PCV Ch4

Chapter 4
Group theory
Group theory will be presented in this chapter as a tool for spectroscopy. Indeed group theory
makes it possible to
• construct and classify molecular orbitals,
• classify electronic, vibrational, rotational and nuclear spin wave functions,
• predict which states are allowed,
• predict physical properties (existence of electric dipole moment, optical activity etc.),
• predict selection rules (electric dipole transitions, configuration interaction. etc.)
However, group theory does not make any quantitative predictions. The interest of group
theory lies in simplifying some problems like those mentioned above that arise in molecular
spectroscopy.
4.1 Symmetry operations
4.1.1 Definition of a group
A group G is a set of elements A, B, C, ... connected by a combination rule (written as a

product, for example A · B) which has the following properties:
1. the closure: for all elements A and B of the group G, A · B = C is also an element of
the group G.
2. the associativity: the combination rule must be associative, i. e. A·(B·C) = (A·B)·C.
3. the identity: there must be an element, the identity E (also called unit), such that
E · R = R · E = R for all elements R of the group.
4. the inverses: each element R must have an inverse R−1 which is also a group element
such that R · R−1 = R−1 · R = E.
101
102 CHAPTER 4. GROUP THEORY
In group theory, the elements considered are symmetry operations. For a given molecular
system described by the Hamiltonian Ĥ, there is a set of symmetry operations Ôi which
commute with Ĥ:

Ôi , Ĥ = 0. (4.1)
Ĥ and Ôi thus have a common set of eigenfunctions and the eigenvalues of Ôi can be used
as labels for the eigenfunctions (see Lecture Physical Chemistry III). This set of operations
defines (with the multiplication operation) a symmetry group. In molecular physics and
molecular spectroscopy two types of groups are particularly important, the point groups and
the permutation-inversion groups.
4.1.2 Point group operations and point group symmetry
The point groups adequately describe molecules that can be considered as rigid on the
timescale of the spectroscopic experiment, which means molecules that have a unique equi-
librium configuration with no observable tunneling between two or more equivalent configu-
rations.
The symmetry operations of the point groups are:
• the identity E which leaves all coordinates unchanged.
• the proper rotation Cn by an angle of 2π/n in the positive trigonometric sense (i. e.
counter-clockwise). The symmetry axis with highest n is chosen as principal axis. If
a molecule has a unique Cn axis with highest n, the molecule has a permanent dipole
moment that lies along this axis (e. g. H2 O, NCl3 in Figure 4.1). If a molecule has
several Cn axes with highest n, the molecule has no permanent dipole moment (e. g.
CH4 ).
C2 C3
O N Cl
H H Cl Cl
Figure 4.1: C2 rotation of H2 O and C3 rotation of NCl3 .
• the reflection through a plane σ; the reflections are classified into two categories:
PCV - Spectroscopy of atoms and molecules

4.1. SYMMETRY OPERATIONS 103
– the reflections through a vertical plane, i. e. a plane which contains the symmetry
axis z are noted σv ,
– the reflections through an horizontal plane, i. e. a plane perpendicular to the
symmetry axis z are noted σh .
• the inversion i of all coordinates through the inversion center.
• the improper rotation Sn or rotation-reflection which consists in a rotation by an

angle of 2π/n around the z axis followed by a reflection through the plane perpendicular
to the rotational axis. Figure 4.2 shows the S4 improper rotation of allene as an example.
(4)
H(3) z H(4) H(3)z H H
(2)
z
(1)
H
C(3) C(3) C(1)
C4
sxy
C(2) C(2) C(2)
(1) (1) (3)

C C C
(1) (2) (2) (1)
H H H H H(3) H(4)
Figure 4.2: Decomposition of the improper rotation S4 of allene.
A molecule having an improper operation as symmetry operation, i. e. a reflection, an im-

proper rotation or an inversion, cannot be optically active and is therefore achiral.
4.1.3 Permutation-inversion operations and CNPI groups
As mentioned already, the point groups are well suited to describe rigid molecules. However,
for floppy systems, especially clusters with tunneling splitting as shown in Figure 4.3, or when
the transition between two states does not hold the same symmetry, another, more general
definition is required.
H(1)
F(2) H(2) F(1) F(2) H(1) F(1)
H(2)
Figure 4.3: Tunneling process in (HF)2 .
To circumvent this problem, the complete nuclear permutation inversion (CNPI) groups have

been developed, originally by Christopher Longuet-Higgins and Jon T. Hougen (see Bunker
and Jensen, Molecular Symmetry and Spectroscopy, 1998). Their concept relies on the fact
that the symmetry operations, i. e. the permutation-inversion operations leave Ĥ unchanged.
The symmetry operations of the CNPI groups are:
• the permuation (ij) of the coordinates of two identical nuclei i and j which denotes
the exchange of the nucleus i with the nucleus j (see Figure 4.4 for examples),
• the cyclic permutation (ijk) of the coordinates of three identical nuclei i, j, and k,
i. e. the nucleus i will be replaced by the nucleus j, j by k and k by i (see Figure 4.4
for example),
(12)
O O
H(1) H(2) H(2) H(1)
H(1) (12) H(2)

F(2) H(2) F(1) F(1) H(1) F(2)
(123)
N N
H(2) H(3)
H(3) H(1) H(1) H(2)
Figure 4.4: Examples of (i j) and (i j k) permutations.
• all possible circular permutations of n identical nuclei (for example, the (1 2 3 4 5 6)

permutation in benzene),
• the inversion E ∗ of all coordinates of all particles through the center of the lab-fixed
frame,
• the permutation followed by an inversion (ij)∗ = E ∗ · (ij) of all coordinates of all

particles
• the cyclic permutation followed by an inversion (ijk)∗ of all coordinates of all

particles,
• all possible circular permutations followed by an inversion of all coordinates of n iden-

tical nuclei.

4.2. IMPORTANT CONCEPTS IN A GROUP 105
The permutation operations only affect identical nuclei, therefore the molecular Hamiltonian
is left unchanged upon these operations. Moreover the molecular Hamiltonian depends on
distances rather than positions, hence the inversion operation also leaves Ĥ unchanged.
The CNPI groups represent a more general description that can also be applied to rigid
molecules. Indeed each point group is isomorphous to a CNPI group although the symmetry
operations are not identical (for example, the inversion i of a point group symmetry is not
the same as the inversion of a permutation inversion group E ∗ ).
———————————————————
Example: the point group C3v is isomorphous to S3 = {E, (1 2 3), (1 3 2), (1 2), (1 3), (2 3)}, which
means that there is a one to one correspondence between the two sets of operations.
———————————————————
However, one disadvantage of the CNPI groups is their size which can become very large. For
example, the CNPI for CH4 contains 48 symmetry operations, and that of benzene 1036800!
In the case of non-rigid systems, this problem is usually solved by using a subgroup, i. e. a
subset of the group which forms a group under the same combination rule. These subgroups
are called molecular symmetry (MS) groups.
In the case of rigid molecules, most of the time the point groups are used. In the following,
we will consider rigid molecules only and restrict ourselves to point group symmetry, but all
concepts can be extended to the CNPI and MS groups.
4.2 Important concepts in a group
4.2.1 Order, conjugated elements and classes
The order of a group is equal to the number of elements in the group. The discrete
(or finite) groups have a finite order (for example C2v is a group of fourth order), while
continuous groups have infinite orders (C∞v for example).
Let us consider two operations Ôi and Ôj that leave Ĥ unchanged when applied individually.
Hence, they must also leave Ĥ unchanged when applied in succession. The notation Ôi · Ôj
means that Ôj acts first, and Ôi second. In other words, Ôi · Ôj must be a symmetry operation
Ôk if Ôi and Ôj are symmetry operations, which is a corollary of the closure property of a
group. Very often it is useful to build the so-called multiplication table which summarizes
all possible Ôi · Ôj combinations.
———————————————————

Example: the C2v group of H2 O
The symmetry operations are E, C2z , σ xz and σ yz . From Figure 4.5 one can verify that the successive
application of any two operations of the C2v point group is equivalent to the application of a third
group operation. For instance: σ xz · σ yz = C2z , σ xz · σ xz = E, C2z · σ xz = σ yz , etc.

