Photoluminescence and conductivity of self-assembled π-π stacks of perylene bisimide dyes

Article abstract of DOI:10.1002/chem.200600889

The self-assembly of a new, highly fluorescent perylene bisimide dye 2 into π stacks, both in solution and condensed phase, has been studied in detail by NMR spectroscopy, vapor pressure osmometry (VPO), UV/Vis and fluorescence spectroscopy, differential

Full text of DOI:10.1002/chem.200600889

DOI: 10.1002/chem.200600889
Photoluminescence and Conductivity of Self-Assembled p–p Stacks of
Perylene Bisimide Dyes
[
a]
[a]
[a]
[b]
Zhijian Chen, Vladimir Stepanenko, Volker Dehm, Paulette Prins,
[
b]
[c]
[c]
[c]
Laurens D. A. Siebbeles, Joachim Seibt, Philipp Marquetand, Volker Engel, and
Frank Würthner*
[
a]
Dedicated to Professor Klaus Müllen on the occasion of his 60th birthday
Abstract: The self-assembly of a new,
highly fluorescent perylene bisimide
dye 2 into p stacks, both in solution
and condensed phase, has been studied
in detail by NMR spectroscopy, vapor
pressure osmometry (VPO), UV/Vis
and fluorescence spectroscopy, differ-
ential scanning calorimetry (DSC), op-
tical polarizing microscopy (OPM) and
X-ray diffraction. The NMR and VPO
measurements revealed the formation
of extended p–p stacks of the dye mol-
ecules in solution. The aggregate size
determined from VPO and DOSY
NMR measurements agree well with
that obtained from the concentration
and temperature-dependent UV/Vis
spectral data by employing the isodes-
served upon aggregation which is ac-
companied by an increase of the fluo-
rescence lifetime and depolarization.
The observed absorption properties
can be explained in terms of molecular
exciton theory. The charge transport
properties of dye 2 have been investi-
gated by pulse radiolysis-time resolved
microwave conductivity measurements
and a 1D charge carrier mobility up to
mic model (equal K model). In the
condensed state, dye 2 possesses a hex-
agonal columnar liquid crystalline (LC)
phase as confirmed by X-ray diffrac-
tion analysis. The columnar stacking of
this dye has been further explored by
atomic force microscopy (AFM). Well-
resolved columnar nanostructures of
the compound are observed on graph-
ite surface. A color-tunable lumines-
cence from green to red has been ob-
2
ꢀ1 ꢀ1
0.42 cm V
s
is obtained. Consider-
ing the promising self-assembly, semi-
conducting, and luminescence proper-
ties of this dye, it might serve as a
useful functional material for nano-
Keywords: charge carrier mobility ·
dyes/pigments
· liquid crystals ·
A
H
R
U
G
perylene bisimide · self-assembly
Introduction
Functional dye assemblies possessing energy and charge
transport properties have attracted considerable attention in
[
1]
materials as well as in biological sciences in the past years.
Inspired by the pivotal role of self-assembled chlorophyll
[
a] Z. Chen, V. Stepanenko, V. Dehm, Prof. F. Würthner
Universität Würzburg, Institut für Organische Chemie
and Rçntgen Research Center for Complex Material Systems
Am Hubland, 97074 Würzburg (Germany)
Fax : (+49)931-888-4756
E-mail: wuerthner@chemie.uni-wuerzburg.de
[2]
dyes in the photosynthesis of green plants, efforts have
been devoted to the rational control of the self-assembly
process of artificial functional dyes in particular with respect
to applications in organic electronic devices. For example,
columnar liquid-crystalline and amphiphilic tubule-forming
hexabenzocoronenes have been successfully prepared that
[
b] P. Prins, Prof. L. D. A. Siebbeles
Opto-electronic Materials Section, DelftChemTech
Delft University of Technology
2
628 BL Delft (The Netherlands)
[3]
exhibit promising optical and electrical properties. Further-
[
c] J. Seibt, P. Marquetand, Prof. V. Engel
Universität Würzburg, Institut für Physikalische Chemie
Am Hubland, 97074 Würzburg (Germany)
more, well-ordered multi-component assemblies of p- and n-
type semiconducting p-electron systems have been con-
structed for efficient charge carrier separation upon illumi-
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author: Additional spec-
troscopic data, aggregation studies and other measurements of PBI 2.
[4]
nation.
436
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 436 – 449

FULL PAPER
Among functional dyes, perylene tetracarboxylic acid bisi-
ing was also observed for a calix[4]arene-functionalized PBI
derivative, which contains alkoxyphenyl groups at imide po-
[5]
mides (PBIs) have been investigated quite intensively due
to their favorable properties, such as light fastness, intense
[19]
sitions.
The fluorescence quenching in these chromo-
[6]
photoluminescence, and outstanding n-type semiconductiv-
phores is attributed to photoinduced electron transfer from
[
7]
ity. Thus, these dyes and their assemblies find application
the electron-rich alkyloxyphenyl groups to the perylene bisi-
[
8]
[15a,20]
as functional units in artificial light harvesting systems,
mide unit.
In contrast, recently we have observed that
[9]
photoinduced electron transfer systems, and organic elec-
tronic devices such as organic light emitting diodes
PBI dye 2 (Scheme 1), which contains trialkylphenyl groups,
instead of trialkoxyphenyl groups at the imide positions, and
[
10]
[11]
[21]
(
OLED),
organic thin-film transistors (OTFT),
and
its aggregates exhibit unique fluorescence properties.
[12]
solar cells. Different types of intermolecular forces such
as hydrogen bonding, p–p stacking and metal–ligand inter-
actions have been applied to direct the formation of desira-
Here we present our detailed investigations on self-assem-
bly, both in solution and condensed state (liquid crystallini-
ty), of PBI dye 2, and report on optical and charge transport
properties of its p–p-stacked aggregates.
[5]
ble supramolecular structures of PBIs. Among these non-
covalent interactions, p–p stacking plays an important role
in the self-assembly of PBI dyes and can determine the tex-
ture of thin films prepared by solution processible techni-
ques such as spin-coating or zone-casting for device fabrica-
Results and Discussion
[13]
tion. In some cases it has been shown that p stacks can be
Synthesis: The synthesis of the trialkylphenyl-functionalized
PBI 2 was achieved in three steps (Scheme 1). First, the pre-
cursor 5 was synthesized according to Sonogashira coupling
transferred on substrates by simple solution casting
[14]
method. Another important property of PBIs that is de-
[22]
termined by p–p stacking is their thermotropic or lyotropic
reaction
zene
between 1-dodecyne and 3,4,5-triiodonitrobe-
in 90% yield. Subsequent reduction of the triple
[15]
[23]
liquid crystallinity, which leads to the formation of colum-
[16]
nar dye stacks that can enable 1D charge carrier mobility.
bonds and the nitro group of 5 was accomplished by Pd-cat-
alyzed hydrogenation to give aniline 6. Finally, the conden-
sation of 6 with perylene tetracarboxylic bisanhydride in the
It has been found that in such liquid crystalline (LC) phase,
the transport of charge carriers along the column axis is sev-
eral orders of magnitude faster than that in perpendicular
[
24]
presence of Zn
A
H
R
U
(OAc) as catalyst
afforded the desired
2
[17]
1
direction.
PBI 2, which was fully characterized by H NMR spectrosco-
py, MS spectrometry, and elemental analysis. The PBIs 3
and 4 were used as reference compounds.
In general, the liquid crystallinity of polyaromatic com-
pounds arises from micro-segregation of a large-size rigid ar-
[25]
[18]
omatic core and flexible, non-bulky alkyl chains. In our
[15a]
previous work,
tridodecyloxyphenyl groups were intro-
Temperature and concentration-dependent aggregate forma-
duced at the imide positions of a PBI to obtain the LC PBI
tion in solution: The aggregation behavior of PBI 2 was first
studied by temperature-dependent H NMR spectroscopy.
1
dye 1 (Scheme 1) which exhibits high charge carrier mobili-
[
15d]
ty.
However, neither in monomeric nor in aggregated
In CDCl solution, a simple pattern of sharp signals for the
3
state is dye 1 fluorescent. Noteworthy, fluorescence quench-
perylene protons was observed, indicating that the com-
pound is in monomeric form in
this polar solvent (Figure 1).
However, when it was mea-
sured in nonpolar deuterated
methylcyclohexane
(
[D ]MCH), drastic changes
14
compared with the spectrum in
CDCl3 were observed, appa-
rently due to aggregation. In
[
D ]MCH at 299 K, the spec-
14
trum displayed a large number
of signals that overlap with
each other in the range be-
tween 9.2 to 6.5 ppm due to dif-
ferent aggregated species. The
complex pattern of the spec-
trum and high or low-field shift
of proton signals, compared
with those of monomeric spe-
cies, can be interpreted in terms
of shielding or deshielding of
the protons by p systems of
[
15a]
Scheme 1. Chemical structures of the trialkoxyphenyl-containing PBI dye 1,
trialkylphenyl-functionalized
Cl ], CuI, Et N, 80 8C; 4 h,
, Pd/C, EtOH/EtOAc; RT, 24 h, 53%. c) Perylene-3,4:9,10-tetracarboxylic acid bisanhydride, Zn-
2
, quinoline, 1808C, 2 h, 56%.
[
25]
PBI 2, and the two reference PBIs 3 and 4. Synthesis of PBI 2: a) [Pd
0%. b) H
(OAc)
A
H
R
U
G
3
)
2
2
3
9
2
A
C
H
T
R
E
U
N
G
Chem. Eur. J. 2007, 13, 436 – 449
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
437

