Document zQjMYkvZX7nrmqq1G103RnLLg

A|>R[L, 1969 < ( CHEMICALS 283 toxic gases mom pvc m HOWSEMOLO FERES By G. W. V. STARK, Ministry of Technology and Fire Offices' Committee Joint Fire Research Organization fttUF smoke and toxic gases produced from combustible I materials in a building on fire can reduce the chances of escape of the occupants and put their lives in jeopardy. Records show that roughly half the deaths in fires in the United Kingdom arc attributable to smoke and gases. The principal toxic gas front burning traditional building materials, such as- wood, is carbon mpnoxidc,.-but .newer combustible, building materials, such as plastics, may evolve additional toxic gases. The hazards due to ccllulosic and other combustible building materials arc being examined, with the aim of quantifying the toxic gas and smoke hazards. Some results obtained on the effect of the addition of PVC to ccllulosic fire loads on the toxic gases produced in small scale laboratory fires are presented here. Fig. I. Model coMportment showing wall linings and cellulosic crib i* 4 Tests and Results The tests were made in a 0.9 m cubical compartment lined with asbestos, the front of which could be opened for loading the chamber with combustibles, and fitted at the top with a vent of the full width of the compartment and of adjustable depth. Fig. 1 shows the compartment with the front open containing the ccllulosic fire (a crib of wood fibre insulating board) and wall linings of rigid PVC. The ccllulosic crib was ignited electrically, and samples of gas were withdrawn from time to time during the test lor analysis of permanent gases and hydrogen chloride. The temperature of the issuing combustion gases was measured during a test, their total rate of evolution was calculated using the flow equation of Kawagoe (the ventilating condition in these tests required the discharge coefficient to be altered to 0.9) and the rate of evolution of the toxic gases was then obtained from the composition of the combustion gas.C* Tests were made with ccllulosic fire loads of 6.4 and 12.8 kg in the crib and with vents of 5 cm (equivalent in an average toom to an open fan-light) and 10 and 15 cm (equivalent in *n average room to an open door). In most tests the three fixed walls of the compartment were lined with 1.5, 3 or 6 mm thick sheets of rigid PVC. A few tests were made with a board consisting of 0.5 mm rigid PVC sheet laminated on to 4.5 mm hardboard. The elfcct of a difTcrent arrangement of the plastic in relation to the fire, for example, in furniture incorporating plastics was examined by further tests in which sticks of rigid PVC were 'incorporated in the ccllulosic crib; the weights of plastic used were 2.8 and 5.6 kg, the latter being fhe weight of the h5 mm wall lining. Temperature records were made for the gases emerging from the vent. In all tests burning started with the crib 'mouldering and producing dense white smoke for a few r"nulcs, which was followed by a (lash ignition of tire smoke lr;d a resulting sharp rise in temperature, up to about 700C. fr tests with the 5 cm vent, this was followed hv smouldering 'umbustion, the temperature of the emergent gases remain- at a steady low value of about 23!>'C. With (he 10 cm 'cr|t. smouldering alternated with Hash ignition in most tests, 'hilc at the largest vent size the Hash ignition alter initial U'uouldering was generally followed by continuous flaming "'nihustion. Combustion usually proceeded to completion in at the two larger vent sizes, hut in many tests at 'the v>ial!cst vent size the fire went out before combustion was i0,np]ctc. lb.. -- -on-ms of rvnhUtnn of iiVf!rn-i'n rbb'rgt,. -i,,, I eirl,.-,.. evolved later than carbon monoxide, the delay increasing with decreasing vent size. Tn tests with the 5 cm vent, the delay was a minimum of 30 minutes, but usually much longer, for tests with the PVC present as wall linings. Shorter delays occurred when the PVC was incorporated in the crib, but the rates of evolution were lower than when the plastic was present as a wall lining. (Fig. 2). In all tests carbon monoxide was evolved a few minutes after ignition. The rate of evolution was little affected bv wall linings of PVC but was reduced in the test in which the larger weight of plastic was incorporated in the crib. (Fie. 3). Discussion The results obtained in these tests may be used to assess the ctfcct of PVC in burning compartments on the toxic gases escaping into the remainder of a building, for the_ worst condition in which no gases escape outside the building. The risk will then vary with the mode of burniiuyvid the. time from ignition during which occupants arc expofWto the gases. The growth of a flaming fire is rapid and $ipe or rescue must take place promptly to avoid cxfAAirc to excessive heat or toxic gases. However, gas evolutwfl could occur more slowly, and without excessive heat pVMuclion for longer times if the (ire started by smotilddyiig and continued wiih restricted ventilation before discoyffy ; lor this report exposure times of up to one fyour befojwyscapc or rescue have been assumed. mSr 'Hie assessment of the effect of hydrogen chloride on escape is not straightforward. Atmospheres containing 100 parts t>ei million or more of hydrogen chloride pro intolerable to breathe.'-i while about 1.500 parts per million could prove lethal lor exposures of half an hour.'d) 'Hie lower concen tration could augment the clients of smoke and other ga-ein impeding escape, but on the other hand the irritancy o! hydrogen chloride could gi\e early warning lx' the occupant- CHEMlCAl/y iS4 ( TOXIC GAM \ UtOH PVC IN HOUSEHOLD FIRES (COtlcIltlk-ll /l.'i'l Mj) ( RUBBER AND PLASTICS AC.R F'Z' Cv'oluiinn of hydrogen chloride. Fuel: Series A: II kg PVC 'cAb "pvr IMF-t; v/ood fibre insulating board crib. Series B: 5.6 kg PVC Im.irpordted jn (2.8 kg wool fibre insulating board crib Fig. 3. Evolution of corbon monoxide. Fuel: 12.8 kg wood fibre insulating board crib. Series A: PVC woll lining. Series B: PVC in wood fibre insulating board crib. Series C: wood fibre insulating board crib cnly remote from the compartment on fire. Thus the presence of hydrogen chlotnl,- could not be relied upon to ensure escape, above \!tli*c*< for hazardous levels of hydrogen chloride - can ho compared with the value of 3.01)0 parts per .......on of Ciuhmi monoxide which could prove lethal in halt an hour: emh.in monoxide is odourless and therefore may provide no uitnutig of its presence/3) The results Indicate that when the ventilation is small (5 cm deep vcm. equivalent to an open fan-light) hydrogen chloride from PVC would make little or no contribution to the hazard presented by carbon monoxide from traditional eolluh.de combustibles. Assuming that the rate of burning i-. controlled solely by ventilation, it can be calculated 1'iuni our results that about 30m3 of carbon monoxide wool.I ho generated in the first hour from a room 4 m x 4 m t. ^jgh ventilated through an opening l m x 0.5 m at the lop ,,f the room. This quantity of carbon monoxide could contaminate 10.000 m3 to a level hazardous to life: as i|,n volume is many times the volume of the average dwell mu space, the hazard is a real one. The abow assessments are made on the -'ssumption that the size of the burning compartment has no clfect on the rate of combustion ,,| (|,c plastic. Some large scale tests are in progress in a vniiipartmcnt. measuring 4mx8mx3m high, using siniil.u pl.i-.ucs materials to examine the clfcct of scale. Conditions w,.,,lil however be different if the ventilation were substantially higher than that of an open fan-light. Provided enough I'VC were present in a room on fire, the hydrogen ehlon lc formed from its combustion would add to the risk po ,, uted by carbon monoxide from traditional ccllulosic iu;u,iU1|s which need not differ much from that given above hi u ., namjng fire developed quickly a substantial part of the phi .1|i: would be burnt in the first hour. Therefore. since other workers have shown that for a wide range of conditions the chlorine content of burning PVC is released quantitatively as hydrogen chloride'3- *) the maximum level of risk due to hydrogen chloride in this case can be calculated It can thus be shown that the hydrogen chloride from .i coating (usually less than 0.1 mm) of the plastic on wallpaper in a room does not add much to the risk presented by iKnormal content of cellulosic matter. For greater loads PVC full account must be taken of' the actual rale os evolution of hydrogen chloride..which has been shown to van with time. A fuller account of the programme of tests on the combustion of cellulose and PVC. including the scries d full scale experiments will be published later. Acknowledgement The paper is Crown copyright, reproduced by permission of the Controller HM Stationery Office. It is conirihiitv.) bv permission of the Director of the Fire Research .Stata-. of the Ministry of Technology and Fire Offices' Committee REFERENCES : 1. Kawneue. K.. l-'irc hchyviour in rooms. Ministry of Conitmrtion P*u in" Research Institute Report No. 27. Tokic*. 