Document 3Q4yOZp8bLe3m4y4V42Z2MxyD

Part 1 - Experimental Methodology (excerpt) ETHYL CORPORATION RESEARCH AND DEVELOPMENT DEPARTMENT BATON ROUGE, . LOUISIANA EXTRACTION OF VINYL CHLORIDE FROM PVC BOTTLES BY THE BOTTLE CONTENTS January 17, 1975 Authors: G. A. Daniels D. E. Proctor SPl-02842 Experimental Program An essential part of the experimental program is the analytical methods for VCM. These methods utilize flame ionization detector gas chromato graphy and are included as Appendix II. One method is used for analyses of vinyl chloride in VCM-containing liquids and another for determining the VCM content of PVC in the form of powder, pellets or bottle walls. The method for VCM in liquid measure ments involves the use ox a long, small diameter column which provides adequate separation of VCM from other components of commercial products and solvents. The use of this method allows detection of as little as 10 ppb VCM. The penalties for using this type column are long elution and bake-out times between sample injections. The method for VCM in PVC is an adaption of the prior method and involves injection of THF in which the PVC sample has been dissolved. A quantity of VCM corresponding to about one-quarter of a ppm in the dissolved sample can be detected. However, variable baseline interferences at the 1 ppm level adversely affect the accuracy. Tables I to VI present the experimental data. The initial experiments included the use of standard FDA-recognized food-simulating solvents. A VCM level of 100-400 ppm in the compound was used to provide sufficient VCM to minimize analytical error. Table I shows the various solvents used; also note that two brands of vegetable oil are included in this series. Tins experiment was begun January, 1974, in response to an FDA request for room temperature storage VCM extraction data and the study continues. In April, 1974, 2-ounce capacity bottles made from <1 ppm VCM PVC were added to the series. The two-ounce bottles were used for several reasons: (1) to maximize VCM level potential in the product so it might be detected, (2) to reduce quantities of solvent used in migration testing, and (3) to allosv filling of several bottles with each solvent so any particular bottle need not be resampled. Table II is a table of results from another VCM extraction series begun in April. Both 16-ounce and 2-ounce bottles are used for storing weak VCM solvents in this series. The only real value of this series is to illustrate erratic results probable when using weak VCM solvents and poor sampling procedures. Largely on the basis of this experiment, methods were considerably altered. It was decided that subsequent experi ments would involve using only strong VCM solvent (50% vol. ethanol) so that losses in sampling would not occur. Also, a procedure of filling numerous two-ounce capacity bottles was begun so that repeated sampling of the same bottle is unnecessary and so that duplicate and triplicate samples could be submitted for VCM analyses. The remaining results in Tables III, IV, V, and VI are for 2-ounce bottles with 50% ethanol. SPI-02843 8 Room Tem perature VCM M igration - 72*F E th yl 8237 O ctyl T in Stabilized PVC Compound r 0 0-- N0 CM O O <M O CM O e- in in M4 -- so dm d O co d ^ 0 v> d 2 o > S O IT) ^ OvO OOOOh* o' V"do V" d o O'"* ^ O' I v1 i--n -- in iOn' -- 00 h O N O O * o * V w Pi O OO0' 'OO0' 'HO^OC'OOhO*OO CO O* v* CM* V* M< V* cm o V co* o V cm' o V cm' v 32 n- ** cCcooO rcoo CO CO CCOO CcoO CO CO fc 4> uu VV u V r- "2 >-l *-< "5 --1 --I S gs g g= S rt .5 << < .5 rt rt ,-- Or^OOSOO^O u oo o>.oo>oo> yw o d > rt o o> H0N.M0M0M.M0M0N.M0M0M.N0N0N.N0M0M.M00MM o_ O N o o O (M o-- o-- O M o-- o O (M oo O CO oo Oo _o _ O N d w-> r-l .-I <* <; _i.-i.-<000^ V oR oC oR crj crtCrt nuy* wo .ac Jrci ifl x** J~ ^-- ^ m WWWWW W < < < aaa'rt rt rt ._____ M OOO I I 1 0 0 0,0" o ICiCnooooinininnfon C C C 1(1 l/l (/) U U U nV n Uii E Xyo ri C <o rmi d *> cc O E& DV I I J_S M P o ca w a4) 00 (4 1 t o O 5co 5.2^2 rt t^o o 6 ij V oE. E V *H&- i "5o y 1Ec-4 w *, ov 5 rt o + AO f*H dCoO v SO .O 75 o'"S s5 5w 5*> *R. - o* E rnt >4 o u c(o4 VcUo a >u M4 o >P4 o ^3 co vcOo v r>t4 .I *s XryVj TV"J .