Electron delocalization in cross-conjugated p-phenylenevinylidene oligomers

Article abstract of DOI:10.1002/chem.200204671

The synthesis, structure, and electronic properties of a series of cross-conjugated p-phenylenevinylidene oligomers with one to four double bonds are reported. The X-ray crystal structure of the compound with two double bonds reveals a nonplanar conformat

Full text of DOI:10.1002/chem.200204671

FULL PAPER
Electron Delocalization in Cross-Conjugated p-Phenylenevinylidene
Oligomers
Mark Klokkenburg,[a] Martin Lutz,[b] Anthony L. Spek,[b] John H. van der Maas,[c] and
Cornelis A. van Walree*[a]
Abstract: The synthesis, structure, and
electronic properties of a series of cross-
mode does not adequately reflect the
electronic interaction between the re-
peating units is, however, not very
strong, which has the consequence that
spatial extension of the molecular orbi-
tals does not lead to a red shift of the
highest occupied molecular orbital
lowest unoccupied molecular orbital
(HOMO LUMO) electronic transi-
tion. This is related to the feature that
the modest narrowing of the HOMO
LUMO gap with the chain length is
accompanied by a relatively large re-
duction of electron repulsion. This find-
ing implies that care should be taken in
the use of electronic spectra for the
evaluation of conjugation phenomena.
ˆ
C C bond order and has a local nature.
conjugated
p-phenylenevinylidene
In contrast, electronic spectra and elec-
trochemical data, as well as AM1 and
PPP/SCF calculations, reveal that the
cross-conjugated compounds basically
behave as linearly p-conjugated systems
in the sense that molecular orbitals are
delocalized over the entire structure and
systematically change in energy. The
oligomers with one to four double bonds
are reported. The X-ray crystal structure
of the compound with two double bonds
reveals a nonplanar conformation with
torsion angles about the C(phenylene)
C(vinylidene) and C(phenyl) C(vinyl-
idene) formal single bonds of 39.5(2)8
and 30.5(2)8, respectively. Admixture of
quinoid character in the ground state is
observed. Infrared and Raman spectro-
scopy do not provide a clear picture of
the degree of electron delocalization in
Keywords: conjugation
¥
cyclic
voltammetry semiempirical
¥
calculations ¥ UV/Vis spectroscopy
¥ vibrational spectroscopy
ˆ
the series, since the C C stretching
which polyacetylene, poly(p-phenylenevinylene)s, and poly-
thiophenes are well-known examples, have by far received the
most attention. However, alternative modes of conjuga-
Introduction
An abundant amount of work has been directed towards the
design of organic compounds and materials with attractive
optical, optoelectronic, and conductive properties.[1] These
properties have in common that they rest on the occurrence of
electron delocalization. Linear p-conjugated systems, of
tion,[2, 3] such as s conjugation found in polysilanes,[4]
s p
conjugation in oligo(cyclohexylidene)s[5, 6] and silylene-ary-
lene polymers,[7, 8] and homoconjugation in diphenylpropanes,
diphenylsilanes,[9, 10] and 7,7-diarylnorbornanes,[11] can also
form the basis of interesting electronic features.
A form of conjugation which has only sparsely been
addressed in materials research is cross-conjugation, which
has been defined as the situation in which two unsaturated
fragments are not conjugated to each other, but are both
conjugated to a third unsaturated moiety.[12] Examples of
simple cross-conjugated compounds are 3-methylene-1,4-
pentadiene, benzophenone, and 1,1-diphenylethene. Recent-
ly, interest in this type of conjugation has flourished, as
illustrated by the development of dendralene-based[13, 14] and
enediyne-based[15 17] systems. This, however, does not alter the
fact that the understanding of cross-conjugation is still limited.
The principal question of how conjugated cross-conjugated
compounds are, that is, to what extent electrons are delocal-
ized in such systems, still needs a solid answer. In the few cases
where a systematic set of compounds was available it was
[a] Dr. C. A. van Walree, M. Klokkenburg
Debye Institute
Department of Physical Organic Chemistry
Utrecht University
Padualaan 8, 3584CH Utrecht (The Netherlands)
Fax : (‡31)302-534-533
E-mail: c.a.vanwalree@chem.uu.nl
[b] Dr. M. Lutz, Prof. A. L. Spek
Bijvoet Center for Biomolecular Research
Crystal and Structural Chemistry
Utrecht University
Padualaan 8, 3584CH Utrecht (The Netherlands)
[c] Prof. J. H. van der Maas
Department of Vibrational Spectroscopy
Faculty of Chemistry
Utrecht University
Sorbonnelaan 16, 3584CA Utrecht (The Netherlands)
3544
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200204671
Chem. Eur. J. 2003, 9, 3544 3554

3544 3554
deduced that cross-conjugation does not result in pronounced
p-electron delocalization and that electronic properties are
essentially determined by the longest linear p-conjugated
fragment present.[14, 18] These findings were, however, merely
based on electronic absorption spectra and were not the result
of a broad study.
acetophenone in the presence of CeCl3.[20, 21] Subsequent acid-
catalyzed dehydration afforded 2. Homologues 3 and 4 were
synthesized in a procedure starting with the reaction between
4-bromophenyllithium and either 4-bromoacetophenone or
1,4-diacetylbenzene to give the respective products 6a or 6b
in good yields (98%). Lithiation of 6a and 6b with three and
four equivalents of n-butyllithium, respectively, followed by
treatment with acetophenone resulted in 7a and 7b. Dehy-
dration of 7a and 7b gave the oligomers 3 and 4, respectively.
NMR analysis and GC-MS (after dehydration) showed that
during the synthesis of alcohols 5, 7a, and 7b a large number
of compounds had been formed, including a substantial
amount (30 40%) of aldol condensation products. Appa-
rently, for dilithiocompounds nucleophilic addition is not
particularly effective and these compounds also act as a base.
In the synthesis of 5 cerium(iii) chloride was used to favor
nucleophilic addition and to suppress side reactions.[20, 21]
Despite the strong oxophilicity of this reagent, this procedure
did not result in a high yield.
In this contribution the synthesis and the characterization
of cross-conjugated oligo(p-phenylenevinylidene)s 1 4, a
series of oligomers derived
n=1: 1
n=2: 2
from 1,1-diphenylethene, are
described. They are topological
isomers of well-studied oli-
go(p-phenylenevinylene)s, the
n=3: 3
n=4: 4
n
important difference being that the phenylene and phenyl
moieties are connected by a single sp2-hybridized carbon atom
instead of by two sp2-hybridized carbon atoms. Insight into the
p-electron system of the oligo(p-phenylenevinylidene)s is
obtained by the analysis of its properties as a function of chain
length. Hereto IR, Raman, UV, and fluorescence spectro-
scopy are used, since the energy and intensity of characteristic
transitions usually depend strongly on the extension of the p
system. In addition, the electrochemical behavior of 1 3 is
monitored by cyclic voltammetry, while a single-crystal X-ray
structure of 2 could be obtained, through which the molecular
structure is accurately established. The experimental results
are correlated with PPP/SCF- and AM1-calculated properties.
This study on oligo(p-phenylenevinylidene)s leads to new
insights in the behavior of cross-conjugated compounds,
insights which may lead to applications of cross-conjugated
systems in materials science.
Molecular structures: In the oligo(p-phenylenevinylidene)s
rotation about the C(phenyl/phenylene) C(vinylidene) for-
mal single bonds is not free because of steric interactions
between the ortho hydrogen atoms of neighboring aromatic
groups. Therefore, for a given compound a number of
geometries are possible which differ in the orientation of
the vinylidene groups with respect to a phenylene moiety;
these groups can be situated on either the same side or on
different sides of the ring. Compound 2 can adopt either Ci or
Cs symmetry, while 3 can occur in two different C2 forms or in
a C1 form. Six structures are possible for 4: two different
isomers of Ci symmetry, two different isomers of Cs symmetry,
and two different isomers of C1 symmetry. Previous research
has established that 1,1-diphenylethene (1) belongs to the C2
point group.[22, 23]
Results and Discussion
In order to find out which structure prevails, 2 was
subjected to single-crystal X-ray diffraction. This compound
crystallizes in the orthorhombic Pbca space group with four
molecules per unit cell. (See also the Experimental Section.)
As depicted in Figure 1, the molecule has an exact, crystallo-
graphic Ci symmetry and thus adopts an undulating, stretched
structure. (In the Cs geometry, with the two vinylidene groups
occupying the same face of the phenylene ring, the structure is
much more compact.) Although in the X-ray crystal structure
the ortho hydrogen atoms H3and H7 avoid each other, there
are still close contacts between the hydrogen atoms on C5 and
Synthesis: While 1,1-diphenylethene (1) was obtained from a
commercial source, oligomers 2 4 were synthesized through
a series of conversions involving the reaction between
phenyllithium compounds and acetophenone derivatives,
followed by dehydration of the formed alcohols (Scheme 1).
