Experimental evidence for carbonyl-π electron cloud interactions

Article abstract of DOI:10.1039/b608628d

In this work we present some experimental evidence of the existence of carbonyl-π electron cloud interactions. Such interactions are analogous to anion-π interactions, which have been predicted to be energetically favourable in the case of electron deficient aromatic rings. UV-Visible spectroscopy and cyclic voltammetry results obtained for 9,10-anthraquinone, 1,1′-bis-9,10-anthraquinone, poly(9,10-anthraquinone-1,4-diyl) and other 1,4-diaryl substituted anthraquinone derivatives are described. It was found that the steric hindrance occurring between the carbonyl groups and the adjacent aromatic substituent forces the plane of the anthraquinone moiety and that of the aromatic substituent to adopt a nearly orthogonal conformation, resulting in relatively strong carbonyl-π interactions that affect both the UV-Vis absorption spectrum and the reduction potential of the compound. Moreover, in the case of thiophene substituted derivatives, the torsion angle between the anthraquinone moiety and its aromatic substituent is smaller and therefore carbonyl-π interaction effects are not observed in these compounds. The Royal Society of Chemistry and the Centre National de la Recherche Scientifique.

Full text of DOI:10.1039/b608628d

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Experimental evidence for carbonyl–p electron cloud interactionsw
a
a
b
Julien E. Gautrot,* Philip Hodge,* Domenico Cupertino and
a
Madeleine Helliwell
Received (in Montpellier, France) 20th June 2006, Accepted 22nd August 2006
First published as an Advance Article on the web 12th September 2006
DOI: 10.1039/b608628d
In this work we present some experimental evidence of the existence of carbonyl–p electron cloud
interactions. Such interactions are analogous to anion–p interactions, which have been predicted
to be energetically favourable in the case of electron deficient aromatic rings. UV-Visible
0
spectroscopy and cyclic voltammetry results obtained for 9,10-anthraquinone, 1,1 -bis-9,10-
anthraquinone, poly(9,10-anthraquinone-1,4-diyl) and other 1,4-diaryl substituted anthraquinone
derivatives are described. It was found that the steric hindrance occurring between the carbonyl
groups and the adjacent aromatic substituent forces the plane of the anthraquinone moiety and
that of the aromatic substituent to adopt a nearly orthogonal conformation, resulting in relatively
strong carbonyl–p interactions that affect both the UV-Vis absorption spectrum and the reduction
potential of the compound. Moreover, in the case of thiophene substituted derivatives, the torsion
angle between the anthraquinone moiety and its aromatic substituent is smaller and therefore
carbonyl–p interaction effects are not observed in these compounds.
Introduction
conjugation occurring along the polymer backbone of 1. On
the other hand, in such a conformation the oxygen atoms of
the carbonyl groups are sufficiently close to the aromatic rings
of the neighbouring moieties as to allow interactions to occur
between the lone pairs on the oxygen atoms and the neigh-
bouring p-system, i.e. carbonyl–p interactions are a possibility.
Supramolecular structures are generally stabilized by a range
of favourable non-bonding interactions, for example, hydro-
gen bonds and the p-stacking of aromatic rings. Understand-
ing the various interactions is, amongst other things, crucial to
the development of conjugated materials for electronic appli-
cations. In this article we present experimental evidence that
supports the presence of favourable interactions between the
oxygen atom of a carbonyl group and an aromatic ring.
In the course of our studies of high electron-affinity poly-
mers we prepared a range of anthra-9,10-quinone-based poly-
mers including poly(1,4-dianthraquinoyl) (1). Previous studies
by Yamamoto et al. showed that the highest transition in the
UV-Vis spectrum of 1 is shifted to lower energies compared to
1
that of anthra-9,10-quinone (2). They attributed this shift to
an increase in the conjugation length along the polymer
backbone. Our results reveal a similar trend in the evolution
of the UV-Vis spectra. However, geometry optimization re-
0
sults obtained for 1,1 -bis-9,10-anthraquinone (3), a simple
model of polymer 1, together with the crystal structure of its
diamino derivative 4, revealed that the torsion angle between
two anthraquinone units was close to 901 (851 from the
geometry optimization results obtained for 3 and 831 from
2
the crystal structure of 4). Such a large dihedral angle is
expected due to the major steric hindrance introduced by the
carbonyls of each anthraquinone moiety. This should prevent
a
Department of Chemistry, University of Manchester, Oxford Road,
Results and discussion
Manchester, UK M13 9PL. E-mail: juliengautrot@yahoo.fr;
Philip.Hodge@manchester.ac.uk
Avecia, Blackley, Manchester, UK
Carbonyl–p interactions
b
w Electronic supplementary information (ESI) available: UV-Vis spec-
tra of fully reduced 1, 2 and 3, UV-Vis spectra of models 5–8,
evolution of the CT band of model 7 with the dielectric constant of
the solvent, CV traces of models 5–8 and crystallographic data for
compounds 6 and 9. See DOI: 10.1039/b608628d
Evidence of carbonyl–p interactions is scarce in the literature.
