Chapter

Twenty FRESNEL DIFFRACTION

One of your commissioners, M. Poisson, had deduced from the integrals reported by the author
[Fresnel] the singular result that the centre of the shadow of an opaque circular screen must,
when the rays penetrate there at incidences which are only a little oblique, be just as illuminated
as if the screen did not exist. The consequences have been submitted to the test of a direct
experiment, and observation has perfectly confirmed the calculation.
—Dominique Arago to the French Academy of Sciences1

Important Milestones
1816 Augustin Fresnel developed the theory of diffraction using the wave theory of light.
1817 Using Fresnel's theory, Poisson predicted a bright spot at the center of the shadow of an opaque disc—
this is usually referred to as the Poisson spot.
1818 Fresnel and Arago carried out the experiment to demonstrate the existence of the Poisson spot, validating the
wave theory.
1874 Marie Cornu developed a graphical approach to study Fresnel diffraction—this came to be known as Cornu's
spiral.

20.1 INTRODUCTION To appreciate the implications of this principle, we con-

sider the incidence of a plane wave on a circular hole of
In Chap. 18 we had mentioned that the phenomenon of radius a as shown in Fig. 20.1. In Sec. 18.3 we showed that
diffraction can be broadly classified under two categories: the beam will undergo diffraction divergence and the angular
Under the first category comes the Fresnel class of diffrac- spreading will be given by
tion in which either the source or the screen (or both) is at a
finite distance from the diffracting aperture. In the second l
Dq ~
category comes the Fraunhofer class of diffraction (dis- 2a
cussed in the last two chapters) in which the wave incident
Thus, when a >>> l, the intensity at a point R (which is deep
on the aperture is a plane wave and the diffraction pattern is
inside the geometrical shadow) will be negligible; on the
observed on the focal plane of a convex lens, so that the
other hand, if a ~ l, there will be almost uniform spreading
screen is effectively at an infinite distance from the aperture.
out of the beam, resulting in an (almost) uniform illumination
In this chapter. we will discuss the Fresnel class of diffrac-
of the screen. This phenomenon is a manifestation of the fact
tion and also study the transition to the Fraunhofer region.
that when a >>> l, the secondary wavelets emanating from dif-
The underlying principle in the entire analysis is the
ferent points on the circular aperture so beautifully interfere to
Huygens–Fresnel principle according to which
produce (almost) zero intensity in the geometrical shadow and a
Each point on a wave front is a source of secondary large intensity inside the circular region (see Fig. 20.1). However,
disturbance, and the secondary wavelets emanating if a ~ l, then the aperture almost acts as a point source, result-
from different points mutually interfere. ing in a uniform illumination of the screen (see Fig. 12.3).

1
The author found this quotation in Ref. 1.

gha80482_ch20_303-324.PMD 303 2/2/2009, 1:06 PM

 304 Optics
u
S of the nth circle will obviously be given by
1/ 2
Edge of the ÈÊ lˆ
2 ˘
R geometrical rn = Í Á d + n ˜ - d 2 ˙
shadow Ë 2¯
ÎÍ ˚˙
P 1/ 2
Ê nl ˆ
= nld Á1 +
Ë 4 d ˜¯
a d
O S¢ or
rn ª nld (1)

where we have assumed d >>> l; this is indeed justified for

practical systems using visible light. Of course, we are assum-
Fig. 20.1 Diffraction of a plane wave incident normally on ing that n is not a very large number. The annular region
a circular aperture of radius a. between the nth circle and (n – 1)st circle is known as the nth
half-period zone; the area of the nth half-period zone is given by
We will first introduce the concept of Fresnel half-period An = p r2n – p r2n–1
zones to have a qualitative understanding of the Fresnel dif-
fraction pattern; this will be followed by a more rigorous ª p [nld – (n – 1)ld] = pld (2)
analysis of the Fresnel class of diffraction and its transition Thus the areas of all the half-period zones are approxi-
to the Fraunhofer region. mately equal. Now the resultant disturbance produced by
the nth zone will be p out of phase with the disturbance
produced by the (n – 1)st [or the (n + 1)st] zone. This can
be easily seen from the following consideration: For infini-
20.2 FRESNEL HALF-PERIOD tesimal area surrounding a point Qn in the nth half-period
ZONES zone, there is a corresponding infinitesimal area surround-
ing point Qn – 1 in the (n – 1)st half-period zone such that
Let us consider a plane wave front WW¢ propagating in the
z direction as shown in Fig. 20.2. To determine the field at an l
QnP – Qn–1P =
arbitrary point P due to the disturbances reaching from dif- 2
ferent portions of the wave front, we make the following which corresponds to a phase difference of p. Since the ar-
construction: From point P we drop a perpendicular PO on eas of the zones are approximately equal, one can have a
the wave front. If PO = d, then with point P as center we draw one-to-one correspondence between points in various zones.
spheres of radii d + l/2, d + 2l/2, d + 3l/2, . . ., these spheres Thus, the resultant amplitude at point P can be written as
will intersect WW¢ in circles as shown in Fig. 20.2. The radius u(P) = u1 – u2 + u3 – u4 + . . . + (–1)m+1um + . . . (3)
where un represents the net amplitude produced by the
secondary wavelets emanating from the nth zone; the alter-
Qn nate negative and positive signs represent the fact that the
W
resultant disturbances produced by two consecutive zones
are p out of phase with respect to each other. The amplitude
Qn – 1
produced by a particular zone is proportional to the area of
the zone and inversely proportional to the distance of the
zone from point P; further, it also depends on an obliquity
O
d+
nl factor which is proportional to 21 (1 + cos c), where c is
2 the angle that the normal to the zone makes with line
d QP; this obliquity factor comes out automatically from rigor-
ous diffraction theory.2 Thus we may write
P
W¢ n th half-period zone An 1 + cos c
u n = constant (4)
Fig. 20.2 Construction of Fresnel half-period zones. Qn P 2

2
See, e.g., Ref. 2.

gha80482_ch20_303-324.PMD 304 2/2/2009, 1:06 PM

 Fresnel Diffraction 305
u
where An represents the area of the nth zone. It can be shown Using Eqs. (8) and (10), we may approximately write
that if we use the exact expression for rn, the area of the zones
increases with n; however, this slight increase in area is u1 um
u ( P) ª + m odd
exactly compensated by the increased distance of the zone 2 2 (11)
from point P. In spite of this, the amplitudes u1, u2, u3 . . . u u
u ( P ) ª 1 - m m even
decrease monotonically because of increased obliquity. Thus 2 2
we may write
If we can neglect um in comparison to u1, then the Eq. (11)3
u 1 > u2 > u3 > . . . (5) gives the remarkable result that
The series expressed by Eq. (3) can be approximately summed u1
due to a method by Schuster. We rewrite Eq. (3) as u(P) ª (12)
2

u1 È u1 u ˘ È u3 u5 ˘ implying that the resultant amplitude produced by the en-

u(P) = + Í - u2 + 3 ˙ + Í - u4 + ˙ + . . . (6) tire wave front is only one-half of the amplitude produced
2 Î2 2˚ Î2 2˚
by the first half-period zone.
where the last term is either 21 um or 12 um-1 - um according
to whether m is odd or even. If the obliquity factor is such
that 20.2.1 Diffraction by a Circular Aperture
un > 1 (un-1 + un+1 ) (7) We may use the above analysis to study the diffraction of a
2
plane wave by a circular aperture. Let point P be at a dis-
then the quantities inside the brackets in Eq. (6) will be nega- tance d from the circular aperture (see Fig. 20.1). We assume
tive; consequently, that the radius of the circular aperture a can be increased
from zero onward. As a increases, the intensity at point P will
1 1
u ( P) < u1 + um (m odd) also increase until the circular aperture contains the first half-
2 2
(8) period zone; this happens when a = l d . The resultant
1 1 u u
u ( P) < u1 + um-1 - um ª 1 - m (m even) amplitude at point P is u1 which is twice the value of the am-
2 2 2 2
plitude for the unobstructed wave front [see Eq. (12)]. The
where we have assumed that the amplitudes of the fields pro- intensity is therefore 4I0, where I0 represents the intensity at
duced by consecutive zones differ only slightly. To obtain point P due to the unobstructed wave front. If we further
the upper limits, we rewrite Eq. (3) in the form increase a, then u(P) will start decreasing and when the cir-
cular aperture contains the first two half-period zones (which
u2 È u2 u ˘ Èu u ˘ happens when a = 2 l d ), the resultant amplitude (= u1 – u2)
u(P) = u1 – - - u3 + 4 ˙ – Í 4 - u5 + 6 ˙ (9)
2 ÍÎ 2 2˚ Î2 2˚ is almost zero. Thus, by increasing the hole diameter,
the intensity at point P decreases almost to zero. This inter-
where the last term is now - 12 um-1 + um , when m is odd, and
esting result is once again due to the validity of the
– 12 um; when m is even. Since the quantities inside the brack- Huygens–Fresnel principle and hence would be valid for
ets are negative, we obtain sound waves also. We may generalize the above result by
noting that if
u2 um-1 u u
u ( P) > u1 - - + um ª 1 + m
2 2 2 2 a= (2n + 1)l d n = 0, 1, 2, . . . (maxima)
m odd
(10) the aperture will contain an odd number of half-period zones
u u u u and the intensity will be maximum; on the other hand, if
u ( P) > u1 - 2 - m ª 1 - m
2 2 2 2
m even a= 2nl d n = 1, 2, . . . (minima)

3
If one assumes a form of the obliquity factor as given by Eq. (4), then it decreases from 1 to 12 as m increases from 1 to •; this
implies that |um| can never be smaller than u1 / 2 . However, when m is large, a slight shift of point P on the axis will change the
amplitude from u1/2 + um /2 to u1/2 - um /2 ; the changes will occur with such great rapidity that one can only observe the average
value which will be u1/2 .

gha80482_ch20_303-324.PMD 305 2/2/2009, 1:06 PM

 306 Optics
u
the aperture will contain an even number of half-period zones zones, then the field at point P is
and the intensity will be minimum. To have a numerical ap- u(P) = up+1 – up+2 + . . .
preciation, we note that for d = 50 cm and l = 5 ¥ 10–5 cm,
u p +1
the radii of the first, second, and third zones are 0.500, 0.707, ª (13)
and 0.866 mm, respectively. As a corollary of the above 2
analysis, we can consider a circular aperture of a fixed radius Thus, we should always obtain a bright spot on the axis be-
a and study the intensity variation along the axis. Whenever hind a circular disc (the more rigorous theory also predicts
the distance the same result—see Sec. 20.4.2). This is called the Poisson
spot. In 1816 the French physicist Augustin Fresnel devel-
a2 oped the mathematical theory of diffraction using the wave
d = n = 0, 1, 2, . . . (maxima)
(2n + 1) l theory of light. Simeon Poisson, the famous mathematician,
point P (see Fig. 20.1) will correspond to a maximum. Simi- used Fresnel's theory to predict a bright spot at the center of
larly, when the shadow of an opaque disc. Poisson was a great sup-
porter of the corpuscular theory of light, and he said that
a2 since the bright spot is against common sense, the wave
d = n = 1, 2, . . . (minima)
2nl theory must be wrong. Shortly afterward, Fresnel and Arago
point P will correspond to a minimum. The intensity distribu- carried out the experiment to demonstrate the existence of the
tion on screen SS¢ at off-axis points can be approximately Poisson spot [see Fig. 20.3(b)], validating the wave theory.
calculated by using the half-period zones, but such a
calculation is fairly cumbersome. However, from the symme-
try of the problem, one can deduce that the diffraction
20.3 THE ZONE PLATE
pattern has to be in the form of concentric circular rings with A beautiful application of the concept of Fresnel half-period
their centers at point P. zones lies in the construction of the zone plate which consists of
a large number of concentric circles whose radii are proportional
to the square root of natural numbers and the alternate annular
20.2.2 Diffraction by an Opaque regions of which are blackened (see Fig. 20.4). Let the radii of the
Disc—The Poisson Spot circles be 1 K, 2 K, 3 K, 4 K, . . . where K is a constant
If instead of the circular aperture we have a circular disc [see and has the dimension of length. We consider a point P1 which is
Fig. 20.3(a)] and if the disc obstructs the first p half-period at a distance K2/l from the zone plate; for this point the blackened

Opaque disc
(a) (b)

Fig. 20.3 (a) When a plane wave is incident normally on an opaque disc, a bright spot is always formed
on an axial point. This spot is known as the Poisson spot. (b) The Poisson spot at the center of
the shadow of a penny; the screen is 20 m from the coin, and the source of light is also 20 m
from the coin [Photograph reprinted with permission from P. M. Rinard, “Large Scale Diffrac-
tion Patterns from Circular Objects,” American Journal of Physics, Vol. 44, p. 70, 1976; Copyright
1976, American Association of Physics Teachers].

gha80482_ch20_303-324.PMD 306 2/2/2009, 1:06 PM

 Fresnel Diffraction 307
u
Thus OP = 50 cm. Similarly the darkest point would be at a distance

(0.05)2
= 25 cm
2 ¥ 5 ¥ 10-5

Example 20.2 Consider a zone plate with radii

rn = 0.1 n cm
For l = 5 ¥ 10–5 cm, we will calculate the positions of various foci.
The most intense focal point will be at a distance

r12 0.01
= = 200 cm
l 5 × 10 −5
The other focal points will be at distances of 200/3, 200/5,and
200/7 cm, etc. Between any two consecutive foci there will be dark
points on the axis corresponding to which the first circle will con-
Fig. 20.4 The zone plate. tain an even number of half-period zones.
The zone plate can also be used for imaging points on the axis;
e.g., if we have a point source at S, then a bright image will be
rings correspond to the 2nd, 4th, 6th, . . . half-period zones. Thus,
formed at P, where point P should be such that [see Fig. 20.5(b)]
the even zones are obstructed and the resultant amplitude at P1
[see Fig. 20.5(a)] will be l
SL + LP – SP = (18)
2
u1 + u 3 + u5 + . . . (14)
the point L being on the periphery of the first circle of the
producing an intense maximum. For point P3 (which is at a zone plate [see Fig. 20.5(b)]. If the radius of the first circle is
distance K2/3l) the first blackened ring contains the 4th, 5th, r1, then
6th zones, the second blackened ring contains the 10th, 11th
SL + LP – SP = a 2 + r12 + b 2 + r12 - (a + b)
and 12th zones, etc.; thus the resultant amplitude is
(u1 – u2 + u3) + (u7 – u8 + u9) + . . . (15) Ê r12 ˆ Ê r12 ˆ
ª a Á 1 + 2 ˜ + b Á 1 + 2 ˜ - ( a + b)
which would again correspond to a maximum, but it would not be Ë 2a ¯ Ë 2b ¯
as intense as point P1. Between points P1 and P3 there will be a
r12 Ê 1 1 ˆ
point P2 (at a distance K2/2l) where the resultant amplitude is ª Á + ˜ (19)
2 Ë a b¯
(u1 – u2 ) + (u5 – u6 ) + . . . (16)
implying that corresponding to P2 the first blackened ring
contains the 3rd and 4th half-period zones, etc. Obviously,
P5 P4 P3 P2 P1
point P2 will correspond to a minimum. Thus, if a plane wave
is incident normally on a zone plate, then the corresponding
focal points are at distances (a)

2 2 2
K K K
, , ,... (17)
l 3l 5l L
from the zone plate. Elementary calculations will show that r1
the zone plate suffers from considerable chromatic aberra-
tions (see Prob. 20.5). S P

Example 20.1 Assume a plane wave (l = 5 ¥ 10–5 cm) to

be incident on a circular aperture of radius 0.5 mm. We will calcu- (b)
late the positions of the brightest and darkest points on the axis.
Fig. 20.5 (a) For a plane wave incident on a zone plate, the
For the brightest point, the aperture should contain only the first
maximum intensity occurs at points P1, P3, etc.
zone, and thus we must have (see Fig. 20.1) The minima occur at P2, P4, . . .. (b) Imaging of a
(0.05)2 = OP(5 ¥ 10–5) point object by a zone plate.

gha80482_ch20_303-324.PMD 307 2/2/2009, 1:06 PM

 308 Optics
u
Thus Eq. (18) becomes Further, in the Fresnel approximation [see Eq. (9) of Chap.19]
the above integral takes the form

zz
1 1 1
+ = (20)
a b f 1 ikz
u(x, y, z) ª e A(x, h)
where f = r12/l represents the focal length. Equation (20) re- il z
sembles the lens law. A very interesting demonstration
experiment of the zone plate can be carried out by using mi- Ï ik ¸
¥ exp Ì [ x - x)2 + ( y - h) 2]˝ dx dh (23)
crowave sources (l ~ 1 cm) and, instead of the dark rings, Ó2z ˛
having aluminum rings on a perspex sheet of dimension
~40 cm ¥ 40 cm.
20.4.1 Diffraction of a Plane Wave
20.4 FRESNEL DIFFRACTION—A Incident Normally on a Circular
MORE RIGOROUS Aperture
APPROACH We assume a plane wave incident normally on a circular ap-
erture of radius a as shown in Fig. 20.7. The z axis is normal
In Sec. 19.2 we gave a more rigorous analysis of the diffrac- to the plane of the aperture, and screen SS¢ is assumed to be
tion of a plane wave by different types of aperture. We normal to the z axis. It is obvious from the symmetry of the
considered a plane wave (of amplitude A) incident normally problem that we will obtain circular fringes on screen SS¢ ;
on an aperture as shown in Fig. 20.6. Using the Huygens– however, it is very difficult to calculate the actual intensity
Fresnel principle, we showed that the field produced at point variation on the screen. Therefore, for the sake of mathemati-
P on screen SS¢ (which is at a distance d from the aperture) cal simplicity, we will calculate the variation of intensity only
is given by along the z axis. Obviously, it will be more convenient to use the
circular system of coordinates. In this system, the coordinates
A eikr
i l ÚÚ r
u(P) = dx dh (21) of an arbitrary point M on the aperture will be (r, f), where r
is the distance the point M from the center O and f is the
where the integration is over the area of the aperture. Now, if angle that OM makes with the x axis (see Fig. 20.7), and a
the amplitude and phase distribution on the plane z = 0 is small element area dS surrounding point M will be r dr df.
given by A(x, h), then the above integral is modified to Thus, using Eq. (21) we get
2p a
1 eikr iA eikr
u(P) =
il ÚÚ A(x, h)
r
dx dh (22) u(P) ª -
l Ú Ú r r dr d f (24)
0 0

y S

h
S

P P
h
r
M
r rj
x O
S¢
Q
d xd h M x
S¢
d
x
d

Fig. 20.7 Diffraction of a plane wave incident normally on

a circular aperture of radius a; point Q is an arbi-
Fig. 20.6 A plane wave incident normally on an aperture. trary point on the periphery of the aperture.

gha80482_ch20_303-324.PMD 308 2/2/2009, 1:06 PM