Photon Number Conserving Models of H II Bubbles during Reionization

TL;DR

This paper introduces a new Monte Carlo model for H II bubble growth during reionization that conserves photon number, addressing a key shortcoming of traditional excursion set models and improving the accuracy of bubble size predictions.

Contribution

The authors develop a photon number conserving Monte Carlo model for H II bubble growth, correcting a fundamental flaw in existing excursion set based models.

Findings

Significant improvement in photon number conservation with the new model

Substantial differences in bubble size distribution compared to traditional models

Robustness of improvements across different initial conditions

Abstract

Traditional excursion set based models of H II bubble growth during the epoch of reionization are known to violate photon number conservation, in the sense that the mass fraction in ionized bubbles in these models does not equal the ratio of the number of ionizing photons produced by sources and the number of hydrogen atoms in the intergalactic medium. E.g., for a Planck13 cosmology with electron scattering optical depth $\tau\simeq0.066$, the discrepancy is $\sim15$ per cent for $x_{\rm HII}=0.1$ and $\sim5$ per cent for $x_{\rm HII}=0.5$. We demonstrate that this problem arises from a fundamental conceptual shortcoming of the excursion set approach (already recognised in the literature on this formalism) which only tracks average mass fractions instead of the exact, stochastic source counts. With this insight, we build an approximately photon number conserving Monte Carlo model of bubble growth based on partitioning regions of dark matter into halos. Our model, which is formally valid for white noise initial conditions (ICs), shows dramatic improvements in photon number conservation, as well as substantial differences in the bubble size distribution, as compared to traditional models. We explore the trends obtained on applying our algorithm to more realistic ICs, finding that these improvements are robust to changes in the ICs. Since currently popular semi-numerical schemes of bubble growth also violate photon number conservation, we argue that it will be worthwhile to pursue new, explicitly photon number conserving approaches. Along the way, we clarify some misconceptions regarding this problem that have appeared in the literature.