Supermassive Black Hole Winds in X-rays: SUBWAYS V. Properties of hot coronae in quasars at intermediate redshift

Abstract

We present the X-ray analysis of coronal properties in a statistically representative sample of 23 mostly radio-quiet AGN from the SUBWAYS campaign (SUpermassive Black holes Winds in XrAYs), focusing on quasars at redshifts $0.1 < z < 0.4 $ and bolometric luminosities $2 \times 10^{44} <L_{bol}(erg/s) < 2 \times 10^{46}$. The main aim of this work is to investigate the properties of the hot corona through the study of the hard X-ray band emission, including a proper treatment of the soft X-ray band. High-quality X-ray spectra from XMM-Newton, complemented by NuSTAR data extending up to 30-40 keV in the rest frame, are available for this sample. The soft X-ray band (0.3-2 keV) spectrum is best fitted by a warm corona model with a median temperature of 0.40 keV, and an optical depth in the range $τ$=20 - 40, consistent with previous results on lower luminosity sources. The hard X-ray band is well described using a hot corona model, with a median high-energy cut-off of 87 keV, at the lower end of the distribution of typical values found in Seyfert galaxies (100 - 200 keV). The derived median value of the optical depth ($τ$ = 1 - 5) suggests the presence of a moderately optically thick corona. Combining the SUBWAYS results with literature samples at low and high redshift, we assemble the largest sample to date of AGN with E$_{cut}$ and accretion parameter measurements, finding a significant anticorrelation of E$_{cut}$ with both $λ_{Edd}$ and $L_{bol}$ with the median E$_{cut}$ decreasing from 250 - 300 keV at low accretion rates and luminosities to 90 - 100 keV at high accretion rates and luminosities - consistent with enhanced coronal cooling, possibly driven by pair-production. These results favor cooler, optically thicker coronae in luminous AGN compared to those in lower-luminosity Seyfert galaxies.

Figures (10)

• Figure 1: a: Spectrum of LBQS1338-0038 fitted with Model-1. Data points are PN (black), MOS1 (red), MOS2 (green), FMPA (blue), and FMPB (cyan). The warm corona component (compTT) is represented by the dashed blue curve, while the reflection-only component from xillver is shown by the dotted orange curve. The primary cut-off power-law is shown in gray, and the total best-fit model for Model-1 is in magenta. b: Residuals with respect to a simple black body+power-law model from Matzeu23, without any reflection or high-energy cut-off. c: Residuals with respect to the fit with Model-1. d: Residuals with respect to the fit with Model-2. The improvement in the fit is significant when reflection and high-energy cut-off are included ($\Delta Cstat\sim30$ for two more parameters), while both Model-1 and Model-2 provide statistically acceptable fits.
• Figure 2: Comparison of the power-law photon index (left side) and of the reflection parameter (right side) among the two models, xillver (x-axis) and xillverCp (y-axis). The single observations are marked as Source with a circle, while multiple observations for a single source are distinguished by the markers and they are explicit in the legend. In black the median value with errors at $90 \%$ confidence level computed with parametric fit using a log-normal function (see \ref{['sec:hotcoronaproperties']} for further details).
• Figure 3: Left: High-energy cut-off distribution for the hot corona using xillver. Right: Electron temperature distribution for hot corona from xillverCp. Detections are shown in orange, while lower limits are shown in violet, stacked over the detections. The median values, taking into account lower limits, are marked by the dashed line with $68\%$ confidence level (light grey areas) and are computed by using the parametric fit using a log-normal function(see \ref{['sec:hotcoronaproperties']} for more details).
• Figure 4: Left: Comparison of the warm corona electron temperatures. On the x-axis, the temperatures obtained from nthcomp model, while on the y-axis, the ones obtained by compTT model, are shown. Right: Comparison of the warm corona optical depths obtained by using nthcomp on the x-axis and compTT on the y-axis. For the former, $\tau$ is computed using \ref{['longair_slab']} (slab geometry) and considering $\Gamma$ and $kT_e^{warm}$ values from nthcomp.
• Figure 5: Left panel: $E_{cut}$ as a function of redshift for the low redshift sample (blue) as compiled by Bertola22, the SUBWAYS sample (red filled circles) the intermediate redshift AGN (in red open circles), and the high redshift sample (magenta diamonds) (Intermediate and High-z sources infos in \ref{['sample_infos']}). The median point per redshift bin is overplotted with larger markers of the same colors of the redshift sample and yellow contours. Right panel: The $E_{cut}$ as a function of the X-ray luminosity (i.e., the compactness-temperature ($l-\theta$) diagram, converted into observable quantities) from the same samples. Yellow regions mark the runaway pair-production limits for a BH mass of $10^8 M_\odot$ under slab (hemisphere) geometry, while dotted lines indicate the same limits for a BH mass of $10^9 M_\odot$. Dark-blue down triangles show the binned results from Ricci2017 on the BAT Sample at low-z.