Starburst-driven winds may have a strong influence on structure formation at high redshifts, on the "porosity" of star-forming galaxies (i.e. the probability for ionizing photons to escape their host galaxies), hence the nature of the extragalactic UV and infrared background, and on the chemical and thermal evolution of galaxies and their environment. Due to space limitations, this review only addresses the last issue.
4.1. Heating and Enrichment of the ISM and IGM
Hot Metal-Enriched Gas. Nuclear starbursts inject both mechanical energy and metals in the centers of galaxies. This hot, chemically-enriched material, the driving engine of galactic winds, is eventually deposited on the outskirts of the host galaxies, and contributes to the heating and metal enrichment of galaxy halos and the IGM. Surprisingly little evidence exists for the presence of this enriched wind fluid. This is due to the fact that the wind fluid is tenuous and hot and therefore very hard to detect in the X-rays. The current best evidence for the existence of the wind fluid is found in M 82 (Griffiths et al. 2000; Stevens et al. 2003), NGC 1569 (Martin, Kobulnicky, & Heckman 2002), and possibly the Milky Way (e.g., Koyama et al. 1989; Yamauchi et al. 1990). The ratio of alpha elements to iron appears to be slightly super-solar in the winds of both NGC 1569 and M 82, as expected if the stellar ejecta from SNe II are providing some, but not all of the wind fluid.
Selective Loss of Metals. The outflow velocities in LIRGs and ULIRGs do not appear to be correlated with the rotation velocity (or equivalently, the escape velocity) of the host galaxy, implying selective loss of metal-enriched gas from shallower potentials (Heckman et al. 2000; Rupke et al. 2002). If confirmed over a broader range of galaxy masses (e.g., Martin 1999; but see the contribution by Martin at this conference for a word of warning), this result may help explain the mass-metallicity relation and radial metallicity gradients in elliptical galaxies and galaxy bulges and disks (e.g., Bender, Burstein, & Faber 1993; Franx & Illingworth 1990; Carollo & Danziger 1994; Zaritsky et al. 1994; Trager et al. 1998). The ejected gas may also contribute to the heating and chemical enrichment of the ICM in galaxy clusters (e.g., Dupke & Arnaud 2001; Finoguenov et al. 2002, and references therein).
Dust Outflows. Galactic winds also act as conveyor belts for the dust in the hosts. The evidence for a large-scale dusty outflow in our own Galaxy has already been mentioned in Section 3.1 (Bland-Hawthorn & Cohen 2003). Far-infrared maps of external galaxies with known galactic winds show extended dust emission along the galaxy minor axis, suggestive of dust entrainment in the outflow (e.g., Hughes, Gear, & Robson 1994; Alton et al. 1998, 1999; Radovich, Kahanpää, & Lemke 2001). Direct evidence is also found at optical wavelengths in the form of elevated dust filaments in a few galaxies (e.g., NGC 1808, Phillips 1993; NGC 3079, Cecil et al. 2001). A strong correlation between color excesses, E(B - V), and the equivalent widths of the blueshifted low-ionization lines in star-forming galaxies at low (e.g., Armus, Heckman, & Miley 1989; Veilleux et al. 1995; Heckman et al. 2000; Rupke et al. 2003) and moderate-to-high redshifts (e.g., Rupke et al. 2003; Shapley et al. 2003) provides additional support for the prevalence of dust outflows. Assuming a Galactic dust-to-gas ratio, Heckman et al. (2000) estimate that the dust outflow rate is about 1% of the total mass outflow rate in LIRGs. Dust ejected from galaxies may help feed the reservoir of intergalactic dust (e.g., Coma cluster; Stickel et al. 1998).