One-Directional Fluidic Flow Induced by Chemical Wave Propagation

Article pubs.acs.org/JPCB

One-Directional Fluidic Flow Induced by Chemical Wave Propagation in a Microchannel Miyu Arai, Kazuhiro Takahashi, Mika Hattori, Takahiko Hasegawa, Mami Sato, Kei Unoura, and Hideki Nabika* Department of Material and Biological Chemistry, Faculty of Science, Yamagata University, 1-4-12 Kojirakawa, Yamagata 990-8560, Japan S Supporting Information *

ABSTRACT: A one-directional flow induced by chemical wave propagation was investigated to understand the origin of its dynamic flow. A cylindrical injection port was connected with a straight propagation channel; the chemical wave was initiated at the injection port. Chemical waves propagated with a constant velocity irrespective of the channel width, indicating that the dynamics of the chemical waves were governed by a geometry-independent interplay between the chemical reaction and diffusion. In contrast, the velocity of the one-directional flow was dependent on the channel width. Furthermore, enlargement of the injection port volume increased the flow velocity and volume flux. These results imply that the one-directional flow in the microchannel is due to a hydrodynamic effect induced in the injection port. Spectroscopic analysis of a pH indicator revealed the simultaneous behavior between the pH increase near the injection port and the one-directional flow. Hence, we can conclude that the one-directional flow in the microchannel with chemical wave propagation was caused by a proton consumption reaction in the injection port, probably through liquid volume expansion by the reaction products and the reaction heat. It is a characteristic feature of the present system that the hydrodynamic flow started from the chemical wave initiation point and not the propagation wavefront, as observed for previous systems. between two liquids causes complicated fluidic flow at the interface. Chemical reactions can also couple with diffusion under nonequilibrium conditions. This is known as the reaction− diffusion (RD) systemthe origin of spatiotemporal pattern formation in nature. A chemical wave, which is a type of RD phenomenon, involves an autocatalytic chemical reaction that creates local chemical and/or physical gradients at the propagation front. In the Belousov−Zhabotinsky (BZ) reaction, the wave propagation can be initiated by touching a Ag wire to an excitable medium containing several reagents.19 The autocatalytic reaction in HBrO2 proceeds at the initiation point. The increased HBrO2 concentration at the initiation point causes its diffusion into the surroundings, where the autocatalytic reaction in HBrO2 proceeds in the surroundings of the initiation point. As such, the position of chemical reaction propagates with time. Under certain well-controlled conditions, the experiment produces a train of BZ waves departing from the initiation point.20,21 Since the propagation front involves the chemical reaction that releases the reaction product and the reaction heat, a chemical and/or physical gradient is always formed between the front and behind of the

1. INTRODUCTION Spatiotemporally controlled and self-powered fluidic flows under nonequilibrium conditions are ubiquitous in nature, ranging from supernovae in space1 to “coffee-rings” on our tables.2 A physicochemical understanding on these fluidic flows helps us to learn the origin of the robustness, exquisiteness, and simplicity of nature. Furthermore, the spatiotemporal control of fluidic flow offers self-powered pumps for installation in nextgeneration smart miniature devices.3 In view of chemical systems, many kinds of fluidic flow under nonequilibrium conditions have been reported, some of which can be used as a model system for natural phenomena. One simple example is the Rayleigh−Taylor instability, which is seen when an interface between two liquids with different densities is placed in a gravitational field with a direction opposite to the density gradient.4−8 The opposed gradient imposes the instability and fluidic flow at the interface, which is considered as the acceleration mechanism for thermonuclear flames in supernovae.1 In addition, Rayleigh−Bénard9−11 and Bénard− Marangoni12−15 instabilities are induced in the horizontal liquid layer with perpendicular temperature gradients. Although these instabilities can occur without the accompanying chemical reactions, coupling between the instability and chemical reactions causes another fluidic mode known as the chemohydrodynamic flow.16−18 The chemical reaction at the interface © 2016 American Chemical Society

