Dipole-Moment-Driven Cooperative Supramolecular Polymerization

Article abstract of DOI:10.1021/jacs.5b00504

While the mechanism of self-assembly of π-conjugated molecules has been well studied to gain control over the structure and functionality of supramolecular polymers, the intermolecular interactions underpinning it are poorly understood. Here, we study the mechanism of self-assembly of perylene bisimide derivatives possessing dipolar carbonate groups as linkers. It was observed that the combination of carbonate linkers and cholesterol/dihydrocholesterol self-assembling moieties led to a cooperative mechanism of self-assembly. Atomistic molecular dynamics simulations of an assembly in explicit solvent strongly suggest that the dipole-dipole interaction between the carbonate groups imparts a macro-dipolar character to the assembly. This is confirmed experimentally through the observation of a significant polarization in the bulk phase for molecules following a cooperative mechanism. The cooperativity is attributed to the presence of dipole-dipole interaction in the assembly. Thus, anisotropic long-range intermolecular interactions such as dipole-dipole interaction can serve as a way to obtain cooperative self-assembly and aid in rationalizing and predicting the mechanisms in various synthetic supramolecular polymers. (Graph Presented).

Full text of DOI:10.1021/jacs.5b00504

Article
pubs.acs.org/JACS
Dipole-Moment-Driven Cooperative Supramolecular Polymerization
Chidambar Kulkarni,†,§ Karteek K. Bejagam,§ Satyaprasad P. Senanayak,§ K. S. Narayan,§
S. Balasubramanian,*,§ and Subi J. George*,†
†New Chemistry Unit, Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore 560064, India
§Chemistry and Physics of Materials Unit, Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore 560064, India
S
* Supporting Information
ABSTRACT: While the mechanism of self-assembly of π-
conjugated molecules has been well studied to gain control
over the structure and functionality of supramolecular polymers,
the intermolecular interactions underpinning it are poorly
understood. Here, we study the mechanism of self-assembly of
perylene bisimide derivatives possessing dipolar carbonate groups
as linkers. It was observed that the combination of carbonate
linkers and cholesterol/dihydrocholesterol self-assembling moi-
eties led to a cooperative mechanism of self-assembly. Atomistic
molecular dynamics simulations of an assembly in explicit solvent
strongly suggest that the dipole−dipole interaction between the
carbonate groups imparts a macro-dipolar character to the assembly. This is confirmed experimentally through the observation of
a significant polarization in the bulk phase for molecules following a cooperative mechanism. The cooperativity is attributed to
the presence of dipole−dipole interaction in the assembly. Thus, anisotropic long-range intermolecular interactions such as
dipole−dipole interaction can serve as a way to obtain cooperative self-assembly and aid in rationalizing and predicting the
mechanisms in various synthetic supramolecular polymers.
the system apparently lacks hydrogen bonding along the
stacking direction and yet shows a cooperative mechanism of
self-assembly.29,30 Recently, cooperativity has also been
observed in π-conjugated systems with a variety of other
intermolecular interactions operating in the growth direc-
tion.31−34 In order to rationalize the observed mechanisms in a
broader sense, we had previously proposed the necessity of
anisotropic long-range interactions between monomers (or
small stacks) for a cooperative self-assembly.35 Dipolar
interaction is one such, and it has also been used to construct
supramolecular polymers.36 The effect of dipolar groups
present in a molecule on the mechanism of supramolecular
INTRODUCTION
■
Synthetic macromolecules assembled from monomers capable
of exhibiting intermolecular interactions such as hydrogen-
bonding, π-stacking, amphiphilic, and electrostatic interactions
have been studied to mimic their biological counterparts and
for materials applications.1−6 The investigation into the
mechanism of supramolecular polymerization has gained
importance in the past decade to allow for a better control
over its structure and functions.7−13 The mechanism of non-
covalent one-dimensional (1-D) polymerization has been
broadly classified into two types: isodesmic and cooperative.14
A cooperative or nucleation−elongation mechanism has been
observed in biopolymers such as actin15 and flagellin.16
Thermodynamic aspects of self-assembly and their relationship
with mechanism have been studied extensively.14,17,18 In their
seminal review, Meijer and co-workers suggested electronic,
structural, and hydrophobic interactions as the cause of
cooperativity in supramolecular polymers.14 Hitherto, a
comprehensive rationale for the observed mechanisms of
polymerization based on structural motifs in the monomer is
lacking.
polymerization has seldom been studied. Recently, Wurthner
̈
and co-workers have shown the cooperative supramolecular
polymerization of dipolar merocyanine dyes.37
Perylene-3,4,9,10-tetracarboxylic acid bisimides (PBIs) are
well-known electron-deficient organic semiconductors and have
been widely employed as active materials in organic
electronics.38,39 PBI derivatives are known to follow an
isodesmic mechanism of self-assembly.40−43 However, there
are only a few reports (by the groups of Wurthner, Meijer, and
̈
Benzene-1,3,5-tricarboxamide19−21 derivatives are seminal
examples of cooperative supramolecular polymerization involv-
ing intermolecular hydrogen-bonding in the stacking direction.
