A close look at the black hole masses and hot dusty toruses of the first quasars with MIRI-MRS

TL;DR

This work leverages JWST MIRI-MRS and NIRSpec to study the four most distant luminous type-1 quasars, updating BH masses with rest-frame optical/IR tracers H$\alpha$, Pa$\alpha$, and Pa$\beta$ and confirming consistency with Mg II-based estimates. It simultaneously detects and models hot dust in accretion tori using SKIRTOR, finding a face-on geometry with a compact, centrally concentrated dust distribution and a torus mass of $(1-4)\times10^{6}\,M_\odot$, comprising a small fraction of host dust and potentially depleting on $\sim$few Myr timescales at current accretion rates. The broad-line region shows high-density, high-ionization conditions without evidence for extinction, and C$\,$IV-based BH masses remain biased high relative to IR tracers. Collectively, the results indicate no major deviations of $z>7$ quasar BH masses or accretion processes from their $z<3$ analogs, though SMBH growth under canonical radiative efficiency $\epsilon\approx0.1$ remains challenging unless radiative efficiency varies over time or accretion histories include low-$\epsilon$ phases.

Abstract

The presence of supermassive black holes (SMBHs, $M_\text{BH}\sim10^9 M_\odot$) at $z>7$ remains a puzzle. While their existence appears to require exotic formation or growth processes, it is possible that BH mass estimates are incorrect due to differences from the low-$z$ quasars where BH mass scaling relations are calibrated. In this work, we employ JWST MIRI-MRS spectroscopy to measure the rest-frame optical/IR properties of the four highest-redshift known luminous type-1 quasars at $7.08\leq z<7.64$. We use three new broad lines to measure updated BH masses, H$α$, Pa$α$ and Pa$β$, finding them to be in the range $(4-15)\cdot10^8 M_\odot$. Our black hole mass estimates from all tracers agree with each other and with previous, less accurate, ground-based measurements based on MgII. The flux ratios of the H lines deviate from expectations for case A and B recombination in the same way as in $z<3$ quasars, indicating similar physical conditions in the Broad Line Region. Rest-frame near-IR continuum emission from a hot dusty torus surrounding the accretion disc is unambiguously detected in all four objects. We model the emission with SKIRTOR and constrain the inclination (face-on) and the opening angle ($θ=40-60^\circ$) of the tori. These constraints are consistent for the four objects and with expectations from luminous quasars. We estimate a total dust mass $(1-4)\cdot10^6 M_\odot$ in the tori, corresponding to $(0.2-7)\%$ of the total dust in the quasar host galaxies. Given observed accretion rates, these SMBHs will deplete their tori in only $\sim5$ Myr. Overall, we confirm that $z>7$ SMBHs in quasars could not have grown from stellar-remnant BHs if the radiative efficiency of accretion is $10\%$. We also find no evidence that inferred BH masses and accretion processes in $z>7$ quasars differ significantly from their near-identical counterparts at $z<3$.

Figures (12)

• Figure 1: MRS spectra of the four $z>7$ quasars in our sample; flux is in black and uncertainties are in grey. Channel 4 is not shown for J1342+0928 because the observations were badly affected by cosmic ray showers, prohibiting the extraction of a spectrum. Vertical lines denote the location of detected emission lines.
• Figure 2: Broad H$\alpha$ lines detected in the four $z>7$ quasars. Blue shows the broad components (BCs). The broadest line components for J0313$-$1806 and J1120+0641 have FWHM$\sim12000$ km s${}^{-1}$. In J1342+0928, CR shower residuals affected the observations (see e. g. around $\lambda_\text{obs} = 5.8\mu$m) and only one broad component could be identified.
• Figure 3: Black hole masses calculated from five different single-epoch virial mass estimators: H$\alpha$, Pa$\alpha$ and Pa$\beta$ (this work), C IV and Mg II (this work, Yang21Farina22). Measurement uncertainties are shown as coloured error bars while grey error bars reflect the intrinsic (systematic) scatter in the mass estimators. Despite implementing a correction of blueshift Coatman17, C IV remains a poor and potentially biased estimator. The other tracers agree within scatter, indicating no bias from ground-based measurements of the Mg II emission line.
• Figure 4: Flux ratios of the broad H$\alpha$, Pa$\alpha$ and Pa$\beta$ lines. Tracks show the output of a simple Cloudy model with parameters of broad line region density $\log n_\text{H} (\text{cm}^{-3}) = 10, 12, 14$ from light to dark, and ionisation parameter increasing in the direction of the arrows over $-7<\log U<0$. Hydrogen BLRs are not consistent with case B nor case A recombination, but not in a direction which would correspond to extinction: the dark blue point shows expectations for case B with E$($B$-$V$)=0.5$.
• Figure 5: Optimal fit to the torus emission for the four quasars (red). The NIRSpec spectra (blue) are used to estimate the SED of the accretion disc, which is then processed through SKIRTOR to predict the torus emission. The fit is conducted over the MRS spectra (black) after subtracting the detected broad emission lines.