We continue our study of the spectral energy distributions ( SEDs ) of 11 AGN at 1.5 < z < 2.2 , with optical-NIR spectra , X-ray data and mid-IR photometry .
In a previous paper we presented the observations and models ; in this paper we explore the parameter space of these models .
We first quantify uncertainties on the black hole masses ( M _ { BH } ) and degeneracies between SED parameters .
The effect of BH spin is tested , and we find that while low to moderate spin values ( a _ { * } \leqslant 0.9 ) are compatible with the data in all cases , maximal spin ( a _ { * } = 0.998 ) can only describe the data if the accretion disc is face-on .
The outer accretion disc radii are well constrained in 8/11 objects , and are found to be a factor \sim 5 smaller than the self-gravity radii .
We then extend our modelling campaign into the mid-IR regime with WISE photometry , adding components for the host galaxy and dusty torus .
Our estimates of the host galaxy luminosities are consistent with the M _ { BH } –bulge relationship , and the measured torus properties ( covering factor and temperatures ) are in agreement with earlier work , suggesting a predominantly silicate-based grain composition .
Finally , we deconvolve the optical-NIR spectra using our SED continuum model .
We claim that this is a more physically motivated approach than using empirical descriptions of the continuum such as broken power-laws .
For our small sample , we verify previously noted correlations between emission linewidths and luminosities commonly used for single-epoch M _ { BH } estimates , and observe a statistically significant anti-correlation between [ O iii ] equivalent width and AGN luminosity .