Document NE09kEq38E9v3aQVzenb2K22b

MAJOR OBJECTIVES i The major objectives of this research problem are: 1) To develop the required analytical instrumentation and techniques for pyrolyzlng PVC resin, compounds, copolymers, and other polymers, and 2) To expand this technology Into areas involving thermo-oxidation and combustion. LIMITED OBJECTIVES The objectives of this study were to study the combustion products ires vinyl chloride monomer in Joint work with personnel from Dow Chemical, Uniroyal and Borden Chemical. This task force is a part of the overall committee sponsored by the Manufacturing Chemists Association to investigate the burning of vinyl chloride monomer. SUMMARY OP CONCLUSIONS A) Conflicting evidence exists on phosgene as a combustion product. Borden, Uniroyal and BFG have reported none present; Dow Chemical has repo ted the following levels of phosgene for the following conditions: diffual^m flame premlstitflaae VCH in an open dish VO burning In a ruptured tank ppm (in air) nil 10-200 15-20 40 (All gas samples were obtained directly above the flame in each case) B) The following conclusions were derived from studies carried out at BFG: 1) When vinyl chloride in burned, 52% of the monomer unit is converted to HC1 This accounts for 89% of the available HC1; the remaining lljE was assumed lost on the combustion apparatus and adsorbed on the generated soot particles. In addition, a small portion of the VCN escaped combustion. )0S60 BFG08449 *T 2 2) A typical vinyl chloride flame releases )-5% unoxidlzed carbon In the form of a light soot. 3) Hie ratio (welgjht/weight) of carbon dioxide to carbon monoxide lr. vinyl chloride combustion gases was found to be 5.81. 4) The overall combustion profile for vinyl chloride is: Combustion Products Weight % HC1 (determined) HC1 (assumed lost) Carbon (as soot) Carbon (as COj) Carbon (as CO) 52 6 4 30 8 5) The relative concentration of HC1, CO, C02 ana pnosgene iucws iaza/ m a sample of air obtained directly above a vinyl chloride flame was: Component gm Concentration MAC HC1 CO CO2 Fhosgene* 27,000 10,000 58,000 40 5 50 -- 0.5 6) Based on the Haxlmum Allowable Concentrations (MAC, above) of these combustion products, the overall toxicity from burning vinyl chloride Is due to the gross amount of MCI In the combustion gases. The concentration of phosgene Is 80X the MAC value, carbon monoxide Is 2001 and HC1 Is greater than 50001 the MAC value. On the other hand, the toxicity nay be due to the carbon monoxide content. This has been tHtenae with the combustion gases from PVC which also contain a hl^Stnl of BC1. Animal testing Wild be required to differentiate the two situations. * Value obtained by Dow for a ruptured tank fire. 2037500 BFG08450 i\5 1 INTRODUCTION 3 In a previous report , the question of the presence of phosgene In the combustion gases from vinyl chloride monomer (VCM) was dealt with. At that time, It was reported that no phosgene (down to 10 ppm) could be detected when vinyl chloride was burned under diffusion flame conditions (no premixing of VCM and air). Three other laboratories have also Investigated the "phosgene question" and have reported the following results: Borden Chemical - phosgene not detected Uhlroyal - phosgene not detected Dow Chemical - not detected for VCM diffusion flame, 10-40 ppm detected for VCM burning In an open dish or In a ruptured tank (sampled Immediately above the flame); 40-200 ppm detected for a premixed VCM flame As a result of the Dow report, a number of additional experiments wre carries out In an attempt to duplicate their work. The same combustion apparatus was used as previously descrlbedl except for the following modification. R. Daniels of Dew In a personal communication has reported that HC1 (but not phosgene) Is removed from combustion gases when the gases are directed through mossy zinc. The VCM combustion gases were treated in this Banner and again analyzed for phosgene. Unfortunately, no phosgene could be detected. Prom a toxicological standpoint, the amount of phosgene found by Dow is Insignificant in view of the gross amount of HC1 that is present. HC1 Determination 11m combustion apparatus has bean previously described* in detail. Briefly, It consists of a tota^^oabustian burner containing VCM and air lines, a chimney with an outlet at aa all glass trap for analyzing HC1 or phosgene and an outlet line canmv||^Ito a vacua pump, amll vent holes were drilled into the base of the burner boosing so that air could be drawn through the entire system. Por typical BC1 detarmlnatlon, the apparatus was well seasoned with combustion products before any analyses were carried out. The decomposition gases were bubbed through 3 gas dispersion tubes (In parallel) into a standard sodium hydroxide solution. The off gases from this scrubber were directed Into a silver nitrate solution In order to Insure that a complete titrmtian had taken place. A back titration of the sodium hydroxide solution with a standard HC1 solution completed the analysis. The amount of HC1 In the combustion gases was then related to the amount of VCM burned. As a result of this analysis, it was found that 52% of monomer by weight Is converted to HC1. 