Energy-dissipative self-assembly driven in microflow: A time-programmed self-organization and decomposition of metastable nanofibers

Article abstract of DOI:10.1246/cl.150292

Energy-dissipative self-assembly in microflow enables us to approach an energetically unstable self-assembled structure, which undergoes morphological transition in a cascade manner from fibers, sheets, and finally "running down" to thermodynamically stable dots, with switching intermolecular interactions.

Full text of DOI:10.1246/cl.150292

Received: March 31, 2015 | Accepted: April 18, 2015 | Web Released: April 25, 2015
CL-150292
Energy-dissipative Self-assembly Driven in Microflow:
A Time-programmed Self-organization and Decomposition of Metastable Nanofibers
Munenori Numata,* Akiko Sato, and Rie Nogami
Department of Biomolecular Chemistry, Graduate School of Life and Environmental Sciences,
Kyoto Prefectural University, Shimogamo, Sakyo-ku, Kyoto 606-8522
(E-mail: numata@kpu.ac.jp)
O
O
O
Energy-dissipative self-assembly in microflow enables us to
O
O
O
O
O
(a)
O
O
O
O
approach an energetically unstable self-assembled structure, which
undergoes morphological transition in a cascade manner from
fibers, sheets, and finally “running down” to thermodynamically
stable dots, with switching intermolecular interactions.
O
O
N
H
O
O
O
O
O
O
O
O
O
H
O
O
N
O
O
OPV
H2O
O
(b)
Microflow Chip
Under equilibrium conditions, entire self-assembled supra-
molecular structures are mainly governed by intermolecular
interactions of adjacent component molecules. As reported in our
previous papers, nonequilibrium environments generated by evap-
oration1a or diffusion1b of solvents, as well as flowing liquid-liquid
interfaces,1c can also contribute to the controlled self-assembling
process of molecules or polymers, where the final self-assembled
structures are affected not only by molecular design but also by
such an environment surrounding molecules. In conventional
solution chemistry, it is still difficult to generate self-assembling
fields under far from equilibrium in terms of reproducibility. In
view of this situation, we have so far focused on a microflow space
as a platform for controlled self-assembly under nonequilibrium.
On the basis of the supramolecular strategies accumulated
over several decades,2 it has been widely accepted that the
cooperative intermolecular interactions play an important role for
the creation of fine self-assembled architectures. Such appropriately
designed interactions always lead to the formation of a sole
assembly through “error-checking” processes under equilibrium
(closed systems). If self-assembly occurs under “opened environ-
ments” to which external energy is supplied continuously, self-
assembled structures could acquire a dynamic nature arising from
instability. We have found herein that even a nonappropriately
designed molecule can acquire self-assemble ability under far from
equilibrium to afford a temporal self-assembled structure, in a
different way from that occurred in a closed system, which further
exhibited morphological transitions with switching interactions
under thermodynamic conditions.
Self-assembly in flow
THF solution
Flexible
Vial fiber
H2O
(c)
In vial
Decompose
Assembly
Flexible fiber
0 min
Bundle (Tape-like)
5 min
Fan-shape
10 min
Dots
24 h
Figure 1. (a) Chemical structure of OPV, (b) its calculated model
and the present concept of energy-dissipative self-assembly in
microflow, (c) illustration of a series of morphological transitions
with time progress.
(a)
(b)
Herein, we have designed oligo(p-phenylenevinylene) (OPV)-
based amphiphilic molecules in which a hydrophobic OPV scaffold
is connected to hydrophilic gallic acid groups with amide spacers
(Figure 1a).3 Thereby, it is expected that π-π stacking and/or
hydrogen bonding would govern its self-assembly under rapid
water diffusion and the shear force effect in the microflow.
In this study, on the basis of our previous work, a flow-focus
type microchip was designed and used as a self-assembly field for
OPV. A THF solution of OPV (0.1 mM) was injected from the
central “leg” and the solution was squeezed by water injected from
the side “legs” (Figure 1b). Under the laminar flow, diffusion of
solvents proceeds uniformly, and therefore, all OPV molecules can
be hydrated at the same time. Hence, the injected OPV molecules
effectively self-assemble under the effect of the anisotropic polarity
gradation generated across the stream. The final flow rate was set to
20 ¯L min¹1, and the final solvent composition was THF/water =
4/6, v/v (final concentration; 0.04 mM). The eluted solution was
Figure 2. (a) Time-dependent UV-vis spectral changes of the eluted
solution, (b) Plot of ¦Abs. at 405 nm with respect to time progress
(3.0 mm cell, r.t). For details see the Supporting Information.
collected in a vial and subjected to spectroscopic (UV-vis,
fluorescence, and IR) measurements, XRD analyses, and AFM
observations.