2. I'atty. I*. A., Industrial h\"icne anil toxicology, New York. 176.'. I** * Wiley. . 3. Henderson, Y.. and lla^card, H. W., Nouou* "ascs and the * of respiration inllfiencin" their action, New York. 1943 '2nd l*l*i*~* Kcinhold Tuhlishin" Corp. , 4. Ka\h;sh. I). J., and 5<:rk, C. W. V., The ceneration of carh"* "* ^ hy lire^ tn coinpurf mcnt<. Joint Tire Research Or-^ni/ah"* Research Note No. 614/1966. 5. Shecnan. I).. `Ihcrm.tl diuntdilinn of polymers* over the ,^ . 60<l:*C. Royal Aircraft Establishment Technical Note Chrin. T;trnhoroti"h. 1956. . f. 6. Tsuchija. V., and Sunit. K.. Thermal decomposition product' u f ' vinyl chloride, j. Appl. Cftt-m., 1767, 77 (12). 364*6. THE ECONOMlf; and PERFORMANCE OF LEAD STABILISERS {ClHlClUiU-ii ,>5) nvity or tne l dabil ISC| 'M--r-e---m- ---e--lJy *h 'iOg'h*. In certain , 'lets such as the production of n profiles. sh< . i illusion, and rigid injection n which havi u become major growth areas in sumption <>l rv< ,, ii,aSic .-ad sulphate is often u sole stabdi-,. i I-. ..mse of the higher ncessmp ter involved A -! hi ,,,,,:i| protection. a..g..a..tils! ulir.i-v may he ohi.,,,,.1 |,v ivplaciiu: pa"rt' or ;dl of tin hv rl4v.. - ' acting as an antioxidant in conjunction with the no,n'-" pigments used in rigid PVC can prolong the life *'l fabricated article. Summarising it can be said that the stnhtl<ver confer on the plastic the necessary properties f^r itv * use. Where non-toxic properties and transpufeiK' * ' required the use of lead stabilisers should not bv ``i sidcrcd. However, if beat stability, resistance- to d-. lion by ultra-violet light or good electrical propc.'Sa ' required lead stabilisers should he used as they . these properties at the lowest cost jo the comi'o'1' ` ; This is particularly so when high, temperature pfev'" 25266003 30 . (p-^ Xml. 4 Jr-/g,3sri-3 Jo JL~ & GASES FROM BURNING PVC v4< ' <L0~ ^. d C. A. Clark B. F. Goodrich Chemical Company Development Center, Avon Lake, Ohio INTRODUCTION From the small beginnings about thirty five years ago -* and particularly since the end of World War II -- the plastics industry b?s become a major visible contributor to our life style. This is quite evident in the improvement In quality and availability of those utilitarian and aesthetic things we consider necessary for living Coday. . Those of us vho have played the.part of contrlbutpr to thl*.growing use'of plastlcs-iln today's society are well acquainted with'the excltcbent 6f these resin, protesting and product developments. Our greatest goal has' been, and is, to .bring to Che area of' product development the best combination, of those attributes that will serve Che product's needs. These can be Itemized in many ways but it will suffice to say chat we seek to optimise desirable physical strength; chemical, flacie and electrical resistance; total economics of materials and processing; as well as the aesthetic properties of color and transparency. Safety for processors and consumers has been "part and parcel" of these product developments and the cany acceptances in the market place are tribute to the success of these endeavors, particularly in the vinyl plastic Industry. Ve see tough transparent containers that serve ir. an outstanding way the package requirements for concent security and safety, chat provide content visibility and that resist breakage which can result in Injury and economic loss to the consumer. In like manner, vc may also mention the flame resistance and toughness of such products as wire and cable Insulation and sheathing, phonograph records, flooring, pipe and building products. From this background on the detailed attention paid by you to plastic product development, it may seeo strange' that the whole industry, and vinyl plastics in particular, has been targeted, mostly in the lay press, as a serious threat to environmental quality. Please note Figure l. When you dig Into the background of these expressed con cerns you find that most of them have to do with the nature of gases evolved when Polyvinyl Chloride is forced to burn. This may be either the accidental burning of the PVC content of a building along with the more easily combustible other materials of construction or the burning of PVC along with other waste products In an incinerator. Some of the questions raised are shown in Figure 2. It is our intent here to take a look at the results reported by scientists fc6n> different areas of the world so that we all nay have a better understanding of just what does happen when PVC burns. That PVC is difficult to ignite is a well recognised face and Hilado^- reports the flash-lgnitIon temperature of 39iC (735F) and seif-lgnition ceoperature of 4 34C <8 50F). Ve would consider the burning of PVC In line with the candle-type oechanfsm described by Fenicorc^,3,t and others^. In this way ve can model the burning by considering that exterior hca: Initiates thermal degradation or pyrolysis of the polymer In the condensed or solid phase to produce fuel for cocbua t ion in the gas phase. pyrolysis of the material, in che absence of oxygen, then, provides us with a means to determine the fuel components that would oxidize in the presence of oxygen. PYROLYSIS The decomposition of PVC resin will produce a cheroogravimetric plot such as that in Figure 3. This is* a graphical picture of PVC weight loss when heated In an inert atmosphere. Kere.ve can see that a large weight loss (approx Irzacely 3S per cent) occurs around 300C (572F) followed by a slower loss of volatiles (approximately 38 per cent)- out to 6003C (lii2F) and leaving a snail amount (approximately 4 per .cent) of ash. Ve are, of course, curious .abouc the chemical nature of these pyrolysis products. Before ve .look Into the reported 'analyses of pyrolysis and combustion products of PVC, 1 believe it is necessary chat w* look at these results with the knowledge that each investigator developed pethod of thermal treatment and product analysis as dictated by his own approach and instrumentation available. I .suppose I am really saying that each study is so=e*-T*t indivi dual but that til represent the results of planned scientific experiments. Some of the major variables In these studies are the method and rate of application of the heat flux, static or dynamic gas flow, and the methods and techniques of the gas analyses. Since space does not allow the full incorporation of these analytical methods, the reader is directed to the full reports a$ indicated in Che references. In 1939, Strocberg^ reported on thermal degradation of PVC in a vacuum by first stripping the resin of KC1 at temperatures up to 350C (662F) for 30 minutes. The resulting residue was then heated in a furnace at 400C (?52F) for 30 minutes end the volatile products from this second heacing were identified by mass spectrometry. Strocberg identified approximately 25 hydrocarbon products with ethylene and benzene predominant. Earlier work by Bradt and Hahler^ showed thac KCl is the pri mary decomposicIon product from 127C to 300C (2S1F to 572F) along with some benzene; and, above 300C (572F), a great variety of hydrocarbons evolve. No chlorine cor chlorinated hydrocarbons were reported in the gases. At a later time, Ohtaml and Ishikava pyrolyzed rVC In a nitrogen atmosphere and analyzed the degradation products by infrared and ultraviolet spectroscopy. They found that: (l) the pyrolysis products, other then rXl, consisted of arociatic and aliphatic hydrocarbons, and (2) the types of pyrolysis products were Influenced by the stereoregularicy (tseticity) ir. the original polymer. Additional qualitative or quantitative measurements were not made. Noffz and associates^ used a high frequency pyrolyzer to degrade PVC as well as a number of other polymers. They separated the products from PVC by ga3 chromjtogrspny and reported that Che hydrocarbon degradation produces consisted only of aromatic compounds. Neither aliphatic hydrocarbon noc chlorine was identified In the volatiles. In a more recent study 0`Kara^ degraded ?VC and PVC plastisols with a radiant type furnace at 6C0C (lil2f) In a helium gas stream and analyzed the products by a cc-bir.j:ion of gas chromatographic and mass s p-jee ror>c c r ic cechnicue. 