Benzene:

1
2

5
4

We will solve Schodinger equation for this molecule by considering only p-orbitals of six carbons under the Huckel
approximation. Huckel approximation, though quite crude, provides very useful results.
Since a-bonding in planar molecules has different symmetry from the ^-bonding, corresponding molecular orbitals
separate in the Hamiltonian -do not have cross off-diagonal element and thus can be solved separately. Besides,
a-bonding is much stronger than ^-bonding, ^- and ^ D molecular orbitals lie within the a ? a D gap. Thus HOMO and
LUMO orbitals are either of ^,^ D -type or nonbonding n-type. All that allows considering only p z -orbitals in
conjugated molecules for treating their spectroscopic features, while s and p x ,p y orbitals will be responsible
primarily for a-bonding, i.e. molecules shape.
Huckel approximation can be summarized with following statements:
1) < p i |p j >= N ij - all overlap integrals are zero
2) < p i |H|p i >= J - diagonal elements equal atomic energies of carbon p z electrons
3) < p i |H|p j >= Kneighbors - the interaction energy is nonzero only for neighboring carbons.
Both, J and K, are negative and can be paraterized for a typical conjugated carbon and carbon-carbon bond. Such
a parameterization can be also introduced for hetero atoms, S, O, N
If we write the secular equation for benzene

det

J?E

K
0

J?E

J?E

J?E

J?E

J?E

= 0, : ?4K 6 + 9K 4 J ? E 2 ? 6K 2 J ? E 4 + J ? E 6 = 0.

The solutions can be found E = J + K,E = J ? 2K,E = J ? K,E = J + 2K. For an arbitrary cyclic molecule with identical
conjugated bonds, the solutions can be written as E = J + 2Kcos 2^
and graphically represented as enegries
n k
horizontally leveled at the corners of the appropriate polygon symmetrically placed on one of its corners:

"Graphical solution" to cyclic aromatics

E = + 2 cos(2 k/n)
One can see that there are degenerate solutions but symmetries of the appropriate orbitals can be found by
finding appropriate eigen functions.
Lets do that analysis for benzene using group symmetry:
Benzene belongs to the ooint group:D 6h When six p^ orbitals are taken as the basis set for a representation of this
group we obtain @ ^ , The character matrix is also shown on the top. C v2 s and a v s were chosen to go through
carbon atoms and C vv 2 s and a d s go in between.
D 6h E 2C 6 2C 3 C 2 3C v2 3C vv 2

A 1g

A 2g

?1

?1

?1

?1

B 1g

?1

?1

?1

?1

?1

?1

B 2g

?1

?1

?1

?1

?1 ?1

E 1g

?1

?2

E 2g

?1

?1

?1

A 1u

?1 ?1

?1 ?1 ?1

?1

A 2u

?1

?1

?1 ?1

?1 ?1

B 1u

?1

?1

?1

?1

?1

?1

B 2u

?1

?1

?1

?1

?1

?1

E 1u

?1

?2

?2 ?1

E 2u

?1

?1

?2

?2

@^

?2

?6

@ 3N 36

?4

12

2S 3 2S 6 a h 3a d 3a v

?1 ?2

?1

That representation can be reduced into irreducible for this point group:
@ ^ = B 2g + E 1g + A 2u + E 2u
A bit faster approach is to recognize that D 6h =D 6 C i and thus use a smaller group, D 6 , instead:
2

D 6 E 2C 6 2C 3 C 2 3C v2 3C vv 2
A1

A2

?1

?1

B1

?1

?1

?1

B2

?1

?1

?1

E1

?1

?2

E2

?1

?1

@^

?2

That representation can be reduced into irreducible for this point group:
@ ^ = B 2 + E 1 + A 2 + E 2 which can be easily assigned g and u subscript based on the required negative character
with respect to operation a h in D 6h . Thus @ ^ =B 2g + E 1g + A 2u + E 2u
We have not found the eigen functions nor energies yet, just their symmetries. Now the task is easier than before.
We have two options:
a)Trained eye can draw the nodal planes and assign the molecular orbitals by analysis of their characters with
respect to appropriate symmetry operations as shown below.I prematuraly put energies next to the MO, but note
that encreaing number of the nodal planes corresponds to increasing energy of MO. Also, a 2u with energy J + 2K is
the lowest because both, J and K, are negative. The approach is handy but might be confusing when trying to find
those c 1 and c 2 coefficients to finish the construction.

