Analysis of Natural and Synthetic Rubber by Infrared Spectroscopy

H. L. DINS3IOHE'
AND

DON C. SMITH, Naval Research Laboratory, Wushington, D. C .

tures, quantitative analysis of natural-GR-S blends, and determination of the nitrile content of Buna-N copolymers. An accuracy corresponding to * 1% average error has been obtained for these determinations. The methods can be extended to analogous problems with q minimum of alteration, and the same order of accuracy can be expected in general. A discussion of the correlations between spectral absorption and molecular structure is also included.

This report describes an investigation of the applicability of infrared spectroscopic methods for quantitative determination of the elastomer composition of industrial rubber products. Generally satisfactory techniques for obtaining a valid spectrum of the elast omercontent of any rubber specimen have been developed and evaluated, have been applied in developing detailed analytical procedures for identification and qualitative analysis of elastomer mix-

means for identification of organic materials and analysis of complex mixtures ( 2 , 8, 13, 16). Kot only are spectroscopic methods rapid and generally conservative in sample requirements, but the results are often more specific than those obtained by chemical analysis, distillation, etc., and in many cases furnish detailed structural information of importance to research and development. The advantages of infrared analysis lie in the facts that (1) the spectrum of a molecule is a unique physical property which is not altered by any changes in which molecular identity is maintained, and that (2) the individual features of the spectrum relate in a definite manner to the chemical bonds comprising the molecule and can thus be interpreted regardless of the t o t a1 configuration. Procedures for accurate determination of the identities and proportions of the elastomers present in a finished product are needed throughout the rubber industry for control and specification. With the ever-groning number and variety of rubber substitutes and special-purpose synthetics incorporated in manufactured goods either singly or as blends, the problem facing the rubber analyst has become increasingly complex and well beyond the scope of conventional chemical methods. In addition, analvtical methods allowing study of polymer structure should prove significant in the development of new polymers and in the evaluation of polymerization techniques. Several investigators have recognized the advantages of employing a physical means of investigation by which the various types of chemically inert molecular structures in polymers may be studied without changing their identities or proportions. Infrared methods have been significant, for example, in exploring thc mechanism of polymer formation (15) and in studying vulcanization (12) and oxidation ( 5 ) . Considerably less has been donc to develop infrared methods into routine procedures for specification of rubber products. This is undoubtedly due, in a large part, to the experimental difficulties encountered in spectroscopic examination of vulcanized and reinforced rubbers. An attempt t o overconie these difficulties was made by Barnes and co-workers (4j n l i o developed a procedure for the analysis of nat urd-Buna-S blends in tread, carcass, and tube stocks. V-hile their results Tvere satisfactory for the purposes of the study, the accuracy did not appear as good as might be expected and the data presented did not permit conclusive evaluation of the procedure, particularly in regard t,o extension of the method to other elastomers. This paper describes the results of an extensive study of the applicability and limitations of infrared methods for the identification and quantitative analysis of the elastomer content in commercial articles of natural rubber and some of the common synthetics. The first phase of the work was concerned with

XFKARED spectroscopy is finding increasing application as a

deterniining valid absorption spectra of the various pure elastomers. These reference spectra were required to evaluate the spectral basis for identification of elastomers and for later use in identifying any spectral effects not directly related t o elastomer composition. h second phase was concerned both v,ith the development of a satisfactory technique for spectral examination of a compounded and cured rubber and with the determination of the extent to vihich the spectral absorption so obtained is related to elastomer composition, as distinguished from other spectral effects of sample history or state, etc. The necessity of removing fillers n hich render the sample too opaque for quantitative spectral study and the additional advantages of eliminating all effects of nonrubber components from the spectral results dictated the use of procedures involving isolation of the elastomer content as a' preliminary step. In this connection, solvents for cured elastomers and methods for removal of colloidal carbon from rubber solutions have received special attention. 1third phase was concerned with the development and evaluation of detailed procedures for certain analyses pvhich are of practical importance and which illustrate the general feasibility of quantitative applications.
GCnI ELASTOMERS

The problem of determining reference spectra-Le., the correct spectrum 6f the pure rubber hydrocarbon-is best approached by the study of gum rubbers Therein the elastomer chains are alieady large compared with the polymerizing units (or monomers) but where modifying effects such as may accompany compounding, curing, aging, etc., have not yet been introduced. Thirteen gum samples comprising seven different classes of elastomers were used for this purpose. The samples were 97W pure elastomer, the remainder being chiefly antioxidant, and are listed in the tabulated summary of studies on gum elastomers in Table I. Sample Preparation. The gum samples were prepared for spectroscopic examination by (1)solvent extract ion of antioxidant and all soluble impurities, (2) solution of the elastonier in a suitable solvent, (31 removal of excess solvent by evaporation, and (4) preparation of films on rock salt or potassium bromide plates for spectroscopic study. Conunercial grade ethyl alcohol (95%) was used for extraction in the standard X.S.T.M. rubber extraction assembly (1) consisting of 400-1111. flask, siphon cup, and lvater-cooled tin-coil condenser. Solubility and ensuing gelation prohibited the use of other solvents such as benzene, chloroform, and acetone. A 2-gram sample v a s cut into thin strips and placed in parallel r o w on a filter paper, so as not t o be in contact when the paper was rolled. Two or three 80-ml. portions of solvent n'ere used and the progress of extraction was followed by noting the coloration. On the basis of subsequent spectra, continuous extraction for 24 hours was shown to be sufficient. Solutions of the various elastomers were made a t room temperature in order t o minimize possible decomposition or alteration of the elastomer. Solvents used were standard technical grade

* Present addreas, University of

Minnesota, Minneapolis, Minn.

11

12
Table I.
Samule Elastomer class Natural rubber

ANALYTICAL CHEMISTRY
Summary of Investigations of Gum Rubber
Preparation Extrac1tion" Solution Solvent methodb Xylene nitroethane S (8:3) Solution C characteristics Gray, W Film
No.d

---~~~.~~~

Representative Samule Nd. type 1 Natural smoked sheet

Lee pure gum cement Tube reclaim0

..
M M M

..
p-Cymene xylene (4:l)

Colorless Yellow, W + Yellow, \V+ Yellow, W + Tan, W Brown, G Brown, W Brown. Brown, Brown, Brown,

1
2 3

+ p-Cymene + xylene p-Cymene + xylene

20 22 70 149 432 434

66

R
S
S

Isobutylene-isoprene copolymer Chloroprene polymers

GR-I Keoprene-GN

1 1 2 3 1 2 3 4 1

L L L L
L

..

Benzene nitroethane Xylene Ethylene dichloride

S
B

Keoprene-ILS

Ethvlene XyGne Ethylene Ethylene

dichloride +dichloride nitroethane dichloride

S
B B B S H B B B B

W G
P

67 71 71 82 120 121 50 D 47 157 59 D 52

61

Spectral Study Wave length Spectral range, charactermicrons isticse 2-14 E 2-14 E 14-23 f 2-14 E E 2-14 I 14-23 E 2-14 E 2-14 E 2-14 A 2-14 . I 2-14 A 2-14 A 7-14 2-14 E 2-13 14-23 2-14 2-14 2-14 14-23 2-14 2-14 2-14 2-14 2-14 2-14 2-14 2-14 2-14 2-14 6-14 2-14 2-14 2-14 2-14 2-14 6-14 6-12 2-14

...

..

...

E E

..
L
L
L
L L

Butadiene-styrene copolymer Bu tadiene-acrylonitrile copolymers

GR-S Chemigum-N-1 Hycar-OR-15 Hycar-OR-25 Perbunan-35

1
1 2 3 1 2 3 1 2 3

..

L
L
L

..

Ethylene dichloride nitroethane Xylene nitroethane Xylene Xvlene nitroethane Ryethy1 ethyl ketone Methyl ethyl ketone Xylene nitroethane Methyl ethyl ketone Methyl ethyl ketone Nitroethane xylene Methyl ethyl ketone Methyl ethyl ketone

+ + + +

S
B B

Brown, P Tan, W Brown, W Yellow, P Yellow, A '' Yellow, W Yellow, P Yellow, W Yellow, W Yellow, P Yellow, mT Yellow, W Yellow, P Brown, W

157 61 29 30 32 94 94 41 56 97 53 114 119 33 24 26 27 28 97 D 62

E E E
E . 1 E E E E E l

...

Ethylene polysulfide polymer

Butaprene-KF Thiokol-Fa

1 1

Xylene Xylene

+ nitroethane + nitroethane

H S

E E E E E
E E E E

Copolymer of butadiene, acrylonitrile, a n d unsaturated ketone

5

Tbiokol-RD

1
2

L L

Xylene nitroethane Ethylene dichloride

d
6

S B

Brown, N Brown, P

...

..

b S.

L. Extraction by 95% ethyl alcohol. M. Extraction by 68-32 chloroform-acetone.

H. R. B. W.

Mechanical shaking a t room temperature. Same as S, followed b y heating. below 90" C. Refluxing, above 1 0 0 ' C. Blenderization,,after preliminary swelling a t room temperature. Complete solution; no gel. P. Over 60% solution; remainder gel. G. Nearly all gel; very little solution. N. Solvent ineffective; very little gel or solution. Removal of filler b y centrifugation.

attributable t o impurities. f N o appreciable absorption in this region. Contained red iron oxide filler.

E Entire ?pectr:im :atisfactor!: and considered \.slid. . I . Specrrurn satisfactory except for additional band obviously

D.

Sample discArded; no spectrum obtained.

+.

and contained no nonvolatile matter. Of the solvents investigated only ethylene dichloride satisfactorily dissolved all elastomers studied, including neoprenes. hlethyl ethyl ketone was very effective for Buna-N copolymers, but not for the others. Natural rubber, Thiokol-FA, GR-I (Butyl), and GR-S (Buna-S) dissolved readily in several solvents (see Table I). The average requirement was 40 to 60 ml. of solvent per gram of rubber. The solutions obtained were transparent but exhibited light color, which probably resulted from oxidation. Solution of the more soluble elastomers was accomplished merely by 6 t o 8 hours' mechanical shaking of the rubber with solvent. This simple technique did not prove satisfactory for neoprene and Buna-X types because of the persistence of gelatinous structure despite use of excessive quantities of solvent. Solution was accomplished far mole rapidly and effectively by mixing rubber and solvent in a \Taring Blendor. [The Waring Blendor, Cenco K O . 17,233 with 90. 17,244 (small-size) jar was used. iillrubber gaskets were replaced by gaskets of felt between outer ones of lead.] This technique readily reduced gels to thin homogeneous mixtures and permitted satisfactory pieparation of all samples in a minimum quantity of solvent. Generally, the rubber was allowed to soften and swell in the solvent for 2 hours, and then blenderized for 10 to 20 minutes. Tube reclaim was treated as an exception t o the gum samples, as it contained a mineral filler and was considerably less soluble. I t was extracted by a 32-68 blend of acetone-chloroform, and was dissolved by refluxing with a 20-80 blend of xylene-p-cymene. The filler was removed by centrifugation. The solution of polymer was concentrated to a viscous mucilage suitable for film spreading by evaporation on an A.S.T.M. gum

bath at SO" C. Stirring (by hand) adequately eliminated gel formation due to localized surface evaporation. While vacuum distillation \\-as also employed successfully, the former method was preferred as it permitted simultaneous evaporation of several solutions. Films of uniform thickness were prepared on sodium chloride or potassium bromide plates by spreading the concentrated mucilage betvxen parallel spacers, as shonn in Figure 1. Thirty minutes' standing was allowed for final solvent evaporation. The appropriate film thickness for useful spectroscopic measurements was determined by trial preparation of films of several thicknesses and subsequent spectral measurement at the common 6.0 mu carbon-hydrogen band where the absorption should be between 80 and 90%. Three sets of spacers-0.05, 0.10, and 0.20 mm.-enabled suitable films to be obtained in all cases. In the case of Buna-S samples, solvent removal was generally stopped prematurely in order t o avoid serious gelation, and films of suitable thickness were obtained by spreading successive films on one another. All spectroscopic measurements were made using a large recording prism spectrometer of high resolving power (10). Spectra as recorded, consisted of a pair of traces measuring (as percentage full-scale) the energy transmitted by the sample and by a blank plate as a function of wave length (on a nonlinear scale). A percentage transmittance spectrum of the sample \vas obtained by measuring ratios of corresponding ordinates, and the spectrum was transcribed (using calibration data) to linear wave-length charts. Under normal operating conditions wave-length measurements were considered accurate to *0.01 mu from 2 to 15 mu, and per cent transmittance measurements were reproducible to about ~ 1 . 0 % .

