Document 5LL0153M139ryD73OXjmD0Ox0

1.5 DKPARTMF.XT OF I .AMOR ' Vk IIIMI M'11.11 >jlctv ukI llv.iltlt i>ri i (it >i i n DEC 17 27S m.uiiiinutun. pc /V- o Mr. R. Jx. Wheeler, Jr. Vinyl Chloride Resins Manager Union Carbide Corporation 270 Par* Avenue New York, New York 10017 JftH 02 R N WHEcU^ ft- Dear Mr. Wheeler: This is in response to your joint letter with Mr. John Whittlesey dated June 17, 1975, petitioning Cor modifi cations of the Exposure to Vinyl Chloride Standard, 29 CFR 1910.1017 (formerly 1910.93 recodified May 28, 1975). i There are no plans presently to amend the Vinyl Chloride Standard. A revised program directive is contemplated, although we are not certain as to the date it will be available. Please be assured that your comments and suggestions are greatly appreciated and they will be considered fully in the revision of the program directive. j In the meantime, the following administrative decisions have been made: 1. 29 CFR 1910.1017(a) and (b)(6) Scope and application (2), (3) and (b) Definitions (6) The standard defines a fabricated product as being one which is "made wholly or partly from polyvinyl chloride, and which does not require further processing at temp eratures, and for times, sufficient to cause mass melting of the polyvinyl chloride resulting in the release of vinyl chloride." "Release of vinyl chloride" means the release of an amount of vinyl chloride which would likely result in employee exposure at or above the action level without regard to the use of engineering controls. Products which can be classified as fabricated products are exempt from the provisions of the vinyl chloride standard. All other products are subject to the re quirements of the standard. If the employer uses or manufactures a product which is not a fabricated product, he must initiate monitoring procedures. If the monitoring reveals that the employees are not exposed to vinyl chloride at or above the action level, the employer's operations will be exempt from the n 0 /COO 33 2, n prcvisicns of the standard. So>/ev:, if the sanitating reveals exposure at or above the action level, the eoployer cust icpl scant the procedures specified in the standard. 2. 23 C77. 1310.1017(b) Definitions (5). ii She petition requests that the definition of "azargency" be revised to include specific exhspies, such as lire and explosion. We acres that the definition tight veil i be expended to include exazples. Again, this tatter will be addressed in a prograz directive and not ac an atend- cest to the standard. She definition of a "massive release* as being "greater 100 parts par till ion (ppt) " found in the current Prograz Directive 5200-23, will aica be addressed in a f revised directive. ?7e agree that the 100 ppt should be changed. * 3- 23 C?2 1310.1217(d) Monitoring (4) - Sbe intent of paragraph (d)(4) is that tf.s esttlover shall he 35% confident that his .monitoring result is within 23*, 235 or 505 of the actual value depending on the concentration. ' Sherefore, .an czpioyar using a' cached which has proven vinyl chloride datsetion accuracy of 234 or less need take only cne(zsaaure-- tent regardless of the actual vinyl chloride'senator esneentratton. In concentration ranges where accura-- ci.es of 334 or 30% are required, the esploycr need taka only one teasuretant if the netted accuracy is less than t-ha specified accuracy. !7ith tschads of unknown accuracy or having errors greater than the specified accuracy- recuirezsnts, repeated ceasure-- tents* are necessary. In these cases, one any use- the coefficient of variation (C7) as a par acetar tc judge whether or net a stripling procedure is adequate to. seat the standard. Tha C7 in percentage units is i defined as the standard deviation of the .cached, trr.es 100, divided by partissihio exposure licit. The recti; C7 cf the procedure is obtained by dividing the required accuracy hy 1.35 (Z value for 35 5 csr.fidcr.ce} Thus> for accuracies of 135, 255 and 53 5, ccthsd I*' values should be lass than 12.0%, 17,0%, and 23.:.5 respectively JOXOOS* r 60 e 0 b ::zzv zzzzzzz-zzznvz v zr zzrzzzz: :kia 'ZZZZZZZZ TZilZA cO {T-dd *C ) sZJinziJY ZZ>VZ ZZZYZXZD ! ZAZ}' SZTZZZZ7 ZZtiZAZZ&i ZZZIZTTS :p*sra s>q Acs Bay paon Bay.pcTTay qp 'ancones; epyacyqe tAstta yanpysE: cdd C*0 KSy*, S5T SCTEPUCC PiA pCyeCV'T Pq pSTlC OCT acys<2 tAey aAXcd paptrorq eqp casqA cuctpbspts see saeqp (SO *T5*7 s**^`S ; (s)iICT*OIST SS *S -puaTsaaTrvBos r sy. BuTayaaq 'paeparp.s a^q Ac psa&Ace suoyqBaade aseqp ac *pspuasB aq 7TTA saoT?TAo:t Bayayszq cqq prep 785577 3- ey qy 'sacyaaaqj , Beypacdsarap :c;pi paaado uaq* aeyaeTrp pAtrra yo raoypEspueeaec q5yq ?7a*A Ae= saoeyrp -ucs cSoacps prap uoypsqypar sees'aesq esyr s?q aaaqs soeuarteeeuce ry^Ts'sed sqq yo bzeab q pyneqs saeAoyd -e pur ansae cp spuspysep acy Apyaapaedde aqp. sAa^y* st aeeqz * apyaoysc tacta ppta psapeee ay ScTasa ^o ApyyyqT ssed ain AEq oua yuuesad it a7 Bayaysrp pasasaA cp anepasarq AypucTeyyyns sy py psqp pspreyptrp arq pTaoyqe tauta uc psyydqice sae^ aa soaapyA aqi psapase ay ease Aa^p cctpa qpy* saeuErpsqns snepartarq eqp ye capita qp yo peepaide q Tytrqs sseAcyds pa^p sapTAcad OiST 7 pa'T epypn ?c5? Apsyt?S yrucTPodnseo atyp ye (L)(q)s acypeas * sot q yatpb Buyaycaq aaBByrp . cp eansodac ye teas'? b Ayyeads qoa saop paeparqs aq Bayaypay, (PU70 7*07e7 SS3 SS ** `qasaeaytvBsa c py SayacpyaoE a-aq* Bays,* core ye.a a assay try syr pa?' ueyryAsa SATpesayp creaBcad aqq ay paaapys -ucs sq yyyM Bay acp rues ec spuassse anoA pr.q parparps aqp pasen cp Aypaassac caByc cu saa aaaqp 'pradsa c; % r n i L< i 1 He hope that the above clar ifirstior.s will sat your petition for notification and naencscnt o riayi chloride standard. Ac previously stated there are no plans presently to formally an-nd standard. There will he an addendum or aociii, tion of the Procraa Directive *200-25. Should you wish to discuss any natter further do not hesitate to contact us or centers of ay staff. 