Different Gelation and Self-Sorting Properties of Two Isomeric Polyamides Owing to the Parallel versus Anti-Parallel Alignment of Backbone Dipoles

Article abstract of DOI:10.1002/chem.201605819

Two isomeric bottlebrush polyamides P-1 and A-1, with the same repeating monomer dipole units aligned along the polymer backbone in pseudo-parallel and pseudo-antiparallel fashion, respectively, were synthesized and characterized. Both polymers can form thermoreversible gels with aromatic solvents but P-1 was found to show inferior gelation strength compared with that of A-1. Furthermore, despite their close structural resemblance, a 1:1 mixture of the P-1 and A-1 polymers was shown to exhibit self-sorting in the gel state. Gel formation was found to be a kinetically trapped process through hydrogen bonding, π–π stacking interactions, and side chain interdigitation. The different gelation and self-sorting properties can be explained by the local dipole–dipole interactions originating from the different modes of backbone dipole alignment. In single gel systems, the antiparallel-aligned dipoles in A-1 facilitated a more compact molecular packing owing to the enthalpically more favorable polymer chain association. On the other hand, the parallel-aligned dipoles in P-1 gave rise to a less stable head-to-head packing, which had difficulties to convert to the more stable head-to-tail packing in a kinetically trapped environment. In the mixed gel system, it is the unfavorable hetero-polymer mismatch dipole–dipole interaction that inhibited the mixing of the A-1 and P-1 polymers and led to self-sorting.

Full text of DOI:10.1002/chem.201605819

A Journal of
Accepted Article
Title: Differential Gelation and Self-Sorting Properties of Two Isomeric
Polyamides Due to the Parallel vs Anti-Parallel Alignment of
Backbone Dipoles
Authors: Chui-Fan Leung and Hak-Fun Chow
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201605819
Link to VoR: http://dx.doi.org/10.1002/chem.201605819
Supported by

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
Differential Gelation and Self-Sorting Properties of Two Isomeric
Polyamides Due to the Parallel vs Anti-Parallel Alignment of
Backbone Dipoles
Chui-Fan Leung,[a] and Hak-Fun Chow*,[a]
Dedicated to Prof. Dieter Seebach on the occasion of his 80th
birthday
new materials with unique functions.[2]
The gelation of macromolecules can be considered as a
special case of self-assembly, in which the molecules
themselves not just interact with each other to form aggregates,
but the resulting assemblies can create a network that spans the
volume of the liquid medium, and immobilizes the solvent
molecules to form solid-like gels. Depending on the nature of the
driving force that leads to gel formation, polymer gels can be
classified as chemical and physical gels. In the former case,
network formation is established through strong covalent
linkages of polymer chains, and therefore the resulting gel
cannot be reversibly converted back to the sol.[3] In contrast, the
driving forces that hold physical gels are non-covalent in nature,
which are weaker and can be broken down easily. Hence, the
gel can return to the sol state at elevated temperature. In the
literature, most polymer gels are chemical ones which are
prepared by extensive crosslinking of polyfunctional monomers.
On the other hand, examples of polymer physical gels are
fewer.[4] In contrast to low molecular weight gelators, in which
the gelation mechanism has been extensively studied,[5] the
gelation mechanism of polymer physical gels is less understood.
This is an intrinsic problem when dealing with the structurally
more complex polymer systems, as there are many parameters,
such as multiple functional group and side-chain interactions,
backbone conformation and chain mobility that are in action. We
are interested in deciphering the gelation mechanism of
dendronized oligo- and poly(amide-triazole)s, and have
identified several factors that can influence their gelation
properties.[6] Hence, it was shown that the gel-to-sol melting
temperature (Tm) of a series of H-bond-mediated oligo(amide-
triazole) gels increased monotonically with increasing number of
amide units along the polymer chain (Figure 1a).[7] Poly(amide-
triazole)s with a structurally more rigid backbone held by
intramolecular H-bonds were more likely to be a gelator than
structural similar poly(amide-triazole)s without backbone rigidity
(Figure 1b).[8] In a study of the polymer side chain effect, it was
noted that the size of the side chain of a series of poly(amide-
triazole)s had an impeding effect on the strength of self-
association (Figure 1c).[9] Hence, a bulky side chain led to
polymer dissolution and a small side chain led to polymer
precipitation, and that an intermediate size side chain was the
optimal one to confer the resulting poly(amide-triazole) with
gelating properties. More intriguingly, the relative orientation of
the triazole dipole units along the polymer backbone also played
a subtle role in controlling their supramolecular properties
(Figure 1d).[10] Hence, the dendronized poly(amide-triazole) with
the triazole dipoles aligned in pseudo-parallel (hereafter parallel)
fashion along the polymer backbone showed different
organogelating properties as compared to the corresponding
isomeric dendronized poly(amide-triazole) with the triazole
dipoles aligned in pseudo-antiparallel (hereafter antiparallel)
manner. It was proposed that the presence of a macro-dipole
Abstract: Two isomeric bottlebrush polyamides P-1 and A-1 having
the same repeating monomer dipole units aligned along the polymer
backbone in pseudo-parallel and pseudo-anti-parallel, respectively,
were synthesized and characterized. Both polymers can form
thermoreversible gels with aromatic solvents but P-1 was found to
show inferior gelation strength as compared to that of A-1.
Furthermore, despite their close structural resemblance, a 1:1
mixture of the P-1 and A-1 polymers was shown to exhibit self-
sorting in the gel state. Gel formation was found to be a kinetically
trapped process via H-bonding, π−π stacking interactions and side
chain interdigitation. The differential gelation and self-sorting
properties can be explained by the local dipole-dipole interactions
originated from the different modes of backbone dipole alignment. In
single gel systems, the antiparallel-aligned dipoles in A-1 facilitated
a more compact molecular packing due to the enthalpically more
favorable polymer chain association. On the other hand, the parallel-
aligned dipoles in P-1 gave rise to a less stable head-to-head
packing, which had difficulties to convert to the more stable head-to-
tail packing in a kinetically trapped environment. In the mixed gel
system, it is the unfavorable hetero-polymer mismatch dipole-dipole
interaction that inhibited the mixing of the A-1 and P-1 polymers and
led to self-sorting.
