page 1 lecture 17 the applications of tomography: ltao, mcao, moao, glao claire max ay 289 uc santa...
TRANSCRIPT
Page 1
Lecture 17Lecture 17
The applications of The applications of
tomography:tomography:
LTAO, MCAO, MOAO, GLAOLTAO, MCAO, MOAO, GLAO
Claire MaxAY 289
UC Santa CruzMarch 7, 2013
Page 2
Outline of lectureOutline of lecture
• What is AO tomography?
• AO applications of tomography
– Laser tomography AO
– Multi-conjugate AO (MCAO)
– Multi-object AO (MOAO)
– Ground-layer AO (GLAO)
• Much of this lecture is based on presentations by Don
Gavel, Lisa Poyneer, and Olivier Guyon. Thanks!
3
Limitations for AO systems with one guide star
• Isoplanatic AngleLimits the corrected field
θ0h
r0
4
Limitations for AO systems with one guide star
• Cone effect
5
Limitations for AO systems with one guide star
• Cone effect1. Missing turbulence
outside and above cone
2. Spherical wave “stretching” of wavefront
More severe for larger telescope diameters
h
r0
Fundamental problem to solve: Isoplanatic Angle
If we assume perfect on-axis correction,
and a single turbulent layer at altitude h,
the variance (sq. radian) is :
σ2
= 1.03 (θ/θ0)5/3
Where α is the angle to the optical axis,
θ0 is the isoplanatic angle:
θ0 = 0.31 (r0/h)
D = 8 m, r0 = 0.8 m,
h = 5 km θ0 = 10”
h
Fundamental problem to solve: Isoplanatic Angle
If we assume perfect on-axis correction, and a single turbulent layer at altitude h, the variance (sq. radian) is :σ2 = 1.03 (θ/θ0)
5/3
Where α is the angle to the optical axis, θ0 is the isoplanatic angle:θ0
= 0.31 (r0/h)
D = 8 m, r0 = 0.8 m, h = 5 km θ0
= 10”
h
Page 8
Francois RigautFrancois Rigaut’’s diagrams of s diagrams of tomography for AOtomography for AO
90 k
m
“Missing” Data
Credit: Rigaut, MCAO for Dummies
Page 9
What is Tomography ?What is Tomography ?2. Wider field of view, no cone effect2. Wider field of view, no cone effect
90 k
m
Credit: Rigaut, MCAO for Dummies
Tomography lets you reconstruct turbulence in the entire cylinder of air above the telescope mirror
Page 10
Concept of a metapupilConcept of a metapupil
• Can be made larger than “real” telescope pupil
• Increased field of view due to overlap of fields toward multiple guide stars
11
• Each wavefront sensor measures the integral of index variation along the ray lines
• The line integral along z determines the kz=0 Fourier spatial frequency
component
• Projections at several angles sample the kx,ky,kz volume
kX
kZ
Fourier slice theorem in tomography(Kak, Computer Aided Tomography, 1988)
How tomography works: from Don How tomography works: from Don GavelGavel
12
kX
kZ
• The larger the telescope’s primary mirror, the wider the range of angles accessible for measurement
• In Fourier space, this means that the “bow-tie” becomes wider
• More information about the full volume of turbulence above the telescope
How tomography works: from Don How tomography works: from Don GavelGavel
Page 13
y
How tomography works: some How tomography works: some mathmath
• Assume we measure y with our wavefront sensors
• Want to solve for x = value of δOPD)
• The equations are underdetermined – there are more unknown voxel values than measured phases ⇒ blind modes. Need a few natural guide stars to determine these.
• wherey = vector of all WFS measurementsx = value of δOPD) at each voxel in turbulent volume above telescope
x
A is a forward propagator
Page 14
y
x
Solve for the full turbulence above Solve for the full turbulence above the telescope using the back-the telescope using the back-propagatorpropagator
y
x
AT is a back propagator along rays back toward the guidestars
y = vector of all WFS measurements
x = value of δOPD) at each voxel in turbulent volume above telescope
Use iterative algorithms to converge on the solution.