0X0053 t ft f rn i i I i i] D 0 II n U.S. DEPARTMENT OF LA.30EI Ccsasasionai 5aA:v me Hssiih Ai.-runiscriticn i 1 . J/ 3 V.lSHINCTC.'.'. 0.0. 10013 o V * \% Mr. Rayr.or.c S. Scheme* Attorney Air Produces and Chemicals, Five Executive Mail 'Swedesfeed Read Wayne, Pennsylvania 13C37 Inc. 0ZC2 * STS ^VV r.~ Dear SchsncM: This is in response no your letter of June 30, 1373 to Assistant Secretary John 5. Star.cer petitioning for the accif icaticr. os IS CPU 1310.1017(h)(5), (formerly 29 C"R 1910.93q(h)(5) recodified May 23, 1973), Exposure to Vinyl Chloride, Occupational Safety and Health Standards. 29 cm 1910.1017(h)(5) defines a fabricated product as Seine one which is "made wholly or partly from polyvinyl chloride, and which does not require further processing at temperatures, and for times, sufficient to cause mass melting of the polyvinyl chloride resulting in the release of vinyl chloride.'' "Release of vinyl chloride' means the release of an amount of vinyl chloride whi.cn would lively result in employee axmosure at or amove the action level without regard to the use of engineering controls. Products which car. he classified as fahricatac products are exempt from the provisions of the vinyl chloride standard./- All other products are suhiect to the recuiraments of the standard. Thera are no plans presently to formally modify the vinyl chloride standard. Therefore, we hope that the id ocv* c*iriz -ulaticr. will sattsry veu: Should you nave furtner questions, pi ease or memcers ZZ ZL*f * ^ f n O *\ 0 1 S3 | r i] C.S. DfiP.-UV ' 'l wn .vs nr l.\30P. CiSJdiii wti'Wl* W.vSHlN, rc.N. n 3t.w i- : sis n Mr . A. 3. Hrooxaan, Manager Research, Dev sic ssten t and Decnnical fir estene ?lastics Company ?o ttscswn , Pennsylvania 1346 4 Dear Mr. Scocknan: In rasccnsa to your ieccar of for rr.cd iiica tier*, of She Vinyl Chi or ida 3 following decarninaticns a v a sear, cade: . r-a i 29 era 1510.1217;!:) ( 2 ) Definitions In regards to ode defiritcr. a: "r.assive release" in ? Directive 5200-33, we agree that the da f initisn 3noul .icdiiiad. 'inis will be addressed in a future Ovograr: tive. In ail probability the stipulation of ICO sot: i be rasoved. 29 C?S 1510.1017(b)(5) Definitions a [j [] : n. t iu n w no c o C3 & I aowav r, the reveals exposure at or above the action isvs: , one sits.oyer .oust marten: o r ocecuc a s specified in the standard. r--> 3. 29 C?Z 13131.1327(0) (4) (iii) Aespiratory protection .I " Secardi.no your suggestion to add a Type C, Supplied Air Aespiracor, Pressure lemand type, with roll or halo racer piece to this section: Id an employer can show that a respirator provides equal or greater protection,than those specified in the standard, t he aav he granted an interim order or a variance from the standard. Tour company received such an interim order, dated hay 30, 1973. I 4 . 29 C79. 19LQ . 1017 (:< ) Medical surveillance Theca is no CSKA regulation requiring an employee to submit to a medical examination. If the employee refuses any medi examination required to he provided hv ohe employer, the employer shall inform the employee of the possible health consequences of such refusal and obtain a signed statement from the employee indicating chat the employee understands the risk involved by refusal to he examined. We gtaatly appreciate your sharing data, experience ar.d Knowledge*with us. At the present time there are no plans to formally amend ar modify the vinyl chloride standard. We hope that the above clarification of the regulation will satisfy the request in your petition. Should you have fur: cuestior.s olease do not hesitate to contact us. " /% a Ccrr7 T VAlxar l,V2S/75 - - . \r-- Casslev, Jr. Arnold \r ' /a e -- 0760083 J . JOSEPH . SELLER jeroxe a. szcexar csarlee a. xzehav VTIiULM 3. 3QRGHESA.VI. JR. ROBERT 3- TSESRaR watre v. 3 lack Oavho l. hill xartir w. sercovxc: JO HR 3. XIORIS carole c. Harris XICHAEL ?. XORROHZ JARRE 3. SOLOMOR johr a. dubecz CHRIS7IRE A. XEaGHER SH1RLXT 3. rsaiXOTO PETER L. 3B 1a CRPZ * LAWRERCE ? 3ALPSIR OZBORaH shcr trireer C. DOUGLAS JaRRSTT sowars l. xobwez ROBERT L. PLESHRER JORATHAR ?. 1EVXRE SHEILA A. XILLaR * ossa sa> atm law omen KLI2H XMD H2C2MA^ 1130 17** STUBS" Jf. w. STIITS lOOO 'VASKI^GTOy, D- C. 20036 May 11, 1981 ?2U?SO$(Z aoa-4j7-uoo C.OL2 AJ30RZ33 ,,iri.'<.