THE JOURNAL OF CHEMICAL PHYSICS 122, 074307 2005

A 3000 K laboratory emission spectrum of water

Pierre-Franois Coheur
Service de Chimie Quantique et Photophysique, Universit Libre de Bruxelles, 50 Av. F.D. Roosevelt, B-1050 Bruxelles, Belgium

Peter F. Bernath
Department of Chemistry, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada and Department of Chemistry, University of Arizona, Tucson, Arizona 85721

Michel Carleer and Reginald Colin

Service de Chimie Quantique et Photophysique, Universit Libre de Bruxelles, 50 Av. F.D. Roosevelt, B-1050 Bruxelles, Belgium

Oleg L. Polyanskya
Sektion Spektran und Strukturdokumentation, University of Ulm, Helmholtzstraasse 22, D-89069 Ulm, Germany

Nikolai F. Zobov and Sergei V. Shirin

Institute of Applied Physics, Russian Academy of Science, Uljanov Street 46, Nizhnii Novgorod, 603950 Russia and Department of Physics and Astronomy, University College London, London WC1E 6BT, United Kingdom

Robert J. Barber and Jonathan Tennysonb

Department of Physics and Astronomy, University College London, London WC1E 6BT, United Kingdom

Received 15 October 2004; accepted 19 November 2004; published online 7 February 2005 An emission spectrum of hot water with a temperature of about 3000 K is obtained using an oxy-acetylene torch. This spectrum contains a very large number of transitions. The spectrum, along with previous cooler laboratory emission spectra and an absorption spectrum recorded from a sunspot, is analyzed in the 500 2000 cm1 region. Use of a calculated variational linelist for water allows signicant progress to be made on assigning transitions involving highly excited vibrational and rotational states. In particular emission from rotationally excited states up to J = 42 and vibrational levels with up to eight quanta of bending motion are assigned. 2005 American Institute of Physics. DOI: 10.1063/1.1847571
I. INTRODUCTION

Vibration-rotation bands of hot water vapor are prominent in the spectra of ames and cool stars.1 As early as the 1890s these steam bands were recorded from the infrared emission obtained during the combustion of hydrocarbons.2 They can also be seen in rocket plumes and in jet engine exhausts.1 Absorption of hot water vapor appears in the near infrared spectra of M-type stars3 and brown dwarfs.4 In the laboratory, the rst modern data for hot water vapor were recorded by workers5,6 at Meudon Observatory near Paris using an oxygen-hydrogen torch with a temperature of about 2900 K. The water emission in the 2800 9000 cm1 spectral region was recorded with a high resolution Fourier transform spectrometer. The rotational analysis was carried by the traditional methods of pattern recognition and combination differences. This pioneering effort provided some assignments up to J = 35, which lies at 11 656 cm1 in the ground vibrational level. The main difculty in assigning the hot water spectrum, apart from dealing with an irregular jumble of lines from a
a

Permanent address: Institute of Applied Physics, Russian Academy of Science, Uljanov Street 46, Nizhnii Novgorod 603950, Russia. b Electronic mail: [email protected] 0021-9606/2005/122 7 /074307/8/$22.50

light asymmetric top, is the presence of anomalous centrifugal distortion. As the water molecule rotates, it experiences a substantial geometrical distortion, particularly of the bending angle, from centrifugal forces. The usual Watson rotational Hamiltonian diverges and an excessive number of centrifugal distortion terms need to be retained to t the experimental data.7 This divergence of the Hamiltonian limits the utility of predictions of line positions with higher J and, in particular, higher Ka values than those already included in the t. It is this anomalous centrifugal distortion that limited the range of the rotational assignments in the original torch spectrumnot the signal-to-noise ratio of the spectra. Various schemes711 have been devised to reformulate the rotational Hamiltonian to improve convergence. These efforts have enjoyed some success, but the most satisfactory approach to assigning the hot water spectrum lies in abandoning perturbation theory altogether. The observation of the very dense water spectrum near 10 m in a sunspot12 Water on the Sun prompted the application of a variational approach13 based on solving the full vibration-rotation Schrdinger equation to predict the energy levels with high quality ab initio potential surfaces.14,15 The variational ap 2005 American Institute of Physics

122, 074307-1

Downloaded 12 Aug 2005 to 129.97.47.16. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

074307-2

Coheur et al.

