| Depth (5-sig) | either z'=21.5 or i'=23.5 (n.b. Sloan filters, Vega mags) |
|---|---|
| Area | 1500 sq. degs |
| Region | 10h<RA<15h, -10<Dec<+10 |
| Timescale | completion by end 2005 |
| Clear nights | approx. 40-50 (depending on filter and exact VST efficiency) |
Notes: 1. This proposal covers high-redshift z=7 quasars and brown dwarfs. For the brown dwarfs z'=21.5 is preferred and required. For the quasars i'=23.5 is slightly preferred over z'=21.5. Also for the quasars strictly speaking a survey depth 0.5 mag. less is just sufficient. But by going slightly deeper the reduced photometric errors will improve completeness and reduce contamination.
2. An area in the NGP is specified, because the publicly available SDSS spectroscopic coverage of this area adds further value to the dataset. For this reason this is the probable target of the UKIDSS Large Area Survey first 2-year plan. However other equatorial fields near the SGP e.g. near 22h-4h are also feasible.
This submission sets out the opportunities for VST to extend the power of the UKIDSS Large Area Survey (LAS) for finding quasars of very high redshift 5.8<z<7.2, as well as the coolest brown dwarfs, beyond type T. The strategy for finding quasars in the redshift interval 7.2<z<8 is also discussed.
The presentation is centred around the quasar work. But the opportunities for studies of very cool brown dwarfs are emphasised. In addition the high quality images (better sampling than UKIDSS, better seeing and depth than Sloan) will add value to the multiwavelength analysis of galaxies in the Sloan+UKIDSS database.
The motivation for this work is set out in the UKIDSS Science Case, under sections 2(ii), 2(vii) and 2 (x). Further details are provided in the recent paper by Warren and Hewett 'WFCAM, UKIDSS, and z=7 quasars'. For the quasars, briefly, the goal is to study redshifts higher than any reached, exploring the epoch of reionisation. The current redshift record z=6.3 belongs to Sloan, but their filters prevent them reaching beyond z~6.4. At z=6.3 we appear to be on the threshold of the epoch of reionisation, which is predicted to occur over a short period, and is in the nature of a phase transition. So the redshift range 6<z<8 is of great interest. With the proposed survey, by end 2005 UKIDSS+VST+VLT should have discovered between 5 and 7 quasars 6.4<z<7.2, and ESO astronomers will be the first to study this fundamental epoch in the history of the universe. Many additional VLT science projects arising from these discoveries can be expected.
As explained in the above references quasars up to z=7.2 can be picked out using (i'-Y) + (Y-J) or (z'-Y) + (Y-J) colours. Beyond z=7.2 selection is much harder.
Concentrating firstly on i', the current UKIDSS proposal is limited to Y<19.0 because it has a colour cut i'-Y>3 and Sloan only reaches i'=22.0. What is needed is a VST survey reaching at least 1 mag deeper than Sloan. With a VST survey to i'>23.0 we can search to Y=20.0 (the 5-sigma limit of the UKIDSS LAS is Y=20.5). By reaching deeper many more high-redshift quasars will be found. Here is a comparison of the effect:
| current | 4000 sq.degs | Y=19.0 | i'=22.0 | finds 10-11 quasars 5.8<z<7.2 |
|---|---|---|---|---|
| with VST | 1500 sq.degs | Y=20.0 | i'=23.0 | finds 15-23 quasars 5.8<z<7.2 |
| with VST | 1500 sq.degs | Y=20.0 | i'=23.0 | finds 5-7 quasars 6.4<z<7.2 |
i.e twice as many quasars from only the first 1500 sq. degs of the UKIDSS LAS. Depending on success the VST area covered could be increased, to cover as much as possible of the total 4000 sq. degs of the LAS.
