Document GKN5Q11yV23vdwbB9kee7rb8N

RESEARCH ANO IX VC LOPMtNT DEPARTMENT UNION CARBIDE CORPORATION CHEMICALS AND PLASTICS P.O. BOX *361. SOUTH CHARLESTON. W. VA. 25303 l).C 1 81975. December 15, 1975 (s * r L <D ' * ' + Managing Editor International Scientific Communications, 808 Kings Highway Fairfield, Connecticut 06430 Inc. Dear Sir: I respectfully submit a paper titled "A Sealed System for the Analysis of Vinyl Chloride" for your consid eration to publish in the AMERICAN LABORATORY and INTERNATIONAL LABORATORY. This paper is original work and has not been sub mitted for publication elsewhere but it was presented at the Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy in March, 1975. The work was performed at the Union Carbide Corporation, Chemicals and Plastics Division, Technical Center, South Charleston, West Virginia. The paper is co-authored by Messrs. A. E. Gabany, Jr. and H. Senman. Your consideration of this paper for publication is highly appreciated. Yours truly, AEG,jr./dk . Attachment A. p. Gabany, Jr. Analytical Chemist UCC 021126 A SEALED SYSTEM FOR THE ANALYSIS OF VINYL CHLORIDE A. E. Gabany, Jr.f Research and Development Department H. Senman, Measurement and Control Technology Department Chemicals and Plastics,' Union Carbide Corporation South Charleston, West Virginia 25303 ucc 021127 SUMMARY Recent interest by the Food and Drug Administration (FDA) to define the migration of vinyl chloride monomer from food con tacting surfaces to food-simulating solvents had lead to the development of a unique sealed system to obtain the analysis. The sealed system consists of a glass jar to contain the polymer sample and extraction media, a special sealing device, and a sample holder. After the system is assembled, it is torqued tight, then incubated, at an elevated temperature. Upon cooling to ambient temperature the extraction media is sampled for a gas chromatographic analysis. Experimental data are presented,verifying that there is no loss of vinyl chloride from the aqueous and organic solvents up to 250F. Additional analytical data are presented which confirm that vinyl chloride monomer migration into food-simulating solvents at the 50 ppb level is readily obtainable with this system. INTRODUCTION Vinyl chloride (bp. - 13.9C) is widely used in polymeri zation reactions to make resins, homopolymers and copolymers, which are used in a diverse gamut of end products that have become an integral part of our daily lives. Until recently, it was generally believed that these polymerization reactions were complete and that essentially no residual monomer existed in the resins. However, with the advent of gas chromatography, it has been demonstrated that ucc 021128 detectable levels of residual monomers, eg. vinyl chloride, can exist in the resin after a polymerization reaction. Also the question of what is a detectable level has undergone much scrutiny and in the past year there has been considerable progress in lowering the detectable level. Residual vinyl chloride monomer (VCM) determinations alone have become an industry-wide challenge,but its migration into food products through contact with the packaging materials has becom an even more important subject for the Food and Drug Administration and the food-packaging industry. FDA's three-fold criteria for clean resins to assume "no reasonable expectation of migration into food" are 1) no detectable VCM in a food contact article, 2) no detectable extractions of VCM from the article into food-simulating solvent, and 3) a sound theoretical basis for predicting no migration below the detectable level. This paper presents a method to define the migration of vinyl chloride monomer from food contacting surfaces to foodsimulating solvents using specific extraction conditions defined by the Food and Drug Administration. Included in these conditions are contact periods which are maintained at above ambient tempera tures. These conditions are as follows; 1. Extraction of the polymer with water at 250F for two hours, followed by an additional lower temperature extraction whereby the extraction system is maintained at 120,'F and subsequently analyzed for vinyl chloride monomer at 72-, 98-, and 120-hour intervals. ucc 021129 -3 - 2. Three percent aqueous acetic acid extraction at 212F for tw hours followed by an additional lower temperature extraction as described with the water extraction. 