600 was chosen co provide pyrolysis products In a nan.-.-r similar to those existing under fire conG: ions. Jc is A; 'notf J that this temperature is above the se 1 f-lgnit Ion :.:p<r;ur (850r) mentioned previously for ?V(4. G23 252660Q3 0`H.ir. reports no dcgrnd.tloo products resulting froo plsstlcl-er Interaction In the plnstlsoU. He also reports that a stoichiometric amounc of HC1 Is released <53.3 per cent) froo ?VC based on the chlorine In the polymer. The remaining product, ue re hydrocarbons (~3B per cent) and a chlor Ine-free carbonaceous ash of 3 to * per cent of the original polymers. Combining these background sCudles uc can now put labels on the thermogravlmecric plot we saw In Figure 3. Please note Figure 4. here we see Chat onburnable HC1 is the primary product up to 300C (572F). Above 300C we see th* production of burn able volatile hydrocarbons and ash. In the O'Hara^ study the hydrocarbon fraction was' further analysed in a gas-chroeatograph to provide a qualitative and quantitative analysis. In Table l ve see the results of three runs Chat demonstrate the reproducibility of this technique. Here ve see Chat the pyrolysis of PVC at 600C (1112F) produces a series of aliphatic and aromatic hydrocarbons. The chrorwTtograph areas shown closely reflect the mole per cent contcnr. The amount of aromatic products is greater than that of aliphatlcs, and benzene la the major organic component of the wore than 25 identified. Only one chlorinated product, chlorobenzene, was found in minor amounts. No chlorine was found. In another recent study, 0*Maca** described a highly reproducible method for HCl analysis from pyrolyzed PVC. He also showed-that the outflow of HCl in the pyrolysis gas could be reduced by'compounding with basic fillers such-as calcium carbonate, potassium carbonate.and magnesium oxide, among others. Under tils conditions of short time flane exposure, the combustion of a number of pyrolysis residues In air revealed that the amounc of HCl absorbed by the basic filler was retained in che non-carbonaceous ash. COM3USTION Now that ve have considered the nature of the gaseous products supplied by pyralyclng PVC In the absence of oxygen, let's cake a look at reports of scientific investigations of the, perhaps, more practical exposure, combustion. Host of the combustion studies of PVC prior to 1959 analyzed che gases by colorimetric and gravimetric methods. An extensive review of these original scientific investigations was published in 1963 by Underwriters* Laboratories** and Includes information on the combustion and pyrolysis of other polymers and natural building materials. It Is consistently reported that combustion of PVC produces large amounts of hydrogen chloride, carbon dioxide and carbon monoxide and mInoc amounts of hydrocarbons. This was accompanied by severe oxygen depletion and, very occasionally, traces of phosgene and vinyl chloride. Recent studies of che old colorimetric ar.d gravimetric methods*5.!4 have shown chat the presence of large amounts of hydrogen chloride can interfere and cause erroneous reporting of Che presence of phosgene. In 1967, Tsuchiya and Suml*5 decomposed PVC aC 350, 600 and 8C0C (662, 1112 and 1472F) in a ceramic boat inserted in a furnace through which they passed helium gas for pyrolysis, <r jfr in excess for complete combustion. When products of d^cocipos it ion wore analyzed, primarily by gas chromatography, they found chat the main decomposition product In both cases war hydrogen chloride. The amount of hydrogen chloride produced In an inert atmosphere approached the theoretical value expected for PVC. Inasmuch as these Investigators measured only gaseous hydrogen chloride, the amount determined in an oxidizing atmosphere -as slightly lower because water vapor produced in an atmosphere of air absorbed some of the hydrogen chloride. t'nJcr combustion conditions, the main change froo pyroly*U f.a*t effluent was the presence of large amounts of carbon dioxide and monoxide with a corresponding reduction In hydro carbon content. Ocher oxidation products such as acid, i l di-hyd,: s, alcohols and ketones were not found. Neither fitlarine nor phosgene wjs detected though sensitive detection method, were used. In a more recent report, Boettner and associates*^ reported combustion p.as analyses on three PVC polymers, one copo1vrer with vinyl acetate and three related compounds. They the collected gases f ro= combustion -- lnlclally by Infrared spectroscopy' _..d, in detail, by a coahlnatlon of gas chromatography and a mass spectromecec. The combustion for quantitative gas analyses was done In a furnace by heating, after an initial heating at room temperature, from 200C to 600C (392 to 1112F) at a rate of 3C per minute. Air was supplied to the sample at a rate which, uSen Integrated over Che.entlrc run, would result in about twice the aeount of oxygen necessary to convert all carbon to carbon dioxide and at higher and lower levels as well. Under these conditions, the resin samples completely disappeared with no measurable residue and the compounds left a residue related to the presence of inorganic matter in the recipe. Boettner reported Chat combustion gases from PVC resin contained approximately 50 different chemical compounds and quantitatively analyzed raore than 20 of these. Although there re many analyses In this report -- including the effects of varying air supply and heating rate -- a typical analysis comparing a pure rcsln and a wire and cable compound (51 pec cent FVC) Is shoun In Table 2. Here we note the major constituents of che combustion gases Co be carbon dioxide, hydrogen chloride and carbon monoxide. The other gaseous components are mainly hydro carbons and che most abundant one Is benzene. , Ue note that che hydrocarbon* produced from the compounds are much greater than that- from PVC resin and this Is attributed to che presence In the^gases of breakdown materials from the phthalate plasti cizer. These latter are a series of hydrocarbon* similar to those produced by PVC. Hydrogen chloride in the gas reflects essentially complete conversion of the chlorlce in the PVC. The only other chlorine compounds reported.in che gas arc vinyl chloride and-nethyl chloride In/quantities -less chan 1- tag per gram of sample -oxidized. Although'decectors. capable of measuring 0.1 ppm were used-, no phosgene was found. One of the more recent reports on PVC cosbustlon studies `was made by W. D. Woolley*? In 1971. In this work gases from vdecomposition of PVC in air and nitrogen are analyzed by gas chromatography. PVC resin and unptesclcLzed PVC sheet samples -are compared and a special effort was made to detect phosgene and other oxygenated organic compounds. Decomposition of the 15 eg saoplcs took place in a tube furnace at 300, 400, 450 and 500 (572, 752, 842 and 932F) with a dry air or nitrogen flow of 100 ml per minute. Woolley essentially verified the gas analyses of Boettner*^ In that the organic components consist mainly