In addition, many π-conjugated systems studied in the literature
that are likewise capable of intermolecular hydrogen bonding
have also been shown to follow the cooperative mechanism of
self-assembly.22−28 However, there are a few exceptions where
Yagai) on the cooperative self-assembly of PBIs; these are
mainly driven by either intermolecular hydrogen bonding or
intramolecular hydrogen bonding together with π-stack-
ing.44−47
Received: January 16, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b00504
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
In the present study, we investigate the effect of dipolar
groups and different self-assembling moieties appended to PBI
on the mechanism of self-assembly. Here we observe
cooperative self-assembly of PBI derivatives containing no
apparent hydrogen-bonding groups. Extensive spectroscopy,
molecular dynamics (MD) simulations, and bulk dielectric
measurements were performed on these molecules to shed light
on the origin of cooperativity. Herein, we show that dipole−
dipole interaction between monomers along the stacking
direction is the primary cause of cooperativity. The interaction
is shown to arise from the dipolar nature of carbonate linkers
and the rigidity of cholesterol self-assembling moieties. Thus,
we present a unique example of non-hydrogen-bonded and
dipole-moment-driven cooperativity in a supramolecular
system.
chiral swallowtail, so as to be able to examine the effect of this
moiety on the mechanism of self-assembly. Synthesis of
molecules 1−4 was carried out by coupling of the
corresponding chloroformates49 with N-Boc-ethanolamine,
followed by deprotection to yield the respective amines.
Condensation of the respective amines with perylene-
3,4,9,10-tetracarboxylic acid dianhydride yielded the target
1
molecules. All molecules were characterized by H and 13C
NMR, IR, and mass spectroscopy (see Supporting Information
(SI) for experimental details and characterization).
Self-Assembly Studies. UV/vis absorption spectrum of 1
in chloroform (c = 1 × 10−5 M) shows prominent peaks at 525
(ε = 74 200 M−1 cm−1), 488, 458, and 260 nm. The absorption
features between 450 and 530 nm have been attributed to the
vibronic bands of S0→S1 transition of PBIs along the long axis
of the molecule.50 The corresponding fluorescence spectrum
shows a mirror image of the absorption spectrum with a
maximum at 538 nm (SI, Figure S6a,b). These absorption and
fluorescence spectral features are characteristic of π−π*
transitions of PBIs devoid of intermolecular interactions.39
The UV/vis absorption spectrum of 1 in methylcyclohexane
(MCH) shows the loss of vibronic features, with an absorption
maximum at 473 nm and a broad band centered around 555
nm. The fluorescence spectrum of 1 in MCH shows the
quenching of monomer emission (at 538 nm by 39 times) and
the emergence of a new band at 640 nm (inset of SI, Figure
S6b). The fluorescence excitation spectrum of 1 in MCH
obtained by monitoring the emission intensity at 565 and 650
nm shows well-resolved vibronic bands (400−550 nm) and a
broad band centered around 470 nm, respectively (see SI,
Figure S7). These spectral features indicate that the
fluorescence of 1 at 565 nm is due to the residual monomers
and that at 650 nm arises from the aggregates. The
hypsochromic shift of the absorption maximum and quenching
of monomer emission points to the face to face H-type of
aggregate formation of 1 in MCH.42 Similar H-type aggregate
formation has been observed for 2 (Figure S8), 3 (Figures S9
and S10) and 4 (Figure S11) in apolar solvents like MCH (see
SI). Nanostructures of aggregates of 1−4 have been
characterized through dynamic light scattering studies and
atomic force microscopy. These studies reveal 1-D aggregation
of molecules 1, 2, and 4 (see SI, Figures S12−S15, for details).
On the other hand, 3 forms spherical particles. It has been
observed earlier that some of the bischolesterol derivatives self-
assemble into 1-D structures, and further higher order
aggregation leads to formation of spherical particles.51
The cholesterol moiety possessing chiral centers has been
utilized for the organization of various chromophores.52−54 It
has an additional role in acting as a source of chiral bias to
organize the π-conjugated molecules into helical assemblies by
imparting van der Waals interaction to stabilize the
assemblies.55 Circular dichroism (CD) spectra of 1 in MCH
and 1,1,2,2- tetrachloroethane (TCE) (95:5, v/v) (c = 1 × 10−5
M) showed a bisignated positive Cotton effect with positive
and negative maximum at 485 and 450 nm, respectively (Figure
2a), suggesting the right handed chiral organization of the PBI
core due to the presence of cholesterol moiety.56 Further, by
heating the solution, CD effect (mdeg) decreased in magnitude,
and beyond 70 °C it vanishes (Figure 2a), suggesting the loss of
chiral organization of the assembly. The corresponding
temperature-dependent UV/vis absorption spectra show the
evolution of vibronic features (characteristic of monomers)
with an increase in temperature (Figure 2b). This observation
RESULTS
■
The molecular design involves utilizing the optical properties of
the PBI core to probe the mechanism of self-assembly by
varying the linker group (either carbonate or ether) and self-
assembling moiety (cholesterol or alkyl groups, Figure 1).
Figure 1. Molecules under study. (a) PBI is functionalized on both
imide nitrogens with either carbonate or ether linker (marked in red).
Cholesterol (1 and 3), dihydrocholesterol (2), and chiral swallowtail
(4) are used as the self-assembling motifs. (b) Schematic
representation of the assembly of different PBI derivatives. PBIs are
depicted by red blocks. Blue and cyan cylinders represent carbonate
and ether linkers, respectively. Self-assembling moieties are shown in
green. Small yellow arrows indicate molecular dipoles along the
carbonate CO axis. Macro-dipole moment is shown as a large yellow
arrow along the stacking direction.