1 M.M.O'Mmrm, Part VIII, Status Report, 12/4/19*9- 20975003 BFG08451 Ash Content Burning vinyl chloride has a characteristic cool, carbonaceous flane. it was of Interest to detemlne the aaount of carbon In terms of the amount of vinyl chloride consumed. Experiments were carried out In which the off gases from the vinyl chloride flame were directed through a piece of glass wool. This effectively trapped all of the unbumed carbon from the flame. The following table summarizes the results of two typical runs for two different burning rates. The results Indicate that at a higher feed rate of vinyl chloride Into the burner with a constant air supply, the amounts of carbon In the flame increase. Total VCM Burned Rate of VCM Consumption Amount Carbon Deposited % Uhbumt carbon 4.00g 0.133 g/mln. 0.226 g 5.5* 8.80g .088 g/mln. 02..283*5 g Ratio of COp/CO This important value was obtained by a gas chromatographic analysis of the off-gases from the VCM flame. A combination of columns (Porapak S, 6' x 1/4" and 5A Molecular Sieve, 1' x 1/4*) enabled the complete separation of nitrogen, oxygen, CO2, CO, VCM, and ethylene. The off-gases from the flame were sampled at two points In the combustion apparatus. One sample was taken at the top of the combustion chlsney with no air flow through the system; the other was taken directly above the flame. Q.C. analyses of these ssmples showed that the weight ratio of CO2/CO was 16.5 for the first sample and 5.8 for the second. It was f~lt that the I0.5 ratio was too high and was not representative of the actual fire conditions. The conditions undvr which this sample (16.5 ratio) were taken were such that the combustion products were released to a eonflnod area (no air stream to remove products from combustion chley). The result ytfcla was a secantery reaction involving the combustion of carbon monoxide.. .tbm^ttftinordlnarlly high ratio of CO2/CO. The coneentrmtnBFrf 00, 00^, I, and 02 are presented In the following table, along with the appropriate reaponae factors. Bssponse Pactor Wltf* * Oxygen Carbon dioxide Bltrogan Carbon monoxide 1.164 0.906 1.000 0.977 14.7 5-8 78.5 1.0 An oxygen balance could not be calculated due to the loss of combustion water In the sampling apparatus. 29133 BFG08452 Unfortunately the HC1 content of this air sample could not be calculated due to the poor resolution of this component on the gas chromatograph. This analysis nevertheless clearly defines the. toxicity of the combustion gases from VCM. Assuming that Dow's phosgene data Is correct, the concentration of this s-s In the Immediate vicinity of the flame* Is 40 ppm or 80 times the accepted Maximum Allowable Concentration; the concentration of carbon monoxide near the flame Is 10,000 ppm or 200 times the MAC value. Based on the amount of CO and CO? In the air sample, the concentration of HC1 is estimated to be 27,000 ppm or 5400 times the MAC value. Toxicity of Combustion Products The toxicity of the combustion products from vinyl chloride might be expected to be due to the high content of HC1. However, an important piece of work should be interjected at this point. H.H.Comlsh? has shown that the toxicity of the combustion products from poly (vinyl chloride) which also contain a high level of HC1 are due not to this gas but rather due to carbon monoxide. This work shewed a definite relationship between animal mortality (mice, rats, guinea pigs, rabbits and monkeys) and the concentration of blood carboxyheaoglobln. A similar situation may be In effect for the combustion products from vlr.yl chloride although the Maximum Allowable Concentration values indicate that HC1 Is the greatest contributor to the overall toxicity. Additional Studies The analytical results from this study have been extended to actual fir? conditions Involving VCM. L.B.Crider has Initiated a computer study in which the concentration of HC1 and carbon monoxide are described In terns of: 1) distance from the fire 2) type of geographic terrain where fire occurs 3) wind candltl--* ~nd 4) temperature lone This result* of *udy will be Bade available to MCA as a guide fer dealing with future VCM fire*. Ruptured vessel fire 2 H.H. Coralsh. Arch. Environ. Health, Vol. 19, July 19*9, p. 15-21. 20975005 BFG08453