Astron. Astrophys. 346, 82-86 (1999)

2. Observations and data reduction

We obtained K-band (2.0-2.5 µ) spectra of V616 Mon using the Cooled Grating Spectrometer (CGS4) on the United Kingdom 3.8-m Infrared Telescope on Mauna Kea, during the night of 1997 November 13. The 40 l/mm grating was used with the 300 mm camera and the 256256 pixel InSb array. The bright star BS2233 was also observed through-out the night in order to remove telluric atmospheric features. A journal of observations is presented in Table 1.

Table 1. Journal of observations

In order to minimise the effects of bad pixels, the standard procedure of oversampling was used. The spectra were sampled over two pixels by mechanically shifting the array in 0.5 pixel steps in the dispersion direction, giving a full width half maximum resolution of 47 Å (610 km s^-1 at 2.31 µ). We employed the non-destructive readout mode of the detector in order to reduce the readout noise. The slit width was 1.2 arcseconds which corresponds to 2 pixels on the detector. In order to compensate for the fluctuating atmospheric emission lines we took relatively short exposures and nodded the telescope primary so that the object spectrum switched between two different spatial positions on the detector. Throughout the observing run the slit orientation was north to south in the spatial direction.

The CGS4 data reduction system performs the initial reduction of the 2D images. These steps include the application of the bad pixel mask, bias and dark subtraction, flat field division, interlacing integrations taken at different detector positions, and co-adding and subtracting the nodded images (see Daly & Beard 1994). Extraction of the 1D spectra, wavelength calibration, and removal of the telluric atmospheric features was then performed using IRAF. A more detailed description of the data reduction procedure is provided in Shahbaz et al. (1996).

First the individual spectra of V616 Mon were averaged using the variance in the spectrum as the weight. This resulted in a final summed spectrum which had a signal-to-noise ratio of 30. We then cross correlated all the summed V616 Mon spectrum with the template star spectrum in order to determine the velocity shift. Given the poor resolution of the data this was not done on the individual spectra of V616 Mon before they were averaged. We applied the appropriate velocity shift and then binned all the spectra onto a uniform velocity scale. Fig. 1 shows the summed spectrum of V616 Mon and the K3V template star spectrum, normalised and shifted for clarity. The V616 Mon spectrum shows the CaI triplet and ¹²CO bands in absorption and doubled-peaked Br in emission.

Fig. 1. This figure shows the spectrum of V616 Mon and that of the template K3V star. The 12CO bandheads can clearly be seen. Also shown is the residual spectrum after optimally subtracting the template star from the V616 Mon spectrum. The lower-most spectrum is of an F6V star (BS2233) which indicates the location of telluric absorption features. All spectra have been normalised by dividing by a spline fit to the continuum. The stars indicate bad pixels.

The peak separation of the Br emission line arising from the accretion disc is 1204156 km s^-1. Note that this is comparable with the peak-to-peak separation in the H and H emission lines (MRW). The equivalent width (EW) of Br in the K3V star spectrum is -1.60.3 Å whereas in the residual spectrum [i.e. the spectrum of the accretion disc; see Fig. 1 and Sect. 3] it is 14.61.3 Å. The ¹²CO bands are the strongest features and will be used in the next section to determine the fraction of light arising from the secondary star.

© European Southern Observatory (ESO) 1999

Online publication: May 6, 1999
helpdesk.link@springer.de