Dialkyl carbonates are known to possess dipole moment in
different conformations48 and are thus utilized as the source of
molecular dipole moment. An ethylene spacer is chosen to be
placed between the imide nitrogen and the carbonate to
provide appropriate solubility and rigidity to the molecule. 1
and 2 contain dipolar linkers (carbonate) and a rigid steroid-
based self-assembling moiety. In contrast, 3, possessing ether
linkers, differs only in the absence of dipolar groups, thus
allowing us to study the effect of linkers on the mechanism of
self-assembly. In 4, the steroid moiety of 1 is exchanged for a
B
DOI: 10.1021/jacs.5b00504
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Self-assembly of 1. Temperature-dependent CD (a) and UV/vis absorption spectra (b) of 1 in MCH/TCE (95:5, v/v) at every 10 °C. All
studies were done at a concentration of 1 × 10−5 M in a 10 mm cuvette. Inset of (a) shows a schematic representation of the assembly at high and
low temperatures. Arrows indicate spectral changes with an increase in temperature.
indicates that the CD spectral changes (with temperature) are
indeed due to the disassembly of the aggregates. CD spectra of
dihydrocholesterol appended derivative with carbonate linker
(2) (Figure 3 and SI, Figure S16) and cholesterol tethered
Mechanism of Self-Assembly. The temperature depend-
ence of the degree of aggregation (α) has been employed in the
literature to ascertain the mechanism of self-assembly.26,29,59
Utilizing the same methodology, we study the self-assembly
mechanism of 1−4 by monitoring changes in either CD or UV/
vis absorption spectra at a particular wavelength (characteristic
of the assemblies) as a function of temperature. The normalized
change in CD monitored at 480 nm for a solution of 1
(carbonate linker and cholesterol motif) as a function of
temperature (cooling curve) is clearly non-sigmoidal (Figure
4a).
The experimental data are in agreement with the fit obtained
by the Eikelder−Markvoort−Meijer (EMM) model for a
nucleation−elongation mechanism for a one-component
system,60,61 indicating a cooperative (or nucleation−elonga-
tion) mechanism of self-assembly of 1. The temperature in the
cooling curve at which the self-assembly begins is termed as
elongation temperature (Te). The relevant thermodynamic
parameters for the self-assembly of 1 at various concentrations
have been obtained from the fitting of the cooling curves (SI,
Table S2). In addition, temperature-dependent UV/vis studies
also exhibited a critical point in the cooling curve (SI, Figure
S21 and Table S3), reaffirming the cooperative nature of the
self-assembly of 1. It is observed that as the total concentration
of the solution increases, Te increases, while other thermody-
namic parameters like ΔG0 and cooperativity factor are nearly
identical. Similar effects of concentration on Te have been
reported in literature for other systems.19
Dihydrocholesterol- and cholestero-based ALS ( where A =
aromatic, L = linker, and S = steroidal groups) systems have
been shown to differ significantly in their gelation properties.62
Since dihydrocholesterol is a rarely used self-assembling motif,
it is interesting to note its effect on the mechanism of self-
assembly. Thus, the cooling curves obtained for 2 (carbonate
linker and dihydrocholesterol motif) by monitoring changes in
CD spectra (Figure 3a) and UV/vis absorption spectra (Figure
3b) at different concentrations also showed a non-sigmoidal
behavior (Figure 4b and SI, Figures S24 and S25); the
experimental data are well described by the nucleation−
elongation model. Cooling curves monitored at two different
wavelengths (475 and 560 nm) show similar Te, indicating that
both absorption bands correspond to the same aggregate (SI,
Figure S27).
Figure 3. Self-assembly of 2. Temperature-dependent CD (a) and
UV/vis absorption spectra (b) of 2 (c = 8 × 10−6 M) in MCH/TCE
(95:5, v/v).
derivative (3) (SI, Figure S17) in the self-assembled state show
positive and negative Cotton effect, respectively, indicating a
chiral supramolecular organization. The sign and magnitude of
the CD spectra vary with aging and cooling rate,57 as observed
in other cholesterol- and dihydrocholesterol-based gelators,54,58
probably due to the presence of multiple stereocenters (which
compete among themselves) in the self-assembling moiety.
Molecule 3 also exhibited a bisignated Cotton effect, suggesting
the chiral organization of molecules in the assembly (SI, Figure
S17). On the other hand, carbonate linker possessing 4 was CD
silent (SI, Figure S18), probably due to the high flexibility of
the chiral swallowtails, which hinders any chiral organization in
the assemblies (vide infra, MD simulations).
Apart from temperature, concentration can also be used to
vary the extent of aggregation, and such concentration-
C
DOI: 10.1021/jacs.5b00504
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 4. Mechanism of self-assembly of 1−4. (a,b) Degree of aggregation (α) as a function of temperature (cooling curve) for a solution (95%
MCH and 5% TCE) of 1 and 2, respectively, by monitoring the CD effect at 480 nm (for 1) and 485 nm (for 2) at a cooling rate of 2 K/min (c = 8.0
× 10−6 M). (c) Cooling curve obtained by monitoring changes in UV/vis absorbance of 3 at 570 nm at a cooling rate of 1 K/min (c = 2 × 10−5 M in
80% MCH and 20% TCE). (d) Normalized change in UV/vis absorbance of 4 as a function of temperature, monitoring the absorbance at 570 nm
(cooling rate 2 K/min, c = 4.69 × 10−4 M, in MCH). The solid blue and red lines are the fits obtained from nucleation−elongation and isodesmic
models, respectively. The structures of linker and self-assembling group corresponding to each of the molecules are shown as insets.
dependent studies have been extensively used by Wurthner and
2) show cooperativity comparable to conventional hydrogen-
bonded systems. In addition, 2 shows: a lower cooperativity
factor σ, more negative ΔG0, and lower solubility compared to
1 (Table 1). These differences can be mainly attributed to the
structural planarity of dihydrocholesterol compared to choles-
terol, which enhances van der Waals interaction between the
moieties in an assembly.62
̈
co-workers to ascertain the mechanism of self-assembly.37,42
Thus, concentration-dependent UV/vis absorption studies of 2
in MCH/TCE (95:5, v/v) were performed. The transition from
monomers to aggregates occurs within a narrow concentration
range of 2.5 × 10−6−3.3 × 10−6 M (SI, Figure S28), thus
indicating the cooperative mechanism of self-assembly.
