Document By2OgveLK8k0yQx7VQ5KJZVL8

Inter-office memo > TENNECO CHEMICALS, INC. To H. B. Carr At Burlington Date March 31> 1969 From F. W. Kanzler At Burlington Copy to Subject Technical Memorandum Gas Chromatographic Determination of Impurities in Vinyl Chloride Monomer CONCLUSIONS: 1. A gas chromatographic method has been developed for the analysis of light end impurities in both virgin and re covered vinyl chloride monomer. 2. The analysis is accomplished by syringe injecting a gas sample from the headspace of a sample bomb into a F & M 5754A gas chromatograph equipped with a Tributyl phosphate column and a hydrogen flame ionization detector. A de tailed outline of the analysis is described in B.A.C.M. #1 which is attached as Figure 1. 3* The sensitivity of this method is to less than one part per million. 4. Preliminary studies indicate that the method provides good correlation with other laboratories (Table 1). 5. The time required for the analysis is approximately one hour. RECOMMENDATIONS: " " - * 1. The method should be further developed to provide capa bility for the analysis of heavy end impurities in MVC. 2. The correlation studies should be continued. 3. The gas sampling valve on the F & M 5754a should be modi fied for use with MVC in order to provide a more rapid and reproducible method of injecting the gas phase sample. COLORITE 016413 Gas Chromatographic Determination of Impurities in Vinyl Chloride Monomer Page 2 DISCUSSION: 1. Column Selection A literature search provided information about several liquid phases and solid supports used for MVC or related analyses. These were: Liquid Phase Silicone Gum Rubber B,B' - Oxidipropionitrile Diethylene Glycol Succinate Tricresyl Phosphate Tributyl Phosphate Solid Support Anakrom ABS Chromosorb w/AW Chromosorb w/AW Chromosorb p/AW and W/AW Chromosorb p/AW It was decided to use a Tributyl Phosphate liquid phase on Chromosorb p/aW solid support, as it offers excellent separation of the following light end impurities: Ethane Ethylene Propane Propylene Acetylene Isobutane Allene (Propadiene) Butane Butene-1 Methyl Acetylene Methyl Chloride Butene-2, trans and cis 1,3 Butadiene Houston also uses this column. However, a 30$ liquid phase of Tributyl Phosphate on Chromosorb P as used in Houston seems to be somewhat high, for a "conditioning" of the column was observed over a period of almost two weeks. For anticipated future heavy ends analysis the higher boiling Tricresyl phosphate liquid phase may be super ior to Tributyl Phosphate as it permits analysis at higher oven temperatures. It is of interest to note that of the three solid supports checked, only Chromosorb p/AW would render a good sepa ration. This is probably due to the relatively large COLORXTE 016414 Gas Chromatographic Determination of Impurities in Vinyl Chloride Monomer Page 3 surface area (4.0 m^/g for Chromosorb p versus 1.0 m12/g for Chromosorb W) and to the smaller porosity (1.1 cc/g for Chromosorb P and 2.8 cc/g for Chromosorb W).l 2. Operating Parameters Column: 12' x 1/4"S3 20% Tributyl Phosphate on Chromosorb p/AW 60 - 80 mesh Carrier Gas: Helium, at a flow rate of 47 cc/min. with a column "in" pressure of 50 psig and a column "out" pressure of 14.7 psia Hydrogen: 60 cc/min. flow, 40 psig line pressure Air: 350 cc/min. flow, 20 psig line pressure Oven Temperature: Ambient (about 25C) Injection Port Temperature: 130C Detector Temperature: 260eC 3. Identification of Impurities The identification of the impurities was arrived at by: a. Injection of pure gases and subsequent identifi cation of the sample impurities by matching the retention times of the various gases. b. "Spiking" of samples with pure contaminants and comparing the peak intensities of the "spiked" with the "un-spiked" sample. 4. Contamination Sources Difficulties were encountered on initial routine analysis by the occurrence of a very high level "background" con tamination after the elution of the vinyl chloride. This contamination resulted in an extreme loss of sensitivity for low level (< 25 ppm) impurity analyses. At this stage of the method development a "Push-Pull" gas inject valve was used to introduce the sample onto the column. To find the source of contamination the following steps were under taken: a. Column "Bleeding" - A check on column bleeding was negative. b. Detector - Upon dismantling of the detector assembly it was found that the copper alloy solder of the detector jet exhibited a high degree of corrosion. A new set of detector jets was secured and installed but resulted in the same background contamination. To eliminate the corrosion problem, caused mainly by the eluting MVC, a threeway ball valve was in 1 A. M. Filbert, M. L. Hair, J. of Chromat. Science, Vol. 7, February, 1969. COLORITE 016415 Gas Chromatographic Determination of Impurities in Vinyl Chloride Monomer Page 4 stalled at the exit of the column before the de tector. This allows the detector to be by-passed during the elution of the vinyl chloride and has reduced the corrosion. Hewlett and Packard has been notified of this problem and is investigating the use of more inert detector jet solder, c. Gas Inject Valve - The gas inject valve was by passed by injecting a sample with a gas syringe directly onto the column. This step resulted in an elimination of the background contamination. The cause was traced to an interaction of the MVC with six silicone rubber O-rings which seal the various chambers of the gas inject valve. Since the use of the gas inject valve offers the best assurance of reproducible injects, more inert Viton O-rings were installed. Again, the result was contamination. Hewlett and Packard was contacted and after a visit to their laboratories, they have agreed to supply us with a Teflon lined gas inject valve, which has shown no contamination during re peated MVC injects. 5- Calibration and Sampling Owing to the difficulty of finding a suitable internal standard, calibrated MVC standards are used for the quan titative determination of the impurities. Until the installation of the Teflon gas inject valve, syringe injection will be used for this analysis. The calibration and analysis is carried out by transferring a calibration gas or sample gas to a special manufactured "Glass Bag". The sample is transferred from the sample bomb to the glass bag and subsequently withdrawn with a syringe through a serum stopper contained on the glass bag. The sample size of 2.5 ml was found to yield re producible results. Due to the excellent peak resolution the calculation is based on the peak height. A typical chromatogram is shown in Figure 4 of the attached B.A.C.M. #1. FWK/r.U ! 'J W. Fred w. J Kanzler COLORITE 016416 TABLE 1 CORRELATION BURLINGTON VS. HOUSTON Component Propane Propylene Isobutane Allene Butane Butene-1 Methyl Acetylene Methyl Chloride Cis-Butene-2 1,3 Butadiene 85486 Burl Houston PPM 84 11 6 55 101 88 63 44 41 42 ND ND ND ND 2 1.2 38048* Burl Houston PPM 1 11 33 49 87 115 6 12 62 56 19 <1 ND <1 ND 17 3 85011 Burl Houston PPM 50 46 26 26 38047 Allied 714601 PPG 9902 Burl Houston Burl Houston Burl Houston PPM PPM PPM 87 <1 0.2 ND ND 10 7 42 ND 0.1 Not 1 Calc <1 0.5 <1 ND 15 18 <1 0.4 <1 0.1 23 <1 1 <1 ND 36 <1 0.7 <1 0.1 6 12 <1 ND <1 ND <1 Tr 149 150 50 52 <1 ND - ND - 41 21 65 COLORITE 016417 *In future cases. such as this. Houston will be notified and a sample will be sent to them for repeat analysis. COI.ORITE 0164X8