Lecture 2 — Symmetry in the solid state -

Part II: Crystallographic coordinates and Space Groups.

1 Coordinate systems in crystallography and the mathemati-

cal form of the symmetry operators

1.1 Why don’t we employ Cartesian coordinates?

In crystallography, we almost only employ coordinate systems with basis vectors coinciding
with either primitive or conventional translation operators. When primitive translations
are used as basis vectors points of the pattern related by translation will differ by integral
values of x,y and z. When conventional translations are used as basis vectors, points of the
pattern related by translation will differ by either integral or simple fractional (either n/2
or n/3) values of x,y and z. This advantage far outweighs the convenience of using Cartesian
coordinates.

It is important to remark the distinction between points of the space and vectors. Points them-
selves are not vectors, but “differences” between points are uniquely associated with vectors.
We may then write

p2 − p1 = v (1)

and its inverse, i.e., the sum of a point with a vector, yielding another point.

p2 = p1 + v (2)

Once an origin point “o” and a basis for the vector space are chosen, the coordinates of a point
p are the components of the difference vector p − o, a.k.a., the position vector. A coordinate
transformation therefore involves two things: (a) An origin change and (b) a basis change 1 .

Naturally, once a choice of basis has been made, that basis can be used to express vectors other
than position vectors in terms of their components.

In crystallography, basis vectors have the dimension of a length, and coordinates (position vector
components) are dimensionless.
1
For a full treatment of the coordinate transformations in crystallography, see [2]

1
1.2 Notation

We will denote the basis vectors as ai , where the correspondence with the usual crystallographic
notation is

a1 = a; a2 = b; a3 = c (3)

We will sometimes employ explicit array and matrix multiplication for clarity. In this case, the
array of basis vectors is written as a row, as in [a] = [a1 a2 a3 ]. The corresponding column array
will be denoted as [a]T .

Components of a generic vector v will be denoted as v i , where

v 1 = vx ; v 2 = vy ; v 3 = vz ; (4)
 
v1
Components will be expressed using column arrays, as in [v] =  v 2 , whereas the row vector
v3
will be denoted by [v]T . As a reminder of this, we use subscripts for row arrays and superscripts
for column arrays.

A ‘vector is then written as

∑
v= ai v i (5)
i

Note that a vector not involving an origin choice as written in eq. 5 is an invariant quantity, i.e.,
it does not depend on the choice of coordinates.

1.3 Determination of distances and angles: the metric tensor

In Cartesian coordinates, the scalar product between two vectors takes the familiar form

∑
v · u = [v]T [u] = a2 δij v i uj (6)
i,j

where a2 = |a| is the length of the basis vector in whatever units (with dimensions) are employed
(it is generally 1 for Cartesian coordinates with a dimensionless basis). We shall now see how

2
eq. 6 can be generalised to non-Cartesian coordinates. Remembering the generic expression of
v and u we can write

∑ ∑ ∑
v·u= ai v i · aj v j = [ai · aj ] ui v j (7)
i j i,j

The quantities in square bracket represent the elements of a symmetric matrix.

Gij = ai · aj (8)

This matrix is known as the metric tensor, and has the dimension of a length square. The metric
tensor Gij enables one to calculate the dot product in any coordinate system, as

∑
v·u= Gij v i uj (9)
i,j

Once the dot product is known, one can easily determine distances and angles between points
in any coordinate system. Here is how the metric tensor can be constructed given the lattice
parameters, and how distances and angles are calculated:

A dummies’ guide to calculating lengths and angles

• You are generally given the lattice parameters a, b, c, α, β and γ. In terms of these, the
metric tensor can be written as
 
a2 ab cos γ ac cos β
G =  ab cos γ b2 bc cos α  (10)
ac cos β bc cos α c2

• To measure the length v of a vector v:

 
v1
v 2 = |v|2 = [ v 1 v 2 v 3 ]G  v 2  (11)
v3

• To measure the angle θ between two vectors v and u:


v1
1
cos θ = [ u1 u2 u3 ]G  v 2  (12)
uv
v3

3
2 The mathematical form of the symmetry operators

One of the merits of introducing coordinates is that we can express symmetry operators in a
mathematical form. As we here recall briefly, a symmetry operator define a correspondence
between points of the space, to that to each point p(1) is associated another point p(2) that
receives its attributes.

xi (2) = ti + Dji xj (1) or, in matrix form [x(2)] = [t] + D[x(1)] (13)

In eq. 13, ti are the components of the translational part, and D represents the rotational
part.

