Sie sind auf Seite 1von 10

PHYS2023 Wave Physics 31/10/2007

7 Fraunhofer diffraction
7.1 More wave propagation around obstructions
eynman, in the introduction to chapter 30 of his Lectures on Physics,
remarks that no one has ever been able to define the difference
between interference and diffraction satisfactorily. This chapter is
therefore a continuation of chapter 6, inasmuch as the interference fringes
observed by Young are the diffraction pattern of a double slit. On the other
hand, few opticians would regard the interference occurring in the Michelson
interferometer as a form of diffraction. It is perhaps fair to regard diffraction
as interference that accompanies division of the wavefront; Feynmans
observation is therefore based on the difficulty we sometimes have in
distinguishing between division of amplitude and division of wavefront.
Fortunately, perhaps, diffraction itself is generally further divided into
Fraunhofer and Fresnel diffraction, and the Fraunhofer diffraction with
which we shall generally be concerned is that which occurs under particular
conditions in which the confusion between division of amplitude and division
of wavefront is almost eliminated. Fraunhofer diffraction, according to the
most formal definition, is that which is observed in the image plane of the
source, and we have already noted that all optical paths from the source to its
image are equal in length. Fraunhofer diffraction, therefore, is entirely (or at
least, almost entirely) concerned with the transverse disruption, i.e. division,
of the wavefront. It tends to be associated with plane waves, small angles and
distant images, although these are not essential. We shall examine further the
definition of Fraunhofer diffraction in section 7.7.
Fresnel diffraction may be regarded as any that cannot be considered of the
Fraunhofer type, and typically includes that observed close to the diffracting
objects, near to the source, and so on. These are the conditions in which the
distinction between diffraction and just plain interference is less clear. The
only problem, however, is that of nomenclature, and the physics is simply that
waves arriving via different routes will exhibit the effects of interference.
The algebra of Fresnel diffraction, however, is generally rather messy so
in this chapter, we shall in any case restrict our discussion to diffraction of the
Fraunhofer type.
7.2 Diffraction by a single slit
We have already examined one of the classic diffraction arrangements in our
treatment of the double slit in section 6.2. For our second example, we shall
consider diffraction by a single slit alone. This is conceptually slightly more
complicated for, while each slit of Youngs pair could be considered to be
infinitely narrow, we must in the single case allow the slit some width.





H. J. Pain, The Physics of Vibrations
and Waves, pp. 366-407
A. P. French, Vibrations and Waves,
pp. 284-298
I. G. Main, Vibrations and Waves in
Physics, pp. 320-339
E. Hecht, Optics, chapter 10
R. P. Feynman, Lectures in Physics,
vol 1, chapter 30
31/10/2007 PHYS2023 Wave Physics
52 Fraunhofer diffraction

Indeed, we shall model our single, broad slit as if it were formed from a line
of narrow slits.
We shall adopt a geometry characteristic of Fraunhofer diffraction, in
which a flat mask is illuminated with plane waves and observed after the
diffracted waves have travelled a distance that is large in comparison with the
diffracting object. Paths from the object to each point on the observation
screen will therefore be approximately parallel, and we shall describe the
diffraction pattern through its dependence upon the diffracted angle , as in
equation (6.7).
As with Youngs double slit, our analysis uses Huygens method, and we
simply sum the contributions from secondary sources that are imagined to lie
in the plane of the mask, as shown in figure 7.1. The slit, of width d, is
illuminated at right angles with plane waves of wavelength , and the
diffracted light is observed at an angle . Taking as our reference the path via
a point at the centre of the slit, other paths are longer or shorter by a distance
s(y) = - y sin. The phasor for such a path is therefore represented by the
complex quantity
( ) ( ) sin exp
0
iky a y a = (7.1)
where a
0
represents the source amplitude, which is the same for all secondary
sources within the slit. The diffracted amplitude is now found by adding the
contributions from all such sources,

