Document Rp7JDKexLzbjQJpMBq94aNVLV

r 6 A ' - ;-;muMM>.iiw mt mthom raa limamwcmpum m mm ~ : ;* **..*'--* * . *7 C* J< TB#ik ; Ail 13, 1*74 ft !?: TS*. ' ' '* | 7' ^ Tlfcit ffLftfKfTttf fttt Saport Idwry ................ ........................................................................... t Conclusion* . . . . .............................................................. .... 1 Future Acelcn.................................... ........................................................................... ... 2 Baceusttaded Action .... ....................... ................. 2 Introduction ....... 2 Sxpcrlneatal Hood space Method of Analysis (1) toeponee Factor for VCM ............................................. .............................. til) Meed Space Analysis, of Goon SOnS Hocla .. (ill)Tlna Dapendancy of VCM equlllbrlM in Hoad Specs Method . . (iv) Conpnrioon of Heed Space end Solvant Method .. 5 11 IS W BxpcrloeaUl Direct Method of Analysis (1) Baproducabillty of Analysis......................... ............................. .... <il) Boapones fectof for VCM Based on Direct Method ............................ (lli)Ccnparison of Head Space Method sad Direct Method of Analysis IS *5 IS Acknowledgment .................... .... ................................................................................ ... Appendix (1) Beepanes Factor Calculations .................... ... . ............................. 19 Appendix (2) Analytical procedure S#r this Hoad Space Analysis .... 20 Appendix ()) Analytical procedure for Direct Method of Analysis. ... 23 BFG24230 i V * The head ipif* oethod of analysis for rMitel vinyl chloride none--r if I viable aethod of analysis for residual vinyl chloric* In Goon resins. The work don* la evaluating tfilo aathod hu shown it to bo auch nor* reproductible than tho solvent aethod now bolng ue*d by Goodrich production plant*. Sou indicates tho hood specs method of analysis will give reproducible results with a deviation of 207, with in the broad rang* covered by this ana^sls. (1 to 2000 PfH vinyl chloride) The head spec* nethod has the advantage of being a faster and cleaner analyses than the solveac aethed now being used. The head specs analysis is noro sensitive than tho solvent aothod. This aothod of analysis has given good results In the low ports per nil lion range (1. to .3 PPM) and should bo usable into the high ports per billion range. The hood spec* aethod of analysis is upsot by other low boll Materials whoa present in the staples being analysed. These*aetorlaIs will disrupt the equilibria* in the head space introducing an error in the results. The extent of this error has not yet boon fully investigated. * t The direct aethod of analysis, although not as fast as the head specs aothod, is faster than tho solvent aethod. This is due to the saapl* prepare* tloo requiring froa % to 1% hours for the solvent aethod coapared to 2 to 3 alautae for tho diroct aethod. Tho direct aethod of analysis is not as sensi tive as tho teed space aethod duo to tho smaller saapl* used for analysis. However, it is aoro sensitive than the solvent aethod currently la use. Tho direct aethod of analysis has tho advantage of giving auch better separation of all tho eoaponents coaing froa a saaple bolng analysed. This is duo to tho tonperatur* prograalag of the chreaatograph during the analysis. Tho drawback to tho diroct aethod of analysis is tho operation of tho chrcae tograph. > Tho procedure for this analysis requires tho chrenatograph coloan to be cooled before tho analysis is started. A coaporlsoa of tho hood space aethod (aa equilibria eyetea) with the direct aMthod ( a non-equilibria eyetea) os applied to tho analysis of Goon 103CPT76 shone excel lent agreeasnt. This is a strong indication but not absolute proof that those two asthods are accurate froa a standpoint of absoluteness. Work done in this investigation has shewn chat aoro work is needed to per fect any of tho three asthods of analysis for residual vinyl chloride. All throe analyses are based on gas chreswitograph evaluation*. Tho gas chrone to * graph is aalaly a quantatlvo instwont and it is very iaportant to taow tho conpoeltlbo of the eaaploo being analysed. Tho ccaposltlon of our aaay Goon rosins vary greatly and without a thorough investigation of oach rosin there is reason to doubt the data genera tod froa any analysis duo to interferences froa other aaterlals eluting along with the vinyl chloride. -JS* - ' -v. _ *iv*C* The tost aothod proposed by A. Borone Is a fester analysis than tho solvent aethod now being weed by HO Production Plants. Aa analysis can bo run every four (A) aiantes with this aothod, coapared to fifteen adautaa needed for tho solvent aethod. 39294BFG24231 23214003 2 ) 2. The head space analysis 1# cleaner then the solvent nsthod in that no bank fluehlng it tN44 to elialnate iolveete it4 ilia them n m castMlutlon tram lalmt laowltiii. 3. The bead apaca analyala has shown good reprodwclbillty both at cha tha Technical Caster and at Avon Labs Conaral Chsnlcai place. Thara la alao a good eorralatlon of results between tha Xarhals'' Center and Conaral Chtnlcal. * Tha elnpliclty of tha hand apaca analysis allows tha coating laboratory to give analytical results of a sanple twenty-five wloutee aftar recalving tha eaaple cooparad to one hour to ana and a half hour* for tha eluant netbod. # 3 Bncauaa of tha largo tenpin alia (1.0 to 0.3 ca) uaed in tha head apaca ana lya la, le ahould glva graatar accuracy la tha low porta pr allllaa range chan the aolvaat ayatan. *. Thara la good agraanaat batvaan Cha head apnea nsthod and tha direct nathod of analyala la tha faw croaa chacka run. 7. The direct nathod glvca better aeparatlea of the volacllca la a raala tenpin. Thla aaparatloo allow a nore accurate neesurenent of tha realdual vinyl chloride peak. Data Indicator thara la tone latarfaraaca In tha aolvant nathod of analyala due to other volatile conponenco la 0013 reals, thla interference cauaaa cha analyala to glva a