Series in Applied Chemistry and Materials Science

Serie en chimie appliquee et science des materiaux

Introduction to Materials Science

Jean P. Mercier, Gerald Zambelli, Wilfried Kurz
ISBN: 2-84299-286-5

2002 Editions scientifiques et medicales Elsevier SAS. All rights reserved.

23, rue Linois, 75724 Paris cedex 15
www.elsevier.fr

All rights reserved. No part of this publishing may be translated, reproduced, stored in a retrieval system
or transmitted in any form or by any other electronic means, mechanical, photocopying, recording or
otherwise, without prior permission of the publisher.

Imprime en France par Louis-Jean, 05003 Gap ISBN : 2-84299-286-5

Depot legal: 592 septembre 2002 ISSN : en cours
 Series in Applied Chemistry and Materials Science

Serie en chimie appliquee et science des materiaux

Introduction
to Materials Science
Jean P. Mercier, Gerald Zambelli,Wilfried Kurz

ELSEVIER
Paris, Amsterdam, New York, Oxford, Shannon, Tokyo

A member of Elsevier Science

VI

Copyright credits
We want to thanks the publishers and societies for their permission to reproduce the data
mentioned hereafter:
- Academic Press (New York): figure 5.10.
- Abington Publishing (Cambridge), figure 13.12.
-Applied Science Publisher (Barking, Essex-UK): figure 13.10.
- Butterworth Scientific Limited (London): figures 5.8. and 5.11.
- Cambridge University Press (Cambridge): figures 2.25., 16.4. and 16.5.
- Clarendon Press (Oxford): figures 2.23., 2.24. and 11.6.
- Cornell University Press (Ithaca, N.Y.): figure 12.25.
- Elsevier Science Ltd (Kidington, U. K.), figure 17.17.
- Groupe Frangais des Polymeres (Strasbourg): figures 10.4., 10.5., 16.11. and 16.12.
-Hachette (Paris): tables 18.2., 18.3., 18.4. and 18.5.
- Houghton-Mifflin (Boston): figure 6.19.
- Iliffe (London): figures 5.8. and 5.12.
- Imperial Chemical Industries (London): figure 16.7.
- Kabelwerke Brugg Product Limited and Sonderdruck aus 'Internationale Seilbahn':
figure 11.14.
- Macmillan Publishing Company (New York): figures 8.16. and 12.9.
- Masson (Paris): figure 15.8.
- McCraw-Hill (New York): figure 7.15.
- Oliver and Boyd (London): figure 11.4.
- Oxford University Press: figure 5.12.
-Pergamon (Oxford): figures 12.5., 12.6., 12.12., 12.27., 13.7. and 16.17.
and table 15.7.
- Plenum Press (New York): figures 10.14. and 12.11.
- Rossignol: figures 16.14. and 16.15.
- Royal Society: figures 2.23. and 2.24.
- Scientific American: figures 2.9., 4.9., 12.30., 14.24., 14.27., 14.29. and 14.30.
- Sonaca, figures 16.9. and 16.13.
- Societe d'Edition Scientifique (Paris): figures 5.16. and 14.16.
- Springer Verlag (Heildeberg): figure 17.5.
- Sulzer Freres S.A. (Winterthur): figure 1.9.
- Toyota Motor Company (Japan): figure 10.19.
- V a n Nostrand Reinhold (Wokingham, U.K.): figures 3.10., 3.11., 3.12., 3.14.,
and 9.12.
-Verlag Stahleisen (Dusseldorf): figure 12.21.
- V o n Roll: figure 11.13.
- W i l e y : 7.9., 7.12., 7.25., 8.22., 10.16. and 16.2.
 VII

Preface
Notwithstanding its major impact on our industry and society, the revolution in materials
science that we are witnessing today is by no means spectacular to the layman. However, to
feel its dramatic effects one merely has to observe the tremendous evolution of materials,
which constitute the common denominator to all important technological developments, wha
tever the sector considered.
Even more discrete are the intellectual efforts carried out by research teams of in the area of
the solid state since the advent of quantum mechanics in the 1920's. Their work allows us to
juggle with the microscopic structure of materials, whether organic or inorganic, in order to
create new solids or to transform radically those which already are part of our daily life.
Currently, the overall understanding of the different types of material heavily leans on
concepts related to atomic and molecular structure and to microstructure. In certain cases, ab
initio calculations make it possible to understand and to predict the macroscopic properties.
This is notably the case in relation to the electronic structure of some inorganic solids, close
ly followed by organics. Of course, the development of the calculation methods and the capa
city of computers is no stranger to these latest studies. But are these powerful computing tools
themselves in a large part not a consequence of material development?
A unified teaching in Materials Science is essential, particularly where, at the present time,
empiricism is strongly being shaken by the intensity of scientific innovation in this area. This
work fits well in this global vision. Its exposition is well balanced between fundamental
notions and applications and all illustrated with concise tables and well-chosen illustrative
examples. Likewise, the three main classes of materials, as per the classification adopted by
the authors (metals, organic polymers, ceramics), are treated following the same approach. The
link is established, whenever possible, between the macroscopic behaviour of the material and
the properties on an atomic scale, also passing via the microstructure when it applies. The
concepts of thermodynamics are omnipresent to describe the situation.
The didactic approach of this book is perfectly suited to engineering students, as well as to
physicists or chemists who are not specialized in materials but who, nevertheless, wish to be
aware of this discipline. The work should also be much appreciated by specialists in a parti
cular aspect of materials science wishing to have a global view on the subject and to position
their activity in a wider context. It will be equally very useful to those who left University ben
ches more than ten years ago.

Jean Paul Issi

Professeur l'Universite Catholique de Louvain
VIII

Acknowledgments

This book constitutes an upgraded version of the third edition of Introduction a la science
des materiaux first published by the Presses polytechniques et universitaires romandes
(Lausanne) in 1987. The basic translation was carried out by Alan Gay. We want to thank him
very much for this.
We are very grateful to Professor Anthony Kelly, CBE, FEng, FIM, FRS for the time he
spent in revising and improving the manuscript.
We want to thank sincerely all the colleagues and members of the Universite catholique de
Louvain and of the Ecole polytechnique federate de Lausanne as well of members of acade
mic and industrial communities for their contribution and in particular:
R.E. Bauman, J. Bauvois, C. Bailly, W. Benoit, J.J. Biebuick, J.C. Charlier, S. Demoustier-
Champagne, W. Form, M. Gerl ( t ) , L. Haeny, H. Hofmann, Y. Houst, Ch. Huet, B. Ilschner,
L. Ingel, J.P. Issi, Y. Kaenel, D. Landolt, J. Lemaitre, A. Mocellin, J. Moreau, J.B. Nagy,
B. Nysten, L. Piraux, B. Poulaert, M. Rappaz, Y. Reverchon, G. Smets ( t ) and J. Vogt.
We are indebted to the following individuals who have made contributions to this edition:
Frans Gerardis, Frederic Fabre, Bryan Neal, Elsbeth Schlosser.
Thanks and apologies to others whose contributions we have overlooked.
 IX

Foreword

This book, which constitutes the first English edition of the French bestseller Introduction
la science des materiaux, fits into a global and unified vision of this discipline.
This book is addressed not only to students but equally to all engineers and scientists who
are willing to acquire a basic understanding of the 'Materials Science'. The objective of the
work is to present in a coherent approach the general characteristics and properties of the prin
cipal materials.
An effort has been made with the aim of unifying theories and with integrating the charac
teristics of the assembly of materials, but avoiding to isolate them in categories as is often the
case in earlier publications. Such an approach should contribute to opening the frontiers which
still exist between metals, organic polymers and ceramics. The work presents the diversity of
the structures and microstructures of materials, their behaviour under different conditions or
in different environments, with a particular emphasis on mechanical properties.
To facilitate reading and to illustrate the fundamental principles, a certain number of results
are given in the form of examples at the end of each chapter.
Three main subjects are considered in this work.
- A basic description of the principal materials (metals, ceramics and organic polymers);
- A concise study of structural imperfection in crystals, phase equilibria and transitions
with the introduction of the concept of alloys and a description of microstructures;
- An analysis of the behaviour of materials under mechanical constraints and during rup
ture, and also their physical properties.
A concise description of those physical properties which are the most important for the
materials science engineer is equally given in this work. It is completed by a study of degra
dation and corrosion as well as by a description of the main composite materials and by an
introduction to nanomaterials and nanostructured materials.
 Chapter 1.

Materials

1.1. Objectives
Familiarisation with the different classes of materials and their proper
ties.
Description of the microstructures of materials and their behaviour.
Definition of the Materials Science.
Analysis of the life cycle of materials and definition of the Concept of
Resources.
Materials have always fixed the level of development of our civilisation. The
first steps of the humanity were thus marked by the Stone Age, the Bronze Age and
the Iron Age. Today, materials play a decisive role in all technological changes.
Whatever his speciality, the engineer can neither conceive of, nor build new objects
without taking into account the properties of the materials of which he makes use. It
is generally the behaviour of materials which limits the performance of machines and
equipment.
The properties of materials are defined by the nature of their chemical bonds,
their atomic ordering and their microstructure. The study of the relationships between
the organisation on the atomic scale, the microstructure and the properties of materi
als is the realm covered by Materials Science. As well as their behaviour, the use of
materials also depends on their availability, their cost and the methods of synthesis
and production. The use of materials is also influenced by processing techniques,
their facility to be shaped and their environmental acceptability.

1.2. Materials
1.2.1. How is a material defined?
The materials making up the surrounding world consist of discrete particles,
having a submicroscopic size. Their behaviour is determined by atomic theories. The
states of organisation of materials range from the complete disorder of atoms or
molecules of a gas under weak pressure to the almost complete order of atoms in a
monocrystal.
In this introductory work materials are defined as solids used by man to produce
items which constitute the support for his living environment.
Indeed, no object exists without materials. All sectors of human activity depend
on materials, from the manufacture of an integrated circuit to the construction of a
hydroelectric dam. They appear in our bodies to strengthen or replace our damaged

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

2 Introduction to Materials Science

biomaterials. Materials are also as indispensable to our society as food, energy and
information. Their essential role is too often forgotten.
The definition employed in this introductory work is limited to solid materials. It
excludes liquids and gases, as well as solid combustibles.

1.2.2. Complementarity of materials

Materials are commonly classified according to their most typical properties:
materials with high mechanical strength, electricity conductors, permanent magnets,
etc. To demonstrate the multiple functions fulfilled by materials, in the following
paragraph, the composition of the various elements of an electric line under high
tension will be examined.
The suspension cable carrying the electric current must be a good conductor of
electricity to minimise losses. A very high electric tension is used to increase the
capacity of such cables. For reasons of economy, the cables are, in general, sus
pended on pylons using air as an insulator. The cable must be light with a very high
tensile strength in order to limit the number of pylons. In the atmosphere, the best
electrical conductors are very pure metals, copper and aluminium. However these
metals in the pure state do not have good mechanical properties. It is therefore neces
sary to use a conducting cable made up from several materials. The core of the cable
is made of steel wires, resistant to mechanical stresses, but with a lower electric
conductivity. The transport of the electrical energy is mainly assured by pure alu
minium wires, which are lighter than copper and arranged around the steel cable
core.
The pylons are made of steel to resist the weight of the cable. The steel has to be
protected against corrosion (rust) by painting with a polymeric material or by a me
tallic coating such as zinc. Insulating elements are necessary to fasten the cables to
the metal pylons. These are made of porcelain (a ceramic material) while concrete
(another ceramic material) is used to anchor the pylons into the ground.
Half a dozen materials are therefore needed to construct a high voltage line and
by a suitable combination, a functional system can be built. This example of limited
complexity shows that a technical realisation generally involves an important number
of materials.

1.2.3. The three classes of materials

Materials are classified according to various criteria such as their composition,
their structure or their properties. Here, distinction is made between three large
groups of materials (figure 1.1.). This classification is based on the atomic structures
and on the nature of bonds:
metals and their alloys (metallic bonding);
organic polymers (covalent bonding and secondary bonding);
ceramics (ionic bonding and covalent bonding).
This classification can be examined with the help of the Periodic Table of the
Elements {Mendeleev's Table - figure 1.2.).
The majority of the elements are metals (approximately 70) (to the left and in
the centre of Mendeleef s table). The non-metals, such as oxygen occupy the right
hand side of the Periodic Table. In the intermediate region between metals and non-
 Materials 3

Metals + Ceramics METAL^XOYS Metals + Polymers

Cobalt + WC Fe, AI Cu, S f c d C Q r d s + r u b b e r

(hard metals) steels (Pneumatic tyres)

Reinforced concrete 1^ b r a s s
1 ^

CERAMICS POLYMERS

AI2O3, Thermoplastics,
S13N4, SiC, elastomers,
mineral glasses organic fibers

Ceramics + polymers
Glass fibers +
polyesters
Carbon fibers +

Figure 1.1. The three classes of materials: metals, ceramics and organic polymers with
some possible combinations of composite materials.

2
1 He
1,01 4,003
3 4 5 6 7 8 9 10
2 Li Be C 0 F Ne
6,94 9,01 10,81 12,01 14,01 16,00 19,00 20,18
11 12 13 14 15 16 17 18
3 Na Mg AI Si S Cl Ar
23,00 24,31 26,98 28,09 30,97 32,06 35,45 39,95
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
4 Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Br Kr
39,10 40,08 44,96 47,90 50,94 52,00 54,94 55,85 58,93 58,71 63,54 65,37 69.72 72.59 74.92 78,96 79,91 83,80
37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54
5 Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te I Xe
85,47 87,62 88,91 91,22 92,91 95,94 (99) 101,1 102,9 106,4 107,9 112,4 114,8 118,7 121,8 126,6 126,9 131,3
55 56 57 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86
6 Cs Ba La Hf Ta W Re Os Ir Pt Au Hg Tl Pb Bi Po At Rn
132,9 137,3 138,9 178,5 180,9 183,8 186,2 190.2 192,2 195,1 197,0 200,6 204,4 207,2 209,0 (209) (210) (222)
87 88 89
7 Fr Ra Ac
(223) (226) (227)

58 59 60 61 62 63 64 65 66 67 68 69 70 71
6 Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu
140,1 140,94 144,2 (147) 150,4 152,0 157,3 158,9 162,5 164,9 167,3 168,9 173,0 175,0
90 91 92 93 94 95 96 97 98 99 100 101 102 103
7 Th Pa U Np Pu Am Cm Bk Cf Es Fm Md No Lw
223,0 (231) 238,0 (237) (242) (243) 247) (249) (251) (254) (253) (256) (245) (256)

Figure 1.2. The Periodic Table of the Elements with indication of atomic weight (Men-
deleef s Table). The metals are shown in bold type and the non-metals in standard type.
There are a certain number of elements such as boron, silicon, which have intermediate
behaviour. These elements are shown in Italics.
4 Introduction to Materials Science

metals, there occurs a certain number of elements such carbon and silicon (semicon
ductor) which escape this simple classification.
At ambient temperature, most of the metals are monoatomic solids. The most
commonly used metals are iron, aluminium and copper. The metallic alloys are often
combinations of two or more metals, as in the case of the brasses (alloys of copper
and zinc), but they can equally contain non-metallic elements. Most of the steels
(alloys of iron and carbon) are examples of this type of alloy.
Metals and their alloys are normally good conductors of heat and electricity and
are opaque to visible light, which they reflect. They are, in the main, hard, rigid and
can undergo plastic deformation (chapter 10). A significant number of metals have a
high melting temperature (T ).m

The organic polymers are materials made up of molecules forming long chains
of carbon atoms to which are attached elements such as hydrogen or chlorine, or
groups of atoms such as the methyl radical (-CH3). Other elements such as sulphur,
nitrogen, silicon, etc., can also participate in the composition of the chain.
The best-known organic polymers are poly(vinyl chloride) (PVC), polyethylene
(PE) and polystyrene (PS). Section 18.3. gives the principal abbreviations used in
organic polymers with their chemical formula. They are often known by their 'trade
name': polymethylmethacrylate (PMMA or 'Plexiglas'), polyamides (PA: 'Ny
lons'), polytetrafluoroethylene (PTFE or 'Teflon'). The organic polymers (or
ganic glasses, rubbers, etc.) have widely diversified physical properties. They are
almost always electrical and thermal insulators. They are light and easily formable.
Contrary to metals, the non-oriented polymers have little rigidity and for the most
part do not withstand temperatures above 200 C.
Ceramics are inorganic materials which, as a general rule, result from the com
bination of a limited number of metallic elements (Mg, Al, Ti, etc.) with non-metallic
elements, most commonly oxygen. Originally, the term ceramic was restricted to
oxides (silica S i 0 , alumina A 1 0 , etc.), but this classification has been enlarged to
2 2 3

include other combinations of atoms such as tungsten carbide (WC) or silicon nitride
(S13N4).
The ceramic materials are especially known for their fireproof character, that is
to say by their high mechanical and thermal resistance at high temperature, but the
use of ceramics is far from limited to this type of application. Ceramics are generally
very hard materials and widely used as abrasives. Most ceramics are electric and
thermal insulators, but they include some of the best thermal conductors, such as
diamonds. A large number of ceramics also have important applications in electrical
equipment as well as in electronics. Mineral glasses, which are combinations of
oxides (S1O2+ N a 0 + CaO, etc.) in an amorphous structure, are also classed as
2

ceramics. Generally, glass and ceramic are fragile materials not usable where
mechanical and thermal shocks are important.
The three types of materials can be combined to form composite materials (fig
ure 1.1.). A composite material is constituted by two or more different materials,
which combine in a synergetic way their specific properties. This is the case of epoxy
resins (a polymer) strengthened with glass fibers, which forms lightweight compos
ites with high mechanical strength. Concrete, an agglomeration of cement and
gravel, represents another example of a composite material.
 Materials 5

A subdivision of materials into three main classes, based especially on atomic

and structural characteristics and properties, is convenient but is arbitrary in charac
ter. Thus diamond, which can be considered to be a three-dimensional polymer, is an
organic material because it consists of carbon atoms. However, the hardness and the
mechanical properties connect it rather with ceramic. Also, electric conductivity is no
longer the exclusive property of metals because certain oxides ( V 0 , Y B a C u 0 )
2 2 3 7

and some organic materials also conduct electricity.

Any classification of materials possesses such an arbitrary character, for there is
no discontinuity between the three classes of materials. Other classifications, based
on specific material properties such as semi-conductivity, can also be justified.

1.2.4. Materials properties

A material exhibits a set of properties, which defines its behaviour. A property
of a material is determined by analysing the reaction of the material to some outside
influence, generally by means of a normalised standard test. According to the type
of outside influence, three categories of properties are recognised.
Mechanical properties, which reflect the behaviour of materials, de
formed by a set of forces.
Physical properties which describe the behaviour of materials subjected
to the action of temperature, electric or magnetic fields, or light.
Chemical properties which characterise the behaviour of material in a
reactive environment.
The realisation of objects and structures designed by the engineer is frequently
limited by the properties of the available materials. Any important technological
progress is often linked to the development of materials endowed with improved or
new properties. It is possible to increase in a substantial way the efficiency of gas
turbines (aircraft jet engines) by working at a higher temperature. The development
of more successful engines is thus linked with finding metal alloys more resistant to
creep or ceramics having better resistance to thermal shock.
Certain properties are derived directly from the atomic arrangement and the
nature of the bonds. This is the case for the opaqueness of metals, the transparency of
glasses or the extensibility of rubbers.
Other properties of materials are strongly dependent on their microstructure.
This microstructure consists of grains or particles of microscopic size, having a de
fined morphology.

1.2.5. Use of materials

It is crucial to choose materials which are best adapted for the applications envi
sioned. The following factors should be taken into account:
The main functions of the objects and structures planned; it is, for ex
ample, necessary to determine the mode of loading, the temperatures as
well as the general conditions of use.
The intrinsic behaviour of materials: strength, wear and corrosion re
sistance, conductivity, etc.
6 Introduction to Materials Science

The behaviour of materials during manufacture and processing.

The environmental behaviour of the materials.

The cost of various possible solutions.

The possibilities of recycling or for elimination by incineration with en

ergy recovery.

In the development of a technology, it is common to substitute one material for

another, either for reasons of performance or for economic or ecological motives.
Consider the example of car bodywork. Initially such bodies were constructed from
wood. Progressively metal replaced wood and was then partially supplanted by or
ganic polymers. If metal was substituted for wood, it was above all because it is
easier to shape with controllable properties and was less sensitive to humidity; all
these factors simplified production lines manufacturing and increased reliability.
To save energy, the weight of moving structures and objects was often reduced.
It is for this reason that polymers are now used in the automobile industry since they
have a density of approximately 1,5 g e m (or t m ) whereas that of sheet steel,
3 3

constituting at present the major part of bodies, is about 8 g e m . Car bodies are
- 3

lightened by using thinner high resistance sheet steel produced by a better knowledge
of hardening mechanisms (fine grain steels) or by the use of sheet aluminium. How
ever, the lighter materials have poorer mechanical properties than those of traditional
materials. For those applications, which involve high tensile stresses, the rigidity-to-
weight ratio must be considered in order to make a choice of materials. The value of
this ratio is practically the same for steel and for aluminium. On the other hand, for
thermoplastic polymers, rigidity-to-weight ratio is about 10 times less. This simple
analysis shows that, in spite of their low densities, polymers cannot be used where
significant stresses will be involved.
The materials selection entails other criteria (weight, corrosion resistance, and
appearance...) besides mechanical strength. Nevertheless, the use of organic polymer
materials has developed with an accelerated rhythm. Especially, due to the introduc
tion of high modulus fibers (carbon, glass, etc.), it is possible to obtain mechanical
properties superior to those of ferrous materials while keeping a density lower than
2 g e m (or t m ) . These materials, called composite materials, have entered progres
-3 -3

sively into different key markets (aviation, car manufacture, and sports goods). How
ever, the development of composites has been slowed by production and manufac
turing costs rising higher than that of the classical materials. They are also severe
problems of recycling for composite materials.
The manufacture of watch glasses also underwent a considerable evolution in
the course of the years. At first the fragile mineral glass was replaced by an organic
glass (a polymer) more resistant to the impact but which could be scratched. At pre
sent, a monocrystal platelet of synthetic sapphire is also used. This ceramic material
associates transparency with impact and scratch resistance.
These examples show that materials are in constant development. A material
developed for a specific application often has uses in other domains of the technol
ogy. This increasing development of materials, associated with a better theoretical
understanding of their behaviour, led in the 1950s to the birth of Materials Science.
This concept is directly related to the notion of microstructure.
 Materials 7

1.3. Materials science

1.3.1. Microstructure
At present, the blades of gas turbines keep good mechanical properties and re
sistance to oxidation up to a temperature slightly higher than 1000 C. These blades
are made of nickel based metal alloys containing a dozen elements, including
aluminium, titanium and chromium. The resistance of these alloys to high tempera
ture mechanical stresses essentially results from their microstructure.
As already mentioned in paragraph 1.2.4, microstructure describes the totality of
grains or particles of microscopic size, which characterise the structure of a large
number of materials. The microstructure of a material is therefore formed by the
assembly of grains, particles or defects observable by microscopy (optical and elec
tronic).
To understand the behaviour of a material, it is necessary to establish a connec
tion between phenomena taking place on microscopic and submicroscopic scales and
the properties of the material. This connection relates to phenomena occurring on
scales across nine orders of magnitude from the nanometer (10~ m) to the meter 9

(figure 1.3). Several steps of magnification are therefore necessary to pass from the
scale of a gas turbine blade to the observation of the microstructure and the
determination of the atomic arrangement. These successively involve the optical
microscope (for observation from mm to ), the electron microscope (for observa-

Machine f ^ ^ ^ 3 I 3m E
? e

, ^ "
Piece JV J 30 mm So
V*^ Jj
/^Y V~^y Optical 2
"\ \?J^j[ microscope w g
Grains \ . [o,3mm g

V-s. *^
Grain indaries : \ V f3 s
yT^<-\ S Electron ^ 00
/ , . '.^"S microscope ^

' /
II I 1
Precipitates in
t . 1 do 1.1

1 matrix , \30Bm
U ^~

atoms the
Matrix between
Figure 1.3. Relation ^ structural
^ andI 3 nm / \size in some of the main
microstructural
domains of science and technology.
XXJ^X% X R
*y
8 Introduction to Materials Science

tion from 100 to run) and X-Ray diffraction for the determination the structure on
an atomic scale (observation on a nm scale). The elements of the microstructure of a
material are generally defined by several parameters:
composition;
atomic arrangement;
relative quantity;
morphology;
size.
The microstructure determines the properties of a very large number of materi
als. Thanks to the controlled modification of microstructures during manufacturing or
processing, the engineer in materials can obtain a broad range of properties. A varia
tion in microstructure as a function of time during the use of a material leads to an
important modification of its properties (ageing phenomenon). To obtain reliable
materials, it is essential to master the ageing phenomena, which are the result of
physical modifications or chemical reactions.

1,3.2. Materials Science

Materials are subjected to complex and multiple stresses. Their behaviour for a
specific application is represented by a set of properties. The behaviour of materials
during manufacture or shaping (for example, fluidity, forming properties...) is also
important.
The concept of Materials Science arose from the necessity of acquiring a knowl
edge of the fundamental laws, which determine properties. Materials Science seeks
to establish the relations existing between composition, atomic or molecular organi
sation, the microstructure and the macroscopic properties of materials. Materials
Engineering concerned with manufacturing, transformation and shaping processes,
complements Materials Science.
A fundamental knowledge of materials was not required when man contented
himself with some clay, some wood and some wool to satisfy most of his necessities.
The empirical approach and experience accumulated by the metalworkers and ce
ramists over thousands of years no longer suffice to satisfy contemporary needs and
to meet the complex requirements of modern technology. An unified, quantitative
and fundamental approach to a description of the behaviour of the engineering
materials has become absolutely essential.
Materials Science has a general character and a multi-disciplinary approach
requiring the knowledge of chemists and physicists for basic sciences and those of
the engineer (chemical, mechanical, electrical, civil) for applications and manufac
turing. Materials science emerges as a coherent whole, coupled with Materials Engi
neering which has the objective the producing materials of well defined properties.
Materials Science treats the whole of materials (metals, ceramic, and polymers) in an
unified way with the same theoretical background and using the same experimental
tools. As schematised on figure 1.4., there are four main aspects Materials Science
and Technology: synthesis, manufacture and processing, composition and structure,
properties and performance. The behaviour in manufacture and in use coupled with
economic factors characterise the performance of a material. Closely linked are four
aspects of Materials Science. The material is elaborated during synthesis (polymer)
or manufacturing (metals, alloys, ceramics, etc.). Processing concerns the shaping of
 Materials 9

a material and the preparation of a finished object according to its behaviour. For
example, the production of a car body involves successively rolling of the sheet steel
from a bar of steel, the stamping of the sheet steel to form the body and a series of
finishing operations (painting, etc.).
Performances

Properties
Synthesis
Manufacturing
Processing

Composition
Structure

Figure 1.4. The four basic aspects of materials science and technology.

To obtain optimal properties, it is essential to master the structure and composi

tion of the material and consequently to have access to a series of sophisticated
analysis techniques.
It is the numerous contributions of Materials Science and Technology, which
has completely remodelled the world, which supports us by freeing man of a huge
number of constraints, linked to our environment. Our way of life has been radically
transformed within a few decades largely due to the contributions of Materials
Science and Engineering which lead to the creation of the tools of the modern life:
automobiles, aircraft, bridges, cable cars, computers, telecommunications equipment,
satellites etc.

1.4. The life cycle of materials and resources

1.4.1. The materials cycle. Recycling
The production capacity of materials is directly bound up with natural resources
and the possibility of waste recycling. After extraction, raw material undergoes vari
ous physical and chemical modifications, which end in the production of a material.
The accumulation of equipment and consumer goods produces a localised surplus of
used materials, and this leads to the problem of the materials recycling. Figure 1.5.
describes the life cycle of materials, with its multiple phases of successive process
ing, which go from resource exploitation to waste formation. An optimal manage
ment of this cycle remains very difficult to achieve in practice. Thus mastering it
becomes, in the long run, more and more important for safeguarding the environ
ment.
Recycling is usually easy for metals. For other materials, like organic polymers,
recuperation remains difficult because of their complex chemical structures which
make difficult to separate the constituents of the materials and prevent alloy forma-
10 Introduction to Materials Science

tion. For ecological reasons and under public pressure, different processes for recy
cling of polymers have been proposed. In general, these produce only l o w p e r
f o r m a n c e materials and the processes remain difficult to very difficult to improve.
Only by standardising the polymers which are used will an efficient recycling
method be found.
An example of current development is that of polymers used in automobiles.
Because of the significant tonnage used, they are selected not only for performance
but also as a function of their recycling possibilities. This limits the number of poly
mers used in a vehicle. For commodity polymers such as those at present used in
packaging, it is actually more economical to incinerate them to recover their energy
content, but this approach evidently poses an ecological problem.
On the other hand, the recycling of mineral glasses (ceramics) and paper (cellu
lose), which encouraged by public authorities in the majority of the European and
North American countries, is now almost completely integrated into daily life.

Materials )"^~^-^
Refining Manufacturing

{
./
Raw X
// X

\
( materials ) ' (ComponentsJ

Extraction Recycling Assembly
Manufactured /
\
\ products /

( Waste j

Figure 1.5. The Life Cycle of Materials ( from: Materials and Man's Needs, 1974 and A.
Kelly, 1994).

1.4.2. Resources and reserves

The resources of an element consist of the quantity of that element available in
the earth's crust, its oceans and atmosphere, and which could be extracted in the
future. All the resources are not therefore usable at the present time.
To calculate the resources of the earth in chemical elements, their average con
centration in the continental earth's crust only is estimated to a depth of about 1 km.
This limited portion of the earth's crust corresponds to a total mass of about
1 0 t o n s . Average concentrations of t h e elements in the earth's crust is ordi
18

narily much too small to make profitable the production of most metals. Only depos-
 Materials 11

its, i.e. zones where concentration of a mineral has resulted from geologic processes
of enrichment, are worth exploitation.
This portion of the resources, which may at present be economically exploited,
is called a reserve. Reserves are therefore more important than resources. The limit
between the two is governed by economic and technological factors related to their
exploitation. These vary with time and also depend on the economic strategy of dif
ferent states and of large industrial groups. Nine elements constitute 99.4 % of the
mass of the earth's crust (table 1.6.), among these, two important metals: iron and
aluminium, which are at present produced on a large-scale. Concentration averages
of the other metals in the earth's crust which do not appear in Table 1.6 is less than
0.01 % (100 gt" ) This is the case of copper, which is produced in practically the
1

same tonnage as aluminium.

Table 1.6. Average distribution of the main elements in the earth's crust,
the oceans and the atmosphere.

Continental crust (1 km) Oceans Atmosphere

Mass ~ 1 0 tons
18
Mass ~ 1 0 tons
17
Mass ~ 1 0 tons
16

(% weight) (% weight) (% weight)

0 47 85 79
Si 27 10 19
Al 8 CI 2 Ar 2
Fe 5 Na 1
Ca 4
Na 3
3
Mg 2
Ti 0,4

About 96 % of the volume of the earth's crust is composed of oxides, which

constitute an inexhaustible supply for the manufacture of ceramic. Organic polymers,
produced from coal and hydrocarbons are also resources existing in large quantities.
The extraction of resources and the production of materials demand considerable
quantities of energy. The price of materials is therefore strongly dependent on the
price of energy. From an energy point of view, the organic materials are particularly
favoured, because the energy requirement for their synthesis (including their intrinsic
energy contents) and for their transformation is much lower than that for metals or
ceramics production. Certain metals, aluminium in particular, require a huge amount
of energy. Their use has slowed and been partially replaced by organic materials
because of the cost of energy.
Figure 1.7. gives an estimation of the cost price of the main materials while
figure 1.8 indicates the production cost by unit of weight of the various manufactured
objects. Building materials such as concrete, which are the least expensive, are most
used. Metallic materials are employed in practically all the fields of industrial activ
ity, in particular for automobiles, because they exhibit a better combination of me
chanical properties (rigidity, high strength, deformability, and toughness). Thermo
plastic polymers, particularly well adapted to mass production, are widely used in
household electrical appliances. Elastomers (rubbers) constitute another important
category of polymeric materials (tyres, joints, flexible hoses, and elastic supports...).
12 Introduction to Materials Science

10-

S 10 4

% 3
Sructural composite
'53 materials

10 2
Polymers
Specials
materials
10
13
Construction
c materials
j 1 1 1 CllV^X
Ceramics and
Metals mineral glasses
and alloys
10"

Type of material

Figure 1.7. Cost price of various types of material per unit weight (from Ashby)

Mineral glasses, which are inexpensive ceramic materials, are mainly used for
their transparency. Technical ceramics (zircon (Zr0 )), silicon nitride and carbide,
2

etc. are much more expensive and reserved for technical applications at high tem
peratures (1900 - 2300 C). An important application for glass fibers composites is
shipbuilding (especially for pleasure boats), whereas the use of use of carbon based
fibre composites, much more expensive, is limited mainly to civil aviation for flaps
and tail planes and the military and aerospace industry. Speciality materials such as
the noble metals (gold, platinum, etc.), diamond and hard metals (cermet) are gener
ally reserved for high technology applications.
The manufacture of items requiring very small quantities of material can use
more expensive material in an economically valid way, whereas in making manu
factured goods of greater size less expensive materials must be used to be profitable.
This explains why (figure 1.8) an object as common as a toothbrush, on a weight
basis, costs more than a Rolls Royce car.

1.5. Summary and conclusions

Materials play an essential role in our society. All important technological pro
gress is influenced by improvements in the properties of existing materials, by ap
pearance of new materials or of more successful and less expensive manufacturing
processes.
Materials can be classified in three categories: metals, organic polymers and
ceramics. Each category of materials possesses a series of specific properties char
acterised by its response to the action of external stress. Since certain properties are
 Materials 13

Biomedical
5
Contact lens
Sports
goods Heart valve
.a 10
% Denture
Automobile
1 3
industry
Artificial hip
2

Aeronautics
-
Packaging Aerospace

Tooth brush

I
10

a Shipbuilding
s Electrical
\ appliances

Offshore
2 -' " Construction Ferrari
materials
6 Rolls Royce
Manufactured products
Limousine
Family car
Small car

Figure 1.8. Cost price of manufactured products by unit weight. In addition, the cost of
automotive and biomedical materials (from Ashby).

common to several classes of materials, one material, frequently, for economic or

technical motives, is substituted for another for a definite application.
The properties of a large number of materials are determined by their micro-
structure. This depends on the composition and on the atomic or molecular arrange
ment, but also on the physical and chemical treatments undergone by the material
during manufacture and shaping. On the other hand, any evolution of the micro-
structure occurring in the course of service leads in general to property modification.
Materials Science is considered as a unified and fundamental approach to the
description of the behaviour of materials. This science, still very young, continues to
progress. One of the objectives of the science of materials is to understand the varia
tion in properties as a function of their microstructure and to show the phenomena
responsible for these modifications.
A brief study of the life cycle of materials demonstrates the importance of re
sources and reserves which, together with the price of energy necessary for extrac
tion, processing and recycling, determines the price paid for finished goods.

1.6. Illustrative example: development of prostheses

Major technical revolutions are directly associated with the development of new
materials. This is notable in the case of the spectacular progress realised during
14 Introduction to Materials Science

recent years in orthopaedic surgery. Prostheses are now replacing important and vital
parts of the human body.
Extremely severe conditions are imposed on materials used for the construction
of prostheses. It is necessary in particular that materials be biocompatible, that is that
they are neither toxic for the body nor carcinogenic, and that they do not produce any
hostile reaction which could cause rejection. They have also to satisfy a series of
requirements concerning mechanical performance and longevity. They must, in par
ticular, have good resistance to fatigue and wear as well, as to corrosion in aggressive
media.
The example of the hip prosthesis, among the most wide-spread at present
(-700 000 a year), can illustrate these problems of medical engineering. These pros
theses are made with an anchoring tail surmounted by a spherical knob to be inserted
into a cup fixed on the hip. This prosthesis, represented in figure 1.9, must be able to
make circular movements in three dimensions.

Figure 1.9. Artificial hip prosthesis made from hot forged stainless cobalt steel and al
lowing three-dimensional circular movements.

From 1940, doctors have studied the possibility of implanting prostheses to re

solve the problem of premature wear of the hip joint resulting from an accident or
degenerative osteoarthritis. The first attempts to replace the head of the thighbone by
prosthesis used a cobalt alloy. This solution had to be abandoned because the carti
lage of the femoral cavity was quickly damaged.
In 1950, total prostheses consisting of a metallic rod and a spherical head able to
be inserted into a cup fixed in the femoral cavity were introduced.
The main problems to be resolved were related to the anchoring of the prosthesis
in the bony tissue and to the wear of parts rubbing against each other. From 1960,
 Materials 15

prostheses were developed capable of working for a period of ten to twenty years and
were especially useful for older patients. The anchoring of the tail with a cement
based on organic polymers (PMMA) joined the bone to the prosthesis. This cement
ensured a good transmission of effort. Wear resistance was optimised by using a
femur head in forged stainless steel (Fe, Cr, Ni, Mo) and a polymeric cup made of
very high molecular weight polyethylene.
Parallel to the introduction of new materials, the study of the global concept of
the prosthesis played an important role in the development. Thanks to recent progress
in Materials Science, new improvements have been introduced. For example, the
fatigue resistance of the metallic alloys constituting the anchoring tail has been sub
stantially increased by selecting a fine grained cobalt alloy (Co, Ni, Cr, Mo, Ti). This
alloy grew from a double fusion procedure, carried out under vacuum and followed
by hot forging. Nowadays, use of titanium alloys is becoming widespread. These are
lighter with an equivalent performance to the previously used metallic materials and
with better biocompatibility.
Various developments have continued to improve the sealing of the anchoring
tail, in particular by using porous materials having a structure close to that of
bone and which living bony tissue can infiltrate. Other improvements concern the
choice of material for the cups like polyacetal (Delrin) and are characterised by
improved wear resistance.
The problem of prostheses having a high wear resistance is even more critical
for implantation in young patients since, in these cases, the prostheses normally have
to stay in place for long period. At present, for the cup and the spherical head, a ce
ramic material based on alumina (A1 0 ) sintered under high pressure is used. This
2 3

material presents the advantage of being biocompatible and having good wear and
corrosion resistance but it is however fragile. Composite materials based on car
bon/carbon fibres also offer interesting possibilities for the development of frictional
elements. This new type of material has properties close to those of alumina but with
better mechanical strength and improved impact resistance. The conception of an
artificial hip involves different types of materials. The choice of these materials is the
result of compromise since a large number of factors must be taken into account. It is
necessary to consider, for example, not only the specific characteristics of each type
of material but also the combination of technical and economic constraints tied in
with the manufacture and implanting of prostheses. Prostheses development
constitutes an excellent example of multidisciplinary activities in which Materials
Science plays an essential role. It is therefore not surprising that there are over 300
different types of artificial hip on the market.
At present various important and vital parts of the human body are replaced by
prostheses. These are often less spectacular than an artificial hip or a knee. For ex
ample, there is the implantation of polypropylene fibre mesh to strengthen the ab
dominal wall for the treatment of umbilical hernias. Not biodegradable and well
tolerated by the body, this mesh is gradually infiltrated by cells and becomes, in a
sense, an integral part of the tissue which it mechanically reinforces.
At present, prostheses for the very complex soff organs (kidney, heart, liver,
6

pancreas, skin...) are being developed since transplanting living organs is still lim
ited by the difficulty of finding donors.
16 Introduction to Materials Science

1.7. Exercises
1.7.1. Which are the different types of materials besides the three traditional classes?

1.7.2. Which categories of materials would be chosen for producing a cup? Which
properties should be considered?

1.7.3. To which class of building material does wood belong to ?

1.7.4. Which material could be substituted to replace the steel ball of a pen ? Please
give advantages and inconveniences.

1.7.5. The manufacture of a light bulb involves different materials. Indicate a part of
the light bulb realised with a material belonging to each of three classes.

1.7.6. Name four constituents of a automobile made with of polymers.

1.7.7. Besides the mechanical properties and the physicochemical characteristics of

materials, which other factors should be considered for the manufacture of an ob
ject?

1.7.8. Define the resources and the reserves of an element.

1.7.9. Which are the two most abundant metals in the earth's crust (to 1 km in depth)?

1.7.10. What material could be used to insulate a furnace heated to 1000 C? Why?

1.8. References and complementary reading

D. ALTENPOHL, Materials in World Perspective, Springer, Berlin, 1980.
M.F. ASHBY, D.R.H. JONES, Engineering Materials, An Introduction to their Properties and
Applications, Pergamon , Oxford, 1980.
M.F. ASHBY, D.R.H. JONES, Engineering Materials 2, An Introduction to Microstructures, Processing
and Design, Pergamon , Oxford, 1980.
M.F. ASHBY, Materials Selection in mechanical Design, Buttworth, Heinemann, Oxford, 1992.
J. BLACK, Orthopaedic biomaterials in research and practice, Churchill Livingston, New York, 1988.
W. D. CALLISTER, Materials Science and Engineering, an Introduction, 4th Ed. Wiley, New York,
1991.
M. GERL et J.P. ISSI, Physique des materiaux, (Tratte des materiaux , vol. 8), Presses polytechniques et
universitres romandes, Lausanne, 1997.
. ASTERLING, Tomorrow's materials, The Institute of Metals, London, 1989.
J.E. GORDON, The New Science of Strong Materials or Why You Don 7 Fall Through the Floor, Penguin,
New York, 1976.
A. GUINIER, La structure de la matiere, Hachette, Paris, 1980.
K. KOVACH (ed.), The Rational Use of Potentially Scarce Metals, NATO Sei. Com., Brussels, 1976.
A. KELLY, The changing Cycle of Engineering Materials, Second Finniston Lecture, London, 1994,
Interdisciplinary Science Review, 19, 3, (1994).
Materials and Man's Needs, Materials Science and Engineering, National Academy of Sciences,
Washington, D. C , 1974.
A. NARATH, A. ARDEN Jr., J.H. BIRELY, M. COHEN, W. KOHN and W. P. SLICHTER, Eds.
Materials Science and Engineeringfor the 1990s, National Academy Press, Washington D.C., 1989.
R. SAAGER, Metallische Rohstoffe von Antimon bis Zirkonium, Bank von Tobel, Zurich, 1984.
R. B. SEYMOUR, C. CARRAHER Jr., Polymer Chemistry, an Introduction, 3nd Ed. Dekker, New
York, 1992.
E. WINTERMANTEL , SI-WI HA, Biokompatible Werhtoffe und Bauweisen, Springer, Berlin, 1996.
 Chapter 2.

Atoms, molecules
and the chemical bond

1.2. Objectives
To describe the atom at an elementary level.
To understand the origin of the interatomic forces responsible for bond
formation.
To describe succinctly the three principal strong bonds and the secon
dary bonds.
To define the molecule.
To calculate internal cohesive forces.

In this chapter, the basic building blocks of materials, atoms and molecules, are
presented. Atoms consist of sub-microscopic elementary particles (nucleons and
electrons). Among these, it is above all the peripheral electrons that play an impor
tant role in the study of the properties of materials. The interactions between periph
eral electrons are the origin of the forces of interatomic and intermolecular bonds
and lead to the formation of a rigid, condensed state called solid.

2.2. Atomic structure

The atom is built up from an aggregation of elementary particles having differ
ent electric charges. It consists of a nucleus, made up of neutrons and protons, sur
rounded by electrons. The number of electrons in the neutral atom defines the atomic
number Z, i.e. the number of positive charges (protons) in the nucleus. The atom has
an overall neutral charge because the number of electrons (negative) corresponds to
the number of protons (positively charged).
The nucleons (protons and neutrons) have practically the same mass. This is
used as the unit of atomic mass: m= 1.66 10" g. As this value is extremely small,
24

the mole is used in practice as the unit of atomic mass. A mole of a substance
contains a number of atoms or molecules equal to 6.02 10 (Avogadro's number, 23

A), that is to say, as many particles as there are atoms in 12 g of the C isotope of12

carbon. A mole of hydrogen atom has a mass equal to 1 gram. The atoms of a given
element have the same atomic number Z, while their mass can vary according to the
number of neutrons present in the nucleus (isotopes).
The mass m of the electron is much smaller than that of a nucleons m
e

(mjm = 1/1800). The electron is, therefore, essentially a charged particle without

2002 Editions scientifiques et medicales Elsevier. All rights reserved

Introduction to Materials Science

18 Introduction to Materials Science

influence on the atomic mass. The electric charge of the electron and of the proton is
equal to 1.6-10" C, while neutrons are neutral particles.
19

In 1924, de Broglie suggested to associate a wave with any particle in move

ment. Particle-wave duality was proposed initially to reconcile the Wave Theory and
the Corpuscular Theory of light. According to de Broglie's principle, the relation
(2.1) gives the wavelength , associated with any particle in movement:

(2.1.)
mv

In this expression m and are respectively the mass and the speed of the particle
and Planck's universal constant h (h = 6.63 10~ Js).
34

In the Quantum Theory, the behaviour of the electron is completely described

by its associated wave function . By means of this, it is possible to determine the
probability, dp, of finding the electron in an element of volume dV:

= 2
(2.2.)

where y/ (y/*yf when the wave function is complex) represent a density of probabil
ity. In the Quantum Theory, the behaviour of an elementary particle, such as the
electron is completely described by its wave function. In the Quantum or Wave The
ory, it is not possible to know with certainty the position of the electron in the atom.
The electron does not occupy, as in the Bohr-Sommerfeld's earlier theory, a definite
orbit but it remains most of the time confined in a definite portion of space called an
orbital.
To illustrate this concept, consider the first orbital, Is, of the hydrogen atom.
The variation of the probability density yf with respect to the distance r from the
nucleus is shown in figure 2.1(a). This probability density is a function, which has
spherical symmetry with regard to the nucleus, y/ reaches a maximum value when
r = 0, that is when the electron becomes superimposed on the nucleus, and tends
towards zero when r = <*>. The limit of the orbital is not therefore defined with cer
tainty. By the integration of (2.2.), the regions of the space in which the probability
of finding the electron is equal to 50, 90, 99 % can be calculated, (figure 2.1. (b)).
Generally, the surface that bounds the region of the space in which the probability of
finding the electron is equal to 95 % is chosen as the limit of the orbital.
It is not because the probability density yf passes by a maximum when r = 0 that
the electron has a maximum probability of being superimposed on the nucleus. To
calculate the most probable orbit, one must consider the volume in which the electron
has the possibility of moving. To determine the most likely orbit of the electron, the
probability of finding the electron inside a spherical shell of radius r and of thickness
dr is calculated. This probability, which is equal to the product of the volume of the
shell ATirdr multiplied by the probability density at a point, , defines the radial
2

probability density P(r) = 47try/.

The variation inP(r) shown in figure 2.1. (c) allows the most probable orbit to
be defined. It should be noted that P(r) becomes equal to zero if r = 0 and passes
through a maximum if r = 0.0529 nm. This value is equal to the radius a for the orbit
0

of hydrogen predicted by the Bohr theory.

 Atoms, molecules and the chemical bond 19

Probability
density 2

(a)

(b)

Radial probability
density
4?rr 2 2

(c)

0.1 0,2 0,3 0,4 (rim)

Radius r

Figure 2.1. First atomic orbital of hydrogen: (a) probability density as a function of the
distance r from the nucleus; (b) spheres in which the probability of finding the electron is
equal to 50,90 and 99 %; (c) radial probability density P(r) = 47tr \f/. a = 0.0529 nm.
2
0

Each atom possesses a certain number of electronic orbitals, which are charac
terised by defined energy levels. At 0 , and without any external excitation, elec
trons only occupy the lowest energy orbitals. A specific atomic orbital is completely
characterised by three quantum numbers , I and mi.
The principal quantum number defines the size of the orbital. It can have all
positive, integral values with the exception of zero: ~ 1, 2, 3, etc. The lowest en
ergy state is associated with the value n=\. These energy levels are often designated
by letters K, L, , N, etc. which correspond to n= 1, 2, 3, 4, etc. respectively.
The angular quantum number I determines the shape of the orbital. It is associ
ated with the momentum of the electron in its orbit. This number is related to the
value of n. It can only have values between 1 = 0 and / = - 1. The energetic sub-
levels corresponding to each value of the angular quantum number / = 0, 1, 2, 3, etc.
20 Introduction to Materials Science

is designated by a letter s, p, d, f, etc. This letter is placed after the principal quantum
number n. The energy level 2p thus corresponds to a quantum number = 2 and
/ = 1. In the absence of a magnetic or electric field, other than that created by the
nucleus, all orbitals having the same values of and / have the same energy.
The shape of the orbital is determined exclusively by the value of the angular
quantum number /. This does not therefore depend on the value of the principal
quantum number n. All s orbitals (/ = 0) have spherical symmetry (figure 2.1. and
2.2. (a)). The superior order orbitals (p, d...) have complex shapes. Thus the orbi
tals (figure 2.2. (b)) are each made up of two lobes separated by a plane where the
electronic density is zero (nodal plane). The even higher order orbitals have corre
spondingly more complex shapes.

(b)

Figure 2.2. Schematic representation of two types of atomic orbitals: (a) s orbital, (b)
orbital.

The third quantum number is the magnetic quantum number m which defines h

the spatial orientation of the orbital. The magnetic quantum number can have integral
values from -1 to +1. The value of mi defines the number of orbitals corresponding to
each of the energetic sub-levels s, p, d,f, ... Therefore, for the sub-level s, 1= 0 and
mi = 0, and there is only a single orbital. In the case of sub-level p, I = 1 and mi=-l,
0 and 1 and so there are three distinct orbitals p , p , p corresponding to this sub-
x y z

level. These have nodal planes {(y,z), (z,x) and (x,y)} oriented orthogonally to each
other (figure 2.2 (b)). In the nodal planes containing the nucleus, the electronic den
sity is zero. The wave functions change their sign when crossing a nodal plane.
To understand the physical significance of the magnetic quantum number, a
representation, taken from the classical Bohr-Sommerfeld theory, is often used. The
circulation of an electron in its orbit is thus likened to an electric current, which
flows and gives rise to an electric field.
The first three quantum numbers n, /, and mi characterise respectively the size,
shape and orientation of the orbital. It only remains to introduce the fourth quantum
number m related to the electron spin. In classical theory, this corresponds to the
s

rotation of the electron about itself, which also gives rise to a magnetic field. The
 Atoms, molecules and the chemical bond 21

quantum number m can have two values +1/2 and - 1 / 2 , associated with the spin
s

momentum of the electron.

The electrons are distributed in the atom starting in the lowest energy levels
(figure 2.3). Thus in the hydrogen atom, there is one electron in the I s orbital, which
is given the notation I s (table 2.4). Helium has two electrons in the same Is orbital,
1

for which the notation Is is used. These electrons must have opposite spins to re
2

spect Pauli's Principle, which states that two electrons cannot have four identical
quantum numbers. The electronic layer which corresponds to the principal quantum
number = 1 is therefore complete. This electronic structure confers great stability
on the helium atom, which behaves like a monatomic gas and has practically no
chemical reactivity.

s d f
10_ 9 - 8
9- 8 o - - 7
8~ 67 - 6
7- 6 5 5
6-
5 4
4
5-
4
3
4-
3

3-

2-

I -

Figure 2.3. The different energy levels of atomic orbitals. It should be noted that the
separation between the energy levels decreases as increases. From - 3 an overlapping
of levels can be observed and thus the 45 level has a lower energy than the Id level.

The chemical behaviour of the atom involves only the electrons in the outer
layer or that which is immediately underneath.
Table 2.4. describes the classification of the first elements of the Periodic Table
as a function of the filling of the electronic layers. When several orbitals of the same
energy level are available, electrons avoid being placed in the same orbital but ar
range themselves in such a way that their spins are parallel, in order to minimise the
interaction between electrons (Hund's Rule). This rule tends to favour the arrange
ment of electrons in different orbitals because two electrons in the same orbital must
have opposite spins.
This preferential arrangement of electrons is the origin of the magnetic proper
ties of a material. Any atom, ion or molecule with one or more unpaired electrons, is
22 Introduction to Materials Science

Table 2.4. Classification of the first elements of the periodic table as a

function of electron shell filling.

Elements Is 2s 2p 3s

He

Li 3

Be 4
tu
5 tu t
C 6 tin rrm 2
N 7
tin
8 tu HS [M
F 9 tu tu tfl
Ne
10
1 1
Na tui uul rful uul Pfu
paramagnetic. Substances that have all their electrons paired are diamagnetic. In a
certain number of cases, paramagnetic atoms or ions interact in a co-operative man
ner, giving rise to various more complex forms of magnetism, such as ferro- and fer-
rimagnetism and antiferromagnetism (chapter 15.).

2.3. Bonds
2.3.1. General points
The formation of polyatomic groupings known as molecules implies that the
atoms are able to combine to form aggregates, which have energies lower than that of
the separate atoms. Chemical bonding is an indispensable concept for explaining the
cohesion of material and it has a fundamental influence on material properties.
Early theories made distinctions between different types of bonds. In addition,
they proposed a difference between strong chemical bonds (greater than 40 kJ moF ) 1

and weak or secondary bonds (4 to 40 UmoF ). Today, all these distinctions have
1

been superseded by wave mechanics, which has led to a unified view of chemical
bonding.
It remains convenient, however, for technical reasons, to continue to make dis
tinctions between ionic, covalent, metallic and secondary bonds. Still, these models
are limiting cases and the more complex physical reality is intermediate between
them.
 Atoms, molecules and the chemical bond 23

The electrons, which participate in chemical bonding, belong in general to the

outer electronic layer of the atom and are most weakly bound to the nucleus. These
electrons are known as valency electrons. The bonding forces are essentially electro
static in nature and arise from a re-arrangement of the valency electrons around the
atomic nuclei involved in the bond.

2.3.2. Ionisation and electroaffinity

The peripheral electrons (valency electrons) are more or less bound to the atom.
The degree of bonding is function of the degree of filling of the outer electron shell.
The gain or loss of one or more electrons changes the neutral atom into an ion, which
is respectively negative (anion) or positive (cation).
The ease with which a cation can be formed depends on the Ionisation potential,
i.e. the energy required to remove one electron from the atom. This ionisation poten
tial is greatest for the rare gases, which can neither easily gain nor lose an electron
(table 2.5.). The ionisation potential is low for the alkaline metals (Li, Na, etc.).
These atoms have outer electron shells with a very low degree of filling, and have a
tendency to lose an electron easily to give a positive ion.
The ease of formation of an anion is in direct relation to its electroaffinity, i.e.
with the change in energy resulting from the capture of an electron. Thus the ele
ments close to the rare gases such as the halogens (F, CI, etc.), which have ionisation
potentials almost as large as those of the rare gases, possess a strong affinity for
electrons. In the case of fluorine, for example, the ionisation potential is equal to
17.4 eV and its electroaffinity is -3.6 eV. This means that atoms such as fluorine
have a tendency to capture an electron to form an anion.

2.3.3. Ionic bonding

The most stable electronic structure is that of the rare gases, which in general
have eight peripheral electrons, with the exception of helium which only has two
(table 2.4.); in practice the rare gases are chemically inert. During the formation of a
chemical bond, the atoms have a tendency to form an outer electron shell similar to
that of the rare gases by sharing electrons either by losing or gaining them.
Thus, when atoms with a low ionisation potential (electropositive) such as the
alkaline metals (Li, Na, K, etc.) are associated with the elements having a high elec
troaffinity (electronegative) such as the halogens (F, CI, Br, etc.) (table 2.5.), the
former tend to lose an electron and form a positive ion while the latter gain an elec
tron in order to complete their outer electron shell and form a negative ion.
Ionic bonding results from the association of positive and negative ions. During
the reaction between an alkaline metal and a halogen, the transfer of an electron from
the alkaline metal to the halogen leads to the formation of a cation and an anion:

(2.3.)

Anion
24 Introduction to Materials Science

Table 2.5. Ionisation Potential (loss of an electron) and Electroaffinity

(gain of an electron) for the first elements of the Periodic Table.

Elements Ionisation Potential Electroaffinity

[kJ moF ] 1
[kJ m o f ]1

1 1312 -73
He 2 2372 21
Li 3 520 -60
Be 4 900 -240
5 801 -29
C 6 1086 -113
7 1402 58
0 8 1314 -120
F 9 1681 -318
Ne 10 2081 29
Na 11 496 -52
Mg 12 737 230
Al 13 578 -48
Si 14 787 -134
15 1012 -75
S 16 1000 -205
CI 17 1251 -343
Ar 18 1521 35

In bold type, the values for the rare gases. The negative values correspond
to exothermic effects. (From HUHEE, 1978)

During the formation of ionically bonded compounds, an exchange of a certain

number of electrons occurs between the different elements, leading to an electrically
neutral assembly. The number of electrons transferred is directly related to the con
cept of valency.
To define precisely the notion of valency in ionic compounds, it is essential to
introduce first the notion of oxidation number. This important concept will also be
used in section 16.3. where corrosion phenomena will be analysed. The oxidation
number of an ion is equal to the number of electrons gained or lost during the for
mation of the ionic compound. Thus, an oxidation number can be either positive or
negative. An atom, which gives up one or more electrons, is oxidised and its oxida
tion number is positive. Should the opposite occur, the atom that acquires one or
more electrons is reduced and the oxidation number is negative. An example is the
formation of sodium chloride, NaCl. By losing an electron, the sodium atom obtains
the electron configuration of neon and forms a positive ion, Na . The chlorine, by
+

losing an electron, becomes a negative ion CI" and takes the electronic structure of
argon. These two steps take place simultaneously. By the exchange of an electron,
the two ions N a and CI" find themselves bound by an electrostatic bonding force. In
+

this process, the sodium atom, which gives up an electron, is oxidised and the chlo
rine atom is reduced. The oxidation number of sodium, which has lost an electron, is
+1 while that of the chlorine atom, which has gained an electron, is - 1 .
The valency of the elements is a very important concept as it determines the
chemical composition of substances. In the case of ionic compounds, the valency
(electrovalency) of the various elements depends on the number of electrons ex-
 Atoms, molecules and the chemical bond 25

changed to form an electrically neutral assembly of atoms. In the case of ionic sub
stances, the valency of an element is identical to its oxidation number. Thus, in so
dium chloride, the valency of sodium is equal to +1 and that of chlorine is - 1 . In
CaCl , the calcium atom loses two electrons and its valency is +2. The concept of
2

valency will be discussed later ( 2.3.4.) during the treatment of covalent bonding.
To form a stable bond, it is necessary to minimise the energy of the system that
is to say that the reaction must occur with an energy release. By examining the en
ergy balance (Born-Haber cycle) for the reaction between sodium in the solid state,
Na(s) and chlorine in the gas phase Cl (g), it can be shown that the formation of a
2

molecule of NaCl in the gas phase is an exothermic process:

Na(s) + 0.5 Cl (g) -> N a , Cl"(g) - 3.29 eV

2
+
(2.4.)
The sodium chloride molecules are, in effect, detected in the gas phase at high tem
perature.
At ambient temperature, it forms a crystalline solid consisting of sodium ions
alternating with chlorine ions. The crystal structure of solid NaCl, limited to only one
spatial dimension is shown in figure 2.6. In such an arrangement, isolated molecules
are not formed, but rows of alternating positive and negative ions appear. The build
ing block of such a substance is therefore the ion, that is to say the atom that has lost
or gained an electron. It should be noted that ionic substances are electrically neutral,
i.e. they consist of the same number of positive and negative ions when these have
the same valency.

Figure 2.6. Reticular row of ions in a NaCl crystal.

The energy of formation of the crystal of ionic NaCl from gaseous chlorine and
solid sodium can equally be calculated by making an energy balance.

Na(s) + 0.5 Cl (g) -> N a , C P (s) - 4.26 eV

2
+
(2.5.)
In comparing (2.4.) and (2.5.), it can be seen that the energy of formation of
crystalline NaCl is, as an absolute value, greater by 0.97 eV than that of gaseous
NaCl. This release of energy (reticular energy) derives from the electrostatic inter
action that occurs to a greater or lesser extent between all the anions or all the cations
of the crystal. This effect increases the thermodynamic stability of the system.
Ionic bonding is non-oriented because the field of attraction between the ions
has spherical symmetry. Ionic bonds are encountered in the alkaline halides (NaCl,
KCl, etc.) and alkaline earths (MgCl , CaCl , BaCl , etc.) as well as in the majority of
2 2 2

oxides (ceramics) such as MgO, CaO, etc.

2.3.4. Covalent bonding

Covalent bonding results from the sharing of one or more pairs of electrons by
two atoms of which the outer electron shell is incomplete. This pair becomes com
mon to the two atoms, which thus complete their outer electron shells and acquire a
rare gas configuration.
26 Introduction to Materials Science

In the case of single covalent bonding, the two atoms have a pair of electrons in
common. Thus the hydrogen molecule, H , is made up of two hydrogen atoms H:
2

(2.6.)
+ {^) C^^ represented by

A bar represents the single covalent bond. Starting from a carbon atom (four
electrons in the outer shell) and four hydrogen atoms, a methane molecule possessing
four single bonds is formed:

represented by C (2.7.)

Multiple bonds ocur when several pairs of electrons are shared. This is the case
with the carbon-carbon double bond of ethylene involving two pairs of electrons:

(2.8.)
/^^\ v ^ / v N r e
P r e s e n t e c
*
( j y i J

Valency in covalent bonding is equal to the number of electrons that an atom

shares with one or more other elements. Thus the hydrogen atom sharing one elec
tron with another hydrogen atom (2.6) or with a carbon atom (2.7) is monovalent. In
methane (2.7.), where it has four electrons in common with four hydrogen atoms, the
carbon is tetravalent, the same as in ethylene (2.8).
Covalent bonding, as ionic bonding, results from a re-distribution of electrons
reducing the total energy of the system. Ionic bonding arises from a transfer of elec
trons from an atom with a low ionisation potential to an atom with high electro-af
finity. In the case of covalent bonding with two identical atoms, there is no reason, a
priori, to have a preferential transfer of electrons since two identical atoms have the
same ionisation potential and the same electro-affinity. In this case, the bonding
results from a sharing of the peripheral electrons of the atoms, which move in orbi
tals (Molecular Orbitals - MO) common to the two atoms. The formation of this type
of bond can only be explained on the basis of the quantum theory. In particular, the
LCAO (Linear Combination of Atomic Orbitals) method, introduced in 1930 by
Mulliken, is applied. This method still remains widely used today. The molecular
orbitals are formed by the linear combination of atomic orbitals (AO).
In the example of the formation of the hydrogen molecule H from two atoms of 2

hydrogen, each hydrogen atom has only one stable orbital (Is), occupied by one
electron and capable of forming a covalent bond. When two atoms of hydrogen ap
proach each other, the two atomic orbitals combine to form a bonding molecular
orbital y/ , which is characterised by a lower energy than that of the two atomic orbi
h

tals (Is) and an antibonding molecular orbital , which has a higher energy level
 Atoms, molecules and the chemical bond 27

(figure 2.7.). The molecular orbitals, like the atomic orbitals, can accept at most two
electrons, so long as they are of opposite spins. The energy of the system decreases
when the two electrons occupy the lower energy molecular orbital \>. Under these
conditions, the electronic density of the molecular orbital > is greater than that of the
sum of the electronic densities of the individual atomic orbitals, considered sepa
rately. The electrons are shared between the two nuclei, and the force of attraction
between the two electrons and the two nuclei is greater than the force of repulsion
between the electrons and the nuclei themselves. This produces a bonding force and
the hydrogen molecule H is formed. If the electrons occupy an antibonding orbital,
2

the electron density between the nuclei is less than that of the individual atomic or
bitals. In these circumstances, a repulsive force occurs between the atoms when they
approach one another.

(b)

Figure 2.7. Schematic representation of the formation of the molecular orbitals of the
molecule H : (a) atomic orbital of two separate hydrogen atoms; (b) equations represent
2

ing the formation of antibonding (a) and bonding orbitals. Two arrows in opposite
&

directions on the bonding orbital \fc represent the two electrons with opposite spins re
sponsible for the bond.

It should be noted, as a general rule, that the electrons first occupy the molecular
orbitals of lowest energy and those that are bonding orbitals. It is only when the
bonding orbitals are filled that the antibonding orbitals are occupied with electrons.
Such is the case of the ion H e that has two electrons in the bonding orbital and only
2
+

one in the antibonding orbital. A single electron is responsible for the covalent bond
in the H e ion; therefore this bond is only half as strong as that in the hydrogen
2
+

molecule. It is easy to understand why the He molecule cannot exist. This molecule
2

would have two electrons in each of the bonding and antibonding orbitals and would
equal a bond of zero energy.
Therefore bonds become established by the overlapping of orbitals lying in the
same plane. So that two orbitals can combine, they must be of similar energy and
possess the same elements of symmetry. Indeed, this overlapping is more effective if
the orbitals are of comparable sizes and thus of similar energies. The overlapping of a
small orbital (e.g. Is) with a much larger orbital (such as 2p) does not lead to optimal
overlapping of the orbitals and does not produce a strong bond because of the more
diffuse distribution of the electrons in the larger orbital.
28 Introduction to Materials Science

By elementary reasoning, it can be shown that the interaction between an elec

tron e and two protons H i and H of two atomic nuclei produces an increase in the
(_) +
2
+

electronic density between the two protons. This results in the formation of a bonding
orbital b that produces bonding energy.
When the electron is in the zone between the nuclei defined by two hyperbolic
elements (shaded in grey tint in figure 2.8. (a)), the attractive force between the elec
tron and the two protons has two components, one in opposite directions and one
parallel to the nuclei axis. The resultant of these two forces works to pull the protons
closer. The higher the electron density between the nuclei, the stronger is the bond
between them.

Separation region

Bonding region Bonding region

(a) (b)

Figure 2.8. Representation of the interaction between an electron e and two protons Hi
(-) +

and H of two atomic nuclei. The electron is placed (a) inside and (b) outside of the field
2
+

of attraction created by the pair of protons. The components of the attractive forces be
tween the protons and the electron are opposed in the case where this is in the bonding
region.

On the other hand, when the electron is outside the surface shaded in grey
(figure 2.8. (b)), the resultant of the interaction forces between the electron and the
protons tends to separate the nuclei. This elementary reasoning shows that both
covalent and ionic bonding result from a redistribution of the electronic density,
which leads to a reduction of the energy of the system, as shown in figure 2.9. for the
H molecule formation.
2

Covalent bonding also occurs between two non-identical atoms. In this case the
electronic sharing is no longer symmetrical. Under these conditions, an electric di-
pole, which characterises the asymmetry of the electronic distribution, is observed.
This is a polar covalent bond, which is intermediate between a pure ionic bond and a
pure covalent bond.
A permanent electric dipole appears in a molecule when a transfer of electrons
from an atom or group of atoms in the molecule to another atom or group of atoms
takes place. This asymmetric sharing of electrons depends on the capacity of atoms
to attract electrons and is measured by their electronegativity. The higher the electro
negativity of an atom is, the greater the capacity of the atom to attract electrons.
Fluorine and chlorine are examples of atoms having high electro-negativity.
Electroaffinity measures the capacity of an atom to capture an electron to form an
anion ( 2.3.2.). It should not be confused with electronegativity determining the
attractive power of an atom for electrons in a covalent chemical bond. For further
 Atoms, molecules and the chemical bond 29

200

2 -200

-400

d
0 5 10
Radius (Borr)

Figure 2.9. Formation of the hydrogen molecule H2. The progression between (a) and (e)
represents the calculated sequence for the distribution of electronic density as a function
of inter-nuclear distance r. The variation in the energy as a function of r expressed in
Bohr (0.53-10~ m) is given in (f) (according to Wahl, 1970).
10

information on the concept of electronegativity, the work of Mercier and Godard, and
K. P. Vollhardt and N.E. Schore (section 4.2. in the references) can be cited.
A polar covalent bond is formed when a strongly electronegative atom (F, CI) is
bonded with a less electronegative atom such as carbon or hydrogen. Hydrogen chlo
ride and fluoride in the gas phase are examples of polar molecules:

+ -
(2.9.)

The electric dipole of a polar molecule is characterised by Direction and Mo

ment.
30 Introduction to Materials Science

= 1 (2.10.)
/ represents the distance separating the two electric charges: equal in absolute value
but with opposite signs. The electric charge <5 is expressed in electrostatic units (esu).
The electric moment is normally given in Debye (D), which is equal to 1.0 10" esu 18

cm. The electric moment of the hydrogen fluoride molecule in the gas phase is
1.98 D.
Geometric factors also play a significant role in establishing covalent bonds.
This is particularly important for orbitals having directional characteristics, such as
the orbitals. These give rise to two types of bond. In one case, the axes of symme
try of the atomic orbitals are aligned according to the inter-nuclear axis (figure
2.10 (a). In the other situation, the symmetry axes of the orbitals are oriented perpen
dicular to the inter-nuclear axis (figure 2.12. (b)). The first type of overlapping gives
rise to sigma () bonding, the second to pi () bonding.

abond

bond

Figure 2.10. Overlapping of atomic orbitals to form molecular orbitals: (a) overlapping
of two 2p orbitals aligned along the internuclear axis to form a abond as in the F mole
2

cule ; (b) overlapping of two 2p orbitals having their symmetry axis perpendicular to the
internuclear axis and forming a bond.

2.3.5. Hybridisation of the carbon atom

The structure of the carbon atom is given in the sixth line of table 2.4. The Is
and 2s layers are full and it has two unpaired electrons in two of the three 2p orbitals.
The structure is represented by the formula ls 2s 2p . The carbon atom, in its basic
2 2 2

state, has only two unpaired electrons and should normally be bivalent. The carbon
atom also exists in an excited state characterised by the promotion of an electron
from the 2s orbital into a vacant 2p orbital:

EH ED ED ED ED <2.n.)
This structure, represented by the formula \s 2s 2p , explains the tetra-valency
2 x

of the carbon atom. In this case, however, the four bonds cannot be identical because
they are formed from one s orbital and three orbitals. This is contrary to experience,
which indicates that four identical bonds are observed for tetravalent carbon.
 Atoms, molecules and the chemical bond 31

To explain the tetravalence of carbon and orientation of the bonds formed by

this element with other atoms, the concept of hybridisation of orbitals (another ap
plication of the LCAO {Linear Combination of Atomic Orbitals method), must be
invoked. A hybrid orbital results from the linear combination of two or more atomic
orbitals. This hybridisation of the carbon atom and the formation of the methane
molecule are described in figure 2.11 (a) and (b). To arrive at the tetrahedral structure
of carbon, encountered in a large number of organic compounds, the s orbital and the
three orbitals are combined to form four hybrid sp orbitals. These, with their 25 %
s character and 75 % character, are oriented towards the corners of a tetrahedron
(figure 2.11 (c)). In the methane molecule, C H , the s orbitals of the four hydrogen
4

atoms combine with the sp orbitals to give four covalent C bonds which are a
bonds enclosing an angle , of 109.5 . A realistic representation of the molecular
structure of methane (CH ) is given in figure 2.11(d).
4

2s 2p + 2p + 2p
x y z 4 orbitales 3

Figure 2.11. Representation of the formation of hybrid sp orbitals for the carbon atom C
3

and for the methane molecule, CH : (a) formation of four hybrid sp orbitals starting
4
3

from the 2s and 2p atomic orbitals of carbon; (b) formation of the methane molecule,
CH ; (c) representation of the methane molecule and calculation of the angle between
4

the H-C-H bonds; (d) atomic model.

32 Introduction to Materials Science

The carbon atom can also bind itself to another carbon atom as in the ethane
molecule, C H (figure 2.12. (a) and (b)). This has six carbon-hydrogen bonds and
2 6

one carbon-carbon a b o n d , as in diamond (figure 2.12. (c)), which is obtained by

replacing the four hydrogen atoms of methane with carbon atoms. Therefore dia
mond consists of the regular arrangement of a large number of carbon tetrahedrons,
covalently bound, which explains its very high hardness. On the atomic scale, the
diamond structure practically extends to infinity in three-dimensional space. It can
therefore be thought of as a giant macromolecule. Carbon atoms can also bind to
gether to form long chain hydrocarbons as in the polyethylene molecule (figure
2.13.).

(a) (b) ()
C

Figure 2.12. The ethane molecule, C H shown as (a) planar formula and (b) spatial
2 6

form, together with (c) the crystalline network of diamond .

Figure 2.13. Spatial presentation of the polyethylene molecule (C H ).

2 4

By hybridisation of the 2s and 2p orbitals, it is equally possible to obtain carbon-

carbon double bonds. Thus by combining a 2s orbital with two 2p orbitals, three
suborbitals are obtained which form between them an angle of 120 (figure 2.14. (a)).
This leads to the creation of three bonds with 33 % s character. Each carbon atom
in the ethylene (ethene) molecule is configured sp (figure 2.14. (b) and is bonded to
3

the other carbon atom and to two hydrogen atoms by three bonds. This leaves each
carbon atom with an unemployed 2p atomic orbital. These two orbitals combine with
z

each other to form a bond. The (x,y) plane, containing the nuclei of the two carbon
atoms and the four hydrogen atoms and passing through the symmetry axis ( C - C ) of
the molecule, is a nodal plane, i.e. a plane where the electronic density is zero.
There are twelve electrons in the ethylene molecule: eight electrons originate
from the two carbon atoms (sp ) and four electrons from the four hydrogen atoms.
2

Ten electrons form the five bonds of the C H molecule. The remaining two form a
2 4
 Atoms, molecules and the chemical bond 33

3 hybrids sp orbitals
1

Ethylene

Figure 2.14. Representation of the formation of hybrid sp orbitals for carbon and for the
2

ethylene molecule, C H : (a) formation of three planar hybrid sp orbitals starting from
2 4
1

one 2s and two 2p atomic orbitals; (b) formation of the ethylene molecule, C H with its
2 4

aand bonds; (c) representation of the ethylene molecule, C H .

2 4

bond (figure 2.14. (b)). The double bond therefore results from the superposing of a
a bond and a bond. The double bond is shorter than the single bond (134 pm
instead of 154 pm) because two pairs of electrons exert more restraint on the nuclei
than only one pair. Note that the a bonds of ethylene are oriented at 120 (figu
re 2.14. (c).
The bond has a energy of formation somewhat less than that of the a bond
because of a reduced overlapping of the atomic orbitals. This explains why the heat
of formation of the double bond (602 kJ m o ! ) is not equivalent to twice the heat of
-1
34 Introduction to Materials Science

formation of the single bond (2-346 kJ mol" ) (table 2.17.). The formation of two
1

single a bonds at the expense of a double bond ( + ) is therefore a process which is

liberating energy and favoured from a thermodynamic point of view. It is the process
taking place in the formation of linear poly-carbon chain compounds (polymers) such
as polyethylene:
Initiator
n CH =CH
2 2 ^ _CH -CH --CH --CH ~CH --
2 2 2 2 2 (2.12.)

Table 2.17 gives the energies, multiplicities and lengths of a number of covalent
bonds. Given values are averages. As shown very clearly by the study of the structure
of ethane (figure 2.12. (b)) and ethylene (figure 2.14. (c)) molecules, the covalent
bonds are directional, that is , they only form in the spatial directions predetermined
by the hybridisation of atomic orbitals. It is the mutually fixed orientations of the
covalent bonds which determines the valency angle.
The valence of carbon does not always have an integral value (1,2 or 3). This is
the case for benzene (C H ) for which two limiting formulae can be proposed (figure
6 6

2.15. (a)). The electronic structure of benzene (figure 2.15. (b) and (c)) is intermedi
ate between those two represented in figure 2.15. (a).
The case of benzene is a mesomeric phenomenon: the real state of the molecule,
called resonance hybrid, is intermediate between a certain number of limiting ficti
tious states (in the example considered, there are two shown in figure 2.15 (a)),
known as resonance forms or mesomers. The molecule is not a mixture of different
isomers separable one from another by classical organic chemistry techniques. A
mesomer has no real existence; it is only a form of writing. The ensemble of the
mesomeric forms describes the real structure and the behaviour of the molecule. The
mesomeric phenomenon plays an important role in organic chemistry. In fact, the use
of mesomeric forms is the result of our inability to transcribe the real structure of
more complicated molecules into a formula containing only single, double or triple
bonds.
The phenomenon of resonance is linked with a delocalisation of the electrons.
This is particularly the case for benzene. In this molecule (figure 2.15.), the six car
bon atoms in the ring are bound to each other by six a bonds resulting from the com
bination of six sp orbitals. Six orbitals are not involved in the formation of the a
1

bonds. They combine amongst themselves to form a polynuclear orbital in which

electrons circulate. The formation of this molecular orbital involves a significant
release of energy and stabilises strongly the benzene molecule with regard to its
limiting forms (figure 2.15. (a)). It is therefore possible to consider that, in the case of
the benzene molecule, two adjacent carbon atoms are bonded to each other by a
bond and a half bond.
A similar structure is found in graphite. It (figure 2.16.) is formed by the super
posing of planes where the carbon atoms are situated at the corners of regular hexa
gons. In these planes, the atoms are bound to each other by bonds onto which are
superposed the effects of electrons circulating in polynuclear orbitals. Each carbon
atom therefore has a valence equal to 1+ 1/3 ( l a b o n d + 1/3 bond). The structure of
graphite is strongly anisotropic. In the hexagonal planes, the C C distance is equal
to 0.142 nm, while the distance between two adjacent planes (0.335 nm) is much
greater. These are bound to each other by much weaker bonds resulting from
interactions between the electrons circulating in adjacent planes.
 Atoms, molecules and the chemical bond 35

I
FL

II

I

(a)

Electrons

I 120

120 120

109 pm I 139 pm
Electrons
(b)

Figure 2.15. Mesomeric structure of benzene: (a) two limiting formulae; (b) and (c) real
structure of the molecule with an indication of the circulation of the electrons.

(a)

Figure 2.16. (a) Crystallographic structure of graphite; (b) three limiting mesomeric for
mulae for three hexagonal aromatic rings of the graphitic plane.

This anisotropic structure of graphite leads to the considerable anisotropy of the

mechanical and physical properties. The lubricating properties of graphite result from
the easy cleavage of the crystal parallel to the hexagonal planes. The presence of
polynuclear orbitals, situated in the hexagonal planes and allowing the electrons to
circulate, explains the high electrical conductivity of graphite parallel to these planes.
The electrical properties of graphite are also strongly anisotropic, the conductivity
36 Introduction to Materials Science

being between 100 and 1000 times less in the direction perpendicular to the hexago
nal planes.
It is also because graphite contains mobile electrons, which interact with pho
tons, that it is an opaque solid, whereas diamond (section 4.4.), an electrical insulator
not containing mobile electrons, is a transparent solid.
This comparison between graphite and diamond shows how the nature of the
bonds exerts an overwhelming influence on the structure and properties of materials.

TABLE 2.17. Heat of formation, multiplicity and lengths of some

typical covalent bonds in organic chemistry.

Bond Bond multiplicity Length Energy

[pm] [kJ mof ] 1

CC 1 154 346
C=C 2 134 602
c=c 3 120 835
C 1 109 411
CF 1 135 485
CCI 1 177 327
CBr 1 194 285
C0 1 143 357
C=0 2 120 798
C 1 147 305
C 2 - 615
C=N 3 116 887
oo 1 166 139
1 96 459
1 103 391

(From HUHEEY, 1978)

2.3.6. Metallic bonds

Metals, which represent almost three-quarters of the elements (figure 1.2.), are
characterised by low ionisation energy and low electroaffinity. The valency elec
trons, which are few, are thus weakly bound to the nucleus of the metal atom. In
practice, the valency electrons are not bound to a specific atom but are delocalised
and spread throughout the metal.
The metallic bond is therefore characterised by a structure of positive ions in a
gas of electrons. This electron gas assures the bonding between the positive ions
(figure 2.18.). The metallic bond is strong, with identical (isotropic) behavior in all
directions. This type of bond favours the formation of simple crystalline structures,
highly symmetrical and compact. The presence of mobile electrons produces the high
electrical and thermal conductivity of metals.

2.3.7. Energy bands

As was shown in paragraph 2.3.4. for the hydrogen molecule, two atomic orbi
tals, belonging to two different atoms, are able to combine to form one bonding and
one anti-bonding molecular orbital. In the same way, three atomic orbitals belonging
to three different atoms can merge to form three orbitals with three nuclei: one
bonding, one non-bonding and one anti-bonding. So, by combining the atomic
 Atoms, molecules and the chemical bond 37

orbitals of atoms of a metallic crystal, polynuclear orbitals are obtained with

very close energy levels (figure 2.19.) When the number of atoms becomes very
large, the energy levels of the polynuclear orbitals form a more or less continuous
energy band. The formation of stable energy bands depends on the degree of super
position of the atomic orbitals in the solid.

Figure 2.18. Schematic representation of the electronic structure of a metal. (+) = posi
tive ions. () = electrons.

Orbital

antibonding / ~
antibonding
atomic ^/'non-bonding / '

bonding
bonding

1 centre 2 centre 3 centre centre

Figure 2.19. Diagram of energy bands in metals.

It is possible to generalise the theory of energy bands to the whole range of

solids: metals, insulators, and semi-conductors. Figure 2.20. shows the distribution of
electrons in energy bands when these are in their lowest energy states, i.e. practically
when 0 . By combining the atomic orbitals, an array of energy bands is ob
tained. The occupied band with the highest energy is the valency band. The bands of
lower energy are completely filled with electrons while the energy bands above the
valency band are vacant. It is at this level that the distinction is made between the
different types of solids: metallic, electrically insulating or semi-conducting.
In the case of metals (figure 2.20. (a)), the last occupied band is not completely
filled and the valency electrons remain delocalised because, whatever the tempera
ture, they can occupy the vacant orbitals which have an energy level slightly higher
than that of the last occupied orbital. A metal remains an electrical conductor even at
0 K ( 15.2.3).
For non-metallic materials (insulators and semi-conductors), the valency bands
are completely filled. The vacant band, which has an energy level immediately supe
rior to that of the valence band, is known as the conduction band as it is in this band
that the electrons responsible for the electrical conductivity of non-metallic materials
38 Introduction to Materials Science

Conduction band
/
t F Forbidden band(E )D

- Valence band

-m Subjacent band

(a) Metal (b) Electrical insulator (c) Semiconductor

Figure 2.20. Diagram of energy levels in the band theory and the distribution of electrons
in their fundamental state (-> 0 ), (a) Metals: last occupied band is partially filled and
the electrons are free to move under the influence of an electric field, even at 0 K. At a
given temperature, the conductivity of non-metallic materials is a function of the width
of the forbidden band (E ). lfE >2 eV, the solid is insulating (b), if E = 1 eV then at
D D D

high temperature it is a semi-conductor.

are found. The energy difference E between the last occupied band and the first
D

vacant band is called the forbidden band because no electron can have an energy
between that of the level above the valence band and that of the level below the con
duction band. In order that a non-metallic solid can be capable of conducting elec
tricity, it is necessary for the electrons to have sufficient energy to cross the forbid
den band and to jump from the last occupied band (valency band) to the first vacant
band (conduction band).
The non-metallic materials therefore conduct when the electrons move in the
conduction band and, also as a consequence of this, due to the mobility of the sites
left vacant (holes) by migration of electrons from the valency band.
The difference between an insulator and a semi-conductor results basically from
the width of the forbidden band. If this is large, (E >2 eV - figure 2.20(b)), then no
D

electron has enough thermal energy to pass from the valence band of to the conduc
tion band at ambient temperature and the solid behaves as an electrical insulator
(electrical conductivity ^ ^ ^ ) . 1

When the forbidden band is in the region of 1 eV, a small number of electrons
can be excited towards the conduction band if the temperature is sufficiently high and
a semi-conductor (figure 2.20 (c)) is produced. The electrical conductivity of semi
conductors varies from 10~ to Q ^ c n f . Silicium and germanium are both semi
9 2 1

conductors and have the same crystalline structure as that of diamond, which is an
insulator. What distinguishes them is only the width of the forbidden band: 0.7 and
1.1 eV for silicium and germanium respectively compared with - 5 . 4 eV for diamond.
At 0 and in the absence of impurities, these three solids, silicon, germanium and
diamond, are insulators because no electron possesses sufficient thermal energy to
pass into the conduction band. The properties of semi-conductors will be further
discussed in paragraph 15.2.3. The electrical conductivity of a semi-conductor, in
general , increases with temperature, while it decreases in the case of metals. This
difference is important.
Sometimes a simple change in crystalline structure is enough to modify ED and
to change the covalent or metallic character of a bond. Thus grey tin (diamond cubic
 Atoms, molecules and the chemical bond 39

structure with coordination index 4), the stable form of tin below 13.2 C is a semi
conductor. White tin (tetragonal; coordination index 8) stable above this temperature,
is a metal. Similarly, vanadium oxide, V 0 is an insulator at ambient temperature.
2

Above 66 C, vanadium oxide becomes an electrical conductor. This completely

reversible change in the conductivity of V 0 results from a minor change in the
2

crystalline structure. Above 66 C, the atoms of vanadium are equidistant and the
electrons are delocalised and able to conduct electricity. Below this temperature, the
vanadium atoms are regrouped in V-V shorter pairs and bonding is covalent. The
energy band model is also use to explain the electrical conductivity of graphite.
By subjecting a material to high pressures the electrical conductivity can be
changed. Certain substances, which, at normal pressure, are insulators, become con
ductors under very high pressure, since polynuclear orbitals are formed when the
atoms move closer to each other.

2.3.8. Secondary bonds

Strong bonds, such as the covalent bond, involve a very significant displacement
of the valency electrons. Besides these strong bonds, which have just been discussed,
weak bonds exist, involving smaller energies between 4 and 40 kJ- mol" . These are
1

therefore at least an order of magnitude weaker than those of covalent bonds. Weak
bond formation only induces a small change in the position of valency electrons. The
secondary bonding forces are mainly the result of electrostatic interactions between
electric dipoles.
Permanent dipoles have a tendency to align themselves relative to one another
and the bonding force, which results from this electrostatic interaction, is about 10
times smaller than that of the weakest covalent bond.

(2.13.)

Polar molecule Polar molecule

Introduction of permanent dipoles brings about a substantial rise in the melting

and boiling points. For this reason, chloromethane (CH C1 - bp = - 2 4 . 2 C) has a
3

boiling point considerably higher than that of methane ( C H - bp = - 1 6 4 C) or that

4

of ethane ( C H - C H - bp = -88.6 C).

3 3

Hydrogen bonding is the best example of a very strong dipole-dipole bond. It is

important not to confuse this linkage with the covalent bond between two atoms
forming H . Hydrogen bonding occurs only when hydrogen is covalently bound to a
2

small and strongly electronegative atom such as fluorine, oxygen or nitrogen. A well-
known example is the hydrogen bonding of water and the carboxylic acids. This
explains the abnormally high boiling point of these molecules. In figure 2.21. (d), a
three dimensional structure of water, stabilised by hydrogen bonds is given.
Dipoles are also able to modify the electron distribution of a weakly polar mole
cule and induce a dipole.

(2.14.)

Polarizable molecule Polar molecule

40 Introduction to Materials Science

This interaction leads to the formation of a secondary bond, which is weaker than the
bond formed through permanent dipoles.
The weakest secondary bonds are those responsible for the condensation and
solidification of apolar molecules (methane) at low temperatures and non-combined
atoms such as the rare gases.
The formation of this type of bond is explained by the existence of instantaneous
perturbations in the distribution of electrons around the nucleus leading to the
formation of instantaneous dipoles. These result from the movement of the electrons
around the nucleus. The cohesive forces resulting from the interactions between these
instantaneous dipoles fluctuate in phase:

Atomic core

(2.15.)

Alignment of
instantaneous dipoles

These attractions between instantaneous dipoles are widely known as van der Waats
forces. They induce bonding interactions that are some 100 times weaker than cova
lent bonds.
The mutual polarisation of the electron cloud is easier when the peripheral elec
trons are away from the nucleus. The bonding forces are practically proportional to
the atomic size. This explains why the melting and boiling points of the rare gases
and of simple molecules increase with the size of the atoms or molecules.
The secondary bonds have a determining influence on the physical and me
chanical properties of organic polymers, which are molecular solids. In polymer
materials, both strong covalent bonds, which form intermolecular links between
chain atoms, and secondary bonds responsible for cohesion between chains, are in
volved. This latter type of bonding is important for the mechanical properties. Thus
in the case of poly amides (figure 2.21 (b)), a very strong hydrogen bond provides
cohesion between chains and is responsible for the relatively high melting point ( T >
m

220 C in general) for this type of material. If the NH group is replaced by an oxygen
atom unable to promote a strong hydrogen bond, a polyester is obtained (chapter 5),
giving a macromolecule with a lower melting point, of the order of 200 C.

2.3.9. Characteristics of molecules

Molecules are aggregates made up of atoms interconnected by strong covalent
bonds that form intramolecular links. Molecules are tied to each other by relatively
weak bonds (intermolecular bonds) called secondary bonds. Molecular solids make
up an important group of materials containing the majority of organic polymers.
In the solid state, molecules are joined together by secondary bonds between 10
and 100 times weaker than intramolecular bonds . It follows that molecular solids, in
general, pass into the liquid state at relatively low temperatures (100 - 250 C) and
that the lightest molecules pass into the vapour phase without any change in mo
lecular structure.
 Atoms, molecules and the chemical bond 41

H
H- n , H
^9
x
Hydrogen bonding

(a) H
- 0
H
\

H*/" * " H 1

"^ ' - ^0
0


0

\ / \
CH 2 HC
2 C=0
HC 2 C=0"H-N
C=OnnH-N CH 2
/ \ /
H-N CH 2 HC 2
\ / \
CH 2 H C^
2 CH 2

H C^2 CH 2 2

<>
b
CH 2 HC
2 CH 2
/ \ /
HC 2 CH HC 2 2 x

CH / 2 HC
2 N-H
HC 2 N-H0=C
N-H"0=C CH 2

/
^ TT A
0=C CH 2 H C^ 2

CH 2 H C^ 2 CH 2

HC 3 CH 2 HC 2

Figure 2.21. Examples of hydrogen bonding (a) water molecules (H 0); (b) a polyamide 2

66 (Nylon) crystal reticular plane.

The majority of organic substances are soluble insolvents that are in general low
molecular weight organic molecules. Solvents are, for the most part, liquid at
ambient temperatures, and are capable of separating molecules to make them pass
into solution. As the molecules are, in the main, made with oriented covalent bonds,
they have a well-determined structure (chapter 5). The macromolecules (polymers)
( 4.4.1.) are commonly soluble in organic solvents and can also be melted, but their
molecular weight is too high to allow them to pass into the vapour phase.
Metals and the majority of mineral solids do not have a molecular structure and,
generally, only melt at high temperatures (above 500 C). This is a consequence of
their being made up of atoms or ions linked together by strong bonds.

2.4. Internal cohesive forces

2.4.1. Attraction and repulsion
All the inter-atomic forces, which bind atoms together, are essentially elec
trostatic. These bonding forces involve only peripheral electrons. The strength of the
42 Introduction to Materials Science

links varies considerably from one type of bond to another, from several kJ, to
several hundred kJ per mole.
To calculate the heat of formation of a bond, it is necessary not only to take ac
count of the attractive forces responsible for the bond formation, but also to consider
the repulsive forces arising when atoms come into contact (figure 2.22.). We restrict
ourselves here to the case of ionic bonding to demonstrate this key point. When the
distance r between two ions of opposite sign is large, these can be considered as
point charges between which only columbic attractive forces are involved. If r
decreases and reaches approximately the size of the atomic radius, the ions no longer
behave as point charges because the peripheral electrons carried by the ions repel
each other. There exists, therefore, for the ions, an equilibrium distance r 0

corresponding to the point where the attractive and repulsive forces counterbalance
each other. This position determines the mean length of the stable bond.

E=E +E A R

Interatomic distance r Interatomic distance r

(a) (b)

Figure 2.22. Variation of potential energy (a) and the central electrostatic attractive (Ea)
and repulsive (E ) forces (b) as a function of the distance r between two neighbouring
R

ions.

To understand the origin of this repulsive force, which is also electrostatic in

nature, it is necessary to make use of quantum mechanics because the interior of an
atom is practically empty. At first sight, it is difficult to imagine why the electron
trajectories of two neighbouring atoms cannot overlap. Electrons can only occupy a
certain number of well-defmed orbitals and each orbital only accepts up to two elec
trons (Pauli's principle). When two ions approach, the peripheral electrons of one ion
interfere with those of the neighbouring ion. To reduce the interatomic distance, it
would be necessary to force the peripheral electron shells to accept some electrons
coming from another atom, an impossible task because they already contain their
maximum number of electrons.

2.4.2. Calculation of ionic bond strength

Coulomb's Law gives the force f A acting between two point electric charges:

/,=- (2-16.)
JA a
r
 Atoms, molecules and the chemical bond 43

In this expression, X is a constant that depends on the electric charge of the ions
and therefore on the number of electrons involved in the bond (valency electrons); r
is the distance between the centres of the ions and the exponent a is equal to 2 (Cou
lomb's Law). A negative sign is conventionally given to the attractive f o r c e d , which
occurs between ions of opposite signs. The smaller the distance between the ions, the
greater the attractive force between them.
The repulsive force fR, only arises when the distance between atoms is very
small (of the order of 'atomic radiV. It can be calculated approximately using a
relation analogous to that of the law of attraction:

(2.17.)

As the repulsive effect is, over short distances, greater than the attractive effect,
the values of the exponent b must necessarily be greater than that of the exponent a.
For ionic bonds, the values are 7 < b < 10. The following relation gives the resultant
of the forces acting on the two neighbouring atoms:

(2.18.)
r r
The attraction constant X is simply calculated by Coulomb's law. This is not the
case for the repulsion constant Y. This can nevertheless be determined from the fact
that at the inter-atomic equilibrium, r , / = 0 and I/ l . / r :
0 A
=

Y=XC a)
(2.19.)

Substituting (2.19) in (2.18) gives:

"/ \{a-b)
f=L \ L \ _i (2-20.)
4UJ
Figure 2.22. describes the variation in the attractive and repulsive forces as well
as their resultant / as a function of the interatomic distance r. Any change in the in
teratomic distance r relative to r leads to the appearance of a force which tends to
0

restore the atoms to their equilibrium positions. For small variations in the inter
atomic distance, the force induced by a distortion of the bond is proportional to the
separation r - r, both in extension and compression. The various inter-atomic forces
0

can therefore be represented by more or less rigid springs.

The elementary treatment, which has just been developed for ionic bonds, is also
capable of being extended to other types of bonds. In a general way, it is possible to
say that inside solids, attractive and repulsive forces tend to balance out. The effec
tive action range of the repulsive forces is always smaller than that of the attractive
forces. For the cases of covalent and secondary bonds, the values of the exponent a
are between 6 and 10, and those of b between 10 and 12. Consequently for atoms in
solids, there always exist defined equilibrium positions at which the resultant of the
attractive and repulsive forces is zero.
In covalent solids, the bonds are in general oriented in well-defined directions
and the interactions between atoms not directly linked are weak. Covalent solids are
made up of molecules that are aggregates of atoms linked by strong covalent bonds.
The secondary bonds gives the cohesion between molecules. With the exception of
44 Introduction to Materials Science

three dimensional systems (cross-linkedpolymers), organic polymers are made up of

macromolecules, which are covalent chains joined to each other by secondary bonds.
For more information, readers are referred to section 4.4.

2.4.3. Size of the atom

This study of the attractive and repulsive forces leads to a simplified model for
the structure of the atom. The centre of the atom can be considered to consist of a
stiff impenetrable sphere made up of an atomic nucleus and inner completed layers of
electrons tightly bound to the nucleus and constituting the core of the atom. The
valency electrons (electrons in the incomplete outer layers) revolve around this rigid
core. It is only these peripheral electrons that interact with those of other atoms. Con
sequently it is possible to attribute relatively well-defined radii to the various atoms.
However, this dimension cannot be determined a priori and varies between an anion,
a proton or a neutral atom.
As an example, consider the case of the ionic bond, which has just been studied
in the preceding section. In condensed matter (solid or liquid), the distance between
two neighbouring positive and negative ions is measured by X-Ray diffraction.
However it is only by comparing the values obtained for different pairs of ions
( 17.2.1.) that it is possible to determine the ionic radius. For ions having the same
number of peripheral electrons ( 17.2.2.) the radius decreases as the nuclear charge
increases. In general, the radius of cations is significantly smaller than that of anions.
For ions carrying the same electrical charge, the radius varies relatively little when
the number of electrons and the nuclear charge increase simultaneously ( 17.2.2.).
The atomic radius depends on the nature of the bond in which the atom is in
volved, so that the radius of a metallic ion is always significantly greater than that of
an ion participating in an ionic bond. The ferrous ion, F e , for example, has a radius
2+

of 0.074 nm, while the metallic ion ( 17.2.3.) has a radius of 0.124 nm. In the metal,
the atom is really present in its ionic form. However, a major difference exists be
tween metals and ionic compounds. In ionic compounds, the ions of opposite sign
attract each other, while in metallic materials, the ions are of the same sign and so
repel each other. In a sense, this characteristic leads to an inflation of the metallic
structure and it is remarkable that, in this case, the hard sphere model still applies.
The atomic radius of metals hardly varies as a function of the number of electrons, Z.
Thus lithium (Z = 3) has exactly the same atomic radius as uranium (Z = 92). This
justifies the fact that the density of metals increases considerably with the atomic
number, Z. The atomic radius also varies according to whether covalent or ionic
bonds are involved. For example, the atomic radius of covalent fluorine is estimated
as 0.071 nm while the F" ion has a radius of 0.135nm.
The atomic radius also varies with the multiplicity of the bond. In the case of
carbon, for example, the CC distance is equal to 0.154 nm for a single bond,
0.134 nm for a double bond and 0.120 nm for a triple bond. As might be supposed,
the CC bonds in benzene (0.139 nm) and in graphite (0.142 nm) fall between the
values found for single and double bonds.
In the case of two atoms joined by secondary valency forces, the distance at
which the repulsive and attractive forces, induced by the peripheral electrons, are in
equilibrium defines the van der Waals' radius of the atoms ( 17.2.4). The calcula
tion of the van der Waals' radius is analogous to that for the ionic radii described
 Atoms, molecules and the chemical bond 45

previously. As the attractive forces are much weaker in the case of secondary bonds,
the repulsive forces come into play quicker and the van der Waals' radii are noticea
bly greater than the atomic radii in covalent bonds. Thus the van der Waals' radius of
chlorine is equal to 0.178 nm, while its atomic radius in covalent compounds is only
0.099 nm.
The atomic radius in covalent, ionic or metallic compounds must not be con
fused with the van der Waals' radius fixing the limit between molecules in molecular
solids such as organic polymers.

2.5. Summary and conclusions

The atom is built up from elementary particles (nucleons and electrons) and
obeys the laws of quantum mechanics. Accordingly, it is not possible to know pre
cisely the position of electrons which are confined within defined spatial regions
known as orbitals.
Four quantum numbers (, I, mi, m ) are required to characterise the electron
s

circulating within a defined orbital. This contains, at the m o s t , two electrons with
values of m = 1/2. Progressive filling of the orbitals produces the classification of
s

the elements in the Periodic Table. The chemical properties of the elements are only
a function of their peripheral electrons (valency electrons).
Basically, chemical bonds are electrostatic in nature. A stable bond forms only if
energy is released simultaneously. Quantum mechanics leads to a unitary view of
chemical bonding, but for didactic reasons, three types of strong bonds can be identi
fied: ionic, covalent and metallic bonds.
ionic bond has an isotropic character and forms between ions of opposite
signs by the transfer of valence electrons. In the covalent bond, there is a sharing of
electrons between neighbouring atoms, and in general, it can only be established in
defined spatial directions. In the case of the metallic bond, certain peripheral
electrons are not attached to a defined atom but delocalised in the solid. They form
an electron cloud, which assures the cohesion between metallic ions. Metallic
bonding also has an isotropic character, like the ionic bond, but it is not polarised.
Molecules are aggregates of atoms joined by strong covalent bonds. In the solid
state, secondary bonds resulting from electrostatic attraction between induced and
permanent dipoles link molecules to each other.
It is the nature of bonds, which mainly determines the properties of condensed
phases (solid or liquid) and the classification of materials can be made according to
the types of bonds. Thus the ductility of metals mostly results from the non-polar
ised, isotropic character of the metallic bond. The fragility of a certain number of
organic and ceramic materials can be related to the oriented nature of covalent bonds
which makes plastic deformation of these materials difficult at ambient temperature.
The brittle character of ionic materials (another type of ceramic) is equally induced
by the presence of ions of opposite signs. At low temperature, the ions inhibit the
mechanism of plastic deformation by dislocations (chapter 7).
As is shown in chapter 14, a certain number of physical properties such as elec
trical conductivity, transparency and reflectivity of (non-oxidised) surfaces are di
rectly linked to electron mobility.
46 Introduction to Materials Science

The properties of molecular solids such as polymers are mainly determined by

the weakness of their secondary bonds, which are responsible for the cohesion be
tween the chains. However, by orienting the chains by drawing ( 12.3.3), polymer
materials (fibres) are obtained and have mechanical properties comparable with those
of high performance steels.

2.6. Illustrative example: carbon fibre

The first known use of carbon fibre dates from the end of the nineteenth century
when Thomas Edison used a fine rod of carbonised bamboo as a filament for the first
incandescent lamp. Since then, carbon fibre has become the reinforcement used in
some recent sophisticated composite materials, notably for applications in aeronauti
cal and sporting goods.
Carbon fibre mostly consists of graphite that is the stable crystalline and low-
density form of carbon ( p = 2.25 g e m ) . As shown in figure 2.16, graphite is char
-3

acterised by a highly anisotropic crystalline structure. In the direction parallel to the

hexagonal planes, graphite possesses mechanical properties as high as those of dia
mond. To produce high strength fibres, it is therefore necessary to align the hexago
nal graphitic planes parallel to the direction of stress since in the axis perpendicular
to these planes, the rigidity of graphite is about 30 times less. This preferential
alignment is achieved by starting from an oriented precursor.
One of the most commonly used precursors is polyacrylonitrile (PAN) fibre,
which is pyrolysed in a controlled manner under mechanical constraint. At tempera
tures close to 200 C, a stable ladder structure is formed (figure 2.23). If this pyroly-
sis takes place in a partially oxidising atmosphere, a number of ketone (CO) groups
are generated (figure 2.23. (c)).
Elimination of water molecules and cyanic acid (HCN) takes place by heating
the cyclised polymer at high temperature (1500-2000 C) in an inert atmosphere, and
the oxygen, nitrogen and hydrogen atoms present in the previously oxidised fibre are
removed. By this treatment, an oriented graphitic structure is obtained: carbon fibre.
An overview of the process for the manufacture of fibre from PAN precursors is
illustrated in figure 2.24. Under these conditions, fibres which have a rigidity be
tween one fifth and one half of that of diamond or perfectly oriented graphite are
synthesized, while the rigidity of the non-oriented precursor (polyacrylonitrile) is 100
times less. Fibres with a very high modulus are obtained by carrying out the final
thermal treatment at a temperature at which the plastic deformation of carbon is pos
sible. Thus if the final treatment is made at a temperature of 2700 C, followed by an
elongation of 30 %, a carbon fibre is obtained with a elastic modulus (E = - 7 0 0 GPa)
greater than that of steel ( = 210 GPa).
Carbon fibres production with a polyacrylonitrile fibres precursor is a relatively
expensive process. Currently there is a move towards processes using pitch (a residue
of oil and coal distillation - a cheaper raw material). When molten pitch is treated at
temperature above 350 C, it acquires the mesomorphic structure of a liquid crystal.
When pitch is spun through a multi-hole spinneret die, it forms an oriented multi
filament fibre constituting another precursor for carbon fibre.
 Atoms, molecules and the chemical bond 47

(a)
^CH /CH ^CH /CH CH X

^CH CH CH2 ^CH X

CH 2
V
2 2
V
2

\ ..
J
.
Cy<
..
^c^ ^c* ^c^ ^c^ ^
(b) I I I I I
^CH CH /CH
X ^CH /CH /
^CH CH
2 CH N
CH 2

2
X
2
X
2

jo 2

^C^ ^C^ ^ c ^ ^c^ ^

(0 I I I I I
CH CH CH ,CH /
N
C "CH "C 2 "CH C 2
X

II II II

Figure 2.23. Chemical transformations of poly(acrylonitrile) fibre before its conversion

to carbon fibre: (a) drawn fibre and, as shown by the arrows in figure 2.23 (a), cyclisa-
tion (b) of the drawn poly(acrylonitrile) fibre with formation of a CN bond; (c) partial
oxidation of the cyclised drawn fibre (from Watt, 1970 and Kelly, 1986).

Figure 2.24. Schematic representation of the production of carbon fibre starting from a
polymeric precursor: (a) unstreched polymer chains (b) polymer fibre (c) graphitisation
under tensional stress (from Watt, 1970 and Kelly, 1973).

To then form the carbon fibre, a controlled thermal treatment is applied to the
oriented precursor fibre, which is first oxidised at a temperature below its softening
temperature to make it non-meltable. The oxidised fibre is then graphitised at tem
peratures between 1500 and 3000 C. A treatment at 2000 C gives a carbon fibre
having an elastic modulus of 380 GPa and a tensile strength of 2 GPa.
The microstructure of carbon (or graphite) fibres is described in figure 2.25.
This microstructure, which is made up of graphitic hexagonal planes aligned parallel
to the fiber axis and wrapped around pores, contains numerous defects.
48 Introduction to Materials Science

As shown in figure 2.25, the degree of organisation of the outside part of the
fibre is significantly greater than that of the centre. Carbon fibre is a microporous
solid. Using certain specific treatments, it is possible to produce fibres having an
important proportion of open pore structure (-0.7 c m g ) with a mean pore diameter
3 _1

of about 70 nm.

Axis of
the fibre

Figure 2.25. Schematic cross section of carbon fibre based on electron microscopy and
X-rays. It should be noted that the graphitic planes are generally aligned parallel to the
direction of the fibre, but with an important number of defects. (From S.C. Bennett,
1976; Hull, 1981; D.J. Johnson, 1982, 1987).

In general, two main types of commercialised carbon fibre exist:

High modulus fibres (E = - 5 0 0 GPa; R = -2300 MPa).
m

High strength fibres (E = - 2 3 0 GPa; R = -4000 MPa).

m

It is above all the thermal treatment at high temperature, which favours the for
mation of the longest crystallites with the highest elastic modulus and the smallest
strength R . Carbon (or graphite) fibres are used in a relatively large number of
m

industrial applications.