^V waresjts atJtzcT xtjkssx 202/457-1116 Mr. John R. Lawrence The Society of the Plastics Industry, Inc. 355 Lexington Avenue New York, New York 10017 Re: Employee Training Under 0SHA Vinyl Chloride Standard Dear John: The purpose of this lepter is to inform you of the recent completion of a case involving the employee training provisions of the Occupational Safety and Health Administra tion (0SHA) vinyl chloride standard. As you may recall, in September, 1973, an 0SEA inspector issued a citation to Hooker Chemical Corporation. The citation noted Hooker's alleged failure to provide training on vinyl chloride for employees who worked in the calendering, compounding and warehouse areas. Hooker contested the citation for its Burlington, New Jersey facility on the oasis that employees working in these areas were not required to he trained on the hazards of vinyl chloride. Both Hooker end OSHA's monitoring of vinyl chloride in the calender and compound facilities indi cated concentrations significantly helcw the 0.5 parts per million, (ppm) action Level. no c o ft Mr. John R. Lawrence May 11, 1981 Page Two KeZUTH JLND EmC30sLAOT In May, 1979, the administrative law judge issued a decision that adopted Hooker's position and vacated the OSHA citation. The judge held that the training provisions of the standard do not apply to areas where the vinyl chlor ide level in the ambient air is below the action level of 0.5 ppm. Although not requested to do so by any party, a commissioner on the Occupational Safety and Health Review Commission directed that the case be reviewed. In respond ing to the review proceedings, OSHA changed its position and filed a letter stating that the judge's decision should be affirmed. Significantly, OSHA stated that the training requirements were not applicable in this case because Hooker's compounding and calendering operations simply were not processes that could result in hazardous exposure to vinyl chloride. Because no party sought review by the Commission, and the Commission did not consider the issue one of com pelling public interest, it chose not to review the case. Thus, under the Commission's order of March 31, 1981, the judge's decision remains intact. While the case has limited precedential value, it does clarify what is a regulated area and when training requirements must be met. In addition, the proceedings indicate that OSHA is unlikely to issue citations in circum stances similar to that which 'existed at Hooker's Burlington facility. Because the conclusion of this litigation may be of general interest, you may wish to further distribute this latter- to the PVC Safety Group. As always, if you have any comments or questions, please feel free to contact me. Cordially yours. Peter L. de la Cruz co: Mr. Jerome ?. Carroll Ms. Fran Lichcenberg Mr. Thomas R. McGrath EXHIBIT VIII BFG TECHNICAL DOCUMENT A PHYSICAL MODEL FOR THE DIFFUSION OP TDTCL CHLORIDE MONOMER FROM FVC UNDER VARIOUS CONDITIONS OP STORAGE by H. M. O'Hara L. B. Crider R. L. Bowles C. J. Tamanek 3.?.Goodrich Chemical Company Avon Lake Technical Center P.0. Box 122 Avon Lake, Ohio 44012 August 25 1975 Abstract An empirical model, baaed on large scale laboratory experimental data, has been developed to show the relatlonahlp between the amount of residual vinyl chloride monomer (RVCM) In F7C resin and the concentration of atmo spheric vinyl chloride monomer (AVCM) under various conditions of storage. The variables studies in the experimental work leading to the development of this model include temperature, time, ventilation rate, RVCM content of the resin, mass/volume ratio and resin type (varying in porosity). The results obtained from a statistically designed series of 24 exper iments show the interactions of all the variables and permit (1) the effects of the variables to be established and (2) the derivation of a model which permits the AVCM levels to be predicted for any combination of the variables. The developed model has been used to predict the AVCM level is SFG warehouses at Avon Lake, Louisville and Pedricktown. We have found excellent agreement between predicted and measured AVCM levels at these three locations' and under a variety of storage conditions. Additionally, che model has been used to show chat--under- less-chan the most severe storage conditions, the OSHA action level will not be exceeded (500 ppb AVCM) when the RVCM content of the scored resin does not exceed 18 ppm. Even "under the most extreme storage conditions the OSHA action level is not exc'eeded .when the RVCM content .o^thei'resiii^S^^j^-ppm or-less no i, * r Table of Contents I. Technical Objectives rage 1 II. Introduction . . . . III. Statistically DesignedStudy ............................................................................. 3----- 17. Experimental ..... ...................................................................................... 3 (A) Description of Model........................................................... 3 (B) Analytical Methodology .......................................................................... ~SHI (C) Theoretical Diffusion of 7CM in Physical Model . . . . .------- 11 7. Diffusion of 7CM Proa PTC Basins in Model . . . . ... . .