Extreme Contrast Adaptive Optics with Extremely Large Telescopes by pptfiles

VIEWS: 8 PAGES: 55

									Extreme Contrast Adaptive Optics with Extremely Large Telescopes
Richard Dekany
Caltech Optical Observatories 23 July 2004

2004 Michelson Summer School Frontiers of High Contrast Imaging in Astrophysics

1

Outline
• Extremely large telescopes
– AO Scaling Laws – Thirty Meter Telescope

• Current and near-term ExAO state-of-the-art
– Palomar AO Coronagraph – ELT AO Technical Challenge

• ELT ExAO
– Architectural Elements – Performance Model • Menagerie of Worrisome Phenomena (10-6, 10-8, 10-10) • High-leverage Component Technologies – Potential Science Reach – Reaching the Fundamental Limits • Strategies for Low-Q Operation (e.g. IFU‟s) • Passive v. Active Speckle Suppression – Ground / space ExAO comparison summary
2

Extremely Large Telescopes

3

Lessons of History

•

Plot of largest optical/IR telescope size vs. time reveals exponential growth
– Remarkable given various social, economic, and technical factors
Log10 collecting area (meters2)

•

Extrapolating from Keck 10 m:
• 10 m 1993 • 25 m 2034 • 50 m 2065 • 100 m 2097 – History does not explain how future gains will be made

Courtesy J. Nelson 4

Hale

1949

Large telescope projects 1950-2020

1990 Keck1 HST

1995
Keck2 MMT HET Gemini (x2) VLT (x4) Magellan ….others

SIRTF

2000

LBT (x2) GTC NGST

2005

2010 TMT TPF

2015

OWL 2020

Adaptive Optics (AO) Scaling Laws
• AO significantly extends the science gain of large telescopes
– Signal-to-noise • AO off ~ D • AO on ~ D * (D/r0) ~ D2 r0-1, for unresolved background limited target • The AO gain, (D/r0) is typically 30 - 60 in the near-IR – With such promising return, it must be hard, right?…

• Required number of degrees of freedom ~ D2 l-12/5
– Required closed-loop bandwidth ~ l-6/5 – Required wavefront measurement photon flux ~l-18/5 – Required level of control of systematics ~l

• Note, scaling laws to reduce residual wavefront error (~l) are typically steeper than for increasing aperture diameter

6

Thirty-Meter Telescope
• Public / private collaboration of ACURA, AURA, Caltech, and UC • First light ~2015 w/ general-purpose AO • ExAO is currently a top priority 2nd generation instrument

7

ExAO contrast metric
• Approximate smooth-halo contrast estimate
– Collected planet flux grows  S * D2; where S is the Strehl ratio – Halo flux per AO diffraction-limited resolution element  (1-S) – Contrast within a resolutions element  Q  S D2 / (1-S)

• Practical contrast limits within today‟s small working angles are usually dominated by speckle noise from quasi-static errors
– Sources are typically non-common-path errors • Thermal induced telescope/instrument changes • Gravity gradients • Chromatic errors • Local turbulence effects

8

Photon-noise-limited ExAO contrast metric

TMT/MCAO
248 nm

TMT ExAO
Keck / XAOPI

248 nm

Keck

Gemini ExAOC
PALAO
165 nm

PALM-3000
85 nm

Adapted from J. Graham

9

TMT draft science capabilities
Field 10” FoV 20‟ FoR 10” FoV 5‟ FoR 2” FoV 2” FoV 5” FoV 30” FoV Mode n-IFU N-Slit 1-Slit n-IFU C = 108 - 2x1010 1-Slit 1-Slit Imaging Spatial l/D r0/D(/2) l/D ~l/D l/D l/D r0/D l/D Spectral R ~ 4,000 150 < R < 6,000 5,000 < R < 100,000 2,000 < R < 10,000 50 < R < 300 20,000 < R < 100,000 50,000 < R < 100,000 5 < R < 100 Wavelength (µm) 0.6 - 5 0.3 - 1.3 5 - 28 0.8 - 2.5 0.8 - 2.5 1-5 0.3 -1.3 0.6 - 5