Searches of the Cambridge Structural Database (CSD) show
that carbonyl–p interactions are present in some protein
3
crystals as well as in some other organic crystals and between
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 1801–1807 | 1801

View Article Online
aromatic analytes and silica supported polyacrylate deriva-
tives (although in this latter case the exact nature of the
4
interactions was not fully investigated). In the case of electron
rich aromatic rings the carbonyl–p interactions were found to
be energetically favourable only with an edge on approach of
the carbonyl, while a face approach was favoured with elec-
tron deficient aromatic rings.
Recent studies have shown that anion–p interactions exist.
Examples of systems hosting such interactions have been
5
found in the CSD and have been the subject of recent papers,
6
but most of the work on this subject still remains theoretical.
The interaction between p-deficient systems such as 1,3,5-
6
d
6b,e,f
triazine or hexafluorobenzene
and anions (or even het-
eroatoms in molecules such as H O or HF) was found to be
2
energetically favourable. The magnitude of such interactions is
similar to what can be calculated (and observed) in the case of
6a
Fig. 1 Qualitative prediction of the evolution of UV-Vis spectra,
considering carbonyl–p interactions only (all other effects are not
included and may direct the overall changes).
the much more studied cation–p rich systems. Moreover,
energy minima were found to lie along the principal symmetry
axis of the molecule with distances from the centroid typically
˚
falling between 2 and 3 A.
oxygen atoms of the carbonyls. Any effect that will stabilise
the semiquinone state (first reduced state in which one
electron has been transferred) will shift the half-wave poten-
tial to a more positive potential. A carbonyl–p interaction
will therefore shift the half-wave potential to more positive
potentials in the case of electron deficient aromatic substi-
tuents, and to more negative potentials in the case of electron
rich ones.
A model of the carbonyl–p interaction can be extrapolated
from the studies of anion–p interactions, although the systems
differ in several parameters such as the distance and orienta-
tion of the carbonyl double bond with respect to the principal
axis (or rather pseudo axis considering the loss of symmetry in
our system) of the aromatic ring. Keeping this in mind, the
carbonyl–p interaction can be predicted to be energetically
favourable in the case of electron deficient aromatic rings and
destabilising in the case of electron rich rings. Such a model is
consistent with results reported in the literature that were
The UV-Vis spectra of polymer 1, anthraquinone (2) and its
dimer 3 in N-methylpyrrolidone (NMP) are presented in Fig.
2. Consider first the spectrum of 2. Three transitions occur
3
,4
briefly presented above.
7
above 260 nm, consistent with the literature. The two intense
bands near 270 and 320 nm are known to be due to p–p*
transitions with quinonoid and benzenoid characters, respec-
8
tively. The third band, near 400 nm, is very weak and is
Unsubstituted anthraquinone derivatives
Now let us return to anthraquinone dimer 3 and polymer 1.
How can the carbonyl–p interaction account for the UV-Vis
spectrum of 3 and its electrochemical behaviour? In which way
is this interaction necessary to explain the experimental results
observed? In order to answer these questions, it is necessary to
make a simple assumption: the electron density around the
carbonyl oxygen atoms is increased in the LUMO. Compared
to the HOMO of neutral 3, the carbonyl–p interaction will
more strongly perturb the LUMO than it does to the
known to be due to the forbidden n–p* transition.
The highest p–p* transition is red-shifted from anthraqui-
none (2) to dimer 3. The shift is about 12 nm, which
corresponds to an apparent stabilization energy, DE, of
6
f
HOMO. Electronic transitions being vertical, no molecular
reorganisation has sufficient time to occur, so the geometry of
the molecule is not modified and the main parameter that
changes is the electron density near the oxygen of each
carbonyl group. The effect of such a phenomenon is to modify
the optical band gap. The direction of this change will depend
on the nature of the aromatic substituent. In the case of 3 and,
more generally, with electron deficient aromatic substituents,
both the HOMO and LUMO are stabilised, but the LUMO is
more stabilised than the HOMO, as it has just been explained.
The optical band gap decreases, and the highest p–p* transi-
tion is red-shifted (Fig. 1). The trend is the opposite in the case
of electron rich aromatic substituents.
The electrochemical behaviour will also be affected by the
carbonyl–p interaction. During the reduction of the anthra-
quinone moiety, the electron density increases around the
0
Fig. 2 UV-Vis spectra of 9,10-anthraquinone 2 (solid line), 1,1 -bis-
9,10-anthraquinone 3 (squares) and poly(9,10-anthraquinone-1,4-diyl)
1 (triangles) in NMP.
