Adrian Mulholland

Molecular Structure and Bonding (II)

Lecture 1 Synopsis:

Ref: P.W. Atkins, Physical Chemistry, 6th edn., pages 406-410.

The Variational Principle, The Variational Method and The Linear

Combination of Atomic Orbitals (LCAO) Approximation

Learning Objectives:

· To Learn the Variation principle.

· To understand how it leads to the Variational method.

· To understand the linear combinations of atomic orbitals

(LCAO) approximation in which a molecular orbital is
approximated as a linear combination of atomic orbitals.

· To understand how the Variational method is used to determine

the coefficients of the atomic orbitals in the LCAO
approximation.

· To understand how secular equations and secular determinants

arise in the application of the Variational method and to
understand in detail the procedure for solving secular equations.

The Variation Principle:

1
 If an arbitrary or approximate wavefunction is used to
calculate the energy of a system then the value obtained is always
greater than or equal to the exact ground state energy of the
system.

This principle can be stated mathematically as:

ψ Hˆ ψ dv
 *
*
Eapprox  Eexact (1)
 ψ ψ dv

Another way of saying this is

The energy of an approximate wavefunction is never less than the

true energy

The Variation principle is the key to finding wavefunctions of

molecules.

Why calculate wavefunctions?

To understand and predict the structure and reactions of

molecules, we need to know about how electrons are arranged in
them. To study electrons, we need to use quantum mechanics. In
principle, all the information we need about a molecule is
contained in its wavefunction (the wavefunction is a function of
the coordinates of all the electrons and all the nuclei). The trouble
is, these are difficult to calculate. We can make a first
simplification by concentrating on the wavefunction for the
electrons separately from that of the nuclei (the Born-Oppenheimer
approximation). The electronic wavefunction for a molecule will
tell us all about its electronic structure - i.e. how the electrons are

2
arranged in the molecule. But how do we find the electronic
wavefunction?

The Variational Method:

This is a method, based on the variational principle, for

finding approximate wavefunctions. It is our main weapon in the
struggle to find electronic wavefunctions for molecules and
enables us to calculate molecular orbitals, molecular electronic
wavefunctions and potential energy curves and surfaces. The
method works as follows:

1. Construct an approximate wavefunction  which depends on

some parameters ci.

2. Calculate the energy using equation (1). The energy calculated

in this way, E(ci), will depend on the parameters ci which are
part of the approximate wavefunction.

ψ Hˆ ψ dv
E (ci )   *
*

(2)
 ψ ψ dv

3. Choose the parameters ci so as to minimize the energy

E
0 (3)
ci
This minimized energy is still above the true ground state
energy. The wavefunction with this set of parameters is the best
wavefunction of the form which has been chosen. It gives the

3
lowest possible energy, as close as possible to that of the exact
ground state wavefunction.

As an example, consider HCl. We can think of a wavefunction in

which the bond is formed between H and Cl by pairing an electron
in the hydrogen in the H 1s orbital with one in the Cl 3pz orbital.
This describes covalent bonding. Let’s call this wavefunction cov.

H
Cl
cov

The energy of this wavefunction is cov.

However, this wavefunction doesn’t give the full picture of

bonding in HCl: because the electrons are paired, it doesn’t allow
both electrons to be simultaneously on the chlorine atom. It is
only a partial description. Chlorine is highly electronegative, so it
is quite likely that a full description would allow more polar
structures to play a part in the full description of HCl.

We can write down another wavefunction which goes to the ionic

extreme, and has both electrons on the chlorine atom, i.e.
representing the H+Cl– structure. In this wavefunction, ion, both
electrons are in the chlorine 3pz orbital.

Both of these wavefunctions, cov and ion, give only a partial

description of the bonding in HCl, each representing one extreme.
A better overall description - a better wavefunction- is given by
combining the two:
 = ccovcov + cionion

4
The coefficients ccov and cion are numbers which give the relative
contribution of the ionic and covalent forms. The variation
principle tells us that the best wavefunction (of this sort) is the one
with the lowest energy. This means we can find the best
wavefunction by varying the coefficients until we find the
wavefunction which has the lowest energy - this is the variational
method.
We might find, for example, that the best wavefunction at a
particular internuclear distance is  = 0.3 cov + 0.95 ion
(indicating that the ionic form plays a bigger role than the covalent
form).

E2

cov

ion
E1

The energy of the better wavefunction resulting from a

combination of the original wavefunctions is lower than the
energies of either of the original wavefunctions. The values of
the coefficients is found by minimizing the energy with respect
to them - the variational method.

The variational method is the most important way of finding

molecular wavefunctions.

5
 Molecular Orbitals and the Linear Combination of Atomic
Orbitals (LCAO) Approximation

To find electronic wavefunctions of molecules, usually the

problem is broken down to simplify it. The overall
wavefunction () is treated as a product of molecular orbitals
():

= |123….n|

Each MO is a wavefunction for one electron, and each MO spreads

over the whole molecule. We will see that individual MOs can tell
us a lot about molecular behaviour.