3. Eight percent ethanol in water extraction at 150F for two hours followed by the previously described lower temperature extraction. 4. Extraction of the polymer with n-heptane at 150"F for two hours followed by an additional lower temperature extraction whereby the extraction system is maintained at 120`F and analyzed for vinyl chloride monomer at 6-, 8-, and 10-hour intervals. As is indicated, these extraction conditions require an extremely well-sealed system which prevents the escape of vinyl chloride monomer (up to 16 psig at 250F) for the designated tim periods. Also, a suitable system would have to provide proper sampling capability to the analyst. EXTRACTION SYSTEM AND INSTRUMENTATION A system meeting these requirements, developed and fabri cated by the Measurement and Control Technology Division of Union Carbide Corporation, is exhibited in Figure I. The system consists of a glass jar to contain the polymer sample and extraction media and a special sealing device and sample holder. The glass jar containing the sample and extraction media is capped with a 0.020" thick Teflon diaphragm: the Teflon provides an inert barrier on which there is no absorption of vinyl chloride monomer.- ucc ! I 021130 4 A Silastic septum provides a seal of the system during and after sample withdrawal. The 0.125 inch thick Silastic septum is placed on the outer Teflon surface and a 0.25 inch thick aluminum plate containing two sampling ports is then placed on the Silastic septum. These sampling ports, located near the periphery of the thicker center section, permit the withdrawal of a sample with a microsyringe. A depression located in the center of the top of th aluminum plate serves as a guide for the acceptance of a tightening screw for the assembly. A specially designed aluminum screw-type cap installed on the glass jar serves to hold the assembly together prior to being placed in the holder. The aluminum plate is situated above the threads in the cap and thus is not tightened firmly against the \ Silastic septum when the cap is screwed on the jar. A special holder is used to seal the system. After the sample assembly is placed in the holder, the screw at the top center is tightened into the center depression of the aluminum plate to seal the unit. A bottom pad of 0.125 inch thick Silastic, on which the glass jar sets gives enough flexibility to allow the application of 50 inch-pounds torque through the screw onto the aluminum plate. The assembled sealed system is torqued tight prior to incubation at elevated temperature. For safety reasons, the entire system is placed in a canvas bag prior to applying the 50 inch-pound torque, after which the bagged system is placed in a suitable wire- formed cage container. The wire cage container containing the assembly is then heated in an air-circulating oven maintained at 1F of th desired temperature. UCC 021131 As stated previously, vinyl chloride has a sub-ambient boiling point of -13.9F. It is also a hydrophobic compound; there fore, the sampling of the aqueous media for analysis required some modification of the system. The procedure used was the addition of 150 ml of aqueous extraction solvent and 15-ml of n-heptane to the system prior to sealing and subjection to the designated heat treat ment. Upon cooling to ambient temperature the system was shaken vigorously thereby extracting the migrated and/or spiked vinyl chloride into the heptane layer which was sampled for analysis. Any vinyl chloride present is distributed in the aqueous, organic