For the synthesis of 2, compound 5 was first prepared by
treatment of 1,4-dilithiobenzene[19] with two equivalents of
OH
O
b
a
Li
Li
Li
+ 2
2
Me
Me
2
À
C11; the interatomic distance H5 H11 of 2.24 ä is signifi-
5
cantly shorter than the sum of van der Waals radii of 2.40 ä. It
is conspicuous that the torsion angles about the C(phenyl-
ene) C(vinylidene) and the C(phenyl) C(vinylidene) single
bonds are different; the C1-C2-C4-C5 torsion angle is
39.5(2)8, while the C5-C4-C6-C11 angle amounts to 30.5(2)8.
The nonplanar structure will reduce conjugation, but this
disadvantage is not very severe since the overlap in a p system
typically varies with the cosine of the torsion angle.
OH
Me
O
c
Br
X
Br
Br
+
Me
n
X = Br, MeOC
b
n = 1: 6a; n = 2: 6b
OH
Me
n
n
ˆ
The olefinic C C bond length, 1.329(2) ä (Table 1), is
n = 3: 7a; n = 4: 7b
n = 3: 3; n = 4: 4
ˆ
relatively short when compared to the average C C bond
Scheme 1. Synthesis of oligo(p-phenylenevinylidene)s 1 4. Reagents and
conditions: a) CeCl3; b) p-toluenesulfonic acid, toluene; c) BuLi (3equiv
for 6a, 4 equiv for 6b), acetophenone (2 equiv).
length of 1.339(11) ä in ™conjugated∫ (defined as having a
torsion angle of 0 208 or 160 1808) phenyl-substituted
Chem. Eur. J. 2003, 9, 3544 3554
www.chemeurj.org
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3545

C. A. van Walree et al.
FULL PAPER
essentially identical heat of for-
mation and that there is no ther-
modynamical driving force favor-
ing the formation of only a single
isomer. The presence of an excess
of the Ci form of 2 may then be
related to preferential crystalliza-
tion or a to kinetic effect during
the synthesis. The AM1 calcula-
tions also showed the orbital en-
ergies of isomers to be identical,
which indicates that the degree of
conjugation (see below) is inde-
pendent of the actual molecular
Figure 1. Displacement ellipsoid plot (50% probability) of the X-ray crystal structure of 2. Symmetry
operations i: 1 À x, 2 À y, Àz.
structure.
À
Table 1. C C bond lengths in 2.
Structural features of the AM1 optimized structures were
qualitatively in agreement with the X-ray crystal structure of
2. All structures were twisted about the C(phenyl/phenyl-
ene) C(vinylidene) single bonds by about 408. Note that
during the optimization the C(phenylene) C(vinylidene) and
C(phenyl) C(vinylidene) torsion angles were constrained to
be equal, so that the difference between them in the X-ray
crystal structure of 2 could not be reproduced. In the AM1
structure of 1, which was optimized without any restrictions,
the torsion angle around the formal single bonds was however
also close to 408, which suggests that AM1 calculations give a
value of 408 anyway. In line with the X-ray crystal structure, a
small amount of quinoid character of the phenylene units is
revealed; in all isomers of 2 and 3 the AM1-calculated
™central∫ aromatic bonds lengths are 1.391 1.393 ä, while
the other bonds are 1.402 1.403ä long. In addition, in the
Bond
Length [ä]
Bond
Length [ä]
i
À
À
C1 C2
1.396(2)
C2
1.329(2)
1.395(2)
1.390(2)
1.381(2)
C1 C3
1.385(2)
1.494(2)
1.489(2)
1.400(2)
1.383(2)
1.386(2)
À
À
95(C2)2 C31.3
C4
À
À
C4 C5
C4 C6
À
À
C6 C7
C6 C11
À
À
C7 C8
C8 C9
À
À
C9 C10
C10 C11
alkenes.[24] Furthermore, the two formally single bonds C2 C4
and C4 C6, having respective lengths of 1.494(2) and
1.489(2) ä, belong to the long fraction of C(sp ) C(aryl)
bonds, for which the average length is 1.470(15) ä.[24] From
the bond lengths, the double bond in 2 thus seems to be rather
isolated in comparison to planar p systems. It is however clear
that delocalization is much more pronounced than in real
isolated p systems. Although the bonds in the phenylene ring
À
À
2
À
À
phenyl groups the C C bonds adjacent to the double bond are
longer than the other bonds in that ring, which in their turn
are not very different from each other.
i
À
have normal lengths, the C1 C3 distance is shorter than the
À
À
C1 C2 and C2 C3distances. This bond length alternation
points to the admixture of some quinoid character in the
ground state and must be based on a small degree of electron
delocalization. Quinoid character is not observed for the
Vibrational spectroscopy: The structural and conjugative
properties of 1 4 were further investigated with IR and
Raman spectroscopy. IR spectra of oligomers 2 and 3 are
depicted in Figure 2, while Raman spectra of 1 and 4 are
À
À
outer phenyl rings, in which the C9 C10 and C8 C9 bonds are
À
À
relatively short and the C6 C7 and C6 C11 bonds long. This
suggests that electron delocalization is most pronounced in
the central divinylbenzene-like unit.
shown in Figure 3. The frequency of the most prominent
27]
signals along with their assignments[25
are compiled in
Table 2. The following vibrations are of particular impor-
À1
ˆ
To establish whether the crystal picked for the X-ray study
was representative for the bulk of 2 as isolated, a powder
X-ray pattern that was simulated from the single-crystal data
was compared to a measured powder diffractogram (data not
shown). The presence of a few weak lines in the experimental
diffractogram which were not anticipated on basis of the
simulation means that it is not possible to claim that the
isolated material was 100% structurally homogeneous. There
may be some Cs isomer present. We nevertheless believe that
the highly prevailing geometry of 2 is the undulating stretched
structure displayed in Figure 1.
AM1 calculations on compounds 1 3 were performed to
investigate how molecular properties depend on the molec-
ular geometry and on the chain length. Obtained heats of
formations were 66.41 (1, C2), 110.80 (2, Ci), 110.85 (2, Cs),
155.16 (3, C2), 155.21 (3, C1), and 155.22 (3, C2') kcalmolÀ1.
These data show that all isomers of a given compound have an
tance: the vinylidene C C stretch mode near 1604 cm , the
n12 breathing mode of the phenyl end groups at 998 cmÀ1, the
CH2 wag motion near 906 cm , the p-phenylene C H out-
of-plane motion at 855 cmÀ1, and the (monosubstituted)
À1
ˆ
À
À1
À
phenyl C H out of plane mode near 775 cm .
Although peak positions are very similar, the IR spectra of
2 and 3 are remarkably different (Figure 2). The spectrum of 2
contains sharp signals that give rise to speculation that 2 exists
as a well-defined compound, that is, the Ci structure from
Figure 1 must be very dominant. In contrast, the bands in the
IR spectrum of 3 are broad (particularly in the interval 1300
950 cmÀ1). To find out whether this was due to a structural
inhomogeneity or to solid-state phenomena, IR spectra of
solutions of 2 and 3 in chloroform were also recorded (not
shown). In these spectra the bands of 2 and 3 were equally
broad, a few bands occurred at different positions, and a few
additional bands were present in the spectrum of 3. Although
3546
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3544 3554

Cross-Conjugated p-Phenylenevinylidene Oligomers
3544 3554
Table 2. Compilation of most important infrared and Raman signals of
oligomers 1 4. All data are in cmÀ1
.
1
2
3
4
IR Raman IR Raman IR Raman IR Raman Assignment[a]
[b]
À
3054 3058
1609 1607
15731571
1490
3052 3057
1604 16031610630316106303
3030 3056
3031 3056
1491
n C H
ˆ
n C C
n8b
[c]
1572
1572
1490
1490
n19a
14431441
1327
1026
14431443
n19b
1327
1025
1323
1026
1323
1026
998
n14
n18a
998
998
997
n12 monosubstituted
ˆ
896
906
855
774
908
855
777
908
854
778
w
CH2
À
g C H para-substituted
À
770
g C H monosubstituted
[a] n: stretch mode; w: wag mode; g: out-of-plane bending mode. The numbers
of stretch modes refer to benzene modes.[26] [b] The most intensive of a number
of peaks. [c] Overlap with the n8a mode possible.
these observations might possibly be explained by a different
point group of 3, which possesses C2 symmetry in the stretched
structure, it is more probable that the broad signals of 3 in the
solid state are caused by both solid-state phenomena and a
structural heterogeneity. It can thus not be ruled out that, in
addition to the stretched structure, other geometries are also
present. Apart from intensity differences originating from a
change of the number of phenyl end groups relative to the
number of phenylene moieties, such as differences observed
in the signals near 855 and 775 cmÀ1, the IR spectrum of 4 (not
shown) is very similar to that of 2. This renders it likely that 4
also occurs mainly in the stretched Ci structure.
Figure 2. Infrared spectra of A) 2 and B) 3 in the frequency range 2000
750 cmÀ1
.
ˆ
In both the IR and Raman spectra of 1, the C C stretch
mode is found near 1607 cmÀ1. This is a conspicuously low
frequency as revealed by a comparison with data from
À1 [27]
ˆ
compounds such as styrene (n C C ˆ 1631 cm ), a-meth-
ylstyrene (1630 cmÀ1),[27] cis-stilbene (1629 cmÀ1),[23, 28] trans-
stilbene (1638 cmÀ1), and related compounds.[29] The low
ˆ
frequency shows that the force constant of the C C stretch
motion is small, which can be tentatively interpreted to be a
consequence of a considerably delocalized electron distribu-
tion in 1.[22] However, AM1 calculations provide a different
picture. In line with the experimental data, the (unscaled) 1,1-
ˆ
diphenylethene C C stretch mode was calculated to be
situated at a lower frequency than that of, for example,
trans-stilbene (in which, in the optimized structure, the phenyl
ˆ À
rings were rotated out of the C C H plane by 21.58): 1853
À1
ˆ
versus 1877 cm . Also, the AM1-calculated C C stretch force
constant in 1, 7.28 mdynäÀ1, is smaller than that in trans-
stilbene, 7.59 mdynäÀ1. Interestingly, these force constants
are not consistent with the bond orders. In 1 these are 1.884
ˆ
and 0.996 for the C C bond and the adjacent formal single
bonds, respectively. In trans-stilbene the corresponding bond
orders were 1.845 and 1.030. The bond orders thus indicate
ˆ
that the C C bond is stronger in 1 than in trans-stilbene,
whereas the force constants indicate the opposite. This
apparent discrepancy can be understood by considering the
(p) bond order, prs, between two atoms r and s, which is
defined as in Equation (1).
X
prs
ˆ
nicricsi
(1)
Figure 3. FT-Raman spectra of A) 1 and B) 4.
i
Chem. Eur. J. 2003, 9, 3544 3554
www.chemeurj.org
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3547

C. A. van Walree et al.
FULL PAPER
Here cri and csi represent the eigenvectors of r and s in the ith
molecular orbital, which is occupied by ni electrons. Accord-
ing to Equation (1) the bond order reflects the number of
electrons present at r and s; it does not however necessarily
reflect the degree of bonding. This bonding appears to be less
ˆ
effective in 1 than in trans-stilbene, since the two C C
eigenvectors in the highest occupied molecular orbital
(HOMO) in 1 (see Figure 6 below) are markedly different,
whereas those in the HOMO of trans-stilbene are equal,
giving rise to efficient overlap. The unequal p coefficients at
ˆ
the two C C atoms in the HOMO of 1 are caused by the
asymmetric substitution pattern. The eigenvector at the
olefinic carbon atom bonded to the phenyl groups is much
smaller because of the interaction with these groups. Hence,
ˆ
Figure 4. Relative intensities of the Raman C C stretch mode at
À1
1604 cmÀ1 (solid circles) and C H stretch mode near 3060 cm (open
À
ˆ
although the low C C stretch frequency of 1 suggests that in
squares) of 1 4 as a function of chain length. Intensities were normalized
to the n12 mode of the phenyl groups at 997 998 cmÀ1 and were determined
by measuring the peak height. The drawn line is a guide to the eye.
the basic building block of the oligo(p-phenylenevinylidene)s
strong electron delocalization between the double bond and
the phenyl groups occurs, in reality the low frequency is the
result of the asymmetric nature of the double bond. As far as
ˆ
we are aware, this is the first time that this low C C stretch
ˆ
À
feature that at n ˆ 1 the C C intensity is lower than the C H
frequency of 1 has been rationalized in this way.
Table 2 shows that in both IR and Raman spectra, extension
of the structure on going from 1 to 2 leads to a small shift of
ˆ
intensity. These observations suggest that the polarizability of
ˆ
the C C bond changes appreciably in going from 1 to 2,
whereas upon moving to the longer compounds only additive
effects are operative. Conjugation effects are thus only
present when going from n ˆ 1 to n ˆ 2, the length at which
the conjugation plateau seems to be reached already. This is
however not in agreement with results presented below. Since
the IR and Raman frequencies were also chain-length
independent, it must be assumed that vibrations are local in
nature. Since a similar conclusion was also drawn for p-
phenylenevinylene oligomers,[34] this is likely to be true in
general for conjugated materials incorporating phenylene
units.
4
5 cmÀ1 of the C C stretch mode towards lower wave-
numbers. Although one may be initially inclined to think that
this reflects a small increase in electron delocalization, it is
due rather to the fact that the vibrational spectra of 1 were
obtained in the liquid phase and those of 2 in the solid phase.
ˆ
In the solid state (at 80 K) the C C stretch mode of 1 is
situated at 1604 cmÀ1 in the Raman spectrum and at 1606 cmÀ1
in the IR spectrum.[25] The CH2 wag frequency of 1 differs
ˆ
from that of 2 for the same reason; in the solid state it occurs
at 905 cmÀ1.[25] It then appears by inspection of Table 2 that
throughout the whole series of compounds all major vibra-
tions are found at essentially the same frequency. This implies
that either the degree of conjugation does not change in going
from 1 to 4 or that the frequencies of the vibrations are not
susceptible to the degree of conjugation. The latter explan-
ation would have the consequence that all vibrations taken
Cyclic voltammetry and orbital energies: To obtain informa-
tion about the effect of the oligomer length on the redox
properties and orbital energies, compounds 1 3 were studied
by cyclic voltammetry. No cyclic voltammogram could be
obtained for 4 owing to its poor solubility. As shown in
Figure 5 for 1, the oligomers display two highly irreversible
oxidation waves, which correspond to the formation of the
radical cation and the subsequent conversion into the
ˆ
into account, including the C C stretch mode, have a local
character. In this context it is noteworthy that only extremely
ˆ
small frequency shifts (particularly of the C C stretch
vibration) have been found in Raman data of oligomeric
model compounds of poly(p-phenylenevinylene).[30, 31]
Since they reflect polarizability phenomena, Raman inten-
sities are in principle better suited to probe conjugation
phenomena than IR and Raman frequencies.[32] This concerns
ˆ
C C stretching modes in particular. In Figure 4 the intensities
ˆ
À
of the C C stretch signals, together with those of the C H
stretching mode, are plotted against the chain length. The
intensities were normalized to the n12 mode of the phenyl end
groups, which represents the most appropriate reference
signal. Starting at n ˆ 2 it is seen that the intensities of the
ˆ
À
C C and C H vibrations increase linearly with the number of
vinylidene and phenylene groups. The steeper slope of the
ˆ
C C signal obviously originates from the larger polarizability
of this group. However, when the region between the origin
ˆ
and n ˆ 2 is considered, the relationship between the C C
Figure 5. Cyclic voltammogram of 1,1-diphenylethene (1). The potential is
given relative to the standard calomel electrode (SCE).
intensity and n is superlinear.[33] This is exemplified by the
3548
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3544 3554

Cross-Conjugated p-Phenylenevinylidene Oligomers
3544 3554
dication. Reduction processes were not observed in the
available potential window.
As can be seen from Table 3, the first oxidation potential
becomes lower with the oligomer length. This implies that the
HOMO energy is raised and the radical cation becomes more
stable as a function of chain length. It thus appears that the
Table 3. First and second oxidation potential (in V versus SCE) of
oligomers 1 3 as determined by cyclic voltammetry.[a]
Compound
Eo1x
Eo2x
1
2
3
1.88[b]
1.72
1.67
2.17
1.90
1.8[c]
[a] Owing to the irreversibility of the oxidation processes, peak maxima
are given as the potentials. [b] Value in agreement with literature data.[35]
[c] Value could not be determined accurately because of low solubility.
oligo(p-phenylenevinylidene)s are conjugated in the sense
that electron delocalization occurs upon chain elongation,
thereby raising the HOMO level. The decrease in the first
oxidation potentials is 0.16 and 0.05 V when going from 1 to 2
and from 2 to 3, respectively. These steps are not as large as
observed for linear p-conjugated compounds. For instance, for
a series of phenyl end-capped oligothiophenes the first steps
in oxidation potential with n are 0.28 and 0.18 V,[36] while for
the reduction of oligo(p-phenylenevinylene)s the steps are
0.26 and 0.14 V.[37] The smaller step size indicates that the
electronic interaction between building blocks in 1 3 is not as
pronounced as in linear p-conjugated compounds. The second
oxidation potential decreases with the chain length as well. In
Figure 6. PPP/SCF-calculated HOMOs and LUMOs of compounds 1 4,
along with their energies. Circle radii are proportional to the size of
coefficients.
is close to reaching its limit in the system with four double
bonds. Not surprisingly, these observations are qualitatively
reflected by the nature of the orbitals. The HOMOs and
LUMOs have identical eigenvectors (there is only an addi-
tional nodal plane in the LUMO), and this accounts for the
equal effect of chain extension on the HOMO and LUMO. In
addition, the HOMOs and LUMOs tend to be localized in the
center of the compounds. For 3 and 4 the contribution of
p-type atomic orbitals in the terminal phenyl groups to the
frontier orbitals is already small. This rationalizes the fact that
attachment of another unit only has a limited effect.
