Tutorial 5: Reciprocal Lattice

Contents
1 Introduction 1

2 Example: FCC to BCC 1

3 Miller Indices 3

4 Bragg Diffraction 5
4.1 Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
4.2 Structure Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
4.3 Body Centered Tetragonal Lattice . . . . . . . . . . . . . . . . . . . . . . . 6

1 Introduction
~ can be reached by means of the primitive vectors
In a Bravais lattice, all lattice points R
{~ai } and a set of integers {ni }:
X
~ =
R ni~ai . (1)
i

The sum extends over all spatial dimensions of the system. The lattice points of the
~ are defined via
reciprocal lattice, denoted G,
~ ~ ~ ·R
~ = 2πn
eiG·R = 1 or G (2)

and can be reached, from an arbitrary origin, via the primitive vectors

~bi = 2π ~aj × ~ak (3)

~ai (~aj × ~ak )

and a set of integers (h, k, l) known as Miller indices:

~ = h~b1 + k~b2 + l~b3 .

G (4)

Note that the denominator in Eq. (3) is nothing but the volume of a primitive unit cell.

2 Example: FCC to BCC

Statement of the Problem Find the reciprocal lattice of a FCC lattice.

1
 𝒃𝟐 𝒃𝟑
𝒂𝟐 reciprocal
𝒂𝟑
𝒃𝟏

𝒂𝟏

Figure 1: FCC to BCC. The primitive vectors of the real space FCC lattice and the
reciprocal space BCC lattice are indicated.

Solution The first step, always, is to find the primitive vectors ~ai . You may convince
yourself that the vectors
a a a
~a1 = (x̂ + ŷ) , ~a2 = (x̂ + ẑ) and ~a3 = (ŷ + ẑ) . (5)
2 2 2
are the simplest choice for the FCC lattice (Fig. 1). The primitive unit cell spanned by
them contains one atom (no basis) and its volume is
a3 a3
VP = ~ai (~aj × ~ak ) = (x̂ + ŷ) · [(x̂ + ẑ) × (ŷ + ẑ)] = − . (6)
8 | {z } 4
ẑ−ŷ−x̂

Here we randomly assigned i = 1, j = 2 and k = 3 but any other choice (cyclic permuta-
tion) yields the same answer.
We now have all we need to apply Eq. (3). Starting with ~b1 we obtain
2
~b1 = 2π ~a2 × ~a3 = − 8π a [(x̂ + ẑ) × (ŷ + ẑ)] = 2π (x̂ + ŷ − ẑ) . (7)
VP a3 4 a
Likewise, making a cyclic permutation:
2
~b2 = 2π ~a3 × ~a1 = − 8π a [(ŷ + ẑ) × (x̂ + ŷ)] = 2π (x̂ − ŷ + ẑ) (8)
VP a3 4 a
and, performing a second permutation:
2
~b3 = 2π ~a1 × ~a2 = − 8π a [(x̂ + ŷ) × (x̂ + ẑ)] = 2π (−x̂ + ŷ + ẑ) . (9)
VP a3 4 a
We have thus found the primitive vectors of the reciprocal lattice.
What kind of lattice does this describe? With a little imagination (and Fig. 1) you
may see that it is a BCC lattice. The corners of the cubic unit cell1 are at

~b1 + ~b2 = 4π x̂, ~b2 + ~b3 = 4π ẑ and ~b3 + ~b1 =

4π
ŷ (10)
a a a
1
The BCC lattice is a Bravais lattice and is described by the obtained primitive vectors. Alternatively
one may think of it as two, mutually displaced SC lattices (two base vectors). In this case the primitive
unit cell is a cube and contains two atoms.

2
while the body-center lattice point is at

~b1 + ~b2 + ~b3 = 2π (x̂ + ŷ + ẑ) . (11)

a

3 Miller Indices
~ [Eq. (4)]. The integers h, k and l can
The points of the reciprocal matrix are found at G
also be used in the real space lattice:

~ = h~a1 + k~a2 + l~a3 .

R (12)

In this case the vector R~ is also denoted [hkl] (square brackets). The set of all vectors
equivalent to [hkl] are denoted hhkli (angled brackets). So for example in a cubic lattice

h100i = [100] , [010] , [001] , · · · (13)

because the three primitive vectors are equivalent.

More importantly, the Miller indices can be used to define planes in the lattice. The
real space plane perpendicular to the reciprocal lattice vector G, ~ given in Eq. (4), is
denoted (hkl) (round brackets). This is especially useful in a cubic lattice since its recip-
rocal lattice is also cubic. Our following discussion will thus be limited to the cubic lattice
where the plane (hkl) is also perpendicular to R ~ = h~a1 + k~a2 + l~a3 . A comprehensive way
to imagine said plane is to consider where the plane cuts the Cartesian axes, namely:

~a1 ~a2 ~a3

, and . (14)
h k l
~ with the limitation that
In fact, the notation (hkl) includes all planes perpendicular to R,
only planes with at least three lattice points in them are legal. Take for example the (100)
plane in a cubic lattice. It intersects with the x-axis at ~a1 and not at all with the y- and
z-axis (~a2,3 /0 = ∞). Hence the notation (100) includes all planes parallel to the y − z
plane as shown in Fig. 2. The distance between those planes can be found from

2π
d= , (15)
~ min |
|G

𝟏𝟎𝟎 𝟎𝟏𝟏 ഥ𝟏𝟎

𝟏 𝟏𝟎𝟐

Figure 2: Planes. For the Miller indices (100), (011), (1̄10) and (102) one or two of the
group of planes is shown.