F. Würthner et al.
[27]
py) NMR experiments were performed. The translational
2
ꢀ1
diffusion coefficient D [m s ] of molecular species can be
estimated by the Stokes-Einstein equation: D=k T/(6phR),
ACHTREUNG
B
where T is the temperature, h is the viscosity of the solvent,
R is the hydrodynamic radius of the molecule or assembly
and kB is the Boltzmann constant. According to this equa-
tion, the D value is inversely proportional to the hydrody-
namic radius (R) of a molecule. Thus, the size of species
could be compared based on their diffusion coefficients
when the other parameters are identical. For PBI 2, only a
single apparent diffusion coefficient was observed at differ-
ent concentrations for all the aggregated species. The same
observation was made in a recent DOSY NMR study on the
aggregation of proteins and was interpreted as a result of
ensemble averaging of the diffusion coefficients of different
[28]
species on macroscopic scale. Such averaging could also
be feasible for monomer–dimer equilibrium or equilibria in-
[29]
volving higher aggregated species.
If only monomer–
1
Figure 1. Temperature-dependent 600 MHz H NMR spectra of PBI 2
(
dimer equilibrium exists, the D values for both monomer
and dimer species should be nearly identical, as shown by
our previous studies on a coordination dimer of terpyridine-
ꢀ
3
3
1.1”10 m) in [D14]MCH at 299 to 370 K and in CDCl at 298 K. (The
signals labeled as a, b are assigned to perylene ring protons and signal c
to the protons of phenyl groups at imide positions.)
[30]
substituted perylene bisimide.
On the other hand, for
highly aggregated species which have much larger R values
than monomers, obviously, a change of D values should be
observable on DOSY NMR spectra.
neighboring dye molecules that are stacked closely in aggre-
gates (could be as close as 3.4 Š according to the crystal
structures of unsubstituted PBIs ). The H NMR spectra
[5]
1
As the monomeric form of 2 can only be obtained at very
low concentrations (see UV/Vis studies), a reference PBI 3,
observed for 2 in [D ]MCH are significantly different from
1
4
ꢀ
1
those reported for a related PBI derivative in CDCl where
which has a nearly identical molecular mass (1528 gmol )
3
ꢀ
1
fast exchange between monomer and aggregates took
as 2 (1552 gmol ) and does not aggregate owing to its ex-
[26]
place. In the latter case, the spectra were of simple pat-
tern that are comparable with those of monomer species
and only high-field shifts of the perylene proton signals
were observed. Thus, the complex spectral pattern of 2 in
tremely bulky substituents (verified by UV/Vis spectroscopy
ꢀ
2
ꢀ1
and VPO at 10 molL in MCH), was used for compari-
son. For all the aromatic signals of 2, D value of 1.57
ꢀ
10
2
ꢀ1
ꢀ3
”10 m s was observed at a concentration of 1.1”10
m
[
D ]MCH at 299 K indicates that the formed aggregates are
(Figure S2 in the Supporting Information), while the refer-
14
kinetically relatively stable on NMR time scale at this tem-
perature. Upon gradual increase of temperature from 299 to
ence compound 3 showed diffusion coefficient of 3.40”
ꢀ
10
2
ꢀ1
ꢀ3
10 m s at 1.3”10 m. The large difference between the
D values of 2 and monomeric 3 indicates that the hydrody-
namic radius of the present species of 2 is much higher than
that of the reference compound. These data clearly indicate
formation of extended aggregates of dye 2 in MCH. Further-
more, a gradual decrease of the D value from 1.57
360 K, the sharp signals of aggregates at 299 K have increas-
ingly broadened and at 370 K (near the boiling point of the
solvent) the spectrum displayed a relatively simple pattern
with three broad signals. One of them appears at 7.0 ppm
which corresponds to the chemical shift for imide phenyl
protons in the monomer and the remaining two high-field
shifted signals belong to perylene protons. These spectral
changes indicate the “melting” of the aggregates. At higher
temperatures than RT, more rapid exchanges between mo-
nomer and aggregated species take place and only averaged
proton signals can be observed, as the spectrum at 370 K
shows. Furthermore, in the concentration-dependent
ꢀ
10
2
ꢀ1
ꢀ11
2
ꢀ1
”10 m s to 7.36”10 m s was observed, when the
ꢀ
3
concentration of the solution was increased from 1.1”10
to 0.11m (Figure 2), indicating the increase of the hydrody-
namic radius of the aggregates.
This diffusion method can also provide the average aggre-
gation number per stack as the cube root of molecular
weight of aggregates has been suggested to be proportional
1
ꢀ1
H NMR spectra of 2 in [D ]MCH at 298 K (Figure S1 in
with D by simplistically taking all oligomers to be hydro-
1
4
[
27,31]
the Supporting Information), gradual changes in spectral
shape and an increase of signal number for aromatic protons
were observed with increasing concentration of 2. These
temperature and concentration-dependent NMR studies
reveal formation of p–p-stacked oligomeric aggregates of
PBI 2 in nonpolar MCH.
dynamically spherical.
Thus, the average aggregate size
3
of 2 can be estimated as NDOSY ꢁ(D /D) , since the differ-
ref
ence of the molecular mass of 2 and reference compound 3
is negligible, where Dref and D are the translational diffusion
coefficients obtained from DOSY NMR measurements for
reference compound 3 and PBI 2, respectively. It is worth to
note that this estimation is only an approximate approach
since the hydrodynamic shape of the aggregates deviates
To further substantiate the formation of extended PBI 2
aggregates in solution, DOSY (diffusion ordered spectrosco-
438
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 436 – 449

Perylene Bisimide Dyes
FULL PAPER
average aggregate sizes of 9, 7, and 6 aggregated molecules,
respectively, for dye 2. The results of DOSY NMR and
VPO experiments confirm convincingly that in MCH ex-
tended aggregates of 2, rather than dimer, are preferentially
formed at different concentrations and temperatures.
To obtain more insight into the thermodynamics of the
aggregation process of PBI 2, concentration and tempera-
ture-dependent UV/Vis spectroscopic studies were per-
formed. At very low concentrations (non-aggregated state),
the UV/Vis spectra (Figure 3) displayed absorption bands
between 400 and 550 nm for the S –S transition of the PBI
0
1
chromophore with lmax =517 nm and well-resolved vibronic
structure that can be ascribed to the breathing vibration of
the perylene skeleton which strongly coupled with the S –S
0
1
32]
[
transition polarized along the long axis of the molecules.
Upon increasing the concentration, the apparent absorption
coefficients decrease gradually and the spectra become
broader and less structured, reflecting the electronic cou-
pling between the chromophores. Furthermore, a hypsochro-
Figure 2. DOSY NMR spectra of a) PBI 2 at c=0.11m and b) the refer-
ꢀ
3
ence compound 3 at c=1.3”10 m in [D14]MCH. The diffusion coeffi-
2
ꢀ1
cients D [m s ] are plotted in a logarithmic scale against chemical shift
d [ppm].
from spherical shape. The average numbers of aggregated
dye 2 were estimated at three different concentrations and
are given in Table 1.
Table 1. Average number of aggregated molecules N of PBI 2 at differ-
ent concentrations and temperatures obtained from DOSY, VPO and
UV/Vis spectroscopic measurements.
ꢀ
1
T [8C]
c
T
[molL
]
N
VPO
N
DOSY
N
UV/Vis
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
3
2
1
3
3
3
25
25
25
40
50
60
1.1”10
2.2”10
1.1”10
5.0”10
5.0”10
5.0”10
10
20
99
10
42
93
11
8
9
7
6
6
Owing to the strong convection effect at higher tempera-
tures in solution, it is difficult to obtain the DOSY spectra
at higher temperatures. Therefore, temperature-dependent
aggregation of 2 was studied by vapor pressure osmometry
Figure 3. a) Concentration-dependent UV/Vis absorption spectra of PBI
ꢀ7
ꢀ3
2
(2.1”10 to 1.1” 10 m) in MCH at room temperature. Arrows indi-
cate the changes upon increasing the concentration. Inset: Fraction of ag-
gregated molecules aagg for PBI 2 as a function of concentration in MCH
obtained by fitting the apparent absorption coefficients at l=517 nm to
the isodesmic model. b) Temperature-dependent UV/Vis absorption
(
VPO), which is based on the colligative properties and pro-
vides average molecular weight of the aggregates formed.
ꢀ
3
ꢀ1
At 40, 50, and 608C and a concentration of 5.0”10 m ,
ꢀ4
spectra of 2 from 0 to 908C (5.0”10 m). Arrows indicate the changes
4
4
3
average molecular weight of 1.4”10 , 1.1”10 , and 9”10
Dalton were obtained, respectively, which correspond to
upon decreasing the temperature (90 to 08C). Inset: aagg of 2 versus tem-
perature.
Chem. Eur. J. 2007, 13, 436 – 449
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
439

