The dynamic ionized gas halos of spiral galaxie

A key element in understanding how galaxies evolve over time lies in understanding the evolution of the interstellar medium (ISM) and star-formation driven gaseous halos. In particular, studies of diffuse ionized gas (DIG) are crucial in tracing the disk-halo connection in normal spirals. This work examines the origin, excitation, and kinematics of these layers.;In an effort to add observational evidence to the notion of a dynamic halo, Halpha images of a sample of edge-on galaxies are presented. These data are compared with existing HI and radio continuum maps with a specific goal of attempting to find correlations consistent with a "chimney" model of the ISM. Comparisons among these three tracers have revealed, for the first time, correlations on the scale of individual extraplanar filaments in a galactic disk. NGC 5775 in particular, shows correlations between Halpha filaments, H I shells, and radio continuum spurs that are consistent with a chimney model.;In order to explore how DIG layers are energized, we have assembled deep long-slit spectra for a sample of edge-on galaxies. In each case, photoionization models fail to explain both the observed runs with height off the galactic midplane of [S II]/[N II] and [O III ]/Halpha. Composite photoionization/shock models are generally more successful than photoionization models alone in explaining the observed line ratios. We also consider whether the data can be explained by a pure photoionization model with an increasing gas temperature with z. The runs of most of the line ratios in each of the galaxies are consistent with such an increase. However, in most cases, the unusually high ionization fractions of O++ implied by the data cannot be reconciled with predictions of a pure photoionization model.;Finally, to better understand halo kinematics and rotation, a model has been developed which assumes clouds ejected from the disk cycle ballistically through the halo. From this model, synthetic velocity profiles in z have been generated for the purpose of comparing to velocity centroid data from the previously mentioned spectra. In each case, DIG kinematics cannot be explained by a purely ballistic model.