v01. 39, No. 2
INDUSTRIAL AND ENGINEERING CHEMISTRY Kraemer, E. O., IXD.EXG.CHEM.,30, 1200 (1938). Long, J. H., Tech. Assoc. Papers, 22, 296 (1939). Lum, J. C., and Keating, T. J. (to Westinghouse Elec. & Mfg. Co.), U. S. Patent 2,367,954 (Jan. 23, 1945). Malm, C. J., Fordyce, C. R., and Tanner, H. A , IND.ENQ. CHEM.,34, 430 (1942). Miller, B. C. (to Bert C. Miller, I n ? , . ) , U. S.Patent 2,117,199 iMav 10. 1938). Ibid.,
A
00.
+- --CH--CH-CH--C€I,-
--CH-CH=CH----CII,---
I
i3j
OOH l'ermination of the reaction chain occurb by combination of the involved in the propagation step. Thermal decompo-ition of the hydroperoxide accelerates the reaction, probably by formation of a free radical capable of starting an oxidation chair,. Peroxide decomposition occurs side by side with peroxide form&tion. This complex decomposition is not well understood, but, in general, the hydroperoxide group reverts to hydroxyl. The active oxygen reacts mainly w t h double bonds t o form epoxides. There is some evidence that peroxide decornpoqition ma7' w!io lead t o the formation of carbonyl groupb. 1 adicals
3d00
I
2000
W
I
is00
1230
lib0
Id00
I
BOO
FREQUENCY, CM
Figure 1. GR-S w-ithout Antioxidant
It is of interest t o determine the amount of combined oxygen required t o bring about changes in the solubility and viscosity of the polymer. Table IV shows the changes 111 oxygen content, benzene solubility, intrinsic viscosity, and antioxidant content. The values recorded for oxygen uptake are probably accurate t o about = t O . l ~ o . An oxygen uptake of less than 0.5yo was suficient to cause significant changes in the solubility and viscosity of the polymer. Oxidation caused a marked decrease in the antioxidant content. It appears that the amount of gel formed is roughly proportional to the amount of combined oxygen. This may be related to the observation of Shelton and Winn (19) that the increase in 200y0stress of GR-S during oxidation is roughly proportional to the amount of combined oxygen. Table V gives the oxygen content for GR-S and polybutadiene samples after aging a t room temperature for long periods. These were highly purified samples of the polymer hydrocarbon and contained no antioxidant. The original samples were entirely colorless; the oxidized samples were colored yellow to yellow-orange. These observations on the effect of oxidation on uncured GR-9 indicate that both chain scission and cross linking occur as the result of oxidation of the polymer. Except during the early stages of oxidation, the rate of cross linkingis more rapid than chain scission. Since aging properties. are fundamentally related t o the chemistry of the polymer hydrocarbon, the same reactions should occur during oxidation of vulcanized polymers, though probably a t a different rate. I n general this prediction is confirmed by the effect of oxidation on GR-S vulcanizates (18, 23). Although experimental studies of the effect of aging on physical properties are useful for practical purposes, they give little information as to the mechanism oxidation. HYDROPEROXIDE THEORY OF AUTOXIDATION
The studies of Farmer and eo-workers (4, 6, 6) on the oxidation of simple olefins, such as cyclohexene, and more complex olefins, such as methyl oleate and natural rubber, indicated that the course of oxidation of unconjugated olefins may be represented as follows: Peroxidation takes place exclusively or almost exclu-
- -CH-CII==CH-CH2--
I
+-CH-CM--@H--CH2 I
'0,
OH
OOH - CH-CI-I=CH-CH,--
I
4)
+ --CH=CH--
+
OOK -CH-CH=CH-CHg--
.C
--CH-CH-
0I 11
(5)
O'/
In vie\$-of the important effect of chain scihsion and cross iirrking on t,he physical properties of a polymer, the course by which these reactions occur is of particular interest. The attcmpts which were made by Farmer and his eo-wvorlrers (4, 7 ) and also by Taylor and Tobolsky (22)to explain these reactions were directed along two somewhat different lines of thought. The first theory assumes that both chain scission and cross linking occur through reactions of the peroxide or its decomposition products. The aecond theory considers t'hat the role of oxygen is to form free radicals, which may disproportionate to cause chain scission or a ttaclr t,he double bond to bring about cross linking. Application of the hydroperoxide t'heory of Oxidation to sgrr-
TABLEIV. OXYGENUPTAKEDURIXG OXIDATIONOF GR-S CO~VTAINING PHE~YL-B-T\TAPHaHYLARIISEh T I O X I D A N T (PBNA: Hr. a t
looo C. 0 24 48
zz 1ii
Increase in % Oxygen 0 0.1
% Benzene Solubility 100
100
0.5
0.7 i n
i:i
88 66 50
% PBSk
Intrinsic Viscosity 1.76 0 96 0 99 0.66
Found 3.98 2.65 1.66 1.26
.. ..
