J. Phys. Chem. C 2007, 111, 11959-11964
11959
Control of Catalytically Generated Electroosmotic Fluid Flow through Surface Zeta Potential Engineering Shyamala Subramanian and Jeffrey M. Catchmark* Engineering Science and Mechanics, College of Engineering, Agricultural and Biological Engineering and School of Forest Resources, College of Agricultural Sciences, 109 Agricultural Engineering Building, The PennsylVania State UniVersity, UniVersity Park, PennsylVania 16802 ReceiVed: March 15, 2007; In Final Form: May 17, 2007
The catalytic decomposition of hydrogen peroxide at bimetallic junctions has been shown to produce a local electric field capable of driving both electroosmotic fluid flow and electrophoretic forces on charged particles. The direction of the electroosmotic fluid flow depends on the metal surface zeta potential. We demonstrate that the direction of fluid flow can be controlled by the use of self-assembled monolayers to engineer the surface charge. In particular, silver disks were patterned on a gold-coated silicon surface. In the presence of hydrogen peroxide, gold behaves as an anode generating protons, and silver behaves as a cathode consuming protons, thus forming a local electric field extending from the gold surface to the silver surface. This electric field distribution results in an outward (away from the silver disk) electrophoretic force on a negative tracer and an inward force on a positive tracer. By selectively functionalizing the gold surface with either negatively charged carboxylic-acid-functionalized monolayers or positively charged amine-functionalized monolayers, we demonstrate the ability to align the direction of the electroosmotic fluid flow with the direction of the electrophoretic force on a given charged particle. A method for photolithographically patterning the monolayers on the gold surface while protecting the silver surface with an inert polymer layer has also been developed. This ability to engineer the direction of local catalytically produced forces provides the possibility of creating numerous fluidic and sorting devices.
Introduction The creation of localized forces capable of directing the movement of nanoscale objects or fluids is fundamental to many applications including active integrated mechanical systems such as MEMS/NEMS, fluidic devices, and sensors. Because of the complexity of fabricating intricate mechanical structures on the nanoscale, methods for producing controlled forces chemically are actively being explored. It has recently been shown that spatially localized catalytic reactions can produce forces capable of propelling both nanoscale and microscale objects such as rods, gears, and spheres.1-3 These forces arise from ion gradients generated near the surface at the interface of the catalytic regions. For example, a surface containing adjoining regions of silver and gold as shown in Figure 1 will decompose hydrogen peroxide via the following reactions (eqs 1-3)
overall: anode: cathode:
2H2O2 f 2H2O + O2
(1)
H2O2 f O2 + 2H+ + 2e-
(2)
H2O2 + 2H+ + 2e- f 2H2O
(3)
Observations on model systems suggest that silver is the cathode in the reaction, reducing H2O2 to H2O and consuming protons (H+) in the process.4,5 At the gold anode, H2O2 is oxidized to generate protons, thus forming a continuous selfgenerating ion gradient. * To whom correspondence should be addressed. E-mail: jcatchmark@ engr.psu.edu.
Figure 1. Illustration of the catalytic decomposition of hydrogen peroxide on the gold and silver surfaces. Gold acts as the anode, producing protons, and silver acts as the cathode, consuming protons. The protons move from the gold end to the silver end, and this results in a proton-gradient-generated electric field in solution.