Introduction
Self-assembly of macromolecules is a universal, complex yet
crucial phenomenon. It is a spontaneous behavior that produces
order in a system. Polymer crystallization, gelation, adsorption,
and phase separation can all be viewed as different
manifestations
of
macromolecular
self-assembly.
Mechanistically, this process is based on the collective and
cooperative non-covalent interactions of polymer molecules.
Understanding the self-assembly of bio-macromolecules, such
as that of amyloidogenic proteins/peptides, occurred in living
systems can give us a better clue on the cause of many neuro-
degenerative diseases.[1] In materials science, studying the
aggregation of conjugated polymers can help improve their
optoelectronic properties, and can illuminate ways to fabricate
[a]
Ms C.-F. Leung, Prof. Dr. H.-F. Chow
Department of Chemistry and Center of Novel Functional Molecules
The Chinese University of Hong Kong
Shatin, NT, Hong Kong SAR
E-mail: hfchow@cuhk.edu.hk
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
Figure 1. Examples of oligo- and poly-(amide-triazole)s that exhibit supramolecular gelation properties.
(resulting from the cumulative effect of the parallel-aligned
dipoles) in the parallel-aligned polymer and the lack of such a
macro-dipole in the antiparallel-aligned polymer led to the
observed differential gelation behavior. The cumulated macro-
dipole concept had been used to rationalize the quaternary
structure[11] and enhanced catalytic property[12] of many α-helical
proteins. However, recent research findings questioned the
validity of the macro-dipole concept, and pointed out that the
local dipole-dipole interaction is much more important to account
for the observed protein properties.[13] In this context, it is
deemed necessary to revisit the mechanism of dipolar polymer
gelation by focusing on the local dipole-dipole interactions. In
this paper we reported new and interesting findings on this
polymer backbone dipole alignment effect by studying the
gelating and self-sorting properties of two isomeric polyamides,
namely P-1 and A-1, with para-alkoxybenzamide dipole units
aligned in parallel and antiparallel fashions, respectively, along
the polymer backbone (Figure 2a). It was found that (a) both
polymers are good organogelators of aromatic solvents, but they
exhibit differential gelation strength, (b) gel formation involves
the kinetic trapping of polymer molecules, (c) the mode of
polymer backbone dipole alignment has a strong effect on the
gelation strength; the difference in their gelation properties can
be readily rationalized by the differential local dipole-dipole
interactions arising from the parallel vs antiparallel dipole
alignments, and agrees well with the corresponding enthalpic
and entropic parameters obtained from differential scanning
calorimetry (DSC) and diffusion wave spectroscopy (DWS), (d)
despite their close structural similarity, the P-1 and A-1
polyamides exhibited self-sorting during mixed gel formation,
and (e) the dipole-dipole interaction effect appears to be
general; previously it was demonstrated that the effect was
associated with the triazole unit, now it is also found with the
para-alkoxybenzamide dipole. The results reported here
highlight the multifaceted nature of polymer physical gels, in
which not only thermodynamic factors, but kinetic ones are also
of importance in determining their gelation mechanism.
Furthermore, the subtle but influential backbone dipole
alignment effect on the self-assembling and gelating properties
of polymers, a factor that is of less significance in low molecular
weight gelators due to the absence of molecular entanglement,
is clearly demonstrated.
Results and Discussion
Design of polymer structure
In order to evaluate the cumulative effect of backbone dipoles on
the supramolecular and gelation properties of the corresponding
polymer, it is important to set out several key design principles.
First, the backbone of the polymers should be relatively rigid in
order to minimize back folding.
A
bottlebrush polymer
architecture with several long alkyl side chains (i.e. R1 and R2)
were installed to ensure the backbone will possess a relatively
rigid conformation.[14] Second, the monomer dipole should be
strong enough to enable its effect to be observed. Third, the
This article is protected by copyright. All rights reserved.

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
property, the corresponding parallel- and antiparallel-aligned
dimers, P-5 and A-5, respectively, were also synthesized.
Structural Characterization
1. Nuclear magnetic resonance spectroscopy. All compounds
1
prepared were characterized by H and 13C NMR spectroscopy
and, except P-1 and A-1, by mass spectrometry. The
spectroscopic data agreed well with the proposed structures.
Due to their isomeric nature, the 1H NMR spectra of P-1 and A-1
are very similar, though careful observation revealed the
presence of some minor oligomeric species in A-1 (see SI
section 3, Figure S2). This was probably due to a slight deviation
from the exact 1:1 reaction stoichiometry between the AA 3 and
BB 4 monomers. Such lower molecular weight oligomers may
facilitate in triggering the gelation of A-1, but should pose little
influence on the thermodynamic and kinetic data obtained from
the experiments since they are present in small amount.
2. Gel permeation chromatography (GPC). The two polymers
were subjected to GPC analysis coupled to a multi-angle laser
light scattering (MALLS) detector in order to obtain the molecular
weight distribution and absolute molecular weight information
(Figure 3 and Table 1). As disclosed in the introduction section,
the gelation strength of the polymer is dependent on the total
number of H bonding units in the polymer chain. However, since
there is a plateauing effect once the number of H-bonding unit
becomes very large (~20), therefore there is no absolute need to
have the prepared polymers having exactly the same MW. As
can be seen, the two polymers possessed similar GPC profiles,
although that of A-1 contained larger amount of low molecular
weight oligomers. It should be noted that the relative % of the
low molecular weight oligomers is exaggerated in the plot as the
x-axis is on a logarithm scale of polymer Mw. The Mw (GPC
based on PS standards) values of P-1 and A-1 were comparable
(17.7 vs 14.6 kD), and their polydispersity indexes (PDI) were
both around 2.0 – a typical value expected from a step-growth
polymerization. The Mw values obtained from MALLS were found
to be 18.0 and 16.2 kD for P-1 and A-1, respectively.
