- Email: [email protected]
High-Resolution Spectroscopy of Jet-Cooled 32SO2 and 34SO2: The ã3B1–X1A1, 210 and 110 Bands
Journal of Molecular Spectroscopy 203, 151–157 (2000) doi:10.1006/jmsp.2000.8151, available online at http://www.idealibrary.com on
High-Resolution Spectroscopy of Jet-Cooled 32SO 2 and The a˜ 3B 1–X˜ 1A 1, 2 01 and 1 01 Bands
34
SO 2:
Cheng-Liang Huang,* Shan-Shan Ju,* ,1 I-Chia Chen,* ,2 Anthony J. Merer,† Chi-Kung Ni,‡ and A. H. Kung‡ *Department of Chemistry, National Tsing Hua University, Hsinchu, Taiwan 300, Republic of China; †Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, British Columbia, Canada V6T 1Z1; and ‡Institute of Atomic and Molecular Sciences, Academia Sinica, P. O. Box 23-166, Taipei, Taiwan 106, Republic of China Received April 4, 2000; in revised form May 5, 2000
Laser-induced excitation spectra of the two bands a˜ 3 B 1 –X˜ 1 A 1 , 2 01 and 1 01 of 32SO 2 and 34SO 2 have been recorded in a supersonic jet at a resolution of 0.015 cm ⫺1. The rotational and electron-spin fine structure has been analyzed for both isotopic species. Analysis of the rotational and electron-spin fine structure yields precise values of the rotational constants A, B, and C and the spin constants ␣ and  for both 32SO 2 and 34SO 2 in the states a˜ 3 B 1 (010) and (100). No interaction between these two vibrational states with any nearby triplet state is observed for rotational levels with J ⱕ 8 and K ⱕ 2. © 2000 Academic Press Key Word: sulfur dioxide. I. INTRODUCTION
The photochemistry and spectroscopy of sulfur dioxide have received much attention because of the practical importance of sulfur dioxide in atmospheric chemistry and coal combustion processes. Strong absorption of sulfur dioxide in the wavelength region 300 –330 nm is assigned to the band systems B˜ 1 B 1 and A˜ 1 A 2 –X˜ 1 A 1 (1– 6). At longer wavelengths, bands with intensity about 0.001 times as strong as those systems were assigned by Merer (7) to the transition a˜ 3 B 1 –X˜ 1 A 1 . Brand et al. (8, 9) recorded spectra of the a˜ 3 B 1 system under roomtemperature conditions and performed rotational analyses of the (000), (100), and (110) bands. They observed minor local perturbations in K⬘ ⫽ 17 and 18 for state (100) and K⬘ ⫽ 15 for (110); interaction with the nearby b˜ 3 A 2 state was proposed as responsible for the frequency shifts of these transitions. Hallin et al. (10) recorded the a˜ 3 B 1 –X˜ 1 A 1 , 0 – 0 bands for S 16O 2 and S 18O 2 at dry-ice temperature; they analyzed the electron-spin fine structure of the a˜ 3 B 1 state with a full spin and rotational Hamiltonian and gave corrected values for the off-diagonal spin constants. Later, Hallin et al. (11) recorded several vibrational transitions of S 16O 2 and S 18O 2 at higher energy and identified many local rotational perturbations that are unambiguously caused by levels of the b˜ 3 A 2 state. Brand et al. (9) suggested that the large anharmonic displacements occurring in bands at higher energy are caused by vibronic interaction with the neighboring b˜ 3 A 2 state. Zen et al. (12) recorded the laser-induced phosphorescence spectra of four isotopic variants of SO 2 in solid neon; they observed an 1
Current address: Institute of Atomic and Molecular Sciences, Academia Sinica, P. O. Box 23-166, Taipei, Taiwan 106, Republic of China. 2 Author to whom all correspondence should be addressed. (E-mail: [email protected]).
extra vibronic progression for the asymmetric species 16OS 18O which they assigned as transitions to the b˜ 3 A 2 state. Katagiri et al. (13) explored the potential energy surfaces using ab initio methods and identified three nearby low-lying triplet surfaces 1 3 B 1 , 1 3 B 2 , and 1 3 A 2 . In studies of dissociation dynamics of SO 2 excited by 193-nm photons to the C˜ 1 B 2 state, a triplet state 1 3 A 1 is suggested as providing a possible pathway that leads to dissociation (13, 14). Hence, the role of triplet states in photodissociation at large energy is important. The most precise information about the low-lying triplet states comes from analysis of high-resolution triplet–singlet absorption spectra. In this paper we report spectra of the a˜ 3 B 1 –X˜ 1 A 1 , 2 01 and 1 01 bands of SO 2 obtained at high resolution. Supersonic jet spectra were used in order to obtain extremely low rotational temperatures so as to resolve the three spin components and to avoid spectral congestion. We expected that any interaction with low-lying vibrational levels of the b˜ 3 A 2 state would be readily observable. II. EXPERIMENTAL
Sulfur dioxide (Scott 99.8%) seeded (1–5%) in argon was expanded through a pulsed valve (General Valve, diameter 0.5 mm, operating at 30 Hz) to form the supersonic jet. The stagnation pressure was maintained at about 2 atm. The vacuum chamber was pumped with a turbomolecular pump (Osaka TG2000); with the pulsed valve operating, the pressure of the chamber could be maintained below 4 ⫻ 10 ⫺5 Torr. The laser beam intersected the molecular jet ⬃2.5 cm downstream from the orifice of the nozzle. Under these conditions, the rotational temperature of SO 2 was estimated to be ⬃10 K for a mixture with 5% SO 2/Ar and ⬃5 K for a 2% mixture. Argon was used as a buffer gas to decrease the Doppler width of spectral lines in the jet expansion.
151 0022-2852/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
152
HUANG ET AL.
FIG. 1. Spectrum and assignments of a˜ 3 B 1 (010)–X˜ 1 A 1 (000) of SO 2. A line marked with an asterisk is assigned to 34SO 2.
