Hunting for the First Explosions at the High-Redshift Frontier

The James Webb Space Telescope (JWST) has spectroscopically confirmed galaxies up to $z\sim14$, 300 Myr after the Big Bang, and several candidates have been discovered at $z\sim15-25$, with one candidate as high as $z\sim30$, only 100 Myr after the Big Bang. Such objects are unexpected, since theoretical studies have not predicted the existence of detectable galaxies at $z\sim30$. While any $z\sim30$ candidates may be contaminants at lower redshifts, we explore whether such extreme redshift sources could be consistent with hyper-energetic transient events linked to the formation of the first, metal-free, stars. Specifically, we consider pair-instability supernovae (PISNe), a predicted class of extreme thermonuclear explosions that leave no remnant behind. Using cosmological simulations, we investigate an overdense cosmic region, where star formation and subsequent PISNe occur at $z\sim30-40$, even within standard cosmology. Assessing the likelihood of such a region, the corresponding number of PISNe at $z\gtrsim20$, and their observed flux, we find that JWST has a non-negligible chance to detect a PISN event at extremely high redshifts. If a transient event were confirmed at $z\sim30$, this would provide a direct glimpse into the epoch of first star formation, dramatically extending the empirical reach of astronomy.

💡 Research Summary

The authors address the surprising emergence of photometric candidates at redshifts as high as z ≈ 30 reported by JWST, a regime where standard galaxy formation models predict virtually no detectable sources. They propose that these extreme‑redshift detections could be the signatures of pair‑instability supernovae (PISNe), the catastrophic explosions of very massive, metal‑free Population III stars. To test this hypothesis they perform a suite of cosmological “biased‑region” simulations using the GIZMO code with a modified initial power spectrum: the amplitude of primordial fluctuations (σ₈) is boosted from the Planck value of 0.829 to 1.5. This artificial enhancement creates a rare, highly overdense patch of the early universe that collapses early enough to form Pop III stars and PISNe at z ≈ 30–40.

The simulated volume is a 6 h⁻¹ cMpc box (≈3 arcmin on the sky at z ≈ 30), matching the field of view of a single JWST NIRCam pointing. The run contains 2 × 512³ particles, giving dark‑matter particle masses of 1.76 × 10⁵ M⊙, gas particle masses of 3.16 × 10⁴ M⊙, and star‑particle masses of ≈2 × 10³ M⊙. Star formation follows a stochastic prescription: gas with n_H > 100 cm⁻³ and T ≤ 10³ K can spawn a star particle with an efficiency η = 0.05 for Pop III and η = 0.1 for Pop II. Each star particle represents a small stellar population (≈600 M⊙) rather than an individual star.

Pop III stars are assigned a Larson‑type initial mass function (IMF) with parameters α = 0.17, β = 2, and a low‑mass cutoff of 20 M⊙, spanning 1–150 M⊙. The PISN progenitor mass range (140–260 M⊙) is integrated over this IMF, yielding an average of N_PISN/M_III ≈ 9.4 × 10⁻³ M⊙⁻¹ (or ≈1.2 × 10⁻² M⊙⁻¹ for a pure power‑law IMF). By measuring the Pop III star‑formation rate density (SFRD) in the simulation and applying the above conversion factor, the authors compute the observable PISN rate per unit solid angle as a function of redshift. The resulting rate peaks at z ≈ 30–35 with dN/dt ≈ 10⁻⁴ yr⁻¹ per JWST field, corresponding to roughly one event per 10 yr of observer time.

The paper then evaluates the likelihood that JWST has already surveyed a region comparable to their biased box. Using halo‑mass functions from the GUREFT and Sheth‑Tormen analyses, the most massive halo in the simulation (M_h ≈ 1.2 × 10⁸ M⊙) has a peak height ν ≈ 5, implying a comoving number density of ≈10⁻⁶ cMpc⁻³. The combined area of existing deep JWST programs (CEERS, JADES, PRIMER, COSMOS‑Web) totals ≈2500 arcmin², which translates to a comoving volume of ≈2.5 × 10⁶ cMpc³ at z ≈ 30. This volume is sufficient to expect at least one such overdense region with a probability of order ten percent, making the presence of a PISN‑hosting patch plausible.

Flux predictions are based on state‑of‑the‑art radiative‑transfer models of PISNe, which give absolute UV magnitudes of M_UV ≈ −22. At z ≈ 30 this corresponds to observed AB magnitudes of ≈28–29 in the NIRCam F200W filter, just at the detection limit of a 10 ks exposure. However, the long observable duration (≈10 yr in the observer frame) and the possibility of stacking multiple epochs increase the effective sensitivity, rendering detection feasible.

In discussion, the authors stress that the combination of (i) a rare but existent overdense region, (ii) a boosted Pop III formation rate, (iii) a non‑negligible PISN production efficiency, and (iv) JWST’s deep, wide‑field capabilities, collectively raise the probability of catching a PISN at z ≈ 30 from essentially zero to a modest, yet measurable, value. They acknowledge uncertainties: the true IMF of Pop III stars, the exact frequency of such extreme overdensities in the real universe, and the possibility that the photometric z ≈ 30 candidates are lower‑redshift interlopers.

The paper concludes that if JWST were to confirm a transient event at z ≈ 30 consistent with a PISN, it would constitute the first direct observation of the epoch of first star formation, extending empirical cosmology into the first 100 Myr after the Big Bang and providing a powerful constraint on models of early structure formation.