J. Chem. Phys. 122, 074307 2005

proach has yielded the majority of the new assignments and energy levels as tabulated by Tennyson et al.16 Highly excited levels of water can also be obtained through the analysis of overtone spectra. In contrast to the spectra obtained with furnaces and torches, the overtone data are rotationally cold. Overtone spectra have allowed the assignment of vibrational levels with up to eight quanta of OH stretching at 25 120 cm1 to be observed.17 Highly excited bending levels, which are expected to show interesting effects,18,19 are not detected by overtone spectroscopy. The highest pure bending level that is reliably assigned16 is 060 , although tentative assignments of some levels up to 0 10 0 has been made on the basis of perturbations.20 Very recently another source of highly excited energy levels of water has been developed by Coudert et al.21 Coudert et al. have tted the rotational energy levels obtained mainly from a far infrared emission spectrum of water vapor excited by a radio frequency discharge. The lines associated with the rst eight vibrational levels were tted with the theoretical approach of Coudert.11 Very few new levels were seen, but the accuracy of the levels was signicantly improved. The goal of the current research is to obtain a new spectrum of a torch over a wider spectral range and then apply the variational approach to extend the assignments of the water levels to higher J, Ka levels. In addition, the torch spectrum allows us to assign different vibrational levels, particularly new pure bending levels. The temperature of the water vapor in an oxy-acetylene or oxy-hydrogen torch can reach 3000 K, but the line width is about 0.05 0.10 cm1 because of pressure broadening at 1 atm. The pressure can be reduced, but this also reduces the signal-to-noise ratio. The torch unfortunately has extensive emission from extraneous molecules such as CO, CO2, and OH, in addition to water. Furnace sources allow the spectrum of pure water vapor to be recorded, but then the temperature is limited to a maximum of about 2000 K by the softening of the ceramic walls of the conning tube. Carbon tube furnaces can also reach 3000 K but the water gas reaction, H2O g + C s CO g + H2 g , prevents their use as a source of hot water emission. Despite its limitations, the oxy-acetylene torch at atmospheric pressure was used for all of the work reported in this paper. As shown below this approach yields a considerable amount of information.
II. EXPERIMENTAL PROCEDURE

FIG. 1. Raw emission spectrum of the oxy-acetylene ame in the 500 2500 cm1 spectral region.

resolution was 0.03 cm1 30 cm maximum optical path difference, mopd . A CaF2 window and beam splitter were used to enhance the signal-to-noise ratio above 1900 cm1, where an InSb detector was used. In that spectral region, a 2 mm aperture was chosen and the spectral resolution was set to 0.05 cm1 18 cm mopd . In both regions 512 scans were co-added, thereby producing emission spectra with very low noise, see Fig. 1. The line positions and intensities in the spectra have been determined using the WSPECTRA program.22 In order to get rid of some weak ringing, the lines were rst identied in spectra that were apodized using a NortonBeer weak function. The ts were then performed on the unapodized spectra using the sinc instrument line shape function to determine line positions and intensities, see Fig. 2. In the tting procedure a Voigt molecular line shape function was used, and both Gaussian and Lorentzian contributions were tted. The WSPECTRA program also automatically tted the baseline. It is worth pointing out that the spectra have not been corrected for the response of the optics and the detectors. The relative

Hot water vapor was produced in an oxy-acetylene torch at atmospheric pressure. Emission spectra of the ame were recorded using a Bruker IFS 120 M Fourier transform spectrometer between 500 and 13000 cm1, using a variety of combinations of lters and detectors. For the 500 2000 cm1 region investigated here, two different settings were used: KBr entrance window and beam splitter were used to record the spectra in the lower wave number region, along with a HgCdTe detector. With this setup, the entrance aperture was 4 mm in diameter and the spectral

FIG. 2. Example of a WSPECTRA t of our data. The black line and gray line are, respectively, the observed and the calculated spectrum, with the residuals below on an expanded scale.

Downloaded 12 Aug 2005 to 129.97.47.16. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