Received: March 18, 2016 Revised: May 10, 2016 Published: May 11, 2016 4654

DOI: 10.1021/acs.jpcb.6b02850 J. Phys. Chem. B 2016, 120, 4654−4660

Article

The Journal of Physical Chemistry B propagating reaction front. Thus, the propagation front could be an active field to create an instability-induced fluidic flow. Both experimental and numerical experiments have revealed fluidic flow at the front of propagating chemical waves.22−25 For example, temperature mapping with the aid of magnetic resonance imaging revealed that the increased temperature at the reaction front of the chlorite−thiosulfate system clearly induced fluidic flows with several modes.26 Since the change in temperature and chemical composition at the reaction front alters the density, the buoyancy force acts as an accelerator or a decelerator for chemical wave propagation, depending on the relative angle between the propagation direction and the gravitational force.25 Furthermore, convective flow induced by the Marangoni effect at the propagating front can act as a microfreight that traps and transports materials floating on the propagation medium.19 As such, the propagation reaction fronts have a spatiotemporally dynamic feature that can be explained by considering both hydrodynamic instability and RD. Although most reports have focused attention on the behavior of dynamic reaction fronts, the initiation point would be another dynamic point that involves successive chemical reactions. The reaction front of a BZ wave involves an autocatalytic reaction in HBrO2 with a certain reaction heat, which produces a chemical and/or physical gradient. In contrast, at the initiation point, the oxidation reaction of an organic acid ignites after the reaction front departs. This reaction at the initiation point also gives reaction products and reaction heat, which can demonstrate a dynamic spatiotemporal behavior. Coupling between the dynamic behaviors between the front and initiation point would open up novel hydrodynamic modes. To make clear the presence of a hydrodynamic effect originating from the chemical reactions at the initiation point, we have used proton wave propagation with the bromate−thiosulfate system, in which proton production and consumption reactions proceed successively at the initiation point. In the analogy of the iodate−thiosulfate system,27 the proton production and consumption reactions would proceed as follows, respectively. BrO3− + 3HSO3− + H+ → Br − + 3SO4 2 − + 4H+

Figure 1. Observation setup. The polydimethylsiloxane (PDMS) microchannel was placed on a glass substrate. The microchannel was filled with the propagation medium containing the substrates, pH indicator, and tracer microspheres; H2SO4 was gently added into the injection port. The chemical wave propagation and flow behavior were observed under a fluorescence microscope with green light excitation. The velocities of the chemical wave and flow were defined as positive when they move from the left (injection side) to the right (channel side) in this configuration.

mM NaOH, a pH indicator (phenol red), and microspheres; the solution was poured into the microchannel through the injection port. Temperature and pH of the poured initial solution were ca. 22 °C and 8, respectively. The volume of poured solution was tuned from 10.06−18.56 mm3 depending on the channel width. The microspheres (hydrophilic fluorescent polystyrene (PS) microspheres with a diameter of 1 μm, Fluoro-Max R0100, Thermo Fisher Scientific Inc.) were added to visualize the fluidic motion in the microchannel with fluorescence microscopy (BX-53, Olympus, Japan). Phenol red was used to visualize the propagation front and to spectroscopically measure the solution pH. Then, 1 mL of 0.1 M H2SO4 was gently added to the injection port, which was previously found to be enough to initiate and propagate the proton wave.28 The observation of the propagation front and fluidic flow was conducted 4 mm from the injection port to avoid geometrydependent proton flux modulation.28

(1)

BrO3− + 6S2 O32 − + 6H+ → Br− + 3S4 O6 2 − + 3H 2O (2)

The proton production reaction 1 is autocatalytic with respect to the protons that propagate with time and are observed as the proton wave. When the proton concentration at the initiation point exceeds a certain threshold, the reaction switches to the proton consumption reaction 2. Since the consumption reaction involves water production, it can be expected to induce the hydrodynamic behavior ignited from the initiation point. This work derives a fluidic flow in the microchannel that propagates the proton wave, which is discussed in terms of the reaction products and reaction heat.

