We have made a detailed study of the opportunities for UKIDSS to extend the explorable redshift range substantially beyond that reachable by Sloan. Figure 2.3 (left) (section 2.7) shows how Fan et al. (2000) found the highest-redshift quasar currently known, z=5.8. At redshift z~6, in i'-z', quasars are redder than all stellar objects except brown dwarfs, but bluer than brown dwarfs in z'-J. Redshift z~6 quasars can be found by searching in the selection box shown. In principal this method works up to about z=6.4, but reaching beyond z=6 may be difficult for Sloan because of the limited depth of the near-infrared data provided by 2MASS, and because the quasars rapidly dim in z' beyond z=6 due to absorption in the Ly-alpha forest. More z=6 quasars will be found by employing the deeper J data from UKIDSS, but the redshift limit can only be extended substantially by searching at longer wavelengths.
Using z' rather than i' as the dropout filter is an order of magnitude more difficult. This is because in the red colour, which would now be either J-H or J-K (or indeed H-K), the quasars are no longer bluer than the brown dwarfs. This is illustrated in two plots in Figure 2.4. Using combinations of the z'JHK filters, quasars in the redshift range 6<z<7.2 have very similar colours to cool M stars, and L dwarfs. Either of the two-colour diagrams shown in Figure 2.4 become effective at higher redshift still, beyond z=7.2, but three things act against such a search: (1) the colour limit z'-J>3.2 is very red, requiring very deep z' data, (2) the quasars are fainter, as further away, and (3) the quasars are rarer, because of the decline in space density. Clearly a z'JH survey is too ambitious as the first step beyond Sloan - there may not be any z>7 quasars.
Figure 2.4. Two-colour diagrams showing how the z'
filter would be used as the dropout band. LHS is the
z'JH diagram, and RHS is the z'JK diagram. The
symbols and colour coding are as in the main text. Using either J-H
or J-K as the red colour, quasars do not stand out clearly from the
cool M stars and L dwarfs. Quasars at redshift z>7.2 could be selected
by identifying objects lying in the box shown. Note that in J-H and
J-K, brown dwarfs become bluer as they cool along the sequence L to
T.
The discussion above explains the motivation for introducing the new Y filter, and surveying the redshift range 5.8<z<7.2, as laid out in the main text.¹ In fact it is difficult to see how this redshift range can be explored in any other way. This approach has the advantage that the optical data already exist, and that the depth is relatively easily achievable, because the filter is sampling the quasar continuum.
WFCAM in any case is not really competitive for surveys in the z' band, because optical cameras with much wider fields of view are being planned. On a similar timescale to the UKIDSS surveys one can envisage deeper i' and z' data becoming available. The J and H data of the UKIDSS LAS will reach deep enough to make a survey for quasars in the redshift interval 7.2<z<8 feasible. If UKIDSS finds quasars at redshifts as high as z=7, this will provide the argument for a wide-field very deep survey in z', using VISTA for example. As noted in the main report, a survey covering 4000 sq. degs to z'=22.7 would be required.
¹ We also considered a narrow filter Yn at 1.05-1.07micron, where the sky is darker. This filter is faster, but because of the longer wavelength does not provide discrimination between T dwarfs and quasars.