Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.arcetri.astro.it/~salinari/GSF/Wavefront%20Control.pdf Äàòà èçìåíåíèÿ: Thu Nov 27 19:50:39 2003 Äàòà èíäåêñèðîâàíèÿ: Sun Apr 10 04:10:13 2016 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: m 35

Wavefront Control
Prospects for the evolution of astronomical telescopes in the next two decades
(P. Salinari INAF-Arcetri, Florence, I)

???
1610 Galileo's Telescope 16 mm lens 0 DoF
Munich Dec. 1-3 2003 Global Science Forum

2004 Large Binocular Telescope 2x 8400 mm mirrors 2x 672 DoF adaptive secondaries
Wavefront Control (P. Salinari)

2025 What telescope? How many DoF?
1


Section 1: Introduction
Narrowing down the subject Establishing priorities Reminding the logics of some basic choiches

Munich Dec. 1-3 2003 Global Science Forum

Wavefront Control (P. Salinari)

2


Talk subject Astrophysics is a purely observational science, that needs multiple sources of information using ·Electromagnetic spectrum (gamma rays to radio) ·Particle spectrum (neutrinos to atomic nuclei) ·Gravitational waves spectrum (low to high frequency)

, X, UV, Visible, NIR, MIR, FIR, SMM, MM, CM, M (almost) all the above observations can be done from the Space and from the ground
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 3


What are the (visual, IR ) science requirements? Cosmology, evolution of galaxies:
·wide spectral coverage (R to K minimum) ·very high sensitivity (in imaging and spectroscopy) ·moderate angular resolution (tens of mas) ·large samples ("wide" field, several arc-minutes)

Star (and planets) formation and evolution:
·Wide spectral coverage (0.5 to 30 µm) ·high angular resolution (a few mas) ·high contrast (10-15 magnitudes, for stellar studies)

Extrasolar (terrestrial) planets:
·Wide spectral coverage (320 to 2500 nm minimum) ·Sensitivity to V>30 ·high angular resolution (a few mas) ·extreme contrast >25 mag
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 4


Sensitivity and angular resolution . . . . . . are related . . .
when the dominant source of noise is background photon statistics and the sources are not resolved S/N D/ (D= telescope diameter, = resolved angle) if /D (diffraction limited case) S/N D2

. . . angular resolution and contrast are also related . . .
in at C the diffraction limited case, if angular separation between two sources is , = n/D, or n = D/ , the contrast C (fraction of light scattered by source 1 location of surce 2 is (approximately) 1/n2 or C (/D)2/ 2

. . . moral: most of the above science requirements are on angular resolution
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 5


CELT, ~ 30 m, ~ 300 K

Space versus ground (visual, IR)

· Space is the only choice for wavelengths where the atmosphere is opaque and the winning choice (in terms of sensitivity) in the thermal IR ( > 3 µm) where the background level is much lower. · (A very large groundbased telescope can be competitive in the thermal IR only if both, high angular and high spectral resolution, JWST are required) James Webb · Where the background is comparable Space Telescope (Visual, Near IR), a larger telescope 6.5 m, ~ 30 K on ground is potentially more sensitive From ~ 2011 provided its angular resolution is comparable or higher than that of the (smaller) telescope in space
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 6


Why ground? Mainly because of cost. The HST with ~ 0.1 arcsec and about 1/2 the sky background in the visual range provides deeper images of the Universe than a seeing limited ( ~ 1 arcsec) 8 m telescope on ground. If the telescope on ground has the same ~ 0.1 arcsec (or better) It becomes more sensitive.
Wavefront Control (P. Salinari)

A 2.4 m telescope in space ( HST), ~ 2.5 Billion $

An 8 m telescope on ground (VLT), ~ 100 Million $
7

Munich Dec. 1-3 2003 Global Science Forum


On ground: filled, non-filled, segmented pupils
The interferometric combination of two or more separate telescopes (VLTI, Keck Interferometer) is a way of obtaining very high angular resolution ( ~ /B), but at the cost of: 1. Poor "image" quality 2. Modest optical efficiency 3. Modest sensitivity

