Microflow-driven temporal self-assembly of amphiphilic molecules

Article abstract of DOI:10.1246/cl.2012.1689

A dynamic oilwater interface generated in a double-Yshaped microfluidic device allowed amphiphilic molecules to self-assemble spontaneously for a set period of time, leading to the creation of discrete supramolecular coordination polymers. After elution f

Full text of DOI:10.1246/cl.2012.1689

doi:10.1246/cl.2012.1689
Published on the web December 8, 2012
1689
Microflow-driven Temporal Self-assembly of Amphiphilic Molecules
Munenori Numata,*1 Momoko Takayama,1 Sunao Shoji,2 and Hitoshi Tamiaki2
1Department of Biomolecular Chemistry, Graduate School of Life and Environmental Sciences,
Kyoto Prefectural University, Shimogamo, Sakyo-ku, Kyoto 606-8522
2Graduate School of Life Sciences, Ritsumeikan University, Kusatsu, Shiga 525-8577
(Received August 14, 2012; CL-120849; E-mail: numata@kpu.ac.jp)
A dynamic oil-water interface generated in a double-Y-
shaped microfluidic device allowed amphiphilic molecules to
self-assemble spontaneously for a set period of time, leading to
the creation of discrete supramolecular coordination polymers.
After elution from the device, these 1-D structures gradually
dissociated into their monomer units, regenerating the initial
state. The self-assembled structures were maintained only under
the flow conditions in the device.
From a limited set of components, nature generates a diverse
range of self-assembled architectures depending on the physio-
logical environment. These architectures are created in a desired
space and moment and exert their functionalities within a
regulated time; subsequently, they dissociate into their initial
components, thereby acting as renewable molecular sources for
further self-assembly. In this sense, nature’s self-assembled
architectures can be characterized in terms of both temporal and
discrete structures; they emerge spontaneously at a desired space
but maintain their forms only through the consumption of energy
supplied from external systems (i.e., far from equilibrium
conditions). One of the goals of supramolecular chemistry is to
reconstruct such a nonequilibrium self-assembling system, as
characterized by flash assembly, a temporal assembled structure,
and circular chemical conditions.1 Inspired by nature, herein we
propose a novel self-assembling system, based on a dynamic
oil-water interface generated temporally in a microfluidic device
in which amphiphilic molecules could self-assemble within a
controlled period of time (Figure 1).2,3
To provide a dynamic interface, we developed microfluidic
devices having a channel pattern (width, 160 ¯m; depth, 40 ¯m)
with Y-shaped junctions at both the up- and downstream ends,
separated by 80 mm. When an organic solvent and water were
charged continuously into the two channels at a constant flow
rate, a dynamic liquid-liquid (oil-water) interface was generated
at the upstream junction; it was maintained through the channel
and disappeared at the downstream junction. Therefore, mole-
cules flowing through the device would experience the interface
only temporarily. If an appropriately designed amphiphilic
molecule having 1-D self-assembling ability was to be injected
into the channel, it would begin to self-assemble at the upper
junction to form 1-D assemblies that would separate from the
interface after a period of time at the downstream junction,
losing the self-assembling ability.4 This self-assembly event
would be influenced by the local concentration at the upper
junction and the flow time required to reach the downstream
junction.
Figure 1. (a) Photograph and schematic representation of
the microflow system. (b) Chemical structure of Chl-4Py and
schematic representation of the 1-D self-assembly of injected
amphiphilic molecules at the dynamic interface.
or nicotinic acid (Chl-3Py, Scheme S1)11 moiety at position 3 of
the chlorophyll ring.3b,5 These amphiphilic molecules, which
feature a perpendicular orientation between the TEG units and
the coordination moieties, were suitable for 1-D self-assembly®
formed mainly through Zn-pyridine coordination®at the dy-
namic interface (Figure 1). In addition, we prepared Chl-OMe
and Chl-OH, which could form somewhat stiffer 1-D assembled
structures through strong ³-³ stacking in addition to Zn-
oxygen coordination (Scheme S1).5b,5d,11
1,2-Dichloroethane (DCE) solutions of Chl-4Py at three
different concentrations (0.5, 2.5, and 12.5 mM) were prepared.
Each solution was charged into a channel, with distilled water
charged into the other, using a syringe pump, at a flow rate of 10,
30, or 50 ¯L min¹1. In these cases, the residence time of the
injected solution was calculated to be 5.48, 0.80, and 0.48 s,
respectively. The collected DCE solutions were subjected to
spectroscopic measurements and microscopic observations.
