Electrochemical Generation of Gradients in Surfactant Concentration

Anal. Chem. 2009, 81, 772–781

Electrochemical Generation of Gradients in Surfactant Concentration Across Microfluidic Channels Xiaoyang Liu and Nicholas L. Abbott* Department of Chemical and Biological Engineering, University of Wisconsin—Madison, 1415 Engineering Drive, Madison, Wisconsin 53706-1691 We report the generation and manipulation of spatial gradients in surfactant and micelle concentration across microfluidic channels by combining use of a redox-active surfactant with electrochemical methods. The approach is founded on the observation that 11-ferrocenylundecyltrimethylammonium bromide (FTMA) behaves as a surfactant in aqueous solution (e.g., self-assembles to form micelles at a critical concentration of 0.1 mM in aqueous 0.1 M Li2SO4) whereas oxidized FTMA remains dispersed in a monomeric state up to concentrations of at least 30 mM. By flowing aqueous FTMA solutions through microfluidic channels (width of 80 µm, depth of 72 µm, and length of 42 mm) and by applying potentials of 0 V (vs Ag|AgCl; cathode) and +0.3 V (vs Ag|AgCl; anode) to gold electrodes lining both sidewalls of the microfluidic channels, we measured lateral gradients in concentration of oxidized FTMA and reduced FTMA to be generated across the microfluidic channels by splitting the exiting stream into four channels. These measurements revealed the lateral concentration profile of FTMA to be consistent with the presence of slowly diffusing micelles of FTMA in a spatially localized region near the cathode and monomeric FTMA only near the anode. The lateral concentration profiles of reduced and oxidized FTMA, and thus the patterning of micelles within the microfluidic channels, were manipulated via changes in the inlet FTMA concentration, potentials applied to the electrodes, and flow rate. These experimental measurements were compared to a simple model, which assumed fast electrode kinetics, lateral transport of FTMA by diffusion only (no migration), and local micelle-monomer equilibrium within the bulk solution. This comparison revealed qualitative but not quantitative agreement between model and experiment. Calculations of ionic conductivity and associated experimental measurements support the proposition that Ohmic resistance to the passage of current along the channel (between the working and the counter electrodes) contribute, in part, to the lack of quantitative agreement between the model and the measurements. The capability to generate and manipulate lateral concentration profiles of surfactants and micelles across microfluidic channels, as demonstrated by the 772

Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

results presented in this paper, offers the basis of new principles for continuous separation processes and microanalytical systems, and more broadly, new methods to generate gradients in concentration of analytes that interact with surfactants. A wide range of physicochemical phenomena are known to be influenced by the presence of gradients in the concentration of surfactants at mobile interfaces (e.g., oil-water and water-air).1-4 These gradients in surfactant concentration can give rise to interfacial stresses that result in convection at the interface. Such convective processes often dominate interfacial transport of molecules, typically giving rise to far greater fluxes of molecules than result from molecular diffusion. These socalled “Marangoni flows”5 have been reported to accelerate the spreading of fluids,6 enhance the stability of surfactant-laden films, emulsions, and foams,1 lead to the assembly of ordered patterns of nanoparticles,7 and increase rates of mass and energy transfer at interfaces.8 Although a variety of experimental approaches exist to generate gradients in surfactant concentration at interfaces, including localized deposition of monolayers of surfactants,3 these approaches typically result in transient interfacial flows that diminish as the concentration of surfactant equilibrates uniformly throughout the system. Recently, we reported a new methodology to generate gradients in surfactant concentration at interfaces that employs redoxactive surfactants that undergo reversible changes in oxidation state.9,10 Because the change in oxidation state of the redoxactive surfactant leads to a substantial alteration of the am* To whom correspondence should be addressed. E-mail: [email protected]. Phone: +1-608-265-5278. Fax: +1-608-262-5434. (1) Pesach, D.; Marmur, A. Langmuir 1987, 3, 519–524. (2) Leenaars, A. F. M.; Huethorst, J. A. M.; Vanoekel, J. J. Langmuir 1990, 6, 1701–1703. (3) Peng, J. B.; He, S. X.; Dutta, P.; Ketterson, J. B. Phys. Rev. A 1989, 40, 7421–7423. (4) Cazabat, A. M.; Heslot, F.; Troian, S. M.; Carles, P. Nature 1990, 346, 824–826. (5) Marangoni, C. S. M.; Stefanilli, P. Nuovo Cimento Ser. 2 1872, 516, 239. (6) Nikolov, A. D.; Wasan, D. T.; Chengara, A.; Koczo, K.; Policello, G. A.; Kolossvary, I. Adv. Colloid Interface Sci. 2002, 96, 325–338. (7) Cai, Y.; Newby, B. M. Z. J. Am. Chem. Soc. 2008, 130, 6076. (8) Mcgrew, J. L.; Bamford, F. L.; Rehm, T. R. Science 1966, 153, 1106. (9) Gallardo, B. S.; Gupta, V. K.; Eagerton, F. D.; Jong, L. I.; Craig, V. S.; Shah, R. R.; Abbott, N. L. Science 1999, 283, 57–60. (10) Bennett, D. E.; Gallardo, B. S.; Abbott, N. L. J. Am. Chem. Soc. 1996, 118, 6499–6505. 10.1021/ac801933v CCC: $40.75  2009 American Chemical Society Published on Web 12/16/2008

