Journal of Research of the National Bureau of Standards Vo!' 60, No.5, May 1958 Research Paper 2863 Heat and Ultraviolet Aging of Poly{vinyl chlorider Charles F. Bersch, Mary R. Harvey, and Bernard G. Achhammer
F?ur poly~vinyl chloride) polymers l?repared with different initiators were exposed to ultravIOlet radIant energy and to heat, III a vacuum and in air. The gaseous products evo!ved were analyzed by mass spectrometry. Changes in chemical structure were followed by Illfrared spectrophotometry. Benzene was among the products evolved in most cases and acetone was produced during exposure to heat in air. Catalyst fragments or othe; incorpor~ted. impurities . a ff ~cted i,nitiation at very mild conditions. Susceptibility to degradatIon Illcreased with Illcreasmg oxygen content and unsaturation of the un treated polymer. Pyrolysis studies indicated a two tage degradation: (1) dehydrochlorination, and (2) decomposition of the resultant polyene chain. Color formation was attributed to both oxidation and conjugated un saturation because exposure in air following exposure in a vacuum, and vice versa, caused bleaching of the degraded polymer. 1. Introduction others have found autocatalysis only in certain atmospheres [1 , 3, 7, 15] . Color formation and los The degradation of poly(vinyl chloride) has been of hydrogen chloride are usually considered as studied extensively [1 to 16] 2 because of the com related, but Wilson [11] uggested that the color mercial importance of this material. The mecha developed on degradation was at least partially due nism of decomposition has not been resolved and the t? the action of light on oxygen-containing imp uri initiation step is especially elusive. Druesdow and t16S. Another so urce of controver y has been Gibbs [7] suggested peroxide and catalyst residues whether color in the degraded polymer is bleached in as possible sites for initiation of degradation. Fox the pre ence of oxygen [5 to 10] . [4] and his coworker proposed random splitting out This paper describes an attempt to clarify some of hydrogen chloride, and Arlman [8, 13] suggested aspects of the degradation process in poly (vinyl end initiation plus random oxidation. Baum and chloride). The effects of several different initiator Wartman [16] reported that the sites for initiation on the stability of poly(vinyl chloride) on expo ure are chain end at about 150° C and tertiary chlorine t? heat and to ultraviolet radiant energy were inves atoms above 190° C. tigated. everal exposure were conducted in a Dehydrochlorination is generally accepted as the vacuum to isolate breakdown of the basic polymer propagation step. An allyl-chloride-type reaction from oxidative degradation. Analyse of the gaseous is considered a the probable mechanism. However, degradation products and the polymer residue were there are other products from and changes occurring made at various stages in the treatments to follow in the polymer. Ylack [15] suggested that, in the the chemical structural changes. Color changes presence of oxygen, dehydrochlorination and con e were studied in the hope of correlating them with the quent development of conjugated double bonds tructural changes or conditions of exposure. compete with oxidation of the conjugated system. Both chain scission and cI'osslinking have been 2. Material and Procedures observed [1, 2, 3, 7], often simultaneously. 2.1. Material The identification, measurement of the amount evolved, and rate of evolution of the products on Table 1 lists the four poly(vinyl chloride) polymers exposure to laboratory weathering or pyrolysis [18 used in this investigation. The polymer catalyzed to 22] are among the most frequently used methods with benzoyl peroxide and 2-azo-bis(isobutyroni for investigating the degradation process. The trile), respectively, were polymerized from vinyl evolution of hydrogen chloride has been studied in a chloride monomer distilled directly from a torage vacuum, oxygen, air, nitrogen, hydrogen and hydro cylinder in to glass tubes containing the catalysts. gen chloride. Wilson [11] found anoth~r product a The tubes containing the monomer, frozen under condensable gas, which he identified as water. 