5.2. Red QSOs

If QSOs are high luminosity Seyfert 1s, high luminosity Seyfert 1.8s, 1.9s and 2s should exist ([475]). Type 1.8 and 1.9 QSOs should appear as QSOs with a reddened continuum and a high Balmer decrement; they are called ``red QSOs''.

Most radio quiet QSOs have been discovered by UV-excess surveys which are biased against reddened QSOs. The identification of red QSOs suffers from a redshift bias in optical surveys because at higher redshifts optical spectra sample shorter wavelengths where there is greater extinction ([473]). [360] showed that UV-selected QSOs indeed suffer little extinction (0 to 0.55 mag.).

In radio loud QSOs, the optical continuum steepens and the narrow-line EWs and Balmer decrements of the broad-line components increase with increasing viewing angle (decreasing flux ratio R of the core to the extended radio components), implying that reddening is considerable (A_V ~ 2-4 mag) in most lobe-dominated QSOs ([206]; [16]); the broad lines are reddened more than the narrow lines, locating much of the dust responsible for absorbing the broad-line emission between the broad- and narrow-line regions ([193]; [16]). H is the only broad line that has been detected in the two QSOs 3C22.0 ([344]) and RXJ1343.4+0001 ([377]). All these objects are the QSO equivalent of Seyfert 1.8 and 1.9 galaxies.

Webster et al. (1995) discovered many sources with unusually red optical continua (B-K) among the QSOs of the Parkes sample of flat-spectrum radio sources and interpreted this result in terms of extinction by dust. They suggested that existing surveys of bright optical QSOs may be missing ~ 80% of QSOs with a given intrinsic optical magnitude. [34] showed that many of the reddest steep spectrum radio loud QSOs have non-stellar images in K, consistent with underlying gE galaxies, suggesting that contamination by starlight accounts for much of the reddening of the measured B-K colours for these objects; however, [283] showed that this may not be the case for flat spectrum QSOs the colours of which could be explained by the addition of a red synchrotron component to an underlying blue continuum which is identical to that in radio quiet QSOs.

X-ray selection is a way of finding red QSOs efficiently: hard X-rays (2-10 keV) penetrate even tens of magnitudes of optical extinction with minimal absorption; even the lower energy band of ROSAT (0.5-2.5 keV) is not strongly affected by optical extinction of up to ~ 2 magnitudes. [219] have looked for red starlike objects associated with ROSAT sources and found seven red QSOs; their H / H ratio, optical slope and X-ray colours all indicate that they are absorbed by A_V ~ 2 mag. Their spectra could be classified either as type 1.8s or 1.9s; they have relatively low intrinsic luminosities and, even if corrected for absorption, they would be bright Seyfert 1s or weak QSOs.

Fiore et al. (1999) have made a survey in the X-ray band 5-10 keV; the resulting sample contains 14 objects, 13 of which are associated with AGNs; one is a Liner, 6 are unreddened QSOs, while seven are Seyfert 1.8s or 1.9s; none are Seyfert 2s which is somewhat surprising since, in this X-ray band, objects with N_H < 10²⁴ cm^-2 should not be affected by extinction and, as we have seen above, about half of all Seyfert 2s have such small column densities ([353]).