Chemistry Reference
In-Depth Information
energy,
b
3u
)), because they are symmetrical with respect to the
C
2
axis
along the long molecular axis and they exhibit positive overlap between the carbon
components of the two adjacent
σ
∗
(
π
(
a
g
,
σ
∗
(
) orbitals. This lowers the energy consider-
ably. The other two orbitals are based on the
b
1g
and
b
2u
orbitals,
π
σ
∗
(
π
(
b
1g
,
b
2u
)).
σ
∗
(
These orbitals are the two remaining symmetry-adapted combinations of the
)
orbitals. By having negative overlap between the carbon components of the two
adjacent
π
σ
∗
(
) orbitals they now lie around 1 eV higher in energy than their
a
g
and
b
3u
counterparts (see Fig. 6.17).
The
π
π
∗
(
π
(
b
3g
,
a
u
)) orbitals are two of the symmetry-adapted combinations of
π
∗
(
the
) orbitals. By being symmetrical with respect to the
C
2
axis along the long
molecular axis, the
π
orbital of the central carbon atom of the C(CN)
2
substituent
cannot mix into these symmetry-adapted combinations, and thus, in practice, the
π
∗
(
π
) orbitals do not delocalize toward the C
6
H
6
ring.
The next peak, at approximately 5.6 eV from the LUMO minimum, arises from
π
a
π
-type orbital delocalized over all the molecule but originating from the highest
π
-type orbital of C
6
H
6
(
b
2g
). Superposed to this contribution there is a second one,
of
σ
C
−
H
antibonding. Thus, the first and fourth peaks are considerably more concentrated in
the benzene ring whereas all other peaks are strongly based on
σ
-type and also localized on the C
6
H
6
moiety, whose main character is
π
CN
orbitals. This
is clear from the comparison of the C and N PDOS in Fig. 6.5(d).
The origin of the different peaks of the N PDOS of TTF-TCNQ (Fig. 6.4(b)) can
be understood exactly as for the neutral TCNQ (Fig. 6.5(d)). C1
s
and N1
s
spectra of
TCNQ are shown in Figs. 6.14(b) and (c), respectively. Note that the contribution of
σ
∗
(
π
(
a
g
,
b
3u
)), which should appear (see the dotted line in Fig. 6.14(c)) in between
π
∗
(
a
u
π
∗
(
+
σ
∗
(
those of
b
2u
)), both clearly visible in
the N1
s
spectra, does not appear at all. That the larger peak in the N1
s
spectra is
partially associated with
,
b
1u
) and
π
(
b
3g
,
a
u
))
π
(
b
1g
,
σ
∗
(
(
b
1g
,
b
2u
)) is clear not only from the energy separation
from the first peak but, more importantly, from the angular dependence of the peaks,
as discussed next.
Figures 6.16(b) and (c) show the angular dependence of the C1
s
and N1
s
NEXAFS spectra, respectively. The most salient feature of Fig. 6.16(c) is the in-
tensity increase of the
π
π
∗
(
+
σ
∗
(
π
(
b
3g
,
a
u
))
π
(
b
1g
,
b
2u
)) peak for increasing
θ
E
π
∗
(
b
2g
) peaks remain nearly
unchanged. For TCNQ the
C
2
molecular axis forms an angle of about 36
◦
with
the
c
∗
-direction. Since CN bonds form an angle of 60
◦
with this
C
2
molecular
axis within the molecular plane, the intensity associated to
π
∗
(
a
u
,
values. However, the benzenic-type
b
1u
) and
σ
∗
(
π
(
b
3g
,
a
u
)) should
π
∗
(
increase for larger
θ
E
. However, the intensity associated with
π
) orbitals, the
π
∗
(
benzenic-like orbitals and
a
u
)), should exhibit the opposite behaviour
because they are perpendicular to the
π
(
b
3g
,
σ
∗
(
π
) orbitals. From Fig. 6.16(c) it is
π
∗
(
+
σ
∗
(
clear that the
π
(
b
3g
,
a
u
))
π
(
b
1g
,
b
2u
)) peak increases with regard to the
σ
∗
(
benzenic-type orbitals for increasing
θ
E
values because of the increasing
π
) and
