Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 355, 1041-1048 (2000)

Previous Section Next Section Title Page Table of Contents

3. Spectroscopy

Low-dispersion spectroscopy has been performed on 1999 Jan. 22 with the EFOSC instrument at the 3.6m telescope on La Silla (ESO). A 300 grooves/mm grating has been used, giving a dispersion of 2.1 Å/pix and covering the wavelength range 3800-8000 Å. With the use of a [FORMULA] slit (adapted to the seeing during that night) the FWHM resolution is 19 Å. The slit was oriented along the parallactic angle always. A total of seven consecutive spectra have been taken with exposure times of 1800 sec each except for the third spectrum with 2700 sec. Due to the faintness of the object and the suspected short orbital period the length of the exposures was a compromise between achieving a reasonable S/N ratio and minimizing the phase coverage of each exposure. One-dimensional spectra were extracted and processed using standard MIDAS routines. Wavelength calibration was done using He-Ar spectra taken before and after the sequence of spectra. The standard star GD108 was used for flux-calibration.

The spectrum of RX J0537.7-7034 has a steep, blue continuum and a moderately strong HeII [FORMULA]4686 emission line. The Balmer series appears in absorption except for H[FORMULA] which seemingly is filled (Fig. 3). No other notable features are seen in the individual spectra. The phase-averaged sum of all spectra (Fig. 3) allows to identify a few other emission lines, among them HeII [FORMULA]4541, the CIII /NIII [FORMULA]4630-4650 complex, and possibly also OVI [FORMULA]5290.

[FIGURE] Fig. 3. Sum of all 7 spectra corresponding to an exposure time of 13 300 sec. The HeII [FORMULA]4686 emission line is prominent, and other detected lines are also marked. Note that the Balmer series is in absorption. With the present signal-to-noise ratio the existence of possible emission cores within the absorption troughs remains unclear.

Despite the low S/N of the individual spectra we attempted a measurement of the HeII [FORMULA]4686 line strengths and velocities. The result is given in Table 2 and plotted in Fig. 4. The HeII [FORMULA]4686 emission line shows a surprisingly clear radial velocity variation. Fitting a single sinusoidal curve yields a semiamplitude of K = 115[FORMULA]20 km/s and a period of P = 3.30[FORMULA]0.15 hrs ([FORMULA] = 0.95). The errors have been determined by a two-dimensional [FORMULA] fitting, i.e. the maximum and minimum possible periods have been determined for various values of the phase zero-point. If we use the standard procedure with a [FORMULA]-minimization only along the y-axis (radial velocity), we get K = 125[FORMULA]40 km/s and a period of P = 3.45[FORMULA]0.25 hrs ([FORMULA] = 1.1). The latter solution is plotted in the lower panel of Fig. 2. Though the errors for the spectroscopic period are large, the spectroscopic period matches one of our tentative photometric periods, namely the shortest one at 0.147275 d. We therefore adopt this value as the best orbital period, and the ephemeris for this is:


[FIGURE] Fig. 4. Radial velocity curves of the HeII [FORMULA]4686 emission line (filled squares) and of the H[FORMULA] (open squares) and H[FORMULA] (open triangle) absorption lines as measured from the seven spectra taken in January 1999. The systemic velocity of 346 km/s has been subtracted. The emission line data are fitted with a sinusoidal curve (dashed line), resulting in a best-fit period of 3.3 hours. The velocity of the Balmer absorption lines seems not to be sinusoidal, and the amplitude is considerably larger than those of the HeII line.


Table 2. HeII [FORMULA]4686 velocities (after subtraction of a systemic 346 km/s) and equivalent widths.

The photometric data of all 5 nights, folded with this period, are shown in Fig. 5. This demonstrates more clearly that while the light maxima have nearly identical brightnesses, the minima vary by slightly more than 0.2 mag from one night to the other and to a somewhat lower degree even from one orbital cycle to the next (January 24 and 25, 1999). Also, the irregular brightness fluctuations are clearly visible.

[FIGURE] Fig. 5. Photometric data folded with the adopted period of 0.147275 d. The top panel shows [FORMULA].

Due to the uncertainty of our final period and the 2 day difference between spectroscopy and photometry we are unable to establish a relative phasing between radial velocity and intensity modulation.

Attempts to measure the absorption line velocities of H[FORMULA] and H[FORMULA] are very difficult due to both, the poor S/N and the possible distortion by emission cores (H[FORMULA] is completely filled, and H[FORMULA] is at the very edge of our wavelength coverage). Nevertheless, it is obvious that these Balmer absorption lines exhibit a considerably larger velocity amplitude than the HeII emission line (Fig. 4). Also, the velocity curve does not seem to be sinusoidal, and possibly out of phase with respect to HeII [FORMULA]4686. A formal (one-dimensional) fit of a circular orbit yields a period of 3.33 hrs, consistent within the errors to the HeII emission line period.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: March 21, 2000