Document By2ayoqeM4pvMOoK3q3rm7XwX

i Journal of Analytical Toxicology, Vol. 4, July/August t980 ^^1 &" Phosgene in the Thermal Decomposition Products of Poly (Vinyl Chloride): Generation, Detection and Measurement James E. Brown and Merritt M. Birky Center for Fire Research, United States Department of Commerce, National Bureau of Standards, Washington, D.C. 20234 iC work reported by Woolley (6J<and Tsuchiya et al (5) seems to Abstract lend support to Barrow's conclusion. However, Woolley's negative results regarding phosgene are limited by an An analytical study was mads to determine whether instrumental detectability of 50 ppni, a level that would be carbonyl chloride (phosgene) Is formed during the quite significant toxicologically (10). thermal decomposition of poly(vinyl chloride), PVC. On the other hand, Baltatis et al (11) reported the presence Four methods of decomposition were studied: (1) ther of phosgene in the combustion products of two PVC mal degradation of PVC in a resistiveiy heated lumace, composite materials which were burned in a closed container. (2) electrical overloading of a PVC clad wire, (3) electri These workers reported that burning 0.5 g of PVC from con cal arcing between electrodes partially covered with PVC, and (4) electric arc initiated flaming combustion veyor belts and cables gave respectively 0.0019 and 0.0135 per f in a cup furnace. Results are reported which show that significant quantities of phosgene can be generated cent of COCI2 in a 5 liter chamber. Earlier Coleman and Thomas (12) reported COCI2 concentrations up to 3 ppm and from PVC by the electric arc method. Lesser amounts to I ppm respectively for unstabilized PVC and stabilized were found In the other scenarios. The measurements, PVC composites burned at 900"C in a closed environment. identification and quantification of phosgene in Ihe In fatal fires involving PVC, various claims have been made decomposition product, were obtained through the use which attribute at least part of the toxicological hazard to of gas chromatography, infrared spectroscopy and mass spectroscopy. While Ihe study was not mecha nistic in nature, phosgene is postulated to result from secondary reactions of the PVC products. Reaction mechanisms are suggested. phosgene (13,14). The bases for the claims, however, are not well documented in those reports. Nevertheless, the reports do raise a fundamental question that the analytical measurements referred to above do not clearly resolve, either as a conse quence of instrumental limitation or the limitation imposed by the type of decomposition scenario chosen. Introduction Based on these uncertainties and the toxicological signifi cance of phosgene, a study was carried out in which analyses Vinyl chloride polymers arc frequently used in many appli for COCI2 were performed on decomposition products cations, perhaps the most familiar application is for electrical insulation. These polymers are also widely used in construc tion, in appliances, in furnishings and in transportation vehicles. Total production for 1978 is estimated to be 2.3 million metric tons (1). Buildings, including home furnishings and appliances, provide the major market for the production. The role of poly(vinyl chloride), PVC, in leading to human fife fatalities has been the subject of a number of investiga tions (2,3,4). In addition, the thermal degradation of PVC has been studied extensively from an analytical point of view (5,6) and various reviews of such studies have been prepared (7,8,9). Toxicological studies involving the exposure of obtained by four methods, each of which represents a potential real situation. The four methods are: (1) direct thermal degradation of PVC in a resistive furnace; (2) electrical overload of a copper wire wrapped with PVC insulation; (3) an electrical discharge between wires covered by PVC insulation; and (4) electric arc initiated flaming combustion in a cup furnace. The first method is equivalent to the analytical type of studies already referenced in the literature. The most significant difference between this study and other work is that more sensitive analytical techniques were used in this study. More importantly, the second and third methods were BFG39416 animals to the thermal degradation products of PVC have also been reported (2,3). Unanswered by these various studies is whether or not not previously represented and were chosen based on expected modes of degradation involving electrical overload and arcing conditions (e.g. reference (7)). carbonyl chloride* is a product of thermal degradation of PVC. A toxicological study by Barrow et al (2) suggests that the toxicity of PVC decomposition products is due principally to HC1, with little contribution from other components. The Carbonyl chloride whose formula is COCl^ has three well known synonyms: phosgene, chloroformyl chloride, and carbonic acid dichloride. In the follow ing text the synonym phosgene and the formula are used interchangeably to denote this compound. 166 Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission. 25237008 Thermal degradation products of polyvinyl chloride (PVC) food-wrap films were studied under simulated supermarket conditions using a commercial wrapping machine with either a hot wire or a cool rod cutting device. A sampling hood was constructed around the wire/rod to confine and allow collection of thermal degradation products produced. Compounds analyzed and normal concentration ranges found included hydrogen chloride (1 -10 fig per cut), plasticizer (1 -50 fig per cut), benzene and toluene (each <5-20 ng per cut), acrolein (25-150 ng per cut), and carbon monoxide (2-4 fig per cut) using the hot wire. Room air samples, collected during hot-wire cutting without the sampling hood, had <0.25 ppm hydrogen chloride. Using the cool-rod cutting device hydrogen chloride, benzene, and toluene were not detected. Plasticizer was detected (25-86 fig per cut) using the cool rod. Thermal degradation products from PVC film in food-wrapping operations* EDWARD A. BOETTNER and GWENDOLYN L. BALL The University of Michigan, School of Public Health, Department of Environmental and Industrial Health. Ann Arbor, Ml 48109 BFG39417 introduction Clear plasticized polyvinyl chloride (PVC) film is used extensively for packaging food products. The' film is thermally cut for each package from a long roll using either a 0.508-mm (0.020-inch) diameter hot wire maintained at 200350 C or a 9.53-mm (0.375-inch) diameter cool rod maintained at 135 0 C. The portion ofthe film in contact with the wire or rod during cutting may be subject to thermal degradation and/ or volatilization of film components. The thermal degradation products of unplasticized PVC have been thoroughly studied.*1* Between 240 and 310 C, PVC undergoes nearly quantitative dehydrochlorination making hydrogen chloride (HC1) the major lowtemperature thermal degradation product Above 350 C, oxidative reactions occur and carbon monoxide and carbon dioxide are the major products formed. Methane, benzene, and toluene are the major hydrocarbons produced, benzene in the same temperature range as HC1, and other hydrocarbons at somewhat higher temperatures. PVC films contain 20-30% plasticizer, those most commonly used being di-(2-ethylhexyl) adipate (DOA), di(2-ethylhexyl) phthalate (DOP), and acetal tributyl citrate (ATBC). These plasticizers generally have boiling points in excess of 300 C, but they are liquids having small but significant vapor pressures at room temperature, and would be expected to volatilize to some extent at both the hot wire and cool rod temperatures. Vandervort*2' studied the 200-230 C thermal degradation products of PVC wrapping films using a micro furnace with ceramic sample boats and quartz combustion tubes. The major products were found to be benzene, toluene, chlorobutene, l-chloro-2-ethyl hexane, 2-ethyl-1-hexanol, benzyl chloride, and hydrogen chloride. This experimental procedure discriminated in favor of benzyl chloride as it 'This research was supported by Borden Chemical. could decompose in the presence of hot metal, e.g. a hot cutting wire. In this same study the author found (race amounts of hydrogen chloride (well below the 5 ppm TLV*1*), but no detectable benzyl chloride, in field samples taken during the actual film