Water Spectra in the Region 4200–6250 cm−1, Extended Analysis of ν1+ν2, ν2+ν3, and 3ν2 Bands and Confirmation of Highly Excited States from Flame Spectra and from Atmospheric Long-Path Observations

Abstract Water vapor infrared spectra have been recorded at room temperature in the range 4200–6250 cm −1 at resolutions (FWHM) between 0.0053 and 0.0080 cm −1 . The use of a White-type multireflection cell made large pressure × pathlength products possible up to 31.27 mbar×288.5 m. The high signal-to-noise ratio allowed us to observe lines with intensities as small as 10 −26 cm −1 /molecule cm −2 at T =296 K. Among about 5100 recorded water lines, about half of which are reported for the first time, 2351 lines have been assigned to the second triad of H 2 16 O (bands ν 1 +ν 2 , ν 2 +ν 3 , and 3ν 2 ). This has allowed the determination of line positions and corresponding upper rovibrational states with considerably improved accuracy. The assignments of certain highly excited states have been confirmed by the analysis of flame spectra and hot emission spectra. New values of effective Hamiltonian parameters for the upper states {(110), (030), (011)} have been determined. The generating function model was used in the data reduction to account for the anomalously strong centrifugal distortion of the rovibrational levels and resonance interactions. The RMS standard deviation of the least-squares fit of the assigned H 2 O data was 5×10 −3 cm −1 for line positions and 7×10 −3 cm −1 for energy levels up to J max =20 and K a ( max ) =13. Particular attention was paid to water lines in the transparency window 4200–5000 cm −1 , in which existing databases are not sufficient. In this region, 1395 lines of four isotopic species of water have been recorded and over 900 accurate line positions of nine bands of H 2 16 O (ν 1 , ν 3 , 2ν 2 , ν 1 +ν 2 , ν 2 +ν 3 , 3ν 2 , 4ν 2 −ν 2 , 2ν 2 +ν 3 −ν 2 , ν 1 +2ν 2 −ν 2 ) are reported in this range. A comparison of laboratory spectra with long path atmospheric spectra (20 km slant path in the mountains) in this region shows that many lines missing from available spectroscopic compilations (or considerably shifted compared to observations) are important for a proper interpretation of atmospheric observations. A comparison of the observed data with the best available predictions from the molecular electronic potential energy surface is discussed.