Document 06bR0vJ5w659bKxqqJ0wnBqkV

i - A -I H ^1$ >18 1 chloride and ence of Freonn; Q^gjethyl v 14 hlor^^knd if dimethyl ms; (1) iyl ether, (3) September, 1975 Pittsburgh PCS 12x30 activated carbon is found to be the most suitable of the commercially available carbons tested for personnel sampling of vinyl chloride, vinylidene chloride, and methyl chloride. The carbon is desorbed with CS3 at dry ice temperature or with a thermal desorption technique. Monitoring personnel exposure to vinyl chloride, vinylidene chloride and methyl chloride in an industrial work environment 1. W. SEVERS and L. K. SKORY Dow Chemical U.S.A., Midland, Michigan 48640 Introduction Late in 1973, information became available linking angiosarcoma of the liver (a rare form of cancer) with occupational exposures to high levels of vinyl chloride. Reacting to this, OSHA issued an Emergency Temporary Standard for vinyl chloride exposures limiting them a a max imum of 50 ppm. Subsequently, a reduction of this standard was proposed and became effec tive whereby the permissible eight hour timeweighted average exposure may not exceed 1 ppm, with a ceiling of 5 ppm averaged over no more than a 15 minute sampling period.1 A study was undertaken to develop a reliable method for long and short term personnel sampling and analysis of vinyl chloride. Methyl chloride and vinylidene chloride were also in cluded in this investigation because of their chemical similarity to vinyl chloride. Breakthrough Studies Field use of commercially available adsorbent carbon tubes gave significantly differing results than sophisticated chlorohydrocarbon analyzers thus indicating that existing adsorbents were not adequate to quantitatively define vinyl chloride exposures to workers. Available carbons were obtained and studied for breakthrough time by passing 200 ppm (parts per million by volume in air) vinyl chloride through a 30 ml bed of carbon at a flowrate of approximately 2 1pm (liters per minute) representing typical conditions that existed a year ago. Comparison of the parts per million breakthrough versus time is shown in Table I.2-3 Dow Experimental Carbon (XF4175L), Pittsburgh MSC-V 14X40 mesh and Columbia Carbon showed the best break through characteristics; however, the Dow Ex perimental Carbon and the MSC-V are not commercially available at this time. Replicate samples of Columbia Carbon TS-570 were tested against those of Pittsburgh w 7\ iffl Loren W. Severs, received a B.A. in chemistry from the University of California, Irvine in 1972, Joined The Dow Chemical Company, Analytical Laboratories, Midland, Michigan, after graduation. Worked on anal ysis of trace contaminants in water and air. Recently served as consultant to the Industrial Hygiene Labora tory to develop methods of collection and analysis of organic vapors to air. American Industrial Hygiene Association Journal Lyman K. Skory graduated from Michigan State Univer sity in 1949 with a 8*S. de gree in Chemistry and from Pennsylvania State Univer sity with an M.S. degree in Organic Chemistry and joined The Dow Chemical Company. Midland. Michi gan, in 1951. He is currently a Group Leader in the Dow Industrial Hygiene Labors* tory. Midland, He works closely with OSHA and NlOSH in health standards programs and with Product Stewardship for Dow cus tomers. 669 R&S155676 { TABLE I Companion in Parts par Million of Commercial Carbons; 30 ML Bad Volume, 200 ppm Vinyl Chloride, 1.97 Ipm Flowrate TISBER:--------------------- ---------------------------------------------- ------------------ --- WESTVACO TIME NUCHAR WVH (HOURS) 8x30 MESH SCIENTIFIC COCOANUT COLUMBIA PITTSBURGH CARBON CARBON 8 PL 6x14 MESH SMALL mesh 4x10 MESH DARCO 4x12 MESH PITTSBURGH MSC*V 14x40 MESH WfTCO GRADE 360 10x20 MESH DOW EXPERIMENTAL CARBON XF-4175L 0.5 <0.05 <0.05 <0.05 <0.05 < 0.02 <0.1 1.0 <0.05 <0.05 <0.05 0.24 < 0.02 <0.1 1.5 0.08 <0.05 <0.05 0.86 0.02 <0.1 2.0 0.53 0.21 <0.05 1.77 0.12 <0.1 2.5 5.22 0.S5 <0.05 0.51 <0.1 3.0 26.30 5.45 0.05 3.15 11.80 <0.1 <2 2 4 4 12 40 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 PCB 12X30 Activated Carbon.