High signal-to-noise echelle spectroscopy of quasar absorption-line systems in the direction of HS 1946+7658

We have obtained a high signal-to-noise (40 less than or equal to S/N less than or equal to 80) high-resolution (FWHM = 20 km s(-1)) spectrum of the radio-quiet QSO HS 1946+7658 (z(em) = 3.051) with the echelle spectrograph on the KPNO 4 m telescope. We detect 11 metal systems in the direction of this QSO, including two Mg 11 systems, five C IV systems, two damped Ly alpha systems, and two associated systems. We use the apparent column density technique and profile fitting to measure the heavy element column densities and to assess the effects of absorption-line saturation. Profile fitting indicates that three of the C IV systems are narrow with b < 8 km s(-1). This implies that T < 50,000 K, and therefore these systems are probably photoionized. The abundance patterns in the damped Ly alpha systems are strikingly similar to those observed in low-metallicity Milky Way stars and suggest that these absorption lines are due to galaxies in early stages of chemical enrichment (see Lu et al.). The prominent associated system at z(abs) = 3.0496, 3.0504 is detected in H I, C II, C IV, Si II, Si III, Si IV, Al II, Al III, and N V. The high ion column density ratios in this associated system imply that the gas is more highly ionized than the Galactic halo. The high degree of ionization is not surprising given the extraordinary luminosity of the QSO (Hagen et al.). To study ionization and abundances in this associated system, we compare the observed column densities to a series of CLOUDY models in which photoionization by the QSO is the dominant ionization mechanism. For the input radiation held, we have used the various spectral energy distributions of HS 1946+7658 observed by Kuhn et al. The model that best fits the observed column densities of singly and doubly ionized species has solar relative and absolute abundances, but models with absolute metallicities a few times greater than solar fit comparably well. None of the models produce enough Si IV and C IV, but the QSO flux near the ionization potentials of these ions is uncertain. This photoionization modeling raises a dilemma: the high metallicity implied by the best models suggests that the associated absorption occurs near the nucleus of the QSO where the star formation rate is likely to be enhanced, but the ionization parameter from the best model combined with a density upper limit from C II* implies that the distance between the QSO and the associated absorber is greater than 300 kpc. We briefly discuss possible explanations of these discordant conclusions, including the possibility that the gas was enriched near the QSO nucleus and then ejected. By combining our HS 1946+7658 data with similar quality observations from the literature, we find that the number density of C IV systems with rest equivalent width W-r > 0.03 Angstrom is dN/dz = 7.1 + 1.7. This is similar to 3 times larger than the number density implied by a sample of C IV systems with W-r > 0.15 Angstrom, which indicates that the number of C IV systems per unit redshift is dominated by very weak lines. An attempt to detect a turndown in the C IV equivalent width distribution, which could indicate that some Ly alpha clouds contain no metals, resulted in ambiguous conclusions due to the small sample size. We briefly discuss recent oscillator strength revisions which are likely to have a significant impact on QSO absorption studies.