'In liquid nitrogen, were sealed under vacuum. Poly his work on sec-butyl chloride as a prototype of merization was accomplished by immersion of the poly (vinyl chloride), Kenyon [1] found carbon tubes in a water bath at 40 ° C for 3 days. For the dioxide, methane, hydrogen chloride, and masses at gamma-initiated polymer, the vinyl chloride monomer 95, 97 , 110, and 112 among the gaseous products was fractionally distilled under e entially equilib- analyzed by mass spectrometry. Benzene has been TABLE 1.-Description of poly (vinyl chloride) materials identified as a product of pyrolysi [18 to 22]. The complexity of the situation has led to many 1'·PVC ______Vinyl chloride bulk-polymerized in a vacuum by exposure to gamma (00·60) radiation. apparent contradictions. Some investigators have bp·PVO ______Vinyl cbloride bulk·polymerized in a vacuum, witb proposed that the loss of hydrogen chloride is auto initiation by O.IO-mole percent benzoyl peroxide. azo-PVO ______. __ _ Vinyl cbloride bulk-polymerized in a vacuum, with catalytic [3 to 7], others deny this [13], and still initiation by 0.02-mole percent 2·azo-bis(isobuty ronitrile). 1 ,['his paper was presented before the D ivision of Polymer Ohemistry at tbe IOI-PVC ______. ___ _ Geon 101 , supplied by the B. F . Goodrich Chemical l ~Oth meeting of the American Chemical Society in Atlantic Oity New Jersey 00. Sept. 16-21, 1956. ' , , Figures in brackets indicnte literature. referenccs at the end of this paper. 481 rium conditions. A 33-in. column packed with glass Several additional experiments and analyses were helices was used. The middle third of the distillate made. 101- PVC was exposed to 100°C, in air, for was redistilled into glass tubes, which were sealed as 25 lu', in an attempt to induce hydroperoxide forma above, and the tubes were then inserted in a 0.3-curie tion, prior to a sequence of exposures and analyses, cobalt-60 source. The center of each tube was 1.5 as described above. 101- PVC and azo-PVC were in. from the center of the source. The time of exposed to 100° C, in a vacuum, for 200 hr, with exposure was approximately 270 hr. mass spectrometric anal:rses of the gaseous products The three laboratory-prepared polymers were in after 1, 3, 5, 10, 20,30, 50 , 100, and 200 h)'; lOI- PVC the form of solid bars when polymerization was was used in a similar experiment in which ultra completed. These bars were pulverized with a stain violet radiant energy was used in place of h eat. less-steel mortar and pestle under liquid nitrogen. 101- PVC was exposed to 100° C in a vacuum for The fractions of these polymers and of the commercial 150, 288, and 400 hI'; to ultraviolet radiant energy polymer that passed through the U . S. Standard No. at 50 ° C il1 air, for 400 lu' ; and to ultraviolet radiant 325 Sieve (nominal opening 44Jl ) were used for these el1ergy at 45 ° C, in a vacuum, for 150, 288, and experimen ts. 400 lu'. The polymers were studied in powder form. Samples of 101- PVC and 'Y-PVC were pyrolyzed Poly(vinyl chloride) films were not used because they in a small evacuated tube furnace. The gaseous retain the polar sol ven ts necessary to dissolve the products were passed directly into a l1l.ass spectrom polymers. This can interfere with both the mecha eter. The temperature of pyrolysis was raised in nism of degradation and the interpretation of analyti steps from 87 ° to 429 ° C, and the l1l.ass spectra cal data. Pressed pellets of the polymer are not were recorded at each step. The equipmen t, method, satisfactory because they provide a very small surface and some of the results of this pyrolysis have been to volume ratio in addition to being partially de described by Bradt [19 , 20]. graded as a result of the heat required in the pelleting operation. 3. Results A typical sample used for an experiment weighed 3.1. Analysis of Original Polymers from 0.15 to 0.30 g. The choice of sample size was a compromise between the desire for a monoparticle Table 2 shows the results of th e oxygen-content layer in the exposure tube and the necessity for determinations on th e four polymers. The two sufficient pressure from the gaseous degradation methods used are similar, and the results can be products to make mass spectrometric analysis compared directly. practical. Infrared absorption spectra of the untreated 2.2. Equipment polymers are shown in figme 1. The 'Y-PVC polymer (dotted line) shows a very weak absorption The polymer powder samples were exposed, in at 6.27 Il, which is attributed to stretching vibra quartz tubes, to various conditions of heat or ultra tions of conjugated C = C structures [30, p. 31]. violet radiant energy or both, in a vacuum and in The double-bond structUl'e was probably formed as air. The exposure tube, exposure chamber, and a result of some dehydrochlorination of the polymer general procedure have been described previously while exposed to gamma radiation. Mass spectro [24]. The only alteration was the use of F- 1 metric analysis of the gases above the 'Y-PVC fluorescent sunlamps in place of RS sunlamps for the polymer immediately after polymerization showed a ultraviolet radiant energy exposures. Six F- 1lamps small amount of hydrogen chloride. 