1
802 | New J. Chem., 2006, 30, 1801–1807
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006

View Article Online
ꢁ
1
.1 kcal mol . An even larger shift of 44 nm is observed in
3
the case of polymer 1, which corresponds to a DE of 10.4
kcal mol . Such stabilization energies are comparable to the
ꢁ
1
interaction energies reported for various halides interacting
5
with hexafluorobenzene
b,6a,e
ꢁ1
(typically, near 13 kcal mol
ꢁ
1
for chloride and 18 kcal mol for fluoride, depending on the
method used for the calculations), although lower in magni-
tude. Obviously the present work is not a quantitative study,
but these relatively small values of DE can be qualitatively
rationalized. Firstly, the stabilization of the HOMO dE
tends to increase the optical band gap, which decreases DE
Fig. 1). Secondly, the position of the oxygen atom is not
1
(
optimal for maximizing the interaction energy, since it is not
placed along the axis passing through the centroid and
orthogonal to the plane of the neighbouring aromatic
6
b,f
ring.
The carbonyl–p interaction can also account for the
differences observed in the cyclic voltammetry (CV) traces
of 2 and 3 (see Fig. 3a and b). The CV trace of 2 in NMP
shows two reversible reduction steps with respective half-
wave potentials at ꢁ0.68 and ꢁ1.44 V (against Ag/AgCl),
corresponding to two successive one-electron reductions,
9
consistent with the literature. The CV trace of 3 in NMP,
however, displays four successive reduction steps (presum-
ably of one-electron each). The first two are reversible and
can be compared to the reduction of two isolated anthra-
quinone moieties: instead of displaying one two-electron
reduction step, the two electrons are added successively.
Moreover, the first reduction step occurs at slightly more
positive potentials (ꢁ0.66 V) than for 2. This is consistent
with the hypothesis of a carbonyl–p interaction occurring in
3
. The second reduction step occurs at more negative poten-
tials (ꢁ0.92 V), which is also consistent with the carbonyl–p
interaction if one considers that the entity that is now
reduced is an anthraquinone moiety substituted with an
electron rich aromatic ring (Fig. 3c).
Now it is interesting to look at the UV-Vis spectra of 1, 2
and 3 in their fully reduced states (in which each quinone
moiety has been reduced to the corresponding dianion). For
this purpose, solutions of 1, 2 and 3 in NMP were treated with
1
0
an excess of sodium hydrosulfite in alkali aqueous solution.
A UV-Vis spectrum was then recorded for each compound.
1
0
All three UV-Vis spectra displayed two absorption bands,
1
1
presumably p–p*, as suggested by Carsky et al., and implied
by the high extinction coefficients of these transitions. The
position of the lmax of the highest transition (near 500 nm) is
almost unchanged when going from anthraquinone (2), to its
dimer 3, and then to polymer 1. This result is predicted by the
carbonyl–p interaction model. Indeed, for these dianions, the
charge density centered on the oxygen atoms is very high in
both the ground and excited states. Therefore, both of these
states are stabilized in a similar way and the carbonyl–p
interaction has no noticeable effect on the position of this
transition. The hypothesis of a significant conjugation, despite
the high torsion angle, between the anthraquinone units is not
satisfactory since it does not explain the fact that the highest
transition of the UV-Vis spectra of the fully reduced species
does not shift when increasing the number of anthraquinone
units in the chain.
0
Fig. 3 CV traces obtained for (a) 9,10-anthraquinone 2 and (b) 1,1 -
ꢁ1
ꢁ1
bis-9,10-anthraquinone 3 (3 mmol L ) in TBAPF (0.1 mol L )
6
ꢁ1
solution in degassed NMP (scan rate: 0.1 V s ), vs. Ag/AgCl; (c)
successive reduction steps for dimer 3.
Diaryl-substituted anthraquinone derivatives
In order to probe further the carbonyl–p interaction model,
and check its validity for explaining the UV-Vis and electro-
chemical behaviour of 1-aryl and 1,4-diarylanthra-9,10-qui-
nones, we decided to investigate the properties of other
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 1801–1807 | 1803

View Article Online
given their relative proximity (the O1–S2A distance being
˚
2
.892 A). The large torsion angle seen in 9 suggests that the
carbonyl–p interaction is still relatively important in this
molecule, while it is much weaker in the case of model 6, for
which the oxygen atoms of the anthraquinone carbonyls are
not lying above the aromatic substituents.
0
Fig. 4 X-Ray crystal structures of (a) 1,4-bis(4 -bromophenyl)-9,10-
0
anthraquinone 9 and (b) 1,4-bisthien-2 -yl-9,10-anthraquinone 6; the
lower fraction disordered component has been omitted for clarity.