-- 15 71. Results and Conclusions ..................................... .......... IS 711. Acknovledgaents.............................................................. . 26 Till. Bibliography ......................................................................... . ; ~ 27 XX. Appendix........................................................................................................................ 28 O s 3 t 1 SJ.W. _________________________________ __ ^ ___________ ^ A-VINYL CHLORIDE FREE NITROGEN PURGE GAS INLET B-55 GALLON DRUMS HOUSED!IN ENVIRONMENTAL CHAMBER C-SAMPLING SYSTEM D-GAS CHROMATOGRAPH WITH SAMPLING LOOP (1) n2 carrier gas (2) AIR (3) HYDROGEN I--------------- 1 * Section A Table II Basic DescTiotlon of Physical Model (Safer to Figure 1) Description inlet nitrogen purge gas; maxinmn flow ~ 20 liters/minute B environmental chamber containing ten air-tight 55-gallon metal drums C 55-gallon drum sampling system D flame-ionization gas chrooeeograph/daea handling system I A - Nitrogen Purge Gas Inlet System B - 55 Gallon Drum* in Environmental Chamber 'O T O fK .O gt I n n f ,a } n o 0 II L. D Gas Chromatograph with Sampling System n p 10 n ;i Table III I Gas Chromatograph and Data Handling System Used to Monitor AVCM Levels ? j i Gaa Chromatograph System - Hewlett-Packard Model 5711 Flame Ionization Gas Chromatograph Method 10' X 1/3" Porapak Q, mesh 30/100; Isothermal at 150*C; nitrogen carrier gaa at 30cc/minute; detector - 250*C; injection port - 150#C Gas Sampling Valve - lOcc sample loop Da;a handling System System - Hewlett-Packard Model 3380 Integrator-Calculator *3 )$U): TT Calibration of the system was accomplished through the use of comaercial.lx available standards (Precision Gas Sampling Corporation) and dynamic dilution of those standards. A typical calibration (checked periodically) and a regression analysis of that data is sumaariaed in Table 17. This analysis' Indicates good linearity over the range from 128 * 2000 ppb of 7CM in the atmosphere. The slope from the linear regression analysis was used as input to the integrator/calculator in order to calculate all data. It is important to note that as the sampling proceeded to a range outside of the calibration range shown in Table 17, new calibrations were carried out to reflect these different ranges. At no time was a calibration extrapolated beyond the exper imentally determined range and used as a basis for analysis. (C) Theoretical Diffusion of 7CM in the Physical Model It was necessary to check out the performance of the system in terms of theoretical diffusion. The main reason for doing this was to provide a final and independent check of the model, the sampling system and the analytical system interacting together. It was further felt that this approach would lead to a definition of any 701 system absorption (or leakage) problems. The ab initio derivation of the mathematical model which describes the build-up of a gas in a dynamic system is described below. if: Go 3 concentration of incoming air scream Ct 3 concentration in drum at any time (t) 7 3 volume o f drum M 3 flow into and out; of drum t S minutes c f o 12 Table 17 VCM Calibration of Gas Chromatograph Calibration Data VCM Concentration ppb 128 423 623 322 1109 2292 Integrator/Calculator Response 2030, 2024, 2013 6819, 6847, 6846 9736, 9755, 9880 12331, 12278, 12479 16703, 16336, 16800 34740, 34910, 34841 .. Linear Regress*rm Analysis ppb * a (response) - 12.36 a = 0.0663 .0004 correlation: 0.9997 - 13 - r' VCM into drum * 7CM out of drum + accumulation ^Ggd^* * Vdc 7 Vdt MC0 - MCt Ve 7 <Co * Ct) Co - Cc M/7 dt Cl) integrating this over the limits from 0 -- C and 0 -- t yields: - n (C0 - Ct) * M/7(t) + A f i. or j Ct . C,, - A.-*TM r! at t o, C a - A Co . * Gt - C0 [1 - -M/V<t)j In essence, equation (1) describes che build-up of VCM- in a 55-gallon drum when VCM at a concentration of CQ is flowed into the drum* at a rate equal to M. At any time * t, the concentration of VCM in the drum is Ct. To test this equation, a 1 ppm VCM standard was flowed into a drum (216 liters) at a rate of 1.82 liters/minute. It can be seen from Figure 6 that there is not a satisfactory agreement between the theoretical and che experimentally deter mined build-up of VCM in Che model. It can be seen frcxa Figure 6 Chat for any concentration of AVCM, che time increment between the two curves is approxi mately 40 minutes. This induction period was analyzed to be a diffusion *Flow into che drum and flow out of the drum is equal. VCM, ppb 0 14 Figure 6. THEORETICAL vs. EXPERIMENTAL BUILD-UP of VCM in 55 GALLON DRUM H 15 - problem within Che large drum. This inherent analytical flow was corrected by installing 3 inch mixing fans in each of the 55-gallon drums. The data presented in Figure 7 was obtained after this experimental modification was made. This treatment unequivocally demonstrated that- VCM could be monitored in real time as it diffused from PVC resins under controlled ventilation con ditions in a dynamic closed system. It also demonstrates the validity of the entire sampling methodology and VCM standards. -- - V. Diffusion of VCM From PVC Resina In Model This section contains a suumary of all diffusion experiments carried out in