The creation of an ion gradient in close proximity to the surface has the effect of producing a nonuniform electric field driving electroosmotic fluid flow. The velocity of fluid movement under the influence of an electric field is given by the Helmholtz-Smoluchowski equation
Ueo ) -ζsE/η
(4)
where Ueo is the electroosmotic velocity, is the dielectric constant of the medium, ζs is the zeta potential of the surface, E is the electric field, and η is the viscosity. If a charged particle is located within the electric field resulting from the catalytically produced ion gradient, its motion will be influenced by an
10.1021/jp072095y CCC: $37.00 © 2007 American Chemical Society Published on Web 07/24/2007
11960 J. Phys. Chem. C, Vol. 111, No. 32, 2007
Subramanian and Catchmark
Figure 2. Illustration of the catalytically generated electrophoretic and electroosmotic forces acting on a functionalized polystyrene microsphere in the vicinity of the silver-gold interface. Because of the negative zeta potential of the gold surface, the electroosmotic fluid flow is always inward toward the silver catalyst. (a) Negatively charged carboxyl-functionalized microspheres experience an outward electrophoretic velocity that counteracts the inward electroosmotic fluid flow. The spheres come to rest and form a ring at a distance from the edge of the silver catalyst where the two forces on the sphere balance. (b) Positively charged amidine-functionalized microspheres experience an inward electrophoretic velocity. This combined with the inward electroosmotic velocity pushes the particles toward the catalyst, so that no ring formation occurs.
electrophoretic force. The electrophoretic movement of a particle in solution is governed by the similar equation
Uep ) -ζpE/η
(5)
where Uep is the electrophoretic velocity and ζp is the particle zeta potential. The catalytically generated electric field E shown in eqs 4 and 5 has been evaluated theoretically based on the catalytically produced ion gradients leading to electroosmosis and electrophoresis. The behavior of the particle was shown to be a function of the tracer zeta potential, the surface zeta potential, and the electric field created.6 The generation of these forces and their impact on particle and fluid flow has been examined using a number of model systems. Bimetallic nanorods consisting of a variety of materials have been studied. Their motion has been explained via electrochemical decomposition of hydrogen peroxide simultaneously on both ends of the rod in accordance with eqs 1-3. Improved directionality of the nanorod system has been achieved by inserting a magnetic nickel segment between the gold and the silver ends. An external magnet was used to align the rods to the magnetic field while the self-generated ion gradient propelled the rod.7 Other geometries such as circular gold gears with the catalyst patterned selectively on the spokes have also been examined to achieve rotational motion. The catalytic forces produced on the spokes rotated 100-µm-diameter gears at an angular velocity of 1 rotation/second in a dilute solution of hydrogen peroxide.2 The bimetallic nanorod system has also been inverted to explore the possibility of creating localized fluid flow.3,8 In particular, silver disks of various diameters have been patterned onto a gold-coated surface and the motions of tracer particles exhibiting different zeta potentials examined in solutions containing low concentrations of hydrogen peroxide. Various kinds of tracers were used, including silica spheres, negatively and positively charged polystyrene spheres, and gold nanorods. In a dilute 0.5% solution of hydrogen peroxide at pH 5.5, gold exhibits a negative zeta potential of approximately -20 mV. Negatively charged carboxyl-terminated spheres form a ring at some distance from the silver catalyst. One hypothesis is that the ring of accumulated spheres is formed as a result of the balancing of the electroosmotic and electrophoretic forces on the spheres. Electroosmotic fluid flow toward the catalyst creates a fluidic surface current that pulls the spheres inward from a distance beyond the point of ring formation. This fluid flow acts on the tracer particle independently of its zeta potential. Once the fluid flow brings the tracer particle close to the silver
