Document 5bnxy5N32eJamR1zrOYMg1yYR

j Uw-^ V v Combustion Products from Vinyl Chloride Monomer M. M. O'MARA, L. B. CRIDER, and R. L. DANIEL B. F. Goodrich Chemical Company, Avon Lake, Ohio, and The Dow Chemical Company, Freeport, Ttxas $ By means of a variety of analytical techniques, the combustion profile of vinyl chloride monomer (VCM) has been determined. This profile includes flame temperatures, soot content, and a combustion gas analysis. Depending on the amount of VCM-air premixing prior to combustion, the temperature of a VCM flame ranges from 950 to 1466'C. Similarly, the soot or unburned carbon content of a VCM flame varies from 3 to 6 weight percent. An analysis of the combustion gases from VCM reveal the following composition: HC1 27,000 ppm; CO* 58,100 ppm; CO 9500 ppm; phosgene 40 ppm; and VCM trace. From a hazard standpoint, the gross quantity of hydrogen chloride is the main source of danger in a VCM fire. Introduction molecule was greater than or equal to 0.40: T-HF, COMBUSTION PRODUCTS from chlorinated hydrocarbons arc of consider able importance from a toxicological stand point. This class of organic compounds, when burned or thermally degraded in an oxidizing atmosphere, can produce carbon monoxide, hydrogen chloride, and phosgene. The highly chlorinated hydrocarbons (carbon tetra chloride, etc. ] have been well characterized in this respect. For example, Sjoberg1 has de termined the amount of phosgene and hydro gen chloride released from a series of chlorin ated hydrocarbons when brought into contact with hot metal surfaces, glowing charcoal, and an open flame. The results of this study showed that, at temperatures above 600C. the major combustion product was hydrogen chloride. In a more recent study Jav= has found that, at decomposition temperatures between 600 to 700C. the presence of phosgene from chlorinated hydrocarbons was noted only when the following empirical ratio (a) for the Address any inquiries to M. M. O'Mara. B. F. Gxodrich tlhrrtiiral Company. Avon Lake. Ohio 44012. This laboratory and analiliral procedure dr\el nimeot work Ha< done in conjunction with joint 'ludio of ite Satriv and Fire Protection Committer of the ManufACinrioa Chemists' Association. (No. of Cl atoms) - (No. of H atoms) 3 ~ ------------------------------------ --------- (No. of C atoms) Little has been reported in the literature on comparable studies involving the combustion of vinyl chloride monomer (VCM). Rubier* has analyzed the combustion products from a number of halogenated aerosol propellants and has reported that, in a hot flame at 1000C. 100 gm of VCM will generate the following leveis (for a.room having a 50-m* volume): CO 63.2 ppm: COC1- 0.05 ppm: and HC1 126 ppm. Under these specific con ditions, HC1 would seem to contribute the greatest to the overall toxicity of the com bustion products. (Recent work on the toxi city of the combustion products from poly vinyl chloride has shown that carbon mon oxide and not HCI is of major toxicological importance under some conditions.4) Current interest in the combustion products from large amounts of VCM has prompted the present study. Rubier'$ results on the com bustion of VCM is of limited use in this respect owing to the fact that, in his study. VCM was decomposed in an external fire. Our interest was in the self-ignition and com bustion of VCM in the absence of an external fuel. In this respect, we have determined the 153 U R L 17952 ^ 154 March, 1971 oxidative coupling of aniline. We have iden tified the following components: URL 17953 M/E=182 and Figure 1. Effect of feed composition and burning rate on phosgene production. combustion products from VCM under both diffusion and premixed flame conditions. The temperature of the flame and the amount of soot generated in a VCM flame was also measured. Phosgene Determinations Linch ct a/.5 have reviewed a number of methods for carrying out an analysis of phos gene in air. Many of the colorimetric methods listed are so sensitive to HC1 that they were of little use in the current study. While at tempting to determine phosgene in a VCM flame with the colorimetric detector 4,4-bis(diethylamino) benzopbenone, we found that the combustion gases from VCM com'ert this reagent to a form (probably the HC1 salt) that is totally insensitive to pure phosgene. Phosgene analyses by the aniline method described by Crummett6 were also attempted. In this method, the combustion gases were scrubbed in an aqueous aniline trap; any phosgene present is converted into 1,3diphenylurca which then can be determined spectrophotometrically. In carrying out this analysis, we found