A Pair of Early- and Late-Forming Galaxy Cluster Samples: a Novel Way of Studying Halo Assembly Bias Assisted by a Constrained Simulation

TL;DR

This paper introduces a novel observational approach using constrained simulations to detect halo assembly bias by comparing early- and late-forming galaxy clusters with similar mass but different large-scale biases.

Contribution

It presents a new method to identify assembly bias in galaxy clusters through matched halos in constrained simulations, overcoming previous observational challenges.

Findings

Detected a ~3σ difference in large-scale bias between early- and late-forming clusters.

Validated the assembly bias signature using properties of cluster galaxies and their spatial distribution.

Established a pathway for future observational detection of assembly bias using observed cluster samples.

Abstract

The halo assembly bias, a phenomenon referring to dependencies of the large-scale bias of a dark matter halo other than its mass, is a fundamental property of the standard cosmological model. First discovered in 2005 from the Millennium Run simulation, it has been proven very difficult to be detected observationally, with only a few convincing claims of detection so far. The main obstacle lies in finding an accurate proxy of the halo formation time. In this study, by utilizing a constrained simulation that can faithfully reproduce the observed structures larger than $2\,$Mpc in the local universe, for a sample of 634 massive clusters at $z\le 0.12$, we find their counterpart halos in the simulation and use the mass growth history of the matched halos to estimate the formation time of the observed clusters. This allows us to construct a pair of early- and late-forming clusters, with similar mass as measured via weak gravitational lensing, and large-scale bias differing at $\approx 3\sigma$ level, suggestive of the signature of assembly bias, which is further corroborated by the properties of cluster galaxies, including the brightest cluster galaxy, and the spatial distribution and number of member galaxies. Our study paves a way to further detect assembly bias based on cluster samples constructed purely on observed quantities.