catalyst, in the vicinity of the ion charge gradient, the electrophoretic force begins to act on the particle as well. The direction of the electrophoretic force is dependent on the zeta potential of the tracer (see eq 5). In the case of negatively charged carboxyl-terminated spheres, the electrophoretic force is directed outward in the opposite direction of the electroosmotic fluid flow. At the point where these forces balance, the tracer particle velocity reaches zero. The accumulation of tracer particles at that point forms a defined ring. This process is shown schematically in Figure 2a. In the case of amidine-terminated positively charged spheres, a ring is not formed around the silver catalyst. Positively charged tracers move inward toward the catalyst first following the electroosmotic fluid flow. Once the tracers reach the position where the ion gradient is present, the electrophoretic force begins to act on the particle as well. Because the particle is positively charged, the direction of the electrophoretic force aligns with the direction of the fluid flow, and the tracers continue to flow onto the silver catalyst and accumulate on its surface. This is shown schematically in Figure 2b. A key to engineering devices such as fluidic pumps, particle sorting devices, sensors, and roaming vehicles lies in the effective spatial and temporal control of the magnitude and direction of the electroosmotic and electrophoretic forces resulting from the catalytically produced ion gradient. To date, control over the resultant direction of the motion has been achieved only via novel device design.2,7 In this work, we demonstrate the ability to control the direction of the electroosmotic fluid flow by modifying the surface zeta potential, as described in eq 5. Systems studied to date have employed surfaces that exhibit negative zeta potentials. One method for controlling the surface zeta potential is to use self-assembled monolayers attached to the gold surface. Selfassembled monolayers have been extensively studied and used as a means of modifying and controlling surface properties such as wettability, etch resistance, charge distribution, lubrication, adhesion, and chemical functionality. Advantages of using selfassembled monolayers include the ability to modify surface properties by simply using different end groups, the ability to cover the entire surface of the substrate uniformly and reproducibly, and the stability of many types of monolayers in various solutions. The most widely studied self-assembled monolayer is organosulfur [HS(CH2)nX] on gold surfaces because of the affinity of the sulfur-gold bond, the speed of monolayer formation, the wide range of possible end groups, the high stability and packing density, and the near-crystalline nature of the films formed.9
Zeta-Potential-Controlled Electroosmotic Fluid Flow In this work, we implement organosulfur surface functionalizations to modify the zeta potential of the gold surface. Specifically, we assemble two kinds of monolayers containing carboxylic acid and amine end groups. These surface modifications vary the zeta potential of the gold surface from -60 to 50 mV while the silver catalyst surface properties are kept constant. Monolayer formation on the silver is prevented using a photolithographically patterned resist that protects the silver surface and remains inert during the surface functionalization of gold. We show that the photolithographic process does not influence the surface functionalizations via infrared (IR) analysis. The motion of 2-µm-diameter polystyrene spheres with both positively charged amidine functionalization exhibiting a zeta potential of 50 mV and negatively charged carboxyl functionalization exhibiting a zeta potential of -60 mV is examined and is consistent with the prediction of a zeta-potentialassociated change in the direction of electroosmotic fluid flow. Values of the zeta potential of the carboxylic- and amineterminated surface functionalizations were obtained from ref 10 where the zeta potential measurements were conducted using a electrophoretic light scattering spectrometer under varying pH conditions. The reported zeta potentials of carboxylic-acidterminated self-assembled monolayers were approximately -60 mV, and the zeta potentials of amine-terminated self-assembled monolayers were approximately +50 mV, both measured at pH 5.5.10 This was the measured pH of the hydrogen peroxide solution used to assess the behavior of spheres under the influence of surface-functionalized gold surfaces. Experimental Section Test structures consisting of 50-µm-diameter silver disks on gold-coated silicon surfaces were fabricated using standard ultraviolet (UV) photolithography techniques. First, a silicon wafer was coated with 80 Å of chromium and 500 Å of gold using vapor-phase deposition. The chromium layer was used for better adhesion of the gold to the silicon substrate. The silver catalysts were patterned on the gold layer using a bilayer liftoff process. A nonphotosensitive sacrificial layer of LOR 5A (MicroChem Corp.), of 4000-Å thickness, was first spin-coated at 4000 rpm for 40 s on the gold-coated silicon wafer. The wafer was then baked at 190 °C for 10 min. Then, photoresist SPR 1813 (MicroChem Corp.) was spin-coated at 4000 rpm for 40 s on the sacrificial layer and baked at 110 °C for 90 s. The photoresist was exposed using a Karl Suss MA6 contact aligner with an unfiltered mercury lamp UV source exhibiting an output power of 12 mW/cm2 for 6 s. The film was then developed in CD26 (0.1 N solution of tetramethyl ammonium hydroxide in water) for 1 min 30 s. CD26 dissolved the exposed portions of the photoresist and the underlying LOR