Crystal structure

Diffraction and Reciprocal Space

The significance of Reciprocal Space
Elastic Scattering by Matter
Diffraction Methods

1. Crystalline structure. Reciprocal space

1.1 Crystal structure

The world "crystal" will have for us a different meaning than from daily life. In solid state
physics, a crystal structure is a unique periodic arrangement of atoms or molecules. This periodic
arrangement can manifest at the microscopic scale as in the beautiful facets of some natural or
artificially grown crystals. Nevertheless such order exists in many materials not exhibiting any
apparent sign of it, as in a copper wire. A crystal can be described as composed of a lattice and a
basis.
Definition 1.1.1 — Lattice. An ensemble of mathematical points all totally equivalent which
are defined by translation symmetry through the translation vector :

~R = n1 a~1 + n2 a~2 + n3 a~3 (1.1)

with n1 , n2 and n3 integer values and ~ai , i = 1, 2, 3 called the primitive vectors or cell
parameters which lie in different directions and span the lattice.

The space in which the lattice is defined is called the direct or real space, as real objects
like atoms or molecules lie on it.
Definition 1.1.2 — Basis. Whatever number of atoms or molecules associated to each one of
the points of the lattice.

Figure 1.1: A crystal with a non atomic basis.

 4 Chapter 1. Crystalline structure. Reciprocal space

A basis can contain just one atom which in such a case will usually be defined by coordinates
~r1 = (0, 0, 0)) or can contain thousands of atoms defined by their vector coordinates ~ri .
In crystallography lattices are also referred as Bravais lattices. Two Bravais lattices are
considered as equivalent when the symmetry operations generating the two lattices are the same.
Then, there are only 14 possible Bravais lattices in three-dimensional space(see Fig. 1.2).

Figure 1.2: The 14 Bravais lattices allowing to describe all 3D crystals. The balls represent not
atoms but lattice points. Of course in various particular crystals one atom can be associated to a
lattice point.

Definition 1.1.3 — Crystal. A crystal is then defined by a Bravais lattice having associated
to each one of its points the same basis:
Crystal = Bravais Lattice + Basis

For a given crystal, the choice of lattice and basis is not unique as illustrated in Figure 1.3.
The vectors ~ai , the cell parameters, define the cell. If in the crystal only one atom is included
in the cell, this is a primitive cell. All unitary cells should have the same surface. A particular
and important unitary cell is the Wigner-Seitz cell defined as the locus of points in space that are
closer to a lattice point than to any of the other lattice points.
The lattice points can be grouped in crystallographic planes (or lines in 2-dimensions)

1.2 Diffraction and Reciprocal Space

A number of techniques allow to obtain experimental evidence of the periodicity of atomic
structures:
(1) x-ray diffraction (2) neutron diffraction (3) electron diffraction (4) High resolution
electron microscopy (5) Atomic force microscopy
(1) and (2) can probe the full crystalline structure, whereas (3), (4) and (5) tend to probe the
atoms closer to the surface or even just the surface atoms.
The principle of a diffraction experiment is depicted in Figure 1.4. As it will be later
discussed, there are different geometries for such experiments depending on the sample and
radiation characteristics.
1.2 Diffraction and Reciprocal Space 5

Figure 1.3: Different choices of crystal lattices. Here every dot represents an atom (or a molecule).
In (1) and (2) the basis is of only one atom. The associated cells are then unitary cells. In (3)
a basis of 2 atoms is needed and in (4) four atoms are needed to complet the basis. (5) is the
Wigner-Seitz cell which is a unitary cell.

Figure 1.4: Principle of a diffraction experiment. The incoming beam characterized by a

wavelength λi is scattered by the sample in the direction 2θ with a wavelength λs . Diffraction is
an elastic process and then λi =λs . In practice the experiment is designed to determine all the
scattering angles needed to obtain the sample structural information.

