THE UNITED KINGDOM INFRARED TELESCOPE
ANNUAL REPORT
1995 AND 1996

2. Scientific Results during 1995 and 1996

2.2. Selected Scientific Results

2.2.1. IRCAM and the Oldest Objects in the High-Redshift Universe

J.S. Dunlop (ROE)

Recently Steidel and co-workers (Steidel et al. 1996) have discovered a substantial population of star-forming galaxies at . This has revolutionised optical cosmology by allowing the study of radio-quiet galaxies at high redshifts. However their selection method, which depends on a Lyman continuum break superimposed on an otherwise blue far-UV continuum, can shed little light on the evolutionary state of the most passively evolving systems of a given epoch. This is unfortunate since, given the ease with which a relatively small starburst can mask the true properties of an underlying galaxy, it is the reddest/most-passive systems at any redshift which are of greatest interest for constraining the first epoch of galaxy formation and indeed the age of the Universe.

Radio-based selection has long provided an alternative and effective method of locating high redshift galaxies which, at least in principle, should not be so directly biassed towards star-forming sources. Indeed, this method should be biassed towards the precursors of old elliptical galaxies since at low-redshifts it is well-established that the hosts of powerful radio sources are elliptical galaxies with well-evolved stellar populations. Despite this, identification of high-redshift objects on the basis of extreme radio power has also yielded optically active objects with the majority of high-redshift radio galaxies studied to date displaying complex elongated optical/UV morphologies, relatively blue optical-ultraviolet continuum colours, and strong emission lines. However, the fact that the optical-ultraviolet properties of high-redshift radio galaxies are known to correlate with radio power suggests that any radio-based search for ``normal'' elliptical galaxies at high redshift should be confined to milli-Jansky flux levels. Accordingly, over the past few years we have investigated the properties of weak radio galaxies with mJy from the Leiden Berkeley Deep Survey, and, using IRCAM1 and IRCAM3 on UKIRT, have isolated a sample of 10 extremely red objects with R-K > 5 and K > 18 for intensive study.

The two reddest objects in this sub-sample have proved to be of particular interest; 53W091 with R-K , and 53W069 with R-K . Follow-up IRCAM3 J-band and H-band photometry indicated that the large R-K colours of both these objects are primarily due to a strong 4000Å break which red-shifted to a rest wavelength of around 1m (rather than, for example, overall reddening of the spectral energy distribution - hereafter SED - by dust, as has often proved to be the cause of red colours in high-redshift objects). Thus our UKIRT observations strongly suggested that in these two objects we had identified the first clear examples of well-evolved elliptical galaxies at .

While a red R-K colour can be taken as indicative of an old stellar population, deep optical spectroscopy is vital for the reliable dating of these objects for four reasons. First, a spectroscopic redshift is required. Second, it is necessary to show that the red colour of the object arises from a lack of young stars rather than, for example, from a dust-reddened active nucleus. Third, the shape of the rest-frame ultraviolet spectrum of a galaxy is extremely sensitive to the age of the stellar population. Fourth, for high-redshift galaxies it should be possible to use evolutionary synthesis models to derive relatively robust age estimates from ultra-violet SEDs. This is because, for the potential age range of interest (< 8 Gyr for z > 1) the ultraviolet SED is completely dominated by stars near the main-sequence turnoff point on the HR diagram and disagreements over, for example, the strength and colour of the AGB or HB are unimportant.

  
Figure 1: The rest-frame spectrum of 53W091 compared with an instantaneous starburst at an age of 3.5 Gyr (green line) and a transposedIUE spectrum of the elliptical M32.

We have now obtained deep optical spectra of both of these red mJy radio galaxies using LRIS on the Keck telescope. The UKIRT-based redshift estimates for both galaxies have proved to be remarkably accurate, with 53W091 yielding a spectroscopic redshift of , and 53W069 lying at . Our redshift determination and spectral dating of 53W091 were published in Dunlop et al. 1996. In brief, the ultraviolet SED of this source is, as illustrated in Figure 1, very similar to those of low-redshift ellipticals such as M32, and essentially identical to that of an F6V star. Both these comparisons suggest an age of Gyr, a result confirmed by spectral synthesis modelling. The SED of 53W069 is, if anything, slightly redder and preliminary analysis yields a best fit age of 4 Gyr.

The existence of such old well-evolved objects at challenges the conventional Einstein-de Sitter cosmology in which, for a Hubble constant the Universe is predicted to be less than 3.2 Gyrs old at . Certainly, almost regardless of adopted cosmology, the ages of these objects indicate that at least some giant elliptical galaxies formed the bulk of their stars at redshifts z > 5. What is beyond dispute is that our UKIRT-based study of mJy radio galaxies has yielded the oldest known objects in the high-redshift Universe.

I gratefully acknowledge my collaborators in this work: John Peacock, Hy Spinrad, Arjun Dey, Rogier Windhorst, Raul Jimenez, Daniel Stern and Ian Waddington.

References

Dunlop et al. 1996, Nature, 381, 581
Steidel et al. 1996, ApJ, 462, L17