cutting process. In a study conducted at the Liberty Mutual Research Center*4* a hot wire film cutting device was operated in a sealed chamber.while concentrations of hydrogen chloride, plasticizer, and total particulates were measured at different wire temperatures and at different heights above the wire. Hydrogen chloride again was found to be well below the TLV, and plasticizer concentrations did not seem sufficient to present a health hazard. A correlation existed between plasticizer and particulate concentrations produced. A Borden study*1* of supermarket breathing-zone HC1 levels using both hot wire and cool rod devices showed less than 1.5 ppm HC1 using the hot wire and less than 0.1 ppm HC1 using the cool rod. Substitution of the cool rod for the hot wire virtually eliminated smoke production during the cutting process. The present study was undertaken to tnfurther identify and quantitate major fume components produced by PVC films under simulated supermarket conditions using both the hot wire and the cool rod, co *\I '-'O description of samples O Five PVC film samples were provided by Borden Chemical and were plasticized as indicated in Table I. Three samples of RMF-61HY were used during the course of the study. Solvent extracts of each film were prepared and analyzed by gas chromatography in order to detect the presence of major film additives. In addition to the plasticizers listed in Table I, films CW-65 and PC-63 produced an additional chromatographic peak which was identified by mass spectrometry as tributyl aconttate. Tributyl aconitate may be an impurity in commercial acetyl tributyl citrate. American Industrial Hygiene Association JOURNAL Copyright 1980. American industrial Hygian* Association (41) 7/80 513 04& Polyvinyl Chloride Film Thermal Decomposition Products as an Occupational Illness I. Environmental Exposures and Toxicology Robert Vandervort and Stuart M. Brooks, M.D. o ver the past several decades a variety of packaging materials and techniques have been utilized to wrap fresh meat With the Wrapping and Labeling Equipment/Hand Wrapping Procedure advent of refrigerated, self-service meat display cases came the A wide variety of hand operated, semi-automatic, and popularity of clear plastic packaging films. Among film materials automatic meat wrapping and labeling equipment are in use. which have been widely used are cellophane, rubber Semi-automatic and hand wrapping machines utilize hot or heated hydrochloride, polyethylene polystyrene pdyvinylidene chloride, wire film cut-off. Price labels are printed, heated to activate the and polyvinyl chloride.1 Since its introduction in 1963-64, hot-melt adhesive, and then either manually or mechanically ap polyvinyl chloride (PVQ has grown to be the predominant fresh plied to the wrapped package of meat meat packaging material in use today. Usage estimates vary con Although some of the newer- wrapping equipment provides a siderably, but approximately 60-90 million pounds of PVC film are means to control sealing temperature and cut-off wire tempera used in the wrapping of meat each year. Another 30 to 35 million ture, most of the older equipment lacks one or both of these pounds are used to wrap produce and other products.1 features. Hand wrapping equipment is durable, relatively un Working Environment complicated, and can be expected to give several years of service. Consequently, many vintage machines are still in use today. A Meat wrapping is usually performed in the meat departments of retail supermarkets. According to union and industry estimates, there are between 75,000 and 100.000 persons employed as meat "typical" hand wrapping machine is shown in Figure 1. The hand wrapping of meat involves: (1) pulling out a desired length of film; (2) wrapping the film around the tray or cut of JO C/T 0 wrappers in the United States.1 The large majority of these meat wrappers work in the nation's 40,000 to 45,000 supermarkets, with the remainder employed by butchershops and packing houses.* Meat wrapping end labeling stations are typically located inside the mildly refrigerated meat departments. Minimal ventilation is meat; (3) severing the film from the supply roll using the hot-wire cut-off; (4) folding the film ends under the package; and (5) sealing the folded ends under the package by touching the package to the heated sealing pad. Most machines apply a mechanical tension to the film so that it retreats a few millimeters from the hot wire after CO "'T O O Qu supplied to these work areas to reduce refrigeration requirements cutting. This tension is intended to prevent the film from and transportation of contaminants from other sections of the smoldering on the wire and causing it to become dirty and smoke. establishment into the meat cutting and wrapping area. This tension also makes the film somewhat more difficult to pull from the supply roll and is often found to be improperly adjusted t From OB AssociitM. Inc, Salt take City (Mr. Vandervort, formed/ Industrial Hygienist NlOSHI and the Department of Environmental Health, University of Cin in the field. The sealing pad is maintained at a temperature suffi cinnati College of Medicine, Cinc'nnati, Ohio 45267 Of. Brooks). cient to cause the film layers to adhere to one another after brief 188 BFG39418 Polyvinyl Chloride Film Thermal Decomposition Products as an Occupational Illness 2. Clinical Studies Stuart M. Brooks, M.D. and Robert Vandervort In 1973, Sokol et al reported three cases of workers employed as meat wrappers who developed respiratory symptoms when ex posed to fumes of polyvinyl chloride (PVO film cut-with a hot wire.1 The patients were middle-aged women who smoked cigarettes and demonstrated reversible airway obstruction on pulmonary function testing. The patients were not studied at work, but reported to a hospital where examinations and laboratory studies were performed. In an epidemiologic study conducted in 152 supermarkets in Houston, Tex., approximately 10% of the meat wrappers studied reported multiple respiratory symptoms at work which included wheezing shortness of breath, and chest pain or tightness.1 A significantly greater prevalence of symptoms suggesting eye. nose and throat irritation was also found. Andrasch et al conducted a questionnaire survey of 165 meat wrappers in the Portland, Ore. metropolitan area.1 Fiftyseven percent of the symptomatic workers reported respiratory symptoms and 30% reported symptoms suggestive of nose and throat irritation. Several workers reported the occurrence of acute asthmatic attacks while at work. Bronchial inhalation provocation studies with emissions from hot-wire cutting of PVC films con ducted on 11 selected meat wrappers demonstrated rhinorrhea in seven wrappers, cough, tightness of the chest and sore throat in five, dyspnea, exhaustion and wheezing in four, and throat soreness in two subjects. These studies were conducted in a laboratory away from work.Nine of 13 patients who were then challenged with emissions from heating price label material developed immediate acute asthmatic reactions. Four of these subjects also experienced a delayed asthmatic reaction which oc curred six to eight hours after exposure. The authors concluded that the emissions from thermally activating price labels were the principal cause of "meat wrappers asthma," but that the entire spectrum of this disease had to be interpreted as a complex From (he Department of environmental Health. Unhrenrty of Cincinnati. Cincinnati. Ohio 45267 (Dr. Brooksl and D6 Associate*. Inc, Salt Lake City (Mr. Vandervort for* merty industrial Hygienist N!OSH). 