of aromatic and aliphatic hydrocarbons. He detected approximately seventy-five organic materials, about twenty-five more chan BoetCner. Although che products are modified by the presence of oxygen, no oxygenated species were detected. Phosgene was particularly looked for and was not detected. The organic gaseous decomposition products from FVC polymer were found to be similar in identity -- but greater in quantity -- compared to those from the rigid PVC sheet. This could be reflecting the difference In the resin content of the samples: 100 per cent for the PVC polymer and approxi mately 90 per cent for the rigid PVC sheet. In 1971 Kobayashl and associates*^ reported on analyses of combustion gases from PVC and PVC products in a comparative study with natural high polymer materials such as wood, paper, wool and silk. Their experiments were made by heating one gram samples In an electric furnace over a range of 300 to 700C (572 to 1292F) through which air flowed at a controlle-c rate (1.0-l.S liters per minute) and at normal (21 per cent) and reduced (10.5 per cent) oxygen levels. The gases were analyzed by means of selected detector tubes. Following is a quote from the report: "As for the gas composition generating from the FVC products at large, a major part consists of hydrogen chloride and carbon dioxide excepting hydrocarbons and water content, followed by carbon monoxide, but neither phosgene, chlorine, nitrogen dioxide, oxygen sulphide nor hydrogen sulfide were determined by the detector tubes." Special effort was made to determine If phosgene v?r< present in the combustion gases. The detection limit vas determined Co be 0.2 pea. No phosgene was resorcad in any the combustion gases. * C/I Ci 05 O o <k Carbon monoxide and carbon /(oxide from combust ion of th** natural polymers generally showed -jp in the f.-Jses at lewr temperatures jnd at higher levels chan with ~-`C. This wa* 62*1 related co the caster Ignition of these ..oducts and the degree chat hydrogen chloride docs not evolve with these natural polymers In comparison with PVC. In the comparison analysis with natural products, Kobayashl reported hydrogen cyanide In combustion gases from i/ool, and both hydrogen cyanide and airmonia In gases from burning silks. The maximum concentrations were rcjchcd at 400C (752F) and were as follows: wool -- hydrogen cyanide (1.2 per cent); silk -- hydrogen cyanide (0.36 per cent) and acaonla (0.5 per cent). For comparison the maximum concentra tion of hydrogen chloride in gases from burning FVC products was from 0.65 per cent at 500C (932F) for flexible auto cover leather to 12.0 per cent at 400C (752F) for PVC resin. TOXICITY At the start we should recognise that Che products of combustion of all organic products arc toxic and this Includes wood, paper, natural fibers, etc. This Is, of course, because of two main occurrences in a fire: production of carbon monoxide and dioxide and reduction of the oxygen concentration. The latter is cociaon to all organic combustion and this Includes PVC. Specific studies of the toxicity of gases from burning FVC have been conducted by various Investigators over the year$l2. The most recent was reported by Cornish and Abac'9 of the University of Michigan. This investigation utilized Che same samples of PVC as chose described in t-he previously mentioned study of PVC combustion by Bocctner16. All animals exposed to the combust ion gases were male Sprague-Dawlcy rats. Autopsies were performed on all animals that died. , Cornish and Abar reported, ''Without question, the major cause of death in this study of four polymers was the presence of relatively large amounts of carbon monoxide in the pyrolysis gas stream." They found relatively minor or no tissue damage in test animals exposed to hydrogen chloride from FVC decom position. CONCLUSIONS 1. Phosgene and chlorine are not combustion products of PVC. 2. When PVC is forced to burn, the price products arc carbon monoxide, carbon dioxide, hydrogen chloride and water. 3. Minor gaseous products from burning PVC are aromatic and aliphatic hydrocarbons, predominantly benzene. A. Carbon monoxide is the prime toxic component produced. 5. Hydrogen chloric, .row PVC pyrolysis or combustion Is essentially stoichiometric. ACKNQULEX.EMrlNT The author wishes to thank Mr. Llnwood B. Crider snd pr. Michael O'M.ira for their onny suggestions and helpful contri butions In the preparation of this paper. REFERENCES1 2 3 * 5 6 7 8 9 * 1. Hilado, C.J., Flammability Handbook for Plastics, Technocalc Publishing Co., 1969, p.39. 2. Fenloore, C.P. and Martin, F.J., Combustion and Flame, 10, 1966, P.135. 3. Fenioore, C.P. and Jones, C.U., Ibid, p.295. A. Fenimore, C.P. and Martin, F.J., Hod. Plastics, 43, 1966 P.1M. 5. Learmonth, C.S. et ai, Br. Poly. J., 2* July, 1969, p.J49. 6.^Strombcrg et ai, J. Polym. Sci., 35, 1959, p.355. 7. Bradt, P. and Mohler, F.L., J. Research Natl. Bur. Standards, 55, 1955. p.323. 8. Ohtani, S. and Xshlkawa, T., Kogyo Kagaku Zasshl, 65, 1962, p.1617. . 9. Noffz, D., Benz. V., and .Pfab, V. , Z\ -Anal. Chca.,- 235, 1968, p.T21. lb. O'Hara, M.M., J. Polym. Sci. A-l, 8, 1970, p.1887. li. O'Hara, H.M., J. Polym. Sci. A-l, 9, 1971, p.1337. 12. Dufour, R.E., Bulletin of Research No. 53, Undervr iters' Laboratories, 207. East Ohio Street, Chicago. 111. 60611, 1963. 13. Linch, A.L., S.S. Lord, Jr., Kubitz, K.A. amd DeSrunner, M.R., Amec. Ind. Hyg. Assoc., J., 26, 1965, p.C65. 14. O'Kara, H.H., Crider, L.B. and Daniel, R.L. , Amec. Ind. Hyg. Assoc. J., 32, 1971, p.153. is. TsuchLya, Y. and Sumi, K., J. Appl. Chea., Londca, 1967. p.366. 16. Boettner, E.A Ball, G., and Veiss, B., J. Appl. Polym. Sci., 13, February, 1969. p.372. 17. Woolley, V.D. , Br. Polym. J., 2* July, 197]., p. 186. 18. Kobayashl, Y. , Kasahlro, H. and Kurata, H.,, Japan Plastics, 5, January, 1971, p.40. 19. Cornish, H.H. and Abar, E.L., Arch. Environ. Health, 19, July, 1969, p.15. FIGURE I NEW YORK TI WES, OCT. 22, 1969--"NEW PLASTIC CAUSES AIR HAZARO" CINCINNATI POST, OCT. 31, I969- -- "DISCARDED PLASTIC MATERIALS GUMMING UP INCINERATOR" . NBS NEWS RELEASE NOV. 1969---"HAZARDOUS PRODUCT PRODUCED 3Y ELECTRICAL IKSULAT10N" INDUSTRIAL RESEARCH FEB. 1970--"DECOMPOSITION OF PVC YIELDS HAZARDOUS GAS" MODERN PACKAGING JAN. I 970---"PVC AN0 THE CYCLAMATE SYNDROME" DETROIT NEWS FEB. 2V, 1970---"NEW PLASTIC IS TOXIC WHEN BURNED" NEW YORK DAILY NEWS JUL. 16, I97I-"PLASTICS FIRMS SUE OVER TAX" C ! E NEWS NOV. 22, 1971---"PLAST t C CONTAINER TAX RULE0 OUT IN NEW YORK" 625 FIGURE 2 [> HOW DOES PVC BURN? 0 WHAT ARE THE MAJOR DECOMPOSITION PRODUCTS FROM PVC? 0 WHAT ARE THE MAJOR COMBUSTION PRODUCTS FROM PVC? 0 IS PHOSGENE A COMBUSTION PRQ0UCT? 0 IS CHLORINE A COMBUSTION PRODUCT? 0 ARE THE COMBUSTION PRODUCTS FROM PVC TOXIC? (VJ CT. cn o o cn (( FIGURE s 0 FIGURE 4 WON'T BURN 626 FIGURE 5 ( CONCLUSIONS |. PHOSGENE AND CHLORINE ARE NOT COMBUSTION PROOUCTS OF PVC. 2. WHEN PVC IS FORCED TO BURN, THE PRIME PRODUCTS ARE CARBON MONOXIDE, CARBON OIOXIOE, HYOROGEN CHLORIOE ANO WATER. 3. MINOR GASEOUS PRODUCTS FROM BURNING PVC ARE AROMATIC AND ALIPHATIC HYDROCARBONS, PREDOMINANTLY BENZENE. V. CARBON MONOXIDE IS THE PRIME TOXIC COMPONENT PRODUCED. 5. HYDROGEN CHLORIDE FROM PVC PYROLYSIS OR COM3USTI ON IS ESSENTIALLY STOICHIOMETRIC. TABLE 1. PVC PYROLYSIS PRODUCTS OVERALL DEGRADATION PROFILE . HC1 Hydrocarbons Carbon Char %. wt. 58.3 37.0 4.7 HYDROCARBONS PROFILE. (Gas Chromatographic Area) 'ABLE 2. COMPARISON OF COMBUSTION PRODUCTS PVC RESIN VS. WIRE L CARLE COMPOUND (POLYMER r. VS, PLASTIC G) COMDUSTION PRODUCTS, mg/g Material PVC Resin Wire E- Cable Compound (S 17, PVC)___________ HC1 co2 CO Methane Ethylene Ethane Propylene Propane Vinyl Chloride 1-Butene Butane Isopentane 1-Pentene Pentane Cyclopentene Cyclopentane 1-Hexene Hexane Methylcyelopentane Benzene Tolue'ne Residue** 583 729 442 4.6 0.58 2.2 0.47 0.84 0.60 0. 18 0.26 0.02 0.06 0.16 0.0s 0.05 0.05 0.12 0.14 36.0 1.3 273 616 67 6.6 2.3 3.0 2.0 1.7 3.3 1.1 l. 1 0.15 4 0. 35 0.58 0. 