101- PVC were mounted in banks of three in planes 3 in. above (solid line) has a weak band at 6.051l assigned to and below the polymer samples. C = C stretching of a nonconjugated natme [30, 2.3. Exposures and Analytical Procedures p. 31]. The presence of double bonds in the un treated 10l- PVC may also be due to loss of hy The major experimental work consisted of exposure drogen chloride during polymerization [15, 17]. of the same samples of the foUl' polymers to the fol Azo- PVC (dashed and dotted line) shows a weak lowing tlu'ee conditions, in sequence : (1) Ultraviolet absorption at 9.83Jl, which was not identified, and radiant energy, in a vacuum, at 45° ± 2° C for 100 a generally lower transmittance. There also appears 0 hI'; (2) 100 ± 2° C, in a vacuum, for 100 hI' ; and (3 ) to be a very weak absorption at about 6.27 Jl, as in 100° ± 2° C, in ail', for 100 hI'. The gaseous degra dation products evolved dUl'ing each stage were TABLE 2. Oxygen content of poly(vinyl chloride) polyrners analyzed by mass spectrometry. Color determina tions were made after each treatment, using the Polymer Oxygen Munsell colors of the Intersociety Color Council National Bureau of Standards (ISCC- NBS) Method lO l·PVC ______. __ .. ____ "0%90 -y- P\-C ______._____ b.21 of Designating Colors. Samples of the polymer bp- PVC ______._._. ___ .. ___ b. 65 powder were removed from duplicate tubes after each azo- PVC ______.______b.43 exposure for infrared analysis. The powder removed for infrared analysis was pelleted with potassium aBy the method of Walton, McCulloch, and Smith, J. Rcscarch N B S 40, 443 (1948) RP1SS9. Oxygen content reported bere refers to < 44·1' particles used in bromide [23], and a sodium chloride prism was this investigation; the oxy gen content of the bulk lOl- PVC was 0.18 percent as used to determine the infrared spectrum in the measured by the same method. bBy a modification of t he Unterza ucber direct microdetermination of oxygen, 2- to 15-Jl region. S. Ishihara. 1Illoublis bed .
482 WAVE NUMBERS, em-'
650
80
0~ --- \ w- "--- ' z 60 \ '"
WAVELEN GTH , fL FIG HE 1. I njmred) pectl'a oj the poly (vinyl chloride) polymers studied. ____ lOl-PV O; ______oy-pvc; _____ bp-PVC: ______azo-PVO. 'Y-PVO. In addition to absorptions at 5.57 and 5.64J.L appeared at each stage of the degradation. The assigned to carbonyl vibration in the catalyst colors developed aftcr the exposure to ultraviolet used [30 , p. 111], bp- PVO (dashed line) has a band radiant energy arc shown for purpo es of comparison. at 6.22 J.L attributed to O= C tretching vibrations. The intensity of the color for any sample correlate The band at 10.08J.L is probably due to unreactcd with the total amount of product reported for that catalyst [3 0, p. 106] . sample, but not with the amount of hydrogen An ab orption around 12J.L in the region associated chloride reported for that ample. with hydropel'oxide absorption by Shreve et al. The 101- PVC, exposed at 100 ° C in air for 25 hI' [25] was present in about equal amounts in all the prior to expo nre to ultraviolet radiant energy, polymers. No changes were observed in thi ab gave resul ts esse ntially the same a obtained with sorption either as a result of degradative treatment 101- PVC without pretreaLment. The pretreatment of the polymer or during the attempt to induce apparently did noL induce formation of hydro hydroperoxide formation. Krimm and Liang [31] peroxides in the polymer. assign this band to chain stretching in the polymer. The same samples that were exposed to ultra violet radiant energy as described above were further 3 .2 . Mass Spectrometric Analysis of Degraded exposed to heat in a vacuum. Tho results obtained Polymers with bp-PVC, hown in table 4, arc typical of tho e The gaseous degradation products evolved by the from the other polymers with two exception. The polymers on exposure to ultraviolet radiant energy