Galaxy bias and primordial non-Gaussianity: insights from galaxy formation simulations with IllustrisTNG

TL;DR

This study uses galaxy formation simulations to investigate how large-scale perturbations and primordial non-Gaussianity influence galaxy bias, revealing that common models underestimate the bias related to PNG, thus affecting constraints on early universe conditions.

Contribution

It introduces a refined model for galaxy bias parameters considering stellar mass dependence, improving upon the traditional universality relation in the context of primordial non-Gaussianity.

Findings

The universality relation $b_\phi = 2\delta_c(b_1 - 1)$ poorly describes galaxy bias.

A new relation $b_\phi(M_*) = 2\delta_c(b_1(M_*) - p)$ better fits the data.

Existing models may underestimate the constraining power on local PNG.

Abstract

We study the impact that large-scale perturbations of (i) the matter density and (ii) the primordial gravitational potential with local primordial non-Gaussianity (PNG) have on galaxy formation using the IllustrisTNG model. We focus on the linear galaxy bias $b_1$ and the coefficient $b_\phi$ of the scale-dependent bias induced by PNG, which describe the response of galaxy number counts to these two types of perturbations, respectively. We perform our study using separate universe simulations, in which the effect of the perturbations is mimicked by changes to the cosmological parameters: modified cosmic matter density for $b_1$ and modified amplitude $\mathcal{A}_s$ of the primordial scalar power spectrum for $b_\phi$. We find that the widely used universality relation $b_\phi = 2\delta_c(b_1 - 1)$ is a poor description of the bias of haloes and galaxies selected by stellar mass $M_*$, which is instead described better by $b_\phi(M_*) = 2\delta_c(b_1(M_*) - p)$ with $p \in [0.4, 0.7]$. This is explained by the different impact that matter overdensities and local PNG have on the median stellar-to-halo-mass relation. A simple model of this impact allows us to describe the stellar mass dependence of $b_1$ and $b_\phi$ fairly well. Our results also show a nontrivial relation between $b_1$ and $b_\phi$ for galaxies selected by color and black hole mass accretion rate. Our results provide refined priors on $b_\phi$ for local PNG constraints and forecasts using galaxy clustering. Given that the widely used universality relation underpredicts $b_\phi(M_*)$, existing analyses may underestimate the true constraining power on local PNG.