We present a survey of 10 galaxy groups and 23 galaxy clusters aimed at explaining the presence of warm, ionized filaments in the cool cores of galaxy clusters. By combining deep, high spatial resolution H-alpha, far-UV, and X-ray data from the Maryland Magellan Tunable Filter, Hubble Space Telescope, and Chandra X-ray Observatory, respectively, we have assembled the most complete picture of these mysterious filaments to date. This extensive database has allowed us to shed new light on two critical questions: i) Where does the cool gas in these filaments come from? ii) What process or processes are responsible for ionizing the cool filaments.

As a pilot project, we obtained high-resolution H-alpha and far-UV data for Abell1795 and find that the previously-discovered H-alpha filament is, in fact, two very thin, intertwined filaments extending 50 kpc and with a width < 700 pc. The clumpy UV morphology and UV/H-alpha flux ratios of these filaments suggest that they may consist of chains of star-forming regions. Based on these data we conclude that the H-alpha emission is a result of photoionization by young stars and that the cool gas filaments are a byproduct of the intracluster medium cooling onto a fast-moving central galaxy.

When we consider the full sample of 23 galaxy clusters, we find several strong correlations between the X-ray and H-alpha data. In general, complex, extended H-alpha filaments are found in clusters with cool, low-entropy cores. Furthermore, the morphology of the warm gas is correlated with the soft X-ray morphology and the filaments are found to occupy regions where, locally, the intracluster medium is cooling fastest. Finally, we find a strong correlation between the mass of of gas cooling below X-ray temperatures and the mass of gas in the warm filaments. These results provide strong evidence that the ionized filaments are a result of highly-asymmetric runaway cooling in the intracluster medium.

By extending this sample to include 10 galaxy groups we are able to probe more than 2 orders of magnitude in mass and cooling rate. We find that there is only a weak correlation between the presence of ionized filaments and the total mass of the system. Instead, the presence of ionized filaments appears to depend almost entirely on the core properties, specifically whether there is a cool, low-entropy core. We find that groups are, in general, cooling more efficiently that clusters, due to their lower starting temperature. This can only be the case if cool-core groups are experiencing exclusively weak AGN feedback, which we show is the case.

Finally, using new far-UV data from the Hubble Space Telescope, we find that 12/15 systems with H-alpha emission are consistent with being photoionized by young stars. The three remaining systems are under-luminous in the far-UV for their H-alpha luminosity, suggesting an alternative ionization source such as fast shocks or significant internal reddening. When we supplement this sample with UV data from GALEX and IR data from Spitzer we find a correlation between the star formation rate and the ICM cooling rate for 32 systems. The inferred efficiency of stars forming out of the cooling ICM is 14(+18/-8)%, which is consistent with the Universal fraction of baryons in stars. These results suggest that, for the majority of cases, young star formation provides sufficient UV flux to ionize the cool filaments.