Details of this laser system are presented elsewhere (15, 16); a brief summary is given here. A cw Ti:sapphire ring laser (Coherent 899/29) pumped with an Ar ⫹ laser (Spectra Physics, Beamlok 2060) was used to generate a narrow linewidth seeding beam near 765 nm; this cw beam was then amplified in four stages, consisting of a dye preamplifier and a Ti:Al 2O 3 amplifier pumped with a Nd:YAG laser (Spectra Physics GCR190), to generate pulses with energy of 30 –35 mJ per pulse. The linewidth of the amplified beam was estimated to be less than 250 MHz. To excite SO 2 to the state a˜ 3 B 1 (010), this infrared beam was passed through a BBO doubling crystal to generate UV pulses with an energy of about 4 mJ per pulse.
FIG. 2.
To excite to the state a˜ 3 B 1 (100), we used a frequencymixing technique. Rhodamine 6G dye was used in the singlemode ring laser to generate a green beam in the wavelength range 540 –560 nm. Rhodamine 560 dye was used in a threestage amplifier to amplify the seeding beam from the ring laser. The output of this amplifier was mixed with the output from a seeded Nd:YAG laser (Spectra Physics GCR190) in a KDP crystal (INRAD M3) to generate pulses at 363–380 nm. The energy of the UV pulses was about 4 –7 mJ/pulse and the beam had a diameter of 2 mm where it crossed the molecular jet. Total SO 2 emission was detected with an EMI 9820QB photomultiplier tube after which the signal was amplified about 10 times with a preamplifier (Philips Scientific 6954). The scattered light from both UV and infrared pulses was attenuated with an interference filter (CVI, 400 ⫾ 10 nm) and a cutoff filter (GG385). A photon counter (Stanford Research Systems 400) counted the signal; the counting interval was set to 150 s. The radiative lifetime of SO 2 is reported to be (8.8 ⫾ 2.5) ms (17); hence most of the excited molecules have moved away from the detection zone before emitting. A spatial filter consisting of two lenses collected the total emission of the SO 2. At the focal point of the second lens, a slit width of ⬃2.5 mm reduced the observed Doppler width to give lines with a full width at half-maximum (FWHM) of 0.015– 0.02 cm ⫺1. Because the emission was so weak, about 60 –100 laser shots were averaged for each data point. We calibrated the absolute wavelength of the ultraviolet laser radiation using transitions of SO 2 recorded at resolution of ⬃0.10 cm ⫺1 (18). Their absolute wavelengths were calibrated to within ⫾0.02 cm ⫺1, using optogalvanic lines of neon and I 2 fluorescence transitions. At 765 nm, the fluorescence
Spectrum and assignments of a˜ 3 B 1 (100)–X˜ 1 A 1 (000) of SO 2. A line marked with an asterisk is assigned to 34SO 2.
Copyright © 2000 by Academic Press
32
SO 2 AND
34
SO 2 IN a˜ 3 B 1 –X˜ 1 A 1 , 2 01 AND 1 01
153
TABLE 1 Observed and Calculated Line Positions (cm ⴚ1) of 32SO 2 a˜ 3B 1 (010)–X˜ 1A 1 (000)
intensity of I 2 is weak, so that e´talon fringes in combination with optogalvanic lines served for absolute calibration. When recording high-resolution spectra, we recorded fringes of an e´talon inside the ring laser simultaneously with the spectrum of SO 2. Because a phase-conjugated mirror was used in the amplification stages to eliminate amplified spontaneous emission, the wavelength of the amplified laser pulses can deviate from that of the seeding cw beam (15, 16); the extent of deviation depends on the wavelength of the laser pulses and the ambient temperature. The free-spectral range (FSR) of the e´talon was first determined to be 0.4995 cm ⫺1 using known SO 2 transitions (calibrated using the low-resolution data); a calibration curve was then set up with about 20 transitions of SO 2 to determine any possible deviations of the FSR with wavelength. Within the small frequency range and the spectral resolution of our experiment, the measured variation of the FSR of the e´talon fringes due to this conjugate mirror is negligible. Overall, the precision of the line positions is ⫾0.002 cm ⫺1; the absolute accuracy is estimated to be ⫾0.02 cm ⫺1. The uncertainty is attributed mainly to the width of the lines in the optogalvanic neon spectra.
III. RESULTS AND DISCUSSION
Excitation spectra of the a˜ 3 B 1 –X˜ 1 A 1 , 2 01 and 1 01 bands of SO 2, at a resolution of 0.015 cm ⫺1, are illustrated in Figs. 1 and 2, respectively; neither spectrum has been normalized for variations in the intensity of the excitation laser. The vibrations 1 and 2 are the symmetric stretching and the bending motions; they both transform as a 1 in the point group C 2v . Partial rotational analyses of the 2 01 and 1 01 bands at room temperature have been given by Brand et al. (9), and more complete analyses at dry-ice temperature, using the full-spin and rotational Hamiltonian (10), have been carried out by Hallin et al. (11). The very low J lines have not been assigned in either of these previous analyses because they are weak compared to the higher J lines, and blending is severe in the band centers; the low J lines are the only lines that appear in our new spectra, so that the analyses are complementary. The basic features of a˜ 3 B 1 –X˜ 1 A 1 band, where the 3 B 1 state is close to case (b) coupling, are as follows. The overall appearance is governed by the product of the upper and lower state vibronic species, which is B 1 ; the band is therefore a
Copyright © 2000 by Academic Press
154
HUANG ET AL.
