Cooperative supramolecular polymerization: Comparison of different models applied on the self-assembly of bis(merocyanine) dyes

Article abstract of DOI:10.1002/chem.201202679

Three new molecular building blocks 1 a-c for supramolecular polymerization are described that feature two dipolar merocyanine dyes tethered by p-xylylene spacers. Concentration- and temperature-dependent UV/Vis spectroscopy in chloroform combined with dynamic light scattering, capillary viscosimetry and atomic force microscopy investigations were applied to elucidate the mechanistic features of the self-assembly of these strongly dipolar dyes. Our detailed studies reveal that the self-assembly is very pronounced for bis(merocyanines) 1 a,b bearing linear alkyl chains, but completely absent for bis(merocyanine) 1 c bearing sterically more bulky ethylhexyl substituents. Both temperature- and concentration-dependent UV/Vis data provide unambiguous evidence for a cooperative self-assembly process for bis(merocyanines) 1 a,b, which was analyzed in detail by the Meijer-Schenning-Van-der-Schoot model (applicable to temperature-dependent data) and by the Goldstein-Stryer model (applicable to concentration-dependent data). By combining both methods all parameters of interest to understand the self-assembly process could be derived, including in particular the nucleus size (8-10 monomeric units), the cooperativity factor (ca. 0.006), and the nucleation and elongation constants of about 10³ and 10⁶ M^-1 in chloroform at room temperature, respectively.

Full text of DOI:10.1002/chem.201202679

DOI: 10.1002/chem.201202679
Cooperative Supramolecular Polymerization: Comparison of Different
Models Applied on the Self-Assembly of Bis(merocyanine) Dyes
Gustavo Fernꢀndez, Matthias Stolte, Vladimir Stepanenko, and Frank Wꢁrthner*[a]
Abstract: Three new molecular build-
ing blocks 1a–c for supramolecular
ACHTUNGTRENNUNGpolymerization are described that fea-
1a,b bearing linear alkyl chains, but
completely absent for bis(merocyanine)
1c bearing sterically more bulky ethyl-
hexyl substituents. Both temperature-
and concentration-dependent UV/Vis
data provide unambiguous evidence for
a cooperative self-assembly process for
bis(merocyanines) 1a,b, which was ana-
lyzed in detail by the Meijer–Schen-
ning–Van-der-Schoot model (applicable
to temperature-dependent data) and by
the Goldstein–Stryer model (applicable
to concentration-dependent data). By
combining both methods all parame-
ters of interest to understand the self-
assembly process could be derived, in-
cluding in particular the nucleus size
(8–10 monomeric units), the coopera-
tivity factor (ca. 0.006), and the nuclea-
tion and elongation constants of about
103 and 106 mꢀ1 in chloroform at room
temperature, respectively.
ture two dipolar merocyanine dyes
tethered by p-xylylene spacers. Con-
centration- and temperature-dependent
UV/Vis spectroscopy in chloroform
combined with dynamic light scatter-
ing, capillary viscosimetry and atomic
force microscopy investigations were
applied to elucidate the mechanistic
features of the self-assembly of these
strongly dipolar dyes. Our detailed
studies reveal that the self-assembly is
very pronounced for bis(merocyanines)
Keywords: cooperative systems
dyes/pigments · merocyanine dyes ·
self-assembly supramolecular
polymers
·
·
AHCTUNGTRENNUNG
Introduction
to highly directional dipole–dipole interactions between
these dipolar chromophores. Subsequently, these strong and
directional intermolecular interactions could be utilized for
the construction of well-defined self-assembled architec-
tures, including discrete cyclic and bimolecular p stacks[12] as
well as extended supramolecular polymers[13] based on bis-
The spontaneous self-assembly of synthetic molecular or
macromolecular components[1] continues to be a versatile
means for the creation of a wide variety of ordered, micro-
meter-sized superstructures with well-defined geometries
that can find application in materials science, molecular
electronics or biomedicine.[2] Among them, supramolecular
polymers[3] have been extensively studied as adaptive and
stimuli-responsive materials,[4] lying at the interface between
conventional polymers and reversible non-covalent architec-
tures. To date, the self-assembly of supramolecular polymers
driven by hydrogen bonding,[5] metal coordination[6] and
host–guest interactions[7] has undergone an increasing devel-
opment. Other emerging approaches make use of aromat-
ic,[8] charge-transfer[9] or dipolar[10] interactions.
AHCTUNGTRENNUNG
(merocyanine) dyes.[14] All of those merocyanine dye assem-
blies described to date were demonstrated to grow sequen-
tially in a stepwise mechanism, in part even involving path-
ways under kinetic control.[10,15] Owing to the superior regu-
lation of self-assembly processes in natural systems by coop-
erative growth pathways as demonstrated in actin,[16]
flagellin[17] or virus assemblies,[18] there is an increasing inter-
est in the identification of cooperative self-assembly path-
ways also in the field of advanced functional supramolecular
polymers.[19–21]
More than a decade ago our group discovered that mero-
cyanine dyes form centrosymmetric dimer associates with
remarkably high binding constants (ca. 106 mꢀ1) in solvents
of low polarity.[11] This binding process could be attributed
Herein, we describe the first example of a cooperative
self-assembly process driven by dipole–dipole interactions
between merocyanines based on newly synthesized bis-
ACHUTNGREN(NUG meroCAHTUNGTRNEcNUGN yanine) dyes linked through p-xylylene benzene
spacers. For the analysis of the cooperative growth we com-
pare, and this is to our knowledge unprecedented, two dif-
ferent models: a) the nucleation–elongation growth model
of Meijer and co-workers (which is related to the one of
Oosawa and Kasai[22a]), the evaluation of which is based on
temperature-dependent data, and b) the cooperative growth
model of Goldstein and Stryer,[23] the evaluation of which is
based on concentration-dependent data. Based on carefully
acquired concentration- and temperature-dependent data
sets for bis(merocyanine) dyes 1a–c, we will show how both
[a] Dr. G. Fernꢀndez, Dr. M. Stolte, Dr. V. Stepanenko,
Prof. Dr. F. Wꢁrthner
Universitꢂt Wꢁrzburg
Institut fꢁr Organische Chemie and Center for
Nanosystems Chemistry
Am Hubland, 97074 Wꢁrzburg (Germany)
Fax : (+49)931-31-84756
E-mail: wuerthner@chemie.uni-wuerzburg.de
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202679.
206
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 206 – 217

FULL PAPER
models are interrelated and which parameters of interest
can be derived more conveniently with the different models.
Our spectroscopic analysis is supported by dynamic light
scattering (DLS), capillary viscosity and atomic force mi-
ACHTUNGTRENNUNGcrosACHTUNGTRENNUNGcopy (AFM) investigations to acquire a complete pic-
ture also for the morphologies that originate upon self-as-
sembly of bis(merocyanine) dyes 1a–c.
Results
Synthesis: Bis(merocyanines) 1a–c were synthesized from
commercially available 1,4-bis(bromomethyl)benzene and
easily accessible n-dodecyl, 2-ethylhexyl or n-octadecyl pyri-
dones as depicted in Scheme 1a. A nucleophilic substitution
reaction of 1,4-dibromomethyl benzene with 4-picoline af-
forded the corresponding bis(pyridinium) salt in quantitative
yield. A subsequent condensation reaction with the corre-
sponding n-dodecyl, n-octadecyl or 2-ethylhexyl pyridones
in the presence of N,N’-diphenylformamidine (DPFA) and
KOAc yielded the target p-bis(merocyanine) dyes 1a–c in
moderate yields. The relatively low solubility of the pyri-
dine-based precursors and bis(pyridinium) salt and compet-
ing decomposition reactions at elevated temperatures in the
reaction medium appears to be responsible for the moderate
yield. Accordingly, the highest yield was achieved for 1c,
which bears solubilizing 2-ethylhexyl substituents. Dyes
Figure 1. MALDI-TOF spectrum of 1a (298 K, matrix: DCTB).
replacement of long, linear alkyl chains by more bulky
branched alkyl groups counteracts the self-assembly. Nota-
bly, reduced binding strength for the dipolar aggregation of
related mono(merocyanines) equipped with bulky alkyl
chains has been described before.[11b] The impact of the size
of the alkyl groups in the series 1a–c on the propensity for
aggregation appears to be, however, far more pronounced.
The aggregation behaviour of 1a,b in solution was studied
by temperature-dependent UV/Vis experiments in chloro-
form. These experiments have recently been recommended
as more accurate to determine the self-assembly mechanism
and the corresponding thermodynamic parameters in com-
parison with concentration-dependent studies.[24] When
these experiments were carried out on solutions of either 1a
or 1b at a concentration of 9ꢄ10ꢀ6 m or above in chloro-
form, only partial dissociation into monomeric species was
observed upon increasing temperature (Figure S2 in the
Supporting Information). Accordingly, more dilute solutions
had to be investigated to observe the significant spectral
changes required to get insight into the self-assembly mech-
1
1a–c were characterized by H NMR, IR and UV/Vis spec-
troscopy, MALDI-TOF and HR-ESI mass spectrometry, and
elemental analysis (see Supporting Information).
Self-assembly in the gas phase and in solution: MALDI-
TOF experiments provided first evidence supporting a su-
pramolecular polymerization of 1a and 1b. Besides the mo-
lecular peak (944 and 1112 m/z for 1a and 1b, respectively),
peaks corresponding to the self-assembly of up to nine
(Figure 1) or three (Figure S1 in the Supporting Informa-
tion) monomeric units for 1a and 1b, respectively, are clear-
ly noticeable. On the other hand, no sign of self-assembly in
the gas phase was observed for 1c, which suggests that the
Scheme 1. a) Synthetic route to bis(merocyanines) 1a–c. b) Schematic representation of the self-assembly of p-xylylene-tethered bis(merocyanine) dyes
1a–c.
Chem. Eur. J. 2013, 19, 206 – 217
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
207

