LARGE-OPTICS TECHNOLOGY LESSONS FROM BALL’S NGST AND TPF PROGRAMS STEVE KILSTON BALL AEROSPACE & TECHNOLOGIES CORP. HUBBLE’S SCIENCE LEGACY: FUTURE OPTICAL/UV ASTRONOMY FROM SPACE APRIL 5, 2002 S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 1 Ball’s Two Large-Optics Space Astronomy Programs TPF NGST TRW-Ball S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 2 Main Topics • Key System Performance Capabilities – – – – • • Wavelength Coverage Angular Resolution System Sensitivity and Dynamic Range Usable FOV TPF Visible-light Coronagraph Description Relevant Technology Concerns – – – – – – – – Instrument Technologies Spacecraft Technologies Launch and Deployment Operational Reliability far from Earth Robustness Risks Costs Technology Development Roadmaps S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 3 System Performance Dependencies PERFORMANCE Wavelength Coverage Sensitivity Min. Sun Angle Sky Coverage Optical FOV Signal-to-Noise Ratio Detector Fill Factor System RMS Noise Thermal Stability COST Detector Array Size Slew Time System MTF ACS -Related Jitter Pixel Subtense Quantum Efficiency Orbit Number of Spacecraft Detector Size System Vibration Ground Segment Costs Space Segment Costs Launch Vehicle Sizes Optical Wavefront Accuracy Optical Throughput System Component Costs Launch Costs Angular Resolution Focal Length Instrument Cost Data Rates & Processing Complexity Spacecraft Bus Cost Target Flux Spectral Band Detector Noise Momentof-Inertia Integration Time On-Orbit Lifetime Stiffness & Damping Spacecraft Mass Key: Thermal System ACS Size and Type Aperture Diameter S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS Other Ground Segment Factors Propulsion System Specified Value P. 4 Driving Design Factor Comparison of NGST, TPF, and Hubble-2 Performance NGST (~ 6-m aperture) Wavelength Coverage TPF Hubble-2 (Visible-light, (8-m aperture, 4 x 10 m aper.) SUVO paper) 0.6 - 28 µm 0.3 - 1.0 µm 0.11 - 1.0 µm Angular Resolution 30 mas (λ/14 , 0.75 µm) 10 x 25 mas (λ/140) 15 mas (λ/14 , 0.5 µm) System Sensitivity MAB = 31 V = 34 MAB = 32 Usable FOV, arcmin 5x6 >2x2 12 x 12 S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 5 Ball TPF Architecture Study Team Astrophysics, Biomarkers, Optical Science Lead, Team Lead, StarLight Contractor, Concepts, Wavefront Control, Planets, Astrophysics, S/C, Optical Systems, Formation Metrology, Controls, Orbits Interferometry Flying, Metrology, Mechanisms Mid-Wave IR Imaging and Instruments Large Optical Ground Systems Design Arch. Concepts, Beam and Wavefront Control, Fourier Transform Spectroscopy Center for Astrophysics Planet Theory, Biomarkers Biomarkers, Analysis, Interferometry Planet Atmospheres, Chemical Evolution, Biomarkers Arch. Concepts, Opt. Sys. Analysis Astrophysics,Metrology, Interferometry, Formation Flying Exoplanetary System Detections S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS Large Optics, Integration and Test IR Systems, Launch, Ops, Propulsion, Detector Systems S/C Controls, Formation Flying Gossamer and Inflatable Structures P. 6 Some Leading Lights of the Ball TPF Architecture Study Team S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 7 Ball TPF Visible-light Coronagraph -Major Features • • • • • Radial slices thru PSF 10 15 10 10 10 5 10 10 deg cone 20 deg cone 0 deg Sky FOV (arcsec) Light intensity (ph/s/cm 2 /um) • -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Earth-like planet 0 0 2 4 Sky FOV (rad) Image of star and planet at focal plane -6 6 x 10 -1 -0.5 0 0.5 1 Sky FOV (arcsec) S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS 6 5.5 5 4.5 4 3.5 3 2.5 2 • • • • • • • Science Directly detect Earth-like planets in the habitable zones surrounding ~ 150 solar-type stars within 20pc Observe Biomarkers from 0.4-1.0 µ m General Astrophysics research utilizing the largest telescope mirror in space Optics 10 x 4 Meter Monolithic Primary, 4-cm thick + backplane, 940 kg Active Optics for Wavefront Control Off-Axis Design, plus shaped-pupil options – very low diffraction Spacecraft Launch in 2015 Earth Trailing Drift-Away or L2 orbit Delta IV H – direct insertion launch 3-Axis Stabilized Spacecraft Large Thermal Shield Launch Mass ~ 6,000 kg 5-Yr Design Life (10 yrs of Expendables) P. 8 Main Technology Issues Affecting Performance • Optical Architecture & Size -- for adequate performance & to fit in launcher • Active Optics -- wavefront quality for large, thin optics; mechanical or thermal • Wavefront Sensing -- to guide active optics and/or a deformable mirror • Coatings, Filters, & Passbands -- UV