1. 2. 3. 4. 5. 6. 7. 8. •

AO (---) AO AO AO AO ---AO

Notes:
– – – FoV = Field of View, FoR = Field of Regard (fields quoted by diameter) N >> n >> 1 (/2) Indicates GLAO option - to be evaluated

10

Adaptive Optics Modes
AO Mode
(w/ corresponding science capability ) Wavelength range Enabled science Components/ Instrument feed
•Multi Lasers •Deployable AO •MEMS •a) 0.005” IFU •b) 0.025-0.040” IFU •Cryogenic DM or •Adaptive Secondary •NGS or multi-lasers •MidIR Echelle Spectrometer •MidIR Imager •Optical multiobject spectrograph

Priority

MOAO
a.Small-Field (#1, #6) b.Multi-Objects on wide-field (#4)

a)0.65- 5m b)1-2.5 m

•Galaxy chemistry •Star forming chemistry

1st light, if successfully demonstrated

MIRAO
Mid IR (#3)

7-28m

•Star forming regions, protoplanetary disks •Characterize planetary systems; AGNs

1st light

GLAO
Wide Field (#2)
(Ground Layer)

0.31-1.0m

•Large sample galaxy spectra

Option on 1st light wide-field instrument Not yet known

ExAO
Extreme (#5)

0.8-2.5m

•Exo planet imaging •Protoplanetary disks

•MEMS •Coronagraph or Nulling Interferometer Planet Imager

MCAO
Multiconjugate (#8)

0.8-5m

•Dark ages •Early galaxies, AGNs •Nearby galaxies resolved star pop. and nuclei •Galactic Center •Star forming regions

•Multi Lasers •Tomography •Single or multi- DM •IFU (with imaging)

2nd light, assuming MOAO validation

11

TMT focal plane

12

ExAO state-of-the-art

13

Keck outer working angle N=18 Palomar outer working angle N=16 Palomar occulting spot

Gemini ExAOC outer working angle N=64 (2009) AEOS outer working angle N=32 (2004)

ExAO state-of-the-art

PALM-3000 outer working angle N=64 (2007) Courtesy B. Macintosh and S. Metchev 14

ExAO today
• ExAO today is saddled with the heavy yoke of general-use AO systems
– Today‟s AO: • Is designed to optimize faint guide star Strehl ratio over wide FoV • Relies on non-common-path and non-common-wavelength wavefront sensing • Uses 70 yr-old coronagraph technology • Tolerates hysteretic and temperature dependent deformable mirrors • Is devoid of any real-time metrology • And Nyquist samples the focal plane

• One can hardly imagine setting out to design a worse ExAO system • ELT ExAO systems are likely to:
– Be highly specialized to the specific scientific requirements (ie, search young systems for hot exo-Jupiters in emission; find water on exo-Earths at Eps Eri; etc.) – Pursue brand new architectures – Require successive generations of prototypes and demonstrations – Require large amounts of telescope time 15

Palomar Adaptive Optics

Open Loop FWHM ~0.70 arcsec Strehl ~2% at K Log Stretch

Closed Loop FWHM 0.090 arcsec Strehl ~80% at K 165nm Wavefront Error

• • •

Facility instrument at Palomar observatory for last ~4 years The most requested instrument at Palomar Natural guide star AO system
– 16x16 subapertures – Bright guide star Strehls as high as 80% at 2.2 mm • Maximum frame rate 2000Hz (<7e- read noise) – Limiting magnitude ~13.5mV, 10-15% Strehl at 2.2 mm • Read noise 3.5e- at < 500 fps

•

Science Camera
– J, H, and K imaging and 0.025 and 0.040 arcseconds/pixel – Coronagraph 0.41 and 0.91 arcsecond spot – J, H and K spectra at R~1500

16

PALAO High Strehl Images
 5 sec images Ten  s (2.145 mm) K  0(0.5 mm) ~ 11 cm r  Strehl (Mean/sample Stdev): 75% +/- 2%  RMS Wavefront error: 185 +/- 10 nm  Wavefront error Max/Min: 198/168 nm
 Strehl estimates are a lower estimate of truth  Ignore 2% Strehl loss to reflection  Ignore spiders an ~3% effect  Mean Wavefront 165 nm! 

