KOSMOS. Optical Design

Transcription

1 KOSMOS Kitt Peak-Ohio State Multi-Object Spectrograph Optical Design Revision History Version Author Date Description 1.1 Ross Zhelem Initial Draft 1.2 Paul Martini July 20, 2010 Minor Edits, Disperser Designs Page 1

2 Table of Contents KOSMOS Document v Overview Optical Design Optical Layout Optical Prescription Baseline Optical Properties Field of View Pupil aberrations Image Quality Design Optimization Details Imaging Mode Spectral Mode Instrument Throughput Throughput Budget Internal Transmittance AR Coatings Tolerance Analysis Fabrication Tolerances Alignment Tolerances Asphere Manufacturability Parasitic Light Analysis Scattered Light Stray Light Ghost Light Thermal Performance Focus Shift Bonded Interfaces Alignment and Testing Acceptance Testing at Vendor s Site Alignment and Testing at the OSU Lab Alignment and Testing Equipment Disperser Designs Direct Vision Triple Prism VPH Grism Page 2

3 100 Overview KOSMOS Document v1.0 KOSMOS (Kitt Peak Ohio State Multi-Object Spectrograph) is a wide field, multi-slit imaging spectrograph for the 4-m Mayall telescope located on Kitt Peak near Tucson, AZ. The telescope scale is 6.6 arcsec/mm, f-ratio is f/7.9. The telescope will deliver a 100 square arcminute field of view to the KOSMOS instrument and the optical design will collimate and then reimage this input beam onto a 4096x2048 CCD with 15 micron pixels at a plate scale of 0.29 arcseconds per pixel. Page 3

4 200 Optical Design 210 Optical Layout KOSMOS Document v1.0 The KOSMOS optics are based on a similar instrument (OSMOS) built for the MDM 2.4m Hiltner telescope. Both the Hiltner and Mayall telescopes have comparable f/ratios. The choice of similar sized linear field of view allowed use of the as-designed OSMOS collimator. The collimator produces a pupil of the same dimensions in the collimated space for both telescopes. The collimator is an f/7.9 double Gauss design with a 14 degree FOV and contains a total of five lenses, including a doublet. The optical design is shown in Figure 1 and the optical prescription is provided in Table 1. The first collimator element is 73 mm from the telescope focal surface, which provide ample space for the slit wheel. The collimated beam diameter is 54 mm and the pupil is located 68 mm from the vertex of the last collimator element. The total collimated beam space is 170 mm and will be used for dispersers and filters (not shown on drawing). Fig.1 Layout of the KOSMOS optical configuration Optical design with ray traces for field angles up to 5 off axis. Light enters the instrument from the left, where the first surface corresponds to the focal surface of the 4m Mayall telescope. This is the location of the slit wheel in the optical path. The disperser and two filter wheels are in the collimated beam space between the fifth and sixth lenses. The last lens on the right is the Dewar window. The optical prescription is provided in Table 1. The KOSMOS camera was designed to deliver the necessary plate scale and maintain constant image quality over the entire area of detector. The camera is an f/2.7 Petzval design with an 26 degree FOV. The camera contains a total of 10 lenses, including two triplets, a doublet and singlet field flattener. The first camera lens is aspheric with the maximum departure from a best Page 4

5 fit sphere of 170 microns over the clear aperture. The last lens is the window of the standard KPNO Dewar. The optical design produces a final plate scale of 0.29 /pixel on the 4096x2048 CCD with 15 micron pixels currently intended for use with KOSMOS. Multiple bonded elements are used in the camera design to minimize the amount of air-glass surfaces. 220 Optical Prescription Element Lens Surface Radius Thickness Material Diameter Slit Mask Field Lens COL-1 1 Infinity 15 BSL7Y Meniscus COL BSM51Y COL BAL15Y 86 Collimator doublet COL CAF Collimator singlet COL-5 8 Infinity 20 CAF Camera Triplet #1 (aspheric) CAM-1* BSM51Y 114 CAM CAF2 114 CAM BAL35Y Camera Triplet #2 CAM BAL35Y 146 CAM CAF2 146 CAM BSM51Y Camera Doublet CAM BSM51Y 120 CAM CAF Infinity 11 Field Flattener CAM PBM18Y Dewar window CAM Infinity 6.35 SILICA Infinity CCD 25 Infinity 0 61 x 31 Table 1. KOSMOS Optical Prescription. *Even Asphere, coefficient on r 4 : E-8, coefficient on r 6 : E-12 Page 5