products from 'Y-PVO were too small to measure are shown in table 3. The 'Y-PVC gave 50 to 80 in the mass spectrometer, and the benzene evolved percent less gaseous products thau the other poly by bp-PVC was about three times as much as wa mers, and showed less discoloration as a result of evolved by the other polymers. the exposure. The benzene and large amount of The mass spectrometer docs not mea nre hydrogen carbon dioxide from the bp-PVC polymer were chloride accurately because of adsorption on the attributed to breakdown of the catalyst used . walls of the expansi.on chamber [26] . The presence Hydrogen was a product of 101- PVC only, and of water vapor and carbon dioxide in the ma spectrometer background and as product of poly TABLE 3. Gaseous products evolved by poly(vinyl chl07'ide) (vinyl chloride) degradation further complicate exposed 10 u ltraviolet radiant energy in a vacuum at 45 0 C for 100 hI' analysis. Since efforts to evaluate these factor were unsuccessful, the values reported for hydrogen Amount evolved pel' gram of sam ple- chloride, water, and carbon dioxide mu t be con- Product T A BLE 4. Gaseous products evolved by bp-PVC polymer lOl-PVO bp-PVC or-P VC exposed to 1000 C in a vacuum for 100 h1' Prior treatment: 100-hr exposure to ultraviolet in a vacuum at 45 ° 0 Micromoles M icromoles "'licromoles Hydrogen chloride ______Water ______4. 3 4. G 0. 9 1.0 1.2 1. 2 Amount evolved pel' gram Carbon dioxide ______1.0 13. 5 1.1 of sa mple- Carbon monoxide and/ or nitrogen ______Product Benzene ______1.1 0. 0 0. 9 0. 0 . 6 . 0 A B H ydrogen ______. 6 . 0 . 0 Flna! color • ______10. 0 YR 8/4 5.0 YR 5/5 7.5 YR 9/2 J\1icromoles .1Vicromoles (pale orange (ligbt brown) Hydrogen chloride ______(pale yellowisb Water ______28. 1 17. 4 yellow) plnk) 1.9 9.1 Oar bon dioxide ______0. 9 1.1 • Accordlng to the Munsell System and the ISCO-:' 2000 650 80 ~ 0 . w ~ 60 ~ 40 z WAVELENGTH,j-L FIGURE 2. I nfrared spectra of lot-PVC subjected to degradation. ___ Untreated; ______100·br exposure to ultraviolet in a vacuum at 45° C.; ______lOO·br exposure to ultraviolet in a vacuum at 45° C, plus lOO·br exposure at 100° C, in a vacuum, plus lOO·hr exposure at 100° C in air. sidered as approximations. Comparison of the The products evolved by 101- PVC after 3 hr of results of experiments A and B in table 4 gives an exposure to ultraviolet radiant energy are shown in indication of the poor reproducibility obtained table 5. Traces of hydrogen chloride were reported when measuring hydrogen chloride and water after 50 and 100 hI' of exposure, and hydrogen chlo simultaneously in the mass spectrometer. The ride was the major product after 200 hI' of exposure. example given is an extreme case for duplicate The results after 3 hI' of exposure to heat are also measurements. listed in table 5. The amount of hydrogen chloride The polymer samples previously exposed to ultra evolved after 200 hI' at 100° C was about 13 times as violet radiant energy and to h eat in a vacuum much as was reported for a 200-hr exposure to ultra were finally exposed to heat in air. The water and violet radiant energy. carbon dioxide frozen from the air by liquid nitrogen Azo-PVC was also exposed to 100° C in a vacuum wh en the air was being pumped from the tubes for 200 hr. Except for considerable amounts of un prior to mass spectrometric analysis masked th ese identified nitriles [27] found in the first 30 hI', the gases as possible degradation products. No hy amounts and types of products were similar to those drogen chloride was detected. About 0.1 micromole found with the other polymers. of benzene per gram of sample was found in the Uninterrupted exposures of 101- PVC to 100° C products of y-PVC and bp-PVC, but not in the in a vacuum and to ultraviolet radiant energy at products of 10l- PVC. Hydrogen was again a prod 45 ° C in a vacuum for 150, 288, and 400 hr resul ted uct of 101- PVC. From 2 to 23 micromoles of ace in hydrogen chloride, water, hydrogen, carbon tone per gram of sample were found in the products monoxide, and carbon dioxide, plus small amounts of all the polymers except y-PVC, where the results of unidentified hydrocarbons. Benzene and toluene were inconclusive because one sample showed no were found after exposure to heat but not after ex acetone and th e other was accidentally contami posure to ultraviolet radiant energy. nated with acetone. 