17

Palomar AO Comparison

3 arcseconds

Log Stretch

Perfect Image

(simulated wave diffraction w/ spiders and 1.5% ghost)

Simulated AO Image RMS wavefront = 130nm Strehl=0.86

(using Caltechs Arroyo C++ lirary)

Measured image on sky RMS wavefront=165nm Strehl= 0.80

 Simulations performed using Arroyo (Caltech)  Wavelength Ks (2.145 mm, width=0.3mm)  second exposures 5  Excellent agreement with simulations!  Difference of 100nm is consistent with AO caliration errors

18

Palomar Performance

• • •

Excellent agreement with simulation High-Contrast imaging:
– AO corrected image is only a factor of 3 worse then perfect case for field angles greater then 0.5 arcseconds

Spectra (and optical communication):
– A factor of 2.4 improvement in 80% enclosed energy

19

Ks = 11.3 at 2.6″ (3.0*10-5)
HD 166435

6 x 60 sec on-source exposure with coronagraph; Ks band (2.16µm); V = 6.9; Strehl  65%.

Ks = 13.6 at 3.3″ (3.6*10-6)
Slide courtesy of Stan Metchev
20

AO challenge for ELTs

RMS wavefront error vs. Telescope Diameter
400 350 300 250 200 150 100 50 0
AO experience ELT LGS Trajectory

RMS wavefront error

mV < 7 mV = 13 mV > 16 mV < 7 (proposed) mV = 13 (proposed) mV = 16 (proposed)

1

10 Telescope Diameter (m)

100
21

AO development

D/sigma vs Year
1000.0

D/sigma * 10^6

mV < 7
ELT LGS Trajectory

mV = 13 mV > 16 mV < 7 (proposed) mV = 13 (proposed) mV > 16 (proposed)

100.0

AO experience

10.0 1980 1990 2000 Year
22

2010

2020

ExAO architecture elements
• A respectable first stage AO system
– Typically > 85% Strehl ratio (enables linearization of the residual phase errors)

• Excellent diffraction suppression
– Many techniques exist (e.g., occulting or phase-mask coronagraph, nulling beamcombiner, Gaussian pupil apodization, cats-eye apodization)

• Dedicated system for nanometer wavefront control
– Second stage high-order AO – Dark hole (Malbet), dark speckle (Layberie), black speckle (Dekany), ripple sensor (Angel, Traub)

• New detection architectures
– Polarization, multi-wavelength backends have make it to the telescope – IFU‟s, interferometers, statistics engines et al. have not yet

• Calibration, calibration, calibration
– Data analysis pipeline and algorithms directly drive the hardware architecture (integrated experiment design)
23

ExAO performance model

24

Key ExAO concepts
• Working region
– Area of focal plane between inner and outer working angle, where wavefront corrector exhibits beneficial control (a function of wavelength)

• Phase ripple
– A single frequency component of a two-dimensional wavefront phase power spectrum – Phase ripple variance, sk2, is integral of power spectrum from k to k+dk

• Q value
– Planet photoflux [photons/m2/sec] divided by stellar photoflux within a single focal plane resolution element • Q = 4 Cplanet/sk2 • Unpublished Palomar AO results (Boccaletti, 2002) hint that Q ~ 1/60 detections possible with existing systems and careful calibration
25