phiphilicity of the molecule (and thus the tendency of the surfactant to lower surface tension) and because the change in oxidation state can be controlled reversibly at electrodes located near an interface, the approach made possible the generation of steady-state and time-varying interfacial gradients in surfactant concentration and associated Marangoni flows.9-12 In this paper, we build on this methodology to investigate the use of redox-active surfactants and electrochemical methods to generate spatial gradients in surfactant and micelle concentration within the bulk of aqueous solutions. We employ the use of microfluidic channels in this study as they provide a convenient means to generate and characterize these gradients, as well as define experimental systems in which the gradients might be exploited for microscale separations and analysis. The redox-active surfactant used in our study is 11ferrocenylundecyltrimethyl-ammonium bromide (FTMA) (Figure 1A).9-27 Past studies have demonstrated that FTMA can be electrochemically reduced and oxidized over numerous cycles within a window of electrochemical potentials that does not lead to electrolysis of water.16 Whereas the reduced state of FTMA (in which the ferrocene is not charged, left side of Figure 1A) exhibits properties that are typical of those of small amphiphiles (such as a critical micelle concentration at 0.1 mM in aqueous 0.1 M Li2SO4), oxidized FTMA (in which ferrocene is oxidized to the ferrocenium cation, right side of Figure 1A) does not measurably self-associate over the experimental conditions investigated to date (up to 30 mM in 0.1 M Li2SO4).17 Recently, we determined the electrochemical characteristics of a thin film electrochemical cell containing FTMA.28 Following application of oxidizing and reducing potentials to each electrode within the thin film electrochemical cell, we measured steady-state currents of equal magnitude at each electrode, consistent with the absence of undesirable Faradaic processes at the electrodes (e.g., hydrolysis of water) and an overall electrochemical rate process that was limited by mass transport of FTMA between the electrodes and not by electrode kinetics. Within the thin-film electrochemical cell, however, it was not possible to characterize the concentration profiles of FTMA and oxidized FTMA between the electrodes. (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27) (28)

Bai, G.; Graham, M. D.; Abbott, N. L. Langmuir 2005, 21, 2235–2241. Bai, G. Y.; Abbott, N. L.; Graham, M. D. Langmuir 2002, 18, 9882–9887. Aydogan, N.; Abbott, N. L. Langmuir 2001, 17, 5703–5706. Aydogan, N.; Gallardo, B. S.; Abbott, N. L. Langmuir 1999, 15, 722–730. Aydogan, N.; Rosslee, C. A.; Abbott, N. L. Colloids Surf., A 2002, 201, 101–109. Gallardo, B. S.; Hwa, M. J.; Abbott, N. L. Langmuir 1995, 11, 4209–4212. Gallardo, B. S.; Metcalfe, K. L.; Abbott, N. L. Langmuir 1996, 12, 4116– 4124. Hattori, T.; Tanaka, S. Electroanalysis 2003, 15, 1522–1528. Hays, M. E.; Abbott, N. L. Langmuir 2005, 21, 12007–12015. Hoshino, K.; Goto, M.; Saji, T. Chem. Lett. 1988, 547–550. Hoshino, K.; Saji, T. J. Am. Chem. Soc. 1987, 109, 5881–5883. Hoshino, K.; Saji, T.; Suga, K.; Fujihira, M. J. Chem. Soc., Faraday Trans. I 1988, 84, 2667–2676. Rosslee, C. A.; Abbott, N. L. Anal. Chem. 2001, 73, 4808–4814. Saji, T.; Hoshino, K.; Aoyagui, S. J. Am. Chem. Soc. 1985, 107, 6865–6868. Saji, T.; Hoshino, K.; Ishii, Y.; Goto, M. J. Am. Chem. Soc. 1991, 113, 450– 456. Tsuchiya, K.; Sakai, H.; Ohkubo, T.; Abe, M. Abstr. Pap.sAm. Chem. Soc. 2004, 228, U468-U468. Tsuchiya, K.; Sakai, H.; Saji, T.; Abe, M. Langmuir 2003, 19, 9343–9350. Liu, X.; Graham, M. D.; Abbott, N. L. Langmuir 2007, 23, 9578–9585.

Figure 1. (A) Molecular structures of 11-ferrocenylundecyltrimethylammonium bromide (FTMA) in reduced (left) and oxidized (right) states. (B) Expanded view of a microfluidic channel with two independent electrodes lining the vertical walls of the channel. (C) Schematic illustration of the microfluidic channel used to electrochemically generate and measure lateral gradients in the concentration of redox-active surfactant generated across microfluidic channels.

The study reported in this paper sought to investigate the formation of lateral concentration gradients of oxidized and reduced FTMA within aqueous solutions by using microfluidic channels with walls that were lined with electrodes (Figure 1B). The broad motivation for using the microfluidic channels was two-fold. First, by exploiting laminar flow within the microfluidic channel in combination with a channel outlet geometry that split the flow into four independent streams that could be analyzed for concentrations of surfactant and micelles (Figure 1C), we sought to establish a methodology that would yield direct evidence for the existence of lateral concentration profiles of micelles and monomeric surfactants generated across the channels. Second, we sought to demonstrate the generation of lateral gradients in the concentration of surfactants across microfluidic channels because this geometry is a promising one for the realization of continuous microscale separations and analytical processes in which selective transport of analytes across the channels would be driven by the lateral gradients in concentration of the surfactants. We note that a wide range of solutes (analytes) are known to interact with surfactants through different physical mechanisms. For example, a number of past studies have established that small organic molecules can be solubilized by micelles23 whereas polymers and Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