TABLE 2 Observed and Calculated Line Positions (cm ⴚ1) of 32SO 2 a˜ 3B 1 (100)–X˜ 1A 1 (000)
type-C band of an asymmetric top, with the selection rule ⌬K a ⫽ ⫾1, . . . , ⌬K c ⫽ 0, . . . . The selection rule on the total angular momentum J (defined by J ⫽ N ⫹ S, where N is the rotational angular momentum) is ⌬J ⫽ 0, ⫾1. Since S ⫽ 1 for a triplet state, this means that triplet–singlet transitions must obey the rotational selection rule ⌬N ⫽ N⬘ ⫺ J⬙ ⫽ 0, ⫾1, ⫾2. Additional O-form and S-form branches are therefore present, which do not occur in a type-C singlet–singlet band. Now the three spin sublevels of a 3 B 1 vibronic state transform as the spin–vibronic representations A 1 , B 2 , and A 2 ; transitions to these three from an A 1 spin–vibronic state, such as the ground state, are only allowed for the A 1 and B 2 components, so that the relative intensities of the branches of a 3 B 1 – 1 A 1 band are governed by just two transition moments. These two transition moments correspond to the intense allowed singlet–singlet transitions from which the triplet–singlet transition gets its intensity through spin– orbit coupling. In the present case the B 2 transition moment, derived from the C˜ 1 B 2 – X˜ 1 A 1 system at 235 nm, dominates (10), which has the effect of enhancing the O- and S-form branches (19). The most abundant isotopomer, 32S 16O 2, only possesses A 1 and A 2 rovibronic levels since the Pauli principle restricts the possible nuclear-spin wavefunctions for equivalent nuclei with zero
TABLE 3 Observed Line Positions (cm ⴚ1) of 34SO 2 a˜ 3B 1 (010)–X˜ 1A 1 (000)
Copyright © 2000 by Academic Press
32
SO 2 AND
34
SO 2 IN a˜ 3 B 1 –X˜ 1 A 1 , 2 01 AND 1 01
TABLE 4 Observed Line Positions (cm ⴚ1) of 34SO 2 a˜ 3B 1 (100)–X˜ 1A 1 (000)
spin. As a result, half the rotational levels, and therefore half the lines, are missing. For example, if K a ⫽ 0, the odd J levels are missing in the X˜ 1 A 1 ground state and the even N levels are missing in the a˜ 3 B 1 excited state. The assignments shown in Figs. 1 and 2 were made using
155
ground state rotational combination differences calculated from the microwave spectrum together with the results of the high J analyses (10). The branch notation used in these figures is ⌬K ⌬J F⬘i , where the spin component label F⬘i , i ⫽ 1–3 indicates levels with J ⫽ N ⫹ 1, N, and N–1, respectively; the ⌬N label is not shown, as can be deduced at once from the ⌬J and F i labels. Because the resolution of the present spectra is higher than that of the previous work, and the rotational temperature is extremely low, very few of the lines are blended; also, in the few places where lines with a common lower level, but going to different spin components of a given (N, K) upper level can be seen, the spin components are clearly resolved. The linestrengths for these low J lines happen to favor transitions to the F 3 spin components over those to the F 1 spin components, so that not many F⬘1 lines appear. Tables 1 and 2 list the wavenumbers of the assigned lines and their residuals from the least-squares fits. Some weak lines, marked with asterisks in Figs. 1 and 2, remained unassigned after all the expected branches had been accounted for. Closer inspection shows that they form a pattern that is very similar to that of the strongest assigned lines, but displaced by a small frequency offset. Combination differences prove that they are the corresponding lines of the 34SO 2 isotopomer; 17 of these have been assigned for the 010 vibrational
TABLE 5 Rotational and Spin Constants a (cm ⴚ1) of 32SO 2 and 34SO 2
Uncertainties (3) are in units of the last significant figure and average deviations between observed and calculated transitions are 0.002– 0.003 cm ⫺1. Ground state parameters of both isotopic species are taken from Belov et al. (22) and those of 32SO 2 a˜ 3 B 1 (000) from Hallin et al. (10). c Parameter is constrained to the value for 32SO 2. a
b
Copyright © 2000 by Academic Press
156
HUANG ET AL.
FIG. 3. Plot of spin splitting vs. rotational quantum number N for K⬘ ⫽ 0, 1, and 2 of a˜ 3 B 1 (010) of 32SO 2. Diamond, circle, and triangle denote the observed value for F 1 , F 2 , and F 3 components, respectively, and dashed line denotes the value calculated with the best fit parameters.
level and 14 for the 100 level. The intensity ratio of about 25 between the lines of 32SO 2 and 34SO 2 is consistent with the known natural abundance of 34S, which is 4.2%. The line frequencies and assignments for these features are given in Tables 3 and 4. No lines of 34SO 2 were assigned in Refs. (9) and (11). We have used Sears’ program ASYTOP (20, 21) to obtain the rotational and spin constants of the upper state vibrational levels by least-squares. The ground state constants were held fixed at the microwave values (10), while for the upper states we allowed the band origin, the three rotational constants, and the two spin–spin interaction constants, ␣ and , to vary. For 32 SO 2 the three spin–rotation constants were also varied, but for 34SO 2 we had to fix the spin–rotation constants at the values for 32SO 2 since so very few lines were observed. No centrifugal distortion was included since the highest observed N value is only 9. The final values of the constants are given in Table 5,
and the line frequencies calculated from them are included in Tables 1– 4. The upper state spin and rotational constants for the 010 level listed in Table 5 are consistent with those given by Brand et al. from their room-temperature spectra (9), except for the spin–rotation parameter b which is not well determined in Ref. (9). Despite the smaller size of our low-temperature data set, our values appear to be more accurate because our resolution is a factor of 6 higher. The 010 level is not perturbed at low K, but the 100 level suffers from two significant perturbations, one vibrational and one electronic, in the K ⫽ 5 and 6 levels (11). As a result, Brand et al. made no rotational assignments for the 100 level below K ⫽ 6 and have not attempted to derive values for the spin constants. Our value of the rotational constant B (0.2965 cm ⫺1) is lower than their value (0.2979 cm ⫺1) by an amount that is well outside the error limits. Since the perturbing levels lie about 30 cm ⫺1 above the 100 level for K⬘ ⫽ 0, their effects on the levels with K⬘ ⫽ 0–2 in our data set should be quite small. The difference between the B values presumably arises because lines that were not recognized as perturbed may have been included in the data set of Ref. (9). The five spin constants for the 100 level of 32SO 2 can be obtained from our data (see Table 5). The values are similar to those of the 010 and 000 levels, indicating that there is no appreciable mixing with the nearby 3 A 2 state. The value of the spin–spin constant  is negative in all three levels, in contrast to Ref. (9). The upper state term values of the 010 level, less the rotational energy BN(N ⫹ 1), are shown plotted against N in Fig. 3. This figure shows the spin splittings in the upper state, in a similar fashion to Fig. 4 of Ref. (10). The constants for 34SO 2 listed in Table 5 are based on much smaller data sets than those for 32SO 2 and are correspondingly less well determined. Only rotational and spin constants ␣ and  were allowed to vary. The rotational constants change as expected with isotope substitution; A and C are slightly smaller in the heavier isotopomer, while the B constant, representing rotation around the axis containing the S atom, hardly changes at all. The isotopic shifts of the band origin for the levels (010) and (100) are 0.36 and ⫺0.15 cm ⫺1, respectively, in agreement with the spectra obtained for SO 2 in solid neon (12). The values 1 ⫽ 906.325 cm ⫺1 and 2 ⫽ 360.527 cm ⫺1 are obtained for 32 SO 2, whereas for 34SO 2 only the difference between 1 and 2 can be obtained; the value 545.137 cm ⫺1 is similar to the difference 1– 2 for 32SO 2, 545.798 cm ⫺1. At low-rotational temperature, the relative intensities of the branches follow the formula given by Hougen (19) for linestrengths of a C 2v prolate rotor quite closely. From comparison of the relative intensities of various branches, we confirm that the B 2 transition dipole moment (R a ), is much greater than the A 1 transition moment (R b), in agreement with Refs. (8) and (10).