F. Wꢁrthner et al.
anism. Figure 2 depicts the temperature-dependent UV/Vis
experiments of 1b at a concentration of 3.1ꢄ10ꢀ6 m in
chloroform. As the temperature increases, depletion of the
aggregate band at lmax =462 nm (H-band) is accompanied
Figure 3. Concentration-dependent UV/Vis experiments (8ꢄ10ꢀ5–1ꢄ
10ꢀ7 m) of 1b (CHCl3, 298 K). Arrows indicate the spectral changes upon
dilution. M and H denote the corresponding monomeric and H-aggregat-
ed species, respectively.
Figure 2. Temperature-dependent UV/Vis spectra of 1b (CHCl3, 3.1ꢄ
10ꢀ6 m) from 278 to 338 K. Arrows indicate the spectral changes upon in-
creasing temperature. M and H denote the corresponding monomeric
and H-aggregated species, respectively, whereas D corresponds to the in-
termediate (thermodynamically disfavoured) antiparallel dimeric aggre-
gate species.
sorption band of the monomeric dyes at about 585 nm only
a small shoulder centred at about 490 nm is discernible that
is absent at lower concentration (ca. 10ꢀ7 m). According to
our previous work[11] this band can be attributed to centro-
symmetric dimer pairs of co-facially stacked merocyanines.
No sign of narrower and more hypsochromically shifted H-
band as expected for extended p stacks of these chromo-
phores can be found throughout the whole concentration
range applied. These results are clearly opposed to those ob-
served in the previously described merocyanines 1a and 1b,
equipped with longer alkyl chains, and are also in good
agreement with the absence of aggregate peaks in MALDI-
TOF experiments of 1c.
The size of the associates of 1a–c in chloroform has been
investigated by dynamic light scattering experiments (DLS)
at a concentration of about 8ꢄ10ꢀ5 m. For 1c, a very narrow
distribution of sizes ranging from 2 to 10 nm is observed
(Figure S6 in the Supporting Information), indicating that
1c is mainly present in its monomeric state. In contrast, a
variable distribution of particle sizes, ranging from 40 to
870 nm for 1a (Figure 4a) and from about 100 to 600 nm for
1b (Figure S7 in the Supporting Information) was found.
These dimensions are in accordance with previous UV/Vis
experiments and with those expected for polymeric frag-
ments of 1a,b. Interestingly, the size distribution was dem-
onstrated to be strongly dependent on the scattering angle,
which suggests that non-spherical aggregates are formed.[26]
Spherical assemblies, such as micelles or hollow vesicles, are
characterized by a size distribution independent of the scat-
tering angle. Therefore, our combined results in solution
would immediately rule out the formation of spherical spe-
cies and suggest the formation of one-dimensional aggre-
gates.
by the emergence of a new bathochromically shifted transi-
tion centred at lmax =589 nm (M-band). The appearance of
an isosbestic point at 505 nm is suggestive of a transition be-
tween two well-defined species. Accordingly, the low-tem-
perature species can be assigned as an H-aggregate formed
by self-assembly of 1b monomers, whereas the high-temper-
ature species can be assigned to free monomers. The tem-
perature-dependent spectral changes of dye 1a, equipped
with slightly shorter n-dodecyl chains, are almost identical
to those of 1b, exhibiting an aggregate band at 462 nm and
a monomeric transition at 589 nm with an isosbestic point at
504 nm (Figure S3 in the Supporting Information). Based on
previous studies in our research group on referable dyes,[13]
the spectral changes upon aggregation can be attributed to
the formation of face-to-face p stacks with H-type excitonic
coupling.[25]
The existence of a self-assembly process under thermody-
namic control for 1a and 1b has also been supported by
concentration-dependent UV/Vis experiments (8ꢄ10ꢀ5–1ꢄ
10ꢀ7 m) in chloroform at 298 K. For both dyes, the spectral
changes are reminiscent of those observed in variable-tem-
perature experiments, that is, the transition from a mono-
meric band at lmax =587 nm to a hypsochromically shifted
aggregate band (lmax =462 nm) as the concentration increas-
es (Figure 3 for 1b and Figure S4 in the Supporting Infor-
mation for 1a). Again, the presence of isosbestic points de-
notes a transition between two defined species.
The presence of self-assembled structures of large size
should give rise to viscous solutions, as observed by a large
number of supramolecular polymers assembled by different
non-covalent interactions.[13,27] Indeed, upon increasing con-
centration of 1a in chloroform from 0.18 to 0.40 mgmLꢀ1,
the values of reduced viscosity (hred) were observed to grad-
ually increase, reaching a maximum value of 185 mLgꢀ1
For comparison, the self-assembly behaviour of dye 1c, in
which the linear alkyl chains were substituted by branched
2-ethylhexyl counterparts, was investigated by UV/Vis ex-
periments in chloroform. UV/Vis studies revealed that even
at high concentrations (ca. 10ꢀ4 m) monomeric dyes prevail
(Figure S5 in the Supporting Information). Besides the ab-
208
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 206 – 217

Cooperative Supramolecular Polymerization
FULL PAPER
2855 cmꢀ1 for dye 1c, suggestive of a more coiled conforma-
tion.
The morphology of the dye assemblies formed on surfaces
has been examined by atomic force microscopy (AFM)
under different experimental conditions. A 10ꢀ4 m solution
of 1a in chloroform was spin-coated onto highly oriented
pyrolytic graphite (HOPG), mica or silicon wafer. On
HOPG, adsorption layers along with irregular structures are
observed owing to strong interactions of this particular sub-
strate with the alkyl chains of 1a,b (see Figures S9 and S10
in the Supporting Information). However, the structures ob-
served on mica or silicon wafer are identical, despite the
fact that mica is very hydrophilic (contact angle close to 08)
and silicon wafer hydrophobic (contact angle of 858), which
rules out any significant influence of the substrates on the
aggregate morphology. The images of 1a displayed ribbon-
like aggregates of several micrometers length with height
and width values of 3–6 nm and 90–150 nm for silicon wafer,
respectively, and 4.0–4.5 and 80–150 nm for mica, respective-
ly (Figure 5a–c). Notably, their outstandingly straight shape
Figure 4. a) Distribution of the hydrodynamic radii (RH) of aggregates of
1a (CHCl3, 0.08 mm) at scattering angles between 10 and 908. b) Plot of
reduced viscosity (hred
) versus concentration of 1a (CHCl3, 0.18-
0.40 mgmLꢀ1).
(Figure 4b). Although higher concentrations could not be
accomplished, these viscosity values are already in the range
of other supramolecular polymers reported in literature[13,28]
and provide strong support for the formation of extended
colloidal species in solution. Unfortunately, the slightly
lower solubility of 1b relative to 1a prevented us from de-
termining the viscosity in concentrated solutions. At concen-
trations above 0.35 mgmLꢀ1 the aggregates of 1b precipitate
out of solution with the simultaneous drop in the viscosity
values.
Self-assembly in the bulk and on surfaces: IR spectroscopy
was utilized to verify whether the linear alkyl chains play a
role in the self-assembly of bis(merocyanines) 1a and 1b, as
previously observed in systems assembled by hydrogen
bonding[29] and van der Waals forces.[30] For both dyes, the
appearance of two bands corresponding to CH2 stretching
Figure 5. AFM height images of a film prepared by drop-casting a chloro-
form solution of 1a (c=1ꢄ10ꢀ4 m) on a,b) silicon wafer and c) mica, and
d–f) 1b (c=8ꢄ10ꢀ5 m) onto mica. g) Image depicts a cross-section analy-
sis along the white line 1–1’ in image (f). The z scale is a) 15 nm,
b,c) 20 nm, d) 15 nm, e) 5 nm and f) 3 nm.
vibrations at about 2919 cmꢀ1 (nanti) and 2850 cmꢀ1
ACTHNUTRGNEN(UG nsym) (Fig-
ure S8 in the Supporting Information) is diagnostic of the in-
terdigitation of the alkyl chains and can be regarded as a
first sign of the formation of layered structures from 1a and
1b. Interestingly, the sharpness and intensity of these vibra-
tions increase with the length of the solubilizing alkyl
groups (1c!1a!1b) (Figure S8 in the Supporting Informa-
tion). Dyes 1a and 1b feature rather intense stretching
peaks at 2919 and 2850 cmꢀ1, indicative of an all-trans (out-
stretched) conformation of the alkyl chains,[31] whereas these
vibrations are less intense, broader and shifted to 2924 and
and the presence of sharp and well-defined edges suggest
that the associates might be crystalline in nature. This crys-
talline appearance becomes more obvious on silicon wafer,
as evidenced by their sharper edges and higher straightness,
symmetry and anisotropy (Figure 5a,b). A typical close-up
shot recorded on mica reveals the appearance of multiple
stratified sheets that subsequently grow into thicker fila-
ments (Figure 5c). The aggregates of 1b on mica are very
similar to those of 1a, showing a ribbon-like appearance
Chem. Eur. J. 2013, 19, 206 – 217
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
209