F. Würthner et al.
ꢀ
1
mic shift of lmax of about 27 nm (1000 cm ) and a progres-
sive evolution of a new broad bathochromically shifted band
at 538 nm is observed. At 490 and 529 nm, there appear to
be isosbestic points, which, however, shift slightly from 490
to 492 nm and 529 to 531 nm upon increasing concentration.
These spectral features are highly indicative for the forma-
tion of face-to-face stacked dye aggregates.
(average aggregation number) at different concentrations
and temperatures can be calculated from the equations 1
and 2, respectively, where K is the aggregation constant and
cT is the concentration of 2. From these equations, the aver-
age numbers of aggregated dyes versus concentration at dif-
ferent temperatures were calculated (Figure 5). The ob-
tained values are shown in Table 1 and they are in good
agreement with those determined from DOSY and VPO
measurements.
In general, the p–p stacking of disk-shaped molecules
generates one-dimensional aggregates reversibly in solution
and the aggregation behavior can be ascribed to the strong
p–p interactions among discotic molecules. Such open-
ended linear aggregates are usually polydisperse and their
formation can be described by equal K or isodesmic chemi-
[33]
cal equilibria, in which a constant K value is assumed for
all the binding processes. Nonlinear least-square regression
analysis of the concentration-dependent UV/Vis spectral
data showed that the aggregation behavior can be well de-
scribed by this model. The value of K at room temperature
4
ꢀ1
was determined as 9.7”10 Lmol , which is about two
orders of magnitude smaller than that of the previously re-
[15a]
ported structurally similar PBI dye 1
in same solvent, in-
dicating that the chemical nature of the substituents play an
important role in the aggregation strength.
In the temperature-dependent UV/Vis spectra of 2 (Fig-
ure 3b), the spectral changes are highly comparable with
those observed in concentration-dependent measurements.
[34]
A “melting” temperature of 478C was calculated for the
phase transition from aggregated dyes to molecularly dis-
solved species for 2. Temperature-dependent UV/Vis spectra
of dye 2 were recorded at different concentrations and ag-
gregation constants K were determined at seven different
temperatures (Table S1, Supporting Information). The en-
thalpy and entropy contributions to the Gibbs free energy
changes for the aggregation of 2 were evaluated from a
Figure 5. a) Molar fraction of n-mer aggregates a
n
of PBI 2 as a function
ꢀ
1
vanꢁt Hoff plot (Figure 4). A plot of lnK versus T shows a
perfectly linear relationship (correlation coefficient 0.999)
and the standard enthalpy and entropy are determined to be
of total concentration c . b) Calculated average aggregation size N of 2
T
versus total concentration c
T
at different temperatures from 283 to 343 K
with 10 K interval.
o
ꢀ1
o
DH
3
=
ꢀ
ꢀ56.3ꢂ1.1 kJmol
and DS
=
ꢀ95.8ꢂ
1
ꢀ1
.6 Jmol K . The negative enthalpy and entropy values
indicate that the self-assembly process is enthalpy driven.
Based on isodesmic model, the molar fractions of n-mer
aggregates an and the average dye numbers per stack N
ꢀ
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ꢁ
n
2Kc þ 1ꢀ 4Kc þ 1
an ¼ ncTnꢀ1
K
nꢀ1
T
T
ð1Þ
ð2Þ
2
2
2K c
T
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
2
N ¼ ð1 þ 4Kc þ 1Þ
T
Although UV/Vis spectroscopy was performed at lower con-
centration range than for DOSY NMR and VPO measure-
ments, the data summarized in Table 1 show that in most
cases the aggregate size derived from UV/Vis spectroscopy
at high concentrations agrees well with those determined
from NMR and VPO measurements. Thus, a combination of
different methods, such as UV/Vis spectroscopy, diffusion
NMR and VPO, that have different concentration windows
can be used to obtain a consistent picture for the formation
of p–p-stacked assemblies, as convincingly demonstrated for
dye 2 in nonpolar organic solvent.
Figure 4. Van’t Hoff plot for the temperature dependence of the aggrega-
tion constant K of 2 in MCH.
440
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 436 – 449

Perylene Bisimide Dyes
FULL PAPER
Self-organization properties of dye 2 in condensed state:
Aggregation studies of dye 2 in solution (see above) have
revealed strong p–p interactions among the dyes, thus, it is
expected that this dye also possesses self-organizing proper-
ties in condensed state, that is, mesophase formation.
Indeed, differential scanning calorimetry (DSC), optical po-
larizing microscopy as well as X-ray diffraction studies with
the bulk samples of 2 unequivocally confirm liquid crystal-
linity (LC) of this dye. The DSC thermogram (Figure S3,
Supporting Information) of this compound shows a crystal-
LC phase transition with an endothermic peak at 858C with
ꢀ
1
DH = 71 Jg (melting of the alkyl chains) and a LC phase-
ꢀ
1
isotropic phase transition at 3008C with DH = 12 Jg . The
clearing point of 2 is about 708C lower than that observed
for the trialkoxyphenyl-substituted PBI 1. No subsequent re-
crystallization took place as no transitions corresponding to
the first endothermic peak were observed on cooling and
second heating, indicating that the recrystallization is kineti-
cally suppressed.
Figure 7. X-ray diffraction pattern for a sample of PBI 2 cooling from
1158C to RT with the Miller indices of the small angle diffraction peaks.
Under the polarizing optical microscope, a dendritic tex-
ture was observed for the LC phase of 2 (Figure 6). The X-
ray diffraction (XRD) measurements were performed for
the bulk sample of 2 after cooling from 1158C to room tem-
perature (Figure 7). The six observed Bragg reflections in
the small angle region can be indexed according to a two-di-
mensional hexagonal lattice with satisfactory accuracy
(
Table S2 in the Supporting Information). The cell parame-
ter of the hexagonal phase of 2 is 3.35 nm, which is nearly
identical with that reported for the structurally similar dye 1
[15a]
containing trialkoxyphenyl groups.
The diffused halo dis-
Figure 8. a) Molecular structure of PBI 2 with fully extended alky chains.
b) Top view of a columnar stacking model of dye 2 with a rotational dis-
placement of 608 of the molecular planes. c) Side view of the stack (for
clarity phenyl groups, instead of trialkylphenyl groups, are shown in the
molecules).
played at around 208 can be attributed to the disordered
alkyl chains. The absence of sharp peak in this region re-
veals a disordered stacking along the column axis. Owing to
the elongated nature of 2 and the fact that, generally, colum-
nar hexagonal phases possess a circular cross-section, a rota-
tional displacement of the individual dye units along the p–
p-stacking direction in LC phase can be implied, as the
model in Figure 8 shows. Molecular modeling study revealed
a length of about 4.9 nm for compound 2 with fully extend-
ed conformation of the alkyl chains. The observed diameter
of the column by X-ray analysis is 30% shorter. This can be
attributed to the alky chain tail folding and interdigitating
into neighboring columns, as pointed out by Percec et al. for
[35]
other systems.
In the lowest energy conformation of 2
that is assessed by molecular modeling, the two phenyl rings
at imide positions are not co-planar with the perylene bisi-
mide unit (Figure 8). Due to the bulkiness of the two phenyl
rings, a rotation displacement is demanded for the p–p
stacking of dye 2.
The unique columnar stacking feature of PBI 2 was fur-
ther verified by atomic force microscopy (AFM) studies. For
this purpose, thin films were prepared by spin-coating of so-
lution of PBI 2 on highly ordered pyrolytic graphite
(
HOPG). Tapping mode AFM images of the samples re-
vealed multilayer structures (Figure 9a) with the height of
.3 nm between the top and the adjacent layer. A high reso-
2
lution image (Figure 9b) showed that the top layer is com-
posed of long and bended columns which formed a finger-
print-like structure. Cross-section analysis of the lateral
structure and 2D Fourier transformation of the AFM image
revealed a mean value of 3.9ꢂ0.3 nm for the distance be-
tween neighboring columns which is in good agreement with
that observed by X-ray diffraction analysis for the liquid
Figure 6. Optical texture of dye 2 between two crossed polarizers ob-
tained by cooling the sample from isotropic phase.
Chem. Eur. J. 2007, 13, 436 – 449
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
441

F. Würthner et al.
crystal phase of 2. When the thin film is prepared from
CH Cl solution, in which dye 2 is less aggregated, structures
2
2
with much shorter rods were observed (Figure 9c), revealing
the solvent effect on the topology of the thin film. Based on
these results, a structure model of the aggregates is pro-
posed in Figure 9e. According to this model, the height dif-
ference between the layers (2.3 nm) should be smaller than
the diameter of the columns which is in agreement with our
observations. As in the case of liquid crystal phase, the col-
umns are segregated by the surrounding alkyl chains. The
length of the columns ranges between 20 nm to several hun-
dred nanometers which corresponds to columnar stacking of
60 to several thousand dye 2 molecules on graphite surface,
assuming a p–p distance of 3.5 Š as observed previously in
[36]
several crystal structures of N,N’-diphenyl PBI dyes.
Figure 10. a) Concentration-dependent fluorescence spectra normalized
at 518 nm (excited at the quasi-isosbestic point at 469 nm) of PBI 2 (2.1”
ꢀ
7
ꢀ4
1
0
m to 2.1” 10 m) in MCH solution and solution-cast thin film
(
dashed line) at RT. The arrows indicate the changes upon increasing
concentration. Inset: UV/Vis absorption spectrum of a spin-coated thin
film of 2 annealed at 1508C for 3 h. b) Temperature-dependent fluores-
ꢀ
6
Figure 9. AFM images of thin films spin-coated from a solution of 2 onto
cence spectra of 2 in n-hexane (1.0”10 m, excited at 469 nm) from ꢀ5
to 508C. The arrows indicate the changes upon decreasing temperature.
Inset: The integrated fluorescence intensity of the sample at different
temperatures.
ꢀ4
HOPG: a) Topography image (from MCH solution, 4”10 m). b) High
ꢀ
4
resolution topography image (from MCH solution, 4”10 m). c) High
ꢀ
4
2 2
resolution topography image (from CH Cl solution, 4”10 m). d) Cross-
section analysis along the line in image a), and e) proposed model for the
packing of columnar stacks on HOPG surface.
Optical properties of the aggregates of dye 2: Most of the
reported crystal structures of PBI dyes show a parallel pack-
ing of molecules along the long molecular axis with longitu-
The concentration-dependent fluorescence spectra of 2
(Figure 10a) display an emission band at 500–650 nm for the
monomeric dye 2 with well-resolved vibronic structure at
[5]
ꢀ6
dinal and transversal offsets between adjacent dyes. Only
in very few cases rotational displacements that lead to a
low concentrations (<10 m). Upon aggregation, a new
broad band emerges at 600–850 nm (aggregate band) and its
intensity increases strongly with increasing concentration.
The long tail of this band reached the near infrared (NIR)
[5,36]
screw-type stacking of dyes have been found in crystals,
as observed for the LC phase of the present dye 2. As a con-
sequence, novel optical and electrical properties might arise
for the aggregates of dye 2. In Figure 10a, the UV/Vis spec-
trum of a thin film of 2 (inset) and concentration-dependent
fluorescence spectra of 2 in MCH solution and of a thin film
as well are depicted. Remarkably, the absorption and the
fluorescence spectra of PBI 2 in thin film match perfectly
well with those obtained in high concentration in MCH
ꢀ
6
region already at a concentration of 3”10 m. Taking into
account that the monomer band is absent, the emission
spectrum of a spin-coated thin film is very similar to that ob-
served for concentrated solutions and only slightly shifted to
lower energy (Figure 10a, dotted line). To shed more light
into the fluorescence properties, temperature-dependent
fluorescence spectroscopy studies were performed in the sol-
[37]
(
Figure 3). This is very indicative for similar packing fea-
vent n-hexane.
For this solvent, the highest aggregation
tures, in particular, with respect to the p–p stacking of the
dye 2 in solution and in the condensed phase. Accordingly,
not only XRD (columnar LC phase) and AFM (columns de-
posit from solution), but also optical spectroscopy (local
contacts) provides evidence for identical or at least very
similar arrangements for the aggregated dyes in solution, on
surface and in the condensed state.
constant is given which enables the formation of aggregated
species even in very dilute solution. Accordingly, at low
temperatures the broad aggregate band at 600–850 nm can
be observed together with the monomer band (Figure 10b).
Upon increasing the temperature, this band is slightly hypso-
chromically shifted and the intensity gradually decreased. Si-
multaneously, the intensity of the monomer band increased
442
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 436 – 449