higher level of realdual vinyl chloride than la actually praaanc In tha aaaple. J JQ^giS^SSSS2LLLfi&l l. An Invaatlgation la currently underway to lavaatigata tha degree of error Introduced into the head apaca analyala duo to other low boll conponanta la tha head apaca of tha aaaple vial. Beyond thla, no further work la planned. ISSSBSSdSSlSBi 1) Exparlaanta ahould ha dealgaad to teat tha abaolutaaeaa of the three Introduction Two new teat acthods have bean developed for cha datexulnatloa of reeidual vlnylchlorlda In Gaon reaina. Thla report eovers tha work In developing thana aatboda and applying then to actual analyaaa. Tha first nathod, propoeed by A. B. larane, la besad on tha reeidual vinyl chloride in a resin reaching aa equilibrium in a cloeed contaiaer. This squill* brltat distributee the vinyl chloride between tha main aaaple and tha head apace gaa of the container. Tha vinyl chloride equilibrates very rapidly idiea the y BFG24232 23214004 ) J I i. 1 l Is hu tid ibov* thi gists transition tmperstvtc of tiw r-i tv** inaly* la la conducted by sampling tha hand apnea gas of tha son lad cofU:n*r ^n-i cal culating tha vinyl chloride in tha head apnea, knowing tha e#nil I i:>r :un vnngliiona tha raaldual vinyl chloride In the ream enn than be calculated fr>ai th 1 a hand space analysis. This nethod it vary clean compared to tha sotvent nathod now being used In 8FG productIrn plants. That Is to say, there nr* r,n interferences frost solvents or contaminants from solvents. Work at tha Avon Lak Technical Center and tha Avon Lake Ceneral Chemical Plant has shown this anatysis to be practical for production plant use. Tha second method la a more direct analytic. A saall sample of raain to ba analysed la weighed Into a gtaaa tuba. This tuba la Inserted directly into tha Injection port of a 3710 Hewlett*Packard gas chromatograph for analysis. Tha heat of tho injection port drlvaa off tha raaldual vinyl chlorlda and tha chromatograph carrier gas sweaps it into tha column for analysis. This nethvd has noe yet baan tried in any production plant, however, data genarated at tha Technical Center .hows it to ba a viable method for residual vinyl chlorlda analysis. This method of analysis can also ba uaaa for raaldual vinyl chlorlda in Ceon compounds. Ic will alao detect other components auch aa vinyl acaeata or vlaylidine from copolymers. These two new methods for analysis of residual vinyl chloride ware suggested separately, the vapor space analysis by Berens of the Research Center and tha direct method by myself at the Technlcel Center. Ve had started an evaluation of tha direct method of analysis at tha Technical Centar, prior to hearing A. Berens' vapor apace analysis. In a meeting with A. Serene' it was decided a more swanlagful evaluation of these two amthode would ba obtained if both methods were evaluated togv.her. In this way data would be generated which could be used to directly compare the two methods. (At chat time no data had bean generated using the head space analysis.) Ve could not Include the sol* vent nethod in this evaluation as ve did not have the backflushing equipment re* qulred for this analysis. A few comparisons of the three methods have been node, using the Avon Lake General Chemical Plant's equipment. la all of the work done on the two analyses, no attapt has been asde to optimise ehe conditions for running either test. We have only established test conditions which give benchmark for the evaluation of the two aschods. Z am referring here to chronetograph conditions such as flow rate of the carrier gas temperatures(column, injection port, end detector), colunn packing end colum length. All of these parameters will Influence the ability to detect vinyl chloride end separace it from ocher volatile aaterlals used in the nanufacturlng of polyvinyl chloride reelne. At first glance, this does not seem important, however, data presented below will show -that good separation of residual vinyl chloride from other polymerisation asterla la does beseem Important whan naesurlng in the low parte per nllllon range. It must also be pointed out that no attempt has been nada to establish the absolute* ness of eleher teat. The vapor space method relies on an equilibria being sate* bllshed in ehe sample vial. Then from aa equilibria constant the total residual vinyl chloride in tha easels is calculated from the vinyl chloride analysed in the head space of tha vial. To verify that the level of residual vinyl chloride found in this manner la lndaed raal, a aanpla of raain with a known level of vinyl chloride la needed ae e control, ho such sample la aval labia. BFG24233 23214005 4 la the dirsat method of analysis, a soap!* of resin is hoot*4 ins id* of tho gi chromatografh In an of fort to Orivo off oil of tho residual vinyl chtortOo aoaoasr fro* tho sample. With thio method <loro lo o question about any vinyl chloride being loft In tho sample. In on attempt to determine this, staples hovo boon rorun Mini thlo aothoO to too If any aero vinyl chloride cm Id ho OrIvon off* To Oato vo hovo not boon ablo to Ootoct any aonoaor cooing oil of tho saapla which had boon previously to*tod. When a sample rooln lo . .n using both methods of analyalo, vary good agrooaont la ovldont botvoon tho two independent oota of data. It lo known that vinyl chloride has poor sensitivity in a flan* Ionisation detector (FID!. That la, tha dataccor response for vinyl chlorldo la auch lover than for nose hydrocarbons. This is probably duo to tho stability of tho vinyl chlorldo aolcculo which in turn does not yield o high concentration of charged particle* in tha FID detector. Tho vapor spec# and tho dlrece aothod of analysis ovorcoa* this problem by puttlag nor* vinyl chlorlda into tha dataccor and thereby increasing the sensi tivity of tho two aothods well above tha aolvont aothod. This increase la aonaltlvity is about two to three thousand elaes gras tar than the solvent awchod duo only to tho increase in saapla slsa. For exaaple, a main containing 100 ppa residual