Galaxies are central to our understanding of the universe. They trace the geometry of space-time and give birth to stars and planets. Galaxy formation theory accurately describes the gravitational amplification of structure in the matter distribution over cosmic time. Unfortunately, most of the (dark) mass density in the universe is not normal baryonic material, so the theory does not directly predict the observable feature of galaxies -- starlight. Research in Crystal Martin's research focuses on the astrophysics of galaxy formation and evolution, trying to understand in detail why the star formation rate varies widely among galaxies. Central to this work is the idea of feedback from supernova explosions, which inject energy, momentum, and heavy elements into the surrounding interstellar gas. The new conditions dictate whether the gas clouds form another generation of stars. Observations demonstrate that supersonic shock waves, driven by the combined energy of many supernova explosions, sweep interstellar gas into large shells. The Figure below shows an example of the chaos that ensues when these shells break out of a galaxy carrying the heavy elements synthesized by massive stars. Polluting the surrounding intergalactic gas this way affects the formation of new galaxies because most of the normal matter in the universe has yet to cool down and condense into galaxies. Such extreme feedback, capable of ejecting material from a galaxy, requires an unusually high density of supernova explosions, probably only reached when two galaxies collide. The resulting net outflow of gas from a starburst galaxy is called a galactic wind. The impact of this reheating, or feedback, on the galaxy population and the intergalactic medium is quite pronounced. For example, it limits the neutral, condensed fraction of baryons to about 10% of the total. It also remains a mystery why the nucleosynthetic products of star formation are found widely dispersed in the intergalactic medium outside galaxies. Gas cooling and reheating are thought to determine the basic form of the galaxy luminosity function. The shape of the luminosity function differs from that of the halo mass spectrum. The deviation is in the sense that star formation must be most efficient at a mass scale similar to that of our own Galaxy and become much less efficient in both the smallest halos and the largest halos. The high-mass cutoff was originally thought to arise from the long cooling times of the gas in massive (i.e. hot) halos but this overproduces luminous galaxies. Some mechanism to suppress halo gas cooling or continually reheat it is needed. Recent suggestions include radiation pressure from active nuclei and radio jets. Starburst-driven winds are expected to have the largest impact on dwarf galaxies owing to their shallow gravitational potential. They may explain the absence of young galaxies with very low mass after reionization, the faint-end slope of the galaxy luminosity function, and the mass-metallicity relation among galaxies.
Nature or Nurture:
Graduate student Taro Sato used several hundred spectra from the Keck telescopes to study the impact of galaxy environment on the star formation history of moderate-redshift (z~0.4) galaxies. By using an emission-line selected sample of galaxies on the outskirts of a massive galaxy cluster, he was able to include lower mass galaxies than previous studies at similar epochs. Our results suggest that galaxy - galaxy interactions trigger star formation well outside the cluster core. The unusually high fraction of composite "e(a)" spectra compared to field samples, however, suggests that some cluster-specific mechanism is also at work. We think this is likely related to the dynamical assembly of the cluster because this particular cluster, Abell 851, is accreting several groups of galaxies.
To quantify the impact of galactic winds on galaxy evolution, Crystal led a series of observational programs with the HIRES and ESI instruments at the Keck telescopes to determine whether current dynamical descriptions of the outflows are accurate. Models predict, and observations confirm, that the hot wind is laced with cooler material likely entrained at the interface between the hot, supernova-heated wind and the gaseous galactic disk. The kinematics of the cool material can be observed in absorption (via resonance lines) imprinted on the starburst continuum. Graduate student Collen Schwartz analyzed extremely high-resolution spectra and found low outflow velocities in dwarf starbursts compared to the speeds in more luminous starbursts -- see Schwartz & Martin (2004) . This result was at first surprising since the nearly uniform x-ray temperatures imply a terminal speed for the hot wind that is indepdendent of galactic mass. Adding new measurements from Keck echellete spectra, Crystal recently showed that cool gas is accelerated to the predicted speed of the hot wind in ultraluminous galaxies and argued that dwarf starburst winds simply lack enough momentum (essentially mass in this case) to accelerate the cooler gas to the velocity of the hot wind (Martin 2005) . Postdoc Akimi Fujita used a powerful computer at Los Alamos National Laboratory and a state-of-the-art hydrodynamics code to simulate the interaction of the wind and with the galactic halo (Fujita et al., in prep).
