C IV absorption in a new sample of 55 QSOs - Evolution and clustering of the heavy-element absorption redshifts

We have obtained spectra with a uniform signal-to-noise ratio of 20: 1 and with a resolution of either 0. 8 or 1. 5 A of 52 QSOs in the redshift range 1. 8 <= zₑm_ <= 3. 56. Most of the spectra cover the wavelength range from just longward of the C IV λ1549 emission line down to the Lyman-α emission line. In some cases additional spectra of the Lyman-α forest region were also obtained. The QSOs were selected for observation using criteria which were independent of knowledge of the QSOs' absorption spectra. The spectra were supplemented with published spectra of three QSOs which had been obtained with the same observational parameters to produce a uniform sample of 55 QSOs. Lists of absorption lines were generated for each QSO using identical methods and using criteria which ensure a high probability that the lines are real. Redshifts were determined by standard methods. A total of 229 absorption redshifts was found in which the C IV λλ1548, 1550 doublet was identified. Uniform samples were isolated consisting of redshifts for which the rest equivalent widths of both C IV lines exceed 0. 15 A. All of the redshifts in these samples could have been detected in any of the spectra. Some of these fall into clumps which are clearly not independent entities. Therefore, additional samples (called the "Poisson samples") were constructed in which clumps containing more than one system on a scale of less than 1000 km s^-1^ were treated as one system. An additional sample of 40 redshifts based on the Mg II λλ2796, 2803 doublet was also generated; these redshifts will be discussed elsewhere. The following principal results were obtained from a study of the C IV redshifts: 1. There is no significant tendency for the absorption redshifts to cluster around the emission redshift of the QSO, regardless of the radio properties of the QSO. 2. An application of the Bahcall-Peebles test shows that the large-scale distribution of redshifts along the line of sight is consistent with that expected based on the intervening hypothesis for the origin of the lines. 3. The distribution of the number of absorbers per QSO shows a marginally significant departure from the Poisson distribution expected for randomly distributed intervening absorbers. This is true at the 89% confidence level. It is proposed that the small range in redshift (DELTAz ~ 0. 6) available in the spectrum of any given QSO causes the detail of the large-scale distribution of galaxies along the line of sight to play an important role in the statistics. 4. The number density of absorption systems per unit redshift range N (z) decreases with increasing redshift in the range 1. 3 <= zₐbs_ <= 3. 4 according to the relation N (z) is proportional to (1 + z) ^-1. 2+/-0. 7^ if expressed as a power law. All of the redshift samples show the same qualitative behavior. 5. There is a systematic change in the mean C IV doublet ratio with redshift. Taken in conjunction with the preceding result this is most naturally interpreted as a decrease in the mean C IV abundance with redshift. 6. The two-point correlation function for the C IV redshifts is generally flat, as would be expected for absorbers randomly distributed along the line of sight. In addition, there is significant clustering of the redshifts on scales up to 600 km s^-1^ in the rest frame of the absorbers. Analysis shows that the observed clustering is unlikely to be due to relative motions of clouds in galaxies, but instead is a manifestation of galaxy-galaxy clustering. The amplitude of the clustering is compared with the predictions of published simulations of the evolution of clustering in "cold dark matter" cosmological models. The observed clustering amplitude is significantly larger than that predicted by any of these simulations. Thus, there is a need to invoke "biasing" in order to understand the galaxy clustering at z ~ 2, as well as at z = 0. 7. There is no significant correlation between the number of absorbers in the line of sight to a given QSO and its intrinsic luminosity. Such a correlation is to be expected if gravitational lensing significantly influences the luminosities of the apparently brightest QSOs. 8. Overall, both the clustering properties and the evolution in redshift of the absorbers responsible for the C IV lines are significantly different from those of the "Lyman-α clouds. " Thus, a recent claim that the two kinds of absorbers are part of one population cannot be true.