vinyl chloride, analysed by tha solvent aothod would put approximately l.S naaograaa of vinyl chlorlda on tho chraaatogrsph detector. This aaao rosin aaaplo analysed by the vapor space aothod would put approximately 4,000 nanogroao of vinyl chloride on the*chromatograph detector. By the direct aathod of analysis tha detector would aoo 2,000 naaograaa of vlnly chlorldo. Those fignras are based on a 30 aicroliter saapla of 91 rsaln/ aolvont solution for tho solvent method; 0.3 ce air saapla taken over a two gran resin saapla for tho vapor spaco analysis, and a 20 ailllgraa raala saapla using .tha direct aschod of analysis. With this increased teat sensitivity, it should now bo possible to look at residual vinyl chlorlda aoooaar in tho ports per billion range. It also atlove tho use of a gas chraaatogrsph with a thermocouple dataccor (TC) to bo used for rooidual vinyl chloride analysis. If this typo of detector (TC) wore used it would not bo necessary to calibrate the chraaatogrsph Mch day as air (not dacactad with PX9) could now bo used aa an internal standard. Ho work has bean dona with TC detector hero at the Technics 1 Center as wo do noe havo a chraaatogrsph with a 1C detector available eo us. larans has dona work with this type of laserwant and dasonatraced eh* feasibility of It. Thora are scat problas with these two new methods of analysis which must bo mantionod. Za the vapor spaco analysis interference from other low boll materials used in polymerisation will disrupt tho vapor apace equilibria. This diatruptlon will cause tho analytical results to ho lower than eh* scutsl rooldnsl vinyl chloride in the resin. This problem con easily bo spotted during tho one lysis by the appearance of other pooka la tho chromatogram. Low boll materials are mot common In mast Coon rosins and therefore this problem does not put much restrlc clan on this aathod. BFG24234 23214006 3 ' The problem with th direct method of analysis la cm of chmf.ogidphteeraCiM When the resin ipU la Inserted Into lh injection poet of the of cho chrometogrsph there la sot aa instantaneous release of all tha residual vinyl chloride. Actually, tha vinyl chloride la released over a period of tlwa determined by tha temperature of tha lajoctlon port and tha length of flaw it tobaa tha raala to ranch thia temperature. To compensate for th:a tlm dependent ralaaee of all of tha vlayi chloride, a smell aactlon of tha chroamtograpn cotton la froaan with dry lea prior to Introducing tha sample. Title atl"u. for a cot lection i of tha vinyl chloride at tha froaan aactlon of tha coltaan, ano than aa heat 1* applied to tha coltaan during the courao of tho analysis a plug type flow of tha 1 vinyl chloride la achieved which than glvaa a aharpar chromatogram peak. Thia allows for a nore accurate noaauraatant to be nada. 4 *i ! Ixoarlmental Kaaulta AlUUUttSJlSS&8l (U, yimLflUorMfl* Preliminary work in developing tho vapor space analyala was started by < determining a response factor for vinyl chloride on the Hewlett-Packard 3710 gat chromatograph. This was dona using mixutres of vinyl chloride gas and air. Calibration gas at 19 ppm vinyl chloride in air was'purchased and othar concen trations wore node up at tha Technical Cantar from compressed air and a 17. mix ture of vinyl chloride la air. (1) ftasponse factors determined from th cali bration gas mixCures are based on die chromatograph peak area par nanogram of , vinyl chloride. Nanograma of vinyl chloride per cc of calibration gas vara ^ calculated using the method described in tha Appendix (1). Table #1 is a tabulation ef the VCM response factors obtained for the calibration gaa adxturea. Tha data In Table 1 indicate a change In tha response factor for low concentrations of vinyl chloride. It appears char tha detector response in creases as the vinyl chloride concentration decreases. However, the calibrating gaa uoed for tho fine determinations la different from that used for cho remaining runs. Tha first eight data points represents prmsixad sample of vinyl i- chloride and air which was purchased rather than mixed ae tho Technical Center aa tha other eea^laa were. Aa error in our calibration gas could bo causing b this difference la response.1 23214007 (1) Oeorge Lees prepared ell of Cho calibration gas mixtures ned in lo method for making tho different concentre clone la reported by M. 0 mre fob. 29, 1974 - Analytical system for measuring tha adsorption of VCM on Charcoal. 39294BFG24235 . 232.14008 9ZK0d% *rt M*P U` cri to*x WIKI *M1C MIC , tl cm ' .. . *MKC *MO `MU *C00C *I0M ni* `CMC n 9* i, *; : wm'1* -*i Vr_- - < tom i* ' C*MM ** V cm ; r ' C`MM MP Mi r i<*a*c Mi M M-ac M PM M-O-C Mi PM mipm . Mi PM A K-ii-C . 'Mil `Mil `Mil `Mil CMC - 1 twx*c n" Ma< 1 v-ti . s1>< >J MPPM~ ' i*-ttC w pm n-tx- l * ) J *PMI om t *. M*f . fin * - m w on tt`i c*m * *m .1 *' ", `Ml *** Ml n I'M CWt C'm C'ltt r fit C*li C'li C*li r U " -3, MINJ CM CMC CMCC*0M C*MC COM C1*0M crooi ctopi* V croot it*pp it`0* Ulf il`M 01`t* M*M M* M*M m*n+ i*u n# : . ** ?*V n-tw mv' .... MP PM**! rr-M-c M.*| v...04*PM. i-* lfilC M *1 Oi' PM';'- - Oftt-C wt 00 PM' M*0Z*C *1 Oi PM K*iX-C *M `I Oi PM it-a-c P Oi PM M-iS* * Oi PM cc-icc SBC* Oi PM K*tt<C *. -c u^n*:tt 1WK M(` Oi PM U-ii-C M *1 01 PM ll-il-C M *| 01 PM M-il-C *1 01 PM M"ll*i M *1 01 PM M-tt* M 1 PM 01*01* MC Ot PM N`U1 MC 01 PM CI*ISaC Ml* 01 PM K*at -wr *1*| ftp 1M| mT9 - . :*1 r - V- vjj | i4 t ' 1 1 .I 7 ) At this point It not felt m cmM determine Aii mi happening is this loo nsnogran range by naming another calibration curve nixing loo parta por million vinyl chloride and air using tha mm technique aa used for tho higher coneanera* tiona. Prior to doing thla to would keep with In our onporiaontal plan and run amm roain aaaplca using tho average raapooao factor found in Table 1 (2.4) to aoo how wall tho hood space analysis would reproduce. A soaplo of Coon 103EPF-76 woa obtoinod fron Avon toko Genera 1 Chemical Plant. This resin sample was run seven times using the hood space analysis. Tho residual