- Email: [email protected]
Line Shift Investigations for Different Isotopomers of Carbon Monoxide
JOURNAL OF MOLECULAR SPECTROSCOPY ARTICLE NO.
190, 226 –231 (1998)
MS987595
Line Shift Investigations for Different Isotopomers of Carbon Monoxide B. Sumpf,*,1 J. P. Burrows,† A. Kissel,* H.-D. Kronfeldt,* O. Kurtz,* I. Meusel,* J. Orphal,† and S. Voigt† *Optisches Institut der Technischen Universita¨t Berlin, Sekretariat PN 0 –1, Hardenbergstr. 36, 10623 Berlin, Germany; and †Institut fu¨r Umweltphysik, Universita¨t Bremen, P.O. Box 330440, 28334 Bremen, Germany Received September 18, 1997; in revised form April 17, 1998
Line shift coefficients for five lines of five different isotopomers in the fundamental band of CO in the spectral region near 2058 cm21 were measured using a three channel lead salt diode laser spectrometer. The study includes the lines P(3) of 13 17 C O, R(3) of 13C18O, P(9) of 12C18O, P(10) of 13C16O, and P(21) of 12C16O, and covers collisions with N2, O2, H2, D2, He, Ne, Ar, Kr, and Xe. Line shifts of the isotopomers 13C16O, 12C18O, 13C18O, and 13C17O were determined for the first time. Within the experimental uncertainty no significant dependence of the shift effect on the isotopomer was found. The R-branch line under study shows a smaller line shift coefficient than a P-branch line with a similar rotational quantum number. With increasing mass of the noble gas perturber the absolute size of the shift coefficient increases. Moreover self- and nitrogenbroadening coefficients for the isotopomer lines were determined. Compared to previous measurements no significant deviations between different isotopomers were observed. © 1998 Academic Press INTRODUCTION
Carbon monoxide belongs to the most important atmospheric trace gases. The concentrations in different layers of the troposphere, the stratosphere, and the mesosphere vary strongly. Due to a short atmospheric lifetime of about 2 months, changes in sources and sinks in the environment are reflected in the atmospheric concentrations faster compared to other molecules (1). Prerequisite for the application of spectroscopic methods for the determination of CO concentrations is the accurate knowledge of line positions, line strengths, line broadening, and line shift coefficients. For line positions and strengths information for all atmospheric important isotopomers is available (most abundant 12C16O, natural abundance 98.7%; 13C16O, 1.1%; 12 18 C O, 0.2%; 12C17O, 0.04%; 13C18O, 0.0023%; and 13C17O, 0.0004%). However, for line broadening and shift, data are only known for the most abundant species. Draegert and Williams (2) published comprehensive line broadening investigations for the major isotopomer for collisions with CO, H2, D2, N2, He, Ne, Ar, Kr, and Xe in the fundamental band. Self-broadening coefficients in the first overtone band are known from Hunt et al. (3). Collisions with diatomic perturbers in the (1 4 0) band of 12C16O were analyzed by Bouanich and Brodbeck (4) and LeMoal and Severin (5). A tunable diode laser spectrometer was used by Hartmann et al. (6) to study the nitrogen and argon broadening in the fundamental band. Theoretical calculations of the collisional induced effects were carried out by Hartmann et al. (7). 1
Present address: ELIGHT Laser Systems GmbH, Warthestr. 21, 14513 Teltow, Germany.
Recently, Mannucci (8) extended the knowledge of the broadening by investigating collisions with helium and hydrogen, whereas Hamdouni et al. (9) studied broadening with N2 as buffer gas. In the microwave region Beaky et al. (10) carried out experiments concerning the H2 broadening of two transitions. Recently, Sinclair et al. (11) presented data on line broadening, shifting, and mixing for collisions with N2. Within the uncertainties of the experimental data a good agreement of the values for a given perturber and a given rotational quantum number can be stated. A pronounced decrease of the broadening coefficients with increasing rotational quantum number J0 and no vibrational dependence were observed. Line shift investigations for the major isotopomer 12C16O are less frequent. The fundamental band of 12C16O was studied by Mannucci (8) (collisions with He and H2), Hamdouni et al. (9) (N2), and Bouanich et al. (12) (N2, O2, He, Ne, Ar, Xe), and the first overtone band by Bouanich et al. (13) (He, Kr, N2, O2). The data show significant differences between the fundamental and overtone bands and within one band a different rotational quantum number dependence for R- and P-branch lines. Up to now no data concerning line shifts for other isotopomers have been known. Recently, the first analysis of the nitrogen broadening of 13 16 C O lines in the (2 4 0) band using a Fourier Transform Spectrometer were carried out by Voigt et al. (14). The broadening coefficients given in this work are on average 6% smaller than those given for 12C16O by other authors [Le Moal et al. (5), Hartmann et al. (6), Hamdouni et al. (9), and Sinclair et al. (11)]. Even slightly larger values are reported by Bouanich and Brodbeck (4). In the work by Voigt et al. (14), additional
226 0022-2852/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
LINE SHIFT FOR DIFFERENT ISOTOPOMERS OF CO
227
experiments were carried out to exclude systematic uncertainties. Broadening coefficients for transitions of 12C16O were determined. A good agreement with previously published data (5, 6, 9, 11) can be stated. Therefore it seems to be necessary to carry out new experiments with another experimental method to determine the source of the deviation mentioned above. The aim of this work was to determine for the first time line shift coefficients for rare carbon monoxide isotopomers for collisions with N2, O2, H2, D2, He, Ne, Ar, Kr, and Xe. Moreover, self- and nitrogen-broadening coefficients were measured for these less abundant isotopomers. EXPERIMENTAL SETUP AND DATA TREATMENT