Figure 2. (a) Structure of parallel and antiparallel aligned polymers and dimers.
(b) Orientation of the overall dipole and its vǁ and v┴ components.
dipole parallel-aligned and anti-parallel-aligned polymers should
be of similar molecular weight and distribution profile such that a
meaning comparison can be made. After putting these into
considerations, it was decided to use para-alkoxybenzamide as
the dipole unit, which was calculated to have a dipole moment of
4.9 Debye according to DFT calculation [see Supporting
Information (SI) section 1, Figure S1]. It is noted that the
orientation of dipole moment of this unit does not coincide with
that of the polymer backbone due to the vectorial contributions
from both the amide CONH and the para-alkoxyaryl dipoles
(Figure 2b). Nonetheless, the overall dipole could be resolved
into two components; one of which (vǁ) is parallel (i.e. backbone
dipole) and the other (v┴) is perpendicular to the polymer
backbone.[15] It should be noted that orientation of vǁ is fixed in
space along with that of the polymer backbone, while that of v┴
can tumble as a result of rotational freedom of the aryl−C(O)
bond. Hence, for both the P-1 and A-1 polymers, the density and
the orientation distribution of the v┴ components are essentially
the same. The differential gelation properties are therefore
principally due to the difference of the orientation distribution of
the vǁ components (i.e. parallel vs antiparallel backbone dipole
alignment). It is anticipated that the dipole moment will change
slightly upon H-bonding, but the above conclusions remain the
same. The resulting P-1 polymer was prepared via condensation
polymerization from an AB-type monomer 2, while the
corresponding A-1 polymer was obtained from condensation
copolymerization of AA-type 3 and BB-type monomers 4 (see SI
section 2 for details). For comparison of the infrared absorption
Figure 3. GPC profiles (THF, 40 °C) of P-1 and A-1 polymers.
This article is protected by copyright. All rights reserved.

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
analysis (see SI section 6, Figure S5). The FT-IR spectra of the
two dimers in 5% toluene solution showed two IR adsorptions at
3440 and 1668 cm-1, which could be attributed to the non-H-
bonded N−H and C=O amide stretching frequency, respectively.
On the other hand, these two peaks were red-shifted to 3280
and 1628 cm-1 in the 3% polymer gels in toluene. Hence gelation
was driven by H-bonding.
Table 1. Molecular weights and PDI of P-1 and A-1 polymers.
P-1
GPC-MALLS
A-1
GPC-UV
17.7
GPC-UV
14.6
GPC-MALLS
16.2
Mw (kD)
Mn (kD)
PDI
18.0
13.2
1.37
8.7
7.3
12.3
3. Differential scanning calorimetry (DSC). Having
established that the backbone dipole alignment did have an
effect on the self-assembly and hence the organogelating
strength of the resulting polymers, the thermodynamic
parameters of the gel-sol transitions were then extracted using
DSC analysis. The DSC traces of toluene gel samples of P-1
and A-1 were shown in Figure 4. At 3.0 % gelator concentration,
where both polymer samples were in the gel state, the heating
profile of P-1 showed one broad peak centered at 51.7 °C. The
broadness of the peak reflected the heterogeneous micro-
structure of the polymer gel. The DSC profile of A-1 was more
complex, giving two peaks at 53.5 and 64.7 °C. The peak at
64.7 °C was relatively sharp, suggesting this was due to the
melting of a more crystalline state. These two prominent
transition temperatures of A-1 were both higher than that of P-1.
For the 1.0% gel samples, the DSC curves were similar, even
though P-1 now appeared as a partial gel. It was noted that the
melting peak at 65.7 °C of A-1 was further sharpened. On the
other hand, the DSC cooling curves (cooling rate = 1 C/min)
looked simpler. The gelation temperature (Tgel) now appeared at
35.9 and 32.3 °C for the 3% and 1% P-1 samples, respectively.
For the 3% and 1% A-1 samples, the Tgel values were 56.5 and
54.3 °C, respectively. This result indicated the presence of
hysteresis and is typical of physical gels,[16] highlighting that the
formation of such polymer physical gels via ambient temperature
cooling is a kinetic-trapping process.[17] Indeed, when the cooling
rate was increased to ~4 C/min in the DSC scans, the melting
temperature of both P-1 and A-1 decreased by about 1−2 C
(see SI section 7.II, Figure S6). Apparently, the slightly faster
cooling rate (4 C/min) did not provide the polymer chains with
enough time to search for a more stable crystalline state. The
relatively complex DSC profiles also confirmed that the gel-sol
process actually involved several phase transitions instead of a
single one. This observation is consistent with the widely
accepted model that polymer physical gels consist of crystalline
regions (also known as junction zones) as well as amorphous
regions composed of dangling chains.[18]
2.04
2.01
1.31
Organogelation property
1. Minimum gelation concentration and gel-to-sol transition
temperature. The two target polymers were found to form
strong organogels with most aromatic solvents (Table 2). The
gels were prepared by ambient temperature cooling of boiling
solutions of the polymer, and gelation generally took place within
5 min. The minimum gelation concentration (MGC) values of A-1
are consistently lower than those of P-1 in all aromatic solvents
examined. A crude estimation of the gelation strength was
obtained by plotting the gel-to-sol melting temperature (Tm) in
para-xylene at different gelator concentrations, and the result
also confirmed the better gelation strength of A-1 (see SI section
5, Figure S4). It was noted that significant difference in the Tm
(~30 °C) was found at a lower gelator concentration (3%). Hence,
preliminary results indicated that the monomer dipole alignment
did affect the self-association strength of the gel network.
