Small hosts, big appetites: unveiling rapid and early low-mass black hole growth in cosmological zoom-in simulations of dwarf galaxies

Abstract

Dwarf galaxies are ideal laboratories to probe the interplay between galaxy formation and the growth of black holes (BHs) in the early Universe. Mounting observational evidence reveals the presence of BHs in low-mass galaxies across cosmic time, with $\textit{JWST}$ uncovering a likely population of $\textit{overmassive}$ BHs at $2 \lesssim z \lesssim 11$. Simulations struggle to reproduce this high-redshift regime, motivating revisions to models of BH accretion and feedback from active galactic nuclei (AGN). To address this, we present high-resolution cosmological zoom-in simulations of a dwarf galaxy based on FABLE physics, introducing novel sink-based BH accretion models and relaxing the fiducial assumption of strong supernova feedback. BHs accrete more efficiently in the sink-based runs compared to the `traditional' Bondi-based counterparts, with AGN feedback leading to early, rapid quenching maintained by fast, hot and metal-enriched outflows. These outflows pollute the outer circumgalactic medium, yielding flat metallicity gradients down to $z=0$. We further assess the performance of two widely used virial estimators and find significant departures from the true dynamical mass, especially during the high-redshift dwarf assembly. Since our galaxy is dark-matter-dominated at all times and radii, BH growth, tied to the baryon cycle, shows no clear correlation with global dynamical properties. Efficient AGN feedback, produced by overmassive BHs relative to extrapolated local $M_\bullet - M_\star$ relations, indicates that dormant BHs residing in local, quenched dwarfs might be the relics of some of the high-redshift $\textit{JWST}$ BHs.

Figures (12)