c1

c1

-c1

-c1

--c 2

c1

c1

-c1

c1

c1

c1

-c1

-c2

c2

c1

c1

-c1

-c1

c1

+
c1

e 2u

--c

c1

b2g

c1

-c1

e1g

a2u

+ 2

-c1

-c2

b) More straight forward (also a bit more tedious) approach is by using the projection operator:
P W => R e W R D O R
Again, well resort to a smaller group, D 6 , instead of D 6h . Then projection operators would look shorter :0). For
example: P A 2 = 1O E + 1O C 6 + 1O C ?16 + 1O C ?13 + 1O C 3 + 1O C 2 ? 1O C v2 1 ? 1O C v2 2 ? 1O C v2 3 ? 1O C vv2 1 ? 1O C vv2 2 ? 1O C vv2 3 .
Using which on p 1 results in
P A 2 p 1 = p 1 + p 2 + p 6 + p 3 + p 5 + p 4 + p 1 + p 5 + p 3 + p 2 + p 4 + p 6 = 2p 1 + p 2 + p 3 + p 4 + p 5 + p 6
3

After normalizing we obtain:S 4 a 2u =

p 1 + p 2 + p 3 + p 4 + p 5 + p 6 . The same way it can be dobe with others.

Some attention is required for doubly degenerate species. The projection operator for E 2 :
P E 2 = 2O E ? 1O C 6 ? 1O C ?16 ? 1O C ?13 ? 1O C 3 + 2O C 2 .when applied to say p 1 gives:
P E 2 p 1 = 2p 1 ? p 2 ? p 6 ? p 5 ? p 3 + 2p 4 .That gives only one symmetry adapted orbital,
S 5 = 1 2p 1 ? p 2 ? p 6 ? p 5 ? p 3 + 2p 4
2 3

Another has to be found by generating a similar one after applying P E 2 say to p 2 :

2p 2 ? p 3 ? p 1 ? p 6 ? p 4 + 2p 5

2p 1 ? p 2 ? p 6 ? p 5 ? p 3 + 2p 4 ?6

2 3

3 p
2
2

? p3 + p5 ? p6

Analogously:P E 1 = 2O E + 1O C 6 + 1O C ?16 ? 1O C ?13 ? 1O C 3 ? 2O C 2 and

P E 1 p 2 = 2p 2 + p 3 + p 1 ? p 6 ? p 4 ? 2p 5 and
S 3 = 2p 2 + p 3 + p 1 ? p 6 ? p 4 ? 2p 5 ?

2p 1 + p 2 ? p 3 ? 2p 4 ? p 5 + p 6 6 =

2 3

3p
2 2

+ 32 p 3 ? 32 p 5 ? 32 p 6 =

B 2g S 1 =

3 p
2
2

+ p3 ? p5 ? p6

p 1 ? p 2 + p 3 ? p 4 + p 5 ? p 6

1
6

E 1g

S2 =

2p 1 + p 2 ? p 3 ? 2p 4 ? p 5 + p 6

1
2 3

S3 =
A 2u S 4 =

1 p
2
2
1

+ p3 ? p5 ? p6

p 1 + p 2 + p 3 + p 4 + p 5 + p 6

E 2u S 5 =

2p 1 ? p 2 ? p 3 + 2p 4 ? p 5 ? p 6

2 3

S6 =

1 p
2
2

? p3 + p5 ? p6

So we found the eigen functions which graphically are represented below with shaded areas corresponding to a
negative sign and the size of each circle resembling the value of coefficient in the eigen function.

b2g

e2u

e1g

a 2u

The energies of these MO can be found by evaluating appropriate < S i |H|S i >:
E(b 2g ) = < S 1 |H|S 1 >=< 1 p 1 ? p 2 + p 3 ? p 4 + p 5 ? p 6 |H| 1 p 1 ? p 2 + p 3 ? p 4 + p 5 ? p 6 >=
6

< p 1 |H|p 1 ? p 2 + p 3 -p 4 +p 5
=

+ < p 3 |H| p 1

1
6

? p 6 > ? < p 2 |H|p 1 ? p 2 + p 3 -p 4 +p 5 -p 6 > +

? p 2 + p 3 ? p 4 + p 5 -p 6 > ? < p 4 |H| p 1 -p 2

< p 5 |H| p 1 -p 2 +p 3

+ p3 ? p4 + p5 ? p 6 > +

? p 4 + p 5 ? p 6 > ? < p 6 |H|p 1 -p 2 +p 3 -p 4

+ p5 ? p6 >

matrix element for highlighted orbitals are zero

1
6

1
6

< p 1 |H|p 1 ? p 2 ? p 6 > ? < p 2 |H|p 1 ? p 2 + p 3 > + < p 3 |H|?p 2 + p 3 ? p 4 > ?