Table 2. MGC values of P-1 and A-1 in various solvents.[a]
Solvent
P-1
A-1
THF
S
S
DMF
S
S
DMSO
S
S
CHCl3
S
S
Ethyl acetate
Chlorobenzene
ortho-Dichlorobenzene
Toluene
2% (OG)
5% (OG)
S
5% (TG)
3% (CG)
S
3% (TG)
3% (TG)
5% (TG)
1% (CG)
1% (CG)
2% (CG)
para-Xylene
Anisole
[a] S: very soluble; OG = opaque gel; TG = transluent gel; CG =
transparent gel.
2. Fourier-transformed infrared spectroscopy (FT-IR). To
confirm that H-bonding interaction was responsible for the
gelation, the two polymer samples P-1 and A-1 as well as the
non-gelating dimers P-5 and A-5 were subjected to FT-IR
This article is protected by copyright. All rights reserved.

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
parameters including binding energy, number of binding sites,
density and rotational freedom of zippers which were taken as
rigid rods in crystalline junction zones. Based on statistical
mechanics, the observed concentration effect on ΔHm is closely
related to the zipper density. Hence, the reduction of ΔHm could
be attributed to higher chain mobility and fewer crystalline
junction zones at a lower gelator concentration.
4. Powder small angle X-ray diffraction analysis (SAXRD).
The two polymers were subjected to SAXRD analysis in order to
reveal their packing arrangement of the crystallization zone at
the molecular level. The SAXRD pattern of a 3% P-1 xerogel in
para-xylene showed two peaks at 2θ = 4.2° and 19.2° (Table 4
and SI section 8, Figure S9). The prominent peak at 4.2o was
originated from a spacing of 21.0 Å, which was slightly shorter
than twice the span of the 5-butylundecyl side chain (~2 × 13 Å).
The other peak at 19.2o was much broader, and corresponded to
a d-spacing of 4.6 Å, which fell into the range of inter-atomic
distance of π−π stacking. This result supported the presence of
the more crystalline interdigitation of the branched hydrophobic
side chains, in addition to the slightly amorphous packing of the
backbone aromatic moieties. For A-1, whilst the inter-distance of
π−π stacking was indistinguishable from that of P-1, the
crystalline hydrophobic spacing was found to be shorter (18.8 Å).
The tighter molecular packing of A-1 could then account for its
higher ΔSm value obtained in the DSC experiment.
Figure 4. DSC heating (1.0 °C/min) and cooling curves (1.0 °C/min) of (top)
3% toluene gels and (bottom) 1% toluene gels of P-1 and A-1 polymers.
Table 3. Enthalpy (ΔHm) and entropy changes (ΔSm) of melting of toluene
gels of P-1 and A-1 polymers.[a]
Table 4. SAXRD Data of xerogels prepared from P-1 and A-1.
ΔHm (Jg−1
)
ΔSm (Jg−1K−1
)
First peak
d-spacing (Å)
21.0
Second peak
d-spacing (Å)
4.61
2θ (°)
4.21
4.70
2θ (°)
19.23
19.24
P-1 (3% in toluene)
A-1 (3% in toluene)
P-1 (1% in toluene)
A-1 (1% in toluene)
25.8
31.4
15.2
18.2
79.6 × 10−3
96.7 × 10−3
47.8 × 10−3
56.1 × 10−3
P-1
A-1
18.8
4.61
[a] Gels were aged at 20 °C for 15 minutes prior to measurements. The
integration region was taken from 20 to 80 °C, which encapsulated all the
various phase transitions.
The respective enthalpy changes (ΔHm) and entropy
changes (ΔSm) of the gel melting process were tabulated (Table
3). It was found that melting of the A-1 gel, which
macroscopically behaved as a better organogelator, was more
endothermic than that of the P-1 gel at the same concentration,
i.e. self-assembly of A-1 led to a thermodynamically more stable
gel aggregates than that of P-1. In addition to having a larger
ΔHm, the melting of the A-1 gel also exhibited a larger ΔSm value,
suggesting that packing of the A-1 gel was more order than that
of the P-1 gel.
Both the ΔHm and ΔSm values of 1.0% gels of polymers were
smaller than those obtained at 3.0%, as opposed to the
assumption of constant melting enthalpy change predicted by
the Eldridge and Ferry’s equation,[19] but this could be
qualitatively accounted for by a zipper model proposed by
Nishinari.[20] This zipper model correlated ΔHm with a number of
This article is protected by copyright. All rights reserved.

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
Figure 6. Differential self-association modes of P-1 and A-1 polymers.
The differential gelation properties observed for P-1 and A-1
must be due to a difference of their macromolecular geometry
and/or electronic environment, which led to the formation of
clusters of different packing order and stability. However, there
is little difference of the spacing between the R1 and R2 polymer
side chains in P-1 and A-1, and therefore the geometry factor
should be less influential. On the other hand, the dipole
alignment effect should not be ignored due to its directional and
multiplicative nature. Adhered to our original objective, the local
dipole-dipole interaction instead of the macro-dipole interaction,
will be used to rationalize the observed behavior. There are two
possible packing modes for P-1, namely, the energetically more
favorable head-to-tail (Figure 6a) and the less favorable head-to-
head (Figure 6b) associations. However, in a kinetically arrested
gel system, there is little chance for one of the chains in the
head-to-head mode to tumble over in order to attain the more
stable head-to-tail association. Hence, gels prepared from P-1
should possess both head-to-tail and head-to-head packings.
The unfavorable head-to-head arrangement would prevent close
interaction between the polymer chains, and was consistent with
its less-compact packing nature as shown by the SAXRD data
and its smaller enthalpy of melting (ΔHm) value from DSC data.
On the other hand, there is only one association mode for A-1.
