Unraveling the structure and exciton coupling for multichromophoric merocyanine dye molecules

Article abstract of DOI:10.1039/c7cp00115k

The relative orientation of chromophores is a key factor in determining the relationship between the structure and the functionality in molecular multichromophore ensembles. In the case of structurally flexible molecular systems in solution, the task to c

Full text of DOI:10.1039/c7cp00115k

PCCP
View Article Online
View Journal
PAPER
Unraveling the structure and exciton coupling for
multichromophoric merocyanine dye molecules†
Cite this:DOI: 10.1039/c7cp00115k
Federico Koch,a Matthias Stolte,bc Andre Zitzler-Kunkel,c David Bialas,c
´
Andreas Steinbacher,a Tobias Brixner*ab and Frank Wurthner*bc
¨
The relative orientation of chromophores is a key factor in determining the relationship between the
structure and the functionality in molecular multichromophore ensembles. In the case of structurally
flexible molecular systems in solution, the task to clarify the relevant effects of accessible chromophore
orientations with spectroscopic observations is very demanding. In this study, we address this issue by
investigating a series of differently connected multichromophoric systems composed of highly dipolar
merocyanine dyes that are systematically varied in their substitution pattern and the number of
chromophores attached to a bridging benzene ring. Combining electro-optical absorption (EOA) and UV/Vis
spectroscopy with density functional theory (DFT) as well as exciton theory discloses conformational
preferences and rationalizes the optical properties of the interacting chromophores. Our findings suggest for
all multichromophoric systems there is a relative orientation of the chromophores which compensates for
the individual dipole moments of the merocyanine dyes by pointing preferably in opposing directions. These
orientations furthermore rationalize the observed spectral properties by partly excitonically-coupled subunits.
Received 6th January 2017,
Accepted 3rd February 2017
DOI: 10.1039/c7cp00115k
rsc.li/pccp
to be deduced. Furthermore, control of conformation via inter-
molecular interactions can serve as a rational approach for
Introduction
Chromophore–chromophore interactions play an important the supramolecular design of functional organic materials for
role in determining the functionality of supramolecular photonics and electronics.14
assemblies1–5 and functional materials.6–9 This is exemplified
The determination of molecular structures, both configu-
by the p-stack architectures of structures such as photo- ration and conformation,15–24 is commonly performed via X-ray
synthetic light-harvesting systems or DNA which are determined crystallography in the solid or liquid crystalline state25–27 and
by particular non-covalent interactions among dyes and with the nuclear magnetic resonance (NMR) spectroscopy in the solid
molecular matrix environment, for example, a protein or a state or in solution.28–31 These contemporary analytical techni-
membrane.10–12 On the other hand, in artificial synthetic multi- ques may be limited in the determination of a molecular
chromophore ensembles, the structure originates from conforma- structure, e.g., ambiguous couplings in the NMR analysis can
tional preferences which may include interchromophore occur or the molecular systems do not crystallize well enough for
interactions. In this case, attractive interactions between single dye a thorough X-ray analysis. However, even if molecular crystals
units determine the chromophore arrangement and orientation. are accessible, the crystalline structure depends on conditions
Therefore, these interactions have a high impact on the resulting during crystallization such as the solvent and the ambient
functional properties such as exciton coupling.13 Analyzing the temperature and might not represent the preferential molecular
molecular conformation in liquid media allows the interplay structure in solution or other environments. Generally, mole-
between structural arrangement and chromophore interactions cules in solution display a higher structural degree of freedom
than in the solid phase, and are additionally influenced by
interactions between the solute and solvent. Hence, even a
a
¨
¨
¨
Institut fur Physikalische und Theoretische Chemie, Universitat Wurzburg,
dynamic distribution of molecular structures is typically present.
Next to the aforementioned experimental techniques, theo-
retical calculations give access to molecular structures. In this
case information is obtained by geometry optimization, e.g., in
DFT calculations,32,33 providing relative energies of different
molecular conformations. Hence, a molecular geometry to start
the optimization has to be chosen carefully to avoid local
¨
Am Hubland, 97074 Wurzburg, Germany.
E-mail: brixner@phys-chemie.uni-wuerzburg.de; Fax: +49 931 31-86332;
Tel: +49 931 31-86330
b
c
¨
¨
Center for Nanosystems Chemistry (CNC), Universitat Wurzburg, Theodor-Boveri-Weg,
¨
97074 Wurzburg, Germany
¨
¨
¨
¨
Institut fur Organische Chemie, Universitat Wurzburg, Am Hubland, 97074 Wurzburg,
Germany. E-mail: wuerthner@chemie.uni-wuerzburg.de; Fax: +49 931 31-84756;
Tel: +49 931 31-85340
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7cp00115k minima. Furthermore, the obtained structures might not
This journal is ©the Owner Societies 2017
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
represent the actual geometry in solution as the molecular
environment is typically neglected. More advanced theoretical
calculations are capable of molecular structure optimization in
media by modeling the surrounding medium or by including
correction parameters based on experimental results.34,35 In
consequence, these methods have higher computational costs
especially for larger and multichromophoric systems.
When confronted by the challenge of the structure determi-
nation of multichromophoric molecular systems in solution,
for which NMR spectroscopy and X-ray crystallography do not
provide unambiguous observations, linear absorption spectra
have been proven to contain valuable structural information,
e.g., by analyzing the orientation and distance dependence
using exciton theory13 or by the analysis of the aggregation
process itself.6,36–40 Additionally, ultrafast laser pulses41,42
in combination with X-ray absorption43 have been shown to
provide information on the molecular structure.
