Introduction 1

Conventional Adaptive Optics System

A conventional (linear) adaptive optics system,

whether it is used for imaging or whether it is used for
laser beam propagation, consists of three principal
subsystems: a wavefront sensor to detect the optical
disturbance, an active mirror or deformable mirror to
correct for the optical disturbance, and a control
computer to monitor and decode the sensor information
for the active mirror.

Telescope
Tip/tilt
Light Mirror
from
Object

Control Deformable Imaging

Computer Mirror Detector

Adaptive optics imaging system.

Telescope
Light from
Tip/tilt
Wavefront
Mirror
Beacon

Laser to
Target

Control Deformable
Laser
Computer Mirror

Wavefront
Sensor
Adaptive optics laser projection system.
2 Adaptive Optics

Image Spread with Atmospheric Turbulence

Linear systems theory shows how an image is composed of

an object convolved with the point spread function PSF
of the imaging system. Atmospheric turbulence
degrades the PSF and smears the image.

The PSF is the image of a point

source of light. The imaging process
experiences diffraction, and the
object is convolved with the PSF.
The resultant image is a blurred
version of the true object.

Adding aberrations to the optical

system results in a broadening of
the PSF and increased blurring.

Adaptive optics can compensate

for the aberrations and reduce
blurring.

What is adaptive optics? Answer

for the common man––Atmo-
spheric turbulence screws up the
image. Adaptive optics unscrews
it.
Introduction 3

The Principle of Phase Conjugation

All systems of adaptive Initial phase error

optics generally use the
principle of phase e-iφ
conjugation. An optical
beam is made up of both
an amplitude A and a Propagate to
phase φ component and is conjugate plane
described mathematically
by the electric field
Aexp(–iφ). Adaptive optics Amplitude and phase
reverses the phase to repeated possibly with some
provide compensation for linear magnification
the phase distortion. The e-iφΜ
reversal of the phase,
being in the exponent of
the electric field vector, e+iφΜ
means changing the sign
of the term behind the Adaptive optics applies
imaginary number. This the phase conjugate
mathematical conjugation
corresponds to phase
conjugation of the optical e+iφΜ− iφΜ = 1
field, just what is needed
to compensate for a
Resultant phase error
distorted phase. removed

While Horace Babcock is generally thought to be the

“inventor” of adaptive optics with his paper “The
possibility of compensating astronomical seeing,”
[Publ. Astron. Soc. Pac. 65, 229, (1953)] his exact idea
was never put into practice. It wasn’t until the
technological developments in electro-optics in the late
1960s and early 1970s that made a working adaptive
optics system possible.
4 Adaptive Optics

Point Spread Function

for an Astronomical Telescope

For an uncompensated astronomical telescope the point

spread function is limited by the diffraction of the
optics and the atmospheric turbulence. The PSF spot
has a central core width and an angular width proportional
to λ/D, where D is the telescope pupil diameter. A halo
surrounding the core has a width with an angular size of
roughly λ/r0, where r0 represents the strength of
atmospheric turbulence.

∼λ/D
~λ/D

∼λ/r
~λ/r00
Introduction 5

Fried’s Coherence Length

Fried’s coherence length is a widely used descriptor of

the level of atmospheric turbulence at a particular site.
For a fixed wavelength λ, astronomical seeing is given
by the angle λ/r0. For a known structure constant profile
[ Cn2 ( z) , where z is the altitude] and a flat-Earth
assumption, the coherence length is given by

−3 / 5
 
r0 = 0.423 k2 sec ζ ∫ Cn2 ( z)dz  ,
 Path 

where k = 2π/λ, ζ is the zenith angle (0 deg is straight

overhead), and the integral is over the path to the ground-
based telescope from the source of light.

Under turbulence, the resolution is limited by Fried’s

coherence length rather than the diameter of the telescope.
Since r0 ranges from under 5 cm with poor seeing to more
than 20 cm with good seeing, even in the best conditions, a
large diameter telescope without adaptive optics does
not provide any better resolution than a telescope with a
smaller diameter.
6 Adaptive Optics

Astronomical “Brightness”

The term brightness represents the brightness of an

object in the heavens. As the object such as a star is
observed, the amount of light (number of photons)
collected by an aperture (such as the human eye) per
second is astronomical brightness. The visual
magnitude mv of a star is a logarithmic measure of the
star’s brightness in the visible spectrum. Smaller numbers
represent brighter stars; negative numbers represent even
brighter stars. One expression that accounts for
atmospheric absorption relates visual magnitude to
brightness:
− mv

(
Bastro = 4 × 106 10) 2.5 photons / cm 2 sec .

8
10

7
10

6
10
Photons/cm sec

5
2

10

4
10

3
10

2
10

1
10
12 10 8 6 4 2 0 -2
Visual magnitude mv

It takes about one millisecond for light to pass

vertically through the Earth’s atmosphere.
Introduction 7

Isoplanatic Angle

Light traveling from a wavefront beacon should traverse

the same atmosphere as the light from the object of
interest. When the angular difference between the paths
results in a mean-square wavefront error of 1.0 rad2, the
angular difference is called the isoplanatic angle. For a
given structure constant profile ( Cn2 ( z) where z is the
altitude), and a flat-Earth assumption, the isoplanatic
angle is given by

−3 / 5
 
θ0 =  2.91k2 sec8 / 3 ζ
 Path
∫Cn2 ( z) z5 / 3 dz 

,

where k=2π/λ, ζ is the zenith angle, and the integral is

over the path from the ground-based telescope to the
source of light above the surface.

The graph illustrates the isoplanatic angle versus

wavelength for the Hufnagel-Valley H-V model and the
Stragic Laser Communication SLC model of
turbulence.
103

SLC, ς = 0û

HV5/7, ς = 0û
102
Θ0(µrad)

101
SLC, ς = 60û

HV5/7, ς = 60û

100
0.5 1.0 5.0 10.0
Wavelength (µm)
8 Adaptive Optics

Zernike Polynomials

Optical phase can be represented by a 2D surface over

the aperture. The deviation from flat (or some other
reference surface) is the wavefront error sensed by the
wavefront sensor. A very useful infinite-series
representation of the wavefront is the Zernike
polynomial series. Radial (index n) and azimuthal (index
m) polynomials are preceded by Zernike coefficients Anm
and Bnm that completely describe the wavefront up to the
order specified by the largest n or m. The series is written

∞
1 r
Φ(r, θ) = A00 +
2
∑A
n=2
0
n0 ℜ n  
 R
∞ n
r
+ ∑∑(A
n =1 m =1
nm cos mθ + Bnm sin mθ )ℜm
n  ,
 R

where the azimuthal polynomials are sines and cosines of

multiple angles and the radial polynomial is

n−m

r 2
( n − s)! r
n −2 s

∑ ( −1)
s
ℜm
n  =   .
 R n+m  n−m  R
s =0 s ! − s !  − s !  
 2   2 

The series is especially useful in adaptive optics because

the polynomials are orthogonal over a circle of radius R,
common to many optical system geometries. For R
normalized to unity, the first few radial terms are given
here.

n=0 n=1 n=2 n=3 n=4

m=0 1 2r2–1 6r4 – 6r2 + 1
m=1 r 3r3-2r
m=2 r2 4r4 – 3r2
m=3 r 3

m=4 r4
Introduction 9

Zernike Polynomials (cont’d)

A few terms are graphed to visualize their relationship to

third-order optical aberrations.

Zernike polynomials associated with low-order modes.

Reflecting telescopes of the Cassegrain design have a

central obscuration, which requires an extremely large
number of Zernike coefficients––more than can be
adequately described. A set of annular Zernike polynomials
may be obtained from Gram-Schmidt orthogon-
alization, and this series is generally used for optical
systems with central obscurations.
10 Adaptive Optics

Atmospheric Turbulence Models

One of the most widely used models for the atmospheric

turbulence structure constant as a function of altitude is
the H-V model:
2
W 
Cn2 ( h ) = 5.94 × 10 −23 h10   exp ( − h )
 27 
+ 2.7 × 10 −16 exp ( −2h / 3 ) + A exp ( −10 h ) ,

where h is the altitude in kilometers, and Cn2 is in units of

m-2/3. The parameters A and W are adjustable for local
conditions. For the most common H-V 5/7 model (leading
to r0 = 5 cm and θ0 = 7 µrad), the structure constant at the
surface A is 1.7 × 10-14, and the wind velocity aloft W is
21.

For conditions other than the 5/7 model, one can calculate
A and W from

A = 1.29 × 10 −12 r0−5 / 3 λ 2 − 1.61 × 10 −13 θ0−5/3 λ 2 − 3.89 × 10 −15 ,

W = 27(75θ0−5/3 λ 2 − 0.14)1 / 2 ,

where the coherence length r0 is in centimeters and the

isoplanatic angle θ0 is in microradians.

Other models are layered, such as the SLC-Night model:

Altitude (above ground) Cn2

h ≤ 18.5 m 8.40 × 10 −15
18.5 < h ≤ 110 m 2.87 × 10 −12 h−2
110 < h ≤ 1500 m 2.5 × 10 −16
1500 < h ≤ 7200 m 8.87 × 10−7 h −3
7200 < h ≤ 20,000 m 2.00 × 10 −16 h−0.5