Astron. Astrophys. 319, 413-429 (1997)

6. Broad band spectral evolution

The emission from quasars appears to be `scale-free': from the lowest to the highest luminosities the quasar continuum and emission lines scale almost linearly with luminosity and are independent of redshift (Blandford 1990). The fraction of the power emitted in the different energy bands remains similar, although a slowly decreasing X-ray loudness has been found for higher luminosity quasars (Worrall et al. 1987). However, an exact quantitative assessment seems to be difficult due to selection biases and apparent changes in the source population at different redshifts.

The X-ray loudness has been used frequently in the past for the discussion of the relative fraction of X-ray to optical emission in an evolving quasar source population. In Fig. 17 we show a plot of versus optical luminosity. A regression analysis yields for the steep spectrum sources while for flat spectrum sources we get . In both cases the probability levels for a correlation are . The two slopes are consistent with each other in their mutual 1  errors but the larger value of the regression constant for flat spectrum quasars shows that, in agreement with the results of Sect. 5.2, flat spectrum quasars are X-ray louder than the steep spectrum objects. Similar slopes have been found for the optically selected LBQS quasars (Green et al. 1995) and from Einstein data with different analysis techniques by Avni & Tananbaum (1986), Wilkes et al. (1994), and Avni et al. (1995).

Fig. 17. X-ray loudness versus optical luminosity. The full line is the regression slope for steep spectrum objects, the dashed line that for flat spectrum quasars.

However, a slope of implies a non-linear relation , inconsistent with the value of found in chapter 5 (see the discussion).

Again, it cannot be ruled out that this correlation is introduced by a redshift dependence of the luminosities and in fact, Fig. 18 indicates there might be a redshift dependence of the X-ray loudness. We obtain fits of the form for the flat spectrum objects and for steep spectrum sources, however both with relatively large probability levels of and , respectively and thus a strong redshift dependence cannot be confirmed. Correspondingly, for simple redshift independent averages we get , , , and

Fig. 18. X-ray loudness versus redshift z.

These values are in good agreement with those found in previous studies considering the generally large errors of smaller samples. They seem to be considerably smaller than those for radio quiet quasars found by Green et al. (1995), demonstrating that radio-quiet quasars are less X-ray bright at a given optical luminosity. But it must be noted that the different selection criteria of the two samples makes it hard to compare them directly.

© European Southern Observatory (ESO) 1997

Online publication: July 3, 1998
helpdesk@link.springer.de