EDP Sciences
Free Access
Issue
A&A
Volume 615, July 2018
Article Number L8
Number of page(s) 9
Section Letters to the Editor
DOI https://doi.org/10.1051/0004-6361/201833396
Published online 19 July 2018

© ESO 2018

1. Introduction

There is great interest in building unbiased catalogues of quasars for use in answering a range of important astrophysical questions. Further constraints on these objects are needed to further our understanding of the quasar phenomenon itself and the growth and occurrence of supermassive black holes through cosmic time. Quasars are also used as probes of intervening material, and have been shown to play a role in re-ionisation of both hydrogen and helium, and effect the ultraviolet (UV) background levels throughout the universe (Hewett & Foltz 1994; Weymann et al. 1981). Most current quasar surveys, however, rely on their specific intrinsic properties, such as strong UV emission (Schneider et al. 2010), distinct near-infrared (NIR) and mid-infrared (MIR) colours (Maddox et al. 2012; Secrest et al. 2015), X-ray output Brusa et al. (2010) or prominent radio emission (Ivezić et al. 2002).

In this Letter we apply an astrometric approach, identifying quasars as apparently stationary sources on the sky, based purely on the astrometric measurement from the Gaia mission (Heintz et al. 2015), and present a pilot study. Such a technique was already proposed by Koo et al. (1986) but has so far not been widely applied. Our goal here is to quantify the efficiency and completeness of this selection technique. Identifying quasars based only on their zero proper motions has the potential to open a route of selecting quasars in an unbiased way, and might even lead to the discovery of new types of quasars or other types of extragalactic point sources.

2. Astrometric selection of quasars

The Gaia data release 2 (DR2; Gaia Collaboration 2018) catalogue consists of more than 1.3 × 109 sources down to G ≈ 21 mag, for which the five-parameter astrometric solution (positions, parallaxes, and proper motions) has been determined (Lindegren et al. 2018). The Gaia G filter is very broad, covering the spectral range from 400 to 1000 nm, and therefore quasars in a wide range of redshifts should be included in the catalogue.

We extract all sources within a radius of one degree from the north Galactic pole (NGP) centred on (α, δ) = 12h51m26s.0 + 27°07m42ps.0 from the Gaia DR2 catalogue (see Fig. 1). We then limit our search to sources with 18 < G < 20 mag, for which the associated uncertainty is up to 1.2 mas yr−1 in the respective proper motion components. This is motivated by our pre-study (Heintz et al. 2015) in which we (based on pre-launch simulations of the Gaia data) found that the expected contamination of apparently stationary stars is lower than ≈20% at the Galactic poles, but increases significantly when observing closer to the Galactic plane or at magnitudes brighter than G < 18 mag. This criterion also rejects the faintest Gaia sources (G > 20 mag), which still have significant errors on their measured proper motions. We then select all point sources with total proper motions, , consistent with zero at the 2σ confidence level (i.e. S/N µ = µ/µerr < 2). Finally, we identify the counterpart to each source in the Sloan Digital Sky Survey (SDSS) and require that all Gaia sources have morphologies consistent with being point sources (class = 6 in the SDSS) to limit our search to quasars only (i.e. excluding Seyferts and potential contaminating extended galaxies). This results in about 2% of the sample being removed due to extended morphology. Matching the Gaia sample to the SDSS with a matching radius of less than 1 arcsec also allows us to investigate the properties of our sample in optical colour-colour space.

thumbnail Fig. 1.

Location on the sky of all point-like Gaia sources with proper motions and 18 < G < 20 mag (black dots) within one degree of the NGP. The subset of these with proper motions consistent with zero within 2σ are shown by the blue circles and those that are already spectroscopically confirmed quasars are shown by the red star symbols.

Open with DEXTER

In total, we find that there are 2634 point sources observed by Gaia, all with measured proper motions and 18 < G < 20 mag, within one degree of the NGP, of which 100 sources (≈4%) have proper motions consistent with zero (within 2σ). These are marked with blue circles in Fig. 1. Cross-matching our extracted catalogue with the SDSS data release 14 quasar sample (SDSS-DR14Q, Pâris et al. 2018) and the NASA/IPAC Extragalactic Database (NED), we find that 34 quasars are already spectroscopically confirmed within the same magnitude limit and region on the sky, of which 32 (≈95%) also have S/N µ = µ/µerr < 2. This number is expected based on the statistics of the 2σ cut. For the remaining two, the measured proper motions are 2.76 ± 1.09 and 1.22 ± 0.58 mas yr−1. We also discover two spectroscopically confirmed stars, observed as part of the SDSS-APOGEE survey (Alam et al. 2015). An extract of the full sample of Gaia sources with zero proper motions is presented in Table A.1. We also examine the additional requirement that the sources have parallaxes consistent with zero within 3σ, but only five sources (GQs 1255+2707, 1248+2658, 1247+2655, 1247+2706, and 1251+2804) were outside this criterion so we chose to include them for completeness.

