🔭 astrophysics

Constraining reionization-era Ly $\alpha$ escape with JELS-MUSE: a highly complete H $\alpha$ -selected sample at $z\sim6.1$

Using a highly complete, H $\alpha$ -selected sample of 24 star-forming galaxies at $z \approx 6.1$ observed by JWST and VLT/MUSE, this study measures a mean Ly $\alpha$ escape fraction of 0.07 and demonstrates that the observed scatter is driven by intrinsic ISM properties like dust attenuation and stellar mass rather than IGM variations, providing a crucial baseline for interpreting Ly $\alpha$ suppression during the Epoch of Reionization.

Original authors: A. L. Patrick, K. J. Duncan, Z. Li, C. A. Pirie, A. M. Swinbank, L. C. Keating, S. R. Flury, R. Begley, P. N. Best, M. Brinch, A. C. Carnall, F. Cullen, J. S. Dunlop, E. Ibar, J. Matthee, D. J. McLeod

Published 2026-09-03✓ Author reviewed ⓘ

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: A. L. Patrick, K. J. Duncan, Z. Li, C. A. Pirie, A. M. Swinbank, L. C. Keating, S. R. Flury, R. Begley, P. N. Best, M. Brinch, A. C. Carnall, F. Cullen, J. S. Dunlop, E. Ibar, J. Matthee, D. J. McLeod, A. Puglisi, H. M. O. Stephenson, J. P. Stott

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the earliest chapters of the universe, a vast fog of neutral hydrogen gas blanketed the cosmos, rendering it opaque to light. For hundreds of millions of years, this fog prevented the first stars and galaxies from being seen. Then, a great transition occurred known as the Epoch of Reionization. During this era, the intense radiation from young, hot stars began to strip electrons from the hydrogen atoms, clearing the fog and turning the universe into a transparent, ionized state. This process is one of the most significant transformations in cosmic history, yet the exact mechanics of how the first galaxies managed to punch holes through the surrounding gas remain a subject of intense study. To understand this, astronomers look for a specific type of light: Lyman-alpha emission. This is a bright, ultraviolet glow produced by hydrogen gas as it cools after being heated by star formation. Because this light interacts strongly with the neutral hydrogen fog, its ability to escape a galaxy and travel across the universe serves as a sensitive probe. If the light escapes easily, it suggests the galaxy has cleared a path through its own gas and the surrounding fog. If the light is trapped or scattered away, it indicates the path remains blocked.

A team of astronomers has now taken a precise census of this phenomenon at a critical moment in time, just as the cosmic fog was beginning to lift. By combining data from two powerful telescopes, they studied a group of twenty-four star-forming galaxies located approximately 13 billion light-years away, corresponding to a time when the universe was about 900 million years old. Their goal was to measure how much of this Lyman-alpha light successfully escaped these galaxies to reach us. Unlike previous studies that often focused only on galaxies already known to be bright in this specific light, this team started with a complete list of galaxies identified by a different, more fundamental signature: the glow of hydrogen-alpha light. This approach ensured they did not miss the galaxies that were struggling to let their light escape, providing a much more representative picture of the population as a whole.

The researchers found that half of the galaxies in their sample were successfully detected in Lyman-alpha light, while the other half were not. This 50 percent detection rate is a crucial benchmark. It tells us that at this specific moment in cosmic history, the conditions inside these galaxies were varied; some had clear paths for their light to escape, while others did not. By analyzing the light that did escape, the team calculated that, on average, only about 7 percent of the Lyman-alpha photons produced inside these galaxies actually made it out into the wider universe. The rest were absorbed by dust or scattered by gas before they could leave. This low average number is significant because it suggests that even at the end of the reionization era, the majority of galaxies were still quite effective at trapping their own light.

To understand why some galaxies let their light out while others did not, the team compared the properties of the successful emitters against those that failed. They discovered that the galaxies which managed to escape were generally less dusty, appeared bluer in color, and were less massive than those that failed. These characteristics point to a specific physical mechanism: the intense radiation and stellar winds from young stars likely cleared out channels in the surrounding gas, creating temporary tunnels through which the light could escape. However, the study also revealed a surprising level of chaos. Galaxies with very similar sizes, dust levels, and colors showed vastly different abilities to let light escape. Some of the most similar-looking galaxies had escape rates that differed by a factor of ten or more. This suggests that the ability to escape is not determined by a single, steady property of a galaxy, but rather by the complex, small-scale geometry of the gas and the specific angle from which we are viewing it.

