The students attending this lecture course are living at a special time! During your lifetime you can reasonably confidently expect to witness the location of the sources responsible for cosmic reionization, to determine the redshift range over which they were active and perhaps even witness directly the `cosmic dawn' when the first stellar systems shone and terminated the `Dark Ages'!
We have made such remarkable progress in the past decade, reviewed
here, that such a bold prediction seems reasonable, even for a cautious
individual like myself! We have extended our fundamental knowledge of
how various populations of galaxies from 0 < z < 3 combine
to give us a broad picture of galaxy evolution, while extending the
frontiers to z 
7
and possibly beyond. Certainly many issues remain, including the apparent
early assembly of certain classes of quiescent galaxy, the abundance
patterns in the intergalactic medium and the apparent `downsizing'
signatures seen over a variety of redshifts. However, the progress has
been rapid and driven by observations. Accompanying this is a
much greater synergy between theoretical predictions and observations
than ever before.
We have seen that the question of `First Light' - the subject of this course - is the remaining frontier for observations of galaxy formation. The physical processes involved are poorly understood and thus observations will continue to be key to making progress. In this final lecture, I take out my crystal ball and consider the likely progress we can expect at optical and near-infrared wavelengths with current and near-term facilities ahead of those possible with the more ambitious new observatories such as the James Webb Space Telescope (JWST) and a new generation of extremely large ground-based telescopes.
Panoramic imaging with large optical and near-infrared cameras
on telescopes such as Subaru, UKIRT and VISTA will enable
continued exploration of the abundance of luminous drop-outs and
Ly
emitters over 5 <
z < 7, reducing the currently troublesome
issues of cosmic variance. Deeper data will continue to be
provided from further fields taken with ACS and NICMOS.
With further spectroscopic surveys on large telescopes, we
can expect improved stellar mass density estimates at z
5-6
(Eyles et al 2006).
However, strong gravitational lensing may still be the only
route to probing intrinsically fainter sources, particularly beyond
z 7.
More detailed characterization of the properties of the
most massive galaxies at z
5-6 will also be
worthwhile, in addition to continuing to deduce statistical properties
such as abundances and luminosity functions.. Laser-guide
star adaptive optics is now in widespread use on our 8-10 meter
class telescopes and, with integral field spectrographs, can be
used to probe the resolved dynamics of distant galaxies
(Genzel et al 2006).
Application of these techniques to the
most intense star-forming lensed systems at z > 4
(Swinbank et al 2006,
Figure 61) is already providing unique insight
into the physical
state of the earliest galaxies. Extending the same techniques with adaptive
optics will advance progress and address process on remarkably
small scales ( 200 pc).
 |
 |
Figure 61. Detailed studies of a lensed
z |
A number of special purpose ground-based instruments are also being
developed, both with and without adaptive optics, to probe
for z > 7 dropouts and Lyman
emitters.
These include:
10-18
cgs in a night of VLT time. This corresponds
to a limiting star formation rate of
1
M
yr-1 in quiet regions of the airglow spectrum at z
7.7
and 9.9. Given the modest field and the need for dedicated
spectroscopic follow-up, a long term campaign is envisaged
to conduct a reliable census.
emitters and
the abundance of intrinsically faint examples.
Ellis et al (2001)
found the lensed emitter at z
5.7 was less
than 30 milli-arcsec across with a line width of only 100-200
kms sec-1. If such tiny emitters are the norm at z
> 7, the GGS may go much deeper than extant
spectroscopic lensed searches
(Stark et al 2006b).
The exciting redshift range 7 < z < 12 will also be the
province of improved drop-out searches using the
instrument WFC3 slated to be installed on Hubble
Space Telescope in 2008-9
11.
The infrared channel of this instrument spans 850 - 1170 nm
with a field of 2
arcmin at an angular resolution of
0.13 arcsec pixel-1. This resolution is coarser than
that of adaptive optics-assisted instruments such as F2T2.
The principal gain over ground-based instruments will be
in deep broad-band imaging free from airglow. The survey
efficiency is about an order of magnitude better than that of
NICMOS. WFC3 also has two infrared grisms which will be
helpful in source discrimination.
