Conclusions

We studied sub-DLAs at z < 0.6 and z > 1.5 using spectra from HST-COS and Magellan-MIKE, respectively. Prior to this work, properties of sub-DLAs in these redshift regimes remained largely unexplored. Our observations have increased the sub-DLA metallicity sample atz <0.6 by_∼4 and added _∼ 50 % more to the metal- licity data atz >1.5, providing improved constraints on the metallicity evolution of sub-DLAs at these redshift regimes. All of the low redshift sub-DLAs in our sample are found to be of near or super solar metallicity in agreement with the predictions from galactic chemical evolution models. Among the sub-DLAs observed at z >1.5, one system atzabs = 1.76 was found with [Zn/H]>−0.06 and, more surprisingly, two

systems with [Zn/H]= +0.25 dex and [Zn/H]= ₋0.02 dex were found at zabs > 2.

These two systems are the most metal-rich sub-DLAs known so far at zabs & 2 and

our observations suggest that metal-rich sub-DLAs appear at high redshifts as well. In addition to determining the S or Zn based metallicities for these absorbers, we measured abundances of various elements from our spectra. Combining our data with those for sub-DLAs from the literature, we studied various chemical and kine-

matic properties of the sub-DLA population as a whole and compared their behavior with that of DLAs. The main conclusions derived from this work are summarized as follows:

• DLAs and sub-DLAs show evolution in their NHI-weighted mean metallicities

over the redshift ranges for which their data are available (0< z.5 for DLAs and 0< z.3 for sub-DLAs). At 0< z .3, sub-DLAs, on average, have been found to be more metal-rich than DLAs and show metallicity evolution at a rate similar to that for DLAs. We also find that while metallicity evolution in DLAs does not resemble the expected mean trend for chemical enrichment in galaxies, the sub-DLA data are consistent with the chemical evolution models at all redshifts probed so far. We note that our study of the metallicity evolution in DLAs and sub-DLAs is based on the most comprehensive sample of depletion- free metallicity measurements.

• Simple photoionization calculations suggest that while there is a significant amount of ionized gas in some of our absorbers, the ionization corrections to the element abundances are relatively modest (. 0.2 dex). Thus, ionization corrections cannot explain the high sub-DLA metallicities.

• The observed difference between the metallicities of DLAs and sub-DLAs can not be explain by the presence of a selection (based on strong metal lines) bias in the low-z samples. The high redshift samples are selected based on neutral gas absorption cross-section only, and therefore, should not have any effect of such a selection bias.

• For DLAs and sub-DLAs, the metallicity of the absorbing gas is anti-correlated with the H I column density. This anti-correlation is evident even for a non- homogeneous sample of metallicity measurements based on several different

heavy elements (Zn, S, Si, Fe). This provides evidence against the existence of a bimodal metallicity distribution in DLAs and sub-DLAs.

• DLA and sub-DLA absorbers show a correlation between metallicity (based on Zn) and depletion (estimated from the abundance ratio between Fe and Zn), suggesting that more metal-rich absorbers are also likely to have a higher dust- to-gas ratio. There could be a stronger dust obscuration bias for DLAs than for sub-DLAs if total dust content and therefore dust extinction depends strongly on metal column density.

• DLAs and sub-DLAs show different [Mn/Fe] vs. metallicity trends, suggest- ing different stellar populations in the galaxies traced by these two classes of absorbers.

• The velocity dispersion (∆v90) vs. metallicity data for DLAs and sub-DLAs taken together appear to show a correlation suggesting that more massive ab- sorber galaxies are likely to be more metal-rich as well. Furthermore, the ∆v90 vs. metallicity relations for sub-DLAs and DLAs appear to be different from each other. If ∆v90 is an indicator of the mass of the absorbing galaxy, then our finding suggests that the populations of galaxies traced by DLAs and sub-DLAs obey different mass-metallicity relations and that sub-DLAs are likely to trace more massive galaxies than DLAs.

• We present preliminary evidence that at least in the case of some metal-rich sub- DLAs, the interstellar cooling rate (and also possibly the SFR) can be higher than that seen in DLAs.

Thus, the DLA and sub-DLA quasar absorber line systems, in addition to the difference in their H I column densities, appear to be distinct from each other in various aspects such as metallicity, stellar populations and gas kinematics. It is very

likely that the gas-rich and metal-poor DLAs trace different populations of galaxies than the metal-rich but gas-poor sub-DLAs. It is possible that the sub-DLAs arise in more massive galaxies while DLAs represent gas-rich dwarf galaxies with low star formation rates. Continued study of the properties of DLA and sub-DLA quasar absorbers across cosmic time can thus give new insights into the processes driving galaxy evolution.