The team also looked at the environment surrounding these galaxies to see if the cosmic fog itself was the primary culprit. Using computer simulations of the early universe, they confirmed that the galaxies in their study were all located in regions where the surrounding fog was similarly thin and ionized. This means that the differences in how much light escaped were not caused by one galaxy sitting in a dense patch of fog while another sat in a clear bubble. Instead, the variation was almost entirely due to the internal conditions of the galaxies themselves. This finding is vital for interpreting observations of even more distant galaxies, where the fog was thicker. It establishes that the scatter in how much light escapes is a natural feature of galaxy evolution, not just a result of the environment.

By anchoring their measurements to a complete sample of galaxies rather than just the brightest ones, the researchers provided a more honest accounting of the universe's ionizing budget. Their work suggests that while the average galaxy at this time was not a particularly efficient leak of ionizing light, a small subset of galaxies with the right internal conditions could be extremely effective. These rare, highly efficient galaxies might have been the primary drivers that finished the job of clearing the cosmic fog, even if they were outnumbered by the less efficient majority. The study serves as a solid reference point for future observations, helping astronomers distinguish between the effects of a galaxy's own structure and the overwhelming influence of the intergalactic medium as they look back to the very first moments when the universe became transparent.

Technical Summary: Constraining reionization-era Ly𝛼 escape with JELS-MUSE

Problem and Motivation
The Epoch of Reionization (EoR) marks the transition of the intergalactic medium (IGM) from neutral to ionized, a process driven primarily by star-forming galaxies. A critical quantity in understanding this era is the Lyman continuum (LyC) escape fraction ( $f_{\text{LyC}}^{\text{esc}}$ ), which determines the rate of ionizing photon injection into the IGM. However, direct LyC observations are impossible at $z \gtrsim 4.5$ due to IGM opacity. Consequently, the Lyman-alpha (Ly $\alpha$ ) escape fraction ( $f_{\text{Ly}\alpha}^{\text{esc}}$ ) serves as a primary indirect tracer, as Ly $\alpha$ photons share similar escape physics with LyC photons regarding neutral gas column density and dust geometry.

A significant challenge in characterizing $f_{\text{Ly}\alpha}^{\text{esc}}$ at high redshift is distinguishing between intrinsic variations in galaxy interstellar medium (ISM) conditions and attenuation caused by the IGM. Previous studies often relied on Ly $\alpha$ -selected or UV-selected samples, which introduce biases toward galaxies with favorable escape conditions or specific luminosity ranges. Furthermore, at $z \gtrsim 6.5$ , the increasing neutral fraction of the IGM suppresses Ly $\alpha$ visibility, making it difficult to isolate intrinsic ISM drivers. There is a need for a highly complete, unbiased sample at the transition redshift ( $z \sim 6$ ) to establish a benchmark for intrinsic Ly $\alpha$ escape before IGM suppression becomes dominant.

Methodology
This study utilizes the JELS-MUSE Large Area Survey, combining data from the James Webb Space Telescope (JWST) and the Very Large Telescope (VLT)/MUSE.

Sample Selection: The parent sample consists of 24 star-forming galaxies at $z \approx 6.1$ , selected exclusively via H $\alpha$ emission from the JWST Emission Line Survey (JELS). This narrow-band imaging approach (using F466N and F470N filters) ensures a flux-limited, complete census of star-forming galaxies down to $\log(L_{\text{H}\alpha}) \gtrsim 41.5$ erg s $^{-1}$ , avoiding the biases inherent in Ly $\alpha$ or UV continuum selection.
Spectroscopy: Ly $\alpha$ spectroscopy was obtained using VLT/MUSE over a contiguous 63 arcmin $^2$ field in COSMOS, covering the same region as the JELS footprint. This provides integral-field Ly $\alpha$ coverage for all 24 H $\alpha$ -selected sources.
Data Processing:
- Continuum Subtraction: A spectral median-filtering approach was applied to MUSE datacubes to remove continuum emission while preserving line flux.
- Extraction: An optimal spatial aperture of $r = 0.6$ arcsec ( $\sim 3\text{--}4$ kpc at $z=6.1$ ) was determined via curve-of-growth analysis to maximize signal-to-noise (SNR) while minimizing contamination from extended, low-surface-brightness haloes of uncertain origin.
- Detection Threshold: A false-positive analysis using 500 random spatial positions per subcube established an empirical detection threshold of $\Sigma_{\text{Ly}\alpha} > 5.09$ (98th percentile) for detections and $\Sigma_{\text{Ly}\alpha} > 6.41$ (99.5th percentile) for strong detections.
- Flux Measurement: Ly $\alpha$ fluxes were measured by fitting skewed Gaussian profiles to account for resonant scattering asymmetry. For non-detections, conservative $5\sigma$ upper limits were calculated.
Escape Fraction Calculation: $f_{\text{Ly}\alpha}^{\text{esc}}$ was computed as the ratio of observed Ly $\alpha$ luminosity to dust-corrected H $\alpha$ luminosity, assuming a Case B recombination ratio of 8.7. Dust correction was applied to H $\alpha$ based on SED-derived $A_V$ , while Ly $\alpha$ was left uncorrected to reflect the total fraction of photons escaping the galaxy (lost to dust or scattered out of the line-of-sight).
Statistical Analysis: Population-averaged escape fractions were derived using reverse Kaplan-Meier survival analysis to incorporate non-detections (upper limits) and a stacked-flux method. Correlations with galaxy properties (UV magnitude, slope, stellar mass, dust extinction) were assessed using Spearman rank and censored Kendall $\tau$ statistics.

