Gas Chromatographic Analysis of Atmospheric Sulfur Compounds Using Cryogenic Sampling and A Modified Thermal Desorber

Transcription

1 ANALYTICAL SCIENCES VOL.13 SUPPLEMENT Gas Chromatographic Analysis of Atmospheric Sulfur Compounds Using Cryogenic Sampling and A Modified Thermal Desorber Y.C. CHEN and J.G. Lo Department of Nuclear Science, Tsing Hua University, Hsinchu 30043, Taiwan, R. 0. C. A refined cryogenic sampling method and a modification of a commercialized thermal desorber for automatic analysis of atmospheric sulfur compounds with a gas chromatography coupled to the sulfur chemiluminescence detector is reported. Twenty eight reduced sulfur compounds were analyzed on a 30m x 0.32 mm i.d., 4 µm film thickness, Quadrex 007 series (dimethyl silicon phase) capillary column. The problems of retention time shift and peak splitting caused by the water were significantly improved by connecting a 70-cm deactivated column before the analytical column. For one-liter air sample with 80% relative humidity at 3 0 C, the retention time shift can be diminished to 1% for the compounds with boiling point higher than ethanethiol. Keywords: Thermal desorption, sulfur chemiluminescence detector The measurement of atmospheric sulfur species is necessary for various research fields. The task is, however, very difficult because the concentration of the sulfur species in the atmospheric is usually very low, and because the high reactivity of sulfur compounds made its measurement difficult. The alkyithiols contain thiol group which is likely to react with inner surface of the container; Tangerman reported that the methanethiol could be converted into dimethyl disulfide upon exposure to 100 C for one minute.' Various methods for preconcentration of gaseous organic samples have been reviewed by Namiesnik.2 Most of the gaseous sampling methods are useful and have been applied to collect gaseous sulfur species; however none of these methods is widely applicable for all kinds of gaseous sulfur species. Among these methods, the use of solid adsorbents sampling tube are widely used. Molecular Sieve 5A3~, Chromosorb 102 and XAD-type polymers, and the Tenax GC and molecular sieve 5A multi-bed adsorbent tube6 have been used to collect the gaseous sulfur species. However, no single adsorbent exhibits both excellent trapping efficiency and good recovery yield for the simples with wide range of boiling point. The techniques of cryogenic trapping are usually applied to collect the samples containing wide range of boiling point. The coolants such as liquid nitrogen, liquid oxygen and dry ice are frequently used to concentrate a variety of sulfur compounds.'"10 These methods involved the use of an empty U-shape Teflon or deactivated Pyrex-glass tube in accompany with liquid nitrogen (or other coolants)to collect the volatile sulfur compounds, and subsequently a boiling-water bath was used to thermally desorb the samples into the analytical column. These methods are difficult to be automated on the commercialized thermal desorber. The retention time shift, peak splitting, sample loss and damage of column due to the water are the most serious problems in using the cryogenic sampling techniques to collect air samples and subsequently analyzed with capillary gas chromatography. In a 1- L air sample at 25 C and with 100% relative humidity, about 20µL of water can be cryogenically collected. Some drying agents including calcium chloride were used in accompany with cryogenic sampling. Tangerman' reported that the calcium chloride ( CaCl2 2H2O) was a good drying agent for sulfur compounds, the recovery yield of the compounds been test was greater than 90%. Among the common laboratory used drying agents, the most effective drying agents are probably molecular sieve 5A and silica gel. For these reasons, various drying agents were first considered to be test for overcoming the

2 200 ANALYTICAL SCIENCES VOL.13 SUPPLEMENT 1997 interference caused by the water vapor, however the results shown later were disappointing. In 1989, Benner and Stedman published a selective sulfur chemiluminescence detector (SCD)", which is based on the detection of chemiluminescence produced from the reaction of SO with ozone. The SCD exhibits many advantages with which overcome the major drawbacks of the flame photometry detector. The main advantages of SCD are: linear response, the selectivity is to hexane, detection limit is typically below l0"12 g/s and no quenching effect caused by the co-eluted hydrocarbons or sample metrics. With its advantages over the FPD detector, the SCD is gaining its popularity in various sulfur