Semi-analytic forecasts for JWST - III. Intrinsic production efficiency of Lyman-continuum radiation

The James Webb Space Telescope is expected to enable transformational progress in studying galaxy populations in the very early Universe, during the epoch of reionization. A critical parameter for understanding the sources that reionized the Universe is the Lyman-continuum production efficiency, xi(ion), defined as the rate of production of ionizing photons divided by the intrinsic UV luminosity. In this work, we combine self-consistent star formation and chemical enrichment histories predicted by semi-analytic models of galaxy formation with stellar population synthesis (SPS) models to predict the expected dependence of xi(ion) on galaxy properties and cosmic epoch from z = 4-10. We then explore the sensitivity of the production rate of ionizing photons, (N) over dot(ion) to the choice of SPS model and the treatment of stellar feedback in our galaxy formation model. We compare our results to those of other simulations, constraints from empirical models, and observations. We find that adopting SPS models that include binary stars predict about a factor of 2 more ionizing radiation than models that only assume single stellar populations. We find that UV-faint, low-mass galaxies have values of xi(ion) about 0.25 dex higher than those of more massive galaxies, but find weak evolution with cosmic time, about 0.2 dex from z similar to 12-4 at fixed rest-UV luminosity. We provide predictions of (N) over dot(ion) as a function of M-h and a number of other galaxy properties. All results presented in this work are available at https://www.simonsfoundation.org/semi-analytic-forecasts-for-jwst/.