The capabilities of lead salt diode laser spectrometers have been well known for years. In the laser spectroscopy group in Berlin experiments were carried out using a self-built dual channel diode laser spectrometer whose basic principles are described in detail by Sumpf et al. (15, 16). A resolution of 5 3 1024 cm21 (estimated from measurements for the applied diode laser) and in best cases a signal-to-noise ratio of about 1000:1 (typically 200:1) was achieved. Recently, the features of the system were illustrated by line broadening and intensity investigations for SO2 (17, 18, 19), H2S (20, 21, 22), NO2 (23), and C6H6 (24). To extend the properties of the system a third channel was implemented to record reference spectra simultaneously. Thus, it is now possible to determine line shifts directly by comparing line positions in reference and sample channels. The reference spectrum is measured at a pressure where the Doppler lineshape dominates, whereas in the sample channel the pressure is changed. Recently, this system was used for the determination of line-shift coefficients in the n3 band of NO2 for collisions with different noble gas perturbers (25). The diode laser is operated in pulsed mode. After mode separation in a grating monochromator the beam is divided into three channels by wedged BaF2 beam splitters. In the experiment a 30-cm-long linear cell was used as reference cell and 3or 30-cm cells (depending on the strength of the lines under study) are used as sample cells. The infrared light was detected with MCT photodiodes, amplified I, and digitized within two storage oscilloscopes. The two oscilloscopes were synchronized by implementing a self-built digital trigger selector. Details are described in Ref. (25). To improve the signal to noise ratio, 30 spectra under identical conditions were measured, transferred via a GPIB interface into the computer, and averaged. Simultaneously the pressures in both cells were measured with a Leybold capacitron CM3 with 10- and 100Torr heads (uncertainty of 0.5%). The data were transferred into the computer via a RS 232 interface before the first and after the last laser shot. Isotopically enriched carbon monoxide samples were taken from a lecture bottle purchased from Cambridge Isotopes Lab-
FIG. 1. Data analysis for the P(10) line at 2057.857 cm21 of 13C16O. Top trace: sample cell, l 5 5 cm, pCO 5 0.64 Torr, ptotal 5 64.30 Torr. Below: reference cell, l 5 30 cm, pCO 5 0.63 Torr. Below: residual sample cell. Bottom: residual sample cell.
oratories (USA) with a stated carbon monoxide purity of 99%. The mixture contains 89.1% 13C16O, 9.9% 13C18O, 0.9% 12 16 C O, and 0.1% 12C18O. As shown in our experiment, a small portion of 13C17O was also in the bottle. The perturbers were taken from Messer Griesheim minicans with the following purities: helium, 99.999%; neon, 99.99%; argon, 99.999%; krypton, 99.99%; xenon, 99.99%; hydrogen, 99.999%; deuterium, 99.7%; oxygen, 99.998%; nitrogen, 99.999%. The experiments were carried out by filling about 1 Torr CO into the reference cell and 1 Torr CO in the sample cell. Measuring the spectra in both channels simultaneously, the position of the line was determined in the reference channel and the sample one. Typically, there were no observable differences. After this, the pressure in the sample cell was raised up to about 100 Torr in 8 steps. The relative wavenumber calibration was carried out with a confocal air spaced e´talon with a free spectral range of 10.4 3 1023 cm21. The spacing was determined by calibrating the system using well-known CO line positions by Maki and Wells (26), which were also used for absolute calibration and assignment. The spectra were analyzed applying nonlinear least-squares fit implementing evolution strategies, as described by Heiner et al. (27). As approximation for the Voigt profile an algorithm given by Humlicek (28), which provides an optimum between the necessary spectral resolution and the time of the fit, was used. In all the experiments only small deviations between measured and predicted Voigt profiles were observed. If pronounced discrepancies occurred, more complex line profiles, which include the Dicke effect (low pressure region) and the speed dependence effect (high pressure, heavier buffer gases) would have to be taken into account (29, 30). For example, in
Copyright © 1998 by Academic Press
228
SUMPF ET AL.
experimental data points and the fitted profile in the sample (upper trace) and reference cell (bottom trace), together with the residuals, are shown for the P(10) line of 13C16O at 2057.857 cm21. In this plot a small negative shift is obvious. The line position in the sample cell (pCO 5 0.64 Torr, pN2 5 63.66 Torr) differs by 2(0.25 6 0.04) z 1023 cm21 from the line position in the reference cell (total CO pressure 0.63 Torr). The magnified residuals in the reference channel show larger deviations which may be caused by the restricted number of points in the line center and saturation problems for this strong line. Individual fits were carried out for each pressure. After this, the determined Lorentzian half-width Dn˜ L and the line shift Dn˜ were plotted versus the perturber pressure. The excellent linearity is obvious in Fig. 2. Figure 2A illustrates the determination of the nitrogen-broadening coefficient for the P(10) line, whereas Fig. 2B shows the similar plot for the line shift coefficient. The uncertainty for the Lorentzian line width is below the size of the symbols. The line shift was determined with an uncertainty of about 5 z 1025 cm21, which is about a factor of 10 smaller than the distance between our data points. FIG. 2. Plot of the line width and line shift versus the nitrogen pressure for the P(10)-line at 2057.857 cm21 of 13C16O. (A) Line width versus PN2. (B) Line shift versus PN2.