-The bed size was reduced to 1 gram of carbon contained in a stainless steel tube (14 cm x 0.25" O.D. x 0.028" wall) fitted with Swagelok caps. The test concentrations was reduced to 50 ppm vinyl chloride and the pump flowrate was re duced to 0.2 1pm. Table II shows the Pitts burgh PCE carbon to be the most suitable for eight hour time-weighted average personnel sampling. The 1 gram slugs of PCB carbon were also tested at 25 ppm with flowrates of 0.5 and 1.0 1pm. Table III gives the results of those tests. Based on these data, the following sampling conditions were recommended for the 1 gram slug of Pittsburgh PCB 12X30 Acti vated Carbon: (1) 0.2 1pm for 8 hours, (2) 0.5 lpm for 10 to 60 minutes, and (3) 1.0 1pm for 10 minutes. Further breakthrough testing was con ducted with different sized tubes for short time periods. The NIOSH method for vinyl chloride in air4 uses a 150 mg divided charcoal tube. This charcoal tube was tested with 1 ppm and 5 ppm vinyl chloride at a flowrate of 1 lpm. Figure l shows breakthrough within the first minute and approximately 12% breakthrough after 10 minutes. The flowrate for this size tube was reduced to 0.2 lpm and the experi ment repeated at the 5 ppm level. Figure 2 shows that the breakthrough under these conditions is nearly 3% after 15 minutes. The NIOSH method also states that 5 liters of 200 ppm vinyl chloride can be col lected at 0.05 lpm without significant amounts of vinyl chloride being found on the back sec tion. Duplicate test runs under these conditions showed that substantial breakthrough occurs well before 5 liters of sample has been collected (Figure 2). A direct comparison of the carbon in the NIOSH recommended tubes and PCB 12X30 was made by repacking the NIOSH tube with 150 mg of PCB carbon and repeating the ex periment. Figure 1 shows that the two carbons have similar breakthrough characteristics under these conditions. However, this may be due to carbon bed size than type of carbon. The tubes could be used in a 5 ppm atmosphere at 0.05 lpm for 15 minute sampling; however, only 0.75 liter is collected. Even with a GC sensi tivity of 1 pg/sample, a less than value of 0.52 ppm is obtained, which is above the action level (0.5 ppm) and much higher than the de sired 0.1 ppm sensitivity. We, therefore, con cluded that the 150 mg NIOSH recommended tube was inadequate for vinyl chloride sampling. Two larger tubes of carbon were tested under similar conditions. A 600 mg charcoal TABLE II percent Breakthrough for Columbia and Pittsburgh PCB Activated Carbon* at 50 ppm Vlnyt Chloride and 0.2 Ipm Flowrate TIME (HOURS) 5.0 * 5.5 6.0 6.5 A0 7.5 COLUMBIA CARBON TS-570 1.0 2.5 12.0 24.0 46.0 -- PITTSBURGH PCB 12x30 < 0.3 <0.3 < 0.3 < 03 <0.3 . <-3 370 TABLE III Percent Breakthrough for One Gram of Pittsburgh PCB 12x30 at 25 ppm Vinyl Chloride With 0.5 and 1 Ipm Flowrate* TIME (HOURS1 1.0 1.25 1.7 1.8 2.5 2.9 3.3 3.8 0.5 LPM____________ 1-0 LPM <0.06 -- -- -- 0.4 <0.06 0.8 < 0.06 -- 0.30 -- 2.80 .-- 13.00 -- l`J75 24 f 22 20 - 18 - 16 14 12 10 - 8- 6 4 2hs 0 Figure 1-Vinyl ch tube at 70 65 60 55 50 JUZ) 45 O3 40 35 <toD 30 CD # 25 20 15 10 5 0 -i--i20 Figi nyl chic reduce tub 5CJO3* ^^tur mg a di UUO11) tube. A : glass t app cly 700 A/mr 1*1 Hyxir oow EXPERIMENTAL 60 CARBON 5H XF-U75L <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 akthrough 3% after 15 so states that 5 can be col'icant amounts the back sec'.ese conditions ough c::'urs ; bece cilected c: in the 2X30 c with ne ex.arbons cter:.-t!cs under mu;. ';i due to ion. The tubes inere at 0.05 wever, only a GC sensii value of 0.52 the action :r than the deicrefore, conrecommended loride sampling, were tested mg charcoal la Gram of ppm Vinyl i Flowrat** l.o LPM m 1375 22 20 18 16 JZ g> 14 I 12 0<o) CO 600 mg commercial tube 25 ppm 700 mg PCS, 12x30, 50 ppm * 700 mg PC8, 12x30, 25 ppm Figure 1-Vinyl chloride breakthrough, 150 mg tube at 1 Ipm. Hr 0 20 30 40 50 60 Minutes Figure 3-Vinyl chloride breakthrough, 600 and 700 mg tubes at 1 Ipm. Figure 2-Vinyl chloride breakthrough, 150 mg tube at reduced flowrates. Figure 4-Methyl chloride breakthrough. 