12.2.1 qso_candidates table

The qso_candidates table gathers products addressing the classification and characterisation of sources identified as quasar candidates. The name ‘quasar’ is used here as a generic term but the nature of the sources covered by a given DPAC module will vary depending on the type of product, and cover among other things sources considered as broad emission line or broad absorption line nucleus-dominated sources (i.e. QSO or BAL), galaxies with active galactic nuclei (i.e. AGN) and blazars.

This table integrates sources stemming from various processing modules. The exact selection rules applied to the outputs of these modules are specified in Table 12.1, and the logic of the table integration is as follows:

• •

  First of all, the eligibility criteria of each module are applied in order to construct the list of sources eventually gathered in qso_candidates. Each criteria is checked separately and sources satisfying any of the criteria are included in the table.

• •

  All the parameters available from any of the modules are then used to populate the fields of the sources resulting from the selection described in the first point. As a consequence, it is possible that a source has parameters from a module even though the source does not pass that module’s selection criteria.

These principles imply that parameters can end up in the table even for sources which were not eligible according to the selection rules of the module that produced them. In order to isolate those sources originally selected based on the eligibility rules of a given module, a bit field source_selection_flags is provided in the table. Section 12.5 also gives hints about how this field and others can be used in ADQL queries.

Finally, it should be noted that the on-board source acquisition rules of Gaia also determine whether extragalactic sources appear in the final catalogue. For point-like quasar sources, this selection is essentially related to the magnitude and affects the completeness at the faint end. For AGN in resolved host galaxies, several effects can affect a source’s detection due to its spatial extent compared to the pixel and PSF sizes.

The outcome of this selection results in the following sources and products from the respective contributing modules:

• •

  Sources for which the Surface brightness as observed by Gaia has been analysed and fitted as described in Chapter 9. About a million of such sources are present in the table. For those, a flag indicating whether or not a host galaxy was detected is given (host_galaxy_detected). This flag is raised for about 65 thousand sources. For a quarter of those, surface brightness parameters of the host galaxies, following a Sérsic profile fitting, were provided. An additional flag provides further information about the processing status (host_galaxy_flag). In the rest of this chapter, this sample is referred to as the ‘Surface brightness’ sample.

• •

  Sources classified as ‘quasar’ by any of the DSC classifiers (‘Allosmod’, ‘Specmod’, or their combination in ‘Combmod’). As explained in Section 11.3.2, the corresponding selection favours completeness over purity. As such, about 5.5 million sources stem from the results of this classification. In the qso_candidates table, a class label (classlabel_dsc) is provided but it should be borne in mind that it corresponds to that of the ‘Combmod’ classifier only. As such there are about 300 thousand fewer sources having this particular class label set to ‘quasar’ than sources considered eligible by DSC in the table overall. Likewise, the two tabulated classification probabilities (classprob_dsc_combmod_quasar and classprob_dsc_combmod_galaxy) are those of the ‘Combmod’ classifier only. In the rest of this chapter, this sample is referred to as the ‘DSC’ sample. For science cases requiring a purer dataset, an additional class label was computed and tabulated as classlabel_dsc_joint. We hereafter refer to this sample as ‘DSC-Joint’. Finally, note that the two class probabilities hosted in the qso_candidates table are also provided with the DSC results provided in the top-level astrophysical_parameters table.

• •

  Sources for which a reliable spectroscopic redshift was computed by the QSOC algorithm based on the BP/RP low-resolution spectra, as described in Section 11.3.14. This amounts to about 1.8 million sources with an eligible redshift provided in the table (redshift_qsoc), although not all of those are considered equally reliable: a quality flag is provided to indicate this (flags_qsoc). About 300 thousand sources meet the high reliability criteria (flag set to 0). In the rest of this chapter, this sample is referred to as the ‘QSOC’ sample. Apart from those, redshifts determined by QSOC but considered non-eligible will also be tabulated for sources selected through other modules (see Table 12.2 and Section 12.3.6).

• •

  Sources classified as ‘AGN’ by a classifier based on the analysis of photometric lightcurves, and described in Section 10.3. For these both a class label (vari_best_class_name) and a class probability (vari_best_class_score) are provided. Details about how these are computed are given in Section 10.3. This classification contributes to about a million sources. In the rest of this chapter, this sample is referred to as the ‘Vari-Classification’ sample. Note that part of this information is redundant with the classification results provided in the vari_classifier_result table.

• •

  For a subset of the Vari-Classification sample (about 872 thousand sources), additional characteristics are provided based on a Special Object Study (SOS) analysis described in Section 10.4. See also table qso_candidates for further details. Note that these parameters are also tabulated separately in the table vari_agn: the content is strictly the same and is repeated in the qso_candidates table for convenience purpose. In the rest of this chapter, this sub-sample is referred to as the ‘Vari-AGN’ sample.

• •

  The sources published in Gaia EDR3 in the table agn_cross_id, and corresponding to the Gaia-CRF3 (Gaia Collaboration et al. 2022), are also added to the table as a further convenience, and indicated with the flag gaia_crf_source. A complementary flag astrometric_selection_flag is also added in order to qualify each of the sources in the qso_candidates table based on an analysis similar to the one that lead to the selection of the Gaia-CRF3 source list (see Section 12.4). In the rest of this chapter, this sample is referred to as the ‘Gaia-CRF3’ sample.

• •

  The class label of sources classified in Self-Organised Maps (SOM) by the OA module described in Section 11.3.12 is tabulated in (classlabel_oa). This information is in principle redundant with that provided in the oa_neuron_information table, but it is here limited to those sources selected through the other modules listed above. As such, there are many other sources labelled as extragalactic by OA that did not end up in qso_candidates, and these are discussed in Section 11.3.12. In the rest of this chapter, this sample is referred to as the ‘OA’ sample.

The Surface brightness sample is based on an input list of quasar candidates that was built from external catalogues. A companion table qso_catalogue_name is provided to complement this. The table compiles the list of catalogues from which such sources were selected. The code used to identify each catalogue is described in catalogue_id. See Section 9.2 for further details.

The qso_candidates table contains 6 649 162 entries in Gaia DR3. Many of them overlap between the modules listed above, but a significant fraction of those is also unique to a given module. Table 12.2 gives an overview of the sources contributed by each module, and their overlaps are indicated in Table 12.3. Section 12.3 provides further statistics about the source distribution among modules.