Interestingly, the cooperativity parameter (σ, obtained from
The self-assembly mechanisms of 3 (ether linker and
cholesterol motif) and 4 (carbonate linker and chiral
swallowtail) were also studied using temperature-dependent
UV/vis spectroscopy.63 A solution (c = 2 × 10−5 M in MCH/
TCE (8:2, v/v)) of 3 showed the gradual evolution of the
extent of aggregation with decrease in temperature (Figure 4c).
The cooling curve could well be described by the isodesmic
model,59 yielding an enthalpy of polymerization (ΔH) of −92.0
kJ mol−1 and melting temperature (Tm, temperature at which α
= 0.50) of 316.5 K. Similarly, the cooling curve for 4 (in MCH,
c = 4.69 × 10−4 M) showed a sigmoidal behavior and could be
fit to an isodesmic model (Figure 4d),64 yielding the
fitting the experimental cooling curve to the EMM model60,61
)
for 2 was found to be consistently (by both CD and UV/vis
measurements) lower than that for 1 (Table 1), suggesting a
higher degree of cooperativity in the former system compared
to that in the latter. σ is of the order of 10−3 for 1, whereas for
systems containing intermolecular hydrogen bonds, it is found
in the range of 10−3−10−6.19,22−28 Despite the lack of any
apparent hydrogen-bonding motif, the present systems (1 and
Table 1. Comparison of Thermodynamic Parameters for the
Assembly of 1 and 2 (in 95% MCH and 5% TCE)
ΔG0
cooperativity
factor, σ
thermodynamic parameters ΔH = −53.5 kJ mol−1 and Tm
=
ΔHe
ΔSe
molecule
(kJ/mol)
(kJ/mol)
(kJ/mol)
300.4 K. Thus, the linker group of the monomer significantly
affects the mechanism of self-assembly.
1
2
−41.61
−89.98
−0.027
−0.17
−33.45
−37.25
2.4 × 10−3
2.0 × 10−4
Since 4 is isodesmic and weakly aggregating, another
derivative of PBI (PBI_carb_dd) possessing linear dodecyl
chains in place of branched chains (4) was synthesized to
investigate its effect on the mechanism of self-assembly.
PBI_carb_dd self-assembles into J-aggregates (SI, Figure
Values presented are average of those obtained by fitting to the EMM
model60,61 for 1 and 2 (from CD and UV/vis studies at different
concentrations). The values for 1 are averages of those presented in SI,
Tables S2 and S3. ΔG0 was calculated at 298.15 K.
D
DOI: 10.1021/jacs.5b00504
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 5. MD simulations of the assemblies of 1, 3, and 4. (a) Snapshot illustrating the arrangement of molecules in the assembled state. Linkers are
highlighted in yellow and magenta to aid in the visualization of the helical packing. Self-assembling groups are represented with thin sticks, and
hydrogens are omitted for clarity. (b) Distance distribution between the linkers of neighboring molecules. (c) Macro-dipole along the stacking
direction as a function of time. Solid horizontal lines (yellow for 1, blue for 3, and maroon for 4) are drawn to represent the mean dipole moment
value for each system. (d) Weighted probability distribution of macro-dipole obtained from (c). (e,f) Distribution of the cosine of the angle between
the dipole moment of each molecule (μ) with the normal to the PBI plane (n). All the analysis was performed on the last 12 ns of the MD trajectory.
S32) and exhibits a cooperative mechanism of self-assembly
(SI, Figure S33).
pre-formed oligomeric stacks, each containing 40 molecules of
1, 3, and 4 were performed in explicit cyclohexane at 298.15 K
for 20 ns (see SI, Figures S35−S37 and Tables S4 and S5, for
more details).71
Molecular Dynamics Simulations. Computational studies
were undertaken to gain further insight into the molecular
organization of the assemblies and to elucidate the role of
different moieties on the self-assembly mechanism. Quantum
chemical calculations are computationally expensive in
determining the equilibrium structure of assemblies due to
the large number of atoms. On the other hand, MD simulations
can describe the molecular packing in these supramolecular
assemblies.65−70 In particular, MD simulations are well suited to
treat the effect of solvents and temperature, so as to realize
experimental conditions. Thus, MD simulations were carried
out to investigate the molecular packing in the aggregated state
for 1, 3, and 4 and to understand the mechanism of their self-
assembly in terms of intermolecular interactions. Simulations of
Snapshots at the end of the MD simulations for cholesterol
appended molecules 1 and 3 exhibit helical packing of linker
groups (carbonate or ether) around the PBI core (Figure 5a).
This, in turn, is reflected in the ordered helical organization of
the peripheral cholesterol groups and PBI core, in agreement
with the experimental observation of bisignated CD signal
(Figure 2). On the other hand, the organization of linker
groups (carbonate) in the assembly of 4 is ill-defined (Figure
5a) which can be attributed to the flexibility of the chiral
swallowtail. Coincidentally, chiral organization in 4 is
experimentally determined to be absent as well (SI, Figure
S18). The arrangement of linkers in the assembly can be further
E
DOI: 10.1021/jacs.5b00504
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
quantified by the distance between them across neighboring
molecules in an assembly (Figure 5b). Assemblies of 1 and 4
exhibit a bimodal distribution of interlinker (carbonate)
distances, centered at 5.5 and 9 Å. For 1, the intensity at 5.5
Å is larger than at 8.9 Å, suggesting that the majority of
carbonate groups are more closely arranged in an assembly of 1
compared to that of 4. A unimodal probability distribution
centered at 5.1 Å is observed for the assembly of 3 (ether
linker), suggesting close and uniform packing of linkers. (For
further discussion see section “Origin of the bimodal
distribution of the linker distances” in SI and Figures S42
and S43.) The inclination angle of the carbonate vector, in two
helices, with respect to the normal of PBI plane is shown in SI,
Figure S40. The average of this quantity can be taken as an
order parameter for structural ordering. This quantity averaged
over the core 30 molecules for 1 and 4 is 0.149 and 0.005,
respectively, clearly suggesting the ordered arrangement of
molecules in the former over the latter. Thus, the differences in
the structural organization can be attributed to the rigidity or
flexibility of the self-assembling moiety.
