How Massive Can a Population III Starburst Be? Simulating the First Galaxies with High Lyman-Werner Background

Abstract

Observing the first generation of Population III (Pop III) stars is one of the most demanding challenges in astronomy. Indeed, Pop III stars are expected to predominantly form within faint minihalos at early times with a top-heavy initial mass function, resulting in efficient metal enrichment and a fast transition to Pop II-dominated systems. However, recent surveys with the James Webb Space Telescope (JWST) have identified galaxies at the end of the Epoch of Reionization (EoR) with possible signatures of significant Pop III star formation even at these later times. We here explore the physical conditions required to produce massive Pop III starbursts during the EoR, using cosmological radiation-hydrodynamic zoom-in simulations. We specifically focus on galaxies with a virial (dynamical) mass of $M_{\rm vir} \approx 10^{8} M_{\odot} $ at $7 \lesssim z \lesssim 8$, i.e., the atomic-cooling halos that could be potential sites for such maximal Pop III starbursts. In particular, we vary the strength of Lyman-Werner (LW) background radiation up to $J_{\rm LW} \leq 10^4J_{21}$, further imposing a high star formation efficiency (up to $ε_{\rm ff} = 1.0$). Our results show that Pop III starbursts, observable in strongly-lensed survey fields like GLIMPSE, can occur in the presence of a sufficiently high LW flux (with $\gtrsim 10^3J_{21}$), leading to delayed, but intense Pop III star formation. However, even for such high LW fluxes, the Pop III starburst mass is limited to $M_{\star, \rm Pop~III} <10^6 M_{\odot}$, as strong internal metal enrichment occurs after the first Pop III supernova explosions within the simulated galaxies. While the conditions favoring observable Pop III starbursts are expected to be rare, we anticipate that future and ongoing large-volume surveys leveraging gravitational lensing, such as VENUS, will detect multiple cases of Pop III starbursts in the EoR.

Figures (10)

• Figure 1: Mass evolution within our simulated galaxies. The left panel showcases the evolution of virial mass ($M_\mathrm{vir}$, dashed), gas mass ($M_\mathrm{gas}$, dotted), and stellar mass ($M_\star$, solid) within the virial radius of the target halo as a function of time (or redshift), with the thick gray line indicating the threshold virial mass for an atomic-cooling halo ($M_\mathrm{ACH}$; Prole2024) as a reference. The right panel further shows the evolution of the stellar mass within the target halo, separating the total stellar mass (dashed) from its Pop III contribution (solid). Different simulation runs are shown in different colors, with increasing line widths indicating increasing values of the strength of the LW flux ($\bar{J}_{21,0}$) in different simulation groups (NoLW, Low, Intermediate, High).
• Figure 2: Spatially resolved properties of the simulated galaxy from the LW1e3E100 run at $z \approx 7.67$ (top row), when the first star formation occurred, and $z \approx 7.59$ (bottom row), capturing the immediate post-starburst moment. From left to right, the panels show DM density, hydrogen number density, gas temperature, and gas metallicity along the same line of sight within $8R_{\rm vir}$ from the center of the galaxy. For reference, we indicate a length scale of 10 kpc at the top of each panel, and also delineate the halo virial radius (dashed circles). After the initial starburst, SN feedback from Pop III stars (cyan X-symbols) has destroyed the dense gas structure at the center. The strong outflows driven by intense SN feedback consequently expel most of the Pop II stars (magenta triangles) and gas clouds; the SN feedback also rapidly heats up and enriches nearby gas clouds.
• Figure 3: Gas properties of the target galaxy for the LW1e3E100 run before the starburst (left), mid-starburst (middle), and after the first SNe have exploded (right). Before the initial starburst, the $T-n$ phase diagram bifurcates at high densities ($n_{\rm H} \gtrsim 10 \rm \, cm^{-3}$) into a hot track ($T_{\rm gas} \gtrsim 10^4 \rm K$), which is fully exposed to strong LW flux, and a cold one ($T_{\rm gas} < 10^4 \rm K$), where the gas is self-shielded by $\rm H_{2}$. When gas particles reach $n_{\rm H} \gtrsim5\times 10^2 \rm \, cm^{-3}$, a small abundance of $\rm H_{2}$ molecules can trigger the $\rm H_{2}$ self-shielding effect, consequently cooling down a fraction of the dense gas. Eventually, intense SN feedback from the starburst has heated nearby gas to $T_{\rm gas} \lesssim 10^8 ~\rm K$, thus destroying the central dense gas structure. Most of the dense gas is expelled through outflows, while second-generation stars are still forming inside the remaining dense gas.
• Figure 4: Illustrating the modeling of external metal pollution. Here, we show the Lagrangian region of high-resolution particles, containing the target halo for Pop III star formation at its center (sky blue circle). The star symbols represent star particles external to the target halo (out-stars), which are the sources of the external metal enrichment. We mark the extent of the metal bubble (black dashed line), reached at $\Delta t$ after the start of the enrichment process. The red arrow ($d_{\rm out-tgt}$) shows the distance between the out-stars host halo and the center of the target halo. The cone (yellow dashed line) indicates the solid angle ($\Omega$) considered for tracking the metal bubble front, covering an area of $2R_{\rm vir}$ at the distance $d_{\rm out-tgt}$.
• Figure 5: Distance from the out-stars host halo to the metal bubble front as a function of time, followed until internal enrichment is triggered within our simulated galaxy in order to constrain the external metal enrichment process. Each solid colored line shows the evolution of the bubble front for select metallicities: $\log(Z_{\rm gas}/~Z_{\odot}) > -6$ (orange), $\log(Z_{\rm gas}/~Z_{\odot}) > -5$ (blue), and $\log(Z_{\rm gas}/~Z_{\odot}) > -4$ (black). The dashed lines show the distance to the center of the target halo (black) and to its outer boundary (red). For comparison, we show the analytic result using the galactic outflow model of Dijkstra2014. We also indicate the time when the Pop III starburst occurs within our simulated galaxy (marked as a green shaded region), demonstrating that metal bubbles from nearby enrichment events cannot reach the target halo on time to prevent metal-free star formation.