Che statistically desi&ied study. More detailed information (concen tration/ time data, etc.) on all experiments can be found in the Appendix. The purpose of this section is to provide a rapid means for comparing the various experiments carried out in this study. In order to make this com parison so that order of magnitude trends can be seen, the following data is provided: (1) resin type (A, B, C . . .) (2) residual VCM in resin (RVCM) (3) exposure temperature (*C) (4) total exposure time (hours) (5) ventilation rate (turnovers/hour) (6) resin mass/storage volume ratio (lbs./ft.3) (7) maximum AVCM level reached and time of maximum (ppm, hrs.) These data are summarized in Table V. O f'f A* 0 O 4 'y 16 rn :J Figure 12. AVCM vs. RVCM and WAREHOUSE LOADING AVCM (ppb) 25 - Figure 13. PREDICTION of AVCM in WAREHOUSE SAMPLES PREDICTED AVCM (ppb) / -26storage conditions varied in temperature, ventilation rate, loading and KVCM (all were approximately at one month's aging 720 hours), Che model predicts the different AVCM results quite well. The somewhat low predictions can probably be explained by the fact Chat the warehouse samples were bagged (slowing the VCM diffusion slightly) whereas the model was based on material being scored in bulk fora. The accuracy of the model is 100% in being able to predict the "true1* AVCM from a set of storage conditions and resin possessing a certain HVCM. Acknowledaments The authors gratefully acknowledge Che considerable contributions of H. T. Kim for his assistance In development of the experimental design and the development of a mathematical model evaluating effective diffusivities of VCM in PVC resins; to A. R. Serena for consultation during the course of this work; to K. R. Rlcland for assistance in the statistical analysis of the data and to a large number of people in BFG Manufacturing who cooperated in providing large quantities of the specific resins required for this study.. CO Bibliography (1) A. &. Berens, L. B. Crider, C. J. Tonanek and J. M. Whitney, "Analysis for Vinyl Chloride in PTC Powders by Head-Space Gas Chromatography" to be published in J. of App. Poly. Sci. (2) A. R. Berens, Polymer Preprints 15 (2) 203 (1974). 28 IX. Appendix Table 7 of this report provides a aumaary of all experiments carried out In order to develop the empirical model. The actual tine/concentration data for all experiments is contained in this section; experiment numbers refer Co those in Table 7. All data is sunsaarized in Table 71. All data is contained in Data Book 001*95 (4/28/75). 20 vo: yr- !i ~Hi ,n r : i d i cw -J H : . ik0`N^<A**'4rt*4r<<>A4ro4>rO4Q9<OdOflO4 "I n <o a wto a <* o <* a o o :a 8 k -Jo U 4 +* o 9l^f<049iAt0NC9dNiA ^N'flONinM,,^N^ <AiAMT i o i ?y cj a u n 0>4Nn^<Oc^>di>N 4 mO 4 N<0> <Qn ^O Nrt ooaoooooo VV tftQw"lQwiQQQO ONMINNin^<n<0 i/tOirtO OOOOOOO ........................V V "1 2o i/>o m o m a OCOOO'dOOO 900000000 vvvvv ` vvv tAOtftO>Ow\Olrt **0*%Qv"i0rt0w>QQQ0 g*4N<<nn^<v>4Nc9 - 29 - N 8 N rt N ^ 404oN<vNrot^<<n0N4O OOQ^^i^f'tOirtO 4nW (Mn^rtO^fOr 1 O r? O (*w> <o< O 0 Nd^ONO< 9*49mN*<0000 <n ** o <o OO<n NN>^*i0m*N4 ifl n < I r v> O Q O O >< *4 p4 ( I N404p44#4rfirf (ni^<nOQmf4MiA gj 0 4S0^H^nNin4 HMm4NQin<3t(89 ^ MmoOONmnMft 4isnQnw>*f N nn^ ! c* o j r*. Ovinv^ 0Q^4(nrt0mc40 NrtOflc>irtiflr'Nirt nncBOM^in^ Nwnn*/ i 00m4<N0ntn;^rt<nn*0A44NnOin i OinnnrtrtONOO 4 rt H N N l a& jo Ha 207C012S 8 Qm Ol JI 04'^O'dN<"<>NirH(N40>Qr<a4ri,%0N^4H^w^ nnOxivnintfO Orta) I A ^ <** ^ GO ? * <*4 1* . Q f N 1^ O D >40 O N ^ (N O 000<rt<^<n0x% r A MODEL FOR THE DIFFUSION OF VINYL CHLORIDE MONOMER FROM PVC UNDER VARIOUS CONDITIONS OF STORAGE limvood B. Crider Michael M. O'Marz Rohm L. Bowks 8fGoodricft Chemical Division Avon Lake Technical Center Avon Lake, Ohio 44012 \ i i t ! . 'I .1 ,H :b >i INTRODUCTION within the pest several /ears significant advancements have been made within the PVC industry to minimize em ployee exposure to vinyl chloride monomer (VCD . Im provements in PVC manufacturing processes eo reduce the residual vinyl ehlorid# monomer IRVCM) content of finished resins has been a key factor in maintaining acceptable VCM levels in work areas during packaging, processing, shipment and storage. The diffusion of vinyl chloride monomer out of finish pvced resins has been adequately studied at tem peratures above the glass transition taoperature (Tg). The results of thsse studies have been particularly useful in reducing RVOt during manufacturing and to predicting RVCM escape to the environment during pro csssing. This new knowledge has also led to a rapid. simpLe gas chromatographic methodl** for the determin ation of RVCM in PVC from the analysis of the vapor phase (head space! over pvc powders in a closed con tainer. This somewhat ideal behavior does not exist at temper.aturss below Tg, howsver, and consequently one cannot use axisting equilibrium data to predict the concen tration of VQt in the air space above PVC resins and in particular under non-static conditions such as ex ists during the storage and transportation of bagged or bulk resin, tn addition to the relative slow dif fusion rate ae or near ambient temperatures* other factors chat affect