Mechanically induced gelation of a kinetically trapped supramolecular polymer

Article abstract of DOI:10.1021/ma502047h

The stimuli-induced gelation of a urethane-functionalized ditopic ureidopyrimidinone (UPy) compound is presented, and the mechanism by which the gelation proceeds is proposed. In a 40-120 mM solution in chloroform, the compound can exist in two different aggregated states, namely a low viscous mixture of (cyclic) oligomers or a fibrous gel. As evidenced by IR, NMR, and WAXS, the liquid state is stabilized by hydrogen bonds between the UPy and the back-folded chain, while the fibrous gel is stabilized by lateral hydrogen bonds within stacked UPy dimers. Controlled preparation techniques allow for pathway selection to arrive at one of both states. The remarkable long-term stability of the low viscous state (over 2 months for a 80 mM solution) is in contrast to the fast transformation into a gel by stirring in a few hours. Other mechanical stimuli like shaking, sonicating, and stirring for a shorter period, as well as freezing and thawing the solution, yield weaker gels than those obtained by long stirring. Heating the gels and slow cooling reversibly yield the nonviscous solution. This shows that the formation of UPy-urethane hydrogen bonds kinetically traps the UPy polymers, thereby preventing their lateral aggregation. The application of mechanical stress or freezing disrupts this interaction, allowing for the formation of a stacked nucleus on which further material can grow, eventually leading to gelation of the solution.

Full text of DOI:10.1021/ma502047h

Article
pubs.acs.org/Macromolecules
Mechanically Induced Gelation of a Kinetically Trapped
Supramolecular Polymer
Abraham J. P. Teunissen, Marko M. L. Nieuwenhuizen, Fransico Rodríguez-Llansola, Anja R. A. Palmans,
and E. W. Meijer*
Laboratory of Macromolecular and Organic Chemistry, Eindhoven University of Technology, P.O. Box 513, NL-5600 MB Eindhoven,
The Netherlands
*
S Supporting Information
ABSTRACT: The stimuli-induced gelation of a urethane-function-
alized ditopic ureidopyrimidinone (UPy) compound is presented,
and the mechanism by which the gelation proceeds is proposed. In a
4
0−120 mM solution in chloroform, the compound can exist in two
different aggregated states, namely a low viscous mixture of (cyclic)
oligomers or a fibrous gel. As evidenced by IR, NMR, and WAXS,
the liquid state is stabilized by hydrogen bonds between the UPy
and the back-folded chain, while the fibrous gel is stabilized by
lateral hydrogen bonds within stacked UPy dimers. Controlled
preparation techniques allow for pathway selection to arrive at one
of both states. The remarkable long-term stability of the low viscous
state (over 2 months for a 80 mM solution) is in contrast to the fast
transformation into a gel by stirring in a few hours. Other
mechanical stimuli like shaking, sonicating, and stirring for a shorter period, as well as freezing and thawing the solution, yield
weaker gels than those obtained by long stirring. Heating the gels and slow cooling reversibly yield the nonviscous solution. This
shows that the formation of UPy−urethane hydrogen bonds kinetically traps the UPy polymers, thereby preventing their lateral
aggregation. The application of mechanical stress or freezing disrupts this interaction, allowing for the formation of a stacked
nucleus on which further material can grow, eventually leading to gelation of the solution.
INTRODUCTION
Recently, our group reported on the autoregulatory behavior
of a urethane-functionalized ditopic ureidopyrimidinone
(urethane-UPy) compound 1 (Figure 1) on the catalytic
activity of 2,7-diamido-1,8-naphthyridine in a Michael
■
External stimuli such as temperature, pH, or light are widely
used to influence material properties. Particularly interesting
1
−3
are mechanical forces; the aligning of polymeric chains by
18
4
addition. During the course of these investigations, we
stretching, inducing symmetry breaking by stirring a solution
5
,6
discovered that solutions of compound 1 in CHCl became
3
during crystallization,
influencing the charge transport
thixotropic under the influence of mechanical agitation or
freezing. Preliminary molecular modeling studies (Figure 1)
suggested that this observation can be related to the presence of
an intramolecular UPy−urethane hydrogen bond in mono-
meric 1.
We herein report our detailed studies on the stimuli-induced
gelation of urethane-UPy 1, with the aim to understand the
mechanism by which the gelation proceeds. As a reference, we
study ester-UPy 2, which lacks the urethanes and, as a result,
also the ability to form a similar intramolecular hydrogen bond.
Remarkably, the application of different types of stimuli on
solutions of 1 (temperature or mechanical forces) results in
different outcomes: either a nonviscous solution or gels of
different strengths (Figure 2). The different states of 1 were
studied with a combination of NMR, IR, rheology, and WAXS,
properties of semiconductors by shear stress during fabrica-
7
tion, and the shear-induced gelation of a mixture of telechelic
8
polymers are examples that highlight their potential. The
behavior of supramolecular systems is sensitive to external
stimuli as well, and many elegant examples using a variety of
external stimuli have been presented.
An intriguing research area is the switching of nonaggregated
precursors into aggregated structures to induce gel formation.
