Anal. Chem. 2004, 76, 5965-5967
Analysis of C2H4 in C2H6 and C2H5D with VUV Absorption Spectroscopy and a Method To Remove C2H4 from C2H6 and C2H5D Hsiao-Chi Lu, Hong-Kai Chen, and Bing-Ming Cheng*
National Synchrotron Radiation Research Center, 101 Hsin-Ann Road, Hsinchu Science Park, Hsinchu 30077, Taiwan
The photoabsorption cross section of C2H4 was measured in the spectral region 107-183 nm and those of C2H6 and C2H5D were accurately determined in the spectral region 107-162 nm using radiation from a synchrotron as source of VUV light. Typically, C2H4 present as a minor impurity in samples of C2H6 and C2H5D distorted the absorption cross section in curves of C2H6 and C2H5D in the onset region. We completely eliminated C2H4 from C2H6 and C2H5D using adsorption on activated Pd/ charcoal at 195 K. By this means, we detected no C2H4 in samples of C2H6 and C2H5D according to their absorption spectra. The detection limit of C2H4 in C2H6 and C2H5D is less than 0.03 ppm with VUV absorption spectroscopy. Ethylene plays a role in many aspects, such as a gaseous phytohormone, which has a profound effect on plant growth and physiology, a pollutant in the atmosphere, and a constituent of comets and the outer planets. Thus, monitoring of ethylene is important in atmospheric chemistry, astronomy, plant physiology, and storage of produce. Gas chromatography,1 mass spectrometry, and IR spectrophotometry are the most selective and useful methods used routinely for ethylene determination. Along with these conventional techniques, analysis of trace amounts of C2H4 has been developed with very high sensitivity by cavity ring-down spectroscopy,2,3 photoacoustic spectroscopy,4-6 and amperometric sensor.7,8 However, these methods require a reference sample, free from analyte gas, or a tedious sampling procedure, which may be difficult to provide in some applications. Therefore, it is worthwhile to establish another spectroscopic method with fingerprint character for analysis of ethylene. VUV absorption * To whom correspondence should be addressed. E-mail: bmcheng@ nsrrc.org.tw. (1) Efer, J.; Mu ¨ ller, S.; Engewald, W.; Knobloch, Th.; Levsen, K. Chromatographia 1993, 37, 361-364. (2) Jongma, R. T.; Boogaarts, M. G. H.; Holleman, I.; Meijer, G. Rev. Sci. Instrum. 1995, 66, 2821-2828. (3) Mu ¨ rtz, M.; Frech, B.; Urban, W. Appl. Phys. B 1999, 68, 243-249. (4) Na¨gele, M.; Sigrist, M. W. Appl. Phys. B 1999, 70, 895-901. (5) te Lintel Hekkert, S.; Staal, M. J.; Nabben, R. H. M.; Zuckermann, H.; Persijn, S.; Stal, L. J.; Voesenek, L. A. C. J.; Harren, F. J. M.; Reuss, J.; Parker, D. H. Instrum. Sci. Tech. 1998, 26, 157-175. (6) Bijnen, F. G. C.; Zuckermann, H.; Harren, F. J. M.; Reuss, J. Appl. Opt. 1998, 37, 3345-3353. (7) Jordan, L. R.; Hauser, P. C. Anal. Chem. 1997, 69, 558-562. (8) Jordan, L. R.; Hauser, P. C.; Dawson, G. A. Analyst 1997, 122, 811-814. 10.1021/ac0494679 CCC: $27.50 Published on Web 08/21/2004
© 2004 American Chemical Society
spectroscopy may be a good candidate that can be applied to analyze the trace of ethylene. Ethane is the second most abundant minor constituent of the outer planets; its concentration is ∼1/100 that of methane.9 To understand the photochemistry and D/H ratio in the upper atmospheres of giant planets, absolute absorption cross sections of C2H6 and its isotopomers have been extensively investigated in the vacuum ultraviolet (VUV) region either by single-photon absorption10-18 or by electron impact excitation.19-24 We have determined its absorption cross section in the VUV for the purpose of modeling the isotopic fractionation of C2H5D in the upper atmosphere of Jupiter.18 In previous work, we noticed that it is difficult to measure accurately the absolute absorption cross sections of C2H6 and its isotopomers in the VUV, especially in the onset region, because values are small and susceptible to distortion from impurities that absorb strongly in that region. Typically, commercial samples of alkanes contain alkenes in small proportions; for instance, the nominal “research purity” 99.99% of C2H6 supplied by Matheson Gas contains the greatest impurity C2H4 at less than 100 ppm with specification. It is possible that C2H4 may have a profound effect on distortion of the VUV absorption spectrum of ethane. Thus, precise analysis of a minute quantity of C2H4 in C2H6 requires verification in VUV absorption spectroscopy. To avoid susceptible data in the measurement of (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19)
