FLAMINGO: Tracing the co-evolution of hot gas and black holes in galaxy groups and clusters

TL;DR

This work investigates why halo gas fractions in groups and clusters scatter around the median gas fraction for a given halo mass. Using the FLAMINGO hydrodynamical simulations, it dissects the roles of central and satellite supermassive black holes (BHs) and traces their growth histories to explain present-day gas content across halo masses. It finds a mass-dependent coupling: central BHs are anti-correlated with $f_{ ext{gas}}$ at $M_{500}\,<\,10^{13}~M_ ext{sun}$ but positively correlated at higher masses, driven by early gas ejection and late gas reaccretion; satellite BHs contribute an anti-correlation and exhibit a proximity effect with centrals. The results tie the scatter to early assembly history and BH growth timing, offering observable predictions for upcoming X-ray/SZ surveys and informing feedback implementations in large-scale simulations.

Abstract

The gas mass fraction of galaxy groups and clusters is a key physical quantity for constraining the impact of feedback processes on large-scale structure. While several modern cosmological simulations use the gas fraction-halo mass relation to calibrate their feedback implementations, we note that this relation exhibits substantial intrinsic scatter whose origin has not been fully elucidated. Using the large-volume FLAMINGO hydrodynamical simulations, we examine the role of both central and satellite supermassive black holes (BHs) in shaping this scatter, probing higher halo masses than previously possible. For haloes with M500 < 10^13 Msun, we find that central BH mass correlates strongly and negatively with gas fraction, such that higher BH masses give rise to lower gas fractions at fixed halo mass, consistent with previous studies. Interestingly, however, for 10^13 Msun < M500 < 10^14.5 Msun the correlation reverses and becomes positive, with overmassive BHs residing in haloes with above-average gas fractions. By tracing progenitor BHs and haloes through cosmic time, we show that this behaviour is driven by the expulsion and subsequent re-accretion of halo gas, regulated by the timing of BH growth and feedback. Specifically, haloes that collapse earlier form BHs earlier, leading to earlier gas expulsion and re-accretion and a high gas fraction compared to haloes of the same present-day mass that formed later. Our results demonstrate that present-day scatter in the gas fraction-halo mass relation is strongly shaped by the early growth history of BHs and their haloes, a prediction that can be tested with future observational measurements.

Figures (9)