A major stumbling block at the moment, even at z
5-6,
is efficient spectroscopic follow-up of dropout candidates.
As we discussed in Section 6, photometric
redshifts have unfortunately become
de rigeur in statistical analyses of luminosity densities
and luminosity functions (Bouwens et al 2006), yet their precision
remains controversial. For z- and J-band dropous beyond z
7, photometric
redshifts will be even less reliable. Spitzer detections will
be harder and fewer, and the typical source may have only
2-3 detected bands. Candidates may be found in abundance
but how will they be confirmed?
The various 8-10m telescopes are now building a new generation of cryogenic near-infrared multi-slit spectrographs. It can be hoped that long exposures with these new instruments will be sufficient to break this impasse.
A number of facilities are being planned, motivated by the progress discussed in these lectures. These include:
 |
Figure 62. The 6.5m James Webb Space Telescope: Then (2013 in orbit) and Now (2005, full scale model). |
 |
Figure 63. The proposed US-Canadian Thirty Meter Telescope (www.tmt.org) now in the detailed design phase. |
These will complement redshifted 21cm line tomography with radio facilities and address the key questions of the escape fraction of photons from star-forming sources and how they create ionized bubbles which merge to cause reionization.
It is quite likely that, by 2013, the redshift range containing the earliest galactic sources, estimated at present to be 10 < z < 20 perhaps, will have been refined sufficiently by special-purpose instruments on our existing 8-10 meter class telescopes. Thus one can surmise that that both JWST and future ELTs will be used for much more challenging work related to the physical process of reionization, as well as the chemical maturity of the most luminous sources found at high redshift.
 |
Figure 64. Flexibility in the deployment
strategy for the multiple
integral field units (yellow squares) for the proposed TMT
infrared multi-object adaptive optics assisted instrument
IRMOS. The IFUs can be distributed around the full
field of 2 arcmin in classic multi-object mode. Alternatively,
the IFUs can be configured together to map a small
6 × 6 arcsec field at high angular resolution in `blind'
mode. Such flexibility will be important in mapping
Ly |
An obvious partnership between JWST and TMT,
for example, which would complement the 21cm
studies, would be to (i) search for the extent and
topology of faint Ly
emission in ionized bubbles around JWST-selected luminous star
forming galaxies and, (ii) pinpoint early sources
for spectroscopic scrutiny so as to identify
signatures of Population III stars.
At the present time there are so many imponderables in our
knowledge of the earliest sources, that even the design
parameters for the ELT instruments is a considerable challenge.
How big are the faintest
Ly emitters? What are
the typical line widths in km sec-1? How big are
the ionized bubbles at a given redshift? And, crucially,
what is the surface density of various types of star-forming
galaxies. Flexibility in the design of survey strategies
will be crucial with instruments such as TMT's IRMOS
(Fig. 64).
Any information we can glean on the properties
of z 10 sources
in the next 5 years will
be valuable in optimizing how to move forward when
these magnificent new generation telescopes are
made available to us. Indeed, it is foolhardy to wait!
Time and again, we can retrospectively look back
at what we thought we would accomplish with our
planned facilities and we always find that we achieved
more than we expected!
I thank Daniel Schaerer, Denis Puy and Angela Hempel for inviting me to give these lectures in such a magnificent location with an enthusiastic group of students. I also thank my fellow lecturers, Avi Loeb and Andrea Ferrara and all of the foregoing for their patience in waiting for the completion of my written lectures. I thank my close colleagues Kevin Bundy, Sean Moran, Mike Santos and Dan Stark for their help and permission to show results in progress as well as Ivan Baldry, Jarle Brinchmann, Andrew Hopkins, John Huchra and Jean-Paul Kneib for valuable input. Finally, I thank Ray Carlberg and his colleagues for their hospitality of the Astronomy Department at the University of Toronto where the bulk of these lectures notes were completed.
9 http://www.ast.cam.ac.uk/~optics/dazle/ Back.
10 http://odysseus.astro.utoronto.ca/ggs-blog/?page_id=2 Back.
11 http://www.stsci.edu/hst/wfc3 Back.