CHANGBSCAT - Chandra Galactic Bulge Survey Full X-Ray Point Source Catalog
Among the goals of the GBS are constraining the neutron star (NS) equation of state and the black hole (BH) mass distribution via the identification of eclipsing NS and BH low-mass X-ray binaries (LMXBs). The latter goal will, in addition, be obtained by significantly enlarging the number of BH systems for which a BH mass can be derived. Further goals include constraining X-ray binary formation scenarios, in particular the common envelope phase and the occurrence of kicks, via source-type number counts and an investigation of the spatial distribution of X-ray binaries, respectively.
The GBS targets two strips of 6 degrees by 1 degrees (12 deg2 in total), one above (1o < b < 2o) and the other below (-2o < b < -1o) the Galactic plane in the direction of the Galactic center at X-ray, optical and near-infrared wavelengths. By avoiding the Galactic plane (-1o < b < 1o) the authors limit the influence of extinction on the X-ray and optical emission but still sample relatively large number densities of sources. The survey is designed such that a large fraction of the X-ray sources can be identified from their optical spectra. The X-ray survey, by design, covers a large area on the sky while the depth is shallow, using 2 ks per Chandra pointing. In this way, the authors maximize the predicted number ratio of (quiescent) LMXBs to cataclysmic variables. The survey is approximately homogeneous in depth to a 0.5-10 keV flux of 7.7 x 10-14 erg cm-2 s-1.
As of Paper I, the authors had covered about three-fourths (8.3 deg2) of the projected survey area with Chandra observations providing 1234 unique X-ray sources. In Paper II, the authors find 424 additional X-ray sources in the 63 Chandra observations that they report on there. In the papers, the authors discuss the characteristics and the X-ray variability of the brightest of the sources as well as the radio properties from existing radio surveys. They point out an interesting asymmetry in the number of X-ray sources as a function of their Galactic l and b coordinates which is probably caused by differences in average extinction towards the different parts of the GBS survey area.
The Galactic Bulge Survey: completion of the X-ray survey observations Jonker P.G., Torres M.A.P., Hynes R.I., Maccarone T.J., Steeghs D., Greiss S., Britt C.T., Wu J., Johnson C.B., Nelemans G., Heinke C. <Astrophys. J. Suppl. Ser., 210, 18 (2014)> =2014ApJS..210...18J (arXiv:1312.3192) = Paper II The Galactic Bulge Survey: Outline and X-ray Observations Jonker P.G., Bassa C. G., Nelemans G., Steeghs D., Torres M.A.P., Maccarone T.J., Hynes R.I., Greiss S., Clem J., Dieball A., Mikles V.J., Britt C.T., Gossen L., Collazzi A.C., Wijnands R., In't Zand J.J.M., Mendez M., Rea N., Kuulkers E., Ratti E.M., van Haaften L.M., Heinke C., Ozel F., Groot P.J., Verbunt F. <Astrophys. J. Suppl., 194, 18 (2011)> =2011ApJS..194...18J = Paper I
For Paper I sources (sources with source numbers from 1 to 1234), the authors reprocessed and analyzed the data using the CIAO 4.2 software developed by the Chandra X-ray Center and employing CALDB version 4.3. The data telemetry mode was set to very faint for all observations except that with ID 8687 since there was a bright source present in archival ROSAT observations of this region. For that observation, the authors used the standard 'faint' mode. The authors also reprocessed and reanalyzed the data that were obtained as part of the Bulge Latitude Survey (Grindlay et al. 2005, ApJ, 635, 920; Hong et al. 2009, ApJ, 706, 223; Chandra observation IDs 7160-7162, 7166-7168, 8199-8204, 9562-9564) that falls in the area they targeted in the GBS using exactly the same reduction and analysis as for their own GBS observations. Because the original Bulge Latitude Survey Chandra observations had an exposure time of approximately 15 ks, the authors selected 2 ks segments to allow for a comparison with their GBS observations. They selected data stretches of 2 ks length taking the start time of the observation from the header of the data, plus 100 s as the starting point of the 2 ks stretches. These data were also obtained using the 'very faint' mode.
For Paper II sources (sources with source numbers from B1 to B424), the authors reprocessed and analyzed the data using the CIAO 4.3 software developed by the Chandra X-ray Center and employing CALDB version 4.4.6. The data telemetry mode was set to very faint for all observations The 'very faint' mode provides 5 x 5 pixel information per X-ray event. This allows for a better screening of events caused by cosmic rays.
In their analysis, the authors selected events only if their energy falls in the 0.3-8 keV range. They used wavdetect to search for X-ray sources in each of the observations using data covering the full 0.3-8, 0.3-2.5, and 2.5-8 keV energy bands separately. They set the sigthresh in wavdetect to 1 x 10-7, which implies that, for a background count rate constant over the ACIS-I CCDs, there would be 0.1 spurious source detection per observation as about 1 x 106 pixels were searched per observation. As in most cases, a source is not detected in a single pixel, this estimate is very conservative. Furthermore, as explained below, some additional selection criteria were applied, further lowering the number of spurious sources.
All sources for which Poisson statistics indicated that the probability of obtaining the number of detected source counts by chance, given the expectation for the local background count rate is lower than 1 x 10-6 were retained. This would be equivalent to a > 5-sigma source detection in Gaussian statistics. Next, the authors deleted all sources for which wavdetect was not able to provide an estimate of the uncertainty on the Right Ascension or on the Declination, as this often indicates that all counts fell in a single pixel which could well have been due to faint afterglow events caused by cosmic ray hits. In addition, the authors imposed a three count minimum for source detection.
Since the Chandra GBS observations were designed to overlap near the edges, the authors searched for multiple detections of the same source either in one of the energy sub-bands or in the full energy band. They considered sources with positions falling within 3" (Paper I) or 5" (Paper II) of each other to be likely multiple detections of the same source. The 5" radius for Paper II sources is larger than the of 3" for Paper I sources as the authors found out that some multiple detections of the same source still remained for sources detected with large off-axis angles: see Hynes et al. 2012, ApJ, 761, 162 for a discussion of the 18 sources from Paper I (source numbers = 136, 153, 191, 225, 230, 233, 273, 285, 303, 323, 382, 384, 464, 483, 515, 571, 635 and 851) that were in fact multiple detections of the same source. This means that in Paper I, the authors actually found 1216 unique sources. The HEASARC has added a parameter to its version of the GBS source catalog called source_flag which is set to 'D' for these 18 duplicates.
In the last quarter of the GBS area reported in Paper II, the authors found that 26 sources were detected more than once. Out of these 26 sources, 23 sources were detected two times, and 3 sources were detected three times. Two of the sources detected twice were already detected and reported in Paper I (155 and 314). The properties that are listed for the sources that were detected multiple times are those for the detection that gave rise to the largest number of X-ray counts. In this combined table, the authors list the number of times that sources were detected in different Chandra observations, except that (the HEASARC believes) the number for Paper I sources does not include re-detections in the later observations first reported in Paper II. Besides the multiple detections of 155 and 314, 14 additional sources detected once in the cycle 13 Chandra observations were previously detected and listed in Paper I. These sources are 15, 17, 25, 44,60, 69, 79, 137, 221, 266, 312, 355,374, and 439. In most cases the off-axis angle of the source position was larger during the new observations and,given that a similar number of X-ray counts was detected in each instance, the source position provided in Paper I is the most accurate X-ray position available. The main exception where the authors consider the newly derived position to be more accurate is source 314. It was detected at 10.8 arcminutes off-axis with 8 counts in the Chandra detection leading to its discovery. The new detection in Paper II was 5.9 arcminutes off-axis in ObsID 13581. The new best-fit source position is (RA, Dec) = (266.6461515, -31.8136964) (J2000) which is 2.6 arcseconds from the previously reported position (the latter is retained as the position of this source in the present table, notice). The HEASARC has added a value of 'P' for the source_flag parameter for source_number = 314 to alert users to this more accurate position.
A unique sequential identification number for each source in this catalog in order of decreasing source counts separated into the two published source lists. Paper I sources have no letter prefix, e.g., source number 1 has the largest number of counts of any source reported in Paper I, whereas Paper II sources have the letter prefix of 'B', e.g., source number B1 has the largest number of counts of any source reported in Paper II. In the two papers, the authors use an initial prefix of 'CX' for these sources: thus source number B1 is referred to as source 'CXB1' therein.
This HEASARC-created flag parameter is set to 'D' for the 18 sources from Paper I (source numbers = 136, 153, 191, 225, 230, 233, 273, 285, 303, 323, 382, 384, 464, 483, 515, 571, 635 and 851) that were in fact multiple detections of another source, according to Hynes et al. (2012, ApJ, 761, 162). This means that in Paper I, the authors actually found 1216 unique sources. The HEASARC has added a parameter to its version of the GBS source catalog called source_flag which is set to 'D' for these 18 duplicates:
136 is a duplicate of 61 153 is a duplicate of 93 191 is a duplicate of 360 225 is a duplicate of 38 230 is a duplicate of 33 233 is a duplicate of 345 273 is a duplicate of 597 285 is a duplicate of 394 303 is a duplicate of 304 323 is a duplicate of 292 382 is a duplicate of 280 384 is a duplicate of 177 464 is a duplicate of 175 483 is a duplicate of 1155 515 is a duplicate of 16 571 is a duplicate of 270 635 is a duplicate of 636 851 is a duplicate of 646
This flag is set to 'P' for source number 314 to indicate that a more accurate position than the one listed in the table based on the observations analyzed in Paper I is listed in Section 2.1 of Paper II (see also the Overview above).
The source designation registered with the Dictionary of Nomenclature of Celestial Objects, using the prefix of 'CXOGBS' for Chandra X-Ray Observatory Galactic Bulge Survey and the J2000 equatorial coordinates.
The Right Ascension of the X-ray source in the selected equinox. This was given in J2000 decimal degree coordinates to a precision of 10-7 degrees in the original reference.
The Declination of the X-ray source in the selected equinox. This was given in J2000 decimal degree coordinates to a precision of 10-7 degrees in the original reference.
The Galactic Longitude of the X-ray source.
The Galactic Latitude of the X-ray source.
The 3-sigma error in the Right Ascension of the X-ray source position, in arcseconds, as provided by wavdetect. It does not take into account the typical Chandra boresight uncertainty of 0.6 arcseconds (90% confidence).
The 3-sigma error in the Declination of the X-ray source position, in arcseconds, as provided by wavdetect. It does not take into account the typical Chandra boresight uncertainty of 0.6 arcseconds (90% confidence).
The 95%-confidence total positional uncertainty, in arcseconds, taking the boresight error into account and based on the observed counts and off-axis angle and using the formula (4) of Evans et al. (2010, ApJS, 189, 37) which is also given in section 2.1 of Paper II.
The total number of counts N detected in the X-ray source, in the full 0.3-8 keV energy band. The uncertainty in the counts can be estimated using the Gehrels relation: sigma(N) = 1 + sqrt (N + 0.75), cf., Grimm et al. (2005, ApJS, 161, 271).
The Chandra observation identification number of the observation in which the X-ray source was detected.
The off-axis angle of the X-ray source from the pointing center, in arcminutes.
The number of individual Chandra pointings in which the X-ray source was detected. For Paper I sources, this number does not include the number of re-detections in the later observations reported in Paper II.
The hardness ratio of the X-ray source, defined as the ratio between the count rate in the 2.5-8 keV band minus that in the 0.3-2.5 keV band to the count rate in the full 0.3-8 keV energy band (after Kim et al. 2004, ApJS, 150, 19). The hardness ratio is calculated for the detection where the off-axis angle was smallest, if the source was detected multiple times. This quantity is given only for sources for which more than 20 counts were detected and which did not suffer from pile-up, i.e., not for the 3 brightest Paper I sources (1, 2 and 3) or any of the sources with 19 or fewer counts. The extreme hardness ratio values for the sources 67, 77, and 72 are caused by the fact that the detections of the sources that were most on-axis had very low count rates. Given that the detection most on-axis was the one used for the calculation of the hardness ratio, the hardness ratio of these sources was done using a few counts only and as a result it carries a large uncertainty.
The uncertainty in the hardness ratio of the X-ray source