Selecting damped Lyman α systems through CaII absorption - I. Dust depletions and reddening at z~ 1

We use the average E(B-V) and ZnII column densities of a sample of z~ 1 CaIIλλ 3935, 3970 absorption-line systems selected from the fourth data release of the Sloan Digital Sky Survey (SDSS) to show that on average, with conservative assumptions regarding metallicities and dust-to-gas ratios, they contain column densities of neutral hydrogen greater than the damped Lyman α (DLA) limit. We propose that selection by CaII absorption is an effective way of identifying high column densities of neutral hydrogen, and thus large samples of DLAs at zabs= 0.5Å), is ~20-30 per cent that of DLAs, after correcting for the significant bias against their detection due to obscuration of the background quasars by dust. On average these absorbers have E(B-V) >~ 0.1mag the dustiest absorbers show depletions of refractory elements at a level of the largest depletions seen in DLAs. For the first time we can measure the dust-to-metals ratio in a sample of absorption-selected galaxies, and find values close to, or even larger than, those observed locally. All of these properties suggest that a substantial fraction of the CaII absorbers are more chemically evolved than typical DLAs. There is a trend of increasing dust content with Wλ3935 this trend with strong-line equivalent width is also observed in an equivalent, but much larger, sample of MgII absorbers. Such a trend would result if the dustier systems are hosted by more massive, or disturbed, galaxies. Follow-up imaging is required to provide conclusive evidence for or against these scenarios. From consideration of the E(B-V) distribution in our sample, and assuming CaII absorbers represent a subset of DLAs, we calculate that dust obscuration causes an underestimation in the number density of DLAs by at least 8-12 per cent at these redshifts. Finally, the removal of broad absorption line (BAL) quasars from the SDSS quasar sample increases the sensitivity of the detection of reddening by intervening absorbers. To this end, we describe a new, automated, principal component analysis (PCA) method for identifying BAL quasars.