Merocyanines have been investigated frequently in the con-
text of nonlinear optics due to their outstanding molecular
properties, e.g., high-ground state dipole moments, tunable
(hyper-)polarizabilities, and strong and narrow charge-transfer
(CT) bands in the visible region.44,45 Merocyanines consist of a
donor (D) and an acceptor (A) moiety which are linked by a
polymethine bridge, which is variable in length. These donor–
acceptor (D–A) systems can be described by two resonance
structures, namely a neutral polyene structure and secondly,
a zwitterionic resonance structure, due to delocalization of
p-electrons in the merocyanines over D, A, and the polymethine
linker.46–49 The excited-state dynamics of merocyanine dyes are
drawing the attention of researchers since they feature vibrational
wavepacket motions,50–52 isomerization reactions,50–55 and the
existence of intramolecular charge-transfer (ICT) states.56–58 Allow-
ing versatile options of structural variation, and therefore enabling
the fine-tuning of electronic and optical properties, merocyanines
provide access to a large field of applications, e.g., in biology and
medicine,57,59 nonlinear optics and photorefractive materials,47,60
in organic solar cells,61,62 and field-effect transistors.63–65
Fig. 1 (a) Chemical structures of the studied mono-merocyanine 1 (red)66
and multichromophoric merocyanines 2p (blue), 2m (brown), and 3
(green). Labels next to the chemical structures describe the number of
chromophores. To distinguish the bis-merocyanine 2, the position of the
chromophore is labeled m (meta) and p (para). The chromophores of 3 are
Fig. 1a displays a series of merocyanines, which is system-
atically varied in the number of identical chromophores
(labeled 1–3) and the substitution pattern (para for 2p or meta all linked to each other in the meta position at the benzene unit. (b) Linear
absorption spectra of all studied dyes dissolved in dichloromethane,
for 2m and 3) at a benzene bridge. Considering the linear
absorption spectra in Fig. 1b, distinct differences in the relative
intensities of the absorption spectra of the multichromophoric
displaying strong differences in their absorption maxima and relative
absorption band intensities.
systems can be observed compared to 1. The monomer (1, red in
Fig. 1b)66 exhibits an absorption maximum at 633 nm, corres- chromophores for multichromophoric merocyanines. The
ponding to the S0–S1 transition and blue-shifted vibrational origin and the nature of the observed chromophore inter-
progressions.67 While 2p (blue in Fig. 1b) displays a 12 nm actions in the linear absorption spectra cannot be inferred
bathochromically shifted absorption maximum at lmax = 645 nm easily. In our recent time-resolved spectroscopic study of the
with a small additional absorption band at 602 nm, the meta- aforementioned dyes, we could infer a stabilization of the ICT
linked dye, 2m (brown in Fig. 1b), displays a hypsochromically states with increasing number of chromophores, possibly due
shifted absorption maximum at lmax = 626 nm with an intense to electrostatic interactions.58 However, only the spectroscopic
additional absorption band at 593 nm. The trimer 3 (green in information about excited-state lifetimes is not sufficient to
Fig. 1b) displays an absorption maximum at lmax = 623 nm with deduce the interaction mechanism. Hence, we address the
an additional absorption band at 598 nm, and a third batho- question of how chromophore interactions are reflected in the
chromically shifted band at E652 nm. Consequently, the changes linear absorption spectra of multichromophoric merocyanines.
in linear absorption spectra display interactions between Since such chromophore interactions strongly depend on the
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2017

View Article Online
PCCP
Paper
relative chromophore orientation, the predominant molecular reaction of the latter compounds with sodium thiocyanate
structure in solution has to be determined to be able to resolve in ethanol afforded the thiocyanate compounds 5a–c. The
the origin of these interactions. In this study, we approach these subsequent condensation reaction of 5a–c with malononitrile
questions experimentally by electro-optical absorption (EOA). under basic conditions yielded the corresponding bis- and
Additionally, to rationalize spectroscopic observations, density tris-thiazolebenzene compounds 4a–c in nearly quantitative
functional theory (DFT) calculations are performed. Based on yields and with good purity, thus these compounds were used
these results, we identify the relative chromophore orientations for the final step without further purification. The target bis- and
and the arising optical properties of multichromophoric mero- tris-merocyanine compounds 2p, 2m and 3 were then achieved
cyanines on the basis of exciton theory in solution.13,68
by the Knoevenagel condensation reaction of the thiazole deriva-
tives 4a–c with indolenine aldehyde 8; the latter was prepared
according to the literature,66 in acetic anhydride in 10–51%
isolated yields. The synthetic procedures and characterization
data for the new compounds are reported in the ESI.†
Results and discussion
Synthesis
The phenyl-substituted mono-merocyanine 1 was reported Electro-optical absorption spectroscopy
previously and synthesized according to the literature.66 The
new benzene-bridged bis-merocyanines 2p and 2m and tris-
merocyanine 3 were synthesized in four steps according to the
route displayed in Scheme 1.
In the first step of the synthesis, the respective commercially
available 1,4-(7a), 1,3-(7b) di- and 1,3,5-(7c) triacetylbenzenes
were brominated at the terminal methyl group under acidic con-
ditions. 1,4-Diacetylbenzene (7a) was brominated with elemental
bromine in glacial acetic acid,69 while the bromination of
1,3-diacetylbenzene (7b) and 1,3,5-triacetylbenzene (7c) was per-
formed with N-bromosuccinimide (NBS) and p-toluene sulfonic
acid ( p-TsOH) in acetonitrile,70 leading to the corresponding
bromoacetyl derivatives 6a–c. The nucleophilic substitution
Electro-optical absorption (EOA) spectroscopy determines the
very small difference in absorption of a molecular solution with
[eE(j, n~)] and without [e(n~)] an externally applied electric field E,
measured with light polarized parallel (j = 01) and perpendi-
cular (j = 901) to the field direction of E. The relative change in
absorption due to the effect of the applied external electric field
can be described by the quantity L(j, n~), which is defined by
eEðj; n~Þ ꢀ eðn~Þ
eðn~Þ
1
E2
Lðj; n~Þ ¼
ꢁ
:
(1)
This depends on the orientation of the molecules due to their
ground-state dipole moment, mg, the shift of the absorption
band proportional to the dipole moment difference Dm = me ꢀ mg,
with me being the excited-state dipole moment, and the electric
field dependence of the transition dipole moment meg(E).66 The
analysis of the EOA response is conducted via multilinear
regression of the EOA spectra, in which spectral features as well
as molecular properties are included. A detailed description of
the multilinear regression is given by Liptay and Wortmann.71–73
Additionally, information about the dipole moment mg can be
retrieved by the linear combination
Lp(j, n~) = 6[L(01, n~) ꢀ 3L(901, n~)],
(2)
2
which is proportional to mg .