100 ppm at 0.2 Ipm for 150 and 600 mg tubes. tube is manufactured by the makers of the 150 mg NIOSH tube. A tube was also made from a disposable glass transfer pipette containing approximately 700 mg of the Pittsburgh PCB American Induitrlil Myjien* Allocution Jcurrul 12X30 Activated Carbon. Figure 3 shows the results for testing these tubes at 25 and 50 ppm vinyl chloride at 1 1pm. Either tube could be used for the 15 minute sampling and at 673 R&S155678 r'ciucc'i flowrates ;0.2 or 0.1 : TZlj ; 'T ended pericas c: u:r.e '7-3 dours) a: -d'esa levels. Two items of interest are evident from Figures 1 and 3: (1) the breakthrough time and percentages are not directly proportional to those seen for the 150 mg tube, indicating ad ditional factors in adsorption (possibly tube dimension and/or a change of carbon), and (2) the breakthrough curves for 25 and 50 ppm are almost identical, as are those for 1 and 5 ppm on the 150 mg tube. These data indicate that the percentage breakthrough for a given flowrate and tube size is independent of the concentration for at least the 1-50 ppm range. Similar breakthrough testing was per formed for methyl chloride and vinylidene chloride. Methyl chloride was tested at 100 ppm (TLV) for various flowrates and tube sizes. Figure 4 shows 150 mg tubes are not suitable for methyl chloride sampling at the current allowable levels, although the PCB carbon appears to have better retention capa bilities. The 600 mg size tube might be used for short term samples. Reducing the flowrate below 0.2 1pm would increase the time of sampling. However, by using the 1 gram slug of PCB 12X30 and reducing the flowrate to 0.1 1pm. sampling times of up to 80-90 min utes are feasible (Figure 5). This figure also shows the increase in sampling time possible by decreasing the flowrate from 0.5 to 0.2 to 0.1 1pm. A sample test of 200 ppm at 0.2 1 pm flowrate is also known for comparison. If methyl chloride and vinyl chloride are to be sampled in the same area, the methyl chloride will be the limiting factor in choosing a sampling period and flowrate. Only by increasing the amount of carbon and/ or greatly reducing the flowrate can eight hour sampling be attempted. (Methyl chloride breakthrough was also determined on the 1 gram slug of PCB 12X30 at the 50 ppm level. The breakthrough curves are the same as those shown in Figure 5 for 100 ppm. This further demonstrates that per centage breakthrough for a given flowrate and tube size is independent of concentration.) The breakthrough characteristics for vinylidene chloride are much less severe than for methyl chloride and vinyl chloride through out testing. Figure 6 shows that the 150 mg size tube may be used for short term samples(10-15 minutes) at 1 1pm. However, the PCB 12X30 carbon begins to show definite ad- v."iiiEes o'-*r is caroon "csec -re merdai race. Reducing me dowrate w,,:-. . course, ailow tor longer samoie timer ou: iong enough for eight hour samples. The 600 mg size tube has very' good retention capability for vinylidene chloride. A test of 31 ppm vin ylidene chloride samples at 1 1pm showed less than 0.08% breakthrough after 75 minutes. Thus, with a reduction in flowrate to 0.1 or 0.2 1pm, eight hour sampling is quite pos sible. (Tliese data suggest that the type of car bon has been changed in this size tube because a similiar test run a year ago showed substan tial breakthrough after 60 minutes. The 150 mg and 600 mg tubes were SXC Lot 104 and Lot 105, respectively. SKC has confirmed a Figure 5-Methyl chloride breakthrought 100 ppm at various flows for 1 gram Pittsburgh PCB 12X30 activated carbon. change of carbon between the two lots. This change in carbon was recommended to the supplier after NIOSH had received our recom mendations.) One gram of the PCB 12X30 carbon in a stainless steel slug has been run under these