In 3 dimensions, the matrix D has determinant 1 or −1. In the former case, it describes a proper
rotation, in the latter an improper rotation, such as reflection or inversion.

Improper rotations are operations that change the handedness (left- to right- hand or vice versa).
It can be shown that all improper rotations can be obtained by composing a proper rotation
with the inversion. We shall see more improper rotations later on.

2.1 Dual basis and coordinates: the reciprocal space

Let us assume a basis vector set ai for our vector space as before, and let us consider the following
set of new vectors.

∑
bi = 2π ak (G−1 )ki (14)
k

From Eq. 8 follows:

∑ ∑
ai · bj = ai · 2π ak (G−1 )ki = 2π Gik (G−1 )kj = 2πδij (15)
k k

Note that the vectors bi have dimensions length−1 . Since the bi are linearly independent if the
ai are, one can use them as new basis vectors, forming the so-called dual basis. This being a
perfectly legitimate choice, can express any vector on this new basis, as

∑
q= qi b i (16)
i

4
As we just said, we can write any vector on this new basis, but vectors expressed using di-
mensionless coordinates on the dual basis have dimensions length−1 , and cannot therefore be
summed to the position vectors. We can consider these vectors as representing the position vec-
tors of a separate space, the so-called reciprocal space.

Position vectors in reciprocal space are linear combinations of the dual basis vectors with
dimensionless components. Their dimension is length−1 . If a primitive basis is used for “di-
rect” (normal) space, then reciprocal lattice vectors are reciprocal-space position vectors
with integral components. See next lecture for a fuller discussion).

The dot product between position vectors in real and reciprocal space is a dimensionless quantity,
and has an extremely simple form (eq. 17):

∑
q · v = 2π qi x i (17)
i

In particular, the dot product of integral multiples of the original basis vectors (i.e., direct or
real lattice vectors), with integral multiples of the dual basis vectors (i.e., reciprocal lattice
vectors) are integral multiples of 2π. This property will be used extensively to calculate Fourier
transforms of lattice functions.

Recap of the key formulas for the dual basis

• From direct to dual bases (eq. 14)

∑
bi = 2π ak (G−1 )ki
k

• Dot product relation between the two bases (eq. 15)

ai · bj = 2πδij

• Dot product between vectors expressed on the two different bases (eq. 17)
∑
q · v = 2π qi xi
i

2.1.1 Dual basis in 3D

In 3 dimensions, there is a very useful formula to calculate the dual basis vectors, which makes
use of the properties of the vector product:

5
 a2 × a3
b1 = 2π (18)
a1 · (a2 × a3 )
a3 × a1
b2 = 2π
a1 · (a2 × a3 )
a1 × a2
b3 = 2π
a1 · (a2 × a3 )

Note that

( )1/2
v = a1 · (a2 × a3 ) = abc 1 − cos2 α − cos2 β − cos2 γ + 2 cos α cos β cos γ (19)

is the unit cell volume. In crystallographic textbooks, the dual basis vectors are often written as
a∗ , b∗ and c∗ .

3 Symmetry in three dimensions

3.1 Key differences with respect to 2D

• Whereas in 2 dimensions, inversion is indistinguishable from 2-fold rotation around z,

in 3D these become two distinct operators. Likewise, a mirror line in 2D can be inter-
preted either as a mirror plane or as an in-plane 180◦ rotation. As we have already
mentioned, in 3D these operators have a different behaviour with respect to handedness.
180◦ rotations (and, in general, all the rotations around an axis) are proper rotations, and
their matrices have det = 1. Mirrors and the inversion are improper rotations, and their
matrices have det = −1.

• There are several new improper rotations, most notably the inversion operator ((x, y, z), →
(−x, −y, −z)) and all the possible composition between inversion and axes of order
higher than 2 (the composition between inversion and a 2-fold axis is a mirror plane).
These are known as roto-inversions.