detection12"13 In this work we report a refined cryogenic sampling method and the modification of a commercialized thermal desorber for automatic analysis of atmospheric sulfur compounds with gas chromatography coupled to a SCD. The interference from the water and the method to improve the retention time shift is described. Fig. 1 Cryogenic sampling devices for collecting volatile compounds. Experimental Sampling apparatus Figure 1 shows a schematic diagram of the sampling devices used for cryogenically trapping atmospheric samples. The wall of the container was made from 1 mm thickness of stainless-steel plate. A 10-cm, 318" o.d. and 114" i.d. stainless steel tube was welded near the bottom of the container. For the rate limit of heat transfer from the wall of stainless-steel tube and the glass sampling tube, the time necessary to reach a equilibrium temperature in the sampling tube and the variation of temperature during sampling has to be evaluated. The temperature variation at the center of the cryogenic adsorbent tube were measured by inserting a P-type dermal couple into the tube, the response time of the thermal couple was is. Temperature was recorded with a PC-based analog-todigital converter. Gas chromatography The analysis was performed on a Shimadzu model GC- 15A gas chromatograph (Shimadzu, Japan) with a model 350 flame-based sulfur chemiluminescence detector ( Sievers, USA). The flame ionization detector (FAD) came with the gas chromatograph was used as the interface for the SCD detector. Figure 2 shows the diagram of hyphenation of instruments for the analysis of sulfur compounds. The hydrogen and Fig. 2 Schematic diagram of the thermal desorber and GC-SCD system. air flow rates in the FID were 210 and 380 ml.min"1 respectively. The oxygen pressure supplied to the ozone generator was 5 psi. The tip of the sampling probe was posited 6 mm above the flame jet. The 30 m x 0.32 mm i.d., 4.im film thickness Quadrex (Quadrex Corp., UK) 007 series fused silica capillary column ( dimethyl silicon phase ) was used for sample analysis. A pressure-fitted glass connector ( J&W Scientific Co., USA) was used to connect the analytical column and the 70 cm in length, 0.53 mm i.d. deactivated fused silica microtrap from the cryogenic module. Sample analysis A Tekmar model Aerotrap 6000 thermal desorber (Tekmar, USA) was used for the sample pretreatment. All the sample paths in the thermal desorber were replaced with either PTFE Teflon tubing or a deactivated fused silica column. In the original design of the thermal desorber, a moisture control unit was installed prior to the sample enter the internal trap. The moisture control unit is a pre-cooled 1116" o.d. nickel tube with a rough inner surface which is designed to diminish the interference caused by water,

3 ANALYTICAL SCIENCES VOL.13 SUPPLEMENT Table 1 The Operation conditions of Tekmar thermal decnrher fnr the analvcic of ciilfiir rmmnrnmrl however the moisture control unit may cause serious sample loss and thus was bypassed. A 1/16" o.d. PTFE Teflon tube, packed with 5 mm, mesh Tenax TA (Chrompack, Netherlands) with a piece of deactivated quartz-glass wool plugged at each end, was used as the internal trap to collect the sample desorbed from the sample desorber. For samples desorption, the internal trap was cooled with liquid nitrogen to -160 C, then the sample tube was heated to 200 C and swept with 20 ml.min'' nitrogen for 4 minutes. When the sample was completely desorbed from the sample desorber, the cryotrap module was activated to be cooled to -160 C. Then the internal trap was heated rapidly, and with the six-way port Valco valve (Valco, USA) rotated the sample was swap by the carrier gas to the cryotrap module. The cryotrap module, within which a 70cm, 0.53 mm i.d. deactivated fused silica capillary column filled with 5 mm, 60/80 mesh Tenax TA (as shown in the enlarged portion of Fig. 2) was installed, was used to further focus sample plug desorbed from the internal trap into a smaller sample band. After sample been completely desorbed from the internal trap, the module was heated with full power to the preset temperature to desorb. sample into the analytical column. The sample pretreatment and analysis started automatically after sample tube been placed in the thermal desorber. Table I summarizes the conditions for sample pretreatment in the Tekmar thermal desorber. Retention volume test The method for testing the specific retention volume can be referred elsewhere'. All the Pyrexglass sampling tubes (6 cm x 114" o.d.) were deactivated with dichlorodimethylsilane and packed with 0.2 gram of Tenax TA. Prior to the analysis, the sample tube was conditioned under 280 C with 80 ml.min' nitrogen flow for 8 hours. The dead time at each condition was calculated according to the retention time of methane peak. Recovery of sulfur compounds on drying agent 0.2 gram of each drying agent was packed in a 1/4` o.d. deactivated Pyrex-glass tube. The drying tubes were conditioned under 200 C and purged with 80 ml.min' of purified nitrogen for 4 hours, and then capped with PTFE Teflon end-cap in both end for later use. During sampling, the drying tube was connected to the inlet of the liquid-nitrogen-cooled Tenax TA sampling tube with a 1/4' PTFE Teflon connector. 