14 0.16 0.24 0.49 0.14 10.0 0.94 15? * RE; Boettner et &1, J. AppL Sci. t 13, February 1969. p. ** Residue is what remained in combustion boat. CO Methane CO2 Ethylene Ethane Propylene Propane Butane, butene 1 Butadiene, diace tylenney} and hydrocarbons Bemene Toluene Chlorobenzene Xylene Allylbenzene C9hI2 C9H1Z Indane Indene iene^ Ethyltolue Methylindane Methylindenee Napchalene Dimezhyli rid a n e hie thylnaphtha le ne Me thy Lna ph '.hale ne"7 Ace naphtnale ne _j Run 1 0.6 8.8 0.4 3.5 4.8 2.9 3.0 2.5 1.9 51.2 5.3 1.1 1.9 1.3 0.4 0.2 0.2 0.9 0.5 1.1 4. 8 0.7 0.8 1.2 Area, % Run 2 2.2 7.8 0.3 3.5 5.1 2.6 1.9 2.1 1.8 50.8 S. 9 1. 1 2. 1 1.4 0.5 0.2 0.3 1.4 0.5 1.3 4. 6 0. 8 0. 7 1.0 Run 3 0.8 9.3 0.2 3.8 5.5 3.0 2.3 2.1 1.3 50. 1 5.8 1.3 2. 3 1.6 0.5 0.3 0.3 0.9 0.5 1.0 4. 5 0. 8 0. 7 1. 1 RE: O * Ma r a, M. M. , J. Polym. Sci. A - 1, 8, 1970. p. 189 627 01 Cl Cfc o o -3 EG ETAI. lo- Toxicity of edema in rats exposed to the pyrolysis prod ucts of polyurethane foam. 'I he quantity X,' Pyrolysis Products of and nature, of pyrolysis products of Teflon depends on degradation temperature. War isrojv } iituO -U of v\/mJvfu-nl On" Vinyl Plastics Hook Co.. 1903. Tests for Moivn.y in Herbert it. Cornish. Chi), anil ily. and Kwon' reported no ndvenv: effect.-: in rats exposed lo products of Teflon heated to 400 C, while pyrolysis nt 4-00 C producr-ri acute: pulmonary hemorrhage and edema. -ncr J Clin Path, uvhnj. Ellen L. Ahar, MA. Ann Arbor, Mich Coleman ot al- .suggest that at boO to 700 C, .iiwanish, H.: bul'inuu' nstrntion of IniraoclUiU.- Polyvinyl chloride polymers and lormulations s of Pancreatic Islcis. were pyrolyzed in a stream ot air by gradually o> ltVk>. raising the temperature from ambient to approxi- in Microscopical IxvhU- _, ... , , ipocam-l Mossv Kib,e mately GOO C. The pyrolysis air stream was diluted .ulmle Silver Procedure. with twice its volume of room air, and rats were 241 1967. exposed to it. Exposure lo an air stream con- ciural Localization o' taining the pyrolyzed products of 1 to 2 gm of carbonyl fluoride may be a major toxic product of Teflon degradation. MacFarland and Lcong*5 stuclied the tox icity of pyrolysis products of polyurethancand polyurethane-coated nylon. Acute as phyxia from occlusion of the upper re spiratory tract was apparently the mechanism of death of animals exposed to pyrolysis -d Sulfide-Silver Mc-.U.kI. polyvinyl chloride polymer resulted in the death products of polyurethane, whereas with ny Line 281 '967. c( 50% of the animals. Most deaths were due to lon -products, death occurred from slowly ?r. S.: Trials to Mluy c2rV|0n monoxide (CO), and carboxyhemoglobin or Lltrastrurtur.il l jcoyjp) levels correlated well with the amount .totals, Acta Hts.oi . ^ plastic pyrolyzed. Little histological evidence ... and Pratt R i.'.lsl of-lung damage was evident. When oxygen (O;) vrica ISSS Edition. ed Sj was a^ded td the air stream to prevent deaths stt Co.. I960. p hL';>. from CO, pulmonary edema and interstitial hetr.or- developing pulmonary edema. Tha-r.e: re ported the chemical analysis of products of combustion of halogen containing epoxy re sins, while Leong and MacFarland' reported on the hazards from thermodecoreposition of 'hysicochemical Itaticnnk rhage developed. The lungs of some animals epoxy resins. On the basis of animal studies tv of Mercury and It- exposed lo high levels of pyrolysis products ol . they suggest that pyrolysis products of epoxy id Sci 65:454-460 (April; v;r>y[ chloride-vinyl acetate copolymer also showed resins may constitute a hazard under cir local edema and intra-alveolar hemorrhage. Poly aiion by Mercuric i'hlo- vinyl chloride formulations, containing additives koto J 50;25S-24S iJunci and inert materials, vie re in general less toxic per T robe: Its Use for 3^m > sample pyrolyzed. ,:cs.........sirs'. J Hisincheir. ) 1966. al: The Localization ol red Radioactive Mercury 1 HE WIDESPREAD use of plastics in rcstrunt Res 3:>:;?> < Doc, the manufacture of commercial and house cumstances realizable in practice. Boettner and Weiss0 described an analyt ical system for identifying the volatile pyrol ysis products of a virgin homopolymer polyvinyl chloride fPVC) resin. In another study, Boettner et al10 noted that three distinct breakdown temperature ranges oc curred as temperature was increased. The' hold products raises the question of poten- pyrolysis products released during these T.; and Nakao. K.: Dis- tial hazards associated with burning these three weight loss periods differed markedly. ie Brain and Its hcrxeilu- n.ajer:a^s This is not a new problem; it was Organic Mercury Poison- .. The present study is an extension of their `-406 (May) 1960. investigated by Berger et ai1 in a study ot work to include toxicological investigations .-.racellular Distribution oi thermosetting plastic materials used for in of the pyrolysis products of polyvinyl chlo ner Methcxyethylmeivuiy sulating electrical conduits. They deter- ride plastics. The samples and sample desig rmccol 16:1645-iocw i.Sepu m;nC(j products produced during pyrol- nations are those of Boettner et al.K' riraceUuiar Distribution ol of phenolic and melamine plastics-- dter a Single Injection oi carbon monoxide (CO), cyanides, and am- -m Phcrmocol 17:551-59) ntonia. Watson et al- reported the chemical identification of pyrolysis products of foamed polyvinyl chloride. Zapp's* investi gations revealed pulmonary congestion and >w society in general r responsibilities. In S.S.: Our Changing ctice, Arr.er J Public SulimiUed for publication Dec 26, 1966; accepted Ina 20, 1969. l'i'om the Department of Industrial Health, School of Public Health, Ann Arbor, Mich. Keprint requests to 3616 School of Public Health, University of Michigan. Ann Arbor, Mich 48104 (Dr. Cornish). Materials and Methods Materials.--The materials consist o: four pol ymers produced by two companies and three formulations which utilize three of the poly mers. The materials were classified as follows. Sample A. A polyvinyl chloride homopoly mer in which approximately 20% of the parti cles fall within the range of KXB to 250-. with most of the remainder falling below 100.- in size. Sample B. Similar to A but of somewhat Arch Environ Health--Vol 19, July 1959 ' ..es O Q C ( 16 PYROLYSIS PRODUCTS--CORN/SH & ABAR P'i smaller particle sire, with 85% of the particles in the boat and combustion tube was weighed i within the 75a to 200a range, and the remain at the end of the exposure to determine the ; Table 1.--Effect der below 75a. extent of pyrolysis. Periodic measurements of ! Sample C. A polyvinyl chloride homopolymer hydrochloric acid (HC1) and CO were made Quantity Weight of cxc. (powder) produced by a different company. both in the entrance and exit air streams of the Pyrolyz^tj Residue Sample D. A copolymer of vinyl chloride exposure chamber, utilizing a gas sampling j;. (<4'V) (s^) T: (~ i and vinyl acetate (powder) in the approximate pump and detector tubes. Unfortunately. HCl : i ratio of S5:15. detector tubes did not go above 500 ppm. Pre- ; 0.8 <0.01 1 Sample E. A formulation commonly used in vious studies by Boettner et al.w however. < electrical wire insulation utilizing polymer C show that HCl is released over a relatively i (small cubes, approximately 5 mm). short time during pyrolysis. Thus, maximum 1I i Sample F. A formulation commonly used for concentrations can be estimated with a fair!; 1-0 <0.01 1 floor tile utilizing copolymer D (flufTy powder degree of accuracy. -| i j containing asbestos). In early studies the level of HCl in the air 1 Sample G. A formulation commonly used in leaving the-exposure chamber (desiccator) was * .2 wire insulation utilizing polymer B (sheets cut markedly lower than that entering the cham-: i into approximately 5-mm cubes). her. Thus, in subsequent studies, exposure: I Animals.--All animals included in the pres chambers in which only the animal's nose pro-: ent study were male Sprague-Dawley rats. The truded into the pyrolysis airstream were uti--; 2 -< ` <0.02 animals ranged in weight from 200 to 300 gm lized. Plastic baby bottles, attached to small' and were maintained on an ad libitum diet of nose funnels, opened into a large vinyl tubing, commercial rat chow and water except during through which the airstream was passed. This: I.S the exposure periods. At autopsy, animals were type of exposure .'eliminated the absorption o: 0.03 anesthetized with ether, and blood was removed HCl, and possibly other vapors in the urine by heart puncture for the determination of which accumulated in the bottom of the small A^'.