Figure 4.5: C2v group operations and their effect on a water molecule. The dot indicates
schematically the coordinates (x, y, z) of an electron.
The multiplication table of the C2v point group with four symmetry operations (E, C2z , σ xz , σ yz ) is
thus a 4×4 table.
1st operation (right)
C2v E C2z σ xz σ yz
E E C2z σ xz σ yz
C2z C2z E σ yz σ xz
2nd operation (left)
σ xz σ xz σ yz E C2z
σ yz σ yz σ xz C2z E
Table 4.1: Multiplication table of the C2v point group.
———————————————————
A group G is said abelian or commutative when all operations commute:
Ôi · Ôj = Ôj · Ôi . (4.2)

For example, the C2v group is abelian (see Table 4.1). However, not all groups are Abelian.
An example of a non-abelian group is the point group C3v .
———————————————————
Example: CH3 Cl in the C3v group
The symmetry operations are E, C3 , C32 , σ a , σ b , σ c , hence the group is of order 6. With the help of
Figure 4.6, one can derive the multiplication table of the C3v point group. One sees that the group is
not Abelian because not all operations commute (e. g., C3 · σ a = σ c and σ a · C3 = σ b ). Moreover, not
all operations are their own inverse (e. g., C3 · C3 = E).

Figure 4.6: The operations of the C3v point group with the example of the CH3 Cl molecule
represented as a Newmann projection (adapted from F. Merkt and M. Quack in Handbook of
high-resolution spectroscopy, 2011).

1st operation (right)
C3v E C3 C32 σa σb σc
E E C3 C32 σa σb σc
C3 C3 C32 E σc σa σb
C32 C32 E C3 σb σc σa
2nd operation (left)
σa σa σb σc E C3 C32
σb σb σc σa C32 E C3
σc σc σa σb C3 C32 E
Table 4.2: Multiplication table of the C3v point group.

———————————————————
If Ôi , Ôj , and Ôk are all group elements and if they obey the relation
Ôk · Ôi · Ôk−1 = Ôj , (4.3)
then Ôi and Ôj are called conjugated elements. All conjugated elements in a group form
a class.
———————————————————
Example: Elements of the point group C3v that belong to the same class as C3
We consider C3 = Ôi and apply each operation Ôk of C3v according to Equation (4.3) with the help
of the multiplication table in order to find the conjugated elements of C3 .
Ôi Ôk Ôk−1 Ôk · Ôi · Ôk−1 = Ôj Ôi Ôk Ôk−1 Ôk · Ôi · Ôk−1 = Ôj
C3 E E E · C3 · E = C3 C32 E E E · C32 · E = C32
C3 C3 C32 C3 · C3 · C32 = C3 C32 C3 C32 C3 · C32 · C32 = C32
C3 C32 C3 C32 · C3 · C3 = C3 C32 C32 C3 C32 · C32 · C3 = C32
C3 σa σa σ a · C3 · σ a = C32 C32 σa σa σ a · C32 · σ a = C3

σc σb
b
C3 σb σb σ b · C3 · σ = C32 C32 σb σb σb · C32· σ b = C32

σa σc
c
C3 σ c
σ c c
σ · C3 · σ = C32 C32 σ c
σ c c
σ · C32· σ c = C32

σb σa
C3 and C32 are conjugated; they are elements of the same class of order 2.
Similarly, one can show that σ a , σ b and σ c form a class of order 3.
———————————————————
The order k of an element Ôi is the smallest integer k ≥ 1 with Ôik = E. This property

exists for finite groups only.

With this definition, one can easily find that a rotation Cn is of order n and a reflection σ is
of order 2. All elements of a class have the same order.
4.2.2 Representations and character table
Up to now, we have described each symmetry operation Ô of a group with respect to one
specific molecule. Now, we would like to get a more general picture and represent each
operation of the group with a n×n matrix B, n being the dimensionality of the representation.
This matrix represents how the vectors or functions chosen as basis set (of dimension n)
transform upon the application of Ô. Therefore, the matrix B depends on the coordinate
system, i. e. the vectors chosen to describe the system. Given a coordinate system ẽ, the
matrix representation B (ẽ) of the operation Ô fulfils
y = B (ẽ) x. (4.4)
With a new coordinate system ẽ = Sẽ, the transformation is y = Sy , and x = Sx.
Therefore
y = Sy = SB (ẽ) x = SB (ẽ) S −1 x = B (ẽ ) x

(4.5)

The matrix B (ẽ ) = SB (ẽ) S −1 forms a new, equivalent representation of the operation Ô.
The trace of a matrix remains unchanged upon an unitary coordinate transformation. Thus

Tr(B (ẽ ) ) = Tr(B (ẽ) ). (4.6)
The trace of a matrix representing an operation is also called the character of the opera-
tion χ. An interesting property is that all elements of a class have the same character.
———————————————————
Example : one-dimensional representation (n = 1) of C2v
Case 1: one can use the functions Ψ1 = x, Ψ2 = y, or Ψ3 = z.
Ψ1 = x : Ex=x Ψ2 = y : Ey =y Ψ3 = z : Ez =z
C2z x = −x C2z y = −y C2z z = z
σ xz x = x σ xz y = −y σ xz z = z
σ yz x = −x σ yz y = y σ yz z = z

Therefore x, y, and z correspond to the following representations designated by Γ:
Γ(z) 1 1 1 1
Γ(x) 1 −1 1 −1
Γ(y) 1 −1 −1 1
Case 2: instead of using x, y, or z, more complicated functions can be used to generate a one-
dimensional representation, such as x2 , y 2 , z 2 , xy, xz or yz as listed in the fourth column of the
character table. For example, one takes the functions Ψ4 = x2 and Ψ5 = xy:
Ψ4 = x2 : E x2 = (E x)(E x) = x2 Ψ5 = xy : E xy = (E x)(E y) = xy
C2z x2 = (C2z x)(C2z x) = x2 C2z xy = (C2z x)(C2z y) = xy
σ xz x2 = (σ xz x)(σ xz x) = x2 σ xz xy = (σ xz x)(σ xz y) = −xy
σ yz x2 = (σ yz x)(σ yz x) = x2 σ yz xy = (σ yz x)(σ yz y) = −xy
It is easy to verify that χ(xy) = χ(x) × χ(y) , a result that can be written as a direct product. To
evaluate a direct product, one multiplies the characters of each class of elements pairwise and
obtains as direct product a representation of the group:
Γ(xy) = Γ(x) ⊗ Γ(y) = (1 -1 1 -1) ⊗ (1 -1 -1 1) = (1 1 -1 -1).
Case 3: one can also look at the transformation properties of rotations and for example take Ψ6 = Rz
as illustrated in Figure 4.7.

Figure 4.7: The Rz rotation of water.
Ψ 6 = Rz : E R z = Rz
C2z Rz = Rz
σ xz Rz = −Rz Direction of rotation reversed.
σ yz Rz = −Rz Direction of rotation reversed.
Rz transforms as follows:
Γ(Rz ) 1 1 −1 −1
———————————————————

Representations of higher dimensionality can be obtained by looking at the transformation

properties of two or more functions. Indeed, to construct an n-dimensional representation of
a group, one takes n linear independent functions or vectors Ψi , i = 1, ..., n spanning a given
n-dimensional space. Applying the group operations on Ψi leads to a transformed function
which is a linear combination of the original functions:
n

ÔΨi = bji (Ô)Ψj . (4.7)
j=1
———————————————————
x
Example: two dimensional representation y of the C2v group

⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜ x ⎟ ⎜ x ⎟ ⎜ 1 0 ⎟⎜ x ⎟ ( x)
E⎝ ⎠=⎝ ⎠=⎝ ⎠⎝ ⎠, with χEy = 2
y y 0 1 y
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜ x ⎟ ⎜ −x ⎟ ⎜ −1 0 ⎟⎜ x ⎟ ( x)
C2z ⎝ ⎠=⎝ ⎠=⎝ ⎠⎝ ⎠, with χCyz = −2
2
y −y 0 −1 y
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜ x ⎟ ⎜ x ⎟ ⎜ 1 0 ⎟⎜ x ⎟ ( x)
σ xz ⎝ ⎠=⎝ ⎠=⎝ ⎠⎝ ⎠, with χσyxz = 0
y −y 0 −1 y
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜ x ⎟ ⎜ −x ⎟ ⎜ −1 0 ⎟ ⎜ x ⎟ ( x)
σ yz ⎝ ⎠=⎝ ⎠=⎝ ⎠⎝ ⎠, with χσyyz = 0
y y 0 1 y
The two-dimensional representation of y has thus the following characters:
⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞
⎜ 1 0 ⎟ ⎜ −1 0 ⎟ ⎜ 1 0 ⎟ ⎜ −1 0 ⎟
2 × 2 Matrix ⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎝ ⎠
0 1 0 −1 0 −1 0 1
x
Γ(y) 2 −2 0 0
———————————————————
If the matrices of all elements of a representation of a group can be simultaneously brought
into block-diagonal form by a given coordinate transformation, the representation is said to
be reducible, if not, it is irreducible.
The character table of a group lists all irreducible representations and gives for each rep-
resentation the character of each class of elements.