3
where G ~ = h~b1 + k~b2 + l~b3 in which the greatest common divisor
~ min denotes the vector G
(gcd) was taken out. For the (100) planes we have G ~ min = ~b1 = 2π/ax̂ and hence d = a,
as expected.
In addition there are planes which are not parallel to (hkl) but, due to the symmetry
of the lattice, are equivalent. The set of all such planes is denoted {hkl} (curly brackets).
For the cubic lattice we have, e.g., {110} = (110) , (101) , (011) , (11̄0) , (1̄01) , (01̄1).
A negative Miller index is indicated by a bar over the number. So instead of writing
(−100) we use the notation (1̄00).

Statement of the Problem In a cubic lattice, sketch the planes (011) and (1̄10).

Solution See Fig. 2. The simplest way to construct them is to use Eq. (14) to find
the three intersections and complete the plane. Remember that a 0 index means no
intersection with that axis. Also note that planes may be moved, as was done for (1̄10)
in Fig. 2.

Statement of the Problem Consider a cubic lattice.

a) For general (hkl), find an expression for the interplanar distance d.

b) Find d for the planes (011) and (102) (see Fig. 2).

a) In a simple cubic lattice the real space primitive vectors are

~a1 = ax̂, ~a2 = aŷ and ~a2 = aẑ (16)

and hence, using Eq. (3), the reciprocal primitive vectors are given by

~b1 = 2π x̂, ~b2 = 2π ŷ and ~b3 =

2π
ẑ. (17)
a a a

The reciprocal lattice points are

~ = 2π (hx̂ + kŷ + l~z) .

G (18)
a

Using Eq. (15) we find the interplanar distance to be

a
d= √ . (19)
h2 + k 2 + l2

√ √
b) Plugging in (011) and (102) leads to d = a/ 2 and d = a/ 5 respectively. The fact
that the planes for larger Miller indices are closer can be visually appreciated in Fig. 2
and is in agreement with Eq. (14).

4
 𝑘 𝑘′

Figure 3: Bragg diffraction. The incoming plane wave ~k is scattered to ~k 0 by the

different planes. The deeper-going wave travels 2d sin θ further than the wave reflected at
the surface.

4 Bragg Diffraction
4.1 Concept
Consider a plane wave with wave length λ and wave vector ~k (|~k| = 2π/λ) incident on a
lattice. This situation is illustrated in Fig. 3 on a square lattice for simplicity. The out-
coming plane wave has wave vector ~k 0 (|~k 0 | = |~k|; but different direction2 ). Constructive
interference will occur if the additional phase accumulated by the deeper-going wave is
an integer multiple of 2π:
 
~k − ~k 0 · d~ = 2πn, (20)

~ = d is perpendicular to the plane3 . Note: ~k − ~k 0 = G.

where |d| ~ Using ~k · d~ = |~k||d|
~ sin θ
(and likewise for k~0 with θ → −θ) on the lhs we obtain

~ sin θ = 2d 2π sin θ.
 
~k − ~k 0 · d~ = |~k||d|
~ sin θ + |~k||d| (21)
λ
The condition for constructive interference can thus be written as

2d sin θ = λn, (22)

a formula known as the Bragg condition.

In a typical experiment the angles θ for which constructive interference occurs are
measured. While assorting them to the integers n, one looks for the value of d such that
Eq. (22) is satisfied.
2
This means that energy (~ω = ~c|~k|) is conserved but momentum (~~k) is not, due to the rigidity of
the lattice.
~ ~0 ~
3
Alternatively, the condition is ei(k−k )·d = 1

5
4.2 Structure Factor
If the lattice has a basis the Bragg condition alone is not enough. In fact, reflection from
different atoms inside the same unit cell (which are reached via the base vectors) may
result in destructive interference. For this one must consider the structure factor
~ ~
X
Shkl = fj e−iG·dj . (23)
j

The sum extends over all base vectors and fj is a coefficient known as the form factor. It
essentially is the Fourier transform of the charge or mass distribution in the lattice.
The intensity of a wave scattered from the plane (hkl) is proportional to Shkl . Hence
if Shkl = 0, constructive interference does not occur even though the angle θ and distance
d satisfy Eq. (22).
If there is only one base vector (d~0 = ~0) the sum in Eq. (23) disappears and we get
Shkl = f0 for any (hkl). The structure factor is therefore irrelevant if a baseless description
is used.