Fig. 2 displays the EOA response and the corresponding Lp
values of all investigated dyes along with their linear absorption
spectrum measured in chloroform. The results of the multi-
linear analysis are collected in Table 1. Depending on the
substitution pattern, remarkable changes in the shape intensity
of the EOA response are observable, which are discussed in
detail as follows.
Dye 1 exhibits a typical electro-optical response for a
strongly dipolar push–pull chromophore close to the cyanine
limit: it exhibits an increase in absorbance for parallel polarized
light with respect to E (Fig. 2a; open symbols) and a decrease for
perpendicularly polarized light (filled symbols) due to the prefer-
ential orientation of the dipolar molecules parallel to the electric
field. This so-called positive electrodichroism [L(01) 4 L(901)] is
representative of a relative orientation (a) of the transition dipole
moment meg in the range of 01 to 54.71 with respect to mg, and
one can easily draw the conclusion that a is close to zero, as is
Scheme 1 Synthesis of benzene-bridged bis-merocyanines 2p and 2m,
and tris-merocyanine 3.
This journal is ©the Owner Societies 2017
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
Fig. 2 Left column: linear absorption (dashed) and EOA (j = 01: open symbol; j = 901: filled symbol) spectra as well as the multilinear regression (solid
lines) of 1 (a, red), 2p (c, blue), 2m (e, brown), and 3 (g, green) in chloroform at 298 K. Please note the color code, i.e., red-dashed lines for reference dye 1
are included in all figures for comparison. Right column: linear absorption spectrum (dashed) and the linear combination Lp (open symbols) of all dyes 1
(b, red), 2p (d, blue), 2m (f, brown), and 3 (h, green) in chloroform at 298 K. Horizontal solid lines correspond to the calculated Lp value determined by
multilinear regression. For comparison, the results of 1 are included in all panels (red).
the case for many structurally related push–pull dyes.44,45,66 limit. Furthermore, the determined Dm values can be used
The linear regression and the wavenumber independence of Lp together with the meg values to classify dyes by the square of
(Fig. 2b) confirm that only homogeneously polarized transitions the resonance parameter, c2, which is calculated via75
are present and the quantitative analysis with a = 01 (J) is valid.
The obtained ground state dipole moment of 43.7 ꢁ 10ꢀ30 C m
is in accordance with previously reported values of other
dipolar merocyanine dyes with similar donor and acceptor
moieties.45,49,75,76 Additionally, as the maxima of the EOA and
UV/Vis spectra almost coincide, the change in the dipole
0
1
1
2
Dm
B
C
A
c2 ¼
1 ꢀ
:
(3)
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
@
4meg þ Dm2
2
moment upon optical excitation (Dm) is very small and the Dyes are classified into polyene-like (c2 E 0; Dm 4 0 C m), cyanine-
chromophore’s resonance structure is almost in the cyanine like (c2 E 0.5; Dm E 0 C m), or betaine-like (c2 E 1; Dm o 0 C m).
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2017

View Article Online
PCCP
Paper
Table 1 (Electro-)optical properties determined by linear UV/Vis absorp-
low-energetic spectral edge of the absorption band is detectable
for both L(01) and L(901). This could indicate that the applied
electric field results in small changes in the conformational
distribution and slightly shifts the conformational equilibrium
(Scheme 2).
The electro-optical response of 2m (Fig. 2e) is again reduced
in comparison to 1 and the decrease in dipolarity can be further
visualized by a smaller Lp (Fig. 2f). In contrast to 1 and 2p, 2m
exhibits negative electrodichroism [L(01) o L(901)], which
indicates that meg has an angle between 54.71 and 901 with
respect to mg. Our analysis of 2m does not properly reproduce
tion and EOAa spectroscopy for all dyes in chloroform
Dye
1
2p
2m
3
lmax/nm
637
652
17 200
48.6
632
17 700
49.2
630
20 400
59.1
emax/m2 molꢀ1 10 200
m
eg/10ꢀ30 C m 34.8
mgb/10ꢀ30 C m 43.7 ꢂ 1.0 (J) 12.2 ꢂ 0.7 (J) 20.7 ꢂ 2.5 (>) —c
meb/10ꢀ30 C m 41.3 ꢂ 1.6 (J) ꢀ1.1 ꢂ 1.6 (J)
—
—
—
c
c
c
Dmb/10ꢀ30 C m ꢀ2.4 ꢂ 1.2 (J) ꢀ13.3 ꢂ 1.2 (J) —c
a
Corrected to give ‘‘gas phase’’ dipole moments by Onsager cavity field
correction.74 b Calculated under the approximation that - (meg, mg) = 01
c
and - (meg, mg) = 901 indicated by (J) and (>), respectively. Were not
determined as the experimental data could not be fitted appropriately the spectral band structure of the EOA spectra, indicating
by multilinear regression analysis.
heterogeneously polarized transitions which superimpose to
give a linear absorption spectrum as expected for conforma-
Hence, the values c2 = 0.52 and Dm = ꢀ2.4 ꢁ 10ꢀ30 C m allow the tionally heterogeneous ensembles. Still, the Lp of the EOA
classification of 1 as a cyanine-like dye.