V O L U M E 20, NO. 1, J A N U A R Y 1 9 4 8

13
mixes and types of CUE, given in Table 1 1 1 ,which were considered representative of industrial products in variety and character. The method used for preparation and study of these specimens comprised: (1) preliminary milling of the sample, (2) extraction of organic additives, (3) solution of the elastomer, (4) removal of the Iiller, and (5) evaporation of the solvent to obtain a transparent film of pure polymer far spectroscopic examination. Experimental. Extraction and solution of the sample mere facilitated bv preliminarv milhng. This comprised several passes

ETAL SPACERS ROCKSALT PLATE

-I-

Form was~Tsedwith the standard A.S.T.M. extraction assembly ( 1 ) far extraction of organic additives and impurities. Continuous ext,raction for 24 hours with two or three 80-ml. portions of solvent was satisfactory.
WAVELENGTH ( I N MICRONS1 +

-4-

Figure 1. Method of Film Spreading

Discussion of Results. Sever81 satisfactory spectra were obtained over the 2- to 14-micron range for each elastomer studied. Beyond 14 mu the absorption was too weak for analytical purposes and the results have not been included. A set of these reference spectra shown in Figure 2 establishes the fact that the seven different classes of rubber examined possess charaeteristioally different spectra. The qualitative speotrsl differences correspond to basic differences in polymer structures and furnish a n adequate basis for spectroscopic identification of each class. On the other hand, dose similarity exists among the spectra of elastomers which are members of the same basic class of polymer or copolymer, as for the two neoprene polymers or the five Buna-N copolymers (not a11 shown). The ditrerenoes here are quantitative and, though less obvious, are significant if the proportions of the various structures comprising the elastomer are of interest. In addition t o the fact that the use of pure samples assured good reference speotra, the validity of the spectra obtained was established by their close reproducibility during extensive duplication of sample preparation and spectral measurement. For nearly all preparations duplicate Iilm were studied i n order t o determine the reproducibility in Iilm prepartrtltion and spectroscopic measurement. In mveml cases the entire preparation was carried out in duplicate in order that the reproducibility of the over-all method of study might be established. Finally, a number of samples were prepared by alternative procedures, using different solvents aud solution methods, in order that the influence, if any, of particular preparation features on the final spectrum might be shown. I n all cases (in whioh the method of preparation was adequate) virtually perfect reproducibility of spectra was obtained, showing definitely the reliability of the over-all method of study and the sole dependence of the spectrum on theelastomer itself. The spectra obtained, therefore, are considered accurate reference spectra of the pure rubber hydrocarbons.
COMPOUNDED AND CURED ELASTOMERS

Figure 2 . Spectra of Pure Gum Elastomers

The problem of developing suitahle procedures for preparing a commercial rubber ptoduct for spectral examination and of determining t o what extent it is possible to obtain accurate, reproducible spectral data for cured elastomers was solved by extensive experimental work involving both considerable duplication and numerous variations in procedure. In investigating this twofold problem eighteen compounded specimens were studied. These specimens, listed in Table 11, were compounded from the gum elastomers previously studied and included a variety of

sfddered because of the-dificultv of su6smuent solvent removal

Donents and gum-forming impurities.

Of the-solvents iivesti-

'all curkd elastomers except. Neoprene-ILS and Buna-N t h e s , and, despite traces of free hydrochloric acid which persisted on.

14
-

ANALYTICAL CHEMISTRY
Table 11.
Sample Representative type Katural rubber Sample
SO.

Summary of Investigations of Compounded and Cured Rubber

Preparation5 Filler separaGond Spectral Study
W S "C-

Elastomer class Natural rubber

Compositionb F-2-1 F-2-1 F-4

Solvent

Solution characterlSllCS*

Film
NO.]

length range, microns 2-14 2-14 2-14 14-22 2-14 2-14 2-14 2-14 2-14 14-22 2-14 2-14 2-14 2-14 2-14 14-22

- _

Spectral characteristics#

1 2 3 1 2 1 2 3 1 2 3 2
1

Tetrachloroethane p-Cymene o-Dichlorobenzene

.. ..

Yellow, W Yellow, W Yellow, I T Amber, I T Amber, 1 3 ' .4mber, W Amber, JT Amber, W Amber, Amber, Amber, Amber, Amber, N Dark, U Brown, P Brown, P Brown, P Brown, P Brown. P Brown, I T Brown, 1%Brown, JY Brown, W .4mber, P Amber, P Amber, P Amber, P Amber, P
IT IT-

Tube reclaim Isobutylene-isoprene copolymer CR-I

(Sone) Tetrachloroethane (None) Tetrachloroethane F-1002 F-1002 F-1010 F-206 F-206 F-206 F-206 F-210-2 F-207 F-207 F-207 F-207 F-207 F-207 F-207 F-602 F-606 F-606 F-606 Tetrachloroethane p-Cymene o-Dichlorobenzene Tetrachloroethane w-Cvmene b-D;chlorlbenzene o-Dichlorobenzene o-Dichlorobenzene Tetrachloroethane Methyl isobutyl ketone p-Cymene Pvridine ~~~~~-Dichlorobenzene o-Dichlorobenzene o-L>ichlorohmzene
I
~

172 267 386 434 164 276 21 1 268 388 437 182 285 424 405 416 435 D 240 281 355 355 396 425 435 180 438 2-16 288 391 173 287 368 426

E E E
h

E E E E
h

F C

Chloroprene polymers

Keoprene-E Neoprene-GN Neoprene-ILS

F F F F F

E A E E E
h

1 2
3 4 5 6 7

...

..

F F

F F

2-14 2-14 2-14 2-14 2-14 2-14 14-22 2-14 14-22 2-14 2-14 2-14 2-14 2-14 4- 5 2-14 2-14

E E E E E
h

Butadiene-styrene copolymer

GR-S

1
2 3 4

Tetrachloroethane Tetrachloroethane p-Cymene o-Dichlorobenzene

..
F F F

E
E A

Butadiene-acrylonitrile copolymers

Chemigum-N-I

1 2 3 4 5

F-1003 Tetrachloroethane F-1003 p-Cymene F-1003 Pyridine F-1003 o-Dichlorobenzene F-1003 o-Dichlorobenzene

F F F F F

...

E E

E X

a b

All samples milled and extracted with 68-32 chloroform-acetone. See Table I11 for composition a n d cure of vulcanizates. Solution b y refluxing. d F. Filtration with Celite; C = centrifugation. Complete solution of specimen. e W. P. Partial solution (over 40%). N. S o appreciable solution. U. Extent of solution uncertain.
c

I D. Sample discarded; no spectrum obtained. E. Entire spectrum satisfactory. A. Spectrum satisfactory except for additional bands obviously attributable t o impurities. X. Unsat,isfactory spectrum, due mainly t o impurities. h K O appreciable absorption in this region.

Table 111. Composition and Cure

Specimen Designation F-2-1 F-4 F-206 G u m Elastomer Specific Gravity
100

Sulfur 3 .O 3.0 1.0

Zinc Oxide 5.0 5.0

10.0

Stearic Acid 1.0 0.5 0.5

Acceleraton

Antioxidants

Plasticizers, Softeners, Tackifiers Reogen. 2 . 0 Circo Oil, id .'d ' F. F wood rosin 5 0 D i b k y l phthalaie, '20 Circo X Oil, 10 .O Dibutyl sebacate, 10.0 Circo X oil, 5.0

Katural smoked sheet 0 . 9 2 Katural smoked sheet 0.92 Neoprene-E .. Neoprene-ILS Neoprene-GX Butyl (GR-I)

100

100

Methyl zimate, 0 . 1 Age Rite Hipar, 1 . 0 Rotox, 1.0 Captax, 0 . 5 Neozone D. 1 . 0 ..... Neozone D , 2 . 0

F-207
F-210-2 F-1002

..
1.25

100

..

10.0
10.0

0.5
0.5 3.0 3.0 0.5

.....
S o . 552, 0 . 2

100 100
100

..
1.5

..

5.0
5,O 5.0

F-1010-1 Butyl (GR-I) F-606 F-602 F-804 F-801 F-803 F-1003 F-1302 F-407 F-402 F-416 F-205 F-10 F-10-1 F-620 F-421
a

..
0.94 0.94 1.34
1:03
I .

1.5

Buna-S (GR-8) Buna-S (GR-S) Thiokol-FA Thiokol-ST Thiokol-RD Chemigum-Pi-1 Perbunan-35 Hycar-OR-15 Hycar-OR-25 Hycar-OR-25 Butaprene-NF S a t u r a l smoked sheet Tube reclaim GR-S Buna-N
'

100
100

..
3.4

Monex. 1 . 0 Captax, 0 . 5 Captax, 1 . 5 Tuex, 1 . 5 Selenac, 2 . 0

Neosone D , 2 . 0 Permalux, 1 . 0 Keosone A, 2 . 0 Parazone, 1 . 0 Permalux, 2 . 0

..... .....

.....
Reogen, 2 . 0

.....
....
L.P. s o . 2 . 2 0 . 0 Dibutyl sebacate, 2 0 . 0 Circo X Oil, 1 0 . 0 Dibutyl sebacate, 1 7 . 5 Tributoxyethylphosphate? 1 7 . 5 S.C., 15.0 tributoxyethylphosphate, 1 5 . 0 S.C.,1 5 . 0 Tributoxyethylphosphate. 15 .O Tribu(oxyethy1 phosphate, 25

5.0
10.0

1.0
0.5 0.5 1.0 1.0 0.5 0.5 0.5 0.5

100
80 100 100 100 100 100 100

..
1.5 2.0 2.0
1.25 1.25

..

1.03

5.0 5.0 5.0

5.0 5.0

0.5

Methyl simate, 0 . 1 Age Rite Hipar, 1 . 0 Rotox, 1. O D.P.G., 0 . 1 ..... Altax, 0 . 3 G.M.F., 1 . 5 ........... Altax, 1 . 0 Neoeone D , 2 . 0 Tuads, 0 . 5 ..... Altax, 1 . 0 ..... Altax, 1 . 5 Altax, 1 . 5 Altax, 1 . 5 Tetrone A, 0 . 7 5 Altax, 1 . 0 Bardol, 5 . 0 Thionex 0 . 5 Altax. lY0 Thionex, 0 . 5 Bardol, 5 . 0 Thionex. 0 . 5 E t h y l zimate. 0 . 1 Altax, 1 . 5

1.00
0.98

5.0

.....

.. .. ..

2.0
2.5 2.0 2.0 2.0 1.5

..... .....
Age Rite 'rk&, 1.O 4 g e Rite resin D, 1 . 0 Age Rite resin D , 1 . 0

100 100
100 100 100

.. .. ..

3 .O 5.0

i.0
1.0

.....

5.0
5.0
5.0

1.0

1.0

.....

.....
Triethyl phosphate, 15.0

.....

Basic oxides in this column are curing agents.