that, when the combustion products from VCM were directed into this solution, a precipitate was formed. The ultra violet spectrum of this material in methanol exhibited an absorption at 254.5 mji, in dicating it to be 1,3-diphenylurea. The mass spectrum of this material, however, un equivocally showed that not even trace amounts of 1,3-diphenylurea were present. Rather the composition of this precipitate consisted of materials that result from the H2N-^3"NH M/E = 184 and polymeric forms thereof. The high level of HC1 in the combustion gases is the cause of this oxidative coupling of aniline. Indeed, aniline can only undergo oxidation when present in the form of a salt (that is, aniline hydrochloride).7 Additional evidence further points up the inadequacy of the aniline method for determining phosgene in the presence of large amounts of HC1. It has been shown that the conversion of phos gene to 1,3-diphenylurea is pH-sensitive and that, at pH = 1.97, the efficiency of this re action is only 9.3%.8 To circumvent these problems, a gas chro matographic analysis of the combustion gases, aimed at determining phosgene, was devised. For this particular analysis, either of two chromatographic columns was used. Chro matographic conditions are summarized in Table I. In addition to these analyses, phos gene determinations were also carried out with a Model 21/31 Drager multi-gas de tector equipped with CH283 Drager tubes. Phosgene analyses were made on VCM flames under a number of experimental con ditions. The following flames were probed: microburner, copper tubing burner contain ing VCM and air feed lines, glass tubing burner, VCM burning in a watch glass and VCM burning in a ruptured 1-liter metal cylinder. In each case, low levels (15 to 150 ppm) of phosgene were detected. Extensive data collected, using the all-glass burner, in dicate that an optimum air/VCM ratio is required for the production of phosgene; ratios above or below this optimum tend to American Industrial Hygiene Association Journal 155 URL 17954 Table I Gas Chromatographic Conditions for Phosgene Determination Safflower oil column: IS % lafflower oil on 30/60 CHR W, nan .w.; 12' X '/*" copper column; deiecior, i.c.. 36*C carrier aai, helium, 40 ec/min. Silicone oil column: 20% Dow Comma tilicone oil 703 on 50/80 CHR W, non a.w.; 12' X /* copper column; detector, thermal con ductivity, 58*C; carrier gas, helium, 94 cc/min. Table II Chrnmatoeraphic Parameters for CO and CO, Analysis Pnrapak column: 6' X 14" i.. Poracak S. 80/100 mesh- detector, thermal conductivity, rarrier gas: helium MoWubr lirve: 1' X m. molecular sieve 5A. 45/60 mesh: aJ] other chromatographic parameters identical to above (tee text for further information) suppress phosgene formation. The relation ship of the rate of feed of VCM to the burner, the rate of air feed, and the amount of phos gene produced is illustrated in Figure 1, Air samples above a watch glass and above a ruptured metal tank containing burning VCM were collected and analyzed for phosgene. These samples were of special significance in that they simulated actual burning conditions. The air sample obtained above the watch glass contained 15 to 20 ppm of phosgene, and that above the ruptured metal cylinder contained 40 ppm of phosgene. The Drager detector was insensitive to phosgene in these experiments due to the high level of HCI in the combustion gases. How ever, this problem was easily circumvented byattaching an HCI scrubber to the tube. This scrubber consisted of a small section of glass tubing containing 20-mesh granular zinc. The Drager tube values varied 25% from the gas chromatographic values. The Drager tube when equipped with the zinc scrubber would be a valuable monitoring device for use at the site of a vinyl chloride fire. VCM Flame Temperature The temperature of a vinyl chloride flame was measured under diffusion and premixed conditions. Under diffusion conditions, only VCM was fed to the burner (65 to 90 cc/ min). Ambient air was the only source of oxygen. For a premixed flame, air (120 to 190 cc/min) and VCM (65 to 90 cc/min) were both fed to the flame; ambient air was also available to sustain combustion. Tem perature measurements were made with a Platinel thermocouple (0.008 inch in di ameter). For a VCM diffusion flame, the highest temperature recorded was 950C, while for a premixed flame, the highest tem perature recorded was 1466C. The diffusion flame best simulates an actual VCM fire in that there is little premixing of air and VCM in the pre-ignition step of a flame. The fact that a VCM fire has a char acteristic "cold" flame is evident from