Small phase error PSF
• Using a 2nd order complex wave expansion, the halo can be described as the sum of individual haloes for independent error processes having unique spatiotemporal behavior

ptotal = a 
Fitting error
Wavefront variance rad2/m1/2

(

i

Fi

2

+ A'

2

)

• Example power spectra (following Rigaut et al. 1998)
Measurement error
Wavefront variance rad2/m1/2

Boiling wind errors

Wavefront variance rad2/m1/2

fc Spatial frequency, f

fc Spatial frequency, f

1/r0 Spatial frequency, f [m-1]

26

PSF halo widths
• The high-contrast point spread function (PSF) can be modeled as the superposition of haloes due to various error processes
– Each halo has it‟s own envelope, width (w), and speckle lifetime • For frozen wind and boiling wind, w = l/r0 • For scintillation, w = Sqrt(lz); the Fresnel length • For photon noise, WFS read noise, w = l/dx; dx = actuator spacing (assumed same as WFS sensor spacing for pupil sensor) • Interference effects average away over many speckle lifetimes – Each halo contributes to reduce the SNR of planet detection

• The contribution of the total wavefront variance attributable to a single phase ripple of spatial frequency k is
s2k = s2 / Nspeck in halo = s2 * (l/w)2/(l/D)2 = s2 * (w/D)2
27

Minimizing integration time
• Optimize AO system design for subaperture diameter, dx, and sensor sample time, dt, for different observation cases
SNR = Planet signal var( photon) + var( speckles)
16 (S F p A)  F   s  16 Fsky A + 16 S F p A + 16  N speckles   
sci 2 2 2

SNR 2 =

 Fssci   s +  N   i  speckles 
2 i

T

 (s
i

4 i

ti )

where S is Strehl ratio, Fp is planet flux, A is telescope area, Fsky is sky background flux, Fssci is parent star flux at science wavelength, s2i is wavefront variance from ith error process, w is ith halo width, ti is the coherence time of the ith process, and T is total integration time
28

Speckle noise
• Speckle noise (e.g. Racine, et al. 1999)
– Fundamentally different than photon noise • Speckle noise variance based upon the square of speckle photoflux • Smooth-halo photon noise variance based upon speckle photoflux
•

•

PALAO PSF stability (Apr 04) over 1 minute • 15 five-second K-band images taken on a 6th magnitude star in 0.9” (visible) seeing. The images are log stretched and 3 arcsec on a side. • The average Strehl is 80% +- 2%, equivalent to 165 nm +- 9nm of RMS wavefront error Coronagraph contrast (~ 5 x 10-4) dominated by speckle noise
29

ExAO issues I
Effect (10-6, 10-8, 10-10)
Aliasing in the wavefront sensor Aliasing in the science array Boiling wind (e.g. nonpredictable phase errors) Complex occulter index of refraction Chromatism Deformable mirror fitting error Detector charge diffusion and amplifier glow (science and/or WFS) Direct scintillation halo

Potential mitigation
Spatial filtering Focal-plane WFSing Spatial filtering Higher correction bandwidth Better understand and/or materials Meteorological monitoring Common-band WFSing Higher spatial bandwidth Improved detectors

Mitigation maturity
Moderate (simulations) Moderate (concepts) Moderate (simulations) Moderate Poor Moderate Moderate (concepts) Moderate Moderate

Active amplitude correction Two-conjugate correction Space and Ground Ground only

Poor Moderate (simulations)

30

ExAO issues II
Effect (10-6, 10-8, 10-10)
Dispersion displacement Flat-field stability

Potential mitigation
Lateral dispersion corrector Optimized spectralwidth of WFS Improved detectors

Mitigation maturity
Moderate (concepts) Moderate (concepts) Moderate

Fourth-order terms in the wavefront expansion
Frozen wind lag (e.g. predictable phase errors) Index of refraction inhomogeneities Multispectral error Non-common path phase errors