Copyright © 2000 by Academic Press
32
SO 2 AND
34
SO 2 IN a˜ 3 B 1 –X˜ 1 A 1 , 2 01 AND 1 01
IV. SUMMARY
Excitation spectra of the 2 01 and 1 01 bands of the a˜ 3 B 1 –X˜ 1 A 1 transition of 32SO 2 and 34SO 2 have been recorded in a supersonic jet at a resolution of 0.015 cm ⫺1. Analysis of the rotational structure has provided precise values of the upper state rotational and spin constants. The spin splittings of these excited vibrational states display a pattern similar to that of the zero-point level. In this region, states with J ⱕ 8 and K⬘ ⱕ 2 show no evidence of interaction with the nearby b˜ 3 A 2 state, unlike the rotational and vibrational states at higher energy. Rotational lines of the less abundant isotopomer 34SO 2 have been detected and assigned for these two bands; the vibrational frequencies 1 and 2 of 32SO 2 are 906.325 and 360.527 cm ⫺1, respectively, and the interval 1– 2 in 34SO 2 is 545.137 cm ⫺1. ACKNOWLEDGMENTS C.-L.H. and I-C.C. thank National Science Council of Republic of China for financial support under Contract NSC 88-2113-M-007-033. C.K.N. and A.H.K. thank NSC and China Petroleum, Taiwan for partial support. We thank the Ministry of Education, Taiwan for support of this cooperative research between Canada and Taiwan.
REFERENCES 1. J. C. D. Brand and R. Nanes, J. Mol. Spectrosc. 46, 194 –199 (1973). 2. Y. Hamada and A. J. Merer, Can. J. Phys. 52, 1443–1457 (1974).
157
3. Y. Hamada and A. J. Merer, Can. J. Phys. 53, 2555–2576 (1975). 4. J. C. D. Brand, J. L. Hardwick, D. R. Humphrey, Y. Hamada, and A. J. Merer, Can. J. Phys. 54, 186 –196 (1976). 5. A. Fischer, R. Kullmer, and W. Demtro¨der, Chem. Phys. 83, 415– 428 (1984). 6. R. Kullmer and W. Demtro¨der, Chem. Phys. 92, 423– 433 (1985). 7. A. J. Merer, Discuss. Faraday Soc. 35, 127–136 (1963). 8. J. C. D. Brand, V. T. Jones, and C. di Lauro, J. Mol. Spectrosc. 40, 616 – 631 (1971). 9. J. C. D. Brand, V. T. Jones, and C. di Lauro, J. Mol. Spectrosc. 45, 404 – 411 (1973). 10. K. E. J. Hallin, Y. Hamada, and A. J. Merer, Can. J. Phys. 54, 2118 –2127 (1976). 11. K. E. J. Hallin, Y. Hamada, and A. J. Merer, unpublished manuscript. 12. C.-C. Zen, I-C. Chen, Y. P. Lee, and A. J. Merer, J. Phys. Chem. A 104, 771–776 (2000). 13. H. Katagiri, T. Sako, A. Hishikawa, T. Yazaki, K. Onda, K. Yamanouchi, and K. Yoshino, J. Mol. Struct. 413– 414, 589 – 614 (1997). 14. P. C. Ray, M. F. Arendt, and L. J. Butler, J. Chem. Phys. 109, 5221–5230 (1998). 15. C. K. Ni and A. H. Kung, Opt. Lett. 21, 1673–1675 (1996). 16. C. K. Ni and A. H. Kung, Appl. Opt. 37, 530 –535 (1998). 17. F. Hegazi, F. Al-Adel, A. Hamdan, and A. Dastageer, J. Phys. Chem. 98, 12169 –12175 (1994). 18. [Spectra recorded with SO 2 in a supersonic jet using UV laser light generated by mixing the output from a dye laser with 1064-nm radiation from a seeded YAG laser.] 19. J. T. Hougen, Can. J. Phys. 42, 433– 451 (1963). 20. T. J. Sears, Comput. Phys. Rep. 2, 1–32 (1984). 21. T. J. Sears, Comput. Phys. Commun. 34, 123–133 (1984). 22. S. P. Belov, M. Y. Tretyakov, I. N. Kozin, E. Klisch, G. Winnewisser, W. J. Lafferty, and J.-M. Flaud, J. Mol. Spectrosc. 191, 17–27 (1998).