20 ng of each sulfur compound was spiked into the sample stream for the recovery yields test. The recovery yields were calculated by comparing the chromatographic peak area with and without drying agents. Results and Discussion Retention volume test The specific retention volume of ten sulfur compounds on the Tenax TA is presented in Fig. 3. The specific retention volume of butanethiol at room temperature can not be obtained directly, therefore to obtain the retention data the extrapolation is necessary. The specific retention volume of H2S and CH3SH under room temperature is less than 100 ml. While the specific retention volume of sulfur species on Tenax TA at -196 C is estimated to be greater than 6 x 105 L. Variation of temperature in sampling tube Because of the rate limited of heat transfer between the wall of sampling tube and the stainless tube, the temperature in the sampling tube should be evaluated. Figure 4 shows the actual temperature of the sampling tube during sampling. When the sampling tube were cooled by the liquid nitrogen, an ultimate temperature of approximate -140 C arrived after two minutes (the pump was turned of at this moment). When a pump was turned on, the ambient air wormed up the sampling tube and temperature increased initially. As the sample stream in accompany with water vapor was trapped in the sampling tube, the ice plug caused flow rate to decrease and the temperature decreased again. For safety reason, effective temperature in the sampling tube was estimated to be -110 C and the estimated specific retention volume of H2S at -110 C are 12 L. So 1-L sampling volume is well safe to

4 202 ANALYTICAL SCIENCES VOL.13 SUPPLEMENT 1997 Fig.3 Specific retention volume ( ml/g ) of sulfur compounds on Tenax TA adsorbents. 1, hydrogen sulfide; 2. methanethiol; 3. t-butanethiol; 4. etanethiol; 5, dimethyl disulfide; b. i-propanethiol; 7.s-butanethiol; 8, n-propanethiol; 9. i-butanethiol;10. n-butanethiol. Fig. 5 Comparison of chromatogram of sulfur compounds (A)standard chromatogram from directly injecting of sample into thermal desorber. (B)sample collected on the Tenax TA sampling tube at room temperature. (C)sample collected with cryogenic sampling trap. Column condition: 45 C for lomin, then 5.5 C/min to 240 C. Carrier gas: 1mLJmin N2. Sample identification: l.ethanetluol; 2. i-propanethiol; 3, thiophene; 4. n-butanethiol; 5, dimethyl disulfide. Table 2 The recovery yields of ten sulfur compounds on Tenax TA with cryogenic sampling and in accompany with various sample treatments. Temperature is 25 C and relative humidity is 80% Fig. 4 Variation of temperature in the liquid nitrogen cooled cryogenic sampling tube. collect the most volatile species (H2S) without any breakthrough. Effects of water vapor Figure 5 shows the effect of water vapor on the retention time shift of sulfur species. Fig. SC is a standard chromatogram obtained directly from the injection of standard gas into the thermal desorber and performs the analytical procedure. Fig. 5B is obtained by using the Tenax TA sampling under room temperature followed by the routine analytical procedure. The Tenax TA retains water vapor poorly, therefore the retention time shift caused by water was is found in the Fig. 5B. The retention of euanethiol on Tenax TA is poor under room temperature (see Fig. 3), breakthrough of euanethiol might happen and therefore the peak height of euanethiol is small in Fig.5. In Fig. SA, one liter of gaseous sample is collected on Tenax TA and with cryogenic sampling, and followed the routine analytical procedures. All the samples were collected without any loss, but the water vapor was collected along with sample solutes. When the sample was desorbed from the cryotrap module, the water vapor condensed at the inlet of column and causes to plug the column. The column flow rate decreases, therefore the retention time increases. Furthermore, the flooding water in the column causes the early peak to split.