2.7. Exercises
2.7.1. Distinguish between atomic number and atomic weight.
2.7.2. Knowing that Avogadro's number N is equal to 6.02 1 0 atoms mol" , calcu
A
23 1

late:
The mass in grams of a copper atom;
The number of atoms in a gram and in a cm of copper. 3

2.7.3. Give the formula of filling of the electronic layers in carbon and iron atoms.
2.7.4. In which state is a metal atom in a crystal?
2.7.5. Explain why secondary bonds are weaker than primary bonds.
 Atoms, molecules and the chemical bond 49

2.7.6. What polymer chain is produced by the vinyl chloride polymerisation?

2.7.7. How does the potential energy of a system evolve as the distance between two
ions of opposite charge tends to its equilibrium value?

2.7.8. Alumina ( A 1 0 - a ceramic) has a density p = 3.7 g cm" .

2 3
3

Calculate the number of atoms per gram of alumina.

Calculate the number of atoms per cm . 3

2.7.9. Describe the electronic structure and the type of and s hybridisation of the
carbon atoms and the heteroatoms of the vinyl chloride (chloroethene), propylene
(propene) and ethane molecules having the following structures:

H2P=CHCI H2C=Cfr-CH 3 H3C CH 3

Vinyl chloride Propylene (propene) Ethane

2.7.10. Give the valency of the metals and non-metals in the following compounds:
CaCl , TiCl , A 1 0 , S i 0 .
2 3 2 3 2

2.8. References and complementary reading

P.W. ATKINS, Physical Chemistry, 2nd ed., Oxford University Press, Oxford, 1982.
S.C. BENNET, Ph.D. Thesis, Leeds University, 1976.
J. K. BURDETT, Chemical Bonds: A Dialog, Wiley, New York, 1997.
A.L. COMPAGNION, Chemical Bonding, Mc Graw-Hill, New York, 1964.
I. N. ERMOLENKO, I. P. LYUBLINER, . V. GULKO, Chemically Modified Carbon Fibers, VCH,
Weinheim (Germany), 1990.
J.E. HUHEEY, Inorganic Chemistry, 2nd ed., Harper, New York, 1978, p. 842.
D. HULL, An Introduction to Composite Materials, Cambridge Solid State Series, Cambridge University
Press, Cambridge, 1981.
D. J. JOHNSON, Chem. Ind. 18, 692, (1982).
D. J. JOHNSON, J. Phys. D: Appl. Phys. 20, 286, (1987).
A. KELLY and N. H. MACMILLAN, Strong Solids, 3rd ed., Clarendon Press, Oxford, 1986.
C. KITTEL, Introduction to Solid State Physics, 7th ed., John Wiley, New York, 1996.
B.H. MAHAN, College Chemistry, 2nd ed., Addison Wesley, Reading Mass., 1966.
J. P. MERCIER et P. GODARD, Chimie Organique. One initiation, 2eme ed., Presses polytechniques et
universitaires romandes, Lausanne, 2001.
P. ROBERT, Materiaux de VElectronique, Traite d'Electricite, Vol. II, Presses polytechniques romandes,
Lausanne, 1979.
J.C. SLATER, Electronic Energy Bands Metals, Phys. Rev., 45 (1934) 794.
K.P. VOLLHARDT and N.E. SCHORE, Organic Chemistry, 2nd ed., Freeman, New York, 1994.
A.C. WAHL, Chemistry by Computer, Scientific American, 222 (1970) 54.
W. WATT, Proc. Roy. Soc, A 319 (1970).
B. L. YAKOBSON, R. E. SMALLEY, American Scientist, 85, 324,1997.
 Chapter 3.

Structure of solids

3.1. Objectives
To distinguish between crystal and glass.
To characterise the various types of atomic arrangements in crystal
space lattices.
To introduce the technique of X-ray diffraction, used to determine the
structure of materials.
The classification of the more or less ordered arrangement of atoms or mole
cules in the solid state will be briefly analysed and the concept of a crystalline struc
ture defined. The ordered state, which characterises a crystal, is reproduced by the
periodic repetition of a structural unit. To link the crystal structure to the behaviour
of materials, it is necessary to introduce a notation for the crystal planes and direc
tions.
The crystalline state is found in the majority of metals and in a significant num
ber of ceramics and organic polymers. The amorphous state on the other hand is very
common in an important number of ceramics (mineral glasses) and organic materials
(organic glasses, rubber).

3.2. Physical states of materials

3.2.1. Order and Disorder
All substances can exist in three physical states: solid, liquid or gas. Other states
such as plasma will not be dealt with here. It is the balance between cohesive energy
(attracting atoms) and thermal energy (tending to separate them), which determines
the physical state of a substance.
The thermal energy , which results from the continual movement of the at
tn

oms, is proportional to the absolute temperature ().

E = kT
th (3.1.)
In this expression, k (1.381 1 0 JK" ), is the Boltzmann constant. This is the
-23 1

quotient of the gas constant, R , divided by the Avogadro number A.

k = RINA (3.2.)

The cohesive energy can be defined as the energy, which would be necessary to
dissociate the system into its basic constituents, i.e. to vaporise the substance. The
cohesive energy is considerably higher for materials with strong bonds (metals, ce-

2002 Editions scientifiques et medicales Elsevier. All rights reserved-

Introduction to Materials Science

52 Introduction to Materials Science

ramies) than for atoms (rare gases) or molecules ( F , Cl , etc., organic molecules) 2 2

tied to each other by weak bonds.

The thermal energy is proportional to the absolute temperature (3.1.), while to
a first approximation cohesive energy is independent of temperature. This explains
the changes of material structures and states with temperature. Figure 3.1 sketches
this evolution. When the thermal energy is high relative to the cohesive energy (high
temperature), all structured or ordered states of the atoms are excluded.

Distance estimated inatomic radius rrj

0 3-10 r 0 100 r 0
CO

Gas Perfect V Real

Gas \ y Gas
State of the matter

Liquid Liquides.

a
Solid Amorphous Real Perfect
Solid Crystal crystal

Complete Short Long Perfect

disorder range range order
order order

Mineral Metals Semi-conductors

Materials: -s Glasses and Ceramics

Organic Polymers

Figure 3.1. Evolution of atomic organization with increasing temperature.

Materials are disordered in the gaseous state, of which the perfect gases repre
sent the limiting form. Neon, at normal pressure (1 atm) and at ambient temperature,
provides an example of a gas close to the ideal state. Under these conditions, each
neon atom occupies a volume equivalent to that of a cube of sides 3.3 nm. In the case
of perfect gases, it is possible to neglect interactions between atoms and molecules.
The physical properties (compressibility, thermal conductivity, etc.) therefore depend
uniquely on the thermal (kinetic) energy of the atoms or molecules. In a perfect gas,
the position of an atom or molecule does not depend on that of another atom or
molecule taken as a reference.
When the thermal energy is low, the motion of the atoms or molecules is slow.
The constituent elements of the material tend to arrange themselves according to a
rigorous structure minimizing the potential energy of the system. The substance is
then in the crystalline state which is the orderedform of the solid.
The liquid constitutes an intermediate state between the gas and the crystal. If a
gas is compressed at a temperature below its critical temperature (T ), then at a given c
 Structure of solids 53

pressure, the atoms or molecules organize themselves into much denser regions.
During this transformation, called condensation, the atoms or molecules approach
one another leading to the transformation of the gas into a condensed state: the liquid.
This change of state, condensation is accompanied by an increase in order, a decrease
of entropy and a release of (latent) heat. For the basic notions of the thermodynamics
used in this treatment, the reader is referred to specialized works such as those of
Smith or Ragona.
The volume of a liquid is generally greater by several percent than that of the
crystal. When the structure of liquids is investigated by X-ray diffraction (sec
tion 3.4.), the nearest neighbours of an atom or a molecule are found to occupy a
position close to the ones they occupy in the crystal. On the other hand, the second
and third neighbours become much less localized. Beyond the third neighbours, the
order is highly perturbed. In other words, liquids are characterised by short-range
order, while maintaining a density close to that of crystal characterised by long range
order. Melting is accompanied by the appearance of voids with a maximum size of
the order of an atom. It is the presence of these voids that is responsible for the par
tially disordered (amorphous) structure of liquids.
The concentration of voids increases considerably with temperature and can
reach up to 10 % of the total volume of the substance. As it will be seen in chapter 7,
voids also exist in real crystals. In this case, the voids are called vacancies and have a
dimension close to that of an atom. The free volume in crystals is much lower (1000
to 10 000 times) than that in liquids. The voids present in liquids have high mobility
and the atoms or molecules are in continuous, erratic movement due to micro-
brownian motion. The voids allow liquid to flow under the influence of relatively
weak forces, characteristic of their low viscosities. At a given temperature, the free
volume of a liquid in equilibrium is constant.
It is possible to distinguish between the ordered state of a crystal and the short
range ordered structure of a liquid by means of the following experiment. Pour into a
container, in one go, about a hundred small balls (approximately 3 mm diameter)
representing atoms or molecules. This assembly of balls is arranged in a disorganised
way in the container and its structure (figure 3.2.) is similar to that of liquids. By
gently shaking the balls, simulating thermal agitation, the progressive formation of
regularly arranged zones of balls, simulating crystal domains, is observed. If the balls
are placed in a container with a hexagonal base and internal dimensions which are
multiples of the ball diameter, then after gentle, repeated shaking, the balls arrange
themselves in a single organised zone representing the structure of a monocrystaV or
4

'single crystal'.

3.2.2. Structure of Solids

On the atomic scale, in a perfect or ideal crystal the regular arrangement of
atoms extends practically to infinity. A silicon monocrystal used in the manufacture
of integrated circuits attains a length of 1 to 2 m with a diameter of 0.2 to 0.3 m. In
such a crystal, which has almost the structure of a perfect crystal, the atoms are situ
ated at predetermined sites at distances extending over 1 billion atomic diameters. A
crystal is characterised by a regular and periodic stacking of atoms bound to one
another by cohesive forces. While still maintaining its regular and periodic structure,
54 Introduction to Materials Science

Figure 3.2. Transition between the compact, long range ordered state of atoms in a crys
talline solid and the disordered state of a gas passing via the short-range order of the liq
uid.

a real crystal such as exists in the majority of materials, contains a large number of
imperfections (chapter 7.).
Two examples of different crystalline structures for carbon have already been
described: diamond (figure 2.12. (c)) and graphite (figure 2.16.). Diamond and
graphite are characterised by very different mechanical and physical properties.
Diamond is the hardest known material and its principal industrial application lies in
the area of abrasives and cutting tools, while graphite, which cleaves easily due to its
different bonding, is a solid lubricant. This example demonstrates the direct relation
ship existing between crystalline structure, chemical bonding and the macroscopic
behaviour of materials.
Crystalline materials exist in mono crystalline or poly crystalline forms
(figure 3.3.). The degree of perfection of monocrystalline materials results in general
from the use of sophisticated crystallisation techniques. However, the existence of a
monocrystalline structure does not necessarily constitute an advantage for many
applications. Metallic materials, as well as many ceramics, have a polycrystalline
structure made of a multitude of microscopic size crystals (grains) linked to each
other by boundaries representing less ordered zones (grain boundaries). Each grain of
a metal is a monocrystal. The grain structure also exists in the case of semi-
crystalline polymers. In this case, each grain is made up of a spherulite built up from
an assembly of crystalline platelets and an amorphous phase. This topic is treated in
more detail in chapter 10.
By changing the microstructure (chapter 1.), it is possible to modify the behav
iour of materials considerably. Thus, for example, a reduction in grain size leads to
an increase in the mechanical strength of the material at ambient temperature.
 Structure of solids 55

(a) (b)

Figure 3.3. Schematic view of crystals. The crystalline structure is represented by a grid
system: (a) single crystal (monocrystal); (b) polycrystal with, in heavy black lines, the
grain boundaries.

The long-range order is absent in some materials. This is the case for mineral
glasses, the majority of organic polymers and metallic glasses. These are glassy sub
stances, which have an amorphous structure, analogous to that of liquids.
The glassy state (amorphous solid) is an unstable form of matter, which tends to
crystallise when the thermodynamic and kinetic conditions allow (chapters 8. and 9.).
As indicated at the beginning of this chapter, nearly all substances are capable of
existing in four states: crystalline or amorphous solid, liquid or gas. However, the
macromolecules do not exist in gaseous form. The thermal energy that would be
required to vaporise them would take them at a very high temperature, the covalent
bonds between the atoms in the chain would be broken and the macromolecule
structure would be destroyed. Compared with metals, organic polymers are sub
stances which crystallise with difficulty. Certain organic polymers have a molecular
structure, making it impossible for them to crystallize. Cross-linked polymers (three-
dimensional systems) provide an example of macromolecules of almost infinite size
unable to pass into the liquid state (chapter 5.). The formation of monocrystals is
quite exceptional for polymers. In the solid state, polymers exist as a semi-crystalline
grain structure (chapter 10.) or as a completely amorphous form (organic glass,
rubber) (chapter 5.). In the classification of solids, rubber occupies a special position
because of its great ability to deform reversibly. The behaviour of rubber will be
covered in detail in chapter 6.
Ceramics lie between metals and polymers in their tendency to crystallise. It is
possible to meet with crystalline ceramics, such as certain oxides (MgO), but other
ceramics have a glassy structure. These mineral glasses are obtained by fusing mix
tures of oxides ( S i 0 + N a 0 + CaO). Nowadays, the term glass characterises the
2 2

structure of a large number of amorphous solids. Thus, not only are mineral glasses
(window glass) and organic glasses (PMMA or Plexiglas) produced, but also me
tallic glasses as will be seen in the illustrative example in chapter 4. It should be
noted that not all glasses are necessarily transparent (chapter 15.).
56 Introduction to Materials Science

3.3. Description of the crystalline state

3.3.1. Crystal Structure

The crystal is the commonest state of matter for materials. Crystallography de
scribes the architecture of crystals, i.e. the three-dimensional distribution of atoms
and the geometric relationship determining their positions. Crystalline solids are
characterised by atoms or molecules arranged periodically in the three spatial direc
tions in a strict sequence defining the crystal structure. A crystal structure is
characterised by its symmetry.
The fundamental concepts, describing the crystal structure, are the space lattice,
the pattern, the structural unit or motif and the unit cell.

A space lattice is made up of a three-dimensional array of points

(nodes), of infinite size, obtained by the displacement in space of three
non-coplanar vectors, a, b and c. They determine the directions and the
distances between the nodes in the lattice.
The lattice is formed by the juxtaposition of identical parallelepiped
unit cells of sides a, b and c. The unit cell therefore has a simple geo
metric form (cube, parallelepiped). The absolute values a, b and c of the
three vectors, which form the sides of the unit cell, are known as the
unit cell parameters. The unit cell contains the structural unit which
repeats itself forming the crystal.
The structural unit or motif makes up the central building block, the
repetition of which in the spatial lattice creates the crystal. It retains the
overall symmetry characteristic of the whole lattice.

Any plane passing through three non-collinear points of a space lattice is a

lattice plane. It contains a double infinity of points. All straight lines passing through
two lattice points contain a single infinity of points and constitute a lattice row.
With two non-collinear vectors a and b having the same origin, it is possible to
construct a two-dimensional lattice by displacing these vectors in the plane which
they form. As shown in figure 3.4, the choice of the vectors a and b is not unique. As
in the space lattice, the two-dimensional lattice can equally be formed by the juxta
position of identical unit cells in the form of a parallelogram of side a and b.
Figure 3.6. (c) also shows different ways of designing the unit cell in 2D. The
unit cell ABCD is known as a primitive cell as it only has points at the cell corners
and contains only one motif. The two other cells shown (EFGH and IJKG) are multi
ple cells as they contain points at the interior and on the edges of the cell. If the
primitive cell contains all the elements of symmetry of the lattice, it is chosen as the
unit cell to represent the lattice. If not, the smallest multiple cell exhibiting all the
elements of symmetry of the lattice is chosen. In the case of the crystal plane of fig
ure 3.6., the EFGH cell is chosen as the unit cell.
Figure 3.5 represents the space lattice of sodium chloride. The face-centred
cubic (fee) unit cell is placed in front of the sodium space lattice. On the whole, this
unit cell contains four sodium cations and four chloride anions, taking into account
the sharing of ions between the neighbouring unit cells. The dimensions of the unit
 Structure of solids 57

Figure 3.4. Two-dimensional lattice starting from two non-collinear vectors a and b hav
ing the same origin. The vectors a', b' and a", b" are equally the basis of the same two-
dimensional lattice. The unit cell (1) and (2) also describe the two-dimensional lattice.

Figure 3.5. Representation of the cubic lattice of the NaCl crystal. In front, the unit cell
delimiting the structural unit.

cell are of the order of 0,3 nm, which means that there are more than three million
unit cells in one mm.
Figure 3.6. gives a two-dimensional description of the crystalline structure of
NaCl. The plane considered here passes through one of the faces of the cubic space
lattice. In figure 3.6. (a), the crystalline structure is represented by use of the compact
sphere model the structural unit (or motif) containing two Cl~ and two N a ions is +

represented by the square ABCD. In figure 3.6. (b), the two-dimensional lattice,
corresponding to the crystalline structure of figure 3.6. (a), is represented. This lattice
has been built up by placing the lattice points at the centre of the chloride ions. It
would have been equally possible to place the lattice points at the centre of the
sodium ions or alternatively at a position equidistant from the centres of the chloride
and sodium ions, etc. as there are an infinite number of ways of choosing the origin of
the lattice without it being modified.
58 Introduction to Materials Science

(a)

Structural unit
o Lattice point

Figure 3.6. Two-dimensional description of the crystalline structure of NaCl: (a) lattice
plane of the NaCl crystal; the structural unit (or motif) is represented by a light grey
square; (b) lattice representing the crystalline structure with different unit cells. The cell
EFGH is generally used as it contains all the symmetry elements of the NaCl crystal,
which is face-centred cubic.

There are only fourteen ways to distribute points regularly in space (figure 3.7.).
These constitute the fourteen spatial lattice systems (Bravais' systems). They are
distributed in seven crystalline systems, which differ from one another by the lattice
symmetry. These are the cubic, tetragonal, orthorhombic, monoclinic, triclinic,
trigonal (or rhombohedral) and hexagonal systems. To obtain the fourteen Bravais'
systems, it is necessary to take account that some different lattices have the same
symmetry and to add the body-centred, based-centred and face-centred variants to the
primitive cubic, tetragonal, orthorhombic and monoclinic systems. This presentation
will be limited to the simplest crystalline systems: cubic and hexagonal.

3.3.2. Miller indices

In crystallography, the origin of the lattice is generally chosen to coincide with a
centre of symmetry. The position of a point in space will be indicated relative to the
lattice point chosen as the origin. The coordinates are specified with the aid of three
non-coplanar vectors, a, b, c defining the unit cell. Thus a lattice point having the
position p, q, r, has the coordinates in the crystal:
 Structure of solids 59

Cubic system

lice) F(cfc)
a = b = c - ===9(

Tetragonal system

I
a = b*c - ===9(
Orthorhombic system

C I F
a *b *c- ===9(
Monoclinic system Triclinic system

C a*b*c-
a *b c- a*=y=9F
Hexagonal system
Trigonal or Rhorn bohedral system

a - b - c
===<1'20

90
\ bc
==90-= 120

Figure 3.7. The fourteen spatial Bravais' systems (P: Simple; C: Based-centred/: Body-
centred; F: Face centred). A cell contains 1 structural unit; C and / cells, 2 structural
units and F cells, 3 structural units.
60 Introduction to Materials Science

= pa
y = qb (3.3.)
= rc

a, b and c are the absolute values of three non-coplanar vectors, a, b, c defining the
unit cell. For example, for the cubic primitive unit cell of figure 3.8. (a), the points A,
B, C have respectively the coordinates 100, 010 and 001, the points D, E, F 110, 011,
101 and the point G, 111. In the body-centred cubic system (cubic system /; figure
3.8. (b)), the atom situated at the centre of the cube has coordinates 1/2, 1/2, 1/2.

Am
*
Am
I (bee) F (fee)
(a) Cubic system

(b) Hexagonal system

Figure 3.8. Cubic and hexagonal unit cells with indications of the different lattice pa
rameters: (a) the cubic system; (a) the simple (P), Body-centred (I) and Face-centred (F);
(b) the hexagonal system. The parallelepiped highlighted in heavy lines shows the primi
tive unit cell, which, by rotation of 120 about the r-axis, forms the hexagonal unit cell.

In the hexagonal system, four reference vectors are often introduced: , a , a 2 3

which are coplanar and c which is perpendicular to the plane xy.

Miller indices are generally used to index the lattice planes. For example, con
sider a lattice plane (figure 3.9.) cutting the reference axes at points (, , Z), which
are at distances equal to Aa, Bh, and Cc from the origin respectively. If a, b and c are
orthogonal, the plane has the equation:
 Structure of solids 61

y y 1

(3.4.)
Aa Bb Cc

Figure 3.9. Indexing of the plane defined by points X, Y, and Z.

In the equation (3.4.), the absolute values of the unitary vectors are respectively
equal to a, b, c.
Taking the reciprocals of A, and C and multiplying them by so that h, k and
/ are the smallest possible whole numbers, the Miller indices, h, k and / are obtained:

h= n_
A
k 1 (3-5.)

The Miller indices are identical for a family of parallel planes (figure 3.10.). It
should be noted it has no meaning to distinguish a particular plane in an infinite
structure.
Miller indices are determined simply by applying the following procedure. The
direction is considered and, starting from the point chosen as origin, the number of
parallel planes encountered along the distance a separating two points of the system
is counted. This is equal to the absolute value of the Miller index h in the direction x.
By continuing in the same way for the y and directions, the indices k and/ can be
determined.
This process can be illustrated for the cases of planes oriented perpendicularly to
the plane of the page containing figure 3.1 O.These planes are represented by the line
of their intersections with the plane of the figure. The points represent lattice points.
For the family of planes represented in the top left-hand corner of figure 3.9., the
number of planes encountered following the direction is equal to 2 and the Miller
index following this direction , h = 2. In the y, direction k= 1 and in the direction
perpendicular to the plane of the figure, / = 0 as the different families of planes
shown in figure 3.10 are parallel to z. The planes are defined by their Miller indices
(210) for the example studied. Other examples are also given in figure 3.10.
62 Introduction to Materials Science

Figure 3.10. Indexing of lattice planes perpendicular to the plane of the figure and paral
lel to the r-axis. In the figure different families of planes are shown with their respective
Miller index notations (from Steadman, 1982).

The lattice planes can also intersect one of the axes in its negative region. In this
case, the corresponding Miller index is negative. The parallel planes are placed (fig
ure 3.11.) in such a way that the reference axes and y are situated on both sides of
the plane passing through the origin. ( 1 2 0 ) or ( 1 2 0 ) notes the indexing of the family
of parallel planes in figure 3.11. It should be noted that the Miller indices (hkl) and
( M / ) b e l o n g to the same family of planes. This description is also used to designate
the family of planes, which are equivalent in a cubic crystal. Thus the planes
(100),(010), (001),(),(), and (OOT)are equivalent and designated by the
notation {100}.

Figure 3.11. Example of a family of planes [(120) or ( 120)] with negative Miller indi
ces. The unitary vectors a and b are on either side of the plane crossing the origin
(from Steadman, 1982).

In the case of hexagonal lattice systems, for preference the Miller-Bravais indi
ces are used to designate the planes. Figure 3.12. represents a hexagonal system
having its axis perpendicular to the plane of the paper; the two other axes are in the
plane of the figure. Although having different Miller indices, the three families of
planes represented in figure 3.12. (a) make up equivalent planes that can de derived
from each other by rotation through an angle of 120 about the axis z. With the aim
 Structure of solids 63

Figure 3.12. Hexagonal lattice with two or three axes at 120 in the plane of the page
and an axis c perpendicular to it: (a) indexing according to Miller; (b) indexing according
to Miller-Bravais (from Steadman, 1982).

of making the indexing of equivalent planes in the hexagonal system more

uniform, a third reference vector a is introduced in the plane
3 and a 2

supplementary index / (figure 3.12. (b)). The Miller-Bravais indices are defined in a
similar way to the Miller indices. They are written as h, k, i, /. The new index i is not,
however, independent of the indices h and k since the position of a point in the plane
is completely determined by two vectors. Therefore the following relationship links
h, k and r.
h+k+i =0 (3.6.)

or:
(h + k) (3.7.)

The three series of equivalent planes represented in figure 3.12. (b) have Miller-
Bravais indices equal to (1120), (1210) and (2110) respectively. The families of
equivalent planes having similar indices h, k, i can be derived by a cyclic permuta
tion. The notation {hkil} designates these. The family of equivalent planes shown in
figure 3.11 is represented by the symbol {1120 } .
64 Introduction to Materials Science

3.3.3. Notation of directions

The indexing of crystalline directions is more direct than that of planes. Con
sider (figure 3.13) a vector R with origin and end R:
R = Aa + Bb + Cc (3.8.)
Please note that vector R' = 2R defines the same direction as the vector R.

Figure 3.13. Unit cell (a,b,c) with the vector R showing a crystallographic direction.

The notation of directions with the indices u, v, w are therefore obtained by

dividing the coordinates of R (OA, OB and OC) by a number chosen in such a way
that u, mdw are integers and as small as possible. The notation of the direction is
given by the indices between brackets [uvw]. Please note that the direction [vw] is
in the opposite sense to [uvw] while the two families of planes (hkl) and QikT) are
identical. In the cubic system the directions [ 1 1 1 ] , [ 1 1 ] , [ 1 1 1 ] , [ 1 1 1 ] , [ 1 ] ,
[ 1 1 1 ] , [ 1 1 1 ] , [ I] are symmetrically equivalent. By analogy with the notation
used for equivalent planes, the whole total of equivalent directions in the cube are
represented by the notation (111).
Figure 3.14 illustrates the method to determine the indices w, v, w. The value of
the index u is equal to the number of points in the direction x. The indices and w are
obtained in a similar way. Since the directions are defined in the plane of the page,
the index w is zero.
The determination of a direction in the hexagonal system involves four indices
w, v,y, and w in order to keep the same notation system as for the planes (Miller-
Bravais). In this case, to keep a coherent approach, a relation similar to that in 3.6 is
defined:

u+v+j=0 (3.9.)

Figure 3.15. shows several examples for the notation of directions using the
four-index notation in the hexagonal system. For simplicity, the examples are limited
 Structure of solids 65

Figure 3.14. Examples of the indexation of crystalline directions in a plane (from Stead-
man, 1982).

Figure 3.15. Determination of directions in the hexagonal system. Notations with the
four Miller-Bravais indices.

only to those directions in the hexagonal plane having a symmetry axis c, perpen
dicular to the plane of the paper. The technique used is not as obvious as that used to
index the planes. It is necessary to proceed, as is shown in figure 3.15., with a vector
summation to ensure that relation (3.9.) is satisfied.

3.4. Determination of the crystalline structure

3.4.1. Principles for the determination of structures
The resolving power of an electromagnetic radiation, such as light, depends on
its wavelength. Thus the limit of resolution of an optical microscope, briefly de-
66 Introduction to Materials Science

scribed in section 10.1, is fixed by the wavelength of visible light, which is of the
order of 500 nm.
To determine the structure of crystalline solids on an atomic scale, it is indispen
sable to use rays which have a wavelength close to, or less than the interatomic dis
tance, i.e. radiation such as X-rays ( ~ 0.1 nm), high energy electron beams ( ~
0.005 nm) or slow neutrons ( ~ 0.1 nm).
The method most often used to study the atomic structure of solids is Electron
Microscopy. The resolving power of the electron microscope is of the order of
atomic size. However, as electrons are strongly absorbed by the material, it is impos
sible with this technique to analyse deeply into a crystal. In transmission electron
microscopy, very thin specimens, of the order of some tens to some hundreds of nm,
are used. In scanning electron microscopy thicker samples are used because the im
age is formed by the emission of secondary and back scattered electrons. Transmis
sion electron microscopy (TEM) and scanning electron microscopy (SEM) are
briefly described in chapter 10.
The organic or mineral glass lens used in light optics and the electromagnetic
lens used in electron microscopy, deviate the light or electronic rays, thus enabling a
real image of an object to be reconstituted. Effective lenses do not exist for X-ray or
neutron focussing. In these two cases, it is therefore impossible to obtain an image of
an object as in optical and electron microscopies.
The most commonly used method for studying crystalline structure is X-ray
diffraction. Because of their low coefficient of absorption, X-rays are able to pene
trate deeply into the material. The diffraction of slow neutrons is also an interesting
technique but it requires sophisticated and expensive experimental equipment. In
fact, the two methods are complementary: X-ray diffraction allows heavy atoms to be
located, while neutron diffraction is more adapted to determining the position of light
atoms and to distinguishing between atoms of similar masses. Electron diffraction is
also used to study crystalline structures. This technique usually coupled with trans
mission electron microscopy, is used to analyse crystals of very small sizes.

3.4.2. X-ray diffraction

X-rays, like light, are electromagnetic radiation. Whenever electromagnetic
radiation hits a charged particle, such as an electron, it induces a periodic
displacement of this electric charge. This becomes the centre of a spherical wave
propagating evenly in all directions in space. The incident radiation is therefore
partially diffused by the charged particle and all the surrounding electrons oscillate.
Its intensity is proportional to the number of electrons. The spherical waves diffused
by different electrons interfere with one another, i.e. in certain directions in space
their intensity is strengthened while in other directions it is cancelled. The
constructive and destructive interferences between spherical waves are at the origin
of X-ray diffraction by crystals.
When electromagnetic radiation strikes a lattice plane of a crystal, the incident
radiation is partially reflected by the atomic plane. Since the intensity absorbed by a
atomic plane is very weak, partial reflection occurs in a cascade from all the atomic
planes of the crystal. However, only reflection of the incident radiation for a well-
defined value of the incident angle is observed. This phenomenon is the result of
interference between radiation reflected by all the parallel lattice planes.
 Structure of solids 67

For the reflected waves to be in phase at the moment they arrive at the detector
(figure 3.16.), the lengths of the different trajectories taken by the radiation must be
equal to an integer, n, of the wavelength . Consider the trajectories taken by the rays
striking the atoms and B; the difference in the path lengths (ABC) between the
rays hitting the two atoms is equal to 2d sin . The reflected intensity will therefore
be a maximum for an angle of incidence satisfying the following equation (Bragg's
Law):
wA = 2</ sin
AA/ (3.10.)

<- Crystal

- -
Figure 3.16. Conditions for X-ray diffraction from crystalline planes (Bragg's Law)

In this equation is a positive integer (n = 1, 2, 3, etc.) and represents the

angle formed between the incident radiation and the family of diffraction planes. The
diffraction angle depends on the distance dhu between the planes, and so the Miller
index (hkl) can be determined. In the case of a cubic system, the distance d ui between h

the planes is given by:

d
C3.ll.)

where a represents the parameter of the cubic unit cell.

A crystalline structure contains a great number of lattice planes. For a definite
crystalline structure, it is, a priori, possible to calculate d xi as a function of the unit
h

cell parameters and obtain the value of the Miller indices of the family of planes
considered, the values of dhxi being related to the angles of diffraction by Bragg's
Law.
In general, the directions of diffraction only depend on the geometry and the size
of the lattice. Even when a monocrystal is not available, it is still possible to
determine the crystalline lattice by the Debye-Scherrer method using an agglomer
ated powder in the form of a cylinder as a sample. The experimental set-up for this
method is shown in figure 3.17.
68 Introduction to Materials Science

XR diffraction cone

Not diffracted beam

(a)

7 Diffracted beam trace

(210) (111) (100) A (110) (200)

(b)
X
)( (
x/lR -2q

Figure 3.17. (a) Schematic representation of an X-ray camera used for the powder
method. The detecting film is fixed onto the internal cylindrical part of the camera, (b)
Diagram of an X-ray pattern of a crystalline powder (primitive cubic lattice) recorded on
the film. A corresponds to the non-diffracted X-ray beam path. On the diagram, different
lattice planes corresponding to the lattice of the crystalline powder are recorded.

The sample is placed at the centre of the camera and the detecting film is fixed
onto the internal cylindrical part. The orientation of the crystalline grains of the pow
der is completely random and there is only a small fraction of the grains with lattice
planes correctly oriented for diffraction. The radiation diffracted by a family of par
allel planes ((110) in this case) makes a cone of angle 2 cutting the detecting film
(figure 3.17. (b)) according to an arc of a circle. In this way, a diffraction spectrum is
obtained made up of a set of lines characteristic of the crystalline lattice of the ana
lysed substance. Thus for a substance with a cubic lattice, the following relationship
is obtained:

5 0 = ^ ( + + )
2 2 2 2 (3.12.)
4a 21
'

In the primitive cubic crystals, each lattice plane which corresponds to the vari
ous possible values of the sum h + k + I ( 1 , 2, 3, 4, 5, 6, 8, etc.) is recorded on the
2 2 2

film in the camera as an arc of a circle. In figure 3.17. (b), the various lattice planes
involved have been shown for the powder spectrum. For other crystalline lattices, the
diffraction spectrum is more complicated.
Each type of crystal has its own specific powder diagram, and this is a 'finger
print' by which it can be identified. A computer file exists covering all known crys
talline materials (American Society for Testing Materials, ASTM,). By comparing
the intensity of intense rays (hkl) with the same rays in a reference mixture, the
concentrations of the constituents of a material are determined.
When an X-ray beam irradiates a substance, an emission of secondary X-rays
occurs which is characteristic of the atoms present. The re-emitted radiation is practi-
 Structure of solids 69

cally independent of the nature of the chemical bonds of the atoms and the analysis
of the re-emitted X-rays enables the chemical elements present in the irradiated sub
stance to be determined. This analytical technique is known as X-ray fluorescence
spectroscopy.
In concluding this section, it should be mentioned that X-rays are equally used
to study the structure of non-crystalline solids and liquids. In this case, however, a
clear diffraction image is not observed but only somewhat diffuse signals, which
nevertheless allow the average distances between near neighbouring atoms to be
determined.

3.5. Summary and conclusions

The balance between the cohesive energy bringing atoms together, and the
thermal energy, tending to separate them determines the physical state of a substance
(solid, liquid or gas). The thermal energy is proportional to the temperature while the
cohesive energy is to a first approximation independent. This situation explains the
change of state with temperature.
The arrangement of atoms in solids can have an ordered or a disordered charac
ter leading to two types of structure.
The amorphous structure, is dense and with little order and similar to
that of liquids. Over short distances (some atomic diameters), a certain
periodicity in the distribution of atoms (short range order) still exists.
This amorphous structure is often found in organic polymers (organic
glasses, rubbers) and also in some categories of ceramics (mineral
glasses), and a few, rapidly quenched metal alloys (metallic glasses).
The crystalline structure, is characterised by a periodic distribution of
atoms ordered over large distances (long range order). Practically, all
the metallic materials and a large number of ceramics exist in the solid
state as an assembly of microcrystals (grains). This grain structure is
also present in semi-crystalline organic polymers.

A certain number of monocrystalline materials are also encountered, such as

silicon used in electronics. Its manufacture is described in the illustrative example in
this chapter.
The crystalline structure, which has long-range order, can be precisely described
geometrically. Crystallography distinguishes seven fundamental crystalline systems.
This elementary treatment has been limited to the analysis of the cubic and hexagonal
systems, which are most common in materials.
It is important to make a clear distinction between the notion of a three-dimen
sional lattice, which is a regular geometric distribution of points in space and the
crystal structure produced by the translation of a structural unit or motif along the
lattice. The unit cell is the minimum volume element of simple geometry (cube,
parallelepiped) containing all the symmetry of the structure. The unit cell contains
the structural unit, which repeats itself forming the crystal The planes and crystallo-
graphic directions are specified in terms of Miller or Miller-Bravais indices.
70 Introduction to Materials Science

The crystal structure of materials is determined by methods based on the dif

fraction of electromagnetic radiation, such as X-rays or slow neutrons having a
wavelength close to , or less than the inter-atomic distances.

3.6. Illustrative example: the silicon mono-crystal, nerve

centre of any computers
At the present time, mono-crystalline silicon remains the indispensable material
to the Microelectronic Industry for integrated circuit manufacture. Because of its
semi-conducting characteristics, it is found in computers as well as in watches, hi-fi
audio equipment and in robotics and computerization machinery.
Silicon is the second element in order of abundance in the earth's crust (table
1.6) where it occurs in the form of oxides either simple (silica, S i 0 ) or complex
2

(silicates). Technical silicon is obtained by the reduction of the oxide S i 0 (sand)

2

with carbon or C a C in an electric furnace.

2

High purity silicon, as used for the manufacture of integrated circuits, is

obtained by the reduction of silicon tetrachloride (SiCl ) with hydrogen. Ultra pure
4

SiCl is prepared by technical silicon chlorination.

4

A strict control of electrical conductivity can only be obtained by starting from

ultra pure silicon, which is an insulator at ambient temperature. Industrial semi-con
ductors (chapter 14.) are obtained by adding to (doping) the ultra pure silicon mono-
crystals (an insulator at ambient temperature), a very small quantity of phosphorus
atoms (10 parts per million = 1 0 ppm = 10~ ) or of aluminium.
5

The manufacture of a silicon mono-crystal is carried out by an already old

method, due to Czochralski (figure 3.18.). A charge of ultra-pure silicon (purity de
gree: 99.99999 %) is placed in a quartz crucible and melted under an inert (or reduc
ing) gas in an electric furnace. After stabilisation of the fused silicon at 1420 C, a
mono-crystalline seed crystal, fixed to a vertical axis, is brought into contact with the
surface of the bath. The seed is then slowly pulled up, leading to solidification of a
single crystal having the same orientation as that of the seed (figure 3.19.). To
achieve structural homogeneity of the mono-crystal, the crucible and the nucleus are
spun contrariwise. Adjusting the withdrawal speed and the bath temperature control
the diameter of the crystal.
In 1952, at the beginning of the development of this technique, the diameter of
the silicon mono-crystals was of the order of 20 mm, while today values greater than
200 mm and lengths of 2 m, representing a volume of 50 to 100 litres, are achieved.
This represents a remarkable technical progress, considering that a monocrystal of
this type has almost perfect order without grain boundaries and with very few defects
and foreign atoms. The principal characteristics of these mono-crystals are
impressive (table 3.20.).
The mono-crystals are then cut transversally to their withdrawal axis into circu
lar 1 mm thick sheets, using diamond saws. Finally the sheets are polished to a mirror
finish to obtain a plane surface suitable for the engraving of integrated circuits by
microlithography.
 Structure of solids 71

1 Withdrawal system 4 Nucleus 7 Crucible

2 Observation window 5 Mono-crystal 8 Heating device
3 Entry and exit for 6 Molten mass
inert gas

Figure 3.18. Czochralski's method of pulling large silicium mono-crystals.

The demand for mono-crystalline materials increases daily and involves more
and more products. These monocrystals are destined for industrial applications requi
ring specific properties such as scratch-resistant watch glasses (sapphire (A1 0 ) 2 3

mono-crystal plates).

Figure 3.19. Progressive extraction of a mono-crystal from the surface of the fusion bath,
seen through an observation port.
72 Introduction to Materials Science

Table 3.20. Main characteristics of a silicon

monocrystal used in microelectronics.

Purity 99,99999 %
Length 2m
Diameter 0,17 m
Volume 451
Weight 106 kg
Moles of atoms 3765
Number of atoms 2-10 27

3.7. Exercises
3.7.1. Make the distinction between short and long range order. In what types of
materials are they found?

3.7.2. What difference can be drawn between a crystalline structure and a crystalline
system?

3.7.3. In the cubic crystalline system (figure 3.8. (a)), indicate the position of the
atoms at the points A, B, and G. Index the directions of the lines passing through the
origin and points A, and D. Designate two planes of the cubic lattice intercepting
the atoms and C.

3.7.4. What are the indices of the intersection line of the plane (111) with the plane
(101) in the cubic crystalline system?

3.7.5. For the cubic crystalline systems, the distance between two parallel planes
having the same Miller indices, d kh is related to the unit cell parameter a by the rela
h

tion:

a
(3.11.)
-Jh + k + l
2 2 2

Apply this relation to the (110) planes.

3.7.6. Give the list of the directions of the family of diagonals of a cubic system.

3.7.7. What are the indices for the planes parallel to the axis a given by the rectan
2

gular faces of the prism of the hexagonal unit cell (figure 3.8(b))?
3.7.8. What is the value of the angle a between the directions (111) and (001) in the
cubic crystalline lattice?
3.7.9. A sample of F e (Jbcc) powder is placed in an X-ray diffractometer with
a

wavelength = 0.154 nm. The diffraction angle 2 0 of the planes {110} is equal to
44.7 . Calculate the parameter a for the lattice of the bcc crystalline system of F e . a

The diffraction is first order.

3.7.10. The recording of the X-ray diffraction pattern for a cubic crystal powder gave
diffraction rays ( with = 1, 2, 3, etc.) for the following angles 20: 40, 58, 73, 87,
 Structure of solids 73

Table 3.21. Diffraction Planes for the crystalline systems (bcc) and (fee).

Sum h + k + I
2 2 2
Planes families {hkl}

bcc bcc

2 110
3 ___ 111
4 200 200
c
J

6 211 ...

/
8 220 220
Q
y

10 310

100.5 and 114.8 . Determine whether the crystalline structure is (fee) or (bcc). The
wavelength is = 0.154 nm.
It should be noted (table 3.21) that, for cubic body-centred crystals, the only ef
fective diffraction lines present correspond to planes having the sum h + k +1 equal 2 2 2

to 2, 4, 6, 8, 10, 12, etc. For the cubic face-centred system, a much greater number of
destructive interferences occur and only planes having the sum h + k + I equal to 3,
2 2 2

4, 8, 11, 12, 16, etc. give rise to diffraction.

3.8. References and complementary reading

D. R. ASKELAND, The Science of Engineering Materials, 2 Ed. Chapman & Hall, London, 1990.
n d

C. DELPHINE, Introduction la Cristallographie, Dunod, Paris, 1971.

P. DUCROS, Radiocristallographie, Dunod, Paris, 1971.
J.P. EBERHART, Analyse structurale et chimique des materiaux, Dunod, Paris, 1989.
Ch. HAMOND, The Basics of Crystallography and Diffraction, Oxford University Press, Oxford, 1997.
A. KELLY, G.W. GROVES and P. KIDD Crystallography and Crystal Defects, revised edition,Wiley,
New York, 2000.
D.V. RAGONA, Thermodynamics of Materials, Wiley, New York, 1995.
L. H. SCHWARTZ, J. B. COHEN, Diffraction from Materials, Springer-Verlag, Berlin, 1987.
. B. SMITH, Basic Chemical Thermodynamics, Clarendon Press, Oxford, 1990.
R. STEADMAN, Crystallography, Van Nostrand-Reinhold, Wokingham, Berkshire, U.K., 1982.
M. VAN MEERSSCHE, J. FENEAUX-DUPONT, Introduction a la cristallographie et a la chimie
structurale, 3e ed., Peeters, Louvain-La-Neuve, 1984.
 Chapter 4.

Structures of materials

4.1. Objectives
To give an appreciation of the structure of the main metallic, ceramic
and organic materials.
To describe the close-packed crystalline structures of metals and ionic
solids.
To introduce the concept of polymers
To bring in the structure of solids with covalent bonds: organic poly
mers and non-ionic ceramics.
Most solids crystallise or may do so. Crystal is the most stable physical state on
earth at low temperature. It is the common structure of metals and a large number of
ceramics. Certain ceramics and practically all the organic materials commercially
available to date are polymers. This chain structure, which considerably increases the
melt viscosity, drastically reduces the possibility to crystallise and leads to amor
phous or semi-crystalline materials.

4.2. Structure of metals

4.2.1. Close-packed cubic and hexagonal structures
The metallic bond is strong and non-oriented. Metal atoms tend to adopt the
most compact arrangement possible with, as a consequence, the formation of simple,
high density, close-packed crystal structures. It has been shown in Chapter 2 that it
was possible to regards the atom as a relatively hard and practically impenetrable
sphere. This model is generally adopted to describe the crystal structure of metals.
Figure 4.1. shows the structure of the close-packed arrangement of spheres (atoms)
of the same diameter, built up by successively stacking of layers of atoms. In figure
4.1. (a), the atoms occupy the lattice points of the basal plane of a hexagonal system.
Each atom is in contact with six neighbours occupying the corners of a regular
hexagon. The introduction of a second layer with maximum stacking is completed by
placing atoms in the hollows formed by groups of three atoms of the first layer. It can
be seen in figure 4.1 (b) that atoms of the second layer occupy only 50 % of the
hollows of the first layer. This last layer shows the existence of two types of hollows
designated in figure 4.1 (b) by the letters a and c. The letters and are used to
represent the atoms of the first and second layers.
There exists, therefore, two possibilities of placing the atoms of the third layer.
They can be put into the hollows of type a situated directly above the atoms of the

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

76 Introduction to Materials Science

first layer. The structure (1) of figure 4.1. (c) is obtained. It is a. hexagonal close-
packed structure (hep), characterised by alternating layers made up of spheres in
position ABABA... In this hexagonal structure with maximum stacking, the atoms of
layer A occupy the lattice points of a hexagonal system (plane (0001)) (figure 4.2).
The atoms of the layers occupy the lattice points of a second hexagonal lattice,
interpenetrated by and parallel to the first. From a crystallographic point of view, the
primitive structural unit of this crystal structure consists of two atoms: eight atoms of
layer A and C shared among eight adjacent structural units, and one atom of layer
shown in figure 4.2 (b) by a black sphere.

Figure 4.1. Compact stacking of layers of identical spheres, (a) One plane of atoms hav
ing compact hexagonal stacking with indication (black dots) of atoms in the second
layer, (b) Two planes of atoms having compact stacking with indication of atoms in the
third layer (ABAB type). The presence of two types of hollow, a and c, can be seen in
this second layer, (c) Three planes of atoms having compact stacking: (1) hexagonal
stacking (ABABA...), (2) face-centred cubic stacking (ABCABCA...). The plane having
compact stacking corresponds to the reticular plane (0001) in the case of hexagonal
stacking and to the reticular plane (111) for the face-centred cubic stacking. The hexagon
(1) and the cube (2) in bold lines represent the envelope of the atoms and not of the lat
tice.

A second configuration exists, as compact as the one just described but with a
different symmetry. This is obtained by placing the spheres of the third layer in the
hollows c situated above the unoccupied interstices of the first layer (figure 4.1 (b)).
A stacking of three superimposed layers ABC is thus obtained, with a successive
displacement before recommencing with the layer A. This stacking is therefore char-
 Structures of materials 77

acterized by the sequence ABCABC... The structure (2) of figure 4.1 (c) shows that
this stacking corresponds to a face-centred cubic structure (fee) of which the diago
nal [111] is perpendicular to the planes (111) of figure 4.3. This cubic structure has
four atoms per structural unit.
[0001]

(a) (b)

Figure 4.2 Unit cell for the hexagonal close-packed crystalline system (hep): (a) model
of atoms; (b) hep unit cell in a space lattice. The (0001) planes correspond to the planes
A of the compact hexagonal structure (figure 4.1). The right prism of height c and having
as base a rhombus of side a, shown in bold lines, constitutes the primitive unit cell of the
hexagonal lattice. Two atoms can be counted per structural unit, i.e. six atoms for the
compact hexagonal unit cell.

(a) (b)

Figure 4.3 Unit cell of the face-centred cubic (fee) lattice: (a) model of atoms (4 atoms
per unit cell); (b) two superimposed fee unit cells (point lattice) with indication of the
close-packed (111) planes which correspond to the A, and C planes of figure 4.1.

For each of these two close-packed structures hep and fee, the number of atoms
immediately adjacent to a given atom is equal to 12. In both cases a coordination
number CN = 12 is defined. For atoms of the same diameter, there is no structure
more compact than that of these two types of crystal. It can be shown by a simple
calculation that 74 % of the space in the lattice is filled by spherical atoms (packing
factor = 0.74).
There exists a third crystalline structure frequently observed in metals: this is the
body-centred cubic structure (bcc) (figure 4.4) which is less compact than the
78 Introduction to Materials Science

structure hep or fee. An atom situated at the centre is in contact with each of eight
atoms situated at the corners of cube, making the coordination number (CN) there
fore equal to 8. The packing factor is 0.68 and the unit cell contains two atoms.
The crystalline structures of metals currently used are given in table 4.5. In this
table, the lattice parameter is also given, i.e. the edge a of the basic cube (cubic sys
tem) and the height c of the hexagonal prism (hexagonal system), with the dimension
a of the base hexagon.

(a) (b)

Figure 4.4 Unit cell of the body-centred cubic (bec) lattice: (a) model of atoms (2 atoms
per unit cell); (b) bec unit cell with point lattice.

Table 4.5. Crystal structure and lattice parameters of some pure metals
at ambient temperature (unless otherwise indicated)

Metal Structure Lattice parameters

a c
[nm] [nm]

Al fee 0,405
Cu 0,362
Fe 0,366 [a]
Ni 0,352

Cr b c c
0,288
Fe a 0,287
Mo 0,315
T\ 0,330 [b]
W 0,316

Mg hep 0,321 0,521

0,295 0,468
Zn 0,266 0,495

[a] at 910 C; [b] at 882 C

All metals with a pronounced metallic bond crystallise according to one of the
three types of structure: fee, hep and bcc (table 4.5.). It must be emphasized that the
crystal structure exerts a considerable influence on the physical properties of metals.
In particular, the possibilities for plastic deformation (chapter 13) are more limited
for hexagonal systems than for face-centred cubic ones. Metals crystallising in the
hexagonal system (Mg for example) are more difficult to draw into the form of wires
or rods than fee metals such as copper or aluminium.
 Structures of materials 79

Certain metals have the ability to change the crystal structure as a function of
temperature (polymorphism). This is the case, for example, of iron which exists un
der structures designated by the symbols , 7, or for titanium a and .

4.2.2. Interstitial sites

The octahedral sites (figure 4.6. (a)) are formed by six spheres situated at the
corners of an octahedron. The deposition of layer on l a y e r s (figure 4.1. (b)) leaves
50 % of the hollows in the first layer vacant. These interstices make up the octahedral
sites.
In crystalline systems, the empty spaces between spherical atoms constitute
interstitial sites. In certain alloys (chapter 8.) or during the formation of intermediate
compounds between atoms of different size, these sites act as locations for atoms of
small diameter. There exist two types of interstitial sites: tetrahedral and octahedral
sites, shown for an fee structure (figure 4.6.).
The tetrahedral sites (figure 4.6. (b)) are formed by the compact stacking of four
spheres, the centres of which are at the corners of a tetrahedron. The tetrahedral sites
form when a second layer is placed on a plane of maximum stacking (figure 4.1. (b)).
Each sphere of layer is in contact with three spheres of layer A. These four spheres
border a tetrahedral interstice. There also exist tetrahedral and octahedral sites in the
two other types of crystalline structure that have been analysed.

(a)
Figure 4.6. Interstitial sites in the fee structure: (a) octahedral sites; (b) tetrahedral sites.
The interstitial sites are represented by black dots.

4.3. Structure of ionic crystals

4.3.1. Factors influencing the structure of ionic solids
The ionic bond, which is strong, can be found in a significant number of crystal
line ceramics, especially in metallic oxides such as rutile ( T i 0 ) , corundum (A1 0 )
2 2 3

or zircon (Zr0 ). Like the metallic bond, the ionic bond (chapter 2) is not oriented. In
2

solids made up of ions of different sign, the crystal adopts a compact structure which
allows the interactions between ions of opposite signs to increase to a maximum and
those between ions of the same sign to be minimised. This condition considerably
limits the choice of possible structures.
80 Introduction to Materials Science

Ionic solids, which have the same number of positive and negative charges, are
electrically neutral and there are two important factors to be considered in the analy
sis of their structure.
It is necessary to take account of the valence of the ions, i.e. the number
of positive (+) or negative (-) charges that they carry. This determines
the overall formulation of the ionic compound A ~ X , B X \ . . . + 2 _
2

C ~Y , etc. according to whether the ions are respectively mono-, b i - ,

2
3
3
2+

or trivalent.
It is equally important to take into consideration the relative size of the
ions which determines the manner of stacking and the coordination
number. In applying the model of a rigid sphere, it is possible to predict
the most favourable way of stacking from an energy point of view.
Whatever the type of system considered, the energy released by the
formation of the crystal is the highest when the distance r between ions
of opposite sign is the smallest.

4.3.2. Crystals from ions with identical charges in absolute value

In this case, the structural unit is composed of an anion and a cation. The crystal
structure, which forms, is that which releases the most energy during its creation. The
energy of formation of ionic crystals is a function of the ratio between the radius of
the cation R and the radius of the anion R and it is for this reason that the
c a

coordination of the ions in the ionic crystals varies as a function of RJR a

If RJRa > 0.72 , the most stable crystal form is a simple cubic structure of the
CsCl type (figure 4.7(a)). Its structural unit is composed of one CI" anion and one C s +

cation. In this case, the degree of coordination of the ions is equal to 8. Each C s i o n +

of the system is thus surrounded by eight CI" ions and, conversely, each C r i o n is
surrounded by eight Cs . The crystal structure can be built up starting from either the
+

cation C s o r the anion CI".

+

If 0.33 < RJRa < 0.72, the preferred crystal form is a face-centred cubic structure
(fee). The coordination number for the ions is 6, the model compound in this case is
sodium chloride, NaCl (figure 4.7 (b)). The structural unit of this face-centred cubic
structure is composed of four anions Cl~ and four cations Na . +

er cs +
CP Na +

(a) (b)

Figure 4.7 Evolution of the crystalline structure of monovalent ionic compounds as a

function of the diameter of the atoms: (a) cesium chloride, coordination 8; (b) sodium
chloride, coordination 6; (c) zinc sulfide, coordination 4.
 Structures of materials 81

When the relation RJR < 0.33, the coordination of the ions in the preferred
A

structure is 4. This is the case for zinc sulphide, ZnS, the structure of which is shown
in figure 4.7 (c). As in the previous case, the ZnS structural unit is composed four
anions and cations.
The structure of simple ionic crystals can also be deduced from the compact
structure of metals. In the case of sodium chloride, the N a cation fits almost exactly+

in the octahedral interstices of the close-packed face-centred cubic system formed by

the Cl~ anions. The crystalline system of NaCl is similar to a slightly dilated close-
packed cubic system. For the ZnS crystalline structure, the stacking of sulphide ions
S ~ forms a close-packed face-centred cubic system in which the Z n ions occupy
2 2+

half of the interstitial tetrahedral sites.

4.4. Structure of solids with covalent bonds

4.4.1. Polymers
Polymers are made up of long chains of atoms joined to each other by covalent
bonds. Polymer chains are composed of atoms having a valency equal to, or greater
than, 2. The main elements forming the basis of polymer structures are oxygen (bi
valent), carbon and silicon (tetravalent). Carbon and silicon form four single bonds
oriented along the angles of a tetrahedron (figure 2.11 and 2.12). Carbon-carbon
double bonds (figure 2.14) are also present in certain polymers (figure 5.3).
The carbon atom can form the simple polymer structure shown in figure 4.8 (a).
This linear chain contains a significant number of unoccupied covalent sites (two per
carbon atom in the chain and three per terminal carbon atom). Two limiting cases are
possible. When all the non-substituted covalent sites are occupied by a monovalent
element, such as hydrogen, a linear polymer (polymethylene (CH )) is obtained 2

(figure 4.8 (b)). If the vacant sites of the carbon-carbon chain (figure 4.8 (a)) are
substituted by other carbon atoms, the three-dimensional structure of diamond is ob
tained. Its crystalline structure is shown in figure 4.9 and its crystalline lattice has
been given in chapter 2 (figure 2.12. (c)).
A polymer is a macromolecule obtained by the repetition of a building block
(structural unit). In the case of polymethylene, the building block is the methylene
group (-CH -). This polymer is a semi-crystalline solid, synthesized by the polymeri
2

sation of ethylene. This small molecule (monomer) is a supercritical gas at ambient

temperature. In the presence of a catalyst, the double bond in ethylene opens and
forms two carbon-carbon single bonds with two other ethylene molecules. This
chemical reaction leads gradually to the formation of a polymer molecule:
H 2 H 2 H 2 H 2 H 2

H=CH 2 ^ C ^ C ^ C ^ C ^ C ^ ^ ^
H 2 H2 H
2 H 2 H2

This polymer material obtained by the polymerisation of ethylene (chapter 5) is

polymethylene. Although correct, this name is never used in practice and the polymer
is known as polyethylene (PE). This thermoplastic polymer is processed in the mol
ten state and solidifies when cooled down to ambient temperature. The melting of
this semi-crystalline polymer occurs at a relatively low temperature (110 - 130 C)
82 Introduction to Materials Science

because the interchain secondary bonds are weak. During fusion, the linear chain
structure is maintained by the strength of the carbon-carbon covalent bond. The fu
sion-solidification process can be repeated many times without altering the chain
structure of the polymer.

Carbon chain

Polymethylene molecule (CH )n ( polyethylene)

2

OC OH

Figure 4.8. Structure of a macromolecular chain: (a) carbon skeleton; (b) polymethylene
molecule (polyethylene, (-CH -)).
2

A fundamental difference exists between the structure of polyethylene and that

of diamond. Polyethylene is a linear macromolecule, composed of a limited number
of chain carbon atoms depending on its molecular weight. Each polyethylene chain is
terminated by a methyl group (-CH ). In the case of diamond, no individual molecule
3

exists and the polymer structure extends in the three spatial dimensions and has the
same dimensions as that of the crystal. As an example, the carbon skeleton of a car
bon-carbon chain in the diamond structure is shown in bold lines in figure 4.9. The
diamond crystal is composed of a very large number (practically infinite) of chains of
this types linked to each other by covalent bonds.

Figure 4.9. General view along the [111] axis of a diamond shown in exploded (non
compact) form highlighting in bold lines one carbon-carbon chain (after Bundy, 1974).

Four types of polymer structures can be identified: linear, branched, two-dimen

sional and three-dimensional polymers. In figure 4.10., the different chain structures
 Structures of materials 83

found in the case of polyethylene are shown. The linear and branched macromole
cules (figure 4.10. (a) and (b)) are soluble and fusible. The three-dimensional cross-
linked structure (figure 4.10. (c)) is insoluble and cannot pass into the liquid state.
However, at ambient temperature, the chain segments between cross-links are able to
crystallise. This cross-linked structure has a much greater resistance to heat. Because
of this characteristic, three-dimensional polyethylene is used as the electrical insula
tion on underground high voltage (200 kV) power lines. On the other hand, graphite
is an example of a two-dimensional insoluble macromolecular 100% structure
(figure 2.16.).

(a) Linear polyethylene

(b) Branched polyethylene

Short chain branching .

Branching ^
connection point

(c) Cross-linked polyethylene

Figure 4.10. Schematic representation of various polyethylene chain structures.

Polymer are also synthesized by combining silicon and oxygen, (figure 4.11.).
The formation of polysiloxane chains (figure 4.11. (a)) is favoured by the high heat
of formation of the Si bond (452 k J m o r - table 2.17.). The SiSi bond
1

(222 kJ m o l ) found in the polymer structure of crystalline silicon, is analogous to

-1

that of diamond. Crystalline silicon is actually the most utilised material for the
manufacture of integrated circuits as well as photovoltaic solar panels. In a polysi
loxane chain (figure 4.11. (a)), there remain two unfilled positions on each silicon
atom in the chain. They can be occupied by monovalent organic groups such as the
84 Introduction to Materials Science

methyl group (-CH ). In this way, a linear organo-mineral

3 macromolecule,
poly(dimethylsiloxane), is obtained (figure 4.11 (b)) which forms the basic structure
of silicone oils and rubbers. Silicone rubbers make up a class of elastomers having a
high temperature resistance (350 - 400 C).
Oxygen anions (-0") linked to the chain silicon atoms by a polar covalent
bond can also occupy the vacancies on the polysiloxane chain. These oxygen anions
combine with mineral cations (Na , Mg , A l ) to form ionic bonds. In this case,
+ ++ + + +

mineral silicates (ceramics) are produced such as pyroxene (figure 4.11 (c) ). There
are only strong bonds in the structure of this mineral (covalent, polar or ionic). These
mineral polymers are considerably more rigid than the organic polymers where the
bonds between the polymer chains are made of secondary weak bonds,
(a) Polysiloxane chain

(b) Poly(dimethylsiloxane) (rubber or silicone oil)

r

(c) Chain silicate: (MgSi0 ) (pyroxene-ceramic)

3 n

M +
M +
M +
M +

r r r r r
|r jiji|i
0" o- o- o- cr
M M+
M +
M + +

r r
-|i|ijiji|i
r r r
CT 0" o- o-
M M M

r r r r r
M M
+ + + + +

-|ijiji |t-
o- cr o- cr o-
M +
M +
M+

+
M +

Figure 4.11. Organo-mineral and mineral polymers derived from silicon.

It is also possible to link the polysiloxane chains by means of polar covalent

bonds (figure 5.13.). This leads to three-dimensional silica with the general formula
S i 0 which exists in an amorphous form (glassy silica) or in a crystalline state (cris-
2

tobalite, quartz,...). In the following chapter, the structure of silica and the silicates
 Structures of materials 85

which make up an important class of ceramics will be covered in more detail. Poly
mer structures are thus encountered in organic materials and in a large number of
ceramics.

4.4.2. Principal characteristics of solids with covalent bonds

Solids with isotropic bonds (metals, ionic materials) form compact crystal
structures where the organisation is determined primarily by the volume occupied by
the atoms and their relative size (steric effect). In the case of solids with covalent
bonds, the situation is completely different. Indeed, the covalent bond can only form
in well-defined orientations. Consequently, in solids with covalent bonds, the struc
ture is first determined by the geometry of the orbitals and subsequently by the steric
effect. These solids have relatively low packing. This characteristic can produce
materials combining low density with high mechanical strength.
The most typical example of this type of material is diamond, which has a den
sity of 3.52 g cm" (tm~ ) only. A solid with close-packed stacking made of carbon
3 3

atoms with a diameter of 0.154 nm would have a density of 7.5 g cm" (tm" ). Despite
3 3

its low density, diamond is the solid with the highest stiffness and hardness.
Another characteristic of the covalent bond is to favour the formation of long
chain structures (polymers). In organic materials, these polymer structures are
mostly macromolecules, soluble in organic solvents. In this case, the inter-chain
cohesive forces are secondary valency forces between 10 and 100 times weaker than
the intramolecular bonding forces linking chain atoms together.
The intermolecular bonding forces constitute the polymer weakest link. This
explains their low level of mechanical properties. Their rigidity is in general 10 to3

10 times less than that of diamond, which contains only covalent bonds (figure 4.9).
6

Organic substances must have a polymer chain structure to exhibit mechanical

resistance. This very important point will be covered in detail in chapter 13. This is
not the case for ceramics, which, even in the absence of a polymer structure, possess
high stiffness and strength. In these solids, practically only strong bonds (polar
covalent and ionic) are found. The basic structure of most common ceramics (chapter
5) is the tetrahedron (Si0 ) ~ associated with a variety of metallic cations. This is the
4
4

case for zircon (ZrSi0 )> a material with a higher hardness than that of quartz.
4

In the liquid state, metals and ionic materials are characterised by a low viscos
ity ( 1 0 to 10" Pas), close to that of water. Under appropriate thermodynamic con
_1 2

ditions, these materials crystallise rapidly and give almost exclusively crystalline
solids. On the other hand, molten organic and inorganic polymers have a high viscos
ity (10 to 10 Pas) and are difficult to crystallize. In this case, amorphous (glassy or
5 6

rubbery) or semi-crystalline structures are often encountered.

4.5. Summary and conclusions

The crystal structure and properties of materials are directly dependent on the
characteristics of their chemical bonds.
Metallic and ionic bonds are isotropic and form dense crystalline solids with a
simple structure and a high degree of symmetry.
86 Introduction to Materials Science

Metals, where atoms are rigid spheres of the same diameter, adopt a more com
pact structure than ionic ceramics where ions of different charge and size are in
volved. Metals form crystals with relatively simple crystal structures: hexagonal
close-packed (hep), face-centred cubic lattice (fee) and body-centred cubic(Z?cc).
In materials with covalent or polar covalent bonds, the atoms are generally con
nected by strong and oriented bonds. These solids no longer adopt a close-packed
structure. This characteristic leads to the formation of light materials. In this case,
chain structures (polymers) are frequently formed going with highly viscous melt and
a drastic reduction in crystallisation possibilities. These physico-chemical character
istics often produce materials with completely amorphous structures such as glasses,
rubbers.
Organic polymers are in general composed of long chains, mostly containing
carbon atoms. Weak bonds link these polymer chains to each other. This structure
gives materials of low rigidity. Organic polymers can be made of linear or branched
macromolecules, which can melt, or of two- or three-dimensional structures which
cannot. Some ceramics also have polymer structures, but their inter-chain bonds are
generally strong (ionic or polar covalent) which leads to solids with high mechanical
strength.

4.6. Illustrative example: metallic glasses, materials

for new technologies
The most stable state of materials at low temperature is the crystal. Nevertheless,
all substances are, in principle, capable of existing in the glassy state (chapter 8) if
they are cooled rapidly enough to fix them in a disordered structure close to that of
the liquid state. Under normal conditions however, the formation of glasses is
observed only for materials, which crystallise slowly or cannot crystallise. This is the
case notably for molten quartz which, on cooling, forms a glass (amorphous silica) or
for polymers with rigid chains, such as polystyrene or poly(methylmethacrylate)
(Plexiglas) which are glassy in the solid state.
For pure metals, which are characterised by a very high rate of crystallisation,
the formation of a glass from the melt has not, so far, been observed. Alloys crystal
lise more slowly than pure metals because of diffusion phenomena. It has been pos
sible, by hyper-quenching, to obtain metal alloys in the amorphous state. With spe
cialised techniques, cooling rates of between 10 and 10 degrees per second are
5 9

achieved. The quenching can be carried out, for example, by squeezing droplets of
metal alloys between two chilled copper plates or by melt quenching onto a Cu
wheel. A continuous cooling process has been developed (figure 4.12), using a
quenching canon which projects a very fine jet of molten alloy onto a chilled copper
drum. In this way, a continuous tape of metallic glass can be produced at speeds of ~
30 ms~\ it is only possible to obtain amorphous metal alloys by working with very
small thicknesses. The amorphous metallic tapes actually produced have a thickness
of less than 50 and a width of 5 to 20 cm.
In contrast to organic glasses, metallic glasses are ductile. They are charac
terised by quite exceptional mechanical strength. Certain amorphous alloys (for
example the composition Fe oCr Mo6B28) reach an elastic limit of greater than 3500
6 6
 Structures of materials 87

Figure 4.12. The melt quench apparatus (planar flow casting process) enables a glassy
metallic tape to be obtained, thanks to high cooling rates (> 10 K ). (A) Fusion of raw
6 _1

materials in an induction furnace. (B) Storing of molten alloy in a hopper. (C) Valve. (D)
Projection of a continuous film onto the rotating mould through a nozzle. (E) Solidifica
tion of the amorphous alloy in less than 10" s. (F) System working in tandem with the
3

feed system to adjust the tape thickness. (G) - (H) Transport systems and winding of
tape, (after Allied Metglas Products).

MPa, while that of high strength steel has an elastic limit of the order of 2500 MPa.
These glassy metallic alloys are also more corrosion resistant than crystalline alloys
since they do not have grain boundaries or inhomogeneities which are, in normal
alloys such as stainless steel, the starting point for localised corrosion by the
formation of minute holes on the surface (corrosion by pitting) (chapter 16).
The discovery of metallic glasses is fairly recent (Duwez, 1960). There is a large
potential for use of glassy metallic alloys. Several industrial applications for these
alloys are known, notably for making homogeneous soldering alloys for the aeronau
tical industry. To obtain a solder having a sufficiently low melting point to join two
metal pieces without melting them, it is necessary to introduce significant amounts of
boron (-weight 3 %) into the amorphous alloys.
Another interesting application of these amorphous alloys is in electrical trans
formers. The amorphous alloys possess magnetic properties more appropriate for the
manufacture of transformer cores than polycrystalline sheet-steel. This allows re
placement of the transformer core by ultra-thin sheets of glassy metal alloy, bringing
about a significant reduction of losses. The magnetic domains of ferromagnetic mate
rials (chapter 15) are delimited by Bloch walls. In transformers using thin sheets of
soft steel, the movement of the Bloch walls is partially restricted, inducing energy
losses in the transformer in the form of heat and acoustic emissions. In amorphous
metallic alloys, there are no grains with different orientations, nor precipitates, nor
other structural inhomogeneities. The movement of the Bloch walls is greatly
facilitated and the losses by hysteresis are much smaller. This leads to an appreciable
energy saving and a reduction in noise levels.

4.7. Exercises
4.7.1. In a close-packed crystal structure, what is the number of atoms in contact with
each neighbour?
88 Introduction to Materials Science

4.7.2. What is the density of atoms in the [111] and [110] directions of the aluminum
crystal?
4.7.3. Which planes possess the greatest density of atoms in the (fee) and (bcc) crys
tals?
4.7.4. How many atoms per c m are found in the (110) plane of chromium?
2

4.7.5. Calculate the density of aluminium.

4.7.6. Iron undergoes an allotropic transformation (polymorphism) at 912 C passing
from the body-centred cubic (bcc) to the face-centred cubic (fee) system on heating.
Knowing that the radii of the (bcc) and (fee) iron atoms are equal to 0.124 nm and
0.127 nm respectively, calculate the relative volume variation during the allotropic
transformation.
4.7.7. In the case of close-packed crystals (fee and hep), calculate the radius of the
octahedral interstitial site relative to the radius r of the atoms.
4.7.8. Show that the carbon atom of radius 0.062 nm cannot be inserted into an octa
hedral interstice of the face-centred cubic y-iron crystal without deformation of the
lattice ( t h e atomic radii of Fe are given in exercise 4.7.6.).
4.7.9. Corundum is a crystal of aluminium oxide A 1 0 (ceramic) with a hexagonal
2 3

crystalline lattice. How many A 1 0 groups ( A l ions and O " ions) can be found in
2 3
3+ 2

this hexagonal lattice ?

4.7.10. Show that the valency angle between the covalent bonds of tetrahedral
carbon is approximately 109 .

4.8. References and complementary reading

ALLIED METGLAS PRODUCTS, The Markets and Applications for Metglas Alloys, Parsipany (N.J.),
1985.
D. ASKELAND, The Sciences and Engineering ofMaterials 2 . ed., Chapman & Hall, London, 1990.
end

CR. BARRETT, T. B. MASSALSKI, Structure of Metals, 3rd. ed., Pergamon, Oxford, 1980.
F. P. BUNDY, Superhard Materials, Sei. Am., 231,62, (1974).
P. CHAUDHARI, B. C. GIESSEN, D. TURNBULL, Metallic Glasses, Sei. Am., 242, 84, (1980).
H.A. DA VIES, The Formation of Metallic Glasses, Phys. Chem. Glasses, 17, 159, (1976).
W. D. KINGERY, H.K. BOWEN, D.R. UHLMAN, Introduction to Ceramics, 2nd ed., John Wiley, 1976.
W. KLEMENT Jr., R.H. WILLES, P. DUWEZ, Metallurgy of Non Crystalline Structure in Solidified
Gold-Silicon Alloys, Nature, London, 187, 869 (1960).
. H. RAY, Inorganic Polymers, Academic Press, New York, 1978.
H. YANAGIDA, K. KOUMOTO, . MIYAYAMA, The Chemistry of Ceramics, Wiley, Chichester,
U.K., 1996.
 Chapter 5.

Organic polymers and ceramics

5.1. Objectives
To introduce the main types of polymer materials.
To describe briefly the molecular structure of the important organic
polymers.
To give a general survey of the structure of silicates.
To describe the amorphous and semi-crystalline ceramics.
To present the new high strength, highly heat-resistant ceramics.
Organic polymers are synthesised by polymerisation or poly condensation of
monomers, which are low molecular weight, reactive molecules. As has just been
seen in chapter 4. Organic polymers are linear, branched or cross-linked, and they
give rise to different types of synthetic materials: amorphous or semi-crystalline
thermoplastics, elastomers, and high-strength fibres, etc., which are essential to many
modern technologies.
Historically, ceramics are essentially materials derived from oxides by firing at
high temperature. Traditional ceramics make up the products for mass consumption:
bricks, refractory materials, pottery etc. Alongside these traditional materials, a new
generation of ceramics, such as silicon carbide and nitride, has developed. These
technical ceramics are extensively used in advanced technologies: abrasives, ferri-
magnetic ceramics, glass-ceramics etc. They are very hard refractory materials used,
for example, in diesel engines, for artificial bones, etc. Ceramics become extremely
important for new technical developments.

5.2. Organic polymers

5.2.1. Basic concept
As described in chapter 4., a polymer is a macromolecule obtained by the repe
tition of a building block (constitutional repeat unit) consisting of a group of atoms
joined by covalent bonds. Thus polystyrene, which is a well-known polymer, is
written as:
. . . -CH CHCH CH-CH CH-CH CH- .
2 2 2 2

(5.1.)

The building block of this polymer is made up of the following group of atoms:

2002 Editions scientifiques et medicales Elsevier. All rights reserved..

Introduction to Materials Science

90 Introduction to Materials Science

-CH 9

(5.2.)

The polymers are synthesised by joining monomer molecules with covalent

chemical bonds. Polystyrene is synthesised by polymerising styrene molecules. Sty-
rene is a liquid at ambient temperature, while its polymerisation yields a glassy, brit
tle solid.
As a result of the action of various initiators, the double bond of the styrene
opens to form two carbon-carbon single bonds with two other styrene molecules, thus
leading to the formation of a polystyrene molecule:

H C=CH
2

Monomer <

Reactive site Reactive site

Styrene

(5.3.)

CH CHCH CHCH CH
2 2 2

Polymer J

Polystyrene molecule

Each carbon atom of the vinyl group ( C H C H - ) of the styrene joins itself to a
=

carbon of another styrene molecule. By this reaction, a linear polymer chain is

gradually formed. A polymerisation reaction is, as a general rule, highly exothermic
as each double bond with a heat of formation of 602 kJ mol" (table 2.17.) is replaced
1

by two single bonds with heat of formation of 2-346 = 692 kJ mol" . 1

Chain-reaction polymerisation reactions take place in the presence of a small

amount (< wt 1 %) of an initiator generating highly reactive sites that are radicals,
anions or cations depending on the type of polymerisation. The active sites are pre
sent in very low concentration (10~ to 10~ mol in radical polymerisation). Each
7 9 1

reactive site adds in a very short time (generally less then one second) a large number
of monomer molecules (10 to 10 ). In chain-reaction polymerisation, high molecular
3 5

weight molecules form practically immediately at the start of the chemical reaction.
When a small proportion (less than one percent) of a monomer containing two
vinyl groups, such as p-divinylbenzene, is introduced into the styrene, then a three-
dimensional, interconnected structure is formed as shown in figure 5.1,which is in
soluble in all organic solvents, while linear polystyrene is completely soluble in
many organic solvents such as benzene or toluene.
Thus, the behaviour of a linear polymer differs considerably from that of a three-
dimensional network.
 Organic polymers and ceramics 91

Divinylbenzene

Figure 5.1. Three-dimensional network synthesized by polymerization of a mixture of

styrene and p-divinylbenzene. The black lines represent parts of the polystyrene chain.

The majority of industrial polymers are linear or slightly branched macromole-

cules. A polymer such as polystyrene is made up of a chain where methylene (-CH -) 2

and phenylsubstituted (C H -) carbon groups alternate

6 5

A large number of polymers (the vinyl polymers) are similarly built up. Thus by
replacing the phenyl ring of polystyrene with a chlorine atom, Polyvinylchloride
(PVC) is obtained:

C CH- -c -CEHC CH- -C CH- -C CH- (5.4.)

H 2 J H 2

CI
H 2 I
CI
H 2 I H 2 j
CI CI CI
Poly(vinylchloride) (PVC)

The structural unit of Polyvinylchloride is shown between brackets in (5.4.). The

monomer, which is used to synthesize this polymer, is vinyl chloride:

CH2=CH
I (5.5.)
CI
Vinyl chloride

Vinyl chloride monomer is a gas, which liquefies under normal pressure at a tem
perature of - 1 2 C. By opening the double bond in vinyl chloride with a initiator, a
polymer chain is formed:
92 Introduction to Materials Science

CH =CH
2

(5.6.)
C 1
Cl CI CI CI
Vinyl chloride Polyvinyl chloride) (PVC)

More generally, vinyl polymers

(5.7.)

X X X X

are obtained by the polymerisation of vinyl monomers:

CHr=CH (5.8.)

Replacing X with different chemical groups yields a wide variety of polymers. In the
two examples considered here, replacing X by a phenyl group or by a chlorine atom
gives polystyrene and Polyvinylchloride respectively.
The sequence of the basic units is not always as regular as shown in (5.7.). In
fact, three types of structural units are formed:

- f i r f - S r f - - f i r f - f - S , f - k - f i r f 1
- 5
- 9
- )
X X X X X X
head-tail head-head tail-tail

The three structures shown in (5.9.) are examples of positional isomers. By far the
most common structure is the head-to-tail configuration. For polystyrene, it is practi
cally 100 %. The largest deviation from the head-to-tail arrangement is 15 % for
certain halogenated polymers.
More over, head-to-tail vinyl polymers also exist in different structures known
as stereo isomers or configurational isomers. As shown in figure 5.2. where the chain
carbon atoms are placed in the plane of the paper (zigzag plane conformation), three
main forms (isomers) are identified. In the isotactic structure (a) all the identical
substituents are located on the same side, in front of or rear of the plane of the paper.
In the case of the syndiotactic polymer (b), the identical substituents alternate in front
of and rear of the plane of the paper. In the heterotactic polymer (c), the identical
groups are distributed more or less randomly.
The technical characteristics of polymers vary considerably as a function of their
configuration. Isotactic and syndiotactic polymers are, in most cases, semi-crystal
line, giving them a greater resistance to temperature ( 15.5.2.). The heterotactic
polymers are in general amorphous; they form (organic) glasses. The synthesis of
isotactic and syndiotactic polymers necessitates more sophisticated catalysts. One of
the great achievements of macromolecular chemistry has been moreover the syn
thesis of isotactic polypropylene by Ziegler-Natta catalysts (Nobel Prize for Chemis
try, 1963).
 Organic polymers and ceramics 93

CH H 3 CH B 3 3 CH 3 CH 3

"-J \ / \ / \ / \ /
^ c c c c c
(a)
^ ^ c ^ \
/ \ / \ / \ / \ / \
H H H H H H H H
Isotactic polypropylene
<:H CH H H CH CH
\/ \/ \7
3 3 3 3 3

\ J \ J
(b) \ c / c
\ c / - c
\ c / c
\ c / c
\ c / c
\

/ / \ /\ /\ /\
H H H H H H H H
Syndiotactic polypropylene
^CH 3 ^CH CH ^H
3 3 ^CH 3 ^CH 3

(c)
C C C C C

/ \ / \ / \ / \ / \
H H H H H H H H
Heterotactic polypropylene

Figure 5.2. Three configurational isomers of polypropylene (PP). The black plain bonds
point out of the plane of the CC chain while the grey, hatched bonds point to the rear
side.

When the carbon atoms of the structural unit bear identical substituents, only
one configurational isomer exists. This is the case for polyethylene (two hydrogens
on each of the carbons of the structural unit), for poly (-1,1-difluoroethylene) (two
hydrogens on one carbon and two fluorines on the other carbon of the structural unit)
and for poly(tetrafluoroethylene) (Teflon) (two fluorines on each of the carbons of
the structural unit) (figure 5.3.). Polymers of this type are semi-crystalline.

Polyethylene Poly( 1,1 -difluoro-ethylene) Polytetrafluoroethylene

(Teflon)

Figure 5.3. Structural units of different disubstituted polymers.

It should be noted that the zigzag plane conformation used to describe the con
figurational isomers of the vinyl polymers is only found in the crystalline state for
polyethylene. As it will be seen in paragraph 5.2.4., the presence of a bulkier sub-
stituent, such as the methyl group of polypropylene, favours the formation of a spiral
conformation.
Hydrocarbon polymers can also be synthesised with monomers containing dou
ble bonds (diene monomers) as:
94 Introduction to Materials Science

R
I 3 4
(5.10.)
H2p=C C=CH 2

l 2

where R represents a hydrogen atom (H) in butadiene, a methyl group ( C H ) in iso- 3

prene or a chlorine atom (CI) in chloroprene.

These polydienes containing four types of structural units are described in figure
5.4, which represent an example of geometrical isomers. The polydienes make up the
main macromolecules involved in nearly all rubbers. Among these should be cited
natural rubber (poly(cis 1-4 isoprene)) (figure 5.4, 1-4 eis unit) which is the main
constituent of tires. In figure 5.5. is represented a chain segment of an industrial
polybutadiene, synthesized by a radical route and which includes the different struc
tural units of figure 5.4. Polydienes are very flexible chains with very week cohesive
forces between them ( 6.3.5.). This explains why rubber experiences very large,
reversible deformations (several hundred of percent) without breaking.

CH
/
/ \ /\ =
2
2

/ c = c \
3 \ A 4
/
R R CH 2

eis 1-4 unit trans 1 - 4 unit

R R
1 U
H C j:
2
H c=j:
2

CH=CH 2 CHCH 2

3 4 3 4
isotactic or syndiotactic 1 -2 unit
isotactic or syndiotactic 3-4 unit

Figure 5.4. The different structural units for polymers of the polybutadiene type.

Figure 5.5. Short chain segment of an industrial polybutadiene, synthesized by a radical

route and including the different structural units of figure 5.4. A polybutadiene segment
in a common rubber contains approximately 200 structural units. Its fully extended
length is about one but in the unstretched rubber the mean distance between the
chain ends reaches only a few percent of its extended length because the chain is highly
coiled in the amorphous rubber.
 Organic polymers and ceramics 95

Alongside the vinyl polymers and the polydienes obtained by polymerisation of

monomers having one or two double bonds, there also exist polymers containing
atoms (O, S, N, etc.) other than carbon in the main chain. These polymers are usually
synthesized by step growth polymerisations. In step growth polymerisation, the chain
growth proceeds by addition of monomers bearing reactive groups of different
chemical types able to react with each other. The chain builds itself by a succession
of condensation reactions between the functional groups present on the monomer
molecules and on the polymer molecules already formed. The chains remain active
during all the polymerisation process and the molecular weight grows at a slow rate
over a long period of time. At the start of the reaction, only low molecular weight
material (dimers, trimers, etc.) forms.
Step growth polymerisation constitutes an extension into macromolecular
chemistry of condensation and addition reactions developed in organic chemistry. In
a polycondensation, chemical reaction is accompanied by a simultaneous elimination
of a small molecule. In a polyaddition, the simultaneous elimination reaction is ab
sent. This elementary treatment will be limited to polycondensation.
The polycondensates, often known as technical or engineering plastics, have in
general properties which are noticeably superior to those of the vinyl polymers fre
quently identified as commodity plastics. It should also be noted that the polycon
densates are the basis of a series of textile fibres.
A well-know polycondensation reaction is the synthesis of the poly amides,
which involves usually two types of monomer: a di-acid, like hexanedioic acid
(adipic acid) and a diamine, hexane-l,4-diamine (hexamethylene diamine) for exam
ple. Initially, at the beginning of the reaction, dimer molecules are formed by the
addition of diacid molecules to diamines molecules with elimination of water mole
cules.

H O C O (CH ) 2 4 COOH + H N(CH ) NH

2 2 6 2

Hexanedioic acid 1,4-Hexane-diamine

(5.11.)

HOCO (CH )4
2 CONH(CH ) NH 2 6 2 + H 0 2

Dimer

The dimer reacts with a di-acid molecule to form a trimer molecule:

HOCO(CH ) CONH(CH ) NH
2 4 2 6 2 + HOOXCH^COOH
Dimer Hexanedioic acid

(5.12.)

HOCO(CH ) CONH(CH ) NHCO(CH2)4COOH + H 0

2 4 2 6 2

Trimer

or add a diamine molecule, to yield another trimer molecule:

96 Introduction to Materials Science

H N(CH )6NH
2 2 2 + HOCO(CH )4CONH(CH )6NH2
2 2

1,4-Hexane-diamine i| Dimer
(5.13.)

H N(CH2)6NHCO(CH )4CONH(CH2)6NH2 + H 0
2 2 2

Trimer

H N(CH ) NHCO(CH2)4CONH(CH2)6NH
2 2 6 2 +HOCO(CH2)4CONH(CH )6NH2 2

Trimer ,. Dimer

J (5.14.)
H N(CH ) NHCO(CH ) CONH(CH ) NHCO(CH ) CONH(CH ) NH
2 2 6 2 4 2 6 2 4 2 6 2 + H 0 2

Pentamer

On completion of the reaction, a polycondensate of high molecular weight is

formed, which contains between 50 and 100 structural units. The general equation for
the polycondensation reaction is written as:

(n + 1) HOCO(CH ) COOH + (n + 1) H N ( C H ) N H
2 4 2 2 6 2

Hexanedio'fque acid Hexane-1,4-diamine

(5.15.)

( )4[( )6( )4](2)

2 2 2 6 2 +(2n+ 1) H 0 2

Polyamide 6-6

In the polymerisation reaction schematically described in (5.3.), each macro-

molecule is formed in one step and in an extremely short time, in general less than
one second. With polycondensation, however, the macromolecule builds progres
sively by a series of successive condensation reactions. Polycondensation reactions
are much slower than polymerisation and, in general, are not favoured from a ther
modynamic point of view. To obtain a high degree of polymerisation, it is usually
necessary to displace the reaction equilibrium by removing the small ( H 0 ) molecule 2

by distillation.
The polyamides derived from diamines and diacids are often designated two
numbers. The first figure indicates the number of carbon in the diamines and the
second that of the di-acid. Thus, the polycondensation product of hexane-l,4-diamine
(6 carbons) and hexanedioic acid (6 carbons) described above is polyamide 6-6.
Please note that the term Nylon is a registered trademark of the Du Pont company,
although practically, polyamide 6-6 is very often called Nylon 6-6.
In some cases, the amine and acid functions are found on the same molecule. In
this case, only one figure, corresponding to the number of carbons in the monomer
molecule, is used to characterize the polyamide. For example, the polyamide 11 (Ril-
san ):

( C H ) i - C -
2 0
(5.16.)
 Organic polymers and ceramics 97

is synthesised from an aminoacid containing eleven carbon atoms (-aminoundeca-

noic acid):

H N(CH ) -C-OH
2 2 1 0 (5.17.)
In figure 5.6., the structure of several much-used polycondensates are shown.
The chemical composition of polyesters is close to that of the polyamides. Only the
imino (-NH-) group has to be replaced by an oxygen. In contrast to vinyl polymers
(figure 5.2.), polycondensates only exist in one configuration and, in general, are
semi-crystalline.
H O Q

-N-
(a) amide group (b) ester group

(CH )n-
2

(c) polymethylene (aliphatic) group (d) phenylene (aromatic) group

- <CH )
2 4 c NH (CH ) NH
2 6

A
(e) Aliphatic polyamide: 6-6 polyamide (Nylon 6-6) (PA 6-6)
O H

"V ^-

Aromatic polyamide (Kevlar)

(f) Semi-aromatic polyester: poly(ethyleneterephthalate) (PET)

Figure 5.6. Examples of polymers containing hetero-atoms in the main chain (polycon
densates). Each unit contains two sub-units A and B.

5.2.2. Dimensions of macromolecules

Most polymerisation reactions are random processes: the chains do not all grow
to the same length. Synthetic organic polymers are hardly ever substances of
homogeneous molecular weight as are in general organic molecules of low molecular
weight, such as benzene for example (M= 78).
98 Introduction to Materials Science

This leads to the concept of average molecular weight, which can be defined in
several ways. Thus the number average molecular weight M , is equal to: n

=
( )
5 18

where X represents the mole fraction (or number fraction) of the n molecules of
x x

degree of polymerisation and of molecular weightM*. J&is defined by:

=JL- (5.19.)

The weight average molecular weight, M is obtained by statistically weighting

w

the molecular weight as a function of the weight fraction of the macromolecules w x

M =^w M
w x x (5.20.)

where w represents the weight fraction of the n molecules with a degree of polym
x x

erisation and of molecular weight M : x

w x = - ^ - (5.21.)

The weight average molecular weight M is biased by molecules of higher molecular

w

weight and in general:

M >M w n (5.22.)

The relationship M /M is known as the polymolecularity

w n index I and is a
measure of the width of the molecular weight distribution. For a homogeneous
polymer, 7 = 1 . The experimental results of I close to unity are only encountered in a
few special cases. In polycondensation, I is in general equal to or slightly greater than
2. For current vinyl polymers, 2 < I < 5.
M a n d M are technically important parameters. M mostly governs the solid
w n n

state properties: it is necessary that theM value exceeds a critical value, M between
rt c

10 000 and 50 000 so as to obtain a polymer with satisfactory mechanical properties.

The value of M depends on the nature of the polymer. The lower limit of M =
c c

10 000 is generally valid for the poly condensates while the upper limit, M = 50 000, c

is mainly characteristic of vinyl polymers.

The rheological behaviour of molten polymer, which influences polymer pro
cessing, is a function of M . The viscosity of organic polymer melts is as high as that
w

of inorganic polymers ( - 10 Pas), and normally varies as a function of Af".

4 5

Often, the objective is to produce polymers with a polymolecularity as low as pos

sible so as to have the lowest possible melt viscosity for easy production while
keeping the value of M high enough to exceed the critical M value and maintain
n c

valuable properties. However, in certain cases, polymers with a high polymolecular

ity index are wanted in order to increase the melt strength and facilitate certain trans
formation processes such as blow-moulding.
 Organic polymers and ceramics 99

5.2.3. Main Types of Polymer Materials

Two main classes of polymer materials exist.

The thermoplastic polymers are made up of limited size, linear or
slightly branched macromolecules. The structure of this class of poly
mers is illustrated in figure 4.10. (a) and (b) for polyethylene. This
category of materials is made up of macromolecules, which as a general
rule are soluble in organic solvents able to disjoin one polymer chain
from one another. At high temperatures (from 175 - 350 C maximum
depending on the polymer type) thermoplastic polymers pass into the
molten state.
The reticulated polymers, which have a three-dimensional structure, are
made of cross-linked polymer chains as shown in figure 4.10. (c) for
polyethylene. In fact cross-linked, polymers consist of a single macro-
molecule. This super macromolecule reaches macroscopic dimensions.
Thus, after vulcanisation, the rubber chains constituting a car tire make
only one macromolecule which has practically the same dimensions as
the tire. The cross-linked polymers form the class of thermoset poly
mers as the cross-linking is generally carried out at high temperature
(about 1 5 0 - 2 0 0 C).

In the solid state, below about 100 - 150 C, thermoplastic polymers exist either
in an amorphous state as an organic glass or in a semi-crystalline state. The ability of
polymers to crystallise is in direct relation to the molecular structure fixed during
synthesis. In order for a polymer to crystallize, its chain has to be of a very regular
structure (isotactic or syndiotactic).
Polymers very seldom form monocrystals. They are in general characterized by
a microcrystalline structure dispersed in an amorphous phase', this will be described
in chapter 10. The crystal fraction varies strongly with the polymer structure. Thus,
very flexible linear polyethylene (figure 4.10. ( a ) ) achieves a degree of crystallinity
of greater than 85 %. Because of the presence of branches, the degree of crystallinity
of branched polyethylene (figure 4.10. (b)) does not exceed about 50 %. In rigid
chain polymers, such as poly(ethyleneterephthalate) (PET) (figure 5.4. (g)), the
crystallinity degree is limited to 30 - 40 %. Because of a difference of refractive
index between amorphous and crystalline phases, semi-crystalline polymers are usu
ally not transparent, but translucent. Their semi-crystalline structure confers on them
greater strength and resistance to temperature and to chemical agents such as organic
solvents. In contrast, the organic glasses consist of a single phase and, consequently,
they are transparent. The organic glasses are frequently used for their optical proper
ties in optical devices as contact lenses (sect. 5.5.).
In thermoplastic polymers, the intermolecular bonding forces are secondary
valence forces, and the transition to the melt takes place at temperatures much lower
than those necessary for metals or ceramics. With a moderate increase in tempe
rature, thermoplastics pass into the molten state and can be formed by continuous
fabrication processes such as extrusion, injection moulding etc. This physical
characteristic allows thermoplastic polymers to be processed at relatively low tem
peratures (175-350 C) with very high production rates for finished articles. It is this
100 Introduction to Materials Science

ability to be processed easily under energetically favourable conditions which has

largely contributed to the economic success of polymer materials. In principle, these
forming processes can be repeated many times without altering the molecular
structure or the technical properties of the polymer. In theory, thermoplastic poly
mers are easy to recycle but these recycling possibilities are drastically limited by the
incompatibility of polymers of different chemical structure, which cannot mix in the
melt.
The cross-linked polymer materials are sub-divided into two categories: elas
tomers (rubbers) and highly cross-linked thermosets.
Elastomers are obtained starting from linear polymers in general having a mo
lecular weight between 100000 and 500 000 and characterized by very weak inter
chain secondary bonds (4 k J m o l ) . At ambient temperatures, the non-vulcanised
-1

rubbers are very viscous liquids. To obtain a solid exhibiting high elastic behaviour
in tension or shear, a certain number of cross-links (covalent bridge bonds) are ne
cessary. These confer a three-dimensional network structure to the material and en
sure a reversible mechanical deformation (chapter 6.). The uncross-linked ('green')
material is first shaped into a product. The cross-linked bonds, which fix the shape of
the fabricated product, are introduced afterwards by a chemical reaction known as
vulcanisation at the final stage of the fabrication. By way of an example, figure 5.7.
illustrates the structure of a cross-linked bond in c/s-l,4-polyisoprene (natural rubber;
figure 5.4) vulcanised with sulphur. In this case, covalent disulphur cross-links (-S -) 2

are mostly formed.

Figure 5.7. Structure of the cross-links (disulphur cross-links) in the eis 1-4 polyisoprene
(natural rubber) vulcanised with sulphur as is found in car tyres.

A very important technical characteristic of rubber materials is their conside

rable extensibility, which reaches 600 to 700 %, that is some 6 to 7 times their origi
nal length. This is directly related to the presence of inter-chain secondary bonds,
which allows large chain segments to move under the action of weak external forces.
In elastomers, the cross-link density, i.e. the number of cross-links per structural unit
of the polymer material is low. Thus in the elastomers currently used for car tyres,
there is typically about one cross-link per 100 structural units. As reported in chapter
6, the elasticity modulus of the rubbers is a direct function of the cross-link density.
 Organic polymers and ceramics 101

In the highly cross-linked thermoset polymers, the cross-link density reaches a

level some 10 to 100 times higher than in the case of elastomers. In these materials,
the cohesive energy between the chains is generally very much higher (20 -
30 kJmol" ). Because of their physicochemical characteristics, thermoset polymers
1

in general achieve mechanical strength levels (table 6.2.) and heat resistance signifi
cantly higher than in thermoplastics.
On the other hand, thermosets processing is much slower and less simple than
that of thermoplastics. The forming of a thermoset polymer must take place before
the cross-link formation. This cross-linking is a chemical reaction analogous to the
vulcanisation reaction for rubbers and is not instantaneous. Due to their three-
dimensional structure, the thermosets are insoluble and will not pass into the liquid
state. Therefore, these materials are not recyclable. A current example, easily
experienced, is that of epoxy adhesives which are used to glue. These epoxy
adhesives are two component systems composed of a resin, which is a low molecular
weight linear polymer, and a hardener able to react with the resin to form a three-
dimensional network. At the moment of application, the two components are mixed
and the adhesive is applied. It then hardens (cross-links) over a period of approxi
mately half an hour to a day. After hardening, the adhesive has become insoluble and
cannot melt.
A large number of thermoplastics and thermosets with a high cross-link density
are brittle materials. It will be seen in chapter 10 how it is possible to obtain polymer
materials with high mechanical strength and high impact resistance by combining
thermoplastics with elastomers.

5.2.4. Spatial configuration of crystalline polymers and lattice structure

In the crystalline polymer lattice, the chain adopts a configuration corresponding
to a minimum potential energy. In figure 4.8. (b) (exploded view) and in 5.8. (a)
(condensed view) the polyethylene chain is in a trans zigzag conformation. All the
chain atoms are co-planar with valence angles of 109.5 and a C - C distance of 0.154
nm. In this conformation, the hydrogen atoms are disposed one above another. The
distance of 0.253 nm, which separates two adjacent carbon atoms, is slightly greater
than the van der Waals' diameter of the hydrogen atom (0.234 nm). Therefore, there
is no significant steric interaction between these hydrogen atoms. All this favours a
compact packing of the atoms in the crystal.
For the other vinyl polymers, because of the hindrance introduced by bulkier
substituents, it is not possible to maintain the same trans zigzag structure. For exam
ple, in polytetrafluoroethylene, the hydrogen atoms are replaced by fluorine with a
larger van der Waals' diameter (0.270 nm). The trans zigzag configuration is no
longer possible without introducing serious steric strain, and, instead, the chain takes
a helical conformation. This structure, illustrated in figure 5.8 (b), produces a con
siderable stiffening of the polymer, which significantly increases the melting point
which passes from 130 C for linear polyethylene to 330 C for poly(tetrafluoroethy-
lene). The exceptional temperature resistance of the latter polymer comes mostly
from this structural difference. It is commercialised under the trade name of
Teflon, and is extensively used as a non-stick coating for pans.
102 Introduction to Materials Science

(a) (b)

Figure 5.8. Structure of the polyethylene (a) and poly(tetrafluoroethylene) chains in the
crystal. Note the formation of a spiral in the case of poly(tetrafluoroethylene) while in
polyethylene the carbon atoms are arranged in a zigzag in a reticular plane (after Gordon,
1963, and Bunn and Howells, 1954).

Isotactic polymers such as polypropylene are able to adopt a trans-trans or trans-

gauche conformation (figure 5.9.). Because of the important steric effect of the
methyl groups which have a van der Waals' diameter of 0.4 nm, the trans-trans con
formation, which would lead to a zigzag structure is less stable. Isotactic polypro
pylene chain takes up a trans-gauche helical conformation in the crystalline lattice
(figure 5.10.).

(a) trans-trans (b) trans-gauche

OCH,

Figure 5.9. Trans-trans and trans-gauche conformations of polypropylene. In the case of

polypropylene, the trans-gauche conformation minimizes the interaction between the
non-bonded groups. This explains the spiral conformation of the polypropylene chain in
the crystal. For reasons of clarity, the hydrogen atoms in the chain are not shown.

A polymer containing hetero atoms (polycondensates) takes frequently a trans

zigzag conformation in the crystal. In the polyamides, this compact structure is stabi
lised by very strong hydrogen bonds formed in the reticular plane containing chain
atoms. This lamellar structure (figures 2.21. (b) and 5.11.) leads to the high me
chanical strength and thermal resistance of aliphatic and aromatic polyamides. Aro
matic polyamides such as poly(/>phenylene terephthalamide), commercialised by Du
Pont under the trade name of Kevlar, form extremely stiff fibres used to reinforce
composite materials (chapter 16.). Their stiffness is about half of that of carbon
fibres. Kevlar fibres are used in bullet proof jackets and as a replacement for asbestos
fibre in certain brake linings.
 Organic polymers and ceramics 103

Figure 5.10. Helical structure of crystalline isotactic polypropylene. Projections parallel

(a) and perpendicular (b) to the helix axis (after Wunderlich, 1973).

Figure 5.11. Illustration of the lamellar structure of poly(p-phenylene terephthalamide)

(Kevlar) in the crystal. For reasons of clarity, the hydrogen atoms are not shown. The
hydrogen bonds are shown with dotted lines (after Tadokoro, 1984).

The chain packing in the crystal lattice is strongly dependent on the conforma
tion of the chains. The maximum compactness is obtained when the chain adopts a
trans zigzag configuration as in polyethylene and polycondensates. The polyethylene
crystal (figure 5.12.) attains a compactness factor of 0.80, calculated on the basis of
the van der Waals' diameters of the atoms. This compactness factor is greater than
that found for the hep and fee metals structures, since the presence of covalent bonds
leads the van der Waals' sphere to interpenetrate. The polyamide 6-6 compactness
factor also reaches a very high value (0.76).
104 Introduction to Materials Science

Figure 5.12. Structure of the orthorhombic unit cell of crystalline polyethylene; represen
tation as an exploded view (after Gordon, 1963, and Bunn 1945).

In the case of a helical structure, the packing is generally lower. The compact
ness factor of isotactic polypropylene is 0.66 only. Certain helical structures do how
ever achieve a high degree of compactness. This is, in particular, the case for
poly(tetrafluoroethylene) which is characterized by a compactness factor of between
0.78 and 0.80.

5.3. Ceramics
5.3.1. Classification and general characteristics
Ceramics are among the oldest materials used by man, since the natural cer
amics (rocks) are the essential constituents of the earth's crust. Their name come
from the Greek Keramos, which signifies baked potters clay. Clay was the raw mate
rial used to make the first synthetic ceramics (pottery). Kaolin (china clay), a
hydrated aluminium silicate (Si 0 )Al (OH)4, is utilized to produce porcelain.
2 5 2

Ceramics are equally among the most recent materials synthesized by modern
technology: ferrites ( M g F e 0 ) for magnets, ceramics with high strength and heat
2 4

resistance (SiC, S i N ) for internal combustion engines, bioceramics (A1 0 ), high

3 4 2 3

tenacity ceramics ( Z r 0 + Y 0 ) , barium titanate ( B a T i 0 ) used as a dielectric in

2 2 3 3

condensers as well as different oxides (ZnO, S n 0 ) employed as detectors of in

2

flammable gases.
The general characteristics of ceramics are their refractory character (resistance
to very high temperatures), their hardness, their chemical inertness and in general,
their very low thermal (in polycrystalline form) and electrical conductivities. Other
specific properties can also be conferred on them such as ferrimagnetism, semi-
conductivity or superconductivity. Ceramics are not generally ductile and have a low
resistance to thermal shock. These two major deficiencies seriously limit their
 Organic polymers and ceramics 105

applications. As a result from recent research, the fracture toughness and shock-
resistance of zirconia ( Z r 0 ) has been substantially improved.
2

5.3.2. Silicates
Natural or synthetic silicates make up an important class of ceramic materials.
The basic element of their structure is the S i 0 tetrahedron (figure 5.13 (a)). As
4

mentioned in paragraph 4.4.2, in the simplest silicates, the S i 0 tetrahedron is present 4

in the form of ions (Si0 ) ~ associated with metallic cations which form strong ionic
4
4

bonds with the oxygen. Two examples of this type of silicate: forsterite ( M g S i 0 ) 2 4

and Zircon (ZrSi0 ).

4

(Si0 ) ' 4
4
(Si 0 ) -
2 7
6

1 independent tetrahedron 2 tetrahedra associated

at an apex

(c)

Tetrahedra associated in a simple chain,

pyroxene type (Si0 ) " 3
2

Tetrahedra associated in a double chain

Amphibole type(Si O ) ~
4 u
6
106 Introduction to Materials Science

(e)

Tetrahedra associated in a lamella

(phyllosilicate)(Si 0 ) "
2 5 2

Figure 5.13. Structure of the silicates.

(a) Silica glass amorphous structure

(b) Crystalline silicates ordered structure (cristobalite).

The tetrahedra are associated in hexagonal cycles
as in phyllosilicates.
Si

Figure 5.14. Two-dimensional illustration of the structure silicates. For reasons of clar
ity, only three of the four oxygens associated with silicon are shown.

The S i 0 tetrahedra are able to share a certain number of oxygen atoms to form
4

structures, which are either of small dimensions, or may be of considerable size

(polymers). In disilicates, two S i 0 tetrahedra share an oxygen and form the (SizO?) "
4 6
 Organic polymers and ceramics 107

anion (figure 5.13. (b)). As already seen, (figure 4.11. (c)), S i 0 tetrahedra can share
4

two oxygens to form simple linear chain structures (pyroxene) (figure 5.13. (c)). S i 0 4

tetrahedra can share three oxygen atoms to make structures with double chains (am-
phibole - figure 5.13. (d)) or in layers (figure 5.13. (e)) forming the tetrahedral layer
(T) of the phyllosilicates. Finally, silicates with a three-dimensional structure are
found in silica ( S i 0 ) which exists both under amorphous (figure 5.14. (a)) or
2

crystalline (figure 5.14. (b)) structures.

Tetrahedric layer

Octahedric layer

Tetrahedric layer
0,922 nm
Secondary bonds

Elementary
lamellae

Si Mg OH
Figure 5.15. Structure of talc (Si OioMg (OH)2). The projection follows a plane perpen
4 3

dicular to the elementary lamellae. This should be compared with figure 5.13 (e), which
represents a projection parallel to the tetrahedral layer.

In certain cases, the architecture of the silicates becomes very complex. So in the
phyllosilicates, the metallic cations (Mg , Al ), with a coordination degree of 8, are
2+ 3+

arranged in layers (octahedral layer O). In figure 5.15., the structure of talc
(Si Oi Mg (OH) ) is shown. In this phyllosilicate, the elementary lamella is made of
4 0 3 2

two tetrahedral layers associated with an octahedral layer. This layer contains a
number of OH~ ions, which complete the octahedral coordination of magnesium.
Two main families of phyllosilicates exist: the TOT phyllosilicates made up of
two tetrahedral layers and one octahedral layer and for which the model is talc, and
the TO where the elementary lamella consists of a tetrahedral layer associated with
an octahedral layer. Kaolin, S i 0 A l ( O H ) , which is the basic constituent of clay, is
2 5 2 4

a phyllosilicate TO and so is made up of a tetrahedral layer and an octahedral layer.