to form an undercut profile as shown in Figure 3a. The silver metal was deposited in the gaps and on the bilayer resist structure. The excess metal was removed by dissolving the photoresist in acetone and the sacrificial layer in CD26, resulting in patterned metal features. Sonication of the sample in CD26 facilitated the easy and fast removal of the excess metal and sacrificial LOR layer. Thiol-terminated molecules spontaneously form self-assembled monolayers on gold and silver surfaces in a similar manner.9 We have found experimentally that the formation of self-assembled monolayers on silver inhibits the catalytic reaction, i.e., no bubble formation associated with oxygen evolution is observed in the presence of hydrogen peroxide solution. The lack of observed catalytic activity could result from the structural differences in the self-assembled monolayers on the gold and silver surfaces. It has been reported that thiol
J. Phys. Chem. C, Vol. 111, No. 32, 2007 11961
Figure 3. Photolithography-based fabrication method for patterning and protecting silver disks on a gold-coated silicon surface. Step a: The silver catalyst is patterned on a gold surface via a lift-off method using sacrificial resist LOR 5A. Step b: LOR 2A is spin-coated and patterned preferentially on top of the silver. Step c: The top layer photoresist is dissolved in acetone. The gold surface is modified using self-assembled monolayers. Step d: The LOR 2A layer is stripped.
molecules on silver are more perpendicular to the surface with lower intermolecular distances than self-assembled monolayers formed on gold. Moreover, it has also been observed that the ionic nature of the Ag-S bond leads to the transformation of the Ag lattice and to the formation of a layer of Ag2S on the surface upon long exposure to thiol solution.11 The denser nature of the surface functionalization and the change in silver properties could explain why catalytic activity is not seen when self-assembled monolayers are allowed to grow both on silver and on gold. To prevent the formation of a monolayer on silver, we have developed a method to preferentially assemble thiol molecules having carboxylic acid or amine end groups on the gold surface by protecting the silver surface. This was accomplished by patterning the silver surface with a thin layer of the nonphotosensitive polymer LOR 2A (lift-off resist). LOR 2A belongs to the same family of nonphotosensitive resists as used in the bilayer lift-off process. LOR was patterned because previous studies have shown that LOR is not dissolved in ethanol solutions used for functionalizing gold surfaces.12 In addition, both CD26 and H2O2 do not remove the functionalization from the surface of gold, nor do they impact the end group functionality. Figure 4 presents FTIR spectra of three gold samples with solution-deposited 16-mercaptohexadecanoic acid self-assembled monolayers. The first sample was untreated. The second sample was placed in a solution of 0.1 N TMAH (CD26) for 10 min, and the third sample was placed in a 1% hydrogen peroxide solution in water for 25 min. The d+ and d- signals identified as the symmetrical and asymmetrical stretching modes of C-H are at 2850 and 2920 cm-1, respectively. The d+ and d- modes indicate the presence of the self-assembled monolayer on all three substrates. Similar profile shapes and intensities indicate that the structure and uniformity of the surface functionalizations were unaffected by the solutions.13 The gold surface containing the silver dots was spin-coated at 4000 rpm for 40 s with a thin layer of LOR 2A. The wafer was baked at 150 °C for 5 min. The lower baking temperature and time facilitated the removal of LOR in the subsequent steps. Photoresist SPR 1813 was coated and exposed as discussed above for the bilayer process as shown in Figure 3b. The mask
11962 J. Phys. Chem. C, Vol. 111, No. 32, 2007
Subramanian and Catchmark the sample in a bath of CD26 for 10 min. The CD26 used did not alter the properties of the surface functionalization formed on the gold surface as described above. This gold surface modified using the self-assembled monolayer containing the silver dots was used as the test structure for observing the effect of the end group of the surface functionalization on the change in direction of fluid flow on the gold surface. Results and Discussion
Figure 4. FTIR spectra of three gold samples with a solution-deposited 16-mercaptohexadecanoic acid self-assembled monolayer. Sample 1 is untreated. Sample 2 has been placed in a solution of 0.1 N TMAH (CD26) for 10 min and Sample 3 has been placed in a 1% hydrogen peroxide solution in water for 25 min. The d+ and d- identified as the symmetrical and asymmetrical stretching modes of C-H are at 2850 and 2920 cm-1, respectively. The d+ and d- modes indicate the presence of self-assembled monolayers on all three substrates. Similar profile shapes and intensities indicate that the monolayer structure and uniformity were unaffected by the solutions.
