Document ZJX9xbwbxe3O4n5mRv2JK71YJ

Purr. ' Appl. Chem, , Vol.53, pp.597-601. Percanim Press Ltd. 1981. Printed in Great Britain. SI'JPAC VOLATILE COMBUSTION PRODUCTS OF POLYVINYL CHLORIDE Edward A. Boettner and Gwendolyn L. Ball Department of Environmental and Industrial Health, The University of Michigan, Ann Arbor, Michigan 48109, USA Abstract - The decomposition products that are evolved when heating or burning polyvinyl chloride or its plastics, both in air and in the absence of air, have been analyzed by several investigators. In recent years, much of this work has b'een directed to those compounds kncwn to be toxic to man. As many as seventy-five compounds have been detected, more than fifty of which have been identified. Quantitative analyses have been carried out on many of these under various burning conditions. The effects of varying the temperature, the rate of burning, and the amount of air available are summarized. INTRODUCTION The Polymer chemist has been interested in the pyrolysis products of polymers for a number of years, using the data generated to elicit information concerning polymer structure, as well as to determine the effects of various additives. More recently, the environmentalist has disc laved an interest in tne products of combustion, from the standpoint of the environmen tal and nealth Hazards of such products. In our present environment, plastics are combusted intentionally (for disposal) or accidentally, such as in fires involving homes, industry, ana transportation, with the evolution of toxic products. One of the first polymers that received the attention of environmentalists was polyvinyl chior'de (PVC). This interest stemmed from the orior knowledge that hydrogen chloride was a principal product of combustion, and from the fact that a considerable amount of this poly mer goes into applications where an exposure to high temperatures is a possibility. Examples of this are plastic formulations for floor tiles and wire insulation, both of which can be involved in accidental fires. Besides the hydrogen chloride, it was soon learned that amounts of caroon monoxide and benzene, both toxic compounds, are also generated in such com bustion processes. Also speculation of phosgene generation continues to arise. Esch and Dyer (1) report on a study of firemen exposed to the combustion products of PVC in the course of fignting fires, and describe the symptoms and signs of an "intoxication syndrome" associated with this type of exposure. A review of other studies of this problem of human exosoures to 3VC fire products is incorporated In a report by Tewarson (2). In 1963, we began a program to determine the volatile combustion products of various plas tics, concentrating our effort on those evolved compounds that are known or suspected to be toxic. Our first investigations were directed to polymers and plastics of polyvinyl chloriae, resulting in a publication that appeared in 1969 (3). Since that time, further work has been carried out on the same polymer, the most recent on meat-wrap film (4). Concur rently with our research, other researchers have also investigated the combustion products of nVC polymers and plastics, and this report includes not only our own findings but a re view of selected data published by others. As many as seventy-five compounds (5), mainly aromatic and aliphatic hydrocarbons, have been identified as products of combustion of polyvinyl chloride. From the polymer itself no oxy genated compounds have been reported (except CO and CO,) even when burned in air, but a few sue compounds (6, 7, 11) can be found in the formulated plastics, probably originating ,crom the plasticizer or other additives. The types of combustion used in the laboratory studies have varied considerably, both in temperature and in the amount of air used. In the lap:.- case: some experiments have been carried out in a completely inert atmosphere (5, 6, 3. 12), wm le in others, air (3, 5, 7, 8) and oxygen-enriched air (3) have been used. Like wise, tne temperatures have varied from as low as 1Q0C (9) up to the temperatures of an ei" **ic arc (10), Generally though, the temperatures have ranged from 25QC, where the den. crochloriration oecomes measurable, to 600C, where most formulations of PVC will be totally consumed. For heating up to 600C, a tuoular combustion furnace Is frequently used witn ' combustion tube of quartz or Vycor. This provides a chamber that is resistant to tne cnemical action of the combustion gases and permits the dis .lacement of these gases by 597 VC5742 598 EDWARD A. r.OETTNER and GWENDOLYN L. BALL flowing controlled amounts of air o other gases through the tube. In c-r work, we gener ally used a flow of about 0.1 - 1.0 liters per minute. Others have usee not plates or open cone heaters within a closed chamber, such as the NBS Smoke Chamber. The rates of heating have varied from a steadily increasing programed rate (3, 8) up to Instantaneous heating (5, 6. 