6 All dimensions in Table 1 are in mm. The slit mask is located at the focal surface of the 4m Mayall telescope. KOSMOS is fed by light from the telescope (a Ritchey-Chretien design), parameters of which were supplied by NOAO. The complete prescription is included in the ZEMAX file. 230 Baseline Optical Properties Instrument Design Focal Station Focal Surface Scale Slit Mask Collimator Design Collimator Focal Length Pupil Location Collimated Beam Diameter Collimated Beam Length Camera Design Camera Focal Length Imaging Plate Scale (De)magnification CCD Operating Modes Wavelength Range Field of View Design Image Quality (FWHM) Imaging mode: All refractive, one aspheric surface used in camera f/7.9 RC at Mayall 4m telescope, radius of curvature 3330mm 6.6 arcsec/mm 103 mm diameter (section 240), TBD shape/material/long slit etc f/7.9, 14 degree FOV Double Gauss optimized for telescope input 430 mm 68 mm from collimator, 102 mm from camera 54 mm 170 mm f/2.7, 26 degree FOV Petzval with pupil relief and focal plane relief mm 0.29 arcsec/pixel (15 micron pixels) 2.93x reimaging of telescope focal surface 4096x2048 (15 micron pixels) Multi-object and single object spectroscopy, imaging nm 12 arc minute diameter circle, cropped by CCD to 100 square arc minutes (section 240) U-band: 0.4 arcsec (zenith) BVRI: 0.25 arcsec (section 320) Polychromatic: 0.32 arcsec Spectral mode: single object 0.2 arcsec (section 330) multi-object 0.5 arcsec (section 330) Throughput TBD Dispersers VPH grism, Surface Relief Grism, Direct Vision Triple Prism Filters 4 inch square, two 6 position filter wheels tilted by 8 degrees Refocus between filters None required (section 320) 240 Field of View Field of view is limited in the field direction (along a slit) by the edge of the 2K x 4K CCD (Fig.2). To ensure 100 square arc minute field coverage, the maximum field of view will be 5.82 Page 6

7 field direction KOSMOS Document v1.0 in the spectral direction while constrained to 4.95 in field direction. This field of view configuration maps onto the footprint in the telescope focal plane, which is defined by the 103 mm diameter in spectral direction and 90 mm in field direction Kx2K spectral direction Fig square arc minute imaging field of view on 4K x 2K CCD is shown in red 250 Pupil aberrations According to the online optical prescription of the Mayall telescope, the entrance pupil is located on the primary mirror. Collimator reimages the pupil at the paraxial distance 68 mm from the last lens. First order size of the pupil is 54 mm circular. The collimator was originally designed for a smaller aperture telescope. When used with the Mayall telescope at the same linear field of view in the telescope focal plane, pupil aberrations increase for objects over 3 arc minute off axis. This results in field and wavelength dependent pupil shift (Fig. 3). The collimated beam footprints for objects within 3 arc minutes radius overlap on the disperser to ±1mm due to chromatic aberration. Monochromatic aberrations shift the pupil radially on the disperser by an average of 5mm. The spectrum of an off-axis object is produced by a slightly different area of the disperser. Page 7

8 Fig. 3. Pupil shift vs field of view Page 8

9 300 Image Quality 310 Design Optimization Details KOSMOS Document v1.0 The KOSMOS camera was initially designed for perfect collimated input using five object fields uniformly distributed over the CCD area. The camera was then fed by the Mayall telescope and OSMOS collimator. Object fields correspond to different image heights in the telescope focal plane or slit mask. Imaging mode is characterized in the next section 320 by image quality dependence on object field on sky. The camera and collimator were designed for perfect input, are well matched in performance, and deliver excellent image quality over the imaging field of view (Fig. 2). In spectral mode, light from every object is dispersed between the edges of the CCD in the spectral direction. Due to the significance of the spectral mode, KOSMOS was reoptimized for the best spectral performance. This was accomplished by dispersing light into 5 equidistant spectral fields according to Fig Fig. 4. Object fields #1 through to #7 and spectral fields (color coded) for spectral optimization. 320 Imaging Mode Imaging performance is shown at the best focus for V-filter or Stroemgren y filter. Page 9

