Astron. Astrophys. 345, 121-126 (1999)

3. Example: An X-ray flare on G 131-026

In the direction of the PSPC pointing ID 700101 (P.I. Turner) with the UV Cet type star G 131-026 in the field of view, we identified a total of 117 X-ray sources above our detection threshold (). All of the X-ray sources were cross-referenced with the SIMBAD and NED databases as well as with the UV Cet type flare stars and related objects catalogue (Gershberg et al. 1998). In addition, we inspected the HST Guide Star Catalogue and the Digitized Sky Survey images. In all cases, 40 arc sec circles around each X-ray source were chosen to identify possible optical counterparts to the X-ray sources (Neuhäuser et al. 1995).

The star G 131-026, also known as LTT10045, LP404-33, and CRSS J0008.9+2050, has a spectral type of M4.5e and is located at the distance of (Reid et al. 1995). We have identified it with the very strong X-ray source RX J0008.8+2050 ().

The procedure described above for variability testing has been applied to this source. Our method has divided the observation into six time segments in the broad energy band , while the dataset of the corresponding background can be represented by one segment, i.e. no significant variations have been observed in the background at the same time. As seen from Fig. (1) during the observation this star has also shown another flare event possibly triggered by the first one with relatively smaller amplitude. It should be noted that the more powerful flare is observed during two successive observational intervals lasting 161 and 1420 seconds each. The time gap between them was equal to 110 sec. In order to estimate flare parameters we have converted the observed count rates to the fluxes and luminosities for each segment subtracting background and radiation of the star in the quiescent level (the leftmost segment in the second panel of Fig. 1).

Fig. 1. ROSAT PSPC observations of the UV Cet type star G 131-026: In the upper panel, calibrated amplitudes (i.e. photon energy) of registered photons are plotted vs. arrival times. The ploting photon energy vs. arrival times and ignoring of observational gaps in the data set have been used for purposes of clarity. The second panel shows the Bayesian-blocks decomposition of these data. The next two panels show histograms of the same data binned into equal time intervals of 25 and 400 sec, respectively (except the last bin). In the last panel, the smaller flare near the end of the observation, is almost undetected, due to long time bins.

The count rates of the flare in each energy band are expressed as

where is the quiescent star radiation count rates in the particular energy band, determined from data outside the flare event. This quantity can be used to determine the so called equivalent duration of the flare

where is given by Eq. (5) for each kth segment. expresses the energy of the flare in terms of the quiescent star's energy.

An estimate of the luminosity of the quiescent star, in each band, can be made by multiplying count rates by an energy conversion factor and taking into account the distance of the star. Assuming that the spectrum of the star in the quiescent level is consistent with a one-temperature Raymond-Smith spectrum (Raymond & Smith 1977), a thermal spectrum from an hot, optically thin plasma of solar abundance, we use 1 keV, i.e. , as temperature of the X-ray emitting plasma, which is typical for late-type stars (Neuhäuser et al. 1995). Because foreground absorption is negligible for our object, we use an energy conversion factor of (Zimmermann et al. 1998) for the PSPC observation.

Furthermore, flare detection and parameter estimation procedures have been applied to different energy bands. Namely, so called hardness ratios (X-ray colors) are defined as follows: If are the count rates in the bands soft (0.1 to ), medium (0.5 to ), and hard (0.9 to ), respectively, then

Ie., hardness ratios range from -1 to . If no counts are detected, e.g, in the soft band, then , but one can estimate a lower limit to by using the upper limit to the soft band count rate in the formula above.

These quantities have been computed for different observational phases, i.e. for the star in the quiescent level as well as for flare radiation alone. The results are presented in Table 1. The first column gives the Julian date at flare onset in different energy bands. Column 2 gives the energy bands at which the flare is detected (Broad: 0.1-2.4 keV, Soft: 0.1-0.4 keV, Medium: 0.5-0.9 keV and Hard: 0.9-2.0 keV). Column 3 is the normalized count rate at flare maximum as determined from Eq. (5). Eq. (6) was used to determine the equivalent duration in seconds, listed in Column 4. The count rates may be converted to the flare energies in different energy bands by multiplying by quiescent luminosity of the star given in Column 5 which is the logarithm of X-ray luminosity of the star at quiescent level and determined by the method outlined in this section. Columns 6 and 7 contain hardness ratios of the flare emission plus quiscent radiation, while the hardness ratios of the flare radiation and of the star in quiescent level alone are given in Columns 8,9 and 10, 11, respectively. These hardness ratios correspond to count rates observed in the segment including flare maximum and, for star only, the segment just before flare onset. The estimates of variations of hardness ratios are calculated by variations of count rates in each energy band. A more sophisticated treatment of best-fit parameter values and ranges is possible using full posterior probabilities distribution, but these simple intuitive estimates will be adequeate here. Obviously, the X-ray emission was harder during the flare, i.e. the plasma was hotter.

Table 1. Flare parameters for RX J0008.8+2050

It is worthwhile to note that there is some time delay between the onset of the flare in different energy bands (see, Fig. 2). Namely, the flare radiation has started in the soft band, then in the medium band, and more later in the hard one. The ending of the flare was in the reverse order. Of course, this phenomenon might be different from one flare to another and a correct conclusion can be drawn only on the basis of statistics of significant number of flares which will be the subject of our subsequent paper (Hambaryan et al., in prep.).

Fig. 2. Bayesian-blocks of the ROSAT PSPC observations of the UV Cet type star G 131-026 for observational intervals including the flare in different energy bands: Soft: 0.1-0.4 keV, Medium: 0.5-0.9 keV, Hard: 0.9-2.0 keV and Medium+Hard: 0.5-2.0 keV.

© European Southern Observatory (ESO) 1999

Online publication: April 12, 1999
helpdesk.link@springer.de