..
..
38
TABLE V. OXYGEN UPTAKEDURING AIR OXIDATIONOF POLYMER HYDROCARBON AT ROOMTEMPERATURE I N ABSENCE OF ANTIOXIDANT Sample
Aging Time, Months
% Oxygen
Color of Aged Sample
Found 9 8
Yellow .~~~
11.7
Yellow Yellow-orange Yelloiv-orange
15.0
15.8
~
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
February 1947
177
i k L
thetic elastomers is complicated by the fact that polymerization of dienes usually occurs by both 4 2 and 1,4 addition t o the conjugate system. In the case of isoprene 3,4 addition also occurs (9). There is evidence that the C-H bond energy decreases in the order primary > secondary > tertiary. Further, the bond energy decreases when the hydrogen atom is attached to a carbon atom adjacent t o the double bond (90). On this basis oxidation of synthetic elastomers derived from butadiene and isoprene will occur a t points in the polymer chains a t which secondary or tertiary hydrogen atoms are adjacent t o a double bond. Such points are indicated by an asterisk:
SAMPLE
BUTADIENE POLYMERS
-&H-CH,-
-~H-CH=CH-&H~--
AH bH*
ISOPRENE POLYMERS
CBa
nh
-~H-cH~-
-eH8-h=CH--6H8-
-CH8
SAMPLE HEIITED
AH, Accordingly, in synthetic polymers derived from butadiene and isoprene, both secondary and tertiary hydroperoxides may be formed on oxidation. Milas and Surgenor (16) recently presented evidence that thermal decomposition of tert-butyl hydroperoxide occurs in the following manner: (CH3)sCOOH ICH,)&OOH
95-100' C.
250" C.
(CHJsCOH
+0
(6)
+ *OH
(CHa)aCO*
(7)
(CHa)jCO*+(CH~)LC=Of- CHI. CHa.
+ .OX1 +CHaOH
3600
2600
15100
12150
Figure 3.
Id30
IlbO
FREOUENCY,
a00
85C
CM:'
GR-S during Early Stages of Oxidation
The latter mode of decomposition seems to be a general reaction of tertiary hydroperoxides. George and Walsh (10) showed that the products obtained by oxidation of cyclopentane and cyclohexane derivatives a t 80-100" C. may be readily explained on the basis of the formation and decomposition of tertiary hydroperoxides. A similar decomposition of tertiary polymer peroxides may lead to chain scission and the formation of a ketone group: DOH
I
--CHz--C-CHz-R
--+ -CHz-
8
-CH,-R
AH
AH
&I*
CH,
+ .OH
(8)
I/
ABSORPTION SPECTRA OF OXIDIZED POLYMERS
I
I
3000
2OCO
'
1500
I 1250
I
1
I
1100
1000
900
FREQUENCY, C M . - I
Figure 2.
