Igniting galaxy formation in the post-reionization universe

TL;DR

This study uses the FIREbox large-volume cosmological simulation with FIRE-2 physics to identify and characterize halos that ignite star formation after reionization. By comparing recently-ignited halos (stellar age $t^{\star}_{\rm age} \leq 100\,\mathrm{Myr}$) to carefully matched failed halos, the authors quantify how the interstellar medium and halo structure govern ignition. They find that all recently-ignited halos exhibit cold-dense gas enhancements, and a majority also show elevated halo concentration; together these factors strongly distinguish ignition from failure, while HI content alone is not predictive after controlling for mass/redshift. Their results support a picture in which gas compression in more concentrated halos promotes cooling and star formation, with ignition potentially occurring as late as $z\sim 2$, offering opportunities for observational tests of this process in the near future. The work also provides a quantitative framework for the dark halo occupation fraction across cosmic time, highlighting the importance of high-resolution, volume-complete simulations to interpret the faint-end galaxy population.

Abstract

It is widely believed that the ultraviolet background produced during the epoch of reionization conspires against the formation of low-mass galaxies. Indeed, this mechanism is often invoked as a solution to the so-called `missing satellites problem.' In this paper we employ FIREbox, a large-volume cosmological simulation based on the Feedback In Realistic Environments (FIRE-2) physics model, to characterize the mechanisms governing galaxy ignition in the post-reionization era. By carefully matching recently-ignited halos (with stellar ages below $100$ Myr at the time of selection) to halos that failed to form any stars, we conclude that the presence of cold-dense gas and halo concentration help incite the process of galaxy formation. Concretely, we find that $100\%$ of recently-ignited halos experience cold-dense gas enhancements relative to their matched failed counterparts. Likewise, approximately $83\%$ display enhancements in both cold-dense gas and Navarro-Frenk-White concentration ($c_{\rm NFW}$), while the remaining $\sim17\%$ exhibit enhanced cold-dense gas content and suppressed $c_{\rm NFW}$ values. Lastly, our simulation suggests that galaxy ignition can occur as late as $z=2$, potentially allowing us to observationally catch this process `in the act' in the foreseeable future.

Figures (9)

• Figure 1: Archetypal recently-ignited and (matched) failed halos in our simulation. The panels show two halos at $z=2$ matched to have nearly identical virial and gas masses ($M_{\rm vir} \sim 1.29 \times 10^9 M_\odot$ and $M_{\rm gas} \sim 1.17 \times 10^8 M_\odot$). The field of view encompasses the virial radius. Top panels: Recently-ignited halo with a single star particle, denoted as a white star-shaped symbol, with stellar mass $M_\star = 6.26 \times 10^4 M_\odot$ and stellar age $t^{\star}_{\rm age}= 34.6$ Myr. Bottom panels: Matched failed halo. Left-to-right panels: Dark matter, gas and HI gas surface mass density maps. Although matched in $M_{\rm vir}$, $M_{\rm gas}$ and $z$, the recently-ignited halo is more concentrated than its failed counterpart ($c_{\rm NFW} = 6.59$ versus $2.36$). The ignited halo exhibits a dense, complex horseshoe-shaped central gas concentration, particularly salient in the HI component. The gas in the matched failed halo is extended and diffuse, with a barely detectable central HI contribution. This figure shows the only ignited halo in our sample found at $z=2$ with stellar age below $100$ Myr, with no other such halos identified at lower redshifts.
• Figure 2: Ignited halos on the cosmic web across epochs. Top-left-to-bottom-right:$z=0, 2, 4$ and $6$, indicated by the large green annotations. On the background, each panel shows a surface-density map of our comoving simulation volume. The white symbols represent the location of ignited halos, color coded by stellar age, with the palette centered at $t^{\star}_{\rm age} = 100$ Myr. Symbol sizes increase with decreasing stellar age to emphasize recently ignited halos, which become more numerous with increasing $z$. We also differentiate objects with stellar ages below $100$ Myr from the rest by switching from circular to star-shaped symbols.
• Figure 3: Halo abundances for various populations in our simulation (see Table \ref{['table:samples']} for definitions). Violet, black and gray lines represent all, all failed, and failed (gaseous) halos, while green and blue denote ignited and recently-ignited (gaseous) halos. Note: we exclude subhalos from our analysis. Left-to-right panels: $z = 0, 3$ and $6$. Top panels: Halo mass functions. Bottom panels: Halo fractions (HF), defined here as the ratio of the halo mass function of a given population and that of all halos. For the failed, ignited, and recently-ignited populations, the normalization of the halo mass function diminishes with cosmic time. Both halo mass function and HF shift to higher $M_{\rm vir}$ values at lower redshift. This mass-shift aside, the HF levels remain stable with cosmic time for every population -- except for the recently-ignited population, which diminishes quickly with cosmic time, becoming negligible below $z=3$ (i.e., only the blue curve plummets with decreasing redshift). The horizontal dotted, dot-dashed and dashed black lines indicate the $M_{\rm vir}-$thresholds where the various HFs reach $1\%$, $10\%$ and $50\%$ respectively.
• Figure 4: The evolution of the failed, ignited and recently ignited halo populations (gray, green and blue respectively -- see Table \ref{['table:samples']} for definitions). Bands represent running medians with widths spanning the first and third quartiles. Recall that we focus exclusively on centrals with a resolved gas component, and our threshold for recent ignition is $t^{\star}_{\rm age}=100$ Myr. Upper-left to bottom-right: number density, virial mass, gas mass, HI-gas mass, fraction of mass in gas, and fraction of gas mass in HI. The dashed horizontal black line denotes $\Omega_b/\Omega_m$, the universal gas fraction. Although approximately one order of magnitude less numerous, the ignited population follows their failed counterparts across redshifts (see text for a more thorough discussion of their differences). The recently-ignited population behaves differently. Their abundance plummets faster with cosmic time. Below $z\sim5$, these halos have larger $M_{\rm vir}$ values. However, their gas masses and gas fractions, HI-content and HI-to-gas ratios increase dramatically with cosmic time relative to their older and failed counterparts. Recently-ignited halos tend to have copious gas reservoirs, particularly in the HI phase.
• Figure 5: Stellar ages and HI content of ignited and recently-ignited halos (see Table \ref{['table:samples']} for definitions). Top panel: Halo abundance versus stellar ages for the ignited halo population. Each color refers to a different epoch, as indicated by the key. The vertical dashed lane at $t^{\star}_{\rm age} = 100$ Myr demarcates the threshold dividing the recently-ignited halos and their older counterparts. In general, ages increase and abundances decrease with cosmic time. Bottom panels: Stellar ages of ignited (black circles) and recently-ignited ( star-shaped symbols) halos as a function of virial mass, color coded by the HI-to-total gas mass ratio, at $z = 3$ and $6$ ( left-to-right). The horizontal solid and dashed black lines indicate the age of the universe and our $100-$Myr threshold. By $z=3$, the ignited population comes in two varieties: an old and HI-deficient set and a slightly younger sample with a mix of HI fractions, with only a few halos with stellar ages below $100$ Myr. Ignited halos at $z=6$ also break into two populations, though the older sample has stellar ages below $1$ Gyr, with intermediate gas fractions and a mix of intermediate and rich HI content. The younger tail is heavily populated with HI-rich objects.