and gas phases of the system with a constant distribution coefficient. The distribution ratios obtained for vinyl chloride on test samples of the various aqueous solvent extractants which had 1 been spiked with 0.5 ppm vinyl chloride are listed in Table I. In all cases, the highest concentration of vinyl chloride is in the heptane layer, however the increased solubility effect from the aqueous organic coextractant is also evident. The heptane layer was sampled for analysis from the glass jar by injection of a microsyr*inge through the septum port. A hypo dermic needle of a 10 or 50--Ml syringe is long enough to penetrate through the Silastic septum and Teflon barrier into the heptane layer. After the microsyringe is filled with sample and is withdrawn, the system is self resealing due to the Silastic septum. The recovered sample is injected into a gas chromatograph for its subsequent analysis. UCC 021132 The gas chromatographic analyses were performed on a Hewlett-Packard (F and M) Model 5750 gas chromatograph equipped with a hydrogen flame ionization detector. The analytical column was a 6-ft. x 1/8-in. O.D. S.S. Tubing packed with Chromosorb 102, 60/80 mesh, which was operated isothermally at 10QF; however, approximately three minutes after a 10- fi 1 sample was injected, the temperature was manually increased to 250F to elute the n-heptan solvent. The injection port was held at 100F (maximum) and helium-, the gas carrier,was regulated at 30 cc/minute. Using these para meters, vinyl chloride elutes at approximately 2.4 minutes. DISCUSSION AND EXPERIMENTAL RESULTS The system was checked to determine if it would indeed hold the vinyl chloride monomer (VCM) at the elevated temperatures. An experiment was performed which involved filling five (5) systems with 150 ml of distilled water and 15 ml of n-heptane. A stock solution of 100 ppm vinyl chloride in heptane was used to spike each container as shown in Table 2. The standard samples included no i vinyl chloride, 0.5 ppm, 1.0 ppm, and two 2.0-ppm samples, one of which was a control maintained at room temperature. The other standards were heated at 250F for two hours, and then allowed to cool to ambient temperature. Ten ^1 of the heptane layer from each sample was obtained as described previously and injected into the gas chromatograph for analysis. In the last column are listed the peak heights obtained for the vinyl chloride peak for each sample. UCC 021133 7 Partial chromatograms showing the vinyl chloride peak for the five samples obtained on a Chromosorb 102 column operated at 100"C are shown in Figure 2. The peak on the left represents the vinyl chloride obtained on the ambient control sample, the next is the zero ppm heated standard, then the 0.5 ppm, the 1 ppm and to the last which was the 2.0 ppm heated standard. These data in graphic form (Figure 3) demonstrate that the detector response versus the concentration of vinyl chloride in the calibration standards is a linear relationship. More importantly, from the data, the syst m was judged acceptable to perform the required analysis. Polymer samples simulating can coatings were analyzed using seventeen 2-inch diameter aluminum disks which had been coated on both sides with the test vinyl polymer. This arrangement provided \ a polymer surface area of 106 square inches per 165 ml of liquid or a surface-to-volume ratio of 0.7 square inch to one milliliter. -Since the usual food container has a ratio of approximately one square inch to ten ml, this apparatus provides approximately a seven fold exaggeration of the usual surface-to-volume ratio. Experimental data given in Table 3 indicate that the migration of vinyl chloride from the test polymer is less than 50 ppb to the simulated food solvents. Notice that generally all of the peak heights obtained on the extracted polymer are less than 1 mm with the exception of the acetic acid where there is a 5 and a 2-mm reading. These latter measurements actually