III.

Diffraction and Crystal Structure

A. The Laue Condition and Bragg’s Law

B. The Structure Factor Calculation
C. Experimental Details of X-ray Diffraction
D. Atomic Clusters
 A. The Laue Condition and Bragg’s Law
In a diffraction grating for visible light, constructive interference
between light rays passing through slits of the grating leads to
light intensity ONLY at certain locations on the screen:

d
n=2
n=1
n=0
n = -1
 n = -2

n  d sin 
grating
screen
 Starting Assumptions

For x-rays, electrons, and neutrons incident on a crystal,

diffraction occurs due to interference between waves scattered
elastically from the atoms in the crystal.

If we treat the incident waves as plane waves and the atoms as

ideal point scatterers, the scattered waves are spherical waves.
We will assume they are also isotropic.
 Physical Model for X-ray Scattering
Consider two parallel plane waves scattered elastically from two
nearby atoms A and B in a solid material:

 A P
k 
k


O

B
 f  
i ( k r  t )
 incident  e i ( k r t )
 scattered  e
r
 
elastic scattering: k   k f = atomic form factor
(scattering power of atom)
 Phase Difference Between the Waves
For the spherical waves scattered from atoms A and B (of the same type):

f   f  
i ( k r  t   )
A  e i ( k r  t )
B  e
r rB

where: r = position of detector relative to A

rB = position of detector relative to B

 = position of B relative to A
 = phase difference between  A and  B
        
 k    k       (k   k )    k scattering
vector
phase lag phase advance
from O  B from A  P
 Sum of Scattered Waves
fj  
 
 
i ( k r t   j k )
So the wave scattered from the jth atom is:
j  e
rj

 = position of atom j relative to A
Thus the total scattered wave at the detector is:

fj     fj    

 e
 
i ( k r t )    i ( k r  t ) i (  j k )
 
i ( j k )
e e e
all atoms rj R all atoms

For a small sample, the distances rj are all essentially the same ( R). Thus
we see that constructive and destructive interference between the scattered
waves that reach the detector is due to the atomic
 sum. The  detector location
is determined by the scattered wave vector k  and thus k .
 Summing Over All Atoms

c 
Now assume a crystal whose lattice has base b
  
vectors a, b , c , with a total number of
atoms along each axis M, N, and P, 
a
respectively:

Thus the amplitude of the total wave at the detector is proportional to:

  M 1 N 1 P 1    
e   e
i (  j k )
 i[( ma  nb  pc )k ]

all atoms m 0 n 0 p 0

which can be rearranged to give:

3
M 1   N 1   P 1    M 1   
  e ima k
e inb k
e ipc k
  e ima k

m 0 n 0 p 0  m 0 
 Evaluating the Sum
We need  

 
iMa k
only M 1   M 1   1 e
evaluate the
sum inside
e
m 0
ima k
 e
m 0
ia k m

1 e
 
ia k

the brackets:

Now the intensity of

I 
2
  * 

1 e
 
 iMa k
1  e 
 
iMa k

1  e 1  e 
the total wave at the    
detector is given by: ia k ia k

which
I
2e
 
iMa k
e
 
iMa k

 

2  2 cos Ma  k 
 
 
simplifies    
to give: 2e ia k
e ia k
2  2 cos a  k

I
 
1  cos Ma  k  
 
1  cos a  k  
 The Bottom Line
Now use the familiar identity: cos 2 x  cos 2 x  sin 2 x  1  2 sin 2 x

1  cosMa  k  sin 2 Ma  k 
  2 1  
I   
1  cosa  k  sin 2 a  k 
which gives the result: 2 1 


This is the same as the intensity of an M-slit diffraction grating. If M is large

( 108 for a macroscopic crystal), it has very narrow, intense peaks where the
denominator goes to zero. In between the peaks the intensity is essentially
zero. Remembering that there are three of these terms in the intensity
equation, the peaks occur when:
     
a  k  2n1 b  k  2n2 c  k  2n3 n1, n2, n3 integers

 
Compare these relations to Ghkl  a  2h
the properties of reciprocal  
lattice vectors (from ch. 2):
Ghkl  b  2k
 
Ghkl  c  2l
 Summary: the Laue Condition
Replacing n1n2n3 with the familiar hkl, we see by inspection that these
three conditions are equivalently expressed as:
     The Laue condition
k  Ghkl  hA  kB  lC (Max von Laue, 1911)

So, the condition for nonzero intensity in scattered x-rays is that the

scattering vector k is a translation vector of the reciprocal lattice. Since
each reciprocal lattice point indexed by hkl corresponds to a family of (hkl)
lattice planes, we see that the incoming x-rays scattered from the (hkl) lattice
planes undergo constructive interference at only one position of the detector.

