CRYSTAL DIFFRACTION

Introduction

Solid state physics is mainly concerned with the manner in which waves are
propagated through crystals without scattering, or if there is scattering, then in what
direction will it occur? The waves may either be X-rays or electrons (the so called
matter waves) or may be phonons (quantized lattice vibrational waves) but the basic
principle is the same.
Bragg Construction

The Braggs construction is used to determine the angle at which X-ray diffraction lines
or spots are to be observed. In this case, it is assumed that the X-rays which enter the
crystal at an angle to a certain set of parallel planes as shown in Fig. 19

F
A D C

d
G H
B
Fig. 19: Bragg reflection

For any angle θ, the emergent rays which have been reflected by successive planes
will not be in phase with one another, and they will cancel out. A diffraction beam is
only observed if the reflected rays are in phase with one another. The extra path
travelled by for example ray ABC compared to that travelled by DEF which has been
reflected by the plane above must be a whole number of wavelengths. If the separation
of the planes is d, then the path difference can be shown to be GB+BH=2dsinθ, or to
observe a diffraction beam at an angle θ, we must have 2dsinθ=nλ ….. (1),

where n is the order of diffraction which is an integer and λ is the wavelength of the
incident beam. Eqn. (1) is the Bragg formula and the angles θ for which it holds are
called Bragg angles. If Eqn. (1) cannot be satisfied for any type of plane in a crystal
then diffraction cannot occur and the X-rays will pass through the crystal undeviated.
Also, if the wavelength λ, is too long then Eqn. (1) can never be satisfied since we must
n
always have  sin  or , since 0 ≤ sin θ ≤ 1, implying that 0 ≤ θ ≤ 90.
2d
Thus, λ must not be greater than twice an inter-planar spacing, otherwise diffraction
2d
will not occur. i.e  
n
Diffraction by discrete lattice

Let us assume the crystal as a lattice of discrete atoms where each atom can interact
with any incident wave. Further, it can be assumed that as each atom is excited by a
wave, it reradiates the wave in all directions in phase with the initial excitation at that
particular atom as seen in Fig. 20 (a).

Beam not
incident Diffracted Beam
Diffracted Beam
normally to
the line of atoms

Transmitted
Beam
without

Diffracted Beam In (b)

phase, hence
(a) constructive interference

N.B: Circles represent the wave

crests
At any instant, thus the tangents Beam not in phase hence
Which are common to all circles Destructive interference
Indicate that the waves are all in (c)
phase

Figure 20

The amplitude of the beam in any direction is obtained by summing the contributions
in that direction for each atom, taking into account the phase differences between
them.
Wave vectors

In normal circumstances, a wave is usually characterised by its frequency and

wavelength. In mathematical treatment however, it is much more convenient to treat
waves using wave vectors, k, instead of wavelength λ. The wave vector k is denoted
2
as k  which is a vector directed along the path of the wave propagation. It is

convenient to use k instead of λ, since k can be resolved while λ cannot be resolved
directly. Also if two wave vectors k1 and k2 combine, then the resultant wave vector
is k1+k2. All in all, it is the wavelength λ which is more readily visualized rather than
its reciprocal represented by k. In a large perfect crystal, the algebraic sum of
amplitudes of waves in any arbitrary direction will be zero and its only in special
directions e.g those where the Bragg conditions holds where this will be non-zero,
implying that a diffraction beam is observed. The larger the crystal, the sharper the
diffraction beam, since both constructive and destructive interferences will be more
efficient.

If there are atomic irregularities in the crystal structure there will be fewer atoms
contributing to interference and hence the intensity of the peaks will be reduced,
especially at higher angles of diffraction. This effect will be similar to that obtained by
increasing temperature. Let us express the amplitude of a given wave at a point
denoted by a wave vector r with respect to an arbitrary origin O as A exp[i(ko.r  t )]
…………………………………………………(1), where ko is the wavevector and ω is
the angular frequency of the radiation at a time t. Assume we have a regular array of
atoms situated at a position r’ relative to O as shown in Figure 21.

Fig. 21
The atoms are excited or illuminated by a plane wave which has an amplitude of
A exp[i(ko.r 't )] at each value of r’. Further, we assume that the atoms re-radiate
spherical waves in phase with the incident radiation. Also, although the wavelength
of the secondary wave will be unchanged, their direction will in general be different
from ko, for example k. It is desired to calculate the amplitude of the secondary wave
at some point r which is far from the atom, so that the section of the wave front can be
treated as a plane. The vector from the point r to an atom at r’is r-r’. As it leaves the
atom, the wave has an amplitude which is proportional to exp[i(ko.r 't )] . Thus, the
amplitude of the wave at observation point r is A' exp(iko.r ' ) exp[ik (r  r ' )] exp(it )
……………………………………………(2)
This can further be written as A' exp(ik.r ) exp[i(k  ko).r ' ] …………………….(3)

Where the term in t has been dropped.

If the observation point r is far from the atom, then the wavevectors k from all atoms
to r may be assumed to be in the same direction so that the effect at r due to all atoms
in the crystal is obtained by summing a set of terms over all lattice points r’, i.e total
amplitude at r of wave vector k is A' exp(ik .r )  exp[i (ko  k ).r ' ]
attatomsatr '

…………………………………. (4)

This can be written as amplitude(k )  A' exp(ik.r )  f (r ) exp[i(ko  k ).r ' ]dr' …… (5)
allspace

Where f(r’) is a periodic function which has the value of unity at a lattice site and
which is zero elsewhere. We now wish to investigate the conditions under which eqn
(5) is not zero, which are the conditions under which a diffraction pattern observed.

Fourier Transforms

In Fourier transform, if f(α) is the amplitude of a complicated waveform with a period

of 2 which extends from ±∞, f can be expressed as a sum i.e:

f ( )  A1 sin   A2 sin 2  ...  An sin n  ...B1 cos  B2 cos 2  ...

…………………….(6)

Where the extra terms in 2α, 3α etc are harmonics of the fundamental as shown in the
figure below:
 
1
The coefficient An 
  f ( ) sin nd ……………………………………………(7)


Where n=integer and the same applies to Bn. The integral given by eqn (7) is used since
 

 sin m sin nd is zero unless m=n and sin m cos nd  0 for any m,n.
 

Thus, if eqn (6) is substituted into eqn (7), the result is the contribution which the term
nα makes to the function f(𝛼).

If f is not periodic in the angle α, we can change to a one-dimensional spatial variable

x so that f(x) is periodic with a characteristic wavelength λc. This may be analysed into
2
waves of the form, An sin(2nx / c ) or An sin( nkc x) , where kc  . An can be expressed
c

kc
kc
as An 
  f ( x) sin(nk x)dx by
c substituting α=kcx. This can be expressed in 3-

kc

dimensions by using the complex notation which includes sine and cosine terms as
 f (r ) exp(ink .r )dr or  f (r ) exp(ik.r )dr ……………………………………………….(8)
c

The periodic function f(r) is the atomic lattice which has a fundamental wavelength
 f  2 k which is equal to the unit cell dimension. Eqn (8) is only non-zero if k is of
f

the form, nkf, i.e an integral number of waves of wavelength 2 , and these must
nk f
fit into the length c  2 . For the wave to fit into the lattice, an integral number of
kf
wavelengths of each component of the original wavevector is needed to fill the
corresponding dimension a1, a2 and a3 of the unit cell lattice as shown in Fig. 22.

Figure 22.
In this case, n1λ1=a1, n2λ2=a2 and n3λ3=a3. This can further be written as 2n1  a1 or
k1
2n 2  a or 2n3
k1  n1  2  ; k 2  n2  2  ;  a3 or k 3  n3  2  ….(9)
 a1  k2  a2  k3  a3 
2