Galactic winds are ubiquitous in most rapidly star-forming galaxies in both the local and high-redshift universe. They shape the galaxy luminosity function, flattening its faint-end slope compared to the halo mass function, and affect the chemical evolution of galaxies, determining the mass-metallicity relation, regulating star formation over cosmic time, and polluting the intergalactic medium (IGM) with metals. Although important, the physics of galactic winds is still unclear. Many theoretical mechanisms have been proposed. Winds may be driven by the heating of the interstellar medium (ISM) by overlapping supernovae (SNe), cosmic rays, the radiation pressure by continuum absorption and scattering of starlight on dust grains, or the momentum input from SNe. However, the comparison between theory and observation is incomplete. The growing observations of emission and absorption of cold molecular, cool atomic, and ionized gas in galactic outflows in a large number of galaxies have not been well explained by any models over a vast range of galaxy parameters. A full understanding of these issues requires both better theoretical explorations and comparisons with new and existing observations. Here, I develop theoretical models of both radiation pressure- and supernova-driven galactic winds, and compared these models with observations.
First of all, galactic winds driven from uniformly bright self-gravitating disks radiating near the Eddington limit, which is relevant to rapidly star-forming galaxies and gravitating AGN disks. I show that uniformly bright self-gravitating disks radiating at the Eddington limit are fundamentally unstable to driving large-scale winds. I apply this theory to galactic winds from ultra-luminous infrared galaxies (ULIRGs) that approach the Eddington limit for dust, and find that the asymptotic velocity of the wind is v_infty ~ 1.5v_rot and that v_infty ~ SFR^{0.36} for hydrodynamically coupled gas and dust, where v_rot is the disk rotation velocity and SFR is the star formation rate. These results are in agreement with recent observations, but neglect the potential of a dark matter halo, bulge, or extended passive disk. A more realistic treatment including the large scale gravitational potential shows that the flow can either be unbound, or bound, forming a ``fountain flow" with a typical turning timescale of t_turn ~ 0.1-1 Gyr, depending on the ratio of the mass and radius of the galactic disk to the mass and break (or scale) radius of the dark matter halo or bulge. Importantly, I note that because t_turn is longer than the star formation timescale (gas mass/SFR) in the rapidly star-forming galaxies and ULIRGs for which our theory is most applicable, if rapidly star-forming galaxies are selected as such, they may be observed to have strong outflows along the line of sight with a maximum velocity v_max comparable to ~ 1.5v_rot, even though their winds are eventually bound on large scales.
Secondly, galactic superwinds may be driven by very hot outflows generated by overlapping supernovae within the host galaxy. We use the Chevalier & Clegg (CC85) wind model and the observed correlation between X-ray luminosities of galaxies and their SFRs to constrain the mass loss rates (\dot{M}_hot) across a wide range of star formation rates (SFRs), from dwarf starbursts to ultra-luminous infrared galaxies. We show that for fixed thermalization efficiency and mass loading rate, the X-ray luminosity of the hot wind scales as L_X ~ SFR^2, significantly steeper than is observed for star-forming galaxies: L_X ~ SFR. Using this difference we constrain the mass-loading and thermalization efficiency of hot galactic winds. For reasonable values of the thermalization efficiency (<~ 1) and for SFR >~ 10 M_sun/yr we find that \dot{M}_hot/SFR <~ 1, significantly lower than required by integrated constraints on the efficiency of stellar feedback in galaxies, and potentially too low to explain observations of winds from rapidly star-forming galaxies. In addition, we highlight the fact that heavily mass-loaded winds cannot be described by the adiabatic CC85 model because they become strongly radiative.
Furthermore, efficient thermalization of overlapping supernovae within star-forming galaxies may produce a supernova-heated fluid that drives galactic winds. For fiducial assumptions about the timescale for Kelvin-Helmholz (KH) instabilities from high-resolution simulations (which neglect magnetic fields) we show that cool clouds with temperature from T_c ~ 10^2-10^4 K seen in emission and absorption in galactic winds cannot be accelerated to observed velocities by the ram pressure of a hot wind. Taking into account both the radial structure of the hot flow and gravity, we show that this conclusion holds over a wide range of galaxy, cloud, and hot wind properties. This finding calls into question the prevailing picture whereby the cool atomic gas seen in galactic winds is entrained and accelerated by the hot flow. Given these difficulties with ram pressure acceleration, we discuss alternative models for the origin of high velocity cool gas outflows. Another possibility is that magnetic fields in cool clouds are sufficiently important that they prolong the cloud's life. For T_c = 10^3 K and 10^4 K clouds, we show that if conductive evaporation can be neglected, the KH timescale must be ~ 10 and 3 times longer, respectively, than the values from hydrodynamical simulations in order for cool cloud velocities to reach those seen in observations.