Near-field limits on the role of faint galaxies in cosmic reionization

Abstract

Abstract Reionizing the Universe with galaxies appears to require significant star formation in low-mass haloes at early times, while local dwarf galaxy counts tell us that star formation has been minimal in small haloes around us today. Using simple models and the ELVIS simulation suite, we show that reionization scenarios requiring appreciable star formation in haloes with Mvir ≈ 108 M⊙ at z = 8 are in serious tension with galaxy counts in the Local Group. This tension originates from the seemingly inescapable conclusion that 30–60 haloes with Mvir > 108 M⊙ at z = 8 will survive to be distinct bound satellites of the Milky Way at z = 0. Reionization models requiring star formation in such haloes will produce dozens of bound galaxies in the Milky Way's virial volume today (and 100–200 throughout the Local Group), each with ≳105 M⊙ of old stars (≳13 Gyr). This exceeds the stellar mass function of classical Milky Way satellites today, even without allowing for the (significant) post-reionization star formation observed in these galaxies. One possible implication of these findings is that star formation became sharply inefficient in haloes smaller than ∼109 M⊙ at early times, implying that the high-z luminosity function must break at magnitudes brighter than is often assumed (at MUV ≃ −14). Our results suggest that the James Webb Space Telescope (and possibly even the Hubble Space Telescope with the Frontier Fields) may realistically detect the faintest galaxies that drive reionization. It remains to be seen how these results can be reconciled with the most sophisticated simulations of early galaxy formation at present, which predict substantial star formation in Mvir ∼ 108 M⊙ haloes during the epoch of reionization.