Figure S1a shows UV-vis spectra of the solutions (flow
sample) immediately after elution from the flow channel in
comparison with a vial mixing sample. We could not recognize
any significant spectral changes in comparison to the sample
prepared in a vial. However, for the flow sample, the absorption
at 405 nm and the corresponding fluorescence peak at 473 nm
gradually increased with time (Figure 2a), while no such spectral
change was recognized for the vial sample (Figures S1b and S1c).
The spectral changes can be further clarified by the plot relative
to time progress, as shown in Figure 2b. The absorption and
Chem. Lett. 2015, 44, 995–997 | doi:10.1246/cl.150292
© 2015 The Chemical Society of Japan | 995

(c)
(e)
(g)
(a)
2 µm
5 µm
5 µm
2 µm
(d)
(f)
(h)
(b)
5 µm
10 µm
2 µm
10 µm
Figure 3. Morphological changes of the metastable fibers created in the flowing self-assembly field; (a), (b) flexible fibers recognized
immediately after elution (0 min); (c), (d) straight fibers and their bundled structures appeared within 5 min after elution; (e), (f), (g) fan-shape sheet
structures appeared within 8-10 min after elution; (h) thermodynamically stable dot-like structures appeared after finishing the transformations. For
further AFM images and their height profile analyses see the Supporting Information.
fluorescence intensity abruptly increased from 0 to 5 min, and then
(a)
(b)
90
80
70
60
50
40
30
20
24 h
vial
0 min
30 min
24 h
90
80
70
60
50
40
30
20
the absorbance gradually increased from 5 to 30 min, finally
reaching a constant value after 30 min. The observed spectral
changes imply that OPV self-assembled in microflow through π-π
stacking interactions, and it was gradually loosened with time after
elution from the channel.
To highlight the morphological changes caused by the
relaxation of π-π stacking, the eluting solution was cast on mica
substrates with a regular time interval and observed by AFM (for
detailed sample preparations, see Supporting Information). From
the vial mixing sample, it was found that OPV self-assembled
to afford spherical aggregates, the average size of which can be
estimated to be 8 nm from the height profiles (Figure S12). The
aggregate was stable and no further morphological change
occurred, i.e., it was a thermodynamically stable structure. In
contrast to this, as revealed from Figures 3a and 3b, flexible fibers,
the diameter of which can be estimated to be 1-2 nm, were initially
formed in the microflow. A series of AFM images clearly indicated
that the fibers underwent drastic morphological changes from
flexible fibers to straight fibers, and fan-shaped sheets (Figure 1c
and Figures 3c-3h). These morphological changes occurred within
approximately 5 min after elution from the channel. The time scale
corresponds to that of the abrupt increase in the absorption.
Furthermore, the magnified images revealed that the sheet
structures are composed of many fibers and that their ends were
decomposing (Figures S8-S10). The sheet structures completely
disappeared within 30 min to afford short fibers with 1-3 ¯m
length, which finally transformed to spherical dots, as observed in
a vial sample. From the time scale, the decomposition can be
associated with the gradual increase in absorption.
1653
1558
1550
Wavenumber / cm-1
1635
1541
1650
1600
1500
1650
1600
1550
1500
Wavenumber / cm-1
Figure 4. (a) Time-dependent IR spectral changes of the microflow
samples, (b) comparative IR spectra between a flow (after 24 h) and a
vial sample.
recorded after 24 h was in good accordance with that of the vial
mixing sample, in which the peak at 1653 cm
¹1
completely
disappeared. The peak shift in amide I to lower wavenumber
regions consistently indicates that the series of morphological
changes shown in Figure 3 were driven by the hydrogen-bonding
interaction. A similar time-dependent peak shift was also observed
in the region of amide II (1558 and 1541 cm¹1). Combining with
the observed hyperchromic effect, the main interaction in the fibers
constructed in the microflow would be switched from π-π stacking
to hydrogen-bonding interactions with time progress in the vial.