operation a challenge • Stability: Thermal & Mechanical -- for efficient use of AO; shields, damping • Testability -- large test chamber, demonstration of end-to-end performance • Contamination -- mild requirements for visible-light system, strict for UV • Detectors -- CCD successors, better UV QE; maybe λ info., photon counting • Radiation-hardening of Electronics -- vital for deep space operation • Observing Efficiency -- slew rates, settle times, field-of-regard • Orbit and Launcher -- driven by thermal stability, data link, servicing needs? • Mission Time Frame -- affects technology development possibilities • Dual-use Missions -- TPF as Hubble-2 affects design, does astrophysics S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 9 Large Space Optics Issues • Segmented vs. monolith -- monolith difficult, but minimizes diffraction – Phasing of segments, phase stability and maintenance – Segment latches for deploy/redeploy • Elliptical vs. circular -- elliptical easier to package, maximizes resolution • Off-axis vs. on-axis -- clear aperture needed for coronagraph • Deployable vs. rigid structures • Material -- Glass vs. composite vs. Be; elastic memory materials – Precision, lightweight (<25 kg/m 2) glass honeycomb developed by Kodak - baseline for TPF coronagraph-- lowers mass, thus mission cost; sets best operating temperatures • Optical quality excellent for large UV telescopes • • – Difficult to achieve optical quality soon w/ thin films (stretched membranes, shells) – Nonlinear behavior in zero-g environment Laboratory development needs to begin now if monolith (long-lead item) – Start with 0.5 - 1.0 m flat mirror, move on to half-scale powered mirror (2 x 5 m) – Validation in space? Flight system test & optical alignment techniques Passive vs. controlled surface and/or Deformable/Correction optics S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 10 Segmented Optics Construction • Monolithic primary mirror built from segments drawing on AMSD technology • AMSD mirror/actuator system areal density = 15 kg/m2 (TPF allocates 35 kg/m2) • Core and faceplate segments fused together to form the monolithic PM blank • Existing Kodak facilities and capabilities will be utilized to produce segments • New Kodak facilities will be needed to process and test the full primary mirror • 30 cm ULE blank • The TPF monolithic PM will have a segmented front and back plane – – – – – – Segmented core Segmented back plate 19 mm deep core Low Temp Fusion of segments Proof of concept blank Core strut thickness of 0.02” (0.5 mm) demonstrated S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS TAKE PICTURES. FURTHER. P. 11 TPF Primary Mirror Technology Development Plan OBJECTIVES Proof-of-Principle Produce 1/2 Meter Flat •Develop processing technique using a flat lightweight optic with segmented face sheets. •Demonstrate low spatial frequency surface control with actuators Segment Demonstration Produce 1+ meter powered segment • Repeat above objectives on larger lightweight optic with power. Sub-scale Development Design, and build 1/4 scale Primary Mirror (PM) TAKE PICTURES. FURTHER. S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS • Achieve TPF requirements on subscale PM Full-scale Development Design and build full scale Primary Mirror P. 12 Active Optics • One stage of correction for NGST, two for a TPF coronagraph – Actuators between primary mirror back surface and backplane, ~ 10/m2 • Stroke - tens of µm; setting resolution ~ 7 nm – Deformable mirror (DM) for mid spatial frequencies, ~ 250 x 100 elements • Stroke - ~ 100 nm, compensates for primary mirror actuator inaccuracy • Setting resolution (iterative setting avoids absolute accuracy reqts.) – Xinetics DMs have achieved resolution of 25 pm • Temporal and thermal stability requirements – Xinetics DMs stable to 100 pm over weeks < 10 mK temperature stability needed • Need accelerated life tests to prove space durability • Other types and suppliers of DMs are entering the arena DM • Comparison to ground-based Adaptive Optics – Much lower temporal frequency required in space -- DM fixed for hours – Wide field of view possible in space -- no isoplanatic angle – Atmospheric dispersion in visible/UV limits spectral bandwidth • Potential use for correcting amplitude non-uniformity S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 13 Mirror Coatings and Spectral Passbands • Tradeoff of throughput vs. spectral range (especially UV) – Silver gives high reflectivity, but no UV performance – Aluminum allows operation in the UV – Reflectivity ~ 92-96% per mirror