Wavefront Sensors 25

Curvature Wavefront Sensor and Error

The curvature sensor is

an image-plane measurement
of local wavefront curvature
[the second derivative of
the wavefront ∇2 Φ(r)] de-
duced from two specific out-of-
focus images. A point-by-point
subtraction of the images is
proportional to the wavefront
curvature term minus the derivative of the wavefront at
the edge d Φ (r)
dn :

d Φ(r)
· ¸
I 1 (r) − I 2 (r) = C ∇2 Φ(r) −
dn

To provide an accurate mea-

surement of wavefront cur-
vature, the blur from the
turbulence must be small
compared to the area where
the curvature measurement
is taken. With p being the
offset from the focal plane of
the system and f as the focal length of the system, a con-
servative estimate of the blur requirement leads to
λf 2
p≥ .
λ f + r 20
Variance of a single curvature measurement:
p2
σ2Curv. Sens. =
f 4 Np
where Np is the photon count. Because the curvature
sensor directly measures the Laplacian of the wavefront,
bimorph mirrors are generally used for closed-loop
compensation, as they possess Laplacian influence
functions.

Field Guide to Adaptive Optics, Second Edition

26 Wavefront Sensors

Pyramid Wavefront Sensor and Error

The pyramid wavefront sensor is a pupil-plane

wavefront sensor.
Another pupil-plane wavefront sensor uses a pyramidal
prism in the image plane to create four subbeams that
are then optically relayed to a detector. The intensity at
position ~r( x, y) in each of the subbeams in the detector
plane ( I 0,0 > I 0,1 > I 1,0 > I 1,1 ) is used to find the x and y
wavefront slopes at ~r ( x, y):
I 0,0 (~r ) − I 1,0 (~r ) + I 0,1 (~r ) − I 1,1 (~r )
S x (~r ) = and
It
I 0,0 (~r ) + I 1,0 (~r ) − I 0,1 (~r ) − I 1,1 (~r )
S y (~r ) =
It
I t is the average intensity over the detector plane.

One advantage of the pyramid technique over a

Shack–Hartmann sensor is that the spatial resolution of
the sensor is the size of the detector pixel in contrast to the
larger lenslet subaperture size of the Shack–Hartmann.
The wavefront error variance associated with the pyramid
technique is approximated by
1 2( e n )2
σ2 = + rad2
2N N
where N is the number of photons per pixel and e n are the
read-noise electrons.

Field Guide to Adaptive Optics, Second Edition

Deformable Mirrors 27

Photodiodes

Photodiodes are p–n junctions and have a response

time that is a combination of drift time, diffusion time
and resistor–capacitor (RC) time. The drift time is the
time required to collect carriers in the depletion region,
the diffusion time is the time required for collection of
carriers in the undepleted bulk region, and the RC time
is the response time due to the equivalent RC circuit.
Photodiodes can be operated in either photovoltaic or
photoconductive mode. The photovoltaic mode has low
noise, as is uses a zero-volt bias, which does not induce
a dark current. Its response time is dominated by the drift
time, as the carriers are not excited by a bias voltage.
This mode is used for extremely low light level and low-
frequency applications. The photoconductive mode uses
a reverse bias voltage that increases the width of the
depletion region and decreases the junction capacitance,
increasing the response speed and linearity. This mode
has a constant dark current and increased noise, and its
response time is dominated by the diffusion time, as the
carriers in the undepleted region have farther to travel.
The responsivity R of a photodiode is the ratio of the
photocurrent to the incident power and is shown below for
silicon. InGaAs is generally used for IR wavelengths.

Field Guide to Adaptive Optics, Second Edition

28 Deformable Mirrors

Photodiode Noise

Photodiode noise comes from shot noise and Johnson

noise. Shot noise depends on the variance of the
photocurrent and dark current:
q
Is = 2q( I p + I d ) f

Here q is the electron charge (1.6 × 10−19 C), I p is the

photocurrent, I d is the dark current, and f is the noise
measurement bandwidth. Shot noise dominates when the
photodiode is operated in photoconductive mode.
Johnson noise is generated by thermal carriers and is
given as
s
4 k BTK f
IJ =
Rp

Here kB is the Boltzmann constant (1.38 × 10−23 J/K), TK is

the temperature in Kelvin, and Rp is the parallel or shunt
resistance. Johnson noise dominates when the photodiode
is operated in photovoltaic mode, as this mode has no dark
current and therefore very low shot noise.
The total noise current I n is the root-sum-square (RSS) of
the noise sources:
q
In = I s2 + I J2

Johnson and shot noise are both white noise processes;

however, 1/ f noise is also present where the power
spectrum is inversely proportional to the frequency. 1/ f
noise is commonly referred to as pink or flicker noise and
has negligible magnitude for frequencies above 1 Hz, so it
is typically ignored for noise performance analysis. This
type of noise has been widely observed in many dynamic
systems, and there is currently no universal theory for
its origin.
Noise equivalent power (NEP) is the ratio of the total
noise current to the responsivity and represents the level
of incident intensity that generates a photocurrent equal
to the noise current:
In
NEP =
R

Field Guide to Adaptive Optics, Second Edition