Examples of gelation induced by stirring or shaking often make
use of colloidal suspensions.
formation of a copper metallogel and the gelation of a small
organic compound under the influence of shaking have been
observed in synthetic molecular systems. Especially interesting
is a reported system where different thermo-, chemo-, and
9
−11
1
2,13
Moreover, the shear-induced
14
15
1
6
mechanical stimuli lead to gels of different strength.
Frequently, gelation is initiated by disruption of an intra-
molecular interaction, allowing the formation of a nucleus on
which further material can grow. In many cases, however, the
exact underlying mechanisms are not well understood.
Received: October 3, 2014
Revised: November 10, 2014
17
©
XXXX American Chemical Society
A
dx.doi.org/10.1021/ma502047h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
compound was suspended in 1 mL of chloroform in a 32 × 11.6 mm
vial; this was heated until the compound was completely dissolved and
then allowed to cool to room temperature. The obtained solution was
then stirred for 3 h at 900 rpm using a 6 × 1 mm stirring bar, followed
by a resting time of at least 2 h. Deconvolution of NMR spectra was
performed using MestReNova software version 7.1.1-9649. All liquid
NOE/ROE samples were degassed by the application of three freeze−
pump−thaw cycles. Data processing was performed using
1
13
VNMRJ.3.2.a software. All nonpulse field gradient H and C NMR
spectra were recorded on a Varian Mercury 400 MHz NMR, while all
pulse field gradient measurements (i.e., NOE/ROE/DOSY) were
recorded on a 500 MHz Varian Unit Inova. All NMR measurements
were conducted at 25 °C. Polarized optical microscopy (POM) was
performed on a Sondag Optische Instrumenten microscope.
Simulations were performed using Maestro software version 9.4, the
structures were minimized using an AMBER force field in chloroform,
hydrogen bonding atoms were connected using flexible constrains.
Synthetic Procedures and Characterization. 8-((tert-
Butoxycarbonyl)amino)octanoic Acid (4). 8-Aminooctanoic acid
(3.0 g, 18.8 mmol) was dissolved in a mixture of DCM (5 mL) and
water (5 mL). NaOH (1.5 g, 37.5 mmol) was added, and the solution
was cooled to 0 °C. A solution of di-tert-butyldicarbonate (4.1 g, 18.8
mmol) in DCM (20 mL) was slowly added, and the reaction mixture
was stirred for 24 h at room temperature. The solution was acidified
using concentrated HCl (3 mL), and the aqueous phase was extracted
with DCM (2 × 40 mL). The combined organic phases were dried
using MgSO , and the solvent was removed under vacuum. Yield: 2.32
4
1
g, 8.97 mmol. η: 47.6%. H NMR (400 MHz; CDCl ) δ: 9.2 (1H, bs,
3
COOH), 4.54 (1H, s, NH), 3.09 (2H, m, NH−CH ), 2.33 (t, CO−
Figure 1. Upper frame: molecular structure of 1 and 2. Lower frame:
minimized conformation of a monomeric cycle of 1.
2
CH ), 1.66 (NH−CH −CH ), 1.44 (11H, m, CH , CH ), 1.32 (6H,
2
2
2
3
2
13
m, CH ). C NMR (100 MHz, CDCl ) δ: 179.12, 155.99, 79.10,
2
3
4
0.53, 33.95, 29.93, 28.92, 28.40, 26.53, 24.58. MALDI-TOF-MS:
+
calcd mass 259.18 g/mol; found 282.27 g/mol (M + Na) .
Dodecane-1,12-Diyl Bis(8-((tert-butoxycarbonyl)amino)-
octanoate) (5). 8-((tert-Butoxycarbonyl)amino)octanoic acid (142
mg, 0.548 mmol) and dodecane-1,12-diol (55 mg, 0.272 mmol) were
dissolved in hexane (5 mL), and Novozym 435 beads (15 mg) were
added. The solution was put on a rotary evaporator and gently rotated
for 3 h at 70 °C at a pressure of 800 mbar. When necessary, extra
hexane was added. The solution was cooled down, hexane (20 mL)
was added, and the enzyme beads were filtered off and rinsed with
additional hexane. The solvent was evaporated under vacuum to yield
1
the desired compound. Yield: 185 mg, 0.27 mmol. η: 99%. H NMR
(
(
(
7
400 MHz; CDCl ) δ: 4.48 (2H, bs, NH), 4.05 (4H, t, O−CH ), 3.11
3
2
4H, m, N−CH ), 1.62 (8H, m, CH ), 1.44 (22H, m, CH , CH ), 1.31
2
2
3
2
13
28H, m, CH ). C NMR (100 MHz, CDCl ) δ: 173.89, 155.97,
2
3
9.02, 64.41, 40.52, 34.31, 30.00, 29.53, 29.50, 29.24, 29.03, 28.91,
Figure 2. Influence of external stimuli on the self-assembly of 1 in
28.64, 28.42, 26.60, 25.91, 24.88. MALDI-TOF-MS: calcd mass 684.53
g/mol; found 707.53 g/mol (M + Na) .
+
CHCl . Slowly cooling the hot solution results in a nonviscous liquid.
