Gas chromatography flow rates for determining ... - Wiley Online Library

1 downloads 0 Views 135KB Size Report
The effects of the gas chromatography flow rate on the determination of the deuterium/hydrogen. (D/H) ratios of natural gas utilising gas ...
RAPID COMMUNICATIONS IN MASS SPECTROMETRY Rapid Commun. Mass Spectrom. 2008; 22: 2521–2525 Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/rcm.3641

Gas chromatography flow rates for determining deuterium/hydrogen ratios of natural gas by gas chromatography/high-temperature conversion/isotope ratio mass spectrometry Wanglu Jia*, Ping’an Peng and Jinzhong Liu State Key Laboratory of Organic Geochemistry (SKLOG), Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou, GD 510640, P.R. China Received 6 March 2008; Revised 22 April 2008; Accepted 15 June 2008

The effects of the gas chromatography flow rate on the determination of the deuterium/hydrogen (D/H) ratios of natural gas utilising gas chromatography/high-temperature conversion/isotope ratio mass spectrometry (GC/TC/IRMS) have been evaluated. In general, the measured dD values of methane, ethane and propane decrease with increase in column flow rate. When the column flow rate is 1 mL/min or higher, which is commonly used for the determination of D/H ratios of natural gas, the organic H in gas compounds may not be completely converted into hydrogen gas. Based on the results of experiments conducted on a GC column with an i.d. of 0.32 mm, a GC flow rate of 0.6 mL/min is proposed for determining the D/H ratios of natural gas by GC/TC/IRMS. Although this value may be dependent on the instrument conditions used in this work, we believe that correct dD values of organic compounds with a few carbon atoms are obtained only when relatively low GC flow rates are used for D/H analysis by GC/TC/IRMS. Moreover, as the presence of trace water could significantly affect the determination of D/H ratios, a newly designed inlet liner was used to remove trace water contained in some gas samples. Copyright # 2008 John Wiley & Sons, Ltd.

The deuterium/hydrogen (D/H) ratios of natural gas components (e.g., methane, ethane and propane) are valuable in geological and environmental studies1–4 due to their great usefulness for source evaluation. Since the technique of hightemperature conversion (TC) was developed and proved to be able to convert H in organic compounds into hydrogen gas quantitatively,5 subsequently commercially available instruments6 using gas chromatography/high-temperature conversion/isotope ratio mass spectrometry (GC/TC/IRMS) have been of more benefit to the determination of D/H ratios of natural gas3,4,7–10 than the previously available off-line methods. Several studies have discussed the TC conditions that could have significant effects on the measured D/H ratios of various samples.5,6,8,11 Basically, before the ceramic tube (Al2O3) in the TC reactor can be used for routine analysis, the tube must be graphitised (conditioned) by passing a relatively large volume of methane at the working temperature (>14008C) through it.8,9,11 This conditioning process is *Correspondence to: W. Jia, State Key Laboratory of Organic Geochemistry (SKLOG), Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Guangzhou, GD 510640, P.R. China. E-mail: [email protected] Contract/grant sponsor: Natural Science Foundation of China; contract/grant number: 40403005.

especially essential for hydrogen isotopic analyses of organic compounds with low carbon numbers.11 Another important parameter for the TC process is the flow rate of the GC carrier gas used to determine the residence time of organic compounds in the TC reactor. A detailed study has shown that when the flow rate of carrier gas passing through the TC reactor was less than 1.1 mL/min, propane could be converted into hydrogen gas quantitatively.5 Other studies have suggested that the measured D/H ratios by GC/TC/IRMS were correct only when the GC flow rate was relatively low, and that the acquired D/H data for some short-chain alcohols were very close to the calibrated values obtained using a TC/ EA (high-temperature conversion/elemental analyser) when the GC flow rate was 0.8 mL/min.11 However, based on the published D/H data for methane obtained by GC/TC/IRMS, when GC columns with identical i.d. (0.32 mm) are used, the GC flow rates can vary from 0.4 mL/min to 2.6 mL/min,3,4,7–9 and are usually much higher than that used for the analysis of propane5 and short-chain alcohols.11 Hence, the question arises as to whether the GC flow rate has a significant effect on the measured D/H data for methane, ethane and propane. In this work experiments were carried out to determine the D/H ratios of natural gas components by GC/TC/IRMS under different GC flow rates, and to determine the optimal flow rate for our instrument configurations. Copyright # 2008 John Wiley & Sons, Ltd.