GR-S with Antioxidant
Experimental study of the mechanism of oxidation of GR-S and other butadiene polymers is made difficult by the fact that these polymers become insoluble as the result of oxidation. Since infrared absorption spectra can be readily measured on a thin polymer film,this methodis of particular value in such cases. The infrared absorption spectra curves shown in Figures 1and 2 were obtained by heating production GR-S and the GR-S hydrocarbon a t 105" C. i n an air oven and a t 40' C. in ultraviolet light (Atlas Fadeometer). These curves are plotted as transmission, measured as a galvanometer deflection, against frequency in wave number. The present infrared data were obtained with the object of studying the oxidation of unsaturated polymers from a qualitative standpoint.
,
178
Vol. 39, No. 2
I N D U S T R I A L A.ND E N G I N E E R I N G C H E M I S T R Y
band a t 970 cm.-'; however, the general absorption occurring in this region was a complicating
factor. DISCU s SION
SAMPLE HEA-ED 24 HRS. 41 IOC'
c.
The band which appears a t 3600 em.-' is associated with vibrations of the 0-H group. The intense absorption a t 1700-1720 cm.-'is undoubt,edly due to the presence of the C-0 group, but i t is difficult to determine the functional groups which may be involved. The frequency usually associated with the various carbonyl groups (1, 3, I d , I S , 16, 24) is given as follows: Monomeric acids Esters Aldehyde2 and ketones Associated acids
1770 c m . 1 1780-1725 om.--* 1725- 1690 cni. -1 li40-1700 rnl. --I
These characteristic frequencies are generally valid for t'he higher rrembers of a homologous series. Conjugation, however, tends to lower the assigned F i g u r e 4. Hevea R u b b e r Changes on Oxidation frequencies. If i t is assumed that any acid groups produced by oxidation of the polymer are distributed a t random, an absorption bandshouldappear a t the position correThe pronounced changes which occur in the spectrum of the sponding to the carbonyl band for monomeric acids a t 1770 cm-1, GR-S hydrocarbon as the result, of oxidation are shown in The absence of this band during the early stages of oxidation Figure 1. I n general, neTy bands were found t o appear a t 890, ivould seem t o indicate that no appreciable number of carboxyl R 1175,1720,and 3600 cm.-', strong general absorption T V ~ observed groups are present. There seems t o be some evidence that, this in the region 1000-1300 cm.-', and a decrease in absorption ocband is present in th(. later stages of oxidation. Although the obcurred a t 914 cm.-l The sample heated 20 hours a t 105' C. apserved bands a t 1720 and 1700 em.-' are probably due t o ketone peared to have undergone a much higher degree of oxidation than or ketone and aldehyde groups, i t is difficultt o determine the sigthe sample heated 48 hours a t 40" C. in ultraviolet light. nificance of this doublet. This doublet may indicate that both The changes in the spectrum produced by oxidation of regular aldehyde and ketone groups are present, or that the frequency of production GR-S containing phenyl-8-naphthylamine as antipart of the carbonyl groups has been shifted by conjugation. The oxidant are shown in Figure 2. Heating for 20 hours a t 105" c. fact that GR-S becomes yellow as the result of oxidation indicaused only slight changes i n t,he spectrum. However, heating cates the presence of conjugate carbonyl groups or carbonyl for 24 hours at' 40" C. i n ultraviolet light brought about changes groups conjugated with double bonds. I n either case a ketone in the spectrum similar to those observed vvhen oxidation was carbonyl 17-ould he involved. Both the shape and the width of carried out i n the absence of antioxidant. the band in the region 1770-1700 cm.-' suggest that several csrFigure 3 s h o w the changes which occurred i n the spect,rum of bony1 groups are present' in the later stages of oxidation. The the GR-S hydrocarbon during the early stages of oxidation. Kew bands a t 996 and 914 are attributed to side vinyl groups and bands appeared a t 890, 1700, 1720, and 3600 em.-' after 2 hours the band a t 967 em.-' to the internal double bonds present in the of heating a t 100" C. Furt'her heating increased t'he intensity of polymer (26). The decrease in intemity of the 914 cm.-i band the band a t 3600 cm.-l and the bands which first appeared a t indicates that' double bonds are undergoing saturation or scission. 