may not represent vinyl chloride at all because of the sensitivity level at which they were ucc 021134 measured. The flame Ionization detector electrometer was set at almost its maximum sensitivity which greatly amplifies even th slightest noise. Additionally, with the samples tested, acetic acid attack on the uncoated aluminum surface presents an unresolved difficulty. Table 4 depicts the calculations used to establish th apparent ppb vinyl chloride migration from the gas chromatographic data. The flow scheme on the left depicts the aqueous extraction which was spiked with 0.5 ppm vinyl chloride. The vinyl chloride peak height response was 177 mm as compared to the 1 mm peak height observed from the sample on the right. From these data a factor was determined equating the vinyl chloride concentration in units of detector response. Hence, in our example the factor calculated \ to be 2.8 ppm/mm. This factor when multiplied by the peak height obtained for the sample under investigation determines the concen tration of vinyl chloride in the system. CONCLUSIONS i To summarize then, we have shown that a sealed system for the analysis of vinyl chloride has been developed and fabricated which is capable of sealing vinyl chloride monomer up to 16 psig at 250F. Additionally, this system has been successfully employ d for the analysis of vinyl chloride monomer migration from PVC coatings into food simulation solvents using gas chromatography as the means of detection. Vinyl chloride migration into food simulating solv nts at less than 50 ppm level is readily obtainable with this system. UCC 021135 A.-E. Gabany,'Jr TABLE I DISTRIBUTION COEFFICIENT OF VINYL CHLORIDE MONOMER IN THE SYSTEM (AT 25C) Head Space (Gas) n-Heptane Phase Aqueous Phase (as reference) WATER 0.05 7.7 1.0 3% ACETIC ACID 0.03 5.2 1.0 8% ETHANOL 0.03 4.0 1.0 ucc 021136 A. E. Gabany,' Jr TABLE 2 LABORATORY PREPARED SYNTHETIC SAMPLES OP . VINYL CHLORIDE IN WATER AND N-HEPTANE DESCRIPTION OF THE SAMPLE 1. Blank at 250"F, 2 hours 2. 0.5 ppm VCL at 250F, 2 hours 3. 1 ppm VCL at 250F, 2 hours 4. 2 ppm VCL at 250F, 2 hours 5. 2 ppm Ambient Control ML OF N-HEPTANE 15.0 14.2 13.4 11.8 11.8 ML OF 100 PPM VCL PEAK HEIGHT OF VCL (CM) -- N.D. 0.8 2.8 1.6 5.6 3.2 11.2 3.2 11.0 ucc Q21137 A. E. Gabany,'Jr.' TABLE 3 MIGRATION OP VINYL CHLORIDE MONOMER (VCM) INTO SIMULATED FOOD EXTRACTS DESCRIPTION OF EXTRACT A. WATER: at 250F for 2 hours followed by 120F for: 72 hours 96 hours 120 hours Control (0.5 ppm VCM) 3% ACETIC ACID IN WATER: at 212F for 2 hours followed by 120F for: 72 hours 96 hours 120 hours Control (0.5 ppm VCM) 8% ETHANOL IN WATER: at 150F for 2 hours followed by 120F for: 72 hours 96 hours 120 hours Control (0.5 ppm VCM) VCM (PEAK HEIGHT, MM) MIGRATION OF VCM*, PPM 1.5 <1.0 <1.0 177 <0.05 <0.05 <0.05 5 <1 2 140 <0.05 <0.05 <0.05 <1 <0.05 <1 <0.05 <1 <0.05 116 ucc 0211.38 A. E. Gabany, Jr'. TABLE 3 (CONTINUED) D. n-HEPTANE: at 150F for 2 hours followed by 120F for: 6 hours 8 hours 10 hours Control (0.5 ppm VCM) <1 <1 <1 41.5 <0.05 <0.05 <0.05 Apparent migration of VCM from test coating calculated using volume of extraction solvent and total surface area of coating \ ucc 021139 A. E. Gabany, Jr. TABLE 4 CALCULATIONS FOR VINYL CHLORIDE IN N-HEPTANE LAYER OF AQUEOUS EXTRACT POLYMER + 0.5 PPM VCL 150 ML WATER (0.5 MG VCL IN SOLUTION) 1 15 ML HEPTANE POLYMER 150 ML WATER 1 15 ML HEPTANE FACTOR - A VINYL CHLORIDE, PPM A PEAK HEIGHT 0.5 PPM ,, ---------------- - 2.8 x 10 PPM/MM 176 MM PEAK HEIGHT X FACTOR - PPM VCL IN POLYMER 1 MM X 2.8 PPB/MM - 2.8 PPB ucc 021140 A. E. Gabany, Jr.. FIGURE 1' VINYL CHLORIDE EXTRACTION SYSTEM ucc I 021141 A. Gabany, FIGURE 2 Gos Chromatograms of Aqueous VCI Standards ro A. E. Gabany^ Jr.. FIGURE 3 ppm VCI ucc' 021143 Managing Editor bcc: Mr. R. A. Bleidt Dr. J. J. Bressinski , Mr. G. A. Gillis Mr. J. B. Johnson Mr. R, W. Martin Mr. R. V. Sealey Mr. L. S. Slaughter Mr. H. Senman Mr. W. E. Whitehurst Decemb r 15, 1975 \ ucc 021144