OM Catalogue Draft - XMM-Newton

DRAFT -- DRAFT -- DRAFT

XMM-OM Serendipitous Ultraviolet Source Survey catalogue: SUSS6.0

XMM-OM-SUSS6.0 is a new 2023 release of the XMM OM Serendipitous Ultraviolet Source Survey catalogue, superseding the previous version, XMM-OM-SUSS5.0.

The catalogue was compiled for the Optical Monitor instrument on board the ESA XMM Newton observatory. This version, the 6^th release of the catalogue, spans the period of observations from XMM-Newton revolution 34 (February 2000) to revolution 4207 (November 2022). A refereed paper on the first version can be found in Page M.J. et al. (2012).

In addition to covering a larger observation period, this sixth release reflects a change in philosophy with regard to the origin of the incorpoated data. In previous releases, the data were generated via a bespoke processing of the OM Observation Data Files (ODFs) while in this new release, we have been guided by the XMM user community and we have sought to harmonize the contents of the catalogue with those of the OM data in the XMM-Newton Science Archive (XSA), which derive from the standard XMM-Newton pipeline processing system. While the bespoke processing and pipeline systems are fundamentally very similar, they are not identical and the differences lead to some differences in the output.

All data have been fully processed at the European Space Astronomy Centre (ESAC, Madrid, Spain) using the XMM-Newton pipeline processing system (PPS) that invokes the XMM Science Analysis Software (SAS) - the vast majority of observations having been processed using SAS versions 18 and 19.

The number of observations (OBSIDs) included in the catalogue is 12057. The total number of entries (detections) in this release is 9,920,390. They correspond to 6,659,554 unique sources, of which 1,225,117 have multiple entries in the source table, corresponding to different observations.

The Catalogue is a FITS file consisting of two tables (please note that users wishing to use the TOPCAT tool to manipulate the catalogue, need to select the table format option, "FITS" to load the catalogue). The first table (SRCLIST) contains the source detections, with positional and photometric data (count rate, magnitude and flux) and quality flags for each measurement. The contents of the 114 columns are shown below:

In the second table (SUMMARY) we give a summary of the observations from which the sources have been detected and measured. The contents of the 35 columns in this table are as follows:

The meaning of the quality flags (columns filter_QUALITY_FLAG and filter_QUALITY_FLAG_ST) in the SRCLIST table are as follows:

^* - see known issues

The filter_SKY_IMAGE columns indicate if a source has been detected in stacked (sky) images obtained with that filter. The catalogue contains 1,422,957 such detections.

A cross-match of the whole catalogue with the GAIA EDR3 data set, using a search radius of 3 arcsecs, gives 80.4% matches (73.2% within a matching radius of 1 acrsecs). This increases to 81.4% within 3 arcsecs if we consider only SUSS6 detections with the cleanest (0) quality flag. The mean angular offset of all matched OM-GAIA sources is 0.50 arcsecs. 90% of the matches lie within 0.96 arcsecs and 95% have separations < 1.21 arcsecs. Detections in the V filter have 88.9% matches to GAIA within 3 arcsecs while corresponding values for the UVW1 and UVM2 filters are 81.9% and 73.7%, respectively.

Comparing the observations from the new SUSS6 catalogue that are common to the previous SUSS5 catalogue, we find, 92.8% of the SUSS6 detections have a match in SUSS5, within 1 arcsec (94.0% within 2 arcsecs). Considering V band detections, this becomes 94.4%, and 90% of observations have > 83% of SUSS6 detections with a SUSS5 match within 1 arcsec. The mean systematic difference in measured count rates between matched detection, in all filters, is less than 1% while in all filters, around 90% of matched detections have count rates in agreement to < 5% - one must bear in mind that there are differences in calibration (principally the time-dependent sensitivity degradation) between data used in SUSS6 and in SUSS5.

The XSA (https://www.cosmos.esa.int/web/xmm-newton/xsa) provides access to the OM Catalogue as a complete download of the file (~1.48 Gb, compressed fits format) as well as through a search interface whose results can be stored by the user in several formats. It should be noted that this interface allows the user to access only part of the information contained in the catalogue: source coordinates, magnitude and flux and their corresponding errors and quality flags. A reduced version of the complete catalogue (708 Mb, compressed fits format) can be downloaded as well.

The ESA-Sky utility allows the user to display the sources from the catalogue on top of sky images from different instruments at all wavelengths. This includes also images obtained with OM in two color combination, Visible (U, B and V OM filters) and UV (UVW1, UVM2 and UVW2 OM filters). It should be noted that OM images have a low dynamic range.

Known Issues and watchouts

A number of low-level issues, outlined here, affect the XMM-OM-SUSS6 catalogue. Fixes and improvements to these will be made in future catalogue releases, where possible.

• Flagging of detections is not foolproof. Problems can arise in crowded regions where sources may be contaminated by nearby objects but not always be flagged (observations containing more than 10000 detections are excluded from the SUSS6 catalogue). Another cause is fields that contain read-out streaks caused by bright stars that lie outside the window - in these cases, sources detected along a read-out streak may have compromised parameters or even be spurious but not be flagged accordingly. Similarly, sources detected on smoke rings (out-of-focus ghost images of bright stars), are often flagged but where the bright star causing the smoke ring lies outside the window where the smoke ring falls, they are generally not flagged. Likewise issues can occur in scattered light features caused by reflections, within the detector, of light from bright stars outside the OM field of view, which can form complex, extended, generally low-intensity features in the images where spurious detections may be found.
• In the case of very bright sources, the photometric data is nullified because the applied corrections for coincidence loss are high and not reliable. However, other quantities remain in the table, such as the major and minor axes, which may appear confusing to users.
• A small number of detections are erroneously associated with another detection from the same observation, and are asigned the same unique source number. These arise when detections of the same source in multiple exposures of the same observation, have slightly different positions. This can happen, for example, because of systematics in the rectification process (when a mean adjustment in RA and DEC is made to all sky positions in an exposure so as to minimise the spatial offsets when matched against an astrometric reference catalogue) or where centroiding is biased due to source counts in the source wings being outside the window, for sources very near window edges. These multiple detections can be divided into more than one unique source within the observation but be erroneously assigned the same unique source number in the catalogue. Although, in principle, the separate unique sources may, in reality, belong to the same object on the sky, if identified as separate unique sources (correctly or not), they should not acquire the same unique source number within an observation, but as noted, this does happen in a small number of cases.
• Following the accidental observation of Jupiter by the OM in July 2017, a small region of the photocathode, near the centre of the OM field of view, has reduced sensitivity - this is referred to as the Jupiter patch (JP). Much of this region is currently identified via 'bad pixels' in the bad pixel calibration table, and sources that are detected within the bad pixel region are flagged as such in the catalogue. However, the region of reduced sensitivity is larger than that currently translated into the bad pixel map and sources in the shallow wings of the JP, beyond the mapped bad pixel area, are not flagged. In fact, the wings of the JP include the location of the pn boresight in the OM image (i.e. where sources (generally, XMM-Newton targets), placed at the pn boresight, appear within the OM field of view). The impact of the sensitivity loss at the pn boresight is filter-dependent but ranges from about 8% in the V filter, to around 3% in the UV filters (see CAL-TN-0019.pdf) at the epoch of the Jupiter observation. The degradation at the pn boresight has since continued to decline, by about 1% per year in the V band.
• Currently, detections that lie within the bad-pixel-marked area of the Jupiter Patch are identified via flag 11. Previously, this flag was erroneously combined with flagging of very bright sources (these bright sources are indicated by flag 12). Nevertheless, there are some sources that that are affected by a subset of OM bad pixels within the OM bad pixel table which also give rise to the same flag value as sources in the Jupiter patch.
• In around 1% of entries for detections, the seconds part of the RA component of the IAUNAME is erroneously rounded up to 60.0, or the arcseconds part of the Declination component is rounded up to 60.
• Cross analysis with the previous SUSS5 catalogue, for the common set of observations, shows that some 92.8% of SUSS6 detections have a match in the SUSS5 catalogue within 1 arcsec. Equivalently, 7% are unmatched to a SUSS5 detection within 1 arcsec. In reverse, about 5% of SUSS5 detections are not matched with detections in SUSS6. One reason for this is a difference, unrecognized at the time of SUSS6 creation, in the way the bespoke pipeline (used to create SUSS5) and the routine XMM pipeline (used to generate this new SUSS6 catalogue), process OM data for a given filter when part of the field is only covered by one exposure in that filter. Both pipelines source-search individual exposure-level detector images, and also stacked sky-aligned images in a given filter when multiple overlapping exposures are present, to find fainter sources than in the per-exposure data - these fainter sources are added to the list of sources detected in the individual exposures. However, OM processing in the routine XMM pipeline also source-searches a sky image made from a single exposure in a given filter (erroneously, given that it is the same data, just rotated to sky coordinates), ensuring there are always the same number, or more detections than from the individual exposures. Additionally, differences may arise due to (i) differences in calibration (the routine pipleine processing has used different SAS versions and calibration files while the bespoke processing uses a single SAS version and a fixed set of calibration files) and (2) randomisation effects at the pixel scale, involved in the image creation.