Germany OK 1 1.00 040
t10187 Engine Manilold Pressure Amenca OK 1.70 u tc
/1 101 0 Engrne Manifold Pressure Amenca OK 1.10 0.15
AWM34873 Engine Orl Temperature Germany OK 1.50 0.15
/16047 001 Engine Speed (RPM) Japan Broken, no glass 400.00 20.00
/12179 Exhaust Temperature Japan OK 7.O0 1.60
oL 18 Exhaust Temperature Japan Cracked glass 0.15 0.15
18134 Flap Posrtron lndtcator Britarn OK 340 080
/5460 001 Fuel Contents Bntarn OK 7.00 150
/5460.002 Fuel Contents Britarn OK 7.00 1.50
4WM34877 002 Fuel Contents (0-3800 ltr) Germany OK UJU u.tc
AWM34877.003 Fuel Contents (0-3800 ltr) Germany OK 050 0.15
AWM34877 001 Fuel Conrents (0400 ttr) Germany OK 0.50 0.15
/8132 FuelFlow(No 1&2Eng) Bntarn OK 6.00 2.OO
t16234.O04 Fuel Pressure (0-25) Amenca OK 4.50 1.00
116234.005 Fuel Pressure (0-25) Amenca OK 370 0.15
/AWM34851 Orl Pressure (0-10 kg/cm') Germany OK 080 0.15
n245 Oxygen Flow Germany OK 1.30 015
AWM34844 Oxygen Flow Germany OK 2.60 0.15
/12896 Oxygen Pressure Germany No case, no glass 18 00 2.50
/8138 Oxygen Regulator Britain OK 2.90 0.40
120198 Hydraulrc Pressure (O-22kg/cm2) Germany OK 1.00 015
AWM34842(x2)
-
Hydraulic Pressure (0-3,0-1 5 kg/c^mz) Germany OK 0.70 0.15
n238 Hydraulic Pressure (0-160 kg/cm1 Germany OK 0.40 0.15
nus QNH lndicator (Air Pressure rn Mrlftbars) Germany No glass 340.00 30.00
nu7 QNH Indrcator (Air Pressure In Mtlfibars) Germany No glass 200.00 20 00

299
 Table II (continued)
Stored Instruments
Readinqs in mF/hr
Accn. No, lnstrument Country Condition Surface 100 mm
n325 Rate of Climb Germany OK 050 u.tc
/16047.OO1 Rate of Clrmb Japan OK 1.00 0.15
t10184 Engrne RPM Bntarn OK 0.80 0.15
n219 Engrne RPM Bntarn OK 7.00 100
naEl Engine RPM Germany OK 0.40 0.15
t8143 Engrne RPM Britain No case, no glass 7.00 1.00
AWM34841 Engrne RPM Germany OK 080 0.15
AWM34878 Engrne FIPM Germany OK UJU u.lc
/1 1 008 Suctlon Amenca OK 070 015
110194 Suctron (0-10" Hg) Bntarn OK 400.00 600
t8072 Temperature (0-150'C) Japan OK 0.15 015
t8429 rempeftlrure (u-iJ5u-u) Japan OK 4.00 100
AWM34892 Thermal Self-trip Sw[ch Germany No glass 55.00 100
/8076 Turn & Bank Bntarn Broken glass 4.00 100
AWM34846 Turn & Bank Germany OK 3.50 0.15
AWM34889 Turn & Bank Germany OK 3.50 015
oL 17 Turn & Bank Japan Broken, no glass 85.00 14 00
4WM34881 Vaflometer Germany OK 0.50 015

300
Conservation of One of Alexander Calder's
Largest Mobiles

Albert Marshall and Shelley Sturman

D ep artm ent of Cons ertt ation

National Gallery of Art
Washington, D.C.
U.S.A.

Abstract The paper also discusses the conservation meas-

ures needed to restore proper movement to the
Alexander Calder began making his moving sculp- mobile. These measures included the tungsten-
ture, the mobiles, in the 1930s. As long as these inert-gas welding of losses to the aluminum and a
sculptures were small and light, the grinding of surface hardening with a spray of titanium diox-
one metal surface against another was minimal, ide to the contact surfaces ofthe sculpture.
but as size andweight increased, so did the issue
of wear. The 23 m (76fu, 418 kg (920Ib) mobile The Development and Construction
Untitled (1976), commissionedfor the grand of Calder's Mobile
atrium of the East Building of the National Gal-
lery of Art in Washington, D.C., was enlarged Alexander Calder's grand mobile has become
andfabicatedfrom Calder's 30.5 cm (l ft)
the unofficial mascot of the East Buildine of
maquette.
the National Gallery of Art. Upon enteririg the
museum, one is immediately engaged by the
When the mobile was loweredfor cleaning and in-
spection after I0 years ofexhibition, an apprecia-
expansive space frame and the slowly rotating,
ble loss of metal was obserted at certain points oJ' colorful mobile. In the late 1960s when archi-
contact on the sculpture. These losses are repre- tect I.M. Pei was making drawings for the East
sentative of a common problem in our industrial Building, plans for commissioning artworks
society. Over the years, various methods oflubri- to complement the architecture were set in mo-
cation have been developed to slow the inevitable tion. In June of 1973, Alexander Calder was
progression of wear. In a sculpture that moves, commissioned to design, fabricate and install a
the effect of wear is not always taken into ac- large mobile sculpture for the central, interior
count, and the object can slowly grind itself court of the National Gallery's East Building.
away. If a lubication system is not built into This sculpture became the ultimate example of
an art object, it cannot be accomplished easily Calder's art, and the study of its manufacture
within the ethical constraints of the conservation and its conservation provided an outstanding
profession. opporhrnity to understand and work with late-
20th-century art materials (Figure 1).
This paperpresents the problems encountered in
making such a large sculpture that hangsfrom a
Calder, who referred to the piece only as "my
single point and moves in air currents without a
object," produced a maquette of the mobile
motor. The construction eventually involved use
ofadvanced materials, such as aircraft specifica- shortly after his initial meetings with gallery
tion aluminum and plasma-spray application of officials. Engineers at the Biemont kon Works
molybdmum to the probable points of wear. in Tours, France, who had helped Calder create

301
 weight originally projected by Biemont. In its
frnished state, the mobile consisted of 13 panels
and 72 arms, weighing 41 8 kg (920 lb) and
spanning approximately 23 m (76 ft). It was
constructed of aluminum-skinned honeycomb
for the flat panels (planes, blades and wings),
with filled and shaped epoxy edges. These were
hung on aircraft specification 2024 aluminum
tubes. In anticipation of heavy wear at the con-
tact points, five of the loops bearing the most
weight were made of steel, 12 intermediate
weight-bearing loops were made of surface
treated aluminum, and the remaining loops
Figure 1 Alexander Calder: Untitled,, 1976. were made of untreated aluminum. Matisse
National Gallery of Art, Washington, D.C., Gift of used a plasma-spray application of molybde-
the Collector's Committee, (1977.76.1). NGA
num to surface harden the intermediate weight-
photograph by Philip Charles.
bearing aluminum loops.

many of his earlier outdoor stabiles and mo- Figure 2 shows a loop with plasma-sprayed
bile/stabile combinations, determined the pro- molybdenum hardfacing intact. Plasma spray is
jected weight of the mobile to be more than an industrial technique that applies molten met-
three tonnes. This estimate assumed that the al to the workpiece by passing powdered mate-
mobile would be 32 times larger than the rial through an inert gas plasma at temperatures
maquette and that it would be constructed from as high as 15,000"C. According to Herbert
traditional steel and solid metal plates. Struc- Herman, plasma spraying "can melt and apply
tural elements necessary to support the mo- a variety of materials, including refractory ce-
bile's weight and the motor required to rotate it ramics, at a high rate. The technique carries
within the large open area would affect the de- much less risk of degrading the coating and sub-
sign of the dramatic space frame. The massive strate than many other high-temperature proc-
sculpture being developed at the Biemont foun- esses do, because the gas in the plasma flame
dry was far removed from the artist's delicate,
almost weightless model that responded to the
slightest puff of air.

At this point Paul Matisse, an artist and engi-

neer, was consulted. Matisse, who is the grand-
son of the artist Henri Matisse and who knew
Calder from childhood, proposed a novel idea.
His suggestion to construct the mobile from alu-
minum-skinned honeycomb, a high-strength,
lightweight, aircraft material, solved the weight
problem, retained the qualities of Calder's origi-
nal, and received the artist's blessing. In May
of 197 5 Calder wrote to Matisse: "You go
ahead and build the thing I will come to
see it when you've finished- it."' By November
1976, Matisse had the huge mobile ready for
inspection. Eight days after he had seen the
mobile in the large warehouse that Matisse
had rented for its construction, Calder died.
Figure 2 Exposed area on aluminum loop showing
With Matisse planning and supervising every mo ly bdenum pl asma-sprayed hardfacing. N GA
step, the mobile was completed at one-sixth the photograph by Philip Charles.

302
is chemically inert and the target can be kept drops from the skylight frame modified the
fairly cool."' paint on the top surfaces of the panels. Some of
the arms had a white precipitate on their under-
The loops were attached to the arms with epoxy sides, possibly from the dripping condensate.
construction adhesive (Tracor, Trabond 2l0l). Dark finger marks could be seen on the bottom
The arms had been carefully bent by Matisse portions of the red blades, which are within
to show the exact curvafure under load as the reach of tall museum visitors.
arms of the maquette produced by Calder. Once
all of the arrns were assembled and the panels The mobile was delivered with a manufacturing
made, the sculpture was painted in Calder's defect. This went unnoticed because the mobile
trademark colors, matt surface black, blue and could not be completely assembled until it was
red, in a silicone modified alkyd binder devel- permanently installed in the East Building.
oped by Keeler and Long of Connecticut. Therefore a slight twist in the orientation of
the wing arm assembly loop of cross arm #1
Today, the mobile rotates slowly around caused wing assembly lAto hang almost per-
the central court, moved only by air passing pendicular to cross arm #l instead ofnearly
through vents high up in the walls. It was Alex- parallel (see Figure 3 for numbering of mobile
ander Calder's last major work of art, though panels). In the interest of time, Matisse, to-
he did not live to see it installed in the National gether with gallery engineering staff, effected
Gallery of ArL a temporary solution to the misalignment in
April 1978: a steel pin was placed in a newly
The Condition of the Mobile in 1988 drilled hole in the loop of cross arm #l so that
wing arm #4 would hang in its proper orienta-
In 1988, the mobile was removed from exhibi- tion. Sometime during August 1988, this pin
tion for the first time since it was originally in- dislodged, causing the misalignment of the arm
stalled in the newly completed East Building to recur and two of the black panels to cross
more than l0 years earlier. This gave the object over each other.
conservation department, working in concert
with Paul Matisse, an opportunity to examine These two conditions were already known and
the entire sculpture and treat existing condi- their solutions were straightforward. The third
tions as well as one unforseen condition. problem was more of a surprise. Upon disas-
sembling the work we found that in five of the
Three components of the sculpture's condition loops the aluminum had worn, creating small
were addressed in the treatment: cleaning the depressions of up to 3 mm (1/8 in.). As con-
dirt, grime and finger marks; correcting a manu- ceived, the mobile not only rotates on its main
facturing defec! and attending to the loss of axis, but the individual arms also pivot as they
material caused by wear at the
contact points ofthe loops. The
lrst two components were an-
ticipated before examining the
disassembled mobile; the last
component was discovered only
after the mobile's lowerins and
disassembly.

All ofthe upper surfaces of the

horizontal panels had a thick
accumulation of dirt and dust.
This had collected from the
construction that occurs during
the special exhibitions and the
general dust in the building's
atmosphere. Condensation Figure j Schematic drawing of the mobile w,ith pans labeled.

303
hang from the loops. (Figure 4 shows a loop (42 n.) from the end that holds the wing
with a depression from wear.) The depressions section. Next, a 30.5 cm (12 in.) joint bar of
restricted movement of the loops, because the 6061-T6 aluminum tube was machined by
two loops keyed into one another. Continued Paul Matisse to fit precisely into the cross affn
wear would aggravate this arm locking problem allowing space for adhesive. The arm was
and eventually cause more problems for the temporarily fitted together and hung by the
sculpture and the public walking beneath it. center loop. A loop representing wing arm #4
was placed on the short end loop and weights
equal to the normal loads of 90.5 kg (199 lb)
and 1 16 kg (255 lb) were applied to each end of
the arm. The cut end of the cross affn was then
rotated about its axis until the proxy loop for
the wing arrn came to its proper orientation.
The rotation turned out to be about five de-
grees. This position was marked, the arm was
disassembled and then reassembled with the
same epoxy construction adhesive as w€rs
originally used.

Once we discovered the losses of aluminum

in the loops, it was Matisse's wish that they be
filled with aluminum and hardfaced in a similar
manner to his original surface treatment. As
mentioned above, Matisse had originally ap-
plied molybdenum by plasma spray. While
plasma spray was not commercially available at
the time, we found that thermal spray, a similar
technique, was used in the area. (Discussions
Figure 4 Depression, approximately 3 mm, in
with the welding firm led us to their supplier,
aluminum loop, caused by wear. NGA photograph by
Philip Charles. Eutectic Corporation of Flushing, New York.
Their representatives and engineers recom-
mended Eutectic MetaCeram 25040, a titanium
Conservation Treatment of the Mobile dioxide [typical hardness: Rb 57] with Eutectic
MetaCeram 2103 I for the bond coat.)
The hrst part of the treatment involved an
elaborate "choreography" of lowering and disas- Thermal spray has been used industrially
sembling the sculpture. Matisse had a very de- since the 1940s. It involves spraying a material
tailed script for installing and de-installing the through an oxy-acetylene flame onto the work-
piece. As the sculpture is lowered the pieces piece at temperatures lower than those used for
land on different floor levels, and when a part plasma spray.
is removed it must be replaced with an equal
weight to keep the rest of the mobile level. The loops were restored by frrst removing
Once all was disassembled, a cursory washing the paint from nine pairs of loops. The original
was undertaken with Orr,us (an anionic deter- molybdenum on the loops of wing arm joint
gent made by Proctor and Gamble) and water #1 (the one bearing the least amount of weight)
to remove most of the dirt. All of the pieces was found to be intact leaving 16 surfaces that
were dried and taken to a storage area. (A draw- required a new hardfacing. First, the five loops
ing of the mobile with parts labeled and repair with aluminum losses were filled by aluminum
to cross arm #l indicated is seen in Figure 3.) tungsten-inert-gas welding. Then these five
loops and the remaining 11 loops were grit
The manufacturing defect mentioned above blasted with steel pellets to further clean and
was rectified by cutting cross arrn #l 105 cm texture their surfaces. All of these areas were
thermal sprayed with a bond coat of Eutectic

304
 I'artiste a rdalisd des cBuvres de plus en plus gran-
des et lourdes, accroissant d'autant le probldme
de leur usure. Le mobile de 23 m, pesant plus de
400 kg, Sans titre (1976), qui a dte command,!
pour le grand atrium de la National Gallery of
Art de Washington, D.C. a 6te agrandi etfobri-
qut d partir d'une maquette de quelque 30 cm
rtaliste par Calder.

Lorsque, apris dix ans, il afallu descendre ce

mobile pour le nettoyer et I'inspecter, on a cons-
tatt une usure appreciable du metal d certains
points de contact. Et cette usure tdmoigne d'un
probldme que I'on rencontrefrequemment dans
notre soci2tt industrielle. Au fil des ans. diverses
mdthodes de lubrification ont ttt mises au point
pour ralentir I'inevilable progression de I'usure.
Or, dans le cas plus pafticulier d'une sculpture
en 2quilibre instable, I'usure n'est pas toujours
prise en compte, et il se peut donc que ses elfets
graduels finissent par compromettre I'existence
Figure 5 During treatmeilt, before inpainting. NGA
photograph by Philip Charles. m€me de I'rpuvre. Si I'artiste n'a pas integrd un
systime de lubrification d I'rpuvre au moment de
sa crtation, les spdcialistes de la restauration
MetaCeram 2103 I, a nickel/aluminum/molyb- pourront dil/icilement decider, d cause de consi-
denum alloy in a 90:5:5 ratio. (Figure 5 shows dtrations liees aux rdgles de la profession, de
a loop with thermal-sprayed hardfacing before I'ajouter par la suite.
inpainting.) Then, a topcoat of Eutectic
MetaCerarn 25040, titanium dioxide, was Dans le cadre de la prtsente communication,
applied by thermal spray. The surface of the nous dtcrirons les problimes entourant lafabri-
thermal spray coat and the edge ridges were cation d'une sculpture aussi grande qui, accro-
smoothed. All of the areas that had lost oaint chte en un seul point, doit se mouvoir sous
were repainted with the same formula paint as I'action des courants d'air, sans moteur. Sa cons-
the original, a poly-silicone. The sculpture was truction a tventuellement exigd I'utilisation de
given a thorough cleaning with dilute Vulpex matdriaux de pointe, tel l'aluminium employe en
(potassium methyl cyclohexyl oleate) and water. aeronautique et du molybddne pulvbris1, qui a tte
appliqut sur les points d'usure probables.
After months of collaborative effort among con-
servators, engineers, machinists, and fabricator, Les mesures de conservation ntcessaires oour
assurer le bon mouvement du mohile seront Agale-
the restored mobile was reinstalled followins
ment ahorddes. Au nombre de ces mesuret,-/iEr-
the original script prepared by Matisse. The
rent la reparation de I'aluminium par le proctdd
awaited rehanging coincided with the celebra-
TIG (Tungsten Inen Gas) et le durcissement de la
tion of the 1Oth anniversary of the opening of surface par la pulverisation de dioxyde de titane
the East Building. sur les points de contact.

R6sum6 References
Les restauration d'un des plus grands mobiles L Calder, Alexander, Letter to Paul Matisse,
d'Alexander Calder May 3, 1975, Archives of the National Gallery
of Art, Washington, D.C.
Alexander Calder a commenct d faire ses sculp-
tures en tquilibre instable, ses mobiles, dans les
anndes 30. Lorsque ces sculptures etuient petites 2. Herman, Herbert, "Plasma-sprayed
et ltgdres, le frottement des surfaces metalliques Coatings," Scientifi c American,September
entre elles 6tait minimal mais, avec le temps, 1988, p. I 12.

Jtr)
Conservation of Weathering Steel Sculpture

John Scott

Sculpture Conservator
New York Conservation Center
New York, N.Y.
U.S.A.

Abstract understanding and practical abilitv that any

conservator active in this area should possess,
After a long histow in decorative arts and archi- and that the preceding sections sketch.
tecture, steel has perhaps become the predomi-
nant material of2)th-century sculpture. Its Introduction
design and production now engage sculptors
much as carving, modeling and the foundry. Ac- The care of contemporary art often includes
c or dingly, today's conservators of sculpture are the care of steel sculpture. Special problems are
interested in the history, technologt, deteiora-
encountered in the conservation of the many
tion, presertation and restoration ofsteel.
sculpfures in "weathering steel" (e.g., Cor-'ien,
discussed below), a material that develops a re-
This paper focuses on high-strength, low-alloy
markable appearance and corrosion resistance
weathering steel commonly known as "corten,"
after the United States Steel Corporation product
through outdoor weathering. Thus, we conser-
"USS COR-TEN," which appeared in contempo-
vators and our colleagues in the field of contem-
rary sculpture in the mid-1960s and continues in porary art need historical information on the
use today. Some weathering steel sculptures have development and design requirements of weath-
deteriorated severely, sometimes within a short ering steel, its marketing history and the circum-
period after installation. Others weather well for stances surrounding the adoption of weathering
many years despite being given little care. steel as a sculptural medium. Good conserva-
tion planning requires, as well, some basic un-
A bief histoical discussion of the development derstanding of weathering steel's distinctive
and general adoption ofweathering steel is con- corrosion process. This introductory exposition
cluded with an account of the material's adoption addresses these topics, then presents preserva-
as a sculptural medium. That introductory ac- tion and maintenance methods and ethical
count situates the matter of weathering steel's considerations.
appearance and special corosion resistance in a
histoical context of metallurgy, economics and
art. A chronological review ofscientific theories
Development and Design
about the protective patina is presented, includ- Requirements
ing the most current accounts. A discwsion of
consertation methods focuses on ad hoc practices Iron-copper alloys have interested Westem
now in widespread use, which are evaluated with metallurgists since at least the early l Tth cen-
respect to practical efectiveness as well as aes- fury. Throughout the 19th century such alloys
thetic and histoic appropriateness. A concluding received increasing attention, and the earliest
discus si on of ethics presumes the considerable useful work on the mechanical properties of

301
wrought "copper steel" appeared around the last outdoors as intended, weathering steel must
decade of the century. ("Steel" is a general term drain well and dry quickly after wetting. Size is
for iron alloys containing less than2Yo carbon. important small structures can be less suscepti-
Late l9th and early 20th century copper steel ble to destructive corrosion than larger struc-
contained from loh to 2o/o copper in addition to tures if they are more effectively dried by sun
carbon and other alloying elements.) and wind.

Corrosion characteristics are always of interest When weathering steel drains poorly or re-
in materials science, and early in the 20th cen- mains wet, it develops no protective rust, and
tury, systematic data were collected on the cor- under such conditions rapid destructive corro-
rosion resistance of copper steel. While some sion proceeds unless wetting is prevented.
copper steel was found to be as much as 25o/o Added to steel's usual structural requirements,
more corrosion resistant than most copper-free drainage and quick drying are fi.mdamental for
steel, other similar steel containing only about good design and fabrication in weathering steel.
0.20Yoto 0.25o/o copper was found to have even
better resistance to atmospheric corrosion. Marketing
In submerged or buried condifions, however, Soon after 1933, U.S. Steel strongly promoted
the presence ofcopper in any amount proved its new "low-alloy high-tensile" product Cor-
of little benefit. It was recognued early that fre- Ten for the manufacture of railroad equipment.
quent repetition in wetting and drying the ex- Subsequently many diverse applications have
posed metal surfaces was an important factor in been developed, and with close attention to de-
maintaining exceptional corrosion resistance. sign constraints, weathering steel is now widely
employed in architectural and civil engineering
Beginning in 1916, the American Society applications. Cor-Ten is still marketed by U.S.
for Testing Materials (ASTM) conducted tests Steel's successor, U.S. X Corporation.
that clarified the benefit of incorporatng0.2%
to 0.5o/o copper and phosphorus into steel. In A 1937 U.S. Steel marketing booklet promotes
1929,the United States Steel Corporation (U.S. the use of Cor-Ten through three basic themes,
Steel) started developing proprietary corrosion which are still persuasive to the present day.
resistant "low-alloy" steel (having special prop- First and second are the economic benefits of
erties imparted by small additions of certain Cor-Ten due to its exceptional strength and sig-
elements). In 1933 U.S. Steel introduced its nificant maintenance reduction. Cost savings
"low-alloy high-tensile" steel product are promised based on using less Cor-Ten than
Cor-Ten A, and later, Cor-Ten B. a given strength specification would require in
carbon steel, and also based on the elimination
Today similar weathering steel (meeting of protective painting, due to Cor-Ten's very
ASTM specifications A242, A325 or ,4.588) is low corrosion rates. The third marketing theme
provided by several manufacturers. is that a well developed rust layer on Cor-Ten
steel is aesthetically pleasing.
U.S. Steel continued testing low-alloy steel
compositions, and in l96l its researchers When used properly, weathering steel is indeed
Larrabee and Coburn published corrosion rate economically beneficial due to its exceptional
data on I 5.5-year trials of 270 fypes of steel strength and minimal maintenance. The dark,
with systematic variations in the copper, nickel, tightly adherent rust patina that develops
chromium, silicon and phosphorus composi- through weathering is protective, given suitable
tion. Alloy variation also affects strength, hard- conditions, and is compatible with a broad
ness, working and other properties. Weathering range of contemporary styles.
steel was the earliest type of today's "high-
strength low-alloy" (HSLA) steel. In the early 1960s Cor-Ten was promoted in
the architectural and civil engineering literature
The design requirements specific to weathering as combining structural and aesthetic utility,
steel are not complex. To weather and endure and its adoption led to sculptural applications

308
within a few years. In 1966, Cobum promoted For many contemporary artists keenly aware of
unpainted Cor-Ten steel for use in highway relations between art history, art criticism and
applications, transmission towers and architec- art patronage, long-term survival of artworks
ture. He confirmed that Cor-Ten had been has been of great concem. According to conver-
tested with regard to atmospheric corrosion sational reminiscences from sculptors active in
resistance for 30 years, and stated that it had the mid-1960s and later, the advent of weather-
proved four to six times as resistant as carbon ing steel seemed to permit iron-working artists
steel. Coburn wrote that about 80 percent of to place sculpture outdoors with every expecta-
weathering steel's corrosion takes place during tion that it would last as long as bronze.
the first two or three years of exposure, with a
loss of about 38 pm (1.5 mil) metal thickness, In fact, when sculptors "discovered" weathering
and that the expected total loss of metal thick- steel in the mid-1960s, information was already
ness dwing the subsequent 20 years is only being disseminated on its susceptibility to cor-
about 12.5 pm (0.5 mil). These figures were rosion, and its consequent special design re-
for urban and semi-rural atmospheres, as ma- quirements.
rine and industrial atnospheres had proved
quite aggressive to weathering steel. Some artists and sculpture fabricators have un-
derstood these requirements and some have not.
Perusal of diverse marketing and design litera- ln 1970, U.S. Steel published Steel In Sculp-
ture published during the period 1934 to pre- ture, mak;ng important information more acces-
sent indicates that before the mid-I960s very sible, but many artists remained uninformed.
little mention was made of the special design re- As is often the case, new developments were
quirents for minimizing weathering steel corro- probably missed through lack of adequate com-
sion. While steel industry scientific staff may munication among artists, manufacturers and
have often advised designers directly, manufac- technical professional s, and an unfamil iarity
turers' literature gave such information little with existing literature. Certainly weathering
place before the early 1970s. By the late 1970s steel has been enthusiastically adopted by many
and early 1980s, manufacturers' literature fea- contemporary sculptors.
tured special design constraints and recom-
mended specific fabrication materials and A Distinctive Corrosion Process
methods.
The modem technology of weathering steel is
Highway designers readily adopted weathering an outgrowth ofearlier copper steel technology.
steel in the early 1960s, but in many cases they Current understanding of weathering steel's
neglected to provide for adequate drainage, so protective rust follows a long history of hy-
some parts of their structures retained moisture. potheses regarding similar phenomena.
As a result of disastrous consequences, civil
engineers and steel industry scientists were re- The conosion resistance of copper steel was
evaluating the usefulness of weathering steel actively investigated beginning in the late
and discussing its special requirements by the l9th century. Early explanations were based
mid-I960s. A similar development took place on galvanic cell chemistry and noted the impor-
in architecture. tance ofwater and air. By 1900 and soon after,
observation had established that copper steel in
The birth of modern steel sculpture between buried or submerged conditions showed no spe-
World Wars was followed by tremendous de- cial corrosion resistance. It was recognaedthat
velopment after 1945. David Smith and other complete drying-out at frequent intervals was
sculptors through the 1950s and 1960s adopted necessary to maintain copper steel's best corro-
the materials and techniques of iron-working. sion resistance. These observations have
Soorl steeljoined bronze and stone as the pre- subsequently proved applicable to weathering
eminent media of sculpture, and, influenced by steel's corrosion resistance.
architectural and other appli cations, sculptors
began using weathering steel in the late 1960s. In 1934 Gregg and Daniloffreported the
theories (1929 to 193 I ) of Carius, a German

309
scientist. Carius postulated barrier mechanisms process, but form water-insoluble sulfates, thus
closely related to earlier galvanic theories of removing sulfate ions needed for autocatalysis
corrosion protection. Carius taught that a and slowing the generation of corrosive iron
virtually impervious copper surface layer was sulfate. A similar article by Fyfe in 1969 adds
formed by dissolution of copper flom the cop- that the precipitated sulfates help to consolidate
per steel, followed by galvanic redeposition of a protective rust.
the copper at the steel surface, and formation of
alayer ofcupric oxide over the copper. Accord- ln 1966 Horton and Goldberg published eight
ing to Carius, corrosion was controlled by the excellent photomacrographic black-and-white
dual copper/copper oxide layer, which effec- images detailing cross-sections of the protec-
tively prevented air and water fiom reaching tive rust, noted a two-layer structue, described
the steel surface. Gregg and Daniloff men- layer compositions, and identified the predomi-
tioned that iron oxides forming simultaneously nant constituents as alpha- and gamma-
with the copper and copper oxide film must pre- FeOOH. However, despite a competent review
vent formation of a completely protective film, of pertinent literature, Horton and Goldberg
but otherwise accepted Carius' theory. While it were unable to suggest formation or functional
would appear that well developed light micro- mechanisms for the protective rust.
scopical methods widely available in Carius'
era had actually revealed the presence ofcop- Our current understanding of weathering steel
per layers on copper steel, Horton and Gold- rust derives from work published in two 1969
berg later stated that Carius' copper fikn had articles. In that year, Schmitt and Gallagher of
never been observed, and that copper films arti- U.S. Steel critiqued the Copson and Schikorr
ficially applied to steel had not resulted in the approaches as too simple to fully explain "the
degree of corrosion resistance observed in cop- many effects of alloying element combina-
per steel. Carius' galvanic and barrier mecha- tions," as indicated by U.S. Steel research pre-
nisms were ignored by pre-eminent weathering sented in 1961 by Larrabee and Coburn.
steel corrosion and passivation researchers little Schmitt and Gallagher mention a model,
more than a decade later, but more advanced "cathodic control," in which the rate of anodic
models of electrochemical processes and barri- iron oxidation depends on the rate ofcathodic
ers to corrosive agents remain fundamental in oxygen reduction. They suggest that the alloy-
explaining steel corrosion and passivation. ing elements have some unexplained effect on
the physical and electrical properties of the rust
In 1945, Copson proposed that HSLA steel films, properties that keep the rate of oxygen
containing copper, nickel and chromium forms reduction very low. This critique of earlier
sulfates of these elements during the corrosion thinking introduced modem electrochemical
process. He suggested that these compounds, concepts to the dialogue on weathering steel
less soluble than iron sulfates, precipitate and corrosion. The Schmitt and Gallasher
clog pores in rust (apparently still regarded as
steel's most commonly occurring rust, com-
x These concepts are also fundamental to cathodic pro-
prised mainly of magnetite, Fe:O+), thus re-
stricting water and air access to steel surfaces. tection methods for corrosron control, which first drew
Hence the rust is a protective barrier. attention around 1929 and which by 1949 were in wide
use and under enthusiastic developmental research. The
basic mechanism ofcathodic protection rs the sacrificial
Schikorr proposed in 1964 that in steel corro-
corroslon of a metal anode electrically joined to a
sion the reaction of iron with dissolved atmos-
cathodic structure berng protected. Electrochemical po-
pheric oxygen and sulfur dioxide produces tential, derived from the reduction ofa sunounding elec-
ferrous sulfates, which then react autocatalyti- trolyte as the anode corrodes, is applied to the cathodic
cally with water and dissolved oxygen to structure, where the excess ofpotential passivates the
produce iron hydroxide and sulfuric acid. The structure's surface.
acid reacts with iron to produce more sulfates
and water, and the cycle continues. According Today cathodic protection and protective coatlng are our
to Schikon, copper, chromium and nickel in major tools against corrosion of buried and submerged
weathering steel participate in the corrosion metals.

310
discussion also showed that the time was ripe ongoing worldwide study of exposed samples
for contribution of major significance from
a of weathering steel.)
Okada et al., also published in 1969.
Recently, Kihira, Ito and Murata re-examined
Okada and his colleagues made electrochemical weathering steel rust sfucture using computer-
measurements of the polarization charac- aided X-ray microanalysis and laser Raman
teristics of rusted weathering steel, and used spectroscopy. They reported that a thin layer of
polarized light microscopy in conjunction with phosphatic rust forms at the steel surface in the
the electron microprobe to examine rust formed initial stage of rust formation, with subsequent
in atmospheric corrosion. They reported a two- gradual formation of mostly amorphous iron
layered structure: an inner layer of "spinel- oxyhydroxide under the phosphatic layer and
type" iron oxide under an outer layer ofiron a mostly crystalline layer above.
oxyhydroxide (FeOOH). They suggested that
the amorphous oxide next to the steel surface Okada et al. laid the foundation for our current
is a kind of oxide polymer with unit structues understanding of weathering steel's protective
linked by hydroxide bridges, and that this rust. Subsequent work has been concemed with
structure acts as a barrier to oxygen and water. disceming finer physical structures, more accu-
Okada also reported that prolonged exposure to rately characterizing components and explain-
moisture without intermittent drying results in ing additional aspects of formation (and
the loss of the apparently protective inner layer. degradation) processes for the protective rust.
In 1991 Wurth summarized current thinkins
Subsequently, Misawa et al. studied weathering on the subject:
steel rust using infrared (IR) and far-IR spec-
troscopy, X-ray diffraction, and scanning elec- The protective iron corrosion products on
tron microscopy (SEM). They suggested in weathering steels form a two-layer structure.
1974that the inner layer is an amorphous ferric The inner layer of this strucfure consists of
oxyhydroxide, FeOl (OH)r-z* (which has much fine, tightly adherent amorphous and poorly
bound water), which can transform to alpha- crystallized fenic hydroxides. Among these
ferric species, the amorphous ferric oxyhy-
FeOOH. They inferred that formation of amor-
droxide peOl(OH)r-zxl is thought to be im-
phous oxyhydroxides could be catalyzedby
portant in preventing both the flow of Oz
copper, chromium and phosphorus.
and corrosive anions to the metal surface and
the flow of ferrous species into solution. The
A succinct 1975 arlicle by Pourbaix and de loose and often flaky outer layer is composed
Miranda summarized the then current under- mainly of this two-layer structure.
standing of the structure and the formation
mechanisms of protective rusts on weathering The sequence of reactions forming the two-
steel. Omitting earlier theories, they reviewed layer structure changes, depending on the
the work of Okada and of Misawa, then pre- environment to which the steel is exposed.
sented results from their own electrochemical Parameters such as pH, presence of solution
investigations, which modeled formation of the anions, and availability ofoxidizers deter-
protective rust. By analogy to the surface films mine which species are formed and how
ofcorrosion products, which they had produced quickly they decompose....
on weathering steel in galvanic cells at their
laboratory, Pourbaix and de Miranda identified Sculpture Preservation and
the protective rust's inner layer as alpha- Maintenance
FeOOH, goethite. They stated that when the
rust is wet, alloying elements, particularly cop- Important factors affecting the condition of
per, dissolve from metal into crevices, where weathering steel sculpture are design, fabrica-
they inhibit the formation of magnetite, Fe3O4 tion. location of site. installation and ambient
(non-protective rust) and catalyze the formation environment. The basic requirements for proper
of additional goethite, which consolidates the function of unpainted weathering steel surfaces
inner protective layer. (lncidentally, in 1983 are that they must have regular cycles of
Pourbaix and Pourbaix announced an wetting and complete drying. Thus, all designs

3ll
for unpainted HSLA steel must provide for However, abrasive blasting is not always possi-
good drainage, and all faying (i.e., closely fif ble or advisable. For instance, in many munici-
ted) surfaces and other features that tend to palities, local law prohibits outdoor silica,
attract and retain moisture, such as interior sur- alumina or similar abrasive blasting for reasons
faces and voids, must be well sealed. On the of public health and sanitation. The cost of
other hand, exposed surfaces that drain well blasting may also be a detenent. Finally, com-
and that are effectively dried by both sun and plete blasting may unnecessarily disrupt enjoy-
air currents do not require sealing. Notwith- ment of well-weathered sculpture. In some
standing the importance of technical aspects of instances, cleaned or repaired areas are small
media choices, the artist's primary responsibili- in relation to the sculpture as a whole, or they
ties are to observe and communicate in excep- may not be visible in some views.
tional ways. Physical properties of media
chosen by contemporary artists often prove Rusting of newly fabricated or blast-cleaned
unexpectedly problematic, and so present new weathering steel is sometimes stimulated or
challenges for care and maintenance. modulated by rinsing with salt solution. Avoid
chloride ion, one of metal's worst enernies;
An ad hoc approach, using mostly local infor- chlorides are very detrimental to weathering
mation and skills for preserving weathering steel's protective rust.
steel sculpture, has prevailed among sculptors,
sculpture fabricators, iron-working craft smen, Painting and Clearcoating: Painting and
restorers. owners. curators and conservators coating ofsteel are also facilitated by blasting.
since the mid-1970s. A variety of measures, Weathering steel sculptures are often painted,
some more effective than others, have been clearcoated or waxed in an ad hoc marurer.
taken, sometimes with valuable advice from However, coating originally unpainted weather-
steel industry experts. It is useful to consider ing steel sculpture is an ineffective conserva-
the most common measures, those we should tion measure because it adds the problems of
adopt and improve, and others we should weathered coatings to the usual problems of
avoid. Common ad hoc measures include the weathering steel. Thus, repair funds are con-
following: sumed without optimum benefits, more effec-
tive drainage and sealing measures may be
Cleaning: Careful cleaning using dry brush, delayed or diminished and necessary mainte-
broom, low-pressure water blast, organic sol- nance becomes more difficult and frequent.
vents or detergent solutions will often remove
paint, marker graffiti, or detritus, and these Both painting and clearcoating markedly alter
methods should always be tested hrst. Steam surface appearance. Painting imposes non-
cleaning can be very effective. Blasting (high- original color, while clearcoats and wax satu-
pressure hydro-blast, or air-blast abrasion using rate (darken) the patina color. Painting and
fine grain silica, alumina or plastic blast media) clearcoating also alter surface texture. Al-
is an ad hoc measure commonly used to elimi- though corrunon in ad hoc preservation of
nate marring by destructive rust, repair scars, weathering steel sculptures, such painting and
graffiti, old paint, and so forth. coating are seldom appropriate. Well-meaning
preservationists have compromised the historic
Stimulation of New Rust: Development of and aesthetic integrity of many weathering steel
new, visually homogeneous rust on weathering sculptures.
steel sculptures is often stimulated by blast-
cleaning. New rust normally appears within However, some sculptures are located on un-
days after blasting begins on outdoor installa- suitable sites where local moisture persists or
tions. At rates dependent on ambient condi- local activities invite graffiti or other destruc-
tions, the patina develops over several weeks tive contamination. ln such cases inobtrusive
from a bright orange-red to light red-brown. physical barriers should be installed to protect
Within a few months, a dark brown color usu- the sculpture. Following conservation, after
ally develops, which within a few years new patina has developed, it may be necessary
becomes very dark. to clearcoat only the threatened areas in order

312
to create a moisture barrier or to simpli$ hastened by hydro-blast or wire-brushing. Fre-
frequent surface cleaning. quently the visual effect of weld repair is miti-
gated by abrasive blasting ofthe entire affected
Refabrication: This practice is often necessary surface, or of the complete sculpture. Then the
to reinforce or replace deteriorated parts or en- entire blasted area weathers together and re-
tire sculptures, or to allow sealing or elimina- gains its palina, and the appearance of the weld
tion of corrosion-prone features. It is important area matches the rest of that surface. As dis-
to ensure that appropriate materials are used in cussed above, blasting is not always advisable.
refabrication. If possible, the weathering steel
used for the repair should be the same type as Although the bright contrast of welds with
that originally used in the sculpture. adjacent unaffected mature patina is alarming
at frst, the appearance ofproperly welded areas
Assembly hardware, such as screws, bolts, improves quickly. After four to six months of
washers and nuts, must be compatible with the normal weathering in urban atmospheres,
steel. Sometimes the original hardware is not welded areas are quite dark. After a year, it
compatible and this often results in corrosion of may be difficult to pick out a weld on the sur-
the hardware and surrounding metal. Currently, face of a large structure. As time goes on the
manufacturers recommend ASTM 316 stainless welded area blends very well.
steel, or weathering steel ASTM A325 for
bolts, washers and nuts. When possible, dis- Sealing Corrosion-prone Sites
similar metals should be isolated by elechically
insulating hole liners and washers to avoid bi- Before sealing at all edges, faying surfaces
metallic corrosion. Hardware joints and holes prone to corrosion must be cleared as com-
must be sealed, since they otherwise tend to pletely as possible of loose or impacted matter
retain moisture and corrode. (very high pressure hydro-blasting can give a
reasonably thorough effect), and dried thor-
Except in non-welded structures, arc welding oughly by compressed air or heat. Joints and
is the appropriate method of repair joining, and gaps can often be sealed effectively using
should follow manufacturer's specifications. weather-resistant, flexible and long-lasting
Generally, single-pass and structural welds may caulks, such as RTV silicones. When necessary
be made using E70-group, mild-steel covered and appropriate, deteriorated sculptures may
electrodes, as long as the weld ensures proper be taken apart. The faying surfaces are then
composition enrichment from adjacent steel. cleared thoroughly by blasting and are well
More care must be taken, and different weld sealed with appropriate paint, which also serves
methods and fill materials used, when the ex- as primer forjoint sealant. Sculptures are reas-
posed weld is broad or must closely match the sembled with "formed in place gaskets" of
surrounding surface in color and corrosion re- caulk sealing the joints. Sometimes instead of
sistance. Manufacturer speci fi cations should be priming and caulking, faying surface edges
obtained and followed; certified welders should are sealed by welding.
be employed and closely supervised.
Drainage Enhancement This is a common
The heat of welding destroys weathering steel strategy for dealing with water pooling, which
patina close to the weld. Heat from the front causes destructive corrosion. It is importrant that
face of a steel plate will usually destroy patina the sculpture's foundation site be well drained,
on the back face. The resulting scars are quite and that an effective moisture barrier exists be-
unsightly and bright orange. After covered elec- tween the base of the sculpture and the surface
trode welding, burnt patina will be obscured beneath it. In the sculpture itself, weep-holes or
temporarily by a blotchy white welding fume slots allow water to drain away rather than to
residue. The burnt surface soon regenerates its pool. Such drains must be properly located and
protective rust under the darkening residue of of correct size: large enough to avoid clogging
the burnt patina, which separates from the yet small and inobtrusive. Weep-holes and slots
surface in paper{hin fragments. Eventually that tend to collect detritus must be cleared
the burnt material crumbles away; this may be regularly to promote proper drainage.

313
Corrosion-prone surfaces in areas of water Public art may suffer popular or ofhcial rejec-
pooling that cannot be drained may need to be tion; this is especially true of modern or con-
sealed with aclear, water-excluding coating. temporary art if the public or the government
When such areas are out of sight, the aesthetic associate the art with an unfamiliar or elite
impact of coatings is minimal. Of course, taste. A display site may seem ill chosen.
such coatings must be maintained in good Safety issues often are important in dealing
condition. with deteriorating structures. The daunting pros-
pect of an artwork's conservation may over-
It is also very important to frequently clear whelm those entrusted with its care. Curators,
away accumulations of rust, dirt, dust and other planners, and administrators can easily lose en-
detritus, which tend to build up in gaps, comers thusiasm for conservation projects that require
and on horizontal surfaces ofweathering steel substantial amounts of time and money. Plan-
structures. Such accumulations become slow- ning often omits consultation with expert con-
drying poultices, which keep weathering steel servators, relying exclusively upon local
surfaces actively corroding. administrators, committees, tradespersons, ar-
chitects or engineers. In these and similar situ-
Appropriate conservation requires thorough ations, the worst of the ad hoc measures are
and periodic examination, condition evaluation, sometimes employed. Unfortunately, many
documentation and reporting by a qualified con- weathering steel sculptures and other contempo-
servator. The conservator should specify and rary artworks with apparently severe condition
implement effective treatment procedures and problems have been altered, removed from
a regular maintenance prognm for cleaning, context or destroyed due to inadequate
drainage enhancement, dry sealing of faying consideration and triage.
surfaces, and structural repair as necessary.
Localized clearcoating, though sometimes un- To enhance appropriate and effective conserya-
avoidable, should be limited. Complete paint- tion planning and implementation it is very
ing or clearcoating should be strictly avoided. It important to obtain genuine expertise and prac-
may be possible to protect grade-level or buried tical understanding conceming specific art
features by cathodic means. The conservator media, and appropriate technology for preserva-
should also supervise or personally carry out tion, restoration and maintenance. Planners
both the recommended treatment procedures must also draw upon historically and techni-
and maintenance program. cally accurate assessments of aesthetics,
significance, structure and condition. The con-
Ethics in Conserving servator's role in consulting or project manage-
Contemporary Art ment encompasses developing knowledge and
providing information, guiding and deve loping
Simplistic problem-solving measures, such as conservation specifications, supervising conser-
sandblasting, painting, relocating or removing, vation activities, and helping all concerned to
characterize the ad hoc approach to caring for maintain frrm commitment to the preservation
weathering steel sculpture. Unfortunately, de- and integrity of each work of art.
struction of deteriorating artworks is another ad
hoc solution, efficient in eliminating problems, Conclusion
but incompatible with the conservator's most
fundamental responsibilities. Ethically minded In most cases weathering steel sculpture can be
conservators will refuse the often-offered roles preserved by proper cleaning, replacing deterio-
of arbiter, facilitator or confessor in the altera- rated parts, enhancing drainage, sealing mois-
tion or disposal ofart. To help us transform de- ture-prone features and following a regular
structive attitudes and educate decision makers program of maintenance. These measures,
through our own practice ofresponsible conser- when planned and implemented with a good
vation, conservators must understand why inap- understanding of weathering steel, can stabilize
propriate measures are often proposed. the sculptural structure and patina. The appear-
ance of a stable and well-maintained

3t4
weathering steel sculpture is enhanced with the
passage of time.

It seems that the surface and other remarkable

qualities of weathering steel are being closely
studied by corrosion scientists outside the
conservation field. In addition to this research,
it would be really useful for conservation
scientists to undertake sustained studies on
the care of weathering steel. For now however,
conseryators with some rudimentary under-
standing of weathering steel must employ
provisional solutions to pressing conservation Figure 2Atrnosphere and Environment X Louise
Nevelson, Fairmount Park, Philadelphia. Photo :
problems.
J. Scott (detail).

Notes for Figures 1r2,,3 and 4 comprises 30 boxes of 12.5 mm (0.5 in.) weath-
ering steel plates screwed and bolted together
Afmosphere and Environment Xby Louise
with a welded base, all on a concrete footing.
Nevelson ( 1970), shown in Figures | , 2 and 3,
This open structure incorporates many water
catchments and faying srrfaces both broad and
naffow. The boxes were originally held to-
gether by unsealed, uninsulated screws of mild
steel, which corroded rapidly. At the edges of
broad faying surfaces considerable gaps devel-
oped within several years of installation, due
to rust expansion in continually moist interiors.
Adjacent narow (12.5 mm) faying surfaces dry
well, so edges remain in close contact. Non-
conserved exemplars of this edition located in
Philadelphia, New Haven and elsewhere are in
very poor condition after 20 years ofexposure.

Night Presences IVby Louise Nevelson

(1912), shown in Figure 4, comprises several
smooth closed forms fabricated of weatherins

Figure I Atrnosphere and EnvironmentX Louise

Nevelson, 1970, Princeton. The John B. Putnam Jr. Figure 3 Atmosphere and EnvironmentX Louise
Memorial Collection. Photo: The Art Museum, Nevelson, 1970, Princeton. Photo: The Art Museum,
Princeton Universitv. Princeton Universitv. Y.L. Kim.

315
steel with all seams welded and no faying sur-
faces, all set on a concrete footing. The entire
structure drains very well, except that detritus
collects and expansive rust forms at the bottom
of a few narow gaps between vertical ele-
ments. While there has been some restoration
of the horizontal surface of the base, the sculp-
ture is today in fine condition after 20 years of
exposure.

rJ I" micron
Figure 5 Weathering steel and patina. PLM
brightfield, no analyzer. Photo taken by J. Scott at
the N.Y. Conservation Center, Inc., New York City.

the light path, the patina region becomes two: a

bright and largely bireftingent outer area next
to the mounting medium, at the patina's sur-
face, and a dark, largely isotropic inner area,
next to the steel (Figure 6). These exemplifu
weg+gs ng steel patina's characteristic
tWd-layer structure. which altemates mostly
amorphous oxyhydroxides with more crystal-
line ferric hydroxides (Figure 7). Also visible
are parallel strata comprising these layers, and
tiny faults running through the strata.

The patina in another mounted sample is exam-

ined using the scanning electron microscope
(SEM), magnified 1,000 times (Figure 8).Note
that darker regions are more dense than lighter
regions. Shatification is evident, with many
faults, but next to the metal at left is a very
continuous dense layer.
Figure 4 Night PresenceslY Louise Nevelson,
Park Ave. at 92 St., New York City. Photo: J. Scott
(detail).

Notes for Figures 5, 6, 7 ,8, 9 and 10

Polished cross-sectional surface samples,of

long-exposed weathering steel il!4;trate the
structure of protective patina. V

Examination is through the use of metallograph

and brightfield polarized light. The patina in a
mounted sample ca. 65 pm wide at center is a
somewhat shiated grayish region next to the
r-r L micron
bright steel (Figure 5). The darkest region, on Figure 6 Weathering steel and patina. PLM bright-
the other side of the gray band, is the mounting /ield, crossed polars. Photo taken by J. Scott at the
medium. Examined with the analyzer placed in N.Y. Conservation Center, Inc., New York Citv.

