Localization/delocalization of charges in bay-linked perylene bisimides

Article abstract of DOI:10.1002/chem.201103954

The copper-mediated Ullmann coupling of 1,7-dibromoperylene bisimides afforded structurally perfect singly-linked perylene bisimide (PBI) arrays, whilst the homo-coupling of 1,12-dibromoperylene bisimides gave doubly-linked and triply-linked diperylene bi

Full text of DOI:10.1002/chem.201103954

FULL PAPER
DOI: 10.1002/chem.201103954
Localization/Delocalization of Charges in Bay-Linked Perylene Bisimides
Wei Jiang,[a] Chengyi Xiao,[a] Linxiao Hao,[a] Zhaohui Wang,*[a] Harald Ceymann,[b]
Christoph Lambert,*[b] Simone Di Motta,[c] and Fabrizia Negri*[c]
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

Abstract: The copper-mediated Ull-
mann coupling of 1,7-dibromoperylene
bisimides afforded structurally perfect
singly-linked perylene bisimide (PBI)
arrays, whilst the homo-coupling of
1,12-dibromoperylene bisimides gave
doubly-linked and triply-linked dipery-
lene bisimides. The interactions of
three bay-linked diperylene bisimides
that differed in their linkage (singly,
doubly, and triply) were investigated in
their neutral and reduced forms
(mono-anion to tetra-anion). UV/Vis
absorption and fluorescence spectros-
copy revealed different degrees of in-
teraction, which was explained by exci-
ton coupling and conjugation effects.
The electrochemical properties and
spectroelectrochemistry also showed
quite-different degrees of PBI interac-
tions in the reduced mixed-valence spe-
cies, which was apparent by the obser-
vation of CT bands. The interpretation
of the experimental findings was sup-
ported by spin-restricted and -unre-
stricted DFT and time-dependent TD-
DFT calculations with the long-range-
corrected CAM-B3LYP functional. Ac-
cordingly, the degree of interaction in
both the neutral and reduced forms of
the bay-linked PBIs was qualitatively
in the order doubly linked<singly
linked !triply linked, owing to the dif-
ferent degrees of twisting and flexibili-
ty between the two PBIs moieties.
Only triply linked diPBI showed com-
pletely delocalized wavefunctions over
the entire p-system.
Keywords: charge transfer · copper ·
cross-coupling
·
perylenes bisi-
mides · Ullmann reaction
Introduction
localization. The structural distortion that goes along with
charge-localization is often named “Peierls’ distortion”, in
particular in the case of solid-state materials.[4] In organic
and inorganic molecular chemistry, the localization versus
delocalization of charge, along with possible charge-transfer
phenomena, is frequently explained by the concept of
mixed-valence (MV) chemistry.[2b] In the simplest case,
a two-state model is used to outline the electronic situation:
two redox centers in different redox states are connected
within one molecular unit. The two diabatic, formally non-
interacting states, in which a charge is localized on either of
the two redox centers, may then be coupled through an elec-
tronic coupling interaction to yield two adiabatic states
(ground and excited) that represent the physically observa-
ble states. In the adiabatic ground state, the charge may
then be localized on one or the other redox center (a so-
called Robin–Day class II situation)[5] or be delocalized be-
tween both redox centers (a Robin–Day class III situation).
For Robin–Day class II systems, the rate of electron transfer
between the two redox centers is determined by the elec-
tronic coupling (V) and the barrier (and thus the shape of
the ground-state potential) between the state minima along
the reaction coordinate.[6] According to Marcus, the barrier
is associated with a fraction (approximately one quarter) of
the reorganization energy (l).[7] The latter energy refers to
the structural changes that are needed for making charge-
transfer possible. By using these terms and the harmonic ap-
proximation for the diabatic states, resonance energy is
given by V2/l, which leads to delocalization, the larger V
and the smaller l is. Thus, if complete delocalization of
a charge is wanted, one aims at increasing V and decreasing
l, which is particularly difficult for large p-systems; as for
class II systems, V usually gets smaller and l increases the
larger the p-system is.[8] Both trends could be counteracted
by forming stiff p-systems, which led us to consider perylene
bisimides (PBIs). We also note that PBIs have hitherto not
been employed as redox centers in a mixed-valence-chemis-
try context.
Functional p-systems with well-defined structures have
evoked considerable interest over the past couple of decades
in light of their potential applications in molecular electron-
ics.[1] “The larger the p-system, the better the delocalization/
stabilization of charge” is a commonly held belief in organic
chemistry. Whilst this statement is true for very small p-sys-
tems (e.g., naphthalene versus benzene), it is not true for
very large systems, in which charge-localization is frequently
observed.[2] Herein, we report that, in a series of very large
di(perylene bisimides) (which have four times more p-
carbon atoms than naphthalene), triple linkage leads to
charge-delocalization whereas double- or single linkage
leads to charge-localization.
Typically, the reason for charge-localization is a competi-
tion between the gain in resonance energy by charge-deloc-
alization and the gain in energy by structural distortion.[3] To
these effects, the additional gain in energy through solvent
interactions and ion-pairing with counterions favors charge-
[a] Dr. W. Jiang, C. Xiao, L. Hao, Prof. Z. Wang
Beijing National Laboratory for Molecular Sciences
Institute of Chemistry, Chinese Academy of Sciences
Beijing 100190 (P.R. China)
Fax : (+86)10-62650811
E-mail: wangzhaohui@iccas.ac.cn
[b] H. Ceymann, Prof. C. Lambert
Institut fꢁr Organische Chemie, Universitꢂt Wꢁrzburg
Wilhelm Conrad Rçntgen Research Center for
Complex Material Systems
Am Hubland, 97074, Wꢁrzburg (Germany)
E-mail: christoph.lambert@uni-wuerzburg.de
[c] Dr. S. Di Motta, Prof. F. Negri
Dipartimento di Chimica “G. Ciamician”
Universitꢃ di Bologna, Via F. Selmi, 2
40126 Bologna, Italy and INSTM, UdR Bologna (Italy)
E-mail: fabrizia.negri@unibo.it
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201103954.
&
2
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charges in Bay-Linked Perylene Bisimides
FULL PAPER
Perylene bisimides (PBIs), which are among the most-in-
tensively investigated chromophores, have been widely used
in dye chemistry,[9] supramolecular assemblies,[10] and opto-
ACHTUNGTRENNUNG
electronic devices,[11] because they offer a variety of desira-
ble features, such as exceptional electro-optical properties,
photochemical stabilities, as well as flexible tunability. PBI
derivatives with an extended conjugated core exhibit p-elec-
tron delocalization, and thus a low-lying LUMO energy
level, which makes them desirable as n-type semiconduc-
tors[12] or as near-infrared (NIR) dyes.[13] Recent efforts on
the preparation of covalently linked PBI oligomers that are
connected by a variety of bridges, such as phenylene, ethy-
nylene, or thieno
[3,2Àb]thiophene units, etc., have focused
on the modulation of the structures and properties and the
realization of various molecular devices.[14]
Copper-mediated coupling reactions have been widely
À
À
À
À
used in the formation of C N, C S, C O, and C C
bonds.[15] Very recently, some of us synthesized doubly-
linked PBI chiral nanoribbons from tetrachloro-PBI by
using a simple copper-mediated reaction.[16] Furthermore,
a series of fully conjugated triply-linked PBI graphene nano-
ribbons of up to four units with very broad spectra and
strong electron-accepting ability have also been developed
by the combination of copper-catalyzed Ullmann-coupling
[17]
À
reactions and C H transformations (Scheme 1). However,
these triply-linked oligo-PBIs (nꢀ1) had structural isomers,
owing to the two possible coupling sites, which led to diffi-
culties in the synthesis and separation of higher homologues.
Consequently, to expand the versatility and utility of this
synthesis, structurally perfect singly-linked PBI arrays along
the bay regions may be a fascinating candidate for exploring
new connectivity and for further regiospecifically construct-
ing triply linked PBI oligomers by ring-fusion reactions.