773

DNA19,29-33 can interact with cationic surfactants through electrostatic and hydrophobic forces in a manner that is dependent upon the extent of self-association of the surfactants. This paper is organized into four parts. First, we describe the experimental considerations and methods that underlie our approach. Second, we demonstrate the generation of gradients of concentration of monomeric FTMA and micelles of FTMA across microfluidic channels, and we compare the experimental measurements with the model predictions.28 Third, we investigate the influence of flow rate, applied electrical potentials, and electrolyte concentration on the formation of the lateral concentration gradients of FTMA to identify the origins of differences observed between the model predictions and the experimental measurements. Finally, we illustrate the level of control over the formation of lateral concentration gradients of surfactant within microfluidic channels that is afforded by the experimental approach reported in this paper and discuss its potential utility in analytical chemistry. EXPERIMENTAL PROCEDURES Materials. The redox-active surfactant FTMA was purchased from Dojindo Corporation (Gaithersburg, MD) and was used without further purification. Lithium sulfate monohydrate (SigmaAldrich, St Louis, MO) was used as received. Polydimethylsiloxane (PDMS) prepolymer (Sylgard 184) was purchased from Dow Corning (Midland, MI). Methods. Sample Preparation. FTMA was dissolved in aqueous Li2SO4 (0.1 M or 0.5 M, adjusted to pH 2 with H2SO4). Water used to prepare these solutions was distilled and then passed through a Milli Q Plus (Millipore Corp.) purification train (R ) 18.2 MΩ, γ ) 72.0 mN/m at 25 °C). Prior to addition of the FTMA, all solutions were deoxygenated by bubbling argon for 30 min. Residual oxygen dissolved within the aqueous solution causes FTMA to slowly oxidize (over hours). All solutions used in the experiments described below were characterized for concentrations of reduced and oxidized FTMA at the end of each experiment (see below for methods). Fabrication of Microfluidic Channels. The microfluidic channels used in the experiments reported in this paper had depths of 72 µm, widths of 80 µm, and lengths of 42 mm. Microfluidic channels were designed with one inlet and four outlets (Figure 1B,C) using Adobe Illustrator software and printed on a chrome mask at 32000 dpi (Fineline Imaging, Colorado Springs, CO). Silicon wafers (4 in.) were spin-coated with positive photoresist (Shipley 1827) and then soft baked at 90 °C for 5 min. The coated wafer was selectively exposed to UV light through the mask using MA6/BA6 contact aligner (Karl Suss). After the exposed photoresist was removed by a developing step, the patterned wafer was hard baked at 120 °C for 30 min, and then was etched using a plasma etcher (STS Multiplex ICP).34 (29) Abbott, N. L.; Jewell, C. M.; Hays, M. E.; Kondo, Y.; Lynn, D. M. J. Am. Chem. Soc. 2005, 127, 11576–11577. (30) Hays, M. E.; Jewell, C. M.; Lynn, D. M.; Abbott, N. L. Langmuir 2007, 23, 5609–5614. (31) Jewell, C. M.; Hays, M. E.; Kondo, Y.; Abbott, N. L.; Lynn, D. M. J. Controlled Release 2006, 112, 129–138. (32) Hays, M. E.; Jewell, C. M.; Kondo, Y.; Lynn, D. M.; Abbott, N. L. Biophys. J. 2007, 93, 4414–4424. (33) Pizzey, C. L.; Jewell, C. M.; Hays, M. E.; Lynn, D. M.; Abbott, N. L.; Kondo, Y.; Golan, S.; Talmon, Y. J. Phys. Chem. B 2008, 112, 5849–5857.

774


A PDMS replica was prepared by pouring a 10:1 (by weight) mixture of PDMS prepolymer and curing agent over the patterned wafer. After the PDMS was cured overnight at 60 °C, the PDMS replica was peeled from the master, and then films of titanium and gold with thicknesses of ∼100 and ∼1000 Å, respectively, were deposited sequentially onto each side wall of the microfluidic channel by using an electron beam evaporator (CHA Industries). Deposition of the metals onto the side walls of the microfluidic channels was achieved by shadow deposition with an angle of incidence of the metal of 45° (measured with respect to the surface normal of the floor of the channel). The layer of titanium was used to promote adhesion between PDMS and the gold working electrodes. The pressure within the electron beam evaporator was less than 1 × 10-6 Torr during each deposition. Microfluidic System. Aqueous solutions of FTMA were fed to the microfluidic system from an upstream reservoir via a plastic tube (inner diameter 0.51 mm, Thermoplastic Processes, Stirling, NJ). A reference electrode (Ag/AgCl) was placed into this reservoir. A salt bridge containing 0.1 M Li2SO4 (glass U-tube) connected the upstream reservoir of aqueous FTMA to a second reservoir containing a counter electrode (platinum mesh). Downstream of the microfluidic channel lined with the working electrodes, the flow was split into four channels (Figure 1C). Each outlet channel was connected to a separate syringe in a 10 channel syringe pump (KDS230, KD Scientific Inc., Holliston, MA) to draw FTMA solutions through the microfluidic channel at a desired flow rate and collect and analyze the composition of the outlet streams. A bipotentiostat (Pine Instruments) was used to control the electrical potentials applied to each of the two working electrodes, relative to the reference electrode potential. In a typical experiment, a potential was applied to one electrode within the microfluidic channel to drive oxidation of FTMA at one wall of the microfluidic channel, and a reducing potential was applied to the electrode located at the opposite wall of the channel. Measurement of Concentrations of Oxidized and Reduced FTMA. The concentrations of oxidized and reduced FTMA in aqueous solutions fed to the microfluidic channel and collected in the four outlets streams were determined by measurement of UV-visible absorption spectra. As demonstrated previously, the absorption spectrum of a mixture is closely approximated as the linear combination of the spectra of oxidized FTMA and reduced FTMA.23 The UV-visible absorption spectra were measured using a NanoDrop ND-1000 UV-vis spectrophotometer (Thermo Fisher Scientific, Wilmington, Delaware) at wavelengths between 220 and 750 nm. Measurements were performed using samples with volumes of 2 µL and an optical path length of 0.2 mm. Each sample was measured three times. We note that the above-described experimental procedure leads to the determination of the total concentration of surfactant in solution (both monomers and micelles). RESULTS AND DISCUSSION Measurement of Lateral Concentration Gradients of FTMA Across Microfluidic Channels. The first experiments performed in our study sought to demonstrate the generation of lateral (34) Ren, X. Q.; Bachman, M.; Sims, C.; Li, G. P.; Allbritton, N. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2001, 762, 117–125.