vinyl chloride monomer was calculated for each na using the average response factor obeained fron Table 1. Table 2 is a tabulation d tha data obtained fron this experiment. These data shows tho test to give very reptoducible results for this aanple of resin. The spread of results is only 4 ppn and tho maximum deviation fron the average results Is 2.6 ppn. Coopering these data with sons results obtained by the solvent atethod, tho vapor apace aethod looks vary good. Per exaaple, a staple of 110X334 resin was run by the solvent method 9 tines. Tha spread of the test results was 34 ppn, (fron 3 ppn to 37 ppn) sad tho nenianni deviation fron an average was 24 ppn.. Thee resin tested here was also in the range of 30 ppn residual vinyl chloride.!2) At thla point our work was interrupted due to the loss of the duranstograph in the Mass Spec tronetry Lab. We were given another Hewlett-Packard 5710 instru ment to use for several days. This instrument wee located in the pollution lab at the Technical Center. When we set this instrument up for tho vapor space J analysis, it was noted that there wes a two fold increase in response to vinyl chloride, compared to the response on the lastnaMot in the Mass Spec Lab. In vestigation of dlls showed that water in the air supply to the detector wes suppressing the response of the Mass Spectrometry instrueant. This problem with the 1detriment in the Hess Spec. Lab was corrected by switching the air supply fron plant air .to compressed air from e cylinder. - .. ,-r . :.w 4v Using the instrument in the pollution isb we again began our work by running calibration gas mixtures to establish a response factor for this instru ment. Zn an attempt to define the change in response at low concentrations of vinyl chloride, (See Table 1) we ran samples of calibration gas at 5 ppn vinyl chloride prepared at the Technical Center. Sy using different volimes of this 5 ppn sample and the Independently prepared standard (19 ppn vinyl chloride) we would get an overlap of data. . Zf die data showed difference in response between these chromatograph runs, it would Indicate we had an error in one of the calibration gases. Zf both calibration gases showed s high response in the low aanogram range it would now indicate the response to vinyl chloride was not linear in the low aanogram araa. (2J Minfllgriny k.S. Morgan in hie letter to S.O. DeCaplta 3/19/74. 23214009 J BFG24237 Ai."taia at 1MW.74 ; C.r~.,,.<l.< 14 .j. cc ton 'JJaahar x . -> --_ taala ItfaaeiflcaClaa 8aspla Wal|ht (w) VoW . 8fli laaiiaal . Hn oc raak ftaapoaaa *ei (ce) Araa Factor (awe) 1-30-74-7 103KFF-76 >la #1 1313 bra. 1.81941 .3 2634 2.4 30 3-30-74-8 " " 1.60302 .3 2720 2.4 3-31-74-13 " " " 1.63742 .3 V 2686 2.4 31 3-31-74-14 - " > ' ' 1.70644 .3 2416 2.4 2* 3-31-74-13 j .^>5; HfV-'V-.,. 3-31-74-16 ,. 2736 2.4 1.8*607 .3 2376 2.4 29 3-31-74-17 - " " . 1.96078 .3 2434 v 2.4 Aoorata vmUmI VCI 29.1 Haal-- darlaclaa trm mmram 2.1 "? . > \ \ -C. v f ' * *i- .- .-v. *#** BFG24238  1 Xt w IU decided UII^IM CM calibration turve to higher levels of chloride. To 4o this a calibration got dM at tM Technical Center at 2(000 PPM vinyl chloridewas Inclu4a4 lathis experlnent* TM data generated in Chic experlnent ar* tabulated la Tafela #3. ** represent sixty eight (64) chraaatograph rune ualag five <5) different concentrations of vinyl chloride In air. By charging different amounts of each viajrl chloride concentration tothe chronatograph, aavenecan (17) points veto aatafeliahed on a calibration curve. Theae seventeen (17) potnee epan a range of 9 aaaograaa to 5,000 nenograas of vinyl chloride in the chronatograph detector. / Theae chronatograph rune were node over a period of two days. It vaa noted that InatruMat drift fren day to day vaa effecting experlaMotal reaulta. By run ning aanplea of the 789 FPM vinyl chloride calibration i gee both day*/a ain percent (6X) deviation vaa aeon in the reaponae factor. IOC of 789 Tfm calibration gaa vaa run both daya. The average of four (4) rune on the flrac day gave a reaponao factor of 5.7 counta per nanograa. On tM second day 1 CC of the 789 PPM calibration gaa gave an average of 5.4 counta per anaogran. Thia day to day iastnaont drift vaa noted several other tiaaa during the course of ocher portions of thia study. To> J coopanas to for this, a calibration run vaa node each day ~ .before running aanplea. a-.. - - .. ft*.* . yt. -- y * v a .. . The data'ln Table #3 Chou there la an error in tha preparation of sene of the vinyl chloride and air nixcures. There is also a definite loss of sensi tivity by the PIO detector as the vinyl chloride concentration increases. Data fron Table #3 are ausoHrised in Table #4. When the response factor for all seventeen (17) points on the calibration curve are averaged, the responso factor is 8.2 counts per uanogran, of vinyl chloride. (The nexlaus deviation fron this naan of 6.2 is 40X). 'J Bov does this deviation affect the analytical results? By using data generated fren an analysis of rssin with low residual vinyl chlorld*, (sanpia vt., peek area, gas sanple also) and using Che tvo response factors at 6.2 as the naan and 8.7 as tM osmium deviation, a residual vinyl chlorld* ves calculated using the fowls shovn in Appendix 2. Ualag tM 8.2 response factor gave a residual vinyl chlorld* analysis of 6.09 ppn and using tM 8.7 response factor result* in an answer of 4.34 ppn. This is a 30K deviation la results. Consequently, if only ono response factor is obtained for laacrmaat calibration It would M possible to have a ML error in data so obtained. This error could also M expected in tM solvaet nathod of analysis where an internal standard is used since tM response to vinyl chloride is changing ooer tha broad teat range but it is not necessarily cMagtng proportionally for tM intarns 1 standard. ) BFG24239 :Tt.ObTZZZ ) ganograns Vinyl chloride On Detector 4.8 6.3 9.6 H24*7 48 127 177 254 * 400 507 1000 loi4 1400 2000 2535 5070 tble 4 inry of Data In t*tU is Chrcgntograph Peak Area Jksponse factor (Afn_opunt Mnecrans ' 42.1 48.3 79.8 89.5 106 332 ~ 637 886 1226 2601 2JJ44 6048 - 0644 - 8292 -T ^ 10888 10716 20184 I8f.7, 7.0 / {.