Document da4ZXnr535oqE5M0a7G392eaQ

/)/) 3oc <_<. A Literature Study of the Combustion Hazards of Polyvinylchloride (PVC) and Acrylonitrile Butadiene Styrene (ABS) by JL ' JL* Judith E. Hall and Eric L. Tollefson, P.Eng. Department of Chemical and Petroleum Engineering The University of Calgary June 20, 1981 - Research Associate -'"Professor BFG103Q7 iw iz x rz Table of Content S umma ry 1 Int roduc tion - 1 ! Stroke Particulates 1I - 1 Optical Methods 11-2 Gravimetric Methods - Il"3 Relative Rates of Sncke and HC1 Release from PVC ll-l) Toxic Gases Adsorbed on Soot ll~5 Smoke Detectors 1 1 1-1 1 1 1-2 Decomposition of PVC Decomposition of ABS IV Toxicology IV--1 PVC Toxicology 1V--2 ABS Toxicology JV-- 3 Real Fires Involving PVC 1V-1 Real Fires Involving ABS and Plastics Known to Produce Hydrogen Cyanide V-l Rates of Smoke and Acid Gas Release in Rooms V-2 Behaviour of PVC E ABS Pipes in Fire Tests Conclusions Acknowledgements References Appendi cS Page 1 7 8 13 U 15 16 17 21 21 2k 27 30 31 32 33 37 38 331 1*8 21127002 BFG11308 Summa ry i The Increasing use of plastics to replace traditional materials in, for example, furniture, draperies, carpeting, electrical and thermal insulation as well as piping in plumbing systems has given rise to much concern in the event of a fire. Many pl'astics are very flammable, burn rapidly and produce intense heat, heavy smoke and toxic gases which together are very dangerous to life. The recent MGM Grand Hotel fire is an example of this type of firep most of the victims dying from smoke inhalation even though far-removed from the site of the fire. This report is the result of a literature search on the combustion hazards associated with the use of two plastics widely used in plumbing systems, namely, polyvinyl chloride (PVC) and aery 1 onitri1e-butadiene-styrene (ABS) . These two plastics, along with polystyrene, comprise about one third of the total plastics sold in the United States. Of the common plastics these three are most prone to produce extremely dense smoke on combustion. ABS is very flam mable and like other nitrogen-containing polymers releases hydrogen cyanide as a combustion product. PVC is much less flammable but will burn if combustion is supported by other fuels. Large amounts of hydrogen chloride gas (HCl) are released by either heating or burning PVC. Heavy smoke production during a fire is dangerous becadse: (a) it reduces visibility thus impeding escape, (b) particulates cause difficulties in breathing if inhaled, and (c) toxic or irritant gases not only interfere with vision and breathing but also damage lungs, cause confusion, unconsciousness and death. ABS, PVC, and wood produce comparable amounts of smoke when smouldering but, under flaming conditions, within one to two minutes after ignition the smoke from these two plastics is about 28-kO times as dense as from wood. Ventilation does not decrease the smoke density as it does in the case of wood. In the case of a rapid electrical overload in PVC-1 nsu1ated wire, smoke which is usually the first indication of a fire, is only noticeable after significant quantities of HCl have been released. Ionization detectors have been found to be quite insensitive to PVC smoke and have been reported to give an alarm after hazardous levels of HCl had already been exceeded. The rapid production of very dense stroke in PVC fires, the fact that 70% of the hydrogen chloride vapor which can be produced by a sample occurs in less than one minute at 200-3C0oC, and the relative insensitivity of ionization C O O LZXTZ > fwiwi awu*"w-w M w JMJi t BFG10309 oieoioja Aliiilfi qjoqs Aj3a e jaqqe em apa Sun i 6uisnea uidd 0001 MMM 3J0ui jo uidd Q01 i *l3A3l qe aiqejaioqui si spijoiqo usSoupAq pue sanjeA uoi qejquaauoa [eqqai aqq uo paseq apixououi uoqjea se aixoq se sauna aAiq-Aquaa qnoqe aje aptxoip uaSojqiu pue apiueAo uaBojpAq -A[[e;e^ jo Aisnoijas jaqqia saiqtaedea Jiaqq aanpaj oq uaSAxo qo asn jo qjodsuejq `ax.eqdn joq uiaqsAs s,Apoq aqq qqiM ajaqjaquj ue apijo^qa uaSojpAq pue apixoip uaSojqiu `apiueAa uaBojpAq `apixououi uoqjeo suoiqejado 6ui qq6i q-aj i q ui jo 3deasa Bujjnp ajaqdsouqe aqq qo quaquoa 13H aqq 6uianpaj u; atqen^eA aq p[non saajnos jaqqo jo suiaqsAs jai^uijds uiojq jaqe.n qeqq satidun siqq *dejq jaqew e qnoqqiM aAiAjns uea Aaqq qeqq sqanpojd sisAiojAd j/\j qo qunoue aqq saunq saij aAiAjns oq ajqe aje sieunue aqq `sqsaq Aqiaixoq adorns ui ioh dejq oq pasn si jaqe* uaq;q uoiqsnquioa $gy uiOJq queaixoq jofeui aqi Si N3H leqi suop uaaq aAeq qeqq siuauiijadxa' waq aqq uiojq quapiAa uiaas os[e p(noM qj *13H s! saseS sisAiojAd 3/\j ui queaixoq jofeui aqq qeqq aqeaiput saipnjs ieai6oioqqedoqsiq pue ieai6o[OisAqj sieuiiue qsaq aqq spje.-voq saiduies aqq jsao UMejp si jje uaqM Aj iea i qeuiuejp sasij osie o/\d } Aqiaixoq aAiqeiaj aqj. 'poo ueqq ajxoq ajou qanui ajaM o/\j pue say pauiuijajap sbm aip oq s[eujiue aqq jo *oS pasnea qaiqM a^duies jo qunouie aqq uaqM qnq pooM ueqq aixoq ssa^ Aiquajedde aje sqanpojd sisAiojAd Say pue 3Ad pssn si qqeap jo auiiq,, uaqM a(duiexa joj -suoiqipuoa [equauijadxa oq aAiqjsuas Aj3a aq uea sqpsaj qeqq quajedde si qi pue paipnqs AiaAjsuaqxa uaaq aAeq sieijaqeui quajaj^ip ujoj^ paonpojd sa^aus saiqiaixoq aqj_ uoiqsnquiaa sqi paanpojd ^0N Pue N3H squncuie aqq qaa^qe AeuJ S9V aqq qo uo 1 q 1 sodiroa aqq 6uiAueA os sgtf qo qjed a[ijquO[Ajae aqq uiojq paanpojd aje saseS asaqj. * (^0N) 3P!*1P uaSojqiu qo qunoiue ja^e'-us e se [[a.M se N3H 6u 1 aq paujjoq qanpojd aixoq jofeui aqq `3Q00C asodu-oaap oq sui6aq say sieijaqe'j aiueSjo jaqqo qsou op se apjxoua-j uoqjea saanpojd 3,\d i uoiqsnqu.-oa `uaoAxo qo aauasajd aqq Ui pue 30OOS 3V `Jnaao oq uojqaeaj siqq joq Ajessaaau qcu si usSAxo qo aauasajd aqq '3Q0LZ lncqe qe aAiqeqiquenb 6uiu-oaaq pus 30Qg( pue qH uaa.Mqaq 6uiuui6aq uo 1 q 1 soduoaap [eujjaqq e si 3AJ mojq ^plJO^qo uaSojpAq qo ssoq auiiq adeasa aiqeuoseaj e aj iq e qans oq pasodxa uosjad e Auap oq auiqj.'oa [|e `a^ous 3Ad oq sjoqaaqap III exposure time. The effects of HC1 will not he evident immediately and carbon monoxide can cause death as long as 3~15 months after a person apparently has recovered. The situation is further complicated by the fact that inhalation of a complex mixture of toxic cases and smoke in an actual fire situation will result in additive or possibly synergistic (greater than additive) responses that are not well studied or understood. Information is available concerning the apparent increase in the toxicity of carbon monoxide when blood alcohol is greater than 0.1S (volume). In actual fire victims the carbon monoxide levels are elevated but very often are not sufficiently high to have caused death indicating that other toxic substances are important factors. In actual PVC fires, even though there was little smoke in their early stages and the fires were small, the firefighters involved exhibited symptoms due to Inhalation of HC1 and often required medical treatment. The role of cyanide poisoning in fires seems to be less well understood but is nevertheless considered to be serious in terms of its highly toxic effects. Calculations based on experimental data show that if 100 pounds of ABS pipe was burned in a 10,000 ft^ apartment it would result in a maximum concen tration of HCN approximately 20 times the lethal amount. The same calculations for 100 pounds of PVC pyrolyzed in the same apartment show that a concentration of HC1 as high as 57,385 ppm could be reached, about 57 times the concentration that will cause lung edema on very short exposure.- The use of plastic pipe in high-rise buildings is of concern due to the fact that it will melt or burn through thus allowing toxic gases, smoke and fire to penetrate fire partitions. It is evident from this literature search that the use of plastic pipe is not a safe practice in many situations and that metal pipe would be less hazardous. Choice of materials, and development of building practices that will contain any fire which breaks out, will minimize smoke production, smoke toxicity and smoke exposure are clearly desirable. BFG10311 S ootittT Z I. Introduction Smoke from smouldering or burning materials is at least as hazardous to life and property as uncontrolled fire (l). Indeed, it Is widely recognized that more than half the deaths attributed to fire are caused by .smoke inhalation rather than by heat or burns (l). With the increasing use of plastics m bui1dings , includIng furnishings, upholstery, piping systems and electrical and thermal insulation, there is increasing concern about the fire- and smoke-producing hazards of these materials. Possible increases in danger to fire fighters and rescue workers resulting from the substitution of plastics for more traditional materials is also a concern (2,3,M> Such concerns have been reinforced by a number of recent fires. A good example is the November 1S80, fire at the HGH Grand Hotel in Las Vegas. In this fire 679 people were injured and 8k people lost their lives. Sixtyfour of the victims were found on the upper floors far-removed from the site of the actual fire. The source of ignition was reported to be electrical and large amounts of foam plastic padding and other plastics were included in the fuels (5)- All deaths were attributed to smoke inhalation (6) but the.lethal component or components of the smoke were not specifically identified. The fact that the concentration of carboxyhemogiobin in most of the victims was not high enough to have caused death (7) indicates that other toxic gases or smoke particles must also have been involved. This opens the question as to whether or not the burning of plastics increases the lethal potential of fires. Although a detailed consideration of the potential fire hazards of the numerous plastics now in widespread use in beyond the scope of this report, much can be learned about the dangers which can exist from considering two very common plastics, PVC (polyvinylchloride) and ABS (acrylonitrilebutadiene-styrene). The largest single use of both PVC and A3S is for piping systems. PVC is also widely used for electrical conduit, electrical Insulation and telephone ducts (8). Lesser amounts are used for wallpaper, window frames, and upholstery. A thorough search of Chemical Abstracts for information on the combustion and pyrolysis (thermal degradation) of ABS and PVC was-carried out. The results are presented in this report. Wl* JPktll'lHIJL l| I"gfE. I BFG10312 s m L z irz The Hazards of The Py ro I ys i s/Com,Sus 11 on Process The fire hazard of a material is described by: 1. its combustibility (how readily and how rapidly it burns or releases heat) 2. the amount of smoke it produces, and 3- the toxic gases released during heating and combustion (9). These factors are in turn related to the chemical composition and weight of the material which will determine the total heat, smoke and toxic cases that can be released under specified conditions. The exposed surface area, surface characteristics, thermal exposure and the location of the material within the fire system will also affect the fire hazard associated with it. The combust ibility can be described by ease of ignition, rate of heat release, and total heat release (9). The ease of ignition and the rate of heat release from some plastics is very high compared with such traditional materials as wood (9,10,11) but other plastics are less combustible than wood (10). Smoke evolution rs a concern relevant to a significant percentage erf the plastics used. In the United States of the 1*4,791 ,000 metric tons of'plastic sold in 1977 were the three types most prone to produce smoke: PVC, ABS and polystyrene (12). Smoke is hazardous because: (13) 1. if dense it impedes escape, rescue or fire-fighting by obscuring vision 2. it is irritating and can be destructive to tissues of the eyes, nose, throat and lungs 3. it may contain lethal or incapacitating concentrations of toxic or suffocating gases *i. certain of its constituents particularly acid gases, may damage property. The particulate and gaseous fractions of smoke can be considered separately, and "smoke" in this report refers to the particulates only. Combustion Properties of ABS and PVC 1. Smoke Very dense smoke is rapidly produced from, both of these plastics and unlike with wood becomes denser under flaming conditions. Indeed for PVC the time until *n observer would find It difficult to see an exit sign through 10 ft. of sroke In a 12.5 *20x5 ft. room was extimated to be 2.1 21127007 .1 tfJPWJJWWP.1 HEY** BFG10313 -3- minutes under smouldering conditions, and 0.5 minutes under flaming conditions (\k). The corresponding times for A3S (cycolac) have been reported as 2.98 minutes (smouldering) and 0-57 minutes (flaming) (15). This short time would make escape difficult or impossible in many situations. The cor res pond i ng values for various woods ranged from 8-15 minutes (smouldering) and A -10 minutes (flaming) (lM- 2. Toxi c Gases In addition to carbon monoxide which is often present when any organic substance burns, the major products of concern from ABS and other nitrogencontaining substances such as wool, nylon, orlon, and acrilan are hydrogen cyanide and nitrogen dioxide (16,17). '"hen PVC is subjected to thermal degradation, under smouldering or combustion conditions, hydrogen chloride gas is