316
 depicts a richly textured and fissured outer
patina surface.

i:i ,r'.$'":.ir*
r\i
,-' l 'i'!

Figilre 7 Weathering steel and patina. A) steel, B)

patina inner, and C) patina outer. (Diagrammatic
representation oJ Figure 6). Diagram: J. Scott.

Figure 9 Weatheing steel and patina. SEM 750x.

Photo taken by J. Scott at Lucius Pitkin, Inc., New
York City.

Figure 8 Ll/eathering steel and patina. SEM 1,000x.

Photo: J. Scott/Mark Wypyski (Wypyski is a Re-
search Associate, Department oJ Objects Conserva-
tion, The Metropolitan Museum of Art, N.Y.).

The comer of an unmounted sample is Figure l0 lVeathering steel and patina. SEM 200x.
examined using SEM, magnified 750 times Photo taken by J. Scott at Lucius Pitkin, Inc., New
(Figure 9). The comer between the polished York City.
face and the outer surface ofthe patina is
viewed directly, at an angle about 135' from Appendix
each surface. Thejagged profile ofthe interface
between steel and patina on the polished face A Note on Terminology
appears in the lower half of the image. Above Certain terms and ideas are introduced in the
the dark (dense) inner patina layer is the lighter forms they actually took in the articles dis-
(less dense) outer layer, and the outer surface is cussed. This is to emphasize the chronological
seen above the comer. nature of the account of weathering steel.

Examining the unmounted sample with SEM, For instance, U.S. Steel originally referred to
magnified 200 times, the viewing angle is their weathering steel product as "Cor-Ten,"
moved past the corner, away from the polished and later as "CorTen." More recently the term
face, which appears in the lower part of the "USS CORTEN" has appeared, probably to
image (Figure l0). The balance of the image avoid possible loss of trademark status with the

JI I
product name's slide into the generic. Today Certaines sculptures faites d partir de ce matt-
we use the term "weathering steel," while dur- riau se sont strieusement ddtdriordes, parfois peu
ing the 1930s to the 1960s technical parlance fa- de temps apr,Ds leur installation, tandis que d'au-
vored "CorTen" or sometimes "high-strength tres rbsistentfort bien aux intempdries pendant de
low-alloy steel" or "HSLA steel." Today not all nombreuses annies, et ce, m€me si elles ne font
HSLA steel is weathering steel. I'objet que de peu de soins.

Nous dicrirons briivement I'histoire de la tech-

Weathering steel had its genesis in the technol-
nique du corten, ainsi que la commercialisation
ogy of"copper steel," a term still current,
de ce produit et ses applications, tout en souli-
which predates the 20th century in common use
gnant I'intdr€t qu'il suscite, de faqon paralldle ou
and which refers to a specific type of industrial
divergente, au sein de I'industie et chez les ar-
product. Another current term, "copper-bearing tistes. Nous fournirons, d'autre part, des explica-
steel," has a broad meaning, but it is more re- tions quant aux connaissances scientifiques que
cent and was not seen in the technical literature I'on posside actuellement sur la nature de la
ofthe 19th or early 20th centuries. sudace d'oxyde protecteur, connaissances qui
aideront d mieux comprendre levice inhtrent et
Acknowledgement manifeste que prdsente le corten. Les probldmes
usuels de conservation que pose ce matdriau sont
Field examinations and consulting assignments, clairement dus d desfacteurs li6s d sa conception
together with major conservation projects for et d des facteurs environnementaux.
Princeton University (The Putnam Collection
of Modem Sculpture) and for the Board of Edu- Nous ferons, par ailleurs, quelques sugges tions
cation of the City of New York (Public Art for au sujet des methodes de conservation les plus
p er ti ne n t es. E t, no u s fo urni ro ns un e il lus tration
Public Schools), have helped the author to
leam, develop, apply and evaluate conservation des mesures de restauration ponctuelles qui sont
actuellement trds nlpandues, tout en |valuant tant
methods for weathering steel sculpture.
leur fficaciti sur le plan technique que leur
pertinence du point de we esthdtique et histoi-
R6sum6 que. Ces propos seront etayes d'etudes de cas
portant sur des objets qui sont actuellement
La conservation de sculptures en acier en traitement.

Utilisi depuis longternps dans les arts dtcoratifs Une bibliographie d jour de documents scienti-
et en architecture, I'acier en est sans doute venu
d constituer le matiriau principal des sculptures
fiques, techniques et critiques portant sur le sujet
accompagnera enfin cette communication.
duXX sidcle. Les sculpteurs travaillent disor-
mais d sa conception et d sa production, au m'Ame
titre qu'ils s'inttressent de longue date aux tech- References and Bibliography
niques de la sculpture, du moulage et de la fonte.
De m€me, les spdcialistes de la restauration des A sampling of pertinent literature regarding
sculptures d' auj ourd' hui se preoccupent-ils de weathering steel was compiled for this paper.
mieux connaitre I'histoire et la technologie de Not all references are cited in the text.
I'acier, et de mieux comprendre le phdnomdne
de la degradation de ce matffiau, ainsi que les Aitchison, L. and W.I. Pumphrey, Engineering
techniques permettant de le prdserver et de Sreels (London: Macdonald & Evans, 1953).
le restaurer.
Boyer, H.E. and T.L. Gall, eds., Metals Hand-
Dans le cadre de la prdsente communication, book,l}th edn. (Metals Park, Ohio: ASM Inter-
nous parlerons surtout de I'acier patinable d national, 1990).
haute nlsistance faiblement alli6, mieux connu
sous le nom de ( corten > (du < USS COR-TEN > Brockenbrough, R.L. and B.G. Johnston, Steel
de la U. S. Steel) qui, ayantfait son apparition Design Manual (Pittsburgh: U.S. Steel Corp.,
dans la sculpture contemporaine au milieu des
1968).
anntes 60, continue d'€tre utilist auiourd'hui.

318
Buck, D.C., "Influence ofVery Low Eichhorn, K.J. and W. Forker, "The Properties
Percentages of Copper in Retarding the Corro- of Oxide and Water Films Formed during the
sion of Steel," Proceedings of ASTM, vol. 19, Atmospheric Exposure of Iron and Low Alloy
no.2, 1919. (Abstracted in Iron Age, 104, Steels," Corrosion Science, vol.28, no.8, 1988,
r9r9,p.44). pp.745-758.

Cathodic Protection, a Symposium by the Evans, U.R., Metallic Comosion, Passivity

Electrochetnical Society and the National and Protection, 2nd edn. (New York: Edward
Association of Conosion Engineers Arnold &Co.,1946).
(NewYork: NACE, 1949).
Forker, W. et al., "Untersuchungen zur aus-
Chizhmakov, M.B. and M.B. Shapiro, "[Jse of bildung von deckschichten auf niedriglegierten
Physical Methods for Investigation of Corro- staehlen und baustahl mit einem wechselt-
sion Resistant Steel and Alloys," Chemical and archverfahren," NeLte Huette, vol. 18, no.4,
Petroleum Engineering, vol. 23, nos. I l- 12, April I 973. pp. 235-240.
July 1988, pp.625-628.
Fromhold, A.T. and S.J. Noh, "The Transport
Coburn. S.K.. 'A Low-Cost Maintenance-Free of Ions and Electrons through Microscopically
Structural Steel for Highway Applications," Inhomogenous Passive Films; Breakdown
Highway Research Record, 1 10, Washington, Implications," Corrosion Science, vol. 29,
D.C., 1966. nos.2-3, 1988, pp. 237-255.

Coburn, S.K., "Increasing Container Service Fyfe, D. et al., "Atmospheric Corrosion of

Life with Painted USS CORTEN Steel," Pro- Fe-Cu Alloys and Cu-containing Steels,"
ceedings, Second Container Technologt Con- 4th International Congress on Metallic
ference, Brighton, England, Dec. 1978, p.177. Corrosion, Amsterdam, Sept. 7-14, NACE,
Houston, 1969.
Cobum, 5.K., Conosion Sourcebook (Houston,
TexasAvletals Park, Ohio: ASMAIACE, 1984). Glueck, G., "Sculptor's Ordeal With Steel: It's
Pretty but Temperamental," New York Times,
Copson, H.R., "A Theory of the Mechanism of August 23, 1991, pp. Cl3, Cl1 .
Rusting of Low Alloy Steels in the Atmos-
phere," Proceedings of ASTM,vol.45, 1945, Gregg, J.L. and B.N. Daniloff, Alloys of lron
pp. 554-580. and Copper (New York/London: McGraw-
Hill, 1934) [N.8.: good bibliography
Cor-Ten, the Low-Cost, High-strength, Corro- r627-1934.1
sion-resisting Steel of Many Uses (New York:
United States Steel Corporation, 1937) lCopi- Hayes, J.M. and S.P. Maggard, "Economic
ous listings of USS researchers' and marketers' Possibilities of Corrosion Resistant Low Alloy
articles and lectures on USS Corten, 1933.] Steel in Short Span Bridges," Proceedings of
Nation al En gin eering C onferen ce, Denver,
Dinkaloo, J. "The Steel Will Weather CO, May 5-6, 1960, pp. 59-60.
Naturally," Architectural Record, Aug. 1992,
pp. 148-150. Hayne, F.H., J.W. Spence and J.B. Upham
(Environmental Protection Agency Environ-
Dinkaloo, J. "Bold and Direct Using Metal in a mental Research Center in Research Triangle
Strong Basic Way," Architectural Record, July Park, N.C., U.S.A.) "Effects of Air Pollutants
1964,pp.135-142. on Weathering Steel and Galvanized Steel: A
Chamber Study," Materials Performance,
Duennwald, J. and A. Otto, "Investigation of vol. 15, no. 4 (Houston: National Association
Phase Transitions in Rust Layers Using Raman of Corrosion Engineers) April 1976, p. 48.
Spectroscopy," Corrosion Science,vol. 29, no.
9,1989,pp. 1167-1176.

319
Heidersbach, R. and F. Purcell, "Analysis of Kihira, H., S. Ito and T. Murata, "Quantitative
Corrosion Products with the Use of the Raman Classification of Patina Conditions for Weath-
Microprobe," Microbeam Analysis I 984, ering Steels Using a Recently Developed Instru-
Proceedings ofthe l9th Annual Conference ment," Corrosion, vol.45, no.4, April 1989,
of the Microbeam Society, July 16-20, 1984, pp.347-352.
Bethlehem, PA, Ronig, A.D. and J.I. Goldstein,
eds. (San Francisco: San Francisco Press, Inc., Kihira, H., S. Ito and T. Mwata, "Behavior of
1984). Phosphorus During Passivation of Weathering
Steel by Protective Patina Formation," in:
High Strength Low Alloy Sreels Publication Passivation of Metals and Semiconductors,
ADV 1 809R3 -5M-21 48 (Cleveland: Republic Proceedings of the Sixth International Sympo-
Steel. 1974). sium on Passivity, Sapporo, Japan, September
24-28,1989, Part 1 (of2), Corrosion Science,
Honzak, J.G.V., "Macroscopic Structure of vol. 31, part l, 1990, pp.383-388.
the Rust Layer Formed in the Atmospheric
Corrosion of Steel," British Corrosion Journal, Kim, Y.L., "Problems in Cor-Ten Steel Sculp-
vol. 8, no. 4, July 1973,pp.162-166. ture," Preprinfs, AIC 7th Annual Meeting, May
30-June l,1979, Toronto, pp. 59-65.
Horton, J.B. and M.M. Goldberg, "Distribution
of Alloying Elements in the Rust Layers Komp, M.E., "Atmospheric Corrosion Ratings
Formed Naturally on Corrosion Resistant Low- of Weathering Steels Calculation and Sig-
Alloy Constructional (Weathering) Steel," 4th nifi cance," Materia ls -P erformance, v ol. 26,
International Congress on Metal lic Corrosion, no. 7, July 1987, pp. 42-44.
Amsterdam, September 7 -1 4, 1969 (Houston:
NACE. 1969). Kruger, J, "The Nature of the Passive Film on
Iron and Ferrous Alloys," Coruosion Science,
Horton, J.B. and M.M. Goldberg, "The Rusting voI.29, nos. 2-3, 1988, pp. 149-152.
of Low-Alloy Steels in the Atmosphere,"
American lron and Steel Institute Regional Kukurs, O. et al., "Structure of Rust Layer on
Technical Meetings 1965 (New York: Ameri- Low-Alloy Steels," Protection of Metals,vol.
can kon and Steel Institute, 1966) pp. l7l-195. 21,no.3, May-June 1985, pp. 349-353.

HSLA Steel for Architectural Applications U .5. Larrabee, C.P. and S.K. Coburn, "The Atmos-
Steel Publication ADUSS-8 8-6659-02 (Pitts- pheric Conosion of Steels as Influenced by
burgh: United States Steel Corporation, 1977). Changes in Chemical Composition,"
First International Congress on Metallic
Kandeil, A.Y. and M.Y. Mourad, "Effect of Corrosion, (London: Butterworth, I 96 I )
Surface Texture on Corrosion Behavior of pp.276-285.
Steel," Surfoce & Coatings Technolog,,,vol.
37 , no. 2, 1989 , pp. 237 -250 . Leidheiser, H. Jr., "Corrosion Behavior of Steel
Pretreated with Silanes," Corrosion, vol. 43,
Kawasaki River Ten Atmospheric Corrosion no. 6, June 1987, pp. 382-387.
Resistant Steels (Tokyo: Kawasaki Steel
Corporation, 1977). Leiderheiser, H. Jr. and I. Czako-Nagy,
"Mossbauer Spectroscopic Study of Rust
Keane, J.D. et al., Remedial Painting of Weath- Formed During Simulated Atmospheric
ering Steel: A State-of-the-Art Survqt (Pitts- Corrosion," Cotosion Science, vol. 24, no. 1,
burgh: Steel Structures Painting Council, 1984). 1984, pp. 569-577.

Keiser. J.T. et al.. "Characterization of the Lloyd, B. and M.I. Manning, "Episodic Nature
Passive Film Formed on Weathering Steels," of the Atmospheric Rusting of Steel," Corro-
Corrosion Science, vol. 23, no. 3, I 983, sion Science,vol.30, no. l, 1990, pp.77-85.
pp.25l-259.

320
Materials Handbook, l3th edn. (New York: Pourbaix, M. and L. de Miranda, "Weathering
McGraw Hill, 1991). Steel's Performance and the Effect of Copper,"
ATB Metallurgie, vol. 29. no. 4, 1983,
Metals and Alloys: Steel, WeatheringPlblica- pp.7.1-7.18.
tion 5, Folder 2424 ( Bethlehem, PA: Bethle-
hem Steel Co.). Pourbaix, M. and A. Pourbaix, "Recent
Progress in Atrnospheric Corrosion Testing,"
Misawa, T. et al., "Mechanism of Atmospheric Corrosion, vol.45, no. l, 1989, pp. 7l-83.
Rusting and the Protective Amorphous Rust on
Low-Alloy Steel," Conosion Science, vol. 14, Pourbaix, M. (J.A.S. Green, trans., R.W. Stae-
no. 4, April 197 4, pp. 27 9-289. hle, trans. ed.) Lectures on Electrochemical
Corrosion (New York/London: Plenum Press
Misawa. T. et al.. "Corrosion Science in for CEBELCOR, Brussels, 1973).
Rusting of Iron and Weathering Steel,"
Boshoku Gijutsu, vol. 37, no. 8, 1988, Priest, H.M. and J.A. Gilligan, Design Manual
pp.50l-506. for High-Strength Steels USS #ADVL-215-54
(Pittsburgh: U.S. Steel Corporation, 1954).
Modern Steels and Their Properties,6th edn.,
Handbook 268-E (Bethlehem, Pennsylvania: Products and Publications of United States
Bethlehem Steel Co., l96l). Steel Corporarror (New York: U.S. Steel
Corp., 1929). [This pamphlet includes an exten-
Okada, H. et al., "The Protective Rust Layer sive bibliography of publications by U.S. Steel
Formed on Low-Alloy Steels in Atrnospheric and its affiliates on copper steel.]
Corrosion," 4th International Congress on Me-
tallic Corcoslor, Amsterdam, September 7-14, Raman. A.. "Characteristics of the Rust from
1969 (Houston: NACE, 1969). Weathering Steels in Louisiana Bridge Spans,"
Corrosion, vol. 42, no. 8, Aug. 1986,
Paint Lasts Longer on USS COR-TEN Steel pp.447-455.
U.S. Steel Publication ADUSS-87-2 133-02
(Pittsburgh: United States Steel Corporation, Raman, A., "Atmospheric Corrosion Problems
r97s). with Weathering Steels in Louisiana Bridges,"
Degradation of Metals in the Atmosphere,
Performance of Weathering Steel in Highway ASTM Technical Publication STP no. 965
Bridges (Washington, D.C.: American Iron & (Philadelphia: ASTM, 1988) pp. 16-29.
Steel Institute, 1982).
Raman, A., S. Nasrazadani and L. Sharma,
Popova, V.M. et al., "Corrosion Resistance of "Morphology of Rust Phases Formed on Weath-
Low-Alloy Steels in the Atrnosphere," Protec- ering Steel in Various Laboratory Corrosion
tion of Metals, vol. 18, no.2, March-April Tests," Metallography, vol.22, no. l, January
1982, pp. 129-133. 1989,pp.79-96.

Pourbaix, M. and L. de Miranda, "On the Na- Raman, A. and S. Nasrazadani, "Packing Corro-
ture of the Rust Layers Formed on Steels in sion in Bridge Structures," Corrosion, vol. 46,
Atmospheric Corrosion as a Function of Alloy no. 7, July 1990, pp. 601-605.
Composition, Environmental Composition,
Temperature and Electrode Potential," inl. Pas- Riederer. J.. " Rostender Stahl : Materialverhal-
sivity and Its Breakdown on lron and lron Base ten und Restaurierungsprobleme," Nur Rost
Alloys, Staehle, R.W. and H. Okada, eds. (Munstger und Marl: Landschaftsverband,
(Houston: NACE, 197 6) pp. 47 -48. Westfalen Lippe, 1986) pp.26-29.

Ruth, U., "Aussenskulptur und Umwelt heute-

zu den praktischen Aspelcten," Museumkunde,
vol. 51, no. 3, 1986, pp.144-148.

321
Sakai, T. and S. Tokunaga, "Maintenance-Free Strekalov, P.V., "Wind Regimes, Chloride
Service of Weathering Steel by Rust Stabiliza- Aerosol Particle Sedimentation and Atmos-
tion Accelerating Treatment," Corros ion/79 pheric Corrosion of Steel and Copper,"
(Houston: NACE, 1979). Protection of Metals,vol.24,no.5, May 1989,
pp.630-641.
Schamweber. D. et al.. "Electrochemical and
Surface Analytical Investigation of Weathering Szaver, T. and J. Jackobs, "Pitting Corrosion
Steels," Corrosion Science, vol. 24, no. 2, of Low Alloy and Mild Steels," Corrosion
1984,pp.67-82. Science, vol. 16, no. 12, 1976, pp. 945-949.

Schmid, E.V., "The Weatherability of CorTen Tamba, A. and G. Bombara, "Die Porositaet
Steel," Applica, vol. 87, no. 12,1980, pp. l0-13. und Hydrophiliedes Rostes von niedriglegierten
staehlen." Archiv Fuer das E is enhuett enwes en,
Schmitt, R.J. and W.T. Gallagher, "Unpainted vol. 43, no. 9, Sept. 1972,pp. 7 13-7 19.
HSLA Steel for Architectural Applications,"
Material Protection, vol. 8, no. 12, December Thomas, N.L., "Protective Action of Coatings
1969. on Rusty Steel," Journal of Protective Coat-
ings and Linings, vol. 6,no.12, Dec. 1989,
Schwitter, H. and H. Boehn, "Influence of pp. 63-69.
Accelerated Weathering on the Corrosion of
Low Alloy Steels," Journal of the Electro- Tosto, S. and G. Brusco, "Effect of Relative
chemical SocieQ,vol. 121 , no. 1, June 1980, Humidity on the Corrosion Kinetics of HSLA
pp.15-20. and Low Carbon Steels," Conosion, vol. 40,
no. 10, Oct. 1984, p.507.
Shastry, C.R., J.J. Friel and H.E. Townsend,
" S ixteen-year Atmospheric Corrosion Perform- The Use of Unpainted Low Alloy-High
ance of Weathering Steel in Marine, Rural and Strength Steel Publication 3 3 (Ottawa:
Industrial Environment s," D e gr ad ati o n of M et - Canadian Good Roads Association, 1968).
als in the Atmosphere, ASTM Technical Publi-
cation STP no. 965 (Philadelphia: ASTM, USS COR-TEN High-Strength Low-Allolt Steel
1988) pp. 5-15. U.S. Steel Publication ADUSS-88-7888-01
(Pittsburgh: United States Steel Corporation,
Shikorr, IMerkstoffe u. Korrosion, 15,1964, le80).
p.457.
Vedenkin, S.G., "Effect of Composition on At-
Staehle, R.W. and H. Okada, eds. Passivigt and mospheric Corrosion of Low Alloy Structural
its Breakdown on lron and lron Base Alloys Steels," Protection of Metals, vol. 11, no. 3,
(Houston: NACE, 1976). May-June 197 5, pp. 259-27 0.

Steel in Sculpture U.S. Steel Publication Wurth. L.A.. "The Role of Chromium in
ADUSS-88-4830-01 (Pittsburgh: United States Weathering Steel Passivation," Mqteriqls
Steel Corporation, 1970). Performance, vol. 3, no. l, January 1991,
pp.62-63.
Storad, A.S., "Coating Systems for High
Strength Low Alloy Steel, Are They Neces-
sary?" Corrosion/S3, Paper 288 (Houston:
NACE. 1983).

JZZ
Testing and Development of Conservation Processes
Elaboration et mise ir I'essai des m6thodes de conservation
An Evaluation of Eleven Adhesives for
Repairing Poly(methyl methacrylate) Objects and Sculpture

Don Sale

Keeper (Cons erv ation)

As s is t an t
Sainsbury Centre for Visual Arts
University of East Anglia
Norwich, England

Abstract Stress-crazing due to adhesive contact was inves-

tigated following the American Societyfor Test-
Perspex or Plexiglas, poly(methyl methacrylate) ing and Materials (ASTM) tests F 791-82 and
(PMMA), isfound in collections rangingfrom F 484-83. The tmsile strength of butt-joints was
fine art and design to social history. First used determined. The effects of detergent (0.5% Syn-
for aircraft glazing in the Second World l(ar, peronic 'N'non-ionic) and solvent (petroleum
PMMA is suitablefor sculpture, furniture, jew- spirit, boilingrange 100"C to 120"C) degreasers
elry, clothing and industial objects because of its on the tensile strength of adhesives were com-
optical and physical properties. When PMMA ob- pared. PMMA test specimens were takenfrom a
jects are broken it is difficult to choose an appro- 3 mm clear cell-cast (non-directional) ICI Per-
priate adhesive because the material is often spex sheet.
transparent and susceptible to stress-crazing or
cracking when exposed to adhesives or degreas- HXTAL NyL-|, Norland 0A65 and a I : I mixture
ing solvents. Since PMMA may be stressed in the of Acryloid 8-67 and Acryloid F-10 may be suit-
manufacture ofthe sheet and/or thefabication of able for repairing certain types of damage.
an object, crazingis highly probable.
Introduction
A series of tests were performed investigating the
properties of I I adhesives from five categories: The widespread use of poly(methyl methacry-
late) @MMA) in linctional as well as decora-
o Two-component all-acrylic cold setting tive objects guar€lntees its presence in museum
Tensol 70 and Acrifix 90 collections; hence its significance for conserva-
tors. Soon after commercial availability n 1934
c Two-component epoxy resin cold setting this thermoplastic was applied as aircraft glaz-
Ablebond 342-I and HXTAL NYL-| ing material superior to glass due to its light
weight, optical and physical properties. By the
o Single-component ultraviolet light curing late 1930s and early 1940s sculptors such as
Norland 0A65 and Norland OA68 Naum Gabo and Moholy-Nagy were machining
and thermoforming this material into numerous
o Polymer in solvent shapes. The many colors and surface decoration
Tensol 12, Acryloid B-72 and a I: I mixture techniques possible (i.e., engraving, frosting or
of Acryloid 8-67 and Acryloid F-10 printing) extend the application potential of this
material. Fumiture, jewelry, lamp fittings, illu-
o Cyanoacrylates minated sigrrs, juke boxes, thermoformed prints
Loctite 406 and Loctite 460 and architectural models illustrate a few

325
ex€Imples of objects made of PMMA. A partial Strength in PMMA depends on manufacturing
list of commercial names includes Perspex technique as well as molecular weight and mo-
(ICI) and Plexiglas (Rohm and Haas), and lecular orientation. PMMA is amealed or "nor-
molding powders, such as Diakon (ICI), Lucite malized" to reduce internal stresses, making it
(Du Pont) and Vedril (Montecantini). more resistant to stress-related damage, such
as crazing.- In addition, annealing produces a
Properties of PMMA more highly polymerized product, with a very
low content of free monomer. This imoroves
The many applications of PMMA result from resistance to ultraviolet lisht induced
the optical and physical properties ofthe poly- degradation.3
mer. The refractive index is approximately
1.49, which is slightly lower than some glasses, Differences in physical properties of PMMA
while the light transmission capability is 92o/o, result from the two main methods of manufac-
which is slightly higher than glass.' A property ture, casting and extrusion. PMMA may also be
that is frequently exploited is that oftotal inter- injection molded but the material so produced
nal reflection; this allows the transmission of has similar properties to extruded material. A
light around comers as well as light enhanced higher molecular weight polymer (molecular
designs made by"releasing light at cuts made weight : l0o) is used for casting. When cast,
into the material.' the material forms a three-dimensional network
of interlocking polymer chains, which means
The physical properties allow production in a that it has no directional preference in shength.
number of forms including sheet, block, mold- It is a stronger and more "fgryiving" material
ings, rod, tubing and unusual castings. As it is than the e*truded material.3'lshee"ts and
a rigid material with a glass transition tempera- blocks, and unusual shapes or even rods (spe-
ture (Tg) of 104'C many shapes can be cut cifically those over 76 mm in diameter) are
with hand-held or power saws, drills and lasers. cast.' Extruded material is made of lower
Cut edges can be highly polished. A multitude molecular wBight polymers (molecular
of non-planar forms can also be created by heat- weight : l0)). The.shorter polymer chains have
ing due to the wide temperature range over a linear alignment.r') This material allows the
which softening is observed.' In order to production of sharp-edged thermoformed ob-
achieve sharp forms, air and mechanical pres- jects at lower temperatures than the higher mo-
sure as well as vacuum assistance are used. lecular weight resin would allow. However, the
extruded material is more susceptible to failwe
PMMA can vary in additive content, purity, mo- and weaker in the direption perpendicular to the
lecular weight and polymer arrangement. Dyes molecular orientation.' It is also more easily
and pigments, ultraviolet light absorbers and stressed and more su^sgeptible to crazingby sol-
plasticizers can be added to PMMA and it may vents and adhesives.''o Rods, tubes and mold-
be coated to resist surface abrasion. Differing ings are generally extruded, but sheets can also
purities are also available (e.g., specialty medi- be produced by extrusion. While extruded and
cal and inexpensive grades). PMMA can also injection molded material have similar proper-
differ in intemal stress levels as a result of ties and behavior, extruded material is made
manufacturing processes (i.e., stresses cau_sed from monomer-polymer syrups and injection
while setting/curing in sheet or rod form).' molded material is made from rigid pellets of
Stresses may be introduced by techniques, such PMMA.
as cutting, drilling or thermoforming. Due to
the high coefficient of thermal expansion, Cast and extruded PMMA in objects cannot be
stresses may also result from material being easily distinguished by conservators. However,
tightly bound with inadequate allowance for di- extruded rod and tubine has onlv been available
mensional changes.' In addition, adhesives can since the l970ss and eitruded sireet since
introduce stresses in PMMA by solvent absorp- 1978.7 One method of determining if a sheet is
tion or heat production as a result of exother- extruded is to pass a beam of light from a fluo-
mic reactions.' rescent tube through the edge ofa sheet and
view the sheet from the opposite edge while

326
facing the light. Ifthe sheet is extruded the exposed to adhesives or degreasing solvents is
viewer should see parallel ridges, which have perhaps the major concem of the conservator.
resulted from the material passing between While there are published charts in the techni-
calender roll ers during manufacture. Likewise, cal literature of damage resulting from solvents
if a light beam is projected through the material and adhesives on stress-free material, no publi-
at the correct aqgle, the ridges may produce cation could be found sivins materials suitable
lines on a wall.) The only reliable method of for stressed PMMA.I1f,2 O'id"r material is
differentiation between cast or extruded mate- more likely to be damaged when exposed to sol-
rial involves buming and is thus destructive. vents because it absorbs moisture from the envi-
Extruded material drips, bums quietly and ronment (0.4Yoto 0.5o/o at ambient RH), which
smells like acrylate due to the low molecular increases the likelihood of crazing.a Solvent va-
weight and orientation. Cast material ignites pors surrounding objects in enclosed environ-
but does not drip and the flame makes a crack- ments, as well as vapors within a closed form,
lins sound due to the extensive molecular entan- can also cause crazing. A firrther problem is
glJments.5'8 Gel permeation chromqtography that internal stresses are not generally detect-
may also distinguish the two forms.t According able or quantifiable until relieved by crazing
to PMMA manufacturers, polarized light will and/or cracking, when it is too late. Adhesive
show stress but cannot differentiate between and material manufacturers advise annealing in
cast and extruded material.5'8 order to relieve potential stress prior to adher-
ing; numerous charts outlining annealing proce-
Repair Difficulties dure s are available.' Annealing might decrease
the potential of crazing, however, there are
The conservator faces a number of problems many practical and ethical considerations con-
when repairing PMMA objects. The three main ceming annealing artifacts.
difficulties arise from the transparency, the ther-
moplastic properties and the susceptibility of Only two solvents typically used by conserva-
the material to stress-craze or crack when in the tors are safe to use on PMMA; these are water
presence ofsolvents and adhesives. The diffi- and petroleum spirits. Solvents to avoid (be-
culties of making sfong, visually satisfactory cause they dissolve material) have Hildebrand
j oints also confront commercial manufacturers. solubility p?{ameters close to that of PMMA
Generally, monomer-polymer solutions make (18.8 MPa"'),'as shown in Table I.
the best "weld-joins." These contain either a
photocatalyst that induces hardening by ultra-
violet light or a promoter cqntaining a peroxide Table I
to catalyze polymerization.' While materials Hildebrand Solubility Parameters of Solvents
and procedures for repairing transparent glass that Cause Problems for PMMA
are well documented in the conservation litera-
Solvent Hildebrand Likely to
ture, many of these materials are unsuitable due
Solubility Dissolve
to the thermoplastic properties of PMMA. In Parameter PMMA
addition, it is suggested in the plastics and adhe- (MPA r/2)
sives literature that many adhesives would prob- ethvl acetate 18.6 ves
ethvlene dichlonde 20.0 yes
ably damage stressed PMMA. There are few
toluene 18.2 yes
suggestions for adhesives in the consery4tion
litJrature that directly relate to plastics.2'e Fur- acelone 202 yes
methvl ethvl ketone 19.0 ves
ther, many thermosetting adhesives, such as
methanol 29.6 no, craze
epoxies that bond adequately to PMMA, can
ethanol 26.0 no, ctaze
cause damage in application or upon re-
rsopropanol z5.c no, ctaze
moval.'" An additional problem is that it can be
difficult to remove old adhesives; mechanical
Such ketones as acetone and methyl ethyl ke-
removal appears to be the safest option.
tone will also dissolve the polymer. Aliphatic
alcohols and those typically used by conserva-
As mentioned earlier, the susceptibility of
PMMA to stress-crazing or cracking when tors, such as methanol, ethanol, isopropanol

52t
and industrial methylated spirits or IMS (95% Materials
ethanol and methanol) are known to cause
5o/o
t'12
cr-ing.2't Aliphatic hydrocarbon solutions PMMA
will cause crazing after long exposure;t t those The PMMA chosen for this study was clear
with high aromatic contents, such as VM & P ce ll-c,qst (non-directional ) I C I Perspex sheeting.
Naphtha and Stoddard Solvent, will cause dam- 000 ." It was 3 mm in thickness, although
age in a shorter time. there was slight variation in this. The material
was chosen because it has no preferred direc-
Another consideration when choosing an adhe- tion in strength and therefore should give repro-
sive is that the glossy surface of the material is ducible test results. In addition, it is a high
easily disfigured. This is a difficult problem to quality, additive-free material that has been an-
overcome because adhesives that adequately nealed by the manufacturer to remove intemal
bond will also damage the surface. As a result stress. It is reported to contain less than lo/ofree
a surface-protecting masking material, such as monomer. Finally, the surface-protecting mate-
pressure-sensitive or water-based adhesive rial is pressure-bound polyethylene with no
tape, is required when joining with adhesives. adhesive coating that could affect the testing.
There are concems about adhesive residues
left from masking tapes and an investigation is Adhesives
warranted. Finally, since the surface is easily Eleven adhesives, representing five categories,
scratched, it is difficult to remove excess rigid were evaluated. The choice was based on rec-
adhesive. ommendations from PMMA manufacturers,
adhesives and conservation literature. Many of
Cyanoacrylates, Acryloid B-72 in acetone and the adhesives tested did not fit all study aims
Weld On Number 3 (which contains methylene (e.g., reversibility). Two adhesives of each type
chloride and trichloroethylene) are examples of were tested for comparative purposes. The
potentially damaging adhesives that have been adhesives tested are those shown in Table II.
successfully used. Time may prove that these
adhesives cause damage and the risks preclude
their use when conserving historic objects.
Table II
Adhesives Used in the Testing Program

Study Aims Adhesive Cateoories Adhesive Tested

Two-component all acrylic Tensol 70 (lOl, U.K.)
The aims were to furd and evaluate adhesives cold setting Acrifrx 90 (Rohm, U.K )
that would; not cause crazing or cracking of
Two-component epory restn Ablebond 342-1
stressed PMMA; form load-bearing bonds;
HXTAL NYL-1 (U.S.)
make joins with a good appearance. (Join ap-
Singlecomponent ultravrolet Norland 0A65 (U.S.)
pearance is dependent onjoin integrity, refrac- hght curing Norland 0A68 (U.S.)
tive index and transparency.) In addition, the Polymer in solvent Tensol 12 (lol, U.K.)
physical properties of the adhesives would Acrylord B-72
include ease of use, having a lower Tg than 1:1 mixture of Acrylord 8-67
PMMA, and" ideally, being reversible or at and Acrvlord F-10
least detectable by analysis. Cyanoacrylate Loctrte 406 (U.K.)
Loctite 460 (U.K.)
Adhesive bond strength was one of the impor-
tant issues in the evaluation, and thus tensile Water-based adhesives, such as poly(vinyl
properties of the PMMA test material were in- acetate) or ethylene vinyl acetate copolymer
vestigated. They were determined mainly for emulsions or dispersions, were not evaluated
three reasons: to compare data to the technical because they are not transparent. The wider
literature, to review the effect of the sample range in variables of the polymer-in-solvent
preparation procedure and to compare the data adhesive group w:uranted the inclusion of
for the adhesives. three adhesives.

328
Properties of the Adhesives polymer deposition as solvent evaporates and is
absorbed into the PMMA. It contains a solvent
Two-component All-acrylic Cold Setting known to attack PMMA and it will extract sol-
The two-part Tensol 70 and Acrifix 90 are com- uble colorants; it was expected to cause
posed of a mobile solution of monomer in poly- stressed PMMA to craze.''
mer in which a liquid ca(alyst is added to
initiate polymerization. ' ' IC I designed the Ten- The Acryloid acrylic resins (refractive indices
sol adhesives to attack surfaces rapidly, which 1.48 and 1.49) h.ave a history of use and testing
may indicate that the methacrylate monomer in conservation.la Acryloid B-72 wasapplied
acts as a solvent or that a solvent is present. It in a 3: I solution of ethanol and xylene and it
was assumed that Acrifix 90 would behave was anticipated that this would cause stressed
similarly to Tensol 70. A significant point PMMA to craze. Xylene was added to the solu-
about Acrifx 90 is that it has a low refractive tion because Acryloid B-72 would not dissolve
index (1.44), well below that of the other a{!re- in ethanol alone. Of the solvents acetone, tolu-
sives evaluated and the ICI Perspex ( I .49). " ene and xylene, xylene seemed the most appro-
These adhesives were expected to cause priate to add. In a previous study it was found
stressed PMMA to craze. that xylene caused less weight loss than toluene
and it th,e^refore appeared to be a less effective
Two-component Epoxy Resin solvent.'o Also. Horie classes toluene as a sol-
Ablebond 342-1 and HXTAL NYL-I are famil- vent and xylene as a non-solvent for PMMA.l9
iar to most objects and sculpfure conservators;
they have been extensively tested and have re- . The l:1 mixture of Acryloid 8-67 and Acryloid
fraitive indices of 1.56 and l.szrespectively.l4 F-10 was evaluated as an adhesive, even
They contain no solvents but before polymeriza- though it has no documented use as such, be-
tion, epoxy re;ins may act as solvents for cause these resins are soluble in petroleum spir-
some plastics." its (a 'safe' solvent) and can be combined with
eaci", other.20 The refractive indices are 1.49
Single'component LJltraviolet Light Curing and L48 respectively.la Preliminary testing in-
There is no documented use in conservation of dicated that Acryloid 8-67 npetroleum spirit
the ultraviolet light curing adhesives tested in was brittle and bonded weakly to PMMA. It
this study, although two similar adhesives (Nor- was assumed that this behavior was a conse-
land optical adhesives 0A,6l-and 04,63) have quence of the relatively high Tg of 50'C. The
been tested for glass repair.'t Norland adhe- Acryloid F-I0, with a Tg of 20'C, was added to
sives 0A65 and 0A68 were tested because decrease the brittleness and increase adhesion.
they contain no solvents. They have been de- In addition, the manufacturers' literatwe sug-
o gests that Acryloid F-10 is good for solvent-
scribed as urethane-related prepolymers,' and
are presumed to be thermosetting in character, sensitive finishes. The Acryloid 8-67 was used
because of their lack of solubility. in peffoleum spirit with a boiling range of
100'C to l20oC and the Acryloid F-10 was
These adhesives were chosen for investigation used as supplied at 40o/o solids in a 9-:1 solution
after consideration of manufacturers' technical of mineraf spirits and Aromatic 150.20
information. They differ in their flexibility,
strength in bonding to plastic, and re,{active Cyanoacrylate
indices ( 1.52 and I .54 respectively). ' ' Cyanoacrylates are used in industry to join
plastics and have been considered as potential
Polymer in Solvent adhesives by conservators. Loctite 406 is rec-
There are considerable differences among the ommended for joining rubbers and plastics, and
three adhesives in this category. Tensol 12 is a forms strong bonds rapidly. Loctite 460 is a
commercial product, produced specifically for low-bloom, low-odour adhesive of a high mo-
joining PMMA. It is composed of an acrylic lecular weight, which is suitable for bonding
polymer dissolved in dichloromethane and dark shiny plastics that will.pq"critically
methyl methacrylate monomer. Itworks by damaged if crazing occurs.'''"

329
Degreasers samples were not annealed. A hole was drilled
The effect ofdegreasers on adhesive tensile in one end of each sample to hold the load. The
strength was compared using 0.5% Synperonic samples were conditioned in the laboratory for
'N' non-ionic detergent in deionized water and at least a week before testing. Laboratory condi-
'AnalaR' petroleum spirit with a boiling range tions during all the testing were 20oC + l.5oC
of 100'C to 120'C. The petro[epm spirit has an and48o/o t 5% RH.
aromatic content of up to lo%." These very dif-
ferent degreasers were chosen to represent the Test Procedure
two potentially safe solvents that could be used
onPMMA. The apparatus for testing the adhesives was
constructed to accommodate a group of seven
Sample Preparation and Testing samples exposed to one adhesive (Figure l).
Procedures The samples, numbered one to seven, were
placed on the apparatus approximately 125 mm
Four tests were carried out with the intention of apart. Samples numbered one were controls and
measuring the following effects: had no load applied. Samples numbered two
through seven were stressed with specific
o Stress-crazing due to adhesive contact weights: samples two and th'ree with 250 g,
samples four and five with 500 g, and samples
o Tensile properties of the PMMA test material six and seven with 1,000 g.

. Tensile properties ofthe adhesives on

tension-loaded butt-j oined samples

o The effect ofdegreasers on the tensile

properties of the butt-joined samples

Crazingof PMMA Stressed at

Specific Loads
This test procedure was intended to identifo ad-
hesives that damage PMMA by causing crazing
or cracking; it was modelled on two ASTM -
methods, namely, F 7gl-82 andF 484-83.24'2s
PMMA samples were sfressed with a load, and
adhesive was applied to the critically stressed
area. The primary purpose was to determine if
an adhesive would cause a stressed sample of
PMMA to craze, crack or break. The secon-
dary purposes were to determine if a critical Figtre I Stress-crazing testrng apparatus.
amount of stress resulting from the load was re-
quired for an adhesive to cause crazing or crack- A light source was used to highlight crazing;
ing and if an adhesive caused damage to this is because light readily passes through
PMMA at any level of stress. PMMA unless there is a disruption, such as a
gap caused by acraze. The light was placed at
Sample Preparation a 45o angle below the horizontal plane ofthe
sample being examined. Crazing was observed
Seven randomly chosen PMMA test specimens at a 45" angle above the samples (i.e., at 90' to
were prepared for each adhesive. Samples were the light path of the source below) with the
cut to the dimensions of 178 mm long by unaided eye, and with 30 times magnification
25.4 mm wide from the 3 mm test material. (Figure 2).
The rough sawn edges were not smoothed and

330
 samples. It was set with a cross-head distance
of121 mm, a cross-head speed of 2 mm/minute
and a chart speed of 100 mm/minute. A
5O0-kilogram load cell was used for the testing.
Laboratory conditions were the same as above.

Samples, 178 mm by 25.4 rnm, were cut from

the 3 mm sheet of Perspex. (These dimensions
equalled those of the adhesive joined samples.)
Before placing the samples in the cross-head of
the tensile tester, the samples were marked
across the breadth with lines, 25.4 mm from
each end, which left a central region, 127 mm
in length, equal to the cross-head distance.
Figure 2 Testing apparatus showing observation of Tensile strength of buttjoined PMMA and
stress-crazing.
the effects ofdegreaser on the tensile properties
of butt-joined Perspex samples were deter-
Ten minutes after loading the samples, they
mined follolying. in part, ASTM tesr merhod
were examined for crazing and adhesive was
D 3163-73.'o Butt-ioints were investisated be-
applied with a brush or pipette to the surface
cause it was felt that they more closel! resem-
above the fulcrum; the area of application was
bled the types ofrepairs anticipated for objects.
approximately 13.9 mm by 25.4 mm. Follow-
ing this, the samples were examined for
The tensile properties for the cyanoacrylates
craztng. Sample examination times were imme-
were not measured because crazing of samples
diately after adhesive application and at inter-
took place at the lowest applied load. The adhe-
vals of l0 minutes, 30 minutes, I hour, 2 hours,
sives were also spread under the masking tape
4 hours, 8 hours and 24 hours.
protecting the surface, and did not wet the
Most of the sample groups were removed from
roughjoin surfaces.
the loading apparatus 24 hours after adhesive
Twelve randomly selected pairs of Perspex sam-
application. However, Ablebond 342-l and
ples were joined with each adhesive. Samples
HXTAL NYL-I samples remained in place and
were 89 mm by 25.4 mm and thus joined sam-
loaded for 72 hours, to ensure that the adhe-
ples were 2 x 89 mm: 178 mm in length. The
sives were cured. The samples used to evaluate
join surfaces were equally roughened with cir-
Norland 0,4.65 and Norland 0A68 were ex-
posed to ultraviolet light for 30 minutes to cure
cular saw marks; sample sides were jagged due
to the band-saw cuts. As above. the samples
them. This was 10 minutes longer than speci-
were not annealed.
fied by the manufacturer to ensure a complete
cure. (N.8. The lights were removed when ex-
Join surfaces were degreased before adhesive
amining the samples and were kept in place un-
til 30 minutes of exposure had been achieved.) application. For each adhesive tested, six
pairs ofsurfaces to bejoined were degreased
The bulbs were positioned 76 mm above the
sample surfaces to cure the adhesive. The light
with 'AnalaR' petroleum spirit (100'C to
120"C) and the other six pairs with 0.5% Syn-
source consisted of two l5-watt fluorescent
peronic 'N'. The samples degreased with the
black lights.
latter detergent were then rinsed twice with
deionized water. The samples were air-dried
Tensile Strength of
under a dust cover for two davs in ambient
PMMA Test Material conditions.
An Instron Tensile Testing Machine Model
In order to protect the glossy surfaces of
1026 was used for determining the tensile
the Perspex from the adhesives. transparent
strength of the test material and the butt-joined
pressure-sensitive tape was applied adjacent to

JJI
the join surfaces. The tape was burnished to en- Before joining the samples with Acryloid B-72
sure a strong bond. The lengthwise edges of the and the 1:1 mixture of Acryloid 8-61 and
samples were not masked. Two kinds of mask- Acryloid F-10, two dilute layers of the adhe-
ing tapes were compared and little difference in sives were applied to the join surfaces. Solvents
effectiveness was observed. Prior to adhesive were allowed to evaporate for two days be-
application, the degreased, masked joint ends tween applications and before the final concen-
were butted together and secured on one side trated adhesive solution was applied. After final
with a tape hinge placed over the masking tape. application, the solvents from these adhesives
were allowed to evaporate for three weeks
The adhesives, prepared fol lowing manufactur- before removing the tape.
ers' instructions, were applied with a brush or
pipette to the join surfaces. After application, Results and Discussion
samples were placed on a flat surface and ex-
cess adhesive was allowed to remain on the sur- The results indicate that crazing or cracking of
face of the masking tape. When bonds were stressed PMMA after adhesive contact depends
secure according to the manufacturers' instruc- first, on the amount of stress applied and sec-
tions, the masking tapes were removed. The ond, on the composition of the adhesive. None
joined samples were marked with a line ofthe adhesives caused the unloaded control
25.4 mm from each end similarly to the samples to crack even though the samples were
procedure described above. probably weakened with internal stress caused
by the preparation procedure. The relative ex-
The ultraviolet light curing adhesives were tent or density of the crazing or cracking was
treated following the procedure outlined in the not evaluated.
section "Crazing of PMMA Shessed at Specific
Loads." It is interesting that the low-bloom, low-odour
cyanoacrylate, Loctite 460, took longer to

Table III
Crazing of PMMA Stressed at Specific Loads

Adhesive No Load 250 o 500 o 1.000 o

Ablebond 342-1 none none none none
HXTAL NYL-1 none none none none
Norland 0A65 none none none none
Norland 0A68 none none none none
1 1 mrxture ot none none none none
Acrylord 8-67 and
Acrvloid F-10
Acryloid B-72 none none one crazed both crazed
one nour two hours
Tensol 70 none none both crazed both crazed
10 mrn. rmmedratelv
Tensol 12 none both crazed both crazed both crazed
10 mrn. rmmedralelv immecliatelv
Acrifix 90 none both crazed both crazed both ctazeo
one hour rmmedrately rmmedrately, broke
10 mtn.
Loctrte 406 none both crazed both crazed both crazed
one hour 10 mrn. immediately, broke
10 mrn. and 4 hours
Loctrte 460 none both crazed, both crazed both crazed
one hour and four 30 mrn. rmmediately, broke
hours 30 min.

Stressrrazrng of cell-cast PMMA (lcl Perspex) due to adhesve contact at specifred loads. Each adheslve was applied to seven
samples. one unloaded and two at each of the three loads In some Instances the two samples at one load behaved ditferently.

JJZ
cause damage; perhaps it contains a less active Tensile Strength of PMMA
solvent. It was surprising that the Acryloid Test Material
B-72 solution caused so little damage because
ethanol and xylene are known to attack The load at break of the test material was diffi-
PMMA. cult to cgppare with that reported by the manu-
facrurer.' ' There were fwo reasons for this.
In summary, the 500 g load appeared to be the First, it was too difficult to prepare the samples
critical load in the testing. When assessing the in the manner specified by the manufacturer's
data, a trend was evident: the five adhesives standard and it was felt that the sample size,
that did not cause damage at this load did not preparation and test procedure should model
damage samples at the heavier load of I,000 g. the one chosen for testing the adhesives. A sec-
ond consideration was that the samples, which
Samples stessed by the heaviest load, 1,000 g, were probably weakened by stresses induced
uazed when exposed to six adhesives; the same during preparation, were not annealed. The data
adhesives also damaged samples loaded at obtained from 12 randomly chosen samples of
500 g. test material were:

While three of the adhesives weakened the sam- . mean load at break 336 t l0 kg
ples to breaking point, it is interesting to note
that those exposed to Tensol 12 did not break . nominal ultimate tensile strensth
even though samples exposed to this adhesive t
43.88 3.07 MN/#
crazed at the stress level induced by the lowest
test load of 250 g. Nevertheless, five of the ad-
o extension at break 3.97 + 0.1 9 mm
hesives tested did not damage the samples at
the 1,000 g load; these were the two epoxies,
the two ultraviolet light curing adhesives, and
o (UTS, see Appendix l)
the 1 :1 mixture of Acryloid B-67 andAcryloid
r (relative error, see Appendix 2)
F-10. These should be further investigated.