Copyright © 2000 by Academic Press
High-Resolution Spectroscopy of Jet-Cooled 32SO 2 and The a˜ 3B 1–X˜ 1A 1, 2 01 and 1 01 Bands
34
SO 2:
Cheng-Liang Huang,* Shan-Shan Ju,* ,1 I-Chia Chen,* ,2 Anthony J. Merer,† Chi-Kung Ni,‡ and A. H. Kung‡ *Department of Chemistry, National Tsing Hua University, Hsinchu, Taiwan 300, Republic of China; †Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, British Columbia, Canada V6T 1Z1; and ‡Institute of Atomic and Molecular Sciences, Academia Sinica, P. O. Box 23-166, Taipei, Taiwan 106, Republic of China Received April 4, 2000; in revised form May 5, 2000
Laser-induced excitation spectra of the two bands a˜ 3 B 1 –X˜ 1 A 1 , 2 01 and 1 01 of 32SO 2 and 34SO 2 have been recorded in a supersonic jet at a resolution of 0.015 cm ⫺1. The rotational and electron-spin fine structure has been analyzed for both isotopic species. Analysis of the rotational and electron-spin fine structure yields precise values of the rotational constants A, B, and C and the spin constants ␣ and  for both 32SO 2 and 34SO 2 in the states a˜ 3 B 1 (010) and (100). No interaction between these two vibrational states with any nearby triplet state is observed for rotational levels with J ⱕ 8 and K ⱕ 2. © 2000 Academic Press Key Word: sulfur dioxide. I. INTRODUCTION
The photochemistry and spectroscopy of sulfur dioxide have received much attention because of the practical importance of sulfur dioxide in atmospheric chemistry and coal combustion processes. Strong absorption of sulfur dioxide in the wavelength region 300 –330 nm is assigned to the band systems B˜ 1 B 1 and A˜ 1 A 2 –X˜ 1 A 1 (1– 6). At longer wavelengths, bands with intensity about 0.001 times as strong as those systems were assigned by Merer (7) to the transition a˜ 3 B 1 –X˜ 1 A 1 . Brand et al. (8, 9) recorded spectra of the a˜ 3 B 1 system under roomtemperature conditions and performed rotational analyses of the (000), (100), and (110) bands. They observed minor local perturbations in K⬘ ⫽ 17 and 18 for state (100) and K⬘ ⫽ 15 for (110); interaction with the nearby b˜ 3 A 2 state was proposed as responsible for the frequency shifts of these transitions. Hallin et al. (10) recorded the a˜ 3 B 1 –X˜ 1 A 1 , 0 – 0 bands for S 16O 2 and S 18O 2 at dry-ice temperature; they analyzed the electron-spin fine structure of the a˜ 3 B 1 state with a full spin and rotational Hamiltonian and gave corrected values for the off-diagonal spin constants. Later, Hallin et al. (11) recorded several vibrational transitions of S 16O 2 and S 18O 2 at higher energy and identified many local rotational perturbations that are unambiguously caused by levels of the b˜ 3 A 2 state. Brand et al. (9) suggested that the large anharmonic displacements occurring in bands at higher energy are caused by vibronic interaction with the neighboring b˜ 3 A 2 state. Zen et al. (12) recorded the laser-induced phosphorescence spectra of four isotopic variants of SO 2 in solid neon; they observed an 1
Current address: Institute of Atomic and Molecular Sciences, Academia Sinica, P. O. Box 23-166, Taipei, Taiwan 106, Republic of China. 2 Author to whom all correspondence should be addressed. (E-mail: [email protected]).
extra vibronic progression for the asymmetric species 16OS 18O which they assigned as transitions to the b˜ 3 A 2 state. Katagiri et al. (13) explored the potential energy surfaces using ab initio methods and identified three nearby low-lying triplet surfaces 1 3 B 1 , 1 3 B 2 , and 1 3 A 2 . In studies of dissociation dynamics of SO 2 excited by 193-nm photons to the C˜ 1 B 2 state, a triplet state 1 3 A 1 is suggested as providing a possible pathway that leads to dissociation (13, 14). Hence, the role of triplet states in photodissociation at large energy is important. The most precise information about the low-lying triplet states comes from analysis of high-resolution triplet–singlet absorption spectra. In this paper we report spectra of the a˜ 3 B 1 –X˜ 1 A 1 , 2 01 and 1 01 bands of SO 2 obtained at high resolution. Supersonic jet spectra were used in order to obtain extremely low rotational temperatures so as to resolve the three spin components and to avoid spectral congestion. We expected that any interaction with low-lying vibrational levels of the b˜ 3 A 2 state would be readily observable. II. EXPERIMENTAL
Sulfur dioxide (Scott 99.8%) seeded (1–5%) in argon was expanded through a pulsed valve (General Valve, diameter 0.5 mm, operating at 30 Hz) to form the supersonic jet. The stagnation pressure was maintained at about 2 atm. The vacuum chamber was pumped with a turbomolecular pump (Osaka TG2000); with the pulsed valve operating, the pressure of the chamber could be maintained below 4 ⫻ 10 ⫺5 Torr. The laser beam intersected the molecular jet ⬃2.5 cm downstream from the orifice of the nozzle. Under these conditions, the rotational temperature of SO 2 was estimated to be ⬃10 K for a mixture with 5% SO 2/Ar and ⬃5 K for a 2% mixture. Argon was used as a buffer gas to decrease the Doppler width of spectral lines in the jet expansion.
151 0022-2852/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
152
HUANG ET AL.
FIG. 1. Spectrum and assignments of a˜ 3 B 1 (010)–X˜ 1 A 1 (000) of SO 2. A line marked with an asterisk is assigned to 34SO 2.