(20) (21) (22) (23) (24)
Strobel, D. F. Astrophys. J. 1974, 192, L47-L49. Okabe, H.; Becker, D. A. J. Chem. Phys. 1963, 39, 2549-2555. Raymonda, J. W.; Simpson, W. T. J. Chem. Phys. 1967, 47, 430-448. Lombos, B. A.; Sauvageau, P.; Sandorfy, C. J. Mol. Spectrosc. 1967, 24, 253-269. Pearson, E. P.; Innes, K. K. J. Mol. Spectrosc. 1969, 30, 232-240. Koch, E. E.; Skibowski, M. Chem. Phys. Lett. 1971, 9, 429-432. Custer, E. M.; Simpson, W. T. J. Chem. Phys. 1974, 60, 2012-2020. Mount, G. H.; Moos, H. W. Astrophys. J. 1978, 224, L35-L38. Kameta, K.; Machida, S.; Kitajima, M.; Ukai, M.; Kouchi, N.; Hatano, Y.; Ito, K. J. Electron. Spectrosc. Relat. Phenom. 1996, 79, 391-393. Lee, A. Y. T.; Yung, Y. L.; Cheng, B.-M.; Bahou, M.; Chung, C.-Y.; Lee, Y.-P. Astrophys. J. 2001, 551, L93-96. Lassettre, E. N.; Skerbele, A.; Dillon, M. A. J. Chem. Phys. 1967, 46, 45364537. Lassettre, E. N.; Skerbele, A.; Dillon, M. A. J. Chem. Phys. 1968, 48, 539. Lassettre, E. N.; Skerbele, A.; Dillon, M. A. J. Chem. Phys. 1968, 49, 23822395. Johnson, K. E.; Kim, K.; Johnston, D. B.; Lipsky, S. J. Chem. Phys. 1979, 70, 2189-2197. Robin, M. B. Higher Excited States of Polyatomic Molecules; Academic Press: New York, 1974; Vol. I and 1985; Vol. III. Dillon, M. A.; Tanaka, H.; Spence, D. J. Chem. Phys. 1987, 87, 1499-1501. Au, J. W.; Cooper, G.; Burton, G. R.; Olney, T. N.; Brion, C. E. Chem. Phys. 1993, 173, 209-239.
Analytical Chemistry, Vol. 76, No. 19, October 1, 2004 5965
the absorption cross section in the VUV for C2H6; the removal of C2H4 from C2H6 also needs to be established. However, the purification of C2H4 from C2H6 is not a trivial task: attempts to purify the sample with vacuum distillation failed to eliminate C2H4. In the present work, special attention and effort has yielded success in obtaining accurate values of cross sections for C2H6 and C2H5D in the VUV range. EXPERIMENT The experimental setup for spectroscopy is similar to that described previously.25 The absorption cross section was measured with a double-beam apparatus. VUV light, produced in the National Synchrotron Radiation Research Center in Taiwan, was dispersed with a high-flux cylindrical grating monochromator (focal length 6 m) beam line. The intensity of light, partially reflected from a LiF window oriented at 45° to the beam line and located before the gas cell, was monitored as a reference. Both reflected and transmitted beams, passed through similar optical components so that they have similar spectral distributions, were converted to visible light with sodium salicylate coated on a glass window before detection with a photomultiplier tube (Hamamatsu R943-02) in a photon-counting mode. The wavelength was calibrated with absorption lines of CO,26,27 O2,28 and H2O;29 the accuracy of measured wavelength is limited by the scan step. Two gas cells with path lengths l ) 38.5 cm and l ) 8.9 cm were used in this work. To maintain a constant gas pressure during data acquisition, we connected a reservoir of volume 1.38 L (or 0.62 L) to the long (or short) gas cell. The density of the gas was determined from the pressure recorded with a capacitance manometer (MKS, Baratron) and the temperature monitored with a thermocouple. Absorption spectra of gaseous samples were recorded at pressures in a range 0.040-2063 Torr at 298 K. The absorption cell was evacuated to ∼3 × 10-8 Torr before filling with a fresh sample to a specific pressure for each measurement. At each wavelength, absorbance determined at 5-11 pressures was plotted against number density and fitted with a least-squares