Large groups of quasars in an ultraviolet-excess survey

Abstract

This thesis describes the objectives, design, construction and analysis of the Chile-UK Quasar Survey, a new wide-field spectroscopic redshift survey of ultravioletexcess quasar candidates. The main objectives of the survey are the construction
of a large, homogeneous catalogue of quasars with redshifts 0.4 < z < 2.2, and the use of the catalogue to investigate the large-scale structure of the early Universe, particularly at large scales ('.-' 100 h' Mpc).
The survey covers 140 deg 2 , with a 100 x 100 programme area centred at 10h40m, +50 00', and a 6?3 x 6?3 control area centred at 11h14'n, —29°00' (131950 coordinates). Quasar candidates have been selected on 16 < b :~ 20 and (Uj - b) < —0.3 from UK Schmidt Telescope photographic plates in the U and J bands digitized on the SuperCOSMOS machine at the Royal Observatory Edinburgh. The methods developed for astrometric and photometric calibration of the SuperCOSMOS data are described in detail. Positional uncertainties of Er = 0.69 arcsec r.m.s. and magnitude uncertainties of zXb 3 = 0.10 mag and Auj = 0.12 mag are achieved across the whole survey area. A function relating positional uncertainty to the efficiency of light collection for spectrographs with optical-fibre input is derived and used to show that the efficiency for this survey is > 80 per cent.
Survey observations using a multi-object spectrograph with optical-fibre input on the 2.5-m du Pont telescope at the Las Campanas Observatory in Chile are reported. Spectra of 4588 candidates from 54 deg 2 of the survey area have been obtained in 12 nights of observations conducted in 1996 March, 1997 March and 1998 May. Of these spectra, 3134 have been analysed, resulting in the identification of 320 quasars with redshifts in the range 0.228 < z < 2.768, and 93 narrow emission-line galaxies with redshifts in the range 0.015 < z < 0.397. The distributions of the quasar and NELG samples as functions of position, redshift, apparent magnitude, colour and luminosity are described in detail. It is argued that the quasar sample is highly homogeneous and at least 70 per cent complete. Identifications of objects without emission lines are discussed.
A minimal spanning tree analysis of a homogeneous sample of 227 quasars drawn from the survey programme area catalogue identifies two large groups of quasars with high statistical significance. The first group contains 6 quasars with median redshift z = 1.236 and mean separation 26 ± 9 h' Mpc. The null hypothesis that the quasars in this group are distributed randomly is rejected by the (m, a) test of Graham et al. (1995) at the 0.0011 level, and the probability that this is a random significant event is 5 per cent. The shape of the group suggests a single filament with a length of - 110 h' Mpc and a width of " 15 h' Mpc. This group coincides with the core of the Clowes & Campusano (1991) group, the reality of which is confirmed by this independent survey. The second group is a new discovery. It contains 13 quasars with median redshift z = 1.507 and mean separation 32 ± 7 h' Mpc. The (m,a) test finds that this group is significant at the 0.0020 level, and the upper limit on probability that this is a random significant event is 13 per cent. The shape of the second group resembles three intersecting filaments with an overall size of - 150 x 80 x 85 h-3 Mpc3, making it one of the largest structures in the Universe found so far.
The use of measures of the Civ A 1549 broad emission line to determine the intrinsic luminosity of quasars is investigated, using spectra from a luminosity-limited sample of survey quasars. The observational evidence and theoretical basis for a
power-law correlation between luminosity and equivalent width (the Baldwin Effect) are reviewed. A model that predicts a power-law correlation between luminosity and the velocity width of broad emission lines is described. The Baldwin Effect
correlation is detected at the 2.4-a level. A positive slope for a luminositysrelocity width correlation is detected in agreement with the proposed model but at lower significance.