Astron. Astrophys. 358, 57-64 (2000)

8. Discussion and conclusions

The analysis of this paper is based on the working hypothesis that AGN variability is due to star collisions occurring in the neighborhood of the central black hole. These collisions will reproduce the observed AGN luminosities (in the range erg s^-1) and light curves only if their rate, , is such that more than one (and no more than one hundred) collisions occur every year. A fixed range in values implies a definite relationship among the dimensions of the star cluster, its density and the mass of the central black hole (see Sect. 4); high central star densities are required especially around black holes with masses approaching the smallest allowed value (). In this framework the theoretically derived variability values are compatible with the observational results of Paltani & Courvoisier 1997. More in detail, as it is apparent from Fig. 4, the results of our analysis clearly do not follow the slope -1/2 (typic of a series of discrete, independent events), but reproduce very well the observed trend in spite of the simplicity and of the strong approximations of the model. A more detailed and realistic physics both of star collisions and of matter accretion can change the variability values but can not change this trend. In addition the dispersion of theoretical points in Fig. 4 is analogous to that present in the experimental data shown in Paltani & Courvoisier 1997. A fact which simply reflects the existence of different stars distributions in different galaxies.

It is also important to stress that the agreement between our results and those of Paltani & Courvoisier 1997 is not due to the presence of many parameters in the model, since most of them, as explained in Sect. 5, do not influence the result shown in Fig. 4. In fact, the dispersion of the points in Fig. 4 is only due to different concentrations of stars near the galaxy center, which induce different distributions of collision energies (see Sect. 5). This means that the model presented here is able to fit observations because of the form of the keplerian velocity. Peaked density distributions favor high velocity stars and hence strong collisions. The difference in the event intensity can be estimated by comparing the light curves corresponding to two different configurations present in Fig. 4 having the same collision rate value but different density distribution. The first one is the light curve of Fig. 1; the second one is shown in Fig. 6. The two curves are generated with the same collision rate and hence both have a shape compatible with the observed AGN light curves. A difference in the star density distribution ( and , respectively) produces a big difference in the collision intensity and hence in the produced luminosity, as the vertical scales of the two figures show.

Fig. 6. Light curve for the case of collision rate and for .

© European Southern Observatory (ESO) 2000

Online publication: June 26, 2000
helpdesk.link@springer.de