Resonance Raman spectroscopy is a sensitive tool for
detecting intermolecular interactions between chlorophyll units.6
Therefore, we used it to observe the self-assembly of Chl-4Py at
the dynamic interface through analysis of the flowing DCE
solution in the microfluidic device. For a bulk DCE solution
containing Chl-4Py, the stretching modes of the chlorophyll
¹1
rings appeared at 1554 cm
upon excitation at 405 nm,
assignable to monomeric Chl-4Py. When the pyridyl group in
Chl-4Py was coordinated to the central Zn atom, however, the
Raman peak shifted to a higher wavenumber (by ca. 5 cm¹1).
With focus on these signals, we recorded Raman spectra at the
start, middle (40 mm from the Y-junctions), and end of the
For this study, we designed Zn-chlorophyll-based amphi-
philic molecules featuring dendritic tetra(ethylene glycol) (TEG)
units at position 17 and an isonicotinic acid (Chl-4Py, Figure 1)
Chem. Lett. 2012, 41, 1689-1691
© 2012 The Chemical Society of Japan
www.csj.jp/journals/chem-lett/

1690
1559
1554
1553
1555
500
400
300
200
100
0
500
400
300
200
100
0
(a)
(b)
50
30
0.48
0.80
1000
1200
1400
1600
1800
2000
1000
1200
1400
1600
1800
2000
Raman shift / cm-1
Raman shift / cm-1
1559
1554
500
400
300
200
100
0
(c)
(d)
(a)
(b)
(c)
10
5.48
Interface
Organic phase
1000
1200
1400
1600
1800
2000
0.5
2.5
Concentration / mM
12.5
Raman shift / cm-1
Figure 2. Resonance Raman spectra of Chl-4Py solutions
(2.5 mM) flowing within the microchannel (10 ¯L min¹1) in (a)
upstream, (b) midstream, and (c) downstream regions; red and
green lines are spectra recorded at the interface and within the
DCE phase, respectively; excitation at 405 nm. (d) Schematic
representation of the data points; each letter (a)-(c) corresponds
to the respective Raman spectrum.
Figure 3. AFM images of the self-assembled structures of
Chl-4Py; each sample was prepared immediately after eluting
from the microchannel (5 ¯m © 5 ¯m, mica substrate, dynamic
mode).
dried them under reduced pressure immediately. The AFM
images in Figure 3 (for selected magnified images, see
Figures S2 and S8a)11 reveal that Chl-4Py self-assembled to
form unique 1-D structures only when we employed its 2.5 mM
solution in DCE.8 Importantly, the sizes of the 1-D structures
were almost uniform in each of the samples, but their shapes
were affected by the applied flow rate. In the case of a
10 ¯L min¹1 flow rate, Chl-4Py self-assembled to form tape-like
structures, with an almost-flat cross section over the range;
higher flow rates (30 and 50 ¯L min¹1) gave 1-D structures
having curved surfaces (Figures S2 and S3).9,11 In contrast, other
concentrations of Chl-4Py (0.5 and 12.5 mM in DCE) gave only
dot-like structures. AFM observations of homogeneous DCE
solutions cast onto mica surfaces revealed similar dot-like
structures, implying that they were formed during the removal of
the solvent from the mica surface. Thus, the self-assembly of
Chl-4Py at the dynamic interface occurred within a narrow
concentration window. At a concentration of 0.5 mM, intermo-
lecular interactions were suppressed, even at the interface, owing
to the low local concentration at the initial Y-junction; in
contrast, at the higher concentration of 12.5 mM, most of the
Chl-4Py units eluted from the channel without effective contact
at the interface, leading to the creation of dot-like structures,
arising from monomeric Chl-4Py.
