The optical-infrared relation for active galactic nuclei: The role of contaminations

TL;DR

This work tackles the calibration of the optical–infrared (OPT–IR) luminosity relation for quasars, explicitly accounting for host-galaxy contaminations such as stellar light, ISM, and polar dust to improve its reliability as a cosmological probe. Using a ~400-quasar sample from SDSS DR16Q across four redshift bins, the authors perform energy-balanced SED fitting with CIGALE, incorporating a SKIRTOR AGN torus plus a polar-dust component, and compare photometric luminosities with those derived from the SED fits. They find that photometric L_{IR} and L_{OPT} tend to be upper limits to the SED-based torus and disk luminosities, but that incorporating disk, polar dust, and stellar contamination brings the two estimates into near 1:1 agreement with reduced intrinsic scatter, especially at z < 1.5; the results also reveal luminosity-dependent dominance: the optical is disk-dominated above ~$10^{45}$–$10^{46}$ erg s$^{-1}$ and the IR is torus-dominated above ~($1.6$–$2$)×$10^{45}$ erg s$^{-1}$, depending on redshift. The paper concludes with two empirical OPT–IR calibrations and argues that, in the high-luminosity regime, simple photometric methods can yield reliable disk and torus luminosities, supporting the use of the OPT–IR relation as a cosmological probe for bright quasars.

Abstract

The main objective is to calibrate the OPT-IR luminosity relation for quasars, focusing on accurate estimations of dusty torus and accretion disk luminosities. We analyzed contaminations related to host galaxies, particularly from polar dust, the interstellar medium, and stellar emission that affect the optical and infrared. We used a sample of nearly 400 quasars with photometrical observations and spectroscopical redshift divided into four redshift bins (0.7-2.4). Full spectral energy distribution (SED) fitting was performed with the CIGALE code, and results were compared with simplified photometric luminosity estimates. The impact of non-active galactic nucleus components and the role of polar dust in the fitting process were assessed. We show that for sources with a disk luminosity above 10^45 [erg/s], the photometric estimates are consistent with SED-based values. While polar dust contributes marginally to luminosity, its presence significantly alters SED fitting, particularly the torus opening angle and cold dust properties. In the optical domain, stellar emission is the dominant contamination. In the infrared, disk emission and cold dust play major roles. We propose two empirical calibrations for the OPT-IR relation. We conclude that the optical band is dominated by the accretion disk component above 10^45 or 10^46 [erg/s] depending on redshift, while IR luminosity is dominated by the dusty torus emission above 1.6 $\times$ 10^45 or 2 $\times$ 10^46 [erg/s] depending on the redshift. In this high-luminosity regime, simplified photometric methods yield reliable disk and torus luminosity estimates. The aim of the analysis we present is to test the parameter space in order to build a well behaving OPT-IR nonlinear luminosity relation for quasars that could serve as a cosmological probe.

Figures (2)

• Figure 1: Flowchart showing the data selection process as described in Sect. \ref{['sec:Data']}. The $\chi^2$ values were calculated for the best-fit SED in CIGALE. The L$_{disk}$ values are from the best-fit estimation of the SKIRTOR model, as described in Sect. \ref{['sec:estimates_lum']}.
• Figure 2: Comparison of luminosities estimated from 1) our photometric method (presented on X-axis) and 2) CIGALE estimates without and with contaminations for 4 redshift bins. The dashed line represents the 1:1 relation, while the solid black line corresponds to the best fitted OLS regression. In the fourth row, outliers were identified as objects lying below the $1\sigma$ threshold (indicated by green lines) and are marked with star symbols. These outliers were excluded, and a second OLS regression was fitted to the remaining data points (gray dash-dot line). Additionally, in the fourth row, the red dashed line indicates the luminosity cut described in Section \ref{['sec:Lum_Cut']}. The $\sigma$ is the spread of the relation.