7, 9, 12). PRODUCTS TESTED The types of PVC tested fall Into two categories, l.e. the polymer itsel*', and plastic for mulations. In the few reports where this information Is available, the tolymers varied in molecular weight from 45,000 to 250,000. The plastic formulations teste: vary considerably, including those intended for wire Insulation, floor tile, and other const'uction uses; but generally contain a plasticizer and other fillers to optimize the product for its particular usage. The amounts of polymer vary from 30 to 60* in the plastic formulations. QUALITATIVE ANALYSIS In our research we were primarily concerned with the "volatile products" of combustion, and therefor optimized collection and analytical techniques for such ccmpouncs. As a result, we found 5 compounds (3), 52 of which were identified. Woolley (5) made a nore extensive col lection, using a liquid nitrogen trap, and as a result found 75 comoouncs, identifying 21 of these in his publication. His chromatograph!c peak No. 49 was identifies as 0-xylene, the highest boiling compound we collected, and it appears that the additional compounds found are those less volatile than the xylenes. He identified his peak No. 75 as naphthalene. The principal products of combustion of PVC polymers are hydrogen cnlorice, carbon dioxide, carbon monoxide, and water. The organic products formed are principally aliphatic and aro matic hydrocarbons, with benzene predominating. Only very small amounts ;f two chlorinated compounds were found, i.e. methyl and vinyl chloride. None of the identified products were oxygenated compouncs, except for carbon dioxide and carbon monoxide, instating that the breakdown process is non-oxidative (3, 5). The compounds reported in our earlier work were present in both the polymers and tne plas tic formulations, with the indication that they originated primarily fren the polymer, and it was postulated at that time that additives contribute mostly to higher soiling combustion products. Since then we nave worked with PVC food-wrap films, and in these plastics have found both unmodified plasticizer and traces of acrolein and styrene. O'Vara (6) compared tne high temperature (6G0C) pyrolysis (in helium) of a PVC polymer with a plastic contain ing 63* of the polymer and 37% of o-dioctyl phthaiate. He found products characteristic of both the PVC and the phthaiate plasticizer, but no products resulting from. the interaction of the two. QUANTITATIVE ANALYSIS Quantitative determinations in our research were carried crut on twenty-two compounds, se lected on the basis of their quantity or environme.ntal concern. Of greatest importance are hydrogen chloride, carbon monoxide, benzene, toluene, vinyl chloride, phosgene and benzyl chloride. The amounts of twelve of these compounds are tabulated in Table 1 for two com binations of polymer and plastic. Plastic G contains 51% of polymer B arc plastic con tains 57% of polymer C. It will be noted that the aliphatic compounds in the plastics have all increased in quantity by factors ranging from 1.2 to 6.1 times that in the polymer, al though the amount of polymer has decreased. Two points should be made regarding the vinyl chloride concentrations shown. First, the increase in vinyl chloride from the plastics may not be real, as it nas been determined that acetaldehyde (a combustion product of the plas ticizer) co-elutes on the chromatograohic column used. Second, altnougn we believe a por tion of the vinyl chloride reported actually results from thermal deqradation, the residual monomer may have been higner in the PVC produced in the 1960's than in present-day PVC. CHANGE IN PRODUCTS WITH TEMPERATURE It has long been known that the decomposition of PVC begins with dehydroc.nlorination, with this process taking place in a narrow temperature range around 280t'C. Thermogravimetrlc analysis shows the orocess beginning in measurable quantities at about 253'C and beinq com pleted by 300C. The 50% weight loss of the polymer up to 300C has been snown by collect ing and analyzing the combustion gases to be almost entirely due to the fermation of hydrogen chloride (and a small amount of benzene), and this accounts for