Higher Strehl Contrast-optimizing amplitude and phase control laws
Predictive phase correction of multilayer atmosphere More uniform materials, better pointing control Common-band WFSing Common-path WFSing Improved calibration/metrology

Poor Poor
Moderate (concepts) Poor Good Moderate (concepts) Poor

Non-common path polarization effects

Slow F/# telescopes, polarizers Vector field AO coronagraph design
Space and Ground Ground only

Poor

31

ExAO issues III
Effect (10-6, 10-8, 10-10)
Residual tip/tilt jitter Scintillation in WFS Telescope pointing errors (Beam walk using a T/T mirror) Uncorrectable dynamic telescope errors WFS calibration instability WFS star and background photon noise WFS read and dark current noise

Potential mitigation
Better control Amplitude correction Sensing of higher moments Better telescope pointing Adaptive secondaries (to minimize beam walk) Improved ACS, telescope stiffness, wind shielding WFS‟s insensitive to seeing changes Active thermal control Optimized system design Improved detectors

Mitigation maturity
Good Moderate, for clearing inner halo Moderate

Moderate Moderate Moderate Good Moderate

Space and Ground

Ground only
32

Cases to be considered
‘Fundamental’ errors
Planet photon noise Sky photon noise WFS photon noise Boiling wind Residual dispersion displacement

‘Expanded’ list of errors
Fundamental errors + Frozen wind WFS read noise Scintillation Residual chromatism Multispectral error

Required innovation
-Predictive control Noiseless detectors Amplitude correction Common-band sensing Common-band sensing

Sensing / Science mode
R-band / H-band

Comment
Traditional AO (chosen for maximum sky coverage) Limited by current deformable mirrors

R-band / R-band

H-band / H-band

Limited by current detectors
33

Performance model example
(with estimated H-band speckle coherence times)

• Planet photon noise • Sky photon noise (msky = 16)
• Extrasolar exozodical light well-resolved for D=30 m, so not significant

• WFS photon noise (tphot = dt, the system update rate)
• Atmospheric phase estimate imperfect due to WFS photon statistics

• Scintillation (tscint = 0.024 sec)
• Due to amplitude fluctuations arising from high-altitude turbulence • We will assume strong high-altitude turbulence

• Frozen wind (twind = 0.009 sec)
• Correction is late due to finite AO correction bandwidth
– Solution: By definition, can be eliminated with predictive controller

• Boiling wind error (tboil = 0.200 sec)
• Component of error not predictable

• WFS detector read noise (tphot = dt)
• Includes dark current shot noise, etc.
– Solution: Photon-counting detectors
34

Integration time by effect (30 m Sun-Jupiter-analogue)
Error terms
Tsi [hours]
R/H Expanded list of errors (dx = 0.33m, dt = 0.083 ms) H/H Fundamental errors (dx = 0.19m, dt = 0.17 ms)

Multispectral error

7.5

--

Residual chromatism
Dispersion displacement Scintillation Detector read noise Frozen wind

0.6
0.4 0.4 0.4 0.3

-0.0006 ----

WFS photon noise
Boiling wind Sky photon noise Planet photon noise Total integration time

0.15
0.01 0.002 0.0002 9.8

0.3
0.1 0.004 0.0003 0.4
35

Tint and Q / Expanded error terms
Integration time [hours] and Q value (Tint / Q) to achieve SNR = 5 for various ground-based aperture diameters and contrast levels 10 meter @ 10-8 10 meter @ 10-9 30 meter @ 10-9 30 meter @ 1.7 x 10-10

Expanded error terms

Sensing band / Science band

R/H
dxopt = 32 cm dtopt = 0.083 ms

7.4 / 0.001

740 / 0.0001

9.8 / 0.001

340 / 0.0001

R/R
dxopt= 31 cm dtopt= 0.12 ms

2.4/ 0.001

240 / 0.0001

3.1 / 0.001

110 / 0.0002

H/H
dxopt = 125 cm dtopt = 0.11 ms
(May violate quadratic expansion assumption)

4.6 / 0.001

460 / 0.0001

5.7 / 0.0001

200 / 0.0001
36

Tint and Q / Fundamental error terms
Integration time [hours] and Q value (Tint / Q) to achieve SNR = 5 for various ground-based aperture diameters and contrast levels 10 meter @ 10-8 10 meter @ 10-9 30 meter @ 10-9 30 meter @ 1.7 x 10-10

Fundamental error terms

Sensing band / Science band

R/H
dxopt = 16 cm dtopt = 0.15 ms

6.1 / 0.002

610 / 0.0002

8.1 / 0.002

280 / 0.0003

R/R
dxopt(10m) = 8.8 cm dtopt(10m) = 0.23 ms dxopt(30m) = 11 cm dtopt(30m) = 0.11 ms

0.10 / 0.008

9.5 / 0.0008

0.24 / 0.003

8.2 / 0.0006

H/H
dxopt = 19 cm dtopt = 0.17 ms

0.35 / 0.003

35 / 0.0003

0.4 / 0.002

15 / 0.0004
37

Scope of target list
100

Integration time [hours]

10 R/H Expanded H/H fundamental

1

0.1

0.01 0 1 2 3 4 5 6 7 8 Guide star magnitude [mV]

Integration time vs. guide star magnitude for R/H expanded and H/H fundamental error cases, using optimized dx and dt pairs. We again consider a Sun-Jupiter analogue at 10 pc, Cplanet = 10-9, D = 30m, 45 degree zenith angle.
For most cases, systems optimized for solar analogue good to mv = 6.
38

Better NIR detectors?
40 35 30 25 20 15 10 5 0 0 10 20 30 40 50

Integration time [hr]

H-band WFS read noise [electrons, rms]

Integration time as a function of WFS read noise for H/H operation and the expanded error list. For each value of read noise, an optimal dx and dt were determined. For zero read noise, the optimal dx = 0.20 m and dt = 0.075 msec, growing for read noise = 50 e- rms to dx = 1.7 m and dt = 0.290 msec. Note, for large values of dx, the Strehl ratio in practice falls due to wavefront fitting error, violating the assumption that the quadratic phase used here.

39

Larger telescopes
1000.000 100.000 10.000 1.000 0.100 0.010 0 10 20 30 40 50 60 70 80 90 100 Telescope diameter [m]

Integration time [hr]

R/H expanded H/H fundamental

Integration time vs. telescope diameter for R-band sensing/H-band science (expanded list of errors) and for H-band sensing/ H-band science (fundamental errors). The target system is a Sun-Jupiter analogue at 10 pc (Cplanet = 10-9) and the desired SNR = 5. Each case has been separately optimized (R/H has dx = 0.33 m and dt = 0.083 msec, H/H has dx = 0.19 m and dt = 0.16 msec). Exoearth times typically 50x greater, but use similar architectures w/ similar D dependence. Note, 8-10m’s can’t reach mature 5AU exojupiters at 10pc in reflection
40

Integration time vs. telescope diameter for solar analogue exoearth @ 10 pc
(l = 0.85 microns, r0 = 0.47 m, t0 = 10 ms, vwind = 15 m/s, sc2 = 0.006, tscint = 20 ms, Z = 20 km, toil = 90 ms, sread = 3 e-, n = 4 pix, q = 0.1, msky = 21.5 / asec2, R = 5, SNR = 5, Cplanet = 1.7 x 10-10)

Total Scintillation WFS read noise WFS photon noise Frozen wind Boiling wind Planet photon noise Sky photon noise

dt = 0.00007 sec, dx = 0.16 m

dt = 0.00014 sec, dx = 0.20 m

30 m telescope requires 70 hours

30 m telescope requires 5.7 hours
41

Exoplanet astrometry and photometry
at 30 m fundamental limits • With repeated imaging observations, we can deduce
– From orbital characteristics • Equilibrium temperature • Tidal locking • Resonances among sibling planets – From phase function • Presence of a cloudy atmosphere • Albedo rotation rates

• Mean radius of habitable zone at 15 pc
– 31 l/D (R-band) and 13 l/D (H-band) – Aggressive apodization possible due to large collector and high angular resolution

• For nearest few stars, binary exoearths or exoearths „moons‟ of exojupiters could be resolved (but SNR still low)
– @ 3 pc, resolution of 0.01 AU  25 Rjupiter = orbital radius of Callisto

• Within 15 pc, there are hundreds of plausible candidate stars for TMT-based exoearth search at R = 5 (down to mV = 6)

42

ELT ExAO Potential Science Reach

43

Bright stars in 15

44

Ground-based exoearth spectroscopy
at 30 m fundamental limits

• Photon-noise limited R = 5 spectroscopy (visible and near-IR) would enable:
– The presence of a clear atmosphere (e.g. Earth via Rayleigh scatter), a deeply clouded atmosphere (e.g. Venus via Mie scatter)

• R = 20 spectroscopy might be reachable
– Require long integrations and careful calibration of Telluric effects – Notable exceptions possible in sub-classes of exoearths • e.g. H2O steam lines (seen in brown dwarfs from the Earth‟s surface) • Other plausible, unearthly atmospheres can be imagined

• Biomarkers (e.g. O2, O3, CH4) are likely not available with 30 m SNR from Earth‟s surface
– High spectral resolution (R=70) implies prohibitive integration times – Telluric confusion may not be soluble at such small SNR • Technique using orbital Doppler shifting of narrow lines, used to study brown dwarfs, generally not available due to low R
45

Ultimate science reach for 30m
• Fundamental limits for AO allow direct detection of exoearths with TMT but not biomarker studies • Potential number of systems
– Thousands for hot, young exojupiters (R = 10-1000) – Hundreds for mature exojupiters (R = 10-100) – Scores for exoearths (R = 5) • Each requires tens of hours of observation
– Observations favor R = 2 R planets, e.g. waterworlds

• High-resolution spectroscopy is very difficult except for nonterrestrial atmospheres (e.g. steam lines or severe pressure broadening)

46

Reaching the Fundamental Limits

47

Reaching the ground fundamental limits
• New techniques
– Speckle noise suppression • Post-processing
– Chromatic techniques – Photon statistical techniques

• Active
– Higher-Strehl ratio operation – Complex amplitude optimizing control laws (not phase conjugation)

• New components
– Deformable mirrors with 104 - 105 actuators – Stable back-end instruments – Focal plane wavefront sensors

• Prototype systems
– Develop H/H or R/R band AO systems optimized for high contrast – Many currently uncontrolled error processes must be addressed by design (partial list follows) – Typical development cycle for 8-10m telescope is 5 years and $10M • Likely to need several generations to get from 2 x 10-4 to 10-8
48

Speckle discrimination in post-processing

• Published techniques for PSF subtraction
– Achromatic techniques • PSF calibrator star
– COME-ON Plus (c. 1997)

• Multiple “roll angles”
– Field (Keck) and pupil (Palomar) rotation (c. 2000)

• Centro-symmetric PSF subtraction (2002) – Chromatic techniques • Discreet multispectral discrimination
– TRIDENT – 3 channel (2001) – Several discrete channel successors (2004+)

49

Active speckle suppression
• The large penalty for speckle noise arises when bright focal plane speckles are allowed to build up (typ. 1000‟s photons)
– This suggests one strategy: avoid speckle noise by running closed-loop correction so fast that speckles typically contain only a few photons

• Wavefront sensing in the focal plane
– Minimizes speckle noise (as well as many other error sources) – Decoupling of wavefront sensing (into the focal plane) allows more flexibility in DM technology (at the pupil plane) • Phase and wavelength information are both needed • Concepts:
– An interferometric technique has been suggested by Angel (2002) – Superconducting tunnel junctions (STJ‟s) appear well-suited, but remain small format – New field, open to new architectures

