Cosmic Evolution of Dust in Galaxies: Methods and Preliminary Results

Abstract

We investigate the redshift (z) evolution of dust mass and abundance, their dependences on initial conditions of galaxy formation, and physical correlations between dust, gas, and stellar contents at different z based on our original chemodynamical simulations of galaxy formation with dust growth and destruction. In this preliminary investigation, we first determine the reasonable ranges of the most important two parameters for dust evolution, i.e., the timescales of dust growth and destruction, by comparing the observed and simulated dust mass and abundances and molecular hydrogen (H 2 ) content of the Galaxy. We then investigate the z-evolution of dust-to-gas ratios (D), H 2 gas fraction (f H 2 ), and gas-phase chemical abundances (e.g., A O = 12 + log(O/H)) in the simulated disk and dwarf galaxies. The principal results are as follows. Both D and f H 2 can rapidly increase during the early dissipative formation of galactic disks (z ∼ 2-3), and the z-evolution of these depends on initial mass densities, spin parameters, and masses of galaxies. The observed A O -D relation can be qualitatively reproduced, but the simulated dispersion of D at a given A O is smaller. The simulated galaxies with larger total dust masses show larger H 2 and stellar masses and higher f H 2 . Disk galaxies show negative radial gradients of D and the gradients are steeper for more massive galaxies. The observed evolution of dust masses and dust-to-stellar-mass ratios between z = 0 and 0.4 cannot be reproduced so well by the simulated disks. Very extended dusty gaseous halos can be formed during hierarchical buildup of disk galaxies. Dust-to-metal ratios (i.e., dust-depletion levels) are different within a single galaxy and between different galaxies at different z.

Modeling Dust Evolution in Galaxies With a Multiphase, Inhomogeneous Ism

“…To balance dust growth at high temperatures and to lower depletions in the diffuse ISM in models MRN3nm and EMRN3nm to the observed level would require an unreasonably short timescale t dest diff of destruction in the diffuse medium in Equation (16). We therefore conclude that the assumption of the maximum coefficient for all temperatures, commonly used in recent hydrodynamical simulations with dust (e.g., Bekki 2015aBekki , 2015bMcKinnon et al 2016), tends to overestimate the dust production rates in hydrodynamic numerical simulations of galactic evolution.…”

“…A disadvantage of the model of Bekki (2013) is the current implementation of dust processing in the ISM in which growth and destruction processes do not depend on the local conditions and occur on constant timescales. In subsequentworks, the dependence of the growth timescale on local conditions is included by scaling it with thetemperature and density of the particles (Yozin & Bekki 2014;Bekki 2015a). A new, live dust particle model has been presented by Bekki (2015b), who decoupled the gas and dust particles and implemented additional gas-grain interactions and radiation pressure on grains.…”

“…Recently, dust evolution has been incorporated in the moving-mesh simulation code AREPO and used in zoom-in cosmological simulations of Milky Way-sized galaxies (McKinnon et al 2016). Their assumptions for the dust model are similar to those outlined in Bekki (2015a), with the exception of the destruction timescale, which they related to the local supernova (SN) rate. While their simulations agree with a number of observables, they overpredict the dust-to-metal ratio in the circumgalactic medium.…”

“…The timescale of dust growth in the ISM is inversely proportional to the metallicity, resulting in the existence of the critical metallicity fordust growth (e.g., Zhukovska 2008;Zhukovska & Gail 2009;Asano et al 2013a). While some simulations include a dependence of the growth timescale on the local temperature and density (Bekki 2015a(Bekki , 2015bMcKinnon et al 2016), they assume a fixed sticking coefficient over the entire simulation volume. With this assumption, grains can grow by accretion even in the warm or hot gas if they stay asufficiently long time there (∼1 Gyr).…”

Globular cluster formation with multiple stellar populations from hierarchical star cluster complexes

“…In order to investigate the accretion processes of AGB ejecta and cold ISM, we use our original simulation code (Bekki 2013;Bekki 2015b, c) that can be run on GPU clusters. The code enables us to investigate chemical evolution, dust formation and evolution (Bekki 2013), formation of molecular hydrogen on dust grains (Bekki 2015b), photo-electric heating of gas by dust and star formation in galaxies (Bekki 2015c). Since the details of the code are already given in our previous papers, we just briefly explain the code in the present study.…”