--.. V trot Mo Xy(j o E" v o 60 2a *h -fpi T3 r*H Ou dc6o v d J81" E 5S > o MO 25 S cS ^33 O y (0 'O- TJ "* > oo * <ei* * H Og ? .5 coo o o CM ^N CO r4! CMfOlAsDt,,,,0, `^CO^li, t,-`*^fMCOlA^Dfs" to SPI-02844 4 > -I 4 -i * lr r r 19 < 4 -4 j' 80 t ( o 0*-0< o CO O fO <SJ (N3 CO o m O' ro o in o r- rt ** 00 o o C4 n *< o o nj cn O I H sO O o to o o o CO eo in o r- O' (SJ o & --4 V o ro rsi o o rt W4kO> RVPJ rsl (O vO r O' ^4 t- oo sO ro CO O o o ** ro O -rt oM Tf o O' <\J in H CO O' rt"4 ** nj H cn ro o (S3 f>3 o o o o O' <NJ tO o ro H ro ro r- o o rj r- sO o ^ ** ro in O' -oo rJ j=> o o o ^4 CP m nJ O ^4 nra rJ O' o rs) o O' O' rrttr<t ^*4 >o (VJ o o ^4 o 4rt rt-4 o o o ^4 ^4 Rw V o0 rJ o o rt . B w O ^ *> Pc o P< m r**4 _ 1] * ** S N IOQ.irOt Ort Ot'NOl'NOI^NOt'N O vO O' (s)<MNOOONfMIM g 0 rrt O rc4 (0a4 oart oart 0crt 1 iS fi 3 fi s * o (0 w w w w ww o N o N o IM o . IN om S!. Vi. ^ sc 3J K tn K S os a eo co co DO (4 RH 00 to> o ret R u o 4) O -4r)t crt 0v n? RV a o m v TL tro8t .8 g.5 I* 43 fl 2S CD rt 1 n v o n V urt *. 20o 'rRot* _*0, mo rt -M rt 6 t *o u T3 s2 Vy U 4- |"6 tsrt fi R S t* vi a 4 u? .oo '?p v H *(-Vi 2U rtS o? sn c " J- rst ,*2 O, .2 WM h> * N 9I~)t A ^Mf^,4<lA'OlwCOO*`0~*fs3 10 9. SPI-02845 I TABLE III Run 4562 120*F Accelerated VCM Migration - 50% Ethanol Ethyl 8237 - 82 PPM VCM - 2-Ounce Capacity Stock Cylinder 10. Storage Time 3 days 3 days 3 days 1 week 1 week 1 week 2 weeks 2 weeks 2 weeks 2 weeks 2 weeks 2 weeks 2 weeks 2 weeks 2 weeks 2 weeks 4 weeks 4 weeks 4 weeks 2 mos. 2 mos. 2 mos. 4 mos. 4 mos. 4 mos. 5 mos. 5 mos. 5 mos. PPM VCM Contents 0.49 0.49 0.54 0. 70 0. 82 0.66 " 1.01 1.37 1. 10 1; 30 1.08 1. 14 1. 13 1.05 1. 05 1.02 1. 73 2.09 1.96 3.84 3.58 3.51 4.28 3.78 4. 10 4.76 4.24 4.55 PPM VCM Bottle 65 56 51 36 37 29 26 31 13 14 15 (1) Standard vinyl pulp cap liners used. (2) No resampling of container, triplicate samples analyzed. SPI-02846 11. TABLE IV Run 4564 Room Temperature VCM Migration -50% Ethanol Ethyl 8237 - 101 PPM VCM - 2-Ounce Capacity Stock Cylinder Storage Time 2 weeks 2 weeks 2 weeks 4 weeks 4 weeks 4 weeks 2 mos. 2 mos. 2 mos. 4 mos. 4 mos. 4 mos. PPM VCM Contents 0.28 0.22 bad chart - 0. 88 1.13 0.76 1.25 1.00 0.90 1.43 1.62 1.65 (1) Standard vinyl pulp cap liners used. (2) No resampling of container. PPM VCM Bottle 97 97 94 SPI-02847 ! 12. TABLE V Run 4566 120*F Accelerated YCM Migration -50% Ethanol Ethyl 8237 - 302 PPM VCM - 2-Ounce Capacity Stock Cylinder Storage Time 5 days 5 days 2 weeks 2 weeks 2 weeks 4 weeks 4 weeks 4 weeks 6 weeks 6 weeks 6 weeks 3 mos. 3 mos. 3 mos. PPM VCM Contents 3.7 12. 1 (reject) 5.7 4.6 "4.4 11.0 11.7 11.25 14. 1 14.7 13.8 18.5 17.3 15.6 (?) PPM VCM Bottle 236 245 204 208 211 149 148 141 130 143 107 36 43 44 (1} Foil liner heat-sealed over bottle mouth. '(2) No resampling of container. (3) Chromatograph syringe needle breaks foil seal during sampling. SPl-02848 13. TABLE VI - Run 4567 1 Room Temperature VCM Migration - 50% Ethanol Ethyl 8237 - 304 PPM VCM Bottle - 2-Ounce Capacity Stock Cylind Storage Time 2 weeks 2 weeks 2 weeks 1 month 1 month 1 month 2 months 2 months 2 months 3 months 3 months 3 months PPM VCM Contents 1.77 1.70 1. 63 " 2.94 2. 84 3.00 5. 11 4.6] 4. 95 6.60 6. 86 6. 70 PPM VCM ` 267 270 272 236 251 237 (1) Foil liner heat-sealed over bottle mouth. (2) No resampling of containers. (3) Chromatograph syringe needle breaks foil seal during sampling. SPI-02849 14. Experimental Procedure The procedure that evolved from the succession of experiments can be briefly stated. First, a small batch of PVC molding compound is thoroughly mixed to insure that all bottles blown from it will contain like concentrations of VCM.' Triplicate bottles from each batch are submitted for analyses at the time the bottles are filled with solvent. The analytical values are averaged to represent the beginning VCM level of the test bottles. Immediately after each bottle is filled with solvent to its rated capacity, a foil disc is heat-sealed over the mouth of the bottle which is then capped to protect the seal. (However, experiments indicate that commercial closure losses are negligible when 50% ethanol is used