Figure 2. Lateral concentration profiles of reduced and oxidized FTMA generated electrochemically across a microfluidic channel by application of 0.3 V (vs Ag/AgCl) to the left electrode (anode) and 0 V (vs Ag/AgCl) to the right electrode (cathode). The electrodes lined the walls of the microfluidic channel. The inlet solution contained 0.9 mM reduced FTMA and 0.1 mM oxidized FTMA (in 0.1 M Li2SO4, pH ) 2). The flow rate was 0.04 mL/h. (0) Concentration of reduced FTMA at inlet; (]) concentration of oxidized FTMA at inlet; ([) measured concentrations of oxidized FTMA at outlet; (9) measured concentrations of reduced FTMA at outlet; (dotted line) calculated oxidized FTMA concentrations; (dashed line) calculated reduced FTMA concentrations.

gradients in the concentrations of reduced and oxidized FTMA within a microfluidic channel through direct measurement of the concentrations of these species in the outlet streams. In particular, we aimed to demonstrate concentration profiles of FTMA that would be expected to result in lateral concentration gradients of micelles of FTMA across the microfluidic channel. These initial experiments were performed by feeding an aqueous solution containing 0.9 mM reduced FTMA and 0.1 mM oxidized FTMA in 0.1 M Li2SO4 to the microfluidic channel at the flow rate of 0.04 mL/h and by applying an oxidizing potential (0.3V vs Ag/ AgCl) to the gold electrode on one side of the channel and a reducing potential (0V vs Ag/AgCl) to the electrode on the other side of the channel. The choice of potentials was guided by the results of our previous study using a thin film electrochemical cell.28 The results of this experiment are shown in Figure 2, where the concentrations of oxidized and reduced FTMA in the inlet and outlet streams are plotted as a function of lateral position across the microfluidic channel. In Figure 2, the anode is located at the lateral position defined as x ) 0 µm and the cathode is located at x ) 80 µm (recall that the width of the channel is 80 µm). The four outlet streams that fractionate the flow through the microfluidic channel collect fluid that passes across cross-sectional areas defined by 0 µm < x e 20 µm; 20 µm < x e 40 µm; 40 µm < x e 60 µm; and 60 µm < x < 80 µm. The concentrations of oxidized and reduced FTMA within these streams is plotted in Figure 2 at the midpoint of these cross-sections (i.e., x ) 10 µm for flow that passes between x ) 0 µm and x ) 20 µm). For the purposes of comparison, we also plot in Figure 2 the inlet concentrations of oxidized and reduced FTMA. These inlet concentrations were

uniform across the channel. Below, we make three observations regarding the data shown in Figure 2. First, inspection of Figure 2 reveals that lateral gradients in the concentration of oxidized and reduced FTMA were generated within the microfluidic channel during passage of the aqueous solution of FTMA through the channel. The concentration of reduced FTMA at the outlet of the microfluidic channel rises from 0.05 mM at x ) 10 µm to 0.96 mM at x ) 70 µm whereas the concentration of oxidized FTMA decreases from 0.75 mM at x ) 10 µm to 0.4 mM at x ) 70 µm. The qualitative nature of these lateral concentration profiles is consistent with oxidation of FTMA at the anode located at x ) 0 µm and reduction of the oxidized FTMA at the cathode located at x ) 80 µm. Second, as noted above, past studies have demonstrated that the critical micelle concentration of reduced FTMA is 0.1 mM in aqueous 0.1 M Li2SO4 (pH 2), whereas oxidized FTMA does not self-associate under the conditions of our experiments. Because the concentration of reduced FTMA decreases below 0.1 mM at x ) 10 µm, we conclude that there exists a region of solution near the left side of the microfluidic channel (proximal to the anode) that is depleted of micelles of FTMA during passage of the solution through the microfluidic channel. Furthermore, the results in Figure 2 suggest that to the right of this region depleted of micelles, there exists a lateral concentration gradient of micelles across the microfluidic channel. This interpretation of the lateral concentration profile of reduced FTMA in terms of the existence of a lateral gradient in micelle concentration across the channel is supported by the observation of a non-linear concentration profile for the FTMA (the concentration of FTMA measured at x ) 10 µm is higher than that estimated by linear extrapolation of the concentrations of FTMA measured at x ) 30 µm, 50 µm and 70 µm to x ) 10 µm). As described in more detail below by reference to a simple model for mass transport of surfactant across the channel, because monomeric and micellar FTMA will diffuse at different rates, at steady state, the presence of a local concentration of micelles will lead to necessarily higher gradients in concentration of FTMA. Third, although we also observe a lateral gradient in concentration of oxidized FTMA to be generated across the channel, the slope of the gradient for oxidized FTMA is substantially less than for reduced FTMA. This lower gradient in concentration of oxidized FTMA as compared to reduced FTMA is consistent with our expectation that oxidized FTMA does not self-assemble to form micelles and thus can diffuse more rapidly than reduced FTMA at concentrations above the CMC of reduced FTMA. We also note that the concentration profile of oxidized FTMA is not linear across the channel. We return to this point below. The results above establish that lateral gradients in the concentrations of reduced and oxidized FTMA are generated in the microfluidic system by application of potentials of 0.3 V (vs Ag/AgCl) and 0 V (vs Ag/AgCl) to the anode and cathode, respectively. We next sought to determine if the concentration profiles were quantitatively consistent with a previously described model that successfully described the current passed at the electrodes of a thin film electrochemical cell containing FTMA.28 Details of the model can be found in our previous publication.28 In brief, the model is a one-dimensional, steady-state model that Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