Page 15
Tomography movie: RagazzoniTomography movie: Ragazzoni
Page 16
LGS Related Problems: LGS Related Problems: ““Null Null modesmodes””
• Five “Null Modes” are not seen by LGS (Tilt indetermination problem)→ Need 3 well spread
NGSs to control these modes
• Tilt Anisoplanatism : Low order modes (e.g. focus) > Tip-Tilt at altitude
→ Dynamic Plate Scale changes
Credit: Rigaut, MCAO for Dummies
Page 17
Outline of lectureOutline of lecture
• Review of AO tomography concepts
• AO applications of tomography
– Laser tomography AO
– Multi-conjugate AO (MCAO)
– Multi-object AO (MOAO)
– Ground-layer AO (GLAO)
Page 18
The multi-laser guide star AO zoo The multi-laser guide star AO zoo (1)(1)
Narrow field, suffers from cone effect
Narrow field, cone effect fixed
Page 19
The multi-laser guide star AO zoo The multi-laser guide star AO zoo (2)(2)
Both correct over a wide field, at a penalty in peak Strehl
Page 20
The multi-laser guide star AO zoo The multi-laser guide star AO zoo (3)(3)
Correct over narrow field of view located anywhere w/in
wide field of regard
Quite modest correction over a very wide field of
view
Central parts of the globular cluster Omega Centauri, as seen using different adaptive optics techniques. The upper image is a reproduction of ESO Press Photo eso0719, with the guide stars used for the MCAO correction identified with a cross. A box shows a 14 arcsec area that is then observed while applying different or no AO corrections, as shown in the bottom images. From left to right : No Adaptive Optics, Single Conjugate and Multi-Conjugate Adaptive Optics corrections. SCAO has almost no effect in sharpening the star images while the improvement provided by MCAO is remarkable.
Credit: ESO
MCAO on-sky performanceMCAO improves image quality where there is no single nearby bright guide star
Page 22
Outline of lectureOutline of lecture
• Review of AO tomography concepts
• AO applications of tomography– Multi-conjugate adaptive optics (MCAO)– Multi-object adaptive optics (MOAO)– Ground-layer AO (GLAO)
Page 23
What is multiconjugate AO?What is multiconjugate AO?
Deformable mirror
Turbulence Layers
Credit: Rigaut, MCAO for Dummies
Page 24
Deformable mirrors Turbulence Layers
Credit: Rigaut, MCAO for Dummies
What is multiconjugate AO?What is multiconjugate AO?
Page 25
The multi-conjugate AO conceptThe multi-conjugate AO concept
Telescope
DM1 DM2
Turb. Layers#2 #1
Atmosphere
WFS
UP
Credit: Rigaut, MCAO for Dummies
Page 26
Difference between Laser Difference between Laser Tomography AO and MCAOTomography AO and MCAO
• Laser Tomography AO can be done with only 1 deformable mirror
• If used with multiple laser guide stars, reduces cone effect
• MCAO uses multiple DMs, increases field of view
Page 27
WFSs
Guide Stars
DM2
Telescope
High Altitude Layer
Ground Layer
DM1WFC
““Star Oriented” MCAOStar Oriented” MCAO
• Each WFS looks at one star
• Global Reconstruction
• n GS, n WFS, m DMs
• 1 Real Time Controller
• The correction applied at each DM is computed using all the input data.
Credit: N. Devaney
Page 28
WFC2
DM2
WFS2
WFC1
DM1
WFS1
Telescope
High Alt. Layer
Ground Layer
Guide Stars
Layer Oriented MCAOLayer Oriented MCAO
• Layer Oriented WFS architecture
• Local Reconstruction
• x GS, n WFS, n DMs
• n RTCs
• Wavefront is reconstructed at each altitude independently.
• Each WFS is optically coupled to all the others.
• GS light co-added for better SNR.