almost all the chlorine in the polymer. The same is the case for plastics, wnere the hydrogen chloride represents al most all of the chlorine in the polymer used in the formulation. In some cases however, wnere tne formulations contain reactive inorganic compounds, the amount of hydrogen chloride VC5743 --- - `"Tr 'itiirllitnfciitiiiTi Volatile combustion products of polyvinyl chloride 599 TABLE 1. Comparison of combustion products of two plastics with the combustion products of their polymers Combustion products, mg/g Compound Polymer B* Plastic G Polymer C Plastic E HCl CO Methane Ethylene Ethane Propylene Propane Vinyl chloride 1-Butene ' Butane Benzene Toluene * see Reference 3. 583. 442. 4.6 0.58 2.2 0.47 0.84 0.60 0.18 0.28 36. 1.3 273. 67. 6.6 2.3 3.0 2.0 1.7 3.3 1.1 1.1 10. 0.94 584. 403. 5.8 0.33 2.5 0.56 1.1 0.52 0.28 0.39 29. 1.1 333.90. 6.8 2.0 2.9 1.4 1.4 2.6 0.58 0.74 11. 1.0 measured is less than the amount liberated from the polymer. Chang and Salovey (12), in studying the thermal degradation of PVC in an inert atmosphere, reported a greater tempera ture range for HC1 evolution. They stated that the liberation of this gas begins at about 2C0C and the rate increases to reach a maximum at 340C. Woolley (5) measured the release of HCl as 94% and 96% of theoretical in nitrogen and in air, respectively, at 300C. Liebman, et al. (8) determined that the maximum rate of weight loss for the dehydrochlorin ation process was at 290C in air and at a slightly higher temperature in nitrogen. The other major product evolved in heating to 300C is benzene. This compound is the pre dominant organic in the thermal degradation of PVC polymer, with 25 to 50 mg per gram of polymer formed in burning up to 600C (see Table 2). Most of this benzene Is evolved before 300C along with the nydrogen chloride. The maximum generation of the other hydrocarbons takes place between 350 and 450C, with the quantities generally less than 1.0 mg per gram of polymer. CHANGES IN PRODUCTS WITH VARYING AIR SUPPLY The types of thermal degradation or combustion used by various Investigators have varied from a complete absence of oxygen to air with added oxygen. This does not change the number or types of products formed and, except for CO; and CO, does not change the quantities ap preciably. The amounts of HCl and benzene generated do not change, and their temperature of evolution changes only slightly. However, Woolley (5) has shown that the rate at which hyarogen chloride is released increases with increasing amounts of air. In our experiments where we varied the amount of air available when burning a polymer (3), it was found that all of the hydrocarbons other than benzene and vinyl chloride decrease with increasing air. Benzene remained constant while vinyl chloride increased slightly. VC5744 600 EDWARD A. BOETTNER and GWENDOLYN L. BALL TABLE 2. Variation of combustion products of Polymer A* with Temperature Compound 2528Q"C co2 CO Methane Ethylene Ethane Propylene Propane Vinyl chloride 1-Butene Butane Isopentane 1-Pentene * Pentane Cyclopentene Cyclopentane 1-Hexene Hexane Methyl clopentane Benzene Toluene --... 0.04 ... 0.06 ... 0.04 0.02 -- ----... ... ... ... ... ... 24. 0.12 it see Reference 3. Combustion products, mg/g 280350C 350430*0 430510C 9.7 20. 0.20 0.33 0.12 0.11 0.08 0.25 0.04 0.03 -- 0.01 0.01 0.02 0.01 0.01 0.01 ... 6.6 0.18 181. 46. 1.3 0.39 0.94 0.31 0.44 0.17 0.08 0.20 0,005 0.03 0.08 0.01 0,02 0.02 0.05 0.02 0.35 0.55 244. 151. ' 1.8 -- 0.41 0.11 0.02 --0,02 0.001 0.01 -- -- -- 0.01 -- 0.16 0.03 510583*0 237. 181. 0.31 -- ---- -- --- -- -- ---- ---- -- 0.01 PHOSGENE: A PRODUCT OF COMBUSTION? For years, it has been claimed by some that phosgene is one of the products of thermal de gradation or combustion of PVC. This claim has been very pooular with the press and fire investigators when reporting on fires involving furnishings and insulation containing this polymer. In some cases, it has been suggested that this compound contributed to the deaths of people killed by the toxic fumes of a fire. Although researchers who have been concerned with the identification of the products of combustion of ?VC have attempted to identify phosgene, they have not generally been successful. Coleman and Thomas (14) have reported finding small amounts (up to 10 ppm), but others, although stating that they have attempted to find it. have been unsuccessful (3, 5, 11, 13). The limits of detection have varied from 0.1 ppm to 50 ppm. Recently, Brown and Birky (10) have published results from a study to determine whether phosgene is a