Journal of Molecular Spectroscopy 219 (2003) 132135 www.elsevier.

com/locate/jms

Emission spectrum of hot HDO below 4000 cm1q

Andrew Janca,a Keith Tereszchuk,a P.F. Bernath,a,b N.F. Zobov,c,d S.V. Shirin,c,d Oleg L. Polyansky,c,d and Jonathan Tennysonc,*
Department of Chemistry, University of Waterloo, Waterloo, Ont., Canada N2L 3G1 b Department of Chemistry, University of Arizona, Tucson, AZ 85721, USA c Department of Physics and Astronomy, University College London, London WC1E 6BT, UK Institute of Applied Physics, Russian Academy of Science, Uljanov Street 46, Nizhnii Novgorod 603950, Russia Received 24 October 2002; in revised form 12 December 2002
a

Abstract Fourier transform emission spectra were recorded using a mixture of H2 O and D2 O at a temperature of 1500 C. The spectra were recorded in three overlapping sections and cover the wavenumber range 18003932 cm1 . This spectrum is analyzed together with a previously reported one spanning the 3802190 cm1 range [Parekunnel et al., J. Mol. Spectrosc. 101 (2001) 28]. This analysis leads to 4409 newly assigned HDO emission lines. This work particularly extends data on the (200) and (120) states of HDO for which newly determined energy levels are presented. 2003 Elsevier Science (USA). All rights reserved.
Keywords: Water vapor; Infrared spectrum; Emission; Line assignments; Hot bands

1. Introduction There have been a number of recent signicant advances in the study of the spectrum of water [1]. In particular the use of laboratory emission spectroscopy [2] has yielded a large number of new lines which, when combined with absorption spectra recorded in sunspots, has yielded a wealth of new data on hot H2 16 O [38]. Signicantly less attention has been paid to the spectrum of HDO, particularly that of hot HDO. However HDO is an important species in its own right for reasons as diverse as the study of D/H ratios in various locations in the Universe to tests of the failure of the BornOppenheimer approximation, where the symmetry breaking eect of single deuteration is expected to be strongest [9]. Recently, in a paper henceforth referred to as I [10], we reported the observation and assignment of the rst emission spectrum of hot (T 1500 C)
q

HDO. This spectrum spanned the 3802190 cm1 region. In the present work we report on a new HDO emission in the 18003932 cm1 region. Since this spectrum overlaps our previously reported one and since the assignment of new transitions leads to further new assignments, we report here a combined analysis of both spectra which therefore cover the range 3803932 cm1 . Most of this region has been the subject of a recent detailed study at room temperature by Toth [11].

2. Experimental details The hot HDO spectrum was recorded in November 1997 at the University of Waterloo with a Bruker IFS 120 HR Fourier transform spectrometer. The spectrometer was operated with a KBr beamsplitter and an InSb detector. A longwave pass lter at 4000 cm1 was used. Forty scans were co-added at a resolution of 0.01 cm1 . A CaF2 window was used on the emission port of the spectrometer. The water vapor was heated in the center of a 1.2 m long, 5 cm diameter alumina tube sealed with cooled CaF2 windows. The tube was placed inside a furnace and heated to 1500 C. A slow ow

Supplementary data for this article are available on ScienceDi-

rect. Corresponding author. E-mail addresses: [email protected] (P.F. Bernath), [email protected] (J. Tennyson).
*

0022-2852/03/$ - see front matter 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0022-2852(03)00015-8

A. Janca et al. / Journal of Molecular Spectroscopy 219 (2003) 132135

133

Fig. 1. Sample portion of the hot HDO emission spectrum. Line assignments are for HDO transitions. Starred lines are due to D2 O.

of water vapor was maintained through the cell at a pressure of 2.5 Torr. An equimolar mixture of H2 O and D2 O was used to provide the vapor. The thermal emission from the cell was focused into the emission port of the spectrometer with a parabolic mirror. The lines were measured with the WSpectra program of M. Carleer and have an estimated accuracy of 0:001 cm1 for strong unblended lines. The spectrum was very dense with H2 O, HDO, and D2 O lines present. A sample portion of the spectrum, with the line assignments discussed below, is presented in Fig. 1. Lines were found throughout the 18003932 cm1 region. Because a substantial fraction of the lines in this spectrum are actually due to H2 O, it was calibrated against a previously reported hot water scale in this region [6,7]. A similar spectrum, not analyzed here, was recorded for fairly pure D2 O which allowed us to identify lines associated with emission of D2 O.

