Traces of HCl and HF in the atmosphere of Venus

Abstract

High-resolution interferometric spectra of Venus show weak, narrow absorption lines which are assigned to the 2-0 vibration-rotation band of HC1 and the 1-0 and 2-0 bands of HF. The HC1 lines are consistent with 2-mm Amagat of that gas in the optical path, at temperatures near 240° K and pressure near 0.1 atm, in a spectral region where the effective CO« path is 1.7 X 10® times as large. The less-extensive HF data indicate about 0.02 mm Amagat of that gas. J. and P. Connes (1966) have described the development of a Michelson interferome-ter for high-resolution planetary infrared spectroscopy. Further advances have since been made; they were described at the Bellevue Conference on Instrumental Spectrosco-py (April, 1966) and will be published together with the proceedings of this conference. The instrument was used in June and early July 1966, at the Observatoire de Saint-Michel to study the planet Venus. Three spectral regions, corresponding to three atmospheric windows, 3950-5300 cm" 1 , 5500-7200 cm-1 , and 7600-8500 cm-1 were isolated by interference filters. All observations were made in daytime, most of them in the mornings. Each recording took about 3 hours (more when it had to be interrupted because of clouds). Altogether, six solar and twenty-eight Venus spectra were considered of sufficient quality to be included in the final results. The accuracy of the spectra can best be described by comparing them to the Michigan Atlas (Mohler, Pierce, McMath, and Goldberg 1950) of the solar spectrum. The resolution (after suitable apodization) is constant throughout the range and equal to 0.08 cm" 1 , which is three times better than that of the Atlas (about 0.25 cm" 1). The signal-to-noise ratio of the solar spectra is comparable to that of the Atlas; for the Venus spectra, it is slightly inferior. The spectra are displayed by an automatic plotter which gives on a single chart one solar and (with a shifted base line) two planetary traces. Each of these is an average computed from several interferograms recorded on different days. One planetary trace contains all interferograms taken in early morning (at low elevations) and the other all interferograms recorded near the meridian. In this way it is possible to discriminate clearly among the three classes of absorption lines: (1) telluric absorption, which falls at the same position in the solar and planetary spectra, and with intensity depending on the zenith angle; (2) solar Fraunhofer lines, which are shifted to new positions in the planetary spectra by the Doppler effect but which retain their intensity; and (3) many thousands of lines which appear only in the Venus spectra (also at Doppler-shifted positions) and which are formed in the atmosphere of the planet. The accuracy of the wavenumber scale in the spectra (derived solely from the known figure for the Hg 198 5461 Â line, which is used to control the interferometer motion) has been checked by comparing the measured values for a set of thirty-five sharp and intense CO lines of Venus to the laboratory values as measured by Rank, Birtley, Eastman, and Wiggins (1960). After reduction to vacuum and correction for both the translational Doppler effect and the diurnal Doppler effect due to Earth rotation, a mean error of about 1 X 10" 3 cm" 1 was found with a 2 X 10" 3 cm-1 rms scatter. This accuracy is distinctly superior to the one achieved in the Michigan Atlas. The complete spectra (together with Jupiter spectra) are being prepared for publication. They will be made available both in atlas form by the Journal des Observateurs and also on magnetic tape in IBM format, since only in this way can the full accuracy of the wavenumber calibration be realized.