5 ANALYTICAL SCIENCES VOL.13 SUPPLEMENT Table 3 Effect of length of retention gap on the percent relative humidity of sample was 70% ( the negative sign means the retention time increase in comparison with standard chromatogram) Table 4 Effects of relative humidity on the percent of retention time shift. Sample volume one liter. ( the negative sign means the retention time decrease in comparison with standard chromatogram) Recovery of sulfur compounds on drying agents Table 2 shows the recovery yields often sulfur compounds with various sample treatment. The recovery yields of sulfur compounds on the test drying agents were typically lower than 60%, with an exception of 83% for COS on silica gel. The results illustrate that most samples loss on the drying agent and none of the drying agent is applicable to remove water during cryogenic sampling. The water vapor and other small molecules (the molecular diameter smaller deviation of retention time shift of sulfur compounds. than 4.5A) are likely to diffuse into the internal cavity of the molecular sieve 5A, and the trapping efficiency decreases as the molecular diameter increases (8, 29 and 44% for H2S, COS and CH3SH respectively). The sulfur compounds of which molecular diameter is larger than the internal cavity of molecular sieve 5A are likely to lose due to the adsorption on the surface. Tangerman' reported that calcium chloride is suitable for removing water in sampling sulfur gases (recovery yield was typically greater than 90%), which is contrary to our results (typically the recovery yields are below 40%). The relative humidity of the air sample in Tangerman's report was not described, and the discrepancy might be attributed to the difference in the relative humidity of air samples. Effects of length of retention gap The concept of retention gap has been widely applied to overcome the problems of peak splitting and the retention time shift caused by solvent flooding. Different length of deactivated column was installed before the analytical column for solving these problems, and the results are shown in Table 3. A 70-em deactivated column can diminish the problem of retention time shift effectively. A note should be made here. In the original concept of retention gap, the length of retention gap should be able to accommodate the solvent zone within the retention gap, and the solute band is further focused by either "solvent effect" or the "stationary phase focusing effect" depending on their retention ability' 5. However in our system the deactivated column does not function actually same as a retention gap. The water in the analytical column has two effects on retention time of samples. The first is that the much water may plug the column, which causes the flow rate to decrease. The second is that the water the water covers the surface of stationary phase, which causes e solute to "slip" over the "retention site" and hence the retention time decrease. When the water vapor was thermally desorbed from the cryotrap module, the oven was still under low temperature and the water vapor was condensed again near the pressure-fit connector used to connect the deactivated column and the analytical column. The flow rate decreased. By adjusting the length of the deactivated column, these two effects will reach a "balance" Table 4 shows the effect of relative humidity on sample retention time for one-liter sampling volume. During the analysis, the 70-cm deactivated column was added before the analytical column. For relative humidity up to 80%, the retention time shift for light sulfur species was less than 5% and 1% for the compounds with boiling point higher than ethanethiol. To identify the compounds with low boiling point, the SF6 can be spiked into real samples as a internal standard. The recovery yields of ten sulfur compounds (without use of drying tube during sampling) were presented in last column of Table 2. The lowest recovery yield was 58% for H2S. The reason for the low recovery yield of H2S is not clear, but the dissolution of H2S in water cause to the sample loss to some degree might be expected. Sample analysis The standard chromatogram of twenty eight sulfur