examples belonging to the latter family of molecules. For this
1
0
The UV-Vis spectra of models 5, 6, 7 and 8 were recorded in
purpose, models 5, 6, 7 and 8 were synthesized. The spectro-
scopic and electrochemical properties of these molecules were
found to depend greatly on their conformation. An X-ray
crystal structure was obtained for model 6 and compound 9, a
dibromo derivative of model 5. These results revealed that the
torsion angle, f, between the anthraquinone moiety and the
aromatic substituent is still large in the case of phenyl sub-
stituents (59 and 691) while it decreases significantly in the case
of thienyl substituents (36 and 381) (Table 1, Fig. 4), whereas
the oxygen to ring-centroid distances are approximately
equivalent for both compounds 6 and 9 (Table 1). This
reduced torsion angle may be due to the smaller size of the
sulfur atom compared to the aromatic CH, together with the
larger angle between the linking single C–C bond and the
adjacent aromatic double CQC bond (see angles C20–C19–C4
and C16–C15–C1 in Table 1), as well as potential chalco-
gen–chalcogen interactions between the oxygen of the carbo-
chloroform and compared to the spectrum of 9,10-anthraqui-
1
0
none 2 in the same solvent. For all these models a weak
broad transition could be observed at long wavelengths (with
onsets of absorption ranging from near 400 nm for 5 to 600 nm
for 8), which was assigned to a charge transfer phenomenon.
The study of the evolution of the position of this transition in
various solvents revealed that it was sensitive to the dielectric
constant of the milieu. At equal extinction coefficients, the
transition is red-shifted in solvents displaying a high dielectric
constant (for model 7 a red-shift of 9 nm can be observed when
going from hexane to chloroform, for an extinction coefficient
ꢁ1
ꢁ1 10
of 1000 L mol cm ). The second highest transition is
ascribed to the highest p–p* transition. In the case of deriva-
tives 5 and 7 (lmax of 284 and 310 nm, respectively) this
transition is blue-shifted compared to the highest p–p* of
anthraquinone (2) (lmax of 326 nm), while it is red-shifted in
the case of thienyl derivatives 6 and 8 (lmax of 326 and 367 nm,
respectively). The blue shift observed for the highest p–p* of
models 5 and 7 can be rationalized thanks to the carbonyl–p
interaction: the presumably large torsion angle in these mole-
cules allows a strong carbonyl–p interaction to take place, and
the relatively high ionisation potential of the aromatic sub-
stituents means that this interaction is energetically destabilis-
ing. On the other hand, the reduction of the torsion angle in
the case of thienyl derivatives will tend to reduce the car-
bonyl–p interaction and increase the extent of conjugation.
Therefore, a red-shift is observed for the highest p–p* transi-
tion of these derivatives.
1
2
nyl and the sulfur atom of the neighbouring thiophene ring,
˚
Table 1 Important distances (A), angles and dihedral angles (1)
obtained for compounds 6 and 9 (standard deviations are in brackets)
Compound 6
Compound 9
C20–C19–C4
C16–C15–C1
C5–C4–C19–C20
C14–C1–C15–C16
centroid–O1
130.6 (0.3) C26–C21–C10
129.4 (0.2) C3–C4–C7
36.4 (0.5) C11–C10–C21–C26
38.0 (0.4) C3–C4–C7–C20
120.4 (0.3)
120.5 (0.3)
69.5 (0.5)
59.4 (0.5)
2.668
2.649
2.660
centroid–O1
centroid–O2
centroid–O2
2.671
1
804 | New J. Chem., 2006, 30, 1801–1807
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006

View Article Online
Table 2 Electrochemical results obtained for the model compounds
opportunity to test the theoretical results obtained for anion–p
systems.
c
c
d
d
E
1/2 /mV
DE /mV
1/2(Epa + Epc) /mV
DE /mV
5
6
7
8
ꢁ980
ꢁ850
ꢁ970
ꢁ910
180
200
390
840
ꢁ1610
ꢁ1430
ꢁ1710
ꢁ1920
280
260
150
490
Experimental
a
b
Methods and materials
a
Results obtained for the different models in degassed anhydrous
For purifications with flash column chromatography Merck
9385 silica gel 60 (230–400 mesh) was used. Thin layer
chromatography (TLC) was carried out using Merck silica
gel (60–254 mesh) coated on PET plates. NMR spectra were
recorded on a Varian Inova 300 MHz spectrometer (abbrevia-
tions used: s, singlet; d, doublet; t, triplet; q, quadruplet; m,
multiplet; bp, broad peak; bd, broad doublet; bm, broad
multiplet; obs, observed; req, required). FT-IR spectra were
recorded on a Perkin Elmer spectrometer with a He/Ne
ꢁ
1
DCM solutions (electrolyte: TBAPF , 0.1 mol L ; scan rate: 0.1 V
6
ꢁ
1
b
), vs. Ag/AgCl. Quasi-irreversible peaks. First reduction
c
s
step. Second reduction step.
d
The CV traces of 9,10-anthraquinone (2), as well as the
1
0
different models, were recorded in DCM. Table 2 gathers the
results obtained. Quantitative electrochemical results within
the anthraquinone family were rationalised using Hammet’s
theory (which relates the rate constant of a reaction to the
6
33 nm (o0.4 mW) laser. Mass spectrometry was carried
out using a Micromass Trio 2000 instrument for EI/CI, a
Micromass Platform instrument for electrospray and a Micro-
mass ToF Spec 2E instrument for MALDI ToF. Melting
points were measured using a Gallenkamp melting point
apparatus. Elemental analysis was performed by the Micro-
analysis lab of the Department of Chemistry, University of
Manchester. DSC measurements were carried out on a Seiko
Instruments DSC 220G instrument. UV-Vis spectra were
recorded using a Unicam UV 300 spectrometer (abbreviations
used: sh, shoulder). Cyclic voltammograms were recorded
using a CH Instruments Electrochemical Workstation. The
working electrode was glassy carbon, the counter electrode
was a platinum wire and the reference electrode was Ag/AgCl.