10 Fig.5. KOSMOS imaging performance in broadband filters Fig. 6. KOSMOS imaging performance in narrowband filters, namely, Stroemgren uvy, Hα, Hβ. 330 Spectral Mode Spectral performance is shown for object and spectral fields defined in section 310 (Fig.4). The goal of spectral optimization is to deliver uniform performance in multi-object mode while maintaining excellent image quality for a spectrum of an on-axis object. The single object case Page 10

11 corresponds to the object field #1. Plots for object fields #1, #4, #5 contain 3 spectral field curves due to obvious symmetry. The table of baseline optical properties (section 230) contains values averaged over fields and wavelengths. Fig. 7. KOSMOS spectral performance Page 11

12 Fig. 7. KOSMOS spectral performance (continued) Page 12

13 Fig. 8. KOSMOS performance with seeing Page 13

14 KOSMOS on sky performance is the convolution of the geometric image quality with atmospheric seeing. Seeing was taken into account via a quadrature sum. The results are shown in Fig. 8 for an on-axis object at 0.9 arcsec seeing, an off-axis object at 0.9 arcsec seeing, and an on-axis object at 0.6 arcsec seeing. The latter case provides evidence for effective use of 2 pixel spectral slit. Simulated performance with atmospheric seeing is uniform over the entire CCD area and wavelength region. Page 14

15 400 Instrument Throughput 410 Throughput Budget The main factors affecting throughput are: KOSMOS Document v1.0 Internal absorption in the glass lenses Reflection losses for bonded surfaces Reflection losses for the AR coatings This budget does not include any losses at the filters since these depend on the particular kind of filter used. Spectral efficiencies are also disperser specific; they should be taken into account separately. Loss Source Notes 1.7% Internal absorption at 500 nm 1.7% Reflectance of bonded surfaces at 500 nm 23.8% Broadband AR (BBAR) coating 72.8% Total Throughput Table 2. Throughput budget 1.5% each, 18 surfaces including dewar window The throughput table indicates that the throughput is sufficiently high to satisfy the 40% requirement for the entire system including disperser and detector. 420 Internal Transmittance Internal transmittance is determined by absorption of light in the bulk of material as it passes through. For bonded elements, it includes Fresnel losses on bonded surfaces. Calcium Fluoride is chosen as a crown material; its internal absorption is small. To maximize throughput in the blue wavelength range, a set of Ohara i-line glasses is used. They ensure high transmittance over the design wavelengths (Fig. 9). Page 15

16 430 AR Coatings Fig. 9 Internal transmittance of KOSMOS optical elements The transmittance of different AR coatings on all KOSMOS lenses is shown in Fig. 10. BBAR performance is modeled as constant over the entire wavelengths range. Single layer coatings and uncoated transmittance are given for reference. Overall throughput of KOSMOS optical system is shown in Fig. 11. It does not include filters, dispersers or CCD quantum efficiency. Design performance of modern BBAR coatings is very efficient (Fig. 12). It confirms that the throughput goal in the Table 2 can be achieved. Fig. 10. Transmittance of AR coatings on KOSMOS optics Page 16

17 Fig. 11. KOSMOS throughput with different types of AR coatings Fig. 12. Theoretical reflectivity of a BBAR coating Page 17