V O L U M E 20, NO. 1, J A N U A R Y 1 9 4 8
Table 11. (Contd.) Summary of Investigations of Compounded and Cured Rubber
Preparationa Filler Composiseparationb SolventC tiond F-407 Tetrachloroethane F F-407 Cyclohexane F F-407 Diisobutyl ketone F F-407 Pyridine F F-407 o-Dichlorobenzene F F-407 o-Dichlorobenzene F F-402 Tetrachloroethane F F-402 Diisobutyl ketone F F-402 o-Dichlorobenzene F F-416 o-Dichlorobenzene F F-1302 Tetrachloroethane F F-1302 p-Cymene F F-1302 o-Dichlorobenzene F F-1302 F-1302 F-205 F-205 F-804 F-804 F-804 F-801 1 2 3 4 5 F-803 F-803 F-803 F-802 F-803 Pyridine o-Dichlorobenzene Tetrachloroethane o-Dichlorobenzene Tetrachloroethane p-Cymene o-Dichlorobenzene o-Dichlorobenzene Tetrachloroethane Diisobutyl ketone p-Cymene . Pyridine o-Dichlorobenzene Solution characteristics' N Dark, U Dark, U Dark, P hmber, P Amber. P N Dark, U Amber. P Amber,' P N D?rk, U Amber P Dark P Amber, P N Amber, P Red, W Red, W Red, W Red, W Spectral S t u d y Wavelength Soectral range, chsractermicrons isticso 2-14 2-14 2-14 21-14 2-14 2-14 2-14 2-14 2-14 14-22 4- J
2-14

15

Table 11.
Sample

Elastomer class Butadiene-acrylonitrile copolymer ( c o n t d . )

Representative type Hycar-OR-15

Hycar-OR-25

Perbunan-35

Sample No. 1 2 3 4 5 6 1 2 3 4 1 2 3

Film

No./

D 194 212 339 339 401

1)

...

--I4

Butaprene-NF Ethylene polysulfide polymers Thiokol-ST Copolymer of butadiene, acrylonitrile and unsaturated ketone Thiokol-RD

F F F F F F
F F F

:97 428 457 D 282 375 437 427 D 394 186 29 1 403 458 438 D 231 269 369

...

E E
E

X X

..

X E X E
h

..

...

..

...

2-14 2-14 2-14 2-14 2-14 14-22 2-14 2-14 4- 5 2-14

...

E
A

E E E E

..

N Dark, U Dark, P Dark, P Amber, P

...

..

...

A E

refluxing, was satisfactorily inert toward the rubber hydrocarbon. p-Cymene dissolved all samples except Buna-N types, but its use a t prolonged refluxing led t o oxidation and serious gum formation which contaminated the sample. Pyridine and nitropropane were ineffective for Buna-r\', and several high-boiling ketones precipitated gums in the sample during refluxing. A 4-gram sample and 250 ml. of solvent were refluxed for 24 to 36 hours in an A.S.T.M. condenser and flask assembly or in a similar all-glass assembly, using a water-cooled condenser (Figure 3) designed for this purpose. The glass condenser can be used interchangeably with the A.S.T.M. tin-coil condenser and is pre-

ferred for chemical inertness and ease of cleaning. Complete solution of some of the less soluble elastomers was best accomplished by successive treatment with two or three 100-ml. portions of solvent. Blenderization decreased the time required for solution, but was not essential since serious gelling never occurred. Absence of appreciable residue after an hour of settling indicated complete solution of the elastomer. Comparatively coarse fillers such as clays, iron oxide, etc., were separated by simple centrifugation. This method was ineffective for colloidal carbon blacks. The procedure used by Barnes ( d ) , employing dilution with n-hexane, was tested on a series of natural

of Compounded Specimens
Inhibitors, Waxes Extenders Carbon Blacks Parts b y weight Basic Oxidesa, Clays Total P a r t s by Weight 113.1 112.5 238.5 243.5 222.7 151.0 220.5 150.0 113.1 150.9 162.5 170.5 168.5 273.5 208.25 208.25 279.75 181.5 164.5 110.5 164.5 Physical Measurements Cure Min. 15, 300' F. 30, 300' F. 40, 300' F. 40, 3 0 0 ' F.

Sp. Gr.

Dur. 30-40 35-40 65 70 60-65 45 45-50 45

.....
Heliozone, 2 . 0

. .. . . ..
P-33, 0 . 5 Thermax, 100 Thermax, 100 P-33, 91 P-33, 30

.....
L.C. m a g n d i u m oxide, 10.0
L.C. magnesium oxide, 1 0 . 0 L.C. magnesium oxide, 4 . 0
I . . . .

..
0.97

TensileI Lb. 3200 2125 1800 1875 1225 1800 1550 1500

Elonnation

%
725 775 300 300 650
825

.....

..

.....
Acrawax, 2 . 0

..
1.39

30, 300' F.
45, 300" F. 45, 310 F.

.....
Paraffin, 3 . 0 Heliozone, 1, O Mineral rubber, 7 . 5

..
1.34

.....
Furnex, 1 6 . 0 Nicronex, 1 8 . 0

Superex clay

785 750

.....
....
I

.....

Furnex, 40 Furnex, 60 Furnex, 40 Furnex, 50 Thermax, 130 Thermax, 70 Furnex. 70 Wyex, 35 Furnex. 35 P-33.50 Btatex B, 50

.....
..... .....
Selastic SR-66, 100

..... .....

.....

..... ..... ..... ..... .....

.....

07, 375' F.
20, 300 F.

.. ..
1.52

...
60-65 65 55 65 60 45-50 45-50 75-80 48 60

..
665 1300 1375 1600 1675 1700 2000 950 600 3200

..
400 325 300 260 610

40, 300' F. 30, 290' 15, 300' 30, 300' 15, 300'

.....

F. F. F.

F.

... . .. ..
1.22 1.20 1.43 i:i3

..... .....
.....

30, 300' F. 30, 300' F. 30, 300 F. 50, 300' F. 15, 300' F.
15, 300" F.

700
550 125 300 500

..... ..... ..... ..... .....

..... ..... .....

.....
Btatex B, 50 Furnex, 50

15, 300 F.
30, 300 F.

174.1

.. .. ..

... ... ...

..
..

.. ..

..

16
and Buna-N samples and found unreliable because of selective precipitation of the elastomer. Filtration with adsorbents such as bentonite, fuller's earth, alumina, and silica gel was also unsatisfactory. Celite analytical filter aid (Johns-1Llanville Co.), however, permitted complete removal of colloidal carbon in every case investigated. After the polymer solution had been diluted to 300 ml. with benzene, about 50 t o 100 cc. of Celite were added and the slurry was filtered through a semifine paper. Filtration proceeds most rapidly if sufficient Celite is added so that free carbon does not clog the paper. Trial filtrations with small portions of the slurry readily allow a proper Celite content t o be found. Quantitative work with solutions of mixed elastomers conclusively showed no selectivity toward the elastomers and the solution holdup in the filter aid can be nearly completely recovered by adding hot benzene and refiltering.

ANALYTICAL CHEMISTRY
adequate evidence of the validity of all spectra obtained! In addition, almost perfect spectral reproducibility was found for the spectra of each elastomer obtained from duplicate preparations in o-dichlorobenzene and in most cases from additional preparations in p-cymene and in tetrachloroethane. The overall method was thus shown to be reliable and the spectral results obtained were shown to be nearly independent of the sample history-Le., of compounding and curing effects and of the method of preparation for study. The complete spectral identity of cured and uncured rubbers is not inconsistent with current concepts of the vulcanization process, which is considered to introduce comparatively few cross linkages in building the three-dimensional vulcanizate. Since the infrared spectrum of a polymer is determined almost entirely by the small recurring structural units and not by the gross size or configuration, the net result of vulcanization, consisting mainly of changes which involve relatively few of the recurring polymer units, does not alter the rubber spectrum to any appreciable extent. Only two minor differences between the spectra of gums and their vulcanizates (Figures 2 and 4) were noticed. In general, oxidation during vulcanization produces hydroxyl absorption near 3.0 mu and carbonyl absorption between 5.6 and 6.0 mu. These extra bands do not interfere, hon-ever, with any bands useful for analysis. A more consistent difference is the poorer definition of bands themselves and the increased background absorption beyond 8 mu in the spectra of most cured rubber. This effect has been observed in homologous series of simple hydrocarbons and is consistent with the increased molecular weight of the vulcanizate. This absorption was not reproducible for different sample preparations and, though serious only in natural vulcanizates, it cannot be disregarded in quantitative work. The method of study leaves little to be desired from the spectroscopist's viewpoint, since isolation of the rubber hyd1ocarbon assures valid results. The method applies to all elastomers investigated and t o blends of these elastomers, and a single procedure remarkably free of exacting techniques may be used. Routine application requires about 6 man-hours per sample.
IDENTIFICATION AND QUALITATIVE ANALYSIS

The fact that the spectra of the various elastomers, when obtained by the procedures described, are unique and experimentally reproducible regardless of the past history of the sample constitutes a sound basis for characterization of the elastomer content of commercial rubber products. The individual elastomers are identified by the following absorptions:

u
Figure 3.

All-Glass Reflux Assembly

The elastomer solutions were concentrated by vacuum distillation at 100' t o 110' c.,using a water aspirator and conventional Claisen-type flask immersed in an oil bath. The viscous concentrate was diluted with a few milliliters of benzene and transferred t o a small container for storage; the benzene was removed later by evaporation a t room temperature or in a vacuum oven a t 60" C. Films for spectral examination were spread as before and 40 minutes in a vacuum oven a t 80' C. were allowed for final solvent evaporation. Films of natural and Thiokol rubber were sticky but satisfactorily rigid and those of other elastomers were solid and nearly free of tackiness.
Discussion. The work on cured specimens is summarized in Table 11, and representative spectra of the cured elastomers are shown in Figure 4. Comparison with Figure 2 shows the spectra of a pure gum elastomer and its vulcanizate t o be nearly identical, except for minor differences which are discussed below. This significant result was obtained consistently throughout more than fifty spectra of separate sample preparations and alone is

Natural and neoprene are distinguished from all other classes by a broad band a t 12 mu, and are differentiated from one another by the exact position of this band, by the 7.25 mu band in natural, by the strength and position of the 6.0 mu band, and by the appearance of the 9 mu region. Butyl (GR-I) is distinguished by a strong unique band a t 8.13 mu and by a strong doublet a t 7.20, 7.30 mu. Buna-N (Chemigum, Hycar, Perbunan) and Buna-S (GR-S) are characterized by strong bands a t 10.3 and 10.9 mu (common t o butadiene polymers), and Buna-S is readily identified by unique bands at 6.23, 6.70, and 14.3 mu, while BunaN is identified by a band a t 4.47 mu. Thiokol-FA and -ST possess identical spectra, characterized by a unique envelope of strong bands a t 8.36, 8.64, 8.98, 9.32, and 9.68 mu. ThiokoLRD possesses typical Buna-N bands a t 4.47, 10.25, and 10.86 mu, but is readily set apart by a strong band a t 5.86 mu and by a doublet at 7.25, 7.38 mu. Thus, identification of a rubber specimen containing a single elastomer is always possible simply by empirical comparison of the spectrum with reference spectra, such as those of Figures 2 and 4. For mixtures of elastomers the ease of qualitative analysis depends upon the degree of differences existing in the spectra of the components. For example, less than 1%of a nitrile (BunaN ) copolymer may be identified in any mixture by absorption a t

V O L U M E 20, NO. 1, J A N U A R Y 1 9 4 8
I

17
sorption a t 10.36 mu, while the latter are predominantly 1,2addition, with absorption a t 10.95 mu.
QUANTITATIVE ANALYSIS OF NATURAL-GR-S MIXTURES

4.47 mu, since all other elastomers are highly transparent in this region. Less than 5% of Thiokol-FA may be detected in any mixture by its characteristic bands in the region of 7.5 to 10.5 mu, since these bands, though not isolated, are much stronger than those of other elastomers in this region. Since blends of more than two components are seldom encountered in commercial products, it is generally possible to identify any component which exists in 6%, or greater, concentration. The limit of detection of neoprene in natural, however, is probably nearer 10yo, since great spectral similarity exists for these elastomers. The major component(s) is best identified by examining a film which transmits about 10 to 20% a t 6.8 mu. Thicker films may then be examined over selected wave-length regions to establish the presence or absence of minor components. Exact spectral correspondence may not always be found for some elastomers because of differences in component proportions and modes of addition in copolymers or oxidation of the sample. These differences do not hinder the analysis and, in addition, furnish valuable information if interpreted in terms of molecular structure. Thus, the relative intensity of the 4.47 mu nitrile absorption distinguishes betn een high- and lownitrile Buna-Pi; and emulsion polymers may be differentiated from sodium polymers since the former are predominantly 1,4-addition, with ab-

WAVELENGTH

(IN

MICRONS)+
8

1011

12131415
1 0 0

50

Basis of Method. Quantitative spectral analysis of a binary mixture generally involves (1) selection of a spectroscopic measurement or function which is sensitive to the composition, (2) construction of a working curve expressing the relationship between the spectroscopic function and sample composition for a set of accurately known mixtures, and (3) subsequent analysis of samples by applying analogous measurements to this working curve. The working curve method requires the utmost in experimental reproducibility of the over-all procedure, and the results obtained with uncured and cured elastomers show the methods presented to be fully adequate in this respect. The analytical problem relates, therefore, largely to determining t,he most satisfactory spectroscopic function for the particular analysis. The case of natural-GR-S mixtures is considered here, but since similar spectral differences exist for other binary mixtures analogous procedures and similar results may be obtained in other cases. Comparison of the spect,ra of natural and GR-S indicates that the former can be detected in the mixture by methyl absorption at 7.25 mu or by 11.95 mu absorption due to the isoprene structure, -(CH,)C=CH-, and that the latter can be detected by the 6.70 mu phenyl absorption or by butadiene absorptions a t 30.35 and 10.95 mu due to 1,4- and 1,2-addition, respectively. The 7.25 and 6.70 mu bands appear most suitable for natural and GR-S determination, since both are independent of olefinic bonds which might be expected to change under certain conditions-for example, with vulcanization, molecular weight, and mode of addition. Since the styrene content of GR-S is maintained as nearly constant as possible by all manufacturers, the number of phenyl and methyl groups present, and hence the intensities of their absorptions, are proportional to the GR-S and natural content, respectively, of the mixture. The use of the shorter wave-length bands avoids any effects due to variable background absorption a t longer wave lengths. The spectroscopic function to be employed generally includes the transmittance, T, of the sample as the experimentally measured quantity. This is usually expressed in terms of extinction, E, by the equations:

Figure 4.

