Gaia space mission and quasars

Quasars are often considered to be point-like objects. This is largely true and allows for an excellent alignment of the optical positional reference frame of the ongoing ESA mission Gaia with the International Celestial Reference Frame. But presence of optical jets in quasars can cause shifts of the optical photo-centers at levels detectable by Gaia. Similarly, motion of emitting blobs in the jet can be detected as proper motion shifts. Gaia's measurements of spectral energy distribution for around a million distant quasars is useful to determine their redshifts and to assess their variability on timescales from hours to years. Spatial resolution of Gaia allows to build a complete magnitude limited sample of strongly lensed quasars. The mission had its first public data release in September 2016 and is scheduled to have the next and much more comprehensive one in April 2018. Here we briefly review the capabilities and current results of the mission. Gaia's unique contributions to the studies of quasars are already being published, a highlight being a discovery of a number of quasars with optical jets.


THE GAIA MISSION
Gaia, an ongoing mission of ESA (gaia.esa.int), is conducting a massive all sky optical survey of sources with 0.0 ≤ V ≤ 20.9 (for current review of mission properties and its status see Prusti et al. 2016). Brown et al. (2016) published an initial sky chart with 1,142,679,969 sources, which already makes Gaia the largest all-sky survey of celestial objects to date. Gaia is primarily an astrometric mission, positional accuracy at the end of the mission is expected to be between 5 and 400 µas (Fig. 1), while the spatial resolution is ∼ 0.1 arc-second (as). This impressive accuracy of position of optical photo-center is mostly photon-noise limited and in many cases supersedes the accuracy of VLBI positions for radio emitting sources. Gaia is continuously scanning the skies. In 5 years of official mission lifetime it will observe each source in the sky transiting its focal plane from 50 to over 200 times, median is 72. Sampling is uneven, with the shortest separations corresponding to two observations within the 6-hour spin period of the satellite. Gaia includes three main instruments: astrometry is done in white light (G band, between 330 and 1050 nm), photometry is collected via weakly dispersed spectra in the blue (BP, blue photometer, 330-680 nm) and red (RP, red photometer, 640-1050 nm) bands, each sampled in wavelength by 45 pixels. 76% of energy of a sharp line from a point source is spread in the wavelength direction over ∼ 1.6 pixels (BP) and ∼ 3.8 pixels (RP). Finally there is a spectroscopic instrument on-board which is collecting spectra in the 845-872 nm range for objects brighter than V ∼ 15.5 and at a resolving power of 10,500. The main goal of spectroscopic instrument is to determine radial velocities, but for the very bright stars (V ≤ 12) it will also measure abundances of chemical elements with lines in wavelength range of the spectrograph. Magnitude and wavelength limitations of Gaia's spectroscopic instrument are being complemented by ground-based optical stellar spectroscopic surveys which are aimed to obtain stellar parameters and accurate radial velocities, as well as abundances of individual elements for stars which are not within the reach of Gaia. They can however observe only a small fraction of such sources. The ongoing surveys include RAVE (Kunder et al. 2017, www.rave-survey.org), LAMOST (Cui et al. 2012, www.lamost.org), Gaia-ESO (Gilmore et al. 2012, www.gaia-eso.org), and Galah (www.galahsurvey.org, de Silva et al. 2015, see also Martell et al. 2017).

GAIA AND QUASARS
A vast majority of quasars are fainter than V ∼ 15.5 so they are not observed by the spectroscopic instrument aboard Gaia, but the mission's astrometry and spectrophotometric BP and RP measurements are extremely relevant for quasar studies. Fig. 1 shows that one may expect a positional accuracy of ∼ 70 µas at V ∼ 17 and ∼ 400 µas at V ∼ 20 at the end of the official 5-yr mission lifetime. These figures may actually be too pessimistic, as Gaia performs better than expected on the faint end, and the mission itself may be extended up to 2022, therefore increasing the photon budget and time span of its observations. Fig. 2 shows accuracy of integrated magnitudes over the G, BP, and RP bands for each transit. It demonstrates the potential of Gaia's ∼ 72 observations of each quasar over a ∼ 5 yr timeframe to assess temporal variability for a complete population of quasars down to V ∼ 20 in white light (G band). For bright sources (V < 18) one can also study colour changes on timescales from hours to years, using their integrated BP and RP magnitudes.
BP and RP instruments do not collect only integrated magnitudes but also spectrophotometric information. Fig. 3 shows a typical quasar spectrum (black tracing), and its BP (blue) and RP (red) spectrophotometry collected by Gaia over a 5-year mission. The panels show the signal for objects with apparent magnitude V = 20 which have 4 different redshifts between 0.5 and 3.5. Vertical stripe at ∼ 8600Å is the wavelength range of Gaia's spectroscopic instrument. In our view the figure, which summarises simulations of Proft et al. (2015), clearly demonstrates the ability of Gaia to determine photometric redshifts for a complete population of quasars down to V ∼ 20. For brighter objects even changes in their spectral energy distribution (SED) should be detectable. Proft et al. (2015) demonstrate that variability of SDSS 154757.71+060626.6, which has Sloan red magnitude r ′ = 17.5 and redshift z = 2.03, is clearly within the reach of Gaia.
Gaia mission observes any object bright enough to trigger its instruments. So the initial quasar catalogue with 1,248,372 sources, which was prepared before the launch by Andrei et al. (2014), is only a lower limit of what we may expect to be observed. Properties of the catalogue are summarised in Table 1. Note the relatively low astrometry precision for most of the quasars. Position is of course more accurately known for 4866 VLBI sources which are part of the catalogue and for additional 38,699 sources with available radio position, although of lower precision.
Quasars observed within the first data release have been discussed by Lindegren et al. (2016) and by Mignard et al. (2016). Main results can be summarised as follows: Some 135,000 quasars from the list of Andrei et al. (2014) were included in the astrometric solution, and their positions were determined with an (inflated) standard uncertainty of 1 mas. Accurately known positions from VLBI were used to align the Gaia DR1 reference frame with the extragalactic radio frame. A special astrometric solution has been computed for selected extragalactic objects with optical counterparts in Gaia DR1, including the ones from the second realisation of the International Celestial Reference Frame (ICRF2). Formal standard error for 2191 such quasars (with 17.6 < V < 20.7) does not exceed 0.76 mas for 50% and 3.35 mas for 90% of the sources. Alignment of the Gaia DR1 reference frame with the ICRF2 is better than 0.1 mas at epoch 2015.0. The two frames do not rotate to within 0.03 mas/yr. There are now 11,444 objects with VLBI positions, i.e. 3.5-times more than in ICRF2 ). Lindegren et al. (2016) note that for sources with good optical and radio astrometry they found no indication of physical optical vs. radio offsets exceeding a few tens of mas, in most cases they are less than 1 mas. This in encouraging, as it will permit to build a very accurate common Gaia+radio reference frame in the future. But some objects may require further checks. In particular, Makarov et al. (2017) publish a list of 188 objects (out of 2293 ICRF2 sources), most of them with z < 0.5, with offsets up to 1 arcsec, the latter corresponds to a distance of ∼ 7.5 kpc at z ∼ 0.5. They propose that 89 of them may be AGNs (quasars and BL Lac) dislodged from their host galaxies' centers.

OFFSETS BETWEEN GAIA AND VLBI POSITIONS AND EVIDENCE FOR OPTICAL JETS
In a much larger study  used the whole 1.1 billion object dataset from Gaia DR1 and the VLBI absolute astrometry catalogue RFC 2016c to find 6055 secure matches with AGNs. Kovalev et al. (2017) used this large sample to check on offsets between radio and optical positions. In 2957 AGNs they were able to determine the direction of the radio jet. They find a significant prevalence of optical and radio offsets in directions along or opposite the one of the radio jet (Fig. 4). This suggests that strong, extended optical jet structures are present in many AGN. Position offsets along the jet require strong, extended parsec-scale optical jets, while small (< 1 mas) offsets in direction opposite to the jet can be due to extended VLBI jet structure or a "core-shift" effect due to synchrotron opacity.

CONCLUSIONS
Gaia is living to its promise of revolutionising virtually any field of astronomy. In case of quasars two of the instruments on board are of particular interest: spectrophotometric BP/RP instrument and astrometric instrument. The former allows studies of SED variability for a complete sample of sources with V < 18, and of general brightness variations and photometric redshift determination for all sources brighter than V ∼ 20. The emphasis is on completeness, and on the ability to discover and characterise yet unknown quasars in this magnitude range. Both present a major advance to current lively research efforts to characterise quasar variability (e.g. Peters et al. 2016;Marziani et al. 2017;Rumbaugh et al. 2017). An obvious importance of astrometry of quasars is to harmonise optical and radio reference frames. Some 135,000 quasars, which were already analysed within the first data release, already allowed to achieve an unprecedented accuracy of alignment of VLBI radio and optical Gaia reference frames. A small offset between optical and radio position, which is seen in some quasars with well determined directions of radio jets, points to presence of strong, extended optical jets on parsec scales. Spatial resolution of Gaia allows to discover large numbers of gravitationally lensed quasars (Lemon et al. 2017), especially when photometric and spectrophotometric measurements of Gaia are combined with mid-infrared selection, as enabled by the WISE mission (Agnello 2017).
Gaia observations are ongoing, with the second public data release planned for April 2018. It will include a five-parameter astrometric solutions for all sources with acceptable formal standard errors (> 10 9 are anticipated), and positions (R.A., Dec.) for sources for which parallaxes and proper motions cannot be derived. Next, it will include G and integrated BP and RP photometric magnitudes for all sources. For sources brighter than G RV S = 12 also median radial velocities will be published. Finally for stars brighter than G = 17 there will be estimates of the effective temperature and, where possible, line-of-sight extinction, based on the above photometric data. Of less interest to quasar studies, a list of photometric data for a sample of variable stars and epoch astrometry for a pre-selected list of over 10,000 asteroids will be provided.
There are further planned intermediate public data releases, with the final database to be published some 3 years after the end of observations, which may happen anytime between 2019 and 2023. By then a complete magnitude limited sample of quasars, many of them new, together with their variability, redshift and in many cases optical jets or multiple gravitationally lensed images are bound to be discovered.