Adaptive optics and its applications - NISCAIRnopr.niscair.res.in/bitstream/123456789/30645/1/IJPAP 43(6) 399-41… · Keywords: Adaptive optics, Wavefront sensor, Deformable and

red with paraeen para and

IlS should be inide two copieserent names.f the correctedl if floppy gets

data should be: except Bold,lis (a ~ X 8 e <I>

used should be

and above is

y where. e.g.11 A K, Indian J~.I for underlinedlpl Phys, shcllc!,& Appl P ny.;

Iy TAB and not

lB only, not by

ning columns or

e entered In

Indian Journal of Pure & Applied PhysicsVol. 43, June 2005, pp. 399-414

Adaptive optics and its applicationsD Mohan, S K Mishra & S K Saha*

Instruments Research & Development Establishment, Dehra Dun, India

*Indian Institute of Astrophysics, Bangalore, India

[E-mail: [email protected]]

Received 4 March 2005; accepted 1 April 2005

The adaptive optics (AO) system combines technologies that enable corrections in real time for the deleterious effects ofthe atmosphere allowing terrestrial telescopes to achieve their near diffraction-limit. Such a system introduces controllablecounter wavefront distortions, which both spatially and temporally follow that of the atmosphere. AO system has advantagesover post-detection image restoration techniques that are limited by noise and recovers near diffraction-limited images andimproves the point source sensitivity. This system may become standard tool for the new generation large optical imagingsystems. Development of such a technique was made possible due to the significant improvements in technologicalinnovation over the past few decades. After a brief introduction on the deleterious effects of the atmosphere on the image,this article discusses in detail about required components in order to develop the AO system and its applications in variousfields, particularly in observational astronomy.

Keywords: Adaptive optics, Wavefront sensor, Deformable and Tip-tilt mirrors, Telescope, Turbulence

IPC Code: GO 1M I 1100

1 IntroductionThe twinkling of stars in the night sky has been

viewed with awe and admiration by the mankind fromtimes immemorial. Ever since the invention oftelescope, the phenomena of atmospheric turbulence,which cause the twinkling of stars, has frustrated theastronomers who need to look into deep space tounravel the secrets of the universe. The turbulence isbecause the various layers of the atmosphere havedifferent refractive indices depending on thetemperature and pressure at that height. In suchsituation the resolution of conventionalastrophotography with large telescopes is not limitedby the diffraction on the finite aperture, but with thesize of the atmospheric turbulence cell. Typically thesize of a tubulence bubble is 10 cm. Diffraction on thebubbles causes an angular blur of :::::)jD = 500 nm/lOem :::::an arc second, in which A is the wavelength ofthe observation. This is the maximum resolution thatcan be obtained irrespective of aperture size of thetelescope. The higher resolution can be achievedinspite of the image degradation by such cells eitherby using a passive method called speckle imagingtechniques I or by using a direct method known asadaptive optics. An alternative arrangement, however,was made by deploying Hubble space telescope above

the atmosphere in order to make sharp images, but itssize is small compared to a dozen 4-10 metre classtelescopes operating on the ground. Consistent effortsof scientists, spread over. two decades, have led to thedevelopment of the AO technology for compensatingatmospheric turbulence effects by introducing anoptical imaging system in the light path, which sensesthe perturbations and compensates in real time.

2 Diffraction Limited ResolutionAn idealized astrophysical source of

monochromatic radiation, propagating in the absenceof the atmosphere, is known as plane wave havinguniform magnitude and phase across the telescopeaperture. All telescopes have inherent limitation totheir angular resolution due to the diffraction of lightat the telescope's aperture. When a continuum of wavecomponents pass through an aperture, thesuperposition of these components result in a patternof constructive and destructive interference. Forastronomical instruments, the incoming light isapproximately a plane wave since the source of thelight is so far away. In this far-field limit, Fraunhoferdiffraction occurs and the pattern projected onto thefocal plane of the telescope will have littleresemblance to the aperture.

400 INDIAN J PURE & APPL PHYS, VOL 43, JUNE 2005

Let us examine how does the diffraction pattern ofa star look like and what is the limit of a telescope'sresolution. For wavefunction \{I, at point P, in theimaging plane, the intensity is give bl,

... (1)

and

u =ka8 ... (2)

where k = 2nl'A is the wave number, a is aperture size,and e is the angle between telescope axis and rangefor the point P.

This intensity pattern of constructi ve anddestructive interference rings is known as the Airydiffraction pattern (see Figure 1), 84% of the totalintensity is located within the central circle or the

. Airy disk. The dark destructive interference ringsoccur at the minimum of JI(u), where u = 3.83, 7.02 ...or 8= 0.61A1a, 1.12A1a, ....

The limit for the telescope's resolution is set by thediffraction at the aperture of the telescope. For a pointsource, like a star, the resulting image is a Airypattern. The Rayleigh criterion for resolution of twopoint sources is that the central maximum of oneimages lies at the first minimum of the second image.Thus the limit of the angular resolution is,

~6= 1.22A I D ... (3)

The Airy disk has angular radius ~e,so the radiusof the central disk is,

p = )"fiDin which f is the focal length.

... (4)

0.9

.~ 0.8

~ 0.7Q)

E 0;6"0Q) 0.5

.~cu 004

E0 0.3Z

0.2

0.1

ci-10 -5 a 5

Scaled u value,rad

Fig. 1 - Airy diffraction pattern

The point spread function(PSF) of the telescope isthe modulus square of the Fourier transform of theaperture function. The resolution at the image plane ofthe telescope is determined by the width of the PSF.

3 Atmospheric Turbulence. Diffraction limited resolution is an ideal condition

for the image which never happens in real life mainlydue to atmospheric turbulence that is considered to behighly turbulent medium. Heating of the Earths'satmosphere by solar radiation causes turbulentmotions. According to the Kolmogrov's theory offluid turbulence.', when the kinetic energy of the airmotions at a given length-scale is larger than theenergy dissipated as heat by viscosity of the air at thesame scale the kinetic energy of large scale motionswould be transferred to smaller and smaller motions;motions at small scales would be statisticallyisotropic; at the small scales, viscous dissipationwould dominate the break-up process. During daytime, large warm packets of air closer to the groundmove up due to buoyancy and initiate convectioncausing the turbulence near the ground. Theydissipate their kinetic energy continuously andrandomly into smaller and smaller packets of air, eachhaving unique temperature. These packets are callededdies. Convection changes with isolation anddisappears during night time. However, horizontalcirculation of air starts. Kolmogrov law represents thedistribution of the turbule sizes, from millimeters tometers, with lifetimes varying from milliseconds toseconds. An important property of eddies is that theyexist in a variety of length-scales and theirdistribution is random. There exists an upperlimit, Lo, decided by the process that generatesturbulence and a lower limit, lo- decided by the size atwhich viscous dissipation overtakes the break-upprocess .

The refractive index values fluctuate with time dueto fluctuations in temperature and pressure and arerandom in nature. The refractive index fluctuationsfollow a power law with large eddies havinggreater power". To a first order approximation, therefractive index of air is related to the pressure andtemperature as",

pn = 7.9 X 10-2 T + 1 ... (5)

1o aouter-srefract

10

where P and T are given in units of atmosphere andkelvin respectively.

Fmterms,a cell (

OPD(

for vispressu

Asdegreeperturlhundr,tempe]wind,the atigive rihence,of thecumu!.inhornatmosjgradieproducimagecan 1(scintieffect

3.I'StalAtn

and tbone al:meantelesccessentitempe]differecurrenstructubetweedistant

where

~telescope issform of thenage plane ofof the PSF.

eal conditionil life mainlysidered to bethe Earths's~s turbulent's theory ofgy of the airger than thethe air at thecale motionslIer motions;

statistically: dissipationDuring day

) the ground: convectionound. Theyuously ands of air, eachts are calledclarion and., horizontal.presents theillimeters toliseconds to: is that they

and theiran upper

It generatesv the size atIe break-up

ith time dueure and arefluctuationslies havingimation, theIressure and

... (5)

osphere and

MOHAN et al.: ADAPTIVE OPTICS AND ITS APPLICATIONS

From Eq. (5), the phase variations, expressed interms of the optical path difference (OPD),caused bya cell of length L, would be:

pOPD(waves)=7.9x--

2LI:!.T "" 2LI:!.T ... (6)

AT

for visible wavelengths at standard temperature andpressure.

A small temperature fluctuation of one tenth of adegree would thus generate strong wavefrontperturbations over a propagation distance of a fewhundred metres. Naturally occurring variations intemperature «1°C) cause random changes in thewind velocity, which we view as turbulent motion inthe atmosphere. Further, the changes in temperaturegive rise to small changes in atmospheric density, andhence, to the refractive index. This index changes isof the order of 10-6 and can accumulate. Thecumulative effect can cause significantinhomogeneities in the index profile of theatmosphere. Environment parameters, viz., thermalgradients, humidity fluctuations and wind shearproduce atmospheric turbulence. The wavefront of animage will change in the course of propagation. Thiscan lead to image jitter, intensity fluctuations(scintillations) and image spreading; in aggregate, theeffect will blur the image.

3.1'Statistical model "Atmosphere is a non-stationary random process

and the seeing conditions evolve with time, therefore,one also need to know the statistics of their evolution,mean value and standard deviation for a giventelescope. Fluctuations in the air refractive index areessentially proportional to fluctuations in the airtemperature. These are found at the interface betweendifferent air layers. Wind shears produce eddycurrents of various sizes in the atmosphere. The indexstructure function is the variance of the differencebetween the two values of the refractive index at adistance, p, apart" and is given by

o, (p)== (In(r)- n(r + p )12) == C; p2/3 ... (7)

where 10« p « Lo.

10 and Lo denote the respective inner-scale andouter-scale lengths of the eddies and C/ is therefractive index structure constant.

401

The value of the initial subrange, 10<«p« <La, inwhich p is the vector between the two points ofinterest, would be different at various locations at thesite.

Diffraction and refraction of the wavefront byatmospheric inhomogeneities cause the most severetime varying effects. The temporal structure functionis simply obtained by substituting Inl for p in Eq. (7),in which v is the velocity of the wind and 't is timeconstant. The refractive index is fairly wavelengthindependent from red to thermal infrared, therefore,the above quantity is wavelength independent.

3.2 Turbulence and wind profile modelsThe spatial and temporal characteristics of

wavefront distortions convey essential information.However, the numerical evaluation of the criticalparameters requires the knowledge of the Cn

2 andwind profiles as a function of altitude. Refractiveindex structure constant Cn

2 is a measure of thestrength of the turbulence. It is not constant but keepsvarying with seasons, daily and hourly. It also varieswith the geographic location and with altitude and canbe perturbed very easily by artificial means such asaircraft. Its unit is m-213• Since most of the aboveparameters are directly or indirectly related with C/,therefore, for a particular optical path it needs to bemodeled.

The most widely preferred two models" are:(i) SLC-Day model and (ii) Hufnagel-Valley model.

This turbulence model has two free parameters, Aand W. The first is normally set to 1.7 x 10-14m-213,while the second, which represents the average windspeed, can be adjusted to achieve the desired low-altitude shape. The most common value for W is21m1s, which yields an expression referred as the HV-21 model. C/ variations with altitude are shown inthe Figure (2a).

SLC-Day turbulence model is described as,Cn

2(h) = 0 : 0 m <h-cl Om=4.008xlO-13h-1.504 : 19m<h<230m=1.300xlO-15 : 230m<h<850m=6.352xlO-7 h-2.966 : 850m<h<7,000m=6.209xlO-16h-0.6229: 7,000m<h<20,000m ... (8)

The Hufnagel-Valley turbulence model is describedas,

C/(h) = 5.94xlO-53(WI27)2h1oexp(-hl1000)+2.7xlO-16exp(-hl1500)+A exp(-hl100) ... (9)

402 INDIAN J PURE & APPL PHYS, VOL 43, JUNE 2005

16,nr----~----------~--------~Goo. HV21 I ( ).

-.- SLC Day a

12

1~)~·":----'O~·"---'O"::·"-=::::='O~'''---'0••.·,,--..J10·UCn2 (m'2J3)

20

18~16 -14

E 12C.,

10"t:l2..,

8<:6

:/I0

5 10 15 20 25

(b)

30 35

Wind velocity (rn/s)

Fig. 2 - (a) Variations of C,?, (b) Wind velocity profile

The wind velocity profile most often applied toturbulence problems is the Burfton model. Thisfunction is described by Eq. (10) and illustrated inFigure 2(b).

The Burfton wind velocity model is expressed as,

v(h) = Vg + 30 exp { - [h - 9400)l4800]2} ... (10)

where Vg is ground wind speed parameter, usually

Vg = 5 m/s.

4 Critical Atmospheric ParametersThe perturbations in the wavefront produce effects

similar to optical aberrations in the telescope and thusdegrade the image quality. When a very smallaperture is used, a small portion of the wavefront isintercepted and the phase of the wavefront is uniformover the aperture. If the amplitudes of the small scalescorrugation of the wavefront are much smaller thanthe wavelength of the light, the instantaneous imageof a star is sharp and resembles the classicaldiffraction pattern (see Figure 1). But as the windmoves the eddies past the aperture, the tilt of the

intercepted wavefront changes. This change in tiltcauses random motion of the stars image at the focalplane. As the aperture size increases, there is decreasein the sharpness and amplitude of the motion. When alarge aperture is used, the amplitude of the randomvariation of phase across the intercepted wavefront islarger.r This leads to the blurring of the image. Theimage motion and blurring together are 'referred to asatmospheric seeing. Knowledge of the followingatmospheric parameters is essential to any systemdesigner and needs to be characterized accurately. Allthese parameters 7 are very much dependent on theturbulence strength constant (Cn

2) and the wind

velocity.

4.1 Atmospheric coherence lengthAtmospheric coherence length", widely known as

Fried's parameter, r» is defined as the transverseaperture through which the beam can be transmittedwith optical phase distortion mean square valuewithin 1 rad", It is a function of the turbulencestrength constant C/, wavelength of the light,altitude, slant angle of beam direction and of course,path length. Seeing angle through atmosphere isinversely proportional to the Fried's parameter and isgiven by,

ro =3.0(C~Lk2tI5 ... (11)

It is important to note that the quality of seeing ischaracterised by Fried's parameter, i.e.,

8, = O.976}.,/ ro .. .(12)

14

8

in which Os is the seeing disk that determines theimage quality.

The seeing disk is defined as the full width at halfmaximum (FWHM) of a Gaussian function fitted to ahistogram of image position in arc sec. The seeingfluctuates on all time scales down to minutes andseconds. At a given site, r» varies dramatically nightto night. It can be a factor 2 better than the median orvice versa. Figure 3 depicts the night time variationsof r» on 28-29 March 1991 at the Cassegrain focus ofthe 2.34m Vainu Bappu Telescope (VBP). It is foundthat average observed ro is higher during the later partof the night than the earlier part.

4.2 Atmospheric time constantChanges in the refractive index in different

portions of the aperture result to the phase changes inthe value of the aperture function. Turbulence can be

Eo

i 12Q5E~mc.

_(J) 10\JQ)·clJ..

Fig. 3-Kavaluncurve is

thought j

the windspatial ffunction)temporaldependenpoint spnatmosphefrequencythe transvgiven by:

'Fhe' irrexposure 1

short-expoatmospheraberration:long-expos

4.3 IsoplanaiTurbule

statisticall;therefore i

viewing diJatmospheridirection:atmospherjangular off

The staisoplanatic

mge in tiltat the focalis decreaseon. When athe random/avefront isimage. Theferred to as

followingany systemurately. Alllent on the

the wind

'I known astransverse

transmittedluare valueturbulencethe light,

d of course,aosphere isneter and is

... (11)

of seeing is

... (12)

errnines the

vidth at halfin fitted to aThe seeing

minutes andtically nightie median orie variationsrain focus of). It is foundthe later part

in different,e changes inlence can be

MOHAN et al.: ADAPTIVE OPTICS AND ITS APPLICATIONS 403

141 I__ _ Uncorrected__ Zenith corrected

Eo

i 12-Q)

E~!1la..en 10"0.~•..u..

816 18 20 22

uTFig. 3 - Night time variations of ro at the 2.34 m VBT site,Kavalure, India, on 28-29 March, 1991(Ref. 9). The solid linecurve is for the zenith distance corrected value, while the

dotted curve is for the uncorrected value

thought as fixed phase screens, which are driven bythe wind in front of the telescope. By measuring thespatial properties of the phase screens (structurefunction) and the wind velocity, we calculate thetemporal behaviour of the perturbations. The timedependence of the aperture function indicates that thepoint spread function (PSF) is time dependent. Theatmospheric time constant also known as Greenwoodfrequency", is a function of the Fried parameter andthe transverse component of the wind velocity and isgiven by:

rTO =0.31-0-

vThe images of astronomical objects taken with

exposure time equal to or shorter than 'fo, are calledshort-exposure images. They correspond to stationaryatmospheric conditions. At longer exposure> 'fo, theaberrations are averaged, and for exposures » 'fo thelong-exposure PSF is obtained.

" ... (13)

4.3 Isoplanatic angleTurbulence and its structure function are

statistically the same everywhere in the long-exposuretherefore atmospheric PSF is independent of theviewing direction (isoplanatic). But the instantaneousatmospheric phase aberrations do depend on thedirection: telescope beam as projected on theatmospheric layer at 10 km shifts by 0.5 m for anangular offset of 10 arc seconds.

The standard definition of the atmosphericisoplanatic angle is a cone within which the beam

does not possess different phase variances due toturbulence and is expressed as

eo = 0.43 roL

... (14)

24

This parameter limits the distance between guidestar and the scientific objects. It turns out that formost objects there is no suitable (bright and close-by)guide star, hence artificial laser guide stars arerequired. Alternatively, a 3-dimensional correction ofturbulence is essential which leads to multi-conjugateadaptive optic technology.4.4 Image jitter

The first order tilt aberration due to turbulence isalso known as image jitter or dancing and is given by:

(jJ = 1.83C;X1l6 L17/6 ... (15)

This is due to refractive transmission of the beamby eddies of sizes bigger than the beam diameter.When it is slow it is called drift and when frequencyis high it is called jitter. Frequency of wander isapproximately 1I10thof the Greenwood frequency.4.5 Scintillation

The temporal variation of higher order aberrationsdue to turbulence causes dynamic intensityfluctuations called scintillations and is derived as:

a = 2.01C6/5 XlI5 r8/5S 1! L ... (16)

This is due to refractive transmission of the beamby eddies of sizes of the order of ~(AL).4.6 Thermal blooming

This effect is represented by the blooming'distortion number NB and it is caused by the resonantabsorption of the high power laser energy withatmospheric molecules". This is due to the non-linearresponse of the atmosphere. There is a critical power,which can be transmitted through the atmosphere forwhich this effect does not occur. This effectessentially needs to be compensated during thepropagation of the high power laser beams throughthe atmosphere. AO system partially corrects thiseffect. For a zero wind case, thermal blooming occurs,since the lowest index of refraction occurs near thecenter of the beam, where the beam intensity is thehighest. This atmospheric negative lens causes thebeam to defocus. An important case occurs when thewind or an artificial disturbance produced by the windslewing, causes the beam to take on a characteristic

404 INDIAN J PURE & APPL PHYS, VOL 43, JUNE 2005

crescent shaped pattern. Wind slewing causes thebeam to take on a characteristic crescent shapedpattern.

S Adaptive OpticsFor long range and high resolution imaging it is

essential to compensate for atmospheric turbulence inreal time. Adaptive optics is the only technologyavailable currently for real-time correction ofatmosphere turbulence. Optical phase conjugation isthe method used for compensation of atmosphereinduced aberrations. It is a multi-disciplinary subjectand a late entry among the list of current technologies.In recent years, the technology and practice ofadaptive optics have become, if not common place, atleast well-known in the defence and astronomicalcommunities. The purpose of AO system is to (i)sense the wavefront perturbations, and (ii)

h . I' 1112compensate t em III rea time ' .Figure 4(a) depicts the plane wavefront that

generated at the laboratory with a laser sourceoffering zero volt to the tip-tilt mirror. While Figure4(b) depicts the wavefront tilt measured with samesource after applying 1 V by computer to the saidmirror':'. These images are grabbed by a CMOSimager based Shack-Hartman (SH) sensor". Theseplane and tilted wavefront (Figure 4 (a)) Planewavefront (b) Tilted wavefront recorded at thelaboratory (Courtesy: V. Chinnappan). resemble tothe wavefront arriving to a detector from a distant starbefore and after passing through the turbulence of theatmosphere respectively. The laboratory experimentshows only tilt as a major error, while in the case ofatmosphere, the wavefront tilts have complicatedcontours. Nevertheless, a reverse situation can becreated by employing the AO systems in order toimprove the throughput of the large telescope.

There is no single inventor of adaptive optics. Thetechnology has evolved over years due to contributionof numerous scientists and engineers. In 1953,Babock" proposed scheme to correct for the rapidlychanging atmospheric seeing effects.

Although his suggestions have been thoroughlyresearched by the US military, only since the mid-1980s Babock's ideas were developed for theastronomical use.

For unresolved sources, adaptive optics attempts toput as many photons in as small an image area aspossible, thus enhancing the image contrast againstthe sky background thereby improving the resolution,and allowing better interferometric imaging with

(a)

. ,,~. ".,

(b)

Fig. 4 - (a) Plane wavefront (b) tilted wavefront recordedat the laboratory

telescope array. For resolved sources, the improvedresolution extends imaging to fainter and morecomplex objects. 'Active' optics is a technique similarto adaptive optics, but its purpose is to correct forwavefront distortions caused by the relatively slowmechanical, thermal and optical effects in thetelescope itself. These corrections are made atfrequencies -1 Hz.

The main components of an AO system are:telescope, combination of flexible mirrors such as tiltmirror (TTM), deformable mirror (DM) whosesurface can be electronically controlled in real time tocreate a conjugate surface enabling to compensate thewavefront distortion, wavefront sensor (WFS),wavefront phase error computationI2

,16-18, and a lasersource to generate a artificial guide star (beacon). Atypical adaptive optics imaging system (AOIS) isillustrated in Figure 5.

Performance of such imaging system is close to thediffraction limit of the input aperture and can only beachieved in the limit of:

• Thprethe

• Th. D~col

• Amatirr

5.1 nle!Non

transmiopticallong rathe sigsignal.telescqbest 0

parabolneed tcmaintaistructuito suppbe cap,move aearth.5.2 Steer

Fastactiveapplicai

,

Jcorded

rnprovedid more.e similarrrect forely slow

in themade at

tern are:ch as tilt) whoseII time tonsate the

(WFS),d a laseraeon), A\.01S) is

ise to the1 only be

MOHAN et al.: ADAPTIVE OPTICS AND ITS APPLICATIONS 405

'-

Fig. 5 - Adaptive optic imaging system

• The angular separation between the turbulenceprobe and the target object should be smaller thanthe isoplanatic angle.

• The spacing between the control elements on the. DM should be well matched to the turbulencecoherence length.

• A sufficiently high update rate should bemaintained i.e. less than inverse of the coherencetime.

5.1TelescopeNormally Cassegrain type telescope used for

transmitting the beacon as well as receiving theoptical signal for the WFS and imaging camera. Forlong range AOIS, beacon laser is needed to improvethe signal-to-noise ratio (SNR) for the wavefrontsignal. There are two major components of thetelescope: primary mirror and secondary mirror. Thebest combination of the mirrors is primary asparabolic and secondary as hyperbolic. These opticsneed to be supported in some suitable structure tomaintain alignment with each other. The supportstructures are indeed an engineering issue. In additionto supporting the optics, the structure also required tobe capable of tracking astronomical objects as theymove across the sky because of the rotation of theearth.5.2Steering/tip-tilt mirrors

Fast steering mirrors (FSM) are effectively used inactive and adaptive optics for various dynamicapplications such as precision scanning, tracking,

pointing laser beam and image stabilization. FSM, asshown in Figure 6, is a mirror mounted to a flexuresupport system that may be tilted fast about its axisindependent of the natural frequency of thespring/mass system in order to direct an image in x-yplane.

In adaptive optics the fast steering mirror is used asone of the two main phas ~ correctors for beam orimage stabilization by correcting beam jitter andwander introduced by atmospheric turbulence as wellas thermal and mechanical vibrations of opticalcomponents. Steering mirrors with high bandwidthoperation can be electronically controlled to tiltaround two orthogonal axes (tip-tilt movements)independently. Beam wander is the first orderwavefront aberration that limits the beam stabilizationand pointing accuracy onto the distant targets. Two,three and four actuator based steering mirrors aregenerally designed to cater for the dynamicapplication in mind with appropriate dynamic range,tilt resolution and frequency bandwidth. The simplestdesign is the two-axis tilting mirror with two PZTactuator stacks pushing the tilt platform/mirrorsubstrate at 90 degrees around the central pivot.However, the two-axis tilt mirror suffers from thethermal instabilities and cross-talk between the tiltingaxes at high frequencies.

5.3 Deformable mirrorsAO system requires novel devices to implement the

phase shift operation necessary for wavefront control.

406 INDIAN J PURE & APPL PHYS, VOL 43, JUNE 2005

----------------,=----------Fig. 6 - Fast steering mirror

The phase of the wavefront can be controlled eitherby changing the propagation velocity or the opticalpath length. Refractive index varying devices such asSLMs and other ferroelectric or electro-optic crystaldevices have been used with limited success toimplement phase control. Frequency response andamplitude limitations have been limiting factor forthese devices. Reflective surface modifying devicessuch as segmented mirrors and continuous surfaceDMs are very successful in several high endapplications. In such Deformable mirrors, two kindsof piezo actuators are used namely as Stacked andBimorph actuators".5.3.1 Discrete Stacked Actuator deformable mirror

Stacked actuator DM contains a thin deformablefacesheet rrurror on a two-dimensional array ofelectrostrictive stacked actuators supported by ruggedbaseplate as shown in Figure 7. In some casesactuators are not produced individually, but rather amulti-layer wafer of piezo-ceramic is separated intoindividual actuators. When some voltage Vi is appliedto the i-th actuator, the shape of DM is described bythe influence function r, (x, y) multiplied by Vi. Itresembles a bell-shaped (or Gaussian) function forDMs with continuous face-sheet (there is, however,some cross-talk between the actuators, typically15%). When all actuators are driven, the shape of theDM is equal to

... (17)

IltiI '.'

I

MirrorSupport

tjlC~v

mountinit must tfixed ingroo",,:es.5.3.3 Micr

The nconsistsstretchedApplyinjresponsefigure, •conducti:parts (i)and (ii) tlmembranin Figureand condilayer of eReflectiv(technolog

GIMS Faeeshcct

~iElectrc-distonive ActuatorStack

~ t:J<,Baseplate

Fig. 7 - Discrete stacked actuator Deformable mirror

A multi-channel high-voltage amplifier must havea short response time, despite a high capacitive loadof DM electrodes. For high bandwidth applicationssuch DMs are preferred and further it could be easilycooled. Both zonal and modal wavefrontreconstruction techniques can be applied with stackedactuator DMs.5.3.2 Bititorph deformable mirror (BDM)

A bimorph mirror is made from two thin layers ofmaterials bonded together. One layer is a piezo-electric material such as PZT and the other is theoptical surface, made from glass, Mo or Si or bothpieces may be PZT material, with the outer surfacebetween the two layers and acts as a commonelectrode. When a voltage is applied to an electrode,one layer contracts and the opposite layer expands,which produces a local bending much like that of abimetal strip. The local curvature being proportionalto voltage, these DMs are called curvature mirrors.The PZT electrodes need not be contiguous. Thegeometry of electrodes in BDM as shown in Figure8is radial-circular, to match best the circular telescopeapertures with central obscuration. In this way, for agiven number of electrodes (i.e. a given number ofcontrolled parameters) BDMs reach the highestdegree of turbulence compensation, better thansegmented DMs. Mirror surface Poisson solution ofthe applied voltage at a particular point. BDM verywell suits with the curvature type wavefront sensor.Modal wavefront reconstructor is preferred withBDM control. There is no such simple thing asinfluence functions for bimorph DMs. The surfaceshape as a function of applied voltages must be' foundfrom a solution of Poisson equation which describesdeformation of a thin plate under a force applied to it.The boundary conditions must be specified as well tosolve this equation. In fact, these DMs are madelarger than the beam size, and an outer ring ofelectrodes is used to define the boundary conditions-slopes at the beam periphery. The mechanical

MirrorSupport

c~

e mirror

r must haveacitive loadapplicationsild be easily

wavefrontvith stacked

iin layers ofis a piezo-other is the. Si or bothuter surfacea commonn electrode,'er expands,ke that of aproportionalure mirrors..guous. TheI in Figure 8ar telescopes way, for al number ofthe highestbetter thansolution ofBDM very

ront sensor.ferred withtie thing asThe surfaceust be foundch describesapplied to it.ed as well tos are madeIter ring of, conditions-mechanical

MOHAN et al.: ADAPTIVE OPTICS AND ITS APPLICATIONS 407

Fig. 8 - Actuator distribution in bimorph DM

Control electrodes

AI-co,.''1tOO membrane

SubstratePCB

Controlvoltages

Fig. 9 - Actuator distribution of micro-machined DM

mounting of a bimorph DM is delicate: on one hand,it must be left to deform, on the other hand-it must befixed in the optical system. Typically, 3 V-shapedgrooves at the edges are used.5.3.3 Micro-machined DM

The micro-machined deformable mirror (MMDM)consists of a thin flexible reflective membranestretched over an array of electrostatic actuators.Applying voltages to these actuators, individualresponses superimpose to form the necessary opticalfigure, can locally deflect electrically grainedconductive membrane. The mirror consists of twoparts (i) the die with the flexible mirror membraneand (ii) the actuator structure. A low stressed nitridemembrane forms the active part of MMDM as shownin Figure 9. In order to make the membrane reflectiveand conductive, the etched side is coated with the thinlayer of evaporated metal, usually aluminium or gold.Reflective membranes, fabricated with thistechnology have a good optical quality. Assembly of

the reflective membrane with the actuator structureshould ensure a good uniformity of the air gap so thatno additional stress or deformations are transmittedonto the mirror chip. All components of a MMDMexcept the reflective membrane can be implementedusing PCB technology. Hexagonal actuators areconnected to conducting tracks on the back side of thePCB by means of vias (metalized holes). These holesreduce the air damping, extending the linear range ofthe frequency response of a micro machined mirror toat least 1kHz, which is much better than for similardevices mounted over plane silicon dies.

Influence function is primarily determined by therelative stiffness of actuators and face sheet. Ifactuators are very stiff compared to the face sheet, theclamped-clamped is more apptopriated and vice-versaif the relative stiffness are reversed. Stiffer actuatorstructures reduce inter-actuator coupling but requirehigh central voltages. A more practical approach is toreduce the stiffness of the face sheet material by

408 INDIAN J PURE & APPL PHYS, VOL 43, JUNE 2005

reducing its thickness and/or elastic modulus and byincreasing the inter actuator spacing. Figure 10depicts the test patterns of MMDM at differentconditions. In order to estimate the influence functionof the deformable mirror actuators, a VeecoInterferometer has been used. During testing andcharacterization of DM, voltages are applied to singleand multiple actuators of the deformable mirror.Interferrograms for following cases were recorded(see Figure 10):

(a) absence of any voltage to its actuators, showsastigmatism shape; (b) all actuators are applied withequal voltage, shows spherical shape; (c) one of theadjacent actuators to the central actuator isapplied voltage, shows comatic shape; (d) centralactuator is applied voltage, shows defocus shape.5.3.4 Liquid crystal DM

Wavefront correction in AO is generally achievedby keeping the refractive index constant and tuningthe actual path length with a mirror. An opticallyequivalent alternative is to fix the actual path lengthand tune the refractive index. This could be achievedusing many different optical materials; a particularlyconvenient class of which is liquid crystals becausethey can be made into closely packed arrays of pixelswhich may be controlled with low voltages.

Electrically addressed NLCs are generally used forthe wavefront correction in conventional AO system,whereas optically addressed SLMs are also being usedto develop an unconventional AO with all opticalcorrection schemes. NLC are having lower framerates so that it is not so appropriate for theatmospheric compensation under strong turbulentconditions. Second type of LCs i.e., ferroelectric LCsare optically addressed in which the wave plateswhose retardance is fixed but optical axis can beelectrically switched between two states. Phase onlymodulation with a retarder whose axis is switchable ismore complicated than with one whose retardance can

be varied. The simplest method involves sandwichinga FLC whose retardance is half a wave in betweentwo fixed quarter wave plates. FLCs have theadvantage that they can be switched at kHz framerates, but the obvious disadvantage that they arebistable. The use of binary algorithm in WFcorrection is the simplest approach to develop closedloop control. The basic wave front correctionalgorithm is: whenever the WF error is greater than1...12,then correction of 1.../2is applied.

5.4 Multichannel DM driver electronicsElectronics for the actuator system are the most

complex, and by far the most expensive part of thesystem itself, typically accounting for the 2/3 of itscost. In an extreme example, the first 2000 channelmirror built had approximately 125 electroniccomponents per control channel just for the driver.These-drivers are incredibly safe but so complex as tobe unreliable. Main component of the single channeldriver electronics is high voltage operationalamplifier. Additionally, there are requirements of afeedback loop which limits the available current andshuts the driver down in case of the actuator fails orshort circuit. This prevents damaging the mirror bypower dissipation in the actuator. Also required is avoltage driver, frequently with an AID converter ontheO/p to provide the main system computer withmoment to moment information on the status of eachcorrector channel. Today analog inputs are generallyinsufficient since most wavefront controllers aredigital, so each channel has its own D/A convertersfor the input. The actuator is a low loss capacitorwhich must be charged and discharged at theoperating rate, typically up to 1 kHz. The requiredpeak current, fpeak is derived as

fpeak = 2nj CV ... (18)

In which C is the capacitance of the actuator and Vthe control voltage of the stroke.

(a) (b) (c) (d)

Fig. 10 - Test patterns of MM DM (a) Astigmatic (b) Spherical (c) Comatic (d) Defocus

Thus, tdeduced a

Ppeak = -h.Thus, e

peak ratinchannel isIn fact resdissipatiorassumptioiand 114of

5.5 Jitter serA po.

photoelecnspot into (manufactUlPSDs incharacterizthe manufawith laserscalibration,outstandinglinearity fesimple opeand Y posattached t(processesabsorption.

The Quawith twoFig. 11(a). Ithe spot sbbeing' biggTypically, tlacti ve sensithe exact mean output siilluminationthe output

I

Fig. 11 - (a

;andwichingin between

, have thekHz frame

at they areun in WFvelop closed

correction'greater than

ue the most! part of there 2/3 of its,000 channeli electronicr the dri ver.omplex as tongle channel

operationalrements of a~current anduator fails orhe mirror byrequired is aconverter onnnputer with.tatus of eachare general! ymtrollers areIA convertersoss capacitorirged at theThe required

... (18)

ictuator and V

MOHAN et al.: ADAPTIVE OPTICS AND ITS APPLICATIONS 409

Thus, the peak power consumption, P peak can bededuced as,

Ppeak = -/2. VmaJpeak ... (19)

Thus, each driver is a linear power amplifier withpeak rating of 1-10 W per channel. Certainly, everychannel is not operating at its full rating all the time.In fact reasonable thumb rule is that average powerdissipation in a driver package is obtained on theassumption that each channel is operating at 1/3 Vmax

and 114 of Imax.

5.5 Jitter sensor (Quadrant Detector)A position-sensing detector (PSD) is a

photoelectric device that converts an incident lightspot into continuous position data. Many industrialmanufacturers and laboratories around the world usePSDs in their daily work. PSDs are able tocharacterize lasers and align optical systems duringthe manufacturing process. When used in conjunctionwith lasers they can be used for industrial alignment,calibration, and analysis of machinery. It providesoutstanding resolution, fast response, excellentlinearity for a wide range of light intensities andsimple operating circuits. In order to measure the Xand Y positions from the PSD, four electrodes areattached to the detector and an algorithm thenprocesses the four currents generated by photoabsorption.

The Quadrant detector is a uniform disk of siliconwith two gaps across its surface as shown inFig. lI(a). For optimum performance and resolution,the spot size should be as small as possible, whilebeing bigger than the gap between the cells.Typically, the gap is of the order of 10-30 11mand theactive sensing area is 77 or 100 mrrr' (depending onthe exact model). When illuminated, the cells generatean output signal proportional to the magnitude of theillumination. It is the electronic card, which digitizesthe output signal, and the host computer then

......•..

{JDQ [Jj;r L.I]\][7, .OuodlGnf L0,,* •• , 1.0-.,.1 tUKIX [),o4 •• ~~r

Fig. 11 - (a) Quadrant detector. (b) Beam wanders relativeto the X or Y direction

processes the signal. The computer and softwareperform basic calculations of the position and powerof the monitored beam. Let A, B, C, D are the fourquadrants respectively, and R is the radius of theincident beam illuminating the detector. The beamposition is calculated using the following formulas:

x = (B+D)- (A+ C)A+B+C+Dand ... (20)

y= (A+B)-(C+D)A+B+C+D

where P (total power) = A+B+C+D.

The output position is displayed as a fractionalnumber or as a percentage figure, where thepercentage represents the fraction of beam movementrelative to the X or Y direction as shown in Fig. l Ib,

. 5.6 Wavefront sensorThe problem of measuring wave-front distortions is

common to optics (e.g. in the fabrication and controlof telescope mirrors), and typically is solved with thehelp of interferometers. Why do not use standard laserinterferometers in AO Wavefront Sensors20

-22

(WFSs)? First, an AO system must use the light ofstars passing through the turbulent atmosphere tomeasure the wavefronts, hence use incoherent (andsometimes non-point) sources. Even the laser guidestars are not coherent enough to work in typicalinterferometers. WFS must work on white-lightincoherent sources. Second, the interference fringesare chromatic. We cannot afford to filter the stellarlight, because we want to use faint stars. WFS mustuse the photons very efficiently. Third,interferometers have an intrinsic phase ambiguity of2n, whereas atmospheric phase distortions exceed 2n,typically. The WFS must be linear over the full rangeof atmospheric distortions. There are algorithms to"unwrap" the phase and to remove this ambiguity, butthey are slow, while atmospheric turbulence evolvesfast, on a millisecond time scale: WFS must be fast.

5.6.1 Shack Hartmann wavefront sensorIt is based upon the class+al Hartmann test. The

major advantage of this sensor is its high opticalefficiency, white light capability and operation withcontinuous or pulsed light sources. It has a highspatial and temporal resolution, large dynamic range

410 INDIAN J PURE & APPL PHYS, VOL 43, JUNE 2005

and no 2n ambiguities. All these reasons make ShackHartmann Sensor more preferred over other kind ofphase measuring sensors.

In the Shack Hartmann Wavefront Sensor (SHWS),the entire wavefront is divided into a large number ofsamples, called sub-apertures by a two dimensionallenslet array and forms an array of spots (Fig. 12). Arelay lens re-images these arrays of focal spots onto aHigh Frame rate CCD camera. Wavefrontmeasurement by SHWS is based on the measurementslocal slopes of a distorted wavefront a¢ / an relativeto a reference plane wavefront. The local slope isproportional to shift of the spot center L1S. Bymeasuring these small shifts, the local gradient of thewavefront is measured by the CCD camera, which isinterfaced to a fast transfer rate frame grabber foralmost real time data acquisition and data analysis. Byintegrating these measurements over the beamaperture, the wavefront or phase distribution of thebeam can be determined. In particular the space-beamwidth product, M2, can be obtained in singlemeasurement. The intensity and phase informationcan be used in concert with information about otherelements in the optical train to predict the beam size,shape, phase and other characteristics anywhere in theoptical train. Moreover, it also provides the magnitudeof various Zernike coefficients to quantify thedifferent wavefront aberrations prevailing in thewavefront. Fig. 9 shows schematically the SHWSprinciple.

The displacement of the spot center (x:' y: )within the sub-aperture with respect to a referenceposition (x~ ,y~) is measured. Local gradient of thewave-front <I>(x,y)is obtained according to

Lensleta Detector ImageWave-tront

Fig. 12 - Shack Hartmann WFS

... (21)

where S, = XC-x"S, = Yc-Yr,j is the focal length of thelenslets.

5.6.2 Curvature wavefront sensorThe curvature wavefront sensor'" measures the

intensity II in an intrafocal plane and the intensity 12in an extrafocal plane as shown in Fig. 13 andcompares these intensities to determine the curvatureof the wavefront. As the normalised difference,(11-12)/(11+12), is used for the comparison, and II and12 are measured simultaneously. The sensor is notsusceptible to the non-uniform illumination due toscintillation.

s = 1~'-I2 = F(F -I) {OW (x, y) s, _P(x, y)V2W(x, y)}II + 12 I an

... (22)

where first term with the curly bracket, is thewavefront derivative in the outward directionperpendicular to pupil edge, F the focal length of thetelescope and P(x,y) =1 within pupil and 0 outside.The type of detectors required for CS must havealmost zero read out noise and very short integrationtime. Either APD or specially optimised CCD/CMOSis used for this purpose.5.6.3 Pyramid WFS

Recently, a new kind of wavefront sensor has beenproposed which is able to change in a continuous waygain and sampling, thus enabling a better match of thesystem performances with the actual conditions on thesky. This sensor, named Pyramid Wavefront Sensor,is a novel concept device, as shown in Fig. 14 and itconsists of a four-faces optical glass pyramid thatbehaves like an image splitter. When the tip of thepyramid is placed in the focal plane of the telescopeand a reference star is directed on its tip, the beam of

Fig. 13 - Curvature wavefront sensor

lightlocatre-inof tlpyrasysteaboufromreco-5.7~

AIandany I

andThishysnimpctogeibandtempachie

TIwaveactuscalctotherchaninacexcelof cmiff(

zonaor seby \\I

apertthe vof rm

Lesome

... (21)

gth of the

sures thentensity 12~. 13 and~curvaturedifference,and II andisor is notion' due to

12W(X,Y)}

... (22)

ket, is the1 directionength of thei 0 outside.must have

t integration:::CD/CMOS

sor has beenitinuous waymatch of thelitions on thefront Sensor,<ig. 14 and itpyramid thathe tip of thethe telescope" the beam of

12

fisor

MOHAN et al.: ADAPTIVE OPTICS AND ITS APPLICATIONS

Fig. 14 - Pyramid wavefront sensor

light is then splitted in four parts. Using a relay lenslocated behind the pyramid, these four beams are thenre-imaged onto a CCD camera, obtaining four imagesof the telescope pupiL Since the four edges of thepyramid act like a knife-edge (or Foucault) testsystem, these images contain essential informationabout the optical aberrations introduced in the beamfrom the atmosphere. These parameters can be used torecover the astronomical images.5.7Wavefront reconstructor/controller

Adaptive optics controls the wavefront, temporallyand spatially both, in closed loop: the WFS measureany remaining deviations of the wavefront from idealand send the corresponding commands to the DM.This is why small imperfections of DM (likehysteresis or static aberrations) are not veryimportant: they will be corrected automatically,together with atmospheric aberrations. A highbandwidth specilised digital board implements thetemporal wavefront control whereas spatial control isachieved by the parallel computing.

This board calculates the aberrations from thewavefront-sensor measurements and generates theactuator commands of the deformable mirror. Thecalculation must be done fast (within 0.5 to 1 ms),otherwise the state of the atmosphere may havechanged rendering the wavefront correctioninaccurate. The required computing power needed canexceed several hundred million operations for each setof commands sent to a 250-actuator deformablemirror. Adaptive optic systems could be controlled byzonal or modal methods. In zonal control, each zoneor segment of the mirror is controlled independentlyby wavefront signals that are measured for the sub-aperture corresponding to that zone. In modal control,the wavefront js expressed as the linear combinationof modes that best fit the atmospheric perturbations.

Let x(t) be the input signal (e.g. a coefficient ofsome Zernike mode) and yet) - the signal applied to

411

x(j) l~m·1 am hFig. 15 - Measurement of error signal

the DM then the measured error signal by WFS asshown in Fig. 15 is

e(t)= x(t) - yet) ... (23)

The error signal must be filtered before applying itto DM, otherwise the servo system would be unstable.In the frequency domain this filter G(f) is called open-loop transfer function.

5.9 Reference sourceOne of the major problems applying adaptive

optics to astronomical observations is to locate abright reference object within isoplanatic patch, whichis required to measure the wavefront errors. Thesources are in most of the cases are too faint, hencetheir light is not sufficient for the correction. In orderto palliate the limitations of low sky coverage, analternative source is found in the form of introducingthe laser guide star source/" as a reference to measuresuch errors by means of a wavefront sensor, as well asto map the phase on the reference pupiL A pulsedlaser is used to cause a bright compact glow in theupper atmosphere, which serves as the source ofmeasuring the turbulence of the atmosphere.Concerning the flux backscattered by a laser shot, twobasic problems: (i) the cone effect which arises due tothe parallax between the remote astronomical sourceand artificial source (located 90 krn high in the caseNa 1D laser), and (ii) the angular anisoplanatic effectsare needed to be looked into.

5.10 Error budgetThe overall performance of an AO system is

estimated by the Strehl ratio, which is determined bythe residual wavefront" error.

S = expt-o,") ... (24)

where cr/ is residual wavefront varies over the pupiland to first order is made up of the contribution fromthe fitting error, temporal bandwidth limitation andWFS error. Hence,

2 2 2 2Or = crdm + crtemp +crwfs ... (25)

412 lNDIAN J PURE & APPL PHYS, VOL 43, JUNE 2005

The fitting error arises from the non-zero spacingbetween the DM actuators and is given by:

... (26)

where r, is the actuator spacing referenced to thetelescope entrance pupil and a is a factor related tothe geometry of the adaptive device. Its value is 0.4for a smooth modal influenced function.

The temporal error results from the limitedtemporal bandwidth of the control system:0temp2 = (fglf3db)5/3 •.. (27)

where fg is the Greenwood frequency and f3dbis the3db bandwidth of the AO control system.

Performance of an Adaptive Optic system isgreatly affected by the wavefront measurement errorof the Shack Hartmann wavefront sensor (SH-WFS).The error associated with the Shack-Hartmann WFSis mainly due to the limitation of the accuracy of thecentroid determinations for each sub aperture.Assuming the target is unresolved by each lensletwith the resulting spot near the center of a quad arrayof the detector it is given by

0wfs2 = 0.35(nI2SNR)2 ... (28)

where SNR is the radiometric signal to noise ratio ofthe WFS detector.5.11 Reduction of the data

Reduction of AO images may be processed withthe methods developed for high resolution imaging,speckle interferometry and other image processingalgorithms'<". Prior to use such algorithms, the basicoperations, namely, dead pixel removal, debiasing orflat fielding, sky or background emission subtraction,suppression of correlated noise, etc., are to beperformed. Characterization of AO PSF foranisoplanatic imaging is an important task; variationof point spread function (PSF) limits strongly thedeconvolution methods for processing of wide field ofview (FOV) images. Fusco et al.28 have derived ananalytical expression of this PSF degradation in theFOV and applied to a posterior processing of saidimages; according to them, the technique restored thestar parameters with a better precision.

6 Adaptive Secondary MirrorAnother way to correct the disturbance in real time

is usage of adaptive optics secondary mirror (ASM)that makes relay optics obsolete/". The other notableadvantages are: (i) enhanced photon throughput thatmeasures the proportion of the light which is

transmitted through an optical set-up, (ii) introductionof negligible extra IR emissivity, (iii) causes no extrapolarization, and (iv) non-addition of reflectivelosses": Due to the inter-actuator spacing, theresonant frequency of such a mirror may be lowerthan the AO bandwidth. The ASM system uses a SHsensor with a array of small lenslets, which adds twoextra reflective surfaces to the wavefront sensoroptical beam3]. An filS AO secondary with 336actuators is installed on the 6.5m Telescope of MultiMirror Telescope (MMT) observatory, Mt. Hopkins,Arizona.

7 Applications of Adaptive OpticsAdaptive optics systems are being used for the

military applications. The requirements of AOsystems for such applications are different from thatof astronomical AO systems. Those defence AOsystems requiring a large number of modes, a narrowoptical bandwidth, mostly monochromatic, and a hightime bandwidth, put less stringent requirement thansystems requiring wide optical bandwidth.

AO technology is being applied in medicine byophthalmologists. It allows for clearer examination ofeye's retina leading to an early diagnosis withprecision. Construction of fundus camera equippedwith such a technology allows one to image amicroscopic size of single cell in the living humanretina. Liang et al.32 have demonstrated that a humaneye with adaptive optics correction can resolve finegratings that are invisible under normal viewingconditions. Roorda and Williams33 were able toidentify the short, middle, and very long wavelengthsensitive cones in the living human eye by combiningretinal densitometry with the high resolution imagesavailable with adaptive optics.

AO system is being employed routinely at severallarge reflectors in order to gather new information.These are in the form of (i) imaging of the nucleus"ofM31, (ii) surveying of young stars and multiple starsysterns ", (iii) resolving the galactic center'",(iv) imaging of Seyfert galaxies, QSO host galaxies",and (v) mapping of the circumstellar environment'[etc. The discovery of compact cluster, R136a(HD38268) that was thought to be the most massivestar with a solar mass of - 2500M, of Doradus nebulain the Large Magellaic Clouds by means of speckleimaging technique." is an excellent achievement inthe field of high resolution imaging. In one of the IAUconferences, one full day was spent in discussingabout its probability of a black hole". The

obsetrevea12.8'

AlIII asmetemethcanaster:systeplaneradiathanto 1Cominterlearthhave

Fi:81C6.5 nbe twthe II0.1 afainrevery

Tbimag.of A(featuisun'sbrancfor s1

-starvrgraun

Fig. 16

InraIe1eerHvoor36ltiIS,

he~O.1at••.0:)w.ghIan

by.oflithoed: aoannan.ine'mg

to19thlingiges

era IIon..US34

starer36 ,es37 ,

t38n ,136a.sivebula.ckleit inIAUssingThe

MOHAN et al.: ADAPTIVE OPTICS AND ITS APPLlCA TIONS 413

observations with adaptive optics system" haverevealed over 500 stars within the field of view12.8" x 12.8".

Another most successful application of AO systemin astronomy has been in studying of the planetarymetereology'f, for example, Neptune has complexmethane clouds that can be discernible. Such a systemcan also be used for detecting moons aroundasteroids. Search for planets outside of our solarsystem is an exciting field in astronomy. The firstplanet around another star, 51 Peg was inferred fromradial velocity measurements of the primary. Morethan 80 such stars have been identified which appearto have Jupiter like mass planets in orbit.Combination of AO and very long baseline imaginginterferometer may in future be able to image theearth like planets. A few brown dwarf around starshave also been imaged using AO systems'r'.

Figure 16 depicts an example of the AO image of81 Ori B taken with adaptive secondary mirror at the6.5 meter MMT. Without AO, this object appears tobe two stars, but with AO turned on it is revealed thatthe lower star is a close binary having separated by0.1 arcsec; the brighter one is a guide star, and thefainter one slightly to the right (see white arrow) is avery faint companion.

The most fascinating use of AO systems is to getimage of the sun, particularly sunspots. Combinationof AO and image processing can pull up the sharpestfeatures to make more impressive insights into thesun's surface. AO systems are employed in otherbranches of physics as well. AO systems are usefulfor spectroscopic observations, as well as for photon-starved imaging with future very large telescopes, andground based long baseline optical interferomerers't".

Fig. 16 - Adaptive optics image of81 Ori B taken with adaptivesecondary mirror at the 6,5m MMT (Courtesy: L. Close)

8 Concluding RemarksEarth-bound astronomical observations are strongly

affected by turbulent air motions in the atmospherethat set severe limit to angular resolution; thedeployment of space-bound telescopes may providethe answer, but the size and the cost of such a ventureis its shortcoming. Adaptive optics technology has theability to improve the point source sensitivity; anexact knowledge of points spread function can bederived. It is going to be indispensable equipment tolarge telescope and will play larger role in thedevelopment of the very large telescopes of 30 mclass in future. The highest ever high angularresolution images from a single aperture, 10 m KeckII telescope, at near IR wave band have beenobtained. Using several deformable mirrors andseveral guide stars to compensate for turbulence in a3-D way, multi-conjugate AO system extends theimage compensation to large field of view (FOV).Ragazzoni et at. 46 have used this tomography andfound advantageous over classical AO approach. Atthe fall of the next decade (post 2010 A.D.),observations using AO system on new generationextremely large telescopes of 100 m class, willrevolutionize in mapping ultra-faint objects likeblazer that exhibits the most rapid and the largestamplitude variations of all AGN47

, exo-planets etc.High-frequency corrections at the secondary minorare the natural next step in the development of suchtelescopes to consider the throughput, polarization, IRemissivity, stray light, conjugation, reliability etc.

AcknowledgementWe express our gratitude to Dr L Close for

providing the AO image of e 1 Ori B. We also thankV Chinnappan for creating the images of planewavefront as well of wavefront tilt.

_ ReferencesI Labeyre A, Astron & Astrophys, 6 (1970) 85.

2 Hecht E, Optics (1987) 333.

3 Kolmogrov A, Turbulence, eds. S K Friedlander &L Topper, (1961) 151.

4 Tatarski V J, Wave propagation in turbulent medium,(Dover, New York) 1967.

5 Ishimaru A, Wave propagation and scattering in randommedia, (Academic Press, New York) 1978.

6 Parenti R, Sasiela R J, JOSA, 'A', 11, (1994), 288-309.

7 Tyson R K, Principles of Adaptive Optics, Academic Press,1998.

8 Fried D L, Limiting resolution looking down through theatmosphere, JOSA 56 (l0) 1966.

414 INDIAN J PURE & APPL PHYS, VOL 43, JUNE 2005

9 Saha S K & Chinnappan V, Bull Astron Sac Ind, 27 (1999) 30 Lee J, Bigelow B, Walker D, Doel A & Bingham R, Pub327. Astron Sac Pac, 112 (2000) 97.

10 Greenwood D P, J Opt Sac Am, 67 (1977) 390. 31 Lloyd-Hart M, Pub Astron Soc Pac, Il2 (2000) 264.11 Beckers J M, Annual Rev Astron & Astrophys, 31 (1993) 13. 32 Liamg J, Williams D R & Millar D T, J Opt Sac Am A, 1412 Roggetmann M C, Welesh B M & Fugate R Q, Rev. Modem (1997) 2884.

Phys, 69 (1997) 437. 33 Roorda A & Williams D R, Nature, 397 (1999) 520.13 Chinnappan V, Private communication, (2003). 34 Davidge T J, Rigaut F, Doyon R & Crampton D, Astron J,14 Shack R V & Plattt B C, J Opt Soc Am, 61 (1971) 656. 113 (1997) 2094.15 Babcock H W, Publ Astron Soc Pac, 65 (1953) 229. 35 Bouvier J, Rigaut F & Nadeau D, Astron & Astrophys, 32316 Saha S K, Bull Astron Soc Ind, 27 (1999) 443. (1997) 139.

17 Saha S K, Rev Modem Phys, 74 (2002), 551. 36 Davidge T J, Simons D A, Rigout R & Becklin E E, &

18 Roddier F (ed.), Adaptive Optics in Astronomy, (Cambridge Astron J J, 114 (1"997) 2586.

Univ. Press), 1999. 37 Hutchings J, Crampton D, Morris S & Steinbring E, Pub19 Sechaud M, Adaptive Optics in Astronomy, ed. F Roddier, Astron Sac Pac., 110 (1998) 374.

(Cambridge Univ, Press), 1999. 38 Roddier C, Roddier F, Northcott M J, Graves J E & Jim K,20 Hardy J W, Adaptive Optics-A progress review, Proc. SPIE Astrophys J, 463 (1996) 326.

1542 (1991) 229. 39 Weigelt G & Bair G, Astron & Astrophys, 150 (1985) LIS.21 Geary J, "Introduction to wavefront sensors", SPIE, (1995). 40 Cassinelli J, Mathis J & Savage B, Science, 212 (1998)22 Roussets G, Adaptive Optics in Astronomy, ed. F Roddier, 1'497.

(Cambridge Univ. Press) 1999.41 Brandle B, Sams B J, Bertoldi F, Eckart A, et al., Astrophys

23 Roddier F, Roddier C & Roddier N,Curvature sensing: a J, 466 (1996) 254.new wavefront sensing method, Proc. SPiE 976 , (1988)203.

42 Poulet F & Sicardy B, Bull Astr Am Soc, 28(1996), 1124.24 Esposito S & Riccardi A, Astron. & Astrophys, 369 (2001)

L9. 43 Close L, http://athene.as.arizona.eduJlclose/AOPRESS,

25 Roddier F, Adaptive Optics in Astronomy, ed. F Roddier,(2002).

(Cambridge Univ. Press,) 1999. 44 Labeyrie A, Astrophys J, 196 (1975), L71.

26 Foy R & Labeyrie A, Astron & Astrophys, 152 (1985) L29. 45 Mourard D, Bose I, Labeyrie A, Koechlin L & Saha S,

27 David Dayton et al., Opt Exp, 10, 25 (2002) 1508. Nature, 342 (189), 520.

28 Fusco T, Conan J, Mugnier L, Michau V & Rousset G, 46 Ragazzoni R, Marchetti E & valente G, Nature, 403 (2000),Astron & Astrophys Suppl, 142 (2000) 149. 54.

29 Bruns D, Barnett T & Sandler D, Proc SPlE, 2871 (1997) 47 Ulrich M, Maraschi L & Urry C, Annu Rev Astron &890. Astrophys, 35 (1997) 445.

ludicVol.

lInlHe

mealwallsrrnpoagricMorelayerOCCUI

envinas tbwiththrougranupowenucle:

DaJflowDarcyGenetbodie.throujDarcythis typracticlaw.rectilirformerperme:the po