A complete sample of 21-cm absorbers at z ∼ 1.3: Giant Metrewave Radio Telescope survey using Mg II systems

We present the results of a systematic Giant Metrewave Radio Telescope (GMRT) survey of 21-cm absorption in a representative and unbiased sample of 35 strong MgII systems in the redshift range: z(abs) similar to 1.10-1.45, 33 of which have W-r >= 1 angstrom. The survey using similar to 400 h of telescope time has resulted in nine new 21-cm detections and stringent 21-cm optical depth upper limits (median 3 sigma optical depth per 10 km s(-1) of 0.017) for the remaining 26 systems. This is by far the largest number of 21-cm detections from any single survey of intervening absorbers. Prior to our survey, no intervening 21-cm system was known in the above redshift range, and only one system was known in the redshift range 0.7 <= z <= 1.5. We discuss the relation between the detectability of 21-cm absorption and various properties of UV absorption lines. We show that if MgII systems are selected with the following criteria, MgII doublet ratio <= 1.3 and W-r(Mg I)/W-r(Mg II) >= 0.3, then a detection rate of 21-cm absorption up to 90 per cent can be achieved. We estimate n(21), the number per unit redshift of 21-cm absorbers with W-r(Mg II) > W-o and integrated optical depth T-21 > T-o, and show that n(21) decreases with increasing redshift. In particular, for W-o = 1.0 angstrom and T-o > 0.3 km s(-1), n(21) falls by a factor of 4 from < z > = 0.5 to < z > = 1.3. The evolution seems to be stronger for stronger Mg II systems. Using a subsample of systems for which high-frequency Very Long Baseline Array (VLBA) images are available, we show that the effect is not related to the structure of the background radio sources and is most probably due to the evolution of the cold neutral medium filling factor in MgII systems. We find no correlation between the velocity spread of the 21-cm absorption feature and W-r(Mg II) at z similar to 1.3.