It is unknown whether the differing times of Tensile Strength of Butt-joined PMMA

sample damage in equally loaded pairs resulted
The load at break, the ultimate tensile strength
from different amol.rnts of intemal stress. How-
and extension at break were determined for
ever, when two samples at the same loads be-
samples butt-joined with the nine adhesives
haved differently after adhesive application, a
shown in Table IV. In Figure 3 a plot of these
trend appeared; the second sample to receive ad-
data (the load at break versus the extension at
hesive cracked before the first one. This may
break) is shown. Useful extrapolations, such as
indicate that the second sample was weakened
estimations of bond strength and the impact of
by an increased volume of solvent vapor re-
degreaser, can be made from the data. How-
leased from the adhesive in the test environ-
ever, as with all tensile testing there were large
ment. This observation raises an important
deviations. This is due to the small number of
consideration when using adhesives in the vi-
samples being tested; typically six per group.
cinity of PMMA objects: objects can become
stressed by adhesive vapors and crazing or
cracking could occur without contact.
All samples broke at the adhesive join at a
load well below the tensile strength determined
for the Perspex (336 tl0 kg). Tensol 70, Acri-
Evaluation of Test
fx 90 and Tensol 12 were strong, but despite
The test methods used to determine which being specifically manufactured for joining
adhesives caused craztng of stressed PMMA PMMA, they were not the strongest according
were useful, but the sample size required is to these tests. The samples joined with the
large. Ifsamples had been annealed or aged epoxy resins, Ablebond 342-l and HXTAL
either naturally or artificially, different results NYL-I broke at similarly high loads but with
might have occrured. smaller deviations from the mean. The samples

JJJ
 Table IV
Tensile Strength of Butt-joined PMMA and Degreaser Impact

Adhesive Load at Break (kg) Ultimate Tensile Strength Extension at Break (mm)

Petroleum Synperonic N Petroleum Synperonic N Petroleum Synperonic N

Snirit Soirit SDirit
Ablebond 342-1 143 t16 101 t7 18.74 x2.44 13.15 I 1.1 8 1.48!017 1.0s I 0.09
f5t '6)
HXTAL 82x17 65 t 15 10.76 !2.26 8.44 ! 2.11 0.81 j 0.19 0.62 I 0.14
NYL.1 (61

Norland 42 !7 5.48 r 1.70 5.45 10.98 0.23 1 0.2 0.22 + O.O2

oA65 A\ (6)
Nodand 51 15 44r5 6 62 10.73 5.77 !0.75 0.48 1 0 07 0.42 10.05
oA68 (5) (6)
1 1 Mrxture of 15 !2 1312 2!034 1.73 I 0.26 0.21 10.01 0.21 !0.02
Acryloid B-67 (6) not dry (6) not dry
and Acrvloid F-10
Acrylord 1216 913 1 54 t 0.85 1.1 1 r 0.38 0.21 + 0.05 0.1910.02
B-72 (6) not drv (6) not dry
Tensol 68r15 69 t18 887 !2.04 8.98 12.43 0.74 !0.21 0.70 + 0.18
70 {6} {6)
Tensol 80126 86 111 10 39 1 3.53 1120 11.57 0.84 !0.27 0.89 r 0.12
12 I5l {6)
Acnfrx 98125 101 !27 1277 !3 45 13 17 r 3.69 'l .02 x 0.29 1.04 1 0.31
90 f5) r6)

Load at break In kg, ultimate tensrle strength (nomrnal) or UTS In MN/m2- and extension at break In millrmeters of butt.Jotned PMMA
samples; the effect of two degreasers used before adhesive application, petroleum sprnt and 0.5% Synperontc'N' in deionized
water rs compared. Mean of data group rs presented, plus or minus sample standard devration.'' The relative error"'" rs presented
with the UTS. The quantrty of samples used for each measurement rs grven In parentheses

-See Appendrx 1
.'See Appendrx 2

degreased with petroleum spirit and joined with It was anticipated that the Acryloid B-72 would
Ablebond 342-l had, a considerably higher form stronger bonds than the l:1 mixture of
bond strength than any other adhesives. Acryloid 8-61 and Acryloid F-10 because the
3: I ethanol and xylene solution was expected
The adhesives not produced by PMMA manu- to attack the PMMA more than the petroleum
facturers broke in four fairly distinct ranges. spirit mixture. Solvent retention may have been
The epoxies formed the strongest bonds break- responsible for masking this difference.
ing at two different load ranges; of these Able-
bond 342-1 formed considerably stronger The extension at break ofthe two epoxies,
bonds. The ultraviolet light curing adhesives Ablebond 342-1 and HXTAL NYL-I, varied
broke at loads only slightly greater than half significantly depending on the degreaser. An
those of the epoxies. Though they broke at simi- explanation of this probably relates to the im-
lar weights, Norland 0A68 broke at slightly pact of degreaser as will be discussed. Of the
higher loads than Norland 0,{65; this was not ultraviolet light curing adhesives, Norland
surprising as the manufacturer's literature indi- 04.65 extended approximately half that of
cated that 04,68 formed stronger bonds to plas- Norland 04.68. This difference was not antici-
tic. The acrylic polymers in solvent, Acryloid pated because Norland 0A65 is manufactured
B-72 and the I : I mixture of Acryloid 8-67 and to be more flexible than Norland 0A68. Per-
Acryloid F-10 formed very weak bonds. This haps Norland 0A68 was not fully polymerized.
was anticipated but the load at break was prob-
ably lower because the solvents had not evapo- However, the extension at break is mainly a
rated from the adhesives, even after three function of the load at break as shown in
weeks; these bonds stretched while the other Figure 3.
adhesives broke cleanly.

334
Degreasers
Exten6lon vf,€us l6d at break
1.5
With the exception of the
epoxies, the two degreasers 13
appeared to have little im- 12
pact on the adhesive bonds. 11
The epoxy-joined samples a

degreased with pefroleum 09

n o8
spirit broke at significantly
o7
higher test loads and ex-
06
tended more than the deter- 6
05
gent-degreased samples; ri o4
this suggests that the deter-
gent solution had some im- o2 E ttrD
pact on join integrity. This ol
might be because detergent o

remained on the surface 50 80

@d rt brs.k , Crg)
forming a barrier, or that the
detergent attracted moisture
to the surface, thereby reduc- Figure 3 Extension at break as a/unction ofload at break.
ing bond strength. It should
also be considered that pe-
troleum spirit may be a more effective de- caused stressed PMMA to qaze. Five of
greaser. (lt may be significant that the epoxies the adhesives tested did not damage PMMA
are the only adhesives that exhibited a signifi- stressed at any of the test loads; these were the
cant difference in bond strength resulting from two epoxies, the two ultraviolet light curing ad-
the degreaser.) This finding reiterates the cau- hesives and the l:l mixture of Acryloid 8-67
tions that have been sussested about the use of and Acryloid F-I0. Of these, it appears that
detergents on plastics.
lfi 8 three may produce satisfactory joins with differ-
ing requirements based on their refractive indi-
Evaluation of Test ces ( 1 .52 to I .49) and the test data; these are
HXTAL NYL-1, Norland 0465 and the l:1
The tensile testing procedure could be modified mixture of Acryloid 8-67 and Acryloid F-10 in
or a much larger number of samples could be petroleum spirits. It is imporiant to match re-
tested in hopes of reducing the high enors in fractive indices when joining fransparent mate-
the data. In addition, it should be noted that as a rials, and Ablebond 342-l and Norland 0A.68
result ofthejoin surfaces being slightly irregu- were eliminated for clear material because they
lar due to the circular saw marks, stronger have refractive indices that differ significantly
bonds may have been created than would have from PMMA.
occured on smooth surfaces; this is because the
adhesives may have been physically bound to Differing applications are possible with the
the surfaces. However, breaks on objects are three favored adhesives due to the distinct
generally not planar or devoid of surface ranges of load at break observed, as well as
inegularity. their other properties. For example, the ther-
moset HXTAL NYL-l was similar in strength
Conclusion to the commercially produced adhesives for
PMMA and could be used for load-bearing
Though it is anticipated that PMMA found in bonds. However, although HXTAL NYL-1 is
artifacts and objects may not behave the same detectable it could only be removed mechani-
as new cell-cast PMMA, useful extrapolations cally. Norland 0A65, also presumed to be a
can be made from this test data. The PMMA thermosetting resin, would be useful for
manufacturer-produced adhesives should not be quickly "tacking" difficult joins. The testing
used to conserve historic objects because they also indicated that it is rubbery and new

JJ)
adhesive could be pulled off the PMMA. More Sculpture Conservation and Joyce Townsend,
investigation is warranted conceming this prop- Conservation Scientist, for continuous encour-
erly and the aging behavior ofthis adhesive. agement and assistance in all aspects of the
study. I would also like to thank The Henry
The 1:1 mixture of Acryloid 8-67 and Acryloid Moore Foundation for funding the post in
F-10 made very weak bonds that may be suit- which I carried out this research. A large part
able for repairing delicate objects, such as jew- of this research would not have been possible
elry, sculptors' maquettes and architectural without the assistance of the late Gerry Hedley,
models. This was the only adhesive tested that who showed me how to use the tensile tester
appears to be removable by solvents that will and assisted in preliminary organization of the
not attack PMMA. While it is known that Acry- data, and The Courtauld Institute for allowing
loid 8-67 is often diffrcult to remove with pe- me to use their tensile testing machine. In addi-
troleum spirits alone, requiring the addition of tion,I would like to thank Dr. Jonathan Ashley-
polar solvents or even heat, it is hoped that the Smith, Deparfinent Head of Conservation at the
addition of Acryloid F-10 will increase Victoria and Albert Museum for allowing me
solubility. time to work on the study while I was em-
ployed, and Graham Martin, Head of the Scien-
It must be emphasized that the results of these tific Section, and Boris Pretzel, Conservation
tests are not conclusive because the PMMA Scientist, for their assistance in using spread
tested is new material of a high quality with a sheets for determining the statistical data of this
high molecular weight and a homogeneous study. Finally,I would like to thank The Gabo
structure of extensive molecular entanglements. Trust and The Samuel H. Kress Foundation for
With the exception of the stresses caused by the the joint sponsorship that allowed me to present
sample preparation and testing procedures this my research at this conference.
material is sffong and resistant. It is expected
that old, extruded material, annealed material,
or material cast in an artist's studio may behave Appendix 1
differently. PMMA that is not totally polymer-
ized, contains additives, such as colorants or UTS(N) or ultimate tensile strength
plasticizers, or has a coating, may also act dif- (nominal) : F/A where:
ferently when exposed to these adhesives. F : the mean load at break of the group of
samples tested
In summary, the three adhesives that met many A: the initial cross-sectional area ofthe
of the five criteria used to evaluate the adhe- samples determined by multiplying the
sives in this study were HXTAL NYL-I, Nor- mean depth in millimeters times the mean
land 0,4'65 and a 1:l mixture of Acryloid 8-67 width in millimeters of l0 randomly
and Acryloid F-10. These adhesives did not selected samples.
craze or crack stressed PMMA, they formed
load-bearing bonds at different ranges and they
have refractive indices closer to PMMA than Appendix 2
the two other adhesives that did not cause craz-
ing. ln addition, these adhesives have desirable The relative error of the UTS(N) was
physical properties including a lower Tg than determi4ed with the_ following formu la :
PMMA. [(Sr-/L)2 + (Sw/w)2 + (so/o')2]l/2 where:
L : the mean load at break in kilograms of the
Acknowledgement sample group
W : the mean width in millimeters of
Research was carried out in the Sculpture l0 randomly selected samples
Conservation Department of the Tate Gallery. D: the mean depth in millimeters of
I would like to thank the following people l0 randomly selected samples
from the Tate Gallery: Jackie Heuman, Senior S : sample standard deviation of the sample
Sculpture Conservator, Derek Pullen, Head of group results

336
Suppliers of Materials R6sum6

Ablebond 342-1, HXT AL NYL- l, Synperonic L'ivuluation dc 11 adhdsifs en vue de leur

'N' non-ionic detergent utilisation pour Ia riparation d'objas a de
Archival Aids Ltd., sculptures en poly(mdthacrylate de mdthyle)
P.O. Box 5, Spondon, Derby,
DE27BP, England. Les poly(mdthacrylate de methyle) Perspex ou
Plexiglas se retrouvent tant dans des rzuvres is-
Acrifix 90 sues des beaux-arts ou du design que dans des ob-
jets qui timoignent de I'histoire sociale. D'abord
(Rohm Plastics) Righton Ltd.,
utilist comme revdkment d'avion, durant la se-
Unit 4, Bush Industrial Estate,
conde guerre mondiale, ce produit convient par-
Station Road,
ticulidrement bien, du fait de ses propridtts
London, Nl9 5UN. optiques et physiques, d I'exdcution de sculptures
et d la fabication de meubles, de btjoux, de v€te-
Acryloid (Paraloid) B-72,8-67 and F-10 ments et d'objets industiels. Il est ntanmoins fort
(Rohm and Haas Ltd.) U.K. distributor, dfficile de trouver un adhbsif pour rdparer un ob-
Chemicryl Ltd., jet en PMMA, car ce matdriau, souvent transpa-
Hockerill Street, rent, devient plus sensible d lafissuration et au
Bishop's Stortford, Hertfordshire, craquellement sous contrainte lorsqu'il est mis
CM232DW, England. en prdsence d'adhisifs ou de solvants de dtgrais-
sage. De plus, tant lafeuille de ce matdriau (au
Loctite 406 and 460 moment de sa fabrication) que I'objet lui-m€me
(Loctite U.K.), Loctite Holdings Ltd., (au moment de son exdcation) auront sans doute
Watchmead, Welwyn Garden City, dijd subi de telles contraintes, ce qui augmentera
Hertfordshire, d'autant le risque defissuration.
AL7 lJB, England.
Des essais ont ttd effectutis pour thtaluer les pro-
pidtds de I I adhdsifs relevant de I'une ou I'autre
Norland Optical Adhesive 0465 and 0,{68
des cinq catigoies suivantes :
Tech Optics Ltd.,
Unit 6, Cala lndustrial Estate, c la prise dfrcid d deux iltments tout acrylique :
Tannery Road, Tensol 70 et Acryfix 90;
Tonbridge, Kent,
TN9 lRF, England. o la prise d dettx 6l6ments de rdsine
(In U.S.A. from Norland Products Inc.,
d froid
tpoxy : Ablebond 342-1 et HXTAL NYL-I;
695 Joyce Kilmer Avenue,
New Brunswick, N.J. 08902.) o la prise d un seul tltment induite d I'ultra-
violet : Norland 0,4'65 et Norland 0468;
Perspex, Clear cell-cast sheet, 000, Tensol 70
and Tensol l2 . un polymAre dans un solvant : Tensol I2,
(ICI) Amari Plastics, Acryloid B-72 et mtlange d'une partie
2 Cumberland Avenue,
d'Acryloid 8-67 pour une partie
Park Royal, London,
d'Acryloid F-10;
NWl0 7RL, England.
o les cyanoacrylates : Loctite 406 et Loctite 460:
Petroleum Spirit 100'C to 120'C 'AnalaR'
BDH Laboratory Chemicals, Lafissuration sous contrainte rdsultant d'un
Freshwater Road, Dagenham, Essex, contact avec des adhesifs a etd etudi1e en ayant
RM8 lRF, England. recours aux essais F 791-82 et F 484-83 dc de
I'Ameican Societyfor Testing and Materials
(ASTM). La rdsistance d la tension des extrtmitds

JJI
rdpartes a 6tt mesurde. Et les ffits des agents 8. REhm, personal communication with staff of
de ddgraissage d dttergent (0,5 pour 100 de Syn- Techrucal Department. nfinm Ltd., Bradbourn
peronic <N> non ionique) et d solvant (essence de Drive, Tilbank, Milton Keynes, MK7 SAU,
pdtrole dont I'intentalle de distillation se situe en- 1990 and 1991.
tre I00 et I 20"C) sur la rdsistance d la tension
des adhdsifs ont dtd compards. Les dchantillons 9. Shields, J., Adhesives Handbook(London:
utilisds ata fins de ces essais provenaient d'un Butterworths, 1984).
panneau de perspex de 3 mm transparent, fabri-
qu,! par les Impeial Chemical Industries (ICI) et
10. Blank. S.. "An Introduction to Plastics and
mould en cuve (non orienti).
Rubbers in Collections," Studies in Conserva-
L'HXTAL NYL-L,le Norland 0A65 et le mtlange tion, vol. 35, 1990, pp. 53-63.
d'une partie d'Acryloid 8-67 pour une partie
d'Acryloid F-10 pourraient €tre utilisds pour ll. 'Perspex' Cell-Cast Acrylic Sheet:
rtparer certains genres de dommages. Properties qnd Fabrication Techniques, PX
l27,4th edn. (ICI, P.O. Box 34, Darwen,
References Lancashire, BB3 IQB: Imperial Chemical
Industries PLC, 1989).
| . 'Perspex' Cast Aoltlic Sheet for Glazing.
PX TD236 (Welwyn Garden City, Herfford- 12. 'Tensol' Cements for 'Perspex' Acrylic
shire: Imperial Chemical Industries PLC, 1986). Sheet,6thedn. (ICI, P.O. Box 34, Darwen,
Lancashire, BB3 lQB: Imperial Chemical
2. Brydson, 1.A., Plastics Materials (London: Industries PLC, undated).
Butterworth Scientific, London, 1982).
13. Technical Information, Acrifix 90,
3. Cousens, D., "'Perspex' a Technical Appre- 2-Component Polymerization Adhesive (Rdhm
ciation," a one-day course, Imperial Chemical GMBH Chemische Fabrik. Postfach 4242.
Industries PLC, Chemicals and Polymers Kirschenallee, D-6100 Darmstadt l: R6hm
Group, Welwyn Garden City, Hertfordshire, Plastics,1989).
I 988.
14. Tennent, N.H. and J.H. Townsend, "The
4. Normalizing and Stress-Relieving of Significance of the Refractive Index of
'Perspex' Cell-Cast Acrylic Sheet, PX TD230, Adhesives for Glass Repair," IIC Adhesives
9th edn. (ICI, Chemicals and Polymers Group, and Consolidanfs, Paris, 1984, pp. 205-212.
Welwyn Garden City, Hertfordshire: Imperial
Chemical Industries PLC, undated). 15. Robson, M., "Clear, Colourless Adhesives
for Glass," Conservation News, vol. 30, 1986,
5. Lombard, M. and others. Imperial Chemical pp. l4-16.
Industries PLC, personal communications
with staff of Acrylics, Technical Department, 16. Robson, M., "A Comparative Study of a
P.O. Box 34, Darwen, Lancaster, BB3 lQB, Range of Glass Adhesives," Poster Session
1990 and 1991. Paper in: Recent Advances in the Conseruation
and Analysis of Artifacts, University of Lon-
6. Chemical Resistance in General Use: don, Institute ofArchaeology, London, 1987.
'Plexiglas' GS and 'Plexiglas '-Y7(Riihm
GMBH Chemische Fabrik. P oslfach 4242 - 17. Norland UI/ Curing Adhesives: Technical
Kirschenallee. D-6100. Darmstadt l: Riihm Dala (Norland, 695 Joyce Kilmer Ave., New
Plastics,1986). Brunswick, N.J. 08902.: Norland Products Inc.,
undated).
7. Perspex: The First Fifiy Years 1934-84
(P.O. Box 34, Darwen, Lancashire, BB3 IQB:
Imperial Chemical Industries PLC, 1984).

338
18. Sale, D. Jr., "The Effect of Solvents on 24. "Standard Practice for Stress Crazingof
Four Plastics Found in Museum Collections," Transparent Plastics," Standard No. F 79 1-82,
in: Modern Organic Materzals, Scottish Plastics (Philadelphia: American Society For
Society for Conservation & Restoration, Testing and Materials (ASTM), 1989).
Edinburgh, 1988, pp. 105-l 14.
25. "Standard Test Method for Stress Crazng
I 9. Horie, C.Y., Materials.for Conservation of Acrylic Plastics in Contact with Liquid
(London: Butterworths, I 987). or Semi-liquid Compounds," Standard
No. F 484-83, Plastics (Philadelphia: American
20. Acryloid Thermoplastic Acrylic Ester Society for Testing and Materials, (ASTM),
R es i n s fo r I n d u s tr i a I F i n i s h in g (Phil adelphia: 1989).
ROhm and Haas, 1987).
26. "Standard Test Method for Determining the
21. 'Loctite' Instant Adhesives (Watchmead Strength of Adhesively Bonded Rigid Plastic
Welwyn Garden City, Herrfordshire Lap-Shear Joints in Shear by Tension Load-
AL7 lJ: LoctiteU.K., 1988). ing," Standard No. D 3263-73, Plastics (Phila-
delphia: American Society for Testing and
22. Loctite, personal communication with staff Materials, (ASTM), 1989).
of Technical Departrnent, Loctite U.K., Watch-
mead, Welwyn Garden City, Hertfordshire 27. Quattro Pro, Spreadsheet Program with
AL7 lJB, 1990. Statistical Analysis Functions, Boland, Scotts
Valley, Califomia, 95067-000 l.
23. BDH, personal communication with staff
of Technical Departrnent, BDH Laboratory 28. Pretzel, B., personal communication,
Chemicals, Freshwater Rd., Dagenham, Essex, Research Scientist, Victoria and Albert
RM8 IRF. Museum, 1991.

339
Labelling Plastic Artefacts

Julia Fenn

Royal Ontario Museum

Toronto, Ontario
Canada

Abstract such as vulcanite, poly(vinyl chloride), cellulose

acetate and cellulose nitrate are capable of
The high isk of using inks, paints or varnishes bleaching or bleeding inl<s and attacking paint
when applying registration numbers to plastic and varnish media so that in a surprisingly short
artefocts is becoming increasingly apparent. Plas' time the numbers become illegible or crumble
ticizers, solvents, pigments and dyes can interact away. This type of attack affects not only tradi-
ineversibly with the plastic substrate, accelerat- tional varnish and ink-marking systems but also
ing stress-cracking, deformation, discolouration attacks threads, wires and other materials used
and even complete disintegration. The long-term for tie-on labels or exhibition labels, which
nature of much of the damage, which may take are adjacent to unstable plastics in display
months to become visible, and the lack of informa- cc6es.
tion about conditions under which molecular
complexes form, mean that damaging combina- This paper will detailfurther experiments with
tions cannot be identified by tests on an incon- combinations of different inks and polymers that
spicuous area ofthe artefact. As a result, thefirst appear to minimize the risk of attack on the arte-
step towards recording a plastic artefact for fact and show greater resistance to degradation
posterity could be the one that initiates its products from the artefacts.
destruction.
Introduction
Each type of plastic is susceptible to damage
from a dffirent range of materials. So, theoreti- When an artefact is acquired by a museum vir-
cally, damage could be avoided by using custom- tually the first treatment it receives is the appli-
ized labelling solutions for each type ofplastic. cation of a registration number. Unfortunately
However, this depends on being able to identify methods of labelling that are satisfactory for
the plastic accurately, which is very dfficult wilh- most museum materials are not safe for plastic
out expensive analytical equipment. Even assum- artefacts. Solvents, dyes, plasticizers and pig-
ing a reliable identification system, there are still
ments in the labelling materials can cause irre-
variables caused by manufactuing techniques,
versible degradation, which may eventually
additives, contaminants and aging; several
destroy the artefact firnctionally and aestheti-
polymers are lcnown to alter their solubility
parameters wilh age. cally. So this veryfrst step towards identifiing
a plastic artefact as a permanent recordfor

Another aspect ofthis interaction that is not posterity may be the very one that initiates its
always considered is the effect ofthe plastic arte- destruction.
facts on the labels. Artefacts made from mateials

341
Common Labelling Techniques: infi nitesimal amounts of absorbed plasticizers,
Effects of Solvents and Plasticizers which form molecular complexes with applied
solvents. For example, plastics that have been
The most common museum method of label- stored in contact with other polymers, or mod-
ling is to apply the regishation number to a ern pieces that have been kept in their original
small insulating patch of lacquer on the arte- "shrink-wrap" packaging, may have absorbed
fact, followed by another layer oflacquer to plasticizers at the contact area, which can cause
protect the ink from mechanical damage. Unfor- local alteration of their solubility. Testing the la-
tunately it is not always possible to predict belling solution on an inconspicuous area ofthe
how, or when, a vamish will affect a plastic. object will be inconclusive if the test area is not
Dissolution is only one of several types of comparable with the rest of the artefact. Other
irreversible damage that can be caused by con- possible causes ofdifferential solubilif are
tact with organic solvents or exposure to their exposure to volatile corrosion inhibitors, moth-
vapours. Even if the lacquer does not immedi- balls, cleaning fluids or adhesives.
ately bond irreversibly to the plastic substrate,
the artefact may gradually swell and distort. There is a cellulose nitrate comb in our teach-
This type of damage is most often seen in flex- ing collection that had an old stain caused by
ible plastics, such as poly(vinyl chloride) belts a pressure-sensitive price tag. When an attempt
and raincoats, but we also have examples of was made to clean off the adhesive stain using
rigid cellulose nitrate knife handles that were isopropyl alcohol, which had already been
labelled with an acrylic varnish in ethanol. This tested without visible damage on other parts
particular cellulose nitrate was not directly sol- of the comb, the plastic instantly softened and
uble in ethanol, but within a year the surface un- blanched around the stain. Pressure-sensitive
der and adjacent to the registration number had labels tend to be even more harmful to plastics
rippled (Figure l). The vamish layer was easily than they are to other museum materials. Cellu-
removed mechanically because it had not fused lose nitrates sometimes develop quite colourful
to the substrate, but the distortion in the cellu- stains especially from rubber-based pressure-
lose nitrate was permanent. Volatile solvents sensitive materials (Figure 2). Even polyole-
have been found to be retained for months, fins, which resist most adhesives, can absorb
even years in an apparently dry varnish; it is pressure-sensitive adhesives so deeply that only
sheer luck if they don't eventually interact the most drastic cleaning agents will (parlially)
with some part of the polymer substrate. remove them. The effects can be seen on poly-
ethylene kitchenware where the commercial
labels have been left in place too long.

q$$.$i-$

Figure I A 1950s'cellulose nitrate knife handle

shows distortion that occured within three months
oJ applying the registration number. Figure 2 French ivory pocket knife with orange-
coloured stain from contact with rubber-based
Some cellulose nitrates are not soluble in either adhesive.
ethanol or etheq but dissolve readily in combi-
nations of the two solvents. In the same way Light damage is another cornrnon source of
solubility ofthe plastic can be affected by changes in solvent sensitivity. Ebonite (highly

)+L
lulcanized rubber) is resistant to most solvents, Crazing or stress-cracking is another phe-
but after photo-oxidation its surface can be dis- nomenon associated with exposure to solvent
rupted by contact with any liquid. Casein is vapours, especially in rigid plastics such as
normally swollen and cracked by prolonged polystyrene, polycarbonate or poly(methyl
contact with water but its moisture sensitivity methacrylate). Absorption of high, local con-
can be greatly increased if the artefact has centrations of solvent vapours apparently en-
been overexposed to ultraviolet light (Figures 3 courages compression cracking and the release
and 4). of strains already incorporated into the plastic
during manufacture. The damage often does not
become visible until most of the absorbed sol-
vent has dissipated, days or even months later.
We were able to demonstrate this delayed reac-
tion by using the traditional labelling method
on modern crystal polystyrene mugs (Fig-
ures 5 and 6). A small patch of Acryloid B72 in
Camsco Super Hi-Flash (a commercial mixture
of mineral spirits, which has just enough aro-
matic content to dissolve the acrylic varnish
but not dissolve the polystyrene) was painted
below the rim. Within a week small cloudy
patches marred the translucency of the polysty-
; rene; these appeared to be enhanced polishing
Figure 3 A 1950s' casein buckle shows surJhce marks. Shortly afterwards, small cracks ap-
damage caused by a paper label experimentally peared around the rim and the base. The crack-
wetted with water adjusted to the same pH as ing has continued, and now nearly three years
Acryloid 33. later the polystyrene is virtually opaque.

The stress-cracking phenomenon is quite

complex and is not fully understood as the fol-
lowing explanation illustrates. Solvents can
apparently initiate stress-cracking in several
ways: by encouraging the migration of plasticiz-
ers and antioxidants, or by removing or intro-
ducing moisture. Evaporation chill may also
be a factor: some plastics such as poly(methyl
methacrylate) expand and contract to a greater
degree than metals; at the same time they have
poor thermal conductivif so the rapid lowering
of the temperature in one area by the evapora-
tion of very volatile solvents could conceivably
Figure 4 Same casein buckle shows detail o/
aggravate existing stress sufficiently to initiate
cracking within the wetted area.
craztng. However, not all plastics are affected
by the same solvents regardless of their volatil-
It was disconcerting to observe in our experi- ity, and it is doubtful if dimensional changes
mental samples of coloured casein and urea- and loss of flexibility can be the sole cause of
formaldehyde buttons that increased solubility stress-cracking. A brittle plastic, such as poly-
often occurred well before there was any notice- styrene, may be remarkably resistant to dimen-
able fading of the dyes or loss of surface gloss, sional changes caused by extreme temperature
indicating that it is not always possible to judge changes and nevertheless be extremely suscepti-
probable sensitivity by appearance. Light dam- ble to stress-cracking when exposed to vapours
age also appears to increase the solvent and from hydrocarbon solvents. Manufacturers use
moisrure sensitivity of some narural plastics. the weight changes in plastics caused by expo-
such as oriental lacquer''' and amber.' sure to solvents as one means of predicting

J+J
 strain is relieved, by which time the artefact
may be ruined.

Unforfunately, there seems to be no one solvent

that can be safely used to make a vamish suit-
able for all plastics. Even aliphatic hydrocar-
bons, which seem to be promising because of
their low solubility, cannot be recommended
because they: 1) cause swelling distortion of
poorly rulcanized ebonite, 2) leach plasticizers
used in flexible vinyl-type plastics, and3) craze
polystyrene, which is one of the most corruron
plastics. Alcohols are potentially harmful to
polycarbonate, poly(methyl methacrylate) and
some cellulose nitrate, cellulose acetate and
polystlrene resins.

Most plastics are not damaged by a short expo-

sure to cold water of neutral pH, but there are
Figure 5 A I9B4 crystal polystyrene mug just
exceptions. One example is any plastic with a
after appl ic ati on of re gis tration numb er. high proportion of poorly combined hygro-
scopic filler, such as wood flour. Early "compo-
sition" plastics, such as Bois Durci and many
moulded shellacs, fall into this category. De-
grading cellulose nitrates and acetates as well
as light-damaged plastics, such as ebonite or
urea-formaldehyde, are also potentially vulner-
able to staining, distortion or accelerated degra-
dation if wetted. Variations in the pH usually
aggravate the damage: cellulose nitrate and ace-
tate are susceptible to both acid and alkaline hy-
drolysis and casein and other "protein" plastics
are particularly sensitive to alkalinity. In spite
of these problems, water is often more satisfac-
tory than organic solvents because its effects
are more predictable.

Other Methods of Attaching Labels

Figure 6 Same crystal polystyrene mug
three months later shows stress-cracking.
There are several methods of numbering plastic
artefacts that avoid the use ofsolvents, for ex-
which solvents are likely to attack the polymer. ample, by labelling the container and not the ar-
However some plastics, such as poly(methyl tefact. However, in museums where the artefact
methacrylate), when exposed to certain sol-
is so often separated from its storage container
vents, such as isopropyl alcohol, show no
while on display or loan, this may not be satis-
weight changes and yet have been found to suf- factory. Moreover several plastics, such as cel-
fer subsequent loss of impact resistance and lulose acetate, degenerate more rapidly when
crazrng. This unpredictability suggests that it
kept enclosed; storage in vented cabinets is
is never wise to expose a plastic to any solvent preferable for such polymers. Possibly a more
vapour. Testing a solution on an artefact, espe- sati sfactory method for exceptionally important
cially ifseveral solvents are investigated in se- pieces might be to label individually custom-
ries, may itself initiate stress-cracking. Once ized supports that can remain with the artefact
craztng has begun it usually continues until the at all times.

344
Methods of attaching registration numbers me- Inks and Paints:
chanically by sewing or tying numbered tags Damage Caused by Dyes and Pigments
are also in common use, especially on garments
or textile accessories made from sheet plastics, Dyes and pigments in direct contact with pias-
such as poly(ethylene terepthalate) (Mylar), tics can act as catalysts or photosensitizers- or
plasticized poly(vinyl chlorides), and cellulose cause irreversible staining. Plastic manufactur-
acetate. Some of these sheet plastics are notch- ers found that nigrosine, an azine dye still used
sensitive so tears can easily be initiated during in fabric marking pens and some newsprint, ac-
subsequent handling from the points where the celerates the evolution of nitrogen dioxide from
needle has punctured the sheet. Ifit is essential cellulose niffate.5 Even melamte-formalde-
to attach a label this way, the danger of tearing hyde, which is used as the surfacing material in
can be reduced by using a very sharp leather counter-tops because of its excellent resistance
punch to create tiny circular holes without to chemical attack, can be damaged by inks.
ragged edges or by using a heated needle so The Royal Ontario Museum recently acquired
the plastic at the edge of the needle hole melts. a pink melamine bowl on which the ink stamp
However, both methods are risky and irre- from a long-departed paper label had left a col-
versible. The type ofthread also has to be taken ourless etched copy ofitselfin the surface of
into consideration; silk, wools, and cellulose the bowl (Figure 7).
threads are rotted by acid degradation products
and certain plasticizers, nylon monofilament
has a sharp cutting edge in addition to poor
chemical resistance, and metal wires can cata-
lyse polymer breakdown and leave ineradicable
stains from corrosion products.

One private collector has solved the problem of

attaching labels to artefacts such as bowls by
drilling a tiny hole in the artefact so that a tag
can be attached. It is doubtful if conservators
could bring themselves to do this, but in fact
it is probably less damaging than some other
methods considered quite acceptable. Neverthe-
less such treatment will do nothing for the Figure 7 Etched stamp mark in a I960s '
m e I am in e -fo rm a I de hy de bow l.
appearance of translucent plastics, such as
poly(methyl methacrylate) or polystyrene,
whose light-transmitting properties accentuate
Old numbering systems where inks or paints
were applied directly to the surface of the ob-
even minor flaws. It will also decrease the me-
ject are no longer in favour because ofcorro-
chanical strength of notch-sensitive or highly
stressed plastics. The development of tiny sive acid inks and the diffrculty in removing the
number. But, on plastic artefacts, a varnish
cracks around holes can sometimes be seen in
layer creates a larger area ofsolvent contact
museum vitrines made from polycarbonate or
poly(methyl methacrylate) even after they have without offering any improved protection or re-
been annealed to reduce stress. Another related
versibility so it seems an unnecessary complica-
technique is the "scrimshaw" method where the
tion. Providing that the inks or paints are very
identification number is scratched into the arte- carefully chosen for compatibility, applying the
fact's surface and may have ink or other colour- numbering medium directly to the surface of
the artefact (rather than to a plastic insulating
ing agents rubbed in to enhance visibility. The
same objections apply to this method, although
layer) offers several advantages:
ifthe scratches are very shallow they can at
o a much smaller area of contact
least be mechanically removed by light
abrasion, but direct contact with inks can
be harmful. o much less solvent

345
o better adhesion: the discontinuous lines pigments were evaluated because either one
forming the numbers often have much better colour or the other is legible on artefacts of
resistance to flexing and dimensional changes any colour. Pure carbon black does not react
than solid patches of relatively rigid varnish harmfully with any plastic and has excellent
chemical resistance, but the presence ofcon-
o no unsightly patches of discoloured vamish taminants depends on the source and type of
combustion. Lamp black is often greasy and
o easier mechanical removal ivory/bone black may contain imperfectly
bumed organic material. Alkaline agents in
these types ofcarbon pigments have been asso-
There are basically two types of ink that will
adhere to plastics: solvent dispersed dyes,
ciated with the breakdown of cellulose ni-
trates.5 The most commonly used carbon black
which adhere by dissolving slightly into the
surface of the plastic, for example, lieezer pen
in commercial hobby paints seems to be fur-
nace black produced by high temperature ther-
inks, and particulate pigments suspended in a
resin medium, which bond to the plastic, for mal decomposition of petroleum products.
example, alkyd hobby craft paints or India inks Carbon in the form of graphite pencil lead is
made from carbon black dispersed in shellac. also used by some museums as a temporary
marker for light-coloured plastics. The only dis-
In the dispersed dye inks, dark colours such as advantages are a tendency to smudge and possi-
black are usually made up of several different ble indentation ofsoft or degraded surfaces.
colours, which tend to react differently to long- Unfortunately pencils do not write well on soft
term contact with the various constituents of elastomers or vinyl sheets. Other black pig-
the substrate, resulting in bleeding or the devel- ments, such as those containing iron salts, cata-
opment of a halo around the ink. Polyolefins, lyse degradation especially in cellulose plastics,
nylons and all highly plasticized polymers are and are often very acidic.
particularly susceptible to this kind of damage.
The dispersed dye inks are also prone to differ- The white pigments are more of a problem;
ential fading and occasionally the haloed area many of the white inks used on our collections
may become hard and brittle, probably because in the past have deteriorated badly. Experimen-
of the photo-tendering effect of the dyes. Al- tal samples of zinc oxide, calcium carbonate,
though such inks may fade or bleed to illegibil- titanium dioxide as well as lead carbonate and
ity over time, they usually leave some sort of barium sulphate (both of the latter pigments
stain or distortion. are under restrictions because oftheir toxicity)
were applied in various media to old samples of
Removal of dyes is a risky procedure requiring cellulose nitrate and ebonite for observation.
solvents. The longer the ink remains in contact, Most of them proved to be unacceptable be-
the more difficult it is to remove as it sinks fur- cause ofhygroscopic degradation products or
ther into the matrix. Casein, urea-formaldehyde incompatible pH.
and melamine-formaldehyde as well as light-
coloured cast phenolic resins tend to be perma- Zinc oxide (Chinese white) was once widely
nently stained with a shadow of the ink number used as an activator in lulcanized rubbers and a
unless it is removed within two to three years. whitening agent in plastics. The comparatively
This is something to consider when buying an- good survival of celluloid ivory imitations, in
tique plastics because dealers seem to be almost comparison with translucent celluloid tortoise-
as fond ofball-point pens and felt-tip pens as
shell imitations, is usually attributed to the
they are ofpressure-sensitive labels. None of superior acid-scavenging and light-screening
the dye inks, or wax crayons and china mark- properties of zinc oxide over aniline dyes. How-
ers, which have a similar effect, can be recom- ever, zinc oxide is chemically reactive and its
mended for labelling museum artefacts. behaviour when applied as an ink can be very
damaging, especially if it is not well combined
Particulate pigments in a suitable medium in a water-resistant medium. On cellulose
are more promising. Only black and white
nitrate, zinc oxide inks in media such as gela-
tine, gum Arabic and methyl cellulose were

346
transformed into transparent wet patches on the alkyd medium, which were applied to ebonite
surface. The rate at which this happened de- jewellery in the 1960s, have yellowed, but they
pended on the ambient humidity and how badly are still perfectly legible and are adhering well.
the cellulose nitrate was degrading. Even water- However, attempts to use the same type of ink
proof alkyd paints coloured with zinc oxide and to apply new registration numbers to the same
exposed to emissions from degrading celluloid artefacts caused major discolouration ofthe ink
eventually discoloured and became hygro- and disruption of the artefact surface because of
scopic (Figure 8). Similar results were obtained changes in the properties of the ebonite during
with calcium carbonate, which is often incorpo- the last two decades of imperfect storage.
rated into inks and paints as an extender. Its
acid-scavenging properties have usually been Photosensitization is another problem with
regarded as an asset, but in this context they ac- some white pigments. Although zinc oxide and
celerate degradation. Metallic aluminium pig- titanium dioxide can have a protective light-
ments that had been used in an old repair on a screening effect, they are also reputed to sensi-
cellulose nihate artefact also liquified and the tize polymers to ultraviolet light, encouraging
results were repeatable with oil-based alu- stress-cracking and embrittlement. especially
minium inks. No such effects have occurred in nylons and the polyolefins.*'o The ellect has
with titanium dioxides, even on badly degrad- also been studied in paint films where the poly-
ing plastics, during the two years that they have mer medium erodes at an accelerated rate ex-
been under observation. posing the Bigment to mechanical damage
(chalking).' Chalking can be desirable in com-
On ebonite, all the white pigments discoloured mercial outdoor paints because of the self-clean-
to some extent, but at high relative humidities ing effect on grimy surfaces that constantly
(above 68oh), zinc oxide darkened rapidly until erode, revealing a clean undersurface (and
it was barely visible against the ebonite (Figure increasing sales by the rapid attrition ofthe
9). Examples of lead white in a waterproof paint). However, it is not an advantage on mu-
seum artefacts where handling could dislodge
registration numbers with degraded media. Of
griTf,* fixfdtl$ the two crystalline forms of titanium dioxide,
anatase is the more photo-active. Rutile tita-
,1..:i:.r,:s...
et,b lrio! ti"r'{rt{$xfJ nium dioxide is also available coated with sili-
{!$* iS*{
{rl$9iSl cone or non-porous inorganic minerals, which
f.\\. * eic\ 5x ?
help to reduce photo-activity to some extent.
ii\$\N:qi;,,,f *rr& - Bru. sr$t It is probable that photo-actination would be
more likely to affect the ink/paint medium than
the artefact beneath, but there is no guarantee
that harmful degradation products will not mi-
Figure 8 Effect of emissions from decaying
cellulose nitrate on three typical labelling media grate to the substrate.
/br plastics.
At present the best white pigment appears to
be coated rutile titanium dioxide, but it is also
advisgble to keep the artefact protected from
ultraviblet light, which is very harmful to most
plastics regardless of the presence of photosen-
sitizing inks. Most labelling media have some
light-screening effects so it is advisable to ap-
ply the number in an area protected from light
or there will be differential fading (Figure 10).

Aggressive Plastics
Figure 9
Various white pigments in gelatine
medium on degraded ebonite show relative Labelling media can attack artefacts, but the
discoloration within one month at ambient humidity. effect of plastic artefacts on the labelling media

J+t
 and flaky and most of the black dye-based
inks showed severe fume fading. There was se-
vere discolouration of many of the resins, espe-
cially epoxy and polyurethane varnishes. Alkyd
vamishes yellowed, but not to the extent of ob-
scuring the legibility of the inks, and their adhe-
sion was initially unaffected although some
degraded before the end ofthe year. Not sur-
prisingly, cellulosic resins, such as methyl cel-
lulose and cellulose nitrate, showed very poor
resistance as did archival paper labels. All me-
tallic wires corroded rapidly, nylon monofila-
Figure I0 Dffirential fading on casein button ment was eventually weakened to breaking
where pigments have blocked the light. point by direct contact with both plastics, and
cellulose threads were only slightly more resis-
tant. Polypropylene sheet and cords seemed to
also needs to be considered. Numbering sys- be unaffected, but the manufacturers did not
tems must be able to accommodate dramatic recommend its long-term use under such ag-
dimensional changes caused by flexing, tem- gressive conditions. The severiry of the damage
perature and deterioration without flaking off; in so short a time was surprising and honifying
otherwise the registration numbers will quietly since the decaying piece of cellulose nitrate
disappear from the artefacts in storage. Some of was not very large and the conditions of the
the old gelatine and gum Arabic inks used on experiment were very similar to conditions in
our cellulose nitrate in the past have flaked to which some of our early cellulose nitrates are
illegibility, whereas shellac-based inks are still kept. Damage caused by the ebonite was less
legible even where the celluloid is disintegrat- dramatic except where inks and tying materials
ing beneath them. were in direct contact with the acidic surface.

