2020ApJ...895..116E

We measure the ionizing photon production efficiency (ξ_ion) of low-mass galaxies (10^7.8-10^9.8 M_☉) at 1.4 < z < 2.7 to better understand the contribution of dwarf galaxies to the ionizing background and reionization. We target galaxies that are magnified by strong-lensing galaxy clusters and use Keck/MOSFIRE to measure nebular emission-line fluxes and Hubble Space Telescope to measure the rest-UV and rest-optical photometry. We present two methods of stacking. First, we take the average of the log of Hα-to-UV luminosity ratios (L_Hα/L_UV) of galaxies to determine the standard log(ξ_ion). Second, we take the logarithm of the total L_Hα over the total L_UV. We prefer the latter, as it provides the total ionizing UV luminosity density of galaxies when multiplied by the nonionizing UV luminosity density. log(ξ_ion) calculated from the second method is ∼0.2 dex higher than the first method. We do not find any strong dependence between log(ξ_ion) and stellar mass, far-UV magnitude (M_UV), or UV spectral slope (β). We report a value of log(ξ_ion) ∼ 25.47 ± 0.09 for our UV-complete sample (-22< M_UV< -17.3) and ∼25.37 ± 0.11 for our mass-complete sample (7.8 < log(M_*) < 9.8). These values are consistent with measurements of more massive, more luminous galaxies in other high-redshift studies that use the same stacking technique. Our log(ξ_ion) is 0.2-0.3 dex higher than low-redshift galaxies of similar mass, indicating an evolution in the stellar properties, possibly due to metallicity or age. We also find a correlation between log(ξ_ion) and the equivalent widths of Hα and [O III] λ5007 fluxes, confirming that these equivalent widths can be used to estimate ξ_ion.