Chandra Source Catalog — Release 2.1

Ian Evans and Rafael Martínez-Galarza for the Chandra Source Catalog team

Version 2.1 of the Chandra Source Catalog (CSC 2.1) was released to the community on April 2, 2024. By combining Chandra's sub-arcsecond on-axis spatial resolution and low instrumental background with consistent data processing, the catalog provides carefully curated, high-quality, and uniformly calibrated and analyzed tabulated parameters (including positional, spatial, photometric, spectral, and temporal source properties) as well as science-ready X-ray data products. The sensitivity limit for compact sources in CSC 2.1 is unchanged at ~5 net counts (on-axis) for most observations. CSC 2.1 is an incremental release that adds imaging observations released publicly from 2015 through 2021 to the previous CSC 2.0 catalog version, increasing the overall size of the catalog by about 30%. Besides adding more recent observations, CSC 2.1 includes several enhancements that provide significant improvements over the previous release. In particular, the catalog is tied to the Gaia-CRF3 realization of the International Celestial Reference Frame for the best sky positions for catalog sources.

The catalog includes measured and derived properties for 407,806 unique compact and extended X-ray sources in the sky, allowing both statistical analysis of large samples as well as detailed studies of individual sources. Extracted properties are provided for 1,304,376 individual observation detections (2,143,847 when including photometric upper limits) identified in 15,533 Chandra ACIS and HRC-I imaging observations released publicly through the end of 2021 (see Figure 1). Photometric properties for 1,717 highly extended (≳30″) sources are provided, together with surface brightness polygons for several observation-specific contour levels. The total cumulative sky coverage of CSC 2.1 is 730.37 deg², and multi-band limiting sensitivity is computed for the entire sky coverage of the catalog at a resolution of ~3.22 arcsec x 3.22 arcsec.

For each X-ray detection and source, the catalog provides a detailed set of more than 100 tabulated positional, spatial, photometric, spectral, and temporal properties (each with associated lower and upper confidence limits) measured in 5 energy bands for ACIS and a single energy band for HRC-I. As a result of this multiplexing, the catalog databases include approximately 1700 columns of information, split across several tables. Furthermore, a Bayesian aperture photometry code produces robust photometric probability density functions for all sources, even in crowded fields and for low count detections. Release 2.1 uses a Bayesian Blocks analysis to identify multiple observations of the same source that have similar multi-band photometric properties, and these observations are analyzed simultaneously to improve S/N.

The catalog contains roughly 40 different types of science-ready source- and field-based data products, totaling about 44 TB, including calibrated data files (photon event lists, multi-band images, backgrounds, and exposure maps), derived measurements (position error Markov chain Monte Carlo draws, extended source contours, aperture photometry Bayesian marginalized posterior probability density functions, and Bayesian Blocks properties), and data analysis products (PSFs, spectra, instrument responses, and light-curves). Since these data products are pre-computed by applying all of the appropriate calibration steps (e.g., matching astrometry, merging observations, applying exposure corrections, removing background) included in the catalog pipelines, they can be used directly by the end-user to significantly simplify the effort required to perform detailed scientific analyses of properties for scientifically meaningful samples of sources without manipulating large volumes of data.

Data Access and User Interfaces

Data access and documentation for the catalog are available through the catalog website (cxc.cfa.harvard.edu/csc/). The catalog documentation describes the content and organization of the catalog data and the data processing steps, and the documentation includes detailed descriptions of the tabulated source and detection properties and data products. The What is new in CSC 2.1 update discusses the algorithmic improvements in CSC 2.1 compared to CSC 2.0. We suggest reviewing the Catalog Organization & Concepts page to make the most effective use of the catalog, while the Caveats and Limitations page provides important information about the catalog that should be reviewed by all users prior to using the catalog data.

Multiple user interfaces to the catalog are accessible via the catalog website’s How to access CSC 2.1 page. The Quick Search web interface supports position-based and crossmatch searches and returns a limited set of source properties, making it ideal for simple queries such as determining whether a source detected in another waveband has a Chandra X-ray counterpart. World Wide Telescope provides a visual interface based on the American Astronomical Society’s World Wide Telescope that exposes the outlines of the stacked observations and the locations of the catalog sources, allowing the user to explore the sky coverage and content of the catalog. Clicking on a source in this tool brings up a box with a basic set of source properties.

For more sophisticated queries and to provide access to the full functionality of the catalog, we recommend using the CSCview application. CSCview is a downloadable Java application that allows arbitrary sets of tabulated properties to be retrieved based on combinations of user-specified constraints on any set of properties plus positional searches or crossmatches. These searches return tabulated results that may be saved to a file or SAMP'ed to another application. Catalog queries can be entered via a simple forms-based interface or alternatively via the ADQL query language. CSCview also provides options to retrieve any desired science-ready FITS data products.

