Presentation File

High-contrast AO for imaging extrasolar planets

(formerly known as Extreme AO)Bruce Macintosh (LLNL)

Outline

• Science motivation for Extreme AO: Imaging extrasolar planets

• Fourier optics with perfect wavefronts – coronagraphs• Fourier optics with phase errors – High-contrast AO PSFs• ExAO system design: the Gemini Planet Imager

Formation history is encoded in distributions

Core acceretion + migration predictions (Ida&Lin 2004)

Orbital scattering in 3 body systems;Chatterjee et al. astro-ph/0703166

5 AU 50 AU

Disk fragmentation efficient at 10-20 AU

Ma

yer et a

l. 20

02

20 AU

Qmin=1.7

Qmin=1.4

160 yr 350 yr

Doppler

Direct detection & spectroscopy of brown dwarfs

Mclean et al 2003

Lafrienere et al 2007 (Gemini Planet Survey) etc.

GDPS

Uncertainty in luminosity of young planets

Marley et al 2006 astro-ph/0609739

Previous models

Low-entropy core accretion models

Extreme AO regime

Current AO surveys

Voyager “family portrait”

Conventional AO limited by scattered light

Strehl ratio S

Halo intensity 1-S

“Extreme” AO (ExAO)

gain > S/(1-S)

High-contrast AO PSF

• Fraunhoffer regime: focal plane and pupil plane are connected by Fourier transforms

• (x,y) = pupil plane coordinates– Natural coordinate system is in units of

telescope diameter

x=x[m]/D

• (= focal plane coordinates– Natural coordinate system is in units of

/D– XD

• Spatial frequency 1/a <=> angular scale /a

• Upper case / lower case = fourier transform pairs– Upper case for pupil plane

• e() = FT[E (x,y)]• P,p = PSF (intensity)

E

e

FT

Pupil electric field from aperture and phase

Pupil plane Focal plane

( , )

2

( , ) ( ( , ) )

( , ) ( , )

i x ye FT A x y e

p e

Φ=

=

( , )( , ) ( , ) i x yE x y A x y e Φ=

E(x,y)

e

A = aperture Φ = phasea, = fourier transforms of above

Simple case: uniform phase


2

( , ) ( ( , )) ( , )

( , ) ( , )

e FT A x y a

p a

= =

=

( , ) ( , )E x y A x y=

E(x,y)

e

A = aperture Φ = phasea, = fourier transforms of above

A |a|2

For small phase errors: Taylor expansion (Sivaramakrishnan et al 2002, Perrin et al 2003)

)2

1(

),(),(2

),(

K+Φ

−Φ+=

= Φ

iA

eyxAyxE yxi


( , )

2

( , ) ( ( , ) )

( , ) ( , )

i x ye FT A x y e

p e

Φ=

=

2( , )

2 *

* * ** * *

*

* * *

* *

* * * *12

( , ) ( ( , ) ) ( ...)2

* *( * ) ...

2

( , ) ( , ) ( , ) ( , )

* * * *( ( * ) ...)( ( * ) ...)

2 2

[ ( ) ( )]

( )( )

( ) (

i x y Ae FT A x y e FT A Ai

aa i a

p e e e

a aa i a a i a

aa

i a a a a

a a

a a a a

φ φφ

φ φ φ φφ φ

φ φφ φ

φ φ φ

Φ Φ= = + Φ − +

= + − +

= =

= + − + − − +

=

+ ∗ − ∗

+ ∗ ∗

− ∗ ∗ + ∗ ∗ )

...

φ⎡ ⎤⎣ ⎦+

PSF expansion

PSF terms

0 1 2

*0

* *1

* *

* *2

* * * *12

...

[ ( ) ( )]

2 Im[ ( )]

( )( )

( ) ( )

p p p p

p aa

p i a a a a

a a

p a a

a a a a

φ φ

φφ φ

φ φ φ φ

= + + +

=

=− ∗ − ∗

= ∗

= ∗ ∗

⎡ ⎤− ∗ ∗ + ∗ ∗⎣ ⎦

• Diffraction pattern term

Airy pattern

aa*=|FT(A)|2 is the diffraction term

Two-d Airy patterns

Coronagraphs

• Invented by Bernard Lyot in 1930 for studying the corona of the sun without waiting for an eclipse

How can we control diffraction?

PSF=aa*=|FT(A)|2A

PSF

Coronagraph 1: Gaussian apodization

Coronagraph 101: Blackman or Kaiser apodization

A=0.42-0.05 cos[2(r+0.5)] +0.08 cos[4(r+0.5)]

• More complex functions can have higher contrast or better throughput

• Apodizers in general are hard (impossible) to manufacture

Apodization in 2d

Shaped-pupil coronagraphs (Kasdin et al. 2003)

Pupil PSF

Lyot coronagraph (Lyot, 1933)

Starlight

Lyot coronagraph (Lyot, 1933)

Planet

Sivaramakrishnan et al 2001 has a nice 1-d analysis of how this works

Many new coronagraphs in recent years

• Explosion of coronagraph concepts in recent years• Lyot family:

– Basic: Lyot 1939 MNRAS 99, 538; Sivaramakrishnan et al 2001

– Band-limited: Kuchner & Traub 2003

– Apodized: Soummer 2005 Ap.J. 618, L161

• Apodizers:– Shaped-pupil: Kasdin et al 2003, Kasdin et al 2005 Applied Optics

44 1177, etc.

– Phase-induced apodizer: Guyon et al 2005 Ap.J. 622, 744

• Interference / wave-optics– 4-quadrant phase mask: Rouan et al 2000 PASP 777 1479

– Nulling interferometer/coronagraphs: Mennesson et al. 2004 Proc. SPIE 4860, 32

• Optical vortices, many others…• Most practical coronagraphs only work at > 3-5 /D• Control of phase errors has been neglected

PSF terms

0 1 2

*0

* *1

* *

* *2

* * * *12

...

[ ( ) ( )]

2 Im[ ( )]

( )( )

( ) ( )

p p p p

p aa

p i a a a a

a a

p a a

a a a a

φ φ

φφ φ

φ φ φ φ

= + + +

=

=− ∗ − ∗

= ∗

= ∗ ∗

⎡ ⎤− ∗ ∗ + ∗ ∗⎣ ⎦

• Diffraction pattern term

• Pinned speckle term– Antisymmetric– Traces the diffraction pattern;

vanishes when diffraction is negligible– See Bloemhof 2003, Perrin et al 2003

• Halo term– ~=||2 (power spectrum of Φ– Symmetric– Dominant source of scattered

light in high-contrast AO!• Strehl term

– Removes power from PSF core

d

/d

White noise

White noise

AO architecture and terms

Collimating Lens

Tip/TiltMirror

WavefrontSensor

Dichroic

Science Camera

D = primary mirrordiameter

DM conjugate to telescopeprimary

d=actuator spacing

d

WFS conjugateto DM & primary

Atmosphere parameters:Coherence length r0

Wind velocity v

DeformableMirror

Phase

Spatial frequency

Power spectrum

Spatial frequency

Spatial frequency

Phase

Spatial frequency

Power spectrum

Spatial frequency

Spatial frequency

Phase

Spatial frequency

Power spectrum

Spatial frequency

Spatial frequency

Phase

Spatial frequency

Power spectrum

Spatial frequency

Spatial frequency

Phase

Spatial frequency

Power spectrum

Spatial frequency

Spatial frequency

Phase

Spatial frequency

Power spectrum

Spatial frequency

Spatial frequency


Collimating Lens

Tip/TiltMirror

WavefrontSensor

Dichroic

Science Camera



d=actuator spacing

d



Wind velocity v

DeformableMirror

Phase Power spectra

Phase Power spectra

Band-limiting for anti-aliasing: spatial filter

PS

F in

tens

ity

Position (arcsec)

/dap

Spatial filter (Poyneer and Macintosh 2004) implementation

WavefrontSensor

DeformableMirror

Dichroic

Science Camera+Coronagraph

Focal stop spatial filter

/d=0.9”

Phase Power spectra

AO Tim

elag

WFS measurement

Inner working distance ~3-5 /D

Fitting error

Outer working distance ~N /D

Random intensity of all the Fourier components produces speckles

(ExAO PSF movie goes here)

As speckles average out (~D/vwind)planets can be detected


Collimating Lens

Tip/TiltMirror

WavefrontSensor

Dichroic

Science Camera



d=actuator spacing

d



Wind velocity v

DeformableMirror

ExAO 0 nm static errors, 5 MJ/500 MYr planet, 15 minute integration




ExAO and the Gemini Planet Imager

2003: Basic ExAO feasibility study and Keck strawman

2004: Gemini “Extreme AO Coronagraph” Conceptual design begins

2005: CfAO team selected

2006: (June): Project start

First light: 2010Team

LLNL: Project lead + AO

AMNH: Coronagraph masks&design

HIA: Optomechanical + software

JPL: Interferometer WFS

UCB: Science modeling

UCLA: IR spectrograph

UdM: Data pipeline

UCSC: Final integration&test

Calibration Module

LOWFS

Referencearm shutter

LO pickoff

Phasing Mirror

Apodizer Wheel

Woofer DM & Tip/Tilt

Linear ADC

F/64 focusing ellipseDichroic

Focal Plane Occultor Wheel

IR spectrograph

Collimator

Beamsplitter

Polarization modulator

Lyotwheel

Lenslet

Stage

Pupil Camera

Zoom Optics

Prism

HAWAIIII RG

Pupil viewing mirror

WFS

Lenslet

WFS P&C& focus

SF

FilterWheel

CCD

FilterWheel

Entrance Window

IR Self-calibration interferometer

Artificial sources

Pinhole

IR CAL WFSCAL-IFS P&C & focus

WFS collimator

DewarWindow

Filter Wheel

Polarizing beamsplitter and

anti-prism

Gemini f/16 focus

MEMS DM

AO

Coronagraph

High order high-speed AO (LLNL)

MEMS deformable mirror

Woofer DM

Calibration/ Alignment

Unit

Spatially Filtered WFS 0.7-0.9 m

GPI Window

Focal stop spatial filter

/d=0.9”

Commercial computer Fourier (predictive)

control

Keck AO (1999)

GPI

(2010)

Deformable mirror

349 actuators

(240 active)

4096 actuators

(1809 active)

Subaperture 56 cm 18 cm

Control rate 670 Hz 2000 Hz

Wavefront sensor

Shack-Hartmann

400 – 1000 nm

Spatially-filtered SH

700-900 nm

Strehl @ 1.65 m

0.4 >0.9

Guide star mag

R<13 mag. I<9 mag.

(V<11 aux.)

Superpolished optics (2 nm RMS)

Apodized-pupil Lyot coronagraph (Soummer 2005)

Apodizer

Hard-EdgedMask

Lyot Mask

Soummer 2005

Integral field spectrograph (James Larkin, UCLA)

DetectorLenslet Array

Collimator Optics

Camera Optics

Focal Plane

Pupil Plane

Rotating ColdPupil Stop

Filters

R.I. TelephotoCamera

Lenslet

Spectrograph

Collimated lightfrom Coronagraph

Prism

Window

Low spectral resolution (R~50)

High spatial resolution (0.014 arcsec)

Wide field of view (3x3 arcsec)

Minimal scattered light

UCLA Spectrograph format

• Each spectrum is 16 pixels long, one of YJHK, =50

• 68,000 spectra on a 2048x2048 detector 4.5 pixel spacing

• 2.8 x 2.8 arcsecond field of view, 0.014 arcsecond pixels

Single Spectrum

UCLA Broad-band ExAO snapshot

UCLA ExAO spectral data cube

James Larkin, UCLA

O1 O2

O3 O4

Fresnel optics effects (more complicated than simple Fraunhoffer model) cause speckles from aberrations near focus not to subtract as well

Marois et al. 2006, Spie

GPI mechanical design

GPI enclosure

Electronics

Gemini Cassegrain

support structure

Optics structure

Gort

GPI optical structure

VLT Planetfinder: SPHERE

Monte Carlo models of science performance(Graham&Macintosh)

Monte Carlo models of science performance(Graham&Macintosh)

ExAO can detect a significant population of planets

Radial velocity detections

GPI detections

Extrasolar planets

H=8-11 mag

H=5-8 mag

H=4-6 mag

Space AO: Terrestrial Planet Finder

• Terrestrial Planet Finder Coronagraph (was 2020, now deferred)

• Original baseline: 8x3m mirror with advanced AO to correct internal errors

• Coronagraph works at 4 /D -> 0.08 arcseconds for 8-m telescope– Earth at 10 pc = 0.1 arcsec

• Various interim 2-4 m class missions proposed with more advanced coronagraphs– 2-3 /D coronagraph allows smaller

telescope

• Some visible-light spectroscopy of Earthlike planets

Extrasolar planets

H=8-11 mag

H=5-8 mag

H=4-6 mag

TPF space coronagraph

Extrasolar planets

H=8-11 mag

H=5-8 mag

H=4-6 mag

Small TPF

A very large coronagraph

TPF Occultor (Webster Cash et al)

References

• Angel, R, “Ground based imaging of extrasolar planets using adaptive optics’, 1994 Nature 368, 203 (Original exoplanet paper)

• Burrows, A., et al., “A nongray theory of extrasolar planets and brown dwarfs”, 1997 Ap.J 491, 856 (Planet models)

• Sivaramakrishnan, A., et al., “Ground-based coronagraphy with High-Order Adaptive optics”, 2001 Ap.J. 552, 397 (Lyot coronagraphs)

• Kasdin, N.J., et al, 2003, “Extrasolar planet finding via optimized apodized pupil and shaped pupil coronagraphs”, Ap.J. 582, 1147

• Kuchner, M, and Traub, W., “A Coronagraph with a Band-limited Mask for Finding Terrestrial Planets” 2002 Ap.J. 570, 200 (improved Lyot coronagraph)

• Sivaramakrishnan, A., et al, “Speckle decorrelation and dynamic range in speckle noise limited imaging”, 2002 Ap.J. 581, L59 (2nd-order PSF expansion)

• Perrin, M., et al. “The structure of the High Strehl Ratio Point-Spread Functions”, 2003, Ap.J. 596, 702 (high-order PSF expansion)

• Poyneer, L, and Macintosh, B., “Spatially-filtered wavefront sensor for high-order adaptive optics”, 2004, JOSA A 21, 810 (aliasing + WFS)

• Guyon, O., et al. “Theoretical Limits on Extrasolar Terrestrial Planet Detection with Coronagraphs”, 2006 Ap.J.S. 167, 81

• Cash, W., et al, “The New Worlds Observer: using occulters to directly observe planets”, 2006 Proc. SPIE 2625

Technology

Presentation File