Simulating the H2 content of high-redshift galaxies

Abstract

We introduce a sub-grid model for the non-equilibrium abundance of molecular hydrogen in cosmological simulations of galaxy formation. We improve upon previous work by accounting for the unresolved structure of molecular clouds in a phenomenological way which combines both observational and numerical results on the properties of the turbulent interstellar medium. We apply the model to a cosmological simulation of the formation of a Milky Way-sized galaxy at z = 2, and compare the results to those obtained using other popular prescriptions that compute the equilibrium abundance of H 2 . In these runs we introduce an explicit link between star formation and the local H 2 abundance, and perform an additional simulation in which star formation is linked directly to the density of cold gas. In better agreement with observations, we find that the simulated galaxy produces less stars and harbours a larger gas reservoir when star formation is regulated by molecular hydrogen. In this case, the galaxy is composed of a younger stellar population as early star formation is inhibited in small, metal poor dark-matter haloes which cannot efficiently produce H 2 . The number of luminous satellites orbiting within the virial radius of the galaxy at z = 2 is reduced by 10-30 per cent in models with H 2 -regulated star formation.

Towards the complete census of molecular hydrogen in a simulated disc galaxy

“…A direct comparison between the KMT equilibrium model and non-equilibrium models shows that they diverge at low metallicities (Krumholz & Gnedin 2011), that non-equilibrium models are capable of maintaining H 2 at lower densities (Tomassetti et al 2014), and that the nonequilibrium models are clumpier and closer to the KS relation (Pallottini et al 2017).…”

A simple model for molecular hydrogen chemistry coupled to radiation hydrodynamics

“…They explore the nature of the relationship between star formation and H 2 , test star formation recipes, and see how H 2 affects the gas composition of galaxies. The semi-analytical models use equilibrium equations to find the H 2 fraction (Fu et al 2010;Somerville et al 2015;Xie et al 2017) while hydrodynamic simulations use either equilibrium equations (Pelupessy et al 2006;Robertson & Kravtsov 2008;Kuhlen et al 2012;Halle & Combes 2013;Hopkins et al 2014; Thompson et al 2014) or a series of non-equilibrium chemical networks (Gnedin et al 2009;Gnedin & Kravtsov 2011;Christensen et al 2012;Tomassetti et al 2014;Baczynski et al 2015;Richings & Schaye 2016;Hu et al 2016;Capelo et al 2018;Lupi et al 2018;Pallottini et al 2017;Katz et al 2017). Equilibrium calculations have the advantage of speed but use the assumption that the chemical species are in equilibrium with their environment.…”

Supernova blast waves in wind-blown bubbles, turbulent, and power-law ambient media

“…Schaye et al 2015). In general, this long-known 'over-cooling problem' appears when the main momentum creating stages, the SedovTaylor and the pressure driven snowplough phase, stay unresolved and become artifically short (Balogh et al 2001;Stinson et al 2006;Creasey et al 2011;Tomassetti et al 2015). The thermal energy is radiated away too quickly and the momentum input is unresolved as too much mass is accelerated to too low velocities (Hu et al 2015), in particular if the time step is not reduced accordingly (Dalla Vecchia & Schaye 2012;Kim & Ostriker 2015).…”