• Figure 1: The present-day supermassive black hole mass, ${M}_{\text{BH}}$, as a function of halo mass, ${M}_{500}$, for the total (top), central (centre), and satellite (bottom) BH populations in the FLAMINGO high-resolution $(1 \ \text{Gpc})^3$ volume. Note that the satellite BH mass is defined as the sum of the masses of all satellite BHs within ${R}_{500}$. A given hexbin contains at least three data points and is coloured by the median mass of the most massive BH (${M}_{\text{mmBH}}$) of the haloes that occupy that space in the relation. The solid black line represents the median trend and the grey shaded region encompasses the 10-90th percentile range. The inset plot shows the median satellite BH contribution to the total BH mass budget as a function of halo mass, with the interquartile range indicated by the shaded region. Each population exhibits large scatter at fixed halo mass, a key driver of variations in halo gas fractions discussed later. At the most massive cluster scales, satellite BHs contribute a total mass on par with the central BH, highlighting their potential role in shaping the halo gas content.
• Figure 2: The present-day relation between halo gas fraction, $f_{\text{gas}}$, and halo mass, ${M}_{500}$. Each hexbin is colour-coded by the median BH mass ratio, ${M}_{\text{BH}} / {M}_{\text{BH, median}}$, for the total BH population. This quantifies the scatter in BH mass by taking the ratio between a halo's total BH mass and the median total BH mass at fixed halo mass. The dot-dash lines encloses the 10-90th percentile range. The colour bar limits correspond to the 16th–84th percentile range of median ${M}_{\text{BH}} / {M}_{\text{BH, median}}$ values for haloes with ${M}_{500} \geq 10^{12.5} \text{M}_{\odot}$. The solid line shows the running median of $f_{\text{gas}}$ in bins of 0.1 dex in ${M}_{500}$, while the horizontal dashed black line represents the universal baryon fraction in FLAMINGO. Marginal histograms of $f_{\text{gas}}$ and ${M}_{500}$ are displayed to the right of and above the main panel, respectively. The Spearman rank correlation coefficient, $\rho$, quantifying the $f_{\text{gas}}$-${M}_{\text{BH}} / {M}_{\text{BH, median}}$ relation, is indicated below the main panel. This relation exhibits substantial intrinsic scatter, which is reduced following the onset of AGN feedback. At low halo masses, a clear negative correlation emerges: overmassive BH populations reside in haloes with below-average gas fractions, and vice versa. Above ${M}_{500} \approx 10^{13} \text{M}_{\odot}$, however, the trend reverses, with overmassive BH populations instead associated with haloes of higher-than-average gas fractions.
• Figure 3: The present-day $f_{\text{gas}}-\text{log}_{10}({M}_{500})$ relation colour-coded by the median ${M}_{\text{BH}} / {M}_{\text{BH, median}}$ for the central (top) and satellite (bottom) BH populations. The colour bar limits are set by the interquartile range of hexbin medians for haloes above ${M}_{500} \leq 10^{12.5} M_{\odot}$ for each BH population. The black solid line indicates the running median, the horizontal dashed black line represents the universal baryon fraction in FLAMINGO. The running Spearman rank correlation coefficient is displayed below each panel. Consistent with previous work, haloes with ${M}_{500} \lessapprox 10^{12.5} \, \text{M}_{\odot}$ that host overmassive central BHs tend to have below-average gas fractions, and vice versa. Beyond this halo mass, however, the correlation becomes positive, where overmassive central BHs reside in haloes with higher gas fractions. The opposite trend is evident in satellite BHs, suggesting an anti-correlation between central and satellite BH mass: haloes with overmassive central BHs are likely to have an undermassive population of satellite BHs.
• Figure 4: Relations between halo gas fraction, $f_{\text{gas}}$, and halo mass, ${M}_{500}$, across the redshift range $3 \leq z \leq 0$, colour-coded by the median ${M}_{\text{BH}} / {M}_{\text{BH, median}}$ for central BHs at $z = 0$. The colour-bar limits are set by the interquartile range of hexbin medians for haloes with ${M}_{500} \geq 10^{12.5} \text{M}_{\odot}$. The solid black curves show the running median, calculated in bins of 0.1 dex. The corresponding Spearman rank correlation coefficient, calculated in bins of 0.1 dex, is shown below each main panel. The horizontal dashed line shows the universal baryon fraction in FLAMINGO, the dot-dashed enclosed region in the bottom right panel shows haloes within the mass bin $10^{13.75} \text{M}_{\odot} \leq {M}_{500} < 10^{14}\text{M}_{\odot}$, whose BHs we track through cosmic time in Section \ref{['sec:Black_Hole_Tracking']}. For $z < 1.5$, a strong negative correlation is evident, with all overmassive BHs residing in haloes with low gas fractions, independent of halo mass. For $1.5 \leq z \leq 0.25$, this trend transitions, as an increasing fraction of massive haloes with overmassive central BHs exhibit elevated gas fractions. By $z=0$, haloes above ${M}_{500} \approx 10^{13}\text{M}_{\odot}$ predominantly host overmassive central BHs and have above-average gas fractions. In Section \ref{['sec:Black_Hole_Tracking']}, we show that this reversal is driven by the ejection and subsequent re-accretion of halo gas.
• Figure 5: The redshift evolution of central BHs and their haloes within the $z=0$ mass range $\mathrm{13.75 \leq log_{10}({M}_{500}) \leq 14}$ split into the 90th (upper, red) and 10th (lower, navy) percentiles of the $f_{\text{gas}}$ distribution at $z=0$. Solid lines show the median trend per sample, and shaded regions enclose the 16-84th interquartile range. (Top) The evolution of median halo mass, ${M}_{500}$, with redshift. (Centre) BH mass, ${M}_{\text{BH}}$, as a function of redshift. Dashed lines show the upper (pink) and lower (light blue) median satellite BH population masses. Dotted lines indicate the total accreted BH mass per sample. The inset panel shows the normalised number of BHs per redshift interval against the redshift at which these BHs were formed, $z_\text{seed}$. (Bottom) Gas fraction within ${R}_{500}$, $f_{\text{gas}}$, as a function of redshift. Haloes that collapse earlier form central BHs earlier, leading to earlier gas expulsion and re-accretion, and resulting in a relatively high gas fraction compared to haloes of the same present-day mass that formed later.