3. RESULTS The pH of the mixed solution prior to proton wave propagation is nearly 8; thus, the phenol red is a bright red color. Under the fluorescence microscope with green light excitation, phenol red at this pH condition shows a slight emission. Since fluorescent PS microspheres are also designed to have an emissive nature under green light excitation, both the microspheres and phenol red were clearly observed before reaching the proton wave propagation to the observation area (Figure 2A). The proton wave reached the left edge of the observation area 1100 s after the injection of H2SO4. When the solution pH

2. EXPERIMENTAL SECTION Proton wave propagation experiments were performed in a straight polydimethylsiloxane microchannel (Fluidware Technologies Inc., Japan) (Figure 1). The straight channel with a depth of 100 μm is connected with cylindrical injection ports with a diameter of 2 mm. The channel width (w) was changed from 300 to 2000 μm. The solution for the proton wave propagation consisted of 70 mM Na2S2O3, 100 mM KBrO3, 2 4655


Article

The Journal of Physical Chemistry B

Figure 2. (A) Fluorescence microscopy images of a proton wave propagating from left to right. The interface between the red and black background color is the proton wavefront. Tracer microspheres are also observed under the fluorescence microscope. (B) Temporal change of the position of the microspheres. Note that immobile microspheres settled on the bottom of the glass substrate.

Figure 3. (A) Velocities of proton wave propagation (black) and fluidic flow (red) as a function of channel width. (B) Flow distance measured by the average transported distance of the microspheres. (C) Volume flux calculated by the flow distance and cross-sectional area of each microchannel.

and their movement stopped and returned to Brownian motion. It should be noted here that the microsphere underwent the Brownian motion both before 3240 s and after 3590 s. Thus, the fluidic flow was found to occur only for a limited duration between 3290 and 3590 s. The average velocity of the fluidic flow was ca. 6 μm/s. Since this value is three times higher than that of proton wave propagation, it can be assumed that there is no relation between the proton wave propagation and the fluidic flow. Note that the immobile bright spots in Figure 2B are the fluorescent microspheres that settled on the bottom of the glass substrate. Also, although the addition of H2SO4 into the inlet causes rapid fluidic flow in the microchannel, it calms down within a few hundred seconds (see the Supporting Information). Thus, the influence of the rapid flow upon the H2SO4 addition can be ignored for both the wave propagation (>1000 s) and the microsphere movements (>3000 s) events. To distinguish the origins of the proton wave and fluidic flow, we have conducted experiments using microchannels with different values for w (Figure 3A). The proton waves propagated at almost the same velocity irrespective of w. This result is reasonable because the proton wave originated from the RD mechanism, which is governed by the chemical reaction and the diffusion rates. Since these two rates are completely dependent on the nature of the constituent reactive species, the propagation velocity should not be dependent on w. On the

decreased below 6, the color of phenol red changed to pale yellow and its emissive nature under green light disappeared. As a result, the background color of the left side (where the proton wavefront reached) turned black (no emission). On the other hand, fluorescent PS microspheres can be visualized in both high- and low-pH regions. Thus, the propagation of the proton wave from left to right can be clearly observed without harming the visualization of the fluorescent microspheres. With a microchannel of w = 1000 μm, the proton wave passed through the observation area in the time period 1100−2030 s, yielding an averaged propagation velocity of 2.2 μm/s, which agrees with a previously reported value under similar conditions.28 It should be also noted that the effect of H2SO4 diffusion (proton flux) from the inlet is negligible since our observation area is far enough (4 mm) away from the inlet, and the observed wave propagation can be assumed to be governed by a purely reaction diffusion mechanism.28 After the passage of the proton wave, a sudden and rapid fluid flow appeared in the microchannel that was visualized by the microspheres (Figure 2B). Only Brownian motion was observed at around 3240− 3290 s, indicating no fluidic flow at this time. At 3340 s, the movement of some microspheres toward the right side indicates the start of the fluid flow. At 3390 and 3440 s, the microspheres were blurred in the images due to their significantly fast fluidic flow compared with the exposure time (ca. 1 s). The microspheres reached the right edge at 3590 s 4656