074307-3

Emission spectrum of hot water

J. Chem. Phys. 122, 074307 2005

line intensities are therefore not reliable over large spectral ranges and should be used with care. The wave number calibration was performed using the laboratory hot water vapor measurements of Tereszchuk et al.23 in the 4880 7550 cm1 region, which is not presented in this paper. The consistency of the calibration was veried using the CO line positions in the 1-0 band, as archived by the National Institute of Standard and Technology.24 Our measurements agree with the NIST data to within 5 103 cm1, which is satisfactory considering the fact that the CO pressure-induced wave number shifts were not taken into account in the comparison. The value of 5 103 cm1 is taken as a worst-case estimate of the absolute accuracy of our measurements. A statistical uncertainty originating from the quality of the spectral t also has to be taken into account for each line. This statistical uncertainty is of the order of 104 cm1 for strong and welldened lines, but reaches signicantly larger values for the numerous weak or blended lines. As is obvious from Figs. 1 and 2, the spectra are very dense over the entire wave number region investigated, showing in addition to water lines, emission features of CO, CO2, and OH. The OH lines were identied in the spectra by comparing the calibrated list of measured line positions between 5000 and 13000 cm1 to calculated values, obtained using the spectroscopic data of Colin et al.25 and Melen et al.26 Vibrational levels up to v = 8 were considered in the comparison, and the very weak satellite lines were neglected. CO lines were similarly identied by using the HITEMP database27 as a reference, with v = 8 as the highest vibrational level, and both the 12CO and 13CO isotopes were considered. Observed CO lines are due to rotational transitions in the v = 0 and 1 vibrational levels, with only ve lines in the v = 2 vibrational level. A search for CO2 lines was made using data taken from the carbon dioxide spectroscopic database CDSD system of the Institute of Atmospheric Optics of the Siberian Branch of the Russian Academy of Sciences.28 This database contains transitions from HITRAN, HITEMP, and GEISA databases and can be used to simulate spectra at different temperatures, pressures, optical path lengths, and line shape parameters. It transpires that the CO2 lines at 3000 K in the 500 2000 cm1 region are approximately three orders of magnitude weaker than the OH lines in this region. This means that contrary to our expectations, CO2 lines are barely detectable in the spectrum analyzed in this paper. Indeed the only transitions identied lie below the cutoff used for measuring lines. The strong vibrational spectrum of CO2 starts at about 2200 cm1 and will be taken into account in our future analysis of water molecule in the stretching mode region. In the region of interest in the current paper, out of the 10 100 measured lines, 363 were assigned to CO and 141 were assigned to pure rotational transitions of OH.
III. THEORETICAL ANALYSIS

Spectra of hot water vapor in the 500 2000 cm1 region considered in this paper have been recorded in the laboratory pure rotational spectrum T = 1800 K, 373 934 cm1 region,

and v2 band spectrum T = 1800 K, 933 2500 cm1 region in emission and in sunspots T 3200 K, 722 1011 cm1 region in absorption.12,29,30 It was decided to analyze the current data along side these older spectra. Figure 3 compares a portion of these spectra. The laboratory emission spectra from heated cells were recorded at a lower temperature than the present spectrum, so are less suitable for searching for highly excited transitions of water. Indeed the previous v2 band spectrum contains some 40% fewer lines than the oxy-acetylene torch spectrum. However in the 373 934 cm1 region, the older emission spectrum has about 1000 more lines than the torch spectrum. The higher pressure in the torch spectrum means that many weak lines are obscured by stronger ones. The sunspot spectrum contains more than three times the number of lines, in absorption, than the present emission spectrum over the region in which they overlap. The sunspot spectrum thus remains the richest in terms of hidden information or number of lines per wave number. The main drawback of the sunspot spectrum is its limited wave number range due to absorptions in the Earths atmosphere. Many of our assigned lines lie outside the range of the sunspot spectrum. Furthermore, the high density of the lines, up to 50 per wave number, requires very precise frequency predictions from theoretical calculations when dealing with the weaker lines. Almost all the strong and medium absorption lines were assigned in our previous work,13,30 but few of the weaker lines. These hot water spectra contain information on highly excited vibrational and rotational levels. To be explicit, we report below emission arising from levels with up to eight quanta of bending excitation and levels rotationally excited to J = 42. Our theoretical model therefore must be capable of dealing with this high degree of excitation. In analyzing the spectra we used a calculated variational linelist for hot water. The BT1 linelist was constructed explicitly to provide reliable models for the opacity of water in the atmosphere of cool stars. It therefore considered all levels of the molecule with J 50 and lying up to 30 000 cm1 above the J = 0 ground state. The linelist used the recent, spectroscopically determined potential energy surface of Shirin et al.31 and special care was taken to ensure complete convergence of these levels as this has proved to be a problem with previous hot water linelists see Ref. 30 . Full details of this linelist, including a database containing all 620 106 transition intensities will be presented elsewhere.32 The rst step in analyzing the torch spectrum, after marking the OH and CO lines, was to make trivial assignments, that is, assignments made using our hot water linelist and previously known experimental energy levels of the water molecule.16 As the observed linewidths are relatively large a signicant number about 20% of the experimental lines have double or even triple trivial assignments. In fact the number of multiply assigned lines was even higher after making our initial trivial assignments. In this initial analysis we used the theoretical linelist with an intensity cutoff about half that of the weakest experimental lines and did not compare theoretical and experimental intensities. In making our

Downloaded 12 Aug 2005 to 129.97.47.16. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

074307-4

Coheur et al.