used here was such that, after development, LOR and photoresist were left behind only on the silver dots, as seen in Figure 3c. The photoresist was stripped in acetone, and any excess LOR on the gold was removed in a reactive ion plasma consisting of 20 sccm of O2 and 7 sccm of Ar at 20 mTorr and 200 W. A thin layer of inert LOR 2A was left behind on the silver catalyst, which prevented the self-assembled monolayer from forming on the silver. Once the silver was protected, the sample was cleaned by rinsing in acetone, isopropyl alcohol, and ethanol. The sample was immersed in a 1 mM solution of thiol in ethanol for approximately 12 h, which caused a 2-nm-thick monolayer of the surface functionalization to be deposited on the gold. Throughout this process, the LOR did not react with the ethanol and thiol solution. The sample was removed, rinsed in ethanol and water, and dried in nitrogen. LOR was stripped by placing
Table 1 lists the different molecules used for surface functionalization of the gold and the tracers used in the experiments. The tracer solution was prepared by diluting 1-2 drops of the as-purchased tracers in 1 mL of deionized (DI) water. This solution was mixed with 1% hydrogen peroxide, the fuel for the catalytic reaction, in a 1:1 ratio. The final tracer solution contained the tracers in 0.5% hydrogen peroxide in DI water. A drop of this solution was placed on the sample surface and viewed with a bright-field microscope with red filter, which blocks any residual UV light that might decompose the hydrogen peroxide. Videos were recorded using a USB microscope and imaging software. We have studied the movement and behavior of two kinds of tracers, carboxyl and amidine-terminated latex spheres, on gold surface modified using two kinds of surface functionalizations, carboxylic-acid-terminated thiol molecule and amineterminated thiol molecule of comparable lengths. All observations were performed using a dilute solution of tracers in 0.5% hydrogen peroxide solution in DI water. The observations were taken 5-10 min after the tracer solution had been placed on the sample to allow bubble formation due to catalytic activity to subside. The catalytic dots used were 50 µm in diameter and had a spacing of 500 µm. A gold surface modified using mercaptohexadeconoic acid, also referred to as a carboxylic-acid-terminated self-assembled monolayer, has a measured surface zeta potential of -60 mV at pH 5.5. In the presence of the self-generated electric field, located primarily on the gold surface near the gold-silver interface, the ions in the double layer move inward toward the silver edge. The ions drag the surrounding water molecules and cause an inward fluid flow toward the catalyst edge. The tracers present in the fluid are also swept along with the fluid and migrate toward the catalyst, radially, from all directions. The dashed arrows in Figure 5 show the direction of electroosmotic fluid velocity. The optical microscope image shown in Figure 5a was taken 5 min after a 0.5% solution of hydrogen peroxide containing carboxyl-terminated tracers had been placed on the
TABLE 1: List of Zeta Potential Values, Measured at pH 5.5, of Self-Assembled Monolayers Used to Modify the Gold Surface and Zeta Potential Values of Tracers Employed To Study Fluid Flow surface functionalizations of gold surface and tracer particles HS(CH2)15CO2H, 16-mercaptohexadecanoic acid HS(CH2)11NH2, 11-amino-1-undecanethiol, hydrochloride carboxyl-stabilized 2-µm-diameter polystyrene spheres amidine-stabilized 2-µm-diameter polystyrene spheres
terminology
ζ potential at pH 5.5 (mV)
carboxylic-acidterminated selfassembled monolayers (SAMs) amine-terminated self-assembled monolayers (SAMs)
-60
carboxyl tracers
-60
amidine tracers
40
50
source
processing steps
Sigma Aldrich
1 mM solution in ethanol
Dojindo Molecular Technologies Inc. Interfacial Dynamics
1 mM solution in ethanol
Interfacial Dynamics
diluted with DI water