All measurements were carried out under an argon atmo-
sphere. Molecular modeling results were carried out by the
Computational Chemistry Group at Avecia Research Centre,
Blackley. All chemicals and solvents were purchased from
nature of the substituents near the reactive center) and s
p
9
a,13
values.
However, the results obtained for the various 1,4-
diarylanthraquinones do not seem to follow any trend gov-
erned simply by Hammet’s theory. The half-wave potential for
the first reduction of 5 is shifted to more negative potentials
compared to 2, while it is shifted to more positive potentials in
the case of 6. In the latter case, the presence of a more
extended conjugation decreases the aromaticity of one of the
aromatic rings of the anthraquinone moiety. This tends to
shift the half-wave potential to more positive potentials, since
less energy is required to break the aromaticity of one of the
1
4
benzene rings to form the anthracenoid moiety. The loss of
aromaticity of the substituted benzene ring of 6 is reflected by
0
1
the variations in the C–C bond lengths within this ring. On
the other hand, such a phenomenon is not found in the case of
5, since this molecule does not display any extended conjuga-
tion. Moreover the relatively important shift (ꢁ0.1 V) of the
half-wave potential, compared to 2, cannot be rationalised by
Aldrich, Lancaster, Fluka, Avocado or Strem. They were used
15
without further purification unless stated. Pd(PPh
synthesised as described in the literature.
3
)
4
was
p
simply using s values. Comparison of the results obtained for
5
with those obtained for other derivatives, such as 2-methyl
and 2,3-dimethyl-9,10-anthraquinone, using Hammet’s theory
General procedure for the preparation of
1,4-diaryl-9,10-anthraquinones
show that a much smaller shift is expected (using a s of ꢁ0.1
p
for phenyl groups). The carbonyl–p interaction explains this
phenomenon well, since its effect, in the case of electron rich
aromatic rings, is to shift the half-wave potential to more
negative values. Unfortunately, the large DE (difference be-
tween the anodic and cathodic potentials, Epa and Epc) values
obtained for models 7 and 8 do not allow the comparison of
these results. This phenomenon is not fully understood yet.
The bistriflate of 1,4-dihydroxy-9,10-anthraquinone or 1,4-
0
0
bis(5 -bromothien-2 -yl)-9,10-anthraquinone (1.00 eq), aryl
boronic acid or ester (2.00 eq) and palladium(0) tetrakis
(
triphenylphosphine) (3 mol%) were placed in a round-bot-
tomed flask (3-neck, 100 mL), fitted with a condenser, under
nitrogen. THF (40 mL, degassed with argon) and aqueous
sodium carbonate (1.0 M, 15 mL, degassed with argon) were
added via a septum. The mixture was heated under reflux
overnight and poured onto aqueous hydrochloric acid (0.1 M,
200 mL). This aqueous phase was extracted with dichloro-
methane (2 ꢂ 50 mL), washed with water (3 ꢂ 50 mL), dried
over magnesium sulfate, filtered off and the solvent evaporated
under vacuum.
