The Complex Optical to Soft X-Ray Spectrum of Low-Redshift Radio-quiet Quasars. II. Comparison with Free-Free and Accretion Disk Models

Abstract

We compare the optical to soft X-ray spectral energy distributions (SEDs) of a sample of bright low-redshift (0.048 < z < 0.155), radio-quiet quasars, with a range of thermal models which have been proposed to explain the optical/UV/soft X-ray quasar emission: (1) optically thin emission from an ionized plasma, (2) optically thick emission from the innermost regions of an accretion disk in Schwarzschild and Kerr geometries. We presented ROSAT PSPC observations of these quasars in an earlier paper. Here our goals are to search for the signature of thermal emission in the quasar SEDs, and to investigate whether a single component is dominating at different frequencies. We find that isothermal optically thin plasma models can explain the observed soft X-ray color and the mean optical-ultraviolet (OUV) color. However, they predict an ultraviolet (1325 Å) luminosity a factor of 3 to 10 times lower than observed. Pure disk models, even in a Kerr geometry, do not have the necessary flexibility to account for the observed OUV and soft X-ray luminosities. Additional components are needed both in the optical and in the soft X-rays (e.g., a hot corona can explain the soft X-ray color). The most constrained modification of pure disk models, is the assumption of an underlying power-law component extending from the infrared (3 μm) to the X-ray. This can explain both the OUV and soft X-ray colors and luminosities and does not exceed the 3 microns luminosity, where a contribution from hot dust is likely to be important. We also discuss the possibility that the observed soft X-ray color and luminosity are dominated by reflection from the ionized surface of the accretion disk. While modifications of both optically thin plasma models and pure disk models might account for the observed SED, we do not find any strong evidence that the OUV bump and soft X-ray emission are one and the same component. Likewise, we do not find any strong argument which definitely argues in favor of thermal models.