Spectra of Cured Elastomers

where I o and I represent the energy transmitted by the blank and sample, respectively, K is a proportionality constant, C is the concentration, and t is the sample thickness. Since it is impractical, if not impossible, to measure or control the thickness of rubber films with sufficient accuracy, it is convenient to eliminate t from the spectroscopic function. This can be done by employing the ratio of extinctions determined at different wave lengths, since only t is independent of wave length. It would be convenient, then, to employ as the spectroscopic function the ratio E7.25/E8.70 There the superscripts refer to wave length. I n practice, however, this ratio can be determined accurately only for mixtures which contain appreciable concentrations of both components, for only under these conditions is it possible t o spread a film of given thickness which yields transmittance values in the accurately measurablc range of 20 t o 65% a t each wave length. This difficulty may be resolved by selecting, as an internal standard, a third band common t o both components and of such intensity that the transmittance, both a t the internal standard wave length and at the analytical wave length of the minor component, will always fall within the accurately measurable range. For this purpose the common band of medium strength a t 7.60 mu is especially well suited, since it occurs in both natural and GR-S a t so nearly the same intensity as to be almost

18

ANALYTICAL CHEMISTRY
the measured extinction ratios against the known composition of each sample, and applying the corresponding values to the working curve (Figure 5 ) drawn through the average value of each group of points. For each blend the several analyses differ from the average only rarely,by more than 2% and the average deviation for the 58 results is 1%. The reproducibility of the analytical values for the three separate preparations of each sample is equal to that obtained for duplicate films of any one preparation. This proves conclusively the adequacy of the preparation method on the one hand, and shows the importance of the quality of the films on the other. While some films were perfectly smooth, others were somewhat uneven, and this probably accounts for what little variation was obtained. For evaluation of the effect of film thickness, a number of films of greatly different thickness were purposely studied, although fairly close control of thickness %-as easily possible.

independent of the composition of the blend. The extinction ratios E6.70/E1.60 and E7.26/E73Q thus may be used as functions only of composition to determine GR-S and natural, respectively, and together enable analysis of all mixtures from 0 to 100%. Both ratios are linear with composition over the useful range (see Figure 5). Analytical Procedure. The following method is presented for routine analysis of finished rubber articles composed of natural rubber and GR-S. The procedures are in no way exacting, and the time requirement is about 6 man-hours per sample.

hIEASURENENTS AND C A L C U LATIOXS. Percentage trans-

SAMPLE PREPARATIOS. h 4-gram sample, which has been milled to pass a S o . 20 screen, is extracted for 24 hours Kith tn-o 80-ml. portions of a 32-68 blend of acetone-chloroform. The extracted sample is completely dissolved in 250 ml. of distilled o-dichlorobenzene by refluxing for 24 hours, and the solution is diluted to 300 ml. with benzene. Carbon blacks are removed bv Celite filtration. and mineral fillers, if present, by centrifugati"on. A thick mucilage for film spreading is obtained by vacuum distillation of the solvent, and films are spread Table IV. Lnalysis of Standard Natural-GR-S Blends on sodium chloride or potasNatural GR-S sium bromide plates. If the Analysis C Analysisd sample is fairly viscous, the Samplea Measurements 6 7.25 6.70 films can be checked for thickBlend NO. 7.60 mu 7.25 mu 6.70 mu 7.60 % 7.60 70 ness immediately (since oF-620 (100% GR-S) 1 00.0 1.68 36.0/0.445 52,0/0.285 18.0/0.745 0.64 dichlorobenzene is transparent 2 0 . 6 3 0 0 . 0 1 . 7 6 52.0/0.283 66.5/0.178 32.0/0.498 a t this wave length) by 1 (90% GR-S) (10% natural) 35.0/0.456 20.0/0.700 10.5 1.54 e 32.5/0.488 1.07 1 measurement a t 7.60 mu 2 22.5/0.647 21.5/0.667 12.3jO.910 1 . 0 3 9.5 1.41 8 where the transmittance should 3a 39.5/0.404 37.0/0.432 29.0/0.542 1.07 10.5 1.34 e 3 b 47.0j0.328 1 . 0 7 1 0 . 5 1 . 3 6 e 49.5/0 306 38.0/0.417 be between 40and 60%. After 45.0/0.349 42.0/0.377 3c 33.5j0.474 1.08 10.7 1.36 e adjustment to an appropriate 3d 1 . 0 8 1 0 . 7 1.37 3 2 . 0 j 0 . 4 9 2 29.0/0.635 21.0/0.672 e thickness the film is dried 40 2 (80% GR-S) (20% natural) 1 30.0/0.523 1 7 . 5 / 0 . 7 5 5 20.0/0.700 1 . 4 4 20.0 1 . 3 4 minutes in a vacuum oven a t 34.0/0.470 2 1.49 21.5 1.42 20.0/0.700 2 1 . 5 j 0 . 6 6 7 80" C., or for 4 hours a t room 1.39 1.33 18.0/0.742 19.0 3a 28.0/0.656 17.0/0.774 3b 38.0/0.424 1 . 4 1 1 ,35 2 7 . 0 / 0 . 5 7 3 1 9 . 5 25.0/0.603 temperature before analytical 1 . 4 0 1 .35 1 9 . 0 49.5/0.305 39.0/0.409 3 7 . 5 / 0 . 4 2 8 3c m e a s u r e m e n t s a r e made. 1.46 2 0 . 5 5Z.Oj0.284 7d 62.5/0.zn6 5 0 . 0 / 0 . 3 0 1 I .38 Films of decided irregularity 1 38.0/0.420 16.3/0.788 3 (70% GR-S) (30% natural) 29.0/0.538 1.87 31.0 1.28 72.0 must not be used. 2 22.5/0.647 15.7/0.804 1.24 69.0

mittance of each film is measured a t 7.60, 7.25, and 6.70 mu. Care must be taken to measure a t the center of the bands. Duplicate films and measurements are recommended for accuracy, and transmittance values outside the range 20 to 65'3 are not used. The extinction ratio E'. ZS/E7.60 or E6.iO/E7.60 (01. both) is obtained, depending upon the suitability of the measurements, and plotted against composition for a series of standard mixtures. Subsequent samples are analyzed by the converse procedure of applying the measured ratios to the appropriate working curve. Experimental Results.
\\-OR[< KITH STANDARD BLEKDS.

4 (607, GR-S) (40% natural)

5 (50% GR-S) (50% natural)

6 (407, GR-S) (607, natural)

7 1 (307, GR-S;) (70% natural)

8 (20% GR-S) (80% natural)

The analytical method was tested by applying it to eleven standard samples of accurately Thclse known composition. were made by curing appropriate mixtures of two master batches, natural (F-10) and GR-S (F-620) (see Table III), compounded with 30% carbon black as filler. Samples werc prepared in triplicate, and several films of each preparation were studied. The analytical results, presented in Table IV, were obtained by plotting

9 (10% GR-S) (90% natural)

3a 3b 3c 3d 1 2 3a 3h 30 3d 1 2 3a 3b 3c 3d 3e 1 2 3a 3b 3c 3d 3a 3b 3c 3d 4a 4b 1 2 3a 3b 3c 3d 1
2

40.5/0.394 50.0/0.301 35.0/0.458 46.Oj0.339 29 510 530 38.5/0.415 48.0/0.320 36 5 / 0 438 70 5 , O . 152 39 0/0 408 32 . r / O ,495 25 ;/0,595 31.5/0.514 27.?/0.558 87.a'O.242 59.5/0.224 41.0/0.384

20.0/0.701 27.0/0.567 14.5/0.842 23.0j0.642

31.5/0.500 41.0/0.388
-

20.0/0.703 46.0/0.339 -

37.0/0.428 25.0/0.603 34.0/0.470 31 . O / O . 512 67.0j0.174 35,0/0.456 33.0/0.482 26.0/0.587 31.0i0.506 26.5/0,576 57.0/0.244 58,510,232 41.5/0.382 24.5/0.610 45.5/0.340 35.0/0.456 43.010.368 46 .O '0.335 3; 0'0.428 4 3 , 5,'O ,342 40.5/0.393 32.0;0.492 39.0;0.410 42.0'0.378 65.5lO.184 43 0 0 365 44 O f 0 355 43 5 0 361 47 .5/0 324 51.510 286 47 0/0 330 50.5/0.297 6 0 , .5/0.218 64.5/0.191 60.0/0.22? 67.U0.17~ 3 3 . 5 0.273 49.5/0.305
53.0/0.260 65.5/0.!85

1.78 1.88 1.84 1.89

29.0 31.5 30.5 31.7

1.27 1.29

71.5 73.0

1.24 1.13 1.13

69.0 60.0 60.0

26.0j0.584 27.0/0.566

2.40 2.53

45 0 48 0

1.17 1,14 1.12 0.975 0.990

0.980

1 03

1.03 1.01

1.00
0.885 0.88L5 0,900 0.890 0.890 0,875 0.770 0.795

20.5/0.690 41.5/0.380 31.0j0.506 38.5/0.413 42.0/0.376 32.5/0.490 36.0/0.444 32.0/0.494 23.0/0.642 29.5/0.532 33.5j0.472 59.0j0.232
25.5jO.595

3a 3b 3c 311 3e F-10 (100% natural)

a b
C

1 2

27.010.570 27,5/0,558 32.0/0.494 34.0I0.4515 30.0/0.524 25.5/0.595 38.010.420 41.0'0.387 3 7 . 5,.0.428 45,010,346 2 8 . 0 / 0 349 24.0/0.620 22.5j0.647 36.0j0.445

0.765 0,770 0 800 0.790

0.610 0.620 0.650
0.635

0.61: 0 630 0.500 0.520 0.495 0.515 0.495

0.500 0 490

63.0 61.0 59.0 47.5 48.5 48.0 52.0 50.0 52.0 50.5 40.0 40.0 41.0 40.5 40.5 39.5 31.0 32.5 30.0 31.0 33.0 32.0 18.0 18.5 21.0 21.5 18.2 19.5 09.5 10.5
08.5

10.0
08.3

08.0

09.5

0.400 0.415

00.5 02.0

Sumbers refer to sample preparation: letters indicate several films measured. Percentage transmittance, T,and extinction, E,a t each Wave length given as T / E . Using indicated ratio of extinction values and curve 2 of Figure 4. d Using indicated ratio of extinction values and curve 1 of Figure 4. e Extinction ratios (though measurable) beyond linear range of working curve. - Absorption too great for accurate measurement. I Preparations 1and 2 of blend 7 ruined prior to film preparation.