Conclusion
To our knowledge, these results are the first to show that
intramolecular carbonyl–p interactions do exist and that their
effect on the UV-Vis spectra and electrochemical behaviour of
the molecule can be rationalized qualitatively using theoretical
calculations on anion–p systems. However, it is important to
stress that this study remains qualitative. Molecular modeling
of the models studied will help to rationalize the phenomenon
quantitatively. Ultimately, the family of 1,4-diaryl-9,10-
anthraquinone could be expanded further and could offer an
1,4-Diphenyl-9,10-anthraquinone (5). The title compound 5
was prepared from the bistriflate of 1,4-dihydroxy-9,10-
anthraquinone (1.50 g, 3.47 mmol), phenylboronic acid (850
mg, 6.96 mmol), palladium(0) tetrakis(triphenylphosphine)
(242 mg, 0.21 mmol), THF (40 mL, degassed with argon)
and aqueous sodium carbonate (1.0 M, 15 mL, degassed with
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 1801–1807 | 1805

View Article Online
ꢁ
1
mg, 73%). IR (NaCl, cm ) 2926, 2853, 1676, 1654, 1560,
argon). Recrystallisation from toluene afforded yellow crystals
ꢁ
1
1
1541, 1457, 1307, 1247, 1076, 921, 799 and 738; H NMR
(
1.12 g, 90%). Mp 109–110 1C; IR (KBr, cm ) 1670, 1540,
1
1
8
7
317, 1245, 957, 756, 729 and 697; H NMR (CDCl , ppm) d
(CDCl , ppm) d 8.17 (2H; dd, J = 3 and 6 Hz; H-5), 7.77 (2H;
3
3
.07 (2H; m; H-5), 7.69 (2H; m; H-6), 7.57 (2H; s; H-2),
3
.52–7.42 (6H; m) and 7.36–7.30 (4H; m); C NMR (CDCl3,
s; H-2), 7.75 (2H; dd, J = 3 and 6 Hz; H-6), 7.72 (4H; m), 7.65
0
1
(4H; m), 7.43 (2H; d, J = 4 Hz; H-3 ), 7.34 (6H; m), 7.10 (2H;
0 0
ppm) d 127.1, 127.4, 128.5, 133.0, 134.1, 136.7, 141.4, 142.6,
d, J = 4 Hz; H-4 ), 2.02 (8H; t, J = 8 Hz; H-1 ), 1.31–1.00
0
0 0
ꢁ
1
44.3, 146.3, 184.3; MS (EI/CI) 361 g mol , C26
13
C
1
H
16
O
2
(40H; m), 0.82 (12H; t, J = 7 Hz; H-8 ) and 0.66 (8H; m);
NMR (CDCl , ppm) d 14.4, 22.9, 24.1, 29.5, 30.4, 32.1, 40.8,
53.7, 55.5, 120.0, 120.2, 120.3, 123.1, 125.0, 127.1, 127.2, 127.4,
ꢁ1
requires 360 g mol ; UV-Vis spectrum (chloroform, nm)
3
lmax, 265 (21 950; sh), 284 (8080; sh) and 323 (3550; sh).
1
1
27.9, 133.3, 134.2, 134.3, 137.0, 137.5, 140.9, 141.1, 141.7,
46.3, 151.2, 151.7 and 184.0 (30 obs, 32 req); MALDI 1150 g
0
1
,4-Dithien-2 -yl-9,10-anthraquinone (6). The title com-
pound 6 was prepared from the bistriflate of 1,4-dihydroxy-
,10-anthraquinone (1.50 g, 3.47 mmol), 2-thiopheneboronic
ꢁ1
ꢁ1
mol , C80
C, 83.6%, H, 8.1%, S, 5.6%; found: C, 83.7%, H, 8.1%, S,
92 2 2
H O S requires 1150 g mol ; microanalysis: calc:
9
acid (900 mg, 6.94 mmol) and palladium(0) tetrakis(triphenyl-
phosphine) (242 mg, 0.21 mmol), THF (40 mL, degassed with
argon) and aqueous sodium carbonate (1.0 M, 15 mL, de-
gassed with argon). Recrystallisation from toluene gave red
5
.2%; UV-Vis spectrum (chloroform, nm)lmax, 335 (54 100),
67 (32 700; sh) and 439 (5500; sh).
3
X-Ray crystallography
ꢁ1
needles (0.87 g, 67%). Mp (DSC) 235–236 1C; IR (KBr, cm
1
)
1
681, 1595, 1457, 1304, 1254, 929, 850, 829 and 723; H NMR
Single crystals suitable for X-ray crystallography of 6 and 9
were obtained from toluene, mounted in inert oil and trans-
ferred to the cold gas stream of the diffractometer.
(
CDCl , ppm) d 8.10 (2H; m; H-5), 7.71 (2H; m; H-6), 7.68
3
0
(
2H; s; H-2), 7.45 (2H; d, J = 5.1 Hz; H-3 ), 7.15 (2H; dd, J =
13
0
0
3
and 5 Hz; H-4 ) and 7.06 (2H; d, J = 3 Hz; H-5 ); C NMR
, ppm) d 126.3, 126.5, 127.1, 127.4, 134.1, 134.2, 137.1,
37.5, 142.8 and 183.9; MS (EI/CI) 372 g mol , C22H O S
Crystal data for 6. C22
12 2 2
H O S , M = 372.44, monoclinic,
(
CDCl
3
ꢁ1
˚
a = 3.973(3), b = 10.960(7), c = 36.13(2) A, b = 91.289(10)1,
1
12 2 2
3
U = 1572.7(17) A , T = 100 K, space group P2 /n (no. 14), Z
ꢁ
1
˚
requires 372 g mol ; microanalysis: calc: C, 70.9%, H, 3.3%,
S, 17.2%; found: C, 70.1%, H, 2.9%, S, 16.1%; UV-Vis
spectrum (chloroform, nm) lmax, 326 (10 880), 364 (4000; sh).
1
ꢁ
1
=
4, m(Mo Ka) = 0.353 mm , 8807 reflections measured,
244 unique (Rint = 0.042) which were used in all calculations.