Article

The Journal of Physical Chemistry B other hand, it is clear that the velocity of the fluidic flow was highly dependent on w; a decrease of w accelerated the fluidic flow. Judging from the difference in the w-dependence of the proton wave and fluidic flow, our assumption that the fluidic flow is not caused by only proton wave propagation would be reasonable. Figure 3B shows the average flow distance as a function of w, which was measured by the averaged onedirectionally transported distance of the microspheres. The velocity showed similar w-dependence; narrower channels induced fluidic flow with a higher velocity and longer distances. From the flow distance and cross-sectional area of each microchannel, we have evaluated the volume of liquid moved by the fluidic flow (Figure 3C) according to a simple geometrical calculation: (flow distance) × (channel crosssectional area) = (volume flux). Although the data become relatively scattered at larger values of w during the numerical operations, the statistical analysis of the data showed that the volumes were not dependent on w within the experimental error. These results imply that the fluidic flow would originate by something not affected by the geometric feature of the channel region. Since a common feature for microchannels with any value of w is the design of the injection port (cylindrical hole with a diameter of 2 mm), one possible origin of the windependent fluidic flow would be the hydrodynamic phenomena induced in the injection port. Thus, the hydrodynamic phenomena induced inside the injection port could propagate into the channel region, where the narrower channel with a lower cross-sectional area yielded to push the fluid in the channel region at a faster velocity and over a longer distance. In the above discussion, we have directly related the microsphere movement to the fluidic flow in the microchannel. To clarify its validity, we have conducted a depth-controlled microscope observation by focusing the observation depth at the bottom, middle, and top of the microchannel. In the case of the proton wave propagation velocity, the velocity histograms demonstrate almost the same distribution irrespective of the observation depth (Figure 4). The proton wave propagates as a plane wave in the microchannel. This is rather reasonable in the same analogy to the above-mentioned w-independent velocity. However, a depth dependence was seen for the velocity of the microspheres. At the bottom and top of the microchannel, the velocity of the microspheres was in the range of 0−4 μm/s. The microspheres moved faster in the mid-depth region, with a maximum velocity of ca. 6 μm/s (e.g., the microsphere shown in Figure 2B). Thus, the velocity distribution of the microspheres was parabolic with a higher velocity at the middepth. Since such a velocity distribution is characteristic of a fluidic flow under laminar flow, our assumption that the microsphere movement represents the fluidic flow inside the microchannel would be reasonable. The velocity of all microspheres was positive, indicating that all microspheres moved in the same direction away from the injection port. Considering that most hydrodynamic flows such as Marangoni flow coupled with a chemical wave are convective, this is one characteristic feature of our system to show one-directional fluidic flow. The absence of the Marangoni effect is also reasonable because our microchannel is fully filled with water and there is no air/water interfaces. To verify our hypothesis that the injection port is responsible for the observed hydrodynamic phenomena inside the microchannel, we used microchannels with an enlarged injection port (diameter = 5 mm). Figure 5A shows the

Figure 4. Depth-dependent velocities of (left) the proton wave and (right) the microspheres.

temporal change of the microsphere positions by the fluidic flow in the microchannel with the 2 mm injection port. As can be seen in Figure 2B, the microsphere moved for ca. 1500 μm during the fluidic flow over a few hundred seconds. By using the microchannel with an enlarged injection port, we were able to increase the volume of liquid inside the injection port from 4.9 to 29.9 mm3 without changing any other geometry. The temporal changes of the microsphere positions (Figure 5B) also represent the appearance of fluidic flow in the microchannel with an enlarged injection port. The period of the fluidic flow was ca. 500 s, which was found to be similar to the results for the normal channel (Figure 5A). However, a significant difference can be noted when the flow distances were compared between these two systems: 1500 μm for the normal microchannel and 15 000 μm for the enlarged microchannel. Thus, it was verified that the fluidic flow observed in the channel region originated and was strongly affected by the hydrodynamic phenomena inside the injection port.