-nal deaths. i carboxyhemoglobin (COHb) and other blood chamber. It also prevented animal crowding Nunretor, Ceaths; den: i i studies. In most cases, blood was also obtained for COHb measurement in animals dying dur and the adsorption of pyrolysis products on animal fur. Hydrochloric acid and CO concen animals under the.se j ing the exposure periods. In some cases, blood trations were measured in samples taken from ' LC-,,was approx:tv i for COHb determinations was obtained from the moving air stream. In all exposures, anir. the tail vein. Tissues were promptly removed Blood Analysis___Blood carboxyhemoglobin of nasal and eye irn i i and small sections fixed in 4% formaldehyde was determined by the micro-method of Scho- charge, salivation, ar and stained with hematoxylin and eosin. Ani lander and Roughtonn-and was always done in Convulsions were a < mals surviving the exposures and not killed duplicate or triplicate. Serum glutamic oxalonr- animals with high were maintained for a minimum of two weeks ctic transaminase (SCOT) was determined by S;nce work O' and observed for latent effects. the method described by Steinberg et ah'- jncjjc-i;c(} tjiat \ Exposure Techniques.--The exposure cham Hemoglobin levels were measured spectropho- D", -,, " ` ' \ . 1 ` . i i ber utilized in early studies was a 7-liter desic tome.tr.tcally in d*ilutie ammon*ia. * ,v 3t-i'*ol^' i. i cator -with the animals maintained on a grid 4 . rna-.s were exposed v- to 5 cm above the bottom. Four rats were normally exposed at one time. The weighed Results early pyrolysis prod: pie A, and remove; plastic was decomposed in a ceramic boat placed in a combustion furnace programmed The results of the first study utilizing affected by large qu; ''ere killed 24 hot;: for increasing tr-mprniture (three degrees per minute). This programming technique was re varying (sample amounts of PVC homopolymer A) arc 'shown in Table 1. Thr exposed animals si: loss during the expe i cently described by Boettner and Weiss.5 Tem perature was measured in the combustion boat by a thermocouple and reached a maximum of 570 C after J15 to 140 minutes at which time pyrolysis was essentially complete. Animals were removed at or near the end of this time period unless death occurred earlier. During pyrolysis. 2 cu ft of air per hour passed over the plastic. 'The pyrolysis stream was cooled animals were exposed in a 7-liter dcssicator. Note that pyrolysis was essentially com plete, and that quantities of ash remaining in the ceramic boat were negligible. Al though the level of CO in the air waf greator than 3.000 ppm for some time dur- ing each of the exposures, the COHb level; rot completely res: oay. L:\er and kid "ere not significant! atvlmalj^sHematocri tc;:d- PT wore SC pjv-lio.p^tf-fyito. loagicmael nes: varied greatly among the different groups id^y, and sple an'! diluUd by the addition of 4 cu ft of air per However, the COHb levels correlate well a.-:weexposed art i hour prior to entry into the animal chamber. Temperature >n the animal exposure chamber d;d net rise above 20 C. with an average of 27 ('. The ash remaining in the ceramic boat or with the amount of plastic pyrolv/od. "Expo sures to increasing amounts of sample demonstrate this correlation.. The quantity of pvrolyzcd plastic which, killed 30'',', of `hr In subseque: -re iCD '-ed by so into `.he pytv r:bee: Cne me:'-.; Arch pnriron -- W/ 19. 'Itllv 1909 rviu,,.Ysis i'i;ouucrs--coi:.\'is!i a- ahai: 17 i' was wi'iplu'd ; cL-U-rmine the | Table 1.--Ettcct ol Pyrolysis Products oI Polyvinyl Chloride Homopolymer (Sample A) ur--wonts of made streams of ilio pas s impling riima'i'ly, IU`I f>00 ppm. I'ro- : ltovvtvt*r. or a relatively hus. maximum -d with a fair ; Quantity pyrolyzccl (qm) 0.8 1.0 We^ht of Residue (gm) <0.01 <0.01 Exposure Time (min) 130 135 CO Max Cone (ppm) >3.000 10-15 min >3.000 10-15 min Rat No. 1 ? w A 1 2 3 4 Blood Hti (gm/100 cc) 1C.2 1 5..*i 16,7. 17.1 15.3 15.7 ie.i 16.2 COHt> (vol %) 14 14 e 15 7 38* 8 Mortality* 0 *4 1 4 HC1 in the air desiccator) was : ring the chamidies. exposure imal's nose pro roam were utiaehed to small 1.2 1.4 <0.01 135 >3.000 15-20 min 125 >3.000 15*20 min i 2 3 4 i 2 3 4 15.3, 16.2 15.0 15.7 17.6 16.6 16.6 15.7 65* 78* 2 11 4 65 65* 19 3 55* 4 68* -go vinyl tubing as passed. This ic absorption of rs in the urine ora of the small nimal crowding l.S 0.03 120 >3.000 15*20 min Animal deaths. ^Numerator. deaths: denominator, total. i 2 3 4 42* 55*. 4 40* 4 45* sis products on and CO conecniplcs taken from animals under these exposure conditions : LC-,,) was approximately 1.2 gm of PVC. In all exposures, animals showed symptoms The results of exposure of animals to PVC homopolymer B is shown in Table 2. All animals exposed to 0.8 gm of pyrolyze-d -boxyhemoglobin method of Schos ; done in ;lu-o..c oxaloacs determined by einberg et al.i: ared spectropho- ia. of nasal and eye irritation with nasal dis charge, salivation, and chromodacryorrhea. Convulsions were a common occurrence in the animals with high COHb levels. Since the work of Boettner and Weiss Itad indicated that HC1 was released from PVC at relatively low temperatures, ani mals were exposed in the desiccator to the plastic survived. Maximum COHb levels were 17 and 21 vol% in tw-o of four rats. At 1.2 gm of plastic, one animal with 66 vol% COHb died. The surviving animals had con siderably lower COHb levels. At 1.4 gm, three of four animals did not survive the exposure. The single surviving animal again had a relatively low COHb level of 13 early pyrolysis products of 2.0 gm of sam vol%, as compared with 65 to 75 vol% in ple A, and removed before they became the other three animals. affected by large quantities of CO. Animals study utilizing were killed 24 hours after exposure. The homopolymer exposed animals suffered a slight weight Table 1. The loss during the exposure period which was -liter dessicator. not completely restored by the following ;ssentially com- day. Liver and kidney body weight ratios f ash remaining were not significantly different from control negligible. Al animals. Hematocrit values were not al in the air was tered, nor were SGOT levels which were some time dur- utilized as a measure of tissue damage. he COHb levels Histopathological examination of lung, liv different groups. er, kidneys, and spleen showed no differences 3 correlate well U'tween exposed and nonexposed rats. ayrolyzed. Expo In all subsequent studies, the animals ts of sample A were exposed by the insertion of only the n. The quantity nose into the pyrolysis air stream as de idled 50% of the scribed in the: methods. An attempt was made to overcome the acutely toxic effects of CO and possibly allow other aspects of toxicity to become evident. Animals were exposed to PVC pyrol ysis products as previously described with air passing over the sample at 2 cu ft/hr, but with replacement of the air dilution stream by oxygen (4 cu ft/hr). In separate studies, 3.5 gm of samples A and B were pyrolyzed and animals exposed to the pyrolyzed air streams. All animals survived these exposures which, without oxygen (0.0 would have been lethal. Calculations of HC1 re lease would predict levels above 6,000 ppm for 15 to 30 minutes during this exposure. Animals were killed 24 hours after expo sure and lung tissue taken for histopatho- Arch Environ Health--Vol J9. July 10C,9 X tn. <r. o ot-* (18 \ PYROLYSIS PRODUCTS--CORNISlA ADAR Table 2.--Effect ot Pyrolysis Products of Polyvinyl Chloride Homopolymer (Sample B) and Formulation (Sample G) Teble 3............. r Quantity Pyrolyzed (gm) Weight of Residue (grn) Exposure Time (min) CO Max Cone (ppm) Rat No- Blood ---------:--------------- --*---------------------------* Hb COHb (grrt/100 cc) (vol %) Quan-.l-.y pyrolyzed Res-d . (grn)__________ Morta'it; Sample 8 0.8 <0.01 130 >3.000 1 2 3 4 17.1 16.2 15.7 13.7 17 1.2 <0.0 21 0 7 4 5 1.2 <0.01 130 >3.000 5*10 min 1 2 3 4 17.1 16.6 15.3 17.1 66* 1.5 0.0 19 94 11 1.4 <0.01 130 >3.000 10-20 min Sample G 6.0 0.93 105 800 Animal deaths. tNumerator, deaths; denominator, total. 