Key Results

Detection Fraction: Ly $\alpha$ emission was detected in 12 of 24 sources (50 $\pm$ 10%). Using the canonical definition of Ly $\alpha$ emitters (EW(Ly $\alpha$ ) > 25 Å) within the magnitude range $-20.25 < M_{\text{UV}} < -18.75$ , the Ly $\alpha$ emitter fraction is $X_{\text{Ly}\alpha} = 33 \pm 12\%$ .
Escape Fraction Distribution: The detected sources exhibit a wide range of escape fractions, $f_{\text{Ly}\alpha}^{\text{esc}} = 0.04\text{--}0.83$ .
Population-Averaged Escape Fraction: Incorporating non-detections via reverse Kaplan-Meier analysis yields a median escape fraction of $\langle f_{\text{Ly}\alpha}^{\text{esc}} \rangle = 0.07^{+0.04}_{-0.03}$ . This is consistent with an independent stacked-flux estimate of $0.08^{+0.02}_{-0.02}$ .
IGM Uniformity: Reionization simulations matched to the survey area and redshift indicate that the IGM transmission across the field is broadly uniform and highly skewed toward unity. The observed scatter in $f_{\text{Ly}\alpha}^{\text{esc}}$ is therefore attributed to intrinsic ISM and circumgalactic medium (CGM) variations rather than IGM patchiness.
Drivers of Escape:
- Global Properties: No statistically significant differences were found between Ly $\alpha$ detections and non-detections in terms of UV magnitude ( $M_{\text{UV}}$ ), stellar mass ( $M_*$ ), star formation rate (SFR), or dust attenuation ( $A_V$ ) when tested with Kolmogorov-Smirnov statistics.
- Correlations: Qualitative trends suggest that galaxies with higher $f_{\text{Ly}\alpha}^{\text{esc}}$ tend to have lower nebular dust extinction ( $E(B-V)$ ), bluer UV slopes ( $\beta$ ), and lower stellar masses. However, these correlations are not individually significant at high confidence due to the sample size and intrinsic scatter.
- Scatter: The large scatter in $f_{\text{Ly}\alpha}^{\text{esc}}$ among galaxies with similar global properties implies that Ly $\alpha$ escape is governed by small-scale, anisotropic neutral gas geometry (e.g., "picket-fence" channels) and kinematics, which are not captured by integrated broadband photometric diagnostics.

Significance and Claims
The paper claims that this work provides a robust empirical benchmark for intrinsic Ly $\alpha$ escape at the end of the Epoch of Reionization ( $z \approx 6.1$ ). By utilizing an H $\alpha$ -selected sample, the study avoids the selection biases that plague Ly $\alpha$ -selected surveys, offering a more representative view of the star-forming galaxy population.

The authors posit that the substantial scatter in $f_{\text{Ly}\alpha}^{\text{esc}}$ observed in this single-field, narrow-redshift sample is intrinsic to galaxy ISM conditions. This finding is significant because it establishes that the decline in Ly $\alpha$ visibility observed at $z \gtrsim 6.5$ cannot be attributed solely to IGM suppression; a significant component of the variance is driven by the stochastic nature of ISM escape channels. Consequently, interpreting Ly $\alpha$ observations at higher redshifts requires accounting for this intrinsic variance.

The measured population-averaged escape fraction of $\sim 7\%$ is consistent with recent JWST-based H $\alpha$ -anchored studies but lower than luminosity-function-based estimates or Ly $\alpha$ -selected samples, suggesting that previous methods may overestimate the typical escape fraction due to selection biases. The results support a scenario where a small subset of galaxies with favorable, transient ISM conditions (likely driven by feedback-regulated low-column-density channels) dominates the ionizing photon budget, rather than a uniform population of efficient leakers.

Technical Summary: Constraining reionization-era Ly𝛼 escape with JELS-MUSE

More like this