VCM release include variable ven tilation rates# mass-to-voluma ratios, RVCM eoneene of the resin and residence time in the storage or ship ment compartment. The primary objective of the resaareh reported in this paper was to develop a physical model for the diffu sion of VCM from PVC resin under various conditions of exposure during storage and shipment. The developed model is based on a statistically significant number oe large seals diffusion experiments to snow the in teraction of ail the above variables and to allow the prediction of atmospheric vinyl chloride monomer (AVCM) levels for any combination of the variables. EXPERIMENTAL It can be postulated that some reasonable understand ing of these storage variables and their interactions can be developed from a pnystcai modeling of a stor age or shipment compartment. It can also te rational ised that the accuracy of such a modal will be much improved if the experimental design is on a large scale that can onysically simulate all of the storage variables but under highly controlled conditions. Any experimental design of this type, however, must be thoroughly tested to eseablisn .the validity of the test measurements and to assure the absence of any errors as may occur through poor test controls, adsorption of'/CM by the system, and possible interfering components chat could lead to spurious results. Using a Computer Optimised Experimental Design CCOEO), 24 large scale experiments were selected to study the variables listed below eo describe their effects and interaction on AVCM in a storage compartment or ware house. The variables and ranges included: 1.) Temperature - 74-iiSr 2.) ventilation Rate - 0.5-3.0 Tumovers/Hour 3.) Loading (PVC Voiuae/Storaga Volume! - 11-34% 4.) Resin RVCM - .01-123 ppm 5.) Storage Time - 1-170 Hours In order to meet the demands of the statistically de signed (COED! diffusion study# ehe experimental phase of this program had to be capable of controlling ehe following experimental parameters: 1.) Temperature 2.) Ventilation Rate 3.) Mass/Volume Ratio The basic physical model that was chosen to meet these requirements was one based on sealed metal storage bins (SS-gailon drums! housed in an environmental chamber where the temperature could be varied and controlled. A schematic of the entire experiment is shown in rigure 1. The schematic is divided ineo four key areas: (A) the ventilation system, (R) the environmental chamber* (Cl the sampling system and (0) the gas chro matograph with data handling system. Clean, dry ni trogen (A) enters the environmental chamber (B) through a spile manifold. The ten drums were piped into the two manifolds through ten flow regulators. In this way, veneilation rates could be independently set in each drum. The exhausted atmosphere from ehe drums was directed eo a sampling system (Cl that provided a split stream: one for vent and one for sampling. The sampling scream was intarfacad to a gas sampling valve/ gas chromatograph/data handling system (0). Sequential sampling of eacn of the drum* was possible with this system. The environmental chamber is a Conrad Model WD--624 Temperature Humidity Chamber with an internal volume of 938 cubic feet. Temperature control inside the chamber was * 2C; a maximum exposure temperature of 10QC is possibLe within the chamber. Calibration of the system was accomplishsd through the use of commercially available standards (Precision Gas Sampling Corporation) and dynamic dilution of those standards. A regression analysis of the calibration data indicates good linearity over the range from 1232000 ppb of VCM in the atstosphere. The slope from the linear regression analysis was used as input to the integrator/calculaeor in order to calculate all data. It is important to note that as the sampling proceeded to a range outside of tne calibration range, r.ew cali brations were carried out to reflect these different ranqes. At no tice was a calibration extrapolated be yond the experimentally determined range and used as a basis for ^alysis. It was necesiary to check out the performance of the system in terms of theoretical diffusion. The primary reason for doing this was to orovide a final and inde pendent check of tr.e model, the sampling system and the analytical system interacting together. It was further felt that this approach would lead to a defi nition of any VCM system absorption (or leakage! proa- *0 O > 03 lcm*. The ab in itio derivation of the mathematical model which describes the build-up of a gas in a dy namic system is described below. If: 0o Ce V Me then: VCM into MCode m Vdc/dt m dc/de m do Co - ce concencraeion of incoming air scream concencraeion in drum it any time (el volume of drum flow ineo and ouc of drum minutes drum * VCM out of drum * accumulation MC t * Vdc MC0 MCr M (C0 - <=t> V M/V de id integrating c hi s over che limics from 0--and 0--^ yields: - n (Co - Ce ) - aM/V(t) * A Ce,; e t o, c <=t - * -H/V(t> c0=o o,& - In essence. equation (1) describes Che build-up of VCM in a 55-gallon drum when vcm at a concencraeion of Cq is flowed ineo the drua ae a race equal co M. At any ciae e, the concencraeion of VCM in che drua is Cz. To case