When the initial contact was unfavorable due to dipole-dipole
repulsion (Figure 6c), one of the A-1 polymer chains can slither
over the distance of one unit of dipole separation to reach the
more stable packing (Figure 6d). The polymer chain does not
need to tumble the whole molecule over to attain an
enthalpically more favorable association. This rationale was also
consistent with the finding that the entropy of melting (ΔSm) of
toluene gel of A-1 is larger than that of P-1, since P-1 can pack
either in head-to-tail or head-to-head fashion, and is therefore
less order or entropically higher as compared to A-1 gel, as the
latter could pack in a uniformly head-to-tail fashion. However, it
should be emphasized here that the packing outcomes were
mainly due to the kinetic entrapment of large polymer molecules
in the gel state, and that the same conclusions may not be valid
for low molecular weight gel systems, in which the molecular
mobility is much higher.
Figure 5. (a) Purposed model of the crystallization zone of the polymer gels of
A-1 and P-1 and (b) Schematic diagram (side chains are not drawn) showing
the formations of gel network via cluster aggregation, cluster formation via
bundle association and formation of bundles via polymer chain aggregation
(aromatic solvent molecules are omitted for clarity).
5. Gel packing model. On the basis of the above results, a
packing model of the polymer gels is proposed (Figure 5). First,
the backbone of the polyamides is stiffened due to repulsions
among the bulky hydrocarbon side chains, and thus the
bottlebrush-shaped macromolecule exists in an extended worm-
like conformation. The polymer chains then aggregate through
strong H-bonding and π−π stacking interactions along the
backbone to produce bundles with an inter-chain spacing of
approximately 4.6 Å. Interdigitation of the hydrophobic side
chains of the woven fibrillar layers then gives rise to a higher-
order layered micro-crystalline cluster in which the inter-layer
distance takes the value of approximately 20 Å. The aromatic
solvent molecules are entrapped within the matrix via π−π
stacking interaction with the aromatic moieties of the polymer
and/or with other solvent molecules. These crystalline clusters of
different sizes then loosely linked together to form the gel
network. Melting of the crystalline zone will then lead to a
discontinuous enthalpy change and appears as
a sharp
transition, while the enthalpy required for cluster-cluster
movement is small and heterogeneous, giving rise to rather
broad peaks in the DSC profile.
This article is protected by copyright. All rights reserved.

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
6. Scanning electron microscopy (SEM). Xerogels of the
polymers were prepared and subjected to SEM examinations in
order to reveal the morphology of their microstructures. All
xerogels appeared as clusters of random fibrillar layers
throughout the entire structures (see SI section 9, Figure S10).
Scrutiny of their SEM images unraveled that A-1 tended to have
a few fibrous structures whereas P-1 possessed slightly denser
morphology.
hydroxyl groups on the silica. The size of the microspheres was
comparable to the mesh size of the gel network as observed by
SEM such that the measured results reflected the change of the
gel network structure at the micron scale, which was equivalent
to the dimension of the bundle structure inside a large cluster.
As the density of the silica particle was about twice as that of
toluene, the non-sticky microspheres tended to sediment in pure
toluene. However, in the presence of the polymer gelator, the
silica particles were able to disperse homogeneously to form a
turbid suspension due to the gel network structure.
Preliminary results carried out on 3% gel samples found that
the silica beads were too rigidly bound even at the high
temperature region (70 °C). Hence only 1.0% partial gel samples
of the polymers were subjected to detailed DWS analysis. The
mean square displacement (MSD) vs lag time plots at various
temperatures were shown (Figure 7). First, both P-1 and A-1 did
not show significant change of MSD even at lag time of 10 s at
25 °C. Secondly, the increment in MSD intensified as the
temperature increased, indicating that the transient constraints
(H-bonding, π−π stacking and hydrophobic interactions)
surrounding the microspheres were loosened at higher
temperature, which correlated to bundle-bundle crosslink
dissociation. Thirdly, at 25 C, the MSD value (10─8 μm2) of the
A-1 gel was 10 times smaller than that (10─7 μm2) of the P-1 gel,
while at higher temperature (55 C), rather surprisingly, a
reversal of this trend was observed. However, it was also noted
that as the temperature increased, the gel network collapsed
and the silica particles began to precipitate. Hence DWS data
obtained at higher temperature and long lag time may subject to
a higher degree of uncertainty.
7. Rheological studies. To further understand the
macromolecular motion in the gel state, classical bulk and micro-
rheology (diffusion wave spectroscopy) were carried out. The
former offers insights on the macroscopic gel mechanical
properties at 10−100 microns that are typically the scales of
cluster-cluster sliding, while the latter examines the microscopic
flow and creeps down to hundredth to several microns, which is
within the domain of bundle-bundle interactions.
7A. Bulky rheology. For both 1.0% and 3.0% toluene gels of
the two polyamides, the storage modulus (G’) values are about
5−10 times of the corresponding loss modulus (G”) at the same
concentration, manifesting their solid-like behavior (Table 5 and
SI section 10, Figure S11). The obtained G’/G” ratios were
consistent with the claim by Almdal, who suggested that a gel
should possess a G’ value greater than the corresponding G”
value at least by one order of magnitude.[21] As expected, gels
possessed a higher modulus at a higher concentration, as
explained by the increase in density of the non-covalent
crosslinks. It was also found that the storage modulus of gels of
A-1 was higher than that of P-1 at the same concentration.
Structural strength of the noncovalent network could be also
estimated by the flow point at which the structure was
irreversibly deformed and the gel started to flow (i.e. the
crossing over point of curve G’ to curve G”). Indeed, gels of A-1
could sustain three- to six-fold more stress than that of P-1 at
the same concentration.
Table 5. Rheological data of of P-1 and A-1 polymer gels in toluene.[a]
LVE[b]
G‘ (Pa)
Flow point[c]
G“ (Pa)
152
Stress (Pa)
23.2
P-1 (3%)
A-1 (3%)
P-1 (1%)
A-1 (1%)
946
4339
194
728
77.5
Figure 7. MSD values as a function of lag time in (a) 1.0% toluene partial gel
of P-1 and (b) 1.0% toluene gel of A-1.