> 6.9 5.0 5.0 4.8 6.5 5.1 6.1 4.0 )5.9 5.5 4.2 4.0 n Aveng* RtSpOM* fiatOr J When it la raallaad Chat the raapaaaa for vinyl chloride la noc llnaar for chla broad range of analysis, this problai can bo dealt with. ly dividing Cha raaponaa curvo laCo aaallor aagqanta cha deviation can bo aupproaaod co a ractoar lew value, for anapla. If Table d* la divided into throe (3) porta aa ahowa aad aa average raapoaae factor calculated for each pare, the deviation ia reduced fron 301 to 19Z.^Thia now glvea a ll doviaeloa la the analytical rdaulta. falag the aaeo data aa above and raaponaa faceora of M for the average and 8.7 for the naxlmaa deviation, the raaulea now obtained are 4.78 PPM VC1 and 4.34 PPM VC1. At the high end of thareepoasa curve (la the area of 4000 PPM VC1) the raaponaa factor haa norn effect on the analytical raaulea. Za thla range a deviation of IA in raaponaa will give n 191 deviation la analytical raaulta. (113 Head Space AnaIrate of caon 80X3 loalnt The neat atop in our experiacntel plan uaa to eapoad the head apace analyala into a typical production plant operation. Through John Hhttaey, cha Avon Lake General Chanlcal Plant agreed to try Che heed apace nethod of analyala. The analyses warn run an aenplea of Caen 80X3 ream which were being dried la a ana 11 fluid bad alio la J.tfhitnay'a Lab. Saoplaa warn Ukan ever claw aa the resin una being dried. Zn thla way a Urge range of raelduel vinyl chlarlde waa analysed. Sanplas were analysed for raelduel vinyl chloride la triplicate. The General Chanlcal Plant lab aaalyead tangles bath by the vapor apace nethod and by the solvent nathod. Also, the Tacbnlcel Canter uaa given a sbailer sat ef aenplea for analyala by the vapor apaee,Method. Ail aaalyaea worn run la duplicate. The vapor space analysis ia the Conor*lChsnlrsi Plant lab uaa restricted to sona axtaat. Va could not uaa the conditions for analysis which we bad w *ad at Tachnlcal Cm tar, but had to uaa tha conditions tee up for the solvent nethod. That --------- - and aa flows and cotuan could not be changes. 23214013 BFG24241 1II t 12 ) It* data Item thle experiment art tabulated in Table #3 below. table 5 Staple Identification 80X5 Run #3 U " * " 18 " " 2A * 2g " * 3A " 38 * " " bA 48 " " 5A " * " 5B Technical Center Reaulta 1584 150; 1424 1423 696 660 342 343 199 197 AL Oen. Chealcal Reaulta 2008 2055 1650 1730 743 755 390 391 223 221 % Different ' Between 1C A AL0C < 244 16* lOJf ' , J1iF I*p* AL Oen. Chealcal Solvent Method Reaulta 2)68 2363 2015 `SI 902 433 433 . 198 180 Saaplea are in duplicate. 1A 4 18 are two aanplea at the came point on the drying curve ' i'** * !hm data ehowe raehar good correlation between the Technical Cancer reenlCe and Avon Lake General Chenical riant'a reaulta although the General Chan!cal rnaulea are conalatently higher than the technical Center reaulta. Zt uea noted during the anal?ala that the aya tea uaed in Che General Chemical lab (eolveat ayatan) uea not giving good aoperacion of other volatile ccnponenta in the reain. There were at laaat too aide peaka on the vinyl chloride peak idiich were being integrated into the vinyl peak area. Thle could very well account for the high reaulta fron the General Chemical analyala. (It uea later determined by J. Whitney that one of theae aide peaka uea xylene uaed in charging catalyat to the 80X3 polymerization.) e The ebove experiment did not cover the low range of realdual vinyl chloride due to a problen with the heat exchanger in the fluidizing alio. Zt uea therefore agreed to nake another run ainilar to thle flrat one and attanpt to get eamplee down in the low PIN realdual vinyl chloride. Again Caen 800 una the reain need. Data fsnn theae tuna are tabulated in Table #8 below. J 39294 . BFG24242 23214014 ) Siapli Identification 80X5 Run #4 1A It 2A 21 3A 38 4A 48 5A 5B 6a 6b 7A & 8t 9A 9B 10A 108 1U UB MI* 6 . a.MWwHBW. lUwlti Bocalta Off MOlt 1833 ' 1535 1514 922 ft* 64? 599 433 396 294 287 not n*n not run 136 144 101 94.1 22.8 21.1 5.7 5.9 1689 lffpr 1643 154? Im 423 ' 410 269 264 200 198 126 129 93.0 95.4 18.1 17.6 Ut .-5 6 Difference T:C./aioc . = 36 H 33 .556 .26 if io6 r H 196 16 3 AL Jan. Chen. Solvtnt Method llltlltl 1 3852 3118 2474 2467 1597 1599 956 974 457 420 298 294 354 314 140 135 91 102 24 26 lots t&An 1 lost tbon 1 J.v ' - ~ . Thcee data show vary good agraonent batwaan.fcha General Cfcaaical lab rasults and eh* tachnlcal Cancar data. On* reason for ehl* bactar agrsaaant was dao to a cha&go in cha'chroaacograph.conditions :1a cba Gaaaral Choadcal float's lab. Xha carrier gcc flow had ehaogad vdabogh to glva batear laparatlm of tfco rlayl chlorlda Crow tha other volatllas lh' the'resin. * This agrecaoat coaid ba layrovod ora if tha conditions war* chaagnd oa tha ALCC chroaatograph and brought acre la Iin* with tha condition! wa hava suggastad. filll Tiaa Ba--iMiane^ a VCM *nilHhrlni in Head 8oaca Two saall scalo cxpcrlaants war# coaductad nast. Tha first ai to verify that aqulllbrlun was aatablishad in tha aaapla vial in fiftaaa alautas. Data froa this axparlasat aro shorn la Tahlo #7 bdlov. lhaaa data doarly shova wa hava raachad aquHibrlua with tha viayl chlorida la tha band spaca aad raain la tan (10) aa A. Barana statad. Tha data also daaoaatraea that this aaaillbrita la bald for at loaat twenty-five (25) alautaa. h* t J (JI * BFG24243 7fcle 7 14 Staple IlentIflestIon *10 min. In air oven 14 nln. In sir oven 15 nln. In air ov-n 17 sin. In air oven 20 nln. In air oven 25 nln. In air oven Staple V 2.5955 2.53*2 2.3550 2.5333 2.4633 2.3732 Sample 31 re .3 cc .3 or .3 cc .3 co .3 cc .3 cc Average Maximum deviation Resin used was a production sample of Ceon 80X5 Response factor for th.: run was 5-27* fir'.a p*r M'lllon /Inyl 71 78 7* 85 73 77 34 79 ** flv) Comparison of Heed Space and Solvent Methcd; The second experiment of Interest was an analysis of a resin saapls end powder coopound Bade from