this equation, a 1 ppa VCM standard was flowed into a drua (216 liters) at a race of 1.32 litecs/min- uee. The daea from this experiaene is presented in Figure 2 and shows a satisfactory aqreeaene between the cheoreeical and experimentally determined concen tration in the modeling systea. The aneiyeicei methodology used to daeermine the VCM in the exie vanes from the drua experiments is as fol lows i Gas Chromatograph System - Hewlett-Packard Model 5711 Flame Ionization Cal Chromatograph Method - 10* X 1/3* PorapeJr Q, aesh 30/100: isothermal ae 130C; nitrogen carrier gas ae lOcc/ainute: detec tor - 250C: injection pore - 150C Gas Sampling valve - lOce sample loop Data Handling Systea System - Hewlett-Packard Model 3330 Integrator-Calculator RESULTS AMO CONCLUSIONS The results obtained from the 24 COCO designed experi ments and a multiple correlation analysis of the re sults permitted 1) the effects and interactions of the variables on AVCM to be established and 2) the deriv ations of a modal which permits the AVCM levels to be predicted for any combination of the variables studied. The resulting model for PVC resins with a porosity of 0/2 is: . . ,,l. V-1.M * Mm "'..le-*"**1* "j where: AVCM atmospheric concentration of VCM in parts per billion Ro initial RVCt of the resin in parts pec million L 3 loading: (resin volume/warehouse volume) X 100 V ventilation rate in turnovers per hour T - temperature in F t time in hours As expected. AVCM decreases as: 1.) RVCM of the resin decreases. 2.) Volume of resin being stored decreases. 1.) Ventilation rate increases. 4.) 5.) Storage temperature decreases (for time periods of 1-9 days); however, if the resin has been stored for longer periods of time (10. days), then che AVCM is lower at higher stor age temperatures. (This is che cumulative result of a higher temperaeure causing mors rapid diffusion at the beginning of aging; therefore, more rapid depletion of RVCM and, chus, lower AVCM at longer agings.) Storage time increases. Using the relationships described in the model, maxi mum RVCM resin levels can be determined which, will assure that AVCM levels will be less than 500 ppb OSHA action level for a variety of storage conditions. (See Figures 3-7 which show the effects of temperature, time, ventilation rate, loading and RVCM on AVCM.) When PVC is stored under the least severe conditions (Temperacurs - 30F, Ventilation * 0.5 tumovers/hour 304 loading of a warahouse volume with VCM containing material, the RVCM of PVC resin must be no greater than 18 oom to assure chat the action level is noe exceeded at 5 days. If che 500 ppb (OSHA action level) muse be met even un der the most adverse seorage conditions (Temperature 120F, Veneilaeion 0.5 turnovers/hour, 804 loading of a warehouse volume with VCM containing macerial), than the resin RVCM muse be no higher than 8.5 oom. Figure 8 shows how well the model predicts the AVCM levels in 3FCoodrich warehouse samples obeained in April, 1975 ae Avon Lake. Ohio, Louisville, Kentucky and Pedriekeown, New Jersey. As the seorage conditions varied in temperaeure. ventilation raee, loading and RVCM (all were approkimaeely at one month's aging 720 hours), the model predicts the diffareae AVCM re sults quite well. The somewhat law predictions can probably be explained by the face that the warahouse samples were bagged (slowing the VCM diffusion slighely) whereas the model was based on material being scor ed in bulk form. The accuracy of the model is t. 1004 in being able to predict the `true* AVCM from a see of seorage condi tions and resin possessing a certain RVCM. ACKNOWLEDGMENTS The authors gracefully acknowledge the considerable contributions of H. T. Kim for his assistance in de velopment of the experimental design and the develop ment of a mathematical model evaluating effective diffusivities of VCM in PVC resins; to A. R. Berens for consultation during the course of this work; to C. J. Tomanek for assistance in most of the experimental work; to M. R. Ritiand for assistance in the statisti cal analysis of the deea and to a large number of peo ple in 3FG Manufacturing who cooperated in providing large quantities of the specific resins required for this study. BIBLIOGRAPHY (1) Berens, A. R., The oiffusion of Vinyl Chloride in Polyvinyl Chloride*, ACS-Oivisioa of Polymer Chemistry, Polymer Preprints, IS (2):2Q3, 1974. (2) Berens, A. R., Crider. L. 3., Tomanek, C. J. and Whitney, J. M., `Analysis of vinyl Chloride in ?vc Powders by Head-Spacs Gas Chromatography*. J. of App. Poly. Sci., 19, 3169-3172, 1975. .OTTCGIS? 19S rr Figure I. A-VINYL CHLORIOE FREE NITROGEN PURGE GAS INLET VCM, ppb Figure 2. THEORETICAL vs. EXPERIMENTAL BUILO-UP of VCM in 55 GALLON DRUM AFTER EXPERIMENTAL MODIFICATION 1000 -- Figure 3. AVCM vs. RVCM and WAREHOUSE LOAOING x-''* x/* / -- THEORETICAL x - EXPERIMENTAL / loot I I I i ' I I___ L O 10 20 20 AO 50 50 TO SO SO 100 110 TIME (Min.) 117 o o \\ Figure 4. AVCM vs. RVCM and TEMPERATURE Figure 5. AVCM vs. RVCM and VENTILATION AVCM(ppb) AVCM(ppb) Figure 6. AVCM vs. RVCM and VENTtLATION Figure 7. AVCM vs. RVCM and WAREHOUSE LOAOING O /w 19t rr Figure 8. PREDICTION of AVCM in WAREHOUSE SAMPLES Simula Lion ot. AVCM Concti itralion During Transporter ion of PVC 4/7/83 Pane 3 2. Analysis of Model Variables (Cont'd) 2.3 'L'ruck load inn (L) Equation (T) shows chat, as might be expected, increased truck load inn produces higher concentrations of AVCM. In discussions with transporta tion experts the assumption was made that 80" loading is the maximum that could be expected. In this study, therefore, the truck Loading was set at 80%. 