Hot Gas in Star-Forming Galaxies
Starburst-Driven Superwinds
A significant fraction of all the heavy elements created by stars in galaxies are ejected by some process into the inter-galactic medium (IGM), with dwarf galaxies losing 80% of their metals and even massive galaxies like our own losing 10% of their metals (Tremonti et al. 2004).
Hot metal-enriched superwinds from starburst galaxies show the clear kinematic evidence for galaxy-scale outflows in both the local and high-redshift Universe (e.g. Heckman et al. 2000, Adelberger et al. 2003) of any of these classes of outflow. Such outflows have been observed in star-forming galaxies of all masses and environments, from dwarf starbursts to ultraluminous merging galaxies, as is required to explain the observed galaxy mass-metallicity (M-Z) relationships. Almost all starburst galaxies show some evidence for outflow (Lehnert and Heckman 1995), and a substantial fraction (25%) of all (massive) star formation occurs in starbursts within the local universe. Starburst activity was more common in the past. The local space-density of starburst-driven winds of a given size and apparent power is approximately an order of magnitude greater than similar large-scale uncollimated winds from AGN.
Hot gas around normal disk galaxies - Red: H-α (WIM), Green: R-band (starlight), Blue: Diffuse soft X-ray (3 million-degree gas). The region covered by each image is 20 × 20 kpc. Intensity scale in square-root.
Superwinds are driven by merged core-collapse supernova ejecta and stellar winds, which initially create a 108 K metal-enriched plasma within the starburst region. This over-pressured gas expands and breaks out of the disk of the host galaxy, converting thermal energy into kinetic energy in a bi-polar outflow, which can potentially reach a velocity of 3000 km s-1. This tenuous wind-fluid sweeps up, entrains, accelerates and possibly shock-heats the cooler, denser ambient disk and halo. Theoretical models predict that the entrained, cool gas is accelerated to lower velocities than the hot metal-enriched gas (e.g. Chevalier and Clegg 1985, Strickland and Stevens 2000). These models also predict that the majority of the energy (90%) and metal content in superwinds exists in the hot 106 K phases, with the kinetic energy of such gas being several times the thermal energy. X-ray observations are thus of singular importance in studying this phenomenon, as they provide a probe of the most energetic phases in these outflows, in particular of the metal-enriched phases most likely to escape into the IGM. Yet all existing observational velocity measurements of superwinds are of entrained cooler material, e.g. warm neutral and ionized gas with measured outflow velocities in the range 200-1000 km s-1. Whether this material escapes into the IGM is uncertain, since vobs ~ vescape for the host galaxy (see Heckman et al. 2000).
X-ray surface brightness profiles. We observe exponential X-ray surface brightness profiles with scale height ~2-4 kpc. Spectral hardness, Q, varies only weakly in halo, at z > 2 kpc. These observations are consistent with the wind hitting and shocking pre-existing exponential halo medium.
Despite a wealth of multi-wavelength data on starburst-driven outflows, the fundamental parameters of the absolute element abundances and velocity of the hot gas have not been measured. Observations with Chandra and XMM-Newton detect thermal X-ray emission from hot gas in superwinds that extends out to 5-30 kpc from the plane of edge-on starburst galaxies (e.g. Strickland et al. 2004a,b), but these observations lack the spectral resolution (ΔE = 100 eV at E = 1 keV) to robustly determine the metal abundance and kinematics of this hot gas. High-resolution X-ray spectroscopy (with Δv ~ 400 km s-1 or ΔE = 1 eV at 0.65 keV) is the only method by which these parameters can be measured and by which the efficiency of winds in ejecting metals can be quantified.
References
Adelberger, K., et al., 2003, ApJ, 584, 45Heckman, T. M., Lehnert, M. D., Strickland, D. K., and Armus, L., 2000, ApJS, 129, 493
Strickland, D., et al., 2004a, ApJS, 151, 193
Strickland, D., et al., 2004b, ApJ, 606, 829
Tremonti, C., et al., 2004, ApJ, 613, 898