this resin. Ceon 103EP T76 resin was compounded into a customer's formulation at the customer's plant. Saaples of the resin and compound were analysed for residual vinyl chloride. The daea froa these analyses are shown in Table PS. This particular rasln sample gave two chroaaeograph peaks In the area of the chroaatogran where vinyl chloride elutes. A mass spectre of these two peaks identified one peak as vinyl chloride and the other as methanol. The methanol present in this resin sample is in a large enough quantity to Interfere with the vinyl chloride analysis. Aa to how much interference is being encountered is not known. A similar experiment has been run using 1031? F76 resin without methanol in It. This resin was compounded at the Technical Center into (Seen 85707 powder compound. Saaples of the resin were analysed before compounding and also saaples of compound that reached different tanperetures in the compoiadiag process were analysed. These data are also shown in Table #8. The data show a significant amount of the residual vinyl chloride is lost in the compounding process. There are not enough data to make any Judgment as to how consistent this loss would be in a production operation. Data la also presented in Table #8 of analyses of these samples using the solvent method. The most striking thing about this data is the poor correla tion of th e headapace method with the solvent method data. Several more experiments had been planned before concluding this work. One experiment wee to find out the absoluteness of this method. The other es mentioned above was to look at the effect of other volatile materials 1m resin to see what effect they heve cm the vapor space analyst*. Those exper iments could not be done es we havo no lnstnaeeat at this time to conduct these tests. (3) An experiment has been set up to investigate the effect of other volatile components in resin samples. This work is not being done as we do not heve an inser\Ment available at this time. BFG24244 O Qi 1 J it- **r > nuu 15 Mead tort Hsthod Meiduil VC1 teaiduai VC1 teaiduai VC1 Solvent Method Mmt Method tit Identification4 Tsch. Crtttr --AMClffftUl--- Ttch. Csnr.r 103SP 176 tula - A low F76 lesln - A Compound froa hopper Coapound froa hoppor Cajimid froa Cooler Coapound froa Coolor 3 3.2 2.7 2.7 1.7 1.6 22 21 94-32 3 - 16 103SP P76 Kesia - S 9707 Compound 0 li'0r ... -- - * - 9707 Coapound 9 230F 90 91 94 SO 37 S3 * 39 36 1.9 1.9 * 2.7 91 96 92 0 64 3 22 9 Kepresaats duplicate runs on duplicate aaaples Ixnorlaontal tesulta ,- Direct Method of Anclvclt ' .rr IS':" t*i* . .** ; \ ; ay AeorodueebllltY of Analysis . .. Dm data preeeaeed hare vara genera ced using the procedure described la the Appendix of thla repore. Aa indicated in the analytical procedure, the difficulty with thla teat la the operation of the chroaatograph. The cooling of the chroaatograph colaan would be a problea it thla teat were used in pro, Auction planta. Thla teat alao requires a aore accurate weight aeaaureaane of the seaple to be analysed. An error of one (1) allllgraa in weighting would glee a 101 error in the analytical reaulta. The Initial evaluation of thla teat aethod wee a determination of the reproducibility of the aethod. A aeaplo of Coon 110K233 renin waa chromato graphed nine (9) tiaaa in two daya. Thla reala waa reported to have 90 PfM rnaldual vinyl chloride by the Avon Lake General Chemical Plant using the solvent method of analysis. Since no calibration curve had been established for this teat aa yet, all calculations are based on the general Chemical lab results of 90 VIM residual vinyl chloride. Bata from this experiment are tabulated in Table #9 below. l BFG24245 23214017 M<ngihiUttL f Olrtrt Hathad ot *mW1i af Oaon 110X831 oc oc raak ^Ayparaat iaaala Wlht -- Am 1 2 02944 .02224 9102 9440 17.90 90.02 2 .02292 9744 92.99 4 01000 2972 07.40 9 .01090 4799 91.74 .02994 7 .02192 4742 40.00 .01910 4424 91.11 t .01272 9149 90.97 Staadard Davlatlou l, ft. Thla la only a ralatlva tti raala. baaad on a raportad KM (eacntntiM la Tha raaulta ibow la this Ubla vara arrlrad at by fladlat tha avaraga araa par allllgran of aaapla aad dlrldlag thla aaafcar by 90 (90 flM TCI raportad voo- oaar laval la tha aaapla) to obeala tha araa eooata par allllgraa of vlayl chlor- 14a. This auabar vaa thaa uaad to datarvlaa tha raaldual vinyl chlorIda la aach aaapla na. Thla axparlaaat vaa rapaatad aalag diffaraat raala. A aaapla of Gaoa 222 raala vaa na fiva(S) tlaaa. Tha data froa thla aaparlaaat ara ahova la Tabla #10 balov. Thla raala vaa raportad to hava 30PfM raaldual vlayl chlorlda by Area taka Gaaaral fliaa'-nlcaapl;*a.**' ' *, < - ft. <*-* <?* ^ OC Run Bo. 1 2 3 4 9 Tabla 10 * . . laproduclblllty of Diract Wathod of Analyala * r*. * i. . ^' " " -iV Utina oaap 222 Raala . ,-x - . ** /'& .r >'t* aaapla Idantlflcatlon Walaht ( 1 AlWFUf IHIffil Oaan 222 raala a a *a a aa a .02457 .01955 .02502 .01960 .01949 V. - kus* 1032 1397 1O09 1308 1201 * 29.81 32.07 28.19 30.92 29.05 V:Standard dariation 3.i*; *v * aa foataata, Tabla 9. 4 - > BFG24246 >k 1 I ) 17 At*In cheeo dote i(mw the ttit aethed M |lt tHillM reproducibility. <11) toermo fcctora for VCM * la the n**t experiment v* iiuklithf^ e Allkmlia nrv* for this MtM of analyai*. The procedure for this vua elaller to tho oao wood for tho vapor pact anelycle. Vinyl eh lor ldo and air alxturea vna proparad aad lajoctod lato tho chrootftograph at different voltaea. prior to iajactiag any ?eupl* tho coltan oven vet coolod to apprcxlootely 40*% and tho fim too to throa lachoa of tho coltaaa froton with dry lco. A tiaM delay of oao and one-heIf (IU oinutoa wi wood to fraoaa the coltam. Tho aoapla of `tallhratloa faa ooa than la loo tad aad tho chroaatograph ovon ooa haatad vary rapidly to 50*C to 200*6 at 14* par olnuta. The vinyl chloride a luted froo tho colon la ala aad oao-hotf (1U aioutoa. Tho data froo thla eallhraeioa aro tabulated la Table #11 below. There data are very olallar to that generated for the vapor apeca aathod. noro la ao la* eraaaa la detector eenoitivlty aa tho vlayl chloride coaceatratlea oa the detector dacraaaoa. table U Calibration Curve for the Piroot Xottod of Analytic Chroaatogrtfh tan to. sutolo oat Staple Vinyl Chloride in laaple Chreaatefiepa Avoraae aeapCMO Ptetei* Ceentc/Stnoane SUo (ee) (Btnocruu) reaic Aroa Peak ifot S-29-74-J 3.29.74*10 3-27-74-1 3-*7-7*-2 3-27-7*-3 )*t*%J* j-27.74.5 33.