the amount of soot generated. From a number of experiments on VCM diffusion flames, it ap pears that the amount of soot produced during combustion ranges from 3 to 6% of the VCM consumed. This range corresponds to 7 to 15% of the available carbon in the VCM molecule. HCI from Burning VCM Initial studies with phosgene determina tions indicated that the combustion gases from VCM contained a high level of HCI In order to better quantify this value, the combustion gases from a VCM flame were directed into a standardized NaOH solution. Back titration of this solution afforded the quantity of HCI produced during VCM com bustion. In terms of the amount of VCM burned, this level of HCI is approximately 52%. Although a theoretical yield of 58% is expected, the above analysis did not account for the HCI lost on the walls of the com bustion apparatus or for the VCM that es caped combustion. CO/CO, in VCM Combustion Gases This analysis was carried out through the use of two chromatographic columns. (Col umn parameters are summarized in Table II). Briefly, the separation is achieved by placing a 6-foot Porapak S in series with a 1-foot 5A molecular sieve column. The air sample is injected into this two-column sys tem: after elution of oxygen, the sieve column which contains the nitrogen and CO com- URL 17955 156 March, 1971 Table III CO/CO! Analysis of VCM Combustion Gases Component Ch CO NCO Volume S% 14.71 5.81 78.53 0.95 ponents is isolated by means of a 6-port valve. The Porapak column is then heated to 245C, resulting in the elution of C02, ethylene, and VCM. On completion of this analysis, the molecular sieve column is put back in series and heated until the CO and N* elute. In this way, a complete separation and analysis of CO. C02, N2 and Os are obtained. Re sponse factors were obtained for each of these components so that the analysis could be ex pressed on a volume basis. Air samples were also obtained above a VCM diffusion flame; Table 111 contains a summary of a typical analysis. It should be noted that this analysis has not been cor rected for HC1 or combustion water content. It is also recognized that the results of each of these analyses (except the HC1 analysis) are dependent on the actual conditions em ployed. For example, the relative levels of phosgene, CO, and CO= would be different in a fire where oxygen depletion is a factor. However, oxygen depletion would tend to suppress the formation of COCl2 and CO, so the values cited herein represent near max imum concentrations. Total Combustion Analysis The various analyses cited in this paper can be combined to give an overall combustion profile for VCM. Based on the C0/C02 and phosgene gas analyses and on the hydrogen chloride determination (52% of the mon omer), an approximate analysis of an air sample obtained immediately above a VCM flame would be, in ppm) : HC1 27,000 ppm; C02 58,100 ppm: CO 9,500 ppm; phosgene 40 ppm; and VCM trace. Conclusions It is felt that, from a hazard standpoint, the gross quantity of HC1 in the combustion gases from VCM is the main source of danger. While it is recognized that, in the very near vicinity of a VCM fire, dangerous amounts of phosgene may be present, it is noted that at this point the atmosphere will already have been rendered insupportable by the high con centration of HC1. The pungent odor of HC1 would also act as a warning device to clear the area or to obtain the necessary breathing apparatus to deal with the fire. Acknowledgments This work was. in part, a study for the Manufacturing Chemist Association, Task Group C (Vinyl Chloride Monomer Trans portation Safety) of the Safety and Fire Protection Committee. We wish to acknowl edge the contributions of J. E. Newell, Uni royal Chemical Division of Uniroyal, Inc., and G. Kitazawa, Chemical Division of Borden, Inc., for their important contribu tions to this work. References 1. Sjobebo, B English). Ktm. Tid. 64 : 63-70 (1952) (in 2. Jay. P.: Chim. Jnd. 92: Nd. 5, p. 533 (1964). 3. Kubleb, H.: Aerosol Age, p. +4 (April 1964). 4. Cobnish, H. H.: Arch. Environ. Health 19: 15-21 (July 1969). 5. Linch, A. L.. S. S. Lord, J*., K. A. Kubitz, and M. R. DkBbvNNEb: Phosgene in Air--Development ol Im proved Detection Procedures. Aw*r. Ini. Hyg. Asioe. ]. 26: 465 (1965). 6. Cbummett, W. B.: Anal. Cktm. 28: 410 (1956). 7. Diskbens, L.: The Chemical Technology of Dyeing and Printing, Vol. II. p. 113. Reinhold Publishing Corp.. New York (1951). 8. Crvmmett, W. B.. and J, D. McLean: Anal. Chem. 37: 494 (1965). 9. Zysx, E. 0.: Engelhard Industries Technical Bulletin, Vol. IV, No. 1. pp. 5-8 (June 1963). Received May 18- 1970 v