• There is a new type of roto-translation operator: the screw axis, resulting from the compo-
sition of a rotation with a non-primitive translation parallel to it.

3.2 The new generalized (proper & improper) rotations in 3D

The inversion We indicate it with the symbol I, and represent it graphically by a small circle (◦),
which can be combined with other symbols, if required (see for instance the roto-inversion

6
 3̄ below).

The roto-inversions obtained by composition of an axis r of order higher than two with the
inversion, as I ◦ r. These operators are 3̄ ( ), 4̄ ( ) and 6̄ ( ), and their action is
summarized in Fig. 1. The symbols are chosen to emphasize the existence of another
operator inside the ”belly” of each new operator. Note that 3̄ ◦ 3̄ ◦ 3̄ = 3̄3 = I, and 3̄4 = 3,
i.e., symmetries containing 3̄ also contain the inversion and the 3-fold rotation. Conversely,
4̄ and 6̄ do not automatically contain the inversion. In addition, symmetries containing both
4̄ (or 6̄) and I also contain 4 (or 6).

Clearly, more orientations of all the allowed

axes are possible than in the 2D case, so some
kind of convention has to be established to -
+/-
+ -

draw the stereographic projections. Generally,

- + + + +/-

the axis of highest order is chosen to be per-

+ - +/-
-
pendicular to the projection plane).

3.3 The 32 point groups in 3D

Figure 1: Action of the 3̄ , 4̄ and 6̄ operators and their
3.3.1 The 3D point groups with a 2D pro- powers. The set of equivalent points forms a trigonal an-
tiprism, a tetragonally-distorted tetrahedron and a trigonal
jection
prism, respectively. Points marked with ”+” and ”-” are
above or below the projection plane, respectively. Posi-
We can derive 27 of the 32 3D point groups tions marked with ”+/-” correspond to pairs of equivalent
points above and below the plane.
directly from the 10 2D point groups by a pro-
cess of “extrusion”. The procedure is outlined
with an example in fig. 2. Basically, one looks for possible 3D groups that have a 2D point group
as a “projection” onto a plane perpendicular to the highest-order axis.

3.3.2 The other 3D point groups: the 5 cubic groups

In the 3D point groups we have seen so far, in-plane operators are proper or improper 2-fold
rotation, all forming angles of 90◦ with the highest-order axis. The missing groups will therefore
involve at least two higher-order axes, either at 90◦ with each other or set at different angles. It
can be shown that only two such rotations can exist, both related to the symmetry of the cube.

1. 3-fold axes set at 70.53◦ (cos γ = 13 ), as the diagonals of a cube. Composition of two
such rotations in the same direction gives a 2-fold axis through one of the cube faces.

7
 Composition in opposite direction yields another 3-fold axis. By subsequent composition
and graph symmetry, one retrieves all the four 3-fold axes and three 2-fold axes through
faces and diagonals of a cube.

3. 4-fold axes set at 90◦ , as through the faces of a cube. Compositions in any direction gives a
3-fold axis through the cube diagonals. By subsequent composition and graph symmetry,
one retrieves all the four 3-fold axes and three 4-fold axes, plus six 2-fold axes through the
cube edges.

From these two groups, composed with proper

rotation only, plus compositions with the in-
version, one can obtain the 5 cubic point
groups. Their Hermann-Mauguin symbols
are similar to those of the other groups, with
the cube faces as primary symmetry direc-
tion (first symbol), the cube diagonal as sec-
ondary and the cube edges as tertiary. The
cubic symmetry directions are shown in fig. 3.
The Schoenflies symbol (see extended notes)
is T (for tetrahedral) or O (for octahedral) de-
pending on the absence or presence of proper Figure 2: Schematic representation of the method em-
ployed to generate, from 2D point groups, 3D groups that
4-fold rotations. have that 2D group as a projection.

23 (Schoenflies notation T ). Corresponds to the group described in item 1 above, and is the
symmetry of a ”chiral” tetrahedron (e.g., with faces marked with a 3-fold propeller).

