In this thesis I use the UKIDSS Ultra-Deep Survey to investigate the evolution in clustering of the galaxy population across the redshift range $0<z<3$. Selected in the K-band, the sample does not suffer as greatly from biases against passive and dusty objects that plague optically-selected samples, and so represents the first analysis of a relatively unbiased galaxy sample above z=1. The unique combination of area (0.77 sq. deg.) and depth (K(AB)<24) allows the analysis of both rare, intrinsically bright objects and faint objects, reaching ~L* at z=3. The UDS data are complemented by similarly deep optical data from Subaru and mid-infrared data from Spitzer. These data have enabled accurate photometric redshifts to be computed, which are used extensively throughout this thesis. The combined optical and infrared data are used to firstly select passive and star-forming samples over a wide range in redshift 1<z<2.5. The results of their number counts and clustering properties motivate a more in-depth study of the respective clustering strengths of passive and star-forming galaxies. The main result of this thesis is that passive galaxies are found to be more strongly clustered than star-forming galaxies to at least z=1.5, irrespective of rest-frame K-band luminosity. Furthermore, within either the passive or star-forming sample, luminosity segregation is found to be very mild. These results therefore indicate that even at fixed stellar mass, passive galaxies are more strongly clustered. Clustering strength is directly related to the mass of the typical host dark matter halo. Galaxies which are observed to be passive at z<1.5 are therefore hosted by more massive halos than their star-forming counterparts. This finding is consistent with downsizing in galaxy formation. Downsizing is also found in the clustering strengths of the star-forming sub-samples. The correlation lengths of star-forming galaxies smoothly decline towards z=0. Galaxies of a given K-band luminosity or stellar mass are therefore found in less massive halos in the low-redshift universe. Passive galaxies on the other hand appear to have constant r0 values over 0<z<1.5. If confirmed, this may indicate a preferred mass of dark matter halo for the transition from star-forming to passive. Above z=1.5 the clustering strengths of passive and star-forming galaxies are found to be more similar. If this trend continues it suggests that they may become equal within 2<z<3. This would therefore constitute the epoch over which the passive/star-forming bimodality becomes distinct. These results are then examined more closely in a marked-correlation function analysis and tested against the one of the latest theoretical galaxy formation models. Due to difficulties in defining a suitable quantity for `passivity' from photometric data, the results from the marked correlation functions are largely inconclusive at this stage. However, the potential in their future use is demonstrated. My results highlight two known issues in the semi-analytic galaxy formation model: the over-production of high-redshift sources and red satellites. In addition I offer tentative evidence for further possible discrepancies, in clustering properties at high-redshift.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:530910 |
| Date | January 2010 |
| Creators | Hartley, William G. |
| Publisher | University of Nottingham |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://eprints.nottingham.ac.uk/11558/ |
Page generated in 0.0014 seconds