My main research interests are centred on trying to understand how galaxies form in the high-redshift Universe at all scales. I mainly use detailed cosmological radiation-hydrodynamics simulations to study how the various feedback processes shape galaxies and their environment during the first few billion years of their evolution. I am interested in understanding how gas flows from the large-scale structure of the Universe all the way to the interstellar-medium of galaxies to fuel their star formation and the growth of the supermassive black holes at their centres. This assembly process results in various form of feedback (supernovae, black hole feedback, radiative feedback, …) that significantly alter the state of the gas in the circumgalactic (CGM) and intergalactic medium (IGM). In particular, part of the ultraviolet radiation produced by massive stars and active galactic nuclei (AGN) propagates outside of the galaxies it originates from, carving in the process ionized bubbles in the IGM. As the primeval galaxies grow and assemble their stars, these ionized regions will grow and overlap, eventually filling the whole universe in less than a billion year, around z ~ 6. This Epoch of Reionization is the stage of most of my past and current research.
Galaxies during the Epoch of Reionization
An important element of reionization models is the amount of ionizing photons escaping the galaxy from which they have been emitted. The current understanding of reionization is that at least at early times, small galaxies in low mass haloes were the major contributors to the global ionizing budget. It is thus crucial to assess how much radiation can escape those small galaxies. I run simulations trying to understand the processes that set this “escape fraction” in high redshift galaxies.
Evolution of the SFR, gas outflow rate and escape fraction for a 109 M☉ halo.
In Trebitsch et al. (2017), I used simulations of small galaxies (in haloes with masses around 108 to 109 M☉), with very high resolution (around 10 pc) to show that the ionizing escape fraction varies with time, from nothing to almost 100%, with the same timescale, but slightly delayed with respect to the star formation. This is most likely due to the fact that ionizing radiation is trapped inside the star forming clouds prior to the first supernovae explosions. As soon as the first supernovae goes off, it clears the way for ionizing photons to escape. After star formation shuts down, there is no massive star left to produce ionizing radiation, and the escape fraction goes down again as the gas cools in the halo. Some animations illustrating this cycle can be found on this page.
Escape fraction for the same galaxy including different feedback processes
I extended this work in Trebitsch et al. (2018) to test the impact of putative AGN feedback in theses galaxies. I focus on five simulations of the same galaxy only changing the feedback processes included, alternatively turning on and off supernovae explosions and black hole feedback. One key result of this work is that in very low mass galaxies, feedback from the AGN cannot affect dramatically the escape of ionizing radiation, simply because the central black hole never grows enough. This is mostly due to the strong supernova feedback that is powerful enough to limit the central fuelling. Additionally, I found that the AGN itself does not provide much additional radiation, and therefore we should not expect massive black holes in dwarf galaxies to contribute significantly to the EoR.
In a paper currently under review, I apply the same methodology to a more massive, L* galaxy.
Galaxy — Black Hole coevolution
Lyman alpha blobs
Lyman-α blobs are very large, luminous, Lyman-α emitting nebulae, usually found at high redshift. While these objects denote the presence of large quantities of neutral hydrogen around galaxies (Lyman-α photons are emitted by the de-excitation of an hydrogen atom), the mechanism powering the Lyman-α emission is still unclear.
Various scenarios have been suggested to explain the origin of this emission. Among them, I studied the idea that Lyman-α blobs are tracers of the cosmic web. In this picture, the Lyman-α radiation is produced by the infall of cosmological filaments on (group of) proto-galaxies. As the gas falls in the dark matter halo, it will radiate its gravitational energy as Lyman-α photons.
Polarization signal around the modeled Lyman-α blob.
This scenario has been studied in details in a paper by Rosdahl & Blaizot (2012). I used a Monte-Carlo radiative transfer code called MCLya to investigate the Lyman-α properties of one of the blob they simulated. I showed in 2016 that the radiative transfer of Lyman-α radiation has only a small impact on the size and the shape of the blob.
One of the goal of this project was to get a theoretical understanding of the polarization properties of Lyman-α radiation emitted by the infalling gas. This was triggered by the observation that Lyman-α emission in a very massive blob was polarized (Hayes et al., 2011), which has often been interpred as a proof that the Lyman-α photons are produced in the center of the blob. In my work, I have shown that a similar polarization can be predicted even if most of the Lyman-α radiation is produced by infall of the intergalactic gas.