SIMBAD references

Recent studies of quasi-stellar objects (QSOs) with ROSAT suggest the existence of a significant population of soft X-ray weak QSOs (SXW QSOs) where the soft X-ray flux is ∼10-30 times smaller than in typical QSOs. Why are these QSOs soft X-ray weak, and what is their relationship to broad absorption line QSOs (BAL QSOs) and X-ray warm absorber QSOs? As a first step in a systematic study of these objects, we establish a well-defined sample of SXW QSOs that includes all α_ox≤-2 QSOs from the Boroson & Green sample of 87 Bright Quasar Survey QSOs. SXW QSOs comprise ~11% of this optically selected QSO sample, and we find soft X-ray weakness in both radio-quiet and radio-loud QSOs. From an analysis of C IV absorption in the 55 Boroson & Green QSOs with available C IV data, we find a remarkably strong correlation between α_ox and the C IV absorption equivalent width. This correlation suggests that absorption is the primary cause of soft X-ray weakness in QSOs, and it reveals a continuum of absorption properties connecting unabsorbed QSOs, X-ray warm absorber QSOs, SXW QSOs, and BAL QSOs. Many of our SXW QSOs have ultraviolet absorption that is intermediate in strength between that of X-ray warm absorber QSOs and that of BAL QSOs, and their X-ray absorption is also likely to be of intermediate strength. From a practical point of view, our correlation demonstrates that selection by soft X-ray weakness is an effective (≳80% successful) and observationally inexpensive way to find low-redshift QSOs with strong and interesting ultraviolet absorption.

We have also identified several notable differences between the optical emission-line properties of SXW QSOs and those of the other Boroson & Green QSOs. SXW QSOs show systematically low [O III] luminosities and equivalent widths as well as distinctive Hβ line profiles. They tend to lie toward the weak [O III] end of Boroson & Green eigenvector 1, as do many low-ionization BAL QSOs. Unabsorbed Seyfert galaxies and QSOs with similar values of eigenvector 1 have been suggested to have extreme values of a primary physical parameter, perhaps mass accretion rate relative to the Eddington rate (M{dot}/M{dot}_Edd). If these suggestions are correct, it is likely that SXW QSOs also tend to have generally high values of M{dot}/M{dot}_Edd. Finally, we present and discuss correlations between α_ox and other QSO observables after removal of the SXW QSOs.