Escape fraction of Lyman-continuum and Lyman-alpha

It is commomly believed that galaxies at z ≳ 5 are the dominant sources for cosmic reionization. The escape fraction of Lyman-continuum (LyC), or hydrogen ionizing photons from these galaxies is hence an important question for understanding the reionization process. The escape fraction cannot be measured directly, so high-resolution cosmological simulations including as many details as possible serve the best tool for this problem.

Similar to LyC photons, Lyman-alpha (Lyα) is also affected by neutral hydrogen in the galaxy. Understanding the Lyα-LyC connection is theoretically challenging: it demands ultra-high resolution in cosmological simulations, as the mean free paths of Lyα photons are extremely short.

Binary stars contribute significantly to reionization

I develop a Monte Carlo radiative transfer (MCRT) code to solve LyC transport and ionization balance, which runs on octree grids (an adaptive mesh refinement structure) constrcuted from meshless simulations like FIRE.

In Ma et al. (2015, MNRAS, 453, 960), we used three high-resolution cosmological zoom-in simulations of z ≳ 5 galaxies from the FIRE project, and found that the escape fractions of ionizing photons from these galaxies are
≲ 5% on average, far below what is required (~20%) for cosmic reionization.

Figure: Left: Gas density around stars of different ages. Young stars are embedded in their "birth clouds", until a few Myrs later due to feedback (from Ma et al. 2015).

Top right: Ionizing photon emissivities for single-star and binary stellar population models. Binaries produce more photons at later times (from Ma et al. 2016b).

Bottom right: The effect of binary stars on ionizng photon escape fraction, which can be boosted by a factor of ~3-10 by binaries.

This is because most of the ionizing photons are produced by the youngest stars, which are always embedded in dense molecular clouds, so that these photons cannot escape (see Figure above). Only after a few Myr can feedback clear sightlines to allow ionizing photons to escape. However, according to canonical single-star stellar population models like STARBURST99, the ionizing photon emissivity declines rapidly after 3 Myr.

In Ma et al. (2016b, MMRAS, 459, 3614), we revisit this calculation using binary stellar population models BPASS. Binary stars have prolonged ionizing photon production rates. This boosts the escape fraction by a factor of ~3‒10. We pioneer the inclusion of binaries in understanding cosmic reionization, which has been followed by other groups.

How do LyC escape fractions depend on mass and redshift?

Figure: The dependence of LyC escape fraction on stellar mass, predicted by FIRE in the high-redshift Universe simulations (from Ma et al. 2020b).

In Ma et al. (2020b, MNRAS, 498, 2001), we use the entire sample of my FIRE in the high-redshift Universe simulations to study the dependence of LyC escape fraction on halo mass, stellar mass, and redshift.

We find f_esc peaks around M_* ~ 10⁸M_⊙ and decreases both at the high-mass and the low-mass end, due to increasing dust attenuation and inefficient star formation and feedback, respectively (see figure on the left).

We also present the typical topology of LyC-leaking regions (see figure below). Most leaked photons are from vigorously star-forming regions that usually contain a feedback-driven kpc-scale superbubble surrounded by a dense shell. The shell is forming stars while accelerated, so new stars formed earlier in the shell are already inside the shell. Young stars in the bubble and near the edge of the shell can fully ionize some low-column-density paths pre-cleared by feedback, allowing a large fraction of LyC photons to escape.

Figure: A strong LyC leaker at z ~ 5. Left: Galaxy-scale gas image. All star formation in the past 10 Myr occured in regions A-C, while only region A is leaking LyC photons. Center: Zoom-in image of region A. The white points are stars 3-10 Myr old, while the color points show stars younger than 3 Myr, color-coded by age. Right: The same as the central panel, expect that the color points are coded by the f_esc of individual stars (from Ma et al. 2020b).

Lyman-alpha escape fraction & LyA-LyC connection

With Aaron Smith, we integrate my octree-based ray-tracing module into his Lyα MCRT code COLT (Smith et al. 2015) to explore the Lyα escape fraction and Lyα-LyC connection using the FIRE simulations. A comprehensive case study using one simulated galaxy is published in Smith, Ma et al. (2019, MNRAS, 484, 39).

In an ongoing study, we run the Lyα MCRT code COLT on the entire FIRE in the high-redshift Universe simulation sample. In future work, we will also model nebular emission lines, including Balmer lines, He II, O III, and so on. We aim to address the following questions:

• How do the Lyα luminosities, line profiles, and peak velocities depend on M_* and other galaxy properties?
• What are the systematic uncertainties of using Lyα from z ≳ 5 galaxies to probe the reionization history?
• Why do LyC escape fractions correlate with Lyα and nebular emission line properties (e.g. [O III]/[O II]) as empirically suggested by observations of LyC-leaking galaxies at lower redshift?
• Can we use these indicators to indirectly measure the LyC escape fractions from z ≳ 5 galaxies?