3
Shaking or sonicating for any period of time or stirring for up to 30
min results in a transparent, weak gel. A similar weak gel is obtained by
8,8′-(Dodecane-1,12-diylbis(oxy))bis(8-oxooctan-1-ami-
nium) 2,2,2-Trifluoroacetate (6). Dodecane-1,12-diyl bis(8-((tert-
butoxycarbonyl)amino)octanoate) (193 mg, 0.28 mmol) was dissolved
in a solution of 30% TFA in DCM (10 mL total). The mixture was
stirred at room temperature for 3 h. The solvents were coevaporated
with toluene (2 × 5 mL) to yield the desired product. Yield: 137 mg.
quickly freezing the solution in liquid N . Weak gels formed by any of
2
these procedures can be converted into nontransparent stronger gels
by stirring for an extended period of time (±3 h). UPy groups are
represented by blue squares while red squares represent the urethane
groups. Cyclic structures have been omitted from the cartoon.
1
+
η: 99.6%. H NMR (400 MHz; CDCl ) δ: 7.86 (6H, bs, NH ), 4.09
3
3
(
1
4H, t, O−CH ), 2.98 (4H, bs, N−CH ), 2.32 (4H, t, CO−CH ),
2
2
2
13
.75−1.54 (12H, m, CH ), 1.47−1.21 (28H, m, CH ). C NMR (100
the results of which reveal that a UPy−urethane hydrogen bond
is present in solution but is absent in the gel. On the basis of
these results, we propose a mechanism that rationalizes the
mechanically and temperature-induced gelation.
2
2
MHz, CDCl ) δ: 174.38, 67.06, 39.95, 34.23, 29.51, 29.05, 28.83,
3
2
8.42, 28.32, 27.25, 25.72, 24.56. MALDI-TOF-MS: calcd mass 684.99
g/mol; found 486.46 g/mol (monoprotonated diamine without TFA).
N-(6-Methyl-4-oxo-1,4-dihydropyrimidin-2-yl)-1H-imida-
zole-1-carboxamide (7). 2-Amino-6-methylpyrimidin-4(1H)-one
EXPERIMENTAL SECTION
Materials and Methods. All used solvents were of analytical
grade; all chemicals were purchased from Sigma-Aldrich and used
without further purification. Immobilized Candida Antarctica Lipase B
(1.0 g, 7.99 mmol) was dissolved in DMSO (10 mL), di(1H-
■
imidazol-1-yl)methanone (1.68 g, 10.36 mmol) was added, and the
mixture was stirred for 24 h at 80 °C. The solution was cooled to room
temperature, and acetone was added to precipitate the product. The
precipitate was filtered, and the residue washed with acetone. The
product was dried under vacuum to yield the desired product. Yield:
1.59 g, 7.25 mmol. η: 91%. Note: due to its extremely low solubility in
(
Novozym 435) was obtained from Novozymes A/S. Unless noted
otherwise, all measurements were performed in chloroform. Unless
noted otherwise, all gels were made by the following procedure: The
B
dx.doi.org/10.1021/ma502047h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
a
Scheme S1. Synthetic Route toward Compound 2
a
Reagents and conditions: (i) DCM, H O, NaOH, di-tert-butyl dicarbonate, RT; (ii) hexane, Novozyme 435, 70 °C, 800 mbar; (iii) DCM, TFA,
2
RT; (iv) DMSO, 80 °C; (v) DMF, triethylamine, RT.
most solvents, this compound is hard to characterize. IR (ATR): v
̀
=
appearance and mechanical properties. A transparent, weak gel
is obtained after roughly 10 h as a result of shaking or
sonication for several seconds as well as stirring for a short
period of time (<30 min). In contrast, a much more rigid,
nontransparent strong gel is formed directly by stirring for an
extended period of time (±3 h). Both the weak and the strong
gel showed no visible change for extended periods of time (>1
month for 80 mM gels). Upon heating the samples above 50
3
1
175, 3075, 2648 (bs) 1701, 1645, 1601, 1509, 1479, 1375, 1334,
320, 1276, 1233, 1224, 1190, 1169, 1090, 1065, 1026, 983 cm .
Dodecane-1,12-diyl Bis(8-(3-(6-methyl-4-oxo-1,4-dihydro-
−1
pyrimidin-2-yl)ureido)octanoate) (2). 8,8′-(Dodecane-1,12-
diylbis(oxy))bis(8-oxooctan-1-aminium) 2,2,2-trifluoroacetate (137
mg, 0.28 mmol) was dissolved in DMF (10 mL). Triethylamine (0.2
mL) and N-(6-methyl-4-oxo-1,4-dihydropyrimidin-2-yl)-1H-imida-
zole-1-carboxamide (300 mg, 1.37 mmol) were added, and the
mixture was stirred at room temperature for 24 h. The solvent was
removed in vacuo, and chloroform (50 mL) was added. The organic
phase was filtered and subsequently dried under vacuum to yield the
°
C, followed by slow cooling, stable low viscous solutions (>2
months for an 80 mM solution) were obtained again. These
solutions can be reconverted into weak or strong gels upon
agitation, showing that the processes are fully reversible.