produced rapidly and quantitatively sometimes before smoke is evident (18). From one gram of PVC, 0.5S3 g of hydrogen chloride is produced. This facile production of large quantities of acid gas in fire situations is perhaps the most striking difference between PVC and more traditional materials. Some benzene is formed at the same time (19,20,21). When acrylonitrile is pyrolyzed or burned, hydrogen cyanide and nitrogen dioxide are .formed. If hydrogen cyanide is inhaled, it causes a reflex stimulation of breathing which in turn will lead to a greater concentration of the gas entering the body. It also inactivates certain respiratory enzymes thus preventing the use of oxygen by tissues (22). Concentrations as low as 100 ppm can be lethal. Safe concentrations of nitrogen dioxide are below 5 ppm. Nitrogen dioxide is irritating to mucous membranes and when inhaled will cause damage to tissues in the upper respiratory tract. Inhalation of very low concentrations may cause l'ttle or no damage to the upper respiratory tract but may result in death hours later due to pulmonary edema (abnormal accumulation of serous fluid in the lungs) (23). If inhaled, hydrogen chloride oas dissolves in water to become hydrochloric acid and will severely damage the upper respiratory tract leading first to breathing difficulties, then to pulmonary edema and death. Hydrogen chloride also acidifies the blood and destroys hemoglobin thereby interfering with transport of oxygen (2k). However, when it is first inhaled it will tend to cause the larynx to close (laryngospasm) thus interfering with breathing but 21127003 .i mj .him,. .liar *. BFG10314 H but protecting the lungs. Vhi1e PVC decomposes, benzene is also formed and it will depress the larynoospasm allowing both solids and aerosols (including acids) to enter the lungs (22). The main action of carbon monoxide CCOl after (t Is inhaled, even at lewconcent rat ions, is to Bind reversibly with her-.oglo5rn (H5l to form carbeocyhemoglobin (CCHb). Thus the % CCHb of a person's blood rs related to the concentration of carbon monoxide in the atmosphere and the time of exposure. Carbon monoxide displaces oxygen in the blood and leads to anoxia and death, if the reaction is not reversed. In general, most persons will not show toxic symptoms below 20* CCHb. Death occurs at approximately 6Q.-7Q* COHb.(25)_. l/hen carbon monoxide is inhaled there is also a loss in nerve conduction velocity and a drop in blood pressure which depletes the Blood flow to the brain resulting in confusion and failure to respond to stimuli .(22), However, even after exposure to a potentially lethal or comatose leyei of carbon monoxide, 80* of individuals will recover immediately or within a few- days. Then perhaps 3-15 months later, 15~2Qt of those people suddenly develop dementia and the "shakes" dying shortly afterwards (22), The statement that carbon monoxide is the primary killer in fires has been disputed (.22) But about 53-* of the fire deaths doe to smoke inhalation occur within 12 hours, usually at the fire scene with fairly slight respiratorytract damage. The remainder of the deaths attributed to smoke inhalatron occur after the victims are taken to hospital and treated. In these cases the respiratory tract damage continues to develop after they are removed from the fire and would Implicate gases other than carbon monoxide and/or the inhalation of particulates as the cause of death (22), Carbon dioxide is also present In smoke, but not normally- in sufficiently- high concent rati ons to cause any toxic symptoms. However, inhalation of carbon dioxide will stimulate respiration which in turn wfli increase Inhalation of any toxic gases present. Organic aldehydes and acids which are both irritat ing and damaging to mucous membranes are also known to be produced from some materials (26}. Human responses to oxygen depletion and various concentrations of gases N relevant to this report are found in Appendix 1 (25). 127009 Mum1 BFG10315 3. F1 ammab i 1 ? t y In general polymers with an aliphatic backbone such as polyethylene and polypropylene are very flammable, but their tendency to generate smoke is minimal. The addition of flame retardants, especially halogen - containing compounds, whi1e reducing the tendency to burn, Increases the evolution of smoke (10) and introduces strong mineral acid cases Into combustion products. Halogen - containing polymers such as PVC are usually relatively non flammable, with high oxygen Indices for PVC) but display high smoke generation when combustion is supported by the presence of other fuels which may include plasticizers. The oxygen index is the percentage of oxygen in an oxygen-nitrogen mixture which will support sustained combustion. The lower the oxygen index, the more flammable is the material. Representative oxygen indices are given in Appendix 2. Polymers with aromatic rings such as polystyrene and ABS not only generate dense smoke (10) but tend to be very flammable. In fact, in several studies, ABS was one of the most flammable polymers tested (11,27). The burning characteristics and "combustibility" of a sample can be described In terms of rate of heat release as wel1 as ease of ignition, and total heat release. The sudden release of heat indicates a rapid spread of flame across the exposed surface. ABS exhibits a very rapid rate of heat release (see Figure 1),, is very easily ignited, and burns quickly compared to wood or PVC (9). Factors Affecting Survival An Important factor in a fire situation is a person's escape time; that Is the time interval Between the person being alerted to the existence of the fire and the point at which escape is no longer possible. The rate of production and distribution of smoke and toxic gases is therefore very Important C2l. It has Been estimated that a sleeping person needs 12-15 minutes to awaken, react to danger and take appropriate action (2), With the very rapid production of smoke Cand toxic gases)_ from /n B S and ?V'C the person wj 11 often not have that nuch time. Judgment may also be impaired on awakening as a result of breathing asphixiants such as carbon monoxide while still a s1ce p. 2X1270X0 'IP HIUPP BFG10316 Time (Minutes) Figure 1. Rate of Heat Release (3). Approaches to the Problem The question of the relative hazards and toxicities of smokes from different sources can be approached in two ways. The most obvious is to study real fires but there are serious difficulties in identifying sources of individual toxicants in such complex situations. One important study of residential fires in Boston employed a sampling system built into the firefighters' turnout coats to determine the concentrations of various gases in the smoke (26). Some autopsy data are also available on fire victims relating to toxic gases. For instance, the COHb levels and cyanide levels in blood are an indication of the exposure to carbon monoxide and hydrogen cyanide at the fire scene (28). Symptoms described by survivors are also documented the case of PVC fires and provide an indication of what gases may have been present (2,l|). T T n /? T T Z BFG103H / The importance of the blood alcohol level of fire victims should not be overlooked but often is either not determined or reported. If it is greater than 0.1% volume, there is probably a strong synergistic response with carbon monoxide that increas'es the apparent toxicity of carbon monoxide (22) and possibly also hydrogen cyanide to the brain. Indeed in one study of fire deaths in .Maryland, of the victims had a blood alcohol content of greater than 0.1% (23). The second approach is to carry out laboratory scale experiments to determine the amount of smoke produced, the concentrations of the toxic gases, and the toxicity to animals of smokes produced from specific materials. There are many papers in the literature on these subjects but findings are often confusing or apparently contradictory and may be difficult to relate to real fire situations. The wide range of factors involved in a real fire is difficult to reproduce in the laboratory. There are two distinct processes in a real I fire: (l) pyrolysis (the thermal deconposition/volati1ization of the sample without any flame being present at the immediate pyrolysis site) and (2) combustion (consumption of the sample by burning with a continuous flame). In the laboratory, pyrolysis and combustion are investigated separately. Hazards associated with the particulate and toxic gas fractions of smoke can be considered separately. Data are not generally available on the biological effects due to the inhalation of smoke particulates but a concentration of about 1,000 mg/m^ air will cause immediate respiratory difficulty; much lower concentrations of 20-50 mg./m3 will cause light obscuration and can thereby interfere with escape (26). 11. Smoke Particulates Tests for smoke-producing characteristics of materials employ one of two measurements techniques: gravimetric or optical. That is, some tests are based on determining the weight of smoke particles deposited on a filter under specified conditions; other tests measure the fraction of L light absorbed or obstructed by the smoke. Both methods measure the airborne particulate fraction of the smoke and not the gaseous fraction. Results are apparatus dependent and are often reported for commrcia 1 products for which the composition is not given. Cross comparisons are 127012 N imi'u l riuj. HEW*!" BFG10318 difficult but comparisons of materials for each method are valid. The best known gravimetric methods are the ASTH El62 (30) and the Arapahoe smoke tests (31). The most widely used optical methods are the KBS smoke test (32) and the ASTM D28^3 test (also-called Rohm and Haas XP2) (33) which measure the density of smoke accumulated in an enclosure. In some experiments a modified KBS chamber is used which allows different levels of ventilation. The ASTM ESk test ('k) ana the OSU release rate test (35) measure, by optical means, the density of smoke flowing past a specific location. I 1 - I Results Using Optical Methods (V NSS Test Specific optical densities (see Appendix 3) which can be measured with this apparatus are independent of exposed area of the specimen, chamber volume, and length of light path, but depend on specimen thickness, chemical and physical properties and exposure conditions. The exposed surface area of- the sample is 6.56 square inches; the chamber volume is 18 cubic feet (1 k). Maximum specifi-c optical densities are shown in Table 1 for a number of materials. , Table 1. Maximum Specific Optical Densities (Dm) for Selected Materials (lk) Thickness Dm Smoldering Flaming Polyethylene (UCC - DXH - 100) PVC (UCC - QYTQ) t 1 526 315 780 ABS (Cycolac) Douglas fir 250 mils lI O 00 780 380 156 Note that for woods Dm is lower under flaming conditions compared to smoldering conditions, but exactly the opposite is true for PVC. The maximum measurable optical density was reached under both smoldering and flaming conditions for ABS. Times at which maximum specific optical density- occurred (T 1 and at which the specific optical densities (Dg) would be equal to 16 in a 12.5 x 20 x 8 ft. room,the value at which an observer would find it difficult to see an exit sign through ten feet of smoke are shown in Table 2 (lk). Under smouldering conditions the times to obscuration for wood and PVC are very similar but under flaming conditions they are much shorter H* h* *oN to sm BFG10319 -9- for PVC. It is interesting to note that doubling the thickness of the PVC sample did not change the time to obscuration even though it roughly doubled the time to maximum specific optical density. Note also that wood and polyethylene show significantly longer times to reach the specific optical density (Ds = 16) than do the PVC and ASS samples tested, an important matter in terms of escape. Table 2. Times (in minutes) to Obscuration and Maximum Specific Optical Densities (14) Smoldering Tm T for D s =16 Flaming Tm T for D s =16 Douglas Fir Polyethylene PVC rigid - filled PVC rigid unfilled - 1/8" thick PVC rigid unfilled - 1/4" thick ABS *ABS 20 17 30 14 33 11 2.1 5.5 1.6 2.1 2.1 3-0 19 9 11 5 10 6.5 4.6 4.0 0.5 0.5 ^ 0.6 0.6 * Data front reference 35 Table 3- Effect of Ventilation on Maximum Smoke Densities from Some Smoldering and Burning Haterials (14) M.ater 1 al Smoke Densities at Specific Ventilat ion Rates(Air Changes/hr) 03 6 12 20 Smoldering Rigid PVC, filled Douglas Fir 4SD 380 335 300 235 225 160 120 150 90 Fleering Rigid PVC, fi1 led Douglas Fir 530 155 535 65 535 70 **75 375 60 25 > mvmii,i ku. ' *ii ii ymgmmmtm q WMMW- BFG10320 The effect of ventilation rate on the development of maximum smoke density is illustrated in Table' 3- It should be noted from the data that ventilation decreased the maximum smoke densities under smouldering conditions but that there was little difference observed under flaming conditions for PVC (lk). Similar observations were made by other workers (37,13)* A5S also produced very dense smoke rapidly (27,3^) under flaming conditions. At 1.5 minutes after ignition,- 176, already a much denser smoke than that which would cause obscuration of vision. Specific optical densities are dimensionless but the time required to reach a given smoke density must depend on material surface area and chamber or room temperature. It is therefore, of Interest to note that the ratio of sample surface area to chamber volume in the NBS test corresponds to 13 feet of sealed 3_inch pipe in a k,000 cu.ft. room. Therefore, burning even a short length of ABS or PVC pipe in a room would rapidly produce dense, obscuring smoke. (2) Asm D2843 (Form and Haas XP2) This test is limited to flaming conditions. The sample surface-to-chamber volume ratios are similar to the NBS test. It has been.reported that the amount of smoke produced is almost proportional to the weight of material consumed (38). The amount of smoke produced is very dense for PVC and ABS (in agreement with NBS results) (15). The times to obscuration (D = 16) for PVC sheets Q in, x 1 in. * 0.0.158 in. or 1 in, x 1 In. x 0,Qk2k in.), were 0.03 and 0.08 minutes shorter than found In NBS tests 05), Both PVC (QYTQ.) and ABS (cycolac) reached the maximum specific optical densities measurable with this apparatus. Various polyethylenes had = 10 - 21k. For PVC (QCA - 2k6o) of 15, 20 and kO mils thick, the respective maximum specific optical densities were reported to be 99, 82, and 319 {15)- (3) OSU Release Hate Test ' This test measures the density of the smoke flowing out of an exhaust Stack under flaming conditions. The sample has 100 sq. in. of exposed surface^ area (9). Results show that the rate of smoke release and total amount of smoke released for samples of PVC and ABC are many times greater than for cak (12) (Table k). C In f "r I ...................... ....................... ......... - - bfGV0321 Material Table k. OSU Release Rate Tests Studies (17) Orientation Applied Heat Flux w/,cm 2 Max. Smoke Release Rate ,. 