Herein, we report the synthesis of three bay-linked PBIs
that differed in their linkage, namely singly, doubly, and
triply linked diperylene bisimides (2, 9, and 10, where n=0;
Scheme 1). Furthermore, we investigated their static optical
properties by UV/Vis absorption and fluorescence spectros-
copy as well as their redox properties by cyclic voltammetry.
UV/Vis/NIR spectroelectrochemistry was used to character-
ize their radical anion states. By comparison with DFT-com-
puted electronic couplings, equilibrium structures, and ab-
sorption spectra, we will draw conclusions about the elec-
tronic structure of this series of diperylene bisimides radical
anions.
Scheme 1. Three bay-linked PBI-based arrays (n=0–x).
Results and Discussion
Synthesis of bay-linked PBIs: Initially, model substrate 1-
bromoPBI (5) underwent homo-coupling with a stoichiomet-
ric amount of nanosized copper powder in dry DMSO at
608C to directly afford the corresponding singly-linked
diPBI (2) in 84.4% yield (Scheme 2). We envisaged that 1,7-
dibromoPBIs could be used as a building block to construct
well-defined singly linked PBI-based arrays along the bay
regions through copper-mediated Ullmann reactions. As ex-
Scheme 2. Synthesis of singly-linked diPBI (2) by the Cu-mediated
homo-coupling of 1-bromoPBI.
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

Z. Wang, C. Lambert, F. Negri et al.
Scheme 3. One-pot coupling of 1,7-dibromoPBI for the synthesis of singly-linked PBI arrays.
pected, using 1,7-dibromoPBI (6)[18] as the precursor afford-
ed a homologous series of well-defined oligomers, from the
dimer up to the octadecamer, as detected by MS
(Scheme 3). Separation by column chromatography on silica
gel afforded the dimer (2) and trimer (3) in 14.7% and
9.8% yield, respectively.
The ladder-type arrays should have planar and rigid p-
structures that facilitate electron-delocalization and enhance
conjugation.[19] As an alternative to 1,7-dibromoPBIs, we an-
ticipated that a 1,12-dibrominated derivative with two bro-
mine atoms in the same bay-region could be used to con-
struct more-rigid, planar p-extended superstructures as
À
a result of the formation of two C C bonds. Therefore, we
examined the homocoupling of 1,12-dibromoPBI (8), which
was readily prepared from the reaction of 1,6,7,12-tetrabro-
moPBI (4Br-PBI, 7), CuI, and l-proline at 608C in 32.5%
yield (65.1% based on the recovered 4Br-PBI). By using the
same procedure, doubly linked diPBI (9) was obtained as an
orange–red solid in 68.7% yield. Surprisingly, a small
amount of triply linked diPBI (10) was also isolated as
a side-product, presumably owing to the efficiency of the
À
Ullmann coupling and C H transformations (Scheme 4).
Our efforts towards the direct coupling of 1,12-dichloro-
ACHTUNGTRENNUNG
PBI[17e] to produce the desired doubly linked diPBI were un-
successful, which indicated the low reactivity of 1,12-di-
ACHTUNGTRENNUNGchloroPBI.
Optical spectroscopy: The singly linked dimer (2) and
trimer (3) were red–violet solids and were soluble in various
organic solvents. The electronic and photophysical proper-
ties of the oligomers were studied at room temperature by
UV/Vis absorption and emission spectroscopy in CHCl3.
Both singly linked dimers (2 and 3) showed broader and
more-complicated absorption bands than the parent PBI
monomer (1, Figure 1). Whilst the spectra of compounds 2
and 3 were more alike, they still showed a pronounced red-
shift of the onset of absorption (590 nm for compound 2 and
620 nm for compound 3 versus 540 nm for compound 1).
This red-shift and broadening could be interpreted by a com-
bination of effects: first, owing to the flexibility of the singly
linked dimers and trimers, a number of slightly different re-
Scheme 4. Synthesis of doubly linked (9) and triply linked diPBI (10).
ciprocal orientations of the PBI units were expected, which,
in turn, may be associated with different exciton couplings
of localized PBI transitions that overall broadened and red-
shifted the spectrum.
The significant Stokes shift compared to parent PBI mo-
nomer 1 (lem =534 nm, ffl =1.0),[20] and the low fluorescence
quantum yield (lem =605 nm, ffl =0.15 for compound 2;
lem =632 nm, ffl =0.26 for compound 3) were also a conse-
quence of this exciton coupling, if one considered that sever-
al twisted conformers contributed to the spectrum. Indeed,
&
4
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charges in Bay-Linked Perylene Bisimides
FULL PAPER
Figure 1. UV/Vis absorption spectra: PBI (1, black); singly linked diPBI
(2, red); triPBI (3, green); doubly linked diPBI (9, orange); and triply
linked diPBI (10, violet) in CHCl3.
the angle between the two PBI p-systems was about 708 for
the gas-phase-optimized geometry but, owing to the flexibil-
ity of the singly linked derivatives, we expected that
a number of different reciprocal orientations of the two PBI
units would be populated at room temperature, which fol-
lowed neither the simple J- or H-type behavior of chromo-
phore aggregates.[21] The magnitude of the exciton coupling
was estimated from the energy separation between the two
lowest-computed energy transitions of compounds 2 and 9
(see the stick spectra of neutral diPBIs in the Supporting In-
formation, Figures S10 and S11).
In contrast, the absorption spectrum of doubly linked PBI
(9) was very similar to monomer 1, just slightly more diffuse
with an onset at about 550 nm (Figure 1). This startling ob-
servation became easily comprehensible if one considered
the structure, which was much-more-rigidly twisted than in
compound 2 (computational results indicated an almost-or-
thogonal orientation of the two PBI p-systems; Figure 2). In
contrast to compound 2, the broadening was reduced be-
cause the structure of compound 9 was forced to be rigidly
twisted by the presence of the double link.
Figure 2. Equilibrium structures (CAM-B3LYP/6-31G*) of the neutral
forms of compounds 2 (a, d); 9 (b, e); and 10 (c, f).
which was bathochromically shifted by about 4350 cmÀ1 rela-
tive to that of compound 1, and showed pronounced vibron-
ic progressions. These properties could hardly be explained
by excitonic interactions and were due to extensive conjuga-
tion over the whole p-system of compound 10. These data
were also supported by the MO coefficients of the
À
HOMOÀ1 and the LUMO, which included the three C C
single bonds that connected the two PBIs to form a rigid,
almost-planar “step ladder” superstructure (Figure 2). The
fluorescence spectra showed a small Stokes shift (170 cmÀ1)
but an almost negligible quantum yield (ffl =0.02).
Electrochemistry: The CVs were measured in CH2Cl2 with
Bu4NPF6 (0.1m) as the electrolyte and were referenced
against the Fc/Fc+ redox couple (Table 1). The CVs of com-
pounds 2, 9, and 10 showed four well-defined, single-elec-
tron reversible reduction waves (Figure 3). Differential
pulse voltammetry experiments showed that each wave cor-
responded to the transfer of a single electron (see the Sup-
porting Information, Figure S4). Compared with PBI, the
half-wave reduction potentials versus Fc/Fc+ were À0.83,
À0.98, À1.14, and À1.25 V for compound 2, À0.80, À0.96,
À1.07, and À1.22 V for compound 9, and À0.46, À0.73,
Moreover, the fluorescence spectrum of compound 9 was
not substantially shifted from that of compound 1 and the
fluorescence quantum yield was the highest among all of the
linked diPBIs considered (ffl =0.44). Triply-linked PBI (10)
behaved significantly different to compounds 2, 3, and 9. It
featured the longest absorbance maximum (684 nm),[17a]
Table 1. Optical and electrochemical properties of bay-linked oligo-PBIs versus PBI (1).
abs
abs
fluo
[b]
[c]
[c]
[c]
[c]
[d]
[e]
Compound
lmax
[nm][a]
e
lonset
[nm][a]
lmax
[nm][a]
ffl
E1r
E2r
E3r
E4r
ELUMO
Eg
[a]
A
]
1
2
3
9
527
533
537
531
684
80900
66200
108500
91600
87000
540
590
620
550
705
534
605
632
552
692
1.00
0.15
0.26
0.44
0.02
À0.96
À0.83
À0.76
À0.80
À0.46
À1.22
À0.98
À0.96
À0.96
À0.73
–
–
À3.91
À4.04
À4.11
À4.07
À4.36
2.30
2.09
2.01
2.25
1.85
À1.14
À1.12
À1.07
À1.49
À1.25
–
À1.22
À1.66
10
[a] Measured in dilute CHCl3 (1.0ꢄ10À5 m); [b] average deviation for ffl (Æ0.04) was determined by using compound 1 (N,N’-di(2,6-diisopropylphenyl)-
perylene-3,4:9,10-tetracarboxylic acid bisimide, ffl =1.00 in CHCl3) as the standard; [c] half-wave potential in CH2Cl2 versus Fc/Fc+; [d] LUMO estimated
by the onset of the reduction peaks and calculated according to ELUMO =À(4.8+Eonset); [e] calculated by the onset of absorption in CHCl3 according to
Eg =1240/lonset
.
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

Z. Wang, C. Lambert, F. Negri et al.
tetra-anion, whilst the band at 10000–17000 cmÀ1 vanished.
These observations suggested the presence of practically
noninteracting PBI units in compounds 2 and 9, whose spec-
troscopic properties were additive. In other words, any
charge that was added to compound 2 or compound 9 was
completely localized on either of the two PBI units. Howev-
er, in particular for the di-anion of compound 2 but also visi-
ble in compound 9, a weak transition at even-lower energy
than those mentioned above appeared at about 5000 cmÀ1,
which was too near to the accessible spectroscopic window
of the solvent to be completely visible. A similar band, al-
though much-less intense, was also visible at the same
energy as the mono-anions and tri-anions of compounds 2
and 9. These bands resulted from the interaction of the two
PBI units and were charge-transfer bands in nature (see
below). Those CT bands in the mono-anions and tri-anions
of compounds 2 and 9 were so-called intervalence charge-
transfer (IV-CT) bands, which are typical of MV class II
compounds. Unfortunately, these bands were very weak and
their maxima were beyond the accessible spectroscopic
window. Thus, we were unable to evaluate the electronic
couplings from these IV-CT bands by, for example, Mullik-
en–Hush theory.[22]
In contrast to compounds 2 and 9, diPBI 10 showed differ-
ent spectra for each redox state, which could not be recon-
structed from independent units. For the mono-anion,
a highly structured band in the range 6000–11000 cmÀ1 ap-
peared that was replaced by a much-more-intense and sharp
band between 7500 and 11000 cmÀ1 for the di-anion. Again,
the tri-anion showed bands down to 5000 cmÀ1 that disap-
pear upon reduction into the tetra-anion, which itself had
a series of new bands at 10000 cmÀ1 and higher energy. This
behavior clearly showed that, in the triply linked and rigid
diPBI 10, the charges that were introduced by stepwise re-
duction interacted strongly, thus confirming the delocaliza-
tion of these charges over the whole diPBI.
Figure 3. Reductive CVs of compounds
2 (red), 9 (orange), and 10
(violet) in CH2Cl2 (electrolyte: 0.1m Bu4NPF6; scan rate: 100 mVsÀ1
;
versus Fc/Fc+).
À1.49, and À1.66 V for compound 10, thereby indicating the
much-higher electron-affinity of triply-linked diPBI as a re-
flection of the electronic coupling between the adjacent PBI
units. Compounds 2 and 9 displayed much-smaller shifts,
which further supported the electronic decoupling of these
non-conjugated oligo-PBIs. The LUMO levels, which were
estimated from the onset of reduction potentials, were
À4.04 eV for compound 2, À4.07 eV for compound 9, and
À4.36 eV for compound 10, which was significantly smaller
(nearly 0.3 eV) than the highly twisted diPBIs (2 and 9).
Spectroelectrochemistry: To gain further insight into the
electronic structure of the charges in compounds 2, 9, and
10, we performed UV/Vis/NIR spectroelectrochemical ex-
periments in CH2Cl2/Bu4NPF6 (0.2m). Thus, specific electri-
cal potentials were applied to a polished platinum disk elec-
trode and UV/Vis/NIR spectra were recorded. Owing to the
relatively large redox-potential separations between the four
redox processes for each compound, there were potentials
at which a specific radical ion could almost exclusively be
generated (in all cases, more than 80% of all molecules).
Thus, the recorded spectra at these potentials were taken as
those of the “pure” radical ion without further deconvolu-
tion. These spectra are shown in Figure 4 versus the wave-
numbers for a better comparison with the computed spectra.
First inspection of the spectroelectrochemistry of com-
pounds 2 and 9 revealed very similar behavior up to the
spectra of the tetra-anion, which was due to their twisted
structures. For both compounds, reduction into the radical
mono-anion resulted in a decrease of the lowest-energy
transition at about 20000 cmÀ1 to roughly 50% of its origi-
nal intensity (if one takes the integral of the band), whilst
new transitions appeared between 10000 and 17000 cmÀ1.
Upon reduction to the radical di-anion, the band at
20000 cmÀ1 disappeared almost completely whilst that be-
tween 10000 and 17000 cmÀ1 doubled in intensity. Further
reduction to the tri-anion again led to an approximate 50%
decrease in intensity of the band at 10000–17000 cmÀ1 and
to the appearance of a new band in the range 14000–
22000 cmÀ1. This latter band doubled on reduction to the
However, a detailed experimental analysis of the spectra
was difficult because the number of states involved and
their electronic origin was unknown. Furthermore, the spin
multiplicity of each redox state (e.g., singlet versus triplet
for the di-anion, doublet versus quartet for the tri-anion,
and singlet, triplet, or quintet for the tetra-anion) was un-
known but may result in distinguishable absorption spectra
that could be used to compare the theoretical and experi-
mental data to allow for unequivocal peak assignment. Thus,
we performed quantum-chemical calculations at the DFT
level to address these questions.
Computational results: As mentioned above, our structure-
optimizations afforded a twisted structure for diPBI 2, an
almost-orthogonal structure for compound 9, and an almost-
planar structure for compound 10. The striking observation
of an even-stronger twist in compound 9 showed that multi-
ple linkages could induce a stronger twist because of more
structural strain.
The different degrees of electronic communication in
singly linked, doubly linked, and triply linked diPBIs could
&
6
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charges in Bay-Linked Perylene Bisimides
FULL PAPER
Figure 4. Comparison between the observed and computed electronic spectra of compounds 2 (a, d); 9 (b, e); and 10 (c, f). Top: CH2Cl2/Bu4NPF6 (0.2m),
versus Fc/Fc+. Bottom: TD-CAM-B3LYP/6-31G*-computed spectra were red-shifted by 2500 cmÀ1
.
be quantified by considering the energy splitting between
the LUMO and LUMO+1 orbitals of the neutral species
(for the optimized molecular structures, see Figure 2 and
Figure 5).[23] Such splitting, which corresponded to twice the
electronic coupling (V) according to the energy-splitting-in-
dimer (ESD) method combined with Koopmans’ Theory
(KT),[24] was 7813 cmÀ1 (CAM-B3LYP/6-31G*) for com-
pound 10. Thus, the electronic coupling between the two
PBI chromophores was 3900 cmÀ1, which should be com-
pared with the intramolecular reorganization energy (li)
that was associated with the PBI-anion formation (about
0.28 eV or 2300 cmÀ1 in the gas phase).[25] Therefore, triply
linked diPBIs were expected to be Robin–Day class III sys-
tems in the gas phase and in nonpolar solvents, such as
CH2Cl2.
For doubly linked and singly linked diPBIs, the situation
was more critical because the electronic coupling diminished
to 809 cmÀ1 for the singly linked diPBI and to 255 cmÀ1 for
the doubly linked diPBI, thus indicating that these systems
should localize the charge on a single PBI unit (Robin–Day
class II) rather than allowing full delocalization.
Single-electron localization was predicted for both the
mono- and tri-anionic forms of the singly- and doubly linked
diPBIs, which was in accordance with their modest electron-
ic couplings compared to compound 10. The bond lengths
for compound 9, which showed the charge-localization, are
given in Figure 6. Symmetry-breaking of the atomic struc-
ture of the mono-anion (red), which indicated localization
of the electron on a single PBI unit, was clearly evident by
comparison with the bond lengths of the neutral (black) and
di-anion structures (blue). Similarly, for the tri-anion, the
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

Z. Wang, C. Lambert, F. Negri et al.
Figure 5. CAM-B3LYP/6-31G* frontier molecular orbitals of neutral
compounds 2, 9, and 10; the larger electronic communication in com-
pound 10 compared to compounds 2 and 9 resulted in a large energy split
in the LUMO/LUMO+1 and HOMO/HOMOÀ1 pairs of orbitals.
bond lengths indicated localization of one electron on one
PBI unit and of two electrons on the other PBI unit. Similar
results were obtained for compound 2 (see the Supporting
Information, Figure S5), whilst a gradual change in geome-
try with delocalization over the two PBI units was predicted
for the triply linked diPBI (see the Supporting Information,
Figure S6).
Figure 6. CAM-B3LYP/6-31G* equilibrium bond-lengths for the neutral
and anionic species of compound 9.
The small electronic communication between the PBI
units in compounds 2 and 9, which was responsible for
charge-localization in the mono- and tri-anions, also implied
a biradical character of the di-anions. To determine the role
of the biradicaloid contributions in the di-anion (and tetra-
anion) forms of compounds 2, 9, and 10, the stability of the
CAM-B3LYP/6-31G* closed shell (CS) wavefunction was
determined at the optimized geometry. Instability was only
found for compounds 2 and 9 in their di-anionic forms,
whose UCAM-B3LYP broken symmetry (BS) equilibrium
geometry was determined and used to compute the electron-
ic absorption spectra. The BS stable structures of com-
pounds 2 and 9 were more-stable than the CS structures by
10.21 and 11.67 kcalmolÀ1, respectively, and were almost de-
generate with the triplet state (see the Supporting Informa-
tion, Table S1). A comparison of the computed bond lengths
for the singlet CS and BS structures, along with the triplet
states, is given in the Supporting Information, Figures S7–
S9.
state of the other anionic forms was confirmed to be a singlet
state (for tetra-anions) and a doublet for the tri-anions, with
higher spin states being typically more than 1.0 eV above
the lowest-energy state (see the Supporting Information,
Table S1). Therefore, electronic absorption spectra were
computed for the most-stable spin state of all of the species
considered except for the di-anion for which the spectrum
of the triplet state was also evaluated for compounds 2 and
9 (Figure 4).
Figure 4 shows a remarkable agreement between the com-
puted and observed spectra, which suggested that the opti-
mized atomic structures and, more specifically, the localiza-
tion versus delocalization pattern (provided by CAM-
B3LYP calculations) was not only consistent with the com-
puted magnitude of the electronic couplings (see above),
but also with the electronic absorption features that were
observed in the spectroelectrochemical investigation.
The finer details of the computed spectra can be found in
the Supporting Information, Figures S10–S12. Reduction to
the mono-anions of compounds 2 and 9 resulted in a de-
crease in the intensity of the band at 20000 cmÀ1 and the ac-
Whilst for the di-anions of compounds 2 and 9, the triplet
and singlet states were predicted to be very close in energy,
the triplet state of the di-anion of compound 10 was
11.86 kcalmolÀ1 (0.51 eV) above the singlet state; therefore,
we safely concluded that the di-anion was its singlet state.
Apart from the di-anionic species, the lowest-energy spin
&
8
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charges in Bay-Linked Perylene Bisimides
FULL PAPER
companying appearance of a number of new electronic tran-
sitions (see the Supporting Information, Figures S10 and
S11) in the 15000 cmÀ1 region, which was in agreement with
the experimental results. Upon further reduction, the ap-
pearance and disappearance of the electronic transitions was
closely reproduced by the calculations. The correlation be-
tween the computed and observed data confirmed the simi-
larity of the spectroscopic signatures of compounds 2 and 9
with the charge-localization that was induced by their de-
creased electronic communication.
The spectra of the di-anionic forms of compounds 2 and 9
deserved an additional comment: for each species, both the
computed spectrum for the triplet state and the computed
spectrum for the BS singlet state (Figure 4). These two spec-
tra were very similar, on account of the similar predicted ge-
ometries for BS singlet and triplet states. Nevertheless, for
the singlet state, a weak feature that was not computed for
the triplet state was found at about 7500 cmÀ1. This feature
was clearly visible as a weak isolated band in the spectrum
of compound 2, whilst it was just a shoulder in the spectrum
of compound 9. This band had the same origin for both
compounds and was due to a charge-transfer (CT) excitation
from the localized molecular orbitals, as shown in Figure 7a
for the singly-linked compound. In this respect, the term
“charge transfer” might be somewhat misleading because in
total, no charge was transferred and there was no change in
the dipole moment upon excitation. However, the individual
MO contributions of the excitations were clearly CT in
nature, which prompted us to call this type of excitation
“quasi-CT”. Interestingly, the experimental spectra of the
di-anion showed a feature at 5000 cmÀ1, which was slightly
more intense for compound 2 than for compound 9 and
which could be attributed to the CT transition. Therefore,
even considering possible inaccuracies in the description of
BS singlet states owing to spin contamination, based on the
comparison between computed and experimental spectra,
we concluded that only the singlet state of the di-anions of
compounds 2 and 9 accounted for the weak CT band that
was observed in the spectroelectrochemical study.
A similar weak feature, which was shifted to even-lower
energy (about 1000 cmÀ1) was computed for the mono- and
tri-anionic forms of compounds 2 and 9; this feature was
due to an IV-CT excitation, as shown in Figure 7b for the
singly linked diPBI. Indeed, the experimental counterpart of
the computed CT band was observed in the low-energy por-
tion of the spectra (Figure 4). The analogy of MO excita-
tions in the mono- and tri-anions of compound 2 (Figure 7b)
also supported our assignment of the low-energy band of
the di-anion (Figure 7a) as a CT excitation.
The electronic transitions of compound 10 were deter-
mined by the strong electron interaction between the two
PBI units; therefore, they were remarkably different from
those of compounds 2 and 9. The close agreement between
the computed and observed spectra suggested that the
degree of electronic coupling in compound 10 was also real-
istically accounted for by the CAM-B3LYP calculations.
Figure 7. a) The localized nature of the UCAM-B3LYP/6-31G* SOMO
and LUMO (a and b) orbitals in the BS biradical structure of the di-
anion of compound 2. The two electronic excitations that were responsi-
ble for the low-energy (computed value: 7500 cmÀ1, observed value:
5000 cmÀ1
) quasi-CT transition are indicated by dashed red arrows.
b) The localized nature of the SOMO and LUMO (a) orbitals of the
mono-anion of compound 2 (left) and the SOMO and LUMO (b) orbi-
tals of the tri-anion of compound 2 (right). These orbitals were involved
in the electronic excitation that was responsible for the low-energy (com-
puted value: about 1000 cmÀ1) CT transition, as indicated by the dashed
red arrow. A schematic representation of the electron localization on the
singly linked diPBI, as determined by UCAM-B3LYP/6-31G* calcula-
tions, is shown in the top-left corner of each graph.
Conclusion
We used the copper-mediated Ullmann coupling of 1,7-di-
bromoperylene bisimides to construct structurally perfect
singly-linked perylene-bisimide arrays along the bay regions,
whilst the homo-coupling of 1,12-dibromoperylene bisimides
afforded doubly linked and triply linked diperylene bisi-
mides. Three bay-linked diperylene bisimides that differed
in their linkage (singly, doubly, and triply linked) were used
to investigate the PBI interactions in their neutral and re-
duced forms. Quantum-chemical calculations at the DFT
and TDDFT levels were used to assist the interpretation of
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!

Z. Wang, C. Lambert, F. Negri et al.
were performed under an argon atmosphere in a CH2Cl2/Bu4NPF6 (0.2m)
solution. The potentials were applied in 20 mV steps by using an EG&G
Princeton Applied Research Model 283 potentiostat. UV/Vis/NIR spec-
tra were recorded with a JASCO V-570 spectrophotometer in reflection
mode at the polished working electrode, which was adjusted to 100 mm
above the cell bottom by using a micrometer screw.
experimental results. UV/Vis absorption and fluorescence
spectroscopy revealed different degrees of interaction,
which may be understood in terms of exciton coupling and
conjugation effects, with the latter effect dominating the in-
teractions in the triply linked compound (10). The electro-
chemical properties (cyclic voltammetry) as well as spec-
troelectrochemical analysis showed quite-different degrees
of PBI interactions in the reduced molecules. Again, the
singly- and doubly linked diPBIs were electronically almost
decoupled but showed weak IV-CT excitations in the NIR,
which resulted from weak interactions between the two PBI
moieties of the mono- and tri-anions of compounds 2 and 9,
respectively. In this regard, the observation of such a quasi-
CT band in the di-anions of compounds 2 and 9 was surpris-
ing in view of their symmetry, but it was explained by a mix-
ture of two CT excitations of the a and b electrons. Howev-
er, the anions of compound 10 showed excitations that were
due to the delocalized wavefunction of the p-system. In con-
trast to our initial expectations, the degree of interaction in
both the neutral and the reduced forms of the bay-linked
PBIs did not follow the order of the number of linkers but
rather qualitatively followed the order: doubly linked<
singly linked !triply linked, owing to the different degrees
of twisting between the two PBIs moieties.
Synthesis of singly-linked diPBI (2): A Schlenk flash was charged with 1-
bromoperylene bisimide (5, 100 mg, 0.13 mmol), copper powder (Aldrich,
particle size<100 nm, 99.8%, 81 mg, 1.27 mmol), and dry DMSO
(20 mL) under an argon atmosphere. The mixture was heated at 608C
with vigorous stirring for 8 h. Next, the cooled mixture was poured into
water and extracted with CH2Cl2. The organic layers were separated,
washed with brine, dried over Na2SO4, and purified by column chroma-
tography on silica gel (petroleum ether/CH2Cl2, 1:2 v/v) to afford com-
pound 2 as an red–violet solid (76 mg, 84.4%). 1H NMR (400 MHz,
CDCl3, TMS) d=8.86–8.91 (m, 8H), 8.72 (d, 3J
ACTHNUTRGNE(NUG H,H)=8.0 Hz, 2H),
8.33–8.35 (m, 4H), 7.45–7.50 (m, 4H; phenyl H), 7.30–7.35 (m, 8H; phe-
nyl H), 2.64–2.82 (m, 8H; isopropyl H), 1.15–1.20 (m, 24H; isopropyl H),
1.06–1.10 ppm (m, 24H; isopropyl H); 13C NMR (100 MHz, CDCl3,
TMS): d=163.69, 163.59, 163.49, 163.36, 146.11, 146.05, 145.83, 142.27,
135.63, 135.42, 135.03, 134.87, 133.83, 132.65, 131.66, 130.60, 130.51,
130.47, 130.11, 129.83, 129.40, 128.25, 127.88, 124.72, 124.55, 124.49,
124.42, 124.21, 124.14, 124.01, 123.67, 31.93, 29.51, 29.46, 24.37, 24.28,
22.99, 14.46 ppm; MS (MALDI-TOF): calcd: 1418.6 [M]À; found: 1419.0;
elemental analysis calcd (%) for C96H82N4O8: C 81.22, H 5.82, N 3.95;
found: C 81.05, H 5.68, N 3.89.
Synthesis of singly-linked PBI-based oligomers: A Schlenk flash was
charged with 1,7-dibromoperylene bisimide (6, 200 mg, 0.23 mmol),
copper powder (Aldrich, particle size<100 nm, 99.8%, 147 mg,
2.30 mmol) and dry DMSO (50 mL) under an argon atmosphere. The
mixture was heated at 608C with vigorous stirring for 5 h. Next, the
cooled mixture was poured into water and extracted with CH2Cl2. The or-
ganic layers were separated, washed with brine, dried over Na2SO4, and
purified by column chromatography on silica gel (petroleum ether/
CH2Cl2, 1:2 v/v) to afford dimer 2 (24 mg, 14.7%) and trimer 3 (16 mg,
9.8%) as red–violet solids, whilst the residual fractions contained a tetra-
mer and other oligomers up to an octadecamer (detected by MS).
Experimental Section
Materials and methods: All chemicals and solvents were purchased from
commercial suppliers and used without further purification unless other-
wise specified. DMSO was freshly distilled from CaH2. N,N’-di(2,6-diiso-
propylphenyl)perylene-3,4:9,10-tetracarboxylic acid bisimide (PBI, 1) and
N,N’-di(2,6-diisopropylphenyl)-1,6,7,12-tetrabromoperylene-3,4:9,10-tet-
racarboxylic acid bisimide (7) were synthesized according to literature
procedures.[26] 1-Bromoperylene bisimide (5) and 1,7-dibromoperylene bi-
simide (6) were prepared according to our previously reported proce-
dure.[18]
1H (400 MHz) and 13C NMR spectra (100 MHz) were recorded on
a Bruker ADVANCE 400 NMR Spectrometer. J values are expressed in
Hz and chemical shifts (in ppm) are given downfield of tetramethylsilane
(TMS) with the residual protonated solvent used as an internal standard.
The signals were designated as follows: s (singlet), d (doublet), t (triplet),
q (quartet), dd (doublet of doublets), and m (multiplet). MS (MALDI-
TOF) were determined on a Bruker BIFLEX III Mass Spectrometer. El-
emental analysis was performed on a FLASH EA1112 elemental ana-
lyzer. UV/Vis spectra were measured with a Hitachi (Model U-3010)
UV/Vis spectrophotometer in a 1 cm quartz cell. Fluorescence excitation
and emission spectra were recorded with a Hitachi FP-6600 FL fluorime-
ter at room temperature. Fluorescence quantum yields were determined
by optical dilute method with N,N’-di(2,6-diisopropylphenyl)perylene-
3,4:9,10-tetracarboxylic acid bisimide (PBI) in CHCl3 as a reference
(ffl =1.0).[20] Cyclic voltammograms (CVs) were recorded on a Zahner
IM6e electrochemical workstation at a scan rate of 100 mVsÀ1, with
glassy carbon discs as the working electrode, Pt wire as the counter elec-
trode, Ag/AgCl electrode as the reference electrode, and ferrocene/ferro-
cenium as an internal potential marker for the calibration of potential.
0.1m Bu4NPF6 in CH2Cl2 (HPLC grade) was used as the supporting elec-
trolyte.
Trimer 3: 1H NMR (400 MHz, [D6]DMSO, TMS) d=9.19 (t, 3J
8.0 Hz, 4H), 8.71–8.77 (m, 8H), 8.57 (s, 2H), 8.45 (s, 2H), 8.38 (t, 3J-
(H,H)=8.0 Hz, 4H), 7.21–7.44 (m, 18H; phenyl H), 2.46–2.66 (m, 12H;
ACHTUNGTRENNUNG(H,H)=
AHCTUNGTRENNUNG
isopropyl H), 0.82–1.04 ppm (m, 72H; isopropyl H); 13C NMR (150 MHz,
CDCl3, TMS): d=163.70, 163.56, 163.45, 163.22, 163.18, 163.13, 146.05,
146.00, 145.85, 142.77, 141.99, 135.97, 135.85, 135.06, 134.98, 134.89,
134.26, 133.48, 132.73, 132.69, 130.73, 130.60, 130.59, 130.47, 130.30,
130.12, 129.87, 129.50, 128.38, 124.75, 124.71, 124.52, 124.44, 124.34,
124.25, 124.04, 123.93, 123.73, 29.55, 29.50, 29.45, 29.43, 29.35, 24.31,
24.26 ppm; MS (MALDI-TOF): calcd: 2126.9 [M]À; found: 2127.6; ele-
mental analysis calcd (%) for C144H122N6O12: C 81.25, H 5.78, N 3.95;
found: C 81.08, H 5.69, N 3.79.
1,12-Dibromoperylene bisimide (8): A mixture of tetrabromoperylene bi-
simide (7, 500 mg, 0.49 mmol), CuI (558 mg, 2.92 mmol), and l-proline
(393 mg, 3.41 mmol) was heated in dry DMSO (10 mL) at 608C under an
argon atmosphere for 24 h. Next, the cooled mixture was poured into
HCl (20 mL, 1m), and extracted with CH2Cl2 (3ꢄ50 mL). The organic
layers were separated, washed with brine, dried over Na2SO4, and puri-
fied by column chromatography on silica gel (petroleum ether/CH2Cl2,
1:1 v/v) to afford compound 8 as an orange–red solid (138 mg, 32.5%,
65.1% based on the recovered 4Br-PBI). 1H NMR (400 MHz, CDCl3,
TMS) d=8.94 (s, 2H), 8.84 (d, 3J
(H,H)=8.0 Hz, 2H), 8.67 (d, 3J
ACHUTGTNRENNUG ACHTUNGTRENNUNG
8.0 Hz, 2H), 7.53 (t, 3J(H,H)=8.0 Hz, 2H; phenyl H), 7.39 (d, 3J
G
ACHTUNGTRENNUNG
8.0 Hz, 4H; phenyl H), 2.73–2.82 (m, 4H; isopropyl H), 1.18–1.22 ppm
(m, 24H; isopropyl H); 13C NMR (100 MHz, CDCl3, TMS): d=163.66,
162.96, 145.97, 138.02, 135.19, 133.37, 131.63, 130.54, 130.22, 130.03,
127.92, 124.63, 124.55, 124.03, 123.64, 123.22, 29.63, 29.57, 24.37 ppm; MS
(MALDI-TOF): calcd: 866.1 [M]À; found: 866.2; elemental analysis calcd
(%) for C48H40Br2N2O4: C 66.37, H 4.64, N 3.22; found: C 66.25, H 4.58,
N 3.21.
Spectroelectrochemistry experiments were performed in a cylindrical
quartz cell with a three-electrode setup that consisted of a platinum disc
working electrode (6 mm diameter), a gold-coated metal plate as counter
electrode, and a Ag/AgCl pseudo-reference electrode. Measurements
&
10
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charges in Bay-Linked Perylene Bisimides
FULL PAPER
Doubly linked diPBI (9): A Schlenk flash was charged with 1,12-dibro-
moperylene bisimide (8, 200 mg, 0.23 mmol), copper powder (Aldrich,
particle size<100 nm, 99.8%, 147 mg, 2.30 mmol) and dry DMSO
(20 mL) under an argon atmosphere. The mixture was heated at 608C
with vigorous stirring for 12 h. Next, the cooled mixture was poured into
water and extracted with CH2Cl2. The organic layers were separated,
washed with brine, dried over Na2SO4, and purified by column chroma-
tography on silica gel (petroleum ether/CH2Cl2, 1:2 v/v) to afford com-
pound 9 as an orange–red solid (112 mg, 68.7%). Compound 10 was also
obtained from the crude mixture as a violet solid (15 mg, 9.2%).
spectra. All of the quantum-chemical calculations were performed with
the Gaussian09 package.[34]
Whilst non-covalently linked aggregates of perylene bisimides needed
more-elaborate methods to cover their electronic structure in the excited
state,[35] our above-outlined procedure appeared to be sufficiently reliable
to treat the electronic aspects of the covalently linked PBIs used herein.
Compound 9: 1H NMR (400 MHz, CDCl3, TMS) d=8.92 (d, 3J
8.0 Hz, 4H), 8.83 (d, 3J(H,H)=8.0 Hz, 4H), 8.03 (s, 4H), 7.48 (t, 3J-
(H,H)=8.0 Hz, 4H; phenyl H), 7.36 (d, J
7.29 (t, 3J
(H,H)=10.0 Hz, 4H; phenyl H), 2.83–2.90 (m, 4H; isopro-
pyl H), 2.61–2.68 (m, 4H; isopropyl H), 1.22 (d, 3J
(H,H)=8.0 Hz, 24H;
isopropyl H), 1.13 (d, 3J
(H,H)=8.0 Hz, 12H; isopropyl H), 1.03 ppm (d,
3J(H,H)=8.0 Hz, 12H; isopropyl H); 13C NMR (100 MHz, CDCl3, TMS):
d=163.82, 163.67, 146.37, 145.60, 140.72, 137.95, 135.47, 134.11, 132.47,
130.56, 130.19, 128.94, 127.39, 124.61, 124.50, 124.44, 123.73, 123.18, 29.50,
24.56, 24.40, 24.34, 24.21 ppm; MS (MALDI-TOF): calcd: 1416.6 [M]À;
found: 1416.9; elemental analysis calcd (%) for C96H80N4O8: C 81.33,
H 5.69, N 3.95; found: C 81.25, H 5.67, N 3.92.
ACHTUNGTRNE(NUNG H,H)=
Acknowledgements
AHCTUNGTRENNUNG
3
For financial support of this research in China, we thank the National
Natural Science Foundation of China (Grant Nos: 91027043, 21021091),
the 973 Program (Grant No. 2011CB932301), NSFC–DFG joint project
TRR61, and the Chinese Academy of Sciences. The research in Germany
was supported by the Deutsche Forschungsgemeinschaft through the
Graduiertenkolleg GRK 1221. The research in Italy was supported by
PRIN Project 2008 JKBBK4 “Tracking ultrafast photoinduced intra- and
inter-molecular processes in natural and artificial photosensors”.
G
ACHTUNGTNER(NUNG H,H)=8.0 Hz, 4H; phenyl H),
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
[1] a) R. E. Martin, F. Diederich, Angew. Chem. 1999, 111, 1440;
Angew. Chem. Int. Ed. 1999, 38, 1350; b) M. Gross, D. C. Mꢁller, H.
G. Nothofer, U. Scherf, D. Neher, C. Brauchle, K. Meerholz, Nature
2000, 405, 661; c) J. M. Tour, Chem. Rev. 1996, 96, 537; d) T. Dutta,
K. B. Woody, S. R. Parkin, M. D. Watson, J. Gierschner, J. Am.
Chem. Soc. 2009, 131, 17321; e) A. Tsuda, A. Osuka, Science 2001,
293, 79; f) M. Baumgarten, M. Mꢁller, A. Bohnen, K. Mꢁllen,
Angew. Chem. 1992, 104, 482; Angew. Chem. Int. Ed. Engl. 1992, 31,
448; g) J. Wu, M. D. Watson, N. Tchebotareva, Z. Wang, K. Mꢁllen,
J. Org. Chem. 2004, 69, 8194.
[2] a) A. Heckmann, C. Lambert, Angew. Chem. Int. Ed. 2012, 51, 326;
b) J. Hankache, O. S. Wenger, Chem. Rev. 2011, 111, 5138.
[3] J. L. Brꢅdas, D. Beljonne, V. Coropceanu, J. Cornil, Chem. Rev.
2004, 104, 4971.
[4] J. L. Brꢅdas, G. B. Street, Acc. Chem. Res. 1985, 18, 309.
[5] M. B. Robin, P. Day, Adv. Inorg. Chem. Radiochem. 1967, 9, 247.
[6] a) B. S. Brunschwig, C. Creutz, Chem. Soc. Rev. 2002, 31, 168;
b) S. F. Nelsen, Chem. Eur. J. 2000, 6, 581; c) K. D. Demadis, C. M.
Hartshorn, T. J. Meyer, Chem. Rev. 2001, 101, 2655; d) J. P. Launay,
C. Coudret, C. Hortholary, J. Phys. Chem. B 2007, 111, 6788.
[7] a) R. A. Marcus, Pure Appl. Chem. 1997, 69, 13; b) R. A. Marcus, N.
Sutin, Comments Inorg. Chem. 1986, 5, 119.
Computational details: Model structures for the bay-linked PBIs, which
featured methyl substituents instead of isopropyl units on the phenyl
rings, were optimized with density functional theory (DFT) calculations
by using the CAM-B3LYP long-range-corrected functional[27] and the 6–
31G* basis set. Because the problem of poor asymptotics of the xc poten-
tial is thought to be a consequence of the self-interaction energy error,
which is in turn responsible for the over-delocalization, CAM-B3LYP
was expected to also provide a more-realistic description of the localiza-
tion of charge in our systems than the B3LYP functional. Note that the
B3LYP functional predicts fully delocalized structures for all of these sys-
tems and localization was not recovered by the inclusion of solvent ef-
fects with the polarizable continuum model (PCM) approach.[28]
The tendency of CAM-B3LYP to overestimate the excitation energy (in
particular for valence excitations)[29] is well-known and, therefore, all of
the computed spectra were rigidly red-shifted by 2500 cmÀ1. Moreover,
CAM-B3LYP was recently shown to be preferable to other functionals,
and in particular to B3LYP, for TD-DFT calculations of excitation ener-
gies and properties, especially those that involve charge-transfer (CT)
states, because it can give a balanced description of local, Rydberg, and
CT excitations.[30] CAM-B3LYP was also recently used to simulate the
CD spectra of doubly linked diPBIs.[16]
[8] a) P. Chen, T. J. Meyer, Chem. Rev. 1998, 98, 1439; b) S. F. Nelsen,
H. Q. Tran, J. Phys. Chem. A 1999, 103, 8139; c) B. S. Brunschwig, S.
Ehrenson, N. Sutin, J. Phys. Chem. 1986, 90, 3657; d) S. F. Nelsen,
D. A. Trieber, R. F. Ismagilov, Y. Teki, J. Am. Chem. Soc. 2001, 123,
5684.
From a computational point of view, the correct description of a biradical
contribution requires one to go beyond the single determinant approxi-
mation, whose limit is readily established by checking the stability of the
wavefunction. An alternative to the computationally expensive multi-de-
terminant approaches is to relax the constraint of identical spatial occu-
pation for the a and b electrons, as in the unrestricted approach. Spin-re-
stricted closed-shell (CS) calculations for the singlet states were per-
formed with the CAM-B3LYP functional. The stability of the singlet
state CS wavefunction was tested for each compound (the keyword
stable=opt was used)[31] in the dianion and tetra-anion states; for those
molecules that showed instability of the CS wavefunction, we determined
the stable broken symmetry (BS) wavefunction[32] (open-shell biradica-
loid configuration) and we also optimized the molecular structure at the
unrestricted UCAM-B3LYP (BS) level. In addition, the lowest-energy
triplet states (T1) were optimized at the UCAM-B3LYP level.
Orbital pictures were prepared with the Molekel 4.3 visual software.[33]
Electronic excitation energies and oscillation strengths were computed
for the thirty lowest singlet excited states of the investigated compounds
with time-dependent (TD) CAM-B3LYP calculations. In plotting the
computed electronic spectra, a Lorentzian linewidth of 0.2 eV was super-
imposed onto each computed intensity to facilitate the comparison with
the experimental spectra. The computed spectra did not include the vi-
bronic structure that was associated with electronic bands and, as
a result, they show a smaller number of bands than the experimental
[9] K. Hunger, Industrial Dyes: Chemistry, Properties, Applications,
Wiley-VCH, Weinheim, 2003.
[10] a) F. Wꢁrthner, Chem. Commun. 2004, 1564; b) T. van der Boom,
R. T. Hayer, Y. Zhao, P. J. Bushar, E. A. Weiss, M. R. Wasielewski,
J. Am. Chem. Soc. 2002, 124, 9582; c) Y. Che, A. Datar, K. Balak-
rishnan, L. Zang, J. Am. Chem. Soc. 2007, 129, 7234; d) C. Addicott,
I. Oesterling, T. Yamamoto, K. Mꢁllen, P. J. Stang, J. Org. Chem.
2005, 70, 797; e) X. He, H. Liu, Y. Li, S. Wang, Y. Li, N. Wang, J.
Xiao, X. Xu, D. Zhu, Adv. Mater. 2005, 17, 2811.
[11] a) M. P. O’Neil, M. P. Niemczyk, W. A. Svec, D. Gosztda, G. L.
Gaines, M. R. Wasielewski, Science 1992, 257, 63; b) J. J. Dittmer,
E. A. Marseglia, R. H. Friend, Adv. Mater. 2000, 12, 1270; c) B. A.
Jones, M. J. Ahrens, M. Yoon, A. Facchetti, T. J. Marks, M. R. Wa-
sielewski, Angew. Chem. 2004, 116, 6523; Angew. Chem. Int. Ed.
2004, 43, 6363; d) B. A. Jones, A. Facchetti, M. R. Wasielewski, T. J.
Marks, J. Am. Chem. Soc. 2007, 129, 15259; e) Y. Che, H. Huang, M.
Xu, C. Zhang, B. R. Bunes, X. Yang, L. Zang, J. Am. Chem. Soc.
2011, 133, 1087.
[12] a) X. Zhan, Z. Tan, B. Domercq, Z. An, X. Zhang, S. Barlow, Y. Li,
D. Zhu, B. Kippelen, S. R. Marder, J. Am. Chem. Soc. 2007, 129,
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
11
&
ÞÞ
These are not the final page numbers!

Z. Wang, C. Lambert, F. Negri et al.
7246; b) J. H. Oh, W. Y. Lee, T. Noe, W. C. Chen, M. Kçnemann, Z.
Bao, J. Am. Chem. Soc. 2011, 133, 4204.
[23] F. Cisnetti, R. Ballardini, A. Credi, M. T. Gandolfi, S. Masiero, F.
Negri, S. Pieraccini, G. P. Spada, Chem. Eur. J. 2004, 10, 2011.
[24] V. Coropceanu, J. Cornil, D. A. da Silva, Y. Olivier, R. Silbey, J. L.
Bredas, Chem. Rev. 2007, 107, 926.
[25] a) E. Di Donato, R. P. Fornari, S. Di Motta, Y. Li, Z. Wang, F.
Negri, J. Phys. Chem. B 2010, 114, 5327; b) S. Di Motta, M. Siracusa
F. Negri, J. Phys. Chem. C 2011, 115, 20754.
[13] a) C. Kohl, S. Becker, K. Mꢁllen, Chem. Commun. 2002, 2778;
b) N. G. Pschirer, C. Kohl, F. Nolde, J. Qu, K. Mꢁllen, Angew.
Chem. 2006, 118, 1429; Angew. Chem. Int. Ed. 2006, 45, 1401.
[14] a) Q. Yan, D. Zhao, Org. Lett. 2009, 11, 3426; b) Z. Yuan, Y. Xiao,
X. Qian, Chem. Commun. 2010, 46, 2772.
[15] a) J. Hassan, M. Sꢅvignon, C. Gozzi, E. Schulz, M. Lemaire, Chem.
Rev. 2002, 102, 1359; b) D. Ma, Q. Cai, Acc. Chem. Res. 2008, 41,
1450; c) S. V. Ley, A. W. Thomas, Angew. Chem. 2003, 115, 5558;
Angew. Chem. Int. Ed. 2003, 42, 5400; d) I. P. Beletskaya, A. V. Che-
prakov, Coord. Chem. Rev. 2004, 248, 2337; e) F. Monnier, M. Taille-
fer, Angew. Chem. 2009, 121, 7088; Angew. Chem. Int. Ed. 2009, 48,
6954; f) Y. Li, L. Tan, Z. Wang, H. Qian, Y. Shi, W. Hu, Org. Lett.
2008, 10, 529; g) Y. Li, C. Li, W. Yue, W. Jiang, R. Kopecek, J. Qu,
Z. Wang, Org. Lett. 2010, 12, 2374.
[26] a) Y. Geerts, H. Quante, H. Platz, R. Mahrt, M. Hopmeier, A.
Bohm, K. Mꢁllen, J. Mater. Chem. 1998, 8, 2357; b) H. Quante, K.
Mꢁllen, Angew. Chem. 1995, 107, 1487; Angew. Chem. Int. Ed. Engl.
1995, 34, 1323; c) L. Fan, Y. Xu, H. Tian, Tetrahedron Lett. 2005, 46,
4443; d) W. Qiu, S. Chen, X. Sun, Y. Liu, D. Zhu, Org. Lett. 2006, 8,
867.
[27] T. Yanai, D. P. Tew, N. C. Handy, Chem. Phys. Lett. 2004, 393, 51.
[28] a) M. Renz, K. Theilacker, C. Lambert, M. Kaupp, J. Am. Chem.
Soc. 2009, 131, 16292; b) M. Kaupp, M. Renz, M. Parthey, M. Stolte,
F. Wꢁrthner, C. Lambert, Phys. Chem. Chem. Phys. 2011, 13, 16973.
[29] a) D. Jacquemin, V. Wathelet, E. A. Perpꢆte, C. Adamo, J. Chem.
Theory Comput. 2009, 5, 2420; b) F. Santoro, V. Barone, R. Improta,
J. Comput. Chem. 2008, 29, 957.
[16] Y. Zhen, W. Yue, Y. Li, W. Jiang, S. Di Motta, E. Di Donato, F.
Negri, S. Ye, Z. Wang, Chem. Commun. 2010, 46, 6078.
[17] a) H. Qian, Z. Wang, W. Yue, D. Zhu, J. Am. Chem. Soc. 2007, 129,
10664; b) Y. Shi, H. Qian, Y. Li, W. Yue, Z. Wang, Org. Lett. 2008,
10, 2337; c) H. Qian, F. Negri, C. Wang, Z. Wang, J. Am. Chem. Soc.
2008, 130, 17970; d) H. Qian, W. Yue, Y. Zhen, S. Di Motta, E. Di
Donato, F. Negri, J. Qu, W. Xu, D. Zhu, Z. Wang, J. Org. Chem.
2009, 74, 6275; e) Y. Zhen, H. Qian, J. Xiang, J. Qu, Z. Wang, Org.
Lett. 2009, 11, 3084; f) Y. Zhen, C. Wang, Z. Wang, Chem. Commun.
2010, 46, 1926; g) Y. Wu, Y. Zhen, Y. Ma, R. Zheng, Z. Wang, H.
Fu, J. Phys. Chem. Lett. 2010, 1, 2499; h) H. Wang, H. Su, H. Qian,
Z. Wang, X. Wang, A. Xia, J. Phys. Chem. A 2010, 114, 9130.
[18] W. Jiang, Y. Li, W. Yue, Y. Zhen, J. Qu, Z. Wang, Org. Lett. 2010,
12, 228.
[30] M. J. G. Peach, P. Benfield, T. Helgaker, D. J. Tozer, J. Chem. Phys.
2008, 128, 044118-1-8.
[31] a) D. Fazzi, E. V. Canesi, F. Negri, C. Bertarelli, C. Castiglioni,
ChemPhysChem 2010, 11, 3685; b) S. Di Motta, F. Negri, D. Fazzi,
C. Castiglioni, E. V. Canesi, J. Phys. Chem. Lett. 2010, 1, 3334.
[32] a) A. Kikuchi, H. Ito, J. Abe, J. Phys. Chem. B 2005, 109, 19448;
b) P. M. Lahti, A. S. Ichimura, J. A. Sanborn, J. Phys. Chem. A 2001,
105, 251.
[33] S. Portmann, H. P. Lꢁthi, Molekel, version 4.3, http://www.cscs.ch/
molekel/, Chimia 2000, 54, 766.
[19] a) A. D. Schlꢁter, M. Lçffler, V. Enkelmann, Nature 1994, 368, 831;
b) C. Xu, A. Wakamiya, S. Yamaguchi, J. Am. Chem. Soc. 2005, 127,
1638; c) U. Scherf, K. Mꢁllen, Makromol. Chem. Rapid Commun.
1991, 12, 489; d) Y. Wu, J. Zhang, Z. Fei, Z. Bo, J. Am. Chem. Soc.
2008, 130, 7192.
[34] M. J. Frisch et al., Gaussian 09, revision A.02, Gaussian, Inc., Wall-
ingford CT, 2009.
[35] a) W. Liu, V. Settels, P. H. P. Harbach, A. Dreuw, R. F. Fink, B.
Engels, J. Comput. Chem. 2011, 32, 1971; b) H. Zhao, J. Pfister, V.
Settels, M. Renz, M. Kaupp, V. C. Dehm, F. Wꢁrthner, R. F. Fink, B.
Engels, J. Am. Chem. Soc. 2009, 131, 15660.
[20] C. You, R. Dobrawa, C. R. S. Mçller, F. Wꢁrthner, Top. Curr. Chem.
2005, 258, 39.
[21] M. Kasha, H. R. Rawls, M. Ashraf El-Bayoumi, Pure Appl. Chem.
1965, 11, 371.
[22] a) N. S. Hush, Coord. Chem. Rev. 1985, 64, 135; b) N. S. Hush, Elec-
trochim. Acta 1968, 13, 1005; c) S. F. Nelsen, R. F. Ismagilov, D. A.
Trieber, Science 1997, 278, 846.
Received: December 17, 2011
Published online: && &&, 2012
&
12
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Charges in Bay-Linked Perylene Bisimides
FULL PAPER
Charge of the linked brigade: The
degree of interaction in the neutral
and the reduced forms of bay-linked
perylene bisimides (PBIs; see figure)
that differ in their linkage is due to the
different degrees of twisting and flexi-
bility between the PBI moieties.
Charge Transfer
W. Jiang, C. Xiao, L. Hao, Z. Wang,*
H. Ceymann, C. Lambert,*
S. Di Motta, F. Negri* ....... &&&& — &&&&
Localization/Delocalization of Charges
in Bay-Linked Perylene Bisimides
Charge Transfer
The copper-mediated Ullmann coupling of 1,7-dibromo-
perylene bisimides afforded structurally perfect singly-
linked perylene bisimide (PBI) arrays, whereas the homo-
coupling of 1,12-dibromoperylene bisimides gave doubly-
linked and triply-linked diperylene bisimides. The degree
of interaction in the neutral and the reduced forms of bay-
linked perylene bisimides, which differed in their linkage,
was due to the different degrees of twisting and flexibility
between the PBI moieties and was evidenced by experi-
mental and computed results. For more details see the Full
Paper by Z. Wang, C. Lambert, F. Negri et al. on page &
&&&ff:
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
13
&
ÞÞ
These are not the final page numbers!