A spectral decomposition of the variable optical, ultraviolet and X-ray continuum of NGC 5548

Abstract

We present an analysis and decomposition of the broad-band optical/UV/X-ray/γ-ray spectrum of the Seyfert 1 galaxy NGC 5548. The spectrum consists of an average of simultaneous optical/IUE/Ginga observations accompanied by ROSAT and CGRO/OSSE data from non-simultaneous observations. We show that the overall optical/UV/X-ray/γ-ray spectrum can be deconvolved into three basic continuum components: a cool multitemperature blackbody, a hard thermal Comptonized component and an EUV/soft X-ray component well described by a thermal Comptonization continuum. Assuming that the optical/IUE spectrum comes from a cold disc, the maximum disc temperature is very well constrained by the data to be kTdisc = 3.2+0.2-0.2 eV. This rules out models explaining the soft excess as a far tail of the disc spectrum. We show that the soft excess inferred by the data requires a separate continuum component, which is consistent with thermal Comptonization in optically thick (τ ∼ 30), warm (∼0.3 keV) plasma. This Comptonization component contains a significant fraction of the source energy. The plasma parameters of the hard continuum (τ ∼ 2, kTHC ∼ 50 keV) are consistent with those suggested for the average spectrum of Seyferts. On the basis of the broad-band spectral model, we also re-analyse the simultaneous IUE/Ginga campaign. We find that the fluxes in all three continuum components are positively correlated. The total flux emitted in the hard component is positively correlated with both the spectral index and the solid angle of cold matter seen by the hot source. Such a correlation suggests that the variability mechanism is related to changes in the geometry of the continuum-emitting regions, and an excess in the amount of reflection requires deviations from a simple plane disc picture. The variations in the hard spectral index can then be explained by increased cooling of the hot plasma caused by the increased number of seed photons. We suggest that the geometry variations may be related to a transition region between a cold and a hot disc.