Surprisingly, not only mechanical agitation caused gelation of
the solutions. Freezing the hot solution by immersion in liquid
1
desired compound. Yield: 191 mg, 0.243 mmol. η: 86%. H NMR (400
MHz; CDCl ) δ: 13.11 (2H, s, CH CN−H), 11.88 (2H, s,
3
3
CH NH(CO)NH), 10.11 (2H, S, CH NH(CO)NH), 5.81
2
2
(
2H, s, CHCCH ), 4.05 (4H, t, (CO)O−CH ), 3.25 (4H, m,
3
2
NH−CH ), 2.22 (6H, m, CH ), 1.54 (6H, m, CH ), 1.55−1.20 (40H,
m, CH ). C NMR (100 MHz, CDCl ) δ: 173.92, 172.99, 156.57,
N resulted in formation of a transparent weak gel immediately
2
2 3 3
13
2
3
after thawing of the solvent. The gels formed via this approach
were similar in appearance as those formed by shaking and
sonication. Although the viscosity and rate of formation were
visibly lower, weak gels could also be formed by seeding a
solution of 1 (40 mM) with gel of (equal) concentration, which
shows that the formation of the gels is driven by a nucleated
growth mechanism. Although continued stirring is the only
preparation method that results in the formation of opaque
strong gels, it should be noted that also the formation of the
weak transparent gels is a gradual process, and the kind and
duration of the applied stimuli influence their eventual strength.
Samples containing urethane-UPy 1 at concentrations of c =
40 and 80 mM in other solvents such as tetrahydrofuran,
dichloromethane, dichloroethane, toluene, or 1,2-dichloroben-
zene did not gelate under the influence of stirring, likely as a
result of the poor solubility of 1 in these solvents.
1
2
54.70, 148.16, 106.70, 64.31, 39.97, 34.34, 29.53, 29.23, 28.64, 26.41,
5.97, 24.83, 18.93. MALDI-TOF-MS: calcd mass 786.50 g/mol;
+
found 787.52 g/mol (M + H) .
RESULTS AND DISCUSSION
■
Gelation Studies. Dissolution of either 1 or 2 in hot
CHCl resulted in nonviscous solutions after slow cooling.
3
While solutions of 2 never showed an increase in viscosity in
the concentrations we examined (c = 1−150 mM), solutions of
1
on the type of stimulus applied. Figure 2 summarizes different
preparation methods (slow cooling by standing under ambient
conditions, quickly freezing a hot solution in liquid N , and the
application of different mechanical forces) and how these affect
the material properties of 1 in chloroform.
could be converted into gels of different strength depending
2
The different methods of mechanically agitating solutions of
(40−120 mM in chloroform) yielded gels that differ in
Liquid State Obtained after Slow Cooling. To elucidate
the different conformations formed by 1 in chloroform, H
1
1
C
dx.doi.org/10.1021/ma502047h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
NMR spectra were measured at various concentrations. As
shown in Figure 3, at least eight different urethane signals can
be observed at c = 10 mM. The signals numbered 1, 2, 3, 6, 7,
and 8 are already present at concentrations as low as 1.25 mM.
analysis of a c = 15 mM solution, which showed that the signals
associated with monomeric cycles displayed a larger diffusion
coefficient than those associated with dimeric cycles (see
Figures S10 and S11 for details).
The stability of the monomeric cycles formed by 1 and 2 was
studied by determining their effective molarities (EMs), defined
as Kintra/Kinter, and a measure for the stability of a cycle. For
urethane-UPy 1 the urethane signals associated with mono-
meric cycles (i.e., signals 1, 2, and 8) were used. Their
combined EM has a value of ±8.5 mM. Since ester-
functionalized 2 obviously lacks the urethane protons, a similar
deconvolution using the concentration-dependent splitting of
UPy proton H was performed (see Figure 5). In order to
d
directly compare the results of 2 with 1, deconvolution was also
performed for UPy-urea proton H in 1.
d
1
Figure 3. H NMR spectrum of 1 at c = 10 mM, showing eight
different urethane signals. The NMR spectrum is shown in black, while
the deconvolution of the different signals is shown in blue.
Deconvolution of the signals allowed for the quantification of
the peak areas, from which the corresponding concentrations
were calculated. The concentrations corresponding to all signals
except signal 5 saturate upon increasing the total concentration
(
Figure 4). In contrast, signal 5 displays a short period of
exponential growth after which it shows a linear increase, which
continues to at least c = 100 mM.
Figure 5. Concentration of monomeric cycles of 1 and 2 versus the
total concentration. The position of the plateau that is reached
indicates the effective molarity of the compounds.
In compound 1, the EM values derived from the urethane
and the UPy protons are very similar, confirming that proton
H represents only the monomeric cycles. In contrast, the EM
d
of ester-functionalized 2 reveals a lower value of ±5.8 mM.
These results suggest that the monomeric cycles formed by 1
are stabilized by an additional intramolecular interaction. It
should be noted that, although only one set of signals for cyclic
structures of 2 is observed, it cannot be excluded that multiple
species are present.