the case of collisions between CO and N2 Voigt et al. (14) estimated that the deviations between Voigt and Galatry profiles would be less than 2% in the line center. In the pressure range under study with the resolution and signal-to-noise ratio we have, the Voigt profile seems to be a good approximation for carbon monoxide lines. A typical result of the fit is given in Fig. 1. Here the
RESULTS
The determined experimental data are compiled in Table 1. For the lines P(3) of 13C17O, R(3) of 13C18O, P(9) of 12C18O, P(10) of 13C16O, and P(21) of 12C16O, the self- and nitrogenbroadening coefficients and the shift coefficients for collisions with nitrogen, oxygen, hydrogen, deuterium, helium, neon, argon, krypton, and xenon are given. Concerning the selfbroadening coefficients, one has to distinguish between two different cases. In Table 1, 12CO is marked as perturber if the perturber was taken from a minican with CO in natural abun-
TABLE 1 Line-shift and Line-broadening Coefficients for Different Isotopomers of Carbon Monoxide and Collisions with Different Perturbers
Copyright © 1998 by Academic Press
LINE SHIFT FOR DIFFERENT ISOTOPOMERS OF CO
FIG. 3. Self-broadening coefficients versus the rotational quantum number umu (P-branch lines, m 5 2J0, R-branch lines, m 5 J0 1 1). ■, this work; h, Hunt et al. (3); 3, Draegert et al. (2); E, Bouanich et al. (4).
dance, and 13CO is quoted if the perturber was taken from the 13 CO-enriched gas mixture. The deviations for the two ‘‘different’’ perturbers are obvious. In all cases the given uncertainties do not overlap. In general the broadening coefficient of a line is larger if the carbon isotope of absorber and perturber is the same. A simple explanation of this behavior could be given assuming resonant collisions between the molecules as major process for the line broadening. Here the rotational constants vary by about 5% if the carbon isotope changes. Therefore the probability for resonant collisions decreases. The given uncertainties for the broadening coefficients originate from confidence estimations calculating the weighted slope of the plots of the Lorentzian half-width and the line shift versus the pressure of the perturber. In Fig. 3 the self-broadening coefficients (i.e., 12CO with
229
FIG. 4. Nitrogen-broadening coefficients versus the rotational quantum number umu. ■, this work; E, Voigt et al. (14); 1, Sinclair et al. (11); p, Hartmann et al. (7); h, Bouanich et al. (4); {, Hamdouni et al. (9); ‚, Le Moal et al. (5); F, Hartmann et al. (6). 12
CO or 13CO with 13CO) versus the rotational quantum number m (m 5 2J0 in the P-branch, m 5 J0 1 1 in the R-branch) are plotted. We compare our data with the data by Hunt et al. (3), Draegert and Williams (2), and Bouanich and Brodbeck (4) for the main isotopic species. Within the uncertainty of our data a good agreement can be stated. No significant deviations are observed, the value for the 12 16 C O-P(21) with the largest uncertainty shows the largest deviation. Concerning the nitrogen broadening (Fig. 4) the situation is similar. Here we include in our comparison the data by Bouanich et al. (4), LeMoal et al. (5), Hartmann et al. (6, 7), Hamdouni et al. (9), Sinclair et al. (11), and Voigt et al. (14). Within our uncertainties no deviations between the different isotopomers are significant. The majority of
FIG. 5. Line-shift coefficients for diatomic perturbers (A, nitrogen; B, oxygen). ■, this work; h, Bouanich et al. (12); 3, Bouanich et al. (13); ‚, Hamdouni et al. (9); 1, Sinclair et al. (11).
Copyright © 1998 by Academic Press
230
SUMPF ET AL.
FIG. 6. Line-shift coefficients for noble gas perturbers (A, xenon; B, krypton; C, argon; D, neon; E, helium). ■, this work; h, Bouanich et al. (12); 3, Bouanich et al. (13); E, Manucci (8).
lines under study shows a nitrogen-broadening coefficient near those given for lines from 12C16O (5, 6, 9, 11). The values given by Voigt et al. (14) are slightly smaller compared to all our other data. Based on the data determined with our diode laser spectrometer, no dependence of the broadening coefficient from the isotopomer could be stated. Nevertheless, for a more conclusive discussion of the isotopic dependence, more experiments using the Berlin diode laser spectrometer are planned for the near future, as well as experiments concerning line broadening with different perturbers. In Fig. 5 the line-shift coefficients as function of m are plotted for collisions with the diatomic perturbers nitrogen (Fig. 5A) and oxygen (Fig. 5B). The determined nitrogen shift coefficients agree with the previously measured data for the (1 4 0) band for 12C16O by Bouanich et al. (12), and the absolute values are slightly smaller compared to the results given by Hamdouni et al. (9) and Sinclair et al. (11). In all cases only a weak quantum number dependence was observed.
Typically, the shift values for R-branch lines are smaller compared to P-branch lines with the same m value. The R(3)-line for 13C18O measured in this study confirms this tendency compared to the P(3)-line for 13C17O. The magnitude of the shift data for the (1 4 0) transition is as expected, about twice as small as compared to (2 4 0) transitions as studied by Bouanich et al. (13). The measured oxygen shift coefficients in this work are close to the data given by Bouanich et al. (12). Also the vibrational dependence is as expected. All experimental data for line-shift coefficients in the case of collisions with noble gases are compiled in Fig. 6. Arranged according to the size of the effect (Fig. 6A, the xenon shift data; Fig. 6B, the krypton values; Fig. 6C, the argon shift coefficients; Fig. 6D, the neon shifts; Fig. 6E, the results for collisions with helium), the data are summarized. Again the xenon-shift coefficients agree well with the data published by Bouanich et al. (12). Based on our data no dependence of the shift coefficients on the isotopomer is observable. The measured R-branch line shows a significantly smaller size of the
Copyright © 1998 by Academic Press
LINE SHIFT FOR DIFFERENT ISOTOPOMERS OF CO
shift effect. The shift coefficient for collisions with krypton in the fundamental band of CO was measured for the first time. Therefore no comparison with previously published data is possible. The vibrational dependence is well reproduced, relative to the data by Bouanich et al. (13). For argon and neon the data determined in this work confirm the (1 4 0) band data by Bouanich et al. (12). Within the uncertainties the agreement is excellent. In the case of collisions with helium, data by Manucci (8) could be included into the discussion. As measured by Manucci for the P(9)-line of 12C16O, a positive shift is observable also for the P(9)-line of 12C18O in this work. The P(10)-line for 13C16O is close to the negative shift data given by Bouanich et al. (12). Summarizing the comparison of the data for rare isotopomers determined in this work with previously measured values for the major isotopomer, a dependence of the shift coefficient on the isotopomer is not observable. A more complete coverage of the different isotopomers and the full quantum number range requires more experiments. Unfortunately, our experiments were restricted to a very small wavenumber region due to the mode properties of the available diode laser. CONCLUSION
Line-shift and line-broadening coefficients for five lines from different isotopomers of carbon monoxide were determined experimentally. Within the uncertainties the self- and nitrogenbroadening coefficients agree with previously published data for the major isotopic species 12C16O. In the case of self-broadening, differences occur perturbing a 12CO-line with 12CO and 13CO or a 13CO-line with 12CO and 13CO, respectively. For the first time, line-shift data for several rare isotopic species of CO were determined applying N2, O2, H2, D2, He, Ne, Ar, Kr, and Xe as perturbers. The line-shift studies for krypton and deuterium represent the first experimental data for the (1 4 0) fundamental band of CO. Based on our experimental data for the rare isotopomers and the previously published data by other authors for the major isotopic species 12 16 C O, no dependence of the shift coefficients from the isotopomer is observable. REFERENCES 1. M. A. K. Khalil, EOS 76(36), 353–354 (1995). 2. D. A. Draegert and D. Williams, J. Opt. Soc. Am. 58, 1399 –1403 (1968). 3. R. H. Hunt, R. A. Toth, and E. Plyler, J. Chem. Phys. 49, 3909 –3912 (1968).
