 |
Below is a revised summary of the science requirements for ACAM, updated following (1) a Sep 06 meeting of the astronomy group to discuss the proposed instrument and (2) Nov 06 discussions with the optical designers at Astron.
Because the field of view (fov) at aux port is small, it is used mainly for target-of-opportunity observations (SN, GRBs etc.). Significantly enlarging the field would make it attractive for a much wider range of science programmes. These potentially include some of the science currently done at prime focus, thereby reducing the number of instrument changes (PF + BLT) required.
A large increase in the fov would also greatly improve the quality of those imaging observations which can only be done at aux port, either because they can't be scheduled in advance, or because they require imaging at awkward intervals e.g. observations of SN, exoplanet transits.
The best seeing at the WHT is 0.5 arcsec, so the scale at the CCD should not be coarser than 0.25 arcsec/pixel. 2k*2k pixels would then deliver a fov of 8.3 arcmin, and 4k*4k pixels would deliver 16.7 arcmin. These correspond to gains in area over the current aux-port camera of x20 and x40 respectively. Larger fields of view would be vignetted by the Cass A&G optics.
Building a new camera for aux port provides an opportunity to add new functionality, in particular:
Ideally the camera should allow for observation over the full optical range 0.33 - 1.0 micron, but there will be a trade-off between image quality (particularly in the UV), field diameter and cost.
The aim of investing in ACAM is to increase the future scientific impact of the WHT, i.e. to give astronomers access to part of the observational parameter space likely to yield high-impact science, or to do more efficiently (lower cost, or more effective time on sky) what we do now.Although we can't predict with certainty which high-impact science programmes we should be catering for in 2 - 5 years time, we can (1) look at current imaging programmes, and the technical considerations which limit them, and (2) investigate what additional functionality can be provided at reasonable cost.
As regards the latter, an astronomical observation is a measurement of intensity I as a function of RA, dec, wavelength (w), polarisation (p) and time (t), so added functionality means extending the range over which these 6 parameters (RA, dec, w, p, t, I) can be measured, or increasing the resolution with which they can be measured.
The current aux-port camera offers:
Parameter Range of parameter Resolution in parameter
--------- ------------------ -----------------------
RA, dec 1.8' fov seeing (> 0.5")
Wavelength 3300 - 10000 A 1000 A for broad-band filters
50 A for narrow-band filters
Polarisation none none
Time limited by scheduling ~ sec
constraints
Intensity upper limit determined limited by availablity
by minimum integration of comparison objects
time ~ sec
Polarimetry is of interest to a significant fraction of users, but implementation is complicated by reflection at the 45-deg flat feeding aux port. A lot of effort would be required for the specification, design, construction and commissioning of a polarimetry module, and polarimetry will not be discussed further here. However, the final design should not exclude the possibility of introducing at some future date a wheel for polarisation components, if leaving this option open does not compromise other aspects of performance.
The other areas where functionality might be improved are: field of view; wavelength coverage (by complementing LIRIS, for spectroscopy); resolution in wavelength; and resolution in time.
The table below summarises the characteristics of some science projects (1 - 5) and areas of enhanced functionality (6 - 8) likely to deliver high-impact science over the next few years. These are discussed individually following the table.
Science Scheduling Field PSF Filters Spectroscopy
reqd (') reqd (") reqd also useful?
------- ---------- -------- -------- ------- ------------
(1) SN, GRB ToO several 0.5-1 UBVRI Yes, very
(2) Exoplanet Awkward several ~ 1 BVRI Yes, very
transits intervals,
part nights
(3) Current PF Regular typically 0.5-1 BVRI mainly Maybe
science < 10 + U, narrow
(4) Narrow-band Regular, typically 0.5-1 narrow Maybe
imaging of service < 10
galaxies
(5) Grav lensing Regular 6 - 10 0.7 R,I
in gal clus.
(6) High-thruput Regular few 0.5-1 --
spectroscopy
(7) Simultaneous Regular few 0.5-1 --
opt+IR spec.