As plastics degrade, some of them become in- After two years of exposure, shellac, poly(vinyl
creasingly corrosive giving off volatile materi- acetate) resins (but not emulsions) and acrylic
als, such as nitrogen dioxide, sulphur dioxide, resins (both emulsion and solvent types) have
hydrogen sulphide, hydrogen chloride, formic neither embrittled nor discoloured.
acid, acetic acid and camphor. Such emissions
inevitably attack labelling media, not only on These initial results suggest that neutral aque-
the plastic artefacts but also on adjacent ous dispersions of acrylic polymers are the saf-
collections. est media in which to mix pigments that are to
be applied directly to plastic surfaces. Colloidal
Various resins and labelling materials used in dispersions with their finer particle size gener-
our own and other museums were applied to a ally have superior prope{ies and fewer addi-
mylar strip and enclosed (but not sealed) in a tives than tlie emulsions,8 but both dispersions
small Plexiglas display case with a piece of de- and emulsions that have been thickened with
cayrng cellulose nitrate llom a 1930s' handbag. organic solvents or cellulose ethers should be
A set of controls was kept in a similar case in avoided because ofthe increased risk ofsolvent
the same ambient environment. Materials that damage or decreased moisture resistance in the
proved resistant to the emissions were sub- dried film. Dispersions are manufactured for
sequently tested in direct contact with decaying the paint industry where requirements are con-
cellulose nitrate. A similar experiment was set tinually changed and new lines are introduced
up using an ebonite bracelet from the 1870s, and discontinued frequently. So it is advisable
which was emitting both hydrogen sulphide to assess dispersions carefully before using
and sulphur dioxide. them. Desirable properties for labelling media
are a neutral pH; a low organic solvent content;
Within six months many of the inks and paints tough, slightly flexible films; and a glass transi-
exposed to cellulose nitrate had become brittle tion temperature above room temperature. At

348
present we are using Acrysol WS 68 (Rohm being much more available and easier to use,
and Haas), which satisfies most of the condi- but the alcohol solvent poses a risk to highly
tions except that the film is rather soft. stressed plastics as well as those that are sol-
uble or partly soluble in alcohol. Some light-
Conclusion : Guidelines coloured plastics that will not be subjected to
for Labelling Plastic Artefacts extensive handling can also be numbered safely
using a soft (2B to 48) graphite pencil.
Each plastic has its own sensitivities, which
are affected by manufacturing techniques, addi- There is still an element of risk that can be
tives, age and use. It is difficult for institutions further minimized by careful placement of the
without sophisticated analytical equipment to registration number.
distinguish one type of plastic from another,
since many simple tests are destructive and un- . Where possible apply the label to a non-
reliable.'The staff responsible for applying reg- plastic component of the artefact, preferably
istration numbers cannot be expected to spend not directly adjacent to plastic components.
a great deal of time devising separate methods
for labelling each individual plastic artefact, . Avoid the areas that are most likely to be
taking its specific sensitivities into account. To exposed to light, abrasion or handling.
simpliff matters the following are suggested
for labelling artefacts made wholly or in part o Avoid areas that are already visibly degraded,
from plastics: that is, dulled, faded, cracked, distorted or
excreting plasticizers.
Use tie-on tags madefrom Teflon tape where
possible. Plumbers' tape, which comes in sev-
o Avoid areas that are likely to be under stress,
eral grades, can be used as the tag or twisted
that is, adjacent to glue lines, repairs, inter-
to form the tie, and it is much cheaper than
faces between different plastics, areas that are
other forms of poly(tetra fluoroethylene) tape.
Mylar D, chemically known as poly(ethylene constantly flexed, such as Tupperware lids or
terepthalate), can be used to make chemically polypropylene-membrane hinges. Anything
resistant tags, but it has extremely sharp edges that might swell the polypropylene and relax
that may cut through the cord attaching it to the biaxial orientation of the molecules will
the artefact. reduce flex resistance. Knife handles also are
usually under stress because the metal tang is
Number the artefoct directly without an inter- inserted in the plastic while it is hot, so that as
vening layer ofvarnish; use a suitable it cools, it shdnks around the metal to give a
ink/paint of known composition. To avoid tight fit. The label in Figure I has been placed
problems with constantly changing constituents on the most stressed area of the artefact.
in a commercial ink, it is safer to make one.
Finely divided carbon black or coated rutile Finally, in case the numbering should initiate
titanium dioxide without impurities can be damage in spite of all precautions, avoid plac-
mixed into a neutral acrylic dispersion, such as ing labels adjacent to the manufacturers' marks,
Acrysol WS 68. Homemade paints are more dif- where critical information can be destroyed
ficult to use than commercial mixtures because frst.
they lack the additives that control flow, homo-
geneous pigment dispersion and drying time. R6sum6
Nevertheless it is possible, with a little practice,
to adjust the viscosity with water to permit La pose des numdros d'enregistrement
small, neat lettering that will not spread into sur les objets en plastique
degraded surfaces and can, if necessary, be re-
moved mechanically after drying. Shellac is Il est de plus en plus |vident que la pose de
the next best choice ofmedium for carbon and numiros d'enregistrement sur les objets en
titanium pigments and it has the advantage of plastique prdsente des risques dlevds. Il peut en
effet alors se produire une interaction ithtersible

349
 entre, d'une part, les plasffiants, les solvants, les poussdes portant sur dffirentes combinaisons
pigments et les teintures et, d'autre part, le sup- d'encres et de polymdres qui ne risquent pratique-
port de plastique, et un tel phtnomdne pourrait ment pas d'abimer I'objet, tout en resistant mieux
fort bien accdldrer lafissuration sous contrainte aux produits de ddgradation qui en proviennent.
du mattriau, sa ddformation ou sa ddcoloration,
voire sa ddsintdgration compldte. Le temps que References
mettront les dommages d apparaitre parfois
des mois, ce qui les situe dans le long-lerme et Umney, Nicholas, "Oriental Lacquer,"
le manque d'informations sur les conditions -
1.
prdsi- Conservation News, vol. 32, March 1987.
dant d laformation des complexes moltcalaires pp.23-25.
font qu'il devient impossible d'effectuer ne serait-
ce que des essais sur des parties cachdes de l'ob- 2. Webb, Marianne, "Light Degradation of
jet pour tenter de ddterminer les combinaisons Oriental Lacquer," IIC-CG Abstracts of the
dommageables. Ironie du sort, si I'on procddait d I 3th Annual Conference, Victoria, B.C., 1987,
de tels essais, cette premidre ttape du processus p.55.
visant d conserver I'objet pour la posterite
pourraitfort bien constituer I'amorce de 3. Williams, R. Scott, Janet Waddington and
sa destruction. Jul ia Fenn," lnfra-red Spectroscopic-Analysis
of Central and South American Amber Ex-
Chaque genre de plastique etant vulnerable d posed to Air Pollutants, Biocides, Light and
une strie particuliire de matiires, il serait thto- Moisfure," Collections Forum, vol. 6, no.2,
riquement possible d'emp€cher qu'ils ne s'abi- Fall 1990, pp. 65-75.
ment en Afubhssant, pour cltacun, un mode
d' dtiquetage particulier. Reste ndanmoins qu'il 4. Allen, Norman S., "Effects of Dyes and
faudrait d'abord pouvoir les identifier cotecte- Pigments," rn: Comprehensive Polymer
ment, et qu'il estfort dfficile de lefaire sans ap- Science, ed. G. Allen, vol. 2, chap.20,
pareils d'analyse, qui sont par ailleurs coilteux. (London: Pergamon Press) pp. 579-595 .
Et m€me si un tel systime d'6tiquetage,fiable,
etuft mis en place, il faudrait aussi considtrer 5. Nitrocellulose. The Chemical and Physical
I'incidence d'autresfacteurs comme les tech- Properties, Technical pamphlet (Wilmington,
niques defabrication, les additifs, les contami- Delaware: Hercules Inc., 1979) p.28.
nants et le vieillissement. Sans compter que les
param,itres de solubilitd de plusieurs polymdres 6. Davidson, R.S. and R.R. Meck, "The
se modifient avec le temps. Photodegradation of Polyethylene and
Polypropylene in the Presence and Absence of
L'autre aspect de cette interaction, d savoir Added Titanium Dioxide," European Polwer
I'effet que peut avoir I'objet en plastique sur Journal,vol. 17, 1981, pp. 163-167.
l'6tiquette, n'est, par ailleurs, pas toujours pris
en compte. Les objets qui comportent des ma- 7. Hoffman, E. and A. Saracz, "Weathering of
tidres comme la wlcanite, le polychlorure de Paint Films: I. Chalking Caused by Zinc Oxide
vinyle (PCV), I'acitate de cellulose et le nitrate in Latex Paints,"; "Weathering of Paint Films:
de cellulose isquent ainsi defaire blanchir ou II. Chalking Caused by Anatase Titanium
couler les encres et d'attaquer lapeinture et le Dioxide in Latex Paints," Journal of the Oil
vernis, si bien que, dans un laps de temps relati- and Colour Chemists Assoc., vol. 52, 1969,
vement coun, les numdros devimdront illisibles, pp. ll3-132; pp. I 130-1 144.
s'ils ne s'effitent pas compldtement. De tels
dommages seront non seulement causts aux 8. Koob, Stephen P., "Consolidation with
6l6ments (encre et vernis) de marquage clas- Acrylic Dispersions," Preprints of the
siques, mais encore aux fils de textile ou defer 9th Annual Meeting of the American Institution
et aux autres mafiAres qui, servant d attacher les for the Conservation of Historic and Artistic
etiquettes volantes et les etiquettes d'exposition, ll/orks, May 1981, pp. 86-94.
auront iti placds dans les vitrines d'exposition d
p roximitt de p I as tique s instables.
9. Coxon, Helen C., "Practical Pitfalls in the
Identification of Plastics." this volume.
Dans le cadre de la prdsente communication,
nous ferons 6tat dans le dttail d'expdiences

350
Degradation Rates for Some Historic Polymers and the Potential
of Various Conservation Measures for Minimizing Oxidative Degradation

David W. Grattan

Canadi an C ons e rvation I ns titut e

artme nt of Communic atio ns
D ep
Ottawa, Ontario
Canada

Abstract In this paper the results of these measurements

are considered in relation to various conserttation
Polymers in theform of plastics, rubbers or tex- procedures, with special reference to refrigera-
tiles cause major headaches for museums. Their tion and the use ofAgeless* oxygen absorber.
instability and unpredictability make the work of
conservation, display and storage a dfficalt chal- *Ageless is a trade name of the Mitsubishi
Gas
lenge. At the Canadian Conserttation Institute, we Chemical Companlt Inc. of Japan.
are well aware of these dfficulties and also of the
limited nature of the options that we have to cope Introduction
with them. For many polymers, degradation is
caused by a chemical reaction with ox-vgen. There Rubbers or plastics (i.e., materials composed of
may be manyfactors that initiate this process and carbon chain polymers) degrade principally by
many complex stages involved but the outvvard
oxidation. Although other forms of decay, such
signs are quite straightforward. Polymeric materi-
as depolymerizalion or the acidic breakdown of
als react with oxygen and become increasingly
cellulose nitrate and cellulose acetate,do occur,
fragile. They may change in volume and colour for many polymers the principal pathway in-
and may also become acidic or release volatile
products. volves the absorption of oxygen ffom the at-
mosphere. Following this step, which is a
These effects, which resultfrom chemical reac- chemical reaction, various oxidation products
tions, can be slowed down in several ways. The are formed, and the polymer backbone is often
oxygen can be excluded, the temperature can be fragmented. As a result, the chemical and physi-
reduced, the active chemical species within the cal nature of polymers alter, and this affects the
polymer can be trapped or destroyed chemically plastic or rubber materials of which they form
by antioxidants or stabilizers, and light (espe- the major component. There may be several di-
cially UV) can be excluded. This list is not ex- verse outward sigas. Plastics may shrink, be-
haustive, but it does cover the main options. come brittle and decrease radically in strength.
These changes result in distortion and cracking
Although a number of the above countermeasures of objects. Materials may alter in colour either
have been explored, little is knovry about the ac- because dark-coloured oxidation products form
tual rates of decay of histoic materials. Thefrst or because the colourant oxidizes or fades. The
step to measuring the success of any consertation production ofoxidation products, such as acids,
procedure appears to be the measurement ofoxi-
aldehydes, ketones and peroxides, causes poly-
dation rates, and oxygen absorption measure-
mers to become more polar. Thus, their solubil-
ments for ambient and low temperature studies
are cunently being used. ity characteristics alter and this often means

351
that plasticizers, originally dissolved within, reduced in the low oxygen levels created by
come out of solution and migrate to the surface. Ageless, but would also probably be reduced in
atmospheres in which some oxygen remained
If oxidation can be stopped, many (although (i.e., ineffectually scavenged). The real ques-
not all) ofthe degradation processes will cease. tion ofwhether the very low oxygen concentra-
The simplest way to do this is to deny the poly- tions produced by Ageless would allow a
mer access to oxygen by using a nitrogen or ar- slowly oxidizing material (i.e., a typical mate-
gon atmosphere or placing the polymer under rial) to continue degrading would remain
vacuum. Oxygen absorber sachets, such as unanswered.
Ageless (manufactured by Mitsubishi Gas
Chemical Company Inc. of Japan) or Freshpax It was concluded that the problem could only
(marketed by the Ludlow Corporation of the be tackled indirectly. To do this, the following
U.S.), provide a simple way of removing oxy- information was required:
gen from closed atmospheres, and the investiga-
tion of this approach forms the topic of o The rate ofoxygen absorption ofhistoric
this paper. polymeric materials under ambient and cold
conditions
Oxygen absorbers do not, however, remove all
the oxygen. A small proportion remains (with o The actual level ofoxygen in a closed
Ageless this is at minimum 0.01%) and it is vi- envelope containing an oxygen absorber, and
tal to know whether this is sufficient to prevent how this level changes over time
significant oxidation. In addition, most contain-
ers leak over time; for instance, most oxygen- . The leak rate of plastic envelopes
barrier films are slightly permeable. Thus, one
of the functions of the absorber is to trap the Experimental
oxygen that leaks in.
Ageless Oxygen Absorber: This material is a
A short-term direct test demonstrating th'bt mixture of finely divided iron, molecule sieves,
materials stored in low oxygen conditions do sodium chloride and moisture that binds oxy-
not degrade would be most useful. This would gen chemically in the presence of moisture. It is
have to be accomplished by measuring directly marketed in sealed paper sachets that are perme-
the accumulation of oxidation products or the able to oxygen but less so to moisture. Differ-
alteration in properties. However, there are ent formulations are sold for various ranges of
problems with this approach. For such an ex- relative humidity (RH), and Ageless Z,the for-
periment, objects would be placed in oxygen- mulation used in these experiments, functions
free conditions and attempts would be made best in the RH range of 0o% to 85%. Other for-
to observe differences in the accumulation of mulations of Ageless are intended for higher
oxidation products or alteration in properties humidity. Sachets are prepared in various sizes
compared to similar objects stored in air. But according to the volume of oxygen that can be
degradation is quite slow and the resulting absorbed.
changes slight. Analyses ofthese changes
would have to be very sensitive to small differ- Ageless Eye is an indicator that tums pink in
ences and the test objects would also have to less than 0.01% oxygen.
be homogeneous in condition (particularly in
their state ofdeterioration). This direct ap- Ageless can create oxygen-free conditions
proach was thus considered unlikely to in heat-sealed oxygen-barrier fi lm enclosures.
prove fruitful. Flexible films allow the consumption of oxy-
gen without pressure differences developing
One way of avoiding these problems might across the barrier. Filrns simply flex to take up
be to use rapidly deteriorating materials that the volume change as oxygen is consumed. In a
would quickly show oxidative changes. How- rigid enclosure, a pressure difference develops,
ever, since such materials have large appetites which encourages gas leakage.
for oxygen, oxidation rates would certainly be

3s2
Heat-sealed envelopes containing Ageless sa- only, whereas the Condor film was taken from
chets were prepared from a number of barrier a standard production run.
films. The gaseous contents were analyzed by
gas-chromatography over a period of a few Experiments
months. The aims were to measure oxygen de-
pletion and moisture content. A series of experi- Ageless Z was tested in envelopes composed
ments were carried out. At first, there was no of each of the five types of barrier films. Two
control of RH, but in the later experiments, envelopes were prepared for each film type at
conditioned silica gel sealed in Remay pouches each intended relative humidity. Relative hu-
was added to produce a range of relative hu- midities of 5lYo, 650/o and 77oh were regulated
midities within the envelopes. by placing silica gel in the envelopes, and one
set of envelopes w€IS prepared with no humidity
Barrier Films: Du Pont Canada Inc. of Missis- control. Thus 40 envelopes, each varying in
sauga, Ontario, and Condor Laminations Ltd. volume from around 150 mL to 200 mL, were
of Agincourt, Ontario, supplied batches of prepared.
films for testing. The Du Pont films included
Oxybar 1,2,3 and4. Oxygen permeance data The sealed envelopes contained:
(supplied by Du Pont and measured following
ASTM D1434) for these films are shown in o An Ageless sachet
Table I. (N.B. Permeance is defrned as the ratio
of the gas transmission rate to the difference in o Four sampling serum cap vials (there was
partial pressure of the gas on the two sides of only one vial per envelope ofopaque
the film.) Oxybar 4 fihn)

Table I . Ageless Eye (not included in the opaque

Permeance of Barrier Films Supplied Oxybar 4 film envelopes)

o l5 g ofconditioned silica gel in heat-sealed

Remay pouches. The use and preparation of
these is explained in more detail below.

Ageless: Samples of Ageless were received

from Mitsubishi packed in heat-sealed enve-
lopes composed of barrier frlm. The Ageless
* Permeance units are rn mL oxygen per 100 square
sachets were kept thus until needed for experi-
inches of banier film, per 1 atmosphere partial pres- ments. Envelopes were prepared in the open
sure difference f rom one side of the film to the other
for 24 hours at 23"C. laboratory and were heat-sealed quickly with
an Audion Elektro Sealmaster 420.
Composition of Films: No composition data
or permeance values are available for the Con- Ageless Eye: Indicator capsules of Ageless Eye
dor film. The Oxybar films were taken from (also received from Mitsubishi) were added to
batches produced for experimental purposes each envelope of transparent film.

Table II
The Composition of Oxygen Barrier Films
(lnformation is based on data supplied by Du Pont.)
Film Barrier Sealinq Lavel
Cxvbar 1 polv(vinVlrdene chlonde) oolvethvlene
Oxvbar 2 colv(ethvlene vinyl acetate) oolvethvlene
Oxvbar 3 onented nvlon oolvethvlene
Oxvbar 4 (ooaoue frlm) alumtnium metallized polvester polyethylene

353
Silica Gel: A 50:50 mixture of 3-9 mesh Unfortunately argon elutes at the same point as
grade 41, and 3-8 mesh grade 59 silica gel was oxygen with this column, and thus oxygen was
conditioned for six months with satwated solu- determined by subtraction of the argon. This
tions of sodium chloride to give 77o/oRH, mag- proved to be very inaccurate because argon
nesium acetate to give 65% RH and calcium (0.934% of the atrnosphere) is present in over-
nitrate to give 5l%o RH. Portions of 15 g were whelming amounts compared to oxygen and
heat-sealed in Remay (non-woven textile) thus this method was discontinued. Some analy-
pouches for inclusion in the envelopes with ses of water were successful with the CTR-l
the Ageless. column and estimates of RH were carried out.
They were very inaccurate.
Sampling Procedure: The atmospheres
in the sealed envelopes (i.e., containing Age- In a second series of expe"riments a HayeSep A
less) were sampled at regular intervals using porous polymer column"' ( l0 metres of
four 3.5 mL screw cap semm vials. At pre- 0.3 mm diameter HayeSep A, 80/100 mesh)
determined times, labelled vials were capped separated oxygen from argon and nitrogen. A
by screwing a neoprene/teflon membrane in typical analysis is shown in Figure L Water
place. The vials remained in the envelopes analysis, and hence RH determination, was not
until time of analysis. possible with the HayeSep A column.

Measurement Procedure: All sampling and Conditions for the HayeSep A column were:
transfer operations were carried out in an oxy- Column temperature 0'C, injection port
gen-free atmosphere. For this pulpose, a nitro- temperature 25'C and detector temperature
gen glove bag was placed on top of the gas 135'C, flow rate l7 mL per minute ultrapure
chromatograph with the injection port sealed helium carrier sas.
within. To conduct an analysis, the envelopes
containing Ageless were placed in the glove
bag, which was then closed. The bag was in-
flated and purged with dry nitrogen, and to
ensure that oxygen had been purged out, the
atmosphere was sampled repeatedly by making
injections into the gas chromatograph. This was
carried out until the oxygen peak had almost
disappeared in the gas chromatogram.
0.017o OXYGEN
Analytical Procedure: Analysis was per-
formed on a Hewlett Packard 5840 A dual
column instrument with thermal conductivity Figure ITyptcal chromatogram shov'ing, in order
detectors and, initially, with an Alltech concen- ofdescent, (1) large nitrogen peak, (2) scale change
tric column model CTR-I. The outer column to a more sensitive setting at Zf , 61 small oxygen
was 2 m x 63 mm activated molecular sieve, peak almost at detection limt, and (4) argon peak.
and the inner,2 mx32 mm "Porapak mixture."
(This is presumably composed of two or more Calibration: This was performed by measuring
unspecified grades of Porapak.) This column is response in area units, as given by the integra-
intended for the analysis of oxygen, water, tor function of the chromatograph, versus injec-
carbon dioxide, nitrogen, carbon monoxide, etc. tion volume measured at about 757 torr
Conditions for the CTR-1 column: Column tem- pressure and 23.5'C. Calibration data is shown
perature 100"C, injection port temperature in Table III, and plots of peak area versus sam-
80"C and detector temperature 135'C. Ul- ple size for various gases are shown in Figures
trapure helium carrier gas flowed at 45 rnl- 2-3. and,4.
per minute.

3s4
 Table III
Response of Gas Chromatograph in Area Units
for Various Gases
Gas Response
(area units / microlitres)
Arqon 23,49A
Nrtrooen )) )AA
Oxyoen 20,476

Results

Figure 2 Calibration of HayeSep A column for The results of analysis by gas chromatography
argon. are shown in Table IV.

Table IV
Oxygen Content of Envelopes as Given by Gas

" -
Chromatography

Film Intended RH Average Percent

Oxygen Analysis by
HaveSep A'
Oxybar 1 517o 0.000%
65% 0.000%
77"/" 0.000%
uncontrolled (0.05 10.03%)
Oxibar 2 51% 0 000%
OC-/o 0 000{.003%
Figure 3 Calibration of HqveSep A columnfor 77"/" 0.0004.005%
nitrogen. uncontrolled (0.039 t 0.03%)
Oxrbar 3 51"k 0.000%
651" 0.000-0.0040/.
7710 0.000-0 0090/.
uncontrolled (0.026 10 03%)
Oxybar 4 51% 0.004-0.03%
65% 0.o07-0.02"/"
77Vo 0.000%
uncontrolled (0.52 i 0.03%)
Condor 51% r0 017 I 0.03%)
65% t0.018 r 0.03%)
77"/" (0 01310 03%)
uncontrolled (0 026 10 03%)

'Frgures In parentheses denote CTR-1 readtngs.

Figure 4 Calibration of HayeSep A columnfor
oxygen.
Discussion of Oxygen Analysis
The limit of detection was given by the height
The oxygen analysis using the CTR-l column
of the smallest observable peak. ln practice this
was very imprecise because of the interference
indicated a limit of about 0.003% oxygen. The
of argon with the oxygen peak. Thus the values
value of argon was typically around l.2o/o, as
quoted in Table IV are for the analysis with the
would be expected from a 20% reduction in
HayeSep A column (except where these data
volume (i.e., from loss of the oxygen).
are unobtainable). Oxygen analyses obtained

355
with the CTR-I column are shown in brackets. ofa deteriorating object in air over a reasonable
There is an uncertainty of at least t 0.03% for period of time.
all the values quoted above.
Measurements of Oxygen Uptake of
Those data obtained with the HayeSep A col- Some Historic Materials
umn (which does separate argon from oxygen)
show very much lower oxygen levels, which Measurement of the rate of oxidation of historic
are in most cases below the detection limit of materials has been performed with an oxygen
0.003%. The uncertainties in the values given absorption apparatus namely the Warburg
in the HayeSep A column result from differ- respirometer. Intended- for use for determining
ences in the amount of oxygen measured. In the oxygen demand by living organisms, the
some instances there was a small amount of Warburg has been successfully used here to
oxygen in the glove bag and this may have measure the comparatively low rates of uptake
contaminated the syringe during transfer opera- of oxidizing polymers.
tions: at these very low levels it could happen
very easily. In most instances the oxygen con- Experimental
tent of all vials was about the same regardless
of when sealing took place. It seems logical The respirometer, shown in Figure 5, consists
that oxygen cannot diffuse out of the sampling of a25 mL conical flask connected to a mer-
vials during transfer. Thus it also seems logical cury manometer. The oxidizing material is
that the lowest reading for oxygen is always placed in the conical flask and the apparatus im-
likely to be the most accurate. mersed in a thermostatically controlled bath at
25'C. After allowing time for equilibration the
This analysis was carried out one year after the apparatus is sealed with the mercury level in
vials had been sealed. Thus in most instances the manometer set at a known point (15 cm on
the level of oxygen was below the limit of de- the 0 cm to 30 cm scale), and the atmospheric
tection (0.003%) and remained so for the pe- pressure noted. At intervals of about one month
riod of the experiment regardless of the relative the manometer is carefully read (usually over a
humidity. range of pressures to compensate for the inaccu-
racy of single point readings) and the reading
The water analysis was found to be unreli- corrected for atmospheric pressure variation.
able and it is thought that the control exercised In this way (i.e., by measuring the reduction in
by the silica gel as conferred by the pre- total gas volume in the apparatus) the amount
equilibration with- saturated salts is a better ofoxygen absorbed can be calculated per unit
guide to envelope humidity than is - the meas- weight of polymer. If a significant proportion
ured water-peak area. This poor water analysis of the oxygen is removed then it is relatively
is not surprising since the water peak "tails" simple to restore the atmospheric gas composi-
badly in the chromatogmms. tion by passing air through the apparatus. It can
then be re-sealed for further measurements.
Ageless Eye went pink in all sealed envelopes
very rapidly and remained in that condition The apparatus was modified to enable oxygen
until analysis. The colour change was most uptake measurements to be made at lower tem-
rapid in the envelopes of higher humidity. It peratwes (ca. -20'C). For this purpose, a tap
therefore has proven to be a reliable indicator that could be closed between measurements
ofoxygen content so far. was inserted between the manometer and the
conical flask. Without the tap, the reduction
It is clear that Ageless is capable of producing in pressure brought about by the lowering of
very low levels of oxygen. However, for this temperature caused mercury to be sucked out of
technique to be considered useful as a method the manometer into the conical flask.
ofanaerobic storage, oxygen levels must be sig-
nificantly lower than the oxygen consumption

356
 Figure 7 Orygen uptakefor shoe sole (NSA-I),
PVC pad (FH-1), ABS cup (FH-4), pofuurethane
foam (FH-2) and (FH-5), ABS "spacer" (FH-3)
(see Table V).

Discussion
For all samples, the amount of oxygen that
Figure 5 The Warburg respirometer. would normally be consumed in a 100-day
period is well in excess of the static amount of
Results oxygen available in the sealed envelopes. There
is a problem however, in that the concentration
Figures 6 andT show the rate of oxygen uptake of oxygen within the envelopes results from
for a number of items of different polymeric a dynamic equilibrium. Oxygen diffusing
materials. (Note that the graphs are labelled in through the film is consumed by the Ageless.
descending order, and that more information More oxygen than is suggested by the static
can be found in Table V, which also gives value ofthe percentage oxygen content enters
quantitative oxygen absorption data. Note also the sealed bag and may therefore be available
that the oxygen uptake is estimated for a period to the oxidizing object.
of 100 davs.)
There are thus three main problems to resolve:

o The leak rate of different oxygen barrier films

is an important factor. It is important to
quantify how much oxygen actually flows
through each type ofbarrier.

o The competition for oxygen between Ageless

and an oxidizing organic object should be
considered. Ifan object can compete success-
fully for oxygen it will continue to decay.

. lf a given bag leaks oxygen at a certain rate,

how much extra Ageless is necessary to
remove the leaked-in oxygen over a
Figure 6 Oxygen uptakefor lamp base (LB-A),
rubber cape QIryS-l) (see Table V). l0- to 20-year period?

357
 Table V
Meaurements of Ox-vgen [Jptake of Some Historic Materials
(It is assumed that at room temperature [25"CJ one mole of orytgen occupies 24,450 cc.)

Sample Polymer Sample Perriod of Amount of 02 Oxygen

Code Test (days) Consumed per Absorbed as
Gram Polymer Percentage of
(moles) Air Volume
oer 100 davs
Rubber cape Butyl rubber wrth CWS.1 r08 53 x10-7 o 40%
(WW ll, 1940s) TtOz ftller
Plastrc lamp Polyethylene with LB-A I 5.4 x 10-6 75%
base (1970s) TtOz ftller
Rubber shoe rs-1 ,4-polyisoprene NSA-1 490 1 .76 x 10-5 2.9"/.
sole (1 950s) plus styrene-
butadiene
copolymer. Ca,
Al srhcates,
TiOz ftller
Plastrc ear cup Poly FH4 490 9.2 x 10-6 1 5"/"
In prlot's helmet (acrylonrtrile/
(1 9s0s) butadiene/styrene
wrth TtOz ftller
PVC pad from Poly(vnyl chlonde), FH-1 490 1 .2 x 10-s 2.Oo/.
pilot's helmet dr(2€thyl hexyl)
(1 950s) phthalate
(plastrcrzer), trace
TrOz
Grey plastic spacer Poly FH.3 490 12 x10'6 0.27"
from pilot's helmet (acrylonrtnle/
(1 950s) butadrene/styrene)
wth TtOz ftller
Black foam "rubbe/' Polyurethane FH-2 490 26x10-6 o 44"/"
from prlot's helmet
(1 950s)

Black foam "rubber" Polyurethane FH.5 490 6.1 x 10-' 1 .Oo/"

trom prlot's helmet
(1 950s)

0.08 (permeance units, mL oxygen per

Leak Rate: Examinations of number of
a 100 square inches of barrier film, per
bagged objects have led to a very rough rule- I atmosphere partial pressure difference
of-thumb that for every gmm of object, we from one side of the film to the other for
can normally expect about 6 square cm 24 hours at 23'C)
( I square in.) of barrier wall (as objects become
lareer this value decreases) and about 3 mL of x li 100 (to correct the value for
air space once the oxygen is consumed (as ob- I square inch of barrier filrn)
jects get larger this value tends to decrease).
In practice this latter value of 3 mL per g may x 100 (to correct the value for a
well vary from I mL to l0 mL depending on 100-day period)
the shape ofthe object.
x2lll00 (to correct the value for
Thus for a 100-day period with Oxybar I film the partial pressure of
barrier, the leakage rate per gram ofobject oxygen approximately
(using the Du Pont permeance values quoted 21%o of -the atmosphere)
in Table D would be calculated thus:
: 1.7 x 10-2 mL oxygen passing through
the barrier in 100 days
per gram ofobject.

358
But, as stated above, for each gram ofobject Of these, probably the most important consid-
we have assumed there are 3 mL of gas space eration is the huge surface area of Ageless. This
(after the initial oxygen is consumed), thus the arises from its small particle size and porous na-
percentage ofoxygen that would leak in equals: ture. The alkyl radical, produced as a reactive
intermediate in the degradation of organic mate-
100x 1.7 xl0'2 l3:0.6% rials, may compete with Ageless for oxygen. It
is only ever present at very low concentration,
This means that if an average-sized envelope but reacts with oxygen exceedingly rapidly. De-
of Oxybar 1 (containing no Ageless) was to be spite this, it is thus quite reasonable to suppose
filled with pure nitrogen, after 100 days it that little of the leaked-in oxygen will end up in
would contain 0.6% oxygen. the oxidizing material, because the surface pre-
sented by the Ageless is so large and the reac-
The values for the other Du Pont barrier films, tion is also very rapid.
calculated similarly, are shown in Table VI.
Capacity of Ageless: Ageless sachets are rated
Table VI by the manufacturer in terms of the number of
Estimated Percentage of Leaked-in Oxygen cc. of oxygen that can be absorbed. Mitsubishi
suggests that the amount of Ageless to be em-
for Envelopes Composed of Vqrious Barrier ployed in a given storage envelope should be
Films at 23"C
based simply on the amount of oxygen initially
present. Allowing for oxygen leakage is not
Barrier Film Percent Oxygen Percent Oxygen
at 50olo RH at 0% RH
suggested. Presumably this is because for food
Oxvbar 1 0.6 0.6
packaging, only a relatively short storage pe-
Oxvbar 2 06 o.2 riod is necessary. For conservation, however,
Oxybar 3 AF o.2 leak rate is an important consideration.
Oxvbar 4 04 0.4
Consider what might happen to a 100-9 object
in a sealed envelope composed of Oxybar I .
Competition for Orygen: As shown above,
the rate ofoxygen leakage is such that ifthe According to the rule-of-thumb explained
organic material were to compete for it success- above, for an approximately 125-9 object we
fully, the oxidation rate would still be signifi- expect 380 mL of air to be present initially. In
cant, despite the low oxygen concentration in this volume of air, 80 mL of oxygen is present,
the envelope and the presence ofAgeless. The thus the volume becomes 300 mL after oxygen
principal reaction by which oxygen is absorbed, removal.
the interaction of alkyl radicals with oxygen to
form peroxy radicals, is very rapid. Other fac- Over 100 days, we know from Table VI that
tors that play a part are the following: 0.6% oxygen leaks in, which means that

o The overall oxidation rate ofthe stored object (0.6/100) x 300 : 1.8 mL of oxygen enters
the storage envelope.
. The concentration of oxygen-absorbing
species in the stored object Over a year, it means that 6.6 mL of oxygen en-
ters, and over l0 years 66 mL is expected to
o The rates of diffusion of oxygen into various leak in. For an envelope intended for long-term
materials use oxygen leakage is therefore significant and
may overwhelm the Ageless capacity unless
o The amount of Ageless added and its surface allowed for.
area
Conclusion
r The rate of reaction of Ageless plus oxygen
If Agelesscan bring oxygen levels down to
an amount that is significantly less than that

359
consumed by decaying objects over a 100-day Acknowledgement
period, further oxidation is probably prevented
or at least severely limited. Oxygen that perme- The author thanks Elizabeth Moffatt of
ates through the filrn will be competed for by Analyical Research Services of CCI for the
the decaying object and the Ageless. It is prob- analysis of both plastic and rubber samples.
able that the Ageless will trap most of the
leaked-in oxygen. R6sum6

In order to maintain an environment with a Le rythme de ddgradation de certains polymires

very low level ofoxygen, oxygen permeation d'importance historique et les diverces mesares
into the envelope should be limited as much as de conservation qui permefrraient de riduire au
possible. This can be achieved by placing the minimum la ddgradation orydative
"object envelope" within an outer envelope.
Each would contain an Ageless sachet. The Si ce n'est ddjd clair pour les spdcialistes de
la restauration, il leur sffira de parcourir
outer envelope would maintain the oxygen
les prdsents rdsumds pour constater que les po-
level at a low level around the inner envelope
lym,ires plastiques, caoutchoucs et textiles
so that oxygen diffusion into the inner envelope -
posent des probldmes sdriew au personnel des -
would be negligible. (The actual rate of permea- musdes. La conservation, I'exposition et la mise
tion would be reduced by a factor of 840 com- en risen;e de ces matiires instables et imprtvisi-
pared to an envelope exposed directly to the bles reprdsentent ainsi un d6fi de taille. Et d
atmosphere. This is because the oxygen partial I'Institut canadien de conservation, nous nous
pressure difference across the qnvelope wall sommes bien rendus compte de ces dfficultts,
would de,crease from 2. I x l0-' affn to de m€me que du nombre limitt de solutions qui
2.5 x l0-- atm.) peuvent €tre mises en euvre pour tenter de les
r'lsoudre. La ddgradation de nombre de po-
For the inner envelope the amount of Ageless lymdres tient bien souvent d une rtaction
to be added could be calculated in the manner chimique avec I'oxygdne. Si ce processus peut
suggested by Mitsubishi is, it would be s'amorcer par suite de nombreuxfacteurs, et s'il
govemed by the amount of -that
oxygen initially pre- comporte parfois plusieurs etapes complexes, il
sent with a small excess. On the other hand, for n'en demeure pas moins que ses signes exttieurs
ne laissent aucun doute. Les matidres polymeri-
the outer envelope, it would also be important
ques rtagissent avec l'oxygdne, puis elles de-
to consider the leak rate ofthe barrier film over
viennent de plus en plus fragiles. Il leur anive
a long period. This is shown for the films
m'Ame parfois de changer de volume ou de
described in Table Vtr. couleur, de devenir acides ou de liberer des
substances volatiles.

Table VII Il est ndanmoins possible de freiner ces effets,

Oxygen Permeation through Dffirent Films qui rdsultent de rtactions chimiques, en usant de
for a l0-Year Period diverses techniques
- en dliminant I'oxygdne, en
rdduisant la tempdrature, en captant les espdces
chimiques actives prdsentes dans le polymdre ou
en les dttruisant chimiquement avec des antioxy-
Film Permeation Ovel Permeation Over
dants ou des stabilisants, ou en bloquant la lu-
10 years', 50o/o RH 10 years', 0olo RH
(mL) (mL) midre (l' ultraviolet, tout particuliirement). Cette
Oxvbar 1 ol ol liste, qui n'est pas exhaustive, couvre nlanmoins
Oxybar 2 61 23 les pincipales options qui s'offrent ata
Oxybar 3 460 23 spicialistes de la restauration.
Oxvbar 4 ea
Bien qu'un certain nombre des ces mesures aient
. mL oxygen per 100 square rnches of barrier film, ttd dtudiees, on sait ntanmoins peu de choses du
per 0.21 atmosphere partial pressure difference for rythme rdel de dtgradation des matdriaux d'im-
portance historique. Or, pour tvaluer le degrd de
1 0 years at 23'C
rtussite de toute technique de conservation, il

360
semble indiqut de mesurer d'abord les tara References
d'oxydation, et I'on a d,ljd commenct d se servir
d'itudes portant sur des mesures de I'absorption l. Catalogue, Chromatographic Specialties'9 I
de I'oxygdne pour des tempiratures ambiantes (Brockville, Ontario, Canada: Chromatographic
et faibles. Specialties, 1991) p. 43.

Dans le cadre du pr,lsent article, nous consid6- 2. G.E. Pollock, D. O'Hara and O.L. Hollis,
rerons les r,lsultats de ces mesures en tenant
"Gas Chromatographic Separation of Nitrogen,
compte des diverses mdthodes de conservation,
Oxygen, Argon and Carbon Dioxide Using
dont la rdfrigtration et I'emploi de I'Ageless*,
Custom-made Porous Polymers from High
un absorbant d' oxygdne.
Purity Divinylbenzene," Journal of
*Ageless est une marque deposee de la Mitsubishi Chromatographic Science, vol. 22, 1984,
Gas Chemical Company Inc.
pp.343-347.

361
 A Field Trial for the Use of Ageless
in the Preservation of Rubber in Museum Collections

Yvonne Shashoua snd Scott Thomsen

Department of Conservation
The British Museum
London, U.K.

Abstract and returned to their storage areas. The same

number of objects were enclosed without Ageless
Rubber deteriorates when exposed to ultraviolet present to act as controls.
light, oxygen or ozone. The deterioration shows
either as cracking and embrittlement or as soften- Allobjects will be visualb, examined and photo-
ing and poor mechanical propenies, depending graphed regularly in order to monitor the rate of
on the environment. degradation.

Several approaches to the preservation ofrubber Introduction

have been investigated at The Bitish Museum
during the last two years. These have involved The Ethnographic Collections of The British
the use ofconsolidants, plasticizers and protec- Museum contain a number of rubber objects;
tive coatings. Since the results ofthese studies these may be composed solely of rubber or may
were not encouraging, it was decided to pursue a
incorporate wood, metals or textiles. The rub-
strategy based upon preventive conservation. The
ber is present in the form of natural rubber
main constraint was that the obiects concerned
bands, colowed balloons, rubber tubing or mo-
had to be kept in their existing storage areas. The
most useful approach appeared to involve sealing tor car tyres and inner tubes. Since rubber is
the objects into an oxygen-depleted micro envi- largely composed of unsaturated polymers, that
ronment created by the use ofAgeless, an oxygen is, polyisoprene, polybutadiene, and poly(sty-
scavenger. rene-co-butadiene), it is susceptible to oxida-
tion, which results in either softening or
Afeld fial was set up in the Museum's Ethnogra- hardening of the substrate. Attack by ozone
phy store to determine v'hether storing rubber ob- causes embrittlement and subsequent cracking
jects in an oxygen-depleted micro environment of the objects. Prolonged, uncontrolled oxida-
noticeably reduced their rate of deterioration tion and attack by ozone results in disintegra-
and, rf this was the case, whether such a proce- tion of rubber and serious damage to objects.
dure was practical. A laboratory trial involving
natural and accelerated ageing suggested that There are several techniques that can be used to
Cryovac BDF 200 pofuolefin batierfilm was protect rubber against degradation. Physical
most effective at excluding oxygen when made
methods that involve introducing a barrier be-
into heat-sealed bags incorporating Agel es s. En-
tween rubber and afinosphere include coating
closing rubber bands in this manner considerably
the surface with paraff,rn or microcrystalline
extended their usefut lifetime. Based on these re-
sults, selected ethnographic artefacts were en-
waxes or an epoxy paint. However, this
closed in oxygen-depleted micro environments method dramatically alters the appearance

363
of the object. Chemical antioxidants and antio- done into its efficiency at prolonging the life of
zonants, such as dialkyl p-phenylenediamines, organic artefacts by inhibiting the oxidation re-
have been introduced into rubber formulations action. For this reason it was felt appropriate to
since the 1930s. Black rubber, such as that used conduct such a study in The British Museum.
in tyre manufacture, has good resistance to oxi-
dation due to its ability to trap oxygen radicals. The aim of the study was to determine whether
However, it is extremely difficult to introduce storing rubber objects in an oxygen-depleted en-
these inhibitors into finished rubber objects. vironment, achieved using Ageless, noticeably
For this reason, it seems unlikely that the intro- reduces their rate of deterioration. The work
duction of additives shows much potential for falls into two parts; a laboratory trial was used
museum objects. Previous research projects at to select the most suitable materials and tech-
The British Museum have investigated the use niques for encasing rubber objects and monitor-
of consolidation to preserve the degraded rub- ing the rate of degradation, and a pilot scheme
ber surfaces by impregnation of weakened rub- that used the results of the trial to store a selec-
ber with plasticizers and by application of , tion of rubber objects from the Ethnographic
protective coatings to inhibit further attack.""" Collections. The findings of the laboratory trial
Although some of the materials used proved ef- will be the main concem of this paper.
fective at stabilizing the substrate, they tended
to alter the appearance ofthe surface to an unac- Degradation of Rubbers
ceptable degree. This was particularly notice-
able when the surface was pigmented. It was The mechanisms that govern the deterioration
felt that non-intervention in the method of con- of rubbers are complex. All chemically unsatu-
servation would be preferable. rated rubbers, synthetic as well as nafural, are
susceptible to oxidation by atrnospheric oxy-
Storage of rubber objects in an inert atmps- gen, a reaction that is geatly accelerated by a
phere has been recommended by Clavir.' This rise in temperature, the presence of metallic
would minimize the exposure of susceptible ob- ions, and by absorption oflight, especially in
jects to oxygen and ozone. However, the practi- the ultraviolet region of the spectrum. It is not
cality and cost of purging a large storage or thought that relative humidity is an imporNant
display area with an inert gas, and excluding factor at room temperature.
oxygen for a prolonged period would be pro-
hibitive. Two types ofageing processes. defined by
Buist,o are the most relevant to rubber objects
In March 1990, David Grattan reported that stored in museums. Shelf ageing, which occurs
oxygen-depleted environments could be effec- at ambient temperatures in the dark, is autoxida-
tively produced by enclosing a chemical oxy- tion. Oxidation is carried out via a chain reac-
gen scavenger known as Ageless*, i4 a sealed tion mechanism that involves free radicals
oxygen-impermeable film envelope.- formed from attack on the rubber hydrocarbon.
Free radicals are highly reactive molecular frag-
Ageless has found widespread application in ments that have one unpaired electron available
the food industry for prolonging the shelf-life for bonding. They can react with oxygen to pro-
of dry foods and wines. The major ingredient of duce peroxy radicals by which the oxidation
Ageless sachets is finely divided active iron, process fuels itself. Ketones, aldehydes and
which forms iron ox-ides and hydroxides on ab- acidic groups are likely to be formed as by-
sorption of oxygen.5 It is claimed to reduce the products of the oxidation process. Shelf ageing
oxygen concentration of an air-tight container can result either in hardening or softening of
down to 0.01% (100 ppm) or less. Ageless has rubbers depending on their formulation and on
been used in museums for the treatment of in- the conditions of storage.
sect infestations, but little research has been
Atmospheric cracking is the term applied to the
reactions that occur when rubber artefacts are
*Ageless is produced by the Mitsubishi Gas exposed to ultraviolet light. In addition to the
Chemical Company Inc. of Japan. oxidation mechanism described previously,

364
ozonolysis, or attack by ozone, takes place. Fe Fe"- + 2e'
Ozone is formed in the atmosphere by the ->
chemical reaction between atomic and molecu- t/zoz + Hzo + 2e-
lar oxygen. It is now thought that a concentra- ->2oH-
tion of less than five parts of ozone per hundred Fe2* + 2 (oH)- Fe(oH)z
million is sufficient to cause ozonolysis.'
Fe(OH)z +t/cOz +t/zHzO
->
Ozone reacts with non-stretched rubber by add- ->Fe(OH):
ing to the double bonds to form ozonides (Fig- Different types of Ageless are available depend-
ure l). When all the surface double bonds are ing on the water activity of the material to be
consumed the reaction ceases. A grey film or packaged. Ageless Z is recommended for the
frosting, l0 to 40 molecular layers thick, may preservation of material that has a water activ-
appear on the surface. If this film is disrupted ity (humidity divided by 100) of 0.85% or less,
by stretching the rubber, unsaturated molecules and so it is the most suitable grade for use with
are exposed and the energy required to promote rubber. Ageless is also available in different
tearins and further crack srowth becomes avail- sizes depending on the volume of oxygen to
able.8-Cracks form at righ"t angles to the surface be scavenged. Ageless Z-zUl,which absorbs
of the rubber and eventually cause failure. 200 mL of oxygen from 1 litre of air, was used
in the laboratory trial.
R R''
Selection of Packaging Materials
C= C +Os
Five main criteria were used to select a suitable
R' R''' packaging material.

R R'' o The material must be sufficiently stable so

\/ that it does not accelerate degradation ofany
c-c
/\,/\
R' Os R"'
component ofthe enclosed object, or lose
cohesiveness during the period ofthe trial.

R' R" o The material must be highly impermeable to

\ oxygen, and should buffer the object against
C=O + C=O
/ extreme changes in relative humidity.
R' R"'
o The material must offer mechanical
and other speoes Includrng polymeric
peroxides and hydroperoxides protection to the object.

o The material must be transparent so that

Figure I Attack by ozone on rubber. study of the artefacts during the trial is
possible without removing the object from its
Laboratory Evaluation of Ageless in microclimate.
the Preservation of Rubber
o The material must be amenable to resealing
Selection of Ageless in order to facilitate rezular examination of
Ageless is manufactured in the form of small sa- test pieces.
chets containing an oxygen absorbent that is
based on finely divided iron. The iron is sealed The materials that were considered to have suit-
within several layers of a gas-permeable plastic able properties and to be readily available were
film that controls the rate offlow ofoxygen glass in the form of air-tight Kilner preserving
and moisture to the absorbent. Iron oxide and jars, poly(vinylidene chloride) (PVDC) film,
hydroxides are formed on reactign with oxygen self-closing polyethylene sample bags and
via the fol lowine half-reactions.e Cryovac BDF 200 film. Cryovac BDF 200 film
is a thin, multi-layered co-extruded polyolefin

36s
formulation, which is supplied for use with can be converted into CIE L* a* b* colour co-
Ageless oxygen absorber. ordinates using a computer programme, where
L* is a measure of the lightness of the sample,
In order to predict the chemical stability of the ax relates to redness/greenness, and b* deter-
packaging materials under their prgBosed condi- mines yellowness,/blueness.'' None of the mate-
tions of use, Oddy tests were used.'' Samples rials tested appeared to yellow after light- or
of polyethylene bags were wrapped around cou- heat-ageing. However, colour measruements
pons offfeshly cleaned silver, copper, and lead. suggest that Cryovac BDF 200 had a higher de-
The test pieces were sealed in glass tubes with gree ofcolour stability, the greatest change in
ground-glass stoppers and aged for 28 days at colour occurring in the b* (yellowness,/blue-
60'C. In order to assess any adverse reactions ness) value (Table I). The breaking strengths of
between polymer films and rubber, an adapta- aged and unaged strips of each film were deter-
tion of the Oddy test was carried out. A strip of mined in order to assess the suitability of the
polyethylene film was wrapped around a piece packaging materials to maintain a physical bar-
of motor car inner tube and a rubber band, be- rier between the objects to be enclosed and the
fore sealing in a glass tube and ageing, as had environment of the store. Dumb-bell-shaped
been done with the metal coupons. The proce- test pieces were conditioned at 650/o RH and
dure was repeated with PVDC and Cryovac 22"C for two hours prior to tensile testing to
BDF 200. After ageing, all metal coupons and breaking point on a J.J. Lloyd tensile tester.
samples of rubber were examined by optical mi- This procedure was carried out in accordance
croscope for evidence of corrosion or other de- with the methods described in British Standard
terioration. None was detected and it was 27 82 P art 3, lg7 6.t2 Elongation and breaking
concluded that all the films were sufficiently strength values for aged and unaged films were
stable for use in the trial. read from the trace.

In order to investigate the colour stability and

retention of mechanical strength of the poten- Table I
tial packaging materials, samples of PVDC, Colour Measurements on Aged Packaging
polyethylene sample bags and Cryovac BDF Materials for Rubber
200 films were light- and heat-aged. Glass
was assumed to be sufficiently stable for the
Material Colour Difference
(Delta E)
trial. Films were light-aged for 28 days in a
Microscal light-fastness tester fitted with an Cryovac BDF 200
Heat-aged 0.04
MBTL 500 bulb. This exposed the films to an Lrght-aged 0.07
emission spectrum similar to that of daylight
Polyethylene
but approximately 100 times more intense.
Heat-aged 0.11
Heat-ageing was effected by sealing films in Lrght-aged 1.17
a test-tube with a ground-glass stopper, main-
PVDC
tained at 70oC for 28 days. Aged and unaged Heat-aged 0.42
films were examined by ultraviolet/visible spec- Lrght-aged 209
troscopy and by human eye in order to deter-
mine the extent of discolouration of the films, Delta E = [ (Delta L')2 + (Delta a")2 + (Delta b.)2 ]1'?

and were tensile tested to quantify their loss in where L' = lrghtness
strength with age. a'= rednesJgreenness
b* = vellownesdblueness
Ultraviolet reflectance spectroscopy is a tech-
nique by which the percentage of light reflected
PVDC lost approximately 27o/o of its elonga-
by the sample at each wavelength in the visible
spectrum can be monitored. A change in the tion after light- and heat-ageing (Table II) com-
shape ofthe spectral trace produced indicates a
pared with a loss of almost 5% by Cryovac
change in hue and lightness of the material, the BDF 200 andTo/o by polyethylene. Breaking
technique being more sensitive than the human stress did not alter geatly on ageing. From
eye to small changes. Reflectance spectra data these results it was concluded that either

366
 vacuum pump. The pipette was withdrawn after
Table II flushing or evacuation, and the last side ofthe
Mean Elongation at Yieldfor Rubber bag was sealed. The time taken to heat-seal the
Packaging Materials films was critical. If the heat sealer was left in
contact with Cryovac BDF 200 for more than
Elongation (mm) two seconds, the frlm shrank, causing pinholes
and tearing of the film. Kilner jars were evacu-
Cryovac
BDF200 PVDC Polyethylene ated or flushed by placing the jar in a self-
seal polyethylene bag and using the pipette
Unaged 44.9 37.0 27 I technique. The jar was closed while inside the
Heat-aged 44.7 25 8 26 2 bag. Ageless was put in some containers.
Lrght-aged 44.6 27 0 25 6
The time required for sufficient oxygen to pene-
trate the container (or inner bag where double
bags were used) to change the colow of the
Cryovac BDF 200 or polyethylene films would Ageless Eye from pale pink to dark blue was
be suitable for the field trial. recorded using a stopwatch (Table III).

Selection of Sealing System for Orygen The results suggested that all containers' sys-
Impermeable Films tems maintained their low oxygen environment
In order to evaluate the most effective method for longer when they were purged with nitrogen
for creating an oxygen-depleted environment in than ifthey were evacuated or left unpurged.
which to store a museum object, a series of en- Flushing with nitrogen prolonged the life of
casing methods were examined. Enclosures Ageless and did not produce the closely fitting
were prepared using polyethylene self-sealing packaging associated with using a vacuum,
bags, poly(vinylidene chloride) film, glass Kil- which can cause physical damage to a friable
nerjars and Cryovac BDF 200 film. Bags were object.
prepared from PVDC and Cryovac BDF 200
films by heat-sealing the edges using a domes- Enclosing the inner, object-containing bag in a
tic heat-sealing device. Double seams with similar outer bag, also containing one sachet of
reinforcement seams at each corner were found Ageless, was found to prolong the useful life-
to produce the strongest and most leak-resistant time of the oxygen absorber, reducing the rate
bags. A sachet of Ageless 2-200 was placed in of oxygen diffusion into the microclimate in
the bag or Kilner jar and, in the case of double which the object was to be stored. Cryovac
bags, a sachet was added between each layer. BDF 200 proved to be the most effective bar-
rier film from which to prepare bags. It proved
The effectiveness of the containers to maintain most resistant to oxygen diffusion under all test-
an oxygen-depleted atmosphere was deter- ing conditions. The Ageless contained within
mined by measuring the time required for Age- double Cryovac BDF 200 bags, which had been
less Eye, the oxygen indicator supplied with flushed with nitrogen, was still active after six
Ageless, to change from pale pink (less than months, the length of this trial. Kilner jars also
0.1% oxygen) to dark blue (greater than0.So/o provided an excellent barrier to oxygen, but
oxygen). Ageless Eye is reusable. On moving would not be practical for large objects.
from an oxygen-rich to an oxygen-free atmos-
phere it regains its pale pink colouration after Many of the ethnographic objects in the collec-
two to three hours at 25'C. Bags and Kilner tion contain wood, textile or metal components
jars were either left unpurged (controls), in addition to rubber. These secondary compo-
flushed with oxygen-free nitrogen or vacuum nents are more sensitive to extreme changes in
packed. Bags were flushed or evacuated via a relative humidity than rubber, and may undergo
glass Pasteur pipette, which was inserted before dimensional changes resulting in the formation
the bag was heat- or self-sealed, and attached ofundesirable stresses on the object, or in the
via silicon tubing to a nitrogen cylinder or acceleration of corrosion in the case of metals.