= J ? 2K

Energies of other orbitals are calculated the same way:

S1 =

p 1 ? p 2 + p 3 ? p 4 + p 5 ? p 6

1
6

Ee 1g

S2 =

2p 1 + p 2 ? p 3 ? 2p 4 ? p 5 + p 6

1
2 3

S3 =
Ea 2u = S 4 =

1 p
2
2
1

+ p3 ? p5 ? p6

p 1 + p 2 + p 3 + p 4 + p 5 + p 6

Ee 2u

S5 =

2p 1 ? p 2 ? p 3 + 2p 4 ? p 5 ? p 6

2 3

S6 =

1 p
2
2

? p3 + p5 ? p6

b2g

-2 : b2g = 0.408(p1 p2+p3p4+p5p6)

e2u

- : e2u(1) = 0.289(2p1 p2p3+2p4p 5p6); e2u(2) = 0.5( p 2p3+p5p6)

e1g

: e1g(1) = 0.289(2p1+ p2p32p4p5+p6); e1g(2) = 0.5( p 2+p3p5p6)

a1u

2 : a1u = 0.408(p1+ p2+p3+p4+p5+p6)

We see that MO energies are symmetric with respect to value of J. All planar conjugated hydrocarbons with such
property are called alternant. Alternant hydrocarbons can be recognized by their ability to have all carbons labeled
in two colors alternatingly (thus the name), i.e. when each neighbor has a different color.
1
The ground state configuration is a 2u 2 e 1g 4 = X A 1g . Note that this ground state energy is lower than that of
cyclohexatriene (benzene with three localized double bonds). The effect of sharing ^-electrons, the so-called
delocalization or resonance energy, equals 2K.
The lowest excited state configuration is a 2u 2 e 1g 3 e 2u 1 (with energy 2K above the ground state) from which
following states can be constructed:
e 1g e 2u = B 1u + B 2u + E 1u .each of which can be either a singlet or a triplet, i.e.:

X:1A1g

b2g

b2g

b2g

e2u

e2u

e2u

e1g

e1g

e1g

a1u

a1u

a1u

B2u, 1B1u, 1E1u

B2u, 3B1u, 3E1u

At this level of comlexity we cannot choose which excited state is the lowest energy, 1 B 1u , 1 B 2u , or 1 E 1u (all we know
is that they each have corresponding triplet of lower energy 3 B 1u < 1 B 1u , 3 B 2u < 1 B 2u , and 3 E 1u < 1 E 1u ). We have to
analyse other higher energy one-electron excited state configurations and find if there are any of the same
symmetry as the three under consideration. Configurational interaction between configurations of the same
symmetry lowers the appropriate lower energy state while increasing the corresponding high energy state energy.
The next one electron excited state configurations (with energy 3K) are:a 2u 2 e 1g 3 b 2g 1 and a 2u 1 e 1g 4 e 2u 1 .
Neither of them has a proper symmetry. The former makes: e 1g b 2g = e 2g , while the latter: a 2u e 2u = e 2g

E2g

E2g

B2u

We shall include even higher one electron excited state configuration with energy 4K. There is only one,
a 2u 1 e 1g 4 b 2g 1 , which produces a 2u b 2g = b 2u . Thus, out of the three states, 1 B 1u , 1 B 2u , and 1 E 1u , only 1 B 2u has
two configurations. Consequently, it will be the lowest energy state.