-2277.-774*-.7 3-27-74-8 3-27-7*-9 3^7.74.10 3-27-7*-U 3-27-74-12 3-T7-74-13 3-27-74-1* 3-27-74-15 3-27-74-l 3-27-74-17 3-27-74-18 3-27-74-19 3-27-T4-W 3-27-74-21 3-27-74-22 3-27-74-23 3-27-74-2* stea 19 m vci .9 24.08 74 Std 19 PIN vci .9 24.00 74 7# 5.14 1(4 19 Pm VCI std a 19 pm vci 1 1 46.17 .,40.17 i3 * * std 19 pm vci 1 46.17 14* std 19 wn vci std s 19 m vci Std 79 PM vci sta 79 pm vci 1 1 .5 .9 46.17 46.17 100.19 100.19 1*2 142 24f 241 1*9 3.00 std 79 pm vci std 79 pm VCI 9 .9 100.19 ' 100.19 2*3 .2*5 24* t'M std 79 pm vci 1 *800.3 520 sd a 79 pm vci - 1 . 600.3 322 std 79 pm vci 1 600.3 91* std a 79 rm vci std a 79 pm vci 1 1 200.3 600.3 920 916 9*0 2.99 std a m pm vci .9 49.3 287* std a MS rm vci .9 549.9 2099 std a 486 pm vci std a 486 pm vci std a 484 rm vci .3 .9 1 ... *739 207* 287* 2076 2.34 *10* std a 466 pm vci 1 1739 *076 std a 484 rm vci std a 484 pm vci 1 1 aa acta mimm 1 oriod *739 *10* 4099 2.39 sta a 789 rm ki std a 789 rm vci 1 1 r. soax ; 8866 "4012 \ 8.42 ... V ' >; *#- * riT-- a# all veapaMO factor* ' V --?* * * ail , * J BFG24247 3 1 i a 1 mil rwrtiin nf Am next sfrlaMt run wi mftrlMi H tka too Mlhada of uilyili. (mini ifl a( Goon ruin war* raa wlai k*tk af MMlytU. Dm <iu tram thit MfcrlMt ar* ukIu< la tikU #12 klor* -v- v nut u ' *U;v' A coacarloon of Boaalta of tlx Two iiathoda af Amlralo jT Siapla IfjV; idantlflcttlcn Kaaulta pra Vapor Spaoa Amljili (tin veil f 103BP 176 Bln #1 151$ hra. 30 31 31 29 30 29 27 flaaalta Fro* Dlrtot Method or Analyaia (PM fCl) * - *. .'. 30 29 20 24 10322 176 Bln #1 1350 hr*. I I; >,. 46.45 47 it IOaon 110X233 ~ " * '****^~* *;, .50 w ....... - '*?''*- ' As / - r* ' / o r. v***......... .. *,* ft-tfi*- . , ?;cCa*a"ldar`lat th` it t*U d`ata a`toa* ,*!* *blftu mm gMenMd Mias cm my dltteurwaatt aa_t_h_a_d_a_._o_f_a__a_a_l_y_a_la_jf_-c__h_o____ _____ my _toad corralatloa batmaa cha ti--Mt Mrtrtir.^yto xmliWu loiaddttaaaattltaasa' Mtha*tWbottthamNthMW*o wm * i>boalutaMlM/ BaaaMr^tfcTdata m att aaffUime ta uy Oat i iVlM for tha throo J^ala^aaarlao U Mli #12 ara la feat tha i*tahi 'this ^t^gaaiwratad It^l&ta* it iffmn^^mmra Ufwrtw la dad ta aacabtlah aqr taat MtBod`aldi tba 4afrMla<iaMaW M^^ ** alfe~ elan m taw it today (tha coatlnaiat 11pwinySwiawaMat raatrlatlonj. x as rafarrlag to all of tha aakaaaa which Mm Woto'Uahc darlas thia i*****1" atloa( itah aa Klimlairia^Na)MitMidhf jlljji ailh.lha rlayl ahlarlda* ***""*'"* *&& %e i~ -i' *3&! S7^ ' 1. tha author acfeaawladfaa. aaat of tha |H liroaatafrath for bio holy 1* adltlas'a** 4^*. > yV . BFG24248 ozowzes t'f* ' it J ir. t CO Huron nttii ditilitli ruun for 19 ffo VCl I9m of 1C1 is i000,000 m if aoapU OCI f* ; r- horo i fli tfco n'i of vet fa 1 oo of eollbraeloo |t oMa$t k-1> * 10 * of VCl lo 1 ce of 19 ppo oalfancloa > w-ta" ti* f&WBy-81 r V, or 40.17 .collbntloo gM iirUiTof ,19 fpo VCl ' - to rooT ' : r?' AV; *;. f- *` W*^jt2'`J*:` * f*'* 'T7 *-'4- r ^.' v t;.4 * < . ^ 23214021 . .afcj *v.^* :VX . ...... ..... ^;j. . i- . - J -S*o-1,',.^ 5 '!' ~Ji+r %', Ti-iHi/.Cr'tr * - ' J3* V.. .r - * .''**.**'* 4 #. . **.: * f V# . v.*W- . -k. _|V- * w --W.r: . ,`f"H * *>W-:. :*vL-,' ^..vrjHSr^. / - - 1/ . V .; J* ".-`j-^5^ i ' `;^vV ) f 4* ,/ /-y' Jiff**- V>1-' D- BFG24249 Appendix (2) Analytical Procedures 1. The heed apace Analysis Dm |ii chreaetocraph (CC) to bo used for the mljrili is wt up with th- following conditional * e) 1/1 inch x ft. stainless stool coltani packed with M/100 aesh Porapak Q* ` b) Tho goo flows are adjusted to obtain tho roc--anJad settings proscribed by the CC aanufacturer (Hewlett-Packard) he litas 30 cc/alo hydrogen 30 ce/aia - air 240 cc/ain. c) CC taaperatura settings ere: Injection port . 130C detector 230C oven UM Afeor tho abewe conditions have'-been sot ond the chrasstogrsph hoc stobillsod at those conditions tho lnstnaeat Is ready for calibration (dctarnlastlcn of lnstnasnt rooponso factor). *. * ,, Prior .to analysing any rosin aaaple, a rooponso factor asst bo calculated for tho gsscfaroaatograph which Is .to ho wood. Tttlsleiftoeaby nslag two gsa wtx? tarns of vinyl chloride and air of haoun concentration. - One' saaplo at a law concentration (about 20 ppa pci) and one saaplo at a high .coacontrstlon (about 1000 ppa PCI). . h The calibration gas alxturoo are bougie os coaprossod gas la a cylinder. To nan than a glass saaplo vessel of shot 1 liter Is needed. Per our work wo need a 1 liter cylindrical separator funnel with a stop cock at both ends. A glass septvn stea ana also fitted to one end of the vessel for rsnowing syringe staples frea the vessel. The soaple vassal la fitted with a septus and both stop cocks are opened. The low concentration vinyl chloride gas is purged through the vessel at about * 3 liters per aAaute for about one alaute. . Then both stop cocks are closed seal ing the vessel with the gas saaplo in le. Tha stop cocks are closed in each e aenaer as to elialaate a pressure buildup in the vessel. That la, the step cock nearest the gas cylinder le dosed first and than the exit stop cock la closed. The needle of a lcc gas syringe i tote the senple vessel and the syringe plunger drawn >eek`ns fer ns 1 go. The syringe to left la this position for five seconds .to1 alien le gas to fill the syringe, After the five second delay, the naedldis n the septan and the syringe plwger is nosed to read aaactly 1 ec. .-fhU.gnd'Jjiiaple la than injected into the chreastograph Injection port aid derfoT -- BFG24250 23214023 ISZPZQdS (M3Q r Jjr: nit >* wn naqa i iwnfij *i*n iff 1 nrn> i*n* > nMtnnn < q.qond tlwni aqy *T popTAyp l^TA jo -- ifmy P JL (c) Fr? ;*- TK* - - * " v- (MI aaf 9 Iq) tou mini > i|dffiii aq nf tqm apyaotqa 1***+ jo aqfyoA aqa yo aayaaamsaaaa (j) ni< tlaiiXi aqa try tpIMIV ! M illtnU* it ,,* *1 ,:' ' - '-: wr t--. T -aa^jnvL^.