2. Observations and analysis
2.1. ROSAT observations
The survey observations of RX J0203 were obtained between 1991 January 23 and 24. For 1.33 days the field was monitored during 19 individual scans with observing times from 10 to 28 seconds accumulating to a net exposure time of 424 seconds. Using the improved survey analysis RASS2 (Voges et al. 1997) 229 source photons could be detected which yield a best-fit position = 02 03 48 :s7 and . The peculiar nature of RX J0203 became obvious from its almost 100% flux modulation reaching a maximum of 2 cts s-1 and its very soft X-ray spectrum indicated by the average hardness ratio . HR1 is defined as (H-S)/(H+S), with H and S being the counts above and below 0.4 keV, respectively. The only optical counterpart within the (2 ) error-circle is a blue stellar object (). A finding chart is reproduced in Fig. 1.
RX J0203 was the target of follow-up pointed ROSAT observations in July/August 1993 and in August 1995 using the PSCP and the HRI detectors, respectively. The effective exposure times of 3.04 and 2.57 hours were split into 12 and 5 sections spread over several epochs. For further investigation the exsas (Zimmermann et al. 1994) software package was used. Source photons from the event tables were extracted using concentric apertures centered on the source position. The background was estimated within a ring around the source cleaned from probable contaminating sources. The source counts were then background - subtracted, corrected for dead time and vignetting and finally binned in time and pulse height channel. During both pointings RX J0203 was observed on-axis, hence the PSPC light curve may be affected by shading of the source by the entrance window support structure with the space-craft wobble period.
A large data set of 115 hours of CCD photometry was collected between 1992 and 1997 with the 1 m telescope at the Mount Laguna Observatory, the 0.4 m telescope of the Braeside Observatory and the 0.7 m reflector associated to the Astrophysical Institute Potsdam, Germany.
The Mount Laguna observations were obtained with a Texas Instrument CCD which was operated in fast read-out mode. Broad-band filters that were designed to reproduce the Johnson - Cousins bandpasses were used. The observations were obtained under photometric conditions with seeing varying from 1.5 - 2.5 ".
The reflector in Potsdam was equipped with a TEK CCD. Except from one night the observations were taken with a Johnson V filter. A 22 binning was applied at a moderate seeing of 3.5 arcsec. After standard calibration (bias, flat-field) of the frames differential magnitudes of RX J0203 were derived with respect to star # 1 marked in Fig. 1.
The Mount Laguna and Braeside data was reduced by calculating aperture magnitudes using IRAF routines. The quality of the Potsdam photometry, when RX J0203 was in general fainter on a brighter sky-background could be significantly increased using profile-fitting magnitudes computed with the DoPHOT photometry package (Mateo & Schechter 1989). On 1996 February 9 a field of standard stars in M67 (Chevalier et al. 1991) was observed from which we derived a V magnitude for star #1 of 13:m 29 mag.
Finally three short runs of high-speed (1 sec) photoelectric photometry were obtained with the Multi-Channel multi Color Photometer (MCCP) mounted at the 2.2 m telescope at Calar Alto in August 1993. A detailed overview of all observations is given in Table 1.
Table 1. Log of optical photometric, spectroscopic and ROSAT X-ray observations of RX J0203.8+2959 a) MLO: Mount Laguna Observatory AIP: Astrophysical Institut Potsdam BO: Braeside Observatory DSAZ: German-Spanish Astronomical Center b) TI, TEK, SITe: CCD cameras MCCP: Multi-Channel multi-Colour Photometer CTS: Cassegrain Twin Spectrograph B&C: Boller & Chivens Spectrograph PSPC: Position Sensitive Proportional Counter HRI: High Resolution Imager c) Photometry: UBVJRIC filter or WL white-light Spectroscopy: Wavelength range and resolution (Å) X-ray: Energy range and resolution (keV) d) optical: number of exposures X-ray: RASS: number of scans, pointings: number of observing intervals (OBIs)
2.3. Spectroscopic observations
Time-resolved low resolution spectroscopy of RX J0203 was obtained on 1992 August 20 and 21 with the 3.5 m telescope at the DSAZ (German-Spanish Astronomical Center) at Calar Alto, Spain. The telescope was operated with the Cassegrain Twin Spectrograph (CTS) which splits the light at 5500 Å. In the blue and red channels low dispersive grisms ( = 144 and 160 Å mm-1) were employed. The resulting spectra cover the whole optical range (3500-9600 Å) with a moderate resolution of 7 Å FWHM. The 41 spectra were taken through a 1.5" slit while the seeing was variable around 1.5". Observations of standard stars during the nights provide a flux calibration with an estimated accuracy of roughly 30%. The wavelength scale was calibrated by HeAr arc spectra taken before and after a sequence of spectra. Systematic shifts during a sequence could be compensated using the positions of the night-sky lines. Both observations cover 85 % of the binary cycle with a phase resolution of . High resolution phase-resolved spectroscopy was carried out in 1993 on August 17 and October 21 with the 3.5 and 2.2 m telescopes, respectively, at Calar Alto. The observations at the 3.5 m were again obtained with the double-beam spectrograph equipped with high resolution gratings giving an effective resolution in the blue and red bands of 2 Å. The 10 spectra taken at new moon show a high signal to noise ratio but span only 30% of the orbit. The observations at the 2.2 m telescope were carried out with the Boller&Chivens spectrograph operated with a grating yielding a resolution of 2 Å. The 17 spectra cover more than one orbital cycle.
© European Southern Observatory (ESO) 1998
Online publication: September 14, 1998