conditions-for eight hours with no detectable breakthrough. Carbon Disulfide Desorption A desorption technique for vinyl chloride was devised that would ensure the integrity of the sample. An experiment in one of our labora tories indicated a 98% recovery was obtained when the carbon was slowly added to the CSs cooled in a dry ice/acetone slurry.2 The follow ing procedure has since proven to be satisfac tory: (1) cool 10 ml of CSi to dry ice tempera ture; (2) slowly add the carbon to the cold CS-; (3) agitate for 30 minutes, keeping cold; (4) -rrz :na neat gem samples hav. 101%. This sair. applied to san period of tim a recoverv rar of 83% 8% wide range of accounted for. Desorpti. is accomplishe ice bath can be low boiling ga: 79 24 1E0 A 1&0 1 11 i Xta. 12 - 0 1_* 0 10 Figure 6-Vinylic at 1 Ip Recoveries for a range of 95-1 Standard Prepar A technique wa. used at Dow for gas is condenser in dry ice. The reweighed to cc lok nut is carefu ampule and the intc ureti is t' ,cn b of; een c gj yl chL bv lgtiu- Ur ft.lll r* Cf--L j Wui* C mes. bur not es. The 600 ion capability )f 3 L ppm vinn showed less 5 minutes, ware to 0.1 s quite pos- type of cartube because -ed substans. The 150 .ot 104 and infirmed a 0 2 1pm / j____ i 120 l 10 gh, 100 ppm Pittsburgh in. lots. This d to the our recomi 12X30 oeen run urs with no loride was rity of the ir laboras obtained to the CS2 Th^tollowe s^mc:e tmperahe cold CS-; :old; (4) i.ir.ysr:::? v:::- zns-.yzi:. *.r*.v;.;:nr t:ot. : w:; tee oa.h. .. '/ns :ound tna: y' tnus minimizing neat generation during nescrpticn. replicate samuies have riven a recovery range of 93- 101 %. This same desorption technique has been applied to samples of methyl chloride. Over a period of time, 25 known samples have shown a recovery range of 70-97 % with an average of 83% 8% (one standard deviation). The wide range of recovery values has not yet been accounted for. Desorption of vinylidene chloride samples is accomplished by the same procedure. A wet ice bath can be used, unless one or more of the low boiling gases are to be determined also. ID 40 0 00 70 W MiftWtM Figure 6-Vinylidenc chloride breakthrough, 20 ppm at 1 Ipm for ISO mg tube. Recoveries for vinylidene chloride in CS? have a range of 95-100%. Standard Preparation A technique was adapted from a procedure used at Dow for many years.5 Vinyl chloride gas is condensed in a tared, narrow neck ampule in dry ice. The ampule is flame sealed and reweighed to constant weight. A small Swagelok nut is carefully placed over the neck of the ampule and the whole assembly gently lowered into a measured volume of CSj. The ampule is then broken by shaking the bottle. No bubbles of gas are seen during this operation, indicating that the vinyl chloride is completely dissolving in CS2. Methyl chloride standards can be made by following the same procedure. rianzards 0: v;r.vi;:.er.s c/iioride cr; drcly prepare- by injecting z measured vmume of vinylidene chloride into a Known vomme or CSj witn a microiiter syringe. There was some question about standards made by injecting a known volume of gas into a known volume of CS2 in a sealed serum vial at room temperature and wet ice temperature so the above technique was utilized. Condens ing the vinyl chloride in dry ice and using a 10 ml glass syringe cooled to dry ice temperature to prepare a standard proved difficult to duplicate. Sample Storage The flood of inquiries from industrial hygienists, both from industry and government regarding personnel samples taken at one location and forwarding to a laboratory at another location for analysis, set the priority for determining sample storage characteristics. The backlog of samples taken in our own company and limita tions of available manpower and instrumenta tion further substantiated the need to know how much absorbed vapor was being lost during storage or shipment of carbon tubes. Sample of 10 liters of vinyl chloride at 1 ppm on the PCB 12X30 activated carbon were stored in polyethylene tubing for one to three weeks; 