367
 Table III
Efficiency of Enclosure Systems for Preventive Storage of Rubber

Barrier Material Single (S) Atmosphere Time for Ageless

or Double (D) Eye to Detect
Layer Oxygen (Hours)

Polystyrene Box Air 0.04

Cryovac BDF 200 5 Air 24.61

D Au 72.05
e Normal 1M.16
D Normal 4336+
S Vacuum 96.02
D Vacuum 4336+

Polyethylene D Air 0.14

D Air 48.32
S Normal 96.77
D Normal 100.10
S Vacuum 24.O3
D Vacuum 24.18

PVDC s Air 6.19

D Air 48.45
J Normal 24.52
D Normal 120.37
s Vacuum 53.21
D Vacuum 25.37

Kilner Jar Air 4336+

Normal 4336+

Since the Ethnography store is humidity con- 40oC + loC for one hour. At the end of this
trolled to55o/o * 5o%, under normal circum- period the bags were returned to the laboratory
stances there should be little cause for concem. environment, the dial hygrometer readings re-
However, if the enclosed objects were to be corded, and the time for the relative humidity
exposed to a dramatic rise in temperature then to equilibrate with the surroundings noted.
it would be possible for condensation to form
inside the bags. The results suggested (Table IV) that relative
humidity increased in all the bags, and took
In order to assess the effects of a rapid rise in longest to return to laboratory conditions in the
temperature on the relative humidity inside a PVDC enclosure, and least time to equilibrate
bag containing an oxygen-depleted atrnosphere, in the polyethylene bag. This may be attributed
bags prepared from single layers of Cryovac to the very low moisture yapour transmission
BDF 200 and PVDC and self-seal polyethylene value of PVDC (1.5 g m-'to 5.0 gm-" per 24h,
bags were examined. A sachet of Ageless for 25 pm films at 90% RH and 38'C) coln-
2-200 and a dial hygrometer were placed in pared w.ith that for polyethylene ( l5 g m-' to
the bags prior to sealing in the laboratory 20 grn" per 24h), which controls the rate of
(20'C + 2"C and 60% RH t 5% RH). The escape of excess moisfure. " These results indi-
prepared bags were placed in an oven at cate that if large variations in relative humidity

368
 Light-Aged Rubber Band (mag. x 2000)

Table IV
EJfect of Exposing Packaging Bags to Sudden
Temperature Increases

Packaging Material

Cryovac Polyethylene PVDC

RH after one hour 77 70 80

at 40"C (1%)

Time for RH to 126 60 135

return to 607" 1570
(minutes)

Light-Aged Rubber Band Plus Ageless (mag. x 2000)

are of concem, buffering agents, such as silica
gel, could be incorporated into the bagging sys-
tem. However, it should be remembered that
some moisture is required to activate Ageless
to absorb oxygen.

Relating Accelerated Ageing to

Natural Ageing
In order to quantify the effectiveness ofenclos-
ing rubber objects some lirther laboratory trials
were made. Thick (l cm wide) rubber bands
were washed in the non-ionic surfactant
Synperonic 'N' in order to remove any separat-
ing agents, such as talc. They were then Figure 2 Scanning electron micrograph ofozone
stretched on a polystyrene former so that their cracking on a rubber band after light-ageing
length increased by 50%. (normal atmosphere) compared to one aged in an
oxygen-depleted atmosphere, created with Ageless.

The rubber bands, still sfretched on the formers,

were sealed in double bags made fiom Cryovac SEM examination (Figure 2) showed severe
BDF 200. One sachet of Ageless and an Age- ozone cracking to be present on the surface of
less Eye was placed in each bag, and bags were the rubber band that had been light-aged under
either flushed with nitrogen or left unpurged. open conditions, compared with the unchanged
The enclosed rubber bands were light-aged un- surfaces ofa band that had been lighfaged in
til the first signs of deterioration were evident. an oxygen-depleted atmosphere. The results of
Samples of rubber bands stretched in a similar breaking stress measurements suggested that en-
manner but left unbagged were also light-aged closing the rubber bands in an oxygen-depleted
in order to compare the rates ofdegradation be- atmosphere significantly reduced the loss in
tween bagging and open storage. Stretched sam- strength caused by ozone attack. Optimum pro-
ples were also naturally aged. tection was afforded by flushing the Cryovac
BDF 200 bags with nitrogen prior to sealing.
The first signs ofdeterioration occurred after The naturally aged band showed the first signs
336 hours. After this period of exposure, all ofdegradation after 2,208 hours (92 days)
bands were examined by Scanning Electron compared with 130 hours for an openly light-
Microscopy (SEM) and their breaking sffess aged band. Rubber bands light-aged in oxygen-
determined using tensile testing. The procedure depleted atmospheres have not shown visible
detailed in British Standard 903 Part A21989 signs of deterioration to date, six months into
was followed. the trial.

369
Preventive Storage of Ethnographic oxygen-depleted conditions were estimated in
Objects in an Oxygen-Depleted Atmosphere order to calculate the number of sachets of Age-
In the stores, The British Museum Ethno- less Z required in each case. Very small and
graphic Collections are arranged by geographic fragile objects were placed in polystyrene
origin rather than on the basis of materials. boxes for mechanical protection prior to enclos-
Hence the rubber objects are to be found in nu- ing in Cryovac BDF 200 bags. Since polysty-
merous locations. Objects are kept in wooden rene is highly permeable to oxygen, this would
drawers or boxes on a racking system and are not reduce the exposure ofthe object to oxy-
only exposed to daylight when being studied or gen. Bags were prepared as described earlier,
on display in the Museum. The store is humid- but using a larger commercial heat sealer in
ity conholled (5 5% ! 5%). place of the small domestic equipment used for
the laboratory trials.
The objects containing rubber were identified
from the computerized catalogue of the Ten of the objects were enclosed in double
collections. The conditions of the rubber arte- bags, each ofwhich contained the appropriate
facts were assessed. The earliest object was col- number of Ageless Z sachets and an Ageless
lected in 1855 and predates the discovery of Eye indicator. Each bag was flushed with oxy-
vulcanization. The collections are still growing gen-free nitrogen prior to heat-sealing. The con-
and consist mainly of toys, sandals and musical centration of oxygen was determined by the
instruments. Those objects made from the side- colour of the Ageless Eye. The remaining ob-
walls of motor car tyres, such as sandals, were jects were enclosed in double bags containing
in excellent condition, while those comprised air. All objects were retumed to the store.
of thin strips of inner tube shetched tightly
around a wooden base, as in a child's catapult, The bagged objects have now been put back in
showed signs of deterioration and of bloom. the Department of Ethnography store and they
Objects that contained copper wire, for exam- will be examined after three months using the
ple, a cage for singing birds, were in a more optical microscope and photography. Examina-
advanced state of degradation. This may be tions will be repeated every six months.
attributed to the presence of the copper. which
is known to catalyse attack by ozone.'- Conclusion

Twenty objects representative ofthe technol- The technique of using oxygen-depleted

ogy and types of rubber in the collections were microclimates to inhibit long-term degradation
selected for use in the pilot scheme. They mechanisms for museum objects is new to The
ranged in types and levels of degradation. Half British Museum. It is a passive method of con-
ofthe objects were selected for storage under servation that neither precludes the use ofother
oxygen-depleted conditions and half were treatrnents nor requires special materials. For
stored in air so that comparisons of the rates these reasons the curatorial staffhave been ame-
and extents of degradation could be made. All nable to its use. Laboratory trials have shown it
objects were photographed using colour film to be a relatively straightforward and inexpen-
and black-and-white film in order to record sive method of prolonging the useful life of rub-
their condition prior to the trial. Pencil draw- ber bands. Rubber objects selected from the
ings of all objects were made in order to high- Ethnographic Collections have recently been
light details of deterioration, and micrographs heat-sealed in bags made from Cryovac BDF
were taken ofthe areas ofconcem using a 200 film, which were flushed with nitrogen in
reflectance microscope. order to prolong the life ofthe oxygen ab-
sorber, Ageless. Their physical condition will
Laboratory trials showed Cryovac BDF 200 to be monitored regularly by optical microscopy
be the most stable and efficient barrier film to and photography in order to assess the effective-
oxygen, and heat-sealing to be the most conven- ness of this technique. The effect of enclosing
ient method of sealing the bags made from the these objects and thus preventing accessibility
film. The volumes of the obiects to be stored in for study and handling, particularly when they
are required in the deparrnental student room,

370
will also be evaluated. If successful, the use of Polystyrene Boxes
oxygen-depleted microclimates may be applied
to a wide range of materials stored in museums. Stewart Plastics
Purley Way
Acknowledgement Croydon CR9 4HS Tel:081-686 2231
United Kingdom
The authors thank Paul Aspill of W.R. Grace
Ltd., for kindly providing samples of Ageless, R6sum6
Ageless Eye and Cryovac BDF 200. They ap-
preciate the time taken by John Loadman and L'essai pratique de I'Ageless* pour la
the staff at the Tun Abdul Razak Laboratory to conservation du caoutchouc duns les
discuss the analysis, deterioration of rubber, collections de musie
and developments in the rubber industry. They
also thank their colleagues, Dr. John Mack, Le caoutchouc se ddgrade lorsqu'il est exposd
Keeper of the Department of Ethnography for aux ultraviolets, d I'oxygdne ou d I'ozone. Cette
permission to carry out the trial, John Osborn ddttioration se traduit, suivant le milieu, soit par
for assistance with locating the objects, Susan son craquellement et safragilisation, soit par son
Bradley and Dr. Vincent Daniels for discussing ramollissement et par une perte de proprittbs
the work and the paper, and Andrew Oddy, mdcaniques.
Keeper of Conservation, for permission to
Au cours des deux dernidres anndes, le British
publish the paper.
Museum a ttudi6 diverses mithodes pour assurer
la conservation du caoutchouc, qui reposent no-
Suppliers tamment sur l'emploi d'agents de consolidation,
de plastifiants et de rev€tements protecteurs. Les
Ageless, Ageless Eye and Cryovac BDF 200 rdsultats de ces 2tudes n'ont toutefois pas ttt trds
encourageants, et on a donc choisi de poursuivre
W.R. Grace Ltd. une stratdgie davantage axde sur la conservation
Northdale House prdventive. La principale contrainte venait dufait
North Circular Road Tel: 081-965 061I que les objets devaient €tre laissds d l'endroit
LondonNWlO 7UH Fax:081-961 6480 m€me oi ils dtaient mis en nlserve. Parmi les
United Kingdom techniques qui ont etd envisagdes, la plus utile
semble €tre celle qui sefonde sur la crdation d'un
Poly(vinylidene chloride) Film (Safeway micro-environnement autour des objets grdce d
Non-PVC Cling Film) de I'Ageless, un dtsoxyginant.

On a donc effectut les preparatifs en vue de

Safeway (Head Office U.K.)
mener, dans la rtserve d'objets ethnographiques
6 Millington Road
du musde, un essai pratique qui permettrait,
Hayes
d'une part, d'etablir si l'entreposage des objets
Middlesex UB3 4AY en caoutchouc dans un milieu sans oxygine
United Kingdom contibue ffictivement d ralentir defaQon mar-
qude le rythme de leur dttdrioration et, d'autre
P o lye thy lene Se lf- Seal B ags part, d'ivaluer I'incidence d'une telle technique
sur le fonctionnement de la r,lserve en particu-
Supreme Plastics Ltd. -
lier, lorsque les objets doivent €tre utilisds dans
Supreme House la salle des ttudiants du ddpartement.
Vale Road
Harringay Les objets et les sachets d'Ageless ont itd ddposds
London N4 lQB Tel:081-802 4202 dans des boites en polystyrdne transparent, puis
United Kingdom scellds dans des sacs htanchesfaits d'une pelli-
cule de polyoldfine. Pour cet essai, 50 objets de
la collection ont ttd retenus. Aprds les avoir ainsi

371
 sceiles, on a cherch1 d ilaborer une technique 4. Grattan, D.W., personal communication to
qui permettrait de dtceler leur ddgradation. On Dr. V. Daniels. March. 1990.
a alors puisd d diverses m,lthodes, au nombre
desquelles figuraient, parmi celles qui apparais- 5. Ageless Ox,vgen Absorber: A New Age in
saient comme les plus prometteuses, I'examen Food Preservation (Tokyo: Mitsubishi Gas
d I'rBil nu ou au microscope classique ou dlec- Chemical Corporation, I 987).
tronique d balayage quoique cette derniire
-
mdthode exige le prdl'ivement d'dchantillons, ce
6. Buist, J.M, Ageing and Weathering of
qui ne pourrait pratiquement pas sefaire dans le
Rubber (London: W. Heffer and Sons Ltd.,
cas d'objets de musde la spectromAtie in-
d transformde
- deetFouier (IRTF). Les 1955) p.24.
frarouge
observations faites lors d'expdiences avec des
ichantillons de caoutchouc scellds suivant la 7 .Layer, R.W. and R.P. Lattimer, "Protection

mdthode susmentionnie et expostis d une lampe of Rubber Against Ozone," Rubber Chemistry
donnant le spectre de la lumiire visible, dans un and Technolog,, Rubber Reviews No. 3,
dispositif de vdrification de la rdsistance d la July-August 1990.
lumi'Dre, permettront d'etablir la mdthode qui
convient le mieux pour un musde. Ces tchantil- 8. Lewis, P.M., "Protecting Natural Rubber
lons serviront en outre de groupe ttmoin lors de Against Ozone Cracking," NR Technologlt,
I'essai pratique, et des hchantillons scellLs avec Rubber Developments Supplement P art l,
de I'Ageless et non scell4s avec ce produit seront 1972,p.2.
placts dans la rdsente.
9. Nakamura, H. and J. Hoshiro, Techniques
L'ttude des mtthodes d utiliser pour sceller les the Preservation of Food by Employment of
objets et pour surveiller le diroulement de I'essai
for
an Oxygen Absorber Sanitation Controlfor
sera mente d terme, et I'on connaitra les risultats
Food Sterilizing Techniques (Tokyo: Sanya
des essais de vieillissement acctldrd. Un rapport
Publishing Co., 1983).
intirimaire sur le ddroulement de I'essai sera
prdsent6.
10. Oddy, W.A., "An Unsuspected Danger in
*Ageless est une marque dtposie de la Milsubishi Display," Museums Journal, vol. 73, June
Gas Chemical Company Inc. 1973,pp.27-28.

References I l. Billmeyer, F.W. and M. Saltzrnan,

Principles of Colour Technolog,t (New York:
1. Shashoua, Y., "Inhibitive Treatments for Wiley Interscience, 1966).
Rubber," Newsletter of the ICOM Working
Group on the Conservation of Modern 12. British Standard, Tensile Strength,
Materials,July 1991. Elongation and Elastic Modulus of Plastics
British Standard 2782Part 3. Methods 320A-F.
2. Shashoua, Y., "Evaluation of Adipate Plasti- r976.
cizers and a Protective Coating as Inhibitors for
Natural Rubber," The British Museum Internal 13. Bristow, J.H., Plastics Films (London:
Report, No. VII33, February 1989. Longman Group Ltd., 1974).

3. Clavir, M., "An lnitial Approach to the 14. Naunton, W.J.S., The Applied Science of
Stabilization of Rubber from Archaeological Rubber (London: Edward Arnold Ltd., 1961)
Sites and in Museum Collections," Journal of p. 103.
the llC-Canadian Group, vol. 7, nos. 7 and2,
1982, pp. 3-10.

372
La mise au point d'un traitement cathodique de stabilisation
de vestiges a6ronautiques immerg6s en alliages d'aluminium

Christian Degrigny

Groupe Valectra
Direction des etudes et recherches
Electricitd de France
Saint-Denis
France

R6sum6 phtnomdne d' alcalinisation local au voisinage

de la sudace est impossible; la corrosion pro-
Afin d'dviler la degradation d I'atmosphdre de vient donc de la modifcation des propriLtis de
piices en alliages d'aluminium provenant de la couche passive d'alumine sous polarisation
vestiges aironautiques immergds depuis une cathodique. Le r6le majeur des prtcipitds inter-
longue ptiode en eau douce, un traitement de mitalliques dans I'initiation de la corrosion et
conservation a ttd mis au point. Il consiste en leur htolution au cours des processus ont 6td mis
I'extraction d'ions dangereux pour le mital, en tvidence.
sous po lari sati on cath odique.
On a pu ainsi ddterminer, en solution d'ions
Dans un premier temps, des tests de conosion citrate, un ensemble de conditions d respecter au
acctl4rie ont permis d'dtablir que, en I'absence cours du traitement pour dviter des corrosions
de traitement, les alliages d'aluminium prdcor- inddsirables (dbsadration et agitation de la solu-
rodds subissent une dtgradation rapide au cours tion, maintien d'un pHfaiblement acide et du
de leur remise d I'air libre. Les ions chlorure, potentiel cathodique au-dessus de la zone de cor-
contenus dans les matdriaux, sont responsables rosion cathodique). La cinttique d'extraction des
de l'accdl6ration de la corrosion, qui est en ions chlorure sous polarisation a alors ttd dtu-
outre activde par la prdsence de produits de dibe et les paramdtres ont eti optimisds afin
corrosion instables. d'acc,ll6rer le traitement. Un pritraitement com-
plexant, assurant une meilleure fficacitt d'ex-
L'extraction des iow chlorure est rdaliste en traction, a 6ti dgalement considdri.
solution tamponnte sous polarisation cathodique
d potentiel imposi, mais des corrosions assocides L'application du procddd d des pidces rdelles
peuvent aussi se produire. On pourra ainsi noter, provenant d'ipaves adronautiques a donnd des
entre autres, I'existence d'une corrosion par rtsultats reproductibles. Ainsi, I' extraction des
piqfrres, due aux ions chlorure extraits, ainsi ions chlorure sous polarisation cathodique est
qu' une conos ion cat hod ique. rapide et fficace. Des traitements adaptds, pour
les cas oi des alliagesferreux sont associ4s aux
Le domaine de piqfiration a dtt ddfini d partir des alliages d' aluminium, sont dgalement proposd s.
valeurs caracttristiques des potentiels de piqAre
et de I'adration du milieu. Introduction
Les micanismes et la cindtique de la conosion Une quantit6 non n6gligeable d'6paves a6ro-
cathodique ont 6td itudids en ditail au moyen nautiques se trouvent immerg6es en eau de mer
d'essais potentio-statiques mends avec une micro- ou en eau douce. Or ces pidces immerg6es
cellule d'observation. En milieu tamponnL, tout pr6sentent parfois une valeur historique

3t)
ind6niable, et devraient donc 6tre en partie ou Les alliages d'aluminium sont aussi vari6s par
totalement remont6es si l'on veut qu'elles Ieur composition que par les formes de corro-
retrouvent la place qui leur revient dans la sion qu'ils pr6sentent. Dans le secteur a6ronau-
m6moire du public. Tel est effectivement le tique, les alliages les plus courants sont sans
cas, par exemple, des hydravions de l'6tang de doute ceux qui contiennent du cuivre comme
Biscarrosse (dans les Landes, en France), qui principal 6l6ment d'addition. Leurs trds bonnes
sont les symboles d'une 6poque r6volue de propri6t6s m6caniques, connues dds les anndes
I'histoire de l'a6ronautique. 1912-1913, en ont fait les alliages id6aux pour
les fuselages et les structures. Ils sont toutefois
Souvent m6talliques, ces vestiges sont, pour trds sensibles aux diff6rentes formes de corro-
la plupart, constitu6s d'alliages ferreux et d'al- sion, de sorte qu'ils sont recouverts d'un pla-
liages d'aluminium. Durant leur p6riode d'im- cage d'alliage plus r6sistant aux agressions
mersion, des corrosions intenses par contact ext6rieures. Au nombre des autres alliages ha-
entre ces m6taux se sont 6tablies, aux d6pens bituels, figurent ceux qui contiennent, comme
des alliages d'aluminium. Et puisque ce 6l6ment d'addition, du mangandse ou du
ph6nomdne est 6volutif, il remet en cause la magn6sium, ou encore une combinaison de
conservation in situ et d long terme de ceux-ci, magn6sium et de silicium; ces mat6riaux
et donc des pidces qui en contiennent. peuvent; du fait de leur meilleure r6sistance
ii la corrosion, €tre utilis6s d diverses fins bien
Des tentatives de nettoyage ont 6t6 men6es afin pr6cises (voir le tableau I).
de mettre un terme ir ces corrosions irr6versi-
bles mais elles ont toutes r6v6l6 que, sans traite- Aprds un s6jour prolong6 en milieu subaqua-
ment ult6riew, les pidces en alliages d'alumi- tique, ces alliages pr6sentent, au moment de
nium continuent de se d6grader i I'atmosph.ere, leur sortie de I'eau, des 6tats de surface plus ou
habituellement dans les zones pr6corrod6es'. moins diffdrents. Ainsi, les alliages aluminium-
cuiwe auront subi des corrosions intergranu-
Contact6 i ce propos par le mus6e de I'Air du laires et par piqtres celles-ci 6tant profondes
Bourget (Paris). le groupe Valectra d'Electricite -
et de type transgranulaire tandis que les al-
de France, sp6cialis6 dans les techniques 6lec- -, et aluminium-
liages aluminium-mangandse
trolytiques de conservation-restauration, s' est magn6sium-silicium n'auront 6t6 sensibles
charg6 de proposer une proc6dure de haite- qu'd des formes de corrosion par piqtres de
ment. La nocivit6 des espdces chlorur6es pr6- type h6misph6rique.
sentes dans les alliages d'aluminium ayant 6t6,
d6montr6e, Valectra a opt6 pour une d6chlorura- La couche de corrosion est particulidrement
tion de ces alliages sous polarisation cathodi- complexe sur les alliages aluminium-cuiwe.
dique. Puisque certaines formes de conosion Elle est form6e, en fait, d'une double couche :
associ6es d ce fraitement pouvaient se produire une couche adh6rente, de faible 6pais-
au cours de I'op6ration, des conditions de pro- seur (0,2 pm), et une couche couwante, sur
tection ont 6t6 d6finies, et une optimisation des laquelle s'dtendent des produits de corrosion
paramdfres exp6rimentaux .a permis d'assurer issus de I'action de l'environnement (voir la
un ffaitement plus efficace'. figure I ). On note ainsi la pr6sence de sulfates
d l'int6rieur de cette dernidre couche, compos6e
La description des alliages et le choix essentiellement d'alumine hydrat6e AI(OH):.
du traitement Cette couche de corrosion est de plus pollu6e,
d I'ext6rieur, par du fer (ir cause de la pr6sence
Toute intervention visant la conservation de pidces en acier d proximit6) et par des ions
future d'un objet d6grad6 n6cessite une 6tude chlorure mais surtout, dans le fond des piqtres,
qui, ax6e sur les causes de la corrosion initiale, par du cuiwe et, ld 6galement, par des ions
permettra de mieux pr6voir son 6volution chlorure. Quant aux autres alliages, on ne
ult6rieure, lorsque l'objet sera expos6 trouve, sur la surface m6tallique, que des pus-
d I'atmosphdre. flrles d'alumine hydrat6e recouvrant les piq0res
qui se sont form6es, au fond desquelles des

374
 touch6e. Ce ph6nomdne se d6veloppe, en fait,
d partir des zones pr6corrod6es, sous la couche
d'alt6ration et sutout dans le fond des piqtres,
ld oi il y a une quantit6 importante d'ions
chlorure. La pr6sence de nombreux produits
de corrosion dans le cas d'un alliage alu-
minium-cuivre, - par exemple peut accentuer
les processus. -
Il est donc n6cessaire d'appliquer un
traitement de conservation aux alliages
d'aluminium pr6corrod6s. Pour extraire
rapidement les ions chlorure ins6r6s dans les
mat6riaux, il est propos6 d'avoir recours d
Figure I : La nature de la couche de coruosion la polarisation cathodique, mais on pourra
prdsente d la surface d'un alliage aluminium-cuivre aussi envisager des dissolutions partielles ou
immergt depuis une longue pdriode en milieu globales des produits de corrosion.
subaquatique.
Les conditions du traitement
PI Silicate de calcium et d'aluminium et Al(OH)t
polluts (Cu, Cl) En se r6f6rant au diagramme d'6quilibres
P2 Al(OH)z pollud (S)
potentiel-pH de l'aluminium en eau pure
P3 Peinnre
donc danp des conditions id6ales
- paret
P4 Sulphate fenique, silicates et Al(OHfu polluds -d6fini
(Fe, Cl) Pourbaix3, on constate que, si le pH se situe
P5 Concritions amorphes entre 4 et9,la surface m6tallique est prot6-
g6e (voir la figure 2). Or,lors du traitement
ions chlorwe peuvent encore €tre trds actifs. cathodique, dans cette zone de pH, d'alliages
Des tests de corrosion acc6l6r6e, en atmosphdre d'aluminium d6grad6s et pollu6s par des ions
naturelle d Biscarrosse. ont 6t6 men6s avec des
chlorure. deux formes de corrosion peuvent
apparaitre, ir savoir :
6chantillons d6grad6s; ils visaient d obtenir des
dorur6es sur la stabilit6 d I'atmosphdre de ces
o la corrosion par piqffres, due aux ions
6tats de corrosion d long terme. Ils ont tous
monh6 que l'6tat de corrosion initial 6volue chlorure extraits durant le traitement, qui
plus ou moins vite d I'atmosphdre, et que toute peuvent se refxer en surface;
l'6paisseur peut, dans les cas exfr€mes, 6tre

Tableau I
Les principales familles d'alliages trouvts dans les vestiges adronautiques

Systeme 56rie Teneur de l'6l6ment Caract6ristiques et utilisation

d'addition principal
Al non allr6 lXXX farbles propri6t6s m6canrques
oeu uthses
Al-Cu 2XXX Cu.2it6% recouverts d'un placage d'Al l XXX
(fuselaoes)
Al-Mn 3XXX Mn 0,5 4 1,5 % pour allrages conoy6s
Mn 4"/" pour allraqes moul6s
AFSr 4XXX Si i 3 d22o/" pour allrages moul6s (moteurs et appareils
de bord)
AFMg 5XXX Mg:0,547ol. pour structures el fuselages

Al-Mg-Sr 6XXX M9.0,541,5% pour structures et luselages

Sr .0,541,5%

375
 des 6chantillons non corrod6s, d partir de
aia
deux paramdtres : le potentiel d'amorgage
des piqfires (Ep), caract6ristique du processus
i..----
d'initiation, et le potentiel de corrosion (E"o11),
qui, par son 6volution au cours de f immersion,
i Arro!.3H1 caract6rise la phase de propagation. Ep est
d6fini suivant les courbes densit6 de courant-
i----.- potentiel obtenues d partir d'6tudes potentiody-
t,tsstvlloN
namiques (voir la figure 3). Sa valeur n'6tant
pas reproductible, on la d6termine pour un en-
semble d'6chantillons, ce qui permet de d6finir
un domaine de probabilit6 de piqtration (voir
la figure 4), qui par ailleurs variera selon le
pH (voir la figure 5).

? 1 0 1 t :t 4 a 6 ? 8 I l0 | l2 13 14 l5 t6
pn
% d'6chantillons piqu6s

lt
Figure 2 : Le diagramme d'hquilibres potentiel-pH Scnsibilitc ma xr malc
de I'aluminium en eau pure d 25o C selon Pourbaix'"

o la corrosion cathodique, due d l'affaiblisse-

ment des propri6tes isolantes du film d'oxyde
superficiel sous polarisation cathodique.
20 1,/ / Domainc dc probabilit€
On a donc cherch6 d d6limiter la zone qui, Sensibiliie m de piq0ration
dans le diagramme d'6quilibres potentiel-pH 0 | /,/
0.8 0,7 0.6 0,5 0,4 0.3 0.2 0.1
pr6c6dent, permettrait d' as surer une protection
Potentiel V/IJSS
r6elle aux alliages durant le traitement.

La sensibilit6 des alliages d'aluminium d la

Figure 4 : La ddtermination du domaine de pro-
corrosion par piqfires est d6finie, en utilisant
babilit€ de piqfiration pour I'alliage 3003 en
milieu tamponnd de citrate de sodium : pH-= 5,4;
T: 10" C;vb: l0 mV/min; [NaClJ : ]['M.
vb : vitesse de balayage

Le tracl de l'6volution de Ecorr des alliages

d'aluminium en fonction de la dur6e de leur
Ln immersion montre que, aprds un certain temps,
T
cette valeur p6ndtre dans la zone of I'alliage
devient sensible d la corrosion par piq0res (voir
la figure 6). Selon I'alliage, cette corrosion se
d6veloppera plus ou moins rapidement. Il a
ainsi 6t6 d6montr6 que, aprds deux ir trois se-
maines en eau d6min6ralis6e, une pidce pr6cor-
rod6e en alliage aluminium-cuivre se recouvre
de pushrles qui t6moignent de la pr6sence
DENSITE DU COURANT de nouveaux ph6nomdnes de piqfiration,
tandis qu'il faut quelques mois pour que ce
Figure 3 : La corrosion par piqfrres : ddfinition du ph6nomdne se produise dans le cas d'alliages
potentiel d'amorgage des piqfires. sans cuivre qui sont plac6s en milieu d6sa6r6.

376
 Densrle du rouranl (uA/rnrz)
Potcntrcl V/ESS

\+\10
20

60{J2
40

60
?.014
BO

77 82 10 100 1000 10000 10000

4.7 5,2 5,? 6.2 6.7 72 B?92
1i'nrps (nrrn)
pll

Figure 7 : La polarisation cathodique d

-1,4 V/ESS (dlectrode au sulphate mercareux) pour
Figure 5 : L'influence du pH sur la sensibilitd de
dffiren* alliages d'aluminium en solution tampon-
I'alliage 6082 (Al-Mg-Si) d la corrosion par
nde de citrate^de sodium : pH : 5,4; T: 10" C;
piqfires : T : l0o C; NaCU : l0'M.
[NaClJ = 103M.

Potentrel V/[SS d partir d'essais potentiostatiques (voir la

figure 7). Deux 6tapes apparaissent, en fait, sur
;I les courbes densit6 de courant-temps (appel6es
< courbes de polarisation >) obtenues. La pre-
2014
midre, caract6ris6e par une 6volution lente des
i,l . )-t'-'
. ,.i-,;q. densit6s du courant, correspond d l'hydratation
_
progressive du film d'oxyde initial d la suite
de I'incorporation des ions H- du milieu. Cette
incorporation des H' est d'ailleurs plus rapide
au voisinage des inclusions qui, contenant des
l.:t 6l6ments nobles tels que le cuivre ou le fer,
0
sont pr6sentes dans les alliages d'aluminium.
Le film ainsi modifi6 localement permet alors,
dans une seconde 6tape, I'activation de la corro-
Figure 6 : L'evolution de Econ des alliages d'alu-
minium enfonction de la durde de leur immersion en sion des inclusions : la matrice environnante
solution de cilrate de sodium : pH : 5,4; T = 10" C; est consomm6e et, au terme du processus, les
[NaCt]: I0'3M. inclusions sont 6limin6es (voir la figure 8). Ce
ph6nomdne 6tant plus ou moins important selon
a : hmite infirieure du domaine de piqiration. l'alliage, l'a6ration du milieu et le potentiel
cathodique, I'augmentation correspondante
des densit6s de courant sera, elle aussi, plus
Sous polarisation cathodique, les alliages d'alu- ou moins prononc6e (voir la figure 7).
minium peuvent Otre le sidge d'une corrosion
cathodique qui prendra, selon les cas, la forme Aprds quelques jours de polarisation, la surface
de piq0res ou d'une corrosion g6n6ra1is6e. Afin est habituellement recouverte de piqfires isol6es
d'6viter le risque majeur de corrosion due ri des ou en amas. La corrosion est particulidrement
alcalinisations locales de la solution d proxi- d6velopp6e lorsque les bulles d'hydrogdne,
mit6 de la surface m6tallique, notre 6tude a form6es sur la surface m6tallique, ne sont pas
6t6 men6e en milieu faiblement acide et tam- 6vacu6es, car des ph6nomdnes d'alcalinisation
ponn6 de citrate de sodium. L'6volution de locale apparaissent alors sous celles-ci. Une
1'6tat de surface sur des 6chantillons non agitation cons6quente de la solution permet de
d6grad6s initialement peut alors €tre suivie parer i cet inconv6nient.

JII
 Hydratation du film au voisinage
l2
des inclusions _J l0

ur H2
-H-
0.ll

H.nz H' H- 0ti

04

,w
lncfuslon H2, matrlce -Hz
) fitm d'oxyde 02
0

0.2

04
06
08
I

n2 H2
tf,
l4
I.ti

2.0

malrice consommde 2.2

24
:br r | | | I | ! I | | , | | | , \\1 26
'
2 1 0 I 2 ) 4 6 ? B I t0 lt l2 l:] 14 t5 16
'

Figure 9 : Le diagramme d'iquilibres potentiel-pH

Figure 8 : Le mtcanisme de la corrosion cathodique expdrimental intdgrant la sensibilitd des alliages
des alliages d'aluminium en milieu tamponnL.
d'aluminium ('alliage 6082, dans ce cas) d la
corrosion par piqires et d la corrosion cathodique.

Au terme de cette 6tude, le diagramme

d6ploy6e en acier inoxydable, permet en
d'6quilibres potentiel-pH exp6rimental, fait d'extraire les ions chlorure ins6r6s pro-
adaptd ir nos conditions op6ratoires, a pu
fond6ment dans les mat6riaux. Le potentiel
Otre trac6. La zone qui permettra dtassurer
cathodique et I'agitation peuvent 6tre optimis6s
la protection des alliages d'aluminium pour augmenter I'efficacit6 de cette technique,
durant le traitement de d6chloruration a mais c'est la sensibilit6 des alliages ir la corro-
dis lors 6t6 d6limit6e (voir la figure 9). sion cathodique qui d6limitera les valeurs
r6ellement utilisables. La d6chloruration sous
Looptimisation de la d6chloruration polarisation cathodique obeit d la loi d'extrac-
tion log(Ccr-) : at"'. ou a: constante. plus
Une polarisation cathodique contr6l6e d poten-
rapide que les loisJrabituelles rencontrees
tiel impos6, en milieu faiblement acide et tam-
sans polarisation*''. ce qui con{-rme I'effet
ponn6 de citrate de sodium, agit6 et d6sa6r6,
b6n6fique du potentiel qui est appliqu6 lors du
est propos6e afin d'extraire rapidement les
nettoyage (voir la figure l0).
ions chlorure pr6sents dans les alliages d'alu-
minium. Avant cette 6tape, ces alliages sont
toujours immerg6s dans une solution com-
Des exemples de traitement
plexante (et sans polarisation), afin de dis-
Les objets dr traiter sont, pour la plupart, com-
soudre les produits de corrosion gdnants qui,
posites, de sorte que la stabilisation des alliages
contenant des 6l6ments tels que du cuivre ou
d'aluminium doit se situer dans un cadre de
du fer, peuvent se r6duire facilement sur les
restauration plus g6n6ral. Les systdmes fer-
surfaces m6talliques. De plus, on extrait ainsi
aluminium exigent ainsi deux phases sup-
les ions chlorure superficiels qui n'exigent
pl6mentaires dans la proc6dure de traitement
aucun traitement 6lectrolytique pour
qui est pr6conis6e ci-dessus : un d6gangage
passer en solution.
des pidces ferreuses en solution basique et in-
hibihice de corrosion pour les alliages d'alu-
La polarisation cathodique qui est appliqu6e
minium (m6tasilicate de sodium 0,5 N), d'une
par la suite, grdce d un g6n6rateur de tension
part, et, d'autre part, un nettoyage des surfaces
dont la bome positive est reli6e d une t6le
m6talliques en eau courante mais sous

378
 aluminium alloys are subject to rapid corrosion
|1oler chlor urcs (M),/cm' d'nllagc on exposure to the open atmosphere. The chlo-
rides contained in the materials are responsible
l0 4 for the acceleration ofcotosion, but the pres-
ence of unstable conosion products may also
l)oldn\ntlon
activate the process es.
l0 5

The extraction ofchlorides is achieved in a buff-

\ r,,s((ct )= ,,y'. ered solution under cathodic polarization at con-
t0 b stant potential, but dffirentforms of corrosion
may occur. Those observed are pitting conosion
caused by extracted chloides and cathodic
l0 7 corrosion.
l"
r/l
The pittingfield has been determinedfrom the
Tcrnps (r)

Figure l0 : La courbe de ddchloruration characteristic pitting potentials and the degree of

d'une plaque prdcorrodde en alliage aluminium- aeration of the solution.
cuivre 20 I 4 sans polaisotion, puis avec polarisation
en solution de citrate de sodium. I'anode Atuil en The mechanisms and the kinetics of the cathodic
acier inorydable 3 16. corrosion have been studied in detail by means of
potentiostatic measurements. In a buffered solu-
tion, localized alkalinization close to the sudace
protection cathodique. Ces deux nouvelles
is impossible, thus the corrosion arisesfrom
6tapes se situent entre la phase initiale d'immer-
the loss ofthe oxide layer's passive properties,
sion en solution complexante et le traitement de which, in turn, is caused by cathodic polarization.
d6chloruration de la pidce (entidre) sous polari- The major role of the intermetallics in the initia-
sation cathodique. tion of corrosion has been determined and the
way in which they change with time during the
Conclusion process has been revealed.

Centr6 initialement sw le probldme de la stabi- Conditions requiredfor the treatment and which
lisation des alliages d'aluminium, le traitement ensure the protection ofthe alloys may be estab-
des vestiges a6ronautiques conduit d consid6rer lished in a de-aerated, stined citrate solution
la conservation-restauration de mat6riaux com- with a stable, slightly acid pH andwith a
posites, dont les alliages d'aluminium ne for- cathodic potential above that at which cathodic
ment qu'une partie. Des m6thodes existent corrosion takes place. The kinetics ofextraction
aujourd'hui pour les systdmes fer-aluminium ofchloide ions under polarization has been stud-
ou cuivre-aluminium mais le cas des compo- ied and the conditions have been optimized to ac-
celerate it. A pre-treatment employing chemical
sites m6tal-mat6riau organique, que I'on trouve
complexation to make the extraction more ffie
autant sur les 6paves les plus anciennes que sur
tive is also described.
les plus r6centes, reste pos6.
The application of this treatment to samples from
Abstract sub-aquatic aircraft remains has given reproduc-
ible results. It has bemfound that the qtraction
An Elearolytic Treatment for Stabilizing ofchloide ions under cathodic polarization is
Suhmerged Aluminium Alloy Aircraft Fragments rapid and ffictive. Where iron alloys are com-
bined with aluminium alloys, spectfic treatments
Once retrieved, after long periods of immersion have been proposed.
under water, aluminium alloy artifacts (i.e., air-
craftfragments) must be given conservation treat-
R6f6rences
ment. Our method consists of the extraction of
"threatening" ions using cathodic polaization.
l. Lacoudre, N., Electricite et Arch1ologie,
Paris, Admitech-Electricit6 de France,
It has been established using accelerated corro-
sion tests that without treatment, pre-corroded
r99l (rsBN 2-7240-00rs-3).

379
 4. North, N. A., et C. Pearson, < Washing
2.Degrigny, C., Thdse de doctorat, Universit6 Methods for Chloride Ion Removal from
de Paris-VI, novembre 1990. Marine Iron Artifacts >>, Studies in
Conservation, vol. 23 ( 1978), p. 17 4-17 6.
3. Th6 del'Atlas of ElectrochemicalEquili-
brium in Aqueous Solutions, M. Pourbaix, 5. Macleod. I. D.. ( Corrosion of Corroded
traduit par J. A. Franklin, National Association Copper Alloys: A comparison of new and
of Corrosion Engineers, 1974, avec la permis- traditional methods foiremoving chloride
sion de NACE. ions >>, Studies in Conservation, vol.32 (1987),
p.25-40.

380
The Discoloration of Acrylic Dispersion Media

James Hamm, Ben Gavett, Mark Golden,

Jim Hayes, Charles Kelly, John Messinger,
Margaret Contompasis and Bruce Sffield

Art Cons ervation D epartment

Buffalo State College
Buffilo, N.Y.
U.S.A.

Abstract Heat and ultraviolet induced aging of mock-ups

were also separately pursued. Fur'ther testing has
A n u n ch a ract e i st i c ye I I ow /brown d i scoI o rat i o nl shown that the changes obtained with these meth-
in the background ofa color-field painting (1970) ods are not as significant to the study of the Good-
by Robeft Goodnough initiated the present investi- nough painting as those obtained by natural
gation of acrylic dispersion media. The owner re- aging under ordinary environmental conditions.
calls that the yellowing had become objectionable
as early as the mid 1970s, suggesting that the Introduction
problem had occutedfairly early in the life of
the painting. During recent telephone conversa- Since the early 1930s, various new slmthetic
tions, the artist indicated that, in this peiod, he resins have been employed by artists for rea-
typically would coat his cotton duck canvas over- sons of increased durability, rapid drying, ease
all with Liquitex matte medium before applying of handling, new application techniques, or to
any paint. achieve different effects. Since its in^troduction
around 1953, acrylic emulsion paint', as it has
At present, the color of the Liquitex matte me-
been popularly described, has gradually be-
dium closely resembles deposits from cigarette
come accepted as a viable painting medium
smoke. Initial tests have shown that although
by professional artists. In relatively few years,
cigarette smoke residue may have contributed
somewhat to the appearance of the painting, this
manufacturers have introduced a wide variety
is not the major cause for the yellow/brown of products for use with acrylic paints, such as
discoloration. clear gels of varying viscosity, matte and gloss
media, modeling pastes, drying rate retarders,
Evidence obtained from mock-ups created to du- and varnishes. The gels and media are often
plicate discoloration caused by natural aging is used independently ofthe paint for textural
discus sed. Remarkably, this discoloration occurs effects, coatings, collage adhesives, or as
over a relatively bief period of time under ordi- transparent gel colors, slightly tinted with
nary environmental conditions of temperature, acrylic paints.
relative humidity, and light. The discoloration
can occur in any brand of acrylic media or gel, Some artists have used these clear drying,
but is most noticeable in the unpigmented gels, acrylic dispersion media as coatings or size
matte, and gloss media. It is the result of a detri- layers on bare canvas, over which a painting
mental interaction between the support and is then executed. An example of the use of
the media.
Liquitex matte medium as a coating or sizing
on bare canvas, and the resultant discoloration

381
that occurred, can be seen in a Robert Good- and 83'C) may be useful indicators of relative
nough painting done in 1970, titled Develop- stability, when compared with other samples ex-
ment MC (Figure 1). In 1990, during telephone periencing the same conditions, but evidence of
conversations, the artist indicated that he had discoloration produced at elevated temperatures
sized the entire canvas with Liquitex matte me- does not imply that the same discoloration can
dium before applying the paint. The original ever occur at normal temperatures.
and current owner noticed that the yellowing
had become objectionable by the early- to mid- Presumably, the Goodnough painting has not
1970s. By comparing this painting with other experienced thermal aging tests. The question
paintings and serigraphs done by the artist at then is: why has it discolored and, furthermore,
about the same time, but without the matte me- why in such a relatively brief time? Could it be
dium sizing, we can easily imagine how this contaminated medium, tobacco smoke (or other
painting once looked originally. atmospheric pollutants), or an apparent yellow-
ing tendency of the medium itselfl Is this prob-
lem localized to paintings by this particular
artist or medium, or is it broader, encompassing
other artists or other acrylic media?

Background
The laboratory work at Buffalo State College
involved testing Liquitex acrylic matte and
gloss media under harsh environmental condi-
tions with the hope of duplicating the appear-
ance of the discolored Goodnough painting.
Figure I Development MC, a color-field painting
by Robert Goodnough (1970), measuring Contompasis, and later Suffield, created canvas
6 I cm x 1 22 cm (24 in. x 48 in.), from the collection mock-ups, using washed and unwashed cotton
of Mr. and Mrs. Charles M. Liddle the Third. and linen, coated with the Liquitex media. Test-
ing occurred under conditions of90oC and
It shouldn't be too surprising that changes in 50% RH for a period ofseven days using the
color might have occurred, since the composi- Blue M Humid Flow Temperature and Relative
tion of acrylic dispersion media is quite com- Humidity Cabinet. A conhol support of glass
plex, and some of the individual components was also used for the media tests.
may not age well on their own. Besides the
acrylic resin itself, we may find residual (unpo- The testing showed that the gloss medium dis-
lymerized) monomer, surface active agent colored slightly more when applied to cotton
(emulsifier), protective colloid (thickener), as compared to linen, but overall less than the
water-soluble initiator, plasticizer, pH buffer, matte medium applied to either support. Some
anti-foaming agent, coalescing agent (fikn for- discoloration occurred in the rmaged matte
mer), preservative, drying retarder, freeze-thaw medium applied to linen. Generally, the matte
stabilizer, ultraviolet (UV) inhibitor, optical acrylic medium discolored the most, even when
brightener, matting agent, and water. the substrate was glass.

Suggestions that the raw acrylic dispersions With respect to the latter furding where glass
may be susceptible to yellowing or discolora- is the substrate, the testing conditions were not
tion, especially during thermal aging tests, have the same as those described in the Howells
been made by a few authors. However, as dis- et al. article, which may explain why an appar-
cussed by Howells, Bumstock, Hedley, and ently greater degree of discoloration was ob-
Hackney thermal aging tests cannot be related tained by them. Howells et al. ran tests for
to the aging produced naturally over time, un- 50 days with presumably no RH control, as
like those changes induced by light aging.' none was mentioned, while Contompasis and
Thermal aging tests (at temperatures of 59"C Suffield ran tests for seven davs at 50% RH.

382
In 1989, the Research Laboratory at Golden o Does thickness of application affect the
Artist Colors, Inc. tested several raw acrylic degree of discoloration?
dispersions and several brands of acrylic gel
media (from Aquatec, Golden, Htutt, Lascaux, o Are tinted gels, used as glazes, affected by
and Liquitex) in a QUV Weathering Tester. discoloration? What effect does pigmentation
The test exposed the materials to 400 intermit- ofthe gel have on the degree ofdiscoloration?
tent hours of UVA exposwe at ambient humid-
ity and temperatwes of 30'C to 35oC. The raw With these questions in mind, a series of
dispersion samples were prepared by pouring tests was designed in order to isolate pertinent
them into plastic trays, and the gels by applying factors.
2.5 mm (100 mil) wet thicknesses over gessoed
canvases. Major discoloration was observed
Experimental
and attributed primarily to the UV exposure. In-
deed, a medium formulated with I-IV inhibitors, The mock-up materials and conditions created
tested,under similar conditions but drawn- in the following experiments closely parallel
down'on an altemative support (posterboard), those normally used by artists. We avoided us-
resulted in the elimination of most of the discol- ing enhanced conditions of artificial aging in
oration relative to the control. However, further order to restrict the study to those phenomena
tests with UV-inhibited formulations revealed that would occur under normal environmental
they were capable of discoloring just as dra- conditions.
matically under a few weeks of ambient condi-
tions, when applied to canvas, as they had in Effectiveness of Acrylic Gesso
the initial test under accelerated conditions. The
as a Barrier between Support and Media
conclusion: something other than UV radiation The test was designed to show whether or not
and heat was causing the discoloration.
the acrylic gesso would provide a barrier to any
discoloration derived from the canvas, and also
The first clear indication of a problem possibly
whether or not some discoloration could be ex-
originating with the supports came from one of pected to occur in the gel regardless ofthe na-
these test panels where a small section on the
ture ofthe barrier layer.
edge had pulled away from the canvas before
complete d.ying had occurred. This section re-
Heavy linen canvas was stretched over a
mained clear while the rest of the test strip dis-
30.5 cm x45.7 cm(12 in. x l8 in.) wooden
colored. Ifthe discoloration problem resided
stretcher and stapled in place. The surface was
solely in the medium, then why did this small divided into four equal areas and treated in the
section remain clear? If the problem originates
following manner: the first area was brush
with the support, then why did the rest of the coated with two layers of acrylic gesso, the sec-
test strip discolor, irrespective of whether it
ond with five layers of the same gesso, the third
w€rs over the canvas or the gesso?
area was protected by a25.4 pm (l mil) mylar
barrier, and the fourth was left uncoated. A
This "accident," combined with the previous
3.81 mm (150 mil)thick coating of Golden
observations, raised some imporNant questions:
Regular acrylic gel was applied across all four
areas with a 10.2 cm (4 in.) GardnerA.,ieotec
r How effective is acrylic gesso as a barrier?
adjustable film casting knife. The coated can-
vas was stored under indoor ambient conditions
o What effect would washing the support have?
and evaluated after two weeks.

o How do different types of supports compare? Effect of Pre-washing the Support

This test was designed to determine to what
o How do different brands of acrylic media extent, if any, washing fabric supports would
compare? subsequently reduce discoloration in gels.