The consequence of structural organization of linkers on the
macro-dipole moment (along the stacking direction) is studied
during the MD trajectory (Figure 5c,d). Macro-dipole is
calculated as ∑qiri, where qi is the charge on the atoms and ri
the displacement vector from negative to positive charge,
respectively, and the sum extends over all atoms. The mean
macro-dipole moment for 1 (14.7 D) is an order of magnitude
higher than that for 4 (1.43 D), although in both molecules (1
and 4), the carbonate linker is the dipolar entity. We now study
the orientation of the dipole moment of a molecule. In order to
eliminate edge effects, we discard contributions from five
molecules present at either end of the stack in further analysis.
The distribution of the angle between the dipole moment
vector (μ) of each molecule with the normal of its PBI plane
(n) is shown in Figure 5e,f. For 3 and 4, the distribution is
nearly symmetric about zero (Figure 5f), implying that parallel
and antiparallel orientations of molecular dipoles are equally
likely which does not lead to any significant macro-dipole for
the stack. In contrast, in the assembly of 1, the carbonate
linkers on each branch of the molecules are better aligned with
each other; however, the net dipole from each of these
branches, although in opposite directions, do not cancel fully.
This results in a net macro-dipole moment for a stack of 1
(details in SI). Thus, the difference in the macro-dipole
moment between 1 and 4 can be mainly attributed to the
organization of the linkers.
Figure 6. Bulk dielectric measurements. (a) εr measurement as a
function of frequency. (b) εr variation with temperature for molecule 2
at 10 kHz. Inset in (b) shows a schematic of the device and the dipole
alignment at different temperatures. The oval-shaped dipoles represent
the polarization of an aggregate and not that of individual molecules.
assembly. Classical theory of linear dielectrics based on
Clausius−Mossotti relation is used to estimate the polarizability
at zero frequency of the molecules.73 The effective dipole
moment per molecule (|μ|) estimated from the dielectric
constant measurement of the assembly was in the range of 3−5
D for molecules 1 and 2. In comparison, the |μ| values for
molecules 3 and 4 were in the range of 0.6−0.7 D. The value of
|μ| for molecules 1, 3, and 4 estimated from MD simulation is
10 times lower than the experimentally obtained value.74 This
can be attributed to the difference in the measurement
conditions (such as condensed phase) versus that in the
simulations which were carried out in the solution phase.
However, the trends in dipole-moment values are same across
different molecules obtained from both MD simulations and
dielectric measurements.
Dielectric Measurements. A significant macro-dipole
moment in 1 and none in 4 would be reflected in the dielectric
constants of these molecules. Capacitance measurement is a
direct experimental means of quantifying the extent of
polarization in the system. Thus, we have measured the bulk
(thin-film state) capacitance for assemblies of 1−4 to evaluate
the magnitude of polarization and to correlate with the
mechanism of self-assembly at the molecular level.
The dielectric constant, εr = Cd/ε0A (where C is capacitance,
d film thickness, ε0 vacuum permittivity, and A area of the film
obtained from the capacitive response), is essentially a measure
of the macroscopic dipole moment.72 Molecules 1 and 2, which
exhibit a cooperative self-assembly, display higher polarization
indicated by the higher magnitude of εr (Figure 6a). It should
be noted that this magnitude of εr (≈ 16 for molecule 2) is
around 5 times higher than the εr observed for molecules 3 and
4, which do not follow a cooperative mechanism of self-
The relaxation dynamics involved in the dipole or an
assembly of dipoles is probed by looking into the variation of εr
as a function of frequency. A decrease in εr magnitude at high
frequency is observed for molecules 1 and 2, whereas molecules
3 and 4 demonstrated a flat frequency response (Figure 6a).
This response can be directly correlated to the long-range
interaction which drives the cooperative assembly. Due to the
existence of long-range interaction, the dipoles behave as
clusters or domains which have a relatively slower response
compared to individual dipoles.75 Thus, frequency-dependent
dielectric studies confirm the presence of significant clustering
of dipoles in condensed phase of 1 and 2 and a lack of such
interactions for 3 and 4.
The temperature dependence of dielectric constant reveals
the significance of dipole−dipole interaction necessary to drive
the cooperative behavior. The εr versus temperature curves for
molecules 1 (SI, Figure S45a) and 2 (Figure 6b) indicate a
F
DOI: 10.1021/jacs.5b00504
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
critical temperature where the εr starts to decrease, and for
higher temperatures (390−420 K) εr = 1−2 is obtained. This
behavior can be correlated to the cooperative nature of the
individual molecular dipoles which result in a macroscopic
polarization. As the temperature increases, the thermal-disorder
decreases the dipole−dipole interaction between monomers in
an assembly. It is to be noted here that the assembly is not
completely disassembled at higher temperatures (see SI, Figure
S46); rather, the dipole−dipole interactions have vanished. It is
also observed that ωbandwidth (where εr decreases by a factor of
2) has a monotonous dependence with temperature; i.e., a
slower relaxation at low temperature indicates a larger range of
interaction and coupling of the dipoles in form of clusters, while
the upper limit of relaxation is achieved at high T and
was observed through both MD simulation and bulk dielectric
measurements. However, the absence of either the dipolar
carbonate group, as in molecule 3, or a suitable self-assembling
moiety, as in molecule 4, results in an isodesmic self-assembly
without any significant polarization for the assembly. Thus, with
the aid of careful molecular design, MD simulation, and bulk
dielectric measurements, we are able to provide a proof-of-
concept example for the long-range interaction-driven cooper-
ative self-assembly, as hypothesized earlier.35 It is also to be
noted that polarization or macro-dipole only in the stacking
direction is critical in governing the mechanism of self-
assembly. Presently studies are underway in our laboratory to
apply the same concepts to other π-conjugated molecules and
to look at the generality of the above conclusions.