VLT Interferometer

Keck Interferometer

The "interferometric" combination of Separate, rigidly connected, pupil segments (Keck, LBT) (gaps < segment sizes) is equivalent to a filled pupil of equal size, providing · similar angular resolution (/D) LBT segmented pupil Keck segmented pupil · similar sensitivity (for equal area)
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 8


Summary of section 1

In the visual-NIR range the most efficient and cost effective way of pursuing the currently identified science objectives is that of programming very large filled aperture, ground based telescopes

provided
they are designed to work at (or near) the diffraction limit

Munich Dec. 1-3 2003 Global Science Forum

Wavefront Control (P. Salinari)

9


Section 2: Wavefront Control (with enphasys on the most demanding requirements)
Defining the problem Remainding the state of the art Discussing physical limitations Identifying further developments

Munich Dec. 1-3 2003 Global Science Forum

Wavefront Control (P. Salinari)

10


What does Wavefront Control (WfC) mean? Correcting the wavefront errors due to its propagation through the turbulent atmosphere (on small spatial and temporal scales) means improving · image sharpness · sensitivity · contrast
Effects of increasing WfC bandwidth ( to ~ 50Hz - 3 dB) (real H band images from the 6.5 m MMT telescope, the first ever equipped with an adaptive secondary mirror in 2002)
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 11


terminology
SR FWHM C() Image sharpness (FWHM) (Full Width at Half Maximum) measured 0.065 arcsec (diffraction 0.060 arcsec) Sensitivity: Signal overcomes Noise

Contrast C() ratio of peak intensity to intensity at angle

computed diffraction profile

measured image profile

Strehl Ratio (SR): ratio of measured peak to diffraction peak
Wavefront Control (P. Salinari) 12

Munich Dec. 1-3 2003 Global Science Forum


Ennemies of image sharpness and contrast
1. 2. Atmospheric turbulence (only partially corrected by Adaptive Optics) Diffraction effects · By pupil outher edge (largely curable by pupil shape choice + coronagraphy) · By pupil inner edge (smaller effect, but more difficult to cure) · By secondary support structure (spikes only in a few directions) · By primary (and other) mirror segmentation (a variety of small, but nasty, effects) Vibrations of optical components Non uniform reflectivity (amplitude variations) Scattering by defects, edges, dust . . .

3. 4. 5.

Only N 1 is specific of groundbased telescopes (and is the worst ennemy). I will only discuss point 1 and touch some of the aspects of point 2 All · · · the other effects are tractable by appropriate telescope design choices coronagraphyc techniques severe tolerancing
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 13


Propagation through turbulence
wavefront before reaching the atmosphere

Atmospheric refraction index inhomogeneities (temperature inhomogeities in a turbulent flow, moving with (different) winds at (different) heights Wavefront after propagation hrough the atmosphere: is corrugated (phase error) and does not propagate in the original direction, neither in average (overall tilt) nor locally. This produces image motion (tilt), image blur (local corrugations) and, after some propagation, intensity fluctuations (scintillation).

r0
typical (visual) r0 ~ 10 cm ~ / r0 ~ 1 arcsec
Munich Dec. 1-3 2003 Global Science Forum

for D > r0 images are limited by diffraction on r
Wavefront Control (P. Salinari) 14

r0, "coherence lenght", r0

6/5
0


Distribution of the "structure constant" of the atmospheric refraction index
Annual average profile (black) and typical situation (green) for the distribution of refraction index fluctuations in the atmosphere. Very frequently only a few turbulent layers are present: · a layer near ground · one or two at intermediate height (5-8 km) · one or two near the tropopause (~ 15 km)
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 15

r0







0

CN( z ) dz
2

-3/5


Correcting wavefront errors: Adaptive Optics

Only a special case has been implemented until now: "single conjugate AO" but the multiconjugate systems will arrive soon
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari)

Three basic ingredients: · deformable mirrors · wavefront sensors · control system
16


A short history of WfC (in astronomical telescopes)
From - to 1988 · Number of controlled parameters: · Control bandwidth: · Spatial scales (modes) controlled: From ·N ~ ·B ·S ~ N3 B 1 Hz (human brain) S = D, (rigid modes: tip, tilt, rot.)
best image sizee best image siz ~ 11arcsec ~ arcsec

1989 (ESO NTT introduces "active optics") to present 100 best image sizee best image siz 1 Hz ~ 1/2 arcsec ~ 1/2 arcsec 1m