Molecular modeling reveals an intramolecular UPy−urethane
hydrogen bond in the monomeric cycle (Figure 1). In order to
experimentally verify the presence of this interaction, DPFGSE-
NOESY was performed on a 2.5 mM sample of 1. This
concentration is significantly lower than the EM (8.5 mM), and
the solution will therefore contain predominantly monomeric
cycles. This experiment revealed a NOE contact between
urethane proton H and the alkylidene UPy proton H ,
Figure 4. Concentration-dependent behavior of various urethane
signals of 1 in CHCl . The leveling-off of all signals except 5 is
3
indicative for the formation of cyclic species, while the exponential
growth of signal 5 is indicative for the growth of a linear polymer.
As shown by Ercolani et al., the leveling off observed for all
signals except 5 is indicative for the formation of cyclic
i
a
19
structures of a compound with a high dimerization constant,
a
confirming that the intramolecular UPy−urethane hydrogen
bond is indeed present. In fact, a very similar UPy−urethane
interaction has recently been reported for the crystal structure
condition that is indeed met for the UPy end-groups (K = 6
dim
7
−1
20
×
10 M in chloroform). In contrast, the linear growth
21
observed for signal 5 is indicative for the formation of linear
of the dimeric cycle of a related UPy compound.
As elegantly demonstrated by Butts et al. for strychnine,
1
8
22
polymeric species.
In an attempt to discriminate between monomeric, dimeric
and possibly higher cyclic structures, the growth rate of the
various signals in Figure 4 was examined in more detail (see
Figure S3 for details). The results show that the observed cyclic
species are mainly monomeric and dimeric in nature and
interconvert rapidly. This was further supported by DOSY
analysis of NOESY data can be used to accurately estimate
proton−proton distances. A similar analysis in our case reveals
that the distance between protons H −H (Figure 1 and Figure
i
a
S9) is 110% of the distance between protons H −H , a value
i
h
that is in close agreement with typical distances (111%)
obtained by molecular modeling.
D
dx.doi.org/10.1021/ma502047h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Disrupting the Intramolecular UPy−Urethane Hydro-
gen Bonds. The material properties of most UPy derivatives
reported to date are the result of a hierarchical assembly
process consisting of the dimerization of UPy moieties into
polymers, lateral stacking of these polymers via hydrogen bonds
formed between urethane or urea groups, and last bundling of
23
these stacks into nanofibers (Figure 6).
Figure 7. Representation of the proposed UPy−urethane interaction
in the linear species (not the result of a simulation).
the gel. To verify this, DPFGSE-ROESY measurements were
performed on the linear species in a 60 mM solution of 1 and
on all species in the corresponding strong gel, which was made
by stirring the solution for 3 h. For proper comparison both
measurements were performed on a magic-angle-spinning
setup. The result of the measurements on the linear species
in solution revealed a very similar NOE contact as was observed
for the cyclic species in the 2.5 mM sample. In contrast, the
measurements performed on the strong gel did not show a
NOE contact between proton H and H .
i
a
Additional investigation of c = 2.5 mM and c = 60 mM slowly
cooled solutions of 1 by IR spectroscopy revealed two N−H
−1
Figure 6. Upper frame: typical assembly process of ditopic UPy
compounds. UPy dimerization results in the formation of linear
polymers. Stacking of these dimers, stabilized by hydrogen bonding
between urethane or urea moieties, results in the formation of stacks.
Clustering of these stacks results in nanofibers, giving rise to increased
material properties. Lower frame: representation of the UPy stacks
with the intermolecular urethane hydrogen bonds present (not the
result of a simulation).
stretch bands at ν = 3450 and 3220 cm , of which the 3450
−1
cm vibration can be attributed to a non-hydrogen-bonded
25,26
−1
urethane (Figure 8).
The value of 3220 cm is low
In most cases of mechanically induced gelation described in
the literature, an intramolecular interaction prevents the
molecule from forming supramolecular polymers. The
application of mechanical force results in breakage of this
interaction, allowing for the growth of nuclei which initiate
1
7
further aggregation. As described above, such an interaction is
indeed present in the majority of the cyclic species formed by 1.
This UPy−urethane interaction might be strong enough to
prevent the ring from opening and thus prevent polymerization
and eventually gelation of the solution. However, at the
concentrations at which the mechanically induced gels are
formed (c = 40−120 mM for both the weak and strong gels),
the combined cyclic species make up only a small fraction of
the solution (≈ 20% for the 100 mM solution), while the rest of
the material is present as linear chains. As a result, it is unlikely
that the cyclic species are the (only) cause of the mechanically
induced gelation.
The above analysis prompted us to hypothesize that a similar
UPy−urethane interaction as observed in the cyclic species
would also be present in the linear polymers (Figure 7). This
would occupy the urethane NH and as thus prohibit the
formation of lateral urethane−urethane interactions, shown to
2
3,24
be vital for the formation of stable UPy stacks
and in good
Figure 8. Zoom of FT-IR spectra of 1 in CHCl (rescaled for clarity).
3
agreement with the fact that ester-functionalized 2 is unable to
form gels. In that case, the UPy−urethane hydrogen bonds
would be present in the linear species in solution but absent in
Liquid samples were measured in solution l = 0.05 cm. The strong gel
−1
was measured in ATR mode. The spike at 3050 cm is an artifact
resulting from the absorption of the sample holder.