( ) ( )
( )
( )
( )


sin
2
sin
sin
sin
2
sin
sin exp
d sin exp
0
0
2
2
0
2
2
0
d a
kd
k
a
ik
iky a
y iky a a
d
d
d
d
=
|

\
|
=
(

=
=

(7.2)
where we have written () (kd/2)sin. The function sin/ is known as
the sinc function, and is plotted in figure 7.2. This, then, with appropriate
scaling, is the amplitude of the diffraction pattern from the single slit; the
intensity is simply its square.
A circular aperture may be considered the 2-D equivalent of the 1-D slit,
and its diffraction patterns is an Airy pattern, given by
( )
( )
2
1
(

=
r
r J
r I (7.3)
where J
1
(r) is a Bessel function of the first kind. Although qualitatively
resembling a sinc function of radius, the nodes of the Airy pattern lie at
slightly different positions.

















Fig. 7.1 Diffraction from a single slit



-15 -10 -5 5 10 15
-0.2
0.2
0.4
0.6
0.8
1

Fig. 7.2 The sinc function, sin/


Fig. 7.3 The diffraction pattern from a
circular aperture
y

s
d
PHYS2023 Wave Physics 31/10/2007
Fraunhofer diffraction 53
7.3 The Cornu spiral
The integral of equation (7.2) may of course be performed, at least crudely, in
a graphical fashion by summing phasors as in section 5.4. Figure 7.4 shows
the result for an uninterrupted wavefront in one dimension, and the phasor
corresponding to the integral is that shown in grey joining the two ends. (The
argument of the resultant here is 45 half of that mentioned in section 5.2.1
for an uninterrupted plane because it accounts for only a single dimension.)
The curve is known as a Cornu spiral.
When the line of secondary sources is curtailed by the presence of a slit,
the integral is also truncated, and the resultant phasor joins the ends of the
corresponding chain of component phasors, as shown dashed in figure 7.4.
The Fraunhofer condition is that the aperture be small enough that all the
phasors lie close to the origin where they are parallel.
Figure 7.4 of course shows the phasor construction for the intensity of the
diffraction pattern at = 0; the interesting features of the diffraction pattern
are those for other angles, and the phasor method does not truly offer a
practical method for their calculation. We can imagine approximately what
will happen, however, for adjacent phasors will differ not only by the phase
angle resulting from their different path differences but also by an additional
contribution due to the inclination of the path to the axis. This has the effect
of tightening or loosening the spiral. As we vary , the degree of tightening
or loosening varies, and the magnitude of the resultant shrinks and grows,
following the sinc function of figure 7.2.
An important feature is that, as we increase the width of the slit and
thereby approach an uninterrupted wavefront, the line of parallel phasors
grows in length, and even the slightest curvature causes it to wrap back on
itself and produce a resultant whose magnitude approaches zero. For an
undiffracted beam, in other words, the amplitude and intensity of the
diffracted wave are zero except for = 0. This has an important consequence
in Babinets principle, as we shall see next.
7.4 Babinets principle
An unobstructed plane wave will continue to propagate as a plane wave and,
if initially its wavefronts are parallel to a given reference plane (such as
x = 0), then the wavefronts will remain parallel, and the direction of
propagation will be along the x-axis. The intensity distribution observed as a
function of angle will therefore be a -function, which is zero for all angles
except = 0.
Suppose, then, that we have determined the diffraction pattern for a
circular aperture, obtaining the distribution illustrated in figure 7.3. What
would be the pattern produced by light diffracting around a circular disc of
the same diameter as this aperture? The answer is that, except at = 0, it will
be exactly the same, for the secondary sources for the aperture, together with
those outside the circular disc, combine to form exactly the set of sources that
correspond to an unobstructed plane wave. Since, for 0, the combined
amplitude must be zero, the amplitudes for the disc and aperture must be
equal and opposite; the diffracted intensities will therefore be identical.