It was expected that the XRD diffraction data would provide
further detail information on molecular packing. Unexpectedly, no
diffraction peak was recognized from the sample prepared imme-
diately after elution from the microflow channel. Taking the rather
flexible nature into consideration, it would be natural that there is
no diffraction peak in the sample, probably due to random molec-
ular packing. On the other hand, the sample prepared after 24 h,
which would contain spherical aggregates, exhibited diffraction
peaks with lattice parameters of q = 13.9 (0.45 nm) (Figure S21),
which would be consistently assignable to the distance of adjacent
π-surfaces assisted by hydrogen bonds.5
To obtain further insight into the intermolecular interactions,
IR measurements were performed. As shown in Figure 4, the
sample elapsed approximately 7 min after elution, which would
contain the flexible nanofibers, exhibited two amide I peaks at 1653
¹1
and 1635 cm¹1. The sharp peak at 1653 cm could be assigned
to the stretching band of free amide groups, whereas the peak at
¹1
1635 cm was indicative of hydrogen-bonding interactions. After
Together with all data, the observed morphological transitions
can be correlated to the dynamics at molecular level. Under kinetic
control, the π-π stacking and hydrogen bonds would not work
¹1
30 min, the peak at 1653 cm became weak and the peak at
¹1
1635 cm
became predominant.4 Furthermore, the spectrum
996 | Chem. Lett. 2015, 44, 995–997 | doi:10.1246/cl.150292
© 2015 The Chemical Society of Japan

cooperatively, and then, reflecting the polar environment, π-π
stacking would govern the self-assembly. The resultant nanofibers,
therefore, should be maintained by somewhat disturbed π-π
stacking. It should be emphasized that such an unfavorable self-
assembly never take places in a conventional vial process; rapid and
uniform diffusion of water or flowing energy would endow OPV
molecules with an unusual self-assembling ability, as we have
confirmed in other molecular assembly systems.1,6 The unfavorable
stacking would be a driving force for the intrapolymer rearrange-
ment in the next stage, where π-π stacking is loosened (but it still
works) and the hydrogen bonds become predominant. The abrupt
hyperchromic effect occurring within 5 min should correspond to
this intrapolymer arrangement. Switching the interactions would
cause the first morphological changes from flexible to straight
fibers, which is still maintained by nonoptimum interactions.
Magnified AFM images revealed that the flexible fibers were
composed of many nanofibers (Figure S3). Bundling of these
nanofibers, therefore, was likely to accelerate the transformation
toward the straight fibers. In some cases, further bundling led to the
creation of the highly branched tape-like structures (Figures S6 and
S7). The height (thickness) of the tape-structure was almost
consistent with that of the original fibers (1-2 nm), being indicative
of the regular piling up of the fibers through π-π stacking.
A further featuring aspect of the created fibers is to decompose
from its ends in the time range from 5 to 10 min, which almost
overlaps with the bundling process (Figure S10). The gradual
increase in absorption from 5 to 30 min would be correlated with
this decomposition process. The synchronous decomposition of
many nanofibers in the tape structures would cause split their edges,
thus giving rise to the creation of fan-shaped sheet structures. The
average height (2-3 nm) of the split end is also comparable to that of
the tape and original fibers, suggesting that successive morpho-
logical transformations from the original fiber to the fan-shaped
sheet occurred. It is noteworthy that the bundling (self-assembly of
fibers) would be a trigger to decompose many nanofibers at the same
time, giving a reasonable explanation for the appearance of fan-
shape structures (vide infra). The nanofibers in the sheet were cut
into fragmentized short fibers (e.g., Figure S10), implying that most
π-π stacking still remained in this time range. The relatively gradual
increase in absorption in this time scale supports this notion. In the
final stage, the fragmentized fibers were directly transformed into
spherical aggregates without passing through the monomeric OPV.
This process occurring after 30 min is no longer accompanied by
any absorption changes, but only by hydrogen bond strengthening.
The resultant spherical aggregate should have a similar self-
assembled structure to that created in a vial mixing process.
suppressed (Figure S14). This result also supports the view that
bundling would be a trigger for further transformations.
Finally, we investigated how the flow microenvironment
contributed to the creation of the active nanofibers.6b,6c When
slightly varying the side solvent polarities from water to the mixed
THF/water solvent (THF/water; 1/11, v/v, final solvent com-
position, 45/55, v/v), OPV did not show any self-assembling
ability even in the microflow (Figure S16); generating a clear
water/THF interface is necessary for the effective assembly. We
have also investigated the effect of flow rates on the self-
assembling ability of OPV. Even upon varying the flow rate from
20 ¯L min¹1 to 10 and 40 ¯L min¹1, active nanofibers were formed
and underwent a similar morphological transition as described
above. Interestingly, the fiber length tended to be shorter under
¹1
10 ¯L min (Figure S18), whereas longer and thicker fibers were
observed with increasing the flow rate (40 ¯L min¹1) (Figure S19).
