A quadruple-phase strong Mg II absorber at Z ∼ 0.9902 Toward PG 1634+706

The z = 0.9902 system along the quasar PG 1634+706 line of sight is a strong Mg II absorber [W_r(λ2796) > 0.3 A°] with only weak C IV absorption (it is "C IV-deficient"). To study this system, we used high-resolution spectra from both the Hubble Space Telescope Space Telescope Imaging Spectrograph (STIS) and the Keck I telescope High Resolution Echelle Spectrometer (HIRES). The STIS spectrum has a resolution of R = 30,000 and covers key transitions such as Si II, C II, Si III, C III, Si IV, and C IV. The HIRES spectrum, with a resolution of R = 45,000, covers the Mg I, Mg II, and Fe II transitions. Assuming a Haardt & Madau extragalactic background spectrum, we modeled the system with a combination of photoionization and collisional ionization. Based on a comparison of synthetic spectra with the data profiles, we infer the existence of the following four phases of gas: 1. Seven Mg II clouds have sizes of 1-1000 pc and densities of 0.002-0.1 cm ^-3, with a gradual decrease in density from blue to red. The Mg II phase gives rise to most of the C IV absorption and resembles the warm, ionized intercloud medium of the Milky Way. 2. Instead of arising in the same phase as Mg II, Mg I is produced in separate, narrow components with b ∼ 0.75 km s^-1. These small Mg I pockets (∼100 AU) could represent a denser phase (∼200 cm^-3) of the interstellar medium (ISM), analogous to the small-scale structure observed in the Milky Way ISM. 3. A "broad phase" with a Doppler parameter b ∼ 60 km s^-1 is required to consistently fit Lyα, Lyβ, and the higher order Lyman series lines. A low metallicity (log Z ≲ - 2) for this phase could explain why the system is "C IV-deficient" and also why N V and O VI are not detected. This phase may be a galactic halo, or it could represent a diffuse medium in an early-type galaxy. 4. The strong absorption in Si IV relative to C IV could be produced in an extra, collisionally ionized phase with a temperature of T ∼ 60,000 K. The collisional phase could exist in cooling layers that are shock heated by supernova-related processes.