V O L U M E 20, NO. 1, J A N U A R Y 1 9 4 8
Despite threefold differences in the 7.60 mu extinction, the analytical results proved without exception to be independent of this factor. Finally, the reproducibility is the same for all concentrations on the linear portion of the working curve. In vierr of the facts that these results include all data obtained in development of the method, and that no exceptional degree of regulation was used in any steps of the analysis, it is considered that a quantitative accuracy expressed by an average deviation of 1% is generally obtainable in application of the method to mistures of high-quality natural-GR-S stocks. ANALYSIS OF RECLAIM-CONTAINISG MIXTURES.Analysis of the standard samples described above necessarily constituted favorable conditions for application of the method, in order that limitations pertaining solely to the experimental procedure could be established. The general usefulness of the

19

method, however, depends upon the consistency of the analytical results when stocks of different origin are encountered. Although different stocks will exhibit the same general spectral features, slight quantitative differences in polymer structure may exist and produce corresponding differences in band intensities which \Till affect the analytical results. To determine the maximum effect which would be encountered, attention 17-as directed toward natural-GR-S blends containing a natural reclaim stock. A different GR-S stock was not studied, as little variation of GR-S mould be expected. Six reclaimcontaining blends, made by curing appropriate mixtures of natural (F-lo), GR-S (F-620), and a tube reclaim (F-10-l), were examined. Several preparations and analyses of each of these samples were made using the working curves previously determined (Table V). -4lthough no apparent difficulties were encountered in sample preparation, both the reproducibility and the 1.401 , correctness of the results are much inferior to the accuracy preI I ' I I I 1 1 . 1 I I 2 80 viously reported. The errors are roughly proportional to the re2.60 claim content, and far greater interference is found in the natural 2.40 I20 ( ~ l u reclaim) s determination than in the GR-S determination. For GR-S concentrations of 10 to 70y0 an accuracy of *EI% or I IO 2 20 better can be obtained by using the GR-S working curve and de100 2 00 termining natural by difference. For other compositions as much as 10% error may be obtained. However, if it is assumed that 0 90 I80 reclaim is allvays used with a t least equal quantities of natural 160 rubber, the over-all accuracy of the method may be put a t * 5 7 0 for reclaim-containing mixtures. The seriousness of interference by reclaini is minimized somewhat by the fact that the presence of reclaim can be recognized in important cases. In the absence of reclaim both natural and GR-S can be measured directly only for concentration near 507& and in these cases the concentrations total 1 0 0 ~ within o the limits of error. However, when reclaim is present in sufficient quantity to cause serious interference it becomes possible to determine both components over 1 ' . N A W R A L + 7 0 60 50 40 30 20 IO o a much wider range, and in every case the interference is indicated by a corresponding deviation of total concentration from Figure 5. Calibration for Natural-GR-S Analysis 100%. Although the results obtained with reclaim-containing blends Table 7.'. Analysis of Reclaim-Containing 3lixtures were not unreasonably erroneGR-S Katural ous considering the severity of Bnalyslsd Analysis C Samplea 6 70 Measurements b 7.2;7 the test, an attempt to obtain 7.60 mu 7 . 2 5 mu 6.70 mu 7.60 "0 7 60 % XO. Blend more satisfactory results was F-10 (lOO'?o natural) la 36.010.442 - e 66.3/0,179 0.405 01.0 made by employing a different lb 57.5/0.241 13y0 80% 2a - 0.4 39.5/0.404 68.0/0.168 choice of analytical wave -1 7 0 2 -. 0 1370 80 70 2h 57,5/0.241 lengths. Results based upon F-10-4 (50% natural) (50% la 32.2/0.492 $2.0/0.209 - 0.425 0 2 . 5 measurements a t the 11.95 mu 55.310.258 18.2/0.739 17.4/0.112 2.86 i 0.435 03.5 reclaim) lb 37,810,422 64.3/0.192 0.455 0 5 , o 2a absorption of natural were 2b 51.7/0.286 13% 74.0/0.132 - 0,460 0 5 . 5 found unreliable and other pro40.0/0.398 67.0/0.175 0,440 04.0 F-10-1 (lOOi-o reclaim) la lb 58.2/0.232 24.3/0.614 78.0/0.108 2.64 54.0 0.465 06.0 cedures employing the func2a 33.910.470 53,310,273 0.580 15.0 tions E10.36/E10.10 ElO.3S/E11.40, 2b 48.8/0.311 22.2'0.653 64.5/0.191 2.10 36 0 0.615 18.0 3 45.5/0.342 14.0/0.844 65.0/0.187 2.46 47.0 0.550 13.0 and E10.95/E11.40 yielded satis4 48.8/0.311 15.0/0.824 66.5/0.177 2.64 54.0 0.570 14.5 factory results for high-quality F-10-2 (33.37, natural) (33.3% l a 31.5/0.499 41.5/0.383 0,770 30.5 lb 54.0/0.266 25.30.594 64,510,197 2.23 40.0 0.740 28.5 reclaim) (33.3'3 GR-S) stocks but no improvement in 2a 34.0/0.470 41.8/0.379 - 0.805 33.5 results for reclaim-containing 2b 51.0/0.292 18.3/0.737 58,5/0.234 2.52 49.0 0.800 33.0 3 48.8/0.311 15.5/0.808 55,5/0.256 2.60 52.0 0.825 35.5 mixtures. The original method F-10-5 (25'3 natural) (257, la 63.0/0.200 37.2/0.430 64.5/0.190 2.15 0 950 45.5 38.5 therefore appears to yield the lb 44.2/0.354 reclaim) (50% GR-S) 17.8/0.750 46.5/0.342 46.5 2.12 3 8 . 0 0.965 43.0 best accuracy for infrared la F-10-3 (50% reclaim) (50% 3 4 , 0 / 0 ,470 14.0,/0.860 37.0/0 ,432 1.83 3 0 . 0 0.920 GR-S) lb 53.0/0.276 29.8/0.525 55.8/0.253 1.90 32.0 0.920 43.0 analysis of natural-GR-S 2 49.4/0.306 26.210.582 50.5/0.296 1.90 32.0 0.970 47.0 3 4 7 . 8 1 0 . 3 2 1 22.2/0.653 46.7/0.331 1.03 50.5 2.03 35.0 mixtures.
I
~~

f-10-6, (15R, natural) (15y0 reclaim) ( i o % GR-S) F-620 ( 1 0 0 7 GR-S)

la 59.2/0.228 41.0/0.386 52.2/0.282 1.69 lb 45.5/0.343 26.5/0.575 37.2/0.430 1.68 la 37.210.430 49.210.308 23.2/0.634 0.715 lh 50,510,296 60.5/0.219 31.010.431 0.740 2 50.0/0.301 57.8/0.238 34.0/0.470 0.790 a S u m b e r s refer t o sample preparations. letters indicate several films measured. b Percentage transmittance T, and extinction E a t each wave length given a s T/E. C Using indicated ratio of ex'tinotion values a n h cbrve 2 of Figure 4. d Using indicated ratio of extinction values a n d curve 1 of Figure 4. e Absorption too great for accurate measurement. f Extinction ratios (though measurable) beyond linear range of working curve.

26.5 26.5 01.0 01.5 03.0

1.24 1.25 1.47 1.45 1.56

69.0 70.0
/ / /

DETERMINATION OF ACRY LONITRILE IN BUNA-N COPOLYMERS

Basis of Method. The general procedure described for the qpectral analysis of elaqtonicr mixtures may also be used to

20

ANALYTICAL CHEMISTRY
The extracted sample is dissolved in 300 ml. of distilled o-dichlorobenzene under reflux, the solvent being used successively in two portions, and blenderization being employed if necessary. Carbon blacks are removed by Celite filtration, and mineral fillers by centrifugation. Vacuum distillation is employed for solvent removal. Films are spread on sodium chloride or potassium bromide plates and residual solvent is removed by evaporation at room temperature or more rapidly in a vacuum oven at 80" C. SPECTROSCOPIC MEASURERIEXTS. Percentage transmittance of each film relative to a blank plate is measured at 4.60, 4.47, and 3.50 mu, care being taken t o locate the maxima of the bands (except at 4.60 mu). Films of satisfactory thickness give about 20 t o 25% transmittance at 3.50 mu. The percentage transmittances at 3.50 and 4.47 mu are corrected for film scattering by dividing by the decimal transmittance at 4.60 mu and the extinction ratio E(4.47)/E(3.50) is calculated, parentheses being used to denote the use of corrected transmittance values. This ratio is plotted versus composition for several known samples and subsequent samples are analyzed by the converse procedure of applying the measured ratios to this curve. Films and measurements are made in duplicate as a check on accuracy.

8
\

U 2
v) 0 trj \
)c

ACRYLONITRILE

-+

Experimental Results. The analyrical method was tested by applying it to over fifty samples prepared from a variety of Buna-N determine the structural units comprising a given elastomer. Decopolvniers. Preliminary measurements were made on gum termination of the nitrile (nitrile and acrylonitrile are used intersamples of Perbunan-18, -26, and -35, Chemigum-S-1, -2, and -3, and Hycar-OR-15 and -25. A working curve was conchangeably throughout the discussion) content of Buna-T\' copolymers is considered in detail, but similar methods may structed using the average values of the uncorrected analytical readily be developed for analogous problems, such as determinafunction and nitrile values 1% ere determined by Kjeldahl analysis tion of the styrene content of Buna-S or the ratio of 1,2- to 1,4(total nitrogen assumed to be nitrile nitrogen). All values nere addition structures in polvburadienes. then applied to the working curve and the results are presented Satisfactory procedures for sample preparation of cured and as set 1 in Table VI. Despite neglect of the scattering correction uncured Buna-N copolymers have been presented, so that only and the use of films of widely different quality and thickness, the the spectroscopic function need be considered. The ratio E4.47/- analytical results shoTy a mean deviation of + 1 and a maxiE3.60 obtained by measuring the transmittance at the 4.47 mu mum error only rarely twice this value. The error is in no sense consistent with film thickness, and can be attributed entirely to nitrile band (due to the C=N valence vibration) and a t the 3.50 mu C-H absorption, is a satisfactory function and can be deterdifferences in film quality. The over-all method was then critically tested by using the same working curve to analyze a set of mined accurately for samples containing 0 to 50% nitrile, a hich range includes all Buna-S rubbers of industrial importance. A the same Buna-N stocks V , hich had been compounded a ith carbon working curve is constructed by plotting the spectroscopic black and cured. The agreements of the results (set 4, Table function versus composition for a series of standard samples VI) Jyith those for gum samples are fully as close as the reproduciand subsequent analyses are made by applying analogous measbility obtained in either set for duplicate films of the same prepaurements to the curve. ration. This severe test of the over-all method shows its complete If the transmittance is measured relative to a blank plate a t reliability on the one hand and the importance of obtaining highthe analytical wave lengths, an error due to reflection and scatquality films on the other. tering losses in the sample film will result and the calibration For extension of the calibration below 207' nitiile a specially curve d l not read zero at 0% nitrile (pure polybutadiene). Alprepared set of ten Buna-K gum elastomers (supplied by the Firethough this effect is scarcely noticeable for nitrile concentrations stone Rubber Company) covering the range 0 to 407' nitrile it gives appreciable positive error at low nitrile conwere prepared in duplicate and several films of each preparation above 15y0, centrations. This error may be eliminated entirely by choosing 1Tei-e measured at 3.50, 4.47, and 4.60 mu. Extinction ratios 1 5 ere calculated both with and without correction for scattering as t ' l O O ~ o transmittance" the transmittance of the sample film itself a t a near-by wave length of maximum transmittance-Le., and the average values for each sample 1%-ere plotted versus nitrile values based upon the manufacturer's Kjeldahl analyses to obtain at 4.60 mu-rather than the transmittance of a blank plate at the analytical wave length. With this modification polybuthe norking curves sho7T-n in Figure 6. Curve 1 (uncorrected ratios) TTas identical Ivith that previously obtained for the tadiene will show 0% absorption at 4.47 mu, and valid results IT ill be obtained for low nitrile samples, irrespective of wide variacommercial elastomers and established the equivalence both of tions in general transparency of films. the spectral results and of the chemical analyses for both series of samples. The analytical results obtained by applying Analytical Procedure. SAMPLE PREPARATIOX. For gum the ratios for each film to the proper curve of Figure rubber, a %gram sample is cut into thin Strips and extracted 6 are presented in Table VII. F~~ each curve an average for 24 hours with two 80-ml. portions of 95yc ethanol The exdeviation of * 1% 1Tas obtained for samples containing 15% or tracted sample is alloxyed to softenand ell in 100 of ethylene more nitrile; for samples of less than 15% nitrile no decrease dichloride for 2 hours and then thoroughly mixed in a \Taiing Blendor. Excess solvent is removed by evaporation on an in accuracy resulted when the corrected ratio was used (curve 2), A.S.T.11. gum bath at 70" C., but evaporation must be stopped whereas the uncorrected ratio gave considerable error. The errors before gelling occurs. For cured rubber, a 4-gram sample, which were independent of film thickness and were attributed entirely has been milled to pass a No. 20 screen, is extracted for z4 hours t o differences in film quality. Some films obtained from gelatiwith two 80-ml. portions of a 32-68 blend of acetone-chloroform.
Calibration for Nitrile Determination

Figure 6.