measurements of 2m leads to the conclusion that again the
The bis-merocyanine 2p also displays positive electrodichroism two highly dipolar merocyanine dyes preferentially adopt con-
[L(01) 4 L(901)] in the EOA spectrum, however, of much smaller formations which compensate most of their individual mg
intensity than 1 (see Fig. 2c). This first indication of a smaller mg, within the bis-merocyanine. Hence, both chromophores point
compared to 1, can be further visualized by the tremendously in predominantly opposing directions to reduce the total mg of
diminished Lp (compare Fig. 2d). The multilinear analysis repro- the bichromophore (cf. the structure 2m-II in Scheme 3, in
duces the EOA spectrum well, which indicates the homogeneity of contrast to the predominantly parallel orientations of struc-
the investigated absorption bands, like in the case of 1. The tures 2m-I and 2m-III). However, the dipole moment of 2m is
analysis of the EOA measurement of 2p thus reveals that two not reduced as efficiently as in the case of 2p, which is
highly dipolar merocyanine dyes adopt a mean conformation with rationalized by the fact that the meta-connectivity cannot afford
respect to the central spacer unit which compensates most of an energetically favored conformation with an inversion center
their individual mg to be about 12.2 ꢁ 10ꢀ30 C m. The most such as for 2p.
efficient compensation of mg is realized if both individual
The electro-optical response of 3 is again reduced in com-
chromophores of 2p perfectly point in opposite directions parison to 1, exhibiting again as with 1 and 2p, positive
(cf. structure 2p-I in Scheme 2, in contrast to structure 2p-II). electrodichroism [L(01) 4 L(901)]. Experimental data for 3 could
In this conformation two highly dipolar merocyanines are not be fitted appropriately by multilinear regression analysis
arranged in an antiparallel manner, i.e., they would display
an inversion center which would result in mg = 0 C m. However,
due to structural flexibility and fluctuations, individual chromo-
phores can break this symmetry and generate on average a small
mg, i.e., reduced vastly by more than 70% compared to the mg of 1.
Interestingly, regarding the spectral changes of L(01) and L(901)
for 2p (cf. Fig. 2c), an increasing absorption for L(01) is obser-
vable which displays a similar spectral shape as the absorption
spectrum of 1. Additionally, a decrease of the absorption at the
Scheme 3 Different conformations of 2m through rotation of one or two
Scheme 2 Different conformations of 2p through rotation of one or two
chromophores around the benzene bridge. Black arrows indicate the
direction of the individual ground state dipole moments of the monomeric
subunits. The structures 2p-I and 2p-II depict anti-parallel and parallel
chromophore orientations for 2p, respectively.
chromophores around the benzene bridge. Black arrows indicate the
direction of the individual ground state dipole moments of the monomeric
subunits. The structures 2m-I, 2m-II and 2m-III depict different structures
of 2m with predominantly parallel (I and III) and anti-parallel (II) chromo-
phore orientation.
This journal is ©the Owner Societies 2017
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
(cf. Fig. 2g), which indicates again an overlap of hetero- (DFT, B3LYP/def2-SVP level) calculations were performed for
geneously polarized bands. The reason originates from the the mono- (1) and the two bis-merocyanines (2p and 2m)
overlap of multiple differently polarized transitions of a starting from different chromophore orientations, i.e., parallel
conformationally heterogeneous ensemble in solution. This or anti-parallel, as shown in Fig. S7 and S8 of the ESI.† Fig. 3
assumption is supported by the wavenumber dependence of depicts the results of the geometry optimization and the
Lp (see Fig. 2h) over the whole region of the absorption band. resulting electron densities of the frontier molecular orbitals
Although a quantitative analysis at this point is not possible for 1, 2p and 2m.
without the knowledge of the individual band shapes of the
For 1 the highest occupied molecular orbital (HOMO) is
overlapping transitions, one can deduce that the overall dipo- delocalized along the whole polymethine chain of the mero-
larity of 3 is, like in the case of the bis-merocyanines 2p and 2m, cyanine chromophore, without extending into the phenyl sub-
on average reduced. This is in accordance with the idea that stituent at the thiazole heterocycle. In contrast, the electron
electrostatic interactions between highly dipolar merocyanine density of the lowest unoccupied molecular orbital (LUMO) is
dyes lead to a preferential population of conformations which extended into the spacer. The lowest energy optical transition
compensate most of their individual ground-state dipole can be related to a HOMO - LUMO excitation with a transition
moments mg. An efficient compensation of mg could be realized, dipole moment meg = 34.7 ꢁ 10ꢀ30 C m, which corresponds nicely
e.g., by a propeller-like arrangement, where the donor moiety of to the experimentally determined value of meg = 34.8 ꢁ 10ꢀ30 C m
one highly dipolar merocyanine is in proximity to the neighbor- (cf. Table 1).
ing chromophore.
In the case of 2p, the geometry-optimized structure displays
two merocyanine chromophores which are almost in one plane,
pointing in opposite directions. The bridging benzene unit is
Density functional theory
In the previous section, the predominant chromophore orien- rotated out of this plane, possibly due to steric effects. Con-
tations were revealed for the bis-merocyanines 2p and 2m cerning the electron density, localization on the chromophores,
based on the spectral changes in EOA measurements and similar to 1, is observed for the HOMO of 2p. Due to structural
presumed electrostatic interactions. In order to validate the symmetry, both merocyanines are energetically degenerate,
previously described observations, density functional theory which explains why the HOMO as well as the LUMO electron
Fig. 3 Structure optimization (B3LYP/def2-SVP level) and electron densities of the related molecular orbitals for the intramolecular charge-transfer
transition of 1, 2p, and 2m. For 1 and 2p the corresponding HOMO and LUMO are depicted. In the case of 2m the HOMOꢀ1 and LUMO are shown.
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2017

View Article Online
PCCP
Paper
density is located on both chromophores. The electron distri- corresponding to the S0 - S1 transition and its vibrational
bution of the LUMO spreads from the acceptor moieties of both progressions (for details, see ESI†). The four Gaussians were forced
merocyanine chromophores over the bridging benzene unit, to have the same bandwidth.77 The resulting energetic distances
which is in agreement with the conjugation of the p-electron (see Fig. S9 and Table S2, ESI†) are thus approximately identical.
system. The structural optimization of 2p thus strongly agrees
with the proposed structure (structure 2p-I in Scheme 2) based systems under investigation (2p and 2m), we stick to the point-
on the EOA measurements. dipole approximation [eqn (4)] and thus include two more
To model the excitonic splitting in the bis-chromophoric
The geometry optimization of 2m leads to an energetically Gaussians (again with an identical bandwidth) to fit the linear
favored orientation where the two merocyanine subunits are as absorption spectra of 2p and 2m (for details, see ESI†). This
anti-parallel as possible (similar to 2m-II in Scheme 3). In this takes into account the energy splitting of the lowest vibrational
conformation, the two chromophores point in different direc- S0 - S1 transition. One can already deduce from the shape of
tions, similar to 2p, but in this case the chromophores are the linear absorption spectra, as well as from our EOA results,
twisted. Like in the case of 1 and 2p, the electron density of the that no pure H- or J-type interaction can be expected, since the
HOMOꢀ1 for 2m is localized on the merocyanine unit. How- monomeric subunits and thus their respective dipoles show an
ever, in stark contrast to 2p, the localization is only on one of oblique orientation. In other words, we model 2p and 2m as
the two merocyanine chromophores as a result of the broken quasi-flexible systems which are approximated by the super-
symmetry. The electron distribution of the LUMO of 2m is position of a ‘‘monomer’’ and a ‘‘dimer’’ spectrum.
localized likewise on a single chromophore, with small con-
tributions on the benzene bridging unit.
Discussion of chromophore interaction
and orientation
The presented results suggest that there are interactions between
the chromophores of the bis- and tris-merocyanines under
investigation. This is obvious from the substantial differences
in the linear absorption spectra presented in Fig. 1 already,
however the results of EOA and previously performed TA
experiments58 also show deterministic variations. In the follow-
ing section we will elaborate on the origin of these chromophore
interactions.
In the case of a strong inter-chromophoric coupling, molecular
excitation leads to a delocalized excitation, which can be treated
via exciton theory, following the seminal work of Kasha.13 The
exciton coupling energy V in the dipole–dipole approximation can
be calculated via
À
ÁÀ
Á
^
^
~m1 ꢃ~m2 ꢀ 3 ~m1 ꢃ R ~m2 ꢃ R
1
V ¼
;
(4)
4pe0
R3
with e0 as the dielectric constant in a vacuum, the unit vector
-
-
-
-
ˆ
R = (r2 ꢀ r1/|r2 ꢀ r1|) as the relative position vector of the two
subunits, i.e., point-dipoles, positioned at -r1 and -r2 in relation to
the origin, R as the spatial separation and ~mn as the transition
dipole moment of the first (n = 1) and the second chromophore
(n = 2).13,36 The corresponding values of V derived from
the geometry-optimized structure from the DFT calculations
2p
DFT
2m
DFT
(cf. Table S1, ESI†) are V
= 249 cmꢀ1 and V
= 568 cmꢀ1
,
indicating an exciton coupling energy VDFT for 2m which is
roughly twice the one of 2p.
To compare these values with experimental results, we turn
again to the linear absorption spectra for a more detailed
analysis.77 To model these spectra, we have to take into account
the pronounced vibronic progression, which is common in
these merocyanine dyes.49 Hence, in the first step we model
the linear absorption spectrum of 1 with four Gaussians
Fig. 4 Linear absorption (blue for 2p and brown for 2m), four Gaussians
with the positions of the fitted absorption spectrum of 1 (black, dashed)
with optimized amplitudes (see Fig. S3 and Table S1, ESI†), and additional
two Gaussians to represent the excitonic bands (red, solid) for 2p (a) and
2m (b), respectively. The fit result for both dimers is shown in dashed pink
lines. Details of the fitting routine are given in the ESI.†
This journal is ©the Owner Societies 2017
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
As one can see from the fitting results in Fig. 4, it is possible (and widths) of the four Gaussians used to fit the spectrum of 1
to reproduce the shape of the absorption spectra of 2p and 2m were kept fixed and only their amplitudes were allowed to vary
quite accurately with this approach. For this fit, the positions (see ESI†). The results (cf. Fig. 4) indicate coupling energies
2p
2m
of V = 700 cmꢀ1 and V = 647 cmꢀ1. Both values (Vfit) are
fit
fit
larger than those calculated for the geometry-optimized struc-
2p
DFT
2m
DFT
tures (V
= 249 cmꢀ1, V
= 568 cmꢀ1) and even show a
different trend regarding the para and meta orientations. This
observation might be rationalized due to several issues with the
performed analysis. (1) The VDFT values represent energetically
minimized structures in the gas phase where dipole–dipole
interactions are much stronger as in a polarizable solvent
environment. (2) The point-dipole approximation [eqn (4)]
might be limited when modeling such systems. (3) Already in
the EOA measurements a certain degree of flexibility and
structural distribution in the solution phase could be deduced,
which is not covered in the modeling.
Therefore, we want to evaluate the influence of the relative
chromophore orientation on V, to test the validity of our initial
approach. Thus, we calculated the exciton-splitting energy for
various geometries of the two subunits in 2p and 2m. In these
calculations of V, one chromophore of the bis-chromophoric
systems was fixed in space with the coordinates obtained using
DFT and the second chromophore’s position was varied by its
tilting angle j and rotation angle g. Fig. 5 depicts the chemical
structures and calculated V values [cf. eqn (4) and ESI†]. While V
of 2p only displays minor changes in the variation of g, a
distinct dependence on j can be observed. Changes in j
of E101 cause V E 700 cmꢀ1, and are obtained by Vf2ipt. In
contrast, 2m displays strong variations in V depending on both
angles, j and g. This observation can be rationalized by the
proximity of the two chromophores in 2m. Similarly to 2p, the
2m
fit
estimated value V is accessed by small variations of j and g
of E101 of one chromophore. As only one chromophore is
varied in its orientation to the fixed other chromophore, it
should be pointed out that larger V values are accessible already
by the flexibility of one chromophore and are even more
probable for variations of both chromophores in the dynamic
system in solution. These results show that even in the point-
dipole approximation13 such complex changes in the linear
absorption spectra of bis-chromophoric systems can be
explained and that the structural flexibility in solution is indeed
their origin.
Conclusion
The structure determination of conformationally flexible mole-
cules in solution is a complex task due to limitations of
the available experimental and theoretical tools. Alternative
approaches are accordingly needed to investigate interactions
between chromophores, and thus obtain insights into their
structure and function.
With this goal in mind and focusing on an important class
of dyes, i.e., dipolar push–pull chromophores, we showed in
this study that electro-optical absorption (EOA) spectroscopy
can provide structural insights for multichromophoric highly
Fig. 5 Molecular structure and exciton splitting energy V depending on
the chromophore tilting angle j and rotation angle g for 2p (a) and 2m (b),
respectively. For the calculation of V the optimized structures of the DFT
calculations were set as starting geometries (marked as crosses in the
center of the maps). Subsequently, one chromophore was fixed in this
geometry, while the other chromophore orientation was varied in j and g.
The resulting values were utilized to calculate V via eqn (4).
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2017

View Article Online
PCCP
Paper
dipolar merocyanines (mg = 43.7 ꢁ 10ꢀ30 C m) because the Academic Foundation (Studienstiftung des deutschen Volkes)
spectral changes upon orientation of the molecules by electric for a scholarship. DB thanks the Fonds der Chemischen
fields are dependent on the overall dipole moment which is Industrie for a PhD scholarship.
strongly dependent on the molecules’ conformation. Additionally,
spectral changes indicate the small influence of the external
electric field on the conformational distribution. These observa-
tions led to a conformational assignment for the multichromo-
References
¨
phoric system that compensates most of the ground-state dipole
moment by antiparallel alignment of the individual chromo-
phores due to electrostatic reasons.
1 Z. Chen, A. Lohr, C. R. Saha-Moller and F. Wu¨rthner,
Chem. Soc. Rev., 2009, 38, 564–584.
2 R. Bhosale, J. Mısek, N. Sakai and S. Matile, Chem. Soc. Rev.,
´ˇ
This structural assignment was supported by DFT calcula-
tions. Starting from different geometries, the para-substituted
dimer (2p) displayed the energetically most stable conforma-
tion with the two chromophores pointing in mutually opposite
directions (cf. Fig. 5a). For the meta-substituted dimer (2m) an
orientation whereby both chromophores point in different
directions was also observed (cf. Fig. 5b). A closer proximity
of neighboring chromophores in 2m and 3 with respect to 2p
allows a more efficient stabilization of ICT states via electro-
static interactions.58 For the trimer (3) the EOA analysis is
ambiguous such that we refrain from a more detailed structural
assignment.
2009, 39, 138–149.
3 T. Weil, T. Vosch, J. Hofkens, K. Peneva and K. Mu¨llen,
Angew. Chem., Int. Ed., 2010, 49, 9068–9093.
4 D. M. Bassani, L. Jonusauskaite, A. Lavie-Cambot, N. D.
McClenaghan, J.-L. Pozzo, D. Ray and G. Vives, Coord. Chem.
Rev., 2010, 254, 2429–2445.
5 F. C. Spano, Acc. Chem. Res., 2010, 43, 429–439.
6 F. Wu¨rthner, T. E. Kaiser and C. R. Saha-Moller, Angew.
Chem., Int. Ed., 2011, 50, 3376–3410.
7 Z. B. Henson, K. Mu¨llen and G. C. Bazan, Nat. Chem., 2012,
¨
4, 699–704.
8 S. S. Babu and D. Bonifazi, ChemPlusChem, 2014, 79,
Lastly, by analyzing the effect of chromophore orientation
on the exciton splitting energy (V) as expressed in the UV/Vis
absorption spectra, it was shown that experimentally observed
895–906.
9 K.-T. Wong and D. M. Bassani, NPG Asia Mater., 2014,
6, e116.
values are accessible by slight changes in the relative chromo- 10 K. N. Ferreira, T. M. Iverson, K. Maghlaoui, J. Barber and
phore orientation. The linear spectra thus display a super- S. Iwata, Science, 2004, 303, 1831–1838.
position of monomer-like and exciton-coupled contributions 11 D. Noy, C. C. Moser and P. L. Dutton, Biochim. Biophys. Acta,
validating the conclusions from EOA and DFT. By combining 2006, 1757, 90–105.
experimental results (EOA) and calculations (DFT) we funnelled all 12 R. J. Cogdell, A. Gall and J. Kohler, Q. Rev. Biophys., 2006, 39,
information with the help of exciton theory into a comprehensive 227–324.
understanding of a flexible multichromophoric molecular system 13 M. Kasha, H. R. Rawls and M. Ashraf El-Bayoumi, Pure Appl.
in solution. We are convinced that this novel approach of combin- Chem., 1965, 11, 371–392.
¨
ing UV/Vis- and EOA spectroscopy together with DFT and exciton 14 F. Wu¨rthner, Acc. Chem. Res., 2016, 49, 868–876.
theory will lead to valuable insights into the conformational 15 C. Song and T. M. Swager, Org. Lett., 2008, 10, 3575–3578.
properties of other multichromophore systems of dipolar dyes 16 R. M. Meudtner and S. Hecht, Angew. Chem., Int. Ed., 2008,
that are of relevance for functional materials design.
47, 4926–4930.
17 M. Wolffs, N. Delsuc, D. Veldman, N. V. Anh, R. M.
Williams, S. C. J. Meskers, R. A. J. Janssen, I. Huc and
A. P. H. J. Schenning, J. Am. Chem. Soc., 2009, 131, 4819–4829.
18 K. Miwa, Y. Furusho and E. Yashima, Nat. Chem., 2010, 2,
444–449.
19 F. Schlosser, M. Moos, C. Lambert and F. Wu¨rthner, Adv.
Mater., 2013, 25, 410–414.
20 I. Pochorovski and F. Diederich, Acc. Chem. Res., 2014, 47,
2096–2105.
21 S. Saito, J.-Y. Shin, J. M. Lim, K. S. Kim, D. Kim and
A. Osuka, Angew. Chem., Int. Ed., 2008, 120, 9803–9806.
22 M. J. Marsella, R. J. Reid, S. Estassi and L.-S. Wang, J. Am.
Chem. Soc., 2002, 124, 12507–12510.