18 500 Tolerance Analysis KOSMOS Document v1.0 KOSMOS optics deliver excellent image quality over the entire field of view and wavelength range (section 300). The purpose of tolerancing is to maintain the image quality by setting limits to fabrication and alignment errors. A sensitivity analysis was run using the RMS image diameter averaged over the field as an image criterion. Camera refocus is allowed as a compensator for axial tolerances, namely, radii, air and glass thicknesses. All spherical surfaces will be fitted to the manufacturer s test plates in order to reduce cost/time of optical fabrication. Test plate fits insure very close matches to the design radii without significant effort. KOSMOS optics are based on high blue transmission i-line glasses from Ohara, Japan. Standard glass data refer to visible wavelengths. Dispersion may be different from catalog values close to the blue or red end of the design wavelength range. It is important to order blanks along with melt index measurements. Upon the delivery of blanks, the design will be reoptimized for actual indices of glasses. This strategy eliminates the need to tolerance on index of refraction. 510 Fabrication Tolerances Element Parameter Nominal Tolerance Unit Field Lens Radius 1 Plano 15 Fringes Radius Fringes Irregularity Fringes Irregularity Fringes Thickness mm Wedge 0 30 arc sec Meniscus Radius Fringes Radius Fringes Irregularity Fringes Irregularity Fringes Thickness mm Wedge 0 20 arc sec Collimator Radius Fringes Doublet Radius Fringes Radius Fringes Irregularity Fringes Irregularity Fringes Irregularity Fringes Page 18

19 Thickness mm Thickness mm Wedge arc sec Wedge arc sec Collimator Radius 1 Plano 5 Fringes Singlet Radius Fringes Irregularity Fringes Irregularity Fringes Thickness mm Wedge 0 20 arc sec Camera Radius 1 Section 530 Triplet #1 Radius Fringes Radius Fringes Radius Fringes Irregularity 1 Section 530 Irregularity Fringes Irregularity Fringes Irregularity Fringes Thickness mm Thickness mm Thickness mm Wedge 1 0 arc sec Wedge 2 0 arc sec Wedge 3 0 arc sec Camera Radius Fringes Triplet #2 Radius Fringes Radius Fringes Radius Fringes Irregularity Fringes Irregularity Fringes Irregularity Fringes Irregularity Fringes Thickness mm Thickness mm Thickness mm Wedge arc sec Wedge arc sec Wedge arc sec Camera Radius Fringes Page 19

20 Doublet Radius Fringes Radius 3 Plano 5 Fringes Irregularity Fringes Irregularity Fringes Irregularity Fringes Thickness mm Thickness mm Wedge arc sec Wedge arc sec Field Radius Fringes Flattener Radius Fringes Irregularity Fringes Irregularity Fringes Thickness mm Wedge 0 30 arc sec 520 Alignment Tolerances Element Axial displacement mm Tilt arcminutes Decenter mm Col. Field Lens Col. Meniscus Col. Doublet Col. Singlet Cam. Triplet # Cam. Triplet # Cam. Doublet Cam. Field Lens Collimator assembly Camera assembly Focus 20 2 Focus 30 2 Fabrication and alignment tolerances are similar for KOSMOS and OSMOS systems. Both designs make use of an identical collimator. The KOSMOS camera includes two triplets of equal alignment sensitivity. Their alignment tolerances are twice as high as those for the only triplet of the OSMOS camera. KOSMOS asphere is sensitive to decenter to within 0.03 mm that is commensurable with the allowable limit for triplet #1. Tolerances down to 5 µm are achievable with modern manufacturing techniques. Page 20

21 530 Asphere Manufacturability KOSMOS Document v1.0 The first surface of KOSMOS camera is an even asphere. It corrects the 3 rd and 5 th order spherical aberration in the system. The aspheric profile is shown in Fig. 13 with respect to the best fit sphere (radius of mm). The best fit sphere is tangent at the vertex of an aspheric surface. It is optimized for minimum volume removal over the clear aperture of the surface. Maximum departure from best fit sphere is mm at the radial height 40mm. Fig. 13. Profile of the even asphere in the KOSMOS camera Asphericity of the first camera lens is within manufacturing capabilities of modern metrology and fabrication equipment that is designed to generate aspheric departures up to 200 micron on parts up to 200 mm in size. The asphere will be generated with the help of a small area tool. Upon finishing the large scale profile, the surface will be covered by residual texture of the size of the tool. This is similar to polishing large astronomical mirrors with undersized laps. The technique of tolerancing the aspheric surface by matching to atmospheric seeing is developed via a structure function approach. The KOSMOS camera asphere is toleranced to cause the degradation of image quality by no more than 0.07 arcsec FWHM. OSU has experience in procuring highly aspheric optics from the MODS project. It is important to read and analyze interferograms from manufacturer to evaluate surface quality and verify the structure function specification. The software package FRINGESOFT is available for $2000. It is required to provide efficient vendor support in regard to the fabrication of KOSMOS camera aspheric lens. Page 21