across the visible band – ~ 50-60% throughput after 10 mirrors – TPF coronagraph design uses Al for primary and secondary mirror, Ag for subsequent mirrors • Pickoff mirror could feed UV instruments (no access to DM) • Example of dual-use mission • Durability and technical challenges • System spectral measurement accuracy involve filter and detector performances and stabilities – Good calibration devices and processes are needed S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 14 Wheel Isolation Greatly Reduces LOS Jitter at High Frequencies 2 With isolator included 1 10 1 mas allocation 1 10 Large, slow RW set 10 4 Hz 0 RMS jitter (mas) RMS jitter (mas) 10 0 10 -1 10 -2 10 Small/fast RW set -3 10 0 10 1 10 2 Choose small/ fast RW 10 Reaction wheel speed (RPS) • • • • -1 10 10 -2 1.5 Hz isolation -3 10 0 10 1 10 2 10 Reaction wheel speed (RPS) Line of sight jitter is below the 1 mas RMS requirement over most speeds. Small/fast wheel produces less jitter than the larger wheel HST-class wheel with 1.5 Hz isolation will meet requirements except at a few resonances. These results are open-loop – they don’t engage fine steering mirror pointing loop S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 15 Dynamic mirror distortion due to wheel harmonics is expected to be tolerable • Without isolator: Movie of primary mirror deformation no RWA isolation • Worst-case primary mirror deformation (41 Hz) Movie of star/planet PSF no RWA isolation planet • Gives spatial/ temporal RMS = 8.4nm (~17 nm WF error) • With 1.5 Hz isolator: • Worst-case primary deformation (4.6 Hz) Movie of primary mirror deformation with RWA isolation planet • Spatial/temporal RMS 0.05 nm (~ 0.1 nm WF error) • Doubling wheel speed expected to suppress by > 10x S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS Note : Tip/tilt component removed - deformation only. P. 16 Summary of Initial Design of Thermal Control Subsystem Thermal Design Provides Accurate Control over Primary Mirror Temperature variations Spatial variations (over 24-hour cycle) Controlled to within 0.03 ºC Without full Sunshield, Primary Mirror Spatial variations Exceed 0.12 ºC Active Heater Control of Optical Bench is Required Passive Design Would Result in Primary Mirror Temperatures of -140 ºC Primary Mirror Operating Temperatures of 0 ºC are Favored: Simplifies Manufacturing Testing & Calibration. Also at 0ºC CTE (Coefficient of Thermal Expansion) of ULE, Zerodur are Minimized Active Heating Provides Precise Control over Primary Mirror Temperature Gradients - Bench Heater Power Requirement ~ 700 W ZERODUR CTE Curve S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 17 Observatory Ground Test Overview (Integrated System Test) • • • S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS Several Facilities to choose from – AEDC Mark I φ12.8m x 25m – GRC SPF φ30.5m x 37.2m – Johnson A φ16.8m x 27.4m Test Set up Evaluated for Plum Brook Space Power Facility – Largest Facility – Very Low vibration level – NASA controlled facility Test the Entire Observatory – Only the Sunshield & Array Removed for Testing – Vertical orientation eliminates moments into primary aperture – Vacuum Test at on-orbit thermal environment P. 18 Delta IV (4050-H19) Heavy Launch Configuration Baseline 4 x 10 Primary Mirror • Robust Launch Margins • Direct Injection to Heliocentric Earth Trailing Drift Away Orbit Several Packaging Approaches still need to be explored Currently not taking full advantage of Delta IV capabilities (35% Launch Margin) Need to work trades to maximize primary mirror with adequate launch margin (20 to 25% pre-phase A margins should be acceptable) • • • – – Launch Mass ~ 6,000 kg Launch Capability of 9,255 kg to C3 = 0.3 km 2/sec 2 S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS Alternative 4 x 13 Primary Mirror P. 19 TPF Visible-Light Coronagraph -Technology Readiness and Prospects NASA TECHNOLOGY READINESS LEVELS Technology Current TRL Future work Estimated Completion Date Resulting TRL Large, lightweight optics 3 Lab demo of scale models 2005 6 Wavefront sensing with science camera 3 TPF techology testbed demos 2005 6 Deformable Mirrors 5 Lifetime tests in lab 2003 6 Thermal Control 4 NGST validation 2008 7 2 Lab demo of full scale masks 2003 6 6 NGST validation 2008 8 Binary Pupil Masks with edge adjustment (CG) Integrated model of full optical system S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 20 Risk Matrix for UV/isible Coronagraph Technology Risk Level Large Optics Moderate Wavefront Stability Moderate Amplitude Uniformity Moderate Wavefront