3. Line assignments Our previous study, I, of hot HDO covered the 380 2190 cm1 range. Because both the high end of this range overlaps the present spectrum and the stretching fundamental bands straddle the spectrum, we chose to analyze both together. The 3802190 cm1 region spectrum analyzed previous was merged with the present spectrum in the 18002190 cm1 region to give a total of 23 200 lines. The le containing these lines is given in the supplementary data. The rst step in our analysis is to identify those lines belonging to HDO emission. In our experimental linelist, 3515 lines could be identied as belong to H2 O and a further 5357 lines as belonging to D2 O by comparison with published hot emission spectra of H2 O and unpublished D2 O spectra. Lines were identied as H2 O or D2 O by matching both frequency and intensity. This is

because the line density of the spectra is such that inevitably some HDO lines coincide with lines from the other isotopomers. Particularly for the case of D2 O spectrum this comparison is not always clear cut since the spectrum of D2 O actually contains HDO lines and there are also many blended lines. In our analysis we emphasized the HDO transitions; a signicant number of the lines we analyzed as HDO are actually blended with lines of H2 O or D2 O. A similar, separate analysis was conducted for the 2190 3932 cm1 region for which spectra of hot HDO have not been reported previously. This region contained a total of 8456 lines; we identied 3069 as belonging to hot D2 O emissions and 998 lines as belonging to hot H2 O emissions. The next step in our analysis was to try and make trivial assignments using known HDO rotationvibration energy levels. For this purpose the energy levels for the various HDO vibrational states of interest here with low J and Ka values reported by Toth [1114] and others [1517] were augmented by higher levels for these states and levels for (040) reported by us in I. Six thousand four hundred and thirty HDO transitions were assigned in the 3802190 cm1 region in I; a further 716 lines in the augmented spectrum in this region were given trivial assignments in this region as part of this work. One thousand eight hundred and fteen in the not previously analyzed 21903932 cm1 region were assigned trivially. To identify transitions involving previously unobserved energy levels in the system it was necessary to use variational predictions to aid the assignment process. There is no linelist available to us for hot HDO, but Partridge and Schwenke [18] have computed a room temperature linelist with states extending to energies about 8000 cm1 above the ground state. This linelist was transformed to a temperature of 1800 K using Boltzmann statistics. This list was augmented with energy levels for HDO states with higher rotational excitation provided by Schwenke. These levels could be used to help follow obvious sequences of transitions but the lack of intensities made them unsuitable for assigning completely new bands. Having eliminated trivial assignments and other isotopomers from our list of transitions, the unassigned lines were analyzed using a computer program. Candidate transitions were identied using the variational linelist and then conrmed, or discarded, by the presence or lack of the appropriate combination dierence transitions. In this way we were able to identify numerous new transitions involving known vibrational states. Our new spectra are particularly rich in transitions involving one change in the stretching quantum numbers. In the 3802190 cm1 region we have assigned a further 626 transitions involving states for which at least one energy level was not previously known. This means that in this region a total of 1342 newly assigned HDO

134

A. Janca et al. / Journal of Molecular Spectroscopy 219 (2003) 132135

Table 1 Summary of newly assigned HDO energy levels State (000) (010) (020) (100) (001) (030) (110) (011) (040) (120) (200) Origin (cm1 ) 0.00 1403.48 2782.01 2723.68 3707.47 4145.47 4099.96 5089.54 5420.04 5363.83 a 119 281 253 216 186 168 178 42 48 13 14 b 348 154 72 96 46 74 36 32 8 11 44 c 9 2 2 19 2 2 5 Total 467 435 325 312 232 242 214 74 56 24 58

a, New energy levels conrmed by combination dierences. b, New energy levels not conrmed by combination dierences. c, Energy levels which dier from previous compilations.

