1. 13,14 July 2005 Feasibility/Concept Study Mid Term Status Review CCAT Instrumentation Gordon Stacey and the Instrumentation Subgroup

2. Outline Where we are: the current state-of-the-art Instrument Requirements Need to make compromising decisions that deliver science most efficiently Need to make compromising decisions that deliver science most efficiently Baseline Instruments – first light Submillimeter wave camera Submillimeter wave camera Near millimeter wave camera Near millimeter wave camera Extragalactic Spectrometer Extragalactic Spectrometer Second light and future instrumentation

3. The Present At present, there are a few 10 to 15 m class telescopes of very good surface quality (15 to 25 um rms) in very good submillimeter sites: CSO APEX JCMT HHT 10.4 m CSO: Mauna Kea 10.4 m CSO: Mauna Kea 15 m JCMT: Mauna Kea 15 m JCMT: Mauna Kea 10 m HHT: Mount Graham 10 m HHT: Mount Graham 12 m APEX: Chajnantor 12 m APEX: Chajnantor

4. The Present These telescopes have delivered high sensitivity and ground breaking science with relatively modest arrays CSO: SHARC – 40 pixels CSO: SHARC – 40 pixels JCMT: SCUBA – 128 pixels JCMT: SCUBA – 128 pixels New larger format arrays promise exciting new science SHARC II – 384 pixels – now in use! SHARC II – 384 pixels – now in use! SCUBA II – 5000 (  2) pixels – very soon! SCUBA II – 5000 (  2) pixels – very soon! SHARC-2 CSO GSFC SCUBA-2 JCMT NIST 350 um SHARC-2 CSO SCUBA JCMT Johnstone & Bally

5. The Future We plan to build a very high quality (12  m surface) 25 m class telescope at the best known mid-latitude site: the high peaks above the Atacama plain in Chile Our baseline instruments will have at least 6 times as many pixels as the best near future instruments The combination of better site and larger dish should deliver ~ 10 to 40 times better sensitivity in the short submm bands Combination of sensitivity gain plus array size results in factors of thousands gains in mapping speed

6. CCAT Instrument Requirements Primary science Exploration of the Kuiper Belt Exploration of the Kuiper Belt Star and planetary system formation Star and planetary system formation Survey of distant star forming galaxies Survey of distant star forming galaxies Sunyaev-Zeldovich Effect Sunyaev-Zeldovich Effect The first three require submm (200  850  m) imaging 5’ x 5’ FOV 5’ x 5’ FOV Operation in the 200, 350, 450, and 620  m bands Operation in the 200, 350, 450, and 620  m bands The last requires mm (850  m  2.0 mm) imaging The science of the survey is greatly improved with a submm spectrometer optimized for extragalactic line (searches) Starformation studies enhanced by very high R spectrometers

7. Submm Camera Decision Tree – Field of View The telescope delivers a 20’ FOV – why are we designing to a 5’ FOV? Science: The initial science can be delivered with 5’ FOV cameras Science: The initial science can be delivered with 5’ FOV cameras Image Scale: The telescope delivers a 1.17 meter image for a 20’ FOV – this is quite challenging to couple into a background limited camera Image Scale: The telescope delivers a 1.17 meter image for a 20’ FOV – this is quite challenging to couple into a background limited camera Technology: Current, and near future technology suggests 30,000 pixels is a reasonable goal for the array – this can deliver Nyquist sampled images over a 5’ x 5’ FOV at 350  m Technology: Current, and near future technology suggests 30,000 pixels is a reasonable goal for the array – this can deliver Nyquist sampled images over a 5’ x 5’ FOV at 350  m – tiling a 20’ FOV requires 500,000 pixels at 350 um… – tiling a 20’ FOV requires 500,000 pixels at 350 um…