35
6.3
1833
268
42.0
The MSD data could also be transformed to oscillatory
rheograms by the generalized Stokes-Einstein equation. The
transition from stagnant behavior at short lag time to diffusive
behavior at long lag time resembled the frequency crossovers in
the terminal flow region, where the liquid character prevailed
over solid character as the frequency decreased (see SI section
11, Figure S12). The crossover frequency marked the
characteristic relaxation time of a network, and could be inferred
to the rate of dissociation of the transient crosslinks. In general,
the crossover frequency increased with temperature as the
[a] Measurements were done at 25 °C. [b] LVE = linear viscoelastic region,
G’ and G” were taken as the average over the LVE region. [c] Flow points
were determined by the cross-over point of G’ and G” curves.
7B. Diffusion wave spectroscopy (DWS). Dimethylsilyl-coated
silica microspheres of 1.5 μm in diameter were chosen and
imbedded in the toluene gels in the study. The dimethylsilyl
groups served to circumvent the interference of H-bonding
interaction between the amide moieties on the polymer and the
This article is protected by copyright. All rights reserved.

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
crosslinks were loosened by thermal agitation. The
corresponding relaxation times were then fitted into the Eyring
equation with good correlation coefficients (r2 = 0.85−0.93). The
obtained activation enthalpy (ΔH‡) and entropy (ΔS‡) reflected
the kinetics of the bundle-bundle breaking down process (Table
6 and SI section 11, Figure S13). The activation enthalpy (ΔH‡)
of the breaking down of the transient crosslinking of P-1 (2.8
Jg−1) was found to be lower than that of A-1 (4.6 Jg−1), whereas
the activation entropy (ΔS‡) for P-1 and A-1 was −5.1 and −0.87
mJK−1g−1, respectively. The negative values of ΔS‡ implied that
the transition states had higher packing order than the ground
states, which can be rationalized by the fact that the prepared
gels were in a kinetically-trapped, less order state during
preparation by ambient temperature cooling. As a result, the
various dipole-dipole and H-bonding interactions were not
geometrically optimized and hence the ground states were
structurally less ordered. The fact that P-1 possessed a
significantly larger drop in activation entropy (i.e. ΔS‡ is more
negative) towards dissociation indicated its gel packing was
much less order than that of A-1, a fact that was already
confirmed by SAXRD and gel strength measurements described
earlier. It should be noted that phase transitions probed by the
DWS method are different from those by DSC measurement.
The latter determines the total enthalpy and entropy of all phase
transition processes, while the former focuses only on the
corresponding processes originated from bundle-bundle
interaction.
Figure 8. Proposed model of dissociation and association of transient bundle-
bundle interaction in (a) P-1 and (b) A-1 polymers.
Self-Sorting of P-1 and A-1 polymers
Formation of mixed gel systems has already been reported for
small molecules,[22] polymers and biopolymers.[23] The problem is
complicated by whether the individual component is a gelator on
its own or not. Various models had been proposed to predict the
property of the resulting mixed gel systems. These include the
non-self-association,[24] single-self and cross-association,[25] and
double-self and cross-association models.[26] One key question
of such mixed gel systems is whether the components are fully
miscible with each other or self-sort which results in phase
separation. Generally, the greater the structural difference
between the two molecular species, the higher the tendency of
the mixed system to exhibit self-sorting. This is because greater
structural difference will normally lead to a higher chance of
having immiscible functionalities in the two components. While
self-sorting is a commonplace in small molecules, it is less
prevalent in synthetic macromolecules. Notable examples are
narcissistic self-sortings of norbornene copolymers,[27] of
bisurea-based bolaamphiphilic polymers[28] and of bisurea-based
thermoplastic elastomers.[29] On the other hand, example of
social self-sorting has also been reported in the case of bisurea-
based thermoplastic elastomers.[30] We earlier disclosed that
oligo(triazole-amide)s of significantly different number of H-bond
units did exhibit self-sorting in a 1:1 mixed gel system.[7] In the
case of the P-1 and A-1 polymers reported here, we are
therefore interested to know whether a mere difference in the
monomer dipole alignment can promote self-sorting in their
mixed gel systems.
Table 6. Activation enthalpy (ΔH‡) and entropy (ΔS‡) of intra-cluster
crosslink dissociation of 1% P-1 and A-1 polymer partial gels in toluene.[a]
ΔH‡ (Jg-1) [b]
2.80 ± 0.57
4.60 ± 0.72
ΔS‡ (mJK-1g-1) [b]
−5.10 ± 1.56
−0.87 ± 0.18
P-1
A-1
[a] Only the data obtained from 25−60 °C (before tracer particles completely
sank) were fitted into Eyring equation. Error terms were calculated at one
standard error. [b] Quantities relative to per gram of polyamide.
The lower ΔH‡ value of P-1 towards inter-bundle sliding could be
again rationalized by the dipolar model proposed earlier (Figure
8). Upon application of shear force, the chains were disengaged
and eventually manifested a liquid flow. If the P-1 chains are
aligned in the less stable head-to-head fashion, chain sliding is
enthalpically favorable due to the already repulsive interaction
between the chains (not shown in Figure 8). If the chains are
aligned in a head-to-tail fashion, the P-1 chain only needed to
slither one dipole unit to attain the next stable configuration
(Figure 8a), whereas for A-1 chains, there existed strong dipolar
repulsion in the course of sliding over one dipole unit, and a
stable state could only be attained when the chain slipped over
two dipole units (Figure 8b). As a result, crosslink dissociation
was enthalpically disfavored for A-1.
Preliminary examination of the DSC heating curve of the 1:1
mixed gel did not give any conclusive evidence due to multiple
phase transitions that were very close to each other. However,
self-sorting was highlighted by the cooling curve (Figure 9). Two
distinct peaks of similar size and shape were observed. If the
mixture was homogeneous, one average broad transition peak
would have been expected. As it turned out, the transition
maxima of the mixed gel were shifted towards lower temperature
in comparison to their respective pure gels, suggesting that the
This article is protected by copyright. All rights reserved.