4. DISCUSSION The results shown in Figures 3 and 5 clarified that the fluidic flow in the present system originated from a hydrodynamic effect in the injection port. One probable cause is the progress of the proton consumption reaction in the injection port, which follows the autocatalytic proton production reaction that forms the proton wave propagation. Thus, it is necessary to know whether the period of the fluidic flow is consistent with that of the proton consumption reaction. As shown in Figure 2B, the fluidic flow appears at 3000−4000 s. To know the progress of the proton consumption reaction, we have acquired micro4657


Article


consumption reaction inside the injection port. Since the observed time for the proton consumption reaction in the injection port was the same as that for the appearance of the fluidic flow in the channel region, we conclude that they are closely related to each other. It should be noted here that the times for proton production and consumption reactions can be tuned by the choice of constituents in the reaction solution because of the difference in the chemical reaction rates.27 Since the fluidic flow would be related to the proton consumption reaction, the ignition time for the fluidic flow would be controlled by the design of chemical reactions induced in the inlet. Next, we further considered the origin of the fluidic flow induced at the same period of the proton consumption reaction. Equation 2 indicates that the proton production reaction provides the water molecule as a product. The increase in water volumes were estimated to 0.00033 mm3 and 0.0020 mm3 for microchannels with 2 mm and 5 mm ports, respectively, based on the pH increase of 2−2.6 (Figure 6) and the liquid volume (Vliquid). To maintain the same water levels between the injection port and the other side port, nearly half of the increased water at the injection port should be moved to the other port through the channel region. This water flow could be the origin of the one-directional fluidic flow observed in the microchannels. However, the estimated water production volume is too small to explain quantitatively the observed flow volume in the microchannel. Another important change caused by the proton production reaction is the reaction heat, which also expands the water volume in the injection port and causes the water to flow into the channel region. Although the temperature increase in the present system cannot be measured, previously reported chemical waves generated detectable temperature increases by a few Kelvin to a few tens of Kelvin.25,26,29 A temperature increase of 25 K would be required (within the range of the reported temperature increase) for a volume increase of 0.05 mm3 at the injection port (Figure 3C). Further experiments based on a temperature imaging technique26 would clarify the temporal relationship between the temperature change in the injection port and the fluidic flow in the channel region. Also, simulations considering spatiotemporal profiles of (1) reaction heat production, (2) thermal conduction, and (3) spatiotemporal water production would offer a quantitative understanding on the duration of the fluidic flow. The present study has clarified that the hydrodynamic phenomena in the injection port, possibly brought from the reaction product and heat, is responsible for the one-directional fluidic flow induced in the microchannel region. The initiation point that involved successive chemical reactions could be one of the dynamic parameters that could control the hydrodynamic flow coupling with RD systems, not only the propagation front.

Figure 5. Temporal changes of microsphere positions induced by the fluidic flow in the microchannel with (A) normal (dhole = 2 mm) and (B) enlarged (dhole = 5 mm) injection ports. Vhole is the volume of the injection port, and Vliquid is the volume of liquid (water) inside the injection port.

scopic spectra around the injection port. We obtained information about the pH change around the injection port from a comparison between the acquired spectra and the standard spectra of phenol red at various pH values. Figure 6 shows the pH around the injection port, including the pH before adding H2SO4 at time zero.

5. CONCLUSIONS We investigated the relationship between the chemical wave propagation and one-directional fluidic flow in microchannels. Microscope imaging clearly demonstrated a difference between their dynamics. The chemical wave propagated with a constant velocity during the observation, whereas the one-directional flow appeared for only a limited period. Furthermore, the dependence on w was also different. The chemical wave propagation velocity was independent of w because its origin is an RD effect. In contrast, the flow velocity and flow distance were strongly dependent on w; the flow with a higher velocity

Figure 6. Temporal change of solution pH near the injection port of the microchannel with normal (red) and enlarged (black) injection ports. The blue region in the inset displays the period of pH increase.

