6778 Vol. 61, No.

23 / 10 August 2022 / Applied Optics Research Article

Determination of the effective focal length and

cardinal points for a coaxial system of thick
lenses
Hassan A. Elagha
High Institute of Optics Technology, El-Nozha, Heliopolis, Cairo 17361, Egypt ([email protected])

Received 9 May 2022; revised 10 July 2022; accepted 13 July 2022; posted 13 July 2022; published 2 August 2022

In this work, an analytical method is proposed to determine the effective focal length and the six cardinal points
for a system of coaxial thick lenses. The technique is different from both the Gaussian and matrix methods. We can
apply it without the need to calculate the power and the principal points of the individual lenses of the system. The
method is based on a proposed formula for the focal length of a whole system of coaxial thick lenses. This formula
is derived in detail in the present work. The proposed method represents a simpler alternative to the Gaussian
reduction method. Several numerical examples are solved in this work using the method. In these examples, we have
easily calculated the effective focal length and the six cardinal points for coaxial systems of two, three, and five thick
lenses in the most general case of different refractive indices. © 2022 Optica Publishing Group
https://doi.org/10.1364/AO.463321

1. INTRODUCTION The method proposed in the present work is new and has
In 1841, Gauss published his famous book in geometrical never been published before. It is designed as a simple alternative
optics, Dioptrische Untersuchungen. In this book, he showed to the classical reduction method. In this method, we deal with
that a coaxial system of lenses could be treated as a single lens the system as a whole without any calculations concerning the
without the necessity of treating the successive surfaces of the principal points and the powers of the individual lenses. The
method is based on a new formula for the back focal length of
system. In this case, the simple thin lens formula can be applied
a combination of thick lenses. This formula is different from
to the whole system, provided that the distances of the object
the famous Gullstrand formula [2]. It is derived using paraxial
and the image are measured from two hypothetical parallel
approximations of two new ray-tracing procedures designed to
planes called the principal planes. The points of intersection of
trace meridional rays propagating through a centered system of
these planes with the optical axis are called the principal points
spherical surfaces. These procedures were derived in previous
or Gauss points. Any coaxial system of lenses is characterized by
works [8,9] and will be shown briefly in the next sections, since
six points called cardinal points. They are two focal points, two
we are going to use their equations to derive the new method
principal points, and two nodal points. The nodal points were
formulas. The two procedures are based on a new linear form
introduced by Listing in 1845 [1–5]. Historically, it is widely
of Snell’s law [9]. To show this linear form, let us consider Fig.
agreed that the Gaussian method represents a great advance in
1, in which a meridional ray is refracted at a spherical refracting
the field of geometrical optics. However, from a practical point
surface. In this case, the new linear form of Snell’s law can be
of view, the application of this method to optical systems of
written as
several coaxial thick lenses was not easy for students and optical
designers. For this reason, Smith, in the early 1930s, suggested n o r 1 = n 1r 10 , (1)
the use of matrices in paraxial ray-tracing procedures to facili-
tate the calculations [4–7]. Indeed, his ideas were not widely where r 1 and r 10 are the normal distances between the curvature
accepted until the early 1960s when researchers revived them center C and the incident and refracted rays, as shown in Fig. 1.
again, and became the focus of their attention. Today, there are In fact, Eq. (1) greatly facilitated the derivation of exact ray-
effective computer programs that can be easily used in optical tracing formulas and procedures in simple forms [9]. In all of
design calculations. Nevertheless, both the Gaussian and the these derivations, the locations of the object and the image were
matrix methods are still widely studied by undergraduate and measured relative to the center of curvature C of the optical
postgraduate students all over the world, as they represent the surface (not to its vertex) as in Fig. 1. Additionally, the distances
possible theoretical solution to the problem, and this makes “ t ” separating the successive optical surfaces were replaced by
them an important part of geometrical optics science. the distances “ d ”separating the successive curvature centers.