3 .3. Infrared Analysis of Degraded Polymers TABLE 5. Gaseous products evolved by 101-PVC polymer in The infrared absorption spectra obtained after 3-hr exposures to ultraviolet radiant energy and to 1000 C in a vacuum carrying out the progressive exposure series on 10l- PVC are shown in figure 2. Exposure to heat in a vacuum after exposure to ultraviolet produced Amount evolved per gram of sample in 3 hI' of ex· no substantial change in the spectrum. The changes Product posure to- around 3 J.I. reflect compensation difficulties in the OH region due to water absorption in some potas Ultraviolet Heat sium bromide pellets. Exposure to h eat in air re MicTo moles Micromoles moved the weak band at 6.05 fJ. attributed to double Hydrogen ... __ .... ____ ...... ______-- __ _ 0.36 0.20 bonds, and produced a broad carbonyl absorption Water ______2.38 1. 20 Carhon monoxide and/or nitrogen ______0.53 0. 27 band around 5.85 fJ. . The changes observed are of Oxygen ______.00 .02 Carhon dioxide ______.36 .51 about the same order as, or larger than, those oc C,H,, ______.QO .11 curring in the other polymers under the same con C,R,, ______.00 .41 Benzene ______-- ___ - ____ _ Trace . 02 ditions. 'roluene ______------'l"'race . 10 Styrene ______------0. 00 . 04 Figure 3 shows the 5- to 7 -fJ. region of the spectra Alkyl benzene ______.00 .02 of bp-PVC. The original polymer shows absorp tions at 5.57 and 5.64 fJ. attributed to residual benzoyl 484 PVC (solid line) and 100-PVC exposed to ultraviolet radiant energy in air at 50° C for 400 hI' (dashed line). A wide absorption band extending from 2.7 to 4 j./, with shoulders at about 2. 5, 3.03, 3.50, and 3. 77 j./, is present in the spectrum of the degraded material. BO The absorption occurring between 3.7 and 4 j./, is in dicative of the OR stretching vibration of strongly hydrogen-bonded OR groups [30, p. 142]. The ab sorption around 2.7 to 2.85 j./, has been assigned to the OR valence stretching of thc unbonded hydroxyl 60 group [30, p. 85], and the ab orption in the 2.85- to 3. 15-j./, ;,!2 range may be caused by the OR vibration of 0. w intermolec ular and intramolecular hydrogen bonds. u Z The broad, intense carbonyl absorption in the 5.8- WAV E NUMBERS, em·' 650 100 8 0 ~ "\ !'~' "- 0 I - ~ \ --- w· I \ <.) \ ",...I " /' z 60 \ ( \ 0 2 WAVELENGTH ,flo F I GURE 4. I nfrared spectra of JOl-PVC before and after exposure to 400 hr of ultmviolet in air at 50° C. _ _ _ Untreated; ______treated. 485 WAVE NUMBE RS, em-I 650 100 80 o_~ w u z 60 ~ ':: ::!! 0 2 WA VELENGTH , fL FIGU RE 5. I nfrared spectra of 101-PVC subjected to heat degradation. ----- Untreated; ______150-hr exposure at 100° 0 in a vacuum; ______400-hr exposu re at 100° C in a vacuum. T ABL E 6_ C0101'S developed in degmded poly (vinyl chloride) 3 .5. Pyrolytic Degradation of lOl-PVC and 'Y-PVC I J\lIwlscll des ignation and color a after expo sure to- The relative ratios of the major products evolved Polymer i during continuous pyrolysis of -y-PVC and 101-PVC 100 hr of ultraviolet Plus 100 hr at 100' C Plus 100 hI' at 100° C in a vacuum in a vaCUUln in air at various temperatures as analyzed by mass spec trometry are given in table 7. A small amount of 101-PYC ___ I 10.0 YR 8/4 2.5 R 3 ..5 /4 7.5 YR 7/4 hydrogen chloride was evolved from 101-PVC at (pale orange yellow) (grayish red-dark (ligh t yellowish red) hrown) 127 0 C. Benzene was a product of both polymers 0 'Y-P\·C __ ___ I 7.5 YR 9/2 10.0 R 7 ..5 /4 7.5 YR 9/2 at about 175 0 , along with relatively large amounts (pale yellowish (moderate yellowish (pale yellowish pink) pink) pink) of hydrogen chloride. The maximum evolution of hydrogen chloride from 101-PVC occurred at 220 0 C, bP-PYC ____ 1 5.0 YR 5/5 7.5 P 7/4 10 .0 R 8/2 0 (l ight brown) (pale purple) (pale yellowish whereas for -y-PVC the maximum was at 238 C. I pink-grayish yel- The largest evolution of benzene and measurable lowish pink) amounts of mass 91 , naphthalene, and anthracene azo-PY C ___ I 7.5 YR 8/3.5 7.5 YR 7/4 10 .0 YR 8/4 0 0 (pale orange yellow) (light yellowish (pale orange yellow) were observed in the 220 to 240 C range. Above brown) 300 0 C very little hydrogen chloride was evolved, and mass 91 became the major product. Mass 91 a The I SCC (Intcrsociety Celor Conncil)-NBS l ethod of Designation Oolor. represents the principal decomposition product of alkyl-aromatic compounds. It was accompanied by PVC discolored most on exposure to ultraviolet radi a large number of different hydrocarbons