The two basic and equivalent formulations of the diffraction process are the Bragg law and
the Laue law.
In the Bragg point of view, the incoming beam of wavelength λ is scattered elastically by
two planes of scatterers (atoms or molecules in general) separated by a distance d, corresponding
to a given interreticular distance. As shown in Fig. 1.5., the difference of optical path between
the beam scattered by the two first planes is ∆ = XY Z (beam 2 lags beam 1 by XYZ). If the
beam impinges on the crystal surface with an angle of θ , ∆ = 2d sin θ The diffraction condition
corresponds to a constructive interference between the 2 beams, i.e., ∆ = nλ . Then the Bragg
condition writes:

2d sin θ = nλ

In the Laue description (Fig. 1.6) we consider a vector ~R connecting 2 lattice points of the
periodic lattice. ~R is a lattice vector as it has already been defined. This time we consider the
beam of wavenumber k~0 = 2λπ0 n~0 scattered by the two lattice points in the direction k~1 = 2λπ1 n~1 ,
6 Chapter 1. Crystalline structure. Reciprocal space

Figure 1.5: Bragg law. Two rays (1 and 2) are diffused elastically by the lattice planes.

Figure 1.6: Laue law. Two rays are elastically diffused by two lattice points separated by a vector
vecR.

where n~0 and n~1 are unit vectors. As diffraction is an elastic process λ0 = λ1 ≡ λ , i.e., the
incident and scattered wavenumber moduli are equal, k1 = k0 .
Again we consider the difference of optic paths which is:

λ
δ ≡ δ0 + δ1 = R cos ϕ0 + R cos ϕ0 = −~R~ n1 = ~R(k~0 + k~1 )
n0 + ~R~ (1.2)
2π
The diffraction condition writes this time:

  λ  
δ = nλ ⇒ ~R k~1 − k~0 = nλ ⇒ ~R k~1 − k~0 = 2πn (1.3)
2π

Definition 1.2.1 — The scattering vector. Is the vector difference between the scattered
and incident wavevectors:
~k = k~1 − k~0 (1.4)

The Laue condition writtes then:

~R~k = 2nπ (1.5)

or

exp (i~R~k) = 1 (1.6)

1.2 Diffraction and Reciprocal Space 7

This constitutes, as it will be shown later, an equivalent form of the Bragg diffraction
condition. As a consequence, for each ~R value, the Laue condition associates a diffusion vector
~R which fulfills the diffraction condition. The Laue condition then makes an association between
points (or vectors) between two different spaces.
We will note the ~k vectors verifying the Laue conditions G~ or K
~

{~R} ↔ {K}
~ (1.7)

For a given λ this association is unique.

As {~R} constitute the points of a lattice, {K}
~ are equally points of another lattice.
We call the direct lattice or the Bravais lattice, the lattice constituted by the ~R vectors:
~R = n1 a~1 + n2 a~2 + n3 a~3 (1.8)

where the {~ai } are the real lattice basis.

~ vectors:
We call the reciprocal lattice, the lattice constituted by the K
~ = ha~∗ + ka~∗ + ℓa~∗
K (1.9)
1 2 3

with (h, k, ℓ) integers.

and the {a~∗i } constitute the reciprocal lattice basis. The volume defined by the {a~∗i } is the
reciprocal unit cell.
The {~R} lattice are points of the so called real space {~r} and the {K} ~ are points of the
reciprocal space {k}.~
For a 1-D crystal with a direct lattice ~R = n1 a~1 = n1 ax̂ with x̂ a unitary vector, the Laue
condition, ~R~k = 2πn, implies:

2π
a~∗ = x̂ (1.10)
a
~ should then verify:
In a 3-D crystal, K

 
(n1 a~1 + n2 a~2 + n3 a~3 ) ha~∗1 + ka~∗2 + ℓa~∗3 = 2πn (1.11)

which needs to be valid for any n1 ,n2 and n3 integers. This can be verified when:

~ai a~∗j = 2πδi j (1.12)

with δi j the Kronecker delta function defined as:

 
1 if i= j
δi j = (1.13)
0 if i 6= j

and then the above product becomes:

hn1 + kn2 + ℓn3 = n (1.14)

which can be verified always if the {ni } and (h, k, ℓ) are integers, which is always the case.
 8 Chapter 1. Crystalline structure. Reciprocal space