1995-2003 Single Conjugate AO is added to most large telescopes · N ~ 50-300 best image sizee · B ~ 100 Hz best image siz ·S ~ 1m ~ 1/20 arcsec ~ 1/20 arcsec 2002: an important step: the first "adaptive telescope": MMT
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 17


The very near future of WfC
2004 Multiconjugate Adaptive Demonstrator for VLT (the first "wide field" AO system for the NIR) 2 conjugates 1-2.5 µm 1-2.5 µm 8 natural stars field ~ 22arcmin field ~ arcmin N ~ 500 sky coverage: modest sky coverage: modest B ~ 100 Hz ~ 2006 "Nirvana" visual-NIR , triple conjugate, "layer oriented" AO for LBT 3 conjugates 0.5-2.5 µm 0.5-2.5 µm 2x(12 + 12) natural stars field ~ 22arcmin field ~ arcmin N ~ 2800 sky coverage: >50% B ~ 200 Hz sky coverage: >50% ~ 2006 Gemini South NIR, triple conjugate, 5 Laser Guide guide star system 3 conjugates 1-2.5 µm 1-2.5 µm 3 natural stars + 5 Laser guide stars field ~ 22arcmin field ~ arcmin N ~ 1000 sky coverage: ~ total B ~ 100 Hz sky coverage: ~ total
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 18


Examples of AO systems
2006? Gemini MCAO A triple conjugate post focus system Using Laser Guide Stars For Near IR ~ 1000 DoF 7 (+3) reflections Gemini MCAO optics

2003 NAOS on VLT A single conjugate post focus system, For Near IR 256 DoF, 5 (+3) reflections
Munich Dec. 1-3 2003 Global Science Forum

2002 MMT Adaptive secondary AO: Single conjugate For near IR 336 DoF 0 (+2) reflections
19

Wavefront Control (P. Salinari)


What can we correct (in principle)? · · · Single conjugate: all layers in a single direction (narrow field, high Strehl Ratio) Multiconjugate: various layers (wider field, lower Strehl, correction of scintillation) Ground Layer: only the strongest layer (reduced "seeing", widest field)
Now I Iwill spend Now will spend aafew slides on few slides on the most the most challenging challenging application: application: The study of The study of terrestrial terrestrial exoplanets exoplanets
20

What is left? The correction will never be perfect. Where are the physical limits?

Munich Dec. 1-3 2003 Global Science Forum

Wavefront Control (P. Salinari)


Scattering of light by Residual Wavefront (phase) Error (RWE)
The RWE, i.e. the "leftovers" of the (phase) AO correction, scatters light around the star proportionally to the Phase Power Spectral Density of the total RWE (total includes the effects of "fitting", "phase lag", "photon noise", "aliasing", etc. errors). In other words: · The RWE at spatial wavelength W scatters light of wavelength at an angle = /W at ~ 0.1 arcsec, in V band W ~1 m, in K band W ~4 m, 1-4 m scales are critical · As with a given actuator separation we can correct the wavefront error only at W> 2 , there is always a non-corrected part of the RWE spectrum (W < 2 ), that produces (by aliasing) further contamination of the corrected part. · The correction must extend well beyond the spatial frequency of interest (W= / ). (In other words: << /2 ) · The scattered light intensity I at angle is proportional to the RWE phase variance 2(W) at the corresponding W. I() 2(W) = 2(/ )
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 21


AO halo shape (adapted from Jolissaint and Veran 2001)
Residual phase power spectrum After AO correction Before AO correction Not corrected by AO AO corrected PSF 20 m telescope =1.65 µm r0()=90 cm =90 cm Airy pattern, (next ennemy) Aliased error

Not corrected halo

Critical spatial frequency fc=1/Wc=1/2

Critical field angle c= /Wc= /2 The profile of the "halo" The profile of the "halo" is not Lorenzian. . is not Lorenzian

Munich Dec. 1-3 2003 Global Science Forum

Wavefront Control (P. Salinari)

22


100 m

Diffraction profiles
(R band, 700 nm, courtesy of A. Riccardi)

100 m

3x10

3

10 10-7 AO contrast

2

Black, no gaps

Red, 10mm gaps

Green, 23nm rms wf piston

Coronagraphy can remove (most of) the diffraction structure,

Differential piston needs adaptive segment piston control
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 23


Result: V band AO PSF (theoretical contrast)
Planet/star flux ratio and angular separation for known exoplanets compared with telescope PSF (Lardiere et Al. 2003) A) even a 100 m telescope cannot resolve some of the known planets from their stars. B) A one arcsec field radius includes most known planets C) Exo-earths, at best distance, are about three orders of magnitude below the scattered light background
Munich Dec. 1-3 2003 Global Science Forum

=10 cm

Wavefront Control (P. Salinari)

24


Stellar sample size (versus telescope size and location)
With a ~ 30 m telescope (at "Mauna Kea") one can explore at short wavelength the entire TPF (goal) sample of ~ 100 stars in 1000 hours With a 100 m telescope in "Antarctica" one can obtain R > 1000 spectra of the TPF sample at short wavelength (R to K) in 1000 hours
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari)

TPF~100

25


What are AO requirements for the study of terrestrial exoplanets? · · · · Actuator spacing Segmented correctors Appropriate pupil shape Multiconjugate system ~ 10 cm (differential piston control to a few nm) (control of diffraction pattern) (control of scintillation)

Is it physically possible? YES Is the technology available? Not yet but is not far Could a large telescope designed for the study of terrestrial exoplanets also fulfill the requirements of the other science goals? YES
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 26


What technology developments are needed? 1. Wavefront correctors: The request of segmented correctors means that the Number of DoF of current correctors is already adequate, they simply need to be closely packed. · A requirement already met by deformable telescope mirrors · A modest extra requirement for Piezo stack deformable mirrors 2. Wavefront sensors: Existing wavefront sensors can be used for all cases mentioned. They need larger format fast detectors (using existing technology) 3. Control loop: The segmentation of the correctors can be used for symplifying the algorithms of the control loop, avoiding divergence of the computational problem.
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 27


What is the state of the art for adaptive telescope mirrors?
2002 MMT Cassegrain secondary 336 actuators in operation ~2 years ~3 years Adaptive primary mirror segment = 10 cm (< 400 actuators) (currently in advanced study phase) ~ 2007 working prototype 2004 (in operation) LBT Gregorian secondary 762 actuators in construction

642mm 1600 mm
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari)

911 mm
28


A fully adaptive triple-conjugate telescope design
(optical design by A. Goncharov, Galway, Ireland) The appearence of adaptive telescope mirrors is the basis for a new generation of telescopes: Adaptive Telescopes These entirely relay on fast (adaptive) Wavefront Control for cancelling the "seeing" and for controlling optics and mechanics The control loop can be closed on natural stars (laser guide stars help extending sky coverage) 29

M2: 2.5 m, monolithic Used for tip-tilt correction M4: 2.5 m, monolithic adaptive conj. at 10 km M3: 4 m, monolithic adaptive conj. at 5 km M1: 40 m, parabolic, F/0.7 adaptive segments · Flat field ! · 2arcmin at SR>0.85 in V ! · up to 8 arcmin at < 50 mas rms !
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari)


What else is needed for a new generation of telescopes? We need to make the new technology cheaper than the old one, it has the potential of being cheaper but, understandably, at present is not The cost of actuators and electronics can be dramatically reduced by mass production Adaptive mirrors require: · a smaller quantity of less expensive glass · more relaxed optical fabrication tolerances we need to find the way of exploiting these advantages to reduce the mirror cost much below current cost
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 30


Summary of section 2 The evolution of the ideas, technology and experience on the problem of wavefront control in the last decade makes us nearly ready for diffraction limited visual-IR adaptive telescopes, which could match all science requirements including the most demanding ones The required on-field testing of the new concepts and the modest further technology development can be done in a few years

Munich Dec. 1-3 2003 Global Science Forum

Wavefront Control (P. Salinari)

31


Section 3: current approaches for the next generation of telescopes

Stretching "active" control Hybridizing "active" and "adaptive" control or Investing in the emerging, fully adaptive, technology?