Perylene Bisimide Dyes
FULL PAPER
expectedly, indicating the transition of aggregates into mon-
omeric species.
Most of the observed absorption properties of dye 2 can
be rationalized by molecular exciton theory if rotational dis-
[38]
Remarkable color and intensity changes of the fluores-
cence were observed upon variation of concentration as well
as temperature of the solution of 2. In MCH, depending on
the ratio of aggregated versus total dye molecules aagg, the
color of the fluorescence changed from green (aagg <0.3, N
placement in dye stacking is assumed. According to this
theory, for n 1D p–p-stacked molecules each vibronic state
in S should split into a n-level band. Only the transitions to
1
the top level (H-aggregates) or bottom level (J-aggregates)
is allowed if all the transition dipoles are perfectly parallel
to each other. Thus, compared with nonaggregated mole-
cules, hypsochromic and bathochromic displacement of the
absorption band can be observed in the electronic spectra
for H- or J-aggregates, respectively. The hypsochromic shift
of lmax of 2 is indicative for the formation of cofacial (H-
type) aggregates. However, when rotational displacements
exist among the transition dipoles and the molecular vibra-
tion is strongly coupled with electronic transitions, as ob-
served for compound 2, the conventional selection rule does
not apply anymore and the forbidden electronic transitions
from the ground state to the lower excitonic state become
partially allowed. Thus, both bathochromically and hypso-
chromically shifted absorption bands (related to lmax of mo-
nomer) are observed for the dye 2 aggregates.
<
1.6), to yellow (aagg ꢁ0.7, N ꢁ2.3), orange (a ꢁ0.8, N
agg
ꢁ
3.0), red (aagg > 0.9, N > 4.0), and deep red (thin film)
(
Figure S4, Supporting Information). The fluorescence quan-
tum yield Ffl of 2 was determined as 0.63 in CH Cl and
2
2
0.65 in MCH for highly dilute solutions, where only mono-
mer band can be observed. Upon increasing the concentra-
tion, the fluorescence quantum yield decreased, although
the fluorescence intensity increased continuously due to the
higher concentration. By comparison of the observed fluo-
rescence intensity with the non-aggregated reference PBI 4
ꢀ4
(
F =1.0 in MCH), the F of 2 at 1”10 m in MCH is esti-
fl fl
mated as 0.4 (see ref. [55]), which is about 60% of that ob-
served for highly dilute solution (Figure S5, Supporting In-
formation). Likewise, the decrease of the fluorescence inten-
sity upon aggregation was also observed in the temperature-
dependent fluorescence spectra (Figure 10b, inset). In n-
hexane, at ꢀ58C the intensity of the fluorescence decreased
to about half of that at 508C where only monomeric dyes
are present. From these data, we can calculate a fluores-
cence quantum yield of 35% for the aggregates in n-hexane
which is still a quite high value for an emitter in the red to
near IR region.
Theoretical calculations for aggregate spectra of face-to-
face-stacked dimers of 2 with different rotation angles be-
[39]
tween the transition dipoles were performed based on es-
[40]
tablished models. Basically, a single vibrational monomer
mode is included and the dimer Hamiltonian is built from
the monomer Hamiltonians within a single exciton picture,
including a point-dipole coupling between the dimer excited
states. Details of the model and time-dependent quantum
[39]
To elucidate further the fluorescence properties of the
calculations are given elsewhere. From a comparison be-
tween experimental and theoretical dimer absorption spec-
tra, we have extracted a most likely dimer geometry, where
the transition dipole moments of the monomers span an
angle of b=558 and the relative orientation of the dipole
moments with respect to the monomer-monomer centre of
mass vector has an angle of a =a =708 (Figure 11). For
dye 2 aggregates, fluorescence lifetime t and anisotropy
f
were measured for MCH solutions (Figures S6 and S7, Sup-
ꢀ
5
ꢀ1
porting Information). For a 2.5”10
fluorescence decays monitored at the monomer band
520 nm) and the aggregate band (650 nm) provided tf
m
solution of 2, the
(
values of 3.2 ns and 32.7 ns, respectively. Thus, lifetime of
the aggregate band is 10 times higher than that of the mono-
mer band. Furthermore, fluorescence anisotropy was deter-
mined as 0.11 for the monomer band and 0.003 for the ag-
gregate band. According to the theory, an anisotropy value
as high as 0.4 is possible if no depolarization takes place nei-
ther by energy transfer processes to neighboring dyes nor by
rotational diffusion. Practically, for low viscosity solvents the
rotational diffusion of small fluorophores is fast and depola-
rization takes place. As mentioned, an anisotropy value of
1
2
this geometry, and an assumed monomer separation of
3.5 Š, the calculated absorption spectrum displayed in the
0.11 was observed for dye 2 monomer in MCH. Upon aggre-
gation, rotational diffusion should be slowed down due to
the higher molecular mass of the dye aggregates. Thus, the
anisotropy should increase if no other depolarization pro-
cess takes place within the dye stacks. However, a nearly
zero anisotropy value (0.003) was observed for the aggre-
gate fluorescence band of dye 2. This depolarization of the
aggregate band can be explained in terms of efficient energy
transfer among the stacked dyes. Due to their rotational dis-
placement their transition dipoles are not parallel and the
fluorescence polarization vanishes.
Figure 11. Dimeric stacking of 2 that provides the best agreement be-
tween experimental UV/Vis absorption spectra and calculated ones (see
Figure 12). For clarity, phenyl groups are shown instead of the trialkyl-
phenyl groups. The geometry parameters are a
1 2
=a =708, b=558 and a
center-to-center distance of 3.5 Š between the molecules. See ref. [39].
Chem. Eur. J. 2007, 13, 436 – 449
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
443

F. Würthner et al.
Figure 13. Schematic representation of the excimer formation in the ex-
tended p–p stacks of dye 2.
[40]
charge transfer states, the p–p distance is reduced in this
relaxed structure in comparison with that of the ground
state p–p stacks. Like the conventional excimers, emission
from such excited state dimer will lead to a high energy vi-
brational level on the ground state potential surface, thus, it
has some dissociative character according to the definition
[41]
of Birks.
However, despite similarities, there are also differences
compared with conventional excimers like those of pyrene
[
41d]
in solution.
excimer formation is only feasible at higher concentration
First, while the diffusion-controlled collision
[42]
or for long-lived excited state species like pyrene,
the
ꢀ
4
probability for the relaxation into a excimer-type emissive
state is highly increased by the extended p–p stacking of
dye 2 that only requires small displacements between the
molecules to give an excited state dimeric unit (Figure 13).
Second, the excimer-type band of dye 2 is not totally struc-
tureless (Figure 10) in contrast with the unstructured pyrene
excimer emission that has been ascribed to the rapid dissoci-
ation of the dimer at a repulsive ground state before the
Figure 12. UV/Vis spectrum of 2 measured at 2.2”10 m in MCH (aagg
.96, top) and the calculated dimer spectrum (bottom).
=
0
lower panel of Figure 12 is obtained. The spectrum is deter-
mined for the case of a low resolution, taking the influence
of the surroundings into account. Considering the simplicity
of the employed model, the agreement with the experimen-
tal spectrum (upper panel) is excellent. Noteworthy, that the
extension of the model to a trimer with the same geometric
parameters yields absorption spectra similar to that of the
dimer, thus, it can be anticipated that the dimer model al-
ready provides valuable information about the geometry of
the extended oligomeric aggregate.
For cofacial p–p stacks with rotational displacements be-
tween the chromophores, photoluminescence can be ob-
served; however, the exciton theory cannot properly de-
scribe the emission properties of the aggregates of 2. The
calculated emission spectrum remarkably differs from the
measured one, pinpointing a relaxation process into an emis-
sive state that is different from the Franck–Condon state.
Indeed, the observed spectra exhibit features that are simi-
[42]
molecules can complete a vibration cycle. In the case of
dye 2, dissociation of the molecules in the ground state is re-
placed by a structural reorganization into a attractive p–p
stacked state. As a result, the vibronic structure can still be
observed in contrast to the pyrene excimer band. Indeed,
the fluorescence band of concentrated solutions and thin
films of 2 can be resolved into several overlapped Gaussian
ꢀ
1
bands with energy interval of 1100–1300 cm (Figure S8 in
the Supporting Information). Third, the stabilization of the
excited dimer unit of strongly quadrupolar perylene bisi-
mide dye 2 might be different from those of pure polycyclic
aromatic hydrocarbons like pyrene or perylene. For pery-
lene excimers, it has been reported that exciton interaction
and charge transfer interaction have 70 and 30% contribu-
[39]
[41]
[43]
lar to those typically observed for excimers,
that is, a
tion, respectively, to the total stabilization energy. For ex-
structureless broad band with pronounced bathochromic
shift and much longer lifetime compared to the monomer
species.
Unlike typical excimers where p–p aggregates are formed
only in the excited states, while the ground state interactions
between the p-systems are repulsive (dissociative), extended
cited aggregates of PBI dye 2, charge transfer interactions
could have a significantly higher contribution due to the
presence of two electron acceptor imide groups that embed
the electron-rich central perylene unit. Indeed, according to
our calculations (Figure 11) already in the ground state the
symmetry is broken with one imide acceptor group posi-
tioned more on top of the central perylene core than for the
other. Thus, it is very likely that the excited aggregates of 2
1D p stacks are already given for 2 in the ground state, as
shown by NMR, UV/Vis and AFM studies. The formation
of excimer-type excited state dimer is, however, feasible. As
suggested in Figure 13, upon photoexcitation, pronounced
structural and energetic reorganization processes may take
place between the excited molecule and the neighbor one to
form a relaxed state (MM*) from which emission takes
place. Owing to the resonance contributions of exciton and
[44]
emanate some properties of charge-transfer excitons.
As a consequence of such severe structural reorganiza-
tions in the excited state, energy transfer process, such as ex-
[
45]
citon hopping
slowed down significantly within the extended p–p stacks of
to nearest neighbor molecules should be
[46]
these dyes. Excited dimer species are expected to behave
444
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 436 – 449