The expression ~ai a~∗j = 2πδi j can be used to compute efficiently the reciprocal lattice basis
vectors.
The integers (h, k, ℓ) have a significance which will be later developed. They are called the
Miller indexes.
The reciprocal lattice vectors,K~ are also often referred as (hkl) vectors or G~hkl .
It is easily shown that the reciprocal lattice vectors verifying ~ai a~∗j = 2πδi j can be written as:

a~2 × a~3
a~∗1 = 2π (1.15)
a~1 (~
a2 × a~3 )

a~3 × a~1
a~∗2 = 2π (1.16)
a~1 (~
a2 × a~3 )

a~1 × a~2
a~∗3 = 2π (1.17)
a~1 (~
a2 × a~3 )

where the product a~1 (~ a2 × a~3 ) represents the volume of the unit cell.
We should note that by symmetry of the expression ~ai a~∗j = 2πδi j , the reciprocal lattice of
the reciprocal lattice is the direct lattice. We may also underline that all these considerations are
done independently of the crystalline basis. Only the lattice is here considered.

1.3 The significance of Reciprocal Space

To understand the geometrical significance of reciprocal lattice, lets consider a vector K ~ of
reciprocal lattice. By definition, this vector fulfills the diffraction condition through the Laue
expression and is also an scattering vector

~ = k~1 − k~0
K (1.18)

with k~0 and k~1 the incident and scattered wavevectors respectively. We can rewrite the
preceeding expression:

~ − k~1 = k~0
K (1.19)

squaring this expression we have:

K 2 − 2K
~ k~1 + k12 = k02 (1.20)

As diffraction is an elastic scattering process, k0 = k1 and then:

K 2 = 2K
~ k~1 (1.21)

which can be also written

 2
K ~
K
= · k~1 (1.22)
2 2
1.3 The significance of Reciprocal Space 9

We can derive from this expression has a number of important consequences:

Consequence 1. If we consider as in Fig. 1.7 the vector K ~ of the reciprocal lattice and the
median plane of this vector, we realize that whatever vector k~1 having common origin with K ~
and its appex on the median plane, fullfills the condition expressed with the above equation.

~ of the
Figure 1.7: Diffraction condition and First Brillouin zone. A median plane for a vector K
~
reciprocal lattice is shown in black. Three scattered vectors k1 are shown.

In other words, for any K ~ of the reciprocal lattice, its median plane defines an infinity of
~
scattered vectors, k1 which fulfill the diffraction condition.
This allows to define a volume in reciprocal espace having as limiting surface precisely the
first planes allowing for the diffraction condition. Such a volume is called the first Brillouin zone
which in fact corresponds to the Wigner-Seitz cell of the reciprocal lattice.
Definition 1.3.1 — First Brillouin Zone. Is the Wigner-Seitz cell in reciprocal space.

Consequence 2. By simple translation of k~1 and including also k~0 we can go from Fig. 1.7
to Fig. 1.8

Figure 1.8: A redrawing of the previous Figure by vector translation.

~ corresponds
We see in the right sketch of Fig. 1.8 that the median plane defined through the K
to the diffracting plane in the Bragg scheme. This demonstrates the equivalence between the
Bragg and Laue schemes.
In addition, we conclude that the median planes of each reciprocal lattice vector K ~ corresponds
to a diffracting plane. This provides a clear geometrical significance to the reciprocal lattice :
each vector of the reciprocal lattice is associated to a Bragg plane. Each (h,k,l) correspond to a
10 Chapter 1. Crystalline structure. Reciprocal space

~ = G~hkl represents not a plane, but a family of

Bragg plane. Let’s demonstrate that in fact, each K
~ ~
parallel planes. The Laue condition GR = 2πn means that

~ ~R = 0, ±2π, ±4π, ±6π, · · ·

G (1.23)

~ being the shortest G

with G ~ in the choosen direction. This expression defines a familly of parallel
planes perpendicular to G and having equal distances between them of d = 2~π .
~
|G|

Figure 1.9: A redrawing of the previous Figure by vector translation.

To see this, let us consider Fig. 1.9. In this figure we proceed as follows:
• We plot G~ from a common origin, O, in direct space and reciprocal space.
• The condition G ~ ~R = 0 defines a plane in direct espace cutting O in which all ~R, direct
~
lattice vectors, are perpendicular to G.
~ ~
• GR = constant defines subsequent planes parallel to the preceeding one in which all the ~R
have the same projection on G. ~ This projection,d, is the distance between planes :

~c~R = d
|~R| cos G (1.24)

We have then :

G ~ = 2π =⇒ d = 2π
~ ~R = 2π =⇒ |G|d (1.25)
~
|G|

Definition 1.3.2 — {h, k, l} familly of planes. We note {h, k, l} the familly of parallel planes
associated to a G~hkl vector of the reciprocal space.

