2025MNRAS.538.1830S

In fuzzy dark matter (FDM) cosmologies, the dark matter consists of ultralight bosons ($m\lesssim 10^{-20}$ eV). The astrophysically large de Broglie wavelengths of such particles hinder the formation of low-mass dark matter haloes. This implies a testable prediction: a corresponding suppression in the faint end of the ultraviolet luminosity function (UVLF) of galaxies. Notably, recent estimates of the faint-end UVLF at $z\sim 5{\!-\!}9$ in the Hubble Frontier Fields, behind foreground lensing clusters, probe up to five magnitudes fainter than typical ('blank-field') regions. These measurements thus far disfavour prominent turnovers in the UVLF at low luminosity, implying bounds on FDM. We fit a semi-empirical model to these and blank-field UVLF data, including the FDM particle mass as a free parameter. This fit excludes cases where the dark matter is entirely a boson of mass $m\, <\, 1.5\times 10^{-21}$ eV (with $2\sigma$ confidence). We also present a less stringent bound deriving solely from the requirement that the total observed abundance of galaxies, integrated over all luminosities, must not exceed the total halo abundance in FDM. This more model-agnostic bound disfavours $m\, <\, 5\times 10^{-22}$ eV ($2\sigma$). We forecast that future UVLF measurements from JWST lensing fields may probe masses several times larger than these bounds, although we demonstrate this is subject to theoretical uncertainties in modelling the FDM halo mass function.