3. Selection efficiency and completeness

We now investigate the location of the Gaia sources with zero proper motions in optical colour-colour space. By doing so, we can examine whether these candidate quasars have, for example, UV excess typical of unobscured, low-z quasars (e.g. Sandage 1965; Schmidt & Green 1983). About 70% of the zero proper motion sources have blue (ug < 1) colours (see Fig. 2). For quasars at z ≳ 2.2, the Lyman-α emission line will move out of the u-band, such that the quasars appear redder in ug colour space. At red gr colours (gr > 1), the zero-proper-motion sources have optical colours consistent with M or G dwarf stars. While some of these are likely to be stellar contaminations, removing these candidates will also exclude dust-reddened quasars and broad absorption line (BAL) quasars from the sample, which are found to have very red optical colours and to be systematically missing in most existing quasar samples (Fynbo et al. 2013; Krogager et al. 2015, 2016; Ross et al. 2015; Krawczyk et al. 2015).

thumbnail Fig. 2.

Optical colour-colour plots of the WISE-detected Gaia point sources with proper motions and 18 < G < 20 mag (black dots) within one degree of the NGP. Gaia point sources with zero proper motions are represented by the blue dots and the spectroscopically confirmed quasars are shown by the red star symbols. Typical stellar colours are shown as grey dots.

Open with DEXTER

To assess the efficiency of our selection, we cross-match our sample of Gaia sources with zero proper motions to the all-sky MIR survey based on the WISE satellite (AllWISE; Wright et al. 2010). Mid-infrared selection of quasars is efficient at separating stars and galaxies from quasars and is not affected by dust extinction, while also being sensitive to high-redshift quasars. Of the 100 Gaia point sources with zero proper motions, we identify 76 of the counterparts in the AllWISE catalogue within 1 arcsec. This cross-match might introduce a bias excluding quasars with weak IR emission. Stellar contaminants will also have weak IR emission however, and we find that of the 24 sources excluded in this approach, roughly half have a significant UV excess, whereas the other half have optical colours consistent with the main sequence stellar track. In Fig. 3, we show the zero proper motion Gaia sources in mid-infrared colour-colour space. Overplotted are contours of the SDSS-DR14Q sample for which WISE photometry exists. A simple colour criterion of W1 − W2 > 0.8 has been found to be robust in identifying quasars at most redshifts (Stern et al. 2012). In our sample of zero proper motion sources with WISE photometry, 55 (70%) have W1−W2 > 0.8 (of which 29 are already identified quasars) and have WISE colours consistent with the full SDSS-DR14Q sample. We consider the remaining 26 sources as high-likelihood quasars. All these have also been photometrically identified as quasars by Richards et al. (2009), and we list their estimated photometric redshifts in Table A.1 as well, marked by a “P”. We note, however, that at W1−W2 < 0.8, two spectroscopically confirmed quasars have also been observed, one being a high-z quasar with optical colours consistent with known quasars in this redshift range and the other being a typical UV-excess quasar. These have MIR colours more consistent with typical stellar colours, as also illustrated in Fig. 3. We therefore consider the sources with zero proper motions and W1 − W2 < 0.8 as possible contaminants (excluding the two already known quasars). We then infer a conservative selection efficiency of N QSO/N star ≳ 75%. This is a lower limit due to the population of quasars with blue W1 − W2 colours that also populates our sample.

thumbnail Fig. 3.

WISE colour-colour plot of all Gaia point sources (black dots), the subset with zero proper motions (blue dots), and spectroscopically confirmed quasars (red star symbols) within one degree of the NGP. Overplotted are contour levels linearly spaced in ten steps from 0 to 100% of the full SDSS-DR14 quasar sample with MIR counterparts in the AllWISE catalogue.

Open with DEXTER

We present our main result in Fig. 4 where we show the full sample of Gaia sources with 18 < G < 20 mag and within one degree of the NGP for which a counterpart in the AllWISE catalogue could be identified. It is clear from the figure that the majority of point sources selected on the basis of zero proper motions occupy a distinct region in S/Nµ – WISE colour parameter space. This demonstrates that selecting quasars as stationary sources on the sky is definitely feasible and has a high efficiency of ≳ 75%. The completeness is close to 100% within the defined magnitude limit, since all cosmological objects are selected without any prior assumptions on the spectral energy distributions.

thumbnail Fig. 4.

W1−W2 colour as a function of S/Nµ of the WISE-detected Gaia point sources with proper motions and 18 < G < 20 mag (black dots) within one degree of the NGP. Gaia point sources with zero proper motions are represented by blue dots and spectroscopically confirmed quasars are shown with red star symbols.

Open with DEXTER

4. Discussion and conclusions

Here we have demonstrated the possibility to select quasars as stationary objects in the Gaia DR2 data set. When observing fields well away from the Galactic stellar disk (here the NGP) the contamination from stars is very modest (below 25%) when targeting the most relevant magnitudes (here 18 < G < 20). Therefore, astrometric selection offers both a complete and clean selection of quasars.