The bonding between the elementary lamellae of the phyllosilicates occurs via
secondary valence forces. This explains their lamellar structure and the ease with
which these minerals can be cleaved. Talc is a solid lubricant. Water layers that
penetrate between the elementary lamellae are the origin of the plastic properties of
the clays (kaolin). This characteristic is notably exploited during the shaping of tra
ditional ceramics starting from a water-based paste (barbotine).
108 Introduction to Materials Science

Aluminium forms with oxygen a polar covalent bond very similar to the Si-O
bond and is able to partake in the tetrahedral layer where it replaces the silicon. It
also plays the role of metal (Al ) by participating in the octahedral layer where it is
3+

characterized by a coordination number of 6. In the case of substitution of silicon by

aluminium in the tetrahedral layer, the (Si0 ) ~ anion is replaced by the (A10 ) ~
4
4
4
5

anion. This goes together with the replacement of certain cations in the octahedral
layer by other cations having a higher valence so that the electrical neutrality of the
mineral is maintained. Silica and the silicates, together with hydrated alumina and
certain oxides (CaO, MgO, N a 0 , K 0 ), form the raw materials used for the fabrica
2 2

tion of traditional ceramics (porcelain, bricks, pottery), mineral glasses ( 5.3.3.) and
for cements( 5.3.4.). Silicates like chrysotile (asbestos) were widely used as rein
forcing fibers in asbestos cements and as insulation but, at present, their usis highly
restricted because of their carcinogenic characteristics.

5.3.3. Mineral glasses and glass-ceramics

The mineral glasses, characterized a three-dimensional polymer structure, make
up a very important class of technological ceramic materials.
The simplest glassy structure is that of amorphous silica which can be obtained
by the cooling of molten silica. This is obtained by the fusion of quartz (1610 C),
which is one of the allotropic forms of crystallised silica. As shown in figure 5.14 (a),
amorphous silica possesses a three-dimensional structure consisting of S i 0 4

tetrahedra attached at the vertex, each oxygen being shared by two tetrahedra. The
high viscosity of molten silica is a consequence of its three-dimensional structure.
The phenomenon of flow in this three-dimensional network occurs because a number
of stressed bonds undo and reform in a relaxed position. This flow by exchange
reaction is described in the following equation:

\ /- - \ /-
- O S~"i ~ S~"i ~ - + - ~
>S Si
i
SS ii
O-
- oy 1 2\ ~
2
o- c / 3 1 4 \
V
(5.23)

- \ 0
" " \
-OSiSi- + -SiSiO-
- O O-
1 3
-</ V - 1 4

This exchange reaction is catalysed by a small number of silanol groups, pro

duced by the hydrolysis of the fused silica with traces of water:

" \ - \
- O - S i OSi - + H 0
2 ^ 2 -O^SiOH (5.24.)
-o O- -O

The three-dimensional structure of fused silica makes its crystallization difficult and
slow.
The addition of metallic cations in the form of oxides (modifying agents or
fluxes'. N a 0 , K2O, and CaO) to the fused silica leads to a considerable reduction in
2

viscosity. Thus the addition of 2.5 mole % of K 0 reduces the viscosity of silica at
2
 Organic polymers and ceramics 109

1750 C from 2-10 to 200 Pas. The metallic cations react with the three-dimensional
silica and decrease the crosslink density:

/
-OSi- S r ^ -l o - + 2 - O S i - 0~K (5.25.)
+
2

-O o- -O
Above a certain proportion of modifying agent, the three-dimensional structure of the
silica is broken and isolated islands linked to each other by modifying cations are
formed. These ionic bonds allow more mobility and facilitate the reorganization of
the molten glass and generate a significant reduction in the viscosity. A schematic
representation of the structure of a mineral glass formed by the addition of N a 0 is 2

given in figure 5.16.

Figure 5.16. Two-dimensional representation of the structure of a mineral glass. The

small dark spheres represent silicon atoms, the large white spheres the oxygen atoms and
the small white spheres the Na ions (after Greaves, 1982).
+

The most common glasses, such as, for example, window glass, are obtained
from a mixture of soda ash (Na 0), silica ( S i 0 ) and quick lime (CaO) corresponding
2 2

approximately to the formula ( N a 0 ) ( C a O ) ( S i 0 ) . For special applications, other

2 2 5

system-forming oxides can be added such as boron oxide (B Os) used in the manu
2

facture of glasses with low thermal expansion (Pyrex).

The glass-ceramics (also called vitro-ceramics or pyrocerams) constitute a
relatively new class of ceramics obtained by the crystallization (devitrification) of
already formed glass objects. Certain mineral glasses crystallize when they are
submitted to precise thermal cycles favouring crystal nucleation and growth. The
glass-ceramics have a fine-grained, microcrystalline structure dispersed in a glassy
phase. These materials, characterized by an extremely low thermal expansion
coefficient, are known for their exceptional resistance to thermal shock. They are
used in a series of important technical applications: electrical insulators, components
of pumps, supports for telescope mirrors, etc. It should be noted that because of the
refractive index difference between the glassy and crystalline phases, in general,
glass-ceramics are opaque like semi-crystalline organic polymers.

5.3.4. Concrete
Concrete, currently the most used building material, is a complex mixture of
gravel, sand (quartz), which are natural ceramics, and of a hydraulic cement (binder).
110 Introduction to Materials Science

The binding cement represents some 20 to 40 % of the concrete volume. If the gravel
is left out of the mixture, a mortar is obtained.
The hydraulic cements (binders) are ceramic materials, which have the property
of hardening under water and afterwards of resisting the dissolving action of water.
The first hydraulic cements go back to antiquity and were used by the Romans. They
notably discovered that certain eruptive rocks (puzzolanas) react with calcium
hydroxide (Ca(OH) forming a product practically insoluble in water which
2

constitutes a hydraulic cement. The discovery of the hydraulic cements (Portland)

currently used in civil engineering is relatively recent, going back some 150 years.
The raw materials used for the production of Portland cement consist of a finely
ground mixture of calcium carbonate ( C a C 0 ) and clay (kaolin) or marl (calcareous 3

clay) which are introduced into a rotary furnace. The material is taken progressively
to 1400 - 1450 C, where a partial fusion of the material (clinkering) occurs. The
product obtained is a complex mixture of calcium silicates and aluminates and is
made up of hard, blackish granules known as clinker.
It has the following composition:
C 2 S: (CaO) Si0 , bicalcic silicate (-20 % ) ;
2 2

C 3 S: (CaO) Si0 , tricalcic silicate (50 - 70 % ) ;

3 2

C 3 A: (CaO) Al 0 , tricalcic aluminate ( 6 - 1 2 % ) ;

3 2 3

C 4 AF: ( C a O ) A l 0 F e 0 , tetracalcic aluminoferrate (-10 % ) .

4 2 3 2 3

These different constituents have been identified in the clinker by optical mi

croscopy. It should be noted that the tetracalcic aluminoferrate C AF does not have a 4

constant chemical composition, being a mixture of two products C F ( ( C a O ) F e 0 ) 2 2 2 3

and C A ((CaO) Al 0 ) in variable proportions.

2 2 2 3

The hydraulic cements are a mixture of finely ground clinkers and approxi
mately 5 % gypsum ( C a S 0 , 2 H 0 ) introduced as a setting regulator. The hydraulic
4 2

cement, with gravel and sand added, is mixed with water and applied as a paste. It
should be noted that air is also a constituent of concretes, introduced at the time of
mixing. The proportion is about 3 % if it is not introduced deliberately. The presence
of air improves the resistance to freezing but reduces the mechanical strength.
The setting and the hardening of cement produce several hydrated solid phases
of which the volume is greater than that of anhydrous cement. The constituents of the
cement progressively dissolve in the water and reprecipitate in the form of hydrated
phases, which gradually take over the space occupied by the water and thus build up
a rigid, porous structure. The reprecipitation exothermic (AH = 400 -800 kJg" ) reac 1

tions start spontaneously as soon as the anhydrous cement comes into contact with
water. The hydration reaction for Portland cement is complex and discussion here
will be limited to the C S hydration, which is approximately represented by the fol
3

lowing equation:
2(CaO) Si0
3 2 + 8Hp-> (CaOHSiO^f^O^ + 3 Ca(OH) 2 (5.26.)
C S
3 C3S2H5

The tricalcic silicate C S forms two hydration products, a polysilicate with an

3

approximate formula C S H close to that of a natural mineral called tobermorite, and

3 2 5

calcium hydroxide (portlandite). The bicalcic silicate (C S) gives the same hydration 2

products. The calcium hydroxide crystallizes in the form of hexagonal platelets. The
 Organic polymers and ceramics 111

hydrated silicate forms a slightly crystalline gel composed of fine needles which
develop in the interstices between the granules of Portland cement. The start of ce
ment setting, which goes with an increase in the paste consistency and is
accompanied by a progressive filling of the interstices, is illustrated in figure 5.17.
The development of the tobermorite produces an increase in mechanical strength.

Mixing Water forms a capillary network

between the granules.

Before setting The tobermorite gel appears on the

surface of the cement granules.

Setting The tobermorite invades the interstices

between the granules. The paste thickens.

The tobermorite gel gradually fills the

Interstices between the granules. The
Start paste acquires some mechanical strength.

The mechanical strength increases. The

gel continues to develop over a very long
Consolidation time period in the presence of water

Figure 5.17. Schematic representation of setting of Portland cement (after Huet, 1993).

In the presence of water, the setting of cement continues for very long time
(several years). The microstructure continues to modify with a gradual increase in
mechanical strength. The cements have mechanical properties about four times lower
than those of fused silicates such as glasses. This weakness in mechanical properties
is a consequence of the very heterogeneous structure of the porosity of these mate
rials. This results from the shrinkage induced by the hydration reaction, which con
sumes water. The porosity varies with the amount of added water and is of the order
of 10 % for present cements. The open porous network of concrete is interconnected
and allows water to circulate within the material. As the tensile strength of cements is
very low, they are used preferably in those applications where the material is under
compression (buildings, dams, etc.) or they are reinforced with steel bars or rods or
with fibres.
Besides the Portland cements, other types of binders exist. Phosphate based
cements are interesting for use as orthopaedic materials. The phosphato-calcic ce
ments, which are biocompatible, have re-absorption and osteoconducting charac
teristics favouring a rapid formation of a new bone, which progressively replaces the
consolidating cement.
112 Introduction to Materials Science

It should also be noted that alongside concrete, there are other cement types
particularly resin cements using a polymer as binder.

5.3.5. Technical ceramics

Because they use the Carnot energy cycle, engineering devices operate at
increasingly higher temperatures. Up to now, because of their strength ductility and
their ability to withstand severe thermal shocks, metallic alloys have been used
almost exclusively to fulfil functions where significant mechanical stresses are
involved. Because of their poor toughness, the ceramics have remained confined to
uses such as coating layers or in the form of weekly stressed parts.
The limit for using metallic alloys for high temperature applications in an oxi
dizing atmosphere such as a gas turbine (about 1100 C) is close to being reached in
the future and there will be a tendency to turn towards ceramics because of their
refractory character.
The ceramics foreseen for these technical applications at high temperature are no
longer only ceramics based on oxides. During the course of the last decades, a series
of engineering ceramics has been developed: zirconium oxide, silicon carbide,
silicon nitride, etc. These latter compounds ( S i N , SiC), which are no longer based
3 4

on oxides, are obtained by elementary chemical reactions taking place at very high
temperature. Silicon nitride is obtained by a direct synthesis reaction between silicon
powder and gaseous nitrogen at 1200 C:

3Si +2N 2 - Si N
3 4 (5.29.)
solid gas solid

Silicon carbide is synthesized by the reduction of sand ( S i 0 ) with coke (carbon) 2

in an electric furnace at a temperature of 2700 C:

Si0 2 + 3C ~ SiC + CO (5-28.)
solid solid solid gas

These ceramics are formed by chemical sintering at high temperature, starting

from a powder premix. Thus, in the case of silicon carbide, a mixture of SiC, graphite
and Si powders is heated to a temperature above the melting point of silicon
(1414 C). The in situ formation of SiC welds the granules of primary carbide.
For silicon nitride, the sintering is carried out at 1700 - 1800 C, with a small
amount of oxides ( Y 0 , MgO, A1 0 ), which form a liquid phase. These oxides react
2 3 2 3

with the S i 0 forming a tin layer of liquid silicate on the surface of the Si N parti
2 3 4

cles. The S i N particles (a form) dissolve gradually in the liquid phase and repre-
3 4

cipitate into a crystalline/? form. The liquid phase solidifies on cooling and serves as
a joint between the S i N granules. In this way, the Si N transformed into its form
3 4 3 4

has a significantly higher tenacity.

The ceramics are resistant to very high temperatures (1300 - 1500 C) and are
extremely hard (after diamond, boron nitride and silicon carbide are the hardest
material). Silicon carbide is characterized by a thermal conductivity comparable to
that of certain metal alloys. This is rather exceptional as the majority of
polycrystalline ceramic are insulating materials. On the other hand, silicon nitride,
with a thermal conductivity some fifteen times less than that of silicon carbide, is an
electric insulator. These materials also have low densities (2.5 to 3.6 g e m ) . As -1
 Organic polymers and ceramics 113

already mentioned in paragraph 5.3.1, the use of these ceramics remains limited,
mostly because of the nature of their predominantly covalent chemical links which, at
low temperature, hinder all dislocation movement (essentially, zero plasticity).

5.4. Summary and conclusions

Two main types of non-metallic materials exist: organic polymers and ceramics.
In general, these materials have an atomic organisation more complex than metals.
The organic polymers, which are formed by the repetition of a basic building
block (constitutional repeat unit), are composed of long chains (polymers) of mainly
carbon atoms. In organic polymers, the cohesion between the polymer chains is
brought about by weak bonds (secondary bonds), while the bonding between the
atoms of the chain is covalent and strong.
Two main types of organic polymers exist: thermoplastics made up of macro
molecules of finite size, and thermosets with a cross-linked tridimensional structure.
These cross-linked polymers form a single macromolecule of macroscopic dimen
sions.
Depending on their chemical structure, the thermoplastics exist in the solid state
either in an amorphous glassy or semi-crystalline form. By raising the temperature,
these materials melt. The thermosets are infusible polymers with a structure that is
nearly always amorphous. A distinction has to be made between thermoset polymers
with a high degree of cross-linking which are rigid, brittle materials, resistant to high
temperature, and the flexible elastomers (rubbers) characterized by extremely weak
secondary bonds and a much lower degree of cross-linking (about one cross-link per
100 structural units). These are able to withstand very large reversible deformations.

Organic polymers have a thermal resistance and mechanical strength

significantly lower than those of metals and ceramics. This behavior re
sults from the fact that organic polymer chains are linked together by
secondary bonds 10 to 100 times weaker than the covalent bonds bind
ing the atoms in the chain.
Structures containing carbon become unstable at temperatures 300 -
400 C, in an oxidizing atmosphere (chapter 15.).

Despite their limited thermal and mechanical strength, synthetic organic mate
rials are proliferating at a high rate because of their low density and their easy proc
essing. The costs of energy required for their manufacture are also much lower than
those associated with metals and ceramics produced at much higher temperatures.
Ceramics exist under either a crystalline (MgO) or an amorphous form (mineral
glasses). The structure of ceramics is very complex involving all types of bonding:
ionic, polar covalent and covalent bonds, and much less frequently, metallic and
secondary bonds. A large number of ceramics are metallic oxides. In a certain num
ber of ceramics (silicates for example), linear, two- and three-dimensional polymer
structures are encountered. In ceramics, the bonding between the polymer structures
is in general realized by metallic cations. This characteristic leads to high mechanical
strength.
114 Introduction to Materials Science

Many ceramics are stable up to very a high temperature above 1000 C.

Technical ceramics such as silicon nitride and carbide, zirconium oxide have high
thermal resistance and mechanical strength and are destined for very high
temperature engineering applications. These ceramics are generally shaped using
powder technology (sintering).
With the exception of a certain number of thermoplastics, non-metallic materials
are generally solids showing brittle failure.

5.5. Illustrative example: a contact lens -10 mg to correct sight

At the turn of this millennium, about 60 million people use contact lenses. In
theory, these provide a greater comfort in correcting sight. Among the main advan
tages are a wide field of vision and the absence of misting. To these should be
added the ease of using this type of optical appliance when practicing a range of
activities such as sport, photography, microscopy, etc. but contact lenses are not
always well tolerated and multifocal lenses are not feasible.
Contact lenses are small precision items. From one kilogram of material, some
100000 pieces are produced. Their price per kilogram attains extreme values:
10 $/kg for disposable lenses and for those with long lifetime (1.5 years), 7-10 $/kg
5 6

which is more than the price of most precious metals, gold in particular.
Leonardo de Vinci proposed the principle of a contact lens in 1508, but it is only
a little over 50 years since, the use of this type of correcting lens has developed. It is
not surprising that they did not appear earlier, since their manufacture demands
extremely sophisticated materials.
The first attempt to produce a contact lens was made at the end of the 19 th

century using mineral glass, but these were just a curiosity, and not at all a
commodity. It was only in the 1940s that the first contact lenses were made involving
rigid plastics such as poly(methylmethacrylate) (PMMA or Plexiglas). Since that
time, the materials have not ceased to diversify and an important stage was reached
in the 1960s with the appearance of the soft lens.

Eyelid Light rays

Cornea
Retina

Aqueous, Macula
humor
Contact / Iris'
lens
Eyelid -
Cilary muscles Optical nerve

Figure 5.18. Representation of an eye wearing a contact lens (after Franz, Bauman and
Thompson, 1993).
 Organic polymers and ceramics 115

Contact lenses have a diameter of about 10-15 mm and their thickness at the
centre can be less than 50 (figure 5.18.). The contact lens fulfils first of all an
optical function related to the transparency and the high refractive index of the
polymer material used. The higher the refractive index, the greater is the refraction of
the incident light beam and the greater is the degree of optical correction for a given
geometry. The material has also to fulfil important biological functions. It must be
chemically inert, physiologically compatible and non-toxic. The fixation of the lens
onto the eye depends on the stability of the tear film formed at the interface between
the cornea and the lens (figure 5.19.). This increases when the wetting of the lens by
the tear fluid is high. Other elements have to be taken into account in the choice of
the material. In particular, it has to be adapted to profitable mass-production of high
precision parts.
It is crucial that the material of which the lens is made allows the passage of
atmospheric oxygen to the cornea (figure 5.19.). In effect, the cornea does not have
any blood vessels and takes the oxygen necessary for its metabolism directly from
the atmosphere. Any reduction in the oxygen supply leads to a swelling (oedema) of
the cornea. Thus the closing of the eyelids during sleep, which reduces by a factor of
three the supply of oxygen to the cornea, leads to it swelling by 5 - 6 %. Swelling
greater than this causes problems of vision. It is therefore absolutely indispensable,
both in the choice of material and the conception of the lens, to maximize the trans
mission of oxygen. Hence, materials highly permeable to oxygen are to be chosen
and the thickness of the lens should be minimized.

Figure 5.19. Mechanism for the diffusion of oxygen to the cornea (after Franz, Bauman
and Thompson, 1993).

Two types of contact lens exist:

Rigid lenses are relatively small (diameter of 8 - 10 mm).

Soft lenses with a larger diameter ( 1 3 - 1 5 mm diameter). Because of
their size, the soft lenses extend permanently under the eyelids.
Currently, 85 % of contact lenses used are soft lenses, which are
noticeably more comfortable.

The first polymer rigid lenses were made from poly(methylmethacrylate)

(PMMA - Plexiglas ), (figure 5.20 (a)), an organic glass with excellent optical char
acteristics, easy to process but with low oxygen permeability. In introducing mono
mers containing siloxane and fluorinated groups into the chain of poly(methyl-
116 Introduction to Materials Science

methacrylate) (figure 5.20. (b)), the permeability is multiplied by 750. The majority
of rigid lenses are fabricated with this type of material.

CH 3

-cc
H
2 I OCH 3

(a) Polymethylmethacrylate

CH 3 CH 3 CH 3

C- I CC- CC-
H - c ,OCH
2
3
H I 2
H21 .OR
2

(1) (2) (3)

CH CH3 3

I I R = ~C-CF
-SiSiCH 2 3
3
?2
I I
CH CH3 3

(b) Methylacrylate (1), disiloxanemethacrylate (2), pentafluoroethyl (3)

copolymer

CC- -C C
H I 2 H J 2

c=o
I H C 2
X
C = 0
OCH CH OH2 2 \ /
H C CH2 2

(4) (5)

(c) 2-hydroxyethyl (4) and N-vinylpyrrolidone copolymer

Figure 5.20. Main polymer materials involved in the production of rigid (a), (b) and soft
(c) lenses.

Rigid lenses, with smaller dimensions than those of the soft lenses, do not
extend permanently under the eyelids. At each blinking of the eye, the thicker rigid
lens compresses the tear film situated between the cornea and the lens and renews the
oxygen contained in the tear fluid. This process, which is comparable to an oxygen
pump, is the predominant mechanism for the transmission of oxygen in the case of a
lens made from PMMA.
Silicone rubber was the first material used to make a soft lens. This material
consists of cross-linked poly(dimethylsiloxane) chains (figure 4.11. (b)). In spite of
their extremely high oxygen transmission, silicone contact lenses have not been a
commercial success because of their hydrophobic character. This results in a poor
wetting of the lens by the tear fluid, making these lenses uncomfortable to wear.
 Organic polymers and ceramics 117

Currently, the great majority (85%) of soft contact lenses are made from hydrogels,
which are also cross-linked polymer materials but synthesized with hydrophilic
monomers (figure 5.20. (b)). The main constituent of these soft lenses is hydroxy-
ethylmethacrylate combined with other hydrophilic monomers such as N -
vinylpyrolidone.
These polymer materials, illustrated in figure 5.21., soak up large quantities of
water (between 30 and 80 % ) . Before hydration, the cross-linked polymer is rigid.
The water acts as a plasticiser reducing the glass transition temperature and changing
the rigid cross-linked polymer into a very comfortable hydrophilic elastomer. It
should be noted that the presence of water slightly reduces the refractive index.
Nevertheless, hydrogels have permeability to oxygen some 4 to 7 times less than that
of a rigid contact lens.

Figure 5.21 Schematic representation of the structure of a hydrogel.

Hydrogels are much less robust than rigid plastics and a soft lens has to be re
placed much more frequently than a rigid lens. Initially soft lenses were conceived
for a period of use of one to two weeks. Currently, the tendency is to replace them
daily.
The rigid lens is in general fabricated on a high precision lathe with diamond
tools, starting from a blank. The soft lens (rigid in the dry state) is also made in the
same way but it is necessary to take into account that hydration leads to a volume
increase of some 30 %. After lathing, the soft lens is hydrated and deep cleaned to
eliminate residual monomers toxic to the cornea. Soft lenses are also made by
pouring a mixture of monomers into a mould followed by polymerisation.
In conclusion, there still remain important developments to make in the area of
contact lenses. In particular, research is being directed to finding more permeable
materials. In addition, it is necessary to perfect the manufacturing processes so as to
have longer production runs and lower cost.
118 Introduction to Materials Science

5.6. Exercises
5.6.1. Calculate the approximate value of the enthalpy of polymerisation of a vinyl
polymer.
5.6.2. Calculate the length / of a linear macromolecule of polyethylene (C H4) when
2

= 500, taking into account that the valence angle between the carbon atoms is
equal to 109.5 .
5.6.3. A polyethylene has an average molecular weight of 150000 g mol" . What is 1

the degree of polymerisation?

5.6.4. What is the energy released by the polymerisation of 100 g of ethylene? The
heats of formation are given in table 2.17 and Avogadro's number is equal to
.^ . 1

5.6.5. Polytetrafluoroethylene (Teflon) is obtained by the polymerisation of tetra-

fluoroethylene CF =CF . What is its degree of polymerisation if the molecular weight
2 2

is equal to 33000 gmol" ? How many macromolecules are there in 1 g of polytetra

1

fluoroethylene?
5.6.6. The density of diamond = 3.52 gem" and that of high density polyethylene
3

HDPE is of the order of 0.95 g cm" . Explain this difference in terms of chemical
3

bonding.
5.6.7. What is the difference in structure between poly(dimethylsiloxane) and the
pyroxenes?
5.6.8. Magnesium oxide MgO is a ceramic with the same type of crystalline structure
as sodium chloride NaCl. Calculate the theoretical density of MgO given that the
structural unit is given in figure 4.7. (b).
5.6.9. It is proposed to make a glass starting from a mixture composed of 70 molar %
of S i 0 , 15 molar % of CaO and 15 % molar of N a 0 . Given that, for a mixture of
2 2

oxides to form a glass, the ratio of O/Si < 2.5, determine if the mixture of oxides
proposed will allow a glass to form.
5.6.10. A block of ceramic results from the firing of 250 g of alumina particles
( A 1 0 ) , with a density of 3.9 gem" . Suspended in water, this block of ceramic
2 3
3

weighs 160 g and its dry weight is 295 g. Calculate the amount of interconnected
porosity accessible to water. Calculate the total amount of porosity. Calculate the
volume fraction of closed porosity.

5.6. References and complementary reading

C.W. BUNN, E.R. HOWELLS, Nature, 174, 549, (1954).
C.W. BUNN, Chemical Crystallography, Oxford University Press, London, 1945.
B. CLAVAUD, F. SAUCIER et L. BARCELO, Scientific American, n 244, 32, (1998).
R.W. DAVIDGE, Mechanical Behaviour of Ceramics, Cambridge University Press, Cambridge, 1979.
G.J. FROHNSDORFF, Portland Cements, Blended Cements and Mortars in Encyclopedia of Materials
Science and Engineering, . B. Bever Ed., Pergamon, Oxford, vol.. 5 p. 3847, 1986.
R.P. FRANZ, R.E. BAUMAN et A.A. THOMPSON, Basics of Contact Lenses, Cibavision Tecnical
Series, 1993.
 Organic polymers and ceramics 119

M. GORDON, High Polymers, Structure and Physical Properties, Iliffe London, 1963.
W. GRAEVES, La Recherche, 13, 1184 (1982).
C. HUET, Cours de Materiauxde Construction, EPFL, Lausanne, 1993.
H.S. KAUFMAN, J J. FALCETTA (ed.), Introduction to Polymer Science and Technology, John Wiley,
New York, 1977.
. H. LEWIS ed., Glasses and Glass-Ceramics, Chapman & Hall, London, 1989.
I.J. McCOLM, Ceramic Sciences for Materials Technologists, Leonard Hill, New York, 1983.
J. LEMAITRE, Injectable Calcium Phosphate Hydraulic Cement: New Developments and Potential
Applications, Inn. Technol. Biol. Med. GRIBOI, 95.
J. P. MERCIER et . MARECHAL, Chimie des polymeres, (Traite des Materiaux , vol. 13), Presses
polytechniques et universitaires romandes, Lausanne, 1996.
D. W. RICHERSON, Modern Ceramic Engineering, Dekker, New York, 1992.
F. L. RILEY, Silicon Nitride in Encyclopedia of Materials Science and Engineering, M. B. Bever Ed.,
Pergamon, Oxford, vol. 6, p. 4412,1986.
H. TADOKORO, Structure and Properties of Crystalline Polymers, Polymer, 25, 147 (1984).
H. W. F. TAYLOR, Cement Chemistry, 2 ed. ,Thomas Telford, London, 1998.
nd

M.H. VAN DE VOORDE, C.A.M. SISKENS, W. BETTERIDGE, Ceramic for High Temperatures
Energy Application, Sprechsaal, 115, 1027-1108 (1982).
L.H. VAN VLACK, Physical Ceramics for Engineers, Addison-Wesley, Reading, Mass., 1964.
B. WUNDERLICH, Macromolecular Physics, Volume 1, Academic Press, New York, 1973.
J.F. YOUNG, Cement as Building Materials in Encyclopedia of Materials Science and Engineering, . B.
Bever Ed., Pergamon, Oxford, Vol. 1 p. 566, 1986.
 Chapter 6.

Elastic behaviour of solids

6.1. Objectives
To study the relationship between the forces acting on a solid and the
resulting deformations.
To introduce the concept of modulus of elasticity.
To identify the atomic parameters which determine the elastic behav
iour of solids.
To calculate the theoretical limit for the strength of materials.
To analyse the viscoelastic behaviour of certain materials and to intro
duce the concept of linear viscoelasticity.

Any object subjected to a mechanical force deforms. This deformation, which

leads to the displacement of the atoms from their equilibrium position, produces
forces of retraction opposing the deformation. These tend to re-establish the solid in
its original form when the force is no longer applied.
The behaviour of solids subjected to mechanical stresses is determined by the
nature of the interatomic forces. In this chapter, only the elastic behaviour of ideal
solids submitted to small deformation (generally < 0.1 %) will be studied in order to
simplify the presentation of the concepts introduced. In chapter 11, the tensile beha
viour of real materials will be studied at high deformation.

6.2. Elastic deformation of solids

6.2.1. Introduction
The theory of elasticity treats the relationship between forces applied to an ob
ject and the resulting deformations. In practice, the analysis of the elastic behaviour
of a material is reduced to the study of simple deformations and the determination of
the corresponding elastic constants. These simple deformations are uniaxial elonga
tion, simple shear and uniform (or hydrostatic) compression.

6.2.2. Uniaxial elongation

When a prismatic sample (figure 6.1.) is subjected along an external force F , it
undergoes an uniaxial elongation proportional to its initial length x . 0

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

122 Introduction to Materials Science

Figure 6.1. Uniaxial elongation of an elastic bar of initial cross-section S =yo z and of
0 0

initial length xo. This figure is schematic and does not show the lateral contraction, asso
ciated with the elongation of the sample during the application of the force

This elongation produces a force of retraction F in the solid equal in absolute

r

value and opposite in direction to the appliedforce F:

F + F =0 r

The force of retraction F and the elongation are expressed in general as a func
r

tion of the initial dimensions of the prismatic sample. The relative force per unit
cross section of the sample is the (nominal) stress:

(6.1.)
x
s 0

The ratio of the increase in length Ax, to the initial length x defines the (nomi
0

nal) strain : x

= (6.2.)

x0

The nominal stress (strain) is the stress (strain) determined taking into account
the initial dimensions of the sample. The real stress (strain) considers the dimension
of the deformed sample. The difference between nominal and real values is only
significant at high deformation and will be discussed in chapter 11.
At small deformation (for most materials, e < 0.1 % ) , a linear relationship exists
x

between the stress and the strain e \

x

=

(6.3.)

This expression defines Hooke's Law, in which is the elastic modulus in ten
sion or Young's modulus characterising the resistance (rigidity) of the solid to
uniaxial deformation. In table 6.2., the values of the elastic modulus for a number of
 Elastic behaviour of solids 123

Table 6.2. Typical values of the modulus of elasticity for different

materials at ambient temperature.

Material Material
[GPa] [GPa]

Diamond 1000 Window glass 70

Silicon carbide (SiC) 450 Aluminium 70
Tungsten 400 Concrete 50
Alumina (A1 0 )2 3 400 Magnesium 40-45
Carbon fibre 300 Wood agglomerate 7
Magnesia (MgO) 250 Epoxy resins (cross-linked) 2.8-4.2
Steel 210 Polystyrene 2
Copper 125 Polyamide 6-6 2
Brass, bronze 110 Polypropylene 1.5
Glassy silica ( S i 0 ) 2 95 Medium density polyethylene 0.7
Gold 80 Rubbers -0.001

materials are given. They are expressed in GPa or G N m ( 1 0 ^ m~ ). The variation2 2

observed from diamond to rubber extends over six orders of magnitude. The value
for the modulus of elasticity is a function of the bond energies, the nature of the re
tractive elastic forces, and the structure (amorphous or crystalline) of the material.

6.2.3. Lateral contraction and Poisson's Ratio

For most materials; the elongation Ax of the sample along the tensile axis leads
to an increase in volume. In the case of elastic deformation, this volume increase is
partially compensated by lateral contractions, Ay and , of the sample in the y and
transverse directions perpendicular to the tensile axis.
The strains in the y and directions are given by:
Ay Az
= et =
y (6.4.)
yo z 0

For an isotropic material, and e are equal. y z

The effect of the lateral contraction is generally measured relative to the tensile
strain. Thus Poisson'sRatio is defined as:

V = -3l =-- (6.5.)

where the negative sign makes positive. It is easy to relate the variation in volume
during uniaxial extension to Poisson's Ratio v. The relative variation in volume is
written as:

(6.6.)

with
Vo = xoyoz 0 (6.7.)

and
124 Introduction to Materials Science

(6.8.)
XQ
0
yo

By considering only small strains, it is possible to neglect the infinitely small second
and higher order terms. This means that, taking into account equations 6.6, 6.7, 6.8, it
is possible to write:

Ax Ay Az
A
(6.9.)
A= + + : x + y + z

XQ
yo

By introducing into equation (6.9.) the value of taken from (6.5.), the following
relationship is obtained:

A= (l-2v)
x
(6.10.)

The various values of Poisson's Ratio are summarised in figure 6.3. The upper
limit for the value of is 0,5 for an isotropic material. This corresponds to a zero
variation in volume during extension (A = 0). This limiting value is almost obtained
for rubber (v = 0.49) which deforms under tension in an elastic manner almost with
out any increase in volume. For metals, Poisson's Ratio is close to 0.35 and for ce
ramics, is generally between 0.07 and 0.27.

Metals Ceramics Polymers

0,5
Nat rubber AV = 0

Pb PE
0,4
Ag
Cu AV>0
Al PMMA
PS, PA 6-6
0,3
Fe, Steel, W

Mineral glass
0,2 A1 0 ,WC2 3

MgO
Amorphous S i 0 2

0,1
Diamond

0,0

Figure 6.3. Indicative values of Poisson's Ratio for different types of materials for
elastic strains at ambient temperature.
 Elastic behaviour of solids 125

6.2.4. Simple shear

Shear forces play an important role in the behaviour of materials. Shear forces
are involved when brake blocks are applied to the rim of a bicycle wheel or to the
jaws of a car disc brake.
An example of deformation under shear is shown in figure 6.4. A prismatic
specimen is fixed at one face of the surface .SO to a rigid support, while on the oppo
site face a force F is applied parallel to the plane xy. The ratio between F and So is the
shear stress r = F/So. The shear stress induces a shear strain which is measured by the
shear angle = Ay/z . There exists for this deformation, which occurs at constant
0

volume, a linear relationship between the stress per unit area of surface, the shear
stress r a n d the shear angle :

= Gtgy- Gy (6.11.)

since for small deformations, t g / ~ = Ay/z . G is the shear modulus.

0

Figure 6.4. Simple shear. The force applied per unit area is the shear stressr = F/SO and
is the shear angle (shear strain).

6.2.5. Uniform (hydrostatic) compression

Uniform compression constitutes the third type of simple deformation. This
occurs when a solid is subjected to a hydrostatic pressure (positive or negative). A
linear relationship exists, at small volume change, between the hydrostatic pressure
(stress)p, and the relative change in volume = AVIV, given by:

= - (6.12.)

In this expression, is the bulk modulus. The negative sign in this expression is due
to the fact that the change in volume is negative for a positive p.
126 Introduction to Materials Science

6.2.6. Relationships between the various elastic moduli

Three different moduli, E, G and K, which have been defined, allows the elastic
behaviour of a material to be characterised. These three constants result from the
proportionality, which exists between stress and strain for these three types of simple
deformation:
=
- Gy (6.13.)
--

As already indicated, these three expressions of Hooke's Law are only valid for
small deformations (linear elasticity). This limit for linear elastic strain is approxi
mately equal to 0.1 % for high modulus materials such as metals. Above this limit the
phenomena of permanent (plastic) deformation appears in a large number of cases
(chapter 12).
As will be seen latter, for elastomers it is not isolated atoms, which are displaced
but segments of chains. Elastomers are characterised by very low values for the
and G moduli (of the order of 0.01 to 0.001 GPa). They are materials very deform-
able by uniaxial elongation and in shear. In uniform (hydrostatic) compression, elas
tomers behave like high modulus materials or like liquids, with a bulk modulus
greater than 1 GPa.
The three moduli E, G and K, as well as Poisson's Ratio are related to each
other by the following equations:

E =- ^ ^ = 3K(l-2v) = 2G(l + v) (6.14.)

In other words, among the four elastic constants, only two are independent.
These relationships are only valid for isotropic solids. For the most general case of an
anisotropic solid, twenty-one independent elastic constants exist (Cottrell, 1964).
In isotropic solids, simple shear and uniform compression make up two types of
elementary deformation, because in one case the volume and in the other case the
shape of the sample remains constant. Uniaxial elongation is not a elementary de
formation because, except in the case of rubber, it causes both the volume and the
shape of the sample to vary simultaneously.

Table 6.5. Relationships between K, G and

Physical State G,E

Liquid large 0
Rubber large small (E, G K)
Crystals, glasses large large (, G~X)

In table 6.5, the relative values of the elastic constants are summarised. For
rubber, is very high compared to and G, i.e. for mechanical deformations, the
elastomers can be considered to be essentially incompressible bodies and Poisson's
Ratio is practically equal to 0.5. Equation 6.14 reduces then to:
E~3G (K^oo) (6.15.)
 Elastic behaviour of solids 127

6.3. Thermodynamics - Atomic origin of elasticity

6.3.1. Thermodynamics of isothermal uniaxial elongation

In section 6.2., consideration was limited to the macroscopic aspects of the elas
tic behaviour of solids. It is useful to determine the relationships existing between the
force of retraction, the length and the temperature of a sample subjected to deforma
tion and the magnitude of the fundamental thermodynamic parameters, i.e. free en
ergy (free enthalpy), internal energy (enthalpy) and entropy.
Consider an elastic bar (figure 6.6.) of initial length / , maintained at a length
0

/ > /o by the force F. The application of the first and second laws of classical thermo
dynamics to this deformation will now be examined. The variation of the internal
energy dUresulting from this elementary deformation is given by:
dU=dq-dW (6.16.)

where dq is equal to the amount of heat absorbed by the system during deformation
and dW represents the work done. By considering only the case of reversible proc
esses, the thermodynamic effect dq is related to the entropy variation dS by:

dq=TdS (6.17.)

The work dW done by the system breaks down into two terms related to the work
done by the elastic force of retraction and to the work of the pressure p\

F + dF

dl

F + dF

Figure 6.6. Uniaxial extension of an elastic bar, maintained at a length / by a force F and
then extended by an increment d/ by an increase in force of dF.
128 Introduction to Materials Science

dW=-F dl+pdV r (6.18.)

It should be recalled that the elastic force of retraction F is equal in absolute valuer

and of opposite direction to the applied force F:

F = -F
r (6.19.)
The variation in internal energy during the reversible uniaxial deformation is there
fore given by:
dU = TdS+F dl-pdV r (6.20.)

If the free energy A (Helmholtz's function) is defined in the usual way:

A-U-TS (6.21.)
then the variation in the free energy dA can be related to the entropy and to the work
carried out by the system. Therefore:

dA = - SdT -pdV + Frdl (6.22.)

If the deformation occurs at constant and V, the expression (6.22.) reduces to

dA = F dl or ( 3 A / 3 / ) = F . For constant and V and taking into consideration
r rv r

equation (6.21.), then:

( 3/ ) r f V ( 3/ ) r v (3/ \ v (6.23.)

The elastic force of retraction, at constant and V, is therefore equal to the in

crease in free energy per unit length of elongation of the specimen. The elastic force
of retraction can be split into two components, the internal or enthalpic force of re
traction F ,i , and the so-called entropic force of retraction F . The mechanical en
r r<e

ergy brought to the system by the deformation can be stored in the form of an inter
nal energy increase resulting from the modification of the interatomic distances or
the bond angles. The energy can also be dissipated into the environment as heat with
a corresponding reduction in the system entropy accompanied by an increase in order
resulting from an orientation of the chains.
As A is a state function, dA is an exact differential, i.e.:

" IdA\' 3 f

(6.24.)
a/1 ,3/Jr.
and so:
JdS\ JdFA (6.25.)

By introducing (6.25.) into (6.23.), the following relation can be deduced for F : r

(6.26.)
 Elastic behaviour of solids 129

The relationship (6.26.) is a form of the equation of state for isothermal reversi
ble elastic deformation. It shows that by studying the variations in the force of re
traction as a function of temperature, the internal and entropic contributions to the
force of retraction can be determined.
For all materials except elastomers, the volume increases during uniaxial elon
gation but since this variation is small, expression (6.26.) remains valid, to a first
approximation, for all materials.
This study of the thermodynamics of elastic deformation leads to two limiting
cases: the ideal crystal and the elastomer that will be introduced in the course of this
chapter. For an ideal crystal, the force of retraction remains practically constant as a
function of the temperature. This means that it varies with internal energy or enthalpy
changes. For rubber, there is an increase in the force of retraction proportional to the
absolute temperature and this has therefore an almost exclusively entropic origin.
There is considerable similarity in behaviour between a gas deformed under
compression and rubber deformed in tension. In a perfect gas subjected to isothermal
reversible compression, the entropy per unit volume varies proportionally with pres
sure:

(6.27.)
P 1
\dV;
T [dT;
V

In the same way, ideal rubber is defined as a material characterised by an elastic

force of retraction exclusively entropic and given by the following expression:

(6.28.)
{dl) y
T {dT) ltV

This analogous behaviour of the perfect gas and ideal rubber is not fortuitous.
Gas molecules heat up when they are rapidly compressed and cool down during
expansion. A piece of ideal rubber heats up when stretched and cools down when the
deformation is released. This thermal effect is easy to explain: during the isothermal
deformation of an ideal elastomer, there is virtually no interaction between the chains
and the internal energy does not depend on the degree of elongation. In this case, the
elongation of the sample is only accompanied by a reduction in entropy, leading to
heat emission into the thermostatic bath surrounding the sample. During the opposite
deformation, entropy increases with heat absorption. In practice, it is observed that
the current elastomers have an elastic behaviour close to that of ideal rubber.
The heating up of an elastomer during elongation should not be confused with
the heating of a car tyre subjected to dynamic deformation induced by the movement
of the vehicle. In effect, during the cycle of deformation experienced by the tyre, the
phenomena of elongation and retraction entirely compensate each other and the
global calorific effect should be zero for an ideal rubber. In reality, a hysteresic effect
(chapter 11) occurs and the loading curve is not identical to that for unloading. A
fraction of the energy put into the system during elongation of the rubber is
dissipated in the form of heat during retraction. The internal heating of tyres is
related to the phenomena of internal friction created by the rubbing of segments of
chains against one another, exactly as molecular friction in a deformed fluid leads to
the Joule effect. This effect is not considered in the theory of elasticity presented
130 Introduction to Materials Science

here. This non-elastic response of the rubber also explains the shock-absorbing
capacity of rubber blocks which isolate machine tools from their mountings.

6.3.2. Relation between phenomena occurring on an atomic scale and thermo

dynamic variable
The force of retraction is always induced by the displacement of atoms from
their equilibrium positions. In solids with enthalpic deformation behaviour (metals,
crystalline ceramics, mineral or organic glasses, highly cross-linked thermoset poly
mers etc.), the force of retraction (figure 6.7. (a)) results from small displacements of
the atoms from their equilibrium positions. The cohesive energy of these solids being
high, the elastic forces of retraction are very strong, the modulus of elasticity is high
and the elastic domain limited ( < 0.1 % ) . In this case, the deformation is associated
with a significant increase in internal energy (or enthalpy) of the system, whereas the

(a)

CO

Cu

0,2 0,4 0,6 (%)

200 400 600 (%)
Deformation > Deformation
(b) (d)
Figure 6.7. Atomic mechanisms involved in the elasticity of solids and their relationships
with internal energy U and entropy S variations. Deformation mechanisms (a) and inter
nal energy variation (b) in enthalpic solids (metals and ceramics). Deformation mecha
nisms (c) and entropy variation (d) in entropic solids (elastomers). The different scales of
deformation between (b) and (d) should be noted.
 Elastic behaviour of solids 131

configuration remains practically unchanged since the atoms hardly move from their
equilibrium positions (figure 6.7. (b)).
The solids with entropic elasticity (elastomers) (figure 6.7. (c)) are made up of
an assembly of polymeric chains with a small number of cross-links (~1 cross-linked
unit per 100 structural units in general). This three-dimensional structure ensures the
reversibility of the deformation. The cohesion between the rubbery chains is weak
and of the same order of magnitude as that found for volatile molecular liquids (~4 kJ
m o f ) ; this value is some 100 times smaller than covalent cross-linked bonds.
1

Consequently in elastomers, it is not individual atoms, which are displaced but seg
ments of chains. In this type of material, the forces of retraction are induced by mi-
crobrownian movements which continuously modify, in a random way, the position
of diverse segments of chains and which therefore oppose any orientation of the
chains under the action of an external force. The forces of retraction created by these
microbrownian motions are very weak and they increase in proportion with the ab
solute temperature T. This explains why elastomers are easily deformed and have
very low elastic moduli (E ~ 1 to 10 MPa), the latter increasing as a function of T.
As previously mentioned, this study has been restricted to the perfect crystal
(slightly deformable) and an ideal rubber (very deformable). It should also be noted
that the majority of current materials are solids with enthalpic elasticity. Their elastic
moduli vary by some three orders of magnitude, from diamond ( ~ 10 GPa) to the 3

organic glasses ( ~ 2 GPa). This variation mainly depends on the nature of the
bonds. This explains the difference in the elastic modulus observed between mineral
glass ( ~ 70 GPa) and organic glass (E = 2 GPa). The cohesion between the polymer
chains of organic glasses is made by secondary bonding (-40 kJ mol" ), while only1

strong ionic or polar covalent bonds exist in mineral glasses. The mechanical proper
ties are also influenced by the atomic structure and so amorphous materials generally
have somewhat lower elastic moduli than crystalline solids.
Semi-crystalline polymers (chap. 5 and 10) have a microcrystalline structure
dispersed in an amorphous phase. If the amorphous phase is glassy (T< Tg), the
modulus of these materials is of the same order of magnitude as that of organic
glasses ( ~ 2 to 3 GPa). When the amorphous phase is rubbery (T> Tg), the elastic
modulus varies between 0.2 and 1.5 GPa as a function of the degree of crystallinity.
In this class of materials, the elastic force of retraction can simultaneously have en
thalpic and entropic components.

6.3.3. Elastic deformation of a perfect crystal

This calculation will be restricted to the case of a crystalline solid having a
primitive cubic unit cell (figure 3.8.); a spherical atom of diameter r (section 2.4)
0

fills each node. Each atom occupies a cubic volume roughly equal tor . The study is
0
3

limited to the deformation of a prismatic crystal in tension (figure 6.8.). To simplify

the calculation, a tensile axis parallel to an edge of the basic network is chosen. Thus,
from expression (6.29.), the strain of the solid is the same on both the macroscopic
x

and atomic scales.

Ax _ Ar
(6.29.)
x

x ~ ~~
0
x r
0
132 Introduction to Materials Science

Figure 6.8. Uniaxial deformation of a prismatic crystal of section S and length x in 0 0

which the basic unit cell is primitive cubic and the structural unit is an atom of diameter
r . S = r is the area occupied by an atom. The tensile axis is parallel to the edge of the
0 a 0
2

basic unit cell.

For isotropic materials, the transverse deformation is similar on both the micro
scopic scale and on the macroscopic level:
_ Ar _ Ar y z
(6.30.)
0
r
r 0

This is known as an affine deformation. In crystalline solids, this behaviour is

only observed for small deformations ( < 0.1 %).

In elastically deformed crystalline solids, the force of retraction is induced by

the atoms being displaced from their equilibrium positions. When only small defor
mations are considered, the elastic force of retraction f , between two neighbouring x

atoms parallel to the direction of deformation is proportional to the interatomic

distance Ar : x

f
x = CAr x (6.31.)

fx represents the microscopic retraction force acting between two atoms ,each occu
pying a section S = r (figure 6.8.). It should be noted that f, represents, in reality, a
a 0
2
x

difference in force f relative to the equilibrium position (f = 0). The macros

x iX

copic force of retraction F , which is related to the microscopic force of retraction f ,

r x

is proportional to the number of atoms found in the section So.f.x represents a dif
ference in force f relative to the equilibrium position if, 0). C is a constant of a
x
=

function proportional to the bond energy. Since each atom roughly occupies a section
S = r (figure 6.8.), the macroscopic retractive force F is proportional to the number
a 0
2
r

of atoms found in the section So:

(6.32.)
0
r
0
r

To simplify the notation, in ( 6 . 3 2 . ) h a s been used for/ ,x and r for r . As: r x

_ Ax _ Ar
(6.29.)
x

x
0 r
0
 Elastic behaviour of solids 133

then:
^ SftC Ax
F =-^ r (6.33.)

Introducing the stress , then:

= ^ =- (6.34.)

In this expression, C/r is the modulus of elasticity or Young's modulus E, which was
0

defined in (6.3.) in introducing Hooke's Law ( 6.2.2.):

=
(6.3.)

6.3.4. Theoretical strength and fracture energy

In solids with enthalpic elasticity, the force of retraction is related to the bond
energy. The following calculation relates tensile strength to the modulus of elasticity.
For ionic solids, an equation was established, in chapter 2, which relates the variation
of the bonding force between two ions of opposite sign as a function of the distance
(2.20). When a pair of ions is subjected to elastic deformation, the force of retraction
has (with the exception of the sign) the same mathematical form as the expression
(2.20). It can thus be written:

f r 8

f =JL i_LL (6.35.)

Jr 2

r represents the interatomic distance after deformation and r is the distance at which0

the cohesion and repulsion forces are in equilibrium (f = 0). If only small strains are
considered, it is possible to write the following approximate expression:

f ~df =^-dr
r r (6.36.)
ro
For ionic solids, this expression permits the rigidity constant, C of the bond and
the elastic modulus to be calculated. Introducing f into (6.31) and (6.3.) respectively,
r

the value of C is obtained:

C=H (6.37.)

and the modulus of elasticity E:

=^ = * (6.38.)

The force of retractions increase significantly when the deformation is in

creased (figure 6.10.). The force of retraction passes through a maximum for a value
of strain equal to e , . The maximum value of the force of retraction/^, which deter
x m

mines the theoretical tensile strength of the material, is calculated by putting the first
derivative of (6.35) equal to zero, giving:
134 Introduction to Materials Science

e , = 0.22
x m (6.39.)

The value of e (22 %) is high and the bond strength f , deviates considerably from
x>m r m

the linear relationship (6.36.) From the equation (6.35.), which gives the variation in
the force of retraction as a function of the interatomic distance r, the maximum
strength of the material R is calculated by:
max

X_
Rn :0,54 (6.40.)
f
ro4

2 Jr,m
By substituting equation (6.38) in (6.40), R can be calculated as a function of the
m x

modulus of elasticity E:
E_ (6.41.)
RfT\ il Y
15
A similar relationship can be calculated for other materials with strong bonds.
This relationship is approximate, as it only takes into account the interactions be
tween pairs of atoms.
Is it possible to obtain maximum strengths as high as those predicted theoreti
cally? In practice, the maximum strength R corresponds to the stress in the material
m

at the onset of plastic deformation, i.e. the yield strength R (chapter 11.). Figure 6.9. e

gives the value for the ratio RJE for the three categories of materials. In certain cases
(diamond), very high maximum strengths are achieved close to those predicted theo
retically. This is equally true for metal whiskers, which are fine needles (some in
diameter), obtained under special conditions of crystalline growth which confers on
them a defect-free crystalline structure. For example, an iron whisker attains an ex
traordinary strength of 11 GPa with a maximum elastic elongation of 5 %. This is
very close to the theoretical value (R = /15 ~ 14 GPa). On the other hand, the best
max

Metals Ceramics Polymers

Theoretical limit
10"
Trichites Diamond ,,
SIC PPMA
Ti Alloys Al 0 ,Si N
2 3 3 4
Steels
MgO
Al Alloys
Cu Alloys
- 3

Concrete, cement
Pure metals
1 1 ( r

3
Ultrapure
10" : metals

10" 6

Figure 6.9. Ratio RJE for different types of materials. The theoretical limit of R / max is
equal to 1/15 (after Ashby and Jones, 1980).
 Elastic behaviour of solids 135

steels currently used have a maximum strength of only about a quarter of the theo
retical value. A steel wire, which in theory should be able to support stresses of the
order of 14000 MPa (1400 kg mm" ), in practice does not generally withstand more
2

than 2400 MPa (240 kg m m ) . -2

Polymers also have strengths which approach the theoretical limit R = E/\5. It max

is important to note, however, that, in this case, the strength of these materials is low
because the modulus of elasticity is significantly less than that of other materials, due
to the presence of weak secondary bonds.
As may be seen in figure 6.9, calculation leads, in the case of metals and alloys,
to theoretical strengths which are often hundreds of times higher than measured
values (10000 times for pure metals). This arises because the role played by defects
in lowering material strength has been neglected in the calculations. These defects
will be studied in the next chapter. It is by taking into consideration the predominant
role played by these defects, in particular by dislocations, that theory and experience
can be reconciled. It should be noted that it was this divergence between theory and
practice that was at the origin of the discovery of dislocations.
Fracture energy is another important characteristic of materials. The energy at
the fracture of a bond U, is given by the area under the curve of the variation in the
force of retraction f as a function of deformation (figure 6.10.):

Ur="fr,xar (6.42.)

By multiplying the fracture energy U by the number of bonds broken per unit
R

area of the solid rupture surface, the specific fracture energy G can be calculated. C

The separation of the sample into two parts during rupture forms two fracture
surfaces, each one with a specific surface energy y. If the creation of two surfaces
were the only phenomenon accompanying the rupture, then the specific fracture
energy G would be double the specific surface energy /.
C

G = 2y (6.43.)
C

0,2I 0,4 0,6 0,8

Deformation = Ar/r
0

Figure 6.10. Variation of the ratio of the retraction force f to the maximum value f , of
r r m

the ionic bond as a function of the deformation .

136 Introduction to Materials Science

Measurement of the specific fracture energy G in brittle materials such as ce

c

ramics gives values close to theory (figure 6.11.). On the other hand, ductile materi
als such as metals and certain polymers have fracture energies from 10 to 10 times 3 6

higher than theory. This indicates the existence of other energy absorption processes,
linked mainly to plastic deformation (chapters 7. and 11.). This highlights a very
important characteristic of ductile materials, which will be examined latter (chapter
14.). There are few materials, which reach the theoretical limit of 2y. In practice,
there are almost always other mechanisms (chapter 14.) dissipating energy, thereby
increasing the fracture energy above the theoretical value predicted by (6.43.).

Metals Ceramics Polymers

10 3

Ductile Metals

10 2 Steels
Ti alloys

Al alloys
10 PP
PE
PA
PS
PMMA
EP
Be
I - 1

S3N4
UP

sic
MgO
- Cement
AUO3
Glass

io-
Figure 6.11. Measured values for the specific fracture energy G for different classes of
c

materials (after Ashby and Jones, 1980).

6.3.5. Elastic deformation of ideal rubber

As already seen in figure 6.7. (c), an elastomer consists of an assembly of chains
linked to each other by cross-links which are introduced by vulcanisation after shap
ing of the material. Microbrownian motions constantly excite the polymer segments
in an elastomer, and their amplitude is proportional to the absolute temperature. The
kinetic theory of elasticity in rubber shows that the microbrownian motion, which
excites an elastic chain linked at both ends by cross-links (figure 6.12.), induces a
force pulling the chain ends together. This force of retraction, which is entropic in
origin, is proportional to the distance / separating the two cross-links situated at the
ends of the elastic chains and is given by the expression:

- (6.44.)
3 K T L

r
~ 2
 Elastic behaviour of solids 137

Crosslink
Crosslink

Figure 6.12. Elastic retraction force between the ends of a elastomer segment.

The factor k is the Boltzmann s constant and represents the mean quadratic dis
1

tance between the elastic chain ends before the introduction of the cross-links (i.e.
before vulcanisation).
In the undeformed state, the retractive forces acting on the cross-links of the
network are in equilibrium. When external tensile forces are applied to the rubber
sample, the distance separating the ends of each elastic segment is increased. For the
chain represented in figure 6.12., when the distance / between the two ends of the
segment is increased by an increment Al parallel to /, the elastic retraction force is
increased by a value f given by the expression:
r

r 3kT (6.45.)

During deformation, the mean distance between the cross-links varies in the
same proportion as the macroscopic deformation. Therefore, an affine deformation
occurs here, as well as in the case of the deformation of a monocrystal described in
paragraph 6.3.3. In an elastomer, it is not the chain atoms, which are displaced in an
affine manner, but it is rather the cross-links.
For high modulus materials (metals, ceramics, glassy or semi-crystalline poly
mers), there exists, for small deformations (-0.1 % ) , a linear relationship between the
stress, i.e. the retraction force per unit area, and the strain. The elastic region for
these materials is very small. More important deformation leads to plastic deforma
tion or rupture ( 11.2.3.). For elastomers, which are much more deformable materi
als, the elastic region is larger. In figure 6.13., the stress-strain curve is shown for
rubber in uniaxial compression and tension. The deformation is completely reversible
up to levels of several hundreds of per cent.
By generalising (6.45.) to a rubber sample, a relationship between the nominal
stress and the strain can be calculated from the kinetic theory of rubber elasticity.

(6.46)

As shown in figure 6.13., the experimental points fit the theoretical curve for
relative strains = x/x from 0.4 (uniaxial compression) to 1.5 (uniaxial extension).
0

The Young's modulus is equal to the tangent to the stress - strain curve at
= 1. From the equation of the tangent, it can be demonstrated that the modulus of
elasticity is proportional to the number of moles of elastic chains per unit volume
and to the absolute temperature T:
E = 3nRT (6.47.)
138 Introduction to Materials Science

0,8

^^cf^y Extension
0,0

-0,8

/ Uniaxial
/ Compression

-2,4

I
0,5 1,0 1,5 2,0
Relative deformation - /
0

Figure 6.13. Elastic response of a natural rubber vulcanised with 8 % sulphur, in uniaxial
compression and extension. The stress is expressed in relation to the area of the
undeformed sample (nominal stress). The continuous line shows the theoretical curve
calculated from the gaussian theory of rubbery elasticity; the points correspond to
experimental results. (After Treloar, 1975).

In (6.47), R is the ideal gas constant, is related to the density p . The molecular
weight Me of the elastic segments is given by:

n=-- (6.48.)
e

Finally, the following expression is obtained for the modulus of elasticity of rubber:

E = 3 p R T
(6.49.)
Me

To a first approximation, the modulus of elasticity of an elastomer does not

depend directly on the chemical structure. The modulus of elasticity increases
proportionally with the absolute temperature, as does the force of retraction. As al
ready shown, this is a result of the almost exclusively entropic origin of rubbery
elasticity. The mechanical deformation leads to an alignment of the elastic chains
accompanied by a reduction in the entropy of the system (increased order). During
the deformation of rubber, the bond angles and the distance between the neighbour
ing atoms do not vary. The internal energy (or enthalpy) therefore remains practically
constant irrespective of the deformation (figure 6.7. (d)).
 Elastic behaviour of solids 139

6.4. Relationship between stress and strain in viscoelastic

materials
6.4.1. Viscous behaviour of Newtonian fluids
It has been seen in paragraph 6.2.2. that when an elastic body is deformed in
uniaxial tension, the stress varies proportionally with the strain (deformation), the
slope of this curve defines the Young's modulus. For Newtonian fluids subjected to
simple shear, a linear relationship exists between the stress and the deformation
(shear) rate defining the viscosity :

= ^ (6.50.)
1
dt
It should be noted that in hydrostatic compression, the Newtonian fluids exhibits
elastic behaviour (E = G =~ 0 ; 0).

6.4.2. Characterisation of the mechanical behaviour of materials

The mechanical behaviour of a substance can be characterised in a relatively
simple manner by applying a step deformation:
*(0 = *,(0 (6.51.)

where e $is a constant and u{t) a step function where:

x

Owheni < 0
u(t) = (6.52)
1 when t > 0

The response of an elastic material to a step deformation (figure 6.14. (a)) is also a
step function (figure 6.14. (b)):
( = ) (6.53.)
while in the case of a viscous fluid (figure 6.14. (c)), the response is a Dirac impulse:

(0 = , = , \ 0 0 (6.54.)
dt
where \t) i s th e Dira c function .
Elastic material s an d viscou s fluid s constitut e tw o limitin g cases . Substance s
exist whic h sho w intermediat e behaviour . Thes e ar e th e viscoelasti c materials . Whe n
a viscoelastic body i s subjecte d t o a ste p deformatio n *(/ ) = e w(0>th e stres s whic h
Xj0

must b e applie d t o kee p th e deformatio n constant , diminishe s progressivel y ove r tim e

(figure 6.14 . (d)) .
This behaviou r appear s principall y i n th e cas e o f thermoplasti c o r slightl y cross -
linked organi c polymers . I t i s du e t o th e anisotrop y o f th e fiel d o f inter - an d intra -
molecular forces , whic h i s induce d b y th e presenc e o f strong , covalen t chain - bond -
ing force s alongsid e relativel y wea k interchai n bonding . Whe n a mechanica l stres s i s
applied, th e chain s d o no t mov e instantaneousl y t o th e ne w equilibriu m position s an d
the mechanica l propertie s evolv e ove r time .
140 Introduction to Materials Science

Figure 6.14. Response of different mediums to an instantaneous step deformation (a): (b)
elastic material; (c) viscous fluid and (d) viscoelastic material where is the viscosity of
the fluid and b\t) a Dirac impulse.

In viscoelastic materials, the strains and stresses measured at a particular time /

depend on the whole history of strains and stresses previously experienced by the
material. This is known faded memory, as it retains a memory for the history of the
stresses or strains previously experienced in the course of manufacture or in use, but
the effect of these stresses or strains dies away progressively.
Viscoelastic behaviour is also exhibited in a large number of metals and ceram
ics when they experience temperatures close to the melting point or, for amorphous
materials, close to the glass transition temperature. At these high temperatures (T >
0.5 T , in degrees Kelvin), thermal agitation becomes significant. Polymers and cer
m

tain metals (Sn, Pb) melting at temperatures below 600 are already at an elevated
temperature under ambient conditions (300 K), and this gives rise to the appearance
of viscoelastic behaviour (chapter 12.).
Mechanical models are sometimes used to represent viscoelastic behaviour. The
behaviour of the ideal elastic solid and the viscous fluid are represented respectively
by a spring and a dashpot (figure 6.15. (a) and (b)). By combining springs and dash-
pots in series or in parallel, it is possible to simulate all types of viscoelastic beha-
 Elastic behaviour of solids 141

viour. In figures 6.15. (c) and (d) two basic mechanical models are shown: that of
Maxwell (figure 6.1. (c)) and that of Voigt-Kelvin (figure 6.15. (d)). Below the me
chanical models the different mechanical equations are given describing their be
haviour. As an example, the Maxwell model, consisting of a spring and a dashpot
placed in series, will be considered; a stress is applied during a very short time dt.
During the application period, the first term on the right of the Maxwell equation is
much more significant and the Maxwell equation reduces to:

te lfo
= (6.55.)
at dt
Equation (6.55.) is equivalent to that of an elastic solid. After much longer times than
that necessary for the application of the stress, the first term on the right hand side of
the Maxwell equation is zero and the equation reduces to that for a Newtonian fluid:

^ = (6.56.)
dt

This simple example shows that, as a function of the measurement time, a vis
coelastic body behaves either as an elastic solid or as a viscous fluid. The Maxwell
model and that of Voigt-Kelvin (figure 6.15. (d)) are in general too simple. To repre
sent quantitatively viscoelastic behaviour, a combination of the Maxwell and Voigt-
Kelvin models should be used. It must be emphasised that a large number of me
chanical models exist which are capable of representing the behaviour of a
viscoelastic material in a quantitative manner without having any direct molecular
significance.

1

e [iL

>
I

= 1
= ^
" dt
+
n +

(a) (b) (c) (d)

Figure 6.15. Elementary mechanical models used to simulate viscoelastic behaviour: (a)
elastic solid: spring; (b) viscous fluid: damper (dashpot); (c) Maxwell model: spring and
dashpot in series; (d) Voigt-Kelvin model: spring and dashpot in parralel.

6.4.3. Extension of Hooke's Law to linear viscoelasticity

In elasticity, the relation between stress and strain is independent of the manner
in which the stress or the strain is applied. In viscoelasticity, this is not the case and it
142 Introduction to Materials Science

is indispensable to define the experimental protocol used to apply the stress or the
strain. Thus, the concept of deformation mode is defined. In this treatment, the study
will be limited to two deformation modes: relaxation and creep.
Stress relaxation consists of applying a step deformation to a material and ob
serving the evolution of the stress as a function of time (figure 6.16. (a)). In a creep
experiment, stress is applied as a step function and the evolution of the strain is
studied (figure 6.16. (b)). Relaxation and creep are two important deformation
modes. It is important to avoid confusion between deformation modes and deforma
tion types, which are uniform compression, simple shear and uniaxial extension. In
the general theory of elasticity, the deformation types are related the form of the
stress and strain tensors while the deformation modes are related to their time de
pendence.

ess
1
in

t - t -
t
Retarded stress

Figure 6.16. Two deformation modes encountered in viscoelasticity: (a) stress relaxation;
(b) creep.

It is possible to apply the various deformation modes such as stress relaxation or

creep to each deformation type (uniform compression, simple shear, uniaxial exten
sion) previously introduced (section 6.2.). Stress relaxation in uniaxial extension can
be taken as an example. In this case, ( = *, and, from this experiment the relaxa
0

tion modulus E (t) can be defined:

r

E (t) = ^
r (6.57.)

x,0

In a creep experiment, it is a stress step function which is applied: (*(0 = ,).

The strain then varies as a function of time and the creep compliance Jt) is defined
by the relationship:

J ()
c t = M (6.58.)

For elastic materials, the creep compliance is the inverse of the relaxation modulus
since o and are constants:
x x
 Elastic behaviou r o f solid s 143

Er=_ = 2*&. = (6 .59.)

,0 Jc

In the case of viscoelastic materials, E (t) and 1/ J (i) are distinct functions of time.
r c

The expressions (6.57.) and (6.58.) are a generalisation of Hooke's law, already
defined in (6.3.) for a purely elastic body. This type of law is only applicable up to a
certain value of stress or strain known as the limit of linear viscoelasticity. This limit
is determined by a series of experiments where the initial stress (in creep) or strain
(in relaxation) is progressively increased. To illustrate this concept, a procedure used
to determine the linear viscoelasticity limit for stress relaxation is described in figure
6.17. For example, four relaxation experiments are carried out by applying increasing

(c)

Linear
viscoelasticity
limit

Strain

Figure 6.17. Experimental determination of the limit of linear viscoelasticity: (a) applica
tion of step deformations of increasing amplitude; (b) relaxation curves for the corre
sponding applied stresses. Stress - strain curve determined from (b) for different relaxa
tion times. The relaxation modulus for the different times considered is shown in (c).
144 Introduction to Materials Science

step deformations (figure 6.17. (a) and (b)). A certain number of values for time
(t\... t ) are selected and, for each value of t, a stress - strain curve (figure 17. (c)) is
5

plotted. The limit of linear viscoelasticity can be determined from the linear portion
of these curves.
The limit of viscoelasticity varies as a function of the modulus value. For po
lymeric materials having a modulus larger than 1 GPa, the value is less than 1 %. For
polymers with a modulus of the order of 1 MPa, it exceeds 10 %. The treatment that
has been developed can be extended to simple shear and to uniform compression.
The viscoelastic behaviour in uniform compression, however, differs notably from
that encountered in uniaxial tension and simple shear. In uniform compression, the
viscoelastic behaviour is much reduced since the variation between the bulk modulus
of a glassy polymer and of a rubbery polymer is of the order of 50 %. The Young's
modulus and the shear modulus diminish by a factor of between 100 and 1000 under
these conditions between a glassy and a rubbery polymer.

6.4.4. Bolzmann's Principle

In o r d e r to calculate the relation between stress and strain for viscoelastic mate
rials subjected to more complex stresses and strains, it is necessary to introduce the
concept of the linear superposition of effects, known in viscoelasticity under the
name of Bolzmann's superposition principle.
This was introduced in 1876 for creep. Boltzmann observed that the amount of
creep was a function of all the stresses previously applied to the sample and the effect
of the application of each stress was additive.
The method of introducing this principle is relatively simple (figure 6.18.). If a
viscoelastic material is subjected at time zero to a stress , after time / the strain
;

induced by this stress is equal to:

(= ,
Jc(t) (6.60.)

Figure 6.18. Bolzmann's principle: the strains e {t) = (t- t, )o j are additive in the same
Xti x

way as the stresses *,,.

 Elastic behaviour of solids 145

If a stress is applied at time h to another sample, the strain equation is given

by:
{)=
Jdt-h) (6.61.)

The successive application of two stresses , and to the same sample at 0

times t = 0 and t = U will now be considered. Bolzmann's superposition principle

postulates that the two stresses are acting independently and that the overall strain is
the sum of the two separate strains:
(t) = G ,oJ (t)
x x c + o ,iJc(t
x - ti) (6.62.)

This superposition of strains is represented in figure 6.18.

The equation (6.62.) can be generalised to the case of any particular stress by
breaking it down into the sum of elementary stresses. So that, if between the times
and + dr, the stress varies by an increment da , this elementary stress da will in x x

duce after a time t > , a strain de of value: x

d = J (t-T)da
x c x (6.63.)

The total deformation at time t is obtained by the integration of (6.63.):

= (-) = J J c ( t - r ) ^ d t (6.64.)
0 -co dr

A time value of minus infinity (->) is taken as the lower integration limit, to show
that all previous stresses applied to the sample should be taken into consideration.
In an entirely analogous way, from a knowledge of the strain history *(), the
stress at time / can be calculated:
* d
(= j(i- )-idT T (6.65.)
-co at
It should be noted that the functions and are continuous and derivable functions x

of time.
By applying Bolzmann's principle, the relationship between the characteristic
functions E (t) and J (t) can be calculated and the following approximate expression
r c

obtained:
E (t)= r
S h l ( m 7 C ) 1
(6-66.)
mn J (t)
c

In this expression, m represents the slope of the function, log J (t) - log t. When c

the slope m tends to zero, the function m(mn)lmii = 1, then E (t) = l/J (t). This r c

situation corresponds to the limiting case of ideal elastic behaviour, which is a par
ticular case of viscoelastic behaviour.

6.5 Summary and conclusions

In this chapter, the study of the behaviour of materials subjected to mechanical
stresses has been studied in an elementary manner. All forces (stresses) applied to a
146 Introduction to Materials Science

material result in a deformation (strain) inducing a change in the relative position of

the atoms. This deformation leads to the appearance of a force of retraction, which
tends to restore the sample to its original form. If, for a given deformation, the force
of retraction is constant over time, the material has an elastic behaviour. By consid
ering only small deformations, there is proportionality between stress and strain (lin
ear elasticity: Hooke's law).
It is possible to relate stress and macroscopic deformations to interatomic
bonding forces by considering two model materials: the perfect crystal and the ideal
rubber. A thermodynamic study has shown that the elastic retraction force is associ
ated with an increase in internal energy (or enthalpy) of the solid or with a decrease
in its structural disorder (entropy). This corresponds to two limiting cases described
in this chapter, i.e. enthalpic solid (metals, ceramics, thermoplastics) and entropic
solids (rubbers or elastomers).
A certain number of materials exist (organic polymers at ambient temperatures,
metals exposed to high temperatures) which no longer obey this ideal elastic behav
iour. In fact, these materials have a behaviour characterised as viscoelastic, i.e. in
termediate between that of an elastic solid and that of a viscous fluid. By introducing
the concept of linear viscoelasticity, it is possible to treat this phenomenon in a
quantitative manner and to generalise Hook's law.
Finally, with the help of simple calculations, the theoretical maximum strength
of materials has been estimated and their fracture energy determined. Theory and
practice do not agree in a large number of cases. This lack of agreement, which is
particularly significant for metallic materials, is related, as will be shown in chapter
7, to the existence of crystal defects and, in particular, dislocations. The movement
of dislocations leads to the dissipation of energy, which increases the fracture energy.

6.6. Illustrative example: the ultra-light aircraft or the

realisation of Icarus' dream
Since Leonardo da Vinci, the various efforts made by man to fly by his own
power have in general ended in failure. The recent conquest of air by manpowered
aircraft owes its success to a large extent to the availability of very strong and very
light materials.
It was in 1977 that a Californian team took the British Royal Aeronautical Soci
ety's Kremer prize by carrying out a take-off and a closed circuit flight between two
points half a mile (800 m) apart with the 32 kg Gossamer Condor (figure 6.19.).
Two years later, an improved version, the Gossamer Albatross (wingspan of
30 m and weighing 25 kg) crossed the English Channel and succeeded in flying a
distance of 36 km in almost 3 hours.
The Gossamer Albatross consists of a cabin suspended from a large braced wing
with the wingspan of a 100-seat airliner. A directional sail is fitted at the front and
the pedal-operated propeller, 4 m in diameter is fixed to the rear of the cabin.
With the exception of the braces, cables, and the propeller drive mechanism, the
structure of the Gossamer Albatross is completely made from organic materials. The
main strengthening member is made of epoxy-carbon fibre composite tubes produced
by filament winding. The ribs of the wing and the propeller are of expanded polysty-
 Elastic behaviour of solids 147

Figure 6.19. The Gossamer Condor making the closed circuit flight for the Kremer prize
(after Grosser, photo: T.Akoma, 1981)

rene reinforced with carbon fibres. The extreme end of the ribs of the wing towards
the trailing edge is made of aromatic polyamide fibres (Kevlar - figure 5.11) and the
wing is an ultra thin and ultra light sheet of poly(ethylene terephthalate) (Mylar -
figure 5.6.).
The significant improvements in performance realised during the two years
between the Gossamer Condor and the Gossamer Albatross resulted from the choice
of materials. The Gossamer Condor was constructed of aluminium aviation tubing
and balsa wood. The Gossamer Albatross was almost entirely built from synthetic
organic materials.
To obtain the highest performance in flight, the structure has to be as rigid and
as light as possible. It can be shown that for a given degree of rigidity in flexure of a
beam, the larger the ratio 4Ejp then the lighter the mass of the structure. In this
relationship, is the density and the modulus of elasticity.
The figure 6.20 shows the relative positions of various materials in a modulus of
elasticity - density diagram. The dotted line gives the values of and correspond
ing to constant values (-jE/p = Q for the performance index. If only two properties,
and p , are taken into account, all the materials which lie on the line have, in princi
ple, the same performance.
The dotted line in figure 6.20, which is shown for a value of constant C close to
8 ((GPa) tm~ ), allows the most interesting materials for aviation to be identified.
05 3

As can be seen in figure 6.20; wood, carbon fibre reinforced composites (CFRC),
some ceramics and diamond seem most attractive.
However, the final choice of a material is not made exclusively on a modulus of
elasticity and density basis. Other properties and other aspects such as the cost and
the ease of fabrication of large parts must also to be considered. Inevitably, this
eliminates diamond from materials suitable for aeronautics. Ceramics in general are
too brittle for this use. They cannot sustain mechanical and thermal shocks. Wood is
inhomogeneous because of the presence of knots and it is sensitive to humidity.
However, wood was widely used in the early days of aviation.
148 Introduction to Materials Science

The interesting position of composites based on carbon fibres or aromatic poly

amide (Kevlar) fibres (figure 6.20.) makes them attractive to the aviator, which ex
plains their increasing use in commercial aviation. In high performance gliders, the
use of composites has been widespread for many years. In the area of materials, ul
tra-light aviation is evolving much more rapidly than commercial aviation.
In the future, progress linked to research in materials science combined with
computer aided design will very likely lead to a significant improvement in the per
formance of materials used in the field of aeronautics.

0,1 1,0 1 0

Density (t/m )
3

Figure 6.20. Selection chart for materials for an aircraft wing. The dotted line gives the
values of and corresponding to constant values V/p = C) for the performance index,
(after Ashby, 1993).

6.7. Exercises
6.7.1. A filament of polyamide (Nylon) 66 (modulus of elasticity = 2GPa) of 1
mm diameter is deformed elastically by a 50 load. Calculate the deformation of the
filament under the action of this load as well as the elastic energy stored in the fila
ment as a result of the deformation.
6.7.2. Determine the value of the Poisson's Ratio of a material, which deforms at
constant volume. In which type of material is this behaviour encountered?
6.7.3. During a free fall "bungee" jump (i.e. attached to an elastic cord) the potential
energy gained by the human body is equal to U = mgl. In this expression, m is equal
to the mass, g is the acceleration due to gravity (9.81 m s ) and / the length of the
-2
 Elastic behaviour of solids 149

elastic cord (20 m) of which the diameter is 20 mm and the Young's modulus
0.01 GPa. Calculate the extension of the elastic during a jump with an individual of
65 kg. To simplify the calculation, the height of the free fall can be considered as the
length of the elastic corde.
6.7.4. A tensile stress a i s applied along the axis of a cylindrical aluminium bar of 15
mm diameter. Assuming that the deformation is totally elastic, calculate the maxi
mum applied stress which would limit the lateral contraction of the cylindrical bar to
2 . Poisson's ratio for aluminium is equal to 0.35 and the modulus is 70 GPa.
6.7.5. A iron block has a volume V = 1 d m at atmospheric pressure. What would be
0
3

the change in the volume of this block if it were submerged at 1500 m below sea
level? It can be assumed that the temperature of the water remains constant. The
density of seawater is equal to 1.025 t m" and the compression modulus of iron
3

is 71.7 GPa.
6.7.6. Explain why the temperature of a rubber band increases when it is stretched.
6.7.7. The different forces acting on the tubes of a bicycle frame can be represented
by studying the case of a beam with a circular cross-section of diameter d and length
L subjected to a force F at its free end. In the elastic region, the deflection is:

s = F4
3End 3

where represents the Young's modulus. Neglecting the actual weight of the beam,
show that wood is preferable to aluminium to minimise the mass of the beam for
the same value of tensile rigidity. It is given that the elasticity modulus of wood in
the direction parallel to the fibres is 10 GPa and the density is approximately equal to
0.5 g cm" . For what reason are bicycles no longer made of wood?
3

6.7.8. What is the class of materials of which the specific fracture energy G is clos
c

est to the theoretical value, equal to twice the specific surface energy y?
6.7.9. A sample of rubber, 12 cm long and with a square cross-section of l x l cm, is
stretched at 20 C to a length of 30 cm by applying a stress of 2 MPa. Calculate:
The Young's modulus of this elastomer and the number of moles of
elastic segments per cubic centimetre;
The stresses which must be applied to this sample to stretch it to a
length of 20 cm at 20 C and to a length of 30 cm at 100 C.
6.7.10. Calculate the Young's modulus of an elastomer obtained by the vulcanisation
of a polybutadiene with a molecular weight M of 10 gmol" and after vulcanisation
5 1

with a molecular weight for the elastic segments M equal to 5-10 gmol" .
c
3 1

6.8. References and complementary reading

M. F. ASHBY, Materials Selection in Mechanical Design, Butterworth-Heinemann, Oxford, 1993.
A. H. COTTRELL, The Mechanical Properties of Matter, John Wiley, New York, 1964.
R. W. HERTZBERG, Deformation and Fracture, Mechanics of Engineering Materials, 4th ed., John
Wiley, New York, 1996.
150 Introduction to Materials Science

M. GROSSER, Gossamer Odyssey, The Triumph of Humanpowered Flight, Houghton Mifflin, Boston,
1981.
L. R. G. TREOLAR, The Physics of Rubber Elasticity, 3rd ed., Clarendon, Oxford, 1975.
I. M. WARD, D. W. HADLEY, An Introduction to Mechanical Properties of Solid Polymers, Wiley,
Chichester (U.K.), 1993.
 Chapter 7.

Structural defects in crystals

7.1. Objectives
To make a distinction between a perfect and an real crystal
To describe the point defects.
To discuss dislocation theory and to summarise the main characteristics
of line defects, dislocations.
To describe dislocation movement.
To understand the mechanism of plastic deformation.
To describe the structure of two-dimensional defects, boundaries.
The atomic arrangement of a real crystal as described in chapter 3 differs locally
from the structure of a perfect crystal. The presence of defects in the crystalline
structure considerably modifies material properties.
One of the most important aspects of Materials Science is to characterise these
defects, analyse their effects on properties and to study methods, which allow them to
be induced in materials in a controlled manner so as to obtain an optimum behaviour.
Three types of structural defects are encountered in crystals:
Point defects, which have approximately the volume of an atom.
Linear defects or dislocations, which are disturbances of the crystal
structure situated along a line of atoms.
Two-dimensional or planar defects, which bring into play such im
perfections as those situated at an interface separating two crystals.
This chapter will be restricted to the study of defects, which have at least one
dimension of approximately atomic size. Larger defects such as pores or fissures will
not be considered here, but will be treated in chapter 14.

7.2. Point defects

7.2.1. Vacancies and solute atoms
Despite its name, a point defect (figure 7.1.) occupies a certain volume. The
point defects, in common with all other types of defect, deform the crystal and gener
ate a stress field, which encompasses a volume much greater than that of the defect
itself. The arrows in figure 7.1 show this schematically for the first neighbours.

2 0 0 2 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

152 Introduction to Materials Science

Three principal types of point defects can be identified:

The vacancy due to the absence of an atom from a normal site in the
crystal lattice (A).
The interstitial atom occurs when a small solute atom (B) enters an
empty space in the crystal lattice between the atoms( 4.2.2). When one
of the constituent atoms of the crystal is positioned interstitially, then an
auto-interstitial (') is formed.
The substitutional atom occurs from the replacement of a constituent
atom of the crystal, situated in its normal position, by a solute atom (C,
c).

Figure 7.1. Types of point defects in a crystal: (A) Vacancy; (B) Interstitial solute atom;
(') Auto-interstitial atom; (C, C ) Substitutional solute atoms.

Vacancies constitute an important type of point defect as they facilitate the dis
placement of atoms in a crystal. This phenomenon, known as diffusion, is sketched in
figure 7.2. Diffusion greatly influences the behaviour of materials at high tem
perature (creep, thermal treatment).

(4) (5) (6)

Figure 7.2. Mechanism of vacancy migration of atoms in a crystal lattice: exchange of

positions between atoms and vacancies (1 to 6).

7.2.2. Concentration of vacancies

There is an equilibrium concentration of vacancies for each temperature. This
equilibrium concentration (molar fraction XI) increases with temperature, as can be
 Structural defects in crystals 153

seen in figure 7.3. for copper, and tends to zero as the temperature approaches 0 K.
At a temperature of 1300 K, close to the melting point (1356 K), the concentration is
about 1 for 2500 sites (0.04 % ) . Therefore, at this temperature, there exist 3 . 1 0 19

vacancies per cm of copper (7.5 .10 atoms cm" ). At low temperature, the equilib
3 22 3

rium concentration is considerably reduced: at ambient temperature, is equal to

10~ as against 3.10~ close to the melting point.
17 4

Molar fraction X/

Figure 7.3. Variation of the equilibrium vacancy concentration (molar fraction Xj) of
copper as a function of temperature.

The vacancy concentration of (figure 7.4 (a)) increases suddenly on melting,

which is normally associated with a volume increase of several per cent. As was
mentioned in paragraph 3.2.1., the voids ,which appear in liquids, are spaces of vari
able dimensions while in solids the vacancies all have the approximate volume of an
atom. In liquids, the concentration of voids strongly increases with temperature.
Similarly, the concentration of solute atoms varies considerably with tempera
ture. This solubility also depends on the crystal structure. For carbon in iron
(figure 7.4 (b)), the solubility of interstitial carbon is considerably less in the a phase
(bcc) than in the / p h a s e (fee).
The introduction of a defect into a crystal is accompanied by an increase in
internal energy U or enthalpy (H ~ U for condensed phases). For a vacancy, the
heat of formation AH/is of the order of 1 eV (97 k J m o l ) . This increase is induced
-1

by the displacement of the atoms from equilibrium positions around the defect.
The creation of a defect is accompanied by a significant increase in the en
tropy S of the system as the defect produces additional disorder in the crystal struc
ture. Up to a defined concentration, the formation of defects is accompanied by a
decrease in the free enthalpy G of the system:
AG = AH - AS (7.1.)
The thermodynamic equilibrium is reached, at constant temperature and pres
sure, when the free enthalpy G is a minimum or, what amounts to the same thing,
when the variation of G with concentration (aGlaXi) is zero. This situation is illus
trated in figure 7.5. Knowing the free enthalpy of formation AG/of a mole of defects,
the equilibrium defect concentration can be calculated with the aid of equation (7.2.)
below.
154 Introduction to Materials Science

Figure 7.4. Schematic representation of the variation of the defect concentration as a

function of temperature: (a) vacancies in a pure metal (Cu); (b) interstitial carbon atoms
in iron.

Figure 7.5. Variation of the free enthalpy AG of a crystal as a function of the vacancies
molar fraction X. AH and AS represent respectively the corresponding variations in en
thalpy and entropy. is the vacancies equilibrium concentration at temperature T.

Thus for an crystal containing identical atoms and point defects, the molar
fraction of vacancies can be calculated:
Xi. = nIN = exp(-AG /RT)
e f (7.2.)
In this expression, R is the gas constant (R = 8.314 J m o l ^ K " ) . The formation of
1

vacancies is a reaction, which is accompanied by an absorption of energy (endother-

mic reaction) and an increase in disorder (increase in entropy). It follows that the
concentration of vacancies increases with temperature.
The concentration of vacancies in a crystal may be higher than the equilibrium
concentration. Thus, rapid cooling (quenching) from a high temperature allows a
 Structural defects in crystals 155

relatively high level of vacancies to be retained at low temperature. The word,

quenching, is derived from the fact that, with steel, this operation was almost always
carried out by immersion in water.
Keeping the material for a period of time at a temperature sufficiently high for
diffusion to be important eliminates this supersaturation of vacancies. It is also pos
sible to obtain a higher level of vacancies by irradiating the material with ionising
radiation ( rays for example). The presence of a high level of vacancies can appre
ciably modify the behaviour of metallic materials, especially during phase transfor
mations.

7.2.3. Point defects in ionic solids and organic polymers

In monoatomic crystals such as metals, point defects are created in an independ
ent manner. In ionic crystals, electrical neutrality must be maintained. Defects are
created in pairs with opposing signs. It is possible to distinguish between Schottky
defects, which are composed of an anionic vacancy and a cationic vacancy and
Frenkel defects consisting of an arrangement of a vacancy (anionic or cationic) and
an interstitial ion (anion or cation) (figure 7.6). In the case of Schottky defects, the
vacancy created by the loss of a cation is negatively charged while that created by the
loss of an anion is positively charged. In Frenkel defects a negatively or positively
charged vacancy is formed according to whether the atom in the interstitial position
brings a positive or negative charge in excess to this part of the crystal.

Figure 7.6. Point defects in ionic crystals (example: NaCl): (a) presence of two vacancies
of opposite charge (Schottky defect); (b) formation of a vacancy-interstitial atom pair
(Frenkel defect).

The same types of defect found in metallic and ionic solids can be encountered in
organic polymers. The crystalline structure of organic polymers also contains a large
number of defects as a result of the chain structure. Some point defects in polymers
are due to structural irregularities in the chemical composition of the chain. Several
structural irregularities have already been encountered such as head-to-head chain
formation by two structural units (5.9) or the chain end making up another type of
point defect (figure 7.7.). A third category of point defect consists of junctions where
branches are joined to the main chain (figure 4.10.).
156 Introduction to Materials Science

In polymers exhibiting chirality such as polypropylene (figure 5.2.), tacticity de

fects are produced. A syndiotactic unit (figure 7.8.) in an isotactic polymer chain is
an example.

Chain ends constituting point defect

Figure 7.7. Schematic representation of point defects introduced by the chain ends in a
polymer crystal.

(a) Defect (~)

Figure 7.8. Structural defect in an isotactic polymer chain due to the presence of a syn
diotactic unit.

(a) (b)

Figure 7.9. (a) Polyethylene chain and (b) Reneker defect in polyethylene (after Reneker,
1962).

Similarly, point defects are found in polymers, which are due to irregularities in
the crystal morphology. The point defects (figures 7.9. and 7.10.), described by
Reneker, result from a change in configuration of the chain passing from a trans-trans
to a trans-gauche configuration. This is immediately followed by a gauche-trans
 Structural defects in crystals 157

configuration. This type of configuration has been described for polypropylene in

figure 5.9. After the two inversions shown in figure 7.9. (b), the chain regains a trans-
trans configuration. The trans-gauche configurations have a potential energy 3.5
k J m o r higher than that of the trans-trans configuration, which is the most stable
1

configuration in the polyethylene crystal. At temperatures just below the melting

point, this morphological irregularity propagates along the chain, which imparts
creep movement. This mechanism plays an important role in lamellae thickening
during polymer crystal annealing.
Trans-trans conformation

Gauche-trans-conformation

Figure 7.10. Exploded view of a Reneker defect. This is made up of a trans-gauche con
figuration followed by a gauche-trans configuration.

7.3. Linear defects or dislocations

7.3.1. Dislocation Geometry
A dislocation is a linear crystal defect. The following experiment can be imag
ined in order to visualise the geometry of this defect: a crystal with an ideal cubic
structure (fig. 7.11 (a)) is cut along a plane ABCD by displacing the upper part by
one interatomic distance. Two particular displacement directions exist in this plane:
When the displacement of the atoms is made in a direction perpendicu
lar to the AB line (figure 7.11. (b)) the upper part of the crystal is com
pressed and an extra half-plane of atoms ABEF, terminating inside the
crystal along the dislocation line (line AB) is formed. This extra half-
plane of atoms is inserted into the crystal rather like an edge, from
where the name edge dislocation given to this type of defect comes. On
the outside of the crystal it is possible to observe the formation of a step
C C ' D D ' . The width of this step shows the displacement of atoms along
the sectional plane.
If the displacement of the atoms occurs parallel to the dislocation line
AB (figure 7.11. (c)), a helical distorsion of the crystal having the line
AB as axis is induced. This linear defect is called a screw dislocation.
The external indications of the slippage of the atoms are the steps
ADD' and B C C .
All other displacements in the plane ABCD forming some angle with the line
AB produces a mixed dislocation, i.e. a dislocation having at the same time a screw
component and an edge component.
158 Introduction to Materials Science

Figure 7.12. gives another representation of an edge dislocation. It is a positive

edge dislocation (symbol ) with the supplementary plane above the line. In the part
of the crystal situated above the dislocation line, the crystal is in compression. Below
the dislocation line, on the contrary, the atoms are stretched out and the crystal is in

(a)

(b)

(c)

Figure 7.11. Geometry of simple dislocations in a cubic crystal: (a) no defect; (b) edge
dislocation; (c) screw dislocation. For the two types of dislocations, the vector (Burgers
vector b), which characterises the displacement of the atoms of the crystal, is indicated.

Figure 7.12. Perspective view of the structure of an edge dislocation in a cubic crystal
(after Kittel 1976). The dislocation line, perpendicular to the figure plane, is in the centre
of the drawing.
 Structural defects in crystals 159

tension. The displacement of the atoms from their equilibrium positions in the neigh
bourhood of the dislocation line increases the internal energy of the system. A nega
tive edge dislocation (symbol T) is obtained through a rotation of 180 about the
dislocation line of figure 7.12.
The screw dislocation does not contain a supplementary plane. The lattice planes
perpendicular to the dislocation line are slightly deformed and wind themselves into
a helix along the dislocation line. The pitch of this helix can be left- or right-handed,
thus defining the left or right-handed character of the dislocation.
The concept of dislocations has been introduced in a simple manner by making a
translation of atoms from one part of an ideal crystal parallel to a sectional plane
(figure 7.11 (a), plane ABCD) known as a slip plane. The dislocation moves in the
crystal in this slip plane. Such dislocation movement in a crystal translates itself into
the formation of a step, which permanently changes the form of the crystal. This
phenomenon is known as plastic deformation.
The vector, which characterises the amplitude and the direction of the displace
ment of the atoms in the dislocation, is known as the Burgers vector, b. It is deter
mined by the method of Burgers circuit summarised in figure 7.13. This represents a
lattice plane, perpendicular to the dislocation. The Burgers circuit, which, by con
vention, follows a clockwise direction, is carried out starting from no matter which
atom in the lattice plane and it consists of a equal number of identical steps in all
directions. If the circuit is carried out in a lattice plane of a perfect crystal (figure
7.13. (a)), the vector b is equal to 0. If the circuit is made around a edge dislocation
(figure 7.13. (b)) or a screw dislocation (figure 7.13. (c)), then the path does not close
and the closure defect of the circuit is a vector which defines Burgers vector. In fig
ure 7.13. (c), for a screw dislocation of right-hand pitch, the Burgers vector points in
the negative sense of the dislocation line. For a left-hand pitch screw dislocation, the
situation is obviously reversed.
The Burgers vector and the dislocation line of an edge dislocation form an angle
with each other and define therefore one plane (the slip plane). The Burgers vector
and the dislocation line of a screw dislocation are parallel (figure 7.13. (c)). In this
case, and contrary to an edge dislocation, the slip plane is not defined as there is an
infinity of lattice planes passing per the dislocation line. A screw dislocation has
therefore more freedom to move than the edge dislocation, which only possesses a
single gliding plane.
If a dislocation takes the form of (mobile) loop, it changes character along the
line (figure 7.14). It may be seen that at A, the dislocation has an exclusively screw
character while at it is uniquely edge. Between A and B, the dislocation is of mixed
type, i.e. intermediate between screw and edge. A dislocation line never terminates in
the interior of a crystal. It extends from one surface to another (which can be an
interface such as a grain boundary). It can also be shown that the Burgers vector
remains constant along the whole length of the dislocation. In effect, the displace
ment of the atoms is the same although the dislocation line changes direction. Figure
7.15 shows a view of the plane ABC of the dislocation in figure 7.14. It is therefore
possible to observe the transition between the pure screw dislocation (A) and the pure
edge dislocation (B).
By completely closing the circular loop in figure 7.14., it can be seen
(figure 7.16) that at two points A and A', the dislocation has a pure screw character
160 Introduction to Materials Science

(a)
2 3 4
>
5
3

4 VJ 2
3
12 3

(b)

r 3

i L
1
1 2
X X

(c)

-t
Figure. 7.13. Burgers circuit enabling the Burgers vector b to be determined. The plane
of the circuit is perpendicular to the dislocation line shown by a large arrow, (a) Circuit
in a perfect crystal (vector b - 0); (b) circuit around an edge dislocation; (c) circuit
around a screw dislocation.

while at two other points, and ' , it has a pure edge character (the dislocations are
each case of opposite sense). The rest of the loop is made of mixed dislocation. This
type of mobile dislocation loop plays a key role in plastic deformation, which will be
further examined in figure 7.21. and in chapter 12.

Figure. 7.14. Example of a curved dislocation line having edge and screw components.

Another type of dislocation loop exists which remains immobile in the crystal.
When a metal is cooled from high temperature, a significant number of vacancies
must disappear because the equilibrium vacancy concentration decreases with tem
perature (figure 7.3.). Several mechanisms are involved in the elimination of these
excess vacancies, such as migration to the surface. For the vacancies inside the crys
tal, an elimination mechanism involves regrouping into lenticular clusters in an atomic
 Structural defects in crystals 161

Figure. 7.15. Perpendicular view of plane ABC of figure 7.14. showing the transition
from a pure screw (A) to a pure edge dislocation (B) in a quarter turn. The crosses repre
sent atoms in a lower plane and the circles in an upper plane when compared to the slip
plane. The black circles represent atoms in the upper plane displaced by the dislocation
(after Read, 1953).

plane (figure 7.17.). This vacancy loop has an entirely edge configuration; the
Burgers vector b is perpendicular to the plane of the clusters and can only move
along the surface of the hatched cylinder (e). The vacancy loop remains immobile in
the dislocation plane. Its behaviour is thus noticeably different from that of the
gliding loop in figure 7.14. However, this second type of loop plays an important role
in particle precipitation which takes place during structural hardening (chapter 12.).

Figure. 7.16. Slip dislocation loop formed by 4 1/4 loop of the type in figure 7.15.

The process of dislocation movement in slip planes is the basis of the low tem
perature (T< 0.4 T ) plastic deformation mechanism in most crystalline materials,
m

principally metals. This phenomenon leads to a permanent change in the shape of the
material (figure 7.11.). If the number of slip dislocations is high, the formation of
multiple microscopic steps brings about a macroscopic change in shape.
162 Introduction to Materials Science

oooooooooooooo
oo ooooooooooo
/x oooooo ooooooo
w oooo oooo oooo
ooooooo oooooo
oooo ooooooooo
oooooooooooooo
oooooooooooooo
oooooooooooooo
, ooo o o o o o o o o o o
(b)
i u
ooooo oooo
oooo ooooooooo
oooooooooooooo
oooooooooooooo

ooooo ooooooooo
ooooo ooooooooo
ooooo ooooooooo
(c) oooo oooo
ooooo ooooooooo
ooooo ooooooooo
ooooo ooooooooo
oooooooooooooo
oooooooooooooo
(d) SS83gg88S8

oooooooooooooo
oooooooooooooo

Figure 7.17. Formation of a vacancy loop (prismatic loop) by the progressive accumula
tion of vacancies (a-c) and relaxation of neighbouring atoms (d). The loop is immobile
and does not contribute to the plastic deformation.

The process of plastic deformation is considerably helped by the presence of

dislocations. In effect, in their absence, very high shear stresses would have to be
applied to displace one part of a crystal relative to another, as it would be necessary
to break simultaneously all the bonds adjacent to the slip plane. This process would
demand stresses of the order of one tenth of the shear modulus. As will be seen in
chapter 12., plastic deformation of metal crystals takes place under the action of
stresses much lower than E/10 because of the easy displacement of the dislocations
present in this type of material.

7.3.2. Movement of dislocations

Dislocation movement along slip planes is the basis of the low temperature
(T< 0.4 T ) plastic deformation mechanism in the majority of crystalline materials,
m

principally metals. This leads to a permanent change in the shape of the material
(figure 7.11.). If the number of slip dislocations is high, the formation or multiple
microscopic steps brings about a macroscopic change in shape.
The process of Plastic Deformation is, considerably helped by the presence of
dislocations. In effect, in their absence, very high shear stresses would have to be
applied to displace one part of a crystal relative to another, as it would be necessary
 Structural defects in crystals 163

to break simultaneously all the bonds adjacent to the slip plane. This would require
stresses of the order of one tenth of the shear modulus. As will be seen in chapter 12,
plastic deformation of metal crystals takes place under the action of stresses much
lower that G/10 because of the easy dislocation movement in this type of material.
Figure 7.18 shows schematically how dislocations simplify plastic deformation
in metallic crystals. During the passage of an edge dislocation across a crystal, there
are few atoms involved and few bonds breaking simultaneously.
As an analogy to dislocation movement, let us discuss, the displacement of a
very long carpet laid on a flat surface. The easiest way to move this carpet on the
ground is to make a fold at one end and to push this fold along the whole length as is
described in figure 7.19. Caterpillars use an analogous mechanism to move. This is
also the process involved in dislocation movement.

(a) (d)

(b) (e)

(c) (0

Figure 7.18. Movement of atoms involved in the displacement of an edge dislocation

under the action of a shear stress . When the dislocation emerges from the crystal, a
step (f) is formed of which the width corresponds to the absolute value of Burgers vector.
The dislocation line is perpendicular to the plane of the figure.

The steps on the surface of a prismatic crystal after the passage of dislocations
give no information about the type of dislocations, which have crossed the crystal.
Figure 7.20. shows schematically the movement of an edge dislocation (a, b, d) and a
screw dislocation (a, c, d) caused by the same forces. When the dislocation has
crossed the crystal then in both cases, the formation of the same step is achieved,
with the depth equal to the Burgers vector. The displacement of the dislocation line
occurs always in a direction perpendicular to the dislocation line and is parallel to
the applied stress in the case of an edge dislocation and perpendicular to it for a
screw dislocation.
Figure 7.21. shows the plastic deformation mechanism for a slip dislocation loop
(figures 7.15. and 7.16.) inside a crystal. As the movement of the dislocation in the
slip plane is always perpendicular to the line, the loop increases in size until it
emerges at the surface, at first partially and then completely.
164 Introduction to Materials Science

Figure 7.19. Movement of a carpet by the displacement of a fold formed at one end: (a)
initial position of the carpet; (b) creation of a linear defect; (c) and (d) displacement of
the fold; (e) final position of the carpet, after complete displacement of the linear defect
(the fold).

(d)

Figure 7.20. Formation of a slip step (plastic deformation) in a prism by the movement
of an edge or a screw dislocation: (a) non-deformed prism; (b) passage of an edge dislo
cation; (c) passage of a screw dislocation; the prism deformed plastically by the passage
of either dislocation is shown in (d).
 Structural defects in crystals 165

Figure 7.21. Plastic deformation by the enlargement of a slip dislocation loop of the
same type as shown in figure 7.16.

For metallic crystals (with isotropic bonds), the preferred slip planes for dislo
cations are the crystallographic planes with high atomic densities: {111} planes for
crystals with fee structure, {110} planes for bcc crystals and {0001} planes for hep
crystals.
For ionic crystals, the movement of dislocations is more difficult as the actual
slip system does not uniquely depend on the atomic density. The preferred means of
slip are those that do not bring ions of the same sign into contact. For the fee crystal
of MgO or NaCl, the slip planes are the lattice planes {110}. In crystals with co
valent bonds, the oriented nature of the bonds makes dislocation movement difficult.
It is for these various reasons that ionic and covalent ceramics have few dislocations
with very limited mobility. The plastic deformation of ceramics is difficult and these
materials generally have a brittle behaviour. It is only at very high temperatures that
significant plastic deformation is observed.
The presence of dislocations has been demonstrated in polymers, notably in
mono-crystals. In semicrystalline polymers, the movement of the dislocations is
seriously restricted by the presence of chains and propagation across long distances is
made more difficult because of the presence of inter-crystalline tied chains. These
materials are therefore mostly brittle.

7.3.3. Dislocation density

The dislocation density is a measure of the total length of the dislocation lines
per unit volume. Dislocation density is expressed as:
dislocation length
P= ;; (7.3.)
crystal volume
In metals, the dislocation density increases considerably during plastic deforma
tion. Characteristic values of the density for an annealed metal are 10 to 10 cm .s 6

For the case of highly deformed metals, values of of 10 cm" have been observed.
12 2
166 Introduction to Materials Science

This corresponds to 10000 000 km of dislocation lines per c m of deformed crystal.3

In a highly deformed crystal there are some 1 0 atoms per c m affected by

20 3

dislocations. As 1 cm of a metallic crystal contains about 10 atoms, the number of

3 23

atoms disturbed (~10~ ) remains relatively small and the crystalline structure is
3

mostly conserved.

7.4. Two-dimensional defects

7.4.1. G r a i n Boundaries
Many materials (figure 3.3. (b)) have a poly crystalline structure, i.e. made up of
small crystals (grains). The regions where crystals are in contact are known as grain
boundaries. These constitute transition zones with a more or less disturbed structure,
which permits the geometric and crystallographic fitting together of the crystals.
In the majority of cases, the grains are in the form of polyhedrons with multiple
facets and the boundary structure depends on the orientations of the faces of these
polyhedrons with respect to one another. Figure 7.22. gives a schematic optical mi
croscope view of a section cut through a polycrystalline material after polishing and
chemical etching .The grain boundaries of metallic materials are more readily at
tacked by acids, making them visible with an optical microscope (chapter 10.). In this
figure, continuous lines represent grain boundaries and dotted lines show sub-
boundaries. The differentiation between sub-boundaries and boundaries is generally
made on the basis of the value of the angle of rotation or tilt between grains situ
ated on either side of the boundary. A sub-boundary has a value of less than 10 .

gram
(crystal)
100

Figure 7.22. Schematic optical microscope view of a section cut through a polycrystal
line material after polishing and chemical etching. The grain boundaries are shown with
solid lines and sub-boundaries in dotted lines inside two grains.

Two types of simple boundary exist: tilt boundaries containing the tilt axis of
the adjacent grains (figure 7.23. (a)) and the twist boundaries perpendicular to the
rotation axis (figure 7.23. (b)). The boundaries can be described by their dislocation
systems. A distinction can be made between tilt sub-boundaries, the structure of
which can be considered as a parallel arrangement of edge dislocations and twist sub-
boundaries with a system of screw dislocations. Consideration here will be restricted
to the case of tilt boundaries.
 Structural defects in crystals 167

Figure 7.24. (a) shows the structure of two matching simple cubic grains, with a
misorientation angle of a ~8 degrees. In the grain boundary described in figure
7.24. (b), a regular network of edge dislocations appears. If the angle between the
grains remains small (< 10 ), it is possible to identify individual dislocations, which
are separated by a distance D, given by:

(7.4.)

(a) (b)

Figure 7.23. Types of boundaries: (a) tilt boundary; (b) twist boundary.

Figure 7.24. Simple grain boundary (tilt) in a cubic system: (a) transition zone between
two crystals with different orientations before relaxation of atom positions; (b) regular
distribution of edge dislocations at the boundary after relaxation of the positions of the
atoms in the region of the boundary (after Read, 1953).

In this expression, b represents the absolute value of the Burgers vector. The
introduction of each dislocation leads to an increase in internal energy. Generally, the
internal energy associated with the grain boundary increases with an increasing angle
of misorientation 0.
When the misorientation angle is large (> 10 ), it is no longer possible to
distinguish individual dislocations. The crystal perturbations are too numerous and
the grain boundaries are regions of considerable atomic disorder but of very limited
168 Introduction to Materials Science

thickness (2 to 3 atomic distances). In this case, the boundary energy remains

practically constant while, under certain conditions, some sharp minima for well-
defined values of are observed. There is a partial superposition of two adjacent
crystals and formation of a super structure known as coincidence lattice. For these
particular orientations, the boundary has a less perturbed structure, which reduces the
energy.
Grain boundaries are both sources and traps for point defects and dislocations.
During quenching, excess vacancies are eliminated much more rapidly near grain
boundaries. Grain boundaries also play an important role in plastic deformation,
since they can form dislocations under the influence of a stress field. They also con
stitute obstacles for dislocation movement (chap. 12).

7.4.2. Stacking faults and twins

Various anomalies occur in the stacking of planes, especially in compact crystal
structures (such as fee and hep). An assembly of two fee or hep planes, taken sepa
rately, has the same structure, and anomalies only appear if an assembly of more than
two planes is considered.

Figure 7.25. Twin in a face-centred cubic {fee) structure: (a) plane (110) (points 1, 2, 3,
4); (b) section perpendicular to dense planes (111) shown as horizontal lines with the
atoms of the (110) plane corresponding to rectangles (after Verhoeven, 1975).
 Structural defects in crystals 169

As was seen in chapter 4. (figure 4.1.), the face-centred crystal structure (fee)
has compact hexagonal layers following a sequence ABCABCABC. A stacking fault
is constituted by an abnormal sequence such as, for example, ABCABABC... It can
happen that the structure ABC ABC... jumps to the symmetrical structure CBACBA...
following the sequence ABCABCBACBA... In this case a twin boundary occurs. The
two parts of the crystal have a compact plane in common, which is a plane of sym
metry (figure 7.25.). The atoms of the twin boundary are common to the two crystals
situated on either side and the interfacial energy of this type of boundary is small.