Sensing Low Deformable Mirrors Low Binary Pupil Masks (option) Low Graded FPA Masks (option) Moderate S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 21 Conclusions and Recommendations for UV/Optical Hubble 2 • A large-mirror telescope with wavefront quality excellent for UV can be built and tested, and is feasible for launch in the 2015 timeframe – – – – Technology development should be adequately supported Active-optics primary actuators could enable high-quality segmented mirror DM could set wavefront to accuracy needed for coronagraph (λ/10,000) Needs high mechanical stability throughout an observation • Maintains wavefront accuracy • Therefore thermal & vibrational control – Elliptical monolith primary would fit in launch fairing • Thermal contol to millikelvin levels may be the key technology need – Thermal shield plus computerized network of temp. sensors and heaters – Need to model and demonstrate level achievable • Will be expensive (up to $1.5 B), but can host several instruments – High spatial and spectral resolution, good FOV and UV/optical- coverage S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 22 TPF Technology Studies Roadmap Coronagraph Technology Areas Proof-of-Conc. Studies Large, lightweight optics Lab Demonstrations Wavefront Sensing Integrated Modeling Pupil and FP Masks Wavefront Stability Deformable Mirrors Amplit.Uniformity, Thermal and Vibrational Effects Common Technologies Phenomenology TPF Test Benches - Common Scene Lab Demonstrations Cryogenic Actuators Cryogenic Nulling Angle Tracking/Metrology Formation Flying Control Science Developments Kepler Keck Interferometry Interferometry Technology Areas Coronagraph and Interferometer Comparison Tests LBT Interferometry Integrated Modeling Wavefront Stability Amplitude Uniformity, Thermal and Vibrational Effects Precursor or Parallel Technology Developments Eclipse or similar missions NGST ESA Studies StarLight TPF Architecture Selection S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS SIM P. 23 Dual-use Missions • TPF/SUVO is the main example (if TPF is a visible-light coronagraph) – TPF is charged with doing astrophysics 30%-50% of the time – Both need large, high-accuracy mirrors • TPF in visible needs primary mirror better than λ/30 at UV – High pointing accuracy & stable thermal environment permit long exposures – A pickoff mirror after the Al-coated primary and secondary mirrors can divert the light beam into UV instruments • Even a 4-m TPF/SUVO could detect Earths at 8 pc, and many Jupiters • Operational reliability without servicing for deep-space missions could lead to more dual-use missions (which presumably get higher funding) – Bigger funding levels could increase reliability (not faster, better, cheaper) S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 24 Operational Lifetime • Need to figure out optimum lifetime (science likes long life; costs much) • Operational reliability without servicing for deep-space missions could lead to more dual-use missions (which presumably get higher funding) – HST had 240 serviceable components, needed to replace only 33 in 12 years – Bigger funding levels could increase reliability (not faster, better, cheaper) S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 25 Wavefront Sensing (WFS) • WFS method depends on required accuracy – Shack-Hartmann method limited to ~λ/100 or coarser • Adequate for ground-based adaptive optics – Phase-diversity methods can potentially reach λ/10,000 • Needed for high contrast coronagraphy • Algorithms require demonstration at high accuracy levels • No additional hardware required - uses science camera – TPF/SUVO combination could use this + mirror actuators for UV observing • Integration time as a function of accuracy – Limits sensitivity if WFS time is limited (e.g. ground-based observing) – Limits observing efficiency otherwise • Update frequency depends upon system stability – Achievable stability for TPF coronagraph has not yet been fully quantified S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 26 Contamination • “Easy” in visible/IR - hard in UV (affected by few nm thickness) - Affects options: closed vs. open structure, intermittent heating MOLECULAR: multi-molecular drops ~1-10 nm high PARTICLES: dominant particle sizes ~0.1–1.0 µm MICROMETEORITES: impacts on the telescope system will be common Molecular deposit quantities At launch ~ 0.5 - 1 nm Steady-state on orbit ~ 0.5 - 1.5 nm Molecular contaminant rates — < 0.004 nm / day Level 100 - 200 Level 100 - 200 Quiescent periods Nil Nil During meteor impact — TBD Particle distributions Particle redistribution rates S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 27 Detector Technology -Summary of Characteristics and Types ----------------------------------------------------------------------------------------------------CHARACTERISTICS CCD Si Hybrid MCP APD STJ Microbolo InSb ----------------------------------------------------------------------------------------------------Wavelength coverage ( µm) 0.1-1 0.1-1 0.01-1 0.1-1 0.01-5 any 0.4 - 5 Pixel pitch ( µm) 6-25 17-30 ? ? ? ? 