Professional Documents
Culture Documents
Optical Design Applying The Fundamentals
Optical Design Applying The Fundamentals
Optical Design Applying The Fundamentals
Riedl, Max J.
Optical design : applying the fundamentals / Max J. Riedl.
p. cm. -- (Tutorial texts series)
Includes bibliographical references and index.
ISBN 978-0-8194-7799-6
1. Optical instruments--Design and construction. I. Title.
QC372.2.D4R54 2009
681'.4--dc22
2009022688
Published by
SPIE
P.O. Box 10
Bellingham, Washington 98227-0010 USA
Phone: 360.676.3290
Fax: 360.647.1445
Email: Books@spie.org
www.spie.org
The content of this book reflects the thought of the author(s). Every effort has been made
to publish reliable and accurate information herein, but the publisher is not responsible
for the validity of the information or for any outcomes resulting from reliance thereon.
Tutorial Texts have grown in popularity and in the scope of material covered
since 1989. They no longer necessarily stem from short courses; rather, they are
often generated by experts in the field. They are popular because they provide a
ready reference to those wishing to learn about emerging technologies or the latest
information within their field. The topics within the series have grown from the
initial areas of geometrical optics, optical detectors, and image processing to include
the emerging fields of nanotechnology, biomedical optics, fiber optics, and laser
technologies. Authors contributing to the TT series are instructed to provide
introductory material so that those new to the field may use the book as a starting
point to get a basic grasp of the material. It is hoped that some readers may develop
sufficient interest to take a short course by the author or pursue further research in
more advanced books to delve deeper into the subject.
The books in this series are distinguished from other technical monographs and
textbooks in the way in which the material is presented. In keeping with the tutorial
nature of the series, there is an emphasis on the use of graphical and illustrative
material to better elucidate basic and advanced concepts. There is also heavy use of
tabular reference data and numerous examples to further explain the concepts
presented. The publishing time for the books is kept to a minimum so that the books
will be as timely and up-to-date as possible. Furthermore, these introductory books
are competitively priced compared to more traditional books on the same subject.
It is my goal to maintain the style and quality of books in the series and to
further expand the topic areas to include new emerging fields as they become of
interest to our reading audience.
James A. Harrington
Rutgers University
vii
Contents
Preface…………….. .......................................................................... xv
ix
Chapter 4 Transition from a Thin Lens to a Thick Lens ........... 33
x
Chapter 7 From an Air-Spaced Doublet to a Triplet .................. 65
xi
11.3 Cassegrain Version with a Maksutov-Mangin Mirror
Combination for the LWIR Region ....................................... 98
11.3.1 Pegel diagram ........................................................... 98
xii
15.10 Gregorian Microscope Objective ........................................... 130
15.11 Two Schwarzschild Objectives .............................................. 131
15.11.1 Prescription for the classic configuration
for an object at infinity .......................................... 133
15.11.1.1 Brief historic remark about the golden ratio ........ 134
15.11.2 Prescription for the inverted configuration
for an object at infinity .......................................... 134
15.12 Solid Microscope Objective ................................................... 136
Index………………… ........................................................................169
xiii
Preface
This book is written for engineers and scientists who have some experience in the
field of optics and want to know more about the details and derivations of
equations used in optical design. Such knowledge is especially valuable in the
layout stages of an optical system, when the question is “Where shall I begin?”
The other question may be “How come?” instead of just using a given equation.
The book begins with the derivation of the fundamental law of geometrical
optics, Snell’s law of refraction, and states the paraxial ray trace equations. The
following discussions are organized by subjects, starting with a thin lens and
progressing to increasingly more sophisticated components and multi-element
systems. Each subject is covered in depth to provide a good understanding for
performance and limitations. The often ignored effects of plane-parallel plates
are included as a separate subject.
While the text is based on general optical laws, special emphasis has been
placed on the two major infrared regions, the mid-wave (MWIR) and the long-
wave (LWIR). This is particularly apparent in the discussion about diffractive
hybrids, which have found their place in these long wavelengths areas for the
correction of chromatic aberrations and athermalization. Comments relating to
single-point diamond turning have been included because this process is
predominantly used to produce optical elements for the infrared regions.
The final subject gradually leads the reader from a single element as an
imaging objective for the visible spectrum, the historic Höegh, to a four-element
anastigmat. This is done by successively adding and shaping elements and
selecting suitable glasses for aberration reduction.
Finally, I thank Tim Lamkins, Scott Schrum, and Gwen Weerts of SPIE for
their special support and editorial assistance.
Max J. Riedl
July 2009
xv
Chapter 1
Law of Refraction: The Foundation of
Geometrical Optics
1.1 Introduction
Snell’s law of refraction is the fundamental law that governs geometrical optics.
We begin, therefore, with the proof of this basic rule, as it has been verified by
Fermat. We then demonstrate how this surprisingly simple law can be applied to
graphical ray tracing. With the equations for paraxial ray tracing, we provide the
tool required for the initial optical design phase. These equations are sufficient to
determine the third-order aberrations, which will be used throughout the book.
where n and n′ are the indices of refraction of the media before and after
refraction. The angle of incidence is i, and the angle of the ray is i′, relative to the
normal, after refraction.
1.2.2 Derivation
Travel time with reference to Fig. 1.1 is
PO OP'
t
v1 v2
and
b2 a x
2
h2 x 2
t ,
v1 v2
1
2 Chapter 1
where v1 is the speed of light in medium 1 before refraction, and v2 is the speed
in medium 2 after refraction. We differentiate this expression with respect to x
and set the result equal to zero.
dt x a x
0,
dx v1 h 2 x 2 v b 2 a x 2
2
which means
x
a x .
v1 PO v2 PO'
Law of Refraction: The Foundation of Geometric Optics 3
With
vvacuum v1
n' ,
vmedium v2
we finally obtain
1
sin i' sin i ,
n'
This simple procedure covers flat, spherical, and aspheric surfaces. It also
serves for calculations of chromatic aberrations because the index of refraction is
a function of wavelength.
Draw line 1 from center C (center of the first surface radius) through point
P1, the point where the entering ray meets the first surface. Draw arcs 2 and 3
around point P1. The scales of the radii are chosen in proportion to the media
indices before and after the refraction. Our lens is surrounded by air with n1 1 ,
and the index of the lens material is n'1 n2 1.5 . The ratio between the radii 2
and 3 is therefore 1 to 1.5. Draw line 4 parallel to line 1 through the point where
arc 2 intersects the entering ray. From the intersection of arc 3 and line 4, draw a
line through point P1 and extend it until it crosses the second lens surface. This
point is marked P2 in the figure. Repeat the steps for the second lens surface to
find the direction of the exiting ray. The focus location F of the lens can be
established by tracing a ray near the optical axis, the paraxial region.
The lateral distance from the optical axis and the point where the marginal
ray passes through the focal plane is the transverse spherical aberration. The
distance from the crossover of the ray on the optical axis is the longitudinal
spherical aberration. While this approach may not appear to be too practical, it
gives a good insight and demonstrates the principle of refraction at lens surfaces.
Figure 1.4 serves for additional explanation of the relationships.
Figure 1.4 Principle of graphical ray tracing, which is also suitable for aspheric
surfaces.
i3 i5 i7
sin i i ...... . (1.2)
3! 5! 7!
sin i i , (1.3)
n'i' ni . (1.4)
This is accurate to the sixth decimal place if the angles are not larger than about
41 minutes of arc. Ray tracing based on, but not limited to, this narrow region is
referred to as paraxial ray tracing. The simple paraxial ray-trace equations are
exceedingly useful in lens design. The image location is determined with these
equations. Deviations from the paraxial image position are the measures for
aberrations. By including the second term of the sine series, the so-called third-
order aberrations, known as Seidel aberrations, can be determined. This will be
demonstrated throughout the following chapters.
6 Chapter 1
Figure 1.5 Symbols and sign conventions for paraxial ray-trace equations: (a) an
axial ray and (b) an oblique ray.
Opening
The distance of an object from the first lens surface is l1. For an axial ray,
y1
u1 , (1.5)
l1
Law of Refraction: The Foundation of Geometric Optics 7
u1 0, and y1 h . (1.7)
y1 10
u1 0.1 .
l1 100
Equations (1.8) and (1.9) are repeatedly applied for the i surfaces in the
optical train.
Closing
For the axial ray,
yi
l'i , (1.10)
u 'i
u1 h'
m . (1.12)
u'i h
Example
Given: A lens with the surface radii R1 = 50 (c1 =1/50 = 0.02) and R2 = −50 (c2 =
−0.02). t'1 10 .
Trace an axial ray with y1 20, and an oblique ray with h 12 and y1 10 .
1. Axial ray
y1 20
u1 0.1
l1 200
0.1
u'1 0.066667
1.5
y2 19.333333
l'2 65.909091
u'2 0.293333
u1 0.1
m 0.340909
u'2 0.293333
2. Oblique ray
h y1 25 10
u1 0.075
l1 200
0.175
u'1 0.116667
1.5
An oblique ray from the edge of the object through the center of the aperture
stop is called the chief, or principle, ray. An axial ray from the axial point of the
object through the edge of the aperture stop is the marginal axial ray. To
determine the third-order aberration contributions, only these two paraxial rays
have to be traced. Ray-trace equations in addition to the ones presented here can
be found in Ref. 2.
References
1. J. R. Meyer-Arendt, Introduction to Classical and Mirror Optics, Prentice
Hall, Inc., Englewood Cliffs, New Jersey (1984).
2. W. J. Smith, Modern Optical Engineering, 4th Ed., McGraw-Hill, New York
(2008).
Chapter 2
Best Shape for a Thin Lens
2.1 Concept of Thin Lens
The “thin lens” concept is an extremely useful one for quick preliminary layout
calculations and initial performance analyses. “Thin” means that the thickness of
the lens is ignored in the calculations.
The power of a lens is the reciprocal of its focal length f. The reciprocal of
a surface radius R is called the surface curvature c. With t being the lens
thickness and n the index of refraction of the lens material, we can state
1
n 1 c1 c2
n 1 c c t . (2.1)
1 2
f n
n 1 c1 c2 n 1 c . (2.2)
In this equation, c is called the net curvature and is the difference between the
front and rear surface curvatures, i.e., c c1 c2 . This indicates that
1
Figure 2.1 Bending a lens affects its spherical aberration but not its focal length.
Shown is the effect for a germanium lens with an index of n = 4.
11
12 Chapter 2
the power of a thin lens remains the same as long as the net curvature is kept
constant. Changing the curvatures but maintaining their difference is called lens
bending. This is a powerful tool because the shape of a lens primarily influences
its spherical aberration, as indicated in Fig. 2.1. The relation between lens shape
and spherical aberration of a thin germanium lens for an object located at infinity
is plotted. The ratio of the first curvature c1 and the net curvature c is known as
the shape factor K. We plot spherical aberration as a function of the lens shape K
and identify the surface radii of the lens. Because germanium has an index of
refraction of n = 4 at a wavelength = 10 µm, the result for the best shape is
amazingly simple and easy to remember. The front surface radius is equal to the
focal length of the lens, and the radius of the rear surface is 1.5 times the focal
length.
Figure 2.3 The marginal ray penetrates the image plane a distance away from
the optical axis. This lateral distance is called the transverse spherical aberration
TSC. See also Figure 2.4.
Best Shape for a Thin Lens 13
To derive the general equations for the best shape surface radii, when the
object is located at any position, we begin the derivation with the expression for
the third-order thin-lens transverse spherical-aberration contribution for any
conjugate2:
y4
TSC
u'k
G1c 3 G2c 2c1 G3c 2v G4cc12 G5cc1v G6cv 2 . (2.3)
In Eq. (2.3), the G-sums are functions of the optical material’s index of refraction
n. They are
n 2 n 1 2n 1 n 1 , 3n 1 n 1
G1 , G2 G3
2 2 2
G4
n 2 n 1 , G5
2 n 1 n 1
, G6
3n 2 n 1
2n n 2n
G7
2n 1 n 1 , and G8
n n 1
.
2n 2
dTSC y 4
dc1
u'k
G2c 2 2G4cc1 G5cv .
To find the optimum, we set this expression to zero and solve for c1,
G2c G5v
c1 . (2.4)
2G4
n 2n 1 2 n 1
c1 v, (2.5)
2 n 2 n 1 n 2
14 Chapter 2
It is interesting to observe that the first term of Eq. (2.5) states the surface
curvature c1 for an object at infinity because, for that case, v = 0.
To find the curvature for the second surface, we use the relation
c2 c1 n 1 .
n 2n 1 2 n 1
c2 v ,
2 n 2 n 1 n 2 n 1
which reduces to
n 2n 1 4 2 n 1
c2 v. (2.6)
2 n 2 n 1 n 2
Here, too, the first term is for an object at infinity. Considering that the inverse
object distance can be stated by m / (m 1) , where m is the magnification,
we insert this relation into Eqs. (2.5) and (2.6), to form
n 2n 1 2 n 1 m
c1 (2.7)
2 n 2 n 1 n 2 m 1
and
n 2n 1 4 2 n 1 m
c2 . (2.8)
2 n 2 n 1 n 2 m 1
To demonstrate the usefulness of these expressions, we analyze a lens with a
focal length of 100 mm 1/f 0.01 mm -1 , made from germanium with
n = 4. We choose the magnification to be m = –2.
1 1
R2 225 mm .
c2 0.0044444
m 2
v 0.01 0.0066667 .
m 1 2 1
For a relative aperture of f/2, the semi-aperture y = 25 mm. The final slope angle
u'k v y 0.01 0.0066667 25 0.0833333 .
Now, we can insert all of these numbers into Eq. (2.3) and find the minimum
transverse third-order spherical aberration for this thin lens to be
254
TSC [24 0.0033333 13.5 0.0033332 (0.001111)
0.083333
19.5 0.0033332 0.006667 2.25 0.003333 (0.001111)2
7.5 0.003333 (0.001111) 0.006667 5.25 0.003333 0.006667 2 ]
= 0.994859.
TSC 1 mm .
16 Chapter 2
y 3n 4n 1 2 4 n 12 v v .
TSC (2.9)
8 n 1 n 2 v
2
253 0.01 4
TSC
8 4 1 4 2 0.01 0.006666
2
and
2 n 2 n 1
R1 f, (2.10)
n 2n 1
2 n 2 n 1
R2 f. (2.11)
n 2n 1 4
y 3 n 4n 1 2 n 4n 1 f
TSC . (2.12)
8 n 1 n 2 64 n 1 n 2 f / #
2 2 3
The blur spot is half the size of TSC after refocusing from the paraxial focus
by 3 / 2 TSC f / # . Applied to our f/2, 100-mm focal-length germanium
lens set for infinity, we obtain
2 n 2 n 1 2 4 2 4 1 36
R1 f 100 100 100 mm
n 2n 1 4 2 4 1 36
and
2 n 2 n 1 2 4 2 4 1 36
R2 f 100 100 150 mm .*
n 2n 1 4 4 2 4 1 4 24
n 4n 1 f 4 15 100
TSC 0.217014 mm .
64 n 1 n 2 f /# 64 9 6 23
2 3
The minimum blur spot size is Bmin TSC 2 0.217 2 0.1085 mm when
refocused by 3 2 TSC f /# 1.5 0.217 2 0.651 mm. The best
image location is, therefore, 100 – 0.651 = 99.349 mm from the last lens surface.
Details are shown in Fig. 2.4.
Figure 2.4 Location and size of the minimum blur spot due to spherical
aberration, which is called “the circle of least confusion.”
*
Also see Fig. 2.1.
18 Chapter 2
f
Bchrom , (2.13)
2V f / #
where V is the inverse relative dispersion of the lens material (also known as
Abbe number, named after Ernest Abbe, who introduced this concept). In the
visible spectrum, V nd 1 nF nC , where the indices of refraction refer to
the Fraunhofer absorption lines. In the infrared regions, V nM 1 nS nL ,
with reference to the middle, short, and long wavelengths of the covered spectral
band. For our germanium lens in the LWIR region (8–12 µm), V ≈ 800.
Therefore, Bchrom 100 2 800 2 0.03125 mm for our lens. Note how mild
the chromatic aberration is with such a large Abbe number. In the visible
spectrum, the V values are lower by more than a factor of 10. As will be
discussed in Chapter 4, this attribute is the main justification for correcting
chromatic aberration in the infrared regions with diffractive phase profiles.
Figure 2.5 Computer ray trace confirms validity of thin-lens calculations for on-
axis aberrations (spherical and chromatic).
Best Shape for a Thin Lens 19
Figure 2.6 Thin lens with added field for object at infinity. The principal ray is an
oblique ray that goes through the center of the aperture stop, which is located
here at the lens.
2.3.3 Coma
Figure 2.6 shows the layout of a thin lens for an object located at infinity and
identifies the image height h′ and the half-field angle up.
If we insert the curvatures c and c1 from the thin lens (shaped for minimum
spherical aberration) and the expressions for the G-sums G5 and G8, we get
up f
CC . (2.15)
16 n 2 f/#
2
With a half field angle of up = 0.1 radian, the sagittal coma blur for our lens
becomes
0.1 100
Bcoma CC 0.026042 mm .
16 4 2 22
2.3.4 Astigmatism
Another off-axis aberration is astigmatism, which introduces a blur spot size of 1:
u 2p f
Bastig 2TAC . (2.16)
2 f / #
20 Chapter 2
Figure 2.8 Balanced image position for equal blur spot sizes of the f/2, 100-mm
focal-length germanium lens for the LWIR spectrum with a half field of 5.73 deg.
Best Shape for a Thin Lens 21
0.12 100
Bastig 0.250 mm.
2 2
This does not take into account the effects of field curvature. This can be seen in
Fig. 2.7, which is from a real ray trace of the thick lens. The off-axis blur is
somewhat larger than the predicted 0.42 mm. By slightly refocusing, the total
blur size is about 0.460 mm in diameter. The choice of best position depends on
the application. A 0.460-mm square detector element is large enough to accept all
the energy from a point source within the stated field of view.
For completeness, the best focus position relative to the paraxial focal plane
for the example is illustrated in Fig. 2.8.
References
1. M. J. Riedl, Optical Design Fundamentals for Infrared Systems, SPIE Press,
Bellingham, Washington (2001).
2. W. J. Smith, Modern Optical Engineering, 4th Ed., McGraw-Hill, New York
(2008).
Chapter 3
Best Shapes for Multiple Thin Lenses,
Aspherizing, and the Natural Stop
Position
3.1 Introduction
Expanding on the previous case for a single lens in Chapter 2, we develop
expressions for elements of a multiple lens arrangement, made from the same
material and individually bent for minimum spherical aberration.
n 2n 1 4 n 2 1 j 1 2 n 1
c j1 v. (3.1)
2i n 2 n 1 n 2
In this equation and in the following, j is the element number, and i is the total
number of elements in the set.
23
24 Chapter 3
n 2n 1 4 4 n 2 1 j 1 2 n 1
c j2 v. (3.2)
2i n 2 n 1 n 2
The minimum spherical aberration is
y 3n
TSCi
8i 2 n 1 n 2 v
2
2 2 j i
4n 1 2 4i 2 n 1 v v 3 j j 1 . (3.3)
i j 1
These expressions reduce to the forms for a single lens if we set j = i = 1. With v
= 0, the lens will be focused for an object at infinity.
Table 3.1 lists the radii and the spherical aberrations for an f/2, 100-mm-
focal-length lens, made from germanium with 1, 2, and 3 elements for an object
at infinity. Notice the change in shapes with the increase of elements for an
object at infinity.
From Table 3.1, it can be seen that the single element is afflicted with a
negative TSC, which means that it has undercorrected spherical aberration. (The
marginal ray crosses the optical axis ahead of the paraxial image point.) With two
elements, the sign changes and the lens has overcorrected spherical aberration.
(The marginal ray crosses the optical axis behind the paraxial image). Three
elements amplify this trend. This means that the high index of germanium
“overshoots” the correction with only two elements.
Let us find the index for which the two-element configuration is zero, when
the object is located at infinity. For those conditions, Eq. (3.3) reduces to
y 3 n2 4n 1 16 n 12 1 .
TSC2 (3.4)
32 n 1 n 2
2 4
n = 2.5. That is close to the index for amorphous material transmitting infrared
radiation (AMTIR-1) for the mid-wave infrared (MWIR) with n4 2.514 , and for
the long-wave infrared (LWIR) where n10 2.498 . Also close to the desired value
is zinc selenide with n4 2.433 in the MWIR region, and n10 2.406 for the
LWIR band.
To find a material for which the TSC is zero when the system consists of
three elements, we proceed as above and get 4n 1 36 n 1 8 / 27 . This
2
leads to n = 1.75. That is the index for the rare earth glass LAFN7, for example,
with nd 1.7495 . For the LWIR spectrum, cesium iodide (CsI) has n10 1.74 ,
which is close to the desired value.
Table 3.2 on the next page shows the radii and the spherical aberrations for
lenses with a magnification of m 2 . The negative sign indicates that the image
is reversed, as demonstrated in Fig. 1.5(b). Notice how the shapes change for an
object at a finite distance.
y 4c
TSC
u'
G1c 2 G2cc1 G4c12 , (3.5)
n 2n 1
c1 ,
2 n 2 n 1
c ,
n 1
u ' y ,
and G1 , G2 , G4 , as identified in Sec. 2.2, and the object is at any position, Eq.
(3.5) becomes
Best Shapes for Multiple Thin Lenses, Aspherizing, and the Natural Stop Position 27
y 3 n 4n 1 2
TSCmin spher . (3.6)
8 n 1 n 2
2
To get rid of spherical aberration, we aspherize one surface of the lens and set the
contribution from the asphere equal to the spherical aberration contribution from
the lens with spherical surfaces, but opposite in sign. In other words, we write
TSCasphere
n ' n fy 3 c3 , (3.7)
2
where is the conic constant, n and n′ are the indices of the media before and
after refraction, f is the focal length of the element, y is the semi-diameter of the
lens, and c is the curvature of the surface to be aspherized. This leads to a conic
constant for the first surface of
2 R13TSC
1 . (3.8)
n 1 f y13
If the second surface is to be aspherized, the conic constant for that surface is
2 R23TSC
2 . (3.9)
n 1 f y23
Example
For a 100-mm-focal-length lens (made from silicon) with an f/2 speed, with n =
3.4,
0.01
c2 c1 0.010231 0.006065 .
n 1 3.4 1
Setting this expression to zero yields the curvature of the first surface as
4G8c
c1 . (3.11)
G5
Best Shapes for Multiple Thin Lenses, Aspherizing, and the Natural Stop Position 29
n2
c1 . (3.12)
n 2
1
R1
n 2
1
f. (3.13)
n2
n n 1 1
c2 c1 c (3.14)
n2 1
and
R2
n 2
1
f . (3.15)
n n 1 1
y 4c
TSC
u'
G1c 2 G2cc1 G4c12 . (3.16)
y 3 n3
TSCcoma=0 . (3.17)
2 n 2 1 f 2
2
To apply this formula to our silicon lens with y = 25, n = 3.4, and f = 100, we get
253 3.43
TSCcoma=0 0.275359 .
2 3.42 1 1002
2
30 Chapter 3
We eliminate spherical aberration again with an asphere. This time we have to set
R1
n 2
1
f
3.4 2
1
100 91.349481 ,
2
n 3.42
2 91.3494813 0.275359
κ1 0.111948 .
3.4 1 100 253
This equation indicates the shape of a prolate ellipsoid. The second surface radius
is
R2
n 2
1
f
3.4 2
1
100 147.486034 .
n n 1 1 3.4 3.4 1 1
If we desire to aspherize the second surface instead of the first, the conic constant
for that surface is
2 147.4860343 0.275359
κ2 0.471140 ,
3.4 1 100 253
which identifies again the shape of an oblate ellipsoid.
Remark
The same exercises can be repeated for the other monochromatic aberrations and
extended to include the effects of the aperture stop location. Such exercises give
much insight into the behavior of optical elements. We limit ourselves to the case
of a single lens, shaped for minimum spherical aberration, and determine the stop
location for which coma vanishes. This stop location is called the natural stop
position.
Best Shapes for Multiple Thin Lenses, Aspherizing, and the Natural Stop Position 31
yp
CC* CC TSC , (3.18)
y
with h' u p f , and c 1/ n 1 f . The curvature of the first surface for a
best-shaped lens (one with minimum spherical aberration) is
n 2n 1
c1 .
2 n 2 n 1 f
up f
CC . (3.20)
16 n 2 f / #
2
n 4n 1 f
TSC . (3.21)
64 n 1 n 2 f / #
2 3
From Fig. 3.1, it can be seen that y p u p lstop . (lstop is defined as negative
because it is measured from the first lens surface). Eliminating coma requires that
CC y pTSC y 0 . If the lens has been shaped for minimum spherical
aberration, inserting the relations from above leads to the simple expression for
the natural stop position,
4 n 1 y f / #
2
lstop . (3.22)
n 4n 1
200 n 1
2
lstop .
n 4n 1
Table 3.3 Natural stop position for three thin lenses shaped for minimum
spherical aberration, at which third-order coma vanishes.
We begin our exercise with a thin lens shaped for minimum spherical
aberration, add a reasonable thickness, and modify only the rear surface radius to
maintain the focal length of the lens. We then calculate the spherical aberration
for both cases and compare the results for three f/2, 100-mm-focal-length lenses.
The first lens is made from glass with an index of refraction of n = 1.5 for the
visible spectrum (VIS). The second is for the infrared mid-wave region (MWIR,
3–5 µm), made from silicon with n = 3.4. The third singlet is made from
germanium, n = 4, for the application in the infrared long-wave region (LWIR,
8–12 µm).
1 1 1
n 1 . (4.1)
f R1 R2
For a thick lens, where the thickness of the lens is accounted for, the equation is
stated as
1 1 1 n 1 t
n 1 , (4.2)
f R1 R2 nR1R2
33
34 Chapter 4
n 1 n 1 t nR1 f
R2 . (4.3)
n R1 n 1 f
We recall that the radius of a thin lens, shaped for minimum spherical
aberration, is
2 n 2 n 1
R1 f. (4.4)
n 2n 1
Inserting Eq. (4.4) into Eq. (4.3) leads to the rear radius of a thick lens, shaped
for minimum spherical aberration:
n 1 2 n 2 f 2n 1 t
R2 . (4.5)
n 2n 1 2 n 2
2 n 2 n 1 2 1.5 2 1.5 1
R1 f 100 58.3333 mm
n 2n 1 1.5 2 1.5 1
and
2 n 2 n 1 2 1.5 2 1.5 1
R2 f 100 350mm .
n 2n 1 4 1.5 2 1.5 1 4
We add a thickness of 10 mm and solve for the second surface radius using Eq.
(4.5):
n 1 2 n 2 f 2n 1 t
R2
n 2n 1 2 n 2
1.5 1 2 1.5 2 100 2 1.5 1 10
330.00 mm.
1.5 2 1.5 1 2 1.5 2
These results can be confirmed by using the ray trace equations [Eqs. (1.7) and
(1.8)] and solving for the curvature c2 . We demonstrate this for the thick lens.
Transition from a Thin Lens to a Thick Lens 35
0.085714286
u'1 u2 0.05714285 .
1.5
With the lens thickness of t '1 10 , the ray height at the second surface is
Recognizing that u'2 y1 f 10 100 0.1 , we rearrange Eq. (1.7) and
solve for the curvature of the second surface:
n 4n 1 f
TSCthin lens . (4.6)
64 n 1 n 2 f / #
2 3
The contributions for our three lenses are summarized in Table 4.1.
n2
TSC B i h' ,
n 1
2
n n (4.7)
whereby
n n' 1
B y u' i . (4.8)
2n' Inv
The optical invariant Inv n'2u'2 h' 0.25h' for our f/2 lenses. This changes Eq.
(4.8) to
2n n' 1
B y u' i .
n'
Table 4.1 Spherical aberrations for three thin singlets in different spectral
regions.
Table 4.2 Ray trace and third-order surface aberration contributions for the VIS-
lens with n = 1.5
Surface y t′ u′ i B TSC
1 25 10 -0.142857 0.428575 -4.761967 -0.8747
2 23.57143 -0.250000 -0.214278 -16.415545 -0.7537
Transition from a Thin Lens to a Thick Lens 37
As expected, the spherical aberration for the thick lens is somewhat less
because the added thickness contributes positive spherical aberration. Positive, or
overcorrected spherical aberration, exists when the marginal ray intercepts the
optical axis beyond the paraxial focal plane. Spherical aberration is called
negative, or undercorrected, when the marginal ray crosses the optical axis ahead
of the paraxial focal plane. This occurrence is illustrated in Fig. 4.1.
We follow the same procedure for the two IR lenses. The results are listed in
Table 4.3. The errors for these lenses are even smaller. This indicates that with
the higher indices of refraction for silicon and germanium, and the longer
wavelengths, the thin-lens aberration expressions are quite accurate in the
infrared regions.
Table 4.3 Transverse spherical aberrations for two singlets in the infrared
regions.
Reference
1. W. J. Smith, Modern Optical Engineering, 4th Ed., McGraw-Hill, New York
(2008).
Chapter 5
Achromats
5.1 Introduction
The progression of an achromat from the basic conventional form to a diamond-
turned aspheric hybrid for the infrared spectrum will be demonstrated in this
Chapter. We start with a cemented doublet for the visible spectrum, then change
to the mid-wave and long-wave infrared regions, and proceed with the rest of the
improvements relating to aberration reductions in the long-wavelength spectral
band. For this exercise, we elect a 100-mm focal length with an f/4 relative
aperture and add a field of ±2 deg.
Va
a for the front element, (5.1)
Va Vb
and
Vb
b for the rear element. (5.2)
Vb Va
Va and Vb are the mentioned Abbe numbers for the two elements.
Va Vb
na 1 ca and nb 1 cb ,
Va Vb Vb Va
from which we can extract the net curvature for the front element a:
39
40 Chapter 5
Va
ca . (5.3)
Va Vb na 1
The net curvature of the rear element b is
Vb
cb . (5.4)
Vb Va nb 1
Since the shapes of the elements in an achromat do not influence the
chromatic aberration, we make the first element equi-convex, i.e.,
ca1 ca 2 0.5ca . Further, we give the first surface of element b the same
curvature as the second surface of element a, i.e., cb1 ca 2 . With that,
Va
ca1 ca 2 . (5.5)
2 Va Vb na 1
Vb
cb ,
Vb Va nb 1
Vb
cb 2 ca 2 cb ca1 . (5.6)
Vb Va nb 1
For the front element a (crown) we take glass BK7 with na 1.517 and
Va 64.17. The rear element b (flint) is made from glass F2 with
nb 1.620 and Vb 36.37 . With that we get
64.17
ca1 0.01 0.022324 ,
2 64.17 36.37 1.517 1
36.37 0.01
cb 2 0.022324 0.001223 ,
36.37 64.17 1.62 1
Figure 5.1 Cemented achromat for the VIS spectrum (corrected for on-axis only)
after adding thicknesses and optimizing.
Figure 5.2 Blur-spot size of the optimized achromat for the visible spectrum.
42 Chapter 5
Figure 5.3 The encircled energy with reference to the diffraction limit.
TSC 0.0225 mm .
1
Figure 5.5 The presence of coma indicates the need for further correction. This
is also indicated in the modulation transfer function shown in Fig. 5.6.
44 Chapter 5
Figure 5.6 The modulation transfer function drops off drastically for the 2-deg off-
axis point.
5.2.2 Optimizing
We include the off-axis aberrations and computer-optimize the configuration
from above. The result is shown in Fig. 5.7.
and
As before, the curvatures are identified as follows: ca1 and ca 2 refer to the first
and second surfaces of the front element a, and cb1 and cb 2 classify the curvatures
of first and second surfaces of the rear element b.
The Ks are functions of the element curvatures, the inverse object distances, and
the familiar G-sums, which are, in turn, functions of the index of refraction of the
optical materials used for the elements, as we already know.
Substituting Eq. (5.9) into Eq. (5.8) leads to the quadratic equation
or
Aca21 Bca1 C 0 .
To solve for the first curvature of element a, we apply the standard form for a
quadratic equation:
B B 2 4 AC
ca1 , (5.10)
2A
where
and
We now relate all these expressions to our example, where the focal length is
100 mm, and the field is ±2 deg. The indices of refraction are
na 1.517 and nb 1.620 , and the Abbe numbers are Va 64.17 and
Vb 36.37 . Notice, the relative aperture (f/#) does not have to be known for these
calculations.
Achromats 47
Radii
Ra1 = 60.7691mm,
Ra2 = –35.4673mm,
Rb1 = –35.8950mm, and
Rb2 = –147.7323mm
We use this prescription, add thicknesses to the elements, and optimize. There is
no difference in comparison with the more direct method applied earlier. If one
prefers this approach using the K-factors, it is highly recommended to create a
program for the lengthy calculations.
The indices are na =3.4253 (Si) and nb = 4.0251 (Ge), with the Abbe numbers
Va = 236.5 and Vb = 107.3.
Vb 107.3
b 0.01 0.008305,
Vb Va 107.3 236.5
and
1 1
fb 120.4101 mm,
b 0.008305
and the thin-lens radii for the silicon element a are R1 = 264.971 mm, R2 = R3 =
–R1 = –264.971 mm, and R4 = –971.817 mm.
Since elements for the MWIR spectrum cannot be cemented (the cement
does not transmit in that region), we have two additional freedoms with the
option of making R3 different from R2 and choosing a suitable spacing between
the elements. (This freedom of enlarging the spacing between the elements will
be discussed later). The lens shown in Fig. 5.8 has been optimized after adding
appropriate thicknesses to the elements.
Achromats 49
Figure 5.8 Blur spots for the MWIR achromat with a ±2-deg field.
Figure 5.9 This LWIR achromat is diffraction limited. The circles indicate the size
of the Airy disk. This indicates that the f/# could be reduced without losing image
quality.
50 Chapter 5
The front element is made from AMTIR-1, with na 2.4975 and Va 113.4 .
For the rear element we use zinc sulfide with nb 2.2003 and Vb 22.7. The
result is a thin lens with the following radii: Ra1 = 91.8029, Ra2 = 393.0097, Rb1 =
−780.7366, and Rb2 =1,243.5363. Adding thicknesses leads to a diffraction-
limited lens whose spot sizes are shown in Fig. 5.9.
0 4
VD 2 .
S L 3 5
This relation indicates that the diffractive phase profile is independent of the lens
material. It also indicates by the negative sign that the chromatic aberration is in
the opposite direction of the aberration from the refractive portion of the lens.
The required refractive focal length for the chromatic aberration correction is
V
f R 1 D f , (5.11)
VR
V
f D 1 R f . (5.12)
VD
r1 2λ 0 f D , (5.13)
D
itotal , (5.14)
2r1
λ0
d max . (5.15)
n0 1
V 2
f R 1 D f 1 100 100.845667 mm .
VR 236.5
52 Chapter 5
From Chapter 2 we apply Eqs. (2.10), (2.11), and (2.12) for a “best-shaped” lens
(minimum spherical aberration) and get for the surfaces the radii
2 3.4253 2 3.4253 1
R1 100.845666 98.690419 mm
3.4253 2 3.4253 1
and
2 3.4253 2 3.4253 1
R2 100.845667 165.451409 mm .
3.4253 2 3.4253 1 4
The transverse spherical aberration of this best-shaped lens, with a 100-mm focal
length and an f/4 relative aperture, is
5.6.1.2 Aspherizing
To eliminate this spherical aberration, we aspherize the second surface with the
method described in Sec. 3.1.2. The conic constant for the aspheric second
surface is
V 236.5
f D 1 R f 1 100.845667 12, 025.84579 mm .
VD 2
5.6.1.4 Comments
The phase equation is expressed by
2π
r
λ
S 2 r 2 S 4 r 4 S6 r 6 ...Sn r n , (5.16)
and the quadratic coefficient for the diffractive surface can be calculated from
1
S2 . (5.17)
2 fD
2π 2
P2 R0 S 2 , (5.18)
λ0
1 1
S2 4.158 105 .
2 f D 2 12025.8
2π
100 4.158 10 653.137 .
5
P2 2
0.004
We add a reasonable thickness of 2.5 mm and optimize this hybrid with a 4-deg
total field. The result is summarized in Fig. 5.6. As a reference, it is mentioned
that the diffraction limit for this lens is BAiry 2.44 f/# 2.44 4 4 39 µm.
The size of the square detector element is shown in Fig. 5.11 with 40 µm.
Figure 5.11 Blur-spot sizes of this optimized MWIR hybrid are equal to the size
of the diffraction-limited blur (Airy disk), which is also the size shown for the
square detector element (pixel).
Figure 5.12 Nomogram for hybrids in the MWIR region (object at infinity).
Achromats 55
Figure 5.13 Nomogram for hybrids in the LWIR region (object at infinity).
The example in Fig. 5.12 shows that the first zone radius of an f/1.5 lens with
f = 50 mm, made from zinc selenide, is 6 mm, and there are eight zones.
56 Chapter 5
In the LWIR region (8–12 µm), germanium is a superior material with its
high Abbe number of about 800. This is indicated by the large separation from
the other materials on the lens material scale in Fig. 5.13.
Example 1 shows an f/1 lens with a 50-mm focal length. The first zone radius
is found to be 6.7 mm, leading to a total of 14 zones across the lens diameter.
Example 2 refers to the 100 mm, f/1 germanium lens discussed above. The first
zone radius is already quite large with 28.3 mm. This germanium lens requires
only three steps (four zones) across its 100 mm diameter.
References
1. W. J. Smith, Modern Optical Engineering, 4th Ed., McGraw-Hill, New York
(2008).
2. M. J. Riedl, Optical Design Fundamentals for Infrared Systems, SPIE Press,
Bellingham, Washington (2001).
Chapter 6
Systems with Two Separated
Components
6.1 Introduction
The telephoto, the reversed telephoto, and the dialyte (which is an air-spaced
achromat) belong in this category, and there are others. We limit our discussion
to the case where the object is located at infinity. The basic layout of such a
system is shown in Fig. 6.1.
V f
f A 1 B f , (6.1)
VA b
V b
f B 1 A b , (6.2)
VB f
Figure 6.1 General layout of system with two separated components for an
object located at infinity.
57
58 Chapter 6
b
d 1 f A . (6.3)
f
It can be seen from these expressions that the achromatic doublet is covered
as a special case with d = 0, which occurs when b = f. Equations (6.1) and (6.2)
then take on the form of the equations presented in Chapter 5, Achromats. There
is a maximum useful separation for the elements, and the separation is dependent
on the materials used. We can derive this limit by first inserting Eq. (6.1) into Eq.
(6.3) and then differentiating the new expression with respect to b. Setting the
result to zero brings us the maximum separation d for the dialyte under
investigation. Let us demonstrate:
b b V f
d 1 f A 1 1 B f . (6.4)
f f VA b
VB
b f .
VA
Typical values for the three spectral windows, VIS, MWIR, and LWIR are
listed in Table 6.1.
Figure 6.2 Ranges of element separations for dialytes with a 100-mm focal
length.
Figure 6.2 shows a plot of the element separation d as a function of the back
focal length b for dialytes with a 100-mm focal length. The materials of the
elements are called out in Table 6.1. It is interesting to observe how the spacing
limits increase with wavelength.
V V
b2 d f 1 B b B f 2 0 . (6.6)
VA VA
107.3 107.3
b 2 10 100 1 b 1002 0 ,
236.5 236.5
b 0.5 135.369979 135.3699792 4 4,536.997886 ,
b1 74.3378 mm, and b2 61.0322 mm .†
†
Also see Fig. 6.2.
60 Chapter 6
As an interesting exercise, we use silicon for both elements and make the rear
element an aspheric hybrid. The basic prescription for dialyte is listed in Table
6.2, and Fig. 6.4 summarizes the performance of this air-spaced achromat with a
hybrid element.
A6 = –2.62939×10–10
S2 = –6.5155×10–5
P2 = –1,023.4466
Table 6.3 Diffractive Zones on Surface 3 (for the first diffraction order)
Zone # 1 2 3
Radius r 7.8353 11.0808 13.5712
62 Chapter 6
The next example of a reverse telephoto for the LWIR region, as shown in
Fig. 6.8, uses only germanium for all elements including the filter. It has a 50-
mm focal length and a relative aperture of f/1.5. The field coverage is ±10 deg.
The front element is a hybrid with the phase profile on the aspherized second
surface. The second surface of the rear element is also aspherized. The aperture
stop is located at the front surface of the first element.
Figure 6.7 Reversed telephoto for the MWIR window, after Aldrich.
64 Chapter 6
Figure 6.8 Reversed telephoto for the LWIR with a hybrid front element.
Chapter 7
From an Air-Spaced Doublet to a Triplet
7.1 Introduction
As an informative exercise, we begin with a suggested formulation1 for a Petzval
objective as shown in Fig. 7.1.
65
66 Chapter 7
c
y 2
TAchC . (7.1)
a Vu 'k
1 c
y 2
TAchC
ya total
a V . (7.2)
c
y 2 ya2 a yb2 b yc2c
a V V V V 0. (7.3)
a b c
After substitution and solving for the power of the first element, we get
a
Vc Vb / 2 Va . (7.4)
2 Va Vb Vc
tot
By choosing the same material for elements a and c, Eq. (7.4) simplifies to read
a
Va Vb / 2 . (7.5)
2 Va Vb
tot
Elements a and c are made from silicon with na = 3.4253 and Va = 236.5. The
material for element b is germanium with nb = 4.0251 and Vb = 107.3. The power
of the front element is therefore
a
236.5 107.3 / 2 0.01 0.011229, and f 1 89.057 mm .
2 236.5 107.3
a
a
The power of the rear element is c tot 0.01 . Therefore, f c f tot 100 mm.
a 0.011229
ca1 0.004630 ,
na 1 3.4253 1
b 0.006229
cb 2 0.002059 ,
1 nb 1 4.0251
and
c 0.01
cc1 0.004123 .
nc 1 3.4253 1
The radii are therefore: Ra1 = 215.987 mm, Ra2 = infinity, Rb1 = infinity, Rb2 =
485.673 mm, Rc1 = 242.542 mm, and Rc2 = infinity.
Figure 7.3 Conventional triplet, derived from the basic Petzval form.
Figure 7.4 Hybrid Petzval objective for MWIR with a cold stop.
Reference
1. W. J. Smith, Modern Optical Engineering, 4th Ed., McGraw-Hill, New York
(2008).
2. M. J. Riedl, “The design of an IR Petzval objective using an aspheric and
diffractive element: an exercise with MOE for beginners,” Proc. SPIE 5865,
586501 (2005).
Chapter 8
A Hybrid for Two Wavelengths
8.1 Introduction
A single element that is corrected for spherical aberration for two chosen
wavelengths can be advantageous for alignment purposes in the visible spectrum
when the application is in the infrared region. We design such an element for the
helium-neon and the carbon dioxide lines. The wavelength for a HeNe laser is
0.6328µm, and is 10.6 µm for a CO2 laser. The design principle is to determine
the shape of the lens for the longer wavelength first and then place a diffractive
structure on one side of the lens to add the required power for the shorter
wavelength to maintain focus position. Since the wavelength ratio is large,
10.6/0.6328 ≈ 17, the step height of the diffractive structure for the short
wavelength is
λ 0.6328
d 0.4 µm .
n 1 2.5907 1
This is in the domain of surface roughness for the infrared region.
First, we have to pick a material that is fitting for both wavelengths. We
choose ZnSe with an index of refraction of n0 6328 2.5907 and n10 6 2.4027 .
We elect for our design example 100 mm as a focal length and a relative aperture
of f/# = 2, and begin with a thin-lens pre-design.
69
70 Chapter 8
2 2
D 50
itotal 667 .
2r1 2 0.968
As hinted before, this is a large number of zones, which is most likely more
cost effective to etch than to diamond-turn. The individual zone radii can be
obtained by ri r1 i 0.9682 i . From this, one can see that at the edge of the
lens the zones are only 19 µm apart. We add a thickness of 6 mm to the lens,
aspherize the front surface slightly in the optimization process, and obtain a lens
that is within the quarter-wave Raleigh limit for both wavelengths, as can be seen
from the plots in Fig. 8.1.
Remark
If we consider the zone step height as the P-V surface roughness for the long
wavelength application, and take a ratio of 2 between the P-V and RMS values,
we have δ 0.5 P V 0.5 0.4 0.2 µm .
Figure 8.1 The quarter-wave Raleigh limit has been maintained for both cases.
A Hybrid for Two Wavelengths 71
relatively low amount of the energy that will become stray radiation and reduce
the contrast ratio (MTF) somewhat.
Reference
1. M. J. Riedl, Optical Design Fundamentals for Infrared Systems, SPIE Press,
Bellingham, Washington (2001).
Chapter 9
Athermats
9.1 Introduction
Athermats are lenses that are designed to compensate for the focus shift that
occurs with temperature excursions. The changing parameters of an optical
element are the radius, the thickness, and the index of refraction. The spacing of
the lens from the detector also changes and is a function of the coefficient of
expansion of the housing material.1
1 1
n 1 , (9.1)
R1 R2
where n = index of refraction, and R1 and R2 are the surface radii. Differentiation
of Eq. (9.1) with respect to temperature yields
d 1 dR 1 dR 1 1 dn
n 1 2 1 2 2 . (9.2)
dt R1 dt R2 dt R1 R2 dt
d 1 1 1 1 dn
n 1 α L α L .
dt R1 R2 R1 R2 dt
d dn
. (9.3)
dt n 1 dt
73
74 Chapter 9
dn / dt
f f α L t . (9.4)
n 1
dn / dt / n 1
L is called the thermal glass constant TR. With that, the
focus shift of a thin lens is simply
f TR f t . (9.5)
T T
f f f a b α h t . (9.6)
f a f b
Element a is the front element and b is the rear element. h is the thermal
coefficient of the housing in which the lens is mounted.
Setting f = 0 to eliminate any focal shift requires that the focal length of the
front element a is
T T
fa b a f . (9.7)
Tb α h
T T fa f
fb a b f , or . (9.8)
Ta α h fa f
r1 2λ 0 f D . (9.10)
Therefore, the focal length is
Athermats 75
r12
fD . (9.11)
2λ 0
r1 ( t ) r1 1 α L t , (9.12)
where L is the thermal coefficient of expansion of the lens material and t is the
temperature change from the design temperature t0. The diffractive focal length at
the changed temperature t can be expressed by
r12 t r12
1 2α L α 2L t ,
2
f D t
2λ 0 2λ 0
r12
f D t 1 2α 2L t . (9.13)
2λ 0
f D 2α L f D t , (9.14)
TD 2α L . (9.15)
Due to the different behavior of a diffractive element, the equations for the
elements’ focal lengths (to eliminate any focal shift caused by a change in
temperature) are
T TD
fR R f . (9.16)
TD α h
fR is the refractive focal length of the element, and f D is the diffractive focal
length of the element.
76 Chapter 9
T T fR f
f D R D f or f D . (9.17)
TR α h fR f
A positive temperature change of the refractive portion of the hybrid reduces
the focal length [see Eq. (9.5)]. As the zone radii r of the diffractive portion
(phase profile) increases with a temperature increase, the diffractive focal length
also increases [see Eq. (9.14)]. It is important to remember that the diffractive
focal length is positive when the steps at the zone transition are negative. The
focal length is negative when the steps are cut in the opposite direction. The
direction of the transition steps is illustrated in Fig. 9.1.
2π
r
λ0
S 2 r 2 S4 r 4 S6 r 6 ... , (9.18)
Figure 9.1 Direction of phase profile and effect of sign on focal length.
Athermats 77
1
S2 . (9.19)
2 fD
λ0
d . (9.20)
n0 1
The change in step height with a temperature change is minimal and results in a
very slight shift of the wavelength at which the theoretical diffraction efficiency
is 1. This change can be neglected and can be confirmed with
λ λ 0 d 1 α L t n dn / dt t 1 . (9.21)
Tables 9.1 and 9.2 list the properties of the most frequently selected materials
for optical elements for the infrared spectra.
Table 9.1 Selected materials for the 3–5 µm spectral band (MWIR).
Table 9.2 Selected materials for the 8–12 m spectral band (LWIR).
The design temperature is 20°C, and the temperature extremes are –20°C and
+60°C. This means t = ±40°C. Both lenses are mounted in aluminum housings.
25.9 126.2
fa 100 205.112 mm .
25.9 23
This leads, according to Eq. (9.8), to
fa f 205.112 100
fb 67.225 mm .
f a f 205.112 100
For simplicity, we use a plano-concave and a plano-convex lens, for which
the radii are Ra1 na 1 f a 4.003 1 205.112 615.953 mm and
Rb1 nb 1 f b 2.200 1 67.225 80.670 mm. Ra2 and Rb2 are infinite.
Athermats 79
T TD 36.3 14.2
fR R f 100 573.86 mm
TD h 14.2 23
and
To determine the details of the phase profile for the flat rear surface of the
lens, we begin by using Eq. (9.19) and get the required phase coefficient,
1 1
S2 0.0058713 .
2 fD 2 85.16
2πRn2 2π 1002
P2 S2 0.0058713 36,890.5 .
λ0 0.01
and
The other zone radii n can be found with ri r1 i . The total number of zones
required is i D 2r1 25 2 1.305067 92 . D is the lens diameter. It is 25
2 2
80 Chapter 9
mm for the f/4, 100-mm focal-length objective. The step size at the zone
transition is d λ 0 n0 1 10 2.406 1 7.11 µm.
–6
Material h(×10 ) fR fD R1 (mm) r1 (mm) i
Alum 23 –573.86 85.16 –806.847 1.30507 92
Steel 12 2,295.45 104.55 3,227.403 1.44603 75
Invar 1.3 391.47 134.09 550.407 1.63954 59
Remark
It can be seen from Eq. (9.16) that the lens will be a plane-parallel plate with a
diffractive structure on the second surface, if the thermal expansion coefficient of
the housing h is equal to the thermal glass coefficient TD, which is 2L, i.e.,
twice the thermal coefficient of expansion on the lens material. AMTIR-1 (L =
12 ×10–6) and aluminium 6061 (h = 23×10–6) are such a matching combination.
T T 36.1 24
fR R D f 100 infinite ,
TD α h 24 24
T T 36.1 24
fD R D f 100 100 f ,
TR α h 36.1 24
and
1 1
S2 0.005 .
2 fD 2 100
Athermats 81
2πRn2 2π 1002
P2 S2 0.005 31, 415.927 .
λ0 0.01
2 2
D 25
i 79 .
2r1 2 2
Figure 9.2 shows the blur spots of this flat, purely diffractive element at
–20°C, + 20°C (the design temperature), and at + 60°C.
60
TR 7.1 106 35.676 106 .
2.40266 1
82 Chapter 9
35.676
fD 100 251.1239 .
14.2
1 1
S2 0.001991 .
2 fD 2 251.1239
For the full lens diameter of D = 100/4 = 25 mm, the total number of
required zones is itotal D 2r1 25 / 2 2.307342 30 .
2 2
Remark
ZEMAX applies a normalized radius of R0 =100 mm for the phase coefficient P.
With that,
2πR02 2π1002
P2 S2 0.001991 11,801 .
λ0 0.0106
We put the numbers from above into the ZEMAX program, add a suitable
thickness to the lens, and optimize. The result is shown in Fig. 9.3.
The top row shows the increase of the blur spot for a purely refractive
singlet. The lower row indicates that the blur spot size stays within the diffraction
limit over the entire 120°C temperature excursion.
Danger!
To demonstrate the effect of machining the phase profile in the wrong direction
(which has happened), we show in Fig. 9.4 how the slopes of the exiting rays are
changed. The focal length of our example lens would increase from 100 mm to
485 mm at the design temperature of 20°C.
Athermats 83
Figure 9.4 Devastating effects with the wrong (reversed) phase-profile direction.
See Fig. 9.1 for direction definition.
84 Chapter 9
total power,
j
i , (9.22)
i1
athermalization,
T α ,
i 1
i i h (9.23)
and a chromatization,
j
i
V 0 . (9.24)
i 1 i
nmiddle 1
VR
nshort nlong
and
λ middle
VD .
λ short λ long
The required powers (inverse focal lengths) for the three elements are
1 / V1 T1 h
1 1 / V3 1 / V1 T3 T1
2 , (9.25)
f 2 1 / V1 1 / V2 T1 T2
1 / V3 1 / V1 T3 T1
1 1 / V1 1 / V2 1 / V1
3 2 , (9.26)
f 3 1 / V3 1 / V1 1 / V3 1 / V1
Athermats 85
and
1
1 2 3 . (9.27)
f1
and
T3 2α 3l 2 5.7 106 11.4 106 .
We recall that the thermal coefficient of expansion for the aluminum housing
is α h 23 106 . Equations (9.25), (9.26), and (9.27) lead to the elements’ focal
lengths. They are: f1 = 44.291 mm, f2 = –79.832 mm, and the diffractive focal
length f3 = –19,273.3 mm. This long focal length indicates that the chromatic
aberration to be corrected with the phase profile is very small. Only two
diffractive zones are required. Figure 9.6 is included for comparison with the
three-element design. Note, there is very little difference. The advantage, of
course, is the savings of the cost of one element, weight, and space.
f f TR ah t . (9.28)
2 f / #
2
t . (9.29)
f TR h
Table 9.4 lists the allowable temperature depatures from the design
temperature for three diffraction-limited singlets mounted in an aluminium
housing. The chosen focal length is again 100 mm, and the relative aperture is
f/4.
It is interesting to note that the focal length of the BK7 lens increases with
the allowable increased temperature of 9.4°C from 100 mm to 100.004 mm,
while the focal length of the silicon lens reduces from 100 mm to 99.908 mm.
For the germanium lens, the focal length shortens to 99.729 mm.
§
The Rayleigh criterion states that a wavefront error of a quarter wave does not
noticeably affect the image quality.
88 Chapter 9
Figure 9.8. Performance plots for the lens at 40.7°C. The quarter wave optical
path difference (OPD) is indicated.
Reference
1. M. J. Riedl, Optical Design Fundamentals for Infrared Systems, SPIE Press,
Bellingham, Washington (2001).
Chapter 10
The Ball Lens
10.1 Introduction
A ball lens is an interesting optical element worth a closer look. Figure 10.1
indicates the optical behaviour of such a lens. It can be seen that the principal
planes fall together and are located at the center of the sphere. For the
aberrations, we only consider here the case where the object is located at infinity.
Some remarks will be made for other situations. The focal length is
Rn
f , (10.1)
2 n 1
bfl f R
2 n R , (10.2)
2 n 1
Figure 10.1 Ball lens. The white lines suggest that the lens could be made into a
cylinder for easier mounting.
89
90 Chapter 10
where R is the radius of the sphere, and n is the refractive index of the lens
material.
n1 n'1 n1
TSC1 y1 u'1 i1 i12 , (10.3)
2n'1n'2u'2
recognizing that n1 n'2 1, and calling n'1 n , we find by paraxial ray tracing,
u'1 n 1 y1 nR , i1 y1 R , and u'2 2 n 1 y1 nR . Inserting all these
relations into Eq. (10.3) results in the simple expression
y3
TSC1 , (10.4)
4nR 2
where we substituted y for y1 . For the second surface we go about it in the same
way. The surface contribution is
n2 n'2 n2
TSC2 2 y2 u'2 i2 i22 . (10.5)
2n'2 n'2u'2
y2
2 n y ,
n
y
i2 ,
Rn
and as before,
2 n 1 y
u'2 .
nR
The Ball Lens 91
TSC2
n 2 2n 1 y 3 . (10.6)
4n 2 R 2
n n 3 1 y 3
TSCball . (10.7)
lens 2n 2 R 2
We recall that the blur spot size due to spherical aberration is Bspher TSC 2 , or
angularly expressed
n n 3 1
βspher . (10.8)
128 n 1 f / #
2 3
ball
10.3 Coma
To get the expression for coma, all we have to do is to multiply the spherical
aberration surface contributions with the factor i p1 i1 i p 2 i2 u p R y . The
symbol u p is the half-field angle. This leads to
n n 3 1 y 2u p
CCball . (10.9)
lens 2n 2 R
n n 3 1 u p f
β coma . (10.10)
16n n 1 f / #
2
ball
92 Chapter 10
10.4 Astigmatism
By multiplying the coma expression once more with the factor
i p1 i1 i p 2 i2 u p R y , we get the third-order astigmatic contribution. The result
is
n n 3 1 yu 2p
TAC . (10.11)
2n 2
The minimum astigmatic blur size is Bastig 2TAC . The angular measure is then
ball
n n 3 1 u 2p
β astig . (10.12)
ball 2n 2 f / #
Example
After that somewhat laborious preparation, we demonstrate the use of these
equations by designing a ball lens, made from BK7, for use with a HeNe laser (
= 0.6328µm). The index of refraction of BK7 is, for that wavelength, n 1.5151 .
We employ a relative aperture of f/2.5, and a rather large half-field angle of
u p 0.05 rad 2.8645 deg. The ball radius R = 2.5 mm.
Coma CC = −0.0029 mm
Blur-spot size, linear Bcoma 0.0029 mm
Blur-spot size, angular β coma 0.80 mrad
The focal length f = 3.6767 mm, and the back focal length bfl = 1.1767 mm.
The Ball Lens 93
Figure 10.2 Image locations inside a ball lens, a function of the refractive index.
From Eq. (10.2) it can be seen that if n > 2, the focus falls inside the lens.
Therefore, a ball lens is not suitable for normal use if it is made from zinc
selenide, silicon, or germanium when the object is located at infinity. This is
indicated in Fig. 10.2. The image locations can be derived by using Eqs. (1.7)
and (1.9) from Chapter 1. With n being the index of the ball material, and
choosing the ray height y = 1, Eq (1.7) changes to nu' n 1 R , and from Eq.
(1.9) we get u' 1 / l' . This leads to l' n / n 1 R .
Table 10.1 Image distances inside the ball lens (n>2) for an object at infinity.
For the image to fall on the backside of the ball lens, the object has to be
placed at a distance l in front of the lens. This distance is a function of the lens
material and is expressed by
2R
l . (10.13)
n 2
Table 10.2 lists this distance for selected materials as a function of the ball
radius. Figure 10.3 shows the relative positions for the different materials.
94 Chapter 10
Table 10.2 Object distances for image to fall on the rear side of a ball lens.
Figure 11.1 An f/1.4 triplet, with a 100-mm focal length and a field of ±3 deg.
95
96 Chapter 11
Surface 1 2 3 4 5
TSC +0.008224 +0.149244 –0.629787 +0.088954 –0.004288
Surface 6 7 8 9 10
TSC +0.404949 –0.582071 +0.571206 –0.307748 +0.305024
10
TSCtotal TSC 0.003707 mm .
1
Figure 11.2 This diagram clearly shows the aberration contributions of the
individual lens surfaces.
Seidel and the Pegel Diagrams 97
h' 2
ρ , (11.1)
2 PC
where h′ is the image height, and PC is the sum of the third-order longitudinal
Petzval contribution. For our example lens with a 100-mm focal length and a
3-deg half-field angle, h' f tan u p 100 tan 3deg 5.241 mm . PC is
determined by the same method as the other third-order aberration contributions.
We find PC to be –0.04357. This leads to a Petzval radius of
ρ h' 2 2 PC 5.2412 2 0.04353 315.5 mm . The negative sign
indicates that the curvature is bent inwards, i.e., towards the lens. This inward
bending of the field curvature was especially pronounced in the earlier cameras
where the objectives were made up of but a few elements, or even just one. To
compensate for this defocusing effect at the edge of the image, photographers at
the time used the trick in group pictures to align the assembly in an arc. This is
demonstrated in Fig. 11.3.
Figure 11.3 Petzval curvature “applied” for the object. (Shown in the picture are
some of the author’s professors from the Academy of Applied Sciences in
Munich.)
98 Chapter 11
The individual surface contributions are listed in Table 11.3, and the Pegel
diagram is shown in Fig. 11.5.
Seidel and the Pegel Diagrams 99
Surface 1 2 3 4 5
TSC –0.040580 +2.838347 –4.117814 +0.858031 +0.417371
Surface 6 7 8 9 10
TSC –0.526075 –0.046702 +0.046419 +0.259682 –0.671524
Surface 11 12 13 14 15
TSC +0.784353 +0.001294 +0.127990 +0.026294 –0.103859
Surface 16 17 18 TSC
TSC +0.166109 –0.177425 +0.174250 +0.016161
Figure 11.5 Pegel diagram for the Cassegrain version depicted in Fig. 11.4.
Chapter 12
The Single-Imaging Mirror
12.1 Introduction
Included in this discussion are mirrors with spherical and aspheric surfaces,
which are single-surface reflectors, and the Mangin mirror, which is a
catadioptric element. The advantage of a pure reflector is that it is free from
chromatic aberration and therefore suitable for applications over the entire optical
spectrum. This property allows alignments of IR systems in the visible spectrum
because there is no focus shift.
spherical aberration,
y3 f
TSC , (12.1)
64 f / #
2 2
8f
sagittal coma,
y 2u p up f
CC , (12.2)
16 f / #
2
4f
astigmatism,
yu 2p u 2p f
TAC , (12.3)
2 4 f / #
Petzval curvature,
yu 2p u 2p f
TPC , (12.4)
4 4 f / #
101
102 Chapter 12
spherical aberration,
TSC f
Bspher , (12.5)
128 f / #
3
2
coma,
up f
Bcoma CC , (12.6)
16 f / #
2
astigmatism,
u 2p f
Bastig 2TAC , (12.7)
2 f / #
h' 2 h' 2 2u 2p f 2
ρ f . (12.8)
2 PC 2u 'TPC 2u 2p f
Figure 12.1 Sagittal focus moves along the tangent of the circle as a function of
the half-field angle up.
The Single-Imaging Mirror 103
1 f 1
AC f s ft cos u p .
(12.9)
2 2 cos u p
1
u' ,
2 f / #
and
AC
TAC ,
2 f / #
therefore,
f 1
TAC cos u p .
4 f / # cos u p
u 2p
cos u p 1 .
2
f 2 2 u 2p fu 2p
TAC .
4 f / # 2 u 2p 2 4 f / #
104 Chapter 12
12.4 Examples
We analyze the image from a spherical mirror, modify the shape of the mirror to
a toroid, and compare the results.
To keep the spherical aberration small, we use a slow system of f/10 with a
100-mm focal length. The half-field angle up = 0.05 rad, which is 2.864789 deg.
Figures 12.1 and 12.2 show that the tangential radius Rt has been shortened, and
The Single-Imaging Mirror 105
the sagittal radius Rs has to be increased to make the two focal lengths fs and ft
equal to each other and to f. The relations are
2f
Rt 2 f s (12.10)
cos u p
and
Rs 2 ft 2 f cos u p . (12.11)
106 Chapter 12
Figure 12.4 shows the shape of a square toroidal mirror and indicates again the
generation of it.
up f 0.05 100
Bcoma 0.003125 mm
16 f / # 16 102
2
u 2p f 0.052 100
Bastig 0.0125 mm
2 f / # 2 10
up f 0.05 100
Bcoma 0.003125 mm .
16 f / # 16 102
2
This is the size of the sagittal coma. The tangential coma blur is three times as
large, i.e., 0.009375 mm. This is confirmed in Fig. 12.5, which is the result of a
real ray trace.
1
2 n 1 c1 2nc2 , (12.12)
f
2 n 1 c1
c2 . (12.13)
2n
1 2nR1 f
R2 . (12.14)
c2 2 n 1 f R1
Finding the curvature c1 for which spherical aberration vanishes leads to the
cubic equation
c1 n 3 c1 4n 5 c1 4n 3
3 2 2
0. (12.15)
2 4 8 n 1
The derivation of the aberration expressions and the analytical solution of Eq.
(12.5) are quite cumbersome2 but enlightening. Matters become much easier if
we set the curvature of the first surface equal, but opposite in sign, to the mirror’s
power, i.e., c1 , which means making the first surface radius R1 f . With
this postulation for the thin Mangin mirror, we get
2nR1 2nf
R2 . (12.16)
2n 1 2n 1
The angular blur spots due to the different aberrations for this shape are
spherical aberration,
βspher
2n 3 , (12.17)
128n 2 f / #
3
The Single-Imaging Mirror 111
coma,
up
β coma , (12.18)
16n 2 f / #
2
astigmatism,
u 2p
β astig , (12.19)
2 f / #
chromatic,
β chrom
n 1 , (12.20)
2nV f / #
n2 f
ρ . (12.21)
2n 2
1
Remark
An interesting case exists when n = 1.5. From Eq. (12.17) we see that with n =
1.5, spherical aberration vanishes. This was recognized by Mangin.
Table 12.1 lists the radii required for different materials to eliminate
spherical aberration of the Mangin mirror when the object is located at infinity.
The results are obtained with Eqs. (12.14) and (12.15).
Table 12.1 Surface radii for Mangin mirrors, corrected for third-order spherical
aberration.
Material n R1 R2
Glass 1.5 −1.0 f −1.5 f
Zinc Selenide 2.4 −1.1506 f −1.3980 f
Silicon 3.4 −1.1790 f −1.3409 f
Germanium 4.0 −1.1823 f −1.3169 f
112 Chapter 12
References
1. W. J. Smith, Modern Optical Engineering, 4th Ed., McGraw-Hill, New York
(2008).
2. M. J. Riedl, “The Mangin mirror and its primary aberrations,” Appl. Opt.,
13(7), 1690–1694 (1974).
Chapter 13
Eight Single Optical Elements
as Imaging Objectives
13.1 Introduction
A single optical element is often adequate to meet the image quality expected,
especially from an infrared system, if the field to be covered is moderate. To
provide a reference for a quick evaluation and comparison, eight different singlet
configurations are presented in the form of a summary matrix of the previous
subjects covered. This matrix will indicate whether a single element is sufficient
for the task at hand or if a multi-element objective is required.
To make matters manageable and simple, it has been assumed that the object
is located at infinity and the aperture stop is placed at the objective. To avoid
unpleasant surprises, it is recommended to always check first the size of the
diffraction blur formed by the element. This is especially important for
applications in the infrared region where the wavelengths are five to ten times
longer than in the visible spectrum.
where is the wavelength and f/# the relative aperture or “speed” of the
objective. This diffraction blur is also known as Airy disk, named after Lord
Airy, the British mathematician and astronomer.
113
114 Chapter 13
Before making any statement about aberrations, Fig. 13.2 identifies the
symbols and relations used in the expressions.
2 n 2 n 1 2 n 2 n 1
2 f f 0 0
n 2n 1 n 2n 1 4
n 1 f
1
3 f n2
0
n
4
n 1
2
f
n 2
1
f 0
n3 n 1
n n 1 1
3
n2 n n 1 1
5
n 2
1
f
n 2
1
f 0
n3 n 1
n n 1 1
3
n2 n n 1 1
2n
6 f f 0 0
1 2n
7 2 f - 0 -
8 2 f - –1 -
radius of the first zone r1 2λf d , with being the wavelength, and fd being the
focal length of the diffractive portion of the lens. The other zone radii of the
profile can be obtained with ri r1 i . The step size at the transition from one
zone to the next is d max λ / n 1 .
13.5 Aberrations1
Remarks
1. The aberrations are stated as angular-blur spot sizes , in radians. The linear
blur-spot sizes are B f .
2. The blur size, due to astigmatism, remains the same for all eight
configurations, namely β astig u 2p 2 f / # .
Table 13.3 Angular aberration blur spot sizes of the elements in radians.
n 4n 1 1 up
2
128 n 1 n 2 f / #
2 3 2V f / # 16 n 2 f / #
2
1 up
3 0
2V f / # 8 n 1 f / #
2
1
4 0 0
2V f / #
5 0 0 0
2n 3 n 1 up
6
128n 2 f / # 2nV f / # 16n 2 f / #
3 2
1 up
7 0
128 f / # 16 f / #
3 2
up
8 0 0
16 f / #
2
Eight Single Optical Elements as Imaging Objectives 117
13.6 Examples
We determine the shapes and aberration blur-spot sizes for an element with f =
100 mm, f/# = 1.5, M = 4 m, S = 3 m, and L = 5 m. The half field angle
u p = 0.05 rad.
With silicon as the chosen lens material, the index of refraction n = 3.425,
and the Abbe number V = 236. For a point of reference, we determine first the
diffraction blur size.
# R1 R2 sph ch c a
There are 12 zones for the diffractive hybrid, Element 5. Their radii are listed
below.
The maximum step dimension at the transition radii is d max = 0.00165 mm.
Reference
1. M. J. Riedl, Optical Design Fundamentals for Infrared Systems, SPIE Press,
Bellingham, Washington (2001).
Chapter 14
A Progression of Performance with an
Increase in Lens Complexity
14.1 Objectives
We apply the knowledge gained from the preceding sections by analyzing six
different objectives for the MWIR spectrum with the same focal length (50 mm),
the same f/# (1.5), and the same square field coverage (4 deg × 4 deg). We start
with the simplest form of a lens and reach (in five steps) an objective that is
frequently found in thermal imaging systems.
1. plano-convex singlet
2. best shaped singlet
3. aspheric singlet
4. aspheric hybrid
5. Petzval objective with one aspheric element
6. Petzval objective with one aspheric hybrid
The plot in Fig. 14.1 is based on using square sensor elements (pixels) that
contain 80% of the received energy from a point source at infinity. It is clear that
the simple singlets are not suitable for IR cameras, but are sufficient for
radiometers (pyrometers) if the target is large enough to resolve the projected
sensor element. The single aspheric hybrid with a 70-µm pixel size over a linear
field of view of 3.5 mm in the horizontal and vertical directions may begin to get
interesting for thermal imaging. The total number of pixels amounts
to (3.5 / 0.07)2 2500 , and the angular resolution is 0.07/50 =1.4 mrad. At the
end of the list is the frequently employed Petzval objective, which was discussed
in detail in Sec. 7.4. With a 20-µm pixel size, the focal plane array contains more
than 30,000 sensing elements. The angular resolution is 0.4 mrad.
Remarks
1. For the off-axis calculations, the field was increased by the factor 2 to
cover the pixels located at the extreme of the field’s diagonal.
2. All lenses are made from silicon with the exception of the aspheric hybrid,
which is made from GASIR-1, a material that lends itself to molding. Its
refractive index is 2.5116 at 4 µm, and its Abbe number in the MWIR region
is 197.3.
119
120 Chapter 14
3. For reference, notice that the Airy disk, containing 84% of the energy from
an object point, is 15 µm for the stated wavelength of 4 µm and the assumed
relative aperture of f/# 1.5 .
4. For an increased field of view and a lower f/#, elements have to be added to
these basic designs with a choice of making use of the properties of different
materials.
Nine years later, in 1672, the Frenchman Guillaume Cassegrain came up with
the idea to intercept the reflected rays from the primary mirror before they come
to a focus. For that purpose, a hyperbolic secondary mirror is needed.
Cassegrain’s system is much shorter and has found its place in many
applications—from astronomical telescopes to thermal imaging devices. It is
interesting to note that Isaac Newton, the great British scientist, never gave
Cassegrain any credit for his invention. He referred to it as “just an obvious
extension” of Gregory’s concept. If the field of view is narrow, only spherical
aberration has to be corrected. Dutchmen Horace Dall and Allan Kirkham
conceived of the idea to replace the hyperbolic secondary mirror with an easier-
to-manufacture convex spherical mirror. To keep the system free from spherical
aberration, the primary mirror becomes, in that case, an ellipsoid. This increases
coma somewhat, but that was not of much concern to the two astronomers. In the
1920s, the American George Ritchey and the Frenchman Henri Chretien
modified the basic Cassegrain arrangement so that it is corrected for spherical
aberration and third-order coma. This Ritchey-Chretien configuration is now the
most widely used combination. The famous Hubble Space Telescope is of this
121
122 Chapter 15
type. We shall discuss the details of all these configurations, including the use of
a Mangin mirror as a secondary reflector. Additionally, we will present the
classic Schwarzschild objective and a special configuration, the inverted
Schwarzschild, which is corrected for spherical aberration, coma, astigmatism,
and field curvature.
15.2.1 Equations1
From the geometry of the layout and a paraxial ray trace one finds the radii of the
mirrors. They are
2df
R1 (primary mirror) (15.1)
b f
and
2db
R2 (secondary mirror). (15.2)
b d f
spherical aberration,
f b f b f d b f d b
3 2
βspher , (15.3)
128 fd 3 f /#
3
coma,
2 f b f 2 f d b f d b d f b u p
β coma , (15.4)
32d 2 f f /#
2
and astigmatism,
4bf b f f d b d f b 2 u p 2
β astig . (15.5)
8 fbd f /#
bfd
ρ Petzval . (15.6)
fd b f
2
124 Chapter 15
2
f d b
κ2 . (15.7)
bd f
up
β coma , (15.8)
16 f /#
2
and astigmatism is
d f u 2p
β astig . (15.9)
2b f /#
bfd
ρ Petzval .
fd b f
2
f f b b f d b f d b
3 2
κ1 . (15.10)
f b f
3
Coma is
u p 2 f b f f d b f d b d f b
2
β coma , (15.11)
32d f f /#
2 2
and astigmatism is
Two-Mirror Systems as Telescope and Microscope Objectives 125
u 2p 4bf b f f d b d f b
2
β astig . (15.12)
8dbf f /#
2bd 2
κ1 1 , (15.13)
b f
3
2 f b f 2 f d b f d b d f b
κ2 . (15.14)
b d f
3
u 2p 2 f d
β astig . (15.15)
4b f /#
15.7 Examples
For performance comparison, we analyze several different telescope systems
with the identical focal length of f = 500 mm, a separation of the elements of d =
100 mm, and back focal distance of b = 150 mm. The relative aperture is f/5, and
the half-field angle up = 0.01 radians. The mirror radii are, according to Eqs.
(15.1) and (15.2),
2df 2 100 500
R1 285.714286 mm
b f 150 500
and
2db 2 100 150
R2 120.0000 mm .
b d f 150 100 500
The Petzval radius is the same for all configurations, namely
bfd 150 500 100
Petzval 103.448 mm.
fd (b f ) 2
500 100 (150 500) 2
126 Chapter 15
Table 15.1 Conic constants and angular blur-spot sizes for chosen two-mirror
systems.
The angular blur spot sizes spher , coma , and astig are in milliradians.
Figure 15.3 Blur spots from a Cassegrain system with a Mangin mirror as a
secondary reflector.
Two-Mirror Systems as Telescope and Microscope Objectives 127
One can see what improvement a Mangin mirror brings if the blur-spot sizes
are compared with the ones from a two-sphere mirror system, where the spherical
aberration blur alone is 0.865 mm when refocused to the best position. This can
be confirmed with Eqs. (15.3), (15.4), and (15.5).
We now can calculate the radii of the mirrors with Eqs. (15.1) and (15.2), and
the conic constants, for which we apply Eqs. (15.13) and (15.14) to correct
spherical aberration and coma. Recall that changing the sign of the focal length
for the equations leads to
2 270 2202
κ1 1 0.942751 ,
270 500
3
and
κ2
2 500 270 500 990 550 450 0.358524.
2
Astigmatism is
While this astigmatic blur is very small, the Petzval curvature is steep with a
radius of
For the classical Cassegrain and R/C systems, with the same focal length and f/#,
the Petzval radius is –103.4483 mm.
After refocusing by a small amount, the geometrical blur spots are still
smaller than the diffraction limit because of the relatively slow speed of f/5 and
the wavelength of 4 µm that we had assumed. The Airy disk is
BAiry 2.44λ f /# 2.44 0.004 5 0.0488 mm .
Figure 15.6 Folded Gregorian system, designed with the provided equations. (f =
500 mm, d = 220 mm, b = 270, folded to d/2 = 110 mm.)
130 Chapter 15
Figure 15.8 Blur-spot sizes for the inverted 20-mm-focal-length, f/1.5 Gregorian
objective. The reference circle is the size of the Airy disk.
Two-Mirror Systems as Telescope and Microscope Objectives 131
Figure 15.7 shows the layout, and Fig. 15.8 depicts the linear blur-spot sizes,
compared with the diffraction limit in the visible spectrum.
We now present a folded variation of this objective. For this case, the
effective focal length is 10 mm, and the relative aperture f/# =1.5. The folding
mirror has been slightly aspherized to aid in the correction of the off-axis
aberrations. A possible layout, as a so-called snap-together arrangement, suitable
for diamond turning, is shown in Fig. 15.9. Notice the large free-working
distance.
The first striking difference is the size. Another one that is not quite as
obvious is the fact that in the second form, the “inverted” configuration, there is
no need for a spider to hold the secondary mirror in place as is the case with the
classic one. The first kind has a large free-working distance which is frequently
required. On the other hand, the short free-working distance of the second kind is
no shortcoming for applications in, for example, photolithography.
b 52 f
R1
5 1 f
R2 5 1 f
f
y1
2 f /#
52 f
y2
5 2 y1
2 f /#
Note that both mirror surfaces are spherical and concentric to each other. It is
an interesting exercise to demonstrate that the third-order spherical aberration,
coma, and astigmatism vanish for this configuration. To eliminate the Petzval
curvature and distortion, one can position a field flattener in the focal plane with
a front radius of
n2 1
R3 f
1 n n 1
The diagram shows the relations of the “golden ratio,” which says,
2f x
.
x 2 f x
Therefore,
R2 R1 d R1 2 f
5 1 f 2 f
5 1 f . **
Example
For a system with f = 10 mm and a relative aperture of f/1, we get d = 20 mm, b =
42.3607 mm, R1 = 12.3607 mm, R2 = 32.3607 mm, y1 = 5 mm, and y2 = 21.1803
mm.
**
See the fourth equation on page 133, under 15.11.1.
Two-Mirror Systems as Telescope and Microscope Objectives 135
d 2 f (Same as above)
b 1 2 f
R1 R2 2 2 f
2
κ1 1 2 = 5.828427 (Conic constant for first surface)
2
κ2 1 2 0.171573 (The reciprocal of κ1 )
f
y1
2 f /#
y2
1 2 f
2 f /#
y2 1 2 y1
2
Fractional area obscuration ε 2 1 1 / 5.8 17.3%
Note that both surfaces’ radii are equal. By aspherizing, both become oblate
ellipsoids.
Here is another example for the same focal length and f/#. The system is
shown in Fig 15.12. d = 20 mm as before. b = 24.1421 mm, and R1 = R2 =
28.2843 mm. The conic constants are independent of focal length and f/#. They
are 1 = 5.828427 and 2 = 1/1 = 0.171573.
References
1. W. J. Smith, Modern Optical Engineering, 4th Ed., McGraw-Hill, New York
(2008).
2. Bibliographisches Institut & F. A. Brockhaus AG, Germany (2007).
3. K. Mütze, ABC der Optik, Verlag Werner Dausien, Hanau, Germany (1961).
Chapter 16
The Plane-Parallel Plate
16.1 Introduction
There are two aspects that have to be dealt with when using a plane-parallel plate
inserted in a converging or diverging beam. First, there are the aberrations, and
second there is a relocation of the image. This relocation is a simple longitudinal
shift if the plate is perpendicular to the optical axis. If the plate is tilted, as is the
case with most beam splitters, there is a longitudinal and a lateral displacement,
which is not too easy to assess because they interact. We shall address both cases.
It should also be clear that most prisms behave like thick plane-parallel plates.
This is indicated in Fig. 16.2 with a number of examples. The “unfolding” of the
prism into an equivalent plane parallel plate is called the tunnel diagram. The
first two examples are equivalent to plane-parallel plates inserted perpendicularly
to the optical axis. The third one, the Dove prism, corresponds to a tilted plate,
which of course introduces astigmatism and coma as will be pointed out further
down in the text.
For several reasons, such as cost and absorption, prisms are not much used in
the infrared spectrum. However, wedges, which are thin prisms with a small apex
angle, are employed in this region as beam-steering elements. The deviation of a
ray exiting such a wedge is n 1 , where n is the index of refraction, and α
is the wedge angle. Details are shown in Fig. 16.1.
Figure 16.1 Thin prism (wedge) and the deviation of an incoming ray.
137
138 Chapter 16
16.2 Aberrations1
It is frequently overlooked that a plane-parallel plate also contributes aberrations
when it is inserted in a converging or diverging beam. Using the third-order
surface-contribution equations again, we demonstrate step by step the derivation
for the spherical aberration. The symbols used in the derivation are identified in
Fig. 16.3
where
n n' n
B y u' i ,
2n'n'k Inv
The total transverse spherical aberration is the sum of the two surface
2
contributions TSC TSC
1
1 TSC2 . We make things easier if we set y1 0 ,
2
because then TSC1 0, and TSC TSC
1
2 . Equation (16.1) can now be restated
as
n 1 y
B 2 u'2 i2 i22 h' .
2nInv
2
i22 1 / 2n f /# 1 / 4n 2 f /#
2
With and recalling that the
minimum blur size, when refocused, is half the transverse spherical aberration,
we finally get the blur size of a plane-parallel plate with thickness t and index n.
It is simply
Bspher
n 2
1 t
. (16.2)
32n3 f /#
3
The Plane-Parallel Plate 139
Figure 16.2 Tunnel diagrams of three prisms show the relation to a plane-parallel
plate.
The blur-spot sizes due to coma, astigmatism, axial chromatic, and lateral color
can be derived by the same procedure. They are
coma,
Bcoma
n 2
1 tu p
, (16.3)
8n3 f /#
2
astigmatism,
Bastig
n 2
1 tu 2p
, (16.4)
2n3 f /#
axial chromatic,
Bchrom
n 1 t , (16.5)
2n 2V f /#
n 1 tu p
Blat color , (16.6)
n 2V
These effects need to be taken into account, especially if filters with different
thicknesses and indices are inserted in the converging beam. The Dewar window
of a cooled detector package is another example of that.
16.2.1 Examples
In Table 16.1, we compare the performance of three 2-mm-thick beam splitters
made from different materials, inserted in a converging f/2 cone. For the visible
spectrum we use BK7, for the MWIR range from 3 µm to 5 µm, we contrast
silicon against sapphire.
The Plane-Parallel Plate 141
Figure 16.4 Blur spot sizes of a 2-mm beam splitter, tilted about 45 deg. The
circles indicate the calculated blur-spot sizes.
Table 16.1 Blur spot sizes as a function of material and spectral range.
The calculated sizes, using the developed equations, are compared with the
results from a full computer analysis in Fig. 16.4.
The scaled blur-spot diagrams from the computer confirm the validity of the
simple third-order expressions. The plot for the sapphire substrate shows clearly
the impact of the very high dispersion for that material in the MWIR. Especially
pronounced is the lateral color, which is already indicated in the table.
1 sin 2 u
L 1 2 t . (16.7)
n sin 2 u
l
n 1 t . (16.8)
n
142 Chapter 16
From the relation tan u D 2 f 1 2 f /# , we get the slope angle u for the
insertion in Eq. (16.7). Table 16.2 lists the errors in percent for a glass plate with
a refractive index n = 1.5. Table 16.3 indicates that the difference is much smaller
for silicon, for which the index n = 3.4 in the MWIR region.
Table 16.2 Error resulting from substituting Eq. (16.8) for small angles u. n =1.5
(glass)
Table 16.3 Error resulting from substituting Eq. (16.8) for small angles u. n = 3.4
(silicon).
1 sin 2 ut
T 1 2 t sin ut . (16.9)
n sin 2 ut
ut is the tilt angle, as indicated in Fig 16.6.
Figure 16.9 Blur-spot comparison with one and two plates arranged as shown in
Fig. 16.8. Also indicated is the best longitudinal image position.
plates are 2 mm thick and are inserted in an f/2.5 cone. The remaining aberrations
are negligible, except coma, which is approximately 20 µm [according to Eq.
(16.3)] for the two plates. The results are compared against the performance of a
beam splitter without a compensator. Figure 16.9 compares the blur-spot shapes
and sizes of a single plate with a two-plate arrangement.
Reference
1. M. J. Riedl, Optical Design Fundamentals for Infrared Systems, SPIE Press,
Bellingham, Washington (2001).
Chapter 17
MTF, Limits, and Pixel Sizes
17.1 Introduction
The optical modulation transfer function (MTF) is a well-known concept for
evaluating the performance of a lens.1 It contains valuable information about the
resolving power of the lens and the contrast in the image. Modern infrared
imaging systems use sensor arrays that are made up of small individual sensing
elements, the pixels. The practical minimum size of such a pixel depends on the
wavelength of the collected radiation and sets the spatial resolution limit of the
optical system, referred to as the Nyquist frequency.
2
2
MTFdiffr ( ) arccos 1 , (17.1)
π 0 0 0
1
0 (17.2)
λ f /#
This indicates that for the visible spectrum with = 0.5 µm, the cut-off
frequency in line pair/mm v0VIS 2000 / f /# . In the MWIR region, with = 4
µm, 0MWIR 250 f /# . For the LWIR spectrum, with = 10 µm,
0LWIR 100 f / # . These numbers are reminders of the sizeable differences
among the different spectral regions.
147
148 Chapter 17
With a common postulation that the pixel size should be not smaller than the
Airy disk, we can state that p 2.44λ f /# . The Nyquist frequency is then
1
Ny . (17.3)
5λ f /#
0Ny 1 / 5λ f /# 1
.
0 1 / λ f /# 5
In other words, the normalized Nyquist cut-off frequency is five times lower than
the optical one. This is marked in Fig. 17.1.
Referring to the different spectral regions, Eq. (17.3) states that the limiting
Nyquist frequency, expressed in line pairs per millimeter, is 50/(f/#) for the
MWIR region and 20/(f/#) for the LWIR window.
MTF, Limits, and Pixel Sizes 149
Figure 17.2 Focal plane array. Columns 3 and 4 form a line pair 2p wide. The
spatial frequency is the Nyquist frequency Ny 1 / (2 p) .
References
1. W. J. Smith, Modern Optical Engineering, 4th Ed., McGraw-Hill, New York
(2008).
2. M. J. Riedl, Optical Design Fundamentals for Infrared Systems, SPIE Press,
Bellingham, Washington (2001).
Chapter 18
Details of a Hybrid Lens
18.1 Introduction
Hybrid lenses are being frequently used in infrared systems, primarily for color
correction. We shall analyze one such lens, which is the third element of the
MWIR objective shown in Fig. 18.1.
18.2 Hybrid1
The prescription of the diffractive surface (front surface) is as follows:
1. Radius R1 = 21.246406 mm
2. Conic constant = –0.151513 (prolate ellipse)
3. Phase coefficient P2 = −5975.55048.
Figure 18.1 25-mm focal-length wide-angle objective with a cold stop for the
MWIR region. The relative aperture is f/1.4, and the field coverage is ±12 deg.
The third lens is a germanium hybrid, which will be analyzed.
151
152 Chapter 18
2π
r
λ0
S 2 r 2 S4 r 4 S6 r 6 ... . (18.1)
Note that, for this design, only the quadratic term of the phase function has been
used, and the asphere is a conic section without additional higher deformation
coefficients. This has been done to simplify the analysis without impacting the
correctness of the general approach.
For this equation, r is the zone radius of the diffractive profile and λ 0 is the
design wavelength, which in our example is 4 µm.
λ0
r1 . (18.2)
S2
ri r1 i . (18.3)
2
D
itotal . (18.4)
2r1
λ0
d max . (18.5)
n 1
We now apply all these relations to our example and find
0.004
r1 3.42657 mm .
3.804 104
The other radii are r2 r1 2 4.5858 mm , r3 = 5.6164 mm, r4 =6.4853 mm, and
r5 =7.2508 mm. Equation (18.4) states that
Details of a Hybrid Lens 153
2 2
D 13.6
itotal 4.397 ,
2r1 2 3.24657
which means the fifth zone is not a full zone. Only a fraction of it will be used.
λ0 0.004
d max 0.001322 mm .
n 1 4.02506 1
1 2 S2 2
zdiff S 2 r λ 0 Int r . (18.7)
n0 1 λ0
zasph
1/ 21.24606 52 0.000669 mm
1 1 0.151513 1 1/ 21.24606 52
2
zdiff
1
3.804 10 5 0.004 Int
4 2 3.804 104 2
5
4.02506 1 0.004
1
zdiff 0.009510 0.004 2 0.000499 mm
3.02506
ztotal zasph zdiff = 0.000669 - 0.000449 = 0.00022 mm
If we insert the third radius with 5.616400 mm in the integer (Int) bracket,
we get zdiff 1 4.02506 1 0.016 0.012 0.001322 mm . This is the
maximum step height as stated with Eq. (18.5). Figure 18.3 shows performance
plots of this objective.
Details of a Hybrid Lens 155
Reference
1. M. J. Riedl, Optical Design Fundamentals for Infrared Systems, SPIE Press,
Bellingham, Washington (2001).
Chapter 19
From the Höegh Meniscus to
Double Anastigmats
19.1 Introduction
This chapter outlines the design steps for several photo objectives in an
evolutionary way. Beginning with a single element, the Höegh meniscus, we
progress by adding elements. Two elements lead to the Hypergon, a
symmetrically arranged doublet. In the next step, the two lenses are shaped into
cemented achromats. Finally, these cemented achromats are split into airspaced
achromats, which leads to an arrangement referred to as a double anastigmat. In
the final optimization, we break the symmetry and change the radii of the second
achromat. The configuration is an excellent general-purpose lens with a relative
wide angle coverage.
157
158 Chapter 19
n1 n1' c1h'Inv
TPC1
2n1n1'
and
n2 n'2 c2 h'Inv
TPC2 .
2n2 n'2
1 n ch'Inv
TPC1
2n
and
n 1 ch'Inv
TPC2 TPC1 .
2n
1
n 1 c1 c2
n 1 c c t .
1 2
f n
2
r n
f .
n 1 t
For a given focal length, we solve for the thickness t and obtain
2
r n
t .
n 1 f
From the Höegh Meniscus to Double Anastigmats 159
We select for our example r = −10 mm, n = 1.5, and f = 100 mm. This results in a
thickness of t = 6 mm.
The principle of symmetry results in coma, lateral color, and distortion being
corrected as well but only at unit magnification. Plots of the OPDs shown in
Figure 19.4 indicate that this objective performs rather well.
The third-order and chromatic aberrations are tabulated below. They indicate
that the f/# has to be kept large due to a large spherical aberration.
With fictitious indices and Abbe numbers, and chosen surface radii, lens
thicknesses, and stop position, the starting layout prescription is summarized in
Table 19.4. We use a substrate with a high index of refraction and a low
dispersion for the positive front element A, and a material with a high dispersion
and low index material for the negative rear element B. Such a combination is
known as the “new achromat”.
Table 19.4 Prescription of starting layout. Radii and thicknesses are given in
millimeters. Notice that the outside radii are all the same, namely 26 mm.
With the help of ZEMAX, we vary the radii and substitute real glasses. For
element A, we select Schott glass N-LASF31, with nA = 1.880609 and VA =
41.0098. For element B, we pick TIFN5, with nB = 1.593555 and VB = 35.5145.
Finally, we scale the lens so that we obtain the desired 100-mm focal length. The
result is summarized in Table 19.5.
Optimizing the input data from Table 19.5 yields the configuration shown in
Figure 19.6.
The performance plots for the achromatic double lens are shown in Figure
19.7 in the form of OPDs.
The lens consists of two positive outer elements and two negative inner
elements. The aperture stop is located between the negative elements, as shown
in Fig. 19.8.
For this exercise, we decide on a 50-mm focal length and a relative aperture
of f/5 with a half-field coverage of 20 deg.
For the positive element A we pick SSKN5 as the material, and we select
LF5 for the negative element B. The indices and V values of the chosen materials
are nA = 1.658 and VA = 50.9, and nB = 1.581 and VB = 40.9, respectively. This
combination is again a “new achromat”.
Neglecting the spacing between element A and B, and assuming thin lenses,
the requirements for the achromat are
VA 50.9
A AB 0.01 0.0509 mm 1
VA VB 50.9 40.9
and
VB 40.9
B A 0.0509 0.0409 mm 1 .
VA 50.9
Therefore,
1 1
fA 19.646365 mm
A 0.0509
166 Chapter 19
and
1 1
fB 24.449878 mm .
B 0.0409
and RA2 25.854616 mm . Similarly, we calculate the radii for the equi-
concave element B and find RB1 RB 2 28.410758 mm .
We put this prescription into the computer and add a thickness of 2 mm for
the positive elements and 1 mm for the negative elements. The stop is placed
symmetrically between the two negative elements, which are 2 mm apart. After
setting the half-fields for 14 deg and 20 deg, we make the radii variables but
maintain symmetry (i.e., the radii of the elements after the stop are identified as
pick-ups). For the first optimization, we set the operand for the focal length and
use the default merit function. The result after this first optimization is indicated
in the graphs in Figures 19.9 and 19.10.
For the second optimization, we vary all radii, which breaks the symmetry.
This asymmetric configuration is now called a Dogmar.2 The improvement is
shown in Figures 19.11 and 19.12.
We end this chapter with the understanding that adding elements and varying
curvatures, thicknesses, and spacings, as well as lens element materials, improve
performance even more.
References
1. R. Kingslake, Lens Design Fundamentals, Academic Press, New York
(1978), page 236.
2. W. J. Smith, Modern Optical Engineering, 4th Edition , McGraw-Hill, New
York (2008), page 486.
Index
Abbe number, 18 focus shift
for the infrared spectra, 115 of a diffractive lens, 74
aberrations, 116 of a refractive element, 73
chromatic, 18 frequency
achromats, 39 cut-off, 147
air-spaced normalized, 147
doublet, 65 Nyquist, 149
triplet, 65 spatial, 147
Airy disk, 113
AMTIR-1, 58 Gregorian
aspherizing, 23, 52 microscope objective, 130
a singlet, 26 telescope, 127
astigmatism, 19, 104 G-sums, 15
athermats, 73
half field angle, 115
ball lens, 89 hybrid, 69
bending a lens, 11 lens, 151
best shape, 11 petzval objective, 67
for minimum spherical aberration,
23 image height, 115
impact of housing material, 80
Cassegrain
basic telescope layout, 122 limits, 147
classic system, 124 lens
with two spherical mirrors, 123 thick, 34
chromatic aberration, 18 thin, 11
cold stop, 67 LWIR region, 50
coma, 19
curvature, 11 manufacturing remarks, 109
materials for the 3–5 µm spectral
Dall-Kirkham arrangement, 124 band (MWIR), 77
detector, 16, 21 materials for the 8–12 m spectral
Dewar housing, 67 band (LWIR), 78
dialyte, 57 microscope objectives, 121
diamond-turned hybrid, 50 minimum blur, 17
diffraction limit, 113 multiple lens arrangement, 23
diffractive phase profile, 52 MWIR region, 48
f/#, 115 n, 12
focal plane array, 148 net curvature, 11
169
170 Index