383
dried sufficiently to become clear, an average
period offive days.

Comparison of Manufacturers' Acrylic

Dispersion Gels over Different Supports
Individual wooden stretchers, measuring
15.2 cm x 15.2 cm (6 in. x 6 in.), were
assembled and covered with supports that
are marketed fi2r artists' use: woven polyes-
ter 0.475 kdm' (14 oz.lyd!), heavy linen
036akglm' (10.15 oz.lyd'), and#10
cotton. Untempered Masonite panels mea-.
suring 15.24 mx 15.24 cm x 0.635 cm Figure 5 Eight brands of acrylic gel on an acrylic
(6 in. x 6 in. x ll4 in.) were also prepared. All gesso coated cotton support showing various
materials were selected from single pieces to discoloration at 56 days.
minimize variability. Each support was coated
with three 0.25 mm (10 mil) layers of Golden
Acrylic Gesso according to the procedure
described previously.

After allowing l0 days for the final gesso coat

to dry, each type of support received a 3.81 mm
(150 mil) wet film application of each of six
different commercially available acrylic gels,
yielding atotalof 24 test pieces. The gels were
allowed to dry for five days before measure-
ments began. The overall test period lasted for
90 days under indoor ambient conditions, as be-
fore (Figures 4 to 6). The gels included in this Figure 6 Eight brands of acrylic gel on an acrylic
discussion are: Golden Regular, Grumbacher, gesso coated polyester support showing minor
Lascaux, Liquitex, Speedball, and Uhecht. discoloration at 56 days.

Thickness Variations
To determine the amount of discoloration occur-
ring in a gel due to its thickness of application,
a test was conducted by preparing individual
15.24 cmx 30.48 cm (6 in. x 12 in.) supports
of Masonite and #10 cotton. The supports were
prepared with the same techniques employed
earlier. A Gardner,A{eotec adjustable film
casting knife was then used to apply gel in
a stepped down application ranging from
3.81 mm to 0.635 mm (150 mil to 25 mil)
Figure 4 Some of the
wet film thicknesses, in 0.635 mm (25 mil) in-
crements. CIELab measurements and color dif-
ferences were periodically recorded.
Spectrophotometric measurements were taken
at frequent intervals and recorded as before. For
Effect of Gel Pigmentation on Degree
the purpose of making the calculations, gels
of Discoloration
over the cotton, linen, and Masonite samples
This test was designed to determine the effects
were compared to the same gels over polyester
of discoloration associated with different
of the exact same age, as described previously.

385
supports upon pigmented gels. Five different using linen, Masonite, and other brands of
supports were used: polyester, heavy linen, acrylic gesso have given similar results. Not all
#10 cotton, Masonite 0.635 cm (l/4 in.) untem- gessos behave the same; however, all allow at
pered, and lauan plywood 0.635 cm (l/4 in.). least some discoloration to pass through from
All supports were prepared following the con- the support.
ventions described above.
Effect of Pre-washing the Support
Golden Regular Gel, tinted in the following ra- Washing the fabric supports was clearly
tios with Golden Heavy Body Ultramarine Blue effective at reducing the extent ofdiscoloration,
Acrylic Artist Paint, was applied in 3.8 mm as the curves indicate (Figure 8). Discoloration
(150 mil) applications to each type of support of the gels over the two cotton supports (#12
in the following ratios: 500:1, 100: l, 10: l, and and #10 weights) was reduced from 18.9 and
0:1 (Gel:Paint). CIELab measurements and 12.9 units to 0.16 and 0.9 units respectively at
color differences were recorded periodically 84 days aging.
through the 90-day test period. Color differ-
20l
ences were calculated by comparing the polyes-
la i

ter to the other supports at points of identical

ages. iU rol
6 ro

Results
S,;]
I.l
:^
6l
'l
Effectiveness of Acrylic Gesso o 4-
2l
as a Barrier between Support and Media 0l
a< b i E ts 6 . F... : ::6

It is clear that discoloration did not develop in o t0 20 --30 40 50 60 70 80 90

Age Days
the section of acrylic gel applied over the mylar
barrier, thus indicating that within the limits of f- *lz c"ttoncrrL * *io c.n"n-crri. '; trvy linencrt I
this experiment, the gel does not have a ten- .,, + #rocoaon-y:h + Hvy. Linen,wash
E 11yl*::i
dency to rapidly discolor when separated from
the canvas (Figure 7). Figure 8 Relative color change of Golden Regular
Acrylic Gel applied over acrylic gesso on three
washed and three unwashed cotton and linen canvas
supports.

Washing the linen also reduced the amount of

available discoloring agent, but not to the same
extent as for the two cotton fabrics. Apparently,
these materials from linen are not as easily re-
moved by a single-water wash. It may be that
they are more readily released with the applica-
tion of emulsifiers, solvents, or one of the other
numerous additives found in acrylic dispersions.

Comparison of Manufacturerso Acrylic

Figure 7 Linen support brush-coated in two areas Dispersion Gels over Different Supports
with two layers andfive layers of acrylic gesso, the The four graphs in Figures 9 to 12, which show
third area protected by a 2 5 .4 pm ( I mil) mylar the performance of all the acrylic gels over
barrier, and the fourth left uncoated.
each ofthe foui supports, reveal that discolora-
tion depends more on the support used than on
In the other three areas, discoloration has oc- the brand ofgel selected. Polyester performed
curred whether the acrylic gel is over two, five, the best, with discoloration at 80 days ranging
or no layers of acrylic gesso. Consequently, the befween 0.45 and 2.1 units. It was not deter-
acrylic gesso here is generally ineffective at mined if this discoloration was related to the
providing a barrier to discoloration. Other tests

386
 9 4
::I
o ;
I
"l
O fr ;:l
-9 ,ol
O

30 40 50
Age- Days
60
_l :l ,-
Age' Days

_
I r. oo,o"n + crumbacher * Lu""u-
F 9..*
-+ Grumbacher * Lascaux
Lrqudex + Speedball t Utrecht Ltqulex Sp@dball + Utrecht
-

Figure 9 Discoloration of various acrylic gels over Figure l2 Discoloration ofvarious acrylrc gels over
#10 cotton. polyester.

The shape ofthe curves, as they reflect the rate

polyester, or if it was inherent in the gels
ofchange, appear consistent for each support,
themselves. The other supports in the test re-
regardless of the gel used. Over Masonite, at
sulted in much more discoloration: Masonite
leastT5o/o of the discoloration occurred within
23.7 to 33.5 units, heavy linen 16.2to
the first four days, over heavy linen, within
27.5 units, and #10 cotlon 17.7 to 30.2 units.
20 days, and over cotton, within 27 days.
40

35
While the extent of discoloration may vary sig-
630
nificantly among different gels over the same
support, the rank of the gels, in terms of most
to least discoloration, varied between supports.
-
#
glol F;'r- ^;=*:::#:+--r For example, over Masonite at 82 days of age,
Speedball exhibited the most discoloration, fol-
lowed by Golden, Lascaux, Liquitex, Utrecht,
o+
o _ ro zo 30 40 sFoo F e-o
and Grumbacher. However, Grumbacher discol-
Age' Days
ored the most (27 .47 units) when the gels were
- ooia"n + Grumbacher * Lascaux
applied over heavy linen, with the other gels
'r Lrqu[ex + Speedball r Utrechl tightly grouped between 16.2 and 17.5 units.

Figure l0 Discoloration ofvarious acrylic gels over The graphs in Figures 13 to 18 group the data
heavy linen. by gel manufacturer, showing the relative im-
pact ofeach ofthe supports on any one gel.
These graphs also illustrate an interesting vari-

;;-F;+:=
40I
ation in the rankings of the extent of discolora-
!
,C 301 tion from gel to gel. With fow of the six gels,
tr
o 251
the Masonite support showed the most discol-
3 zol
oration, while cotton showed the most in the
o- other two gels.
i
d
tol
-L
3l
I Thickness Variations
o+
0 '10 20 30 40 50 60 70 Figures 19 and 20 show that increasing the
Age- Days
thickness ofthe gel does increase the
+ Grumbacher -
discoloration.
- Cotaen Lascaux
+ Lrqurlex + SDeedball r Utrecht
Appearances suggested that perhaps the
Figure I I Discoloration of various acrylic gels over thicker areas ofgel discolored even more than
Masonite. the thinner areas, beyond simply a proportional

381
 40

^*l
€ 3ol 6' 301
'u]
*-
; !
I "u1
,ol
P 'zo-
'l
tol ,-?
o t5-',,f O
o
b 101
o I,
_s
o
O .
l "l
)0
----
10 20 30 40
Age, Days
50

+ Cotton * Heavv Lrnen

-
Masonile
lt'l l-:":l * Heavy Lrnen
-
Mason{e
-
Polyester I

Figure I3 Effects ofsupport on discoloration: Figure l6 EJfects ofsuppoft on discoloration:

Golden Regular Gel. Liqufiex Gel.

o
--4- T -i -r-_-__- g
O J
o 20

O ts r I O 15

-9
o ro I / o 10
O a

'l 5

0+ - 0
0 10 20 30 40 50 60 70 80 90 20 30 40 50 60 70
Age Days Age- Days

i con- Heavy Lrnen lv4asonfe

-'1f:6 l I
* "a';- * Heavy Lhen Masonrte Polyester
I
- - -
Figurel4 Effects ofsupport on discoloration. Figure I7 Effec* oJ'support on discoloration:
Grumbacher Gel. Speedball Gel.
40 40

35 35

€ sol
'+*H

3, ,o
;?20
5rs
o
oo 10 t Q10 1.,-u
5 5

0 0
0 r0 20 30 40 50 60 10 20 30 40 50
Age- Days Age- Days

+ Cotton Heavy Lrnen _ Masonte Polyesier *@

- -
I
--+conm
Figure l5 Efects ofsupport on discoloration: Figure I8 Effects ofsuppoft on discoloration:
Lascaux Gel. Utrecht Gel.

increase. However, comparing gel thicknesses approximately 8.5 times. Since the discolora-
of 0.625 mm and 2.54mm (25 mil and 100 mil) tion remains closely proportional between thick-
at day eight, the CIELab units are respectively nesses at widely different moments in time,
1.2 and 9.7, a difference of approximately then the extent of discoloration is simply pro-
8 times, while at day eighty-two the units are portional to the thickness ofapplication only.
respectively 2.1 and 18.0, a difference of

388
 35. 4o-
301 35 ]
6e^

9ro U.u
o
F 15 $.0
o
-b 10 5rs
o o
o5 b
O
10

0
=;*
o 5 10
to 15 20 25 30
3035
35 40 45
Age ' Days
I
50 55 60 65 70 75 80 85 90
5

Age-Days

tr 25 mrl +
100 mrl i
50 mtl *
125 mrl a
75 mrl

150 mrl l--ffi' + rorcetCotor * roolGercorof .' soorc.m I

Figure 19 Discoloration related to gel thiclness Figure 22 Discoloration ofcolored glazes

over #10 cotton. Ultramarine Blue Acrylic Paint: Golden Regular Gel
tul
mixed in various ratios and applied to acrylic gesso
a-+<- pimed heaw linen.
ao] I {
4 ^,^ i ^
-' ,.
: 251 i_
aoI
Y 2oI
o
.F ttl 6q"".^l
O
* .ul
o
P
6
201

0 5 10 15 20 25 3035 40 45 50 55 60 65 70 75 80 85 90 6 rsl
Age - Days
F 101

i*rffiil F5f ff- -: !

40-- 50 =60
+1@mil* il A 1$mil ol
0 10 20 30 70 80
Age Days
Figure 20 Discoloration related to gel thickness
over Masonite. o.o[i u
l= ffi
ioo I 5oo 1 Ge coo,

::l Figure 23 Discoloration of colored glazes

'"1 Ultramarine Blue Acrylic Paint: Golden Regular
Gel mixed in vaious ratios and applied to acrylic
O gesso primed lauan plywood.
o ::
-c?_to-
O
_9
o 101
O
"l € .ol
;1 o z3'
E zot
f, r<r
. 0rGelcoo. + l0lGetcotor ryffiHbordl 6
b 10l
O

Figure 2I Discoloration of colored glazes , -;--r

Ultramarine Blue Acrylic Paint: Golden Regular 0 10 20 30 40 50 60 70 80
Age Days
Gel mixed in various ratios and applied to acrylic
gesso pimed cotton.

Effect of Gel Pigmentation

Figure 24 Discoloration of colored glazes
on the Degree of Discoloration Ultramarine Blue Acrylic Paint: Golden Regular
Figures 2l to 24 indicate that measurable, Gel mixed in various ratios applied to acrylic gesso
hence visible, discoloration generally increases primed Masonite.
as the concenfration of pigment decreases.

389
On the cotton support, the two higher propor- fabric support before stretching markedly
tions ofgel do in fact discolor at an increasing reduces the potential for discoloration ofthe
rate, displaying curves similar to those seen acrylic media, whether gesso is used or not.
earlier. The surprising difference between
these two higher proportions is that the curves A likely mechanism for SID is the following:
closely follow one another, even though there immediately after the application of an acrylic
is five times more gel in one than in the other. medium. water and other ingredientsd move
into the support, allowing unidentified soluble
As we might have predicted, the curves for components in the support to migrate back into
Masonite indicate the greatest degree of discol- the medium as the water evaporates, thus creat-
oration in the higher gel content samples. How- ing the discoloration. The unknown compo-
ever, Masonite's curves are fairly similar to nents are most likely finishing materials used
those of cotton and linen when gel proportions by yarn and/or fabric manufaiturers,9 dirt, o.
are low. other contaminants from crop growing, fiber
storage, or mill operations. They may also in-
In general, at gel:paint proportions over l0: I , clude components of the plant from which the
noticeable discoloration can be expected when fiber is produced.
the support is linen, cotton, or Masonite. The
results recorded for lauan plywood are some- Some questions remain:
what inconsistent with the other furdings and
therefore require further inquiry. o At what point during the drying process of
the acrylic medium does the discolored mate-
Conclusion rial become extracted?

The graphs clearly show that supports typically r Does discoloration change continually
used by artists can create significant discolora- over time?
tion of acrylic media. Modem paintings, espe-
cially color-field paintings, or any painting o Can the extractable discolored material be
incorporating the acrylic media described, may lo
identifi ed more specifi cally?
show a discoloration originating from the inter-
action between tlre acrylic dispersion medium
o How can paintings thaJ.have already
and the support.o This phenomenon is most
noticeable in the unpigmented gels, matte, and discolored be treated?' '
gloss media, but can occur in acrylic paint con-
taining these media. It is a characteristic that Endnotes
can be expected to occur regardless ofthe
brand of acrylic medium used. This phenome- l. The term discolorallon is used here to mean
non should be distinguished from changes in an undesirable change in color.
color, which are due to atmospheric/environ-
mental pollution, or a possible inherent ten- 2. It is worth mentioning that theterms emul-
dency for a given acrylic dispersion to discolor. sion and dispersion often used to describe these
paint media are not interchangeable, although
The graphs also indicate that the discoloration they are related. The resins used in the manufac-
can occur quite rapidly over a few weeks, but ture of these media are commonly marketed as
generally levels offafter a few months. Long- "acrylic emulsions." They are, in fact, accord-
term measurements should be pursued in order ing to the Paint/Coatings Dictionary of the Fed-
to determine whether or not the discoloration eration of Societies for Coatings Technology,
curves remain flat or gradually change produced by emulsion polymerization. This is
(increase or decrease) over the years. a process whereby the monomeric material pre-
sent as the discontinuous phase of an emulsion
The commonly used acrylic gesso does not sig- is polymerized in situ. However, according to
nificantly inhibit the Support Induced Discol- the same source, "after polymerization, a latex
oration (SID).' However, the washing of a

390
[acrylic emulsion] is a solid dispersed in water canvas, engaged panel stretchers, or paint on
and. therefore. is not an emulsion." back ofthe canvas, can worsen the discolora-
tion. Any of these will retard or prevent the
Emulsion is defined as a two phase, all liquid evaporation of water from the reverse, thus
system in which small droplets of one liquid forcing all moisture to migrate to the surface
(the intemal phase) are immiscible in, and dis- of the painting in order to evaporate. This one-
persed uniformly throughout, a second continu- sided evaporation of moisture apparently en-
ous liquid phase, the external phase. The term hances the movement of extractable materials
dispersion has broader meaning. encompassing from the support to the surface of the painting,
any composition or state related to a heteroge- thus intensifuing the discoloration.
neous system of solids, liquids, gases, or immis-
cible liquids. Therefore, an emulsion qualifies 7. A new acronym proposed by the authors to
as a special type ofdispersion. [n the topic at reference this phenomenon.
hand, the polymerized acrylic resin solids
comprise the internal phase, while water com- 8. Some likely candidates include emulsifier,
prises the continuous external phase ofa hetero- surfactant, coalescing solvent, stabilizer, or
geneous system. Technically, the materials these (and other ingredients) working in
we are discussing are called acrylic dispersions, conceft.
even though they are marketed as acrylic
emulsions. 9. Monie (1908) described fabric sizings
(many of which are still used today) containing
3. In the Howells, Burnstock, Hedley, and several types of starches (farina, wheat, or rice),
Hackney 1984 trC Paris Preprints article "Poly- fats, waxes, or glycerine, which may interact
mer Dispersions Artificially Aged," six acrylic with acrylic dispersions media unfavorably.
dispersions in a field of 14 polymer dispersions
were subjected to enhanced environmental con- 10. Suffield (1991) described the identifrcation
ditions using sunlight, fluorescent light, and ele- of components of the soluble extract. Cotton
vated temperatures. The dispersions were then and linen canvases were examined using micro-
tested for changes in weight, color, solubility, chemical tests. The tests indicated the presence
mechanical properties, softening to heat, sub- of starch and lignin in the linen samples and
mersion in water, and pH. Only one resin primarily starch in the cotton. Subsequent thin
showed no significant changes throughout all layer chromatography and infrared spectro-
the tests. The most common change among the photometry also indicated proteins were present
others was yellowing, a characteristic that did in both canvases. After tagging the hydrolyzed
not seem to impact any of the other physical proteins with dansyl chloride in order to visual-
properties of the dispersions. izethe chromatographs with UV light after de-
velopment, the results indicated the presence of
4. American Society for Testing and Materials a mixture of casein and other proteins in both
(ASTM), "Standard Practice for Preparing cotton and linen canvases.
Drawdowns of Artists' Paste Paints,"
(Designation D 4941-89) ASTM Standards, I l. Contompasis and Hamm have tested treat-
Philadelphia, P.A. (1990) pp.916-917 . ment options on the Robert Goodnough paint-
ing in Figure I using UV filtered fluorescent
5. ASTM, "Standard Test Method for Calcula- light as a bleaching tool (Tahk 1979). Results
tion of Color Differences from Insfumentally have shown an improvemurt in appearance, but
Measured Color Coordinates," (Designation D not to a degree sufficient to warrant recommen-
224-89) ASTM Standards, Philadelphia, P.A. dation as a viable treatment option in this case.
(1990) pp.297-300. (See Watherston 1972,p. 838 for a more
aggressive approach.)
6. Coverings or coatings on the back of the
canvas, such as stretcher bars contacting the

391
R6sum6 Des expdiences distinctes, d I'aide de maquettes
4galement, ont ttt mendes pour tvaluer la dbco-
Lu dicolorstion des mddius de dispersion loration caw,le par la chaleur et I'ultraviolet.
acryliques Ntanmoins, des essais ultdieurs ont rtvdld que
les modifications obtenues d I'aide de telles tech-
Une dicoloration j aune-brun plut1t exception- niques n'4taient pas aussi significatives, auxfins
nelle, sur le fond d'une peinture par champs de de I'dtude du Goodnough, que celles qui avaient
couleurs de Robert Goodnough (1970), est d 6td obtenues au terme d'unvieillissement naturel.
I'origine de I'dtude sur les mddias de dispersion dans un milieu ambiant normal.
acryliques qui est dtcrite dans la prtsente com-
munication. Tant le premier proprietaire du ta- References
bleau que le propri2taire actuel se souviennent
que cejaunissement avait dijd atteint le stade LeSota, S. et al., Paint/Coatings Dictionary
d'une coloration indisirable au milieu des an- (Philadelphia: Federation of Societies for
ntes 70, ce qui signifie qu'il se serait produit Coatings Technology, 1978).
assez t6t aprds I'exdcution de la peinture. L'ar-
tiste a pour sa part indiqub, lors d'une conversa- Howells, R., A. Burnstoch G. Hedley and S.
tion ttldphonique rtcente, qu'il avait l'habitude, Hackney, "Polymer Dispersions Artificially
durant cette ptfrode, de couvrir ses toiles (en Aged," Adhesives and Consolidants, Preprints
coton) avec du Liquitex mat avant d'appliquer of the Contributions to the Paris Congress,
toute peinture.
2-8 September 1984 (London: Intemational
Institute for Conservation, 1984) pp.3643.
) I'heure actuelle, le Liquitex mat ressemble de
trds pris d la couleur que laisseraient des ddp6ts
Monie, H., Sizing Ingredients, Size Mixing
defumie de cigarette. S'il est efectivement possi-
ble que de lafumie de cigarette ait quelque peu
and Sizing (Manchester, England: Abel
contibut d donner d la peinture son aspect ac- Heywood & Son. 1908).
tuel, de premiers essais ont ndanmoins riydld
qu'elle ne saurait €tre considirde comme la Suffield, B., "An Investigation into the
principale source de cette decoloration Discoloration of Acrylic Dispersion Media,
jaune-brun. Seventeenth Annual Conservation Training
Programs Conference (New York The Institute
Il estfait 6tat, dans laprdsente communication, of Fine Arts Conservation Center, New York
des nlsultats d'expiiences qui ont permis de University, l99l) pp. 152-166.
reproduire, d I'aide de maquettes, des ddcolora-
tions causies par un vieillissement naturel. Il est Tahk, C., "The Recovery of Color in Scorched
d noter que ces dicolorations se produisent assez Oil Paint Films," Journal of the American
t6t dans un milieu ambiant normal oi la tem- Institutefor Conservation, vol. 19, no. l, 1979,
-
pdrature, I'humiditd relative et la lumidre sont pp. 3-13.
normales. Si toutes les marques de mtilias ou de
gels acryliques sont susceptibles de se dicolorer,
Watherston, M., "Problems Presented by
cette dtcoloration sera plus apparente dans le
cas d'un gel pigmente, mat, et d'un mddium lus-
Color Field Paintings. Cleaning of Color Field
trd. Il s'agit d'une interaction indhirable entre Paintings," Conservation of Paintings and the
le support et le mddium. Graphic Arts,Prepr'tnts of Contributions to
The Lisbon Congress 1972,9-14 October 1972
(London: International Institute for Conserva-
tron,1972) pp. 83 l-845.

392
Methods of Analysis and Identification
M6thodes d'analvse et d'identification
 Practical Pitfalls in the Identification of Plastics

Helen C. Coxon

Royal Ontario Museum

Toronto, Ontario
Canada

Abstract more subjective approach. Several years ofwork

with plastic artefacts, and with plastics both old
The plastic artefacts now making their way into and new collectedfor expeimental purposes,
museums and pivate collections represent a new have shown quite clearly that there are limita-
fieldfor the conservator, one in which research is tions in the use of such methods. Thus, it is
only just beginning, and there are few, if any, essential to be aware of these limitations if identi-
guidelines for treatmenL Unlike materials such as fications are to be made with any degree of
ceramics or metal, which if stored properly will certainty.
remain stablefor long periods of time, plastics
can and will continue to degrade, sometimes with The principal problem appears to be that litera-
startling rapidity. The initiating agentfor such ture regardingplastic often presents tests that
deterioration can be as apparently insignificant workwell on samples of pure polymer, whereas
as the application ofa layer oflacquer before manufactured plastic items are rarely unmodi-
applying an accession number. fied. Fillers and pigments affect the densie and
appearance ofplastics, and interfere with colour
Plastics vary widely in their reaction to factors reactions. Plasticizers also alter the properties in
such as heat, light, cleaning agents and other en- unpredictable ways, and chemical alteration as a
vironmental conditions, and can also have drastic result ofdegradation cannot be ignored. Using
effects on each other. Before we can begin to set art history texts as a means of identification also
up treatments for specific polymers, we need to be has its dfficulties.
able to identify those polymers with accuracy.
These problems will be illustrated with reference
Ll/hile infra-red spectroscow provides the best to actual experience in trying to set up aioherent
means of identification, there is also a need for id entifi c ation sc he me for p lastics.
s i m p I er i de n t ifi ca t i on m e t hod s. p refera b ty ipp t i -

cable by conseryators and collectors on a routine Introduction

basis. Many such methods are suggested in plas-
tics literature, some scientific, some more empii- The so-called "age ofplastics" is convention-
cal. Float tests attempt to distinguish plastics ally dated from the late 1860s, when Alexander
using solutions of known density: bum tests rne
Parkes in the U.K. and John and Isaiah Hyatt in
smell and behaviour when heated or ignited;
the U.S. developed and marketed very similar
chemical spot-tests rely on colour changes in
semi-synthetic materials, known as Parkesine
response to a particular experimental procedure.
Identification by trade name (if lenown) or by sty- on one side of the Atlantic and Celluloid on
listic design and/or type ofobject represent the the other.

395
These days the number of plastics available dealing with in order to be able to attempt to
seems almost infinite, their properties can be devise a suitable featment. Where plastics are
tailored to meet specific criteria, and if we look concerned, we need to know more than that,
around us, more items than not in our everyday however. It is not enough just to look at an
lives are made wholly or partially of plastics. artefact and say "plastic." Since plastics vary
Correspondingly, plastics in an artefactual widely in their response to factors such as heat,
context are increasingly becoming a part of the light, airbome pollutants and potential conserva-
muserlm world. Initially they often appeared in tion treatrnents, it is important to be able to
museum collections more or less accidentally, identify them accurately. For example, the
as incidental components of artefacts collected wrong choice of solvent, or even the use of a
for other reasons, for example, cellulose nitrate solvent at all, can be disastrous, since, unlike
fan sticks, an ebonite pistol grip, or an early artefact materials such as glass or metal, some
composite book cover. Now, however, plastics plastics are inherently solvent-soluble. This
are being collected in their own right. has profound implications for other types of
treatrnent repair, infill, and to a certain
Some six years ago the European department of
-
extent inpainting as many adhesives and
the Royal Ontario Museum began a deliberate
- will not adhere well,
repair materials either
collection ofplastics, both art objects and or will do so by partial dissolution of the
more utilitarian items. The collection currently plastic surface.
includes items made from cellulose nitrate,
acrylic, melamine-formaldehyde, urea-formal- There are other reasons for identification also.
dehyde and phenol-formaldehyde, poly(vinyl It enables us to separate out items that may give
chloride), ebonite, cellulose acetate, and casein. off breakdown products harmful to other materi-
Many of these newly accessioned objects are als. We are used to testing display materials
now on display in the 'Lifestyles' section of with such a reason in mind, but we must now
the European Gallery. With the creation of the apply it also to plastic artefacts. Cellulose ni-
Institute for Contemporary Culture within the trate, for example, gives off acidic nitrogen di-
museum, we can expect the number of plastic oxide as it ages, which can cause bleaching or
artefacts in the collections to increase. fading of pigments and dyes, corrosion of met-
als. and discolouration. distortion or accelerated
There is probably no need to stress the fact breakdown in other polymers in direct contact
that it is going to be some time before we have or in an enclosed space, such as a display case
many definitive answers to the problems we or storage drawer.
are facing in freating these materials. As a new
type of material in the museum field, conserva- Identification also enables us to discover plas-
tors, curators and collectors are onlyjust begin- tics masquerading as natural materials, among
ning to come to grips with the reactions of the most obvious being cellulose nitrate imitat-
plastics to light, heat, atmospheric pollutants ing ivory or tortoiseshell.
and repair materials, and their effects on each
other and on other materials associated Identification
with them.
Information on the identification of poly-
In the past few years of working with various mers can be found mainly in publications
aspects of plastics, particularly the ir identifrca- dealing with the history of plastics, industrial
tion, a number of difficulties and anomalies chemistry joumals, books on polymer science,
have become apparent, which it seems appropri- organic chemistry, ftne art, and the food indus-
ate to bring to the attention of others working to- l-4't0'zo ut *.ll as more specific publica-
in the field. tions.'-o Suggested properties and means of
identification can be broadly grouped into
Why ldentify? foru areas (Table I).

First of all, why identiff? At a very basic Analytical methods such as Fourier Transform
level, we need to know what material we are Infra-red Spectroscopy (FTIR) and Gas

396
 plastics, and are used as references for the iden-
Table I tification of plastic,objects by means of visible
rr+'. r r! I z
Suggested Identification Methods for Plastics characteristicq.

At the outset itseemed as though identification

Visible Characteristics: would not be too much of a diffrculty with re-
- companson with reference ltterature
- trademarks
gard to the museum's holdings, since the initial
- function selection of objects for purchase was more or
- oesrgn less picked straight from the pages ofvarious
- date
- colour books on the art history ofplastics.
- feel: surface texture, welght, etc

However, it can be unwise to rely completely

Simple Physical Characteristics:
- taste and smell on art history or dealer identifications. The
- folding, tearing, etc. (sheet materials only) museum acquired a small box in black plastic,
- densrty
designed by Ren6 Lalique in the 1930s, the de-
Chemical Reactions: sign being identical to that in a photograph in
- pyrolysts one of the aforementioned books on the art his-
-pH
- solubility tory ofplastics, except that the pictured box is
- colour change reacttons red. The author of the book identifies the box
as "probably made from celluloid," that is, cel-
Instrumental Methods:
- Foufler Transform Infra-red Spectroscopy (FTIR), lulose nitrate, and ours was labelled as such by
Gas Chromatography (cC), etc. the dealer from whom it was bought, probably
on the basis of a photograph such as this. How-
ever, our box is actually made from cellulose
Chromatography (GC) are frequently referred acetate this was indicated by the sfiong
to as essential means for identiffing the com- smell of- acetic acid coming from it, and con-
plex mixtures of components present in 1n,fly frmed by means of a chemical spot-test. This
plastics, especially more recent plastics.'' '' may be a case of the same design having been
However, many conservators do not have ready used with different plastics, or it may be an
access to such equipment, and would therefore example of the use of the term 'celluloid' gen-
have use for simpler identification methods. erically, since in the early years of the develop-
ment of cellulose acetate as a replacement for
The initial focus of the identification work, be- cellulose nitrate in certain applications, cellu-
fore plastics became a real presence in the col- lose acetate was often referred to as "non-flam-
lections, was on sheet materials and adhesives. mable celluloid." Either way, the discovery of
However, it became very apparent that manu- the discrepancy does give an indication ofthe
factured artefacts are much less straightfor- kind of problem that may be encountered.
ward. "Case histories" of some identification
problems will demonstrate some of the prob- An acrylic display head, also purchased with
lems or anomalies that may be encountered, reference to a photograph, gives us another
and show that a certain degree ofcaution in case in point. It was accepted that it was
the interpretation of results is advisable. acrylic, that is, poly(methyl methacrylate),
without question, having no information to the
Identification by Visible contrary, until a colleague, while visiting a plas-
Characteristics tics factory in France, discovered in the course
of a casual conversation that students in plastics
Reference Literature for Identilication training schools in Paris were often required to
Dealing first with what is sometimes called duplicate that particular piece as an exercise...
reference literature, these publications usually in polyester.
describe the history or art history ofplastics.
They often include lists of trade names and in- Other problems with relying on reference
formation on properties, illustrations of typical literature include frequent lack of agreement on
designs, and typical functions for the various dates, and a bias towards the history of plastics

397
in one particular country, with corresponding Identification by Physical Properties
gaps in the record. Identification by date also
assumes either that the date is stamped on the The second group of identification possibilities
piece, or that its history is sufficiently well- is based on physical properties. As shown in
documented to be fairly certain of its date of Table II, non-artefactual sheet materials can be
manufacture. initially assessed by means of a range of
physical properties:
Identification by Trade Name 'o appearance
Identification by trade name should also be flexibility
handled warily. Strictly speaking, the name o whether the material holds a crease
Celluloid should be used only for the cellulose o the degree to which it can be stretched
nitrate products of the Albany Dental Plate o how easy it is to tear
Corporation, to whom the name was originally
registered in 1869 to 1870, followed by the These properties are often useful in the identifi-
Celluloid Manufacturing Corporation and, cation of plastic storage materials. However,
more recently, the Celanese Plastics Corpora- this kind oftest is diffrcult to perform on three-
tion. However, just as we tend to call all photo- dimensional pieces, and although some collec-
copiers Xerox machines, so the name has tors routinely use the smell and taste of plastics
become generically associated with all cellu- as a guide to identification, such methods have
lose nitrate products. obvious negative implications.

Bakelite was the trade name originally given to Identification by Chemical Reactions
the resin produced from phenol and formalde-
hyde by means ofa process involving heat and Density Test: One readily performed test can
pressure, patented by Leo Baekeland in 1909. be used to distinguish the polyolefins, that is,
The final product gave a characteristic dark, polyethylene, polypropylene, polyisobutylene,
mottled appearance to manufactured items such from other materials, at least in fairly pure
as electrical fittings. However, it would be a forms. The polyolefins have a density less
mistake to assume that everything labelled than one, so if a small sample is dropped into
Bakelite is phenol-formaldehyde, since that a beaker of water to which a few drops of deter-
trade name now applies to polyethylene, gent have been added to negate the surface ten-
polypropylene, epoxy, phenolic, polystyrene, sion. it will float. while all other materials will
phenoxy, polysulphone, ethylene copolymer, sink, provided there are no air bubbles clinging
ABS, acrylic and vinyl plastics, among others. to the sample.
Conversely, despite some texts that will try to
convince you that phenol-formaldehyde plastic By extension, it is possible to devise a theoreti-
was only available in dull, dark colours, the cal system ofseparating plastics into groups
same resin, used unfilled for items made by based on density, or specific gravity, which is
casting rather than moulding, could be all col- assessed according to whether the material
ours of the rainbow, and was used to imitate sinks or floats in solutions of known specific
jade, amber, onyx, and similar materials. Nor is gravity, such as saturated sodium chloride. In
everything dark and mottled brown necessarily practice, however, additives and fillers in arte-
phenol-formaldehyde. fact materials may alter the density to the extent
that this method becomes, at the best, very
Where plastics are concerned, appearances can unreliable. Tables III a and III b show values
be deceiving. The visible attributes of a piece for various plastics compared with the specific
colour, design, function, feel may indeed - gravities ofsaturated solutions. It is noticeable
-
hold clues to the identity of the material, but we that several plastics have a wide range of possi-
should also remember that plastics were often ble density, some of which bracket the satu-
deliberately fabricated to resemble other materi- rated solutions. It is also possible for moisture-
als, and a wide range of surface finishes was absorbent fillers to affect the flotation response
possible. by absorbing water from the solutions. Out of a
group of casein and melamine-formaldehyde

398
 Table II
Physical Characteristics of Sheet Materials

Material Appearance Stretch Tear Characteristics

Characteristics
Cellophane Transparent, very flexible; Fairly hard to stretch, Easily torn, tears in an
crackling sound when crumpled. medium extensibility. undulating line. Tear
Coated films are water-resistant Permanent may be hard to initiate.
and will mist if breathed on. deformation.
Uncoated films become limo
when wet. Holds a crease well.
Cellulose acetate or Transparent, medium flexibility. Hard to stretch. No Easily torn. Tear is
triacetate Does not crumple. Holds a permanent initiated easily from a cut
crease well. deformatron. edge, less easily from a
torn edoe.
Polyamide Transparent, very flexible. Holds Hard to stretch. No Hard to tear, even when
a crease well. Dermanenl initiated with a scissor
deformation. cut. Edqes of tear stretch.
Polycarbonate Transparent, usually as a sheet
scveral millimctres think
Polyethylene Slightly cloudy. Flexibility Easy to stretch, Fairly hard to tear, even
depends on thickness. Slightly depending on when tear initiated with a
waxy feel. Holds a crease thickness. Permanent scissor cut. Edges of
moderatelv well. extension. tear stretch.
Poly(ethylene Transparent. Very flexible to Hard to stretch. No Tears with ditficulty, and
terephthalate) moderately flexible, depending permanent extension. only if rnitiated with
on thickness. Holds a crease scissor cut. Thinner
moderately well; thinner grades grades tear more easily.
better than thicker. Characteristic rough
edge to tear, especially
in thicker grades, which
separate within the
thickness as if laminated.
Poly(methyl Transparent. Used as sheets
methacrvlate) several millimetres in thickness.
Polypropylene Transparent, hard, and relatively Hard to stretch. No After initiation with
inflexible. Does not hold a permanent extension. scissor cut, tears easily
crease well. in a straight line with a
cloudv edoe to the tear.
Polystyrene Transparent, relatively inflex- Very hard to stretch. Very easy to initiate and
ible. Characteristic metallic No oermanent propagate. Tears cleanly
sound when flexed. Holds a e)dension. but not straight with a
crease well; crease line shows cloudy edge to the tear.
white crazino.
Poly(vinyl chloride) Transparent. Softer feel and Hard to stretch. No Reouires a scissor cut to
(rigid) more flexible than polypropylene oermanent initiate. Fairly hard to
or polystyrene. Holds a crease deformation. tear. Slightly ragged tear
we!!: crease shows a small with some stretching of
amount of white crazinq. edoes
Poly(vinylidene Transparent, very flexible. Fairly easy to stretch. Tear is hard to initiate.
chloride) Sofiest of the sheet materials. Medium extensibility Tears easily, but not in a
Easily crumpled and holds a (less than straight line.
crease well. (May depend on polyethylene.)
orade.)

399
samples, some floated in saturated magnesium element within the plastic is blocking or skew-
chloride, some sank, and some progressed from ing the result, or whether there is a problem in
one state to the other over a period of time. experimental method.

Manufactured plastic objects also present a Burn Tests: Pyrolysis, or 'heat and bum' tests,
considerable problem in terms of their chemical can be applied to plastics in all forms. These
structure. The vast majority of literature refer- tests aim to identify plastics by their reactions
ences apply to the testing of plastics in a fairly to heat and flame, based on observations such
pure form. Unforhrnately, plastics are very sel- as whether a sample melts or decomposes,
dom pure, unaltered materials. They may con- whether it will continue to burn after being
tain not only the original base plastic, but also ignited, what colour the flame is, and the odour
structures altered by breakdown and crosslink- of any fumes given off.
ing, copolymers, similar chemical structures,
and plasticizers and other additives, all of With the exception of the Beilstein Test,13
which combine to make the task of identifica- which identifies poly(vinyl chloride) or
tion a more complex one, since removal of such poly(vinylidene chloride), the results of such
additives is usually difficult, if not impossible, tests are of limited practical value, for several
by simple means. The failure of a test to work reasons. Observations have to be made very
on a plastic that it is supposed to identiff then rapidly, especially in the bum test, and judge-
raises the question as to whether an unknown ments, for example on the colour of the flame,
are very subjective. This is reflected by the lack
of agreement between published sources, as
Table III a demonstrated in the comparative data in
Table IV, so that interpretation of the results
Specific Gravity of Selected Plastics
becomes something of a guessing game.
Specific Gravity Ranges of Plastics
ABS 1.04-1.10 Experimentation revealed fewer easily identifi -
Casern 1.26 Pure
1.35 Plastrc able differences between flames than the
Cellulose acetate 1.25 - 1.35 sources would suggest, and though the results
Cellulose nrtrate 1.34-1.38 Plastrc
usually agreed fairly well with one source or
'1.45 Frlm
Ebonrte 1.08 - 1.25 Unfrlled another, they did not agree with the same
1.25 - 1.80 Frlled source consistently. These procedures also
Nylon 1 01 - 1.16
Phenol-formaldehyde 1 27 - 1.3O Unfilled required the removal of larger samples than
1 36 - 1 46
Cellulose would usually be acceptable for museum
Frlled artefacts.
1 .54 - 1.75 Mrneral
Frlled
1 .75 - 1.92 Mrneral or One variation of this identification method
Glassfibre
Frlled
often suggested, which may be of more use
Polycarbonate 120-122 when dealing with plastics in the sense that it is
Polyethylene 0.91 - 0 95 not necessary to remove a sample, is to apply a
Poly(ethylene terephthalate) 1 38 - 1.41
Poly(methyl methacrylate) 1.16 - 1.20 hot needle in an inconspicuous place, observe
Polypropylene 0 85 - 0.92 the behaviour ofthe plastic, and sniffthe odour
Polystyrene 1.04 - 1.08 of the fumes given off. A number of difficulties
Poly(vrnyl chlonde) 1.19 - 1.35 Plastrcrzed
1.38 - 1.41 Rigrd are apparent here, though the method is useful
Urea- and Melamine- as a basic means of distinguishing between
formaldehyde 1.50 Pure
thermosets and thermoplastics.
1 80 - 2.10 Frlled

Using buttons as experimental samples for the

Specific Gravities of Saturated Aqueous Solutions
Water 1.00
most part, this method was successfully used
Sodium Chlonde 1.20 on casein, shellac, phenolics, acrylic and what
Magnesum Chlonde 1.34 appear to be two grades of polystyrene. How-
Calcium Chloride 145
Zinc Chloride 2.01
ever, there were instances where the identifica-
tion was uncertain. For example, it was difficult

400
 Table III b
Specific Gravities of Selected Plastics in Relation to Standard Saturated Solutions

HzO NaCl .MgOlz CaClz .Zn9lz

Polypropylene . .

-
Polyethylene- . : . :

Nylon

Polystyrene
-
ABb..
. unfrlled frlled

Poly(methyl methacrylate)

Plastrcrzed PVC -
Polycalbonate

. Cellulose acetaie

. prr" . plastic
Casein | |

. unfrlled . cellulose frlled mtneral ftlled mrneral or glass ftbre ftlled.

. Phenol formaldehyde
- : .

plastic ftlm
. Cellulose nrtrate

Polyethylene terephtalate :

. Brgrd PVC -
-
Urea- & Melamrn"-torr"td"nyd" ?ut"

08 0.9 1.0 1.1 1.2 1.3 1.4 1 5 1.6 1.7 1.8 1.9 20 21
Specrfrc Gravrty

to distinguish among urea-formaldehyde, mela- The squeak seems to be due to a small amount
mine-formaldehyde and casein plastics, as they of moisture being tumed into steam, and can be
all smelt similar. heard also from other water-absorbing plastics
such as urea-formaldehyde and casein. Vegeta-
There was one type of button identified, or ble ivory closely resembles a cellulose nitrate
rather separated from the rest, only by the smell plastic in visual appearance, and also tests nega-
ofburnt caramel and the fact that it sque-ked tive for both starch and cellulose. This is a very
when touched with the hot point. Nowhere in specific example, and vegetable ivory is un-
the reference material was there mention of likely to be found in large pieces, but it
bumt caramel as an identifying odour, let alone demonstrates a potential problem.
the squeak, which left alarge question mark.
Sometime later, when a batch of samples was It is obvious, both from these tests and from
analysed by infra-red spectroscopy, it was dis- previous experience, that assessment and de-
covered that these buttons were made from scription of flame colour and odour is very sub-
vegetable ivory, and that such buttons were jective, and this is corroborated by the fact that
rather common at a certain point in time, due to published lists often disagree, or are incom-
the use of vegetable ivory nuts as ballast in plete. Odour identification can be made more
cargo vessels. reliable by having a series of reference samples,

401
 Table IV
Heat and Burn Tests: A Comparison of Results

Material Source-
Source (films Krause and Langes SaudersT Other**
and sheets)
Cellophane a easy a. nla a nla
b contrnues b nla b nla
c light like paper c. nla c- nla
d. burns quickly and completely d. nla d n/a
e burnt paper e. nla e rva
f. neutral t. nla t. nla
Cellulose acetate a. easy a. nla a. nla
or tnacetate b extinguishes b. continues b. continuesls
c- dark yellow, sooty c. yellow c yellow-green^edge2o
d. melts, forms drops d. n/a d. burns slowlf '
e. acetrc aod e. acetic acid e. n/a
f. acrdrc I aodrc f. n/a
Fluorocarbon a. non-rnflammable a. nla a. no flame
b. n/a b extingurshes b. n/a
c. nla c. yellow c. nla
d nla d. bums wrth drfficulty, ciars slowly d. sample deforms slowllP
e hydrogen fluonde or chlonde e. none e. r/a
f stronolv acrdic f. aodlc f. n/a
Polycarbonate a drffrcult a. drtfrcult a nla
b. extrngurshes b. continues b n/a
c. lumrnous, sooty c. yellow, smoky c nla
d. melts, decomposes, chars d. n/a d. n/a
e. non-characteristic e. phenolrc e. n/a
f. neutral, weakly acidic Inrtially t nla l. nla
Polyethylene a diffrcult or easy lumrnous a. nla a. nla
b. n/a b. contrnues b n/a
c. blue centre c. yellow with blue base c. no smoke, blue flame22
d. melts, forms drops d. clearwhen molten d. n/a
e. hke extrnguished candles e bumrng candle wax e. nla
f. neutral f. rva I. nla
Poly(ethylene a. drffrcull a. nla a. nla
terephthalate) b. continues b. continues b. rgnrtes but extngurshes22
c. yellow-orange, sooty c. yellow with blue base c. n/a
d softens, melts, dnps d. melts sharply to freejlowrng hq- d. n/a
e sweet, aromatc urd that can be drawn Into a trbre e. rYa
t e. sweet t. nla
^/a t. nla
Poly(methyl a. easy a. nla a. nla
methacrylate) b. conlinues b. continues b. n/a
c. luminous yellow, blue edge, shghtly sooty c. yellow with blue base c whrte trp?, blue flame
d. softens d nla d. rva
e. sweet, fruity e methyl methacrylate e odour sweet, fruiw,
f neutral L nla floralm
I nla
Polypropylene a. drtfrcult or easy a. nla a.nla
b. contrnues b. clnlrnues b. n/a
c. lumrnous wrth blue centre c. yellow wrth blue base c no smoke, blue flame22
d. melts, forms drops d. becomes clear when molten d. rYa
e. hke extrngurshed candles e. buming candle wax e. n/a
f. neutral t. nla {. rva
Polystyrene a. easy a. nla a nla
b. continues b. contrnues b n/a
c. lumrnous, sooty c yellow wrth blue base c. nla
d. softens d. very smoky d. melts Into clear hqurd .
e. sweet (styrene) e. sryrene e. smells hke mangolds''
f neutral t. ila I nla
Poly(vrnyl a drlfrcult a. nla a. nla
chlonde), b. extrnguishes b extingulshes b. n/a
poly(vnylrdene c. yellow-orange, green border c yellow wrth green base c. nla
chlonde) d. softens. decomooses to black d. n/a d. nla
e. hydrogen chlonde e. acrid e n/a
f. stronqly acrdrc f. aodrc t. nla

402
 Table IV (continued)
*Results from the two major sources are paraphrased to save space. The letters in the Table represent
the categories given in the legend below. In some instances information or findings may not be
available for a particular category and thus will be marked n/a.
a. ease of ignition
b. whether samples continue burning after removal from flame
c. flame colour
d. behaviour during burning
e. ogour
f. vapour acidic or alkaline
**
Other sources are quoted only when their results differ significantly from the major sources.

The numbering of Table sources conesponds to the main list of References.

5. Krause and Lange, 1969, pp.4-8.
7. Saunders, 1966, pp. 15-19.
19. Briston, 1974, p.289.
20. Nechamkin, 1943.
21 . Taylor, 1985.
22. Bird. 1963.

but it is impossible to tell how fillers and addi- and hooked over the lip ofthe test-tube during
tives will affect the odour, and furthermore, heating. Results ranged from a pH of 0.0
from a health and safefy point ofview it does to 0.5, given by poly(vinyl chloride) and
not seem sensible to be inhaling fumes, many poly(vinylidene chloride), to a pH of 9.0 to
of which are toxic. 10.0 given by polyamide (Table V). Here
again, the presence of additives in manufac-
Last but not least in the list of negative com- tured items makes this a suspect method for
ments, a fair degree of damage can be caused to artefacts. As a footrote, since the Merck strips
plastics by a hot needle. are supported by a poly(vinyl chloride) back-
ing, it is impor|ant to avoid heating the strips
pH Tests: The one aspect of these tests that themselves, or the resulting pH may be
proved both fairly accurate and potentially use- too low.
ful for preliminary identification of sheet mate-
rials was the pH determination. Good results Solubility Tests: Some initial consideration
were obtained using Merck non-bleeding pH was given to the identification of plastics by
indicator strips, dampened with distilled water solubility, but an elimination scheme such as
that in Table VI requires a considerable number
Table V of different solvents, some of which are
pH Values ofVapours Releasedfrom Sheet carcinogenic, such as tetrachloromethane and
benzene, and generally no longer found in con-
Materials During Heating
servation laboratories. As in the case of the py-
rolysis tests, sample sizes would also be
Material Approximate pH prohibitively large.
PVC/PVDC 0.0 to 0.5
triacetate
Celf ulose 2.5 Overall, then, various aspects of the test meth-
Polyethylene/Polypropylene 3.0to4.0
Poly(ethyleneterephthalate) 4.O ods mentioned so far may be of some use in
(Mylar) preliminary identification of plastics, but it
Polycarbonate 4.5 to 5.5 must be emphasized that the results are often
Polystyrene 5.5
subjective and, as such, should be interpreted
Polyamrde (nylon) 9.0 to 10.0
with caution.

