Chemical Signatures of Population III Stars in Damped Lyman-$α$ Absorption Systems at $z \approx 6$

1. Problem Statement

The formation and properties of the first stars (Population III or Pop III) remain a major open question in cosmology. While direct observation of individual Pop III stars is currently unfeasible, indirect probes are necessary. Recent observations by Sodini et al. (2024) identified a sample of Damped Lyman-α (DLA) absorption systems at redshift $z \approx 6$ containing OI lines. These systems exhibit a wide spread in elemental abundance ratios, particularly Carbon-to-Oxygen ([C/O]), with some showing high [C/O] values ($\sim 0.6$) potentially indicative of enrichment solely by Pop III stars.

Previous theoretical attempts to model these signatures were limited:

• Analytic models often lacked 3D spatial fluctuations, Lyman-Werner (LW) feedback, and detailed reionization histories.
• Simplified chemical models assumed gas enrichment from single supernovae or arbitrary mixing fractions without tracking the cosmological evolution of halos.

The core problem addressed is whether current semi-analytic models, when incorporating advanced physical effects (3D feedback, reionization, streaming velocities), can reproduce the observed abundance and chemical signatures of these high-redshift DLAs, and if so, what constraints this places on Pop III star formation parameters.

2. Methodology

The authors utilize a state-of-the-art semi-analytic model (SAM) built upon 10 cosmological N-body simulation boxes (each 3 Mpc across, $512^3$ particles, resolution $\sim 8,000 M_\odot$).

Key Physical Components:

• Halo Merger Trees: Generated using rockstar to track halo growth and mergers.
• Pop III Star Formation: Triggered in pristine halos reaching a critical mass ($M_{crit}$). $M_{crit}$ is dynamically calculated based on LW radiation intensity, redshift, and baryon-dark matter streaming velocity ($v_{bc}$), calibrated against recent simulations [45].
• Feedback Mechanisms:
  • LW Feedback: Suppresses H$_2$ cooling in low-mass halos.
  • Reionization: Photoheating suppresses star formation in halos below a characteristic mass ($M_{ion}$).
  • External Metal Enrichment: Supernova-driven metal bubbles ($R_{bub} \approx 150 h^{-1}$ kpc) enrich neighboring halos, transitioning them from Pop III to metal-enriched (Pop II/I) star formation if metallicity exceeds $Z_{crit}$.
• Chemical Yields:
  • Pop III: Assumed to be rapidly rotating $35 M_\odot$ stars (yields from [60]), dispersing metals via stellar winds with [C/O] $\approx 0.6$. A small fraction of non-rotating stars are assumed to explode as supernovae to drive the delay time.
  • Pop II/I: Follow a Kroupa IMF with yields from [69].
• DLA Abundance Calculation:
  • The number of DLAs is estimated by summing the effective cross-sectional area ($A_{eff}$) of halos above a column density threshold, weighted by the baryon-dark matter streaming velocity distribution.
  • $A_{eff}$ is calibrated from $z=3$ simulations [70] and extrapolated to $z \approx 6$.

Model Parameters:
The study varies key parameters from a fiducial set (Table 1), including:

• Pop III star formation efficiency ($f_{III}$).
• Delay time ($t_{delay}$) between Pop III burst and subsequent metal-enriched star formation (modeling supernova feedback).
• Critical mass for ionization suppression ($M_{ion}$).
• Metallicity of Pop II stars ($Z=0.004$ vs $0.02$).

3. Key Contributions

This work represents the first theoretical prediction of Pop III chemical signatures in DLAs that simultaneously includes:

1. 3D Spatial Fluctuations: Accounting for halo clustering and inhomogeneous metal enrichment.
2. Complex Feedback Loops: Integrating LW feedback, reionization history, and baryon-dark matter streaming velocities ($v_{bc}$).
3. Realistic Delay Times: Modeling the suppression of Pop II star formation following a Pop III burst due to feedback.
4. Direct Comparison to Data: Quantitatively comparing predicted [C/O] distributions against the specific OI-selected DLA sample from Sodini et al. (2024).

4. Results

Fiducial Model Performance:
The fiducial model (standard parameters) is strongly ruled out by the observational data. It predicts a severe deficit of high-[C/O] DLAs (by $\sim 3$ orders of magnitude). The fiducial model produces mostly low-[C/O] systems dominated by metal-enriched stars, failing to reproduce the tail of the distribution where [C/O] $\approx 0.6$.

Parameter Sensitivity:

• Pop III Efficiency ($f_{III}$): Increasing $f_{III}$ to 0.01 increases the fraction of Pop III stars and [C/O], but does not fully resolve the discrepancy with observations.
• Delay Time ($t_{delay}$): This is the most critical parameter. Increasing $t_{delay}$ to $5 \times 10^7$ years significantly improves agreement.
  • Mechanism: A longer delay allows Pop III stars to ionize the local IGM before Pop II stars can form. This suppresses subsequent star formation in low-mass halos (due to photoheating), preserving a higher ratio of Pop III to Pop II metals and extending the duration of purely Pop III enrichment.
• Effective Area ($A_{eff}$): The model requires a normalization increase of $\sim 4\times$ to match the total DLA count, likely due to higher neutral gas fractions at $z \approx 6$ compared to $z=3$.
  • A specific variation reducing the mass scale $M_0$ in $A_{eff}$ (enhancing low-mass halo contribution) combined with increased $t_{delay}$ yields excellent agreement with the observed [C/O] distribution.

Chemical Trends:

• High-mass halos ($M_h > 10^9 M_\odot$) consistently show [C/O] $\approx -0.45$ due to well-sampled IMFs of metal-enriched stars.
• High [C/O] ($\sim 0.6$) is only achieved in low-mass systems where the gas is either purely Pop III or where stochastic sampling of the IMF in metal-enriched galaxies fails to produce carbon-poor supernovae.

5. Significance and Conclusions

• Validation of DLA Probes: The study confirms that DLA analogs at $z \approx 6$ are a promising probe for Pop III star formation. The observed high-[C/O] systems can be naturally explained by Pop III enrichment if the delay between Pop III and Pop II formation is sufficiently long.
• Constraints on Physics: The data places strong constraints on the delay time ($t_{delay}$) and Pop III star formation efficiency. The results suggest that feedback mechanisms (specifically supernova-driven delays and subsequent photoheating) play a crucial role in preserving Pop III chemical signatures.
• Model Improvements: The inclusion of 3D fluctuations and feedback mechanisms was shown to be essential; previous models lacking these features predicted Pop III star formation to cease entirely by $z \sim 8$, whereas this model shows it persisting to $z \approx 6$.
• Future Directions: The authors note that perfect agreement requires further refinement, including:
  • Hydrodynamical simulations to better model the DLA effective area (specifically using OI column density).
  • Investigation of inhomogeneous metal mixing within halos.
  • Coupling with larger-scale boxes ($>100$ Mpc) to accurately model reionization topology.

In summary, the paper demonstrates that with physically motivated adjustments (specifically a longer delay time between Pop III and Pop II formation), current semi-analytic models can successfully reproduce the chemical signatures of the first stars observed in high-redshift DLAs, offering a new window into the earliest epochs of galaxy formation.