Credit: N. Devaney
Page 29
MCAO Simulations, 3 laser guide MCAO Simulations, 3 laser guide starsstars
Strehl at 2.2 μm3 NGS, FoV = 1.5 arc min
Strehl at 2.2 μm3 NGS, FoV = 1 arc min
Credit: N. Devaney
As field of view increases, average Strehl drops and variation over
field increases
As field of view increases, average Strehl drops and variation over
field increases
Page 30
VLT Multi-conjugate AO VLT Multi-conjugate AO Demonstrator (MAD) showed that Demonstrator (MAD) showed that MCAO worksMCAO works
• Originally built as a technical demo to exercise in the lab, and then on the sky
• Compare layer-oriented with star-oriented MCAO
• Uses natural guide stars Users love it! Became
a regular VLT instrument
Page 31
Results from ESOResults from ESO’’s Multiconjugate s Multiconjugate AO Demonstrator (MAD)AO Demonstrator (MAD)
Single Conjugate Multi Conjugate
Effective θ0 ~ 20” Effective θ0 ~ 40-50”
Page 32
Science papers with MAD: Orion Science papers with MAD: Orion star clusterstar cluster
• MCAO correction is clearly better than single conjugate AO, but this is the case all across the image (not clear what θ0 is doing)
Page 33
Orion star cluster with MAD, Orion star cluster with MAD, continuedcontinued
Page 34
Jupiter with MAD: single conjugate Jupiter with MAD: single conjugate AO has done as well or better AO has done as well or better than MCAOthan MCAO
MAD MCAO Regular Keck AO
Jupiter is about 30” across
Page 35
Early impressions of MCAO Early impressions of MCAO performance based on MAD performance based on MAD results to dateresults to date
• Can extend θ0 from ~ 20 arc sec all the way to ~ 60”
• Some penalty in peak Strehl, in return for larger corrected field
• But really need three guide stars - hard to find if using “real” stars. (Hence MAD’s lopsided Strehl maps.)
• MCAO should really excel with laser guide stars– Gemini South’s GEMS is first to do this
• Note: Solar AO has demonstrated benefits of MCAO using multiple “regions of interest” as wavefront references
Page 38
GEMS image of star formation in GEMS image of star formation in OrionOrion
• Orion star forming region:
• Compare GEMS MCAO with ALTAIR single conjugate AO on Gemini North Telescope
ALTAIR GEMS
Page 40
Outline of lectureOutline of lecture
• Review of AO tomography concepts
• AO applications of tomography– Multi-conjugate adaptive optics (MCAO)– Multi-object adaptive optics (MOAO)– Ground-layer AO (GLAO)
Page 41
Distinctions between multi-Distinctions between multi-conjugate and multi-object AOconjugate and multi-object AO
• DMs conjugate to different altitudes in the atmosphere
• Guide star light is corrected by DMs before its wavefront is measured
• Only one DM per object, conjugate to ground
• Guide star light doesn’t bounce off small MEMS DMs in multi-object spectrograph
?
1-2 arc min
Page 42
Multi-Object AOMulti-Object AO
• Correct over multiple narrow fields of view located anywhere w/in wide field of regard
• In most versions, each spectrograph or imager has its own MEMS AO mirror, which laser guide star lights doesn’t bounce off of
• Hence this scheme is called “open loop”: DM doesn’t correct laser guide star wavefronts before LGS light goes to wavefront sensors
• In one version, each LGS also has its own MEMS correction
Page 43
Science with MOAO: multiple Science with MOAO: multiple deployable spatially resolved deployable spatially resolved spectrographsspectrographs
• A MEMS DM underneath each high-redshift galaxy, feeding a narrow-field spatially resolved spectrograph (IFU)
• No need to do AO correction on the blank spaces between the galaxies
Page 44
Why does MOAO work if there is Why does MOAO work if there is only one deformable mirror in the only one deformable mirror in the science path?science path?
• Tomography lets you measure the turbulence throughout the volume above the telescope
90 k
m
Page 45
Why does MOAO work if there is Why does MOAO work if there is only one deformable mirror in the only one deformable mirror in the science path?science path?
• Tomography lets you measure the turbulence throughout the volume above the telescope
• In the direction to each galaxy, you can then project out the turbulence you need to cancel out for that galaxy
90 k
m
Page 46
Existing and Near-Term MOAO Existing and Near-Term MOAO SystemsSystems
• CANARY (Durham, Obs. de Paris, ONERA, ESO)– MOAO demonstrator for E-ELT– On William Herschel Telescope; one arm– Had already commissioning run – First NGS (done), then Rayleigh guide stars
• RAVEN (U Victoria, Subaru, INO, Canadian NRC)– And Celia Blain!– MOAO demonstrator for Subaru telescope– 3 NGS wavefront sensors– Field of regard > 2.7 arc min– Will feed an IR science camera – Delivery in 2012
Page 48
VILLAGES at Lick Observatory: VILLAGES at Lick Observatory: Demonstration of Open-Loop Demonstration of Open-Loop PerformancePerformance
• Initial demonstration of open-loop MEMS performance– VILLAGES experiment on 1-meter telescope at
Mount Hamilton (Lick Observatory)– It works!