J. Chem. Phys. 122, 074307 2005

FIG. 3. Upper panel: sunspot absorption upper trace and laboratory emission from heated cell spectra lower trace from work Ref. 30 near 840 cm1. Lower panel: torch spectrum in the same region; vertical lines denote assigned transitions.

nal trivial assignments we retained only lines which give at least 20% of the total line intensity. About 80% of the measured lines were assigned trivially. In analyzing the remaining unassigned lines in the torch spectrum we employed different methods for different unknown levels. These levels can be roughly divided into three groups: a highest J, low Ka rotational levels in the lowest vibrational levels 000 , 010 , and 020 ; b high J =20 30 , intermediate and high Ka, low vibrational levels 000 , 010 , 020 , 030 , and 040 ; and c high bending levels 050 , 060 , 070 , and 080 .

For assigning transitions involving the highest J and low Ka rotational levels we used the method of branches, which has been described elsewhere.30,33 To predict the energy of an unknown level we used not only the calculated value, but also looked at the difference between observed and calculated values for the same branch of levels with lower J values. The BT1 linelist is based on a spectroscopically determined potential energy surface for which only levels with J = 0, 2, and 5 were used in the t. The standard deviation for that t was 0.1 cm1, but for the highest J levels known previously the typical observed minus calculated obs calc

Downloaded 12 Aug 2005 to 129.97.47.16. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

074307-5

Emission spectrum of hot water

J. Chem. Phys. 122, 074307 2005 TABLE I. Wave numbers of the assigned highest J rotational levels in 000 , 010 , and 020 vibrational levels in cm1. J 36 36 36 36 36 36 36 36 36 36 36 36 36 36 36 36 37 37 37 37 37 37 37 37 37 37 38 38 38 38 38 38 38 38 39 39 39 39 39 39 39 39 40 40 40 40 40 40 41 41 41 41 42 42 33 33 33 33 Ka 0 1 1 2 2 3 3 4 20 20 21 21 22 22 23 23 0 1 1 2 2 3 3 4 24 24 0 1 1 2 2 3 3 4 0 1 1 2 2 3 3 4 0 1 1 2 2 3 0 1 1 2 0 1 0 1 1 2 Kc 36 36 35 35 34 34 33 33 17 16 16 15 15 14 14 13 37 37 36 36 35 35 34 34 14 13 38 38 37 37 36 36 35 35 39 39 38 38 37 37 36 36 40 40 39 39 38 38 41 41 40 40 42 42 33 33 32 32 Level 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 010 010 010 010 E 12 290.5959 12 290.5959 12 921.2711 12 921.2711 13 433.3356 13 433.3356 13 935.4309 13 935.4309 19 398.9036 19 398.9036 19 750.1982 19 750.1982 20 101.7967 20 101.7967 20 452.4935 20 452.4935 12 940.1441 12 940.1441 13 585.9949 13 585.9949 14 099.7715 14 099.7715 14 614.1750 14 614.1750 21 553.1329 21 553.1329 13 604.4879 13 604.4879 14 265.3170 14 265.3170 14 778.6740 14 778.6740 15 305.9609 15 305.9609 14 283.4348 14 283.4348 14 959.0419 14 959.0419 15 469.7102 15 469.7102 16 010.5323 16 010.5323 14 976.7975 14 976.7975 15 666.9690 15 666.9690 16 172.5357 16 172.5357 15 684.3814 15 684.3814 16 388.8934 16 388.8934 16 406.0561 16 406.0561 11 953.6676 11 953.6676 12 640.1170 12 640.1170