7.5. Summary and conclusions

The real crystal deviates considerably from the perfect crystal model described
in chapter 3. It contains a large number of defects of which the smallest dimensions
are generally of the order of an atomic diameter. These defects have a well-defined
geometry. They can be point defects, as vacancies or solute atoms, linear defects, as
in dislocations, or two-dimensional defects, like grain boundaries, stacking faults or
twin boundaries.
The concentration of point defects generally increases with temperature. For
vacancies, this explains the much greater mobility of atoms (diffusion) at high tem
peratures. The dislocation is the only type of linear defect. Their displacement is the
basis of plastic deformation (slip) in metals and alloys. Slip, always perpendicular to
the dislocation line, is aided by the localised nature of the defects. During the move
ment of a dislocation in a crystal, there are few atoms involved and few bonds are
broken simultaneously. Due to a high dislocation density and the ease with which
they are displaced, a significant macroscopic deformation can be obtained in a short
time in metals through cold and hot forming processes (forging, rolling, drawing).
Other types of material (ceramics, cross-linked organic polymers and thermoplastics
at temperatures below the glass transition temperature) have few or no dislocations
and are brittle. It is necessary to use thermoforming processes (at high temperature
with respect to their melting point) to work ceramics and thermoplastics.
Grain boundaries are regions of attachment between grains (mono-crystals) of
different orientations. When the angle of misorientation between the grains is only a
few degrees, the boundaries consist of networks of dislocations and are known as
sub-boundaries. With greater misorientation, the defect density is higher and a sys
tem of dislocations can no longer describe the structure of these boundaries. The
grain boundaries are then disordered regions, structurally similar to a glass, but with
a thickness limited to two or three atomic distances.
Only a strict control of the type and number of defects allows materials with
controlled properties to be obtained.

7.6. Illustrative example: structural defects of diamond films

Diamond, which is the hardest material we know, has always fascinated man
because of its exceptional properties. Since the Middle Ages, scientists have tried to
170 Introduction to Materials Science

find the origin of its extreme hardness and have searched to synthesise it. To make
this material requires particular conditions of pressure and temperature. It was only in
the 1950s that experimental conditions were found which led to the industrial
production of diamond. Pressure of the order 10 GPa (100 kbar) and temperatures
close to 1500 C are necessary.
The recent discovery of low pressure manufacturing processes, which enable
diamond films to be deposited on different substrates, has revived the interest in this
material. In effect, numerous applications difficult if not impossible to realise at high
pressure have become feasible. Consequently, it is possible not only to exploit the
hardness of diamond but also to take advantage of other properties of this unusual
material. These include high thermal conductivity, which can be used to cool effi
ciently electronic components (integrated circuits, diodes for power lasers, etc.),
chemical inertness, which allows diamond to be its used in corrosive environments or
in the biomedical area, or its optical and semiconducting properties.
In the well-established diamond machine-tools sector, it is now possible to de
posit continuous diamond films which perform better than classic coatings made
from diamond particles embedded in a metallic matrix.
The deposition of a diamond film from the vapour phase by chemical reaction
(CVD = Chemical Vapour Deposition) is carried out under experimental conditions,
which are far from thermodynamic equilibrium. This process uses an active gaseous
mixture (plasma) mainly containing hydrocarbons and hydrogen and operates at
moderate temperatures of the order of 500 to 1000 C and below atmospheric pres
sure. A complex series of physico-chemical processes leads to the growth of thin
layers. The mechanisms involved are still poorly understood, but initially the forma
tion of small islands is observed, which grow progressively and finally touch laterally
thus forming generally polycrystalline, continuous films. The thickness deposited
varies from several tenths to several hundreds of microns per hour, depending on the
method used to activate the precursor gases.
Graphite is the stable form of carbon under the experimental conditions used to
generate diamond by CVD. Consequently, during the production of the diamond
film, a number of unwanted phases appear which interfere with the crystallisation of
the film. The layers of CVD diamond therefore contain chemical anomalies (chemi
cal bond irregularities, impurities, etc.) and crystalline structural defects, in addition
to grain boundaries. These strongly influence the final properties of the CVD dia
mond films, in particular the thermal conductivity, which decreases significantly as
the number of defects increases. Vacancies and impurities have a great influence on
the optical and electronic properties of diamond. Impurities are responsible for,
among other things, the colour of diamond; for example, nitrogen and boron give rise
respectively to yellow-orange and blue layers.
The study of film structures gives important information about the properties of
diamond layers and the way in which they grow. Two types of crystalline defects are
founds: twins and dislocations. In figure 7.26., planes of symmetry, characteristic of
twinned structures can be observed in the diamond crystals. Figure7.27. (transmission
electron microscopy) shows an assembly of dislocations forming small angle grain
small angle grain boundaries indicated and dislocations appearing in the form of
black lines (in isert).
 Structural defects in crystals 171

The study of these dislocations provides essential information concerning the

growth mechanism of thistype of layer (figure 7.28.) hi particular, the grains are all
oriented in the same way. However, in this case, the presence of dislocations induced
during crystalline growth accommodating the slight misorientation between
neighbouring grains is observed. Defects are also present in natural diamonds. Their
nature and distribution allow the specialist (gemmologist) to determine the origin of
this precious mineral.

Figure 7.26. Micrograph, taken by scanning electron microscopy, of grains in a layer of

CVD diamond produced at low temperature (about 420 C). This image reveals the
presence of twins marked with asterisks, with parallel planes of symmetry.

1 Insert

Figure 7.27. Dislocations in the grains of diamond layers: micrograph taken by

transmission electron microscopy and showing different types of defects: small angle
grain boundaries indicated by the white arrow and dislocations appearing in the form of
black lines in insert.
172 Introduction to Materials Science

Figure 7.28. Topology of the surface of the same layer taken by scanning electron
microscopy. The only visible defects are grain boundaries characterised by a step on the
surface of the diamond layer.

7.7. Exercises
7.7.1. What are the different types of point defects existing in crystals?
7.7.2. The equilibrium concentration of vacancies is given by relation (7.2.). Calcu
late the number of vacancies per cm of copper at 1000 C, given that: the atomic
3

weight of copper m = 63.54 gmol" ; the density p, 8.4 gem" ; the Avogadro's num
n 1 3

ber, A 6.02 10 atoms mol" ; the activation energy for the formation of a va
= 23 1

cancy AG = 0.144-10" J a t o m ; the ideal gas constant R = 8.314 J m o l ^ K " . Com

f 18 -1 1

pare this result with that reported in figure 7.3.

7.7.3. What are the types of stresses acting in the neighbourhood of an edge disloca
tion?
7.7.4. What is the difference between an edge dislocation and a screw dislocation
with respect to the orientation of the Burgers vector b related to the dislocation line?
7.7.5. Determine the Burgers vector b of an edge dislocation associated with slip in
the (110) plane and the (111) direction in the face-centred cubic structure of NaCl.
7.7.6. What is the behaviour of two edge dislocations of opposite signs placed on the
same slip plane when a shear force is applied perpendicularly to the dislocation
line?
7.7.7. Calculate the length of the Burgers vector b for F e (bcc) and Al (fee).
a

7.7.8. The circumference of the earth is approximately 40 075 km. This is equivalent
to the length of the dislocation lines (placed end to end) in 1 c m of copper at 20 C.
-3

Calculate the dislocation density p.

7.7.9. What is the result of the movement of an edge dislocation when it crosses a
crystal and emerged at the surface?
7.7.10. How is the tilt boundary between two simple cubic grains described when the
misorientation angle is smaller than 10 ?
 Structural defects in crystals 173

7.8. References and complementary reading

J. FRIEDEL, Dislocations, Pergamon, Oxford, 1964.
B. HENDERSON, Defects in Cristalline Solids, Arnold, London, 1972.
D. HULL, D. J. BACON, Introduction to Dislocations, 3 ed., Pergamon, Oxford, 1984.
th

J. PHILIBERT, A. VIGNES, Y. BRECHET et P. COMBRADE, Metallurgie: du mineral au materiau,

Masson, Paris, 1998.
W. J. READ Jr., Dislocations in Crystals, Mc Graw-Hill, New York, 1953.
D. H. RENEKER, Points Defects in Crystals of High Polymer Molecules, J. Polym. Set, 59, S 39 (1962).
J. SCHULTZ, Polymer Materials Science, Prentice-Hall, Englewood Cliffs, New Jersey, 1974.
E. L. THOMAS ed., Structure and Properties of Materials vol. 12 in Materials Science and Technology,
R. W. CAHN, P. HAASEN and E. J. KRAMER eds., VCH, Weinheim, 1993.
J. D. VERHOEVEN, Fundamentals ofPhysical Metallurgy, Wiley, New York, 1975.
J. WEERTMAN, J. R. WEERTMAN, Theorie elementaire des dislocations, Masson, Paris, 1970.
 Chapter 8.

Alloys and phase diagrams

8.1. Objectives
To distinguish between constituents and phases.
To introduce the concept of alloys.
To describe phase equilibria at a given temperature.
To determine composition and the relative quantities of phases in
equilibrium.
To define solubility by the study of phase diagrams.
To introduce interfacial energy.
The properties of pure materials are frequently of limited or no interest for tech
nological applications. Even the highly purified silicon is only of value for elec
tronics after the controlled addition of a small amount of a solute element (doping),
otherwise silicon is an insulator.
The majority of materials are mixtures of different types of atoms or molecules.
Sometimes the material constituents (atoms or molecules) are soluble in all propor
tions and the mixtures are homogeneous. Many times, the constituents are only par
tially miscible. The material then contains several phases with different compositions
and structures. The combination of these phases produces various microstructures
having a considerable influence on the properties and technical characteristics of the
materials. Phase diagrams provide the basis for analysing microstructure formation
(chap. 9). They define the equilibrium states of phases (corresponding to state at
infinite time).

8.2. General aspects

8.2.1. Alloys
An alloy is a combination of two or more metallic and, possibly, non-metallic
elements. Initially this was restricted to metals, however the concept of alloys, which
is evolving rapidly, has progressively spread to other materials (ceramics, polymers)
although the term blend is also widely used for polymers. An alloy is now defined as
a combination of atoms, ions or molecules forming a material having different prop
erties from those of the constituents.
An alloy generally consists of several microscopic or nanoscopic phases of
different composition and structure, which contribute synergistically to the material
properties. Carbon steel is a well-known example of a two-phase alloy. With a mix-

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

176 Introduction to Materials Science

ture of 99.2 wt.% Fe and 0.8 wt.% C, the microstructure consists of grains containing
alternate thin lamellae of ductile F e and hard iron carbide Fe C which are the two
a 3

phases in this material (figure 8.1.). A high strength steel such as this is used for the
cables of cable cars. Aluminium is used in aeronautics and automobile because of a
high AE/p ratio (chap. 6). In order to increase strength (hardness), it is alloyed with
elements such as copper, zinc and magnesium, which form precipitates of interme-
tallic compounds in the aluminium matrix. Intermediate or intermetallic compounds
have a definite stoichiometry (atomic composition), as in the Al Cu phase. 2

Figure 8.1. Microstructure of a carbon steel (0.8 wt.% C ) observed with an optical mi
croscope. It consists of two phases: lamellae of Fe and Fe C. The structure is known as
a 3

perlite.

In ceramics, by combining four elements (Si, , , N) "Sialon" type alloys are

produced consisting of two chemical compounds, silicon nitride S i N and alumina 3 4

A 1 0 . These engineering ceramics are destined to replace metal alloys in high tem
2 3

perature applications.
Polymer alloys are a recent innovation and are under continuous development.
By alloying a high modulus of elasticity organic phase with an elastomer, a hard,
resilient thermoplastic is obtained (figure 8.2.), which is used for telephone housings.

8.2.2. Phases and their constituents

Each material consists of a large number of atoms (~10 c m ) or molecules.23 -3

The chemical elements of metals, the molecules of polymers and the chemical com
pounds (such as pure oxides) forming ceramics are called constituents.
Pure Cu is not an alloy. It has good electrical conductivity, but insufficient me
chanical strength to withstand high stresses. Another constituent can be added to
increase hardness, Sn for example, but any improvement of the mechanical properties
is accompanied by a decrease in electrical conductivity. These alloys are called
bronzes and contain two phases: Cu + Cu Sn . x y

The alloy (hard metal) used for drilling bits contains three constituents, Co, W
and C, forming two phases: tungsten carbide (WC) present in the form of tiny grains
and - 1 0 vol.% cobalt metal, which is the ductile matrix bonding the hard, brittle
carbide grains.
Ruby, used in the construction of lasers, is a mono-crystalline alloy of A 1 0 2 3

(sapphire) and 0.05 % C r 0 . T h e characteristic red colour of this crystal is due to the
2 3
 Alloys and phase diagrams 177

Figure 8.2. Microstructure of high impact polystyrene consisting of rubbery particles

dispersed in a phase with high mechanical strength. The partially cross-linked rubbery
particles (a) are composed of two finely dispersed phases of grafted styrene (white) and
butadiene (black line) copolymer while the continuous phase (b) is pure linear thermo
plastic polystyrene in white.

presence of C r ions dispersed in solid solution in the sapphire crystal ( 14.5.5.).

3+

The alloy consists of only one phase although it has two constituents.
An alloy is therefore composed of two or more constituents and contains one or
more phases. A phase is a part of the system in which the composition (constituent
type and concentration) and the atomic arrangement (crystalline or amorphous
structure) are fixed. In a system in equilibrium, each phase has a fixed composition,
which is entirely homogeneous. In non-equilibrium systems, which are frequently
encountered in materials, the composition can vary as a function of time and position
within the material. Each phase is separated from other phases by an interface. The
amorphous and crystalline phases have different atomic arrangements. Elemental
iron (Fe) can exist in the vapour, liquid, face-centred cubic or body-centred crystal
states, four different phases. Under very high pressure, such as that at the centre of
the earth, iron forms a fifth, hexagonal phase.

8.2.3. Solid solutions

The phenomenon of solid substances dissolving in solvents is generally under
stood intuitively. Although more difficult to grasp, the solubility of a solid in another
solid is an entirely analogous case. The mixture of two elements A and in the solid
state forms a solid solution in which the major element is the solvent and the minor
element, the solute. For a crystalline solid solution, the structure of the phase formed
by the principal constituent A is not changed by the addition of the solute even
though the overall chemical composition is different. In chapter 7, it was shown that
there are interstitial solid solutions and substitutional solid solutions. The solubility
range is the range of concentration over which the solution is stable.
In some cases, the constituents A and of the alloy -B have the same crystal
structure. The solubility of A in B, or in A, can be complete so that there is only
one solid phase. This occurs with others in the Ni-Cu system. Both constituents are
soluble in all proportions and keep the same crystal structure (fee phase), even
though the distance between atoms in the lattice varies slightly with composition
since the copper and nickel atomic radii are similar but not equal. When the diff-
178 Introduction to Materials Science

erence in atomic radii between solvent and solute is large, the same crystal structure
will not give complete solubility but only a limited solubility.
Limited solubility is also observed when the crystal structures of the two con
stituents are different, as it is impossible to form a continuous transition from one
crystal structure to another. This limited solubility also occurs when the alloy con
stituents have a strong tendency to chemically react. Intermediate compounds are
then formed which have a definite stoichiometry (atomic composition), as in the
Al Cu phase. Theses are known as intermediate or, if formed between two metals,
2

intermetallicphases (figure 8 . 3 . , ^ phase).

This type of behaviour can also be found in ceramics, although their structure is
more complex.
Crystalline solid solutions of polymers, or of polymers and small organic mole
cules, are rare. However, polymers and small molecules often form solutions in the
glassy and liquid states. PVC dissolves low volatility organic liquids with a molecu
lar weight close to 500 known as plasticizers. Amorphous polymers have little com
patibility and only rarely form miscible blends. A commercialised example of com
patible polymers is Noryl which is a mixture (alloy) of polystyrene and
poly(oxy(2,6-dimethyl)-l,4-phenylene) (table 17.6.) in variable proportions.

f
QA O b
Interface

Figure 8.3. Two cubic crystalline phases: the a phase is a substitutional solution of
atoms in a crystal of A atoms. The phase is a type -B intermetallic or intermediate
compound.

8.2.4. Interfaces between phases

A phase occupies a definite volume, limited in space by an interface, which
separates it from other phases (figure 8.3.). Grain boundaries are a particular type of
interface since they mark the boundary between two crystal grains of the same com
position and structure but with different orientations. The creation of an interface
requires an amount of work to be done and is therefore unfavourable from a thermo
dynamic point of view. It increases the free enthalpy of the system (figure 8.4). Fine
grained systems have many interfaces and a higher free enthalpy than large grain
systems. Grains tend to grow at high temperatures, which reduces the free enthalpy.
This phenomenon is known as grain growth.
An interface is characterized by its specific interfacial energy / [ J n f ] , which, in
2

case of liquids, can be regarded as an interfacial tension [Nm ]. This specific inter
-1

facial energy can be determined by measuring the work required to create a unit inter-
 Alloys and phase diagrams 179

Interface

AI

Figure 8.4. Variation of the free enthalpy G of a system in passing from phase a to phase
via an interface of thickness / of the order of several atomic diameters. If the system
is in equilibrium, G = 0, from which G = G .
a

face. When a condensed phase is in contact with a gas, the interfacial energy is called
surface energy. The interface plays a key role in phase transformations (chapter 9) as
well as in rupture (chapter 14.).

8.3. Single-constituent systems

8.3.1. Phase Diagram of a pure substance

As previously mentioned ( 3.2.1.), all substances can exist in three distinct
physical states: crystalline solid, liquid or gas. A pure substance exists in principle in
these three forms (phases) as a function of temperature and pressure. A phase equi
librium diagram is shown in figure 8.5. This is divided into three areas, in relation to
the stability ranges of the crystal, liquid and vapour phases. The curves marking the
boundaries between the phases define the temperatures at which two phases coexist.
Three curves meet at the triple point T' where three phases coexist in equilibrium:
crystalline solid, liquid and vapour. At the critical point C, liquid and vapour become
indistinguishable.
Above its critical temperature, a gas can no longer be liquefied. When pressure
goes up, the density of the gases increases and can reach values close to that of
liquids. Under these conditions, supercritical gases behave as solvents. During its
synthesis at about 250 C and a pressure greater than 200 MPa (-2000 bars), poly
ethylene remains dissolved in supercritical ethylene.
In general, solids have higher densities than liquids and the melting temperature
increases slightly with pressure. However, some substances (water, Bi and Ga) in
crease in volume during crystallisation and therefore their melting point is lowered
when pressure is increased. On the whole, pressure has little influence on the equi
librium between condensed phases (i.e. liquids or solids), and its effect can often be
ignored (for metals, aTJaP = ~lK/kbar).
In a system composed of constituents, the thermodynamic conditions for phase
equilibrium are expressed by the Gibbs phase rule, which can be written as:
180 Introduction to Materials Science

0
Wi)
Temperature

Figure 8.5. Phase diagram of a pure substance showing the stability areas of the crystal,
liquid and vapour. is the triple point, C, the critical point, T(Pi), the melting point at
pressure Pi and T (P ), the boiling temperature at pressure Pi.
V X

(8.1.)
is the number of constituents of the system and p, the number of phases. Consid
ering the most useful variables (composition, temperature and pressure), + 1 vari
ables must be defined in a system in thermodynamic equilibrium: ( - 1) molar frac
tions, and P. A certain number of these variables can be fixed arbitrarily without
changing the physical state of the system. This is known as the number of degrees of
freedom or the variance f of the system.
When the pressure is fixed, the variance is reduced by one:

+f= + 1 ( ^constant) (8.2.)

By applying the phase rule to the phase diagram of a pure substance (fig
ure 8.5.), = 1. It can be shown that the total of the variance / and the number of
phase is always equal to three. In other words the variance of a system consisting of
a pure substance with three phases in equilibrium is zero. Therefore, there is only one
point (the triple point T' with a unique value for and T) in the equilibrium diagram
(figure 8.5.), where the three phases are in equilibrium. The curves ', T'B and T'C
in figure 8.5. give the temperatures and pressures at which two phases co-exist. In
this case, the variance / i s 1 for points situated on one of the three curves. Either the
temperature or the pressure can be freely chosen (the second variable being fixed by
the choice of the first). The areas between the curves ( ' - T ' B , T ' B - T'C for
example) make up the regions of the phase diagram where only one phase is in
equilibrium. Here the v a r i a n c e / = 2 and, within limits, the temperature and pressure
can be selected independently.
The free enthalpy G of a phase is related to the enthalpy and entropy S by the
well-known thermodynamic equation (see for more details: Smith, 1990):

G = H-TS (8.3.)
 Alloys and phase diagrams 181

As the entropy S in less ordered phases is higher, the free enthalpy G decreases
more quickly with temperature than in more ordered phases. This is why the free
enthalpy of a crystal (figure 8.6.) varies less than the free enthalpy of a liquid. This,
in turn, varies less than the free enthalpy of a gas, which corresponds to the most
disordered state of a substance.

Temperature -

Figure 8.6. Variation at constant pressure of the free enthalpy of the three stable phases
of a pure substance A as a function of temperature. The point A corresponds to the
melting point T , to the boiling point T , and C to a hypothetical transition between
m v

solid and vapour. The upper part of the curves, shown in fine lines corresponds to non-
equilibrium thermodynamic states .

The curves for the variation of molar free enthalpy shown in figure 8.6., inter
sect at two points A and which correspond, at constant P, to the temperatures at
which the solid and liquid phases (A) and the liquid and vapour phases (B) co-exist.
These two points A and B, where the free enthalpies of the crystal and the liquid, and
that of the liquid and the vapour are equal, define the melting point T and the boiling
m

point T . The most stable phase always has the smallest free enthalpy. The point C
v

marks a hypothetical transformation between the crystal and the gas. At the triple
point, the three enthalpy curves intersect at a point and A, and C meet, which
unequivocally determines the temperature and pressure of the triple point.
Numerous pure substances exist in more than one crystalline forms. This phe
nomenon is known as allotropy or polymorphism. Iron, at atmospheric pressure,
exists either in a body-centred cubic ( F e and Q = ferrite) or a face-centred cubic
a 8

(Fe = austenite) form, depending on temperature.

y

8.3.2. Behaviour of glass forming substances

As mentioned in chapter 3, the behaviour of organic and inorganic polymers is
quite particular, as they form highly viscous liquids crystallising with difficulty. A
large number of organic polymers exist with very irregular molecular structures that
are not capable of crystallising. This is so for atactic or heterotactic vinyl polymers
( 5.2.1.), which constitute a class of thermoplastics of considerable industrial im
portance.
182 Introduction to Materials Science

At low temperatures, liquids, which are unable to crystallise, form glasses.

These have the structure of liquids without long range order and form optically iso
tropic and transparent materials. Glasses possess some of the properties of a crystal,
such as a high modulus of elasticity.
The transformation from the crystal to the liquid state is marked by a discon
tinuous change in the volume and enthalpy of the system. The change of a pure
crystalline solid into a liquid occurs at a well-defined melting point T . For glasses, m

the solid-liquid transition (glass transition) is much less well marked.

The mobility of liquids, which is measured by their viscosity ( 6.4.1.), is the
result of a high concentration of voids having approximately the volume of an atom.
The presence of these voids, which have great mobility, allows a liquid to flow under
the influence of weak shear forces. The reduction in volume associated with the
cooling of liquids contains both vibration and configurational terms. The vibration
term is similar to that characterising the volume change in crystalline solids. It results
from a small variation in the mean interatomic distance caused by an anharmonic
non-symmetric character of the atomic vibration (figure 8.7.). The readjustment of
atomic distance with temperature is practically instantaneous.
The configurational rearrangements of the liquid structure result from a varia
tion in the number of voids. This configurational reorganisation is associated with a
readjustment time (relaxation time), which increases exponentially as the temperature
is lowered. The viscosity of the non-crystallisable liquid polymer liquids can reach
very high values ( = 10 Pa s). In comparison, the viscosity of liquid water is only
13

10~ Pas.
3

The lengthening of the reorganization time leads to a significant diminution in

the volume expansion coefficient as shown in figure 8.8. The polymer progressively
acquires the properties of an solid (glass) with a high elasticity modulus (E > 1 GPa).
The volume-temperature variation curve is characterised by a singular point
characterized by two values for the tangent. This temperature marks the extrapolated
limit between the viscous liquid and the amorphous solid (glass) and defines the
glass transition temperature, T . g

Interatomic distance r

Figure 8.7. Progressive change of the mean interatomic distance (from r to r ) with in
0 3

creasing the temperature. The mean interatomic distance increases because of the an
harmonic character of the atomic vibrations. This results from the asymmetric form of
the potential energy function U(r). The slope of the left-hand part of the curve, resulting
from the repulsive forces between atoms, is steeper in absolute value than the slope of
the right hand part induced by the attractive forces ( 2.4.1).
 Alloys and phase diagrams 183

As can be seen in figure 8.8, the glass transition temperature depends on the
cooling rate = |dr / di| and the transition between the glassy solid and liquid is not
as well defined as the transformation temperature (T ) between the crystal and liquid
m

state which is a first order thermodynamic transformation. On cooling a glass-form

ing liquid more slowly, the configuration rearrangement continues on to a lower tem
perature and the glass transition is shifted to a lower temperature. Thus when the rate
of cooling varies by an order of magnitude, the glass transition temperature varies by
3 to 5 for organic glasses. There is always a time scale associated with a T deter g

mination. Therefore, it is more accurate to refer to a glass transition range.

Temperature
Stable state
Non stable state (glass)
Specific volume (v)

Figure 8.8. Variation of the free enthalpy G and the specific volume during vitrification
of a non-crystallisable organic polymer at two cooling rates and v and
a
are
the volume expansion coefficients of the glass (g) and the liquid (/).

In the glass transition range, the variation in free enthalpy progressively deviates
from the curve corresponding to the equilibrium state, and the molecular configura
tion depends on the whole thermal history of the system. A glass is therefore never in
thermodynamic equilibrium.
Even liquids able to crystallise vitrify if they are cooled sufficiently rapidly
(chapter 9.). As shown in the illustrative example in chapter 4, metal alloys form
glasses if cooled at rates of 10 to 10 degrees per second. If an atomic or a low mo
5 9

lecular weight liquid is slowly cooled, a discontinuous volume change occurs during
crystallisation. The free enthalpy variation and the change in volume are discontinu
ous (bold lines in figures 8.9. (a) and (b) respectively). When the cooling rate is high,
crystallisation does not occur. The curves for the variation in volume V and free en
tropy G shown in dotted lines in figure 8.9. are of application. The glass transition
occurs by the blocking of some degrees of freedom related to the translation move-
184 Introduction to Materials Science

1
/^Liquid
!!
1 I /
1 1 / ^Supercooled
(b) '/ liquid
V
\
8.
CO Glass ^ 1/ J

Crystal | |
^ 1 1 J g2
T

Temperature
Stable state
Non-stable state (glass)
Specific volume (V)

Figure 8.9. Variation of the free enthalpy G (a) and the specific volume (b) as a func
tion of temperature for the crystalline, liquid and glassy states. 7j and 7^ represent dif
ferent cooling rates (T\ < T2 ) The bold lines show the curves corresponding to ther
modynamic equilibrium. The non-stable state is shown with dotted lines.

ments of atoms or molecules. In the glassy state, the structure of the liquid state has
been solidified by blocking the translation movements.
It is important to distinguish systems in thermodynamic equilibrium from me-
tastable and unstable systems. A single constituent system in thermodynamic equi
librium is completely defined if the pressure and the temperature are known. A
crystal at <T and a liquid at > T are stable systems. At constant P, a stable
m m

system has a minimum value of free enthalpy G. When a stable system is displaced
from equilibrium by a perturbation, it tends to return spontaneously to equilibrium if
the perturbation ceases. At a given temperature, a system may exist in several states,
each having its own value of the free enthalpy. Close to T , a system can be in the
m

liquid or the solid state and each of these states has a separate function G(T, P). At
any temperature, the state, which has the lower free enthalpy, forms. Under certain
 Alloys and phase diagrams 185

conditions, (T or ), two distinct phases, with the same value of G, can coexist
m

(figure 8.6.) However, phases not corresponding to the minimum value of G can also
be found. When a liquid is cooled to < T , it is frequently observed that the liquid
m

does not crystallise within a limited temperature range (up to several hundred degrees
in some cases). This (supercooled) liquid therefore has a free enthalpy G(T, P) value
greater than that of the crystal and is then said to be in a metastable state.
If a supercooled liquid does not crystallise, it is because there is a threshold,
related to nucleation that must be overcome (section 9.3). In figure 8.9., the curves
for the temperature variation of free enthalpy are shown in bold lines for stable states
and in fine lines for metastable states. A metastable state is separated from a stable
state by a threshold. In an unstable system, the threshold between the unstable and
the stable states is practically zero. Glasses have an excess free volume compared
with the liquid in thermodynamic equilibrium. If a glass is taken to a temperature
close to T , this excess empty volume progressively diminishes. This phenomenon
g

has a small threshold and the glass volume can change so long as it has not reached
the limiting value, which is the volume of the liquid in thermodynamic equilibrium
(stable or metastable).
Inorganic polymer liquids, such as molten silica or window glass, crystallise
with great difficulty and when cooled form inorganic glasses. The glass transition
temperature of window glass is close to 750 K. For inorganic glasses, T varies by g

some 15 when the cooling rate is changed by an order of magnitude.

Semi-crystalline polymers ( 5.2.3) contain both micro-crystalline and amor
phous phases. When a crystallisable liquid polymer is slowly cooled (the rate de
pends on the polymer chain structure), a first transformation, with partial crystalli
sation of the polymer is observed and, at a lower temperature, a glass transition (T ) g

is detected (figure 8.10.). When a crystallisable organic polymer is quenched, the

polymer does not have sufficient time to crystallise and only the glass transition is
observed.

Figure 8.10. Variation of the specific volume as a function of the temperature for a
crystallisable polymer: (1) slowly cooled and (2) quenched. For many polymers, the in
terval T, - T is of the order of 150 to 200 and the width of the melting zone is -50 K.
g
186 Introduction to Materials Science

The heating and cooling curves of semi-crystalline polymers are more or less
similar if they are readily crystallisable, such as linear polyethylene, and if the heat
ing and cooling rates are low. Melting (or solidification) occurs over a relatively broad
temperature range (up to 50 K) due to the presence of thin lamellar microcrystals
(chapter 10). In polymers, the melting point T is defined as the temperature at which
m

the last crystal trace disappears on heating. In general, T varies with the heating rate
m

and differs significantly (up to 100 lower) from the thermodynamic melting point
which corresponds to the melting of a polymer crystal of infinite dimensions. For
organic polymers, T^can only be determined by extrapolation.
The degree of crystallinity greatly varies between polymers and is a function of
chain flexibility: in practice from 80-90 % for a highly flexible linear polyethylene to
25-30 % for the much more rigid poly(ethylene terephthalate) (PET) (figure 5.6. (f)).
Sometimes the crystallisation rate is so sluggish that the polymer never crystallises
from the melt (bisphenol-A polycarbonate, polyethersulphone).

8.4. Phase equilibriums in binary systems

8.4.1. Miscibility of two constituents
This section will consider the conditions under which a two-constituent mixture
in equilibrium can form a solid or liquid solution or exist as two distinct solid or
liquid phases.
Figure 8.11 shows the variation of the free enthalpy of mixing G for a phase a m

as a function of composition at constant temperature and pressure. The values of

at XB = 0 and XB 1 are the free enthalpies of the pure constituents A and B. This
=

thermodynamic quantity varies with temperature as shown in (8.3.). When the system
-B contains two phases, one (a) solid and the other (/) liquid, two curves are ob
tained ,one for each phase, for the variation of the enthalpy G (figure 8.12. (a)) as a
function of the molar fraction X . The shapes of the curves are similar, but the abso-
B

Figure 8.11. Variation of the free enthalpy of mixing G, for a phase a at constant and
as a function of the molar fraction of element B.
 Alloys and phase diagrams 187

lute values for G for both pure constituents (X = 0 or X = 1) and the mixture vary
m B B

with temperature. The molar fraction X is define by the following expression:

B

X = - - n M
(8A)
n +n
B
A B

with rij is the number of moles of constituent /.

For concentrations less than X , the free enthalpy of solid a is less than that of
0

liquid / (figure 8.12. (a)). For concentrations greater than X the free enthalpy is
09

greater (G <G%)- Since at constant and P, the most stable systems are those with
l
m

the lowest free enthalpy, crystal a is more stable than liquid / for concentrations of
X between 0 and X ,but between X and 1 the liquid phase / is favoured.
B 0 0

There is an additional possibility of decreasing the free enthalpy of mixing and

of reaching a more stable system close to X (figure 8.12. (b)). By making a mixture
0

of the two phases and / between X" and X given by the tangent common to the
l
e9

Figure 8.12. Variation at constant and of the molar free enthalpy G of a binary sys
tem: (a) free enthalpy of mixing of a binary system -B containing a solid phase a and a
liquid phase /. The functions G and G , which give the variations in free enthalpy for
l
m

the phases a and / respectively, intersect at Go corresponding to the concentration X ; (b)

0

the part of the diagram corresponding to the heterogeneous mixture of a {Xf) and
1{X[).
188 Introduction to Materials Science

two curves, a two-phase system is formed which is even more stable than either of
the two phases a or/. For concentrations of X between X* < X < X , the most stablel
e

system is formed from a mixture of two phases with compositions of X" and X[
respectively, in such proportions that the overall composition of the system isX, so
that the following relation is satisfied:

X?fa + 4fl =X (8.5.)

For the binary system shown in figure 8.12., there are three distinct regions of
concentration for a given pressure and temperature: for 0 < X< X", the crystal is
the most stable phase (homogeneous region) at equilibrium. If X[ < X < 1, the liquid
is the only stable phase. For concentrations X < X < X[, a heterogeneous mixture
a
e

of two phases is found, made up of solid a and liquid / in thermodynamic equilib

rium. By generalising this concept over several temperatures, a phase equilibrium
diagram, also known as a phase diagram, can be constructed.

Figure 8.13. Binary system with two totally miscible constituents. Variation of the free
enthalpy of mixing for the system (a to f) and corresponding phase diagram (g).
 Alloys and phase diagrams 189

A phase (or equilibrium) diagram (at constant P) for constituents A and B, solu
ble in all proportions in both solid and liquid states, is illustrated in figure 8.13. (g).
At low temperature, the crystalline phase a is stable, while at higher temperature, the
liquid phase becomes more stable. In the temperature range between the melting
points and T of the pure constituents, a cigar-shaped area forms (hatched in fig
B

ure 8.13. (g)) where two phases (crystal a + liquid I) coexist.

The concentrations defining the boundary between the two-phase area and the
liquid phase, (liquidus curve) and the boundary between the solid phase a and the
two-phase region (solidus curve) are determined by constructing the common tangent
to the G and G curves, as in figure 8.12. This is also shown in figures 8.13. (c) and
l
m

(d). At the melting point of the pure constituents A and B, G% and G% are equal, i.e.
at A, G = G% and at 7B, G^ = G (figure 8.13. (b) and (e)).
L
k

The changes in the shape of the free enthalpy of mixing curves G (X) must be m

examined to show the considerable variations in miscibility which occur in some

binary systems. In figure 8.13., the shape of the G (X) curves hardly varies with tem-
m

/

V

(a)

Molar fraction X B

Figure 8.14. Formation of a miscibility gap: (a) variation of the free enthalpy of mixing
of a system containing two constituents A and showing reduced solubility at low
temperature; (b) phase diagram.
190 Introduction to Materials Science

perature T. However, the G (X) curves shown in figure 8.14. for a binary system with
m

a miscibility gap for various temperatures are very different. In figure 8.14., G (X) m

increases with a decrease in temperature. In the free enthalpy G expression above, at

the critical temperature T , there is a minimum in the G (XB) curve. At temperatures
c m

below T , there is a maximum between two minima. AtT<T

c the unique phase with
c

an structure separates into two phases with the same ^structure but different
compositions, shown in figure 8.14. by a \ and a (at ) and a [ a n d a{ (at T ). The
2 2

compositions of a and a are determined at each temperature by the common

2

tangent method on the curves in figure 8.14. (a).

The phase diagram in figure 8.14. (b) shows the stability limit of phase as a
function of T. If T> T , the solubility of A and in is complete. A miscibility gap
c

occurs with < T , shown by the hatched region in the phase diagram. In this region,
c

the two-phase system is more stable than the single-phase system. As mentioned in
paragraph 7.2.2, solubility usually increases with temperature.
An equilibrium diagram not only allows the solubility limits of the elements to
be determined as a function of temperature, but also permits the calculation of the
relative amounts of the phases in isothermal equilibrium. A phase diagram must
always be read isothermally, i. e. parallel to the concentration axis.
In figure 8.15., the cigar-shaped Ni-Cu diagram is shown. Ni and Cu are misci-
ble in all proportions. Insert (b) is an enlargement of part of this diagram and shows
how the relative quantities of phases in equilibrium can be determined by applying
the inverse segments rule. Since f +f = 1, it can be shown, from (8.5.) that the frac
a

tions of phases a and / are respectively given by:

f a = - and//=-2.- ( )
8 6

a+b a + b

Concentrations in molar fractions can be converted into volume fractions if the

molar volumes of the constituents are known.

Figure 8.15. Determination of the relative quantities of phases by the inverse segments
rule: (a) Ni-Cu equilibrium diagram. In (b), an enlargement shows the principle of the
inverse segments rule.
 Alloys and phase diagrams 191

8.4.2. Phase diagrams with complete solubility - Thermal analysis

The simplest phase diagram is formed when the two constituents A and are
miscible in all proportions in the solid and the liquid. The number of binary systems
forming a single crystalline phase in the solid state is relatively limited. For metals,
the conditions required for two elements to form a solid solution are defined by the
Hume- Rothery rules. The following essential conditions must be satisfied:
The elements have the same crystal structure..
The atomic dimensions do not differ by more than 15 %.
The elements have the same valency.
The Hume-Rothery rules are not inflexible laws. Nevertheless, they provide a
useful guide for predicting solubility. Ni-Cu (figure 8.15.) and Au-Ag are two sys
tems where these rules apply. The Hume-Rothery rules are also used for ceramic
alloys. The ceramic alloy MgO-NiO forms a completely miscible system which
obeys these rules. These two oxides crystallise with a face-centred cubic system, the
nodes of the cube being occupied by oxygen anions. The Mg and Ni cations occupy
the interstitial sites, which are of sufficient size to accommodate either cation.
Phase diagrams can be constructed from cooling curves using a technique
called thermal analysis. The temperature variations of the system are measured with
a thermocouple as a function of time during cooling or heating. Any discontinuity in
enthalpy related to a phase change (decrease or increase in enthalpy) with a release or
absorption of latent heat causes a change in the slope of the cooling or heating curve.
Differential Scanning Calorimetry (DSC) is also used for constructing phase dia
grams where variations in enthalpy or thermal capacity are measured relative to a
known inert reference. DSC is a more sophisticated and sensitive technique.
By the use of low cooling or heating rates, the influence of kinetic phenomena
(i.e. effects linked to the rate of phase formation (section 9.3.) can be reduced to a
minimum. For liquids and for systems which change phases rapidly (metals), trans
formation temperatures determined by thermal analysis are practically equal to the
equilibrium transformation temperatures.
The exothermic solidification of metal alloys can be observed by thermal analy
sis (figure 8.16.). A plateau occurs in the cooling curve of pure elements since the
decrease in enthalpy (latent heat release) takes place at one temperature, (T ). In
m

alloys, such a plateau is not observed, but there is a more or less marked variation in
the cooling rate between the liquidus and solidus temperatures (figure 8.16.).

8.4.3. Partial miscibility in the solid state

When the difference in atomic size or valency (for the same crystal structure) is
larger than postulated by Hume- Rothery rules, the free enthalpy as a function of
composition shows two minima separated by a maximum (figure 8.14.) and the cor
responding phase diagram shows a miscibility gap in the solid state. The solid-liquid
equilibrium curve often has a minimum melting point in this case. At this minimum,
the liquidus and solidus curves contact each other with a horizontal tangent (Au-Ni
system) (figure 8.17.). At concentration X , the cooling curve shows a plateau similar
m

to that of pure elements.

192 Introduction to Materials Science

A miscibility gap, representing a zone of phase separation, appears in the system

described in figure 8.17. Such a system, completely soluble at high (in the solid
state with a single a phase), form two phases, a,\ and a , on cooling, which have the
2

same crystal structure but very different compositions at low T.

When the constituents have different crystal structures, two solutions, a and
are formed in solid-state binary mixtures. In each phase, the mutual solubility of A in
and of in A is limited. Solution a is stable for pure A and for low concentrations
of in A, and solution for pure and for low concentrations of A in . These two
zones are separated by a two-phase region + (figure 8.18. c). If the melting points

Figure 8.16. Relation between the thermal analysis curves and the phase equilibrium
diagram for the case of a completely miscible binary system (a); thermal analysis curves
(b) for X = 0, 0.5, and 1; temperature, time, composition diagram(c). The cooling curve
B

for the mixture does not show a plateau section because of the existence of a transforma
tion range / + a (after Eisenstadt, 1971).
 Alloys and phase diagrams 193

of the a and phases are lowered by the presence of the other element, a eutectic
diagram is obtained.
The construction of this phase diagram is shown in figure 8.18 (a) and (c). At the
eutectic temperature, , three two-phase zones interpenetrate:

a) The two-phase region (a + I) situated between the constituent A melting

point ( ) having a stable a structure and that of constituent (T' ) having
B

the same a structure which now corresponds to a metastable state

characterised by T' < T . This diagram is similar to figure 8.15.
B B

b) The two-phase region ( + /) situated between the constituent melting

point with a stable structure and the melting point () corresponding,

for constituent A to a metastable structure with T < T . A A

c) The two-phase region delimited by the two solubility lines of in a

and of A in .
c

0 X m 0,4 0,6 0,8

^ u
Molar fraction X Wl

Figure 8.17. Phase diagram for the gold-nickel system showing a minimum melting point
and a miscibility gap in the solid state. At the minimum, the liquidus and solidus
temperatures are identical.

Three phases are in equilibrium at temperature and at composition XE: the

solid solutions a and and the liquid solution / . As the pressure is fixed and the
E E

number of constituents is 2, then from equation 8.2, the variance is equal to zero. The
eutectic point ( , XE) is therefore unique for this system which contains only three

phases in equilibrium at constant pressure. The pressure effect can be neglected in

equilibrium diagrams of condensed phases.
The silver-copper system presents such a eutectic (figure 8.19.). The thermal
analysis curves at compositions 1 and 4 (pure element and eutectic alloy) are of the
same type, i.e. they contain a plateau (dT/dt = 0) during the transformation. The
cooling curve for composition 2, situated outside the eutectic plateau ,only shows a
reduction in the cooling rate, similar to example 2 of figure 8.16. (b). For composi
tions situated between the points of maximum solubility of the a and phases (8.19.
3 and 5), the cooling curves show first a reduction in the cooling rate due to a and
crystallisation followed by a plateau of variable duration, which is due the transfor
mation of the residual liquid into a two-phase eutectic solid.
194 Introduction to Materials Science

Figure 8.18. Construction of an eutectic diagram for two constituents A and crystallis
ing with two different crystalline structures a and : (a) represents the liquidus-solidus
diagram for the constituents A and crystallising with a structure; (b) represents the
liquidus-solidus diagram for the constituents A and crystallising with structure; parts
of these diagrams are metastable (indicated in dotted lines); (c) The two-phase region de
limited by the two lines gives the maximum solubility of in and of A in^.; (d) com
bination of equilibrium diagrams a, b and c (stable parts only) to form the phase equilib
rium diagram for the eutectic; note the equilibrium between three phases at the eutectic
temperature: / = " ft.
E

A peritectic diagram occurs when the cigar-shaped two-phase regions (a + )

and ( + /) are inclined in the same direction relative to the concentration axis. The
construction of the phase diagram, shown in figure 8.20., is analogous to that for the
eutectic diagram of figure 8.18. At composition X? and temperature , three phases

are in equilibrium: the solid solution a , the solid solution? and the liquid solution
P

/P. In accordance with the phase rule and at constant pressure, a variance of zero can
 Alloys and phase diagrams 195

Ag Concentration CQ ^ (% poids)
U Temps t

Figure 8.19. Example of an alloy forming a eutectic: (a) phase equilibrium diagram for
the silver-copper system and (b) thermal analysis curves 1 - 5 picked out on the equilib
rium diagram. (1): polycrystal of pure a; (2): polycrystal of solid solution a; (3): crystals
of solid solution of a surrounded by the two-phase eutectic ( +)\ (4): eutectic (a+)
polycrystal; (5): crystals of solid solution of surrounded by the two-phase eutectic
(a+).

again be calculated for the peritectic point P. A binary peritectic system is found, for
example, in the iron-carbon diagram at temperatures close to the melting point of
iron (figure 8.23. (b)).

8.4.4. Systems containing intermediate phases

Intermediate phases or compounds (also known as intermetallics in metals),
having generally well-defined chemical compositions appear in many equilibrium
diagrams. For example, in the copper-magnesium system (figure 8.21.) two inter
metallic compounds occur, MgCu and Mg Cu. These intermetallic compounds,
2 2

which exist only in the solid state, have well defined and generally complex crystal-
lographic structures. The structure of MgCu is shown in figure 8.22.
2

In the copper-magnesium diagram (figure 8.21), there are successively three dis
tinct eutectics, between Cu and MgCu , between MgCu and M g C u and between
2 2 2

Mg Cu and Mg. This way of breaking down a complex phase diagram into a set of
2

elementary diagrams makes it much easier it to analyse alloys. In Cu rich alloys

(0 <Xu < 0.33), only the eutectic Cu-MgCu is considered. The other eutectics are
% 2

not involved in this concentration range. The same reasoning is valid for other re
gions of the diagram. The phase diagram also shows that Cu and MgCu can form 2

solid solutions, and Mg and Mg Cu have a very low solubility (zone of solubility less
2

than the thickness of a drawn line).

Some metals form intermediate compounds with carbon, known as carbides,
such as tungsten carbide (WC) and cementite (Fe C). These compounds have very
3

high hardness and are found in hard metals and steels.

Intermediate compounds also form in certain ceramics, the constituents of which
are already chemical compounds with fixed compositions. Thus MgO and SiO form
two stable ternary phases: forsterite ( M g S i 0 ) and enstatite ( M g S i 0 ) ( 8.5.2.).
2 4 3
196 Introduction to Materials Science

8.5. Applied phase diagrams

8.5.1. Metallic alloys: iron-carbon
Iron-carbon alloys, i.e. steels and cast iron, play a major role in current tech
nology. In spite of the appearance of new materials which compete with steels, the
iron-carbon system remains of great importance for industry. The iron-carbon equi
librium diagram is very complex, but it can be broken down into a series of simpler
diagrams.
For the Fe-C system, two possible equilibrium diagrams exist: one stable (Fe-C;
iron-graphite) and the other metastable: Fe-Fe C (Iron-cefnentite). For steels, with
3

little carbon, (< 1 wt.% by) the study can be limited to the metastable system, since
under practical conditions, cementite always forms (figure 8.23.). As already seen,
pure iron exists in two allotropic forms in the solid state: a body-centred cubic form
(a and iron: ferrite) and a face-centred cubic structure ( iron; austenite).

Figure 8.20. Binary peritectic diagram constructed using phase diagrams corresponding
to different possible equilibriums (a to c). The peritectic diagram (d) is obtained by com
bining, as in figure 8.18 for the eutectic, the three diagrams (a), (b) and (c). (e) Three-
phase equilibrium at the peritectic temperature T? and composition X?\ / + = .
P
 Alloys and phase diagrams 197

t\l085C

1000

\ MgCu 2

k
800 MgCu +/ 2

\ 722C \ / ' N/ Mg Cu + /
a
2

\ \ Mg Cu
2
, \ , 650C,
r Mg Cu + / /
600 - 2

a + MgCu 2
522C \^-

485c\/Mg+/
1 MgCu + Mg Cu 2 2
Mg Cu +Mg2

400 1 1 1 1 I 1 1 < 1
0 0,20 0,40 0,60 0,80 1
Cu Concentration \ ~ %
Mg

Figure 8.21. Phase diagram for the copper-magnesium system with formation of two in
termetallic compounds, MgCu and Mg Cu. The concentration X is given as a molar
2 2 Mg

fraction.

Cu Mg

Figure 8.22. Unit cell of the intermetallic compound MgCu (after Westbrook, 1967). 2

These allotropic varieties are stable within well-defined temperature ranges.

Pure a iron is stable up to 912 C (figure 8.23. (c)). At this temperature, it transforms
into / iron. This phase remains stable up to 1394 C (figure 8.23. (b)). At this tem
perature it re-transforms into a body-centred cubic crystal (-) melting at 1538
C. In reality a iron and iron are identical phases and correspond to the same allo
tropic variety. It is only for reasons of usage that distinction is made here.
The addition of a second element to a metal with one or more allotropic trans
formations generally modifies the equilibrium temperatures. According to the me
tastable diagram Fe-Fe C (figure 8.23. (a)), the phase (austenite) is stable up to
3

1495 C (figure 8.23 (b)) at a carbon concentration of 0.17 wt.%. At low tempera
tures, the phase remains stable down to a temperature of 727 C at a carbon con
centration of 0.8 wt.% (figure 8.23. (c)). The considerable variation of the solubility
of carbon within the crystal structure of the solvent is an essential characteristic of
198 Introduction to Materials Science

the iron-carbon system, as already mentioned (figure 7.4. (b)). The carbon is inter-
stitially dissolved in the a and /crystal structures. The austenite can dissolve up to
1.98 wt.% carbon at 1148 C, while -ferrite can only dissolve a maximum of 0.02
wt.% at equilibrium and the -ferrite, 0.09 wt.%.
At high temperatures and low carbon concentrations, the transformation is
described by the peritectic shown in figure 8.23. (b). The introduction of carbon has
the effect of lowering the melting temperature of -iron from 1538 to 1495 C and of
raising the temperature of the iron transformation from 1394 to 1495 C. At
the latter temperature, for an overall carbon concentration of 0.17 wt.%, a peritectic
point exists where three phases are in equilibrium: a solid solution of iron
containing 0.09 wt.% carbon, a solid solution of / i r o n and 0.17 wt.% carbon and a
liquid solution of iron and 0.53 wt.% carbon.

(a) (c)

Figure 8.23. Iron - cementite (Fe C) diagram: (a) metastable diagram used for steel and
3

white cast iron; (b) detail of the peritectic equilibrium zone; (c) detail of the eutectoid
equilibrium zone. The concentration is given in wt.%.

The <-> transformation temperature is lowered from 912 to 727 C by an

increase in carbon (figure 8.23. (c)). At this temperature, a eutectoid transformation
occurs . This is the same as a eutectic transformation apart from the fact that the
transformation starts from a solid solution rather than a liquid solution. Three
phases exist in equilibrium at the eutectoid temperature,: a / s o l i d (austenite) solution
containing 0.08 % carbon, an a solid solution with 0.02 % C (ferrite) and cementite
(Fe C) with 6.7 % C.
3

With carbon contents greater than 1.98 wt.%, (figure 8.23. (a)), another eutectic
transformation is observed, with a liquid solution containing 4.3 % C, a solid solution
of / iron (1.98 % C-austenite), and of cementite (Fe C) which contains 6.7 % C.
3
 Alloys and phase diagrams 199

Alloys of eutectic composition containing cementite (Fe C) are white cast irons. If
3

carbon precipitates during eutectic solidification as graphite grey cast iron is formed.
In the following chapter, the iron-carbon diagram will be studied in more detail
and microstructures resulting from heat- treatments will be examined.

8.5.2. Ceramic alloys: silica-alumina

The study of phase diagrams also enables the phase transformation phe
nomena encountered during the production of ceramics to be understood and pre
dicted. Their production often involves chemical reactions occurring mostly in the
solid state, as in solid phase sintering and equilibrium is rarely achieved.
The constituents of ceramic alloys and phase diagrams are often chemical com
pounds of various metallic elements a n d oxygen. These diagrams shows phase
equilibriums similar to metal alloys. The A l 0 3 - C r 0 3 or NiO-MgO systems have
2 2

phase diagrams similar to totally miscible metallic systems such as copper-nickel.

Figure 8.24. shows the phase diagram for the silica ( S i 0 ) and alumina (A1 0 ) sys
2 2 3

tem. Each phase is designated by a mineral nomenclature (mullite, corundum), which

defines the various crystalline structures, in the same way that silica is often charac
terised by its allotropic forms (quartz, cristobalite, etc.) By fusing 70 wt.% silica with
30 wt.% alumina, a refractory brick is obtained, which after cooling has a micro-
structure of large mullite crystals agglomerated with a eutectic of small crystals of
mullite and cristobalite. According to the phase diagram shown in figure 8.24., the
refractory brick, which only contains two chemical compounds ( S i 0 and A 1 0 ) , 2 2 3

remains solid up to the eutectic temperature of 1595 C.

Figure 8.24. Phase diagram for the alumina (Al 0 )-silica (Si0 ) ceramic system. The
2 3 2

solubility of A1 0 in Si0 and of S i 0 in A1 0 is very low (< width of the line). The
2 3 2 2 2 3

diagram is shown with two concentrations C: in moles and by weight.

200 Introduction to Materials Science

8.5.3. Miscibility in polymer blends - Copolymers

By alloying metals and ceramics, materials can be obtained which have ex
tremely varied physical properties and technical characteristics. For polymers, the
situation is not the same. It is difficult to obtain polymer alloys by mixing several
molten macromolecules of differing chemical structure because, in general, they do
not form a homogeneous blend either in the liquid or in the solid state.
For two constituents A and to be able to form a homogeneous mixture, the
mixing free enthalpy AG must be negative and there should be no intermediate
m

maximum in the free enthalpy-concentration curve as in the case of the alloy given in
figure 8.14. The mixing free enthalpy AG is given by:
m

AG = AH - TAS
m m r (8.7.)

For mixtures of low molecular weight organic molecules, the mixing enthalpy
AH is generally positive (endothermic mixing). This unfavourable effect on the
m

mixing process is entirely compensated by a significant increase of the configuration

entropy which is the result of considerable increase in disorder during the mixing of
A and B. Consequently :

TAS > AH > 0 and AG < 0

m m m (8.8.)

which leads to the miscibility of A and B.

The situation is somewhat different for macromolecules. The intermolecular
interaction are not dependent on the molecular weights and for a given volume, the
enthalpy of mixing AH , which is practically independent of molecular weight, gen
m

erally remains positive. On the other hand, the number of molecular configurations is
reduced and the mixing entropy AS decreases significantly as the molecular weight
m

increases. It can be shown that the mixing entropy AS becomes very small when the
m

chain is long. As AH is usually greater than zero, AG is in general positive and two
m m

macromolecules with different chemical structures cannot be mixed.

Macromolecules with different chemical structures can only be mixed if
specific interactions occur between them such as hydrogen bonds (figure 8.25.) or
complexes resulting from the transfer of electrons between electron donating and
electron accepting macromolecules (figure 8.26.). In this case only, AG can become m

positive leading to homogeneous polymer blend

The synthesis of polymer materials containing several types of structural units is
usually carried out by means of copolymerisation reaction. A copolymer is a macro-

HC

C C

Poly(propeno'fc acid) Polyoxyethylene

Figure 8.25. Specific interactions between the structural units of two macromolecules: by
hydrogen bonding.
 Alloys and phase diagrams 201

CH 3

Polystyrene

Poly(2,6-dimethyl-1,4-phenylene-oxyde)

Figure 8.26. Specific interactions between the structural units of two macromolecules: by
electron donor and electron acceptor complexes.

molecule containing building blocks with different chemical structures. For example,
a copolymer may be produced containing a definite number of structural units of
polystyrene and of polybutadiene, but it is impossible to produce an alloy of polysty
rene and butadiene by physical mixing.
As was shown in chapter 5, a reactive molecule (monomer) is used to synthesise
a polymer. For example, polystyrene, is obtained from styrene by a polymerisation
reaction (figure 8.27. (a)). Polystyrene is an amorphous solid at ambient temperature
while its precursor while styrene is a liquid.
A mixture of monomers is used as the reaction medium to synthesise a copoly
mer. These monomers mixtures (small molecules) form in general homogeneous
liquid solutions at the synthesis temperature (between 50 and 150 C). If styrene and
acrylonitrile are mixed in defined proportions in the presence of a specific catalyst
the monomer molecule add to each other and a copolymer of styrene and acrylonitrile
(SAN) is obtained by the copolymerisation reaction summarised in figure 8.27. (b).
Copolymerisation of monomers A and B, gives a copolymer represented by the
general formula A B (figure 8.27. (c)), which gives the overall composition of the
m n

copolymer. The way in which the chains of the structural units A and are organised
in the copolymer strongly depends on the reaction catalyst.
Indeed, various types of copolymers can be distinguished, such as random co
polymers where the distribution of the structural units A and occurs randomly
(figure 8.28. (a)). Copolymers are also synthesised in which the distribution of the
structural units is organised in regular sequences such as block copolymers (figure
8.28. (b)) and graft copolymers (figure 8.28. (c)).
In graft copolymers, the main chain is exclusively made of structural units A.
The grafted chains only contain the units B. Block copolymers have several se
quences of homogeneous structure so that it is possible to distinguish di-block co
polymers (figure 8.28. (c)) containing two homogeneous sequences and tri-block
copolymers containing three homogeneous sequences.
In the liquid and solid states, random copolymers form a single amorphous
phase. In graft and block copolymers phase separation is on a microscopic scale and
202 Introduction to Materials Science

the sequences of homogeneous structure are regrouped altogether in finely dispersed,

separate phases, forming an emulsion-like microstructure.
Alloying polymers is therefore much more difficult than alloying metals or ce
ramics. Due to the incompatible nature of the chemical constituents, miscible blends
are the exception. However by copolymerisation, structural units of different chemi
cal nature are forced to cohabit in the same polymer chain. This gives rise to an enor
mous variety of polymer materials.
initiator
(a) m(CH =CH)
2 -FCH CH 2

(b) m{ CH =C9
2
+ n(CH =CH) initiator
2 CH CH- 2 CH -Crff
9

CN CN

L SX
styrene + acrylonitrile copolymer SAN

(c) mA + initiator

Figure 8.27. Synthesis of copolymers: (a) homopolymerisation of styrene to form poly

styrene; (b) copolymerization of a liquid mixture of two monomers, A (styrene) and
(acrylonitrile) giving a copolymer A B (SAN); (c) generalisation of the copolymeriza
m n

tion reaction. The molecular structure of the copolymer depends on the nature of the
catalyst.

(a) Statistical copolymer

AA-B-BA-A A- B-B-BB-A-AA

(b) Block copolymers

BiblOCk -

Block A Block

Triblock ^/vvx/vvrv/^-
Bloc A Bloc Bloc A

(c) Graft copolymer

Chain A
v/x/x/xyy\/>j^/\/\y\^
CQ PQ PQ PQ

Figure 8.28. Molecular structure of the principal copolymers: (a) random copolymers:
the structural units A and are distributed randomly. In the graft copolymers (b) and
block copolymers (c), the constituent units A and form homogeneous blocks.
 Alloys and phase diagrams 203

8.6. Summary and conclusions

The production of alloys by mixing several atomic species or chemical com
pounds is of great technological importance as various materials with specific
properties can be obtained. In particular, mixing of metals and ceramics constituents
gives :
solid solutions;
multiphase alloys in which the phases are separated by interfaces;
intermediate phases or compounds having well-defined stoichiometric
compositions.
In reality, phase equilibrium rarely occurs. Nevertheless the study of phase
diagrams provides a sound basis for understanding phase transitions and developing
new materials. For a system in equilibrium at constant temperature and pressure, the
free partial molar enthalpies (chemical potentials) of the various phases are equal.
The diagrams show the equilibrium states for a range of temperature and composi
tion. Pressure has an insignificant influence in condensed phase equilibria. For that
reason, is only given when it takes very large values. The interpretation of these
diagrams provides important information about the system. It is possible to deter
mine:
the composition and quantity of phases in equilibrium;
the solubility limits of the phases;
the transition points or transformation temperatures at phase equilib
rium for different compositions.
In reality, only a small number of basic phase diagrams exist: the cigar-shaped
diagram, the miscibility gap, the eutectic and the peritectic. More complex diagrams
are only a superposition of these basic diagrams.
Organic polymers are particular materials since they do not generally form ho
mogeneous mixtures, neither in the molten state nor in solid form. In general, it is
difficult to speak of polymer alloys in the strict sense of the word.
Usually, the synthesis of polymer materials containing several types of structural
units is carried out by means of a copolymerisation reaction. The copolymers do
occur either in a homogeneous form with a single phase (random copolymers) or in a
heterogeneous form on a microscopic scale (graft and block copolymers) giving rise
to an enormous variety of polymer materials.

8.7. Illustrative example: how can dental caries be treated?

The application of phase diagrams will be carried out in the specific case of
metal alloys containing mercury (called amalgams) and used in dentistry.
In the treatment of tooth decay, the dentist eliminates the damaged tissue with a
drill and prepares a cavity of appropriate size and shape that he fills with a substitute
material. It is vital that this material is biocompatible, resistant to saliva, which con
stitutes an aggressive and slightly acid environment (pH: 5.2 to 6.7), and is capable
of resisting forces of the order of 570 (-57 kgf) for the molars and 150 (-15 kgf)
for the incisors. As the contact areas between the teeth are small, the local stresses
are significant. The substitute material must be hard, tenacious and easy to form with
204 Introduction to Materials Science

precision and without heat. It should have good adhesion to the healthy part of the
tooth to ensure that the filling is sealed to prevent the ingress of food.
No filling material meets all the stringent requirements summarised in the pre
ceding paragraph and there is still much progress to be made in order to perfect den
tal materials. At present, three main types of material are used to fill dental caries.
The oldest and still the very currently used is an alloy of mercury with various met
als, particularly silver and tin (amalgam). Resin composites are also used, which
contain polymer binders with dispersions of inert powder fillers such as silica or
quartz. The polymers used for such applications are thermosets ( 5.2.3), cross-linked
at ambient temperature by means of various catalytic systems causing the resin to
harden and provide mechanical strength. The mixture is introduced into the cavity in
the form of a paste. Cross-linking is usually accompanied by significant shrinkage of
the filling mixture, which occurs inside the cavity being filled and leading to the
formation of cracks, which could allow the ingress of food. These polymer materials
are aesthetically more pleasing than the amalgams and are in particular used for re
pairing front teeth. However, these filling materials are noticeably weaker than the
amalgams.
Another category of products recently developed for reparative dentistry is the
glass-ionomer cements. The dispersed phase consists of fine particles (20 to 50 )
of a glass synthesised by high temperature fusion of a mixture of alumina (A1 0 ), 2 3

silica ( S i 0 ) and calcium fluoride (CaF ). After grinding, pigments and a polyacid
2 2

(binder), such as poly(propeno'ic acid) (poly(acrylic acid)) ( 18.3.1.) are added. A

paste is formed with water, which then hardens by reaction between the fine particles
of glass and the polyacid. These glass-ionomer cements also have a number of defi
ciencies, notably excessive hardening time and poor resistance to acids present in the
mouth.
An important category of dental filling materials remains the amalgams which,
nevertheless, require anchoring points to be drilled in the dental cavity due of their
lack of adhesion. The processes taking place during the application of these repair
materials will now be examined in more detail starting from the relevant phase dia
grams
Mercury is a liquid metal under ambient conditions. The amalgams used in
dentistry contain several constituents. For simplicity only the binary Ag - Hg system
will be studied, which is sufficient to illustrate the most important phenomena. The
phase diagram (figure 8.29.) is complex and is formed from the superposition of two
peritectics and a eutectic (this latter, at a high concentration of Hg, is not shown on
the diagram as it happen at concentrations smaller than the thickness of the line). It is
not necessary to analyse this complicated diagram in detail. At the temperature used
in practice, the partial diagram given in figure 8.30. is sufficient to demonstrate the
physico-chemical phenomena occurring during filling and hardening.
When a fine silver powder is mixed with liquid mercury (-65 wt.%) at ambient
temperature, a paste is obtained, similar to the sugar paste used for icing cakes. This
mixture can easily be prepared. The mercury atoms penetrate the solid silver by dif
fusion and the silver atoms dissolve in the liquid mercury. This change in compo
sition leads to a lowering of the melting point of the silver. It gradually forms a stable
intermetallic compound j \ of composition A g H g with a melting point at 127 C,
3 4)

considerably higher than that of the mouth (37 C). It is the formation of this inter-
 Alloys and phase diagrams 205

metallic compound by isothermal diffusion, which hardens the alloy. As the diffusion
of liquid mercury in the solid is slow, the grains of silver must be sufficiently small
to allow the isothermal phase transformation to take place in a few hours. It is im
portant to avoid hardening while the amalgam is being put into the cavity. This
process of amalgamation is similar to liquid phase sintering which is a technique
used in the manufacture of ceramic materials and hard metals (chapters 4. and 5.).

0 20 40 60 80 100
Ag Hg
Concentration (weight %)

Figure 8.29. Phase diagram for the binary Ag - Hg system. The phase is the compound
AgH and the phase the compound Ag Hg . Mouth temperature (37 C) is indicated by
3 4

the dotted lines.

Initial composition Final composition Initial composition

200 r
U



\

1 - 1
0 20 40 60 80 100
Ag Hg
e Concentration (weight %)

Figure 8.30. Partial view of the phase diagram for the Ag - Hg system (figure 8.29.) be
tween 0 and 200 C.

The amalgam used by dentists is more complex than the example described here.
Previously, dentists used metal powders, mainly composed of a silver tin alloy
(composition Ag Sn), which forms two intermetallic compounds, 71 and 72 when
3

mixed with mercury:

206 Introduction to Materials Science

2 Ag Sn + 8 Hg - 2 Ag Hg (7) + 2 Sn
3 3 4

(8.9.)
Hg + 7-8 Sn -> Sn . Hg ( )
7 8 2

The 72 phase is undesirable because it corrodes easily leading to blackening of

the treated tooth. To overcome this, copper is added to the metallic powder in a dis
persed form as an Ag-Cu eutectic eliminating the y phase by the reaction:
2

+ AgCu -> / + Cu Sn
2 6 5 (8.10.)

More corrosion resistant amalgams are available which avoid the formation of
72. These alloys contain at least five constituents (Ag: 63 - 70 wt.%; Sn: 26 - 28
wt.%; Cu: 2 -5 wt.%; Zn: 0.2 wt.%). These alloys also show a shrinkage close to
zero.
The mercury used in dental alloys is toxic. Periodically rumours arise
concerning the danger of mercury in amalgams. In principle, the formation of the 7
intermetallic renders the mercury harmless. The formation of a chemical compound
between the constituents explains this. In March 1997, W.H.O. (the World Health
Organisation) also concluded that mercury in dental alloys is harmless. Several re
searchers of the C.N.R.S, who have calculated that the total amount of dental amal
gam in French mouths has reached 240 tonnes, have recently called the W.H.O. study
into question. From the equilibrium diagram, it is essential to avoid any free mercury
remaining after the reaction, which forms the intermetallic compound 7 (Ag Hg ). It
3 4

is therefore recommended to use a smaller amount of mercury than that required by

the stoichiometry of A g H g . Under these conditions, a mixture of -AgHg and y
3 4 r

A g H g phases would be formed.

3 4

8.8. Exercises
8.8.1. How does the solidification of a solid solution differ from that of a pure con
stituent?
8.8.2. Calculate the concentration (in mol.%) of the constituents of an alloy, used as
solder, made from 100 g of tin (Sn) and 65 g of lead, knowing their molar weight
(Table 1.1.).
8.8.3. What is the concentration (wt.%) of copper in the eutectic alloy of the binary
Ag-Cu system (figure 8.19.)? How does the composition of the liquid vary during the
solidification of an alloy containing 80 wt.% copper?
8.8.4. At 200 C, an alloy of lead-tin, used for solder and containing 50 wt.% Sn,
consists of two phases: a solid phase a rich in lead and a liquid phase / rich in tin.
What is the degree of freedom (variance) of this alloy at constant pressure (1 arm)?
8.8.5. The maximum solubility of tin in copper in bronze (Cu-Sn alloy) is 15.8 wt.%
at 586 C. What is the atomic fraction of tin atoms at the solubility limit?
8.8.6. A Ag-Cu alloy (25 wt.% copper) is cooled from the liquid state (figure 8.19.).
Determine the temperature at which the first solid phase forms and the type and
composition of this phase. At what temperature is this alloy completely solidified at
 Alloys and phase diagrams 207

equilibrium? What phases can be observed at a slightly lower temperature than that
of complete solidification?
8.8.7. From the equilibrium diagram of the binary Ni-Cu system (figure 8.15.), de
termine:
the melting point of nickel;
the solubility of Cu in the solid Ni phase at 1000 C.
Given a Ni -Cu alloy with 40 mol.% Cu, find :
the liquidus and solidus temperatures;
the mole percentages of the phases present at 1300 C (figure 8.15.).
8.8.8. An Al-Cu alloy containing 4 wt.% copper (see figure 8.31.) is cooled slowly
from the melt to ambient temperature (T ~ 20 C). Describe the solidification stages
and the solid-state precipitation. Which phases are formed?

0 20 40
Weight %Cu
Figure 8.31. Partial view of the Al-Cu equilibrium diagram.

8.8.9. The equilibrium phase diagram for Cu-Zn alloy (brass) is shown in figure 8.32.
Determine the maximum solubility of zinc in copper and the corresponding tem
perature. Which alloys are composed exclusively of a phase at ambient temperature,
given that diffusion is too slow to allow equilibrium to be reached below 200 C

1200
11084 C

0
0 20 40 60 80 100
Cu Zn
Weight %

Figure 8.32. Cu-Zn equilibrium diagram.

208 Introduction to Materials Science

8.8.10. Using the S1O2-AI2O3 equilibrium diagram (figure 8.24.), find the alloy with
the lowest melting point. Indicate the number of phases in equilibrium at the eutectic
point as well as the concentration in wt.% of silica ( S i 0 ) at this point. What is the
2

solubility of A l 0 i n solid S i 0 ? Starting from the liquid state, describe the stages in
2 3 2

the formation of the microstructure of the alloy containing 80 wt.% alumina

8.9. References and complementary reading

D.R. GASKELL, Introduction to Metallurgical Thermodynamics, 2 ed., Mc Graw-Hill, New York, 1981.
nd

J. P. MERCIER et . MARECHAL, Chimie des polymeres, (Traite des Materiaux, vol. 13), Presses poly
techniques et universitaires romandes, Lausanne, 1996.
J. W. NICHOLSON, The Chemistry of Polymers, 2 Ed., RSC Paperback, The Royal Society of Chemis
nd

try,Letchworth (U.K.), 1994.

J. PHILIBERT, A. VIGNES, Y. BRECHET et P. COMBRADE, Metallurgie: du mineral au materiau,
Masson, Paris, 1998.
N. A. PLATE, 1. D. LITMANOVICH, . V. NO Ah, Macromolecular Reactions, Wiley, Chichester
(U.K.), 1995.
A. PRINCE, Alloy Phase Equilibria, Elsevier, Amsterdam, 1966.
. B. SMITH, Basic Chemicalthermodynamic-,4 'ed., Clarendon, Oxford, 1990.
th

R. B. SEYMOUR, Ch. E. CARRAHER, Jr., Polymer Chemistry, An Introduction, 3 ed., Dekker, New
th

York, 1992.
J.H. WESTBROOK (ed.), Intermetallic Compounds, Krieger, Huntington, New York, 1977.
 C h a p t e r 9.

Phase Transformation

9.1. Objectives
To describe the mechanisms responsible for atomic and molecular
movements in condensed phases.
To relate solid-state atomic mobility to transport phenomena in
materials.
To identify the main mechanisms, which control phase transformations
and to present some elements of solidification theory.
To determine the heat-treatment conditions required to obtain a given
microstructure using Transformation - Temperature - Time (TTT) dia
grams.
The phase diagrams described in chapter 8 give the number, composition and
relative quantity of phases as a function of temperature under equilibrium conditions.
Phase diagrams do not provide information on phase transformation kinetics, i.e.
the time required for the formation of new phases when the temperature or pressure
acting on a system is changed.
All phase transformations, such as crystallisation, involve the transfer of atoms
or molecules from one phase into others. Consequently, the rate of microstructure
formation is controlled by the mobility of atoms and molecules in the liquid and
solid states.

9.2. Diffusion
9.2.1. Mechanism of atomic movement
Diffusion phenomena determine the kinetics of most phase transformations.
Diffusion in the solid state will be specifically studied in this chapter. This phe
nomenon is observed in practically all solids, not only in solutions but also in pure
crystalline and amorphous solids (auto-diffusion). Diffusion is a thermally activated
transfer of atoms or molecules.
Two main types of solid solutions exist in crystalline materials, interstitial and
substitutional solutions. Interstitial solutions involve smaller atoms (C, , N),
which occupy empty spaces between atoms in the crystal ( 4.2.2.). These small
atoms are able to move relatively easily along channels formed by the adjacent vacant
interstitial sites. In substitutional solid solutions, dissolved atoms or molecules oc
cupy a site or, in the case of bulky molecules (polymers), several sites of a crystal
line or quasi-crystalline lattice.

2 0 0 2 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

210 Introduction to Materials Science

The movement of a dissolved species might be associated with the displacement

of a solvent atom or molecule. However, such an exchange process between solute
and solvent is exceptional. A much more probable migration event is the exchange of
sites between an atom (or molecule segment in case of a polymer) and a neigh
bouring vacancy (figure 7.2.). The number of vacancies determines therefore the rate
of diffusion. As there are fewer vacancies in a crystal than in a liquid, the diffusion
processes involving vacancies in crystalline or quasi-crystalline structures are much
slower than for a liquid. This is also valid for auto-diffusion.
An atom or a molecule segment close to a vacancy can move only if it has suffi
cient thermal energy to separate slightly the adjacent atoms or molecule segments
from their equilibrium positions. Therefore, the separation of adjacent atoms in
volves necessarily the passage over a potential energy barrier (figure 9.1). Diffusion
is a process which is easier at high temperature and when the binding energy is low.
It is possible to quantify diffusion in substitutional solutions by the jump frequency
v, which gives the number of jumps of an atom or of a molecule segment from one
equilibrium position in the lattice to another.
An atom or a molecule segment can undergo translational diffusion motion rela
tive to its neighbours only if:
a point defect is adjacent to the atom or the molecule segment;
sufficient thermal energy is available to overcome the potential barrier
between the occupied site and the vacancy.

Therefore is the product of two probabilities:

v = cst.Xi.p E (9.1.)

The probability of finding a vacancy adjacent to the atom or the molecule is

equal to the molar fraction of vacancy . p is the probability that the thermal
E

energy is sufficient to allow the displacement to occur. The value of Xj was calcu
lated in chapter 7.:
= nIN = exp(- AG//RT) = Qxp(-AH /RT + AS /R) = A exp(-AH /RT)
f f f (7.2.)

AG/, AH/dina AS/ are respectively the free enthalpy, the enthalpy and entropy of for
mation of the vacancies. For solids, AH/ is close to 100 kJmol" . is the absolute
1

temperature, given in degrees Kelvin (K)

In a condensed phase (solid or liquid), the number of atoms or small molecules
having sufficient energy to overcome a potential barrier AG is given by an exponen
t

tial distribution (Maxwell-Boltzmann):

p = Bexp
E L (9.2.)
\ RT J

Since AG, = AH - TAS, t

A (-AHA (ASA

The second exponential does not involve temperature and may be included in the
constant:
 Phase Transformation 211

(a)

Distance
(b)

Figure 9.1. Representation of the random and reversible movement of an atom in the
neighbourhood of a defect in a crystalline structure (a). Variation of the potential
energy during the movement of the atom (b).

(9.3.)
V j

and C are constants and AH which represents the activation enthalpy for
h

migration of the point defect, is also of the order of 100 kJ mol" . 1

The frequency is only a function of temperature and can be written:

(AH t +
v(7) = v e x p -
0 (9.4.)
RT J

To a first approximation, v is independent of temperature.

0

In a substitutional solid-solution, the variation of the jump frequency (diffusion

rate) with temperature results from the superposition of two phenomena:
the variation of the number of vacancies depending on AH/;
their mobility given by AH . t

For interstitial diffusion, the diffusion sites are already present at the interstices
between the spherical atoms and only the potential barrier given by AH must be t

taken into account.

9.2.2. Diffusion coefficient - Fick's Laws

When there is a concentration gradient in a phase, the mobility of the atoms or
molecules produces a material flux in the direction of the gradient which tends to
reduce the free enthalpy (energy) of the system and to eliminate the concentration
gradient. A simple relationship exists between the flux J, i.e. the number of atoms
or molecules crossing a unit area perpendicular to the axis per unit time and the
concentration gradient (dc/dx) along the axis:

(9.5.)
\dx)
212 Introduction to Materials Science

This equation (Fick's first law) states that the flux [ m ~ V ] is proportional to
l

the concentration gradient dc/dx [mf ]. The concentration is measured in atoms or

4

molecules per metre [m~ ]. The proportionality factor is the diffusion coefficient D
3

[ m V ] . The negative sign has been introduced since the flux and the concentration
1

gradient are in opposite directions (the flux normally occurs towards the lowest con
centration). The diffusion coefficient is a measure of atomic mobility. It is directly
proportional to the jump frequency v(T) and by analogy with equation (9.4), can be
written as:

f Q 1
(9.6.)
D = A>exp -

Q is the activation enthalpy (energy) of diffusion and the absolute temperature in

degrees Kelvin (K). In substitutional diffusion by a point defect mechanism, Q is
equal to the sum of AH and AH . For an interstitial mechanism, only AH must be
( f t

considered.
The variation of D with temperature for solid phases is shown in figure 9.2. In
most cases, the increase in D with temperature is the result of a rise in thermal
energy and the vacancy concentration. The higher rate of diffusion at grain boundaries
and on surfaces is due to an increase in the free volume in the disturbed regions of
the crystal. The slope of the line for the logarithmic variation of D as a function of
l/T enables the activation enthalpy Q to be obtained, while the intercept with the log
D axis gives D (9.6.).
0

The interstitial diffusion of carbon atoms in iron plays an important role in mi

crostructure formation during the heat-treatment of iron - carbon alloys (steels and
cast-irons). At about 1000 C, the jump frequency v(7) of the carbon atoms in a soft
steel (0.1 % C) is of the order of 10 s~ , while the vibration frequency of these atoms
9 1

in the crystal is ~ 1 0 s " \ Therefore, on average, only one oscillation in 10000

13

provokes a jump. As each jump leads to a displacement, which is approximately

equal to an interatomic distance (r = ~0.3 nm), the displacement rate of the atoms is
0

of the order of 0.3 m s (~1 km IT ). These displacements occur randomly as with

-1 1

Brownian motion and the effective displacement between two atoms is negligible.
For organic glasses, a value of v(7) close to 1 s" at the glass transition tem
1

perature T was calculated. In polymers, diffusion movements involve chain seg

g

ments of about the size of a building block and not the overall macromolecule. It
would be highly improbable to find in condensed phases defects of the size of a
macromolecule, which in general contains between 500 and 10 000 building blocks.
In chapter 7. it was shown that the concentration of point defects increases
abruptly at the melting temperature. Diffusion in liquids is much more rapid than in
solids. The high viscosity of molten polymers results from the fact that macromole
cules move segment by segment, by the de Gennes' reptation mechanism described
in figure 12.25.
Fick's first law (9.5.) allows the diffusion rate of a constituent to be calculated
only when the concentration gradient remains constant with time and space. This
situation is only fulfilled in particular cases, such as the diffusion of a substance
through a thin wall separating two large reservoirs with different concentrations
(figure 9.3.).
 Phase Transformation 213

0 ^Temperature 1/7

Figure 9.2. Variation of the diffusion coefficient D on a logarithmic scale as a function

of l/T corresponding to (9.6.) for various types of diffusion in a polycrystalline mate
rial.

Figure 9.3. Steady state diffusion through a wall of thickness Ax separating two reser
voirs where the concentration of the diffusing substance is respectively equal to c\ and

When the local concentration of the diffusing substance varies over time, the
concentration gradient and the material flux are modified. For unidirectional diffu
sion through unit area perpendicular to the axis, the variation of the flux dJ/dx is
given by :

^ = (- *.)
D (9.7.)
dx dx V dx y

The increase over time of the concentration c in a small volume element of thickness
dx is equal to the flux reduction through the same volume element:

= _ L (9.8.)
dt dx
Assuming that D is independent of concentration, the unidirectional diffusion is
given by :
214 Introduction to Materials Science

Distance *

Figure 9.4. Distribution of the concentration c in relation to the distance as a function

of time. The time intervals between t - h and t\ - h are equal. The concentration
0

variation over time is greater as the second derivative d c/dx is higher. At infinite
2 2

time, the first derivative dc/dx and the second derivative d c/dx are both equal to zero,
2 2

which corresponds to thermodynamic equilibrium.

A ( \
(9.9.)
de _ d c
dt [dx 2
t

Equation (9.9.) is Fick's Second Law. The concentration variation as a function of

time is proportional to the second derivative of the concentration with respect to the
distance x. The concentration variation at a point is more rapid the greater the second
derivative of (concentration field curvature) (figure 9.4.).
At equilibrium:
dc
= 0 (9.10.)
ax
and

(9.11.)
dx \dx j

When equilibrium is achieved, the first and second derivatives of the concentration
with respect to the distance are both equal to zero.
Equation (9.9.) has the same form as the equation for heat diffusion (conduc
tion) (Fourier's Law). It can be integrated as a function of the initial conditions and
those at the limits of the system.

9.3. Phase transformation

9.3.1. Generality
The study of phase equilibria (chapter 8.) fixes the compositions of phases at
equilibrium, but gives no indication of the kinetics of phase transformations or the
microstructures formed (in particular the phase size and morphology).
 Phase Transformation 215

For a phase transformation to occur, the free enthalpy (energy) of the system
must decrease during the transformation (AG < 0). This condition is, however, far
from being sufficient since the transformation from one phase into another takes
place mainly by nucleation and growth. At the onset of the transformation, nuclei of
the new phase are formed which are small volume elements having the structure of
the phase being formed and separated from the original, non-transformed phase by
high-energy interfaces (figure 8.4.). These nuclei have a high surface to volume ratio
and initially possess a free energy higher than that of the transforming phase.
The phase transformation with the most rapid kinetics occurs preferentially and
the microstructure, which is formed, is rarely that with the lowest free enthalpy. This
is so for crystallisation, which is the most important physical transformation in
materials. This leads to a great variety of microstructures, particularly in alloys where
several solid phases with different structures can form.
Solid-state phase transformations are also very important in materials. Al
lotropic transformations occur in ferrous alloys (transformation of fee y-iron into the
bcc -phase on cooling). Eutectoid transformation and precipitation of new phases
are other examples of solid-state transformations.
An eutectoid transformation is similar to an eutectic transformation. It occurs
when a solid phase transforms into two other solid phases of lower and higher con
centrations with respect to the transforming phase. For example, the solid solution of
fee 7-iron containing 0.8 wt.% carbon, at a temperature below 727 C forms pearlite,
a two-phase microstructure consisting of alternate lamellae of a-iron and cementite
(Fe C).
3

Formation of precipitates of intermetallic compounds may also occur in the

grains of an alloy. These play an important role in the hardening of metals such as
aluminium.

9.3.2. Nucleation of a new phase

The nucleation of a new phase is the result of atomic or molecular fluctuations
occurring in the transforming phase. Nanoscopic particles having the structure of the
phase being formed appear spontaneously at random through thermal agitation. Most
nuclei having a high specific surface are unstable and disappear. Only some of them
reach a size sufficiently large for them to develop. The initiation of a new phase is
assisted by the presence of microscopic foreign particles (impurities), which are al
ways present in more or less significant quantities in the transforming phase. At the
surface of such foreign crystalline particles, the free enthalpy of formation of a crys
talline nucleus is reduced. As soon as the thermodynamic conditions are favourable,
phase transformation initiates on the surface of these foreign particles. In figure 9.5,
the crystallisation of a molten metal and of a liquid organic polymer is described as
examples of phase transformations. In both cases, crystallisation starts on micro
scopic heterogeneous particles when a certain degree of undercooling is reached.
The seeds grow progressively to form grains. When the thermodynamic condi
tions remain favourable, the transformation occurs throughout the available volume.
Dendrite formation is normally observed in metals ((a), U). In pure metals, these are
no longer visible at the end of solidification ((a), / ) . Each grain is a monocrystal.
5

Their final size is a function of the number of heterogeneous nuclei. In solidified

alloys, the dendrites can be observed after polishing and surface etching with an acid
216 Introduction to Materials Science

Figure 9.5. Crystallisation of molten metal (a) and a liquid organic polymer (b) on het
erogeneous nuclei at time (/i). For metals, dendrite formation ((a) U) is generally ob
served but, for pure metals, these are no longer visible at the completion of solidifica
tion ((a) / ). In polymers, spherulite formation is observed. Polymer spherulites are
5

semicrystalline grains (figure 10.4.). However the size of the grains formed by poly
mer crystallisation is, in general, substantially smaller due to a high density of nuclei.
On completion of solidification ((b) t ), the spherulites impinge.
5

(figure 10.6.). In organic polymers, the spherulites grow and touch each other at the
completion of the transformation. Spherulites (chapter 10.) have a complex, partially
crystalline structure. The size of the grains formed on polymer crystallisation is, in
general, significantly smaller because of a very larger number of seeds.
Phase transformations can also proceed in a homogeneous manner. Thus, the
random formation of transformed nanoparticles may form in the molten material
starting from a homogeneous grouping of atoms without any trace of foreign sub
stance. This mechanism is called homogeneous nucleation. The formation of nuclei
within a homogeneous phase is energetically less favourable and homogeneous nu
cleation proceeds at a higher undercooling than heterogeneous nucleation. This is
why the majority of phase transformations, which occur via a nucleation-growth
mechanism, involve a heterogeneous nucleation process. For homogeneous nuclea
tion to occur, there must be a complete elimination of all microscopic foreign parti
cles (impurities) in the system. This is difficult to achieve.
For reasons of simplicity, only the homogeneous nucleation mechanism will be
described. The appearance within the liquid phase of a nucleus of volume V and ex
ternal area S' is accompanied by a change of free enthalpy AG in which two compo
nents are identified:
 Phase Transformation 217

AG n =AG V + AG S
(9.12.)
The first term on the right hand side of (9.12.) AG corresponds to the crystalli V

sation of the inner part of the crystal particle and is therefore proportional to the its
volume V. AG is the term resulting from the formation of an interface between the
S

liquid phase and the crystal embryo. This term AG varies as a function of the area S' S

of the nucleus interface. Thus:

AG V =VAg v and AG* = S'y (9.13.)

Ag represents the free enthalpy of formation per unit volume of the new phase and ,
v

the specific interfacial energy, i.e. the variation in free enthalpy per unit of interface
formed at constant and ( 8.2.4.). AG is always positive since the formation of
S

an interface is an endothermic process.

As shown on figure 9.6., when the formed particles are small, the free enthalpy
of formation is positive because of a high surface to volume ratio. The absolute value
of AGs is therefore larger than that of AG . When the particle exceeds the critical size
V

r* at which the contributions AG and AG become equal in absolute value, the trans
V S

formation takes place spontaneously with a free enthalpy decrease. Particles of radius
smaller than r* are unstable and spontaneously disappear. Such particles are called
embryos. Particles larger than r* became more stable as their radius increases and
such particles are called nuclei.
From thermodynamics ( 6.3.1.), we obtain:

Ag =
v Ah -TAsv
v
(9.14.)
Ah and As are the enthalpy and the entropy of formation per unit volume of the
v v

new phase, respectively. In crystallisation, both parameters are negative. At the ther
modynamic melting point T , the free enthalpy of formation (AG or Ag ) is zero and
m V v

one has:

(9.15.)
m

/
/
/

Spherical nucleus radius r

Figure 9.6. Variation of the free enthalpy of formation of a spherical nucleus AG as a

function of the radius r at a given supercooling AT(T< T ). m
218 Introduction to Materials Science

The volume term AG , or Ag , should be negative at the transformation tempera

V v

ture. Therefore, a physical transformation can occur spontaneously, only if the tem
perature of the system is lower than the thermodynamic equilibrium temperature. For
crystallisation, the system must be at a temperature < T . This corresponds to an m

undercooling (also called supercooling) AT = T - T. From figure 9.7., it may be m

observed that the volume free enthalpy AG , (or Ag ) is zero at the melting temperaV v

ture T and that it becomes more negative as the degree of undercooling in

m

creases. This negative volume free enthalpy Ag constitutes the driving force of v

phase transformations.
Close to T , Ah and As remain practically constant. Ag varies linearly with
m v v v

temperature (figure 9.7.) and the following equation may be written:

Ag = As AT = Ah
v v v Ah. (9.16.)

V m J 1

For a spherical nucleus of radius r, the volume component of the free enthalpy of
formation of the seed (9.13) is given by:
4
AG =-nr Ag
v
3
v
(9.17.)

The free enthalpy variation arising from the interface creation (9.13) is written
as:
AG = s 4nr y 2
(9.18.)

For a spherical particle, the free enthalpy of formation at temperature is there

fore equal to:
4
AG = ~ 7 c r A g + 4 7 i r 7
n
3
v
2
(9.19.)

&0

ft
\
"03
Ag <0 v

w Ag =0
v

Temperature

Figure 9.7. Variation of volume free enthalpy with T. Close to T , it can be assumed m

that the volume free enthalpy varies linearly.

 Phase Transformation 219

From (9.16.):

AG = - nr Ah
n
3
v + 4nr y2
(9.20.)

The equation (9.19.) shows that for small values of r, the term for the surface
free energy is dominant and that AG is positive. On the other hand, if r is large, it
is the volume free enthalpy term that is predominant and AG becomes negative. A n

critical radius r* (figure 9.6.) must be reached in order that the free enthalpy of the
nucleus formation begins to decrease with r and that a stable phase forms. By
making the first derivative of (9.20.) equal to zero, the maximum value of the curve
AG -r is obtained:
n

* 16 3
% _cs^
=
(9.21.)

and

r * = rm
2 T
= CSt (9.22.)
Ah AT v (AT)

From (9.21.) and (9.22.), for = T , the potential barrier AG* and the seed
m

critical radius r* become infinite. Consequently, the transformation cannot occur at

T= T . For the transformation to occur, the system must be at < T . The trans
m m

formation occurs more easily at high degree of undercooling AT = T - , since as m

AT increases, the critical radius r* get smaller and the thermodynamic barrier is
lowered.
However, in this calculation, no account has been taken of the fact that the
nuclei do not form in the system in an isolated manner. When particles of radius r
appear in the system, the increase of entropy AS going with the random dispersion
of particles amongst the atoms or molecules of the system should be taken into
account. AS, which is the increase of configurational entropy (entropy of mixing),
is positive. The variation of free enthalpy(-TAS ) going with the random dispersion
n

of particles in the system is negative.

By taking account of this, the number of nuclei of radius r per unit volume as a
function of the temperature can be calculated from the value of the free enthalpy of
formation AG of a mole of nuclei:

=cji-exp-^. (9.23.)
RT

This expression is similar to (7.2).

By combining equation (9.21.) and (9.23.), the equilibrium concentration of
nuclei of critical radius r* is calculated as a function of T. Since AG * gets smaller
9 n

with increasing undercooling (9.21.), the number of seeds reaching the critical size
increases considerably when is decreased.
At a given temperature, a constant fraction of the critical nuclei is going to grow
and transform into grains of new phase (crystalline), and additional nuclei of critical
size will appear in the untransformed part. In this condition, a steady state concentra
tion of nuclei is achieved. Therefore, the number of nuclei remains constant at a
given temperature. This leads to the concept of the nucleation rate I that represents
220 Introduction to Materials Science

the number of nuclei, formed per unit volume and per unit time [ i r f V ] and passing
1

the free energy barrier AG*. This rate of nucleation increases considerably with the
degree of undercooling AT.
When an atom, a molecule or a segment of a macromolecule attaches to a
growing nucleus, it has also to overcome a potential barrier AG, analogous to that
involved in diffusion (9.2.). Therefore, the nucleation rate / depends both on a ther
modynamic factor and on a transport term and is given by the following expression:

(AGI+AGA
(9.24.)
I = cstexp- -

The thermodynamic term exp(-AG */RT) represents the driving force for trans
n

formation. It depends on the critical free energy of formation of the nucleus AG * n

(9.21.). The transport term exp(-AG,)/RT characterises the mobility of the atoms or
molecules in the untransformed molten phase. The thermodynamic term is zero when
the supercooling = 0 and it increases when the temperature is lowered. The
transport term (9.6), which depends on the rate of atomic or molecular movements,
diminishes exponentially when the temperature decreases, reaching zero at 0 or
before. The thermodynamic factor and the transport term therefore have an antagonis
tic effect on the nucleation rate /. This passes through a maximum for a defined tem
perature r max lower than T (figure 9.8). / is equal to zero at temperature T when the
m m

seed formation free enthalpy reaches an infinite value (figure 9.6).

Figure 9.8. Rate of nucleation / as a function of the temperature T. It is zero at T and

m

passes through a maximum between T and T= 0 K.

m

For temperatures between T and r , / increases when the temperature de

m max

creases, and the thermodynamic factor exp(- AG * /RT) predominates in this tempera
n

ture range. If < T , the opposite occurs, / decreases when diminishes. The
max

transport exp(-AG,//?7) term then controls the transformation.

The nucleation density [m~ ] determines the size of the grain (figure 9.5) which
3

is an important characteristic of the microstructure.

Sometimes, finely divided substances called inoculants or nucleating agents
are added to the molten material to increase the number of heterogeneous seeds and
control the grain size. The mechanism by which nucleating agents work is not pre
cisely understood. In some cases (chemical seeding), a chemical reaction takes place
 Phase Transformation 221

between the nucleating agent and the molten material with the formation of heteroge
neous nuclei.
Homogeneous or heterogeneous nucleation, which occurs in materials at the
onset of a transformation, are often described as primary nucleation phenomena to
distinguish it from secondary nucleation which occurs in polymers and certain ce
ramics at the growth front.

9.3.3. Crystal growth at atomic scale

After formation of stable nuclei, the growth of the new phase proceeds by addi
tion of atoms or molecules at the interface separating the parent phase from the new
phase. There is a flux of atoms or molecules from the parent phase to the new phase
as well as a flux of atoms or molecules in the opposite direction. There is only
growth if the flux of atoms or molecules is higher towards the phase being formed.
The growth rate is a function of the probability that the atoms or molecules have
been absorbed on the interface of the growing phase.
The probability of absorbing on a surface depends on the bonds formed by the
atoms or molecules when attached on to the surface. An interface, which is very
rough at an atomic scale, favours absorption while a smooth surface is unfavourable.
Growth depends therefore on the interface structure.
On a rough interface (figure 9.9. (a)) characterised by a large number of steps
and surface defects, atoms or molecules are able to attach to many sites and have a
high probability of being absorbed. Materials having a low melting entropy
( A S h / R - 1) such as metals or certain organic substances (camphor, succinonitrile,
etc.) form a rough interface on crystallising. Crystalline growth of these substances is
easy and the growth rate obeys an equation of the type:
R = K\AT
g (9.25.)
In this equation, K\ is very large and a small amount of supercooling is suf
ficient to reach high growth rates. In fact, in crystals with a rough interface, the
growth occurs close to the melting point and the rate-determining step is diffusion,
either of heat in the case of pure metals or of constituents for alloys.
Some substances, polymers in particular, characterised by very high entropy of
fusion (ASJR ~ 100) form atomically smooth interfaces. The attachment of mole
cule segments onto such interfaces is difficult. In fact, with this type of interface,
growth occurs by the successive deposition of monoatomic layers (figure 9.9. (c)).
Molecule segments are aligned in a coherent fashion with the underlying crystalline
structure. The polymer chain is perpendicular to the thickness of the crystal. The
crystal segments are deposited on the underlying layer and connected by loops.
Chain folding occurs in a more or less regular manner.
The rate-determining step (figure 9.9. (c)) is the absorption of the first crystal
line segment of the layer. This first crystalline stem constitutes the secondary seed.
The length of this initial segment, which defines the crystal thickness, increases with
temperature. Once the first segment is deposited, the growth of a mono-atomic layer
occurs directly in the two directions perpendicular to the initial segment. This
process, which repeats successively many times, leads to the crystal growth in a di
rection perpendicular to the mono-atomic layer. In this type of crystallisation, the
growth rate takes an exponential form:
222 Introduction to Materials Science

~lnm

Figure 9.9. Structures of the interface between liquid and crystal:

(a) a rough interface offering numerous sites for atoms to attach to the crystal;
(b) a smooth interface containing a permanent defect (here a screw dislocation)
allowing a relatively easy attachment;
(c) a smooth, defect-free interface requires, in the case of organic polymers, the
nucleation of a new layer each time the interface is entirely covered.
The atoms in the crystal are shown schematically as cubes. In the diagrams (a) and
(b), the arrows indicate the preferential absorption sites. In sketch (c), the arrows c'
and c" indicate respectively the direction of growth of a crystalline segment and of a
mono-atomic layer of the polymer crystal. The atoms in the liquids state are not shown.

R = K exp(-K' IAT)
G 2 2 (9.26.)

This crystallisation mechanism is very slow and demands high degrees of under
cooling ( - 5 0 K).
Intermediate cases exist where the crystal surface is smooth but contains certain
defects. Figure 9.9 (b) shows an example of this type. At the crystal surface, a screw-
dislocation emerges which stays active throughout the growth as the step turns
around the dislocation line. This mechanism is typical of the crystallisation of cer
tain ceramics. The growth rate obeys an expression of the type:
R = K (AT)
G 3
2
(9.27.)
Crystallisation via a screw-dislocation occurs at larger undercooling than that
observed for a rough surface. In general, the crystallisation of ceramics is more diffi
cult than that of metals but much easier than that of organic polymers.
In conclusion, the crystal growth is in a large part controlled by the nature of the
interface on an atomic scale. The crystallisation of organic polymers is particularly
complicated. Their degree of crystallinity always remains less than 100 %. Parts of
 Phase Transformation 223

the chains are incorporated in the folds and the chains can participate in several crys
tal lamellae at the same time. The inter-lamellar chain segments have an amorphous
structure (chapter 10.).
Like the nucleation, the growth rate can be limited by diffusion phenomena.
When the extent of undercooling is high, it is the transport term that is dominant.
The growth rate passes through a maximum at a temperature T ' m a x < T . In general, m

the maximum growth rate occurs at a temperature higher than the temperature T max

where the nucleation is most rapid, i.e.:

T >
m 'max > T max (9.28.)
The combination of nucleation rate and growth velocity determines the
transformation rate ( m V ) of the transformation.
1

9.3.4. G r o w t h of microstructures
In materials, such as metals which crystallise rapidly (figure 9.9. (a)), pheno
mena which occur on the atomic scale are not generally rate determining. They repre
sent only one part of the phase transformation process. The rest consists of pheno
mena related to the structuring of the material on a larger scale. These phenomena
produce crystal grains with various microstructures formed by crystallisation:
dendrites,
eutectic grains,
platelet structures
The dimension of the microstructure, which is visible by optical or electron
microscopy, varies usually from a fraction of up to mm. In figure 9.10., these
microstructures are illustrated at an intermediate stage of formation between nuclea
tion and the growth completion.

(a) (b) (c)

Figure 9.10. Main morphologies observed during phase transformation of metallic

crystal alloys:
(a) dendrites formed during solidification;
(b) spherulitic eutectic grains (note the formation of a grain boundary in top right
hand corner);
(c) crystal platelets occurring in a solid state phase transformation.

In this introductory treatment, it is not possible to review all the microstructures

developed during metal crystallisation. Only two important cases will be considered:
growth leading to dendrite formation and eutectic solidification producing a two-
phase lamellar or fibrous structure.
When a liquid metal is undercooled during transformation, growth usually oc
curs equally in all spatial directions (equiaxial growth). During growth, the crystal
224 Introduction to Materials Science

rejects latent heat thus increasing its temperature and a negative temperature gradient
is established at its growth front (figure 9.11. (a)). This temperature gradient allows
heat to be dissipated. If a protuberance develops, (figure 9.11. (b)) the heat evolving
from the transformation will be dissipated most easily at the apex which will grow at
a greater rate. The undercooled solid-liquid interface is therefore morphologically
unstable. As shown in figure 9.12., this produces equiaxial dendritic growth from an
initially small spherical crystal nucleus. The dendrite branches develop in well-de
fined crystallographic directions: <100) for cubic metals and<1100>for hexagonal
crystals (snow flakes for example).

Solid Liquid

(a)

(b)

Figure 9.11. Development of a protuberance at the solid-liquid interface when the

liquid temperature is below that of the solid by an amount AT: (a) temperature distribu
tion curve; (b) formation of a protuberance.

71
Figure 9.12. Development of a dendrite: (a) spherical crystal nucleus; (b) development
of instabilities at the crystal surface; (c) first stage in the formation of the dendrite; (d)
dendrite in course of development (from Porter and Easterling 1992).
 Phase Transformation 225

In normal eutectic growth (figure 8.19), as for the Ag-Cu alloy, there is
simultaneous formation of two solid phases, a and at the expense of the
liquid phase. The most common microstructure in the case of eutectic solidifica
tion has a lamellar morphology, which results from the growth mechanism il
lustrated in figure 9.13. This involves a constant exchange of A and atoms
between the growth fronts of the a and lamellae. The growth rate strongly in
creases with the degree of undercooling until diffusion phenomena slow down
the process.

Figure 9.13. Development of an eutectic structure, represents the distance between

the lamellae of the same phase. The flux of atoms A and between the lamellae
through the liquid is shown by arrows and J . This flux develops after the rejection
B

by of atoms in the liquid. In a similar way, rejects A atoms.

In general, all the microstructures become finer as the transformation rate in

creases. The acceleration of the transformation is only possible if the heat and
material transfers are rapid, i.e. if the distances to be covered are small. Crystals
usually form with morphologies favourable to transfer phenomena, this explaining
the appearance of crystals in the form of needles, dendrites and eutectic lamellae.

9.3.5. Overall kinetics of phase transformations - AvramFs Theory.

The kinetics of phase transformations (liquid-solid and solid-solid) which pro
ceed by a nucleation-growth mechanism generally obey a transformation law pro
posed by Johnson, Mehl and Avrami,. This theory provided an equation enabling the
extent of phase transformations to be calculated as a function of time.
The development of a new phase within an existing phase a can be repre
sented as follow. Nuclei of phase form within the parent phase and these nuclei
grow at the expense of the phase. At the final stage, the formed grains interact with
each other and the transformation is slowed down (figure 9.14.).
In the simplest case, a uniform growth (spherulitic growth) occurs in the three
spatial directions. A nucleus appearing at time t = 0 will reach, at t > 0, a volume
equal to:

V= nG t'
4 3
(9.29.)
3

G is the radial growth rate and Gt the spherulite radius at time t. A spherical grain
nucleated at time t = reaches, at time / > , a volume:
226 Introduction to Materials Science

T- cste

Log time t

Figure 9.14. Isothermal variation of the volume fraction/of the transformed phase as a
function of the logarithm of time /, according to (9.31).

(9.30.)

Initially, the growth of the new phase occurs freely. This behaviour changes
when impingement of the growing grains becomes important. By taking this into
account and relying on the laws governing nucleation and growth, a general equation
is obtained giving the extent of conversion (volume fraction f) as a function of the
transformation time t (figure 9.14.):

/=l-exp(-0 (9.31.)
is the overall rate constant for the transformation taking into consideration the diff
erent factors involved in the equations describing nucleation and growth. Thus for a
spherulitic growth initiated by a sporadic homogeneous nucleation of rate /, the ex
ponent n = A and it can be calculated that:
K=(n/3)IG )3
(9.32.)

The relation (9.31) is called the Avrami's equation, varies from 1 to 4 as a func
tion of the nucleation processes and of the growth type. varies strongly with the
temperature. Knowing K, the time required to reach a defined extent of conversion
(1, 50, 90 %for example) at a given temperature can be calculated.

9.3.6. Time - T e m p e r a t u r e Transformation (TTT) Diagrams

From Avrami's equation (9.31), curves can be calculated for different values of
the volume fraction transformed as a function of time and of temperature (TTT dia
grams). In order to perform this calculation, two values for the extent of the trans
formation are chosen which can be determined experimentally and characterise the
start (1 % conversion) and end of the reaction (99 % conversion). Figure 9.15. (a)
describes the principle of this method and shows how to determine the values of t s

and //at a temperature T= T\ inside the transformation range. Curve 9.15. (b) estab
lishes the relation between the TTT diagram and the isotherm of transformation at
the temperature T=T\. These TTT diagrams permit the determination of the type of
thermal treatment that should be applied to a material to obtain a definite microstruc
ture.
 Phase Transformation 227

While such diagrams can, in principle, be obtained for all phase transformations,
they are difficult to determine experimentally for the crystallisation of metals and
their alloys because of the high rate of these transformations. Consequently, the TTT
diagrams are mostly used for the characterisation of relatively slow transformations
such as solid state phase transformations of metals, crystallisation of organic poly
mers and certain ceramics.
Figure 9.15. (a) shows that the transformation time tends to infinity when the
temperature = T . At the thermodynamic equilibrium temperature T , the free en
e e

thalpy of formation of the nuclei is infinite (figure 9.6.) and the transformation is
impossible. It can only be initiated at a temperature of < T . At low temperatures,
e

the transport term limits the rate and increases the transformation time. These charac
teristics of the TTT diagrams are based on the nucleation and growth theory, just
summarised.

