PolymerPhotochemistry 2 (1982) 1-12
A S T U D Y OF T H E P H O T O D E G R A D A T I O N POLYVINYL CHLORIDE
R. STEPHEN DAVIDSON
Department of Chemistry, The City University, Northampton Square, London, EC1V OHB, Great Britain and RALPH R. MEEK
Department of Chemistry, The University, Leicester, LE 1 7RH, Great Britain (Received: 28 November, 1980)
ABSTRACT Mass spectrometry has been used to examine the photoinduced dehydrochlorination of polyvinyl chloride films both in the absence and presence of titanium dioxide. Hydrogen chloride evolution was only observable when light of h < 300 nm was used. The presence of titanium dioxide affords some protection. When light of ~, > 3 0 0 nm is used photodegradation of added stabilisers is observed. Preliminary experiments indicate that rigid polyvinyl chloride behaves in a similar way to the film. By use of infra-red spectroscopy it was shown that polyvinyl chloride photodegrades in the presence of air to give carbonyl compounds, hydroperoxides and polyenes if light having )t < 300 nm is used. Added titanium dioxide can retard these processes. From these results and those obtained with light of A > 300 nm, in which only carbonyl compound formation can be observed, it is proposed that the routes for carbortyl compound and hydroperoxide formation are not necessarily inter-related. INTRODUCTION
Polyvinyl chloride is used extensively for outdoor purposes and therefore its weathering performance is a subject of some importance. 1 It has been established that irradiation of the polymer leads to dehydrochlorination, 2"3 and photo-oxidation to produce carbonyl compounds and hydroperoxides. 4a'b From experiments utilising polyvinyl chloride deliberately doped with aromatic carbonyl compounds it was shown that carbonyl compounds photosensitise the dehydrochlorination reaction, and a mechanism involving hydrogen abstraction from a methylene group was proposed as an important sensitisation process, s That triplet benzophenone is involved in this 1 Polymer Photochemistry 0 1 4 4 - 2 8 8 0 / 8 2 / 0 0 0 2 - 0 0 0 1 / $ 2 . 7 5 © A p p l i e d
England, 1982 Printed in Northern Ireland
R. STEPHEN DAVIDSON, RALPH R. MEEK Ph2CO ~
Ph2COTt + ~ C H 2 ~ H ~ -
, Ph2COH + --(~HCH-S.-CH2CH-~-
C1 J . 1
process was established by the finding that the addition of naphthalene, a good quencher of triplet benzophenone, retarded both the reduction of benzophenone and the dehydrochlorination process (Scheme 1).6 Aliphatic carbonyl compounds sensitise the dehydrochlorination of 1,4-dichlorobutane, 7 probably via the excited singlet state of the ketone, and therefore the production of carbonyl compounds via photo-oxidation leads to increased susceptibility to weathering. It has also been shown that phenols, which are present in many polyvinyl chloride formulations, act as prodegradants. 8 A particularly significant finding is that temperature plays an important part in determining the extent to which irradiation with ultra-violet light leads to dehydrochlorination. 9 When the surface of the polymer is maintained at a low temperature (e.g. 0°C), dehydrochlorination is not observed although other degradation reactions (e.g. photooxidation) do take place. By way of contrast, when the sample, having a fairly high surface temperature (80°C), is irradiated, dehydrochlorination was observed. There is little information concerning the effect of incorporating titanium dioxide into the polymer. From previous studies we expect that the photoreactivity of the pigment, and its compatibility with the polymer should determine whether or not added pigment stabilises polyvinyl chloride. In this paper we describe the use of mass spectrometry to monitor hydrogen chloride evolution from samples of polyvinyl chloride irradiated in v a c u o . This is shown to be a sensitive method for following the dehydrochlorination reaction. The results of these experiments, together with the application of more conventional methods have enabled us to assess the effect of wavelength and of added titanium dioxide upon the photodegradation process. EXPERIMENTAL
Film samples were prepared as follows. Tetrahydrofuran (10 ml) was distilled from lithium aluminium hydride into a flask. Commercial polyvinyl chloride (10 g) (Corvic D65/02) was slowly added to the solvent which was rapidly stirred and boiled during the addition. When all the polymer had dissolved the desired quantity of titanium dioxide (types A & B) 1° was added to the solution. Good suspensions of the pigment were obtained by either thoroughly agitating the mixture for 12 h or by employing a Dawe Instrument Ultra-Sonic Probe. Approximately 5 ml of the final solution was drawn down a glass plate to give a coated glass surface, which after evaporation of the solvent gave a film of
A STUDY O F T H E P H O T O D E G R A D A T I O N O F P O L Y V I N Y L C H L O R I D E
3 0 ± 5 tx thickness. The film was removed from the plate and placed in ether in order to remove the final traces of tetrahydrofuran. 4b The ether was removed by applying a vacuum to the film which was maintained at room temperature. The described procedure gave films of reproducible thickness and properties. In order to remove stabilisers from polyvinyl chloride, the tetrahydrofuran solutions of the commercial polymer were added to methanol. The precipitated polymer was filtered off, dried and re-dissolved in tetrahydrofuran. Commercial rigid polyvinyl chloride samples were prepared at Laporte Industries Limited (Stallingborough). A lead free formulation of the following composition was used. Constituents Corvic D65/02 Stanclere T135 Phosclere T26 Estabex 2307 Stanclere 1052 Stanclere 1063 Stearic Acid Titanium dioxide (Pigment A or B) K120N
Weight ratio 100 2.4 0.5 2.0 0.1 1.0 0.3 As required 7.0
These constituents were blended using a Banbury mixer, rolled, diced and finally extruded to give a film of polymer ( - 1 5 0 Ix thick). Mass spectrometry measurements were carried out as follows. Films were placed in a water-cooled cell (see Fig. 1). The cell was evacuated on a vacuum line. The evacuated cell was connected to the mass spectrometer (VG Micromass Q16). The background spectrum was taken just before irradiation was commenced. Irradiations were carried out using a medium-pressure mercury lamp (Hanovia 2 5 0 W ) at a distance of 10 cm from the cell. The volatile products produced on illumination were analysed by two different methods. In the first, the sample cell was 'left open' to the mass spectrometer during the irradiation, and spectra were taken after a known period. Utilising this method the products were being continuously taken into the mass spectrometer due to the vacuum system of the mass spectrometer. In the second method the evacuated cell was sealed off and irradiated for a set period of time, after which the contents of the cell were admitted to the mass spectrometer. The first method was used the most extensively since it showed how the production of hydrogen chloride varied with irradiation time. Experiments were performed with both quartz and pyrex cells. To examine the photodegradation in air, films were prepared which were
R. STEPHEN DAVIDSON, RALPH R. MEEK
g , ~" ~To l lMtss i Spectrom~:er s..tus~,,p via a ~ Rob~towtap
()F~.usl watt" tn
cr~q~rtz "1, pyrex ~umi iu.msb.Ip ~
RV.C. sa~npLe on
Cell (water cooled) for the analysis of photodegraded PVC.
attached to holders which fitted directly into the sample compartment of a Perkin Elmer 580 infra-red spectrophotometer. Ultra-violet spectra for unpigmented samples were examined using a Pye Unicam SP800 spectrophotometer. The samples were irradiated with a medium-pressure mercury lamp (500 W Hanovia) using the equipment previously used for studying the photodegradation of polyethylene and polypropylene.
When polyvinyl chloride is irradiated in a water-cooled quartz cell in vacuo, hydrogen chloride is evolved and this can be readily detected using mass spectrometry. A typical spectrum, obtained after a short irradiation time, is shown is Fig. 2. The peaks due to hydrogen chloride are readily visible at m/e 36 and 38. With the cell open to the mass spectrometer one can monitor the formation of hydrogen chloride by following the intensity of the m/e 36 as a function of time (Fig. 3.). In order to compare the rates of evolution of hydrogen chloride in different experiments it is necessary either to operate the mass spectrometer under the same conditions for every experiment or to have an internal standard to which the height of the m/e peak can be related. The former method proved to be impractical. It was found that degassing the cell on the vacuum line always led to some residual oxygen. The concentration of
A STUDY OF THE PHOTODEGRADATION OF POLYVINYL CHLORIDE
20 Fig. 2.
The mass spectrum of volatile products from irradiation of PVC.
the residual oxygen changed very little after the cell had been on the line for 15 min. Oxygen was therefore used as the internal standard. To obtain a profile of rate of hydrogen chloride formation versus time the m/e 36 peak height divided by the oxygen m/e peak was plotted versus time. In this way constant results were obtained for runs employing samples cut from the same polymer sheet. Figures 4 and 5 show that there is an initial induction period and then the rate of hydrogen chloride formation steadily increases until a maximum is reached. Since in this particular type of experiment the evolved gases are being drawn into the mass spectrometer it appears possible that the profiles in Figs. 4 and 5 show the balance between hydrogen chloride formation and its removal by evacuation into the mass spectrometer. An indication that this was the case, was that after cessation of illumination the concentration of hydrogen chloride was found to decrease at a very fast rate. Two experiments were carried out in order to demonstrate this point. In the first the irradiation was carried out with the cell isolated from the mass spectrometer. The evolved gases were then removed and trapped in a vessel attached to the vacuum line. This vessel was opened up to the mass spectrometer and the decay of the m/e 36 peak
oj o t
~ I I
0 Fig. 3.
m/e 36 for PVC samples irradiated for different periods.
R. STEPHEN DAVIDSON, RALPH R. MEEK 4 o
Irradiation l~me/min Fig. 4.
T h e d e h y d r o c h l o r i n a t i o n of p o l y ( v i n y l chloride). U n p i g r n e n t e d s a m p l e s .
monitored with time (Fig. 6). In the second experiment the evolved gases were left in contact with the cell and then after cessation of illumination the contents of the cell admitted to the mass spectrometer. Once again the decay of the m/e 36 peak versus time was monitored (Fig. 6). Thus, Figs. 4 and 5 show that the rate of evolution of hydrogen chloride does reach a maximal value and the decrease in rates in Figs. 4 and 5 indicate that at an irradiation time of 60-80 min, the rate of production of hydrogen chloride is less than its rate of being pumped away. Figure 4 shows that the stabilisers have a beneficial effect. Interestingly, the removal of the stabilisers led to much simpler mass spectra and it therefore appears that many of the peaks in the spectrum are due to the antioxidants present in Corvic D65/02. From Fig. 5 it can be seen that films containing 10% Pigment B are more stable than those containing 10% Pigment A. A comparision of Figs. 4 and 5 shows that Pigment A at this loading has little effect. In all cases it was found that when a 'Pyrex' irradiation cell was used, hydrogen chloride evolution was not observed after 90 min irradiation of either pigmented or unpigmented films. However, some volatile products were produced showing m/e values at 31, 44, 45, 59 and 74 and 30 min irradiation. After a further 30 min, irradiation peaks at m/e 15, 26, 27, 29, 30, 41, 42, 43, 46, 59 and 74 were observed. Little change was observed on
A STUDY OF THE P H O T O D E G R A D A T I O N OF POLYVINYL C H L O R I D E
Irradiation "time/rain Fig. 5.
T h e d e h y d r o c h l o r i n a t i o n of P V C . P i g m e n t e d s a m p l e s .
i/.\ 5 cO ¢O
contact w i t h the P.V.C.
fl Cl |so(a ~rom ",~e P.V.C.
(HCI g a s in m a s s ~ c t r o m e t e c )
T h e d e c r e a s e of HC1 after the l a m p has b e e n t u r n e d off.
further irradiation. The increase in p e a k heights for m / e 45 and 31 was followed with time for several samples (see Table 1). A correlation between each of the irradiated samples cannot be made since the use of oxygen as an internal standard was not used. These results indicate that irradiation of light having h > 300 nm leads to the photodegradation of antioxidants. The values obtained for the polymer with antioxidants removed indicates that there must be some residual antioxidant in the polymer. Studies of the formation of carbonyl compounds and hydroperoxides were made and the results, when an unfiltered light source was used, are shown in Figs. 7 and 8. Pigment B was found to afford some protection against carbonyl compound formation whereas, at the loading used, Pigment A had little effect. However, both Pigment A and B retard hydroperoxide formation. This suggests that the formation of carbonyl compounds and hydroperoxides may not necessarily be interconnected. A further indication that this might be the case comes from the finding that when light, filtered through 'Pyrex' is used, carbonyl compound formation (Table 2) is observed but not hydroperoxide formation. A few experiments were carried out with rigid polyvinyl chloride. The results of measuring hydrogen chloride evolution using mass spectrometry and an unfiltered light source are shown in Fig. 9. At a loading of 2 parts per hundred Pigment A was found to protect the polymer and Pigment B was even more efficient. Hydrogen chloride evolution was not observed when irradiations were carried out with light filtered through 'Pyrex'.
TABLE 1 PVC SAMPLESI R R A D I A T E D Exposure time (rain) 5
15 30 45 60 75 90 m/e 45 5
15 30 45 60 75 90
WITH WAVELENGTHS GREATER THAN
No pigment (no antioxidants )
No pigment .
. 38 203 263 303 350
. 29 86 120 125 130
. --15 38 40 . -10 18 29 26
Plus Pigment A
Plus Pigment B
52 272 523 638 617
6 55 81 131 119
300 N M
. -136 136 272 272
24 37 65 47 52
A STUDY OF THE PHOTODEGRADATION OF POLYVINYL CHLORIDE (a) Change in carhonyt conc,
O unpigrnent~ X p~us pigmer~ A • ~ s pigmentB
0 Fig. 7.
100 exposure time I h
The infra-red analysis of photo-oxidation of PVC; change in carbonyl concentration.
x ~ x ~ •
exposur~ tlm e/h Fig. 8,
Change in hydroperoxide concentration.
TABLE 2 CARBONYL INDEX (A ABS. 1720 cm-~/ABS. 1430 cm 1).
Exposure time (h)
PVC (no pigment)
PVC (+Pigment A)
PVC (+Pigment B)
65 137 307
0-024 0.049 0'050
a Abs. = absorbance.
R. STEPHEN D A V I D S O N , R A L P H R. M E E K
2phr pi9ment A
Fig. 9. Hydrogenchloridegas evolvedfromrigid PVC (area =20 cm2).
Irradiation of polyvinyl chloride film and rigid polyvinyl chloride with light <300 nm and in the absence of oxygen leads to rapid dehydrochlorination. This process can be retarded by adding titanium dioxide and in the reported results, the more durable Pigment B offered more protection than Pigment A. At first sight it may seem strange that titanium dioxide which absorbs at wavelengths <420 nm does not act as a UV screen. Presumably most of the degradation initially takes place on the surface of the polymer and consequently, unless a very high loading of the pigment is used, there will be very little UV screening. The lack of hydrogen chloride evolution when filtered light is used supports earlier findings that light of A > 300 nm is ineffective in causing dehydrochlorination,s furthermore the temperature of the sample was maintained at a fairly low value (-20°C). The mass spectrometry method also revealed that irradiation with light A < 3 0 0 n m and A > 3 0 0 n m leads to the photodegradation of stabilisers. Irradiation of films of polyvinyl chloride in air leads to photo-oxidation, the products being dependent upon the wavelength used. Thus, hydroperoxides are only formed when light of ~ < 300 nm is used. Furthermore, in these experiments it was also found that dehydrochlorination only occurred when light of h <300 nm was used. Somewhat surprisingly Pigments A and B do behave
A STUDY OF THE PHOTODEGRADATION OF POLYVINYL CHLORIDE
differently in their protective action, e.g. Pigment B retards carbonyl compound formation (Fig. 7) whereas Pigment A does not. Since Pigment B also retards hydroperoxide formation it is unlikely that its retardation of carbonyl compound formation is due to it preventing the break up of peroxidic compounds to give carbonyl compounds. The apparent lack of inter-relation between carbonyl compound formation and hydroperoxides may indicate that the carbonyl compounds are produced via peroxides. -[CH2--~H~ CI-I,
OI2 \ t.
CI J --CH2CH-~CH2CH~-
We thank the SRC for a maintenance grant (to RRM). We gratefully acknowledge the support given by Laport Industries Limited (Stallingborough), via a CASE award and to Mr M. Gamon (Laportes) for technical assistance and for many helpful comments.
REFERENCES 1. MCKELLAR,J. F. and ALLEN, N. S., Photochemistry of man-made polymers, Applied Science Publishers, London, 1979. 2. OWEN, E. D., ACS Symp. Series, Polym. Int. Symp. 1975, 25 (1976) 208. 3. GraB, W. H. and MACCALLUM,J. R., Eur. Polym. J., 10 (1974) 529. 4a. MART~, K. G. and TrLLEY, R. I., Br. Polym. J., 3 (1971) 36. 4b. Moax, F., KOYAMA,M. and OKI, Y., Angew. Makromol. Chem., 64 (1977) 89. 5. OwErq, E. D. and BAILEY, R. J., J. Polym. Sci. A-I, 10 (1972) 113.
R. STEPHEN DAVIDSON, RALPH R. MEEK
6. OWEN, E. D. and WILLIAMS,J. I., k Polym. Sci., Polym. Chem. FAn., 11 (1973) 905. 7. GOLtJB, M. A., J. Phys. Chem., 75 (1971) 1168. 8. FOSTER, R. J., METEOR, J. M., WHrrLING, P. H., GRANT, K. R. and PI-IILLWS,D., J. Appl. Polym. Sci., 22 (1978) 1129. 9. MoP,l, F., KOVAMA, M. and OKI, Y., Angew. Makromol. Chem., 75 (1979) 113. 10. DAVmSON, R. S. and MEEK, R. R., Eur. Polym. J., 17 (1981) 163.