as the solvent. ) The filled and sealed samples are kept at the appropriate temperature for the desired time interval. Triplicate samples are sub mitted for VCM analyses. The seals are broken only when the sample syringe pierces the foil. After sampling, many of the sample bottles are emptied and then analyzed for bottle wall VCM content. Experimental Data Analysis The experimental data shown by Tables HI to VI is particularly useful for analysis because two sets, 4562 (Table III) and 4566 (Table V) were run at 120*F, while sets 4564 (Table IV) and 4567 (Table VI) were run at room temperature. The bottles for data sets 4566 and 4567 were blown tiie day before the bottles were filled, while the 4562 set was filled 78 days after the bottles were blown, and the 4564 set was filled 118 days after the bottles were blown. The compound for data sets 4566 and 4567 was from the same batch, while 4564 and 4562 were from different batches of compound. An analysis of the compound for 4566 and 4567 was not run when the bottles were blown, but a set of bottles were blown in January, 1974, from the same batch of compound. In January the compound analyzed 391 ppm VCM, while the bottles analyzed 333 ppm VCM. If the same amount of VCM were lost in blowing the bottles for 4566 and 4567, the compound would have contained about 356 ppm VCM as the bottles for 4566 and 4567 averaged 303 ppm VCM after blowing. The contents for 4566 and 4567 are shown in Figure 4. All points for the 2 oz. bottles lie below the theoretical no initial loss curve. As the VCM loss during the blowing operation is about 15% (at least for the bottles blown in January), it is necessary to use the equations for Case II to determine the parameters of the model. The data sets 4566 and 4567 have two common parameters: 1. The initial concentration of the compound as bottles for both runs were blown from the same batch of compound. 2. The dimensionless time parameter which describes the VCM loss before filling, as all bottles used in 4566 and 4567 were blown and filled at the same times. SPI-02850 SPI-02851 T /M - DfMENS/OMlSS 16. The parameter D/l2 is different for the two data sets as 4566 was run at 120'F, while 4567 was run at room temperature. Both the bottle content analyses and the bottle wall analyses have been included in the regression, using the appropriate equation for Case II given in Appendix I. The model parameters obtained by using a non-linear least square analysis are given as follows: Runs 4566 and 4567 Parameter D/l2 at 120*F, day1 D/l2 at room temper ature, day1 WG, initial concen tration of VCM in compound, ppm Dimensionless time corresponding to VCM loss before filling Value 2. 30 ' 10"5 2. 88 * lCT* 346 3.52 ir3 95% Confidence Limits Lower Upper 2.23 10r* 2.37 ' icr3 2.64 * i<r* 3. 13 * icr* 336 2.34 ` icr3 357 4.71 ' lO-3 Figure 5 shows the bottle content data for runs 4566 and 4567. The three data points of set 4566 at 2 weeks have been rejected during the regression. Figure 6 shows the comparison of VCM concentration in the bottle wall for data sets 4566 and 4567. Some of the bottle wall data which deviates widely from the curve have been rejected during the regression. A similar regression analysis has been made for the data of runs 4562 and 4564. The delay between blowing and filling has been included. For the 4564 series the 118 days of storage at room temperature plus about 12 days which corresponds to the vinyl chloride loss during blowing was used as the total delay time (130 days total). For the 4562 series the delay time at 120*F was estimated by dividing the sum of 78 days of storage at room temperature plus 12 days for blowing losses (90 days total) by the ratio of the 120*F diffusivity to the room temperature diffusivity determined from the analysis of the 4566 and 4567 data. The D/l2 parameter and the initial concentration of vinyl chloride were estimated from the regression. SPI-02852 0 .0 5 r/M E - >/MEMS/OA/LSS SPI-02853 W EIGHT FPECT/ON VCM 0.1 0.05 F/GU/ZE 6 compopison of bottle well /CM analysis WITH degression CUP7E l'CM LOSS PORING B!OWING included (CASEJJ) 2 OUNCE Bottles 50?a ETHANOL o sep/ee +567 72'f a SEMES 4566 /20`F 0.02 12 MO MOS. 11 52 PAYS was. II 4ms. 1 / r 6 NAS, 11 1 0 0.05 o/ 0./S t/me- dimensionless & 2 yf-TiAtC AT72,r 3 MfS WAEAT/ZDC,f 1 0.20 1 0.24 SPl-02854 19. The results are given as follows: Run 4562 - I20CF Parameter D/12, day"1 Wo. Initial concen tration, ppm Value 1.67 - lO-3 102 95% Confidence Limits Lower_____ Upper 1.53 * 10"3 1.82 * 1(T3 99 / 106 Run 4564 - Room Temperature D/12, day-1 W0, Initial concen tration, ppm 2. 