2522 W. Jia, P. Peng and J. Liu

EXPERIMENTAL Analytical conditions of GC/TC/IRMS The D/H ratios of natural gas were determined by a Deltaþ XL GC/TC/IRMS instrument (Finnigan, Bremen, Germany). A SiC heater (with initial resistance value of 5 V, manufactured by Lufeng Co. Ltd, Zibo, P.R. China) was used in the TC reactor since the original SiC heater installed in the instrument had burned out. The length of the heating zone (120 mm) and the i.d. (9.5 mm) of the SiC heater were designed to be identical to those of the heater from Finnigan. In addition to the SiC heater, the TC temperature controlling units and other instrument configurations are the same as those of the commercial instruments from Finnigan. The 6890 gas chromatograph (Agilent, Santa Clara, CA, USA) was installed with a HP Plot Q column (30 m  0.32 mm i.d.  20 mm film thickness; Agilent), and a temperature profile of 508C (4–8 min) to 1808C (5 min) at a rate of 158C/ min was used for separation of the compounds. The inlet of the gas chromatograph was usually maintained at 1008C, and the carrier gas was He (5.0 grade). A split injection mode was used for sample introduction and the split ratio was set to 10:1. The TC reactor was set to a temperature of 14308C. The isotope ratio mass spectrometer was tuned carefully before the experiments to obtain a relatively stable and low H3 factor (6.5, used for Hþ 3 correction) and a relatively low background (190 mV) at the mass2 collector with full He flow (GC flow rate at 1.5 mL/min), i.e. with the TC heater at working temperature, the ’back flush’ on and the ’open split’ in, being introduced into the mass spectrometer. A new Al2O3 tube (320 mm  0.5 mm i.d., Finnigan) was installed and graphitised by the method introduced by Bilke and Mosandl11 before being used in the measurements. In the back flush mode, a methane flow (0.5 mL/min) was mixed with the carrier gas (He) and passed through the Al2O3 tube for 4–5 min at 14308C. During the D/H measurements of a gas sample, the GC flow rate was altered randomly from run to run to eliminate the effects of fluctuations from the instrument.

Calibration methods The H2 reference gas was calibrated against two International Atomic Energy Agency (IAEA, Vienna, Austria) water standards (Vienna Standard Mean Ocean Water (VSMOW) and Greenland Light Snow Precipitation (GLSP) with dD values of 0% and 189.5%, respectively). The dD values of a natural gas sample (HGS1, 76% CH4, 10.4% C2H6 and 4.3% C3H8) used in the experiments were calibrated against a methane gas standard (dD: 160.8  2.1%, from Arndt Schimmelmann, Indiana University, Bloomington, IN, USA). The glass tube (20 cm  0.6 cm i.d.) containing the methane standard was connected to a glass line equipped with an outlet sealed by a silicone septum in a cap, and the glass line was evacuated to vacuum by a rotary pump. The glass tube was then cracked to release methane into the evacuated glass line. A gas-tight syringe with a PTFE valve (Hamilton, Reno, NV, USA) was used to introduce the methane gas from the outlet of the glass line into the GC inlet of the GC/TC/IRMS instrument. Typically, 5–10 mL natural gas injected (10:1 split ratio) gave a CH4 signal with peak Copyright # 2008 John Wiley & Sons, Ltd.

height of more than 3000 mV in the mass2 chromatogram by GC/TC/IRMS, corresponding to 15–30 nmol CH4 on-column. For the analysis of ethane and propane, 30–45 mL gas was injected since the gas samples had a relatively low content of these two compounds. The dD values of the C1–C3 hydrocarbons in the HGS1 sample were calibrated by making alternate measurements of the HGS1 gas and the methane standard 15 times, and a GC flow rate of 1.5 mL/min was used. A correction factor was calculated by subtracting the average dD value measured for the methane standard from the dD value determined by the off-line method (Arndt Schimmelmann, Indiana University, Bloomington, IN, USA). The measured dD values for the gas components from the HGS1 gas sample were corrected using this factor, and the average corrected results were accepted as the calibrated dD values for HGS1. The standard deviations (SD, 1s, n ¼ 15) of the measured dD values for methane from both the HGS1 gas sample and the methane standard were less than 2.0%, and those for ethane and propane were 2.6% and 3.1%, respectively. The relatively higher SD exhibited by these measured dD values for ethane and propane might be a result of slightly higher baseline with increasing GC oven temperatures. All the dD values in this paper are reported relative to VSMOW.