– While good Strehl is needed to sharpen planet light, modest DM formats (typ. N = 128) allow exploration of habitable zones on exoearth target list
50

IFU‟s for exoplanet study
• The logical extension of multichannel coronagraphic imagers
– R = 20-100 spectroscopic speckle discrimination – An intermediate step toward spectroscopic focal-plane wavefront sensors

• ExAO is new application for IFU‟s
– Requires development of new observational techniques and data analysis – Requires a professional group of exoplanet IFU researchers – We need to learn how to use these things

• Near-term integral field coronagraph (IFC) prototype options
– Lab tests • Rapid prototyping, but does not engage scientific community – Existing AO systems • All require new (presumably, warm) coronagraph relay and have larger than necessary spectral resolution
– Slit spectrographs (most existing AO systems) – AO-fed IFU‟s (e.g. Keck/OSIRIS, 2004)

– Gemini Extreme AO Coronagraph (ExAOC) (2009) • Instrument call includes consideration of an IFU-mode

51

IFU‟s for exoplanet study (cont.)
• Demand for low-Q operation drives ground ExAO concepts
– A photon-counting IFU can be used to determine wavefront amplitude and phase and drive the „sharp-end‟ of an optimized ExAO system (e.g. hierarchical control)

• IFU technology in the path of TMT and other ELT ExAO development • Ground-based experience with ExAO IFU‟s could be extremely useful for TPF coronagraph mission design
– – – – Similar sub-component requirements (Detectors, fibers/slicers, etc.) Similar data sets Similar analysis techniques (Implies similar humans) Difference is only bandwidth of wavefront control loop

52

Why must TPF work at Q=1?
• Low-Q operation aka speckle discrimination is fundamental to all techniques of high-contrast direct detection, and is stock in trade for ground systems • Ground-based observers only have just a few years experience in PSF calibration, but no one on the ground is planning Q=1 instruments
– Q = 0.25 published PALAO (Boccaletti, 2002) – Q < 0.05 reported MMT (Close, private communication, 2003) – Q = 0.016 unpublished PALAO (Boccaletti, 2002)

• Working Q value for ground-based exoplanet study will be < 0.1 for next 20 years
– We don‟t fret about about this, but seek to develop new techniques

• Significant relaxation of TPF coronagraph requirements possible if tightest contrast (1/2 exoearth) requirements planned for Q = 0.1
53

Exoearth study comparison
• Ground 30m
– Cons
• • • • • • Biomarkers not accessible R < 20 (integration time limited) Must master Q = 0.001 One hemisphere accessible Not top ELT priority Many new technologies required to reach fundamental limits
– Speckle suppression

• Space
– Pros
• • • • • • • • • • Biomarkers accessible R > 20 Could work at Q = 0.1 – 1.0 Science in visible or mid-IR Whole sky accessible Enterprise mission – top priority Slow mission development Typically 0.02 nm rms WFE Highly stable Many new technologies required to reach fundamental limits
• Speckle suppression

– Cons

– Pros
• • • • Rapid instrument development Smaller inner working angle Higher spatial resolution Science possible in red and near-IR • Typically 10 nm rms WFE

54

EAO key concepts
• Phase ripple
– A single frequency component of a two-dimensional wavefront phase power spectrum – Phase ripple variance, sk2, is integral of power spectrum from k to k+dk

• Q value
– Planet photoflux [photons/m2/sec] divided by stellar photoflux within a single focal plane resolution element • Q = 4 Cplanet/sk2 • Unpublished Palomar AO results (Boccaletti, 2002) hint that Q ~ 1/60 detections possible with careful calibration

• Working region
– Area of focal plane between inner and outer working angle, where wavefront corrector exhibits beneficial control (a function of wavelength)

55


								
To top