Quasar detection with the GAIA broad-band photometric system

Alexander Gontcharov
Dmitry Docenko

Nordic Baltic Research Course in Astrophysical Photometry

Abstract

It is evident that to achieve the new level in understanding of many phenomena in the universe and to resolve a large number of current astrophysical problems, the higher accuracy in measurements of all kinds is needed. For instance, for kinematic separation of stellar populations in our Galaxy or for measure cosmological acceleration of the Sun, it is necessary to define reliable astrometric coordinate system etc. Possibility of using such reference frame is of great importance nowadays, since suitability of existing coor-dinate frame requires updating. The current precision in position measurements of astro-nomical objects is approaching to the limits of the Fundamental Catalog 5. Establishing of the new non-rotating reference frame is one of the tasks of GAIA project.

Concept and technology study for this satellite was finished by ESA in 1998, and it will be launched in 2009. The GAIA mission includes multi-band photometry of all objects for which astrometric measurements are taken. The whole optical range from near UV to near IR will be covered. Typical color index accuracy is supposed to be 0.02 magnitude. The quasars are the best candidates for forming reference points in new non-rotating reference frame. Among the billion point-like objects observed by GAIA there will be quite a few quasars. Because of low brininess they can be identified only by pho-tometric criteria (unusual colors and variability). For doing that GAIA has Broad Band Photometer (BBP) which divides the visual range into four spectral bands: F33B, F45B, F63B and F83B centered on 330, 450, 630 and 830 nm respectively. From these band color indices of an object will be estimated and conclusions on its origin will be made.

By means of synthetic photometry it is possible to imitate results of GAIA measurements and find suitable method for quasars' identification. For this purpose our work has been done.

In our work we have used the composite spectra of quasars with redshifts from 0.0 to 5.0 (with step of z=0.1), some main sequence stars and subgiant stars at the spectral interval 240 - 1100 nm, curve of CCD quantum efficiency (QE), curves of transparency of BBP four bands (Tj) and curve of interstellar extinction. Using formula

Fj =Integral(QE(l)*f(l)*Tj(l)*IE(l)dl),

where l - wavelength,
f(l) - initial flux of object,
Fj - registered flux,
we are able to estimate fluxes Fj through every BBP filter j. Then, with formula

mj=C-2.5lg(Fj) we estimate magnitudes of every object through every filter. Here C is constant which in our work is of no importance because only color indices are used.

As we have four BBP bands, there are three independent color indices, i.e. one can imagine three-dimensional "color space" where all quasars and stars reside. It turns out that these objects are close to one plane (Fan, 1999), so we can use this plane to get maximum separation of objects. We approximated this plane with (m1-m2, (m1-m4)+(m1- -m3)), where 1,2,3,4 are BBP bands from UV to IR.

If we consider that there is no extinction for both quasars and stars, that is the first approxi-mation for galactic latitudes |b| > 15°-20° (Lindegren, personal communication), then the separation between given stars and quasars is more then 1m, therefore they can be easily distinguished. We were not given spectra of white dwarfs (WD), but from the logical point of view it is seen that they can be mixed with quasars. But as WD are faint objects, seen WD are close enough to Sun to measure their parallax with GAIA.

At the second step we have included interstellar extinction effect with assumption that all objects are equally affected. We have found that sequences of quasars and stars are distinguishable up to mean value of extinction about 7m. Since quasars are faint objects they will not be detected under such extinction, thus observable quasars will not overlap with stars.

The third step is to combine quasars and stars with different reddenings that is more realistic approach. It comes out that if reddening of stars is greater then of quasars for about 1.5 magnitudes, then quasars are indistinguishable from O-B type main sequence stars. But GAIA high accuracy in parallax measurements helps us to overcome this problem. O and B type stars are supposed to be in the Galaxy, having large lumino-sity (unreddened not fainter then 12m) and detectable parallaxes.

The bottom line is that in all three cases we can separate quasars from stars using the color-color diagram or measuring parallaxes of suspected objects.


Get original version in Microsoft Word format