G C1 = E C2 ... Cn Tx Ty Tz Rx Ry Rz
(1)
Γirr χ(1) (C1 ) χ(1) (C2 ) ... χ(1) (Cn )
(2)
Γirr χ(2) (C1 ) χ(2) (C2 )
.. .. ..
. . .
(n)
Γirr χ(n) (C1 ) χ(n) (C2 ) ... χ(n) (Cn )
In a character table, Γ(n) designates the n-th irreducible representation, Ci the i-th class of
elements and χ(n) (Cj ) the character of the elements of class j in the n-th representation.
There are as many irreducible representations as classes. Next to the characters of the el-
ements of the different classes, the character table also gives in the last columns how the
translations Tx , Ty and Tz and the rotations Rx , Ry and Rz transform.
———————————————————
Example: The character table of the C2v group
A1 1 1 1 1 z x2 , y 2 , z 2
A2 1 1 −1 −1 xy Rz
B1 1 −1 1 −1 x xz Ry
B2 1 −1 −1 1 y yz Rx
x, y, and z correspond to the irreducible representations B1 , B2 , and A1 , respectively as indicated in
the third column of the character table. Rz transforms as A2 as indicated in the fifth column of the
character table. One can verify that Rx and Ry transform as B2 and B1 , respectively.
x
The Γ(y) representation is not an irreducible representation of C2v because it is of dimension 2, and
x
C2v has one-dimensional irreducible representations only. Γ(y) is reducible, i. e., it corresponds to a
x
linear combination of irreducible representations: Γ(y) = B1 ⊕ B2 .
———————————————————
Character tables exist for all groups. Many groups have a finite number of representations,
but groups with an infinite number of representations also exist such as D∞h and C∞v .
Important remark: The character of the unity operation (E) is always equal to the dimension
of the representation.

4.2.3 Reduction of reducible representations
There is a systematic mathematical procedure to perform the reduction of representations.

All representations in a character table form a set of orthogonal vectors that span the complete
space:

χ(i) (Ô) × χ(j) (Ô) = hδij , (4.8)
Ô
where h represents the order of the group and Ô runs over all the elements of the group. (Note
that some classes of non-Abelian groups contain more than one element!). Any reducible
representation can thus be expressed as a linear combination of irreducible representations

Γred = cred
k Γ ,
(k)
(4.9)
k
where Γ(k) represents an irreducible representation. The expansion coefficients cred

k can be
determined using the reduction formula Equation (4.10):
1 red
cred
k = χ (Ô) × χk (Ô) (4.10)
h
Ô
———————————————————
1s(1)
Example : two-dimensional representation spanned by 1s atomic orbitals 1s(2)

centred on the H
atoms of a water molecule H2 O in the C2v group (see Figure 4.8)
O
H(1) H(2)
Figure 4.8: 1s atomic orbitals on the H atoms of H2 O.
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜ 1s(1) ⎟ ⎜ 1s(1) ⎟ ⎜ 1 0 ⎟⎜ 1s(1) ⎟
E⎝ ⎠ = ⎝ ⎠ = ⎝ ⎠⎝ ⎠ thus χE = 2
1s(2) 1s(2) 0 1 1s(2)
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜ 1s(1) ⎟ ⎜ 1s(2) ⎟ ⎜ 0 1 ⎟⎜ 1s(1) ⎟
C2z ⎝ ⎠ = ⎝ ⎠ = ⎝ ⎠⎝ ⎠ thus χC2z = 0
1s(2) 1s(1) 1 0 1s(2)
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜ 1s(1) ⎟ ⎜ 1s(2) ⎟ ⎜ 0 1 ⎟⎜ 1s(1) ⎟
σ xz ⎝ ⎠ = ⎝ ⎠ = ⎝ ⎠⎝ ⎠ thus χσxz = 0
1s(2) 1s(1) 1 0 1s(2)
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜ 1s(1) ⎟ ⎜ 1s(1) ⎟ ⎜ 1 0 ⎟⎜ 1s(1) ⎟
σ yz ⎝ ⎠ = ⎝ ⎠ = ⎝ ⎠⎝ ⎠ thus χσyz = 2
1s(2) 1s(2) 0 1 1s(2)

A1 1 1 1 1
A2 1 1 −1 −1
B1 1 −1 1 −1
B2 1 −1 −1 1
⎛ ⎞ ⎛ ⎞ ⎛ ⎞ ⎛ ⎞
⎜ 1 0 ⎟ ⎜ −1 0 ⎟ ⎜ 1 0 ⎟ ⎜ −1 0 ⎟
2 × 2 Matrix ⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎝ ⎠
0 1 0 −1 0 −1 0 1
Γ(1s) 2 0 0 2
Reduction of Γ(1s) = (2 0 0 2) :
(1s) 1
c A1 = (2 × 1 + 0 × 1 + 0 × 1 + 2 × 1) = 1
4
(1s) 1
c A2 = (2 × 1 + 0 × 1 + 0 × (−1) + 2 × (−1)) = 0
4
(1s) 1
c B1 = (2 × 1 + 0 × (−1) + 0 × 1 + 2 × (−1)) = 0
4
(1s) 1
c B2 = (2 × 1 + 0 × (−1) + 0 × (−1) + 2 × 1) = 1
4
⇒ Γ(1s) = A1 ⊕ B2 .
This means that one can therefore construct one linear combination of the two 1s(H) orbitals of H2 O
with A1 symmetry (totally symmetric) and one with B2 symmetry as will be shown in the following.
———————————————————
4.3 Useful applications of group theory
4.3.1 Determination of symmetrized linear combinations of atomic orbitals
To find the symmetrized linear combination of atomic orbitals (LCAO), one uses so-called
projectors P̂ . The projector associated with the irreducible representation Γ is defined by
1 (Γ)
P̂ Γ = χ (Ô) × Ô . (4.11)
h
Ô
The application of P̂ Γ onto the atomic orbitals provides a LCAO of symmetry Γ.

4.3. USEFUL APPLICATIONS OF GROUP THEORY 115
———————————————————
Example : symmetrized LCAOs of the two 1s orbitals on the H atoms of H2 O.
1
P̂ A1 1s(1) = [1 × E 1s(1) + 1 × C2z 1s(1) + 1 × σ xz 1s(1) + 1 × σ yz 1s(1)]
4
1 1
= [1s(1) + 1s(2) + 1s(2) + 1s(1)] = [1s(1) + 1s(2)]
4 2
1
P̂ A2 1s(1) = [1 × E 1s(1) + 1 × C2z 1s(1) − 1 × σ xz 1s(1) − 1 × σ yz 1s(1)]
4
1
= [1s(1) + 1s(2) − 1s(2) − 1s(1)] = 0
4
As expected, no A2 linear combination can be formed from the 1s(H) functions. Similarly one finds
P̂ B1 1s(1) = 0 , and
1
P̂ B2 1s(1) = [1s(1) − 1s(2)] .
2
The two LCAOs of symmetry A1 and B2 can be represented schematically in Figure 4.9.
O O
H(1) H(2) H(1) H(2)
A1 symmetry B2 symmetry
Figure 4.9: Linear combinations of atomic orbital 1s(H) of H2 O of symmetry A1 and B2 .
The symmetrized LCAOs can then be used to determine the chemical bonds that can be formed with
the p orbitals on the O atom. First, one must determine the transformation properties of the p orbitals
on the O atom depicted schematically in Figure 4.10.
z pz py px
O O O
x y H H H H H H
Figure 4.10: p orbitals on the O atom.
The px , py , and pz orbitals of the O atoms transform like x, y, and z, respectively as indicated in
Section 4.2.2. Hence: Γ(px ) = B1 , Γ(py ) = B2 , and Γ(pz ) = A1 as indicated in the third column of the
character table.
Only orbitals of the same symmetry can be combined to form bonding or antibonding molecular

orbitals. The five symmetrized orbitals listed above can be used to form five molecular orbitals
according to the following diagram (Figure 4.11) which does not take the 1s and 2s orbitals on the
oxygen into account because only valence electrons are considered for the formation of chemical bonds.
b*2
a*1
A1
H1s
B2
px py pz
O2p B1 A1
b1
B2
a1
b2
Figure 4.11: Valence molecular orbitals of H2 O built from symmetrized H(1s) “ligand” or-
bitals and the 2p atomic orbitals of O. The labels of the molecular orbitals refer to their
symmetry in lower case letters.
From the electronic configuration of each atom, there are six valence electrons (O ... (2p)4 , H (1s)1 )
to place in the Molecular Orbitals (MOs) following Pauli’s Aufbau-principle gives the ground state
configuration: ...(b2 )2 (a1 )2 (b1 )2 with an overall symmetry A1 . Because four of the six electrons are
in bonding orbitals and two in a non bonding px orbital, one expects two chemical bonds in H2 O.
The energetical ordering of the two bonding MO of B2 and A1 symmetry depends on the HOH angle
α defined in Figure 4.12. Whereas the a1 orbital becomes nonbonding at α = 180◦ , the b2 orbital
remains bonding at α = 180◦ but becomes antibonding at small angles.

