Draft Concepts for the NOAO Extremely Wide-Field
Infrared Imager (NEWFIRM)
NEWFIRM is a wide-field, near infrared imager designed for use at the Cassegrain focus of the Mayall 4-m telescope. The philosophy of this instrument is consistent with the long heritage of wide-field optical and infrared capabilities within NOAO. NOAO's future role will include providing supporting capabilities for science programs on Gemini and other large aperture telescopes. A 1997 Tucson workshop involving 46 scientists from 26 institutions evaluated the requirements for the support capabilites within a number of disciplines. All of the panels identified large-scale surveys as essential to meaningful observing programs on very large aperture telescopes. Moreover, the sensitivity limits required for many of the infrared surveys demand the aperture of the 4-m telescope both to complement wide-field optical surveys and to enable followup programs on Gemini-class telescopes.
The design approach will be to maximize the field of view within the physical constraints imposed by the existing 4-m telescope. While wide field coverage is the primary goal, the design will also evaluate capabilities such as narrowband imaging and low-resolution multislit spectroscopy, which scientifically exploit the 4-m in its own right, rather than as a path to followup on larger apertures.
Design Goals
· 4 meter telescope aperture
· Effective field of view 30 arcmin square
· Pixel scale 0.3 -0.5 arcsec to meet science requirements
· Wavelength coverage 1-2.5 microns
· Maximize optical throughput
· Minimize extraneous background radiation
· Broadband filters (J, H, K, Ks’)
· Narrowband (~ 1%) filter capability over field
· Explore multislit grism spectroscopic capability in K (and possibly H) bands to complement NGOS
Discussion of Issues
Detector
The desired 30 arcmin field of view, coupled with the requirement of a pixel scale better than 0.5 arcsec, determines a 4K X 4K detector format. (A 4K X 4K detector covering a 30 arcmin field yields a pixel scale of 0.44 arcsec). At present, there does not exist a monolithic 4K X 4K IR array suitable for astronomical applications; development of 2K X 2K IR arrays is underway, although they are not yet on the market. Within the relevant timescale for the development of NEWFIRM, the most feasible approach to achieving a 4K X 4K detector focal plane is the use of an array of four 2K X 2K detectors.
The two candidates for detector material are HgCdTe and InSb. Both have been successfully developed into 1K X 1K formats, and a development program for 2K X 2K HgCdTe is underway; several astronomical instruments designed to incorporate these arrays are under development or construction. A brief summary of the detector properties is given below:
|
Detector |
HgCdTe |
InSb |
|
Wavelength |
1 - 2.5 microns |
0.8 - 5.5 microns |
|
Pixel Size |
18 microns |
27 microns (?) |
|
Temperature |
77 -– 85K |
30 –- 35K |
|
Status |
2K X 2K under development |
? |
One may marshal arguments for and against either detector material. InSb covers a much larger wavelength range, and the larger pixels may facilitate the camera design. However, the detector itself must operate at a temperature requiring a cryogenic cooler; a HgCdTe system could, in principle, be cooled with LN2 alone, although operational considerations will probably dictate cryogenic coolers for NEWFIRM regardless of the detector. Thermal radiation control is a greater technical challenge with InSb. The detective quantum efficiency (DQE) of HgCdTe is generally less than that of InSb, although the properties of the 2K X 2K arrays are not certain, and there is concern that the small pixels of the HgCdTe arrays are more susceptible to crosstalk.
Radiation Control
Because the ambient temperature environment generates significant infrared radiation in the sensitivity region of either of the candidate detectors, optical control of extraneous radiation is necessary in any infrared instrument. A desirable goal for any instrument is that the performance be limited by the detector dark current or the sky/telescope background in the lowest background configuration. This goal is generally achieved by the use of foreoptics which reimage the telescope exit pupil at a location where a cold aperture stop can be placed. This effectively restricts the field of view of the instrument to the focal ratio of the telescope.
For the same reason, the filters which define the observational bandpass must also be cooled. They are generally placed very close to the pupil stop for several reasons: it is advantageous to eliminate as much unwanted radiation as possible before entering the camera portion of the instrument; the beam at the pupil stop is often at a minimal diameter, thus keeping the filter size minimized; imperfections in a filter are least likely to map into the image plane.
