Memoirs of mass accretion: probing the edges of intracluster light in simulated galaxy clusters

TL;DR

This study investigates how the outer intracluster light in massive galaxy clusters records the cluster's mass assembly history and merger events. Using high-resolution Symphony dark-matter simulations with the Nimbus star-tagging model, the authors measure two stellar splashback radii from the ICL outskirts and analyze their correlations with continuous MAH and discrete merger history. Employing the MultiCAM framework, they reconstruct individual MAHs from these radii and show that the radii outperform traditional proxies like halo concentration and stellar mass gaps, and are competitive with X-ray-based relaxedness indicators. They also assess the observability of these features with next-generation surveys, finding promising prospects for the secondary radius in deep imaging, and outline observational strategies and limitations for real data.

Abstract

The diffuse starlight extending throughout massive galaxy clusters, known as intracluster light (ICL), has the potential to be read as a memoir of mass accretion: informative, individual, and yet imperfect. Here, we combine dark matter-only zoom-in simulations from the Symphony suite with the Nimbus "star-tagging" model of the stellar halo to assess how much information about the mass assembly of an individual galaxy cluster can be gleaned from idealized measurements of ICL outskirts. We show that the edges of a cluster's stellar profile -- the primary (Rsp*1) and secondary (Rsp*2) stellar "splashback" radii -- are sensitive to both continuous mass accretion histories and discrete merger events, making them potentially powerful probes of a cluster's past. We find that Rsp*1 strongly correlates with the cluster's mass ~1 dynamical time ago, while Rsp*2 traces more recent mass accretion history to a slightly lesser degree. In combination, these features can further distinguish between clusters that have and have not undergone a major merger within the past dynamical time. We use both to predict realistic cluster mass accretion histories with the MultiCAM framework. These outer ICL features are significantly more sensitive to mass accretion and merger histories than the stellar mass gap and halo concentration, and perform comparably to the commonly used X-ray-based tracer of relaxedness, x_off. While our analysis is idealized, the relevant ICL features are potentially detectable in next-generation deep imaging of nearby clusters. This work highlights the promise of ICL measurements and lays the groundwork for more detailed forecasts of their power.

Figures (13)

• Figure 1: An illustration of the procedure of measurements and analysis as described in detail in \ref{['sec:Measure']}, \ref{['sec:continuous']}, and \ref{['sec:discrete']}. We measure the present-day primary and secondary stellar splashback radii of simulated clusters (see \ref{['sec:Measure']}) and probe how they inform the continuous mass accretion histories (MAH) (\ref{['sec:continuous']}) and discrete merger histories of the clusters (\ref{['sec:discrete']}). See the text in \ref{['sec:structure']} for a description of each panel and toy figure.
• Figure 2: Projected stellar density maps for ten example clusters out of our sample of 79 galaxy clusters illustrating the diversity of stellar distributions captured by the Nimbus star-tagging framework.
• Figure 3: An example of the stellar density maps, phase space distributions, and radial stellar mass density profiles and our separation of particle populations for a cluster following the methods described in \ref{['subsec:MeasureRsp']}. In the left column, we show the projected stellar mass density map of the cluster and indicate the relevant radii. In the central column, we show the density of stellar mass in radial velocity phase space separated into particles on infall, first orbit, and second orbit. As particles are only star-tagged after crossing $R_{\rm vir}$, the infall panel is lacking any stars outside that radius. In the right column, we show the profiles and corresponding log derivatives of particles on all orbits, on first orbit, and on second orbit. The dashed circles and vertical lines indicate the resulting measured primary and secondary stellar splashback radii for this cluster.
• Figure 4: Spearman's correlation coefficients between present-day cluster properties and the mass accretion history (MAH) as a function of time. Time $T$ is measured in units of number of dynamical times (where $T$= 0 indicates present-day) and corresponding scale factors are included in the upper $x$-axis for reference. The shaded bands indicate the jackknife-estimated standard error (see \ref{['subsec:correlation']} for further details). We plot our proposed tracers of the MAH (solid lines): the primary stellar splashback $R_{\rm sp\star, 1}$ and the secondary stellar splashback $R_{\rm sp \star, 2}$. Traditional tracers of the MAH are shown for comparison (dashed lines): the central offset $x_{\rm off}$, the concentration $c_{\rm vir}$, and the stellar mass gap $M_{\star \rm 1,2}$. $R_{\rm sp\star, 1}$ is highly correlated with $m(T)$ with a peak 1.16 dynamical times ago. $R_{\rm sp \star, 2}$ is strongly correlated to a slightly lesser degree and peaks more recently, 0.73 dynamical times ago. $x_{\rm off}$ is correlated similarly to $R_{\rm sp\star, 2}$, but $M_{\star \rm 1,2}$ and $c_{\rm vir}$ are both substantially less correlated with MAH than the stellar splashback tracers.
• Figure 5: Example true MAHs, 200 MultiCAM model reconstructed MAH samples, and medians of MultiCAM model reconstructed MAH samples for three clusters. MAHs are plotted as $m(T)$ as a function of $T$, time in units of dynamical times where $T = 0$ is present day and the corresponding scale factors are indicating in the upper $x$-axis. The three clusters shown reflect three percentiles of accuracy in model predictions (where lower percentiles indicate higher accuracy). Accuracy is quantified here by the root mean squared error (RMSE) between the median reconstructed MAH and the true MAH. The 10$^{\rm th}$ percentile cluster has a closer recovery of true MAH in its median reconstructed MAH than the 90$^{\rm th}$ percentile cluster.