Details of this laser system are presented elsewhere (15, 16); a brief summary is given here. A cw Ti:sapphire ring laser (Coherent 899/29) pumped with an Ar ⫹ laser (Spectra Physics, Beamlok 2060) was used to generate a narrow linewidth seeding beam near 765 nm; this cw beam was then amplified in four stages, consisting of a dye preamplifier and a Ti:Al 2O 3 amplifier pumped with a Nd:YAG laser (Spectra Physics GCR190), to generate pulses with energy of 30 –35 mJ per pulse. The linewidth of the amplified beam was estimated to be less than 250 MHz. To excite SO 2 to the state a˜ 3 B 1 (010), this infrared beam was passed through a BBO doubling crystal to generate UV pulses with an energy of about 4 mJ per pulse.
FIG. 2.
To excite to the state a˜ 3 B 1 (100), we used a frequencymixing technique. Rhodamine 6G dye was used in the singlemode ring laser to generate a green beam in the wavelength range 540 –560 nm. Rhodamine 560 dye was used in a threestage amplifier to amplify the seeding beam from the ring laser. The output of this amplifier was mixed with the output from a seeded Nd:YAG laser (Spectra Physics GCR190) in a KDP crystal (INRAD M3) to generate pulses at 363–380 nm. The energy of the UV pulses was about 4 –7 mJ/pulse and the beam had a diameter of 2 mm where it crossed the molecular jet. Total SO 2 emission was detected with an EMI 9820QB photomultiplier tube after which the signal was amplified about 10 times with a preamplifier (Philips Scientific 6954). The scattered light from both UV and infrared pulses was attenuated with an interference filter (CVI, 400 ⫾ 10 nm) and a cutoff filter (GG385). A photon counter (Stanford Research Systems 400) counted the signal; the counting interval was set to 150 s. The radiative lifetime of SO 2 is reported to be (8.8 ⫾ 2.5) ms (17); hence most of the excited molecules have moved away from the detection zone before emitting. A spatial filter consisting of two lenses collected the total emission of the SO 2. At the focal point of the second lens, a slit width of ⬃2.5 mm reduced the observed Doppler width to give lines with a full width at half-maximum (FWHM) of 0.015– 0.02 cm ⫺1. Because the emission was so weak, about 60 –100 laser shots were averaged for each data point. We calibrated the absolute wavelength of the ultraviolet laser radiation using transitions of SO 2 recorded at resolution of ⬃0.10 cm ⫺1 (18). Their absolute wavelengths were calibrated to within ⫾0.02 cm ⫺1, using optogalvanic lines of neon and I 2 fluorescence transitions. At 765 nm, the fluorescence
Spectrum and assignments of a˜ 3 B 1 (100)–X˜ 1 A 1 (000) of SO 2. A line marked with an asterisk is assigned to 34SO 2.
Copyright © 2000 by Academic Press
32
SO 2 AND
34
SO 2 IN a˜ 3 B 1 –X˜ 1 A 1 , 2 01 AND 1 01
153
TABLE 1 Observed and Calculated Line Positions (cm ⴚ1) of 32SO 2 a˜ 3B 1 (010)–X˜ 1A 1 (000)
intensity of I 2 is weak, so that e´talon fringes in combination with optogalvanic lines served for absolute calibration. When recording high-resolution spectra, we recorded fringes of an e´talon inside the ring laser simultaneously with the spectrum of SO 2. Because a phase-conjugated mirror was used in the amplification stages to eliminate amplified spontaneous emission, the wavelength of the amplified laser pulses can deviate from that of the seeding cw beam (15, 16); the extent of deviation depends on the wavelength of the laser pulses and the ambient temperature. The free-spectral range (FSR) of the e´talon was first determined to be 0.4995 cm ⫺1 using known SO 2 transitions (calibrated using the low-resolution data); a calibration curve was then set up with about 20 transitions of SO 2 to determine any possible deviations of the FSR with wavelength. Within the small frequency range and the spectral resolution of our experiment, the measured variation of the FSR of the e´talon fringes due to this conjugate mirror is negligible. Overall, the precision of the line positions is ⫾0.002 cm ⫺1; the absolute accuracy is estimated to be ⫾0.02 cm ⫺1. The uncertainty is attributed mainly to the width of the lines in the optogalvanic neon spectra.
III. RESULTS AND DISCUSSION
Excitation spectra of the a˜ 3 B 1 –X˜ 1 A 1 , 2 01 and 1 01 bands of SO 2, at a resolution of 0.015 cm ⫺1, are illustrated in Figs. 1 and 2, respectively; neither spectrum has been normalized for variations in the intensity of the excitation laser. The vibrations 1 and 2 are the symmetric stretching and the bending motions; they both transform as a 1 in the point group C 2v . Partial rotational analyses of the 2 01 and 1 01 bands at room temperature have been given by Brand et al. (9), and more complete analyses at dry-ice temperature, using the full-spin and rotational Hamiltonian (10), have been carried out by Hallin et al. (11). The very low J lines have not been assigned in either of these previous analyses because they are weak compared to the higher J lines, and blending is severe in the band centers; the low J lines are the only lines that appear in our new spectra, so that the analyses are complementary. The basic features of a˜ 3 B 1 –X˜ 1 A 1 band, where the 3 B 1 state is close to case (b) coupling, are as follows. The overall appearance is governed by the product of the upper and lower state vibronic species, which is B 1 ; the band is therefore a
Copyright © 2000 by Academic Press
154
HUANG ET AL.