method to a line to yield an absorption cross section according to Beer’s law. Data with absorbance () ln(I0/I)) greater than 2.0 were discarded to avoid effects of saturation. RESULTS AND DISCUSSION Resultant absorption cross sections of C2H6 in the spectral region 107-183 nm are displayed in Figure 1a with linear scale presentation. The general shape and values of the absorption cross section curve of C2H6 are similar to those reported by earlier workers. A more detailed discussion on absorption systems of C2H6 is presented elsewhere. To estimate whether C2H4 affects the data of C2H6, we also recorded absorption cross sections of C2H4 in the spectral range 107-183 nm, as shown in Figure 1b for comparison. The corresponding curve and values of cross sections for C2H4 agree satisfactorily with previous reports.30-35 (25) Cheng, B.-M.; Chew, E. P.; Liu, C.-P.; M. Bahou. Lee, Y.-P.; Yung, Y. L.; Gerstell, M. F. Geophys. Res. Lett. 1999, 26, 3657-3660. (26) Tilford, S. G.; Vanderslice, J. T.; Wilkinson, P. G. Can. J. Phys. 1965, 43, 450-456. (27) Simmons, J. D.; Bass, A. M.; Tilford, S. G. Astrophys. J. 1969, 155, 345358. (28) Ogawa, S.; Ogawa, M. Can. J. Phys. 1975, 53, 1845-1852. (29) Wang, H. T.; Felps, W. S.; McGlynn, S. P. J. Chem. Phys. 1977, 67, 26142628.
5966 Analytical Chemistry, Vol. 76, No. 19, October 1, 2004
Figure 1. Absorption cross sections in the wavelength region 107183 nm for (a) C2H6 (solid curve) and (b) C2H4 (dash curve).
The absorption spectrum of C2H6 differs markedly from that of C2H4. For wavelength greater than 143 nm, the absorption cross section of C2H6 decreases continuously with increasing wavelength. Above 154 nm, the value decreases rapidly, by approximately one-tenth in each successive 2-nm interval, and reaches 0.0004 Mb (1 Mb ) 1 × 10-18 cm2) at ∼161.5 nm.36 In contrast, C2H4 has strong absorption in the wavelength region 143-175 nm, including characteristic features possessing absorption cross sections 20-55 Mb in the wavelength region 155-175 nm. Measurement of the absorption cross section of C2H6 near the onset region is consequently extremely sensitive to C2H4 present as a minute impurity that absorbs strongly in that region. Hence, it is convenient to estimate the proportion of impurity C2H4 in a sample of C2H6 by analyzing the absorption features of a sample in a wavelength region 160-175 nm. A curve for the absorption cross section of unpurified 99% C2H5D is depicted in Figure 2b as a semilogarithmic plot, with that of C2H4 shown in Figure 2a for comparison. The curve of unpurified 99% C2H5D has a long tail extending to greater wavelengths and shows characteristic features of C2H4 in the wavelength region 160-180 nm. According to data for analysis provided by the supplier of C2H5D (Isotec Inc.) from gas chromatographic analysis, the sample contains C2H4 at 542 ppm. Cross sections of C2H5D in the onset region are seriously distorted from this contaminating C2H4, which absorbs strongly in that region. As mentioned above, to eliminate C2H4 from C2H5D is not a trivial task: attempts to purify the C2H5D with vacuum distillation failed to get rid of C2H4. To remove alkene from alkane, a general method is hydrogenation, which involves conversion of a carbon-carbon double bond into a single bond. For this purpose, a finely divided metals typically platinum, palladium, or nickelsas catalyst effects addition (30) Holland, D. M. P.; Shaw, D. A.; Hayes, M. A.; Shpinkova, L. G.; Rennie, E. E.; Karlsson, L.; Baltzer, P.; Wannberg, B. Chem. Phys. 1997, 219, 91116. (31) Zelikoff, M.; Watanobe, K. J. Opt. Soc. Am. 1953, 43, 756-759. (32) Person, J. C.; Nicole, P. P. J. Chem. Phys. 1968, 49, 5421-5426. (33) Wilkison, P. G.; Mulliken, R. S. J. Chem. Phys. 1955, 23, 1895-1907. (34) McDiarmid, R. J. Phys. Chem. 1980, 84, 64-70. (35) Xia, T. J.; Chien, T. S.; Wu, C. Y. R.; Judge, D. L. J. Quant. Spectrosc. Radiat. Transfer 1991, 45, 77-91. (36) For absorption cross sections of C2H4, C2H6, and C2H5D, please visit the web site http://ams-bmc.nsrrc.org.tw.