8. Decision Tree: Dichroic Operation? Why not build a dichroic instrument that simultaneously images in two bands, e.g. 350 and 850  m in a single cryostat?  Excellent spatial registration – enables SED science However: Sensitivities, and SEDs are not well matched – the confusion limit is reached 3 times faster at 850  m than at 350  m Sensitivities, and SEDs are not well matched – the confusion limit is reached 3 times faster at 850  m than at 350  m Technology: An optically coupled (SCUBA-2) array is best in the submm, while antenna coupled arrays have better promise at the longer wavelengths Technology: An optically coupled (SCUBA-2) array is best in the submm, while antenna coupled arrays have better promise at the longer wavelengths Fore-optics: Lenses or mirrors? Fore-optics: Lenses or mirrors?  Lenses deliver the image quality and sensitivity for the submm camera, but have unacceptable emissivity for the mm camera  Mirrors achieve adequate image quality over large FOV for the mm camera, with very low emissivity Costs: the arrays are the largest single capital item for an instrument. Folded into the different array technologies, it is logical to construct separate instruments Costs: the arrays are the largest single capital item for an instrument. Folded into the different array technologies, it is logical to construct separate instruments

9. Submm Camera Design First light instrument FOV is 5’ x 5’ FOV is 5’ x 5’  For Nyquist sampling at 350  m this requires a 170  170 pixel array  30,000 pixels, or 6 times that of SCUBA-2 Primary bands are Primary bands are  200, 350, 450  m and 620  m  Driven by similar backgrounds and adequate sampling requirements  Filter wheel to change wavelengths Future instrument will take advantage of the entire FOV

10. Fore-optics Investigated both mirror and lens designs Mirror design maximizes through-put Mirror design maximizes through-put Aberrations kept under control Aberrations kept under control However to obtain a 20’ FOV… However to obtain a 20’ FOV…  Mirror design requires 4 m class off-axis paraboloids  Dewar would likely 8 m  3 m in size For 5” FOV, the design is more modest For 5” FOV, the design is more modest  3 m class off-axis paraboloids  Dewar could be more modest 3 m  1.5 m in size

11. Transmissive Optics System is much more compact The instrument is ~ 0.7 x 1.0 m in size, with a 25 cm dewar window The instrument is ~ 0.7 x 1.0 m in size, with a 25 cm dewar window However, selection of lens material is problematic – bulk absorption hurts both with transmission, and emission However, selection of lens material is problematic – bulk absorption hurts both with transmission, and emission Found a variety of materials that will work (e.g. PE, Quartz, Sapphire, Silicon, Germanium) Selection criterion was essentially the extinction coefficient Other important features Material properties – environmental (H 2 O), structural (window) Material properties – environmental (H 2 O), structural (window) Cost and availability Cost and availability AR coatable? AR coatable? Current design based on Germanium lenses with t > 90%

12. Germanium Lens Design: 5’ FOV  f/8 reimaged to f/4.6 to Nyquist sample 1 mm square pixels at 350  m  This system requires a 170  170 (30,000) pixel array  The change in f/# means the second lens is smaller than the first – can use this as the dewar window Instrument Envelope 2.8 m Lens diameter ~ 44 cm

13. Spot diagram is good – circle is /d at 350  m Image plane is curved so can do better with curved focal plane Can do significantly better if I let lenses grow (modestly) in diameter Germanium Lens Design: 5’ FOV

14. 4 position filter wheel (e.g. 200, 350, 450, 620) The Dewar 21 cm window Array 5’  5’ FOV: 30,000 pixels Cryocoolers Heat Reflecting filters 1.0 m Lyot stop diameter: 12 cm

15. Array Baseline is extension of SCUBA-2 array from NIST 4 x (32x40) pixel subarrays to make 5120 pixels – extend to 30,000. Other similar technologies JPL/Caltech group manufactures sensitive “spider- web” arrays JPL/Caltech group manufactures sensitive “spider- web” arrays CCAT members also have great experience with arrays from GSFC (e.g. SHARC-2) CCAT members also have great experience with arrays from GSFC (e.g. SHARC-2) These arrays will easily deliver the requisite sensitivity with milli- Kelvin cold heads

16. Submm Camera Budget 6 year build expected Detailed design begun in phase II Total budget ~ 10 M$ Dominant items in the budget are: Salaries (6.7 M$) Salaries (6.7 M$) Arrays (~ 3 M$) Arrays (~ 3 M$)