H2SO4 addition decreased the solution pH to ca. 2 due to the progress of the autocatalytic proton production reaction. Then, the pH became stable, indicating that no reactions that changed the proton concentration proceeded during this period. After the stable state, the pH gradually increased for the period 3000−4000 s for both the microchannels with diameters of 2 mm and 5 mm, indicating the progress of the proton 4658


Article


(8) Shim, J.; Yun, J. M.; Yun, T.; Kim, P.; Lee, K. E.; Lee, W. J.; Ryoo, R.; Pine, D. J.; Kim, S. O. Two-Minute Assembly of Pristine LargeArea Graphene Based Films. Nano Lett. 2014, 14, 1388−1393. (9) Velegol, D.; Shori, S.; Snyder, C. E. Rayleigh-Bénard Instability in Sedimentation. Ind. Eng. Chem. Res. 2009, 48, 2414−2421. (10) Mühlenhoff, S.; Eckert, K.; Heinze, A.; Uhlemann, M. Controlled Retardation of Electrochemical Rayleigh-Bénard Convection during Copper Electrolysis. J. Electroanal. Chem. 2007, 611, 241−249. (11) Jerri, H. A.; Sheehan, W. P.; Snyder, C. E.; Velegol, D. Prolonging Density Gradient Stability. Langmuir 2009, 26, 4725− 4731. (12) Uchiyama, H.; Matsui, T.; Kozuka, H. Spontaneous Pattern Formation Induced by Benard-Marangoni Convection for Sol-GelDerived Titania Dip-Coating Films: Effect of Co-Solvents with a High Surface Tension and Low Volatility. Langmuir 2015, 31, 12497− 12504. (13) Uchiyama, H.; Miki, Y.; Mantani, Y.; Kozuka, H. Spontaneous Formation of Micrometer-Scaled Cell-like Patterns on AlkoxideDerived Silica Gels Induce by Bénard-Marangoni Convection. J. Phys. Chem. C 2012, 116, 939−946. (14) Uchiyama, H.; Mantani, Y.; Kozuka, H. Spontaneous Formation of Linearly Arranged Microcraters on Sol-Gel-Derived Silica-Poly(vinylpyrrolidone) Hybrid Films Induced by Bénard-Marangoni Convction. Langmuir 2012, 28, 10177−10182. (15) Cui, L.; Han, Y. Honeycomb Pattern Formation via Polystyrene/Poly(2-vinylpyridine) Phase Separation. Langmuir 2005, 21, 11085−11091. (16) Bratsun, D. A.; Shi, Y.; Eckert, K.; De Wit, A. Control of Chemo-Hydrodynamic Patterm Formation by External Localized Cooling. Europhys. Lett. 2005, 69, 746−752. (17) Almarcha, C.; Trevelyan, P. M. J.; Grosfils, P.; De Wit, A. Chemically Driven Hydrodynamic Instabilities. Phys. Rev. Lett. 2010, 104, 044501−044501. (18) Almarcha, C.; R’Honi, Y.; De Decker, Y.; Trevelyan, P. M. J.; De Wit, A.; Eckert, K. Convective Mixing Induced by Acid-Base Reactions. J. Phys. Chem. B 2011, 115, 9739−9744. (19) Ichino, T.; Asahi, T.; Kitahata, H.; Magome, N.; Agladze, K.; Yoshikawa, K. Microfreight Delivered by Chemical Waves. J. Phys. Chem. C 2008, 112, 3032−3035. (20) Miike, H.; Müller, S. C.; Hess, B. Oscillatory Deformation of Chemical Waves Induced by Surface Flow. Phys. Rev. Lett. 1988, 61, 2109−2112. (21) Miike, H.; Müller, S. C.; Hess, B. Oscillatory Hydrodynamic Flow Induce by Chemical Waves. Chem. Phys. Lett. 1988, 144, 515− 520. (22) Masere, J.; Vasquez, D. A.; Edwards, B. F.; Wilder, J. W.; Showalter, K. Nonaxisymmetric and Axisymmetric Convection in Propagating Reaction-Diffusion Fronts. J. Phys. Chem. 1994, 98, 6505− 6508. (23) Wu, Y.; Vasquez, D. A.; Edwards, B. F.; Wilder, J. W. Convective Chemical-Wave Propagation in the Belouzov-Zhabotinsky Reaction. Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top. 1995, 51, 1119−1127. (24) Pojman, A.; Epstein, I. R. Convective Effects on Chemical Waves. 1. Mechanisms and Stability Criteria. J. Phys. Chem. 1990, 94, 4966−4972. (25) Komlosi, A.; Nagy, I. P.; Bazsa, G.; Pojman, J. A. Convective Chemical Fronts in the 1,4-Cyclohexanedione-Bromate-Sulfuric AcidFerroin System. J. Phys. Chem. A 1998, 102, 9136−9141. (26) Zhivonitko, V.; Koptyug, I. V.; Sagdeev, R. Z. Temperature Changes Visualization during Chemical Wave Propagation. J. Phys. Chem. A 2007, 111, 4122−4124. (27) Rabai, G.; Orban, M.; Epstein, I. R. Design of pH-Regulated Oscillators. Acc. Chem. Res. 1990, 23, 258−263. (28) Nabika, H.; Hasegawa, T.; Unoura, K. Propagation Behaviors of an Acid Wavefront Through a Microchannel Junction. J. Phys. Chem. B 2015, 119, 9874−9882.