3 umts/min-m Cak, 1" * PVC rigid pipe PVC flexible sheet, ABS sheet, 125 mi 1 V 1.0 2.0 2.5 V 1.0 2.6 V 1.0 2.0 H0 1.0 2.5 0.9 5-5 6.9 4.6 m 228 336 274 366 366 Total Smoke Re 1 ease Units/m^ 3 min 10 min 0.9 1.8 9-1 0-9 31.4 228 61.0 2-3 4.6 11.4 22.8 ^57 61.0 731 2.3 137 155 ^57 457 -1548 V = Vertical sample H = Horizontal sample 1 smoke unit in 1 air will produce an optical density of 1.0 for a light path of 1 metre. The very large variation in rate of smoke release per ft 2 of exposed surface for different materials is shown in Figure 21 While the maximum 2 rate of smoke release for red. oak is less than 200 "particles"/min ft the corresponding value for ABS at the same heat flux is over 7900 and for FVC is about 5,500 (9). The use of piloted ignition and excess oxygen assures that essentially all decompositions vapors are burned and would minimize the smoke release In the case of the oak. 21127016 'MPJ.ILMS' 9* BFG10322 2 Figure 2. Rate of Smoke -Release at 2.6 Btu/sec-ft (9) (4) Other Experiments Other workers using different types of apparatus also observed very dense smokes from ABS and PVC compared to wood (39,^0,^!,^2,^3). Hass optical density (MOD) (Appendix 3) relates smoke generation to the weight of sample rather than the area of sample exposed. Results are given in Table 5 (^0). Table 5. Hass Optical Densities of Materials Under Flaming Contitions (frO). Sample 2 HOD cm gm ABS PVC with 28$ plasticizer PVC PE Hardwood (3 thicknes ses) 6900 5800 3^00 2900 300-800 The results from the different experiments and the various calculations all show that much denser smoke is produced under flaming conditions from ASS and PVC than from wood. BFG10323 JJL\\f y r r y 13 11-2 Gravimetric Methods (1) ASTM El62 In this test flaming conditions are used and the particles are collected on filter paper positioned at the top of the stack. Some results for this test are given in Table 6. It will be noted that, as with optical methods, a large difference was observed between wood and PVC (12). Material Red Oak Exterior fir plywood PVC PVC, Fire retardant Table 6. ASTM E1 >2 Results (12) (Flaming Conditions) S i ze 12x18" II II II Thickness (mils) 750 250 U7 147 Smoke Deposit (mg) 0.3 0.3 28.9 10.5 This shows that the optically dense smoke released by these polymers is at least partially due to release of a larger mass of smoke particles. (2) ^rapahoe Smoke Chamber The sample is exposed to the burner flame for 30 seconds with air flowing through a filter at the top of the stack. The burner is then extinguished and air flow continued for a further 30 seconds to give a total collection time of 60 seconds (k1*). Results are reported as percent smoke based on the initial weight of the sample or on the weight loss of the sample (Table j) (12). Because some cellulosic materials (wood) smoulder for several minutes at the end of the test the calculation based on weight loss for these materials would not be valid and they are best compared to plastics using percent smoke based on initial weight 4) - The results confirm that much more smoke is produced from PVC and ABS than from wood. In conclusion it can be stated that both PVC and'ABS produce copious amounts of dense smoke under both smouldering and flaming conditions. Furthermore, in the preceding discussion these plastics were compared with wood which is a traditional building material but is not used for piping or copduit systems. Traditional metal piping does not produce any smoke under no/mal fire conditions. 1127018 Wli "i I VniHLI uu ijwj.mu'm wamnssimMHie BFG10324 Material Table 7- Arapahoe Smoke Test Results (12) (Flaming Conditions) & Smoke based on Initial Weight Weight Loss Cellulose Fibreboard, coreboard Hardboard PVC flooring PVC rigid PVC, flexible, fire retardant ABS, fire retardant 0.570.06 0.21 1.33 2.36 k.02 0.75 0.08 6.23 10.52 12.7^ 20.5*1 1 1-3 Relative Rates of Smoke and HCl Release From PVC In the case of a rapid electrical overload of PVC insulated wire, the evolution of HC1 preceding the detection of visible smoke is as shown in Figure 3 (18). Figure 3. Comparison of Rates of Formation of HCl and Smoke for Different Heating Rates of PVC Insulated Wire. (A=frQ"C/mIn and B=15BC/mIn) When pure PVC film was used the smoke and HCl evolved at the same time regardless of the heating rate used. To complicate matters further, it was found that a commercial sheet of PVC evolved smoke about half a minute before HCl was detected at a heating rate of 20C/min, and, that at decreased heating rates smoke was detected before HCl. Under some circumstances, particularly in the case of a rapid electrical overload, HCl could be a significant hazard before a person became alerted to the fire by the presence of smoke. Firefighters should be aware of the fact that dangerous acid gases BFG10325 - 15 can be present in areas involved with electrical fires even though there cay be little visible smoke. Il-lj Toxic Gases Adsorbed on Soot (2) BCl Irritant gases are found to be adsorbed on soot. HC1 adsorbed on soot would gain access to the lungs where, on combination with water, hydrochloric acid would fc-e formed causing a violent 1 nfl a.~~.atory response, resulting in destruction of lung tissue (2,^5)- Hydrogen chloride from PVC combustion was found in the gas phase, and both loosely and tightly bound to the soot. About 2% of the total hydrogen chloride was found adsorbed on soot. The hydrogen chloride that was loosely bound amounted to 19 mg/g soot on soot particle diameters of 300-1100 A (0.03-0.11 microns). Such particles would agglomerate rapidly to form larger particle clusters. For rigid PVC the size distribution of particles was found to be 0.1-Q.2 microns and the mean particle size was smaller for flaming than for smouldering conditions (1(2). The particle sizes ranging from 0.1 to 2.5 microns diameter (corresponding to soot aged from seconds to one hour) would be retained in the alveolar sacs to an extent of from 20-1(03; of those inhaled (1(6). Assuming there was 1.57 grams of soot/m^ and no HCJ in the gas phase and that the breathing rate of a person is 18 1/min, after one hour of exposure 0-7 grams of soot bearing 13 mg of loosely bound HC1 would be retained in the lower lungs (1(6). If there was a gas phase hydrogen chloride concentration of 150 mg/m^ with no soot present, at a breathing rate of 18 1/min and an exposure of one hour about 108 mg of HC1 would be retained in a person's body (1(6). (It has been estimated that 62% of the hydrogen chloride gas inhaled is retained in the body (1(7)I Therefore, it was concluded (1(7) that gas phase exposure was about eight times as severe as exposure to soot; however the HC1 on the soot would cause damage to the lungs. (2) polyarcr.atic hydrocarbons 21127020 Polyaromatic hydrocarbons have been detected in small amounts from the combustion of various plastics, including PVC (^8,^9,50,51). The production of these compounds was relatively lower under flaming than under non-flaning conditions (1(9) (Table 8) and included the potent carcinogens benzo [a] pyrene and 7,12-dimethyIbenz [a] anthracene. As the humidity of the ventilating air increased the production of polyaromatic hydrocarbons* * inn.p niiuim BFG10326 Table 8. Amount of Polyaromatic Hydrocarbons Found in PVC Pyrolysis Products under Simulated Fire Nonflaming and Flaming Conditions (k9). Amount pg/g of PVC Nonf1aming Flaming Fl uorene Phenanthrene 1-methyl phenanthrene 9-methylanthracene 9,10-dimethylanth racene Fluoranthene Pyrene 1,2-benzof1uorene 2,3-benzofluorene Chrysene, 1,2,-benzoanthracene triphenylene 7,12-dimethylbenzo [a] anthracene Benzo [a] pyrene, benzo [e] pyrene Perylene 3-6 29.1 19.1 15.5 15-6 10.5 11.6 35.3 32.8 16.0 3.5 5-3 7-3 3-6 3-5 3-6 k. 7 6.1 k.5 5.8 10.5 - 2.k 2.k decreased (k9) . The evolution of benzo la] pyrene generally increased with increasing temperature but.at high temperatures it decreased drastically when a high air supply rate was used (^8). Whilst these compounds are not considered to have acute toxicity they could Be of concern to firefighters who are repeatedly exposed to them. It it difficult to assess the dangers due to the amounts of polyaromatic hydrocarbons detected in these experiments. 11-5 Smoke Detectors With respect' to detectors it is interesting to note three papers in the literature (52,53,5*0- Tests showed that ionization detectors are particularly insensitive to the degradation products from PVC, polyurethane foam, "and polyethylene (52). All ionization detectors tested showed a decreasing sensitivity (based on increasing smoke obscuration at alarm) in the order "punk" (a standard calibration source) > polyurethane > polyethylene > PVC (53). PVC cable overloads of i 3-5 times the nominal current were detected by Ionization detectors but the alarm was given at 3*10 times the maximum allowable hydrogen chloride concentration (5k). In the case of PVC, evolution of toxic gas is a major threat before there is sufficient smoke released to activate whatever detectors that may be in m. iLLii! ! **""**' 11 BFG10327 T z n /.? T 7 z ' / place (55). 111-1 Decomposition of PVC 1. Temperature at which decomposition takes place Dehydrochlorinat ion* of fyC is essentially- a. thermal decomposition, Flaming and combustion conditions are not required (55). Dehydrochlorination is reported, to begin be tween' 150-190C (57,58,59,60) becoming quantitative at about 270C (18,61). It was also observed that pyrolysis of PVC at 600C under helium atmosphere resulted in quantitative recovery of HCl and formation of a chlorine-free ash (62). Woolley (63) showed that the rate and extent of dehydrochlorination was essentially independent of the amount of oxygen present In the atmosphere. Removing HCl from the reaction zone by evacuation decreases the rate of dehydrochlorination PVC but does not change the rate of benzene formation (6*1) Until recently the only product reported to be produced during the decomposition of PVC below 220*C was hydrogen chloride (65,66). Voorhees et a 1 (21) established the fact that benzene and hydrogen chloride were formed simultaneously during the thermal decomposition of PVC. At 275C a very rapid weight loss occurred accounting for 60 of the initial weight. The product was about 55% s. hydrogen chloride and 5? benzene (20). A second weight loss occurred at higher temperatures (66). In another experiment in which 1.0 g of PVC was burned 583 mg HCl, .729 mg. of carbon dioxide, kk2 mg. of carbon monoxide, and 36 mg. benzene were formed (66). Other workers found that burning 3 9 PVC in a closed container (0.12 1 air) at ^00-800 gave a gaseous mixture containing lA-3% HCl, 10.61 CO, 10.6 and *.0t organic materials (67). Burning PVC in air gave 0.^96 g HCl, 0.001 g C0C?2 (disputed by others), 0.229 g CO, and 0.*i33 g C0^ per gram of PVC (67). Although a very small amount of phosgene had been reported (67) subsequent workers have been unable to detect either phosgene or chlorine (68,69). A mixture of smaller amounts of saturated and unsaturated hydrocarbons is also formed, in addition to benzene (70,71,60,20,72,73,7*1,75,69) At temperatures greater than 300* eighty compounds were formed in relatively small amounts from the combustion of PVC (76). Various chlorobenzenes (77.78) and forma 1dehyde (75) have also been detected. At 200C no hydrocarbons were found (77). Carbon monoxide appeared at about 500*C (79,80) and at higher temperatures was the major product (8l). The higher the pyrolysis temperature the greater the variety of gases produced (82):. B*!WWK#L'ws ifliuUWW.i MlI IJUJI.iME. BFG10328 K 2. Rate at which toxic cases are produced Several factors affect the rate of dehvdrochlorination. The rate increases with increasing temperature (59) arid with Increasing air flow (83, 72.56). From the rate constants of decomposition the time required for 20, 40 and 60% dehydroch1 orination of PVC in and air between 200-300C was calculated (72). These results appear in Table g'. Table 3.