Science Discussion of a New Optical Spectrograph for Kitt Peak National Observatory


Held in the Main Conference Room the afternoon of July 22, 1998.



The goal of the meeting was to reexplore the science drivers for the original High Efficiency Spectrograph Project and to review the status of some new technologies that may benefit the performance of a new spectrograph.


In attendance:

T. Armandroff, S. Barden, G. Jacoby, C. Harmer, D. Harmer, R. Green, S. Strom, C. Claver, D. Joyce, N. Suntzeff, J. Najita, A. Dey, T. Lauer, J. Valenti, B. Bohannan, B. Jannuzi, H. Abt, K. Hinkle


Current Spectroscopic Facilities:

Taft presented an overview of currently available spectroscopic facilities.

KPNO Facilities





RC Spec

















1 degree




History of High Efficiency Spectrograph:

Sam briefly discussed the history of the High Efficiency Spectrograph Project which was initiated in early 1994. It involved a study to see if a dual beam spectrograph could be produced with maximum efficiency at the cost of field of view. The design focused on a new technology grating for the red beam and a concave-holographic grating for the blue channel. Spectral coverage was to be complete from the atmospheric cutoff to the redward limit of CCD sensitivity at a resolving power of about 3000 (2 Ang. at 6000 Ang.). The primary science driver was to observe faint, rare objects. Recent discussion about the viability of the science drivers for the High Efficiency Spectrograph led to a reinvestigation to the science goals of a new optical spectrograph for Kitt Peak National Observatory.


New Grating Technology:

Sam presented a tutorial on the technology of Volume-Phase Holographic Gratings. These are gratings that function via Bragg-diffraction in a volume of material in which the refractive index is modulated. Some of the benefits of this technology over surface-relief gratings (ruled and holographically generated) are:

  1. Potentially higher diffraction efficiencies are possible.
  2. "Blaze" profile is tunable due to the Bragg nature of the gratings rather than fixed as in a surface grating.
  3. Ability to fabricate complex grating structures impossible with surface relief gratings (eg. two gratings in one assembly).
  4. Ability to make "custom" gratings more easily than with a surface grating.
  5. Potential to produce gratings at least as large as 24 inches in size.
  6. Ability to clean the gratings (the grating volume is actually encapsulated between two substrates).
  7. Ability to utilize anti-reflection coatings on the grating surfaces due to the encapsulated nature of the grating.
  8. Our recent scattered light tests also show these gratings to be at least comparable, if not superior, to ruled, surface relief gratings in terms of scattered light. (This was not mentioned at the meeting, but is an item of interest.)

For the current discussion, one can assume that gratings of up to 8-inches in size will be available for a new optical spectrograph. (We are also exploring the possibility of fabricating up to 12-inch gratings.)


Strawman Concepts:

Sam also presented some spectrograph concepts that might be possible with the new grating technologies. A very wide field spectrograph may encounter optical challanges not unlike those faced by the GMOS spectrograph. As a starter, the GMOS optical design may provide a basic idea of how our spectrograph may look like.

The tunable nature of the Volume Phase Gratings can allow a reconfigurable spectrograph, in which the grating can be tilted and the camera angle adjusted, to utilize a single grating element for a variety of dispersions and to optimize the diffraction efficiency for a particular wavelength.


Prospects for Delivered Image Quality (DIQ) at the Mayall and WIYN Telescopes at Kitt Peak:

Chuck gave an overview of what type of image quality we might expect to be in place at both the WIYN and Mayall telescopes within the time frame of a new spectrograph. WIYN currently delivers about 0.8" median images without tip-tilt. The Mayall delivers about 1.1" median at present. The inclusion of an active support system for the primary mirror should improve the seeing by about 0.1 to 0.15", bringing the 4-meter to just slightly worse than the WIYN. Tip-Tilt implementation looks promising for narrow fields, but with only a modest improvement of another 0.1" realistically expected. The field of view with a tip-tilt system will likely be restricted to about 5 arc-minutes in the optical. For comparison, CTIO's Blanco telescope is currently achieving about 0.9" median seeing after many years of image improvements.


Science Cases:

Jeff Valenti presented a case which would drive the resolving power of the desired instrument to greater than 60000 (0.1 Ang at 6000 Ang.). He also gave some science drivers for a more modest resolution spectrograph. A survey of the stellar activity of young stars in clusters requires resolving powers of between 3800 (1 Ang. at 3800 Ang.) to 7600 (0.5 Ang. at 3800 Ang.). Very large fields of view (>40 arc-min) are desired. Spectral coverage would be minimal but centered on the Ca II H and K lines. Hydra can perform observations of the brighter sampling of stars. Low efficiency of Hydra will limit the achievable magnitude of these surveys.


George Jacoby discussed the science of planetary nebulae in the buldge of M31. Spectral coverage is a driver with coverage of the 3727 line considered a requirement out to beyond 7000 Ang. Spectrophotometry over this range in wavelength will require an ADC to compensate for atmospheric dispersion. Desired S/N of at least 20 (per resolution element?) is a minimum. Resolution of better than 7 Ang. with a 1 Ang. goal is needed (R of about 5000 to 6000). FOV requirements are minimal with a goal of 3'. Reaching planetaries in the nearest elliptical galaxies is a scientific goal which requires good efficiency in the spectrograph (about a factor of 2 to 3 times better than current RC spectrograph with the KPC10 grating). GMOS may be capable of doing this specific project.


Chuck Claver presented two science cases. The first involves the study of the white dwarf cooling luminosity function. Limiting magnitudes are V=23. Target density (selected via imaging surveys) is about 50 to 100 objects per square degree of which only 10% are actually white dwarfs. A resolving power of up to 5000 is needed (1 Ang. at 5000 Ang.). Spectral coverage should extend from 4500 to 9000 Ang. Goal is to survey an area of 10 square degrees. None of the currently available facilities could perform this survey in an efficient manner. GMOS has too narrow of a field of view. V=23 limiting magnitude requires excellent efficiency at the required resolving power.

Chuck's second case concerned white dwarf stars in open clusters. Magnitude range is V=18 to 23. Requirement is to fit profiles of H-beta and higher Balmer lines. A resolving power of 1000 is needed (6 Ang. at 6000 Ang.). There are about 10 to 20 white dwarfs per cluster. Excellent image quality and low scattered light is required due to the presence of neighboring bright stars. Maybe GMOS is capable of conducting this program.


Taft Armandroff talked about low surface brightness galaxies (>22 mag.). The aperture of the telescope is less of an issue than spectrograph efficiency, scattered light, and detector performance. A novel approach of charge shuffling on a CCD in synchronization with telescope beam switching allowed a group at MDM to observe an object at V=24.0 with a resolving power of 4000 in 8 hours. Similar techniques with a new optical spectrograph at Kitt Peak can be used to observe dwarf galaxies and cirrus clouds.


Arjun Dey presented a goal to to an optical survey of objects that are emitters at sub-mm wavelengths. There are about 2000 such objects per square degree on the sky, so a 10 arc-minute field is adequate. The optical magnitude of such objects is between V=23-25. A spectral resolving power of 4000 to 5000 is required. Wavelength coverage should be in the red to look for the [OII] line at high redshift. Good efficiency at 1.1 microns is desired.


Tod Lauer gave a discussion on the study of large scale structure evolution at redshifts around 1 or less. Galaxy clusters are 1 to 15 arc-min. in size. A 30 arc-minute field might allow the simultaneous observation of more than 1 cluster at a time. Good red performance at 1 micron is desired. A resolving power of 5000 (2 Ang. at 1 micron) is needed.


Richard Green put forward two scientific cases. The first is a search for low Fe/H objects. These are very rare objects with about 1 in 1000 emission line galaxies falling into this category with an abundance of less than 1/50th solar. Spectral coverage from 3727 redward to about 7500 is required. A resolving power of 2000 (2.5 Ang. at 5000 Ang.) is adequate. Spectrophotometry requires the presence of an ADC unit to compensate for atmospheric dispersion. Target density is about 12 objects in a 40 arc-minute field.

His second example stressed the desire for a high efficiency efficiency with a resolving power of about 5000 (1 Ang. at 5000 Ang.) and spectral coverage down to the atmospheric cutoff around 3000 Ang.


Steve Strom discussed spectroscopy of scattered light from young stellar objects that could yield photospheric spectra of YSOs; current spectra are mostly dominated by disks, so not as useful.



Future Directions:

Volunteers to assist Sam Barden and Taft Armandroff in developing the

initial concept over the next year are: Steve Strom, Nick Suntzeff, and Di Harmer.

We have access to optical designer time in FY99 to start conceptual design efforts. We hope to compare different options of design, location, field of view, and efficiency tradeoffs. A narrow-field design will be pursued to explore the maximum efficiency that one might achieve with an optimized spectrograph. In parallel, a very wide field design will provide us with a estimate in the tradeoffs between field and efficiency. Study of location of the instrument at either the Mayall Cassegrain or WIYN Nasmyth focal planes will be made to evaluate which location may provide the most significant scientific return and/or best operational performance.

[Note that the WIYN option has been rejected on the grounds that such an instrument would be impossible or very difficult to implement due to severe space constraints.]