This technique offers the possibility to take major steps ahead on some very interesting problems relating to the quasar phenomenon. We mention a few examples here. First, obtaining a more complete picture of dust obscuration in quasar hosts will be possible with a sample of quasars selected using proper motions. Second, the redshift dependence of the frequency of BAL quasars can be determined. Third, using a purely astrometrically selected sample of quasars, we can get an independent gauge of the metallicity distribution of intervening galaxies, in particular the damped Lyman-α absorbers. Fourth, the identification of quasars via zero proper motion also provides unbiased measurements of number densities of various absorbers, such as C IV, Mg II, and H I. Such a sample will still be subject to a flux limit, but this is easier to model than the combined effect of a flux limit and the effect of dust reddening on the quasar-selection efficiency in optical quasar surveys. We also note that the Gaia DR2 data have been applied to find new gravitationally lensed quasars (Krone-Martins et al. 2018).

An interesting case is the confirmed quasar SDSS J125209.59+265018.4 (GQ125209+265018 in Table 1). In Fig. 2, this is located as the object on the stellar track at ug = 1.5. In Fig. 3, it is one of the two sources with blue WISE colours at W1 − W2 < 0.8. This illustrates well the capability of the selection of quasars from astrometry to find quasars that are otherwise difficult to photometrically identify.

Table 1.

Point sources within one degree of the NGP with proper motions consistent with zero (within 2σ) and 18 < G < 20 mag.

When the full Gaia data are released, the errors on the proper motions will decrease and it will therefore be easier to disentangle objects that are truly stationary (quasars) from stars with low proper motions. This will also make it possible to search for stationary sources at even fainter magnitudes. Also, since Gaia astrometry exists for most of the sky, this proper motion criteria could help reduce the contamination in other quasar surveys. Since Gaia covers the full sky, the selection can also be carried out for a large sample of sources; however, with the caveat that the contamination from apparently stationary stars increases significantly closer to the Galactic plane. We can also estimate the expected contamination of, for example, the WISE W1−W2 colour selection, where it can be seen from Fig. 4 that 15% of the sources with W1 − W2 > 0.8 have significant proper motions at more than 5σ.

Acknowledgments

We would like to thank the referee for a clear and constructive report. KEH and PJ acknowledge support by a Project Grant (162948–051) from The Icelandic Research Fund. The Cosmic Dawn Center is funded by the DNRF. LC is supported by DFF – 4090-00079.

Appendix A: Table

Table A.1.

Point sources within one degree of the NGP with proper motions consistent with zero (within 2 σ) and 18 < G < 20 mag.

Appendix B: Thumbnail images of all sources

Thumbnails of all point sources within one degree of the NGP with proper motions consistent with zero (within 2σ) and 18 < G < 20 mag. East is up and north is to the right.

thumbnail Fig. B.1.

50 × 50 arcsec2 thumbnails around each stationary source.

Open with DEXTER

References

All Tables

Table 1.

Point sources within one degree of the NGP with proper motions consistent with zero (within 2σ) and 18 < G < 20 mag.

Table A.1.

Point sources within one degree of the NGP with proper motions consistent with zero (within 2 σ) and 18 < G < 20 mag.

All Figures

thumbnail Fig. 1.

Location on the sky of all point-like Gaia sources with proper motions and 18 < G < 20 mag (black dots) within one degree of the NGP. The subset of these with proper motions consistent with zero within 2σ are shown by the blue circles and those that are already spectroscopically confirmed quasars are shown by the red star symbols.

Open with DEXTER
In the text
thumbnail Fig. 2.

Optical colour-colour plots of the WISE-detected Gaia point sources with proper motions and 18 < G < 20 mag (black dots) within one degree of the NGP. Gaia point sources with zero proper motions are represented by the blue dots and the spectroscopically confirmed quasars are shown by the red star symbols. Typical stellar colours are shown as grey dots.

Open with DEXTER
In the text
thumbnail Fig. 3.

WISE colour-colour plot of all Gaia point sources (black dots), the subset with zero proper motions (blue dots), and spectroscopically confirmed quasars (red star symbols) within one degree of the NGP. Overplotted are contour levels linearly spaced in ten steps from 0 to 100% of the full SDSS-DR14 quasar sample with MIR counterparts in the AllWISE catalogue.

Open with DEXTER
In the text
thumbnail Fig. 4.

W1−W2 colour as a function of S/Nµ of the WISE-detected Gaia point sources with proper motions and 18 < G < 20 mag (black dots) within one degree of the NGP. Gaia point sources with zero proper motions are represented by blue dots and spectroscopically confirmed quasars are shown with red star symbols.

Open with DEXTER
In the text
thumbnail Fig. B.1.

50 × 50 arcsec2 thumbnails around each stationary source.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.