231
4. J.-P. Bouanich and C. Brodbeck, J. Quant. Spectrosc. Radiat. Transfer 13, 1–7 (1973). 5. M. F. Le Moal and F. Severin, J. Quant. Spectrosc. Radiat. Transfer 35, 145–152 (1986). 6. J. M. Hartmann, M. Y. Perrin, J. Taine, and L. Rosenmann, J. Quant. Spectrosc. Radiat. Transfer 35, 357–363 (1986). 7. J. M. Hartmann, M. Y. Perrin, J. Taine, and L. Rosenmann, Appl. Opt. 27, 3063–3065 (1988). 8. A. J. Mannucci, J. Chem. Phys. 95, 7795–7805 (1991). 9. A. Hamdouni, A. Barbe, J.-J. Plataux, V. Langlois, V. Dana, J.-Y. Mandin, and M. Badaoui, J. Quant. Spectrosc. Radiat. Transfer 50, 247–255 (1993). 10. M. M. Beaky, T. M. Goyette, and F. C. DeLucia, J. Chem. Phys. 105, 3994 – 4004 (1996). 11. P. M. Sinclair, P. Duggan, R. Berman, A. D. May, and J. R. Drummond, J. Mol. Spectrosc. 181, 41– 47 (1997). 12. J.-P. Bouanich, F. Rachet, and A. Valentin, J. Mol. Spectrosc. 178, 157–164 (1996). 13. J.-P. Bouanich, D. Bermejo, J. L. Domenech, and R. Z. Martinez, J. Mol. Spectrosc. 179, 22–31 (1996). 14. S. Voigt, S. Dreher, J. Orphal, and J. P. Burrows, J. Mol. Spectrosc. 180, 359 –365 (1996). 15. B. Sumpf, D. Go¨ring, R. Haseloff, K. Herrmann, and J. W. Tomm, Collect. Czech. Chem. Commun. 54, 284 –296 (1989). 16. B. Sumpf, Y. Heiner, K. Herrmann, F. Ku¨hnemann, V. Pustogov, and J. Waschull, ‘‘Pressure Broadening of Gases with and without Line Overlapping studied with Tunable Diode Laser Spectroscopy,’’ Proceedings of the International Conference on LASERS ’93, Lake Tahoe, Nevada, pp. 297–304, STS Press, McLean, VA, 1994. 17. B. Sumpf, O. Fleischmann, and H.-D. Kronfeldt, J. Mol. Spectrosc. 176, 127–132 (1996). 18. B. Sumpf, M. Scho¨ne, and H.-D. Kronfeldt, J. Mol. Spectrosc. 179, 137–141 (1996). 19. B. Sumpf, J. Mol. Spectrosc. 181, 160 –167 (1997). 20. J. Waschull, F. Ku¨hnemann, and B. Sumpf, J. Mol. Spectrosc. 165, 150 –158 (1994). 21. B. Sumpf, I. Meusel, and H.-D. Kronfeldt, J. Mol. Spectrosc. 177, 143– 145 (1996). 22. B. Sumpf, I. Meusel, and H.-D. Kronfeldt, J. Mol. Spectrosc. 183, (1997). 23. B. Sumpf, V. V. Pustogov, and H.-D. Kronfeldt, J. Mol. Spectrosc. 180, 150 –153 (1996). 24. J. Waschull, Y. Heiner, B. Sumpf, and H.-D. Kronfeldt, J. Mol. Spectrosc. 184, 434 – 441 (1997). 25. B. Sumpf, S. Bouazza, A. Kissel, and H.-D. Kronfeldt, J. Quant. Spectrosc. Radiat. Transfer 59, 99 –108 (1998). 26. A. G. Maki and J. S. Wells, ‘‘Wavenumber Calibration Tables from Heterodyne Frequency Measurements,’’ NIST Special Publication 821, 1991. 27. Y. Heiner, O. Stier, V. Tu¨rck, J. Waschull, B. Sumpf, and A. Ostermeier, J. Quant. Spectrosc. Radiat. Transfer 56, 769 –782 (1996). 28. J. Humlicek, J. Quant. Spectrosc. Radiat. Transfer 27, 437– 444 (1982). 29. P. R. Berman, J. Quant. Spectrosc. Radiat. Transfer 12, 1331–1342 (1972). 30. H. M. Pickett, J. Chem. Phys. 73, 6090 – 6094 (1980).