16 * 10"4 160 " 1.70 * 10"* 149 2.62 * 10*4 171 The VCM concentration in the bottle contents is shown in Figures 7 and 8. The bottle wall analyses are compared in Figure 9. A reasonable comparison of the experimental data for runs 4566 and 4567 with the theoretical curve for no VCM loss can be made if half of the VCM lost during blowing were added to the observed values (the other half of the VCM diffuses to the exterior surface and does not appear in the contents) and if the time corresponding to the blowing operation is added to tiie actual time after filling. This approximation will yield numbers slightly greater than the true values. The results of this comparison are plotted in Figure 10. Conclusions Experimental extraction data, both the VCM concentration in the 50% ethanol contents and the VCM concentration in the bottle wall, determined using 2 oz. bottles are consistent with the diffusion controlled extraction model. This extraction model can be used to predict the effect of different bottle sizes and bottle contents on the VCM level in the bottle contents. Predictions for 2 oz. cylinderical and 16 oz. cylinderical bottles for several materials are shown in Figures 1 and 2. Bottle weight as a function of bottle capacity is shown in Figure 11. Using the average curve from Figure 11, the following values for the maximum VCM concentration in 50% ethanol solution can be calculated for bottles made of PVC compound containing 1 ppm VCM. SPI-02855 Part 2 - Additional Comments on Methodology The PVC compound used in the preparation of the sample bottles was compounded at Ethyl's Tiptonville, Tennessee plant. This compound was blown into the 16-ounce and 2-ounce bottles in our Baton Rouge laboratory. The bottles were made under conditions and on equipment that are used normally in commercial PVC bottle manufacturing plants. The bottles are of a quality (appearance and properties) comparable to those commercially produced. The 16-ounce bottle is a tapered oval design with approximately 3. 9 inch x 1.8 inch maximum oval dimension and 6. 75 inch height with a standard 24 mm screw cap neck finish. The 2-ounce bottle is a nominal 1.35 inch diameter, 3. 15 inch height flat bottom cylinder with a standard 8 mm screw cap neck finish. The 16-ounce bottles weighed 36 grams and the 2-ounce bottles weighed 10.5 grams. The aqueous solutions were made with distilled water that was checked for zero VCM content. Solutions were mixed and stored in clean glass vessels. The heptane and vegetable oil were purchased locally in glass or PE containers. The PVC bottles were not flushed prior to filling. The bottles were filled at 72F by decanting the solvents from the storage vessel into the bottles. Bottles were filled to a point slightly above the shoulder of the bottle, which is the nominal capacity of the bottle. The bottles were immediately sealed after filling. The bottles were sealed with the following closures: Run Bottle Size ____________ Seal 4555 and 4536 4555 and 4556 4562 and 4563 4566 and 4567 16 oz. oval 2 oz. cyl. 2 oz. cyl. 2 oz. cyl. Cap with vinyl coated pulp liner Cap with polyethylene plug liner Cap with vinyl coated pulp liner Heat sealed foil liner and cap The room temperature storage work was carried out at a constant 72 *F. The elevated temperature work was made in an air circulating oven at a constant 120 F. After completion of the storage period, sample bottles were refrigerated at 35-40F until VCM analysis was made on the contents. Representative chromatograms from each of the series is attached. SPI-02856  Part 3 ' Anaiytical Procedures (excerpt from "Extraction of Vinyl ,(f1rlde from PVC bottles by Bottle Contents", January 17, The Determination of Vinyl