E
dx.doi.org/10.1021/ma502047h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
compared to the value of 3300 cm typically found in
Article
−1
2
5,26
polyurethanes.
This may be caused by the fact that the
urethane group in our system acts only as a hydrogen bond
donor (the CO part of the urethane is not participating in
any hydrogen bonds). In contrast, IR studies on the strong gel
formed by extended stirring showed the absence of both peaks
−1
and the appearance of a new N−H peak at ν = 3319 cm ,
which can be attributed to the formation of lateral urethane−
2
5,26
urethane hydrogen bonds.
According to Painter et al., three different signals
corresponding to urethane CO stretch vibrations can be
2
6
observed: a signal corresponding to a non-hydrogen-bonded
−1
CO at ν = 1721 cm , a disorderly bound CO at ν = 1699
−1
−1
cm , and an orderly bound CO at ν = 1684 cm . As shown
in Figure 8, signals corresponding to a free and disorderedly
hydrogen-bonded urethane are observed for the c = 2.5 and c =
Figure 9. Polarized optical microscopy pictures of a weak gel made by
sonication for 10 min and a strong gel made by stirring for 3 h (c = 60
mM). Both gels were allowed to equilibrate for 24 h before analysis.
Magnification is 4×; scale bar represents 0.2 mm.
6
0 mM solutions, in agreement with the observations in the
N−H regime. Also similar as in the N−H regime, the free C
O is absent in the strong gel while a band for the ordered state
has emerged. It should be noted that a signal corresponding to
the disorderly bound CO is also still present. While this
could originate from disordered sections of the gel, it is also the
position of the CO vibration in the “urea” part of the UPy
addition, the thixotropic nature of the gel was examined by
alternating between low (1%) and high (100%) strain. The gel
heals partly in ±40 s, and this process is reproducible (Figure
1
0). The slow drift in G′ and G″ is attributed to incomplete
2
5,27
moiety.
Similar measurements were also performed for the weak gel,
formed by sonication or stirring for 10 min. These revealed
identical NMR and IR spectra as observed for the solution. This
suggests that only a very small fraction of the linear UPy
polymers is stacked, in good agreement with the low strength of
this gel. In order to investigate if the strong gel formed by 1
indeed comprises bundles of stacked polymers, WAXS was
performed on the dried gel (c = 80 mM). As a reference,
compounds with urethane−urethane linker lengths of C and
8
C
1
instead of C were synthesized according to a literature
12
17
6
procedure and measured as well. The results revealed signals
which have previously been attributed to interchain distances in
polyurethanes, interdimer UPy−UPy distances, and stack−
stack distances in UPy nanofibers. Hence, evidence is presented
to support the proposed structure of the gel as presented in
Figure 6 (see Figure S13 for more details).
In order to study the gels directly (i.e., without drying),
polarized optical microscopy (POM) was used to characterize a
weak gel made by sonication and strong gel made by stirring (c
Figure 10. Thixotropic properties of the c = 80 mM gel of 1 in CDCl3
formed by stirring for 3 h and a resting period of 2 h. The
measurement was performed using a 25 mm parallel plate setup,
consisting of strains of 10 s 1%; 20 s 100%; 200 s 1%; 20 s 200%; and
100 s 1%. ω = 1 rad/s. See Supporting Information for further
experimental details.
=
60 mM, Figure 9). POM analysis revealed that the weak gel
does contain some birefringent sections but is generally very
unordered and heterogeneous. In contrast, the strong gel is not
birefringent at all but appears much more homogeneous. This
suggests that the continued sheer stress applied by extended
stirring allows the heterogonous sections of the weak gel to re-
form in a more ordered and dense network, very similar as has
healing of the gel and/or evaporation of the solvent, leading to
shrinkage of the gel and a reduced contact area with the setup.
Proposed Mechanism for the Gelation of 1. The results
presented herein indicate that a UPy−urethane interaction is
present in both cyclic and linear species in solution but is
absent in the strong gel. This interaction most likely kinetically
traps urethane-UPy 1, thereby preventing the formation of
16
been reported in the literature.
Mechanical Properties of the Strong Gel. With the
molecular structure of the gels understood, we focused on
studying their material properties using oscillatory rheology.
The fragile structure of the weak gel made it impossible to
obtain reproducible results. Therefore, further experiments
were performed on a strong gel, formed by stirring an 80 mM
solution of 1 in chloroform for 3 h. Frequency sweep
measurements (ω = 0.5−100 rad/s, γ = 1%) revealed that
the storage modulus is larger than the loss modulus over the
complete range of the experiment, showing the elastic
properties of the gel. Strain sweep experiments show that the
crossover point lies at a strain of 12% (ω = 1 rad/s). In
2
3,24
lateral aggregates, necessary for gelation.
This might partly
be the result of sterical hindrance as a result of back-folding of
the polymeric chain but likely more significantly arise from
occupation of the urethane moieties in the UPy−urethane
hydrogen bonds.