192 response from exposure to both emissions from hot-wire cutting of PVC film and from thermal activation of price labels. Other in vestigations of respiratory illness in meat wrappers have also been reported.1"* The 8ureau of Occupational Safety & Health received com plaints associated with exposure to air contaminants generated by the hot wire cutting of PVC meat packaging film as early as the summer of 1969. Beginning in 1972, The National Institute for Oc cupational Safety & Health conducted clinical studies of meat wrappers in response to requests for Health Hazard Evaluations in supermarket meat departments. The present report contains the results of these clinical studies. Investigation of many of these cases was in collaboration with the Department of Environmental Health, University of Cincinnati College of Medicine. Materials and Method Twenty-four meat wrappers, eight meat cutters and 20 control subjects (office personnel and store clerks) make up the population studied. Data concerning age, sex and smoking history of the subjects are shown in Table 1. While meat wrappers were commonly women, meat cutters were usually men. The three groups were of similar age, but there was a greater number of meat wrappers who were cigarette smokers. The mean duration of wrapping in the meat wrappers was 12.5 years (range 1 to 26 . years). Investigations were performed at or near the workplace in each case. The meat cutters worked in the same general area as the meat wrappers. Each subject was administered a modified respiratory questionnaire which included questions on past medical history, cigarette smoking, previous occupational ex posures, current respiratory status and adverse effects allegedly arising from work Pulmonary function testing was performed on each of the 53 workers at the beginning and end of the work shift. Thirty-seven subjects were tested using a waterless, high fidelity, precaiibrated spirometer equipped with an air temperature probe (Ohio Medical BFG39419 25237009 Vinyl Chloride Formation from the Thermal Degradation of Poly(Vinyl Chloride) I. B. WAKEMAN and H. R. JOHNSON Tcnncco Chemicals, Inc. Corporate Research and Development Department Piscataway, New Jersey 08854 The volatile products from the thermal degradation of poly(vinyl chloride) (PVC) resins and compounds are shown to contain trace amounts of vinyl chloride. Data presented show the effect of temperature and resin type on the amount ofvinyl chloride formed. At the maximum temperatures involved in PVC processing which may reach 210*C, vinyl chloride monomer (VCM) evolution amounts to less than 1 ppm (resin basis). A technique employing a thermogravimetric balance and charcoal adsorption of volatiles is described for studying thermal degradation of PVC. The volatiles are analyzed for vinyl chloride by gas chromatography. Peak identity was confirmed by mass spectrometry. 25237010 INTRODUCTION Numerous studies and surveys have been published describing the thermal degradation and pyrolysis products of po!y(vinyl chloride) (PVC). These studies have shown that at elevated temperatures essentially quantitative dehydrochlorination of PVC occurs, result ing in the formation ofa complex mixture ofaromatic and aliphatic hydrocarlxms with benzene being the largest' single component (1-8). Stromberg, ctal. (1) studied the decomposition of PVC under vacuum at temperatures up to 400C and measured the variations in the decom position products as a function of temperature. They did not indicate the formation of any vinyl chloride monomer (VCM) under these conditions. Boettner, et al. (2) investigated the thermal degradation of PVC in air up to 600C\ Using infrared spcctroscop.y and gas chromatographv-mass spectroscopy, they identified ap proximately 50 volatile degradation pr<ducts and mea sured the rate of formation of the major products as a function of temperature. Their data indicate that up to 6(X) ppm of VCM is formed from PVC homopolymer and up to 3000 ppm from PVC compounds. Although no mention is made of residual VCM in the polymer, the bulk of the VCM was detected in the 280-430C range and it is inferred that VCM is produced during the thermal decomposition. Woollev (3) reported low levels of VCNl in PVC pyrolysis products but did not measure residual VCM, and it is not dear whether any VCM was actually formed during the decomposition. Lewis (4) has recently reported degradation studies ofa homopolymer having a residual VCM content of < 0.5 ppm and found a maximum VCM formation of 35 ppm at 350C. The recent findings (9) of carcinogenicity on pro- trations of vinyl chloride in the air and of cases of angiosarcoma in vinyl chloride workers have caused seri ous concern about the effects on workers involved in all phases of PVC manufacture and fabrication. Although present day manufacturing processes have reduced the residual VCM content of PVC resins and compounds to very low levels (10), the possibility of forming VCM by degradation of the polymer during calendering, extrusion, and thermoforming operations is still a cause of concern. There is very little likelihood of VCM formation dur ing thermoforming of PVC for food packaging applica tion since this operation is carried out under mild conditions(90-120C for a few seconds). It has been demonstrated previously that no detectable VCM is formed in up to 1 h exposure at 130C, using an analyti cal method capable of detecting 1 part per billion of VCM (10). During calendering and extrusion operations, how ever, temperatures may reach the 175-210C range for brief time periods. We know of no published informa tion involving VCM formation under these conditions. Our present work involved the investigation of the thermal degradation of PVC over the 130-500C tem perature range with special emphasis on VCM formation at 210C, the upper limit used in PVC fabrication. , , | 1 . { EXPERIMENTAL Thermal degradation studies were carried out using a duPont Model 950 .thermal balance which provided a convenient method for accurately controlling the tem perature and heating rates. .Samples of 80-100 mg were placed in a platinum foil boat suspended on the thermal balance sample holder. The furnace tube surrounding BFG39420 364 Tsuchiya & Sumi: Thermal Decomposition Products of Polyvinyl Chloride THERMAL DECOMPOSITION PRODUCTS OF the cat POLYVINYL CHLORIDE /^T) mil siti By YOSHIO TSUCHIYA and KIKUO SUM! ^ dia the qu; the When plastics are involved in a fire they may yield toxic decomposition products. Some quantitative data on the decomposition products of plastics are available in the literature, but it is difficult to assess the danger from the different amounts of various products because of the absence of a suitable method of evaluation. The authors have proposed a method of evaluation based on pyrolysis followed by gas chromatographic analysis and have used it to assess toxicity from various thermal decomposition products of polyvinyl chloride. Hydrogen chloride was found to be the main toxic decomposition product. poi J (hi, col for ten 10' flo- gra Introduction The increasing use of plastics and other organic polymers raises the possibility that when involved in a fire they may yield toxic decomposition products in quantities sufficient to produce a dangerous atmosphere. Some quantitative results on the decomposition products of plastics are available in the literature, but it is difficult to compare data because of differences in experimental methods and differences in presentation of experimental data. It is also difficult to assess the danger from the different amounts of various decomposition products because of the absence of a suitable method of evaluation. In view of the above factors, the authors bei-eve that a need exists for a systematic study to provide quantitative results on the volatile decomposition products of a wide variety of plastics instead of the static system used by other investigators,1-1 because the decomposition conditions during an experiment can be more closely defined in a flow system than in a static system. Material Experimental The PVC used in this study was a commercially available, general purpose resin in powder form. It did not contain any plasticiser. Thermal decomposition PVC was decomposed in a furnace through which inert gas or air was passed and the gaseous products were collected in a flask. The sample was weighed in a ceramic boat, to which a stainless steel wire and a piece of iron were attached, so that cot tio the Fo the *P to of i the hy< the det me chi sut in both inert and oxidising atmospheres. A need also exists for a method of assessing danger from the different quantities it could be moved inside a 19-mm diameter tube to the hot zone of the furnace by means of an external magnet. The mi 1 of decomposition products. This investigation was carried furnace temperature was maintained at isothermal values of col out to meet these needs. 