corresponds to the classical independent dipole (εclassical‑dipole
)
relaxation limit (see SI, Figure S47b). These findings suggest
that temperature-dependent εr measurements can be used as an
electrical probe for monitoring the presence of dipole−dipole
interactions, which are critical in governing the mechanism of
self-assembly. In the case of molecules 3 and 4, which exhibit
isodesmic mechanism, a typical Arrhenius behavior is observed
for εr(T) (see SI, Figure S45b,c). Thus, cooperativity or
otherwise in the assembly mechanism can also be observed
using dielectric measurements.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthesis and characterization details for 1−4; computational
and dielectric measurement details; supporting figures and
tables. This material is available free of charge via the Internet
at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Authors
*bala@jncasr.ac.in
*george@jncasr.ac.in
DISCUSSION
■
The different mechanisms of self-assembly observed in the
present systems can be rationalized at a molecular level, based
on MD simulation and bulk dielectric measurements as
summarized below. For molecules 1 and 2, the carbonate
linkers are closely packed in an ordered manner leading to an
enhanced interaction between them, which is reflected in the
magnitude of its macro-dipole moment and dielectric constant.
Between 1 and 2, the latter shows higher cooperativity, more
negative free energy of polymerization, and higher dielectric
constant. This could be mainly due to the planar and rigid
conformation of dihydrocholesterol leading to stronger
carbonate−carbonate interaction in the assembly compared to
molecule 1. In the case of 3, although the assembly is ordered,
it does not lead to a high macro-dipole moment since the
dipole moment of ether linkers is small; no appreciable long-
range interaction is possible, and the molecule assembles in an
isodesmic manner. Similarly, although 4 possesses dipolar
carbonate groups, because of the lack of ordered organization
wrought by the flexibility of chiral swallowtails, no macro-dipole
moment is observed, resulting in its isodesmic polymerization.
Cooperative self-assembly is observed for PBI_carb_dd
containing dodecycl chains as the self-assembling moiety and
carbonate groups as linkers; this again suggests that the linear
dodecyl chain can pack better through van der Waals
interactions, leading to strengthening of the carbonate−
carbonate interaction. Thus, the rigidity/flexibility of the self-
assembling moiety reinforces the dipolar interactions between
the monomers in the assembly.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank Department of Science and Technology
(DST), India, for financial support. S.J.G. and S.B. gratefully
acknowledge Sheikh Saqr Career Award. C.K. thanks Prof. A. J.
Markvoort and Prof. ten H. M. M. Eikelder from TU/e for
suggestions on modeling experimental data, and Mr. V. K.
Bharadwaj for aid in synthesis of molecule 1. C. K.
acknowledges Mr. Kumaraswamya from Malvern for DLS,
and Mr. Hitesh Khandelwal and Ralf Bovee from TU/e for
MALDI-TOF measurements. The authors also thank Prof. E.
W. Meijer for comments on the manuscript.
REFERENCES
■
(1) Hoeben, F. J. M.; Jonkheijm, P.; Meijer, E. W.; Schenning, A. P.
H. J. Chem. Rev. 2005, 105, 1491.
(2) Brunsveld, L.; Folmer, B. J. B.; Meijer, E. W.; Sijbesma, R. P.
Chem. Rev. 2001, 101, 4071.
(3) Babu, S. S.; Praveen, V. K.; Ajayaghosh, A. Chem. Rev. 2014, 114,
1973.
(4) Aida, T.; Meijer, E. W.; Stupp, S. I. Science 2012, 335, 813.
(5) Stupp, S. I.; Palmer, L. C. Chem. Mater. 2013, 26, 507.
(6) Zelzer, M.; Ulijn, R. V. Chem. Soc. Rev. 2010, 39, 3351.
(7) Faramarzi, V.; Niess, F.; Moulin, E.; Maaloum, M.; Dayen, J.-F.;
Beaufrand, J.-B.; Zanettini, S.; Doudin, B.; Giuseppone, N. Nat. Chem.
2012, 4, 485.
(8) Ogi, S.; Sugiyasu, K.; Manna, S.; Samitsu, S.; Takeuchi, M. Nat.
Chem. 2014, 6, 188.
CONCLUSION
■
(9) Gilroy, J. B.; Gadt, T.; Whittell, G. R.; Chabanne, L.; Mitchels, J.
̈
We have studied the mechanism of self-assembly of four PBI
derivatives appended with different linkers and self-assembling
moieties, in an attempt to elucidate the molecular features
governing the mechanism of supramolecular polymerization.
For molecules following a cooperative mechanism of self-
assembly (1 and 2), significant macro-dipole or polarization
M.; Richardson, R. M.; Winnik, M. A.; Manners, I. Nat. Chem. 2010, 2,
566.
(10) Hudson, Z. M.; Lunn, D. J.; Winnik, M. A.; Manners, I. Nat.
Commun. 2014, 5, No. 3372.
(11) Kang, J.; Miyajima, D.; Mori, T.; Inoue, Y.; Itoh, Y.; Aida, T.
Science 2015, 347, 646.
G
DOI: 10.1021/jacs.5b00504
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(12) Ogi, S.; Fukui, T.; Jue, M. L.; Takeuchi, M.; Sugiyasu, K. Angew.
Chem., Int. Ed. 2014, 53, 14363.
(13) Jain, A.; George, S. J. Mater. Today 2015, DOI: 10.1016/
(46) Yagai, S.; Usui, M.; Seki, T.; Murayama, H.; Kikkawa, Y.;
Uemura, S.; Karatsu, T.; Kitamura, A.; Asano, A.; Seki, S. J. Am. Chem.
