Astron. Astrophys. 348, 211-221 (1999)

3. Observations and data reduction

The observations were carried out with the ESO Coudé Echelle Spectrometer (CES) in three different periods: October 22-27, 1992, June 6-8, 1993, and February 5-9, 1995. In Oct. 92 and Feb. 95 the CAT 1.4m telescope was applied, whereas in June 93 the 3.6m telescope was feeding the CES through a 35m long fiber and an image slicer (D'Odorico et al. 1989). On all occasions the detector was a front illuminated Ford Aerospace CCD with 15 µm pixels. The CES camera has a dispersion of 1.88 Å mm^-1 at 6700 Å providing a spectral coverage of about 58 Å and 0.0282 Å pixel^-1. Different settings for the central wavelength were applied on the various nights in order to avoid any signatures in the spectrum resulting from a possible improper flatfielding.

The main targets for the June 93 observations were some halo turnoff stars, for which the results have been presented by Smith et al. (1998) including a description of the reductions of the image slicer observations. The Oct. 92 and Feb. 95 observations have been reduced in much the same way. Subtraction of background, flat field correction and extraction of spectra were performed with standard tasks in IRAF. The wavelength calibration was based on 28 thorium lines well distributed over the wavelength region (6676-6734 Å). A second-order polynomial was adopted for the dispersion solution resulting in a typical rms deviation from the fit of 0.0015 Å. Furthermore, the FWHM of the thorium lines were measured and found to vary by less than 4% along the spectrum. This near-constancy of the instrument profile is important in connection with the determination of instrumental and stellar line broadening from Fe I lines near the LiI line. For the Oct. 92 and Feb. 95 observations the FWHM of the Th lines corresponds to a resolution of , whereas the June 93 image sclicer spectra have a resolution of . Finally, the spectra were normalized to an approximate level of 1 by fitting a fifth-order cubic spline function to the continuum, and corrected for the radial velocity shift using the accurate wavelengths (6703.567 and 6705.102 Å) given by Nave et al. (1995) for the two Fe I lines close to the the Li line.

Table 2 lists the number of observations and exposure times as distributed over the three periods. Note, that HR 3578 has been observed in all three periods and HR 8181 in two of them. The individual spectra show excellent agreement and have therefore been co-added to obtain the final spectra shown in Fig. 1 for the 6702-6710 Å region. These are the spectra used for the determination of the ⁶Li/⁷Li ratio, but in the case of HR 3578 we have also derived the ratio separately for each observing period in order to check for possible systematic differences.

Fig. 1. The observed spectra in the 6702-6710 Å region. Spectral windows used in defining the continuum and in the analysis of the lines are marked with C and L, respectively. Line identifications are shown for HR 3578

Table 2. Number of observations and exposure times for the program stars

© European Southern Observatory (ESO) 1999

Online publication: July 16, 1999
helpdesk.link@springer.de