Astron. Astrophys. 363, 1065-1080 (2000)

3. Observations

High resolution ( 70000) 2 µm region spectra of a number of LPV's were observed in the 1980's using the FTS at the 4 meter Mayall telescope at Kitt Peak National Observatory. For about 20 Miras and semiregular (SR) variables, time series spectra with a sampling of about 0.1 in phase or better exist. Most of these observations are described in detail in HHR, HSH, Wallerstein et al. (1985), and Hinkle et al. (1997). Table 2 is a list of spectra that were examined for H₂. As will be discussed below, time series of Mira variable spectra show the H₂ lines to be strongest near minimum light. As a result, phases near minimum light have been selected for the Mira variable spectra in Table 2. Most of the spectra cover the 1.5-2.5 µm region. For some of the fainter stars the spectrum was only observed over the 2.0-2.5 µm region.

Table 2. Minimum light 2 µm spectra. Classification in Column 5 is based on major spectral features in 2 µm region. m+=strong H₂O, m=moderate H₂O, m-=weak H₂O, s=weak H₂O and weak CN, c=strong CN (see text). References (Column 11): (1) HSH, (2) Lambert et al. 1986, (3) Hinkle et al. 1989, (4) Hinkle et al. 1997, (5) Lebzelter et al. 1999, (6) Lebzelter et al. 2000, (7) Wallerstein et al. 1985.

For the Mira Cygni a sample of spectra covering the entire light cycle is used. For the most part these are observations of HHR. New observations at an apodized resolution of 0.07 cm^-1 and signal-to-noise ratio approaching 100 have also been obtained and are listed in Table 3. These observations, like the majority of the other data, cover the 1.5 to 2.5 µm region.

Table 3. Additional Time Series Spectra of Cygni

The signal-to-noise ratios reported in Table 2 and Table 3 are for the peak signal. The peak signal in these spectra typically occurs near 2.2 µm with the signal-to-noise ratio in the 1.5-1.8 µm region less by as much as a factor of two, the actual amount depending on the phase-dependent energy distribution. Fortunately, the peak signal-to-noise region corresponds to the region where the most interesting H₂ lines are found.

All the spectra were observed with one of two Fourier transform spectrometers at the coudé focus of the Kitt Peak 4 meter telescope. Observations prior to October 1978 were obtained with a prototype spectrometer similar to the instrument described by Ridgway & Capps (1974). The remainder of the spectra were observed by the facility spectrometer described by Hall et al. (1979). Adequate discussion of the instrumentation and general characteristics of the data may be found in the papers listed above. We re-emphasize that the photometric accuracy and frequency calibration of spectra obtained with the FTS are limited by noise rather than instrumental characteristics. Frequencies are referenced to the spectrometer's laser frequency and after correction for the index of refraction of air, the spectral frequencies require only the addition of a small ( 1 km s^-1) correction for collimation differences between the reference and signal beams.

All spectra discussed in this paper have been apodized by function I2 of Norton & Beer (1976). Apodizing damps the wings of the intrinsic sinc function FTS instrumental profile, creating an instrumental profile similar to that of a grating spectrograph. In this process the apodizing function lowers the resolution and increases the signal-to-noise ratio. The prototype FTS suffered from instrumental apodization resulting from systemic variation in reimaging with path difference. This results in the resolution being slightly lower than the theoretical value presented in Table 2 for the 1976 and 1977 observations.

All velocities in this paper are heliocentric. Velocities listed for lines are from the line core. Accuracies per line are better than 0.5 km s^-1. Measurement techniques are described by HHR.

© European Southern Observatory (ESO) 2000

Online publication: December 5, 2000
helpdesk@link.springer.de