Chloride Monomer in Various Solvents and in Alcoholic Bcvcmt/cs Stores in Polyvinyl Chloride Containers T. G. Mungall Introduction Upon request from the Plastics Research Group a method was developed for the determination of vinyl chloride monomer in 50-50 ethanolwater, heptane, 3% acetic acid and alcoholic beverages stored in polyvinyl chloride containers. This report covers the related development work and includes con firmatory data which were also gathered at their request. A copy of .this me'.hod was turned over to the Gas Chromatography Services Lab and they have been running samples submitted to them by the Plastics Research group using this method for the past several months. Summary and Conclusions 1. A detailed method for the determination of vinyl chloride monomer (VCM) in various solvents including 50-50 ethanol-water, heptane, 3% acetic acid and alcoholic beverages is attached to this report. The method should also be applicable to the determination of VCM in other solvents. 2. The method, as written, can be used to determine as little as 0.01 ppm VCM in solution with a relative standard deviation of 16%. 3. A computer or electronic integrator can be used, with this method, for samples containing 0.05 ppm or more VCM. It is necessary, however, to hand calculate the data for concentrations below 0.05 ppm. 4. Vinyl chloride monomer recovery factors for absolute calibration were in agreement regardless of the solvent medium. Factors were obtained for VCM in 50-50 cthanol-watcr solvent at various concentration levels from 0.01 ppm to 0.27 ppm, in heptane solvent from 0.01 to 3.6 ppm and in 3% acetic acid solution at the 0.25 ppm level, 5. Recovery data on VCM from "spiked" samples of 50-50 cthanol- watcr solution, actual J cthanol-watcr and other ToTvcTvtr ' as well as aiconoiic beverages stored in pVc, arc included in~i.:is rep erf! %kj-__ Sample chromatograms are inrbwterl in shffw the relative peak areas obtained with varying concentrations of VCM. SPI-02857 Discussion In April, 1973, a request was made by the Plastics Research Croup for the development of a method by which VCM in 50-50 ethanol-water and alcoholic beverages could be determined at the 1 ppm level or lower. Later on this request was extended to include the determination of VCM in heptane, 3% acetic acid solution, water and PVC (both the resin and finished products). Initial work in this area was done using a Varian Aerograph 2800 equipped with a flame ionization detector (FID) and fitted with a 1/8" x 15' stainless steel column packed with 16.7% TCEP (tris-2-cyanocthoxy propane) on 70/80 mesh Chromasorb W, A. W. The limits of detection with this method were about 0. 1 ppm VCM in liquids and 2 ppm in PVC. Upon a request from the Plastics Research Group for greater Bensitivity it was decided to try an electron capture detector. Electron capture proved to be little or no improvement over flame ionization as far as sensitivity for VCM is concerned; however, we were able to achieve the desired sensitivity by modifying our original FID method. The limits of detection of this new method, provided the peaks were hand-calculated, was 0.01 ppm VCM in solutions and 0.25 ppm in PVC. In this new method a different column was chosen to give a better separation between the VCM and heptane peaks than the TCEP column afforded and the flame ionization detector was operated at a much higher sensitivity. The size of the sample injected into the instrument was also several times larger than the former method. Due to time considerations the gas chromatography service lab began integrating peak areas by means of the IBM S/7 computer. This practice resulted in a higher detection limit of approximately .0. 03 ppm VCM in the 50%/50% ethanol-waiter solvent. The Plastics Research group requested data showing sensitivity, accuracy and precision of the routine analysis as applied to the 50%/50% ethanol-water system. Such data were obtained on the Bendix 2200 instrument by three different lab assistants. . The S/7 computer was used for data processing. First, a standard was prepared: VCM in 50-50 ethanol-water solution at the 0.27 ppm level. This was prepared three different times; on 8-14, C-20 and 8-30-73 and each standard run two or more times. The overall relative standard deviation for this standard was 5% and the vinyl chloride factor thus derived was 0. 