The AM1 calculations yielded the same picture with regard
to the orbital energies and constitutions as the PPP/SCF
calculations. As an illustration, AM1 gave HOMO energies of
À8.89, À8.69, and À8.62 eV and LUMO energies of À0.015,
À0.21, and À0.28 eV for 1, 2 (Ci), and 3 (C2), respectively.
38]
common with linear p-conjugated systems,[36 the decrease
is faster than that of the first oxidation potential. Unfortu-
nately, the limited number of data points did not allow a
reliable determination of the coalescence point of the two
oxidation processes.
The cyclic voltammetry results are supported by PPP/SCF
and AM1 calculations. HOMO and lowest unoccupied
molecular orbital (LUMO) plots of 1 3 obtained from the
PPP calculations, along with the energies, are shown in
Figure 6. It is important to mention that in all compounds the
double bonds and aromatic rings together form a single
delocalized system. However, the frontier orbitals are quite
ˆ
strongly localized on the double bonds. In particular the CH2
carbon atoms contribute appreciably. A similar pattern is
exhibited by the frontier orbitals of dendralenes.[39] Also, the
phases of the eigenvectors at the double bonds behave in
exactly the same way as in dendralenes: in the HOMO the
Electronic spectra: UV spectra of 1 4 are depicted in
Figure 7, while maxima and absorption coefficients are
compiled in Table 4. To provide insight into the spectra,
PPP/SCF-calculated absorption maxima, oscillator strengths,
and configuration interaction data are listed in Table 5. The
results of these PPP calculations are in excellent agreement
with the experimental data.
The spectra are complicated by the superposition of many
transitions. For example, in line with previously reported
calculations,[40] the PPP data point out that in the spectrum of
1 two forbidden transitions are situated at lower energy than
the HOMO LUMO transition. These transitions appear as
the absorption onset in the region 270 295 nm. For the longer
homologues these transitions are almost completely obscured
by the HOMO LUMO transition. Furthermore, the inter-
ˆ
C C units are p bonding but have an alternating phase with
ˆ
ˆ
neighboring C C units, and in the LUMO the ™C C∫ units
are p antibonding while having the same phase as their
neighbors. This makes it possible to regard oligo(p-phenyl-
enevinylidene)s as dendralenes in which the double bonds are
connected by phenylene spacers.
The PPP/SCF calculations nicely show that the HOMO is
destabilized by chain extension, while the LUMO is stabilized
by the same amount. However, as the chain becomes longer
the effect is gradually dampened. When moving from 3 to 4
the difference is only 0.03eV, which indicates that conjugation
Chem. Eur. J. 2003, 9, 3544 3554
www.chemeurj.org
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3549

C. A. van Walree et al.
FULL PAPER
study the HOMO LUMO transitions are of principal
interest. They are found at 256 (1), 271 (2), 272 (3), and
271.5 (4) nm. This shows that there is a moderate red shift in
going from 1 to 2, whereas surprisingly no further change
occurs when the number of repeating units increases. At first
sight, this suggests that conjugation in the oligo(phenylene-
vinylidene) series 1 4 is limited to only two units. However,
from the cyclic voltammetry data and the PPP-generated
orbitals, it appeared that in the series 1 4 the HOMO and
LUMO energies change until n ˆ 4. It will be shown below
that in this case the HOMO LUMO transition energy is not
a measure of the degree of conjugation. Stronger delocaliza-
tion than anticipated from the UV maxima is also indicated by
the intensity of the HOMO LUMO transition, which is
indicative of the occurrence of conjugation and which
increases progressively with the chain length (Table 4).[41]
Fluorescence spectra of the oligomers initially show a red
shift with increasing length (Figure 8, Table 4) but saturation
Figure 7. UV spectra of 1 4 in cyclohexane. Spectra were normalized at
the low-energy maxima and were offset vertically in going from 1 to 4. See
Table 4 for absorption coefficients.
Table 4. Electronic absorption and fluorescence data for oligomers 1 4.
[a]
[c]
[d]
lmax
e[b]
lfl
Ffl
1
2
3
4
292.5, 283.0, 256.0, 233.5
284.0, 271.0, 251.5, 240.0
296.5, 272.0, 248.0
271.5, 255[f]
10.4
18.5
311[e]
336
361
0.0012
0.0015
28.8
0.0016
[g]
[g]
361
[a] Absorption maxima in nm, determined from the second derivative
spectrum. Only transitions at wavelengths higher than 230 nm are reported.
[b] Absorption coefficient, in units of 103 mÀ1 cmÀ1, at 250 (1) and 270
(2, 3) nm. [c] Fluorescence maximum in nm. [d] Fluorescence quantum
yield. [e] Corrected value, see Experimental Section. [f] Value subject to
error because of limited signal-to-noise ratio. [g] Not determined due to
limited solubility.
pretation of the spectra is complicated by a change in the
sequence of orbitals and the different structures of the
compounds. This is illustrated by the transition predominantly
built up from the HOMO ! LUMO ‡ 1 (1 2') and HOMO
À 1 ! LUMO (2 1') excitations, which is forbidden for the Ci
compounds 2 and 4, but carries considerable oscillator
strength in the C2 species 1 and 3. Nevertheless, in the present
Figure 8. Fluorescence spectra of
response) and 2 4 (corrected) in cyclohexane at room temperature.
1 (not corrected for the detector
is reached for compounds 3 and 4. In other words, the first
excited state of the long compounds is better stabilized.
Furthermore, within the series
there is a difference in the
appearance of the fluorescence
spectrum. Whereas the fluores-
Table 5. PPP/SCF-calculated electronic transitions of oligomers 1 4.[a, b]
l [nm]
f
CI[c]
cence spectrum of 1 is struc-
1
2
3
4
262
262
242
226
0.00
0.00
0.33
0.51
0.54(1 3') ‡ 0.54(3 1 ') ‡ 0.43(2 4') ‡ 0.43(4 2')
tureless, the spectra of the high-
er homologues exhibit a distinct
vibrational splitting, which be-
comes more prominent with
extension of the structure.
From the approximate spacing
of 1600 cmÀ1 it appears that the
electronic transition is coupled
0.54(1 4') ‡ 0.54(4 1') ‡ 0.43(2 3') ‡ 0.43(3 2')
0.98(1 1')
0.69(1 2') ‡ 0.69(2 1')
264
256
238
231
0.00
0.61
0.00
0.85
À 0.62(1 6') ‡ 0.62(6 1') À 0.28(3 6 ') À 0.28(6 3')
0.94(1 1')
À 0.68(1 2') ‡ 0.68(2 1')
0.63(2 2') À 0.51(1 3') ‡ 0.51(3 1 ')
255
2530.47
235
0.91
0.92(1 1') ‡ 0.26(2 2')
0.66(1
0.27
1.15
2
') ‡ 0.66(2 1') ‡ 0.24(1 4') ‡ 0.24(4 1')
0.61(2 2') À 0.49(1 3') ‡ 0.49(3 1 ')
0.55(2 3') À 0.55(3 2 ') ‡ 0.40(1 4') ‡ 0.40(4' 1)
ˆ
to the C C stretch vibration
232
(compare with Table 2). This is
in line with the large contribu-
255
254
252
232
1.15
0.00
1.17
1.59
0.87(1 1') ‡ 0.26(1 3')
À 0.65(1 2') ‡ 0.65(2 1')
ˆ
tion of the C C pp orbitals to
À 0.61(2 2') À 0.47(1 3') ‡ 0.47(1 3') ‡ 0.27(1 5') À 0.27(5 1')
À 0.50(3 3 ') ‡ 0.46(2 4') ‡ 0.46(4 2') À 0.31(1 5') ‡ 0.31(5 1')
the HOMO and LUMO (Fig-
ure 6). It is noteworthy that
under the structureless enve-
lope of the emission of 1 a
À
[a] Calculations were run on geometries with torsion angles of 408 about the C C single bonds. [b] Not all
transitions of zero oscillator strength situated at lower wavelengths than the 1 1' transition have been listed.