The Miller indeces (hkl) define then a family of parallel planes associated to the reciprocal space
vector, G~hkl . The interreticular spacing in this family of planes is dhkl = 2~π
|Ghkl |
 1.4 Elastic Scattering by Matter 11

1.4 Elastic Scattering by Matter

Diffraction is the elastic scattering of radiation by matter. Any radiation can be in principle
scattered by matter. Due to practical reasons, X-rays and neutrons are the most extended used
radiation for crystalline structure determination. Electrons are also frequently used.
Lets first consider the a Bravais lattice as a density fonction:

 
ρ(~r) = ∑ δ ~r − ~R (1.26)
~R

with ~R being all vectors of the direct lattice.

Let us compute the Fourier Transform of this lattice:

Z Z
( n o )
  N(∞) if ~k ∈ K
~
i~k~r ~ ~
G(~k) = ρ(~r)e d~r = ∑ δ ~r − ~R eik~r d~r = ∑ eik~r = (1.27)
~R ~R 0 otherwise

We recall here that normalization factors in the Fourier transform definition can respond to
different conventions.
We can then write:
 
G(~k) = ∑ δ ~k − K~ (1.28)
~
K

or in other words, the reciprocal lattice is the Fourier transform of the direct lattice.

Diffraction by a Bravais lattice If we consider now the scattered radiation by a solid hav-
ing scattering centers in each of the direct lattice points, we can consider that for a far observer,
each one of the scattering points is emitting a planewave with an amplitude proportional to f0 ,
which we define as the diffusion amplitude of a scattering center. The total diffused amplitude
reaching a far observer will be:

Z
~
A(~k) = f0 ρ(~r)eik~r d~r (1.29)

Here we have considered the incident amplitude, A0 normalized to one.

We then find out that the scattering amplitude by a Bravais lattice is proportional to the
reciprocal space lattice:

A(~k) = f0 G(~k) (1.30)

And the detected intensity :

I = |A(~k)|2 = f02 |G(~k)|2 (1.31)

To further progress we need now to include:

• A base of atoms or molecules
• Atoms having finite dimensions (and not only mathematical points)
12 Chapter 1. Crystalline structure. Reciprocal space

Diffraction by a crystal.
We apply basically the same formalism with the scattering amplitude written as:

Z
~
A(~k) = A0 f (~r)eik~r d~r (1.32)

where here:
• A0 is the amplitude of the incident radiation (we choose here not to normalize for more
generality).
• f (~r) is the scattering amplitude per unit of sample volume. This magnitude will depend
both on the nature of the scatterer (which atom or molecule) and of the radiation(x-ray,
neutrons,...).
In a crystal:
3
~r = ∑ n j a~ j +~ri (1.33)
j=1

with n j and i being integers. The sum represents a lattice point whereas the different ~ri are
the atoms of the basis.
Replacing and factorizing, we have:

" #" #
N1 N2 N3 n n
i~k(n1 a~1 +n2 a~2 +n3 a~3 +~ri ) i~k(n1 a~1 +n2 a~2 +n3 a~3 +) i~k~ri
A = A0 ∑ ∑ ∑ ∑ fi e = A0 ∑ e ∑ fi e
n1 =1 n2 =1 n3 =1 i=1 n1 ,n2 ,n3 i=1
(1.34)

where n is the number of atoms in the base. The value of fi will depend on the radiation and
on the nature of the atom sitting in position ~ri .
The first factor is called L(~k) or form factor. The second factor is called F(~k) or structure
factor. We have then:

Adi f f racted = A0 L(~(k)F(~k) (1.35)

The form factor. To evaluate L(~k) we replace the expression of ~k = kx a∗1 + ky a∗2 + kz a∗3 and
~R = n1 a1 + n2 a2 + n3 a3 and using ~ai a~∗j = 2πδi j we have:

N1 N2 N3
L(~k) = ∑ ei2π n1 kx ∑ ei2π n2 ky ∑ ei2π n3 kz (1.36)
n1 n2 n3

in this expression, the ni are integers as we are summing over all the lattice points, whereas
the k j are real numbers. We are considering here the possibility of having a finite crystal by
taking ni values comprised between 1 and Ni (any translation of this will be equivalent).
Each of the sum terms correspond to a geometrical series of the form:

a1 − ran
a + ar + ar2 + · · · + arn = (1.37)
1−r
with

a1 = 1 an = ei2π (Ni −1)−ri r = ei2π ri (1.38)

1.4 Elastic Scattering by Matter 13

We have then:

   
~ 1 − eiN1 2π kx 1 − eiN2 2π ky 1 − eiN3 2π kz
L(k) = (1.39)
1 − ei2π kx 1 − ei2π ky 1 − ei2π kz

Every one of the numerator or denominator terms in the above equation can be written as:

ix ix ix ix x
(1 − eix ) = e 2 (e− 2 − e 2 ) = −2ie 2 sin (1.40)
2
As we are just interested on the modulus of this function and not the phase we have:

sin2 (πN1 kx ) sin2 (πN2 ky ) sin2 (πN3 kz )

|L(~k)|2 = (1.41)
sin2 (πkx ) sin2 (πky ) sin2 (πkz )

If we consider a one-dimensional case to simplify, we have:

sin2 (πN1 kx )
|L(kx )|2 = (1.42)
sin2 (πkx )

Figure 1.10: The 1-D form factor squared for N=5 (a) and N=50 (b). The x-axis represents kx .

Figure 1.10, shows the shape of the above fonction for different values on N1 .
It is possible to show that:

sin2 (πN1 r1 )
lim → δh (1.43)
N→∞ sin2 (πr1 )
14 Chapter 1. Crystalline structure. Reciprocal space

In 3-dimensions this leads to:

|L(~k)|2 → δhkl (1.44)

This results means that for an infinity sized crystal, the diffracted intensity is only different of
zero when the diffraction condition is fulfilled. In other words for a crystal of large dimensions :

Adi f f racted = A0 L(~(k)F(~k) = A0 F(G~hkl ) (1.45)

For submicrometric crystals ( for grains in fact smaller than about 0.1 to 0.2 µm), it is not
possible to consider that the delta function approximation is valid. The diffraction peaks have a
fullwidth at half maximum which is inversely proportional to the crystallite domain size. This is
the so called Scherrer effect.
Structure factor. For a large enough crystal, the peak intensities will be then evaluated using
the structure factor :

n
~ = ∑ fi eiK~ri ~
F(K) (1.46)
i=1

~ = ha~∗ + ka~∗ + l a~∗ and n the number of atoms in the

~ the reciprocla lattice vectors, K
with K 1 2 3
considered cell (the basis).
If we consider only one atom per unit cell, then:
~ 2 = fi2
|F(K)| (1.47)

All family of planes will diffract with an intensity which only is affected by the k-dependence
of fi (see next section).
Let us consider 2 examples:

Example 1. Structure factor of a bcc monoatomic lattice.

We have atoms at positions ~ri and ~ri + 12 (~a +~b +~c).

h i
~ ri
~ = eiK~
F(K) fi 1 + eiπ (h+k+l) (1.48)

The term between braces becomes zero when :

h + k + l = 2n + 1 (1.49)

This is called an extinction condition or a systematic-absence condition, as all hkl familly of

planes verifying this condition, will show zero associated diffraction intensity. Extinctions can
appear when we have a basis of more than one atom.

Example 2. Structure factor of a bcc diatomic lattice.

We consider here the CsCl salt, which has a bcc structure with the Cs+1 ion in one of the
basis sites and the Cl −1 occupying the other one. Then we have:
~ = fCl −1 + fCs+1 eiπ (h+k+l)
F(K) (1.50)
1.4 Elastic Scattering by Matter 15

where we have taken arbitrarely ~ri = (0, 0, 0). Then we have:

~ = fCl −1 + fCs+1 (−1)h+k+l 6= 0

F(K) (1.51)

if fCl −1 6= fCs+1

The atomic form factor Also called atomic scattering factor, the term fi is not in general a
constant and depends on the atom considered as scatterer and of the radiation.
In the case of x-ray diffraction, the electromagnetic wave is going to interact with the electrons
of the atom, which become the real scatterers. For a given atom with atomig number Z the
scattered amplitude by the atom will be given by:

Z Z
~ ~′
~ =
fi ((k)) ∑ f j (~r′j )eikr j d 3~r′j (1.52)
j=1 atomi

Where ~r′j is the coordinate position of electron j and f j (~r j ) is the diffusion density of that
electron. The dominant electron spatial symmetry and the extension of the electronic claude
determines the x-ray atomic form factor which is nothing else than the Fourier transform of the
electronic density of the considered atom.
For k = 0 the above sum gives fi (0) = Zi , the number of electrons in the atom.
If we consider as a first approximation that the electronic cloude density can be described by
a gaussian profile of full width at half maximum(FWHM) L, the atomic form factor will just be
the Fourier transform of that gaussian profile, i.e., another gaussian of FWHM 1/L. In figure
1.11 are depicted showing that hevier atoms

Figure 1.11: Atomic form factor for x-ray scattering for different atoms and ions. Heavy atoms
or ions are easy to "see" by x-rays, while light atoms as hydrogen are difficult to detect

The isotropic image of the electronic density is less acurate when d or f orbitals are involved.
Neutron scattering. Not having charge, neutrons interact with the atomic nuclei. The
electron spatial extension is several orders of magnitude more important than the nuclei ones and
 16 Chapter 1. Crystalline structure. Reciprocal space

Figure 1.12: Atomic form factor for x-ray scattering for different ions having the same number
of electrtons as Ne. One can observe that Si4+ , having the highest nuclear charge and then the
less diffuse electronic cloud exhibits the largest FWHM. In the case of a punctual atom, the
extension of the electronic cloud will be zero and hence the atomic form factor will have and ∞
FWHM

we can assimilate it to a delta fonction:

ρ(r) = bδ (~r) (1.53)

Consequently, we can consider that the neutron form factor will be equivalent to the Fourier
transform of a delta function, i.e., a constant independent of k.

Z
~
f (~r) = b δ (~r)ei~rk d 3~r = b (1.54)

Considering adequate prefactors, b can be expressed in units of length and is called the
Fermi length. Its value changes from one isotope to the another and varies erratically with Z
(see Fig. 1.13).
Neutrons having a half integer spin, they can interact with the atom magnetic moment, S~A .
If neutrons are spin polarized, with their spin S~n and hS~A S~n i =
6 0 then there is also a magnetic
component of neutron scattered originated by the magnetic interaction. The magnetic scattering
part due to the neutron scattering with the electron spin, will exhibit on its side k-dependence.
atomic form factor neutron x-rays

1.5 Diffraction Methods

The application of structure determination are manyfold and include:
• The begining of material science. As a first step to understand the physico-chemical origin
of the properties of a material, we need to determine their structure.
• To study the evolution of systems as a fonction of the thermodynamic conditions of Pres-
sure, Temperature, Magnetic field. This is particularly important in geology, planetology,
thermodynamics, physics of materials...
• The application of all this concepts to many domains including history of arts or archeology,
material engineering or the one being presently the most developped: structural biology.
1.5 Diffraction Methods 17

Figure 1.13: Atomic form factor for neutrons and x-rays

We can clasify technics of diffraction depending on the sample type, the source characteristics
or the detection methods.
We can distinguish the use of monochromatic sources (λ is fixed) or white sources (also
colled polychromatic sources). Then, depending if we study with these sources a single crystal
or a polycristal, we will face different geometries and detection modes.
Let us first consider the diffraction of a single crystal by a monochromatic source. A very
useful tool to discuss this diffraction method is the so called Ewald construction.Ewald came
up with a geometrical construction to help visualise which Bragg planes are in the correct
orientation to diffract (Fig. 1.14). We consider the reciprocal espace and we plot a sphere having
radious ~ki = k~f . One needs to fix one point of the sphere surface on one point of the reciprocal
lattice that we take as the origin (point O). The sphere needs allways keep this point fixed in
reciprocal espace and we can make turn it around O (which correspond to rotate the crystal
in a diffraction experiment). When pivoting the sphere around O, the sphere touches another
point of the reciprocal lattice, we will find a situation as shown in the figure, in which a vector
k~1 − k~0 = δ k = K.
~ In fact, any two point of the lattice touching the sphere will be connected by a
vector K ~ of the reciprocal lattice. The Ewald sphere will then define the diffracting angle 2θ .
The Ewald sphere allows then to determine all reciprocal lattice planes attainable for a given
λ and provides the diffracting angle for each of ones (or equivalently, the crystal orientation
needed).