Fig. 7.4 Phasors from sources across an
uninterrupted wavefront form a Cornu
spiral. The resultant phasor joins the
ends of the spiral
















Fig. 7.5 Phasors corresponding to the
diffracted amplitude for a single slit,
shown for several angles








E. Hecht, Optics, section 10.3.11
31/10/2007 PHYS2023 Wave Physics
54 Fraunhofer diffraction

Babinets principle that the Fraunhofer diffraction patterns of
complementary masks are identical proves to be of great utility in the
calculation and interpretation of diffraction patterns. It also offers helpful
insight into the phenomenon of Poissons bright spot, whose observation was
another key indication of the wave nature of light.
7.4.1 Poissons bright spot
Simon-Denis Poisson, reviewing Fresnels work for the French Acadmie
des Sciences, realized that an implication of the diffraction theory would be a
bright spot in the centre of the shadow of a small disc. The Acadmie
committee was sceptical, generally favouring the corpuscular nature of light,
but had the scientific integrity to seek experimental investigation, which was
carried out under Franois Arago. The bright spot was indeed observed, and
Fresnel was subsequently awarded the Grand Prix of the Acadmie.
7.5 The diffraction grating
Huygens theory of wave propagation and its refinements by Huygens,
Kirchoff, Fresnel and Fraunhofer has already allowed us to explain
experimentally observed diffraction patterns. As is usual in the development
of a theory, the first efforts are to confirm the principles and investigate its
mechanisms; then, once confidence and understanding have been achieved, it
becomes a tool for the design of useful applications. One of the most
significant designs to apply the theory of wave interference is the diffraction
grating a regular series of narrow slits that may be regarded as an extension
of Youngs double slit, but whose practical utility is far greater.
The diffraction grating is shown in figure 7.6. A series of slits, each of
width w and with centres separated by d, fill an aperture of width a. Light of
wavelength is incident at right angles on the array, and diffracted light is
observed a large distance away at an angle to the incident direction.
Proceeding exactly as in section 7.2, we may write the diffracted amplitude as
( ) ( ) ( )

=
2
2
0
d sin exp
a
a
y iky y a a (7.4)
where a
0
(y) is the transmission (here always either 0 or 1) of the grating at a
distance y from the axis. Explicitly, this hence becomes

( ) ( )
( )
( )
K +
+
+
=

2
2
0
2
2
0
2
2
0
3
3
2
2
1
1
d sin exp
d sin exp
d sin exp
w y
w y
w y
w y
w y
w y
y iky a
y iky a
y iky a a


(7.5)
where y
1
, y
2
, y
3
are the coordinates of the centres of the slits, taken in turn
across the whole aperture. This expression may be simplified by realizing that
each integral is identical but for a phase factor exp(-iky
n
sin):






E. Hecht, Optics, pp. 494-5
http://www-groups.dcs.st-
and.ac.uk/~history/Mathematicians/Fresnel.
html




















Fig. 7.6 The diffraction grating
y

s
a
d
w
PHYS2023 Wave Physics 31/10/2007
Fraunhofer diffraction 55

( ) ( )
( )
( )
K +
+
+
=

2
2
0
sin
2
2
0
sin
2
2
0
sin
d sin exp
d sin exp
d sin exp
3
2
1
w
w
iky
w
w
iky
w
w
iky
y iky a e
y iky a e
y iky a e a

(7.6)
This may be written as the product of an integral and a geometric series
( ) ( )


)
`

=
=

2
2
0
2
2
sin
d sin exp
w
w
N
N n
iknd
y iky a e a

(7.7)
where we have substituted
nd y
n
(7.8)
with n = -N/2 N/2, and Nd = a. The geometric progression is evaluated in
the usual fashion, giving

( ) ( )
( ) ( )
( )
2
sin
sin
2
sin 1
sin
1
2 sin 2 sin
2 sin 1 2 sin 1
sin
sin 2 sin 1 2 2
2
sin

kd
d N k
e e
e e
e
e e
e
ikd ikd
d N ik d N ik
ikd
d N ik d N ik N
N n
iknd
+
=

+ +

+
=

(7.9)
If we count the number of slits from -N/2 to N/2, we find that there are in fact
N = N+1, as there is one with the number n = 0. Equation (7.9) may therefore
be written more helpfully as

2
sin
sin
2
sin
sin
2
2
sin

kd
d N k
e
N
N n
iknd

(7.10)
and, using the result for the integral calculated in section 7.2, the diffracted
amplitude is therefore
( )
2
sin
2
sin
sin
2
sin
sin
2
sin
sin
0

kw
kw
d a
kd
d N k
a

= (7.11)





E. Hecht, Optics, pp. 476-481
H. J. Pain, The Physics of Vibrations
and Waves, pp. 373-376
I. G. Main, Vibrations and Waves in
Physics, pp. 324-7
R. P. Feynman, Lectures on Physics,
vol. 1, pp. 30-3 30-6
31/10/2007 PHYS2023 Wave Physics
56 Fraunhofer diffraction

The diffracted amplitude is therefore that for a single slit of width w,
multiplied by the function of equation (7.10), which turns out to be a series of
narrow peaks occurring every time that the denominator approaches zero
i.e., when

m
kd
=
2
sin
(7.12)
If we write k in terms of the wavelength , we hence obtain
m d = sin (7.13)
and the integer m serves as a label for each diffraction order.
We shall see in chapter 12 that the shape of the peak corresponding to each
order is simply the diffraction pattern for a single slit of width a the overall
width of the diffraction grating. The diffracted intensity distribution for a
typical diffraction grating is shown in figure 7.7.
The utility of the diffraction grating rests in its ability to direct different
wavelengths to different angular positions, as we may see by rearranging
equation (7.13):

d
m
= sin (7.14)
A beam of light containing several wavelengths is therefore dispersed so
that an appropriately placed photographic film or detector will record its
spectrum. Usually, a lens is used to focus the parallel rays that leave the
grating at a given angle onto a particular point on the film or detector. In this
arrangement, the diffraction is recorded at the position where an image of the
incident beam of light (for example, from a distant point source) would have
been focussed. We therefore record the diffracted wave in the image plane of
the source which we mentioned earlier is the criterion for the observation of
Fraunhofer diffraction.
7.5.1 Phased array radar
A classic example of the application of the Huygens principle of wave
propagation is in the giant phased array radar that has replaced the golf-
balls near Fylingdales moor in North Yorkshire. The three faces of this 35 m
high pyramid each contain an array of 2,560 radar transducers that fill a
roughly circular area some 25 m in diameter. The beam produced is steered,
not by moving this vast antenna, but by introducing electronically an
appropriate delay between the different transducers. Each array can thereby
be steered though 120 of azimuth and 85 of elevation.
It is instructive to check these figures. Each transducer accounts for 0.2 m2
of the antenna face, giving a typical separation between transducers of 45 cm.
Frequencies of 420-450 MHz correspond to wavelengths of 0.690.02 m. If
the transducers are driven in phase, then equation (7.14) gives

45 . 0
69 . 0
sin
m
= (7.15)









-1.5 -1 -0.5 0.5 1 1.5
0.2
0.4
0.6
0.8
1

Fig. 7.7 Fraunhofer diffraction intensity
from a diffraction grating








Fig. 7.8 Phased-array radar for early
warning of ballistic missiles. Each of
the 2,560 transducers radiates pulses of
0.25-16 msec at a power of 340 W in
the range 420-450 MHz. (Raytheon)
m=1
m=2
m=0
m=-1
m=-2

PHYS2023 Wave Physics 31/10/2007
Fraunhofer diffraction 57
and we see that only m = 0 gives a valid solution: the antenna will produce a
single beam.
If a diffraction grating is illuminated at an angle to the normal, equation
(7.13) is modified to become
( ) m d = sin sin (7.16)
where, clearly, the zeroth-order beam (m = 0) emerges at = . Inserting an
appropriate phase shift between transducers of the radar array has the same
effect, so we can use this expression to establish how far the main beam may
be deflected before higher diffraction orders are produced. The critical angle
will be given by setting m = 1, = -90, giving
53 . 0 90 sin
45 . 0
69 . 0
sin = =
o
(7.17)
and hence = 32: somewhat less, if our estimates are valid, than the 60
required for the full field of view, suggesting that echoes from the m = -1
beam will have to be taken into account when the beam is steered through
more than 32 from the axis; this signal will, however, be visible to the array
on one of the other two faces, and may thus be distinguished.
We may also estimate the angular resolution of the radar, for the main
beam will have a width determined by its single slit (or, strictly, single
circular aperture) diffraction pattern. For a slit width of 25 m and a
wavelength of 0.69 m, equation (7.2) would predict the first zero in the
diffraction pattern to be when
= = sin
2
kd
(7.18)
giving sin = /d, and hence = 1.6. At the specified range of 5,000 km,
this corresponds to an area of a few 10
10
m
2
; even a large missile would
therefore intercept and reflect only a few mW of the 800 kW radiated. Note
that, although we have here assumed a rectangular aperture, a more careful
calculation for a circular aperture would yield roughly similar results.
7.6 Wavefront reconstruction and holography
An instructive perspective on the diffraction grating is that it transmits the
incident illumination only in those regions where the phase is also correct for
a diffracted order. This insight, recognized by Dennis Gabor many years
before the invention of the laser, led him to propose that any propagating
wave could be reconstructed using a given source by using a mask that
transmitted only those parts of the incident wavefront that matched the phase
of the original wave. If the recorded wave were light that had been scattered
by an object, then the reconstructed wave would form a virtual image of the
object, even though the object itself had been since removed. Further, the
virtual image would change with viewing angle, exactly as for the original
three-dimensional object. This is the principle behind holography.










D. Gabor, Nature, 161 (4098) 777-8
(15 May 1948);
Nobel lecture (1971):
http://nobelprize.org/physics/laureates/1971
/gabor-lecture.pdf
31/10/2007 PHYS2023 Wave Physics
58 Fraunhofer diffraction

Production of the required mask (or hologram), which at first sight would
appear a significant challenge, can be achieved simply by recording the
interference pattern between the original propagating wave and the wavefront
that will be used for reconstruction. With reference to figure 7.6, it is
apparent that, were the diffracted beam already to exist, then it would
interfere with the plane wave illumination to produce fringes that coincided
exactly with the slits of the grating. A photographic plate placed at the
position of the grating would record these fringes, and could subsequently be
developed to convert them into transmissive regions.
It may be intuitively helpful, although mathematically challenging, to
regard any scattered light as having approximately plane wavefronts over a
sufficiently small region, within which interference with the plane reference
wave to be used for reconstruction will produce regular fringes exactly as
described above. When recorded and illuminated with the reference beam,
this section of the hologram will function just as a diffraction grating,
producing a diffracted ray that retraces the path of the original, hence
reconstructing the image.
Holograms that function as described so far, by modulating the transmitted
amplitude of the reference beam, are known as amplitude holograms and,
provided that the apparatus can be held stable to within fractions of a
wavelength during the recording process, are relatively straightforward to
produce. A variation on this principle is the phase hologram, whereby the
entire reference wave is transmitted but its phase is modulated, by the varying
thickness of the hologram, so that it at all points has the phase of the original
object wave. A photographically-produced amplitude hologram can be
converted to a phase hologram by appropriate bleaching during the
development process.
Because amplitude holograms reproduce only certain points of the original
wave, it is clear that some information regarding the original wave must have
been lost. This turns out to be apparent in the production of a second, real
image that, in the diffraction grating analogy, corresponds to the negative
diffraction order. The phase hologram, by reconstructing the entire
wavefront, does not suffer from this (although the real image can, of course,
be useful in its own right).
It will be apparent from the above explanation that holograms may only be
recorded and replayed using monochromatic light: with a range of
wavefronts, multiple fringe patterns would overlap and destroy the recording
process, while different reconstructing wavelengths will produce images of
varying sizes. Information theory again explains this through the reduction of
three-dimensional information into a two-dimensional hologram. The solution
is therefore to record the hologram within a volume of photographic
emulsion: such volume holograms may indeed be viewed in white light.
7.6.1 Phased arrays at visible wavelengths: computer-created holograms
Centimetre-sized liquid crystal displays, when illuminated with plane waves
from a laser, have the effect of imposing a controllable phase shift at each of
the million or so pixel positions. As with the phase array radar, these act as
Huygens secondary sources and, with suitable programming, can produce








Fig. 7.9 Image produced by diffraction
from a spatial light modulator. Note the
zeroth-order beam and negative order
image to the bottom-right.(Holoeye)
PHYS2023 Wave Physics 31/10/2007
Fraunhofer diffraction 59
diffraction patterns of considerable complexity. This is an example of
computer-generated holography.
An example of diffraction from a computer-controlled pixel array,
demonstrating the wavelength dependence of the diffraction, is shown in
figure 7.9. Spatial light modulators such as this are used both for holographic
image formation and for the production of dynamically-adjustable light
beams for the remote, contact-free manipulation of biological cells and the
like, using a technique known as optical tweezers, based around a force that is
mediated entirely optically.
7.7 Definition of Fraunhofer diffraction
We have seen that the calculation of diffraction patterns is relatively
straightforward when a flat mask is illuminated with plane wavefronts and the
diffraction is observed as a function of angle. This is an example of the
Fraunhofer diffraction condition, which merits a little further examination.
It is instructive to consider how our classic diffraction arrangement could
be achieved in a finite space and using a point source. As shown in figure
7.10, the addition of two lenses is required. The first, placed a focal length
away from the source, produces a parallel beam of plane wavefronts from the
spherical waves emanating from the source; the second, placed a focal length
from a screen (or photographic film or photodetector) images parallel rays of
diffracted light onto individual points.
Now, if we were to omit the diffracting mask from this arrangement, we
should simply have a pair of lenses placed so as to form on the screen an
image of objects in the plane through S, and indeed the point source at S
would form a point image on the screen at O. This is what we mean when we
say that the screen is in the image plane of the source.
The presence of the diffracting mask causes some light to appear at points
other than O on the screen: we shall see the Fraunhofer diffraction pattern of
the mask. Since diffraction causes light, which would otherwise have been
focused on a single point, to be spread elsewhere, the diffraction pattern is
referred to by lens designers as the point spread function. (For real lenses,
aberrations may also contribute to the point spread function.)
It is clear that moving the diffracting mask to the left and right between the
lenses will have no effect upon the diffraction pattern (as a function of angle,
at least); and a little thought will show that, provided it remains within the
beam, neither will movement up and down. It turns out that, remarkably, the
pattern is the same wherever the mask is placed even if it is outside the pair
of lenses provided that the paraxial approximation remains valid (and the
mask remains within the beam!).
The mathematical statement of the Fraunhofer condition is that
the length of the path from the source to a point P is a linear function of
the mask coordinate y.
The practical interpretation of this is that
Fraunhofer diffraction is observed in the image plane of the point source.













Fig. 7.10 Fraunhofer diffraction

S
screen
P
O
31/10/2007 PHYS2023 Wave Physics
60 Fraunhofer diffraction

7.8 The resolution of an imaging system
Figure 7.10 offers an immediate insight into the resolution with which a lens
may form an image, for the image of each point of an object will be spread by
diffraction from the effective aperture corresponding to the finite size of the
lens. Practical imaging systems must optimize the combination of diffraction,
which is improved by increasing the lens diameter, and aberration, which is
better with smaller lenses. It is interesting that the diffraction limit of the eye
is approximately the same as the sampling limit on the resolution, determined
by the density of rods and cones on the retina.

Das könnte Ihnen auch gefallen