Perylene Bisimide Dyes
FULL PAPER
as traps, which should reduce the exciton mobility compared
with ordered one-dimensional stacks and limit the perfor-
mance of liquid crystalline materials compared to their crys-
talline counterparts (in organic crystals, the excimers have
Table 2. Charge carrier mobilities of PBI 2 in crystalline and LC phases.
[
a]
[b]
TRMC,iso
Phase
ꢀm
[
min,iso
ꢀ1 ꢀ1
ꢀm
[cm V
min,1D
ꢀm
[cm V
ꢀm
TRMC,1D
2
2
ꢀ1 ꢀ1
2
ꢀ1 ꢀ1
2
ꢀ1 ꢀ1
cm V
s
]
s
]
s
]
[cm V s ]
crystal
0.047
0.14
0.14
0.42
[47]
LC phase
0.0026
0.0078
0.0076
0.023
been recognized as a sort of self-trapped excitons ).
[
a] ꢀmmin,iso =Ds/Deop·E
p
·
A
H
R
U
G
·
A
C
H
T
R
E
U
N
G
(1/f)·-
(1/Weop). Ds/Deop =charge in conductivity at the end of the pulse, E
energy needed to create one electron hole pair, f=fill factor=mass of
A
H
R
U
G
p
=
Charge transport properties of PBI 2: The semiconductive
properties of present PBI 2 were explored by using the
pulse-radiolysis time-resolved microwave conductivity (PR-
s
compound/volume holder, Weop =W ·fraction of charges created in side
chains.
[48]
TRMC) technique. This technique probes the (change in)
conductivity of solid samples upon the generation of charge
carriers by irradiation of the sample with high energy elec-
tron pulses. In this way, the motion of charge carriers can be
probed within solid samples of organic materials without
employing electrodes. Since both positive and negative
charges are generated during the high energy pulse, it is not
possible to distinguish between the contribution of positive
and negative charges to the observed conductivity. The con-
ductivity of 2 in a temperature range from room tempera-
ture to 380 K is shown in Figure 14. The conductivity was
found to decrease by more than an order of magnitude
upon the transition from the crystalline phase to the LC
more, not all generated charges contribute to the conductivi-
ty signal and the actual mobility value can be obtained by
considering the charge carrier decay during the high-energy
electron pulse. Taking these effects into account, the one-di-
mensional TRMC mobility calculated for PBI 2 in the crys-
2
ꢀ1 ꢀ1
talline phase is 0.42 cm V s . This mobility value is among
the best values for molecular semiconductors and compares
[7]
well with that of another PBI dye reported previously.
Marder and co-workers have recently reported on the
charge carrier mobility of the tridodecyloxyphenyl-substitut-
ed PBI 1 in the LC phase measured by SCLC and TRMC
[48b]
[15d]
phase, as observed for other LC organic materials.
techniques.
With SCLC method a mobility value of
2
ꢀ1 ꢀ1
0
.2 cm V
s
was obtained for the LC phase of PBI 1,
2
ꢀ1 ꢀ1
whereas by TRMC method a mobility of 0.0078 cm V
s
was assessed. The latter value is comparable with the lower
2
ꢀ1 ꢀ1
limit of the mobility (0.0026 cm V s ) that we have ob-
served for PBI 2 in the LC phase by TRMC. Obviously, the
TRMC mobility cannot be directly compared with the mobi-
lity determined by SCLC as significantly different mobility
values were found for the same system (LC phase of PBI 1)
[15d]
with these two different techniques.
On transformation of the crystalline to the LC phase the
structural disorder in the material increases which leads to a
reduction in the mobility. At this point we like to speculate
that the lowering of the charge carrier mobility in LC phase
of PBI 2 is at least partly due to the formation of dimeric
traps that are of similar nature as the exciton traps discussed
before. The formation of p dimers from radical cationic and
anionic species has been amply discussed in the literature,
respectively, for oligothiophene and naphthalene bisimide
semiconductors in solution and amorphous solid state, but
to our knowledge this discussion has never been extended
to liquid crystalline semiconductors, despite their obvious p–
p-stacked nature which requires only minor structural reor-
ganization to provide a localized more stable dimeric spe-
cies. Because such dimeric species (of polaronic or bipolar-
onic nature) constitute traps, they offer a reasonable explan-
ation for the lower charge carrier mobilities of LC com-
pared to crystalline phases. Thus, the polaronic hopping mo-
bility can be described by the Equation (3)
Figure 14. Temperature dependence of the dose-normalized conductivity
of PBI 2 for increasing temperature (
) and decreasing temperature (*).
&
In accordance with our DSC study, no increase of the con-
ductivity could be observed upon cooling because the LC
phase prevails. An estimate for the lower limit of the mobili-
ty of charge carriers can be made by assuming that all the
charges generated during the high energy pulse contribute
to the conductivity signal. For this situation, and assuming
charge carriers transport in three dimensions, a lower limit
[49]
2
ꢀ1 ꢀ1
of the mobility of 0.047 and 0.0026 cm V
s
is calculated
for the crystalline phase and the LC phase of 2, respectively.
However, our packing model derived from X-ray analysis
suggests that the pathways for charge carrier motion along
the stacks of the PBI 2 dye is one-dimensional. As a conse-
quence, the mobility for 2 is not isotropic and the mobility
value of interest is the mobility along the stacking direction
of the dye. Thus, the lower limits of the one-dimensional
ꢀ
ꢁ
A
l
m ¼ T exp ꢀ
ð3Þ
3
=
2
4
kT
2
ꢀ1 ꢀ1
[50]
mobility are 0.14 and 0.0078 cm V
s
for the crystalline
where A is a constant and l the reorganization energy. As
the lowest energy absorption of the aggregates of PBI 2
phase and the LC phase of 2, respectively (Table 2). Further-
Chem. Eur. J. 2007, 13, 436 – 449
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
445

F. Würthner et al.
occurs at around 540 nm (Figure 3) and the excimer fluores-
cence is observed at about 680 nm (Figure 10), a spectral
shift of about 0.5 eV is given which can be taken as an ap-
proximation for the reorganization energy of a charge carri-
er. For a reorganization energy of 0.5 eV, the polaronic hop-
ping mobility increases approximately by a factor of 2 upon
increasing the temperature from 280 to 380 K. This tempera-
ture dependence is only slightly stronger than for the ob-
served mobilities in the LC phase of PBI 2 (Figure 14, cir-
cles). Taking into consideration the simplicity of the polar-
onic hopping model for a description of charge transport,
the agreement between the theoretical and measured tem-
perature dependence is fairly good. At higher temperatures
enhanced dynamic structural fluctuations may have a nega-
tive effect on the mobility, which partially compensates the
thermally activated behavior, as expected on the basis of po-
laronic hopping motion only.
Wide angle X-ray diffractograms were obtained at room temperature on
a Siemens powder diffractometers (CuKa radiation). The vapor pressure
osmometry measurements were performed on a KNAUER osmometer
with a universal temperature measurement unit. Benzil was used as stan-
dard and a calibration curve for each temperature (40, 50 and 608C) in
terms of R in Ohm versus molal osmotic concentration (moles per kg
MCH) was constructed up to 0.01 molal.
1
DOSY NMR spectroscopy: H DOSY measurements were performed at
2
98 K on a Bruker DMX 600 (600 MHz) NMR spectrometer equipped
with a BGPA 10 gradient generator, a BGU II control unit and a conven-
15
31
1
tional 5 mm broad band ( N, P)/ H probe with z axis gradient coil capa-
ble of producing pulsed magnetic field gradients in the z direction of
ꢀ
1
52 Gcm . The NMR spectra were recorded in [D14]MCH solutions in
mm Norell NMR tubes. The spectral data were acquired using the lon-
gitudinal eddy current delay sequence with bipolar gradient pulse pairs
5
[
51]
for diffusion (BPP-LED) and additional sinusoidal spoil gradients after
the second and fourth 908 pulses were used. The temperatures were cali-
brated with a probe of 4% MeOH in CD OD. The fluctuation of the
3
temperature was less than 0.1 K during the measurements. The strength
1
of the pulsed magnetic field gradients was calibrated by H DOSY ex-
periments with a sample of 1% H
2
O in 99% D
2
O, doped with GdCl
3
(
0.1 mg per mL) to achieve short spin-lattice relaxation times, using the
known value of the diffusion coefficient for H
2
O at 298 K in this H
2
O/
Conclusion
2
D O mixture. The diffusion coefficient D was obtained by curve fitting
2
2
2
ACHTREUNG
according to the equation I=I
0
exp[ꢀDg g d (Dꢀd/3)], where I is the ob-
served intensity for the gradient strength g and g is the gyromagnetic
ratio of the observed nuclei, D represents the diffusion time reduced by
half of the delay (t=300 ms) between the two gradients of a bipolar gra-
dients pulse pair and d is the duration of the short bipolar gradient
pulses. For each spectrum, the average value of D obtained from differ-
ent aromatic signals (at least 3, depending on the numbers of the aromat-
ic signals) was used.
A new, highly soluble PBI dye 2 formed extended fluores-
cent aggregates in nonpolar solvent. The aggregation behav-
ior is concentration and temperature-dependent and can be
described very well with the isodesmic model. Furthermore,
dye 2 possesses liquid crystalline properties. Thus, in con-
densed state a hexagonal columnar LC phase was observed
and well-defined columnar nanostructures could be obtained
by simple solution casting. X-ray diffraction and AFM stud-
ies indicated that the columnar stacking is accomplished by
rotational displacement among the molecules which is addi-
tionally substantiated by theoretical calculations. Such
unique p–p stacking of 2 leads to rather interesting and re-
markable absorption and emission properties of the dye ag-
gregates indicating strong electronic coupling and charge
transfer interaction among the self-assembled dyes. PR-
TRMC conductivity measurements revealed that dye 2 is
among the best molecular semiconductors in the crystalline
columnar phase. Consequently, the fluorescent nanowires of
this dye are ideal candidates for further investigations as n-
type semiconducting materials and self-assembled nanoemit-
ters.
Fluorescence measurements: The steady state fluorescence spectra and
the fluorescence anisotropy were measured under ambient conditions on
a PTI QM4/2003 spectrofluorometer equipped with two Glan-Thomson
polarizers under magic angle and front-face setup due to the high optical
density of the samples. All the fluorescence spectra were corrected. The
fluorescence quantum yields were determined by the optical dilution
method using fluorescein (Ffl =0.92 in 1n aqueous NaOH) as stan-
[
52]
dard. The given quantum yields are averaged value of data obtained at
three different excitation wavelengths. The fluorescence anisotropy r is
defined as r = (IVVꢀGIVH)/(IVV + 2GIVH), where G=IHV/IHH and I is the
fluorescence intensity at a specific wavelength and the indices are related
to the vertical or horizontal orientation of the excitation (first index) and
the emission polarizer (second index) with respect to the excitation–emis-
sion plane. G is an instrument factor which compensates for polarization
effects of the emission optics. Correction for the G factor was obtained
automatically. Fluorescence intensities were averaged for 60 s at 2 points
per s, bandpass was set to 6 nm, and the measurements were performed
at room temperature.
Fluorescence lifetimes were measured under ambient conditions accord-
[53]
ing to the method described in the literature. By using a PTI Laser-
Strobe fluorescence lifetime spectrometer system containing a GL-3300
nitrogen laser (337.1 nm, pulse width 600 ps, pulse energy 1.45 mJ) cou-
pled with a dye laser GL-302 (pulse width 500 ps, pulse energy 220 mJ at
Experimental Section
5
50 nm) as an excitation source and a stroboscopic detector. Laser
General methods: NMR spectra were recorded at 300 K on Bruker
Avance 400 (400 MHz) or DMX 600 (600 MHz, for temperature depen-
dent measurements) spectrometers and the spectra were calibrated
against TMS (d=0.0 ppm). For the temperature-dependent measure-
ments the spectra were calibrated against the chemical shift of residual
methylcyclohexane (MCH) in [D14]MCH at d=1.60 ppm. The solvents
for spectroscopic studies were of spectroscopic grade and used as re-
ceived. UV/Vis spectra were measured on Perkin–Elmer Lambda 40P
spectrometer equipped with a Peltier system as temperature controller.
Differential scanning calorimetry (DSC) measurements were performed
by using a TA Q1000 calorimeter. Optical textures at crossed polarizers
were obtained with an Olympus BX-41 polarization microscope equipped
with a Linkam THMS 600 hot stage and a temperature controller unit.
output was tuned within the emission curves of the laser dyes supplied by
the manufacturer (PLD 421, 500, 579, 665, 735). The time resolution fol-
lowing deconvolution of experimental decays was 200 ps. The instrument
response function was collected by scattering the exciting light of a
dilute, aqueous suspension of colloidal silica (Ludox). Decay curves were
evaluated using the software supplied with the instrument applying least
square regression analysis. The quality of the fit was evaluated by analy-
2
sis of c (0.9–1.1), DW factor (> 1.75) and Z value (< ꢀ1.96, confidence
level 0.95) as well as by inspection of residuals and autocorrelation func-
tion.
Atomic force microscopy (AFM): AFM measurements were performed
under ambient conditions using a MultiMode Nanoscope IV system oper-
446
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 436 – 449