One mechanism for the gelation that we considered is the
mechanically induced breakage of laterally stacked UPy
aggregates. This would lead to a large increase in the number
of aggregates, which could act as nuclei for further aggregation.
However, as mentioned above, it is unlikely that many lateral
aggregates are formed while the UPy−urethane hydrogen
F
dx.doi.org/10.1021/ma502047h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
bonds are present in the cyclic structures or supramolecular
polymers. Furthermore, if the above mechanism is correct, it
would effectively lead to a multiplication of similar structures
and would therefore not explain the absence of the UPy−
urethane interaction in the gel as shown by NOE and IR. For
the initiation of the gelation process, therefore, another
mechanism is required.
in the formation of a gel made of fibers of stacked UPy dimers,
with the UPy−urethane interaction no longer present. These
results show that the UPy−urethane hydrogen bonds kineti-
cally trap the compound, while mechanical force disrupts these
interactions, allowing for the formation of a stacked nucleus on
which larger structures can grow, leading to gelation of the
solution. Fast cooling of the solutions results in gelation even
when no mechanical stress is applied. Experimental observa-
tions show that this phenomenon only happens when the
solvent is (nearly) frozen. Likely explanations are therefore
either a nucleation process initiated by the solvent crystals or
gelation as a result of the increase in the dielectric constant of
the solvent as a result of the decrease in temperature. The work
presented here provides further insight into the influence of
mechanical stresses on supramolecular systems and will
contribute to the development of stimuli-responsive materials.
For this reason it is more likely that the linear structures in
solution are initially nonstacked one-dimensional polymers.
Stretching or breaking the linear polymer by the applied
mechanical stress likely results in breakage of the UPy−
urethane interactions. In turn, the loss of this interaction allows
for the formation of a nucleus containing stacked UPy dimers,
stabilized by urethane−urethane hydrogen bonds. This nucleus
then induces the growth of larger structures, eventually leading
to the mechanically induced gelation of the solution. The fact
that only continued stirring leads to the formation of the strong
gel can be explained by a higher degree of physical cross-links.
This not only leads to a stronger gel but also to larger
aggregates, explaining the opaque appearance of this gel. In
other words, due to the high viscosity, the weak gel is also
kinetically trapped structure on its way to a strong gel.
The formation of weak gels as a result of quenching has been
ASSOCIATED CONTENT
Supporting Information
■
*
S
WAXS, rheology, and FT-IR analysis. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
E-mail e.w.meijer@tue.nl (E.W.).
1
6
■
reported before, where the low mechanical strength of the gel
was attributed to the formation of many nucleation sites as a
result of the fast cooling, thereby leading to large amorphous
sections. This is a very plausible explanation for the weak
material properties of our gel as wel. In contrast to the reported
system, our system does not gelate as a result of slow cooling,
suggesting that nucleation only takes place at or around the
melting point of chloroform. The gelation as a result of fast
cooling by immersion in liquid nitrogen might be explained as
the result of a shift in activation barriers for the formation of the
structures found in the slowly cooled liquid and gel state as a
function of temperature. However, the slowly cooled liquid
samples also form a similar weak gel as a result of cooling in
liquid nitrogen. Furthermore, fast cooling the hot or ambient
solutions by immersing in ice−water or cold chloroform (T =
*
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work is financed by The Netherlands Organisation for
Scientific Research (NWO-TOP grant: 10007851), the Dutch
Ministry of Education, Culture and Science (Gravity program
0
2
24.001.035), and the European Research Council (FP7/
007−2013, ERC Advanced Grant No. 246829).We thank Dr.
Wesley R. Browne for useful discussions about the gelation
induced by fast cooling.
−45 °C) did not result in gelation, similar as removal of the vial
REFERENCES
■
from the liquid nitrogen before crystallization of the solvent
had occurred. For this reason it is more likely that, rather than a
change in activation barriers, the crystal growth of chloroform is
responsible for the observed gelation. If the chloroform crystals
indeed act as nuclei for the gelation, their large number might
also explain why the gelation as a result of cooling proceeds in
several minutes, instead of the hours that are necessary for the
gelation as a result of mechanical agitation to take place.
Another plausible explanation is related to the large increase in
the relative permittivity (i.e., dielectric constant) of chloroform
(
1) Stuart, M. A. C.; Huck, W. T. S.; Genzer, J.; Muller, M.; Ober, C.;
Stam, M.; Sukhorukov, G. B.; Szleifer, I.; Tsukruk, V. V.; Urban, M.;
Winnik, F.; Zaucher, S.; Luzinov, I.; Minko, S. Nat. Mater. 2010, 9,
1
(
01−113.
2) Kumar, G. S.; Neckers, D. C. Chem. Rev. 1989, 89, 1915−1925.
(3) Schmaljohann, D. Adv. Drug Delivery Rev. 2006, 58, 1655−1670.
(4) Misra, A.; Stein, R. S. J. J. Polym. Sci., Polym. Phys. Ed. 1979, 17,
235−257.
5) Kondepudi, D. K.; Kaufman, R. J.; Singh, N. Science 1990, 250,
(
9
75−976.
2
8,29
(6) Noorduin, W. L.; Izumi, T.; Millemaggi, A.; Leeman, M.; Meekes,
as a result of the decrease in temperature.
It is possible that
H.; van Enckevort, W. J. P.; Kellog, R. M.; Kaptein, B.; Vlieg, E.;
at a temperature close to its freezing point the polarity of the
solvent increases to a point where the solubility of 1 is strongly
decreased, thereby inducing gel formation.
Blackmond, D. G. J. Am. Chem. Soc. 2008, 130, 1158−1159.
(7) Giri, G.; Verploegen, E.; Mannsfeld, S. C. B.; Atahan-Evrenk, S.;
Kim, D. H.; Lee, S. Y.; Becerril, H. A.; Aspuru-Guzik, A.; Toney, M. F.;
Bao, Z. Nature 2011, 480, 504−508.
CONCLUSIONS
(8) Mahmoudi, T.; Karimkhani, V.; Seok Song, G.; Soo Lee, D.; J.
Stadler, F. Macromolecules 2013, 46, 4141−4150.
■
Ditopic UPy 1 forms a variety of different cyclic species in
slowly cooled solutions, primarily comprising different
conformations of monomeric and dimeric cycles. NOESY and
FT-IR data show that the majority of these cycles are stabilized
by an intramolecular UPy−urethane hydrogen bond. A very
similar interaction is also observed in the linear polymers
formed at higher concentration. As observed by WAXS,
NOESY, and FT-IR, the application of mechanical force results
(9) Carnall, J. M. A.; Waudby, C. A.; Belenguer, A. M.; Stuart, M. C.
A.; Peyralans, J. J.-P.; Otto, S. Science 2010, 327, 1502−1506.
10) Xu, J.-F.; Chen, Y.-Z.; Wu, D.; Wu, L.-Z.; Tung, C. H.; Yang, Q.-
Z. Angew. Chem. 2013, 52, 9920−9924.
11) Paulusse, J. M. J.; Sijbesma, R. P. J. Polym. Sci., Part A: Polym.
(
(
Chem. 2006, 44, 5445−5453.
(12) Cabane, B.; Wong, K.; Lindner, P.; Lafuma, F. J. Rheology 1997,
41, 531−547.
G
dx.doi.org/10.1021/ma502047h | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
(
13) Xie, D.; Wu, H.; Zaccone, A.; Braun, L.; Chen, H.; Morbidelli,
M. Soft Matter 2010, 6, 2692−2698.
14) Piepenbrock, M.-O. M.; Clarke, N.; Steeds, J. W. Soft Matter
010, 6, 3541−3547.
15) Herpt, J. T.; Stuart, M. C. A.; Browne, W. R.; Feringa, B. L.
Langmuir 2013, 29, 8763−8767.
16) Weng, W.; Beck, J. B.; Jamieson, A. M.; Rowan, S. J. J. Am.
(
2
(
(
Chem. Soc. 2006, 128, 11663−11672.
(
17) Cravotto, G.; Cintas, P. Chem. Soc. Rev. 2009, 38, 2684.
(
18) Llansola, F. R.; Meijer, E. W. J. Am. Chem. Soc. 2013, 135,
6
549−6553.
19) Ercolani, G.; Mandolini, L.; Mencarelli, P.; Roelens, S. J. Am.
Chem. Soc. 1993, 115, 3901−3908.
20) Sontjes, S. H. M.; Sijbesma, R. P.; Genderen, M. H. P.; Meijer,
E. W. J. Am. Chem. Soc. 2000, 122, 7487−7493.
21) Lafitte, V. G. H.; Aliev, A. E.; Horton, P. N.; Hursthouse, M. B.;
Hailes, H. C. Chem. Commun. 2006, 2173−2175.
22) Butts, C. P.; Jones, C. R.; Towers, E. C.; Flynn, J. L.; Appleby,
L.; Barron, N. Org. Biomol. Chem. 2011, 9, 177−184.
23) Appel, W. P. J.; Portale, G.; Wisse, E.; Dankers, P. Y. W.; Meijer,
E. W. Macromolecules 2011, 44, 6776−6784.
24) Kautz, H.; van Beek, D. J. M.; Sijbesma, R. P.; Meijer, E. W.
Macromolecules 2006, 39, 4265−4267.
25) Coleman, M. M.; Lee, K. H.; Skrovanek, D. J.; Painter, P. C.
Macromolecules 1986, 19, 2149−2157.
26) Lee, H. S.; Wang, Y. K.; Hsu, S. L. Macromolecules 1987, 20,
089−2095.
27) Dankers, P. Y. W.; Adams, P. J. H. M.; Lo
Hest, J. C. M.; Meijer, E. W. Eur. J. Org. Chem. 2007, 3622−3632.
28) Philippe, R.; Piette, A. M. Bull. Soc. Chim. Belg. 1955, 64, 600−
27.
29) West, R. C. In Handbook of Chemistry and Physics, 60th ed., 2nd
printing; Chemical Rubber Co.: Boca Raton, FL, 1980.
(
(
̈
(
(
(
(
(
(
2
(
̈
wik, D. W. P. M.; van
(
6
(
H
dx.doi.org/10.1021/ma502047h | Macromolecules XXXX, XXX, XXX−XXX