value grows to 0.8 cm1 for 000 state, 0.4 cm1 for 010 and 0.3 cm1 for 020 . Although these errors are much larger than the average spacing between measured lines, the differences increase smoothly with J. This means we can predict the positions of unknown higher J levels with an accuracy of about 0.02 cm1. This is sufcient for unambiguous assignments. Some of the lines linking levels with the highest J values were found only in the laboratory pure rotational spectra,30 as they are obscured by stronger neighboring lines in the higher linewidth torch spectra. We were unable to assign these levels previously due to the lack of sufciently accurate predictions from variational calculations. The present analysis allowed us to determine about 100 different levels and increase the maximum J value from 35 to 42 from 11656 to 16406 cm1 for the ground vibrational level, from 32 to 39 for 010 , and from 31 to 36 for 020 . The newly determined highest J, low Ka rotational levels for the 000 , 010 , and 020 vibrational levels are presented in Table I. The majority of the assigned levels belong to states with high J =20 30 , intermediate or high Ka, and low bending vibrational excitation. Within a set of levels with a given J and vibrational state the obs calc changes smoothly with changing values of Ka and Kc. This means one can predict unknown levels with an accuracy sufcient for assignments. Our previous work on hot water spectra gave us enough known levels for each set to make this method of assignment possible. Our energy levels are given in the EPAPS archive.34 Most of the determined levels in the two groups mentioned above were obtained from a single pure rotational transition, as in each case this line is the strongest for the given level. It should be noted that the corresponding vibration-rotation transitions are largely not present in the torch spectrum since such transitions involving levels with high Ka are too weak to be seen. The tting procedure employed to obtain the spectroscopically determined potential31 used experimental energy levels for bending states up to 060 . For levels belonging to the 040 , 050 , and 060 vibrational levels the residuals of the t were small and changed smoothly with quantum numbers. Assignments could therefore be made using theoretical predictions. Most levels determined were conrmed by combination differences, making them reliable. About 20 levels of the 070 bending level have been previously determined from overtone spectroscopy,20,35 along with the 634 level of the 080 vibrational level.20 As these levels were not used in tting, the residuals for them grows to 0.5 cm1 for 070 and 1 cm1 for 080 . This is due to the bending vibrational levels approaching the barrier to linearity. Furthermore, for these states the obs calc depends signicantly on the value of Ka but has almost no dependence on J for a given Ka. This allowed us to assign 17 additional levels in 070 and 9 in 080 . The assignments to 070 can be considered as reliable, especially those conrmed by combination differences. The assignments involving 080 are more tentative. The assigned rotational levels in the 060 , 070 , and 080 vibrational levels are presented in Table II. For the 070 and 080 bending levels no transitions

Downloaded 12 Aug 2005 to 129.97.47.16. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

074307-6 TABLE I. J 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 33 34 34 34 34 35 35 35 35 36 36 36 36 37 37 37 37 37 37 38 38 39 39 32 32 33 33 34 34 35 35 36 36

Coheur et al. Continued. Ka 2 3 3 4 7 15 16 17 18 20 21 21 22 22 24 24 0 1 1 2 0 1 1 2 0 1 1 2 0 1 1 2 2 3 0 1 0 1 0 1 0 1 0 1 0 1 0 1 Kc 31 31 30 29 26 18 17 16 15 13 13 12 12 11 10 9 34 34 33 33 35 35 34 34 36 36 35 35 37 37 36 36 35 35 38 38 39 39 32 32 33 33 34 34 35 35 36 36 Level 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 010 020 020 020 020 020 020 020 020 020 020 E J 13 115.9905 13 115.9905 13 639.8400 14 085.0896 15 228.6773 17 420.0613 17 778.7798 18 130.4003 18 492.5424 19 211.6078 19 581.1881 19 581.1881 19 939.7786 19 939.7786 20 645.8124 20 645.8124 12 557.4625 12 557.4625 13 264.2962 13 264.2962 13 177.6869 13 177.6869 13 904.3971 13 904.3971 13 814.3780 13 814.3780 14 560.2728 14 560.2728 14 467.5545 14 467.5545 15 231.8067 15 231.8067 15 653.8554 15 653.8554 15 137.1779 15 137.1779 15 823.2561 15 823.2561 12 913.3436 12 913.3436 13 511.4038 13 511.4038 14 128.3021 14 128.3021 14 763.9356 14 763.9356 15 418.0188 15 418.0188 0 1 1 2 3 3 3 4 4 4 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 Ka 0 1 1 2 1 2 3 2 3 4 0 1 2 2 3 4 4 5 5 1 1 2 2 3 3 4 5 0 1 3 4 4 5 6 6 7 7 1 2 3 3 4 4 5 5 6 6 8 1 2 3 3 4 6 6 Kc 0 1 0 1 2 1 0 3 2 1 5 4 4 3 2 2 1 1 0 6 5 5 4 4 3 3 2 7 6 4 4 3 2 2 1 1 0 8 7 6 5 5 4 4 3 3 2 1 8 8 7 6 5 4 3