1559-128X/22/236778-09 Journal © 2022 Optica Publishing Group

Corrected 1 August 2022
 Research Article Vol. 61, No. 23 / 10 August 2022 / Applied Optics 6779

(5) On applying this procedure for a centered system of spheri-

cal surfaces, only Eqs. (2) and (3) are applied for each one
of the successive surfaces; then Eq. (4) is applied just one
time for the last surface #m to calculate the desired image
distance K m .
(6) For the first surface of the series ( m = 1 ), we do not need to
use Eq. (2), since it is clear (see Fig. 1) that

r 1 = K o sin θo , (6)
Fig. 1. Geometrical meaning of r 1 and r 10 .
where K o and θo are already given.
(7) We can see numerical solved examples applying the
2. NON-PARAXIAL RAY-TRACING PROCEDURES previous procedure in [9].
A. First Procedure
This procedure is based on Eq. (1), and it works for meridional
B. Second Procedure
rays. It is designed to easily calculate the position of the final
image formed by a centered system of spherical surfaces; see Fig. The second procedure was also derived in a previous work (see
2. In this case, we can calculate the distance K m separating the Sec. 6.D of [8]). It consists of the following two equations:
final image Im formed by the surface #m from the curvature  
r m−1 dm−1
center C m of this surface using a simple procedure. rm = 1− , (7)
This procedure consists of three equations that can be written µm−1 km−1
as
µm
 
1 rm rm r m−1
r m−1 = sin sin−1 − sin−1 + sin−1 .
rm = + dm−1 sin θm−1 , (2) Km rm Rm µm R m µm−1 K m−1
µm−1 (8)
rm rm This procedure can be easily applied to solve meridional ray-
θm = θm−1 − sin−1 + sin−1 , (3) tracing problems. Equations (7) and (8) are applied to each
Rm µm R m
surface of the system.
1 µm
=− sin θm . (4)
Km rm C. Paraxial Approximation
Remarks: In the previous procedure, Eqs. (7) and (8) can be reduced to a
single paraxial form written as
(1) In the previous equations, m is an integer referring to the
order of the surface in the series. The radius r is that shown n m−1 nm n m − n m−1
in Eq. (1) and, for the m th surface, it is represented as r m . + = . (9)
Km dm−1 − K m−1 Rm
(2) The relative refractive index at the m th surface is denoted
by µm , where µm = n m /n m−1 . For example, in Fig. 1, we The derivation of Eq. (9) can be found in Sec. 6.D of [8].
have µ1 = n 1 /n o . Equation (9) can be used to trace a paraxial ray through a system
(3) The distance dm represents the separation between the two of centered spherical surfaces. In this case, we apply it as a recur-
successive centers of curvatures of the m th and (m + 1) rence relation to the successive surfaces of the system. We will
th surfaces, while tm is the separation between the vertices apply Eq. (9) latter in the present work.
of these two surfaces. The relation between dm and tm is
given by the equation
3. SYSTEM OF COAXIAL THICK LENSES
dm−1 = tm−1 − (R m−1 − R m ). (5) A. Introduction

(4) In this work, we apply the sign convention mentioned in Consider a system of coaxial thick lenses as shown in Fig. 3. The
Appendix A. incident (parallel) ray OP is refracted at the first surface making
an angle θ1 with the optical axis. The final surface ( #m ) refracts
the ray towards the rear focus F making an angle θm with the
axis.
The extensions of the incident ray OP and focal ray q F
intersects at the point e 2 at the rear principal plane of the sys-
tem. In this case, the distance f = P2 F is the rear focal length
of the whole system.
Now

h = s 1 tan θ1 = f tan θm , (10)

Fig. 2. Ray tracing through a centered system of spherical surfaces. and, for very small angles, we have
 6780 Vol. 61, No. 23 / 10 August 2022 / Applied Optics Research Article

1 µm µm − 1
+ = , (21)
K m dm−1 − K m−1 Rm
where µm = n m /n m−1 ; then
µm (µm − 1)K m − R m
= . (22)
dm−1 − K m−1 Rm K m
Thus,
K m−1 − dm−1 Rm K m
Fig. 3. System of coaxial thick lenses. = . (23)
µm R m − (µm − 1)K m
1 1 Substituting from Eq. (23) into Eq. (20), we get
= (θm /θ1 ). (11)
f s1 θm K m (µm − 1)
=1+ . (24)
Thus, if we can determine the ratio (θm /θ1 ) , we can then use θm−1 R m − (µm − 1)K m
Eq. (11) to directly calculate the rear focal length of the whole Thus, we get
system.
θm Rm
= i, (25)
θm−1
h
B. Determination of the Ratio (θm /θ1 ) K m K m − (µm − 1)
Rm

To determine the ratio (θm /θ1 ) , we reduce the procedure of

i.e.,
Section 2.A to the paraxial form. This means that, in Eqs. (2),
(3), and (4), we consider that θm , θm−1 , and r m are very small θm Rm
= . (26)
(almost equal to zero). In this case, Eqs. (2)–(4) can be rewritten θm−1 R m − (µm − 1)K m
as
Equation (26) can be written for all values of m , and we can
r m−1 get the ratio (θm /θ1 ) from
rm = + dm−1 θm−1 , (12)
µm−1
θm θm θm−1 θ2
       
rm rm = × ... . (27)
θm = θm−1 − + , (13) θ1 θm−1 θm−2 θ1
R m µm R m
Therefore,
1 µm
=− θm . (14) θm
m
  Y Ri
Km rm = . (28)
θ1 R i − (µi − 1)K i
Dividing Eqs. (12) and (13) by θm−1 , we get i=2

rm r m−1
= + dm−1 , (15) C. Paraxial Ray-Tracing Formula
θm−1 θm−1 µm−1
For a system of m centered spherical surfaces, the distance s m of
θm rm rm the image formed by the m th surface can be calculated using the
=1− + . (16) formula
θm−1 θm−1 R m θm−1 µm R m
µm 1 µm − 1
Equation (16) can be written as + = . (29)
sm tm−1 − s m−1 Rm
θm µm − 1
 
rm Equation (29) was derived in Sec. 4.C of a previous work
=1− . (17)
θm−1 θm−1 µm R m [10]. The distance s m is measured relative to the vertex of the
surface #m , and µm is the relative refractive index of the m th
From Eq. (14), we get surface, where µm = n m /n m−1 . In fact, Eq. (29) is a recurrence
r m−1 relation with which we can calculate s m for the m th surface
K m−1 = − . (18) if we know s m−1 for the preceding surface # m − 1 . We will
θm−1 µm−1
see this in numerical examples later in this work. Additionally,
Substituting from Eq. (18) into Eq. (15), we get Eq. (29) will be applied in deriving the final formula of the back
rm focal length of the lens system.
= −K m−1 + dm−1 . (19)
θm−1
D. Formula for the Back Focal Length f of a Coaxial
Substituting from Eq. (19) into Eq. (17), we get System of Thick Lenses
θm µm − 1 Equation (28) can be written in the form
 
= 1 + (K m−1 − dm−1 ) . (20)
θm−1 µm R m m  
Y Ki
θ1 /θm = 1 − (µi − 1) . (30)
Now we rewrite Eq. (9) in the form Ri
i=2
 Research Article Vol. 61, No. 23 / 10 August 2022 / Applied Optics 6781

Based on Sec. 4.C of [10], we find that the distance K i for the If the two lenses are thin, then t1 = 0 , t2 6 = 0 , and t3 = 0 .
surface # i is always given by All values of si can be easily obtained using Eq. (29). In fact, Eq.
(29) can be rewritten in the simpler form
K i = si − R i . (31)
µm − 1 −1
 
1
Substituting from Eq. (31) into Eq. (30), we get sm = µm + . (39)
s m−1 − tm− 1 Rm
m   
Y si
θ1 /θm = 1 − (µi − 1) −1 Remarks:
i=2
Ri
(1) Equations (36) and (39) can be used to calculate the back
m 
Y si
 Ym 
µi (µi − 1)
 focal length f for a system of coaxial lenses. We first calcu-
= µi − (µi − 1) = si − . late the values of s 1 , s 2 , s 3 , ... , s m using Eq. (39) and then
i=2
Ri i=2
si Ri substitute for these values into Eq. (36) to get f .
(32) (2) The symbol m in Eq. (36) is the total number of optical
surfaces in the lens system.
Substituting from Eq. (29) into Eq. (32) we get
m  
Y 1
θ1 /θm = si . (33) F. Numerical Example #1
i=2
s i−1 − ti−1
Let us calculate the back focal length f for a bi-convex thick
Thus, lens (immersed in air) with the following data:
m
Y 1 R 1 = 8 cm, R 2 = −10 cm, t = 2 cm, n o = 1, n 1 = 2, n 2 = 1.
θm /θ1 = [s i−1 − ti−1 ]. (34)
s
i=2 i To solve this example, we first need to calculate the relative
Now, substituting from Eq. (34) into Eq. (11), we get refractive indices using the equation µm = n m /n m−1 .
Thus, µ1 = 2 , and µ2 = 0.5 . Now we use Eq. (39) to
m
1 1 Y1 calculate s 1 for the first surface and then s 2 for the second
= [s i−1 − ti−1 ] one. It should be remembered that always s o = ∞ (for parallel
f s 1 i=2 si
incidence). Thus,
(s 1 − t1 )(s 2 − t2 )...(s m−1 − tm−1 )
2 − 1 −1
 
= 1
s 1 s 2 ... s m s1 = 2 + = 16 cm,
∞ 8
    
1 s 1 − t1 s 2 − t2 s m−1 − tm−1 −1
. (35)

= ··· 1 0.5 − 1
sm s1 s2 s m−1 s 2 = 0.5 + = 4.11764705 cm.
16 − 2 −10
Therefore, we finally get
Now, substituting for the values of s 1 and s 2 into Eq. (36),
m−1 
we get the back focal length f 2 as follows:

1 1 Y ti
= 1− . (36)
f s m i=1 si 1

t1

1/ f 2 = 1− = 0.2125.
Equation (36) is the desired formula for the back focal length s2 s1
f of a system of coaxial thick lenses. To calculate the value of f Thus, for this lens, f 2 = 4.70588234 cm .
for the system, we need first to calculate the value of si for each
surface of the system. This can be easily done using Eq. (29). Remarks:
(1) This lens is surrounded by one medium (air), so the back
E. Special Cases and front focal lengths are numerically equal.
(2) If we calculate f using a Gullstrand equation, we get the
(1) For a single thick lens, Eq. (36) will have the form same result.
 
1 1 t1
= 1− . (37)
f s2 s1
4. SUMMARY OF THE NEW PROCEDURE
s 1 and s 2 are the distances of the images formed by the
For a system of coaxial thick lenses, we can determine the effec-
first and second lens surfaces measured to their vertices,
respectively. If the lens is thin, the thickness t1 is zero, and tive focal length, and the six cardinal points of the system by
the equation of the lens reduces to 1/ f = 1/s 2 . applying the following four steps.
(2) For a system of two coaxial thick lenses, Eq. (36) will have Step #1: Calculation of f 1 , f 2 , and β2 for the system
the form (1) We calculate the values of s m for the successive system
1 1

t1

t2

t3
 surfaces by applying the following equation to each of the
= 1− 1− 1− . (38) successive surfaces from left to right:
f s4 s1 s2 s3
 6782 Vol. 61, No. 23 / 10 August 2022 / Applied Optics Research Article

Fig. 4. Ray penetrates a coaxial system of lenses from left to right. Fig. 5. Ray penetrates a coaxial system of lenses from right to left.

µm − 1
 −1
1
s m = µm + . (40)
s m−1 − tm−1 Rm
(2) To get the back focal length f 2 for the system, we substitute
for the obtained values of s (i.e., s 1 , s 2 , ... , s m ) into the
equation
m−1  
1 1 Y ti
= 1− . (41)
f 2 s m i=1 si

(3) The front focal length f 1 can be calculated from

Fig. 6. Six cardinal points for the most general case of a coaxial
no
f1 = − f2, (42) system with different refractive indices. The drawing shows the case
nm no > nm .
(4) The effective focal length f e of the coaxial system is given
[2] by
1 no nm x = f 1 + f 2 + β2 . (48)
=− = . (43)
fe f1 f2
(5) To get the shift β2 of the back principal point P2 from the Step #4: The six cardinal points for the system
right surface vertex v2 (Fig. 4), we just substitute into the Finally, we determine the positions of the six cardinal points
equation of the system measured relative to the left vertex V1 of the left
surface of the system [1,2]. These points are F1 , F2 , P1 , P2 ,
β2 = s m − f 2 . (44) N1 , and N2 . The distances between these points and the right
vertex V1 are given by the equations
Step #2: Calculation of β1 for the system
V1 F1 = f 1 + β1 , (49)
(1) We rewrite the system data for a ray passing through the sys-
tem in the reverse direction, i.e., from right to left; see Fig. V1 F2 = f 2 + β2 + t, (50)
5. Then we apply the following equation to these successive
surfaces (one by one) from right to left to get the values of s 0 V1 P1 = β1 , (51)
:
V1 P2 = β2 + t, (52)
µm 0 − 1 −1
 
1
s m0 = µ0m 0 + . (45)
0
s m−1 − tm−1 R m0 V1 N1 = w, (53)
Now, to get the shift β1 of the front principal point P1 from V1 N2 = x + t. (54)
the left surface vertex v1 of the system (see Fig. 5), we substitute
into the equation Remarks:
β1 = − s m0 − f1, (46) (1) Equations (49–54) give the distances of the six cardinal
points measured relative to the vertex V1 . If the sign of
where s m0 is the distance of the final image formed by the sur- the distance is positive, the corresponding cardinal point is
face # m in the system, i.e., it is the last value of s 0 obtained by located to the right of the vertex V1 and vice versa.
Eq. (44). (2) t is the total thickness of the lens system.
Step #3: The nodal points N1 and N2 for the system (3) If the lens system is immersed in a single medium (air in
The shifts w and x of the nodal points N1 and N2 from the most practical cases), the nodal points N1 and N2 will
vertices v1 and v2 (Fig. 6), respectively, are given [1] as
coincide with the principal points P1 and P2 since, in this
w = f 1 + f 2 + β1 , (47) case, f 1 + f 2 = 0 .
 Research Article Vol. 61, No. 23 / 10 August 2022 / Applied Optics 6783

   
1 t1 t2 t3
1/ f 2 = 1− 1− 1−
s4 s1 s2 s3

= (−0.08834081) × (0.95238095)

× (−0.46428570) × (1.14358975)

= 0.04467120,

i.e., f 2 = 22.38578771 mm .
Then we calculate f 1 using Eq. (42) to get
Fig. 7. Coaxial system of two thick lenses. f 1 = − 22.38578771 mm.

The effective focal length f e for the system is given by substi-

(4) According to the sign convention applied in this work tuting into Eq. (43) to get
(Appendix A), it was necessary to add a negative sign to s m0
f e = 22.38578771 mm.
in Eq. (46) because, in step #2 of this summary, the light ray
passes through the system in the reverse direction, i.e., from Then we determine the shift β2 of the rear principal point P2
right to left. using Eq. (44), to get
(5) If the numerical value of β2 is negative, the back principal
point p 2 is to the left of the vertex v2 of the right surface β2 = s 4 − f 2 = −33.70558391 mm.
and vice versa. The same for β1 with respect to v1 . Step #2: Calculation of β1 for the system
First, the given data are rewritten for a ray penetrating the sys-
tem in the reverse direction, i.e., from right to left as
5. NUMERICAL EXAMPLE #2
R 10 = 5 mm, R 20 = −5 mm, R 30 = 7 mm, R 40 = −7 mm,
Let us determine the six cardinal points of a coaxial system of two
thick lenses [2]; see Fig. 7. t10 = 1 mm, t20 = 10 mm, t30 = 1 mm, n 00 = 1, n 01 = 1.5,
The following data are given:
n 02 = 1, n 03 = 1.5, n 04 = 1.
R 1 = 7 mm, R 2 = −7 mm, R 3 = 5 mm, R 4 = −5 mm, The signs of R were reversed according to the sign con-
vention (Appendix A). We also need, to calculate the relative
t1 = 1 mm, t2 = 10 mm, t3 = 1 mm, n o = 1, n 1 = 1.5, refractive indices for the four surfaces in the direction from right
n 2 = 1, n 3 = 1.5, n 4 = 1. to left. Thus, we get µ01 = 1.5 , µ02 = 0.666667 , µ03 = 1.5 , and
µ04 = 0.666667 .
Now, applying Eq. (45), to each of the four surfaces of the sys-
First, we need to calculate the relative refractive indices for the
tem from right to left (see Fig. 7), we get
four surfaces. Thus, we get µ1 = 1.5 , µ2 = 0.666667 , µ3 =
1.5 , and µ4 = 0.666667 . s 10 = 15 mm, s 20 = 4.82758625 mm,
Solution:
We apply the steps of Section 4. s 30 = −12.30468733 mm, s 40 = −24.205075 mm.
Step #1: Calculation of f 1 , f 2 , f e , and β2 for the system
To calculate β1 , we just substitute into Eq. (46) to get
We apply Eq. (40) to each of the four successive surfaces to get
the values of s . Thus, we get β1 = −s 40 − f 1 = 46.590861 mm.

1 1.5 − 1
−1 Step #3: The nodal points N1 and N2
s 1 = 1.5 + = 21 mm, The nodal points for this system coincide with the principal
∞ 7
points as the lenses are surrounded by the same medium (air),
 −1 which means that f 1 + f 2 = 0 .
1
s 2 = 0.66666667 + 0.047619 = 6.82926836 mm. Step #4: The six cardinal points
21 − 1
To get the positions of these six points measured relative to the
Similarly, we easily get left vertex V1 of the system, we just apply Eqs. (49–54) as

V1 F1 = 24.20507329 mm, V1 F2 = 0.6802038 mm,

s 3 = −6.96428549 mm, s 4 = −11.3197962 mm.
V1 P1 = 46.590861 mm, V1 P2 = −21.70558391 mm,
Now, using Eq. (41), we get
V1 N1 = 46.590861 mm, V1 N2 = −21.70558391 mm.
 6784 Vol. 61, No. 23 / 10 August 2022 / Applied Optics Research Article

The effective focal length f e of the system is given by

Eq. (43) as

f e = 6.357777877 mm.

The shift β2 of the back principal point P2 is given by

Eq. (44) as

β2 = s 6 − f 2 = −5.27691778 cm.

Step #2: Calculation of β1 for the system

First, we rewrite the given data for a ray penetrating the
surfaces from right to left according to the sign convention as
Fig. 8. Coaxial system of three thick lenses.
R 10 = 20 cm, R 20 = −20 cm, R 30 = −10 cm, R 40 = ∞ cm,

It should be noted that t in these equations is the total thick- R 50 = 10 cm, R 60 = −8 cm,
ness of the lens system, t = 12 mm .
t10 = 1 cm, t20 = 0.5 cm, t30 = 1 cm, t40 = 2 cm, t50 = 1 cm,
6. NUMERICAL EXAMPLE #3 n 00 = 1.5, n 01 = 1.8, n 02 = 1.2, n 03 = 1.7, n 04 = 1.2,
Let us determine the six cardinal points for a coaxial system of
three thick lenses that is shown in Fig. 8. This problem repre- n 05 = 2, n 06 = 1.3.
sents the most general case where the index of refraction is not The relative refractive indices for the six surfaces are
the same for all spaces.
The following data are given: µ01 = 1.2, µ02 = 0.66666667, µ03 = 1.41666667,

R 1 = 8 cm, R 2 = −10 cm, R 3 = ∞ cm, R 4 = 10 cm, µ04 = 0.70588235, µ5 = 1.66666667, µ06 = 0.65.

R 5 = 20 cm, R 6 = −20 cm, Now we apply Eq. (45) to the six surfaces (from right to left) as
we did in the previous examples to get
t1 = 1 cm, t2 = 2 cm, t3 = 1 cm, t4 = 0.5 cm, t5 = 1 cm,
s 10 = 120 cm, s 20 = 26.59217908 cm,
n o = 1.3, n 1 = 2, n 2 = 1.2, n 3 = 1.7, n 4 = 1.2,
s 30 = −424.02393312 cm, s 40 = −300.01689271 cm,
n 5 = 1.8, n 6 = 1.5.
s 50 = 26.30654322 cm, s 60 = 7.80635699 cm.
The relative refractive indices for the six surfaces are
Then β1 is given by Eq. (46) as
µ1 = 1.538462, µ2 = 0.6, µ3 = 1.416667,
β1 = −s 60 − f 1 = 0.45875425 cm.
µ4 = 0.705882, µ5 = 1.5, µ6 = 0.833333.

Solution: Step #3: The nodal points N1 and N2 for the system
We apply the steps of Section 4 as follows: The shifts w and x for the nodal points N1 and N2 , respec-
Step #1: Calculation of f 1 , f 2 and β2 for the system tively, are given by Eqs. (47) and (48) as
Applying Eq. (40) to the six surfaces from left to right as we
w = f 1 + f 2 + β1 = 1.7303094 cm,
did in the previous examples, we easily get
x = f 1 + f 2 + β2 = −4.00536263 cm.
s 1 = 22.85714391 cm, s 2 = 6.99695139 cm,
Step #4: The six cardinal points
s 3 = 7.07901448 cm, s 4 = 5.22532942 cm,
The six cardinal points can be obtained by applying
s 5 = 6.33913196 cm, s 6 = 4.25974861 cm. Eqs. (49–54) as

Now, substituting these values of s into Eq. (41), we get the V1 F1 = −7.80635699 cm, V1 F2 = 9.75974861 cm,
back focal length as
V1 P1 = 0.45875425 cm, V1 P2 = 0.22308222 cm,
f 2 = 9.53666639 cm.
V1 N1 = 1.7303094 cm, V1 N2 = 1.49463737 cm.
The front focal length f 1 is given by Eq. (42) as
It should be noted that the total thickness of the lens system is
f 1 = − 8.26511124 cm. t = 5.5 cm .
 Research Article Vol. 61, No. 23 / 10 August 2022 / Applied Optics 6785

The effective focal length f e of the system is given by

Eq. (43) as

f e = 5.100590992 mm.

Now we get β2 by applying Eq. (44) as

β2 = s 10 − f 2 = −1.48059907 cm.

Step #2: Calculation of β1 for the system

First, we rewrite the given data for this case (from right to left)
according to our sign convention as
Fig. 9. Coaxial system of five thick lenses.
R 10 = 5 cm, R 20 = ∞ cm, R 30 = 10 cm, R 40 = −10 cm,
7. NUMERICAL EXAMPLE #4
R 50 = −10 cm, R 60 = 20 cm, R 70 = ∞ cm, R 80 = −20 cm,
Let us determine the cardinal points for a coaxial system of five
thick lenses that is shown in Fig. 9. The following data are given: R 90 = 25 cm, R 10
0
= −15 cm,
R 1 = 15 cm, R 2 = −25 cm, R 3 = 20 cm, R 4 = ∞ cm, t10 = 0.4 cm, t20 = 0.2 cm, t30 = 0.6 cm, t40 = 0.5 cm,
R 5 = −20 cm, R 6 = 10 cm, R 7 = 10 cm, R 8 = −10 cm, t50 = 1 cm, t60 = 2 cm, t70 = 1 cm, t80 = 2 cm, t90 = 1 cm,
R 9 = ∞ cm, R 10 = −5 cm, n 00 = 1, n 01 = 2, n 02 = 1, n 03 = 1.5, n 04 = 1.3, n 05 = 1.8,
t1 = 1 cm, t2 = 2 cm, t3 = 1 cm, t4 = 2 cm, t5 = 1 cm, n 06 = 1.2, n 07 = 1.6, n 08 = 1.3, n 09 = 1.7, n 010 = 1.2.
t6 = 0.5 cm, t7 = 0.6 cm, t8 = 0.2 cm, t9 = 0.4 cm, The relative refractive indices for the six surfaces are
n o = 1.2, n 1 = 1.7, n 2 = 1.3, n 3 = 1.6, n 4 = 1.2, µ01 = 2, µ02 = 0.5, µ03 = 1.5, µ04 = 0.86666667,

n 5 = 1.8, n 6 = 1.3, n 7 = 1.5, n 8 = 1, n 9 = 2, n 10 = 1. µ05 = 1.38461539, µ06 = 0.66666667, µ07 = 1.33333333,

The relative refractive indices for the 10 surfaces are µ08 = 0.8125, µ09 = 1.30769231, µ0 10 = 0.70588235.
µ1 = 1.41666667, µ2 = 0.76470588, µ3 = 1.23076923,
Now we apply Eq. (45) to the 10 surfaces (from right to left)
µ4 = 0.75, µ5 = 1.5, µ6 = 0.72222222, µ7 = 1.15384615, to get

µ8 = 0.66666667, µ9 = 2, µ10 = 0.5. s 10 = 10 cm, s 20 = 4.8 cm, s 30 = 5.60975609 cm,

Solution: s 40 = 4.06993051 cm, s 50 = 5.72969687 cm,

We apply the steps of Section 4 as follows.
Step #1: Calculation of f 1 , f 2 and β2 for the system s 60 = 3.42295708 cm, s 70 = 1.8972761 cm,
Applying Eq. (40) to the 10 surfaces as we did in the previous
examples, we easily get s 80 = 0.72295535 cm, s 90 = −1.69664849 cm,

s 1 = 50.99999971 cm, s 2 = 25.99999973 cm, s 0 10 = −2.00978487 cm.

s 3 = 23.13252991 cm, s 4 = 16.59939743 cm, Then we get the shift β1 from Eq. (46) as

s 5 = 34.48594746 cm, s 6 = 346.30729079 cm, β1 = −s 10

0
− f 1 = 8.13049406 cm.

s 7 = 63.13312312 cm, s 8 = 13.51583557 cm, Step #3: The nodal points N1 and N2 for the system
The shifts w and x for the nodal points N1 and N2 , respec-
s 9 = 26.63167114 cm, s 10 = 3.61999189 cm. tively, are given by Eqs. (47) and (48) as
Now, substituting these values of s into Eq. (41) we get the w = f 1 + f 2 + β1 = 7.11037583 cm,
back focal length f 2 for the system as
x = f 1 + f 2 + β2 = −2.5007173 cm.
f 2 = 5.10059096 cm.
Then we calculate the front focal length f 1 from Eq. (42) as Step #4: The six cardinal points
To get the positions of these six points measured relative to the
f 1 = − 6.12070919 cm. left vertex V1 of the system, we just apply Eqs. (49–54) as
 6786 Vol. 61, No. 23 / 10 August 2022 / Applied Optics Research Article

V1 F1 = 2.00978487 cm, V1 F2 = 12.31999189 cm, (4) The distance S1 separating the image I and the vertex V
of the surface is positive when the image is located to the
V1 P1 = 8.13049406 cm, V1 P2 = 7.21940093 cm, right of the refracting surface and vice versa.
(5) The distance K o separating the object and the center of
V1 N1 = 7.11037583 cm, V1 N2 = 6.1992827 cm. curvature C for a refracting surface is positive if the object
The total thickness of the lens system is t = 8.7 cm. is to the left of C and vice versa.
(6) The distance K 1 separating the image and the center of
curvature C for that surface is positive if the image is to the
8. CONCLUSION right of C and vice versa.
This work proposes a new reduction method for a system of
coaxial thick lenses. The method is based on a new formula for Disclosures. The author declares no conflicts of interest.
the focal length of a coaxial system of thick lenses. This formula Data availability. No data were generated or analyzed in the present
is derived in detail in the present work. The procedure of the research.
method is simple and can be easily applied without any details,
and we do not need to consider the power and the principal
REFERENCES
points of the system individual lenses. In the present work,
we have applied this method to the case of five coaxial thick 1. F. L. Pedrotti and L. S. Pedrotti, Introduction to Optics (Prentice-Hall,
1993).
lenses in the most general case and obtained the results without 2. E. L. Dereniak and T. D. Dreniak, Geometrical and Trigonometric
any difficulties. The application of the method requires only Optics, 1st ed. (Cambridge University, 2008).
numerical substitution in a single ray-tracing recurrence relation 3. K. K. Sharma, Optics, Principles and Applications (Academic, 2006).
(Eq. 40) and, finally, we obtain the desired results using very 4. E. Hecht, Optics, 4th ed. (Addison Wesely, 2002).
simple equations without any complications. 5. R. Kingslake and R. B. Johnson, Lens Design Fundamentals
(Academic, 2010).
6. A. E. Conrady, Applied Optics and Optical Design (Dovers, 1992).
7. A. Gerrard and J. M. Burch, Introduction to Matrix Methods in Optics,
APPENDIX A 1st ed. (Wiley, 1975).
Sign convention 8. H. A. Elagha, “Generalized formulas for ray tracing and longitudinal
spherical aberration,” J. Opt. Soc. Am. A 34, 335–343 (2017).
(1) Light rays always pass from left to right. 9. H. A. Elagha, “Analytical and graphical ray-tracing techniques based
(2) The curvature radius, R , is positive for a convex surface on a classical principle for photon refraction,” J. Opt. Soc. Am. A 34,
2222–2230 (2017).
and vice versa.
10. H. A. Elagha, “Exact ray tracing formulas based on a trigonometric
(3) The distance So separating the object and the vertex V of a alternative to Snell’s law,” J. Opt. Soc. Am. A 29, 2679–2687 (2012).
surface is positive when the object is located to the left of V
and vice versa.