To further characterize the self-assembly of Chl-4Py, we
recorded UV-vis spectra of the DCE solutions immediately after
they had eluted from the channel (Figure S4).11 Unexpectedly,
we did not observe any changes in the UV-vis spectra of any of
the samples subjected to the flow conditions, suggesting the
absence of strong ³-³ stacking interactions among the Chl-4Py
molecules.10 The small but obvious Raman shifts in Figure 2
indicated that coordinative interactions played a critical role
in the 1-D self-assembly process. Similar Raman shifts were
evident for the same DCE solutions subjected to UV-vis spectral
analysis (Figure S5).11 From these results, the unique 1-D
structures observed in the AFM images appeared to maintain
water-DCE dynamic interface (Figure 2d). We found that Chl-
4Py existed in its monomeric form at the upstream junction,
whereas it tended to assemble at the middle and downstream
junction. Figures 2a-2c (for magnified spectra, see: Figure S1)11
display the spectra obtained from each position in the flow
channel marked in Figure 2d. At the upstream junction, the
¹1
Raman signal appeared at 1555 cm at the interface, consistent
with monomeric Chl-4Py. In contrast, in the middle and at the
downstream junction, the Raman signal at the interface appeared
at 1559 cm¹1, but that in the DCE phase had not undergone such
a shift. The relative intensity of the Raman signal in the DCE
phase decreased upon progressing from up- to downstream; the
corresponding intensity at the interface increased, with the
accompanying peak shift. These real-time measurements of
Raman spectra strongly supported the view that Chl-4Py
underwent self-assembly at the dynamic interface through Zn-
pyridine coordination. It should be emphasized that, unlike in
bulk solutions, the effect of the specific surface area becomes
predominant in a microchannel. Furthermore, as characteristic
features of the microflow, the diffusional mass-transfer proper-
ties of molecules are also enhanced in microspace. Thus, the
enhanced self-assembling ability of Chl-4Py in the microspace
would be essential for the effective intermolecular interactions
among Chl-4Py at the interface, leading to the creation of a 1-D
assembly.7 As a reference experiment, when water was carefully
added to the DCE solution of Chl-4Py, forming a centimeter-
scale interface, we could not observe any Raman spectral
changes in the DCE solution even after leaving it to stand for
several days, supporting the view that the self-assembling event
occurs rapidly in the microspace.
To obtain further evidence for Chl-4Py self-assembling at
the interface, we recorded AFM images of the eluted DCE
solutions. To avoid morphological transformations in DCE, we
cast the eluting solutions directly onto fresh mica surfaces and
Chem. Lett. 2012, 41, 1689-1691
© 2012 The Chemical Society of Japan
www.csj.jp/journals/chem-lett/

1691
partially supported by Grants-in-Aid for Scientific Research on
Innovative Areas “Reaction Integration,” Scientific Research (B)
(No. 22310069) and Exploratory Research (No. 23651118).
References and Notes
1
2
a) J.-M. Lehn, Supramolecular Chemistry: Concepts and
Perspectives, VCH, Weinheim, Germany, 1995. b) M. Fujita,
Chem. Soc. Rev. 1998, 27, 417.
For recent development of microchannel, see: a) P. J. A. Kenis,
R. F. Ismagilov, G. M. Whitesides, Science 1999, 285, 83. b)
K. P. Brazhnik, W. N. Vreeland, J. B. Hutchison, R. Kishore, J.
Wells, K. Helmerson, L. E. Locascio, Langmuir 2005, 21, 10814.
c) Y. Uozumi, Y. M. A. Yamada, T. Beppu, N. Fukuyama, M.
Ueno, T. Kitamori, J. Am. Chem. Soc. 2006, 128, 15994. d) R.
Karnik, F. Gu, P. Basto, C. Cannizzaro, L. Dean, W. Kyei-Manu,
R. Langer, O. C. Farokhzad, Nano Lett. 2008, 8, 2906. e) S.
Suga, D. Yamada, J.-i. Yoshida, Chem. Lett. 2010, 39, 404. f) M.
Pumera, Chem. Commun. 2011, 47, 5671. g) D. Kiriya, M. Ikeda,
H. Onoe, M. Takinoue, H. Komatsu, Y. Shimoyama, I. Hamachi,
S. Takeuchi, Angew. Chem., Int. Ed. 2012, 51, 1553.
Figure 4. Schematic representation of the concept of temporal
self-assembly at the flowing interface; AFM images of the self-
assembled structures of Chl-4Py before and after contact with
the interface (5 ¯m © 5 ¯m, mica substrate, dynamic mode). The
image of 1-D assembly was recorded after reinjection of
recycled DCE solution.
3
4
Our preliminary communications: a) M. Numata, Y. Takigami,
M. Takayama, Chem. Lett. 2011, 40, 102. b) M. Numata, D.
Kinoshita, N. Taniguchi, H. Tamiaki, A. Ohta, Angew. Chem.,
Int. Ed. 2012, 51, 1844. c) M. Numata, Y. Takigami, M.
Takayama, T. Kozawa, N. Hirose, Chem.®Eur. J. 2012, 18,
13008.
For conventional 1-D self-assembly of amphiphilic molecules,
see: a) T. Shimizu, M. Masuda, H. Minamikawa, Chem. Rev.