J. Chem. Phys. 122, 074307 2005 TABLE II. Wave numbers for rotational term values in the 060 , 070 , and 080 vibrational levels in cm1. Level 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 E 8 869.9538 8 998.0996 9 004.6170 9 271.2847 9 139.0734 9 344.3799 9 628.7737 9 438.2526a 9 725.7365 10 049.2800 9 210.1607 9 377.2237 9 556.7927a 9 565.5203 9 847.2247 10 171.0870 10 171.2898 10 519.2778b 10 519.2778b 9 400.6406 9 533.5828 9 698.0404b 9 714.4540 9 991.0166 9 992.3307 10 316.3808 10 666.2485 9 487.9479 9 714.7602 10 162.1936 10 484.2790 10 486.3693a 10 837.7072 11 199.7595 11 199.7370 11 564.2145 11 564.2145 9 697.2256 10 047.3069a 10 348.1658 10 356.0418b 10 672.4910 10 678.8623b 11 032.0251 11 032.4511a 11 395.4898 11 394.9593 12 263.3487 10 143.6997 10 254.3984 10 558.6984b 10 574.3192 10 894.2476b 11 612.7597 11 613.0694b

have been observed to the 000 rotational level. However it is possible to obtain an estimate for the vibrational band origin for these two levels by looking at the systematic differences between the observed and calculated rotational term values. Table III presents estimates for these band origins and a summary of the energy levels determined in this work. A complete tabulation of these levels is given in the EPAPS archive.34 The many trivial assignments made in this work can be

Downloaded 12 Aug 2005 to 129.97.47.16. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

074307-7 TABLE II. J 10 10 10 10 10 10 10 11 11 11 11 11 11 11 12 12 12 12 12 12 12 13 13 13 13 13 13 14 14 14 15 16 16 17 1 1 2 2 3 3 3 3 3 4 4 4 4 5 5 5 5 6 6 6 6 7

Emission spectrum of hot water Continued. Ka 1 2 3 4 6 8 8 0 1 1 2 3 4 5 1 1 2 2 3 4 5 0 1 1 2 3 5 0 1 2 0 0 1 0 1 1 1 2 0 1 1 2 3 1 1 2 3 1 1 2 2 1 1 2 3 1 Kc 10 9 8 7 5 3 2 11 11 10 9 8 7 6 12 11 11 10 10 9 8 13 13 12 11 10 8 14 14 13 15 16 16 17 1 0 2 1 3 3 2 1 0 4 3 3 2 5 4 4 3 6 5 5 4 7 Level 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 060 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 E 10 069.0056 10 483.2868 10 785.3000 11 143.4201a 11 853.8368 12 720.4711 12 720.4944 10 259.4551 10 283.9173a 10 656.5089 10 826.3404b 11 084.3958b 11 392.2185b 11 755.8094b 10 518.0698 10 942.1613b 11 003.6709a 11 112.0853b 11 326.8032a 11 676.3099b 12 038.3255b 10 756.2885 10 773.1396b 11 246.8263a 11 411.2944b 11 690.6398a 12 356.5256 11 036.1947b 11 043.9614 11 607.5558a 11 330.4470b 11 654.8647b 11 658.8880a 12 004.1736b 10 334.3666 10 341.0547 10 374.8919 10 703.3386a 10 224.5606b 10 435.5297 10 475.6581 10 775.3009 11 136.7986b 10 516.1570 10 582.9251 10 871.3547a 11 233.8187b 10 616.7746 10 716.5816a 10 990.9667 10 990.1665b 10 737.4070 10 875.8198 11 134.1762 11 496.9058b 10 878.1735
a

J. Chem. Phys. 122, 074307 2005 TABLE II. J 7 7 7 7 8 8 8 8 8 9 9 9 10 10 10 11 11 11 12 12 13 2 3 3 4 6 7 9 10 10 11

a b

Continued. Ka 1 2 3 3 0 1 1 2 2 1 1 2 1 2 3 0 1 4 0 2 0 2 0 1 1 2 1 1 1 2 1 Kc 6 6 5 4 8 8 7 7 6 9 8 7 10 9 8 11 10 7 12 11 13 1 3 2 4 5 6 8 10 9 10 Level 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 070 080 080 080 080 080 080 080 080 080 080 E 11 060.8235 11 300.7657b 11 660.8681 11 666.9403 10 897.3083b 11 039.2005 11 270.5126a 11 490.4559a 11 482.0729 11 221.7144 11 504.7708b 11 689.3420 11 424.9390a 11 938.9806a 12 302.1714b 11 573.2541b 12 043.1979a 13 778.6454 11 779.5362 12 479.7768b 1 2064.1500 12 148.2928b 11 390.5763b 11 813.6815b 11 851.5431b 12 585.1205b 12 417.0858a 12 880.4466a 12 769.9046a 13 416.0819b 13 443.8321b

The levels conrmed by combination differences. The levels obtained from one transition.