The FLAMINGOS layout shown below displays the essential optical components of an infrared imager. The foreoptics section, consisting of a field lens/collimator, images the telescope pupil onto the pupil stop. The filters are located close to this location. The camera assembly then reimages the telescope focal field onto the detector.
The length of the foreoptics section is roughly equal to the product of the pupil image diameter and telescope focal ratio, so the choice of pupil size is important in determining the overall physical scale of the instrument. Keeping the pupil small permits the use of smaller (and less expensive) filters, but also places constraints on the ability of the foreoptics to produce a good image of the telescope pupil on the stop. Furthermore, the filter performance will be affected by the size of the pupil diameter.
Filters
Infrared filters are constructed of interference layers deposited on (usually) both sides of a substrate which is transparent in the bandpass region. This geometry is necessary for the filters to withstand the cryogenic vacuum environment of the instrument. The substrate is usually chosen to provide long wavelength out-of-band blocking; in general, filters in the 1- 2.5 micron range are on glass or silica substrates, and those at longer wavelengths utilize Si or Ge. A wide variety of materials is used for the actual interference layers, depending on the wavelength of the filter.
In addition to the standard broadband photometric filters (J, H, K, Ks’), narrowband filters with as little as 1% fractional bandwidth can be fabricated. A recent consortium purchase from OCLI yielded filters of excellent quality 60 mm diameter. Barr Associates claims to be able to produce IR interference filters up to 125 mm diameter. The dichroic beamsplitters in SQIID are effectively interference filters, the largest of which is 125 mm diameter.
Although the instrument beam from any point in the telescope field is
effectively collimated at the pupil image, it passes through the pupil at an
angle = tan-1
([2a + d]/2F), where a is the off-axis radius at the telescope focus, d is the
pupil stop diameter, and F is the field lens -– pupil distance. The effective
bandpass of an interference filter will be effectively shifted to shorter
wavelengths by an amount
= 1 - cos(
/n) for light passing through at an angle
off-axis; n is the effective index of
refraction of the materials in the interference layers. Laboratory tests (see
figure below) suggest that n ~ 2.0 for the materials used for filters in the 1
- 2.5 micron range. An off-axis angle of 10 degrees will shift the effective
wavelength shortward by 0.5% , beyond the permissible amount for a 1%
narrowband filter. Some relief is possible by designing the filter so that its
center wavelength is larger than the nominal value by half the total wavelength
shift from the center to the corner of the field.
The scientific arguments for NEWFIRM identify narrowband imaging as an important capability; furthermore, it is one which other large-format IR imagers under development may find difficult to accommodate.
Optical Design Concepts
All-Transmissive Baseline Concept – FLAMINGOS
FLAMINGOS is a wide-field near-infrared imager and multi-object spectrograph under construction by Richard Elston at the University of Florida. It is designed to cover a 10 arcmin field of view at the Mayall 4-m telescope and use a 2K X 2K HgCdTe array. The conceptual optical layout, designed by Charles Harmer, is shown below:
Simplified Layout of Charles Harmer's FLAMINGOS Design
A straightforward scaling of the FLAMINGOS concept to NEWFIRM would entail significant difficulties. At the f/8 Cassegrain focus of the 4-m a 30 arcmin field is 270 mm on a side, or 380 mm diagonally. The window and first field lens would have required diameters of at least 400 mm. The field lens in FLAMINGOS is BaF2, which would probably be impossible to obtain in this size. The maximal off-axis beam angle through the pupil is 10 degrees, sufficient to affect narrowband imaging at the corners of the field (although the wavelength shift would probably be acceptable at the field edges).
Finally, it is difficult to accommodate a mosaic of detectors in this concept. Unlike the optical CCDs being used in Mosaic imagers, the IR arrays under development are not designed to be edge-buttable, and one certainly cannot count on this capability being developed in the near future.
Quattroflamingos: A 4-Shooter Concept
One approach to covering a very large field is to break up the focal plane, as in the original "4-shooter" CCD camera [Gunn et al 1987, Opt. Eng., 26,779]. A pyramidal mirror splits the telescope focal plane into four subfields, each of which is then imaged with a FLAMINGOS-like instrument. Each of these four arms is still a substantial instrument, but the smaller requirements on field lens size (200 mm vs 400 mm) greatly increase the probability that these optics can even be fabricated.