350, 600 or 850'. Helium was used to provide an inert tec Polyvinyl chloride (PVC) is the plastics material that has atmosphere in one series of experiments and air was used to probably received the most attention from fire authorities provide an oxidising atmosphere in the other series. After because of its relatively wide use and the possibility of its some preliminary studies a gas flow rate of 450 cm1 per min thermal decomposition leading to the formation of toxic gaseous products. This polymer was. therefore, selected for was adopted. This corresponded to a mean linear velocity of about 160 cm per min. The sample weights selected for de< in the present study. this study were 0-5 g in helium atmosphere and 0-5 and 0-25 we Coleman & Thomas'- determined the combustion products g in air. The amount of air used during each experiment or of PVC and other chlorinated elastics using a static system. was slightly in excess of that required for complete com The specimens were decomposed in a flask over a temperature bustion of 0-5 g of PVC to carbon dioxide, water and hydro ch range of 300 to 1000'. The ratio of plastics to air was varied gen chloride. The smaller weight of 0-25g was also used in an hy by using different weights of sample, and the combustion products were analysed by conventional methods. Schries- effort to examine the influence of effectively increasing air supply on the formation of decomposition products. du heim1 employed a similar method over a temperature range The higher boiling components of the decomposition of 250 to 550' and analysed the products of combustion by a products were collected in a U-tube filled with glass beads and mass spectrometer and chemical methods. Stromberg et at.1 maintained at O'. The unreacted gas and the remainder of studied the mechanism of thermal decomposition of PVC in the volatile decomposition products were collected in a 3- vacuum. The volatile products were analysed by mass spectro metry and found to consist airr.os: entirely of hydrogen chloride and smalt proportions of benzene, toluene and other litre flask that had been evacuated prior to the experiment. The sample was kept in the furnace until the gases filled the Cw At flask. Each experiment lasted about 8 min. w- hydrocarbons. Gilbert & Kipling4 investigated the carbonisa tion of PVC and other vinyl polymers by decomposing the specimens, under vacuum, in an inert atmosphere and in air. After removal of hydrogen chloride the residue was de composed and the products were analysed by gas chromato graphy. In the present investigation the thermal decomposition products of PVC were determined, firstly, in an inert atmos phere and, secondly, in air. A flow system was adopted Analysis Gas chromatography was the main method used in the analysis of decomposition products. As these were composed of widely differing materials, three chromatographic con ditions had to be employed (two with a thermistor-type thermal conductivity cell detector, with different columns). First a molecular sieve 5A (60- to 80-mesh) column of length 2 m and diameter i in. (0-6 cm) was used with the 3-i j. J. appl. Chem., 1967, Vol. 17, December Ct BFG39421 fwiri'iriii mi,, i > i ^ Building Research Establishment Curront Paper crae T**wi , Yn . (R8) CP 5/74 January 1974 Smoke and toxic gases from burning plastics G W V Stark Fire Research Station ROBERT I. MEYER r. GOCCCIC?. Cl - - OK 'ii.CV-.-C--- BUIIOING RESEARCH ESTABLISHMENT Departmont of the Environment A cn w co o N BFG39422 measurement of j8gSSnS8Sra COMPOUNDS Hardy Sze On Chan Department of Chemistry National University ofSingapore Kent Ridge Singapore 0511 (Received October 31,1983) (Revised May 1.1984) EMISSI0N ABSTRACT The development of a novel Mlcro/TG method for measuring HC1 emission from PVC compounds and optimization of operation conditions are described. The main problem with conventional large scale methods is the loss of HC1 due to leaks and condensation in the system leading to erroneous results. The Micro/TG method overcomes many of the problems inherent in the large scale method. Comparison with results obtained by the large scale method showed that the total balance of chlorine content determined by the Micro/TG method is closer to the theoretical in all of the formulations tested. Any chlorine in the form of organic-chloride (RC1) will not be detected by the Micro/TG method, a limitation shared by the large scale method. The Micro/TG method offers other advantages such as automation, short experimental time and direct measure ment of chloride ion concentration by ion selective electrode. INTRODUCTION Plasticised pvc is widely used for the insulation and sheathing of cables because of its low flammability (when used with flame retardant plasticisers), satisfactory electrical properties and low cost [1,2]. The emission of large quantities of hydrochloric acid (HC1) when it burns, however, is an unacceptable disadvantage for some ap plications. Even if the PVC compound does not actually ignite in a fire, it may still be heated sufficiently to evolve HC1. Quite small amounts of 106 JOURNALOF FIRE SCIENCES, VOL. 2-MARCH/APRIL 1984 0734-9041/84/02 0106--17 304.50/0 1984 Technomic Publishing Co., Inc. HydrochloricA HCI released a electrical and < creases the o\ which react wi methods for n methods have laboratory met using a thermo Current Metho While the ge cable compoun proposed by LI cepted, some measurement o turers. The I.E. compounds (0.5 furnace either a at 800 C for a quartz tube by. collecting bottle tration is deterr The actual expe differ mainly in method for mea large scale meth sources of errors 1. Loss of volat: 2. Condensation 3. Carry-over of 4. The formatior may give spu: Because of tl oratories is repoi emission results measuring HCI ducibility. 10 (^EXPERIMENTS ^Materials Used QBreon S125/14 pX)P CJ BFG39423 'I8 Combustion Gases Generating from Polyvinyl Chloride and Its Products ---- Model Experiment in Case of Fire----- By Yoshitaka Kobayashi*), Masahiro Hori*) and Hisako Murata*) `A.** 1. Introduction IN building fires, death has been caused in many cases recently by generating gases rather than by the fire itself. The increase of disasters resulting from such generating gases depends to some extent on building construction but to a . greater extent on the change of composition of generating gases. The change of gas composition has been brought about by the fact that building materials have been changing from natural materials such as wood, etc. to synthetic high polymer materials. With the development of high polymer materials in recent days, a variety of plastics or synthetic fibers has been utilized for interior materials or furniture, with many of these materials having thermal decomposing properties or burning properties differing from those of natural high polymers1'. It is considered from the differ ence of the gas composition, there fore, that many problems are in volved in application in considera tion of fire prevention or escape. Among synthetic high polymer materials, polyvinyl chloride (PVC) is one of those which are most commonly utilized. Building ma terials made chiefly of the resin are floor tiles, roofing, walling, gutter-, various piping, partitions, accordion curtains, chair upholstery, etc., and more recently, PVC is used for a *) Engineering Faculty of Yoko hama National University diversified range of application in cluding large building exteriors and interiors as well as for wooden houses by laminating onto metal, wooden and other plastics bases. Some studies were conducted on the thermal decomposition of PVC in vacuum and air at an elevated temperature1*41, but in this experi ment, investigation was especially made on various processed PVC products recently marketed in Japan. The method of elevating tem perature was adopted according to the fire resistance test (JIS 1304) for building construction elements, and combustion gaseq generating at each temperature level were determined. Comparative study was also made for natural high polymer materials such as wood, paper, etc., and PVC products. Be sides, stress was especially placed on generation of minute toxic gases when burning conditions such as air supply, etc. were changed. 2. Method of Experiment 2-1. Materials Tatted and Device Materials tested were PVC raw material and 8 kinds of PVC prod ucts presented in Table 1. Among PVC building materials are in cluded flexible and rigid products in plate or film form, and the ma terial.: '.