and longer distance appeared in the narrower channel. However, the flow volume was constant. These results were explained by a simple geometrical calculation based on the flow distance and channel cross-sectional area. This simple idea implied that the observed one-directional flow was not initiated in a w-dependent region; we then focused on the effect of the injection port. We conducted the same experiments with a microchannel with an enlarged injection port, where a larger flow volume appeared, as we expected. From these experiments, we conclude that the observed one-directional flow in the microchannel originates from a hydrodynamic effect in the injection port. Spectroscopic analysis of the pH indicator revealed the simultaneous behavior between the pH increase and the one-directional flow. These results can associate the progress of a proton consumption reaction and the onedirectional flow phenomenon, probably through the liquid volume expansion by the reaction products and the reaction heat. Our findings offer unique hydrodynamic effect coupling with the chemical wave propagation, where it would be a characteristic feature that the hydrodynamic effect was not caused at the reaction front but at the initiation point.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.6b02850. Control experiments on the effect of droplet injection on the fluidic flow (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Tel.: +81-23-628-4589. E-mail: [email protected]. jp. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This study was supported in part by a Grant-in-Aid for Young Scientists (A) 25708012 from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japanese Government.

■

REFERENCES

(1) Cabot, W. H.; Cook, A. W. Reynolds Number Effects on Rayleight-Taylor Instability with Possible Implications for Type-la Supernovae. Nat. Phys. 2006, 2, 562−568. (2) Hu, H.; Larson, R. G. Marangoni Effect Reverses Coffee-Ring Depositions. J. Phys. Chem. B 2006, 110, 7090−7094. (3) Ortiz-Rivera, I.; Shum, H.; Agrawal, A.; Sen, A.; Balazs, A. C. Convective Flow Reversal in Self-Powered Enzyme Micropumps. Proc. Natl. Acad. Sci. U. S. A. 2016, 113, 2585−2590. (4) Yang, J.; D’Onofrio, A.; Kalliadasis, S.; De Wit, A. RayleighTaylor Instability of Reaction-Diffusion Acidity Fronts. J. Chem. Phys. 2002, 117, 9395−9408. (5) Arosio, P.; Müller, T.; Mahadevan, L.; Knowles, T. P. J. DensityGradient-Free Microfluidic Centrifugation for Analytical and Preparative Separation of Nanoparticles. Nano Lett. 2014, 14, 2365−2371. (6) Yang, X.; Tschulik, K.; Uhlemann, M.; Odenbach, S.; Eckert, K. Enrichment of Paramagnetic Ions from Homogeneous Solutions in Inhomogeneous Magnetic Field. J. Phys. Chem. Lett. 2012, 3, 3559− 3564. (7) Morris, M. L.; Baird, L. M.; Panigrahi, A.; Gross, M. C.; Deacon, R. M.; Benkoski, J. J. Surfactant Sculpting of Biologically Inspired Hierarchical Surfaces. Soft Matter 2013, 9, 9857−9866. 4659


Article

The Journal of Physical Chemistry B (29) Nagy, I. P.; Pojman, J. A. Multicomponent Convection Induced by Fronts in the Chlorate-Sulfite Reaction. J. Phys. Chem. 1993, 97, 3443−3449.

4660


One-Directional Fluidic Flow Induced by Chemical Wave Propagation

Recommend Documents