The use of a large pupil (100 mm diameter) will increase the length of the instrument arms, but maintain a sufficiently small off-axis angle through the pupil to permit narrowband imaging over the entire field. This tradeoff will be considered in the optical design of NEWFIRM.
Unlike the original "4-shooter", which utilized a relatively shallow-angle pyramidal mirror at the back of the instrument, this concept requires a 45 degree pyramid. If this pyramid were placed at the telescope focus, there will be some vignetting of the fields imaged onto the four separate detectors. Furthermore, scattering from the edges of the pyramid or from anywhere within the "glass box" containing the pyramid and the entrance windows to the four arms is a potential source of unwanted radiation.
Since a tightly coupled mosaic is impossible to obtain by either physical butting of the detectors or a pyramidal mirror, an "open mosaic" is required. Obtaining a filled image will then require several observations at different telescope positions for any inter-array spacing up to the size of a detector. This is a standard observing protocol in background-limited IR observations in any case. The most efficient format is a spacing just less than an array dimension, but a practical value is 0.85 -– 0.9, to yield sufficient field overlap for registration of adjacent fields. This concept is being utilized in other wide-field IR imagers such as those on VISTA and UKIRT
A concept for NEWFIRM is an open mosaic of four arrays, each covering 15 arcmin square, separated by 13.5 arcmin. Observations at four telescope positions offset by 14.25 arcmin will cover an area of 0.93 deg2, as illustrated below.
The requirement for an open mosaic permits the pyramid to be located in front of the telescope focus without vignetting any of the four subfields. The focus may be sufficiently far from the pyramid to be enclosed within the cryostat of each of the four arms of the instrument. This will allow a cold focal plane mask for imaging baffling or multiobject spectroscopy. A possible problem with this approach is that moving the beamsplitting pyramid well in front of the telescope focus significantly increases the horizontal extent of the instrument, and additional folding within the instrument may be required. The bar in the figure is 1 meter in length.
4-Shooter Concept with Pyramidal Beamsplitter
The concept shown above and below is NOT an optical design, but a conceptual scaleup of the FLAMINGOS design from a 10 to 15 armin field in which the pupil stop has been increased to 100 mm diameter and the first field lens moved closer to the focal plane. Since the telescope field in this concept is 50% larger than in FLAMINGOS, but the detector is the same size, the actual camera design may be more complex.
If the field lens is sufficiently close to the telescope focus, its diameter may be sufficiently small that the pyramidal mirror is unnecessary, and the four arms of the instrument may be installed on a common optical bench. The unused on-axis field may be used for acquisition or guiding.
4-Shooter Multibarrel Concept
This approach will be seriously investigated, as it makes the instrument
structurally more compact, has less potential for scattered light, and also leaves
the telescope focus volume clean, allowing for the possibility of inserting
slit masks for multiobject spectroscopy. The common optical bench support would
facilitate coalignment of the four arms, including the tilt (approximately 0.7
deg) required by the curvature of the telescope focal plane. The design is
sufficiently compact that the the filter and focal plane mask wheels may
determine the instrument envelope.
The VISTA Approach
VISTA (Visible & Infrared Survey Telescope for Astronomy) is a proposed 4-m telescope for very wide-field optical (1.6 X 1.4 deg) and near-IR (0.25 deg2 within a 1.1 deg field) imaging. The nine IR arrays are arranged in an open mosaic, a conceptual basis for the proposed NEWFIRM geometry. This system uses a large cryogenic Schmidt camera, with the telescope focal surface well in front of the telescope primary. This approach would probably be incompatible with the space envelope at the existing f/8 R-C focus of the KPNO 4-m. The large beam angles passing through the pupil stop are incompatible with the goals of narrowband imaging, and the location of the front corrector plate very near the telescope focus would preclude a spectroscopic capability.
F/6 Optical Design
One concept under consideration for NGOS is the use of a new f/6 forward focus located approximately 0.8 m in front of the primary. This may be driven by the limited physical envelope within the present Cassegrain cage and the need for access to the focal region for interchanging slit masks and other tasks. If the NGOS optical studies suggest that this geometry is required, a modified version of the VISTA Schmidt camera approach should be considered for the NEWFIRM optical study, along with the refractive concepts referred to earlier.