(8) Time-resolved Regular few 0.5-1 BVRI
imag/spec
(2) Exoplanet transits
One of the biggest growth areas in astrophysics at the moment (and probably for several years to come) is the search for, and study of, planets orbiting other stars (exoplanets). One particularly promising approach involves accurately timing transits of a known exo-planet across the star. Changes in the orbital period can then reveal the presence of smaller (e.g. earth-like) planets. But to detect the partial occultations, simultaneous observation of nearby comparison stars is crucial, and this requires a reasonable field of view (several arcmin), as well as scheduling of several observations a few days apart. This will be important for follow-up observations of exoplanets discovered by SuperWASP, preferably soon after discovery (i.e. << 1 semester away), to scoop competing teams. The PSF is not so important, in fact observations may need to be carried out with the telescope defocussed. The field needs to be large enough to include several comparison stars, in order to achieve mmag photometric accuracy. If a frame-transfer CCD is used, very high photon rates can be recorded, and correspondingly high S:N achieved. If spectroscopy were also available, one could both time the transits and look for absorption features in the atmophere of the transiting planet.
(3) Current PF science
WHT PF offers:
03A - 04B 05A - 06B
Allocations 18 11
Allocations
using U filter 8 2
Allocations needing
full PF field ~12 ~3
Between these two periods (03A - 04B and 05A - 06B)
interest in using PF for cosmological
surveys (e.g. gravitational lensing, surveys for candidate objects
for GTC/OSIRIS) appears to have diminished, and there is
more emphasis on individual, lower-redshift
objects (e.g. TNOs, brown dwarfs in open clusters, satellites of
M31).
In most cases, the 05A - 06B users could have obtained the data they need with an ACAM offering 10-arcmin field, PSF < seeing (particularly near the field centre) and throughput comparable to PF. Reasonable throughput and PSF in U are desirable, but mainly at the field centre.
(4) Narrow-band imaging of galaxies
The main science driver behind an earlier attempt to upgrade
aux-port (the WAPCAM project - Knapen/Hough/Beckman)
was determination of where star formation takes
place in galaxies, building on measurements over the last decade
of evolution with redshift of the mean rate of star
formation.
An ambitious programme of narrow-band imaging of z < 0.1
galaxies is required, to complement a planned study with
GTC/OSIRIS of galaxies out to z ~ 1.
An 8-arcmin field is required.
(5) Gravitational lensing in galaxy clusters
The WHT is not currently used for gravitational lensing work,
partly because of perceived variation of the PSF (including ellipticity)
across the PF imaging field,
which hinders analysis of the weak shear caused by large-scale structure.
The reasons for this are not clear, but the fact that the CCDs are
not coplanar, and that the mosaic has to be realigned relative to
the instrument each time it is mounted, are probably contributory factors.
A field at aux port of several arcmin, with a stable, single CCD detector, and reasonable (< seeing) PSF, would be attractive for studies of the much stronger gravitational shear caused by individual clusters of galaxies.
(6) High-throughput spectroscopy
The throughput of WHT/ISIS/R158R/Mar2 is about 0.18 in R band.
For other gratings, it is lower.
Depending on the detailed design, ACAM could, in terms of photons/A/sec,
be the highest-throughput spectrograph on the mountain-top (and
probably in the northern hemisphere, as far as Dutch and Spanish users are
concerned) for some time to come,
suitable for low-resolution spectroscopy of
very faint targets e.g. extragalactic SN and GRB,
high-redshift objects.
GTC's ELMER would have higher throughput, but this is unlikely to be commissioned soon, if the planned first-light instruments (OSIRIS etc.) are ready.
Allowance should be made in the optical design for a future upgrade to accept multi-slit masks, if this does not compromise other aspects of instrument performance.
(7) Near-simultaneous optical/IR spectroscopy
Obtaining LIRIS and optical spectra (and images)
during the same night is attractive (a) for objects which may be
varying (solar system?), and (b) to observers who don't want to have
to do in two observing runs what they could achieve in one.
The pixel scales of LIRIS and ACAM (as proposed) are well matched
(both 0.25 arcsec/pixel).
(8) Time-resolved imaging/spectroscopy
Deployment of a frame-transfer CCD (like AG3)
at a secondary focus, fed by a 45-deg
flat after the field lens, would allow
high-time-resolution imaging of objects which are varying on a
short time scale, or for which short exposures are required to avoid
saturation.
