Optimal phase contrast in common-path

interferometry

Jesper Gluckstad and Paul C. Mogensen

We have developed an analytical model for the design and optimization of common-path interferometers
CPIs based on spatial filtering. We describe the mathematical analysis in detail and show how its
application to the optimization of a range of different CPIs results in the development of a graphical
framework to characterize quantitatively CPI performance. A detailed analytical treatment of the effect
of curvature in the synthetic reference wave is undertaken. We show that it is possible to improve the
linearity and fringe accuracy of certain standard interferometers by a modification of the Fourier filter,
and we propose and analyze a dual CPI system for the unambiguous mapping of phase to intensity over
the complete input phase range. 2001 Optical Society of America
OCIS codes: 070.2580, 070.6110, 100.2650, 100.5070, 120.3180, 120.5050.

1. Introduction complex. Although special cases such as, for exam-

The imaging and visualization of phase objects, wave-
front disturbances, and phase aberrations are of con-
siderable interest in the field of optics. A variety of
techniques are applied in such diverse fields as opti-
cal component testing, wave-front sensing, and
phase-contrast microscopy113 in which a qualitative
or quantitative analysis of an optical phase distur-
bance is required. In general, because a phase dis-
turbance cannot be viewed directly, a method must be
sought to extract the desired information indirectly,
for example, through the generation of fringe pat-
terns in an interferometer. A number of such visu-
alization techniques can be grouped under the
heading of common-path interferometry, in which a
portion of the signal is perturbed to act as a synthetic
reference for the unperturbed portion of the signal
wave front. Although the implementation of a
common-path interferometer CPI is relatively
straightforward and they exist in many different
forms such as the point-diffraction, dark-ground fil-
tering, field absorption, and phase-contrast method,
an analytical description of their operation is more
J. Gluckstad and P. C. Mogensen are with the Department of
Optics and Fluid Dynamics, Ris National Laboratory, Roskilde
DK-4000, Denmark. The e-mail address for J. Gluckstad is
jesper.gluckstad@risoe.dk.
Received 13 March 2000; revised manuscript received 11 Sep-
tember 2000.
0003-693501020268-15$15.000
2001 Optical Society of America
ple, the Zernike phase-contrast method have been
treated extensively,1,2,3,9 a comprehensive generic ap-
proach to CPI analysis is lacking.
Our overall aim in this paper is to provide a rigor-
ous analytical framework describing the operation,
design, and optimization of this generic class of in-
terferometers. The approach we use is based on a
generalization of the phase-contrast approach of
Zernike,1,2 which is not restricted by the so-called
small-scale phase approximation that limits the
original method. The generalized phase-contrast
approach1217 provides a straightforward methodol-
ogy for the optimization of a broad range of different
CPI types, and it supports and enhances the more
empirical methods for CPI optimization that are gen-
erally employed. Because our approach is an exten-
sion of the Zernike method, we use this as the
starting point for our analysis. In Section 2 we clar-
ify the limiting conditions for the application of the
Zernike method and explain the requirements for a
generalized description. In Section 3 we discuss
how our approach, based on the generalized phase-
contrast method, can be used for the optimization of
the entire class of generic CPI systems. We provide
a detailed description of the mathematical approach
that we adopted, and we analyze the importance of
the synthetic reference wave SRW amplitude profile
to obtain optimal performance from a CPI. In Sec-
tion 4 we consider the parameters that can best quan-
tify the performance of a general CPI system.
Adopting criteria originally proposed by Anderson,18
we use the simultaneous optimization of the visibility

268 APPLIED OPTICS Vol. 40, No. 2 10 January 2001

and the peak irradiance as the two key parameters
describing the output of any CPI. We demonstrate
that our analysis can be used for a number of differ-
ent CPI types with different Fourier plane filtering
parameters. Through the introduction of a com-
bined filter parameter, we develop a graphical
method for the optimization of the visibility and the
peak irradiance for a given phase distribution at the
input of the CPI. We apply our analysis to some
well-known CPI systems and consider how they
might be further optimized.
Of particular importance in the development of
CPI systems is the desire to produce a linear mapping
between the input phase of the wave front and the
output amplitude. The research of Henning19,20 ex-
amined the extension of the linear phase to the in-
tensity mapping range for the Zernike method. In
Section 5, using the analysis developed in Sections
2 4, we demonstrate that it is possible to generalize
Hennings phase-contrast method to obtain a wider
functionality. We take this opportunity to look in
detail at the difference between ambiguous and un-
ambiguous phase-to-intensity mapping and discuss
the possible improvements to some currently applied
CPI systems. In Section 6 we conclude by examin-
ing the benefits of our approach compared with other
techniques.
Previously we have described the application of a
graphical phasor chart for the design and analysis of
generalized phase-contrast systems.15 This ap-
proach is particularly useful in the understanding of
the optimization of the CPIs undertaken in this re-
search and as such we include a brief summary of the
application of the phasor chart in Appendix A.
In short, the aim in this paper is to provide a de-
scription of a rigorous, completely general analytical
approach for the optimization of the entire class of
CPI systems based on spatial filtering.
2. Generalization of the Zernike Description
The Zernike phase-contrast method2 is a well-
established technique for the visualization of phase
perturbations by use of a Fourier plane phase-
shifting filter. However, its successful operation in
a linear regime requires that the spatial phase dis-
tribution x, y at the input be limited to the so-
called small-scale phase approximation where the
largest phase is typically taken to be significantly less
than 3.1 If the phase distribution at the input is
thus restricted, then a Taylor expansion to first order
is sufficient for the mathematical treatment so that
the input wave front can be written as
expix, y 1 ix, y. (1) scale Zernike approximation, generate significant er-
The light corresponding to the two terms in this
small-scale phase approximation can be separated
spatially by use of a single lens where the phase
distribution is located in the front focal plane and the
corresponding spatial Fourier transformation is gen-
erated in the back focal plane of the lens. In this
first-order approximation the constant term repre-
sents the amplitude of on-axis light focused by the
lens in the back focal plane, and the second spatially
varying term represents the off-axis light. Zernike
realized that a small phase-shifting quarter-wave
plate acting on the focused light makes it possible for
one to obtain an approximately linear visualization of
small phase structures by generating interference be-
tween the two phase quadrature terms in approxi-
mation 1:
I x, y 1 2x, y. (2)
It should be noted that a three-quarter-wave plate
works equally well to produce contrast, but the plus
sign in approximation 2 is negated, leading to the
so-called negative phase contrast. A substantial im-
provement in the visibility of the Zernike phase-
contrast approximation requires strong damping of
the focused light in addition to the phase shift that is
required to generate the contrast.1
In the general case, in which we are not limited to
a small-scale input phase perturbation, we cannot
assume that a series expansion to first order as in the
Zernike approximation is a sufficient representation
of a given phase perturbation. Higher-order terms
in the expansion need to be taken into account, so the
expansion takes the form
expix, y 1 ix, y 12 2x, y 16 i 3x, y
1
24 4x, y . . . . (3)
However, here the spatially varying terms cannot be
considered as separate from the supposedly focused
light represented by the first term in this Taylor-
series expansion, as is implied by the Zernike ap-
proach, and all these spatially varying terms
contribute to the strength of the on-axis focused light.
For a significant modulation in the input phase, this
contribution of the spatially varying terms can result
in a significant modulation of the focal spot amplitude
in the back focal plane of the lens. These terms can
in fact result in either constructive or destructive
interference with the on-axis light, although the net
result will be an attenuation of the focused light am-
plitude that has only a maximum value for a perfect
unperturbed plane wave at the input.
There are, however, examples in the literature in
which the contribution of the higher-order terms in
the Taylor expansion is neglected, and it is assumed
that only the first term in the Taylor-series expansion
contributes to the strength of the focused light.311
In particular, some frequently cited derivations of
phase contrast,7,9 although correct within the small-
rors if they are extended to cover larger-scale phase
perturbations. However, the fact that certain re-
sults derived for small-scale phase-contrast imaging
including, for example, the dark-field method are
expressed by use of a general phasor notation,
expix, y, may explain why some flawed analyses
continue to propagate in the phase-contrast litera-
ture. For phase-contrast or dark-field imaging of
10 January 2001 Vol. 40, No. 2 APPLIED OPTICS 269

Fig. 1. Generic CPI based on a 4-f optical system lenses L1 and
L2. The input phase disturbance is shown as an aperture-
truncated phase function x, y, which generates an intensity
distribution Ix, y in the observation plane by a filtering opera-
tion in the Fourier plane. The values of the filter parameters A,
B, determine the type of filtering operation.
large-scale phase objects, use of a first-order Taylor-
expansion-based analysis that is actually valid only
within the regime of the small-scale phase approxi-
mation is unacceptable.
For phase objects that fall outside the limits of the
Zernike approximation we must find an alternative
mathematical approach to that of the Taylor expan-
sion given in Eq. 3. We choose a Fourier analysis
as a more suitable technique for separating com-
pletely the on-axis and higher-spatial-frequency com-
ponents. We obtain the following form for expix,
y, where x, y :
expix, y dxdy expix, ydxdy 3. Mathematical Analysis of the Common-Path
higher-frequency terms. (4)
In this Fourier decomposition the first term is a
complex-valued constant linked to the on-axis fo-
cused light from a phase object defined within the
spatial region , and the second term describes light
scattered by spatially varying structures in the phase
object. When we compare Eq. 3 with Eq. 4 it is
apparent that the first term of Eq. 3 is a poor ap-
proximation to the first term of Eq. 4 when one is
operating beyond the Zernike small-scale phase re-
gime.
An important issue for one to consider when ana-
lyzing the effect of spatial filtering on the light dif-
fracted by phase perturbations is the definition of
what spatially constitutes focused and scattered
light. In the above description of Zernike phase con-
trast it was assumed that the focused light is spa-
tially confined to a somewhat unphysical delta
function. As we know, any aperture truncation that
occurs within a practical optical system will lead to a
corresponding spatial broadening of the focused light.
Therefore it is essential that we define the terms
focused light and scattered light explicitly for such a
system. In this context it is necessary to look more
carefully at the sequence of apertures confining the
light-wave propagation through a typical optical
setup.
A commonly applied architecture that provides an
efficient platform for spatial filtering is illustrated in
270 APPLIED OPTICS Vol. 40, No. 2 10 January 2001
Fig. 1 and is based on the so-called 4-f configuration.
We obtain an output interferogram of an unknown
phase object or phase disturbance by applying a trun-
cated on-axis filtering operation in the spatial-
frequency domain between two Fourier-transforming
lenses L1 and L2. The first lens performs a spatial
Fourier transform so that directly propagated light is
focused into the on-axis filtering region, whereas spa-
tially varying object information generates light scat-
tered to locations outside this central region. We
can describe a general Fourier filter in which differ-
ent phase shifts and amplitude damping factors are
applied to the focused and the scattered light. In
Fig. 1 we show a circularly symmetric Fourier filter
described by the amplitude transmission factors A
and B for the scattered and for the focused light,
respectively, and by the relative phase shift .
These filter parameters can be chosen to replicate a
large number of commonly used filter types i.e.,
phase-contrast, dark-central-ground, point-diffraction,
and field absorption filtering. By applying a given
Fourier filter and a second Fourier lens, we obtain an
interference pattern in the observation plane. The
focused on-axis light acts as the SRW in the CPI
system, which interferes with the scattered light and
generates the output interference pattern. In Sec-
tion 3 we discuss the importance of the SRW and
show how it influences, among other things, the
choice of the Fourier filter parameters.
Interferometer
Having described the generic optical system that
makes up the CPI, we now turn to a detailed analyt-
ical treatment of the important elements in this sys-
tem. Assuming a circular input aperture with
radius r truncating the phase disturbance that is
modulated onto a collimated, unit amplitude, mono-
chromatic field of wavelength , we can describe the
incoming light amplitude a x, y by
a x, y circrrexpix, y (5)
at the entrance plane of the optical system shown in
Fig. 1 using the definition that the circ function is
unity within the region, r x2 y212 r, and
zero elsewhere.
Similarly, we assume a circular on-axis centered
spatial filter of the form
H fx, fy A1 BA1 expi 1circ frfr, (6)
where B 0; 1 is the chosen filter transmittance of
the focused light, 0; 2 is the applied phase shift
to the focused light, and A 0; 1 is a filter param-
eter describing the field transmittance for off-axis
scattered light as indicated in Fig. 1. The spatial-
frequency coordinates are related to spatial coordi-
nates in the filter plane such that fx, fy f 1xf,
yf and fr fx2 fy212 Ref. 1.
Performing an optical Fourier transform of the in-
put field from Eq. 5 followed by multiplication with
the filter parameters in Eq. 6 and a second optical
Fourier transform corresponding to an inverse Fou-
rier transform with inverted coordinates, we obtain
an expression for the intensity Ix, y that describes
the interferogram at the observation plane of the 4-f
setup:
Ix, y A2expi x, ycircrr
BA1 expi 1gr2, (7)
where gr is the SRW and the terms and are
given by
r21
x2y2r
expix, ydxdy
expi ,
. (8)
The generally complex-valued and object-dependent
term that corresponds to the amplitude of the fo-
cused light plays a significant role in the expression
for the interference pattern described by Eq. 7. Re-
ferring to the discussion in Section 1, we are now able
to confirm that the frequent assumptionthat the
amplitude of the focused light is approximately equal
to the first term of the Taylor expansion in approxi-
mation 1 generally can result in misleading inter-
pretations of the interferograms generated at the CPI
output.
Of similar importance in the analysis of Eq. 7 is
the term gr that describes the spatial profile of the
SRW, diffracted from the aperture formed by the on-
axis centered filtering region. It is the interference
between this SRW term, carrying the information
about the filtering parameters, and the imaged phase
object that generates the output interferogram. It is
important to obtain an accurate description for the
SRW and thereby an accurate derivation for Eq. 7.
The zero-order Hankel transform1 followed by a se-
ries expansion in the spatial dimension r is used here
to describe the SRW. This is a relatively simple ap-
proach, which to the best of our knowledge21 has not
previously been applied to this problem.
For a circular input aperture with radius r, we
can describe the radius of the corresponding central
phase-shifting region of the Fourier filter character-
ized by the parameters B and in terms of a radial
spatial-frequency range fr. Thus we can obtain the
following expression for the SRW by use of the zero-
order Hankel transform:
fr
gr 2r J12rfrJ02rfrdfr. (9)
0
To simplify the analysis, we introduce the term ,
which explicitly relates the radius of the central fil-
tering region R1 to the radius of the mainlobe of the
Airy function R2 resulting from the Fourier trans-
form of the circular input aperture alone. This re-
lationship defining is shown graphically in Fig. 2.
Fig. 2. Schematic representation of the relationship between the
profile of the input aperture-diffracted Airy function and the spa-
tial profile of a generic Fourier filter see Fig. 1. The term is
defined as the ratio of the radii of the filter R1 and the mainlobe of
the Airy function R2.
We can thus express in terms of r and fr such
that
R1R2 0.611rfr, (10)
where the factor of 0.61 arises from the radial dis-
tance to the first zero crossing of the Airy function
corresponding to half of the Airy mainlobe factor of
1.22.1 If we make this substitution in Eq. 9 and
then perform a series expansion in r, we obtain the
following expression for the SRW:
gr 1 J01.22 0.612J21.22
rr2 0.61342J31.22
0.61J41.22rr4 . . . . (11)
In this expansion, the SRW is expressed in radial
coordinates normalized to the radius of the imaged
input aperture. This can be scaled easily to allow for
a magnification within the imaging system, although
for the remainder of our analysis a direct imaging
operation is assumed. From Eq. 11 it is apparent
that the SRW will change as a function of the radius
of the central filtering region. In addition, it is clear
that the SRW profile is not necessarily flat over the
system output aperture. This is an important, yet
often neglected, factor when one is determining the
performance of a CPI.
In Fig. 3 we plot the input normalized SRW ampli-
tude as a function of the output radius coordinate r
normalized to the system aperture radius r for a
range of different values. It can be seen that as
increases so does the strength of the SRW, which
from qualitative arguments can be expected see Fig.
2. However, the curvature also increases with in-
creasing , thus distorting the wave-front profile of
the SRW. From the point of view of optimizing a
CPI, we want the curvature of gr to be negligible
over a sufficiently large spatial region of the system
aperture centered around r 0. First, should be
chosen so that gr 1, corresponding to the first
zero crossing of the Bessel function J01.22, where
10 January 2001 Vol. 40, No. 2 APPLIED OPTICS 271

Fig. 3. Plot of the spatial variation of the normalized SRW am-
plitude gr as a function of the normalized CPI observation plane
radius for a range of values from 0.2 to 0.627. This plot shows
that a large value of produces significant curvature in the SRW
across the aperture, which will cause a distortion of the output
interference pattern. In contrast, a low value of generates a flat
SRW, but at the cost of a reduction in the SRW amplitude.
0.627. Second, it is important to keep as small
as possible to make sure that the scattered object
light is not propagated through the zero-order filter-
ing region. Finding a minimum applicable value
is less apparent, but obviously a small value will
reduce the strength of the SRW to an unacceptably
low level, compromising the fringe visibility in the
interference with the diffracted light. The term
can thus have a significant impact on the resulting
interferometric performance and is as important as
the filter parameters A, B, and in the design of a
CPI.
Depending on the accuracy needed for the descrip-
tion of the interferograms, one can choose to include
a number of spatial higher-order terms from the ex-
pansion in Eq. 11. The influence of the higher-
order terms has the largest impact along the
boundaries of the imaged aperture. For values
smaller than 0.627 and when one is operating within
the central region of the image plane, spatial higher-
order terms are insignificant and we can approximate
the SRW with the first and space-invariant term in
Eq. 11:
gr central region 1 J01.22. (12) cause of the relative flatness of the SRW as shown
We can then simplify Eq. 7 to obtain
I x, y A2expi x, y tering region, reducing potential errors in the in-
K BA1 expi 12, (13) terpretation of the interferometric fringe patterns
where K 1 J01.22. Thus the influence of
the finite on-axis filtering radius on the focused light
is included effectively as an extra filtering parameter
so that the four-parameter filter set A, B, , K
together with the complex-object-dependent term
effectively defines the type of filtering scheme that we
are applying.
In this context it is of interest to compare our anal-
ysis of the SRW with a previous treatment of the term
272 APPLIED OPTICS Vol. 40, No. 2 10 January 2001
denoted by . In Ref. 22 a liquid-crystal point-
diffraction interferometer the phase-shifting type is
analyzed in detail. In estimating the impact of the
term , the authors apply numerical calculations on
an example amplitude object. The amplitude of an
unperturbed focused and filter-truncated beam cor-
responding to K above is normalized to unity for
different values, making it difficult to validate the
impact on the resulting interferograms. However,
the authors cite other papers in which 0.5 Ref.
23 and 33,24 and, based on ad hoc arguments,
they choose 3 in their own application despite the
fact that 3 for a system that is categorized as a
point-diffraction interferometer. This lies far out-
side the analytically derived maximum value of
0.627, which we suggest in this paper.
In Ref. 18 a general CPI architecture is analyzed
carefully. In this paper the impact of the term
corresponding to in our analysis is evaluated by
numerical calculations and computer simulations
showing the difference in the obtained K values.
Based on a series of numerical simulations, Ander-
son18 concludes that a choice of 0.4 consistently
provides good interferometric results. Our analy-
sis is in good agreement with this, and it can be
seen from Fig. 3 that a value of 0.4 represents
a good compromise between the two extremes of low
curvature and large amplitude for the SRW in a
CPI.
Having determined a suitable operating range for
the CPI in terms of the production of a good SRW, we
now examine the role that the remaining filter pa-
rameters play in the optimization of a CPI.
4. Optimization of a Generic Common-Path
Interferometer
In Section 3 the deleterious effects of SRW curva-
ture on the performance of a CPI were discussed,
and our desire to minimize this effect while keeping
a sufficient amplitude strength prompted the choice
of 0.4 for the ratio of the radius of the central
filter region to the mainlobe of the Airy function
corresponding to a value of K 12 in relation 13.
This choice should provide high fringe accuracy be-
in Fig. 3. In addition, a small value almost guar-
antees that any scattered light from the input does
not propagate through the zero-order centered fil-
that are generated. Keeping 0.4 and thereby
K 12 fixed throughout the remainder of this
analysis, we can now focus on the optimization of
the remaining filter parameters A, B, . To op-
timize these CPI filter parameters we adopt the
three criteria outlined by Anderson,18 aiming for a
high fringe accuracy, b optimal visibility, and c
maximum peak irradiance.
From relation 13 we can see that the filter pa-
rameters A, B, can be combined to form a single
complex-valued term C, the combined filter term,
such that
C CexpiC BA1 expi 1. (14)
Therefore relation 13 can be simplified to obtain
I x, y A2expi x, y iC K C2, (15)
where
BA1 1 2CcosC C212,
sin1BA11CsinC. (16)
Because it is a complex variable, the combined filter
term C, which effectively describes the complex filter
space, can be considered to consist of a vector of phase
C and length C, as shown in Eq. 14. Thus to
obtain an overview of the operating space covered by
all the possible combinations of three independent
filter parameters A, B, , we can now choose instead
to consider a given filter in terms of the two combined
parameters C and C. However, in Eq. 15 it can
be seen that the filter parameter A also appears in-
dependently of the combined filter term C. Fortu-
nately, we can resolve this problem by considering
that the term BA1 from Eq. 14 must be constrained
in the following way:
BA1 1 f A 1, B C 1,
BA1 1 f A 1, B 1,
BA1 1 f B 1, A C 11. (17)
These constraints arise from the adoption of Ander-
sons maximum irradiance criterion.18 This implies
that we want to minimize unnecessary absorption of
light in the Fourier filter, which reduces both the
irradiance and the signal-to-noise ratio in the CPI
output fringe measurements.
Any given filter can be defined explicitly by a given
value of the two parameters C and C. Therefore
we can use a single plot to display graphically the
location of a given filter within the complex filter
space. Such a plot is shown in Fig. 4 in which we
plot the magnitude of the combined filter parameter
C versus its phase C. There are different families
of operating curves in this complex filter space, each
of which can be traced out when a term such as BA1
is kept constant while is varied or vice versa these
form the gridlike structure in Fig. 4. When the op-
erating curves for C are plotted in this way, it is
relatively simple to identify particular operating re-
gimes for different classes of filters. For example,
we are particularly interested in the operating curve
for a lossless Fourier filter, a filter in which BA1
1, because this is a class of filter for which optical
throughput is maximized. The lossless operating
curve is shown as the bold curve in Fig. 4. We can
derive the expression for the shape of the lossless
operating curve by using the identity
expi 1 2sin2expi 2, (18) ing methods. We can use the complex filter space
Fig. 4. Complex filter space plot of the modulus of the combined
filter parameter C versus the phase C over the complete 2 phase
region. Use of these combined parameters defined in Eqs. 14
and 16 allows us to visualize simultaneously all the available
combinations of the terms A, B, and . The bold curve is the
operating curve for a phase-only lossless filter, whereas the fine
grid represents the operating curves for different values of the
filter terms A, B, and . The operating regimes are marked for a
number of CPI architectures, including A, Zernike; B, Henning;
C, generalized phase-contrast methods; D, dark central ground;
E, field absorption filtering; F, G, point-diffraction. Full filter
details for these techniques are summarized in Table 1.
and combining Eqs. 14 and 18 we obtain an ex-
pression for the lossless operating curve, for which C
is defined for two distinct regions over the 2 phase
region:
C 2cosC for C 2; 32,
C 0 for C2; 32. (19)
The development of this complex filter space plot
makes it extremely simple to compare different filter
parameter selections that have been proposed and
independently dealt with in the scientific literature.
In Fig. 4 we superimposed the operating points for
some of the different CPI filter types that are com-
monly used. The filter type, parameters, and corre-
sponding labels and references are summarized in
Table 1. The filter space plot provides not only a
convenient method of comparing different CPI archi-
tectures, but also forms a useful basis for exploring
the effectiveness of new CPI configurations.
Referring to Fig. 4, it can be seen that many of the
filter types used in CPIs lie on or close to the lossless
operating curve defined by Eq. 19. The arrows on
the plot indicate the general regime in the complex
filter space in which a given filter type can be oper-
ated.
To the best of our knowledge, this identification
and labeling of known cases from the literature in a
single general filter space representation has not
been demonstrated previously. We believe that it
provides a useful basis for understanding the inher-
ent differences or similarities between existing filter-
10 January 2001 Vol. 40, No. 2 APPLIED OPTICS 273
 Table 1. Comparison of Filter Parameters for the Different CPI Types Highlighted in Fig. 4
Method
Zernike phase contrast
Henning phase contrast
Generalized phase contrast
Dark-central-ground filtering
Field absorption filtering
Point-diffraction Smartt interferometer
Phase-shifting point-diffraction interferometer
plot as a framework on which to optimize the visibil-
ity and the irradiance of a CPI.
A. Optimization of the Visibility and Peak Irradiance
Returning to the CPI optimization criteria of Ander-
son,18 we can denote the maximum and minimum
intensity levels at the interferometric output as Imax
and Imin for any given CPI. The visibility V is ex-
pressed in terms of the output intensity minima and
maxima such that
Imax Imin
V .
Imax Imin
The peak irradiance is defined as the maximum in
the intensity at the output of the CPI and is therefore
simply characterized by the term Imax.
For any given input phase perturbation with a dy-
namic range of , we can derive an expression for
the visibility and the peak irradiance by combining
Eqs. 15 and 20:
V max cos min cosK C1 K C
max cos min cos1,
Imax A21 K C2 2K Cmax cos, (22)
where
min cos mincos C,
max cos maxcos C, 2; 2. 32. These are shown in the form of normalized
These expressions for the visibility and the peak ir-
radiance can be simplified in some special cases,
which we discuss below and in Section 5. However,
first let us consider an analysis of the general case as
given in the full versions of Eqs. 2123. It is ap-
parent that a general analysis over the entire domain
covered by the complex filter space plot of Fig. 4 is
complicated by the appearance of the variables
and in the expressions for the visibility and the
peak irradiance. The variable K is held fixed K
12 in this analysis for high fringe accuracy, as was
discussed in Section 3. Instead of arbitrarily choos-
ing combinations of and before an optimization
analysis for the filter parameters, we model the input
phase perturbations by a random distribution of pha-
sors over the phase range generating a spatial
average or wave-front Strehl ratio29 corresponding to
274 APPLIED OPTICS Vol. 40, No. 2 10 January 2001
Filter Parameters Reference Figure 4 Label
A 1, B 0.05; 1, 2 1, 2 A1, A2
A 21, B 1, 4 19, 20 B1, B2
A B 1, 2 sin1 12 1217 C
A 1, B 0 46 D
A 1, B 1, 0 18 E
A 1, B 1, 0, K 1 23, 2527 F
A B 1, 0; 2, K 1 22, 28 G
. For example, we can choose to analyze phase
perturbations modeled by a uniform random distri-
bution of phasors. In this case, the two terms and
are related by the absolute value of a sinc function
in the following way30:
uniform sinc 2. (24)
Using this expression to substitute for in Eqs. 21
and 22 we can produce expressions for the visibility
(20) and the peak irradiance, which depend only on the
input phase modulation depth and the combined fil-
ter parameter. After we derive expressions for the
visibility and the peak irradiance, it is instructive to
compare their variation as a function of the phase
modulation depth. We are then able to compare sys-
tematically the performance of a generic CPI system
for different ranges of phase perturbations at the
input. Using the complex filter space plot as a
framework, we are able to display the variation in the
visibility and the peak irradiance at the CPI output
as a function of the complex components C and C of
(21) the combined filter parameter. By using a pair of
complex filter space plots, we can thus show the op-
timum filter type for a given phase perturbation to
maximize the visibility or the peak irradiance. In
Fig. 5 we show a set of contour plot pairs for the
visibility and the peak irradiance with a range of
input phase distribution depths from a 8 to d
(23) gray-scale contour plots in which the maximum value
is represented by the darkest region. All four sets of
complex filter space plots cover the same range for
the parameters C and C, from 0 2 rad and 0 4,
respectively. The operating curve for the lossless
filter given by Eq. 19 is shown on all the plots. It is
evident in Fig. 5a that a small does not provide
much overlap between regions of high visibility and
strong peak irradiance. For the larger values of ,
such as those in Figs. 5b and 5c, an overlap can be
found between the regions of high visibility and peak
irradiance. However, there is no overlap between
regions in which these values are maximized, illus-
trating that there will always be a trade-off between
the absolute optimization of these two terms. It is of
interest to note that, as the value of increases
further to 32 see Fig. 5d and beyond, then the
overlap between high values of the visibility and the
peak irradiance diminishes.
Fig. 5. Set of complex filter space plots for the optimization of the visibility and the peak irradiance with different input phase
distributions. The four pairs of plots illustrate the variation in the visibility and the peak irradiance at the output of a CPI for a uniform
random-phase distribution of dynamic ranges of a 8, b 3, c , and d 32, for a constant value of K 12. The gray scale of
the contour plots defines the relative visibility or peak irradiance with the darkest regions representing the maximum value that can be
achieved when the Fourier filter parameters are varied. The operating curve for a lossless filter is shown on all the plots.

The upward shift toward higher values of C and the Imax A21 K C2 2K C. (28)
broadening of the maximum visibility region in the
filter space domain can be understood when we derive Another interesting observation related to the contour
the expression for C in the case of optimal visibility: plots in Fig. 5 is that C 2 2 provides a good
compromise filter in most cases, both for obtaining high
CV1 K 1 expi i , 2; 2. visibility and for strong peak irradiance. Moreover, this
(25) corresponds to a lossless phase-only filter with and
A B 1 as is evident from Eq. 14 or Eq. 18. Using
In this case, we use the fact that the minimum in- numerical simulations, Anderson18 also found the loss-
tensity will have to be zero within the interval to less -phase filter to be a good overall compromise for a
obtain optimal visibility. When we combine Eq. 21 wide range of input phase disturbances.
with Eq. 25, the reason for the shift and the broad-
ening of the maximum visibility region in Fig. 5 with B. Optimal Unambiguous Phase-to-Intensity Mapping
increasing becomes clear. In the limiting range For phase perturbations in the range we have
where 3 2, we can disregard the phasor part of the possibility of optimizing both the visibility and
Eq. 25 and we then find that the irradiance while simultaneously obtaining a
monotonically increasing mapping of the input phase
CV1 CV1 K 1, 2. (26) to the output intensity. Unfortunately this excludes
We find the corresponding expressions for the visibil- relatively simple filters such as the lossless -phase
ity and the peak irradiance in this limiting case to be filter that was convenient in the above analysis in
which the unambiguous mapping of phase to inten-
V 2K C1 K C1, (27) sity was not the primary concern.

10 January 2001 Vol. 40, No. 2 APPLIED OPTICS 275

Fig. 6. Optimization of the visibility and the peak irradiance for an unambiguous direct mapping of the input phase to the output intensity
for a range of uniform random-phase distributions of dynamic ranges of a 8, b 3, c 2, and d 34. A truncation of the allowed
values for C arises from the unambiguous phase-to-intensity mapping requirement. As the depth of the input phase distribution
increases, the allowed values of the combined filter parameter are further constrained.

To obtain the required monotonic increase in the
phase-to-intensity mapping we must map the respec-
tive minima and maxima of the input phase distur-
bance to the corresponding limits in the output
intensity distribution. We therefore obtain the fol-
lowing expression for the visibility:
V I 2 I 2I 2 I 21
2 sin 2sinCK C1 K C to ensure an unambiguous phase-to-intensity map-
2 cos 2cosC1, (29)
where the corresponding peak irradiance is given by
Imax I 2 A2expi 2 iC K C2. (30)
As in the analysis of Subsection 4.A, we can again
derive an analytic expression for the combined single
filter parameter C that generates optimal visibility
for the unambiguous phase-to-intensity mapping
over the range :
CV1 K 1 expi i 2. (31)
This corresponds to a filter consisting of the following
amplitude and phase values obtained by use of Eq.
16:
BA1 1 2K 1 cos 2 K 212,
sin1BA1K 1 sin 2. (32)
Another interesting constraint enforced on the filter
ping is that the available phase space for C is reduced
as increases, approaching its maximum value of
. In fact, it can be shown that the range of C that we
can use for a given is
C 2; 2, (33)
where the single value C 2 is obtained for the
largest possible unambiguous range .
The constraints on the values that C can take
given by expression 33 can be identified easily on
the complex filter space plots for the visibility and the
irradiance. In Fig. 6 we show sets of such plots in

276 APPLIED OPTICS Vol. 40, No. 2 10 January 2001

which the visibility and the peak irradiance are max-
imized for different input phase depths from 8
to 34, with the additional constraint of an unam-
biguous mapping of the phase to the intensity. As in
Subsection 4.A, we model the phase perturbation us-
ing the relationship given in Eq. 24. In Fig. 6a we
show a pair of plots of the visibility and the peak
irradiance for a phase perturbation 8. The
constraints on the values that C can take are shown
by the sharp vertical boundaries on both plots for C
values of 16 and 1516 rad. In all the examples
shown in Fig. 6, we are limited to working within the
range of C ; therefore a restricted C axis is used
to improve clarity. The appropriate portion of the
operating curve for lossless filtering is shown on all
the plots. As the value of increases to 3 and
2 as shown in Figs. 6b and 6c, respectively, it
can be seen that the available portion of the phase-
space domain in which a filter matching our criteria
can be placed is further constrained. In the plots of
Figs. 6a 6d it can be seen that the trends in peak
irradiance and maximum visibility observed are the
same as those exhibited for the general optimization
of the filter in Fig. 5. However, it should be noted
that there are differences in the position of the peak
irradiance regions and the extent of the maximum
visibilities in the contour plots of Fig. 6 when com-
pared with those of Fig. 5.
5. Optimization of the Phase-to-Intensity Mapping
Linearity
Thus far we have developed a rigorous approach for
the optimization of the visibility and the peak irradi-
ance in a generic CPI and extended this optimization
to the removal of ambiguity from the phase-to-
intensity mapping. In this section we consider
improving the linearity of the unambiguous phase-
to-intensity mapping described in Section 4. In-
spired in part by Hennings phase-contrast
method,19,20 we look for ways of improving the linear-
ity of the phase-to-intensity mapping because the
user of a CPI is interested primarily in the correlation
of a given output intensity to a particular phase per-
turbation at the input.
The intention of the Henning method was to extend
the linearity of the Zernike phase-contrast method by
introducing field absorption to the Fourier filter, in
this case setting the filter parameter A 12.
The phase shift of the filter was also modified from
2 to 4. Using our analysis, we can
confirm that this change of filter parameters extends
the linearity as intended. Starting from our derived
relationship for the dependence of the CPI output
intensity on the input phase distribution and filter
parameters, given in relation 13, we obtain the fol-
lowing expression for the Zernike case:
IZernikex, y expi x, y i 12
3 2sin x, y cos x, y. Ix, y A21 K C2
Here we use a positive phase-contrast filter with A
B 1 and 2 the alternative negative phase
simply generates contrast reversal and we take K
1. This is a basic assumption in the Zernike case
in which the phase perturbation x, y is small
around the zero value corresponding to the average
phase in Eq. 8. Thus for small phase pertur-
bations, Eq. 34 simplifies to
IZernike 3 2sin cos 3
1 2 2 for 3 0. (35)
This expression has the same limiting value as the
well-known form found in many optics textbooks see,
for example, Ref. 1. The quadratic term in Eq. 35
quickly becomes important and enforces a natural
limitation to the range of linearity of the Zernike
phase-contrast scheme.
We can demonstrate the elegance of the Henning
method for extending the linearity19,20 by again start-
ing from relation 13 and making the following sub-
stitutions: A 12, B 1, 4, and K 1
positive Henning phase-contrast filter. Thus we
obtain
IHenningx, y 12expi x, y
2 expi4 12
1 sin x, y. (36)
For small phase perturbations around zero, Eq. 36
simplifies to
IHenning 1 sin 3 1 for 3 0. (37)
Comparing this expression with that given for the
Zernike method in Eq. 35, we can see that linearity
has been extended significantly.
Both Zernikes and Hennings phase-contrast
methods inherently assume that we apply the con-
straint K 1 to relation 13 to arrive at the ex-
pressions derived in Eqs. 35 and 37 and that are
found in many optics textbooks.1,7,9,20 However, as
we have seen throughout this paper, having K 1
is, in the majority of cases, not the best choice for an
operating condition i.e., we found K 12 more
suitable for high fringe accuracy. In addition, is
implicitly less than unity for even small-scale phase
perturbations, as can be verified by the phasor inte-
gration in Eq. 8.
A. Generalization of the Henning Phase-Contrast Method
In light of the analysis presented above, we thus
suggest modifications to Hennings method so that we
can guarantee high linearity as derived in Eq. 37 as
well as high fringe accuracy. The first modification
to Hennings method is simply to introduce K 12
and rederive the filter parameters required to arrive
at Eq. 37.
From Eq. 15 we obtain
(34) 2K Ccos x, y C, (38)
10 January 2001 Vol. 40, No. 2 APPLIED OPTICS 277
and, except for a trivial overall intensity scaling, we
can now identify the combined filter parameter C
that best solves the problem:

K 12 f C 2 expi2. (39)
This corresponds to the following choice of the basic
filter parameters:

BA1 5 f A 15, B 1,
sin 25.
1
(40)
and the resulting intensity distribution Fig. 7. Graphical phasor chart analysis for the generalized Hen-
ning method. The zero point of the quadratic intensity scale I is
Ix, yHenning 251 sin x, y. (41) fixed at the point where 32 on the unit phase circle. The
K12 intersection of I with the unit circle directly yields the output
intensity value, which corresponds to the input phase of the CPI.
This result can be improved further if we also take This mapping is approximately linear for the range 3 about the
into consideration the correct input-dependent aver- points 0 and indicated by the dashed curves.
age value of 1 so that we obtain

C K 1 expi2. (42)
The filter parameters now have the values
BA1 1 K 212 f A 1 K 212, representation of the generalized Henning method,
B 1, sin11 K 212. (43) I at the point
which generates the resulting intensity distribution
I x, ygeneralized 21 K 211 sin x, y. intensity mapping in the generalized Henning case
Henning
We call this technique with improved linearity the
generalized Henning method. In practice, it would
require an adaptive Fourier filter controlled by a
sensing unit that continuously measures the
strength of the focused light intensity, 2, and pro-
vides this as feedback to determine the selection of
the parameters in the adaptive filter. With such a
filter in place, we would be guaranteed optimal lin-
earity for any given phase perturbation covering a
range up to 3, which is the approximately lin-
ear region of the term 1 sin .
From Eq. 42 it can be seen that the inclusion of
the C-parameter solutions for the generalized Hen-
ning method would correspond graphically to a trac-
ing of a vertical line parallel to the C axis and fixed
at C 2 in the filter space plot of Fig. 4. It
should be noted that the filter operating points form-
ing the grid in Fig. 4 have an irregular distribution,
which explains why an apparently simple C solution
generates the need for such an adaptive filter in
which one component is fixed and the other two are
functions of , as shown in Eq. 43.
It is possible to illustrate the operating regime for
the generalized Henning phase-contrast method by
use of a two-dimensional phasor chart. This chart
was developed originally as a graphical method for
the analysis of CPI systems.15 It has been applied
subsequently to the optimization of systems based on
the generalized phase-contrast technique, including
among others the analysis of array generators.16 A
brief summary of the phasor chart and its use are
given in Appendix A.
In Fig. 7 it can be seen that, for the phasor chart
we fix the zero point for the quadratic intensity scale
32 on the unit circle corre-
sponding to the filter parameters given by Eqs. 42
and 43. It is then possible to directly read out
corresponding pairs of values for the phase-to-
(44) as is illustrated in Fig. 7. Using the phasor chart,
we can verify graphically that the range for linear
mapping is approximately 3 around 0 indi-
cated by the dashed curve on the chart. An addi-
tional linear mapping region of 3 around is
also indicated. We can confirm the linearity ob-
tained in the phase-to-intensity mapping in the pha-
sor chart by considering the mapping as the graphical
intersection points of the quadratic intensity scale
with the dashed curve. By simple trigonometry we
find that the squared length L2, measured along the
intensity scale intersecting the unit circle at any
given point, is given by
L2 cos2 sin 12 1 sin . (45)
This expression has exactly the same functional form
as obtained in Eq. 44.
B. Linear Phase-to-Intensity Mapping over the Phase
Unity Circle
Now that we have demonstrated that it is theoreti-
cally possible to improve the linear range of the
phase-to-intensity mapping for the CPI through a
generalization of the Henning method as described in
Subsection 5.A, it is interesting to contemplate how
linearity might be extended to generate a mapping
for a larger range of input phase values than 3.
One possible solution for the extension of a linear
phase-to-intensity mapping regime would be the ad-
dition of a second CPI working in conjunction with

278 APPLIED OPTICS Vol. 40, No. 2 10 January 2001

 Table 2. Dual CPI Configuration for Linear Phase-to-Intensity Mapping
over the 2 Phase Circle

Phase Segment CPI-1 CPI-2

4:4 Linear detection Low-intensity mode

4:34 High-intensity mode Linear detection
34:54 Linear detection High-intensity mode
54:74 Low-intensity mode Linear detection

Fig. 8 we can now explicitly identify four linear phase

detection intervals over the complete 2 phase circle,
and these intervals and the corresponding operation
Fig. 8. Graphical phasor chart for the CPI with filter parameters modes of both CPIs are given in Table 2.
given by Eqs. 46 and 47. The zero point of the quadratic in-
This particular configuration of the two CPIs is not
tensity scale I is fixed at the point where 0 on the unit phase
circle. As in Fig. 7, the intersection of I with the unit circle
the only possibility for establishing linear phase-to-
yields the output intensity for a given input phase. This phase- intensity mapping over the whole unit circle. As we
to-intensity mapping is approximately linear for the range 3 can easily confirm by use of the phasor chart, any pair
about the points 2 and 32 indicated by the dashed of CPIs with the intensity scale zero points displaced
curves. by 2 on the unit circle is capable of doing the task.
Therefore the CPI designer can, within the optical
constraints given, choose a dual CPI configuration
the first in the same optical system and thus viewing from this range of options that could optimize addi-
the same input phase disturbance. tional tasks such as the energy throughput or sim-
Looking at the phasor chart mapping of the gener- plicity of filter fabrication or operation.
alized Henning method in Fig. 7, we can see that it is We conclude this section by considering an exam-
indeed possible to cover the whole unity phase circle ple of a CPI pair configuration that is both relatively
with an additional interferometer with its intensity simple to implement and energy efficient for each of
scale zero point fixed at the point corresponding to the CPIs when K 12 is used for high fringe accu-
0, as shown in Fig. 8. For this CPI, we obtain racy. The phasor chart for this setup is shown in
the combined filter parameter Fig. 9. The linear mapping regions are labeled on
the respective intensity scales for each CPI where 1Q
C K 1 expi, (46) corresponds to the first quadrant, 2Q is for the second
quadrant, and so forth. There are some interesting
corresponding to the filter parameters
points that arise from an examination of this dual
, CPI configuration. First, the zero points for the in-
tensity scales for both CPIs must be placed 2 apart
1
BA 1 2K 1 K 212, (47) on the unit circle as already described if we wish to
obtain a linear phase-to-intensity mapping over the
and the resulting intensity distribution complete 2 phase range. Second, because the zero
Ix, y 2A21 cos x, y. (48)
The reason we do not calculate explicitly the individ-
ual filter values A and B for this interferometer can
be understood when we look at the phasor chart in
Fig. 8. It can be seen that, depending on which K
circle we are operating on, we need the possibility for
both field absorption to scale outward by a factor
A1 or dark-ground absorption to scale inward by a
factor B to arrive at 0 for the zero point on the
intensity scale. Choosing K 12 for high fringe
accuracy, however, we can omit these considerations
and keep B 1, varying only a single filter parame-
ter, the field absorption, according to Eq. 47.
Combining a CPI based on the generalized Henning
method with a second CPI as shown in Fig. 8, we show
Fig. 9. Graphical phasor chart for a symmetric dual CPI config-
that the whole unit circle can be covered by an al-
uration with both CPIs shown on the same chart. The zero point
most linear phase-to-intensity mapping. In Figs. 7 of the quadratic intensity scales I for each of the two CPIs are fixed
and 8 we used dashed curves to indicate how each CPI at the points 54 and 4 on the unit circle. The
covers two ranges extending over 4 centered at intensity scales are labeled to indicate the region for which a linear
different points. Denoting the two CPIs as CPI-1 mapping is provided i.e., 1Q corresponds to the first quadrant, 2Q
corresponding to Fig. 7 and CPI-2 corresponding to is the second quadrant, and so forth.

10 January 2001 Vol. 40, No. 2 APPLIED OPTICS 279

intensity points for both CPIs are positioned sym-
metrically about the x axis and therefore equally
displaced from 0, the same degree of field ab- optimization. We conclude that, in cases in which the
sorption is required for both CPI Fourier filters, how-
ever the phase shifts of the two filters are different.
We selected K 12 to fulfill high fringe accuracy
requirements, and because 1, it is apparent that mance. Using the complex filter space plots that we
the CPI input can lie anywhere within the confines of
the K 12 circle. The proximity of the zero in- less operating curve, for a filter in which there is no
tensity point to the K value describing the input absorption, provides an extremely good first choice for
phase disturbance determines the extent of the field
absorption that is required. In Fig. 9 we can see that
the distance of both zero intensity points from the
region enclosed by K 12 circle is minimized for However, the inclusion of a certain degree of field ab-
this dual CPI configuration, hence reducing the
amount of field absorption that is required, whereas
the symmetry ensures the same light throughput and
signal-to-noise ratio for the two CPIs.
The final point to note from the phasor chart in Fig.
9 is that the filter term B is not actually required for
the operation of the CPI while we operate using the
high fringe accuracy requirement of K 12. With
B 1 there is no attenuation in the central filtering
region which need act only as a phase-shifting ele-
ment. The phasor chart illustrates this point clearly
showing that, while working within the confines of
K 12, we can achieve the mapping of the zero whole 2 phase circle and demonstrated through use
intensity point of the quadratic intensity scale any-
where on the unit circle solely by varying the two
filter parameters A and . This is generally appli-
cable for any linear phase-to-intensity mapping in a
CPI.
6. Conclusion
We have presented an analytical framework and pro-
vided a complete description for the design and opti-
mization of CPIs based on spatial filtering in the
Fourier plane. We have shown that our method can
be applied successfully to the interpretation and sub-
sequent optimization of a number of different, com-
monly used CPI architectures. We have proposed
that a considerable improvement in CPI design and
the interpretation of experimental results should
arise from the fact that we have a detailed treatment
of the profile of the SRW. To our knowledge, this
paper constitutes the first analytic derivation of the
effect of the curvature of the SRW in CPI systems.
We have derived optimal conditions for fringe accu-
racy, taking account of this curvature, and have
shown that our results agree well with empirical re-
sults from the literature.
Our analytical approach makes it possible to char-
acterize any CPI in terms of a combined filter param-
eter that places all CPI filters in the same phase-space
domain, drastically simplifying the comparative anal-
ysis of different CPI types. We compared a range of
well-known CPI types using the filter space plots that
we developed and indicate how their performance
might be improved. Using the criteria of high fringe
accuracy, high visibility, and peak irradiance as sug-
gested by Anderson,18 we have shown that it is possible
to optimize a CPI system for operation with a given
280 APPLIED OPTICS Vol. 40, No. 2 10 January 2001
dynamic range of phase distribution at the input. We
prioritized these criteria for high fringe accuracy CPI
absolute photon flux in the optical system is not of
critical importance, the visibility is the term that
serves as the best comparative indicator of CPI perfor-
developed in this paper, we have shown that the loss-
a variety of filtering applications. Thus for the oper-
ation of a CPI in a photon-limited regime, one should
always seek to remain on the lossless operating curve.
sorption becomes increasingly necessary for large-
scale input phase perturbations if the visibility is to be
maximized.
We applied our analysis to the extension of the
linearity of the phase-to-intensity mapping in a CPI
and have shown that when the Henning method19,20
is used as a starting point, it is possible to improve
the linearity of some currently applied systems. We
conclude that this generalization of the Henning
method should offer considerable practical improve-
ments. We also discussed the extension of linear,
unambiguous phase-to-intensity mapping to the
of phasor charts that this can be achieved by the
operation of two CPI systems in parallel. We have
also shown that, for high fringe accuracy and linear
phase-to-intensity mapping conditions, the on-axis
amplitude damping term is superfluous, thus simpli-
fying filter design.
In this paper we developed an analysis that can be
used for the optimization of any given CPI based on
spatial filtering, and when used in conjunction with
the previously developed graphical phasor chart
method, we now have a complete approach to the
qualitative and quantitative analysis of the CPI.
Appendix A: Application of the Graphical Phasor Chart
Method to the Analysis of Common Path
Interferometers
It is often a complicated task to obtain an overview of
the output from a CPI like that shown in Fig. 1 for a
given input phase distribution and the set of filter
parameters A, B, . In this appendix we give a
brief description of a graphical phasor chart method
for CPI analysis that we previously developed15 and
summarize how to apply this technique. We explain
how this original phasor chart can be reconfigured for
use with the combined filter parameter C that we
introduced in Section 4. We show how the two types
of phasor chart can be used with the rigorous math-
ematical treatment developed in this paper as pow-
erful visualization tools in the analysis of a CPI.
The primary role of the graphical phasor chart is to
illustrate the relationship between the intensity at
the output of the interferometer and the size of the
input phase perturbation, in other words, the phase-
to-intensity mapping for a given set of filter param-
eters. The basic framework of the original phasor

Fig. 10. Basic layout of the graphical phasor chart for mapping an
input phase to an output intensity I for a CPI with the Fourier
filter parameters A, B, . The key elements of the chart are
labeled for clarity.
chart is shown in Fig. 10 where the individual ele-
ments are labeled for clarity. We now briefly outline
their role in use of the chart.
The input phase deviation within the range 0 2
is given on the unity phase circle. The set of eight
circles on the right-hand side of the phasor chart rep-
resents the range of values for K from 18 to 1. As
described in Section 4, the term K represents a com-
bination of properties relating to the input phase dis-
tribution and the relationship between the radius of
the Fourier filter and that of the mainlobe of the Airy
function in the Fourier plane. In short, it describes
the amplitude that, when focused in the Fourier plane,
is incident on the central filter region generating the
SRW at the output. Finally, the quadratic intensity
scale shown at the bottom of Fig. 10 is the method by
which we map the input phase value to the intensity at
the output. The lower end of the scale is fixed in place
on the chart at a position determined by the three filter
parameters A, B, , and the intersection of the inten-
sity scale with the unity phase circle gives the phase-
to-intensity mapping. The effect of the amplitude
transmission filter parameters A and B can be under-
stood in terms of their effect on the position at which
the zero point of the quadratic intensity scale is fixed
labeled fix point in Fig. 11. Adjusting these param-
eters has the effect of shifting the zero point of the
intensity scale radially by the factor BA1 with respect
to the corresponding K circle along a line of the
angular position, which is given by the third filter pa-
rameter .
A simple example of the application of the phasor
chart serves to demonstrate how it is used. The first
step is to identify the appropriate value of K for the
situation we want to evaluate. In the example
shown in Fig. 11 we have K 12 and 2.
The center point of this K circle is marked on the
chart, and the angle is measured about this point.
The amplitude transmission terms A and B have the
Fig. 11. Application of the phasor chart for K 12 and
2 with a radial scaling for the zero point of the quadratic inten-
sity scale determined by the amplitude transmission terms A
12 and B 1. When the zero point of the intensity scale is fixed
at the desired point, the phase-to-intensity mapping is determined
from the intersection of the quadratic intensity scale with the unity
phase circle.
values 12 and 1, respectively, that effectively scale
the radius of the K circle by a factor of BA1 2.
This scaling serves to fix the zero point of the qua-
dratic intensity scale at the required position. Fi-
nally, the intersection of the quadratic intensity scale
with the unity phase circle occurs at the points , I.
In most of this paper we are concerned with the
combined filter parameter, C BA1 expi 1,
obtained from Eq. 14, and ideally we would like to
modify the graphical phasor chart to work with this
parameter. Therefore the original phasor chart can
be modified to operate directly with this combined
and complex filter parameter C instead of the basic
filter parameters A, B, . This modified phasor
chart is illustrated in Fig. 12. The most noticeable
feature in this modified chart is that all the K
circles are now arranged concentrically within the
unity phase circle. In Fig. 12 it can be seen that the
scale of the unity phase circle has the same orienta-
tion as in the original chart. The phase of the com-
bined filter parameter C is measured about the same
Fig. 12. Modified phasor chart that was adapted to work directly
with the two components of the combined filter parameter: C
and C.
10 January 2001 Vol. 40, No. 2 APPLIED OPTICS 281

Fig. 13. Application of the modified phasor chart to a CPI with
the same zero intensity point as that described by the phasor chart
in Fig. 11.
center point as , but the scales are shifted by 180
the axis labels on the modified chart give both phase
values with first. The same quadratic intensity
scale is used and is given in terms of the basic filter
parameter A as in the original chart. If required,
this can be related to the combined filter parameter C
by use of relation 17.
In Fig. 13 we show how the modified phasor chart
can be applied to analyze graphically the same sys-
tem as used in the example illustrated by Fig. 11. In
this case, the parameter C gives the phase describ-
ing the position for the zero point of the intensity
scale on the modified chart. The radial scaling for
the position of the zero intensity point simply de-
pends on the factor C, resulting in a radial param-
eter R in the coordinate system given by R CK .
Thus we find that C 5 and C 117 yields the 21.
same fix point location as in Fig. 11, which can be
described by the polar coordinate set R, C
52, 117. Because the fix point location is the
same in both types of phasor chart, it can therefore be
transferred directly between them. In most cases
we find it useful first to locate the intensity scale zero
point by use of a modified phasor chart as shown in
Fig. 12 and then transfer the coordinates directly
onto the original phasor chart shown in Fig. 10. It
should also be noted that to find the phase-to-
intensity mapping at the intersection between the
quadratic intensity scale and the unit circle follows
exactly the same procedure in both phasor charts.
This research is funded as part of an award from
the Danish Technical Scientific Research Council.
References
1. J. W. Goodman, Introduction to Fourier Optics, 2nd ed.
McGraw-Hill, San Francisco, Calif., 1996.
2. F. Zernike, How I discovered phase contrast, Science 121,
345349 1955.
3. H. H. Hopkins, A note on the theory of phase-contrast imag-
es, Proc. Phys. Soc. London Sect. B. 66, 331333 1953.
4. S. F. Paul, Dark-ground illumination as a quantitative diag-
nostic for plasma density, Appl. Opt. 21, 25312537 1982.
5. R. C. Anderson and S. Lewis, Flow visualization by dark
central ground interferometry, Appl. Opt. 24, 3687 1985.
6. M. P. Loomis, M. Holt, G. T. Chapman, and M. Coon, Appli-
282 APPLIED OPTICS Vol. 40, No. 2 10 January 2001
cations of dark central ground interferometry, paper AIAA
91-0565, presented at the Twenty-Ninth Aerospace Sciences
Meeting, Reno, Nev., January 1991 American Institute of
Aeronautics and Astronautics, New York, 1991, pp. 1 8.
7. D. Malacara, Optical Shop Testing, 2nd ed. Wiley, New York,
1992, pp. 302305.
8. A. K. Aggarwal and S. K. Kaura, Further applications of point
diffraction interferometer, J. Opt. Paris 17, 135138 1986.
9. M. Born and E. Wolf, Principles of Optics, 6th ed. Pergamon,
New York, 1980, pp. 426 427.
10. C. A. Mack, Phase contrast lithography, in OpticalLaser
Microlithography VI, J. D. Cuthbert, ed., Proc. SPIE 1927,
512520 1993.
11. Y. Arieli, N. Eisenberg, and A. Lewis, Pattern generation by
inverse phase contrast, Opt. Commun. 138, 284 286 1997.
12. J. Gluckstad, Phase contrast image synthesis, Opt. Com-
mun. 130, 225230 1996.
13. J. Gluckstad, L. Lading, H. Toyoda, and T. Hara, Lossless
light projection, Opt. Lett. 22, 13731375 1997.
14. J. Gluckstad, Pattern generation by inverse phase contrast
comment, Opt. Commun. 147, 16 19 1998.
15. J. Gluckstad, Graphic method for analyzing common path
interferometers, Appl. Opt. 37, 8151 8152 1998.
16. J. Gluckstad and P. C. Mogensen, Reconfigurable ternary-
phase array illuminator based on the generalised phase con-
trast method, Opt. Commun. 173, 169 175 2000.
17. P. C. Mogensen and J. Gluckstad, Dynamic array generation
and pattern formation for optical tweezers, Opt. Commun.
175, 75 81 2000.
18. C. S. Anderson, Fringe visibility, irradiance, and accuracy in
common path interferometers for visualization of phase dis-
turbances, Appl. Opt. 34, 7474 7485 1995.
19. H. B. Henning, A new scheme for viewing phase contrast
images, Electro-Opt. Syst. Des. 6, 30 34 1974.
20. G. O. Reynolds, J. B. Develis, G. B. Parrent, Jr., and B. J.
Thompson, The New Physical Optics Notebook: Tutorials in
Fourier Optics SPIE, Bellingham, Wash., 1989, Chap. 35.
J. Gluckstad, Image decrypting common path interferome-
ter, in Optical Pattern Recognition X, D. P. Casasent and T.
Chao, eds., Proc. SPIE 3715, 152159 1999.
22. C. R. Mercer and K. Creath, Liquid-crystal point-diffraction
interferometer for wave-front measurements, Appl. Opt. 35,
16331642 1996.
23. C. Koliopoulus, O. Kwon, R. Shagam, J. C. Wyant, and C. R.
Hayslett, Infrared point-diffraction interferometer, Opt.
Lett. 3, 118 120 1978.
24. H. Kadano, M. Ogusu, and T. Asakura, Phase shifting com-
mon path interferometer using a liquid-crystal phase modula-
tor, Opt. Commun. 110, 391 400 1994.
25. R. N. Smartt and W. H. Steel, Theory and application of
point-diffraction interferometers, Jpn. J. Appl. Phys. 14, 351
356 1975.
26. W. P. Linnik, Ein einfaches interferometer zur prufung von
optischen systemen, Proc. Acad. Sci. USSR Sect. Appl. Phys.
1, 208 211 1933.
27. P. M. Birch, J. Gourlay, G. D. Love, and A. Purvis, Real-time
optical aberration correction with a ferroelectric liquid-crystal
spatial light modulator, Appl. Opt. 37, 2164 2169 1998.
28. C. R. Mercer and K. Creath, Liquid-crystal point-diffraction
interferometer, Opt. Lett. 19, 916 918 1994.
29. A. van den Bos, Aberration and the Strehl ratio, J. Opt. Soc.
Am. A 17, 356 358 2000.
30. G. D. Love, N. Andrews, P. M. Birch, D. Buscher, P. Doel, C.
Dunlop, J. Major, R. Myers, A. Purvis, R. Sharples, A. Vick, A.
Zadrozny, S. R. Restaino, and A. Glindemann, Binary adap-
tive optics: atmospheric wave-front correction with a half-
wave phase shifter, Appl. Opt. 34, 6058 6066 1995;
addenda 35, 347350 1996.