[Paper Review] The XMM-Newton serendipitous survey. VII. The third XMM-Newton serendipitous source catalogue

S. R. Rosen, N. A. Webb+38 more|UCrea (University of Cantabria)|Apr 27, 2015

Astrophysical Phenomena and ObservationsPhysics and Astronomy37 references167 citations

TL;DR

This paper presents the 3XMM-DR5 catalogue, the largest X-ray source catalogue ever produced, based on XMM-Newton observations through 2013. It includes 565,962 X-ray detections from 396,910 unique sources, with improved source detection, astrometry, and spectral/lightcurve extraction, enabling high-precision cross-identification and the discovery of rare or extreme objects across 877 square degrees of sky.

ABSTRACT

Thanks to the large collecting area (3 x ~1500 cm$^2$ at 1.5 keV) and wide field of view (30' across in full field mode) of the X-ray cameras on board the European Space Agency X-ray observatory XMM-Newton, each individual pointing can result in the detection of hundreds of X-ray sources, most of which are newly discovered. Recently, many improvements in the XMM-Newton data reduction algorithms have been made. These include enhanced source characterisation and reduced spurious source detections, refined astrometric precision, greater net sensitivity and the extraction of spectra and time series for fainter sources, with better signal-to-noise. Further, almost 50\% more observations are in the public domain compared to 2XMMi-DR3, allowing the XMM-Newton Survey Science Centre (XMM-SSC) to produce a much larger and better quality X-ray source catalogue. The XMM-SSC has developed a pipeline to reduce the XMM-Newton data automatically and using improved calibration a new catalogue version has been produced from XMM-Newton data made public by 2013 Dec. 31 (13 years of data). Manual screening ensures the highest data quality. This catalogue is known as 3XMM. In the latest release, 3XMM-DR5, there are 565962 X-ray detections comprising 396910 unique X-ray sources. For the 133000 brightest sources, spectra and lightcurves are provided. For all detections, the positions on the sky, a measure of the quality of the detection, and an evaluation of the X-ray variability is provided, along with the fluxes and count rates in 7 X-ray energy bands, the total 0.2-12 keV band counts, and four hardness ratios. To identify the detections, a cross correlation with 228 catalogues is also provided for each X-ray detection. 3XMM-DR5 is the largest X-ray source catalogue ever produced. Thanks to the large array of data products, it is an excellent resource in which to find new and extreme objects.

Motivation & Objective

• To produce a comprehensive, high-fidelity X-ray source catalogue from the full XMM-Newton serendipitous survey data up to 2013.
• To improve source detection and astrometric accuracy through enhanced data reduction pipelines and updated calibration.
• To provide spectral and lightcurve products for the brightest 133,000 sources to enable detailed source characterization.
• To enable robust cross-identification of X-ray sources with counterparts in multi-wavelength catalogues across 228 external databases.
• To assess the reliability of source positions and errors, particularly in the outer regions of the XMM-Newton field of view, where PSF elongation may affect accuracy.

Proposed method

• Utilized an automated, improved XMM-Newton data reduction pipeline developed by the XMM-Newton Survey Science Centre, incorporating updated calibration and source detection algorithms.
• Performed manual screening of all X-ray detections to ensure data quality and minimize spurious sources.
• Extracted source positions, count rates, fluxes in seven energy bands (0.2–12 keV), hardness ratios, and variability metrics for each detection.
• Generated spectra and lightcurves for the 133,000 brightest sources using enhanced signal-to-noise techniques.
• Conducted cross-identification with 228 external multi-wavelength catalogues, including SDSS and Chandra, to link X-ray sources to their counterparts.
• Assessed positional reliability by comparing XMM-Newton source positions with SDSS quasar counterparts and Chandra detections, using error-normalized offsets (x = Δr / σ_total) and examining dependence on off-axis angle.

Experimental results

Research questions

• RQ1 What is the overall size and dynamic range of the XMM-Newton serendipitous source catalogue after 13 years of observations and improved data reduction?
• RQ2 How do the positional uncertainties and detection reliability of XMM-Newton sources vary with off-axis angle, particularly in the outer regions of the field of view?
• RQ3 To what extent do error-normalized offsets between XMM-Newton detections and their SDSS quasar counterparts indicate positional errors or underestimated uncertainties?
• RQ4 Can systematic errors in source positions be attributed to flaws in the rectification process or PSF modeling at large off-axis angles?
• RQ5 How does the inclusion of elliptical error contours improve the assessment of positional uncertainty compared to circularized error estimates?

Key findings

• The 3XMM-DR5 catalogue contains 565,962 X-ray detections from 396,910 unique sources, making it the largest X-ray source catalogue ever produced.
• The median flux of detected sources is approximately 2.4 × 10⁻¹⁴ erg cm⁻² s⁻¹ in the 0.2–12 keV band, with observations spanning 13 years and covering 877 square degrees of sky.
• A notable excess of XMM-Newton detections with error-normalized offsets >3.5 from their SDSS quasar counterparts is observed, particularly at higher off-axis angles (θ > 10′), suggesting potential positional inaccuracies or underestimated errors.
• The excess in large-offset detections is most pronounced in the 10′ ≤ θ < 15′ annulus, where 5.6% of XMM-quasar pairs have x > 3.5, indicating a possible dependence on field-of-view location.
• Even when using elliptical error contours derived from RA and DEC errors to better model the elongated PSF at large off-axis angles, the excess in large offsets persists, indicating no systematic error in the rectification or error modeling process.
• Sources near the field center show slightly overestimated position errors, while those at larger off-axis angles may have underestimated errors or incorrect positions, though no consistent systematic cause was identified.