RESULTS AND DISCUSSION dD values of gas components determined at different GC flow rates The measured dD values of methane from the HGS1 gas sample using two Al2O3 tubes are shown in Fig. 1. Generally, the dD values of methane decrease with increase in GC flow rates. All the measuring procedures and GC conditions for the two TC tubes are the same, but the acquired results show slight differences. For tube 1 (Fig. 1(A)), the measured dD values of methane decrease relatively smoothly with increase in GC flow rates. There are relatively small differences among the measured dD values of methane when the column flow rates range from 1.2 mL/min to 2.5 mL/min for tube 2 (Fig. 1(B)). This phenomenon could be caused by slight differences between the reactor tubes.8 For the two reactor tubes, the dD values of methane measured at GC flow rates of 0.4 and 0.6 mL/min show little difference and they are very close to the value calibrated against the methane standard. The dD values of methane measured at GC flow rates greater than 0.8 mL/min are 6–14% lower than the calibrated values. For some biogenetic gases, the hydrogen gas in the sample is relatively abundant and could disturb the D/H measurements of methane either due to incomplete separation or to the significantly low dD values of hydrogen gas (14008C) in the TC reactor.11 Moreover, organic compounds with only a few carbon atoms, like the C1–C3 components in natural gas, may be more difficult to convert into hydrogen gas than compounds with high carbon numbers. Relatively high GC flow rates, which are generally used to determine the residence time of organic compounds in the TC reactor, may lead to unquantitative conversion of organic H into hydrogen gas. The optimised value for the GC flow rate may vary with both the TC heater being used and the temperature set at the heater. Based on the instrument configuration in this work, a 0.6 mL/min flow rate was chosen so that accurate results could be obtained. The residence time of a sample run at a flow rate of 0.6 mL/min was shorter than that run at 0.4 mL/ min. When the GC temperature profiles were identical, methane eluted at 410 s (Fig. 4) and 620 s; these corresponded to GC flow rates of 0.6 mL/min and 0.4 mL/min, respectively. Significant peak broadening was observed in the chromatograms as a result of relatively low flow rates (Fig. 4). The relatively wide peaks of methane, ethane and propane at a GC flow rate of 0.6 mL/min are still satisfactory, as shown by the isotopic results, with good precisions indicating no disturbance being caused to the peak integrations from peak broadening (Figs. 1 and 3).

Design of a water-removing GC inlet liner When the humidity in the laboratory was relatively high, a significant tailing peak that occurs prior to propane during the D/H analysis became more and more pronounced with an increase in the number of sample runs. This problem also occurred during the D/H analysis of natural gas sampled together with wet core mud in sealed pots. Via injection of 1 mL samples of water with a large split ratio, the tailing peak was assigned as being due to water. Theoretically speaking, such a trace amount of water could also be converted into H2 and CO in the TC reactor and consume the black carbon deposited on the internal surface of the Al2O3 tube during the process of graphitisation. Hence, efforts should be made to prevent water from entering the Al2O3 tube, instead of separating water and propane chromatographically. Small modifications were made to the inlet liner (split/ splitless, Agilent, Forest Hill, Australia): First, glass wool seated in the upper half part of the liner was picked out, and then a small volume of glass wool was separated and plugged into the bottom of the liner; Secondly, some P2O5 granules (total volume occupied is 350 mL, Merck, Darmstadt, Germany), usually used for water-removing in the EA, were carefully packed into the lower half part of the liner; Finally, the glass wool was plugged back into the upper half part of the inlet liner. A working gas sample composed of 10% methane, 5.2% ethane and 3.3% propane (40 mL injected with a split ratio 10:1) was analysed to test the effect of the added P2O5 on the D/H determination (Fig. 5). The gas sample was extracted from the gas cylinder into a U-shaped glass tube (1.5 cm i.d.) before isotopic analysis. The ’U’ tube charged with water was sealed at one end by a silicone septum in a cap (the air at this end was expelled by slightly inclining the ’U’ tube and tightening the cap), and the gas sample was introduced into the ’U’ tube by piercing the septum with a needle connected to the gas cylinder. Thus, the positioning of the gas sample between the septum and