22 600 Parasitic Light Analysis 610 Scattered Light KOSMOS Document v1.0 Parasitic light in an optical system can be generated by multiple effects. It can be scattered by distributed or local features such as glass defects, dust particles, surfaces of mechanical structures, etc. The materials chosen for the KOSMOS optics are highly homogeneous across the blank with regard to index of refraction, air inclusions and striae. Lenses will be fabricated to high scratch and dig tolerances and mounted in enclosed barrels of cylindrical shape. Scattering in the instrument is negligibly small as compared to the telescope. 620 Stray Light KOSMOS will be mounted at the Cassegrain focus of the Mayall telescope. At this location, the focal surface of the instrument may be directly illuminated by the sky. Baffles are mounted on the primary and secondary mirrors of the telescope to block stray light. Their size and relative separation are shown in Fig. 14. Based on the information about the baffles, the focal plane of the telescope is shielded from stray light up to 15 arcminute diameter. The KOSMOS field of view is under 12 arcminutes, therefore, the instrument is protected from stray light. 630 Ghost Light Fig.14. Baffling of the Mayall telescope. Ghost light is caused by undesirable reflection from specular or nearly specular surfaces. In a refractive system these are coated air-glass surfaces. The KOSMOS optical train was analyzed Page 22

23 for the effect of double bounce reflections redirecting light toward the detector. Every pair of surfaces generating a ghost reflection is identified by surface numbers in Section 220. A constant reflectivity of 1.5% for each surface was assumed at all wavelengths. Reflection from the CCD as the first surface is also considered. The effect of reflections can be calculated with the following formula (MacFarlane, M.J. and E.W. Dunham, Optical design of the Discovery Channel Telescope, Ground-Based Telescopes, J.M. Oschmann, Jr., ed., SPIE, 5489, 796, 2004): 2 S I G R1R2, 2 2 (1 ) G where R 1 and R 2 are the reflectivities of the two surfaces, S is the image diameter, is the relative obscuration of the secondary and G is the ghost image diameter. Pupil ghosts are produced by double reflections with focus point which is far from CCD. The intensity is evaluated based on the contribution of 100 objects of equal brightness uniformly distributed in the field of view in imaging mode. The strongest ghost pupil was created by the surfaces 22 and 1 with peak relative intensity 0.14x10-6 (Fig ). In the KOSMOS optical train there are over 144 pairs of air-glass surfaces that generate defocused images on the CCD. Overlap of all these double reflections results in the formation of a ghost pupil (Fig. 17) with roll off edges and central peak intensity 1.3x10-6 with respect to the intensity of the parent images. It is below the requirement of 10-4 for allowed ghost intensity. Fig. 15. Pupil from surfaces 22 and 1 on central 2Kx2K area of CCD. Page 23