3
The atoms S2 and O20 were disordered over two sites whose
occupancies were constrained to sum to unity, with the main
fraction having a final occupancy of 0.91. The final R(F) was
0
0
0
1
,4-Bis(9 ,9 -dioctylfluoren-2 -yl)-9,10-anthraquinone
The title compound 7 was prepared from the bistriflate of
,4-dihydroxy-9,10-anthraquinone (159 mg, 0.37 mmol), 9,9-
(7).
1
0
2
.0425 using 3244 with I 4 2s(I), wR = 0.1048 (all data).
dioctylfluorene-2-pinacolboronic ester (400 mg, 0.77 mmol),
palladium(0) tetrakis(triphenylphosphine) (44 mg, 0.04
mmol), THF (10 mL, degassed with argon) and aqueous
sodium carbonate (1.0 M, 5 mL, degassed with argon).
Chromatography (silica, petroleum ether/DCM 90/10) af-
forded an orange oil that crystallises slowly in the refrigerator
Crystal data for 9. C26 H14 Br O , M = 518.19, triclinic,
2
2
˚
a = 5.8760(7), b = 9.9959(12), c = 17.790(2) A, a =
02.163(2), b = 91.168(2), g = 103.455(2)1, U = 990.8(2)
1
˚
3
A , T = 100 K, space group P-1 (no. 2), Z = 2, m(Mo Ka) =
ꢁ1
4
0
0
.113 mm , 5802 reflections measured, 3963 unique (Rint
=
ꢁ
1
(
273 mg, 75%). Mp (DSC) 76.3 1C; IR (KBr, cm ) 2925,
853, 1676, 1463, 1457, 1445, 1317, 1271, 1244, 825 and 739;
.029) which were used in all calculations. The final R(F) was
.038 using 3131 with I 4 2s(I), wR = 0.91 (all data).
CCDC reference numbers 618745–618746. For crystallo-
graphic data in CIF or other electronic format see DOI:
0.1039/b608628d
2
1
2
H NMR (CDCl
.82 (2H; d, J = 8 Hz), 7.76 (2H; m), 7.68 (2H; s; H-2), 7.67
2H; dd, J = 3 and 6 Hz; H-6), 7.40–7.31 (8H; m), 7.24 (2H;
m), 1.96 (8H; m), 1.29–1.00 (40H; m), 0.81 (12H; t, J = 7.0 Hz;
3
, ppm) d 8.02 (2H; dd, J = 3 and 6 Hz; H-5),
7
(
1
00
13
H-8 ), 0.77–0.59 (8H; m); C NMR (CDCl , ppm) d 14.4,
3
Chemical doping experiments
2
1
1
2.9, 24.1, 29.5, 29.6, 30.5, 32.1, 40.8, 55.4, 119.9, 120.9, 123.0,
23.6, 126.6, 126.8, 127.0, 127.3, 133.6, 133.8, 134.6, 136.5,
40.5, 141.3, 144.7, 150.8, 151.3 and 184.2; MS (MALDI) 986
In a typical experiment, 9,10-anthraquinone 2 solution (4.6 ꢂ
ꢁ ꢁ1
5
1
0
mol L ) in NMP (UV-Vis spectroscopy grade) and a
ꢁ
3
ꢁ
1
ꢁ1
reductant solution of sodium hydrosulfite (4.6 ꢂ 10 mol
88 2
g mol , C72H O requires 986 g mol ; microanalysis: calc:
ꢁ1 ꢁ2 ꢁ1
L
) and sodium hydroxide (4.6 ꢂ 10 mol L ) in deionised
C, 87.7%, H, 9.0%; found: C, 87.5%, H, 9.5%; UV-Vis
water were degassed by gently bubbling argon through for 20
min. The 9,10-anthraquinone 2 solution (2 mL) was syringed
into a spectroscopic cell (1 cm Spectrosil cell) fitted with a
septum and further degassed for 5 min. A first UV-Vis
spectrum was recorded at this stage. The reductant solution
(0.05 mL) was added to the spectroscopic cell and the resulting
mixture quickly shaken. Two further UV-Vis spectra were
recorded just after addition of the reductant and after allowing
the mixture to rest for 10 min. Finally, some reductant
solution (0.05 mL) was added to the spectroscopic cell, the
mixture shaken and a final UV-Vis spectrum recorded. The
spectrum (chloroform, nm) lmax, 268 (63 900), 303 (29 100),
3
10 (28 000; sh), 362 (3700; sh).
Model compound (8). The title compound 8 was prepared
0
0
from 1,4-bis(5 -bromothien-2 -yl)-9,10-anthraquinone (250
mg, 0.47 mmol), 9,9-dioctylfluorene-2-pinacolboronic ester
(
421 mg, 0.98 mmol), palladium(0) tetrakis(triphenylpho-
sphine) (55 mg, 0.05 mmol), THF (40 mL, degassed with
argon) and aqueous sodium carbonate (1.0 M, 15 mL, de-
gassed with argon). Purification by chromatography (silica,
petroleum ether : DCM 90 : 10) afforded a dark red oil (395
1
806 | New J. Chem., 2006, 30, 1801–1807
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006

View Article Online
reference used is an NMP solution in a matched cell. Identical
amounts of reductant were added each time.
6 (a) C. Garau, A. Frontera, D. Quinonero, P. Ballester, A. Costa
and P. M. Deya, Chem. Phys. Lett., 2004, 392, 85; (b) C. Garau, D.
Quinonero, A. Frontera, P. Ballester, A. Costa and P. M. Deya,
New J. Chem., 2003, 27, 211; (c) C. Garau, D. Quinonero, A.
Frontera, A. Costa, P. Ballester and P. M. Deya, Chem. Phys.
Lett., 2003, 370, 7; (d) M. Mascal, A. Armstrong and M. D.
Bartberger, J. Am. Chem. Soc., 2002, 124, 6274; (e) D. Kim, P.
Tarakeshwar and K. S. Kim, J. Phys. Chem. A, 2004, 108, 1250; (f)
I. Alkorta, I. Rozas and J. Elguero, J. Org. Chem., 1997, 62, 4687;
Acknowledgements
We thank Prof. S. Faulkner for his help with the UV-Vis
spectroscopy and his patient discussion about photophysics.
We also thank Dr P. Mackie for the molecular modeling results.
This work was fully funded by Avecia Ltd, Blackley, UK.
(
g) I. Alkorta, I. Rozas and J. Elguero, J. Am. Chem. Soc., 2002,
24, 8593.
1
7
(a) A. Navas Diaz, J. Photochem. Photobiol., A, 1990, 53, 141; (b) J.
Fabian and M. Nepras, Collect. Czech. Chem. Commun., 1980, 45,
References
2
605; (c) Z. Yoshida and F. Takabayashi, Tetrahedron, 1968, 24,
1
(a) T. Yamamoto and H. Etori, Macromolecules, 1995, 28, 3371;
b) T. Yamamoto, Y. Muramatsu, B.-L. Lee, H. Kokubo, S.
Sasaki, M. Hasegawa, T. Yagi and K. Kubota, Chem. Mater.,
003, 15, 4384.
933; (d) R. A. Morton and W. T. Earlam, J. Chem. Soc., 1941, 159;
(e) R. H. Peters and H. H. Summer, J. Chem. Soc., 1953, 2101.
8 M. Nepras, J. Fabian and M. Titz, Collect. Czech. Chem. Com-
mun., 1981, 46, 20.
(
2
2
3
(a) Work of Dr P. Mackie, Avecia Ltd, unpublished results; (b) K.
Ogawa and H. J. Scheel, Z. Kristallogr., 1969, 130, 405.
(a) K. A. Thomas, G. M. Smith, T. B. Thomas and R. J.
Feldmann, Proc. Natl. Acad. Sci. U. S. A., 1982, 79, 4843; (b)
C.-Y. Kim, P. P. Chandra, A. Jain and D. W. Christianson, J. Am.
Chem. Soc., 2001, 123, 9620.
(a) A. Gambaro, P. Ganis, F. Manoli, A. Polimeno, S. Santi and A.
Venzo, J. Organomet. Chem., 1999, 583, 126; (b) H. Ihara, S.
Uemura, S. Okazaki and C. Hirayama, Polym. J. (Tokyo), 1998,
9 (a) P. Zuman, Collect. Czech. Chem. Commun., 1962, 27, 2035; (b)
J. Q. Chambers, Electrochemistry of quinones, in The Chemistry
of the Quinonoid Compounds, ed. S. Patai and Z. Rappoport,
Wiley, Toronto, 1988, p. 719.
10 See ESI.
11 P. Carsky, P. Hobza and R. Zahradnik, Collect. Czech. Chem.
Commun., 1971, 36, 1291.
12 V. I. Minkin and R. M. Minyaev, Chem. Rev., 2001, 101, 1247.
13 (a) P. Zuman, Collect. Czech. Chem. Commun., 1962, 27, 648; (b)
E. Muller and W. Dilger, Chem. Ber., 1973, 106, 1643.
14 (a) M. Buschel, C. Stadler, C. Lambert, M. Beck and J. Daub, J.
Electroanal. Chem., 2000, 484, 24; (b) G. J. Gleicher, D. F. Church
and J. C. Arnold, J. Am. Chem. Soc., 1974, 96, 2403.
15 D. R. Coulson, Inorg. Synth., 1972, 8, 121.
4
5
30, 394.
(a) P. De Hoog, P. Gamez, I. Mutikainen, U. Turpeinen and J.
Reedijk, Angew. Chem., Int. Ed., 2004, 43, 5815; (b) D. Quinonero,
C. Garau, C. Rotger, A. Frontera, P. Ballester, A. Costa and P. M.
Deya, Angew. Chem., Int. Ed., 2002, 41, 3389.
This journal is ꢀc the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 1801–1807 | 1807