Active galactic nuclei at z ∼ 1.5 – I. Spectral energy distribution and accretion discs

Abstract

The physics of active super massive black holes (BHs) is governed by their mass (M BH ), spin (a * ) and accretion rate (Ṁ ). This work is the first in a series of papers with the aim of testing how these parameters determine the observable attributes of active galactic nuclei (AGN). We have selected a sample in a narrow redshift range, centered on z ∼ 1.55, that covers a wide range in M BH andṀ , and are observing them with X-shooter, covering rest wavelengths ∼1200-9800Å. The current work covers 30 such objects and focuses on the origin of the AGN spectral energy distribution (SED). After estimating M BH andṀ based on each observed SED, we use thin AD models and a Bayesian analysis to fit the observed SEDs in our sample. We are able to fit 22/30 of the SEDs. Out of the remaining 8 SEDs, 3 can be fit by the thin AD model by correcting the observed SED for reddening within the host galaxy and 4 can be fit by adding a disc wind to the model. In four of these 8 sources, Milky Way-type extinction, with the strong 2175Å feature, provides the best reddening correction. The distribution in spin parameter covers the entire range, from −1 to 0.998, and the most massive BHs have spin parameters greater than 0.7. This is consistent with the "spin-up" model of BH evolution. Altogether, these results indicate that thin ADs are indeed the main power houses of AGN, and earlier claims to the contrary are likely affected by variability and a limited observed wavelength range.

Reaching the peak of the quasar spectral energy distribution – I. Observations and models

“…This echoes the findings of J12 and of Elvis et al (1994), who use a similar proxy, L bol /L 2500Å . This suggests that the spread is larger than the ∼20 per cent stated in Capellupo et al (2015). This cannot be solely due to the ∼0.1 dex error on our mass estimate.…”

Reaching the peak of the quasar spectral energy distribution – II. Exploring the accretion disc, dusty torus and host galaxy

“…Using this method to accurately measure the accretion rate, spin, inclination, and intrinsic reddening values would require exceptional data coverage and quality. Despite these uncertainties, our measurements of L bol are more accurate than those in other studies that lack X-ray spectra in addition to optical-NIR spectra (Capellupo et al 2015;Castelló-Mor et al 2016).…”

“…This is probably partly for the same reasons as discussed above-fewer Gaussian components in the decomposition lead to less degeneracybut additionally our SED continuum contains only one free parameter (the normalization), versus the Shen & Liu (2012) approach, which uses power-law continua (in some cases with a break included), with both normalization and slope left as free parameters. Mejía-Restrepo et al (2016) used the Capellupo et al (2015) data and models to directly compare results from performing a global decomposition with the AD model to those utilizing local power-law continuum fits, finding the global approach to be more reliable. Although our study is smaller, we have also shown that using our physical model of the underlying AGN continuum facilitates a global spectral decomposition that is at least consistent with alternatives.…”

“…However these models are not well understood (Castelló-Mor, Netzer & Kaspi 2016) and are beyond the scope of this paper. Capellupo et al (2015Capellupo et al ( , 2016 used a thin AD model to constrain spin values in their sample of 39 AGN at z ∼ 1.5. Using the numerical code of Slone & Netzer (2012), and contemporaneous opticalnear-infrared (NIR) spectra from the VLT/X-shooter instrument, they successfully modelled the rest-frame optical-UV SED in all but two objects.…”

The first ultraviolet quasar-stacked spectrum at z ≃ 2.4 from WFC3

“…Furthermore, the standard black body disc model depicted above, does not reproduce the observed soft X-ray emission seen in AGN, consequently other physical parameters are required such as black hole spin and/or substantial Compton upscattering (Laor et al 1997, but see also Capellupo et al 2015).…”