Figure 4.12: Bond angle α.
———————————————————

4.3.2 Symmetry of normal modes
We consider the 3N -dimensional reducible representation Γ3N spanned by the set of 3N

Cartesian coordinates of the N atoms in a molecule and reduce it into irreducible repre-
sentations of the corresponding group. The molecule can also be characterized by its 3N
displacement coordinates i. e. the translations (t), rotations (r) and vibrations (v). There-
fore:
Γ3N = Γt ⊕ Γr ⊕ Γv (4.12)
The representation of the vibrational modes Γv can be deduced from Γ3N subtracting the
representations Γt and Γr as indicated in the character table.
———————————————————
Example: The vibrational modes of H2 O
The total representation is 3 × 3 = 9-dimensional. All irreducible representations of C2v are one-
dimensional, and only three vibrational modes (3N -6) exist in H2 O. The symmetry of these modes will
be obtained by eliminating the six irreducible representations corresponding to the three translational
and the three rotational degrees of freedom of the molecule.

Figure 4.13: Coordinates used to derive the Γ9 representation of H2 O in the C2v group.
In the basis set (or representation) Γ9 = {x1 , y1 , ..., z3 }, the C2v symmetry operations are represented
by 9 × 9 matrices.

The matrix representing the identity E is given by:

⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜ x1 ⎟ ⎜ x1 ⎟ ⎜ 1 0 0 0 0 0 0 0 0 ⎟⎜ x1 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ y1 ⎟ ⎜ y1 ⎟ ⎜ 0 1 0 0 0 0 0 0 0 ⎟⎜ y1 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ z1 ⎟ ⎜ z1 ⎟ ⎜ 0 0 1 0 0 0 0 0 0 ⎟ ⎜ z1 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ x2 ⎟ ⎜ x2 ⎟ ⎜ 0 0 0 1 0 0 0 0 0 ⎟⎜ x2 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎟
E ⎜ y2 ⎟ = ⎜ y2 ⎟
⎜ ⎜ ⎜
⎟=⎜ 0 0 0 0 1 0 0 0 0 ⎟
⎟⎜
⎜ y2 ⎟
⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ z2 ⎟ ⎜ z2 ⎟ ⎜ 0 0 0 0 0 1 0 0 0 ⎟ ⎜ z2 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ x3 ⎟ ⎜ x3 ⎟ ⎜ 0 0 0 0 0 0 1 0 0 ⎟⎜ x3 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ y3 ⎟ ⎜ y3 ⎟ ⎜ 0 0 0 0 0 0 0 1 0 ⎟ ⎜ y3 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎝ ⎠ ⎝ ⎠ ⎝ ⎠⎝ ⎠
z3 z3 0 0 0 0 0 0 0 0 1 z3
Hence the character is χ(Γ9 ) (E) = 9 (in agreement with the dimension of the representation).
The matrix representing the rotation C2z is given by:
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
⎜ x1 ⎟ ⎜ −x2 ⎟ ⎜ 0 0 0 −1 0 0 0 0 0 ⎟⎜ x1 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ y1 ⎟ ⎜ −y2 ⎟ ⎜ 0 0 0 0 −1 0 0 0 0 ⎟⎜ y1 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ z1 ⎟ ⎜ z2 ⎟ ⎜ 0 0 ⎟ ⎜ z1 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ 0 0 0 0 0 1 0 ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ x2 ⎟ ⎜ −x1 ⎟ ⎜ −1 0 0 0 0 0 0 0 0 ⎟⎜ x2 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
C 2 ⎜ y2 ⎟
z⎜
⎟=⎜
⎜ −y1 ⎟ ⎟=⎜ 0
⎜ −1 0 0 0 0 0 0 0 ⎟
⎟⎜
⎜ y2 ⎟
⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ z2 ⎟ ⎜ z1 ⎟ ⎜ 0 0 ⎟ ⎜ z2 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ 0 0 1 0 0 0 0 ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ x3 ⎟ ⎜ −x3 ⎟ ⎜ 0 0 0 0 0 0 -1 0 0 ⎟⎜ x3 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ y3 ⎟ ⎜ −y3 ⎟ ⎜ -1 0 ⎟ ⎜ y3 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ 0 0 0 0 0 0 0 ⎟⎜ ⎟
⎝ ⎠ ⎝ ⎠ ⎝ ⎠⎝ ⎠
z3 z3 0 0 0 0 0 0 0 0 1 z3
Hence the character is χ(Γ9 ) (C2 ) = −1.

From the construction of these two matrices we notice that only atoms that are not exchanged by
the symmetry operations can contribute to the character; therefore in the following, we consider the
atoms that are “invariant” by the operation Ô only to determine χ(Γ9 ) (Ô).
The reflection σ xz exchanges the two H atoms so that only the O atom needs to be considered:
σ xz x3 = x3 ; σ xz y3 = −y3 ; σ xz z3 = z3 ⇒ χ(Γ9 ) (σ xz ) = 1.
The reflection σ yz does not exchange any atom. For each atom, the x coordinate is inverted and the
y and z coordinates are preserved:
σ xz xi = −xi ; σ xz yi = yi ; σ xz zi = zi
⇒ χ(Γ9 ) (σ yz ) = −1 + 1 + 1 − 1 + 1 + 1 − 1 + 1 + 1 = 3

The reducible 9-dimensional representation is therefore:
A1 1 1 1 1 = Γz
A2 1 1 -1 -1 = ΓRz
B1 1 -1 1 -1 = Γx = ΓRy
B2 1 -1 -1 1 = Γy = ΓRx
Γ9 9 -1 1 3
The Γ9 representation can then be reduced using the reduction formula of Equation (4.10)):
1
c A1 = (9 − 1 + 1 + 3) = 3
4
1
c A2 = (9 − 1 − 1 − 3) = 1
4
1
cB1 = (9 + 1 + 1 − 3) = 2
4
1
cB2 = (9 + 1 − 1 + 3) = 3
4
Γ9 = 3A1 ⊕ A2 ⊕ 2B1 ⊕ 3B2 .
From these nine irreducible representations, three correspond to translations (Γx = B1 , Γy = B2 ,

Γz = A1 ) and three correspond to rotations (ΓRx = B2 , ΓRy = B1 , ΓRz = A2 ). The remaining three,
namely 2A1 ⊕ B2 , correspond to the three vibrational modes of H2 O (3N − 6 = 3, because H2 O is a
nonlinear molecule). To determine these modes one can use the projection formula of Equation (4.11).
Let us consider the vibrational mode of symmetry B2 as an example. In practice it is convenient to
first treat the x, y and z displacements separately and then to combine the x, y, and z motions.
For the x-dimension:
1
P̂ B2 x1 = (1Ex1 − 1C2z x1 − 1σ xz x1 + 1σ yz x1 )
4
1
= (x1 + x2 − x2 − x1 ) = 0.
4
The B2 mode does not involve x-coordinates.
For the y- and z-dimensions:
1
P̂ B2 y1 = (1Ey1 − 1C2z y1 − 1σ xz y1 + 1σ yz y1 )
4
1 1
= (y1 + y2 + y2 + y1 ) = (y1 + y2 ) .
4 2
1
P̂ B2 z1 = (1Ez1 − 1C2z z1 − 1σ xz z1 + 1σ yz z1 )
4
1 1
= (z1 − z2 − z2 + z1 ) = (z1 − z2 ) .
4 2
The B2 mode involves both y and z coordinates. Drawing the displacement vectors one obtains a
vectorial representation of the motion of the H atoms in the B2 mode. The motion of the O atom

can be estimated in a same way or reconstructed by ensuring that the center of mass of the molecule
remains stationary.

