NG2dF A Next Generation 2degree Field Instrument for the AAT.

1
NG2dF A Next Generation 2-degree Field
Instrument for the AAT. Submitted to AAO Users
Committee, July 2002. Version 1.1
15/7/02 Authors R. Content T. Shanks
(Department of Physics, University of Durham)
Concept presented to Next Generation AAT
instrumentation workshop at AAO 17 May 2006 Will
Saunders
2

• Summary NG2dF is a multi-slit spectrograph
  designed to sit behind the current 2dF corrector,
  replacing the 2dF fibre coupler. It is designed
  to have up to 10000 multislits over most of the
  2-degree field. NG2dF has a 6-9x advantage over
  instruments such as VLT VIMOS and Keck DEIMOS for
  surveys of tens of thousands z3-4 Lyman-break
  galaxies. NG2dF also has a multiplex gain of an
  order of magnitude over the current 2dF
  instrument plus improved throughput and sky
  subtraction in galaxy surveys at B22. These lead
  to a likely overall gain of a factor of 25 in
  redshift acquisition rate of NG2dF over 2dF.
  NG2dF would therefore produce huge potential
  gains in prime science areas and NG2dF is an
  instrument for which the AAT is uniquely suited.
• NG2dF is designed to use relatively cheap
  components such as 2kx2k CCDs. The total cost of
  the 12 spectrographs is expected to be low, only
  of order 1m. One issue still to be investigated
  is whether it would be cheaper to build a new 2dF
  top-end rather than make interchangeable modules
  for the fibres and NG2dF. Other remaining
  uncertainties are whether the weight of NG2dF can
  be kept low enough for AAT prime. Also we have to
  check the cost of mask manufacture given that
  slits will have to be placed to an accuracy of
  0.3arcsec or 20 microns. We request 20k to make
  a preliminary design study and costing of NG2dF.

3

• NG2dF Science Case T.Shanks
• Surveys of z3 LBGs. The first motivation for
  NG2dF arose from considering how a 4-m telescope
  might compete with the VLT VIMOS or Keck DEIMOS
  for surveys of tens of thousands of U drop-out
  galaxies at z3 and B dropout galaxies at z4.
  For example, there is a proposal to observe 35000
  z3 Lyman-break galaxies with Keck DEIMOS. Z3
  Lyman break galaxies at Rlt25 have a sky density
  of 1 arcmin-2 and z4 Lyman break galaxies at
  Ilt24.5 have a sky density of 0.2arcmin-2.
  Therefore some 10000 z3 galaxies plus 2000 z4
  galaxies are observable in the 2df field-of-view.
  U dropout galaxies usually lie in the redshift
  range between 2.5ltzlt3.5 and therefore the
  wavelength range of interest for detecting
  Lyman-alpha emission is only 4250Alt?lt5500 Å,
  arranged by means of the grating dispersion and
  the cut-off filters. Therefore spectra with
  300-400 pixels of width 3-4 Å (at a resolution of
  6-8Å), is all that might be needed to measure
  redshifts. Assuming a row of 4 targets can be
  fitted across a CCD in the dispersion direction,
  this leaves 200 such rows to be fitted into the
  spatial direction of each CCD for a maximum slit
  width of about 10 pixels or 8arcsec (see design
  overview below for details). There is also the
  possibility of beam-switching microslits and
  charge shuffling which might also be implemented
  if thought to be more efficient for sky
  subtraction while maintaining multiplex
  advantage.
• Exposure times at the Keck are 3hrs suggesting
  that the exposure time on the AAT may be 20hrs.
  Although this exposure time is long, the huge,
  36x bigger, field-of-view advantage of NG2dF over
  even the 20? fov of VLT VIMOS means that the
  NG2dF LBG redshift rate would be 6-9 x faster
  than VLT or the Keck. Just 60-80hrs or a week of
  4-m time would be needed to measure redshifts for
  35000 z3 galaxies, compared to 6-9 weeks of
  8-10-m time!

4

• Blt22 Galaxy Redshift Survey. Although surveys of
  z3 galaxies will become highly important in the
  future for cosmology, it turns out that there are
  many other classes of project where NG2dF now
  shows huge gains over the fibre-based approach of
  2dF. Consider the possibilities for making a
  magnitude limited sample of all galaxies to the
  limit Blt22. Low resolution slit spectra here
  might be obtained in a 1-2 hr exposure. Here the
  galaxy sky density is approximately 3000deg-2 and
  so there are again 10000 galaxies in the 2dF fov.
  Keeping the same configuration as before with
  spectra of length 1200Å but supposing that the 12
  NG2dF gratings and/or the cut-off filters could
  be changed (during the day) then 2400Å of
  spectrum, say, between 3800-6200Å could be
  obtained in 2 exposures on separate nights and
  3600 Å of spectrum between 3800-7400Å in 3
  exposures on separate nights. Again, the
  multiplex gain over the current 2dF system is
  fantastic. Even assuming the same exposure times
  at B22 for fibres and slits, the multiplex gain
  is 5000/400 in the 2 exposure case and 3300/400
  in the 3 exposure case or approximately an order
  of magnitude for NG2dF over 2dF. Taking into
  account the likely factor of 2 advantage in
  exposure time given better sky subtraction
  properties of slits, the NG2dF gain over 2dF is
  then a factor of 25.
• At the B22 limit, NG2dF exposure times are
  likely to be lt2 hrs per 1200 Å spectrum to get
  redshifts. Therefore in only 30 nights,
  half-a-million Blt22 galaxies could be observed.
  These are 2.5 magnitudes fainter than the current
  2dFGRS limit and they would be observed in just
  40 of the time needed for the previous survey!!

5
(No Transcript)
6
(No Transcript)
7

• Other Prime Focus Imaging Spectrographs?
• HET, SALT, f/4 f/4.5
• Single camera instruments

8
Filling factor issues

• SDSS cameras are 200mm diameter to get 50x50mm
  detector size
• gt 125x125mm on sky, filling factor 40-50

9

• Instrumental Outline
• 17,000mm2 detector area (130mm x 130mm)
• f/3.3 collimators
• f/1.3 cameras
• 7 x 4Kx4K x 12µm detectors?
• 1" gt 2.3 pixels
• 100mm beam size
• 7 x 125mm x 125mm 7 x 30' x30' FOV
• 2Å/pixel _at_ 500/mm, R1000
• 4250-5500Å 600 pixels
• 5000-8500A1700 pixels

10

• A few random considerations
• VPH vs ruled gratings
• Off-axis angles in dispersion direction
  dramatically lessens efficiency for VPH gratings.
  Max angle here is very large, 9?. Ruled gratings
  are expensive, limited to certain blazes and
  densities, and have lower efficiency.
• Differential refraction effects during
  extended tracking
• Biggest effects are rotation and shear. If
  entire instrument is rotatable, and if each unit
  spectrograph is orientated to have the slits
  radial, then slit mask is relatively tolerant to
  this distortion, and tracking can be maintained
  for several hours.
• Chromatic distortion in PF corrector is also
  radial.

11

• Camera lens spectrograph?
• 19 x 2Kx2K x 13.5µm detectors
• Use 85mm x f/1.2 camera lenses
• 1" gt 2 pixels
• 60mm beam size
• 3Å/pixel _at_ 500/mm, R1000
• 4250-5500Å 500 pixels
• 5000-8500A1500 pixels
• Off-axis vignetting in camera lens?

12

• Lyman Break Galaxies
• 4000/deg2 , 1000 spectra /beam, 7000 total
• V25 S/N2/angstrom in 40 hours!
• B22 galaxies
• 3000/deg2 , 750 spectra /beam, 5000 total
• S/N7/A in 1 hour, few x 106 galaxies in 100
  nights