transitions can be added to the 6430 transitions assigned in I. In the higher 21903932 cm1 region, 1252 transitions were assigned to states with new energy levels. The le containing the experimental transitions given in the supplementary data has lines identied as belonging to H2 O or D2 O, where appropriate, and gives our line assignments for HDO transitions. In the newly analyzed region, 21903932 cm1 , over 85% of the transitions have been marked in this fashion. Only about 65% of the transitions have been so marked in the lower wavenumber region. This is because there are signicant portions of this region for which data on the emission of hot D2 O are not available. Table 1 gives a breakdown of the 2439 newly assigned energy levels for HDO. Trivial assignments were also made to other states not listed in the table, such (021), (101), (002), and (210), using known data [14,17]. The tables distinguish between levels to which more than one transition has been assigned, and which are therefore conrmed by combination dierences, and those for which only a single transition has been observed. The latter assignments must be considered less secure. We note that the presence of the many H2 O and D2 O lines in the spectrum makes it dicult to conrm assignments using combination dierences. As noted in the table our analysis suggests that for 42 levels the data given in previously published tabulations is not correct. Analysis of our new infrared emission spectrum has allowed us to determine many new energy levels for states involving stretching excitations, in particular (100), (001), (110), (011), (120), and (200). These new levels are tabulated in the electronic archive. This compilation augments that given in I which largely covered states with bending excitation. 4. Conclusions Emission spectra recorded in a mixture of H2 O and D2 O have been analyzed together to give a list 31 656 lines. Of these lines, 10 839 have been assigned to tran-

sitions of HDO; the majority of the remaining lines have been identied as belonging to emissions of H2 O or D2 O by comparison with emission spectra of the species. The new assignments have allowed us to build up a signicant dataset of HDO energy levels which are available both for further spectral analysis and to provide a rigorous test of theoretical models. Acknowledgments This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC). We thank T. Parekunnel for assistance with measurements. Some support was also provided by the NASA Laboratory Astrophysics Program, the UK Engineering and Science Research Council, The Royal Society, INTAS, and the Russian Fund for Fundamental Studies. References
[1] P.F. Bernath, Phys. Chem. Chem. Phys. 4 (2002) 15011509. [2] P.F. Bernath, Chem. Soc. Rev. 25 (1996) 111115. [3] O.L. Polyansky, N.F. Zobov, S. Viti, J. Tennyson, P.F. Bernath, L. Wallace, Science 277 (1997) 346349. [4] O.L. Polyansky, N.F. Zobov, S. Viti, J. Tennyson, P.F. Bernath, L. Wallace, Astrophys. J. 489 (1997) L205L208. [5] O.L. Polyansky, N.F. Zobov, S. Viti, J. Tennyson, P.F. Bernath, L. Wallace, J. Mol. Spectrosc. 186 (1997) 422447. [6] N.F. Zobov, O.L. Polyansky, J. Tennyson, S.V. Shirin, R. Nassar, T. Hirao, T. Imajo, P.F. Bernath, L. Wallace, Astrophys. J. 530 (2000) 994998. [7] K. Tereszchuk, P.F. Bernath, N.F. Zobov, S.V. Shirin, O.L. Polyansky, N. Libeskind, J. Tennyson, L. Wallace, Astrophys. J. 577 (2002) 496500. [8] J. Tennyson, N.F. Zobov, R. Williamson, O.L. Polyansky, P.F. Bernath, J. Phys. Chem. Ref. Data 30 (2001) 735831. [9] N.F. Zobov, O.L. Polyansky, C.R. Le Sueur, J. Tennyson, Chem. Phys. Lett. 260 (1996) 381387. [10] T. Parekunnel, F. Bernath, N.F. Zobov, S.V. Shirin, O.L. Polyansky, J. Tennyson, J. Mol. Spectrosc. 101 (2001) 2840. [11] R.A. Toth, J. Molec. Spectrosc. 195 (1999) 7397. [12] R.A. Toth, J. Mol. Spectrosc. 162 (1993) 2040. [13] R.A. Toth, J.W. Brault, Appl. Opt. 22 (1983) 908926.

A. Janca et al. / Journal of Molecular Spectroscopy 219 (2003) 132135 [14] R.A. Toth, J. Mol. Spectrosc. 186 (1997) 6689. [15] A. Perrin, C. Camy-Peyret, J.-M. Flaud, Can. J. Phys. 64 (1986) 736742. [16] S.-M. Hu, O.N. Ulenikov, E.S. Bekhtereva, S.-G. He, X. Wang, H. Lin, Q.-S. Zhu, J. Mol. Spectrosc. 203 (2000) 228234.

135

[17] O.N. Ulenikov, S.-M. Hu, E.S. Bekhtereva, G.A. Onopenko, X. Wang, S.-G. He, J.-J. Zheng, Q.-S. Zhu, J. Mol. Spectrosc. 208 (2001) 224235. [18] H. Partridge, D.W. Schwenke, J. Chem. Phys. 106 (1997) 46184639.