Mass, metal, and energy feedback in cosmological simulations. Oppenheimer, B. D. & Davé, R. Monthly Notices of the Royal Astronomical Society, 387:577–600, June, 2008.
Mass, metal, and energy feedback in cosmological simulations [link]Paper  doi  abstract   bibtex   
Using GADGET-2 cosmological hydrodynamic simulations including an observationally constrained model for galactic outflows, we investigate how feedback from star formation distributes mass, metals, and energy on cosmic scales from z = 6 -\textgreater 0. We include instantaneous enrichment from Type II supernovae (SNe), as well as delayed enrichment from Type Ia SNe and stellar [asymptotic giant branch (AGB)] mass loss, and we individually track carbon, oxygen, silicon and iron using the latest yields. Following on the success of the momentum-driven wind scalings, we improve our implementation by using an on-the-fly galaxy finder to derive wind properties based on host galaxy masses. By tracking wind particles in a suite of simulations, we find: (1) wind material re-accretes on to a galaxy (usually the same one it left) on a recycling time-scale that varies inversely with galaxy mass (e.g. \textless1 Gyr for L* galaxies at z = 0). Hence, metals driven into the intergalactic medium by galactic superwinds cannot be assumed to leave their galaxy forever. Wind material is typically recycled several times; the median number of ejections for a given wind particle is 3, so by z = 0 the total mass ejected in winds exceeds 0.5Ωb. (2) The physical distance winds travel is fairly independent of redshift and galaxy mass (\textasciitilde60-100 physical kpc, with a mild increase to lower masses and redshifts). For sizeable galaxies at later epochs, winds typically do not escape the galaxy halo, and rain back down in a halo fountain. High-z galaxies enrich a significantly larger comoving volume of the intergalactic medium (IGM), with metals migrating back into galaxies to lower z. (3) The stellar mass of the typical galaxy responsible for every form of feedback (mass, metal, and energy) grows by \textasciitilde30 times between z = 6 -\textgreater 2, but only approximately two to three times between z = 2 -\textgreater 0, and is around or below L* at all epochs. (4) The energy imparted into winds scales with M1/3gal, and is roughly half the SN energy. Given radiative losses, energy from another source (such as photons from young stars) may be required to distribute cosmic metals as observed. (5) The production of all four metals tracked is globally dominated by Type II SNe at all epochs. However, intracluster gas iron content triples as a result of non-Type II sources, and the low-z IGM carbon content is boosted significantly by AGB feedback. This is mostly because gas is returned into the interstellar medium to form one-third more stars by z = 0, appreciably enhancing cosmic star formation at z \textless\textasciitilde 1.
@article{oppenheimer_mass_2008,
	title = {Mass, metal, and energy feedback in cosmological simulations},
	volume = {387},
	issn = {0035-8711},
	url = {http://adsabs.harvard.edu/abs/2008MNRAS.387..577O},
	doi = {10.1111/j.1365-2966.2008.13280.x},
	abstract = {Using GADGET-2 cosmological hydrodynamic simulations including an observationally constrained model for galactic outflows, we investigate how feedback from star formation distributes mass, metals, and energy on cosmic scales from z = 6 -{\textgreater} 0. We include instantaneous enrichment from Type II supernovae (SNe), as well as delayed enrichment from Type Ia SNe and stellar [asymptotic giant branch (AGB)] mass loss, and we individually track carbon, oxygen, silicon and iron using the latest yields. Following on the success of the momentum-driven wind scalings, we improve our implementation by using an on-the-fly galaxy finder to derive wind properties based on host galaxy masses. By tracking wind particles in a suite of simulations, we find: (1) wind material
re-accretes on to a galaxy (usually the same one it left) on a recycling time-scale that varies inversely with galaxy mass (e.g. {\textless}1 Gyr for L* galaxies at z = 0). Hence, metals driven into the intergalactic medium by galactic superwinds cannot be assumed to leave their galaxy forever. Wind material is typically recycled several times; the median number of ejections for a given wind particle is 3, so by z = 0 the total mass ejected in winds exceeds 0.5Ωb. (2) The physical
distance winds travel is fairly independent of redshift and galaxy mass ({\textasciitilde}60-100 physical kpc, with a mild increase to lower masses and
redshifts). For sizeable galaxies at later epochs, winds typically do not escape the galaxy halo, and rain back down in a halo fountain. High-z galaxies enrich a significantly larger comoving volume of the intergalactic medium (IGM), with metals migrating back into galaxies to lower z. (3) The stellar mass of the typical galaxy responsible for every form of feedback (mass, metal, and energy) grows by {\textasciitilde}30 times between z = 6 -{\textgreater} 2, but only approximately two to three times between z = 2 -{\textgreater} 0, and is around or below L* at all epochs. (4) The energy imparted into winds scales with M1/3gal, and is roughly half the SN energy. Given radiative losses, energy from another source (such as photons from young stars) may be required to distribute cosmic metals as observed. (5) The production of all four metals tracked is globally dominated by Type II SNe at all epochs. However,
intracluster gas iron content triples as a result of non-Type II sources, and the low-z IGM carbon content is boosted significantly by AGB feedback. This is mostly because gas is returned into the
interstellar medium to form one-third more stars by z = 0, appreciably enhancing cosmic star formation at z {\textless}{\textasciitilde} 1.},
	urldate = {2019-05-13},
	journal = {Monthly Notices of the Royal Astronomical Society},
	author = {Oppenheimer, Benjamin D. and Davé, Romeel},
	month = jun,
	year = {2008},
	keywords = {cosmology: theory, galaxies: abundances, galaxies: evolution, galaxies: high-redshift, intergalactic medium, methods: numerical},
	pages = {577--600},
}

Downloads: 0