Page 49
Page 50
Both E-ELT and TMT have done Both E-ELT and TMT have done early designs for MOAO systemsearly designs for MOAO systems
• Artist’s sketch of EAGLE MOAO system for E-ELT
• One of the constraints is that the spectrographs are very large!
• Hard (and expensive) to fit in a lot of them
Page 51
Outline of lectureOutline of lecture
• Review of AO tomography concepts
• AO applications of tomography– Multi-conjugate adaptive optics (MCAO)– Multi-object adaptive optics (MOAO)– Ground-layer AO (GLAO)
Page 52
Ground layer AO: do tomography, Ground layer AO: do tomography, but only use 1 DM (conjugate to but only use 1 DM (conjugate to ground layer)ground layer)
single DM conjugated to ground layer
Credit: J-M Conan
GLAO uses 1 ground-conjugated DM, corrects near-ground turbulence
MCAO GLA
O
Page 53
Correcting just the ground layer Correcting just the ground layer gives a very large isoplanatic gives a very large isoplanatic angleangle
• Strehl = 0.38 at θ = θ0
θ0 is isoplanatic angle
θ0 is weighted by high-altitude turbulence (z5/3)
• If turbulence is only at low altitude, overlap is very high.
• If you only correct the low altitude turbulence, the isoplanatic angle will be large (but the correction will be only modest) Telescope
Common Path
Page 54
Ground Layer AO (GLAO) typically Ground Layer AO (GLAO) typically decreases natural decreases natural ““seeingseeing”” by a by a factor of 1.5 to 2factor of 1.5 to 2
• Example: GLAO calculation for Giant Magellan Telescope (M. Johns)
• Adaptive secondary conjugation at 160 m above primary mirror.
• Performance goals:
– λ > 0.8 μm
– Field of view: >10’
– Factor of 1.5-2 reduction in image size. Modeled using Cerro Pachon
turbulence profile. (M-L Hart 2003)
Page 55
Many observatories have Many observatories have ambitious GLAO projects plannedambitious GLAO projects planned
• Near term on medium sized telescopes: SOAR (4.25m), William Herschel Telescope (4.2m), MMT (6.5m)
• Medium term on VLT (8m), LBT (2x8m)
• Longer term on Giant Magellan Telescope etc.
• Is it worth the large investment “just” to decrease “seeing” disk by factor of 1.5 to 2 ?
– Depends on whether existing or planned large spectrographs can take advantage of smaller image
– Potential improved SNR for background-limited point sources
Page 56 Credit: A.
Tokovinin
time
Page 57 Credit: A.
Tokovinin
Page 58
Initial demonstration of GLAO at Initial demonstration of GLAO at MMTMMT
• With GLAO, image width is reduced from 0.70 to 0.33 arc sec and peak intensity is increased by factor of 2.3
• Credit: Christoph Baranec, Michael Hart, and colleagues, U AZ
Page 60
GLAO at the MMT (Rayleigh guide GLAO at the MMT (Rayleigh guide stars)stars)
Page 61
SOAR Telescope
Page 62
SummarySummary
• Tomography: a way to measure the full volume of turbulence above the telescope
• Once you have measured the turbulence there are several ways to do the wavefront correction
– Laser Tomography AO: Multiple laser guide stars, 1 DM, corrects cone effect. Narrow field.
– Multi-conjugate AO: Multiple DMs, each optically conjugate to a different layer in atmosphere. Wider field of view.
– Multi-object AO: Correct many individual objects, each over a small field. Each has very good correction. Wider field of regard.
– Ground-layer AO: Correct just ground layer turbulence. Very large field of view but only modest correction.
• All four methods will be used in the future
Page 63
Backup slidesBackup slides
Alt
itu
de
Cn2
n
(Turbulence Strength) (Index variations)Projection
Cn2
weighted back-
projection
Uniformly weighted back-
projection
5-guidestar
reconstructions
Reconstruction using back-projections with CReconstruction using back-projections with CNN22
x
GS1 GS2S
z
Page 65
Cone (laser guide star) back-projectionsCone (laser guide star) back-projections
Alt
itu
de
n
(Index variations)
3 laser guidestar reconstuction 5 laser guidestar
reconstuction
5 natural star reconstuction
x
GS1 GS2S
z