Soc. 2012, 134, 7983.
(47) Seki, T.; Asano, A.; Seki, S.; Kikkawa, Y.; Murayama, H.;
Karatsu, T.; Kitamura, A.; Yagai, S. Chem.Eur. J. 2011, 17, 3598.
(48) Reddy, S. K.; Balasubramanian, S. J. Phys. Chem. B 2012, 116,
14892.
(49) Chloroformates were either obtained from commercial source
or synthesized by the chloroformylation of the corresponding alcohols
using triphosgene following the reported procedure. See SI for details.
(50) Adachi, M.; Murata, Y.; Nakamura, S. J. Phys. Chem. 1995, 99,
14240.
(51) Abraham, S.; Vijayaraghavan, R. K.; Das, S. Langmuir 2009, 25,
8507.
(52) Svobodova, H.; Noponen, V.; Kolehmainen, E.; Sievanen, E.
RSC Adv. 2012, 2, 4985.
(53) Sugiyasu, K.; Fujita, N.; Shinkai, S. Angew. Chem., Int. Ed. 2004,
j.mattod.2015.01.015.
(14) De Greef, T. F. A.; Smulders, M. M. J.; Wolffs, M.; Schenning,
A. P. H. J.; Sijbesma, R. P.; Meijer, E. W. Chem. Rev. 2009, 109, 5687.
(15) Sept, D.; McCammon, J. A. Biophys. J. 2001, 81, 667.
(16) Abram, D.; Koffler, H. J. Mol. Biol. 1964, 9, 168.
(17) Chen, Z.; Lohr, A.; Saha-Moller, C. R.; Wurthner, F. Chem. Soc.
Rev. 2009, 38, 564.
(18) Zhao, D.; Moore, J. S. Org. Biomol. Chem. 2003, 1, 3471.
(19) Smulders, M. M. J.; Schenning, A. P. H. J.; Meijer, E. W. J. Am.
Chem. Soc. 2008, 130, 606.
(20) Cantekin, S.; Balkenende, D. W. R.; Smulders, M. M. J.;
Palmans, A. R. A.; Meijer, E. W. Nat. Chem. 2011, 3, 42.
(21) Smulders, M. M. J.; Nieuwenhuizen, M. M. L.; Grossman, M.;
Filot, I. A. W.; Lee, C. C.; de Greef, T. F. A.; Schenning, A. P. H. J.;
Palmans, A. R. A.; Meijer, E. W. Macromolecules 2011, 44, 6581.
(22) Wang, F.; Gillissen, M. A. J.; Stals, P. J. M.; Palmans, A. R. A.;
Meijer, E. W. Chem.Eur. J. 2012, 18, 11761.
(23) García, F.; Viruela, P. M.; Matesanz, E.; Ortí, E.; Sanchez, L.
́
Chem.Eur. J. 2011, 17, 7755.
(24) Helmich, F.; Lee, C. C.; Nieuwenhuizen, M. M. L.; Gielen, J. C.;
Christianen, P. C. M.; Larsen, A.; Fytas, G.; Leclere, P. E. L. G.;
Schenning, A. P. H. J.; Meijer, E. W. Angew. Chem., Int. Ed. 2010, 49,
3939.
̈
̈
43, 1229.
(54) Murata, K.; Aoki, M.; Suzuki, T.; Harada, T.; Kawabata, H.;
Komori, T.; Ohseto, F.; Ueda, K.; Shinkai, S. J. Am. Chem. Soc. 1994,
116, 6664.
(55) Ajayaghosh, A.; Vijayakumar, C.; Varghese, R.; George, S. J.
Angew. Chem., Int. Ed. 2006, 45, 456.
(56) Although 1 is completely soluble in MCH, the aggregates could
not be disassembled with increase in temperature. Thus, to control the
extent of self-assembly such that diassembly could be achieved, a
fraction of good solvent (TCE) was used.
(57) The data are not shown. Also, a systematic study of the variation
of CD spectra as a function of cooling rate and aging is beyond the
scope of this work and will be investigated in detail as a separate work.
(58) Lu, L.; Cocker, T. M.; Bachman, R. E.; Weiss, R. G. Langmuir
2000, 16, 20.
(59) Smulders, M. M. J.; Nieuwenhuizen, M. M. L.; de Greef, T. F.
A.; van der Schoot, P.; Schenning, A. P. H. J.; Meijer, E. W. Chem.
Eur. J. 2010, 16, 362.
(60) ten Eikelder, H. M. M.; Markvoort, A. J.; de Greef, T. F. A.;
Hilbers, P. A. J. J. Phys. Chem. B 2012, 116, 5291.
(61) Markvoort, A. J.; ten Eikelder, H. M. M.; Hilbers, P. A. J.; de
Greef, T. F. A.; Meijer, E. W. Nat. Commun. 2011, 2, 509.
(62) Huang, X.; Raghavan, S. R.; Terech, P.; Weiss, R. G. J. Am.
Chem. Soc. 2006, 128, 15341.
(63) Because of the low solubility and the presence of multiple
stereocenters in 3, reliable CD spectra (under thermodynamic
control) could not be achieved; thus, temperature-dependent UV/vis
studies were performed.
(64) The isodesmic fits to the cooling curves for 3 and 4 show slight
deviations from the experimental data. Since these cooling curves are
obtained from UV/vis studies, the deviation could be due to the small
change in absorbance at the monitored wavelength. The fit is sub-
isodesmic in the case of 3; thus, this molecule can be weakly
cooperative, which adds further to the deviation from the fit.
(65) Brocorens, P.; Linares, M.; Guyard-Duhayon, C.; Guillot, R.;
Andrioletti, B.; Suhr, D.; Isare, B.; Lazzaroni, R.; Bouteiller, L. J. Phys.