24 Fig, 16. Radial profile of the pupil shown in Fig. 15. Fig. 17. Pupil ghost in spectral direction. Page 24

25 Field ghosts are focused close to the detector surface. They can produce spurious images of objects or spectral lines. KOSMOS filters are located in the collimated space next to the camera. The double bounce path CCD-filter requires special attention because it yields a nearly perfectly focused ghost. Reflection off the CCD surface is recollimated by the camera, then reflected by the flat surfaces of the filter and reimaged by the camera again. Filters are designed with a tilt of 8 degrees to the optical axis in order to remove ghost images from the detector. The tilt is implemented in the field direction (Fig. 2). Double reflection from the ¼ inch fused silica dewar window generates the defocused image 3.2 mm across. There are two field ghosts of smaller size listed in Table 3. Surfaces of double reflection On axis diameter, mm Max. relative brightness x x10-5 Table 3. Characteristics of the strongest field ghosts Maximum relative brightness is specified for an off-axis location where ghosts are imaged with large aberrations. The cross section of all identified pupil and field ghost images is larger than 6 arcsec and relative brightness is less than the requirement of Field ghosts may be registered on a frame with oversaturated objects or spectral lines. They will be recognized due to their oversized dimensions. The characteristics of ghosts depend on the actual radii of an optical system and their separation. Test plate fits of lens radii will change the optical prescription to a certain extent. Ghost light analysis will be repeated to ensure comparable performance. Page 25

26 700 Thermal Performance 710 Focus Shift KOSMOS Document v1.0 KOSMOS operational temperature range is from -10 to +30 C. In a lens system of multiple elements, optical properties vary with temperature for two main reasons, namely, i) each individual element s index of refraction depends on temperature of the medium in which the lens is immersed ii) both lenses and mount undergo thermal expansion/contraction. To understand the effects of temperature variation, a soak condition thermal analysis of the KOSMOS optics was performed using the rubber tube model in the optical design. Results indicate that the camera requires a small motion with respect to the CCD dewar to compensate for both index of refraction and isotropic change of dimensions. Once the lens is refocused, the original performance is restored because the effects of temperature and pressure perturbations result in defocus as the primary optical aberration. The rate of thermal focus drift is 12.5 micron per 1 degree C or 0.5mm over the operational temperature range. According to Fig. 18 the geometric depth of focus across the field of view is ±0.1 mm. Based on the depth of focus, KOSMOS image quality will be stable over 8 degrees C of temperature change, therefore it can be focused once at the beginning of nighttime observations. However, seasonal variations of focus may be required. Fig. 18. Focus curves for five field positions The thermal performance will depend on the details of the mechanical mounts. Once they are finalized by the vendor, a more accurate model will be created for image quality variations with temperature or pressure. Page 26

27 710 Bonded Interfaces KOSMOS Document v1.0 KOSMOS optics include two doublets and two triplets. Lenses will be cemented with the help of Sylgard 184. This is a silicon encapsulant for electronics components with good transmission and elastic properties. Optical bonds join elements up to 145 mm in diameter made of glasses with different coefficients of thermal expansion (CTE). During assembly of the SOAR Imager triplet there was an issue with bonding the fused silica lens to calcium fluoride. At the same time, the bond between calcium fluoride and BAK2 was trivial due to both reduced differential expansion and eventually the low adhesion of Sylgard 184 to fused silica. KOSMOS lenses are smaller and made of glasses with similar CTE to that of BAK2 glass. Table 4 contains comparison of the main characteristics affecting the stability of an optical bond. OSMOS elements were joined by a layer 0.1 mm thick. KOSMOS materials and parameters are similar. The behavior of a cemented interface depends on details of the process; the strength of a bond may be reduced by surface contamination. It is important to thermally cycle all KOSMOS doublets and triplets within survival temperature range -20 to +50 C. Instrument SOAR Imager OSMOS KOSMOS Bonded materials Diameter mm Bond thickness mm Δ CTE, 10-6 Silica/CaF CaF2/BAK BAL15Y/CaF BSM51Y/CaF BAL15Y/CaF BSM51Y/CaF BAL35Y/CaF Table 4. Parameters of cemented interfaces. Survival temperature range, C -5 to to to +50 Page 27