0338 ppm ViCl/unit (s/7) area (see Table I). A standard was then prepared containing 0. 054 ppm vinyl chloride in 50-50 ethanol-water. This standard was injected repeatedly and the areas for the vinyl chloride peaks obtained on the IBM S/7 (Table II). SPI-02858 3. i Standards at the .027 and .011 ppm levels were prepared and an attempt made to collect the data on the IBM S/7; however, the noise to signal ratio, at these levels was so high that only a few pieces of data at the . 027 pprn level were obtained on this system (sec Table III). The charts were hand-calculated to obtain numbers at the .011 and .027 ppm levels (sec Tables IV and V). Attached are typical chromatograms obtained at the 0. 270, 0. 054 and 0,011 ppm levels of vinyl chloride in 50-50 ethanol-water solution. Vinyl chloride factors have been obtained in heptane and 3% acetic acid as well as 50-50 ethanol-water. Table VI summarizes these data. i SPI-02859 * TABLE I VINYL CHLORIDE FACTORS 50-50 Ethanol-Water Solution {0.270 ppin ViCI level) - IBM System/? Peak Area 8.6.113 .8.3823 8.1879 7.8188 7.7846 7.8890 7.VI10 Factor ` ppm/llnlt S/7 Area .03135 .03221 .03298 .03453 ' . 03*168 .03418 .03629 X = 0.03375 (0.0338) S = 0.00168 where: /l/nyX5 - (rX)s S = [/ n(n - 1) Rel. Std. Dev. + 5.0# where: ! Rel. Std. Dev. = X / l table ii 0.060 0.053 0.060 - 0.055 o.o'ig X = 0.059 s = 0.0079 Rel. Std. Dev. = 13.A# 1 I I i i SPI-02861 TABLE III RECOVERY OF VTHYL CHLORIDE KOKOMER FROM bO-SO KTliAUOh-V.'ATKR SULUTIUi) (0.027 ppm level) IBM System/7 ppm In Solution .031 .026 .028 .026 X = .028 S h .0025 Rel. Std. Dev. = 9.1^ SPI-02862 TADT.K IV RECOVERY OP VJHYT, CHLORIDE MONO KKOM 50-:.>() J':T)IAI.;C)J,-WA,J'j-;H SOUJT.l (0.011 ppm level) Hand Calculated 0.016 . 0.020 0.014 0.014 0.014 0.015 0.014 0.016 0.014 - 0.022 0.019 0.017 0.017 0,016 JC*e 0.016 S = 0.0025 Rel. Std. Dev. = 15.8# .I SPI-02863 I TAl'LE V RECOVERY OF VINYL CHLORinE MONOMER FROM !)Q-bO ETHANOL-WATER SOLUTION i (0.027 ppm level) i _ Hand Calculated i 0,027 a 00..00320^ 0.02q X = 0.028 S = .00263 Rel. Std. Dev. = i9Jf < ' rt * < 4 SPI-02864 TADJIK VI 0. 355 ppm Vinyl Chloride Level 3% Acetic Acid 0. 249 ppm Vinyl Chloride Level .0281 ppm/Area Unit .0319 .0334 ppm/Arca Unit .0280 .0309 .0301 .0339 .0290 .0303 SPI-02865 !M!iIII STAiJDAjcD II r,- ------ o-~- Lilli * ii --1 i'iI '! *i Li.i t ( < AVLLTT-fAC 'Also 'CtOO-DZ?.?. ,, ,... n !; jilt :'ldOCd/;i/l\ .Ak."X '{tsfeoft 'Sf:Z i \i \! i i;: -ifIjjjijajr: i ! i \n ! ! I'l&M&'P.X -ii iS: 'ii Lu ,. M[A Ccticr;c;j: jsm >1 SPI-02866 / 'n''! f |: f i [ nf'f I i 11 [ i f|f; i [ [i! i i n11 ;-Fi-n w y i |ii - i! Hn ( -fi- iU -4;' H-' iii' iMhihJiitLl an: 1,H-1'_.-ihi-_ilj j r "ETE ' nr- I. i. h iTrj iTh' -H4~i -- !--! T H-H- inT.r: -i-r-'-H- -t- i - : !!- ------ t-rH- j. <1 1 y : I i -<t -r-i--f~i ... .I. ... i.i.i... LL' | ; jX---i -- i rl -r-h- Vi 'mmm-zms: mm. se :=-lit /TJ-l.U. i-; I .1r u lit,1 aaj ijL.iTiu iimz'ijjili!.Txt. tEbidHisinai.fat:. .. ;:i.iji: xjtl:. riii: W.t.l . .;_|.J | iI.i1. ti:i!h.i.fH'T.!i".. f-fl- i r i .ii &w TX rrm--mi1-1 Ji_il-...r. -*":. H- rh ..ttn . i-li-"H-i i-t-r- .rrpjrt !-| r` tie -M-r-L -Hjf -!n-! _|_ -Hri~ : ft l-ft 5H fjr exit -;-Li;_i_u Ljj_l_E .1-] -i--t--1_ J^C-_ f-4-4- X!- J- 4 1 .! 1. H- !' i+X iim: TJTLLzp-i: X-t--U rrn- -J jjLL1 ___Ll_:_ . !_ . I .at... I r-r*r- L i ......-.M. - i- -^-;T-LFi-rH--T J UXS rae: {:-h ::iix ro- : I if!" ! r.! .i-t-i ; 11 :-y/tiy.1 amRixrm. LIT' I -1 I I!' * : l. i ! J ii-i W1 .. . lV, -H'-r j ; . ! M: H j-1 i ; i !-,t ; 1-1:1 Li.-;-.:: IT 1 J.;... iy:'y \m 1 -4-1 1 Jt iiii g:^t ft;"!- i!]:r .1. 1-' rtn- -x-kL... 4TI4 J: -.1|.J1.1 rhr SPI-02867 Jio nyco-ozi?. lit; '.bj TV " !-* ` T^: } tv . ': t l.l ;. iJ I:!; j\\. .! i ' 'i: ! 1 : 1 ; '-I J 1 : 1L ' Ij Ii kkl*^r ; i . <j /> ' ./ - I :-W- irn _i _. ri; fc: j fcK i-n gdlm/ti = r!-r- . I.!' iI . .: I..'J m h'i-j m _!.-! ai'j. i . f-: i >* Ii SPI-02869 SPI-02870 SPI-02871 ... * I: ; .. . .j : ;j v o OOC o c oooo to Oc O<\)l O<T >5 C9 2 o t~ o co co z<i o...t.f.i..if.)