[c] Configuration interaction coefficients larger than 0.20 are given.
3550
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3544 3554

Cross-Conjugated p-Phenylenevinylidene Oligomers
3544 3554
vibrational substructure with spacing of 1600 cmÀ1 is hid-
den,[40] which shows that the fluorescence of this compound is
associated with the LUMO HOMO transition as well. In
going to longer compounds in particular, the 0 0 emission
gains intensity. This has been observed previously for linear p-
conjugated oligomers,[36, 42] and it indicates that the (relaxed)
S1 geometry becomes more similar to the ground state. Thus,
although the excited states of 3 and 4 were found to be
situated at equal energies, they must be somewhat shifted
the longer compounds, in fact, overcompensated for) by a
decrease of the stabilizing effect of the Coulomb integral. The
decrease of the exchange integral in its turn has a stabilizing
contribution on the excited state, but this not sufficient to
cancel the destabilizing effect of the Coulomb term. Hence,
although the HOMO LUMO gap decreases with the
oligomer length, the HOMO LUMO transition energy is
not lowered since these molecular orbitals become more
extended and electron repulsion is reduced. Recently, the
Coulomb and exchange terms have been addressed to explain
the chain-length dependence of the electronic spectra of
electron donor acceptor substituted p-phenylenevinylene
oligomers.[44]
The situation outlined in the last paragraph has some
important consequences. Firstly, despite being very frequently
applied as a measure of the HOMO LUMO gap and the
extent of conjugation of a system, electronic spectra are not
unambiguous in this use. This should also hold for linear p-
conjugated systems, in which the Coulomb integral must also
decrease with the chain length. However, here the reduction
of the HOMO LUMO gap with n is usually much stronger,
so that the role of the Coulomb term is less prominent.
Nevertheless, in a given system it is more appropriate to
consider the HOMO and LUMO energies to assess the extent
of conjugation. Secondly, the quite unique situation exists in
cross-conjugated materials that extension of the conjugated
system is not accompanied by a bathochromic shift of the
optical transition. Although it may be premature to state that
is true for all cross-conjugated systems, we note that chain
extension in dendralene-type compounds also leads to
changes in redox potentials but not to significant UV/Vis
shifts.[14, 18]
ˆ
along the C C stretching coordinate with respect to each
other.
Fluorescence quantum yields are low, of the order of 0.0015,
and of comparable magnitude for all compounds. The latter
observation supports the fact that for all compounds the same
type of excited state is involved in the fluorescence process,
despite the fact that in the absorption spectra several
transitions were found to be situated closely together. The
low values of the quantum yields strengthen the idea that the
first excited state is localized at the olefinic bond(s), as excited
states of alkenes are known to deactivate rapidly.
Correlation of HOMO LUMO transitions and orbital en-
ergies: As described above, the HOMO LUMO transition
energies in the UV spectra seem to provide a different
conjugation picture to that suggested by the HOMO and
LUMO energies disclosed by cyclic voltammetry and PPP/
SCF calculations. The orbital energies continue to change
until n ˆ 4, whereas in the UV spectrum saturation is already
observed at n ˆ 2. It should however be realized that in
addition to being based on the HOMO LUMO gap DEH
,
L
the energy of the singlet excited state formed upon HOMO
LUMO excitation, Es, receives contributions from the Cou-
lomb integral JH L and the exchange integral KH L, according
to Equation (2).[43]
Conclusion
Es
ˆ
DEH L À JH ‡ 2KH
(2)
L
L
It has not been easy to obtain a coherent view on the degree of
electron delocalization in oligo(p-phenylenevinylidene)s 1 4,
since techniques which are frequently applied for the evalua-
tion of conjugation effects do not provide unequivocal
information. For instance, whereas the low frequency of the
The last two terms account for the decrease in repulsion
energy when an electron is promoted from the HOMO to the
LUMO. Data for 1 4, extracted from the PPP/SCF calcu-
lations (see Experimental Section), are collected in Table 6. It
is seen that the energy of the singlet state Es decreases from 1
to 2, but subsequently increases. As expected from the data in
Figure 6, the HOMO LUMO gap decreases continuously
with n. The same trend is followed by the Coulomb and
exchange integrals; this is in line with the expectation that in
more extended systems the electronic repulsion is smaller. It
now appears that within the series 1 4 the lowering of the
excited state by narrowing of the band gap is opposed (and for
ˆ
C C stretch mode in 1 suggests that electrons in the double
bonds are strongly delocalized, this frequency is actually due
to the asymmetric substitution pattern. When studied as a
function of the oligomer length neither vibrational peak
positions nor intensities give a reliable picture, since modes
have a local character. Moreover, UV spectra can be
interpreted by the conjugation effectively reaching its limit
at n ˆ 2, but in reality they indicate that the decrease of the
HOMO LUMO gap is accompanied by a relatively strong
reduction in electron repulsion.
Table 6. Singlet energies Es, energy gaps DEH L, Coulomb integrals JH
,
L
and exchange integrals KH of the HOMO LUMO transition of
Despite the fact that the spectroscopic data should be
treated with some caution, they clearly reveal that the largest
changes in the series occur when going from 1 to 2. In this step,
a pronounced red shift of the UV and fluorescence maxima
L
oligomers
1 4 extracted from PPP/SCF calculations. Configuration
interaction effects are not included. All data are in eV.
ES
DEH
JH
KH
L
L
L
ˆ
1
2
3
4
5.22
5.01
5.08
5.18
7.89
7.47
7.31
7.25
4.49
3.75
3.27
2.930.43
0.91
0.65
0.52
takes place, and a superlinear increase of the C C Raman
intensity is found. These results point at a substantial
extension of the delocalized system. In this connection it
should be realized that 2 can be considered as consisting of a
Chem. Eur. J. 2003, 9, 3544 3554
www.chemeurj.org
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3551

C. A. van Walree et al.
FULL PAPER
FP5/FP51 photoelectric apparatus. HPLC was performed by using a
Hypersil ODS (c18) 5u column in combination with a Thermo Separation
Products Series 200 pump and UV detector (monitoring at 264 nm). For gas
chromatography a Varian 3400 gas chromatograph equipped with a DB5
capillary column was used. GC-MS spectra were collected on an
ATI UNICAM Automass System 2 quadropole mass spectrometer (elec-
tron ionization (EI), 70 eV), while probe-analysis mass spectra were
obtained on a Jeol JMS-AX505W mass spectrometer (EI). Elemental
analysis was carried out at Kolbe Microanalytisches Laboratorium in
M¸lheim an der Ruhr, Germany.
central phenylene group with two linearly p-conjugated
double bonds attached; it is somewhat reminiscent of
divinylbenzene. This is also suggested by the X-ray crystal
structure of 2, in which quinoid character is only present in the
phenylene ring and not in the outer phenyl groups. Con-
sequently, in common with other systems,[14, 18, 45] electron
delocalization by linear p conjugation is favored over
delocalization by cross-conjugation.
Nevertheless, in the complete series 1 4 conjugation
extends beyond the divinylbenzene moiety. This is substanti-
ated by a number of features. The frontier molecular orbitals
are delocalized over the entire structure, although they have
the largest contributions from the central part of the
compounds. Simultaneously the HOMO and LUMO energies
vary systematically with the chain length; this is supported by
the oxidation potentials. Furthermore, the AM1 data suggest
a small degree of quinoid character to be present in all
phenylene moieties of the longer compounds. Thus, in
general, the electronic system of oligo(p-phenylenevinyli-
dene)s resembles that of linear p-conjugated oligomers.
However, in a quantitative sense conjugation is moderate
and the electronic interaction between repeating units is not
very strong. In the longest homologue 4, which contains only
four double bonds, the electron delocalization is already close
to its limits. The change of the HOMO and LUMO levels with
n is quite small, so that the reduction of electron repulsion is
relatively strong and optical transitions are virtually chain-
length independent. In this connection it is worth mentioning
that a theoretical model showed that among all acyclic p
systems the 1,1-disubstituted ethene topology leads to a
maximal HOMO LUMO gap.[39] The feature that the
extension of the conjugated system is not accompanied by a
bathochromic shift of electronic transitions renders cross-
conjugated materials interesting candidates for optoelectronic
applications that rely on transparency in the visible region,
such as nonlinear optical materials.
UV/Vis spectra were recorded on a Varian Cary 1 UV-Vis or a Varian
Cary 5 UV/Vis-NIR spectrophotometer. Fluorescence spectra were ob-
tained by using a Spex fluorolog instrument, with excitation wavelengths of
250, 265, 265, and 267 nm for compounds 1, 2, 3, and 4, respectively.
