Perylene bisimide dimer aggregates: Fundamental insights into self-assembly by NMR and UV/vis spectroscopy

Article abstract of DOI:10.1002/chem.201201661

A novel perylene bisimide (PBI) dye bearing one solubilizing dialkoxybenzyl and one bulky 2,5-di-tert-butylphenyl substituent was synthesized and its aggregation behavior was analyzed by NMR and UV/Vis spectroscopy in various chloroform/methylcyclohexane

Full text of DOI:10.1002/chem.201201661

FULL PAPER
DOI: 10.1002/chem.201201661
Perylene Bisimide Dimer Aggregates: Fundamental Insights into Self-
assembly by NMR and UV/Vis Spectroscopy
Changzhun Shao, Matthias Grꢀne, Matthias Stolte, and Frank Wꢀrthner*[a]
Abstract: A novel perylene bisimide
(PBI) dye bearing one solubilizing di-
ACHTUNGTRENNUNGalkoxybenzyl and one bulky 2,5-di-tert-
periments. Based on ROESY NMR, a
well-defined p-stacked dimer structure
was determined and further corrobo-
rated by molecular modeling studies.
By varying the solvent composition of
chloroform and MCH, the solvent ef-
fects on the Gibbs free energy of PBI
dimerization were elucidated and
showed a pronounced nonlinearity be-
tween lower and higher MCH contents.
This observation could be related to a
further growth process of dimers into
larger aggregates that occurs in the ab-
sence of chloroform, which is required
to solvate the aromatic p surfaces.
With the help of a single-crystal struc-
ture analysis for a related PBI dye, a
structural model could be derived for
the extended aggregates that are still
composed of defined p–p-stacked PBI
dimer entities.
butylphenyl substituent was synthe-
sized and its aggregation behavior was
analyzed by NMR and UV/Vis spectro-
scopy in various chloroform/methylcy-
clohexane (MCH) solvent mixtures. In
the presence of no less than 10 vol%
chloroform, exclusive self-assembly of
this PBI dye into p-stacked dimers was
Keywords: dimerization
·
NMR
spectroscopy · perylene bisimides ·
selective solvation · self-assembly ·
UV/Vis spectroscopy
un
N
both
concentration-dependent
1H NMR and UV/Vis spectroscopic ex-
Introduction
sional infinite columnar aggregates by means of p–p interac-
tions.[8] The aggregation degrees may be governed primarily
by three factors, that is, concentration, temperature, and sol-
vent.[8b] However, more detailed elucidation of the struc-
ture–property relationships are made difficult on account of
their random aggregate sizes, that is, they exist as a polydis-
perse mixture of aggregates, such as dimers, trimers and
higher oligomers. In this sense, discrete PBI dimers that are
self-assembled by a single p–p interaction are strongly in
demand to provide in-depth appreciation of PBI aggregation
behavior. To date, it is still a challenge to trap PBI aggrega-
tion at the stage of dimer formation. Alternatively, PBI
dimer aggregates have been produced in the presence of ad-
ditional constraints such as covalent linkages or hydrogen
bonds.[10] The Wasielewski group synthesized a PBI dimer in
which two PBI units are covalently fixed to a rigid back-
Aromatic p–p interactions[1] are one of the important non-
covalent forces and exert significant influence on determin-
ing the structure and properties of supramolecular assem-
blies in many key biological processes.[2] To mimic functional
biological structures, aromatic stacks have recently received
appreciable attention as futuristic functional nanosystems.[3]
Likewise, in thin films a great number of p–p-stacked organ-
ic materials have been employed as effective electronic and
photonic molecular devices, such as organic field-effect tran-
sistors[4] and solar cells.[5] Among the various synthetic p-
conjugated molecules, perylene bisimides (PBIs) have been
shown to be excellent photofunctional building blocks for
light harvesting[6] and the most promising candidates for n-
type semiconductor materials,[7] because of their high photo-
stability, excellent optical properties, and electron affinity.
Our group is particularly interested in controlling PBI
self-assembly and elucidating the relationship between su-
pramolecular structure and emergent optoelectronic proper-
ties.[8,9] In solution, PBI compounds with solubilizing alkyl
and alkoxy chains at the imide positions form one-dimen-
AHCTUNGTRENNUNG
bone.[10a] In this way, the PBI dimer arrangement is restrict-
ed to a reduced conformational space that facilitates good
agreements between theoretical calculation and spectroscop-
ic data.[11] Nevertheless, the backbone-directed parallel ori-
entation and a p–p distance of 4.5 ꢀ are evidently not the
favored arrangement of noncovalent PBI aggregates in solu-
tion or the solid state. A similar situation is encountered for
calixarene-tethered PBIs synthesized by our group.[10b] Even
though the two perylene planes are restricted in a small
space, their intramolecular folding process is dynamic and
the spectroscopic features are strongly influenced by the
electron-rich calixarene units. The first self-assembled PBI
dimers were described by Schenning and Meijer.[10c] In this
case, the PBI p–p stacking appears not to be the dominant
driving force for the dimer formation; instead, discrete di-
[a] C. Shao, Dr. M. Grꢁne, Dr. M. Stolte, Prof. Dr. F. Wꢁrthner
Universitꢂt Wꢁrzburg, Institut fꢁr Organische Chemie and
Center for Nanosystems Chemistry
Am Hubland, 97074 Wꢁrzburg (Germany)
Fax : (+49)931-31-84756
E-mail: wuerthner@chemie.uni-wuerzburg.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201201661.
Chem. Eur. J. 2012, 00, 0 – 0
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

merization is achieved with the aid of fourfold hydrogen
bonding.
Very recently, our group reported discrete dimer forma-
tion of bay-substituted core-twisted PBIs comprising oligo-
ACHTUNGTRENNUNG(ethylene glycol) bridges that surround one side of the pery-
lene plane and therefore prevent aggregation beyond
dimers.[12] In this article, we present an alternative strategy
for the uniform dimer formation of bay-unsubstituted PBIs.
The discrete dimerization behavior of PBIs is realized on
the basis of delicate steric control. A new PBI derivative
with both rigid and flexible substituents, one at each imide
position, has been designed and synthesized. Notably,
Matile and co-workers have recently applied the same con-
cept for naphthalene bisimides (NBIs). However, owing to
the smaller p surface of NBIs, the p–p stacking forces are
significantly reduced and no in-depth elucidation of the
thermodynamics of dimerization was achieved.[13] For our
PBI systems, systematic NMR spectroscopic studies paired
with molecular modeling illustrate in detail the self-assem-
bled dimer formation process and the well-defined dimer
structure. The thermodynamics of PBI dimerization were
Figure 1. Possible equilibrium between a) open-m/o-facet monomer con-
formations and b) m–m, m–o, and o–o dimer combinations.
1
determined by variable-concentration H NMR and UV/Vis
spectroscopic experiments in
a
solvent mixture of
CDCl3/A[D14]MCH (v:v=1:5; MCH=methylcyclohexane).
G
perylene facet, it is called the open-m-facet conformation.
As demonstrated in Figure 1a, both open-m- and open-o-
facet conformations possess only one accessible perylene
facet for p–p interaction as the other facet is blocked by
both the imide substituents. The open-o-facet conformation
has a half-opened o-perylene facet, whereas the m-perylene
facet is fully accessible in the open-m-facet conformation. In
the molecularly dissolved state (nonaggregated), rotation of
Finally, the solvent effect on the Gibbs free dimerization
energy is investigated by altering the solvent composition
and a most interesting further self-assembly process into
larger aggregates is observed upon depletion of the good
solubilizing solvent chloroform.
Molecular design of the PBI building block: In order to con-
fine the self-assembly of PBI to dimers in solution, elaborate
design of the PBI substituents is a prerequisite, since parent
PBIs as well as many other p systems[14] have a strong pro-
pensity to form infinite aggregates. In the absence of bay-
substituents that exert additional distortion effects,[12] only
modifications at the two imide positions may be taken into
account.
ꢀ
the benzyl group around the C N bond serves like a “rotat-
ing door” making this molecule conformationally switchable
between the open-o- and open-m-facet conformations. Both
conformations can make use of the accessible perylene facet
to form p–p-stacked PBI dimers, ideally in a head-to-tail
manner to compensate for the spatial demands of the tert-
butyl moieties. Upon dimerization, the benzyl groups will ef-
fectively be locked in a single conformation away from the
aggregating facet, preventing further aggregation of the self-
assembled PBI dimers.
In this work we chose the 2,5-di-tert-butylphenyl group as
one of the imide substituents, because the ortho-positioned
ꢀ
tert-butyl fragment is so bulky that rotation around the C N
bond is essentially impossible (rotational barrier more than
180 kJmolꢀ1).[15] This feature has been exploited for the
design of NBI receptors for adenines[16] and NBI oligomers
with “locked-in” conformations.[17] Since the 2,5-di-tert-bu-
tylphenyl group is fixed and unsymmetrical, the perylene
plane may be distinguished into two facets. For the purpose
of this discussion, the m- and o-perylene facets are defined
as being on the same side as the m- and o-tert-butyl groups,
respectively (Figure 1a). At the second imide nitrogen, a
benzyl group was installed that may freely rotate around the
On account of the different monomer conformations,
three kinds of p–p-stacked dimers may be envisioned, namely
m–m dimer (m-perylene facet versus m-perylene facet stack-
ing of two open-m-facet conformations), m–o dimer (m-per-
ylene facet in open-m-facet conformation versus o-perylene
facet in open-o-facet conformation) and o–o dimer (o-pery-
lene facet versus o-perylene facet stacking of two open-o-
facet conformations) as shown in Figure 1b. While the o-
tert-butyl moiety points towards the p-surface of the o-pery-
lene facet, the m-tert-butyl group extends outside the m-per-
ylene facet. Therefore, m–m dimers should exhibit enhanced
face-to-face contact when compared to m–o and o–o dimers.
Owing to the potentially larger p–p overlap, we anticipated
that the designed PBI building block would prefer m–m
dimer formation to m–o and/or o–o dimers under a thermo-
dynamically controlled self-assembly process.