2/m3̄ (m3̄ in short, Schoenflies notation Th ). The same generators as 32 plus the inversion. It is
the symmetry of a double tetrahedron yielding a centrosymmetric solid.

432 (Schoenflies notation O). Corresponds to the group described in item 2 above, and is the
symmetry of a ”chiral” cube, for example, with faces marked with a 4-fold propeller. Note
that the 2-fold axis along the tertiary direction is obtained by composition of the 4-fold
axis (say along the z direction) with a 42 2-fold axis (say, along the x direction).

4̄3m (Schoenflies notation Td ). This is the full symmetry of the tetrahedron. It is obtained from
the previous group by replacing the 4-fold axis with 4̄. By the previous argument, the
tertiary 2-fold axes are now replaced by mirrors.

4/m3̄2/m (m3̄m in short, Schoenflies notation Oh ). It represents the full symmetry of a cube
or octahedron.

8
Figure 4: Primitive and conventional cells for the three cubic lattices - Primitive - P (left), Body-centered-I
(middle) and Face-centered-F (right).

3.4 The 14 Bravais lattices in 3D

The procedure followed to derive the 14 Bra-

vais lattices in 3D is closely related to the
one used for the 2D case. However, there
are many more possibilities to obtain “cen-
tered” lattices, i.e., lattices in which the primi-
tive translations are not orthogonal. For exam-
ple, fig. 4 shows the three lattices with cu-
bic symmetry: Primitive, Face Centred Cu-
bic (BCC) and Body-Centred Cubic (BCC)
(you have probably encountered these lattices
already). Here, we simply list the 14 Bravais
lattices with their symmetry (holohedry) and
the crystal classes they support.

Triclinic system (Classes 1 and 1̄, holohedry

1̄, lattice P ). There is no symmetry re-
striction on the basis vectors, which are
Figure 3: The symmetry directions of a cube: primary
therefore allowed to be at any angle with (I) - three 4-fold axes through the faces; secondary (II)
each other. There is a single primitive four 3-fold axes through the corners and tertiary (III) six
2-fold axes through the edges.
lattice and its symbol is P

Monoclinic system (Classes 2, m and 2/m,

holohedry 2/m, lattices P and C)

9
 - No face is centred: the lattice is monoclinic primitive (symbol P ).
- One of the faces containing the “unique” 2-fold axis is centred: the lattice is monoclinic
face-centred (symbol C).

Orthorhombic system (Classes 222, mm2 and mmm, holohedry mmm, lattices P , C, F and
I).

- No face is centred: the lattice is orthorhombic primitive (symbol P ).

- One of the faces is centred: the lattice is orthorhombic A-, B- or C-centred, depending
on which face is centered (this is to some extent arbitrary — the ITC convention
is normally C). The symbols are A, B or C.
- All of the faces are centred: the lattice is orthorhombic face-centred, symbol F .
- The body (middle) of the cell is centred: the lattice is body-centered orthorhombic
(symbol I).

Tetragonal system (Classes 4, 4̄, 422, 4/m, 4mm, 4̄m2, 4/mmm, holohedry 4/mmm, lattices
P and I).

- No face is centred: the lattice is primitive tetragonal (symbol P ).

- The body (middle) of the cell is centred (face-centering can be reduced to body-centering
by a cell transformation): the lattice is body-centred tetragonal (symbol ”I”).

Trigonal system (Classes 3, 3m1, 321, 3̄m1, lattices P and R). This system is peculiar, in that
each class can be supported by two lattices, P and R, with different holohedries.

- The trigonal primitive P lattice is simply the 3D extension of the 2D hexagonal lattice
by a translation along the z axis, and has holohedry P6/mmm. Here, the unit cell is a
hexagonal prism.
- In the trigonal rhombohedral R lattice, the primitive cell is a rhombohedron, i.e.,
a cube ”stretched” along one of the body diagonals (it is easy to see that a rhom-
bohedron has 3-fold symmetry). A conventional larger hexagonal cell (3 times the
volume) can be constructed (”hexagonal” setting; both cells are shown in fig. 5).
Both rhombohedral and hexagonal settings are used and are listed in the ITC.