• Figure 1: Zoom-in sequence of the multi-component structure of our simulated dwarf at $z \sim 3$ for the Sink+Low-z simulation run. From left to right: DM surface density ($\Sigma_\mathrm{DM}$) projection, gas temperature ($T_\mathrm{gas}$) projection and slice of the Voronoi mesh, all centred on the position of the BH. The projection dimensions are $(360 \; \mathrm{ckpc}/h)^3$, $(3 \times R_\mathrm{vir})^3$ and $(3 \times R_\mathrm{sink})^2$, respectively. In the second panel, the white circle marks the virial radius. In the third panel, the white circle indicates the sink radius, i.e. $R_\mathrm{sink} = 0.387 \; \mathrm{ckpc}$, which represents the region from which the BH can accrete gas in all sink-based runs. The black filled circle at the centre of all panels marks the position of the BH particle. Our simulated dwarf is located within a DM filament and is surrounded by numerous smaller-mass subhaloes. The middle panel shows an AGN-driven hot bubble, which expands well beyond the virial radius. The colour-coding in the last panel illustrates the sink-particle BH accretion model described in Section \ref{['sec: methods sink-particle model']}, in which gas cells in the innermost regions of the galaxy contribute preferentially to BH accretion according to the Gaussian kernel $\omega(r)$ (see Equation (\ref{['eq: Gaussian kernel']})).
• Figure 2: Large-scale surface density (upper panels) and temperature (lower panels) 2D projections for two representative runs, namely the Bondi+Low-z run and the Sink+Low-z run. The redshifts $z$ and times $t$ of the selected snapshots are indicated in the upper left-hand corner of the upper and lower panels, respectively. The projection dimensions are fixed at $(100 \; \mathrm{ckpc}/h)^3$. The region delimited by the virial radius is represented as a white dashed circle. Note that the virial radii for the selected redshifts approximately correspond to $R_\mathrm{vir} \sim 9, 11,\, 17,\, 21,\, 33$ kpc. In the Sink+Low-z run, the AGN is able to heat and expel gas well beyond the virial radius, suppressing cosmic inflows.
• Figure 3: Evolution of BH and host galaxy properties as a function of cosmic time, for all simulation runs. Left-hand panels: From top to bottom, the panels show the BH accretion rate $\dot{M}_\bullet$, the Eddington fraction $f_\mathrm{Edd}$ and the BH mass $M_\bullet$ against cosmic time. Right-hand panels: From top to bottom, the panels show the SFR, the integrated gas mass $M_\mathrm{gas}$ and the integrated stellar mass $M_\star$, all calculated within twice the stellar half-mass radius, against cosmic time. All quantities are binned over $\Delta t = 0.3 \, \mathrm{Gyr}$ to even out small-scale fluctuations and emphasise overall trends. With the exception of the Sink+FF runs, other sink-based runs are able to drive gas out of the galaxy at a redshift that depends on the specific parametrisation. After being depleted, the gas reservoir is temporarily and partially replenished before being depleted again by the next feedback burst. Efficient AGN runs are thus able to inhibit the star formation activity of the dwarf, which remains quenched at low redshifts.
• Figure 4: AGN active fraction as a function of cosmic time $t$, for all simulation runs. The AGN active fraction is defined as the fraction of each time bin of width $\Delta t = 0.3 \; \mathrm{Gyr}$ during which the Eddington ratio satisfies $f_\mathrm{Edd} \geq 0.01$. Among the sink-based runs, the Sink+FF runs generally display the highest active fraction -- especially at low redshifts -- as a result of the ability of the BH to sustain self-regulated growth across the entire redshift evolution. Other sink-based runs also display significant AGN active fraction at high redshifts, while due to efficient quenching of their host, BHs in these runs are largely dormant at low redshifts.
• Figure 5: Main panel: Evolution of the dynamical mass as a function of cosmic time for the Sink+Low-z setup. The dynamical mass is defined as $M_\mathrm{dyn} \equiv 2 \times M(<R_\mathrm{e})$, where $R_\mathrm{e}$ is the 2D projected stellar half-mass radius. At each snapshot, we perform 100 random projections, measure the projected axis ratio $q$, and carry out Sérsic fits to determine the Sérsic index $n$ and $R_\mathrm{e}$. We then compare the true dynamical mass measured directly from the simulations ($M_\mathrm{dyn}^\mathrm{sim}$; purple solid line) with the virial estimators from van2022mass (green dash-dotted line) and walker2009universal (pink dashed line). We label these inferred masses as $M_\mathrm{dyn}^\mathrm{obs,\star}$ as they were derived by setting $\sigma \equiv \sigma_\star$ (see Equation (\ref{['eq: vdW M_dyn']}) and Equation (\ref{['eq: Walker M_dyn']})). Shaded regions around each line correspond to the 16th–84th percentile range, capturing the scatter across the projections. For comparison, we also show the 16th–84th percentile ranges obtained by setting $\sigma \equiv \sigma_\mathrm{gas}$ in the same estimators, which we denote as $M_\mathrm{dyn}^\mathrm{obs, gas}$. These are indicated as a light-grey shaded band for the van2022mass estimator and as an unfilled band bounded by light-grey lines for the walker2009universal estimator. Inset: Radial profiles of the 3D mass density $\rho$ for DM (dotted line), stars (solid line), and gas (dashed line) within the virial radius at $z \sim 6$. The median effective radius ($R_\mathrm{e}$) and the 3D DM half-mass radius ($R_\mathrm{hm,DM}$) are marked with vertical grey dashed and dash-dotted lines, respectively. The 16th and 84th percentiles for $R_\mathrm{e}$ are shown as a vertical grey shaded band. As our simulated dwarf falls well outside the stellar mass range of the galaxies used for calibrating the van2022mass estimator, extrapolating this calibration to such low-mass, DM-dominated systems is likely invalid and leads to a systematic overestimation of $M_\mathrm{dyn}$. The walker2009universal estimator recovers the true dynamical mass to within $\sim 10-15 \%$, albeit only for $z \lesssim 2$.