Most recently, the absorption from this cool gas was mapped across 18 ultraluminous starbursts. The large angular extent over which outflowing gas is detected is shocking. Martin (2006) demonstrates that the outflows rotate in a few cases, argues that this rotating portion of the cool outflow must have a low scale height, and concludes that the starburst activity is spread over a much larger area than that covered by the nucleus of the ULIG. The paper also provides empirical estimates of the wind efficiency.
Chemical Enrichment by Starburst Winds:
The Chandra X-ray satellite provides a direct glimpse of hot, supernova-shocked galactic outflows. Crystal and her collaborators measured gas-phase metallicity in a hot galactic wind. See Martin, Kobulnicky, & Heckman (2002) to find out how. Their work suggests most of the heavy elements synthesized by the starburst are expelled from the galaxy, a result which helps explain the overall high level of chemical enrichment observed in the intergalactic medium. Crystal reviewed the role of galactic winds in the chemical evolution of the Universe for the Carnegie Astrophysics Series -- see Martin (2004) .
Finding the First Galaxies:
The nature of the first stars is not well understood, but most models suggest they formed in dwarf galaxies during the Dark Ages when the universe was mostly neutral hydrogen. Sometime after first light, more massive galaxies capable of sustaining star formation collapsed and quickly reionized most of the hydrogen in the intergalactic medium. The Epoch of Reionization thus marked a distinct change in the galaxy population as well as a phase transition in most of the baryonic material. Recently there has been significant progress in establishing when Reionization occurred, and Crystal has undertaken a major initiative to develop observational techinques that can reveal the galaxy population responsible for Reionization. Crystal is using large ground-based telescopes to search for emission-line objects at frequencies matched to hydrogen Lyman-alpha line emission from the pre- and post-reionization galaxies. A pilot project at Keck Observatory with Marcin Sawicki put a significant limit on the evolution (in number and in luminosity) of star-forminggalaxies out to redshift 6. Learn more by reading Martin, & Sawicki (2004) . A large number of high-redshift candidates were found more recently using the ultra-wide field of view of the IMACS camera. on the Magellan telescope. See Martin, Sawicki, Dressler, and McCarthy for a description of the candidates. Our recent follow-up observations of these candidates appear to confirm some! Stay tuned!
Star Formation Thresholds:
Crystal's work with Rob Kennicutt has shown that dynamical instabilities in gaseous disks accurately describe where star formation occurs. Graduate student Colleen Schwartz is studying the role of feedback in regulating the star formation rate in normal, i.e. not starbursts, galaxies. Galaxy environment, in addition to self-regulation, may well play a critical role in determining why some galaxies turn interstellar gas into stars much more efficiently than others. Graduate student Taro Sato is studying the impact of galaxy environment on star formation history, focusing on the outskirts of the rich galaxy cluster Abell 851. He has analyzed hundreds of spectra of cluster galaxies taken with the LRIS and DEIMOS spectrographs at Keck and will measure the dependence of the star formation rate, star formation history, and galactic metallicity on the local density of galaxies.
So, you want to study winds?
Funding from the David & Lucile Packard Foundation, the Alfred P. Sloan Foundation, the UCSB/LANL CARE program, and NASA supports this work. Top figure is the Hubble Ultra Deep Field courtesy of STScI. Bottom figure shows the galactic wind emanating from the nearby dwarf galaxy NGC 1569; image from Martin, Kobulnicky, Heckman 2002.