The Ohio State University College of Mathematical & Physical Sciences Department of Astronomy

Optical Design Review

Brooke Gregory, working with Paul Byard, has made an independent review of the MODS optical design. All of the optical design prescriptions were successfully translated from from Code-V to OSLO. Both design tools agree on the results, producing similar surfaces to well within reasonable fabrication tolerances. Both codes also arrive at the same ends in optimization exercises, and their merit functions agree. Our internal performance metric is based on maps of the predicted D₈₀ (80% encircled energy diameters) as a function of field position and wavelength. Again, both OSLO and Code-V produce the same performance figures. Brooke then did a detailed analysis of the which elements required optimization, and where the main sources of aberrations were.

Only the camera requires optimization in this design. Brooke started with a decentered Classical Schmidt Camera (flat corrector), and then found that he was able to get significant improvement when he allowed OSLO to introduce a meniscus figure on the corrector (as in a Maksutov Schmidt). His design in the end is essentially the same as Byard's decentered Maksutov camera design.

The primary source of aberrations is coma due to field offset in the collimator. Brooke explored a couple of improvement options. One was to minimize chromatic aberration in the camera. By using an achromatic doublet for the corrector with a "perfect" collimator does provide some optimization at the field center in terms of performance with wavelength, but when you consider the full field you get only incremental (few percent) gains. With the real collimator, the same result prevails. The implications of this need to be explored further, but it is clear that chromatic aberrations in the camera corrector are a minor source of aberrations.

Considering an aspheric field flattener also gives only incremental (~3%) improvement, but Brooke did not explore varying the field flattener thickness or spacing relative to the detector. He also worried about chromatism in the singlet field lens, but this turned out to be a negligible effect.

The bottom line is that the collimator is the primary source of aberrations in the system in wide field mode. Brooke suggested three lines of inquiry to be explored in an effort to see if there are practical [and affordable] paths to significant improvement:

1. Can a better reflective collimator design be found? In wide-field mode, coma is clearly the principal limiting factor.

2. Does tilting the elements in the camera lead to any improvement?

3. Consider refractive collimator designs. This is a fundamental change in the MODS design philosophy, but perhaps a doublet design is possible that will work.

Our heartfelt thanks to Brooke for all his work reviewing our design. MODS will be a better instrument for his efforts.

Other Optical Design News

Paul Byard reported briefly on some further explorations based on his iterations with Brooke Gregory, these will be developed further for our next meeting. Paul also reported on recent advancements in enhanced reflection and anti-reflection coatings from the SPIE meeting. In particular, blue-enhanced aluminum coatings being developed by LLNL look very promising as possible reflective coatings for the MODS blue channel (composite Al+Ag coatings, with a buffer layer between the metals and a protective dielectric coating on the Ag). For transmissive optics, progress has been made using solgel coated MgF₂ to create improved anti-reflection coatings that we should keep an eye on.

Mechanical Design

Tom continues to design various of the MODS mechanisms, concentrating on those mechanisms that will impact upon the focal plane space. This week he unveiled his design for the slit mask mechanism (see a picture on the Instrument Views web page).

Slit masks are stored on-instrument in a cassette mechanism that can hold up to 25 slit masks. Masks are deployed into the focal plane with a linear insertion/extraction arm. Each slit mask has a maximum field size (machinable surface) of 230x230mm supported around the outside by a 16mm thick (front-to-back) 18mm pitch metal frame. The depth of the mask frames is sufficient to allow us to introduce 6mm of sag to match the focal surface of the f/15 Gregorian focal plane, as well as to accommodate auxiliary optics, small tilting elements (e.g., a tilted 1-arcminute AO slit mask), etc. Slit masks may consist of custom multi-slit masks machined on-site (the "machine" still TBD), "facility" long-slit masks of various kinds, imaging masks, etc, up to 25 total. The mechanism can change slit masks in about 10-seconds without heroic effort. The "open" position is always available by just retracting the slit mask.

Special mask units may be built that occupy more than 1 adjacent position. For example, an integral field unit might be implemented that occupies 3 or 4 adjacent mask positions. Payload limits are 4.5kg, with ~0.6kg for the frame, and ~1kg for a 230x230mm field lens (6mm thick).

The 25-position mask cassette can be loaded from outside while the spectrograph is on the telescope. Another loading path is accessible when the instrument is on its carrying cart or in the lab.

The entire structure occupies a T-shaped volume, and uses up one of the two remaining quadrants above the focal plane (the others are used by the grating turrets). The final quadrant will probably be used to some extent by the dichroic changer, more on that later.

Bruce raised the question of whether we need a continuously adjustable mechanical slit, or if having a suite of fixed long-slit masks of different widths available. Pat and Rick were tasked with coming up with an answer by next week.

Some other issues raised by Tom:

Shutters

We all agreed we need a shutter for the CCDs. Bruce was tasked with coming up with options and requirements for next time.

Slit Quality

Tom raised the issue of slit quality, namely "how good is good enough". Tom suggests that we test this in the lab. Darren will begin exploring how we might go about this. The goal would be to answer quantitatively how good a slit we need. Among the issues are control of scattered light in the f/15 beam by light glancing off the inside edges of the slit.

Grating Tilt Repeatability

Again, we need to explore this in more detail, but discussion around the table was that the stock spec of 0.1 pixel keeps coming up, and in the end we expect that this is coupled to the flexure control spec. Heroic efforts to achieve tilt repeatability much better than the flexure compensation would be waste of effort.

R. Pogge, 2000 April 17