TABLE 2 Observed and Calculated Line Positions (cm ⴚ1) of 32SO 2 a˜ 3B 1 (100)–X˜ 1A 1 (000)
type-C band of an asymmetric top, with the selection rule ⌬K a ⫽ ⫾1, . . . , ⌬K c ⫽ 0, . . . . The selection rule on the total angular momentum J (defined by J ⫽ N ⫹ S, where N is the rotational angular momentum) is ⌬J ⫽ 0, ⫾1. Since S ⫽ 1 for a triplet state, this means that triplet–singlet transitions must obey the rotational selection rule ⌬N ⫽ N⬘ ⫺ J⬙ ⫽ 0, ⫾1, ⫾2. Additional O-form and S-form branches are therefore present, which do not occur in a type-C singlet–singlet band. Now the three spin sublevels of a 3 B 1 vibronic state transform as the spin–vibronic representations A 1 , B 2 , and A 2 ; transitions to these three from an A 1 spin–vibronic state, such as the ground state, are only allowed for the A 1 and B 2 components, so that the relative intensities of the branches of a 3 B 1 – 1 A 1 band are governed by just two transition moments. These two transition moments correspond to the intense allowed singlet–singlet transitions from which the triplet–singlet transition gets its intensity through spin– orbit coupling. In the present case the B 2 transition moment, derived from the C˜ 1 B 2 – X˜ 1 A 1 system at 235 nm, dominates (10), which has the effect of enhancing the O- and S-form branches (19). The most abundant isotopomer, 32S 16O 2, only possesses A 1 and A 2 rovibronic levels since the Pauli principle restricts the possible nuclear-spin wavefunctions for equivalent nuclei with zero
TABLE 3 Observed Line Positions (cm ⴚ1) of 34SO 2 a˜ 3B 1 (010)–X˜ 1A 1 (000)
Copyright © 2000 by Academic Press
32
SO 2 AND
34
SO 2 IN a˜ 3 B 1 –X˜ 1 A 1 , 2 01 AND 1 01
TABLE 4 Observed Line Positions (cm ⴚ1) of 34SO 2 a˜ 3B 1 (100)–X˜ 1A 1 (000)
spin. As a result, half the rotational levels, and therefore half the lines, are missing. For example, if K a ⫽ 0, the odd J levels are missing in the X˜ 1 A 1 ground state and the even N levels are missing in the a˜ 3 B 1 excited state. The assignments shown in Figs. 1 and 2 were made using
155
ground state rotational combination differences calculated from the microwave spectrum together with the results of the high J analyses (10). The branch notation used in these figures is ⌬K ⌬J F⬘i , where the spin component label F⬘i , i ⫽ 1–3 indicates levels with J ⫽ N ⫹ 1, N, and N–1, respectively; the ⌬N label is not shown, as can be deduced at once from the ⌬J and F i labels. Because the resolution of the present spectra is higher than that of the previous work, and the rotational temperature is extremely low, very few of the lines are blended; also, in the few places where lines with a common lower level, but going to different spin components of a given (N, K) upper level can be seen, the spin components are clearly resolved. The linestrengths for these low J lines happen to favor transitions to the F 3 spin components over those to the F 1 spin components, so that not many F⬘1 lines appear. Tables 1 and 2 list the wavenumbers of the assigned lines and their residuals from the least-squares fits. Some weak lines, marked with asterisks in Figs. 1 and 2, remained unassigned after all the expected branches had been accounted for. Closer inspection shows that they form a pattern that is very similar to that of the strongest assigned lines, but displaced by a small frequency offset. Combination differences prove that they are the corresponding lines of the 34SO 2 isotopomer; 17 of these have been assigned for the 010 vibrational
TABLE 5 Rotational and Spin Constants a (cm ⴚ1) of 32SO 2 and 34SO 2
Uncertainties (3) are in units of the last significant figure and average deviations between observed and calculated transitions are 0.002– 0.003 cm ⫺1. Ground state parameters of both isotopic species are taken from Belov et al. (22) and those of 32SO 2 a˜ 3 B 1 (000) from Hallin et al. (10). c Parameter is constrained to the value for 32SO 2. a
b
Copyright © 2000 by Academic Press
156
HUANG ET AL.
FIG. 3. Plot of spin splitting vs. rotational quantum number N for K⬘ ⫽ 0, 1, and 2 of a˜ 3 B 1 (010) of 32SO 2. Diamond, circle, and triangle denote the observed value for F 1 , F 2 , and F 3 components, respectively, and dashed line denotes the value calculated with the best fit parameters.
level and 14 for the 100 level. The intensity ratio of about 25 between the lines of 32SO 2 and 34SO 2 is consistent with the known natural abundance of 34S, which is 4.2%. The line frequencies and assignments for these features are given in Tables 3 and 4. No lines of 34SO 2 were assigned in Refs. (9) and (11). We have used Sears’ program ASYTOP (20, 21) to obtain the rotational and spin constants of the upper state vibrational levels by least-squares. The ground state constants were held fixed at the microwave values (10), while for the upper states we allowed the band origin, the three rotational constants, and the two spin–spin interaction constants, ␣ and , to vary. For 32 SO 2 the three spin–rotation constants were also varied, but for 34SO 2 we had to fix the spin–rotation constants at the values for 32SO 2 since so very few lines were observed. No centrifugal distortion was included since the highest observed N value is only 9. The final values of the constants are given in Table 5,
and the line frequencies calculated from them are included in Tables 1– 4. The upper state spin and rotational constants for the 010 level listed in Table 5 are consistent with those given by Brand et al. from their room-temperature spectra (9), except for the spin–rotation parameter b which is not well determined in Ref. (9). Despite the smaller size of our low-temperature data set, our values appear to be more accurate because our resolution is a factor of 6 higher. The 010 level is not perturbed at low K, but the 100 level suffers from two significant perturbations, one vibrational and one electronic, in the K ⫽ 5 and 6 levels (11). As a result, Brand et al. made no rotational assignments for the 100 level below K ⫽ 6 and have not attempted to derive values for the spin constants. Our value of the rotational constant B (0.2965 cm ⫺1) is lower than their value (0.2979 cm ⫺1) by an amount that is well outside the error limits. Since the perturbing levels lie about 30 cm ⫺1 above the 100 level for K⬘ ⫽ 0, their effects on the levels with K⬘ ⫽ 0–2 in our data set should be quite small. The difference between the B values presumably arises because lines that were not recognized as perturbed may have been included in the data set of Ref. (9). The five spin constants for the 100 level of 32SO 2 can be obtained from our data (see Table 5). The values are similar to those of the 010 and 000 levels, indicating that there is no appreciable mixing with the nearby 3 A 2 state. The value of the spin–spin constant  is negative in all three levels, in contrast to Ref. (9). The upper state term values of the 010 level, less the rotational energy BN(N ⫹ 1), are shown plotted against N in Fig. 3. This figure shows the spin splittings in the upper state, in a similar fashion to Fig. 4 of Ref. (10). The constants for 34SO 2 listed in Table 5 are based on much smaller data sets than those for 32SO 2 and are correspondingly less well determined. Only rotational and spin constants ␣ and  were allowed to vary. The rotational constants change as expected with isotope substitution; A and C are slightly smaller in the heavier isotopomer, while the B constant, representing rotation around the axis containing the S atom, hardly changes at all. The isotopic shifts of the band origin for the levels (010) and (100) are 0.36 and ⫺0.15 cm ⫺1, respectively, in agreement with the spectra obtained for SO 2 in solid neon (12). The values 1 ⫽ 906.325 cm ⫺1 and 2 ⫽ 360.527 cm ⫺1 are obtained for 32 SO 2, whereas for 34SO 2 only the difference between 1 and 2 can be obtained; the value 545.137 cm ⫺1 is similar to the difference 1– 2 for 32SO 2, 545.798 cm ⫺1. At low-rotational temperature, the relative intensities of the branches follow the formula given by Hougen (19) for linestrengths of a C 2v prolate rotor quite closely. From comparison of the relative intensities of various branches, we confirm that the B 2 transition dipole moment (R a ), is much greater than the A 1 transition moment (R b), in agreement with Refs. (8) and (10).