775

assumes fast electrode kinetics (the composition of oxidized and reduced FTMA at the electrodes is assumed to be described by the Nernst equation), local micelle-monomer equilibrium within bulk solution (the dynamics of micellization for low molecular weight surfactants such as FTMA are typically fast [of order 10 ms] compared to the diffusion times in these microscale systems [or order 1-10s]), and transport by diffusion only (no migration). The latter assumption was based on the use of swamping concentrations (>0.1 M Li2SO4) of electrolyte relative to the concentrations of FTMA. This model, although simple, was accurate in describing the steady state currents passed in the thin film electrochemical cell. In Figure 2, we plot the predictions of the model (dotted lines) using parameters corresponding to the experimental conditions used to perform the measurements in the microfluidic channel. Inspection of Figure 2 reveals that some predictions of the model are in qualitative agreement with the experimental measurements of the concentration profiles of oxidized and reduced FTMA. For example, the model predicts the existence of a region of solution near the anode (x < 16 µm) that is free of micelles, a result that is consistent with our interpretation of the experimental measurements (see discussion above). However, the level of quantitative agreement between the measured gradients and model predictions is not good. For example, the concentrations of oxidized FTMA measured across the channel are substantially higher than the values predicted by the model. Furthermore, the model predictions for the lateral concentration gradient of reduced FTMA are greater than those measured in the experiments. To understand the origins of the quantitative differences between the model predictions and the experimental measurements of lateral concentration profiles of oxidized and reduced FTMA in the microfluidic channel, we investigated the influence of several key experimental parameters on the lateral concentration profiles. For example, whereas our past measurements within a thin film electrochemical cell were performed in stagnant solutions, we sought to determine if factors such as the rate of flow of the FTMA solution through the microfluidic channel (which influences the residence time of the fluid, as well as other factors, such as dispersion) were impacting the lateral concentration profiles. We also sought to determine if factors such as the electrolyte concentration might be affecting the generation of concentration profiles (through Ohmic resistance) or if the overpotentials applied to the electrodes were sufficient to drive the electrode reactions at rates that lead to steady state concentration profiles within the residence time of the solution within the microfluidic system. In addition to providing insight into the differences between the experimental measurements and the model, these experiments help define the experimental parameters that influence the generation of lateral gradients in the concentration of FTMA in the microfluidic system. In the microfluidic channel, if the residence time of the fluid elements between the electrodes is short compared to the time for diffusion of species between the two electrodes, a steady state concentration profile of FTMA across the channel will not be achieved at the channel exit. The ratio of the characteristic time for lateral diffusion across the channel to the residence time of the fluid in the channel is a so-called Peclet number, defined as 776


Pe )

W2V LD

where W is the width of the channel (80 µm), V is the average velocity of fluid passing through the channel (1.93 × 10-3 m/s for a flow rate of 0.04 mL/h), L is the length of the microfluidic channel (42 mm), and D is mass diffusivity (3.2 × 10-10 m2/s for both reduced and oxidized FTMA monomers). For the flow rate of 0.04 mL/h, we calculated the Peclet number to be 0.92, indicating that the time for lateral diffusion is comparable to the time for convection along the channel in the experiments shown in Figure 2. This result leaves open the possibility that the residence time in the experiment reported in Figure 2 was insufficient to generate a steady state concentration profile. To determine if, in fact, the residence time in the experiment reported in Figure 2 was sufficient to generate a steady-state concentration profile of FTMA across the microfluidic channel, we performed a second experiment. In this experiment, we applied the same potential (0 V vs Ag/AgCl) to both electrodes that lined the microfluidic channel. Inspection of Figure 3A reveals that the concentration of reduced and oxidized FTMA measured at the outlet of the microfluidic channel is, in fact, uniform across the microfluidic channel, thus indicating that sufficient time exists during the residence time of the fluid elements in the channel to reach a steady state concentration profile of FTMA across the channel. Although a near-uniform concentration profile of reduced and oxidized FTMA was measured across the microfluidic channel shown in Figure 3A, a surprising result was the observation that the composition of the outlet stream comprised 0.6 mM reduced FTMA and 0.4 mM oxidized FTMA. At equilibrium, the composition of an aqueous solution of FTMA in contact with an electrode with a potential of 0 V (vs Ag/AgCl) comprises almost entirely reduced FTMA. Furthermore, as shown in Figure 3B, application of potentials of either 0.3 V (vs Ag/AgCl), 0.4 V (vs Ag/AgCl), or 0.5 V (vs Ag/AgCl) at both electrodes lining the microfluidic channels did not lead to the generation of solutions exiting the microfluidic channels that comprised almost entirely oxidized FTMA. Similar to the results shown in Figure 3A, we measured almost uniform concentration profiles across the microfluidic channel in each of these experiments (see Supporting Information, Figure S1). We note that an increase in flow rate to a Peclet number substantially greater than 1 (see Figure 3C for results corresponding to Peclet numbers of 1.8 and 3.7) did impact the extent of conversion of FTMA, consistent with the anticipated effects of relative rates of lateral diffusion and linear convection through the microfluidic system. Figure S2 of the Supporting Information shows that non-uniform concentration profiles were obtained for the case of a Peclet number of 3.7, although the shape of the non-uniform concentration profile is not yet understood (it may reflect unequal electrode areas on each side of the channel). Because the results described above suggested that the reason for incomplete oxidation of the FTMA during passage through the microfluidic channel at the low flow rates is not related to insufficient time for oxidation of the surfactant, we investigated the possible impact of the Ohmic resistance of the electrolyte solution on the electrochemical rate processes occurring at the electrodes. We prepared 1 mM reduced FTMA solutions in either 0.1 M Li2SO4 or 0.5 M Li2SO4 (pH ) 2) and flowed the solutions