- Calculated Time (min) for Dehydrochlorination of PVC (72) Temperature C 200 220 2140 260 280 300 20S N2 Air 40% N2 Air 60% N2 Air 251.4 114.9 41.4 24.3 7-9 5.8 1.69 1.55 0.41 0.45 0.11 0.14 619-8 283.3 102.1 60.0 19.4 14.3 4.16 3-82 1.00 1.12 0.27 0.36 1237-8 565-8 204.0 119.8 38.7 28.6 _ 8.3I 7.62" 2.00 2.23 0.53 0.71 The experimental values were found to correlate well with the calculated values (72). It should be noted that time for 60% dehydrochlorination to occur is less than one minute at 300C. The rate of release of hydrogen chloride from 1.5 inch (3*8 cm) PVC conduit at different heat fluxes is shown in Figure 1* ,t9). Figure 1). Rate of Release of Hydrogen Chloride from 1.5" PVC Conduit (9). BFG10329 The relative rates of smoke and toxic gas release is a matter of some concern. In the case of a rapid electric cable overload the evolution of hydrogen chloride precedes the appearance of a measurable amount of smoke (18). At oSS'C heat and smoke release from PVC were very low but significant release of HC1 occurred (9). It is of interest to note here that tests with chlorinated PVC water pipe showed that copious amounts of hydrogen chloride were released at about 5S0C. Heat and sroke release were very small and increased when obvious burning occurred after nearly all the hydrogen chloride had been released (9). 3- Full Scale Fire Tests Full scale fire tests have been carried out in a special compartment (3 x 3 x 2.5 m)-corridor (12 m. long x 1.3 m wide x 2.5 m high) facility designed to represent a room attached to a corridor (9l). The fire load consisted of wood (120-2A0 Kg) in the form of a crib with thick PVC wall linings (rigid PVC sheet 0.63-115 Kgin the compartgent or corridor. Eyen _ at the corridor end of the test rig the temperature of the fire gases in all tests exceeded a bearable temperature within a few minutes after ignition. The assessment of the hazard arising from the presence of CO and HC1 is difficult because of the very high temperatures of the undiluted gases. In practical situations it is often necessary to consider the dilution of gases as would occur during the discharge of fire gases into the air enclosed within a building. The maximum concentrations of hydrogen chloride and carbon monoxide formed when the compartment and corridor gases are diluted with clean air to a bearable temperature (120C) are shown in Table 10 (81*). Table 1CT. Production of CO, HC1 and Smoke from Full Scale Fire Tests with PVC Wall Linings, After Dilution to a Bearable Tem.p erature (8*1) . Fire Load Wood (kq* PVC"(kg) Compartment CO (ppm) HC 1 (ppm) Corridor CO (ppm) HC1 (ppm) V 5 s i b i 1 i ty (m) 123 0.9 2,750 123 2.63 2 32 120 95.0 2.230 120 115-0- 526 2*t0 95.0 2,350 2*40 115.0* 6,730 *?VC used as corridor lining 3 79 3,230 360 3,920 1 ,800 M50 -- 850 -- 566 3,330 8 67 1,^60 9,690 1,890 20,500 ' 1.6 1.6 2.*t 1.2 2.k 2.0 fj K O Si WWJIPMfIW.U W'MWWU IWWIHiJW-g X IBVmikllJUl.MK W-WIMW BUtMWlWt BFG10330 Using a 127 g wood crib and 100 kg PVC v;a I 1 lining in the same apparatus the amounts of HC1 and CO were measured and compared to a control experiment in which the wood only was present (72). The results are shown in f i g u r e 5. -- WCt (.oerd t fvCj Figure 5- Toxic Gas Production from Vood and PVC Fires (Adapted from ref. 72) The rapid and facile release of HC1 when PVC is heated suggests that PVC should not be used where there is a possiblity of it becoming overheated. In service corridors, for example, if a fire should break out, very high concentrations of HC1 vapor are to be expected if PVC piping, conduit and wire insultation have been employed. Firefighters need to be aware of these hazards in fighting such fires. U. Effect of Composition of PVC on Decomposition Products The release rate of hydrogen chloride for different PVC materials is fairly uniform (9), however the temperature at which decomposition occurs and the amount of hydrogen chloride released can vary. Certain inorganic fillers ini PVC compounds can act as HCl adsorbers during pyrolysis and combustion. For example, CaCO^, a typical filler in PVC wire formulations was shown to adsorb a stoichiometric equivalent of HCl when burned (85). Rigid PVC film containing plasticizer and a small amount of stabilizer completed a major part o'f dehydro chlorination at less than 300C, while rigid PVC plate containing about 5% stabilizer evolved maximum concentrations of HCl at k00C (69). PVC artificial leather evolved about 1/10 the amount of HCl compared to PVC raw material (69,86). The amounts of aromatic hydrocarbons were found to be dependent on the presence of additivies (73). Heating PVC plasticized with phthalotes cave dibuty1phthalate (87). *** yyjwn BFG10331 C 'x w /y T T ,y ! I ! -2 Da compos i t i on of ABS V/hen ABS.is pyrolyzed In atmospheres of either nitrogen or air (36) a large weight loss begins to occur at about 300C. The decor,pcsition products which have been reported are hydrogen cyanide (HCN) , nitrogen dioxide (NO^) , methane (CH^) , hydrogen and ethylene (H^) (58,17,$9,71). (n 1974, Chaigneau (17) using a scr.ple of ABS {22% acrylonitrile, 19a" butadiene and 53* styrene) subjected to temperatures of 500-1200C under dynamic conditions in air found the HCN and l.'O^ production shown jn Table 11, Table 11. Decor.pos i t ion Products from A3S (17) Temp C 500 600 700 500 1000 HCN g/lOOg ABS N02 g/lDOg A3S 3.00 0.012 3-56 It.03 It.62 0.015 7.61 1100 4.59 0.01 1200 5.42 Much less of the nitrogen was found as N0j as compared to HCN. amount of HCN was formed at 10D0C under these conditions. The maximum Sumi and Tsuchiya (16) also using dynamic conditions and temperatures ranging from 400-800C found that the amount of HCN produced from ABS pipe increased as the temperature increased in agreement with Chaigneau. They also observed that the yield of HCN was highest in a 50? air - 50? nitrogen atmosphere. IV Toxicology- U'fien an animal or man is placed In contact with a chemical agent it can produce an acute toxic effect by acting as a primary- irritant upon the skin and/or mucous membranes or by Being absorbed into the Blood stream. Absorption of very- low concentrations may affect mental functions, THe most common syndromes associated with exposure to combust ion products are those resulting from impaired oxygen delivery or transport caused 5y impaired pulmonary function, decreased ambient oxygen or carbon monoxide binding to hemoglobin. Animal experiments are done in several different ways, Sehsvioural endpoints, for example, time to incapacitation, are sometimes determined. This is considered to be directly related to escape capability as Incapacitation is defined as the onset of staggering, prostration, collapse or convulsions (SOl. Physiological changes can be monitored by recording h* H 1S3 li N G* BFG10332 changes in respiratory rate, heart rate and rhythm (EKG), brain wave activity ! (EEG) and blood pressure (2A). The time to death for animals exposed to } decomposition products under specified conditions has also been used to i determine their toxicity (91,52,53,3^.55). This does not allow for consideration of slow-acting toxicants such as HC1. The traditional toxicological approach is through determination of the dose required to cause death in 50% of the animals (LD50). ft can be determined for a single toxicant but combustion or pyrolysis atmospheres.are complex and may contain many varied toxicants making it impossible to determine a true "dose-response" relationship. The LC50 (sample weight lethal to 50% of the animals) is sometimes used to assess toxicity (35). It has also been defined in the literature as the concentration of material in the atmosphere inhaled that will produce 50% mortality (36,97). Apparent lethal concentration (ALC50) is defined as that concentration of gaseous pyrolysis products in the atmosphere being inhaled which will produce 50% mortality (S6). Histopathological studies have been carried out (98,99). The type and extent of damage in the tissues of different parts of the animals is a 'good indication of the toxic substances that were inhaled before death. According to Hilado (100) 85.66% of the plastics sold Jn the United States are less toxic than wood. These results were based on "time to death" data for mice in an unvented chamber, and carbon monoxide and methane were the only gases for which analyses were done. It was noted that, in the case of A3S, the co.ncentrat ion of carbon monoxide was too low to have caused death (100,90,101,102,103). On the basis of these tests PVC was said to be less toxic than wood (100). These results have been questioned on the basis that much HCl probably did not reach the animals due to condensation and the fact that it is not a fast-acting toxicant (2A). The latter point is important because only deaths occurring during exposure were recorded (100). 2 When wood is exposed to a 2.5 W/cm radiant heat flux in a non-flaming N mode, animals are incapacitated at a carbon monoxide level of 7^ (22). This is likely due to inhalation of a mixture of aldehydes including acrolein, an extremely strong lachrymator which is lethal to humans'at concentrations of 10 ppm (see Appendix l). At a higher heat flux (7-5 w/cm') which is just below the ignition point the animals do not develop symptoms until a much higher carboxyheroglobln level is present (22). Under these conditions the aldehydes t would form and then be rapidly destroyed. Quite different toxicities were observed using dynamic air flow conditions MHffiR BFG10333 end ter.ptr3tures from kOO - EOO'C. Toxicity factors, based on the concentrations fatal to msn in 30 minutes, the weight of sample, and the volume of each toxic gas produced were calculated and are shewn in Table 12 (16). From these data it can be seen that the toxicity of the decomposition products is primarily due to HCN and HC1 in the case of A5S and PVC, respectively, with carbon monoxide playing a minor role. . Table 12 (16) Calculated Toxicity Factors (1/g) <N SI Materials PolyacryIonitri1e Nylon 6 ABS pipe PVC White Pine CO HCN HC1 7 1 1201 - 17 1 931 - 10 1 267 - 12 1 - 3^3 *t7 3 - - A study of 72 residential fires in Boston (26) in the early 1970's using a sampling system built into the firefighters' turnout coats, showed that acrolein, benzene, hydrogen chloride, hydrogen cyanide, nitrogen dioxide and ( carbon dioxide were present at real fires. The carbon monoxide concentration in only 2$% of the fires exceeded 500 ppm. At this concentration people would experience hallucinations after 30-120 minutes (Appendix l). A lethal concentration to man in 30 minutes would be about .^,000 ppm (16). Carbon dioxide at the levels observed in these fires would not have been sufficient to be a significant hazard in contrast to engineered burns where high carbon dioxide concentrations were observed. In only six fires did the oxygen concentration fall below 18%. A further study in 197& (26) again showed that low oxygen concentrations were not generally significant. This also is in contrast to results from model fires which consistently showed low oxygen concentrations (26). An investigation of 135 fire injuries and deaths, carried out by paramedics taking blood and breath samples from the victims immediatly after their removal from the fire showed that many had low csrboxyhemoglobin levels (COHb) (22); 2k% had COHb greater than 50^, 36* had COHB of 11-^5^ and \\% had COHb of 7`10t. Clearly carbon monoxide alone did not cause a high percentage of these deaths. 'hat did cause them remains unanswered in many cases. V-Tien a mixture of toxic gases is inhaled synergism (meaning more than additive), additive, or antagonistic effects can occur. This is a subject TO II Illl--l 11 " - BFG10334 e rm / yTT> that is not well understood. There appears to be a strong synergistic response between blood alcohol and carbon monoxide. If a person has a blood alcohol content greater than 0.1% (volume) his CCHb need not be higher than }Q% for death to occur (22,23). It is interesting to note that both alcohol and carbon monoxide interfer with the availability of oxygen to the brain. It is therefore, a reasonable expectation that hydrogen chloride which interfers with breathing and hydrogen cyanide v.hich interfers with use of oxygen by tissues r.ay be members of the same synergistic complex. The ratio of the percentage of carbon monoxide in the blood to the percentage of carbon monoxide in the atmosphere is always higher with fumes from the combustion of materials than it is with carbon monoxide alone (2*0. It is possible that this is due to increased respiration caused by other gases present such as carbon dioxide or hydrogen cyanide. A statistical analysis of data on the combined lethal effects of carbon monoxide, carbon dioxide and oxygen depletion showed that generally the effect of a combination of these factors was additive and even when synergism occurred its contribution was minor (lO1*). An increase in the ambient temperature which is likely to occur during a fire has been reported to lower human resistance to toxic fumes (105). This is an area where much work remains to be done if we are to understand what is occurring. In time such studies may indicate why people die in atmospheres in which the maximum tolerance levels for various gases known to be present have not been exceeded. 