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
inter-molecular and inter-cluster interactions were slightly
disrupted. The efficacy of self-sorting could be reflected by the
magnitude of enthalpy and entropy changes of gelation (Table 7).
As can be seen, the magnitude of the enthalpy of gelation from
the 1:1 mixture gel (−17.7 Jg−1) was higher than that of the
average of pure P-1 and A-1 (−22.5 Jg−1). Hence the mixing
process resulted in a positive rise of enthalpy (4.7 Jg−1).
Incidentally, the mixing process also gave a positive rise in
entropy (14.5 × 10−3 Jg−1K−1). At 25 °C, the rise of the Gibbs free
energy of mixing can be calculated as +0.48 Jg−1, indicating the
mixing was an unspontaneous process (additional experimental
results of the self-sorting property can be found in SI, section
7.III).
Figure 10. (left) Unfavorable dipole-dipole interactions between P-1 and A-1
polymers. (right) Such unfavorable interactions still persist even after chain
slippage.
Conclusions
Two isomeric polyamides P-1 and A-1, with the backone-dipoles
aligned in parallel and antiparallel fashion, respectively, were
successfully prepared and characterized. The two polymers
were found to form strong physical gels with a wide variety of
aromatic solvents. Despite their close structural similarity, Tm,
DSC and bulky rheology measurements all revealed that the
gels obtained from A-1 were consistently stronger than those
from P-1. Based on FT-IR and SAXRD experiments, gelation
was due to H-bonding, π−π stacking interactions and
hydrophobic side chain interdigitation. DSC study also revealed
that the polymer gels so prepared by ambient temperature
cooling were in a kinetically trapped state. Gelation involved first
the association of individual polymer chains to produce bundles
which subsequently formed crystalline clusters of different sizes.
These clusters then further aggregated in an amorphous manner
to produce the bulk gel network. DWS study found that the
activation enthalpy of the intra-cluster melting process of A-1
was higher than that of P-1. Incidentally, the two polymers were
also found to exhibit self-sorting in a 1:1 mixed gel system. The
differential gelation properties and the self-sorting phenomenon
could be readily rationalized by local dipole-dipole interactions
that were originated from the parallel or antiparallel alignment of
the backbone dipoles. As it appears, there is no need to invoke
the macro-dipole model in order to explain the observed
differential properties. Admittedly this may be a simplified model,
but it serves the purpose to rationalize the property of a polymer
system which is less complicated than that of a protein system.
In the protein case, additional dipolar influences from the polar
side chains and polar water molecules cannot be simply ignored.
The results reported here indicating that local dipole-dipole
interaction is a critical factor that can strongly affect the self-
association and supramolecular properties of polymers.
Furthermore, a subtly small change in the polymer structure,
such as reversing the direction of the dipole alignment, can lead
to highly efficient molecular self-sorting in polymer gels.
Figure 9. DSC exotherms of 1:1 mixed gel of P-1 and A-1 (blue),
superimposed with that of pure gels of P-1 (black) and A-1 (red) at 3.0%
concentration.
Table 7. Enthalpy and entropy changes of gelation of 3% toluene gel of P-1,
A-1 and their 1:1 mixture.[a]
ΔHg (Jg−1
)
ΔSg (Jg−1K−1
)
P-1 (3% in toluene)
−20.1
−24.8
−17.7
−22.5
−65.6 × 10−3
−77.7 × 10−3
−57.2 × 10−3
−71.7 × 10−3
A-1 (3% in toluene)
1:1 mixed gel of P-1 and A-1 (3% in toluene)
Average value of pure P-1 and A-1 (3% in
toluene)
[a] Gels were annealed at 80 °C for 15 minutes prior to measurements. The
integration region was taken from 20−80 °C.
The origin of the unfavorable mixing of the two polymers can
be rationalized by the unfavorable dipole-dipole interaction due
to hetero-association, leading to reduced inter-chains and inter-
clusters association (Figure 10). The repulsion was arisen from
mismatches of dipole motifs when the parallel and antiparallel
strands are closed to each other, and it also served as an
impetus to drive self-sorting. Sliding of the polymer strands
cannot eliminate such unfavorable interactions.
Experimental Section
Details of the polymer syntheses, compound characterizations and
gelation property studies are given in the electronic supplementary
information (SI).
Acknowledgements
This article is protected by copyright. All rights reserved.

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
[11] R. P. Sheridan, R. M. Levy, F. R. Salemme, Proc. Natl. Acad. Sci. USA
1982, 79, 4545–4549.
We thank the Research Grants Council (RGC Ref No 400712)
for the financial support, Ms Yan-Lam Lau for synthesizing some
of the intermediates, and Profs. To Ngai and Zhaoyan Sun for
their helpful discussion.
[12] W. G. J. Hol, P. T. van Duijnen, H. J. C. Berendsen, Nature 1978, 273,
443–446.
[13] E. G. Baker, G. J. Bartlett, M. P. Crump, R. B. Sessions, N. Linden, C.
F. J. Faul, D. N. Woolfson, Nat. Chem. Biol. 2015, 11, 221–228.
[14] S. S. Sheiko, B. S. Sumerlin, K. Matyjaszewski Prog. Polym. Sci. 2008,
33, 759–785.
Keywords: supramolecular chemistry • polyamide • physical gel
• self-sorting • dipole-dipole interaction
[15] Polymer chains containing only vǁ dipoles are called type-A chains;
while those containing solely v┴ dipoles are called type–B chains. The
dipolar effects of type-A and type-B chains on their properties had been
reported, see a) W. H. Stockmayer Pure. Appl. Chem. 1967, 15, 539–
554; b) K. Adachi, T. Kotaka, Prog. Polym. Sci. 1993, 18. 585–662.
[16] a) T. Sakai, J.-i. Horinaka, T. Takigawa, Polym. J. 2015, 47, 244–248;
b) A. H. Clark, S. B. Ross-Murphy, Adv. Polym. Sci. 1987, 83, 57–192.
