<<

1 Journal of Research of the National Bureau of Standards Vol. 60, No. 2, February 1958 Research Paper 2832

Infrared Spectra of Thermally Degraded Poly(Vinyl Chloride} *

Robert R. Stromberg, Sidney Straus, and Bernard G . Achhammer

The cha nges in chemical structure occuring in poly (vinyl chloride) as a res ul t of heating ina vacuum in t he range 100° to 400° C were studied using infrared spectrophoto metry. The principa l changes occurri ng in t he residue during p yrolysis in a vacuum were the forma­ t ion of unsaturated structures a nd a change from an aliphatic spectrum to one showing aromatic absorption. The data are used to support a previously proposed mechanism of decomposition for poly(vinyl chloride).

1. Introduction 2. Experimental Procedure

The kine Li cs of the thermal decomposition of Three pol~- (v in~- l chloride) polymers were studied. l)oly(vinyl chloride) in a vacuum have been pre­ Th e ~- were prepared from vinyl chloride monomer a viously reported [1).1 Under pyrolytic conditions, follows: pol.v(viuyl chloride) decomposes predominantely into hydrogen c hloride and a colorcd solid residue. ,--P Olymer The color of the residue has been attributed to a Preparation polyene structure exi ting after the d eh~- dro chloriua­ tion [2] . The role of in the decomposition PVC- /, ___ _ Init iator- /, radiation from 0.3 Curie m echanism has not ~r et been resolved although Co- 50 source PVC- bp ___ _ Init iator- O.l mole percent of bcnzo.d hydroperoxide m echanisms have been proposed [3]. pero xid e; 40° C r R ecently group theory has been applied to the PVC- [l,ZO __ _ Initiator- 0.02 m ole percent of 2,2' interpretation of the infrared spectra of poly(vinyl azobis (isobutyronitrije); 40° C chloride) and the spectrum has been analyzed in a comprehensive manner [4]. The head-to-tail ar­ rangement of the monomer units [5], and the alter­ All the polym er samples were powdered by grinding nate arrangements of the chlorine atoms in the plane under liquid nitrogen . The white powders were of the carbon chain [6], with the conclusion that the sieved through a 325 mesh screen to a particle siz e crystallographic repeating unit consists of two less than 44 microns in diameter . monomer units, were further verified b~1 the work of Specimens of these polymer were pyrolyzed in a Krimm and Liang [4]. Polarized spectra were used vacuum for 30 min at temperatures up to 400° C by Krimm and Liang to make band assignments. after outgassing at 110° C for approximatcl~T 1 hour. Other, more limited analyses of the infrared spectra At temperatures above 180° C volatile products were of poly (vinyl chloride) have also been reported [7 , 8]. evolved . The color of the residue changed to li ght The branching of poly (vinyl chloride) has been brown at the lower pyrolysis temperatures and to measured [9J by determining the methyl group con­ dark brown at the more elevated temperatures. centration of the reduced polymer. The application After the exposure to pyrolytic conditions wa of infrared spectrophotometry to the explanation of completed , the degraded powder samples were re­ the degradation process occurring in poly (vinyl moved and prepared for infrared analysis. 1Jntreated chloride) has, however, not been extensively reported poly(vinyl chloride) is soluble in only a few solvents, in the literature. Campbell and Rauscher [3] which include cyclohexanone, tetrabydrofuran, di­ u tilized changes in the infrared spectra of poly(vinyl methyl formamide and to a limited extent, . chloride) in a study of the base-accelerated degrada­ Some of the degraded material had an even more t ion of this polymer. Others [10, 11] limited their limited solubility. Cyclobexanone and tetrahydro­ studies to changes in only one portion of the spec­ furan tend to form hydroperoxicLes, which decompose trum. easily and may initiate decomposition of the polymer. This paper describes the results of an investigation If these solvents are retained in the polymer film, in which infrared spectrophotometry was used to they also can lead to erroneous conclusions about study changes in chemical structure occurring in the the structure of the polymer. This is especially true solid residue during p~'r o l ysis of poly (vinyl chloride) when hydroperoxide breaks down to give carbonyl in vacuum. and hydroxyl groups. Therefore, the solid phase pellet m ethod was used to prepare the material for . Presen ted at the 8th Ammal Pittsburgb Conference of Analytical Chemistry infrared analysis. Potassium bromide (Harshaw and Applied Spectroscop y. Mardh 1957. I Figures ill brackets indicate the literature refe rences at the end of this pape.-. Chemical Co.) was used for the suspending medium 147 because it has a refractive index, n D of l.56 whi ch approximates that of untreated poly(vinyl chloride) (l.55) . The polym er and th e potassium bromide in the ratio of 1 to 50 were mixed in a dry box and the pellets wer e prepared in the manner described pre­ viously [12]. Although the ratio by weigh t of polym er to potassium bromide was th e same for all samples, there was an increase in the molal' ratio of degraded polym er to potassium bromide as a result of the loss of h~T dl'o gen chloride. A P erkin-Elmer :YIodel 2] double-beam infrared spectrophotometer with a sodium chloride prism was used for m easurements in the 5000 to 670 cm- I region. For increased resolution a P erkin-Elmer YIodell2- B instrumen t wi th a lithium fluoride prism was used for measuremrnts in the 3000 to 26C Ocm - 1 regIOn . 2919 F I GURE 3 , POT/ion oj infraTed spec/rum cf Geon-l Ol PIT 3 . Infrared Spectra and Discussion using lithium filw1'il:e prism.

