NuSTAR observations of heavily obscured Swift/BAT AGN: constraints on the Compton-thick AGN fraction

Abstract

The evolution of the accretion history of the Universe has been studied in unprecedented detail owing to recent X-ray surveys performed by Chandra and XMM-Newton. A focus on the most heavily obscured or Compton-thick active galactic nuclei (AGNs) is missing in these studies. These AGNs evade detection even in X-ray surveys owing to their extreme hydrogen column densities, which exceed 10^(24) cm^(−2). Recently, the all-sky hard X-ray survey performed by Swift/BAT brought a breakthrough, allowing the detection of many of these AGNs. This is because of the very high energy bandpass (14–195 keV) of this instrument, which helps to minimise attenuation effects. In our previous work, we identified more than 50 candidate Compton-thick AGNs in the local Universe, corresponding to an observed fraction of about 7% of the total AGNs population. This number can only be converted to the intrinsic Compton-thick AGNs number density if we know their exact selection function. This function sensitively depends on the form of the Compton-thick AGN spectrum, that is the energy of their absorption turnover, photon-index and its cut-off energy at high energies, and the strength of the reflection component on the matter surrounding the nucleus. For example, the reflection component at hard energies 20–40 keV antagonises the number density of missing Compton-thick AGNs in the sense that the stronger the reflection the easier these sources are detected in the BAT band. In order to constrain their number density, we analysed the spectra of 19 Compton-thick AGNs that have been detected with Swift/BAT and have been subsequently observed with NuSTAR in the 3–80 keV band. We analysed their X-ray spectra using the MYTORUS models which properly take into account the Compton scattering effects. These were combined with physically motivated Comptonisation models, which accurately describe the primary coronal X-ray emission. We derived absorbing column densities that are consistent with those derived by the previous Swift/BAT analyses. We estimate the coronal temperatures to be roughly between 25 and 80 keV corresponding to high energy cut-offs roughly between 75 and 250 keV. Furthermore, we find that the majority of our AGNs lack a strong reflection component in the 20–40 keV band placing tighter constraints on the intrinsic fraction of Compton-thick AGNs. Combining these results with our X-ray background synthesis models, we estimate a percentage of Compton-thick AGNs in the local Universe of ≈20 ± 3 % relative to the type-II AGNs population.