Figure 4. Typical chromatograms (mass2) obtained from D/H analysis of HGS1 gas by GC/TC/IRMS. Those numbers bracketed denote the peak width defined by the IsodatNT 2.0 software (Finnigan). Copyright # 2008 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2008; 22: 2521–2525 DOI: 10.1002/rcm

Determining D/H ratios of natural gas by GC/TC/IRMS

2525

CONCLUSIONS The GC flow rate proved to be one of the critical factors that could affect the D/H ratios of natural gas measured by GC/TC/IRMS. The optimised flow rate could be dependent on the instrument configuration, and more probably on the SiC heater used in the TC reactor. According to the experimental results with our instrument, a GC flow rate of 0.6 mL/min can be accepted as the optimal value for determining the D/H ratios of natural gas by GC/TC/ IRMS. Based on published data concerning the effects of GC flow rate,5,11 our results suggest that for GC/TC/IRMS instruments commercially available at this time, relatively small GC flow rates should be preferred in order to obtain the correct dD values11 in determinations of the D/H ratios of natural gas or organic compounds with a few carbon atoms.

Acknowledgements The authors appreciate the efforts of Dr Wei Gangjian and Dr Deng Wenfeng for calibration of H2. Several constructive suggestions provided by two anonymous reviewers were greatly helpful to improve the manuscript. The author also thanks the editor for some useful comments. This work was supported by the Natural Science Foundation of China (Grant No. 40403005). Figure 5. Chromatograms (mass2) obtained from D/H analysis of the working gas utilising GC/TC/IRMS before and after the GC inlet liner was packed with P2O5. Table 1. dD values of a working gas sample measured before and after the GC inlet liner was packed with P2O5 utilising GC/ TC/IRMS CH4

C2H6

C3H8

dD (%)SD (1s, n ¼ 15) Original liner Liner packed with P2O5

141.0  1.3 142.0  2.2

167.9  2.9 168.4  2.3

167.0  2.9 167.2  3.0

the water column was similar to natural gas co-existing with water. The measured dD values for the C1–C3 hydrocarbons after the liner had been packed with P2O5 agreed well with the values measured with the original liner (Table 1). After these modifications, the H2O in the gas sample could be removed completely (Fig. 5). The inlet liner packed with P2O5 was still effective after a few weeks.

Copyright # 2008 John Wiley & Sons, Ltd.

REFERENCES 1. Chen JF, Xu YC, Huang DF. AAPG Bull. 2000; 84: 591. 2. Snover AK, Quay PD, Wei MH. Global Biogeochem. Cycles 2000; 14: 11. 3. Rice AL, Gotoh AA, Ajie HO, Tyler SC. Anal. Chem. 2001; 73: 4104. DOI: 10.1021/ac0155106 4. Kinnaman FS, Valentine DL, Tyler SC. Geochim. Cosmochim. Acta 2007; 71: 271. DOI: 10.1016/j.gca.2006.09.007 5. Burgoyne TW, Hayes JM. Anal. Chem. 1998; 70: 5136. DOI: 10.1021/ac980248v 6. Hilkert AW, Douthitt CB, Schlu¨ter HJ, Brand WA. Rapid Commun. Mass Spectrom. 1999; 13: 1226. 7. Yamadaa K, Ozakic Y, Nakagaw F, Tanakac M, Yoshidaa N. Atmos. Environ. 2003; 37: 1975; DOI: 10.1016/S1352-2310(03) 00030-X 8. Potter J, Siemann MG. Rapid Commun. Mass Spectrom. 2004; 18: 175. DOI: 10.1002/rcm.1311 9. Henning M, Stra˛poc´ D, Lis GP, Sauer P, Fong J, Schimmelmann A, Pratt LM. Rapid Commun. Mass Spectrom. 2007; 21: 2269. DOI: 10.1002/rcm.3083 10. Tarasova OA, Brenninkmeijer CAM, Assonov SS, Elansky NF, Ro¨ckmann T, Brass M. Atmos. Environ. 2006; 40: 5617. DOI: 10.1016/j.atmosenv.2006.04.065 11. Bilke S, Mosandl A. Rapid Commun. Mass Spectrom. 2002; 16: 468. DOI: 10.1002/rcm.599

Rapid Commun. Mass Spectrom. 2008; 22: 2521–2525 DOI: 10.1002/rcm