17. Submm Camera Budget: Capital Equipment ItemCostComments Array ~ 3 M$ Estimate for purchase of SCUBA 2 – like arrays from NIST Estimate for purchase of SCUBA 2 – like arrays from NIST Possible in-house development– leveraging off of future space projects (e.g. SAFIR) Possible in-house development– leveraging off of future space projects (e.g. SAFIR) Array Electronics 0.5 M$ In house manufacture… expertise exists within the consortium Cryostat 250 K Based on similar sized cryostats Filters/Optics 250 K Cardiff (Peter Ade’s group) Fore-optics 250 K Pulse tube cooler 37 K Cryomech baselined Low-T cooler 50 K Chase Research baselined

18. Photometer Budget: Personnel PI/Co-PI Electrical Engineer Lead Engineer Software Engineer 2 Research Associates Administrative Support 2 Mechanical Engineers Graduate Students Summer Students Total personnel 800 K$/year + 33% fringe benefits Travel 30 K/year (on average) Miscellaneous 15 K/year Shipping 30 K totals 6 year build: 6.7 M$

19. Millimetere Wave Camera The millimeter wave camera will be a separate development from the submm camera Foreoptics will be mirror system Backgrounds are much lower so that even small emissivity of Germanium lenses is not sufficent – need off-axis ellipsoidal mirrors Backgrounds are much lower so that even small emissivity of Germanium lenses is not sufficent – need off-axis ellipsoidal mirrors Beam is much larger, so the relatively poor PSDF delivered by the off-axis mirror design is sufficient Beam is much larger, so the relatively poor PSDF delivered by the off-axis mirror design is sufficient It is possible to populate a larger FOV with the same number of pixels at the longer wavelengths. Lenses required to image this field become unaffordably large. It is possible to populate a larger FOV with the same number of pixels at the longer wavelengths. Lenses required to image this field become unaffordably large.

20. Millimeter Wave Camera Bands will likely be 850  m, 1.1, 1.4, and 2.0 mm Bands will likely be 850  m, 1.1, 1.4, and 2.0 mm Designed to match performance (sensitivity) with best astrophysical bands Designed to match performance (sensitivity) with best astrophysical bands  Dusty galaxies  Sunyaev-Zeldovich Effect Detectors will be antenna coupled arrays that promise better performance than optical coupled bolometers These detectors being developed at Caltech and JPL A 64  64 pixel array fully populates 20’ FOV at 200  m Costs nearly identical to those of the submm camera

21. Direct Detection Spectrometers Wide-field imaging spectroscopy is a unique niche of single- dish telescopes  Bolometer array spectrometers cover N times more sky than ALMA can CO J=3->2 M. Dumke et al. 2001 CO J=7->6 Bradford et al. 2003 16x16 array on CCAT ALMA Primary beam at 810 GHz Herschel beam at 810 GHz Example: nearby galaxies -- ALMA resolves out emission, while Herschel under resolves it

22. Direct Detection Spectrometers Broad-band spectroscopy is also a unique niche of single-dish telescopes  Bolometer array spectrometers cover N times more bandwidth than ALMA can Suite of far-IR through mm-wave lines provide physical conditions and redshifts in dusty high-z populations -- broad bandwidth is required. CCAT spectrographs will be sensitive to z~5 ULIRGS. ISO LWS, Fischer et al 1999 Z-Spec first light at the CSO June 2005, Caltech/JPL Bradford, Bock, Glenn, Zmuidzinas et al.

23. Strawman Spectrometers: Long- Slit Echelle 30 cm 40 cm 35 cm 816 micron pitch grating Layout from ZEUS -- Cornell – Hailey- Dunsheath, Stacey, Nikola et al. 128 pixels in spectral dimension 16 pixels in spatial dimension Collimator must be oversized by 12 cm to accommodate angles off the grating!  30 cm diameter collimator  grating 30 cm by 40 cm, to accommodate spatial throughput Grating and collimator ~ 1 meter in all dimensions -- Large but doable

24. 8 x 20 cm = 160 cm collimator focus Full field at f/8 -> 20 cm window Collimated beam + overheads: 25 cm dia (and etalon must be near pupil) Etalon spacing: few cm Faster final focal ratio (2-3) to accommodate large array Array up to 10 cm Layout from SPIFI -- Bradford, Stacey, Nikola et al. 2001 Strawman Spectrometers: Imaging Fabry-Perot F-P field is limited by beam divergence For a 20 cm collimated beam and a 3.5 arcmin field this means: R max =1850 at 450 um R max =990 at 850 um CCAT strawman F-P cryostat is 2 meters in longest dimension -- Large but doable.

