Intermediate- and high-velocity ionized gas toward ζ orionis

We combine near-UV spectra obtained with the Hubble Space Telescope GHRS echelle with far-UV spectra obtained with the Interstellar Medium Absorption Pro le Spectrograph and Copernicus to study the abundances and physical conditions in the predominantly ionized gas seen at high velocity ( 105 km s(-1) less than or similar to v(circle dot) less than or similar to 65 km s(-1)) and at intermediate velocity (-60 km s(-1) less than or similar to v(circle dot) less than or similar to -10 km s(-1)) along the line of sight to the star zeta Ori. We have high-resolution (FWHM similar to 3.3-4.5 km s(-1)) and/or high signal-to-noise ratio spectra for at least two significant ions of C, N, Al, Si, S, and Fe enabling accurate estimates for both the total N (H II) and the elemental depletions. C, N, and S have essentially solar relative abundances; Al, Si, and Fe appear to be depleted by about 0.8, 0.3-0.4, and 0.95 dex, respectively, relative to C, N, and S. While various ion ratios would be consistent with collisional ionization equilibrium (CIE) at temperatures of 25,000-80,000 K, the widths of individual high-velocity absorption components indicate that T similar to 9000 +/- 2000 K-so the gas is not in CIE. Analysis of the C II ne-structure excitation equilibrium, at that temperature, yields estimates for the densities (n(e) similar to n(H) similar to 0.1-0.2 cm(-3)), thermal pressures (2n(H)T similar to 2000-4000 cm(-3) K), and thicknesses (0.5-2.7 pc) characterizing the individual clouds. We compare the abundances and physical properties derived for these clouds with those found for gas at similar velocities toward 23 Ori and tau CMa and also with several different models for shocked gas. While the shock models can reproduce some features of the observed line profiles and some of the observed ion ratios, there are also significant differences between the models and the data. The measured depletions suggest that roughly 10% of the Al, Si, and Fe originally locked in dust in the preshock medium may have been returned to the gas phase, consistent with recent predictions for the destruction of silicate dust in a 100 km s(-1) shock. The observed near-solar gas-phase abundance of carbon, however, appears to be inconsistent with the predicted longer timescales for the destruction of graphite grains.