-tod can be clarified into 4 according to the quantity of .-,tahi|i/.'.T added. lit addition, natural high polymer materials such as wood, etc. and other plastics materials such as melamine resin, etc. were selected 40 and a comparison carried out. The equipment shown in Fig. I was used as a combustion gas gener ating device. Heating is provided by use of an electric furnace, and the method of elevating tempera ture was adopted according to the fire resistance test for building con struction parts. (JIS 1304) (ca. 100`C/min.). Readers are recommended to refer to Fig. 2. Specimens were heated in a magnetic boat placed in a quartz tube having a diameter of 2.5 cm and a length of 30 cm in the elec tric furnace. Supplying air at a constant flow speed by means of a compressor and nitrogen bombe, the specimens were burned and generating gases gathered. The electric furnace is preheated, then 1 gram of specimen is fed into the burning tube, and the air supply quantity is adjusted to a constant flow speed. Heating is immediately conducted, and part of the ex hausted gases including excess air and generating combustion gases is gathered in a glass cylinder at 3fl0j 400, 500, 600 and 700" C for rai2 manalysis. The temperature inside the nace is determined by a thern*$ couple attached to the magne boat. As for PVC raw material, chaff? ing the air flow speed from 1 l/min. to 1.5 l/min., and also changing the oxygen ratio in air supplied from 21 to 10.5 per cent, the influence on the generated gas composition was studied. Experiments were re- JAPAN PLASTICS BFG39424 l>y so Slump. II 11;> i*'* .......... V` "`'i ofli*'st;in ,,r (*,c design: pccliicadnn and viii. ' .anccs of successful pssfnnmmcc car. be improved. Ackficivicftyoincfit The auCiur wishes it. thank Mr IM. Vincent (or his litlp and com* mil,Is (Iuriiu: discussion of liiis paper. Hclotonces i Ointhnir.. K-V. /'it'.ics A 1'Mvmcrs 40 (1972) 59. . Gotham. K.V. IHd-X0 (107 ') .177. Vincent, IM, Tnijocl le:,ts nit:! jrrvice performance of Ihermopt.irt'cs' ''fustics Institute Monof/fpli J971. t Kennedy, A.d. 'i'.oav.cs of creep mui fatigue in mclals* (Oliver and Jioytl, 1VM). .j A ; >. ,kiA V.-r. I C h;io 10 (f<i `.ii-l Wii'an.l I'ut-.i-.lung ( r.`., Ar,;i.tcf.! it.? m;v! * I (ill, II. T!.iM(< i<v * l. r -iKior. I'kv., Oxford, J930;. !*cOM\nn. IU-.. `Mirv. cijnr cn<raion or'.itrn factor/ (Joint Wi!^y So***. Inc, I lew Y<rV, K.V, un<l turner, i'oiytnrr Knf. Set. 13(1971) MV ' Itoor.-mirt, \. uin den, and 'liirncr, CM.. Traiit. J. riauiti hut 7! (I`>(i.tj 109. I peris, r.Citml Silt, O.C. 'Stress analysis of cracks: fr.t. i-.irr an(j toutdii.css If ;imp and its upphratigns' AS l'M Sficcial iccimicAl Publication No. .'SI, i' Mai !i;.ll, G.P., CUrer, L.F.. end Williams, J.C, Plastics A Polymers J7 (1969) 75. 11 Vii;rr'!(, IM. plastic rlcfor.nrliou ond fracture' Tit 97 nivriim of Polymer Science, Case Western Reserve University, Cleveland, Ohio. w Wats, N.H. hiv5 Iiurr-i, D. J. Polymer !>?. Set. 7 (I'.'67) 90. ^4 r...Toxic products from plastics.materials m Tires VV. D. Woolley c sc, ph D, m l fire e i.Fire Research Station, Uulldiag Restart h Establishment, Department ofthe Environment and Fire Offices', . CommitteeJoint Fire F.cseardiOrgcnlearian , Abstract: Over one-halfot the tataliues in fires ore caused by smoke and toxic gases. Week is in progress ; .. ot the Fi'o Research Sutlnn to study the production of toxic gases when plastics materials arts decompos- ' cd under controlled laboratory conditions involving pyrolysis at temperoturas up to 1Q00"C and in full ^ ecola fires using a special compartment with a variable ventilation slit (doer) hading to a corridor. When healed above about UlO'C. PVC releases tho toxic get hydrogen chloride. At 300'C this dehydro .. 'V. chlorination is rapid and quantitative. In fires involving wood and PVC fuels the main toxic products .' . ; -. ero cerbon monoxide mic hydrogen chloride end the contribution of hydrogen chloride to tho over-ill ... ` .toxicity depends upon the fuel loadings end ventilation. Pyrolysis of flexible polyurethane foam at ebout 300'C releases a polymeric 'yellow srr.cho' which decomposes at tempeuturss above COO'C in nitrogen (t>03*C in cir) to giro e range of nitrogen-cent doing compounds, porticuisrly hydrogen cyanide. Fires. .. Involving industrial iostls of from show temperatures in excess of 1000'C (sufficiently high tc ;. ; .decompose Om yellow smoke) ere etrshred within ; fosv minutes. Outing the early stages oi the tiro tho - toxic fretard ficm hydrogen cyanide c:n approach tho samo order of magnitude as that from the carbon monoxide. ' ; / .. . ;V.\ :I. ;; v : , 1 Introduction .; .. -A,. ,V- Last year in the United Kingdom approximately 1000 ; -persons died ns a direct usuU of fire. Evidence indicates' that over onc-half of thie fatalities ware caused by the , combined effects of imoicc and toxic gases and that the .remainder were-due to hums and other injuries^ In : addition, many' thousand:, of perrons were involved in 'potentially Imrardous fire situatisms. ,, The cflccls of r-mohe and toxic r ases can be considered '. separately. The unin effect of sir,eke is that it reduces visibility and iirurc hinders e-;c?p". whereas the effect of toxic gases, con pice \vtJ osygc/. depletion- of the atmos.. pli.-.re and Iie.u ;xf>o:.urc, is to si:!. Certain toxic gases at . t;ub-!cihal levels can directly hinder escape by producing severe eye and inhalation irritancy. Other toxic gases that arc odourless kill rapidly without warning. * ' This paper will outiinc the current v/orl; being carried out at the Fire Research Station to study the r ilcrne of toxic products when building anu iui .fishing iiir, Lai -.Is are involved in fires. As examples the paper will report the results of two well known and topical materia Is, r.nmcly polyvinyl chlori<ic(i'VQ and poly.iruh.mc foam. Essentially, the research is divided i.ito two main areas involving, first,'* small-scale labouriory dreereposition experiments under simulated lire condition' and, secondly, full-scale .lire tCj.S. It is teieV.iili ni imi i' -a (to discuss the experimental facilities and proccdr. s end then give a broau survey of liic results. i -< 2 Exporiniontnl procedures 2.1 glory experiments i -. Or W. O. V/ooiloy ii a frinf lpul ocionlilic Oli'cer i,t liw Pi;;j r.etusfcli Million. L!o'ur.::.n Wooil. (!t-il.e He In the laborntoiy dudies, small s.-nniiics of in.ua'ial arc crndti'itod t.i chcnistry 11 Si A:.v in 1901 r.i.d apont decomposed l\ a linns.ee. r.y:,t.;m under the coruifons iwo yii.".ns iviih tho (;,-:<i,i:i;t -t u;-,u-ii Gcunei! of likely to bcr.i-rorniic.-cd in fir.\. ic temprraturr up i.*and (''c-.ndi. i.iujyiii,; |.liC U'C:..:i Set i 1.1. en-Ji.a. iu i-v.crs-., of iIa if C in tR'.ironiiv:nf; t anging fiom :>: to a fi.rious deficivucy cf oxype... r lii:, is in fact a stiu';.1 c! the 7 SO r lj~ - ... I'sm, !> ! i. fl r , iro i;*. I'iiAr .-fl/.IN BFG39425 # growing industrial challenge of compositions of neoprene and Wypalon via the Oxygen Index Method indicate that both polymers have a high degree of flame resistance. Pyrolysis and combustion lests show that hydrogen chloride evolution from these materials can be minimized by compounding. BY CHARLES E. MCCORMACK ELASTOMER CHEMICALS DEPT., E. I. DU PONT DE NEMOURS AND CO., INC., WILMINGTON, DEL. r mill ixisting Is and service erer ublic concern regarding atmospheric pollution and Pgovernmental regulations aimed at consumer pro: :onion have caused the spotlight to be focused on ' *hat happens to materials when they are exposed to conditions of pyrolysis and combustion. There is evi dence that industry is aware of the challenges and is in I :he process of dealing with them. ! In considering matters such as pyrolysis, combus- j -ion and flame resistance, neoprene, the first general; Purpose synthetic elastomer commercially introduced i 1932, merits special attention. Right from the start, ; was recognized as flame-resistant. Of course, if it ud only flame resistance to recommend it, it would "t have attained the hundreds of millions of lbs/yr. ;utus it enjoys today. It is noteworthy, though, that ctme resistance is the key property in many applica:,,ns such as wire and cable (mining cable, telephone 'r"p wire, appliance