1700 to 1720 cm.-l merged into one very intense band. This The very intense, general absorption in the region 1000-1300 band extended from about 1700 t o 1770 em.-'. General absorpcm.-l is associated x-ith vibrations of the C-0 group and may be tion began a t about 1000 em.-' and gradually extended toward due to acid, ester, hydroxyl, or ether groups. No definite explanahigher frequencies as oxidation proceeded. The intensity of the tion can be offered for the appeuance of a band a t 890 cm.-1, but band a t 914 cm.-l appeared to decrease more rapidly than the olefins of the type RnC=CH, have a band in this region. Studies on the oxidation of GR-S and nat,ural rubber showed that the concentration of peroxidic oxygen is never very high. Thus it is probable that the observed 0-H absorption during the eady stages of oxidation is due primarily t o hydroxyl and not to carboxyl or hydroperoxide groups. Since thermal decomposition of hydroperoxides results in the formation of hydroxyl and lretone groups, the probable presence of these groups in oxidized GR-S is taken as an indication that the initial oxidation product is a hydroperoxide. Other functional groups containing oxygen seem to be indicated by the complex nature of the changes in the absorption spectra, but identification of these groups cannot be made with any degree of certainty on the basis of the present data. Since the C-0 absorption bands for both alcohol and ether groups generally occur in the same region, 3600 2 h O 1200 lis0 iIb0 Id00 sbo abo FREPUENCY, C M - ' there seems to be little hope of establishing the presence or absence of ether-type crcss links in Figure 5. Polyisoprene Changes on Oxidation aboo
zdoo
15b0
Id50
lib0
FREQUENCY,
Id00
CU;'
9bo
ado
February 1947
INDUSTRIAL AND ENGINEERING CHEMISTRY
oxidized GR-S. As already indicated, chain scission may result from peroxide decomposition, which leads to the formation of aldehyde and ketone groups. Although evidence for the presence of these groups has been presented, there does not appear to be much possibility of establishing the mechanism of chain scission by infrared methods alone. Figures 4, 5, and 6 show the effect of oxidation on the spectra of the hydrocarbons derived from natural rubber, polyisoprene, and polybutadiene, respectively. The changes brought about by oxidation were similar to those observed with GR-S. I n the case of natural rubber and polyisoprene the decrease in intensity of the band a t 840 cm.-l indicates saturation or scission of the double bonds in the polymer chain. On the basis of the changes i n the infrared absorption spectrum, the mechanism of oxidation of -all the elastomers studied appears to be similar.
179
L!qLLb L L L L FRESH SAMPLE
SAMPLE HEATED 24HRS AT IOO’C.
2600
3600
1250
Id00
Id00
IlbO
r r r - , lCL,P
“
960
850
CM.7’
FREQUENCY. CM.-l
Figure 6.
Polybutadiene Changes on Oxidation
MECHANISM OF ANTIOXIDANT ACTION
Aging of GR-S for 20 hours a t 105’ C. in the presence of an antioxidant caused little change in the spectrum, whereas a similar aging period i n the absence of an. antioxidant brought about a marked change i n the spectrum. I n addition, the antioxidant provided much more effective protection against thermal oxidation than for oxidation catalyzed by ultraviolet light. These observations suggest that infrared methods may be of value in fundamental studies in the field of antioxidants. There seems to be practically no data in the literature on the consumption of antioxidants during accelerated aging. Table VI shows the effect of air oven aging a t 100” C. on the phenyl-@napht,liylamine content of GR-S.
Diarylamino radicals such as A do not react with oxygen. Because of resonance stabilization, such radicals are probably too inactive to start a new oxidation chain by removal of a hydrogen atom from the polymer. ACKNOWLEDGMENT
The authors wish t o express their appreciation to L. B. Sebrell and The Goodyear Tire & Rubber Company for permission t o publish this work. They also wish t o thank H. J. Osterhof, S. D. Gehman, C. R. Parks, and D. E. Woodford for helpful suggestiens and assistance. LITERATURE CITED
TABLEVI. EFFECTOF ACCELERATED AGIKG ON PHENYL-13NAPHTHYLAYIXE CONTENT OF GR-S % PBNA in Acetone
% PBNA Hr. a t looo C. 0 24 48 96 400
in GR-S (Kjeldahl) 1.43 1.42 1.29 1.32 1.45
..