of varying 0 ant energy and 101-PVC discolored most on subse degrees of saturation. > At 389 C there were products quent exposure to heat. Bp-PVC bleached slightly of every mass number from 78 to 600, the limit of and changed color on exposure to heat after exposure the resolving power of the mass spectrometer. to ultraviolet radiant energy. Beyond mass 200 the amounts of each product were TABLE 7. Ga seous products evolved by 101-PVC and -y-PVC during pyrolysis 'Y-PVC lOl-PVC P eak heights, in scale divisions for Peak heights, in scale divisions for mass numbers- mass numbers- L argest m ass Probable Temp Temp number composi- tion ___I 36 78 91 36 78 91 128 178 I ' C ° 0 127 7 172 24 0.5 179 218 12.5 ]90 164 27 ]94 440 48 128 C IOR , 214 736 52 220 2,294 244 4 14 3. 5 179 CuR" 238 2, 416 94 1 251 338 6 1 250 1, 01 5 45 284 18 9 14 280 1, 290 43 13 13.5 2. 5 179 CuHu 306 126 51 72 40 10. 5 192 OisHIZ 33 1 4 4 26 335 59 37 123 67 26 253 C" R u 38 1 1.8 3.9 26 389 3.5 11.5 89 21 31 596 486 about equal, and between mass 400 and mass 600 was much more stable than tbe other polymers, iL i the spectrum resembled a sine wave with maxima possible that incorporated impuriLies arc a major at all the even mass numbers. factor in in tabili tv under normal service co ndiLion s. The ')'-PVC sample was smaller than the 101-PVC Initiation or degradaLion of 'Y-PVC was probabl.v sample, and this is reflected in the lower values for due to oxygenated or unsaturated sL ructures in Lhe ')'-PVC at each step. polymer. If this is so, it might be said Lhat Lhe earlier and greater instabiliLy of Lhe other polymer 4. Discussion and Conclusions was solely due to the greater amount of oxygenation and unsaturation in these polymers. BoLh were The stability of poly(vinyl chloride) is a function probably of considerable importance, but the very of: (1) the structure of the polymer, including early occurrence of the decomposition products structural deviations and chemically or physically from incorporated impurities and the fact that Lhe incorporated impurities; (2) the processinO' and incorporated impurity was largely responsible for storage history of the polymer; and (3) the conditions the high oxygen content in bp-PVC seem to indicate of exposure, including the ambient atmosphere :md that impurities playa large role in mild conditions of the wavelength and intensity of the incident energy. exposure. The failure of other investigators to find Recognition of these factors, combined with the any effect from incorporated impurities may be at results obtained in this investigation, lead to possible tributed to the transitory nature of the reaction explanations that may account for some of the and to the possibility that at the more stringenL observations r ecorded here and in the literature. conditions usual in degradation studies, other factors 4.1. Initiation such as unsaturation assume greater imporLance. Thus, at relatively low LemperaLures some labil e Initiation of degradation probably occurs at several groups may become active, and a the temperatme different sites, the particular polymer and the condi is raised other labile groups, possibl y more pre tions of exposure being determining factors. The dominant but relaLively dormant at the lower tem. susceptibility to degradation of the polymers studied peratw'e, m.ay become so active as Lo obscUl'e Lhe increased both with the oxygen contents of the effects of the less prevalent structures. A SOlr.e polymers and with the initial unsaturation. The what parallel situation occurred when Baum and effects of the two were not separated, but both are Wartman found unsaturation of primary importance probably factors of some importance in initiation. at 1.50 0 C and LerLiary chlorine assumin g a larger Several investigators [7 , 8, 13] have proposed that role when degradaLion was carried ouL at 190 0 C or oxygen contained in a polymer is a factor in degrada higher. tion, but the actual mechanism by which that The type and conditions of expo ure, especially oxidation, presumably occurring during polymeriza the wavelengLh and intensity of the incident e n e r~y, tion, contributes to the decomposition of thc polymer have specifi.c degraclative eH'ecLs dependent on the has not been isolated. .Although the oxygen con chemical strucLure of the polymer. For example, tents of the polymers studied here may in part be bp-PVC is more susceptible to ultraviolet