Figure 4. Effect of electrolyte concentration on the lateral concentration profile of reduced and oxidized FTMA, when applying an oxidizing potential (0.3 V, 0.3 V) (vs Ag/AgCl) at both working electrodes. The electrodes line the walls of the microfluidic channel. The inlet solution contained 0.85 mM reduced FTMA and 0.1 mM oxidized FTMA, and the concentration of electrolyte was: (A) 0.1 M Li2SO4 or (B) 0.5 M Li2SO4. The flow rate was 0.04 mL/h.

through the microfluidic channel with 0.3 V (vs Ag/AgCl) applied to both electrodes lining the channel. A comparison of the results shown in Figure 4A with Figure 4B reveals that the increase in electrolyte concentration from 0.1 to 0.5 M increased the extent of oxidation of FTMA from 40% to 86%. To provide further insight into this observation, we estimated the conductivity of aqueous 0.1 M Li2SO4 as κ)F

∑ |z |u C ) 0.0237 Ω ·1cm i

i i

i

Figure 3. (A) Lateral concentration profiles of reduced and oxidized FTMA generated electrochemically across a microfluidic channel by application of reducing potential (0 V, 0 V) (vs Ag/ AgCl) at both working electrodes. The electrodes line the walls of the microfluidic channel. The inlet solution contained 0.05 mM reduced FTMA and 0.95 mM oxidized FTMA (in 0.1 M Li2SO4, pH ) 2). The flow rate was 0.04 mL/h. (B) Average outlet concentrations of reduced and oxidized FTMA as a function of potentials applied to both electrodes lining the microfluidic channel. The inlet solution contained 0.9 mM reduced FTMA and 0.1 mM oxidized FTMA (in 0.1 M Li2SO4, pH ) 2). Flow rate was 0.04 mL/h. (C) Average outlet concentration of reduced and oxidized FTMA as a function of flow rates when applying oxidizing potential (0.5 V, 0.5 V) (vs Ag/AgCl) at both working electrodes. The inlet solution contained 0.9 mM reduced FTMA and 0.1 mM oxidized FTMA (in 0.1 M Li2SO4, pH ) 2).

where κ is the conductivity, F is the Faraday constant, zi is the charge of ion i, Ci is the ion concentration, ui is the ion mobility (8.27 × 10-4 cm2/(s · V) for SO42- and 4.01 × 10-4 cm2/(s · V) for Li+ 35). For a path length of approximately 20 cm (distance between working electrode and counter/reference electrodes) and a conduit diameter of 0.508 mm, the resistance of the solution was calculated to be

R)

l ) 4.16 × 105 Ω κA

where A is the cross-sectional area perpendicular to the field vector. For a current of 0.1 µA (measured when applying 0.3 V vs Analytical Chemistry, Vol. 81, No. 2, January 15, 2009

777

Ag/AgCl) at each electrode, we calculate the potential drop between the working electrode and reference/counter electrodes to be ∆E ) IR ≈ (0.1 + 0.1) µA × 4.16 × 105 Ω ≈ 0.1V. This estimate suggests that the Ohmic resistance of the 0.1 M Li2SO4 solution, and thus the associated decrease in the available electrochemical driving force for the electrode rate processes, is likely the cause of the incomplete conversion of FTMA during passage through the microfluidic channel. For aqueous solutions containing 0.5 M Li2SO4, we calculate the potential drop to be only 0.02 V. The above results, when combined, suggest that Ohmic resistance of aqueous 0.1 M Li2SO4 likely influences the lateral concentration profiles of FTMA shown in Figure 2. To test this proposition, we repeated the experiments in 0.5 M Li2SO4 (Figure 5). Inspection of Figure 5, and comparison of those results to Figure 2, reveals that the increase in electrolyte concentration to 0.5 M does lead to closer agreement between the experimental measurements and the model-predicted lateral concentration profiles of reduced and oxidized FTMA within the microfluidic system (the closer agreement between experiment and model is pronounced for the oxidized FTMA). To provide additional support for the role of Ohmic resistance in our measurements, we performed cyclic voltammetry in 1 mM reduced FTMA aqueous solutions containing different concentrations of electrolyte. In these experiments, the reference electrode was located 200 mm from the working electrode so as to mimic the distance separating the electrodes in our microfluidic system. These measurements (see Supporting Information, Figure S3) clearly revealed the impact of Ohmic resistance on the positions of the redox waves measured during the cyclic voltammetry. Our conclusions regarding Ohmic resistance are also consistent with a prior publication36 in which 0.5 M sulfuric acid was used as the electrolyte, and the distance between the external reference

Figure 5. Lateral concentration profiles of reduced and oxidized FTMA generated electrochemically across a microfluidic channel by application of 0.3 V (vs Ag/AgCl) to the left electrode (anode) and 0 V (vs Ag/AgCl) to the right electrode (cathode). The electrodes line the walls of the microfluidic channel. The inlet solution contained 0.8 mM reduced FTMA and 0.2 mM oxidized FTMA (in 0.5 M Li2SO4, pH ) 2). The flow rate was 0.04 mL/h. (0) Concentration of reduced FTMA at inlet; (]) concentration of oxidized FTMA at inlet; ([) measured concentration of oxidized FTMA at outlets; (9) measured concentration of reduced FTMA at outlets; (dotted line) oxidized FTMA concentrations calculated at outlets; (dashed line) reduced FTMA concentrations calculated at outlets. 778