Copyright © 1998 by Academic Press
190, 226 –231 (1998)
MS987595
Line Shift Investigations for Different Isotopomers of Carbon Monoxide B. Sumpf,*,1 J. P. Burrows,† A. Kissel,* H.-D. Kronfeldt,* O. Kurtz,* I. Meusel,* J. Orphal,† and S. Voigt† *Optisches Institut der Technischen Universita¨t Berlin, Sekretariat PN 0 –1, Hardenbergstr. 36, 10623 Berlin, Germany; and †Institut fu¨r Umweltphysik, Universita¨t Bremen, P.O. Box 330440, 28334 Bremen, Germany Received September 18, 1997; in revised form April 17, 1998
Line shift coefficients for five lines of five different isotopomers in the fundamental band of CO in the spectral region near 2058 cm21 were measured using a three channel lead salt diode laser spectrometer. The study includes the lines P(3) of 13 17 C O, R(3) of 13C18O, P(9) of 12C18O, P(10) of 13C16O, and P(21) of 12C16O, and covers collisions with N2, O2, H2, D2, He, Ne, Ar, Kr, and Xe. Line shifts of the isotopomers 13C16O, 12C18O, 13C18O, and 13C17O were determined for the first time. Within the experimental uncertainty no significant dependence of the shift effect on the isotopomer was found. The R-branch line under study shows a smaller line shift coefficient than a P-branch line with a similar rotational quantum number. With increasing mass of the noble gas perturber the absolute size of the shift coefficient increases. Moreover self- and nitrogenbroadening coefficients for the isotopomer lines were determined. Compared to previous measurements no significant deviations between different isotopomers were observed. © 1998 Academic Press INTRODUCTION
Carbon monoxide belongs to the most important atmospheric trace gases. The concentrations in different layers of the troposphere, the stratosphere, and the mesosphere vary strongly. Due to a short atmospheric lifetime of about 2 months, changes in sources and sinks in the environment are reflected in the atmospheric concentrations faster compared to other molecules (1). Prerequisite for the application of spectroscopic methods for the determination of CO concentrations is the accurate knowledge of line positions, line strengths, line broadening, and line shift coefficients. For line positions and strengths information for all atmospheric important isotopomers is available (most abundant 12C16O, natural abundance 98.7%; 13C16O, 1.1%; 12 18 C O, 0.2%; 12C17O, 0.04%; 13C18O, 0.0023%; and 13C17O, 0.0004%). However, for line broadening and shift, data are only known for the most abundant species. Draegert and Williams (2) published comprehensive line broadening investigations for the major isotopomer for collisions with CO, H2, D2, N2, He, Ne, Ar, Kr, and Xe in the fundamental band. Self-broadening coefficients in the first overtone band are known from Hunt et al. (3). Collisions with diatomic perturbers in the (1 4 0) band of 12C16O were analyzed by Bouanich and Brodbeck (4) and LeMoal and Severin (5). A tunable diode laser spectrometer was used by Hartmann et al. (6) to study the nitrogen and argon broadening in the fundamental band. Theoretical calculations of the collisional induced effects were carried out by Hartmann et al. (7). 1
Present address: ELIGHT Laser Systems GmbH, Warthestr. 21, 14513 Teltow, Germany.
Recently, Mannucci (8) extended the knowledge of the broadening by investigating collisions with helium and hydrogen, whereas Hamdouni et al. (9) studied broadening with N2 as buffer gas. In the microwave region Beaky et al. (10) carried out experiments concerning the H2 broadening of two transitions. Recently, Sinclair et al. (11) presented data on line broadening, shifting, and mixing for collisions with N2. Within the uncertainties of the experimental data a good agreement of the values for a given perturber and a given rotational quantum number can be stated. A pronounced decrease of the broadening coefficients with increasing rotational quantum number J0 and no vibrational dependence were observed. Line shift investigations for the major isotopomer 12C16O are less frequent. The fundamental band of 12C16O was studied by Mannucci (8) (collisions with He and H2), Hamdouni et al. (9) (N2), and Bouanich et al. (12) (N2, O2, He, Ne, Ar, Xe), and the first overtone band by Bouanich et al. (13) (He, Kr, N2, O2). The data show significant differences between the fundamental and overtone bands and within one band a different rotational quantum number dependence for R- and P-branch lines. Up to now no data concerning line shifts for other isotopomers have been known. Recently, the first analysis of the nitrogen broadening of 13 16 C O lines in the (2 4 0) band using a Fourier Transform Spectrometer were carried out by Voigt et al. (14). The broadening coefficients given in this work are on average 6% smaller than those given for 12C16O by other authors [Le Moal et al. (5), Hartmann et al. (6), Hamdouni et al. (9), and Sinclair et al. (11)]. Even slightly larger values are reported by Bouanich and Brodbeck (4). In the work by Voigt et al. (14), additional
226 0022-2852/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
LINE SHIFT FOR DIFFERENT ISOTOPOMERS OF CO
227
experiments were carried out to exclude systematic uncertainties. Broadening coefficients for transitions of 12C16O were determined. A good agreement with previously published data (5, 6, 9, 11) can be stated. Therefore it seems to be necessary to carry out new experiments with another experimental method to determine the source of the deviation mentioned above. The aim of this work was to determine for the first time line shift coefficients for rare carbon monoxide isotopomers for collisions with N2, O2, H2, D2, He, Ne, Ar, Kr, and Xe. Moreover, self- and nitrogen-broadening coefficients were measured for these less abundant isotopomers. EXPERIMENTAL SETUP AND DATA TREATMENT