1V -- 1 PVC Toxicology 1. Physiological Responses a). Changes in Respiratory Rate Physiological changes accompanying the decrease in respiratory rate have been wel1-documented in laboratory animals and humans (106). The irritant nature of pyrolysis cases from PVC was assessed using the decrease in the respiratory rate of mice (S8). The animals were exposed under dynamic air flow conditions. For mice the RD^g in mg/1 of pyrolysis products were found to be: (106) (RD^q 's the sample weight that caused at 50% reduction in the respiratory rate) GZ0Z.ZTTZ TSWKMU....WJUI MmilUHM'HL.Jk 4MHMMIM JRW tsmmmmmtmmmmmmmm BFG10335 R050 m9/1 Material pyrolyzed 0.2*1 0-50 0.19 0.45 (l"l5 f P"O Douglas Fir PVC (no plasticizer or fire retardant) FVC (piasticizer) HC1 The concentration of chemicals which caused RD 50 in mice were found to be Intolerable to man. The 0.1 RD^ value for mice W c s unccmfortab1e but tolerated by man (107)- b). Other Physiological Responses Thermal decomposition of polymers can produce chemical species that may contribute little towards sensory irritation but which would have a significant lethal potential. For example, benzene which has been shown to be produced from PVC at the same time as hydrogen chloride (20,21) is a powerful sensitizer of the heart to epinephrine. Thus cardiac arrhythmias (irregularity of the heart) and ventricular fibrillation (rapid and erratic contraction of the individual muscle fibres of the ventricular walls of the heart producing weak and Irregular heartbeats) could be induced, particularly if a person was also experiencing a deficiency of oxygen. Such effects on the heart during exposure of rats to thermal decomposition products of various polymers have been noted (106). The maximum rate of decrease in respiratory rate is not as important as the time of onset of irritation and its duration. The total response was quantitatively related to the total stress imposed on the animals by the Sensory Irritation Stress Index (S1S1) (98). On this basis PVC was more toxic than douglas fir and polyurethane. A comparison of PVC with HC1 suggested that the sensory irritation response was due primarily to HC1 (98). Rabbits were exposed to PVC pyrolysis products under smouldering (Ii00oC) and flaming (800C) combustion for thirty minutes and the arterial pO^, pCO^, pH, central nervous system (EEG) and cardiovascular system {EKG) were monitored (2*0. As in the respiratory rate studies, the intoxication syndrome due to combustion of PVC was found to resemble the results of exposure to HC1. The concentrations of carbon monoxide present were insufficient to have killed the animals. The rabbits die^ during the recovery period, so~.etim.es after 15-p0 days. JO H h* NI d CO With carbon monoxide alone, the EEG activity decreased and cardiovascular rod ifications were not large (2*0- On the other hand, both HC1 and PVC T 111,1 Ml BFG10336 decomposition products caused great decay of the cardiovascular system, lowering of the EEG and rapid appearance of acute pulmonary edema (abnormal accumulation of serous fluid in the lungs). When a water trap to remove HCl was present only the effect of carbon monoxide was observed in the rats and they survived exposure to the pyrolysis products from about five times as much PVC (2k). With wood fumes the effects were the same as for carbon monoxide for the first fifteen minutes, but Just after this period a dramatic decay of the Cord!ovascu1 ar system occurred. Death followed at about the end of the thirtyminute intoxication period, and acute pulmonary edema was observed. Acrolein and other aldehydes known to be present in wood decomposition products (127) could be responsible for this effect (2k). Kishitani (108) reported that EKG (cardiovascular) abnormalities in mice exposed to PVC degradation products appeared prior to the generation of large quantities of smoke. This is a similar effect to that experienced by firefighters at PVC fires (see Section IV-M, and is more evidence that HCl is evolved before detectable amounts of smoke. The COHb of dead mice exposed to PVC degradation products averaged 21.2? which is much below the lethalarount of about 70% (103). Workers acutely intoxicated by inhalation of the combustion gases from PVC had moderately elevated venous CCHb chloride is known to cause this, whereas carbon monoxide results in high pOj (2*i). The results of all of these physiological experiments seem to indicate that the toxicity of P'?C degradation products is mainly due to hydrogen chloride. 2. Mortality a). Time to Death and Lethal Concentration Time to death is a very poor index of toxicity because extremely toxic chemicals are not necessarily fast-acting (111,22). Using time to death as a toxicity index PVC was less toxic than wood (55,112,113). However, when the sample weights lethal to 50? of mice during a 30-minute exposure period and a 10-minute recovery period are determined the values were 12-15 grams of PVC and 6g of douglas fir (THi). Dynamic conditions were used and PVC is clearly much more toxic than wood in this case. Whether static or dynamic conditions were used was found to make a significant difference to the toxicities of materials, with irritant gases such as HCl playing a much greater role under the dynamic conditions (1)5). It is interesting to note that when HCl, CO and C0^ are removed from PVC pyrolysis products the time to death for mice is about twice as long (2*0* T tw rrry. .4WMWMII! BFG10337 . jjwnl! ,.;i. I I, ' 3. Histopatho 1oqIca1 Studies Based on pathological changes observed in mice PVC is much more hazardous than douglas fir or polyurethane (S8). It must be noted that the nouse is an obligate nose breather and man is a voluntary nose breather. Vhen irritation occurs, nan will convert immediately to mouth breathing and the first susceptible epithelium would be the larynx. Damage to this area comparable to that seen in the nose of the mouse exposed to HC1 or PVC pyrolysis products would markedly obstruct the airway either killing the victim or necessitating immediate surgical therapy (115)- A comparison of the damage to tissues from exposure of animals to HC1 and PVC degradation products led to the conclusion that HC1 was the principal contributer to the toxicity in the case of PVC (S9,2k,22). After exposure to PVC combustion products, the development of completely abnormal cells was observed in rats (22). If the PVC contains antimony a massive haemorrhage may occur k-5 days later (22). IV-2 ABS Toxicology 1. Behavioural Endpoints and Mortality Nearly all the experiments with ABS have been done using a method developed at the University of San Francisco referred to as the "USF test" (93). A second method, the Federal Aviation Administration-Civil Aeromedical Institute (FAA-CAMl) method has also been used (116, 117)- The times to various behavioural endpoints and to death were determined. In these experiments the exposure chamber atmosphere was never analyzed for HCN or NO^, but it was noted that the concentration of carbon monoxide was insufficient to have caused death (118,30,101,102,103). This is in contrast to the case of polyethylene where the major toxicant is carbon monoxide (116,119). 2112703; Vhen the "USF" methods were used the introduction of forced air flow (G,H,|) caused a marked reduction in deaths in the case of polyethylene but it resulted in decreased times to incapacitat ion and death in the case of ABS. Table lk (120,116,119). The times to these endpoints also depended on the rate at which the sample was heated, and were shorter when A3S was exposed to a fixed temperature of 800C than when a rising temperature program was used^Table lk) (12 1 ,122,120, 103) The relative toxicities of the pyrolysis products appeared to be fairly insensitive to the temperature program used (97). The effect of oxygen depletion was not discussed in any of these papers although in a control experiment using methods B,E, and F (see Table lk) the oxygen concentration was reduced to 12% after 30 minutes mmmwmmmmmmm mmww BFG10338 Table 1* Results for "USF" Tests on Mice (120,116,119,118,113,55) Method Temp. C B 200-800 E 600 F 800 G 800 H 800 1 6 00 J 600 Air Flow (1/min) 0 0 0 1 3 I 3 Sample PE ABS woods PVC PE ABS PE ABS woods PE ABS PE ABS PE ABS PE ABS Time to Incapacitation (min.) 16.68 13-52 10. 3* 5.95 9.95 5.83 3-57 3-09 3-23 2.91 1.69 5.11 1 -^7 5-31 2.8* A.08 2.80 Time to Death (min) 22.60 17-62 1*.72 16.37 16.62 18.52 11.72 9.51 6.16 no deaths 3.80 no deaths 3-*7 *.53 9-69 5-69 Mortali 8/8 20/20 8/8 8/8 8/8 T8/16 0/8 8/8 0/8 8/8 1/8 8/8 7/8 8/8 21127033 UHMII.M 1 BFG10339 (6>). Using the apparent lethal concentration values as a toxicity index for mice in a sealed chamber A5S and PE appear to be more toxic than wood in contrast to the. results using t igje to dC5tb.i The ALC^q yalues in mg/1 were (116,123,36): ABS,2 PE ABS, 1 Douglas Fir Red Oak 10.45 11.81 20-32 22 JO 64.00 The large discrepancy in results for the two A2S samples probably is doe to their composition. Varying the ratio of acrylonitrile: butadiene: styrene gives a tremendous variety of ABS products. This results In confusion as many papers,do not state the composition of the ABS used. The lethal dose (LDj-qI based on the quantity of material which after combustion caused 50% mortality in mice in 10 minutes was determined and the relative toxicities were ABS > polyethylene > PVC, under these conditions Table 15 024J. ~ In the case of ABS and other polymers containing nitrogen, as the char rield increased from 0-20% the time to death increased But at char yields greater than 20% time to death decreased (33). The higher the char yield, theoretically the higher the fraction of the original carbon retained in the char and consequently the lower the concentration of carbon monoxide produced Cl011. Thus it would seem that the toxicity due to ABS is related to toxicants other than carbon monoxide as already observed in other experiments. Hydrogen cyanide and nitrogen dioxide, both detected and measured in ABS pyrolysis gases Cl71, are, on the Basis of LC^q values, about twenty.-five times as toxic as carbon monoxide (125). Table 15 024) LC50 Values for Plastics ABS 0.034 g PE 0.052 g PVC 0.140 g 21127034 ism mimam BFG10340 IV-3 Heal Fires Involving PVC In PVC fires a study of the medical histories of over 100 firemen showed that they became disoriented and that their COHb was usually less than 20* (22). In another study it was observed that at 17^ COHb some subjects had a momentary lapse of attention and failed to respond to all the stimuli presented (2). This could explain the fact that firefighters, having detected the sharp odor of- HCI do not always seem capable of escaping from the vapors. Several PVC fires have been investigated (19). In one case less than one pound of electrical insulation (plasticized PVC) burned yet five of the eight firefighters experienced shortness of breath, eye irritation, loss of balance, loss of co-ordination, numbness and chest pains and had to be treated in emergency. All exhibited a rapid Irregular pulse and developed a wheezing condition and coughs during the postexposure period. A second fire involving a small amount of PVC pipe and plasticized PVC, resulted in firefighters experiencing respiratory difficulty immediately upon entering the building although a very limited amount of smoke was observed. After the fire was extinguished they experienced the same symptoms as above. Upon returning to the station these firefighters complained of a burning sensation of the neck, wrists and elbows. Twenty-four hours after exposure they experience's a sloughing of skin where they had experienced the burning sensation (19). A third fire involving plasticized PVC electrical insulation also resulted in the incapacitation of the firemen (19). Several important features of these fires and those reported by Esch and Dyer (2) were that all of them were small and easily extinguished, the level of smoke was described as low to moderate, and the Intoxication syndrome described by all of the firefighters was essentially the same. On a larger scale, a PVC fire in 1975 in the New York Telephone Company installation generated smoke so dense that it hindered all phases of fire fighting operations and resulted In over 200 people being treated f6r smoke inhalation (1)). w h* b It is clear that smoke and toxic gases from burning or heating PVC are dangerous to people and that the use of PVC in buildings is a significant concern. CO (jj It is also apparent once again that these involved in fighting fires in which plastics are consumed should be extremely cautious. Such fires, depending upon the conditions, may cause relatively little to extremely dense smoke. BFG10341 Both extremes are dangerous. On the one hand, toxic gases may be quite concentrated with little snake; on the other hand both toxic cases and the effects of dense snoke can be very damaging to the people involved unless proper precautions are taken. IV- k Real Fires Involving ABS and Plastics Known to Produce Hydrogen Cyan? d e In 1977 a fire In a Tennessee jail resulted In k2 deaths. The only thing which burned was the cell padding which was styrene-butadiene rubber covered with neoprene-covered nylon fabric. It had been patched with toluene-diisocysnate polyether-polyurethane covered with PVC-'nylon fabric. In this case, the cyanide could not account for the deaths not already explained by CO In the 10 victims autopsled. In spite of the relative simplicity of this fire, the toxicological significance of elevated HCN in fire fatalities was not resolved (28). In the case of the MGM fire, blood cyanide in one victim was 7-9 pg/ml and COHb was 22.2?. Several other victims with blood cyanide levels of 1-2 pg/ml had COHb l8.2-^S.2?. One with cyanide levels of 2-3 pg/ml had COHb ^0?. None of these COHb levels is considered high (7). In fact an analysis of the data in (7) showed that the COHB was below the accepted lethal concentration in all but one case (Table 16). Whether the cyanide levels in conjunction with the CO levels in the victims were sufficient to have caused death is unknown with one exception. Blood alcohol for these victims was'inot reported (7). Cyanide toxicity is a difficult problem. Just what the toxic concentrations are is not agreed upon. Normal blood cyanide levels for humans range from 0 to 0.22 pg/ml, with an average of 0.05 pg/ml for non-smokers (128). These higher background levels are attributed to cigarette smoking. The above results were obtained on 32 living subjects and 22 non-fire related deaths (128), In acute cyanide poisoning, the blood level is likely to be 5 pg/ml and inhalation of HCN results in signs and symptoms of acute toxicity at blood concentrations at or above 0.2 pg/ml blood (129). Blood cyanide in fatal cases may be below 1 pg/ml when HCN gas Is inhaled (130)- By contrast Caplan et al. (126) considered 0,26-1.0 pg/ml as being sub-toxic concentrations In a report on cyanide concentrations in 25 fire fatalities. Elevated blood cyanide has generally been found with significantly elevated COHB In fire victims, in contrast to the autopsy results from the HGfl fire. 1127036 j lyjigpii BFG10342 Table lb COHb Concentrations in Victims of the MGM Grand Hotel Fire (?) % COHb 5 10 11-20 21-30 31-40 41-50 51-60 > 60 No. of Victims 4 9 17 2* 7 9 1 (% of Victims) (5.6) (12.9) (23-9) (33-8) (9-9) (12.7) (1.4) 21127037 V-1 Rates of Smoke and Acid Gas Release In Rooms Two examples taken from an article in the literature by E.E. Smith are instructi-. on this topic (3). The release rate of HC1 is such that 2.0 ft (60.96 cm) of 1.5 inch (3-81 cm)'PVC conduit when exposed to a temperature of l,000eF (538C) will release over 0.1 lb (45.4 g) of HC1 per minute. At this rate the air in a 10 x 10 x 100 ft (3-05 x 3-05 x 30.5 m) corridor in which the combustion occurs would reach an average concentration (assuming rapid mixing) of 100 ppm HC1 in less than 45 seconds. This concentration of HC1 would be intolerable to breathe. The use of PVC in ducts is also a matter of concern should it suddenly 22 become exposed to fire or intense heat (3). Assume that 100 ft (9.30 m ) of vinyl-covered fibrous glass duct liner is in the concealed space above a 4,000 ft^ (113.49 m^) compartment and that it is suddenly exposed to a 22 heat flux of 2.6 Btu/sec.ft (2.95 watts/cm ). It is assumed that the ventilation system pulls 10S of the compartment's air (400 ft3 or 11.35 m^) 2 per minute into a corridor. In less than one minute 775 "particles"/ft. x 100 ft^ (8,333 "part i cl es"/metre^ x 9-3 n>^} or 77,500 "particles" will be released into the compartment reaching a maximum of 18.5 "particles" per ft^ (660.7/m^). The rate of smoke "particle" emission when there is 30% transmission of light (O.S.U. Apparatus - Section 11-1). was determined to be 198 "particles" per minute (9). In the first minute following the "flash" fire, over 7,000 smoke "particles" will be discharged into the corridor. If these "particles" were uniformly distributed along a 10 x 10 x 50 ft (3.05 x 3.05 x 15.25 m) or 5.000 ft3`(l4l.86 m3) corridor the concentration would be 7,000/5,000 BFG10343 qr l.k "part icles"/f (7000/1^1.86 = kj.lk "particles" per m^) . This concent rat ion produces a degree of obscuration which would make it impossible to see an exit sign at a distance of 5-5 feet (1.68 m) . The major hazard from the duct material would be from the HC1 released, During the first minute of exposure 7-5 g/ft^ x 10Q ft 2 (80.65 g/m2 x 9-3 m^) (9) or 750 g (1.65 lb) of HC1 would be released. Over the next minute, 75 grams (0.1.7 11?) of HCl would be discharged into the corridor where the average concentration could reach k00 ppm, raking it unsafe as an escape route. If the same surface area of red oak were present in the same concealed space the smoke concentrations reached in the corridor would be very different (9)- Because of the larger mass of red oak present the total amount! of smoke could be larger, but, b'ecause of the slower release rate, the maximum number of "particles" per minute discharged into the corridor would be less than 25% of the number from the duct liner (9). Using the amounts of HCN and N0^ measured when ABS is pyrolyzed at 800C (17)., we can calculate the concentrations of these gases in an average apartment taken to be 10,000 ft^, assuming that there are 100 lb of DWV* pipe and fittings in that apartment. These calculations- also assume the composition of the ABS to be 22% acrylonitrile, 19% butadiene and 59% styrene as in (17). From 100 lbs of this ABS at 80Q*C, assuming that all of it burned, 1,733 1 of HCN and 3-3 1 of N0_ (at STP) would be produced. Thus 13 the average concentration of HCN in the 10,000 ft apartment would Be 6,lA0 ppm, about 22 times the concentration (280 ppm) that would be immediately fatal (see Appendix 1). The concentration of N0 woold Be 11.7 ppm, a level which would be mildly irritating to the eyes, nose, and respiratory tract (see Appendix l). V-2 Behaviour of PVC and ABS Pipes in Fire Tests Opposition to the use of plastic pipe in high-rise buildings particularly, is based on the fact that it will burn through and thus might allow fire gases, smoke, and possibly flames to penetrate fire partitions. A paper by P.C. Attwood (109) in which an extensive series of small scale tests are described using a positive pressure of 5 mm within the fire compartment and either a horizontal or vertical penetration of the partitions is quite informative. He discusses several possible situations involving the use of plastic pipe. 'Drain, waste, vent. SeoAZTTZ hr-a-su'VW -'v-w-wa* nfcm BFG10344 Horizontal Pipes 1. Unprotected The wall was 5/8-in.type X gypsum board mounted on both sides of 2 x 6-in, pine studs. The exposed s-ide of the plumbing consisted of a 2-in (50.8 mm) lateral with a water-filled P-trap. This lateral penetrated the wall and joined the stack within the wall cavity. The exposed gypsum was backed by_ a sheet of 2^ ga. sheet metal and the PVC wall section was fire-stopped at the top. (10k) ABS laterals sagged and collapsed within 7 minutes; the pipe burned through in 10 minutes for A3S and 13 minutes for PVC, completely exposing the penetration and igniting the pipe and studs within the wall cavity. 2. Steel-Sleeved Laterals at a *<5* Angle With both 1.5-in (38 mm) and 3"in (75 mm) PVC pipe penetrating walls at a ^5" downward angle protection was provided for up to two hours by the t _ formation of a plug due to the intumescent nature of PVC and to char formation. A 3-in diameter pipe would be the largest one that would form such a plug. Pipe above the plugs had been totally consumed. The same test with 1.5-in A8S pipe resulted in the pipe being totally consumed leaving an open path for fire gases. A test with ABS pipe penetrating a wall upwards at a *i5 angle gave similar results. With l-5"in PVC pipe, penetrating the wall upwards at a angle, it was found that a barrier against flame and gas propagation was formed due to swelling and charring of the pipe as long as the unexposed end of the assembly was unvented. QqI<J_ 3. Mechanical Shut-Off Devices Mechanical shut-off devices which relied on the softening behaviour of thermoplastics at elevated temperatures were also tested. To be effective the closure device should shut as the plastic collapses, creating a seal. A variety of single shut-off devices were tested with failure times ranging from 28 minutes to 115 minutes. Such devices were generally unsatisfactory, and the 115 minute duration of one test was only a result of severe blistering of the metal assembly presumably caused by acidic products (HCl) from combustion of PVC. A double shut-off device was more successful and provided protection for up to two hours, although with ABS, combustion products vented through the plumbing system until the second device pinched off the pipe after 60-70 minutes. (' aw .iuuuurn BFG10343 Vertical Pipe Assemblies 1. Unprotected Pipe A *<-inch (10.2 nun) pipe extending seven feet (2.13 m) above a concrete slab and U inches below it "was used for the tests. A twelve-inch, long steel sleeve was around the pipe. The *t-inch end of the pipe was heated in the furnace. After 1*1 minutes for A3S and 22 minutes for PVC the pipe had softened sufficiently to collapse in a heap around the sleeve on top of the slab creating a temporary seal. The ABS ignited immediately whereas the PVC burned through after 35 minutes. {10A) 2. Stacks Within Vented Chases In each test damage was substantial: a remnant of pipe remained in the top of the chase, the rest of the pipe having collapsed into the bottom of the chase. In one PVC test and the ABS test the pipe in the chase was totally consumed. In two PVC tests the furnace gases had created a path through the PVC ash, thus exposing the penetration. (10*0 3. Mechanical Shut-Offs In the time before a seal is created (30 minutes and 25 minutes for 3-inch * (75 mm) PVC and ABS and 22 minutes for both *)-inch PVC and ABS) hot combustion gases and smoke are vented through the plumbing system. (10*0 *. Unvented Applications Three-inch copper, ABS and PVC drain pipes were connected to water closets with the trap filled with water. In the PVC and ABS tests the pipe within the furnace was consumed and a small amount of smoke and fume seepage was evident around the ungrouted base of the water closet. By the end of two-hour tests all ABS and PVC pipe, flanges and rubber seal rings had been consumed. In the copper tests, the system remained intact except for the seal ring. Severe cracking of the water closet base occurred in each test, but only on surfaces not contacting water. There was no loss of water in any test but smoke seepage occurred through the cracks in the water closet. Considering the rapid rate of dehydrochlorination of PVC at reasonably low temperatures there is a possibility that significant amounts of HCl would be vented through the plumbing system before any mechanical shut-offs would have sealed the pipe, or through the cracks in the case of water closet Unprotected plastic pipes, both horizontal and vertical are obviously hazardous (10* BFG10346 It must be concluded on the basis of the above that integrity of fire partitions penetrated by either ASS or PVC pipe is indeed compromised and that the mechanical shut-off devices tested were generally ineffective despite the author's conclusion (10M that plastic pipe can penetrate fire separations without propagating fire beyond the separation. X V IH L Z X JZ mMAA-MWlLliL .pump HI IMJIJUSWA U1M1I BFG10347 Cone 1 usions Some conclusions derived from this literature search are the following: 1. The main product of pyrolysis or combustion of PVC is hydrogen chloride vapor which is not only extremely irritating to the eyes and nose but causes extensive lung carnage when inhaled. 2. ABS is a very flamm.able plastic which on combustion produces deadly hydrogen cyanide as well as carbon monoxide. 3. The length of time between the initiation of combustion of ABS or PVC in a room and the point at which concentrations of smoke and toxic gases become over powering to a victim is short probably about one minute; ABS and PVC under test conditions produce many times more smoke under flaming conditions than a number of woods that were tested. 5. Ionization detectors are relatively insensitive to the products of combustion of PVC sometimes giving a warning after the concentration of hydrogen chloride has built up to several times the tolerable concentration. 6. Use of plastic pipe plumbing systems in high rises and multi-unit dwelling places can contribute to the rapid production of. dangerous levels of smoke and toxic gases as well as spreading of flames should the components of such a system catch fire. On the basis of experiments described in the literature it is evident that fire proof penetration of fire partitions by plastic pipe has not yet been achieved. BFG10348 H>Z.2TT Ac It now! edoements The authors wish to express their grateful appreciation to Or. E.S. Kail who reviewed and commented on the contents of this document. They also wish to thanks Misses J. Porter, C.l. Seifert and D. Bundgsard who typed the manuscript. The work was funded by the Canadian Foundry Association Cast Iron Soil Pipe Division, Edmonton, Alberta. \ M H- 2 CO BFG10349 P.e fe fences 1. Caski11 , J.R., "Smoke Development During Pyrolysis or Combustion" in Smoke and Products of Cc~bustion, Vol 2. fire and fiammabi1ity series edited by C.J. Hilao'o, Technomic Publishing Co. Inc., Westport, Conn. 1973. p. 1. 2. Dyer, R.F. end Esch, V.H., "Polyvinyl chloride Toxicity in Fires, Hydrogen Chloride Toxicity in Fire Fighers", J. Am. Med. Assoc. 1976, 235, 393. 3. Woolley, W.D., " The Hazard of Building Contents in Fires", Second Symposium on Combustibility and Plastics, Ottawa, Canada, October 30-1, 1979, P- 117- 4. Clark, P.R., "Plastics and Fire: A Fire Fighter's View", Second Symposium on Combustibility and Plastics, Ottawa, Oct. }0~] , 1979. P> 218. 5- "Fire at the MGM Grand, a preliminary report". Fire Journal, March 1981, P. 33. 6. Bush, V.R., "MGH Grand Hotel Fire, Las Vegas, Nevada", Building Standards, Jan-Feb. 1981, 6. 7. Clark-County Coroner-Medical Examiner, MGH Grand Hotel Fire Report. 8. Modern Plastics Encyclopedia 1980, p. 1*. 9. Smith, E.E., "Fire Hazard Characteristics of Duct Materials", ASHRAE 1972, 14(7). 29. 10. lmhof, L.G. and Stueben, K.C., "Evaluation of Smoke and Flammability Characteristics of Polymer Systems", Polym. Eng. Sci., 1973> 13(2), 146. 11. Kourtides, D.A. and Parker, J.A., "Flammability Properties of Some Thermo plastic and Thermoset Resins", SAMPE Q. 1978, 9(3). 36. 12. Hilado, C.J., Cumming, H.J., and Machado, A.M., "Screening Materials for Smoke Evaluation", Modern Plastics 1978, 55(7), 61. 13* Gaskill, J.R., "Smoke Development of Plastics Under Various Fire Parameters", Soc. Plast. Eng.. Tech. Pap., 1972, l8(Pt.l), 264. 14. Gaskill, J.R. and Veith, C.R., "Smoke Opacity from Certain Woods and Plastics", Fire Techno!. 1968, 4, 185- 15. Hilado, C.J., VThe Effect of Chemical and Physical Factors on Smoke Evolution from Polymers" in Flammability of Solid Plastics Vol. 7 fire and fl amrr.abi 1 i ty series ed. by C.J. Hilado, Technomic Publishing Co. Inc. 1974, 93- 16. Sumi, K. and Tsuchiya, Y. "Toxicity of Decomposition Products - mm BFG10350 21127044 Polyacrylonitrile, Nylon 6 and A3S", 3uild. Res. Note Can. Div. Bldo. Res., 1 376, 111. Natl. Res. Counc. 17. Chaigneau, H. and LeMoan, G., "Pyrolysis of Plastic Materials VII. Polyacrylonitrile and Co-polymers", Ann. Pha rm. Fr., 137^. 32, k&$. .18 Comeford, J.J. and Birky, M., "Measurement of Smoke and Hydrogen Chloride Evolution from rVC", Fire Techno!. 1972 , 8(2), 85. 19- Einhorn, I.N. and Grunnet, M.L., "The Physiological and Toxicological Aspects of Degradation Products Produced During Combustion of PVC Polymers", Spec. Rep. - Electr. Pcwer Res. Inst. (Palo Alto) 1 g SO, EPRI EL - 1263, Flair.-abil* 1ty Solid rolym. Cable Dielectr., 3/1 " 3/^7- .20 O'Kara, H.M., "Combustion of PVC", Pure Appl. Chem. 1977, ^9(5), 6^9. .21 Voorhees, K.J., Ryan, P.V., Mickelson, R.W. , Einhorn, I.N. and Futrel1 , J.H., Proceedings 23rd Annual Conference on Mass Spectrometry and Allied Fields, Houston, Texas, 1975, p. 200. .22 Einhorn, I.N., "Methodology for the Study of Toxicology in Combustion : Application to PVC, J. Macromol. Sci., Chem. 1977, A 11(8), 1519- 23- Deichrr.ann, V.B. and Gerarde, H.W., Toxicology of Drugs and Chemical?, Academic Press, N.Y. and London, 1969. 2k. Boudene, C., Jouany, J.H. and Truhaut, R., "Protective Effect of Water Against Toxicity of Pyrolysis and Combustion Products of PVC and Wood, J. Macromol. Sci., Chem., 1977, A 11(8), 1529. 25- Hilado, C.J., Flammability Handbook for Plastics, 2nd edition, Technomic, Westport, Conn., 137^. .26 Trietman, R.D., Geld, A., and Burgess, W.A., "Air Contaminants Encountered by Firefighters", Second Symposium on Combustibility and Plastics, Ottawa, Canada, October 30-1 , 1979, p. 28. 27. Kourtides, D.A., Gilwee Jr., W.J. and Parker, J.A., "Therrochemlcal Characterization of Some Thermally Stable Thermoplastic and Thermcset Polymers",Thermosets High Perform. Thermoplast. 1977, 113- .28 Birky, M.M., Paabo, M. and Brown, J.E., "Correlation of Autopsy Data and Materials Involved in the Tennessee Jail Fire", Fire Safety J., 1980 2(1), 17. 29. Ber1, V.G. and Kalpin, B.M., "Fire-Related Fatalities: An Analysis of Their Demography, Physical Origins and Medical Causes", Fire Standards and Safety, A.F. Robertson, editor, ASTM Symposium, Gaithersburg, Hd., April 5-6, 197&. .N 30. "Standard Test Method for Surface Flammability of Materials Using a Radiant Energy Source", ASTM E 162-176, ASTM Standards, Part 18, 1977. 31. Hilado, C.J. and Cumming, H.J., "Studies With the Arapahoe Smoke Chamber", BFG10351 J. Fire Fla-.-.zbi 1 i ty , 1977, 8, 300- 32. Lee, T.G., "Smoke Density Chamber Method for Evaluating the Potential Smoke Generation of Building Materials", l.'BS Technical 757, Jan. 1973 . 33- "Standard Method for Measuring the Tensity of Smoke from the Burning or CcCC.~?os i t ion of Plastics", AS7M D2S43-/0, AS7M Standards, Part 35,(1377). 34. "Standard Test Method for Surface Burning of Building Materials", ANSI/ AS IM 84^77, ASTM Standards, Part 1 8 (1 977). 35. Smith, E.E., "Measuring Bate of Heat, Snoke and Toxic Gas release", Fire Tech, 2972, 8, 237 36. Kourtides, D.A., Parser, J.A. and Hilado, C.J., "Therrochemica1 Character ization of Some Thermoplastic Materials", J. Fire Flammability, 1977, 8(l), 59- ' 37- Gaski11, J.R., "Hew Plastics Smoke", Soc. Plast. Eng. J., 1972, 23(10), 43. 38. Salto, F., Research Institute(Japan), Research Paper No. 33, 1988. 39- Calcraft, A.M., Green, R.J.S. and McRcberts, T.S., "Burning of Plastics .1. Smoke Formation", Plast. Polym., 197^, 42(16), 200. ^0. Keisters, M., More Realistic Testing of Smoke Generation", Plast. Enq 2976, 32(8), 49. ------------------ -- 4l . 3ankston, C.P., Powell, E.A. , Cassanova, R.A. and Zirin, B.T., "Detailed Keesurment of Physical Characteristics of Smoke Par11cu1ates Gene rated by Flaming Materials", J. Fire Flammability, 1977, 8(4), 395- 42. Zinn, B.T., Powell, E.A., Cassanova, R.A. and Bankston, C.P., "Investiga tion of Smoke Particulates Generated during the Thermal Degradation of Natural and Synthetic Materials", Fire Res, 1977, 1(1), 23- 43- Christopher, A.J., "Some Aspects of Smoke and Fume. Evo 1 ution from-Over heated Non-Metal lie Materials", J. Combust. Toxicol. 1976, 3, 89. lib. Hilado, C.J. and Machado, A.M., "Smoke Studies in the Arapahoe Smoke Chamber", J. Fire FIammsbi1ity 1978, 9, 240. 1)5. Cornish, H.H., Hahn, K.J. and Barth, H.L., "Experimental Toxicology of Pyrolysis and Combustion Hazards", Environ. Health Perspect. 1975. 11, 191. 46. Stone, J.P., Hazlett, R.N., Johnson, J.E. and Carhart, H.W., "The Transport of Hydrogen Chloride by Soot from Burning PVC", J. Fire Flammability 1973, 4, 42. 47. Henderson, Y. and Haggard, H.V., Noxious Cases, 2nd ed., Reinhold Publish ing Corp., N.Y. (1943), p. 71. 48. Morikav.a, T., "Evolution of Soot and Polycyclic Aromatic Hydrocarbons in Combustion", J. Combust. Toxicol. 19/8, 5, 349- VTT?. BFG10352 KS. Liao, J.C. and Browner, R.R., "Determination of Polynuclear Aromatic Hydrocarbons in PVC Smoke Particulates by High Pressure Liquid Chromato graphy and Cas Chrcm.stocraphy - Hass Spectro.metry", Anal. Chan. 1578, 50(12), 1683. ' 50. Korikewa, T., "Evolution of Soot and Polycyclic Aromatic Hydrocarbons in Combust ion", Shcbo Kenkvusho Hckoku 1578, A'5,- 13. 51. Chaigneau, H., 1'Fornation of Some Polycyclic Hydrocarbons in the Course of the iherrr.al Decracation of PVC", C.R. Hebd. Seances Acad. Sci. Ser. C. 1S78, 286 (8), 257. 52. Vagner, J.P., "Smoke Detector Character!sties", Proc. Int. Conf. Fire Safety, 1977, 2, 53* Velker, R.W. and Vagner, J.P., "Particle Size and Hass Distribution of Selected Smokes: Effect on Ionization Detector Response", J. Fire Flamm. J977, 8(1), 26. 5^. Purt, G., "Fire Protection for Fires With PVC", Ver. Foerder. Deut. Srandschutzes Z. 1565, l8(k), 156. 55. Hilado, C.J., Cunnings, H.J. and Casey, C.J., "Toxicity of Pyrolysis Gases from Natural and Synthetic H.aterials", Fire Technol. 1978, 1^(2), 1. 56. Voolley, V.D., "Dehydrochlorination of PVC in Nitrogen and Air", P1 ast. Polym. 1972, k0()A8), 203. 57- Chaigneau, M., "On the Evolution of Various Pollutants from the Thermal Degradation of PVC", C.R. Hebd. Seances Acad. Sci. Ser. C. 1976, 282(13), 583. 58. Bradt, P. and Mohter, F.L., "Hass Spectra of Thermal Degradation Products of Polymers", J. Research N'atl . Bur. Standards 1355, 55 , 323~7. 53. Paulik, J., Paulik, F. and Erdey, L., "Combined Derivatographic and Thermal Gas - Analytical Studies", Kikrochim, Acta. 1388, 886. 60. LeHoan, G. and Chaigneau, H., "Pyrolysis of Materials in Plastics .1. PVC CharacterizatI on of Possible Volatile Toxic Compounds", Ann. Pharm, Fr. 1989, 27, 97. 61. Carel , A.B., "Pyrolysis of PVC and Experimental P(VC) Compounds Using a Thermal Gravimetric Analysis Furnace and Oven Technique with Recovery Studies of Hydrogen Chloride", Proc. Int. Wire Cable Symp. 1978, 27, 239. 62. O'Kara, H.H., "High Temperature Pyrolysis of PVC : gc-ms Analysis of the Pyrolysis Products from PVC Resins and Plastisols", J. Polym. Sci. Part A-1 1970, 8, 1887. 63. Voolley, V.D., "Decomposition Products of PVC for Studies of Fires", Paper CP 11/7^ Building Research Establishment, Fire Research Station, Soreham.vood WPG 23C, Great Britain. f'aim.an, M.B., Papko, R.A. and Pudov, R.A., "Thermal Degradation of PVC (l). TT2 TT ww BFG10353 ' n) Vysokcoo1 Soedin, Ser, A. 1563, 10, k 1 . 65. Forstier, M. , "Combustion Products of Synthetic Materials", Rev. Tech. Feu. 1375, 16. 2k. 66. Autain, J . , "Toxicologic Aspects of FI b ia.T.r.s 1 i ty and Comb us t i on of Polymeric Materials" in Smoke and Froo'ucts of Combustion; Vol 2.fire and flammability series, ed. C.J. Hilaco, Techncmic Publishing Co. Inc., Westport, Conn., 1973, p. 57- 67. Irroto,. M., "Products by Pyrolysis of PVC", Setchaku 157k, 18, 5k0. 68. Michal, J., "Evaluation of the Assumed Toxicity of Some Plastics", l.'ehor 1 avos t . rolyn. Mater. 1976, 129. 63. Ke-bayashi, Yt, Hori, M. and Murata, H., "Combust ion Gases Generated from PVC and its Products. Model Experiments Simulating Fire", Jap. PI as t. 1971, 5, k0. 70. Chang, E.P. and Salovey, R., "Pyrolysis of PVC", J. Polym. Sci., Poly. Chem. Ed. 197^, 12, 2927. 71. Ball, G.L. and Boettner, E.A., "Toxic Gas Evolution from Burning Plastics", Polym. Prepr., Am. Chem. Soc. Div. Polym.Chem. 1973, lk, 86. 72. Woolley, W.D., "Toxic Products from Plastics Materials in Fires", Plast. Polym. 1973, kl, 280. 73- h'offz, D., Benz, W. and Pfab, W. , "Pyrolytic Gas Chromatography Study of High Polymers"., Fresenius Z. Anal. Chem. I968, 235, 121 . 7k. Hi ramatsu, K., "Pyrolysis Products of Polymeric Materials by Mass Spectro metry. IV, Mass Spectrometry of Pyrolysis Products of Poly(vinyl alcohol) and PVC", Shitsuryo Bunseki 15, 29. 75. Hagen, E., "Composition of Pyrolysis Gases of Plastics", Plaste. Kaut. 1968, 15, 711. 76. Takahashi , T. and Serizawa, M., "Pyrolysis of PVC", Enka . Biniiru To Porima 1971, 11, 26. 77- lida, T., l.'akanishi, M. and Goto, K,, "PVC.I. Evolution of Aromatics on Pyrolysis of PVC and Its Mechanism", J. Polym. Sci., Polym. Chem. Ed. 197k, 12. 737- 78. Goto, K. and lida, T., "Pyrolysis of PVC . Formation of Aromatic Compounds", Enka. Biniiru To Porima 1973, 13, 18. 79- Koichi, K. and Nakamura, K., "Toxicities of Combustion Products", J. Fire FI ammabi 1 i ty/Combustion Toxicology Suppl . 197k, 1, 10^. 60. Kishitani, K. and Nakamura, X., "Study on Toxicities of Combustion Products of Building Materials at Initial Stage of Fire", J. Fac. Eng,, Univ. Tokyo, Ser. B. 1377, }k, 295- BFG10354 21127048 r-s. ON -5-. Hilado, C.J., Solis, A.N., Marcussen, V.H., and Machado, A.M., "Effect of Temperature and Heating Rate on Apparent Lethal Concentrations of Pyrolysis Products," J. Combust. Toxicol. 1978, 3, 381. 58. Barrow, C.S., Lucia, H., Stock. M.F. and Alarie, Y.C., Am. Ind. Hygiene Assoc. J. 1979. A0, kC8. 59. Barrcw, C.S., Lucia, H. and Alarie, Y.C., "A Comparison of the Acute Inhalation Toxicity of Hscrocen Chloride vs the Thermal Decompositicn_ Products of PVC", J. Combust. Toxicol. 'S79 8, 3. 100. Hilac'o', C.J., Cu-m.ino , H.J. and Casey, C.J., "Toxicity of Pyrolysis Gases from Plastics and Elastomers", J. Elastomers Plast. 1979, 11, 3- 101 . h'ilao'o, C.J. and Machado, A.M., "Effect of Char Yield and Specific Toxicants on Toxicity of Pyrolysis Gases From Synthetic Polymers", Fire Technol. 1979, 15, 51. 102. Hilado, C.J. and Huttlincer, N.V., "Concentration - Response Data on Toxicity of Pyrolysis Gases from Six Synthetic Polymers", J. Combust. Toxicol. 1978, 5, Si. 103. Hilado, C.J., Soriano, J.A. and Kcsola, K.L., "Effect of Heating Rate on Toxicity of Pyrolysis Gases from So,me Synthetic Polymers", J. Combust. Toxicol. 1977, 4. 533. 10it. Attwood, P.C., "Penetration of Fire Partitions by Plastic Pipe", Second Symposium on Combustibility and Plastics, Ottawa, October 30-1 , 1979 P* 272. 105. Origgi, P., "Flammability of Plastic Materials", Mater. Plast. Elsstomeri. 1978, 120. 105. Barrow, C.S., Alarie, Y.C. and Stock, H.F. , "Sensory Irritation and Incapacitation Evoked by Thermal Decomposition Products of Polymers and Comparison with Known Sensory Irritants", Arch. Fnvir. Health 1978, 33, 79- 107. Kane, L.E., Barrow, C.S. and Alarie, Y.C., "A Short-Term Test to Predict Acceptable Levels of Exposure to Airborne Sensory Irritants", Am. Ind. Hygiene Assoc. J. 1979, ^0, 207* 108. Kishitani, K. , "Study on Injurious Properties of Combustive Products of Building Materials at the Initial Stage of Fire", J. Faculty of Enoineering, U. Tokyo (3), I97I, 31, 1-35. 109. Sakai, T., "Intake of Carbon Monoxide in the Mouse Exposed to a Mixture of Carbon Koncxlde and Hydrogen Chloride", Igaku To Seibutsugaku 1977, 94, ^75- 110. Colardyn, F. van der Straeten, M., Lamont, H. and van Peteghem, T., "Acute Inhalation - Intoxication by Combustion of PVC", Int. Arch, Cccup. Environ. Health 1978, 37 121. 111. Anderson, R.C. and Alarie, Y.C., "Approaches to the Evaluation of the Toxicity of Decomposition Products of Polym.eric Materials Under Thermal ^ 6K>2T yw 1.1^ imM.jnm v. i ; i .mniwv'i 81M*-,Vltfl BFG10355 8 1 . Do 5. re ppe - G i s c s r d , N., " Ccmpo ur. d s Obtained by Cc'bjstlon of Plastics", Trib. Centre Selce Etude Ddc. Faux 158, 21, 2^7* 82. Schriesheim, A., "Method for the Controlled Burning of Combustible Materials and Analyses of the Combustion Gases", J. research N'atl. Bur. Standards 1956, 57, 2^5. 83. Cordon, G.Ya., "Thermal Cehydrochlorinstion of Chlorine Containing High Folym.ers", PI ast i cheskie Messy lg6L (h) , 6. _ 81). Cool ley, V.D., "Behaviour of PVC in Fires", J. Macro-pl. 5c i. Chem. All, 1977, 1509- 85. O'Hara, M.M., "Pyrolysis Gas Chromatography Analysis of PVC.ll. In Situ Absorption of HCI During Pyrolysis and Combustion of PVC", J. Polym. Sci., Part A-1 1371 , 9, 1387- 88. Eitingon, A. 1 . , Solov'eva, T.V., Gribuncva, G. and "aurrova, L.S., "Toxi city of Volatile Products Emitted During the Burning of Polymer Materials Synthesized from a PVC Ease", Gig. Sanit. 1975 . 2, 10^. 87. Antonyuk, O.K. and Aldyreva, M.V., "Substantiation of the Maximum Permis sible Concentration of Dibutyl Phthalate in the Air of Industrial Premises", Gig. Tr. Prof. Zabol. 1973. 8, 26. 88. Chaigneau, M., "Mass Spectrometry Analysis of Compounds Formed by Pyrolysis of Polyacrylonitrile and Copolymers", Ana 1 us Is'1977, 5, 223. 89. Chaigneau, M., "Gas Emitted by the Pyrolysis of Different Plastic Materials at 1 ,000C.", C.R. Acad. Sci . , Ser. C. 137*. 278, 109. 90. Hilaoo, C.J. and Brauer, D.P., "Concentration-Time Data in Toxicity Tests and Resulting Relationships", J. Combust. Toxicol. 1979, 6, 138. 91. Hilado, C.J. and Machado, M., "Toxicity of Pyrolysis Gases from Some Cellular Polymers", J. Combust. Toxicol. 1978, 5, 162. 92. Hilado, C.J., Huttlinger, N.V. and O'/.'eill, B.A., "Effect of Heating r.ate on Toxicity of Pyrolysis Gases from Some Wood Products'.'. J. Combust, loxicoi. 1978, 5. 25. 93. Hilado, C.J. and Casey, C.J., "Pyrolysis of Polymeric Materials.!. Effect of Chemical Structure, Temperature and Heating Rate and Air Flow on Char Yield and Toxicity", J. Fire Flammability 1979, 10, 1*)0. S1). Chaigneau, H. and LeMoan, G.t "Toxicity of Plastics Combustion Products.il.", Ann. Pharm. Fr. 1977. 35, 153- 95. Barrow, C.S., Alarie, Y. and Stock, M.F., "Sensory Irritation and Incapaci tation Evoked by Thermal Decomposition Products of Polymers and Com,pari son with Known Sensory Irritants", Arch, Envir. Health 1978, 33. 79. 96. Hilado, C.J., Marcussen, V.H. and Furst, A., "Apparent Lethal Concentrations of Pyrolysis Products of Some Polym.eric Materials," J. Ccnbust. Toxicol. 1576. 3. 365. 21127050 'WJI1 m* BFG10356 Stress", J. Combust. Toxicol. 13 73, 5, 21 b. 112. Hilado, C.J., Dumming, H.J., Hour tides, D.A. and Gilwee Jr., W.J., "Relative Tcxicities of the Pyrolysis Products From Sorme Thermoplastic and Ther.moset Polymers", Thermosets Hich Perform. The rmopl as t. 1977, ISO. 113- Kourtides, D.A., Gilwee Jr., W.J. and Hilado, C.J., "Relative Toxicity of the Pyrolysis Products from Some Thermoplastic and Therroset Polymers", Pol ym. Eng. Sci . 1978, 18, 67L. liA. Alarie, Y.C., "Toxicologic and Acute Lethal Hatard Evaluation of Thermal Decomposition Products of Synthetic and natural Polymers", presented at Second Symposium on CombustIbi1ity of Plastics, Ottawa, October 30-1, 1979, p. 56. 115- Lucia, H.L., Barrow, C.S., Stock, M.F. and Alarie, Y.C., "A Semi-Quantitative Method for Assessing Anatomic Damage Sustained by the Upper Respiratory Tract of the Laboratory Mouse, Mus musculis", J, Combust. Toxicol. 1 977, A, ^72. 116. Hilado, C.J., Cumming, H.J., Machado, A.M. and Schneider, J., "Comparison of Animal Responses to the Combustion Products Generated by Two Test Procedures, the t'SF/f.'ASA Methodology and the FAA/CAMI System", J. Combust. Toxicol 1977, 325- 117- Hilado, C.J. and Crane, C.R., "Comparison of Results with the USF/KASA and FAA/CAMI Toxicity Screening Test Methods", J. Combust. Toxicol. 1977, ^,56. 118. Hilado, C.J., Cumming, H.J. and Casey, C.J., "Toxicity of Pyrolysis Gases from Plastics and Elastomers", J. Elastomers Plast. 1979, 11, 3- 119- Hilado, C.J. and Cummings, H.J., "The Effect of Test Conditions on Relative Toxicity of the Pyrolysis Products from Some Plastics", Fire Technol . 1977, 13, 325- 120. Hilado, C.J. and Cumming H.J., "Relative Toxicity of Pyrolysis Gases from Materials: Effects of Chemical Composition and Test Conditions", Fire Mater. 1978, 2, 68. 121. Hilado, C.J., Soriano, J.A., Kosola, K.L., Solis, A.N. and Furst, A., "Effect of Heating Rate on Pyrolysis Gas Toxicity of Synthetic Polymers and Elastomers", Proc. V-'est Pharmacol . Soc. 1978, 21 , 171. 122. Hilado, C.J., "Toxicity of Pyrolysis Gases from Materials", Mall Symp. Exhib. 1978, 23, 925. 123. Hilado, C.J., Cumming, H.J. and Furst, A., "The Effect of Test Conditions on Relative Toxicity Test Results and Rankings", Aercsp. Med. Res. Lab. (Tech. Rep.) AmRL-TR (U.S.) 1977, AKRL-TR - 77~97; Proc. Annu. Conf. Envir. Toxicol. 8th; AD - ADS1.33A, 77~S5- 12^. Chaigncau, M. and LeMoan, G., "Toxicity of Plastic Combustion Products 111. rapid Evaluation Method by Determiner ion of LD50 in Mice", Ann. .1127051 BFG10357 harm. Fr. 1577, 35, *<Sl . 125. Hilado, C.J. and Cumming, H.J., "A Review of Available LC50 Data", J_. Combus1 . Tox? col . 1 977. , kl5- 126. Caplan, Y., Thompson, B. and Altman, R., to be published. Reported in reference 28. 127. LipsVs, A.E. and Alvares, N.J., "Chem.lcal Constituents of Smoke and Conjecture as to Their Role in Fean Destruction" in Smoke and Products of Combustion, Vol 2 fire and flammability series ad. C.J. Hilado, Te cisr.cmi c Publishing Co. Inc., Vest port, Conn. 1373, P- 171- 128. Caplan, Y. and Altman, R., "X! c rode te rm i r.a t ion of Cyanide in Fire Fatalities", 2Sth Annual Meeting American Academy of Forensic Sciences, Wash. D.C. 1976, 129. Sunshine, 1 and Finkle, 8., "The Necessity for Tissue Studies in Fatal Cyanide Poisoning", Int. Arch. Gev.erbepa thpl . Gev-e rbehy g . .1 S8*< , 20, 558. 130* Reiders, F. and Sunshine, I., Kethodolocv for Analytical Toxicology, CRC Press, 1971. P- 115. I 31 - s uni, K. and Tsuchiya, K., "Combustion Products of Polymeric Hateri-als Containing Nitrogen in their Chemical Structure", in Smoke and Products o f Comb us tion Vo 1. 2 fire and flammability series ed. C.J. Hilado, Technomic Publishing Co. Inc., Vestport, Conn., 1973, p. 92. 132. Isaacs, J.L., "The Oxygen Index Flammability Test", Flammability of Solid Plastics, Vol. 7 fire and flammability series, Technomic Publishing Co. Inc., Westport, Conn., 197*1, p. 1. ZS0A2TT2 BFG10358 APPENDIX 1 HUMAN RESPONSE TO OXYGEN DEPLETION AND TOXIC GASES (25) Cxy c e .1 Depletion Si q n s arid Symptoms of Tcxicity of Reduced Levels of Cxycen % Oxycen in Air 20% 12-15* io-m 6-8% 6% or below 2-3% No symptoms Muscular co-ordination for skilled movements lost Nausea and vomiting, faulty judgment, rapid fatigue Collapse and unconsciousness, rapid treatment can prevent death Death in 6-8 minutes Death in ^5 seconds Toxic Gases The combined effects of breathing a mixture of gases is not well under stood. The information presented here applies to each gas as the sole toxicant. Carbon Monoxide Concentration of CO (ppm) Symptoms 100 200 300 A00 500 600 1000 1500 2000 3000 5000 12o00 ------------------------------------- ------------------- ------------------- ----------------------------------------------------------------------------------------------------------------------------------------------- - No poisoning symptoms even after 8 hours. Headache after 2-3 hours; collapse after hours. Headache after 1.5 hours; distinct poisoning after 2-3 hours; collapse after 3 hours. Distinct poisoning, frontal headache and nausea after 1-2 hours; collapse a_fter 2 hours; death after hours. Hallucinations felt after 30-120 minutes. Collapse after 1 hour; death after 2 hours. Difficulty in ambulation; death after 2 hours. Death in 1 hour. Death after ^5 minutes. Death after 30 minutes. Immediate death Unconsciousness after 2~3 breaths; death in 1-3 minutes. 1127053 BFG10359 -I.O. COHb (Ca r bcxyhem.ag 1 ob i n ) , * in the blood Sy.Tipt cms 20-30 30-^0 ^0-50 30-o0 . 60-70 70-80 80-30 ----------------------------------- ------------------ ----------------------------------- ------------------------------------ Headache, throbbing in the temples. Severe headache, weakness, dizziness, dimness of vision, nausea, vomiting, col 1 apse. As above, increase in pulse and breathing rates , greeter possibility of asph i xi at 1 c.n , col 1 apse. As above, cc-ma, intermittent convulsions and Cheyne-Stores respiration. Co_.a, intermittent convulsions, depressed heart action and respiratory rate, possible death. Weak pulse, slowing of respiration leading to death within a few hours. Death in less than an hour. Carbon Dioxide Concent ration (ppm) Symptoms 250-350 900-5000 18.000 25.000 AO.OOO 80.000 100.000 120.000 200.000 ---------------------------------------------------------------------------'------------- ------------------- ------------------------------------ ------------------ Normal concentration in air. No effect. Ventilation increased by 50?. Ventilation increased by 100?. Ventilation increased by 300?, headache, weakness. Dizziness, stupor, unconsciousness, distinct dyspnoea, lowered blood pressure, congestion, death within k hours. Headaches and dizziness. Immediate unconsciousness, death in minutes. Narcosis, immediate unconsciousness, death by suffocation. Hydrogen Chloride Concentration (ppm) 1-5 5-10 35 50-100 1,000 Sympt cms Limit of detection by odor. Mild irritation of mucous membranes. Irritation of throat on short exposure. Barely toierable. Danger of lung edema after short exposure. N O 01 lWl.pi mm ..if, 1 BFG10360 Hydrogen Cyanide (23 and 25) Concentration (ppm) Svmpt cos I8-36 ^5-5^ 100 110-135 135 1 Si 2SO ------------------- ------------------------------------------------------------------------------------------------------------- Slight symptoms, headache after several hours. Tolerated for 1/2-1 hour without difficulty. Fatal after 1 hour. Fatal after 1/2-1 hour. Fatal after 1/2 hour. Fatal after 10 minutes.' Immediately fatal. Nitrogen Dioxide (25) Concent rat i on (ppm) Sy~.pt ans 10-20 ------------------- 50 80 90 100-200 250 ------------------------------------------------------------------------------------------- Mildly irritating to eyes, nose and upper respiratory tract. Distinct irritation. Tightness of chest after 3"5 minutes_. Pulmonary edema after 30 minutes. : Very dangerous within 30-60 minutes. Death after a few minutes. Acrolein (CH2=CHCH0) (25) .' Concentration (ppm) Synpt oms 0.8 1.0 5.5 10+ ------------------- ------------------------------------------------------- Lachrymation, irritation of mucous membranes. Irritation Intense irritation. Lethal in a short time. Benzene (25) Concentration (ppm) 500 15DO-LDOO 8000 20000 Symptoms S1ight irritation. Dangerous to life after several Fatal after 30-60 minutes. Fatal after 5 minutes. hours. Parts per million (ppm) Is defined as the volume of the gas In one million parts air-vapor at 25C and 760 nm Hg. M as 0T OI BFG10361 -51- APPEND I X 2 Oxygen Index1 (Defined as the Percentage of Oxygen in an Oxygen-N' I t r og en Mixture Vh i ch Vi 1 1 Just Support Sustained Co~.b us t i on .) (1 0) Polyethylene, high density ASS, Cycolac T, (harbon) ABS/polycarbor.ate Alloy (Cycoloy SOO) 17-^ ' 18.2 20.2 PVC, Plaskon 2005 Chlorinated PVC, Geon 101 Polytetraf1uoroethylene (Teflon) 1*0.3 1*5-0 35-0 Various Woods 22 - 24.5 (1) More complete Indices are given by Isaacs (132) and by Hi 1 ado (25). Effect of Composition of PVC on the Oxygen Index Plastic PVC PVC with 95 ABS impact modifier PVC with 1*15 dioctyl phthalate PVC (505)/ABS (505) PVC (1*85)/ABS (4S5) with 45 Sb203 PVC/ABS alloy Cycovin KA Oxygen Index (lO) 1*0-3 35-3 21 .6 23-6 33-0 27-0 9SUZTT bFG 10362 APPENDIX 3 OPTICAL DEFINITIONS Optica1 Dens i ty Optical Decsity D = 1 og 1 c Optical density is directly proportional to the length of the light path. V *o S ---pecific optical dens--it--y, Ds = M--L logic --* f , where V= A= K= IQ = It = volume of smoLe chamber. area of burning or smouldering material length of the light path. initial light intensity. transmitted light intensity. Dr is dimensionless. Hass Optical Density, MOD = log ---e---- , where ---------------------------------------- *-ni t m = mass loss of sample up to T. T = minimum light transmittance (%) (^0). muiy;j H O rj BFG10363