403
Spot-Tests: Moving onward to colour change give a very convincing positive result. A dusty
reactions, it is important that the tests be as polypropylene lid, partly cleaned, gave a nega-
simple as possible (i.e., ideally they should in- tive result on the clean area, but as the solution
volve: only basic laboratory equipment, such as came into contract with the dust on the other
test-tubes andpipettes; as few reagents as possi- half of the lid, the characteristic blue colour ap-
ble ; and uncomplicated proc edures). Otherwise peared. The point here is that for reliable re-
a certain degree of investment in time, money sults, the surface to be tested must be clean, and
and equipment is required. In doing early re- it is entirely possible for a cracked surface, or
search into published plastics identification one that is oozing plasticizer, to give a positive
methods, the aim was to find a spot-test for result due to the presence ofdust or other con-
each material, along the lines of spot-tests taminants in the cracks or adherine to the sur-
used to identiff metals. These self-imposed face, even after wiping.
limitations rule out a considerable number of
published tests that require complex apparatus The test is extremely sensitive, requiring only
or esoteric or expensive reagents, but even so- a tiny sample, and it is essential to beware of
called "simple" tests may not be so simple in cross-contamination. A pipette accidentally
practice. As with all the other identification touched to the surface ofa cellulose nitrate
methods mentioned so far, some of the results object can subsequently produce a false
have been good, and some have been most positive if used for a non-cellulose nitate ob-
puzzling. ject. A side-effect ofthis sensitivity is that the
presence of cellulose nitrate as a component of
One procedure that generally gives good, defi- an object, even in tiny quantities, can result in a
nite results is a test for the presence ofacetates mistaken identification.
that requires the application of four pre-mixed
reagents in succession (Table VII). The Lalique This can lead to ambiguities in the identifica-
box mentioned earlier was identified using this tion of early film materials. One piece that was
test after the smell of acetic acid alerted us to sent to us for identification gave a positive re-
the likelihood that it was not cellulose nitrate. sult with the diphenylamine test around the
The positive colour may vary from pale pink to edges ofthe sample, but not on the flat sur-
burgundy red, depending on the size of the sam- faces. Removal of the emulsion from the base
ple relative to the volume of reagent. plastic and retesting revealed that the base
plastic was cellulose acetate. The cellulose ni-
To date only one problem has occurred when trate was eventually tracked down to a thin sub-
using this test. One pair of spectacle frames layer between the emulsion and the base
smelt strongly of acetic acid, implying cellulose plastic.
acetate, but showed breakdown pattems charac-
teristic of cellulose nitrate, and the plastic Another drawback to the diphenylamine test is
tested positive for cellulose nitrate and negative the mark left on a glossy surface, and even the
for cellulose acetate. As yet, no answer to this potential for a colour change.
apparent contradiction has been found.
It is also worth noting that this_te_st i_s men-
The Diphenylamine Test for Cellulose tioned by a variety oi.ou...r,7'15'16 brrt there is
Nitratera: The reagent is a solution of dipheny- considerable disagreement as to the concentra-
lamine in concentrated sulphuric acid. The acid tions of diphenylamine and sulphuric acid re-
liberates nitrogen oxides from the surface ofa quired. Given concentrations range from 0.5%
cellulose nitrate object, and these oxidize the diphenylamin e in 90o/o acid to 20% dipheny -
diphenylamine to a dark blue dye, which ap- lamine in 100% acid. Trying a range of solution
pears virtually instantaneously on applying the concentrations revealed some very inconsistent
reagent. This does indeed work very effec- results. The best, most consistent results were
tively on cellulose nitrate. However, any oxidiz- obtained from a 50% solution of diphenylamine
ing agent will give the same result. A surface in75o/o to 85% sulphwic acid.
contaminated with normal museum dust can

404
 Table VI
Schemefor Identification of Polymers by Sotubility (After Briston Ie)

Sample
I

Cellulose tests' posrrve-tS tl

negative
I

add hot toluene soluble add ethyl acetate soluble

I to new samole
I

I t-
Insoluble
I
- - Insoluble
-> @ r----:-

decant toluene,
add ethyl acetate
soluble add amyl
to new sample
I
fornate
@
v
-> insoluble
I
->soluble
in.luut" V
F"rhloseaceECl
I
add amyl acetate soluble - or cellulose I
I to new sample acetate-butyrate
I I
I
I

I
insotubteS
I
I
->
remove ethyl acetate > soluble
then add water
I

insoluble
I

dry,thenaddCCl+
to new sample
->soluble

ru
I I

Insoluble -addacetone
in"otl,ot"

v
I

add droxan to soluble -soluble

new sampte
I
I

insoluble
-
I

add 40% HzSOn

lo new sample
I

->soluble
riiSOruOrg

- Cellophane is eliminated first,

using the following criteria:
a) Burn test gives odour of burning paper.
b) Tear strength very low.
c) Hot wire or match-head applied to sample does not melt it.

405
Conclusion sulphate solution followed by l0% sodium
hydroxide solution, gave dubious results when
It is true of all the wet-chemical tests that they used on a sample from a casein plastic button,
identiff one component of a polyrneric sub- though the same test performed on a casein in-
stance without revealing whether it is a pure painting medium gave a positive result. The
substance or a copolymer. To identify a copoly- strong purple colour seen on the gelatin sample
mer requires deliberate testing for both poly- was present on the casein, but only faintly,
mers, with the added complication that the probably due to the treatrnent of casein with for-
presence of one may mask the other. A positive maldehyde to create the plastic. Pigmentation
test result for one particular plastic tends to in the casein totally obscures the result, since
lead to the assumption that it is unnecessary to the colour is developed on the plastic surface,
test for other plastics in the same sample. In not in the solution, so that the test can only
reviewing the literature there is no mention of a be performed effectively on pale-coloured
tested plastic giving results that seem to indi- objects.
cate the presence of lwo polymers, but there is,
for example, a publication from the 1920s that Another factor that can interfere with correct
states that casein compounded with cellulose identification is the assumption that a hade
nitrate wqs"actually being used more often than name product will always have the same com-
celluloid." Whether or not this is accurate, position. This problem arose in connection with
most of us would look no firther than the posi- Saflip coin holden for the storage of coin col-
tive result for cellulose nifate. The question of lections. According to the manufacturer's infor-
copolymers complicates analysis to the extent mation, and a 1982 analysis, the Saflip holders
that modern, multi-component plastics are were made from poly(ethylene terephthalate),
essentially unidentifiable by such methods. better known as Mylar, but in terms of physical
properties and solvent testing, they behaved un-
A further problem is discovered in cases where like any identified sample of Mylar. Based on
the applied reagent solution produces a colour tearing and folding tests, and solvent suscepti-
change on the sample rather than in the solution bility, the best guess was that they were a lami-
itself. A spot-test for protein, using2%o copper nate ofcellulose acetate on the inside and

Table VII
Testfor the Presence ofAcetates

Requirements: a) solution of 6% potassium hydroxide in methanol

b) saturated solution of hydroxylamine hydrochloride in methanol
c) solution of 1% ferric chloride in water
d) solution of 1Qo/" hydrochloric acid

1. Place sample in test-tube. Add 1 mL of solution a), then 1 to 2 drops of solution b). Shake gently
and leave to stand for at least 3 minutes. Volume of solutions added should be altered in proportion
to amount of sample.

2. Add 1 drop of solution c) and shake again.

3. Add solution d), one drop at a time, shaking gently after each addition. Up to 25 drops may be
required in some cases. A burgundy-red colour indicates the presence of cellulose acetate or
poly(vinyl acetate). Pale yellow is a negative result. Pale purple-red is either a dilute positive result
(repeat test, if possible, using smaller quantities of reagents), or may indicate cyanoacrylate or
cellulose nilrate.

Alternative Procedure: some forms of this test suggest warming the mixture gently after the addition
of solution b). This step probably has the same effect as leaving it to stand for a lew minutes.

406
polystyrene on the outside. A second and identification of plastics, but we should be
more recent instrumental analysis revealed that wary of relying on them solely or implicitly.
Saflip holders are made from glycol-modified
poly(ethylene terephthalate); a small change, R6sum6
seemingly, but enough of an alteration that the
plastic was not correctly identifiable by simple Les difficaltds pratiques lides d I'identijication
means. des plastiqaes

One question yet to be addressed is that of Les objets en plastique qui sont maintenant
sample taking. Here arises the same dilemma conservds dans les musdes et dans les collections
as with the identification of any material from privdes ddfinissent une discipline nouvelle pour
les spdcialistes de la restauration, un domaine
which a sample must be taken. Modem
dans lequel les recherches ne commencent qu'd
analytical equipment allows the removal of
peine et oi il n'existe quefort peu- s'il en est
ever smaller portions of the object under study,
de directives de traitement. Or, contrairement d
-
or in some cases none at all, but the grosser ces matidres qui, tels la ctramique ou le mttal,
identifi cation methods outlined earlier require resteront stables pendant longtemps si elles sont
rather larger amounts for accurate results. The mises en rtserve dans un milieu appropri2, les
diphenylamine test for cellulose nitrate is ex- plastiques risquent de se dtgrader, parfois avec
tremely sensitive and can be performed under a une rapiditi dtonnante. Et lefacteur qui amor-
microscope with a tiny sample. Other tests may cera une telle ddgradation sera parfois, en ap-
require a small shaving from an inconspicuous parence, aussi anodin que cette couche de laque
place. Spot-testing on the surface, as we might que I'on applique avant d'apposer le numiro
do for metal spot-tests, is risky, since many of d'enregistrement.
the reagent solutions will mark a plastic sur-
face. On the other hand. it does not involve re- Les plastiques rdagiront de fagon trds varide face
moving any part of the object. In many cases it d la chaleur, d la lumiire et d d'autres facteurs
may well be possible to further refine the tests li6s atu conditions ambiantes ou aux agents de
to reduce the quantities involved. The impor- nettoyage, et ils auront parfois eux-m€mes des
tant point is not to drown a small sample in effets marquds les uns sur les autres. C'est
donc dire toute I'importance de bien identifier
reagent, or the colour may be indeterminate.
les polymires avant d'amorcer tout traitement.
In finishing, having made all the foregoing
Si la spectroscopie infrarouge demeure la
negative comments, it is necessary to redress
meilleure mdthode pour identifier de tels matt-
the balance a little. From everything said, it iaux, il n'en demeure pas moins que l'on doit
would be natural to think that the conclusion is aussi disposer de mdthodes d'identification plus
that instrumental analysis is the only way to go. simples, qui puissent €tre couramment appliqutes
However, along with the problems, some of tout aussi bien par les spdcialistes de la restaura-
which have been outlined, there have also been tion que par les responsables des collections. Les
some very good results from using the kind of documents portant sur les plastiques offrent nom-
spot-tests and simple physical determinations bre d'exemples de telles mdthodes, dont certaines
discussed earlier. Certainly it is not the claim sont scientifques et d'autres, plus empiiques.
that these methods can replace full instrumental Les essais deflottation sefondent ainsi sur des
analysis, but if all we need to know is what the solutions de masse volumique connue pour distin-
base plastic is, and particularly if we have a guer les divers genres de plastiques; les essais de
good idea of what the plastic might be on stylis- combustion, sur I'odeur qu'ils dtgagent et sur
leurs rdactions d la chaleur et aufeu; les analy-
tic or visual grounds, then it may well be possi-
ses chimiques ponctuelles, sur les changements
ble to prove it with a specific test. Working
de couleur que provoque le procddi expirimental
towards an elimination scheme for the identifi- qui est utilisd. L'identification d'apris la marque
cation of complete unknowns is a project cur- de commerce (si elle est connue) ou d'aprds le
rently underway that may also prove useful. style ou le genre particuliers de I'objet demeure
Overall, the experience has been that many of une approche plus subjective. Et, I'expdrience
the outlined methods are of assistance in the acquise au terme de plusieurs annies de travaux

407
portant sur des objets en plastique, ainsi que sur 5. Krause, A. and A. Lange, Introduction to the
des plastiques tant anciens que nouveaux qui sont Chemical Analysis of PlasDcs, trans. J. Haslam
conservAs d des fins expirimentales, dimontre (London: Iliffe Books Ltd., 1969).
assez clairement les limites de telles mithodes,
limites qu'il est essentiel de connaitre si I'on veut 6. Braun, D., Simple Methods for ldentification
faire des identffications qui seront un tant soit of Plastics,2nd edn. (Munich: Hanser, 1986).
peu certaines.
7. Saunders, KJ., The ldentification of Plastics
La principale dfficalte tiendrait aufait que les
and Rubbers (London: Chapman and Hall,
documents portant sur les plastiques prdsentent
souvent des essais qui s'appliquent bien d des 1e66).
6chantillons de polymdre pur. Et il est plut6t rare
que I'on trouve des articles fabiquts en plastique 8. Haslam, J., H.A. Willis and D.C.M. Squir-
qui n'aient subi aucune modification. Qui plus rell, Identification and Analysis of Plastics,2nd
est, les charges et les pigments influent sur la edn. (London: Iliffe Books Ltd.,1972).
masse volumique des plastiques et sur leur ap-
parence, et interfdrent donc avec les essais qui 9. Baker, M., D. von Endt, W. Hopwood and
sefondent sur les changements de couleur, tandis D. Erhardt, "FTIR Microspectrometry: A Pow-
que les plastifiants modifient les proprietds des erful Conservation Analysis Tool" in: Pre-
polymires de maniire imprevisible. Sans compter prints of the l6th Annual Meeting of the
que la digradation elle-m€rne produit des modifi- A m er ic an I ns ti tute for C o n s erv a ti o n of H is tori c
cations chimiques. and Artistic Works, 7988.

Par ailleurs, le recours aux textes d'histoire de 10. Cardamone, J.M., "Reflectance Absorption
I'art pour identi/ier les plastiques n'est pcts sans
Fourier Transform Infra-red Spectroscopy for
poser des dfficultds d'un tout autre ordre.
Nondestructive Chemical Analysis," rn'. Pro-
ceedings of the l4th Annual IIC-CG Confer-
Dans la prisente communication, nous abor-
derons ces probEmes enfaisant rdfirence d ence, 27-30 May (Toronto: IIC-CG, 1988).
des expiriences rtelles qui ont ite faites en vue
d' 6t ab I ir un sy s t ame d' identifi c atio n c o hd re nt I l. Newman. T.R.. Plastics as an Art Form
pour les plastiques. (Philadelphia: Chilton Book Co., 1969).

References 12. Kaufman, M., The First Century of Plastics

(London: The Plastics & Rubber lnstitute,
l.KaIz,5., Classic Plastics (London: Thames 1963).
and Hudson. 1 984).
13. Williams, R.S., "The Beilstein Test," CCI
2.Baer, N.S. and M.J.D. Low, "Advances in Noles No. 17/l (Ottawa: Canadian Conserva-
Scientific Instrumentation for Conservation," tion Institute, 1989).
in: N.S. Brommelle and G. Thomson, eds.,
Science and Technologt in the Service of 14. Williams, R.S., "The Diphenylamine Spot
Consertation: Preprints of the Contributions Test for Cellulose Nitrate in Museum Objects,"
to the Washington Congress, 3-9 September CCI Notes No. l7l2 (Ottawa: Canadian
(London: IIC, 1982). Conservation Institute, I 988).

3. Delmonte, J., "Effect of Solvents upon I 5. Browning, 8.L., Analysis of Paper, 2nd,
Organic Plastics," Journal of Industrial and edn. (New York: Marcel Dekker Inc., 1977).
Engineering Chemistry, vol. 14, no.6, 1942.
16. Shaw, T.P.G., "Systematic Procedure for
4. Sparke, P., ed., The Plastics Age (London: Identification of Synthetic Resins and Plastics,"
Victoria and Albert Museum. 1990). Journal of Industrial and Engineering Chemis-
try, Analytical Edition, vol. 16, no.9,1944.

408
17. Tague, 8.L., Casein, Its Preparation, 2l.Taylor, Thomas O., "The Use and
Chernistry and Technical Utilization Identification of Plastic Packaging Films for
(New York: D. van Nostrand co., 1926). conservation," in The Book and paper Group
Annual. American Institute for Conservation of
18. Coxon, H.C., "An Evaluation of Tests for Historic and Artistic Works. 19g5.
the Identification of Polymeric Materials Used
in Conservation," 1986, unpublished. 22.8kd,V., "How to Identiff plastic Films,"
Technologt'vol' 9' no' 9' 1963'
1g. Briston, J.H.,plastics Films,plasic, "lastics
Institute Monograph (New York Halstead
Press, John Wiley and Sons, 1974).

20. Nechamkin, H., "A Schematic Procedure

for Identification of Common Commercial
Plastics, " Journal of Indus trial and Engineer-
ing Chemistry, vol. 15, no. 1, 1943.

409
The Identification and Characterization of
Acrylic Emulsion Paint Media

Carol Stringari Ellen Pratt

The Museum of Modern Art Conservation Center

New York, N.Y. Institute of Fine Arts, New York University
U.S.A. New York, N.Y.
U.S.A.

Abstract make informed decisions about preservation

where this is appropriate.
An ongoing project to study the properties of
acrylic emulsion paints is underway at The The range of materials used in this epoch is lim-
Museum of Modern Art, New York. Approxi- itless. The manufacture of synthetic products
mately 150 paintings in the collection were sur- for industrial use escalated after World War II
veyed to study the medium and assess condition. and artists began to experiment with these
Routine examination and testing were done in the materials in their creative endeavors. Acrylic
laboratory and a body ofstandards was prepared
emulsion paints as an artist's medium became
for comparison. Three case studies were chosen widely available in the late 1950s to early
to explore more specific methods of identification.
1960s and were fundamental in resolving some
These methods included microscopic analysis and
st ai n i ng, F ou i e r Tran sfo rm I nfrare d Sp ec tro s -
contemporary aesthetic concerns. Color-fi eld
copy (FTIR), and Pyrolysis-Gas Chromatography painters, pop artists, and hard-edge minimalists
(Py-GC). gravitated quite naturally towards this medium.

Introduction The availability of acrylic paints directly coin-

cided with a prevalent rejection of traditional
Marcel Duchamp, when asked about his deci- values and romantic notions, and an increased
sion to stop painting and pursue chess as an art interest in rational objective geometry. Some of
form, stated: the properties of the emulsion paints germane
to artistic production during this period were
I have not stopped painting. Every picture has their immediacy (fast d.ying and water sol-
to exist in the mind before it is put on canvas, uble), flexibility, and ability to achieve a hard
and it always loses something when it is edge and a slick "contemporary" surface not
turned into paint (Schwarz I 969). prone to cracking.

During the twentieth century, creative expres- Artists such as Kenneth Noland, Jules Olitski,
sion was redefined and existing notions of aes- and Ellsworth Kelly rejected the lyrical expres-
thetics were challenged. Modem works often sionist work of artists such as Willem de Koon-
require analysis beyond material charac- ing and Jackson Polloclq stressing objectivity
terization, and in the case of conceptual works, and pure visual configuration. These artists
there are no existing materials to identify. None- utilized acrylic paints to realize their formalist
theless, conservators are faced with the task of concerns.
undentanding physical attributes in order to

4tl
Concurrent to formalism was an inclination for substitution of a methyl group for the alpha-
mass media and pop culture, where art was pro- hydrogen in the acrylate monomer produces a
duced and simultaneously called its role into methacrylate monomer. The polymer formed
question. These ideas were embodied in the by methacrylate monomer has more steric hin-
work of artists such as Andv Warhol and Rov drance than a poly-acrylate and therefore forms
Lichtenstein. a harder polymer with higher tensile strength
(Figure 2). Since the polymers are insoluble in
Although materials ultimately act as vehicles, water and must be suspended in a water phase,
subservient to ideas, their properties and limita- a surfactant, which is essentially a soap, is used
tions can often influence the realization of a to facilitate this dispersal (Becher 1983). (See
concept. Acrylics facilitated the evolution from footnote for monomer abbreviations.) Base
heavy paint and gestural brushwork to sharp emulsions, such as Rhoplex (or Primal in the
clean surfaces and precise outlines. u.K.) AC-234 (66%PEN34% PMMA),
Rhoplex AC-33 (66% PEN33% Pil/'N 1%
This study investigates the acrylic emulsion acrylic acid), Rhoplex AC-22, or Rhoplex
paints manufactured for use as artists' materials AC-507 (PnBMA/PMA), or any of awide
during this revolutionary period of the 1960s range of acrylic emulsions with varying speci-
and early 1970s. ficity and properties, are purchased by paint
manufacturers.
Composition of Acrylic Emulsions
The base acrylic emulsions are composed of ho-
mopolymers or copolymers of acrylate and/or Tg ('C) Vs NONOUER COmSrTIoll
m - XETAYL EmAYIATE - gAru
methacrylate monomers (Figure 1), which un- U - EIFYL AGYUTB - SOm
S - BU?YL AGYUTE - SOm
dergo free radical emulsion polymerization
(Bennett et al. 1968). By varying proportions
of these monomers a resin can be tailored for
specific properties, such as adhesion, glass
transition temperature, tensile strength, and
elongation (Bondy 1968). For example, the

-20

I
--r00
- cHr- CH - - cH,-c
-l -
I rc UONOER
c- o-cHr ,-c-o-cH3 O 20 {0 60 80 lOOi
o- o.. som xoNom
Figure 2 Copolymer composition and its relation to
Eethyl acrylate {8) nethyl nethacrylate (m) glass transition temperahtre.

CHr
I

- cHr- cH - -cHr-9-
I I

;- o-cHzcHr /t- o - cHzc{3

o' o'
Abbreviations used for monomers:
ethyl acrylate (EA) ethyl Eethacrylate (m)
MA methyl acrylate
CHr EA ethyl acrylate
- cH,- cH -
I
MMA methyl methacrylate
I -cH,-q- I EMA ethyl methacrylate
,.c- o -cH,cHrCH,CH3 c-o-cH,cH,cH.cH,
c;' ot/ iBMA iso-butyl methacrylate
nBMA n-butvl methacrvlate
n-butyl acrylate (nBA) n-butyl Eethacrylate (nBe)
Polymers are distinguished from monomers by the
Figure I Acrylate and methacrylate monomers. prefix P, thus PMA is poly(methyl acrylate).

4t2
acrylic paints. RH is critical, as it will influence The Project
the rate of evaporation and alter the effect of
the co-solvent (e.g., ethylene glycols). Suitable The acrylic research project at The Museum of
rheological properties of the formulation are Modern Art began in 1988 as a survey funded
essential to ensure the proper film formation, by the Institute of Museum Services to examine
which can be adversely affected by excessive approximately 150 paintings purported to be of
foaming due to surfactants and dispersants, or a "synthetic polymer" medium. This generic
cracking due to excess filler material, such as category refers to many synthetic resins used
cellulosics or protective colloids. In recent for- in coatings dating from the early 20th century,
mulations the viscosity of the emulsions including, for example, cellulose esters, vinyls,
has been adjusted to facilitate film formation. PVAs, alkyds, and synthetic lacquers (Bria
Certain monomers, such as styrene and butyl 1981; Lodge 1988). Duringthe course ofthe
esters, have been used in these formulations to suryey, it became obvious that no firm conclu-
improve performance. This has resulted in in- sions could be reached from such a general
creased gloss, higher pigment load, and better study. However, this preliminary study was
resistance to weathering and water sensitivity essential to define the scope of the inquiry.
(Bondy and Coleman 1970). Acrylic resins, and more specifically acrylic
emulsions, were chosen as the topic for further
The acrylic emulsion paint films are readily sol- study.
uble in xylene, toluene, and acetone. In certain
cases, the films are soluble in hydrocarbons and Phase I of the project was the completion of an
even water, which can present problems for extensive examination form for each painting
conservation treatnents. This sensitivity to and the collection ofquestionnaires sent to each
water and petroleum benzine is probably due to artist. Photos were taken of the works and dis-
poor formulations or over-dilution, leaving the tinguishing characteristics were documented.
pigment vulnerable to mechanical removal. The Phase 2 consisted ofan investigation into appro-
solubility also presents a problem for artists priate methods for precise characterization of
who wish to apply a protective coating or acrylic emulsion paints.
varnish. To date there is no one simple
solution for a reversible varnish on acrylic It was diffrcult to actually identify these materi-
paint films. als through visual examination alone, since
their appearance was so dependent on manipula-
Thus, it becomes evident that these formula- tion of the material by the artist. Since the in-
tions are in a constant state offlux and itial media attribution for each painting in the
advancement and can vary considerably from survey was based on information from The Mu-
one product to the next and even from one year seum of Modem Art (MoMA) curatorial files,
to the next. Table I lists some categories of and not scientific analysis, it was imperative
additives, with examples for each and their that a less subjective means of characteriztng
function within the paint film (Gachter and these paint films be pursued. Furthermore, it
Muller 1985). was felt that information about chemical proper-
ties and variations in formulations could help
This briefreview ofthe general structure, facilitate treatment decisions and preventive
constituents, and properties of acrylic paints maintenance programs.
demonstrates the enormous range ofpossibili-
ties for furthering our understanding of these From the 150 paintings surveyed, a small
materials through testing and analysis. To make group, as representative as possible, was cho-
informed decisions about the treatment of these sen for closer examination and testing. As a
paint films, we felt it was necessary to design a practical study, the project advanced our under-
systematic means of identiffing and distin- standing of the artists' working methods, and
guishing them. elucidated potential obstacles encountered
when analyzing an actual work of art, with all
ofthe variables inherent in the creative process.

414
 Table I
Additives : Categories and Functions

Additive Purpose and Effect on Formulation

polyethylene glycol/glycols plasticizer/hygroscopiclfreeze-thaw stabilizer

phenyl mercury compounds/ fungicide/sensitive to hydrogen sulfide/can

mercury sulfide salts cause brown discoloration

caboxylic acids, sodium soaps of anionic surfactants/effective for

stearic and lauric acids, amine soaps, whites and extenders/hygroscopic
phosphates

alkyl phenoxy polyethoxy ethanol/ ionic surfactants/hygroscopic

alkyl aryl ethers (i.e., Triton)

cellulosics, casein, polysaccharides, thickening agents/provide proper

polyacrylic acids, gum arabic, consistency/can affect gloss/water
sodium alginate sensitivity/less surface smoothness

styrene imparts gloss/increases water resistance

sulfonated oils, sulfonated esters emulsifiers

ammonia pH adjuster

titanium dioxide/hydrated clay/calcium extender pigments/TiO2 may have reactive

cabonate/amorphous silica sites for water-free radical formation

soya lecithin dispersant

primary alcohol ethoxylates dispersant, wetting agent

tamol (chemical formula unknown) dispersant

Methodology method of application for achieving different

surfaces. Their paintings will be discussed as
We pursued the identification of acrylic three separate case studies.
resins and emulsions for three paintings using
Fluorescence Microscopy, Fourier Transform A body of material standards was created for
Infrared GnR) Microspecfroscopy, and use as reference and for comparison to samples
Pyrolysis-Gas Chromatography (Py-GC). The from actual paintings. Materials were painted
works selected were Flemish IrII (1973)by Al out onto canvas and/or mylar, sometimes as a
Held,Dec.l8,1979 (1979) by OnKawara, and single layer, other times as a laminate structure
Primary Light Group: Red, Green, Blue (1964 to simulate a painting for cross-sectional study.
to 1965) by Jo Baer. These artists painted Materials collected for standards included:
consistently with acrylic paints during the Liquitex acrylic emulsion gessos, paints, and
1960s and early 1970s, and each used a unique mediums; Golden acrylic emulsion products;

415
Daler Rowney Cryla acrylic emulsion products;
Rohm and Haas Rhoplex acrylic emulsions;
Lascaux acrylic emulsions; Magna acrylic resin
solution paints and medium gels; Lefranc and
Bourgeois Flashe vinyl emulsion paints; Bellini
oil paints; Grumbacher oil paints; Winsor and
Newton oil paints; linseed oil; and various
varnishes.

These standards now serve as the basis for a

library, which we are amassing for research and
as a resource for other conservators.
Figure 4 Detail of an acrylic emulsion paint surface
The standards and the samples from the illustrating pinholing and the afr.nity of the taclqt
surface for particulate matter, at 20x magnification.
actual paintings were all subjected to the same
scrutiny and the same tests. Each painting was
examined frst in visible light then in ultraviolet i:ii\\9;1f:

(UV) light, using two long-wave UV lamps.

Empirical observation of these materials has
proven helpful in the recognition of certain vis-
ual characteristics and autofluorescent or non-
fluorescent properties of synthetic materials.

In UV light, the acrylic paints have a charac- .\

teristic purple appearance, the result ofa reflec-
tion as opposed to a fluorescent emission. This
is distinguishable from oil paints, which emit a \
slight yellowish fluorescence. It is helpful to
have standards prepared to compare fluores- Figure 5 Detail of an acrylic emulsion paintfilm,
depicting the levelling and rounded edges of
cence, as the oil paint fluorescence can be diffi-
brushwork.
cult to recognize without the comparison. For a
discussion offluorescence ofpaint and varnish
see de la Rie 1982. [n contrast to the behavior
are applied too thickly because a dry skin will
form rapidly on the surface, disallowing water
of oils, the purple appearance of the acrylic
evaporation from below, and forcing the forma-
seems to override the primary fluorescence
tion of cracks to release underlying moisture
of the pigments.
and relieve stress.

Under stereomagnification such character-

On the whole, due to the numerous additives,
istics as pinholing or air bubbles, which form
thickeners, and pigment load, acrylic paints
while the paint film dries, can be discerned
tend to dry more matte than solvent-borne paint
(Figure 4). This tendency to froth is caused by
systems.
the addition of surfactants to form the emulsion.
The influence and complexity of the pigments
Impasto formed from acrylic paints is often
were not thoroughly addressed in this study, al-
rounded (Figure 5), as opposed to crisp and
though whenever possible pigment composition
peaked, which is more characteristic of oils.
was determined by polarized light microscopy.
These rounded edges ofthe brushwork are due
In many instances, results of analysis were
to levelling, a property determined by surface
clearly influenced or obscured by the pigment.
tension and the addition of thickeners to
Black and white pigments were chosen as the
decrease rapid capillary absorption into the sub-
focus of the study for the sake of comparison
strate (Patton 1979). Distinctive drying cracks
and to set some arbitrary parameters.
can occur in the surface of the acrvlics if thev

416
Microscopy/Fluorescence Staining conservation, and its accessibility and
The fluorescent staining technique involved ease of use in the laboratory (Talbot 1982;
staining cross-sections and thin sections to de- Wolber.s 1987). For systematic staining of all
termine whether any constituents or functional unknowns and knowns, Rhodamine B was se-
groups within the acrylic film could be lected to detect lipids and fluorescein isothiocy-
discemed. anate (FITC) was chosen to detect proteins
(Figure 6). These stains have been used on tra-
Experimental ditional painting materials to indicate the pres-
E quipment : A Leitz Labor-Lux Metallographic ence of functional groups contained in oil and
Microscope with a xenon light source fitted protein binders. In an attempt to isolate a stain
with a Ploemopak ,{2 filter pack for viewing useful for acrylics, or to arrive at a negative
in normal and UV light was used. The ,{2 cube attribution by elimination, samples from the
has an excitation range from 270 nM to body of standards and the three paintings were
380 nM and a suppression filter from 410 nM tested. Results of staining the acrylics proved to
to 580 nM. Each filter combination produces be quite confusing, and will be outlined in the
different results, due to specific excitation and case study discussions.
emission wavelengths. Broader transmission
results in a brighter, less contrasting, and Although the stains do aid in identiffing broad
generally less specific fluorescent image categories of traditional materials, such as fatty
(Becker 1990). acids, amino acids, and starches, it appears that
the acrylic emulsions are far too complex and
The method of histological staining was the formulations too variable to be able to rely
chosen because of its successful adaptation to on staining techniques. Although the acrylic
polymer backbone itself does not appear to
have sites that would react with histological
stains, there are a number of surfactants that

[!i-g:ryt:i:i: may contain fatty acid components or amine

groups that would react with lipid and protein
stains. Therefore caution should be observed
cr- when making judgements about the nature of
cooH
these materials through staining, and standards
should always be test-stained prior to staining
an unknown.

C2"H3,NrO3Cl mol wt : 4'19.O2

Results of this method for media identification
Rhodamine B appear to be highly dependent on technique, in-
cluding pH, solvent systems, and concentra-
tions (Clark 1983; Lillie 1990). Since acrylic
polymers are readily soluble in the suggested
carrier solvents for the stains, both FITC and
Rhodamine B were tested in water. The FITC
was buffered to a specific pH, as the literature
suggested (McKinney, Spillane, and Pearce
S:C: 1966). After staining, all samples were rinsed
in water with pH adjusted to that of the stain.
In all cases the results were more reproducible
C'H,,NO'S; mol. wt. 389.390 when the stain was carried in water and pH
carefully controlled.
fl uorescein isothiocyanate
Figure 6 Chemical structures of Rhodamine B and Microtomed thin sections proved to be the
FITC stains. preferred method of sample preparation for

417
staining. Thin sections measuring 5 pm were The acrylic paints are consistently alkaline
prepared with a Sorvall JB-4 microtome with a (above pH 8.5) in their liquid state and the dried
tungsten carbide blade. The inherent fluores- films are acidic (below pH a.\. (This seems to
cence of a material is minimized in the thin sec- disprove a widely held belief that the acrylic
tion, and the stain penetrates the entire sample. paints may be buffering the unprimed canvas
This minimizes the tendency for the stain to on color-field paintings.)
remain on the surface, where it can occupy
vacuoles and fissures and autofluoresce. As it became evident that more sophisticated
means of accurately identiffing these binders
Information gleaned from scientists at Kodak would be necessary, two additional methods
and Sigma, and literature on histological stains were chosen, Fourier Transform Infrared
and their application indicated that while many Analysis (FTIR) and Pyrolysis-Gas
of these stains do indeed have an affrnity for Chromatography (Py-GC).
certain materials or functional groups, they are
also used as tracers, for example, to locate Fourier Transform Infrared Analysis
microfrssures in metal swfaces and scratches FTIR was selected as an analytical tool based
on corneas. This suggests that there will be on current work in the fields of Coatings and
non-specific fluorescence, where the surface Forensics (Nielsen and Raaschou 1984; Reffner
is irregular, as is often the case at the interface 1990) and at the Getty Conservation Institute
between two layers, not necessarily due to any (Denick, Landry, and Snrlik l99l). It is an
affrnity with the material. There is a color differ- expedient and accessible technique for the
ence between the positive staining and the depo- identification of organic binders.
sition in vacuoles, but this is highly subject to
misinterpretation. Empirical observation indi- FTIR has several advantages over other specho-
cates that rinsing the sample in the carrier scopic techniques, such as increased energy
solvent minimizes this problem of non- output and spectral resolution (Low and Baer
specific fluorescence. Wetting the surface with 1977). The coupling of a light microscope with
petroleum benzine after staining also quenches the reflecting optics for FTIR microspectro-
the non-specific fluorescence and facilitates scopy allows relating the sample's morphologi-
viewing. cal structure with the chemical composition,
and analyzing intact layered structures.
Ultimately, it seems problematic to use these
stains to draw absolute conclusions without the Another advantage is the extremely small sam-
use of a spectrophotometer, which would mea- ple size necessary. Spectral libraries and com-
sure the wavelength versus intensity of the sam- puter software programs allow comparison and
ple before and after staining. This would allow subtraction of spectra, and information can be
quantification ofthe reaction and reduce subjec- readily shared and reproduced by other laborato-
tivity. Solubility tests on the cross-sections or ries with different instruments. This has been
thin sections can also help to determine the demonstrated in this study by correlating sam-
nature of the materials present. ples run on two separate instruments with
different operators.
pH indicators applied to thin sections proved
helpful in determining the properties of the Experimental
films. Three solutions were used to measure the Instrument 1.' Nicolet Nic-Plan FTIR spec-
liquid acrylic emulsions and the dried films: trometer coupled to a Spectra Tech IR-Plan ana-
lytical microscope at the Philadelphia Museum
methylred: pink pH4.2to6.2 yellow of Art. One group of samples was run neat on a
phenol red: yellow pH 6.8 to 8.2 red
single diamond plate (SDP), 200 scans, with a
bromothymol
resolution of 4 wavenumbers. An additional
blue: yellow pH 6.0 to 7.6 blue
group of samples was run as thin sections on a

418
NaCl window with the same 200 scans and a changes in the analytical parameters (especially
resolution of 4 wavenumbers. of the pyrolysis) can affect the results. For
maximum reproducibility and thus the possibil-
Instrument 2; Spectra Tech Irus plan FTIR ity for exchange of information between labora-
microprobe, at the Getty Conservation Institute. tories, experimental conditions must be
These samples were all run as thin sections on a identical. A recent article (Sonoda and Rioux
barium fluoride window. 200 scans at a resolu- 1990) about the use of Py-GC for identification
tion of 4 wavenumbers. of synthetic binders illustrates the problem of
reproducibility; we found we could not directly
Samples weighing approximately 50 pg compare our pyrograms with those published in
to 100 pg were taken from the paintings and the article.
mounted in Ward's bioplastic polyester embed-
ding medium. A 5 pm thin section area me€rsur- One advantage of pyrolysis is that it requires a
ing approximately 15 pm to 34 pm x 100 pm very small sample size and very little sample
preparation. For our dried pigmented paint
was analyzed. The same procedure was used
for the body of standards. Dispersed samples films, we found that a sample size of approxi-
were prepared by placing loose particles on the mately 200 pg produced the best results. We
plates. A number of them were run for compari- weighed the samples before and after pyrolysis
son to the thin sections. and found that, in general, 50% of the Magna
acrylic solution paints are pyrolyzed,40%o of
One disadvantage to this method is that it is not the acrylic emulsion paints, and25Yo of the oil
a separation technique and therefore little infor- paints. Due to the 50:l split injection, only 2Yo
mation can be obtained about the additives. of the pyrolyzate enters the column, resulting in
Complex mixtures can present problems in in- a final sample size that could be quite small.
terpretation ofthe spectra, and the influence of
pigments on spectral results is a field requiring There are paintings, though, whose smooth, un-
extensive investigation. broken surfaces or extremely thin paint layers,
as in color-field paintings, may not offer a sam-
Pyrolysis-Gas Chromatography pling site that can yield even a 50 pg or 100 pg
Py-GC is especially useful for the analysis of sample without damaging the surface. Depend-
non-volatile samples, such as pigmented acrylic ing on instrumental parameters, it may not be
emulsion paint films, which cannot be com- possible to obtain a large enough sample size
pletely solvent-extracted or dissolved. Pyrolyz- on such paintings. Because our Py-GC instru-
ing a sample thermally degrades it, breaking ment was concurrently used for other types of
down the large polymeric molecules into small sample materials, we were somewhat restricted
volatile molecules, which pass into the gas chro- in setting the instrumental parameters.
matograph. During the GC process, the sample
components are separated by molecular weight Experimental
and by polarity. Lower molecular weight and \trolysis Unit: CDS (Chemical Data Systems,
lower polarity components tend to pass through Inc., Oxford, PA) Model 120, pulse-mode sys-
the column first and thus result in lower tem. Pyroprobe: platinum wire of 0.36 mm
retention times. diameter, formed into a coil, 15 mm long by
3 mm diameter. Pyrolysis temperature can
As each component exits the column, it is de- range from ambient to 1400'C. Solid samples
tected and the response is recorded. The output, were placed tn a quartz tube, plugged at both
a pyrogram (pyrolysis chromatogram), is a plot ends with quartz wool, and inserted into the
of the retention time vs. degree of response. coil. Flash pyrolysis: l0 s at 650'C, although
the temperature reached inside the quartz
A pyrogram of a particular material is distinc- tube could be approximately 100'C to 150'C
tive and can be used as a fingerprint for that less.
sample. Fingerprinting, of course, requires a
reliable compound reference library. Any Gas Chromatograph Interfoce: CDS GC Inter-
face permits insertion of probe for pyrolysis, or

4t9
injection needle for direct gas chromatogaphy. formulations fu rther complicate the interpreta-
The interface is connected directly to column. tion. Extraction and separation techniques for
Interface maintenance heating temperature is these formulations are worthy of further study.
200"C to prevent condensation ofthe pyrolytic Nielsen and Raaschou 1984 discuss the use of a
products on the inner walls of the chamber. wick stick to separate the components of a
The pre-heated carrier gas, helium, flow rate: paint sample for FTIR analysis.
4ml-lmin, sweeps the pyrolysis products from
the interface chamber into the capillary column. Acrylic polymers show strong absorptions in
regions of the infrared spectrum related to es-
Gas Chromatograph Instrument: Perkin ters. The characteristic peaks are: a doublet
Elmer Model 8500 GC. Samples were intro- around 2900 cm-', indicating C-H stretches, a
duced into the column in the split injection sharp intense peak around 1730 cm-' due to
mode, ratio 50:l (49 parts of pyrolyzed carbonyl bonds, two sharp C-H (bends) peaks
sample is vented out and I part enters column). at 1450 cm' and 1385 cm-'. OH bends are
Quadrex Corp. SE-54 fused silica capillary col- distinguished by numerous broad peaks from
umn (50 m x 0.25 mm, 0.5 pm film thickness). I 024 cm-' to I 236 cm-'. which have similar

Column temperature (oven temperature) held at configurations for all acrylic resins. The peak
50oC for I min, then increased at 8oC per min at 1250 cm-' will be a doublet for methacrylate
until temperature reaches 325oC where it stays and a broad single band for acrylates. Compari-
for l0 min. Analysis time: 45 min. son of pure acrylate and methacrylate polymers
from spectral libraries did clearly illustrate
Gas Chromdograph Detector: Flame Ioniza- the prese4ce or absence ofthis doublet at
tion Detector (FID). 1250 cm-'. Copolymers, however, are more dif-
ficult to recognize or determine the monomeric
Results and Discussion of Analvtical components present. Ethyl methacrylate and
Techniques copolymers containing a methacrylate, such as
Acryloid B-72 (PMA/PEMA), show a sharp
The body of standards analyzed with FTIR peak around 1028 cm', which is not present in
and Py-GC illustrated a number of interesting spectra for acrylates (Lomax and Fisher 1990).
points, and allowed us to draw some conclu-
sions about the techniques and how they can The paints studied here are mixtures of these
be applied to the study of acrylic paints. Mono- polymers, and therefore are difficult to distin-
mers, polymers, copolymers, base emulsions, guish in this manner. A number of acrylic
and pigmented emulsions were compared to de- emulsion base resins were compared, such as
termine the degree to which they can be charac- Rhoplex AC-234, Rhoplex AC-33 and Rhoplex
terized. AC-507. The spectra are quite similar to one an-
other and would require a more complex inter-
FTIR: The characteristic spectra of acrylic pretation to elucidate specific differences in
polymers show specific absorptions in the infra- structure.
red region related to functional groups and mo-
lecular arrangement. For a given peak, intensity Although furgerprint comparison is very useful
of absorption is a measure of the concentration for broad characterization, the subtleties are dif-
of a functional group, which allows some quan- ficult to ascertain without a spectral library col-
titative estimation. lected with the same instrument under the same
conditions and sample preparation. Comparing
The infrared spectrum of a mixture is com- the emulsion Rhoplex AC-234 (PEAPMMA)
posed of the spectra of all components com- with Acryloid B-72 (PMA/"EMA) shows
bined. The functional groups for all of the virtually no difference in the fingerprint region,
monomer units used to form the acrylic formu- illustrating that different copolymers of similar
lations are similar, and therefore it is difficult to materials will absorb similarly.
distinguish which homopolyrners form the co-
polymer. Additives and pigments in the paint Proprietary paints were analyzed as standards
and further illustrated that complex mixtures of

420
 Table II
FTIR Absorotion Bands

Sample c-H C=O c-H c-H c-o GH

Held, top black layer 2982 1728 1448 1 383 1236 1 026 850
PMA/drsnersd 2953 1178
Held, black 2982 1736 1600 1448 1383 1 255 1 073 1 026 848
PMA"/thin sectron (dl 1 161

Held, black 351 8 2966 1 738 1 609 1453 1 386 1263 1162 1040 856
GC|/thrn sec{ion 2922 I.l\ 1073
Held, white 3520 z9b5 1729 1 600 1440 tJ/o 1246 1170 1 064 858
GC|/thrn sectron 3420 2941 (d) 1152 1022
Held, white 2982 1728 1448 1 383 1 236 1l(o 1 026 850
PMA /dispersed 2953
Held, whrte 3439 2984 1732 1 600 1448 1 383 1255 1176 1032 850
PMMhrn sectron 331 2 2951 (d)

On Kawara, black 2942 1732 1446 1 384 1234 1175 1 031 913
PMA,/drspersed 2952 1 159 879
11 15
On Kawara, black 2998 1728 1600 1452 1 381 1282 1 161 1068 875
PMA,/thin section 2939 (d)

On Kawara, black 3366 3027 1730 1 601 1498 1383 1283 1 161 1 070 340
GC|/thin seclion 2983 (d) 14s2 1124 1 034
2940
On Kawara, brown 3/.14 2982 1743 1 601 1493 1 381 1 265 1 167 1068 ltl
PMA,/thrn section 2918 /d\ 1448 1033
On Kawara, brown 3027 1732 1 601 1493 1383 1275 1 163 1072 877
GCI/thin secton 2985 (d) 1452 1 026
2936
On Kawara, gesso 3696 3026 1730 1 600 1452 1280 1161 1068 912
PMMhrn sectlon 3620 2998 (d) 1116 1032 877
2932
25'17
On Kawara, gesso 3697 3027 1 798 1 600 1452 1285 1175 1034 877
GC|/thrn sectron 3621 2983 1 730 (d) 1 161
2938 1117
251 8
Jo Baer, whrte 3450 2932 1730 1 602 1452 1400 1 280 '| 161 1 032 960
PMA/dicnarqed 2R6n 1 581 1116
Jo Baer, mock-up u47 2928 1741 1 595 1481 1380 1 260 1't72 1032 938
PMA"/disoersed 2856 1421 11 15
Jo Baer, whrte 3062 1732 1599 1493 1408 1 285 1 163 1072 810
GC|/thrn secton 3027 1452 I tzJ
2929
2857
Lrquitex acryhc gesso 3698 2983 1 730 1 593 1447 1 387 1260 '1161 1 058 477
GC|/thrn sectron JOZ I 2952 1117 1 030
2936
2514
Liquitex acryhc 2982 1724 1 562 1242 1176 1024 848
emulsion'l-io2 2950 10&
PMA,/disDersed
Golden acrylic 2959 1 736 1 468 1 383 1240 1'176 1022 850
emulsion TiO2 2876 1 448 1ofl
PMA,/disoersed
Daler Rowney acrylic 3/40 3027 1732 1 601 1493 1383 1273 1167 1034 863
emulsion TiOz 2956 (d) 1452 1 130

GCI/thin sectron 2932 1072

2889

PMA refers to all samples run with the Instrument at Phrladelphia Museum of Art and and GCI relerc to those samples run at
Getty Conservatron Institute. (d) refers to doublet, which was present at 1600 in all samples run as thrn sectrons.

421
 Table II (continued)
FTIR Absorption Bands

Sample c-H C=O c-H c-H c-o o-H

Lrquitex rron oxde black 3368 2976 1 738 1 600 1 500 1392 1278 1170 1 065 840
acry[c emulsron 2933 (d) 1455 1 130 1 034
GC|/thrn sectron 2901
1 948
Liqurtex rron oxide black no 1732 'tM8 1 381 1 236 1 161 1024 ttSu
acrylrc emulsion trans- 1 130
PMA/drsDersed mtssron
Golden carbon black no 2930 1734 1448 1 381 1160 1 030 840
acrylrc emulsion Irans- 2880 1097 800
PMA/drsoersed mrsston weaK
Daler Rowney rron oxide JJOO 2976 1 738 1 600 1500 1 386 '1278 1170 975 b3u
blac* acrylrc emulsron 2942 (d) 1454 1142
GC|/thin sectron 2889 1 068
1 948

Magna TiOz J4J 2959 1728 1 468 1 383 1271 1066 966 752
acrylic re$n PMA,/dispersed 2928 1242
PMA 2874 1176 '1022
2851 1 155
Magna ivory 32M 2959 201 3 1587 1412 1 269 1037 962 873
black acrylrc resrn 2928 1728 1468 1242 798
PMA/disoersed 2874 1147
Grumbacfier TiOz orl 3414 2918 1743 1587 1398 1242 1 099 978
PMA/drsoersed 2851 1541 1379 1167
2264 1464
Bellinr rron oxrde black orl 2924 1741 1460
PMA./dispersed 2850
weak
B|oplastic 3530 291 I 1740 1601 1493 1 381 1259 1157 1 070 744
embedding medrum 3449 2851 rdl 1120 1 039
Liqurtex gel medium JJOO 2978 '1738 600 450 390 250 1170 1068
1 | 1 1 850
GCl/thin section 2936 (d) 1030

PMA refers to all samples run with the Instrument at Phrladelphra Museum of Art and GCI refers to those samples run at
Getty Conservation lnstttute. (d) refers to doublet, whrch was present at 1600 in all samples run as thin secttons.

organic materials plus the pigments can obfus- at 1724 cr.rr-' in the Liquitex has shifted to
cate the interpretation of the spectra. Black and 1736 cm'' in the.Golden sample. The C-H
white paints from three manufacturers, Daler peaks (2850 ,rn-l to 2960 c#1) have a clear re-
Rowney, Golden and Liquitex, were compared solved shoulder in the Golden, corresponding
(Figure 7). to the base resin Rhoplex AC-507, while the
Liquitex configuration in this area looks more
The whites appear very similar upon initial like Rhoplex AC-234. Poly(methyl methacry-
inspection with clearly resolved characteristic late) and poly(butyl methacrylate) reference
peaks for acrylic polymers. When a more pre- spectra also have this clear shoulder peak.
cise comparison is made between the Liquitex
and Golden whites, which are clearly the same Daler Rowney white has the most distinctive
type of material, several differences are noted. spectrum, which was also true with the Py-GC
The ratios of the peaks are somewhat different. results. The carbonyl peak is at 1738 cm-r, the
The doubfet indicating the C-H peaks at highest wavenumber for this region of the three
2982 cm-' has shifted to 2959 cm-' in the formulations, and the C-H peaks are rqore
Golden sample, and the sharp carbonyl peak complex and have shifted to 2932 cm-' and

422
2889 cm-1. (This complexity may be in part The black colors proved more troublesome due
due to the bioplastic embedding medium, as to the presence of pigments, such as iron oxide
this sample was only run as a thin section.) and carbon, which have very strong absorption
Pigment interference seems to be problematic patterns above 2000 cm-' . and bone black. ,
with FTIR analysis and can influence results. which has a sfong peak between 1000 cm-'
Titanium dioxide has a strong absorption below and 1100 cm-'. The interference from the
800 crr'. However, in the case of the acrylics, pigment seemed to present more of a problem
this has little influence on the primary peaks, with the dispersed samples on the diamond
which characterize the binder in the fingerprint plate, as it was difficult to get the sample thin
resion. enough for maximum transmission. The elastic-
ity of the acrylics cause them to spring back
when pressure is put on to smear them on
the cell.

