next previous
Up: Ultraviolet variability of

8. Conclusions

The analysis of the ultraviolet variability of the AGN in the Seyfert I galaxy Fairall-9 (continuum and emission lines) has been made using IUE spectra of this AGN between 1978 and 1991 (4903 days) with a mean sampling interval of 96 days. These data have been combined with those available at other wavelength domains to derive the SED of F-9. This instantaneous SED has been used as input for the ionizing continuum in photoionization models. The line profile variability has allowed the decomposition of the strongest lines in several components. The comparison between the line ratios in these components for tex2html_wrap_inline8003, tex2html_wrap_inline8005, tex2html_wrap_inline8007 and tex2html_wrap_inline8009 combined with the photoionization model calculations and the results of reverberation analysis has allowed to derive a consistent model for the structure and dynamics of the BLR.

  1. The UV continuum shows a dominant variation with tex2html_wrap_inline8011 and a characteristic time of 182 days. Also faster variability on time scales of the order of 56 days appears present, but this has not been adequately sampled in the available data. The intrinsic spectral index UV-optical is tex2html_wrap_inline8013 and the optical variations don't lag behind the UV ones. In the NIR, The J band flux shows as an extension of the UV-optical continuum with tex2html_wrap_inline8017. On the contrary, the H, K, and L bands variations lag behind the UV ones (Clavel et al. 1989). The tex2html_wrap_inline8025 flux is proportional to the UV continuum at low flux levels, while at higher UV flux levels this proportionality breaks down.
  2. The SED of the F-9 nucleus shows two important flux excesses, extending over the FIR-NIR range and from the J-band to the X-rays. The FIR-NIR excess, together with the delays of the H, K and L bands fluxes with respect to the UV continuum, is consistent with reradiation from dust heated directly by the ionizing continuum, as already suggested by Clavel et al.\ (1989) and Barvainis (1992). The deduced hot dust mass is tex2html_wrap_inline8035 at a distance of tex2html_wrap_inline8037 of the UV ionizing source (Clavel et al.\ 1989).

    The problems for a simple disk model to explain the big blue bump (J-soft X-rays) of F-9 (Clavel et al. 1989 and Courvoisier & Clavel 1991) and the strong correlation observed in the X-rays-UV variations at low UV flux levels for Fairall-9, are similar to those reported for NGC 5548 (Clavel et al. 1992) and NGC 4151 (Perola et al.\ 1986). In these AGN the partial solution includes reprocessing on an accretion disk of the hard X-rays emitted from a region above the disk. The presence of strong Fe Ktex2html_wrap_inline8041 line centered at tex2html_wrap_inline8043 in the Ginga spectra of F-9 supports this model. However, the equivalent width of this line, tex2html_wrap_inline8045, could indicate that the Fe abundance is tex2html_wrap_inline8047 larger than the cosmic one (Zdziarski 1990; Bai 1979). Wills et al. (1985) similarly obtained this Fe overabundance in their Seyfert 1 galaxies sample from the high ratio FeII tex2html_wrap_inline8049. This is consistent with the ratio obtained by us (see Table 8 (click here)).

  3. The line profile variability for tex2html_wrap_inline8051, tex2html_wrap_inline8053, tex2html_wrap_inline8055 and tex2html_wrap_inline8057\ has allowed us to isolate the same four components (represented by gaussians) in all lines and at all levels of brightness: one narrow (i.e. unresolved at the IUE resolution), and three broad components: a central (FWHM=3600 tex2html_wrap_inline8061, velocity same as the narrow line), a red shifted (FWHM=5800 tex2html_wrap_inline8065, tex2html_wrap_inline8067) and a blue shifted (FWHM=5900 tex2html_wrap_inline8071, tex2html_wrap_inline8073) one. The strong correlations obtained between the three broad components and the UV continuum indicates that photoionization is the dominant mechanism in the BLR. The delays between these components and the UV continuum are tex2html_wrap_inline8075, tex2html_wrap_inline8077 and tex2html_wrap_inline8079 for the central, red and blue one, respectively.
  4. Photoionization models (CLOUDY) with a UV-optical cut-off of 3.5 ryd, derived from the observed instantaneous SED, have been applied to the observed lines ratios of the broad components, and indicate that the BLR is situated at tex2html_wrap_inline8081 from the ionizing source, with an hydrogen density of tex2html_wrap_inline8083, a column density of tex2html_wrap_inline8085 and a covering factor of tex2html_wrap_inline8087, with an ionization parameter between 0.003 and 0.089. However, all these optically thick models underestimate the tex2html_wrap_inline8089 ratio and overestimate the tex2html_wrap_inline8091 one. The solution to this discrepancy could be associated with the co-existence of optically thick and optically thin material in the BLR, as suggested by Shield (1995).
  5. From all the results we have propose a model for the structure and dynamics of the BLR: the mass of the central compact object is tex2html_wrap_inline8093. Around this exist two distinct gas zones within the BLR: the gas producing the central component at tex2html_wrap_inline8095 and the one emitting the red and the blue fluxes at tex2html_wrap_inline8097 moving inward to a central source. These results require that both gas zones be localized along the line-of-sight or, alternatively, that the continuum emission must be strongly anisotropic. Besides, the gas emitting the central component is most likely mixed with the dust and the central gas to dust mass ratio is Mass tex2html_wrap_inline8099.

Acknowledgements

Part of this work was supported by a Comunidad de Madrid FPI grant between 1990-1994. M.C. Recondo-González thanks all the VILSPA personal for their kindness and Antonio De la Fuente for his help.


next previous
Up: Ultraviolet variability of

Copyright by the European Southern Observatory (ESO)
web@ed-phys.fr