Paraxial ray analysis of a cat's-eye

retroreflector
J. J. Snyder

The cat's-eye retroreflector is a passive optical system consisting of a secondary mirror placed at the focal

point of a primary lens. We analyze the cat's eye using the paraxial ray matrix approach. The position of
the equivalent reflecting surface and the angular field of view of a realizable cat's eye are functions of the radius of curvature of the secondary mirror. The field of view is maximum for a secondary mirror with a concave radius of curvature equal to the focal length of the primary lens. We further derive the general

dependence of retroreflection errors on misadjustment of the secondary mirror.

Introduction

B.

An ideal retroreflector is a passive optical system that returns each incident light ray at an angle of reflection exactly opposite the angle of incidence. Practical realizations of retroreflectors can be either the corner cube or the focusing system commonly called a cat's eye: a primary lens (or mirror) with a secondary mirror located at the focus of the primary corner cube appears to be more common and more extensively studiedlA although for some applications the cat's eye offers certain advantages. For example, the ability of a properly designed cat's eye to reflect polarized light without modifying its polarization state preserves the effectiveness of optical isolators in preventing laser feedback. Our approach is to describe a general cat's-eye sys5 tem using the paraxial ray matrix method and to develop the equations showing the retroreflection errors induced by misadjustment of the secondary mirror. From the analysis we find that there is an optimum class of cat's-eye configurations for which the angular
field of view is maximum.
Paraxial Ray Matrix

CD J

The output ray vector is related to the input ray vector by

ro'I cD

Ir

(1)

(see Fig. 1). Of the two types of retroreflectors,

the

where r and r' are the (radial) height and slope of the input ray at the entrance plane, and ro and ro' are the height and slope of the output ray at the exit plane. The same matrix correctly describes the transformation of a Gaussian beam by the optical system according to the ABCD law6:
Aq
+ B

Cq + D'

(2)

where q, the complex curvature of the Gaussian beam, is related to the radius of curvature R and size w of the beam by
q = R1 1 . X
W.

(3)

The paraxial ray matrix approach is a convenient tool for describing linear combinations of simple optical elements. An arbitrary, axially symmetric optical system may be described by the matrix

The spot size co is the radius for which the electromagnetic field drops to l/e of the peak amplitude. Useful matrices are those describing propagation
through a distance d,

[ d], 1o 1J and through a thin lens of (positive) focal length f,

The author is with the Joint Institute for Laboratory Astrophysics, National Bureau of Standards and University of Colorado,
Boulder, Colorado 80302. Received 14 February 1975.

Mirror reflections are treated by unfolding the optical ray about the mirror surface so a spherical mirror
OPTICS August 1975 / Vol. 14, No. 8 / APPLIED 1825

A=

2d

2d

2d2

fR

B = 2d - 2d
C =
(a)

2 R

4d + 2d R T
2d+

2d2 '
isRK

(5)

D=1-2d

fR

fR

2d

The optical system fulfills the definition of a retroreflector,

r
( b)

= -r

(6)

if and only if the latter two elements of the matrix are

D = -1

c=1

(7)

The solution of Eqs. (5) and (7) requires the mirror placed at the focus of the lens [see Fig. 2(b)],
d = ,
(c )
Fig. 1. Retroreflectors: (a) corner cube; (b) cat's eye with pri(8)

so the general matrix for the cat's-eye retroreflector

is A B J=
1,

mary lens; (c) cat's eye with primary mirror. (b) and (c) are formally equivalent in the paraxial approximation.

2d

2d2]

(9)

with radius R (R > 0 for concave) is described by the same matrix as a thin lens of focal length R/2,
F 1 01] L-(2/R) 1J

= = -_
rTO{<

A plane mirror is described by the unit matrix

l Cat's Eye

rlj' , I

I
R

[tn. 1(a)

OPTIC AXIS

We consider an optical system consisting of a (thin) lens of focal length f and a mirror of radius R
at a distance d from the lens [see Fig. 2(a)]. A ray

d
I

entering the primary lens with (radial) height r and slope r' is reflected by the secondary mirror. The reflected ray exits the primary lens with height ro and slope r. If the entrance and exit surfaces are taken to be coincident with the lens, the ray matrix for the system is the product of the matrices of the separate components: IC DJ

Iriwl I

[A B

[10
l-/

[dl
ii Lo 1

F
l-2/R

F dl F'0 1. 11 Lo I l1/
0
f 1J
(4)

(b) Fig. 2. Optical system: (a) The primary lens of focal length f is

The elements of the matrix for the optical system are found to be
1826 APPLIEDOPTICS/ Vol. 14, No. 8 / August 1975

separated from the secondary mirror of radius R by a distance d. At the lens surface the input ray of height r and slope r' is transformed by the system into the output ray of height r and slope r'. (b) The cat's-eye configuration. The mirror is located at a distance d = f from the primary lens.

r'o = -2 (1
-

) + C

+ Dri',

(12)

11 I'I 2a

where C and D are the matrix elements given in Eqs.

(5).

The sensitivity of the output-ray slope to a small axial displacement of the secondary mirror from the focal position (focusing error) may be found by differentiating Eq. (12) with respect to the mirror position d and evaluating for d = f.
ad + +( ) (13)

Fig. 3.

The effect of tilt of the secondary mirror.

The slope of

the intermediate ray

Ir I is changed by 2a (dashed line) if the mirror is tilted by a. simplicity, a plane secondary mirror is shown. For

The total angular error in an improperly adjusted

cat's eye is, therefore,
A'VOtId,,~f

2 a
+ 2 + 2

)]Ad.

(14)

A retroreflector appears to each incident ray as a plane, perpendicular, reflecting (and inverting) surface. The distance of the apparent reflecting surface of a cat's eye, measured from the primary lens, is 2 =?R- d.
B d2

(10)

For example, a plane secondary mirror in a cat's eye produces an apparent reflecting surface at a distance d = f in front of the primary lens, whereas a cat's eye with a concave secondary mirror of radius R = f has an apparent reflecting surface coincident with the primary lens.
Alignment Errors

The first term in the brackets in Eq. (14) introduces a wedge between the input and output rays. The effect of this term may be minimized in practice by adjusting the tilt of the secondary mirror to be small compared with the ratio of the primary lens aperture to the focal length. The second term has a lenslike effect, since it is proportional to the input ray height r. This term can be thought of as due to a finite radius of curvature of the equivalent reflecting surface of the cat's eye. Its effect may be reduced only be restricting the primary-lens aperture (increasing the f number). The final term, which is proportional to the input
ray slope ri', will limit the angular field of view of the

In our model, the two possible types of misadjustment of the cat's eye are displacement of the secondary mirror along the optical axis away from the focus position and tilt of the secondary mirror. Displacement of a spherical secondary mirror transverse to the optical axis is equivalent to tilt. In order to analyze the effects of secondary mirror tilt, we consider the general lens-mirror optical system described by the matrix of Eqs. (4) and (5). In the absence of tilt the intermediate ray,

cat's eye. However, because of the dependence of this term on the radius of curvature R of the secondary mirror, we may construct a cat's eye for which the field of view is insensitive in first order to axial positioning errors of the secondary mirror. For this optimized cat's eye the mirror radius is equal to the focal length of the primary lens. The ray matrix for the properly focused cat's eye then has the simple form

and the total error term for a slightly defocused cat's

eye is, from Eq. (14),

Ar, dRf
Example

( a + 2ri) Ad.

(16)

(see Fig. 3), reflected from the secondary mirror, is transformed into the output ray according to

Iroh

r =[ 1, d

-1/, 1 - d/f]

Ir
r/

(11
(11)

A tilt 'of the secondary mirror by a small angle a changes the slope r' of the intermediate ray by 2a, so the total slope of the output ray becomes, from Eqs.
(1) and (11),

As a numerical example we consider a primary lens of 5-cm diam and 20-cm focal length. If we assume the retroreflected wavefront spherical error due to the second term in Eq. (14) may be adjusted to be of the order X/10, the secondary mirror position will be correct to within
IAdI
3.84 Mim. (17)

If we (arbitrarily) require all other error terms to be'

August 1975 / Vol. 14, No. 8 / APPLIEDOPTICS 1827

less than 1% of the sphericity,' the tilt of the secondary mirror must be less than

Ilax I._ - 1.25 x

10-3.

(18)

We optimize this cat's eye by choosing the radius of curvature of the concave secondary mirror to be equal to the 20-cm focal length of the primary lens. The equivalent surface of reflection of the optimized cat's eye is coincident with the primary lens, and the field of view, in our model, is unlimited.7 However, if the secondary mirror were, for example, plane, the equivalent reflecting surface would be located 20 cm in front of the primary lens, and the field of view
would be only
Iima A1.25 x Conclusion
10-3.

to small misadjustments of the secondary mirror were analyzed. We found that the field of view is maximized for a cat's eye having a concave secondary mirror with radius of curvature equal to the focal length of the primary lens. The author is grateful to John Hall for many useful conversations. This work was supported in part by the National Science Foundation through grant 39308X to the University of Colorado. The author is an NRC-NBS postdoctoral research associate.
References
1. R. F. Chang, D. G. Currie, C. 0. Alley, and M. E. Pittman, J. Opt. Soc. Am. 61, 431 (1971). 2. H. D. Eckhardt, Appl. Opt. 10, 1559 (1971). 3. F. Stenman, Comment. Physico-Mathematicae 42, 39 (1972). 4. R. Beer and D. Marjaniemi, Appl. Opt. 5, 1191 (1966).

(19)

5. A. Yariv, Introduction to Optical Electronics (Holt, Rinehart and Winston, New York, 1971), pp. 18-49 and references therein. 6. H. Kogelnik, Appl. Opt. 4, 1562 (1965). 7. In fact, lens aberration (which we do not consider here) or the

We have derived the paraxial ray matrix for the general form of a cat's eye: a secondary mirror located at the focal surface of a primary lens or mirror. The errors in the angle of the retroreflected ray due

secondary mirror diameter would limit the effective field of view of the optimized cat's eye.

Comment on filler in AO

Awards available here

Hellmut Hanle

Sonderprogramm zur Frderung der fachbezogenen Zusammenarbeit zwischen Forschungsinstituten in der Bundesrepublik Deutschland und in den Vereinigten Staaten von Amerika, Sekretariat der Alexander von
Humboldt-Stiftung, D-53 Bonn-Bad Godesberg, Schil-

lerstrasse 12, Germany. The notice in the February issue [Appl. Opt. 14, 484 (1975)] has resulted in numerous applications being made by American scientists for a Senior U.S. Scientist Award. Under this U.S. Special Program self-applications are not
possible.

The basic idea of the program is to distinguish outstanding American scientists with a so-called Senior U.S. Scien-

tist Award in recognition of their past accomplishments in research and teaching. This award entitles the recipient to stay for an extended period at a research institute in the Federal Republic of Germany carrying out research of his
own choice. We would be very much obliged to you if you could pos-

sibly print a correction to this notice in one of the future editions of Applied Optics, pointing out the basic idea of the U.S. Special Program.
1828 APPLIED OPTICS/ Vol. 14, No. 8 / August 1975