23 C. Yuan, S. Saito, C. Camacho, S. Irle, I. Hisaki and
S. Yamaguchi, J. Am. Chem. Soc., 2013, 135, 8842–8845.
Experimental methods
Steady-state absorption spectra were measured with a Jasco
V-670 UV-Vis spectrometer. Electro-optical absorption was per-
formed using a home-made instrument under highly diluted con-
ditions (10ꢀ6 M) in anhydrous chloroform. The electrochromism
eE(j, n~) ꢀ e(n~) was induced by both an alternating and a constant
electric field of about 3 ꢁ 106 V mꢀ1 and was determined by a
lock-in technique.
Acknowledgements
We thank the German Science Foundation (DFG) for funding
within the Research Unit ‘‘Light-Induced Dynamics in Molecular 24 C. Yuan, S. Saito, C. Camacho, T. Kowalczyk, S. Irle and
Aggregates’’ (FOR 1809) as well as within the Graduate College S. Yamaguchi, Chem. – Eur. J., 2014, 20, 2193–2200.
‘‘Control of Electronic Properties of Assemblies of pi-Conjugated 25 M. F. C. Ladd and R. A. Palmer, Structure Determination by
Molecules’’ (GRK 1221). AS thanks the German National
X-Ray Crystallography, Springer US, Boston, MA, 1995.
This journal is ©the Owner Societies 2017
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
26 M. S. Smyth and J. H. J. Martin, Mol. Pathol., 2000, 53, 44 S. R. Marder, C. B. Gorman, F. Meyers, J. W. Perry,
´
8–14.
G. Bourhill, J.-L. Bredas and B. M. Pierce, Science, 1994,
27 Y. Waseda, E. Matsubara and K. Shinoda, X-Ray Diffraction
265, 632–635.
Crystallography: Introduction, Examples and Solved Problems, 45 V. Parthasarathy, R. Pandey, M. Stolte, S. Ghosh, F. Castet,
Springer Science & Business Media, 2011.
28 S. Ganapathy, S. Sengupta, P. K. Wawrzyniak, V. Huber,
F. Wu¨rthner, P. K. Das and M. Blanchard-Desce, Chem. –
Eur. J., 2015, 21, 14211–14217.
F. Buda, U. Baumeister, F. Wu¨rthner and H. J. M. de Groot, 46 B. Kippelen, S. R. Marder, E. Hendrickx, J. L. Maldonado,
Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 11472–11477.
29 S. Nozinovic, B. Fu¨rtig, H. R. A. Jonker, C. Richter and
H. Schwalbe, Nucleic Acids Res., 2009, 1–12.
G. Guillemet, B. L. Volodin, D. D. Steele, Y. Enami,
¨
S. Sandalphon, Y. J. Yao, J. F. Wang, H. Rockel, L. Erskine
and N. Peyghambarian, Science, 1998, 279, 54–57.
30 G. Harald, NMR Spectroscopy: Basic Principles, Concepts and 47 F. Wu¨rthner, R. Wortmann and K. Meerholz, ChemPhysChem,
Applications in Chemistry, Wiley-VCH, 3rd edn, 2013.
2002, 3, 17–31.
31 M. R. Hansen, R. Graf and H. W. Spiess, Chem. Rev., 2016, 48 L. Beverina and G. A. Pagani, Acc. Chem. Res., 2014, 47,
116, 1272–1308.
319–329.
32 R. G. Parr and Y. Weitao, Density-Functional Theory of Atoms 49 F. Wu¨rthner, G. Archetti, R. Schmidt and H.-G. Kuball,
and Molecules, Oxford University Press, 1994.
Angew. Chem., Int. Ed., 2008, 120, 4605–4608.
33 W. Koch and M. C. Holthausen, A Chemist’s Guide to Density 50 P. Nuernberger, S. Ruetzel and T. Brixner, Angew. Chem.,
Functional Theory, Wiley, 2001.
Int. Ed., 2015, 54, 11368–11386.
34 J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 51 S. Ruetzel, M. Diekmann, P. Nuernberger, C. Walter,
105, 2999–3094.
B. Engels and T. Brixner, J. Chem. Phys., 2014, 140, 224310.
35 D. Keefer, S. Thallmair, J. P. P. Zauleck and R. D. Vivie- 52 C. Walter, S. Ruetzel, M. Diekmann, P. Nuernberger,
Riedle, J. Phys. B: At., Mol. Opt. Phys., 2015, 48, 234003.
T. Brixner and B. Engels, J. Chem. Phys., 2014, 140, 224311.
36 O. Ku¨hn and S. Lochbrunner, 2011, arXiv:1108.4834 53 A. Yartsev, J.-L. Alvarez, U. Åberg and V. Sundstrom, Chem.
¨
[physics].
Phys. Lett., 1995, 243, 281–289.
¨
37 D. Mobius, Adv. Mater., 1995, 7, 437–444.
54 M. Kullmann, S. Ruetzel, J. Buback, P. Nuernberger and
T. Brixner, J. Am. Chem. Soc., 2011, 133, 13074–13080.
55 Z. Wang, S. Chu, S. Wang and Q. Gong, J. Chem. Phys., 2012,
137, 164502.
56 P. Pasman, F. Rob and J. W. Verhoeven, J. Am. Chem. Soc.,
1982, 104, 5127–5133.
38 X.-Q. Li, X. Zhang, S. Ghosh and F. Wu¨rthner, Chem. – Eur. J.,
2008, 14, 8074–8078.
39 H. Marciniak, X.-Q. Li, F. Wu¨rthner and S. Lochbrunner,
J. Phys. Chem. A, 2011, 115, 648–654.
40 G. A. Lott, A. Perdomo-Ortiz, J. K. Utterback, J. R. Widom,
A. Aspuru-Guzik and A. H. Marcus, Proc. Natl. Acad. Sci. U. S. A., 57 A. V. Kulinich and A. A. Ishchenko, Russ. Chem. Rev., 2009,
2011, 108, 16521–16526.