Figure 4.14: Determination of the nuclear motion of the B2 mode of water.
The mode can be easily identified as the asymmetric stretching mode.
———————————————————
4.3.3 Symmetry of vibrational levels
The nomenclature to label the vibrational states of a polyatomic molecule is
v
ν1v1 , ν2v2 , · · · , ν3N
3N −6
−6 (4.13)
where νi designate the mode and vi the corresponding vibrational quantum number. Usually
only the modes νi for which vi = 0 are indicated. The notation
(v1 , v2 , · · · , v3N −6 ) (4.14)
is also often used. For the ordering of the modes, the totally symmetric modes come first
in order of descending frequency, then the modes corresponding to the second irreducible
representation in the character table in order of descending frequency, etc.
To find the overall symmetry of the vibrational wavefunction one must build the direct
product
Γvib = (Γν1 )v1 ⊗ (Γν2 )v2 ⊗ · · · ⊗ (Γν3N −6 )v3N −6 . (4.15)
———————————————————
Example: The three vibrational modes of H2 O ν1 is the O-H symmetric stretching mode (ν̃1 =3585 cm−1 )
of symmetry A1 , ν2 is the H-O-H bending mode (ν̃2 =1885 cm−1 ) of symmetry A1 and ν3 is the O-H
asymmetric stretching mode (ν̃3 =3506 cm−1 ).
We consider the state with v1 = 2, v2 = 1, v3 = 3. In the first notation, this will correspond to :

12 21 33 . In the second notation, it will correspond to (2 1 3). The symmetry of this vibrational state
is:
Γvib (H2 O, 12 21 33 ) = A1 ⊗ A1 ⊗ A1 ⊗ B2 ⊗ B2 ⊗ B2 = B2 .

ν1 ν2 ν3
———————————————————
4.3.4 Symmetry of electronic states and labels of configurations
Just as in the case of vibrational wave functions, the overall symmetry of an electronic
wavefunction is obtained from the direct product
Γel = (Γ1 )n1 ⊗ (Γ2 )n2 ⊗ · · · ⊗ (Γm )nm , (4.16)
where Γi is the irreducible representation of orbital i and ni is the occupation number of

orbital i in the considered configuration. As totally filled subshells are always totally sym-
metric they do not influence the overall symmetry and can be omitted in equation 4.16.
———————————————————
Example: Electronic ground state configuration of H2 O and H2 O+ :
... (b2 )2 (a1 )2 (b1 )2 (see Figure 4.11)
H2 O:
A1
Γel = B2 ⊗ B2 ⊗ A1 ⊗ A1 ⊗ B1 ⊗ B1 = A1

(b2 )2 (a1 )2 (b1 )2
Therefore the electronic ground state is labelled X̃ 1 A1
H2 O+ :
... (b2 )2 (a1 )2 (b1 )1
A1
Γel = B2 ⊗ B2 ⊗ A1 ⊗ A1 ⊗ B1 = B1

(b2 )2 (a1 )2 (b1 )1
+2
Therefore the electronic ground is labelled X̃ B1
———————————————————
Example: Electronic ground state configuration of the borane molecule BH3 in the D3h point group.

Figure 4.15: BH3 molecule with its coordinate system.

We follow the same procedure as for H2 O in section 4.3.1, retain the 2s and 2p orbitals on the B atom
and the 1s orbitals on the H atoms. First, symmetrized “ligand” orbitals are constructed from the H
1s orbitals; then these are combined with the orbitals of the B atom to form bonding and antibonding
orbitals. For the ligand orbitals, a 3D-representation is spanned by the three 1s atomic orbitals on
the H atom.
D3h E 2C3 3C2 σh 2S3 3σv
A1 1 1 1 1 1 1
A2 1 1 −1 1 1 −1 Rz
E 2 −1 0 2 −1 0 x, y
A1 1 1 1 −1 −1 −1
A2 1 1 −1 −1 −1 1 z
E 2 −1 0 −2 1 0 Rx , R y
Γ3D 3 0 1 3 0 1
This representation can be reduced using the reduction formula of Equation (4.10):
1
cA1 = [3 × 1 × 1 + 0 × 2 × 1 + 1 × 3 × 1 + 3 × 1 × 1 + 0 × 1 × 1 + 1 × 3 × 1] = 1
12
1
cA2 = [3 × 1 × 1 + 0 + 1 × 3 × (−1) + 3 × 1 × 1 + 0 + 1 × 3 × (−1)] = 0
12
1
cE = [3 × 1 × 2 + 0 + 1 × 3 × 0 + 3 × 1 × 2 + 0 + 1 × 3 × 0] = 1
12
Γ3D = A1 ⊕ E .
Figure 4.16 shows the ligand orbital of A1 symmetry found by intuition. The ligand orbitals of E
symmetry are found by using the projection formula of Equation (4.11):
1
P̂ E 1s(1) = 2 × E 1s(1) − 1 × C3 1s(1) − 1 × C32 1s(1) + 2 × σh 1s(1) − 1 × S3 1s(1) − 1 × S32 1s(1)
12
1
= [2 × 1s(1) − 1s(2) − 1s(3) + 2 × 1s(1) − 1s(2) − 1s(3)]
12
1 1
= 1s(1) − [1s(2) + 1s(3)] .
3 2

To find the second orbital of E symmetry, we can the projector P̂ E to the the 1s(2) and 1s(3)
1 1 1 1
orbitals; we find two further molecular orbitals 3 [1s(2) - 2 [1s(1) + 1s(3)]] and 3 [1s(3) - 2 [1s(1) +
1s(2)]]. The three orbitals are linearly dependent. One can use linear algebra to eliminate one of these
three orbitals and to find an orthogonal set of two orbitals of E symmetry (see Figure 4.16) using the
Gram-Schmidt orthogonalization procedure.

H(1) H(1) H(1)
B B B
H(2) H(3) H(2) H(3) H(2) H(3)
A1’ symmetry E’ symmetry
Figure 4.16: Ligand orbitals of BH3 .
The molecular orbitals are finally found by determining the symmetry of the 2s and 2p orbitals of
the central B atom and combining the orbitals of the same symmetry into bonding and antibonding
orbitals (see Figure 4.17).
2s(C): A1 ↔ A1 ligand orbital
px (C), py (C): E ↔ E ligand orbital
2pz (C): A2
The 2pz orbital of A2 symmetry must remain nonbonding because there are no ligand orbitals of A2
symmetry.
2e'
3a1'
1a2''
pz 2px,y
A2'' E' A1' E'
1e'
2s
A1' 2a1'
Figure 4.17: Valence molecular orbitals of BH3 .
The electronic configuration of BH3 (in total eight electrons) is therefore:
(1a1 )2 (2a1 )2 (1e )4 (1a2 )0 ,

A1
where the 1a1 orbital is the 1s orbital on the B atom. Therefore, the ground state is X̃ 1 A1 .
———————————————————

Example: The case of a linear molecule.

As all homonuclear diatomic molecules, O2 belongs to the point group D∞h . The character table and
direct product tables of D∞h are given in Tables 4.3 and 4.4, respectively.
ϕ ϕ
D∞h E 2C∞ . . . ∞σv i 2S∞ . . . ∞C2
Σ+
g 1 1 ... 1 1 1 ... 1 x2 + y 2 , z 2
Σ−
g 1 1 ... −1 1 1 ... −1 Rz
Πg 2 2 cos ϕ ... 0 2 −2 cos ϕ ... 0 Rx , Ry xz, yz
Δg 2 2 cos 2ϕ . . . 0 2 2 cos 2ϕ ... 0 x2 − y 2 , xy
... ... ... ... ... ... ... ... ...
Σ+
u 1 1 ... 1 −1 −1 ... −1 z
Σ−
u 1 1 ... −1 −1 −1 ... 1
Πu 2 2 cos ϕ ... 0 −2 2 cos ϕ ... 0 x, y
Δu 2 2 cos 2ϕ . . . 0 −2 −2 cos 2ϕ . . . 0
... ... ... ... ... ... ... ... ...
Table 4.3: Character table of the D∞h point group.
⊗ Σ+ Σ − Π Δ Φ ...
Σ+ Σ+ Σ − Π Δ Φ ...
Σ− Σ+ Π Δ Φ ...
Π Σ+ ⊕ Σ− ⊕ Δ Π⊕Φ Δ ⊕ Γ ⊕ ...
Δ Σ+ ⊕ Σ− ⊕ Γ Π⊕H ...
Φ Σ+ ⊕ Σ − ⊕ I . . .
... ... ... ... ... ... ...
Table 4.4: Direct product table of the point groups C∞v and D∞h . For D∞h , the “gerade” (g)
or “ungerade” (u) character is determined as follows: g ⊗ g = u ⊗ u = g and g ⊗ u = u ⊗ g = u.