These results suggest the notion that the mechanical energy arising
from the hydrodynamic effect might also play an important role for
the up-hill self-assembly of OPV.
In conclusion, we have demonstrated that energetically up-hill
self-assembly was effectively achieved in the flowing self-assembly
field. The obtained nanofibers are metastable and have the potential
to generate various kinetic structures in a cascade manner toward
the formation of a thermodynamically stable structure. Microflow
allowed us to select an interaction from multiple recognition sites
in the OPV molecule and to enforce it under kinetic conditions.
Extending the present concept more generally, a guide for
molecular design is to introduce multiple recognition sites in a
way that they do not work cooperatively, where self-assembly can
be accomplished only through energy consumption. We believe that
self-assembly of such a nonoptimum molecules, in combination
with microflow, will open up new opportunities for supramolecular
chemistry.
We thank Prof. H. Tamiaki and Dr. S. Shoji for kindly
supporting our IR spectral measurements.
Supporting Information is available electronically on J-STAGE.
References and Notes
1
a) M. Numata, D. Kinoshita, N. Taniguchi, H. Tamiaki, A. Ohta,
Angew. Chem., Int. Ed. 2012, 51, 1844. b) M. Numata, T. Kozawa,
Chem.®Eur. J. 2013, 19, 12629. c) M. Numata, Y. Takigami, M.
Takayama, Chem. Lett. 2011, 40, 102.
2
a) J.-M. Lehn, Supramolecular Chemistry: Concepts and Perspectives,
Wiley-VCH, Weinheim, 1995. doi:10.1002/3527607439. b) Supra-
molecular Chemistry: From Molecules to Nanomaterials, ed. by
P. A. Gale, J. W. Steed, Wiley, Hoboken, 2012. doi:10.1002/
9780470661345.
The original nanofibers created in the microflow as well as
following structures are metastable. Indeed they are kinetically
self-assembled structures, and each could not be isolated due to
the short life, which is in contrast to our previous works.1 The
following experiments highlight this point. When the eluted
solution was stored under 30 °C, the nanofibers easily decomposed
to afford short fibers or spherical aggregates directly, without
passing through the sheet structures (Figure S13), implying that the
observed transformations took place under the delicate balance
between bundling (self-assembly) and decomposition. In contrast,
the fibers were basically stable under 0 °C for 1 h. Likewise, when
the eluted solution was pipetted through a capillary tube, the
nanofibers decomposed to give spherical aggregates (Figure S17),
implying that the nanofibers are quite fragile and sensitive to
mechanical stimuli. When diluted with THF/water mixed solvent
(THF/water; 4/6, v/v), bundling of the nanofibers was completely
3
4
a) S. Yagai, S. Kubota, H. Saito, K. Unoike, T. Karatsu, A. Kitamura, A.
Ajayaghosh, M. Kanesato, Y. Kikkawa, J. Am. Chem. Soc. 2009, 131,
5408. b) P. G. A. Janssen, J. Vandenbergh, J. L. J. van Dongen, E. W.
Meijer, A. P. H. J. Schenning, J. Am. Chem. Soc. 2007, 129, 6078.
To prepare the sample for IR measurement, at least 240 ¯L of eluted
solution was needed. It was difficult to collect the sample every several
minutes (for details see the Supporting Information).
5
6
S. S. Babu, V. K. Praveen, K. K. Kartha, S. Mahesh, A. Ajayaghosh,
Chem.®Asian J. 2014, 9, 1830.
Time-programmed self-assembly in microflow: a) M. Numata, M.
Takayama, S. Shoji, H. Tamiaki, Chem. Lett. 2012, 41, 1689. Related
papers: b) A. B. Marciel, M. Tanyeri, B. D. Wall, J. D. Tovar, C. M.
Schroeder, W. L. Wilson, Adv. Mater. 2013, 25, 6398. c) V. Foderà, S.
Pagliara, O. Otto, U. F. Keyser, A. M. Donald, J. Phys. Chem. Lett.
2012, 3, 2803.
Chem. Lett. 2015, 44, 995–997 | doi:10.1246/cl.150292
© 2015 The Chemical Society of Japan | 997