Chem. B 2013, 117, 5379.
̀
(25) García, F.; San
́
chez, L. J. Am. Chem. Soc. 2012, 134, 734.
(26) Mayerhoffer, U.; Wurthner, F. Chem. Sci. 2012, 3, 1215.
̈
̈
(27) Aparicio, F.; Matesanz, E.; Sanchez, L. Chem. Commun. 2012,
48, 5757.
(28) Fenske, M. T.; Meyer-Zaika, W.; Korth, H.-G.; Vieker, H.;
Turchanin, A.; Schmuck, C. J. Am. Chem. Soc. 2013, 135, 8342.
(29) Jonkheijm, P.; van der Schoot, P.; Schenning, A. P. H. J.; Meijer,
E. W. Science 2006, 313, 80.
̌
́
Z.; van Dongen, J.; George, S. J.; Xu, H.; Pisula, W.;
(30) Tomovic,
Leclere, P.; Smulders, M. M. J.; De Feyter, S.; Meijer, E. W.;
̀
Schenning, A. P. H. J. J. Am. Chem. Soc. 2007, 129, 16190.
(31) Mayoral, M. J.; Rest, C.; Stepanenko, V.; Schellheimer, J.;
́
Albuquerque, R. Q.; Fernandez, G. J. Am. Chem. Soc. 2013, 135, 2148.
(32) Krieg, E.; Weissman, H.; Shimoni, E.; Bar On, A.; Rybtchinski,
B. J. Am. Chem. Soc. 2014, 136, 9443.
(33) Rest, C.; Mayoral, M. J.; Fucke, K.; Schellheimer, J.; Stepanenko,
́
V.; Fernandez, G. Angew. Chem., Int. Ed. 2014, 53, 700.
(34) Li, L.; Yuan, C.; Dai, L.; Thayumanavan, S. Macromolecules
2014, 47, 5869.
(35) Kulkarni, C.; Balasubramanian, S.; George, S. J. ChemPhysChem
2013, 14, 661.
(36) Wurthner, F.; Yao, S.; Beginn, U. Angew. Chem., Int. Ed. 2003,
̈
42, 3247.
(37) Fernan
Chem.Eur. J. 2013, 19, 206.
(38) Wurthner, F.; Stolte, M. Chem. Commun. 2011, 47, 5109.
́
dez, G.; Stolte, M.; Stepanenko, V.; Wurthner, F.
̈
̈
(39) Wurthner, F. Chem. Commun. 2004, 1564.
̈
(66) Bejagam, K. K.; Fiorin, G.; Klein, M. L.; Balasubramanian, S. J.
Phys. Chem. B 2014, 118, 5218.
(67) Danila, I.; Riobe, F.; Piron, F.; Puigmartí-Luis, J.; Wallis, J. D.;
́
Linares, M.; Ågren, H.; Beljonne, D.; Amabilino, D. B.; Avarvari, N. J.
Am. Chem. Soc. 2011, 133, 8344.
(40) Ikeda, T.; Masuda, T.; Hirao, T.; Yuasa, J.; Tsumatori, H.;
Kawai, T.; Haino, T. Chem. Commun. 2012, 48, 6025.
(41) Wurthner, F.; Thalacker, C.; Diele, S.; Tschierske, C. Chem.
̈
Eur. J. 2001, 7, 2245.
(42) Chen, Z.; Stepanenko, V.; Dehm, V.; Prins, P.; Siebbeles, L. D.
(68) Marty, R.; Nigon, R.; Leite, D.; Frauenrath, H. J. Am. Chem. Soc.
2014, 136, 3919.
A.; Seibt, J.; Marquetand, P.; Engel, V.; Wurthner, F. Chem.Eur. J.
̈
2007, 13, 436.
(69) Roche, C.; Sun, H.-J.; Prendergast, M. E.; Leowanawat, P.;
Partridge, B. E.; Heiney, P. A.; Araoka, F.; Graf, R.; Spiess, H. W.;
Zeng, X.; Ungar, G.; Percec, V. J. Am. Chem. Soc. 2014, 136, 7169.
(70) Klein, M. L.; Shinoda, W. Science 2008, 321, 798.
(71) MD simulations were not performed for 2 because of its
structural similarity to molecule 1.
(43) van Herrikhuyzen, J.; Syamakumari, A.; Schenning, A. P. H. J.;
Meijer, E. W. J. Am. Chem. Soc. 2004, 126, 10021.
(44) Kaiser, T. E.; Stepanenko, V.; Wurthner, F. J. Am. Chem. Soc.
2009, 131, 6719.
(45) van der Weegen, R.; Korevaar, P. A.; Voudouris, P.; Voets, I. K.;
de Greef, T. F. A.; Vekemans, J. A. J. M.; Meijer, E. W. Chem. Commun.
2013, 49, 5532.
̈
(72) Senanayak, S. P.; Guha, S.; Narayan, K. S. Phys. Rev. B 2012, 85,
115311.
H
DOI: 10.1021/jacs.5b00504
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(73) Ashcroft, N. W.; Mermin, N. D. Solid state physics; Holt,
Rinehart and Winston: New York, 1976.
(74) The average dipole moment per molecule obtained from bulk
dielectric measurements for molecules 1, 3, and 4 is 3.4, 0.7, and 0.5 D,
respectively (see SI for details of calculation). The macro-dipole
moment obtained from MD simulation was divided by 40 (oligomer
size) to obtain the dipole moment per molecule; dipole moments from
simulations for 1, 3, and 4 are 0.36, 0.05, and 0.03 D, respectively.
(75) Senanayak, S. P.; Narayan, K. S. Adv. Funct. Mater. 2014, 24,
3324.
I
DOI: 10.1021/jacs.5b00504
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX