The rest ultraviolet to infrared spectral energy distributions of heavily reddened quasars are "V-shaped" and hot-dust poor

TL;DR

This study constructs rest-frame UV-to-IR SEDs for 63 heavily reddened quasars across $0.7<z<2.7$ and $0.4< E(B-V)<1.8$, showing an ubiquitous UV excess best described by a two-component model combining a dust-reddened quasar and a scattered/host contribution. HRQs are markedly hot-dust poor relative to blue quasars, with a mean $\langle L_{Dust}/L_{Disk}|_{2μm} \rangle\approx 1.6\pm0.8$ versus $4.4\pm2.1$ for blue SDSS quasars, implying torus-scale dust depletion likely driven by AGN radiative feedback; the inferred accretion rates are high, with several objects in the super-Eddington regime using a bolometric correction $k_{Bol}=25\{\lambda L_{\lambda}(3000Å)/10^{42}\}^{-0.2}$. A pan-chromatic HRQ composite reveals similar rest-optical SEDs to JWST’s Little Red Dots (LRDs) but bluer UV continua in LRDs, suggesting selection biases and phase differences in obscured AGN evolution. Overall, the work links extinction, scattering, and dust emission to a possible feedback-driven blowout phase, refines HRQ demographics, and motivates broader, less UV-biased HRQ selections to capture a fuller population of dust-obscured quasars.

Abstract

We present a rest-ultraviolet to infrared spectral energy distribution (SED) analysis of 63 heavily reddened quasars (HRQs) at redshifts z=0.7-2.7 and with dust extinctions E(B-V)=0.4-1.8. Our analysis demonstrates that SEDs with red optical and blue UV continua are very common in HRQs, with more than 82 per cent of the sample showing a UV-excess relative to the reddened quasar continuum. We model the SEDs by combining a reddened quasar and an unobscured scattered light component, though contributions from a star-forming host galaxy cannot be ruled out. The average scattering fraction is small (0.3 per cent). Higher scattering fractions are ruled out by the (i-K)=2.5 colour-cut used to select HRQs which pre-dates the discovery of the JWST "Little Red Dot" (LRD) population. Hence, LRDs generally have bluer UV continua. Nevertheless, four HRQs satisfy the LRD UV/optical continuum slope selections and are therefore massive, cosmic noon analogues of LRDs. Analysis of the near-infrared SEDs of HRQs reveals a deficit of hot dust relative to blue quasars, similar to what is observed in LRDs. This suggests HRQs trace a phase where strong AGN feedback processes eject dust from the inner torus. The UV scattering fraction of HRQs is weakly correlated with the amount of hot dust emission and anti-correlated with the line-of-sight extinction, E(B-V). This is consistent with the hot dust acting as the scattering medium, and the line-of-sight extinction being dominated by dust on interstellar medium scales in the host galaxy.

Figures (16)

• Figure 1: The redshift versus 3000Å continuum luminosity for blue SDSS quasars 2020MNRAS.492.4553R2021MNRAS.501.3061T2023MNRAS.523..646T and HRQs (coloured circles). The colour bar illustrates the $\rm E(B-V)$ of each HRQ.
• Figure 2: The Eddington-scaled accretion rate distributions of the HRQ sample (orange) and blue SDSS quasars 2020MNRAS.492.4553R2023MNRAS.523..646T. HRQs generally host higher accretion rates than their blue SDSS counterparts, with HRQs tending towards the super-Eddington regime.
• Figure 3: Example SED fits for HRQs where a UV excess is "confirmed" (top), "inconclusive" (middle) or "rejected" (bottom). The photometric data and associated uncertainties are presented in black. The best-fit SED model and photometry are presented by the blue line and triangles. The dust attenuated quasar component and photometry are presented by the orange line and triangles. In the "confirmed" case, there is clear evidence of a UV excess. In the "inconclusive" case, there is tentative evidence of a UV excess, but the wavelength coverage does not extend blue enough to constrain the model. In the "rejected" case, there is no evidence to support that the two-component model yields a better fit than a single-component reddened quasar SED.
• Figure 4: The scattering fraction, $F_{\rm{UV}}$, vs the $(i-K)_{\rm AB}$ colour, calculated from the best-fit SED models for the 42 HRQs that exhibit a statistically significant UV excess. The $(i-K)_{\rm AB} =2.5\,\rm mag$ colour selection is represented by dashed black line. There is a strong anti-correlation between $F_{\rm{UV}}$ and the $(i-K)_{\rm AB}$ colour. Therefore, this colour selection excludes HRQs with higher scattering fractions from the sample.
• Figure 5: The best-fit SED for the HRQ VIKJ2258-3219 at $z=0.879$ with the S0 galaxy template omitted (top) and with $fragal=0.05$ (bottom). The photometric data and associated uncertainties are presented in black. The best-fit SED model and photometry are indicated by the blue line and triangles. The dust-attenuated quasar component and photometry are indicated by the orange line and triangles. At near-infrared wavelengths the model is unable to reproduce the photometry in the top panel, suggesting that old stellar populations have a significant contribution to the SED in this region. Hence, the inclusion of the galaxy component improves the fit in the bottom panel.