(a)

Log time t

Figure 9.15. Schematic overview of the TTT diagram for phase transformations, (a)
Curves for phase transformation as a function of time and of temperature for two
values of the conversion (1 %, (t ) and 99% conversion (//)); t corresponds to the
s mi

minimum time for the transformation to commence effectively, (b) Relation between
the TTT diagram and the transformation isotherm at temperature T=T h

In general, the nucleation and growth rates of the phase pass through a maxi
mum at Tmax and T respectively below the equilibrium transformation T . Ac
max e

cordingly, the rate constant (9.31) that is a product of the nucleation and growth
rates, will also pass through a maximum at a temperature ' between T and max

Tmax. A maximum transformation rate corresponds to a minimum transformation

time.
228 Introduction to Materials Science

Interpreting the TTT diagram is very straightforward. As an example, a tempera

ture T\ < T is selected and the progress of the transformation analysed. After the
e

material has been rapidly cooled (quenched) from > T down to temperature T\
e

(solid line in figure 9.15 (a)), the material is held at this temperature. The transfor
mation starts at time t = t . In the early stage (figure 9.15. (b)), the transformation
s

takes place at an increasingly rapid rate (Aflat) since the transformation rate is propor
tional to the volume transformed. Finally, the transformation progressively slows
down when impingement of the growing bodies becomes important. The curve in
dotted lines in figure 9.15. (a) shows the end of the isothermal transformation as a
function of the temperature.
A typical example of solid-state transformation is given by the eutectoid trans
formation of steel containing 0.8 %by weight of carbon. Its TTT diagram is shown
in figure 9.16. Starting from Fe (austenite), lamellar eutectoid microstructures
(pearlite) of a Fe (ferrite) and iron carbide Fe C (cementite) are formed (cooling curve
3

(1)). At lower temperatures (cooling curve (2)), the microstructure becomes finer.
This fine lamellar perlite considerably increases the hardness and the elastic limit of
steel. The manufacture of cables for cable-cars (chapter 11., Illustrative example) is an
important application for eutectoid steel with extremely fine lamellae.

(a) (b)

Figure 9.16. (a) Equilibrium diagram for the eutectoid transformation of austenite
(/Fe) into ferrite (aFe) and cementite (Fe C) and (b) TTT diagram for a steel with
3

eutectoid composition (C = 0.8 weight % C). In this case, thick (1) or fine (2) lamellar
E

perlite ( 9.3.4), or martensite (3) ( 9.3.8), is formed.

In practice, continuous transformation diagrams (CT diagrams) indicating the

microstructures formed on continuous cooling at different rates (exercise 9.6.9.), are
often used. In this way, optimal quenching conditions can be determined.

9.3.7. Phase transformation by spinodal decomposition

In binary systems consisting of a homogeneous mixture of two molten compo
nents A and B, local fluctuations in concentration constantly occur as a result of
thermal agitation.
 Phase Transformation 229

By applying the laws of thermodynamics, it can be shown that the mean ampli
tude of these fluctuations depends on the sign of the second derivative of the free
enthalpy (d G/dX ) relative to the concentration (molar fraction XB).
2 2

When the second derivative is positive, the local concentration fluctuations are
accompanied by a local increase in the free enthalpy due to thermal agitation around
the mean composition. The local concentration fluctuations tend to die away. This
occurs in a system in a stable equilibrium state as described in figure 8.11.
When the second derivative is negative (figure 8.14. at ) all local variations in
concentration lead to a reduction in the free enthalpy of the system. In such a system,
the concentration fluctuations have a tendency to increase throughout the volume,
leading progressively to the separation into two phases. This phase transformation,
which takes place gradually, without the initial creation of an interface and without
nucleation, is called a spinodal decomposition. Such a transformation is shown in
figure 9.17. (a) where the gradual evolution of a single phase system with initial
composition XB to a two phase system (a + oQ characterised by compositions
{

and is observed. The interface between the two phases, formed by a concen
tration gradient, develops gradually. The spinodal decomposition evolves sponta
neously to a stable state without passing a thermodynamic barrier. From a thermo
dynamic point of view, this system is unstable ( 8.3.2.), the diffusion coefficient
become negative and diffusion takes place in the opposite direction to that normally
predicted by Fick's law because, here the flux occurs towards the highest
concentration and the concentration gradient increases.
In metastable systems characterised by a minimum in free enthalpy
(d G/dX > 0) any local variation of small amplitude decays. The initiation of a
2 2

phase transformation in such a system only takes place if the concentration

fluctuations are of high amplitude leading to nuclei formation. In this case, from the
beginning of the transformation, two phases of compositions X^ and X# are
formed, separated from the initial phase of composition X by a well-defined
B

interface (figure 9.17. (b)).

The processes involved in the formation of a miscibility gap in a two compo
nents (A and B) system (figure 9.17) will be described to illustrate the mechanisms
of such a phase transformation. In the zone included in the miscibility gap, two dis
tinct phases, a\ and a are formed at the expense of a phase a of the same crystalline
2

structure ( 8.4.1.). The phase equilibrium diagram corresponding to a miscibility

gap formation is shown in detail in figure 9.18. (b). The curve bordering the misci
bility gap, determined by the common tangent method (figure 9.18 (a)), is shown by
the continuous line in figure 9.18 (b). At temperature T\ the equilibrium diagram
predicts that two phases ot\ and a of composition X# and X# are formed. Between
2

these compositions, the free enthalpy (figure 9.18 (a)) curve passes through a maxi
mum and is characterised by two points of inflexion and I" corresponding to con
centrations X,s and %B,S. For concentrations between XB,S and XB,S, the free en
thalpy second derivative is negative and the system is unstable. The two composi
tions XB,S and XB,S therefore define the spinodal transformation zone at temperature
.
By varying T\ the miscibility gap delimited by the solid-line curve in figure
9.18. (b), the region of the spinodal transformation shown in hatched lines in the
same figure is constructed.
230 Introduction to Materials Science

Distance ^ Distance JC
(a) (b)

Figure 9.17. (a) Representation of local variations in composition: (a) of a spinodal de

composition and (b) of a phase transformation by nucleation-growth. In the two dia
grams, XB represents the composition of the initial phase a and XB and X' the compo
B

sitions of the final phases and a .

2

In the temperature and composition zone inside the hatched area, all concen
tration fluctuations bring about a diminution of AG (figure 9.18. (a)) and inevitably
m

lead to spinodal decomposition. In the two regions situated between the solid- and
dotted-line curves, the second derivative of the free enthalpy is positive and, in this
case, the system is metastable and the phase transformation takes place via a nuclea
tion-growth mechanism.
Spinodal decomposition occurs in a certain number of mineral glasses (Pyrex
glass, figure 9.19.) and equally in a certain number of metal alloys such as in Alnico
alloys (Fe, AI, Ni, Co) for permanent magnets.

9.3.8. Transformation without diffusion

The majority of phase transformations involving diffusive motion can be frozen
by fast cooling (quenching). This characteristic makes it possible to obtain structures
which are far from their equilibrium state. There are three important cases to discuss
according to the nature of the initial phase.
When the parent phase is a liquid, a glass is produced providing the quenching
is sufficiently rapid, i.e. the liquid is cooled at a high enough rate to avoid the start
of crystallisation. The critical cooling rate depends on the time required to reach the
 Phase Transformation 231

Figure 9.18. (a) Variation of the molar free enthalpy of mixing in a binary system with
partial miscibility as a function of composition at T\ (b) Phase equilibrium diagram.
The solid-line curve defines the two-phase zone (figure 8.14); the hatched area repre
sents the spinodal region. T is the critical temperature above which the solubility in
c

phase a is complete.

Figure 9.19. Electron micrography of a Si0 -Na 0 glass showing spinodal decomposi
2 2

tion, observed after chemical etching to increase the relief. Note the interconnection
between the different phases.
232 Introduction to Materials Science

Figure 9.20. TTT diagram for the crystallisation of a liquid. At a high quenching rate
(very short cooling time) (a), the liquid solidifies without crystallisation and a glass is
formed. If the cooling rate is sufficiently slow (b), generally a polycrystal forms.

temperature 7 ' where the transformation rate is maximum (dotted cooling curve - fi
gure 9.20). At temperatures below T\ the liquid progressively solidifies without
significant atomic rearrangement and a glass is formed. The real problem, therefore,
is not to know whether a liquid is vitrifiable but to determine at what rate it is ne
cessary to cool in order to prevent crystallisation.
In practice, glass formation occurs in systems crystallising slowly, but all liq
uids are in principle capable of being vitrified. It has been recently possible to vitrify
certain metal alloys by hyperquenching techniques (illustrative example in chapter
4.). As already mentioned, certain organic polymers are structurally incapable of
crystallising (heterotactic polymers, figure 5.2. (c)). Whatever the cooling rate, these
materials always exist in the solid state in a glassy form only.
In a solid crystalline solution, the composition remains unchanged if the
quenching rate is sufficiently high. It is then possible to retain alloy elements in a
supersaturated solution. In this case, as in the case of glasses, the system keeps its
initial composition.
In the allotropic transformation, the parent crystalline phase changes after slow
cooling into another crystalline phase of different composition. Nevertheless, at a
high quenching rate in such systems, an important structural modification can occur
at low temperature without diffusion, i.e. without changing the phases composition.
This is the martensitic transformation.
In the previous chapter, it was seen that at high temperature, iron and carbon
form a solid solution in fee 7-iron (austenite) containing up to 0.8 % by weight C at
the eutectoid temperature (727 C). Carbon is much less soluble in bcc a-iron
(ferrite): 0.02 % by weight C at the eutectoid temperature (figures 8.23. and 9.16.).
If austenite of composition C is cooled slowly, (figure 9.16. (b)), the eutectoid
E

transformation takes place and lamellar pearlite forms. However, austenite quenched
under the condition (3) of figure 9.16. (b), does not keep an identical structure at low
temperature. It transforms into another metastable phase: martensite consisting of a
 Phase Transformation 233

supersaturated iron and carbon solution. Overall, martensite has a bcc structure like
ferrite, but its crystal is deformed by the excess of carbon trapped during the trans
formation from austenite into ferrite.
The crystallography of the martensitic transformation is illustrated in figure
9.21. A tetragonal centred lattice (ferrite crystal deformed along [001] axis) is formed
from fee austenite (white atoms) without significant displacement of the atoms. This
explains why the martensitic transformation occurs without diffusion and almost
instantaneously. The transformation propagates at a rate close to that of sound.

Figure 9.21. Schematic representation of the austenite to martensite transformation in a

carbon steel. The fine lines indicate two face centred cubic unit cells. The tetragonal
centred (deformed bcc) martensite Fe () lattice is shown with thick lines and white
atoms. To keep the figure simple, the two central atoms of the front faces are omitted.
The carbon atoms occupy all sites represented by black circles.

The martensitic transformation starts suddenly at a well-defined temperature Ms

and finished at some lower temperature, M . In the carbon steels, the degree of trans
f

formation is a function of temperature only. The transformation only continues by

lowering the temperature. The martensitic transformation is given by a horizontal
line (isotherm) in the TTT diagram. This transformation can be represented by a se
ries of lines M o, M o giving the temperatures at which the martensitic transformation
5 9

achieves a degree of 50 and 90 %respectively. The martensitic transformation is also

strongly influenced by the steel carbon content. Thus, in non-alloyed steels, Ms
varies from - 5 0 0 C for 0.2 wt.% to - 2 0 0 C for 1 wt.% weight carbon.
The martensite, formed in carbon steels, is a hard and brittle phase. Martensitic
structures equally form in other metallic alloys, such as in nickel steels or in
titanium. In these cases, the martensite is not brittle.

9.4. Summary and Conclusions

Phase transformations, occurring on the processing of materials, influence
greatly their properties. Materials are usually shaped and formed starting from the
liquid state. Their final microstructure develops during phase transformations
occurring during the heat treatment operation.
A transformation always occurs at a temperature below that of the
thermodynamic equilibrium temperature T determined by the phase diagram. The
e

driving force for these transformations is the free enthalpy difference between the
phase being formed and that undergoing transformation. The free enthalpy difference
234 Introduction to Materials Science

between the phases in transformation is proportional the temperature difference

T -T.
e

The transformation rate passes through a maximum at a specific temperature

T' < T because a certain degree of atomic or molecular mobility (diffusion) is
e

required by the phase change. This mobility decreases considerably when the
temperature is lowered, becoming zero at 0 or before.
By varying the temperature and time of treatment, it is possible to obtain a large
number of microstructures. The microstructures developing during the thermal
treatments are those having the fastest transformation kinetics.
In most cases, the phase transformations are initiated by a nucleation process. A
nucleus of a new phase appears after fluctuations of large amplitude of the structure
and/or composition. As a thermodynamic barrier arising from the presence of an
interface exists between the phases, a new phase can never appear at the equilibrium
transformation temperature. At low temperature, slow diffusion processes hinder the
transformation.
During the growth following nucleation, there is absorption of atoms or
molecules at the interface between the phases undergoing transformation. The three
classes of materials behave differently.
Metal crystallisation is rapid and easy. A metal interface is rough on an
atomic scale with numerous sites for attaching atoms.
Ceramics and polymers, on the contrary, have more complex crystal
structures, making atom or molecule absorption more difficult. The
crystal growth of these materials is generally slower and is aided by the
presence of defects, such as screw dislocations. This explains why
ceramics and polymers are easily obtained in the glassy state.

The degree of transformation as a function of time follows an S-shaped curve.

Determination of these S-shaped curves for a series of temperatures allows a TTT
(time - temperature - transformation) diagrams to be obtained. These diagrams
indicate the start and the end of phase transformation such as the solid-state
formation of pearlite in steel. They are therefore useful for defining the thermal
treatment for different steels. Each steel has a different TTT diagram. Another type of
diagrams is the (non-isothermal) continuous cooling diagram (CT). There, the
observed microstructures are induced by the various cooling rates.

9.5. Illustrative Example: 'Snowstorm' in an ingot

The quantity of snow crystals deposited by a snowstorm is impressive. An at
mospheric depression inducing a 50 cm deep snowfall corresponds to the precipita
tion of several million flakes per m , i.e. there are several times 10 flakes produced
2 12

in a depression. The basic structure of these crystals in the form of hexagonal plate
lets of 10 to 20 thickness and 0.1 to 2 mm diameter is shown in figure 9.22.
The hexagonal platelet morphology of these monocrystals in the form of a six-
branched star reflects the hexagonal symmetry of the structure of the ice crystal. The
snowflake is a dendrite having a structure similar to that which forms in metals during
 Phase Transformation 235

Figure 9.22. Hexagonal dendritic platelet structure of a single crystal snow flake.

their solidification. It differs in that the symmetry is hexagonal and not cubic as in
most metals, and in the growth mechanism ( 9.3.3) which takes place in the
atmosphere from water vapour.
A hexagonal platelet dendrite of snow only forms at a temperature of the order
o f - 1 5 C. Above - 8 C, growth occurs perpendicular to the hexagonal bases leading
to the formation of acicular prismatic crystals having their length eight to twenty
times greater than their diameter. Some snowflakes are formed by crystal entangling
often accompanied by partial melting with a more or less significant loss of the
dendritic structure
The morphology and size of the flakes has an important influence on the
mechanical behaviour of the snow. A comparable relation exists between dendrite
morphology and size and mechanical properties of metal alloys.
The grain structure of a casting after solidification is shown in figure 9.23. Note
that one grain is formed by one dendrite with many ramifications. It can be seen that
the grains have different shapes and sizes, depending on their position in the object.
On cooling the melt, displacements of matter occur in the ingot. A higher density of
grains of the solid relative to the liquid leads to some settlement of the grains being
formed. Convection movements also occur resulting from density differences of the
liquid as a function of its temperature and composition. These induce grain swirling
movements, similar to those observed in a snowstorm.

Figure 9.23. Grain structure of a casting after solidification.

236 Introduction to Materials Science

Using computer-aided simulation of the solidification process, the importance of

the effect of liquid convection on the final microstructure can be demonstrated. The
liquid descends along the solidification front, and rises towards the centre of the in
got (figure 9.24. (a)); this leads to the formation of the heterogeneous structure simu
lated in figure 9.24. (b). The calculated microstructure is similar to that observed in
real articles (figure 9.23.)., In the absence of convection (figure 9.25.) in the liquid,
the grain structure is homogeneous as shown by the corresponding computer model
ling.

(a) (b)

Figure 9.24. Computer-aided simulation of the solidification process taking account of

the effect of liquid convection on the final microstructure. (a) Intermediary situation
with indication of isotherms in the ingot. The symbols 7, , T , ... correspond to
2

isotherms in decreasing order of temperature, (b) Final microstructure.

(a) (b)

Figure 9.25. Simulation of the microstructure development in the absence of convection

during solidification.
 Phase Transformation 237

In certain cases, the liquid convection on cooling is increased to give products

that are more homogeneous. For example, in steel continuous casting, a powerful
electromagnetic agitation is used to induce in the centre of the casting a
microstructure of equiaxed grains of uniform size, something, which improves the
properties. Generally, fine dendrites and grains give better mechanical properties. To
obtain this, metals are inoculated before casting, i.e. nuclei forming substances are
introduced to initiate the transformation.

9.6. Exercises
9.6.1. What are the main mechanisms involved in diffusion in solid solutions?
9.6.2. Consider a tube, 30 mm in inner diameter and containing 50 1 0 gaseous 24

nitrogen atoms per m downstream from an iron membrane 0.01 mm thick.

3

Upstream from this, the nitrogen concentration is equal to M O atoms per m .

2 4 3

Calculate the total number of nitrogen atoms crossing the membrane per hour at
700 C given that the diffusion coefficient of nitrogen D in iron at this temperature
is equal to 4T0~ m s~ .
11 2 1

9.6.3. Calculate the dimension of the critical radius r* and the number of atoms in
the spherical nucleus of critical size when copper solidifies by homogeneous
nucleation. The latent heat of fusion of copper is equal to 1628 10 Jm" and its
6 3

specific interfacial energy is 177-10" Jm" . To a first approximation, homogeneous

3 2

nucleation occurs when supercooling A J i s of the order of 0.2 T (K).

m

9.6.4. The diffusion coefficient D involves, in the exponential, two types of energy.
Define these energies.
9.6.5. A diffusion reaction is complete in 5 s at 600 C but requires 15 minutes at
290 C. Calculate the time necessary to carry out this diffusion reaction at 50 C,
given that the diffusion coefficient (9.6) obeys an Arrhenius law.
9.6.6. Cementation is a surface treatment process for steels that consists of exposing,
at high temperature, the surface of a steel piece of an atmosphere rich in carbon
(methane for example). By which physical phenomenon does the carbon atoms
penetrate into the surface? What is the law characterising this phenomenon in a
stationary state? Define the parameters of this law and their units.
9.6.7. In the case of a transformation following Avrami's equation (9.31) with an
exponent = 4 (spherulithic growth at a constant sporadic nucleation rate), calculate
the time required to achieve a 90 % degree of transformation if the volume fraction
converted reaches 25 % after a transformation time of 150 sec.
9.6.8. What are the necessary conditions to obtain a precipitation hardening of the
silver/copper system? Please, consider the equilibrium phase diagram for the Ag - Cu
system in figure 8.19.
9.6.9. Figure 9.26. shows the diagram for the transformation by continuous cooling
of a steel with 0.45 % C. A sample of this steel is austenised at 900 C then cooled
in a continuous manner to ambient temperature. Figure 9.26 shows the two curves
for continuous cooling for the centre and the surface of the sample. What will be the
hardest zone of the sample?
238 Introduction to Materials Science

900 r

i ^^
00
r+F
700

( / V

1\ B
500

300
v

Surface Centre
100

0
10 1
10 10 3
10 5

T i m e / (s)

Figure 9.26. Diagram of the transformation for the continuous cooling of steel with 0.45
%by weight C. F = ferrite, = perlite, = martensite.

9.7. References and complementary reading

Y. ADDA, J. PHILIBERT, La diffusion dans les solides, Presses universitres de France, Paris, 1966.
M. AVRAMI, J. Chem.Phys. 7, 1103, (1939), 8, 212, (1940), 9, 777, (1941).
H.I. AARONSON (ed.), Lectures on the Theory of Phase Transformations, The Metal. Soc. AIME,
(U.S.A.), 1975.
J. BURKE, La cinetique des changements de phase dans les metaux, Masson, Paris, 1968.
G.A. CHADWICK, Metallography of Phase Transformations. Butterworths, London, 1972.
R.W. DAVIDGE, The Structure of Special Ceramics with Particular Reference to Mechanical
Properties, Proa Br. Ceram. Soc, 20 (1972) 364.
H. JONES, Rapid Solidification ofMetals and Alloys, Institution of Metallurgists, (Great Britain), 1982.
W. JOHNSON and R. MEHL, Trans. AIME 135, 416, (1939).
W. KURZ, D.J. FISHER, Fundamentals of Solidification, Trans. Techn. Publications, Zurich
(Switzerland), 1998.
G.E. MOLAU, S.L. AGGARWAL (ed.), Block Polymers, Plenum Press, New York, 1970, p. 79.
M. MORTON, Thermoplastic Elastomers, J. Polym. Sei, C 60, (1977), 1.
D.A. PORTER, K.E. EASTERLING, Phase Transformation in Metals and Alloys, 2nd ed., Chapman &
Hall, London., 1992.
G. STROBL, The Physics of Polymers, Springer, Berlin, 1996.
E. L. THOMAS, ed. Structure and Properties of Polymers, Materials Science and Technology, Volume
12., R. W. CAHN, P. HAASEN, E. J. KRAMER, eds., VCH, Welheim, 1993.
J. D. VERHOEVEN, Fundamentals of Physical Metallurgy, Wiley, New York, 1975.
J. ZARZYCKI, Glasses and the vitreous state, Cambridge University Press, Cambridge, 1991.
 Chapter 10.

Microstructures

10.1. Objectives
To give a brief description of the main microscopic methods.
To summarise the techniques used for microstructures.
To describe the principal microstructures of different materials (metals
and their alloys, ceramics and organic polymers) and to explain their
formation in terms of phase equilibrium diagrams and transformation
kinetics.
It is important to characterise the various microstructures formed in materials
since these determine most of the mechanical, physical and chemical properties.
Diverse microstructures are found in materials: crystalline or amorphous phases,
often containing precipitates of one or more dispersed phases, spherulites, lamellar or
fibrous structures, etc.

10.2. Observation of Microstructures

10.2.1. Preliminary note
The properties of a large number of materials are determined by the micr
ostructure which is made up of an ensemble of organised elements which can be ob
served by microscopic techniques.
Electromagnetic radiation of various wavelengths is used in microscopy. The
resolving power ( 3.4.1) of a radiation is determined by its wavelength. For optical
microscopy, this is of the order of 0.5 . In an electron microscope using higher
energy electron beams, the maximum resolving power is close to the dimension of an
atomic radius, i.e. lower than one nm. Nowadays, electron microscopy is the com
monest method for studying solid structures on an atomic scale. Other techniques are
also used to observe microstructures but in this basic treatment, consideration will be
restricted to optical and electron microscopy.

10.2.2. Optical microscopy (visible light)

Two main experimental techniques are used: transmission microscopy where
the light passes through the sample and reflection microscopy in opaque samples
where the light is reflected at the sample surface. Reflection optical microscopy has
been extensively used to study metallic materials since the middle of the nineteenth
century. Transmission optical microscopy is used to follow the development of a
semi-crystalline microstructure in thin transparent polymer films (figure 10.5.).

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

240 Introduction to Materials Science

Eye piece

Light

(a) (b)

Figure 10.1. Direct observation of microstructures by reflection optical microscopy (a),

of a polished and chemically etched metallic surface. In a pure metal (b), grain bounda
ries are selectively attacked and form grooves which do not reflect light into the micro
scope and appear therefore as black lines.

Sample polishing and chemical etching are preliminary treatments in the de

termination of metallic microstructures by reflection optical microscopy (figure 10.1
(a)). Polishing produces a surface comparable to that of a mirror. Chemical etching of
the polished area induces selective modifications of the surface topography {micro
graphy) and reveals differences of crystallographic orientation and chemical compo
sition. Grain boundaries and other inhomogeneites appear as dark lines or zones
(figurelO.l (b)).

10.2.3 Electron microscopy

To observe the fine details of the microstructure, it is necessary to use magnifi
cations greater than 1000 X (limit of optical microscopy) by using electron micros
copy techniques. Polymer structures, where typical dimensions are generally smaller
than a micron, are usually studied by electron microscopy as are defects structure in
metals and sintered ceramics.
Among all the particles, used to build up an image (protons, photons, X-rays,
electrons, etc.), electrons are the most useful. They are easy to generate and they are
accelerated by an electrostatic field and since they have an electric charge they can
be deflected, i.e. focussed by electromagnetic fields. With electrons accelerated by a
potential difference of 100 kV and having an associated wavelength of 3.7 pm, a
resolution of the order of interatomic distances (approximately 0.3 nm) is obtained.
However, electrons are strongly absorbed by matter and extremely thin samples (-0.1
in metals) must be used. Sample preparation is often delicate and demands a se
ries of special techniques such as electrochemical thinning or ionic bombardment.
A transmission electron microscope (TEM) consists of:
an electron source,
a condenser (electromagnetic lens Li) serving to focus the electron
beam onto the object AB,
an objective lens L magnifying the object,
2
 Microstructures 241

a projector (Lens L , L ) enlarging the image and projecting it ( A B )

3 4 3 3

onto a fluorescent screen, similar to a TV screen or onto a photographic

plate.
The optical path of the transmission electron microscope is basically similar to
that of the optical microscope (figure 10.2.).

LtJ Electron gun

Lens (condenser)

Object
Lens (objective)

First image

Lens
Second image
Lens (projector)

L i A3B3 Image on fluorescent screen

Figure 10.2. Optical path of a transmission electron microscope.

The mechanism of image formation in a transmission electron microscope in

volves electron diffraction or diffusion whereas in transmission optical microscopy,
variations of light absorption produce the image. For example, during the observation
of a crystalline material, electrons in passing through the sample, may interfere with
atoms in the crystal planes and diffract according to Bragg's law (3.10.).This partial
diffraction of the incident electron beam reduces its intensity. This more or less
significant attenuation appears on the screen. The sample image appears in dark on a
light background. This is known as a bright field image. Besides bright field, other
image formation techniques exist which will not be treated here. For amorphous ma
terials, the image formation mechanism remains similar except that the electron beam
is not diffracted in specific directions but more or less diffused in all directions.
The presence of defaults in the crystal structure of materials has been demon
strated by transmission electron microscopy. For example, dislocations can be di
rectly seen. The image of a thin crystal slice (10.3. (a)), containing dislocations
obliquely crossing the sample, is shown in figure 10.3 (b). The dislocations appear as
dark lines.
When the surface of a sample is struck by an electron beam, various secondary
phenomena take place as a result of the beam/sample interaction (emission of secon
dary electrons or Auger electrons, X-rays, etc). The secondary electrons can be used
to form an image with the aid of an appropriate detector. This gives rise to another
type of electron microscope known as a scanning electron microscope (SEM).
242 Introduction to Materials Science

(a) D c
(b)

Figure 10.3. Dislocation images obtained with a transmission electron microscopy: Po

sition of defects (a) in the thin film and the transmitted image (b).

In this technique, a fine electron beam of about 10 nm diameter is focused on the

sample and scanned over the surface. The back-scattered electrons are collected by a
detector, placed laterally, thus allowing the sample surface to be analysed. It is
therefore possible to avoid the difficulties of thin sample preparation. By simulta
neously analysing the emitted X-rays, a topographic description of the surface
chemical composition can also be obtained.
Recently a series of very spectacular advances have been made in this technique.
It is now possible to work while maintaining the microscope chamber under a con
trolled pressure of water vapour (Environmental Scanning Electron Microscopy -
ESEM) and to make measurements on hydrated materials such as wood without
modifying the original microstructure.
Several microscopic techniques have emerged recently in particular scanning
tunnelling microscope (STM) and atomic-force microscope (AFM), which make pos
sible the direct observation of atoms and molecules on surface. A short description
on these new techniques is given in Chapter 17.

10.3. Principal Microstructures

10.3.1. Pure Metal Solidification
When metals solidify, they normally form a poly crystalline solid (figure 9.5.
(a)). The number of active nucleation sites present when solidification starts deter
mines the grain size after solidification.
Metals generally form dendrites (figure 9.5. (a) (at / )). Dendrite formation is
4

explained in detail in paragraph 9.3.4. (figure 9.12.). In pure metals, the dendrites are
not visible when the solidification has finished but a polycrystalline grain micro-
structure is observed as in figure 9.5. (a) (at ts) and figure 10.1. (b).

10.3.2. Polymers crystallised from the molten state

TheSpherulite is the basic structure of semi-crystalline polymers (figure 9.5.
(b)). Polymer spherulites have a polycrystalline and partly amorphous structure. De
tails of the spherulite microstructure can be observed by electron microscopy and
diffraction techniques. A spherulite (figurel0.4.) is composed of very long crystalline
ribbons (lamellae) extending from a central nucleus. The ribbons progressively
spread out into branches as the spherulite radius increases. The polymer chains are
perpendicular to the crystalline lamellae and adopt a partially folded structure (figure
9.9 (c)). The thickness of the lamellae ( 1 0 - 5 0 nm) depends on the crystallisation
 Microstructures 243

temperature. The lamellae width is not predetermined, while their length is close to
the spherulite radius, reaching up to 500 when the nuclei density is reduced.
Amorphous chain segments tie the lamellae to each another. Between the lamellae,
amorphous material, made up of non-crystallisable polymer chains, is sometimes
found.
The spherulite microstructure described on figure 10.4. involves regular chain
folding with adjacent re-entry in the crystal. Other types of morphology are also pos
sible.

Growth
i Intercrystalline tie chain
I (amorphous)

Figure 10.4. Microstructure of a spherulite of an organic polymer (from Groupe Frangais

des Polymeres (GFP), 1982).

100

100

Figure 10.5. Optical micrographs in polarised light of partially crystallised polymers: (a)
spherulites of isotactic poly(l-butylene) during growth; (b) spherulites of crystallised
polyethylene (after Groupe Francaises des Polymeres (GFP), 1978).
244 Introduction to Materials Science

When the primary nucleation density is low, spherulites can be observed by

means of an optical microscope (figure 10.5.). In polarised light, polymer spherulites
are birefringent with characteristic extinction zones (Maltese Cross). As crystalline
growth proceeds by secondary nucleation ( 9.3.3.), polymer crystals grow more
slowly than metallic crystals. For a fast crystallising polymer (polyethylene), the
maximum crystal growth rate is about 100 . In metals this value is close to 100
-1

ms" . 1

Since polymers crystallise slowly, it is necessary to crystallise them at high

degrees of undercooling to obtain a sufficiently fast rate. In this condition, the
primary nucleation is greatly increased (figure 9.8.). The grain size becomes ex
tremely small (< 1 ) and is only visible in an electron microscope.
In this condition, the thickness of crystalline lamellae is also reduced because
the length of the secondary seed is decreased at high undercooling. Even when the
spherulites fill all the material volume, a fraction of amorphous material is always
present, as the degree of crystallinity of polymers varies usually from 30 to 60 and
seldom exceeds 80 %.
As in the case of metals, the grain size of parts made from cast or injected
polymers is not uniform throughout the section. The external skin in contact with the
mould is cooled more rapidly and crystallises with greater supercooling, giving finer
grains. Normally, the degree of crystallinity in the skin is lower. The polymer
microstructure can also be influenced by residual mechanical stresses induced during
the fabrication process.

10.3.3. Structures observed in metal alloys

Four types of microstructures are principally observed in metallic alloys.
Dendrite crystals formed during solidification and visible after
chemical etching in alloys due to local variations in chemical com
position (figurelO.6. (a)).
This segregation results from slow solid-state diffusion which prevents
thermodynamic equilibrium occurring. The alloy elements with higher
melting points are preferentially concentrated at the centre of the
dendrite whereas the external parts are enriched with the elements
having lower melting points (figure 10.7.).
Eutectic microstructures formed from liquid (eutectic transformation)
or solid (eutectoid transformation). Eutectic microstructures, made up
of two or more different phases, have lamellar (figure 10.6. (b)) or
fibrous morphologies.
Mixed microstructures i.e. dendritic and eutectic (figure 10.6 (c)).
Small particles appearing in the solid state by precipitation from a
supersaturated phase (figure 10.6. (d)).
All these microstructures occur under well-defined conditions as a function of
the phase diagram (figure 10.8. (a)) and the transformation kinetics. When materials
solidify, heat is released (latent heat of crystallisation), considerably slowing the
cooling rate (figure 8.16. (b)). When a pure metal is crystallised slowly, cooling stops
close to the melting point during the growth of the dendritic grains (figure 10.8. (b)).
 Microstructures 245

In alloys forming solid solutions (figure 10.8. (c)), solidification does not occur in an
isothermal manner as in pure metals. The dendritic crystals form progressively on
cooling.
When an alloy of eutectic composition (figure 10.9. (c)) is slowly cooled, all the
transformation occurs at a constant temperature T . As for the crystallisation of a
E

pure constituent (figure 8.16. (b)), the thermal analysis curve (fig. 10.9. (c)) exhibits
an isothermal solidification plateau. Two crystalline phases a and form simulta
neously as alternating fine lamellae (fig. 10.6. (b)) or as fibres.

(a) 100 (b) 10

Figure 10.6. Microstructures of metallic alloys: (a) dendrites, (b) lamellar eutectic, (c)
dendrites and inter-dendritic eutectic, (d) precipitates formed in the solid state within
grains. It is important to note the different magnifications of the various micrographs.

Figure 10.7. Diagram illustrating the variations in composition (C\ > C , etc.) of a metal
2

alloy dendrite after solidification in non-equilibrium conditions.

Alloys with compositions either inferior or superior (hypo-eutectic or hyper-

eutectic alloy) to the eutectic composition constitute the intermediate case between a
completely miscible alloy and the eutectic alloy. The thermal behaviour is shown by
the thermal analysis curves in figure 10.9. (b) and (c). The microstructure, shown in
the inserts, contains dendrites a and eutectic grains (a+).
246 Introduction to Materials Science

The solubility of a dissolved element varies considerably with temperature. An

alloy can be homogeneous at high temperature and exist in two phases at lower
temperatures. The name precipitation has been given to this phase transformation of
a supersaturated solid solution. This nomenclature was proposed by analogy with the
terminology used for liquid solutions where similar phenomena are observed.

Seeds Dendrites Grains

Liquid yT^j

<o
3

Pure A element
crystallisation
A Concentration C - B Time t -
(a) (b)
Figure 10.8. Microstructure formation of in a pure metal and in an alloy forming a solid
solution. Part of phase equilibrium diagram (a). Cooling curve (b) (thermal analysis) of a
pure metal, and (c) of a single phase alloy with the evolution of the microstructure shown
in the inserts.

Dendrite + Eutectic a + Eutectic

A
Concentration c Time t T i m e t
(a) (b) (c)
Figure 10.9. Microstructure formation in binary systems forming an eutectic (a). Cooling
analysis curve of a hypoeutectic alloy at concentration c (b) and of an eutectic alloy c
h 2

(c).

A well-known example is that of aluminium which dissolves up to 5.7 wt.% of

copper at 548 C (figure 10.10. (a)). The solubility drops to 0.2 wt.% at 200 C. At
540 C, an Al - Cu alloy containing 4 wt.% of copper forms a homogeneous solid
 Microstructures 247

solution. When this alloy is slowly cooled, precipitation of the excess copper in the
form of intermetallic -Al Cu particles occurs. Precipitation in the solid state is slow
2

and quenching to ambient temperature (figurelO.10. (b), point 1), produces a super
saturated, homogeneous solid solution containing 4 wt.% Cu. By consecutive an
nealing at an intermediate temperature, Al Cu precipitates are produced in a con
2

trolled manner. The microstructure varies as a function of time and temperature

(figure 10.10. (c))
The supersaturated homogeneous alloy is relatively soft, the hardness and yield
stress are low. Al Cu precipitation provokes a hardening, increasing the hardness
2

and yield stress. However, prolonged heat-treatment produces large precipitates,

making the alloy soft and brittle. The precipitation conditions (T and t) must be
closely controlled when a microstructure with maximum hardness is required (figure
10.10. (c)). However, the details of this hardening mechanism are more complex than
described here. There are also other phases than the - Al Cu forming. 2

Supersaturated crystal Al Cu precipitate

2

Figure 10.10. Microstructure formed by solid-state precipitation (precipitation harde

ning). Al - Cu partial equilibrium diagram (a). Heat treatment cycles (b) starting with
homogenisation of the alloy at a temperature inferior to the eutectic temperature, then
quenching and annealing at an intermediate temperature. Four microstructures are shown
in (c) after different annealing times in relation to the hardness H .
v

10.3.4. Principal microstructures of iron - carbon alloys.

There are two important categories of iron - carbon alloys: steels and cast irons.
Carbon steels are iron - carbon alloys containing less than 1.5 wt.% carbon. Alloys
with more than 2 wt.% carbon make up the cast irons.
At 727 C and with a concentration of 0.8 wt.% carbon (figure 9.16.), the Fe - C
diagram shows a eutectoid phase transformation from the (austenite) phase to a
(ferrite) + Fe C (cementite) phases.
3
248 Introduction to Materials Science

The microstructure of the slowly cooled eutectoid steel is shown in figure 8.1.
The alternate ferrite and cementite layers are clearly visible. This lamellar structure,
has an iridescent appearance similar to that of a pearl at low magnification in the
optical microscope, and has been given the name of pearlite (or perlite). The forma
tion of perlite occurs similarly to that of a eutectic microstructure (figure 10.9.), but
instead of a liquid phase, the solid fee /crystal is the phase which transforms.
When a steel of hyper-eutectoid composition (> 0.8 wt.% C) is cooled, cemen
tite (Fe C) forms first and the eutectoid perlite appears later. The pro-eutectoid ce
3

mentite nucleation takes place preferentially on the austenite grain boundaries since
nucleation and growth is easier there. The perlite subsequently appears below the
eutectoid transformation temperature. The presence of carbide networks along the
grain boundaries often makes the hyper-eutectoid steels brittle.
The microstructure of hypo-eutectoid steels (< 0.8 wt.% C) is formed of an a
iron solid solution (proeutectoid ferrite) containing less than 0.02 wt.% carbon and
perlite grains. These steels are widely used in engineering (figure 10.11. (a)).
With appropriate heat-treatments, a wide variety of microstructures can be pro
duced in steels. For example, when austenite (- Fe) is quenched, martensite, which
has a high hardness and yield stress ( 9.3.8.), is obtained (figure 10.1 l.(b))

(b)

Figure 10.11. (a) Typical construction steel (hypoeutectoid) microstructure after slow
cooling (~0.4 wt.% C) and of a quenched martensitic steel of the same composition.

The addition of 4.3 wt.% carbon (figure 8.23.) lowers the melting point of iron
by almost 400 C. This produces low cost eutectic alloys (cast irons) with useful
properties which can be cast in complicated shapes.
There are several major categories of cast iron:
g y lamellar cast iron containing graphite in the form of eutectic la
re

mellae which make it brittle;

grey spheroidal or nodular graphite cast iron with improved mechani
cal properties;
white cast iron where the carbon is present in the form of eutectic iron
carbide (Fe C) lamellae.
3

The lamellar morphology of traditional grey cast iron is due to impurities such
as sulphur. Grey cast irons can be adapted to various requirements by additions of
 Microstructures 249

other elements. Some elements (silicon) favour graphite formation. Adding Mg pro
duces cast iron with spheroidal or nodular graphite, this is ductile and suitable as a
substitute for steel in some applications.
In the car industry, malleable cast iron has been substituted for steel for the pro
duction of various parts such as car suspension arms. Chromium additions lead to the
formation of cementite and give white cast irons. This effect is reinforced by rapid
cooling. White cast iron is made ductile by a treatment at 900 - 1000 C transforming
the unstable cementite into austenite and nodular graphite.

10.3.5. Microstructures of sintered ceramics

Traditional ceramics, such as porcelain, are made from natural raw materials
(kaolin, quartz, feldspar) by a process called sintering. The preliminary grinding step
necessary to obtain raw material in powder form greatly influences the properties of
the final product. The raw materials are usually shaped from an aqueous paste
processed by a variety of techniques (hand moulding, compression, etc.). The shaped
articles are dried and baked ('fired') at a temperature, which varies between 900 and
1500 C as a function of the composition. Traditional ceramics are often porous
materials with complex microstructures which depend on the processing methods.
A porcelain microstructure is shown in figure 10.12. This material, used for
electrical insulation, is made from a mixture of quartz, feldspar and hydrated silicates
of aluminium and magnesium. The microstructure after baking (-800 C) is compli
cated. Several components are identified in figure 10.12:
quartz crystals (white grains) already present in the initial mixture and
which remains unchanged after the ceramic baking;
mullite crystals (grey), an aluminium silicate forming a network struc
ture around quartz crystals;
pores of various size (in black).
The technical ceramics (alumina: A 1 0 ; zirconia: Z r 0 ; silicon carbide: SiC)
2 3 2

produced from synthetic powders with controlled shape and size, have single-phase
microstructures. Porosity is virtually zero and their density reaches 99 % of the theo
retical value.
The sintering process is summarised in figure 10.13. The number of grains
shown in each frame is identical. The frame dimensions decrease a little since the
total volume of material is slightly reduced by sintering. The porosity, shown in
black, is initially significant but diminishes progressively. Overall, the grains grow a
little and the volume decreases slightly. The grain boundaries (figure 10.13. (c)) re
main visible under the microscope with the residual porosity localised in these re
gions. By using multidirectional compression (isostatic compression), densities ap
proaching the ideal value are obtained. The sintered microstructures are largely con
trolled by the size and shape of the granules and by the powder treatment. The pro
duction of powders with well-defined reproducible properties has become an impor
tant industrial activity.
Liquid-phase sintering involves a melting of part of the constituents at the
sintering temperature. A liquid film forms at the grain boundaries and the sintering
process occurs more readily. Structural rearrangements and re-precipitation take
place also during this consolidation phase. A liquid phase sintering process is
currently used in the manufacture of WC and Co based hard metals (WC = 90 %;
250 Introduction to Materials Science

Co = 10 %) which are used to make cutting tools. In this case, a thin film of liquid
Co bonding the tungsten carbide grains is produced by a heat-treatment effected
above the Co-WC eutectic melting temperature.

200

Figure 10.12. Microstructure of a technical ceramic composed of crystals of quartz

(white grains) and mullite crystals (grey) and pores (black) of various size.

(c)
-10

Figure 10.13. Representation of the three steps of the sintering process: (a) compacted
powder; (b) intermediate stage: the pore size decreases; (c) final product: the pores have
practically disappeared. The residual nanoporosity makes the grain boundaries visible.
 Microstructures 251

10.3.6. Microstructures of polymer blends and copolymers

Polymers hardly ever form alloys (or blends) by mixing macromolecules of dif
ferent chemical structure (chapter 8). Most polymers mixtures (blends - alloys) are
obtained by a chemical route, i.e. by copolymerisation.
Single-phase amorphous statistical copolymers (figure 8.28. (a)) do not have a
microstructure. Copolymers containing block (figure 8.28. (b)) and graft
(figure 8.28. (c)) homogeneous sequences, form an emulsion of finely dispersed
phases. These microstructures, shown in figure 10.14., depend on the relative lengths
of the A and blocks:
spheres, at small A or concentration;
cylinders, at intermediate A or concentration;
lamellae, when A and concentrations are around 50 vol.%.
Typical dimensions (sphere and cylinder diameter, lamellae thickness) are between
10 nm and 1 .
ABA triblock copolymers
TT

A
oil
A A,
V
Spheres Cylinders Lamellae Cylinders Spheres
~200 nm
Increasing A %
Decreasing %

Figure 10.14. Variation of the microstructure of tri-block copolymers ABA as a function

of composition (after Molau, 1970)

The physical properties of multiphase copolymers are completely different from

that of single-phase copolymer systems of the same overall composition. As a rule,
when rubbery spheres are dispersed in an amorphous glassy matrix, a rigid thermo
plastic with high impact resistance is obtained. In contrast, a dispersion of glassy
spheres in a rubbery matrix yields a thermoplastic rubber. One example of each type
of material made with copolymers constituted of homogeneous blocks will be suc
cinctly described.
Polystyrene is a thermoplastic with very poor impact resistance. In technical
applications, polystyrene homopolymer is replaced by an impact resistant, rigid
thermoplastic, which is called high impact polystyrene (HIPS). It is composed of a
blend of homopolystyrene and a grafted polystyrene-polybutadiene copolymer
(-20 vol.%) (figure 8.28. (c)). Its microstructure is shown in figure 8.2. The
dispersed spherical nodules have a complex morphology.
The polymer chain arrangements in a nodule microstructure are summarised on
figure 10.15. Solid lines represent polybutadiene chains and dotted lines, polystyrene.
A small number of polystyrene chains are linked at the two extremities to polybuta
diene chains (figure 10.15. (4)). This leads to the formation of a three-dimensional
network constraining the spherical nodules and making them plastically non-de-
formable. This three-dimensional network inhibits nodule fragmentation. The
252 Introduction to Materials Science

nodules retain their shape during the material processing (mostly injection moulding)
at high temperature. The size of the nodule is critical for the mechanical behaviour. It
is adjusted at polymer synthesis. It must be between 1 and 10 . Smaller nodules
have no effect on toughness and coarser particles are harmful for the mechanical
strength. Indeed, a high impact polystyrene is a complex mixture of three-
dimensional (insoluble) polymer nodules with a thermoplastic material forming the
continuous phase.

Figure 10.15. Details of a nodule microstructure in high impact polystyrene (HIPS). The
elastomer phase is in white and the polystyrene phase is indicated in dark. Several details
of the polymer chains organisation in the microstructure are shown: solid lines (1) for
polybutadiene chains (elastomer) and dotted lines (2, 3,4) for grafted polystyrene chains.
Various molecular structures are present: homopolystyrene (3) and grafted polystyrene-
polybutadiene copolymers (1-2, 1-4).

Our second example will deal with biphasic materials which associate the rub
bery behaviour of cross-linked rubbers and the injection mouldability of thermoplas
tics. Such materials, made of homogeneous blocks of different chemical composition,
are called thermoplastic rubbers.
Among the various systems exhibiting this type of mechanical behaviour, the
best known is made of tri-block hydrocarbon copolymers composed of two blocks of
a rigid polymer (like polystyrene) and one block of a very flexible rubbery polymer
(for example polybutadiene) in between. In this kind of material, the rubbery
polymer is in excess (between 60 to 80 vol.% rubber). Their microstructure (figure
10.16) consists of a dispersion of polystyrene glassy spheres in a continuous rubbery
matrix of polybutadiene. The size of the nodule is regulated by the length of the rigid
block.
At ambient temperature, the tri-block copolymer behaves as an elastomer (rub
ber) but when taken to high temperature (T> 100 C), the polystyrene nodules sof-
 Microstructures 253

ten, the biphasic materials becomes liquid and can then be shaped by injection
moulding.
In the three-dimensional networks of a vulcanised elastomer (chapter 5.), the
elastic chains are connected to each other by cross-links (figure 6.12.). In the absence
of cross-links, any mechanical deformation is irreversible and a very viscous visco
elastic liquid is obtained.
In the linear tri-block copolymers described in figure 10.16, the rubbery poly-
butadiene segments are connected to each other by a stiff polystyrene sequence. Each
polybutadiene segment is anchored in two different polystyrene glassy nodules by a
polystyrene chain segment. All the rubbery segments are therefore linked together by
the glassy spheres of polystyrene, making up a rubbery three-dimensional system.

r ~ .
Polystyrene Polybutadiene ^ Polystyrene
M=n 10- 12000 M = 50-70000 M = 10 - 12000
n

Figure 10.16. Microstructure of tri-block poly(copolystyrene-butadiene-styrene) co

polymers, containing approximately 30 wt.% polystyrene (Solid lines represent polybu
tadiene blocks (elastomer) and dotted lines, polystyrene blocks from M.Morton, 1977).

At temperatures between the glass transition temperature of polybutadiene

( ~ - 80 C) and that of polystyrene (~ 100 C), this material behaves like a sulphur
vulcanised network and can be reversibly deformed by more than 100 %. In practice,
for these tri-sequential copolymers of this chemical composition, the service tem
perature varies between about - 4 0 and +50 C. Above the glass transition tempera
ture of polystyrene (-100 C), the polystyrene nodules soften. The material is then
shaped like conventional thermoplastic polymers (high impact polystyrene, etc.). It is
required to use a tri- (or a multi-) block copolymer to obtain a thermoplastic rubber.
Polybutadiene-co-styrene di-block copolymers (figure 8.28. (b)) form also a
microstructure consisting of a dispersion of glassy spheres in a continuous rubber
matrix but the elastic segments of these di-block copolymers are only anchored on
one side to a nodule. A three-dimensional rubbery network is not formed in this case
but instead a very viscous viscoelastic liquid not useful as a material.
254 Introduction to Materials Science

10.4. Summary and Conclusions

The microstructure of a material is composed of different phases of variable
form, size and distribution (grains, precipitates, dendrites, spherulites, lamellae,
pores, etc.). The phases are distinguished from each other by their various crystalline,
semi-crystalline or amorphous structures when observed with an optical or electron
microscope. The engineer can obtain a wide range of properties by controlled micro-
structural modifications produced during processing. To have a clear understanding
of the material behaviour, it is needed to establish relationships between the macro
scopic properties and phenomena occurring on the microstructural scale.
The microstructures formed in materials depend not only on the chemical
composition and structure but also on the atomic mobility and on the presence of
concentration gradients during processing. Microstructure formation is also strongly
influenced by the amount of energy required to create new interfaces.
Appropriate heat-treatments (quenching, annealing) produce a great variety of
microstructures. Frequently, these have an inhomogeneous composition and are al
most always metastable at the service temperature. In the case of metals and alloys,
heat-treatments combined with mechanical processes such as rolling have reached a
very high level of sophistication in order to control the microstructure.
The majority of solidification microstructures are crystalline. However, a small
number of materials having an irregular molecular structure are unable to crystallise,
whatever the heat treatments. These materials have a glassy structure in the solid
state. It is the case of heterotactic vinyl polymers. Glass is always less stable than
crystal when it exists and owes its transparency to a lack of microstructure. The mi
crostructure of organic polymers is mostly controlled by their chemical architecture.
Macromolecules with a regular molecular structure are in general able to crystallise.
However, crystallisation of polymer materials is never complete and, in the bulk,
only semi-crystalline spherulites are formed.
Ceramics are frequently obtained by powder sintering. This explains the
presence of porosity, which is an important microstructural feature of this class of
material.

10.5. Illustrative example: the laser, a tool for the automobile

industry
A laser is a radiation source which makes use of a special light emission tech
nique in which there is phase coherence of the emitted light. Traditional light sources
such as the filament of an incandescent lamp only produce a spontaneous emission
with frequent phase changes leading to interference and significant energy losses. In
a laser, the atoms of a gas or a crystal are excited into a high-energy state by electric
or light stimulation. Part of this energy is re-emitted as an intense, parallel beam of
coherent electromagnetic radiation in a narrow wavelength band. This beam can be
focussed by a lens or a mirror, which permits high energy to be focussed onto a very
small spot for a very short or for long time (pulsed or continuous wave laser).
Various lasers operate over a range of wavelengths from ultra-violet to infrared. The
basic concept of the laser was proposed in 1917 by A. Einstein but it was only more
 Microstructures 255

than 40 years later that the first laser was built by Maiman in 1959. The term laser is
a universally accepted acronym derived from: Light Amplification by the Stimulated
Emission of Radiation.
Nowadays, the laser is an indispensable tool in the treatment of materials and an
important number of industrial applications have been developed (drilling, cutting,
welding). More recently, new surface treatment processes have appeared to improve
the wear, fatigue and corrosion resistance of metallic alloys. In these processes, for
example, a high power C 0 laser is used with a continuous emission of several kW at
2

10.6 wavelength. This type of laser may also be used for welding. To-day, Nd:
YAG lasers or diode lasers with shorter wave length are replacing the C 0 laser in 2

these applications.
The light beam from these lasers is focussed onto a small spot (the exact spot
size varies with the laser power). By sweeping the metallic surface with this spot, a
metallic layer with a fine microstructure is produced. When the laser surface
treatment is combined with metallic or ceramic powder spraying, an alloy layer of a
different composition from the bulk material is formed at the surface and the surface
properties may be drastically modified. In the example described, laser treatments
using powders of very different composition to that of the substrate will be con
sidered, enabling corrosion or wear resistance to be modified without affecting bulk
mechanical properties.
Figure 10.17. describes the process for deposition of a metallic alloy layer onto
he surface of a metal by the interaction of a laser beam with a powder jet. To obtain a
continuous treatment, the zone to be covered is treated by an oscillating beam. When
the laser beam interacts with a metallic surface, part of the energy is absorbed and
serves to melt the solid. The amount of absorption, which influences heating and
cooling depends on the nature of the material and its surface properties, the radiation
wavelength and the intensity. Treatment conditions depend on the laser source
intensity and the displacement rate. This can vary from several millimetres to several
tens of centimetres per second. The temperature gradient induced by the treatment is

Substrate

Figure 10.17. Schematic of the laser cladding process where powder is injected in the
molten pool to form a coating of different composition and properties with respect to the
substrate.
256 Introduction to Materials Science

of the order of 10 K s and the cooling rate varies between 10 and 10 K s " . Very
6 1 3 6 1

fine solidification microstructures are formed (figure 10.18.) under these

conditions which can improve surface properties. This coating process is particularly
interesting since the substrate surface melts superficially and is joined to the coating
material by a strong metallic bond.

Figure 10.18. Fine dendritic microstructure of a Stellite (Co-Mo-Cr-C alloy) deposited

by laser.

This process has other advantages.

The very short interaction time between the substrate and the laser
spot allows only slightly heating of the part with minimal defor
mation.
The coating layer thickness is well controlled reducing machining
operations.
The process is very flexible and can be automated.
The tool, which is a light beam, does not wear out.
The automobile industry currently uses this technique for the fabrication of
valve seats directly onto engine cylinder blocks (Figure 10.19). In the classic process,
valve seats (steel) are fitted onto the cylinder head (Al alloy). The laser treatment
process ensures better heat transfer between the cylinder and the cladded valve seat
leading to lower temperature which reduces wear. The absence of fitted valve seats in
the laser process permits larger diameter valves to be produced thereby improving
engine performance and reducing fuel consumption.

10.6. Exercises
10.6.1. The microstructure of figure 10.20 has been obtained by optical microscopy.
Define the elements of this microstructure.
 Microstructures 257

What do the black lines represent?

What treatment makes it possible to observe this microstructure?
10.6.2. Give the application range for the optical and transmission electron micro
scopes. Explain.

Figure 10.19. An engine cylinder block, laser treated with a hard, corrosion resistant al
loy (from T. Saito, Toyota process).

-100

Figure 10.20. Microstructure of a pure metal or a ceramic obtained by sintering.

10.6.3. Explain why the crystallisation of organic polymers is never complete. Sketch
the structure observed with an optical microscope at various stages of crystallisation
when the number of nuclei is small.
10.6.4. Sketch the interface of a spherulite crystal lamella during growth. What is the
kinetic step in polymer crystal growth. Basing arguments on paragraph 9.3.2, explain
why the lamella thickness decreases with temperature.
10.6.5. Describe the evolution of the spherulite volume fraction during an isothermal
crystallisation and sketch the microstructure development during this phase transfor
mation (figure 9.14.).
10.6.6. Sketch the microstructure evolution of steel containing 0.4 wt.% carbon
(figure 9.16.) and draw the microstructures observed after rapid cooling to and
holding at 740 and 700 C (The steel is cooled to ambient temperature after complete
transformation).
258 Introduction to Materials Science

10.6.7. What microstructure is observed when steel containing 0.4 wt.% of carbon is
cooled from 950 C to ambient temperature under the following conditions:
Very slow cooling to equilibrium;
rapid cooling (quenching).

10.6.8. From the Cu-Zn (brass) phase diagram given in figure 8.32, trace the cooling
curves and draw the microstructures of alloys containing 10 and 20 weight % of Zn
respectively.

10.6.9. The phase diagrams of the Al-Mg (a) and Al-Sn (b) alloys are given in
figure 10.21. Is precipitation hardening of aluminium possible in these systems? Jus
tify your answer.

0 20 40 60 80 100 0 20 40 60 80 100
AI Magnesium weight % Mg Al (b) Sn
(a)

Figure 10.21. Phase equilibrium diagrams for Al-Mg (a) and Al-Sn (b) alloys.

10.6.10 Figure 10.22 Shows the microstructure of a polypropylene (continuous

phase) - ethylene-propylene (dispersed phase) copolymer blend.
Describe and compare the microstructure with that of high-impact
polystyrene (figure 8.2. and 10.15.) discussed in paragraph 10.3.6.
Describe the main differences between their physical
characteristics?

1 urn
I 1

Figure 10.22. Transmission electron micrographs of a ruthenium tetroxide stained poly

propylene - ethylene-propylene) copolymer blend. The ethylene-propylene copoly-mer
forms the dispersed phase. The core-shell structure is typical of a block copolymer, (after
Biebuyck, 1994).
 Microstructures 259

What can be deduced from this microstructure for the solubility

and heat stability of this material.
Is the polypropylene - ethylene-propylene copolymer blend a duc
tile material? Justify your answer by referring to paragraph 10.3.6.

10.7. References and complementary reading

M.F. ASHBY, D.R.H. JONES, Engineering Materials 2, An Introduction to Microstructures, Processing
and Design, Pergamon , Oxford, 1986 and 2 ed., Butterword-Heinemann, Oxford, 1996.
nd

J. J. BIEBUYCK, Personal communication (1994).

R.W. DAVIDGE, The Structure of Special Ceramics with Particular Reference to Mechanical Properties,
Proc. Br. Ceram. Soc, 20 (1972) 364.
T.H. MAIMAN, Nature, Aug. 6, 1969).
J.-P. EBERHART, Analyse structurale et chimique des materiaux, Bordas, Paris, 1989.
F. A. Chadwick, Metallography of Phase Transformation, Butterword, London, 1972.
G.F.P., Initiation a la chimie et la physico-chimie macromoleculaire, Volume 1, Physico-Chimie des Po
lymeres, 1978; Volume 4. Quelques grands polymeres industriels, 1982.
H. JONES, Rapid Solidification of Metals and Alloys, Institution of Metallurgists, (Great Britain), 1982.
W. KURZ, D.J. FISHER, Fundamentals of Solidification, Trans. Techn. Publications, Zurich (Suisse),
1998.
J. L. MARTIN, A. GEORGE, Caracterisation experimentale des materiaux II, Traite des materiaux Vo
lume 3, Presses polytechniques et universitaires romandes, Lausanne, 1998.
G.E. MOLAU, S.L. AGGARWAL (ed.), Block Polymers, Plenum Press, New York, 1970, p. 79.
M. MORTON, Thermoplastic Elastomers, J. Polym. Sei, C60,(\977), 1.
T. SAITO, Toyota SAE Technical Series, 1992.
R. E. SMALLMAN, K. H. G. ASHBEE, Modern Metallography, Pergamon, Oxford, 1986.
 Chapter 11.

Tensile behaviour of materials

11.1. Objectives
To interpret tensile test results and describe the mechanical properties
of a material.
To understand the typical tensile curves for different materials.
To relate nominal and true stresses and strains.
To introduce the concept of strain energy.
In chapter 6., the elastic behaviour of ideal solids subjected to mechanical stress
was described. It was shown that the strength of materials was generally much lower
than that calculated theoretically by only considering the interatomic bond strengths
( 6.3.4.). This divergence between theory and practice is the result of defects (dislo
cations), which were described in chapter 7.
In this chapter, the tensile behaviour of real materials will be studied. In service,
materials are subjected to loads or forces which cause deformations (strain). It is
important to know how materials behave during deformation. The most important
mechanical properties are stiffness (elastic modulus), yield strength and ductility
measured by tensile testing and hardness determined by various types of test.
This chapter is limited to the macroscopic and phenomenological aspects of
mechanical properties. The microscopic aspects of deformation mechanisms and the
hardening processes, which influence the mechanical behaviour of materials, will be
presented in chapter 12.

11.2. Tensile properties

11.2.1. Strength of materials and mechanical properties
Distinction should be made between mechanical properties describing the spe
cific behaviour of materials, and what is called the strength of materials which analy
ses the behaviour of structural elements subjected to mechanical stresses: machine
parts, structural elements in civil engineering, etc.
The strength of materials is an engineering discipline using the specific proper
ties of materials, such as the modulus of elasticity (Young's modulus) or the yield
stress, to calculate the service stresses and strains of structural elements and assure
dimensional correctness. On the other hand, the study of mechanical properties,
considered in this chapter, aims to establish the intrinsic characteristics of materials
when they are deformed.

> 2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

262 Introduction to Materials Science

The application of an external force to a solid provokes initially an elastic de

formation. For a large number of materials (metals, certain polymers), this elastic
deformation, which is reversible, is followed by a permanent, irreversible plastic
deformation. There are two types of limiting stresses: the yield strength Re, giving
the stress in the material at the end of the region of elastic deformation, and the
ultimate tensile strength R , which is the value of the maximum load before rupture,
m

divided by the original area of the specimen. These two strength values are the same
in brittle materials such as ceramics and a significant number of organic polymers
which break without any prior plastic deformation,
Two other important characteristics are also measured in the tensile test: the
ductility of a material, which is the amount of permanent deformation at rupture and
the toughness, which is the energy absorbed by a material at rupture. Materials are
often required with a high elastic modulus and high toughness, i.e. ductile, but rigid
materials with a high yield strength. In practice it is difficult to obtain such a combi
nation of mechanical properties.
As already mentioned in chapter 6, organic polymers at ambient temperature,
and metals and ceramics at high temperatures can have viscoelastic behaviour. Here,
the instantaneous elastic reaction is followed by viscous flow leading to a significant
change in the mechanical properties as a function of the duration of the application of
the stress (creep). To be able to use viscoelastic materials sensibly, it is necessary to
use extrapolation techniques to determine the mechanical behaviour over long times.
This point will be discussed in chapter 12.
The mechanical properties of materials are measured by standard test proce
dures using defined test specimens subjected to specific loading conditions. A test
specimen of the material is made to specified dimensions. These standard tests make
it possible to compare data from different laboratories. Various service conditions
can also be simulated by appropriate test procedures (accelerated aging, wear, fa
tigue, etc.). However considerable care must be taken when translating the results of
laboratory tests to 'in service' performance where real conditions are often much
more complex.

11.2.2. The tensile test

The tensile test is the most common mechanical test. An increasing tensile force
is applied to a bar of standard dimensions until it ruptures after loading at a constant
deformation rate. A series of important mechanical characteristics can be measured
by recording the force applied by the tensile machine to the specimen and its pro
gressive elongation. In most instances, the variation of the cross-section of the
specimen during testing is not known and, as a general rule, the force F and elonga
tion / are measured relative to the initial dimensions of the specimen which gives
the nominal stress :
F (11.1.)
=
So
where So is the initial cross-section of the tensile specimen. In the same way, the
nominal deformation is defined as:
Al
(11.2.)
=
h
 Tensile behaviour of materials 263

where / corresponds to the initial length the specimen. The value of is generally,
0

given as a percentage. Plotting =() gives the tensile curve.

The tensile curve of a ductile metal is shown in figure 11.1. The following four
characteristic parameters can be determined from a tensile curve of this type.
The modulus of elasticity (or Young's modulus) is given by the slope
of the elastic part of the stress - strain curve. As mentioned in chapter 6,
this elastic modulus is a function of the bond energies between the
atoms or molecules making up the material.
The yield strength R giving the value of the stress beyond which the
e

material deforms plastically. As plastic deformation often appears

gradually, the yield strength is difficult to determine precisely, and
usually the 0.2 % offset yield R 0 2 I S used. A line parallel to the elastic
portion of the stress-strain curve and intersecting the strain ordinate at
0.2 % is drawn (figure 11.1) and the point at which this line intercepts
the stress-strain curve determines the 0.2 % offset yield Ro. .
2

The ultimate tensile strength R is defined as the maximum load

m

supported by the specimen divided by the original cross-sectional area.

The rupture elongation e measures the plastic deformation of the speci
R

men after rupture under tension. This is a measure of ductility.

The yield strength R is important since it defines the limiting stress which must
e

not be exceeded in order to avoid the permanent deformation of a part in service. A

large difference between the values of R and Re, together with a high value of R,
m

provides a good degree of security in case the yield stress is locally exceeded in a
loaded part.
For safety reasons machine components are designed so as to maintain stress at a
lower level than yield. The mechanical behaviour thus depends only on the modulus
of elasticity E. For example, when the top of the Eiffel Tower experiences
oscillations of the order of 50 cm amplitude in a strong wind, this deflection is not
catastrophic because it remains well within the elastic limit of the steel used in the
construction. However, even when the applied stress is lower than yield, prolonged
periodic deformations can lead to fracture due to the phenomenon offatigue, which
will be considered in chapter 13. In practice, the surface degradation of materials,
due to corrosion, often adds to the effect of static or cyclic stresses, thereby reducing
the strength.
During the thermodynamic study of uniaxial elongation (section 6.3.), it was
possible to differentiate between two major classes of material: materials with en-
thalpic elasticity (metals, organic and mineral glasses, semi-crystalline polymers) and
elastomers, which have entropic elasticity. The general characteristics of the tensile
curves of enthalpic elastic materials, by far the most numerous, are reviewed below
followed by a description of the behaviour of elastomers in tension.

11.2.3. Stress - Strain Curves of enthalpic elastic materials

When a ductile metal, such as mild steel, is subjected to tensile stresses, the
following effects are observed (figure 11.1.).
264 Introduction to Materials Science

Figure 11.1. Stress - Strain curve for a cylindrical bar of ductile metal subjected to
uniaxial elongation.

The initial deformation is elastic. When the stress is removed the specimen
regains its initial shape and dimensions with a total and almost immediate reversal of
the deformation. In metals, elastic deformation is normally linear. In materials with a
high modulus of elasticity (metals and ceramics), the elastic deformation is usually
less than 0.1 %. In this region, Poisson's Ratio (chapter 6.) is close to 0.3 for
metals.
If the yield stress is exceeded, the specimen length is increased permanently
after the stress is removed. This is due to an irreversible deformation (plastic
deformation).
When the material is subjected to loading and unloading in the region of plastic
deformation, the yield stress increases with successive deformations. This is
illustrated in figure 11.2. Two additional loadings of a 0.2 % carbon steel specimen
were carried out after the initial loading. The conventional 0.2 % offset yield Ro 2

( = 0.2 %) is of the order of 200 MPa. After an additional loading cycle, the
conventional 0.2 % offset y i e l d s 2 r e a c h e s 350 MPa. After another loading cycle, it
is greater than 400 MPa.
Work hardening generally occurs during the plastic deformation of metals and
is marked by an increase in hardness and yield stress. This is due to an increase of
dislocation density during plastic deformation, which tends to block dislocation
movement. This will be covered in more detail in the following chapter.
 Tensile behaviour of materials 265

(MPa)
500

0
0 2 3 (%)
Strain

Figure 11.2. Stress - Strain curve in nominal values for a steel specimen containing
0.2 % of carbon. Two additional loading cycles (1) and (2) were made after the initial
tensile test.

The phenomenon of work hardening has been mainly studied for metals but
similar phenomena occur in other materials such as ductile thermoplastics
(figure 11.3.) although different mechanisms are involved.
A tensile curve can show an inflection during the elastic-plastic transition at the
yield point. This is particularly so for soft steels (figure 11.2.), and results from weak
anchoring of the dislocations by the interstitial carbon atoms.
During the initial elastic deformation and for small amounts of plastic deforma
tion, the elongation of the specimen goes with a homogeneous contraction along the
whole sample length. Beyond the deformation corresponding to the maximum load
of the nominal stress - strain curve, the contraction of the section ceases to be homo
geneous and becomes more significant at some place in the specimen. At this point,
the localised cross-section of the specimen is reduced, i.e. necking (figure 11.1.)
occurs. The strength of the specimen, which is proportional to the cross-section, also
decreases. Therefore, necking appears when the elongation exceeds the maximum of
the tensile curve in figure 11.1. This value corresponds to the maximum load sup
ported by the test specimen deformed in a homogeneous manner.
The reduction in area A is the relative percentage decrease in cross-section
s

measured after rupture and is given by the following relationship:

=-^-00 (11.3.)
So

So is the initial cross-section, and S , the cross-section after rupture. The reduction in
r

area is also a measure of ductility: A varies from zero (brittle materials) up to nearby
s

100 as a function of ductility. When rupture occurs, the specimen suddenly contracts,
releasing stored elastic energy (figure 11.1.). It is accompanied by sound emission as
a result of the transformation of potential into kinetic energy.
Figure 11.3. shows the tensile curve of a ductile polymer, using moderate de
formation (strain) rates of (1 to 100 % m i n ) . In the first part, (points (1) and (2)),
_1
266 Introduction to Materials Science

100
Strain (%)

Figure 11.3. Nominal stress - strain curve for a ductile thermoplastic polymer. Up to (A),
a pseudo-elastic deformation can be observed allowing the apparent modulus of elastic
ity to be determined. (B) Formation of the neck and (C) plastic deformation with orien
tation of chains. The necking propagates across the whole sample and (D) sample break.
Curves of this type are encountered for some amorphous thermoplastics (PC) at < T g

and semi-crystalline thermoplastics ( - PP - PA 6-6).

this curve is similar to that of metals, although stresses are lower and strains greater.
The initial part of the tensile curve is almost linear, but it does not necessarily corre
spond to ideal elastic behaviour. As was mentioned in Chapter 6, polymers are usu
ally viscoelastic.
Necking occurs in ductile semi-crystalline polymers (polyamide, polyethylene,
etc.) or amorphous polymers (polycarbonate). The difference with respect to metals
is that the neck gradually propagates through the entire specimen (figure 11.3.).
Plastic deformation is accompanied by a significant increase in strength, as a result of
the orientation of chains in the stress direction. This is why necking propagates
throughout the sample and plastic deformation can reach up to 500 %. This
phenomenon of consolidation by plastic deformation is used in the manufacture of
textile fibres, which undergo a drawing step just after spinning.
The modulus of elasticity, determined from the slope of the tensile curve at low
strains, is an apparent modulus of elasticity depending on the strain rate. The ap
parent modulus may also be strongly influenced by temperature depending on the
proximity of the glass transition. For these materials, only tests carried out at the
same temperature, at the same strain rate and, for moisture sensitive polymers (poly
amide), at the same relative humidity, are comparable. The variation in mechanical
behaviour with temperature, specific to polymers, is illustrated in figure 11.4 which
shows the behaviour under tension of poly(methylmethacrylate) (PMMA). The tran
sition between brittle and ductile behaviour occurs around 40 - 50 C, i.e. about
60 C below T (-105 C). The slope at the origin gives the apparent modulus of
g

elasticity, which decreases considerably with temperature.

 Tensile behaviour of materials 267

20 30
Strain m % ~

Figure 11.4. Stress - strain curves for poly (methylmethacrylate) PMMA (organic glass)
as a function of temperature (after Andrew, 1968).

10%
200 1 1 1 "

(a)

HDPE
(

(b) g loo

10%

/AI O2 3

1 alloy
(c)
j/. Mineral glass

Organic glass
0,2%
Deformation

NR Vulcanised natural rubber

HDPE High density polyethylene
Organic glass Poly(methylmethacrylate)

Figure 11.5. Stress - strain curves for different types of materials: (a) large deformation
(max 100%); (b) moderate deformation (*-10%); (c) small deformation
>

( < 0,2 %).

268 Introduction to Materials Science

To compare the elastoplastic behaviour of different materials, it is often neces

sary to use different levels of strain ( = 0.2 %, 10 %, 500 %, (figure 11.5.)). The
max

plastic deformation of metals and alloys involves dislocation movement. In ceramics,

the movement of dislocations is severely impeded at ambient temperature by the high
strength of the bonds and by the presence of ions of opposing sign also making
atomic movement difficult. Ceramics, like mineral glasses, are mainly brittle mate
rials which break without yielding. A variety of behaviours is observed in polymers,
depending on the molecular structure and the temperature. Most amorphous glassy
polymers (polystyrene, poly(methylmethacrylate)) are brittle at < T . The semi- g

crystalline polymers (polyethylene, polypropylene) are generally ductile at tempera

tures between T and T . Below T the semi-crystalline polymers become brittle, for
g m g9

example, polypropylene, (T ~ - 1 5 C) or poly(ethylene terephthalate) (T ~ 70 C).

g g

11.2.4. Tensile Behaviour of R u b b e r

As mentioned in chapter 6., elastomers are a special class of materials. The
polymer chains are bound to each other by cross-links (figure 5.5.). The elastomers
are three-dimensional networks (figure 6.7) of elastic polymer segments. The thermal
motion of very mobile chain segments induces the retraction force. The elastic
retraction force is almost exclusively of entropic origin.
The tensile curve at high strain is given in figure 11.6. Since the initial por
tion of the stress-strain curve is not linear, the modulus of elasticity is calcula
ted from the slope at the origin. This is extremely low. This observation obvious when

(MPa)

Figure 11.6. Tensile stress-strain curve for vulcanised natural rubber (non-linear elas
ticity): (a) nominal stress-strain curve, (b) theoretical stress-strain curve (after Treloar,
1975).
 Tensile behaviour of materials 269

the tensile stress-strain curve for vulcanised natural rubber (NR) is compared with
other tensile curves (figure 11.5. (a)). Elastomers rupture at extremely high levels of
deformation ( ~ 700 %) without preliminary plastic deformation. Therefore,
paradoxically, they break in a brittle manner. Non-cross-linked elastomers are very
viscous viscoelastic liquids.
The glass transition temperature of elastomers T is close - 7 0 C. They are
g

always used at temperatures higher than T when cohesive forces between the chains
g

segments are extremely weak. These are able to move relative to each other without
modification of the interatomic distances and without appreciable variation of the
valence angles, i.e. practically without any variation in the internal energy. The
tensile curve for an elastomer in uniaxial compression and at low strain elongation
was shown in figure 6.13.
From the kinetic theory of rubbers, it has been possible to deduce the theoretical
curve of figure 11.6. (b) from expression (6.46). This expression has been verified by
experiment up to strains of the order of 60 % (figure 6.13.). As shown in figure 11.6.,
the difference between experimental results and theory is not significant up to about
500 %. The important discrepancy observed at very high strains ( > 500 %) result of
a rapid increase in the elastic force of retraction between the ends of the rubbery seg
ments (figure 11.7.). When the molecular chains are highly oriented, the tensile force
is progressively exerted on the covalent bonds of the chain. This considerably
increases the retractive force.

11.2.5. T r u e stress and strain

In general, the results of tensile tests are expressed as a function of nominal
rather than true stress values. The true stress ,, which is the ratio between the ap
plied force F and actual cross-section of the sample at a given moment, can be easily
calculated by using simplifying hypotheses. By assuming that the volume of the sam
ple remains constant as in the case of plastic deformation of metal alloys (v = 0,5),
the initial volume (SWo) and the volume after deformation (S-l) will be equal, from
which it follows that:

0 1
Extension ratio

Figure 11.7. Variation of the retractive force between the extremities of a chain as a
function of chain extension ///max. In inserts, variation of the shape of the chain (confor
mation) as a function of extension.
270 Introduction to Materials Science

So/S=I/l =l+
0 (11.4.)

The true stress Gt = F/S is related to the nominal stress = F/S by: 0

/ = (1 + e ) (11.5.)

In the same way, the true strain , is expressed by considering the infinitesimal
increase of the specimen from / to / + al. The real increase is therefore equal to dl/l
and not to d/// and the total true strain e is given by:
0 t

, d / (H.6.)

, = 1^ = 1(/// )
0

The true strain , is related to nominal strain by:

e =1(1 + )
t (11.7.)

In these calculations, strains are expressed as fractions and not as percentages.

Table 11.8 compares nominal and true strains in uniaxial tension and compression.
The deviation between nominal and true strains is minimal below a strain of 0.1 or
10 %. Above this value, the deviation becomes more important.

Table 11.8. Comparison between nominal and true values s of strains in uniaxial
tension and compression.

Compression Tension

-2.0 -1.0 -0.5 -0.2 -0.1 -0.01 0.01 0.1 0.2 0.5 1.0 2.0
-0.86 -0.63 -0.39 -0.18 -0.095 -0.01 0.01 0.11 0.22 0.65 1.72 6.4

Figure 11.9. shows the tensile curve for polycrystalline ductile copper expressed
as nominal and true values of stress and strain. The maximum in the nominal stress-
strain curve does not correspond to the maximum value of the intrinsic strength of the

Figure 11.9. Nominal and true stress - strain curves for polycrystalline copper. The true
curve , - , increases gradually to rupture while the true curve - passes through a
maximum value due to necking.
 Tensile behaviour of materials 271

elastic material. The intrinsic strength of the material increases continuously until
rupture. The maximum of the tensile curve for the nominal stress is only the result of
necking which decreases the cross-section of the specimen. The tensile force required
to deform the specimen is therefore lower in spite of the increase in strength of the
material by work hardening.
In reality, the value R (nominal) is a measure of the strength of the specimen
m

and not that of the material. The deviation between the nominal and real strain
curves, can also be seen in the curves for uniaxial compression. The processes of
plastic deformation do not depend on the loading direction, and there should be no
difference between the tensile and compression curves. When nominal stresses and
strains are used, there is a marked difference (figure 11.10. (a)). This is because
compression causes a lateral dilation which increases the cross-section of the sample
and, consequently, the force required for compression increases. In true terms, the
tensile and compression curves are identical in absolute values (figure 11.10. (b)).

(a) (b)

Figure 11.10 Stress - strain curves in uniaxial extension and compression for a ductile
metal. Comparison between nominal (a) and true (b) curves (after Ashby, Jones, 1980).

11.2.6. Deformation energy and anelastic behaviour

The deformation energy of a material can be deduced from the tensile curve.
Ductile materials, which break after plastic deformation, have a rupture deformation
energy which is much higher than that of brittle materials. The volume deformation
energy U is given by the area under the tensile curve (figure 11.11.):
e

e
= \ (11.8.)
0

Figure 11.11 shows the difference between elastic deformation energy U and e

plastic deformation energy U . p

As the tensile curve for most materials is linear in the elastic region, the volume
elastic deformation energy U is equal to the area of the triangle (0 - 1 - \). Since
e

= , the volume elastic deformation energy U cm be calculated:

e

U)=0,5E{e\) 2
( U 9 )

After plastic deformation, the elastic deformation energy can be calculated in a

similar way from the triangle (2 - - }). U} is considerably larger than U}.

272 Introduction to Materials Science

(2)

) 4
ill
Strain e

Figure 11.11. Calculation of the volume elastic (U ) and plastic (U ) deformation energies
e p

from the stress - strain curve for a ductile material at two degrees of strain (1 and 2).

For a perfectly elastic material, the tensile curve is reversible and the energy
absorbed on loading is completely recovered when the load is removed (spring be
haviour). Most materials have do not have an ideal elastic behaviour, and a part of
the absorbed energy is dissipated within the specimen by internal friction mecha
nisms (movement of dislocations in metals or molecular chains in polymers). In this
case, there is a difference between energy supplied to the system and the energy
recovered. The unloading curve is no longer equivalent to the loading curve, even
though the specimen returns to its initial length. During a deformation cycle, a hys
teresis loop is formed. The loop surface represents the amount of energy dissipated
as heat on loading and unloading (figure 11.12.). This anelastic effect, is a form of
viscoelastic behaviour.

Figure 11.12. Anelastic deformation of a grey cast iron with formation of a hysteresis
loop after loading and unloading. The energy dissipated during a deformation cycle is
proportional to the black area. Two loading cycles were made. In the second case, the
elastic limit was exceeded and consolidation of the material by work hardening occurred,
with some permanent deformation.
 Tensile behaviour of materials 273

In most materials, the anelastic effect is weak but sometimes this effect can be
relatively important. In grey cast iron, a significant anelastic effect is observed due to
the movement of dislocations in the graphite lamellae in the iron matrix. Elastomers
also form hysteresis loops. In this case, energy dissipation result from friction of the
elastomer chains on one another during dynamic deformations. It is this anelastic
effect, which produces heat-build-up in tyres during flexing. Because of this be
haviour, grey cast irons and elastomers are excellent dampers of mechanical vibra
tions and sound.

11.2.7. Hardness measurements

Tensile tests are often performed together with hardness measurements. These
measure the resistance to localised deformation. The methods of measuring hardness
(Brinell hardness, Vickers hardness, etc.) are based on the penetration of a very hard
indenting device, a steel ball or a diamond, into the surface of the material. A hard
ness value is obtained from the dimensions or the depth of the indentation made by
the indenting device under given loading conditions. The hardness measurement is
one of the commonest mechanical tests because it is quick, non-destructive and only
requires a simple apparatus. It can usually be related empirically to the tensile
strength. However, it is a complex test and not easy to interpret.

11.3. Summary and conclusions

The modulus of elasticity E, the yield strength R , the ultimate tensile strength
e

R and the rupture elongation #, are important mechanical properties of materials.

m

The values of these parameters are generally obtained from the nominal stress - strain
curve.
Ideally the true stress - strain curve should be taken but, in practice, this is diffi
cult because of necking. However, the most important mechanical properties, the
modulus and the yield strength R (or the ultimate tensile strength R in the case of
e m

brittle materials) are practically independent of this choice.

The thermodynamic study of the uniaxial elongation of materials (chapter 6.)
demonstrates that they can be divided into two groups: materials with enthalpic elas
ticity, constituting the majority of materials (metals, ceramics, organic and mineral
glasses etc.), and the elastomers, which are materials with entropic elasticity. The
analysis of the tensile curves gives the most important mechanical properties of these
two main categories of materials.
In materials with enthalpic elasticity, there are fragile materials breaking with
little or no plastic deformation, and ductile materials, which show a more or less
important degree of plastic deformation before rupture. Ceramics, a large number of
glassy thermoplastics, as well as thermoset polymers and elastomers, are brittle mate
rials. Metals and some amorphous and semi-crystalline thermoplastics have a ductile
behaviour.
Isotropic polymers have much lower mechanical strength than metals or ceram
ics. They can be significantly oriented and stretched, and then show a significant
improvement in properties. This effect will be considered in more detail in Chapter
12. This capacity for plastic deformation, which is shown by some thermoplastic
274 Introduction to Materials Science

polymers should not be confused with the high degree of elastic defor mobility of the
elastomers which break without any plastic deformation in a brittle manner. The
deformation energy of materials is measured by the area under the stress-strain curve.

11.4. Illustrative example: cables for cable cars

The movement of a modern, high-capacity cable car is by a wheel-train pulled
by a traction cable, moving on two carrying cables (figure 11.13.).
Generally, a cable is manufactured by the cabling of twisted or parallel metal
wires. This spiroid construction is known as 'stranded'. Several strands, with one or
more layers of wires, can be cabled around a central core of natural or synthetic fi
bres. This cabling technique plays a key role in determining the properties of the
cable.

Figure 11.13. Wheel-train of a cable car resting on two carrying cables, pulled by a trac
tion cable (photo Von Roll).

The traction cable of a cable car system must be flexible because it has to be
wound on to the driving drums. This requires that the strands are twisted in the same
sense. The sense and the pitch of the cabling of the various layers of wires are chosen
in such a way that the resultant torsion acting on the whole cable is very weak. The
carrying cable has a smooth surface which facilitates movement of the wheel train.
Carrying cables with a closed construction are produced with profiled wires (e.g.
shaped cross-section) for the outer layer (figure 11.14.). The cabling of the outer
layer of the profiled wires is produced by assembling the wires like a zip. Before the
intrinsic properties of these metallic wires are defined, the operating conditions must
be considered and the stresses acting on the cables calculated.
 Tensile behaviour of materials 275

Figure 11.14. Section of a carrying cable with, respectively, one layer and one and a half
external layers of profiled wires (after Huber 1980).

Table 11.15. summarises the main characteristics of the cable car system at the
Klein Matterhorn, Zermatt, Switzerland. It can be seen that the actual weight of the
cable is important. For this installation, the maximum free length of the two carrying
cables is 2885 m corresponding to 66 tons weight which is more than three times the
weight of a cabin transporting 100 passengers. Any improvement in the tensile
strength of the cable wires would be beneficial because the cable weight could be
reduced.

Table 11.15. Characteristics of the cable car Steg-Klein Matterhorn at Zermatt.

Overall length 3835 m

Difference in altitude 891 m
Average slope 25%
Maximum slope 90.6 %
Number of intermediate pillars 3
Maximum distance between two pillars 2885 m
Transport 2 cabins
Number of people per cabin 100
Speed 10 ms"1

Duration of journey 500 s

Carrying cables per cabin 2
diameter 45.2 mm
weight per unit length 11.47 kg/m
rupture force 219000 kg
Traction cable per cabin 1
diameter 40 mm
weight per unit length 6.21 kg/m
rupture force 102000 kg

The wires are made of carbon steel with an ultimate tensile strength R of be m

tween 1700 and 2200 MPa and an elongation at rupture e of 2.5 to 5 %. The safety
R

factor of the carrying cable is greater than 3.5. This factor means that the maximum
stresses experienced in service are 3.5 times lower than the ultimate tensile strength
of the cable. This is the lowest safety factor used in this type of transport. The trac
tion cable of cable car systems and the carrying cables of chairlifts have a safety
factor of 5 while a elevator cable has a safety factor of 12.
The average lifetime of a carrying cable is 17 years and not only depends on the
intrinsic strength of the steel wire and the cable construction, but also on the mainte-
276 Introduction to Materials Science

nance. Generally, it is not rupture occurring inside cables due to fatigue or excessive
loads which limit the lifetime, but internal damage resulting from corrosion. That is
why the cable is greased during manufacture and this is regularly renewed every
year.
The carbon steels used in the manufacture of cables are of eutectoid composition
(0.8 % C). The high mechanical performance is mainly a result of the heat- treatment,
which refines the perlitic structure (chapter 9., figure 9.16.). To obtain significant
hardening, the wires are quenched and transformed in an isothermal bath (at about
500 C) to form very fine perlite. The wires are then drawn, i.e. worked to further
refine the microstructure and to introduce a large number of dislocations. For many
years, this process called patenting has enabled relatively cheap high-strength steel
wires to be produced with a tensile strength of about 2000 MPa (200 kg m m ) , -2

which are used for cable car system cables and piano wires.

11.5. Exercises
11.5.1. What are the deformation and elongation of a steel wire, of 2.5 mm diameter
and 3 m length, supporting a mass of 500 kg, given that the modulus of elasticity
of the steel is 210 GPa?
11.5.2. A bar of initial length L is stretched in uniaxial plastic elongation to length
0

L\= 2 L . This bar is then stretched by a second deformation to length L = 3 L .

0 2 0

Calculate the nominal and true elongation after each stage.

11.5.3. Calculate the elastic energy U stored in an aluminium alloy wire with a
e

modulus of elasticity = 70 Gpa, after an elastic deformation of 0.01 %.

11.5.4. An aluminium alloy specimen has a tensile strength R m = 300 MPa and a
reduction in area, A of 77 %. Calculate the true tensile stress a .
s r

11.5.5. A steel bar 10 cm in diameter is subjected to an alternating axial load of

500 kN. Knowing that the modulus of elasticity is 200 GPa and Poisson's ratio
is 0.3, calculate the maximum and minimum diameters of the bar during service.
11.5.6. Which type of polymers could be proposed to obtain:
greater tensile strength,
increased rigidity,
greater ductility?
In each case, explain the differences in behaviour of these various materials.
11.5.7. Is it possible to plastically deform in compression, an aluminium bar, 50 mm
in diameter with a yield strength, Re of 150 MPa, using a press with a maximum
capacity of 50 tons?
11.5.8. A standard tensile test is carried out on a copper-nickel alloy sample of initial
diameter 12.5 mm and initial length 50 mm. Using the values in table 11.16, draw the
initial part of the stress a - strain curve and calculate:
the modulus of elasticity E\
the 0.2 % offset yield stress RQX,
the ultimate tensile strength R and the rupture elongation
m R
 Tensile behaviour of materials 277

Table 11.16 Data of the tensile test

Load Elongation
[kN] [mm]

5 0.015
15 0.045
26 0.500
35 1.300
48.5 (a) 7.000
39.5 (b) 18,700

(a) maximum load; (b) Breaking load.

11.5.9. A magnesium alloy wire of 1 mm diameter has a modulus of elasticity of

45 GPa. Plastic deformation occurs when the load reaches 10 kg. At a load of 12 kg,
the total deformation of the thread is 1 % . Calculate the permanent deformation of
the wire after this loading.

11.5.10. The tensile curve of a material is generally recorded at low strain rates, e.g.
ael dt= 10 s" . What difference would be observed in the modulus of elasticity if
3 1

the tensile test were carried out at a strain rate 100 times higher, for:
a metal?
a polymer?

Explain your answers.

11.6. References and complementary readings

E.H. ANDREWS, Fracture in Polymers, Oliver and Boyd, London, 1968.
M.F. ASHBY, D.R.H. JONES, Engineering Materials 2, An Introduction to Microstructures, Processing
and Design, Pergamon, Oxford, 1986 and 2 ed., Butterword-Heinemann, Oxford, 1996.
n d

. F. COUTNEY, Mechanical Behavior ofMaterials, McGraw-Hill, New York, 1990.

R.W. DAVIDGE, Mechanical Behaviour of Ceramics, Cambridge University Press, Cambridge, 1979.
R.W. HERTZBERG, Deformation and Fracture, Mechanics of Engineering Materials, 4th ed., John
Wiley, New York, 1996.
K.J. PASCOE, An Introduction to the Properties of Engineering Materials, 3rd ed., Van Nostrand-Rein-
hold, Wokingham, Berkshire, U.K., 1978.
D. ROSENTHAL, R.M. ASIMOW, Introduction to Properties of Materials, 2nd ed., Van Nostrand-Re in
hold, Wokingham, Berkshire, U.K., 1971.
J. SCHULZ, Polymer Materials Science, Prentice-Hall, Englewood Cliffs, New Jersey, 1974.
E. HUBER, Experiences et Progres lors de la Fabrication de Cables Porteurs en Construction Close, Int.
Seilbahn, Rundschau 7, 303 (1980).
 Chapter 12.

Factors influencing
mechanical properties

12.1. Objectives
To analyse the most important factors influencing the mechanical pro
perties.
To describe the mechanism of plastic deformation at low temperature.
To study the main methods of metals hardening.
To analyse the creep behaviour of metallic materials.
To describe the viscoelastic behaviour of polymers.

Various factors influence the mechanical properties of materials, two of the most
important being temperature and microstructure.
The concept of low or high temperature plays an important role in materials
science. However, this notion is not directly related with the intuitive concept of high
or of low temperature found in daily life. This perception, resulting mainly from our
personal physiological experience, is closely linked to the freezing and boiling tem
perature of water. For materials, the notion of high or low temperature is a concept
related to the melting temperature T (and to the glass temperature T for amorphous
m g

materials) or the rate of reorganisation of atoms induced by the thermal agitation.

The two main types of deformation, which exist for materials at low tempera
tures, i.e. elastic deformation and plastic deformation, were covered in the previous
chapter. Elastic deformation characterises the behaviour of materials in normal con
ditions of use. However, during forming, yield strength is always exceeded. The
capacity for plastic deformation, called ductility, plays an important role in metals
forming. This property is also involved in the design of complex structures, such as
car bodies, which have frequently to absorb impact energy.
At ambient temperature, the capacity for plastic deformation is:
practically nil for mineral glasses, ceramics, highly cross-linked poly
mers (thermosets) and elastomers;
of the order of several percent for hardened alloys and several tens of
percent for pure metals;
of several hundred percent for ductile thermoplastic polymers; thermo
plastics are the only polymer materials undergoing plastic deformation.

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

280 Introduction to Materials Science

Ductitity increases significantly at high temperature. Advantage is taken of this

characteristic to shape steel bars or thermoplastic sheets (thermoforming) at high
temperature. Generally a temperature can be considered to be high when > 0,5 T m

(T in case of amorphous materials) (T, T , T are given in Kelvin). At high tempera

g m g

ture, plastic deformation is observed practically for all materials, with the exception
of the cross-linked thermoset polymers and elastomers, which cannot undergo plastic
deformation because of their network structure (chapter 5.).
Significant mechanical deformations can be induced at high temperature by low
stresses resulting from a time-dependent deformation. This viscoelastic behaviour
defined in chapter 6 is intermediate between that of the elastic solid and of the vis
cous liquid. It is encountered with many thermoplastics at temperatures close to or
above T . The creep of metals is another example of viscoelastic behaviour. In gen
g

eral, this only occurs at high temperatures (> 0.5 T ). At present, this effect limits the
m

working temperature and efficiency of gas turbines described in the illustrative

example and mostly used in aviation.
The main objective of this chapter is to analyse the mechanisms governing the
mechanical strength of materials. This chapter will not explicitly study ceramics,
which fail in a brittle manner. Their strength is strongly influenced by the presence of
cracks or pores. The influence of these defects on rupture will be considered in
chapter 13.

12.2. Concepts of low and high temperatures

At low temperature, the movements of atoms by diffusion are infinitely slow.
By contrast, at high temperature, diffusion takes place rapidly and plastic deforma
tion is favoured. At high temperature, mechanical properties often change with the
duration of application of stress.
The notion of high or of low temperature varies from material to material. At
150 C, steel or industrial porcelain are at 'low temperature' because diffusion
movements are very limited while at this temperature, most polymers are already at
'high temperature' because they are above T and diffusion movements are impor
g

tant.
At low temperature, the strain depends only on the stress <x
(12.1.)

At high temperatures, the strain depends not only on the stress , but also on
the temperature and the time t. The material becomes then viscoelastic (chapter 6.):
=Acx, /, T) (12.2.)

The temperature, at which diffusion becomes important, varies considerably for

the different materials (table 12.1.). To a first approximation, a value ~ 0.5 T is m

considered as the temperature for the initiation of diffusive motions. In metals, at

temperatures around 0.3 to 0.4 T , diffusion become important. For ceramics, mis
m

varies between 0.4 and 0.5 T , while it is between 0.5 and 0.6 T for polymers. The
m m

distinction between low and high temperatures depends also on the stress and on the
duration of its application. For amorphous materials, T should be considered.
g
 Factors influencing mechanical properties 281

Table 12.1. Approximate limits between low and high temperature for various materials.

Metals Ceramics Polymers

Materials 0,4 T 0,5 T 0,6 T

h m m m

K(C) (o K(C)

W 3680 1472(1199)
SiC 3110 1555 (1282)
MgO 3073 1537(1264)
Mo 2880 1152 (879)
A1 0 2 3
2323 1162 (889)
Si N 3 4
2173 1087 (814)
Ti 1943 777(504)
Fe 1809 724 (451)
Ni 1726 690 (417)
Cu 1356 542 (269)
Al 933 373 (100)
aromatic PA: Kevlar 00
913 648 548 (275)
PEEK (a)
613 423 368 (95)
Pb 600 240 (-33)
p (a)
C 573 423 344 (71)
PETP (a)
543 343 326 (53)
/M-PS (a)
513 373 308 (35)
zstf-PP (a)
443 258 266(-7)
PE (HD) (a)
413 153 248(-25)
Hg 235 94(-179)

(a) The symbols are given in the appendix (chapter 18.).

12.3. Low temperature plastic deformation of materials

12.3.1. Comparison of the plastic deformation in metals and ductile thermoplas
tic polymers
It is interesting to compare the plastic deformation of metals with that of ductile
polymers. As already mentioned, the tensile curves show certain similarities in be
haviour, although the implied mechanisms differ appreciably and metals and their
alloys are much stiffer and stronger. In this analysis, the viscoelastic effect present in
thermoplastics will be neglected; this is, in any case, negligible if the temperature is
sufficiently low (table 12.1.).
On the - curve of figure 12.2., it may be seen that the tensile curves of metals
and that of ductile polymers exhibit qualitatively similar characteristics up to point 3.
Between points 1 (yield point) and 2 (ultimate tensile strength), deformation is
homogeneous. Beyond point 2, necking occurs in the specimen. The failure of metals
happens just after the formation of the neck (point 3). Metals harden during plastic
deformation. Dislocations generated by this process become more and more
numerous and closer to each other, causing hardening. In the region between points 2
and 3, the nominal stress (retraction force) decreases. The increase in strength of the
material resulting from work hardening does not compensate completely for the
decrease in the cross-section of the specimen caused by the instability of the plastic
deformation leading to the failure of the specimen at 3.
282 Introduction to Materials Science

Nominal Strain e

Figure 12.2. Comparison between the tensile curves of metals and thermoplastic poly
mers at low temperature. The level of stress is about 10 times higher in the case of
metals. Points (1) and (2) indicate respectively the yield stress and the ultimate tensile
strength. The failure of metals occurs at point (3), while that of the thermoplastic poly
mers at point (5).

For thermoplastic polymers, a decrease in the nominal stress can also be ob

served as the neck forms, but the increase in strength of the material is more impor
tant in this case since the chains orient themselves in the neck along the tensile di
rection. The chain orientation stabilises necking, which progresses through the whole
sample because of the lower strength of the unstreched part of the specimen. Be
tween points 4 and 5, a slight hardening of the deformed specimen is also observed,
which breaks at 5.

12.3.2. Traction and shear

Low temperature plastic deformation of crystalline materials ( 7.3.2) occurs by
successive movement of dislocations on well-defined atomic planes under the action
of shear (slip) forces. By applying a tensile stress ato a crystal, a proportional shear
stress is induced (figure 12.3.). Plastic deformation occurs when the shear stress Tis
greater than the resistance of dislocations to glide in their slip plane. The yield
strength is then exceeded. The shear stress at which plastic deformation starts is
called shear yield stress .

When a monocrystal is subjected to tension, it is the shear stress induced tan-

gentially to the slip plane, which must be considered. The shear stress induced by a
tensile force F in a given plane .S of a monocrystal is given by (figure 12.4.):
c

T = FJS C (12.3.)

where F is the tangential component of the tensile force in the slip plane.
C

If represents the angle between the slip direction d and the tensile axis, and
the angle between the normal to the slip plane and the tensile direction, it can be
written:
F = Fcos/l
C

and
 Factors influencing mechanical properties 283

Slip plane /

I
Figure 12.3. Deformation by uniaxial tension of an ideal crystal by shear along a slip
plane separating two planes with compact stacking of atoms. The shear at 45 to the
tensile axis equals half of the tensile stress: = /2.

Figure 12.4. Tangential component of the force (F ) in the slip plane due to the tensile
c

force F. Vectors d and represent respectively the slip direction and the normal to the
slip plane. So is the cross-section normal to the axis of the monocrystalline cylinder and
S the sheared section.
c

(12.3.) becomes:

= acos^cosA (12.4)

Equation (12.4.) is known as Schmid law and (cos0cosA) is Schmid factor. This
relationship makes it possible to relate a a n d for a given slip system.
284 Introduction to Materials Science

When the normal to the slip plane, the tensile axis and the slip direction are
coplanar, becomes equal to /2 -, and (12.4) can be written as:
= acosAsin/l (12.5.)

Angle

Figure 12.5. The shear stress in a material passes through a maximum at 45 to the
tensile axis ( = constant). In this figure n, d and the tensile axis are coplanar (after
Ashby and Jones, 1980).

The function (12.5) is represented in figure 12.5. To induce a slip (plastic de

formation) in a defined plane, it is necessary to increase the tensile stress until the
shear stress induced in this plane reaches its critical value . The tensile yield

stress then becomes:

R=
e - (12.6.)
cos cos
The maximum value of the shear stress is reached when Schmid factor is equal to
0.5:
=0.5 (12.7.)

For all other orientations of the slip system relative to the tensile axis:
0<<0.5 (12.8.)

As shown in figure 12.6., slip occurs preferentially in the slip plane where the
induced stress Tis greatest, i.e. in the slip plane where Schmid factor is maximum.
Plastic deformation of crystals occurs preferentially in the reticular planes with
high densities of atoms and in the directions where the reticular density is the highest.
These planes with high densities of atoms and the preferential slip directions
constitute slip systems (table 12.7.). The greater the number of slip systems present
in a crystal, the easier is the plastic deformation.
Metals with cubic crystal structure possess twelve slip systems with high
densities of atoms. In fee crystals, there are four different {111} planes with three
 Factors influencing mechanical properties 285

Strain e Schmid' Factor (cos cos )

0 < 4 5 0>45-*
(a) (b)
Figure 12.6. Plastic deformation of anthracene monocrystals: (a) curve for the variation
of the axial stress for monocrystals with different orientations; (b) tensile yield
strength R for a series of monocrystals given as a function of Schmid factor (after
e

Robinson and Scott, 1967).

Table 12.7. Shear yield stresses and slip systems for monocrystals of pure metals
(after Reed Hill, 1973).

Metal Structure Purity Slip plane Slip direction G

[GPa] [MPa]

Al fee 99.99 {111} (.To) 30 1,0

Cu fee 99.999 {111} (10) 50 0,6
Fe bcc 99.96 {011} () 85 28
Mg he 99.996 (0001) (ll20) 20 0,44
Zn he 99.999 (0001) (mo) 45 0,18

independent <> directions in each plane, bcc crystals deform generally according
to six type {011} planes and two < 111> directions. Metals with hexagonal crystalline
structure (Ti, Mg, Be, Zn) have only three slip systems: a (0001) plane with
three < 1120> directions. Metals with cubic structure thus have a greater capacity for
deformation (ductility) than metals with hexagonal structure. Although the bcc
crystal possesses the same number of slip systems as the fee crystal, it is less ductile.
This is due in part to the reduced symmetry and density of atoms in the {011} planes
of the bcc crystal compared with the density of atoms in the {111} planes of the fee
crystal.
286 Introduction to Materials Science

12.3.3. Mechanisms for the plastic deformation of metals at low t e m p e r a t u r e

There are several mechanisms for the deformation of crystalline materials. The
movement of atoms of the crystal during elastic deformation is shown in figure
12.8. (a). In this figure, two mechanisms for plastic deformation are also given: slip
(figure 12.8. (b)) and twinning (figure 12.8. (c)). Plastic deformation by twinning is
less common than that taking place by slip.
Twinning occurs when a part of the crystal tilts under the effect of a shear stress
into a symmetrical position without changing the structure of the lattice. This tilting
leads to the formation of a twin boundary between the deformed part and non-de
formed part ( 7.4.2). Dislocations are also involved in twin formation. Twinning
plays an important role in the plastic deformation of hep metals where the number of
slip systems is limited. A process similar to twinning is also observed in diffusionless
phase transitions such as the martensitic transition.
Here only slip deformation will be considered in detail, as it is the most impor
tant mechanism of low temperature plastic deformation. Identical values of plastic
deformation can be obtained by slippage in a large number of slip planes (homoge
neous plastic deformation) or by involving only a limited number of slip planes (in-
homogeneous deformation). In the first case, a number of narrow steps form at the
grain surface (slip bands). Wide steps are formed in the second case with large dis
continuities formed on the grains surface. As illustrated in figure 12.9., plastic
deformation is practically always inhomogeneous on the microscopic scale.

Figure 12.8. Deformation mechanisms: (a) elastic deformation; (b) plastic deformation
by slip (b) and (c) by twinning (tilting of a part of the lattice in a symmetrical position).

20 nm

200 nm

Figure 12.9. Localisation of plastic deformation in a limited number of parallel planes

(from Eisenstadt, 1970).
 Factors influencing mechanical properties 287

A number of simple relations exist between the dislocation movement and the
macroscopic plastic deformation. If there is a density p of parallel mobile edge dis
m

locations each covering an average distance , then a shear deformation will be

produced equal to:
Y=p bAx
m (12.9.)

where b is the absolute value of the Burger vector. By differentiating equation (12.9.)
with respect to time, the expression for the deformation rate is obtained:
= /dt = bd(p Ax)/dt
m (12.10.)

When the density of dislocations remains constant, this equation simplifies. By intro
ducing the concept of dislocation velocity = dAx/dt, then:

f=p bv
m (12.11.)

The dislocation velocity is a function of the shear stress v.

v = ^J (12.12.)

In this expression and C are constants which depend on the material. Finally, the
following expression is obtained for the rate of plastic shear:

(12.13.)

where varies strongly according to the nature of the material. For materials with
polar covalent bonds and for some bcc metals, is generally less than 20. For fee
metals, can reach values between 100 and 200. The maximum rate of dislocation
movement corresponds to the speed of sound (propagation of elastic waves). This
shows that the effect of stress on the dislocations and on the shear rate is very im
portant. A very small increase in stress can lead to a very significant increase in the
rate of dislocation movement and of plastic deformation.

12.3.4. Plastic deformation mechanisms in thermoplastics

The plastic deformation of polymer materials involves the deformation, move
ment and reorientation of different structural elements: chain segments and crystal
line lamellae.
The great majority of semi-crystalline polymers crystallise as spherulites, which
correspond to grains in metals and ceramics. As seen in figure 10.4., polymer spheru
lites are not monocrystals, but are made up of crystalline lamellae interlinked by
amorphous zones.
The mechanisms occurring during the plastic deformation of semi-crystalline
polymers have been analysed in detail by X-ray diffraction, as well as by optical and
electronic microscopy. In particular, the case of uniaxially deformed polyethylene
has been studied. This treatment will be limited to the analysis of modifications oc
curring in three adjacent crystalline lamellae of a spherulite, interlinked by an amor
phous zone (figure 12.10. (a)).
288 Introduction to Materials Science

Stretching direction

Figure 12.10. Modelling of the plastic deformation of a semi-crystalline polymer: (a)

three crystalline lamellae of a spherulite interlinked by amorphous chains before defor
mation; (b) chain slip in the crystalline lamellae; (c) fragmentation of the crystalline la
mellae into crystal blocks; (d) alignment of the crystal blocks and the amorphous chains.

In the first stage of deformation, modifications take place almost exclusively in

the amorphous zones (rubbery for PE), which are stretched. In the second stage
(figure 12.10. (b)), some slip parallel to the axis of the chains appears in the crystal
line lamellae induced by the highly stretched amorphous chains. In stage three
(figure 12.10. (c)), fragmentation of the lamellae occurs under the effect of shear
stress. This fragmentation of the lamellae is accompanied by the uncoiling of chain
segments initially included in the lamellae. In the fourth stage, the crystalline blocks
formed by fragmentation and the highly stretched amorphous chains align parallel to
the tensile axis (figure. 12.10. (d)).
During the deformation, the spherulites gradually deform and eventually disap
pear and, a fibre is progressively formed having markedly superior tensile strength to
that of the undeformed polymer.
The deformation mechanism of lamellae under shear stress acting on specific
planes is equivalent to that occurring in the plastic deformation of metals. As for
metals, the formation of shear bands can be observed during the deformation of semi-
crystalline polymers.
Plastic deformation by shear also takes place in glassy thermoplastic polymers.
In this case, the process is much less well defined than for crystalline materials. The
location of the deformation varies. In some cases, plastic deformation by shear oc
curs in the entire deformed region. In the other cases, the deformation is much more
localised, and shear bands also form as with semi-crystalline polymers and metallic
materials.
As well as plastic deformation by shear, polymers can also show plastic defor
mation by crazing (figure 12.11.). This mechanism occurs mainly in the glassy ther
moplastics. The crazes (micro-cracks) appear as small crevices starting on the surface
 Factors influencing mechanical properties 289

(a)

Figure 12.11. Crazes formation in an impact polystyrene plastically deformed (a) and
(b), (c), (d) and (e) evolution of the microstructure of a craze as a function of the degree
of opening of the crazes (after Hull, 1973).

of the specimens, and grow in a plane perpendicular to the tensile axis. In homoge
neous systems, these crazes tend to gradually occupy all the section of the specimen.
The crazes are regions where the polymer is stretched, in the form of microfibrilles of
extended chains, parallel to the tensile axis with pores between the microfibrilles. In
this type of plastic deformation, there is cavitation. When crazes occupy the whole
section of the specimen, the microfibrilles support all the tensile load.
In the homogeneous glassy polymers, crazes can lead to larger cracks and ini
tiate specimen failure. Microcracks have a thickness of the order of 100 nm, and due
to their heterogeneous structure, they diffuse light and give an iridescent appearance
to the sample in oblique illumination.
The two main processes, which occur during the plastic deformation of organic
polymers, are plastic deformation under the action of shear stresses and plastic de
formation induced by tensile stresses. With tensile deformation, which occurs espe
cially in the glassy thermoplastics, plastic deformation is very localised and forms
crazes oriented perpendicularly to the tensile axis. These two mechanisms of plastic
deformation are involved to different extents in the process of energy absorption
during the rupture of parts produced from organic polymers.

12.4. Yield strength and hardening mechanisms of

crystalline solids
12.4.1. Materials and the yield strength
The yield strength R (figure 12.12.) is one of most important properties used to
e

determine the dimensions of components. The yield strength marks the limit of forces
permitted without leading to a permanent change in shape and dimensions. A high
290 Introduction to Materials Science

yield strength allows elements with reduced cross sections, and hence of lighter
equipment to be constructed. However, there are applications which require a low
yield strength R . This is notably the case, for example, for sheet steels, which un
e

dergo significant plastic deformation during the forming.

Pure metals have a low yield strength Re but a large capacity for plastic defor
mation. Plastic deformation of metal leads to hardening and increases its yield
strength (figure 11.2.). Other mechanisms also increase the yield strength Re of met
als such as the presence of precipitates. It is therefore possible to obtain different
yield strengths with the same metal. For example, with carbon steels (Fe-C alloys),
the yield strength R varies from 150 to more than 4000 MPa. Figure 12.12. gives the
e

main values for the yield strength or the ultimate tensile strength R for metals, ce m

ramics and some polymers. Ceramics have the highest mechanical strength. Some
oriented polymers (fibres of polyamide (Kevlar) and very high molecular weight
polyethylene) reach a much higher ultimate tensile strength (from - 7 0 0 to - 3 5 0 0
MPa) than that of the non-oriented polymers considered here, but these stretched
polymers have a much lower temperature resistance than that of the metal alloys.
Metals Ceramics Polymers
*

Diamond
Yield strenght or ultimate tensile strength R or R [MPa]

Fe whisker SiC
m

Al 0
2 3

Steel Mineral glass

e

Ti alloys MgO
Al, Cu alloys

Pure metals PMMA

PA
Cement EP
PS
Ultra-pure PE
metals
Expanded
polymers

Figure 12.12. Comparative table giving the yield strength R for plastic materials, or the
e

ultimate tensile strength R, for brittle materials (after Ashby and Jones, 1980).

12.4.2. H a r d e n i n g conditions
The large difference observed between theoretical strength values and those
experimentally determined for the yield strength (figure 6.10.) is due to the presence
of dislocations favouring slip. Therefore two possibilities to increase the strength of
crystalline materials exist:
to eliminate the dislocations and to avoid their formation;
to create obstacles opposing their movement.
 Factors influencing mechanical properties 291

The first possibility is not presently feasible, although, threadlike crystalline

fibres very close to perfection (diameter < 10 ), known as whiskers, almost reach
theoretical strength. In practice, the only realistic way is to influence the mobility of
dislocations.
Two main effects are observed when a stress greater than the yield strength is
applied to a metal. Dislocations move and new dislocations are created during the
plastic deformation induced by applied stress leading to the shearing of crystals. The
increase of the yield strength (hardening) can be induced by the presence of obstacles
opposing the movement of dislocations and thus making any further plastic deforma
tion more difficult. Two categories of obstacles are known:
the atoms of the crystalline lattice, which slow down dislocation move
ment;
disturbances of the crystalline structure.
Strength due to the crystalline lattice has negligible effect on the mobility of
dislocations and on the plastic deformation of metals, because the bonds are isotropic
and all metal ions have the same polarity. Disturbances in the lattice (deformations),
generated by the presence of foreign atoms of different diameter, of other disloca
tions, precipitated particles, interfaces between two phases and grain boundaries
create stress fields capable of interacting with the stress fields of dislocations and
thereby slow down their movement.
For ceramics, it is the crystal lattice which constitutes the major obstacle. The
movement of dislocations is very difficult at low temperature. This is due to the
presence of covalent bonds or ions of alternating signs, as mentioned in chapter 7.
The yield strength of ceramics is generally very high and their capacity for plastic
deformation extremely limited.

12.4.3. Variation of internal energy associated with a dislocation

To analyse the hardening of a metal alloy, it is necessary to determine the in
crease of internal energy associated with a dislocation. It was seen in chapter 7. that
there is, around an edge dislocation, a zone where the lattice is in tension, and an
other zone where the lattice is in compression. In a screw dislocation, there is shear
deformation of the crystal around the line of the dislocation. It can be said that the
creation of a dislocation mobilises a certain amount of elastic energy stored in the
lattice in the neighbourhood of the dislocation line.
In figure 12.13., the deformation field around a dislocation is shown. The de
formation of the crystal is very large at the centre of the dislocation (dislocation
core). Practically, most of the deformation is concentrated in a tube having a diame
ter equal to that of Burgers vector. Outside this tube, the lattice deformation gradual
ly decreases in a manner inversely proportional to the distance r. Therefore, around
each dislocation, a cylindrical volume of elastic deformation is formed. This strain
field interacts with the strain fields of the heterogeneities present in the crystal, which
constitute obstacles to the dislocations movement.
The elastic strain energy of the lattice by a screw dislocation is given the fol
lowing expression:

j Gb l2
R
U In (12.14.)
4n b
292 Introduction to Materials Science

\
\

/
/
/

I b

&
1
0
Radius r

Figure 12.13. Schematic representation of the strain field around an edge dislocation.

R is the maximum radius of the deformation field and / its length. If dislocation does
not encounter any other dislocations, the cylinder of elastic deformation of radius R
is limited by the grain boundary. In all other cases, it can be considered that every
deformation cylinder meets its neighbour halfway between every pair of dislocations,
i.e. R = dl2.
In an annealed crystal, the normal density of dislocations is equal to 10 cm~ , 6 2

which gives an average distance between dislocations of 10 . For a forged metal,

the density of dislocations can reach 1 0 c m and distance between dislocations falls
12 -2

to 10 nm, i.e. approximately 30 b. For these two cases, the radius R is between 15 b
and 15 000 b.
For a characteristic distance R ~ 100 b between dislocations, the following rela
tion is obtained from (12.14):

U~-Gb l2
(12.15.)

The energy accumulated by a dislocation is therefore proportional to its length,

the shear modulus and the square of Burgers vector.
By differentiation of (12.15.), it can be written:

(12.16.)
dl 3 d

Td, which is the energy per unit length of the dislocation, has the dimensions of a
force. It is the line tension, which has a direction parallel to the line of dislocation,
and tends to minimize the dislocation energy by reducing its length. A dislocation
behaves like a rubber band. The higher the shear modulus of the crystal and the
larger the Burgers vector b, the greater is the energy and line tension of the disloca
tion.
 Factors influencing mechanical properties 293

12.4.4. T h e force required for dislocation movement

The force required for dislocation movement in a crystal has to be determined in
order to analyse the interaction between a dislocation and the various obstacles,
which it may encounter during slip.
When a shear stress is applied to a crystal containing an edge dislocation
(figure 12.14.), a force acts on this dislocation. When the force exceeds the critical
value necessary to overcome friction, it provokes the movement of the dislocation.
The force / acting per unit of length of the dislocation can be calculated. The fol
lowing expression is obtained by considering that work (rbLi) necessary to displace
the upper part of the crystal through a distance b is equal to the work (fLl) done by
the force fl acting on the dislocation of length / moving through a distance L:
f=zb (12.17.)

Figure 12.14. Force / per unit length acting on an edge-dislocation and opposing the
force f=rb provoking its slip.

The dislocation will be able to move in the crystal, if the force / = xb is greater
than a critical value f necessary to overcome the friction forces of the material. A
e

critical shear stress (yield strength) therefore exists to initiate plastic deformation:

T =f /b
e e (12.18.)

represents the critical shear stress for the slip plane concerned. For an ultra pure

monocrystalline metal, this value is between 1000 and 10000 times smaller than the
theoretical strength calculated in the absence of dislocations. The critical shear of a
metal monocrystal is thus difficult to measure exactly because the critical shear
deformation is very weak ( y ~ 10" , figure 12.15.). In an ideal crystal containing only
6

a single dislocation, r is a measure of the frictional force of the crystalline lattice

e

called Peierls stress T . In reality, in an ultra-pure monocrystal (figure 12.15.) with a

c

low density of dislocations, T is greater than T . This increase is attributed to the

e c

effect of friction due to the other dislocations and to the presence of a small number
of atoms of impurities, which add to the intrinsic lattice strength.
294 Introduction to Materials Science

(MPa)

Figure 12.15. Microdeformation of a copper monocrystal. x e is the resolved critical

shear stress.

12.4.5. Effect of particles on the movement of dislocations

Any inhomogeneity in the crystal leads to a deformation of its lattice and creates
stress fields. Repulsion occurs when the dislocation line of a crystal intercepts pre
cipitates of another phase. The resistance to the movement of the dislocation varies
locally. The dislocation line bows under the effect of the force b when it meets, for
example, two particles separated by a distance d (figure 12.16.).

Figure 12.16. Passing of dislocations through precipitates: (a) interaction of a disloca

tion line with two particles separated by a distance d; (b) forces acting on the dislocation.

When the shear stress is increased, the dislocation line forms an increasingly
curved arc until the anchorage points of the obstacles fail and allow the dislocation to
pass. At this moment, the shear stress corresponds to a local yield strength x (critical e

shear stress).
(12.19.)
 Factors influencing mechanical properties 295

In this expression T^is the increase in the yield stress, relative to r , resulting from the
c

interaction between dislocations and particles. The critical shear T is the minimum
e

shear stress required to initiate the dislocations movement.

Equalising the forces (figure 12.16. (b)) allows the following to be written:
Tbd=2T cos6 d (12.20.)

By introducing T<t (12.16.), can be calculated as a function the critical angle

of passage of the dislocation through the obstacle.

z = 2Gbcose /3d
p p (12.21.)

In this interaction process, the critical shear is influenced by different parame

ters: the shear modulus G characterising the elastic behaviour of the crystal, the dis
tance d determined by the microstructure, the vector b, which is a function of the
dislocation and the crystalline lattice, and the angle depending on the resistance of

the obstacle cut by the dislocation. If = 90 , the obstacle has no effect. If = 0 ,

obstacle cannot be cut by the dislocation and its effect on the dislocation movement
will be maximum.
In case where the obstacles cannot be cut (Orowan mechanism, = 0), e.g. in

the presence of very hard carbides in steel, the maximum contribution of the obstacle
is achieved for a shear max

Tpmax = 2Gb/3d (12.22.)

In this case, the dislocations can by-pass the anchorage points by forming a loop
between points 1 and 2 of figure 12.17. (a). This loop will progress and emerge from
the crystal by forming a step. The initial dislocation is regenerated by this mechanism
and the process resume. The precipitate particles therefore make up dislocation
sources (Frank-Read source, figure 12.17. (a) - ( f ) ) .

12.4.6. Solid solution hardening

The addition of alloying elements in solution increases the yield strength of
crystalline materials. This reinforcement is due to the force caused by the interaction
between dislocations and alloying atoms in solution. The importance of this interac
tion depends on the effect of the size of the alloying atoms. Each foreign atom can be
considered as a small precipitate, whose size is similar to that of the Burgers vector.
There will therefore be an increase in the yield strength by the reaction of the atoms
in solution opposed to the dislocation movement. This increase in yield strength is
given by:
T = aGJx
s (12.23.)

is a constant close to the unity.

The increase in yield stress T is proportional to the shear modulus G, to the
s

relative size difference of the solute atoms and the solvent, and to the square root of
the concentration X of the solute. Numerous metal alloys such as brass (Cu - Zn),
bronzes (Cu - Sn) or stainless steels are hardened by this mechanism.
296 Introduction to Materials Science

Figure 12.17. Mechanism for the multiplication of dislocations by a Frank-Read source:

(a) a dislocation in shape of a U is blocked at points 1 and 2; (b) under the effect of the
stress , the dislocation line curves; (c) when - 0 the stress is maximum; then the

dislocation line continues to curve (d) forming a loop with two parts of opposite sign
able to cancel each other out (e); a new dislocation between the original points 1 and 2 is
thus created (f) and the process can continue (see section 7.3. for the loop geometry).

12.4.7. W o r k Hardening
The yield strength of a metal is also increased by plastic deformation (chap
ter 10.). This reinforcement by work hardening results from the interaction between
dislocations, when slip becomes more and more difficult as the number of disloca
tions increases.
This basic treatment will be limited to dislocations in the same slip plane. When
two edge dislocations are in such a configuration, they exert on each other an attrac
tion or a repulsion on each other depending on the direction of their Burgers vector.
The algebraic value of the interaction force / ( p e r unit length) between two disloca
tions is given by:

f= fo
Gb
(12.24.)
2/r(l-v)r

This expression can be deduced from an analysis of the elastic stress field
exerted by dislocations. b\ and b are Burgers vectors, and r, the distance between the
2

dislocations. If b\ and b have the same orientation, there is a repulsion between the
2

two dislocations. When Burgers vectors are in opposite directions, there is mutual
attraction of the dislocations, which will tend to get closer and to destroy themselves.
These types of interactions are represented in figure 12.18.
When a Frank-Read source emits a series of dislocations of the same sign in a
certain direction under the influence of an external stress, there is an increase in the
density of the dislocations which block each other, leading to an increase in the yield
 Factors influencing mechanical properties 297

Figure 12.18. Interaction between two parallel edge-dislocations situated in the same slip
plane: (a) dislocations with the same sign repelling each other; (b) dislocations with op
posite signs attracting each other. Zones C and indicate respectively regions in com
pression and in tension in the neighbourhood of the dislocation; (c) process of annihila
tion of dislocations with opposite signs.

strength. The change in the yield strength due to work hardening can be related to

the dislocations density:

TE ~ ccGb<sfp (12.25.)

is a constant close to 0.5.

Work hardening becomes progressively more significant as the values of G and
b increase. This hardening effect is only effective at temperatures < 0.3 T , since at m

high temperature, dislocations can climb and cancel each other ( 12.6.2).

12.4.8. Grain boundary effect

Grain boundaries also constitute obstacles to dislocation motion. Their influence
is all the more effective when the orientation of the crystal planes differs strongly
from one grain to another. Polycrystals with fine grains have higher yield strength
than those with large grains.
It can be shown that the stress exerted on the boundary is equal to , is the
accumulated number of dislocations produced by a source inside the grain and , the
stress applied on the slip plane (figure 12.19.). If Tis high enough, the 'pressure' of
dislocations on the grain boundaries will initiate plastic deformation in the neigh
bouring grain. The contribution of the grain dimension (average diameter of
grains) to the variation of the yield strength can be expressed as follows:

Figure 12.19. Pile-up of edge dislocations in grain boundaries

298 Introduction to Materials Science

(12.26.)

is a constant which depends on the strength of the grain boundaries. This hardening
mechanism is used to obtain high strength in construction steels containing a low
concentration of alloyed elements (fine grain steels).

12.4.9. Polycrystal effect

For a polycrystalline material, the grains have various crystal orientations. The
initiation of slip preferentially occurs in the grains having slip systems with a Schmid
factor close to 0.5 before appearing in the other grains (figure 12.20.). To obtain a
macroscopic plastic deformation, slip must be activated in an adjoining set of grains
which do not all have slip planes and gliding directions oriented at 45 to the tensile
axis. The cooperative slip of randomly distributed grains requires a higher applied
stress relative to that needed to plastically deform a monocrystal oriented in such way
that Schmid factor is equal to 0.5. For a polycrystalline material, only an average
yield strength ; is determined which takes account of the random orientation of the
grains, this involving various values of Schmid factor. The following approximate
relations are then obtained:

T^l,5Te (12.27.)

,~3 (12.28.)

Figure 12.20. Conditions of progressive plastic deformation in a polycrystalline material.

In expression (12.27.), is the critical shear stress of the monocrystal. This

enables it to be understood why a polycrystalline material has an yield strength larger

than that of the whole of its grains which would all be oriented with a Schmid factor
equal to 0.5. No account is taken here of stresses created by the delayed deformation
of grains. This hardening effect depends only on the orientation of the grains and
differs from the same effect obtained by the refining of grains discussed in the pre
vious chapter. This latter type of hardening depends on the grain diameter and resis
tance of boundaries to the passage of dislocations.
 Factors influencing mechanical properties 299

12.4.10. Yield strength of an industrial alloy

Some hardening mechanisms lose their effect at high temperatures due to atomic
mobility. This is the case of work hardening and hardening by grain refinement of the
grains since dislocations are eliminated and grain growth occurs. It is also the case of
hardening by metastable phases obtained after quenching. Martensite (chapter 9.)
considerably hardens carbon steels, but its action is only effective at ambient
temperature where diffusion is too slow to cause transition into a stable phase. For
the development of high temperature resistant alloys, there are only precipitates and
the atoms in solid solution, which can be used as reinforcing agents ( 12.4.5. and
12.4.6.).
These various hardening mechanisms can be utilized individually or in combi
nation to produce high yield strength alloys. The effects of the hardening mecha
nisms are approximately additive. The various contributions to the shear yield
strength in T (12.22.), (12.23.), (12.25.), (12.26.) can be transformed into a tensile
e

yield strength R by using a simple multiplying factor deduced from the Schmid fac
e

tor. Finally, the following relationship expresses the yield strength R : e

R = R + AR + AR + AR + AR
e c s W P G (12.29.)

The different reinforcing elements taken into consideration are, as in the case of
shear: the crystal C, atoms in solid solution S, work hardening W, the precipitation
hardening P, and the size of grains G. By using equations relating to the mechanisms
previously studied and knowing that r is proportional to R : e e

Rc = K G l

AR = K GSX'
S 2
5

AR =K Gbp
w 3
05
(12.30.)
AR = K Gbld
P 4

AR = K /d >
G 5 G
0 s

The different constants K\ to K have values varying with the type of material.
s

Figure 12.21. describes schematically the hardening mechanisms of a metal alloy.

Atom in solution Precipitate

J L Dislocation JL^ Grain boundaries

Figure 12.21. Schematic description of the main contributions to the strengthening of a

polycrystal (after Hornbogen, 1974).
300 Introduction to Materials Science

To conclude this section concerning the hardening of crystals and more espe
cially metals, a table of values is given for the yield strength R of pure iron or iron
e

containing a small quantity of carbon, having undergone various treatments to create

different microstructures (table 12.22.). The yield strength varies by a factor 1000
according to the method of hardening employed.

Table 12.22. The yield strength Re of pure iron and Fe - C alloys after various treatments.

Material Microstructure and Condition Yield strength

R (MPa).
e

Pure Fe Monocrystal 10
Fe-C (C < 0.02 %) Annealed 150
Fe-C (C < 0.02 %) Fine grains 350
Fe Polycrystalline, work-hardened 700
Steel (C = 0.8 %) Perlite (a)
1200
Steel Martensite 2000
Steel Ultra fine perlite, work hardened 4200
Pure Fe (b)
Whisker (ideal crystal) >10000
(a) Lamellar Structure of -Fe + Fe3C.
(b) Theoretical yield strength of the pure Fe crystal: 14 GPa

12.5. Factors influencing the mechanical strength

of organic polymers
12.5.1. Cohesion in organic polymers

Low molecular weight organic compounds do not possess sufficient mechanical

strength to be used as materials except in the field of electronics where strong me
chanical properties are not necessarily required. This lack of mechanical strength
results from the nature of bonding. Organic compounds form molecular solids. In this
case, the cohesive forces results from Van der Waals interactions (chapter 2.) which
are between 10 and 100 times weaker than the covalent bonds connecting atoms in
the molecule.
Organic materials with strong mechanical properties are made up of long ma-
cromolecular chains. This chain structure considerably increases the mechanical
strength as a large number of secondary bonds ties each macromolecule to the sur
rounding chains. The mechanical strength of polymers varies considerably with their
molecular weight. When this is small, the number of secondary valence bonds
formed by each macromolecule is low and failure occurs easily by pulling out of the
molecules (figure 12.23. (a)). When the molecular weight exceeds a critical value of
about 10 000 - 40 000 according to the polymer type, there is a considerable increase
in the mechanical strength. In these materials, failure by molecule disentanglement is
no longer possible. As is shown in figure 12.23. (b), rupture necessarily requires clea
vage of covalent bonds.
 Factors influencing mechanical properties 301

Figure 12.23. Schematic representation of the macromolecular structure of an amorphous

polymer, (a) Low and of (b) high molecular weight. The dotted lines indicate the fracture
lines. The arrows in drawing (b) indicate points where homolytic cleavage of polymer
covalent bonds occurs.

Homolytic cleavage of covalent bonds (12.31) leads to free radical formation,

which can be detected with electron paramagnetic resonance (EPR):
C-CH-C-CH *- CCH*+ H C Ctf-
2 (12.31.)

H 2 H 2 H 2
X X X X
The presence of a large number of secondary bonds in the organic materials
always leads to lower mechanical strength than in metals and ceramics (table 6.2.).
Very high mechanical properties can only be obtained in organic polymers by orien
tation of the chains in the stress direction by drawing.

12.5.2. Influence of t e m p e r a t u r e
The mechanical properties of polymer materials vary strongly with temperature.
In figure 12.24., the variation of the elastic modulus for various types of polymers is
shown as a function of temperature. As polymers are generally viscoelastic, the elas
tic modulus given on the diagram is the stress relaxation modulus iv(10) measured at
an elapsed time of 10 s after application of the deformation. Another reference time
or another deformation mode would give similar results.
For amorphous thermoplastic polystyrene, a rapid variation of E (\0) can be r

observed around the glass transition temperature (T ~ 100 C for the atactic PS). If
g

the polymer is linear or branched, reduction of E (l0) occurs in two steps corre
r

sponding respectively to the excitation of short and long-range molecular diffusional

motions. The first transition involves chain segments counting about 1000 chain
atoms. The second transition concerns the whole macromolecule.
If the polymer is cross-linked, only a single variation is observed resulting from
the excitation of short-range molecular movements involving portions of chains in
cluded between the cross-links.
What differentiates the thermoplastic and reticulated polystyrene is only the
presence of a limited number of cross-links between chains segments (about one
302 Introduction to Materials Science

structural unit in 100 in this case for cross-linked polystyrene). The presence of
cross-links confers to the tridimensional polystyrene, rubbery properties at > T g

(above 100 C in this case).

The mechanical properties of rubber, which were discussed in detail in chapters
6 and 10, will not be covered again here. It should just be remembered that regular
elastomers, which exhibit rubbery behaviour at ambient temperature, have in general
a T close to - 7 0 C.
g

For semi-crystalline polymers (figure 12.24.), a first drop of the elastic modulus
( (10)) with the Tis observed (-100 C for isotactic polystyrene), which is due to
r

the glass transition of the amorphous polymer fraction. Between T 30 and T m m

(-230 C for this polymer) a second rapid variation of ( 1 0 ) takes place, which r

results from melting of the spherulites. At temperatures between T and T 3 0 , the g m

interkristalline amorphous tie chains (figure 10.4.), fastening the crystalline lamellae
together, form a transitory tridimensional network. In this temperature range, the
viscoelastic character is very marked and the polymer does not retain high mechani
cal properties.

[MPa]
I

-
V Semi-crystalline isotactic
I ^^^^^^ polystyrene -

\ Slightly reticulated \
\ ^ polystyrene

AmorphousX

g
l
thermoplastic \ 1
m
polystyrene \

50 100 150 200 250

Temperature PC]

Figure 12.24. Zv(10) versus temperature for thermoplastic, cross-linked atactic and linear
semi-crystalline isotactic polystyrene (after Tobolsky, 1960)

Below T , the yield strength and elastic modulus of amorphous and semi-crystal
g

line polymers vary in a rather similar way. As already mentioned ( 6.3.4), the rela
tion R IE for amorphous or semi-crystalline polymers is very close to the theoretical
e

value (1/15), while that of metals is generally much lower (10" < RJE < 10~ ). 4 2

The elastic modulus and the yield strength of amorphous or semi-crystalline

thermoplastics vary very little for temperatures < T (E~2 GPa - R - 50 MPa). It
g e

is very easy to vary the heat resistance of a polymer by changing its chemical struc
ture, but the level of mechanical properties remains practically the same.
The heat resistance of thermoplastic polymers depends mainly on two factors:
the strength of secondary bonds and the rigidity of the chains. The presence of sub-
stituents capable of forming strong secondary bonds (e.g. hydrogen bonds) con-
 Factors influencing mechanical properties 303

siderably increases the heat resistance. This is the case for the polyamides (Nylon )
which have already been discussed (figure 2.21. (b)).
The rigidity of the chain also plays an important role in the heat resistance. The
motion of a polymer chain is restricted by the presence of its neighbours. Indeed, a
linear polymer chain is only able to move by a snaking motion inside a tube made up
of the atoms of the adjacent macromolecules (figure 12.25.). The more rigid the
chain (because of bulky substituents), the more difficult the snaking motions are
made, which consequently occur at high temperature only.

(a) (b)

Figure 12.25. Snaking of macromolecules. In (a), a molecule surrounded with points

representing the atoms of adjacent macromolecules. In (b), the tube which surrounds the
macromolecule illustrated in (a) (after de Gennes, 1979).

12.5.3. Modification of the mechanical properties of polymers

There are three main methods to increase the mechanical strength of polymers:
chain orientation,
crystallisation,
cross-linking.
Chain orientation during plastic deformation (chapter 11.) induces a large in
crease of the polymer strength. To understand this effect, the crystal structure of
diamond will be compared with that of polyethylene (figure 12.26.). The chain
structure of diamond is close to that of polyethylene but it extends in the three di
mensions of space, this producing extremely high strength and modulus of elasticity.
The polyethylene crystal only has this chain structure in a single spatial direction. In
typical polyethylene, the crystals are small (< 1 ) and randomly oriented; the
mechanical properties are low ( ~ 1 GPa, R = 30 MPA). However, by spinning a very
e

(a) (b)
Figure 12.26. Comparison of the crystal structure of diamond (a) and oriented polyethy
lene (b).
304 Introduction to Materials Science

high molecular weight linear polyethylene from dilute solutions by a procedure

called 'gel spinning', high strength polyethylene fibres were obtained. They are
made up of crystallised extended chains oriented in the fibre axis and possess much
higher mechanical properties ( = 1 1 0 GPa, Re = 4000 MPa).
A similar structure is obtained for aromatic polyamides fibres (Kevlar) spun
from a lyotropic (oriented) polymer solution. These fibres are also crystallised ex
tended chains (figure 5.6. (f)). Parallel to the fibre axis, they have a modulus and
yield strength ( 16.3.2.) comparable to those of metal alloys.
Stretching (uniaxial or biaxial) also increases the properties of amorphous poly
mers. For example amorphous poly(ethylene terephthalate) after drawing of the order
of 500 % reaches a modulus of 9 GPa and a yield strength of 200-250 MPa while the
non-oriented polymer has a modulus of elasticity of 2 GPa and an yield strength of
80 MPa.
Crystallisation does not significantly modify the level of mechanical strength of
unoriented polymers but it increases the heat resistance. Above T a linear amor
g9

phous polymer is a viscoelastic liquid. This is not the case of semi-crystalline poly
mers, which maintain mechanical integrity up to their melting temperature, i.e.
several tens of degrees higher than T . A very characteristic example is that of the
g

polypropylene. The heterotactic (figure 5.2.) amorphous variety with a glass transi
tion temperature close to - 1 5 C does not possess any useful mechanical properties at
ambient temperature. Isotactic polypropylene (figures 5.2. and 5.10) is a semi-crys
talline thermoplastic having a T close to 170 C. It is used up to temperatures of the
m

order of 120 - 130 C. Nowadays, it is one of the most important thermoplastic

polymers. Semi-crystalline polymers having continuous operating capability up to
temperatures of 250 C for long period (over 5 years) are available.
A third method used to increase mechanical properties of polymers is cross-
linking. This modification provides to the polymer a very high heat resistance limited
only by thermal stability of the chemical links. A very good example is highly cross-
linked natural rubber called ebonite or hard-rubber. Ebonite is the oldest synthetic
rigid polymer, first commercialised in the 1940s.
Highly cross-linked thermosets reach Young's moduli of about 10 GPa, while 3
to 4 GPa is the maximum for glassy or semi-crystalline thermoplastics. However, the
yield strength and ultimate tensile strength are not substantially changed by cross-
linking. In a general way, the mechanical properties of non-oriented polymers reach
values between 40 to 70 MPa for yield strength or ultimate tensile strength.

12.6. High temperature deformation

12.6.1. Introduction
The different materials given in table 12.1. are used for extremely diversified
applications over a very broad temperature range. From this table, a series of inter
esting comments can be made.
The notion of high temperature varies greatly from one material to another.
There are several metals or alloys, which are already in a state of high temperature at
ambient or near ambient conditions. This is the case for lead, which is deformed and
worked easily at ambient temperature and of the silver and mercury alloys (amal-
 Factors influencing mechanical properties 305

gams) used in dentistry. So, A g H g , which is one of the main constituents of dental
3 4

amalgams (section 8.7.), behave in the mouth as a material at high temperature (78 %
of its melting point). Aluminium is at high temperature at T> 100 C. It is therefore
unsuitable for use in load bearing structural applications in buildings. It is excluded
from use because of the risk of flow in the case of fire. It is also very important to
note that a large number of polymers, for which the service temperature rarely ex
ceeds 200 C, are practically always used under high temperature conditions. This
explains the more or less generalized viscoelastic behaviour of these materials. This
behaviour also justifies the current development of reinforced polymers discussed in
chapter 16 which can resist very high stresses thanks to high strength, high modulus
fibres, which make up their reinforcing frame. The most effective reinforcing agents
(aramid and carbon fibre) themselves are also oriented polymers with high melting
points.
Even materials with very high melting temperature are not always suitable for
use at high temperature in oxidizing atmospheres. Most of the refractory metals
(tungsten, molybdenum, niobium and titanium) are not interesting, because their
oxide is volatile, and therefore does not form a protective coating at high tempera
ture. In a protected atmosphere, on the other hand, tungsten is used at very high tem
perature. The filament of an electric light bulb can reach a temperature of 3000 C in
argon.
Only materials forming layers of stable oxide ( S i 0 in the case of SiC) opposing
2

the diffusion of the oxygen from the surface are of interest. In the metal alloys, there
are three main elements, which protect against oxidation: Si, Cr, AI. Silicon carbide
(SiC) has very high temperature resistance and can be used to make heating elements
for furnaces up to 1500 C.
The plastic deformation of materials working at the high temperature has some
times to be controlled to improve the heat resistance. It is the case of superalloys
used for turbine blades. These metal alloys melting at very high temperature
(T = 1250 - 1300 C) have a complex microstructure and contain about ten ele
m

ments, the most important of which are: Ni, Cr, Co, Al and Ti. With a careful choice
of composition, thermal treatment and the crystallisation processes (section 12.8.) a
service temperature of 1000 C is achieved. This temperature corresponds to a rela
tive temperature of 0.8 with respect to the melting point.

12.6.2. Creep of metal alloys

When the temperature of a metal exceeds 0.4 T , the movements of atoms be
m

come important because of the increase in the number of vacancies and the diffusion
rate of atoms with (9.4.). In this temperature range when a constant stress is ap
plied, a time dependent plastic deformation occurs even at very weak stress.
At high stress, the creep results from the dislocation climb bypassing obstacles.
In this case, dislocations creep (figure 12.27.) is referred to. At low stress, the creep
occurs rather by a stress-directed atomic rearrangement known as diffusional flow.
The analysis of the deformation mechanisms for a given thermo-mechanical
loading - can be made by means of diagrams with non-dimensional coordinates
known as deformation mechanism maps (figure 12.27.). These maps must be estab
lished for every material and the various grain sizes. The relative yield strength RJG
does not decrease with the temperature because G varies in the same proportions as
306 Introduction to Materials Science

Re. The yield strength Re varies strongly with the hardening mechanism and the rela
tion RJG can be lower or higher than the value shown in figure 12.27.
The function deformation () - time (t), at constant stress or load, is known as
a creepcurve.Figure 12.28.givestwocreepcurvesatconstantloadfortwo temperatures

" 1

//////////////
{

" 3

1<T 5

0 0,5 1

Figure 12.27. Deformation mechanisms map for metals and alloys as a function of the
relative stress /G and the relative temperature TIT (1) Plastic deformation by disloca
m

tion slip; (2) Creep by dislocation climb; (3) diffusional creep (after Ashby and Jones,
1980).

Time t -
Figure 12.28. Creep curves for a constant applied force F, at temperatures T > T\.2

Primary creep (I): decreasing creep rate resulting from the work hardening
process induced by the deformation.
Secondary creep (II): balance between strain hardening and recovery (diffu
sion)
Tertiary creep (III): reduction of the effective cross-section by pores forma
tion and necking.
is the initial elastic strain. In the insert, the furnace is illustrated.

 Factors influencing mechanical properties 307

(T > ). On the loading at time t = 0, an instantaneous elastic deformation * is

2

observed in the specimen. Three different zones can be distinguished in the creep
curve as a function of time. In stage I (primary creep), the creep rate = ad at de
creases constantly to reach a constant value (stage II). This is the regime of secon
dary creep.
In stage III (tertiary creep), cavitation phenomena occur in the material. Under
constant load the real stress supported by the sample increases considerably because
of pore formation at grains boundaries and necking. In this condition, experiments at
constant stresses are therefore difficult to realise. In practice, constant nominal stress
(constant load) is used.
The main microscopic mechanisms responsible for creep are the following:
Dislocation slip (as at low temperature).
Dislocations climb which decreases or offsets the effect of precipitates
on the propagation of dislocations.
Grain boundary slip.
Diffusion of atoms and vacancies.
Among all these mechanisms, the dislocation climb plays the most important
role in practical terms. Figure 12.29. shows the mechanism by which dislocations can
climb and by-pass the precipitates at high temperature. This mechanism involved the
movement of vacancies by diffusion to dislocation lines, provoking a movement of
the dislocation perpendicular to its glide planes (formation of jogs; figure 12.29. (a)).
This phenomenon, known as dislocation climb is the reason why precipitates are less
effective at high temperature. This mechanism of dislocation movement by diffusion
of vacancies also explains how dislocations are destroyed during high temperature
annealing.
In creep, two phenomena compete with each other (figure 12.30.):
hardening by multiplication of dislocations (work hardening)
softening of the material by the annihilation of dislocations (annealing).

Figure 12.29. Diffusion mechanism of voids under the influence of the temperature pro
voking (a) the dislocations climb, by successive jog formation, perpendicular to the slip
plane; (b) deviation around an obstacle by dislocations climb.
308 Introduction to Materials Science

Figure 12.30. Representation of strengthening and weakening processes in competition

in a metallic alloy subjected to creep: (a) increase in the density of dislocations by work
hardening, e.g. formation of networks with a high density of dislocations; (b) decrease in
the density of dislocations by diffusion and dislocations climb at high temperature (an
nealing) (after McQueen and McGregor Tegart, 1975).

In stage I, the number of dislocations created by the plastic deformation in

creases more rapidly than their annihilation through diffusion mechanisms. The creep
rate decreases. When a certain density of dislocations is reached the two mechanisms
counterbalance, and creep occurs at a constant rate (stage II; figure 12.28.).
Grain boundary slip also plays an important role, because the concentration of
vacancies at boundaries is very high. Fine-grained metals have poor creep resistance.
Creep resistant materials, as used in gas turbines blades, have large grains or mono-
crystals, preferably oriented in the direction of maximum stress as in the illustrative
example (section 12.8.).
Metals and highly deformed (hardened) alloys contain a high density of disloca
tions, which increase the energy of the crystal. In a hardened polycrystal, this high
energy level tends to be reduced through several mechanisms:
recovery,
recrystallization,
grain growth (coarsening).

In recovery, dislocations are regrouped in the form of sub-boundary networks

more stable than entangled dislocations. In recrystallisation, there is seeding of new
grains with a low density of dislocations, which grow and completely reform the
 Factors influencing mechanical properties 309

grain structure. Grain coarsening is a phenomenon which reduces the total internal
(grain boundary) surface area of the material. This phenomenon can be regular (with
formation of grains of similar size) or irregular (with formation of some very large
grains among the small). This second case is known as secondary recrystallisation.
The weak creep resistance of fine-grained metal can be exploited for hot
shaping. A very fine grain (diameter of some ) produces superplasticity. It then
becomes possible to plastically deform a sample by several hundred percent. It is,
however, necessary to prevent grain-growth during this operation.
The creep rate, which depends strongly on the temperature, varies proportionally
to the coefficient of diffusion. For the second creep stage, the following relation is
generally observed:

= ADa n
(12.32.)

A is a constant and an exponent whose value for the creep of metals is between
3 and 8 for the creep by climb of dislocations, and 1 for the creep by diffusion under
low stress (figure 12.28.). As the coefficient of diffusion D is given by expression
(9.6.), then finally:

e^A'cfexpi-Q/RT) (12.33.)

The creep-rate increases appreciably with temperature (figure 12.28.) and with
stress, which influences the life-time. The parts undergoing creep at high temperature
must be changed periodically. These are parts with a life-time limited by accumu
lated deformation.
Ceramic materials having high melting points, are of interest for high tempera
ture use. However, their low toughness especially at ambient temperature still limits
their use.

12.6.3. Viscoelastic behaviour of amorphous thermoplastic polymers

Polymers in general, are at high temperature in their field of application. Be
cause of this characteristic, they exhibit generally viscoelastic behaviour ( 6.4.1.).
At temperatures of the order of T - 50 C, this viscoelastic behaviour is practically
g

negligible. It becomes important only at temperatures superior to T - 20 C.

g

Linear viscoelastic behaviour (chap. 6) has been studied in detail, and, as a

model of the viscoelastic behaviour, amorphous polymers will be analysed here. The
viscoelastic behaviour of a polymer can be characterised by the study of the stress
relaxation modulus E (t) (6.55.) as a function of time at several temperatures. The
r

stress relaxation modulus isotherms E (t) of a thermoplastic heterotactic polystyrene

r

of high molecular weight (> 100 000) are shown in figure 12.31.
In these stress relaxation curves, four characteristic viscoelastic regions can be
identified: a glassy region characterised with a modulus of elasticity greater than
1000 MPa, a glass transition region, where the modulus varies rapidly with time and
temperature, and then a rubbery plateau characterised by a modulus between 10" and1

1 MPa. In this zone, the polymer is highly deformable and it is in this temperature
zone that thermoforming is carried out. Finally a rubbery flow region is observed
in which the modulus decreases very rapidly (viscous flow). In this temperature range
310 Introduction to Materials Science

Figure 12.31. Isotherms for the relaxation of polystyrene. The temperature (C) is indi
cated on each isotherm (after Mercier and Van Cutsem, 1971).

Figure 12.32. Time and temperature superposition principle and the process of construc
tion of the master curve, illustrated by the means of 5 stress relaxation isotherms of an
amorphous polystyrene at the reference temperature T = 101,6 C. The experimental
0

temperature (C) is indicated on every isotherm.

 Factors influencing mechanical properties 311

injection moulding is performed. The four viscoelastic regions can also be identified
in the modulus-temperature curves (figure 12.24.).
The isotherms of the modulus E (t) obtained at different temperatures can be
r

superposed by horizontal translation along the logarithmic time axis. Thus, a master
curve is obtained representing viscoelastic behaviour of the polymer at the tempera
ture To chosen as a reference. This transformation is illustrated in figure 12.32. The
isotherms obtained respectively at 105.3, 104.9, 100.4 and 98.6 C can be shifted
onto the isotherm at 101.6 C, allowing the master curve at 101.6 C to be con
structed. In figure 12.33. (a), the complete isotherm at 7 = 1 0 8 C is constructed by
shifting the relaxation isotherms of figures 12.31 along the log time scale. The curve
represents the complete isotherm (master curve) corresponding the whole range of
the viscoelastic behaviour. Such a master curve covers a time interval extending over
more than 10 orders of magnitude. It cannot be obtained by direct experimentation
but only by the construction shown in the figure 12.32.
The procedure followed in the master curve construction can be summarised in a
very concise mathematical form. In a bilogarithmic modulus - time diagram (figure
12.34.), a series of isothermal relaxation curves determined at temperatures , , 0

,... is schematised.
2

I
[MPa"l
1 1 I f ' r 1

io 4
r = 108C
o

Glassy region
s 1 0 3

- Glass transition

(a) 1 10
-
^ Rubbery plateau

.1

I IO" - 1

irr -
2 1
2
io 4
io 6
() S

Time t

J 10 4

- io 2

(b)

IO" 2

1 0 - 4

-20 -10 0 10 20 30 40
Temperature T-T 0 fC]

Figure 12.33. Amorphous polystyrene: (a) Master Curve; (b) curve of the Time Factor
a(T) at TQ, equation (12.35) (after Van Cutsem and Mercier, 1971).
312 Introduction to Materials Science

logr 2 logf. logr 0 log/t

3
3
ISI

Time t

Figure 12.34. Illustration of relaxation isotherms.

The time t required to relax the sample to a particular value of the modulus (e.g.
10 MPa) depends only on the temperature and it can be written:
2

log ti = log t + log a(Ti)

0

log h = log + log a(T ) 2

log t = log t + log a(T )

Q n

Or, more generally:

log t= log to + log a(T) (12.34.)
As the relaxation isotherms can be superposed by horizontal translation, the
function log a(T) depends only on the reference temperature To chosen arbitrarily.
Log a(T) is equal to the algebraic value of the vector necessary to slide the isotherm
E (to, To) measured at the temperature To onto the isotherm E (t, T) measured at the
r r

temperature T. Log a(T) is a function which decreases with temperature. (12.34.) can
be written as:
t = t a(T)
Q (12.35.)
The function a(T) is the quotient of two times; it has no physical dimension and it is
given the name of time factor (figure 12.33. (b)).
The time and temperature equivalence principle can be expressed in the following
way: The viscoelastic functions as E (t,T) obtained at a temperature are trans
r

formed to another temperature by a simple multiplication of the time scale. If time is

expressed in the form of the variable t/a(T), the isotherms obtained at different tem
peratures superimpose to form a single curve E (h, To) which is the master curve at
r

the reference temperature T . So, the superposition procedure described in figure

0

12.34 is equivalent to the statement that the effect of temperature on the viscoelastic
behaviour is to multiply or divide the time scale by the value a(T) of at each tem
perature. When the temperature changes, only the position of the log-log modulus-
time curve is modified, but the shape of the curve is not altered. Materials having this
behaviour have been called thermorheologically simple
The time - temperature equivalence principle represents a great simplification
for the study of viscoelastic materials. Two simple functions, the master curve
E [t/a(T)] and the time factor a(T), summarise in a concise way the experimental
r
 Factors influencing mechanical properties 313

results obtained for the stress relaxation study of a single-phase amorphous polymer
in a extended temperature range.
The equivalence principle can equally be applied to the other deformation
modes such as creep. It can be shown experimentally that the value of the time factor
equation a(T) does not depend either on the mode (relaxation, creep, etc.), or on the
type (strain, shear, etc.) of deformation when the reference temperature T remains
0

the same.
As shown in figure 12.33. (a), the mechanical behaviour of a thermoplastic
polymer depends strongly on the duration of the application of the stress. If short, the
material behaves like an elastic body. As the time of application of the stress in
creases, the behaviour of the polymer tends more and more towards that of a viscous
fluid. This mechanical behaviour strongly depends on the temperature.
Viscoelastic behaviour is in close relation with the auto-diffusion coefficient of
polymer segments. The time and temperature equivalence principle only applies to
materials in thermodynamic equilibrium and for small deformations (linear viscoe
lasticity. For thermo-rheologically simple materials, this coefficient only depends on
the number of vacancies which varies exponentially with temperature. At tempera
ture close to T the mobility of the chain segments is low. Before measurement, the
g)

specimen must be annealed for a sufficient long time in order to eliminate the vacan
cies in excess of their equilibrium value.
The creep behaviour of metals cannot be likened to simple thermo-rheological
behaviour, because the microstructure of the material is often modified during the
creep experiment. However, in the case of creep resulting from diffusion, a parame
ter known as Larson-Miller (LM) can be obtained which gives a relation of equiva
lence between Tin Kelvin and t in hours:
LM=T(cst+\ogt) (12.36.)
where est is a constant. The LM parameter can be considered as a reference tem
perature for a particular alloy. Equation (12.36) is applicable only in the case where
are no significant modifications of the microstructure during creep. The parame
ter is very convenient to estimated the time to fracture of a metallic alloys at given
temperature using data obtained at higher temperature.
An essential difference between the creep behaviour of metals and polymers can
be found at the level of the reversible character of the viscoelastic deformation. In
polymers because of the chain structure, long-range forces exist which lead to a cer
tain reversibility of the deformation when the stress is removed. These long-range
retraction forces are absent in the case of metals and their alloys.

12.7. Summary and conclusions

A large number of metals and some thermoplastic polymers show a ductile be
haviour at low temperature. They fail only after a significant plastic deformation and
their tensile stress - strain curves show an elastoplastic transition. By subjecting a
material to a tensile stress, shear stresses are induced which vary as a function of the
angle between the tensile axis and the shear plane. For crystalline materials, the
movement of the dislocations displaced under the effect of the shear stresses induces
plastic deformation.
314 Introduction to Materials Science

Two main processes are involved in the plastic deformation of the glassy and
semi-crystalline polymer: plastic deformation by shear and plastic deformation by
crazing which involves formation of crazes perpendicular to the tensile axis.
A detailed analysis of the mechanisms implicated in the tensile deformation of
polyethylene shows that plastic deformation of semi-crystalline polymers leads to an
orientation of crystalline and amorphous chain segments in the direction of the ap
plied stress. The process induces the formation of a highly oriented, stronger struc
ture. This is used for the manufacture of resistant fibres.
The mechanical resistance of a crystal depends on the value of the critical shear
stress T and of the orientation of slip planes and directions relative to the applied
e

stress. The ductility of a polycrystalline material depends on dislocation mobility. A

limitation of this mobility increases hardness and yield strength. In metals, isotropic
interatomic bonding does not constitute a major obstacle to dislocation propagation.
Pure metals are thus generally ductile materials with low yield strengths. Different
processes increase the yield strength of metals by blocking dislocation movements:
by addition of solute atoms or by precipitates,
by the work hardening,
by grain refining,
by the formation of hard, metastable phases with limited deformability
such as martensite obtained by quenching carbon steels.

The very strong and oriented interatomic bonds, which exist in ceramics, fre
quently found together with ions of opposite signs, constitute important obstacles to
dislocation movement, preventing plastic deformation of ceramics at ambient tem
perature so producing hard and brittle materials. These characteristics only change at
very high temperature.
The mechanical resistance of organic polymers is generally much lower than
that of the other materials because of the presence of weak interchain secondary
bonds. It is only by orienting the chains that mechanical properties comparable to
those of metals and ceramics, can be obtained but in one direction only.
The mechanical strength of polymers depends above all on the intensity of the
interchain secondary bonds. Their rigidity varies considerably with temperature. The
transition temperatures T and T are important parameters determining the possibili
g m

ties for shaping and use of polymers.

When T> 0.5 T (T for amorphous systems), the behaviour of materials is modi
m g

fied by high temperature deformation mechanisms. Diffusion phenomena become

important and play an essential role in the mechanical behaviour. Purely elastic de
formation is no longer observed, even at low stress.
Materials creep under the action of a constant load. For metals, creep develops
progressively in three stages leading to the rupture of the material. Creep of a metal
lic material at high temperature mostly results from the effect of two opposing
mechanisms: a work hardening by accumulation of dislocations, and a restoration
which destroys dislocations by diffusion.
The behaviour of a large number of polymers, which are almost always used
under high temperature condition, is already viscoelastic at ambient temperature. In
the glass transition region, this behaviour depends only on the auto-diffusion coeffi
cient of the polymer segments. The temperature and the logarithm of time are then
two equivalent variables.
 Factors influencing mechanical properties 315

12.8. Illustrative example: single-crystals for gas turbine

Gas turbines power most current aircraft. They are also used to produce
electricity for peak periods because of the possibility of quick start-up under full
load. With the current development of natural gas, it is likely that their use will be
come more widespread. Gas turbines (figure 12.35.) have three main elements:
a compressor which injects the atmospheric air and which is the
bulkiest part of the machine;
a combustion chamber where the fuel is injected and burned (kerosene
for airplanes, natural gas for stationary turbines);
a turbine which is activated by the hot gas produced by the combustion
and which drives the compressor.

Figure 12.35. Scheme of a gas turbine.

Exhaust gases leave the turbine at great speed and propel the airplane by reac
tion. In electricity generation, the power take-off shaft of the alternator is connected
to the turbine axis by a gear.
The energy efficiency of an internal combustion engine increases linearly with
the combustion temperature (Carnot's cycle). To obtain maximum energy efficiency,
the tendency is to work at the highest possible temperatures. The materials employed
must be suitable for these extreme-operating conditions. An increase of the
power/weight ratio of the reactor is another consideration because it leads to sub
stantial fuel savings. Materials should therefore be both light and strong.
The optimum combustion temperature, reached when the fuel and oxygen are
present in stoichiometric quantity, is close to 2000 C. This temperature is much too
high for present day materials. The temperature of the combustion gas is therefore
limited to a maximum of 1200 -1300 C, which corresponds to a temperature of the
order 1000 C for the cooled turbine blades. The reduction in the combustion tem
perature is obtained by injecting an excess of air.
To understand the reasons for the choice of materials of gas turbine blades
(figure 12.36.), it must be appreciated that these components working under substan
tial centrifugal force in an hot air-gas atmosphere are subjected to significant oxida
tion and corrosion effects due to the impurities contained in the combustion gases.
These mainly consist of sulphides from the fuel and of salts (sodium chloride)
coming from the injected air.
316 Introduction to Materials Science

Wheel

Figure 12.36. Mechanical and thermal loading of a gas turbine blade. Temperature is
highest at the upper part of the blade. The tensile stress due to centrifugal force is
greatest in the root of the blade. The maximum combined load (, ) is situated in the
first third of the blade from below.

During the take-off of an aircraft, turbine blades are subjected to a tensile stress
of about 250 MPa due to centrifugal force. Blades must also resist impacts arising from
the accidental passage of foreign bodies through the turbine, for example birds. Re
sistance to thermal shocks is another important property. Thermal shock results from
the rapid cooling of blades when the engine stops, or from their rapid heating during
the take-off phase. The temperature differences induced by such changes in jet en
gine operating conditions create internal stresses due to the local variations in expan
sion or contraction. These repeated cycles of temperature variation can lead to a
blade breaking by thermal fatigue. The severe operating conditions for the blades of
an aircraft jet engine limit the choice of materials. The main criteria for the alloy
selection are:
Creep strength. Creep limits the life expectancy of the turbine blades.
These are operated at temperatures close to 0.8 T . m

Oxidation and corrosion resistance at high temperature.

Toughness.
Resistance to thermal fatigue.
Initially, the blades for gas turbines were made in stainless steel (Fe, Cr, Ni). In
the 1950s, the introduction of vacuum casting improved the alloys by adding hard
ening (but very reactive) elements such as Al and Ti. The composition of the so-
called super alloys has passed from 4 to 5 elements some 50 years ago, to about 12
elements in 1970. The number of elements has recently decreased. The typical com
position of modern superalloys used for blade manufacture is given in table 12.37.
Each element of the alloy plays a specific role. Nickel is the base metal matrix
and has a compact (fee) crystal structure with a low diffusion coefficient. The ele
ments W, Co, Nb strengthen the matrix by solid solution hardening. The elements Cr
and Al forming dense oxide films protect the alloys against oxidation (figure 15.9.).
The elements Al, Ti, Ta form precipitates of the type Ni (Al, Ti, Ta) giving structural
3

hardening. Carbon forms carbides with W, Nb, Ti and Cr. The mono-crystalline alloy
CMSX4 is characterized by a higher melting temperature than the alloy MAR 200.
This leads consequently to a better high temperature performance.
The main aim of past and present developments has been to improve creep re
sistance. It might be surprising that the gas turbine blades are made of superalloys
having a melting point between 1250 C and 1300 C, while there are other metal
alloys or ceramics capable of resisting higher temperatures.
 Factors influencing mechanical properties 317

Table 12.37. Composition of two superalloys.

MAR 200 CMSX4

Element wt.% Element wt.%

Ni 60.25 Ni 61.7
W 12.5 W 6
Co 10 Co 9
Cr 9 Cr 6.5
AI 5 AI 5.6
Ti 2 Ti 1.0
Nb 1 Ta 6.5
C 0.15 Re 3.0
Zr 0.05 Mo 0.6
0.015 Hf 0.1

Impurities Fe < 0.2 % Cu < 0.05 %

Si < 0.1 % S < 0.008 %
Mn<0,l % Pb < 0.0005 %

Figure 12.38. Oriented structure of a monocrystalline turbine blade obtained with three
dimensional modelling with the computer programme CAFE (Cellular Automaton Finite
Element) by using an grain selection device with of'pig tail* shape (after C. A. Gandin,
M. Rappazetal, 1997).

High melting refractory metals like W, Ta, Nb or Mo, have to be avoided be

cause these quickly oxidize forming volatile oxides (figure 15.9.). Numerous cera
mics also possess good oxidation resistance at very high temperatures. However,
these materials are still too brittle for this application when used in a bulk form. As a
coating, however, they protect the blades quite efficiently. Changes in manufacturing
318 Introduction to Materials Science

techniques have also allowed the thermal resistance of turbine blades to be increased.
During the 1960s, internally cooled blades with higher working temperatures were
introduced. Recently, directionally solidified monocrystalline blades have produced a
significant increase in the operating temperatures. The reduction of the microstruc
tural complexity in present-day single-crystal alloys by elimination of grain bounda
ries has lead to an increase of 50 C in service temperature.
The casting of the monocrystalline blades is difficult and demands a strict con
trol of the manufacturing processes. The blades are dendritic single-crystals con
taining precipitates of intermetallic Ni Al phases. These single-crystals are very dif
3

ferent from the silicon single-crystals presented in section 3.6. Casting takes place
under vacuum in a directionally oriented temperature gradient, so that all the grains,
which initially form, are eliminated in order conserve only one (figure 12.38.), in
particular, by using a 'pig tail' starter.
Now that nickel-based superalloys have practically reached their maximum
performance limit, uni-directionally solidified eutectic alloys offer possibilities for
the development of new jet turbine blades. Their microstructure is a composite of
carbide fibres (e.g. TaC) in a ductile matrix (nickel). These alloys open interesting
perspectives. Ceramics are also being considered for future application.

12.9. Exercises
12.9.1. Figure 12.39. gives the effect of the alloying elements Be, Sn and Zn on the
limit of elasticity R of copper. By considering the difference between the radii of the
e

atoms, what is the reinforcement mechanism induced by each of the alloying ele
ments considered?

250
Be
200

Zn

0
0 10 20 30
Alloy elements (weight %)
Figure 12.39. Effect of three alloying elements Be, Sn and Zn on the yield strength R of
e

copper.

12.9.2. A cylindrical monocrystal of nickel, diameter d = 1 cm, is oriented along its

growth axis [210]. What is the force inducing the plastic deformation of the mono-
 Factors influencing mechanical properties 319

crystal according to the preferred slip mode of the crystalline system fee of nickel
when it is applied along to the growth axis? The critical shear stress % = 5.5 MPa.
12.9.3. What is the behaviour of a thermoplastic hot stretched at a temperature above
the glass transition temperature T l g

12.9.4. What is the variation of Poisson's ratio when a metal alloy stretched in
tension deforms in a plastic way?
12.9.5. What can be done to the microstructure of polymers to reinforce them and to
increase their tensile strength? Indicate at least two mechanisms.
12.9.6. The effect of the size of grains dof a polycrystalline ferritic steel (0.1 %C) on
its yield strength Re is shown in figure 12.40.
Establish the relation between the yield strength and the grain size in
.
Determine the average number of grains per millimetre to obtain an
yield strength R = 300 MPa.
e

400, .

Grain size d~ 1,z

[mm]" ' 1 2

Figure 12.40. Relation between the yield strength R and the grain size d~ for a poly
e
m

crystalline steel containing 0.1 % C.

12.9.7. By what mechanism can edge-dislocations by-pass precipitates at high tem

perature? What is the critical temperature above which this mechanism begins to act
for an aluminium alloy having a melting point T = 570 C ?
m

12.9.8. The fine perlite of steel has a yield strength R greater than that of coarse
e

perlite.
Explain why.
Establish the relation between the yield strength R and the spacing e

between the Fe C lamellae.

3

12.9.9. How does the creep rate vary in the stationary state when the applied stress
increases in a crystalline material? Is the same behaviour observed in a non cross-
linked polymer, above its glass transition temperature T l g

12.9.10. Can a crystalline ceramic be strengthened by work hardening at ambient

temperature? Justify your answer.
320 Introduction to Materials Science

12.10. References and complementary Reading

CH. BUCKNALL, Toughened Plastics, Applied Science Publishing, London, 1977.
W.A. BACKOFEN, Deformation Processing, Addison-Wesley, Reading, Massachusetts, 1972.
CR. CALLADINE, Engineering Plasticity, Pergamon, Oxford, 1969.
P.G. DE GENNES, Scaling Concept in Polymer Physics, Cornell University Press, Ithaca, . Y., 1979.
C. A. GANDIN, M. RAPPAZ et al, 3D Modelling of Dendritic Grain Structures in Turbine Blade
Investement Cast Parts in Superalloys and Various Derivatives 718, 625, 706, E. A. LORIA ed.,The
Minerals, Metals & Materials Society, Warrendale, Pa., USA, 1997.
F. GAROFALO, Deformation et rupture par fluage, Dunod, Paris, 1970.
R.W.K. HONEYCOMBE, The Plastic Deformation of Metals, Arnold, London, 1968.
E. HORNBOGEN, Hochfeste Werkstoffe, Verlag Stahleisen, Dusseldorf, 1974.
D. HULL, Deformation and Fracture of High Polymers, Plenum, New York, 1973.
F.J. HUMPHREYS, . HUTHCALY, Recrystallization and Related Annealing Phenomena, Pergamon,
1995.
R.I. JAFFEE, B.A. WILCOX (eds.), Fundamental Aspects of Structural Alloy Design, Plenum, 1977.
A. KELLY and N. H. MACMILLAN, Strong Solid, 2th. ed., Clarendon, Oxford, 1986.
A. KELLY, R.B. NICHOLSON (eds.), Strengthening Method in Crystals, Applied Science Publishing,
1971.
J.W. MARTIN, Precipitation Hardening, Pergamon, Oxford, 1968.
H.J. McQUEEN, W.J. M C G R E G O R T E G A R T , The Deformation of Metals at High Temperatures, Sei
Am., 232, 116, 1975.
J.P. MERCIER, R. VAN CUTSEM, personnal results.
J.C. POIRIER, Plasticite a Haute Temperature des Solides Cristallins, Eyrolles, Paris, 1976.
R.E. REED-HILL, Physical Metallurgy Principles, 2nd ed., Van Nostrand, New York, 1973.
P.M. ROBINSON, H.G. SCOTT, Plastics Deformation of Anthracene Single Crystals, Acta. Met., 15,
1581, 1967.
C. T. SIMS, N. S. STOLOFF, W. C. HAGEL, Superalloys II, Wiley, New York, 1987.
A.V. TOBOLSKY, Properties and Structure of Polymers, John Wiley, New York, 1960.
 Chapter 13.

Fracture and toughness

13.1. Objectives
To differentiate between tensile strength and toughness.
To show that brittleness in materials is related to crack propagation.
To calculate critical crack length.
To relate plastic deformation mechanisms to the toughness of metals
and a number of thermoplastic polymers.
To describe fatigue behaviour.

Fracture is the separation of a material into two or more parts, under the action
of stress. This separation occurs at more or less rapid rates by the propagation of
crach pre-existing in the material. Fracture is therefore strongly influenced by the
presence of internal defects such as microscopic cracks, porosity, inclusions of brittle
particles and by notches (stress-raisers) resulting from design or manufacturing de
fects.
There are brittle materials (chapter 11.), which break without plastic deforma
tion, and ductile materials, which only rupture after a more or less significant degree
of plastic deformation. The fracture energy (chapter 6.) of a ductile material, i.e. the
toughness, can be three orders of magnitude higher than the theoretical value calcu
lated on the basis of surface energies. This important difference is directly associated
with the ductility of the material. This can decrease significantly at low temperature.
Toughness is also influenced by the deformation rate. Some materials, are ductile at
low loading rates, but are brittle if shock-loaded.
To measure the toughness of a material, the conditions of crack propagation in
terms of the applied forces are determined. All materials contain a large number of
microscopic cracks, which become unstable and grow when the applied force ex
ceeds a critical value. To understand this behaviour, the phenomena occurring at the
crack tip must be analysed in detail.
The study of the behaviour of a crack under stress is the basis of fracture me
chanics.

13.2. Brittle Fracture

13.2.1. Stress concentration
Crack propagation will be analysed in the case of an ideally brittle material i.e.
in the case of a material, which breaks without plastic deformation. Practically, in

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

322 Introduction to Materials Science

most materials, cracks are present at the surface or in the interior. These act as stress
raisers so that locally, stresses can be much higher than the applied stress. To illus
trate this, the stress field acting ahead of an elliptical crack in a thin plate subjected to
tension perpendicular to the major axis of the ellipse will be analyse (figure 13.1 (a)).

(a) (b)

Figure 13.1. (a) Thin plate containing an elliptical crack of semi-axes b and /, subjected
to a uniform stress cr, (b) stress concentration ojaat the tip of an elliptic crack with
1 = 2,5 b

The local stress reaches its maximum o m at the tips of the elliptical crack. The
following relation is obtained for o : m

= + 2-1 (13.1.)

The increase in stress intensity only depends on the ratio lib. The local stress
thus reaches very high values for a long, sharp crack (high lib ratio). For a circular
hole, / = b and a =3a. mThe stress concentration factor o A r i s found by
introducing the radius of curvature = b ll at the tip of an ellipse into equation
2

(13.1.):
/ 0 5

^ = 1+ 2 I (13.2.)
[p_
The local stress o ahead of the crack tip and parallel to the applied stress can
y

be calculated by the theory of linear elasticity. The variation of a along the axis of y

the crack (the distance being measured from the crack tip) is shown in figure
13.1. (b) for the particular case where / = 2.5 b. Accordingly, the stress a varies y

between a maximum value a =6 a forjc = 0 and the value of the applied stress
m

when = oo.
When the crack is ideally sharp (p = 0), a is given by the following simplified
y

expression:
 Fracture and toughness 323

(13.3.)

When the crack tip is very sharp, the radius of curvature reaches a value close
to that of the inter-atomic distance and the maximum stress becomes very high. The
breaking strength of the inter-atomic bonds is then locally exceeded and the crack
propagates.
Crack can propagate in two ways:
by stable propagation, with a crack tip stress just sufficient to maintain
slow crack propagation;
by unstable propagation; when the local stress a exceeds a critical
y

value a , the crack propagates catastrophically.

c

13.2.2. Critical stress for brittle fracture

Two phenomena occur simultaneously during crack propagation:
stress relaxation around the crack, since the cracked section does not
have any mechanical strength;
creation of new fracture surfaces.
Brittle materials are solids with enthalpic deformation behaviour ( 6.3.2.). In
these materials, elastic deformation results from small displacements of the atoms
from the equilibrium positions with a resulting increase in the internal energy of the
system. During fracture, a sudden release of accumulated elastic energy occurs which
is instantaneously transformed into other forms of energy: kinetic energy, surface
energy, heat, etc. This phenomenon can be compared to the behaviour of a watch
spring. When it is wound up, the spring provides the mechanical energy necessary for
the movement during a long period. If the tensioned spring is removed from its case,
the stored elastic energy is immediately released.
Any movement of the crack tip due to the action of a complex stress field leads
to a combination of the three principal modes illustrated in figure 13.2.
Displacement by crack opening (tension).
Displacement by slip (in-plane shearing).
Displacement by tearing (antiplane shearing).
This elementary treatment will be limited to the deformation in tension with
crack opening (mode I).

Mode I: opening Mode : sliding Mode : tearing

Figure 13.2. Principal modes of loading with different crack surface displacements.
324 Introduction to Materials Science

The method for calculating the elastic energy stored in a material with linear
elastic behaviour was seen in chapter 11. If a ,e and respectively represent the
e e

stress, the deformation at the yield strength and the elastic modulus, the elastic stored
energy per unit of volume can be calculated by integration of (11.8):

U = 0.5 0.5^ (13.4.)

Stress relaxation is a process in which energy is released. Conversely, the crea
tion of two new fracture surfaces is a process absorbing energy. The creation of new
surfaces is therefore a constraint on crack propagation.
The basic question is to know under which conditions the release of the elastic
deformation energy leading to the catastrophic propagation of the crack is greater
than the energy required for the formation of the new surfaces. The answer can be
found by analysing the propagation mechanism of a crack of length / at the surface of
a specimen subjected to a tensile stress (figure 13.3. (a)). The surface energy of the
crack, which is proportional to its area, increases linearly with /. Thus the surface
energy U of the crack shown in figure 13.3 is equal to that required to create two
s

new fracture surfaces:

U =
8 Ijle (13.5.)

is the specific surface energy and e, the specimen thickness.

The propagation of the crack involves a release of elastic energy in the vicinity
of the crack. The stress relaxation zone can be represented by an elliptical half-cylin
der with semi-axes, equal to 2 / and / respectively (figure 13.3. (a)). The area of the
stress relaxation zone is then equal to I . The creation of a crack length / thus in
2

volves the stress relaxation in a volume V= nl e and the released energy can be 2

calculated from (13.4.) as:

a nl e
2 2

(13.6.)
2E
The negative sign indicates that the stress relaxation involves a release of energy.

zaz:

(a) (b) (c)

Figure 13.3. Volume involved by the elastic energy release as a function of crack propa
gation during tensile deformation (a) stable crack for a < a (/ < / ); (b) critical crack,
c c

- a (l- lc) (c) broken bar / > l The hatched zone represents the volume of stress relaxa
c c

tion.
 Fracture and toughness 325

By summation of equations (13.5.) and (13.6.), the value of the crack formation
energy Uf is obtained, ^ v a r i e s with the crack length. The crack formation energy
/f (figure 13.4.) passes through a maximum for a given value l of the crack length. c

When / > / , crack propagation is accompanied by a reduction in the energy of the

c

system, and occurs spontaneously leading to catastrophic failure. Therefore, two

types of crack can exist in the system: stable cracks with a length less than l and c

unstable cracks longer than / , which propagate spontaneously.

c

Figure 13.4. Crack formation energy U as a function of the crack length / for a given
f

stress . The surface energy U is always positive and the stress relaxation energy U al
s e

ways negative. When / = / , the variation of U and U are equal in absolute value.
c s e

The phenomena accompanying crack propagation can be visualised by an analy

sis of the forces involved at the crack tip. The crack propagates under the action of a
force F applied to its tip in the plane of the crack and perpendicular to the stress axis.
This force F can be divided into two components F and F of opposite direction e s

(figure 13.5.). F is the force induced by the stress relaxation, which increases
e

linearly with the length / of the crack, while F is the restraining force associated with
s

the creation of new surfaces. F is independent the crack length.

s

The algebraic values of the forces F and F are respectively equal to the values
e s

of the first derivatives (taken with a negative sign) of stress relaxation and the surface
energies, U and U respectively, as functions of /:
e s

_ ^ = ^ ( 1 3 7 )

dZ
and

F =-^L=^
s e (13.8.)
d/

When the two forces are identical in absolute value, the crack length reaches a
its critical value l (critical crack length) which (figurel3.5. (b)) is determined by:
c

(13.9.)

The value of the critical length l given by equation (13.9.) is determined for a crack
c

propagating from the specimen surface. For an internal crack, the critical length is
doubled and l represents the half-length of the crack (figure 13.1.).
c
326 Introduction to Materials Science

Figure 13.5. Crack propagation: (a) forces F and F acting at the crack tip; (b) variation
e s

of the absolute value of the components F and F of the crack propagation force for a
e s

plate of unit thickness.

The critical stress a c can be calculated from equation (13.9.):

(13.10.)

Gc cannot be exceeded in a brittle material containing internal cracks of length 21 or
surface cracks of length /. In 1920, Griffith postulated the existence of cracks in ma
terials and introduced this concept of critical stress a . Any applied stress exceeding
c

the critical stress a produces a rapid propagation of the cracks and leads to rupture.
c

13.2.3. Elements of fracture mechanics

The Griffith's crack theory, introduced in the preceding paragraph, is limited to
brittle materials. Fracture mechanics aims to find a general relationship to enable
the critical stress o to be calculated whatever the type of material.
c

Griffith's analysis can be generalised by considering that the reduction in me

chanical energy by stress relaxation must reach a critical value in order for fracture to
occur. Using G to designate the energy released by stress relaxation per unit area of
crack formed, equation (13.6.) can be written:

G= (13.11.)

For a crack of length / to grow, the applied stress must exceed the critical value a c

given by (13.11.)

(13.12.)

This equation, which applies to any type of material, is similar to Griffith's
equation, which only applied to ideally brittle materials. In the general case, G no c

longer corresponds to the simple surface tension / b u t takes account of all the energy
dissipated during crack propagation. As already mentioned in paragraph 6.3.4, G , c

which is considered as the crack propagation energy, defines the specific fracture
energy. G is a parameter, which measures the toughness of a material and so it is an
c
 Fracture and toughness 327

intrinsic material parameter. For ideally brittle materials, the parameter G is equal to c

(6.43):
G = 2y
c (13.13.)

Under these conditions, equation (13.12.) is equal to (13.10.). The factor 2 intro
duced into equation (13.13.) takes account of the fact that two fracture surfaces are
formed at the crack tip. Equation (13.12.) can also be written:
a (nl)
c
05
= (EG f c
5
(13.14.)

The left hand term of this equation describes the applied stress state whereas the right
hand term describes the material properties. Equation (13.14) defines an important
fracture mechanics parameter: the stress-intensity factor, K:
=() 05
(13.15.)

The stress intensity factor must not be confused with the stress concentration factor
a Ja given in equation (13.2).
Failure occurs when the stress intensity factor, reaches the critical value K : c

K = a (nl)
c c
05
= (EG ) c
05
(13.16)

K can be measured by a standard test on a specimen which already has a pre

c

existing crack of length a. K is another parameter which allows the toughness of a

c

material to be measured. Knowing K the values of G can be determined using C9 c

equation (13.16). The failure mode described in figure 13.3 corresponds to a tensile
deformation (mode I). A test employing this mode gives a factor K\ , The factor K c c

integrates into a single expression G , which measures the energy required to propa
c

gate the crack, and the modulus of elasticity E. The higher the value of and G , the c

more difficult is the crack propagation. It is therefore the ductile materials with high
modulus of elasticity (metals) that are toughest.
The values of G , R and K for the principal materials are given in table
c m c

13.6. Figure 13.7 gives a comparative table for the values of the toughness K of c

various types of materials. There is no direct correlation between the tensile strength
R and toughness G . If the same material is tested in various forms and dimensions
m c

(e.g. glass plates vs. glass fibres), the observed variations in tensile strength can be
explained on the basis of the dimensions of the critical cracks. Indeed, a fibre cannot
contain cracks longer than its diameter and is therefore always stronger than the
bulk. Brittle materials with low values of K do not necessarily have low strength. On
c

the other hand, the fracture propagation energy G of these materials is low. c

The factors K and G constitute the two intrinsic parameters for the measure
c c

ment of material toughness.

13.3. Ductile Failure

13.3.1. Ductility and Toughness
The crack propagation energy corresponds approximately to twice the surface
energy (figure 6.11.) for very brittle materials only, such as mineral glasses. In all the
328 Introduction to Materials Science

Table 13.6. Crack propagation energy G , stress intensity factors

c

K and tensile strength R of principal materials.

c m

Material G c K c R m

Urn -2
MNm~ 3/2
MPa

Mineral glasses 0.01 0.7-0.8 30-90

Glass fibres 0.01 0.7 300-3000
Alumina (A1 0 ) 2 3
0.02 3-5 10 000
Magnesia (MgO) 0.04 3 5000
Silicon carbide (SiC) 0.05 3 3000
Polyester, epoxy 0.1 0.3-0.5 50-90
Polyamide 6 2-4 3 100
Polypropylene (PC) 8 3 30
Polyethylene (EP) 6-7 1-2 10-40
Polystyrene (PS) 2 2 35-80
Poly(methyl methacrylate) (PMMA) 0.3-0.4 0.9-1.4 60-80
Wood 10 11-13 40-80
Steel 100 140 400
Stainless steel 10-100 50 150
Pure copper, aluminium 100-1000 100-350 200-400

Metals Ceramics Polymers

200
Pure metals
Steels
100

50
Al alloys

20

10 Al 0 2 3 + Zr0 2

5 Be Si N<
3

Al 0 2 3

PS
2 S i C , MgO
PP-Nylon

1 PMMA
Glass PE
Epoxy
0.5
Cement
0.2

Figure 13.7. Comparative table of the values of the toughness K for various types of c

materials at ambient temperature (after Ashby, Jones, 1980).

other cases, the crack propagation energy is increased in a more or less significant
manner by other energy dissipation processes.
In metals, dislocation slip is the main energy dissipation mechanism. The
plastic deformation at the crack-tip (figure 13.8.) leads to an increase in the radius of
curvature and a reduction of the stress concentration factor. Moreover, a plastic de
formation zone forms at the crack-tip.
For polymers, a specific mechanism of energy dissipation involving crazing
(micro-cracking) was described in chapter 12. At the crack-tip, a strip of stretched
 Fracture and toughness 329

chains fibrils is formed which slows down the crack propagation. The fibrils length
increases as the crack lips separate by progressive extraction from the as yet unde-
formed polymer chains. Some ductile polymers, such as polyamide 6 (13.6.) have
values of G notably higher than those of ceramics, whereas the values of K are of
c c

the same order of magnitude (figure 13.7.). This is due to the low value of the elastic
modulus of polymers. As noted in figure 13.8., the shape of the plastic deformation
zone depends on the mechanisms of energy dissipation at the crack-tip.

Figure 13.8. Illustration of fracture mechanisms at the crack-tip: (a) macroscopic view
(from left to right) of a brittle material (ceramic), ductile material with a plastic deforma
tion zone (metal) and ductile material with crazing (microcracking) (thermoplastic
polymer); (b) microscopic view of the crack-tip without plastic deformation in a brittle
material, due to the movement of dislocations in metals and resulting from stretching and
rupture of thermoplastic polymer chains.

13.3.2. Plastic zone ahead of the crack-tip

Consideration will be limited here to the case of plastic deformation by shear
yielding due to dislocation movement as in metallic alloys. The plastically deformed
zone at a crack-tip is caused by stress concentration.
At the beginning of this chapter, the evolution of the stress a at the tip of an y

elliptic crack contained in a large, thin plate deformed in tension was analysed
(figure 13.1. (b)). According to equation (13.3.), the local stress of a crack with ra
dius = 0 tends to infinity for JC = 0. In fact, a never becomes infinite since the
y

radius of curvature at the crack tip is at least of the size of the interatomic distance.
For ductile materials, the maximum local stress is limited by the material strength.
The zone of plastic deformation is calculated by substituting the yield strength R to e

o in (13.3.):
y
330 Introduction to Materials Science

( Y' 5

(13.17.)
Oy = Re = [1 + (l/2x ) ] 05

p

V
2x J

In this expression, x defines the limit of the plastic deformation

p zone ahead of the
crack-tip (figure 13.9.). Thus:

(13.18.)
X p
~ 2R]

By taking into account the stress intensity factor given in equation (13.15.):
_ K 2

(13.19.)
XP =
2nR 2

The zone of plastic deformation is especially important when is large, the

crack /, long and the yield strength R , low. e

This elementary treatment is limited to the analysis of the stress field in the
direction of the crack. A complete treatment can be performed to determine the entire
deformed zone. The cross section of the deformed zone is approximately equal to a
circle of diameter x (figure 13.9.). Basic calculations made here show that the yield
p

strength is reached locally at very low stress values when long, narrow cracks are
present in these materials.

Figure 13.9. Distribution of the stress a ahead of the crack as a function of distance. The
y

plastically deformed zone is shown in grey.

When plastic deformation occurs on failure, it is necessary to add a plastic de

formation energy term to the surface energy factor . From (13.13.), the specific

fracture energy Gc can be written as:

G = 2y+ Yp
c (13.20.)

For ductile materials, 2 7 is negligible with respect to y . p

A multiplication of the energy absorption mechanisms increases the toughness

of materials. An example of this is the improvement in toughness of glassy polymers
such as polystyrene by the addition of rubbery particles during synthesis of the
polymer (figure 8.2.). Under impact, the rubbery particles in the glassy matrix initiate
a very dense network of microcracks (crazes), which absorb the energy of the impact.
The rubbery particles stop the transformation of the microcracks into cracks and
 Fracture and toughness 331

therefore prevent the brittle fracture of the material. Polystyrene (High Impact Poly
styrene - HIPS) modified by the rubber particles is ductile (figure 13.10.), whereas
pure polystyrene is brittle.

0 10 20 30 40
Strain e (%) ^

Figure 13.10. Tensile curves of a pure polystyrene (PS) and impact resistant polystyrene
(HIPS) (after Bucknall, 1977).

13.3.3. Fatigue
Structural and machine components are frequently subjected during their service
lives to loads of variable amplitude Such repeated loadings can lead to failure by
fatigue at stresses lower than the tensile strength or the yield strength. Variable
loading occurs, for example, in the rotating parts of engines. Fatigue failure can also
occur in large constructions such as bridges subjected to periodic forces of limited
amplitude.
The fatigue process implies that cracks are developing gradually under the
action of repeated loads of limited amplitude. Measurement of crack grow rates are
commonly performed on precracked samples. A sample containing a crack is sub
jected to repeated cyclic loading with a stress variation and the crack length / is
recorded as a function of the number of loading cycles n. Failure occurs when the
crack reaches the critical length l . c

The crack propagation rate all an is determined from the slope of the experi
mental curve giving the variation of / as a function of n. The variation of the stress
intensity factor AK at various stage of the fatigue test is also calculated from equation
(13.15.):
= () 05
(13.21.)

Figure 13.11, which gives the evolution of the crack propagation rate dl/dn as a
function of the variation of AK, describes the fatigue behaviour of the test sample. If
the surrounding environment is not corrosive, the propagation rate of the crack is low
or even zero below a critical threshold value of the variation of the stress intensity
factor AK or the stress for a given crack length. Above this threshold value, an
0 0

initially rapid and then more progressive increase in the crack propagation rate is
observed, which reaches a constant slope regime:
all an = AMC" (13.22.)
332 Introduction to Materials Science

A and m are characteristic constants of the material. Finally, the critical value of the
stress intensity factor Kc is reached in a single cycle leading to the instantaneous
failure of the specimen.
Microstructure changes, which modify the yield strength or the homogeneity of
the material, can significantly alter the fatigue behaviour. The average value and the
frequency of the load application are also significant factors changing the fatigue
performance. This is particularly so for polymers in which heat-build up occurs under
the effect of cyclic loading thus causing creep of the material. Corrosion (chap
ter 15.), which decreases the critical threshold value AK for crack propagation
0

(figure 13.11. (b)), also plays a important role in the fatigue behaviour of metallic
materials.

Figure 13.11. Various stages of fatigue cracking shown by a curve relating the rate of
cracking dl/dn as a function of the variation of the stress intensity factor (a) Diagram
showing three zones of crack propagation by fatigue. Zone (2) is described by equation
(13.22.). (b) Curve for the rate of fatigue cracking as a function of AK for a Ti-al-
loy (Ti: 8 wt. % - Al: 1 wt. % - Mo: 1 wt. %) in a neutral atmosphere (argon) and a cor
rosive solution (3.5 % NaCl).

13.4. Summary and Conclusions

The failure behaviour depends on the presence of cracfa in the interior or at the
surface of materials and on the conditions of crack propagation. A crack in a material
will propagate catastrophically under an applied stress only if the increase in crack
length is associated with a reduction in the total energy of the system. The relation
ship between the elastic energy released by crack propagation and the energy re
quired for the creation of the fracture surfaces allows the calculation of the crack
critical length. Any crack exceeding this critical length is unstable.
 Fracture and toughness 333

A crack of length /, will propagate catastrophically during a single loading, if the

applied stress exceeds a critical value a . Equally, from a knowledge of the stress
c

applied to a machine component, the maximum size of crack which can be tolerated
without catastrophic failure, can be calculated.
An elementary study of fracture mechanics shows that fracture depends on two
quantities: the applied stress and defect length /. These two values are combined in
the stress intensity factor K. For the same material, the critical stress a can vary
c

according to the size of the defects present in material. Fracture occurs when the
critical value K is reached.
c

The measurement of the critical crack propagation energy G is associated with

c

the stress intensity factor K . G contains not only a surface energy term, but also an
c c

additional parameter (often larger) which takes into account the energy dissipated by
the formation of a plastically deformed zone ahead of the crack tip. The toughness of
a material, measured by K or G , is a fundamental mechanical property in the same
c c

way as the elastic modulus or the tensile strength.

When a material is subjected to repeated loading, cracking can progress at a
stress intensity factor lower than the critical value K . Repeated loading finally leads
c

to fatigue failure. In these circumstances, a significant reduction of the component

lifetime is observed when corrosion is present.

13.5. Illustrative example: The triumph and the tragedy

of the De Havilland Comet
The first jet aircraft were built during the Second World War: the British Glos-
ter Meteor, and the Messerschmitt 262 produced by the German war industry. These
jets, which were much faster (record speed of the Gloster Meteor IV: 976 km/hour)
than propeller driven aircraft and which had a much greater operating range, opened
new perspectives for commercial civil aviation.
The first commercial jet aircraft propelled by kerosene-fuelled gas turbines
supplied and built by the English firm, de Havilland entered service on May 2 , nd

1952. It was the first of a new generation of aircraft with many innovations which
have become standard today: 4 gas turbine engines, high pressure fuelling, hydraulic
flap control, air-conditioned pressurised cabin. It was able to transport 36 passengers
at 750 km/h over 4 000 km.
To obtain reasonable commercial productivity in spite of the poor efficiency of
the gas turbines of the time, the manufacturer had to minimise the airframe weight
and limit the quantity of fuel carried. To achieve maximum weight gain, the fuselage
was built using two different assembly techniques: adhesive bonding, which was at
the time an avant-garde technique, and traditional riveting. A cruising altitude of 10
to 11 km was used to optimise engine efficiency. This altitude, significantly higher
than that of the propeller driven commercial aircraft of the time, imposed much more
severe cabin pressurisation cycles on this new aircraft. The back-pressure in the fu
selage, zero at take-off was 0.58 atm at the cruising altitude, i.e. double the value
used up to then.
At the time of the design of this first commercial jet at the end of the nineteen
forties, notions of fracture mechanics already existed, but they were hardly used by
334 Introduction to Materials Science

mechanical engineers. Moreover, the stress calculation in the vicinity of the rivets
indicated much lower values than in reality. The aluminium alloys were much less
tough and less strong than those of today. All these elements contributed to the
tragedy which took place 2 years after the beginning of service. Two Comet aircraft
of British Airlines exploded and were lost in the Mediterranean Sea, one, on January
10, 1954 close to the Island of Elba and the other, on April 8 of the same year, off
Naples. From the wreckage found off the Island of Elba, it was not possible to de
duce anything about the origin of the catastrophes. But one particular fact attracted
the attention of the investigators: the two accidents occurred just after take-off, at an
altitude of 10000 m, i.e. at the moment when the stresses on the walls of the pres
surized cabin were at a maximum. These observations led the investigators to think
that the sudden failure of the aircraft structure could be due to constructional defects
which generated severe stresses at specific points in the fuselage.
Finally, the hypothesis of a fatigue failure was regarded as plausible, and it was
decided to perform fatigue tests on a Comet aircraft, which had already flown but
was still intact. The fuselage of Comet G-ALYP, nick-named Yoke Uncle, was en
tirely immersed in a vast water-filled container. Only the wings emerged on each side
of the container. The aircraft had already flown 3500 hours, i.e. a flying time compa
rable with that of the two crashed aircraft. The container with the aircraft inside was
built in six weeks. Loads were distributed on the wings and the fuselage to reproduce
the loading conditions on the aircraft corresponding to pressure variations in flight,
the landing and takeoff sequences and, at the same time, the effects of the thrust from
the engines. The cycle of pressure variations imposed on the cabin oscillated between
the values experienced on take-off or landing. The results showed that after a test
time of two and a half times the initial period (9 000 hours), a fatigue crack occurred
at the base of a window in the emergency door (figure 13.12.).
A first conclusion was drawn from this experiment. If the tested aircraft had
continued in service, it would have probably suffered an accident after approximately
9 000 hours of flight. This was significantly higher than those of the crashed aircraft,
but the scatter in the results inherent in this type of simulation test must be taken into
account.
A detailed examination of the fuselage showed that the crack started at a rivet
hole situated at a lower corner of the forward rectangular shaped port escape hatch.
The cracked sheet was made of aluminium alloy. Elements of this sheet taken from
the fuselage and subjected to fatigue tests showed that the same type of fatigue crack
initiated at rivets in the vicinity of the windows. Fatigue tests were also carried out
on cylinders made of rolled and riveted sheets, and subjected to cycles of variable
internal pressure. Other tests were carried out on flat specimens of aluminium alloy
with one drilled hole, subjected to alternating tensile stresses. Figure 13.13 shows the
variation of the breaking strength as a function of the number of cycles for the two
types of tests. It can be seen that the rupture strength falls to one third of the initial
value after approximately 30 000 cycles.
The conclusions of the report on all of the above mentioned tests attributed the
origin of the two Comet aircraft disasters to the low fatigue resistance of the cabin.
Other hypotheses were also retained, because the subsequent examination of the
remains of the Comet which crashed off the Island of Elba did not show conclusively
that failure was from the propagation of a crack induced by fatigue.
 Fracture an d toughnes s 335

Figure 13.12 . Positio no f th e principa lfracture s observe do nCome taircraf tG-ALY P af -

ter fatigu etests .I ngrey ,th ecracke dsectio nwhic hle dt oth e fractur eo f th e structur e
(reprintedwit hpermissio nfrom Lancaste r(2000). )

IMPa]

J 50 0

<J 40 0

30 0

3
S 200

100

10 2 10 3 10 4 10 s

Cycles to failure
Figure 13.13. Variation of the peak stress supported by the specimen as a function of the
number of cycles: (a) alternating tensile stress tests on a flat specimen with a drilled hole
in a fatigue testing machine (b) tests on pressurised cylinders of ~1 m of diameter, made
of rolled and riveted sheets (after Bishop, 1955).

From these disasters, engineers learnt lessons which were put into practice in
future designs of aircraft and machine. It became quickly apparent that it was im
portant to avoid notch effects and stress concentrations. Many other examples of
catastrophic failures observed at the time (turbines, conduits, tanks) showed that the
initiation of these ruptures occurred at stresses considerably lower than the yield
strength R measured for the materials. Significant work concerning the behaviour of
e

notched materials was then undertaken and testing methods for toughness control, i.e.
resistance to crack propagation, were proposed.

13.6. Exercises
13.6.1. A technical ceramic Sialon (S13AI3O3N5) has a tensile strength Rm of
415 MPa. What is the maximum stress that can be applied to it if elliptical cracks
present on the surface have the following average dimensions:
336 Introduction to Materials Science

major axis 21 = 0.50 mm,

minor axis 2b = 4.5 mm.
13.6.2. A ceramic bar of diameter d equal to 5 mm has a fracture toughness
0

Kic = 2.75 M N m . If this bar is subjected to a tensile force F o f 1900 N, determine

_3/2

the maximum permissible length / of an infinitely sharp crack both on the surface and
inside the material before brittle fracture.
13.6.3. An element of the wing of an aircraft is made of aluminium alloy with a yield
strength R = 470 M N m and a toughness K\ = 21 MNm~ . This part is designed to
e
-2
c
3/2

resist a critical stress a equal to 0.5 R An inspection procedure can detect defects
m e

of 3 mm and greater. Is this method sufficiently sensitive to detect critical internal

defects? Briefly explain your conclusion.
13.6.4. The toughness of a material is measured by the energy absorbed during the
rupture of a sample subjected to an impact by a mass m placed at the end of an arm of
a pendulum of length / = 75 cm and released from a standard height (Charpy test).
Calculate the energy absorbed if the mass m = 10 kg is released from an initial
height corresponding to an angle of 120 of the pendulum arm. After the rupture of
the sample placed at the lower part of the instrument, the mass reaches a height cor
responding to an angle of 85 .

13.6.5. Metal alloys with a bcc crystal structure have a ductile-brittle transition when
the temperature decreases. This is the case for ferritic steels (figure 13.14.).
What is the type of failure in a 0.2 % weight carbon steel at ambient
temperature?
What is the embrittlement temperature of this steel?
Why does the fracture toughness decrease with an increase in the carbon con
tent?

Figure 13.14. Influences of the carbon concentration in steels on the impact energy
measured as a function of temperature
 Fracture and toughness 337

13.6.6. A steel suspension plate (toughness K\ = 90 MNm" ) is subjected to a stress

c

= 350 N m . Calculate the minimum dimension / of a crack able to propagate from

-2

the surface. Determine the minimum dimension of an internal crack.

13.6.7. From the modulus of elasticity of glass (70 Gpa) and the surface energy
(0.6 Jm" ), calculate the fracture stress a of a glass plate containing a series of inter
2
c

nal microcracks of 10 maximum size . What will be the fracture stress a of this c

plate if a fine scratch 0.3 mm deep is made on the surface?

13.6.8. A annealed steel containing 0.4 wt. % carbon has a fracture toughness
Kic = 150 MNm~ . After quenching, will the toughness K\ and the yield strength Re
3/2
c

of this steel be greater or smaller than the toughness and yield strength of the steel
before quenching?
13.6.9. What is the difference in the creep behaviour curves (deformation - time t)
of a metal alloy sample subjected to a constant tensile stress or a constant load?
1 3 . 6 . 1 0 . A steel bar is subjected to alternate tensile fatigue cycles
( Cf= 100 MNm" from a = 0 to = 100 MNm" ). The average dimension of the
2 2

surface cracks is / < 2 mm. Estimate the lifetime of this bar if the toughness of steel is
K\ = 30 MNm" and the values of the coefficients of equation (13.22.) are m = 3
c
3/2

and A = l 10~ for expressed in MPa and / in mm.

12

13.7. References and complementary reading

T. BISHOP, Fatigue and the Comet Disasters, Metal Progress, (1955) 79.
W. BROSTOW, R.D. CORNELIUSSEN (ed.), Failure of Plastics, Hanser, Munich, 1986.
C.B. BUCKNALL, Toughened Plastics, Applied Science Publishers, London, 1977.
. H. COURTNEY, Mechanical Behavior of Materials, McGraw-Hill, New York, 1990.
T.V. DUGGAN, J. BYRNE, Fatigue as a Design Criterion, Macmillan, New York, 1977.
D. FRANCOIS, A. PINEAU et A. ZAOUI, Comportement mecanique des materiaux, Hermes, Paris,
1995.
R. W. HERTZBERG, Deformation and Fracture Mechanics of Engineering Materials,^ ed., Wiley,
New York, 1996.
H.H. KAUSCH, Polymer Fracture, Springer, Berlin, 1978.
J.F. KNOTT, Fundamentals of Fracture Mechanics, Butterworths, London, 1973.
J. LANCASTER, Engineering Catastrophes. Causes and effects of major accidents, Abington Publishing,
Cambridge, 1996.
 Chapter 14.

Physical properties

14.1. Objectives
To stress the difference between electrical conductors, semiconductors
and electrical insulators.
To study the electric conductivity of metals and to determine the effect
of the temperature and the addition of various elements on this
property.
To define the superconductive state.
To determine the conditions for obtaining a semiconductor starting from
ultra pure silicon.
To study the thermal conductivity of materials.
To analyse the magnetic behaviour of materials as a function of their
atomic structure and to differentiate between permanent and non-
permanent (soft) magnets.
To relate the optical properties of materials (transparency, brightness,
colour) to photon and electron interactions.
Indeed, the mechanical properties studied in the previous chapters are intrin
sically part of the physical properties. If they have traditionally been treated sepa
rately, it is in fact for reasons of didactic convenience.
The area covered involves traditional concepts of solid-state physics based
in particular on the study of the movement of electrons. The importance of these
theories for Materials Science is due to the fact that the totality of the physical prop
erties, including the mechanical properties, can be interpreted in a unified way with
quantum mechanics. When necessary for didactic convenience, more traditional
models will be taken also account in this chapter
In this work, which is an introduction to the study of the structures and
behaviour of materials, the contents of this chapter are deliberately limited to those
physical properties, which are of most interest to the materials scientist.

14.2. Electric Conductivity

14.2.1. Conductivity and electrical resistance
Electrical conductivity is a property of materials of great technical importance.
The transport of electrical power over long distances and electric resistance heating
can be quoted as significant examples of this type of phenomenon.

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

340 Introduction to Materials Science

In the first case, it is necessary to reduce the energy losses as much as possible
by using materials with high electrical conductivity, such as copper or aluminium.
This objective explains the immense interest in the search for superconductors, which
could be used at temperatures close to ambient. In the case of electric resistance
heating, the material composition has to be adapted to maximise the transformation
of electric power into thermal energy.
A third example, just as important, is provided by semiconductors on which all
modern electronics (Hi-fi equipment, computers, etc.) are based. The various phys-
icochemical and physical concepts on which the phenomenon of electric conductivity
is based will be treated simply in this section.
An electric conductor is a material with free electrons allowing the transport of
an electrical current. When an electrical current of intensity I moves through a con
ductor of resistance R, a potential difference given by Ohm's law, is observed at
the ends of the conductor:
V = RI (14.1.)
The units used for V, R and / are the volt (V), the ohm () and the ampere (A).
Knowing the geometrical dimensions of the conductor, i.e. the cross section S
and the length L, the specific resistance or electrical resistivity of the conductor can
be calculated:
RS 1 (14.2.)
=
=

The resistivity p , which is expressed , reflects the resistance opposed to

the passage of an electrical current through the material, for unit cross section and
length. Its inverse, the electrical conductivity is thus a measure of how easily an
electric current flows through the material.
The passage of an electric current through a conductor is always accompanied
by a release of heat, it is the Joule effect
= VI = Rf (14.3.)

is the thermal power released. This is expressed in Watts (1 W = 1 J s ) .

-1

The conductivity and the electrical resistivity of materials vary very considera
bly according to the nature of their chemical bonds, the composition and the tem
perature (figures 14.1. and 14.2.). As figure 14.3. shows, electric conductivity is
directly proportional to the number of mobile electrons present in materials.
Thus, at ambient temperature, electric conductivity varies from ^ for
8 1

the best metallic conductors (Cu, Ag) up to ^ ^ for the best electrical insula
1

tors such as polytetrafluoroethylene (PTFE) (figure 14.1.).

In certain cases (materials which are superconductors at low temperature), re
sistance falls to zero below a critical temperature. In practice, for superconductors,
conductivity exceeds ^ . Metals (figure 14.4.) have a very high electrical
27 1

conductivity at ambient temperature, which increases with decreasing temperature.

The use of copper and aluminium as materials for carrying electric power is well
known.
Ceramics represent the group of materials showing the greatest variation of
electrical conductivity. They can be classified in three groups: conductors, semi
conductors and insulators. Thus certain carbides (SiC, TiC) have partially metallic
 Physical properties 341

io 11
Metals Ceramics Polymers

io 8

Cu,Ag,Al
Fe
io 5 Hg Ti
Sn Graphite (C)
SiC
IO 2

T3
10~
Ge (pure) CO c
SiC (pure)
GaAs
KT 4
Polyacetylene
(pure)

Mineral glass

PMMA
Diamond PE,PS
Si0 2
PTFE
IO" 1

Figure 14.1. Values of the intrinsic electric conductivity of different classes of mate
rials at ambient temperature.

Isolators

Semiconductors

Conductors

Supraconductors

Figure 14.2. Electrical resistance of materials as a function of temperature T.

342 Introduction to Materials Science

Figure 14.3. Representation of the variation of electric conductivity ( ) as a function of

the density of mobile electrons.

Figure 14.4. Variation of the resistivity as a function of temperature for a normal metal
with (a) a high and (b) a low density of defects and (c) for a superconductive metal. The
temperature T marks the transition between the conducting and superconductive state.
c
 Physical properties 343

bonds and, in a certain number of cases (e.g. TiC), an electrical conductivity close to
that of metals. As already mentioned in chapter 2, pyrolytic graphite has a high elec
tric conductivity in the direction parallel to its hexagonal basal plane. Certain ce
ramics based on oxides of yttrium, copper and barium are superconductive at the
temperature of liquid nitrogen (-196 C).
Certain materials have an intermediate electric conductivity (~10~ to 2

10 Q m ) between that of metals and that of insulators at ambient temperature;

4 _ 1 - 1

these materials are called semiconductors.

The most common semiconductors (Si, Ge) are materials with covalent bonds.
The intrinsic electric conductivity of these un-doped semiconductors is about 10~ to 2

1 \ This is highly sensitive to the presence of impurities, a very important point

_ 1

which will be considered further.

As already highlighted in chapter 1, substances like germanium and silicon,
which have properties intermediate between those of metals and non-metals, do not
fit into any simple classification. By the nature of their chemical bonds and their
physical and mechanical behaviour, the semiconductors like germanium and silicon
tend to be associated more with ceramic materials.
Among insulating ceramic materials, a large number with ionic, covalent or
polar covalent bonds can be identified, such as amorphous silica, quartz, mica
(phyllosilicate) and diamond, which are characterized by an electric conductivity
approximately 10 times lower than that of graphite.
21

Generally speaking, polymers are electrical insulators. Thus, the use of poly
ethylene (PE) as an electrical insulator for high (200 kV) and low-tension cables is
well known.
However, certain conducting polymers like polyacetylene are under develop
ment. These conducting polymers are characterized by the presence of conjugated
double bonds and molecular orbitals, which confer mobility to the electrons.
From the point of view of their electronic structure, only two classes of materials
exist: conductors and insulators. In chapter 2, the elements of the band theory were
analysed. Conductors (mostly metals) are materials whose last occupied band is in
complete. In insulators and semiconductors, on the other hand, electrons occupy all
the energy levels of the valence bands. At zero degree Kelvin, under the action of a
moderate electric field and in the absence of any other external excitation, such as
light for example, a displacement of electrons (electric current) would occur in the
conductors, while in the case of insulators and the semiconductors, there would not
be any movement of the electrons.
There is therefore no qualitative difference between an insulator and a semicon
ductor; what differentiates them is only the width of the forbidden band (figure 2.20).
As mentioned in chapter 2, in a semiconductor, the width of the forbidden band is
close to 1 eV and a certain number of electrons are excited, at ambient temperature
and above, from the valence band into the conduction band. When the width of the
forbidden band is greater than 2 eV, the excitation of the electrons is no longer
possible under ambient conditions and the material became an insulator.
The variation of the electric conductivity of metals and semiconductors as a
function of temperature is very different. Thus, in pure metals (figure 14.4.) at very
low temperature (T < 10 K), conductivity hardly varies with T. At higher tempera
ture, it decreases markedly with the temperature.
344 Introduction to Materials Science

In the case of semiconductors, the conductivity, which is zero at low tempera

ture (infinite resistivity), increases with in certain temperature ranges. Their con
ductivity thus generally evolves in a manner opposite to that of metals. This dif
ference in behaviour results from the fact that in metals, the number of mobile
electrons does not vary with the temperature; their conductivity decreases when the
temperature increases because the electron mobility is reduced. In the semiconduc
tors, it is the number of carriers, which increases with the temperature because an
increasingly large number of electrons are transferred from the valence band to the
conduction band by thermal activation.

14.2.2. Electrical conductivity of metals

To treat the electrical conductivity of metals rigorously, it is necessary to make
use of quantum theories and to apply Fermi-Dirac statistics to the electrons. This
takes account of their indistinguishability and the Pauli's exclusion principle that
strictly limit the number of electrons per energy level (chapter 2).
In this elementary volume, a simplified theoretical approach, developed at the
beginning of the 2 0 century by Drude and Lorentz, will be used. This theory con
th

siders an assembly of free electrons obeying classical statistics (Maxwell-Boltz-

mann). The electrons, which in this model are not influenced by the effects of the
crystallographic structure of the solid, are free to move within the geometrical limits
of the crystal in a uniform electrostatic field.
A representation of the physical model used is shown in figure 14.5. At ambient
temperature and in the absence of an external electric field, the electrons exhibit
random movements in the metal. The electric energy carried by the electrons is
distributed isotropically, because there is no particular preferred direction to their
movement.

C T ~ ~ TD C C CT~~~T

D er cA o y
w <j )
D d d
\> ^ (
q
\> <\ *" / \

s~s -%
_ _
-C

^
-
r*\

r\
.
^> ^
r\ r\ r\
d

(b) -

Figure 14.5. Elementary description of the electrical conductivity of metals by the

Drude-Lorentz corpuscular theory: (a) no electric field E-0trandom movements of the
electrons; (b) 0, oriented movements of the electrons in a direction of decreasing
electric potential (V 2 > V 1).
 Physical properties 345

As soon as an electric field (electric potential gradient) is applied to the ends of a

conductor, the displacement of the free electrons occurs preferentially between the
highest potential V and the lowest potential V\. For a movement through a distance
2

A JC, the potential difference V= V -V\ defines an electric field

2

E = - g r a d F = - AVIAx (14.4.)
The electric potential gradient is at the origin of the electric current of intensity
/, but its conventional direction is opposed to that of the preferential movement of the
electrons. Each electron carries an electric charge e = 1.6 10~ C (coulomb). The
( ) 19

free electrons resulting from the electric field are accelerated over a certain distance
until they are abruptly stopped by collision with the metal ions. This random and
repetitive process is illustrated in figure 14.6.
The mean drift velocity of the electrons is given by the relation:
vd = (14.5.)

In this expression, represents the mobility of the electron ( m V s ) . It is the drift

2 _1 _1

velocity of the electrons subjected to a unit electric field. At low temperature, it is an

indirect measurement of crystal perfection, since is greater when the long range
order in the crystal is high and the number of defects is low.
Taking account of the expression for the drift velocity and the volume density
of mobile electrons N , the electrical current density J (Am ) can be calculated:
e
-2

J = Ne v e
{r)
d (14.6.)

The electric current / in a conductor of section S is given by the relation:

I = JS (14.7.)

(a)
Time t

Figure 14.6. Variation of the drift velocity of electrons between the collisions with the
metal 'ions': (a) polycrystal; (b) monocrystal with very few defects.
346 Introduction to Materials Science

By combining (14.5.), (14.6.) and (14.7.) and taking into account that according to
(14.4), = V/L for a conductor of length Z, then:

I= Ne -^S
{
(14.8.)
e ~

Equation (14.8.) is an expression of Ohm's law (14.1.), the electric resistance of

the conductor being given by:

R=L _ ! _ (14.9.)
S N e ~^ e
(

Starting from (14.2.), the resistivity and the conductivity a of material can be
expressed as:

= -= ," (14.10.)

The quantum theories give an expression for a similar to that obtained from
the Drude-Lorentz theory. Classical theory does not take into account that the
majority of electrons are unable to absorb energy because of Pauli's exclusion
principle. The quantum theory shows that a small number of electrons, moving at
very high speed, is responsible for the high electrical conductivity of metals. This
quantum theory, which associates waves with electrons, shows that these waves are

0,1
Impurity weight %

Figure 14.7. Variation of electric conductivity in function of the impurity concentra

tion in copper.
 Physical properties 347

not attenuated in a perfectly periodic infinite system. It stresses that any disturbance
of the periodicity of the system gives rise to a diffusion of the electronic waves and
thus to a resistivity. The simplified development presented here nevertheless allows a
certain number of interesting remarks to be made.
In a metal such as copper, a great number of mobile electrons exists, because
each atom puts one electron at the disposal of the conductor, which gives a value of
N ~ 10 electrons m~ . The conductivity decreases in a real crystal as a function of
e
29 3

the number of defects, the number of impurities or alloying elements, which it con
tains (figure 14.7.). As shown in figure 14.4., the conductivity a, or its inverse the
resistivity p, varies considerably with temperature. At very low temperatures, the
resistivity remains practically constant. At somewhat higher temperature, it varies
as a function of , then as a function of T. This variation of the resistivity with the
5

temperature and the number of crystal defects, is expressed by Mathiessen's law

(figure 14.8.):
p = Pr + p, (14.11.)

Figure 14.8. Various contributions to the electrical resistance of a metal according to

Mathiessen's law.

The residual resistivity p depends on the degree of physical perfection and the
r

chemical composition of the sample. For a given metal, p can vary according to the
r

method of manufacture, the degree of purity, and the thermal or mechanical treat
ment. The value of p is determined by the defects in the crystal lattice and conse
r

quently in practice does not depend on temperature.

The intrinsic or ideal resistivity p, of pure metals is a function only of in
teractions of the electrons with the thermal vibrations of the lattice. As the latter are
strongly influenced by temperature, p, therefore varies considerably with tempera
ture. At high temperature, it is the ideal resistivity, which controls the electric be
haviour of a pure metal, while at low temperature, it is p , which predominates.
r
348 Introduction to Materials Science

Generally, the electrical conductivity of metals decreases with because the

frequency of collisions between the electrons and the ions in the lattice increases
with the amplitude of the vibrations of the lattice. In practice, at high temperatures,
the resistivity varies linearly with temperature (figure 14.8.):
p = Pb(l +ccAT) (14.12.)

In this expression i s the resistivity at the reference temperature T . For many pure
0 0

metals between 0 and 250 C, the coefficient a is of the order of 4-10" K" . This 3 1

corresponds to an increase in the number of collisions by a factor 2 over a tem

perature interval of 250 C.
In the case of alloys, the coefficient a is smaller, because the atoms in solution
disturb the mobility of the electrons. Certain alloys have a very low coefficient a.
For example, constantan (55 wt% Cu by weight, 45 wt% Ni) has a coefficient
a = - H 0 K" . This type of alloy is used in the manufacture of heating elements and
- 5 1

for certain resistances in electronic circuits.

The resistivity of good conductors, such as copper or silver, remains practi
cally constant at temperatures below 10 (-263 C). For a certain number of metals,
intermetallic compounds and ceramics known as superconductors (table 14.9), a
critical temperature T exists below which the mean free path of the electrons be
c

comes infinite and the electrical resistance zero (figure 14.4.).

Table 14.9. Values for the critical temperatures T

c

of certain superconductors.

Metals, compounds T (K)

C

YBa Cu 07-x
2 3
92
Nb Ge
3 23,4
Nb Sn
3 15
Nb 9,25
Pb 7,2
Hg. 4,15
Sn 3,7
Al 1,18
Zn 0,85
Zr 0,6
Ti 0,4

Thus an electrical current induced in a superconductor, in the absence of

magnetic field, can be maintained almost indefinitely. It can be deduced from this
type of experiment that the conductivity a of this type of material is greater than
10 m .
This value is 10 times greater than that of pure copper at low temperature. The
17

critical temperatures of the series of superconductors discovered at the beginning of

the 2 0 century were too low for many technical applications (table 14.9). Recent
th

extraordinary developments have revealed the possibility of producing materials

which are superconducting at relatively high temperatures, something which might
open the way for exceptional industrial progress.
 Physical properties 349

14.2.3 Semiconductors and semimetals

In fact, from the point of view of electrical conductivity, only two categories of
materials exist: metals which have an incomplete valence band and which conduct
electricity at zero degree Kelvin and insulators with a complete valence band which
do not. At higher temperatures, the situation is however less clear-cut. In order of
decreasing electric conductivity, two intermediate categories can be identified
namely semimetals and semiconductors.
In a semimetal (Bi, graphite, Sb, As), the valence band and the conduction band
overlap slightly and at zero K, a small number of electrons are already in the con
duction band. This situation produces a low density of positive holes in the valence
band and induces a low electrical conductivity at zero K. With temperature, the num
ber of electrons in the conduction band increases, and consequently the number of
free electrons. At ambient temperature, in the semimetals this varies between about
10 and 10 conducting electrons per c m whereas a typical value for a metal such
17 21 3

as copper is close to 10 .
23

It has already been mentioned that there was no qualitative difference between
an insulator and a semiconductor: these materials have a complete valence band and
an empty conduction band, and what differentiates them is the width of the forbidden
band E (table 14.10.). In insulators like diamond, the energy gap is close to 5.5 eV.
D

Under these conditions, conductivity is zero at ambient temperature. When the

energy gap is of the order of 1 eV, a certain electrical conductivity is observed at
ambient temperature. Although this energy is much larger than the average thermal
energy (kT ~ 0.025 eV at ambient temperature), it is sufficiently low to allow, statis
tically, certain electrons to pass into the conduction band. In the case of the semicon
ductors, the number of free carriers at ambient temperature is notably lower than in
the case of semi-metals; it varies from about 10 to 10 conducting electrons per m .
19 23 3

In practice, electric conductivity depends on the density of carriers and their

mobility. In the case of a typical semimetal such as graphite, the density of the free
carriers is approximately four orders of magnitude lower than that of copper whereas
the electric conductivity at ambient temperature is only two orders of magnitude
smaller. This results from the higher mobility of the free carriers in graphite.
In metals, the number of carriers is very high and constant but their mobility is
low; the variation of with the temperature is only the result of a variation of the
electron mobility. In the semimetals and the semiconductors, the electrical conduc
tivity increases with temperature in certain temperature ranges because the number of
carriers is increased. The electrical conductivity of semimetals and semiconductors is
strongly influenced by the presence of impurities, which increases the number of
carriers. As it can be seen in table 14.10 for tin, the metallic or semiconductor
character of an element varies with its crystal structure. At ambient temperature, the
electrical conductivity of the typical semiconductors remains low (Ge) and even very
low (Si). To increase this, it is necessary to add very small quantities (doping) of
elements belonging to chemical groups close to that of the semiconductor in Mende
leev's periodic table. In the case of silicon or germanium, elements of the group V A
(P, As, Sb) or III A (B, Al, In, Ga) are used for doping.
When an electron passes from the valence band to the conduction band by ther
mal excitation in an un-doped semiconductor (intrinsic semiconductor), simulta
neously an electron (negative) and a positive hole appear. The hole will tend to col-
350 Introduction to Materials Science

Table 14.10. Some properties of the Group IV elements at ambient temperature

Elements ED Conductivity Crystal system

(eV) electric

Diamond (C) 5,5 < 10~ 16

cubic (diamond)
Silicon (Si) 1,17 5 -4
cubic (diamond)
Germanium (Ge) 0,75 2 cubic (diamond)
Grey tin (Sn) 0,09 310 5
cubic (diamond)
White tin (Sn) 0 HO 7
tetragonal
Lead(Pb) 0 5-10 6
fee
The values given for the conductivity of semiconductors, very sensitive to the
presence of impurities, constitute only an approximation of the maximum val
ues [Compiled after Ralls, Courney and Wulff; Kittel; Gerl and Issi]

lect an electron from a nearby atom, which is equivalent to the fictitious movement
of a positive charge (figure 14.11. (a)). Two types of charge thus convey the electri
cal current in an intrinsic semiconductor.
By addition to silicon (doping) of a small quantity of a Group V A element, such
as phosphorus P, which has one valence electron more than Si, a surplus electron is
introduced (figure 14.11(b)). This is situated at an energy level slightly lower than
that of the conduction band; only a slight thermal excitation is required to transfer it
into the conduction band and to make it mobile and able to convey the electrical
current. An extrinsic group -type semiconductor (n = negative (electron)) has thus
been created.
By introducing 10 ppm (1 ppm = 1 0 = 10" %) of phosphorus into silicon,
-6 4

approximately 10 mobile electrons per m of the semiconductor are obtained. Al

23 3

though this number is very large, it is nevertheless 10 times smaller than that for the
6

number of mobile electrons in metals. This explains why the electric conductivity of
doped semiconductors remains considerably lower than that of metals, although the
electron mobility in a semiconductor is generally higher than in a metal (typically by
1 to 2 orders of magnitude).
By introducing into silicon small quantities of an element like aluminium, which
contains one electron less than silicon, acceptor centres having an energy slightly
higher than that of the valence band (figures 14.11. (c) and 14.12. (c)) are created.
These are able to accommodate electrons coming from silicon atoms belonging to the
basic lattice. Only a slight thermal activation is thus required to create electron gaps
in the basic silicon lattice. The acceptor centres (aluminium atoms) then become
negatively charged.
A positive charge, resulting from the displacement of an electron of a silicon
atom to an aluminium atom, moves progressively from one silicon atom to another
thus creating a displacement of electron holes within the semiconductor; this gap
displacement is equivalent to a displacement of electrons in the opposite direction. In
this case, electrical conductivity is ensured by the movement of electron holes (figure
14.11. (c)) and type ^-semiconductors (p = positive holes) are referred to.
Chemical compounds combining elements of group III and V and forming a
diamond-type crystal structure can also exhibit semiconducting properties. ZnS, Ga
P, Ga As, In can be quoted as examples.
 Physical properties 351

(a)

Electron path
Vacancy path

Figure 14.11. Simplified diagrams of: (a) an intrinsic semiconductor (b) an extrinsic
semiconductor doped with phosphorus (-type); (c) with aluminium (p-type).

Figure 14.12. Approximate energy levels of: (a) an intrinsic semiconductor (Si forbidden
band: 1.1 eV); (b) -type extrinsic; (c) -type.

As seen in figure 14.13, at low temperature, the electric resistivity of extrinsic

semiconductors initially decreases when increases, because the donor centres (or
acceptors) will release their electrons (or holes) until all the donors (or acceptors) are
ionised (extrinsic region). At the intermediate temperatures, the number of carriers is
constant, but increases as in metals because of a reduction in the mobility of the
carriers with the temperature (depletion zone). At high temperature, decreases
again because the thermal energy is sufficient to directly excite the electrons of the
valence band into the conduction band (intrinsic region).
This basic presentation has been limited to the study of thermally activated
electrical conductivity. The electric conductivity of semiconductors and insulators
can also be activated by electromagnetic radiation, a phenomenon referred to as
photoconductivity.
352 Introduction to Materials Science

Figure 14.13. Variation of the resistivity in an extrinsic semiconductor.

14.3. Thermal conductivity

14.3.1. T h e r m a l conduction
It is possible to have an intuitive idea of the phenomenon of thermal conductiv
ity when a block of metal is cold to the touch, whereas an object in polymer or wood
seems warmer, although all three are at the same ambient temperature. Metal, being a
better conductor than plastics or wood, removes the calories much more efficiently,
leading to our organism experiencing the sensation of cold.
The phenomenon of thermal conductivity plays a very significant role just as
much in technological applications as in everyday life. Thus, many technical proc
esses involve high temperature and require significant heat transfer by means of heat
exchangers, which must be made from materials having the highest possible thermal
conductivity.
In other cases (dwellings, cryogenic applications), it is, on the contrary, neces
sary to minimise the calorific losses. Currently, this phenomenon is becoming in
creasingly important because of the high cost of energy.
If it is easy to experience the phenomenon of thermal conductivity, its physical
interpretation nevertheless remains a very difficult problem. This discussion will be
limited to its most significant aspects.
Heat exchangers make use of three types of energy transfer phenomena: con
duction, convection and radiation. Only conduction will be considered here, because
convection is non-existent in solids, and heat transfer by radiation is only relevant to
transparent materials at high temperature.

14.3.2. Macroscopic aspects of conduction: Fourier laws

Fourier described the essentials of the macroscopic aspects of thermal conduc
tivity in 1822. He was the first to express precisely the proportionality between the
heat flux Q and the temperature gradient d 77 d x .
 Physical properties 353

When there is a difference in temperature between the two ends of a conductor,

separated by dx:, a heat flux Q is established which is analogous to an electric current
resulting from a potential difference. In the case of a unidirectional thermal flow in
an isotropic material, the flux Q expressed in Wm" and the temperature difference
2

d r a r e related as follows (Fourier's first law):

Q=- K (14.13.)
ax
The proportionality coefficient relating the heat flux Q and the temperature gra
dient is the thermal conductivity coefficient ( W m ^ K ) . The negative sign in the
-1

expression (14.13.) indicates that heat is transported from hot to cold. This expres
sion is similar to Fick's first law (chapter 9.).
Fourier's first law only applies if the temperature gradient remains constant over
time. If the temperature gradient is not constant, Fourier's second law gives the
variation of the temperature at a point in the conductor as a function of time:

f!I = D (14.14.)
dt 1
dx2

The coefficient of thermal diffusion D in the expression (14.14) is related to thermal

T

conductivity by equation:

(14.15)

In this expression, C is the heat capacity per unit mass (specific heat) at constant
p

pressure and the density.

The expression (14.14.) gives the rate at which a planar thermal wave is propa
gated in one direction through an isotropic conducting substance. Fourier's second
law is similar to Fick's second law and the same type of mathematical solutions is
applicable in both cases. In the form exposed here, Fourier's laws are valid only for
isotropic materials. In the case of an anisotropic material, the heat flux is no longer
necessarily parallel to the temperature gradient, and Fourier's laws, like Fick's laws
take a tensorial form.
In a large number of cases an analogous behaviour exists between thermal con
ductivity and electric conductivity: metals are good thermal conductors and the ma
jority of electrical insulators (polymers, ceramics) are also poor thermal conductors.
There are nevertheless a certain number of exceptions: electrical insulators like dia
mond, boron nitride or highly oriented polyethylene fibres, conduct heat better than
metals at ambient temperature.
However, thermal conductivity varies less than electrical conductivity. Whereas
the values of the latter are spread over more than 24 orders of magnitude between
conductors and insulators, this is not the case for thermal conductivity. Figures 14.14.
and 14.15. illustrate the ranges of thermal conductivity encountered at ambient
temperature for various materials, from the best thermal conductors (-2000 Wm" 1

K" ) such as copper, diamond and graphite (in the direction parallel to the hexagonal
1

plane) to the best insulating materials (10~ W m K ) which are heterogeneous

5 _ 1 _ 1

solid-gas systems.
354 Introduction to Materials Science

10

-I2

Homogeneous
1
crystalline
Glasses and solids
e 10 ceramics (figure 14.15)

-1 Glass and
> ceramic
Foams, powder foams Polymers
3
"" 2
and powders
Powders
10' 3

Opaque Glass fibers

powder
" 4

Multilayered
isolation
Systems under vacuum

Figure 14.14. Outline of the thermal conductivity of various dense, porous and divided
materials (after Issi, 1981).

Metals Ceramics Polymers

10 j

Diamond
10"

Cu,Ag
A1,W Graphite
Zn
Fe
Cr
Pb, Ni, Steel A1 0
2 3

Stainless steel
G

Mineral glass
Concrete
PE
PA, PTFE

10 1 Wood, mbber,PS

Figure 14.15. Table showing the values of thermal conductivity for various types of
materials in solid, non divided form.

It should be noted that the thermal conductivity of dense crystalline solids

(figure 14.15.) varies relatively little (three to four orders of magnitude). By making
use of organic materials (semicrystalline or amorphous) and especially porous
materials (fibres, foams, powders), it is possible, with certain techniques, to gain a
further three orders of magnitude. Extracting the gas phase from porous materials
enables three more orders of magnitude to be gained.
 Physical properties 355

14.3.3. Mechanisms of thermal conduction

Thermal conductivity is calculated starting from the kinetics of gases:

= Levi (14.16.)
3 v

In this formula, c represents the heat capacity at constant volume and v, the velocity
v

of the heat conducting particles and /, their mean free path. This formula applies
equally in the case of electronic thermal conductivity and the case of phonon thermal
conductivity of the lattice.
As previously seen (figurel4.3.), electrical conductivity is directly proportional
to the number of mobile electrons present in materials. In the case of metals, which
have a density of mobile electrons greater than 10 nf , thermal conductivity is also
29 3

proportional to the same factor. In this case, almost exclusively the electrons convey
thermal energy.
As already seen, solids exist which are known to be electrical insulators but
which conduct heat as well as metals. This observation leads to the idea that there is
at least one other mechanism for thermal conduction besides those involving mobile
electrons.
The mechanism that is mainly for transport of heat in electrical insulators uses
the vibrations of the crystal lattice. In this case, lattice thermal conductivity or ther
mal conductivity by phonons is referred to.
In a solid, the atoms vibrate constantly about their equilibrium position and the
amplitude of the vibrations increases with the temperature. If thermal energy is intro
duced at one end of a crystal, the amplitude of the atomic vibrations increases at this
end. Because of the interatomic forces ensuring the crystal cohesion, any movement
of an atom influences its immediate neighbours and the movement is transmitted
gradually in the form of a wave in the direction of the heat gradient. It is by this
mechanism that heat is carried in electrical insulators.
In the language of quantum mechanics, each vibration of the crystal lattice can
be described as a travelling wave transporting energy. In the same way that particles
(photons) are associated with light waves, phonons are associated with acoustic
waves.
When the temperature rises, phonons are emitted following the increase in the
amplitude of the vibrations. A temperature difference induces a concentration
gradient of phonons and a greater number of them move from the hot source to the
cold source.
This particulate image makes it possible to apply the kinetic theory of gases to
phonons. The variation of phonic conductivity with temperature is illustrated in fig
ure 14.16 where it can be seen in particular that the thermal conductivity of diamond
passes through a maximum at approximately 100 K. The increase of the thermal
conductivity of diamond at low temperatures results from the increase in the heat
capacity of the phonons. The decrease observed at higher temperature results from a
reduction in their mean free path due to the mutual interaction of the phonons, the
heat capacity of the phonons remaining more or less constant at high temperature.
Although proceeding by a completely different mechanism, electronic thermal
conductivity, as with copper, also passes through a maximum at a temperature close
to 10 (figure 14.16.). At low temperatures, electronic thermal conductivity in-
356 Introduction to Materials Science

creases with because the heat capacity of electron gas rises with temperature and
the mean free path remains constant. At high temperatures, conductivity is constant
because of the reduction in the mean free path of the electrons.

Absolute temperature (K)

Figure 14.16. Dependence of thermal conductivity on temperature for: (a) a metal; (b) a
crystalline ceramic (diamond); (c) an amorphous ceramic (S1O2) (after Issi, 1981).

It should be noted that there is a very clear difference between the temperature
dependence of electric conductivity and electronic thermal conductivity. This results
from the fact that the electric charge transported by the electron does not vary with
the temperature, whereas the thermal energy that can be exchanged and transported
by the electron is a function of the temperature.
The mean free path of the phonons is like that of the electrons, limited by any
thing that disturbs the periodicity of the crystal lattice: alloying elements, impurities,
specific defects, dislocations, grain boundaries, other phonons. The limit is reached
for amorphous materials (figure 14.16) and generally glasses have a thermal conduc
tivity significantly lower than that of the crystals. The study of amorphous materials
up to now has been limited to the cases of ceramics and polymers. In practice, in
these materials, only phonic conductivity mechanisms are involved. In general, or
ganic polymers are poor thermal conductors (~ 0.1 to0.7 Wm^K" ). An exception 1

to this behaviour are highly oriented polyethylene fibres which have recently been
shown to have, in the direction parallel to the fibres, a thermal conductivity of about
100 Wm^K" i.e. more than 100 times greater than that of solid polyethylene. This
1

significant discovery opens the way to organic materials that are both electrical in
sulators and thermal conductors. If highly insulating materials are required, porous
materials should be chosen, because the air present in the pores has a very low ther
mal conductivity since there is practically no convection if the diameter of the pores
is lower than one millimetre. In high temperature furnaces, porous ceramics obtained
 Physical properties 357

by sintering are used. Vacuum is an excellent insulator because thermal energy can
only be transmitted through it by radiation. This type of thermal insulation is used in
cryogenics for the containers (thermos flasks) for liquefied gases. Synthetic foams
made from organic polymers are also insulators very much used in cryogenics.

14.4. Magnetic properties

14.4.1. Introduction
The magnetic properties of materials play a significant role in machinery and
electrical equipment such as motors and transformers. Their role is dominant in in
formation storage (recording tapes, computer hard disks), which is carried out by use
of a substrate having specific magnetic properties. In this section, an outline of the
principal magnetic properties of materials will be given.

14.4.2. Definitions
The magnetic induction is the internal magnetic field induced in a substance
by an external magnetic field H. Its value is given by:
= (14.17.)

is a constant of proportionality called magnetic permeability. The unit of magnetic

induction is the tesla (T). The magnetic field is in general created by the circula
tion of an electrical current in a solenoid. It is expressed in ampere per meter ( A m ) -1

and the magnetic permeability in Henry per m (H m ) . -1

To characterize the magnetic behaviour of materials, the relative permeability

r is often used:

= / 0 (14.18.)

is the magnetic permeability of vacuum ( = 410~ H m " ) . In the absence of

0 0
7 1

material, the value of magnetic induction B is given by:

0

B 0 = (14.19.)

Magnetic induction in material is also expressed by:

= ( 0 + ) (14.20.)

is the magnetization vector, which represents the local magnetic field induced
by the external magnetic field H. The magnetization vector is directly proportional
to and it can be written:
= %H (14.21.)
and

= 0 (1 + X r ) (14.22.)

By comparing (14.17) and (14.22) the following expression is obtained for the
magnetic permeability:
358 introduction to Materials Science

= (1 0 + Xr) (14.23.)

represents the relative magnetic susceptibility. This parameter, which is dimen-

sionless, is related to the relative permeability by: r

(14.24.)

14.4.3. Magnetic classification of materials

As emphasized in chapter 2., the circulation of an electron in its orbit (Bohr-
Sommerfeld theory) can be compared to a circulating electrical current. The circula
tion of the electron generates a magnetic moment perpendicular to the plane of the
orbit. The intensity of this magnetic moment is measured by the magnetic quantum
number mi. In the same way, the spin of the electron itself also generates a magnetic
moment, which is quantified by the magnetic spin number m , which can take two
s

distinct values (+0.5 and -0.5). The magnetic moment is parallel to the magnetic
field when m = +0.5 while it is oriented in a direction opposed to that of the ma
s

gnetic field if ms = - 0 . 5 .
When the electron shells are filled, the magnetic moments of the electrons are
mutually compensated and the resulting magnetic moment is zero. In the solid state,
only atoms having an incomplete internal electron shell possess a permanent mag
netic moment, because the external electron shell (valence electron) is completed by
the valence electrons of the neighbouring atoms.
Among the elements having an incomplete electron shell, the first series of tran
sition metals and in particular iron, cobalt, and nickel constitute the most significant
magnetic materials (figure 14.17.).
On the basis of their magnetic behaviour, materials are divided into five groups:

diamagnetic,
paramagnetic,
ferromagnetic,
antiferromagnetic,
ferrimagnetic.

Diamagnetic materials are composed of atoms having all the electron shells
filled. They do not have permanent magnetic moment. Their magnetic susceptibility
is low, negative ( ~ -10" ) and independent of the temperature. The magnetization

6

vector is in the opposite sense to that of the magnetic field vector H. A certain
number of metals (Cu, Ag), the non-metals and the majority of organic compounds
are diamagnetic.
The paramagnetic materials are characterized by a magnetic susceptibility % in r

the range 10~ - 10~ . This behaviour is encountered in substances having permanent
6 3

magnetic moments, which are not mutually coupled. Under the action of an external
magnetic field, the moments tend to align and reinforce the influence of the external
field (M a n d / / are of the same direction). However, the value of magnetic suscepti
bility % remains small because thermal agitation constantly reorients the moments
r

(figure 14.18. (a)).

 Physical properties 359

Elements 3.s 3/> 3d 4s

Fe 26 |TU TU 4 TU 4 t TU
Co 27 TU 1 4 4 4 4 4

Ni 28 TU T4T4T4 TU T i T U ; 4

Cu 29 TU T i t i T I 4 4 4 4 4- t i
Figure 14.17. Electron states in the 3 and 4 electron shells of iron, cobalt, nickel and
s

copper.

The rise in the temperature tends to disturb the organization of the dipoles and,
with some exceptions, (U, Ti), magnetic susceptibility decreases with following
Curie's law:

(14.25.)

C is a constant. A certain number of metals and ceramics are paramagnetic, for
example, fee iron (y-Fe - austenite).

/ \ \ t t t t t t
\ / -~ \ t t t t t t t
t \ \ / \ t t t t t t
/ \ \ \ t t t t t t
(a) Paramagnetic (b) Ferromagnetic

t \ \ t \
\ t \ t \ \
\ \ t \
1
t t \ t \ \
(c) Antiferromagnetic (d) Ferrimagnetic

Figure 14.18. Magnetic arrangement dipoles in a paramagnetic, ferromagnetic, antifer

romagnetic and ferrimagnetic crystal.

Three types of magnetic materials (figure 14.18. (b) to (d)) are characterized by
the presence of intense magnetic dipoles having a high exchange energy (between
neighbouring dipoles) relative to the thermal agitation energy. In these materials, the
dipoles tend to be aligned specifically along certain crystallographic directions to
form magnetic domains (Weiss domains).
If all the dipoles are aligned in a parallel manner in the different domains, a
ferromagnetic material is involved. Up to a given temperature (the Curie tempera
ture, ) the magnetic permeability remains very high ( 1 0 < % < 10 ) and the 3
r
6

magnetic induction is considerably reinforced inside these materials by the

360 Introduction to Materials Science

magnetization vector M. When > , the coupling forces between dipoles are no
longer sufficient relative to the thermal agitation energy to maintain the dipoles
aligned in magnetic domains and the material acquires a paramagnetic behavior. The
Curie-Weiss law expresses this behavior:
~ = C (14.26.)
-
At = , %r becomes practically infinite and the material takes on a ferromag
netic behavior. The most significant examples of ferromagnetic materials are bcc
-iron (ferrite), cobalt and nickel. The Curie temperature of pure -Fe is equal to
770 C.
In antiferromagnetic materials the atomic magnetic moments are equal and
arranged in an antiparallel mode These materials never have a permanent magnetic
moment and their susceptibility, which is low, increases with up to a critical
temperature (Neel temperature). Above this temperature, the domains are de

stroyed by thermal agitation and the material becomes paramagnetic ( is weak and

decreases with T). Chromium and manganese have antiferromagnetic behavior.

When the magnetic moments of the atoms are unequal and form domains where
they are aligned in an antiparallel way, ferrimagnetic behaviour is referred to. The
permanent magnetic moment of the ferrimagnetic crystals is not zero and their mag
netic susceptibility remains high up to their Neel temperature above which these

materials acquire a paramagnetic behaviour.

Ferrimagnetism is the magnetism of a particular class of ceramics called ferrites
which are oxides made up of bivalent and trivalent cations. These materials, which
have varied chemical compositions, are characterized by a cubic or hexagonal struc
ture (these ferrites should not be confused with the -iron (ferrite) bearing the same
name).
Ferrites, which are electrical insulators, have a high magnetic susceptibility. A
very well known example is that of magnetic iron oxide ( F e 0 ) or (FeO)(Fe20 )
3 4 3

consisting of F e and F e ions and which is a natural ferrite. In a similar way,

2+ 3+

( N i O ) ( F e 0 ) is a ferrite. Certain ferrites are used as permanent magnets. Ferrites

2 3

have a significant technological role in data processing, in particular as materials for

data storage. A hard disk or a magnetic tape consists of an inert polymer support on
to which ferrite powder is attached. Given the technical importance of ferromagne-
tism, a section will be devoted to it.

14.4.4. Ferromagnetic behavior: hard and soft magnetic materials

Ferromagnetic materials are permanent magnets or solids attracted by
permanent magnets. Their magnetic susceptibility is very high, up to 10 , in

6

materials used in electrical engineering for the cores of electromagnets.

Permanent magnets are solids having a permanent macroscopic magnetic mo
ment. A permanent magnet consists of atoms, some of which have a permanent mag
netic moment.
There are relatively few ferromagnetic materials (Fe, Co, Ni; alloys of Co - rare
earths). In these materials, the atomic magnetic moments are aligned over long dis
tances and form domains of the same orientation, known as Weiss domains, sepa
rated by interfaces called Bloch walls (figure 14.19.). This concept of a magnetic
 Physical properties 361

domain separated by walls is fundamental for understanding the behaviour of soft

and hard (permanent) magnets.
The magnetisation of a ferromagnetic material involves an orientation of the
elementary magnetic dipoles in the same direction. It is generally carried out under
the influence of an external magnetic field. If all of the magnetic dipoles of material
are oriented in the same direction, the material is magnetized to saturation. This
situation corresponds to a high-energy state, because the induction lines leave the
crystal to close the magnetic circuit (figure 14.19. (a)). A subdivision into domains of
various orientations is energetically more favourable, because it leads to containment
of the magnetic field inside the crystal. The minimum size of the domains is however
limited because of the increase in energy associated with the formation of the Bloch
walls. In the optimum situation, the energy gain associated with the formation of
small domains is compensated by the energy loss related to the formation of the
walls.

(a) (b) (c)

AI' " 1

titttititnttttt'*

Figure 14.19. Weiss domains and Bloch walls of a ferromagnetic material: (a) one Weiss domain;
(b) two Weiss domains; (c) four Weiss domains forming a closed magnetic circuit; (d) Bloch wall
separating two Weiss domains with progressive reorientation of the magnetic moments.

The dimensions of Weiss domains are in general less than 100 m. They can be
observed by various microscopic methods (optical or electronic).Their dimensions
vary from one material to another. They are modified under the action of an external
magnetic field. Their behaviour makes it possible to make the distinction between a
soft magnetic material (non permanent magnet) like pure iron, and a hard magnetic
material (permanent magnet) like the tempered steel and the intermetallic compound
Co Sm. When a ferromagnetic material is subjected to the action of an external mag
5

netic field there is a preferential and progressive development of Weiss domains in

a direction close to the direction of H. This growth of Weiss domains occurs by pro
gressive orientation of the dipoles and elimination of the walls (figure 14.20., points
1, 2, 3, 4). Finally, the orientation of the internal magnetic field becomes parallel to
the applied field. The magnetic induction is at saturation and the curve -
reaches a plateau, because there are no more domains to orient ( = 0). Only
one Weiss domain is found (figure 14.20., point 5).
362 introduction to Materials Science

If the applied field is gradually decreased, disorientated Weiss domains form

again and the material tends to return to its initial magnetic structure. This reorienta
tion is accompanied by the creation and the displacement of Bloch walls, which leads
to a reduction in magnetic induction B.
When, after saturation, the magnetic field is eliminated, in general a certain
preferential orientation remains conferring on the material a permanent magnetic
moment. The value of the residual internal magnetic field B is called the magnetic
r

remanence (figure 14.21, point 3). When the aligning field is removed, it is neces
sary to apply a magnetic field of opposite direction called the coercive field H (fig c

ure 14.21, point 4) to eliminate the effect of the magnetic remanence B . The mag r

netic induction = 0, for = H . c

Figure 14.20. Curve of magnetic induction as a function of the applied magnetic field
(points 1 to 5). Progressive elimination of the Bloch walls to reach the state of satura
tion.

By increasing the negative field beyond H , the ferromagnetic material is satu

c

rated in the other direction. By decreasing the field in the negative direction and by
then increasing it in the positive direction (figure 14.21, items 5, 6 and 2), a hystere
sis loop is formed. The part > 0 and < 0 is also called the demagnetisation
curve. This is an important characteristic of permanent magnets for defining their
magnetic hardness. Any variation of induction in a magnetic material causes, in the
interior, a dissipation of energy, usually in the form of heat, which is an irreversible
process (magnetic loss).
The surface enclosed by the hysteresis loop is a measure of the energy dissipated
by the material during one cycle. The soft magnetic materials have a narrow hys
teresis loop and a high magnetic induction (tab. 14.22 and figure 14.23, inner loop).
This combination of properties allows the creation of a significant magnetic flux
while minimising the energy losses in an alternative field.
The sheet of a transformer core is made from a soft magnetic material, which
easily follows the changes of the magnetic field resulting from the alternating char
acter of the electrical current. The hysteresis loop of this sheet is narrow. The mag
netic loss leading to noise and heating of the transformer is low and the reduction in
the transformer output is minimised.
 Physical properties 363

Figure 14.21. Variation of the magnetic induction as a function of the applied field H.
The complete curve is called a hysteresis loop. B and H are respectively the magnetic
r c

remanence and the coercive field.

Table 14.22. Hysteresis loss and magnetic susceptibilities

of soft magnetic materials.

Materials Hysteresis loss per cycle * atH

r

(JnT )3

-> 0

Commercial iron 500 250

Sheet (Fe-Si 4 %) non oriented 50-150 500
Permalloy 45 (Ni 45 % - Fe 55 %) 120 2700
Mu metal (Ni 75 %; Cu 5 %; Cr 2 %; Fe 18 %) 20 30000
Supermalloy (Ni 79 %; Fe 15 %; Mo 5 %; 0,5 %) 2 10000
Ferroxcube A (ferrite) (MnFe 0 + ZnFe 0 )
2 4 2 4 40 1200

If the material should maintain a high degree of permanent magnetization (per

manent magnet), internal friction must be as high as possible to prevent the Weiss
domains from reforming in a disordered way after saturation. This is achieved by
creating many obstacles in the microstructure, for example precipitates to block the
movement of the walls. This phenomenon of magnetic hardening is analogous to the
phenomenon of mechanical hardening of a crystalline material (chapter 12) which is
obtained by creation of obstacles to the displacement of dislocations.
In general, the permanent magnets are mechanically hard but not very tough
materials. Hard magnetic materials are characterized by a broad hysteresis loop and
high values of B and H . There are other mechanisms of magnetic hardening. For
Y c

example it is possible to split up the magnetic material into sufficiently fine particles
such that each one constitutes a unique Weiss domain. These magnetic powders are
then re-agglomerated by sintering or using an organic binder.
The value of the product BxH varies along the demagnetization curve (figure
14.23) between B x0 and OxH passing through a maximum value ( J ? x i y ) m . The
r c ax
364 Introduction to Materials Science

value of (B x / Z W is a measure of the stored energy in a permanent magnet. The

product ( x H)m&* of ferromagnetic materials increased considerably since the be
ginning of the 20 century. As will be seen in the illustrative example of this chapter,
th

these materials have become essential for advanced technologies. The ferrimagnetic
materials have magnetic properties analogous to those of ferromagnetic materials:
soft and hardferrimagnets exist.

Figure 14.23. Curves of magnetic induction as a function of the applied field H. Small
loop = soft magnet; large loop = permanent magnet. The value of ( 2 * x / / ) characterizes
m a x

the hardness of the magnet.

14.5. Optical properties

14.5.1. Introduction
This section is devoted to the study of the optical properties of materials. It
will be limited to an elementary treatment of the effect of visible light to interpret
certain aspects of materials such as transparency, brightness, and colour.
The optical properties of materials result from the interaction of light with
the electrons. Luminous radiations are electromagnetic waves characterized by their
wave length , their frequency or their energy. The quantum approach shows that
light can also be considered as an emission of photons which are particles of given
energy:
E=hv = hc/X (14.27.)

c and h represent respectively the speed of radiation and Planck's constant.

Figure 14.24 illustrates the spectrum of sunlight. This is much more extensive
than the spectrum to which the eye is sensitive which extends from 700 nm (limit
between red and infra-red) and 400 nm which is the part of the spectrum marking
the border between violet light and ultraviolet radiation. The sensitivity of the eye is
 Physical properties 365

Solar spectra measured on the earth

Eye sensitivity^^.

1000
Wave length ( n m )

Ultraviolet Infrared
Visible

Violet Blue Green Yellow Orange Red

400 450 ) 550 600 700
Wave length (nm)

750 700 650 600 550 500 450

Frequency ( 1 0 1 2
Hz)
I
2,75 2,5 2,25 1,77
Energy (e V )

Figure 14.24. The spectrum of sunlight and the corresponding curve for the sensitivity of
the eye. A light will appear white to us if its spectrum matches that of sunlight, (after
Nassau, 1980).

maximum at 560 nm (yellow-green light) corresponding to the most intense

radiations of the solar spectrum.

14.5.2. Interaction of light with solids

When a light ray passes through a transparent piece of material, various phe
nomena occur as indicated in figure 14.25. At its entry, the beam is refracted i.e.
deviated from its initial trajectory. During its passage through the transparent piece, a
fraction of the light intensity is absorbed. On coming out, the light ray undergoes a
second refraction opposite to the first and it follows its initial trajectory if the envi
ronment is the same. The sum of the transmitted ( I ) , absorbed (I ) and reflected (IR)
T A

intensities must be equal to the initial intensity I . 0

I = I + IA + IR
0 T (14.28.)

A solid is transparent if it transmits visible light with a weak absorption. In an

opaque solid the luminous radiation is not transmitted. A metal piece is an example
of an opaque object. If the surface of metal is polished, it reflects the incident radia
tion (mirror).
366 Introduction to Materials Science

Transparent solids generally consist of only one amorphous, optically homoge

neous (mineral or organic glass) phase or of a monocrystal (scratch-resistant watch
glass made from synthetic sapphire). When the materials have a microstructure
(grains, microcrystalline phases, precipitates, pores) of dimensions comparable with
the wavelength of the light, they become translucent. This results from variations in
the refractive index of the various phases leading to a diffusion of the light. The
semicrystalline organic polymers such as polyethylene and partially crystallised min
eral glasses are translucent, as well as polycrystalline ceramics.
There are several ways to measure the transparency of glass. In particular there
is the optical density (O.D.) which is equal to the logarithm (base 10) of the ratio hl I
of the incident and transmitted light intensity:

Figure 14.25. Interaction of a light ray with a plate of transparent material.

O.D. = log (V7) (14.29.)

Optical fibre technology introduced the use of the decibel (dB) related to the
optical density by:

1 dB = _L(O.D.) (14.30.)
10
The transparency of glasses is influenced by the presence of colour centres
(traces of transition metals) absorbing part of the light spectrum. Local variations of
the refractive index due to impurities diffuse glow and also reduce the transmitted
intensity.
Transparency is a very important property for materials used in optics and
communications technology. This is the case for optical fibres where the
transparency plays a key role in the performance of this type of material. For an
optical fibre to be useful, at least 1 % emitted light intensity has to be transmitted to
the receiver. Losses in optical fibres are generally given as decibel per kilometre
(dB k m ) . An optical loss of the order of 10 dB km" corresponds to a transmitted
-1 1

intensity of 1 % over distance of 10 km. Recent development on the transmission of

data by optical fibres opens new prospects, because only one optical fibre of 10
of diameter is able to convey simultaneously several thousand telephone calls.
The phenomena of refraction and reflection, which are at the base of traditional
optics and which are outside the scope of this book will not be covered. The focus
will be especially on the phenomena of absorption, which influence the transparency
of materials. These absorption phenomena vary considerably with the wavelength of
 Physical properties 367

luminous radiations. This will allow an understanding of why an object is coloured

yellow, orange, red or blue and why certain materials are transparent or opaque.

14.5.3. Absorption and emission of electromagnetic radiation

To explain the mechanism of absorption and emission of electromagnetic radia
tion, the case of a monoatomic gas made up of sodium vapour has been chosen as an
example. The atom of sodium contains eleven electrons. Among these, ten belong to
a complete shell and are strongly bound to the core. Only the valence electron is
sufficiently labile to interact with a luminous radiation. The energy levels of the atom
of sodium are shown in figure 14.26. The ground state (3sm) is indicated at the bottom

Ionisation limit (.5.12 eV)

. Av >5.12eV
eo

4dv

Ground state

Figure 14.26. Energy levels of the sodium atom. Starting from this diagram, the absorp
tion or emission lines of sodium in a condensed state can be calculated. It should be
noted that the return to equilibrium of the valence electron of gaseous sodium by a cas
cade mechanism obeys selection rules, which imply the passage through one of the two
intermediate 3 energy levels

of this figure. The Na atom in its fundamental state is in this energy state. To make
the valence electron pass to a higher energy level, it has to be excited with a radiation
of energy hv determined by the difference in energy between the fundamental and an
368 Introduction to Materials Science

excited levels. For example, by using radiation of energy ' ' of 2,103 eV, the outer
electron of the fundamental level 3si/ can be made to pass to the excited energy level
2

immediately above (3pm). The electron does not remain indefinitely in this excited
state, it can return to the fundamental level with emission of a radiation having
exactly the same energy hv\.
The valence electron can pass to a level higher than 3pm by absorbing a
radiation of higher energy but to be absorbed, the radiation must have a value which
corresponds exactly to the difference between two energy levels. The difference
between the successive energy levels becomes smaller as the energy of the levels
increases. This tends towards a limit corresponding to the ionisation energy which
equals 5,12 eV for sodium. If the valence electron absorbs a radiation of energy
higher than 5,12 eV (which corresponds to an ultraviolet radiation), it is pulled off
the atom and a N a ion is formed.
+

Subsequently, the ion Na and the electron recombine. The electron falls in cas
+

cade to the various energy levels of the sodium atom and returns finally to its initial
energy state. There are selection rules, which require that the sodium atom should
pass through one of the two intermediate energy levels 3pm or 3/? /2. From one of
3

these two energy levels, the Na atom returns to its fundamental level by emitting one
of the two characteristic radiations situated in the yellow zone (sodium doublet).
These are the intense yellow radiations observed if a sodium salt is heated in the
flame of a torch. As an example, figure 14.26. represents some possible energy tran
sitions with, of course, passage through one of the two intermediate energy levels
3pm or 3pm.

14.5.4. Light absorption and colouring of materials

Absorption mechanisms similar to those explained for gases operate also in the
condensed state. When the atoms combine to form molecules or condense to form a
liquid or a solid, the energy levels multiply. These are induced in particular by the
vibrations and rotations of atoms bound to each other. In this case, the fine and in
tense absorption lines found with gases are absent; instead absorption bands are ob
served.
In the case of the condensed phases, the return to equilibrium of the excited
electrons occurs by a cascade mechanism similar to that found for the isolated atoms.
In a certain number of cases, the emission of a luminous radiation is observed which
has, in general, a longer wavelength than that of the absorbed radiation, i.e. with
displacement of the absorption bands towards red (bathochrome effect). Such a phe
nomenon is called photoluminescence. If the light output occurs immediately after
absorption (within a delay of the order 10~ s), the effect is referred to as fluores
8

cence. If it is not instantaneous, then the phenomenon is known as phosphorescence,

which is prolonged in certain cases over several hours. As the energy levels are very
close in the condensed phase, the return to equilibrium also occurs by emission of
low energy radiations of about the same magnitude as that of phonons. A significant
fraction of dissipated energy is thus found in the form of heat.
When a transparent body selectively absorbs visible light, it appears coloured.
The colour of the object is characteristic of the part of the spectrum transmitted.
Thus, the pale blue colour of deep water or ice is the result of a weak absorption in
 Physical properties 369

the red by the water molecule, induced by a deformation vibration of the angle be
tween the oxygen atom and the hydrogen atoms. The pale blue colouring observed
corresponds to the radiations not absorbed by the water. Generally, if a series of
radiations is absorbed, it is always the complementary colour, corresponding to
transmitted radiations, that is observed.
If the absorption of visible light occurs uniformly for all the wavelengths, the
material appears to be not coloured. It should be noted that a non-coloured trans
parent material could selectively absorb a radiation situated outside the spectrum of
the visible light, in the ultraviolet for example. In this case, this material could be
used as filters in photography or for spectacle lenses for example.
The colouring of a substance practically never results from the promotion of a
valence electron because these are strongly bound and their excitation is not, in
general, possible except by using ultraviolet radiations. It is, in fact, almost exclu
sively electrons placed in an exceptional situation (strongly delocalised electrons,
unpaired electrons) that are weakly bound and are more easily excitable, that are
responsible for the colouring of a substance.

14.5.5. Effects of transition metal ions on ceramic colouring

Transition metal ions contribute to a great extent in the colouring of natural or

synthetic ceramics. By way of example, the influence of the Cr ion on the colour of
3 +

the ruby and emerald will be considered.

Mono-crystalline aluminium oxide ( A 1 0 ) (sapphire) used in particular for
2 3

scratch-resistant watch glasses is colourless. Ruby, a red coloured ceramic, is an

aluminum oxide A 1 0 in which a certain number of A l sites have been substituted
2 3
3+

by the C r ion
3+

A transition metal such as chromium has an incomplete 3d layer (chapter 2.).

Each C r ion has three unpaired electrons in this layer which has three excited
3+

energy levels. Their fundamental state is usually designated as 4A2 (figure 14.27.)
and the three excited states known as 2 , 4 , 4 in the classical terminology of the
2

spectroscopists. In the crystallographic lattice of ruby, the chromium ion is sur

rounded by six atoms of oxygen O (ligands) which form a deformed octahedron.
2 -

The energy level of the excited states is strongly influenced by the electric field in
duced by the O " ions adjacent to the C r site. In this complex, the C r - O bonds are
2 3+

63 % ionic, which leads to a high value for the electric field (ligand field) in the vi
cinity of the Cr atom. This influences the energy levels of the 2 , 4 , 47Vexcited
2

states associated with the unpaired electrons and which involves a very strong ab
sorption of the higher energy components (violet and yellow-green) of light. As
shown in figure 14.27., only the red components with a little blue are transmitted,
which gives the ruby its dark red, slightly violet colour.
The allowed transitions between the various excited states and the fundamental
level are shown in figure 14.27. These are controlled by complex selection rules
which prohibit the transition from the fundamental state 4 A to the excited state 2E
2

but which allows the passage from the fundamental state 4 A to the excited states 4Ti
2
370 Introduction to Materials Science

Absorption

Figure 14.27. Energy level of the Cr ion in ruby (A1 0 + Cr ) with indications of the
+
2 3
+

principal absorptions and transmissions (after Nassau, 1980).

a n d 4 T . It should be noted that there are also three transitions from higher
2

energy levels 4Ti and 4 T towards lower energy level 2E and which correspond to an
2

emission of infrared radiations.

Crystal field colour centres arise also in emerald, which is a beryllium and alu
minium silicate. As in ruby, a certain number of A l sites are replaced by C r ions.
3+ 3+

In this ceramic, the chromium ion has a different environment from that encountered
in ruby. The Cr bond is less ionic (a few percent) than in ruby, and the induced
electric field is weaker. This effect lowers energy levels 4T\ and 4 , leading to a 2

displacement of the absorption band towards the yellow-red areas of the spectrum
and giving emerald its blue-green colour. As the 2E level is very little modified, a red
fluorescence, almost identical to that of ruby, is also observed in the case of emerald.
Transition metals are used to colour a large number of pigments and mineral
glasses. Thus, photochromic glasses, used for sunglasses with tinting modulated
according to illumination, can be obtained by doping mineral glasses with E u and 2+

T i ions. These doped glasses develop, under the action of the ultraviolet component
4+

of sunlight, coloured centres which disappear gradually when the light intensity de
creases. The reaction occurring in these photochromic glasses is as follows:

Eu 2+
+ Ti 4+
% T i + Eu
3+ 3+
(14.31.)

The brown and violet colouring is given by the Ti ion, which has a unpaired 3+

electron, while the Eu , E u and T i ions are colourless.

2+ 3+ 4+
 Physical properties 371

14.5.6. Colouring of organic substances

In the organic substances, the electrons are paired to form very strong covalent
bonds. As long as the paired electrons remain localised between two atoms, the ex
cited states have a high energy and the resulting absorption is in the ultraviolet re
gion. This situation is found with the majority of amorphous thermoplastics, which
are transparent and form organic glasses
In certain cases, conjugated double (or triple) bonds exist in the molecule
which allow the electrons (chapter 2.) to be displaced over the whole of the mole
cule. These electrons then occupy the molecular orbitals. They are more weakly
bound to the system and the energy of their excited states is considerably lowered.
Such molecules absorb in the visible region and consequently appear coloured.
Crystal violet provides an example of a coloured substance the formula of which is
given in figure 14.28. This molecule contains several aromatic nuclei. In the mole
cule shown, the colour is reinforced by the presence of three amplifying groups
N(CH3) (auxochromes) which are electron donors and cause the displacement of
2

absorption towards longer wavelengths. The presence of these groups gives the violet
colour of this dye molecule. Organic dyes as well as mineral pigments are much used
to colour organic materials.

14.5.7. Optical properties of metals and semiconductors

In metals and semiconductors, the electrons are no longer bound to only one
atom or ion, and the polynuclear orbitals extend throughout the whole volume of the
material. Metals and semiconductors contain a large number of delocalised electrons
and they have optical properties very different from those of other materials.

Figure 14.28. Molecular structure of crystal violet. This dye molecule contains three
auxochrome groups N(CH3) . The arrows indicate the displacement of the electrons in
2

the molecular orbitals.

372 Introduction to Materials Science

Number of electrons

Figure 14.29. Energy Levels of electrons in a metal at zero degree Kelvin; all the levels
are occupied up to the Fermi level (after Nassau, 1980).

It was seen in chapter 2, that an incompletely filled valence band characterized

metals. At zero degree Kelvin (figure 14.29.), all the energy levels are occupied up to
a certain level called the Fermi level. The other energy levels are empty. Any energy
contribution, however small, takes the electrons to a level higher than the Fermi
level. Metals contain a continuous succession of energy states and can in practice
absorb all radiations and re-emit them instantaneously. For this reason a non-
oxidized metal surface is reflective and not absorbent, which would have given a
black colour like in graphite. The nuances, which are found in the colour of metals
(copper - gold - silver), occur because certain wavelengths are absorbed and re-
emitted more easily than others. Certain very thin metal films (of the order of several
nanometers) become transparent. This is the case for gold, which acquires a green
colour when transparent.
Semiconductors have a completely filled valence band, separated from the
conduction band by a forbidden band (figure 14.30.). The minimum wavelength
that semiconductors can absorb depends on the width of the forbidden band. If this is
lower than the energy of the least energetic component of visible light (red), all the
visible light is absorbed and the semiconductor has a black (CdSe) or grey metallic
(silicon) appearance, depending on whether the re-emission of the photons is slow or
fast.
If the width of the forbidden band corresponds to a wavelength falling into the
range of visible light energy, the material then has a well-defined colour. This is the
case for mercury sulphide HgS which has a forbidden band of 2.1 eV and which is a
bright red pigment, tending to orange. Cadmium sulphide, with a forbidden band of
2.6 eV, is a yellow pigment.
In the case of certain electrical insulators, the width of the forbidden band is
greater than the highest energy of visible light and the material becomes transparent.
This is the case for diamond with a forbidden band of 5.4 eV.
 Physical properties 373

/ " 5
Colouring Colouring of
ofthe J the substance
/ -a . absorbed
radiation > Colourless

ba
J

Energie du photor
ith of the forbidden
- Violet
- Blue Yellow

1
- Green Orange

:tron energy
- Yellow Red
- Red

1
Forbidden band VBlack

1
<-
<
J
0 Fermi level

Filled band /

Number of electrons

Figure 14.30. Energy level of electrons in a semiconductor. The valence band is com
pletely occupied and separated from the conduction band by the forbidden band (after
Nassau, 1980).

Solids with a wide forbidden band are coloured by doping them with electron
donors or acceptors which create intermediate energy levels in the forbidden band.
Thus, by introducing one atom of nitrogen per 100000 carbon atoms into the lattice
of diamond, a band of electron donors situated 1.4 eV above the valence band is
created, which leads to the yellow colouring of diamond.

14.6. Summary and Conclusions

In this chapter, the principal physical properties of materials have been exam
ined. The electric, thermal, magnetic and optical properties depend initially on the
electronic configuration of the atoms and the nature of the associated chemical
bonds. The organization (crystal or amorphous) of the atoms also plays a significant
role.
The electrical conductivity of materials varies by approximately 25 orders of
magnitude from conducting materials such as copper or aluminium to insulators like
PTFE. Generally, the purest metals are the best conductors. Great progress has been
made recently in the field of superconductors by the use of complex ceramics. Mate
rials with intermediate electrical conductivity (semiconductors) play an important
role in microelectronics.
For metals, thermal conductivity is very high and depends, like electrical con
ductivity, on the density of mobile electrons. Other materials like diamond, at
ambient temperature, conduct heat better than the best metal conductors (Cu, Al),
while the density of delocalised electrons is practically zero. In this case, the heat is
transferred in the form of phonons by the vibrations of the crystal lattice. In this case
also, the purest materials are the best thermal conductors.
374 Introduction to Materials Science

The insulation capacity of materials can be increased by dividing them (porous

materials, foams, etc.).
Magnetic properties play a very significant role in electrical and electrome
chanical equipment. Ferromagnetism is at the origin of the performance of hard
magnetic materials (permanent magnets) and soft magnetic materials (non-permanent
magnets). The intrinsic properties of magnetic materials are strongly influenced by
the control of their microstructure. Ferrimagnetism is also very important, in par
ticular for materials used for computer memory.
The optical properties of the materials (colour, reflectivity and transparency)
largely condition our perception of objects. The colour of a substance is directly
related to the presence of more weakly bound electrons producing excited energy
levels close to the fundamental state and allow the absorption of visible light. These
phenomena give rise to colour centres, which are, for example, transition metal ions
often present in ceramics at relatively low concentrations.
Transparency very much depends on the microstructure: in practice, only the
homogeneous materials on a scale down to 0.1 pm are transparent. The transparency
of glasses is strongly influenced by the presence of colour centres and local varia
tions of the refractive index. The impurities diffuse light reducing the transmitted
intensity. New varieties of very transparent mineral glasses are currently being de
veloped for the manufacture of optical fibres.
In conclusion, as highlighted in the study of mechanical properties, the physical
properties are determined to differing extents by the intrinsic properties of the mate
rials (chemical bonds, electronic structure, etc.) and by the phases and the micro-
structures present (grains, impurities, precipitates, etc.).

14.7. Illustrative example: Countless applications for perma

nent magnets
The range of applications for the permanent magnets is innumerable. Thus,
current cars are filled with electric motors, drive systems, measuring devices etc,
using permanent magnets. In figure 14.31, a series of vital accessories for our cars
using this type of materials is shown: starter, fan motor, fuel pump, windscreen
wipers, tachometer, etc. Much apparatus and many machines function thanks to
permanent magnets, such as computers and their peripherals, magnetic lenses, tape
recorders, generators, lifting devices, magnetic filters and separators, etc.
An example drawn from our everyday life: a quartz watch with analogue display
could not function during its years of life if its drive system was not extremely
powerful and required only a minute amount of electric energy due to high tech per
manent magnets.
The saving of electrical energy together with a minimal weight are the economic
drivers for the development of permanent magnets of high magnetic hardness. As
indicated in section 14.4, the hardness of a permanent magnet is measured by the
product (BxH) max, which is obtained starting from the demagnetisation curve
(figure 14.21). As can be seen in figure 14.32, the product (BxH) max of ferromag
netic materials increased by a factor of more than 50 between 1900 and 1980. For
example, a steel containing 3.5 % by weight of chromium is characterized by a value
 Physical properties 375

Sunshine roof

Crankshaft a n d
accelerator position
detector

Figure 14.31. Example of use of permanent magnets in the automobile (after Croat and
Herbst).

I l l I Year
1920 1940 1960 1980 2000

Figure 14.32. Evolution of the hardness of permanent magnets.

376 Introduction to Materials Science

of 2.3 kJ trf for the product (BxH) max whereas AINiCo of the 1950s reached 70
3

k J m " . At the beginning of the 1970s, a ferromagnetic intermetallic compound such

3

as S m C o reached 160 kJm~ . The current values for the latest generation of rare
5 3

earth magnets of the type F e i N d B are close to 300 kJ m~ .

4 2 3

This increase in magnetic hardness can be illustrated by comparing, in a very

simple experiment, the forces of repulsion which these permanent magnets exert on
themselves. In figure 14.33., pairs of magnets of the same mass having their identical
poles opposed have been placed in three separate transparent tubes. These pairs of
magnets are respectively made from ferrite, SmCo and F e i N d B . A cylinder of a
5 4 2

nonmagnetic material, the mass of which is adjusted to keep the distance separating
two opposite poles constant, surmounts each pair of magnets. It is easy to see the
extraordinary progress made in the manufacture of permanent magnets.
The effectiveness of the current permanent magnets can also be perceived if it is
considered that one cm of a SmCo requires a force higher than that of the weight of
3 5

a heavy man to pull this magnet off a steel part.

It seems that the upper limit for these hard magnetic materials is tending towards
values of 800 kJ m . There is still much to achieve for future researchers.
3

Nd^FeHB rod
270 KJrrf 1

SmCo5 rod
160 KJm" 1

Ferrite rod
30 KJm 1

Figure 14.33. Forces of repulsion of various permanent magnets.

14.8. Exercises
14.8.1. Calculate the thermal power loss in a copper wire in an electricity transmis
sion line of length L = 2000 m when the current intensity / is equal to 50 A. The wire
has a diameter d= 1 mm and its electrical resistivity = 1.7 10" Qm. 8

14.8.2. In service, a toaster equipped with a heater (nickel-chromium) consumes at

800 C a power = 350 W It operates under a potential difference V = 220 V. What
is the intensity / of the current circulating in the heating element?
 Physical properties 377

14.8.3. Explain why diamond, an electrical insulator, is one of the best thermal con
ductors at ambient temperature?
14.8.4. A rail of length L =1 km is heated to 40 C before being attached to its sup
port. Calculate the stress which acts on the support when the temperature reaches in
winter - 1 0 C.
s t e e , = 210 GPa, a stee , = 12 10"* C~ l

14.8.5. What is the relation between the electrical and the thermal conductivity of a
metal?
14.8.6. What is the heat loss per day through a brick wall of 4 m by 4 m and 0,25 m
thickness when the internal face of the wall is maintained at a temperature of 20 C
while the external face is at 0 C. The thermal conductivity of brick is 0.5 Wm^K" . 1

14.8.7. A rare earth permanent magnet has a demagnetisation curve of parabolic form
which can be expressed by the formula 5 = 1 . 1 5 - 4.5 10~ H . What is the value of 6 2

the coercive field HQI

14.8.8. What does the product (2?x//) m a x of a permanent magnet represent? Make a
drawing.
14.8.9. A magnetic field = 2500 Am" is applied to a material of relative perme
1

ability = 6000. Calculate the magnetization and the magnetic induction field.
r

14.8.10. Explain why certain mineral glasses lose their transparency if they are
heated between 200 and 400 C.

14.9. References and complementary reading

J.C. ANDERSON, K.D. LEAVER, J.M. ALEXANDER, R.D. RAWLINGS, Materials Science, 2nd ed.,
Walton-on-Thames, G.B., 1981.
J. BARDEEN, L.N. COOPER, J.R. SCHRIEFFER, Phys. Rev. 108, 1175, 1957.
J.G. BEDNORZ, K.A. MLLER, . Phys., 64, 189, 1986.
J.J. CROAT and J.F. HERBST, MRS BULLETIN 13, 37, 1988.
R. DAGANI, Supraconductivity: A revolution in electricity is taking shape, C & En., May 1987, 7.
M. GERL et J.P. IS SI, Physique des materiaux, Presses polytechniques et universitaires romandes,
Lausanne, 1997.
R.M. GERMAN, M.M. GUZOWSKI, D.C. WRIGHT, Color and color stability as Alloy Design Criteria,
J. Metals, 20/21, 1980.
J.P. ISSI, La conduction de la chaleur, La Recherche, 12, 442, 1981.
J.P. ISSI, J. LEBEGGE, Les supraconducteurs, Revue E,9,l. 1978.
C. KITTEL, Introduction to Solid State Physics, 7th ed., John Wiley, New York, 1996.
M.C. LOVELL, A.J. AVERY, M.W. VERNON, Physical Properties of Materials, Van Nostrand
Reinhold, New York, 1981.
Materials Science and Engineeringfor The 1990s, National Academy Press, Washington D.C, 1989.
K.A. MLLER, J.G. BEDNORZ, La docouverte de la supraconductivite haute temperature, La
Recherche, 19, 52 1988.
K. NASSAU, The Causes of Color, Sei. American 243,October 1980, p. 106.
K.M. RALLS, T.H. COURTNEY. J. WULFF, Introduction to Materials Science and Engineering, Wiley,
New York, 1976.
J. ZARSYCKI, Glasses and the Vitreous State, Cambridge University Press, Cambridge, 1991.
 Chapter 15.

Corrosion, degradation
and ageing

15.1. Objectives
To review the various forms of material degradation.
To introduce the electrochemical mechanism of the wet corrosion of
metals.
To understand the phenomenon of passivation.
To identify the oxidation mechanisms of metals at high temperature.
To highlight the decomposition mechanisms of polymers.
To give some examples of degradation of ceramics.

The corrosion and degradation of materials represent a very significant eco

nomic problem. Each year about a quarter of the steel production is destroyed by
corrosion. One of the objectives of Materials Science is to obtain materials more
chemically stable over time and, consequently, more reliable. The study of the
degradation and ageing phenomena is therefore of paramount importance.

15.2. Various forms of ageing

15.2.1. Chemical or physical ageing
The use of materials is very dependent on how their physical and mechanical
properties evolve over time and on the environment in which they are placed.
Materials are practically never in a state of thermodynamic equilibrium and their
properties are expected to change as a function of the constraints imposed by the
environment. This evolution of material properties has physical and chemical origins.
Thus, an increase in the working temperature of a metal alloy is likely to involve an
alteration of the microstructure and modification of the mechanical properties, e.g. by
formation of a precipitate which weakens the material. Equally, the physical ageing
of organic or mineral glasses has to be considered. When glassy materials, which are
always out of equilibrium, are held at temperatures close to their glass transition
temperature, their volume decreases and the tolerances for the manufacture of
precision parts must be modified.

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

380 Introduction to Materials Science

In many cases, the evolution of physical and mechanical properties over time is
accompanied by significant changes in the structure and chemical composition of
materials. Under the influence of temperature, atmospheric agents (e.g. 0 ) , chemical 2

reagents or applied stresses, the structure at the atomic scale and the internal or sur
face microstructure of materials can be modified.
The chemical stability of materials varies considerably according to their nature.
Thus, practically no metals occur in their natural state, except for noble metals like
gold. Metals show a general tendency to return to their original state as sulphide or
oxide. As will be seen later in this chapter, this sensitivity to oxidation is directly
related to the high electrical conductivity of these materials and to the mobility of
their valence electrons.
Organic polymers exhibit considerable thermal instability. A moderate tem
perature rise (250-400 C) causes the rupture of a significant number of bonds in the
material with, depending on the specific case, formation of volatile compounds
(monomers - molecular fragments - C 0 ) or graphitisation. An example of con
2

trolled graphitisation is that of the pyrolysis under stress of poly(acrylonitrile) fibre

leading to the formation of carbon fibre (chapter 2, illustrative example).
Ceramics are thermally the most stable of all materials because they are mainly
made up of oxides and consequently are not able to react with oxygen. The refractory
materials belong almost exclusively to the class of ceramics. Indeed, refractory
metals, W, Mo, etc., are not stable chemically in the presence of oxygen at high
temperature as they form volatile oxides.

15.2.2. Corrosion and degradation

In many cases, corrosion is limited to the surface. The oxidation occurring at the
metal-atmosphere interface can be completely eliminated by isolating the metal sur
face with an organic coating (painting) or with a layer of impervious oxides.
Surface corrosion often takes a much more localised form (figure 15.1.). Often,
it is restricted to well-defined sites on the surface, e.g. where there is a defect in the
protective coating or at a surface defect (grain boundaries for instance), or within a
surface crack. These localised attacks are most dangerous, because they often pro
gress at a high rate and end up affecting a significant area of the material.

(d)

Figure 15.1. Some examples of corrosion: (a) surface corrosion; (b) corrosion in a crack;
(c) corrosion at a grain boundary; (d) corrosion in a particle (precipitate).
 Corrosion, degradation and ageing 381

Depending on their nature, certain types of corrosion are considerably acceler

ated by the presence of static or cyclic stresses. Fatigue resistance (chapter 13) is
reduced by corrosion occurring in the neighbourhood of cracks smaller than the criti
cal size. The presence of certain organic liquids or water containing detergents, leads
to the stress cracking of a large number of thermoplastic polymers e.g. polyethylene.
Corrosion also occurs at the interface of two materials of different composition.
This type of corrosion, which occurs mainly in metals, takes almost always the form
of electrochemical corrosion as will be seen in paragraph 15.3.1.
The corrosion and degradation of materials can take many varied and insidious
forms. The remainder of this chapter will be limited to the study of some important
practical examples.

15.3. Corrosion of metals

15.3.1. Galvanic corrosion
A metal immersed in an electrolyte such as water at a pH < 7, releases metal
ions to the electrolyte. To leave the metal, the ions must overcome the attraction of
the metal and the repulsive effect of the ions already in solution. The passage of
metallic ions from the metal into the electrolyte is limited in time, because releasing
ions charges the metal negatively. The voltage between the metal and the solution
tends to a stationary value called the electrode potential U.
Two pieces of different metals (e.g. iron and zinc (figure 15.2.) connected with
one another and immersed in an electrolyte form an electrochemical cell. The metal
with the most negative potential (Zn) constitutes the anode. The zinc atoms lose two
electrons and pass into solution in the form of Z n ions. The zinc sheet becomes
++

negatively charged. A zinc oxidation reaction occurs at the anode:

Zn -> Zn + 2e~ (15.1.)
The iron piece constitutes the cathode. On the surface of the cathode, a reduction
reaction of the electrolyte ions takes place:
2H + 2e~->H
+
2 (15.2.)
The iron constituting the cathode becomes positively charged. The potentials of the
two electrodes are opposing. To compensate, the cathode receives electrons from the
anode through the external circuit.
Each time two metals of different chemical nature are in contact with an electro
lyte, the phenomenon of galvanic corrosion, also called bimetallic corrosion, occurs.
The metal, which has the lowest negative electrode potential, tends to dissolve in the
electrolyte. Electrochemical corrosion often occurs at a microscopic scale. Hetero
geneities (grains, boundaries, precipitates, etc.) then play the role of electrodes.
Corrosion is considerably accelerated by the presence of ions able to transport
the electrical current in the electrolyte. Thus, pure water (pH = 7) containing only a
very small quantity of H and OH~ ions is much less corrosive than seawater con
+

taining approximately 0.6 m o l l of Cl~ and N a ions.

-1 +

The phenomenon of galvanic corrosion is also used in certain cases for the pro
tection of metals. When a scratch damages the zinc layer of galvanized steel and the
382 Introduction to Materials Science

= 0,3

r - ^0
-0,76V 0
0
-0,44V
0

> H 2

Zn <J
K

/t Fe
1 2H +

Electrolyte

Figure 15.2. Electrochemical cell consisting of an iron and a zinc sheet placed in an
electrolyte solution. The standard electrode potentials are indicated on each electrode.
When the electrolyte activity is unity, the potential difference U between the anode and
the cathode of the cell is initially equal to 0,32 V. Electrolyte composition: H , Zn^, CI".
+

steel so exposed comes into contact with an electrolyte, the steel - zinc combination
forms an electrochemical cell where the zinc layer is the anode. The scratched zinc
layer dissolves slowly and the steel cathode is not attacked.
The dissolution of a metal involves the transfer of electrical charges to the
metal-electrolyte interface and the formation of ions passing into solution. To clas
sify metals according to their aptitude to dissolve, it is necessary to know their elec
trode potential. Metals have been classified relative to a reference electrode (hydro
gen electrode) which has a potential arbitrarily taken as zero. The determination of
standard electrode potential is carried out in an electrolyte consisting of a solution of
a salt of the metal at an activity equal to unity. This classification is shown in fi
gure 15.3. The highest oxidation-resistant metals are characterized by a very high
electrode potential. Placed in an electrolyte, these metals generally have a cathodic
behaviour and constituted the class of the noble metals as silver, gold and platinum.
The most oxidizable metals e.g. Zn, Al have a very negative electrode potential and
generally show anodic behaviour.
In table 15.4., various metals and alloys have been classified in order of corro
sion resistance in sea water.

15.3.2. Mechanism of rust formation

The formation of rust on an iron surface in contact with water or wet air is
another classical aspect of wet corrosion. The principal constituent of ras/is hydra-
ted iron oxide, F e 0 . 3 H 0 . To understand the mechanism of rust formation, a very
2 3 2

simple test can be carried out. A plate of pure iron is immersed in a dilute, deoxy-
genated aqueous solution of NaCl and some drops of a Phenolphthalein solution and
a potassium ferricyanide solution acting as indicators are added. After some time
(figure 15.5. (a)), the appearance of a pink colouring towards the top of the specimen
is observed, indicating the formation of OH" ions while at the bottom of the speci
men, the blue colouring characteristic of the ferrous ions F e ^ is seen. In the interme
diate zone, the formation of a layer of rust is noted.
 Corrosion, degradation and ageing 383

(V)

- - Au Au* + e"

- Ag 5= A g * e~ +

- Cu ^ Cu 2 +
* 2e"

^Ni^Ni * 2
+ 2e"
- v C o = Co * 2
+ 2e~
- Fe^Fe 2 +
+ 2e-
_xCr = ^ C r * 3
* 3e-
Zn^Zn
v 2 +
. 2e-

- Al ^ A l 3
* + 3e _

Figure 15.3. Standard electrode potentials of some pure metals at 25 C.

Table 15.4. Classification of metals and alloys in decreasing order of resistance

to corrosion in circulating sea water.

Noble metals Composition

Cathodic behaviour

Platinum
Gold
Stainless steel (18%Cr,8%Ni)
Titanium
Silver
Bronze (Al) (80 % Cu, 5 % Ni, 9 % Al, 4,5 % Fe)
Inconel
Nickel
Cupronickel (70%Cu, 30%Ni)
Bronze(Sn) (86 % Cu, 5 % Sn, 5 % Zn, 4 % Pb)
Brass (85%Cu, 15%Zn)
Copper
Tin
Lead
Cast iron
Iron, mild steel
Aluminium
Aluminium alloy (96 % Al, 0,2 % Cu, 0,6 % Si, 1 % Mg)
Galvanized steel
Zinc
Magnesium and alloys
Non-noble metals
Anodic behaviour
384 Introduction to Materials Science

Fe

Pink colouration (OH~)- Anodic

corrosion pit
Rust
Cathodic region

Blue colouration ( F O -

(a) (b)

Figure 15.5. (a) Iron plate immersed vertically in a dilute, deoxygenated aqueous solu
tion of NaCl to which some drops of a Phenolphthalein solution and a potassium ferri-
cyanide solution have been added as indicators; (b) attack on the iron by a drop of dilute
aqueous NaCl solution.

How can these results be interpreted? The oxygen from the air dissolves gradu
ally in the aqueous NaCl solution where a significant oxygen concentration gradient
exists, which induces the formation of an electrochemical cell in the system. The
cathodic part of the electrochemical cell is located close to the surface where a re
duction of oxygen occurs:
0 + 2 H 0 + 4 e " ^ 4 (OH)"
2 2 (15.3.)
The lower part of the iron plate constitutes the anode where there is a dissolution
(oxidation) of iron in the form of ferrous ion:
Fe -> Fe +
+ 2 e" (15.4.)
In the intermediate zone, ferrous hydroxide is precipitated:
Fe ++
+ 2 OH" - Fe(OH) 2 (15.5.)
This can undergo a subsequent oxidation:
2 Fe(OH) + H 0 + 1/2 0 - 2 Fe(OH)
2 2 2 3 (15.6.)

Note that this last reaction is not electrochemical. On drying, the iron hydroxide
yields a porous hydrated iron oxide of complex chemical composition (rust).
The same phenomenon occurs when a drop of dilute NaCl solution is placed on
the surface of an iron specimen (figure 15.5. (b)). It is the formation of an electro
chemical cell, similar to that which has just been described, which explains the rapid
progression of corrosion in cracks.

15.3.3. Method for protection of metals by passivation

The formation of a stable and dense oxide layer on the surface of a metal con
siderably modifies its electrode potential and provides an effective protection against
corrosion. The rust layer formed on the surface of iron and carbon steels is porous
and does not protect from later attack. On the other hand, the very dense and imper
meable oxide layer ( C r 0 ) , which develops on the surface of stainless steels
2 3

(Fe, Ni, Cr), does provide such protection. In this case, the oxide coating is suffi-
 Corrosion, degradation and ageing 385

ciently thin and transparent and does not to affect the metal brightness. Moreover, it
is hard and quickly regenerated after mechanical damage.
Aluminium also forms a very dense oxide layer (A1 0 ) protecting this very
2 3

oxidizable metal against later corrosion. The dense oxide coating can also be formed
electrochemically by placing the part to be protected at the anode of an electrolysis
bath (aluminium anodization).
Various methods do exist for the protection of ferrous alloys against corrosion.
The use of organic coatings (painting) has already been mentioned, and also galvani
zation consisting in covering the metal with a thick layer of zinc, which is in turn
protected by a ZnO oxide coating. In fact, in this case, there are two levels of protec
tion. Indeed, when a scratch damages the zinc layer, this functions as the anode and
protection continues to be effective.
Ferrous alloys can also be protected by electrodeposition of metals such as cop
per, nickel and chromium. Note that in the event of damage to the protective
covering, this type of coating generally constitutes a cathode, and that the
electrochemical corrosion of the bare underlying metal is accelerated.
Large parts, such as pipes or boats, can be protected by the use of a sacrificial
anode (figure 15.6.) which is made of a less noble metal (zinc, magnesium) than the
part to be protected. This is connected to the sacrificial anode by a conducting wire
(Cu) and taken to a sufficiently negative potential to inhibit the phenomenon of ano
dic corrosion.

Figure 15.6. Cathodic protection of metals by a sacrificial anode.

15.3.4. Corrosion and oxidation at high temperature

Dry corrosion is produced by the attack of a gas on a metal at high temperature.
Oxidation is by far the most frequent, but other gases ( S 0 , C 0 , water vapour, etc),
2 2

are also able to react at high temperature.

This paragraph will be limited to the case of oxidation. The reaction between a
solid and a gas is very complex, and the thermodynamic and kinetic factors will be
considered respectively.
Table 15.7. illustrates the thermodynamic aspect. The noble metals, gold in par
ticular having the weakest affinity for oxygen, form the least stable oxides. However,
it is not necessarily metals forming the most stable oxides that are the most prone to
oxidation at high temperature. Indeed, the formation of a solid oxide film at high
temperature generally occurs in two stages: creation of a continuous oxide film on
the metal surface and the thickening of this film.
The oxidative behaviour of metals is illustrated in figure 15.8. Initially, chemical
adsorption of an oxygen film occurs. This involves dissociation of the gas, its ionisa-
386 Introduction to Materials Science

Table 15.7. Enthalpy of formation of some oxides

at ambient temperature.

Material Oxide Enthalpy of formation

(kJmol" 0 consumed)
1
2

Aluminium A1 0
2 3
-1045
Titanium TiO -848
Silicium Si0 2 -836
Tantalum Ta 0
2 5
-764
Chrome Cr 0
2 3 -701
Zinc ZnO -636
Silicium Carbide (SiC) Si0 + C 0
2 2 -580
Tungsten W0 3 -510
Iron Fe 0
3 4
-508
Nickel NiO -439
Cobalt CoO -422
Diamond, graphite C0 2 -389
Lead Pb 0
3 4 -309
Copper CuO -254
Silver Ag 0
2 -5
Gold Au 0 2 3 + 80

(after Ashby and Jones, 1986, 1998)

Figure 15.8. Mechanism of growth of an oxide film on the surface of a metal: (a) forma
tion of a film of chemically absorbed oxygen; (b) appearance of seeds; (c) lateral growth
of the seeds; (d) growth of the compact oxide (after Benard, Michel, Philibert and Tal
bot, 1984).

tion and saturation of the free valences of the metal atoms on the surface. In a second
stage, there is an epitaxial seed formation at predetermined points on the surface
(emergence of dislocations, precipitates, etc). The third phase is the lateral growth of
the seeds to form a continuous oxide coating. Afterwards the oxide layer thickens at
the expense of the metal and oxygen in air.
In certain cases (Fe), the oxide surface layer is porous and molecular oxygen
(O2) has access to the metal surface through the pores, and the oxide layer is not
protective, even though with time the oxidation reaction rate decreases (figure 15.9.
(a)). In other cases (AI, Cr, Si), the oxide coating is dense and compact. Oxidation
then occurs, either by chemical adsorption of oxygen at the oxide layer surface and
migration of O " ions to the metal surface, or by circulation of metal ions and elec
2

trons through the oxide coating towards the oxide-gas interface. These two mecha
nisms can occur simultaneously. When the oxide coating is compact, the rate of
oxidation is much slower, the oxide coating is protective and the metal seems non
oxidizable (figure 15.9. (b)). Certain refractory metals (tungsten, niobium, tantalum)
form oxides which are volatile at high temperature. These oxides disappear as soon as
 Corrosion, degradation and ageing 387

(a) (b)

Figure 15.9. Various oxidative behaviours of metals.

they are formed. As already mentioned in chapter 12, these high melting point metals
have very poor oxidation resistance at high temperature (figure 15.9. (c)).
Three elements (chromium, aluminium and silicon) are known to increase the
corrosion resistance of metal alloys at high temperature. In that context, the stainless
steels (Fe/Cr 18 %/Ni 8 % or Fe/Cr 13 % by w e i g h t ) , the superalloys
(Ni/Cr 15 %/Al 5 % by weight, plus six to eight other elements) and the alloys used
for heating elements (Fe/Cr 20 % /Al 5 %) are well known.
Silicon is seldom introduced into metal alloys, because it tends to make them
brittle. At high temperature, silicon carbide tends to be used. Heating elements made
with this ceramic material are able to withstand several thousand of hours at high
temperature (T < 1500 C). This is not possible with metal alloys. It is the presence
of a self-healing, protective coating of S i 0 on the surface of the SiC, that is
2

responsible for the exceptional stability of this ceramic at high temperature in an

oxidizing atmosphere. Some examples of corrosion resistant materials are given in
table 15.10.

15.4. Polymer degradation

15.4.1. General
Polymers are organic compounds and their resistance to degradation is very low
compared to other materials. In general, polymers are not very stable under process
ing conditions, typically at relatively high temperatures ( 1 8 0 - 3 5 0 C).
Various types of degradations take place. The most basic is pyrolysis occurring
in the molten state during processing. Depending on the case, the mechanism pro
ceeds at random or by selective rupture of less stable chemical bonds. Polymers are
also very sensitive to oxidation induced by traces of oxygen present in the com
pounding machines. In practice, at high temperatures, pyrolysis and oxidation
mechanisms are superimposed and interfere with each other.
388 Introduction to Materials Science

Table 15.10. Some corrosion resistant materials.

Materials Systemsfa] Applications

1. Materials resistant to corrosion at low temperature

Glass Si0 , CaO, Na 0

2 2 Containers
Polymers Ex.: polyethylene Containers
Stainless steel Fe (Cr 0 ) 2 3 Chemical and hydraulic
installations
Al alloys AI (AI2O3) Metal constructions, aircraft
Ta Plating on steel Chemical industry
Ti Plating on steel Chemical industry, implants
Co + Cr alloy Co (Cr 0 ) 2 3 Implants (surgery)
Noble metals Pt, Au, Ag Electrical contacts

2. Materials with high temperature resistance in an oxidizing atmosphere

Oxides A1 0 , MgO2 3 Furnaces, crucibles, cutting tools

SiC SiC (Si0 ) 2 Heating elements
Si N
3 4 Si N (Si0 )
3 4 2 Engines components
Superalloys Ni, Co, Cr, AI, etc. Gas turbine blades
Noble metals Pt Heating elements, crucibles for
mineral glasses
[a] The nature of the oxidation resistant layer is specified in brackets when it differs
from the bulk material.

Polymer degradation also occurs during its use at ambient or moderate tem
peratures. These degradations are frequently of photochemical origin. In this case,
photodegradation is usually coupled with oxidation. All these processes are
obviously a function of the chemical structure polymer. For example, the presence
of substituents absorbing specifically in the ultraviolet range induces a particular
sensitivity to photodegradation.
It is generally possible to change the course of degradation reactions by the use
of various stabilizers modifying the degradation mechanism. Indeed, the reactive
intermediates are often free radicals present in very low concentrations which can be
neutralized by various additives. A number of substances present in very small
quantities, transition metals traces for example, promote the formation of free radi
cals and increase degradation, sometimes quite spectacularly and are therefore
degradation promoters that should be neutralised.
The tangible signs of the degradation of a polymeric material are embrittlement
and the weakening of mechanical properties resulting from a molecular weight
reduction. Discolouration is also a concrete indication of polymer degradation.

15.4.2. Pyrolysis reactions

Two limiting cases are considered in polymers pyrolysis. Statistical degradation
proceeds by a rupture of bonds belonging to the polymer chain. In this type of rup
ture, the valence electrons are shared between the new-formed chemical species
(homolytic cleavage), resulting in the formation of free radicals:
 Corrosion, degradation and ageing 389

^CC-
" * C C C ~C '^ r' ^rcC*H + HC 2 Coxr (15.7.)
H 2 I H 2 J H J 2 I
R R R R
These free radicals react via various mechanisms:

v^cC'H + v^ccc/\
H 2 I H 2 J H 2 I
R R R


v/v*C
H 2 I
CH + 2 v ^ C C
H 2 I C O/vn
J V
(15.8) J

R R R

v/V^C C=CH 9 + HCW

H2 I 2
I

Statistical degradation leads to a progressive molecular weight reduction.

Depolymerisation in PMMA is the reverse of the polymerisation reaction:

CH 3 CH 3 CH 3 CH 3

v/v^CCCC* sWCC + H C=C

2 ( . q\
H 2 I H 2 I H 2 I J t 1 5
-^
c=o c==o c=o c=o
0-CH 3 0-CH 3 0-CH 3 0-CH 3

This reaction is in general initiated by a small number of free radicals formed by the
homolytic rupture of less stable bonds, often situated at a chain end. During the de-
polymerisation, the overall molecular weight remains practically unchanged because
only a small number of macromolecules, carrying a radical, depolymerise
simultaneously and form monomers.
The pyrolysis of polymers at high temperature is often a combination of these
two mechanisms, and the proportion of monomer formed in the volatile products
varies from zero to 100 %, depending on the polymer chemical structure. In some
polymers like polyethylene, little formation of volatile compounds is observed and
d e p o l y m e r i s a t i o n does not occur. For other cases, in p o l y m e r s like
poly(methylmethacrylate) described in (15.9.), depolymerisation is complete.
Following heating to a relatively moderate temperature ( 180 C), practically 100 %
of the initial monomer is recovered. In intermediate cases e.g. polystyrene,
approximately 40 % of the monomer is recovered and oligomers (dimers, trimers,
etc) are also formed. Frequently, crosslinking reactions are observed and lead to a
polymer becoming progressively insoluble.

15.4.3. Auto-oxidation. Limiting oxygen index (LOT)

Because of its particular electronic structure (Qz is regarded as a diradical), the
oxygen molecule is particularly reactive. Thermal oxidation concerns many
polymeric materials including polyamides, polyurethanes and polyesters. The case of
polyolefins such as polypropylene is particularly important. The oxidation of
390 Introduction to Materials Science

propylene by oxygen is a chain radical mechanism initiated by the following

reaction:

CH H
2 CH CH3H2 CH3H2 CH
3
2
H H 3

0
3 3
2

I
I ^ 2
I I

(R-)
(15.10.)
CH "2 CH H2
3 3 CH 3
CHfi2 CHf?2 CH
3 3 3
v
c c c + 0 2
c c c
I I I I I

'er

(R-) (ROO ) -

The tertiary hydrogen carried by the CH-CH3 group is labile and easily removed
by molecular oxygen. The macroradical (R') formed is particularly reactive and adds
to an oxygen molecule 0 to give a more stable peroxy radical ( R O O ' ) . This reaction
2

proceeds preferably in an intramolecular mechanism and leads to the progressive

peroxidation of the polymer chain:

CH3JJ2 CH "2
2

CH2 CH"2 CH
3 CH 3
3 3
I 3

I I I I
c

H
(15.11.)

CHJj2 CH^2 ch3 3 3

I I
The peroxidized units are unstable and their decomposition
OOH produces
the pro
gressive fragmentation of the polymer chain. A hydroxyl (O-H) group is formed at
one end of the fragmented chain and a ( C = 0 ) ketone at the other end:

H CH H
CH 2 2 CH

CHJi2 ICH
3
3

(15.12.)
"C OH 3 3
I I I I

HO^ Polymer chain Polymer terminated by
terminated by a a ketone group chain
hydroxyl group
 Corrosion, degradation and ageing 391

Chain fragmentation, which reduces the degree of crystallinity of the polyolefin,

produces a progressive loss of the mechanical properties.
In general, the decomposition of polymers at high temperature is considerably
accelerated by the presence of oxygen. Often, combustion of the polymer is
observed. A significant characteristic of a polymer is its limiting oxygen index (LOI)
that gives the minimum proportion of oxygen necessary to maintain combustion in an
atmosphere made up of N and 0 (table 15.11.). A high LOI indicates that the
2 2

polymer has a great resistance to combustion.

Table 15.11. Limiting oxygen index LOI ( %).

Polymer LOI (%)

Poly(ethylene) (PE) 17
Poly(methylmethacrylate) (PMMA) 17
Poly(styrene) (PS) 18
Poly(bisphenol-A carbonate) (PC) 27
Polyvinyl chloride) (PVC) 49
Poly(tetrafluoroethylene) (PTFE) 95

15.4.4. Photodegradation and photo-oxidation

Photodegradation is the most significant aspect of the climatic ageing (weath
ering) of polymers. They generally contain groups able to absorb the ultraviolet
radiation in sunlight. So, eis 1-4 poly(isoprene), i.e. natural rubber reacts in the fol
lowing way:

CH 3 CH 3

I hv I
v/v-CC=C Cv/v> C=CCv/v/^ + * (15.13.)
H2 H2 H2
The macroradical formed causes cross-linking which produces hardening, em-
brittlement and loss of elasticity of the rubber:

2^v>CC=C-C^ ^C C=C-C-C C=C-G/^ (15.14.)

H 2 H 2 HH I H
t 2

Macroradicals CH 3

Crosslinks

The macroradical reacts with oxygen to form a peroxy radical:

CH 3 <F3 H
O-O-
I I
^cc=ccv^ + o 2
c =
(15-15.)
H
2 H 2
This reaction leads to the formation of peroxide groups which react according to a
mechanism similar to the one explained in (15.12).
392 Introduction to Materials Science

In general, photodegradation, which is almost always coupled with oxidation,

produces a molecular weight reduction or/and crosslinking with loss of mechanical
properties. As the polymers are in general not very permeable to UV, photochemical
degradation is often limited to the surface.

15.4.5. Stabilization of polymers

The reactive species responsible for the thermal and photochemical degradation
of polymers are, in most cases, free radicals. Various stabilizers (antioxidants, UV
stabilizers) are able, either to inhibit the formation of free radicals or to destroy them
and so significantly increase the lifetime of polymers. As an illustrative example, the
action of antioxidants will be described.
Various substances are able to react with free radicals (chain breaking antioxi
dants), which are the drivers for the oxidation reaction, to give inert products. The
most commonly used antioxidants are substituted phenols of which the most well
known is Ionol (di(i-butyl-p-cresol) which has the following formula:

(15.16.)

Ionol

The mechanism of action of Ionol is summarised in figure 15.12. Each Ionol

molecule is therefore able to consume two peroxy radicals (ROO'). However, there is
formation of unstable peroxide groups in the final reaction product, which are still
potentially able to induce a further degradation. Therefore, the role of a second group
of antioxidants called preventive antioxidants is to destroy the remaining peroxide
groups (15.17.).
The preventive antioxidants eliminate the peroxides by non-radical mechanisms.
Triphosphites are a classical example:

4 C9H19

/?-tri-nonyl-phenylphosphite
+ R 0 H - - ROH + 0 = P - O - ^ ^ - C g H !
2

peroxide alcohol p-tri-nonyl-phenylphosphate

(15.17.)

The peroxides are transformed into inert products (alcohol and phosphate). The pre
ventive antioxidants are never used alone because they are unable to react with radi
cals. They are always employed in combination with chain breaking antioxidants.
A new category of highly effective antioxidants, recently introduced, are the aromatic
lactones, which possess very reactive hydrogen. They are able to destroy the very
reactive macroradicals R\ Their mechanism of action, summarised in figure 15.13., is
very simple. The hydrogen atom in a position of the lactone carbonyl group
(Z> C ) reacts with a free radical.
 Corrosion, degradation and ageing 393

Peroxide group

Figure 15.12. Mechanism of action of Ionol

Figure 15.13. Lactones as oxidation inhibitors.

394 Introduction to Materials Science

The formed radical, which is very stable, disappears by addition of a R" or RO'
radical. There are various techniques to inhibit photodegradation, which is extremely
aggressive in certain environments. These techniques include protective coatings, UV
filters and radical de-activating agents (quenchers). It should be noted that the use of
certain pigments (carbon black), which make polymers opaque, also considerably
increase the lifetime of organic materials, and in particular car tyres.

15.4.6 Degradation and stabilization of P V C

The degradation of PVC poly(vinyl chloride) is a special case because it occurs
without chain rupture. Its proceeds via a spontaneous dehydrochloration occurring
during processing at relatively low temperatures (from 160 C) which results in the
formation of short poly(acetylene) segments:

CCH- C- :C- + nHCl (15.18.)

H2 I H H
CI

These short poly(acetylene)segments (electrical semiconductors) strongly colour

PVC even at very low concentrations. PVC is stabilised using salts of organic acids
and bivalent cations (Ca , B a ^ ' Cd , P b , etc.) which neutralize hydrogen chloride
++ ++ ++

as it is formed:

Ba(0-CO-R) + 2HC1 - BaCl + 2 R - C O - O H

2 2 (15.19.)

The hydrogen chloride, which exerts a catalytic effect on the dehydrochloration, is

also very corrosive for the processing machines.
The dehydrochloration is initiated by a certain number of irregular structural
units introduced during the synthesis of the PVC chains. These units, which contain a
double carbon-carbon bond adjacent to a carbon atom carrying a chlorine atom, can
be destroyed by specific reagents such as certain organotin compounds:

CI
v/vnC=C C CM + [H c4CH^-Sn[sCH -OCOR]
3 2 2

H
H? 2

Anomalous unit Organotin derivative

(15.20.)

o^rc=C C CM + [H3C4CH2|4-Sn-Cl
I H 2
J 3 J 2
I
SCH2-OCO-R SCH^OCO-R
15.4.7. Resistance of polymers to corrosion and chemical agents
The resistance of a polymer to corrosion and chemical agents depend on its
molecular structure. A certain number of polymers (polyfluorinated polyolefins) have
excellent resistance to chemical agents and solvents, and are used to store very corro
sive liquids (acids - bases - organic solvents).
 Corrosion, degradation and ageing 395

It should be stressed that the corrosion mechanisms of polymers are very com
plex and, by contrast with metals, are never electrochemical in nature because, in
general, polymers are not electrical conductors.

15.5. Resistance of ceramics to chemical corrosion

While ceramics are normally highly corrosion resistant, there are nevertheless
exceptions. Thus, the corrosion of mineral glasses can be considerable in the pres
ence of sodium hydroxide (alkali), which is able to hydrolyse the silica surface
network:

<j) p
SiSi + OH" * Si OH + Si O" (15.21.)
I I I I
0 0

This hydrolysis reaction results in a tarnishing of the surface. Even in the pres
ence of pure water, a certain degree of ionic exchange can occur with the alkaline or
alkaline-earth cations present in mineral glasses:

Si 0~Na +
+ H 0
2 S i O H + OH" (15.22.)
I I

Special glasses are manufactured for the storage of certain liquids sensitive to
alkalis (e.g. drugs). These glasses are subjected to chemical treatment eliminating the
soluble ions at the surface.
Another category of ceramics, very sensitive to degradation, are building materi
als (concrete, cement, limestone, brick). The calcium carbonate limestone and certain
calcium salts in concrete and cement dissolve very slowly under the action of H C O , 3-

S 0 ~ and S 0 ~ ions contained in rainwaters. This corrosion reaches mainly the sur
3
2
4
2

face capillary network of these materials and under the action of the wetting -drying
cycles, a degraded crust is formed at the material surface (stone cancer).
Another type of degradation of building materials is the destruction, both at the
surface and in the bulk of the material, by the action of freezing-thawing cycles.
Water, which penetrates into the capillary networks, freezes, at temperatures
depending on the pore diameter, and causes material cracking. In temperate climates,
the materials containing capillaries of intermediate diameters (0.2 to 0.5 m) are the
most sensitive to degradation. Water does not freeze in temperate climates in smaller
capillaries.

15.6. Summary and conclusions

In the use of materials, great importance has to be attached to the effects of the
environment, which can cause a considerable change in the physical and mechanical
properties.
396 Introduction to Materials Science

Ageing may have a physical or chemical origin. The ageing mechanisms vary
significantly according to the material type.
In metals, the corrosion, which occurs at ambient temperature, is primarily an
electrochemical reaction between the constituents of the material and oxygen. This
oxidation reaction is catalysed by moisture. The formation of passivated oxide layers
reduces considerably the rate of corrosion.
At high temperature, dry corrosion occurs and, in particular, dry oxidation.
When a dense and compact oxide coating is formed, it protects the material surface
from subsequent attacks.
In organic polymers, degradation results from the rupture of a certain number of
bonds in the chain under the combined or individual action of light, temperature and
oxygen in the air. These very complex reactions occur in general via a radical
mechanism. They are considerably inhibited by substances (stabilizers) able to pre
vent the formation of, or to neutralize, free radicals.
Ceramics, which are mainly made up of oxides, are in general very stable.
Nevertheless, in certain cases, they can undergo severe corrosion under the influence
of chemical pollutants present in the atmosphere and precipitation.

15.7. Illustrative example: corrosion of reinforced concrete

by sodium chloride salts
For a long period, concrete was regarded as a material, which was inert and
stable over time. However, after some years, reinforced concrete constructions often
began to show signs of degradation. This is the case for certain bridges built in the
1950 and in the 1960s. These are often highly corroded with a significant reduction
in their safety.
Under normal conditions, the steel reinforcement of concrete is well protected.
Indeed, the free water held in the pores of the concrete is strongly alkaline (pH ~ 13).
This environment favours the development of protective coatings (passivation),
which form after hydration of the cement. This protective coating consists of
compact iron oxide preventing corrosive attack of the metal.
The alkalinity of the concrete decreases gradually with time by the washing out
of certain hydrated components in the cement. Among these, slaked lime Ca(OH) is 2

the most soluble. Water dissolves it gradually whatever its pH. This dissolution,
which reduces the pH of the material, is accompanied by various changes decreasing
the strength of the concrete. When slaked lime passes into solution without reacting,
there is an erosion process.
When concrete comes into contact with C 0 contained the air, which is the
2

case in almost all constructions, the calcium hydroxide formed during the hydration
of the silicates is neutralized:

Ca(OH) + C 0 - C a C 0 + H 0
2 2 3 2 (15.23.)

The alkaline character decreases until a pH ranging between 8 and 9.5 is

reached. The iron reinforcements thus find themselves more or less quickly in a more
aggressive environment and degradation can start. The kinetics of this carbonation
 Corrosion, degradation and ageing 397

depends mainly on the permeability of the concrete and the composition of the sur
rounding atmosphere.
Many civil engineering works, hardly 20 years old, already show serious degra
dation due to this phenomenon. The corrosion products of the steel reinforcements
occupy a much greater volume than non-degraded metal causing swelling, cracking
or rupture of the concrete structure. This results from an easier access for the aggres
sive agents and an aggravation of the corrosion.
The salts dissolved in water in contact with the concrete structure reduce the
passivation layer of the reinforcement and accelerate corrosion considerably. This
comes mainly from sodium chloride introduced by snow clearing agents on the roads
or seawater in the case of marine structures or those located in coastal regions. These
salts gradually penetrate the concrete through pores in the material. In the case of ma
rine structures, the salt penetration is enhanced by drying and immersion cycles.
It is in general the structural elements in occasional contact with seawater that
are most vulnerable. Infrastructures located at great depth are much less corroded
because of the lack of oxygen. The presence of salts locally destroys the passivation
layer, allowing a direct contact between metal and the electrolyte, leading to the
dissolution of iron:

Fe Fe + +
+ 2e" (15.25.)

To prevent the corrosion of the metal reinforcements, electrochemistry can

provide a certain protection, for example by the use of a sacrificial anode, following
the same principle as adopted for ships or underground pipelines (figure 15.6.). A
metal plate (anode) is connected to the structure forming an electrochemical cell with
the iron (cathode). The anode deteriorates slowly thus protecting the reinforcing iron.
A weak D.C. current can also be used to ensure the protection of steel, even in the
presence of water having a low pH. It is this latter method that will be used for the
protection of the Berlin ring road where the structures are highly corroded.
The galvanization of metal reinforcements can also be envisaged. Under the
initial pH conditions of the concrete (pH ~ 13), the galvanized reinforcement is cov
ered with a compact film of calcium hydroxyzincate (Ca(Zn (OH) ), which protects
2 6

the metal reinforcement from corrosion. When the pH of the concrete decreases,
protection of zinc by passivation is no longer effective. Nevertheless the zinc layer
plays its usual protective role and corrosion is considerably reduced. However, this
protection is not durable because the zinc layer is gradually consumed. The degrada
tion of the concrete is a phenomenon of a great complexity and still cannot be con
sidered as completely under control.

15.8. Exercises
15.8.1. What is the electric potential of an electrolytic half-cell consisting of a
copper plate placed in a solution containing 200g of C u ions diluted in 1000 g of
+ +

water?
15.8.2. Copper is deposited by an electrolytic process on a cathode of area 1 cm , 2

using a current having an intensity of 10 A. By using Faraday's equation:

398 Introduction to Materials Science

ltM a

w =
nF
where
w is the weight of covering deposited or the loss by corrosion deposited
per unit of time (gs );
_1

7, the intensity of the electrical current (A);

M the atomic weight of the metal (g);
a

t, the time (s);

n, the number of electrons produced or consumed;
F, the Faraday's constant = 96 500 (C mol" or A s m o l ) ;
1 -1

Calculate:
the weight of copper deposited per hour;
the time necessary to manufacture a 1 coating of copper.
15.8.3. An electrochemical cell consists of a copper electrode and a zinc electrode of
100 cm . The current density on the surface of each electrode being equal to
2

0.05 A c m , calculate:
-2

the corrosion current;

the current density at the zinc anode;
the loss of zinc per hour and the weight gain of the copper cathode.
15.8.4. Describe briefly the action of light on the natural rubber used in car tyres.
How can this degradation be limited?
15.8.5. Explain the difference between electrochemical oxidation and reduction.
What reaction occurs at the anode and what happens at the cathode?
15.8.6. What is the time necessary to produce a 2 mm thick electrolytic deposit of
nickel on a 10 cm steel surface of when using a current of intensity 5 A in a solution
2

of nickel sulphate?
15.8.7. A 1 cm volume of magnesium is oxidized into MgO. What is the volume of
3

oxide formed?
15.8.8. At the temperature of 1000 C, nickel oxidizes according to the parabolic
equation:

x = ^[kt

where
represents the thickness of the oxide coating;
k is a constant equal to 3.9 1 0 cm s for an atmosphere of oxygen;
1 2 2 _1

/ represents time.

Calculate the time necessary to obtain a nickel oxide film of 1 mm thickness, given
that the transformation rate is not affected by the variation of the thickness of film.
15.8.9. What are the main precautions which should be taken to avoid the electro
chemical corrosion of steel components of a bridge?
15.8.10. What are the differences between the modes of corrosion of metals and
ceramics?
 Corrosion, degradation and ageing 399

15.9. References and complementary reading

A. BAUTISTA and J . A . GONZALEZ, Cement and Concrete Research, 26, 215, 1996.
J.P. CHILTON, Principles of Metallic Corrosion, 2nd ed., The Chemical Society, 1973.
U.R. EVANS, The Corrosion and Oxidation of Metals, Arnold, London, 1978.
D.R. GABE, Principles of Metal Surface Treatment and Protection, 2nd ed., Pergamon, Oxford, 1975.
T. KELEN, Polymer Degradation, Van Nostrand Reinhold, New York, 1983.
O. KUBASCHEWSKI, B.E. HOPKINS, Oxidation of Metals and Alloys, 2 ed., Butterworths, London,
nd

1962.
D. LANDOLT, Corrosion des metaux et chimie de surfaces des metaux, Presses polytechniques et
universitaires romandes, Lausanne, 1993.
A. NEVILLE, Materials and Structures, 28, 63, 1995.
J. F. RABEK, Photodegradation of Polymers, Springer, Berlin, 1996.
J.C. SCULLY, The Fundamentals of Corrosion, 2 ed., Pergamon, Oxford, 1975.
nd

W. SCHNABEL, Polymer Degradation, Hanser, Munich, 1992.

H.H. UHLIG, Corrosion and Protection, Dunod, Paris, 1970.
 Chapter 16.

Composite materials

16.1. Objectives
To review the main types of fibre-reinforced composites.
To establish the mechanical behaviour of fibre-reinforced composites in
the elastic range.
To identify the most important sandwich structures.
Modern technology needs materials combining at the same time rigidity, me
chanical strength, high tenacity and considerable lightness. It is not possible to find
these physical characteristics in a single material, and to achieve this goal it is neces
sary to use a composite material Therefore, composite materials, and in particular,
fibre-reinforced composite materials containing are increasingly playing a significant
role in advanced technologies. This chapter will be essentially limited to a basic
study of these materials, with the emphasis on composites with a polymer matrix.

16.2. Introduction
Composite materials have a very long history. Several thousand years ago, the
craftsmen of ancient Egypt improved the mechanical properties of bricks by adding
short straw to fresh clay. Certain natural materials such as wood (cellulose fibre in a
lignin matrix) and bone (complex mixture of collagen and calcium phosphate) are
composite materials.
There is no simple definition of a composite material. In fact, a great number of
modern materials are alloys. They consist of a mixture, on a microscopic scale, of
several phases, having distinct structure and chemical composition, contributing
synergistically to the physical, chemical or mechanical properties. Strictly speaking,
certain alloys (e.g. impact polystyrene, ABS, carbon steel) can be classified as com
posite materials. Polymers containing finely divided mineral substances ( T i 0 - 2

CaCOs) or ceramics such as porcelain also constitute composite materials.

More and more, the designation composite materials is reserved for materials
made up of a continuous matrix reinforced with high strength fibres. These fibres are
immersed in a matrix that keeps them in place and ensures the cohesion of the whole.
The mechanical properties of fibre-reinforced composites depend on the strength, the
quantity, the dimensions and, above all, the orientation of fibres. This study is re
stricted to composites based on an organic matrix, which are currently the most
widely used. Sandwich structures, combining, on a larger scale, various materials
used to form functional structures, will also briefly be considered.

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

402 Introduction to Materials Science

Composite materials containing fibres enable a wide range of objects, machine

components, industrial structures to be manufactured that would not be feasible with
traditional materials. Most fibre-based composites were already known and used at
the beginning of the 1980s. What brought about the development of this type of ma
terial in the 1990s, was mainly the result of the computer revolution, which was at
the origin of computer-aided design/computer-aided manufacturing (CAD-CAM)
coupled with the numerical control on the machine tools. These new techniques can
be applied to all types of materials, but it is with composite materials that these
methods find their most extensive application.
The technological importance of fibre reinforced composite materials kept
growing during the last decade. Their introduction revolutionised certain areas, in
particular those related to sports, leisure and advanced technologies. Structural com
ponents for commercial aircraft and the complete shells for light aircraft are now
manufactured in fibre reinforced composites. However, the composites have not
replaced aluminium alloys as basic structural materials to the extent that certain ex
ploratory studies in the 1980s were predicting. These advanced materials have been
extensively called on to build the entire shell only in the case of light aircraft.
However, in space technology and for satellites in particular, there is very extensive
use of these materials. Even the more traditional automobile industry is gradually
adopting composites for certain structural parts (suspension springs) or body sections
(hatchback doors).

16.3. Composite materials containing fibres

16.3.1. Elastic behaviour
Only the simple case of composites where continuous fibres are uniaxially
oriented will be considered here. It will also be assumed that the components exhibit
elastic behaviour and that there is perfect adhesion between the fibre and the matrix.
When a tensile force is applied to the section S of the composite, in a direction
parallel to that of the fibres (figure 16.1(a)), the deformation is identical for the fi
bres (/) and the matrix (m) fe = = e = ), and the force F applied to the speci
f m c

men is distributed on the fibres and the matrix.

The force F/ required for elongation of the fibres is given by:
F =eE/Sf
f (16.1.)

island Sfave respectively the Young's modulus and the surface area occupied by the
fibres.
The force F required to deform the matrix is equal to:
m

F = eE S m m (16.2.)

E and S respectively represent the modulus and the surface area occupied by the
m m

matrix.
The force F needed to deform the composite in a direction parallel with fibres is
written:

F F/+ F s(EfSf +
c m ES)
m m (16.3.)
 Composite materials 403

t
- " l b )

1
0,5 1
Fibres volume fraction /

Figure 16.1. Variation of the modulus of elasticity E of a uniaxial fibre reinforced com
c

posite as a function of the volume fraction of the fibres when the fibres are (a) parallel
and (b) perpendicular to the tensile direction. E, and E represent the modulus of the ma
f

trix and the modulus of the fibres respectively.

The modulus of elasticity of the composite is derived from expression (16.3.):

(16.4.)

As Sj/S and S /S are equal to the volume fraction of the fibres / a n d to that of the
m

matrix (1 -f) respectively, then:

E =E (\-f)
c m + Ef
f (16.5.)

In fibre-reinforced composites, the Young's modulus of the fibre is normally

much higher than that of the matrix, and practically all of the tensile force is applied
to the fibre phase. To underline this, a composite containing 50 % by volume of
carbon fibre (E/= 300 GPa) in an epoxy resin (E = 2 GPa) can be considered where:
m

E = 1 + 150= 151 GP
c (16.6.)

Practically, the stress is transferred completely through the fibres.

If the force F is applied in a direction perpendicular to that of the fibres (figure
16.1. (b)), the stress applied to fibres and the matrix is identical:
a =a =
c m Gf ( 1 6 7 )

In this case, the elongation of the composite is equal to the sum of the elongations of
the fibres and the matrix:
Al =Al +Al c f m ( 168)

The relative deformation of the composite is related to the relative deformation of

c

the fibres and the matrix by:

= f+ (\-fi
c f m (16.9.)
404 Introduction to Materials Science

By introducing the elasticity moduli, it can be written:

(16.10.)
E m

Starting from (16.7), we found:

(16.11.)
E c Ef E
m

By calculating the modulus E of the carbon fibre-epoxy composite containing

c

50 % by volume of carbon fibre following (16.11.), then:

E = - - 4 GPa (16.12.)
[0.5/300 + 0 . 5 / 2 ]
c

Therefore, up to a relatively high volume fraction, the fibre reinforcement has

very little influence on the rigidity of the composite (figure. 16.1. (b)). Under these
conditions, there is little benefit from the high modulus of elasticity of the fibres. As
can be see in figure 16.2, identical conclusions can be drawn for the tensile strength,
which is a maximum when the tension acts parallel to the fibres.
To take advantage of the exceptional mechanical properties of the fibre-based
composites, it is absolutely essential to use them under conditions where the tensile
force acts in a direction parallel or quasi parallel to the fibres.
This explains why composites containing unoriented fibres have much lower
mechanical properties compared with bi- or, above all, unidirectional composites.
The elasticity modulus of a fibre-reinforced composite with random orientation (3D)
is equal to 1/5 the maximum modulus of a unidirectional composite of the same
composition. In the case of a statistical distribution (2D) in a plane, the modulus
parallel to the plane reaches 3/8 the maximum modulus of an unidirectional
composite of the same composition.
Until now, only continuous fibres have been taken into account. However, com
posites made up with short fibres have a mechanical strength equivalent to that of
composijtes with fibres of infinite length if the aspect ratio i.e. the ratio of the fibre
length to its diameter (l/d) exceeds a certain critical value of the order of 100. The
critical value l is thus all the greater when the diameter of the reinforcing fibre is
c

high. In these calculations, a perfect adhesion between the fibre and the matrix is
assumed, both of which have an ideal elastic behaviour.