electrode and the working electrodes was several centimeters. In future studies, the potential drop could be minimized by fabricating a reference electrode and counter electrode at the entrance of the microfluidic channel,37 thus minimizing the distance over which ion transport occurs. Manipulation of Lateral Gradients in Surfactant and Micellar Concentration. The experimental results described above establish that it is possible to generate lateral gradients in the concentration of oxidized and reduced FTMA across microfluidic channels, and the experiments also define important factors that impact the realization of those concentration profiles. In the section below, we describe experiments that were performed to demonstrate the versatility and generality of the methodology reported in this paper as a means to realize and manipulate surfactant concentration profiles that may find use in a number of analytical and other technological contexts. A particular merit of the approach that we report in this paper for the generation of lateral gradients in the concentration of surfactants is that the gradients can be easily manipulated by changing the potential applied to the electrodes. This point is illustrated in Figure 6, in which the change in the lateral concentration profile of oxidized and reduced FTMA is measured before and after switching the potential from 0.3 V (vs Ag/AgCl) to 0.0 V (vs Ag/AgCl) at the left electrode (at x ) 0 µm), and simultaneously switching the potentials from 0.0 V (vs Ag/AgCl) to 0.3 V (vs Ag/AgCl) at the right electrode (x ) 80 µm). We note that following the change in potential of the electrodes, the new concentration profile of surfactant across the channel forms within ∼10 s, which is set by the characteristic time for the surfactants to diffuse across the microfluidic channel (L2/2D). Although the concentration profiles shown in Figure 6A and 6B are almost symmetric, we note that the highest concentration of reduced FTMA in Figure 6A is 1.05 mM whereas the highest concentration of reduced FTMA in Figure 6B is only 0.9 mM. These small quantitative differences in the magnitudes of the concentration profiles suggests that the working electrodes evaporated onto the two sides of the walls of the microfluidic channel may not be identical (the areas or thicknesses may differ on the two side walls). We also comment that whereas the concentration profiles shown in Figure 6 are obtained by application of constant potentials to the electrodes, a topic of future study will revolve around the manipulation of the concentration profiles using time-dependent potentials with periods that are comparable to the diffusion time. We predict that such an approach will enable the generation of highly non-linear lateral concentration profiles of surfactants. A potentially useful attribute of the methodology described in this paper is that it permits the generation of spatially localized (in direction perpendicular to channel flow) populations of micelles. As noted in our discussion regarding Figure 2, the CMC of reduced FTMA is 0.1 mM, and thus regions of the microfluidic channel that contain concentrations of reduced FTMA that are less than 0.1 mM do not contain micelles. Although it is possible that the presence of oxidized FTMA (35) Bard, A. J., Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications; Wiley: New York, 2000. (36) Choban, E. R.; Waszczuk, P.; Kenis, P. J. A. Electrochem. Solid-State Lett. 2005, 8, A348-A352.

Figure 6. Lateral concentration profiles of reduced and oxidized FTMA generated electrochemically across a microfluidic channel by application of (A) 0.3 V (vs Ag/AgCl) to the left electrode (anode) and 0 V (vs Ag/AgCl) to the right electrode (cathode). The inlet solution contained 0.65 mM reduced FTMA and 0.25 mM oxidized FTMA (in 0.1 M Li2SO4, pH ) 2); (B) 0 V (vs Ag/AgCl) to the left electrode (cathode) and 0.3 V (vs Ag/AgCl) to the right electrode (anode). The inlet solution contained 0.7 mM reduced FTMA and 0.2 mM oxidized FTMA (in 0.1 M Li2SO4, pH ) 2). The electrodes line the walls of the microfluidic channel. The flow rate was 0.04 mL/h.

within the microfluidic channel may impact micellization of reduced FTMA (through the formation of micelles composed of mixtures of oxidized and reduced FTMA), a previous study using light scattering and surface tension measurements suggests that “mixed micellization” does not occur.28 These prior observations, therefore, allow us to interpret regions of the microfluidic channel that contain less than 0.1 mM reduced FTMA to be free of micelles. The results shown in Figure 7 demonstrate that the fraction of the cross-sectional area of the microfluidic channel that contains micelles of reduced FTMA can be manipulated substantially by variation of the potentials applied to the electrodes in the system (note that the concentration of surfactant in the micellar state was determined by subtracting the CMC from the total surfactant concentration). Similarly, changes in the inlet concentrations of the FTMA can lead to substantial manipulation of the regions of the microfluidic channel that contain micelles (Figure 8). Because a wide range of solutes interact with the micellar state of surfactants but not their monomeric states (such as small drug-like molecules and DNA),23,38-40 the results in Figure 7 and 8 suggest that it should be possible to use the lateral gradients

Figure 7. Lateral concentration profiles of reduced and oxidized FTMA generated electrochemically across a microfluidic channel as a function of potentials applied to the electrodes. (A) 0.5 V to the left electrode (anode) and 0.4 V to the right electrode (cathode); (B) 0.3 V to the left electrode (anode) and 0 V to the right (cathode); (C) 0.1 V to the left electrode (anode) and -0.05 V to the right electrode (cathode). The electrodes line the walls of the microfluidic channel. The inlet solution is 0.5 mM reduced FTMA and 0.2 mM oxidized FTMA (in 0.1 M Li2SO4, pH ) 2). The flow rate was 0.04 mL/h. The hashed regions indicate the micelle-containing volumes of the microfluidic channel (see right axis).

in micellar concentrations generated in the microfluidic channels to drive and manipulate the lateral partitioning of such solutes across the channel. The results above suggest that FTMA, when combined with electrochemical methods, provides a general methodology to generate precisely controlled gradients in the concentration of (37) Satoh, W.; Loughran, M.; Suzuki, H. J. Appl. Phys. 2004, 96, 835–841. (38) Okada, T. J. Chromatogr. A 1997, 780, 343–360. (39) Chan, K. C.; Muschik, G. M.; Issaq, H. J.; Siiteri, P. K. J. Chromatogr. A 1995, 690, 149–154.


