Interstellar Mg II absorption lines from low-redshift galaxies

We have used the GHRS aboard HST to search for interstellar Mg II λλ2796, 2803 absorption from the disks and halos of 17 low-redshift galaxies, using as probes QSOs and extragalactic supernovae whose sight lines pass close to, or through, intervening galaxies. The galaxies studied are of diverse morphological type, reside in different environments, and lie at separations of ρ′ ≃ 2-113 h^-1 kpc from a QSO line of sight. Ten of 11 galaxies at separations 31-113 h^-1 kpc show no absorption to equivalent width limits of W(λ2796) < 40-90 mÅ, which corresponds to N(Mg II) ≃ 1-4 × 10¹² cm^-2. Six galaxies lie at ρ′ ≤ 9 kpc, and of these, four (NGC 4319, the LMC, M81, and the Milky Way) show absorption. Two early-type galaxies (NGC 1380 and Leo I) show no absorption at ρ′ < 9 kpc: these nondetections are surprising because the separations are small and point to the possibility that the existence of extended absorbing halos may be a function of galaxy type. All of the galaxies which produce absorption are plausibly members of interacting systems. For absorbing galaxies probed below 9 kpc, the sight line passes within the optical radius of the galaxy, where the interstellar medium (ISM) is expected to have a high covering factor, and we do not attribute the absorption to the interactions. However, we do find that the environment of the absorbing galaxies affects the characteristics of the absorption detected - the strength of lines, the complexity of line components, the ionization state of the gas - and we warn of the dangers inherent in constructing models of generic halos based on statistical properties of QSO absorption-line surveys. Our data suggest that the covering factor of Mg II absorption is high for galaxies within ≈ 10 kpc, but very small beyond ≈ 30 h^-1 kpc, a result consistent with the size found of Mg II halos deduced for galaxies at redshifts z > 0.2. The low-redshift galaxies observed in this study which show Mg II absorption are probably drawn from the same population as the higher redshift absorbers, although the sample is too small to permit a definitive analysis. There is some evidence for a sharp cutoff in equivalent width as ρ′ increases, suggesting that low-redshift galaxies may be sharply bounded beyond ∼30 h^-1 kpc. The luminosities of most of the absorbers in our program are such that the source sight lines pass through the predicted (Holmberg) optical radii of the galaxies. Hence the region between the optical radius and the gas radius of the z > 0.2 Mg II absorbers, ≈ 10 kpc to 30 h^-1 kpc, remains unexplored at low redshift. Although our data do not enable us to determine whether the outer regions of galaxy disks are responsible for Mg II absorption systems, we note that 21 cm measurements currently available show that H I disks of the brightest spiral galaxies are already close to the sizes of higher redshift Mg II absorbers. We have found a correlation between the absolute blue magnitude of spiral galaxies and their H I radii measured at the N(H I) = 10¹⁹ and 10²⁰ cm^-2 level, and there is evidence that the correlation tends to that of the intermediate-redshift Mg II absorbing galaxies at lower N(H I) limits. Hence the outer regions of galaxy disks may well have sufficient column densities to produce Mg II absorption lines.