10 to 20% loss was noted with increased storage time. The effects of four weeks storage at room temperature were checked with both the stainless steel and glass tubes, and no apparent sample degradation was observed. However, it was found that if samples were removed^rom their tubes and placed in small vials sealed with Polyseal caps, up to 40% of the initially loaded vinyl chloride was lost after one week of storage in the freezer. Samples, therefore, should be kept sealed in the original sample tubes until desorbed. The storage effects on vinylidene chloride in glass tubes proved to be within the error * (5%) of the analytical method. Effects on methyl chloride samples have not yet been determined. Thermal Desorption Thermal desorption was investigated as an alternate to CSi desorption. Thermal desorp tion, if perfected, offers several advtanges over carbon desorption, such as reduced health . hazards. ***'*fu ^ (773 R&S 155680 The experimental apparatus consist of gas source, a dry test meter for measuring gas volumes, a muffle furnace and a Saran* film bag to collect the expelled vapors. The inlet and outlet tubes are of stainless steel as are the sample tubes. Prepurified nitrogen at a flowrate of 500-800 ml/minute is used as the purge gas. The oven temperature was initially set at 225 C, however, the low and erratic recoveries obtained were eliminated by raising the temperature to 430C. The desorption procedure is as follows. Place a 10 liter Saran film bag at the exit end of the outlet tube; connect the sample tube in the line and start the nitrogen flow. Purge the system with approximately 1 liter of nitrogen before placing the tube in the furnace. After placing the tube in the furnace, collect the expelled vapors until the desired volume is obtained. Divert the nitrogen flow from the collection bag and remove the bag for analysis. For samples of approximately 100 liters of 50 ppm vinyl chloride, 4 liters of nitrogen were sufficient to obtain recoveries of 97% 4%. The same was true for 100 liters of 10 ppm. However, for short term samples, e.g., 10 liters collected at 11pm, it was found that more nitrogen was needed. At 25 ppm, 4 liters of nitrogen gave 90% recovery. At the 10 ppm level, 10 liters of nitrogen had to be used to obtain 93% recovery and at 1 ppm, 12 liters of nitrogen were needed to obtain this recovery. Fifteen replicate, samples of 1 ppm vinyl chlo ride were analyzed by this method. The aver age recovery was 92% 3.5% (one standard deviation). Subsequent series of tests have shown a recovery factor of 80% 5%. At tempts were made to find the cause of the decrease. Neither the use of glass tubing nor deactivated and conditioned stainless steel tub ing increased the recovery factor to its original value. Replacement of the inlet/outlet tubing and the nitrogen supply has had the same nega tive effect. Although this new factor is lower than what was originally found, it has been reproducible and could, therefore, be effectively used as the recovery factor for this technique. The recoveries of methyl chloride and vinylidene chloride have also been determined by this technique. A recovery of 75% 5% is obtained (one standard deviation) for 6 liters of both 50 and 100 ppm methyl chloride desorbed with 10 liters of nitrogen. Replicate samples of 10 ppm vinylidene chloride gave a recovery of 88% 5% (one standard devia tion) when desorbed with 10 liters of nitrogen. Analytical Conditions Analysis is performed by gas chromatography/ flame ionization detection. No one column or set of conditions can cover all situations in volving analysis for vinyl chloride. A .variety of columns have been used depending upon the situation. Figures 7-11 show some of the columns That can be used under varying cir cumstances. Breakthrough was determined by period ically injecting air samples drawn from a glass tee placed behind the carbon tube. The follow ing GC conditions were employed. Column: Column Temperature: Carrier Flow: Hydrogen Flow: Detector. Temperature: Sample Size: 6' x Va " stainless steel 20% DC 200 on Chromosorb W, 80/100 Mesh 80C 30 ml/min N* 30 ml/min 240 ml/min 250C 0.5 cc with GSV, or 1.0 ml Pressure-Lok Syringe Sample analysis was originally performed on this same column at 40C. Both vinyl chloride and vinylidene chloride can be deter mined (Figure 7); however, light hydrocarbons may interfere with the vinyl chloride determi nation. Such interferences were eliminated by employing the following conditions (Figures 8 and 9). Column: Column Temperature: Injection Port: Sample Size: 20' x Va " stainless steel Carbowax 4000 on Supelcoport 80/100 Mesh 80C Unheated, on-column injection 2 microliters Other parameters were the same as pre viously stated. Temperature and earner flow are quite critical because the vinylidene chlo ride is not completely separate from the CSi; *&araa ii a trademark of The Dow Chemical Company abroad. 674. Mvttnbar, 117S c Figure 7-Chrc colt chic small changer results in loss These ci exce* r sar dichlorodiflui ether in addk situations sat: achieved und urcs 10 and 1 on Anuricin tnflwtrtu Minutes Figure 7-Chromatogram of components on DC 200 column; (1) vinyl chloride, (2) vinylidene chloride, (3) CS* Conclusions When the vinyl chloride situation arose it was found that existing techniques for personnel sampling and analysis of vinyl chloride expo sures were found to be inadequate because vinyl chloride vapor was easily lost from the sample at several different stages in the collec tion and analysis. This study evaluated numerous adsorbing media and found that one specific cocoanut carbon, Pittsburgh PCB 12X30 Activated Carbon, is the most suitable of the commer cially available carbons tested for personnei sampling of vinyl chloride and related com pounds. One gram of this carbon packed in a stainless steel tube (14 cm x 0.25" O.D. x 0.028" wall) can be used satisfactorily to col lect vinyl chloride, vinylidene chloride and methyl chloride according to the sampling conditions developed herein. The study devel oped improved techniques of desorbing the carbon using CS; at dry ice temperature. Thermal desorption was also investigated as an alternative to CS* desorption. Acknowledgement The authors wish to express their thanks to A. A. Allemang, R. A. Dommer, R. K. Lee, G. J. Roush and G.E. Socha for their assistance in this study. R&S 155682 small changes in either of these parameters results in loss of resolution. These conditions have proved satisfactory except for samples containing methyl chloride, dichlorodifluoromethane and/or dimethyl ether in addition to vinyl chloride. For these situations satisfactory separations have been achieved under the following conditions (Fig ures 10 and 11). Column: Carrier Flow: Column Temperature: 5' x Ve " stainless steel Porapak QS, 80/100 Mesh 40 ml/min N- 70C (methyl chloride, dichlo rodifluoromethane, vinyl chloride) 115C (methyl chloride, di methyl ether, vinyl chloride) Figure 8-Chromatograms of components in air on Carbowax 4000 column; (1) isobutane, (2) vinyl chloride, (3) vinylidene chloride, (4) trans-1, 2-dichloroethylene, (5) methacrylonitrile. (6) acrylonitrile. AmnjA inaviiml kygisn* Awocution Journal t575 14 18 Figure 10-Chromatogram of methyl chloride and vinyl chloride in the presence of Freon12 on Porapak QS column: (1) methyl chloride, (2) Freon-12, (3) vinyl chloride. Minutes . .Figure 9-Chromatogram of components in CS2 on Carbowax 4000 column; (1) vinyl chlo ride, (3) CS,. References 1. Federal Register 39:194, October 4, 1974. 2. Private communication, A. A. Allemang. 3. Private communication, R. K. Lee. 4. Vinyl Chloride in Air, NIOSH Analytical Method. 5. Stenger, V, A., S. A. Shrader and A. W. Beshge- toor: Analytical methods for methy bromide. hid. < Chern. (Anal. Ed.) 11:121 (March 15, 1939). 676 4 6 8 10 12 14 Time, Min. Figure 11 -Chromatogram of methyl chloride and vinyl chloride in presence of dimethyl ether on Porapak QS columns: (1) methyl chloride. (2) dimethyl ether, (3) vinyl chloride. Septimper, 5975