The optics could either (a) duplicate those of the main camera (which are
all glass, after the field lens), or (b) be simplified, given that
the requirements for PSF and wavelength coverage could be relaxed.
This arm could also include simple dispersing optics.
E.g., to search for the signature of the atmosphere of an exoplanet transiting
its host star, high-S:N low-resolution slitless (for comparison stars)
spectroscopy is required. Asteroseismology is another possible application.
Increasing the number of optical components decreases the throughput (particularly in U), and probably raises the risk of ghosting.
Field distortion is relatively unimportant.
Broad-band imaging / ADC
Without an ADC, there will be significant dispersion over the
wavelength range covered by broad-band
blue filters e.g. ~ 1 arcsec within Harris B band (50% points) at
ZD = 50 deg.
This sounds a lot,
but it's probably better that observers work around it
(use narrower bands, observe in the blue only at high elevation) rather
than incorporate an ADC, which will cost a lot to design, build and
commission, and which will severely constrain the optics.
Narrow-band imaging
The central wavelength of the filter should
change across the field by << filter bandpass (this may not be satisfied
if the filter is in a collimated space).
A shift in the mean wavelength of the filter (due to being in a
converging beam) is less of a problem, because ING's collection of
redshifted Halpha filters covers a continuous range of wavelengths.
Throughput
Throughput is reduced at each glass/air surface by 0.5 - 1 %
(depending on the coating).
Comparable instruments e.g. EFOSC, ELMER, CAFOS typically have
~ 7 - 15 glass/air surfaces.
In Astron's baseline design, there are no reflecting surfaces after the 45-deg flat which feeds aux-port. Any additional Al surfaces would reduce throughput by ~ 10% each. A Ag surface has higher reflectivity ~ 98%, but is not useful below 4000 A. X-shooter uses a Ag+SiNx surface, which has useful reflectivity down to the atmospheric cutoff. This can be applied to surfaces as large as the A&G 45-deg flat feeding aux port.
Spectroscopy
Spectroscopy affects the design mainly because the dispersing
element needs to be in a collimated space. The longer the
collimated space, the greater the cost in terms of complexity.
In addition, a slit-mask wheel is required in the focal plane,
and calibration light needs to be fed into the system
(there are several options for this).
Possible upgrades: multi-slit masks for the focal plane,
VPH gratings (high throughput, but more expensive, and there may be ageing
issues).
Detector
With a scale of 0.25 arcsec/pixel, 2k*2k and 4k*4k CCDs would
deliver fields of view of diameter 8.3 and 16.7 arcmin
respectively.
A mosaic of 2 2k*4k CCDs would have a small gap down the middle,
and non-coplanarity could be a problem.
A low level of fringing is required, since many of the observations will be carried out in R and I bands.
Cass/Nasmyth imagers on telescopes elsewhere populate a fairly small region of the instrumental parameter space. The properties of 20 imagers on 2-m - 4-m telescopes worldwide, plus FORS (VLT), GMOS (Gemini) and ELMER (planned, GTC), were compared. Most Cass / Nasmyth imagers offer fields of view <~ 10 arcmin, scale 0.2 - 0.3 arcsec/pixel, PSF <~ 0.6 arcsec (80% encircled energy), wavelength range 3300 - 10000 A, 2k*4k CCDs as detectors (singly, or mosaiced), minimum exposure times 0.5 - 2 sec. Polarimetry is offered on only 4 instruments, all mounted directly at Cass.Several of these instruments also offer spectroscopy. These include CAFOS (Calar Alto 2.2-m), ALFOSC (NOT 2.5-m), EMMI (NTT 3.5-m), MOSCA (Calar Alto 3.5-m), EFOSC-2 (ESO 3.6-m), FORS, GMOS and ELMER. Most offer a selection of long slits (for different seeing conditions), the option of multiple slits, and typically 6 - 15 grisms for spectroscopic resolutions of a few 100 to a few 1000.
ACAM will be unique in offering a wide range of observing options (broad-band and narrow-band imaging, spectroscopy), and flexibility of use, on most nights throughout the year.
The main science requirements driving the optical dseign are:Polarimetry is not required. An ADC is probably not required. The principal trade-offs are:
|