Document yk4ZDe8bwdvy8vLM3e9rwegJ4

Tlank you for the opportunity to visit the Louisville plant and examine in depth the personnel monitoring program for vinyl chloride. Your dual tube program to determine the ircidence of vinyl chloride breakthrough has shown that vinyl chloride is breaking through the primary tube about 30% of the time. However, after close examination of your measure ments, it is clear that only 13% of the time is this breakthrough greater than 35? of the anaunt retained on the primary tube. I judge this situation as serious and in need of explanation and correction but not critical. This cause of the breakthrough is not clear to me. It can be related to poor sampling practice, dirty PMCC tubes (inadequately cleaned) and excessive amounts of water on the PM;c tubes. It is likely that all of these contribute to the problem. It is my judgement that thermal cleaning is a small factor. It is more likely that unusually large quantities of water are responsible for the breakthrough. Unusually large quantities of water are th )se which arise from collecting suspended water (fog and mists) not just the water from humidity conditions. ** PMCC dual tube program from February 1976 through July 1976 generated 215 sets of surements (2 tube combinations). This represents S% of the 2524 personnel monitorings conducted during that period. The program shows that in 63 cases (29?) of the 215 dual tube te ;ts breakthrough greater than 10% occurred. These breakthroughs are not related to the amount of vinyl chloride retained. My calculations show the following distribution among thi 63 breakthrough cases; PI 'H Range c. 1 1 to .43 J >4 to .86 . 7 to 1.63 1. .4 to 3.26 >3 27 Percent of the 63 Breakthrough Incidents i1 14 24 22 19 10 Twq> qualifications need to be imposed upon your results to bring them into proper per- sp ective. First of all, the number of breakthroughs (63) should be reduced to 55 based up< >n a probable artifact of dual tube programs. Eight dual tube sets showed virtually no hing in the primary tube and all of the measured vinyl chloride in the secondary or b>a<; k up tube. It seems that the tubes were reversed and the primary was sampled, labeled or analyzed as the secondary and vice versa. Th n, we must not use 10% breakthrough to the secondary as significant but rather choose 35 . When we use this new criteria we see that 13-5% of the 215 tests showed breakthrough Th s says that 13.5% of the time single tube personnel monitors will be out of the 35:5 uracy interval identified by OSHA for measurements in the range 0.5 to 1_.0 ppm. In otre ds one out of every 10 personnel monitorings is a bad test but which one is it? It is sa e to assume that if we can find the reason for this breakthrough we can reduce tnis un 1 :e r ta i n ty- BFG-I- 646 2- - exa.mi .".ad your personnel monitor program in further depth, looking at the laboratory's jrk, the calibration of the pumps by your people, the selection of the test subjects, the ciresistances of actual field sampling (the incidents of the six hour sampling period) and the storage of the PMCCs before and after sampling. Based upon my examination l make the following comments: a) Proper care Is made to calibrate PM pumps prior to and following the six hour sampling period. b) Good attention is given to the selection of employees to be tested. c) Inadequate attention is paid to observing the employee, his pump and tube during the six hour sampling period. d) Insufficient attention is paid to the water "puffing" during the analysis or PMCC's. When water "puffing" is observed the measurement should be voided. e) The laboratory needs to analyze freshly cleaned PMCC tubes ("20) to determine the average signal level from cleaned tubes. Then subtract out this value as a "blank". I took the time to observe personnel samplings which were in progress. 1 found a sampling vhich was subjected to massive water sprays and mists. This water can be retained on the nMCC tube. 1 found a case when the pump and tube were resting on the window sil with no e around. Finally, in a spot check for the men and pumps of 10 samplings I could find iy 2 tests. Not knowing where the others are, what they may be picking up or if they are m use is a serious question. It is a good idea to pay closer attention and record the conditions during the six hour sampling. Tie incidence of water "puffing" v/as observed in the lab and I feel it tells us that a great deal of water (too much) is on the PMCC tube. When the tube is removed from the Jsndix Flasher after analysis, a puff of steam jets up from the point where the tube ;*ats in the downstream solenoid. In addition to that, occasionally water is so abundant :lat it collects in the connecting tube between the flasher and chromatograph. This is concluded from observations of water puffing following the second, third and fourth iggering of the flasher inject switch after the first "puff" is observed and before any P ICC tube Is installed in the flasher, At present this puffing observation is not even `corded. It should be recorded and it should be the basis of voiding a measurement. is a good idea to continue the present dual tube runs employing new PMCC tubes and cleaning. Fifty or so tubes seem adequate to establish if 300C cleaning will iminate the breakthrough observed from December to July. I : d oes not seem likely that this will eliminate the breakthrough because breakthrough i s n >t pri -aril y associated with low level samples. 1 : is clear from humidity studies at Brecksville that humidity up to 92% n )t alter the perfmrrnance of PMCC's or produce significant breakthrough, t:e likelyhood that water in droplet form (fog, mist) can be filtered by Qosicing quantities of water much greater than that resulting from 92% (relative) does There is, however, the PMCC thus relative humidity BFG-I- 647 r -3- ir. This very large quantity of water could easily account for breakthroughIso be associated with the water puffing observed in the laboratory. It should In the light of suspected high water charges to the PMCC your proposal to try a desicant study is wise. 1 have two recommendations for desicants, silica gel and calcium hydride. The silica gel recommendation comes from work Chuck Titus completed. In testing a vinyl :hloride point source for composition, a silica gel trap was placed upstream of a carbon trap. Vinyl chloride and other monomers were transmitted by the silica gel but water was *etained. We do not know if this case was 100% transmittion of vinyl chloride through the {Mica gel. However, analysis of the silica gel should clear up this point. And I would inalyze the silica gel with the carbon disulfide procedure just as if it was carbon. The ias?s of this silica gel recommendation is a mass spectroscopic analysis of the carbon led which was performed for Chuck by Bob Lattimer of the Brecksville lab. 1 he calcium hydride recommendation comes from work being carried out by John Nikora at the Avon Lake Development Center. John is using a calcium hydride tube ahead of a carbon tube in a gas stream. The stream purges a water sample (and thus becomes virtually saturated with water) to strip vinyl chloride from the water. The calcium hydride removes tie water but does not appear to retain the vinyl chloride which is retained on the carbon bid. This is a preliminary conclusion because the calcium hydride bed has not been analyzed for retained vinyl chloride. John has only prepared various mixtures of VC1 and water (at the part per billion level) and obtained good straight line relationships between analyses and VCI content. You can get more details and keep up to date on this w>rk by discussing the work with John at Ext. 288, Avon Lake Development Center. summary, 1 believe you have a breakthrough problem which is serious but not critical, I do not know if this same problem exists at the other PVC production facilities but 1 w II find out. I recommend you continue your breakthrough program with 50 sets using new ar d clean flasher tubes and 50 tests employing desicant tubes. You and the other plants wi 11 be receiving the PMCC/humfdity report from Brecksville. Finally, 1 plan to examine al 1 BFG data and experiments and prepare the best statement of accuracy and uncertainty T lating to personnel measurements employing the PMCC/flasher method. jb cc E- B. Katzenmeyer, Jr. - Akron R. W. Strassburg -- J. Nikora - Avon Lake DC J. M. Whitney - Avon Lake GC C. H. Lufter C. E. Titus J. W. Born 0. S. Kurtz _ Breck1s1 vi l . e _ 11 _ 11 J. A. K1upar - Cleveland P. H. Lawrence S. 5. Michels R. R. Taylor - Louisvin e n 11 R & D Files Paul M. Zakriski