Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 354, 847-852 (2000)

Previous Section Next Section Title Page Table of Contents

5. Summary and conclusions

The present study clearly proof that there are significant differences in the structure of the X-ray photon trains from different types of sources. A comparison of time scale spectra (Fig. 3), calculated for low apparent luminosity objects, shows twice as large variability for the QSO group as for the S1 group. Larger vertical extent of the average features in the spectra of S1 indicates more frequent occurrence - relative dominance of semi-regular deterministic components.

A study of ampligram amplitude distributions shows that there is another significant difference in the occurrence of deterministic structures in both types of sources. As a measure of information in the distributions the entropy was used. We have seen that there is a lower minimum entropy for lower level wavelet coefficient magnitudes in S1 variability spectra than in QSO spectra, indicating less random and more deterministic or regular contributions in S1 objects. This is also seen in different change in entropy with apparent source luminosity in the two categories (Fig. 10). These results indicate that by studying statistical properties of an X-ray photon train it is possible to identify unambiguously the character of its source.

[FIGURE] Fig. 10. Occurrence of different distribution types (see Fig. 6) for both groups of sources and for different apparent luminosities.

These differences have a natural interpretation. The luminosity "building blocks" or events in Seyfert 1 galaxies, which are responsible for the intrinsic X-ray variability of these sources show more correlations and regularities than those in QSOs. This means that the accretion disk flares or the cluster black hole collisions with accretion disks around the larger holes ("ballistic events", see Pacholczyk & Stoeger 1994) - or whatever causes these discrete events - are themselves more strongly correlated in Seyfert 1 galaxies than in QSOs. If we assume that the black hole cluster scenario, for example, this would mean that subclustering in Seyfert 1 galaxies would be stronger than in QSOs. That might be either because there are more smaller and intermediate size black holes in Seyfert 1 nuclei (QSO nuclei have perhaps evolved much further towards a configuration of just one or a few supermassive black holes surrounded with relatively few smaller satellite black holes, whereas Seyfert 1 nuclei are possibly at a less-evolved black hole nuclear cluster stage, cf. Pacholczyk & Stoeger (1994), and/or because the cluster in QSOs exhibit less tendency towards subclustering for some other reason (they may be more massive and more compact, making subclustering less pronounced or less possible). If subclustering of black holes are common in Seyfert 1 nuclei, their ballistic encounters with accretion disks around other black holes, or with intracluster clouds will be correlated. A spherical subcluster penetrating an accretion disk will exhibit a definite signature - a swarm of little black holes collisions with the disk causing a swarm of small luminosity building blocks followed by fewer larger events from the collisions of the larger black holes in the subcluster, and then finally another swarm of smaller building blocks from the smaller black holes on the back side of the subcluster. More than likely, however, most subclusters will have been tidally deformed into elongated trains of black holes, due to close encounters with larger black holes in the cluster, or with other subclusters. As the black holes in these subcluster trains hit accretion disks in correlated, fairly rapid succession they will generate a detectable series of quasi-periodic series of luminosity events, much like a fragmented asteroid or a comet hitting a planet.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 2000

Online publication: February 25, 2000