5.2 The Era of the Near-Infrared
Working at near-infrared wavelengths provides the opportunity to probe to even higher redshifts; the optical searches being limited to finding objects at z 5. The earliest results in the near-infrared (1-5 µm) were limited to experiments with single-element detectors: Boughn et al. (1986) placed some of the earliest limits at 2 µm by searching for the fluctuations in the sky-brightness on scales 10-30 arcsec - an experiment optimised to detect the signal of low surface brightness PGs at z ~ 20 overlapping on the sky. Collins and Joseph (1988) carried out what was probably the first near-infrared search for discrete PGs, looking for candidate objects with the expected flat spectral energy distribution (see Figure 3). A summary of these early results is presented by Thompson et al. (1994), who conclude that with the limited field size and sensitivity available to near-infrared instruments at that time, the expected SFR of PGs could still be as high as 1000 M yr-1.
An important technological landmark for PG searches came with the advent of CCD-type array cameras for use in the 1-5 µm waveband; in 1986 the 3.8m United Kingdom Infrared Telescope (UKIRT) in Hawaii was the first telescope to have such a general-purpose camera. In fact infrared array devices were constructed as early as 1974 but the technical information came late to astronomers due to the secrecy surrounding the early development which was largely carried out for military applications.
With these new panoramic detectors near-infrared searches began to rival their optical counterparts in limiting sensitivity and areal coverage: Parkes et al. (1994) placed the first really useful limits on the number density and Lyman luminosity of objects at 7 z 9 using narrow-band filters, while Pahre and Djorgovski (1995) achieved similar flux limits by targetting the emission lines of H, [OII], H and [O II] redshifted to z = 2.88 and z = 4.79 using the Keck 10m telescope situated alongside UKIRT in Hawaii. These limits constrain the SFR of PGs to a value ~ 1-10 M yr-1, similar to the constraints achieved in the optical.
One final avenue of investigation previously undertaken concerns radio searches for neutral hydrogen: From the lack of absorption lines at wavelengths shorter than Lyman in the spectra of high redshift quasi stellar objects, it has been known for some time that the mass density in diffuse neutral hydrogen must be very small (Gunn and Peterson 1965). Nonetheless, a number of experiments have been carried out designed to detect redshifted 21cm radiation from clumps of neutral hydrogen corresponding to an evolutionary stage of PGs preceeding the epoch at which stars form (Hogan and Rees 1979; Davies et al.1978; Uson et al. 1991). A big drawback of these experiments as they currently stand is limiting sensitivity - the faintest of which corresponds to a neutral hydrogen mass ~ 1014 M at z 3, which is much bigger than any predicted characteristic mass scale (see section 3).