thermal decomposition product of PVC where electrical fires are the source of combustion. Four methods of decomposition were studied: (1) thermal degradation of PVC in a resistively heated furnace, (2) electrical overloading of a PVC clad wire, (3) electrical arcing between electrodes partially covered with PVC, and (4) electric arc initiated flaming combustion in a cup furnace. Results are reported which show that yields of 0.2 to 1.6 mg of phosgene can be generated per gram of PVC by the electric arc decomposition. Lesser amounts (0.1 to 0.2 mg/gm) were found in the other types of combustion. The Identification and quantitation of phosgene in the decomposition products were obtained through the use of gas chromatography, infrared spectroscopy and mass spectrometry. PVC FOOD-WRAPPING FILMS Since 1973, there has been an interest in the combustion products formed in cutting clear plasticized PVC film used to package meats and other food products in supermarkets. At that time, an occupational health problem, la'ailed "meat wrappers asthma" was described VC5745 Volatile combustion products of polyvinyl chloride 601 (15), with the suggestion that the source of the problem was the thermal decomposition prod ucts resulting from the hot-wire cutting of these food-wrap films. The film, which is about 0.2 imi thick,,is thermally cut for each package from a roll using either a 0.5 mm diameter, hot-wire (200' to 350'C) or a 10 mm diameter "cool rod" (135*C). The portion of the film , in contact with the wire or rod during cutting may be subject to thermal degradation and/or volatilization of film components. We carried out a study to determine what compounds were generated and released in this cutting operation, and found that more than sixteen compounds were liberated, most in extremely small quantities. The types of films tested, the method of containing and collecting the evolved compounds, and the analytical procedures used are described in Reference 4. The principal products which were determined quantitatively were hydrogen chloride, plasti cizer, benzene, toluene, acrolein, and carbon monoxide. Benzyl chloride was not found in the products involved. Its presence had been reported by Vandervort (16) when heating plas ticized film to 230C in an all-glass vessel. However, using gaseous and liquid benzyl chlo ride standards and the hot-wire cutting device, we found that benzyl chloride could not be recovered either in the gas phase or In a concentrated hood wash. The decomposition of this compound in the presence of hot iron may account for its absence. Compounds analyzed and normal concentration ranges found included hydrogen chloride (1-10 yg per cut), plasticizer (1-50 ug per cut), benzene and toluene (each <5-20 ng per cut), acrolein (25-150 ng per cut), and carbon monoxide (2-4 ug per cut) using the hot wire. Room air samples, collected during the hot-wire cutting without the sampling hood, had less than 0.25 ppm hydrogen chloride. Using the cool-rod cutting device hydrogen chloride, benzene, and toluene were not detected. Plasticizer was detected (25-36 yg per cut) using the cool rod. CONCLUSIONS On the basis of what is known about the toxicity of the individual products of combustion, and tne amounts of these products as determined experimentally by our research and that of others, we have concluced that by far the greatest hazard associated with exposure to burning PVC is that stenming from the evolved hydrogen chloride and carbon monoxide. This conclusion does not take into account any possible synergistic effect of other compounds in conbination with the two principal croducts, but toxicity studies to date have not yet established such interactions. REFERENCES 1. R. F. Dyer and V. H. Esch, JAMA 235(4), 393-397 (1976). 2. A. 'ewarson, Factory Mutual~Sesearch. A report prepared for The NationalElectrical Manufacturers Association, Washington, D.C. (Aug. 1979). 3. E. A. 3oettner, G. L. Ball and B. Weiss, J. Appl. Polymer Sci. 13. 377-391 (1960) 4. E. A. Boettner and G. L. Ball, AIHA J. 417 5T7322 W80).------- ~ 5. w. D. Woolley, Br. Polym. J. 3,~TSS'-T93-0971). 6. M. M. D'Mara, J7"Polymer Science: Part A-l, 8, 1887-1899, Wiley & Sons, New York (1970). 7. K. L. Paciorek, R. n. kratzer, J. Kaufman", J.Tiakahara and A. M. Hartstein, J. Appl. Polymer Sci. 18, 3723-3729 (1974). " 3. S. A. Liebman, UT H. Ahlstrom, E. J. Quinn, A. G. Geigley and J. T.Meluskey, J. Polymer Science: Part A-1 , 9. 1921-1935 . Wilev & Sons. New York M9711. " --- M. M. Birky, accepted for publication in J. Anal. Toxicology. Washington VC5746