28 800 Alignment and Testing 810 Acceptance Testing at Vendor s Site KOSMOS Document v1.0 The main acceptance test will be a test of the entire system assembly that simulates real operating conditions. This test will include a light source, narrow band filters, the optical assembly, and a test CCD. The light source should have a broad band spectrum to produce measurable input over the entire wavelength range. A pinhole source ( 5 microns) is suitable for testing the imaging performance of the entire system assembly. It would be translated along a curved path representative of the telescope focal surface to test the assembly off axis and would require a precision XYZ stage with an XY travel of 100 mm and a Z travel (optical axis) of 1mm. A stop aperture with 29 mm would be placed 68 mm from the collimator singlet at the location of the collimated pupil. The test CCD should be able to record the fine structure of the light source image produced by the optics. A 10 micron (or less) pixel size is required to reliably measure the image quality. The test CCD will therefore need to be mounted on a precision XY stage to sample the image quality across imaging field of view. 820 Alignment and Testing at the OSU Lab KOSMOS lenses will be integrated by a vendor into collimator and camera assemblies according to high tolerances (Section 520). Mutual alignment of collimator and camera is less demanding. It will be verified that the slit mask, collimator, disperser wheel cell, camera, and CCD mount plate share the same opto-mechanical axis. The collimator will be focused by means of autoreflection from a mirror mounted in the disperser wheel. The camera will be focused on-axis by reimaging a sieve mask onto the CCD. In order to accelerate focusing and detector alignment it would be useful for the NOAO data taking system to implement a focus series frame taking sequential exposures and placing them on a single frame with charge shift pixels per exposure. Camera is refocused between exposures that results in a focus sequence of multiple objects in the same fits file. Dewar will be aligned in tip/tilt using a sieve mask. It consists of 5 micron pinholes attached to spherical surface following the telescope curved focal plane. Pinholes are uniformly distributed over 103 mm area. The focus sequence will be run with the z filter for the entire sieve mask and the tip/tilt of CCD dewar will be determined. The correction is made with the help of a spacer machined to acquire the necessary compensating wedge. Throughput of the system will be measured in monochromatic filters for various field positions. Chromatic focal shift will be measured to confirm dependence of focus upon wavelength. An illuminator with variable filters will be used to take focus sequences at different wavelengths. Page 28

29 830 Alignment and Testing Equipment KOSMOS Document v1.0 The cost of alignment and testing equipment, the fabrication of test fixtures, and software have been included in the budget and project plan. Page 29

30 900 Disperser Designs KOSMOS Document v1.0 KOSMOS is designed as an imaging spectrograph with dispersers operating in transmission. Given that the optical axes of the collimator and camera are coincident and the camera allows only focus adjustment, the common requirement is to restore the path of the chief ray upon exiting a disperser. The Triple Prism and VPH Grism satisfy this condition. 910 Direct Vision Triple Prism The Triple Prism is shown in Fig. 19 along with filters in the collimated space between the last element of the collimator and the first element of camera. The middle prism is made of Ohara glass PBL2Y. It is surrounded by identical prisms made of S-FPL51Y from the same manufacturer. Cemented surfaces provide dispersion while external facets steer the beam into the camera. Fig. 19 Triple Prism The resolving power of the prism increases if higher index materials are used with larger difference in dispersive power. In both cases the prism angle and size increases as well. The current prism design was used for OSMOS. It can be implemented in KOSMOS due to the fact that collimated beam space is retained. Spectral resolution of the prism is evaluated in Fig. 20 for an on axis object based on a 3 pixels slit. Page 30

31 Fig. 20 Triple Prism resolution 920 VPH Grism VPH Grism is shown in Fig. 21 along with filters in the collimated space between the last element of the collimator and the first element of camera. A VPH grating is cemented between two right angle prisms. The front facet provides the incident angle for the grating according to the Littrow condition. The back surface steers the dispersed beam into the camera. Fig. 21. VPH Grism The resolving power is determined by the grating line frequency. Higher dispersion requires steeper facets or higher index material to satisfy the Littrow condition. The grism shown in Fig. Page 31

32 21, corresponds to the limiting case of an equilateral assembly (60º apex angle) due to packaging. The prism material is BK7 and the grating constant is 1185 lines/mm. The spectral resolution of the grism is R=1600 for an on-axis object based on 1 arcsec slit. The spectral coverage in single object/long slit mode is nm. Absolute peak efficiency is 95% at 450 nm (Fig. 22) and ~60% close to the CCD edge. There is a slight variation of grating efficiency for off-axis objects that should be taken into account when considering the multi-object mode. With spectral coverage of 260 nm per grating, the entire wavelength range nm will require the total of three VPH grisms of similar design. Fig. 22. VPH Grism efficiency Page 32