3.1. Untreated Polymers

The infrared sprctra of th e three poly(vinyl chlo­ ride) polymers studied are shown in fi gure l. These T A B LE L .1ssign1llents oj absorption bands /07' untuated patterns wcrc obtain ed by placing a potassium bro­ poly(vinyl chl'JI'ide) mide pellet containing the un treated poly(vinyl chloride) in the sample beam and a blank potassium bromide pellet in the reference beam. A commercial Freque ncy As~ i g n mem I References I 2 , polym er, Geon 101 was studied with a lithium cm- 1 fluoride prism in the region of 3000 to 2820 em - I, 690 0 - 0 1 stretch"""""""""",."",."""" 34 833 chain stretch"""""""""""".""""", 4 This spectrum is given in fig ure 2. 963 cba in stretch".".""""""""".""."",,, 1 1095 perpend icuhr chain stretch ." , .. " ... ,,"""" 4 The assignments of the principle absorption bands 11 20 parallel chain stretch .. , ''".''''''''''"'''''''' 4 of th e basic polymer are given in table 1. Krimm 1200 OH wag, out of phasc " .. "". ,,"""""""'" 4 1250 OH bend, in phase"""""""""".,,._""" 4 and Liang [4] reported a bsorptions in the region of 1330 OH be nel, out of phase, ""."".,,"",,"""" 4 1 1352 OH , wag, in phase ____ ,, ______. 4 3000 cm - that varied slightly from those given in 1427 OH , be nd, in phasc ____ ,, __ . __ ,, __ ,, __ "' ~,, ____ , ~.4 table 1 and fi gure 2. It is possible that this sm all 2823 OH stretch, in phase __ "" __ ,,,,""""______4 2854 OH , stretch, symmetrical, in phase " __ ",, ____ ,, 4,13 difference is a result of th e nature and perhaps 29 19 Clb stretch, asym metrical, in phase ______4. 13 2977 on stretch , ou t of phase, on n ei~ h bori n g OHOI scatter of the pellet specimen. Krimm and Liang groups""" " __ .,, __ " "'" ,,,.,,"''''''''''

2 B. F . Goodrich Ohemical 00.

650 100

80

0~ w ~ 60

0 2 8 9 WAVELENGTH, fL

FIGUR E 1. Inf rared spectra of the poly(vinyl chloride) polymers stu died . . . , , . , PVC-/, - ---- PVC-bp - . . . - .. . - PVC- azo

