The submillimetre and near-infrared properties of Herschel -ATLAS sources.

Abstract

In this thesis I investigate a sample of galaxies that are detected with the Herschel Space Telescope in the sub-millimetre wavelength range and that also have near-infrared detections with the VISTA telescope in Chile as part of the VIKING survey. The first necessity is to find the near-infrared galaxies that are most likely to be the counterparts to the Herschel galaxies. I accomplish this by using a likelihood ratio method which I modify to allow for an appropriate estimate of the probability of finding a genuine near-infrared counterpart above the magnitude limit to a SPIRE source. This probability is found to be Q0 0.73. 51% of the SPIRE sources have a best VIKING counterpart with a reliability R = 0.8, and the false identification rate of these is estimated to be 4.2%. I expect to miss 5 per cent of true VIKING counterparts. There is evidence from Z - J and J - Ks colours that the reliable counterparts to SPIRE galaxies are marginally redder than the field population. I obtain photometric redshifts for 68% of all (non-stellar) VIKING candidates with a median redshift of ˜z = 0.405. I have spectroscopic redshifts for 3147 ( 28%) of the reliable counterparts from existing redshift surveys. Comparing to the results of the optical identifications supplied with the Phase I catalogue, I find that the use of medium-deep near-infrared data improves the identification rate of reliable counterparts from 36% to 51%. I investigate the evolution of the sub-millimetre luminosity function (LF) using the sample of SPIRE sources with reliable counterparts in VIKING with z 1. I find strong evolution of the 250 μm LF out to about redshift z = 0.6 and possibly out to z = 0.8 in broad agreement with previous studies. A double-power law seems to fit the local LF (z 0.2) slightly better than a Schechter function and we find a flatter slope at lower luminosities as compared to recent studies. i Finally, I construct the star formation rates (SFR) from far-infrared (FIR) and ultra-violet (UV) luminosities of the SPIRE sample with reliable VIKING counterparts (SFRFIR and SFRUV respectively) and show that the contribution of the SFRFIR increases with increasing luminosity. UV observations are hence crucial for all but the brightest SPIRE galaxies in calculating a total SFR. Calculating the slope of the UV continuum and comparing with the ratio LFIR/LUV leads to dust attenuation corrected SFRUV that represent the total SFR well in the low to medium LFIR range.