Copyright © 2000 by Academic Press
32
SO 2 AND
34
SO 2 IN a˜ 3 B 1 –X˜ 1 A 1 , 2 01 AND 1 01
IV. SUMMARY
Excitation spectra of the 2 01 and 1 01 bands of the a˜ 3 B 1 –X˜ 1 A 1 transition of 32SO 2 and 34SO 2 have been recorded in a supersonic jet at a resolution of 0.015 cm ⫺1. Analysis of the rotational structure has provided precise values of the upper state rotational and spin constants. The spin splittings of these excited vibrational states display a pattern similar to that of the zero-point level. In this region, states with J ⱕ 8 and K⬘ ⱕ 2 show no evidence of interaction with the nearby b˜ 3 A 2 state, unlike the rotational and vibrational states at higher energy. Rotational lines of the less abundant isotopomer 34SO 2 have been detected and assigned for these two bands; the vibrational frequencies 1 and 2 of 32SO 2 are 906.325 and 360.527 cm ⫺1, respectively, and the interval 1– 2 in 34SO 2 is 545.137 cm ⫺1. ACKNOWLEDGMENTS C.-L.H. and I-C.C. thank National Science Council of Republic of China for financial support under Contract NSC 88-2113-M-007-033. C.K.N. and A.H.K. thank NSC and China Petroleum, Taiwan for partial support. We thank the Ministry of Education, Taiwan for support of this cooperative research between Canada and Taiwan.
REFERENCES 1. J. C. D. Brand and R. Nanes, J. Mol. Spectrosc. 46, 194 –199 (1973). 2. Y. Hamada and A. J. Merer, Can. J. Phys. 52, 1443–1457 (1974).
157
3. Y. Hamada and A. J. Merer, Can. J. Phys. 53, 2555–2576 (1975). 4. J. C. D. Brand, J. L. Hardwick, D. R. Humphrey, Y. Hamada, and A. J. Merer, Can. J. Phys. 54, 186 –196 (1976). 5. A. Fischer, R. Kullmer, and W. Demtro¨der, Chem. Phys. 83, 415– 428 (1984). 6. R. Kullmer and W. Demtro¨der, Chem. Phys. 92, 423– 433 (1985). 7. A. J. Merer, Discuss. Faraday Soc. 35, 127–136 (1963). 8. J. C. D. Brand, V. T. Jones, and C. di Lauro, J. Mol. Spectrosc. 40, 616 – 631 (1971). 9. J. C. D. Brand, V. T. Jones, and C. di Lauro, J. Mol. Spectrosc. 45, 404 – 411 (1973). 10. K. E. J. Hallin, Y. Hamada, and A. J. Merer, Can. J. Phys. 54, 2118 –2127 (1976). 11. K. E. J. Hallin, Y. Hamada, and A. J. Merer, unpublished manuscript. 12. C.-C. Zen, I-C. Chen, Y. P. Lee, and A. J. Merer, J. Phys. Chem. A 104, 771–776 (2000). 13. H. Katagiri, T. Sako, A. Hishikawa, T. Yazaki, K. Onda, K. Yamanouchi, and K. Yoshino, J. Mol. Struct. 413– 414, 589 – 614 (1997). 14. P. C. Ray, M. F. Arendt, and L. J. Butler, J. Chem. Phys. 109, 5221–5230 (1998). 15. C. K. Ni and A. H. Kung, Opt. Lett. 21, 1673–1675 (1996). 16. C. K. Ni and A. H. Kung, Appl. Opt. 37, 530 –535 (1998). 17. F. Hegazi, F. Al-Adel, A. Hamdan, and A. Dastageer, J. Phys. Chem. 98, 12169 –12175 (1994). 18. [Spectra recorded with SO 2 in a supersonic jet using UV laser light generated by mixing the output from a dye laser with 1064-nm radiation from a seeded YAG laser.] 19. J. T. Hougen, Can. J. Phys. 42, 433– 451 (1963). 20. T. J. Sears, Comput. Phys. Rep. 2, 1–32 (1984). 21. T. J. Sears, Comput. Phys. Commun. 34, 123–133 (1984). 22. S. P. Belov, M. Y. Tretyakov, I. N. Kozin, E. Klisch, G. Winnewisser, W. J. Lafferty, and J.-M. Flaud, J. Mol. Spectrosc. 191, 17–27 (1998).
Copyright © 2000 by Academic Press