Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 318, 134-139 (1997)

Previous Section Next Section Title Page Table of Contents

2. Observations and results

RX J2353.0-3852 was discovered in the ROSAT all-sky survey as a source with intensity [FORMULA] counts s-1. We obtained optical observations of the field around the source at ESO, La Silla as part of an ESO Key Programme on ROSAT identifications (Danziger et al., 1990). The object marked in Fig. 1 was identified as the optical counterpart of RX J2353.0-3852.

[FIGURE] Fig. 1. Finder chart of RX J2353.0-3852. Coordinates (2000): 23 53 00.6 -38 51 45. The bright object to the northeast is nonstellar.

2.1. Optical spectroscopy

On 1992, October 1, classification spectra of RX J2353.0-3852 were obtained with EFOSC on the ESO 3.6-m telescope. Two spectra were obtained, with the B300 and R300 grisms and using a 1."5 slit, yielding a resolution of [FORMULA] Å. The integration time was 10 minutes in each case and the spectra were taken approximately 12 min apart. The data were reduced using the MIDAS software package. The spectra (Fig. 2) are dominated by strong, double-peaked Balmer emission which is embedded in broad absorption from H [FORMULA] onwards. The H [FORMULA] line profile is clearly different in the two spectra. He I [FORMULA] 5876, 6678, 7065, Fe II [FORMULA] 5169 and He I [FORMULA] 5016 and/or Fe II [FORMULA] 5018 are also present and similarly doubled. He II is not observed. There is no evidence of a contribution from a late-type secondary star. Equivalent widths and peak-separation of some lines are given in Table 1.

[FIGURE] Fig. 2. Low resolution spectra of RX J2353.0-3852. Note the strong, double-peaked Balmer emission, embedded in broad absorption troughs from H [FORMULA] onwards. The upper, red, spectrum has been shifted vertically by [FORMULA] erg cm-2 s-1 Å-1.


Table 1. Emission line strengths of RX J2353.0-3852 (see Fig. 2)

An image taken two days earlier at 1992, September 29, 03:55 UT showed RX J2353.0-3852 was at V [FORMULA] 16.5.

2.2. Optical photometry

Time series CCD photometry of RX J2353.0-3852 was obtained using the Dutch 0.9-m telescope and a thinned Tektronix 512 [FORMULA] 512 pixel CCD (ESO CCD #33) in 1993 November and 1994 September according to the observing log given in Table 2. Stellar brightnesses were extracted with an IDL data reduction package written by TMCA and using circular apertures; the resulting differential light curves are shown in Figs. 3 & 4. Although the same differential comparison star was not used in all cases, the light curves could all be converted to the same instrumental magnitude scale. Light curves of other stars in the field are provided for comparison. Typically, there is a dead time of [FORMULA] 45 s between images.


Table 2. Log of observations of RX J2353.0-3852 at ESO, Dutch 0.9-m telescope with CCD # 33. No filter was used.

[FIGURE] Fig. 3. Light curves of RX J2353.0-3852 from 1993 November 15 to 18. The X-axis is in hours, starting at HJD 2449306.50, the Y-axis is in arbitrary instrumental magnitudes. The open circles are light curves of simultaneously observed comparison stars. The comparison star light curve in the top panel has been raised vertically by 0.77 magnitudes. The solid line is the result of a nonlinear, least-squares fit of a multi-component cosinusoid, with coefficients given in Table 3.
[FIGURE] Fig. 4. Light curves of RX J2353.0-3852 from 1994 September 12 and 14. The X-axis is in hours, from HJD 2449607.50, the Y-axis is in arbitrary instrumental magnitudes. The open circles are light curves of a simultaneously observed comparison star which has been raised vertically by 0.77 magnitudes.

A V image of RX J2353.0-3852 taken at 00:21 UT on 1993 November 17 showed the source at V = 16.5.