Referring back to Figure 3.10 we can derive the molecular orbitals of O2 and determine the most
stable electronic configuration to be
(1σg )2 (1σu∗ )2 (2σg )2 (2σu∗ )2 (3σg )2 (1πu )4 (1πg∗ )2 (3σu∗ )0 .

Σ+
g
All fully occupied orbitals contribute the totally symmetric representation Σ+

g to the electronic sym-
metry. The irreducible representations of the electronic states resulting from the above configuration
can be determined from the direct product

Πg ⊗ Πg = Σg ⊕ Σ−
g ⊕ Δg .
Σ−
g appears in square brackets because it is the anti-symmetric part of the direct product, whereas
Σg ⊕ Δg is the symmetric part. Since the total electron wave function must be antisymmetric under
exchange of two electrons, the anti-symmetric spatial part combines with the symmetric spin part
giving rise to the 3 Σ−

g state, whereas the symmetric spatial part combines with the anti-symmetric
spin part giving 1 Σ+ 1

g and Δg states.
———————————————————
4.3.5 Generalized Pauli principle and allowed states
The simplistic expression of the Pauli principle states that two electrons can not occupy the
same spin-orbital. In Section 2.1, we have seen that the wavefunction describing fermions
(particle with half integer spin) must be antisymmetric with respect to the permutation of
two particles, while the wavefunction describing bosons (particle with integer spin) must be
symmetric.
The generalized form of the Pauli principle states that the total wavefunction describing a
molecular system must transform under the group operations Ôj as the irreducible represen-
tation whose characters χirr (Ôj ) are given by
NF

χ (Ôj ) =
irr
(−1)Pi (Ôj ) (4.17)
i
where NF is the number of types of identical fermions in the system. Pi (Ôj ) is the so-called
parity of the permutation of the i-th kind of fermions.
If the operation Ôj applies on bosons, then χirr (Ôj ) = +1.
If the operation Ôj applies on fermions, the parity is “even” (respectively “odd”) if Ôj can
be written as an even (resp. odd) number of permutations (n m). Therefore, χirr (Ôj ) = +1
if the permutation of fermions has an even parity and χirr (Ôj ) = −1 if the permutation of

fermions has an odd parity.

———————————————————
Example: the two fermions H+ in H2 O
The (1 2) permutation has an odd parity. Hence χirr (1 2) = −1. In the C2v group, the operation that
corresponds to the permutation (1 2) is C2z . Two irreducible representations exhibit χirr (C2z ) = −1:
B1 and B2 .
The total wavefunction of water must transform as B1 or B2 . In the Born-Oppenheimer approxima-
tion, the total wavefunction can be written in the product form Ψtot = Φel Φvib Φrot Φns Φes and its
symmetry can be determined by the direct product
Γtot = Γel ⊗ Γvib ⊗ Γrot ⊗ Γns ⊗ Γes .
Since Γel = A1 and Γvib = A1 in the vibrational and electronic ground state, it imposes Γrot ⊗ Γns =
B1/2 . This shows that not all combinations of rotational levels and nuclear spins are allowed.
———————————————————
Example: CO2 in the D∞h
In that case, all operations Ôj apply on bosons only, hence χirr (Ôj ) = +1. The total wavefunction
Φel Φvib Φrot should transform as Σ+

g.
The vibronic (vibrational-electronic) ground state: Γel =Σ+ +

g and Γvib =Σg .
The rotational wavefunction must transform as Σ+

g.
It can be shown that for the CO2 molecule in a rigid rotor approximation, the even values of J are
associated with rotational wavefunctions of Σ+

g symmetry, while the odd values of J are associated
with rotational wavefunctions of Σ+

u symmetry. Therefore only half of the rotational levels, with even
values of J (J = 0, 2, 4...) are allowed in the electronic vibrational ground state.
The antisymmetric CO stretching ν3 vibrational state: by applying the protocol described in Sec-
tion 4.3.2, one can find that Γν3 =Σ+

u . In that case, the rotational wavefunction must transform as
Σ+
u . Therefore only half of the rotational levels, with odd values of J (J = 1, 3, 5...) are allowed in the
ν3 vibrational state.
Rovibrational transitions (0 0 1)←(0 0 0) fulfil ΔJ = ±1, which in the spectrum gives rise to a P and
an R branches. But every other line is absent compared to the spectrum of CO shown in Figure 3.11
because of the missing states.
It can be shown that for an electronic transition to an electronic state of Π symmetry, this does not
hold and all J values are allowed.
———————————————————

4.3.6 Selection rules with group theory
In the dipole approximation, the interaction between molecules and electromagnetic radiation
is assumed to only come from the interaction
V̂ = −μ
ˆlab · E. (4.18)
μ
lab is used here to distinguish the dipole moment of the molecule in the laboratory-fixed
frame from μ
, which is the dipole moment in the molecule-fixed frame as illustrated in Fig-
ure 4.18 in the case of the CH3 Cl molecule.

Figure 4.18: Relationship between the expression of the dipole moment μ

= (μx , μy , μz ), ex-
pressed in the molecule-fixed reference frame (x, y, z), and that μ
lab = (μX , μY , μZ ) expressed
in the space-fixed (X, Y, Z) reference frame for CH3 Cl; the permanent dipole-moment vector
μ
lies along the z axis of the molecule-fixed reference frame.
of the radiation is defined in the laboratory-fixed frame (X, Y, Z),

The polarization vector E
whereas the components of μ
are defined in the molecule-fixed frame (x, y, z) as follows:
N

μξ = qi ξi
i=1
with ξ = x, y, z and qi is the charge of particle i. The space-fixed components μX , μX , and

μZ of μ
lab vary as the molecule rotates while the molecule-fixed components μx , μy , and μz
remain the same.

= (0, 0, E) ,
For a linearly polarized light E
V̂ = −μ̂Z E. (4.19)
As the intensity of an electric dipole transition between an initial state Ψi and a final state
Ψf is proportional to
Ψf |V̂ |Ψi , the selection rule can be written as follows:

Ψf |μ̂Z |Ψi =
0. (4.20)
The wavefunctions Ψi and Ψf are expressed in the molecule-fixed frame (x, y, z); for simplicity,
the selection rules are usually expressed in terms of μx , μy , and μz instead of μX , μY , and
μZ . Therefore one needs to express μX , μY , and μZ as functions of μx , μy , and μz i. e. the
transformation from the space-fixed frame to the molecule-fixed frame.
The relative orientation of the space-fixed and molecule-fixed coordinate systems is given by
the three Euler angles (ϕ, θ, χ) defined by three successive rotations depicted in Figure 4.19:
1. the rotation around Z by ϕ which generates the coordinate system (x , y , z ) in grey in
Figure 4.19
2. the rotation around y by θ which generates the coordinate system (x , y , z ) in red in
Figure 4.19
3. the rotation around z by χ which generates the coordinate system (x, y, z) in black in
Figure 4.19.
The transformation 1 can be written as follows:

⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞

⎜ x ⎟ ⎜ X ⎟ ⎜ cos ϕ sin ϕ 0 ⎟⎜ X ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ y ⎟ = RZ (ϕ) ⎜ Y ⎟ = ⎜ −sin ϕ cos ϕ 0 ⎟⎜ Y ⎟ (4.21)
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎝ ⎠ ⎝ ⎠ ⎝ ⎠⎝ ⎠
z Z 0 0 1 Z
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
x
⎜ x ⎟ ⎜ x ⎟ ⎜ cos θ 0 −sin θ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ y ⎟ = Ry (θ) ⎜ y ⎟ = ⎜ 0 1 0 ⎟⎜ y ⎟ (4.22)
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎝ ⎠ ⎝ ⎠ ⎝ ⎠⎝ ⎠
z z sin θ 0 cos θ z
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎛ ⎞
x
⎜ x ⎟ ⎜ x ⎟ ⎜ cos χ sin χ 0 ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ y ⎟ = Rz (χ) ⎜ y ⎟ = ⎜ −sin χ cos χ 0 ⎟⎜ y ⎟ (4.23)
⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎜ ⎟
⎝ ⎠ ⎝ ⎠ ⎝ ⎠⎝ ⎠
z z 0 0 1 z

