CLARA's view on the escape fraction of Lyman-Alpha photons in high redshift galaxies

TL;DR

This study uses the CLARA radiative transfer code to estimate the Lyman-Alpha photon escape fraction in high-redshift galaxies, aligning simulations with observations and exploring the impact of dust and intergalactic medium effects.

Contribution

Introduces a new radiative transfer code CLARA and a slab model with dust clumpiness to better simulate Lyman-Alpha escape fractions and luminosity functions at high redshift.

Findings

Weak redshift evolution of escape fraction observed.

Model reproduces bright end of luminosity functions at z~5-6.

Over-prediction of luminosity function at z~7 suggests missing IGM effects.

Abstract

Using CLARA (Code for Lyman Alpha Radiation Analysis) we constrain the escape fraction of Lyman-Alpha radiation in galaxies in the redshift range 5<z<7, based on the MareNostrum High-z Universe, a SPH cosmological simulation with more than 2 billion particles. We approximate Lyman-Alpha Emitters (LAEs) as dusty gaseous slabs with Lyman-Alpha radiation sources homogeneously mixed in the gas. Escape fractions for such a configuration and for different gas and dust contents are calculated using our newly developed radiative transfer code CLARA. The results are applied to the MareNostrum High-z Universe numerical galaxies. The model shows a weak redshift evolution and good agreement with estimations of the escape fraction as a function of reddening from observations at z \sim 2.2 and z \sim 3. We extend the slab model by including additional dust in a clumpy component in order to reproduce the UV con- tinuum luminosity function and UV colours at redshifts z>~5. The LAE Luminosity Function (LF) based on the extended clumpy model reproduces broadly the bright end of the LF derived from observations at z \sim 5 and z \sim 6. At z \sim 7 our model over-predicts the LF by roughly a factor of four, presumably because the effects of the neutral intergalactic medium are not taken into account. The remaining tension between the observed and simulated faint end of the LF, both in the UV-continuum and Lyman-Alpha at redshifts z \sim 5 and z \sim 6 points towards an overabundance of simulated LAEs hosted in haloes of masses 1.0x10^10h-1Msol < Mh < 4.0x10^10h-1Msol. Given the difficulties in explaining the observed overabundance by dust absorption, a probable origin of the mismatch are the high star formation rates in the simulated haloes around the quoted mass range. A more efficient supernova feedback should be able to regulate the star formation process in the shallow potential wells of these haloes.