A discrete Fourier transform of the 1993 November data is shown in Fig. 5. A strong peak at 5246 [FORMULA] 55 sec is clearly visible, an inspection of the observational window (inset) suggests that this is not a one-day alias of the true periodicity. Pre-whitening the light curves with a nonlinear, least-squares fit of a cosinusoid at this frequency and its first 3 harmonics, and then repeating the DFT reveals the presence of a further periodicity at 2282 [FORMULA] 18 sec. We repeated the exercise of fitting a cosinusoid at the strongest periodicity, its first three harmonics and also this latter periodicity, the result is shown as the solid line in Fig. 3 and summarized in Table 3.

[FIGURE] Fig. 5. A discrete Fourier transform of the light curves given in Fig. 3. The highest peak is at a periodicity of 5246.59 [FORMULA] 55 Sect. The individual light curves were normalized with a second-order polynomial fit before computing the transform. Inset is the window function, on the same frequency scale.


Table 3. coefficients of a nonlinear, least-squares cosinusoid fit to the light curves of 1993 November 15 to 18.

A DFT of the 1994 September light curves shows no significant periodicities and the light curves are dominated by incoherent variations and noise.

2.3. X-ray observations

A pointed observation of RX J2353.0-3852 was obtained with ROSAT PSPC starting 1993 November 14 21:43 UT with a total coverage of 4566 seconds. Unfortunately, this does not overlap with the optical observations made the same night. We measured a count rate of 0.095 [FORMULA] 0.005 counts s-1 and as SASS hardness ratio HR1 = [FORMULA] 0.14 for the X-ray source. The higher count rate measured in the survey partially reflects the increased gain used but this difference is expected to be [FORMULA] 20%, and RX J2353.0-3852 appears to be a variable X-ray source and was in a low state during the pointed observation.

The X-Ray light curve (Fig. 6) does not appear to correlate with the signals observed in the optical, whether extrapolated directly as shown in Fig. 6, or with the 5246 s or 2282 s oscillations at any phase. The maximum correlation coefficient between the 2282 s signal and the X-ray light curve is [FORMULA] 0.4. However, the duration and S/N of the observation is not sufficient to rule out the presence of such signals.

[FIGURE] Fig. 6. ROSAT light curve of RX J2353.0-3852, 1993 November 15. The observed X-ray events have been accumulated into 120 sec bins. The lower solid line is the extrapolation to the period of the X-ray observations of the cosinusoidal model fit to the optical light curve, summarized in Table 3. The upper solid line is the extrapolation of the 2282 sec component of this model.The X-ray spectrum of RX J2353.0-3852 is shown in Fig. 7.

Using XSPEC (Shafer et al. 1991) we fit the following models to the X-ray spectrum:

  1. A simple bremsstrahlung model: [FORMULA] = 0.95 keV, [FORMULA] = 1.5 [FORMULA] 1020 cm-2, [FORMULA].
  2. A two component Raymond-Smith model: [FORMULA] = 0.15 and 1.24 keV, [FORMULA] = 1.6 [FORMULA] 1020 cm-2, [FORMULA].
  3. A power law: photon index = 2.33, [FORMULA] = 2.6 [FORMULA] 1020 cm-2, [FORMULA].
  4. A combined blackbody and thermal bremsstrahlung model: [FORMULA] = 0.24 keV, [FORMULA] = 0.93, [FORMULA] = 1.4 [FORMULA] 1020 cm-2, [FORMULA] (integrated X-ray flux = 1.14 [FORMULA] 10-12 erg s-1 cm-2 (0.1 - 2.4 keV) in this model).

Fig. 7 shows the folded X-ray spectrum and the last, combined model fit.

[FIGURE] Fig. 7. The folded ROSAT X-ray spectrum of RX J2353.0-3852. The solid line is the combined thermal bremsstrahlung and blackbody model as described in the text. Note the excess flux at 0.2-0.25 keV.

In all cases we were unable to obtain satisfactory fits in the range 0.2-0.25 keV, where there appears to be significant excess flux. This excess is present throughout the pointed observation and is not well fit by a single spectral line. The same excess flux was not present in the background and is therefore not of solar origin. However, Snowden (1992) points out that the response matrix near 0.27 keV is not well determined suggesting an instrumental origin for the excess flux we observe.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1997

Online publication: July 8, 1998