As regards the z' band, the Sloan survey reaches z'=20.0. This is just as good as i'=22.0 in terms of depth, because quasars will be selected by colour z'-Y>1. However using z' the redshift range detectable is restricted to 6.4<z<7.2 i.e. with UKIDSS at present, using the Sloan survey, one can detect the same z>6.4 quasars using either z' or i' as the short filter. Therefore if VST reached 1 mag. deeper than Sloan (i.e. to z'=21) over 1500 sq. degs this would be as effective as i'=23 for z>6.4. However VST is expected to be slower in z' relative to i' than Sloan is, because Sloan uses special long-wavelength sensitive CCDs for z'. So in terms of speed, and because it includes also the redshift range 5.8<z<6.4 a survey to i'=23 is preferred over z'=21.0 for the quasar work. Either is acceptable and the choice could depend on other science proposed for surveys similar in scope, such as the brown-dwarf work mentioned below.
The depths quoted above, i'=23.0, z'=21.0, are the minimum useful depths. The quasar magnitude distribution will be skewed towards the faint end, and selection of nearly all the quasars will rely on an upper limit to the short-wavelength band flux (i' or z') determined at S/N=5. It is highly desirable to reduce the photometric error on this quantity by reaching 0.5 mag. deeper. This will increase the completeness of the sample and reduce contamination. This has the very desirable effect of making follow-up observations on the VLT more efficient. For these reasons the proposed depth is i'=23.5, or z'=21.5.
In summary it can be seen that by reaching >1 mag. deeper than Sloan the VST can greatly enhance the value of the UKIDSS LAS.
As described in the UKIDSS science case, to have a reasonable chance of finding quasars beyond z=7.2 requires extremely deep z' data. A survey over 4000 sq. degs to z'=22.7 is needed to find 3-4 quasars 7.2<z<8 in the UKIDSS LAS. This does not look practical for VST, and really requires a 4m telescope with z-sensitive CCDs (the original VISTA optical camera would have been suitable).
The UKIDSS Large Area Survey (LAS) will reach Y=20.5, J=20.0, H=19.2 K=18.4. As stated above Sloan reaches z'=20.0. L and T dwarfs have Y-J=1.0 to 2.0 and z'-J even larger. Thus at present, at the survey limit, the UKIDSS LAS will detect many L and T dwarfs only at J since the T dwarfs are blue in J-H and J-K. Thus the LAS would be greatly enhanced by a deeper z', say to z'=21.5, 5 sigma. The i' band is not very helpful for the coolest brown dwarfs, as i'-J colours are huge.
The depth z'=21.5 is estimated to require 900sec exposures (1" seeing, 1.2 airmass, assuming VST reaches 0.3 mag. deeper than WFI). Allowing for overheads, and the long nights (at the preferred NGP RAs), this means that 1500 square degrees can be completed in less than 50 clear nights. For brown dwarfs ideally we would wish to go even deeper at z', but that does not seem very realistic. The depth i'=23.5 should be slightly faster (Sloan is 2 mag faster in i' than z', but uses z' sensitive CCDs i.e. with VST i' should be more than 2 mag faster than z').
Since the UKIDSS LAS survey is available to ESO astronomers, as well as this proposed VST survey data, the science of exploring the epoch of reionisation is accessible to any ESO astronomer. We expect the UKIDSS LAS first 2-year plan to include the required Y, J coverage over 1500 sq. degs i.e. the infra-red data would be available by end 2005. So this sets the timescale for obtaining the VST data.
We predict 5 to 7 quasars will be found in the redshift range 6.4 to 7.2, computed using the luminosity functions of Fan et al., and Schmidt, Schneider, and Gunn. VLT will be required firstly to confirm the quasars (Y=20 is challenging for a 4m telescope). Then the VLT would be used to obtain very deep spectra as well as deep imaging below Ly-alpha to quantify absorption (particularly the inverse effect). Near-ir spectra to search for metal lines will be of great interest. Narrow-band imaging to look for the predicted Lya halo (to study the IGM at this time), and deep imaging to look for galaxy companions can also be expected. These will all be challenging observations requiring the photon-gathering power, and range of instrumentation, of the VLT.