-r - ling qtfaafo*--aqa til aqa Of apytTP iday yo:"amq;aa aqa jo ooyaaaTauaaaa <I> ,f-4 . 'piwtqt iXap aoy mq aaoodaaa Mfuanf n apan aat tooyaaitwias tvpiouoy atp aaap ayqa mi aaitaaa aaaqa Mp jo tpm aoj mon qnd ituni oa mt> oj aaqaafoa palim* tit wu ntim tmxojjjf nup aqa yo qn aoy pnej imt qni qdajfoaanaiqa mj tqi *aanya moy uni a* ayqa ea 'qiaatoa--aqa aqa any lim ipp anfai pn n r iptm paaa ea aafanid tip Ipaq inn* iidni tip mij atpaan tfcuii wp foyoonaa in}y 'tpiiom liap ;o atiataannti qlp tip qn* panyj intt* tydan up ana; aa f 70 <tttxt ay 3no Map 'ptw i|nt|*tii a* tinptmi anva aqa luiMne) 'ptitidni tit anu aaaqa oaqit *aana ay aaf aidnra aiq) qayn aona qdtafoa--aqa wo) aqan opfaofqa litj* ooyaaaanaa noa net oqa joy paan aa oanpaaoad anaa aqa fayooiyoy pna oAoq? paan aa altflia aaf anaa aqa qap tint* aydnaa aqa mj aa \ jo aaaana ay ano atip oyaty t *aaaayn aaf t.'ijoma |itl* aayauaaaaooa qfyq aqa nan iiut aooqa oaayf at aanpaaoad tan aqa ftiyno jaaaa* aidant aqa afaaqaaj 'paaaidaaa ait tnu aaaqa aoyyy aavaa nay aaaay aa tyqa o| qiajfaaanajqa aqa ea apyaoiqa iAi|* aoyaajaaaanoa ( np ) n i faylaiqa ya inptiiK fsfifnta aooqa aqa avndta paa aiwpi ana aim aaaafa qiatf npyjoiqa |fp* aqa inyy *aaaya nan qaad apyaoiqa (itp* aqa (an|M go a|)JW*f tqa if fajpaadtp) foamy aaaqa ijaimniMt q u * ,0 * s -- -*. r v ) *^ :S J dJe- 22 fra dM threa a((mt iMpl ef vinyl chUrUi gaa aUiwaa rw^thtM mp<n fciuri in determined, TbM dm fNtm act aeertgod tagsOut te oMiU a response factor for ealenlatlag resides! vinyl chloride it a aaapli. Tt analyse a roala aatpla for reeideal vinyl chloride, Mi|h approximately 2 gram of tha roala into a rial of known wlm (Wo hart wood a t] ca vial flttad with a removable aaptm cap* Thla glvta a wolght to volte* ratio of 1 to 4 which worked wall in our analysis.) foal tha vial with a aaptm ant place la at air ovaa control 1*4 at 906. Laave tha vial la tha oven for fifteen minutes. After the fiftoon minute heating time, remove the vial from tha oven and Craw out a cample of the gaa from tha top of the vial with a gaa syringe. Inject thia ample into the chromatograph. The aample also injected depends on the amount of raaldual vinyl chloride la the roala* In the range of 1000 ppi to 100 ppm *2 ca la aufflalaat. In tha range of 100 ppm to 1 ppm .5 at la cmfflclant. The vinyl chlorldo peak will elute in approximately 3 minutes. One minute after die vinyl chloride peak alataa from the GC celmn another aample can he injected. ^ To calculate tha reaideal vinyl chloride in the reela aample the fallowing data are needed! l; leapenae factor for vinyl chloride 2. Weight of the polyvinyl chloride realm in vial C^e) 2. Sian of the sample injected into the chrcnatograph <ee) 4. Area of thc chnmntogragfc'peek for the vinyl chloride from the aample. 3. ; Voluwmne oef the aample vial yfaoe) ... i/` K' Atmoepherlc pressure (baroma isK*- BA.vf .4- r- - It v -'__>_:_tJ_*'' fhaae data are manblaed in the following apintion to give tha residual vinyl chloride for the sample in part* per million. a 23214024 * y > i i BFG24252 analytical Jroctdntt toTlHfocl Withed of toilnH j; 1hl netiiod also raguirce a nepeaee Utter te be eelcvlatsd prier to anal* sin any atarlae for reeidual viayl chloride. TfcU la dees la the tana aanecr d fa tho vapor apart aualyaie akora. The oaly different* beiaj the way cha chronaeogreph la rua. Tha chttaaeograph la att up with cha sane eeltraa lal condition* at used In cha haad apact ana Wait except Chaa cha actual arc than coaled with dry lea for aoa aod cat half (1*) alaueaa. The ataadard gas aanpla la chaa injected on to tha colun and tha owan canperacure la brought rapidly CO Mrc> AC 50% a taw paratura progran la aeartad oa tha celtara oven to nice Cha tenperatsre to 200% at a rata of lb4 par nlnute. Ualng thla nethod tha vinyl chlorlda alutaa trm cha ultra la appraalaacely 7 alutaa. with thaaa condldona cha thraa acaadard vinyl chlorlda air adapter era rua to product a raapoaaa factor. After thla calibreClan la aonplecad cha Intentone la toady to run aaaplaa. To analysa for roaldual vinyl chlorlda la ratio with thla nethod, a snail aanpla of tha ratio (.01 to .03 gran) la weighed into ala glean tube 93 an long. Tho aanpla la poeltlone^ la cha tuba 30 rat trm cha batten by a gleet. wool plug, with the chroaaeagrdph colrau oven at rota tanpartCara and cha flree two to thraa inchat of tha coluan cooled with dry lea. cha aapara la ranrad frra cha Injection pore and' tha aanpla Cuba dropped In. The pert In Chan xwaaalad with a eeptia and tha ana lya la aeartad. .The oven progran la Initiated whan tha oven temperature ratchet 50%. In approxlnecsly 7 alnataa the vinyl ehlarlda peak will aluta fron tha colurai. Aa aeon aa tha vinyl chlorlda peak hee elated, cha aanpla tube la rrawvod fren tha Injection pact and .thn part raaaalad. .'i ,.k. ./ ' ... Evaluation of thla nethod haa.bean.conductedualng a levlete-yuckard 3710 get chrcnatograph equipped with a Clone ianiaatlea dttactor (flO). Thla chranacograph la ideally suited for thla ana lyala aa tha Injection pore la la a vertical position allowing tha gloat aanpla tuba to ha dropped Into cha part and positioned by gravity. Tha glaaa wool balds the aanpla In tha cancer of the baa cad area of tha Injection part. It la the heat of tha Injection port (130%) which drives off tha roaldual nononer. ;BI' chrcnatograph carrier gat than awaapa the nononar once the coluan for acparatlon and detection. Tha FI9 detector allows tho flow of carrier gas (helliaO to be Interrupted whan tha aanpla Cuba la In serted into cha Injection port without haxudng tha detector. To.calculate tho ports par nilllon <ppu) vinyl chlorlda In Cha main aaaplaa using tha wathod, tha following calculations are nadat <1) Peak area of vinyl chloride Kaepans# factor (2) vinyl chloride fa t of aanpla (in nl vlayl chloride la lias V- S ppn vlhyl Alarlda fa seapU - fr- , ! - " .. I n $i 23214025 BFG24253