5 Search for short and long term time variability

5.1 Variable sources within the XMM-Newton observation

We have extracted the light curves of the 26 strongest X-ray sources within the EPIC-PN field of view. The bin time of the light curve was chosen to ensure a sufficient number of count within each bin (typically larger than $\sim$20). The light curves were searched for variability using the lcstat FTOOLs. Using a $\chi^2$ test, we found that 4 sources have a probability of being constant of less than 0.1%. The light curves of these 4 objects are shown in Fig. 6. Only one lies in the core (source 24). The three others lie outside the half mass radius.

Figure 6: The light curves of the four variable sources found within the field of view. Top The brightest object in the field (source 1). 2nd The star USNO-A2 0375-18249604 identified by Cool et al. (1995) (source 3). 3rd An object which is located within twice the half mass radius (source 15). 4th. The only core source which showed variability within the XMM-Newton observation (source 24).

5.2 Variable sources between the Chandra and XMM-Newton observations

As shown above, all XMM-Newton sources should have been detected by Chandra, providing that their X-ray intensity had remained constant. In the core this is the case, as the only XMM-Newton source not detected by Chandra fell on an ACIS-I CCD gap. Within the half mass radius, 4 XMM-Newton sources were not detected by Chandra and must have therefore varied by at least a factor of $\sim$3-4.

We have converted the count rate of the Chandra sources to XMM-Newton count rates for comparison with our count rate detection threshold. We have found that 7 Chandra sources that lie within the half mass radius should have been detected by XMM-Newton and were not (this number rises to 20 if one considers the whole field of view). Finally, within the half mass radius there are 4 sources (13 in the whole field of view) detected by both Chandra and XMM-Newton but with different luminosities (a factor of two or higher variations). These sources are listed in Table 4.

 

 

Table 4: Sources for which a flux variation of a factor two or higher has been detected between the Chandra and XMM-Newton observations. The source list is declination sorted, as in Tables 1 and 2. The unabsorbed Chandra and XMM-Newton fluxes are computed from the detected count rates using a 0.6 keV blackbody spectral model and are given in units of 10^-15 ergs s^-1 cm^-2.

Source	XMM	Chandra	Half mass
ID	flux	flux	radius
			source
73	7.1	3.3	No
32	12.6	67.2	No
39	14.2	5.3	No
18	25.1	11.2	No
49	4.0	15.0	Yes
24	14.7	4.8	Yes
47	7.7	2.3	No
13	11.0	35.6	No
117	4.0	1.2	No
9	40.0	13.7	Yes
108	4.3	2.0	Yes
8	57.4	20.7	No
23	21.1	43.3	No

From this we conclude that 15 sources contained in the half mass radius have shown variability between the XMM-Newton and Chandra observations spaced by $\sim$1.5 year.

5.3 Variable sources between the XMM-Newton and ROSAT observations

$\omega$ Cen was observed between August 1992 and January 1997 by ROSAT (Verbunt & Johnston 2000). The luminosity limit of the ROSAT observations was about $7\times10^{31}$ ergs s^-1 in the 0.5-2.5 keV range (Verbunt & Johnston 2000). We have computed the luminosity of the XMM-Newton sources in this same energy band. From this, we found that 1 source (source 13) should have been detected by ROSAT. Obviously, XMM-Newton which is more sensitive than ROSAT should have detected all ROSAT sources. This is not the case, as 1 ROSAT core source (source R20) is not present in the XMM-Newton image (it is however detected by Chandra).