Emission spectra were corrected for the spectral response of the detector
by the use of a correction factor file provided by the manufacturer. Since
this file does not cover wavelengths below 300 nm, the fluorescence
emission spectrum of 1 was not corrected. Correction of part of the
spectrum gave a maximum of 311 nm. Fluorescence quantum yields[48] were
determined with naphthalene in cyclohexane (Ffl ˆ 0.23) as the reference.
Solutions were purged with argon for at least ten minutes. When necessary,
fluorescence samples were purified by HPLC. Both UV-Vis and fluores-
cence spectra were recorded in spectrophotometric-grade cyclohexane
(Acros). Cyclic voltammetry measurements were performed by using an
EG&G Princeton Applied Research Model 263A potentiostat/galvanostat
at a scan rate of 50 mVsÀ1. Solutions of 1 3 (<0.5 gLÀ1) in acetonitrile
(freshly distilled from CaH2) were used, with 0.1m anhydrous tetrabuty-
lammonium hexafluorophosphate (Fluka, electrochemical grade) as the
electrolyte. Prior to measurement, the solution was purged with nitrogen
for at least ten minutes. Oxidation potentials were measured relative to a
Ag/AgNO3 (0.1 m) electrode, with Pt working and counter electrodes. By
determining the potential of the ferrocene/ferrocinium couple (E1/2
ˆ
‡0.31 V versus SCE), oxidation potentials were calibrated to the SCE
electrode. Since we were primarily interested in the variation of the
electrochemical behavior within the series
1 4, peak values of the
irreversible waves were taken as the oxidation potential. IR spectra were
measured on a Perkin Elmer 2000 FT-IR spectrometer, operating with
medium apodization at 4 cmÀ1 spectral resolution, in combination with a
Golden Gate Single Reflection Diamond ATR system. FT-Raman spectra
were recorded with a Perkin Elmer System 2000 NIR FT-Raman spec-
trometer equipped with a Nd:YAG laser (1064 nm) and an InGaAs
detector. The Raman spectra were collected with a laser power of 300 mW
at medium apodization and a spectral resolution of 4 cmÀ1. The IR and
Raman spectra were averaged from 16 scans.
A final point to be addressed is how far properties of the
oligo(p-phenylenevinylidene)s are representative for other
types of cross-conjugated compounds. On one hand, based on
other reports dealing with electronic properties of cross-
conjugated compounds, which generally also fail to exhibit
strong shifts of HOMO LUMO transitions,[12, 14, 18] we are
inclined to think that most cross-conjugated systems behave
similarly to oligo(p-phenylenevinylidene)s. On the other
hand, it should be realized that there are two factors which
have an unfavorable effect on the conjugation in oligo(p-
phenylenevinylidene)s. These are the nonplanar structure[46]
and the high resonance energy of the incorporated aromatic
moieties.[38] Although, for instance, dendralenes are also not
planar,[47] somewhat stronger interactions may be expected for
other classes of cross-conjugated systems.
Calculations: PPP/SCF/CI calculations were performed with a home-
adapted version of Griffiths× program.[49] Torsion angles of 408 around the
À
C C formal single bonds were introduced by correction of the bond
resonance energy with cos408. For 1 a full configuration interaction was
performed, while for the other compounds the configuration interaction
included states arising from excitations from any of the six highest occupied
to any of the six lowest unoccupied levels. Exchange integrals K and
Coulomb integrals J were obtained by use of Equation (2) and Equa-
tion (3) in which ET is the energy of the triplet state.
2K
ˆ
ES À ET
(3)
AM1 calculations were run with the MOPAC7 package.[50] Geometries
were optimized with the eigenvector-following routine. In order to obtain
the desired symmetry, symmetry relations were imposed on the dihedral
angles around the C(phenyl/phenylene) C(vinylidene) single bonds. Force
calculations were run to establish that the obtained geometries were
minima on the potential energy surface and to calculate vibrational spectra.
X-ray crystallography: X-ray intensities were collected on a Nonius
KappaCCD diffractometer with rotating anode and MoKa radiation
(graphite monochromator, l ˆ 0.71073ä). An absorption correction was
not considered necessary. The reflections were merged by using the
program SORTAV.[51] The structure was solved with direct methods by
using the program SIR97[52] and refined with the program SHELXL[53]
against F 2 of all reflections. Non-hydrogen atoms were refined freely with
anisotropic displacement parameters. All hydrogen atoms were located in
the difference Fourier map and were refined as rigid groups. The drawings,
Experimental Section
General: NMR spectra were recorded on Bruker AC 300 or Varian Unity
Inova spectrometers (both operating at 300 MHz for 1H and 75 MHz for
13C NMR spectroscopy). Chemical shifts are given in ppm relative to
internal tetramethylsilane. Melting points were determined on a Mettler
3552
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3544 3554

Cross-Conjugated p-Phenylenevinylidene Oligomers
3544 3554
structure calculations, and checking for higher symmetry were performed
with the program PLATON.[54] Crystal data and experimental details are
listed in Table 7. CCDC-198762 contains the supplementary crystallo-
graphic data for this paper. These data can be obtained free of charge via
www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crys-
tallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax:
(‡44)1223-336033; or deposit@ccdc.cam.ac.uk).
was obtained as an off-white solid (0.14 g, 0.50 mmol, 6%); m.p. 1358C
(lit.:[55] 123 124, 131.5, or 1388C); 1H NMR (CDCl3): d ˆ 7.40 7.32 (m,
2
ˆ
10H; Ar-H), 7.31 (s, 4H; Ar-H), 5.51 (d, J(H,H) ˆ 1.2 Hz, 2H; CH2), 5.46
(d, 2J(H,H) ˆ 1.2 Hz, 2H; CH2) ppm; 13C NMR (CDCl3): d ˆ 149.7
ˆ
ˆ
ˆ
(C CH2), 141.4, 140.8, 128.3, 128.2, 128.0, 127.7, 114.2 (C CH2) ppm; IR:
nÄ ˆ 3092, 3052, 1818, 1604, 1571, 1506, 1490, 1441, 1327, 1025, 906, 855,
‡
774 cmÀ1; GC-MS (EI): m/z: 282 [M ].
1,1-Bis(4-bromophenyl)ethanol (6a): A solution of 1,4-dibromobenzene
(3.00 g, 12.7 mmol) in diethyl ether (50 mL) was placed in a two-necked
round-bottomed flask. n-Butyllithium in n-hexane (7.9 mL of a 1.60 m
solution, 12.6 mmol) was added over 10 min. After 30 min, a solution of
4-bromoacetophenone (2.53g, 12.7 mmol) in diethyl ether (20 mL) was
introduced. Subsequently the reaction mixture was stirred for 12 h at room
temperature. The solution was poured into water (200 mL), the layers were
separated, and the water layer was extracted with diethyl ether (4 Â
75 mL). The combined organic layers were washed with water (100 mL),
dried on MgSO4, filtered, and evaporated to dryness. Purity was 95%
(GC); further purification was not considered necessary. Compound 6a was
Table 7. Crystal data for 2 and experimental details.
formula
formula weight
crystal system
space group
a [ä]
b [ä]
c [ä]
V [ä]
C22H18
282.36
orthorhombic
Pbca (no. 61)
10.7113(3)
7.4158(2)
19.3568(6)
1537.57(8)
4
1.220
0.069
600
Z
1
obtained as a yellow solid (4.43g, 12.4 mmol, 98%); m.p. 76 8C; H NMR
1 [gcmÀ3
]
(CDCl3): d ˆ 7.44, 7.26 (AA'BB', 3J(H,H) ˆ 8.7 Hz, 2  4H; Ar-H), 2.12 (s,
1H; OH), 1.90 (s, 3H; Me) ppm; 13C NMR (CDCl3): d ˆ 146.5, 131.4, 127.6,
m (Mo-Ka) [mmÀ1
F(000)
]
À
121.3, 75.6 (C OH), 30.7 (Me) ppm; IR: nÄ ˆ 3563, 2976, 2928, 1904, 1678,
crystal size [mm]
0.03 Â 0.12 Â 0.30
1590, 1483, 1398, 1076, 1007, 916, 833 cmÀ1
.
data collection:
T [K]
qmin, qmax [o]
data set (h, k, l)
measured data
unique data
1,1-Bis[4-(1-phenylvinyl)phenyl]ethene (3): Compound 6a (2.49 g,
7.00 mmol) was dissolved in diethyl ether (60 mL) and placed in a two-
necked round-bottomed flask. This mixture was stirred and n-butyllithium
in n-hexane (14.5 mL of a 1.6 m solution, 23.2 mmol) was added over
15 min. Stirring was continued and after 1.5 h a solution of acetophenone
(2.8 g, 23.1 mmol) in diethyl ether (30 mL) was slowly introduced to the
white suspension. The mixture was stirred for an additional 12 h at room
temperature and poured into water (200 mL). The water layer was
extracted with diethyl ether (4 Â 75 mL). The combined organic layers
were washed with water (100 mL), dried over MgSO4, and filtered.
