Document BvdKDbQaq5qwb9zoJ200kmyEw

f TECHNICAL INFORMATION Fire Properties of Polyvinyl Chloride Figure 1 Ignition Temperatures Ifemptntura, C* 70S PVDC 'Phenolic ' Glass Fiber Melamine 500 / gijss Fiber Nylon 'PVC r Polystyrene (Foam) r ' Polyethylene 'SAN ' Polyurethane ^ Cellulose Acetate 300 Polystyrene (Beads) " PMMA Douglas Fir - -%\ Cotton x ^ Pine 'Paper ' Wool - Cellulose Nitrate Melamine ' Glass Fiber Phenolic Glass Fiber /PVDC 'PTFE Polystyrene , SAN ' Cellulose Acetate PMMA ' PVC ' Nylon ' Polyurethane Polyethylene Pine Cotton Paper Polyvinyl chloride (PVC, or vinyl) pos sesses excellent fire performance proper ties. In particular, it will not bum once the source of heat or flame is removed. This results from PVC having 56.8% chlorine in its base polymer weight. It is well known that chlorine is one of the few elements that confers good fire properties to a polymer.1,2 Samples of unplasticized (rigid) vinyl, such as those found in pipe, siding or ver tical blinds, have higher ignition tempera ture, lower flame spread and lower heat released in a fire than similar samples of wood. When PVC is plasticized to make it into flexible products such as wire coat ings, upholstery, medical blood bags or wall coverings, the fire properties become less favorable, depending on the amount and kind of plasticizer and other additives used. However, most plasticized PVC products in use will not continue to bum once the flame source is removed, even if not additionally fire-retarded. The actual fire properties of PVC have been calculated from the results of smallscale and full-scale tests, and interpreted in terms of overall fire hazard. Cellulose Nitrate FIRE HAZARD Flash Ignitloa MMhMm The fire hazard of a product is deter mined by a combination of factors includ ing its ignitability and flammability, the amount of heat released from it when it bums, the rate at which this heat is re leased, the flame spread, the smoke pro duction and the toxicity of the smoke. Ignitability If a material does not ignite, it will not contribute to fire hazard and thereby can not endanger lives. Most organic materials do, however, ignite and the lower the igni tion temperature, the greaier the hazard. Figure 1 presents the self-ignition and flash-ignition temperatures of a variety of common materials, as measured by a stan dard test (ASTM D1929). With a flash ignition temperature of 346C, PVC is among the least easily ignited polymers. Flammability Once ignited, the greater the flammabil ity of a material, the greater the hazard associated with it. One of the most reliable quantitative small-scale flammability tests is the limited oxygen index test (LOI; ASTM D2863), which gives the limiting concentration of oxygen in the atmosphere necessary for sustained combustion. In general, a material with LOI of more than 21 should not bum in air at room tempera ture, and, one with LOI of more than 25-27 will only bum under extreme condi tions. Table 1 shows that very few com mon materials have an LOI higher than PVC. When PVC is plasticized in order to achieve flexibility, its LOI is lower than that of rigid PVC; the exact value will de pend on the formulation. Phthalate plasti cizer affects the LOI of PVC as shown in Figure 2. Because typical wire and cable compounds contain only 15-35 percent plasticizer, most will not bum in air unless a source of heat or flame is applied. Flame Spread The tendency of a material to spread flame can be measured using the radiant panel test, ASTM E162. Results from this test (Table 2) show that PVC is one of the materials with the lowest flame spread rating; it will not spread flame on its own. cTL030248 ^bV'n." . -i ' ' 'T\ Wayne inie'i rjeP-azeii "tt. irdi,si'v rc \`ew Jemev 074 70 /r'~> f Table I2'4 Limiting Oxygen Indices of Various Materials Material Polyacetal Polyoxymethylene Cotton Natural rubber PMMA Polyethylene Polypropylene Polystyrene Polyacrylomtnle SAN ABS Rayon Cellulose PET PVF Nylon 6,6 Wool Polycarbonate