radiant due to pOl'oxidic structures, none were found; neither cnerg,\T, and the lOl-PVC a nd azo-PVC polymer lVere any other oxygenated structures except in the are more susceptible to heat. case of bp-PVC where an excess . of an oxygen containing catalyst was apparent both in the infrared 4.2. Propagation spectra and in the gaseous degradation products. After initiation the degradaLion of poly(vinyl The presence of unsaturation in undegraded chloride) proceeds in a complex manner. The most poly(vinyl chloride) has been established by others obvious feature is dehydrochlori naL ion, which Adman [8 , 9, 16]. Baum and IVartman [16] have shown [13] has shown to be a free-radi cal reacLion. Py roly that it occurs at chain ends, probably as a r esult of sis indicates that Lwo stages actually occur in tbe clisproportionation, and is a major fa ctor in initiation degradation of poly(vinyl chloride), exclusive of of degradation at 150 0 C. oxidation reactions. DehydrochlorinaLion leaves a The presence of catalyst fragments has also been polyene backbone that further decom.poses into a proposed as providing sites for initiation of degrada mixture of aliphatic, aromatic, and alkyl-aromatic tion [7]. The importance of impurities incorporated materials. The first such product observed was chemically or physically was illustrated here by the benzene, possibly formed by cyclization of unsatu gaseous products that resulted from the decomposi rated chain ends [21]. 'rhe eady appearance of tion of catalysts, soap, etc. when 101-PVC and benzene as a product of degradation uncleI' mild no-PVC were exposed to heat in a vacuum for conditions seems to confirm the mechanism of short periods of time and when bp-PVC was ex cyclization and also is another indication of end ])osed to ultraviolet radiant energy in a vacuum, initiation and a zipper-type r eaction. The failure even over extended periods of time. In the latter to produce more than a small am.ount of benzene is case the decomposiLion of the catalyst could be fol attributed to deCl'eased mobility of the chains as lowed in the infrared absorption spectra as well as unsaturation proceeds and crosslinking occurs. by analysis of the gaseous degradation prod.ucts. Ultimately, under pyrolytic conditions, sufficient Since the exposure conditions were little more strin energy is available for random C- C rupture. This gent than actual service conditions, and since accounts for the multitude of other carbon com 'Y-PVC, which did not contain any catalyst residue, pounds produced in pyrolysis. 461441-58--5 487 In degradation under service conditions oxygen phores is subjected to heat in a vacuum, bleaching may affect both stages of the degradation of poly will also occur as well as a change in hue. A shift (vinyl chloride). Oxidation of both saturated and in the carbonyl absorption occurs at this stage. unsaturated portions of the polymer chain may lead Probably the oxygenated chromophore is removed to chain scission and to an increased concentration slowly, and if the exposure were carried far enough of free radicals, which may accelerate dehydro the color would again increase in intensity. Ultra chlorination. This has been observed frequently violet radiant energy is probably necessary to [1, 3, 5, 7]. The presence of ultraviolet radiant produce oxygenated chromophorcs. energy accelerates oxidation reactions. The production of acetone is probably due to a The authors are indebted to the :Mass Spectrome m echanism similar to that found by George and try Section for the mass spectrometric analyses, to Walsh [28, 29]. This would result in chain scission. the Organic Chemistry Sectioll and to Shigeru Whether the necessary oxygen to form the acetone Ishihara for the oxygen-con tent determinations, and came from the ambient atmosphere or from the oxygen to Shigeru Ishihara. for assistance in some of the contained in the polymers was not determined. The experimental investigations. polymers with higher oxygen contents produced more acetone but acetone only appeared after exposure in air. 5. References The mechanism by which dehydrochlorination [1] A. S. Kenyon, Sym.oosium on polymer degradation mechanisms, NBS Cire. 525, p. 81 (1953). occurs has generally been accepted to be of an allylic [2] M . Imoto and T . Otsu, J. Inst. Poly tech., Osaka City nature [2, 4, 5, 7,9], that is, a chlorine atom in the Univ. [C] 4, 269 (1953). beta position to a double bond requires less energy [3] M . Irnoto and T . Otsu, J . Inst. Poly tech., Osaka City for removal. However, many of the proposed Univ. [C] 4, 124 (1953) ; J. Chern. Soc. Japan (Ind. Chem. Sec.) 54, 771 (1951) . mechanisms were not presented as free-radical chain [4] V. W. Fox, J. G. H endricks, and H . J . Ratti, Ind. Eng. reactions, though the OCCUlTence of a free-radical Chem. 41, 1774 (1949). reaction has been established [12, 13]. The mech [5] A. L. Scarbrough, V;V. L. Kellner, and P. 