6 204 ANALYTICAL SCIENCES VOL.13 SUPPLEMENT 1997 automatically on the commercialized thermal desorber. Poor recovery yield was found when a drying-agent tube was placed at the inlet of the sampling tube with cryogenic trapping. Installing a 70 cm, 0.53 mm i.d. deactivated fused silica capillary column greatly improves problem of the retention time shift and peak splitting caused by the water collected during the cryogenic sampling. The method has been applied to the measurement of atmospheric sulfur compounds in the petroleum industrial areas where the relative humidity are up to 85 % at 25 C. Fig. 6. Chromatogram of sulfur compounds. (A)standard chromatogram; (B)air sample at a waste water retreatinent plant of a petroleum refinery plant. Column condition: 45 C for 10 mien, then 5.5 C/min to 240 C. Carrier gas: lmlfmin N2. Sample identification: 1. hexafluoro sulfide; 2. hydrogen sulfide; 3.carbonyl sulfide; 4. methanethiol; 5. ethanethiol; 6. dimethyl sulfide; 7. carbon disulfide; 8. i-propanethiol; 9. t- butanethiol; 10. n-propanethiol; 11. s-butanethiol; 12. thiophene; 13. i-butanethiol; 14. diethyl sulfide; 15. n- butanethiol;16. t-pentanethiol;17; dimethyl disulfide; methylthiophene; methylthiophene; 20. Tetrahydrothiophene; 21. n-pentanethiol; 22. diallyl sulfide; ethylthiophene; 24. dipropyl sulfide; 25, diethyl disulfide; 26. n-hexanethiol; 27, n-heptanethiol; 28, dibutyl sulfide; compounds is shown in Fig. 6A. The SF6 was spiked as the internal standard for the identification of H2S and OCS. The samples in Fig. 6A are well separated with an exception of s-butanethiol and thiophene. The initial oven temperature was 45 C, which prevents using sub-ambient oven temperature in analyzing the volatile sulfur compounds. Figure 6B shows the chromatogram of the air sample near a waste water treatment plant. Sample volume was one-liter and the relative humidity was 78% at 31 C. In comparison with Fig. 6A, the retention time shift caused by the water was insignificant in Fig. 6B. and the peak splitting is not found. Conclusions A method for sampling and analysis of atmospheric sulfur compounds is described. With a modification in sampling apparatus and the sample pretreatment, the analysis can be performed Acknowledgment The authors thank for the financial support from the National Science Council of The Republic of China (Grant number NSC M ZA) References 1. A. Tangerman, J. Chromatogr., 366, 205(1986) 2. J. Namiesnik Talanta, 35, 567(1988). 3. M.S. Black, RP Herbst and D.R. Hitchcock,, Anal. Chem., 50, 848(1978) 4. Kangas Juhani and Ryosa Hannus, Chemosphere, 17, 905(1988). 5. A. Przyjazny, J Chromatogr., 333, 327(1985). 6. P.A. Steudler and W. Kijowski, Anal. Chem. 56, (1984). 7. S.O. Farwell, S.J. Gluck, W.L. Bamesberger, T.M. Schutte and D.F. Adams, Anal. Chem., 51, 609 (1979). 8. W.E. Beard and W.D. Guenzi, J Environ. Qual., 12, 113(1983). 9. F.J. Sandalls and S.A. Penkett, Atmos. Environ., 11, 197(1977). 10. P.A. Steudler and W. Kijowski, Anal. Chem., 56, 1432(1984). 11. R. L. Benner and D..H. Stedman, Anal. Chem., 61, 1268(1989). 12. K.K. Gaines, W.H. Chatham and SO. Farwell, HRC.CC.,13, 489(1990). 13. F.P. Di Sanzo, W. Bray and B. Chawla, HRC. CC., 11, 255(1994). 14. R.H. Brown and C.J. Purnell, J Chromatogr.,178, 19(1979). 15. K Grob Jr., G. Karrer and M. L. Riekkola, J Chromatogr., 334, 129(1985).