Neoprene rubber Modacrylic Nomex Polysulphone Leather (FR) Polyimide PVDF PVC (rigid) Carbon black rod PVDC Chlorinated PVC PTFE LOI 14 9 15 7 16-17 17.2 17 4 17.4 17.4 17.6-18.3 18.0 18.0 18.3-18.8 18.7-18.9 19.0 20.0 22.6 24-29 25.2 26-28 26 3 26 8 28.5 30-32 34 8 36.5 43.7 45-49 59-63 60 0 60-70 95.0 Note: The fire performance for this test improves as the LOI becomes higher. Heat Release A burning object will spread a fire to nearby products only if it gives off enough heat to ignite them. Moreover, the heat has to be released fast enough not to be dissi pated or lost while traveling through the cold air surrounding anything not on fire. Therefore, the most important flamma bility property, in terms of fire hazard, is the heat released. Figures 3 and 4 present normalized results for total heat released and maximum rate of heat released for a variety of materials, as measured by an Ohio State University rate of heat release calorimeter (OSU-RHR. ASTM E906).6 They show that most materials are more prone to ignite other products than PVC. Smoke The amount of smoke that materials produce when they bum also is important because smoke obscures light and hinders escape from a fire. The most common method for measur ing smoke from burning products is the U. S. National Bureau of Standards (NBS) smoke chamber in the vertical mode (ASTM E662), and results from it are shown in Figure 5. This test is now known, however, to be seriously flawed; the principal deficiencies identified are shown in Table 3. The most important of the problems with the NBS smoke cham ber is that its results misrepresent the smoke found in real fires. One example of this is the effect of sample orientation. Some materials melt or drip when exposed to flame. When samples of such materials are exposed ver tically in the NBS smoke chamber test, the molten portions will have escaped the effect of the radiant heat source. This means that some of the material does not bum during the test. If these dripping products are exposed horizontally, the en tire sample will be consumed. The test, therefore, measures an amount of smoke artificially lower than what would be formed in a realistic scenario (see Table 4). PVC, however, does not melt or drip and the test, thus, gives the same smoke production results in the vertical or hori zontal orientations. Therefore, when the smoke results of other materials are cor rected for the melting effect by changing the orientation, the relative ranking of PVC in terms of smoke production changes. In fact, its smoke production fig ures become very similar to those of other materials regarded as low smoke produc ing, such as polyethylene or polypropylene (Table 5). Table 24 Surface flammability of some materials Material Chlorinated PVC Polyether sulphone PVC Polyester FR polystyrene FR polycarbonate Polycarbonate Red oak Phenolic resin Plywood (fir) Hardboard GRP polyester (21%) FR acrylic Polystyrene Acrylic Polyurethane loam (flexible) Polyurethane loam (rigid) Thickness (mm) 3 3 4 3 3 6 3 19 2 6 6 2 3 2 6 Flame spread index 4 5 10 30-56 59 73 88 99 114 143 185 239 316 355 416 1490 CTL030249 2220 Note The tire performance tor this test improves as the tlame spread ratings become lower Figure 2s PlasUchar (Dtoctyt PMkalate) Weight % Figure 4 Maximum Rate of Heat Release (OSU-RHR) -0 Gypsum Board--: :--PVC (Flexible) Vinyl Asbestos Tile--: --"Oak-Pine ----- Exterior Plywood Polycarbonate -- -1000 ----- Polystyrene (FR) Hardboard----- -2000 ----- ASS Polystyrene----- -- Polypropylene -3000 PMMA----- At an Incident Heat Flux of 2 W cm-2 (in BTU min-' ft 2) Figure 3 Total Heat Released (OSU-RHR) 0 PTFE----L-Wool PVC (Rigid)--- (Carpet) PVC (Flexible) Oak----- -- Particleboard Pine----- ----- Polyurethane (Flexible Foam) ABS----- -- Polystyrene -7000 PMMA------ Polypropylene At 10 Min, per Unit Surface, at the Limiting Incident Heat Flux, per Unit Incident Heat Flux, (in BTU ft 2/W cm-2) Figure 5 Maximum Smoke Density Measured in the NBS Chamber (0,1 PTFE (NF) - PTFE (F) Red Oak (F) Polypropylene (F) - Pine (F) - - ABS (NF) Walnut (F) . Polyethylene (F) Douglas Fir (F) Nylon (F) Pine (NF) Douglas Fir (NF) - Polypropylene (NF) PVC (NF) - Natural Rubber (NF) - Nylon (NF) - Polystyrene (NF) -Red Oak (NF) - Walnut (NF) - Polyethylene (NF) -PVC(F) -700 ABS (F) - - Polystyrene (F) ' Natural Rubber (F) F = flaming mode, NF = nonflaming mode Even when corrections are made, how ever. the NBS smoke chamber cannot ad dress the fact that fire hazard from smoke depends on how much product bums and how much heat is associated with it. This can be done by measuring smoke produc tion from materials using a smoke para meter.8 This is a test, developed by NBS. which measures the amount of smoke pro duced in the form most relevant to fire hazard by combining the light obscuration with the rate of heat released. It has been measured411 on another, newer rate of heat release instrument: the NBS Cone calorimeter (ASTM E5- Proposal P-190).12 Results from this instrument cor relate with full-scale fires.13'15 Table 6 indicates that standard rigid PVC has a smoke parameter less than one tenth of that of other plastic materials such as polyester, polyethylene, or polypropylene, with low smoke production numbers in the NBS smoke chamber test (Figure 5). Flex ible PVC compounds have higher smoke parameters than rigid ones, due to the effect of the plasticizer. Mass Loss In addition to heat and smoke release, fire hazard also depends heavily on the re lease of toxic combustion products into the atmosphere - that is, on the mass loss rate of a material. If a product loses mass very rapidly, a hazardous accumulation of dangerous smoke will develop more easily. The rate of mass loss for most flame retarded materials is relatively simi lar. The other results shown in Table 6 rates of mass loss and of heat release also measured with the NBS Cone cal orimeter, indicate that PVC has both a very low maximum rate of heat release and a low rate of mass loss. Toxicity Finally, fire hazard also is associated with the toxicity of the smoke itself. Fig ure 6 presents some results on the toxic potency of the smoke of a variety of common materials, as measured by the NBS cup furnace toxicity test,18 and com pares them with the intrinsic toxic potency of other poisons and toxic gases, as well as with textbook toxicity categories.|y CTL030250 This figure shows that toxicity is a rela tively minor factor in fire hazard since the intrinsic toxic potency of the smoke of the majority of common materials is very similar, with very few exceptions. This is because the most important toxic product in any fire is carbon monoxide (CO), which is produced by all organic materials when they bum. Furthermore, there is wide agreement today that small-scale smoke toxicity tests can only show broad toxicity categories and cannot really dis tinguish between most materials. This indicates that the toxicity of the smoke of most materials, including PVC, is simply a consequence of the amount of smoke produced. The only exceptions are those very few materials which have an exceptionally high toxic potency. That is why a low mass loss rate will generally mean a low toxic fire hazard, almost irre spective of the toxic potency of the smoke itself. Table 37'10 Deficiencies in the NBS smoke chamber Results do not correlate with full-scale fires. Vertical orientation leads to melt and drip. Time dependency of results cannot be established. No means of weighing sample during test. Maximum incident radiant flux is 25 kW/m2 Fire self-extmguishes if oxygen level becomes <14%. Therefore, composites often give misleading results. Wall losses are significant. Soot gets deposited on optics. Light source is polychromatic. Rational units of m2/kg are not available. Table 4 Effect of Orientation on Smoke Density (NBS Chamber)7 (maximum smoke density; flaming mode) Horizontal Polypropylene 398 Polyethylene 286 Nylon 264 Parattin Wax 228 Vertical 57 35 48 83 HEALTH EFFECTS OF HYDROGEN CHLORIDE The majority of combustion products given off by PVC are the same as those produced by wood or most other common materials, both natural and synthetic.' 3 The one product given off by PVC that is not given off by natural materials is hydrogen chloride (HC1). Studies have been made on the suscept ibility of different animals to various toxic gases. These studies have found that rats are reasonable models for primates as far as incapacitation and exposure dose for post-exposure lethality of irritants (like HC1) is concerned.20'21 The rat is also a good model for asphyxiants such as CO.21"22 On the other hand, the mouse has been shown to be much more sensitive to irritants (particularly HC1) than rats23 25 or baboons.25 Table 7 shows the lethal doses of some of the most important toxic gases present in fires, as measured in those animal species with responses most similar to those of humans - namely, rats and baboons.21 22- 26- 27 These data show that although the mechanisms of action of CO and HQ are totally different, their lethal doses axe very similar. However, HCl has an important feature related to fire hazard; a very pungent odor, detectable at a level of less than 1 ppm,28 while CO is odorless and narcotic. Therefore, HCl will signal people in a fire atmosphere to escape, while CO will narcotize them. Table 7 also shows the highest concen tration of these gases found in two studies involving fire fighters equipped with monitoring devices. Interestingly, the peak CO concentration found when they entered burning buildings was higher than that known to cause lethality at 30 minutes. Similarly, the peak acrolein concentration found is equivalent, after a 30 minute ex posure, to a dose in the range of the 5 minute lethal dose for baboons. On the other hand, the peak HCl concentration found was less than one tenth of the cor responding 30 minute lethal value. This data supports, again, the observation that CO is the prime toxic hazard in a fire. CTL030251 Table 5 Smoke Density (NBS Chamber) Corrected for Melting Effect7 Polystyrene Polypropylene PVC Polyethylene Paraffin wax maiimum optical density (cm2/g) 14,300 5,250 3,400 2,900 2,300 Table 64,11,17 Heat, Smoke and Mass Release Parameters from NBS Cone RHR Calorimeter (Horizontal exposure*) Material 1. Polycarbonate 2. Aircraft panelM 3. Low smoke rigid PVC 4. PVC 5. Low smoke FR PVC wire cpd 6. Carpet" 7. PVC (extrusion) 8. Rigid PU foam" " 9. Particle Board"c 10. PPO/PS with fiberglass 11. Standard PVC wire cpd 12. FRABS 13. PPO/Polystyrene alloy 14. FR Polystyrene 15. FR ABS (with PVC) 16. FR polyurethane TPU 17. PMMA"C 18. Flexible PU foam" " 19. ABS 20. Polystyrene 21. EPDM/SAN 22. Polyester 23. Polyethylene 24. Polypropylene Maximum Rate Heat Release kW/m2 22 40 89 91 92 100 115 160 180 184 204 250 263 315 445 509 650 650 746 859 883 1216 1325 1335 Average Specific Extinction Area at 5 min m2/kg 0 - 47 380 159 - 340 - - 903 1099 1874 1561 1747 1312 606 - - 863 1027 1016 460 337 461 Maximum Mass Loss Rate g/(s m2) 5.0 5.6 8.0 13.7 19.3 5.5 12.7 7.3 12.6 9.1 14 0 17.7 11.3 21.5 11 6 12.6 25.0 28.5 24.2 26.5 24.8 59.4 24.7 26.7 `Incident Flux: 20 KW/m2; thickness: 0.63 cm "Incident Flux: 25 kW/m* Thickness: 1.28 cm Thickness: 2.54 cm Smoke Parameter kWkgxIO"8 0 - 004 0.35 0 15 _ 0 38 _ ' 66 2.23 4 68 4 10 5 50 5 84 3 08 _ _ 6 44 8 82 8.97 5.59 4.46 6.15 Materials: 1,6,8,9,17,18 - Reference 17; 5,11 BFGoodnch compounds; others - References 4,11. Lexan 141-111 polycarbonate; composite with phenolic-polyamide honeycomb and tedlar coating; low smoke PVC sheet extrusion compound; general purpose PVC custom injection moulding compound; PVC fire retarded flexible wire compound designed for low smoke; wool-nylon carpet with rubberized backing; weatherable PVC extrusion compound; GM31 rigid polyurethane foam'8; 0 5 inch particle board; Noryl GFN-3-70 PPO/PS with fiberglas, standard flex ible PVC wire and cable compound (non fire-retarded) Cycolac KJT ABS; Noryl N-190 polyphenylene oxide/ polystyrene. Huntsman 351 FR polystyrene. ABS fireretarded with PVC. FR thermoplastic polyurethane; 0.5 inch Rohm and Haas black poly (methyl methacrylate). 0.5 inch low density flexible polyurethane foam; Cycolac CTB ABS; Huntsman 333 polystyrene; Rovel 701 EPDM SAN copolymer; Celanex 2000-2 polyester, Marlex HXM 50100 polyethylene, Durro 8938 polypropylene CTL030252 f Table 7 wire insulation (containing the chiorine equivalent of 8.700 ppm of HCI) was elec Lethal Exposure Doses for Common Gases trically decomposed in the presence of sorptive surfaces. The peak concentration Gas CO Acrolein HCI HCI HCN LEO ppm min 192,000b 2,500-5,000c 112,000-169,000" ca. 150,000' 4,800' Animals rats baboons rats baboons rats 0DL* ppm" 0.16 0 77 0.77 0 58 Peak in Fire ppm2*" 7,450 98 280 280 9 of HCI found in this simulated plenum was only 10% of the theoretical concentration. One of the consequences of these studies is that toxicity tests carried out in glass or plastic exposure chambers may exaggerate the toxicity of PVC smoke. On `Odor detection level "30 min exposure; within exposure deaths; Reference 27 c5 min exposure; post-exposure deaths; Reference 21 '30-60 min exposure; post-exposure deaths; Reference 20 *5-/5 min exposures, with no deaths; Reference 20 '30 mm exposure; within exposure deaths; Reference 26 these surfaces, HCI does not decay as fast as on real-use surfaces, so that toxic gases are present longer than in real fires. PVC PERFORMANCE IN LARGE SCALE TESTS Table 8" Results of Corner Bum Room Fire Test* Material Crib only Low smoke rigid PVC Chlorinated PVC Rigid PVC FR Acrylic FRABS Wood (oak) Polycarbonate Temp. at door (C) 171 178 169 188 322 748 558 382 Heat generated at 13 minutes (MJ) 32.6 25.6 30.2 31.7 36 6 70.2 89.8 133.7 Peak smoke at door (0D/m)b 1.6 4.6 1.5 8.3 >15.1 >15.1 9.6 >15 1 `A 6.3 kg wood crib was used in all experiments. "ODIm: optical density per meter. cDm: maximum smoke density. Smoke yield (g) 106 187 93 384 398 >900 746 >2500 NBS Smoke Dmc 94 53 780 435 900 106 247 HYDROGEN CHLORIDE DECAY A series of studies was done to investi gate the "lifetime" of HCI in a fire atmos phere.31'34 These studies showed that HCI reacts very rapidly with most common construction surfaces (cement block, ceil ing tile, gypsum board, etc.). Therefore, the peak HCI concentration found in a fire is much lower than would be predicted from the chlorine content of the burning material. Moreover, this peak concen tration soon decreases and HCI disappears completely from the air. Figure 7 shows the HCI concentration-time pattern for several identical experiments where PVC It is probably of much greater impor tance to understand how PVC behaves in a real full-scale fire than to understand the results of small-scale tests, some of which are of questionable validity. A series of tests was conducted in which a comer. floor to ceiling, was paneled with different materials including wood and rigid PVC. Table 8 shows that the PVC panels added nothing whatsoever to the total heat generated and next to nothing to the temperature measured at the door of the room, as compared to the ignition source itself (a wood crib).35 The smoke emitted by the burning PVC caused the room to become somewhat darker than when the crib burned alone, however, the darkness was much less than that generated when the wood panels were burned. The wood also gave off a very considerable amount of additional heat, much hotter temperatures at the door, and a large mass of smoke particulates. Other materials, with the ex ception of chlorinated PVC (CPVC), simi larly generated much more smoke and heat than the PVC panels. It is also interesting to note that the smoke yield in the full-scale tests did not correlate with that in the NBS smoke chamber. The reason for this lack of cor relation is that the largest smoke yield in the NBS smoke chamber was given by those materials burning least extensively in the large-scale test. CTL030253 f Figure 6 Toxic Potency (Lethal Dose) of Substances and of Smoke (LDM in mg/kg) -100,080 Toxic -10,000 SUfhthr Toxic" y^Ejanat (16,000) Moderately -1,000 Toxic . z--JT4 . \?`V_ ' Ethylbenzene (6,250) Naphthalene (2,500) -Methanol (425) 'Aspirin (250) T OPT (OP) Extremely -10 Toxic - Phenobarbital (27) . Paraquat (7) - -------- & /Nlcotlnefl) Cocakw<til5) Strychnlne(0^7) Tiger Snake Poison (0.048) : 0todh(0.flSS) 8sa Snake feTotson (0.0t2) . . CO, (1,400) - CFCII (940) Chloroform (300) Ether (273) HCI (27) CO (22) HCN(0.75) ;TDI(0.5) ` Acrolein (0.4) ' Phosgene (0.17) Categories Botuknus 'Toxin (0.0002) Poitoes Gases Smokes Figure 7 Experiments Fitted Values Large-scale experiments have also been earned out to investigate HC1 decay. These show that, even if massive amounts of PVC are decomposed in a plenum space above a room, no HC1 filters down into the room below unless dnven by an air conditioning system, while other gases do accumulate m the room.11' 14 Even with an air conditioning system, the concentra tions in the room were found to be of vir tually no toxicological concern. Conse quently, high concentrations of HC1 are unlikely to reach rooms other than that of origin and unlikely to affect victims in the post-flashover period. Two other interesting studies have in vestigated the effects of burning PVC electrical products behind a wall (nonmetallic tubing)36 or a ceiling (conduit, non-metallic tubing, wire coating).17 It was found, in both cases, that the tempera tures and concentrations of toxic gases in the room would be lethal long before there would be any effect resulting from the combustion of the PVC products. SUMMARY 1) The most important cause of fire deaths is CO, produced when all organic mate rials bum. 2) PVC is less flammable than most or ganic materials and it will not continue burning unless there is a powerful ex ternal source of heat. 3) When PVC eventually bums, it gives off less heat and does so more slowly than most materials. 4) The smoke produced by burning PVC is within the same range as that of many other materials. 5) The toxicity of PVC combustion prod ucts is comparable to that of most or ganic materials. 6) PVC is unusual in that, when it bums, it releases HCI, which is irritating. However, HCI odor is easily detectable (at less than 1 ppm) and HCI does not incapacitate or become dangerous until it reaches concentrations much higher than those measured in real fires. 7) HCI is unique among common fire gases in that it decays by reacting very rapidly with most construction surfaces and consequently, is not transported easily to other rooms. CTL030254 REFERENCES 1. C.F. Cullis and M.M Hirschler, "The Combustion of Organic Polymers," Oxford University Press, Oxford (1981). 2. M.M. Hirschler, "Developments in Poly mer Degradation," Vol. 5, (G. Scott, edi tor), Applied Polymer. Barking. Chapter 5. p. 102(1982). 3. C.J. Hilado, "Flammability Handbook of Plastics," 3rd. Ed., Technomic Publishing, Lancaster, PA (1982). 4. M.M. Hirschler, Journal of Fire Sciences 5,289(1987). 5. J.M. Avento and I. Touval, "Flame Re tardants (Antimony)," Kirk-Othmer Ency clopedia of Chemical Technology, Vol. 10, 3rd. Ed., Wiley (1980). 6. E.E. Smith, "Ignition, Heat Release and Noncombustibility of Materials, ASTM STP502," (A.F. Robertson, editor), p. 119, American Society for Testing and Materials, Philadelphia (1972). 7. L.H. Breden and M. Masters, Journal of Fire and Flammability, 7. 234 (1976). 8. V. Babrauskas, Journal of Fire and Flam mability 12, 51 (1981). 9. J.G. Quintiere. Fire and Materials, 6. 145 (1982). 10. V. Babrauskas, SPE RETEC conference on "PVC: THE ISSUES." Atlantic City, p. 41 (1987). 11. G.F. Smith, SPE RETEC conference on "PVC: THE ISSUES." Atlantic City p. 124(1987). 12. V. Babrauskas, "Development of the Cone Calorimeter. A Bench-Scale Heat Release Rate Apparatus Based on Oxygen Con sumption," National Bureau of Standards. NBSIR 82-2611 (1982). 