'V. R izzo, anism proposed by Stromberg et al [21] combines Modern Plastics 29, 111 (1952); NBS Circ. 525, p. 95 (1953) . allylic excitation with a free-radical reaction and can [6] C. B. Havens, NBS Circ. 525, p. 107 (1953). be made to fit any or all of the proposed mechanisms [7] D . Druesdow and C. F. Gibbs, NBS Cire. 525, p. 69 of initiation. It also allows for the effects of oxida (1953); Modern Plastics 30, 123 (1953). tion in that oxygenated radicals as well as a chlorine [8] E. J . Arlman, J . Polymer Sci. 12,547 (1954). [9] L. H. Wartma n, Ind. Eng. Chem. 47,1013 (1955). radical freed by allylic excitation may initiate a [10] C. S. Marvel, J. I-I. Sample, and M . F . Roy, J. Am. chain reaction. One minor modification suggested Chem. Soc. 61, 3241 (1939). is that removal of a hydrogen atom to start a chain [11] J . E. Wilson, J . Am. Chem. Soc. 72. 2905 (1950). reaction need not necessarily occur adjacent to a [12] J . Furukawa and T. Kimara, J. CI18I11. Soc. Japan (Ind. Chern. Sec.) 55,133 (1952) . double bond. [13] E. J . Arlman, J. Polymer Sci. 12,543 (1954). Crosslinking is possible whenever two chains [14] H . Batzer and A. Nisch, Makromol. Chem. 16, 69 (1955). with free radicals are in proximity. A moderately [15] G. P . Mack, Modern Plastics 31, 150 (1 953). frequent occurrence of this nature is conceivable. [16] B. Baum and L. H. vVartman, Paper presented at North Jersey Section, American Chemical Society Meeting Scis'lion probably occurs principally as a result of in Miniature, Jan. 28, 1957. oxidative degradation except in pyrolysis, where [17] S. Wakano and K. Yamamoto, J . Chem. Soc. Japan there is usually sufficient energy available to rupture (Ind. Chem. Sec.) 54, 513 (1951). C-C bonds. [18] P. D . Zemany, Anal. Chem. 24, 1709 (1952) . 4.3. Color [I9] P . Bradt, V. H. Dibeler, and F. L. Mohler, J . Research NBS 50, 201 (1953) RP241O. T he degradation of poly (vinyl chloride) is ac [20] P. Bradt and F . L. :Mohler, J . Research XBS 55, 323 (1955) RP2637. companied by a discoloration that has been at [2J] R. R. Stromberg, S. Straus, and B. G. Achhammer, tributed to the polyene system. It was observed by Paper presen ted at ACS National YIeeting, Atlantic som.e investigators that exposure of degraded poly City, N . J ., Sept. 20, 1956. [22] R. R. Stromberg, S. Straus, and B. G. Achhammer, (vinyl chloride) to oxygen led to bleaching of the J . Research NBS 60, 147 (1958) RP2832. color, whereas others failed to find this occurring. [23] M. R . Harvey, J. E. Stewart, and B. G. Achhammcr, In view of the results obtained in this study, it seems J. Research NBS 56, 225 (1956) RP2670. that two sources of chromophores are possible, from [2-!] B. G. Achhammer, M. J . Reiney, L. A. Wall , and F . W. R einhart, J. Polymer Sci. 8,555 (1952). conjugated unsaturation and from oxygenated struc [25] O. D . Shreve, M . R. H eether, H . B. Knight, and D . tures. If a polymer is degraded in a vacuum and Swern, Anal. Chem. 23, 282 (1951) . exposed to air, the color due to unsaturation may be [26] B. G. Achhammer, Anal. Chem. 2<1, 1925 (1952). bleached. A carbonyl absorption in the polymer [27] F. M. Lewis and M . S. Mat heson, J. Am. Chem. Soc. 71,747 (1949). . occurs simultaneously. In the case of degradation [28] P . George and A. D . v;Valsh, Trans. Faraday Soc. 42, 9-1, on exposure to ultraviolet radiant energy with (1946) . oxygen present, color formation seems to be due [29] A. D . Walsh, Trans. Faraday Soc. 42, 269 (1946). either to oxygenation only or to both types of [30] L. J . Bellamy, The Infra-red spectra of complex mole cules (John Wiley & Sons, Inc., New York, N. Y. chromophores. Absorptions clue to carbonyls and 1954) . conjugated unsaturation may be found. Further [31] S. Krimm and C. Y. Liang, J. Polymer Sci. 22, 95 (1956). exposure to oxygen cloes not lead to bleaching. On tbe other hand, if a polymer containing both chromo- WASHINGTON, Dec ~ mber 9, 1957. 488