148 obtained their spectra from a film and the spectrum the order of 300 0 C and higher , there is a marked shown in figure 2 was obtained from a pellet. over-all departure from the spectrum of poly(vinyl A weak absorption band at 1590 cm- 1 is observed chloride) . It was shown in a previous study ill in the spectrum of the PVCr sample. This is that at approximately 300 0 C practically all of the assigned to a carbon- carbon double bond stretching chloride atoms have been removed from the polymer vibration [13]. The double bond structure is as HCl without any apparen t appreciable break­ formed as a result of some dehydrochlorination down of the polymer chain . This removal of the [1] of the polymer while exposed to the 'Y-radiation. highly elec tronegative halogen from the chain could The curve obtained from the PVC-bp sample cause frequenc:v shifts in some absorption bands as exhibits several absorp tion bands not observed in well as other spectral changes r es ulting from t he the other polymers. The bands at 1795 cm-1 and new structural configurations. Substantial break­ 1773 cm- 1 are assigned to the carbonyl vibration of down of the polymer chain occurs at 400 0 C [1] . t he catalyst, benzoyl peroxide [1 31. The band at In the region of the carbon-·hydrogen stretching 1733 cm - 1 is probably from t he carbonyl stretching absorption, 3000 cm- l, a new band is formed at vibration of an or [1 3], formed from 3030 cm- l that mav be associated with th e carbon­ interaction of the catalyst with vinyl chloride. hydrogen str etchilig vibration from an unsaturat,ed Benzoyl peroxide shows an absorption [13] near carbon- carbon linkage [13]. The intensity of thi s 1000 cm - 1 and the band at 990 cm- l is probably band increases wi th increasing temperatures of from um-eaeted catalyst. PYTolysis. N onco nj ugated carbon- carbon double bonds arc s uggested by Lhe absorpLion at 1670 3 .2. Spectra of Poly(Vinyl Chloride) After Pyrolysis cm- l [13]. The absorption band aL ]6]0 cm- 1 is assigned to Lhe carbon- carbon double bond stretch­ Figure 3 shows the spectra of PVC-bp res idue ing vibration for conjugated bonds, eiLhe !" aromat:c after pyrolysis for 30 min in a vacuum at several or aliphatic or both [13]. temperatures . The spectra arc progressively altered The shoulder at 1590 ClU- l is assigned Lo Lhe car­ as a result of increasing the pyrolysis temperature. bon- carbon double bond vibration of an aromatic There is considerable radiation scatter for t he pellets ring [131. A reso na nce splitting of the absorpLion of the samples pyrolyzed at higher LemperaLures bands with t he appearance of one ba nd ncar 1600 wbere t he rcfracti ve index of the more highly de­ cm- 1 and one ncar] 500 cm- l occu rs wh en t here arc graded colored specimens did not match that of three unsaturated carbon- carbon vibrations in a the potassium bromide. The samples were pre­ ring [13 ]. The bands at 1610 and 1500 c:,n - l nrc there­ treated by heating at 110 0 C in a vacuum for fore assigned , at least in part, to an aromatic struc­ approximately 1 hour. The only changes observed ture. The broad absorption bands at 880, 820, 750, in the spectrum of t he materi al after this Lreatm ent and 700 cm- l arc present in the spectra of asphalts, was the disappearance of the peroxide bands at beavy petroleum oil -fractions, coal, eLc., r14] and 1795 and 1773 cm- 1 , indicating r emoval of th e indicate the prese nce of aromatic structures, ,,,,ith unreaeted catalyst by th i pretreatm en t . At tem­ the latter two indicatin g carbon- hydrogen out-of­ peratures up through 250 0 C, the infrared pattern , plane deformations of monosubstituted aromatic although somewhat alter ed, is easily iden tifLed as groups [13]. The absence of aromatic structural that of poly(vinyl chloride). At temperatures of bands in t he untreated poly(vin yl chloride) and

~ w ~ 60 « l- I- ::;: (j) 4 0 Z « a:: I- 20

o 2~----~-----7----~----~~----~----~----~----~~--~L-----~----~----~~--~~

F l GUHE 3. Infmred spectm of PVC-bp after 30 minutes' pyrolysis. -- Unt reated - -- 2550 C - . . . - ... - 3000 C ...... 400 0 C 149 their appearance and increase in intensity as a presence of the conjugated aromatic structure is res ult of higher degradative temperatures can be revealed in the spectra of the specimens pyrolyzed readily observed. at 342 0 and 400 0 C by the bands a t 880, 820, 750, The absorption at 1720 cm- 1 indicates the pres­ 700 cm- I and the shoulder at 1490 cm- I . There is ence of carbonyl vibrations. This band did not also absorption caused by aromatic structure in the appear to any appreciable extent when the pyrolyzed spectrum of the specimen pyrolyzed at 275 0 C. material was stored under an inert atmosphere. As The spectra of the pyrolyzed PVC- azo sample is this spectral change does not occur at room tempera­ shown in figure 5. The same general observations ture in untreated polymer, it appears that the made with the PVC- bp and PVC--y polymers also double bonds in the degraded polymer are subj ect apply to this material. However, this sample is to attack by atmospheric oxygen, forming such observed to have undergone more degradation at a oxygenated structures. given temperature than the other two polymers did. The shifting of the carbon- hydrogen vibration For example, no C- Cl stretching band at 690 cm- 1 of CH2 [3] from 1430 cm- 1 in the spectrum of un­ remains in the specimen pyrolyzed at 253 0 C, while treated polym er t o 1455 cm- I for the more highly this band is relatively intense for the PVC- bp pol~T ­ decomposed samples may be a result of a loss of mer, pyrolyzed at the same temperature. This is in the chlorine atoms or an increase in the CH3 concen­ agreement with the faster rate of decomposition, as tration [1 3]. The band at 1380 cm- I in the most measured by rate of weight loss for this material as bighly degraded material may be assigned to the previously reported [1] . This faster rate was attrib­ carbon- hydrogen deformation frequency for the uted to unreacted catalyst remaining in the sample. aliphatic C- CH3 bond [1 3]. The pyrolysis of poly (vinyl chloride) produces The removal of the band at 690 cm- I, assigned to volatile products and a solid residue. At 400 0 C tbe C- Cl stretching vibration [3, 4], from the some of these products are volatile at the tempera­ spectrum of the sample pyrolyzed at 300 0 C corre­ ture of pyrolysis but not at room temperature. lates with mass spectrometer and weight loss m ea­ This portion is a brown, waxlike material at room surements [1] which showed that almost all the temperature. The spectrum of this wax is similar chlorine atoms wer e r emoved from the polymer at to that of the more highly degraded specimens ob­ this temperature. served earlier and probabl~T consists of fragments of The infrared spectra of the PVC, polymer after the dehydrochlorinated polymer. The C- Cl band pyrolysis at several temperatures are shown in figure at 690 cm- I is absent, accounting for a flat spectrum 4. The changes in the spectra as a result of increas­ observed at the lower wave numbers. A broad ing the pyrolysis temperature are similar to those carbonyl band at 1730 cm- I is probably the result observed for the PVC-bp material. The spectrum of the oxidation of double bond structures following of the specimen degraded at 250 0 C can be readily the pyrolysis. The CH vibration [3] in CH2 is identified as poly(vinyl chloride) . Increasing the shifted from 1430 cm- I in the original material to t emperature to 275 0 C gives a spectrum, from 1400 1470 cm- I . A band at 1380 cm- I , assigned to the to 650 cm- 1 that is different from the spectrum of C- H deformation for the aliphatic C- CH3 bond poly(vinyl chloride). This is probably caused by [1 3], is present in this spectrum as well as in the the almost complete removal of chlorine atoms as spectrum of the residue of the material pyrolyzed a t shown by the absence of a band a t 690 cm- I . The 400 0 C. Although there are weak bands at 700,

80

0 L-____ 2 ~----~---~5 ----~~----~----~----~----~~6 8 9 10 --~~----~----I I 12 ~I13 ----~~14 --~1~I5 .~ WAV ELENGTH • f-L

F I GU R E 4. I nfmred spectm of P VC- -y after so minutes' pyrolysis. ---- 2500 C ---- 2750 C - ... - . . . - 340 0 C ...... 400 0 C 150 650

80 o~

""~ 60

O~--~-----+----~----~----+-----~--~~--~'---~----oT----~----~---,~

F,f:um; 5. InIraTed spectra of PVC- azo after 30 minutes' pYI·olysis. ---- 220 0 C - ... - . . . - 255 0 C ...... '100 0 C

720, and 755 cm- I , the relatively intense bands the oxygenated strucLmes observed. This action assigned to aromatic stru ctures in the spectrum of may further degrade the polymer by reducing the the residue of the material pyrolyzed at 400 0 0 size of the chain. The infrared technique was, are absent from this spectrum .. The differences however, insensitive to measurement of chain between the spectrum of the resIdue and the spec­ length, although the apparent rise in methyl group tram of the waxlike material are attributed to the concentration is an indicatioll of an increascd num ber different molecular weights of these portions. of chain ends. The formation of aromatic structures in the 4. Conclusions spectra of the degraded aliphatic poly(villyl chloridc) correlates with the presence of benzene in th e Specimens pyrolyzed in a vacuum at temperatures volatile products of such decomposition reactions [1]. up to 400 0 0 yielded some spectra that were decidedly This indicates that an aliphatic conjugated structure different from the original material and which varied is presen t at chain ends. This structure can oceLir from each other depending upon the temperature. as a result of dehydrochlorination at the original The volatile products obtained from the poly(vinyl ends, either by initiation or termination, or by sissiol l chloride) polymers at temperatures up to 300 0 0 of an internal portion of a dehydrochlorinated chain . consisted primarily of HOI and a few mole percent The dehydrochlorination reaction that. has b.een of benzene [1]. At 400 0 0, the volatile products previoll sly proposed [1] would allow a free produced from the dehydroehlorinated polymer at the end of a polyene group in either instance. consisted, in addition, of larger quantities of benzene As a resul t of kinetic mo tion, variolls c~ ' eli c 0011- and various hydrocarbons. The principal altera­ ftgurations are formecl. Because of resonance con­ tions in the infrared spectra of these degraded siderations, benzene is the most stable structm c, specimens were loss of chlorine atoms, with complete and as a result of this stabilization, the carbon­ removal at temperatures near 3000 C, establishment carbon bond alpha to the benzene ring is weakened of double bond character, both aliphatic and aro­ and breaks, evolving benzene as a product. The matic, and a possible increase in OH3 concentra­ more complete removal of chlorine at the highcr tion. A variety of carbonyl structures is formed temperatmes, together with the increased opportu­ when the degraded material is allowed to stand nity for chain cleavage at these elevated tempera­ exposed to the air for several days at room tem­ tures, yiclds a more aromatic structure. N[oreovcr, perature. Within the limits of this technique the the material that is volatile at the temperaturc of spectral changes are independent of polymerization pyrolysis but not at room temperature apparently initiators used. consists of short aliphatic segmen ts of the rcsidu c. The mechanism of the decomposition appears to be a removal of chlorline as HCI in sllch a manner as to form a conjugated polyene structure, even when only small quantities of chlorine have been The authors express appreciation to the following removed. No infrared evid ence was found for members of the National Bureau of Standards: hydroperoxides . It appears reasonable, therefore, James E. Stewart for hi suggestions regarding some that oxygen does not play a direct role in ~he of the interpretations and for obtaining some of the dehydrochlorination process of the systems st ud~ e d spectra, and Mary R. Harvey and Mary K . IVba.rton here, but adds to the dOll ble bond structure, producmg for preparing some of the spectra.

151 [8] H. "V. Thompson and P. Torkington, Trans. Faraday 5. References Soc. 41, 246 (1945). [9] J . D . Cotman, Jr., Ann. N. Y. Acad. Sci. 57,417 (1953). [I] R. R. Stromberg, S. Straus, and B. C. Achhammcr [10] A. L. Scarbrou~h, W. L. Kellner, and P. W. Rizzo, Thermal decomposition of poly(vinyl chloride) ill ~ Modern. PlastICs 29, 111 (1952); vacuu.1ll ( pre se n~ e d at the 139th meeting of the mechanlSll1s, p. 95, NBS Circ. 525 (1953) . AmcrIcan Chenllcal Society, Atlantic City, ~. J ., [1 1] V. W. Fox, J. G. H e ndricks, and H . J. Ratt i, Ind. Eng. Sept. 1956) . Chem.41, 1774 (1949). [2] C. S. :Marvel, Syn~hetic polymers of organic chemistry, an [12] M . R. Harvey, J. E. Stewart, and B. C. Achhammer advance~ treatise, vol. I, p. 754, edited by Gilmall J. R esearch NBS 56, 225 (1955) RP2670. ' (John Wiley & Sons, Illc., New York, N . Y., 1943). [13] L. J. B e lla~lY , The infrared spectra of complex molecufes [3] J . E. Campbell and W. H. Rauscher, J. Polvmer Sci. (John Wiley & Sons, Inc. New York N . Y. 1954) 18, 461 (1955). • [14] J. Linnig and J. E. Stewa{·t J. R esea'reh N BS 59 '27 [4] S. Krimm and C. Y. Liang, J. Polymer Sci. 22, 95 (1956) . (1957) RP2771. ' , [5] C. S. Marvel, .J. H . Sample, and M. F. Ray, J. Am. Chem. Soc. 61, 324 (1939). [6] C. S. FlIlla, Chem. Revs. 26, 143 (1940). [7] H . W. Thompson and P. Tarkington, Proc. Roy. SOP. (Londoll) [A] 184,21 (1945) . VVASHINGTON, July 24, 1957.

152

U. S. GOVE RNMENT PRINTING OffiCE: 1956