Perylene Bisimide Dyes
FULL PAPER
ating in tapping mode in air. Silicon cantilevers (OMCL-AC160TS) with
a resonance frequency of ꢁ300 kHz were used. Solution of perylene bisi-
mide 2 in MCH was spin-coated onto a HOPG (highly ordered pyrolytic
graphite) under 7000 rpm.
1.2 (m, 120H, CH
trobenzene): m/z: calcd for C108
emental analysis (%) calcd for C108
1.80; found: C 83.13, H 10.75, N 1.89; UV/Vis (CH
96300), 491 (58100), 460 (21000), 434 (6100), 369 (5000 mol Lcm );
2
), 0.88 (m, 18H, CH ); MS (FAB, matrix: p-octyloxyni-
3
+
H
162
N
2
O
4
: 1551.3; found: 1551.2 [M] ; el-
(1552.5): C 83.56, H 10.52, N
Cl ): lmax(e)=527
162 2 4
H N O
2
2
ꢀ
1
ꢀ1
(
Charge carrier mobility measurements: The conductive properties were
studied with the pulse-radiolysis time-resolved microwave conductivity
fluorescence (CH
2
Cl
2
): lmax =532 nm; quantum yield: Ffl =0.63.
(
PR-TRMC) technique. Pressed pellets (25 mg, fill factor 0.75) of the ma-
terial were irradiated with 10 nanosecond pulses of 3 MeV electrons from
a Van de Graaff accelerator, which results in the creation of a uniform
micromolar concentration of electron-hole pairs. The resulting change in
conductivity was monitored as the microwave power absorbed by the
mobile charge carriers in the material at an electro-magnetic field fre-
quency of 33 GHz. The mobility values were determined from the con-
ductivity at the end of the electron pulse. The lower limit to the mobility
was obtained by assuming that all generated charges contribute to the
conductivity (survival parameter Weop =1) and a pair formation energy of
Acknowledgements
We are grateful to the Deutsche Forschungsgemeinschaft for financial
support for this work within the research graduate school GK 1221 “Con-
trol of electronic properties of aggregates of p-conjugated molecules” at
the University of Würzburg. We are indebted to Dr. Matthias Grüne and
Elfriede Ruckdeschel for the DOSY NMR measurements, Professor
R. B. Neder for XRD measurements, and their helpful discussions.
1
6 eV. This pair formation energy was obtained using the electron density
fractions (f) and the pair formation energies (E ) of the side chains (f=
.79, E 25 eV) and the core of the dye (f=0.21, E =6.72 taken from the
onset of optical absorption) as described elsewhere.
p
0
p
p
[
48]
The one-dimen-
[1] a) F. J. M. Hoeben, P. Jonkheijm, E. W. Meijer, A. P. H. J. Schenning,
Chem. Rev. 2005, 105, 1491–1546; b) Supramolecular Dye Chemis-
try, Vol. 258 (Ed.: F. Würthner), Topics in Current Chemistry,
Springer, Berlin, 2005.
[2] a) T. Pullerits, V. Sundstrom, Acc. Chem. Res. 1996, 29, 381–389;
b) T. S. Balaban, H. Tamiaki, A. R. Holzwarth, Top. Curr. Chem.
2005, 258, 1–38; c) X. Hu, T. Ritz, A. Damjanovic, F. Autenrieth, K.
Schulten, Q. Rev. Biophys. 2002, 35, 1–62.
sional TRMC mobility was calculated assuming that electron-hole pairs
generated in the core of the dye recombine and that only a fraction of
charges generated in the side chains (W
s
) escapes recombination. For
[
54]
compound 2 an escape fraction W of 0.72 can be determined. Using
s
the electron density fractions mentioned above, the fraction of the charg-
es that contributes to the conductivity signal Weop was 0.54. The mobility
data for PBI 2 are collected in Table 2.
[
3] a) W. Pisula, M. Kastler, D. Wasserfallen, J. W. F. Robertson, F.
Nolde, C. Kohl, K. Müllen, Angew. Chem. 2006, 118, 834–838;
Angew. Chem. Int. Ed. 2006, 45, 819–823; b) J. P. Hill, W. Jin, A.
Kosaka, T. Fukushima, H. Ichihara, T. Shimomura, K. Ito, T. Hashi-
zume, N. Ishii, T. Aida, Science 2004, 304, 1481–1483.
4] a) T. van der Boom, R. T. Hayes, Y. Zhao, P. Bushard, E. A. Weiss,
M. R. Wasielewski, J. Am. Chem. Soc. 2002, 124, 9582–9590; b) F.
Würthner, Z. Chen, F. J. M. Hoeben, P. Osswald, C.-C. You, P. Jonk-
heijm, J. v. Herrikhuyzen, A. P. H. J. Schenning, P. P. A. M. van der
Schoot, E. W. Meijer, E. H. A. Beckers, S. C. J. Meskers, R. A. J.
Janssen, J. Am. Chem. Soc. 2004, 126, 10611–10618; c) M. G.
Debije, Z. Chen, J. Piris, R. B. Neder, M. M. Watson, K. Müllen, F.
Würthner, J. Mater. Chem. 2005, 15, 1270–1276.
3
,4,5-Tris(1’-dodecynyl)-nitrobenzene
(5):
3,4,5-Triiodonitrobenzene
(
(
3.0 g, 6.0 mmol), bis(triphenylphosphine)-palladium(ii) dichloride
252 mg, 0.30 mmol), copper(i) iodide (115 mg, 0.60 mmol) were mixed in
triethylamine (80 mL) under argon atmosphere and stirred at 808C. Then
-dodecynene (3.30 g, 19.8 mmol) was added dropwise and the reaction
1
[
mixture was stirred for another 4 h. After cooling to room temperature,
the solid precipitate was removed by filtration and the solution was con-
centrated by rotary evaporation. The residue was purified by silica gel
column chromatography with n-hexane/acetone 10:1 to obtain a brown
1
oil (3.40 g, 93%). H NMR (400 MHz, CDCl
3
, 300 K, TMS): d=8.08 (s,
), 2.46 (t, 4H, J=7.0 Hz, Cꢃ
), 1.49 (m, 6H, CH ), 1.4–1.1 (m, 36H, CH ),
); MS (EI, 70 eV): m/z (%): 615.6 (100) [M] ; elemen-
tal analysis (%) calcd for C42 (616.0): C 81.90, H 10.64, N 2.27;
found: C 81.50, H 11.05, N 2.36.
2
H, Ar-H), 2.54 (t, 2H, J=7.0 Hz, CꢃCCH
CCH ), 1.65 (m, 6H, CH
.88 (m, 9H, CH
2
2
2
2
2
+
0
3
[
[
[
5] F. Würthner, Chem. Commun. 2004, 1564–1579 and references
therein.
6] H. Langhals, J. Karolin, L. B. Johansson, J. Chem. Soc. Faraday
Trans. 1998, 94, 2919–2922.
7] C. W. Struijk, A. B. Sieval, J. E. J. Dakhorst, M. Dijk, P. Kimkes,
R. B. M. Koehorst, H. Donker, T. J. Schaafsma, S. J. Picken, A. M.
van de Craats, J. M. Warman, H. Zuilhof, E. J. R. Sudhçlter, J. Am.
Chem. Soc. 2000, 122, 11057–11066.
H
65NO
2
3
2
,4,5-Tridodecylaniline (6): 3,4,5-Tri(1’-dodecynyl)-nitrobenzene (1.60 g,
.60 mmol) was placed in a flask containing 10% Pd/C (0.4 g) and dry
ethanol (80 mL) and dry ethyl acetate (20 mL). The flask was affixed
with balloons filled with hydrogen gas and the mixture was stirred at
room temperature for 24 h. Then the Pd/C in the reaction mixture was
removed by filtration and the solvent was removed by rotary evapora-
tion. The resulting brown oil was purified by silica gel column chroma-
[
8] a) F. C. DeSchryver, T. Vosch, M. Cotlet, M. van der Auweraer, K.
Müllen, J. Hofkens, Acc. Chem. Res. 2005, 38, 514–522; b) A. Saut-
ter, B. K. Kaletas, D. G. Schmid, R. Dobrawa, M. Zimine, G. Jung,
I. H. M. Van Stokkum, L. De Cola, R. M. Williams, F. Würthner, J.
Am. Chem. Soc. 2005, 127, 6719–6729; c) T. Ishi-i, K. Murakami, Y.
Imai, S. Mataka, Org. Lett. 2005, 7, 3175–3178.
tography with n-hexane/CH
became a white solid when stored in refrigerator (0.96 g, 53%). H NMR
400 MHz, CDCl , 300 K, TMS): d=6.41 (s, 2H, Ar-H), 2.48 (m, 6H, Ar-
CH ), 1.7–1.1 (m, 60H, CH ), 0.88 (m, 9H, CH ), a very broad signal was
observed for NH protons at 4.0–3.0 ppm; MS (EI, 70 eV): m/z (%):
98.1 (100) [M] ; elemental analysis (%) calcd for C42 79N (598.1): C
4.34, H 13.31, N 2.34; found: C 84.37, H 13.32, N 2.32.
2 2
Cl 1:1 to give slightly yellow oil which
1
(
3
2
2
3
2
[
9] a) M. P. OꢁNeil, M. P. Niemczyk, W. A. Svec, D. Gosztola, G. L.
Gaines III, M. R. Wasielewski, Science 1992, 257, 63–65; b) E. H. A.
Beckers, S. C. J. Meskers, A. P. H. J. Schenning, Z. Chen, F. Würth-
ner, R. A. J. Janssen, J. Phys. Chem. A 2004, 108, 6933–6937; c) A.
Prodi, C. Chiorboli, F. Scandola, E. Iengo, E. Alessio, R. Dobrawa,
F. Würthner, J. Am. Chem. Soc. 2005, 127, 1454–1462.
+
5
8
H
N,N’-Di(3,4,5-tridodecylphenyl)-perylene-3,4:9,10-tetracarboxylic acid bi-
simide (2): A mixture of perylene-3,4:9,10-tetracarboxylic acid bisanhy-
dride (0.13 g, 0.33 mmol), aniline 6 (0.40 g, 0.67 mmol) and zinc acetate
(
0.073 g, 0.33 mmol) were mixed in quinoline (15 mL). The reaction mix-
[10] a) A. Kraft, A. C. Grimsdale, A. B. Holmes, Angew. Chem. 1998,
110, 416–443; Angew. Chem. Int. Ed. 1998, 37, 402–428; b) C. Kara-
pire, C. Zafer, S. Icli, Synth. Met. 2004, 145, 51–60; c) J. Pan, W.
Zhu, S. Li, W. Zeng, Y. Cao, H. Tian, Polymer 2005, 46, 7658–7669.
[11] a) P. R. L. Malenfant, C. D. Dimitrakopoulos, J. D. Gelorme, L. L.
Kosar, T. O. Graham, Appl. Phys. Lett. 2002, 2517–2519; b) C. D.
Dimitrakopoulos, P. R. L. Malenfant, Adv. Mater. 2002, 14, 99–117;
c) B. A. Jones, M. J. Ahrens, M.-H. Yoon, A. Facchetti, T. J. Marks,
M. R. Wasielewski, Angew. Chem. 2004, 116, 6523–6526; Angew.
ture was stirred at 1808C for 2 h. After cooling to room temperature, the
mixture was poured into MeOH (30 mL). The precipitate was collected
by filtration, washed with methanol (3”20 mL), and then dried in
vacuum. The crude product was further purified by silica gel column
chromatography (CH
methanol 1:1 (10 mL) to give a red powder (280 mg, 56%). H NMR
2 2 2 2
Cl ) and then slowly precipitated from CH Cl /
1
(
(
400 MHz, CDCl
d, 4H, J=8.2 Hz, Hpery), 6.97 (s, 4H, Ar-H
3
, 300 K, TMS): d=8.74 (d, 4H, J=8.0 Hz, Hpery), 8.66
, 2.66 (m, 12H, Ar-CH ), 1.8–
)
2
Chem. Eur. J. 2007, 13, 436 – 449
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
447

F. Würthner et al.
Chem. Int. Ed. 2004, 116, 6363–6366; d) for a recent review, see: F.
Würthner, R. Schmidt, ChemPhysChem 2006, 7, 793–797.
[31] B. C. Burdett, Studies in Physical and Theoretical Chemistry, 1983,
26 (Aggregation Processes in Solution), 241–270.
[
12] a) C. W. Tang, Appl. Phys. Lett. 1986, 48, 183–185; b) L. Schmidt-
Mende, A. Fechtenkçtter, K. Müllen, E. Moons, R. H. Friend, J. D.
MacKenzie, Science 2001, 293, 1119–1122; c) A. J. Breeze, A. Salo-
mon, D. S. Ginley, B. A. Gregg, H. Tillmann, H.-H. Hçrhold, Appl.
Phys. Lett. 2002, 81, 3085–3087.
[32] M. Sadrai, L. Hadel, R. R. Sauers, S. Husain, K. Krogh-Jespersen,
J. D. Westbrook, G. R. Bird, J. Phys. Chem. 1992, 96, 7988–7996.
[33] a) R. B. Martin, Chem. Rev. 1996, 96, 3043–3064; b) N. J. Baxter,
M. P. Williamson, T. H. Lilley, E. Haslan, J. Chem. Soc. Faraday
Trans. 2 1996, 92, 231–234; c) Supramolecular polymers (Ed.: A. Ci-
ferri), CRC Press, Boca Raton, 2005; d) P. van der Schoot, M. A. J.
Michels, L. Brunsveld, R. P. Sijbesma, A. Ramzi, Langmuir 2000, 16,
10076–10083.
[34] A. P. H. J. Schenning, P. Jonkheijm, E. Peeters, E. W. Meijer, J. Am.
Chem. Soc. 2001, 123, 409–416.
[35] V. Percec, C.-H. Ahn, T. K. Bera, G. Ungar, D. J. P. Yeardley, Chem.
Eur. J. 1999, 5, 1070–1083.
[
13] A. Tracz, J. K. Jeszka, M. D. Watson, W. Pisula, K. Müllen, T.
Pakula, J. Am. Chem. Soc. 2003, 125, 1682–1683.
[
14] P. Jonkheijm, F. J. M. Hoeben, R. Kleppinger, J. van Herrikhuyzen,
A. P. H. J. Schenning, E. W. Meijer, J. Am. Chem. Soc. 2003, 125,
1
5941–15949.
15] a) F. Würthner, C. Thalacker, S. Diele, C. Tschierske, Chem. Eur. J.
001, 7, 2245–2253; b) I. K. Iverson, S. M. Casey, W. Seo, S.-W.
[
2
Tam-Chang, B. A. Pindzola, Langmuir, 2002, 18, 3510–3516; c) J.
van Herrikhuyzen, A. Syamakumari, A. P. H. J. Schenning, E. W.
Meijer, J. Am. Chem. Soc. 2004, 126, 10021–10027; d) Z. An, J. Yu,
S. C. Jones, S. Barlow, S. Yoo, B. Domercq, P. Prins, L. D. A. Sieb-
beles, B. Kippelen, S. R. Marder, Adv. Mater. 2005, 17, 2580–2583.
16] J. M. Warman, A. M. Van De Craats, Mol. Cryst. Liq. Cryst. 2003,
[36] a) F. Graser, E. Hädicke, Liebigs Ann. Chem. 1980, 1994–2011;
b) G. Klebe, F. Graser, E. Hädicke, J. Berndt, Acta Crystallogr. Sect.
B 1989, 45, 69–77; c) P. Zugenmaier, J. Duff, T. L. Bluhm, Cryst.
Res. Technol. 2000, 35, 1095–1115.
[37] In n-hexane, the aggregation constant of dye 2 is 1.2”10 m at RT,
which is about one order of magnitude higher than that in MCH.
[38] a) M. Kasha, H. R. Rawls, M. A. El-Bayoumi, Pure Appl. Chem.
1965, 11, 371–392; b) V. Czikkely, H. D. Fçrsterling, H. Kuhn,
Chem. Phys. Lett. 1970, 6, 207–210; c) P. W. Bohn, Annu. Rev. Phys.
Chem. 1993, 44, 37–60; d) J-Aggregates (Ed.: T. Kobayashi), World
Scientific, Singapore, 1996.
6
ꢀ1
[
[
[
3
96, 41–72.
17] A. Bayer, J. Hübner, J. Kopitzke, M. Oestreich, W. Rühe, J. H. Wen-
dorff, J. Phys. Chem. A 2001, 105, 4596–4602.
18] a) C. Tschierske, J. Mater. Chem. 1998, 8, 1485–1508; b) C. D. Simp-
son, J. Wu, M. D. Watson, K. Müllen, J. Mater. Chem. 2004, 14, 494–
5
04.
[39] J. Seibt, P. Marquetand, V. Engel, Z. Chen, V. Dehm, F. Würthner,
Chem. Phys. 2006, 328, 354–362.
[40] a) A. Witkowski, W. Moffit, J. Chem. Phys. 1960, 33 872–875;
b) R. L. Fulton, M. Goutermann, J. Chem. Phys. 1964, 41, 2280–
2286.
[41] a) J. B. Birks, Rep. Prog. Phys. 1975, 38, 903–974; b) S. A. Jenekhe,
J. A. Osaheni, Science, 1994, 265, 765–768; c) B. J. Schwartz, Annu.
Rev. Phys. Chem. 2003, 54, 141–172; d) A. Hayer, V. de Halleux, A.
Kçhler, A. El-Garoughy, E. W. Meijer, J. Barberꢂ, J. Tant, J. Levin,
M. Lehmann, J. Gierschner, J. Cornil, Y. H. Geerts, J. Phys. Chem.
B 2006, 110, 7653–7659.
[
[
[
[
19] M. O. Vysotsky, V. Bçhmer, F. Würthner, C.-C. You, K. Rissanen,
Org. Lett. 2002, 4, 2901–2904.
20] E. H. A. Beckers, S. C. J. Meskers, A. P. H. J. Schenning, Z. Chen, F.
Würthner, R. A. J. Janssen, J. Phys. Chem. A 2004, 108, 6933–6937.
21] F. Würthner, Z. Chen, V. Dehm, V. Stepanenko, Chem. Commun.
2
006, 1188–1190.
22] a) R. Singh, G. Just, J. Org. Chem. 1989, 54, 4453–4457; b) H.-C.
Zhang, W.-S. Huang, L. Pu, J. Org. Chem. 2001, 66, 481–487; c) S.
Thorand, N. Krause, J. Org. Chem. 1998, 63, 8551–8553; d) M.
Sonoda, A. Inaba, K. Itahashi, Y. Tobe, Org. Lett. 2001, 3, 2419–
2
421.
23] C. Niemann, C. E. Redemann, J. Am. Chem. Soc. 1941, 63, 1549–
552.
[42] N. J. Turro, Modern Molecular Photochemistry, University Science
Books, Mill-Valley, California, 1991, pp. 137–143.
[43] R. Katoh, S, Sinha, S. Murata, M. Tachiya, J. Photochem. Photo-
biol. A 2001, 145, 23–34.
[
1
[
[
24] H. Langhals, Chem. Ber. 1985, 118, 4540–4543.
25] a) J. Hofkens, T. Vosch, M. Maus, F. Kçhn, M. Cotlet, T. Weil, A.
Herrmann, K. Müllen, F. C. De Schryver, Chem. Phys. Lett. 2001,
[44] The formation of photoluminescent charge transfer excitons has
been reported by Forrest and co-workers for perylene tetracarboxyl-
ic dianhydride (PTCDA) crystals. These crystals contain 1D stacks
of tightly packed (p–p distance of 3.2 Š) chromophores that facili-
tate a high degree of p-orbital overlap between adjacent molecules
along the stack, see: a) F. F. So, S. R. Forrest, Mol. Eng. Mol. Cryst.
Liq. Cryst. Sci. Technol. Sec. B 1992, 2, 205–222; b) Z. Shen, P. E.
Burros, S. R. Forrest, M. Ziari, W. H. Steier, Chem. Phys. Lett. 1995,
236, 129–134; c) V. Bulovic, P. E. Burrows, S. R. Forrest, J. A.
Cronin, M. E. Thompson, Chem. Phys. 1996, 210, 1–12; d) S. R.
Forrest, Chem. Rev. 1997, 97, 1793–1896.
[45] L. M. Herz, C. Daniel, C. Silva, F. J. M. Hoeben, A. P. H. J. Schen-
ning, E. W. Meijer, R. H. Friend, Phys. Rev. B 2003, 68, 045203.
[46] Bayer, J. Hübner, J. Koptizke, M. Oestreich, W. Rühle, J. H. Wen-
dorff, J. Phys. Chem. B 2001, 105, 4596–4602.
[47] a) A. H. Matsui, Pure Appl. Chem. 1995, 67, 429–436; b) S. Jurse-
nas, A. Gruodis, G. Kodis, M. Chachisvilis, V. Gulbinas, E. A. Sil-
insh, L. Valkunas, J. Phys. Chem. B 1998, 102, 1086–1094.
[48] a) P. G. Schouten, J. M. Warman, M. P. de Haas, J. Phys. Chem. 1993,
97, 9863–9870; b) P. G. Schouten, J. M. Warman, M. P. de Haas, C. F.
van Nostrum, G. H. Gelinck, R. J. M. Nolte, M. J. Copyn, J. W. Zwik-
ker, M. K. Engel, M. Hanack, Y. H. Chang, W. T. Fords, J. Am.
Chem. Soc. 1994, 116, 6880–6894.
3
33, 255–263; b) S. K. Lee, Y. Zu, A. Herrmann, Y. Geerts, K.
Müllen, A. J. Bard, J. Am. Chem. Soc. 1999, 121, 3513–3520.
26] W. Wang, J. J. Han, L.-Q. Wang, L.-S. Li, W. J. Shaw, A. D. Q. Li,
Nano Lett. 2003, 3, 455–458.
27] For reviews on this technique, see: a) C. S. Johnson Jr., Prog. Nucl.
Magn. Reson. Spectrosc. 1999, 34, 203–256; b) Y. Cohen, L. Avram,
L. Frish, Angew. Chem. 2005, 117, 524–560, Angew. Chem. Int. Ed.
[
[
2
005, 44, 520–554.
28] a) W. S. Price, Fumihiko, Tsuchiya, Y. Arata, J. Am. Chem. Soc.
999, 121, 11503–11512; b) W. S. Price in Encyclopaedia of Nuclear
[
[
1
Magnetic Resonance, Vol. 9 (Eds.: D. M. Grant, R. K. Harris), Wiley,
West Sussex, 2002, pp. 364–373.
29] The formation of dimer or higher aggregates not only depends on
the intrinsic properties of the molecules, but also on external condi-
tions, that is, temperature, solvent, pH value of the environment,
etc.; for examples of distinct monomer–dimer equilibria of dye ag-
gregates, see: a) A. S. Shetty, J. Zhang, J. S. Moore, J. Am. Chem.
Soc. 1996, 118, 1019–1027; b) F. Würthner, S. Yao, T. Debaerde-
maeker, R. Wortmann, J. Am. Chem. Soc. 2002, 124, 9431–9447;
c) J. Wu, A. Fechtenkotter, J. Gauss, M. D. Watson, M. Kastler, C.
Fechtenkotter, M. Wagner, K. Müllen, J. Am. Chem. Soc. 2004, 126,
1
1311–11321. For examples of oligomeric or polymeric dye aggre-
gates, see: d) A. W. Snow, N. L. Jarvis, J. Am. Chem. Soc. 1984, 106,
706–4711; R. E. Hughes, S. P. Hart, D. A. Smith, B. Movaghar,
[49] Formation of p-dimers of oligothiophene radical cations, see:
a) M. G. Hill, J. F. Penneau, B. Zinger, K. R. Mann, L. L. Mi ller,
Chem. Mater. 1992, 4, 1106–1113; b) Y. Yu, E. Gunic, B. Zinger,
L. L. Miller, J. Am. Chem. Soc. 1996, 118, 1013–1018; c) J. A. E. H.
van Haare, M. van Boxtel, R. A. J. Janssen, Chem. Mater. 1998, 10,
1166–1175; d) D. A. Scherlis, N. Marzari, J. Phys. Chem. B, 2004,
4
R. J. Bushby, N. Boden, J. Phys. Chem. A 2002, 106, 6638–6645.
30] R. Dobrawa, M. Lysetska, P. Ballester, M. Grüne, F. Würthner, Mac-
romolecules 2005, 38, 1315–1325.
[
448
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 436 – 449

Perylene Bisimide Dyes
FULL PAPER
1
08, 17791–17795; e) T. Sakai, T. Satou, T. Kaikawa, K. Takimiya,
[55] Note added in proof: While this paper was in print, we had the op-
portunity to perform measurements on a recently developed abso-
lute photoluminescence quantum yield measurement system C9920-
02 from Hamamatsu (Japan). For compound 2, the following abso-
T. Otsubo, Y. Aso, J. Am. Chem. Soc. 2005, 127, 8082–8089; for p-
dimers of naphthalene bisimide radical anions, see: f) J.-F. Penneau,
B. J. Stallman, P. H. Kasai, L. L. Miller, Chem. Mater. 1991, 3, 791–
ꢀ
1
7
96; g) L. L. Miller, K. R. Mann, Acc. Chem. Res. 1996, 29, 417–423.
lute quantum yields were determined: 47% for a 0.1 mmolL ag-
gregate solution in n-hexane and 16% for thin films. Both values
are in good agreement with the results shown in Figure 10 and sub-
stantiate our conclusion that a significant photoluminescence is ob-
served from these H-type dye aggregates. We are thankful to Dr.
Tanja Schüttrigkeit and Mr. Hiroyuki Shirai of Hamamatsu Photon-
ics for the opportunity to conduct these measurements with their
help in Würzburg.
[
[
50] P. Prins, K. Senthilkumar, F. C. Grozema, P. Jonkheijm, A. P. H. J.;
Schenning, E. W. Meijer, L. D. A. Siebbeles, J. Phys. Chem. A 2005,
1
09, 18267–18274.
51] D. Wu, A. Chen, C. S. Johnson, J. Magn. Reson. Ser. A 1995, 115,
60–264.
2
[
[
[
52] J. N. Demas, G. A. Grosby, J. Phys. Chem. 1971, 75, 991–1024.
53] R. G. Bennett, Rev. Sci. Instrum. 1960, 31, 1275–1279.
54] J. M. Warman, K.-D. Asmus, R. H. Schuler, J. Phys. Chem. 1969, 73,
Received: June 22, 2006
Revised: October 23, 2006
Published online: December 4, 2006
9
31–939.
Chem. Eur. J. 2007, 13, 436 – 449
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
449