1 2 3 4 1 2 3 4 16.6 14.5 15.3 17.1 13.0 15.3 12.0 12.7 7458* 13 69* 7 7 9 3 1.8 3_ 4 2.1 0_ 4 2.3 <0.C 0.1 logical examination. Of the animals exposed D with COHb levels of 12.3 to 14.5 vo1?fn-f:s to sample A, all showed evidence of pul These COHb levels do not anpear to h `pTMTM' ce*ibs. monary edema and intra-alveolar and inter high enough to be responsible for the cieath stitial hemorrhage. Several of the hemor which, occurred during and shortly aftilungs of expo rhages were associated with angular fibers exposure. SGOT levels of all rats exposed tjto controls. and sheets of foreign material within the 2.0 gm were slightly elevated, suggestive d Three of fc alveoli. Examination of the lungs of animals tissue damage. |lhc pyrolysis exposed to sample B revealed pulmonary Histological sections of lung, liver. nriformulntion. ( edema and interstitial hemorrhage, hut no kidney of all animals exposed to 2.0 gm ejand two oil. traces of foreign material. product were not different from (hose rjhours. ('arise Sample C is a PVC homopolymer pro control animals. The lungs of some but m3 high as 30 v duced by another company. The results of all animals exposed to 2.5 to 2.S gin cl rats exposed exposure to the pyrolysis products of this pyrolysis product showed focal edema widgm of PVC fc material are shown in Table 3. Exposure areas of acute suppuration and occassior.3ately after ex ranged from 1.2 gm where all animals sur-1 intra-alveolar hemorrhage. I for histopnri vived, to 2.3 gm where all animals died. In Sample E is a formulation of polymer C-jwas no evrc.c general, COHb levels correlate well with the commonly used in wire insulation applies! exposed grou degree of exposure and the mortality data. tions. The results of this study are shown a Sample r. The acute LC-,n under these conditions is Table 5. Even at levels as high as 3.0 gm c` tion of samp approximately 2.0 gm of pyrolyzed plastic. formulation, CO levels did not reach l,CC-5 of inert mate Surviving animals were killed 24 hours af ppm in the air stream, compared with vs1.) 6.0 gm o. - ter exposure. Tissues of exposed animals ues as high as 3,000 ppm with 1.2 gm ci total residue were comparable to control animals. Serum homopolyiner. Hydrochloric acid levels y of the orrgu SCOT levels, however, showed some eleva the 4.0-gm sample air stream were in c.xcesS 1-4 l tions suggestive of tissue damage. of 500 ppm. All animals survived exposure^ posed. All The results of the exposure of animals to ranging from 3.0 to 4.0 gm of pyrolyzed vol%. Tne the pyrolysis products of the vinyl chloride- material. Carboxyhemoglobin levels did r.oi after expos vinyl acetate copolymer (sample D) are rise above 23 vol% in any of the exposed lungs and S shown in Table 4. The pyrolysis of Quantities greater than 2.0 gm resulted in deaths of animals even though they had relatively low COHb levels. Hats died after exposures to pyrolysis products of 2.5 and 2.8 grn of sample animals, and no animals died during the controls. exposure period. Animals not killed for The pyrol COHb levels svore maintained for 2-1 hours, (wire forma then killed and examined for histological, the 16 rats \ evidence of lung damage. In all cases, l!^ cal cvidcnc Arch F.nuirnr. Health--Vol 19, July 1969 Cm Cl V-* 4 o i PYROLYSIS PRODUCTS--CORNISH & ATIAR 19 and Table 3.--Effect of Pyrolysis Products of Polyvinyl Chloride Homopolymer (Sample C) .V.ortaht) 0 4 1 Quantity Pyrolyaed (am) Weight of Residue (m) 1.2 <0.01 1.5 0.02 Exposure Time (min) 130 135 CO Max Cone (ppm) >3.000 >3.000 Rat No. Olood Lung Hb COHb (%) (gm/100 cc) (vol %) (Body Weight) SGOT (units) Mortality!1 1 15.0 7 2 15.0 0 3 16.2 0 ... 4 14.1 0 0 ... 4 1 20.4 30 0.38 2 12.3 64* 0.38 3 13.7 42 0.47 4 14.5 24 0.46 69 127 124 1 4 * i 15.3 42 0.50 99 1.8 ... 135 >3,000 2 3 15.7 14.5 28 0.51 38 0.58 132 123 0 4 4 4 15.0 15 0.47 109 1 13.7 46* 0.41 2.1 <0.01 130 >3.000 2 3 17.6 17.5 21" 44* 0.47 0.50 2 > 4 1 16.2 . 41- 0.50 134 2 15.3 30* 0.50 103 4 2.3 0.02 135 >3.000 3 15.3 39 0.56 4 4 15.7 61* 0.52 164 3 14.5 volf( Controls 6 rats 0.53 67 .ppear to b or the death Animal deaths. tNumerator, deaths; denominator, total. shortly aft* lungs of exposed animals were comparable these animals. The results of the exposure .ts exposed: to controls. to pyrolysis products of 6.0 gm of sample G suggestive c Three of four rats died when exposed to are included in Table 2 since this is a the pyrolysis products of 6.0 gm `of this formulation of sample B. Even at this high ?. , a* formulation. One died during the exposure level of exposure, no animals died and all to g.u gm c and two others died within the next 16 COHb levels were below 10 vol%. rom those c hours. Carboxyhemoglobin levels ranged as some but r.t high as 30 vol%. An additional group of Comment 0 2.S gm c rats exposed to the pyrolysis products of 6.0 The toxicological findings in pyrolysis 1 edema \vh gm of PVC formulation were killed immedi studies will depend, to a large extent, on d occassiom ately after exposure and lung sections taken methodology. This'includes such factors as for histopathological study. ' Again, there presence or absence of oxygen during com of polymer C was no evidence of lung pathology in the bustion, air flow rates, dilution of pyrolysis ation applic exposed group of animals. are shown: Sample F, which is a floor tile formula a as 3.0 gm( tion of sample D, contained large amounts ot reach 1,0: i of inert material. Pyrolysis (up to 550 C) of a red with va 6.0 gm of this material (Table 4), left a ith 1.2 gm i total residue in the boats of 4.6 gm of 76% acid levels i ' of the original material. Thus, only about were in exec 1.4 gm of material was actually decom ved exposuR posed. All COHb levels were below 5.0 of pyrolyzf vol%. The animals were killed 24 hours levels did r.: after exposure. Histological sections of if the expos lungs and SGOT levels were comparable to ed during ti controls. aot killed f. The pyrolysis of 1.2 to 2.5 gm of sample G 1 for 24 hou; fu-ire formulation) produced no deaths among or histologic the 16 rats used in these studies. No histologi cal evidence of lung damage occurred in stream, presence of particulates, and type of exposure chamber utilized. In spite of the variables involved, certain valid compari sons can be made utilizing standardized conditions of pyrolysis and exposures. In several of the early studies described here, animals were exposed in a small desic cator in which urine and feces accumulated during the exposure period of one to three hours, and vapor collected on the chamber walls. Measurement of HC1 gas at the inlet and outlet of this chamber indicated a con siderable loss of HC1 as the air stream passed through the chamber. Although most of this loss may have been due to solution in liquid, some may also have deposited on the animal's fur. Subsequent studies utiiiz- Arch Environ Health--Vol 19, July 1969 JO w JO C: Cn o H- JO < ----------------- --------------- -H ' \ 20 PYROLYSIS PRODUCTS--CORNISHABAR Table 4.--Ellecl oI Pyrolysis Products ol a Vinyl Chloride-Vinyl Acetate Copolymer (Sample O) and Formulation (Sample F) /*Qua ntity Pyrolyzed (gm) Weight of Residue (S'")* Exposure Time (min) CO Max Cone (ppm) Rat No. Blood Hb (gm/100 cc) COHb Lung (%> (Body Weight) SGOT (units) Mortality! Quantity Pyrolyzed (gTM) We R6 f 0> Sample D 2.0 0 2.5 0.16 112 103 8,000 >3.000 5 min >3.000 5 min i 2 3 4 1 2 3 4 12.0 13.7 15.3 14.3 13.3 ... ... 14.5 12 18 12 8 56t 13t 21t S4t 0.57 0.58 0.56 0.52 0.60 1.38 cyst 0.53 0.48 100.5 84.0 143.5 135.5 .. ... ... 0 4 ; 4 4 3.0 C 3.5 c ...2.8 0.10 105 7.000 >3,000 5 min I 2 3 14.1 13.0 32 50t ...t 0.53 0.60 0.58 i, 4 4 12.3 61+ 0.99 4.0 c Sample F 6.0 4.6 130 >3.000 5 min 1 2 '3 4 14.5 14.1 15.3 15.0 ' 5 5 5 s 0.48 0.55 0.50 0.52 80 55 0 72 4 56 6.0 ' Controls 12 rats 0.60 . 