The capabilities of lead salt diode laser spectrometers have been well known for years. In the laser spectroscopy group in Berlin experiments were carried out using a self-built dual channel diode laser spectrometer whose basic principles are described in detail by Sumpf et al. (15, 16). A resolution of 5 3 1024 cm21 (estimated from measurements for the applied diode laser) and in best cases a signal-to-noise ratio of about 1000:1 (typically 200:1) was achieved. Recently, the features of the system were illustrated by line broadening and intensity investigations for SO2 (17, 18, 19), H2S (20, 21, 22), NO2 (23), and C6H6 (24). To extend the properties of the system a third channel was implemented to record reference spectra simultaneously. Thus, it is now possible to determine line shifts directly by comparing line positions in reference and sample channels. The reference spectrum is measured at a pressure where the Doppler lineshape dominates, whereas in the sample channel the pressure is changed. Recently, this system was used for the determination of line-shift coefficients in the n3 band of NO2 for collisions with different noble gas perturbers (25). The diode laser is operated in pulsed mode. After mode separation in a grating monochromator the beam is divided into three channels by wedged BaF2 beam splitters. In the experiment a 30-cm-long linear cell was used as reference cell and 3or 30-cm cells (depending on the strength of the lines under study) are used as sample cells. The infrared light was detected with MCT photodiodes, amplified I, and digitized within two storage oscilloscopes. The two oscilloscopes were synchronized by implementing a self-built digital trigger selector. Details are described in Ref. (25). To improve the signal to noise ratio, 30 spectra under identical conditions were measured, transferred via a GPIB interface into the computer, and averaged. Simultaneously the pressures in both cells were measured with a Leybold capacitron CM3 with 10- and 100Torr heads (uncertainty of 0.5%). The data were transferred into the computer via a RS 232 interface before the first and after the last laser shot. Isotopically enriched carbon monoxide samples were taken from a lecture bottle purchased from Cambridge Isotopes Lab-
FIG. 1. Data analysis for the P(10) line at 2057.857 cm21 of 13C16O. Top trace: sample cell, l 5 5 cm, pCO 5 0.64 Torr, ptotal 5 64.30 Torr. Below: reference cell, l 5 30 cm, pCO 5 0.63 Torr. Below: residual sample cell. Bottom: residual sample cell.
oratories (USA) with a stated carbon monoxide purity of 99%. The mixture contains 89.1% 13C16O, 9.9% 13C18O, 0.9% 12 16 C O, and 0.1% 12C18O. As shown in our experiment, a small portion of 13C17O was also in the bottle. The perturbers were taken from Messer Griesheim minicans with the following purities: helium, 99.999%; neon, 99.99%; argon, 99.999%; krypton, 99.99%; xenon, 99.99%; hydrogen, 99.999%; deuterium, 99.7%; oxygen, 99.998%; nitrogen, 99.999%. The experiments were carried out by filling about 1 Torr CO into the reference cell and 1 Torr CO in the sample cell. Measuring the spectra in both channels simultaneously, the position of the line was determined in the reference channel and the sample one. Typically, there were no observable differences. After this, the pressure in the sample cell was raised up to about 100 Torr in 8 steps. The relative wavenumber calibration was carried out with a confocal air spaced e´talon with a free spectral range of 10.4 3 1023 cm21. The spacing was determined by calibrating the system using well-known CO line positions by Maki and Wells (26), which were also used for absolute calibration and assignment. The spectra were analyzed applying nonlinear least-squares fit implementing evolution strategies, as described by Heiner et al. (27). As approximation for the Voigt profile an algorithm given by Humlicek (28), which provides an optimum between the necessary spectral resolution and the time of the fit, was used. In all the experiments only small deviations between measured and predicted Voigt profiles were observed. If pronounced discrepancies occurred, more complex line profiles, which include the Dicke effect (low pressure region) and the speed dependence effect (high pressure, heavier buffer gases) would have to be taken into account (29, 30). For example, in
Copyright © 1998 by Academic Press
228
SUMPF ET AL.
experimental data points and the fitted profile in the sample (upper trace) and reference cell (bottom trace), together with the residuals, are shown for the P(10) line of 13C16O at 2057.857 cm21. In this plot a small negative shift is obvious. The line position in the sample cell (pCO 5 0.64 Torr, pN2 5 63.66 Torr) differs by 2(0.25 6 0.04) z 1023 cm21 from the line position in the reference cell (total CO pressure 0.63 Torr). The magnified residuals in the reference channel show larger deviations which may be caused by the restricted number of points in the line center and saturation problems for this strong line. Individual fits were carried out for each pressure. After this, the determined Lorentzian half-width Dn˜ L and the line shift Dn˜ were plotted versus the perturber pressure. The excellent linearity is obvious in Fig. 2. Figure 2A illustrates the determination of the nitrogen-broadening coefficient for the P(10) line, whereas Fig. 2B shows the similar plot for the line shift coefficient. The uncertainty for the Lorentzian line width is below the size of the symbols. The line shift was determined with an uncertainty of about 5 z 1025 cm21, which is about a factor of 10 smaller than the distance between our data points. FIG. 2. Plot of the line width and line shift versus the nitrogen pressure for the P(10)-line at 2057.857 cm21 of 13C16O. (A) Line width versus PN2. (B) Line shift versus PN2.
the case of collisions between CO and N2 Voigt et al. (14) estimated that the deviations between Voigt and Galatry profiles would be less than 2% in the line center. In the pressure range under study with the resolution and signal-to-noise ratio we have, the Voigt profile seems to be a good approximation for carbon monoxide lines. A typical result of the fit is given in Fig. 1. Here the
RESULTS
The determined experimental data are compiled in Table 1. For the lines P(3) of 13C17O, R(3) of 13C18O, P(9) of 12C18O, P(10) of 13C16O, and P(21) of 12C16O, the self- and nitrogenbroadening coefficients and the shift coefficients for collisions with nitrogen, oxygen, hydrogen, deuterium, helium, neon, argon, krypton, and xenon are given. Concerning the selfbroadening coefficients, one has to distinguish between two different cases. In Table 1, 12CO is marked as perturber if the perturber was taken from a minican with CO in natural abun-
TABLE 1 Line-shift and Line-broadening Coefficients for Different Isotopomers of Carbon Monoxide and Collisions with Different Perturbers
Copyright © 1998 by Academic Press
LINE SHIFT FOR DIFFERENT ISOTOPOMERS OF CO
FIG. 3. Self-broadening coefficients versus the rotational quantum number umu (P-branch lines, m 5 2J0, R-branch lines, m 5 J0 1 1). ■, this work; h, Hunt et al. (3); 3, Draegert et al. (2); E, Bouanich et al. (4).