I
00.
-CH-CH=CH-CH,
I
OOH
Bolland and Gee, Trans. Faraday Soc., 42,238 (1946). Davies and Sutherland, J . Chem. Phys., 6, 763 (1938). Farmer, Bloomfield, Sundralingam, and Sutton, Trans. Faraday
Sample (Kjeldahl) 0.15 0.37 0.61 0.88 1.01
H
I + GoH?--T\T--CaHb
667 (1943).
%~~~~~~
Extract Nitrotis acid Kjeldahi 1.36 1.35 0.75 .. 0.36 0.12 0.02 0:iz
Oxidation caused a marked decrease in the antioxidant content of the acetone extract. An appreciable quantity of the antioxidant appeared to combine with the polymer. I n Table VI the column under nitrous acid refers to the secondary amine content as determined by titration of the acetone extract with nitrous acid, assuming that phenyl-@-naphthylamine was the only secondary amine present. Since phenyl-P-naphthylamine is not readily attacked by atmospheric oxygen a t 100” C. either i n the crystalline form or in toluene solution, i t would appear that destruction of the secondary amine group may occur as a consequence of termination of the oxidation chain reaction. It is conceded that reaction of the antioxidant with polymer oxidation products may occur t o some extent. If oxidation involves a chain mechanism of the type indicated in reactions 1 to 3, it is suggested that the antioxidant may terminate the oxidation chain in the following way: -CH-CH=CH-CHz-
Barnes, Liddel, and Williams, IND.ENG.CHEM.,ANAL.ED., 15,
+
+ C,~H,-N-C,H, Radical A
Soc., 38, 350 (1942).
Farmer and Sundralingam, J . Chem. Soc., 1942, 121. Ibid., 1943, 119, 125. Farmer, Trans. Faraday SOC.,42, 228 (1946). Feigl, “Qualitative Analysis by Spot Tests”, p. 215, New York, Nordeman Pub. Co., Inc., 1939. Field, Woodford, and Gehman, J . Applied Phys., 17, 388 (1946). George and Walsh, Trans. Faraday SOC.,42, 95 (1946). Harrison and Cole, IND. ENG.CHEM.,36, 702 (1944). Herman, J . Chem. Phys., 8, 257 (1940). Herman and Hofstadter, Ibid., 6, 537 (1938); 7, 461 (1939). Kemp and Straitiff, IND. ENG.CHEM.,36, 707 (1945). Lecomte, Compd. rend., 180, 1481 (1925). Milaa and Surgenor, J . Am. Chem. Soc., 68, 205 (1946). Scott, Trans. Inst. Rubber Ind., 21, 78 (1945). Shelton and Winn, IND. ENG.CHEM.,36, 728 (1944). (21). (22) (23) (24) (25) (26)
Ibid., 38, 72 (1946). Smith and Taylor, J . Chem. Phus., 7, 390 (1939); 8, 543 (1940). Sturgis, Baum, and Vincent, IND.ENCL’CHEM.66, 348 (1944). Taylor and Tobolsky, J . Am. Chem. Soc., 67, 2064 C1945). Tobolsky and Andrews, J . Chem. Phys., 13, 19 (1945). Thompson and Tortington, J . Chem. SOC.,1945, 640. Thompson and Tortington, Trans. Faraday Soc., 41, 252 (1945). UnterBaucher, Ber., 73B, 391 (1940).
PRESENTED before the Division of Rubber Chemistry a t the 109th Meeting of the AMERICAN CHEMICAL SOCIETY,Atlantic City, N. J. T h e work re-
(9)
ported here, contribution 135 from the Goodyear research laboratory, was done in connection with the Government research program on synthetic rubber under contract with the Office of Rubber Reserve, Reaonstruction Finance Corporation.