cord), conveyor belting and me-resistant latex foam. In each of these applica"ns. other properties such as physical strength, oil 'd ozone resistance, have played strong supporting Ics. I he results of some recent studies on the flame -''stance of compositions of neoprene and Hypalon '-judged by the Oxygen Index Method, and on the products of their pyrolysis and combustion with par ticular emphasis on the control of hydrogen chloride evolution are covered in this paper. Oxygen Index Flammability Ratings The higher the oxygen concentration required to support the combustion of a material, the better its flame resistance. This is the basis for oxygen index flammability ratings, which are numerically equal to the minimum volume percentage of oxygen in a slowly rising mixture of oxygen and nitrogen which will sup port candlc-Iikc burning. The test method, as described in ASTM Method D-2863. fills a long felt need for a relatively simple, reproducible test for the characterization of materials. Of particular appeal are its applicability to all types of polymeric materials, regardless of whether or not they normally support combustion in air, and its simple numerical index which permits a ready comparison of materials. Oxygen index levels easily attainable with neoprene arc shown in Table I. All of the values are well above the index level of 27 considered to indicate sell-extin guishing characteristics. No attempt is made here to isolate the effect of various compounding ingredients on oxygen index. It is evident, however, that com- ' j8BER AGE, JUNE, 1972 BFG39426 27 By Marts Meisters* Concepts and trends in polymer fire testing This report provides a framework for understanding the principles that are important in correlating small- and large-scale (ires by describing a selected series of small-scale tests In terms of such fire concepts as ignition, propagation, heat release, and smoke generation. All of the methods have limitations and cannot completely evaluate fire performance, but they offer a baseline for continued development of tests to provide better correlations with real-fire performance. A technical feature Gordon M. Kline, technical editor Plastics are replacing wood and other "natural" materials in many consumer products. Ease of fabrication is the overriding economic factor and guarantees continued mar ket penetration. A question is posed by the shift from nat ural to synthetic materials: Do plastics increase potential fire hazards in the built environment? Answers can be ob tained if realistic fire performance tests are available to compare new and existing materials. Fire performance can be defined as the behavior of a material or product in a real-life fire situation. Many prob lems are encountered in evaluating fire performance. Per haps the most perplexing difficulty results from the fact that no two fires are exactly the same in terms of such pa rameters as ventilation, proximity of burning objects, and enclosure size. No small-scale test can completely evaluate fire per formance. This goal can be sought, however, by determin ing the effects of some of the more important variables on an individual basis. The effect of chemical structure can be understood more thoroughly by thermogravimetiy, a semimicro technique that measures weight loss as a function of temperature. The method can be used to rank thermo-oxidative sta bilities of materials. However, commercial products such as tables and chairs have burning characteristics that are not completely dependent on chemical structure. Endproduct geometry and orientation can have major effects on fire performance. Many small-scale tests are in use. In the past, results were often described by such terms as "flame spread" or "flammability index." The ability to predict actual fire per formance was sometimes implied. Experience has shown that this approach generally is not valid. For example, ASTM D-1692, a standard test method for rate of burning or extent of burning of cellular plastics, specifies a hori zontal sample orientation producing results vastly differ ent from those obtained using a vertical orientation with 'Research Associate, Center for Fire Research. National Bureau of Stan dards. Washington, D.C Dr. Meialen' participation ia die SPi-supported research program on Are control of interior furnishings is sponsored, by Cetanese Plastics Co. and Cclancse Research Co. bottom ignition. This has been illustrated in field installa tions utilizing polyurethane foams, which bum rapidly in a vertical configuration. Theoretical considerations For most solid materials, the combustion cycle consists of five interrelated steps: 1) heatup, 2) dissociation, 3) vola tilization, 4) ignition, and 3) propagation. The first two steps involve the action of heat and possi bly oxygen to dissociate the polymer into fragments, some volatile and combustible. The role of oxygen in polymer decomposition is still a point of controversy, but cannot be ruled out, especially in the case of polymers highly suscep tible to oxidative degradation. Combustible gases in the correct proportion with oxygen are ignited if sufficient heat or outside flame source is available to raise the gas mixture to its ignition temperature. If ignition leads to a steady state process, the resulting flame will feed heat back to undecomposed material, re peating the cycle. Critical points in the propagation se quence are maintenance of a flammable gas mixture com position and exothermic heat release sufficient to continue polymer degradation and pyrolysis. In defining chemical factors affecting polymer fire per formance, heat of combustion offers a reasonable starting point Heat of combustion is related to bond strengths, and can be determined by oxygen bomb calorimetry. In this context it is the total heat energy released when a material is ignited and burned in a pure oxygen atmosphere. An ex amination of heat-of-combustion data, however, indicates that heat of combustion is not sufficient to rank materials in terms of potential fire performance. Based on heat-ofcombustion data (Table I), and assuming equivalent bum- Flfl. 1: Flame spread. 25237017 i i 76 BFG39427 Modern Plastics, September 1975 23 June 1978, Volume 200, Number 4348 JLJa Qcr> ----------- ----- ii-t '-VJKT Return To. R. K. Hinderer D/5456 Cleveland in fires. These and other fires resulted in more detailed fire investigations that, to gether with laboratory experiments with animals, made it possible to begin to identify the lethal factors in fires. Toxic Gases from Fires Epidemiologic Study James B. Terrill, Ruth R. Montgomery, Charles F. Reinhardt According to ancient Greek mytholo gy, Prometheus suffered extreme torture for giving men heavenly fire. However, without fire it is difficult to envision how humans could have advanced beyond the caves. As people crowded together in cities, the occasional, unwanted fire problem escalated drastically. Some dra matic examples include the Great Fire of London in 1666, the Chicago Fire in 1871, and the destructions of Tokyo, Dresden, and Hamburg during World War II. Frequent fires in American cities certed efforts of materials suppliers, fire fighting officials, fire code groups, and the insurance industry. In the United States today, fire exacts an annual toll of some 8000 to 9000 deaths and property losses amounting to some $3 billion to $4 billion (2). The sur vivors may also experience severe an guish. Approximately 70 percent of the 1976 fire deaths occurred in residential fires. To reduce this fire fatality rate, it will be necessary to assess the potential life hazard from dwelling fires or fires in general. The factors leading to fire tox Summary. The major lethal factors in uncontrolled fires are toxic gases, heat, and oxygen deficiency. The predominant toxic gas is carbon monoxide, which is readily generated from the combustion of wood and other ceilulosic materials. Increasing use of a variety of synthetic polymers has stimulated interest in screening tests to evaluate the toxicity of polymeric materials when thermally decomposed. As yet, this country tacks a standardized fire toxicity test protocol. stimulated efforts to test the fire perform ance of materials to minimize potential fire damage. The American Society for Testing and Materials (ASTMl accord ingly devised the first scaled test proce dures: door fire resistance {ASTM E-152 (1940)] and wall fire resistance [ASTM E-l 19 (1917)]. ignition resistance [ASTM D-1692 (1959)], and flame spread (ASTM E-84 (1950); E-162 (I960)] (/). These pro cedures were devised through the con- Dr. Terrill is group leader of the Inhalation Sec tion. Department of Medicine and Environmental Health. Monsanto Company. St. Louis. Missouri 63166. Ms. Montgomery is an information specialist a( the Haskell Laboratory for Toxicology ami Indus trial Medicine. E. t. du Pont dc Nemours and Com pany, Wilmington. Delaware 19898. Dr. Reinhardt is director of the Haskell Laboratory. Dr. Terrill was formerly at the Haskell Laboratory. icity must be examined. This knowledge must then be applied toward developing reliable test procedures that will make it possible to evaluate the potential of a combustible material to create an ex traordinary toxic gas hazard in a fire. The total life hazard in fires results from a composite of at least four inter locking and variable sets of conditions, as shown in Fig. 1 (J). A few decades ago, the main lethal factors were fre quently identified as burns, hot gases, and smoke poisoning. Disastrous fires such as the Cleveland Clinic fire in 1929 (4), the Cocoanut Grove fire in 1942 (5), the Hartford Circus tire in 1944 (6), and the S.S. Norortic fire in 1949 (7) raised serious questions about how people die The first major epidemiologic study was retrospective. The Columbia Pres byterian Hospital team led by Zikria (8) conducted an extensive analysis of au topsy records of New York City fire vic tims during 1966 and 1967. Carbon mon oxide (CO) poisoning was noted in 70 percent of all victims with a primary di agnosis of smoke poisoning or asphyxia. Among victims who died in less than 12 hours, 59 percent of the 70 percent who were tested had lethal or significant ex posure to CO as determined by the blood carboxyhemoglobin (COHb) concentra tion (COHb is stable in postmortem tis sue); CO poisoning was found almost equally in the presence or absence of surface burns (in 50 percent of the \ietims with bums and 39 percent of the vic tims without burns). An ongoing epidemiologic study was started in September 1971 under the joi n auspices of the Johns Hopkins Universi ty Applied Physics Laboratory and the Maryland State Medical Examiner's Of fice (9, 10). This study was limited to autopsied victims who died within 6 hour' after a particular fire. Data obtained from more than 200 victims through the en.l of 1974 (9) indicate that 50 percent died from CO poisoning (COHb > 50 per cent); 30 percent died from CO poisonin'.', plus contributory factors such as heart disease, alcohol, or burns; 10 perem: died from causes other than CO poison ing (probable laryngeal spasm, hums, or heart failure); and 10 percent died iron undefined causes. Hydrogen cyanide (HCN), hydrogen chloride (HCI). and other gases assoc.ued with the thermal decompo^i'ea ot synthetic materials did not appear mohiIicant in the deaths of the Mar) laud |l11' victims. Synthetic materials Id-.:produce such gases were "ihe pm".a articles burning in only 5'.f "I the l;K' reported" (9). SCIENCE, VOL. 200, 23 JUNE I97S 0036-8075/78/0623-I343SOI.25/0 Copyright 1978 AAAS I/O BFG39428 O T A P 'V O 'y J. POLYMER SCI.: Symposium No. 42, 1347-1361 (1973) MECHANISM OF THE THERMAL DEGRADATION OF POLY(VINYL CHLORIDE) B. B. TROITSKII, L. S. TROITSKAYA, V. N. MYAKOV, and A. F. LEPAEV Institute of Chemistry, Academy of Sciences of U.S.S.R., Gorkii, U.S.S.R, SYNOPSIS The kinetics for dehvdrochlorination and benzene formation has been studied during the degradation of two samples of polyfvinyl chloride) (PVQ at 180-200"C under vacuum with and withoutthe removal of hydrogen chloride. The effects of various additives, such as organic phosphites, mercury, silver, gold, and glass upon the thermal degradation of PVC were investigated. Mercury, silver, and glass decrease the rate of autocatalvtic dehvdro chlorination of the polymer in the presence of hydrogen chldride but silver and glass do not affect PVC degradation rate in the absence of HC1 PVC thermal decomposition under vacuum with removal of HO mostly proceeds by a molecular mechanism. At the same time, radical reactions take place. But there are some very endothermic stages of chain propagation, this is why in this case, the radical mechanism of PVC dehvdrochlorination does not compete with the molecular one. In the presence of HQ, the rate of radical reactions increases and predominates over rate of molecular reactions. INTRODUCTION Ii The thermal degradation of poly(vinyI chloride) (PVQ is accompanied ! by a dehydrochlorination reaction, color change, and crosslinking of macro molecules. Some theories explaining these processes have been offered.1 Braun2 has discussed in detail the molecular-ionic mechanism for the thermal degradation of PVC. One of the most important questions of PVC decomposition is the eluci dation of the nature of the weak structures in macromolecules from which the polymer degradation begins. On the basis of investigation of the thermal j decomposition of model compounds and detailed study of various structures | of PVC, it has been proved* that the chloroallylic groups, distributed at j random over the polymer chain, are the most unstable structures. In PVC J macromolecules the amount of double bonds is about 1.6 per 1000 monomer units.3 Geddes4 found the average length of polyenes in degraded PVC to be 10-15 by comparing the quantity of evolved HC1 with the number of formed polyenes which was determined by ozonolysis. On the basis of electron 25237019 Toxic Gas and Smoke Assessment Studies on Vinyi Floor Coverings with the Fire Propagation Test K. G. Marlin and I). A. I'owcll Division ol' llutliling Kosearch. (.'M1KO. Melbourne. Australia The I'ke Propagation Test (US 476 Pari 6) lias been used willi (lie discharge of eoiubuslion products into a small test room, to provide a lire hazard assessment of some polyvinylchloride lloor covering systems, ami includes measurements of tire propagation, smoke. CO and IICI product ion. Particular attention luis Ikiii iim to the change in llie IICI concentration of the riNiiu altnosplicre during the lest and the influence t>f relative humidity and the nature of the surface linings of the room. The results have been discussed as a emitrilnition to (lie under standing of the role of 11 Cl in toxicity of lire atmospiicrcs. Souk* approuctics for improving furtlier the repro ducibility of the lest method arc indicated and support is given to the Japanese proposal that such h test Instandardized lu enable comparison of lire liazard potential of lining systems used in Imildings. B fG 3 9 4 3 0 /A r tA J try *\/ INTRODUCTION The Fire Propagation Test (British Standards Institution US 476 Part 6) svas previously studied1 as a method lor assessing both the lire spread and smoke production potential of modern interior lining materials. It was shown that both parameters could be measured simul taneously in the one test and it was argued that such a proecdurc was necessary lor valid consideration ol" lire hazard potential because smoke production is a by-product of lire and dependent upon the nature and extent of (ire propagation. Both parameters were so much higher under llaming conditions that only this condition and not pyrolysis (mm-llaming) was assessed. Assessment of smoke was done in an unstirred room for which condition correlation of lire propagation with simulation room and corridor lircs had been established.2 It should be noted that the correlation relates to the spread of a developed lire from (he room down the corridor after Hash-over and that the position of the lining, whether Hour or wall covering, is irrelevant, live test method relating to both. A recent review4 indicates extensive research and International Standards activity on toxicity of com bustion products, hut little attention to the mode of combustion degradation. Since the main problem of gaseous combustion products is the hazard to people, and hazard assessment by definition should relate to a fire scenario, it follows that the combustion degrada tion methods associated with toxic gas studies should also relate to lire scenarios. The Fire Propagation Test is one of the simpler methods that conforms to this requirement. As indicated previously,1 the major concern is the haAirtl to |K*oplc in dwelling lircs and the rule of smoke and combustion products in retarding egress. There is evidence ' to suggest that the lloor is the element most frequently damaged in building fires. Sinec it is also known4 that prinJucts made of polyvinylchloride (PVC) may emit copious smoke and hydrogen chloride (HCIH which is highly lachrymatory, as well as carbon monoxide (CO), which is the major established cause of non-burn lire fatalities, it was decided to assess the lire hazard of PVC door coverings with the Fire Propaga tion Test, including combustion propagation, smoke obscuration and principal toxic gas production (IICI and CO). PVC COMBUSTION PRODUCTS It is known that PVC will evolve HO at relatively low temperatures. Small losses of HCI are believed to he responsible for discoloration of uuMuhiUzcd PVC when heated at 76 C for several weeks in relation to weathering studies1' and of stabilized PVC at IXO'C for several hours in relation to processing studies.' At about 300rC the loss of HCI is stoichiometric (58.6%), and weight loss at this temperature may be used to determine PVC resin content In vinyl lloor coverings.4 The rale of loss of HCI has been determined by Woolley'-' at 200 and 3(X)"C. which are considered important tem peratures for surfaces exposed to early development of lire in a room. At 200'C. 50",, of theoretical dohydrochlorinalion occurs in XX2 min in nitrogen ami 403 min in air, while at 300 C these times reduce to \ 0.38 and 0.51 min respectively. While the rale of HCI < emission is influenced by oxygen concentration, de hydrochlorination will occur without the presence of air or llame. The evolution of HCI in actual lircs is known to have caused corrosion of building materials10 and to be implicated in injury (including one fatality) to firemen,11 although studies on civilian lire victims'- have only confirmed carbon monoxide as the toxic gas primarily responsible lor death. It is well known that a wide range of-organic compounds is present in the gaseous products from PVC pyrolysis; however, concentrations of these are low and. compared to HCI anil CO, con- Pur* i Appl. Chcm., Vol . 49, pp. 649-660. Pergamon Press, 1977. Printed in Great Britain. COMBUSTION OF PVC Michael M. O'Hara B.F.Coodrlch Chemical Company, Technical Center, Avon lake, Ohio 44012, USA Aha tract - , examine thermal decomposition and combustion mechanisms of PVC are With respect to thermal decomposition, the results of a deuterium labeling (d-PVC) study indicate that the mechanism of benzene formation Involves an Intramolecular cyclizatlon step rather than an lntermolecular Dlels-Alder condensation step. With respect to combustion, the evolution and ultimate fate of hydrogen chloride and ocher combustion gases generated under NBS Smoke Density Chamber conditions Is described. The combustion of rigid PVC. flexible PVC and rigid PVC-wood mixtures haa been carried out. It was observed that under smoldering conditions, rigid PVC evolves cosibuseton gases In a sequence that agrees with the current mechanism of decomposition. It was~ discovered that under flaming conditions, the concentration of hydrogen chloride In the NBS Smoke Density Chamber rapidly decreases and that the decay follows first order kinetics. Analytical data Is presented which shows that this decay is due to HC1 condensation on the chamber walls and la not an experimental artifact. The data from these experiments Indicate that water generated during combustion plays a key role In this condensatlon. Recent data published on the exposure of teat animals to the combustion gases from PVC are reviewed In light of the above findings. INTRODUCTION The combustion of polyvinyl chloride (PVC) haa been the Subject of considerable research over the past 20 to 30 years. The excellent lgnltlon-reslstant properties of rigid PVC and modified flexible PVC were responsible for many of its Initial uses and for Its continuing, growth rate since the late 1930's to become the second most widely used plastic material in today's world market. The wide use and acceptance of PVC In applications demanding Improved fire performance is well documented (1). More recently however, combustion research relat I, ing to PVC as well as ocher construction materials has been dlrecced toward the by-products of combustion: heat, gases, aerosols, particulates and the effects of those by-products on biological systems. Indeed, one of the purposes of the present symposium Is to delve Into the cosiplexlty of just this Issue, to determine what ta known and what la not known and, hopefully, to outline what remains to be done. In this spirit then, this paper on the com bustion of PVC will prove Into the thermal decomposition of PVC (mechanisms), the combustion products from PVC, methods of altering the combustion profile of PVC and finally Into the question of exposing test animals to the combustion products from PVC. Since the thrust of this paper will be dlrecced toward the analytical chemistry of PVC combustion, it Is relevant to ask what Is Che relationship between combustion toxicology or biological response and analytical chemistry. While some have suggested that the latter can predict the former (2, 3), recent published work on urethane foams, which produced a highly toxic bleyclophosphate escer during combustion, completely invalidates this approach when it Is applied to broad classes of materials. In other words, a knowledge of the combustion gases released from a material does not allow one to predict, a priori, a biological response (4). However as was shown In this same report, ultimately It was analytical chemiscry (specifically, phosphorous -31 nuclear magnetic resonance spectroscopy and chemical icnlzatlon masa spectro metry) that unraveled the cause and effect observed In the biological testing. In essence, both disciplines are necessary and compliment each ocher In providing an Improved underatandlng of combustion toxicology. 25237021 | BFG39431 MICROCHEMICAL JOURNAL 25, 1-7 (1980) Investigations of the Thermal Decomposition of Polyvinyl Chloride by Means of Gas Chromatography Krystyna Karska Central Mining Institute (GIG), Katowice, Poland AND J6ZEF Sliwiok AND J6ZEF Rzepa Institute of Chemistry, Silesian University, Katowice, Poland Received April 12, 1979 INTRODUCTION The course of thermal decomposition of polyvinyl chloride has already been widely investigated. The patterns of decomposition products are fairly differentiated and depend on both the reaction conditions and on the nature of accompanying substances (/ -3,6, 7, 9-13). The obtained reac tion products, hydrochloride being the main one, strongly influence the kinetics and mechanism of the discussed process (3 -5, 8, 10). These facts explain the necessity of further investigations aiming at a better understanding of the decomposition process of polyvinyl chloride. This paper discusses thermal decomposition of polyvinyl chloride with particular attention on chromatographic identification of the obtained reaction products and the role of oxygen in the process. EXPERIMENTAL Characteristics of the Applied Polyvinyl Chloride (PVC) Polymeric, emulsive PVC (in form of a powder): E 68 (in the range of the K number from 68 to 72); flow weight, 0.56 g/cm3; volatile part, 0.3% (mainly moisture); thermal stability at 180C, 32 min; PVC stabilized with the 0.3% addition of NajC03. Conditions of Thermoxidative Separation and of GC Analysis Thermal decomposition of the polymeric PVC E 68 was performed inside the silite oven in a stream of air, its flow rate 1 cm sec-1, the reaction time 15 min, at 300, 400, 500, and 600C. The sample of 1000 5 mg in a quartz vessel was introduced into a quartz pipe placed in an oven at a demanded temperature. The tempera ture regulation of the oven allowed a measurement correctness in the i 0026-265X/80/010001-07S01.00/0 Copyright 1980 by Academic Press. Inc. All rights of reproduction in any form reserved. BFG39432 ZZ04CZSZ Report of Investigations 8235 Coal Mine Combustion Products: Ingredients of Conveyor Belts By Theodore Christos, David R. Forshey, and Arthur M. Hartstein 25237023 UNITED STATES DEPARTMENT OF THE INTERIOR Cecil D. Andrus, Secretary BUREAU OF MINES BFG39433 n ? \r, it:.'..:. \ r". rv \/