dance, and 13CO is quoted if the perturber was taken from the 13 CO-enriched gas mixture. The deviations for the two ‘‘different’’ perturbers are obvious. In all cases the given uncertainties do not overlap. In general the broadening coefficient of a line is larger if the carbon isotope of absorber and perturber is the same. A simple explanation of this behavior could be given assuming resonant collisions between the molecules as major process for the line broadening. Here the rotational constants vary by about 5% if the carbon isotope changes. Therefore the probability for resonant collisions decreases. The given uncertainties for the broadening coefficients originate from confidence estimations calculating the weighted slope of the plots of the Lorentzian half-width and the line shift versus the pressure of the perturber. In Fig. 3 the self-broadening coefficients (i.e., 12CO with
229
FIG. 4. Nitrogen-broadening coefficients versus the rotational quantum number umu. ■, this work; E, Voigt et al. (14); 1, Sinclair et al. (11); p, Hartmann et al. (7); h, Bouanich et al. (4); {, Hamdouni et al. (9); ‚, Le Moal et al. (5); F, Hartmann et al. (6). 12
CO or 13CO with 13CO) versus the rotational quantum number m (m 5 2J0 in the P-branch, m 5 J0 1 1 in the R-branch) are plotted. We compare our data with the data by Hunt et al. (3), Draegert and Williams (2), and Bouanich and Brodbeck (4) for the main isotopic species. Within the uncertainty of our data a good agreement can be stated. No significant deviations are observed, the value for the 12 16 C O-P(21) with the largest uncertainty shows the largest deviation. Concerning the nitrogen broadening (Fig. 4) the situation is similar. Here we include in our comparison the data by Bouanich et al. (4), LeMoal et al. (5), Hartmann et al. (6, 7), Hamdouni et al. (9), Sinclair et al. (11), and Voigt et al. (14). Within our uncertainties no deviations between the different isotopomers are significant. The majority of
FIG. 5. Line-shift coefficients for diatomic perturbers (A, nitrogen; B, oxygen). ■, this work; h, Bouanich et al. (12); 3, Bouanich et al. (13); ‚, Hamdouni et al. (9); 1, Sinclair et al. (11).
Copyright © 1998 by Academic Press
230
SUMPF ET AL.
FIG. 6. Line-shift coefficients for noble gas perturbers (A, xenon; B, krypton; C, argon; D, neon; E, helium). ■, this work; h, Bouanich et al. (12); 3, Bouanich et al. (13); E, Manucci (8).
lines under study shows a nitrogen-broadening coefficient near those given for lines from 12C16O (5, 6, 9, 11). The values given by Voigt et al. (14) are slightly smaller compared to all our other data. Based on the data determined with our diode laser spectrometer, no dependence of the broadening coefficient from the isotopomer could be stated. Nevertheless, for a more conclusive discussion of the isotopic dependence, more experiments using the Berlin diode laser spectrometer are planned for the near future, as well as experiments concerning line broadening with different perturbers. In Fig. 5 the line-shift coefficients as function of m are plotted for collisions with the diatomic perturbers nitrogen (Fig. 5A) and oxygen (Fig. 5B). The determined nitrogen shift coefficients agree with the previously measured data for the (1 4 0) band for 12C16O by Bouanich et al. (12), and the absolute values are slightly smaller compared to the results given by Hamdouni et al. (9) and Sinclair et al. (11). In all cases only a weak quantum number dependence was observed.
Typically, the shift values for R-branch lines are smaller compared to P-branch lines with the same m value. The R(3)-line for 13C18O measured in this study confirms this tendency compared to the P(3)-line for 13C17O. The magnitude of the shift data for the (1 4 0) transition is as expected, about twice as small as compared to (2 4 0) transitions as studied by Bouanich et al. (13). The measured oxygen shift coefficients in this work are close to the data given by Bouanich et al. (12). Also the vibrational dependence is as expected. All experimental data for line-shift coefficients in the case of collisions with noble gases are compiled in Fig. 6. Arranged according to the size of the effect (Fig. 6A, the xenon shift data; Fig. 6B, the krypton values; Fig. 6C, the argon shift coefficients; Fig. 6D, the neon shifts; Fig. 6E, the results for collisions with helium), the data are summarized. Again the xenon-shift coefficients agree well with the data published by Bouanich et al. (12). Based on our data no dependence of the shift coefficients on the isotopomer is observable. The measured R-branch line shows a significantly smaller size of the
Copyright © 1998 by Academic Press
LINE SHIFT FOR DIFFERENT ISOTOPOMERS OF CO
shift effect. The shift coefficient for collisions with krypton in the fundamental band of CO was measured for the first time. Therefore no comparison with previously published data is possible. The vibrational dependence is well reproduced, relative to the data by Bouanich et al. (13). For argon and neon the data determined in this work confirm the (1 4 0) band data by Bouanich et al. (12). Within the uncertainties the agreement is excellent. In the case of collisions with helium, data by Manucci (8) could be included into the discussion. As measured by Manucci for the P(9)-line of 12C16O, a positive shift is observable also for the P(9)-line of 12C18O in this work. The P(10)-line for 13C16O is close to the negative shift data given by Bouanich et al. (12). Summarizing the comparison of the data for rare isotopomers determined in this work with previously measured values for the major isotopomer, a dependence of the shift coefficient on the isotopomer is not observable. A more complete coverage of the different isotopomers and the full quantum number range requires more experiments. Unfortunately, our experiments were restricted to a very small wavenumber region due to the mode properties of the available diode laser. CONCLUSION
Line-shift and line-broadening coefficients for five lines from different isotopomers of carbon monoxide were determined experimentally. Within the uncertainties the self- and nitrogenbroadening coefficients agree with previously published data for the major isotopic species 12C16O. In the case of self-broadening, differences occur perturbing a 12CO-line with 12CO and 13CO or a 13CO-line with 12CO and 13CO, respectively. For the first time, line-shift data for several rare isotopic species of CO were determined applying N2, O2, H2, D2, He, Ne, Ar, Kr, and Xe as perturbers. The line-shift studies for krypton and deuterium represent the first experimental data for the (1 4 0) fundamental band of CO. Based on our experimental data for the rare isotopomers and the previously published data by other authors for the major isotopic species 12 16 C O, no dependence of the shift coefficients from the isotopomer is observable. REFERENCES 1. M. A. K. Khalil, EOS 76(36), 353–354 (1995). 2. D. A. Draegert and D. Williams, J. Opt. Soc. Am. 58, 1399 –1403 (1968). 3. R. H. Hunt, R. A. Toth, and E. Plyler, J. Chem. Phys. 49, 3909 –3912 (1968).