Figure 1.14: The Ewald construction in 2D.

Some consequences are easely derived from the Ewald construction. First of all we note that
if 2k0 is smaller than the smaller distance between two reciprocal lattice points, no diffraction
conditions can be attained. This shows that diffraction needs to be done with small enough
wavelengths or in other words with radiation with not too low energies. But what will be the
18 Chapter 1. Crystalline structure. Reciprocal space

optimum wavelength for a diffraction experiment?

If 2k0 is too big with respect to the characteristic distances in reciprocal espace, 2π/a, then
the number of diffraction peaks will be too important to allow for resolving the different peaks.
An optimum is then than 2k0 is of the order of a relative small number of times 2π/a. That is
2k0 ≃ n2π/a with n of the order of a 3-5 typically. Considering that typical cell parameters range
in inorganic crystal in the 3 − 10 Å range, we end up with a typical optimum value of λ ≈ 1 Å.
Of course for large crystal lattices, which can be of hundreds of Å in biomolecular crystals, a
larger λ would be more appropriate.

In the case of x-rays, Ex−ray = hc/λ which for λ ≈ 1 Å leads to Ex−ray ≈ 10 keV. This is
typically the energy which can be obtained in an x-ray tube or an energy which can be easily
extracted in most synchrotron radiation facilities.

In the case of neutrons, as we have a mass particle, En = p2 /2m and p = h/λ . This time
considering λ ≈ 1 Å we find En ≈ 0.08 eV. This is much smaller than the energy with which
neutrons are expelled from the 235U fission reaction. The energy of neutrons expelled from the
reactor core is described by a Maxwell distribution which peaks at about 1 MeV. Neutrons are
then cooled down through colissions with molecules in a bath (typically a liquid as water) at
ambient temperature (kB Tamb ≈ 25 meV). Neutrons obtained in such a way, which are typically
the ones used in diffraction experiments are called thermal neutrons.

The Ewald construction let us also see the expected result when illuminating a single crys-
tal with a polychromatic beam with wavelengths comprised between λ1 and λ2 . From Figure
1.15 we can appreciate that without any rotation of the crystal, several reciprocal lattice points
will fulfill the diffraction condition and give rise to diffraction at the corresponding diffraction
angle provided by the Ewald construction. This is called the Laue method.

Figure 1.15: The Ewald construction for the Laue Method of single crystal diffraction with a
white beam having wavelengths comprised between λ1 and λ2 .

Powder diffraction
An alternative to the study of single-crystal samples is powder diffraction. Techniques based
on powder diffraction use samples made of a very large number of small crystallites having equal
probability for any particular crystallite orientation. True randomness may need in some cases to
introduce sample rotation orientation.
As it is shown in Fig. 1.16, the use of a powder sample with a monochromatic beam in
angular dispersive mode, leads to the formation of rings on the detector corresponding to each
one of the sample diffracting plane. These rings are called the The azimutal integration of each
ring generates a peak at a given angle. The corresponding diffractograme is presented then as
1.5 Diffraction Methods 19

Figure 1.16: Powder diffraction. The figure illustrates de formation of the Debye-Scherrer rings
for two different diffraction planes.

shown in Fig. 1.17.

Crystal structure determination from powder diffraction data can be challenging due to the
overlap of reflections in a single peak in a powder experiment. The experimental data can
be fitted with a model in order to determine the crystalline structure of the studied sample.
Rietveld refinement is one of the most used method in powder diffraction. The Rietveld method
is a so-called full pattern analysis technique. A crystal structure, together with instrumental
and microstructural information, is used to generate a theoretical diffraction pattern that can
be compared to the observed data. A least squares procedure is then used to minimize the
difference between the calculated pattern and each point of the observed pattern by adjusting
model parameters.

Figure 1.17: X-ray powder diffraction pattern. Crosses are the experimental data. The su-
perimposed line corresponds to the Rietveld refinement of the obtained data. Tick marks are
the different diffraction reflexions included in the model and the bottom line is the residual
(difference between the model and the data).