Evaporation of the solvent left a yellow solid, 7a. This compound was
dehydrated by the procedure described for 2. Evaporation of the solvent at
reduced pressure yielded a mixture of products. Volatile components were
removed by kugelrohr distillation (0.04 mbar, 50 1808C). The residue was
further purified by column chromatography (with hexane/CHCl3 (1:1) as
the eluent). Compound 3 was obtained as a yellowish solid (0.23g,
150
2.1, 23.5
0:11, 0:8, 0:21
14846
1130 (Rint ˆ 0.075)
observed data [I > 2s(I)]
845
refinement:
no. of parameters
100
2
weighting scheme [wÀ1
R1,[a] wR2[b] (obs. refl.)
R1, wR2 (all refl.)
S
]
s2(F0 ) ‡ (0.0594P)2
0.0381, 0.0882
0.0607, 0.0996
1.080
Min. and max. residual densities [eAÀ3
]
À 0.20, 0.13
2
[a] R1 ˆSjjF0 jÀjFc jj/SjF0 j. [b] wR2 ˆ[S[w(F0 ÀFc2)2]/S(w(F02)2)]1/2
.
0.60 mmol, 9%); m.p. 1658C; 1H NMR (CDCl3): d ˆ 7.37 7.34 (18H, m;
2
Synthesis: All reactions involving organolithium compounds were per-
formed in a dry nitrogen atmosphere. Diethyl ether, tetrahydrofuran
(THF), and toluene were distilled from sodium benzophenone prior to use.
Silica (Acros, 0.035 0.070 nm, pore diameter approximately 6 nm) was
used to carry out column chromatography. The purity of compounds was
established by thin-layer chromatography (Merck 60 F254 silicagel), IR
spectroscopy, 1H and 13C NMR spectroscopy, and HPLC (with methanol as
the eluent). 1,2-Diphenylethene (1) was obtained from Acros (99%) and
used as received. Commercially available reagents were used without
further purification.
ˆ
ˆ
Ar-H), 5.51 (d, J(H,H) ˆ 1.2 Hz, 2H; CH2), 5.50 (s, 2H; central CH2),
5.47 (d, J(H,H) ˆ 1.2 Hz, 2H; CH2) ppm; 13C NMR (CDCl3): d ˆ 149.7
2
ˆ
ˆ
ˆ
(C CH2), 149.4 (C CH2), 141.4, 140.9, 140.8, 128.3, 128.2, 128.1, 128.0,
ˆ
127.8, 114.3(C CH2) ppm; IR: nÄ ˆ 3030, 1818, 1658, 1603, 1572, 1507, 1490,
1443, 1400, 1323, 1263, 1072, 1026, 1009, 908, 855, 777 cmÀ1; GC-MS (EI):
‡
m/z: 384 [M ]; elemental analysis: calcd (%) for C30H24 (384.5): C 93.71, H
6.29; found: C 93.58, H 6.22.
1,4-Bis[1-hydroxy-1-(4-bromophenyl)ethyl]benzene (6b): This compound
was prepared from 1,4-dibromobenzene (5.00 g, 21.2 mmol), n-butyllithium
in n-hexane (13.2 mL of a 1.6 m solution, 21.1 mmol), and 1,4-diacetylben-
zene (1.72 g, 10.6 mmol in diethyl ether/THF (8:2)) as described for 6a.
Compound 6b was obtained as a brown solid (purity 91% by GC) in
1,4-Bis(1-phenylvinyl)benzene (2): Firstly, 1,4-dilithiobenzene was pre-
pared according to a modification of the procedure described by Brandsma
and Verkruijsse.[19] In a two-necked round-bottomed flask 1,4-dibromo-
benzene (2.10 g, 8.90 mmol) was dissolved in THF (40 mL). The stirred
solution was cooled to À608C and n-butyllithium in n-hexane (13.1 mL of a
1.42 m solution, 18.6 mmol) was added over 10 min. After 2 h a cold
(À608C), dry suspension of CeCl3 (6.56 g, 26.7 mmol) in THF was added to
the white suspension. The reaction mixture was stirred for 1 h at À608C,
after which a solution of acetophenone (2.14 g, 17.8 mmol) in THF (10 mL)
was introduced over 5 min. The mixture was allowed to warm to room
temperature overnight and was poured into water (200 mL). The aqueous
system was extracted with diethyl ether (4 Â 100 mL) after which the
combined organic layers were washed with water (100 mL), dried over
MgSO4, and filtered. Evaporation of the solvent left the yellow solid 5,
which was not pure and was not fully characterized. The crude 5 was
dissolved in toluene (150 mL) and placed in a Dean-Stark apparatus. A
catalytic amount of p-toluenesulfonic acid was added and the mixture was
boiled for 12 h. Evaporation of the toluene at reduced pressure left a yellow
solid, which proved to be a mixture of products. Compound 2 was isolated
by column chromatography (with chloroform as the eluent) and further
purified by crystallization from methanol (100 mL) at À208C. Compound 2
1
quantitative yield; m.p. 1308C; H NMR (CDCl3): d ˆ 7.43, 7.25 (AA'BB',
3J(H,H) ˆ 8.9 Hz, 2  4H; Ar-H), 7.31 (s, 4H; Ar-H), 2.32 (s, 2H; OH), 1.89
(s, 6H; Me) ppm; 13C NMR (CDCl3): d ˆ 146.9, 146.3, 131.2, 127.6, 125.7,
À
À
121.3, 75.6 (C OH), 30.7 (Me) ppm; IR: nÄ ˆ 3426 (O H), 2976, 1676, 1507,
1076, 1008, 1016, 827 cmÀ1
.
1,4-Bis{1-[4-(1-phenylvinyl)phenyl]vinyl}benzene (4): First, 7b was pre-
pared from 6b (0.40 g, 0.84 mmol), n-butyllithium in n-hexane (2 mL of a
1.6 m solution, 3.2 mmol), and acetophenone (0.29 g, 2.4 mmol) as descri-
bed above for 7a. The impure 7b was not completely characterized. Crude
7b was dehydrated as described before. After evaporation of the solvent at
reduced pressure a mixture of products was obtained, from which volatile
components were removed by kugelrohr distillation (0.04 mbar, 50
1908C). The residue was further purified by column chromatography (with
hexane/chloroform (3:7) as the eluent). Compound 4 was obtained as a
white solid (27.4 mg, 0.056 mmol, 7%); m.p. 2118C; 1H NMR (CDCl3): d ˆ
2
ˆ
7.37 7.33 (22H, m; Ar-H), 5.51 (m, 6H; CH2), 5.47 (d, J(H,H) ˆ 1.1 Hz,
ˆ
2H; CH2) ppm; IR: nÄ ˆ 3031, 1817, 1603, 1572, 1505, 1491, 1443, 1400,
1323, 1119, 1073, 1026, 1014, 908, 854, 778 cmÀ1; MS (EI): m/z: 486 [M ].
‡
Chem. Eur. J. 2003, 9, 3544 3554
www.chemeurj.org
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3553

C. A. van Walree et al.
FULL PAPER
The solubility of 4 was too low to record a 13C NMR spectrum of
satisfactory signal-to-noise ratio.
[27] A. Perrier-Datin, J.-M. Lebas, J. Chim. Phys. 1979, 76, 925 933.
œ
¬
¬
œ
[28] Z. Meic, T. Suste, G. Baranovic, V. Smrecki, S. Holly, G. Kerestury, J.
Mol. Struct. 1995, 348, 229 232.
[29] M. Oelgemˆller, B. Brem, R. Frank, S. Schneider, D. Lenoir, N.
Hertkorn, Y. Origane, P. Lemmen, J. Lex, Y. Inoue, J. Chem. Soc.
Perkin Trans. 2 2002, 1760 1771.
Acknowledgement
[30] B. Tian, G. Zerbi, K. M¸llen, J. Chem. Phys. 1991, 95, 3198 3207.
[31] B. Tian, G. Zerbi, R. Schenk, K. M¸llen, J. Chem. Phys. 1991, 95,
3191 3197.
[32] M. Del Zoppo, C. Castiglioni, P. Zuliani, G. Zerbi, in Handbook of
Conducting Polymers, 2nd ed. (Eds.: T. A. Skotheim, R. L. Else-
nbaumer, J. R. Reynolds), Marcel Dekker, New York, 1998, pp. 765
822.