The combination of sample density and

pigment obstruction resulted !n inadequate
transrnission above 2000 cm-' for these
samples. A number of samples were run on dif-
ferent instruments using a reflectance mode,
which allows analysis of a thick sample or
cross-section directly from the surface with no
sample preparation. Intensity of reflected radia-
tion is dependent on the surface composition
and morphology. In all cases the resolution was
significantly poorer than with transmission
through the sample.

Despite the pigment interference in the black

formulation, several peaks were unquestionably
similar in all formulations, such as the sharp
carbonyl at 1134 cm-', t\e sharp peaks around
I
1446 cm' and 1380 cm-' indicating C-H
bends, and the broad area ofpeaks from
I I 50 cm-' to I 030 cm-' indicating C-O bonds.

Fewer distinctions could be made between

b) Liquitex TiO2 the black formulations than was possible for
the whites due to poor resolution. The Daler
Rowney iron oxide black C.-H peaks appear
at2942 cm-' and 2889 cm-' while the
Liouitex and Golden seem to be a bit lower
-t,2g26cm-1,
at;%0 2850
and The
cm-1.
ff], "
Daler Royney also has a strong peak at around
i/; \lh I170 cm', while both Liquitex and Qolden
llgl;1 show a comparable peak at I150 cm-'.

l.y
$! The samples run as thin sections on barium
fluoride windows produce clearer spectra, but it
appears that the bioplastic embedding medium
c) Daler Rown"y cryiirTio, interferes with the results. The polyester embed-
ding medium seems to penetrate the sample and
the infrared absorption from this material adds
Figure 7 Spectrafor acrylic emulsions. to the spectra. The infrared absorption of

423
polyester is quite similar to that of acrylic poly-
mers, but it is possible to isolate the peaks of
the embedding medium by comparing the same
sample of acrylic run as dispersed specimen.
All samples run as thin sections show absorp-
tion peaks due to the polyester embedding me-
dium (Figure 8). Subtracting the bioplastic
from the spectrum confirmed this with a drop
in intensity of the corresponding peaks. By F

comparison with reference spectra for bioplas- ba rrq (n htbr @t'tqhtd

tic and for identical samples run as dispersed

films, peaks from the bioplastic can be identi- {AVLNUMBFR'

fied: these appear as several weak but sharp Figure 9 Spectrum for Magna TiOz white.
peaks above 3000 cm-', a weak doublet at
i600.--1, and a sharp peak around 1070 cm-1. lower than any of the acrylic polymers. The
carbonyl peak for oil is,found around
1740 cm-' to 1750 cm-', while for tJre.acrylic
polymers.it is found arormd 1720 cm' to
1732 cm'' . There are sharp peaks at 1464 cm-'
and 1379 cm-', which is the aliphatic hydrocar-
:. bon region, while the acrylic poly,mers show
peaks at 1448 cm-' and 1381 cm-'. The C-O ab-
iorption lands of the oilq appear at 1242 cm'\ ,
I I 65 cm- ' and 1099 cm- ' . This area of the
spectrum is similar in both oil and acrylics, al-
though the ratios are differer.rt and the acrylic
polymers lack the 1099 cm-' peak. The most
intense pquk ir this area for the oils appears at
Figure 8 Sp", *o^iol"Wrd's bi oplast ic po b'est er I165 cm-' and the major peak in the,grouping
embedding medium. for the acrylic polymers is I159 cm-'. Although
the black oil shows the aforementioned pig-
ment interference, the fingerprint region for oil
Comparison of the acrylic emulsion paints with
has a distinctly different profile than the acrylic
Magna, a poly(n-butyl methacrylate) solution polymer.
paint, showqd an identical sharp carbonyl peak,
at 1728 cm', the sharp C-H peaks at 1468 cm-'
and 1383 crL'. The spectrum for the white
Magna paint was a bit more resolved than for
the black Magna, which is consistent with
the results for the acrylic emulsion paints,
especially in the broad area of peaks from
I170 cm-' to 1030 cm-'. The C-H peaks were
identical in both black and white formulations
(2959 cm-1, 2932 cm-r,2874 cm-t ,2851 cml)
and verv clearlv resolved. The doublet at
1242 cm-l and 1269 i, sharply defined as
"--1
expected for a poly methacrylate (Figure 9).

All of the acrylic paints can be unequivocally rigu;"" lo sp"rt**'ffiErumbacher Tio: oil.
distinguished from pigmented oil films (Figure
l0). The primary differences are: the C-H
peak area,is shifted to around 2920 cm-' and. Py-GC: Haken and McKay 197 3 have studied
1850.tr.l, approximately 30 cm-l to 60 cm-l pyrolysis of various acrylic homopolymers,
copolymers, and mixtures of homopolymers,

424
 Table III
Py-GC Retention Times

RETENTION TIMES (MINUTES)

MONOMERS
MA c. l

EA 59
MMA o
EMA 7.1

nBMA 10.9
EA-MMA mrxture 5.8 6

HOMOPOLYMERS
MA 4.5 47 J.Z 11.5 14.6
16
24.7 30.5 30.7
EA 4.3 4.6 b
17.4 18.5
26.9 32.8 329 378
MMA 5.9 6.1 192
EMA 3.9 4.3 o 7.3
rBMA 10.8 11 .3 to.z
nBMA 4.5 10.8 11

COPOLYMER
MA.EMA 43 c,1 7.3
15.9 16.7 t/.o
24.4 25.1 25.5 zo.J 27.'l

BASE ACRYLIC EMULSIONS

AC-22 4.2 45 qt o o. l 7.2 9.2
16.7 17.4 184
25.7 zol 26.8 27.1
AC-234 4.2 4.5 5.2 f,. 1 9.3
16.8 17.4 18.5
258 26.2 26.9 27.2
AC-507 4.4 5.7 o.< 7.6 93
20.3 209 23.9 24.7
303 307 31.9 335

ACRYLIC EMULSION PAINTS

LIQUITEX WHITE 42 4.5 5.2 o o. l 7.2 10.7
16.8 t/ c 18.5
25.8 26.2 26.8 24.2 28.5
GOLDEN WHITE 4.2 4.4 qt 5.6 o 6.1 7.2 9.3 11.2
16.8 17.6 18.5
25.9 26.2 26.9 273
CRYLA WHITE 4.4 62 7A v.J
20.4 21 24.1 24.8
30.4 30.7 32 33.5
LASCAUX WHITE 4.5 6.2 9.3 11.1
20.3 20.9 22.7 24 24.4 24.4
305 308 32 33.5

425
 Table III (continued)
Py-GC Retention Times

RETENTION TIMES (MINUTES)

ACRYLIC EMULSION PAINTS
LIQUITEX BLACK 4.4 +.c 5.2 6 62 10.7
16.8 17.6 18.5
25.9 26 26.9 1L5 283 28.5
GOLDEN BLACK 4.2 4.3 4.5 5.2 5.6 6 o. l 7.2 9.2 11.1
16.8 17.6 18.5
zJd zo.a 26.9 27.2
CRYLA BLACK 44 57 6.2 7.8
zu-5 24.2 25
304 30.7 32 33.5
LASCAUX BLACK 4.5 qq 75 9.4 11.2
19.7 21 2 227 24.2 26.6 26.a
31.5 34.6

OIL PAINTS
GRUMBACHER WHITE 4.4 4.6 5.2 o 7.5 Y.O

11.7 12.3 12.9 IJ.I 14.7 15.7 17.6 19.3

22.7 24.4 25.4 26.3 26.7 28.6
31 31.2 32.1 33.2
BELLINI BLACK 4.5 47 ol 7.7 oe
11 .7 12.4 13.8 14.9 15.9 17.4 19.4
21.4 11.4 24.3 28.8
31.2 31.4 JZ.J u
ACRYLIC SOLUTION PAINTS
MAGNA WHITE 4.5 11.3
MAGNA BLACK 4.6 1 1.5

PAINTINGS
HELD WHITE 4.3 4.5 5.J o o.z 7.2
168 17.5 18.5
258 zo.z 26.9 27.2
HELD BLACK 4.3 4.5 6Q o o.z 7.2
16.8 176 18.5
25.8 262 zo.Y 272 28.4 30.8 31
KAWARA BLACK 4.2 4.5 5.3 59 6.1
16.9 17.7 18.8
25.9 26.2 26.9 27.3
BAER WHITE 4.4 4.6 5.8 7.4
11 l l.4 121 12.9 13.3 14.8 15.5 '17.4 19.1
21 22.7 zJ.o 27.4 28.4 28.8
308 31 38.6
BAER MOCK.UP 43 4.5 5 56 0z 7.4 9.5
11 11.4 12.1 137 14.5 15.8 175 19.3
21 245 28.4
30.9 a2 I
AGED LINSEED OIL 4.4 4.4 5.3 6.2 10
11.8 12.6 I J.3 14.9 18.6
20.2 22 1 23.9 25 z5-b 28.8 29.7
32.2 55.2

426
including the methyl, ethyl, and n-butyl esters and cannot be pyrolyzed, about I pl ofa solu-
of acrylic and methacrylic acids. They found tion of 6 drops of monomer in about 2 mL of
that (regardless ofthe alkyl group attached to methylene chloride was injected directly into
the acrylate or methacrylate) there is a constant the port. Consequently each monomer chroma-
high recovery rate of methacrylate monomer togram has a peak at 4.5 or 4.6 min retention
from the methacrylate polymer whether it is a time representing the methylene chloride
homopolymer, copolymer, or mixture. For the solvent.)
homologous acrylate polymer, however, the re-
covery rate of the acrylate monomer is much The pyrograms for the poly methacrylate ho-
lower and is dependent on whether the polymer mopolymers, PMMA, PEMA, PiBMA, and
is a homopolymer, copolymer, or mixfure. The PBMA, all display a tall distinct singular
pyrolysis of an acrylate/methacrylate copoly- peak at corresponding retention times for their
mer gives greater yield of the acrylate mono- parent monomers, indicating a high yield of
mer than does the pyrolysis of either an monomer recovery. The acrylates also display
acrylate homopolymer or a mixture of acrylate corresponding monomer peaks, but with much
homopolyrner and methacrylate homopolymer. lower relative yield, and with the presence of
For example, yield of the methyl acrylate mono- numerous other peaks, probably alcohols, di-
mer from the homopolymer PMA was l4o/o, mers, and trimers. The PMA/PEMA copolymer
and from a mixture of homopolymers PMA and pyrogam displays a high methacrylate yield
PMMA itwas l4Yo, but from the copolyrner and an increased acrylaLe monomer yield as
PMA/PMMA itwas 40%o. described by Haken and McKay.

Besides the monomer, the only pyrolysis prod- Although much information is available about
ucts for methacrylates were colresponding alco- methods for quantitative Py-GC, we made no at-
hols at very low concentrations. Acrylates gave tempt to determine precise quantities and ratios.
a low monomer yield with a high yield of Instead, where possible, we estimated the ap-
alcohols. proximate relative quantities of various types of
monomers recovered, or identified a material as
It is these differences in monomer recovery that an acrylate or methacrylate.
allow differentiation between copolymers and
homopolyrner mixtures using pyrograms. In ad- The Rohm and Haas base acrylic emulsions
dition to monomers, other pyrolysis products of Rhoplex AC-22, Rhoplex AC-33, Rhoplex
PMMA/PMA copolymer could be methane, eth- AC-234, and Rhoplex AC-507 were analyzed
ylene, hydrogen, carbon dioxide, carbon mon- (Figure 1l). These emulsions are based on
oxide, ethyl methacrylate, and ethanol. At EA-MMA copolymer (about 66:34 ratio). Al-
higher temperatures, the decomposition of the though these separate emulsion products may
ester groups of polyacrylates results in other be based on the same copolymer, they differ
products, such as olefins and alcohols. Recom- from each other in viscosity, for example,
bination reactions of degradation products dur- Rhoplex AC-22 50 Pa.s, and Rhoplex AC-33
ing pyrolysis can also occur. In pyrolysis, PMA 600 Pa.s. In fact, the pyrograms for Rhoplex
degrades to methyl alcohol, ethyl alcohol, MA AC-234 and Rhoplex AC-22 are virtually iden-
monomer, and MMA monomer. PEA degrades tical. The pyrogram for Rhoplex AC-507 is
to ethyl alcohol, EA monomer, and EMA mono- similar, but has fewer peaks and different reten-
mer. PMMA, PEMA, and PnBMA degrade to tion times in the early retention time area (0 to
mostly monomer. l0 min); and the dimer and trimer groups shift
to longer retention times compared to Rhoplex
The results of the Py-GC analysis of acrylate AC-22 and Rhoplex AC-234. For Rhoplex
and methacrylate monomers, homopolymers, AC-22 and Rhoplex AC-234 the dimer groups
and copolymers in this study agree with the re- fall in the 16 to 19 min range and the trimers
sults of Haken and McKay. Chromatograms for between 25 and27 nfu. But, for Rhoplex
the monomers MA, EA, MMA, EMA, nBMA, AC-507, the dimer group is between 20 to
and EA-MMA mixtures were obtained. (Be- 25 min and the trimer between 30 and 34 min.
cause monomers are in a volatile liquid form,

427
Peaks for EA monomer (5.9/6.0 min) and 9.3 min, indicating a difference in Rhoplex
for MMA monomer (6.1 min) appear in the AC-507 from Rhoplex AC-22 and Rhoplex
pyrograms for Rhoplex AC-22 and Rhoplex AC-234.
AC-234, as would be expected. The Rhoplex
AC-507 is suggested to be based on Eight proprietary acrylic emulsion paints,
MA-nBMA copolymer (Rohm and Haas 1991). four black and four white, were analyzed
However, the peak for nBMA monomer (Figure 12). Within each brand, except for
(11.0 min) does not appear in the pyrogram the Lascaux, the pyrogram for the white paint
evidence that is contradictory to information
- matches the pyrogram for the black paint, im-
given by the manufacturer. Also, note the plying consistent use of the same base formula-
presence ofan additional distinctive peak at tion for the two colors. For the Lascaux black
and white, the peaks in the monomer area coin-
cide, but there are variations in the dimer and
trimer areas. The Lascaux white has additional
peaks. Between brands, however, differences
do exist in the pyrograms. In general, the Daler
Rowney and the Lascaux are quite similar, ex-
cept that, in the monomer area of the pyrogram,
the Lascaux has an exfra small peak at
1l .l lll .2 min. The Liquitex and the Golden re-
semble each other except that, in the monomer
area, the Golden has an extra peak at 9.3 min.
In the dimer area, both Daler Rowney and Las-
caux have peaks in the low 20s, except that
Lascaux white has two extra peaks at 22.7 and
24.4 min. Lascaux black has a slightly different
configuration. Liquitex and Golden both have
peaks in the mid-teens in the dimer area. In the
trimer area, Daler Rowney and Lascaux are
both in the low 30s. Liquitex and Golden are
both in the high 20s, except that Liquitex has
a) Rhoplex AC-234 two extra peaks at 28.2 and 28.5 min. When
the Rohm and Haas base emulsions are com-
s pared to the paints, Rhoplex AC-22 and
I

I
I

tr
I

l;
ll
|l
tltl
';llll
:
rllll I
I

l fl"ll dR .i
i r, S, !r-*
{
"irii,
': lljllt
It,lll. r
I
ll . ll^ Ll* - -I LL
3g

i!:
b) Rhoplex AC-507

Figure 1 I Pyrogramsfor base acrvlic emulsions. c) Rhoplex AC-22

428
Rhoplex AC-234 are almost identical to Gold- expected there is a strong peak at I 1.0 min, the
en. Liquitex is also quite close except that it retention time for nBMA homopolymer.
lacks the peak at 9.3 min and has two extra
small peaks in the trimer area. Rhoplex AC-507 Compared to acrylic emulsions, traditional oil
matches the Daler Rowney. It is quite close to paints (Figure 14) produce identifiable pyro-
the Lascaux, except for two extra peaks in the grams with a greater number of peaks, greater
paint at 24.0 and 24.8 min. complexity, and a repetitive regularity in the
profile.
The medium for Magna acrylic solution paints
(Figure 13) is poly(n-butyl methacrylate) resin. In general, different paint media (acrylic
Magna paints produce quite simple pyrograms, emulsions, acrylic solutions, and drying oils)
that of a methacrylate homopolyrner. As produce pyrogams with easily recognized fin-
gerprint profiles. More specifically, different
N
acrylic and methacrylic homopolymers and
copolymers can be discerned.

As an addendum, we found that overall, the pig-

ments do not seem to affect the Py-GC results.
Weighing of the sample seemed to indicate that
the pigment was still in the quartz tube after py-
rolysis (40% of the sample by weight). Accord-
ing to the literature metal pigments can affect
retention time. In our results, however, the pig-
ment seems to have a more dramatic effect with
FTIR than with the Py-GC.

Case Study I
Flemish WI (1973) (Figure l5) by Al Held was
chosen as a representative example of the use
of acrylic emulsion. The emulsion was applied
in numerous layers to create a hard edge compo-
sition of intersecting planes, constructing an

b) Binnqv and Smith Liquitex

Figure 12 Pyrograms for acrylic emulsion titanium

white paints.

429
Figure l3 Pyrogramfor Bocour Magna acrylic Figure l5 Al Held, Flemish YII (1973), acrylic
solution titanium white paint. emulsion paint on canvas, 60 l/8 in. x 60 I/8 in.
(152.7 cm x 152.7 cm), Collection, The Museum oJ'
illusion ofdepth and architectural clarity Modern Art, New York, Riklis Collection of McCrory
(Lodge 1 99 1). Held began using acrylic paints Corp or ati on (frac ti ona I gift).

in the late 1950s, slowly evolving a meticulous

achieved the composition he desired and then
technique of applying multiple layers, creating
used a disc sander to remove the "process" as
a masonry effect on the surface. The cornposi-
much as possible. Marks from the sander are
tion is worked out directly on the surface and
pentimenti or compositional changes are visible often evident through the surface. His very re-
cent paintings are conceived in a full scale
in raking light (Figure 16). Prior to 1963 Held
study and then entirely repainted on another
allowed the effects of this process to remain
canvas to achieve a pristine surface without
visible, but eventually the critics influenced his
working methods by referring to traces of the unintended tape edges showing through. Held
process as "scar tissue," indicating numerous himself blocks in the composition with a base
color and then numerous layers of the color are
failures or uncertainties (Sandler 1984). At this
applied by his assistants (Held 1991).
time he began to build up the layers until he

Figure l4 Pyrogram for Grumbacher titanium white Figure l6 Flemish Ytl, detail of surface showing
oil paint. changes in composition.

430
Flemish WIhas a very high gloss, an unusual Results of FTIR microspectroscopy for the sam-
characteristic for the black acrylics, which ples from the Al Held painting confirmed the
often contain fillers and impurities causing suspected acrylic binder in both the black and
them to dry matte (Golden 1990). the white (Figure 17). The general peak pattem
seems to be reproducible and consistent with
The artist recalls that in some of his earlier the standard of Liquitex gel medium, Rhoplex
paintings he used a50/50 mixture of Liquitex AC-234, and Acryloid B-l2.The Held spectra
and Rhoplex, for example, in The Big N (1964, seem to be closer to the Liquitex gel medium
MoMA) and in his fnst acrylic painting, Taxi- than to the Liquitex black iron oxide, which
cab (1959). He recollects that he probably may be due to the Held pigment, which is bone
mixed gel medium in the later painting or ivory black. Liquitex gel medium was found
Flemish WI,but in smaller proportions. The by the Canadian Conservation Institute to be
earlier paintings seem to hold onto more grime 1:l PnBA.1PMMA.
and particulate matter from the atmosphere.
This is probably due to the large proportion of The results of the samples run neat on a single
Rhoplex, which could impart a softer, tackier diamond plate also correspond very closely
surface. to Rhoplex AC-234 (PEA/PMMA). The Held
sample ip slightly moqe complex in the
A test showed that black and white paints were 900 cm-' to 1075 cm-' region of the infrared
both insoluble in water and hydrocarbons, but spectrum, which could be caused by pigment
readily soluble in xylene, toluene, acetone, and interference. All of the sa.mples show 4 very
ethanol. weak peak at approximatbly 3441 cm-', indicat-
ing an aromatic (possibly a styrene component)
Ultraviolet examination revealed the non- or acids from unreacted monomer units.
fluorescent nature ofthe medium and the char-
acteristic purple appearance of the white. This
type of fluorescence was corroborated in the
cross-section that was taken from the edge of
the painting where there are numerous layers
and build-up.

Results of staining were as follows:

. Stained for lipids with Rhodamine B in

ethanol black uppermost layer faintly posi-
tive and-the white layer strongly positive.
Both layers are soluble in the carrier solvent. IAV€NUXB€FS

Same procedure with Rhodamine B in water

a) white paint, dispersed
black uppermost layer negative, white
-layer faintly positive.
. Stained for protein with FITC in acetone
both layers readily soluble in the carrier -
solvent. Black layer spotted positive fluores-
cence, white layer strongly positive. Same
procedure with FITC in pH 9 aqueous sodium
phosphate solution black layer is negative
-
and the white layer faintly positive.

o Stain for pH with methyl red and

bromothymol blue indicated a pH below
4.2 in all layers. - b) black paint, dispersed
Figure l7 FTIR spectrafor Flemish VII.

431
Styrene is commonly used in more recent selects a pre-shetched canvas for his daily
acrylic formulations to adjust their glass transi- work, a carefully painted monochrome on
tion temperatures and to improve gloss. which is noted the day's date. If the painting
is not finished by midnight, it is destroyed.
The Held white sample, pigmented with tita- By October 31,1970 eight hundred and
twenty-three of these works were completed
nium dioxide, closely matched both the Held
(Denizot 1979).
black and the Liquitex gel medium. There is
a very strong absorption overlay from the tita-
The second work chosen for examination was
nium below 700 cm-'. Absorption bands match
closely for the black and white samples liom
Dec.I8,I979 (Figurel9).It is one of thethree
this painting, except the white has an intense
doublet atSll/834 cm-'. This peak also ap-
pears in some unpigmented emulsion formulas,
such as Rhoplex AC-234, Liquitex gel medium,
and Rhoplex AC-33.

When compared with white pigmented

standards, the fingerprint match for the Held
white sample proved to be Liquitex titanium
white acrylic emulsion.

Samples of the white paint and of the black

paint were each run with Py-GC. The pyrogram
for the white matches that of the black and
they both match the pyrograms for Liquitex
(Figure l8). This again confirms information
from the artist that he painted Flemish WI
with Liquitex.

Retention times of painting and standard sam-

ples were identical, the one difference being
the ratios between the monomer peaks of 4.5 a) white paint
and6.2 min. This may reflect some change in
ratios of the homopolymers in the formulation
or in the degradation products of the natwally
aged painting sample. (The standard is not
aged.) The pigments do not appear to affect the
results, and weighing of the sample seemed to
indicate that the pigment remains in the quartz
tube after pyrolysis (40% of the sample by
weight).

Our effort to distinguish between the various

formulations was unequivocally more success-
tul with Py-GC than with FTIR.

Case Study II
The moming of January 4, 1966 in his East
l3th Sheet New York studio, 39-year-old
Japanese-bom On Kawara began The Today
Series of Date Paintings. After stamping the
exact time he woke up on color postcards b) black paint
which he sends to his friends, On Kawara Figure I 8 Pyrograms forFlemish VII.

432
 rabbit skin glue, which is quite reactive and sub-
ject to embrittlement. The artist recalled that he
used rabbit skin glue to prepare the cotton
duck, then applied acoat of commercial acrylic
gesso and then acrylic paint. He said that he
purposely manipulated the paint to achieve a
matte surface that should never be varnished.
The paint is insoluble in water and petroleum
spirits, and readily soluble in xylene, toluene,
acetone, and ethanol.

Under UV light, the dark area is non-fluores-

cent and exhibits the aforementioned violet re-
Figure I9 On Kawara, Dec.18,1979 (1979), acrylic flection or appearance. The white text has a
emulsion paint on canvas, I8 I/4 in. x 24 3/8 in. purplish cast, and a slight whitish inherent
(46.3 cm x 61.6 cm), Collection, The Museum of fluorescence, probably due to pigment
Modern Art, New York, Blanchette Rockefeller
composition.
Fund.

In a cross-section, the presence of three layers

Date Paintings in the collection of MoMA be- explained the slightly greenish color of the
longing to the Today Series by conceptual artist surface, which at first glance appeared black.
On Kawara,who manipulates acrylic paints in Above the ground layer is a brown layer and
a quite different manner from Held. He is then a black or grayish layer. Analysis of the
deeply concerned with the notion of time and pigments revealed a complex mixture of cal-
divisions into units such as hours and days and cite, yellow and orange oxides, carbon black,
their relation to the earth's rotation on its axis. and small amounts of an isotropic blue-green
Each painting is housed within boxes con- pigment.
structed by the artist, many of which are lined
with a newspaper sheet from the specific date The thin section clearly shows the porous
of the painting. nature of the ground and the particle distribu-
tion in the upper layers. Under UV light, the
The artist's meticulous painting technique gesso layer appeared faintly fluorescent with
consistently results in a smooth , matte, flawless characteristic violet overtones, the upper two
surface (Figure 20). This surface could be layers being non-fluorescent.
mistaken for an oil due to its even quality and
no apparent air bubbles. Air bubbles or pin- Results of staining were as follows:
holes are often an indication of the frothing of
an emulsion. In fact, the artist has stated that he o Stained for lipids with Rhodamine B in
carefully wipes out air bubbles with a cloth, in ethanol all layers stained positive. All
order to eliminate this phenomenon (On -
Kawara 1989). However, when viewed under
stereomagnification, these charucteristic bub-
bles become more evident and the rounded
edges of the brushwork appear more like
acrylic, which flows and levels readily. The
painting does not have a particularly plastic ap-
pearance, although the edges of the painting,
which are tautly stretched and painted around
the side, show no sign of stress-cracking. This
would be a difficult feat if using an oil film,
which is inherently much more brittle. The
overall absence of cracking is also particularly
interesting considering the priming layer is Figure 2 0 Detail of surface Dec.l 8, I g7g .

433
 layers were readily soluble in carrier solvent, Held results were similar. Reference samples
and the gesso layer was immediately broken that contained pure iron oxides were run as
up. Using the same procedure with Rho- dispersed samples and often suffered from
damine B in water faint positive in blacks, pigment interference and the thickness ofthe
brightly fluorescent- in gesso. sample. There was a clear gain in energy trans-
mission for thin section samples, but they have
. Stained for protein with FITC in acetone
entire sample brightly positive. All layers-sol-
uble in carrier solvent. Using the same proce-
dure in a pH 9 sodium phosphate aqueous
solution only the gesso layer is positive
-
and brightly positive in area of sizing (rabbit
skin glue).
q
o Stain for pH with methyl red and F'
t
bromothymol blue indicated a pH below !
4.2 in all layers. -
fAVENUMSEFS
The gesso reactions in this sample from the Ut
PHILA0ELPHIAjUST c rrmi s|g^^* ffiY
painting corresponded to the results of staining 9lJ!Ilolstsvarr0N_ . 0 rrE 01 ro !
G/07At

numerous acrylic gesso standards, which con- a) top blacUgray paint

sistently stain positive for both lipids and pro-
teins. This reaction may be because of the
solubility, or due to its complex mixture of
fillers and bulking agents. Surfactants with
hydrophilic goups (fatty acid chains) may
react positively for lipids, or proteinaceous
material, such as casein or gelatin, which are
sometimes added as thickeners. This strong
reaction is peculiar, however, since the percent-
ages of these materials would be very small.

FTIR analysis was done on the three layers of

HAVENUTBEFS
the On Kawara sample, which was run as a thin PHILAOELPNIA TLSEUM
section on a barium fluoride window. The top OF AR] CONSEWAIION
OEFABTftNT . O'IE O3/O6/9t F?t A

two layers produced nearly identical spectra b) middle brown paint

(Figure 21). The priming layer appeats quite
different due to the interference from pigments
and bulking agents in the gesso. This interfer-
ence was exemplified by comparing the On !I
Kawara gesso with a Liquitex gesso standard It
and reference spectra for kaolin and calcium
carbonate. Both the unknown and the.Liquitex
gesso standard have distinct absorbency pat-
terns for kaolin at 367613596 cm-' and for
calcium carbonate at2502 cmj', a broad unre-
solved band around 1446 cm-'to 1387 cm-',
which is also indicative of calcium carbonate NAVENU{8EFS

(Figure22). PHILAOELPHIA TUSIUX

9tJlllolsEfrar r0N
UEPABtrLNI IOTIf, €/6/9'6A3'

Both the black and the brown layers have ab- c) bottom gesso layer
sorption pattems similar to Daler Rowney and
Liquitex black acrylic emulsion standards; the Figure 2 I FTIR spectra for Dec. I 8,1979.

434
Figure 22 Liquitex acrylic white gesso.
ia
c*r Fif
"
the disadvantage of bioplastic interference. :e: r*
A comparison of the top layer run dispersed 9&P:
and as a thin section indicated that the small ----;---^,l"tr-
sharp peafs above 3000 cm-', the doublet at
|
I 600 cm- ', and the sharp peak at | 068 cm- Figure 23 Pyrogram for Dec.l8,l979 top
can be atfributed to the bioplastic embedding black/gray paint.
medium. Computer spectral subtraction and
deconvolution ofthe spectrum can aid in chosen as an example of a more complex prob-
interpretation. lem of identification (Figure 24). Lfte On
Kawara, the artist is also interested in'mathe-
Despite these obstacles, most spectra were matical systems and sequences. She writes of
quite well defined and could be easily matched this series:
to their corresponding standards. When the
spectrum for bioplastic was subtracted from the These paintings form part ofa series of
specfra of the thin section samples taken fiom twelve. There are four colors in the series
On Kawara paintings spectra, it was interesting blue, green, purple, yellow. There are also-
to note that one ofthe wgak but sharp peaks four sizes and shapes. ... The possibilities for
directly above 3000 cm-' remained, which, combination or grouping of the paintings are
the permutations of twelve (831,753,600) or
when compared to reference spectra of coatings
whatever set factor are chosen. The paintings
resins, seems only to appear when the acrylic
here are the three large squares and they use
co- or ter- polymers contain styrene.
the intense color bands. All the paintings are
color in a luminous mode, but this group
Using Py-GC, only the top black/gray paint also renders the primary colors of light: a
layer could be isolated and sampled. The pyro- red (magenta), a green, a blue. They are each
gram for the On Kawara black layer is identical constructed equivalent to one another as a
in fingerprint and retention times to Liquitex color presence (Alloway 1966).

nnn
(Figure 23). This was verified by the artist. The
different pigments have little or no effect on the
results.

One clear advantage of the FTIR method for

this painting was the ability to analyze three
discrete layers of the composite structure.

Figure 24 Jo Baer, Primary Light Group: Red,

Case Study III Green, Blue (1964 to 1965), oil and acrvlic on can-
vas, triptych, approximately 60 l/4 in x 60 l/4 m.
The triptych Primary Light Group: Red, Green, (153 cm x 153 cm) each, Collection, The Museum of
Blue painted in 1964 to 1965 by Jo Baer was Modern Art, New York, Phtlip Johnson Fund.

435
In this work, the spatial effect of figures in removed easily with water, in contrast to the
a field is avoided and a unrty of surface is acrylic emulsion paints.
created. To achieve this overall white field Jo
Baer used a mixture of materials that she Under UV light the surface has a slight white
claims were chosen to avoid yellowing in the fluorescence from the white pigment (found to
whites. These paintings use the theory of Mach be zrnc and titanium), yet it still has the violet
band (from the physicist Ernst Mach) to create overtones characteristic of the synthetic resins.
a retinal phenomenon where the viewer sees There seems to be an underlying layer of white
contrast enhancement, or luminous glow in cer- that fluoresces more brightly than the surface.
tain shapes. It is easy to imagine how discolora- A cross-section was taken of the white layers
tion or cracking could adversely affect the to funher elucidate the structure and stain for
overall impression. media.

According to MoMA files, the medium of the Results of staining were as follows:
painting is oil and synthetic polymer. It was not
clear upon examination if this was correct, or . Stained for lipids with Rhodamine B in
which areas of the painting were oil and which ethanol an unusual pattern of globules
were acrylic. A questionnaire was sent to the - in the white layer. Positive stain
was outlined
artist, the painting was examined, and samples was observed for oil in the matrix, but in the
were taken to determine the nature of the mate- globules, negative. Paint was not readily sol-
rials. According to the artist, the medium used uble in the carrier solvent. Rhodamine B in
was a mixture of titanium white oil paint and water produced the same result. (This uneven
Lucite 44. Lucite 4, spoly(n-butyl methacry- staining of an oil matrix with non-staining
late) resin, was added by the artist to decrease globules was also noticed in a cross-section
the yellowing of the drying oil. from another painting by the same artist.)
When the painting is viewed in a strong light, o Stained for protein with FITC in acetone and
differential areas of matte and gloss are visible water negative in all layers.
across the white field. The overall qualrty of the -
paint is glossy and hard, not plastic like the o Stain for pH with methyl red and
acrylic paints (Figure 25). bromothymol blue indicated a pH below
4.2 tn all layers. -
Fortunately in this case we had the information
from the artist about the materials she used. In
response to the questionnaire inquiring about
her technique, she wrote, "I only used an
acrylic resin (Lucite 44, oIl-compatible) in
white paintings to insure their non-yellowing ...
and only in the white parts (normal varnish
[i.e., damar]/stand oil/turps in blacks and
colors):
1 large tube Winsor &Newton Titanium White
Figure 25 Detail of surface, Pnmary Light Group: 3 oz. Lucite 44 (dissolved in xylene and
Red, Green, Blue. benzine)
3 oz.ber:z,rne" (Baer 1989).
The solubility of the layer was much different
than the acrylics. It was insoluble in water, From this formula we were able to prep are a
hydrocarbons, xylene, and toluene. It required mock-up of her paint formulation to run as a
more polar solvents, such as acetone and etha- standard in FTIR and Py-GC. This was com-
nol, to dissolve it. The surface was quite soiled, posed of 4 parts Winsor & Newton Artist Oil
but the particulate matter was not adhered to Color Titanium White, 3 parts DuPont Elvacite
a tacky surface or embedded. It could be 2044 resin (PnBMA) as we were unable to

436
 was compared to that of the painting sample,
they matched quite closely.

It thus became possible to locate specific peaks

that corresponded to both oil standards and
standards of Magna and Elvacite, which are
both PnBMA resins.

A sample of white paint from the painting

and from the prepared standard were run with
IAVENUNBEAS
Py-GC (Figure 27).The overall profile of the
PHILAOELPHIAJUS€UM M MsffHI N pyrogram of the Baer white paint sample
OF AFT CONSEAVATION IXre M ffi'ld
DEtrABTRNT IOUf, €/d/9TFC6 matches that of an oil paint, with one important
Figure 26 FTIR spectrafor Primary Light Group:
Red" Green, Bfue, overlay of spectrafrom painting
andfrom standard.

obtain Lucite 44 (roughly about40Yo concentra-

tion dissolved in xylene4rydrocarbon 50/50),
and 3 parts hydrocarbon. This mock-up formu-
lation proved extremely helpful when attempt-
ing to interpret the spectra and pyrograms
generated.

In fact, the FTIR spectra generated by the :

unknown sample from the painting would
have been difficult to interpret without the
comparison (Figure 26).lnthe 2900 cm-'re-
gion, the absorption bands had shifted down, in-
dicating an oil, yet the carbonyl group matched
quite closely to the acrylic polymer spectra.
Upon initial examination it appeared to have
b) white paintfrom standard ofBaer's technique
characteristics corresponding to poly(vinyl ace-
tate). When the spectrum of the mock-up paint

k!lE' ird @ M, (bn,

uE (196{-65)

a) white paintfrom the painting c) aged linseed oil

Figure 27 Pyrogramsfor Primary Light Group: Red, Green, Blue.

437
difference. It contains a distinct peak at I 1.0 Py-GC samples were rul with the assistance of
min not found in the pyrograms of oil paints. It Dr. Alexander Shedrinsky at the Conservation
is found, however, in the chromatograms for Center, Institute of Fine Arts, New York
nBMA monomer, nBMA homopolymer, University. IR samples were run by Beth Price,
Magna white and black, and the Baer mock-up, Conservation Scientist at the Philadelphia
thus demonstrating that the pyrogram of the Museum of Art and Michele Derrick at the
Baer painting sample contains peaks for both Getty Conservation Institute.
the oil and the Lucite.
R6sum6
This case study shows the potential that exists
for misinterpretation of scientific data collected L'identijication et la caractdrisation des
from works of art, especially when materials p eintu re s - imu I si on s acryliqu es
are complex and/or have been manipulated by
the artist in an unexpected or unusual way. Un programme d'6tude continu, por'tant sur les
Without such reliable precise information from propridtds des peintures-tmulsiow acryliques, est
Baer about the paint formulation, the pyrogram en cours au Museum of Modern Art de Nau York.
On aprocddd jusqu'd maintenant d I'analyse du
for her white paint might well have been
midium de quelque I50 tableaux de la collection
misread.
du musde, et leur Atat gAniral a iti halui. Les
examens et tests usuels ont 6ti ffictuts en labora-
Conclusion toire, et toute une sdie de normes a eft etublie d
desfira de comparaison. Trois cas ontfait l'objet
Much investigation remains to be done on the d'une analyse plus pousste, qui a permis d'ex-
identification as well as on the practical meth- plorer davantage un certain nombre de mithodes
ods of conservation for acrylic emulsion paint plus pricises d'identification
films. The information obtained by visual ex- - des examens au
microscope ou analyses par coloration locale
amination, UV examination, and fluorescence jusqu'aux techniques de la spectrometie in-
microscopy seems to be useful, albeit limited. frarouge d transformee de Fouier (IRTF) ou
For more specific analysis, FTIR and Py-GC de la p1'rochromatographie en phase gezeuse,
are both effective. Separation and precise notamment.
identification of small components within
the formulations would be a valuable area of References
research.
Alloway, Lawrence, Sys temic Painting,
On a more general note, the results of the initial Exhibition catalogue (New York:
survey and examination of the body of acrylic Solomon R. Guggenheim Foundation, 1966).
paintings in the collection of MoMA have
shown that, overall, they are in quite good con- Baer, Jo, personal communication, 1989.
dition, unless they have been mishandled or
subjected to extreme environments. Of course, Becher, Pau| Encyclopedia of Emulsion
they have only been in existence for less than Technologt, I/o/. 1 (New York Marcel Dekker,
half a century, and we are unable to observe 1983).
their long-term natural degradation. Future
phases of the research project at MoMA will Becker, Eberhard, Fluorescence Micros copy
include artificial aging of paint samples (in (Leitz,1990).
addition to collecting naturally aging samples),
and investigation of practical problems such as Bennett, H., J. Bishop and M. Wulfinghloff,
varnishing and cleaning these surfaces. Practical Emulsions (New York: Chemical
Publishing Co., 1968).
Acknowledgement
Bondy, C., "Binder Desigrr and Performance in
The authors thank the scientists who ran Emulsion Paints," Journal of Oil and Colour
samples and assisted in interpreting the data. Chemists' Association,vol. 51, 1968, p. 409.

438
Bondy, C. and M.M. Coleman, "Film Forma- Lillie, R.D., ed., H.J. Conn's Biological Stains
tion and Film Properties Obtained with Acrylic, 9th edn. (St. Louis, Missouri: Sigma Chemical
Styrene/Acrylic and Vinyl AcetateA/eoVa Co., 1990).
Copolymer Emulsions," Journal of Oil and
Colour Chemists' Association, vol. 53, 1970, Lodge, Robert G., "A History of Synthetic
pp.555-577. Painting Media with Special Reference to
Commercial Materials," AIC Preprints, I6th
Brendley, W.H. Jr. and C. Karl Erhard" Annual Meeting, New Orleans, 1988,
"Water-bome Coatings: Acrylic Polymers," pp. ll8-127.
Paint and Varnish Production, Reprint
(Philadelphia: Rohm and Haas, March 1973) Lodge, Robert G., "Special Report on Acrylic
pp.l-5. and Other Synthetic Media Paintings," McKay
Lodge Conservation Report, No. 3, Spring
Bria, Carmen, "The Development of Synthetic 1991.
Polymer Resins for Artist Use," unpublished
student paper, WinterthurAJniversity of Lomax, Suzanne Quillen and Sarah L. Fisher,
Delaware Art Conservation Program, 1981. "An Investigation of the Removability of
Naturally Aged Synthetic Picture Vamishes,"
Clark, George, History of Staining (Baltimore: Journal of the American Institute for
Williams and Wilkins, 1983). Conservation, vol. 29, 1990, pp. l8 l-19 l.

de la Rie, Rene, "Fluorescence of Paint and Low, M.J.D. and N.S. Baer, "Application of
Varnish Layers," Studies in Conservation, Infiared Fourier Transform Spechoscopy to
vol.27, no. 3, August 1982,pp. l-7. Problems in Conservation" Studies in
Conservation, vol.22, 1977 , p. 116.
Denizot, Pien6,, Word for Word (Paris: Yvon
Lambert,1979). Martens, Charles R' Technologl,t of Paints,
Varnishes, and Lacquers (New York: Reinhold
Derrick, Michele, James Landry and Dusan Publishing Co., 1968).
SEl[k, Methods in Scientific Examination of
Works of Art: Infrared Microscopy (Getty McKinney, Roger M., Janet T. Spillane and
Conservation Institute Workshop, Loyola George W. Pearce, "A Simple Method for
Marymount University, Jan. 1991). Determining the Labeling Efficiency of
Fluorescein Isothiocyanate Products," Journal
Gachter, R. and H. Muller, eds., Plastic of Analytical Biochemistry,vol. 14, 1966,
Additives Handbook: Stabil izers, Proc ess ing pp.42l-428.
Aids, Plasticizers, Fillers, Reinforcements,
Colorants for Thermoplastics (Munich: Hanser Michalski, Stefan, "Paintings Their Re-
Publishers, 1985). -
sponse to Temperature, Relative Humidity,
Shock and Vibration," Art in Transil: Sudies
Golden, Mark, personal communication, in the Transport of Paintings, lnternational
Golden Paints, N.Y., 1990. Conference on the Packing and Transportation
of Paintings, London (Washington D.C.:
Haken, J.K. and T.R. McKay, "Quantitative National Gallery of Art, 1991), pp. 223-248.
Pyrolysis Gas Chromatography of Some
Acrylic Copolymers and Homopolymers," Nielsen, Hans and K. Raaschou, "Forensic
Analytical Chernistry, vol.45, no. 7, June 1973, Analysis of Coatings," Journal of Coatings
pp. 125r-1257. Technologt, vol.56, no.7l8, 1984.

Held, Al, personal communication, SoHo, On Kawara, personal communication, 1989.

New York, August, 1991.

439
Patton, Temple C., "Paint Flow and Pigment Talbot, Robin R., "The Fluorescent
Dispersion," in: A Rheological Approach to Antibody Technique in the Identification of
Coating and Ink Technolog,t (New York: Proteinaceous Materials," Papers Pres ented by
John Wiley and Sons, 1979). Conservation Students at the 3rd Annual
Conference of Art Conservation Training
Reffner, J.A. and W.T. Wihlborg, "Microana- Programs, Queens University, Kingston,
lysis by Reflectance FTIR Microscopy," Ontario, May 1982, pp. 140-149.
American Laboratory, April 1990, pp. 1-6.
van de Wiel, H. and W. Zom, "Water Borne
Rohm and Haas Co. technician, personal Acrylics and Urethanes for the Coatings
communication, August, 1991. Industry," Journal of Oil and Colour
Chemists' Association, vol. 64, no. 7, July
Sandler, Irving, Al Held (New York: Hudson 1981, pp. 264-268.
Hills Press, 1984).
Wolbers, Richard and Gregory Landrey, "The
Schwarz, Arturo, The Complete Worl<s of Use of Direct Reactive Fluorescent Dyes for
Marcel Duchamp (New York: Harry N. the Characterization of Binding Media in
Abrams Inc., 1969). Cross- Sectional Examinati ons," AI C P r epr in ts,
Vancouver Conference, Washington D.C.:
Sonoda, Naoko and Jean Paul Rioux, AIC, r987.
"Identification of Synthetic Materials in
Modern Paintings: I. Polymeric Varnishes and
Binders," Studies in Conservation, vol. 35,
no. 4, 1990, pp. 189-20a.

440
These base emulsions require a number of addi-
tives to disperse the pigment, to create a homo-
geneous mixture with specific film forming and
levelling properties, and to ensure the shelflife
of the product (Martens 1968). These numerous
additives are present in quite small percentages
and are being constantly re-evaluated and
changed by manufacturers, without regulations
or standards.

Earlier formulations contained higher concen-

trations of fi llers, surfactants, and colloids,
which were water-soluble and resulted in low
gloss (van de Wiel andZom l98l). More re-
H 5 microns
cent formulations include surfactants that Figure 3 Scanning Electron Micrograph oJ the
perform optimally atvery low concentrations suryface of an acrylic emulsion paintfilm, at 2,000x
and intemal plasticizers that do not migrate magnifcation, showing distribution of pigment
through the film. aggregates in the acrylic resin matrLr.

Methods of polymer production have been im-

a large part in the appearance and properties of
proved to obtain high specificity in the proper-
the dry film. For each pigment, the CPVC ex-
ties of the emulsion without the need for
pressed as the percentage of pigment volume to
excessive alterations with additives. Nonethe-
total volume of the film is that point at which
less, no precise information is given about the
there is just sufficient binder to wet and fill the
constituents, and this seems to be problematic.
voids between each particle.
Little is known about the long-term properties
of the additives and how they ultimately affect
The second order glass transition temperafure
the aging of the films. It is clear that some of
(Tg) is another critical factor determining the
these materials act as humectants and have an
properties of these emulsions and their ability
affinity for moisture and particulate matter in
to form a continuous film. A hard latex with a
the environment (Brendley and Erhard 1973).
high glass transition temperature may require
Cleaning an acrylic film with moisture can
coalescing agents to ensure its ability to form
often result in foaming, indicating that there
Itlms. The Tg for emulsions range fiom
are materials being leached. Yet it is not clear
-15'C (soft film) to 30'C (hard film). When the
whether this will have a deleterious effect on
glass transition temperature of a material is too
the paint film by leaving it more porous. or per-
low, cold flow can take place or it can remain
haps even an advantageous one, by removing
tacky near room temperature. The acrylic resins
hydrophi lic components.
become glassy or brittle within a very narrow
range below room temperature. These flexible
The drying mechanism of the acrylic paints is
surfaces can also attract particulate matter and
vastly different from haditional oil binders,
deform readily in warm conditions. Acrylics
which auto-oxidize and crosslink. The acrylic
can become glassy at temperatures as high as
emulsions form a film by evaporation of the
5'C (40'F). Mechanical cracking can result if
water phase. Strong capillary tensions draw the
stress is applied at these temperatures. Thus,
water out of the acrylic film, forcing the long
such high or low temperatures present prob-
polymer chains to deform and stack together in
lems for artworks in storage or travel,
a tightly packed polyhedral arrangement
especially if they are subjected to impact or
(Bondy and Coleman 1970). The emulsions are
manipulation (Michalski 1991).
suspensions of spherical resinous particles that
are very close in size to the pigment particles
Relative humidity (RH), air circulation, and
incorporated in the film (Figure 3). The critical
pigment volume concentration (CPVC) plays
temperature affect the initial drying process of

413
 rsBN 0-660-57E51 -9

t|[il[lJJil[ililil|

Canadian Conservation
Institute
Communications
Canada