Finally, there are several interfaces for users looking to integrate CSC 2.1 into their pre-established workflow. Chandra's CIAO data analysis package includes a set of CIAO scripts to search the catalog via the catalog command-line interface. Additionally, Virtual Observatory interfaces that support the IVOA Simple Cone Search (SCS), Table Access Protocol (TAP), and Simple Image Access Protocol (SIAP) can access the corresponding catalog interfaces directly; web forms that allow queries via these interfaces are also provided. The Virtual Observatory interfaces allow Jupyter notebooks to access the catalog tables using the PyVO Python package.

A Historical Perspective

The provision of a catalog of X-ray sources identified by the Chandra X-ray Observatory was a requirement of the original NASA contract for the CXC. We started thinking about what such a catalog should include early in 2001 and had developed an initial catalog concept that separated detections and master sources (because of the variation of the size of the Chandra PSF across the field of view) a year later. Around the same time, we were considering whether, in addition to tables of measurements, the catalog should include “data objects” that could be used directly for further detailed analysis. These data objects later morphed into the now roughly 40 types of FITS data products (for compatibility with existing data analysis tools) that we supply as part of CSC 2.1. A couple of years later, we had developed a very preliminary prototype pipeline to perform source detection and other studies. While existing source detection algorithms could be manually fine-tuned for a specific field with a few iterations, the complexity here was to develop an approach that could robustly perform source detection on almost all observation fields irrespective of source content—and to do so with high detection efficiency, low false-source rate, and very minimal human intervention.

By late 2005 our source detection prototype was working sufficiently well, our catalog and pipeline concepts were settled, and other catalog pipeline prototype algorithms—such as aperture photometry—were coming together. In February 2006, we presented our plans to a well-qualified review committee of both X-ray and non-X-ray astronomers from across the globe. They provided outstanding feedback and had many excellent suggestions as to how to proceed. Our biggest takeaway from the committee was to ensure that the catalog was directly usable by multi-wavelength astronomers with little or no X-ray experience, while at the same time being sufficiently robust that X-ray astronomers familiar with Chandra data and state-of-the-art X-ray analysis techniques would feel comfortable using the catalog data. This advice caused us to reconsider and update the set of source properties that we were planning to include in the catalog to ensure that all our potential audiences would be well supported. The next two years saw us continuing to refine the catalog pipelines and fold in the changes that resulted from the feedback provided by the review committee. The first catalog production run commenced in September 2008, and Version 1.0 of the catalog, which included 94,676 X-ray sources from observations released publicly through the end of 2008, was released in March 2009. In August 2010, an incremental catalog release, CSC 1.1, was published, containing 106,586 sources from observations released publicly through the end of 2009.

Following release 1.0, we received tremendous feedback from the user community with numerous suggestions for future catalog releases. The three biggest limitations of release 1 were (a) source detection was performed on a single observation at a time, (b) the source detection threshold was relatively high, around 10 net counts on-axis, and (c) the source detection approach was unreliable in areas where the X-ray background was spatially variable (which resulted in regions around the cores of bright galaxies—often the most interesting region of the observation—being excluded). Release 2.0 was planned to address these issues and more. Multiple observations of the same field would be stacked (co-added) where possible prior to source detection, source detection would be modified to use multiple algorithms to identify candidate detections, and those detections would be graded using a maximum likelihood algorithm to achieve a lower detection threshold while still ensuring a low false-source rate. The changes required major updates for virtually all aspects of the production pipelines, and these updates required much longer than originally anticipated to develop and test. Release 2.0 production started in April 2015. While the initial stages of processing went well, by mid-autumn of 2015 we had amassed sufficient data to indicate that the maximum likelihood algorithm was providing unreliable detection positions for an excessive fraction of detections. Processing was suspended while we investigated the cause of this issue and developed a fix. Production restarted in March 2016 and CSC 2.0 was finally released in October 2019, with 317,167 X-ray sources and ~1.4 million individual observation detections and photometric upper limits extracted from observations released publicly through the end of 2014. While the operational production schedule was impacted by multiple hardware and software issues, much of the delay resulted from the time required for manual quality assurance reviews that were designed to ensure a reliable catalog.

After the CSC 2.0 release, we immediately started working on a minimal set of updates necessary to add in more recent observations (accounting for the slow changes in the characteristics of the Chandra spacecraft and instruments that happen with age) while also dramatically reducing the number of manual quality assurance reviews compared to CSC 2.0. Version 2.1 of the CSC began production in April 2022. Completion of the release, announced above, brings us to the present day.

Enabling Science

The reason for the existence of the CSC is the science that it enables. The catalog constitutes a rich dataset that facilitates data-driven discovery, population studies, and detailed characterization of individual astrophysical sources. Some of the science highlights of the catalog include studies of the contribution of ultra-luminous X-ray galaxies to the ionizing radiation of galaxies, the identification of dual AGN and their properties as a function of pair separation, and the identification of transient events related to the merger of neutron stars. Release CSC 2.1 promises to bring along a new wave of discovery.