The authors thank A. Schouten for recording a powder X-ray diffraction
pattern and Prof. L. W. Jenneskens and Dr. T. Visser for fruitful discussions.
This work was supported in part (M.L., A.L.S.) by the Council for Chemical
Sciences of the Netherlands Organization for Scientific Research (CW-
NWO).
À
[33] It can not be ruled out that the n8a phenyl/phenylene C C stretch
mode contributes to the signal at 1603cm À1. It is anticipated that this
mode is local in character and can not be held responsible for the
observed superlinear dependence.
[1] Handbook of Conducting Polymers, 2nd ed. (Eds.: T. A. Skotheim,
R. L. Elsenbaumer, J. R. Reynolds), Marcel Dekker, New York, 1998.
[2] M. J. S. Dewar, in Modern Models of Bonding and Delocalization
(Eds.: J. F. Liebman, A. Greenberg), VCH, Weinheim, 1988, pp. 2 61.
[3] S. Winstein, Spec. Publ. Chem. Soc. 1967, 21, 5 45.
[4] R. D. Miller, J. Michl, Chem. Rev. 1989, 89, 1359 1410.
[5] F. J. Hoogesteger, C. A. van Walree, L. W. Jenneskens, M. R. Roest,
J. W. Verhoeven, W. Schuddeboom, J. J. Piet, J. M. Warman, Chem.
Eur. J. 2000, 6, 2948 2959.
[34] G. Zerbi, E. Galbiati, M. C. Gallazzi, C. Castiglioni, M. Del Zoppo, R.
Schenk, K. M¸llen, J. Chem. Phys. 1996, 105, 2509 2516.
[35] D. R. Arnold, X. Du, J. Chen, Can. J. Chem. 1995, 73, 307 318.
[36] J. J. Apperloo, L. B. Groenendaal, H. Verheyen, M. Jayakannan,
R. A. J. Janssen, A. Dkhissi, D. Beljonne, R. Lazzaroni, J.-L. Bredas,
Chem. Eur. J. 2002, 8, 2384 2396.
[6] E. P. A. M. Bakkers, A. W. Marsman, L. W. Jenneskens, D. Vanmae-
kelbergh, Angew. Chem. 2000, 112, 2385 2388; Angew. Chem. Int. Ed.
2000, 39, 2297 2299.
[7] M.-C. Fang, A. Watanabe, M. Matsuda, Macromolecules 1996, 29,
6807 6813.
[37] K. Meerholz, H. Gregorius, K. M¸llen, J. Heinze, Adv. Mater. 1994, 6,
671 674.
[38] J. Roncali, Acc. Chem. Res. 2000, 33, 147 156.
[39] P. W. Fowler, P. Hansen, G. Caporossi, A. Soncini, Chem. Phys. Lett.
2001, 342, 105 112.
[8] G. Kwak, T. Masuda, Macromolecules 2002, 35, 4138 4142.
[9] C. A. van Walree, H. Kooijman, A. L. Spek, J. W. Zwikker, L. W.
Jenneskens, J. Chem. Soc. Chem. Commun. 1995, 35 36.
[10] C. A. van Walree, M. R. Roest, W. Schuddeboom, L. W. Jenneskens,
J. W. Verhoeven, J. M. Warman, H. Kooijman, A. L. Spek, J. Am.
Chem. Soc. 1996, 118, 8395 8407.
[40] K. Gustav, M. Boelke, Z. phys. Chemie (Leipzig) 1983, 264, 950 956.
[41] For noninteracting chromophores an increase of e with n is also
expected since, in that case, identical transitions are superpositioned.
However, the PPP/SCF calculations indicate that it is indeed the
HOMO LUMO transition of the delocalized system which gains
intensity.
[11] A. GarcÌa MartÌnez, J. OsÌo Barcina, A. de Fresno Cerezo, G. Rojo, F.
[42] J. Cornil, D. Beljonne, C. M. Heller, I. H. Campbell, B. K. Laurich,
¬
¬
Agullo-Lopez, J. Phys. Chem. B 2000, 104, 43 47.
[12] N. F. Phelan, M. Orchin, J. Chem. Educ. 1968, 45, 633 637. It should
be noted that this definition is not completely unambiguous since it
can be interpreted to hold as well for linear p-conjugated systems.
[13] S. Fielder, D. D. Rowan, M. S. Sherburn, Angew. Chem. 2000, 112,
4501 4503; Angew. Chem. Int. Ed. 2000, 39, 4331 4333.
[14] M. R. Bryce, M. A. Coffin, P. J. Skabara, A. J. Moore, A. S. Batsanov,
J. A. K. Howard, Chem. Eur. J. 2000, 6, 1955 1962.
[15] R. R. Tykwinski, Chem. Commun. 1999, 905 906.
[16] Y. Zhao, R. McDonald, R. R. Tykwinski, Chem. Commun. 2000, 77
78.
¬
D. L. Smith, D. D. C. Bradley, K. M¸llen, J. L. Bredas, Chem. Phys.
Lett. 1997, 278, 139 145.
[43] J. Michl, E. W. Thulstrup, Tetrahedron 1976, 32, 205 209.
[44] H. Meier, J. Gerold, H. Kolshorn, W. Baumann, M. Bletz, Angew.
Chem. 2002, 114, 302 306; Angew. Chem. Int. Ed. 2002, 41, 292 295.
[45] N. Tyutyulkov, F. Dietz, K. M¸llen, M. Baumgarten, S. Karabunarliev,
Chem. Phys. 1994, 189, 83 97.
[46] PPP calculations on planar structures reveal the same trend as
observed for the twisted structures. For planar 1 4, calculated
HOMO LUMO transition energies are 4.89, 4.58, 4.60, and 4.66 eV,
respectively, while the HOMO LUMO gaps are 7.37, 6.85, 6.65, and
6.57 eV. This shows that the observed effects are not due to the
nonplanar geometries of 1 4.
[17] M. Br˘nsted Nielsen, M. Schreiber, Y. G. Baek, P. Seiler, S. Lecomte,
C. Boudon, R. R. Tykwinski, J.-P. Gisselbrecht, V. Gramlich, P. J.
Skinner, C. Bosshard, P. G¸nter, M. Gross, F. Diederich, Chem. Eur. J.
2001, 7, 3263 3280.
[47] H. Hopf, Angew. Chem. 1984, 96, 947 958; Angew. Chem. Int. Ed.
Engl. 1984, 23, 948 960.
[18] A. M. Boldi, J. Anthony, V. Gramlich, C. B. Knobler, C. Boudon, J.-P.
Gisselbrecht, M. Gross, F. Diederich, Helv. Chim. Acta 1995, 78, 779
796.
[19] L. B. Brandsma, H. Verkruijsse, Preparative Polar Organometallic
Chemistry, Springer, Berlin, 1987, pp. 189 190.
[20] T. Imamoto, T. Kusumoto, M. Yokoyama, J. Chem. Soc. Chem.
Commun. 1982, 1042 1044.
[21] T. Imamoto, T. Kusumoto, M. Yokoyama, J. Am. Chem. Soc. 1989, 111,
4392 4398.
[48] D. F. Eaton, Pure Appl. Chem. 1988, 60, 1107 1114.
[49] J. Griffiths, J. G. Lasch, QCPE Program QCMP054.
[50] J. J. P. Stewart, MOPAC, Version 7.0, 1993.
[51] R. H. Blessing, J. Appl. Cryst. 1997, 30, 421 426.
[52] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giaco-
vazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R. Spagna, J.
Appl. Cryst. 1999, 32, 115 119.
[53] G. M. Sheldrick, SHELXL-97, Program for Crystal Structure Refine-
ment, University of Gˆttingen, Germany, 1997.
[22] A. Perrier-Datin, J.-M. Lebas, J. Chim. Phys. 1979, 76, 935 939.
[23] A. Bree, R. Zwarich, J. Mol. Struct. 1981, 75, 213 224.
[24] Handbook of Chemistry and Physics, 82nd ed. (Ed.: D. R. Lide), CRC,
Boca Raton, FL, 2001.
[25] I. Baraldi, E. Gallinella, M. Scoponi, Spectrochim. Acta 1987, 43A,
1045 1054.
[54] A. L. Spek, Platon, a Multipurpose Crystallographic Tool, Utrecht
University, The Netherlands, 2001.
[55] Beilstein Crossfire, Database BS0202AB, Beilstein Informationssys-
teme, 2002.
¬
[26] G. Varsanyi, Vibrational Spectra of Benzene Derivatives, Academic
Press, New York, 1969.
Received: December 16, 2002 [F4671]
3554
¹ 2003Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3544 3554