Figure 2. Absorption cross sections in the wavelength region 107183 nm for (a) C2H4 (purity 99.99%), (b) C2H5D (purity 99%, without further purification), and (c) C2H5D (purified with activated Pd/charcoal cooled to 195 K).
of dihydrogen across the double bond. This catalytic hydrogenation is generally performed with excess hydrogen to suppress the reverse reaction, the dehydrogenation of an alkane; after hydrogenation, great care is required to remove H2 from the alkane. A further issue in this case is that the product of hydrogenation of C2H4 is C2H6, which is difficult to separate from C2H5D. To remove C2H4 from sample C2H5D, hydrogenation is thus not feasible. As the first step of hydrogenation involves an alkene binding with the surface of the catalyst across the carbon-carbon double bond, adsorption of alkene on a metal catalyst might be applied to effect purification; C2H4 might be retained on the surface of the catalyst and thus become removed from gaseous C2H5D. According to this scheme, we first activated an absorbent, Pd on charcoal, at 425 K for 10 h, cooled it to 195 K, and then passed the sample slowly through it to the gas cell. Figure 2c shows the resulting curve of C2H5D cross section. The tail of the curve beyond 160.7 nm contains only noises without the detectable characteristic feature of C2H4. Hence, C2H4 is totally removed from C2H5D by this adsorption method. For the case of C2H6, Figure 3a displays a curve for the absorption cross section of an unpurified sample (Matheson, 99.99%); this curve shows a weak but long tail extending from 159 to 180 nm. With this gaseous sample passed slowly through a copper U tube cooled to 195 K, we obtained the curve shown in Figure 3b; the tail decreased somewhat, but the pattern of features due to C2H4 remains present. Passing the same sample slowly through a copper U tube packed with activated Pd on charcoal at 195 K, we obtained the curve shown in Figure 3c; in which the tail of the curve beyond 161.5 nm contains only noises but no detectable absorption peak of C2H4. The impurity has clearly been completely eliminated from our sample by this means. Our data in the shorter wavelengths generally agree with previous work within experimental uncertainties. The values reported by Mount and Moos16 are also presented in Figure 3 for comparison. Near the threshold region, our values are smaller because our samples are free from interferences from absorption of trace impurities. In our experimental system, the noise level of measured absorbance is ∼0.005 but can be improved to 0.001. When we
Figure 3. Absorption cross sections in the wavelength region 107183 nm for (a) C2H6 (nominal purity 99.99%, as received), (b) C2H6 (purity 99.99%, after slow passage through a U tube cooled to 195 K), and (c) C2H6 (purified with activated Pd on charcoal at 195 K). The circles are the values obtained by Mount and Moos.16
recorded an absorption spectrum of purified C2H6 at a pressure of 2063 Torr with a path length of 38.5 cm in our gas cell, no absorption feature of C2H4 was detected. Based on a value of 56 Mb for the cross section of C2H4 at 170.1 nm, we derived the detection limit of C2H4 in C2H6 to be as small as 0.03 ppm, which might be improved to 0.006 ppm in our system. This great sensitivity of detection demonstrates the potential and utility of vacuum ultraviolet absorption spectra in analytical applications. CONCLUSIONS Usually, the major impurity in alkane is alkene. For instance, ethylene is the major impurity in an ethane sample. The impurity of alkene can affect alkane in many respects. Concerning the spectroscopy of target molecules in this work, we noticed the contamination of ethylene distorts the VUV absorption spectra of ethane, and the concentration of ethylene in ethane can be analyzed with VUV absorption spectroscopy. It is very difficult to remove ethylene from sample ethane by the vacuum distillation method. In this work, we reported that ethylene can be completely eliminated from a sample of ethane by using the adsorption scheme, passing the impure ethane sample through the activated Pd on charcoal at 195 K. We detected no signal of a VUV absorption peak of ethylene in a purified sample. That means the contamination of ethylene in purified ethane is below the detection limit of 0.03 ppm. The successful results of this work can be applied to other alkane samples, and the findings, we believe, are significant to many fields. ACKNOWLEDGMENT We thank the National Science Council of the Republic of China and the National Synchrotron Radiation Research Center for support.
Received for review April 7, 2004. Accepted July 8, 2004. AC0494679
Analytical Chemistry, Vol. 76, No. 19, October 1, 2004
5967