2005, 105, 1401. b) L. C. Palmer, S. I. Stupp, Acc. Chem. Res.
2008, 41, 1674. c) I. C. Reynhout, J. J. L. M. Cornelissen,
R. J. M. Nolte, Acc. Chem. Res. 2009, 42, 681. d) H.-J. Kim, T.
Kim, M. Lee, Acc. Chem. Res. 2011, 44, 72, and references cited
therein.
a) H. Tamiaki, A. R. Holzwarth, K. Schaffner, J. Photochem.
Photobiol., B 1992, 15, 355. b) V. Huber, M. Katterle, M.
Lysetska, F. Würthner, Angew. Chem., Int. Ed. 2005, 44, 3147. c)
R. F. Kelley, M. J. Tauber, M. R. Wasielewski, Angew. Chem.,
Int. Ed. 2006, 45, 7979. d) V. Huber, M. Lysetska, F. Würthner,
Small 2007, 3, 1007. e) H. Tamiaki, T. Michitsuji, R. Shibata,
Photochem. Photobiol. Sci. 2008, 7, 1225.
a) Y. Saga, H. Tamiaki, J. Photochem. Photobiol., B 2004, 75,
89. b) A. V. Ruban, A. A. Pascal, B. Robert, P. Horton, J. Biol.
Chem. 2002, 277, 7785.
their structures mainly through Zn-pyridine interactions, with
loose ³-³ stacking interactions merely supporting the 1-D
molecular arrangement. Calculated models for the assembly of
oligomeric Chl-4Py support this finding; ³-³ stacking inter-
actions were weak in the 1-D structures because of the long
pyridyl arms (Figure S6).11 Along the same line, Chl-OMe and
Chl-OH could not adopt such a 1-D molecular arrangement; they
did not give any identifiable assembled structure through the
microchannel accordingly.
Because Chl-4Py lacked the ability to form any self-
assembled structures in DCE, time-course Raman spectra
revealed that the shifted Raman peak assignable to the 1-D
assemblies (1562 cm¹1) gradually returned to its original
position (1554 cm¹1), assignable to monomeric Chl-4Py, within
1 h after eluting from the microchannel (Figure S5).11 AFM
images confirmed the dissociation of the assembled 1-D
structures; all of the 1-D structures had transformed into dot-
like structures after 1 h (Figure S8b).11 Importantly, reinjection
of the resultant DCE solutions into the microchannel led to
reconstruction of the 1-D self-assembled structures, depending
on the flow conditions (Figure 4). Thus, Chl-4Py displayed its
inherent 1-D self-assembling ability at the dynamic interface but
could maintain its resulting 1-D self-assembled structures only
under the flow conditions; these architectures are temporal and
renewable structures only under the flow conditions.
5
6
7
8
J. B. Knight, A. Vishwanath, J. P. Brody, R. H. Austin, Phys.
Rev. Lett. 1998, 80, 3863.
We have investigated the effect of temperature on self-assembled
property of Chl-4Py, with changing temperature of water phase at
4 and 65 °C. As a result, we could not observe any identical
structures by AFM observation under such conditions.
To investigate the effect of water dissolving into DCE phase,
we used water-saturated DCE for preparing a Chl-4Py solution.
However, we could not observe any self-assembled structures in
such “wet” DCE solution, implying that observed unique 1-D
structures were certainly formed at the dynamic microinterface.
In summary, we have demonstrated that a microfluidic
device can provide a novel self-assembling field within a
flowing microspace, forming an oil-water interface that is
indispensable for the temporal self-assembly of flowing amphi-
philic molecules. One of the fundamental features of natural
molecular systems®never before realized in an artificial
system®is the use of time as a dimension in supramolecular
assembly processes. We believe that this present system will
open up new opportunities for the development of novel
supramolecular structures.
9
10 To avoid any transformations of the obtained self-assembled
structures, we also injected the DCE solution directly into a flow
cell immediately after its elution from the microfluidic device.
Even under these conditions, however, we could not observe any
shifts of the signals in the UV-vis spectra.
11 Supporting Information is available electronically on the CSJ-
Journal Web site, http://www.csj.jp/journals/chem-lett/index.
html.
We thank Dr. K. Sugiyasu and Dr. M. Takeuchi at NIMS
for kindly supporting our SEM observations. This study was
Chem. Lett. 2012, 41, 1689-1691
© 2012 The Chemical Society of Japan
www.csj.jp/journals/chem-lett/