...f.)..N.. qo ^n oOO Oto- oo oo o o o <t O O ii. c ^ ^ ^ z <COt c*u <t < O- Ow' "ra? n n cm -- n oe ooo X<E OI to -- X CO loftoooiono .. O O O o <f z O--> to on. o o O' *- s. 3-- 1J0 H nJ.c-o4'ljfclOO/J tOlo') H Z to z w o >: x <r au>- Z. J AO A c- Z. Z Z OG C J zx za.: zx >z z z z. =>=>=>=> I '.I i i i-- _j-------: SPI-02872 * ijEf'O.iT FOH GC 09 OH/22/73 * COMPONENT .NAME ** <<?>> UNKNOWN UNKNOWN 1 UNKNOWN VINYL - CL ELUTION PEAK TIME HEIGHT 95.5 116.1 164.9 245. 1 293.5 "0.05 0.01 0.0B 0.06 0A0 13:50 METHOD 9001 noh pun AKF.A FHACTION 0.105750 0.081302 0.129270 0.120300 0.563350 FACTOH 0.00000 1.00000 1.00000 1.00000 1.00000 ANALYSIS ... 0.000000 1.607300 2.555700 2.378200 11.137000 SPI-02873 procedure iI The Dotermlnntion of Vinyl Chloride Monomer in Var:iou.-. Solvents or Llouors SCOPE 1. This'method is desicned to determine as little as 0.01 ppm vinyl chloride monomer in various solvents or liquors that have been stored in polyvinylchloride containers. OUTLINE OF METHOD 2. The solvent or liquor is taken up in a microliter syringe and injected directly into a gas chromatograph equipped with a flamp ionization detector. Calibration is done on an absolute basis by obtaining the peak area for VCM in fixed aliquots of standard solutions. \ APPARATUS 3. (a) Gas Chromatograph - Bendix Model 2200, or the equivalent, equipped with a l'lame ionization detector. , (b) Gas Chromatographic Column - 1/8" x 25', 30jS SE-52 on 80/100 mesli Chromosorb iiP. (c) Microliter Svrlnre - A 50 pi Hamilton gas chromatography syringe No. 705-N or the equivalent. (d) High Purity Nitrogen, Hydrogen and Compressed Air - (e) Volumetric Flasks - .50, 100 and 1000 ml capacity. (f) Rubber Stoppers, Serum Tyne - (g) "Gastjght" Syringe - a 2.5 ml Hamilton "gaslight" syringe #1002 or the equivalent fitted with a 4" No. 25 needle. (h) Data Processing Ennirmont - IBM System/7 or the equivalent, if desired. III SPI-02874 REAOKDTC If. (a) Vinvl Chloride - Jlich purity, available from Ethyl Corporation, baton Rouge, La, (b) Ethyl) Alcohol - 95^ alcohol (190 proof). (c) 3# Acetic' Acid Solution - * . <<0 ..Heptane - Phillips 99 mole # or the equivalent. CAS CHROMATOGRAPHIC CONDITIONS ------------------------------------------------------------- . 5. (a) Carrier Gas: 40 cc/min. < (b) Detector: Flame ionization. (c) Hydropen Flow Rate: Refer to instrument manual for optimum conditions. ' (d) Air Flow Rate: Refer to instrument manual for optimum conditions. ^ (e) Injection Port Temperature: 200C (f) Detector Temperature: 300C (g) Oven Tomnerature: Ambient during the run until the vinyl chloride elutes from the column, then the oven tem perature is programmed 15/min to 155C and held for 12 minutes to clear the column prior to injection of the next sample. CALIBRATION 6. (a) Prepare a standard mixture of vinvl chloride in the solvent by disavITllT 1TTTOT cc of vinvl c>ilo}JL&e--gas-^t-ro-vr* /PS^O-pg ViCl) in UhO'ugn solvent to fill, a 1000 ml volumetric flask. (The flask should be weighed empty and then filled with the solvent and re-weighed so that the actual weight of solvent can be deter- mined). A serum cap is then placed over the mouth of the volu metric flask and the flask inverted. In this position 1.0 cc of vinyl chloride gar, is injected via the gas tight syringe into the solvent and the flask shaken for several minutes in this inverted position to allow the vinyl chloride to go into solution'. This standard solution will contain from 2.5 to 3.5 ppm vinyl chloride depending upon the weight of solvent. SPI-02875 Further dilutions of this standard may be made by pipetin known volumes of the standard into volumetric flasks and diluting wit)) additional solvent. For example, a convenient way of obtuinin 0.05 ppm vinyl chloride in 50-50 ethanol-water is to prepare the standard as described, using about 9^0 G the 50-50 mix, then 1.0 cc of vinyl chloride gas (25;I0 pg ViCl) is injected. This will result in a stock mixture containing 2.70 ppm vinyl chloride. A 10 ml aliquot of this stock mixture is then diluted to 100 ml with fresh 50-50 solution giving an intermediate standard of 0.27 