V O L U M E 20, N O . 1, J A N U A R Y 1 9 4 8
nous preparations !yere of such low transparency that scattering losses contributed half of the measured absorption. Even for such inferior films use of the scattering correction gave results within about 2YG of the correct value. In view of the superior reliability of the corrected function the analysis of both the cured and uncured samples of the eight commercial Buna-N rubbers was repeated, using both corrected and uncorrected measurements. The accuracy of the results (set 3, Table V I ) c o r r e s p o n d e d t o * 1 yo average error for either funct,ion except for one samp l e , P e r b u n a n - 3 5 , where gelation caused larger errors in the results obtained with the uncorrected f u n c t i o n . An additional preparation of the gum samples was made but without the ethanol extraction. The results (set 2, Table 11) exhibited excellent reproducibility but deviated from the p r e v i o u s r e s u l t s in some cases. Extraction appears to be necessary for highest accuracy, although this is surprising in view of the fact that the method is based upon the specific nitrile abso r p t i o n . Ilon-ever, with a de q u a t e sample preparation and avoidance of gelation, an average deviation :1 in results is readily of i obtainable in the application of the method either to gum or to cured Buna-S copolymers.
SPECTRA L-STRCCTURA L R E LATION SHIPS

21
Acrylonitrile Analysis of Commercial Buna-N Copoloymers
hleasurementsa (4.47) 3.50 mu 8 5 . 0 / 0 . 0 7 0 16.7/0.777 90.0/0.046 3 4 . 0 / 0 . 4 6 8 8 6 . 5 / 0 . 0 6 3 26.7/0.574 84.5/0.072 20.0/0.699 74.0/0.130 87.2/0.060 72.5/0.140 Xitrile Determination b 4.47 (4.47) 3.50 % % 0 , 1 0 0 0 7 . 2 0 . 0 9 1 08.2 0.132 0 9 . 4 0 . 1 0 1 08.8 0 . 1 0 8 0 8 . 0 0 . 1 0 1 08.8 0.103 0 7 . 4 0.103 09.0 0.201 1 4 . 0 0.191 0 . 1 9 4 1 3 . 6 0.189 0.222 l 5 , Z 0.205 0.298 0 309 0.298 0.244 0.261 0.285 0.301 0.324 0.321 0,317 0,442 0.444 0.420 0.442 0,407 0,422 0.445 0.475 0.425 0.447 0.427 19.6 20.0 19 6 16.6 17.4 18.6 19 F, 20.6 20.4 20.2 14.8 14.6 15.6

Table VI.
Sample
~

CopolymerC Perbunan-10

Set$ and r u n 4.60 mu 98.07 1 a b 96.Od 99.0% 2 a b 100.0%

1 a b 2 a 1a b c

4.47 mu 83.5/0.78 86.7/0.062 85.5/0.068 84.5/0.072 72.8/0.138 86.7/0.072

70.0/0.155

(3.50) 17.0/0.770 35.2/0.453 27.0/0.568 20.0/0.699

Perbunan-16 Perbunan-18 (20.5%)

98.57 99.59 97.0# -

20.6/0.686 20.9/0.680 47.8/0.320 48.0/0.318 20.0/0.699 2 0 . 7 / 0 . 6 8 4 f7.3/0.761 28 8 / 0 . 5 4 1

22.010.658

59.2/0.227 68.0/0.167 63.8/0 196 77.8/0.108 66.7/0.176 61.0/0.214 60.5/0.218 62.0/0.208 63.8/0.195 58.3/0.234 54.7/0.262 45.8/0.339 62.7/0.203 61.0/0.214 63.5/0.197 45.5:0.342 51.7/0.286 52.8/0.277 47.810.320 51,7/0.287 47.8/0.320 37.5/0.426 50.8/0.294 40.1/0.397 39 . O / O , 408 54.0/0.267 31.6/0.500 35.0i0.456 38.3/0.416 51.1/0.292 44.5/0.352

17.6 17.6 19.0 19.5

2 a b 3 a b 4 a
b c b
C

99.0% 97.57 97.5% 95.57,

87.57,

78.5/0.106 3 6 . 0 / 0 . 4 4 4 3 6 . 5 / 0 . 4 3 8 69.0/0.160 2 1 . 1 / 0 . 6 7 5 21.8/0.662 6 3 . 0 / 0 . 2 0 0 17.8/0.750 1 8 . 3 / 0 . 7 3 8 6 3 . 3 / 0 . 1 9 8 18.9/0.724 19.8,0.704 22.8/0.642 24,7/0.607 -67.0/0.174 1 8 . 3 / 0 . 7 3 8 21.0/0.678

0.242 0.244 0.271 0.282 0.257

Perbunan-26 (26.7%)

l a

d 2 a b 3 a
b

65.0/0.187 47.5/0.323 52.8/0.78 54.0/0.268

4 a b
C

54.5/0.264

Perbunan-35 (33.5%)

l a b
C

2 a b 3 ea b 4 a b
C

21.6/0.664 3 s .4 / 0 . 4 8 2 23.8/0.624 4 8 . 3 / 0 . 3 16 2 4 , 4 / 0 . 6 1 2 64.0/0.194 39.5/0.403 46.3/0.334 2 0 . 6 / 0 . 6 8 6 4 7 . 7 / 0 . 3 2 2 2 3 . 3 / 0 633 22.8/0.642 35.0j0.456 54.8/0.260 27.8/0.566

__ __ _-

25.5/0.593 17.2/0.764 33.0/0.483 33.0/0.485 32.8/0.484 33.7/0.472 15,5/0.810 16.3/0.788 22.8/0.642 2 3 . 2 / 0 . 6 3 2 26.0/0.583 2 6 . 7 / 0 . 5 7 4 17.6/0.755 22.7/0.644 17.8/0.750 20.2/0.696

___

0.643 0.651 0.638 30.3/0.518 0.667 4 6 . 6 / 0 . 3 3 0 0.662 30.0/0.523 0 . 7 3 0 31.8/0.498 0.720 0,650 0.640 34.2/0.466 0.633

__ __ __
-_

26.2 26.4 25.2 26.2 24.6 0.396 25.4 0,410 26.4 0,440 27.8 0 . 4 6 7 25.4 26.4 2 5 . 6 0.378

18.4

24.8 25.4 26.6 27.7

34.6 34.8 34.4 35.4 35.2 37.6 37.4 34.8 34.4 34 2

--

-__ --_ -_

0.610 0.590 0,640 0.647 0.557

- - 31.2

33.2 32.4 34.2 34.5

ChemigumIi-1

(30.8%)

1a b
c

2 a b 3 a b 4 a
b c

46.?/0.328 51.0/0.292 36.2/0.440 9 7 . 0 % 44.0/0.356 98.570 56.7/0.264 9 5 . 0 % 43.3/0.364 96.5% 51.1/0.292 - 47.2/0.326 52.7/0.287 93.57, 51.0/0.292 63.6/0.181 52,710,276 56.8/0.244 98.570 7 4 . 5 / 0 . 1 2 8 97.570 67.8/0.168 9 8 . 0 7 ~ 63.3/0.198 93.5y0 60.0/0.222 - 54.0/0.268 - 64.5/0.191 9 0 . 0 % 64 . O / O . 194 90.0% 94.5% 94 0% 90 07, 94.570

45.5/0.342 57.6/0.240 45.5/0.342 53.0/0.276 54.8/0.260

31.6/0.515 29.0/0.538 31.0/0.510 24.8/0.604 14.6/0.834 16.2/0.788 30.0/0.523 21.6/0.665

20.0/0.700

20.8/0,588

24.5/0.608 -2 8 , 4 / 0 . ,544 16.710.804 1 6 . 7 / 0 . 7 7 7 17.3;0.762 27.8/0.556 28.2/0.550 2 0 . 0 / 0 . 7 0 0 21.0/0.678 2!.6/0.575 27.6/0.560

--

0.540 30.6 0.537 30.4 0.548 30.8 0.458 27.0 0.442 26.2 0 . 5 2 0 29.6 0.55 0.508 3 29.2 1.0

0.448 0.436 0.505 0.493 -

27.0 26.4 29.2 28.8

29.4

--

0.558 31.2 -0.542 3 0 . 6 0.810

ChemigumN-2 (20.7%)

l a
C

__

75.7/0.120 69.7/0.156 64.8/0.188 64.3/0.192

30.510.516 22.3/0.652 20.4/0.690 20.8/0.682

2 a b 3 a b 4 a b
C

71.0/0.150

19.4/0.712 15.0/0.820 26.4/0.578 24 . O / O . 620

-_-

26.5/0.576

0,300 19.2 0,332 21.0 0.310 1 9 . 8 0.245 1 6 . 2 0.253 16.6 0.283 18.2 0.312 2 0 . 1 0,327 20.6 0.330 2 0 . 8 0 . 3 1 3 20.0

0.232
0.240

- -

0.272 0.282
0.260

- - -

17.0 17.4 19.0 19.5

18.4

ChemigumN-3
(40.0%)

l a b 2 a b 3 a b
4 a

b c

53.5/0.270 34.410.462 21.6/0.655 39.5/0.403 32.8/0.484 27.8/0.556 42.2/0.375 5 1 . 2 0,291 34.5i0.462

24.0/0.620 41.8/0.380 35.0/0.456 30.8/0.511

36.5/0.438

0.764 44.0/0.354 0.790 25,8/0.386 15.0/0.824 16.7/0.777 0.806 3 1 . 7 / 0 . 5 0 0 3 3 . 5 / 0 . 4 7 5 0.806 2 6 . 1 / 0 . 5 8 4 27.8/0.556 0,830 22.2/0.654 2 4 . 7 / 0 . 6 0 8 0 . 8 5 0 0.795 33 . 7 / 0 ,472 0,820 44.0/0.356 2 7 . 8 / 0 . 5 5 6 29.5/0.530 0.830

__ -

39.0 39.8 40.2 40.2 41.0 41.7 39.8 40.8 41.0 39.2 39.8
41

0.828

39.8 0.800 4 0 . 0 0,820 40.6 0.840 4 1 . 2 41.0

0.798

Hycar OR-15 (41.8%)

l a b
C

-_
96.0% 98.0% 97.0% 95 570

It remains impossible to assign many of the bands in the spectra of polyatomic molecules to exact molecular vibrations. However, certain functional groups in the molecule frequently control certain vibrations and thus it is possible in many cases to correlate spectral absorption 17-ith molecular structure. In the case o,f rubber spectra, these relationships are incomplete and are mostly empirical, but because of thcir importance in connection with analytical work,

2 a 3 a

4 a b
C

87.0%

26.5/0.574 32.8/0.484 44.8/0.348 31.0/0.508 3 2 . 4 / O ,490 5 4 . 5 / 0 . 2 6 4 55.5/0.256 32.2/0.492 33.4/0.476 3 2 . 8 / 0 . 4 8 4 34.3/0.464 35.4/0.451 54.W0.264 44.8/0.348 39.0/0.410
42,2/0.373 38.8/0.411 49.7/0,303 5 9 . 0 / 0 . 2 3 0 61.8,/0.210 40.5/0.392 39 . O / O , 408 42.2/0.375 44.2/0.355 40.5/0.392 44.0/0.356 42.0/0.377 46.310.344 46.7/0.331 53,810,269

18.0/0.740 24.5/0.611 38.010.420 21,6/0,66.5 46.7/0.330 25.6/0.092 27,210,566 27,210.566 47.8/0.321 30.5/0.516 20.2/0.692 18.0/0.745 28.2/0.549 37.2/0.430 17.8/0.750 22.210.654 21.7/0.664 20.3/0.692 24.2/0.616 Z5.6/0.592

0.776 0.793 n mn 22 6/0.646 0.780 47.7/0.322 0.800 26.4/0.578 0.830 28.5/0.545 0.870 0.798 0.822 35.2/0.454 0.795

0.758 38.6 0.795 39.8 0.825 4 0 . 8 4 2 . 3 0.853 41.7 40.0 40.8 - 3 9 . 8 0.785 39.4 39.4 40.0
41.0

n -

- - -

Hycar OR-25 (33.0%)

l a b
C

__

2 a b 3 a b 4 a
b

95 5% 96.0% 95.5% 92.570

0.506 29.2 C 87 5% Perrentage transmittance and extinotlon a t each analytical mave length given as T / E . Parentheses a t head indirate v ~ l l i i m r n r r w t e d t o 100% transmittance at 4.60 mu.

39.0/0.408 18.5/0.732 23.2/0.634 0 . 5 i 5 23.5/0.629 0 . 5 9 0 0.545 0.559 29.410.532 0.560

0.540 0.552 0.553 0.533 0.544

30.4 31.0 31.0 30.4 30.6 32.0 32.6 30.6 31.4 31.2

0.515 0.535 0,560 0.566

- - - 29.6 30.4 31.2 31.5

- -

22

ANALYTICAL CHEMISTRY
Table F'II.
Sample Copolymere 1 (0.9%)
J

ierylonitrile inalysis of Special Buna-N Cop015 mers

Measurements" 4 47 mu (4.47) 3 50 mu 89.0/0.050 99.I?/0.00 19.0/0.720 92.5/0.034 9 9 . a / 0 . 0 0 17.2/0.764 96.0/0.020 99.5/0.00 09.0/01.05 97.5/0.010 99.5/0.00 19.0/0.720 91.0/0.040 92.7/0.032 20.5/0.688 9 3 . 3 / 0 . 0 3 0 94.5/0.024 29.4/0.532 92.2/0.035 93.2/0.030 1 9 . 0 / 0 . 2 7 0 92.3/0.035 93.3/0.030 20.0/0.699 79.5/0.100 81.7/0.087 25.0/0.602 14.5,'O. 838 85.5/0.068 88.0/0.055 84,010.075 8 5 . 0 / 0 . 0 7 0 20.0/0.699 76.7/0.115 78.5/0.105 23.3/0.632 7 1 . 0 / 0 . 1 4 8 72.5/0.140 15.5/0.810 74.0/0.130 75.2/0.124 1 9 . 0 / 0 . 7 2 0 75.5j0.122 76.5/0.116 20.6/0,686 74.5/0.128 7 8 . 0 / 0 . 1 0 8 3 4 . 0 / 0 . 4 7 0 60.5/0.218 62.2/0.206 14.5/0.840 66.7/0.176 68.0/0.168 21.1/0:676 5 1 . 5 / 0 . 2 8 8 53.0/0.276 16.7/0.777 6 1 . 7 / 0 . 2 1 0 63.5/0.198 24.4/0.612 6 2 . 8 ' 0 . 2 0 2 64.2/0.192 26.0/0.584 57.2/0.242 58.5/0.232 20.6/0.686 58 3'0 234 60.5/0.218 34.0/0.468 41 7/0 286 53 8/0 270 25.0/0.602 43 3 0 363 47 0/0.328 19.0/0.722 3 0 . 0 / 0 . 5 2 3 33.5/0.474 15.5/0.810 3 7 . 8 / 0 . 4 2 2 40.5/0.392 19.5/0.710 3 9 . 0 / 0 . 4 1 0 42.7/0.370 20.0/0.699 3 5 , 5 1 0 . 4 5 0 51.5/0.288 29.0/0.538 35,5/0.450 65.3/0.184 31.0/0.508 29 . O / 0.538 43,310.363 19.5/0.710 24.0/0.620 28.0/0.554 1 6 . 7 / 0 . 7 7 8 3 5 . 0 / 0 . 4 5 6 39.5/0.404 26.7/0.574 26.010.383 29.8/0.526 19.0/0.722 2 7 . 2 , ' O . 566 33,010,482 20.6/0.686 Titrile Determination b 4 47 (4 47) 3 50 13 2) 5 0.0 0 . 0 7 0 04 8 0 . 0 0 0 013 0 2 . 4 0 . 0 5 0.0 0 0 0 020 00 0 0 . 0 0 0,010 0 0 . 0 0 . 0 0 0.0 4.4 0.058 0 4 . 0 0 . 0 4 7 4.2 0.056 0 3 . 6 0.046 4.0 0.049 0 3 . 2 0 . 0 4 2 0.050 0 3 . 2 0 . 0 4 3 4.0 0 , 1 1 9 0 8 . 8 0.105 9 2 0.113 0 8 . 2 0.093 8.6 8.8 0.107 07.8 0.101 0.182 1 2 . 8 0.169 1 3 . 6 0.183 1 2 . 8 0.176 1 3 . 8 0.180 1 2 . 6 0.173 1 3 . 8 0.178 1 2 . 6 0.170 1 3 . 6 0.272 18.0 0.240 17.4 0.260 1 7 . 6 0 . 2 4 8 1 7 . 8 0.260 1 7 . 6 0.252 1 8 . 0 0 . 3 7 0 2 3 . 0 0 . 3 6 0 23.2 0 . 3 4 3 2 1 . 6 0 . 3 3 2 22.0 0.346 2 1 . 8 0 . 3 3 5 2 2 . 0 0.552 2 2 . 2 0 . 3 4 2 2 2 . 4 0.500 2 8 . 8 0.478 2 8 . 2 0 . 4 7 5 2 7 . 8 0 . 4 6 3 27.6 0.503 2 9 . 0 0 . 4 7 5 2 8 . 0 0.645 34.6 0.622 33.6 0.595 32.8 0.575 31.8 0.585 3 2 . 4 0 . 5 6 3 3 1 . 4 0.836 4 1 . 2 0.762 38.6 0.885 4 2 . 8 0 . 7 5 5 3 8 . 4 0 . 7 5 8 3 8 . 6 0.660 3 5 . 0 0.797 40.0 0.780 39.2 0.795 40.0 0.775 39.0 0.810 4 0 . 4 0 . 7 8 8 3 8 . 6 0.825 4 0 . 8 0.795 3 8 . 8

(4.2%) 3 (8.7%) 4 (14.0%)

5

(li.0Yc)
6 (21.7%)

7 (27.0%)
8 (32.3%)

Setd and run 1a h 2 a h 1a 1b 2 a 2 b 1a 1 b 2 a 1 a 1 h 2 a 2 h 1a 2 a 2 h 1a 1h 2 a 2 b 1 a

2 a 2 h

9C (35.270) 10 (38.2%)

1 2 2 1
1

2 1 1
1 1

a a b a h a a h c d

4 6 0 mu 90.0% 93.5% 96.5% 98.0% 98.5% 99.0% 99.0% 99.0% 91.5% 97.5% 99.0% 98.0y0 98.5% 98.5% 99.0% 95,5% 97.5% 98.0% 97.5% 97.5% 96.0% 92.5% 96.5% 96.0% 92.5c7, 90.0% 93.5% 91.070
68.07c

(3 50) 21.0/0.678 18.5/0.732 09.1/01.04 19.3/0.715 21.0/0.678 29.8/0.526 19.2/0.717 20.2/0.695 14.8/0.828 25.7/0.590 20.2/0.695 23.8/0.622
15.8/0.800

--

19.2/0.716 20.8/0.682 35,5/0.450 14.8/0.830 21.6/0.666 17.0/0.768 25.2/0.598 26.7/0.574 21.0/0.678 35.0/0.456 26.0/0.584 20.4/0.690 17.3/0.762
20.8/0.682

.54.5YC 67.07, 85.5% 98.0% 88.0% 83.0%

22.0/0.658 42.0/0.378 57.0/0.244 28.2/0.550 19.5/0.710 30.0/0.523 21.5/0.668 24.8/0.606

GPercentame transmittance and extinct,ion a t each analytical wave length given as T I E . Parentheses a t head of column indTcate values corrected t o 100% transmittance a t 4.60 mu. b Using indicated extinction ratio and appropriate working curve of Figure 5 . c Xitrile content as determined hy Kjeldahl analysis given in parentheses. d Suinhers refer t o sample preparations; letters designate several films studied. Roth preparations are of gum samples. e Gelatinous sample and correspondingly unsatisfactory films.

as vel1 as in the general elucidation of molecular structure, some of the correlations which have been found useful in this investigation are summarized here. Bands due t o C-H vibrations are found in the spectra of all hydrocarbon-type elastomers and these may be considered collectively before passing on to those which characterize the various elastomers. The bands in the neighborhood of 3.4 mu are due chiefly to C-H valency (stretch) vibrations, while those at 7 mu are due to C-H deformation (bending) vibrations. The study of a large number of hydrocarbons has shown that -CH*, methyl, groups absorb a t 3.37 mu and a t 7.25 mu Tvhile >CHt, methylene, groups absorb a t 3.42 mu, and that both methyl and methylene groups absorb a t 3.5 mu and in the 6.81 to 6.98 mu region. Greater spectral resolution has shown that the absorption in the 3.4 mu region is complex ( 7 ) and the presence of additional unresolved bands determines the exact contour of the bands obtained here and confuses the assignment. However, the correlations obtained under moderate resolution agree with the frequencies assigned to the stronger C-H bands by Fox and Martin ( 7 )and are useful in identifying the various types of C-H bands when they are present in significant proportions and in not too great variety. These bands in Figures 2 and 4 are too strong to show the resolvable structure, but results obtained xith thinner films were consistent Kith the expected absorptions for all of the elastomers, except for a minor difference in the case of Butyl 15 here a band a t 3.50 mu was expected but not found. Two additional bands a t 3.67 and 2.32 mu are found in the spectra of most elastomers, as well as in the spectra of most hydrocarbons, and are due t o combinations or harmonics of other C-H frequencies. Natural (Hevea) rubber consists of long chainlike molecules

composed of the recurring isoprene unit, I, and all the strong

CH,
-CH*-C=CH-CH*bands in the spectrum are consistent TTith this structure.
Wave Length, M u Absorbing Structure -CHI >CHz -CHo. >CHz C = C (valency vibration) -CHs, >CHa -CHI CHI

1)

11.95

-b

= CH-

Sheppard and Sutherland ( 1 2 ) suggest that weak absorption around 3 mu is composed of a band a t 3.1 which is to be distinguished from hydroxyl (oxidation) absorption at 3.0 mu. Increased resolution obtained here has located this band a t 3.05 mu, while the oxidation impurity absorption (which always appears stronger in vulcanized and reclaimed samples) appears to be nearer 2.9 mu. The 3.05 mu band is characteristic of the rubber polymer although it is frequently masked by the hydroxyl impurity absorption. I t has previously been assigned (12) either to a combination frequency or to the fundamental of the hydrogen vibration relative to a tertiary carbon. The same authors also report a new band a t 4.98 mu which has not been confirmed here. This band was found for only one sample and apparently originates in an impurity. However, most hydrocarbons absorb weakly in this region and under proper conditions-Le., using a thick film-a band at 4.93 mu would be expected in the spectrum of natural. All other bands reported by Sheppard and Sutherland were also found in this investigation.

V O L U M E 2 0 , N O . 1, J A N U A R Y 1 9 4 8
Weak bands in the region of 5.8 mu are due t o C=O, carbonyl, oxygen resulting from oxidation, and, rn for the 2.9 mu hydroxyl absorption, these impurity bands are stronger in vulcanizates than in gums. The 6 mu olefinic absorption varies with structure, terminal linkages (vinyl groups) absorbing at 6.10 mp, internal double bonds absorbing a t slightly shorter wave lengths and in general at much less intensity, and double bonds at branching centers usually absorbing at slightly different wave lengths than the corresponding unbranched structures. The olefinic absorption for the normal rubber skeleton absorbs a t exactly 6.00 mu, and the weaker absorption at approximately 6.1 mu probably results from olefinic groups modified by removal from methyl branching centers or by proximity to the end of the polymer chain. Absorption around 6.5 mu found by Barnes (3)but not by Sheppard and Sutherland (1.2)was not consistent in the various samples and must be due to an impurity. Since stearates (12) and many soaps ofother fatty acids absorb strongly a t 6.5 mu, the absorption of natural gum rubber in this region probably results from the presence of unextracted vegetable soaps or esters. It is noted that this absorption becomes weaker after vulcanization, which would tend to decompose these impuritj- compounds. For natural rubber, as for the other elastomers, the bands beyond 7.5 mu are generally characteristic of larger molecular groupings, so that correlation of absorption and structure is more difficult. I n addition, correlations which have been found for liquid hydrocarbons are not necessarily applicable to rubber spectra, since considerable variation in position and intensity of absorption occurs with change of state. However, the strong band a t 11.95 mu has been assigned by Thompson and Torkington (15 to type R1R2C=CHRs structures and this corresponds to structure I for natural rubber, provided that R1 or Rg is methyl. Since the bands which characterize other olefinic structures such as those in IV and V (see belox) are absent, all double bonds in natural rubber apparently occur a t branching positions. Absorption a t 11.25 mu may be due to type R-(CHs)C= CH2 structure a t the ends of the elastbmer chains. 2-Methq-11-heptene absorbs strongly a t this wave length and the terminal structure would also account for the n-eak olefinic absorption around 6.10 mu. The intensity of the 11.25 mu absorption varies for different samples and may be dependent upon molecular weight, although this has not been investigated. The question of czs-trans isomerism about the double bond (9) is being considered but must be deferred for the present. Butyl rubber, or GR-I, designates a class of elastomers prepared by low-temperature copolymerization of isobutylene with a very small percentage of a diolefin-e.g., isoprene. The elastomer chains undoubtedly consist chiefly of head-to-tail polyisobutylene units, 11, separated occasionally by a 1,4- addition isoprene unit, I. The small olefin content (about 0.6 mole cc) contributes the necessary property of vulcanizability, but does not contribute significantly to the absorption, so that the spectrum of GR-I corresponds to that of pure head-to-tail polyisobutylene (15), and all the strong bands correlate entirely with structure 11. CH, CH, -C-CH?-C-CHZI

23
In some cases the 3.4 mu absorption was resolved into methyl and methylene bands a t 3.38 and 3.44 mu, but the bands are diffuse and difficult to resolve. A band ordinarily found a t 3.50 mu is missing, which indicates that the absorption in this region probably consists of several unresolved bands. Interpretation of the bands a t 8.56, 10.52, and 10.82 is uncertain, but their origin certainly lies in distortion vibrations of the carbon skeleton (16). h quaternary carbon atom does not seem fundamentally 1,1,2related to these bands, however; 1,1,6-triisobutylhexane, triisobutylethane, 1,l-diisobutyltridecane, and isobutylbenzene all absorb a t 8.55, 10.55, and 10.85 mu (but not a t 8.0 to 8.3 mu) while isopropylbenzene absorbs at 10.85 mu only. Neoprene is a trade-name which designates a variety of elastomers formed by polymerization of chloroprene and composed of the recurring structure 111, which differs from I only in that the methyl has been replaced by chlorine.

-CH*-

8'

I11

=CH-CHS-

dccordingly, the spectrum of neoprene agrees with that of natural if allowance is made for the expected effects of chlorine substitution.
Wave Length, Mu
3.42 3.50 6.02 6.90 6.98)

Absorbing Structure

>CHz >CHn
>CHz

c=c

8.92'
12 10

?
-d=CH-

CH,
Wave Length, M u
ca.

I CH,

(11)

3.40 6.80 7.201 7.301

Absorbing Structure -CHa, > C H I -CHa, >CHg -CHs (deformation vibration; methyls on same carbon atom)

An unresolved band a t 3.30 mu is probably due to a fundamental vibration of hydrogen relative to a tertiary carbon (i), and the 3.38 mu methyl band is absent. Oxidation impurity bands a t 2.90 mu and from 5.5 to 6.0 mu are generally present, especially in cured samples. The 6.02 mu band with a shoulder a t 6.12 mu isanalogous to a similar band in the spectrum of natural and is correlated with the olefinic structure, although Sears (11)suggests that a harmonic of the 12.10 mu band may contribute here too. The 6.90 mu band is the usual C-H deformation and the 6.98 mu band has a similar origin, the shift resulting from chlorine substitution (15)on adjacent carbon atoms. Other reproducible bands are found a t 7.42, 7.66, 8.16, 8.94, 9.20, and 10.00 mu, that a t 8.94 being one of the most distinctive features of the neoprene spectrum (for analytical purposes). Sears (11) quggests that the weak absorption a t 12.85 mu may be due to C-C1 absorption. There seem to be no strong bands betvieen 13 and 20 mu, where C-C1 vibrations might be expected. Neoprene-ILS is a polychloroprene which contains a small amount of acrylonitrile. Except for the nitrile absorption a t 4.47 mu, the spectrum of Neopren-ILS is virtually identical n.ith that of Neoprene-GX and Neoprene-E. Buna-S is a generic designation for the copolymers of butadiene I ! ith styrene, of which GR-S is the particular product from emulsion copolymerization of 3 parts of butadiene with 1 part of styrene. Condensation can occur by either 1,4- or 1,2-addition, giving, respectively, the structural units IV and V, both of which aic shoivn by the infrared spectrum.

two

C
8.13
8.56 10.52 10.82

C-&- C carbon atom)

C

(internal quaternary

-CH~-CH-CHZ-CH-

CH
I1

CsH,

CH,

24
Wave Length, M u 3.42 3.50 6.10 6.23 6.70 10.85 10.97 Absorbing Structure >CH1 >CHI >C=CHz Phenyl, caH6Phenyl, CaH6>CHn -CH=CH-(1,4addition) -CH=CH2 (1.2- addition) (vinyl group) Phenyl, c6H6Phenyl, CoHi1

ANALYTICAL CHEMISTRY
Wave Length, Mu 3.42 6.82 7.08 7.26 7.82 8.36 8.64 8.98 9.32 9.68 10.53 Absorbing Structure >CHz (valency) >CHz (deformation)

::E}

13.16 14,28

Bromatic C-H absorption, which usually appears as a band or doublet a t 3.26 and 3.30 mu, is indicated only by a shoulder in the GR-S spectra and probably is unresolved owing to interference from an additional tertiary C-H band at 3.35 mu. The 6.10 mu band is due to terminal olefinic (vinyl) groups and a weaker band a t 6.00 mu is characteristic of internal olefinic linkages. The bands from 7 to 9 mu are definitely reproducible for GR-S, but they have not been interpreted. Thompson and Torkington (15) have shown the strong 10.35 mu band to be associated with the structure R1CH=CHR2, and the 10.97 (and also the weaker band at 10.05) to be associated with the structure RCH=CH2 of a vinyl group, all these bands arising from C H deformation vibrations. The relative intensities of these bands have been used a t this laboratory and elsewhere (6) t o derive information concerning the relative amounts of 1,4- and 1,2- addition in butadiene polymers and it is noted that 1,4- addition (structure IV) predominates in GR-S as in other emulsion polymers. The 14.3 mu band is characteristic of all monoalkyl benzenes and a band around 13.2 mu seems to appear whenever the alkyl group differs from the methylene structure a t the attachment carbon. Buna-N is a generic designation for a variety of copolymersHycar OR, Perbunan, Chemigum, Butaprene-forme$ by emulsion copolymerization of butadiene with 15 to 40% of acrylonitrile. Condensation can proceed by either 1,4- or 1,2- addit on, giving, respectively, structural unit VI or VII. The spectrum is similar to that of the structurally analogous Buna-S, -CH~-CH=CH-CH~-CHZ-CHVI
I .

Except for the S-H, mercaptan, frequency a t 3.9 mu, which is not present here, very little is known about the characteristic absorption frequencies of sulfur compounds (I$), especially these concerned in structure VIII. The aliphatic ether oxygen groups probably account for the absorption envelope in the 8.5 to 10.5 mu region, but the origins of the discrete maxima, however, are not known. Thiokol-RD, contrary t o t4e usual terminology, is a nonsulfur material formed by interpolymerization of butadiene, acrylonitrile, and an unsaturated ketone. The molecular units of the interpolymer can hardly be predicted with completeness, but the expected Buna-N structures VI and VI1 as well as additional ketonic structures are evident from the spectrum.
Wave Length, M u 3.42 3.50 4.47 5.86 6.00 6.84 to 6.94 7.25
7 .Y 9Q

10.25 10.86

Absorbing Structure >CHz (aliphatic) >CHz, -CHn - C Z N (nitrile) C=O (carbonyl) > C = C < (aliphatic) >CH* a n d -CHs (deformation) -CHn (deformation) -CH=CH-(1,4addition) -CH=CH% (1,2- addition) (vinyl group)

The band a t 5.86 mu is too strong to be due to oxidation impurities and must be ascribed t o the presence of carbonyl oxygen structures incorporated into the elastomer during polymerization.
ACKNOWLEDGMENT

i:
N

iii

-CH2-CH-CH2-CH
I

The authors wish to express their appreciation of the interest and extensive cooperation of Peter King and Ralph Taylor, NRL High Polymer Section, and of other members of their section, especially Ferdinand Thurman, who compounded all the cured samples, and Donald Cohen, who performed the Kjeldahl analyses.
LITERATURE CITED

except that the phenyl bands are all absent and a new nitrile band due t o the C s N valency vibration (14) appears a t 4.47 mu.
Wave Length, W aveL e M u 3.42 3 3.50 4.47 6.10 6.90 10.32 10.86 Absorbing 1g Structure >CHz >CHz -G CEN >C=CHx > CHa -CH=CH(1,4- addition) -CH=CHz (1,2- addition) (vinyl group)
/).

(1) Am. Soc. Testing Materials, Standards Supplement, Part I11 (1943), Section D, 297-43T, Tentative Methods of Chemical Analysis of Rubber Products.

(2) Barnes, Gore, Liddel, and Williams, Infrared Spectroscopy, Industrial Applications and Bibliography, New York, Rein-

Thiokol is commonly used to designate the sulfur-containing elastic resins formed by condensation of sodium polysulfide with ethylene dichloride or its derivatives. Thiokol-FA is one variety made by reaction of sodium polysulfide with dichloroethylformal to give the recurring structure VIII.

LCH2-CH2-O-CH2-O-CHa-CH2-S-S-

8 B

IIII

VI11

hold Publishing Corp., 1944. (3) Barnes, Liddel, and Williams, ISD.ESG.CHEM., ANAL.ED., 15,83 (1943). (4) Barnes, Williams, Davis, and Gieseke, Ibid., 16, 9 (1944). (5) Cole and Field, I n d . Eng. Chem., 39, 174 (1947). (6) Field, Woodford, and Gehman, J . Applied Phys., 17, 386 (1947). (7) Fox and Martin, Proc. Roy. SOC., A175 208 (1940). (8) Gore, J . Chem. Education, 20,223 (1943). (9) Hendricks, Wildman, and Jones, Arch. Biochem., 7,427 (1945). (10) Nielsen, Crawford, and Smith, J . Optical SOC.Am., 37, 296 (1947). (11) Sears, J. Applied Phys., 12,35 (1944). 41, 261 (1945). (12) Sheppard and Sutherland, Trans. Faraday SOC., (13) Sutherland, Thompson, and others, Ibid., 41, 171-295 (1945). (14) Thompson and Torkington, J . Chem. SOC., 11,597 (1944). 41, 246 (1945). (15) Thompson and Torkington, Trans. Faraday SOC., (16) Wright, IND. ENG. CHEM., ANAL. ED., 13, 1-8 (1941).
RECEIVED April 19,1947.

As expected, the infrared spectrum bears no great resemblance to the spectra of the common hydrocarbon-type elastomers.