Figure 4.19: Euler angles θ, φ, χ defining the relative orientation of the space-fixed reference
frame (X, Y, Z) in blue and the molecule-fixed reference frame (x, y, z) in black. Starting from
the space-fixed reference frame, the molecule-fixed reference frame is obtained by 1) rotation
by an angle ϕ around the Z axis, leading to the intermediate (x , y , z in grey) frame; 2)
rotation by an angle θ around the y axis, leading to the second intermediate (x , y , z in
red) frame; and 3) rotation by an angle χ around the z axis.
Using these three equations, the laboratory- and molecule-fixed frames can be linked by the
following transformation:
⎡ ⎤ ⎡ ⎤
x X
⎢ ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ⎥
⎢ y ⎥ Rz (χ)Ry (θ)RZ (ϕ) ⎢ ⎥
⎢ ⎥ = ⎢ Y ⎥ (4.24)
⎣ ⎦ ⎣ ⎦
z Z
⎡ ⎤⎡ ⎤
cos ϕ cos θ cos χ − sin ϕ sin χ sin ϕ cos θ cos χ + cos ϕ sin χ −sin θ cos χ X
⎢ ⎥⎢ ⎥
⎢ ⎥⎢ ⎥
= ⎢ −cos ϕ cos θ sin χ − sin ϕ cos χ −sin ϕ cos θ sin χ + cos ϕ cos χ sin θ sin χ ⎥ ⎢ Y ⎥
⎥ ⎢
⎢ ⎥
⎣ ⎦⎣ ⎦
cos ϕ sin θ sin ϕ sin θ cos θ Z

or its inverse
⎡ ⎤ ⎡ ⎤⎡ ⎤
X cos ϕ cos θ cos χ − sin ϕ sin χ −cos ϕ cos θ sin χ − sin ϕ cos χ cos ϕ sin θ x
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎢ Y ⎥ = ⎢ sin ϕ cos θ cos χ + cos ϕ sin χ −sin ϕ cos θ sin χ + cos ϕ cos χ ⎥
sin ϕ sin θ ⎥ ⎢ y ⎥
⎢
⎢ ⎥ ⎢ ⎥
⎣ ⎦ ⎣ ⎦⎣ ⎦
Z −sin θ cos χ sin θ sin χ cos θ z
⎡ ⎤
x
⎢ ⎥
⎢ ⎥
= λ⎢ y ⎥
⎢
⎥, (4.25)
⎣ ⎦
z
where λ is the direction cosine matrix. λ can thus be used to express the components of a vector in
the laboratory-fixed frame as a function of the components of the same vector in the molecule-fixed
frame, and especially μZ as a function of μx , μy , and μz :
μZ = λZx μx + λZy μy + λZz μz . (4.26)
In the Born-Oppenheimer approximation, the molecular wavefunctions Ψf and Ψi are expressed in the
product form
Ψf = φel φvib φrot φnspin φespin , and (4.27)
Ψi = φel φvib φrot φnspin φespin . (4.28)
The transition moment

Ψf |μ̂Z |Ψi can now be written as follows:

φel φvib φrot φnspin φespin | λZα μ̂α |φel φvib φrot φnspin φespin . (4.29)
α
Equations (4.20) and (4.29) lead to the selection rules for an electric dipole transition.
Spin conservation upon electric dipole transition
The φnspin functions depend on the nuclear spin variables only, and the φespin functions depend on
the electron spin variables only. Their integration in Equation (4.29) can thus be separated:

φespin |φespin
φnspin |φnspin
φel φvib φrot | λZα μ̂α |φel φvib φrot . (4.30)
α
Because electron- and nuclear-spin functions are orthogonal, Equation (4.30) vanishes (transition
forbidden) unless φespin = φespin and φnspin = φnspin , which bring the selection rules:
• ΔS = 0 interdiction of intercombination
• ΔI = 0 nuclear-spin conservation rule

Angular momentum selection rules
The remaining rovibronic (rotational-vibrational-electronic) transition moment in Equation (4.30)

φel φvib φrot | α λZα μα |φel φvib φrot can be further simplified.
A rotation of the molecule in space leads to a change of the Euler angles, and the rotational wave-
functions φrot (ϕ, θ, χ) are expressed in the space-fixed frame as a function of these angles, while the
functions φel , φvib and μα do not depend on the Euler angles: φel (qi , Q), φvib (Q), μα (qi , Q). The
direction cosine elements λZα and φrot only depend on ϕ, θ, χ and the integration can be further
separated in an integral over angular variables and an integral over electronic coordinates and normal
modes:

φrot |λZα |φrot
φel φvib |μ̂α |φel φvib . (4.31)
α
The integral
φrot |λZα |φrot leads to angular momentum selection rules:
• ΔJ = 0, ±1; 0 ↔ 0 angular momentum conservation (see also Chapter 2)
The projection quantum number M of J on the Z axis leads to further selection rules:
• if the polarization is along the Z axis, then ΔM = 0
• if the polarization is along the X or Y axis, then ΔM = ±1
Finally,
• if the dipole moment lies on the z axis, the transition is said parallel and ΔΛ = 0 for diatomic
molecules
• if the dipole moment lies on the x or y axis, the transition is said perpendicular and ΔΛ = ±1
for diatomic molecules
Further selection rules
The integral
φel φvib |μ̂α |φel φvib of Equation (4.31) represents a selection rule for transitions between
electronic and vibrational levels. Depending of the type of transitions investigated, its evaluation can
be simplified using the vanishing integral theorem and group theory: this theorem states that the
product
Ψ2 |Ô|Ψ1 vanishes if the product of the symmetry representation of Ψ2∗ , Ô and Ψ1 does not
(sym)
contain the totally symmetric irreducible representation Γirr of the group, i. e.
(sym)
0 ⇒ Γ(Ψ∗2 ) ⊗ Γ(Ô) ⊗ Γ(Ψ1 ) ⊃ Γirr

Ψ2 |Ô|Ψ1 = (4.32)
This theorem will be used in three cases in the following.
Rotational spectroscopy: φel = φel , φvib = φvib

The integral
φel φvib |μ̂α |φel φvib of Equation (4.31) represents the expectation value of μ̂α . Tran-
sitions are only allowed for molecules with a permanent dipole moment. The angular momentum
selection rules are as above.

Vibrational spectroscopy: φel = φel , φvib = φvib

Since, in the BO approximation, only φel and μα depend on electron coordinates qi , the integration
over qi can be done first:

φvib (Q)|
φel (Q, qi )|μ̂α (Q, qi )|φel (Q, qi )qi |φvib (Q)Q =
φvib (Q)|μ̂el,α (Q)|φvib (Q). (4.33)

μ̂el,α (Q)
Whether Equation (4.33) vanishes or not, can be determined using the vanishing integral theorem.
Therefore the transition is allowed if
(sym)
Γvib ⊗ Γα ⊗ Γvib ⊃ Γirr , (4.34)
where Γα (α = x, y, z) transforms as the components of a vector and thus as α.
———————————————————
Example: the case of H2 O
The vibrational ground state φvib = (0,0,0) has the symmetry Γvib = A1
The components of the dipole moment have the symmetries Γμx = B1 , Γμy = B2 , and Γμz = A1 .
The vibrational state φvib = (1,0,0) has the symmetry Γvib = A1 .

⎧ ⎫ ⎧ ⎫
⎪
⎪ ⎪
⎪ ⎪
⎪ ⎪
⎪ B 1 ⎪ ⎪ B 1 ⎪
⎪
⎪
⎨ ⎪
⎬ ⎪
⎨ ⎪
⎬

Γvib ⊗ Γα ⊗ Γvib = A1 ⊗ B 2 ⊗ A 1 = B 2
⎪
⎪ ⎪
⎪ ⎪
⎪ ⎪
⎪
⎪
⎪ ⎪ ⎪ ⎪
⎩ A ⎪ ⎭ ⎩ A ⎪
⎪ ⎭
1 1
The allowed vibrational transition originates from the z component of the transition dipole moment.
The vibrational φvib = (0,0,1) has the symmetry Γvib = B2 .