Hexagonal system (Classes 6, 6̄, 6/m, 622, 6̄m2, 6mm, 6/mmm, holohedry 6/mmm, lattice
P ). This system supports a single lattice, the primitive hexagonal lattice (symbol P ); the
unit cell is a hexagonal prism.

Cubic system (Classes 23, m3̄, 432, 4̄3m and 43̄m, holohedry 43̄m, lattices P , I and F ). See
(Fig. 4).

10
 - No face is centred: the lattice is primitive cubic (P ), and the primitive unit cell is a cube.
- The body of the cubic cell is centred: the primitive cell is a rhombohedron with angles
between edges α = 109.3◦ (cos α = − 13 ). The lattice is body-centered cubic (BCC)
— symbol I.
- All of the faces are centred: The primitive cell is a rhombohedron with angles between
edges α = 60◦ . The lattice is face-centered cubic (FCC) — symbol F .

3.5 3D space group symmetry

There are 230 space groups in 3 dimen-

sions, resulting from the combination of
each of the 14 Bravais lattices with its sup-
ported classes and systematic replacement
of the proper/improper rotations with roto-
translations. With some practice, it is possi-
ble to construct the complete group diagram
from its symbol for most space groups (for cu-
bic groups, this is a rather tedious procedure).
Here, we will simply introduce the new roto-
translations operators that we have so far not
seen.
Figure 5: The rhombohedral primitive cell (left) and
its projection in the larger, hexagonal conventional cell
(right)
3.5.1 Roto-translations in 3D

These remaining operators will therefore be

compositions of proper rotations with translation parallel to them, and are known as screw
axes. Note that improper screw axes are not roto-translations, because they have invariant points,
and correspond to displaced improper rotations.

The following properties can be easily proven:

• For an axis of order n, the nth power of a screw axis is a primitive translation t . Therefore,
the translation component must be m n
t, where m is an integer. We can limit ourselves to
m < n, all the other operators being composition with lattice translations. Roto-translation
axes are therefore indicated as nm , as in 21 , 63 etc.

• Screw axes can be chiral (e.g., right- or left-handed), whence the name screw. Space groups
that contain chiral axes but do not contain reflections are themselves chiral, and they always

11
 come in enantiomorphic (e.g., left- and right-handed) pairs (e.g., P 41 21 2 and P 43 21 2). If
reflections are present, the space group will contain both types of chirality.

3.5.2 Glide planes in 3D

Glide planes in 3D are not essentially differ-

ent from the 2D case, since they are compo-
m a,b or c n e d
sitions of a mirror plane with a translation in
the same plane, and their square is a pure lat- 3
8
1
8

tice translation. However, since mirror planes

in 3D are truly 2-dimensional (unlike the 2D
mirror ”lines”), the translation can be oriented 31
2 21
62 64
at different angles with respect to the symme- 41 43
61 65
try directions lying in the plane. 42
32
63

3.6 Graphical notation for 3D

symmetry operators Figure 6: Top Graphical symbols for mirror and glide
planes. From left to right, mirror planes m top and side
view, glide planes a, b, and c, top view and side view with
A table with all the graphical symbols for the glide translation in the plane of the sheet (dashed line) and
3D symmetry operators is given in the ITC- orthogonal to it (dotted line). diagonal glide n top and side
view, double glide plane e top and side view, pair of dia-
Volume A [1] on pages 7-10. A the most im- mond glides d top and side view. Bottom roto-translation
screw axes of all orders. Twofold axes are also shown in
portant symbols that we have not encountered projection.
so far are collected in Fig. 6. Another impor-
tant ”novelty” in the 3D case is the use of fractions to indicate the vertical fractional coordinate
of features such as inversion centers and horizontal reflection planes and axes (see, for example,
the diamond glides in Fig. 6.)

References

[1] T. Hahn, ed., International tables for crystallography, vol. A (Kluver Academic Publisher,
Do- drecht: Holland/Boston: USA/ London: UK, 2002), 5th ed.

[2] Paolo G. Radaelli, Symmetry in Crystallography: Understanding the International Tables ,

Oxford University Press (2011)

12