16.3.2. Principal types of fibre reinforcement

The principal types of fibre reinforcement used in the majority of organic com
posites are limited to three groups: glass, carbon and aromatic polyamides (aramid)
fibres. Their main mechanical and physical characteristics are shown in table 16.3.
Glass fibres currently form the most widely used reinforcing agent because of
their mechanical characteristics and moderate cost. The majority (-95 %) of glass
fibres actually used are aluminium borosilicates containing small quantities (~1 %)
of fluxes ( N a 0 + K 0 ) (E glass fibres) which lower the viscosity of glass and allow it
2 2
 Composite materials 405

(GPa) (MPa)

0 45 90 0 45 90
Angle +~ Angle
(a) (b)

Figure 16.2. Effect of the fibre orientation of on the modulus of elasticity E (a) andc

strength R, (b) of an epoxy resin-fibre carbon composite (after Davis, 1984).

Table 16.3. Significant mechanical and physical characteristics of the main types of
reinforcement (virgin filaments) [a].

Glass fibres Carbon fibres Aramid fibres

R S HM[b] HR[c] Kevlar 49 [d]

Young's modulus (GPa) 73 87 -500 -230 131

Tensile strength (MPa) 3450 4600 -2300 -4000 -2800
Compressive strength (MPa) 1200 1300 1300 1800 500
Elongation at rupture (tensile) (%) -4.7 -5.4 -0.5 -1.4 2.6
Diameter (pm) 5 to 25 6 to 10 12
Density (g c m ) -3
-2.58 -2.50 -1.9 -1.77 1.45
[a] There is a great spread in the results given in the literature, the values given here are average values;
[b] HM = high modulus; [c] HR = high strength; [d] Du Pont de Nemours registered trademark.

to be spun at lower temperatures. High mechanical strength glass fibres (S glass

fibres) are also manufactured in which there is practically no boron oxide nor alka
line metal oxide. These compositions have considerably higher melt viscosities and
must be spun at higher temperatures. This significantly increases cost.
glass fibres are spun at a temperature of 1250 C at very high speed (several
decameters s ) through dies made of a platinum/rhodium alloy and are cooled in air
-1

and then with water. Each die has a significant number of holes (between 400 and
4000), approximately 2 mm in diameter. After sizing, the filaments, which have a
diameter between 5 and 25 , are recombined to form the basic fibre. S fibres are
obtained by the same process but they are spun at a temperature higher than 1500 C.
The manufacturing process for carbon fibres has been described in the illustra
tive example of chapter 2. Among the various types of carbon fibres currently mar
keted, two main categories may be identified: high modulus fibres (HT) and high-
strength fibres (HR). Their principal characteristics are also included in table 16.3.
The structure of the poly(/?-phenylene terephthalamide) fibre, marketed under the
name of Kevlar by Du Pont or Twaronby A K Z O i s described in chapter 5. Its
principal characteristics also appear in table 16.3.
406 Introduction to Materials Science

It is interesting to compare the mechanical characteristics of the three families of

reinforcing fibres shown in table 16.3 and figure 16.4. Their tensile strength, in
general, lies within a relatively narrow range, between 3000 and 5000 MPa but the
values for the Young modulus of these fibres are quite different. They vary in a pro
portion ranging from 1 to 7 between glass fibre and the high modulus carbon fibres.
This characteristic has a direct consequence on the rupture elongation of the fibres.
This is obviously the lowest for the highest modulus fibres, such as carbon fibres.

t
6000
S glass fibre

_ 4000

3000
b

1 2000

I 1000

0 2 4 6 %
Tensile strain (%)

Figure. 16.4. Tensile curves for various reinforcing fibres (after Hull, 1981).

For fibres having a lower modulus such as glass fibres, values of rupture elon
gation higher than 5 % can be observed, which is exceptional for a material showing
brittle fracture. Certain trichites (whiskers) exhibit similar mechanical behaviour but
with, in general, a lower rupture elongation. Glass fibre represents an almost unique
case of linear elasticity characterized by such behaviour. This extraordinary me
chanical behaviour, is induced by the extremely high draw ratio of the extruded glass
filament, which aligns the silicate chains parallel to the fibre axis and eliminates most
of the surface defects (flaws) responsible for the brittleness of bulk mineral glass
(chapter 13). Nevertheless, the rupture elongation of fibres decreases with their incor
poration in a composite. Thus for example the rupture elongation of S glass fibres
drops from 5.2 to about 4 %; that of Kevlar fibres from 2.5 to roughly 2 %.
These significant differences observed for the rupture elongation of the various
fibres types is of great importance in the choice of reinforcing agents. Thus, in order
to manufacture flexible and deformable components, glass fibres are preferably used.
These structural parts undergo significant deformations. The ski and the pole for the
pole vault provide two examples of items subjected to important mechanical defor
mations during use. A high proportion of glass fibres is involved in their manufac
ture. On the other hand, in aircrafts where very rigid elements showing little de-
formability are sought, carbon fibres are predominantly used. Various elements of
aircrafts are currently made from carbon fibre reinforced organic composites: tail
planes, certain ailerons, etc.
 Composite materials 407

The compressive strength of various fibres is also an important factor to con

sider when selecting a reinforcing material. Kevlar fibre (table 16.3.), which has
poor compressive strength, cannot be used for certain applications because of this
characteristic. Thus, in the case of flexural deformation of beams, Kevlar fibres can
never be used in the part of a component working in compression.
Other factors play a role in fibre selection. Lightness, electrical and thermal
conductivity are important for certain applications. Another essential element is the
economic aspect strongly favouring glass fibre. Indeed, the other reinforcing fibres
are between 10 and 500 times more expensive. Lowering of the cost of carbon fibre
would promote a much more extensive use of this fibre.
Various surface treatments for reinforcing fibres have been developed to in
crease adhesion between the fibre and the matrix and so facilitate load transfer be
tween these elements. The size, which is more developed for glass fibre, is applied by
spraying with an emulsion containing a number of components and serves several
purposes. Its first a role is to preserve the physical integrity of fibre during
processing. Glass fibre is, indeed, very sensitive to handling which can induce sur
face defects and so decrease its mechanical strength. The coating formulation in
cludes tacky film-forming agents such as emulsions or suspensions of polyvinyl
acetate) in order to protect the surface of the filaments and make it possible to com
bine them to form a multi-filament fibre. An important component of the coating is a
coupling agent, which is a reagent able to bind chemically to both the organic matrix
and the glass fibre surface and which forms, to some extent, a strong link between the
fibre and the organic matrix. The coupling agent is an adhesion promoter. The most
commonly used coupling agents are silanes which mode of action is shown in figure
16.5. Lubricants and anti-static agents are also often included in filament coating
formulations.
For other reinforcing agents, fibre surface treatment processes are less de
veloped. For carbon fibres, surface oxidation treatments have been elaborated; these
form polar groups ( - C O - O H and -OH) on the surface of the fibres, which act as
adhesion promoters. Surface modifications promoting the fibre - matrix adhesion
were also developed for aramid fibres. However, composite ballistic armours made
with aramid fibres have a better penetration resistance when the fibre - matrix adhe
sion is reduced.

16.3.3. Matrices
The three major material classes (metals, ceramics, organic polymers) can be
used as matrices for fibre reinforced composite materials. Because of their lightness
and ease of processing, organic matrices are currently the most widely used. Two
main categories of polymeric matrices are concerned: thermosetting (resins) and
thermoplastic.
Thermosetting matrices (epoxy resin, unsaturated polyester, etc.), are the most
developed. These resins have a certain number of advantages and, in particular, great
fluidity during reactive processing (at the cross-linking step), which favours penetra
tion of the resin, wetting of the reinforcing agent and avoids the presence of bub
bles. It should be remembered, however, (chapter 5) that thermoset shaping has to be
408 Introduction to Materials Science

(a) R-SiX 3 3 H 0 2
- R- S i ( 0 H ) 3

R R R
H0-Si-0H HO-Si-OH HO-Si-OH
(b) .0 N
H' H' H'
"0'' ^0'' "0''

Glass
wsmssm
R R R
O-Si - 0 - S i - O - S i - 0
(c) 1
0

6 0

mmmmm, ciass

Polymer
R R R
( )
d
O-Si-O-Si-O-Si-0
I I
0 0 0
1 t I
'MWMmm. Glass
Figure 16.5. Reaction of a coupling agent at the glass fibre -polymer interface: (a) hy
drolysis of the coupling agent with formation of silanols; (b) formation of a hydrogen
bond between the silanol and the glass; (c) formation of a chemical bond at the fibre sur
face and (d) with polymer; X represents a halogen and M, a cation (after Hull, 1981).

completed before the crosslinking reaction (curing), which forms an insoluble and
infusible three-dimensional network. To avoid any premature chemical reaction
before processing, these resins must be stored at low temperature. Manufacture of
this type of composite requires cold rooms to store fibre felts and fabrics impregnated
with non cross-linked resins ('prepregs').
Thermoplastic matrices (figure 16.6.) are less widely used to prepare composite
m a t e r i a l s . Among these, aromatic poly(etheretherketones) such as P E E K
(T = 145 C and T = 340 C) are very interesting for aerospace applications. The
g m

great advantage of thermoplastic matrices is that they allow thermoforming starting

from a sheet (figure 16.7.), as well as recycling. However, the use of these resins
presents other disadvantages resulting in particular from their great melt viscosity
and their lower temperature resistance when compared to thermosets.

16.3.4. Examples of reinforced materials

The unidirectional fibre reinforced composites (table 16.8.) have mechanical
properties at least equivalent to those of steels for a specific gravity mostly in the
range 1 to 2. They are thus light and strong materials but in o r d e r to take advan
tage of these exceptional mechanical properties, the composite should be stressed in
a direction parallel to the fibres. For the same type of fibre, the mechanical
performance level reached at ambient temperature does not depend on the matrix.
This result is logical since practically the applied load is completely transferred to the
fibres (16.6.).
 Composite materials 409

Poly(/7-phenylene-ether-ether-ketone)

O H
II 1 II I Ii

Poly(N-hexamethylene-adipamide (PA-66; Nylon-66)

Poly(bisphenol-A carbonate) (PC)

Poly(butyleneterephthalate) (PBT)

Poly(p-phenylene sulphide)

Figure 16.6. Principal thermoplastic matrices used in composite materials.

Figure 16.7. Example of components obtained by thermoforming of sheets of carbon fi

bre reinforced PEEK (ICI APC 2).
410 Introduction to Materials Science

Table 16.8. Mechanical characteristics and density of unidirectional fibre reinforced

composites and of some traditional materials [a].

Materials Density
(GPa) (GPa) (gem )-3

PEEK + 60 % continuous carbon fibre 120-170 1.7 1.61

Epoxy + 60 % carbon fibre 145-210 1.4-2.8 1.6
Unsaturated polyester + 60 % aromatic polyamide 70 1.3 1.4
fibre (Kevlar)
Unsaturated polyester + 50-80 % glass fibre 20-70 0.65 - 0.75 1.6-2.0
Epoxy + boron fibre 200 1.4 ~2
Aluminium + boron fibre covered with 4 C 3 200 1.4 2.65
Carbon fibre / carbon Composite 150-200 1.2-1.5 1.3-1.6
Light alloy (aluminium) 80 0.4 2.8
Steel 210 0.5-2 7.8
Unsaturated polyester 2-5 0.02 - 0.07 1.1 - 1.3
Epoxy 2.5 0.015-0.07 1.1-1.5
PEEK 2.8 0.105 -
[a] There is a great spread of results in the literature and the values given here are indicative only. In
the case of composites, the values are given for stresses parallel to the fibres. The percentages indi
cated are by weight.

Figure 16.9. Aerothermal cover. Bismaleimide resin and intermediate modulus carbon
fibres manufactured by Resin Transfer Moulding (RTM) (from B. Poulaert (SONACA),
1996)
 Composite materials 411

Figure 16.9. shows a piece of equipment made from a carbon fibre-thermoset for
aerospace application. The use of thermosets allows very complicated structures to
be manufactured.
The temperature resistance of the various fibre-reinforced composites depends
primarily on the behaviour of the matrix. For composites with an organic matrix the
continuous service temperature seldom exceeds 150 C.
The mechanical behaviour of a composite material varies considerably with the
orientation of the fibres. To obtain a high mechanical strength, it is necessary to ori
ent the fibres parallel to the direction of the highest stresses. In the case of the com
posites, material production and component manufacture generally take place to
gether. This explains the importance of computer aided design and manufacturing
techniques (CAD-CAM) in the production of fibre-reinforced components.

16.4. Sandwich structures

16.4.1. Laminated sheets and shells
To have a high mechanical flexural strength, it is not necessary to produce the
entire structure in a high performance material. Such considerations gave rise to
sandwich structures made up of at least three layers (figure 16.10. (b)). The outer
skins are made out of a material (composite or not) having high stiffness as well as
high tensile and compressive strength. The core of the sandwich structure is made of
a light material with low rigidity and low mechanical strength. The aluminium-balsa
wood panels, which are used in pleasure boats and aircraft, are a typical example of
such sandwich structures. Because of the difference in rigidity between core and
skins of the sandwich structure, there are significant shear stresses at the interfaces
between the various materials making up the sandwich. This often causes failure by
delamination, i.e. by separation of the layers from each other.

(a)

(b)

Stress in Strain in
the thickness the thickness

Figure 16.10. Distribution of forces on a beam in flexion: (a) solid uniform structure and
(b) sandwich structure.
412 Introduction to Materials Science

16.4.2. Examples of sandwich structures

In order to combine rigidity, high mechanical strength and lightness, more com
plex structures than those presented in figure 16.10. (b) are produced. These are, for
example, made up of a core with a corrugated or honeycomb structure (figure 16.11.).
These sandwich structures are encountered both in simple (corrugated cardboard) and
advanced (aeronautics-aerospace) applications. Currently, very complex sandwich
structures such as helicopter rotor blades (figure 16.12.) and a satellite equipment
bays (figure 16.13.) are being produced.

(a) (b) (c)

Figure 16.11. Various sandwich structure constructions: (a) solid core (foam-balsa); (b)
honeycomb; (c) corrugated core.

Figure 16.12. Helicopter rotor blade (after Gole, 1979).

16.5. Summary and Conclusions

By combining materials of various types, composite structures are produced
which have a combination of physical and mechanical properties, which is not found
in each component separately.
This chapter has been limited to fibre-reinforced composites with an organic
matrix that constitutes a rapidly expanding area. Advanced technologies require
materials combining at the same time lightness and superior mechanical properties.
These characteristics are impossible to achieve in practice with single materials.
For fibre reinforced composites, the orientation of continuous fibres, parallel to
the mechanical stress, allows an almost total transfer of the stress to the fibres, which
makes it possible to realise light and strong materials. In a direction parallel to the
fibres, light materials have been obtained with mechanical characteristics equal to
and even higher than those of single high performance materials. An important point
that has not been covered in this elementary presentation is the extraordinary resis
tance of the fibre-reinforced composites to fatigue. This important characteristic
results from the fact that they are bonded structures, which distribute the stresses
much more uniformly.
 Composite materials 413

Figure 16.13. Satellite Equipment Bay. Modified epoxy resin and high modulus carbon
fibres (PAN) for sandwich panel skin. Cyanate ester resin and high modulus carbon fibre
(PAN) for the struts, (from B. Poulaert (SONACA), 2000)
414 Introduction to Materials Science

This chapter has also envisaged the study of laminated structures, which com
bine single materials and composite materials in more complex structures.

16.6. Illustrative example: materials for skis

of the 2 1 century st

The materials used for skis at the beginning of the third millennium are almost
identical to those used in the 1980s. What differentiates the ski of the 2 1 century s t

from the ski of the 1980s, is above all the evolution of its shape which modifies its
behaviour on snow in a spectacular manner.
The most significant parameter of the ski shape is the profile of its side cut. As
indicated on figure 16.14., the side cut is the line describing the curved portion of the
side contour between the widest part of the ski shovel and tail, the narrowest part of
the runner following the bottom edge. The side cut determines the capacity of the ski
to turn. When the bottom edge is in contact with snow, the ski in practice describes a
circle which radius is determined by its side cut. The greater the curvature of the side
cut, the smaller the radius of trajectory followed by the ski. A traditional ski is lim
ited to circles with radii larger than 35 to 40 meters. By carving the side camber of
the skis and by decreasing their length, it is possible to reduce the radius of the turn
by more than 50 %. A ski is then obtained which is easier to control in turns and
consequently more convenient for an average skier. An important function of the ski
is to maintain contact with the piste. Good contact allows the skier to control his
trajectory. The ski must be at the same time relatively soft in flexion and rigid in
torsion. These qualities are partly contradictory.
The evolution of the ski shape goes back to a little more than ten years. It was
mainly the result of progress made in computer-aided design and manufacturing
(CAD-CAM) and the introduction of numerical control of machine tools. It is now
possible to manufacture extremely complex moulds. This transformation of manufac-

Traditional ski

Side cut
Trajectory

Parabolic ski

Tail Shovel

Side cut

Trajectory

Figure 16.14. Parameters of shape and trajectory of the traditional and the parabolic ski
(after to Bauvois and Reverchon, 1998).
 Composite materials 415

turing methods was coupled with digital simulation methods. By working on models,
various technical parameters of the ski were optimised and the use of materials em
ployed rationalised.
This change of the ski shape has also affected its cross section which was rec
tangular in the 1980s. As an example, figure 16.15. shows in an exploded view a
modern ski. The main materials used are shown in this figure. These are also referred
to in table 16.16. An important selection criterion is lightness. The classification of
materials is given in table 16.16.

Figure 16.15. Section of a modern ski. The principal materials are indicated. The
presence of two vibration absorbing systems (VAS) should be noted: the external and in
ternal VAS (after Bauvois and Reverchon (Rossignol), 1998).

The ski is primarily subject to flexing and the reinforcing materials used, which
are positioned on both sides of the core, work in tension in the lower part and in
compression in the higher part of the ski. Furthermore, the ski must be able to resist
high amplitude deformations without either warping or breaking. To satisfy this re
quirement, essential to the correct use of the ski, reinforcing materials having a high
elastic limit are mainly used. The principal reinforcing agent is glass fibre, which is
selected because of its very interesting performance/price ratio; in particular, it has a
high elongation compatible with high amplitude deformations. Glass fibres are used
in the form of fibre cloths sewn onto a polyester veiling. Aromatic polyamide (Kev-
lar 49) fibres are also employed in the form of longitudinal fibre ribbons, possibly
fixed on a polyester sheet. However, Kevlar is only used in the lower part of the ski
because this fibre has a low compressive strength. Carbon fibre, which has lower
rupture elongation, is seldom used because it frequently breaks before the other fibre
reinforcements intervene to take part in the load carrying. Another factor, which
limits the use of carbon fibre, is its very high price.
Aluminium alloys developed initially for aeronautical applications are used to
yield skis with both torsion stiffness and flexion softness, thus enabling precise turns.
The presence of an external vibration-absorbing system (external VAS) stuck onto
416 Introduction to Materials Science

the ski and of an internal vibration-absorbing system (internal VAS) positioned in

side the ski under the core is indicated on figure 16.15. The external VAS is a device
made up of a sheet in Zicral alloy associated with a viscoelastic material and at
tached locally to the upper part of the ski. The internal VAS consists of a rubber with
a high damping coefficient which coats cables made up of twisted steel strands. The
VAS selectively damp the vibrations of the ski reducing its manoeuvrability. It re
mains to mention the insert, which is a paper-phenolic resin laminate, used to in
crease the pullout resistance of the binding screws. As a core, semi-rigid polyure-
thane foam is used. This is produced by injection of the reactive liquid components
(diisocyanates and polyols) into a mould, these components being mixed just before
injection. The blowing agent used does not contain fluorocarbon compounds.
The bottom surface element constitutes, with the steel bottom edge, the struc
ture for protection of the lower surface of the ski. Their principal role is to ensure
sliding of the ski. The running surface is actually made of low friction high-density
polyethylene (HDPE). It is not the ideal material because its scratch resistance is low
but it is the only material to have a frictional behaviour adapted to all types of snow.
For the current skis, an extruded HDPE of molecular weight from 300000 to 500 000
is used.
The extruded material is being replaced progressively by very high molecular
weight (-2 000 000) HDPE. This polymer has too high a viscosity to be extruded and
is processed by sintering, using a powder technology similar to that used for the
manufacture of technical ceramics. It is peeled from a cylinder of the size of a
Gruyere cheese made by sintering. This material has an excellent wax penetration
because of its microporosity resulting from its manufacturing process. The HDPE
must be flame treated before being adhered. The bottom edges are made from a
chrome steel strip with 0,6 % of C (Rockwell hardness 4 9 - 5 1 ) .
The top surface element (Rossitop) is styrene-acrylonitrile copolymer modified
with thermoplastic polyurethane rubber (impact). The sidewalls can be also manu
factured out of modified styrene-acrylonitrile copolymer, reinforced with short glass
fibres. Laminates made up of paper impregnated with a phenolic or melamine ther
mosetting resin are also used for this application.
The various constituent parts of the ski are bond together by an adhesive. Ther
mosetting epoxy resins are mainly used. Just before crosslinking, the thermosetting
polymer goes through a stage of very low viscosity, which allows adhesion without
bubbles. The adhesive properties of these cross-linked resins are remarkable. Joining
is carried out under a pressure from 0.8 to 1.0 MPa at a temperature between 100 and
120 C during approximately 15 min. To avoid any deformation, the skis are then
cooled under pressure to a temperature lower than the glass transition of the adhe
sive. Joining by adhesive makes it possible to obtain a structure particularly resistant
to fatigue.
There are other manufacturing technologies. For example, in the 'in situ'
process, the ski components are first introduced into a mould. The pre-mixed liquid
components of a polyurethane foam are then injected and foamed inside the mould.
After expansion and hardening, this foam becomes the core of the ski and it also
forms the adhesive binder holding together the various elements of the ski. This
process constitutes one possibility to manufacture Cap skis, which do not have side-
walls.
 Composite materials 417

Tables 16.16. Principal materials used in the manufacture of the skis.

Reinforcement and structural materials

Materials Nature Use

Glass fibres. Textile made up of superimposed Reinforcing fibre reducing the
fibre cloths sewn together. ski weight.
KevlarK49 [a]. Aromatic polyamide. Reinforcing agent for the ski
lower part.
Carbon fibres. High-strength unidirectional fibres. Seldom used; small elongation
at break; expensive.
Aluminium alloy Aeronautical aluminium alloy with Structural layer; reinforcement
(ZICRAL [b]). high elastic limit. in torsion.

Fastening, filling and damping materials

Materials Nature Use

Epoxy resin. Two component thermosetting Adhesive fastening the various

polymer (resin + hardener). elements of the ski.

Polyurethane foam (PU). Cellular, expanded thermoset with Semi-rigid light material not
out fluorocarbon compounds. sensitive to moisture, used as
core.

Wood. Natural unidirectional composite of Light material used as core in

cellulose fibres in a lignin matrix. practically 50 % of the skis.

External and internal Viscoelastic material attached to a Damping of vibrations, harmful

vibration absorbing system ZICRAL sheet or coated on twisted to the ski stability.
(VAS). steel cables.

Protective materials

Materials Nature Use

Bottom surface. High density polyethylene of high or Sliding surface and protection of
very high molecular weight. the bottom surface. Made from
low friction HDPE with good
wax penetration.

Bottom edge. Chrome steel with 0,6 % of C Metal strip forming the lateral
(Rockwell C hardness 49-51). termination of the bottom sur
face. Provide protection and
transmission of the steering
forces.

Phenolic or melamine Paper laminate impregnated with a Sidewalls of the skis and manu
laminate. thermosetting phenolic or melamine facture of inserts designed to
resin. increase the pullout resistance of
the binding screws.

Thermoplastic materials. Styrene- acrylonitrile copolymers Upper protection of the ski.

modified with thermoplastic polyure- Short glass fibres reinforced,
thane rubber (impact). also used for the sides manufac
ture.

[a] registered trade marks of Du Pont de Nemours; [b] registered trade mark of Pechiney Aluminium (after
Bauvois and Reverchon, 1998).
418 Introduction to Materials Science

It should be noted that, in spite of the remarkable development of synthetic

materials, wood remains the core material in more than 50 % of skis. Indeed, this
natural material, of low density, microporous and with attractive cost, has a good hot
crush resistance allowing hardening of the adhesives. Its principal defect is moisture
uptake.
The first function of a ski is to ensure good sliding, i.e. to obtain a minimum
friction between the bottom surface and snow. The friction mechanism is complex
because snow is a heterogeneous medium of ice crystals. Above - 1 0 C, the sliding
of the ski is lubricated by the presence of a water film produced by the partial
melting of snow crystals in contact with the bottom surface of the ski. This water
film plays the role of a lubricant ensuring a low coefficient of friction ( ~ 0,02). The
coefficient of friction represents the relationship between the force necessary to
initiate the slip and the normal force applied to the contact surface. This sliding
mechanism involving a water film is only valid above a temperature of - 1 0 C and,
in this temperature range (figure 16.17.), most materials have very similar friction
coefficient.
Below - 1 0 to - 1 5 C, the heat produced by friction is rapidly dissipated and
local melting ensuring the formation of lubricating film can no longer occur. Under
these conditions, lumps of ice adhere to the bottom surface and the friction coeffi
cient increases rapidly for wood (figure 16.17.) or metals. This is not the case for
certain polymers such as polytetrafluoroethylene (PTFE) or polyethylene (PE) for
example, which retain a low friction coefficient even when the temperature drops
below - 1 0 C. For this reason, the bottom surface of the skis is made of HDPE,
PTFE being too soft and much too expensive.

0,6 . 1 1

Waxed wood

o 1 1
' "
0 -10 -20 -30 (C)
Temperature

Figure 16.17. Variations of the coefficient of friction of waxed wood and polytetra
fluoroethylene (PTFE) as a function of the temperature of the ice (after Ashby, Jones,
1980).

The ski represents an excellent example of the use of a sandwich structure com
bining various materials. Each type of material fulfils one or more particular func
tions in an assembly stressed in a very complex way. The quality of a ski depends not
only on the choice of an optimal structure, but also on its shape. Modifications of
length, thickness, size and camber of the ski make it possible to modulate the distri
bution of its flexibility. Its behaviour in turns depends on its shape. The field of ski
manufacturing is in constant technological development.
 Composite materials 419

16.7. Exercises
16.7.1. A powder of alumina (A1 0 ) (2 wt.%) having an average diameter of 1 p m
2 3

is introduced into a molten aluminium matrix. Calculate the number of alumina par
ticles introduced per cm of the composite material.
3

16.7.2. An electrical conductor is made of a steel wire core 0.8 mm in diameter

sheathed by copper. This composite wire, having a total diameter of 2 mm, is sub
jected to a tensile test. Determine:
the metal in which occurs initially a plastic deformation;
the maximum loading supported by the composite wire before a plastic
deformation occurs in one of the two components;
the modulus of elasticity of the composite wire?
16.7.3. For the manufacture of a boat, an epoxy resin reinforced with 45 % oriented
long glass fibre is used. Calculate the modulus of elasticity perpendicular to the glass
fibres.
16.7.4. Calculate the value of the modulus of elasticity of a thin section of a com
posite made of an epoxy resin reinforced by continuous glass fibres laid parallel to
one another and perpendicularly. The volume fraction of the fibres is 0.30 distributed
50 % in each direction
16.7.5. What is the volume fraction of glass fibres which must be introduced into a
polycarbonate matrix to obtain a composite with a modulus of elasticity/density ratio
equal to 20 (GPa/(tm~ )? 3

16.7.6. The tips for cutting tools consist of hard and brittle particles of tungsten car
bide WC in a matrix of cobalt being used as binder. Explain why it is not possible to
use aluminium, which is more ductile than cobalt, to manufacture this matrix?
16.7.7. A safety glass consists of a sandwich structure composed of a polycarbonate
film of 1 mm thickness placed between two glass plates of 3 mm thickness. Calculate
the value of the modulus of elasticity of this safety glass, in parallel and perpendicu
lar to the surface of the plate.
16.7.8. Give an advantage and a disadvantage resulting from the choice of a com
posite reinforced with discontinuous fibres and oriented in a plane and of a composite
with randomly distributed discontinuous reinforcing fibres.
16.7.9. What is the difference between cement and concrete? Quote two important
cases where the use of concrete is limited. Propose two techniques for reinforcing
concrete.
16.7.10. Why in general is the use of thermosetting resins preferred to thermoplastic
polymers for the manufacture of composite materials?

16.8. References and complementary reading

J. BAUVOIS, Y. REVERCHON, Les technologies Rossignol, 1998.
G.R. BELBIN, Thermoplastic Structural Composites - A Challenging Opportunity, Proc. Instr. Mech.
Engrs., 198,47, 1984
420 Introduction to Materials Science

L. A. CARSSON and R. B. PIPES, Experimental Characterization of Advanced Composites, Prentice

Hall, Englewood Cliffs (N. J.), 1987.
. K. CHAWLA, Composite Materials, Springer, New York, 1987.
G. CHRETIEN, D. HAT AT, Initiation aux plastiques et aux composites, Tec. & Doc. Lavoisier, Paris,
1990.
D. DAVID, Les Materiaux composites, conception, caracteristiques, methodes de Contrle, Materiaux et
Techniques, no. 1-2 , 9 1984.
D. DAVID, High Performance Composites in Encycl. Chem. Techn. (Kirk-Othmer), Suppl. Volume, John
Wiley, New York, 1984, pp. 260-281.
J. B. DONNET and R. C. BANSAL, Carbon Fibers, Dekker, New York, 1990.
B. D. DUNN, Metallurgical Assessment of Spacecraft Parts, Materials and Processes, Wiley, New York,
1997.
M. EPSTEIN and R. L. SHISHOO, J. Appl Polym. Sei. 57, 751, (1995).
J. GOLE, Resines renforcees dans Initiation la chimie et a la physico-chimie macromoleculaire, Volume
2, Proprietos physiques des polymeres, Groupe Francais des Polymeres (G.F.P.), 1979, p. 268.
Y. GRENIE, Les Composites multidirectionnels dans les vehicules balistiques et spatiaux, Materiaux et
Techniques, no 1-2,33, 1984.
D. HULL, An Introduction to Composite Materials, Cambridge Solid State Series, Cambridge University
Press, Cambridge, 1981.
A. KELLY, Concise Encyclopedia of Composites Materials, Elsevier, Oxford, 1995.
A. KELLY, . H. MACMILLAN, Strong Solids, 3th. ed., Clarendon Press, Oxford, 1986.
A. KELLY and C. ZWEBEN, eds., Comprehensive Composites Materials, Elsevier Sciences, Oxford,
2000.
E.Y. KIM, S. K. AN and H. -D. KIM, J. Appl. Polym. Sei. 65, 99, (1997).
J. LAMANCHE, Renforts en fibres de verre, article A 3237-7, Collection plastiques et composites,
M. CHATAIN ed., Techniques de l'ingonieur, Paris, 1998.
P. K. MALLIK, Fiber-Reinforced Composites, Dekker, New York, 1988.
R. NASLAIN, Ceramic Matrix Composites in High Temperature Structural Materials, A. G. EVANS and
M. MC LEAN eds , Chapman & Hall, London, 1996.
L. A. PILATO, M. J. MICHO, Advanced Composite Materials, Springer, Berlin, 1994.
E. PLUEDDEMANN, Silane Coupling Agent, 2 ed., Plenum, New York, 1991.
C

B. POULAERT, Proc. Conf on Spacecraft Structures, Nov 2000, Noordwijk, The Netherlands. Proc.
Conf on Spacecraft Structures', ESA SP-386 June 1996 p. 491-499.
M.O.W. RICHARDSON (ed), Polymer Engineering Composites, Applied Science Publishers, London,
1977.
R. E. SHALIN, ed., Polymer Matrix Composites, Chapman & Hall, London, 1995.
R. TALREJA ed., Polymer Matrix Composites in Comprehensive Composite Materials, Volume 1,
A. KELLY and C. ZWEBEN eds., Pergamon, Oxford, 2000.
R. YOSOMIYA, MORIMOTO, A. NAKAJIMA, Y. IKADA, T. SUZUKI, eds., Adhesion and bonding
in composites, Dekker, New York, 1990.
D. ZENKERT, An Introduction to Sandwich Construction, Chameleon, London, 1995.
 Chapter 17.

Nanomaterials and
nanostructured materials

17.1. Objectives
To show that the properties a material change radically when one of its
critical dimensions is decreased to the nanoscale.
To distinguish between nanomaterials and nanostructured materials.
To present selected examples of different classes of nanomaterials.
To briefly introduce new experimental tools used to manipulate
individual atoms and molecules.

Nowadays, an important number of laboratories are involved in, and a huge

amounts of work is devoted to, the development of nanomaterials. Their number has
been growing explosively worldwide in recent years. They have the potential for
revolutionizing Materials Science and already have a significant commercial impact,
which is certain to increase substantially in the future.

17.2. Introduction
In a macroscopic solid, trillions of atoms are involved. By contrast, the number
of atoms and molecules included in nanomaterials is very limited. One nanometer
(one billionth of a meter) roughly contains a few atoms. At the micron scale, there
are only a few thousand atoms in one spatial direction. This means that in nanomate
rials the number of atoms involved is limited in most cases to a maximum of a few
thousand. In a monomolecular film of 10 nm , there are only several hundred atoms.
2

This means that it is possible to envisage molecular or atomic manipulations to create

new entities directly under the microscope. In the fabrication of nanomaterials, the
scientist will have to learn to handle individual atoms. This approach is now feasible
due to the introduction of new experimental tools such as the scanning electron
tunnelling microscope (STM) and the atomic force microscope (AFM).
Small devices are very important for the development of new technologies.
Miniaturisation has been the source of major technical progress: electronics, sensors,
photography, etc. All these developments result mostly from the creation of new
functional micromaterials. In particular, the spectacular development of electronics
has been fuelled by miniaturisation.
The size of a transistor has been reduced to a diameter of 180 nm. Engineers are
now able to put l(f transistors on a few c m of the surface of a silicon single crystal.
2

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

422 Introduction to Materials Science

However, the size of a transistor on a chip is around 10 larger than the dimension of
5

a contemplated molecular electronic component. If the size of a transistor is

compared to a page of this book, a molecular nanodevice would be equivalent to the
dot terminating this sentence. The present-day electronic components will be scaled
down and single-molecule transistors will most probably be introduced in the
forthcoming years.
A clear distinction should be made between nanomaterials and materials assem
bled from nanocomponents. Nanomaterials contain a controlled morphology with at
least one nanoscale dimension. Nanostructured materials, like the bones constituting
our body, are materials constituted of assembled nanoelements (nanocrystals, nanofi-
bres).
In nature, nanomolecular machines sustain life for millions of years. The natural
nanoworld is populated by biological machines such as ribosomes, which are cellular
organisms coded for producing proteins. The field of man-made nanomaterials is still
in its infancy compared with the opportunities offered by biology.

17.3. Why is the transition from micro to nano so important?

At the micro scale, most of the properties remain generally the same as for bulk
materials. On the other hand, the decrease of one or more geometric dimensions
down to the nanoscale completely modifies the material behaviour.
Transition to the nanoscale implies an enormous increase of the surface area
with respect to the volume. The mean dimension of a grain in polycrystalline mate
rials is generally of the order of several tens of microns. In this case, only one in a
thousand atoms are located at the boundary. If the size is reduced to 12 nm, 15 % of
the atoms are involved in the boundary. For a size of 5 nm, 40 % atoms will be on the
interface. Figure 17.1., which shows the percentage of atoms at the surface of hex
agonal close packed full-shell cluster, illustrates this effect.
For a low number of shells, the number of surface atoms becomes important and
their behaviour is completely different from that of the inner atoms. This situation
can account in particular for the catalytic properties of nano-crystalline particles.

Number of
shells around 1 2 3 4 5
one central atom
Number
of atoms 13 55 147 309 561
Percentage
of atoms at 92 76 63 52 45
the surface
Figure 17.1. Representation of hexagonal close packed full-shell clusters. As the number
of atoms decreases, the percentage of atoms at the surface of the cluster increases (after
Hofmann, 2001).
 Nanomaterials and nanostructured materials 423

Electrons, confined in a nanostructure having one or more of its dimensions

close to their associated wavelength (2.1.) (about ten nanometers for a semiconductor
at room temperature) take on peculiar properties. Their behaviour becomes quan
tized. In a metal, the continuous density of states in the conduction band breaks down
into discrete electronic levels. The spacing between levels increases with decreasing
size and a forbidden band gap appears. Metal particles start to lose their metallic
behaviour and tend to become semiconductors.
In semiconductors, the phenomenon is slightly different but similar since a for
bidden band gap already exists. The energy bands gradually convert into discrete
electronic levels and the forbidden band gap increase as the size of the particle de
creases.
It is possible to fabricate flat metal structures by lithographic techniques with a
thickness of 50 to 100 nm; sufficiently low to provide quantum confinement. This
leads to so-called quantum wells, which are ultra thin quasi two-dimensional struc
tures, where the atoms are confined in a slice of a few tens of nanometers deposited
for example on a S i 0 surface. In this case, the electrons are free to move inside the
2

walls of this quasi two-dimensional box with a very high mobility.

A narrow strip sliced from one of these flat structures gives a one-dimensional
quantum wire. The resistivity of a quantum wire is equal to zero. An electric current
flows in a quantum wire in a discrete manner without energy dissipation. Several
examples of quantum wires will be further reported. Cutting up a one-dimensional
quantum wire yields zero-dimensional quantum dots in which the electrons are
physically trapped.
The transition from micro to the nano-scale has a direct bearing on most physi
cal properties, such as modulus of elasticity, electrical and thermal conductivities,
magnetic properties, catalytic phenomena, etc.

17.4. Nanopowders and derived nanostructured materials

17.4.1. Introduction
Powder technology is a very important facet of modern ceramic manufacture.
Powders used in technical ceramics preparation are materials of controlled morphol
ogy and size with a diameter ranging between the micron and the millimetre. Each
particle contains several billions of atoms. Nanopowders have now been developed
for high performance ceramic applications and the preparation of hard metals. They
consist of small clusters of atoms with diameters from 5 to 25 nm. They contain a
maximum of only several thousands of atoms. For example a nanoparticle with a
3 nm diameter contains only around 1000 atoms. Nanopowders will find applications
in many areas and several products are already on the market.

17.4.2. Nanopowder synthesis

Nanoparticles can be prepared by gas-phase synthesis. Cooling of a saturated
dense gas under specific conditions induces nucleation and growth of nanoparticles
as raindrops appear in a cloud. Different synthesis techniques are involved, such as
424 Introduction to Materials Science

inert gas condensation for the production of metal nanopowders, or reactive vapour
condensation for the preparation of ceramic oxide nanopowders.
Diverse nanoproducts such as alumina ( A 1 0 ) andzirconia ( Z r 0 ) h a v e been
2 3 2

prepared by pyrolysis of a precursor, which is an organic (polymer) substance or an

inorganic salt.
Another procedure for producing nanoparticles is by controlled hydrolysis of a
chloride and precipitation from an aqueous solution followed by calcination at
T< 600 C. This technique, which produces very small particles (< 10 nm) forming
larger agglomerates is used, for example, for anatase ( T i 0 ) nanopowder production.
2

17.4.3. Some applications and properties

Nanopowders are finding applications in many areas.
Ceramics made with nanosized powders are more ductile at elevated
temperature compared to coarse-grained ceramics and can be sintered at
lower temperature. For example, nanophased titanium oxide obtained
by compaction at 800 C of grains smaller than 30 nm shows plastic de
formation up to 60 % without breaking. In this case, it seems that nano
phased particles are better able to slide one over the other than coarse
particles.

Nanosized iron and copper powders have hardness about 4-6 times
higher than observed for the bulk material ( 12.4.8). This increase in
hardness is attributed to the absence of dislocations in the nanograms.
Nanophased metals made by sintering of nanoparticles are much harder
than coarse-grained metals. For example, nanostructured copper ob
tained from nanograms of 6 nm size is 5 times harder that conventional
copper

Particles smaller than 50 nm are much too small to scatter light

(400 nm < < 700 nm) and ceramics sintered with nanopowders be
come transparent.
Nanosized hard magnetic particles form magnetic mono-domains
which increases the coercivity compared to large particles in which
energetic considerations favours the formation several magnetic do
mains separated by Bloch walls ( 14.4.4.).
Nanosized magnetic particles based on rare-earth borides and nitrides
such as N d F e i B or S m F e i ( N C ) with single magnetic domains of
2 4 2 7 3

about 100 nm are used for high density information storage and for
mechanical force transfer in ferrofluids.
The colour of nanosized semiconductor pigments ( 14.5.7.) may vary
as a function of their granularity since the width of the forbidden band
is increased in semiconductor nanoparticles. Cadmium selenide (CdSe)
is a red powder used as pigment. In nanosized cadmium selenide, the
colour can turn to yellow.
Ultraviolet-absorbing nanoparticles such as T i 0 and ZnO are used in
2

sunburn lotion.
Nanosized Cu and Ag are used in conducting inks and polymers.
 Nanomaterials and nanostructured materials 425

17.5. Carbon nanostructures

17.5.1. Identification of novel forms of carbon

Carbon can exist in a great variety of allotropic forms. The case of graphite and
diamond has already been discussed in chapter 2. More recently novel varieties of
carbon have been discovered: fullerenes, nanotubes, etc. which are interesting mate
rials because they open new fields of application. Their structures are illustrated in
figure 17.2.
Fullerene molecules are composed of a specific number of carbon atoms. 60 in
the case of the C o which has a diameter of 0.7 nm. Its structure, similar to that of a
6

football, consists of 20 hexagons and 12 pentagons. Pentagons are necessary to close

the hexagonal network. The carbon-carbon sp is predominant with 10-15 % sp
1 3

hybridisation. Its bond length (0.14 nm) is equal to that of the hexagonal graphite
sheet.
The carbon nanotubes are formed of rolled graphitic planes. The simplest struc
ture of carbon nanotubes (figures 17.2. and 17.3. (a)) consists of one single cylindri-
cally rolled layer {single walled nanotubes - SWNT). Multiwalled nanotubes
(MWNT) composed of several cylindrical concentric layers are also found (figures
17.3. (b) and 17.4.). A nanotube is normally closed at each end by the introduction of
pentagons as in fullerene molecules.

Fulierene ( C )
6 0

Nanotube

Figure 172. Schematic structure of fullerene Coo and carbon nanotube.

426 Introduction to Materials Science

(a) (b)

Figure 17.3. Single (a) and multilayered (b) carbon nanotubes under high magnification
(from J. C. Charlier, 2001).

0.1 5 nm
(a) (b)

Figure 17.4. Multilayered carbon nanotubes under two magnifications. In figure (b) ta
ken under higher enlargement, one observes the cylindrical nanostructure of multilayered
nanotubes (about twenty concentric layers). One observes also on the micrograph (b)
occurrence of amorphous carbon deposits. These can be eliminated by oxidation in the
air at 500 C (from J. B. Nagy, 1999).

17.5.2. Fullerene and carbon nanotube synthesis

Fullerenes were first produced by a direct current discharge between graphite
electrodes immersed in an inert gas (helium or argon) (figure 17.5. (a)). Graphite is
vaporised at the high temperature developed by the electric arc. Fullerenes, mostly
C oand C o, condense on the reactor walls. Multi-walled nanotubes are formed si
6 7

multaneously but they appear only on the cathode where the current flows. Single-
walled nanotubes are also achieved by the same procedure but it is necessary in this
case to use a carbon anode containing a small amount of a transition metal catalyst.
 Nanomaterials and nanostructured materials 427

The preparation of single-walled nanotubes (figure 17.5 (b)) was also achieved by
laser ablation of a carbon target containing 0.5 atomic percent of nickel and cobalt
and placed in a tube furnace heated to 1200 C. As a general rule, single-walled
nanotubes never grow in the absence of transition metal catalysts.
Chemical vapour deposition is also used to prepare nanotubes (figure 17.5. (c)).
This process, which is similar to heterogeneous supported polymerisation reactions,
proceeds by decomposition of acetylene, ethylene and propylene at a temperature of
650-800 C in the presence of transition metals catalysts. This cracking reaction
produces hydrogen and C and C3 intermediary reagents. Nanotubes of various di
2

ameters, length and structure can be obtained by this method.

The separation of the nanotubes from the catalyst particles is carried out by a
hydrofluoric acid treatment. The crude nanotubes are freed from amorphous carbon
by a permanganate or air oxidation.

17.5.3. Properties and applications

Properties and applications of carbon nanotubes remain still largely to be dis
covered. Their very regular structure endows them with high mechanical properties
while they remain very flexible. While the elementary filament of carbon fibres has a
diameter comprised between 6 and 10 , the multiwalled carbon nanotubes
(MWNT) have a diameter of the order of 20 nm for a length exceeding several mi
crons. A carbon nanotube is, in a sense, the ultimate carbon fibre having a much
more regular structure and in some cases being practically free from defects.
The carbon nanotube can be either semiconducting or metallic depending on
their diameter and on the orientation of the hexagons with respect to the nanotube
axis.
Single-walled carbon nanotubes of one atom in thickness and around ten atoms
of diameter can be considered as a prototype one-dimensional (ID) quantum wire.
Their electronic properties are of great interest. Indeed, in this defect-free
nanostructure, the electron propagates in a ballistic way without scattering by the
atoms. Therefore no heat dissipation is observed in the nanotube arising from the
transport of electrical current. Their conductance is quantized and independent of the
tube length.
Carbon nanotubes are very flexible. They can be bent through 90 and recover
0

their initial shape when relaxed. Their Young's modulus varies with diameter. For a
nanotube with a diameter smaller than 8 nm, the modulus is equal to 1200 GPa,
slightly higher than for diamond and ~6 times the value for carbon steel. A Young's
modulus as high as 5000 Gpa is predicted. For tubes with a diameter larger than 30
nm, it drops to 200 GPa. The thermal conductivity at 300 of single-walled carbon
nanotubes should be very high and estimated to be of the order of the best thermal
conductors such as pyrolysed graphite (-2000 W m K in a direction parallel to the
_I _1

hexagonal plane). This thermal conductivity is superior to that of the best thermal
conducting metals (Ag: 430 Wm K~ - Cu: 450 Wm^K" ) at ambient temperature.
-1 l 1

Applications are envisaged for carbon nanotubes in energy storage. The utilisa
tion of multi-walled nanotubes intercalated with lithium has been investigated for
rechargeable lithium batteries.
The present challenge is now to make nanotubes of controlled structure, in suffi
cient quantity and at competitive prices to consider large scale applications.
428 Introduction to Materials Science

(a)

laser ablation

oven temperature ~ 1200 C

(b)

CVD

CHN M

oven temperature 500-1000 C

Figure 17.5. Schematic experimental setups for nanotube preparation methods (from
Hongjie in Dresselhaus et al, 2000)

17.6. Template method for the preparation of nanomaterials

17.6.1. Template membranes used
This technique involves the fabrication of nanomaterials within the nanopores of
a porous membrane used as a template. Several porous structures can in principle be
used as templates but in practice two main types of nanoporous membranes have
been utilised: 'track-etch' polymer membranes and porous alumina membranes.
Micro and nano-porous membranes made by the 'track-etch' method are available
commercially from Whatman Company. They have various applications in e.g. sen
sors, virus detection or removal, ultra pure water preparation, etc. Nanoporous track-
etched membranes are made in various pore sizes in the range 15-100 nm.
The preparation method of track-etch membranes is in its principle very easy
(figure 17.6.). It entails irradiation of a thin film (thickness ranging from 5 to 20
usually) of bisphenol-A polycarbonate or semi-aromatic polyester (poly(ethylene
terephthalate)) by accelerated particles such as A r ions produced in a cyclotron
9+

followed by irradiation with UV light. The irradiation produces damage tracks inside
the film. Pores of uniform diameter are then obtained by chemical etching of the
irradiated tracks by aqueous sodium hydroxide at moderate temperature (-70 C) for
various periods as a function of the pore diameter required. Depending on the in
tensity and the duration of the irradiation, pore densities between 10 and 10 cm" are
5 9 2

obtained without major difficulties.

 Nanomaterials and nanostructured materials 429

(a) (b)

Figure 17.6. Principle of fabrication of'track-etch' nanoporous membranes.

Alumina porous membranes can also be obtained by anodic oxidation of alu

minium in acidic solution. These membranes also contain cylindrical pores of uni
form diameter. However, in this case, the pores are arranged in hexagonal arrays and
are perpendicular to the membrane surface unlike track-etch membranes which con
tain slightly tilted pores randomly distributed due of the random nature of the irra
diation process. In this case, a number of pores intersect within the membrane.
However, recent developments have tackled this problem and nanoporous
membranes with parallel pores perpendicular to the surface and with perfectly cylin
drical walls were produced at the laboratory scale. These optimised membranes con
stitute excellent templates for the preparation of nanowires and nanotubes.

17.6.2. Examples of nanomaterial synthesis by the template method

Examples of electronically conductive conjugated polymers are presented in
figure 17.7. Polypyrrole, poly(3-methylthiophene) and polyaniline nanotubes were
synthesized by the template procedure by chemical or electrochemical oxidative
polymerisation.

Polyaniline

Figure 17.7 Examples of electronically conductive polymers synthesized by the template

method.
430 Introduction to Materials Science

Enlargement of a nanopore
Pyrrole monomer Initiator

Polypyrrole
polymerised
nanotubule

Nanoporous membrane

Figure 17.8. Two-compartment reactor used in the polymerisation of pyrrole (from De-
moustier-Champagne et al, 1998).

(a)
20 jim

(b) 2

17.9. Polypyrrole nanotubules made by template polymerisation in 'track-etch'

nanoporous polycarbonate membranes, (a) Overall view, (b) enlargement showing the
hollow tubular structure (courtesy of Demoustier-Champagne).
 Nanomaterials and nanostructured materials 431

As an example, the oxidative polymerisation of pyrrole in nanotubes will be

briefly described. As shown in figure 17.8., a polycarbonate membrane is used as a
dividing wall in a two-compartment reactor cell. In the left compartment, an aqueous
pyrrole monomer solution is introduced and is allowed to diffuse through the mem
brane prior to the introduction of an aqueous oxidant initiator. Polymerisation takes
place into the membrane nanopores and, depending on the polymerisation time (1 to
16 h at ambient temperature) a more or less thick hollow polypyrrole tubule or a
filled solid nanowire is formed. Typical polypyrrole hollow nanotubules are shown in
figure 17.9.
The elastic tensile modulus of polypyrrole nanotubules was determined by
atomic force microscopy. For nanotubes having an external diameter larger than
100 nm, the elastic modulus varies between 1.2 and 3.2 GPa. These values are com
parable to those obtained for polypyrrole films. However, the nanotubule modulus
increases considerably when the diameter decreases. Values as high as 60 GPa were
obtained for an external diameter of 35 nm. This observation shows that the proper
ties are drastically modified by the confinement of the material. Similar results are
observed for the electrical conductivity. Values more than one order of magnitude
larger than the bulk material were obtained for diameters of around 30 nm.
Metals (Ag, Au, Cu, Ni, etc.) nanowires were also obtained by electrodeposition
using a template technique. Electrochemical deposition is carried out using mem
branes coated on one face with a metal film as a cathode for electroplating. Quantum
nanowires of extremely small diameters (30 nm) were prepared by this method.

17.7. Nanomaterials by biological mechanisms

In Nature, many systems are directly connected to nanotechnology. Biological
processes are at the basis of very important polymer syntheses. Among them, cellu
lose (figure 17.10.) has very important application in various fields: papers, textiles,
plastics, fibres, etc. Vascular plants are the major industrial source of cellulose. Most
paper products originate from wood pulp. Cotton is the major resource for textiles.
However, cellulose harvesting in the world's great forests leads to important losses in
the ecosystem and cellulose fibre extraction from wood is very polluting.
Although, it has been known for over 100 years that acetobacter produces extra
cellular cellulose, this process has not been developed until very recently, although it
could be considered as an alternative source for this material. Bacterial nano-cellu-
lose has unique features which distinguish this product from vascular plant cellulose.
In particular, it is obtained in a very pure form without lignin, hemicelluloses and
other by-products. Lignin constitutes the polymer matrix embedding the cellulose
fibres synthesised by trees and plants. These characteristics make bacterial cellulose
very attractive especially for countries like Japan which do not possess important
natural resources in this field. However, cellulose from trees and plants is a very
inexpensive material and bacterial cellulose has to compete with cheap cellulose
products.
The most widely used source of bacterial nano-cellulose is acetobacter. The
bacterial cellulose is produced in both static and stirred aerobic cultures. The medium
is mostly composed of dilute aqueous sugar (glucose, etc.) solution. The current yield
is around 20 g of cellulose for 100 g of glucose. Fermentation normally takes about 30
432 Introduction to Materials Science

CH OH2

( ),'
5 10 5

= 1500 to>6000

Figure 17.10. Cellulose structure.

hours. Next, the bacterial cells are destroyed using a hot caustic treatment and the
remaining cellulose goes through a series of washing and dewatering steps. The
bacterial cellulose fibres made in a stirred medium are recovered in the form a gelati
nous mass.
Bacterial cellulose is produced in the form of highly crystalline (70 %) nano-
fibrils. A typical bacterial nano-fibril ribbon (figure 17.11.) is around 60 nm in width
with an indefinite length. Softwood pulp fibres are generally at least 100 wide.
Owing to fibril thickness, bacterial cellulose has 200 times more specific surface
area than soft wood pulp fibres. This high surface area combined with hydrogen
bonding capabilities is responsible for the exceptional thickening properties of
bacterial cellulose fibres. Bacterial cellulose retains up to 700 % water.
However, the price of the bacterial cellulose remains too high to be considered
as a replacement for the traditional woody cellulose. The productivity of large
fermentors remains low and is complicated by a number of phenomena associated
with the biology of acetobacter. Bacterial nano-cellulose is preferably recommended
for value-added niche markets.

Elementary fibril formation (40-60 nm)

1.5 nm sub-elementary fibril

3.5 nm pore / ^ ^ (

Acetobacter xylinum

1 - 2

Figure 17.11. Schematic representation of nanofibrils of bacterial cellulose produced on

an acetobacter xylinum.

In a static process, the long cellulose ribbons produced by the bacterial cells
interweave to form a very thin and very strong hydrophilic cellulose membrane at the
surface of the fermentor. 1 g of bacterial cellulose may cover a surface of 4000 cm . 2

After a cleaning process and pressure treatment, this cellulose membrane has a very
high shape retention in various environments and a tensile strength five times higher
than a polyethylene or PVC film. Sony Corporation in Japan has developed a high
fidelity audio speaker diaphragm made from this material.
 Nanomaterials and nanostructured materials 433

Bacterial cellulose produced in stirred reactors can also be used as high per
formance filtration membranes by the food industry. Applications are envisaged
either as an additive for paper coating to improve smoothness and printing
characteristics, or as a raw material to prepare very strong paper. Bacterial cellulose
is a non-nutritive non-toxic substance that can also be used as an additive in food as a
thickener or as a texturizing agent. Due to its extensive surface area and its high hy
drophilic character, it is by far more interesting for this kind of application than
wood-plant cellulose.

17.8. Observation and manipulation of atoms and molecules

The direct observation of atoms and molecules has been made possible by the
development of new microscopic techniques: scanning tunnelling microscope (STM)
and atomic-force microscope (AFM).
The electron-tunnelling microscope is a direct application of the well-known
tunnelling effect. When a sharp tungsten tip positively charged (anode) is brought
(figure 17.12) at distance Ar slightly shorter than a nanometer, it produces an intense
electric field and electrons are able to leave the cathode with an energy less than the
potential barrier separating the anode and the cathode through a tunnelling process. A
tiny electric (tunnelling) current can be measured. This tunnelling effect can be ex
plained by quantum mechanics. Only discrete electrons can tunnel between the anode
and the cathode. The tunnelling current falls off very steeply with increasing distance
Ar from the surface. This effect is the basis of the scanning tunnelling microscope,
which allows the study of the topography of a conductor or semiconductor surface
with an atomic resolution.

Solid conductor or semiconductor surface

(cathode)
Q Positive ion -Electron

Figure 17.12. Principle of the scanning tunnelling microscope (STM).

This technique allows an atom or a group of atoms to be picked from a surface

and deposited onto another spot. This is illustrated on figure 17.13. where the motion
of a CO molecule adsorbed on a copper plate is represented. The CO molecule is al
ways attached to copper by the carbon atom.
For a non-conductive substance, atomic force microscope (AFM) is used (figure
17.14). The deviation of a sharp tip on the sample atomic surface is followed by the
deflection of a cantilever. Extremely sharp tips with a curvature radius from 3 to
50 nm are available for this technique.
434 Introduction to Materials Science

oooooooooooooo
Cu surface

Figure 17.13. Transfer of a CO molecule on a copper substrate.

Cantilever probe

(b)

Figure 17.14. Atomic-force microscope (AFM) used with non- conductive materials, (a)
Principle: the deviation of a sharp tip on the sample surface is followed by the deflection
of a cantilever, (b) Enlargement of a tip.

Different modes of operation are used. For strong repulsive force (contact
mode), the tip may be dragged across the surface maintaining a constant force or a
constant distance as it scans. For weak interactions, a non-contact mode (tapping-
mode) is used; the tip is vibrating up and down while scanning the surface under
investigation. This technique enables operation on a scale from microns to nanome
ters. In AFM, the interaction forces between the tip and the sample decrease less
steeply than for STM. An AFM probe responds more to an average force induced by
neighbouring atoms than by individual atoms. The AFM response depends greatly on
the sharpness of the tip.
 Nanomaterials and nanostructured materials 435

AFM allows the possibility of measuring very weak forces (10~ -10~ N) and is
18 16

used to determine the mechanical properties of nano objects. The Young's modulus
of nanotubes have been measured by this technique. Atomic and molecular
manipulations can be used to create new combinations of atoms. They have reached
sufficient reproducibility to enable experimentation directly on a surface. They
constitute a basic tool for nanoscience.

17.9. Summary and conclusions

In less than one century of Materials Science development, we have experienced
a transition from the 'drop hammer to jewellery'. Nowadays, nanomaterials consti
tute the ultimate step of this revolution. Their development is extremely important
because not only do they allow the preparation of nano-assembled materials but they
also pave the way to building nanomachines and nanorobots, which are particularly
interesting for the medical applications of the future.
In this chapter, selected examples of nanomaterials from various classes of ma
terials have been introduced. New properties appear simultaneously with phase
miniaturisation due to the great importance of the specific surface, atom confinement
and quantum effects. The critical dimension for the appearance of new properties is
around 30 nm. It is only below this dimension that substantial changes in material
behaviour take place: increase of the hardness, modification of electrical and thermal
conductivities, colour changes, etc.

17.10. Illustrative example: molecular electronic devices

Molecular devices might someday replace transistors, diodes, conductors and the
various components of integrated circuits. They will be very much smaller than the
smallest transistor currently envisaged and can be produced in a very reproducible
way. All the accumulated knowledge of the organic chemist already acquired in
drugs preparation, colorants and new macromolecular materials could be successfully
applied to this development.
Polynuclear orbitals, in which the -electrons are free to circulate, exist in many
aromatic compounds like benzene (figure 2.15.). It has been proved recently that a
simple aromatic derivative like p-benzenedithiol is able to transport electricity. The
sulphur atom of the thiol group has a great affinity for gold atoms and it is easily
complexed on a gold surface. A /?-benzenedithiol molecule (figure 17.15.) was fas
tened between two gold tips of a modified STM and a potential difference of 5 V was
applied. It was observed that, under these conditions, a weak electrical current of
0.2 was flowing through the molecule. At the molecular level, the density of this
electrical current is enormous considering that the electrons are passing through the
molecule one at a time. This means that one after another 1 0 electrons per second
12

traverse the molecule. What is extraordinary is that all these electrons are flowing
through the /-benzenedithiol molecule passing through the 7i-molecular orbital
avenues without energy dissipation. It is the fact that electrons are not scattered
while crossing the molecule. This means that there is no interference between the flo-
436 Introduction to Materials Science

Gold sensor tip

Gold sensor tip

Figure 17.15. Measuring electrical conductivity using aromatic molecules, (a) Chemical
structure ofp-benzenedithiol. (b) Measuring the electrical conductivity.

wing electrons and the p-benzenedithiol molecule i.e. the electrical resistance of this
organic molecule is equal to zero. An electric current flows ballistically through this
quantum nanowire in a discrete manner without energy dissipation.
In the case of the /?-benzenedithiol molecule, the electric current is carried over a
distance slightly smaller than one nm. However, in conductive polymer chains, -
electrons are free to circulate over much larger distance (several tens of run) and it is
envisaged to use this type of molecule to connect various molecular devices. The use
of carbon nanotube connectors is also contemplated for molecular-scale circuitry.
Various substituents can be introduced into aromatic polynuclear molecules to
direct the electron flux and to adjust their distribution. As shown on figure 17.16. with
three benzene rings, it is possible to create a molecular switch. The central benzene
ring of this aromatic molecule incorporates an electron-accepting nitro group ( N 0 ) 2

and an electron-donating amine group (NH ). As these two substituents are very
2

bulky, the central benzene ring is not completely coplanar with the other two and a

N X

(a) (b)

Figure 17.16. Molecular electric switch built with three aromatic rings, (a) Chemical
structure with indication of the electron transfer, (b) Spatial representation of the switch
molecule. The three aromatic rings are represented in light grey. The hydrogen atoms on
the benzene rings are not shown.
 Nanomaterials and nanostructured materials 437

polynuclear orbital cannot be formed in this condition. The circulation of the electric
current is therefore forbidden. However, this three-nuclear molecule contains a very
strong electric dipole, which can be oriented by an external electric field. Under these
conditions, the asymmetric molecule gets twisted, a partial coplanarity of the three
aromatic rings is established and electric current is free to circulate. When the
electric field is removed, the initial configuration is restored and the electron
circulation is interrupted. The process can be repeated indefinitely at a very high
commutation rate.
By combining in the same molecule, an electron-rich donor and an electron poor
acceptor, it is possible to imagine aromatic molecules in which electrons would flow
in only one direction.
Recently the C o-OPV hybrid constituted of a fullerene C o molecule bearing a
6 6

short oligophenylenevinylene (OPV) moiety has been prepared. Q - O P V films can 0

be used to create a new type of photovoltaic cell. The thin C o-OPV film was sand
6

wiched (figure 17.17.) between an indium-tin layer (ITO) and a aluminium layer
(Al). Under illumination, an electron is promoted into an empty molecular orbital of
the COO - OPV hybrid and a strong electron transfer occurs from the OPV electron-
donor substituent to the electron-acceptor fullerene and a photocurrent is observed.
This last example demonstrates the great potentiality of molecular devices to
replace present microelectronic components. However, substantial developments
have yet to be achieved in order to elaborate new electronic components from mo
lecular devices.

Glass substrate I

HO
iThinfilm ofC -3PV
1
60
Al

Wire to ITO Wire to Al

(a)

(b)

Figure 17.17. Schematic representation of a molecular photovoltaic device, (a) Sche

matic structure, (b) Functioning of the solar cell (after Nierengarten et al, 2001).
438 Introduction to Materials Science

17.11. References and complementary reading

N. ARMROLI, F. BARIGELLETTI, P. CERONI, J. F. ECKERT, J.F. NICOUD, F. NIERENGARTEN,
Chem. Commun. 2000, 599.
R. M. BROWN, Jr., CH. HAIGLER, J. SUTTIE, A. R. WHITE, E. ROBERTS, C. SMITH, . ITOH,
K.COOPER, J. Appl. Polym. Sei, Applied Polymer Symposium, 37, 33, (1983).
S. CUENOT, S. DEMOUSTIER-CHAMPAGNE, B. NYSTEN, Phys. Rev. Let. 85, 1690, (2000).
M.S. DRESSELHAUS, G. DRESSELHAUS and Ph. AVOURIS eds., Carbon Nanotubes, Springer,
Berlin, 2000.
J. DUCHET, R. LEGRAS, S. DEMOUSTIER-CHAMPAGNE, Synthetic Metals, 98, 113, (1998).
A. FERT and L. PIRAUX, J. Magn. Magn. Mater.,200, 338, (1999).
M. FREEMANTLE, Chem. & Eng. News, 77, 13, (1999).
S. GAUTHIER, Appl. Surf. Sei. 164, 84,2000.
. HOFMANN, Nano-Materials, Summer School EPFL, DMX-MPI, 2001.
J. C. HULTEEN, C. H. MARTIN, J. Mater. Chem. 7, 1075, (1997).
A. HUCZKO, Appl. Phys. A 70, 365, (2000).
C. JEROME, S. DEMOUSTIER-CHAMPAGNE, R. LEGRAS, R. JEROME, Chem. Eur. J. 6, 3089, 2000.
C. H. MARTIN, Science 266,1961, (1994), Chem. Mater. 8, 1739, (1996).
J. B.NAGY, Personal communication, (1999).
F. NIERENGARTEN, J. F. ECKERT, J.F. NICOUD, L. OUALI, V. KRASNIKOV, G. HADZIIOAN-
NOU, Chem. Commun. 1999, 617.
J. F. NIERENGARTEN, G. HADZIJOANNOU, Materials To-Day, 4, (1), 2001, p.16.
L. PIRAUX, S. DUBOIS, J. L. DUVAIL, A. RADULESCU, S. DEMOUSTIER-CHAMPAGNE E. FE-
RAIN, R. LEGRAS, J. Mater. Res., 14, 3042, (1999).
M. REED and J. M. TOUR, Pour La Science, August 2000, p. 78.
. H. RIEDER, G. MEYER, L. BARTELS, J. REPP, S. W. HLA, Preliminary Proceeding of STM, p. 208,
19-23 July, Seoul, 1999.
. TANAKA, . YAMABE, . FUKUI eds., The Sciences of Carbon Nanotubes, Elsevier, Amsterdam,
1999.
 C h a p t e r 18.

Appendixes

18.1. Density and mechanical properties of the principal

materials
Table 18.1. gives the approximate values of the density and the following
mechanical properties of the principal materials: modulus of elasticity (or Young's
modulus), offset yield strength ^0,2, ultimate tensile strength R m rupture elongation
e , fracture toughness K (K =
R c c (EG ) ). c
m

Table 18.1. Density and mechanical properties of the principal materials.

Material ^0,2
[tm" ]3
[GPa] [MPa] [MPa] [%] [MPam ' ]
0 5

Metals
2.7 70 40 200 50 100-350
Aluminium (Al) 8.9 125 60 400 55 100-350
Copper (Cu) 7.9 200 50 200 30 150
Iron (Fe) 8.9 215 70 400 65 100-350
Nickel (Ni) 19.3 410 1000 1510 1-50 20
Tungsten (W)
Metal alloys
Carbon steels 7.8 200 220 400 18-25 150
Carbon steels (hardened, tempered) 7.8 200 250-1300 500-800 20-30 50-150
Steel alloys 7.9-8.3 195-215 500-2000 650-2400 2-30 15-150
Stainless steels (ferritic) 7.5-7.7 190-200 240-400 500-800 15-25 50-120
Aluminium alloys 2.6-2.9 70-80 100-620 300-700 5-30 25-45
Copper alloys 7.5-9.0 120-150 60-900 250-1000 1-55 10-100
Cobalt and alloys 8.1-9.1 200-250 180-2000 500-2500 1-60 20-50
Grey pig iron 6.9-7.8 90-150 220-600 400-1000 0-15 5-20
Alloys of nickel 7.8-9.2 130-230 200-1600 400-2000 1-60 50-250
Magnesium and alloys 1.8 40-45 80-300 125-350 6-20 -
Titanium and alloys 4.3-5.1 80-130 200-1300 300-1400 6-30 50-140
Polymers
Polyethylene low density (LDPE) 0.90 0.2 6-20 20 _
1-5
Polyethylene high density (HDPE) 0.95 0.7 20-30 35 _
2-5
Polypropylene (PP) 0.90 0.9 20-35 35 _
3
Polystyrene (PS) 1.0-1-1 3-3.4 35-70 40-70 _
2
Poly (methylmethacrylate) (PMMA) 1.2 3.4 60-110 110 - 1
{continued)

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

440 Introduction to Materials Science

Table 18.1. (Continuation)

Material Ro.2 R Kc

[tm" ]
3
[GPa] [MPa] [MPa] [%] [Mpam ] 05

Polytetrafluorethylene (PTFE) 2.3 0.3-0.8 10-15 20-40 - 3

Polyvinyl chloride) (PVC) 1.3-16 0.01 45-50 - - 2-8
Polyamide 6 (Nylon 6) (PA6) 1.1-1.2 2-4 50-80 100 - 3
Poly(biphenol A carbonate) (PC) 1.2 2.6 50 60 - 1-2.5
Epoxy resin (EP) 1.1-1.4 3 30-100 30-120 - 0.5
Acrylonitrile-butadiene-styrene
(ABS) 1.2 - 55 60 500 4
Elastomers 0.8-0.9 0.01-0.1 - 30 - -
Expanded polymers (foams) 0.01-0.6 0.001-0.1 0.2-10 0.2-10 10-100 -
Ceramics
Alumina (Al 0 ) 2 3
3.9 390 5000 - - 3.5
Magnesia (MgO) 3.5 250 3000 - - 3
Silica (Si0 )
2
2.6 95 7200 - - 1
Silicon carbide (SiC) 2.5-3.2 450 10000 - - 3
Silicon nitride ( S i N ) 3 4
3.2 350 8000 - - 4-5
Normal glass 2.5 70 3600 - - 0.7
Tungsten carbide (WC) 14-17 450-650 6000 - - -
Zirconia (Zr0 ) 2
7.5 145 4000 - - 5-10
Composites
Concrete (compression) 2.5 45-50 20-30 - - 0.2
Concrete (reinforced) - - - 410 2 10-15
Wood// 0.4-0.8 10-15 - 35-55 - 0.5-1
Wood! 0.4-0.8 0.5-1.0 - 4-10 - 10-12
Co-WC (Cermets) 11-12.5 400-530 400-900 900 2 15
EP + boron fibres 2.0 125 - 700-1700 - 45
EP + glass fibres 1.8 35-45 - 100-300 - 40-60
Polymers + carbon fibres 1.5 70-200 - 650 - 30-45

18.2. Variation of atomic radii for the various types of

chemical bonds
Table 18.2. Tests of validity of the concept of ionic radius (after Guinier, 1980).

Alkali metal Distance between + and - Computed values starting

halide ions, measured in the from the ionic radii in
crystal [nm] table 18.3 [nm]

LiF 0.202 0.205

Lil 0.300 0.285
NaF 0.232 0.235
NaCl 0.281 0.281
Nal 0.324 0.315
KF 0.267 0.268
KCl 0.314 0.314
KI 0.353 0.348
RbF 0.282 0.285
Rbl 0.367 0.365

(after Guinier, 1980)

 Appendixes 441

Table 18.3. Average radius of the ions in ionic crystals (nm).

Ionic charge 2- 1- Rare gas 1+ 2+ 3+ 4+

(He) Li Be
Z=2 0.07 0.035
O F (Ne) Na Mg Al Si
0.140 0.135 =10 0.10 0.07 0.05 0.04
S
0.18
CI
0.181
Ar
Z= 18

0.133
Ca
0.10
Sc
0.08
Ti
0.065
Br (Kr) Rb Sr Y Zr
0.195 0.195 Z = 36 0.15 0.11 0.10 0.08
Te I (Xe) Cs Ba La Ce
0.22 0.215 = 54 0.17 0.135 0.11 0.095

(after Guinier, 1980)

Table 18.4. Atomic radii of metals.

Element Atomic; Element Atomic Element Atomic

radius radius radius
(nm) (nm) (Nm)

Li 3 0.152 Ca 20 0.197 Cs 55 0.266

Na 11 0.186 Cr 24 0.125 Ba 56 0.217
Mg 12 0.160 Fe 26 0.124 With 79 0.144
Al 13 0.143 Cu 29 0.128 Pb 82 0.175
19 0.231 Ag 47 0.144 U 92 0.150

(after Guinier, 1980)

Table 18.5. Van der Waals radii of some atoms (after Guinier, 1980).

Element Van der Waals Element Van der Waals

radii (Nm) radii (Nm)

0.117 I 0.210
C 0.180 Ne 0.160
0.157 Ar 0.192
0 0.136 Kr 0.201
CI 0.178 Xe 0.221

(after Guinier, 1980)

18.3. Principal polymer materials

Table 18.6. Formulae and nomenclature of principal organic polymers.

Formula of constituent block Present name and abbreviation

C H CH 2 Polyethylene (PE)
CH CH
2 Polypropylene (PP)
CH 3

CH 2 C H = CH C H 2
Polybutadiene (PB)

{continued)
442 Introduction to Materials Science

Table 18.6. (continuation)

Formula of constituent block Present name and abbreviation

fH 3

Polyisoprene (PP)
- C H C = CH CH
2 2

Polyvinyl chloride) (PVC)

CH -CHC1 2

Poly(vinylidene chloride) (PVDC)

CH - CC1- 2

Polyvinyl alcohol) (PVA)

-CH (pH 2

OH

-CH 2 C Poly (aery lie acid) (PAA)
COOH

CHo-^H Polyvinyl acetate) (PVAC)

O-COCH 3

OCH 3

- C H 2 - C - Poly (methyl methacrylate) (PMMA)

CH 3

CF CF - 2 2 Poly(tetrafluoroethylene) (PTFE)

CH CHF 2
Polyvinyl fluoride) (PVF)

-CH -CF 2 2
Polyvinylidene fluoride) (PVDF)

C H , - C1H Poly(acrylonitrile) (PAN)

CN

-(CH )-C-NH- 2 Polyamide 6 (PA6 or Nylon 6)

O

CH 3

Poly(2.6-dimethyl-1.4-phenyleneoxyde)
(PPO or PPE)
CH 3

- N H - (CH ) - NH C - (CH ) - C-
2 6 2 4 Polyamide 6-6 (PA 6-6 or Nylon 6-6)

-(CH ) 2 1 0 -C-NH Polyamide 11 (PA 11 or Nylon 11)

(continued)
 Appendixes 443

Table 18.6. (Continuation)

Formula of constituent block Present name and abbreviation

CO-^ CO C H CH
2 2 Poly (ethylene terephthalate)
(PET)

- C Q
" ^ =
^ C O O - (CH ) O
2 4
Poly(butylene terephthalate)
(PBT)

Poly(bisphenol-A carbonate) (PC)

CH 3

Poly(p-phenylene-ether-ether-ketone) (PEEK)

CH.
I 3

OSi Poly(dimethyl siloxane)(PDMS)

I
CH 3

9

Poly(diphenyl siloxane) (PDPS)
Si

Poly(phenylene sulphide) (PPS)

Principal copolymers Ethylene-vinyl acetate (EVA)

Styrene-butadiene (SBR)
Acrylonitrile-butadiene (NBR)
Styrene-acrylonitrile (SAN)
Acrylonitrile-butadiene-styrene (ABS)

Dense three-dimensional networks (thermosets) Phenol-formaldehyde resins (PF)

Urea-formaldehyde resins (UF)
Melamine-formaldehyde resins (MF)
Epoxy resins (EP)
Polyurethanes (PU)[a]
Unsaturated polyesters (UP)

[a] There are thermoplastic (linear) and thermoset (cross-linked) polyurethanes.

444 Introduction to Materials Science

18.4. Classification of Principal Ceramics

Table 18.7.

Classes Examples Standard uses

Terra cotta Terracotta Bricks, tiles, pipes

Traditional ceramics Porcelains Crockery

Earthenware Medical
Electric insulation

Abrasives Alumina Abrasive discs

Silicon carbide Cutting tools
Diamond Trepans of drilling

Glass and new ceramics Window glasses Containers

Pyrex Transparent partitions
Vitroceramics Heat-resistant dishes
Enamels

Refractories Alumina Coatings of cupola furnace

Silica Moulds for metal castings
Magnesia Furnace linings
Zircon
Graphite

Technical ceramics Carbides, Nitrides Moulds, bearings

Iron oxides and mixed oxides Components for electronic
(ferrite) engines, sensors
Ceramic metals (cermet) Matrices

Cement Portland cement Construction

Nuclear ceramics Oxides and carbides of Nuclear fuels

uranium and fissile metals

18.5. Reference
A. Guinier, La structure de la mauere, Hachette, 1980.
 General bibliography

C. ANDERSON, K.D. LEARER, R. D. RAWLINGS, J. M. ALEXANDER, Materials Science, 3 ed., rd

Van Nostrand Reinhold, Wokingham, Berkshire, U.K. (1985).

M. F. ASHBY, D. R. H. JONES, Engineering Materials, An Introduction to their Properties and
Applications, 1 ed., Pergamon , Oxford, 1980 and 2 ed., Butterword-Heinemann, Oxford, 1998.
st nd

D. R. ASKELAND, The Science of Engineering Materials, 2 ed. Chapman & Hall, London, 1990.
n d

M. F. ASHBY, D. R. H. JONES, Engineering Materials 2, An Introduction to Microstructures, Processing

and Design, Pergamon, Oxford, 1986 and 2 ed., Butterword-Heinemann, Oxford, 1996.
nd

. F. ASHBY, Materials Selection in Mechanical Design, Buttworth, Heinemann, Oxford, 1995.

J. BENARD, A. MICHEL, J. PHILIBERT, J. TALBOT, Metallurgie generale, 2 ed., Masson, Paris,

1984.
C. B. BUCKNALL, Toughened Plastics, Applied Science Publishers, London, 1977.
W. D. CALLISTER, Jr., Materials Science and Engineering, An Introduction, 4 ed., Wiley, New York, th

1997.
. M. EISENSTADT, Introduction to Mechanical Properties ofMaterials, Macmillan, New York, 1971.
S. R. ELLIOT, Physics ofAmorphous Polymer, Longman, Harlow, Essex, UK, 1989.
J. D. FERRY, Viscoelastic Properties ofPolymers, 3 ed., Wiley, New York, 1980.
th

D. FRANCOIS, A. PINEAU et A. ZAOUI, Comportement mecanique des materiaux, Hermes, Paris,

1995.
D. R. GASKELL, Introduction to Metallurgical Thermodynamics, 2 ed., Mc Graw-Hill, New York, n d

1981
M. GERL et J. P. ISSI, Physique des Materiaux, (Traite des Materiaux, vol. 8), Presses polytechniques et
universitaires romandes, Lausanne, 1997.
U. W. GEDDE, Polymer Physics, Chapmann & Hall, London,1995.
. K. GODOVOSKY, Thermophysical Properties of Polymers, Springer, Berlin, 1992.
A. Y. GROSBERG, A. R. KHOKHLOV, Stastistical Physics of Macromolecules, AIP Press, New York,
1994.
A. GUINIER, La Structure de la Matiere, Hachette, Paris, 1980.
C. HAMOND, The Basics of Crystallography and Diffraction, Oxford University Press, Oxford, 1997.
R. W. HERTZBERG, Deformation and Fracture, Mechanics of Engineering Materials, 4 ed., John th

Wiley, New York, 1996.

F.J. HUMPHREYS, M. HUTHCALY, Re crystallization and Related Annealing Phenomena, Pergamon,
1995.
B. ILSCHNER, Wertetojfwissenschaften, Springer, Berlin, 1982.
A. KELLY, The changing Cycle of Engineering Materials, Second Finniston Lecture, London, 1994,
Interdisciplinary Science Review, 19, 3, (1994).
A. KELLY, Concise Encyclopedia of Composites Materials, Elsevier, Oxford, 1995.
A. KELLY et . . MACMILLAN, Strong Solids, 3 ed., Clarendon Press, Oxford, 1986.
rd

C. KITTEL, Introduction to Solid State Physics, 7 ed., John Wiley, New York, 1996.
th

J. I. KROSCHWITZ, ed., Concise Encyclopedia of Polymer Science and Engineering, Wiley, New York,
1990.
W. KURZ, D. J. FISHER, Fundamentals of Solidification, Trans. Techn. Publications, Zurich (Suisse),
1998.
J. LANCASTER, Engineering catastrophes. Causes and effects of major accidents, Abington Publishing,
Cambridge, 1996.
D. LANDOLT, Corrosion et Chimie de Surface des Metaux, (Traite des materiaux, vol. 12), Presses poly
techniques et universitaires romandes, Lausanne, 1993.
P. K. MALLIK, Fiber-Reinforced Composites, Dekker, New York, 1988.

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

446 Introduction to Materials Science

J. L. MARTIN et A. GEORGE, Caracterisation experimentale des Materiaux II, (Traite des materiaux,
vol. 3), Presses polytechniques et universitaires romandes, Lausanne, 1998.
J. P. MERCIER et . MARECHAL, Chimie des polymeres, (Traite des materiaux , vol. 13), Presses
polytechniques et universitaires romandes, Lausanne, 1996.
J. P. MERCIER et P. GODARD, Chimie Organique. Une initiation, 2 ed., Presses polytechniques et
eme

universitaires romandes, Lausanne, 2001.

J. PHILIBERT, A. VIGNES, Y. BRECHET et P. COMBRADE, Metallurgie: du mineral au materiau,
Masson, Paris, 1998.
D. A. PORTER, K. E. EASTERLING, Phase Transformation in Metals and Alloys, 2 ed.,Chapman &
nd

Hall, London., 1992

D. V. RAGONA, Thermodynamics of Materials, Wiley, New York, 1995.
. M. RALLS, . H. COURTNEY. J. WULFF, Introduction to Materials Science and Engineering,
Wiley, New York, 1976.
D. W. RICHERSON, Modern Ceramic Engineering, Dekker, New York, 1992.
W. SCHNABEL, Polymer Degradation, Hanser, Munich, 1992.
. B. SMITH, Basic Chemical Thermodynamics, Clarendon Press, Oxford, 1990.
L. H. SPERLING, Introduction to Physical Polymer Science, 2 ed., Wiley, New York, 1992.
nd

L. R. G. TREOLAR, The Physics of Rubber Elasticity, 3 ed., Clarendon, Oxford, 1975.

rd

D. W. VAN KREVELEN, Properties of Polymers, 3 ed., Elsevier, Amsterdam, 1990.

rd

J. D. VERHOEVEN, Fundamentals of Physical Metallurgy, Wiley, New York, 1975.

I. M. WARD, DW. HADLEY, An Introduction to Mechanical Properties of Solid Polymers, Wiley,
Chichester (U.K.), 1993.
H. YANAGIDA, K. KOUMOTO, . MIYAYAMA, The Chemistry of Ceramics, Wiley, Chichester,
U.K., 1996.
R. YOSOMIYA, MORIMOTO, A. NAKAJIMA, Y. IKADA, T. SUZUKI, eds., Adhesion and bonding
in composites, Dekker, New York, 1990.
J. ZARZYCKI, Glasses and the vitreous state, Cambridge University Press, Cambridge, 1991.
 Glossary of principal symbols

Symbol Description Unit

A Free energy (Helmholtz function) (JmoF ) 1

A s Reduction in area (ductility) (%)

a, b, c Lattice parameters (nm)
Magnetic induction (T)
b Burgers vector
bcc Body-centred cubic lattice
-
-

B r
Magnetic remanence CO
c Concentration (composition) of component /. wt.%, g/g, at.%
est Constant -

D Diffusion coefficient (mV)

D T
Thermal diffusion coefficient (ms )
d G
Average diameter of grain ()
Elasticity modulus of (Young's modulus) (GPa)
E D Band gap energy (forbidden band energy) (eV)
E (t)
r Tensile relaxation modulus (GPa)
F Applied force (N)
F r
Elastic retraction force (N)
f Number of freedom degree - Variance

fj Molar fraction of phase j -

/A Attractive force (N)

f Repulsive force (N)
fee Face-centred cubic lattice -
G Shear modulus (GPa)
G Free enthalpy (JmoF ) 1

hep Hexagonal close-packed lattice -

Magnetic field (Am" ) 1

Enthalpy (JmoF ) 1

He Coercive field (magnetic coercivity) (JmoF ) 1

HB Brinell hardness -
HV Vickers hardness -
h Planck's constant (6.63 10~ ) 34
(Js)
h,k,l Miller indices for a crystallographic plane -
I Electric current intensity (A)
I Polymolecularity index -
J Diffusion flux (m-Y ) 1

Jc(t) Creep compliance (1/GPa)

Bulk modulus (GPa)
Stress-intensity factor (MPam ) 0 5

Kc Fracture toughness (MPa m ) 0 5

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

448 Introduction to Materials Science

k Boltzmann's constant (1.38" ) 23

(JK ) 1

I Length (m)
I Crack length (nm)
I Critical crack length (nm)
M e Molecular weight of elastic chains (gmoF ) 1

M Number-average molecular weight (gmoF|)

n

M Weight-average molecular weight (gmol ) ]

m Masse (kg)
m e
Electron mass (9.11.10 )31
(kg)
mi Quantum magnetic number -
m, Quantum number of spin -
N A
Avogadro number (6.02 10 ) 23
(moF ) 1

Pressure (Pa)
Power (W)
Dipole moment (D)
Number of phases -
Q Activation enthalpy (energy) of the diffusion (JmoF ) 1

Q Thermal flux (Wm~ ) 2

q,q- Electrical charges (C)

R Gas constant (8.314) (J mol^K" ) 1

R Rate (mol F s ) 1 -1

R Electrical resistance ()
Re Yield strength (MPa)
Rm Ultimate tensile strength (MPa)
Ro.2 0.2 % off-set yield strength (MPa)
s Entropy (J K" moF )
1 1

Temperature (K)
T c Critical temperature (superconductivity) (K)
Td Line tension (dislocation) (N)
T g
Glass transition temperature (C) (K)
T m
Melting temperature (C) (K)
t Cooling rate (Ks" ) 1

Internal energy (J moF ) 1

Ue Volume deformation energy (Jm" ) 3

Ur Fracture energy (Jm" ) 3

V Voltage (V)
w x Weight fraction of constituent -
Xx Molar fraction of constituent -
Atomic number -
a,, Volume expansion coefficient (" ) 1

7 Shear strain -
7 Specific surface energy (Jm" ) 2

Nominal strain -
, True strain -
Viscosity (Pas)
Bond angle ()
XR diffraction angle ()
 Glossary of principal symbols 449

Curie temperature (K)

Neel temperature (K)
Thermal conductivity coefficient (Wm K ) _1 _1

Wavelength (nm)
Magnetic permeability (Hnf ) 1

bond -
Poisson ratio -
Density (gcnf )(tnf )
3 3

Dislocation density (nf ) 2

Electric(al) resistivity ()
Electric(al) conductivity (-' ) 1

bond -
Nominal stress (MPa)

Critical stress (MPa)
<*t True stress (MPa)
Shear stress (MPa)
Relaxation time (s)
Peierl's stress (MPa)

Critical shear stress (MPa)
Xr Magnetic susceptibility -

1 Viscosity Pas
 Index

A Binary systems, phase equilibrium,

Acetobacter xylinum, 432 186
Active sites polymerisation, 90 Blend, 175
Affine deformation, 132 Blok copolymer, 201
AFM, 433 Body-centred cubic (bcc) structure,
77
Ageing, 379
Bolzmann's superposition principle,
Allotropy, 181
144
Alloy, 175
Bonding, 22
Amalgam, 204
-chemical, 22
Amorphous structure, 55, 69, 252,
302 - covalent, 26

Anelastic effect, 273 - ionic, 23

Antibonding molecular orbital, 27 - molecular orbital, 27
Antiferromagnetic materials, 360 - orbital, 27
Atom, 17 - secondary, 22, 40
Atomic - strong, 22
- force microscope, 433 - weak chemical, 22, 39
- movement, 201 Bragg's law, 68
- orbital, 19 Bright field, 241
- origin of elasticity, 127 Brittle Fracture, 321
- radii, 440 - critical stress, 323
- structure, 17 Building block (structural unit), 81,
Auto-diffusion, 209 89
Auto-oxidation, 389 Bulk modulus, 125
Auxochromes, 371 Burgers circuit, 159
Average molecular weight 98 - vector, 159
Avrami's equation, 227
C
Carbon atom, 30
Bacterial cellulose, 431 - fibre, 46, 406
Band, 3 6 - 3 9
- hybridisation, 30
- conduction, 37
- nanostructures, 425
- energy, 36
- nanotubes, 425
- forbidden, 38, 349
Bathochrome, 368 - properties and applications, 427

2002 Editions scientifiques et medicales Elsevier. All rights reserved.

Introduction to Materials Science

452 Introduction to Materials Science

Cast iron, 196, 247 Coordination number, 77

Casting 235 Copolymer, 200, 251
Cell - Synthesis, 202
- multiple, 56 Copolymerisation, 251
- primitive, 56 Corrosion, 379, 380, 381, 385
- unit, 56 -ceramics, 396
Cellulose, 432 - d r y , 386
Ceramic, 4, 105, 249 - electrochemical, 381
- a l l o y s , 199 -galvanic, 381
-classification, 104,444 - protection against, 385
- chemical corrosion resistance, 395 - resistant materials, 388
Chain breaking antioxidants, 392 - w e t , 382
Chain-reaction polymerisation, 90 Coupling agents, 408
Chemical etching, 240 Covalent bond, 26 - 28
- ageing, 379 Crack
- properties 5 - instable crack, 325
Clinker, 110 - formation energy, 325
Coercive field, 362 - propagation, 323, 331
Cohesive forces, 41 - stable, 325
Cohesive energy, 51 Creep, 305
Coincidense lattice, 168 -compliance, 142
Colouring of materials, 368 - c u r v e , 306
- o f organic substances, 371 Critical crack length, 325
Comet aircraft, 333 Critical crack stress, 323, 326
Commodity plastic, 95 Cross-linked polymer, 45, 83, 91,
100, 136, 302
Composite materials, 401
Cross-linking, 100
- elastic behaviour, 402
Crystal growth, 221
- fibre reinforced, 401
Crystal structure, 56, 69
- matrices, 407
- determination, 65
- sandwich structure, 401, 411
Crystalline polymers, 101
- types of fibre, 404
- lattice structure, 101
- w o o d , 401
- spatial configuration, 101
Compound, 195
Crystalline direction, 64
Compression, 125
- systems, 64
Concrete, 4, 109
Crystallography, 56
Conductor (electrical), 340
Conformation, 102 D

Constituent, 176 Deformation

Contact lens, 114 -affine, 132
 Index 453

-elastic, 121 Dry corrosion, 385

- e n e r g y , 271 DSC, 191
- internal energy, 127 Ductility, 262, 279
- m o d e , 142 Ductile failure, 327
- nominal, 262

-plastic, 159, 262
Ebonite, hard-rubber, 304
- simple, uniaxial, 121
Edge dislocation, 157
- thermoplastic, 287
Elastic behaviour, 126
- t y p e , 142
Elastic deformation of a perfect
Degradation, 379,380, 387 crystal, 131
- o f PVC, 394 Elastic deformation of ideal rubber,
Derealisation electrons, 34 136
Demagnetisations curve, 362 Elastic modulus, 122
Dendrite crystal, 216, 223, 244 Elastomer, 100
Density, 439 Electrical conductivity, 340
Dental caries, 203 - metals, 344
Depolymerization, 389 - semiconductors, 349
Diamagnetic materials, 22, 358 - semimetals, 349
Diamond films, 169 - organic molecules, 436
Differential Scanning Calorimetry, Electrical resistivity, 340
191
Electroaffinity, 23
Diffusion, 209
Electrochemical cell, 381
Diffusion coefficient, 211
-corrosion, 381
Diffusional flow, 305
Electrode potential, 381
Dimensions of macromolecule, 97
Electromagnetic radiation, 367
Dislocation, 157:
Electron microscopy, 240
- climb, 307
Electron-tunnelling microscope, 433
- creep, 306
Elements, 3, 8
- density, 165
Energy
- e d g e , 157
- bands, 36
- effect of particles, 294
- cohesive, 51
- internal, energy 291
- f r e e , 128
- l i n e , 157
- internal, 127
- m i x e d , 157
-mouvement, 162 - r u b b e r , 137

- obstacles interaction, 295 - specific fracture, 136

- s c r e w ,157 - specific interfacial, 217

- source, 294 -surface, 179
Drift velocity of the electrons, 345 -thermal, 51
454 Introduction to Materials Science

- volume deformation, 271 G

Energy band, 36 Galvanic corrosion, 381
Engineering ceramics, 112 Galvanization, 2, 385
Equiaxial growth, 224 Gas turbine, 315
Equilibrium concentration of - b l a d e , 316
vacancies, 152
Gel spinning, 304
Equivalence principle, 313 Glass, 55, 109, 182
Eutectic - c e r a m i c , 109
- d i a g r a m , 194
- forming substances, 181
- growth, 225
- ionomer cements, 204
- microstructure, 244
-metallic, 86
Eutectoid transformation, 198,215
- transition range, 183
F - transition temperature, 182, 302
Face-centred cubic (fee) structure, 77 Glass fibre, 405
Faded memory, 140 - coupling agent, 408
Fatigue, 263, 331 - mechanical and physical
Ferrimagnetic materials, 360 characteristics, 405
Ferrite, 360 Glassy state, 55
Ferromagnetic materials, 359 Graft copolymer, 201
Fibre-reinforced, 401 Grain, 166
Fick first law, 212 -boundary, 166
Fick second law, 214 - coarsening, 309
Fluorescence, 368 - g r o w t h , 178
Flux, 211 Grey cast iron, 199
Forbidden band 38 Growth at atomic scale, 221
Fourier laws, 352 - eutectic, 225
Fracture, 321 - microstructures, 223
- b r i t t l e 321

- crack propagation 322 Hard magnetic materials, 361
- ductile 328 Hardener, 101
- e n e r g y , 133 Hardening, 247, 290, 296
- m e c h a n i c s , 321, 327, 326 - grain boundary effect, 296
- m e c h a n i s m s 330 -mechanisms, 289
- stress concentration 322 - polycrystal effect 298
- toughness,252, 262, 327, 439 - solid solution, 295
Free energy, 128 - work 296
Friction coefficient, 418 Hardness, 273
Fullerene molecules, 425 Heterotactic polymer, 92
 Index 455

Hexagonal close-packed (hep) Isotactic polymer, 92

- structure, 76 Isotactic polypropylene, 92, 103
- clusters, 422
High temperature, 280
J
Joule effect, 340
-corrosion, 385
Jump frequency, 210
HIPS, 252
Homogeneous nucleation, 216
Hooke's Law, 126 Kinetics of phase transformations,
Hume- Rothery rules, 191 225
Hund'sRule,21
L
Hybridisation of orbitals, 31
Laminated structures, 414
Hydraulic cement, 110
Larson-Miller (LM) parameter, 313
Hydrostatic pressure, 125
Laser, 254
Hysteresis loop, 362
Lattice plane, 56
I Lattice row, 56
Ideal elastic behaviour, 145 L C A O , 26
Ideal rubber, 129 Life cycle of materials, 9
Inoculant, 220 Light
Inorganic glass, 185 -absorption, 368
Instable crack, 325 - interaction with solids, 365
Instantaneous dipole, 40 Limit of linear viscoelasticity, 144
Insulator, 340 Limiting oxygen index (LOI), 391
Inter-atomic force, 40 Line tension, 292
Interface, 178 Linear polymer, 81
Intermediate phase, 178, 195 Linear viscoelasticity, 143
Intermetallic phase, 178 Liquide phase sintering, 249
Internal energy deformation 127 Liquidus, 189
Interstitial atom, 152 Loading modes, 323
Interstitial site, 79 Low temperature, 280
Interstitial solution, 209 Lyotropic polymer solution, 304
Ion, 23
Ionic
- bond, 42 Macromolecule dimension, 97
- crystal, 79 Magnetic
- radii, 440 - classification of materials, 358
- solids, 79 -field, 357
Ionisation potential, 23 -hardening, 363
Iron-carbon alloy, 196 - induction, 357
Isostatic compression, 249 - loss, 362
456 Introduction to Materials Science

-permeability, 357 - electron, 240

-properties, 357 - optical, 239
- r e m a n e n c e , 362 Microstructure, 7 , 2 4 2
Magnetization vector, 357 - elements, 8
Manipulation of atoms and molecules, - growth, 223
433 - high impact polystyrene, 252
Martensite, 233 - iron - carbon alloys, 247, 249
Martensitic transformation, 232 - metal alloys, 244
Master curve, 311 - non-equilibrium conditions, 245
Materials Science, 7 - observation, 239
Material, 1 - polymers bulk, 242
- classes, 2 - polymer blends, copolymers, 251
- complementarity, 2 - pure metal solidification, 242
- cycle, 9 -tri-block copolymers, 251, 253
- physical states, 51 Miller-Bravais indices, 62
- properties, 5 Miller indices,58, 60
- r e s e r v e s , 10 Mineral glass, 109
-resources, 10 Mineral polymers, 84
Matrices, 407 Mixed dislocation, 157
Maximum strength, 134 Mixed microstructure, 244
Maxwell model, 141 Modifying agents or fluxes, 109
Mean drift velocity (electrons), 345 Modulus

Mechanical - b u l k , 125

-behaviour, 139 -elastic, 122

-properties, 439 -relaxation, 143

- Y o u n g ' s , 122
Mesomeric phenomenon, 34
Modulus of elasticity, 263
Metal, 4, 75, 196, 223, 234, 244
Mole, 17
- a t o m i c radii, 441
Molecular
- lattice parameters, 78
- electric switch, 436
- oxidative behaviours, 387
- orbitals (MO), 26
- structure, 75
-photovoltaic device, 437
Metallic
- solids, 40
- b o n d , 36 Molecule, 40
- glasses, 86 Mortar, 110
Metastable state, 185 Movement of dislocations, 162
Miscibility in polymer blends, 200 Multiple cells, 56
Micrography, 240 Multiwalled carbon nanotubes, 425
Microscopy MWNT, 425
 Index 457

Nanomaterials, 421 Packing factor, 77
Nanopowders, 423 Paramagnetic material, 22, 358
Nanosized Partial miscibility, 191
- metals, 424 Particles precipitation, 244

- hard magnetic particles, 424 Passivation, 384

- pigments, 424 Patenting, 276

Nanostructured materials, 421 Pauli Principle, 21

Necking, 265 Pearlite (perlite), 248

Perfect gas, 129
Newtonian fluid, 139
Performance, 8
Noble metals, 382
Nominal deformation, 262 Periodic table, 3

Nominal stress, 262 Peritectic diagram, 194

Notation of directions, 64 Permanent magnet, 360, 374

Nucleation, 215 Phase, 175

- homogeneous 216 Phase diagram, 186

- n e w phase 215 - binary systems 186

- p r i m a r y 221 -eutectic, 193
- r a t e 219 -peritectic, 194
-secondary 221 Phase diagrams
Nucleating agent, 220 - (Al - Cu), 247
Nuclei, 215 -(AI2O3 - S i 0 ) , 199
2

- ( C u - M g ) , 197
Number of degrees of freedom, 180
- ( F e - C ) , 198
- ( N i - C u ) , 190
Octahedral site, 79 - of a pure substance, 179
Ohm's law, 340 - w i t h complete solubility, 186
Optical microscopy, 239 Phase equilibrium diagram, 175, 179
Optical properties, 364 - binary systems, 186
- m e t a l s , 371 Phase transformation, 214
- semiconductors, 371 Phosphorescence, 368
Ordered state, 51 Phyllosilicates, 106
Organic matrices, 407 Physical
Organic polymer, 4, 85, 89, 216, 243, - ageing, 379
251,300
-properties, 339
Oxidation, 385
Photodegradation, 391
Oxidation number, 24
Oxide film, 386 Photooxidation, 391
Oxides formation enthalpy, 386 Plastics
458 Introduction to Materials Science

- commodity, 95 Precipitation, 246

- engineering, 95 Preventive antioxidants, 392
- technical, 95 Primary creep, 307
Plastic deformation, 159, 262 - nucleation, 221
Plastic zone, 329 Primitive cell, 56
Point defect, 151, 155 Properties, 5
Poisson ratio, 123 Protection of metals, 384
Polishing, 240 Prostheses, 13
Polycrystalline structure, 54, 166, PVC
242, 297
-degradation, 393
Polyethylene chain structure, 83, 104 - stabilization, 393
Polypropylene (PP), 93 Pyrolysis reactions, 388
Polydiene, 94
Polyethylene (PE), 82 Q
- crystal structure, 104 Quantum
Polymer, 99, 89, 200, 243, 251,387, - dots, 423
409 -numbers, 1 9 - 2 1
- chemical agents resistance, 395 - wells, 423
- corrosion resistance, 395 - wire, 423
- c o h e s i o n in, 300 Quenching, 154
-degradations, 388
R
- influence of temperature, 301
Real crystal, 54, 151
- main types, 99
Recovery, 309
- mechanical properties 303
Recrystalisation, 309
- mineral, 84 Recycling, 9
- organic, 89 Reduction in area, 265
- snaking.303 Reflection microscopy, 239
- viscoelastic behaviour, 139, 309 Refractive index, 366
Polymer materials Refractory metal, 387
-formulae, 441 Reinforced concrete corrosion, 396
- nomenclature, 441 Relative permeability, 357
Polymerisation, 90 Relaxation modulus, 143
Polymolecularity index, 98 Reptation, 212, 303
Polymorphism, 181 Repulsive force, 43
Polynuclear orbital, 34 Reserve, 11
Polypropylene, 93, 102 Residual resistivity, 347
Polypyrrole, 430 Resin composite, 204
Polysiloxane chains, 83 Resonance form, 34
Polyvinylchloride (VC), 91 - h y b r i d , 34
 Index 459

Resource, 11 Simulation, microstructure

Reticulated polymer, 83, 100, 136 development, 236
Ribosomes, 422 Snaking, 303
Rough interface, 221 Snowstorm, 234
Rubber, 94, 136, 268, 302 Soft magnetic materials, 361
- tensile behaviour, 268 Solid
Rupture elongation, 263 - ionic, 79
-solution, 177
Rust, 382
- with covalent bonds, 81, 85, 89 ,104
S Solid state, partial miscibility, 191

Sacrificial anode, 385 Solidus, 189

Sandwich structures, 411 Solubility range, 177

Solute, 177
Scanning electron microscope, 241
Space lattice, 56
Schmid law, 283
Specific fracture energy, 135, 326
Screw dislocation, 157
Specific interactions, 200
Secondary creep, 307
Specific interfacial energy, 217
Secondary nucleation, 221
Specific surface energy, 324
Secondary recrystalisation, 309
Spherulite, 242
Semiconductor, 38, 340, 349
Spherulitic growth, 226
Semimetals, 349
Spinodal decomposition, 229
Shear, 125
Stabilizers, 392
- a n g l e , 125
Stable crack, 325
- deformation, 287
Stacking fault, 168
- deformation velocity, 287 Statistical degradation, 388
- m o d u l u s , 125 Steel, 196, 247
- yield stress, 282 Stellite, 256
Sialon, 176 Step growth polymerisations, 95
Silicon mono-crystal, 70 STM, 433
Silica-alumina alloys, 199 Stone cancer, 395
Silicates, 105 Strain, 122
Single-constituent systems, 179 Strength
Single walled nanotubes, 425 -material, 261
Sintering, sintering processes, 249 - ultimate tensile, 263
Skis, 414 - y i e l d , 263,290
-parabolic, 414 Stress, 122
Slip plane, 159 - concentration factor, 322
Slip, 286 - nominal, 262
Smooth interface, 221 - Peierls, 293
460 Introduction to Materials Science

-relaxation 143 -coefficient, 3 5 3 , 3 5 4 , 3 5 6

- relaxation modulus, 142, 302, 309 -mechanisms, 355
- t r u e , 270 Thermal energy, 51
Stress - strain curves Thermodynamic term, 220
- enthalpic elastic materials, 263 Thermodynamics of uniaxial
- r u b b e r , 268 elongation, 127
Stress-intensity factor 327 Thermoplastic
Structure - matrices 408
- amorphous, 55, 86, 300 -polymer, 81, 99
- determination, 66 - rubber, 252
- metal, 75 Thermorheologically simple
behaviour, 312
- polycrystalline, 54
Thermal analysis, 191
- solid, 53
- s u p e r , 168 Thermoset polymer, 100

Structural unit or motif, 56, 94,441 Thermosetting matrices, 407

Substitutional atom, 152 Tilt boundary, 166

Substitutional solid solution, 209 Time factor, 312
Sunlight spectrum, 365 Toughness, 262, 321,327
Superstructure, 168 Track-etch nanoporous membranes,
Superplasticity, 309 429

Surface energy, 135, 179, 324 Transformations

- treatments (fibres), 407 - Avrami theory, 225

SWNT, 425 - by spinodal decomposition, 230

Syndiotactic polymer, 92 - Time - Temperature ( T T T ) , 226
Transition from micro to nano, 422
Transmission electron microscope,
Technical ceramics, 112 240
Technical or engineering plastic, 95 Transmission microscopy, 239
Temperature equivalent principle, 312 Transport term, 220
Template method, 428 True strain, 270

Tensile curve, 263 True stress, 270

Tensile test, 262 TTT diagram, 226

Tertiary creep, 307 Twin boundary, 169

Twinning, 286
Tetrahedral site, 79
Twist boundary, 166
Theoretical strength, 133
Two-dimensional lattice, 56
Thermal analysis, 191
Types of fibre, 404
Thermal conduction 352
Thermal conduction mechanisms 355 U
Thermal conductivity Ultimate tensile strength, 263
 Index 461

Unidirectional fibre reinforced Voigt-Kelvin, 141

composites, 410 Volume deformation energy, 271
Unit cell parameters, 56 Vulcanisation, 100

V W
Vacancy loop, 161 Weathering, 391
Vacancy, 152 Weiss domains, 359
Vacancies, equilibrium concentration, Wet corrosion, 382
152 Whisker, 291
Valency electron, 23 White cast iron, 199
Valency, 23, 24, 25, 80 Wood, 401
Valve seats, 256 Work hardening, 264, 296
Van der Waals force, 40
X
Variance, 180
X-ray diffraction, 66
Viscoelastic behaviour, 140, 309
Viscoelastic body, 139 Y
Viscoelastic regions, 309 Yield strength, 263, 289
Viscous behaviour of Newtonian - industrial alloy, 299
fluids, 139 - iron and F e - C alloys, 300
Vitro-ceramics or pyrocerams, 109 Young's modulus, 122