231
4. J.-P. Bouanich and C. Brodbeck, J. Quant. Spectrosc. Radiat. Transfer 13, 1–7 (1973). 5. M. F. Le Moal and F. Severin, J. Quant. Spectrosc. Radiat. Transfer 35, 145–152 (1986). 6. J. M. Hartmann, M. Y. Perrin, J. Taine, and L. Rosenmann, J. Quant. Spectrosc. Radiat. Transfer 35, 357–363 (1986). 7. J. M. Hartmann, M. Y. Perrin, J. Taine, and L. Rosenmann, Appl. Opt. 27, 3063–3065 (1988). 8. A. J. Mannucci, J. Chem. Phys. 95, 7795–7805 (1991). 9. A. Hamdouni, A. Barbe, J.-J. Plataux, V. Langlois, V. Dana, J.-Y. Mandin, and M. Badaoui, J. Quant. Spectrosc. Radiat. Transfer 50, 247–255 (1993). 10. M. M. Beaky, T. M. Goyette, and F. C. DeLucia, J. Chem. Phys. 105, 3994 – 4004 (1996). 11. P. M. Sinclair, P. Duggan, R. Berman, A. D. May, and J. R. Drummond, J. Mol. Spectrosc. 181, 41– 47 (1997). 12. J.-P. Bouanich, F. Rachet, and A. Valentin, J. Mol. Spectrosc. 178, 157–164 (1996). 13. J.-P. Bouanich, D. Bermejo, J. L. Domenech, and R. Z. Martinez, J. Mol. Spectrosc. 179, 22–31 (1996). 14. S. Voigt, S. Dreher, J. Orphal, and J. P. Burrows, J. Mol. Spectrosc. 180, 359 –365 (1996). 15. B. Sumpf, D. Go¨ring, R. Haseloff, K. Herrmann, and J. W. Tomm, Collect. Czech. Chem. Commun. 54, 284 –296 (1989). 16. B. Sumpf, Y. Heiner, K. Herrmann, F. Ku¨hnemann, V. Pustogov, and J. Waschull, ‘‘Pressure Broadening of Gases with and without Line Overlapping studied with Tunable Diode Laser Spectroscopy,’’ Proceedings of the International Conference on LASERS ’93, Lake Tahoe, Nevada, pp. 297–304, STS Press, McLean, VA, 1994. 17. B. Sumpf, O. Fleischmann, and H.-D. Kronfeldt, J. Mol. Spectrosc. 176, 127–132 (1996). 18. B. Sumpf, M. Scho¨ne, and H.-D. Kronfeldt, J. Mol. Spectrosc. 179, 137–141 (1996). 19. B. Sumpf, J. Mol. Spectrosc. 181, 160 –167 (1997). 20. J. Waschull, F. Ku¨hnemann, and B. Sumpf, J. Mol. Spectrosc. 165, 150 –158 (1994). 21. B. Sumpf, I. Meusel, and H.-D. Kronfeldt, J. Mol. Spectrosc. 177, 143– 145 (1996). 22. B. Sumpf, I. Meusel, and H.-D. Kronfeldt, J. Mol. Spectrosc. 183, (1997). 23. B. Sumpf, V. V. Pustogov, and H.-D. Kronfeldt, J. Mol. Spectrosc. 180, 150 –153 (1996). 24. J. Waschull, Y. Heiner, B. Sumpf, and H.-D. Kronfeldt, J. Mol. Spectrosc. 184, 434 – 441 (1997). 25. B. Sumpf, S. Bouazza, A. Kissel, and H.-D. Kronfeldt, J. Quant. Spectrosc. Radiat. Transfer 59, 99 –108 (1998). 26. A. G. Maki and J. S. Wells, ‘‘Wavenumber Calibration Tables from Heterodyne Frequency Measurements,’’ NIST Special Publication 821, 1991. 27. Y. Heiner, O. Stier, V. Tu¨rck, J. Waschull, B. Sumpf, and A. Ostermeier, J. Quant. Spectrosc. Radiat. Transfer 56, 769 –782 (1996). 28. J. Humlicek, J. Quant. Spectrosc. Radiat. Transfer 27, 437– 444 (1982). 29. P. R. Berman, J. Quant. Spectrosc. Radiat. Transfer 12, 1331–1342 (1972). 30. H. M. Pickett, J. Chem. Phys. 73, 6090 – 6094 (1980).
Copyright © 1998 by Academic Press