78, 141–164.
41 M. Gruebele and A. H. Zewail, J. Chem. Phys., 1993, 98, 58 F. Koch, A. Steinbacher, C. Consani, A. Zitzler-Kunkel,
883–902.
M. Stolte, F. Wu¨rthner and T. Brixner, Phys. Chem. Chem.
Phys., 2016, 18, 19820–19831.
42 A. H. Zewail, J. Phys. Chem. A, 2000, 104, 5660–5694.
43 H. N. Chapman, P. Fromme, A. Barty, T. A. White, R. A. 59 A. Mishra, R. K. Behera, P. K. Behera, B. K. Mishra and
Kirian, A. Aquila, M. S. Hunter, J. Schulz, D. P. DePonte,
G. B. Behera, Chem. Rev., 2000, 100, 1973–2012.
U. Weierstall, R. B. Doak, F. R. N. C. Maia, A. V. Martin, 60 B. Kippelen, F. Meyers, N. Peyghambarian and S. R. Marder,
I. Schlichting, L. Lomb, N. Coppola, R. L. Shoeman,
J. Am. Chem. Soc., 1997, 119, 4559–4560.
S. W. Epp, R. Hartmann, D. Rolles, A. Rudenko, L. Foucar, 61 A. Arjona-Esteban, J. Krumrain, A. Liess, M. Stolte, L. Huang,
¨
N. Kimmel, G. Weidenspointner, P. Holl, M. Liang,
M. Barthelmess, C. Caleman, S. Boutet, M. J. Bogan,
D. Schmidt, V. Stepanenko, M. Gsanger, D. Hertel, K. Meerholz
and F. Wu¨rthner, J. Am. Chem. Soc., 2015, 137, 13524–13534.
J. Krzywinski, C. Bostedt, S. Bajt, L. Gumprecht, B. Rudek, 62 A. Arjona-Esteban, M. R. Lenze, K. Meerholz and F. Wu¨rthner,
¨
B. Erk, C. Schmidt, A. Homke, C. Reich, D. Pietschner,
in Elementary Processes in Organic Photovoltaics, ed. K. Leo,
L. Stru¨der, G. Hauser, H. Gorke, J. Ullrich, S. Herrmann,
Springer, 2017, pp. 193–214.
G. Schaller, F. Schopper, H. Soltau, K.-U. Ku¨hnel, 63 L. Huang, M. Stolte, H. Bu¨rckstu¨mmer and F. Wu¨rthner,
M. Messerschmidt, J. D. Bozek, S. P. Hau-Riege, M. Frank, Adv. Mater., 2012, 24, 5750–5754.
C. Y. Hampton, R. G. Sierra, D. Starodub, G. J. Williams, 64 A. Liess, L. Huang, A. Arjona-Esteban, A. Lv, M. Gsanger,
¨
J. Hajdu, N. Timneanu, M. M. Seibert, J. Andreasson,
V. Stepanenko, M. Stolte and F. Wu¨rthner, Adv. Funct. Mater.,
2015, 25, 44–57.
¨
A. Rocker, O. Jonsson, M. Svenda, S. Stern, K. Nass,
¨
R. Andritschke, C.-D. Schroter, F. Krasniqi, M. Bott, 65 A. Lv, M. Stolte and F. Wu¨rthner, Angew. Chem., Int. Ed.,
K. E. Schmidt, X. Wang, I. Grotjohann, J. M. Holton, 2015, 54, 10512–10515.
T. R. M. Barends, R. Neutze, S. Marchesini, R. Fromme, 66 S. Beckmann, K.-H. Etzbach, P. Kramer, K. Lukaszuk,
¨
S. Schorb, D. Rupp, M. Adolph, T. Gorkhover, I. Andersson,
H. Hirsemann, G. Potdevin, H. Graafsma, B. Nilsson and
J. C. H. Spence, Nature, 2011, 470, 73–77.
R. Matschiner, A. J. Schmidt, P. Schuhmacher, R. Sens,
G. Seybold, R. Wortmann and F. Wu¨rthner, Adv. Mater.,
1999, 11, 536–541.
Phys. Chem. Chem. Phys.
This journal is ©the Owner Societies 2017

View Article Online
PCCP
Paper
¨
67 D. N. Sathyanarayana, Electronic Absorption Spectroscopy and 72 W. Liptay, R. Wortmann, R. Bohm and N. Detzer, Chem.
Related Techniques, Universities Press, 2001. Phys., 1988, 120, 439–448.
68 E. G. McRae and M. Kasha, in Physical Processes in Radiation 73 R. Wortmann, K. Elich and W. Liptay, Chem. Phys., 1988,
Biology, ed. L. A. M. Rosenberg, Academic Press, 1964,
124, 395–409.
pp. 23–42.
74 L. Onsager, J. Am. Chem. Soc., 1936, 58, 1486–1493.
69 J. Schumann, A. Kanitz and H. Hartmann, Synthesis, 2002, 75 F. Wu¨rthner, C. Thalacker, R. Matschiner, K. Lukaszuk and
1268–1276. R. Wortmann, Chem. Commun., 1998, 1739–1740.
70 H.-B. Wang, J. A. Wisner and M. C. Jennings, Beilstein J. Org. 76 F. Wu¨rthner, S. Yao, T. Debaerdemaeker and R. Wortmann,
Chem., 2010, 6, 50. J. Am. Chem. Soc., 2002, 124, 9431–9447.
71 W. Liptay, R. Wortmann, H. Schaffrin, O. Burkhard, 77 U. Selig, P. Nuernberger, V. Dehm, V. Settels, M. Gsanger,
¨
W. Reitinger and N. Detzer, Chem. Phys., 1988, 120,
429–438.
B. Engels, F. Wu¨rthner and T. Brixner, ChemPhysChem,
2013, 14, 1413–1422.
This journal is ©the Owner Societies 2017
Phys. Chem. Chem. Phys.