[17] F. M. Menger, K. L. Caran, J. Am. Chem. Soc. 2000, 122, 11679–
11691.
[1]
a) J. D. Sipe, A. S. Cohen, J. Struct. Biol. 2000, 130, 88–98; b) J.-C.
Rochert, P. T. Jr. Lansbury, T. Peter, Curr. Opin. Struct. Biol. 2000, 10,
60–68; c) F. Chiti, C. M. Dobson, Ann. Rev. Biochem. 2006, 75, 333–
366.
[2]
D. Wang, Y. Yuan, Y. Mardiyati, C. Bubeck, K. Koynov,
Macromolecules 2013, 46, 6217−6224; b) Y. Yao, H. Dong, W. Hu,
Polym. Chem. 2013, 4, 5197−5205.
[3]
[4]
G. Tillet, B. Boutevin, B. Ameduri, Prog. Polym. Sci. 2011, 36, 191–217.
For reviews and monographs on thermoreversible gelation of polymers
and biopolymers, see: a) J. Spěváček, B. Schneider, Adv. Colloid
Interface Sci. 1987, 27, 81–150; b) J.-M. Guenet, Thermoreversible
Gelation of Polymers and Biopolymers, Academic Press, London, 1992;
c) M. Suzuki, K. Hanabusa, Chem. Soc. Rev. 2010, 39, 455–463; d) A.
Noro, M. Hayashi, Y. Matsushita, Soft Matter 2012, 8, 6416–6429.
For reviews and monographs on low molecular weight gels, see: a) P.
Terech, R. G. Weiss, Chem. Rev. 1997, 97, 3133–3160; b) N. M.
Sangeetha, U. Maitra, Chem. Soc. Rev. 2005, 34, 821–836; c) R. G.
Weiss, P. Terech, Eds. Molecular Gels. Materials with Self-Assembled
Fibrillar Networks, Springer, Dordrecht, The Netherlands, 2005; d) M.
George, R. G. Weiss, Acc. Chem. Res. 2006, 39, 489–497; e) P.
Dastidar, Chem. Soc. Rev. 2008, 37, 2699–2715.
[18] a) P. J. Flory, Faraday Discuss. Chem. Soc. 1974, 57, 7–18; b) K. te
Nijenhuis, Adv. Polym. Sci. 1997, 130, 1–235.
[19] J. E. Eldridge, J. D. Ferry, J. Phys. Chem. 1954, 58, 992–995.
[20] K. Nishinari, S. Koide, P. A. Williams, G. O. Phillips, J. Phys. France
1990, 51, 1759–1768.
[21] K. Almdal, J. Dyre, S. Hvidt, O. Kramer, Polym. Gels Netw. 1993, 1, 5–
17.
[5]
[22] L. E. Buerkle, S. J. Rowan, Chem. Soc. Rev. 2012, 41, 6089–6102.
[23] Biopolymer Mixtures; S. E. Harding, S. E. Hill, J. R. Mitchell, Eds.;
Nottingham University Press, Nottingham, England, 1995.
[24] C.-F. Mao, J.-C. Chen, J. Appl. Polym. Sci. 2006, 99, 2771–2781.
[25] C.-F. Mao, J. Polym. Sci. Polym. Phys. 2008, 46, 80–91.
[26] Y. Park, B. Veytsman, M. Coleman, P. Painter, Macromolecules 2005,
38, 3703–3707.
[6]
[7]
H.-F. Chow, T.-K. Chui, Q. Qi, Synlett 2014, 25, 2246–2255.
J. Zhang, H.-F. Chow, M.-C. Chan, G. K.-W. Chow, D. Kuck, Chem. Eur.
J. 2013, 19, 15019–15025.
[27] a) C. Burd, M. Weck Macromolecules 2005, 38, 7225–7230; b) C. R.
South, C. Burd, M. Weck, Acc. Chem. Res. 2007, 40, 63–74.
[28] A. Pal, S. Karthikeyan, R. P. Sijbesma J. Am. Chem. Soc. 2010, 132,
7842–7843.
[8]
[9]
S.-L. Yim, H.-F. Chow, M.-C. Chan, C.-M. Che, K.-H. Low, Chem. Eur.
J. 2013, 19, 2478–2486.
[29] N. E. Botterhuis, S. Karthikeyan, A. J. H. Spiering, R. P. Sijbesma
Macromolecules 2010, 43¸ 745–751.
K.-N. Lau, H.-F. Chow, M.-C. Chan, K.-W. Wong, Angew. Chem. 2008,
120, 7018–7022; Angew. Chem. Int. Ed. 2008, 47, 6912–6916.
[30] E. Wisse, L. E. Govaert, H. E. H. Meijer, E. W. Meijer Macromolecules
2006, 39, 7425–7432.
[10] H.-F. Chow, K.-N. Lau, M.-C. Chan, Chem. Eur. J. 2011, 17, 8395–
8403.
This article is protected by copyright. All rights reserved.

10.1002/chem.201605819
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
Text for Table of Contents
Author(s), Corresponding Author(s)*
Page No. – Page No.
Title
((Insert TOC Graphic here: max.
width: 5.5 cm; max. height: 5.0 cm))
Layout 2:
FULL PAPER
Chui-Fan Leung,[a] and Hak-Fun
Chow*,[a]
Page No. – Page No.
Differential Gelation and Self-Sorting
Properties of Two Isomeric
Polyamides Due to the Parallel vs
Anti-Parallel Alignment of Backbone
Dipoles
Two isomeric polyamides, one with the backbone dipoles aligned in antiparallel and
the other in parallel fashion, were found to show differential organogelation
properties. They also exhibited self-sorting in their 1:1 mixed toluene gel. These
observations can be rationalized by the different local dipole-dipole interactions
manifested in the pure gel and the mixed gel systems.
This article is protected by copyright. All rights reserved.