XMMSLEWCLN - XMM-Newton Slew Survey Clean Source Catalog, v1.6
The first XMM-Newton slew survey catalogue: XMMSL1. Saxton R.D., Read A.M., Esquej P., Freyberg M.J., Altieri B., Bermejo D. <Astron. Astrophys., 480, 611-622 (2008)> =2008A&A...480..611S (SIMBAD/NED BibCode)
Band Energy (keV) Description 0 0.2-12 Total 4 2-12 Hard 5 0.2-2 Softto form three distinct surveys which have been subsequently combined to form one catalog.
There are two related catalogs: a full catalog (the HEASARC Browse table XMMSLEWFUL constructed from the file xmmsl1D6_total.fits.gz)), containing 41,423 detections of sources found with a likelihood of DET_ML > 8, and a "clean" catalog (the HEASARC Browse table XMMSLEWCLN constructed from the file xmmsl1D6_clean.fits.gz) containing 20,163 sources, where all known bad sources have been removed and where the detection limit has been raised to DET_ML > 10 in general and DET_ML > 14 for sources found in images with a higher than usual background. Efforts have been made to identify spurious detections and 3,494 of these have been flagged as such in the full catalog.
There is an overlap between the slew paths which has led to > 951 sources being observed two or more times. The full catalog has an expected spurious fraction of about 20%, determined from simulations of blank fields. A cleaner subset has been created, with sources of likelihood DET_ML > 10 taken from images where the background count rate is <= 3 ct/s and DET_ML > 14 for other fields. This selection is expected to contain < 1% of false sources in the soft band, ~ 9% in the hard band and ~ 4% in the total band. The number of sources (after removal of flagged sources) and the sky areas for the two catalogs are as follows:
Full catalog, sky area = 47,350 deg2:
Band Number of Sources 0 30,808 4 6,306 5 20,136 ALL 37,929
Clean catalog, sky area = 35,350 deg2 down to DET_ML > 10 and 12,000 deg2 down to DET_ML > 14:
Band Number of Sources 0 17,345 4 2,160 5 14,371 ALL 20,163
The median flux is 3.0E-12 ergs/s/cm2 in the total band, 1.1E-12 in the soft band and 9.2E-12 in the hard band in the clean catalog. All sources have been cross-correlated with existing astronomical catalogs and an identification assigned for each source where found. These have been included together with the source category, e.g. Galaxy, flare star, etc., as columns in the catalog together with any ROSAT All Sky Survey (RASS) counterpart. A list of IDs will be maintained. The quoted positions in the slew catalog have a 1-sigma error of 8 arcseconds which for point sources is dominated by the accuracy of the attitude reconstruction during the slewing maneuvers. The statistical position error quoted in the catalog has a mean of ~4 arcseconds (1 image pixel) for non-extended sources. This is usually much larger for extended sources. After correcting for overlaps, 65% of the sky has now been covered.
Detections in the three energy bands have been combined using a match radius of 30 arcseconds to identify sources detected in more than one band. The match radius is approximately 4 sigma and is reasonable for point sources because of the great scarcity of sources (0.8 detections per square degree on average) at the flux levels probed by the slew survey.
Spurious detections are produced due to systematic effects with the instruments and detection software. These sources have been screened and flagged ('T' for 'true' or 'F' for 'false') in the different categories described below. When a source is flagged true, the FLAG_COMMENT column gives a more detailed explanation of the problem for each particular source.
All sources detected in the survey have been correlated with different catalogs in order to identify the XMM-Newton slew sources with previously known objects. The catalogs used for this aim comprises two astronomical databases, a catalog of clusters of galaxies and nine other catalogs (some of them have been queried through the HEASARC astronomical database). Although the astrometric uncertainty of the slew sources was found to be 8 arcsec, the offset radius for the correlations was 30 arcsec (with a few exceptions described below) in order to include sources from catalogs with worse accuracy or truncated coordinates. For the EXOSAT CMA catalog the offset radius was 45 arcsec, and for the Einstein IPC it was 2 arcmin, both due to their high uncertainty in source coordinates. A radius of 5 arcmin was chosen for the clusters catalog due to the extension of this type of objects. As not all catalogs have the same astrometric accuracy distances are rounded to the nearest 0.1 arcmin in many cases. The identification process results in unidentified sources, sources with a single counterpart and sources with multiple matches. A hierarchical selection scheme has been applied for sources with different counterparts. A decision has been made to derive the most plausible identification candidate using the technique described below. Firstly, the SIMBAD and NED astronomical databases have been used for the cross-correlation and results from both databases have been compared in detail. (SIMBAD has been queried in the frame of the Astronomical Virtual Observatory.) When a source has the same counterpart in both databases the one selected for the identification is that which gives more specification about the source category. When contradictory identification candidates have been found, the closest one has been chosen. These two databases provide the large majority (90%) of the total number of identifications. Then, a correlation with a clusters table (Abell and Zwicky Clusters of Galaxies ) has been performed. The final identification for those extended sources (ext > 2.0) with a SIMBAD/NED and also a cluster correlation comes from the clusters table. The rest of the catalogs used for the cross-match are listed below ordered in priority for the preferred identification. For sources with multiple matches in a catalog, the identification was selected as the closest match. These catalogs are: All-Sky Optical Catalog of Radio/X-Ray Sources, Catalog of PSPC WGA Sources, Einstein IPC Sources Catalog, EXOSAT CMA Images/Lightcurves, ROSAT All-Sky Survey Catalog, ROSAT Results Archive Sources for the PSPC, ROSAT Results Archive Sources for the HRI, RXTE Master Catalog, XMM-Newton Serendipitous Source Catalog, Version 1.1.0, INTEGRAL Bright Source Catalog. Identifications are offered for 73% of the sources, while roughly 50% of the clean catalog slew sources have a ROSAT counterpart.
Events are recorded initially in RAW or detector coordinates and have to be transformed, using the satellite attitude history, into sky coordinates. The tangential plane geometry commonly used to define a coordinate grid for flat images is only valid for distances of 1-2 degrees from a reference position, usually placed at the center of the image. To avoid this limitation, slew datasets were divided into roughly one square degree event files, attitude corrected and then converted into images.
Source searching used a semi-standard `eboxdetect (local) + esplinemap + eboxdetect (map) + emldetect' method, tuned to about zero background, and performed on a single image containing just the single events (pattern=0) in the 0.2-0.5 keV band, plus single and double events (pattern=0-4) in the 0.5-12.0 keV band.
The attitude information of the XMM-Newton satellite is provided by the Attitude and Orbit Control Subsystem (AOCS). A star tracker co-aligned with the telescopes allows up to a maximum of five stars to be continuously tracked giving accurate star position data every 0.5 seconds, which operates in addition to the Sun sensor that provides a precise Sun-line determination. Such information is processed resulting in an absolute accuracy of the reconstructed astrometry of typically 1 arcsecond during pointed observations. For the open-loop slews, large slews outside the star-tracker field of view of 3 x 4 degrees, the on-board software generates a three axis momentum reference profile and a two-axis (roll and pitch) Sun-sensor profile, both based on the ground slew telecommanding. During slew maneuvering a momentum correction is superimposed onto the reference momentum profile and, as there are no absolute measurements for the yaw axis, a residual yaw attitude error exists at the end of each slew that may be corrected in the final closed-loop slew. To process slew data, attitude information is taken from a Raw Attitude File (RA F) which stores attitude points every 40-60 seconds. The times quoted in the RAF are offset from spacecraft time by 0.75 seconds, which has to be corrected in the software, otherwise a 1 arcminute offset in source positions is seen along the slew direction. Initially the correction was performed in the slew-specific software but from SAS 7.0 has been included in the general SAS software. The SAS astrometry software interpolates between the rather sparse RAF points using a curve fit. Five RAFs have been found to contain one bad attitude point which is sufficient to totally corrupt the attitude solution over a considerable section of the slew. In these cases the affected sources have been removed from the CLEAN catalog and the errors in their positions have been set to 120 arcseconds. In a further 7 slews a part or all of the attitude seems to be turbulent rather than smooth. It is not known why this occurs and no attempt to fix it has been made. Sources affected by this attitude problem have a position suspect (VER_PSUSP) flag set to true. The error introduced in the source positions can be of the order of 1 arcminute. A sample of point sources from the catalog with matches in the SIMBAD catalog have been analyzed and show that 68% lie within 8 arcseconds of the SIMBAD position and 90% within 17 arcseconds.
1. Assess the effect of the number of photons in an image on the number of spurious detections as a function of detection likelihood.
An average sized slew image was taken and a number of photons inserted at random positions. The number of false sources produced is a strong function of the number of photons in the image, with the peak appearing with a coverage of a few hundred photons over this size image. It is suspected that the peak in the curves is due to peculiarities of the background estimation technique which fits a spline-surface to the image. The majority of the slew images contain about 100 counts where the number of spurious sources produced by the background is small.
2. Equate the number of counts in an image with the measured background rate.
The background rate in XMM images is traditionally measured by finding the count rate above 10 keV (PI > 10000), where the contribution from celestial X-ray sources is expected to be very small. We see a strong correlation between this background estimator and the total number of source photons.
3. Estimate the total number of spurious sources in each energy band.
There were 11449 images which have been searched for sources to form the original XMMSL1 catalogue. For each of these images, in each energy band, the position of the photons has been randomly reset and the images searched again using the same software. The table below gives, for different values of the minimum acceptable detection likelihood, false source numbers which are compared with the total number of sources detected in the catalogue. It can be seen that using a discriminator on the background count rate helps to reduce the fraction of spurious detections while maximizing the number of real sources. The optimum extraction for the hard band and total band catalogs uses a likelihood limit of 10 for sources found in low background images [(PI > 1000) <= 3 ct/s ] and a limit of 14 for sources detected in images with a higher background. This selection has been used to form the "clean" catalog which is the default selection within the XMM-Newton Science Archive (XSA). As can be seen from the table, this is quite conservative for the soft band catalog:
Band DET_ML>8 DET_ML>8 DET_ML>10 DET_ML>10, DET_ML>14, rate<3 rate<3 +(DML>10, rate<3) ALL 4730 (929) 3471 (456) 3015 (195) 2436 (106) 2713 (109) 0 3882 (580) 3037 (348) 2596 (118) 2171 (86) 2384 (89) 4 694 (272) 429 (93) 314 (61) 241 (24) 261 (25) 5 2619 (186) 1994 (69) 2042 (46) 1649 (13) 1888 (13)
Table: Numbers of sources in the original XMMSL1 catalog, and estimated false sources (in parentheses), for various selections of detection likelihood and background rate.
Band Energy (keV) Flux limit (10-12 ergs/s/cm2) Total 0.2-12 1.3 Hard 2-12 3.7 Soft 0.2-2 0.57
This is the official name for sources detected in the XMM-Newton slew survey. It starts with the prefix, XMMSL1, the IAU registered designator, and then encodes the J2000 sky position, e.g. XMMSL1 J010537.6+364858. The name is assigned in two passes. When the three independent energy band source lists are combined to form one catalog the source name is set using the position in the band where the DET_ML likelihood is the highest. A second pass is then performed such that sources which have been observed in more than one slew are given the same name. Again, priority is given depending on the detection likelihood. Note that the combination of XMMSLEW_NAME and observation number (OBSID) is unique. Detections are deemed to be from the same source if their centers lie within 30 arcseconds of each other. Note that the statistical position error (RADEC_ERROR) calculated by the source search software is not used in this calculation. This is because the mean error on the slew attitude reconstruction of 8" dominates the error budget. Given the scarcity of slew sources (0.8 per square degree) on the sky, 30" was found to be a reasonably robust match radius for point sources. It is not so good for extended sources and the catalog may contain multiple detections of the same extended source with different names. Please note that no attempt has been made to assign a unique name from the original XMMSL1 catalog to sources added in the subsequent delta increments.
This is the XMM-Newton observation number assigned to the slew by the satellite scheduling system. Slew observations always begin with a 9 to distinguish them from pointed observations, followed by a 4 digit satellite orbit number and a 5 digit slew designator, e.g. 9031400004, refers to the second slew of revolution 314.
This is a number which uniquely identifies each detection in an observation.
The Right Ascension of the source in the selected equinox. This was given in J2000.0 coordinates and to a precision of 8 significant digits in decimal degrees in the original table. In the event of a detection in more than one energy band, the position is taken from the band with the highest detection likelihood.
The Declination of the source in the selected equinox. This was given in J2000.0 coordinates and to a precision of 8 significant digits in decimal degrees in the original table. In the event of a detection in more than one energy band, the position is taken from the band with the highest detection likelihood.
The statistical error of the source position, in arcseconds, as returned by the source detection software.
The Galactic Longitude of the source.
The Galactic Latitude of the source.
The hardness ratio, defined as
HR1 = (COUNT_RATE_B4 - COUNT_RATE_B5) / (COUNT_RATE_B4 + COUNT_RATE_B5)where COUNT_RATE_B4 is the hard band count rate and COUNT_RATE_B5 is the soft band count rate. This has been calculated for all of the sources which have a positive detection in both bands.
The error in the hardness ratio, calculated as
sqrt(COUNT_RATE_B4_ERROR2 + COUNT_RATE_B5_ERROR2) / (COUNT_RATE_B4 + COUNT_RATE_B5)where COUNT_RATE_B4_ERROR and COUNT_RATE_B5_ERROR are the errors in the hard and soft band count rates respectively.
The start time of the slew containing this source.
The end time of the slew containing this source.
The number of background-subtracted counts, in the total energy band (0.2-12 keV). This number has been corrected for photons scattered outside the source region due to the Point Spread Function (PSF).
The statistical 1-sigma error in the total-band source counts.
The number of background-subtracted counts, in the hard energy band (2-12 keV), corrected for the PSF.
The statistical 1-sigma error in the hard-band source counts.
The number of background-subtracted counts, in the soft energy band (0.2-2 keV), corrected for the PSF.
The statistical 1-sigma error in the soft-band source counts.
The X position of the source in the total-band image in image pixels. Each pixel covers an area of 4.1" x 4.1".
The Y position of the source in the total-band image in image pixels. Each pixel covers an area of 4.1" x 4.1".
The X position of the source in the hard-band image in image pixels. Each pixel covers an area of 4.1" x 4.1".
The Y position of the source in the hard-band image in image pixels. Each pixel covers an area of 4.1" x 4.1".
The X position of the source in the soft-band image in image pixels. Each pixel covers an area of 4.1" x 4.1".
The Y position of the source in the soft-band image in image pixels. Each pixel covers an area of 4.1" x 4.1".
The spatial extension of the source in the total energy band, in pixels. This measures the deviation from a point source of the spatial distribution of the source counts. It is defined as the sigma of a Gaussian which would need to be convolved with the point spread function (PSF) to produce the observed counts distribution. The software (emldetect) fits sources out to a maximum extent of 20 pixels. Each pixel covers an area of 4.1" x 4.1".
The statistical 1-sigma error in the total-band spatial extension, in pixels.
The spatial extension of the source in the hard energy band, in pixels.
The statistical 1-sigma error in the hard-band spatial extension, in pixels.
The spatial extension of the source in the soft energy band, in pixels.
The statistical 1-sigma error in the soft-band spatial extension, in pixels.
The detection likelihood in the total energy band (0.2-12 keV).
The detection likelihood in the hard energy band (2-12 keV).
The detection likelihood in the soft energy band (0.2-2 keV).
The likelihood of the source being extended in the total band.
The likelihood of the source being extended in the hard band.
The likelihood of the source being extended in the soft band.
The background counts per pixel for the total band.
The background counts per pixel for the hard band.
The background counts per pixel for the soft band.
The effective on-axis exposure time in the total energy band, in seconds.
The effective on-axis exposure time in the hard energy band, in seconds.
The effective on-axis exposure time in the soft energy band, in seconds.
The source flux in the total energy band, in erg/cm2/s. Source fluxes have been calculated from count rates based on energy conversion factors assuming a spectral model of an absorbed power-law with NH = 3.0 * 1020 cm-2 and a slope = 1.7 (see XMM Science Survey Center memo SSC-LUX TN-0059 for a general description of the technique). The energy conversion factors used here to convert between the source count rate in ct/s to flux in these energy bands in units of 10-12 erg/s/cm2 are the following:
Band Conv. Factor Total 3.159 Hard 9.144 Soft 1.436
The error in the total-band source flux, in erg/cm2/s.
The source flux in the hard energy band, in erg/cm2/s. Source fluxes have been calculated from count rates based on energy conversion factors assuming a spectral model of an absorbed power-law with NH = 3.0 * 1020 cm-2 and a slope = 1.7 (see XMM Science Survey Center memo SSC-LUX TN-0059 for a general description of the technique). The energy conversion factors used here to convert between the source count rate in ct/s to flux in these energy bands in units of 10-12 erg/s/cm2 are the following:
Band Conv. Factor Total 3.159 Hard 9.144 Soft 1.436
The error in the hard-band source flux, in erg/cm2/s.
The source flux in the soft energy band, in erg/cm2/s. Source fluxes have been calculated from count rates based on energy conversion factors assuming a spectral model of an absorbed power-law with NH = 3.0 * 1020 cm-2 and a slope = 1.7 (see XMM Science Survey Center memo SSC-LUX TN-0059 for a general description of the technique). The energy conversion factors used here to convert between the source count rate in ct/s to flux in these energy bands in units of 10-12 erg/s/cm2 are the following:
Band Conv. Factor Total 3.159 Hard 9.144 Soft 1.436
The error in the soft-band source flux, in erg/cm2/s.
The total band count rate, in counts/second.
The error in the total band count rate.
The hard band count rate, in counts/second.
The error in the hard band count rate.
The soft band count rate, in counts/second.
The error in the soft band count rate.
The J2000 Right Ascension of the total-band source, in degrees.
The J2000 Declination of the total-band source, in degrees.
The statistical 1-sigma error in the total-band source position, in arcseconds.
The Galactic Longitude of the total-band source, in degrees.
The Galactic Latitude of the total-band source, in degrees.
The J2000 Right Ascension of the hard-band source, in degrees.
The J2000 Declination of the hard-band source, in degrees.
The statistical 1-sigma error in the hard-band source position, in arcseconds.
The Galactic Longitude of the hard-band source, in degrees.
The Galactic Latitude of the hard-band source, in degrees.
The J2000 Right Ascension of the soft-band source, in degrees.
The J2000 Declination of the soft-band source, in degrees.
The statistical 1-sigma error in the soft-band source position, in arcseconds.
The Galactic Longitude of the soft-band source, in degrees.
The Galactic Latitude of the soft-band source, in degrees.
This flag is set to 'CLEAN_SAMPLE' to indicate that the source is included in the clean subset, else it is set to 'XXXXXXXXXXXX'.
The source name defined in the total-band list. The format is 'xs' followed by the revolution number, the observation ID and the source position, e.g., 'xs0841_9084100002_12:57:07.5+01:50:42'.
The source name defined in the hard-band list. The format is 'xs' followed by the revolution number, the observation ID and the source position, e.g., 'xs0841_9084100002_12:57:07.5+01:50:42'.
The source name defined in the soft-band list. The format is 'xs' followed by the revolution number, the observation ID and the source position, e.g., 'xs0841_9084100002_12:57:07.5+01:50:42'.
The name of the image containing the total-band detection.
The name of the image containing the hard-band detection.
The name of the image containing the soft-band detection.
If set to 'T' (true), this flag notes that a detection has been found within an extended source and is probably spurious. This flag is used to ensure that only one source is quoted for the large supernova remnants.
This quality flag is used to ensure that only one source is quoted for extended objects; that with the highest det_ml value. Two different cases have been found: (1) Large extended sources (mainly SNR) that often result in multiple detections of the same object. These have been identified by searching for images with a large number of sources, (2) Other spurious sources (mainly due to Cluster of Galaxies) appear because the technique used for merging detections in the different energy bands is optimal for point-like sources. This consists of considering the same source if their centers lie within 30 arcsec of each other, but this offset has to be bigger for extended sources. This is expected to be fixed in a new version of the catalog.
If set to 'T' (true), this flag indicates that a detection lies within the point spread function, or halo, of a very bright source, and is probably spurious.
A halo of false detections is often seen around bright slew sources due to the imperfect modeling of the PSF (as was also seen in the creation of the 1XMM Serendipitous Source Catalogue). Due to the reduced exposure time in slew observations this is only important for very bright sources rate >> 10 ct/s. For flagging this category of spurious, the different sources detected in the same image and containing a very bright source have been inspected. If their centers lie within 30 arcsec but corresponding to the same object, the one with the highest det_ml value is taken as the non-spurious detection and the rest have been flagged 'T'.
If set to 'T' (true), this flag indicates that the source lies within a bright region, caused by high background, and is probably spurious.
One of the characteristics of slew exposures is their low background (on average ~ 0.1 ct/arcmin2). Although high background slews taken at times of enhanced solar activity have been rejected for the processing, some high background sub-images have been detected in observations with a general rate below the threshold used for rejecting entire slews. These sources have been identified and flagged in this category looking for images with a large number of sources (without being included in the VER_INEXT category of spurious sources explained above). It has been checked that no other bright off-axis source could generate the detections. Rates for all images have also been computed.
If set to 'T' (true), this flag indicates that the source has been detected in two consecutive images of the same slew. This can occur if the source lies on the border between slews and implies that the position of the source is unlikely to be accurate.
Some sources have been detected lying on the border of consecutive images of the same slew observation, so the position of both detections of the same object is unlikely to be accurate. These spurious sources have been identified by searching for sources in consecutive images whose centers lie within 30 arcsec.
This flag indicates that the quoted position of the source is likely to be inaccurate. In a few slews, the attitude reconstruction is poor and, in these cases, the actual position of the source is difficult to determine, and can be wrong by an arcminute.
This quality flag is related to the astrometry performance and arises after a careful visual examination of the RAF files. The attitude reconstruction of some slews appears to be "turbulent" rather than smooth, so source positions lying in these poorly attitude reconstructed regions are likely to be inaccurate.
This is a catch-all for problems not included in the other flags. In the catalog, only two sources have this flag set true; in both cases, it is because the exposure time has been calculated as zero for reasons which remain unclear.
This flag is related to spurious detections not included in the other flags. Two sources (both are close to an SNR) which have been discovered to lie outside the images they are supposed to belong to, and hence have a calculated exposure time of zero, have this flag set true. A further 27 detections, due to out of field-of-view reflections from Sco X-1 also have this flag set.
A comment which explains why a particular source flag has been set to true.
Cross-correlations of the positions of the slew sources with astronomical databases and catalogs have been performed. This parameter gives the catalog name of the best match.
An alternative designation for the best-matched source.
The name of the closest ROSAT All-Sky Survey (RASS) source to the position of the XMM-Newton Slew source.
The source type, as determined by SIMBAD, NED, and the other resources used in the cross-matching process. This is directly taken from the catalog in question, and no attempt has been made to rationalize the values.
The astronomical database or catalog from which the best match has been selected, e.g., SIMBAD, NED, etc.
The angular distance in arcminutes between the best match candidate and the XMM-Newton Slew Survey source. The resolution is set at 0.1 arcminutes because this is the resolution in which the values are available from NED, and also because the best match candidate often has a large error in its position.
The angular distance between the best matching ROSAT source and the XMM-Newton Slew Survey source, in arcseconds.
The observing mode of the EPIC-pn camera, where 'FF' means full-frame mode, 'eFF' extended full-frame mode, and 'LW' means large window mode.
The peak background count rate in the image from which the source was extracted, in counts/second. This is measured as the count rate, over the whole image, for events with energy > 10 keV (PI > 10000). It is quoted to a resolution of 0.1 ct/s and is used to determine the overall background environment in which a source was detected. This value is used in the creation of a clean sub-sample of the catalog (see Description above). Generally, the higher this value, the greater the probability that the source is spurious. This is especially true if the detection likelihood of the source is less than 14. The value has been tabulated for images with a background rate greater than 2 ct/s. For quieter images the value has been artificially set to 1.9 ct/s.
The HEASARC Browse object classification, based on the value of the ID_CATEGORY parameter.