In the linear combination of atomic orbitals (LCAO)

approximation, each MO is treated as being made up of the AOs of
the atoms in the molecule. To find the MOs, we need to find the
coefficients which appear in the linear combination, by the
variational method.

Consider a molecular orbital which is written as a linear

combination of atomic orbitals:

ψ  c11  c22  c33  c44  (4)

where  is the molecular orbital and 1 etc. are the atomic orbitals
centred on the different atoms.

6
 is our approximate wavefunction. We will substitute it into eq.2
above and use the variational method to determine the molecular
orbital coefficients c1 etc.

As a simple example, we will expand our molecular orbital as a

linear combination of just two atomic orbitals (this could, for
example, be the two 1s AOs in the hydrogen molecule):

ψ  c11  c2 2 (5)

7
Now substituting into eq. 2 we obtain:

ψ Hˆ ψ dv
E (ci )   *
*

 ψ ψ dv
(c11  c22 ) Hˆ (c11  c22 ) dv
 (6)
 (c11  c22 )(c11  c22 ) dv
c12  1Hˆ 1 dv + c1c2  1Hˆ 2 dv + c2c1  2 Hˆ 1 dv + c2 2  2 Hˆ 2 dv

c12  11 dv + c1c2  12 dv + c2c1  21 dv + c2 2  22 dv
We now define the hamiltonian and overlap matrix elements:

Hamiltonian matrix element:

ˆ 1 dv
H11   1H
ˆ 2 dv = H 21
H12   1H
(7)

Overlap matrix element:

S11   11 dv = 1
S12   12 dv = S 21
(8)

Note the symmetry H12 = H21 and S12= S21 in these matrix elements
or integrals.

Eq. 6 now simplifies to give:

c12 H11 + 2c1c2 H12 + c2 2 H 22

E (ci )  2
c1 S11 + 2c1c2 S12 + c2 2 S 22
(9)
c12 H11 + 2c1c2 H12 + c2 2 H 22

c12 + 2c1c2 S12 + c2 2

This can be rewritten in the form:

 
E c12 + 2c1c2 S12 + c2 2  c12 H11 + 2c1c2 H12  c2 2 H 22 (10)

8
Now differentiating eq.10 with respect to c1 we obtain:
E 2
c1
 
c1 + 2c1c2 S12 + c2  E 2c1  2c2 S12   2c1H11 + 2c2 H12
2
(11)

E
We now set c1
0 to obtain:
E 2c1  2c2 S12   2c1H11 + 2c2 H12
or
2  H11  E  c1   H12  ES12  c2  0 (12)
or
 H11  E  c1   H12  ES12  c2  0

If we differentiate eq.10 with respect to c2 we obtain the equation :

 H 21  ES21  c1   H 22  E  c2  0 (13)

Secular Equations and Secular Determinants

Equations 12 and 13 form the secular equations:

 H11  E  c1   H12  ES12  c2  0
(14)
 H 21  ES21  c1   H 22  E  c2  0

These equation have a "trivial" and useless solution c1= c2 = 0.

The condition that there should exist a nontrivial solution of these
equations is that the secular determinant should be zero:

 H11  E   H12  ES12 

0 (15)
 H 21  ES21   H 22  E 

Everything in this equation is a known number except E. The

equation is therefore an equation for E. In the present case where

9
the molecular orbital was a linear combination of just two atomic
orbitals it is a quadratic equation and has two solution for E.
These two values of E are the molecular orbital energies. We
always get the same number of molecular orbitals as atomic
orbitals we start with.

10
If we multiply out the secular determinant we obtain:
 H11  E  H 22  E    H12  ES12  H 21  ES 21   0
(16)
E 2   H11  H 22  2 H 12 S12  E   H 11 H 22  H12   0
2

which you should now be able to recognise as a quadratic equation

for E. It has two solutions E1 and E2. These are the two molecular
orbital energies.

We now take one of these, E1 , and place it in the secular equations

(eq. 14). These equations now have a nontrivial solution and we
find the ratio c2/c1 by solving them. The absolute value of the
coefficients c1 and c2 can only be found by using the
normalization condition:

ψ
*
ψ dv  1

 (c11  c22 )(c11  c22 ) dv  1

2 2
c1  11 dv + c1c2  12 dv + c2c1  2 1 dv + c2  2 2 dv = 1
c12 + 2c1c2 S12 + c2 2 = 1
(17)

We now add an extra index to the coefficients c1 and c2 to show

that they belong to molecular orbital no. 1. We obtain the
coefficients for molecular orbital No. 2 in the same way by
substituting the second solution of the secular equations, E2 , into
the secular equations.

Molecular orbital no. 1

Molecular orbital energy: E1

11
Molecular orbital: ψ1  c111  c122

Molecular orbital no. 2

Molecular orbital energy: E2
Molecular orbital: ψ 2  c211  c222

12