779

as to separate proteins.42-44 The approach that we propose complements these prior investigations by focusing on the electrochemical generation of molecules capable of undergoing self-assembly. These gradients in surfactant concentration are readily manipulated, and if desired, can be maintained along a channel for an indefinite length. In future studies, we propose to investigate how such gradients can be exploited in several contexts, including (i) to drive transport of analytes across the microfluidic channels, and (ii) as dynamic templates to generate gradients in the concentrations of molecules that interact with surfactants. We end by noting also that a number of methods for the separation of analytes have been proposed previously in which an electric field is applied across a channel in a direction that is orthogonal to the flow. These methods include electric field flow fractionation,48 isoelectric-focusing (IEF) field flow fraction,49-53 and transverse IEF.42-44 In contrast to these past methods, our system is distinguished by the presence of high concentrations of electrolyte (0.1 to 0.5 M Li2SO4) which will minimize transport processes due to migration, and redox-active amphiphiles that can be reduced and oxidized reversibly at electrodes at small potentials (0-0.3 V), thus minimizing changes in pH, Joule heating, and associated convective disturbances. These characteristics lead us to speculate that the gradients in surfactant concentration described in this paper may enable separations and microanalytical processes that have operational characteristics that differ in useful ways from those described previously. Figure 8. Lateral concentration profiles of reduced and oxidized FTMA generated electrochemically across a microfluidic channel by application of 0.3 V (vs Ag/AgCl) to the left electrode (anode) and 0 V (vs Ag/AgCl) to the right electrode (cathode). The electrodes line the walls of the microfluidic channel. The inlet solution contained (A) 0.33 mM reduced FTMA and 0.15 mM oxidized FTMA (in 0.1 M Li2SO4, pH ) 2); (B) 0.18 mM reduced FTMA and 0.1 mM oxidized FTMA (in 0.1 M Li2SO4, pH ) 2). The flow rate was 0.04 mL/h. The hashed regions indicate the micelle-containing volumes of the microfluidic channel (see right axis).

surfactants across microfluidic channels. We note that chemical/physical gradients generated at surfaces or within bulk solutions offer the basis of a variety of approaches leading to the design and discovery of catalysts and drugs and to the development of new analytical methods and measurement tools.41 In the context of microfluidic systems, several methods for generation of gradients in solution composition have been previously reported.42-47 For example, chemical gradients within microfluidic channels can be generated by contacting two inlet streams under conditions that lead to laminar flow.46 Alternatively, pH gradients have been generated across microfluidic flows by electrolysis of water in a fluidic channel so (40) (41) (42) (43) (44) (45) (46) (47)

780

Khaledi, M. G. J. Chromatogr. A 1997, 780, 3–40. Genzer, J.; Bhat, R. R. Langmuir 2008, 24, 2294–2317. Cabrera, C. R.; Finlayson, B.; Yager, P. Anal. Chem. 2001, 73, 658–666. Macounova, K.; Cabrera, C. R.; Holl, M. R.; Yager, P. Anal. Chem. 2000, 72, 3745–3751. Macounova, K.; Cabrera, C. R.; Yager, P. Anal. Chem. 2001, 73, 1627– 1633. May, E. L.; Hillier, A. C. Anal. Chem. 2005, 77, 6487–6493. Xu, C.; Barnes, S. E.; Wu, T.; Fischer, D. A.; DeLongchamp, D. M.; Batteas, J. D.; Beers, K. L. Adv. Mater. 2006, 18, 1427. Dertinger, S. K. W.; Chiu, D. T.; Jeon, N. L.; Whitesides, G. M. Anal. Chem. 2001, 73, 1240–1246.


CONCLUSION The key result of the study reported in this paper is the demonstration that redox-active surfactants can be combined with electrochemical methods to generate gradients in the concentration of monomeric surfactants and micelles across microfluidic channels. A simple one-dimensional diffusion model describes the qualitative features of the lateral gradients in concentration generated within the channel, and several phenomena that prevent quantitative agreement between the experimental gradients and the model predictions were identified (including Ohmic resistance of the solution within the channel). Through manipulation of the potential applied to electrodes that line the walls of the microfludic channel, we demonstrate that the methodology reported in this paper provides a facile method to achieve spatially localized populations of micelles within the microfluidic channel. To our knowledge, the method reported in this paper is unique in its ability to generate steady state gradients in the concentration of surfactants across microchannels. Because a wide range of solutes (small organic molecules23 through to large biomolecules19,32) are known to interact with surfactants, and because these interactions are often dependent on the self-association of the surfactants, the gradients in concentration of reduced and oxidized FTMA realized in the study described herein offer the basis of approaches to generate concentration gradients of (48) (49) (50) (51) (52) (53)

Caldwell, Chmelik, Chmelik, Chmelik, Chmelik, Chmelik,

K. D.; Gao, Y. S. Anal. Chem. 1993, 65, 1764–1772. J. J. Chromatogr. 1991, 539, 111–121. J.; Janca, J. Chemicke Listy 1989, 83, 321–325. J.; Thormann, W. J. Chromatogr. 1992, 600, 297–304. J.; Thormann, W. J. Chromatogr. 1992, 600, 305–311. J.; Thormann, W. J. Chromatogr. 1993, 632, 229–234.

a range of other solutes. Such transport processes driven by spatial gradients in surfactant concentration may provide the basis of new approaches for microscale separations and analytical processes.

SUPPORTING INFORMATION AVAILABLE Concentration profiles of FTMA as a function of applied potentials and flow rates, and cyclic voltammograms. This material is available free of charge via the Internet at http://pubs.acs.org.

ACKNOWLEDGMENT Support from the Petroleum Research Fund of the American Chemical Society (44193-AC7) and NSF (CTS-0553760 and CBET 075921) is gratefully acknowledged.

Received for review September 11, 2008. Accepted November 13, 2008. AC801933V


781

Electrochemical Generation of Gradients in Surfactant Concentration

Recommend Documents