ppm vinyl chloride. A 10 ml aliquot is then taken of this 0.27 ppm ` mixture and diluted to 50 ml with additional 50-50 solution. This 'will result in a standard containing 0.05^ ppm vinyl chloride. (b) Exactly 20 pi of the 0.27 ppm vinyl chloride standard mixture is injected into the gas chromatograph via the system inlet and the area of the vinyl chloride peak is obtained via the System/' or by hand calculating. The factor is then calculated by dividing the concentration (0.27 ppm) by the peak area thus giving an absolute response in terms of ppm/pea): area unit. This procedure is repeated several times/ preferably including a duplicate preparation of the 0.27 ppm standard solution. PROCEDURE 7. (a) Inject exactly 20 pi of the sample solution into the gas chromatograph. This solution can be either 50-50 ethanolwater, 3# acetic acid, heptane or liquor. Obtain the peak area on an IBM System/7 or other suitable integrator. If the concentration of VCM is below 0.05, hand calcu lation from the recorder chart is recommended. CALCULATIONS 8. (a) Calculate the ppm vinyl chloride monomer in the sample as follows: ppm ViCl = A x B where: A = ViCl peak area for sample. (20 pi) B = ppm VICl/unit area obtained from absolute calibration of the instrument. (20 pi of std.) (see Calibration Section) The detection limit for this method is approx. 0.01 ppm ViCl. The sensitivity is 0.01 at the 0.05 ppm level and the accuracy is 13$; at 0,05 ppm level. /r n uo 10-3-73 / l SPI-02876 Revised 11-15- GAS CHROMATOGRAPHIC METHOD FOR TIIE ANALYSTS OF VINYL CHLORIDE MONOMER IN POLYVINYL CHLORIDE INTRODUCTION This method includes the dissolution of polymer in tetrahydrofuran, which contains ethyl bromide standard, and the analysis of the solution via gas chromatography. MATERIALS 1. Gases: All must be high purity, water pumped where applicable, and suitable for use with a gas chromatograph. Hydrogen AiiHelium 2. Chemicals (Reagent Grade): _ Ethyl Bromide Tetrahydrofuran (THF) 3. Magnetic stirrers 4. Stirring bars 5. - 125 ml Erlenmeyer flasks 6. 100 ml Volumetric flask 7. Syringcrs, 50 \il, 10 8. 50 ml Pipette 9. No. 268 Sani-Tab caps 10. Gas Driers (Tek Lab Cat. No. 235) INSTRUMENTS 1. Gas Chromatograph Varian 1740 or equivalent equipped with dual flame ionization detectors. Columns: 16 feet by l/8 inch thin wall (.012 in.) stainless steel tubing packed with 16.7% 1, 2, 3, tris(2-cyanocthyl)propanc, (TCEP), on Chromosorb W, AW, 70/80 mesh. Columns are to be arranged for on-column injection. 2. Strip Chart Recorder 0 - 1 millivolt full scale 1 inch per minute chart speed I SPI-02877 PROCEDURE 1. Standard Solution: Fill to Uic mark, a 1000 ml volumetric flask with TIIF and add to it exactly 40 \iL of ethyl bromide. This provides a 65 ppm solution of ethyl bromide. 2. Weigh exactly 2. 50 g of PVC sample into an Erlenmcycr flask and pipette 50 ml of standard solution into the same flask. Add a stirring bar, cover with a Sani-Tab cap, and magnetically stir the mixture until the sample is dissolved or gives a homogeneous suspension. 3. Inject 5 \il of solution into the gas chromatograph under the following conditions: Injection Temp. Detector Temp. Oven Temp. Helium Flow Electrometer Range = Attenuator = 100*C 250*C 75*C 40 cc/min. (60 lb) 10"namps/mv 2 4. The attenuation may need switching from sample to sample in ar. effort to keep the vinyl chloride peak and ethyl bromide peak on scale. If at any time, the peaks exceed the limit of the chart paper, the sample must be rerun. 5. After the peaks of interest arc collected, the analyst must wait for complete elution of THF solvent before the next sample is injected. 6. At the end of a series of samples, elevate oven temperature to 150* for approximately one hour. 7. Measure Hie areas of the peaks representing vinyl chloride and ethyl bromide and calculate the ppm vinyl chloride using the following arithmetic: Av As x 0.573 x 65 x 17. 6 = C Av = area of vinyl chloride peak As = area of ethyl bromide peak C = ppm, by weight, of vinyl chloride in the sample. PRECISION AND ACCURACY The precision and accuracy of the method has been shown to be 10-20% relative at the 1 ppm level. SPI-02878