F. Wꢁrthner et al.
with a defined lamellar pattern (Figure 5d–f). In contrast to
1a, the lamellar distribution within the aggregates of 1b is
perpendicularly oriented to the main direction of growth
with a periodicity of 6.6 nm (Figure 5g). In sharp contrast,
dye 1c featuring more bulky and shorter branched chains
showed no ordered structures at all on these substrates.
These observations support our previous UV/Vis studies, in
which the absence of any aggregate band indicated the lack
of self-assembled structures of 1c in chloroform. The overall
findings highlight two important issues: 1) linear long alkyl
chains play a crucial role in the self-assembly events of bis-
tal data for bis(merocyanines) 1a–c as well as our earlier re-
sults on structurally related merocyanine dyes. In our earlier
work we could show that related monomeric merocyanine
building blocks[11b] as well as other dipolar merocyanines[32]
form close antiparallel dimer aggregates (D-aggregates). If
the steric features of merocyanines allow, then such dimer
aggregates are prone to form one-dimensional p-stacks in
the crystalline state.[11,32] By this means the second p face of
the dye also becomes co-facially stacked with another anti-
parallel neighbouring molecule, albeit slightly further away
or with a smaller contact surface owing to longitudinal or
transversal dislocations.[11b,32] From our earlier work we can
also relate packing arrangements to optical properties. Ac-
cordingly, a hypsochromic displacement of the absorption
band of the molecularly dissolved dye (lmax ca. 590 nm; M-
band) by about 100 nm corresponds to the formation of a
supramolecular unit of two dye molecules in antiparallel co-
facial arrangement (D-band), whilst a further hypsochromic
shift to even shorter wavelengths, that is, to lmax about 440–
460 nm (H-band), can be attributed to the further assembly
of these dimer units into extended p stacks. With this infor-
mation on optical properties of merocyanine aggregates, it
can be concluded that the assemblies formed from 1a,b in
solution, as well as on solid supports, contain extended
stacks of merocyanine building blocks with an alternated
orientation of the respective
ACHTUNGTRENNUNG(merocyanine) dyes 1a–c and 2) the appearance of the char-
acteristic absorption band of an antiparallel dimer aggregate
species centred at about 490 nm (“D-band”) in the UV/Vis
spectrum of 1c, not detectable for 1a and 1b, suggests that
this absorption band is most likely an initial stage in the
self-assembly process that develops in the case of 1a and 1b
directly into more stable H-aggregates.
Discussion
Structural model for bis(merocyanine) aggregates: In the
first part of our discussion we will derive a model for the ag-
gregate structure (Figure 6) that is based on our experimen-
merocyanine units. Further evi-
dence supporting the forma-
tion of extended dye stacks is
provided by the reduced ab-
sorption strength (hypochro-
micity) and the band narrow-
ing, which are both characteris-
tic features of H-aggregates.[33]
Accordingly, the direct trans-
formation of bis(merocyanine)
dyes 1a,b from monomer spe-
cies to H-aggregates upon in-
creasing concentration or de-
creasing temperature suggests
that an intermediate D-state
(which would relate to the for-
mation of a simple supramo-
lecular polymer) is disfav-
oured.
For dye 1c, featuring shorter
and more bulky branched
chains, such H-aggregates are
not formed in the same con-
centration range. However, the
appearance of a weak band
centred at 490 nm at the high-
est concentration in chloro-
form (1ꢄ10ꢀ4 m) indicates that
a small percentage of merocya-
nine 1c molecules are at least
assembled into D-type aggre-
Figure 6. Proposed self-assembly model for bis(merocyanine) dyes 1a,b with denotation of the nucleation and
the elongation regime. The nucleation regime is related to the intermolecular interaction between two mero-
cyanine dyes, which self-assemble into an antiparallel dimer pair. In contrast, the elongation regime cannot be
related to a single intermolecular interaction, but arises from the combined forces of merocyanine–merocya-
nine and alkyl–alkyl chain interactions.
210
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 206 – 217

Cooperative Supramolecular Polymerization
FULL PAPER
gates. Apparently, the replacement of long linear chains by
small branched ones provokes two major effects: on one
hand, the short branched and even racemic chains cannot
support the growth of structurally defined aggregates by in-
termolecular van der Waals contacts between neighbouring
chains[34] and, on the other hand, an enhanced steric hin-
drance weakens the strength of the non-covalent interac-
tions between the dipolar p surfaces of the dyes.[11b] Both ef-
fects are presumably much more pronounced for the further
growth into extended dye stacks than for the first initial
dimer aggregate formation step, for which steric encum-
brance can be more easily circumvented by a displacement
of the molecules with regard to each other.
Molecular modelling and semi-empirical (AM1) calcula-
tions support the arrangement of 1a,b into supramolecular
lamellae with extended H-aggregated p-stacks as depicted
in Figure 6. As expected, the most stable monomeric confor-
mation keeps the dipolar merocyanines at the largest possi-
ble distance to minimize their electrostatic repulsion
(Figure 6, top left and Figure S11a). Significant stabilization
arises upon co-facial stacking of two dipolar dye units of
neighbouring molecules in a perfectly antiparallel fashion.
The modelling of a set of four molecules (involving eight
dyes), including their long alkyl chains, supports the forma-
tion of a supramolecular chain formed by the pairing of di-
polar dyes into dimeric stacks with their peripheral alkyl
chains nearly perpendicular to the plane formed by the dye
arrangements (Figure 6, top right and Figure S11b in the
Supporting Information). These one-dimensional strands of
dimerized dyes, however, cannot account for the formation
of strongly hypsochromically shifted H-bands as observed in
UV/Vis experiments, because the arrangement shown in
Figure 6 (top right) would give rise to a less displaced D-
band according to our previous studies.[13] However, further
assembly of these one-dimensional strands into two-dimen-
sional sheets (lamellae) as shown in Figure 6 (bottom; see
also Figure S11c in the Supporting Information) is easily fea-
sible and supported by our AFM studies.
The AFM images on silicon wafer or mica indeed display
ribbon-like aggregates with a well-defined lamellar pattern
of several micrometers length with identical morphology
and dimensions. The arrangement of the lamellae within the
ribbon-like aggregates is dependent on the peripheral alkyl
chains. Bis(merocyanine) 1a, equipped with dodecyl chains,
features a lamellar arrangement that is oriented along the
longitudinal axis of the ribbon-like associates, whereas the
lamellar pattern of dye 1b, featuring longer octadecyl
chains, is perpendicular to the main direction of growth (see
Figure 5). Although dipolar interactions between merocya-
nine subunits are the strongest non-covalent forces, the pres-
ence of longer octadecyl chains in 1b will enhance van der
Waals interactions between alkyl chains through interdigita-
tion,[34] which appears to outweigh the dipolar interactions
between merocyanine chromophores when a critical aggre-
gate size has been formed. The fact that the lamellae (which
are formed through dipolar interactions) are oriented per-
pendicularly to the main direction of growth of the aggre-
gates points out that van der Waals forces between lamellae
become more relevant for 1b than in the case of 1a.
Evaluation of self-assembly mechanism from temperature-
dependent data: As a convincing structural model for the
self-assemblies of dyes 1a,b has now been given, the second
part of the discussion will carve out the self-assembly mech-
anism from the data sets of the temperature- and concentra-
tion-dependent UV/Vis spectra. The temperature-dependent
UV/Vis studies of 1a,b in chloroform revealed a self-assem-
bly process under thermodynamic equilibrium when the
temperature is lowered. Only two species are discernible
upon temperature variation: 1) monomeric dyes (M-band,
centred at 589 nm) and 2) H-type aggregates (centred at
462 nm) and the transition from the former to the latter
takes place instantaneously. This behaviour is in contrast to
what has previously been observed for all other bis(mero-
cyanine) dyes investigated in our group. In all earlier exam-
ples, the formation of either an intermediate or final species
with an absorption band centred at about 485–490 nm was
observed in the course of the self-assembly process. This
band was assigned to the excitonic coupling of two adjacent
dyes that are aggregated in an antiparallel fashion (D-aggre-
gate). This process is evident either in the formation of cy-
clic[12a] or random coil polymeric architectures.[13a,b] In the
latter, further contact between folded polymeric chains (D-
state) leads to helical nanorods in which additional excitonic
coupling to other dye molecules that are located in close
proximity to the dimeric unit afford a further hypsochromic
shift leading to H-aggregates (lmax ca. 460 nm).[13a,b] The de-
pletion of the D-aggregate maximum and the appearance of
a new hypsochromically shifted transition in the UV/Vis
spectrum (462 nm) accounts for the formation of extended
supramolecular architectures formed upon interaction of
random-coil supramolecular polymers (spectroscopically
characterized by D-bands). In the temperature-dependent
UV/Vis experiments performed for 1a,b, the absence of the
transition at 490 nm clearly points out that the formation of
the D-aggregated species is not favoured and that more ex-
tended H-aggregates form right from the start of the self-as-
sembly process.
To gain further insight into the self-assembly mechanism,
we investigated in detail the spectral changes observed in
our experiments. To that end, solutions of 1a or 1b in
chloroform were slowly cooled down from 338 to 278 K in
small steps (1 Kminꢀ1) to ensure that self-assembly takes
place under thermodynamic control. For this cooling rate no
hysteresis in the molar absorption coefficient was observed
when heating the solution again to 338 K, which indicates
that the self-assembly process is reversible. Slower cooling
rates (0.5 Kminꢀ1) did not influence the cooling curves. As-
suming the formation of monodimensional non-cyclic supra-
molecular polymers through dipolar aggregation of mero-
cyanines, the simplest model is the isodesmic or equal-K
model.[3b,8b] This model assumes that all equilibrium con-
stants or Gibbs free energy changes are equal for every
binding event. Mathematically, applying this model to a cer-
Chem. Eur. J. 2013, 19, 206 – 217
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
211

F. Wꢁrthner et al.
tain temperature-dependent experiment, the fraction of ag-
gregated species (aagg) can by expressed by Equation (1)[24]
in which DH is the enthalpy release, R is the ideal gas con-
stant and Tm is the melting temperature, that is, the temper-
ature for aagg(T)=0.5.
For the analysis of the cooperativity, we have made use of
the nucleation–elongation model of Meijer, Schenning and
Van der Schoot.[22b,24] This model is related (but not identi-
cal) to the model of Oosawa and Kasai that was originally
developed for the helical assembly of proteins in solACHTUNGTRENNUNGu-
tion.[22a] Recently, this model has been successfully exploited
for the elucidation of cooperative self-assembly processes of
p-conjugated oligomers[20a] and C3-symmetrical discotic tri-
1
aaggðTÞ ffi
ð1Þ
ꢀ0:908DHTꢀT
m
RTm2
1 þ e
AHCTUNGTRENNUNG
amides[20b] by Meijer, Schenning and co-workers and for the
formation of p stacks composed of squaraine dyes by our
group.[35] Meanwhile, even more elaborate models have
been developed by Markvoort et al.[36] and ten Eikelder
et al.[37] for the analysis of cooperative supramolecular
The value for aagg can be easily extracted from the experi-
mental absorption spectra by applying Equation (2) in
which emon and eagg are the molar absorption coefficients for
pure monomeric and pure aggregate species and e(T) is the
apparent absorption coefficient at the given temperature.
AHCTUNGTRENNUNG
aagg values according to Equation (2) for 1b in a tempera-
ture range from 278 to 338 K and at a concentration of 3.1ꢄ
10ꢀ6 m. The plot features two different temperature regimes:
the nucleation (blue fit) and the elongation (red fit) regime,
separated by the elongation temperature. At the highest
temperature the system is exclusively present in its mono-
meric state (aagg ca. 0). Subsequently, within a very narrow
temperature range around 330 K the system “activates” to-
wards the formation of higher H-type aggregates (elonga-
tion regime).
In the elongation regime, the fraction of aggregated spe-
cies (aagg) can be defined by Equation (3) in which DHe is
the enthalpy corresponding to the aggregation (elongation)
process, T the absolute temperature, Te the elongation tem-
perature, R the ideal gas constant and aSAT is a parameter
introduced to ensure that aagg/aSAT does not exceed unity.
eðTÞ ꢀ eagg
ð2Þ
aaggðTÞ ¼ 1 ꢀ
emon ꢀ eagg
By applying these equations to the spectral changes of
1a,b, we noticed remarkable discrepancies between our ex-
perimental data at different wavelengths (461, 589 nm) and
the theoretical curve predicted by the isodesmic model.
Rather, we observed a more distinct transition from mono-
mers towards aggregates in comparison with the isodesmic
plot, which suggests a cooperative process as described by
the nucleation–elongation mechanism.[24] This mechanism
involves a cooperative chain growth (elongation) that can
only take place after an initial critical-sized aggregated spe-
cies has been formed (nucleation). Once these species
(nuclei) have overcome an initial energetic barrier the
ACHTUNGTRENNUNGpropagation of the polymer chain dramatically increases in a
ꢂ
ꢀ
ꢁꢃ
cooperative manner. As a result, the sigmoidal curve corre-
sponding to the plot of fraction of aggregated species
against temperature or concentration reveals a critical point
that does not exist in the isodesmic model and which bears
resemblance to a phase transition (Figure 7). As expected
for a well-differentiated two-state cooperative mechanism,
the equilibrium constant of the nucleation steps (Knuc) are
smaller compared to those of the highly favoured elongation
steps (K).
ꢀDHe
aagg ¼ aSAT 1 ꢀ exp
ðT ꢀ TeÞ
ð3Þ
RTe2
By using Equation (3) at temperatures below the elonga-
tion temperature (T<Te) we could accurately fit the experi-
mental data to the elongation regime (Figure 7, red line)
with an R2 =0.9991. At the given concentration of 3.1ꢄ
10ꢀ6 m the enthalpy release during the self-assembly of 1b in
the elongation process (DHe) and the elongation tempera-
ture (Te) were calculated to be ꢀ55.9 kJmolꢀ1 and 329.9 K,
respectively (see Table 1).
Table 1. Thermodynamic parameters Ka, Te, DHe and aSAT obtained from
the temperature-dependent UV/Vis experiments of 1a and 1b in chloro-
form on the basis of the modified nucleation-growth model.[22b]
c [m][a]
aSAT
Te
[K]
DHe
[kJmolꢀ1
]
Ka
hNn(Te)i
U
1a
1b
2.8ꢄ10ꢀ6
3.1ꢄ10ꢀ6
1.003
1.051
329.6
329.9
ꢀ62.8
ꢀ55.9
0.00122
0.00146
8–10
8–10
[a] c=concentration.
Figure 7. Fitting of aagg data against T in the framework of the nuclea-
tion–elongation model. The blue line corresponds to the nucleation
regime and the red line to the elongation regime. Black squares represent
the experimental data of 1b obtained in temperature-dependent experi-
ments at a wavelength of 589 nm.
On the other hand, in the nucleation regime the fraction
of aggregated species (aagg) can be defined by Equation (4)
in which Ka is the dimensionless equilibrium constant of the
activation step at the elongation temperature.
212
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 206 – 217

Cooperative Supramolecular Polymerization
FULL PAPER
ꢀ
ꢁ
DH
RTa2
Evaluation of self-assembly mechanism from concentration-
dependent data: Our detailed investigation of the coopera-
tive self-assembly process of 1a,b has been complemented
by an analysis of the concentration-dependent UV/Vis data
at room temperature. As expected, the isodesmic or equal-K
model[8b] is unable to describe the aggregation process as
evidenced by an abrupt decrease of the apparent absorption
at 587 nm upon reaching a certain critical concentration
(Figure 8). In contrast, application of the general coopera-
aagg ¼ Ka1=3 exp ð2=3ꢀa 1=3 ꢀ 1Þ
ðT ꢀ TaÞ
ð4Þ
By using this expression at temperatures above the elon-
gation temperature (T>Te) at the given concentration of
3.1ꢄ10ꢀ6 m, a Ka value of about 0.00146 was calculated (Fig-
ure S12b in the Supporting Information). The low value of
this dimensionless constant at the nucleation regime is sug-
gestive of a high degree of cooperativity, and has the same
role as the cooperativity parameter s, which can be calculat-
ed by concentration-dependent experiments, as will be
shown later. In the elongation regime the number-averaged
degree of polymerization, averaged over all active species
hNni can be described by Equation (5).
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
aagg
1
Ka
pffiffiffiffiffiffi
hNni ¼
ð5Þ
aSAT ꢀ aagg
By introducing the value of Ka obtained in the nucleation
and the parameters aSAT and aagg from the elongation re-
gimes, respectively, the number-averaged degree of polymer-
ization hNni can be calculated and plotted at different tem-
peratures, according to Equation (5). At room temperature,
the length of the associates of 1b at the given concentration
has been determined to be about 200 molecules (Figure S13
in the Supporting Information). Taking into consideration
that the length of a dimeric unit within the stack obtained
by molecular modelling is approximately 4.1 nm, the calcu-
lated length of the stack would be 410 nm, which is in very
good agreement with DLS results.
Figure 8. Plot of fraction of aggregated molecules against KcT for s
values of 0.006 and a nucleus size (s) of 5, 8 and 15 according to the gen-
eral nucleation-elongation mechanism. Black circles represent the experi-
mental data of 1b obtained in concentration-dependent experiments at a
wavelength of 587 nm. The best fit is obtained for s=8.
tive model of Goldstein and Stryer[23] afforded a good fit.
According to this model a nucleus of variable size s is
formed following the isodesmic model with first equilibrium
constants Ks up to the formation of the nucleus. With the
size of the supramolecular polymer reaching s, further iso-
desmic association steps show larger equilibrium constants
of K. The mass balance equation for a nucleation-elongation
supramolecular polymerization process can be expressed as
in Equation (7) giving a relation between the equilibrium
monomer concentration c1 and the total concentration of
molecules cT.
The average length of the stack hNni averaged over the
nucleated species at the Te is given by Equation (6).
1
Ka1=3
hNnðTeÞi ¼
ð6Þ
The substitution of Ka in this equation enables the calcu-
lation of the number of aggregated molecules, that is, the
nucleus size, at the elongation temperature. By using Equa-
tion (6) the number of aggregated molecules in the nucleus
was calculated to be between 8 and 10, which indicates that
a relatively large nucleus needs to be formed prior to the
elongation process.
s
1
X
X
n
n
KcT ¼
nsnꢀ1ðKc1Þ þ
nssꢀ1ðKc1Þ ¼
n¼1
n¼sþ1
sðKc1Þ ssꢀ1 ðKc1Þsþ1ssꢀ1 Kc1ðsðsKc1Þ ꢀ 1Þ
s
sꢀ1
For comparison, we also investigated in detail the self-
ACHTUNGTRENNUNGassembly mechanism of merocyanine 1a in chloroform at a
ð7Þ
þ
þ
2
1 ꢀ Kc1
sKc1 ꢀ 1
ð1 ꢀ Kc1Þ
slightly lower concentration of 2.8ꢄ10ꢀ6 m. Similarly to 1b,
the spectral changes could also be accurately fitted to the
modified nucleation–elongation mechanism (Figures S3b,
S4b, 12a, S13a in the Supporting Information) affording
values of DHe and Te of ꢀ62.8 kJmolꢀ1 and 329.6 K, respec-
tively (Table 1). The similar degree of cooperativity and
thermodynamic parameters for the self-assembly of both 1a
and 1b demonstrates that once reached a certain length, the
extension of the linear alkyl chains has little influence on
the self-assembly process.
2
sꢀ1
sðKc1Þ ððsKc1Þ ꢀ 1Þ
ꢀ
2
ðsKc1 ꢀ 1Þ
In Equation (7) s reflects the parameter for the coopera-
tivity defined as Ks/K, and s is the size of the nucleus. Since
it is not possible to determine s alone, but only in combina-
tion with s, accumulative cooperativity was defined as w=
s
sꢀ1. The fraction of aggregated molecules aagg can be calcu-
lated from the relation of KcT (cT =total concentration) and
Kc1 (c1 =concentration of monomers) by using Equation (7)
in combination with Equation (8)
Chem. Eur. J. 2013, 19, 206 – 217
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
213

F. Wꢁrthner et al.
Kc1
KcT
rectly accessible by the nucleation-growth model, and this is
the cooperativity factor s provided by the Goldstein–Stryer
model and its application on concentration-dependent data
points. This parameter is much more descriptive than the
more abstract parameter Ka of the nucleation–elongation
model. Furthermore, from this parameter the binding con-
stant Ks for the nucleation regime can be simply calculated
by Ks =Kꢄs from the larger binding constant K of the elon-
gation regime. By application of the mathematically more
complex Goldstein–Stryer model instead of the hitherto
more commonly applied simpler K2-K model[21,38] to concen-
tration-dependent self-assembly experiments, there is a sig-
nificant gain in information and the possibility for a most
satisfactory merger of the temperature- and concentration-
dependent results to afford all parameters of interest: coop-
erativity parameter s, nucleation constant Ks, elongation
constant K, and nucleus size s at a particular temperature as
well as elongation temperature Te and enthalpy DHe at a
particular concentration.
ð8Þ
aagg ¼ 1 ꢀ amon ¼ 1 ꢀ
The experimental data points for 1b extracted from con-
centration-dependent experiments at 587 nm were fitted
manually to this general nucleation–elongation model for
the best match. The best fits were obtained for a nucleus
size (s) of eight molecules and a degree of cooperativity (s)
of 0.006 (Figure 8). The values of nucleation (Ks) and elon-
gation (K) constants were calculated to be about 9ꢄ103 and
1.6ꢄ106 mꢀ1, respectively (Table 2). These results support the
Table 2. Thermodynamic parameters s, s, Ks and K obtained from con-
centration-dependent UV/Vis experiments of 1a and 1b in chloroform
on the basis of the Goldstein–Stryer model.[23]
T [K]
s
s
Ks [mꢀ1
]
K [mꢀ1
]
1a
1b
298
298
8
8
0.006
0.007
9ꢄ103
1.6ꢄ106
9.5ꢄ103
1.65ꢄ106
earlier conclusion obtained from temperature-dependent ex-
periments that a relatively large nucleus needs to be formed
for the initiation of the supramolecular polymerization pro-
Conclusion
ACHTUNGTRENNUNGcess. On the other hand, the remarkably increased elonga-
tion constant (between two and three orders of magnitude
larger than the initial aggregation constants during nuclea-
tion) suggests the formation of a well-organized supramolec-
ular architecture in which each dipolar dye is surrounded by
two antiparallel neighbour dyes by close face-to-face p con-
tacts and likewise each alkyl chain realized favourable van
der Waals interactions to those of the neighbour molecules.
The changes observed in concentration-dependent experi-
ments in chloroform solution of 1a were also manually
fitted to the general cooperative model of Goldstein and
Stryer[23] (Figure S4b in the Supporting Information). In line
with temperature-dependent experiments, the best fits were
obtained when s is about 0.007 and the nucleus size (s) cor-
responds to eight molecules. Accordingly, both nucleation
(Ks; ca. 9.5ꢄ103 mꢀ1) and elongation (K; ca. 1.65ꢄ106 mꢀ1)
constants are very similar to those obtained for 1b
(Table 2).
The comparative evaluation of concentration- and tem-
perature-dependent UV/Vis data for two slightly different
bis(merocyanine) dyes containing side chains of different
length (C12H25 vs. C18H37) afforded accordingly a consistent
result. Both models were suitable to identify the existence
of a non-isodesmic self-assembly and to quantify the in-
crease of the binding strength after formation of the nucleus,
that is, the cooperativity. The evaluation by the tempera-
ture-dependent nucleation–elongation model of Meijer,
Schenning and Van der Schoot has, as pointed out by these
authors,[24] the advantage of easier data acquisition as well
as of a simpler mathematical treatment for the extraction of
the parameters of interest, including the nucleus size. This
model enables the calculation of the enthalpy release during
the aggregate growth (DHe), the elongation temperature
(Te) and the nucleus size hNni. There is, however, one pa-
rameter that we appreciate very much and which is not di-
In conclusion, we have described three bis(merocyanine)
dyes 1a–c spaced by p-xylylene units and equipped with
alkyl chains of different length and nature. A thorough col-
lection of spectroscopic, microscopic and other experiments
showed that 1c, which bears short and branched alkyl
chains, cannot self-assemble whilst 1a,b, which bear long n-
alkyl chains, readily self-assemble in the gas phase, in solu-
tion, or in the solid state, giving rise to extended H-aggre-
gated species of micrometric dimension. The simultaneous
interplay of dipolar interactions between dipolar merocya-
nine dyes and van der Waals forces between long alkyl
chains enables a two-dimensional growth into lamellar as-
semblies which further grow into ribbons. Most importantly,
we could reveal that the supramolecular polymers of
bis(merocyanines) 1a,b involve
a
highly cooperative
self-assembly process. In a first (nucleation) step, up to
8–10 monomeric units are required to interact through dipo-
lar forces to “activate” the self-assembly process. At this
point, further lateral interactions between the peripheral
alkyl chains of different strands become relevant to facilitate
additional interstrand interactions, thereby activating the su-
pramolecular polymerization process. For the first time, the
nucleation–elongation model for the evaluation of tempera-
ture-dependent data introduced recently by Meijer, Schen-
ning and Van der Schoot was linked with the Goldstein–
Stryer model applicable to concentration-dependent data.
This combined approach provides access to the nucleus size
and the cooperativity parameter, supplemented by the ag-
gregation constants for the nucleation and elongation proc-
esses. Because these parameters are of considerable interest
to quantify self-assembly proACTHNUGTRNENUGcessCAHTUNGTRENNeUGN s, such a combined ap-
proach should be of interest to all scientists looking for
guidelines towards a rational design of highly-organized su-
pramolecular systems.
214
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 206 – 217

Cooperative Supramolecular Polymerization
FULL PAPER
3H), 1.45 (brt, 2H), 1.12–1.25 (m, 30H), 0.72 ppm (brt, 3H); ESI-MS:
Experimental Section
m/z calcd for C25H42N2O2: 402.32 [M+]; found: 403.33 [M++H].
Synthesis of 1,4-bispyridinium salt: A solution of 1,4-dibromomethylben-
zene (5 g, 14 mmol) and 4-picoline (8.18 mL, 84.2 mmol) in MeCN
(100 mL) was heated to reflux for 15 h. After allowing the mixture to
reach room temperature the resulting precipitate was filtered through a
glass frit, washed with MeCN, and dried under vacuum, affording the 1,4-
bispyridinium salt (8 g, 90%) as a white solid that can be subsequently
used without further purification. 1H NMR (MeOD, 400 MHz): d=2.70
(s, 6H), 5.89 (s, 4H), 7.65 (s, 4H), 8.47 (dd, J=6.8 Hz, 8H); ESI-MS: m/z
calcd for C20H22Br2N2 =448.01 [M+]; found: 368.13 [M+ꢀBr], 288.23 [M+
ꢀ2Br].
General procedure for the synthesis of bis(merocyanine) dyes: A suspen-
sion of the corresponding pyridone (3 mmol) and N-N’-diphenylformami-
dine (3 mmol) in Ac2O (3 mL) was stirred at room temperature for
30 min. The mixture was heated to 908C for 1 h to complete the reaction.
After the reaction mixture was cooled down to room temperature, the
corresponding bis(pyridinium) salt (1 mmol) and KOAc (3 mmol) were
added and the mixture was heated to reflux overnight. The solvent was
evaporated to dryness, and the crude product was purified by column
chromatography on silica gel, using mixtures chloroform or dichlorome-
thane/methanol as eluent.
General: Solvents and reagents were used as purchased. Column chroma-
tography was performed on silica gel (Merck Silica 60, particle size 0.04–
0.063 nm) and thin layer chromatography (TLC) was conducted on silica
gel plates (60 F254 Merck, Darmstadt). Melting points were determined
on Bꢁchi Melting Point B545 heating stage and are uncorrected. Solvents
for UV/Vis absorption and AFM studies were of spectroscopic grade and
used as received. 1H NMR spectra were recorded on a Bruker Avance
400 spectrometer at 298 K using partially deuterated solvents as internal
standard. Coupling constants (J) are denoted in Hz and chemical shifts
(d) in ppm. Multiplicities are denoted as follows: s=singlet, d=doublet,
t=triplet, m=multiplet, br=broad. High resolution ESI-TOF mass spec-
trometry was carried out on a microTOF focus instrument (Bruker Dal-
tronik GmbH) in positive mode with CH3CN or CHCl3 as solvent, unless
otherwise stated. UV/Vis absorption spectra were recorded on a Perkin–
Elmer Lambda 950 UV/Vis spectrophotometer with a spectral bandwidth
of 0.20 nm using conventional quartz cells of 1–50 mm path length to
cover a suitable concentration range. Temperature was controlled by a
PTP-1 Peltier element (Perkin–Elmer). MALDI-TOF (matrix assisted
laser desorption ionization (coupled to a time-of-flight analyzer)) experi-
ments were recorded on a Bruker REFLEX spectrometer. Viscosity
Measurements were recorded at 258C with a capillary viscosimeter using
the automated viscosity measuring device AVS 360 (Schott Gerꢂte
GmbH) in order to obtain reproducible run times. The effective capillary
diameter was 0.64 mm. The setup was mounted in a thermostat, which
was controlled by circulating water. The measurements were performed
from concentrated to diluted solutions. Dilution was achieved by using
an automated Titronic universal unit (Schott Gerꢂte GmbH). Dynamic
light scattering (DLS) measurements were carried out at 258C on an
ALV CGS-3 goniometer using a HeNe laser (l=632.8 nm) and an ALV
LSE-5004 correlator. Sample solutions were filtered through a 0.45 mm
hydrophobic PTFE filter. AFM measurements were carried out at ambi-
ent conditions by using a Bruker AXS system operating in tapping mode.
Silicon cantilevers (Olympus, OMCLAC160TS) with a resonance fre-
quency of about 300 kHz and spring constant of about 42 Nmꢀ1 were
used. The 512ꢄ512 pixel images were collected at a rate of two scan lines
per second. Large scale images were recorded at a scan rate of 1 Hz. Sol-
utions of bis(merocyanine) dyes 1a,b in chloroform were either spin-
coated or drop-cast onto mica (Plano GmbH, Germany) or silicon wafer.
Data for merocyanine 1a: Dark red solid; m.p.>2008C decomp; eluent:
CH2Cl2/MeOH (8:1); yield: 35%; 1H NMR (CD2Cl2/MeOD (2:1),
400 MHz): d=0.89 (brt, 6H), 1.28 (40H, m), 1.59 (brt, 4H), 2.47 (6H, s),
3.95 (t, J=7.5 Hz, 4H), 5.48 (s, 4H), 7. 48 (s, 4H), 7.78 (m, 4H),
7.90 ppm (dd, J=8 Hz, 8H); IR (neat): n˜ =776, 871, 945, 1034, 1055,
1159, 1198, 1219, 1321, 1464, 1500, 1523, 1564, 1643, 2195, 2850,
2919 cmꢀ1
; HRMS-ESI (pos. mode, CH3CN/CHCl3 =1:1 vol%): m/z
calcd for C60H76N6NaO4 [M+Na]+: 967.5822; found: 967.5818; elemental
analysis calcd (%) for C60H76N6O4: C 76.2, H 8.1, N 8.9; found: C 75.9, H
8.1, N 9.3.
Data for merocyanine 1b: Dark red solid; m.p.>2008C decomp; eluent:
chlorofom/MeOH (8:1); yield: 16%; 1H NMR (CDCl3/MeOD (2:1),
400 MHz): d=0.73 (brt, 6H), 1.10 (m, 64H), 1.45 (m, 4H), 2.31 (s, 6H),
3.81 (t, J=7.3 Hz, 4H), 5.34 (s, 4H), 7. 34 (s, 4H), 7.61 (m, 4H), 7.76
(dd, J=7.2 Hz, 8H); IR (neat): n˜ =721, 778, 949, 1035, 1057, 1160, 1297,
1335, 1469, 1495, 1523, 1565, 1588, 1641, 2194, 2850, 2919 cmꢀ1; HRMS-
ESI (pos. mode, CH3CN/CHCl3 =1:1 vol%): m/z calcd of C72H101N6O4
[M++H]: 1113.7884; found 1113.7872; elemental analysis calcd (%) for
C72H100N6O4: C 77.6, H 9.0, N 7.5; found: C 77.2, H 9.3, N 7.6.
Synthetic details and characterization: Dodecyl- and 2-ethylhexyl-substi-
tuted pyridones (Scheme 1, second reaction step) were prepared accord-
ing to previously reported synthetic procedures (see reference [11b]) and
showed identical spectroscopic properties to those reported therein.
Data for merocyanine 1c: Dark red solid; m.p.>2008C decomp; eluent:
chlorofom/MeOH (10:1); yield: 36%; 1H NMR (CDCl3/MeOD (3:1),
400 MHz): d=0.70 (m, 12H), 1.10 (m, 16H), 1.71 (m, 2H), 2.30 (s, 6H),
3.76 (m, 4H), 5.31 (s, 4H), 7. 34 (s, 4H), 7.32 (s, 4H), 7.71 ppm (dd, J=
7.0 Hz, 8H); IR (neat): n˜ =728, 773, 861, 945, 1036, 1054, 1152, 1195,
1220, 1289, 1326, 1374, 1395, 1462, 1495, 1521, 1560, 1593, 1643, 2192,
2854, 2924, 2957, 3062 cmꢀ1; HRMS-ESI (pos. mode, CH3CN/CHCl3 =1:1
vol%): m/z calcd for C52H60N6NaO4 [M++Na]: 855.4554; found 855.4565;
elemental analysis calcd (%) for C52H60N6O4: C 75.0, H 7.3, N 10.1;
found: C 74.5, H 7.3, N 9.9.
Synthesis of 2-cyano-N-octadecylacetamide: Octadecylamine (10 g,
37 mmol) and ethyl cyanoacetate (25 mL, 241 mmol) were mixed in a
250 mL round-bottomed flask. The resulting suspension was slightly
heated to 508C for 15 min until the mixture was completely dissolved.
Upon stirring the solution at room temperature for 30 min a white solid
crashed out. After that, the solidified mixture was stirred overnight at
room temperature, and the solid was filtered under vacuum and washed
gently with EtOH and Et2O. The crude product was recrystallized from
EtOH, affording 2-cyano-N-octadecylacetamide (9 g, 72%) as a white
solid. 1H NMR (CDCl3, 300 MHz): d=3.29 (s, 2H), 3.05 (t, 2H), 1.31
(brt, 2H), 1.05–1.1 (m, 30H), 0.65 ppm (brt, 3H); ESI-MS: m/z calcd for
C21H40N2O=336.55 [M+]; found: 337.52 [M++H].
Acknowledgements
We thank the Alexander-von-Humboldt foundation for generous support
of our research (postdoctoral fellowship and subsequent Sofja-Kovalev-
skaja award for G.F.).
Syntheis of 1,2-dihydro-6-hydroxy-4-methyl-1-octadecyl-2-oxopyridine-3-
carbonitrile:
A
mixture of 2-cyano-N-octadecylacetamide (5 g,
14.9 mmol), ethylacetoacetate (1.93 g, 14.9 mmol) and piperidine (12 mL)
was heated to reflux at 1008C for 20 h. The resulting brown solution was
allowed to cool down to room temperature and HCl 32% was carefully
added dropwise until pH 1. The resulting white precipitate was filtered
through a glass frit, washed with water and dried under vacuum. The
crude residue was purified by column chromatography (silica gel,
CH2Cl2/MeOH 10:1) affording 1,2-dihydro-6-hydroxy-4-methyl-1-octa-
decyl-2-oxopyridine-3-carbonitrile (3.2 g, 60%) as a light brown solid.
1H NMR ([D6]DMSO, 400 MHz): d=5.38 (s, 1H), 3.12 (brt, 2H), 2.05 (s,
[1] a) J.-M. Lehn, Supramolecular Chemistry: Concepts and Perspectives,
VCH, Weinheim, Germany, 1995; b) L. J. Prins, D. N. Reinhoudt, P.
Timmerman, Angew. Chem. 2001, 113, 2446–2492; Angew. Chem.
Int. Ed. 2001, 40, 2382–2426.
[2] a) B. M. Rosen, C. J. Wilson, D. A. Wilson, M. Peterca, M. R. Imam,
V. Percec, Chem. Rev. 2009, 109, 6275–6540; b) H. Cui, M. J.
Webber, S. I. Stupp, Biopolymers 2010, 94, 1–18; c) H. N. Tsao, K.
Chem. Eur. J. 2013, 19, 206 – 217
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
215

F. Wꢁrthner et al.
Mꢁllen, Chem. Soc. Rev. 2010, 39, 2372–2386; d) G. R. Whittell,
M. D. Hager, U. S. Schubert, I. Manners, Nat. Mater. 2011, 10, 176–
188; e) T. Aida, E. W. Meijer, S. I. Stupp, Science 2012, 335, 813–
817.
[15] For other recent work on self-assembly pathways under kinetic con-
trol, see: a) A. Ajayaghosh, R. Varghese, V. K. Praveen, S. Mahesh,
Angew. Chem. 2006, 118, 3339–3342; Angew. Chem. Int. Ed. 2006,
45, 3261–3264; b) E. T. Pashuck, S. I. Stupp, J. Am. Chem. Soc.
2010, 132, 8819–8821; c) V. Percec, S. D. Hudson, M. Peterca, P.
Leowanawat, E. Aqad, R. Graf, H. W. Spiess, X. Zeng, G. Ungar,
P. A. Heiney, J. Am. Chem. Soc. 2011, 133, 18479–18494; d) Y.
Tidhar, H. Weissman, S. G. Wolf, A. Gulino, B. Rybtchinski, Chem.
Eur. J. 2011, 17, 6068–6075; e) I. Danila, F. Pop, C. Escudero, L. N.
Feldborg, J. Puigmartı-Luis, F. Riobe, N. Avarvari, D. B. Amabilino,
Chem. Commun. 2012, 48, 4552–4554; f) P. A. Korevaar, S. J.
George, A. J. Markvoort, M. M. J. Smulders, P. A. J. Hilbers,
A. P. H. J. Schenning, T. F. A. De Greef, E. W. Meijer, Nature 2012,
481, 492–496; g) K. Liu, Y. Yao, C. Wang, Y. Liu, Z. Li, X. Zhang,
Chem. Eur. J. 2012, 18, 8622–8628.
[16] a) A. Wegner, Nature 1982, 296, 266–267; b) E. D. Korn, M.-F. Car-
lier, D. Pantaloni, Science 1987, 238, 638–644; c) M.-F. Carlier, J.
Biol. Chem. 1991, 266, 1–4.
[17] D. Abram, H. Koffler, J. Mol. Biol. 1964, 9, 168–185.
[18] a) A. Klug, Angew. Chem. 1983, 95, 579–596; Angew. Chem. Int.
Ed. Engl. 1983, 22, 565–582; b) K. van Workum, J. F. Douglas, Mac-
romol. Symp. 2005, 227, 1–16.
[19] a) V. Percec, G. Ungar, M. Peterca, Science 2006, 313, 55–56;
b) C. A. Hunter, H. L. Anderson, Angew. Chem. 2009, 121, 7624–
7636; Angew. Chem. Int. Ed. 2009, 48, 7488–7499; c) G. Ercolani, L.
Schiaffino, Angew. Chem. 2011, 123, 1800–1807; Angew. Chem. Int.
Ed. 2011, 50, 1762–1768.
[20] a) P. Jonkheijm, P. Van der Schoot, A. P. H. J. Schenning, E. W.
Meijer, Science 2006, 313, 80–83; b) M. M. J. Smulders, A. P. H. J.
Schenning, E. W. Meijer, J. Am. Chem. Soc. 2008, 130, 606–611.
[21] T. E. Kaiser, V. Stepanenko, F. Wꢁrthner, J. Am. Chem. Soc. 2009,
131, 6719–6732.
[22] a) F. Oosawa, M. Kasai, J. Mol. Biol. 1962, 4, 10–21; b) P. Van der
Schoot, in Supramolecular Polymers, 2nd ed. (Ed.: A. Ciferri),
Taylor & Francis, London 2005.
[23] R. F. Goldstein, L. Stryer, Biophys. J. 1986, 50, 583–599.
[24] M. M. Smulders, M. M. L. Nieuwenhuizen, T. F. A. De Greef, P. Van
der Schoot, A. P. H. J. Schenning, E. W. Meijer, Chem. Eur. J. 2010,
16, 362–367.
[3] a) A. Ciferri, Supramolecular Polymers, M. Dekker, New York
2000; b) L. Brunsveld, B. J. B. Folmer, E. W. Meijer, R. P. Sijbesma,
Chem. Rev. 2001, 101, 4071–4097; c) Supramolecular Polymers, 2nd
ed. (Ed.: A. Ciferri), Taylor & Francis, New York, 2005; d) J. D. Fox,
S. J. Rowan, Macromolecules 2009, 42, 6823–6835; e) T. F. A. De
Greef, M. M. J. Smulders, M. Wolffs, A. P. H. J. Schenning, R. P. Sij-
besma, E. W. Meijer, Chem. Rev. 2009, 109, 5687–5754.
[4] a) L. Zang, Y. Che, J. S. Moore, Acc. Chem. Res. 2008, 41, 1596–
1608; b) W. Weng, Z. Li, A. M. Jamieson, S. J. Rowan, Soft Matter
2009, 5, 4647–4657; c) Y. Wang, H. Xu, X. Zhang, Adv. Mater. 2009,
21, 2849–2864; d) B. Rybtchinski, ACS Nano 2011, 5, 6791–6818.
[5] a) R. K. Castellano, D. M. Rudkevich, J. Rebek, Proc. Natl. Acad.
Sci. USA 1997, 94, 7132–7137; b) R. P. Sijbesma, F. H. Beijer, L.
Brunsveld, B. J. B. Folmer, J. H. K. K. Hirschberg, R. F. M. Lange,
J. K. L. Lowe, E. W. Meijer, Science 1997, 278, 1601–1604; c) R. P.
Sijbesma, E. W. Meijer, Chem. Commun. 2003, 5–16; d) H. M.
Keizer, R. P. Sijbesma, Chem. Soc. Rev. 2005, 34, 226–234; e) S. Si-
vakova, S. J. Rowan, Chem. Soc. Rev. 2005, 34, 9–21; f) D. Gon-
zꢀlez-Rodrꢅguez, A. P. H. J. Schenning, Chem. Mater. 2011, 23, 310–
325.
[6] a) U. S. Schubert, C. Eschbaumer, Angew. Chem. 2002, 114, 3016–
3050; Angew. Chem. Int. Ed. 2002, 41, 2892–2926; b) R. Dobrawa,
F. Wꢁrthner, J. Polym. Sci. Part A 2005, 43, 4981–4995; c) V. Bellas,
M. Rehahn, Angew. Chem. 2007, 119, 5174–5197; Angew. Chem.
Int. Ed. 2007, 46, 5082–5104; d) M. Elhabiri, A.-M. Elbrecht-Gary,
Coord. Chem. Rev. 2008, 252, 1079–1092; e) V. A. Friese, D. G.
Kurth, Coord. Chem. Rev. 2008, 252, 199–211; f) M.-O. M. Piepen-
brock, G. O. Lloyd, N. Clarke, J. W. Steed, Chem. Rev. 2010, 110,
1960–2004; g) A. Wild, A. Winter, F. Schlꢁtter, U. S. Schubert,
Chem. Soc. Rev. 2011, 40, 1459–1511.
[7] a) R. M. Yebeutchou, F. Tancini, N. Demitri, S. Geremia, R. Mendi-
chi, E. Dalcanale, Angew. Chem. 2008, 120, 4580–4584; Angew.
Chem. Int. Ed. 2008, 47, 4504–4508; b) A. Harada, Y. Takashima,
H. Yamaguchi, Chem. Soc. Rev. 2009, 38, 875–882; c) A. Harada, A.
Hashidzume, H. Yamaguchi, Y. Takashima, Chem. Rev. 2009, 109,
5974–6023; d) B. Zheng, F. Wang, S. Dong, F. Huang, Chem. Soc.
Rev. 2012, 41, 1621–1636.
[25] A. Lohr, T. Greß, M. Deppisch, M. Knoll, F. Wꢁrthner, Synthesis
2007, 3073–3082.
[8] a) C. A. Hunter, Chem. Soc. Rev. 1994, 23, 101–109; b) Z. Chen, A.
Lohr, C. R. Saha-Mçller, F. Wꢁrthner, Chem. Soc. Rev. 2009, 38,
564–584; c) M. Hasegawa, M. Iyoda, Chem. Soc. Rev. 2010, 39,
2420–2427.
[9] a) J. J. Reczek, K. R. Villazor, V. Lynch, T. M. Swager, B. L. Iverson,
J. Am. Chem. Soc. 2006, 128, 7995–8002; b) E. M. Pꢆrez, N. Martin,
Chem. Soc. Rev. 2008, 37, 1512–1519; c) Y. L. Liu, Y. Yu, J. Gao,
Z. Q. Wang, X. Zhang, Angew. Chem. 2010, 122, 6726–6729;
Angew. Chem. Int. Ed. 2010, 49, 6576–6579.
[10] A. Lohr, F. Wꢁrthner, Isr. J. Chem. 2011, 51, 1052–1066.
[11] a) F. Wꢁrthner, S. Yao, Angew. Chem. 2000, 112, 2054–2057; Angew.
Chem. Int. Ed. 2000, 39, 1978–1981; b) F. Wꢁrthner, S. Yao, T. De-
baerdemaker, R. Wortmann, J. Am. Chem. Soc. 2002, 124, 9431–
9447.
[26] a) J.-H. Ryu, M. Lee, J. Am. Chem. Soc. 2005, 127, 14170–14171;
b) J.-H. Ryu, H.-J. Kim, Z. Huang, E. Lee, M. Lee, Angew. Chem.
2006, 118, 5430–5433; Angew. Chem. Int. Ed. 2006, 45, 5304–5307;
c) G. Fernꢀndez, F. Garcia, L. Sanchez, Chem. Commun. 2008,
6567–6569.
[27] a) S. L. Craig, Angew. Chem. 2009, 121, 2683–2685; Angew. Chem.
Int. Ed. 2009, 48, 2645–2647; b) E. J. Foster, E. B. Berda, E. W.
Meijer, J. Am. Chem. Soc. 2009, 131, 6964–6966; c) W. Weng, Z. Li,
A. M. Jamieson, S. J. Rowan, Macromolecules 2009, 42, 236–246.
[28] a) W. Lu, Y. Chen, V. A. L. Roy, S. S.-Y. Chui, C.-M. Che, Angew.
Chem. 2009, 121, 7757–7761; Angew. Chem. Int. Ed. 2009, 48, 7621–
7625; b) R. Abbel, C. Grenier, M. J. Pouderoijen, J. W. Stouwdam,
P. E. L. G. Leclꢇre, R. P. Sijbesma, E. W. Meijer, A. P. H. J. Schen-
ning, J. Am. Chem. Soc. 2009, 131, 833–843.
[12] a) A. Lohr, S. Uemura, F. Wꢁrthner, Angew. Chem. 2009, 121,
6281–6284; Angew. Chem. Int. Ed. 2009, 48, 6165–6168; b) A. Lohr,
M. Grꢁne, F. Wꢁrthner, Chem. Eur. J. 2009, 15, 3691–3705.
[13] a) S. Yao, U. Beginn, T. Gress, M. Lysetska, F. Wꢁrthner, J. Am.
Chem. Soc. 2004, 126, 8336–8348; b) A. Lohr, M. Lysetska, F.
Wꢁrthner, Angew. Chem. 2005, 117, 5199–5202; Angew. Chem. Int.
Ed. 2005, 44, 5071–5074; c) A. Lohr, F. Wꢁrthner, Angew. Chem.
2008, 120, 1252–1256; Angew. Chem. Int. Ed. 2008, 47, 1232–1236.
[14] For other work on self-assembly of merocyanines dyes, see: a) S.
Yagai, M. Ishii, T. Karatsu, A. Kitamura, Angew. Chem. 2007, 119,
8151–8155; Angew. Chem. Int. Ed. 2007, 46, 8005–8009; b) S. Yagai,
T. Kinoshita, M. Higashi, K. Kishikawa, T. Nakanishi, T. Karatsu, A.
Kitamura, J. Am. Chem. Soc. 2007, 129, 13277–13287.
[29] a) S. Yagai, S. Kubota, H. Saito, K. Unoike, T. Karatsua, A. Kita-
mura, A. Ajayaghosh, M. Kanesato, Y. Kikkawa, J. Am. Chem. Soc.
2009, 131, 5408–5410; b) J. Miao, L. Zhu, Chem. Mater. 2010, 22,
197–206.
[30] a) E. Lee, J.-K. Kim, M. Lee, Angew. Chem. 2009, 121, 3711–3714;
Angew. Chem. Int. Ed. 2009, 48, 3657–3660; b) D.-J. Hong, E. Lee,
H. Jeong, J.-K. Lee, W.-C. Zin, T. D. Nguyen, S. C. Glotzer, M. Lee,
Angew. Chem. 2009, 121, 1692–1696; Angew. Chem. Int. Ed. 2009,
48, 1664–1668.
[31] T. Nakanishi, T. Michinobu, K. Yoshida, N. Shirahata, K. Ariga, H.
Mçhwald, D. G. Kurth, Adv. Mater. 2008, 20, 443–446.
[32] H. Bꢁrckstꢁmmer, E. V. Tulyakova, M. Deppisch, M. R. Lenze,
N. M. Kronenberg, M. Gsꢂnger, M. Stolte, K. Meerholz, F. Wꢁrth-
216
www.chemeurj.org
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 206 – 217

Cooperative Supramolecular Polymerization
FULL PAPER
ner, Angew. Chem. 2011, 123, 11832–11836; Angew. Chem. Int. Ed.
2011, 50, 11628–11632.
[36] A. J. Markvoort, H. M. M. ten Eikelder, P. A. J. Hilbers, T. F. A. de
Greef, E. W. Meijer, Nat. Commun. 2011, 2, 509–517.
[33] F. C. Spano, Acc. Chem. Res. 2010, 43, 429–439.
[34] The importance of alkyl-alkyl chain interactions is well-known in
paraffin crystals, self-assembled monolayers and for biological mem-
branes, see for example, a) A. Ulman, Chem. Rev. 1996, 96, 1533–
1554; b) R. Azumi, G. Gçtz, T. Debaerdemaeker, P. Bꢂuerle, Chem.
Eur. J. 2000, 6, 735–744.
[37] H. M. M. ten Eikelder, A. J. Markvoort, T. F. A. de Greef, P. A. J.
Hilbers, J. Phys. Chem. B 2012, 116, 5291–5301.
[38] D. Zhao, J. S. Moore, Org. Biomol. Chem. 2003, 1, 3471–3491.
Received: July 26, 2012
Revised: October 17, 2012
[35] U. Mayerhçffer, F. Wꢁrthner, Chem. Sci. 2012, 3, 1215–1220.
Published online: November 21, 2012
Chem. Eur. J. 2013, 19, 206 – 217
ꢃ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
217