13. V. Babrauskas, "Bench-Scale Methods for Prediction of Full-Scale Fire Behavior of Furnishings and Wall Linings." Society of Fire Protection Engineers. Boston, Tech nology Report 84-10 (1984). 14. V. Babrauskas, Journal of Fire Sciences 2, 5(1984). 15. V. Babrauskas and J.F. Krasny, "Fire Safety. Science and Engineering. ASTM STP 882" (T.Z. Harmathy, editor), p. 268, American Society for Testing and Mate rials, Philadelphia (1985). 16. M.M. Hirschler and G.F. Smith, "Fire Safety Progress in Regulations. Technology and New Products," Fire Retardant Chemi cals Assoc. Fall Conference Oct. 18-21, Monterey, CA, p. 133 (1987). 17. W.D. Walton and W.H. Twilley, "Heat Re lease and Mass Loss Rate Measurements for Selected Materials," National Bureau of Standards NBSIR 84-2960 (1984). 18. B.C. Levin, A.J. Fowell, M.M. Btrky, M. Paabo, A. Stolte and D. Malek, "Further Development of a Test Method for the Assessment of the Acute Inhalation Toxic ity of Combustion Products," National Bureau of Standards NBSIR 82-2532 (1982). 19. L.J. Casarett, "Toxicology - The Basic Science of Poisons," (L. Casarett and J. Doull, editors), Macmillan, New York, p. 24, (1975). 20. G.E. Hartzell, S.C. Packham, A.F. Grand and W.G. Switzer, Journal of Fire Sciences 3, 195 (1985). 21. H.L. Kaplan, A.F. Grand, W.G. Switzer, D.S. Mitchell, W.R. Rogers and G.E. Hartzell, Journal of Fire Sciences 3, 228 (1985). 22. H.L. Kaplan and G.E. Hartzell. Journal of Fire Sciences 2, 286 (1984). 23. K.I. Darmer. E.R. Kinkead and L.C. DiPasquale, Journal of the American In dustrial Hygienists Association 35, 623 (1974). 24. H.L. Kaplan, R.K. Hinderer and A. Anzueto, Journal of Fire Sciences 5, 149 (1987). 25. H.L. Kaplan, M.M. Hirschler, W.G. Switzer and A.W. Coaker, proceedings of the 13th International Conference on Fire Safety (C.J. Hilado, editor). Product Safety Corp., San Francisco, p. 279 (1988). 26. V. Babrauskas. B.C. Levin and R.G. Gann, Fire Journal, 8/(2), 22 (1987). 27. G.E. Hartzell, A.F. Grand and W.G. Switzer, Journal of Fire Sciences 6. 368 (1987). 28. J.E. Amoore and E. Hautala, Journal of Applied Toxicology 3, 272 (1983). 29. W.A. Burgess, R.D. Treitman and A. Gold. "Air Contaminants in Structural Firefighting," N.F.RC.A. Project 7X008, Harvard School of Public Health (1979) 30. A.F. Grand, H.L. Kaplan and G.H. Lee, "Investigation of Combustion Atmospheres in Real Fires," U.S.F.A. Project 80027, Southwest Research Institute (1981). 31. J.J. Beitel, C.A. Bertelo, W.F. Carroll, R.A. Gardner, A.F. Grand, M.M. Hirschler and G.F Smith, Journal of Fire Sciences 4, 15(1986). 32. C.A. Bertelo, W.F. Carroll, M.M. Hirsch ler and G.F. Smith, Fire Safety Science, proceedings of the 1st International Sym posium (C.E. Grant and P.J. Pagni, edi tors), p. 1079, Hemisphere, Washington (1986). 33. J.J. Beitel, C.A. Bertelo. W.F Carroll, A. F. Grand, M.M. Hirschler and G.F. Smith. Journal of Fire Sciences 5, 105 (1987). 34. F.M. Galloway and M.M. Hirschler, "Mathematical Modeling of Fires. ASTM STP 983" (J.R. Mehaffey, editor), American Society for Testing and Mate rials. Philadelphia p. 35 (1987). 35. G.F. Smith and E.D. Dickens, proceedings of the 8th International Conference on Fire Society (C.J. Hilado, editor). Product Safety, p. 227, San Francisco (1983). 36. l.A. Benjamin. Journal of Fire Sciences 5, 25 (1987). 37. M.M. Hirschler, Journal of Fire Sciences 6,(1988). This report has been prepared by the Technical Committee of the Vinyl Institute as a service lo us members and their customers and is based on litera ture and information believed to be accurate No warranty or guaranty, expressed or implied, is made for the accuracy or completeness of the infor mation provided herein and neither the Vinyl Insti tute nor its members or contributors assume any responsibility for the accuracy or completeness of the information contained in this document. Figures 1. 3. 4. 5. 6 and 7 reprinted with the per mission of ihe Journal of Fire Sciences, 1987. Technomic Publishing. Lancaster. PA, USA CTL030255 c. 1988 The Vinyl Institute