⎧ ⎫ ⎧ ⎫
⎪
⎪ ⎪
⎪ ⎪
⎪ ⎪
⎪ B 1 ⎪ ⎪ A 2 ⎪⎪
⎪
⎨ ⎪
⎬ ⎪
⎨ ⎪
⎬
B2 ⊗ B 2 ⊗ A 1 = A 1
⎪
⎪ ⎪
⎪ ⎪
⎪ ⎪
⎪
⎪
⎪ ⎪ ⎪ ⎪
⎩ A ⎪ ⎭ ⎩ B ⎪
⎪ ⎭
1 2
The allowed vibrational transition originates here from the y component of the transition dipole
moment. The component of the permanent dipole moment along the y axis is zero in H2 O. Nevertheless
a transition can be observed. The condition for a vibrational transition to be observable is that a
change of the dipole moment must occur when exciting the vibration. This is obviously the case when
the antisymmetric stretching mode is excited in H2 O.
One can show that transitions to all vibrational levels are allowed by symmetry in H2 O. However,
overtones are weaker than fundamental excitations as discussed in the case of diatomic molecules in
Section 3.5.1.
———————————————————

Electronic spectroscopy: φel = φel

The left hand side of Equation (4.33) can be written as

φvib (Q)|
φel (Q, qi )|μ̂α (Q, qi )|φel (Q, qi )qi |φvib (Q)Q , (4.35)

μfi
el,α
where the inner integral represents an electronic transition moment μfiel,α obtained by inte-
gration over the electronic coordinates.
A transition is electronically allowed when μfiel,α = 0, which is fulfilled if
Γel ⊗ Γα ⊗ Γel ⊃ Γirr

(sym)
. (4.36)
———————————————————
Example: the case of H2 O
The electronic ground state is X̃ 1 A1

⎧ ⎫ ⎧ ⎫ ⎧ ⎫
⎪
⎪ B ⎪
⎪ ⎪
⎪ B ⎪
⎪ ⎪
⎪ ⎪
⎪
⎪
⎪ 1 ⎪
⎪ ⎪ 1 ⎪ ⎪ A1 ⎪
⎨ ⎬ ⎪ ⎨ ⎪
⎬ ⎪
⎨ ⎪
⎬
Γel ⊗ Γα ⊗ Γel = B 2 ⎪ ⊗ ⎪ B 2 ⎪ ⊗ A 1 = ⎪ A1
⎪
⎪ ⎪ ⎪
⎪
⎪ ⎪ ⎪ ⎪ ⎪
⎪ ⎪
⎪ ⎪
⎪
⎩ A ⎪ ⎭ ⎪ ⎩ A ⎪ ⎭ ⎪
⎩ A ⎪
⎭
1 1 1
Electric dipole transitions to electronic states of B1 , B2 and A1 symmetry are electronically allowed
and transitions to electronic states of A2 symmetry are electronically forbidden in H2 O.
———————————————————
If one assumes that the transition moment function μfiel,α (Q) varies slowly with Q, then
μfiel,α (Q) can be described by a Taylor series and one can in good approximation neglect
higher terms: & '
$ %
3N −6
∂μfiel,α
μfiel,α (Q) = μfiel,α + Qj + ... (4.37)
eq ∂Qj
j eq
In electronically allowed transitions the first term in Equation (4.37) is often the dominant
one and the transition moment (Equation (4.35)) becomes:
$ %

φvib (Q)|φvib (Q) μfiel,α . (4.38)
eq
The intensity of a transition is proportional to the square of the transition moment and thus,
( (2
I ∝ (
φvib (Q)|φvib (Q)( . (4.39)
|
φvib (Q)|φvib (Q)|2 is called a Franck-Condon factor and represents the square of the
overlap of the vibrational wavefunctions. Equation (4.39) implies the vibrational selection
rule for electronically allowed transitions:
Γvib ⊗ Γvib ⊃ Γirr

(sym)
. (4.40)

———————————————————
Example: H2 O X̃ 1 A1 (0, 0, 0) → H2 O C̃ 1 B1
!
Γvib = A1 ⇒ Γvib = A1
Only the symmetric stretching mode ν1 and the bending mode ν2 can be excited. The asymmetric
stretching mode ν3 of B2 symmetry can only be excited if v3 is even.
———————————————————
Electronically forbidden transitions can become weakly allowed if the electronic and vibra-
tional degrees of freedom cannot be separated as in Equation (4.35). The condition for them
to be weakly observable is that
Γ ⊗ Γ ⊗ Γα ⊗ Γvib ⊗ Γel ⊃ A1 . (4.41)

vib el
Γev Γ
ev
———————————————————
Example:
Transitions from H2 O X̃ 1 A1 (0,0,0) to electronically excited states of A2 symmetry (electronically
forbidden) may become weakly allowed (vibronically allowed) if a non totally symmetric mode is ex-
cited.
X̃ 1 A1 (0,0,0): Γel = A1 , Γvib = A1 , Γev = A1 ⊗ A1 = A1
A2 (0,0,1): Γel = A2 , Γvib = B2 , Γev = A2 ⊗ B2 = B1
⎧ ⎫ ⎧ ⎫
⎪
⎪ ⎪
⎪ ⎪
⎪ ⎪
⎪
⎪
⎪
B 1 ⎪
⎪ ⎪
⎪
A1 ⎪
⎪
⎨ ⎬ ⎨ ⎬
Γev ⊗ Γα ⊗ Γev = B1 ⊗ B2 ⎪ ⊗ A1 = ⎪ A2
⎪
⎪ ⎪ ⎪ ⎪
⎪
⎪
⎪ ⎪ ⎪ ⎪
⎩ A ⎪ ⎭ ⎪
⎩ B ⎪
⎭
1 1
The vibronically allowed transition originates from μx .
———————————————————

PCV Ch4

Uploaded by

Copyright:

Available Formats

PCV Ch4

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PCV Ch4

Uploaded by

Copyright:

Available Formats

Chapter 4

4.1 Symmetry operations

4.1.1 Deﬁnition of a group

A group G is a set of elements A, B, C, ... connected by a combination rule (written as a

4.1.2 Point group operations and point group symmetry

• the identity E which leaves all coordinates unchanged.

PCV - Spectroscopy of atoms and molecules

• the inversion i of all coordinates through the inversion center.

• the improper rotation Sn or rotation-reﬂection which consists in a rotation by an

(1) (1) (3)

A molecule having an improper operation as symmetry operation, i. e. a reﬂection, an im-

4.1.3 Permutation-inversion operations and CNPI groups

F(2) H(2) F(1) F(2) H(1) F(1)

Figure 4.3: Tunneling process in (HF)2 .

PCV - Spectroscopy of atoms and molecules

H(1) (12) H(2)

Figure 4.4: Examples of (i j) and (i j k) permutations.

• all possible circular permutations of n identical nuclei (for example, the (1 2 3 4 5 6)

• the permutation followed by an inversion (ij)∗ = E ∗ · (ij) of all coordinates of all

• the cyclic permutation followed by an inversion (ijk)∗ of all coordinates of all

• all possible circular permutations followed by an inversion of all coordinates of n iden-

PCV - Spectroscopy of atoms and molecules

4.2 Important concepts in a group

4.2.1 Order, conjugated elements and classes

PCV - Spectroscopy of atoms and molecules

Example: the C2v group of H2 O

group operation. For instance: σ xz · σ yz = C2z , σ xz · σ xz = E, C2z · σ xz = σ yz , etc.

thus a 4×4 table.

1st operation (right)

Table 4.1: Multiplication table of the C2v point group.

Ôi · Ôj = Ôj · Ôi . (4.2)

PCV - Spectroscopy of atoms and molecules

PCV - Spectroscopy of atoms and molecules

1st operation (right)

Table 4.2: Multiplication table of the C3v point group.

Ôk · Ôi · Ôk−1 = Ôj , (4.3)

C3 E E E · C3 · E = C3 C32 E E E · C32 · E = C32

C3 C3 C32 C3 · C3 · C32 = C3 C32 C3 C32 C3 · C32 · C32 = C32

C3 C32 C3 C32 · C3 · C3 = C3 C32 C32 C3 C32 · C32 · C3 = C32

C3 σa σa σ a · C3 · σ a = C32 C32 σa σa σ a · C32 · σ a = C3

Similarly, one can show that σ a , σ b and σ c form a class of order 3.

PCV - Spectroscopy of atoms and molecules

exists for ﬁnite groups only.

4.2.2 Representations and character table

y = B (ẽ) x. (4.4)

C2z x = −x C2z y = −y C2z z = z

PCV - Spectroscopy of atoms and molecules

Therefore x, y, and z correspond to the following representations designated by Γ:

C2z x2 = (C2z x)(C2z x) = x2 C2z xy = (C2z x)(C2z y) = xy

σ xz x2 = (σ xz x)(σ xz x) = x2 σ xz xy = (σ xz x)(σ xz y) = −xy

σ yz x2 = (σ yz x)(σ yz x) = x2 σ yz xy = (σ yz x)(σ yz y) = −xy

σ xz Rz = −Rz Direction of rotation reversed.

σ yz Rz = −Rz Direction of rotation reversed.

PCV - Spectroscopy of atoms and molecules

Representations of higher dimensionality can be obtained by looking at the transformation

Example: two dimensional representation y of the C2v group

The two-dimensional representation of y has thus the following characters:

PCV - Spectroscopy of atoms and molecules

x, y, and z correspond to the irreducible representations B1 , B2 , and A1 , respectively as indicated in