Statement of the Problem Treat an FCC lattice as a SC lattice with 4 base vectors
and find the planes (hkl) for which constructive interference can occur. You may assume
fj = f for all j.

Real Space and Reciprocal Description From the arbitrary origin d~0 = ~0, the SC
lattice is reached via the primitive vectors in Eq. (16). Now, since we are dealing with a
FCC lattice, we must use a basis to reach the face-centered points. The vectors
ax̂ + aŷ ax̂ + aẑ aŷ + aẑ
d~1 = , d~2 = and d~3 = (24)
2 2 2
take care of this (see Tutorial 4). This description illustrates that a FCC lattice can be
viewed as 4, mutually displaced SC lattices.
Before we apply Eq. (23), the inverse lattice points G ~ must be found and hence we
need the reciprocal lattice primitive vectors. As above, by means of Eq. (3), it is easy to
see that they are given by the vectors in Eq. (17).

Structure Factor We are now ready to use Eq. (23):

~ ~
X
Shkl = fj eiG·dj
j
3  
X 2π
= fj exp −i (hx̂ + kŷ + lẑ) · d~j (25)
j=0
a
= f 1 + e−iπ(h+k) + e−iπ(h+l) + e−iπ(k+l)
 

We must make sure that this expression is non-zero. If all Miller indices are even we
get Shkl = 4f . The same is true for all odd. But if two are odd and one is even or the
opposite, Shkl = 0. This means that only the planes
(111) , (200) , (220) , (222) , (311) , ... (26)
can be seen in a Bragg diffraction experiment but not, e.g., (100).

4.3 Body Centered Tetragonal Lattice

6
Statement of the Problem Given is a tetragonal lattice4 with
lattice constants a = 3Å and c = 4Å and a body-centered lattice
point (see Fig. 4). We consider two types of materials. In type A all 𝒂𝟑
atoms have unity form factor (fj = 1), while in type B the center
atom has form factor −1 (all others are 1). On both materials a 𝒅𝟏
Bragg diffraction experiment (λ = 1.5Å) is conducted.

a) In material A, for which angle θ will the first (n = 1) con- 𝒂𝟐

structive interference occur? 𝒂𝟏

b) Same for material B.

Figure 4: BCT
c) The Bragg experiment is conducted on a mixture of both mate- lattice.
rials. The result is an intensity ratio IA : IB = 4 : 1. Assuming
that the intensity is proportional to the amount of material,
what is the concentration ratio nA /nB ?

Real Space and Reciprocal Description We use the vectors

~a1 = ax̂, ~a2 = aŷ and ~a3 = cẑ (27)

to describe the tetragonal lattice and, in addition to d~0 = ~0, the base vector
1
d~1 = (ax̂ + aŷ + cẑ) (28)
2
to reach the body-centered point. The reciprocal lattice is given by
 
~ = 2π h k l
G x̂ + ŷ + ẑ . (29)
a a c

a) Lets start with the Bragg condition. For the plane (hkl), the distance between two
consecutive planes [see Eq. (15)] is found from
r r
2π a 2 a2
~ =
|G| h2 + k 2 + 2 l2 → d = a/ h2 + k 2 + 2 l2 . (30)
a c c
From Eq. (22) we get r
λ λ a2
sin θ = = h2 + k 2 + 2 l2 . (31)
2d 2a c
The relevant integers (hkl) are found via the structure constant:

Shkl = 1 + eiπ(h+k+l) . (32)

As can be seen, for h + k + l = 2n − 1 (odd) it vanishes. Hence we must choose a

combination whose sum is even, e.g. (110) or (011). Since a < c, the first, i.e. smallest,
angle is obtained for (011): r
λ a2
sin θ = 1+ 2. (33)
2a c
4
A tetragonal lattice results from stretching a cubic lattice along one of its lattice vectors. The cube
thus becomes a rectangular prism.

7
Plugging in the numbers we get
5
sin θ = → θ = 18.21◦ . (34)
16
There are in fact 8 equivalent planes for which this angle is obtained. They are given by
all the different combinations (±1, 0, ±1) and (0, ±1, ±1). This will be important for (c).

b) The only difference is the structure factor. Here we have Shkl = 1 − eiπ(h+k+l) (note
the minus sign) and hence the sum h + k + l must be odd for constructive interference to
occur. The minimal angle is obtained for (001), namely
r
λ a2 3
sin θ = 2
= → θ = 10.81◦ . (35)
2a c 16
Here, only two planes, (0, 0, ±1), result in this minimal angle.

c) The intensity of the reflected light Ij (j = A, B) is proportional to both the concen-

tration nj and the number of reciprocal lattice points that contribute to the reflection.
The number of reciprocal lattice points contributing is equal to the number of planes in
the set {hkl} for the used Miller indices. As mentioned above, we have 8 (2) for material
A (B). The intensity ratio is thus given by

IA 8nA
= . (36)
IB 2nB
Using the given ratio IA /IB = 4 we find that the concentrations nA = nB are equal. The
intensity of the Bragg signal can thus be misleading.