67 Residue in boat plus residue in combustion tube. tAnimal deaths. ^Numerator, deaths: denominator, total. k t / ing an air stream into which only the ani- ' large amounts of pyrolysis products, the mal's nose protruded overcame these diffb animals died quickly due to carbon monox- culties. ide poisoning, thas the cffccLs of HG1 could 6.0. : Control Residue in b iAnirnat <Jcatk ^Numerator. < Without quostion, the major cause of not be demonstrated. However, when onideath in this study of four polymers was the mals.did survive exposure to lesser amounts exposed ani: be pyrolyz presence of relatively large amounts of carbon monoxide in the pyrolysis air stream, The blood carboxyhemoglobin levels, in general, correlated well with the amount of of pyrolysis products, the histopathological findings in the lung were not remarkable. Ir. the two studies where sufficient 0- was .. added to the airstream so that animals did COHb levelThree P studied. The sample E cc pyrolyzed PVC. However, rather marked not succumb to CO effects, pulmonary edc- much chlor: exceptions occurred within groups. For ex- ma and interstitial hemorrhage were noted . homopolym-: ample, in three of four rats exposed to 1.2 gm of polymer A, COHb levels were above 65 vol%. In the fourth rat, the. blood COHb level was only 11.4 vo!%. Other similar in all animals. In one instance, this was associated with the presence of angular fibers and sheets of foreign material within the lung. No attempt was made to remove 5 about 50% The analytic much CO v In the press irregularities can be noted throughout the particulate matter from the air stream. Fer- amount of r. study. When animals were exposed in a haps in this one instance, a sudden crupdcsiccator chamber, it was felt that animal tion of material from the pyrolysis boat produce urn exposed ani crowding might have played a role in these could account for the findings of particulate inconsistent findings. However, similar matter in the lung. 6.0 gm of P' imum COH findings will be noted in the subsequent The lungs of some animals exposed to 2.5 y posed anim.a studies when only the animal's nose was to 2.8 gm of pyrolysis product of copolymer ^of 24 to 64 - exposed to the air stream. Perhaps a partial D showed focal edema with several areas of . of the h explanation may relate to the animal's re- acute suppuration and occasional intra-al- t^four animal- spiratory rate at the time of high CO levels veolar hemorrhage. It is interesting to note biological star in the air stream. This period of high CO that the analytical data of Boottnor et cil10 ^yjexposed gro levels was relatively short compared to the indicate that this pyrolyzed sample (.Dl has Otro!s- thus total exposure time, CO levels being greater HCl concentrations quite comparable to *juipurent. Ano> than 3,000 ppm for less than 20 minutes samples A. 13, and C, but CO production is : ly gospasm. cue. during most exposures, and in some eases such levels may have been reached for only a few minutes. With exposure to relatively approximately one half [hat of the previous samples. Thus in this instance the irritant effects of HCl U'unncs more pronounced in : : This hiveelie Hcahh s NneomU IV. . T,v'..:' Ball. Arch Environ Health--Vo! 19, July 1909 f HOI.YS1S l'lt(JfjlICTS --COIiNIKII A Ar.Alt 21 met (Sample D) and Table 5.--Ellcct o! Pyrolysis Products ot Sample E (Formulation ot Sample C) oGOT ight) (un;ts) 100.5 S-.O 1-13.5 135.5 St ... . ... ' 80 55 72 56 67 -j ' i jMortality; Quantity Pyroly/ed (i">) 0 t t 3.0 7J f . 4 i i.c i t t 3_______ | 4; t 4.0 i o ! 6.0 4 6.0 Weight of Residue (">)* 0.70 0.98 0.96 1.20 1.19 Exposure Time (min) 135 155 135 140 137 CO Max Cone (ppm) 850 1015 mm S>0 10-20 men 2.000 3-5 min 2,000 5*10 min 3,000 few min Rat No. 1 2 3 4 I o 4 i 2 34 1 2 3 4 i .2 3 4 HI) (Em/100 cc) Lung COH b SCOT (vol %) (Body Weight) (units) 13.3 17 0.57 83 12.3 1 l 0.57 74 13.3 18 0.55 too 14.1 12 A 54 to l 13.6 13.3 12.7 <s O / U.M> 7 0.51* 9 0.54 K1 fit. Hi 15.3 23 0.50 14.1 7 0.54 14.5 2 0.53 11.7 4 0.55 67 67 83 93 16.6 14.5 18.1 18.1.. 20 301 71 241 0.75 ... 0.43 98 ... 17.1 14.1 13.7 13.7 15 0.51 11 20 0.51 10 0.57 1 95 72 151 Mortality t 0 4. It 4 r. 4 * 3* Killed Control 8 rats 0.54 59 -olysis products, tht Residue in boat plus residue in combustion tube. ue to carbon monox- tAnimal deaths. ^Numerator, deaths; denominator, total. effects of HC1 coulc References However, when ani- exposed animals since larger quantities can 1. Berger, L., et al: Toxicity and Flame Resist- ure to lesser amount- be pvrolvzed without producing lethal once of Thermosetting Plastics, report 4134. US Dept of the Interior, Bureau of Mines, Oct 19;7. the histopathologics. COKb levels. 2. Watson, H., et al: Thermal Decomposition re not remarkable. It Three PVC formulations were also Products and Burning Characteristics of y.-nc .Synthetic Low-Density Cellular Materials, report 4777. re 'cient O- wa- studied. The analytical data0 indicated that US Dept of the Interior. Bureau of Mines. 1951. i ..iat animals die sample E contained approximately 50% as :ects, pulmonary ede- much chlorine per gram as do the PVC morrhage were notec homopolymers, indicating the presence of 3. Zapp, J.A., Jr.t Hazards of Isocyanates in Polyurethane Foam Plastic Production, Arch Indusir Health 15:324*330 (April) 1967. 4. Waritz, R-S-* and Kwon, B.K.: The Inhalation e instance, this wa- about 50% of materials other than PVC. presence of angula: The analytical data also indicated 30% as reign material withir. much CO was produced from this sample, Toxicity of Products cf Polycetraf.uoroethyler.e Heated Below 500 Degrees Centigrade. Amer Industr Hyg As$oc J 29:19-26 (Jan-Feh) 1966. 5. Coleman, W.E., et al: The Identir.caticn o: was made to removt In the present studies, doubling the usual Toxic Compounds of Polvtetraducroethyler.e (PTFE), Amer Industr Hyg Assoc J 29:33-10 i the air stream. Per- amount of material to be pyrolyzed did not (Jan-Feb) 1968. tnce, a sudden erup produce unusually high COHb levels in the a the pyrolysis bos exposed animals. In fact, the pyrolysis of ir.dings of particular 6.0 gm of PVC formulation resulted in max 6. MacFarland, H.N., and Leong, K.J.: Hazards From the Thermcdecomposition of Plastics, Arch. Er.viron Health 4:591-597 (June) 1962. 7. Thrune, R.I.: Gases Released When Fire Re imum COHb levels of 30 vol% in the ex jaimals exposed to 21 posed animals, compared with COHb levels product of copolyme- of 24 to 64 vol% after pyrolysis of only 1.5 sistant Epoxy Pvesins are Burned. Amer Industr Hyg Assoc J 24:475-480 (Sept-OcO 1903. 8. Eeong, K J., and MacFarland, H.N.: Hazards From Thcrmodecompo.sitton of Epoxy Resins, Arch i with several areas c gm of the homopolymer. However, three of d occasional intra-a; four animals in this group died. Histopathis interesting to not ological studies showed that the lungs of the .ta of Boettner et al\ exposed group were not different from conilyzed sample (D) hs trols, thus the cause of death is not ap- Environ Health 7:075-681 (D^c) 1963. y. Boettner, E.A., and Weiss, B.: An Analytical System for Identifying the Volatile Pyrolysis Prod ucts of Plastics, Amer Industr Hyg Assoc J 25:535540 (Nov-Dec) 1967. 10. Boettner, E.; Ball, G.; and Weiss, B : Analy sis of the Volatile Combustion Products of Vinyl quite comparable i parent. Anoxia, resulting from irritant larynbut CO production i gospasm, could be a factor in these mortalities. If that of the previou This investigation was supported in part by Pub Plastics, J Appl Polymer Science, to be published. 11. Scholander, P.F., and Roughton. F.J.W.: Sim ple Micro-Gasometric Method of Estimating Carbon Monoxide in Blood, J Industr Hyg Toxic 24:215-221 s instance the irritar. lic Health Service research grant UI-00485 from the more pronounced i: National Center of Urban and Industrial Health. Technical assistance was given by Gwendolyn (Oct) 1942. 12. Steinberg, D., et al: A Clinical Method fer the Aassay cf Serum Glutamic-Oxalacctic Transaminase, 1UU. J Lab Clin Med 45:144-151 (July) 1956. Arch Environ Health--Vol 19, July 1969 *4 C/l K <T. c?: o H-