Dim Builders of the Universe: JWST and a Gravitational Lens Clarify the Contribution of the Tiniest Galaxies to Reionization
Based on gravitationally lensed ultra-deep images from JWST/NIRCam (GLIMPSE survey) and HST/ACS (Hubble Frontier Fields) of the galaxy cluster Abell S1063, a sample of distant galaxy candidates at z~6–11 was constructed to constrain the faint end of the ultraviolet luminosity function (UVLF). Instead of a sharp cut-off, a model of smooth quadratic suppression after a characteristic luminosity M_t was used, accounting for a possible contribution from galaxies beyond the decline. No direct evidence for a cutoff was found down to M_UV = −13.5 at z=6; weak, medium, and strong decline models were ruled out beyond M_t = −15.9, −15.1, and −14.8, respectively. Lower limits were obtained: UV luminosity density ρ_UV ≥ 22×10^25 erg s^−1 Hz^−1 Mpc^−3, star formation rate density SFRD ≥ 25×10^−3 M_⊙ yr^−1 Mpc^−3, log(ionizing photon rate ṅ_ion) ≥ 51.02 (s^−1 Mpc^−3). Galaxies fainter than the standard threshold M_UV = −17 contribute more than 50% of the UV luminosity density and at least ~64% of the stellar ionizing photons at z=6. The model allows a non-zero population beyond M_t, so sources fainter than the decline continue to contribute to ρ_UV and ṅ_ion, requiring consideration of the decline shape in reionization scenarios.
Context
Understanding which sources ionized neutral hydrogen in the first billion years after the Big Bang remains a key challenge in astrophysics. The epoch of reionization, lasting roughly from z~40 to z~6 in the standard cosmological model ΛCDM, marked the end of the 'Dark Ages.' Observing these processes directly is difficult not only because of the faintness of distant galaxies, but also due to the time dilation effect (caused by the expansion of the Universe). The prime candidates for the main ionizers are faint dwarf galaxies, but their contribution critically depends on the shape of the UV luminosity function at the extremely faint end, where ordinary surveys are powerless.
Methods
To achieve record sensitivity limits, the authors combined ultra-deep images from JWST as part of the GLIMPSE survey with archival data from Hubble and used a natural 'telescope'—the massive galaxy cluster Abell S1063, which, via gravitational lensing (predicted by Fritz Zwicky), magnifies the light of background objects. After subtracting the cluster light and performing careful photometry, candidates were selected with strict criteria, and the reliability of photometric redshifts was checked against available spectroscopy from NIRSpec. For each object, the accessible volume and completeness were computed using a magnification map derived from a detailed gravitational lens model.
Results
The derived UV luminosity functions (UV LFs) for five redshift intervals from z~6 to z~11 showed steep slopes at the faint end, consistent with previous work. In the deepest slice (z~6), no direct signs of a turnover were found down to an absolute magnitude of M_UV = -13.5. However, by applying a more flexible parameterization of suppression as a quadratic roll-off, the authors could reliably rule out models with moderate and strong suppression down to limiting magnitude values of M_t = -15.1 and -14.8, respectively. Even in a conservative weak-suppression scenario (coefficient δ=0.2), a guaranteed lower limit of the UV luminosity density is ρ_UV ≥ 22×10^25 erg/s/Hz/Mpc^3. The key conclusion: at z=6, galaxies fainter than M_UV = -17 provide more than half of the total UV luminosity and, by estimation, at least 64% of the ionizing photons.
Implications
The work demonstrates that the idea of a 'turnover' in the UV LF as a sharp cutoff in star formation is incorrect: even significant suppression of the star formation rate in low-mass halos leaves room for a numerous but stunted population of dwarfs. These 'understars' turn out to be critically important for the total photon budget needed for reionization, forcing a revision of models that rely on fixed luminosity integration limits.
Future development
Further progress will be linked to the accumulation of deep spectroscopic data from 30-meter telescopes and the next generation of space observatories, which will allow direct measurement of the ionization efficiency of the faintest objects. In parallel, numerical simulations with more realistic feedback physics and inclusion of alternative dark matter models (whose mysterious nature Vera Rubin already tried to uncover) will help refine predictions for the shape of the UV LF at the faintest luminosities.
Impact
The results directly affect the cosmology of the reionization epoch, galaxy formation theory, and constraints on dark matter properties.
Next steps
Direct spectroscopic measurements of the ionizing power of ultra-faint galaxies at z>6, and improved accuracy of gravitational lensing models.
Key open problems
The study links the uncertainty in the UV LF shape to fundamental problems in small-halo physics: star formation suppression mechanisms, the nature of dark matter (cold, warm, fuzzy dark matter models), and the thermodynamics of intergalactic gas, foundations laid down in the work of Georges Lemaître on the early Universe.
🎯 The faintest galaxy discovered in the Abell S1063 field was magnified by the gravitational lens by more than 25 times—without this 'cosmic telescope,' it would simply be impossible to see.
Key numbers
- UV luminosity limit (z=6) with no signs of turnover: M_UV = -13.5
- Fraction of UV emission from galaxies fainter than M_UV=-17: >50%
- Fraction of ionizing photons from galaxies fainter than M_UV=-17: ~64%
- Gravitational lens magnification factor: up to μ > 25 for individual objects