25 -50 QE (%, UV has lower QE) >80 >80 >90 >90 ? ? 85 Read noise (electrons) 3* 3* 0 0 ? ? 15 Dark current (counts/s/pix) 0.01 ? ? ? ? ? 0.01 Overlight damage N N Y N N N N CR degradation low low low ? ? ? low Format size (2001) 4K x 4K 1K x 1K 1K x 1K ? 10x10 100x100 4K x 4K Energy resolution N* N* N(?) N(?) Y Y N Time stamping N* N* Y Y Y Y N Operating temperature (K) 150 30 150 150 0.1 0.1 77 Power supply 5V 5V 1 KV 5V 10 mV 1 mV 15 V Technology readiness high moderate high low low low high ----------------------------------------------------------------------------------------------------* Modified CCDs and hybrid devices that support non-destructive read modes can in principle have less than 1 e/read effective read noise and can therefore be used to detect individual photons Modified CCDs can time-tag photon arrival to within a frame read-time Since there is no read-noise penalty, low dark-current devices can be mated to integral field spectrographs to determine the energy of each photon Backup chart P. 28 S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS Stability: Thermal and Mechanical • Balance between passive and active is a system level trade – Depends on relevant timescales • Performance requirements for TPF coronagraph appear to require both passive and active thermal controls – – – – – • Large sunshields, insulation, and ULE glass needed for passive control Distributed heating elements for active control Open vs. closed loop control Open Loop - thermal modeling can predict required heater settings Closed Loop - network of temperature sensors leads to real time feedback Effects of localized vibration sources (reaction wheels, thrusters) can be suppressed with isolators – Disturbances are high-frequency, allowing strong suppression • • Microsnap (spontaneous release of strain) is poorly quantified and difficult to isolate, since distributed throughout the structure Integrated modeling required for optimal system design – Quantifies performance advantages of thermal & mechanical design options S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 29 Thermal effects on wavefront error • First order finite-element thermal model of primary mirror – With full sunshield – Divided mirror into 500 elements – Examined effects of large azimuth slews (no change in star) separated by 8 hour integrations • Results – Front-to-back temperature difference changed by 1.2 mK over 8 hours • Coherent across full mirror, with rms ripple of ~0.1 mK – We calculate 0.03 Å peak-peak ripple on front surface due to glass only • Assumed ULE glass with CTE of 10 -8/K • The mirror backing structure has not been included yet – Will add a large low-frequency contribution to deformation – Design has not been optimized to control thermal deformations S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 30 Abstract Ball Aerospace and Technologies Corp. and its partners have been considering many large space-astronomy design concepts, especially through our work on the Next Generation Space Telescope (NGST) and Terrestrial Planet Finder (TPF) programs. During those efforts our teams strove to be open-minded and creative about design approaches and suitable technology elements. Our evaluations of the candidate concepts were based on performance, costs, and technology risks and required development. We assessed many performance factors relevant to an optical-UV successor to the Hubble Space Telescope (HST), such as system sensitivity, angular resolution, and usable fields of view. Our teams investigated key technology issues such as attitude control, thermal and vibrational disturbances, ease of launch and deployment, and operational reliability beyond Earth orbit (if that's required). In the case of TPF we considered several large-optics designs operating at optical wavelength bands, with wavefront quality more than adequate for UV operation too. We'll discuss here some of our work's potential implications for choosing an HST optical-UV successor's technologies. S. KILSTON -- LARGE-OPTICS TECHNOLOGY LESSONS P. 31