Since in the D 6h group the dipole moment components behave like E 1u (x,y) and A 2u (z), the only dipole allowed
1
1
transition is C E 1u ? X A 1g which indeed shows a high extinction coefficient, while the lower energy transitions, 1 B 1u
and 1 B 2u , are electronically forbidden. They are observed anyway but due to so call vibronic coupling.
Before that lets consider electric dipole allowed transitions first. The matrix element for the transition moment is:
M evvve"v" =< f De v v v |W|f e vv v vv >=< f De v |W|f e vv >< f Dv v |f v vv >= M eve" < f Dv v1 |f v vv1 >< f Dv v2 |f v vv2 >...
The selection rules in this case are defined by the selection rules for electronic part M eve" and the Franck-Condon
factor, < f Dv v |f v vv > 2 , for vibrational part. The latter would be nonzero if that integrand has the totally symmetric
species, which means that for totally symmetric vibrations all vibrations are allowed:
Av = 0,1,2,3,...
For all other vibrations the totally symmetric integrand appears only for even difference in vibrational number, i.e.
when:
Av = 0,2,4,...
In cases like the lowest energy transition in benzene, which is electronically forbidden, the only way the matrix
element for the transition moment would be nonzero is to step back from the Born-Oppenheimer approximation
and consider mixing nuclear (vibrational) and electronic coordinates, so called vibronic coupling.
The electronic Hamiltonian parametrically depends on on vibrational coordinates:
e
H e = H 0e + > i /H
Q i +..= H 0e + H v
/Q i
The excited state wavefunction f 0f becomes accordinly mixed with other zero-order electronic states :
f e v = f 0f + > k c k f 0k
where the degree of mixing, c k =

<f 0k |H v |f 0f >
E 0f ?E 0k

energy separation between the states

, depends on both, the vibronic coupling element < f 0k |H v |f 0f > and the

E 0f ? E 0k . The electronic transition moment becomes:

M eve" =< f e v |W|f e vv >=< f 0f |W|f 0e vv > + > k c k < f 0k |W|f 0e vv >
While the first term (0-0 transition) is zero, the other terms might be not. Then the intensity of transition would
become nonzero, or as it is often referred to as intensity is borrowed from a neigboring transition. That is why B 1u
transition, beeing closer to the allowed E 1u , is more intense than B 2u . The energy factor E 0f ? E 0k is not the only
one responsible for nonzero c k , the other part , < f 0k |H v |f 0f >, enforces additional selection rules:
e
< f 0k |H v |f 0f >= /H
< f 0k |Q i |f 0f >
/Q i
Q i =0

or Av = 13,5,...
for vibronicaly active mode, the mode for which the symmetry of @ f k @ Q i matches that of the

ground state X. The rules for other vibrational states would follow the rules of ordinary Franck-Condon
1 factor
described above. In case of benzene, vibronically allowed transition should be a 2u and e 1u . For the A B 2u state it
translates to a reqire for either b 1g b 1g B 2u = B 2u or e 2g e 2g B 2u = E 1u . There are 123-6=30 vibrational
modes in benzene, which for the ground state would be realized in 20 fundamental frequencies:
2a 1g + a 2g + a 2u + 2b 1u + 2b 2g + 2b 2u + e 1g + 3e 1u + 4e 2g + 2e 2u . From these modes only e 2g satisfy the selection rules
(X 15 e 2g through X 18 e 2g ) for vibronic transition. By symmetry all four of them can contribute and those that have
e
the highest value of /H
contribute the most. C-H vibrations do not affect the energies of ^-electrons very much,
/Q i
it that should be a mode that shakes the molecular sceleton associated with ^-system. A likely candidate is C-C-C
in plane bending, X 18 e 2g , with energy ca. 500-600 cm ?1 (606 cm ?1 in the ground state and 522 cm ?1 in the
electronically first excited state). Thus the excited vibronic transitions that should be observed are:
1
1
1
A B 2u 18 10 - transition from vibrational ground state in the X A 1g to A B 2u at vibrationally excited state with v"=1 at
X 18 e 2g .
1
1
0
Another
1 possibility is to observe a hot A B 2u 18 1 transition from the vibrationally excited with v=1 at X 18 e 2g X A 1g to
the A B 2u and v"=0 at X 18 e 2g . The latter transition is very temperature dependent (hot). What about combinational
transitions? Again, the other vibrations have to be involved in ring distortion vibrations (preferably C-C stretching
since it affects the Hamiltonian the most) and posses appropriate symmetry, which can be achived by combining
X 18 e 2g with totally symmetric combinations of the ring distortion vibrations such as X 2 a 1g (992 cm ?1 in the ground
state and 923 cm ?1 in the electronically first excited state).
1
As a result, a series of vibronic transitions
A B 2 m 18 1 appears as a famous benzene spectrum near 260 nm.
2u

10