The first experimental confirmation of x-ray diffraction by crystalline solid

came from von Laue’s young colleagues Friedrich and Knipping in 1911.
Despite this, von Laue was more inclined to mathematical rather than
experimental analysis of x-ray scattering. The practical application of von
Laue’s work came only through the efforts of others.
 A Simpler Formulation
Just one year after von Laue’s work, two British physicists developed a
simpler (and easier to use) expression for the x-ray diffraction condition, and
actually used it to determine the crystal structure of NaCl!
This was a father & son team: William Henry Bragg and William Lawrence
Bragg. The father is shown at left below, along with Max von Laue.

The Braggs’ experimental skill and

their simple equation allowed them to
quickly determine the crystal structure
of many common salts and metals.
Max von Laue and the Braggs received
the Nobel Prize in physics in 1914 and
1915, respectively.
 From Laue to Bragg

Now the magnitude of the scattering vector k depends on the angle
between the incident wave vector and the scattered wave vector:
  
 k k k

  k
2
d hkl 
k

Do you see why this

  2 angle is 2 ?
Elastic scattering requires: 
k  k k 

So from the wave vector triangle  4  2
k  2k sin   sin   G 

hkl
and the Laue condition we see: d hkl

For each (hkl) family x-rays will

Leaving Bragg’s law:   2d hkl sin 
“diffract” at only one angle 
 A Final Comment
The interplanar distance between (hkl) planes is d hkl . By inspection we can
see that the distance between (nh nk nl) planes is d hkl / n . This means that
we can write the Bragg condition for these planes as:

  2d hkl / nsin  n  2d hkl sin  n  1, 2, 3,...

However, in practice we need only consider the n = 1 values, since the n = 2

and higher values for the (hkl) planes correspond to the n = 1 value for the
(nh nk nl) planes, and this would be redundant.


1 2 
 k k
Question: can you see a simple
derivation of Bragg’s law by
 
requiring constructive interference
d hkl between x-ray paths 1 and 2?
 B. The Structure Factor Shkl
We know that the scattered x-ray intensity is proportional to:
2 2
  M 1 N 1 P 1    
e  e
i (  j k ) i[( ma  nb  pc )k ]
I 
all atoms m 0 n 0 p 0

where the sum runs over all of the lattice points and we assume that there is
only a single atom at each lattice point.
 
Laue and Bragg remind us that for k  Ghkl or   2d sin 
I  0 at the detector:

BUT if we have a crystal with more than one atom per lattice point (a basis
with two or more atoms), we must sum over all atoms in the basis as well!
 Defining the Structure Factor Shkl
    
f
i ( k r t ) i (  j k )
The sum of the scattered
x-rays was found to be:
 e j e
R all atoms
   
f  f
i (  j k ) i (  j Ghkl )
So the amplitude of the A j e  j e
sum is proportional to: all atoms lattice basis

Structure factor S hkl

 
S hkl   f j e
i (  j Ghkl )
The structure factor is a sum
over all atoms in the basis:
j

Where the position of all    

atoms in the basis is given by:
 j  u j a  v jb  wj c
 An example

S hkl  f e  f  I hkl  S hkl  f 2

For the simple cubic lattice i (0) 2
with a one atom basis:

So the x-ray intensity is nonzero for all values of (hkl), subject to the Bragg
condition, which can be expressed   2d hkl sin  .

a
d hkl 
h 
Now we know for cubic lattices: 1/ 2
2
 k2  l2

2
Substituting and squaring both sides: sin 2  
4a 2
h 2
 k2  l2 
Thus, if we know the x-ray
wavelength and are given or can
measure the angles at which each sin 2 
diffraction peak occurs, we can
graphically determine a for the
lattice! How? h2  k 2  l 2
 Another example
Consider the body-centered cubic
lattice with a one atom basis. This is 
i ( a2 iˆ  b2 ˆj  2c kˆ )Ghkl
equivalent to the simple cubic lattice S hkl  f e i (0)
 fe
with a two atom basis, with atoms at
[000] and [½½½]:

Remembering that the base vectors      

in direct and reciprocal space are Ghkl  a  2h Ghkl  b  2k Ghkl  c  2l
related by:

We obtain:   
S hkl  f 1  ei (h k l )  f 1  ei ( h  k l ) 
Now by inspection (or trial and error) 2f if h+k+l is even
we can see that there are only two S hkl 
0 if h+k+l is odd
possible values for the structure factor:
 The result
So for a crystal with the bcc lattice and a one atom basis, the x-ray intensity is
nonzero for all planes (hkl), subject to the Bragg condition, except for the planes
where h+k+l is odd. Thus, diffraction peaks will be observed for the following
planes:
(100) (110) (111) (200) (210) (211) (220) (221) (300) …

Just as before, if we are given or can

measure the angles at which each sin 2 
diffraction peak occurs, we can
graphically determine a for the lattice!
h2  k 2  l 2

A similar analysis can be done for a crystal with the fcc lattice and a one atom
basis, or in other words for the simple cubic lattice with a four atom basis. A
slightly different rule for the values of (hkl) is generated.
 But how can we determine the lattice type?
Typically all we know is the angles at which diffraction peaks are found. If we
suspect a cubic lattice, how can we determine whether it is sc, bcc, or fcc? This
is your assignment in HW problems 3.1 and 3.7.
(See the table on p. 52 for a list of possible h2 + k2 + l2 values for each lattice)

Once you make a plot for each lattice

type and see which one is linear, you
sin 2  will know the lattice type and can then
find the lattice constant a.

h2  k 2  l 2
 C. Experimental Details of X-ray Diffraction
X-ray sources
1. high voltage tubes (30-50 kV, fixed wavelength)
2. synchrotron facilities ($$, but tunable wavelength, very high intensity)

In a high-voltage tube, electrons are accelerated through a kV potential

difference and collide with a transition metal target (Cu, Ni). As the
electrons abruptly slow down after the collision, this deceleration causes EM
radiation to be given off (“braking radiation” or Bremsstrahlung).

Most of the radiation is continuously

distributed across the energy 

Intensity
spectrum, but sharp lines occur that 
are called characteristic x-ray lines.
They are caused by electrons from
higher energies falling down to an
empty energy level from which an Energy
electron was ejected in a collision.
Characteristic X-ray Spectrum
 Techniques of X-ray Diffraction
Laue back reflection
1. “White” x-rays scatter from a single crystal sample
2. Used to determine orientation of sample and lattice symmetry

x-ray source
sample
film

Spots appear on the film at

locations determined by the Bragg
condition for different (hkl) planes
 Techniques of X-ray Diffraction
Debye-Scherrer powder diffraction
1. Monochromatic x-rays scatter from a finely ground polycrystalline sample
2. Used to determine lattice type and detailed crystal structure

2 Film or
movable
Monochromatic sample
detector
source

Intensity
Sample is finely ground so that
essentially all of the (hkl) planes
that can cause diffraction are
present. As a result, an intensity
peak is measured for each of these
planes: 2
 D. Atomic Clusters (Nanoparticles)
A bulk solid can be built up as more and more atoms are combined:

Atom  Molecule  Cluster  Bulk Solid

N: 1  2-100  10-104 > 105

Important questions:
1. At what value of N does a cluster of (metal) atoms begin to display
bulk-like behavior?
2. What interesting properties do smaller clusters have, and how can
they be used?

The answer to #1 typically depends on what property we are interested in, so

there is no universal answer.
 Cluster Stability
In the early 1980s a very interesting property of metal clusters was
discovered: there are specific cluster sizes that are more stable than others
nearby. For Na and other alkali metals these “magic numbers” are:
N = 8, 20, 34, 58, 92, 138, 196, …

Why? A simple explanation:

1. Treat valence electrons (3s for Na) as “free” to move within the
cluster volume
2. Solve the Schrödinger equation for a simple model potential energy
function (spherical infinite potential well)

R
U = 0 inside U =  outside
 Reason for “Magic Numbers”
The energy level structure that comes from solving the Schrödinger equation
is similar to that of the hydrogen atom, but without the same restriction on the
orbital quantum number l:

orbital degeneracy
0
1g (18) (58)
2p (6) (40)
Energy

1f (14) (34)
2s (2) (20)
1d (10) (18)

1p (6) (8)
1s (2) (2)