2.4 Ventilation rate (V) Ventilation rate as measured by air turn-over/hour cannot be readily ascertained for a truck since it is dependent on numerous factors e.g. truck's material of construction, design of truck, speed at which truck is traveling,and climatic conditions. However, an effective air turn-over rate can be calculated from Equation if actual AVCM data measured inside the truck were available. Kruszynski^'^) an(j gmith(4) have reported the AVCM concentrations inside the nose of a trailer immediately upon opening the doors, as well as immediately after a trailer has been unloaded. As can be seen from Table 1 AVCM concentrations did not exceed the action level limit for those cases where the samples were taken upon opening the doors. However, data where bag was broken and resin exposed did exceed the limit. Since this is the highest AVCM concentration reported, it was used in calculating the effective air turn-over rate. As shown in Appendix 1 the minimum effective air turn-over rate is approximately 6/hour. However, to eu-ure greater safety margin simulations were carried out using 3 air tumovers/hour_ 3. Discussion of Results With the variables fixed at levels given in Section 2 , Equation ^l) can be solved to predict the resin residual VCM (RVCM) which would produce AVCM concentration below the action level limit of .5 ppm. However, one of the var iables, the transportation time, cannot be ascribed a definitive value, and a range of values (4-72 hours) were assumed. Fig. 1 shows the relationship between the resin RVCM and the AVCM concentration inside a truck at 4, 24 and 72 hours. As can be seen from Fig. 1 the AVCM concentrations do not exceed the action limit of .5 ppm provided the resin RVCM content is below 10 ppm. It should be emphasized that Fig. 1 has been predicted using conditions which deliberately maximize the AVCM concentrations inside a trailer transporting PVC resins. Table Simulation of AVCM Concentrations PurinR Transportation of PVC 4/7/83 Pap.e 3 807. 3/HR. 90 F FIG. 1: RELATIONSHIP BETWEEN RESIN RVCM AND THE AVCM CONCENTRATIONS INSIDE A TRUCK AS PREDICTED BY O'MARA ET AL MODEL^1) . C3 ^* TT Simulation of AVCM Concentrations During Transportation of PVC 4/7/83 Page 6 APPENDIX 1 Calculation of Effective Air Turn-over/Hr. Equation 1 (see page 2) can be rearranged to allow the calculation of air turn-over rate (V), viz, V1*56 -(Rj*47L'3/AVCM){|T6ex(-24.1 ,0031Tt)| + .2 exp(-.000116Tt)} ----- 2 As mentioned in Section 2 temperature (T) and truck loading (L) were deliberately set at high values so as to maximize the AVCM concentration and, hence, ensure greater safety margin. Thus T is assumed to be 90F and loading L is set at 80%. The selection of AVCM data to be used In the above equation is made from Table 1, page 4. The criteria for selecting AVCM data are, (1) whether conditions under which AVCM data given in Table 1 were measured were similar to the conditions used in the development of the model, and, (2) ensure that chosen data would minimize the calculated air turn-over rate and, thereby, increase margin of safety. Based on this criteria, AVCM value selected was 1.4 ppm. However, in order to use this value corresponding resin RVCM (Rq) is also required. Since this was not recorded at the time AVCM data was taken, an estimated value of Ro had to be used. Based on statistical analysis of historical resin RVCM data Haller and AhmedO) have shown that in 1974 BFG PVC resins contained 160 ppm of VCM (95% confidence) and 260 ppm (99% confidence). Table 1.1.summarizes the calculated effective air turn-over rates using these values. Table 1.1 Calculated Effective Turn-over Rate Using 1974 AVCM Data Resin RVCM in 1974 (PPM) 160 160 160 260 260 260 Transportation Time (Hours) 4 24 72 4 24 72 Calc. Effective Turn-over Rate/Hr 20 14 6 30 21 9 As can be seen from Table 1.1 the minimum air turn-over is calculated to be 6. Since the lower value maximizes AVCM concentration, it is preferable to use air turn-over rate of 6/Hr. However, to build an even greater safety margin it is desirable to reduce this value even further. Accordingly^/"$;he'TObs-used gubse- quent calculation is 3/Hr. ( 6/2). icoxoi Simulation of AVCM Concencrat ions During Transporcat ion of PVC 4/7/83 Page 7 References (1) O'Mara, M. M. et al. Journal of Vinyl Tech. Vol. 1, 3, page 168 (1979). (2) Kruszynski, R. J., BFG IOC, Dept. 5401, 7/8/75. (3) Ibid, BFG Customer Service Report, Proj. No. 3689, 12/20/74. (4) Smith, T. H., BFG Customer Service Report, 1/3/75. (5) Ahmed, M. J., Haller, H. S., BFG Technical Document, 3/31/83. MJAjHSH:vmt Distribution: P. Dunnigan R. D. Hardesty R. M. Kreager (first page only) R. J. Kruczynski L. P. Tenney CTF M. J. Ahmed H. S. Haller