Munich Dec. 1-3 2003 Global Science Forum

Wavefront Control (P. Salinari)

32


The 20-20 interferometer
(US, Steward Observatory study, a blown up LBT)

21 m Fast, F/0.8, primary (7x LBT primary) The 2.1 m adaptive secondary is made by 7(LBT secondary) segments

Priority on IR interferometry (exoplanets) ·Primary: 7 large active segments ·Adaptive optics: -Segmented Adaptive Secondary -Further TBD correctors in beam combiner
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari)

A 30 m single telescope based on the same approach is under study

33


Thirthy Meter Telescope (TMT) (AURA NIO collaboration of Caltech, UoC, Canada . . .)
Funding for design in progress (17.5 M$ donation + NSF application)

Priority on time: operative in 10 years ·Primary: ~ 900 active segments ·Adaptive Optics: Post-focus (baseline) Adaptive secondary (as an option)
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 34


Euro50 (study of a 50 m by an european consortium lead by Lund University)

Priority on size and AO ·Primary: ~1000 active (a-spherical) segments ·Adaptive optics: - two conjugates, LGSs - Monolithic adaptive secondary (~ 4 m) - Post-focus mirrors
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 35


OWL
(European Southern Observatory, 100 m telescope study) A six mirror, spherical primary design with a 30 m flat segmented secondary that could become a high order corrector all other mirrors are < 8 m >1 arcmin high SR at V > 10 arcmin seeing limited field

30 m Priority on size (and cost containment) ·Primary: >3000 active (spherical) segments ·Adaptive optics: -Post-focus - At mirrors 4,5,6
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari)

100 m

36


The European Large Telescope
(study proposal for FP6 funding) Priority in Europe seems to be is on size and time (but, as usual, with a variety of opinions) Science priorities are those of slide 4 (in reverse order)

No Pictures, is a technology study!

The goal is to define/develop the technology for a large European Telescope (open to the collaboration of non Europeans) within the current FP6 programme (2007). The proposal is in its formation stage, but will include: · Primary mirror segment technologies (spheric and a-spheric) · Adaptive telescope mirrors and post focus correctors · Differential piston sensitive wavefront sensors · Active structure and mirror control · Site
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 37


Summary of section 3 The pressure for a new generation of telescopes is building up All existing studies give very high weigth to high angular resolution and are open to the use of new technology often the driving priority is time while a few years can make the difference in performances and, probably, cost

Munich Dec. 1-3 2003 Global Science Forum

Wavefront Control (P. Salinari)

38


Perspectives 1 The perspective for the next generation of ground-based visual-IR telescopes is splendid: in spite of the strong competition from space (and else) the ideas and the technology developed in recent years (and still in progress) offer the perspective of maintaining Visual-IR telescopes on the forefront of astrophysical research in various fields for several decades In particular we can already identify the technology for a new and challenging field: the detailed study of terrestrial exoplanets (the Higg's boson of astronomy: a well defined goal)
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 39


Perspectives 2 If "terrestrial exoplanets" is the strategic goal, which is the shortest way to it? I believe we need to approach the problem in two steps: 1. An intermediate generation (i.e. a multeplicity) of medium size (30 ±10 m) telescopes equipped with (different) advanced AO systems. This phase is the one that is already starting. It can produce important progress in different fields (including that of exoplanets) and is necessary for the final validation of the advanced AO technology 2. The Big Ones (1 or 2 supertelescopes, in the 50-100 m range) for exo-earths detailed study (and much more, in other fields)
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 40


Perspectives 3 What is needed in the near future (say next 5 years)? · Strong support of AO technology development and of AO testing on existing 8-10 m class telescopes · Exploration of approaches to reduce the cost of the most advanced technology for "step 1", in particular of highly aspheric segments (thick and thin, concave and convex) · Identification (preservation) of excellent mid latitude telescope sites · Carachterization of Antarctica as a Visual-IR astronomical site. Is it really a quasi-space on earth? If yes, we must start learning how to use it!
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 41


Perspectives 4 What could be obtained from the "intermediate generation" (say by 2015-20)? · Further major progress in the understanding of many science fields · Better definition of the science goals of the "Big Ones" · On field comparison of different technical solutions · A strong, convinced motivation for the next step really big telescopes are likely to be really expensive (unaffordable in the the current local-regional funding schemes) They will require a world-wide collaboration.
Munich Dec. 1-3 2003 Global Science Forum Wavefront Control (P. Salinari) 42