used to gauge the accuracy with which any energy levels are determined. For strong lines the trivial assignments agree with the line position with a typical error of 0.003 cm1. This error is thus appropriate for the high J, intermediate or high Ka, and low bending vibrational excitation levels assigned here. For weaker transitions this level of agreement drops and the high J levels and high bending levels are probably only accurate to about 0.01 cm1.
TABLE III. Summary of rotational energy levels for each vibrational level. Experimental origins Ref. 16 , number of new levels N and maximum J value. Level 000 010 020 030 040 050 060 070 080 E 0.000 1 594.746 3 151.630 4 666.790 6 134.015 7 542.437 8 869.954 10 085.9 2 11 253.8 5 N 179 149 75 14 24 58 37 17 9 J 42 39 36 22 21 20 17 12 11

Downloaded 12 Aug 2005 to 129.97.47.16. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

074307-8

Coheur et al.
1 2

J. Chem. Phys. 122, 074307 2005 P. F. Bernath, Phys. Chem. Chem. Phys. 4, 1501 2002 . A. G. Gaydon, The Spectroscopy of Flames, 2nd ed. Chapman and Hall, London, 1974 . 3 S. K. Leggett, F. Allard, C. Dahn, P. H. Hauschildt, T. H. Kerr, and J. Rayner, Astrophys. J. 535, 965 2000 . 4 J. D. Kirkpatrick, I. N. Reid, J. Liebert et al., Astrophys. J. 519, 802 1999 . 5 J.-M. Flaud, C. Camy-Peyret, and J.-P. Maillard, Mol. Phys. 32, 499 1976 . 6 C. Camy-Peyret, J.-M. Flaud, J.-P. Maillard, and G. Guelachvili, Mol. Phys. 33, 1641 1977 . 7 O. L. Polyansky, J. Mol. Spectrosc. 112, 79 1985 . 8 V. I. Starikov and S. N. Mikhailenko, J. Phys. B 33, 2141 2000 . 9 P. Chen, J. C. Pearson, H. M. Pickett, S. Matsuura, and G. A. Blake, Astrophys. J., Suppl. Ser. 128, 371 2000 . 10 V. G. Tyuterev, J. Mol. Spectrosc. 151, 97 1992 . 11 R. Lanquetin, L. H. Coudert, and C. Camy-Peyret, J. Mol. Spectrosc. 206, 83 2001 . 12 L. Wallace, P. Bernath, W. Livingston, K. Hinkle, J. Busler, B. Guo, and K.-Q. Zhang, Science 268, 1155 1995 . 13 O. L. Polyansky, N. F. Zobov, S. Viti, J. Tennyson, P. F. Bernath, and L. Wallace, Science 277, 346 1997 . 14 H. Partridge and D. W. Schwenke, J. Chem. Phys. 106, 4618 1997 . 15 O. L. Polyansky, A. G. Csaszar, S. V. Shirin, N. V. Zobov, P. Bartletta, J. Tennyson, D. W. Schwenke, and P. J. Knowles, Science 299, 539 2003 . 16 J. Tennyson, N. F. Zobov, R. Williamson, O. L. Polyansky, and P. F. Bernath, J. Phys. Chem. Ref. Data 30, 735 2001 . 17 N. F. Zobov, D. Belmiloud, O. L. Polyansky et al., J. Chem. Phys. 113, 1546 2000 . 18 J. P. Rose and M. E. Kellman, J. Chem. Phys. 105, 7348 1996 . 19 M. S. Child, T. Weston, and J. Tennyson, Mol. Phys. 96, 371 1999 . 20 A. Bykov, O. Naumenko, L. Sinitsa, B. Voronin, J.-M. Flaud, C. CamyPeyret, and R. Lanquetin, J. Mol. Spectrosc. 205, 1 2001 . 21 L. H. Coudert, O. Pirali, M. Vervloet, R. Lanquetin, and C. Camy-Peyret, J. Mol. Spectrosc. 228, 471 2004 . 22 M. Carleer, in Remote Sensing of Clouds and Atmosphere, edited by J. E. Russell, K. Schafer, and O. Lado-Bardowski SPIE, Washington, 2001 , Vol. 5, pp. 337342. 23 K. Tereszchuk, P. F. Bernath, N. F. Zobov, S. V. Shirin, O. L. Polyansky, N. I. Libeskind, J. Tennyson, and L. Wallace, Astrophys. J. 577, 496 2002 . 24 A. G. Maki and J. S. Wells, http://physics.nist.gov/PhysRefData/ wavenum/html/contents.html 25 R. Colin, P.-F. Coheur, M. Kiseleva, A. C. Vandaele, and P. F. Bernath, J. Mol. Spectrosc. 214, 225 2002 . 26 F. Melen, A. J. Sauval, N. Grevesse, C. B. Farmer, Ch. Servais, L. Delbouille, and G. Roland, J. Mol. Spectrosc. 174, 490 1995 . 27 L. S. Rothman, C. Camy-Peyret, J.-M. Flaud et al. private communication . 28 See http://spectra.iao.ru 29 N. F. Zobov, O. L. Polyansky, J. Tennyson, J. A. Lotoski, P. Colarusso, K.-Q. Zhang, and P. F. Bernath, J. Mol. Spectrosc. 193, 118 1999 . 30 O. L. Polyansky, N. F. Zobov, S. Viti, J. Tennyson, P. F. Bernath, and L. Wallace, J. Mol. Spectrosc. 186, 422 1997 . 31 S. V. Shirin, O. L. Polyansky, N. F. Zobov, P. Barletta, and J. Tennyson, J. Chem. Phys. 118, 2124 2003 . 32 R. J. Barber and J. Tennyson unpublished . 33 O. L. Polyansky, N. F. Zobov, S. Viti, J. Tennyson, P. F. Bernath, and L. Wallace, Astrophys. J. 489, L205 1997 . 34 See EPAPS Document No. E-JCPSA6-122-010506 for electronic versions of all the torch spectrum between 500 and 2000 cm1, plus assignments, as well the 373 933 cm1 laboratory spectrum and the 722 1011 cm1 sunspot spectrum with updated assignments. Newly determined energy levels for vibrational states 0v20 with v2 = 0 8 are also given. A direct link to this document may be found in the online articles HTML reference section. The document may also be reached via the EPAPS homepage http:// www.aip.org/pubservs/epaps.html or from ftp.aip.org in the directory /epaps/. See the EPAPS homepage for more information. 35 O. Naumenko and A. Campargue, J. Mol. Spectrosc. 221, 221 2003 . 36 S. V. Shirin, N. F. Zobov, O. L. Polyansky, J. Tennyson, T. Parekunnel, and P. F. Bernath, J. Chem. Phys. 120, 206 2004 .

The EPAPS archive also contains a listing of the three spectra of hot water re analyzed here: the oxy-acetylene torch spectrum obtained with T = 3000 K for the 529 2000 cm1 region, the pure rotational laboratory emission spectrum T = 1800 K, 373 934 cm1 region, and the sunspot absorption spectrum T 3200 K, 722 1011 cm1 region.12,29,30 After completing work on the torch spectrum we performed a trivial assignment of the sunspot spectrum and then checked for consistency between theoretical and experimental intensities. Incorrect trivial assignments, those for which the theoretical line was too weak for a relatively strong experimental line, were removed. Even with OH and SiO lines marked, about half of measured lines in the sunspot spectrum remain unassigned. Assignments obtained from the torch spectrum have also been added to the pure rotational laboratory spectrum.
IV. CONCLUSIONS

In this paper we report the observation of a new spectrum obtained from an oxy-acetylene torch with a temperature of about 3000 K. This spectrum is rich in emission from hot water and, in principle, contains a signicant quantity of information. More accurate and more extensive variational linelist has been used to analyze the torch spectrum with different strategies employed for transitions involving three different types of highly excited states. Altogether about 85% of the 10 100 lines measured in the torch spectrum in the 500 2000 cm1 range have been assigned. The majority of the remaining unassigned lines lie in pure rotational transition region 500 1000 cm1 . These lines belong mostly to high J rotational transitions within vibrational states with stretching excitation. Some of these lines could, in principle, be assigned now. However a more productive strategy is to analyze these transitions at the same time as the 2000 5000 cm1 region, which contains the stretching vibration-rotation transitions. This approach will give combination differences to conrm the assignments made. The bending region 1000 2000 cm1 contains further information on the higher bending states. To make progress on these highly excited bending states will require an improved t to the potential energy surface to give more reliable predictions. One method is to include iteratively the determined energy levels of the high-lying bending states to give better predictions and hence more levels. This method has recently been used to analyze successfully the emission spectrum of hot 1800 K D2O.36
ACKNOWLEDGMENTS

Financial support provided by the Fonds National de la Recherche Scientique FNRS, Belgium, FRFC Convention No. 2.4536.01 and the Actions de Recherches Concertees Communaute Francaise de Belgique is acknowledged. This work was also supported by The Royal Society, INTAS, the UK Engineering and Physical Science Research Council, the NASA astrophysics program, the Canadian Natural Sciences and Engineering Council, and the Russian Fund for Fundamental Studies.

Downloaded 12 Aug 2005 to 129.97.47.16. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp