Académique Documents
Professionnel Documents
Culture Documents
Javier Alda
University Complutense of Madrid, Madrid, Spain
Fig. 1 Trigonometric functions. The sine is in blue reaching 1 When the paraxial approach is applied to the Snell law,
when the angle is 90°, the cosine is in red and departing from 0 n sin e ¼ n0 sin e0
1 for 0°, and the tangent is in green, always larger than the sine
function. The paraxial approach for the sine and the tangent is it produces the following paraxial relation:
the dashed straight line. The paraxial approach for the cosine has
a constant value of 1. ne ¼ n0 e0
This equation makes it possible to treat refraction as a
linear transformation of angles. The reflection law
These figures present the departure of the paraxial
approach with respect to the actual behavior. e00 ¼ e
is linear already. This linearization is formalized within
GEOMETRICAL PARAXIAL OPTICS the scope of matrix optics, where a ray within an optical
system is characterized in terms of the height with respect
Geometrical optics is based on a few axioms: the existence to the optical axis and its slope.
of the incidence plane containing the normal to the in-
terface, the incident, the reflected, and the refracted rays; Correspondence Equations and
the Snell law and the reflection law, easily derived from Image-Forming Systems
the Fermat’s Principle; the Fermat’s Principle itself; and
the reversibility of the optical paths. Although no one of The definition of a perfect optical system is based on the
these axioms involves the paraxial approach, the paraxial stigmatism concept. This can be formulated as the
optics is traditionally linked with the basis of geometrical constancy of the optical path along any light trajectory
optics. These axioms are the actual foundations of the
calculation of light trajectories—the goal of the geomet-
rical optics. The first axiom, defining the incidence plane,
has important consequences for the simplification of the
treatment. It allows to treat rotationally symmetric optical
systems, after neglecting the effect of the skew rays, only
by analyzing the ray trajectories in a meridian plane that
contains the optical axis of the system. This axiom is
implicitly used to draw in a two-dimensional plot the ray
tracing of a three-dimensional optical system.
More specifically, paraxial optics appears as the re-
gime where the concept of perfect optical system applies.
There are three conditions for an optical system to be
considered as perfect: Every object point corresponds
Fig. 2 Relative error between the paraxial and the exact
with an image point, then every ray departing from an trigonometric functions. The blue curve is for the sine (always
object point arrives to the corresponding, conjugated, negative, i.e., the paraxial approach is larger than the exact
image point; every plane in the object space is imaged function), the green is for the tangent (always positive, i.e., the
onto another plane in the image space; and a figure lo- paraxial approach is smaller than the exact function), and the red
cated in an object plane produces an image having a size one is for the cosine (plotted on the right).
1922 Paraxial Optics
from the object point to the image point. The optical path The shape of the optical surfaces is usually a sphere
is defined as the following integral due to its easy manufacture and testing. When the diopter
Z is a spherical surface having a radius of curvature of r,
L ¼ nð~
rÞd~r the geometry is the same with that of the previous fi-
C gure. The optical path in terms of the frontal distances s
where nð~ rÞ represents the index of refraction at a point ~r and s’, r, and the angle j becomes
that belongs to the light trajectory, C. The light trajectory qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
may travel along different materials. The index of L ¼ n r 2 þ ðr sÞ2 2rðr sÞ cos j
refraction characterizes the propagation properties of the qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
light, and if these materials are linear, homogeneous, and
þ n0 r 2 þ ðs0 rÞ2 2rðs0 rÞ cos j ¼ K
isotropic, then nð~rÞ is constant along the light trajectory
inside the same material. In this case, the optical path is
When the stigmatism condition about the constancy of
merely the product of the index of refraction times the
the optical path is applied, the result is
geometrical path along the light trajectory in the media.
A general case of stigmatism can be considered from
n n0 1 ns n0 s0
Fig. 3, where a curved diopter is the interface between þ ¼ þ 0
r r0 r r r
two linear, homogeneous, and isotropic media having
index n and n’. where r and r’ are the object and image optical paths
The astigmatism condition is written as the constancy reaching the diopter at the incidence point. In this for-
of optical path for any arbitrary light trajectory from O to mula, it is possible to apply the paraxial approach to the
O’. This condition is: trigonometric functions included in the expressions of r
L ¼ nr þ n0 r0 ¼ K and r’ to obtain the following correspondence equation:
Sign convention
Fig. 5 Definition and use of the focal object and the image focal points. They are the corresponding image and object for the infinity
points on the axis, respectively. If the object or the image are in the infinity, but located off_axis, then the image or the object,
respectively, are on the focal plane.
1924 Paraxial Optics
b0 ¼
f
¼
f z0
¼ 0 ¼
a0 f 0
P
z af f f0
Combining Paraxial Elements Fig. 9 A lens is a combination of two curved diopters with
Within the Paraxial Approach radius, r1 and r2, thickness (t), and fabricated with a material
showing an index of refraction of n.
Paraxial optics is able to deal with a combination of
diopters to find out what are the characteristic parameters
of the combination. The goal is to obtain the location of the where f1’ and f2’ are the image focal distances of the
focal and principal planes of a system formed by a individual systems, n2 is the index of refraction of the
combination of diopters or other optical subsystems. The medium between the systems, n’2 is the index of refraction
first step in the process is to know how to combine two of the image space of the combined system, and t is the
optical systems, each one having their own principal and distance between the image principal plane of the first
focal planes. In Fig. 8, we have presented the graphical system and the object principal plane of the second system.
solution to this problem. To obtain it, we have traced a ray The location of the principal planes of the combined
parallel to the optical axis, coming from infinity and system is given by
reaching the focal point of the compounded system, F’,
after passing through the two optical systems. The f 0t
H20 H 0 ¼
intersection of the input and the output rays provides a f10
point belonging to the image principal plane of the total ft
system, H’. To obtain the object focal and principal planes, H1 H ¼
f2
F and H, we have traced a ray coming from the right and
passing through the object focal point, F. The analytical One of the most useful applications of this method is to
solution relates the involved magnitudes to provide the find the characteristics of a thick lens. A lens is a
following results. The image focal distance is given by combination of two diopters separated by a distance equal
to the central thickness and having an internal index of
1 n2 1 1 t
¼ 0 0þ 0 0 0 refraction due to the material that the lens is made of
f0 n2 f1 f2 f1 f2 (Fig. 9). The principal planes for the subsystems are
located on the vertex of the refracting surfaces. Then if we
assume that the lens is immersed in air, the value of the
focal length is given as
1 1 1 ðn 1Þ2 t
¼ ðn 1Þ þ
f0 r1 r2 n r1 r2
If the central thickness can be considered negligible, then
the first term provides the focal length of the lens in
the thin lens approximation. The previous equation and
the thin lens approximations are sometimes known as the
lens-maker formula.
Prisms
d ¼ e1 e02 a
where the dependence with the index of refraction appears It is important to note that due to the convention sign used
when calculating the angle of refraction at the output in- in the definition of the previous parameters, the slope of
terface of the prism, e2’. By using the paraxial Snell law, it the ray, o, and the angle with respect to the axis, s, have
is possible to relate e1 with e1’ and e2 with e2’. If this opposite signs. In Fig. 11, h, h’, and o, are positive and o’
is done in the previous equation and using the relation is negative. An illustration of the application of the matrix
with the angle of the prism a = e1’ e2, the previous optics is shown in Fig. 12 for the case of a spherical
equation becomes surface having a radius of curvature, r, that provides the
following ABCD matrix
d ¼ ðn 1Þa
!
which is the paraxial form of the deviation of a prism 1 0
A B
immersed in air. ¼ 0 n
C D
n n
n0 r n0
Matrix Optics and Paraxial Optics
A derivation of the matricial relation provides also the
The first equation of paraxial optics that we found in this object and image distance relation. This relation is known
contribution was the linearization of the Snell law. This li- also as the ABCD law
nearization can be extended and fully completed when
using the matricial formulation of paraxial optics. In this As B
s0 ¼
sense, a given optical system can be seen as a transformer Cs þ D
that changes linearly the characteristic parameters of a
given light trajectory (Fig. 11). These characteristic pa- When the elements of the matrix for a curved diopter are
rameters are the height and slope of the ray with respect to replaced in the ABCD law, it is possible to obtain the
the optical axis of the optical system. This matricial re- correspondence paraxial equation for a curved diopter.
lation is written as The main advantage of matrix optics is that the com-
bination of optical elements is easily done by the matrix
h0 A B h multiplication of the individual matrices of the optical
¼
o0 C D o elements. This modularization, along with some exten-
sions applied to laser beam propagation and array optics,
makes matrix optics a powerful tool for the paraxial ana-
lysis of optical systems.
PARAXIAL APPROACH IN
PHYSICAL OPTICS
Huygens – Fresnel Integral in the The paraxial approach applied here is related with the
Paraxial Approximation degree of this expansion. When only linear terms in x’
where r is the radius of curvature. The term 1/r is res- then the sagita is totally neglected. When the next term of
ponsible for the attenuation of the amplitude after pro- the expansion is taken into account, it provides the
pagating from the source. The phase term varies very fast following expression for the sagita
along the propagation and configures the spherical shape x2
of the wavefront. sag ¼
2r
A sphere in a meridian plane is expressed as
A thick lens can be seen as a phase plate introducing a
ðx2 þ z2 Þ ¼ r 2 phase change variable along the transversal direction.[23]
Then any incoming wavefront will change its phase ac-
Then the phase term expressed in Cartesian coordinates
cordingly to the value of this phase screen. To account for
contains a square root. Usually, we are interested in the
the dependence with the transversal coordinate, x, for a
spherical waves that propagate along coordinate, z, being
thick lens, we need to calculate the optical paths at any
x the transversal coordinate. Mostly, the propagation
height from the input plane (tangent to the vertex of the
distance from the source is larger than the transversal
first surface of the lens) to the output plane (tangent to the
dimension of the optical layout. Then the square root can
vertex of the second surface of the lens). This calculation
be properly written as
needs the values of the sagita for both surfaces to build up
rffiffiffiffiffiffiffiffiffiffiffiffiffi
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi x2 the thickness function as
r ¼ 2 2
x þz ¼ z 1þ 2
z x2 x2
tðxÞ ¼ t0 þ
that can be expanded also in powers until second order 2r1 2r2
The phase screen is obtained as (n 1)t(x), where n is the
x2
r ffi z 1þ 2 index of refraction of the material of the lens, assuming it
2z
is immersed in air (Fig. 16). This product contains the
If we only take the first term, the spherical dependence is following term
lost and the phase behaves as a plane wave. This approach
1 1 x2 x2
can be valid for the 1/r term of the amplitude. Then the ðn 1Þ ¼
very first approximation to the spherical wavefront is r1 r2 2 2f 0
given by the following form where we have made use of the lens-maker formula to
write down the focal distance of a thin lens, f ’, within the
1 2p x2
Eðx; zÞ ¼ exp i paraxial approach.
z l 2z
This equation represents a spherical wavefront having its
center at a distance z from the observation plane. It is BEYOND PARAXIAL APPROACH
interesting to notice that this previous equation has re-
placed the spherical wavefront by a parabolic wavefront. The departure of the paraxial approach may be interpreted
This approximation will be related with the paraxial ap- as a departure from a perfect optical system. This depar-
proach previously presented in geometrical optics. ture is known as aberration. The analysis of the optical
aberrations is made in terms of their third-order approach.
Paraxial Thin Lens as a Phase Screen
This analysis can be made in terms of ray tracing and also 2. Klein, M.V.; Furtak, T.E. Optics, 2nd Ed.; John Wiley &
in terms of the wavefront aberration. Both types of ana- Sons: New York, 1986.
lysis are widely used in the fine optimization routines for
designing optical systems.
3. Keating, M.P. Geometrical, Physical and Visual Optics;
Butterworths: Stoneham, MA, 1988.
P
4. Tunnacliffe, A.H.; Hirst, J.G. Optics; The Association of
British Dispensing Opticians: London, 1990.
5. Freeman, M.H. Optics, 10th Ed.; Butterworths: London,
CONCLUSION 1990.
6. Guenther, R. Geometrical Optics. In Modern Optics; John
Paraxial optics is based on the paraxial approach. This Wiley & Sons: New York, 1990; 129 – 212.
approach is a linearization of the trigonometric functions 7. Blaker, J.W.; Rosenblum, W.M. Optics. An Introduction
used in the description of the optical systems and their for Students of Engineering; Macmillan Publishing Com-
related phenomenology. pany: New York, 1993.
Geometrical optics uses very intensively the paraxial 8. Goodman, D.S. General Principles of Geometrical Optics.
approach to provide a first approximation to the behavior of In Handbook of Optics; Bass, M., Ed.; McGraw-Hill: New
light inside optical systems. Paraxial optics allows making York, 1995; Vol. 1. Chapter 1.
9. Pedrotti, F.L.; Pedrotti, S.L. Introduction to Optics, 2nd
paraxial ray tracing. It defines and uses the concepts of
Ed.; Prentice Hall International: Englewood Cliffs, NJ,
cardinal points and planes: focal, principal, and nodal.
1996.
Paraxial optics is very well adapted to describe an optical 10. Hecht, E. Optics, 3rd Ed.; Addison Wesley: Massachusetts,
system as a perfect optical system. Moreover, the departure 1998.
of the paraxial behavior is taken as the departure of the 11. Boreman, G.D. Geometrical Optics. In Basic Electro-
perfect optical system behavior. However, the chromatic Optics for Electrical Engineers; SPIE Optical Engineering
aberrations can be easily treated inside the paraxial Press: Bellingham, WA, 1998; 1 – 22.
approach. Some other important concepts, such as the 12. Herzberger, M. Modern Geometrical Optics; Robert E.
aperture and the field of an optical system, can be Krieger Publishing Company: Huntington, NY, 1980.
understood also inside the paraxial framework. The matrix 13. Loshin, D.S. The Geometrical Optics Workbook; Butter-
optics is a complete formalization of the paraxial approach. worth-Heinemann: Stoneham, MA, 1991.
14. Katz, M. Introduction to Geometrical Optics; Penumbra
The paraxial approach also applies to physical optics
Publishing Company: New York, 1994.
concepts. Then the reflectance and the transmittance of
15. Mouroulis, P.; Macdonald, J. Gaussian Optics. In Geo-
light for small angles are obtained by using paraxial metrical Optics and Optical Design; Oxford University
concepts. There exists a paraxial wave equation that de- Press: New York, 1997; 62 – 93.
scribes the propagation of an optical field when its plane- 16. Ditteon, R. Paraxial Optics I and II. In Modern Geo-
wave spectrum is within a cone of about 30° of half-angle. metrical Optics; John Wiley & Sons: New York, 1998;
This is the case for most of the light beams actually 82 – 166.
propagating along optical systems. The paraxial form of the 17. Buchdahl, H.A. An Introduction to Hamiltonian Optics;
kernel of the Huygens – Fresnel integral yields the Fresnel Dover Publication Inc.: New York, 1970.
and the Fraunhofer diffraction equations. Moreover, when 18. Stavroudis, N. The Optics of Rays, Wavefronts, and
the elements of the ABCD matrix are included in the Caustics; Academic Press: New York, 1972.
19. Welford, W.T. Aberrations of Optical Systems; Adam
paraxial form of the kernel of the Huygens – Fresnel
Hilger: Bristol, 1986.
integral, we finally find the strong links between geo-
20. Mahajan, V.N. Aberration Theory Made Simple; SPIE
metrical and physical paraxial optics. Finally, as another Optical Engineering Press: Bellingham, WA, 1991.
example of this close relation, the description of a thin lens 21. Born, M.; Wolf, E. Principles of Optics, 6th Ed.;
as a phase screen is shown using the paraxial approach in Cambridge University Press: Cambridge, MA, 1998.
the definition of the thickness function of the lens. 22. Mahajan, V.N. Optical Imaging and Aberrations; SPIE
The extensions of the paraxial optics correspond with Optical Engineering Press: Bellingham, WA, 1998.
the analysis of the optical aberrations. This study can be 23. Goodman, J.W. Introduction to Fourier Optics, 2nd Ed.;
carried out with the use of a geometrical point of view by McGraw-Hill: New York, 1996.
means of the exact ray-tracing calculation, or by the 24. Siegman, A.E. Lasers; Oxford Sciences Books: Mill
formalism dealing with the propagation of optical fields Valley, CA, 1986.
25. Brower, W. Matrix Methods of Optical Instrument Design;
beyond paraxial approach.
Benjamin: New York, 1964.
26. Gerrad, A.; Burch, J.M. Introduction to Matrix Method in
Optics; John Willey & Sons: New York, 1975.
REFERENCES 27. Wang, S.; Zhao, D. Matrix Optics; Springer-Verlag:
Berlin, 2000.
1. Ghatak, A.K.; Thyagarajan, K. Paraxial Ray Optics. In Con- 28. Spiegel, R.J.; Liu, M. Mathematical Handbook of Formu-
temporary Optics; Plenum Press: New York, 1978; 1 – 30. las and Tables, 2nd Ed.; McGraw-Hill: New York, 1998.