Optical Design for an Infrared Multi-Object Spectrometer

1. Optical Design for an Infrared Multi-Object Spectrometer R. Winsor, J.W. MacKenty, M. Stiavelli Space Telescope Science Institute M. Greenhouse, E. Mentzell, R. Ohl NASA – Goddard Space Flight Center R. Green National Optical Astronomy Observatories

2. Multi-Object Spectrometers • Punch Plate • Take image • Make punch plate based on image • Install punch plate • Take observation data • Does not work well for spacecraft • Robotically positioned optical fibers • Integral field

3. Multi-Object Spectrometers • Punch Plate • Robotically positioned optical fibers • Mechanically complex and expensive • Limits ability to get spectra on neighboring objects simultaneously • Difficult to apply to an Infrared instrument • Cryogenic vacuum environment • Integral field

4. Multi-Object Spectrometers • Punch Plate • Robotically positioned optical fibers • Integral field • Small field of view • Maps fiber bundle to a vertical arrangement for imaging

5. Micro-Mirror Array (MMA)

6. Micro-Mirror Array (MMA) Using a Texas Instrument’s Digital Micromirror Device (DMD) for IRMOS • Mirrors are individually addressable into one of two tilt configurations (on or off) • Slit lists can be generated and implemented quickly • Flexibility in geometry of slit (good for galaxies) • 16mm square mirrors, 17mm mirror spacing • Allows for input focal ratios as fast as f/3.0 • 848 x 600 Mirror array

7. Micro-Mirror Array (MMA) Design Challenges: • Tilted Focal Plane • Clocked at 45 degrees • Discontinuous surface • Interference effects? • Wavefront error from spillover • Requires a User-Defined Surface for modeling in optical design software

8. Optical Design – Stage One • Designed for the Kitt Peak Mayall Telescope (3.8m) • Convert F/15 from telescope to f/4.6 • Plate scale = 0.2 arcsec/pixel • Seeing is typically 0.8”, and can be as good as 0.6” • Create tilted focal plane • Angle of incidence ~10 degrees at MMA • Spot sizes: FWHM better than 0.6” • Entirely Reflective

9. Merit Function • Start with axial design • Optimize for RMS spot radius • “Unfold” by adding angle of incidence operands • Require a minimum angle of incidence rather than an exact angle • RAID is exact angle of incidence • Use OPGT to set the minimum • Allows configurations that might not be expected to work well

10. Optical Design - Spectrometer • Resolutions (Dl/ l) of 300, 1000, and 3000 in the J (1.1mm), H (1.6mm) and K (2.2mm) bands • Gratings have 50mm diameter active area • Spot sizes better than 0.6” FWHM • F/5.0 beam to detector • Rockwell HAWAII-I detector, 18.5mm pixels • Maintains 1:1 mapping from MMA to Detector • Compact size • Entirely Reflective

11. Merit Function • Multiple Configuration • 5 different grating groove densities were used • 0, 36, 150, 333, 600 • Grating without grooves is a mirror for imaging purposes • Coordinates of Optics downstream of the gratings had to be fixed • Different gratings require different substrate tilts • Zemax does not have built in solves for coordinate break tilts to deal with different grating configurations

12. Merit Function • No operands were used to encourage a pupil or a collimated beam at the grating • Multiple configurations were made to guarantee good performance across all grating configurations • Especially important due to different grating tilts • Not clear that use of such operands would be a better strategy • Is time saved? Only if a merit function with fewer configurations can be developed. • Will solution work after entering new grating information?

13. Merit Function • Angle of incidence at detector was allowed to be variable, but not exceed 30 degrees • RMS spot radius optimized

14. Implementing Merit Function • Start with axial design • Optimize for good spot sizes • Modify Merit function to slowly “Pull” apart the design • So that light paths are realistic • Increase angles of incidence and re-optimize • Repeat until a real solution is found

15. Add Folds • Fold the design into a size that can be packaged into a “compact” dewar • Performed after optimization of each stage • More time consuming to optimize with the other optics • Global coordinates must remain constant during multiple configurations

16. Correcting for Astigmatism • Traditionally, Toroidal surfaces are used. • Relatively easy to fabricate • Different radii of curvature (different power) in the “x” and “y” directions • Use conic values with toroidal shape to correct higher order wavefront error. • In “x” or “y” direction, but not both • Still straightforward to fabricate • Biconic • Different radii and conic values in both “x” and “y” directions

17. Biconic Mirror • Allows compact design • Difficult to Fabricate • Only a handful of vendors are capable of making such a surface • Requires a minimum of a 4-axis diamond machine • Process referred to as “Diamond Machining”

18. Testing • Profilometry • Contact Method (discontinuous mapping) • Several hundred contact points • Relatively Low cost • Adequate? Perhaps. • Computer Generated Holograms (CGH) • Interferometric – continuous surface mapping • Relatively high cost • Complex setup • How are the CGH’s tested? • More than adequate

19. Front End Optics – Side View

20. Front End Optics – Top View

21. Spectrometer Optics – Side View

22. Spectrometer Optics – Top View

23. IRMOS Optics – Side View

24. IRMOS Optics – Top View

26. Acknowledgements Moore’s Nanotechnology division Werner Preub, and the Labor Fur Mikrozerspanung (Institute for Micromachining), University of Bremen, Bremen, Germany Focus Software (Zemax-EE) Texas Instruments (DMD)