ꢀ
C N bond at the methylene linker. Steric repulsion between
the imide carbonyls and the benzyl group drives the benzyl
ring to be preferentially located above or below the pery-
lene plane. When the benzyl ring is bent away from the o-
perylene facet, this conformation is assigned the open-o-
facet conformation; if the benzyl ring rotates away from m-
&
2
&
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Perylene Bisimide Dimer Aggregates
FULL PAPER
Results and Discussion
composition was selected after a series of variations to be
the most suitable one to cover the major range of the aggre-
gation process. Upon increasing concentration, the chemical
shifts of perylene protons P1–P4 undergo a pronounced
high-field shift as illustrated in Figure 2, reflecting the inter-
molecular association of PBI
Synthesis of the desired PBI building block: The desired
PBI derivative was obtained according to the synthetic
route outlined in Scheme 1. Benzyl alcohol 1[18] was convert-
chromophores through p–p in-
teractions. It is worthwhile
noting that the two different
imide substituents and their
pronounced impact on the ag-
gregate structure differentiate
the perylene protons into four
signals P1–P4, and they exhibit
distinct concentration-depend-
ent changes in chemical shift.
Of these protons, P3 experien-
ces the largest up-field shift
(Dd=ꢀ1.38 ppm), followed by
P4 (Dd=ꢀ1.10 ppm) and then
Scheme 1. Synthesis of PBI 6: a) CBr4, PPh3, dry THF, under argon, RT, 18 h, 87%. b) NaN3, DMF, 1058C,
3 h, 99%. c) PPh3, dry THF, RT, 2 h; H2O, 3 h, 75%. d) Imidazole, 1408C, 4 h, 50%.
P2 (Dd=ꢀ0.99 ppm), while the
P1 resonance moves to high
ed to benzyl bromide 2 with an Appel reaction in high yield
(87%). Subsequently, the bromide compound 2 was treated
with sodium azide to quantitatively afford benzyl azide 3,
which was reduced to benzyl amine 4 in a Staudinger reac-
tion over two steps. Condensation of perylene mono-imide
5[19] with benzyl amine 4 furnished the target molecule PBI
6 in a reasonable yield of 50%.
PBI 6 was purified by using recycling HPLC and was fully
characterized by 1H and 13C NMR spectroscopy and ESI-
TOF mass spectrometry. Although the tert-butyl groups pro-
vide satisfying dissolution of PBI 6 in chlorinated organic
solvents like CH2Cl2 or CHCl3,[20] additional long alkoxy
chains on the benzyl ring were introduced simply to raise its
solubility in nonpolar solvents, such as MCH, that facilitate
aggregation.[21]
Concentration-dependent 1H NMR spectroscopic studies:
NMR spectroscopy is the preeminent method to monitor
and quantify the formation of supramolecular architec-
tures.[22] In dilution experiments, it has been commonly ob-
served that the proton resonance signals are shifted to a
higher field upon p–p stacking[23] and to a lower field upon
hydrogen bonding.[24] Due to the sensitivity of each proton
to its magnetic environment, observation of resonance shifts
provides a wealth of structural information about the self-as-
sembled aggregates. Nevertheless, this powerful tool has
been rarely employed to follow the process of PBI p–p
stacking.[12b,25] The intrinsic obstacle is the fact that PBI ag-
gregates often suffer from extremely broadened NMR sig-
nals.[26]
Figure 2. Top: Molecular structure of PBI 6 with designation of its pro-
tons. Bottom: Aromatic range of concentration-dependent 1H NMR
spectra of PBI 6 from 2ꢄ10ꢀ5 to 4ꢄ10ꢀ2 m in CDCl /
3 ACHTNUTRGNE[UGN D14]MCH (v:v=
1:5) conducted with 600 MHz NMR at 208C. The dotted lines follow the
resonance shifts of protons P1–P4 and A6 with the concentration
changes.
field only by ꢀ0.40 ppm. This phenomenon indicates that
the p–p stacking center of PBI 6 is not located at the center
of the perylene unit, but displaced towards the P3 and P4
protons. It can be attributed to the steric demand of bulky
tert-butyl unit, which shifts the aggregation center in the di-
rection to the benzylimide side.
To shed light into the aggregation behavior of PBI 6, the
concentration-dependent H NMR spectra were recorded at
ambient temperature ranging from 0.02 to 40 mm in a sol-
Besides the high-field shift of perylene protons, a signifi-
cant down-field movement of aromatic A6 signal occurred
1
concomitantly, indicative of hydrogen-bond formation.[27] In
[28]
ꢀ
vent mixture of CDCl /
N
this case, it is ascribed to a weak C H···O hydrogen bond
Chem. Eur. J. 2012, 00, 0 – 0
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

F. Wꢁrthner et al.
between the aromatic A6 proton and carbonyl oxygen atom.
However, the A6 proton resides outside the perylene plane
and cannot interact with the carbonyl oxygen atoms of its
own molecule. Accordingly, because of p–p dimer formation
the A6 proton may approach the carbonyl oxygen of the
neighboring PBI molecule. It is surprising and unexpected
ꢀ
to detect the weak C H···O interaction in solution, despite
the existence of chloroform co-solvent in the system that
may also act as a comparable hydrogen-bond donor.[29] Thus,
ꢀ
the weak C H···O interaction cannot be regarded as a
major driving force for PBI 6 self-assembly. As shown in
molecular modeling studies (vide infra), the close intermo-
lecular proximity of the A6 proton and the carbonyl oxygen
is only a result of PBI 6 p–p stacking. Alternatively, the
downfield shift of the A6 proton may be attributed to the
magnetic anisotropy of the nearby carbonyl group, resulting
in identical structural conclusions. The observation of a
Figure 3. Fitting of the concentration-dependent chemical shift (d) varia-
tions of perylene protons P1–P4 and aromatic proton A6 (see Figure 2)
to a dimer model by means of nonlinear least-squares analysis (correla-
tion coefficient more than 0.998).
ꢀ
C H···O interaction provides further evidence for tightly
gated molecules aagg as shown in Equations (3) and (4).
stacked PBI chromophores in solution (as close as 3.2–
3.4 ꢀ)[6b] with a preferential rotational angle of approxi-
mately 308, in accordance with the intrinsic electrostatic at-
traction of PBI units.[9] Because proton A6 is positioned
above the m-perylene facet, it can approach the intermolec-
ular carbonyl oxygen only when PBI 6 is organized into m–
m-type dimers.
The most striking feature revealed in Figure 2, however, is
the clean transition from monomers to dimer aggregates
without further p–p stacking into extended oligomers, even
at higher concentration. During the entire NMR dilution ex-
periments of PBI 6, only a simple pattern of proton reso-
cD ¼ cT aagg=2
cM ¼ cTð1ꢀaagg
ð3Þ
ð4Þ
Þ
Since the aagg values at a certain concentration cT can be
extracted from the nonlinear least-squares analysis, both cD
and cM are obtainable. The aggregation degree aagg of PBI 6
at a concentration of 40 mm reaches nearly 90%, indicating
almost complete PBI 6 dimer formation. From the fitting
procedure of perylene protons, the dimerization constant KD
is generated to be 340–600mꢀ1, whereas the best fit of A6
proton gives a somewhat higher value of (900ꢁ50)mꢀ1. In
spite of these deviations, all the values are in the same order
of magnitude and of high reliability (correlation coefficients
more than 0.998).
ACHTUNGTRENNUNGnances is observed that exhibit significant broadening upon
increasing concentration. This can be perceived in terms of
rapid exchange between monomers and aggregates on the
NMR timescale. In principle, at lower temperature the
dimer species should become trapped in either M- or P-heli-
cal arrangements with concomitant splitting of each of the
P1–P4 signals into two signals. Unfortunately, even at 204 K
we could not observe such separated resonances (Figure S1
in the Supporting Information). Because in a fast exchange
regime the observed chemical shift of a specific proton is
the weighted average of chemical shifts in the native (non-
exchanging) states, nonlinear least-squares analysis of the
concentration-dependent chemical shift variations was per-
formed to probe the thermodynamics of the aggregation
process (Figure 3).[14,30] All of the resonance variations of
the perylene and A6 protons fit well to a model that as-
sumes only dimer formation.[14]
Concentration-dependent UV/Vis spectroscopic studies:
UV/Vis spectroscopy is an alternative means applicable over
a large concentration range to elucidate the thermodynamics
for dyes with characteristic absorption bands in the UV/Vis
range. The same PBI 6 solutions used for the NMR dilution
experiments in the concentration range of 0.02 and 40 mm in
CDCl3/ACHTNUGTRNEUNG[D14]MCH (v:v=1:5) were subjected to concentra-
tion-dependent UV/Vis measurements using cuvettes with
path lengths of 0.01–5 mm at ambient temperature
(Figure 4). At low concentration, the absorption spectra
show well-resolved vibronic bands between 400 and 550 nm
for the S0–S1 transition of the single PBI chromophores and
can be attributed to the breathing vibration of the perylene
skeleton polarized along the long molecular axis. The ab-
sorption maximum resides at 521 nm, followed by a second
For a two-state equilibrium between monomer (M) and
dimer (D) species [Eq. (1)] the equilibrium constant KD can
be characterized by Equation (2).
band at 485 nm and a third at 454 nm. The value of A0!0
/
A0!1 at the lowest concentration is greater than 1.6, charac-
teristic of the molecularly dissolved (nonaggregated) state
of bay-unsubstituted PBIs.[25a,26a]
KD
2 M
D
H
ð1Þ
ð2Þ
G
2
KD ¼ cD=ðcMÞ
Upon increasing concentration, the absorption bands
gradually transform, revealing strong electronic coupling be-
tween PBI chromophores due to aggregation. A new ab-
sorption maximum with a substantially smaller e value arises
Herein, both cD and cM can be expressed by a function of
the overall concentration cT and the molar fraction of aggre-
&
4
&
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Perylene Bisimide Dimer Aggregates
FULL PAPER
fully elucidated by means of concentration-dependent NMR
and UV/Vis experiments. As the same samples prepared in
the deuterated solvent mixture of CDCl3/ACTHNURGTNEUGN[D14]MCH (v:v=
1:5) were used, nonsystematic errors were avoided. Accord-
ing to our general understanding, the spectral changes in di-
lution experiments are attributed to the excitonic coupling[32]
of the dyesꢅ transition dipole moments for the case of UV/
Vis absorption measurements, and to ring-current effects
and hydrogen-bond formation in proton NMR spectra.
Herein, the spectral changes due to the three distinct inter-
actions could be followed simultaneously by using the two
spectroscopic techniques. As demonstrated in Figures 3 and
4, the concentration-dependent NMR and UV/Vis spectra
exhibit spectral changes that are attributed unambiguously
to PBI 6 dimerization.
Figure 4. Concentration-dependent absorption spectra of PBI 6 from 2ꢄ
10ꢀ5 to 4ꢄ10ꢀ2 m in CDCl /
3 ACHTNUTRGNE[UNG D14]MCH (v:v=1:5) recorded at 208C. The
dotted lines are the calculated monomer (M) and dimer (D) absorptions
from available data in terms of the dimer model. Arrows indicate the
spectral changes upon increasing concentration. Notably, the spectrum la-
beled as D shows the absorbance of one PBI 6 dye within the dimer ag-
gregates and the dimer spectrum therefore has a twofold intensity. Inset:
Fitting of the concentration-dependent absorbance at 521 nm to a dimer
model by means of nonlinear least-squares analysis (correlation coeffi-
cient 0.997). For comparison, fitting to the isodesmic model is also shown
(dashed line).
DOSY and ROESY NMR studies: To further corroborate
exclusive PBI 6 dimer formation, diffusion ordered spectro-
scopy (DOSY) NMR experiments were carried out with the
sample at a concentration of 40 mm in a CDCl3/ACTHNURGTNEUGN[D14]MCH
(v:v=1:5) solvent mixture to evaluate the aggregate size in
solution.[33] Although the aggregation degree reaches nearly
90% at this concentration, analysis of the DOSY spectrum
demonstrates a relatively large value of 2.11ꢄ10ꢀ10 m2 sꢀ1
for the translational diffusion coefficient D (Figure S2 in the
Supporting Information), indicating small-sized assem-
blies.[8b] By assuming a spherical shape, the hydrodynamic
radius of the small PBI 6 stacks is evaluated based on the
at 494 nm, hypsochromically shifted by 27 nm. Such a hypso-
chromic shift is typical for H-type p–p stacking of parent
PBIs. Simultaneously, the prominent monomer absorption at
521 nm flattens at higher concentration.
The whole set of concentration-dependent absorption
spectra share several isosbestic points at 532, 505, 471, 460,
and 438 nm, implying an aggregation equilibrium between
only two explicit species in the system, the monomer and
dimer of PBI 6. Nonlinear least-squares analysis of the con-
centration-dependent UV/Vis data exhibited a good fit to
the dimerization model (inset of Figure 4). The resultant di-
merization constant KD calculated from the monomer ab-
sorption maximum at 521 nm is (910ꢁ110)mꢀ1, in close
agreement with the values from our NMR analysis.[31] Based
on the recorded absorptions (aagg from 5 up to 90%), the
spectra of the pure monomer and dimer species are calculat-
ed and shown as dotted lines in Figure 4. The calculated
monomer spectrum exhibits an extinction coefficient of
94500mꢀ1 cmꢀ1 at 521 nm and the calculated dimer spectrum
is characterized by two main bands at 496 nm (e=
39000mꢀ1 cmꢀ1) and 535 nm (e=27800mꢀ1 cmꢀ1). The pres-
ence of two allowed optical transitions in the dimer spec-
trum that are displaced to longer and shorter wavelengths
with regard to the main transition of the monomer complies
Stokes–Einstein equation: D=kBT/ACTHUNGTERN(NUG 6phR), in which kB is
the Boltzmann constant, T is the experimental temperature,
h is the viscosity of the solvent mixture[34] and R is the hy-
drodynamic radius of the species. According to this equa-
tion, the hydrodynamic diameter of PBI 6 aggregates was
derived to be 3.2 nm. Taking into account the intramolecular
distance of 2.4 nm between B2 and C4 protons in open-m-
facet conformation of PBI 6 and the alkoxy chain extended
length of 1 nm, the detected dimension from DOSY NMR
experiment confirms the absence of assemblies larger than
dimers of PBI 6.
For PBI 6 dimers, there are three potential combinations:
m–m, m–o, and o–o dimers as outlined in Figure 1b. To de-
termine if all three dimer conformations coexist or if one of
them is preferred, the spatial proximity of individual protons
was probed by the nuclear Overhauser effect (NOE). Cross
peaks in such two-dimensional NMR spectra indicate
through-space interactions at proximities of 5 ꢀ or less and
are particularly useful for determining the structure of ag-
gregates in solution.[35] Because the mass of dimers of PBI 6
(2074 gmolꢀ1) lies in a range for which the NOESY signal
intensity may be fairly low, a ROESY technique was em-
ployed instead.
with
a rotational displacement of the dyes in the p
stack.[8b,32] Remarkably, owing to the very high intensity of
the PBI monomer 0!0 transition, the important lower
energy absorption band of the dimer could not be experi-
mentally obtained (solid lines), but was only revealed by
our mathematical analysis of the monomer–dimer equilibri-
um.
The ROESY NMR experiment was carried out with a de-
gassed sample of PBI 6 in CDCl3/ACTHNUTRGNE[UNG D14]MCH (v:v=1:2) at a
concentration of 10 mm.[36] Since the fixed di-tert-butylphen-
yl group distinguishes the perylene plane into o- and m-per-
ylene facets, the two distinct tert-butyl protons B1 and B2
residing above and below the perylene plane can be regard-
To the best of our knowledge, this study is the first exam-
ple in which the aggregation behavior of a PBI dye has been
Chem. Eur. J. 2012, 00, 0 – 0
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

F. Wꢁrthner et al.
tons from the benzylimide substituent or perylene core. This
finding precludes not only the existence of m–o and o–o
dimers, but also the possibility of further cofacial p stacking
of m–m dimers. For PBI 6 m–m dimers, the outside facing o-
perylene facets are blocked by the o-tert-butyl and benzyl
groups, sterically shielding the dimer from further p–p
stacking beyond dimer species.
Molecular modeling of PBI dimer structure: To help visual-
ize the structure of the PBI 6 dimer we have performed mo-
lecular modeling studies on the basis of our NMR and UV/
Vis measurements. The molecular modeling studies were ac-
complished with MacroModel using MM3* force-field calcu-
lations on the dimer with surrounding chloroform mole-
cules.[37] We began the calculation with a simple monomeric
PBI building block (PBI 6 without the solubilizing long
alkoxy chains) and performed an energy minimization until
full convergence in search of the optimal conformation. Irre-
ꢀ
spective of the rotation of the benzyl ring along the C C
bond, only two kinds of optimized geometries bear the
lowest calculated potential energy of ꢀ25 kJmolꢀ1, corre-
sponding to the open-m- and open-o-facet conformations
(Figure 1a). The steric repulsion between the benzyl and the
carbonyl groups should be the reason why the benzyl group
is prone to lie outside the perylene plane. After these two
conformations were confirmed as the lowest energy states,
two PBI molecules were placed in a parallel face-to-face
stacked m–m dimeric arrangement and energy minimized.
The dimer conformational search found a global minimum
for the calculated potential energy of ꢀ125 kJmolꢀ1, which
corresponds to the m–m dimeric structure with a certain ro-
tational angle, in accord with experimental observations. For
comparison, the calculated energies of m–o and o–o dimers
according to the same procedure were ꢀ117 and
ꢀ114 kJmolꢀ1, respectively. These values are approximately
10 kJmolꢀ1 higher than that of the minimized m–m dimer.
The main energy difference arises from the van der Waals
term, which is attractive for p–p stacking and roughly pro-
portional to the area of p overlap.[1b]
Figure 5. a) Part of the ROESY NMR spectrum of PBI 6 at a concentra-
tion of 10 mm in CDCl3/A[D14]MCH (v:v=1:2) conducted with 600 MHz
NMR at 208C. Positive and negative signals are represented by blue and
red contours, respectively. b) MM3* geometry-optimized m–m dimer
structure of the designed PBI building block with selected intermolecular
distances and red labeled protons discussed in the text.
CTHUNGTRENNUNG
ed as “probing protons” to detect the stacked PBI counter-
part located over the o- or m-perylene facet in the dimer
structure. As shown in Figure 5a, protons B2 have spatial
cross correlation with protons C2 and M1 from the other
imide side. Because the rigid perylene skeleton separates
the two imide positions intramolecularly far away from each
other, the cross signals between protons from different
imide sides can be safely assigned to the dimer structure of
PBI 6. The presence of cross-peaks for B2 with C2 and M1
is a strong evidence for the exclusive formation of head-to-
tail m–m dimers. In this manner, the benzyl group is orient-
ed away from the m-perylene facet and its methylene pro-
tons become spatially adjacent to the m-tert-butyl moiety of
the nearby PBI molecule. Additionally, the cross-peak be-
tween protons A6 and M1 validates the unique m–m dimer-
ic arrangement and also provides further evidence for the
In the geometry-optimized m–m dimer structure illustrat-
ed in Figure 5b, the face-to-face stacked PBIs have a p–p
distance of 3.46 ꢀ and a rotational angle of 268. The spatial
distance between B2 and C2, B2 and M1, as well as A6 and
M1 are within 3 ꢀ, accounting for the through-space corre-
lation detected in the ROESY NMR spectrum. Since there
was zero contribution of hydrogen bonds for the potential
ꢀ
energy, the weak C H···O hydrogen bonding was not includ-
ed in the calculation. In spite of this fact, the A6 protons
are directly facing the carbonyl oxygen atoms of the adja-
cent PBI. There is a 2.94 ꢀ intermolecular H···O distance
ꢀ
ꢀ
weak C H···O interaction between the A6 proton and car-
and the C H···O bending angle is 1528. As observed in crys-
bonyl oxygen as observed in NMR dilution experiments.
Only in m–m dimers can the A6 protons intermolecularly
approach both M1 protons and carbonyl oxygen atoms.
Furthermore, it should be especially noted that unlike
protons B2 and A6, there is no cross-peak that could dem-
onstrate the spatial approach between B1 and remote pro-
tal structures and biological systems,[28b,e] these findings cor-
ꢀ
roborate the existence of the weak C H···O hydrogen bonds
in the PBI 6 m–m dimer. It is worth noting that although a
starting geometry with initially parallel aligned PBI units
was applied, the geometry-optimized PBI dimers display a
certain rotational angle and a slight longitudinal shift, with
&
6
&
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Perylene Bisimide Dimer Aggregates
FULL PAPER
the best p-overlap area close to the benzylimide side of the
unsymmetrical PBI building block. This complies with the
observation in NMR dilution experiments that perylene pro-
tons P3 and P4 exhibit the largest high-field shift, followed
by protons P2 and then P1. Overall, the dimer structure ob-
tained from the molecular modeling studies is strongly con-
sistent with our robust experimental results.
Information). Although these spectra show a number of rea-
sonably defined resonances, a reliable assignment was im-
possible.
Accordingly, the aggregation behavior of PBI 6 in MCH
was investigated by means of concentration-dependent UV/
Vis spectroscopy. Unlike the aggregation process in the sol-
vent mixture, the self-assembly of PBI 6 in pure MCH upon
increasing concentration suggests the involvement of some
cooperativity.[38] By applying the simple K2–K model,[39] two
equilibrium constants of K2 =2800mꢀ1 for dimerization and
K=14000mꢀ1 for subsequent growth were estimated (inset
of Figure 6b). From the K2 value of 2800mꢀ1, the Gibbs di-
merization energy for PBI 6 in MCH was assessed to be
ꢀ19.3 kJmolꢀ1 that agrees with the extrapolated DDGoMCH
value (vide infra). However, the further cooperative aggre-
gation process is puzzling if we take into account the fact
that extended PBI p–p stacking beyond m–m dimers is steri-
cally prevented and that the UV/Vis spectra of aggregate
solutions in MCH (Figure 6b) show a close resemblance to
those observed for the dimer aggregates in the presence of
chloroform (Figure 4). Furthermore, several isosbestic
points in Figure 6b corroborate the presence of only two
types of spectroscopically distinguished species, that is, mon-
omer and dimer of PBI 6, over the whole concentration
range in MCH.
Aggregation studies in pure MCH solvent: In the previous
variable-concentration NMR and UV/Vis spectroscopic ex-
periments, we employed
a
solvent mixture of
CDCl3/A[D14]MCH (v:v=1:5) to cover the largest possible
CHTUNGTRENNUNG
range of the binding isotherm (aagg from 5 up to 90%).
Chloroform is a “good” solvent for PBIs, in which PBI 6 has
a high solubility but quite weak aggregation property (Figur-
es S3–S4 in the Supporting Information).[21] In contrast,
MCH is a “bad” solvent, in which PBIs bear a low solubility
but very strong aggregation tendency.[21] Indeed, unlike in
the presence of at least 10 vol% chloroform (vide infra, see
Figure 9), a complex pattern of proton NMR signals appears
in pure MCH at higher concentration and the corresponding
DOSY NMR data indicate the formation of large aggregates
beyond dimers (Figure 6a and Figure S6 in the Supporting
Inspired by the crystal structure of a reference PBI build-
ing block 7 (Figure 7b; for further details see Supporting In-
formation and CCDC 876014 at Cambridge Crystallographic
Data Centre), we realize that the di-tert-butylphenyl group
may occupy the free o-perylene facets of PBI 6 m–m dimer.
Such a binding process might be strongly favored owing to a
proper rigidification of the benzyl substituent in an ideal
conformation for accommodation of the di-tert-butylphenyl
subunit of another PBI 6 dimer. Thus, we propose a further
growth process for PBI 6 self-assembly in MCH as shown in
Figure 7c. In this kind of packing arrangement, the solvo-
[40]
phobic effect[21] and the C H···p interaction should play a
ꢀ
significant role.[41] In the absence of chloroform, which fa-
vorably solvates the PBI p surface, PBI 6 dimers need to
find a way to exclude the badly solubilizing MCH molecules
on top of their “naked” o-perylene facets, and the accom-
modation of the di-tert-butylphenyl groups just meets this
demand. This evidently leads to a rather spectacular change
of the self-assembly process from anti-cooperative (in the
presence of “good” solvent chloroform) to cooperative (in
pure MCH) upon variation of the solvent composition.
A noticeable feature of this further growth process is that
it is not readily identifiable by the UV/Vis spectral changes,
because they are still governed by excitonic coupling within
the PBI dimer stacks, whereas excitonic coupling to neigh-
boring PBIs is negligible owing to a large distance and im-
proper angular relationship[32] between the dyesꢅ transition
dipole moments.[42] Thus, from the self-assembly process
proposed in Figure 7c we are able to rationalize why the
UV/Vis spectra in Figure 6b only indicate the existence of
p–p-stacked PBI dimer aggregates. In many aggregation
studies, similar concentration- and temperature-dependent
Figure 6. a) Concentration-dependent 1H NMR spectra of PBI 6 from
10ꢀ5 to 10ꢀ2 m in pure [D14]MCH conducted with 600 MHz NMR at 208C.
b) Concentration-dependent absorption spectra of PBI 6 from 10ꢀ6 to
10ꢀ2 m in pure undeuterated MCH recorded at 208C. Arrows indicate the
spectral changes upon increasing concentration. Inset: Fraction of aggre-
gated molecules aagg plotted as a function of KcT with different s=K2/K
values according to the cooperative model (lines from left to right: s=1,
0.2, 0.1, 0.01 and 0.001), and plot of experimental absorption data of PBI
6 in MCH at 517 nm.
Chem. Eur. J. 2012, 00, 0 – 0
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

F. Wꢁrthner et al.
Figure 7. a) Molecular structure of PBI 7. b) Packing of PBI 7 crystal structure (view along c-axis). The co-crystallized CHCl3 molecules are also shown.
c) Schematic representation of the proposed aggregation behavior of PBI 6 in MCH.
changes in the UV/Vis spectra with isosbestic points are
taken as evidence for a simple dimerization equilibrium.
Herein, we have shown by an in-depth analysis of the bind-
ing isotherm with the cooperative K2–K model and addition-
al concentration-dependent NMR experiments that such in-
terpretations are not necessarily correct. Indeed, we believe
that many dye aggregates, if not the majority, grow into
more extended aggregates in solvents with poor solubilizing
capabilities for aromatic p surfaces (e.g., aliphatic solvents,
protic solvents and water), although the UV/Vis spectra in
these solvents are still in accordance with the presence of
excitonically coupled dimers.[9] This is explained by the pro-
nounced distance dependence of the excitonic coupling[32]
prohibiting further insight into the location of additional
neighboring molecules, the transition dipole moments of
which are not as proximate as those within the p-stacked
dimer units.
whole polarity range.[21] In this study, the Gibbs free binding
energies were correlated with different solvent polarity
scales including relative permittivity er, the Kirkwood–On-
A
ACHTUGNTREN(NUNG 2er+1), ET(30), p*, and cR. The pri-
mary conclusion from this study was that the aromatic p–p
interaction is the strongest in both nonpolar aliphatic and
most-polar protic solvents, and the weakest in chloroaliphat-
ic solvents of intermediate polarity, such as dichloromethane
and chloroform.[21,43d] In a variety of cases, however, it is
better to employ a binary solvent mixture to achieve the de-
sired binding strength, particularly in studies of the folding
process of oligomers and macromolecules.[26a,43a,d,45] By
mixing a “good” and a “bad” solvent, the folding degree
could be modulated and a straight-line relationship between
the Gibbs folding energy and solvent composition was
found in the transition region.[46]
While keeping in mind the special situation in pure MCH
solvent (vide supra), we have further examined the solvent
effect on the Gibbs dimerization energies of PBI 6 by using
solvent titration experiments in a similar way as studied for
solvent denaturation of foldamers.[26a,45] The solvent-depend-
ent UV/Vis absorptions with CHCl3/MCH compositions
from 100:0 to 10:90 were acquired at a fixed concentration
of 10 mm and ambient temperature. Accordingly, the ab-
sorption changes shown in Figure 8a can only be attributed
Solvent-dependent aggregation studies: Solvation is an im-
portant factor in controlling the self-assembly of aromatic p
stacks.[21,43] Due to the complexity of the nature of p–p inter-
actions,[44] the relationship between the solvent polarity and
the strength of aromatic interaction is not simple.[1,2] In
recent work, we have examined the magnitude of PBI p–p
interaction in water and 16 organic solvents covering the
&
8
&
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Perylene Bisimide Dimer Aggregates
FULL PAPER
DDGo ¼ mDDGoMCH þ ð1 ꢀ mÞDDGo
ð8Þ
ð9Þ
CHCl3
DDGo ¼ mðDDGMo CH ꢀ DDGo
Þ þ DDGo
CHCl3
CHCl3
In Figure 8b, the DDGo values were plotted against the
MCH content m in the range of m=0 and 0.9. The points
can be fitted to two linear relationships with the intersection
at m=0.6. At low MCH contents (from m=0 to 0.6), the
linear fitting can be interpreted in terms of Equation (9),
giving rise to an extrapolated DDGoMCH value of (ꢀ16.0ꢁ
0.2) kJmolꢀ1. In contrast, upon utilization of the data points
for higher MCH contents (from m=0.6 to 0.9), the dotted
fitting line crosses the y axis at m=1.0 for an ideal DDGMo CH
value of (ꢀ19.8ꢁ0.1) kJmolꢀ1. In our opinion, this discrep-
ancy points to a selective solvation phenomenon. In the
presence of up to 60 vol% MCH in CHCl3, PBI p faces are
preferentially solvated by CHCl3 molecules. At higher MCH
contents, however, preferential solvation by the decreasing
fraction of CHCl3 molecules becomes disfavored for entrop-
ic reason (i.e., CHCl3 molecules have to be considered like
guest molecules dissolved in MCH) and PBI dimerization
involving the m-perylene facets becomes strongly triggered
by the solvophobic effect of the MCH solvent.
Figure 8. a) Solvent-dependent UV/Vis absorption changes of PBI 6 at a
concentration of 10 mm recorded at 208C, starting in pure CHCl3 and in-
creasing the volume fraction of undeuterated MCH in 10 vol% steps.
Arrows indicate the spectral changes upon increasing the volume fraction
of MCH. b) Plot of the DDG8 values versus the MCH volume fraction m.
The solid and dotted lines are the best linear fitting results of the points
from m=0 to 0.6 and from m=0.6 to 0.9, respectively. The data point of
DDG8=ꢀ19.3 kJmolꢀ1 at m=1 was deduced from the K2 value of
2800mꢀ1 in terms of K2–K model.
1H NMR linewidths: Further fundamental insights into PBI
aggregation can be derived from the signal broadening in
1
NMR spectra. The widths of the H NMR signals of PBI 6
exhibit a pronounced dependence on CDCl3/ACTHNURGTNEUGN[D14]MCH
ratio (Figure 9), concentration (Figure 6a), and temperature
to the solvent effect. Such absorption changes resemble
those of the concentration-dependent measurements. With
increasing MCH content, the monomer bands decrease
gradually and the dimer bands become apparent. The aggre-
gation degree aagg, dimerization constant KD, and Gibbs di-
merization energy DDGo were calculated for each solvent
composition according to Equations (5)–(7), respectively.
ðeA0!0
Þ
ꢀ eA0!0
Þ
mꢀax ðeA0!0 Þmin
ð5Þ
aagg
¼
ðeA0!0
max
1
Figure 9. Partial H NMR spectra of PBI 6 recorded with 600 MHz NMR
at a concentration of 10 mm in pure CDCl3 and [D14]MCH solvents as
well as in solvent mixtures of CDCl3/ACTHNURGTNEUGN[D14]MCH.
aagg
2ð1 ꢀ aaggÞ cT
cD
KD
¼
¼
ð6Þ
ð7Þ
2
2
ðcMÞ
DDGꢂ ¼ ꢀRT ln KD
(Figure S1 in the Supporting Information). For a high
chloroform fraction, a small PBI 6 concentration and the
ambient temperature sharp signals are observed correspond-
ing to a weighted average of monomer and dimer resonan-
ces with fast exchange between each other. For the opposite
extreme (pure MCH and high PBI 6 concentration) a large
If we neglect the possibilities of preferential solvation, the
Gibbs free dimerization energies at the respective solvent
composition should be proportional to their component in
the solvent mixture: see Equations (8) and (9) in which m is
the volume fraction of MCH in the solvent mixture, and
1
number of individual H NMR signals with moderate line-
DDGo
and DDGoMCH represent the Gibbs free energy con-
tribution for the PBI 6 dimerization in pure CHCl3 and
widths appear that can be attributed to coexisting dimers
and oligomers with slow exchange resulting in individual
resonances for each species. The broad NMR signals in the
CHCl3
MCH, respectively.
Chem. Eur. J. 2012, 00, 0 – 0
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!

F. Wꢁrthner et al.
intermediate region can thus be interpreted by coalescence
effects as the crossover from slow to fast exchange process-
es. All three types of discussed spectra (fast, intermediate,
and slow exchange) can be observed by varying the solvent
from pure chloroform to pure MCH at a fixed PBI 6 con-
centration (Figure 9). The spectral changes (chemical shifts,
linewidths) upon increasing the MCH fraction from 0 to
90 vol% strongly resemble those monitored for rising con-
centration of PBI 6 at a constant CDCl3/ACTHNURGTNEUGN[D14]MCH ratio
(Figure 2).
In additional investigations, we kept the dimerization
degree aagg constant (around 50%) and compared the 1H
linewidths under different experimental conditions, starting
dimer formation driven solely by p–p interactions, but also
fully described the well-defined dimer structure in solution.
Furthermore, the p–p stacking of PBI 6 was investigated
in different CHCl3/MCH solvent mixtures. Two regimes with
a linear relationship of the Gibbs free energy versus the sol-
vent composition are found between 0 and 60 vol% MCH
and between 60 and 90 vol% MCH in CHCl3 (for PBI 6 at
a concentration of 10 mm), suggesting a selective solvation
effect. Accordingly, preferential solvation of the p surfaces
by chloroform molecules weakens the PBI p–p interaction
in solvent mixtures as long as the CHCl3 content is of suffi-
cient magnitude (ꢃ40%). In pure MCH, the free p face suf-
fers from bad solvation, which triggers the formation of
larger aggregates according to complex 1H NMR spectra,
DOSY NMR studies, and UV/Vis dilution experiments. The
last set of experiments indeed revealed a cooperative self-
assembly process that cannot be explained by a simple PBI
dimerization process. Based on the single-crystal structure
of a similar PBI building block, a packing model was elabo-
rated for the extended aggregation that does not involve ad-
ditional contacts between PBI p faces beyond those given in
the dimer units, in compliance with the UV/Vis spectra of
the aggregates. To the best of our knowledge, this is the first
time that often observed apparent equilibria between two
spectroscopically distinguished species with well-defined
with a PBI 6 concentration of 2 mm in a CDCl3/ACTHNURGTNEUGN[D14]MCH
(v:v=1:5) mixture at 258C. Upon increasing the concentra-
tion of PBI 6 either the chloroform content (Figure S8 in
the Supporting Information) or the temperature (Figure S9
in the Supporting Information) have to be augmented to
ensure a constant aagg. The narrowing of NMR signals ob-
served for both cases can be attributed to kinetically less
stable dimers at larger chloroform content or higher temper-
ature, resulting in faster exchange rates.
These results suggest that the onset of NMR signal broad-
ening with increasing MCH content and/or PBI 6 concentra-
tion correlates with the selective solvation phenomenon dis-
cussed in the previous paragraph. For a sufficient amount of
chloroform, the PBI p faces are preferentially solvated by
chloroform molecules connected with a fast exchange be-
tween monomers and dimers due to the “weakening” of p–
p interactions by the “good” solvent chloroform. Below a
critical chloroform content (which increases with PBI 6 con-
centration) the exchange rate between monomers and
dimers slows down, owing to the solvophobic effect of
MCH. In addition, dimers start to assemble into larger
oligomers resulting in a further decrease of exchange fre-
quencies in pure MCH solvent.
isosACTHNUGRTENUNGbestic points in concentration- or temperature-depend-
ent UV/Vis experiments were analyzed not only in terms of
a simple dimerization (monomer–dimer) or an isodesmic
growth (monomer–oligomer). Instead, we have provided ex-
perimental evidence for the formation of larger aggregates
composed of p-dimeric units. The formation of such aggre-
gates by various types of complex self-assembly pathways[47]
might indeed be very common, but has been widely over-
looked in the past. Our efforts to program discrete self-as-
sembled PBI stacks are still in progress. Our next target is
to produce uniform and kinetically stable PBI aggregates.
Conclusion
Experimental Section
In this work, we reported a novel PBI building block featur-
ing one rigid and one flexible substituent at the two imide
positions. By virtue of the steric demand of these substitu-
ents, the designed PBI is prone to form discrete p–p-stacked
All solvents and reagents were purchased from commercial sources and
used as received without further purification, unless otherwise stated.
Dry solvents were purified according to literature procedures.[48] Benzyl
alcohol 1 and perylene monoimide 5 were synthesized as described in lit-
erature.[18,19] Thin-layer chromatography (TLC) was conducted on alumi-
num plates coated with silica gel (60 F254, Merck). Column chromatogra-
phy was performed using silica gel (Geduran Si 60 from Merck, particle
size 0.040–0.063 mm) as stationary phase. Recycling preparative HPLC
was performed on a system (JAI LC-9105) equipped with a photodiode
array UV/Vis detector (JAI 3702) and an RI detector (JAI RI-7 s) by use
of a semipreparative NUCLEOSIL 100–5 NO2 column (Macherey&Na-
gel). For HPLC separation, HPLC grade solvents (CHCl3 and MeOH)
from VWR (Darmstadt, Germany) were used. High-resolution electro-
spray ionization (ESI) mass spectra were measured on a MicroTOF
Focus instrument (Bruker Daltronik GmbH) and MALDI-TOF measure-
ments were carried out on a Bruker Autoflex II.
dimers rather than extended oligomeric p stacks. The di
ACHTUNGTNERmNUNG er-
ACHTUNGTRENNUNG
nonlinear least-squares analysis of the concentration-de-
1
pendent H NMR and UV/Vis spectra in terms of a dimer
model. DOSY NMR spectroscopy further confirmed the ab-
sence of aggregates larger than dimer. The observed shifts
of proton resonances in NMR dilution experiments indicate
the formation of a head-to-tail p stack composed of two
PBIs with a certain rotational displacement. Detailed eluci-
dation of the head-to-tail m–m dimer structure with a rota-
tional angle of about 308 is corroborated by means of
ROESY NMR spectra associated with molecular modeling
studies. Consequently, we have not only programmed PBI
UV/Vis absorption studies: The UV/Vis absorption measurements were
performed on a Perkin–Elmer Lambda 35 spectrometer making use of
conventional quartz cells with 0.01–10 mm path length to cover a large
concentration range. The slit width and the scan speed were set to be
&
10
&
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Perylene Bisimide Dimer Aggregates
FULL PAPER
1.0 nm and 120 nmminꢀ1, respectively. The temperature was controlled
with a PTP-1+1 Peltier temperature programmer (Perkin–Elmer). In
order to avoid nonsystematic error, all the samples having a concentra-
tion of more than 0.5 mm were individually prepared by carefully dissolv-
ing the precisely weighed compound in solvents, assuming that the
volume of the compound is negligible after dissolution. The samples with
a concentration less than 0.5 mm were diluted ten times from a corre-
sponding stock solution. For a comparative study with NMR spectra,
UV/Vis dilution experiments were conducted in the deuterated solvents.
(ESI-TOF, pos. mode): m/z calcd for C31H56N3O2: 502.43670 [M+H]+;
found: 502.43631; elemental analysis calcd (%) for C31H55N3O2: C 74.20,
H 11.05, N 8.37; found: C 74.26, H 11.21, N 8.35.
3,5-Bis(dodecyloxy)benzyl amine (4): Triphenylphosphine (0.65 g,
2.5 mmol) was added to a solution of compound 3 (0.61 g, 1.2 mmol) in
THF (10 mL). After the reaction mixture was stirred for 2 h at room
temperature, distilled H2O (0.2 mL, 11.1 mmol) was added and resulting
mixture was stirred for another 3 h. After reaction, the solvent was re-
moved under reduced pressure. The residue was dissolved in CH2Cl2
(20 mL) and washed with 5% NaOH solution (3ꢄ20 mL). The aqueous
layer was extracted with CH2Cl2 (3ꢄ20 mL) and the combined organic
layers were concentrated in vacuum and purified by a short column chro-
matography on silica gel with CH2Cl2/MeOH (v:v=97:3) as eluent. The
1H NMR and MS indicate a mixture of benzyl amine product and tri-
NMR spectroscopic studies: 1H, 13C, DEPT, COSY, HSQC, and
HMBC NMR spectra were recorded in standard 5 mm NMR tubes on a
Bruker Avance 400 and/or a Bruker DMX 600 spectrometer with TMS
or residual undeuterated solvents as internal standard (7.26 for CHCl3
and 77.00 ppm for CDCl3). The 2D ROESY and DOSY data were ac-
quired with the 600 MHz instrument, which was equipped with 5 mm
13C/1H cryoprobe with a z axis gradient coil capable of producing pulsed
magnetic field gradients of 55 Gcmꢀ1. For low- and high-temperature
NMR measurements, the temperature was calibrated using NMR samples
of 4% CH3OH in [D4]methanol and 80% ethylene glycol in [D6]DMSO,
respectively. Prior to 2D ROESY NMR measurements, the deuterated
solvents and the NMR tubes used were degassed by bubbling with dry
argon gas. Alternating-phase 1808 pulses were applied during the mixing
time of 200 ms for the ROESY spectra to suppress unwanted TOCSY
contributions.[49] In the DOSY experiments, the suppression of flow ef-
fects owing to temperature gradients in the coil of the cryoprobe was
achieved in two different ways: 1) using the stimulated echo BPP-LED
pulse sequence[50] (longitudinal eddy current delay sequence with bipolar
gradient pulse pairs for diffusion and additional spoil gradients after the
second and fourth 908 pulse) in which convection in the z direction was
suppressed by sample rotation[51] and 2) by the corresponding double
stimulated echo–pulse sequence[52] (without sample rotation and with
spoil gradients after the second, fourth, and sixth 908 pulse) in which the
double stimulated echo results in a compensation of the flow effects. The
following acquisition parameters were used for both methods: duration d
of a bipolar gradient pulse 4 or 6 ms, diffusion time D 50 ms, eddy current
delay 5 ms. The diffusion time D was kept constant in each DOSY experi-
ment while the sinusoidal diffusion gradients were incremented from 2 to
95% of maximum gradient strength in 32 linear steps.
ACHUTNGRENUpNG henAHCUTNGTRENyNNGU lphosphine oxide byproduct (molar ratio about 1:1), which was
used for next step without further purification. Yield: 0.43 g (0.9 mmol,
75%) as a white solid; TLC (CH2Cl2/MeOH=9:1): Rf =0.47; 1H NMR
(400 MHz, CDCl3): d=6.45 (d, J=2.4 Hz, 2H; Ph-H), 6.34 (t, J=2.4 Hz,
1H; Ph-H), 3.92 (t, J=6.6 Hz, 4H; OCH2CH2), 3.80 (s, 2H; PhCH2NH2),
1.77 (quint, J=6.6 Hz, 4H; OCH2CH2), 1.40–1.47 (m, 4H;
OCH2CH2CH2), 1.27 (brs, 32H; alkoxy-H), 0.88 ppm (t, J=6.8 Hz, 6H;
CH2CH3); HRMS (ESI-TOF, pos. mode): m/z calcd for C31H58N1O2:
476.44621 [M+H]+; found: 476.44612.
N-[3,5-Bis(dodecyloxy)benzyl]-N’-(2,5-di-tert-butylphenyl)perylene-
3,4:9,10-tetracarboxylic acid bisimide (6): A mixture of compounds 5
(116 mg, 0.20 mmol), 4 (240 mg, 0.75 mmol), and imidazole (10 g) was
stirred under an argon atmosphere at 1408C for 4 h. Before cooling, the
reaction was directly quenched with 2.0m HCl (50 mL) and stirred for an-
other 1 h. The resulting precipitate was separated by filtration and
washed with MeOH (3ꢄ10 mL). After drying in vacuum, the crude prod-
uct was purified first by silica gel chromatography (CH2Cl2 as eluent) and
then by using recycling HPLC (CHCl3 as eluent). Yield: 102 mg
(0.10 mmol, 50%) as a deep red solid; TLC (CH2Cl2): Rf =0.34; m.p.
168–1698C; 1H NMR (400 MHz, CDCl3): d=8.24–8.68 (m, 8H; perylene
protons), 7.61 (d, J=8.8 Hz, 1H; Ph-H), 7.49 (dd, J1 =8.8 Hz, J2 =2.4 Hz,
1H; Ph-H), 7.30 (t, J=2.4 Hz, 1H; Ph-H), 6.66 (d, J=2.4 Hz, 2H; Bn-
H), 6.36 (t, J=2.4 Hz, 1H; Bn-H), 5.29 (s, 2H; NCH2), 3.93 (t, J=6.4 Hz,
4H; OCH2CH2), 1.74 (quint, J=6.4 Hz, 4H; OCH2CH2), 1.40–1.45 (m,
4H; OCH2CH2CH2), 1.37 (s, 9H; tBu-H), 1.31 (s, 9H; tBu-H), 1.24 (brs,
32H; alkoxy-H), 0.87 ppm (t, J=6.8 Hz, 6H; CH2CH3); 13C NMR
(100 MHz, CDCl3): d=164.2, 162.9, 160.3, 150.3, 143.7, 139.0, 134.4,
134.1, 132.6, 131.5, 131.1, 129.6, 128.9, 128.7, 128.0, 126.3, 126.2, 125.9,
123.7, 123.0, 122.9, 122.8, 107.3, 100.3, 68.0, 43.7, 35.5, 34.3, 31.9, 31.7,
31.3, 29.8, 29.7, 29.6, 29.5, 29.4, 29.3, 29.2, 26.0, 22.7, 14.1 ppm; HRMS
(ESI-TOF, pos. mode): m/z calcd for C69H85N2O6: 1037.64021 [M+H]+;
found 1037.64073; elemental analysis calcd (%) for C69H84N2O6: C 79.88,
H 8.16, N 2.70; found: C 79.39, H 8.13, N 2.90.
3,5-Bis(dodecyloxy)benzyl bromide (2):
A mixture of compound 1
(1.52 g, 3.2 mmol), carbon tetrabromide (1.60 g, 4.8 mmol, 1.5 equiv), and
triphenylphosphine (1.26 g, 4.8 mmol, 1.5 equiv) in anhydrous THF
(16 mL) was stirred under an argon atmosphere at room temperature
overnight (ca. 16 h). The solvent was removed under reduced pressure
and the residue was washed with MeOH (3ꢄ10 mL). After that, the
crude product was purified by column chromatography on silica gel with
CH2Cl2/n-hexane (v:v=1:1) as eluent. Yield: 1.49 g (2.76 mmol, 87%) as
a white solid; TLC (CH2Cl2/n-hexane=1:1): Rf =0.83; m.p. 48–498C;
1H NMR (400 MHz, CDCl3): d=6.51 (d, J=2.0 Hz, 2H; Ph-H), 6.38 (t,
J=2.0 Hz, 1H; Ph-H), 4.40 (s, 2H; PhCH2Br), 3.93 (t, J=6.8 Hz, 4H;
OCH2CH2), 1.76 (quint, J=6.8 Hz, 4H; OCH2CH2), 1.41–1.48 (m, 4H;
OCH2CH2CH2), 1.27 (brs, 32H; alkoxy-H), 0.89 ppm (t, J=7.0 Hz, 6H;
CH2CH3); HRMS (ESI-TOF, pos. mode): m/z calcd for C31H56BrO2:
539.34582 [M+H]+; found 539.34587; elemental analysis calcd (%) for
C31H55BrO2: C 68.99, H 10.27; found: C 68.72, H 10.42.
N-(4-Iodobenzyl)-N’-(2,5-di-tert-butylphenyl)perylene-3,4:9,10-tetracar-
ACHUTNGRENUbNG oxACHUTGNTRENyNUGN lic acid bisimide (7): A mixture of compound 5 (100 mg, 0.17 mmol),
4-iodobenzyl amine (100 mg, 0.43 mmol), and imidazole (10 g) was stirred
under an argon atmosphere at 1408C for 4 h. Before cooling, the reaction
was directly quenched with 2.0m HCl (50 mL) and stirred for another
1 h. The resulting precipitate was separated by filtration and washed with
MeOH (3ꢄ10 mL). After drying in vacuum, the crude product was puri-
fied by silica gel chromatography (CH2Cl2 as eluent) and reprecipitated
between CH2Cl2/MeOH. Yield: 90 mg (0.11 mmol, 64%) as a red solid;
TLC (CH2Cl2): Rf =0.10; m.p. 388–3898C; 1H NMR (400 MHz, CDCl3):
d=8.54–8.74 (m, 8H; perylene protons), 7.65 (dt, J1 =8.8 Hz, J2 =2.0 Hz,
2H; Bn-H), 7.61 (d, J=8.8 Hz, 1H; Ph-H), 7.48 (dd, J1 =8.8 Hz, J2 =
2.4 Hz, 1H; Ph-H), 7.33 (dt, J1 =8.8 Hz, J2 =2.0 Hz, 2H; Bn-H), 7.13 (d,
J=2.4 Hz, 1H; Ph-H), 5.34 (s, 2H; NCH2), 1.35 (s, 9H; tBu-H), 1.29 ppm
(s, 9H; tBu-H). 13C NMR (100 MHz, CDCl3): d=164.3, 163.3, 150.2,
143.7, 137.6, 136.7, 134.8, 134.7, 132.6, 131.8, 131.6, 131.2, 129.8, 129.3,
128.8, 127.8, 126.6, 126.4, 126.3, 123.8, 123.2, 123.1, 123.0, 93.3, 43.3, 35.5,
34.3, 31.7, 31.2 ppm; HRMS (ESI-TOF, pos. mode): m/z calcd for
C45H36IN2O4: 795.17143 [M+H]+; found: 795.17179; elemental analysis
calcd (%) for C45H35IN2O4: C 68.01, H 4.44, N 3.53; found: C 68.00, H
4.52, N 3.58.
3,5-Bis(dodecyloxy)benzyl azide (3): A mixture of compound 2 (1.08 g,
2.0 mmol) and sodium azide (390 mg, 6.0 mmol, 3.0 equiv) in anhydrous
DMF (20 mL) was heated to 1058C for 3 h. After cooling, the DMF sol-
vent was removed under reduced pressure. The residue was dissolved in
ethyl ether (20 mL) and washed with H2O (3ꢄ20 mL) and brine (3ꢄ
20 mL). The aqueous layer was extracted with ethyl ether (3ꢄ20 mL)
and the combined organic layers were concentrated in vacuum and puri-
fied by column chromatography on silica gel with CH2Cl2/n-hexane
(v:v=1:1) as eluent. Yield: 0.99 g (2.0 mmol, 99%) as a white solid; TLC
(CH2Cl2/n-hexane=1:1): Rf =0.76; m.p. 36–378C; 1H NMR (400 MHz,
CDCl3): d=6.43 (d, J=2.4 Hz, 2H; Ph-H), 6.41 (t, J=2.4 Hz, 1H; Ph-
H), 4.25 (s, 2H; PhCH2N3), 3.94 (t, J=6.4 Hz, 4H; OCH2CH2), 1.77
(quint, J=6.4 Hz, 4H; OCH2CH2), 1.41–1.48 (m, 4H; OCH2CH2CH2),
1.27 (brs, 32H; alkoxy-H), 0.88 ppm (t, J=6.8 Hz, 6H; CH2CH3). HRMS
Chem. Eur. J. 2012, 00, 0 – 0
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
11
&
ÞÞ
These are not the final page numbers!

F. Wꢁrthner et al.
[18] I. Bury, B. Heinrich, C. Bourgogne, D. Guillon, B. Donnio, Chem.
Eur. J. 2006, 12, 8396–8413.
Acknowledgements
[19] H. Kaiser, J. Lindner, H. Langhals, Chem. Ber. 1991, 124, 529–535.
[20] H. Langhals, Chem. Ber. 1985, 118, 4641–4645.
[21] Z. Chen, B. Fimmel, F. Wꢁrthner, Org. Biomol. Chem. 2012, 10,
5845–5855.
[22] a) S. P. Brown, H. W. Spiess, Chem. Rev. 2001, 101, 4125–4156; b) J.
Hu, T. Xu, Y. Cheng, Chem. Rev. 2012, 112, 3856–3891; c) H.-J.
Schneider, F. Hacket, V. Rꢁdiger, H. Ikeda, Chem. Rev. 1998, 98,
1755–1786.
[23] a) J. Wu, A. Fechtenkçtter, J. Gauss, M. D. Watson, M. Kastler, C.
Fechtenkçtter, M. Wagner, K. Mꢁllen, J. Am. Chem. Soc. 2004, 126,
11311–11321; b) J. A. A. W. Elemans, A. E. Rowan, R. J. M. Nolte,
J. Am. Chem. Soc. 2002, 124, 1532–1540; c) D. Zhao, J. S. Moore, J.
Org. Chem. 2002, 67, 3548–3554.
[24] a) F. H. Beijer, H. Kooijman, A. L. Spek, R. P. Sijbesma, E. W.
Meijer, Angew. Chem. 1998, 110, 79–82; Angew. Chem. Int. Ed.
1998, 37, 75–78; b) C. Rether, E. Verheggen, C. Schmuck, Chem.
Commun. 2011, 47, 9078–9079; c) R. Wyler, J. de Mendoza, J.
Rebek, Angew. Chem. 1993, 105, 1820–1821; Angew. Chem. Int. Ed.
Engl. 1993, 32, 1699–1701.
[25] a) A. D. Q. Li, W. Wang, L.-Q. Wang, Chem. Eur. J. 2003, 9, 4594–
4601; b) W. Wang, L.-S. Li, G. Helms, H.-H. Zhou, A. D. Q. Li, J.
Am. Chem. Soc. 2003, 125, 1120–1121.
[26] a) V. Dehm, M. Bꢁchner, J. Seibt, V. Engel, F. Wꢁrthner, Chem. Sci.
2011, 2, 2094–2100; b) M. J. Ahrens, L. E. Sinks, B. Rybtchinski, W.
Liu, B. A. Jones, J. M. Giaimo, A. V. Gusev, A. J. Goshe, D. M.
Tiede, M. R. Wasielewski, J. Am. Chem. Soc. 2004, 126, 8284–8294;
c) B. Rybtchinski, L. E. Sinks, M. R. Wasielewski, J. Phys. Chem. A
2004, 108, 7497–7505; d) M. J. Ahrens, R. F. Kelley, Z. E. X. Dance,
M. R. Wasielewski, Phys. Chem. Chem. Phys. 2007, 9, 1469–1478.
[27] E. Arunan, G. R. Desiraju, R. A. Klein, J. Sadlej, S. Scheiner, I. Al-
korta, D. C. Clary, R. H. Crabtree, J. J. Dannenberg, P. Hobza, H. G.
Kjaergaard, A. C. Legon, B. Mennucci, D. J. Nesbitt, Pure Appl.
Chem. 2011, 83, 1619–1636.
We are grateful to the Deutsche Forschungsgemeinschaft for financial
support of this work within the research graduate school GK 1221 “Con-
trol of electronic properties of aggregates of p-conjugated molecules” at
the University of Wꢁrzburg. We are indebted to Ana-Maria Krause and
Marcel Gsꢂnger for the single-crystal analysis.
[1] a) C. A. Hunter, K. R. Lawson, J. Perkins, C. J. Urch, J. Chem. Soc.
Perkin Trans. 2 2001, 651–669; b) C. A. Hunter, J. K. M. Sanders, J.
Am. Chem. Soc. 1990, 112, 5525–5534; c) C. R. Martinez, B. L. Iver-
son, Chem. Sci. 2012, 3, 2191–2201.
[2] a) E. A. Meyer, R. K. Castellano, F. Diederich, Angew. Chem. 2003,
115, 1244–1287; Angew. Chem. Int. Ed. 2003, 42, 1210–1250;
b) L. M. Salonen, M. Ellermann, F. Diederich, Angew. Chem. 2011,
123, 4908–4944; Angew. Chem. Int. Ed. 2011, 50, 4808–4842.
[3] a) F. J. M. Hoeben, P. Jonkheijm, E. W. Meijer, A. P. H. J. Schenning,
Chem. Rev. 2005, 105, 1491–1546; b) R. Bhosale, J. Misek, N. Sakai,
S. Matile, Chem. Soc. Rev. 2010, 39, 138–149.
[4] C. Wang, H. Dong, W. Hu, Y. Liu, D. Zhu, Chem. Rev. 2012, 112,
2208–2267.
[5] a) A. Mishra, P. Bꢂuerle, Angew. Chem. 2012, 124, 2060–2109;
Angew. Chem. Int. Ed. 2012, 51, 2020–2067; b) T. M. Clarke, J. R.
Durrant, Chem. Rev. 2010, 110, 6736–6767.
[6] a) M. R. Wasielewski, Acc. Chem. Res. 2009, 42, 1910–1921; b) F.
Wꢁrthner, Chem. Commun. 2004, 1564–1579.
[7] a) F. Wꢁrthner, M. Stolte, Chem. Commun. 2011, 47, 5109–5115;
b) J. E. Anthony, A. Facchetti, M. Heeney, S. R. Marder, X. Zhan,
Adv. Mater. 2010, 22, 3876–3892.
[8] a) F. Wꢁrthner, Z. Chen, V. Dehm, V. Stepanenko, Chem. Commun.
2006, 1188–1190; b) Z. Chen, V. Stepanenko, V. Dehm, P. Prins,
L. D. A. Siebbeles, J. Seibt, P. Marquetand, V. Engel, F. Wꢁrthner,
Chem. Eur. J. 2007, 13, 436–449; c) V. Dehm, Z. Chen, U. Baumeis-
ter, P. Prins, L. D. A. Siebbeles, F. Wꢁrthner, Org. Lett. 2007, 9,
1085–1088.
´
[28] a) J. D. Badjic, S. J. Cantrill, J. F. Stoddart, J. Am. Chem. Soc. 2004,
[9] a) R. F. Fink, J. Seibt, V. Engel, M. Renz, M. Kaupp, S. Lochbrun-
ner, H.-M. Zhao, J. Pfister, F. Wꢁrthner, B. Engels, J. Am. Chem.
Soc. 2008, 130, 12858–12859; b) H.-M. 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, 15660–15668; c) J. Seibt, T. Winkler,
K. Renziehausen, V. Dehm, F. Wꢁrthner, H. D. Meyer, V. Engel, J.
Phys. Chem. A 2009, 113, 13475–13482.
[10] a) J. M. Giaimo, J. V. Lockard, L. E. Sinks, A. M. Scott, T. M.
Wilson, M. R. Wasielewski, J. Phys. Chem. A 2008, 112, 2322–2330;
b) C. Hippius, I. H. M. van Stokkum, E. Zangrando, R. M. Williams,
M. Wykes, D. Beljonne, F. Wꢁrthner, J. Phys. Chem. C 2008, 112,
14626–14638; c) A. Syamakumari, A. P. H. J. Schenning, E. W.
Meijer, Chem. Eur. J. 2002, 8, 3353–3361.
[11] a) D. Veldman, S. M. A. Chopin, S. C. J. Meskers, M. M. Groeneveld,
R. M. Williams, R. A. J. Janssen, J. Phys. Chem. A 2008, 112, 5846–
5857; b) H. Yoo, J. Yang, A. Yousef, M. R. Wasielewski, D. Kim, J.
Am. Chem. Soc. 2010, 132, 3939–3944; c) F. Gao, Y. Zhao, W.
Liang, J. Phys. Chem. B 2011, 115, 2699–2708.
[12] a) M. M. Safont-Sempere, P. Osswald, K. Radacki, F. Wꢁrthner,
Chem. Eur. J. 2010, 16, 7380–7384; b) M. M. Safont-Sempere, P.
Osswald, M. Stolte, M. Grꢁne, M. Renz, M. Kaupp, K. Radacki, H.
Braunschweig, F. Wꢁrthner, J. Am. Chem. Soc. 2011, 133, 9580–
9591.
126, 2288–2289; b) G. R. Desiraju, Acc. Chem. Res. 1991, 24, 290–
296; c) F. M. Raymo, M. D. Bartberger, K. N. Houk, J. F. Stoddart, J.
Am. Chem. Soc. 2001, 123, 9264–9267; d) T. Steiner, Angew. Chem.
2002, 114, 50–80; Angew. Chem. Int. Ed. 2002, 41, 48–76; e) A.
Senes, Proc. Natl. Acad. Sci. USA 2001, 98, 9056–9061.
[29] a) J. L. Cook, C. A. Hunter, C. M. R. Low, A. Perez-Velasco, J. G.
Vinter, Angew. Chem. 2007, 119, 3780–3783; Angew. Chem. Int. Ed.
2007, 46, 3706–3709; b) S. L. Cockroft, C. A. Hunter, Chem.
Commun. 2006, 3806–3808.
[30] R. B. Martin, Chem. Rev. 1996, 96, 3043–3064.
[31] In accordance with the values obtained by NMR analysis for differ-
ent protons, also here slightly different KD values were obtained at
different wavelengths: (410ꢁ50)mꢀ1 at 454 nm and (520ꢁ70)mꢀ1 at
485 nm.
[32] a) M. Kasha, H. R. Rawls, M. A. El-Bayoumi, Pure Appl. Chem.
1965, 11, 371–392; b) M. Kasha, Radiat. Res. 1963, 20, 55–70.
[33] Y. Cohen, L. Avram, L. Frish, Angew. Chem. 2005, 117, 524–560;
Angew. Chem. Int. Ed. 2005, 44, 520–554.
[34] In the strict sense, the viscosity of the sample solution rather than
that of the solvent should be used to calculate the hydrodynamic
radius. With higher concentration, the viscosity of the medium be-
comes larger, leading to slightly exaggerated value for hydrodynam-
ic radius. The measurement of a precise viscosity value for the
40 mm solution could not be performed due to the considerable
amount of compound needed.
[13] N.-T. Lin, A. Vargas Jentzsch, L. Guenee, J.-M. Neudorfl, S. Aziz, A.
Berkessel, E. Orentas, N. Sakai, S. Matile, Chem. Sci. 2012, 3, 1121–
1127.
[35] H. Mo, T. C. Pochapsky, Prog. Nucl. Magn. Reson. Spectrosc. 1997,
30, 1–38.
[14] Z. Chen, A. Lohr, C. R. Saha-Mçller, F. Wꢁrthner, Chem. Soc. Rev.
2009, 38, 564–584.
[36] Owing to the signal broadening, the ROESY NMR spectrum of PBI
[15] H. Langhals, S. Demmig, H. Huber, Spectrochim. Acta Part A 1988,
44, 1189–1193.
[16] J. M. Lavin, K. D. Shimizu, Chem. Commun. 2007, 228–230.
[17] D.-S. Choi, Y. S. Chong, D. Whitehead, K. D. Shimizu, Org. Lett.
2001, 3, 3757–3760.
6 was first recorded with the sample in CDCl /
at 1 mm. A more clear ROESY spectrum was obtained with the
sample in CDCl3/A[D14]MCH (v:v=1:2) at 10 mm. In addition, the
sample in CDCl3 at 50 mm has also been measured (Figure S5 in the
3 ACHTUNGTRENN[NUG D14]MCH (v:v=1:5)
CTHUNGTRENNUNG
&
12
&
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Perylene Bisimide Dimer Aggregates
FULL PAPER
Supporting Information). All of the three ROESY spectra exhibit
the same and reproducible cross signals.
[44] S. Grimme, Angew. Chem. 2008, 120, 3478–3483; Angew. Chem. Int.
Ed. 2008, 47, 3430–3434.
[37] Maestro, version 9.1, MacroModel, version 9.8, Schrçdinger, LLC,
New York, NY, 2011.
[38] C. A. Hunter, H. L. Anderson, Angew. Chem. 2009, 121, 7624–7636;
Angew. Chem. Int. Ed. 2009, 48, 7488–7499.
[39] T. E. Kaiser, V. Stepanenko, F. Wꢁrthner, J. Am. Chem. Soc. 2009,
131, 6719–6732.
[40] a) C. A. Hunter, J. Chem. Soc. Chem. Commun. 1991, 749–751;
b) W. B. Jennings, B. M. Farrell, J. F. Malone, Acc. Chem. Res. 2001,
34, 885–894; c) O. Takahashi, Y. Kohno, M. Nishio, Chem. Rev.
2010, 110, 6049–6076.
[45] J. S. Moore, C. R. Ray, Adv. Polym. Sci. 2005, 177, 91–149.
[46] R. B. Prince, J. G. Saven, P. G. Wolynes, J. S. Moore, J. Am. Chem.
Soc. 1999, 121, 3114–3121.
[47] a) P. A. Korevaar, S. J. George, A. J. Markvoort, M. M. J. Smulders,
P. A. J. Hilbers, A. P. H. J. Schenning, T. F. A. de Greef, E. W.
Meijer, Nature 2012, 481, 492–496; b) A. Lohr, F. Wꢁrthner, Isr. J.
Chem. 2011, 51, 1052–1066; c) Y. Tidhar, H. Weissman, S. G. Wolf,
A. Gulino, B. Rybtchinski, Chem. Eur. J. 2011, 17, 6068–6075.
[48] D. D. Perrin, W. L. F. Armarego, Purification of Laboratory Chemi-
cals, Pergamon, Oxford (UK), 1980.
[41] J. N. H. Reek, A. H. Priem, H. Engelkamp, A. E. Rowan, J. A. A. W.
Elemans, R. J. M. Nolte, J. Am. Chem. Soc. 1997, 119, 9956–9964.
[42] It is worth to note that the K value of 14000mꢀ1 is more than one
order of magnitude smaller than the binding constant observed for
extended PBI aggregation in the absence of bulky substituents at
the imide positions (see reference [21]). Accordingly, the proposed
packing arrangement only accounts for the special situation where
extended PBI p–p interaction is impossible.
[43] a) M. S. Cubberley, B. L. Iverson, J. Am. Chem. Soc. 2001, 123,
7560–7563; b) S. Lahiri, J. L. Thompson, J. S. Moore, J. Am. Chem.
Soc. 2000, 122, 11315–11319; c) F. Wꢁrthner, S. Yao, T. Debaerde-
maeker, R. Wortmann, J. Am. Chem. Soc. 2002, 124, 9431–9447;
d) G. A. Breault, C. A. Hunter, P. C. Mayers, J. Am. Chem. Soc.
1998, 120, 3402–3410.
[49] a) T. L. Hwang, A. J. Shaka, J. Am. Chem. Soc. 1992, 114, 3157–
3159; b) T.-L. Hwang, M. Kadkhodaei, A. Mohebbi, A. J. Shaka,
Magn. Reson. Chem. 1992, 30, S24–S34.
[50] D. H. Wu, A. D. Chen, C. S. Johnson, J. Magn. Reson. Ser. A 1995,
115, 260–264.
[51] a) J. Lounila, K. Oikarinen, P. Ingman, J. Jokisaari, J. Magn. Reson.
Ser. A 1996, 118, 50–54; b) N. Esturau, F. Sꢆnchez-Ferrando, J. A.
Gavin, C. Roumestand, M.-A. Delsuc, T. Parella, J. Magn. Reson.
2001, 153, 48–55.
[52] A. Jerschow, N. Mꢁller, J. Magn. Reson. 1997, 125, 372–375.
Received: May 11, 2012
Published online: && &&, 0000
Chem. Eur. J. 2012, 00, 0 – 0
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
13
&
ÞÞ
These are not the final page numbers!

F. Wꢁrthner et al.
Self-assembly control: Proper steric
control can program the formation of
well-defined perylene bisimide (PBI)
dimer aggregates (see figure) in
Perylene Bisimides
C. Shao, M. Grꢀne, M. Stolte,
F. Wꢀrthner* .................. &&&& — &&&&
chloroform (“good” solvent). The fur-
ther growth of PBIs into extended
aggregates can be triggered by the
addition of methylcyclohexane (“bad”
solvent). These processes have been
investigated in great detail and may
serve as a guideline for an in-depth
characterization of dye aggregates.
Perylene Bisimide Dimer Aggregates:
Fundamental Insights into Self-
assembly by NMR and UV/Vis Spec-
troscopy
Widely applied automotive color
pigments…
…based on perylene bisimide have
hitherto been utilized by many
groups to create extended one-
dimensional p stacks by self-
assembly. In their Full Paper on
page && ff., Wꢁrthner et al.
report on the application of steri-
cally demanding groups to termi-
nate self-assembly at the dimer
stage. Only in highly nonpolar
aliphatic solvents is a rather
unusual growth into larger self-
assembled structures observed.
&
14
&
www.chemeurj.org
ꢃ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!