25. 13,14 July 2005 Feasibility/Concept Study Mid Term Status Review Polarizing grid Dichroics (soft edges OK) BLISS-SPICA WaFIRS suite for R~1000 at 40-600 microns. All fits within 100 x 100 x 50 cm envelope. Caltech/JPL Bradford, Bock, Glenn,Zmuidzinas, et al. Single point source spectrograph Broad instantaneous bandwidth: 1:1.65 2-D medium -- compact and stackable Z-Spec R~200-400 at 1-1.6 mm Multiple modules can multiplex in position or waveband. So could cover multiple CCAT atmospheric windows simultaneously. 40cm WaFIRS Grating Modules

26. Cryo-coolers We are base-lining closed cycle refrigerators for all CCAT instrumentation Pulse tube coolers cool down instrument to 4.2 K Closed cycle 4 He system cools detector package to 2 K Closed cycle 3 He system cools detector package to 250 to 300m K For the baseline cameras, the sensitivity is achievable with a head temperature of 225 mK For the baseline cameras, the sensitivity is achievable with a head temperature of 225 mK We get NEPs ~ 10 -16 W/Hz with Zeus at 250 mK We get NEPs ~ 10 -16 W/Hz with Zeus at 250 mK If necessary, ADR can cool system further (60 mK) The end stage coolers are closed cycle 3 He systems or ADRs that are temperature stabile, and vibration free

27. Low T Head “ He-7” system from VeriCold: Based on 4K pulse tube cooler two stage 4 He and 3 He sorption coolers by Chase Research Cryogenic Ltd. Based on 4K pulse tube cooler two stage 4 He and 3 He sorption coolers by Chase Research Cryogenic Ltd. 100 uW cooling @ 300 mK 100 uW cooling @ 300 mK Can go to 225 mK with “He-10” system: Dual stage 3 He sorption cooler Dual stage 3 He sorption cooler 50 uW cooling @ 250 mK in our ZEUS spectrometer 50 uW cooling @ 250 mK in our ZEUS spectrometerADR Typically has 3 He thermal shield Typically has 3 He thermal shield Provides ~ few uW cooling @ 60 mK Provides ~ few uW cooling @ 60 mK Dual stage 3 He cooler used in ZEUS/SPIFI Figure of helium-6 system removed to minimize file size

28. Photometer Cooler Summary Pulse tube cooler from Cryomech Cooling power of 0.7 W at 4.2 K Cooling power of 0.7 W at 4.2 K Power consumption ~ 7 KW Power consumption ~ 7 KW Price ~ 34 to 38 K$ including compressor (water or air cooled) Price ~ 34 to 38 K$ including compressor (water or air cooled) Low T cooler Cooling power 75  W at 250 mK Cooling power 75  W at 250 mK Power consumption negligible Power consumption negligible Price ~ 50 K$ for He-10 system Price ~ 50 K$ for He-10 system

29. Next Generation Instruments First light: 2 cameras and spectrometer “Borrowed” and/or second generation: Heterodyne spectrometer Heterodyne spectrometer Polarimetry Polarimetry Other types of direct detection spectrometers Other types of direct detection spectrometers 40 um camera? 40 um camera? More distant future Full FOV cameras Full FOV cameras  Very Large optics  “multiplexed” design

30. The Full Field of View The submm camera can be “multiplexed” either spatially, or spectrally 4 instruments cover 16’ FOV, or up to 4 bands 16‘

31. Summary We have quite a bit of expertise at Cornell/Caltech/JPL with submillimeter instrumentation Some of the first light instrumentation, especially heterodyne receivers may well be receivers “borrowed” from other facilities Our primary new instruments are: 200 to 620  m camera 200 to 620  m camera 850  m to 2.0 mm camera 850  m to 2.0 mm camera Low resolution direct detection spectrometers Low resolution direct detection spectrometers