diluted with DI water
Zeta-Potential-Controlled Electroosmotic Fluid Flow
J. Phys. Chem. C, Vol. 111, No. 32, 2007 11963
Figure 5. Matrix showing the effects of tracer zeta potential (arranged in columns) and surface zeta potential (arranged in rows) on tracer behavior. The silver disk is seen as a circle in the optical images and as a black block in the illustrations. Top images a and b have a carboxylic-acidterminated self-assembled monolayer (negatively charged) on the gold; hence, electroosmosis is inward. The carboxyl tracers form a ring, and amidine rushes in toward the catalyst. The bottom images c and d have an amine-terminated self-assembled monolayer (positively charged) on the gold; hence, electroosmotic velocity is outward. The carboxyl spheres are pushed away, and amidine tracers form a ring.
sample. The electroosmotic fluid flow is experienced by the particles even as far as 500 µm away from the silver edge (refer to video 1 in the Supporting Information). As the tracers move closer to the silver, they also experience an electrophoretic velocity. Because the tracer particle is negatively charged, this velocity is directed outward, opposite to the fluid flow (illustrated by the dotted arrow in Figure 5a). At a certain distance from the edge, the two velocities cancel each other, and the tracers come to rest. The tracers begin to accumulate at this distance to form a ringlike structure, as seen in Figure 5a. The ring diameter was measured to be ∼114 µm. This is in general agreement with the ring formation on a bare gold surface, which also has a negative zeta potential.8 In the case of the tracer having a positive charge, as with the amidine-terminated spheres with the +40 mV zeta potential, both the electroosmotic fluid flow and electrophoretic velocity experienced by the particle are in the same direction, as illustrated by the dashed and dotted arrows in Figure 5b. In this case, all of the tracers rush in toward the silver and are observed to stick to the silver surface. However, some of the tracers that rush toward the silver feature are also swept upward with the fluid. The fluid is assumed to move inward and upward to maintain fluid continuity (refer to video 2 in the Supporting Information). The attachment of the tracers to the silver could be the result of weak electrostatic attraction between the negatively charged silver surface (because of the formation of a silver oxide) and the positive tracer. With time, the silver disk becomes covered with the tracers, as seen in Figure 5b, where the silver is black in color because of the
accumulation of the tracers. No ring formation is seen in this system, in agreement with studies done on a bare negative gold surface. Gold surfaces modified using 11-amino-1-undecanethiol, also referred to as an amine-terminated self-assembled monolayer, have a positive zeta potential of +50 mV at pH 5.5. In this case, the ions in the gold surface double layer migrate away from the silver edge in response to the self-generated electric field. This causes the fluid near the surface to flow away from the silver edge, as indicated by the dashed arrows in Figure 5c,d. Figure 5c illustrates the case of a carboxyl sphere placed on a gold surface with an amine-terminated self-assembled monolayer. Because the tracer is negative, the electrophoretic velocity is directed in the outward direction, aligned with the electroosmotic fluid velocity. The optical image shows two distinct regions surrounding the silver disk. Region 1, immediately adjacent to the disk, is depleted of tracers, and region 2 extends beyond the depletion region and has a uniform distribution of tracers. The first region is formed because both the forces acting on the tracers are aligned outward and push the tracers away from the catalyst, forming a region of no tracers up to a distance of ∼200 µm. This is almost twice the ring diameter of Figure 5a, as both velocities are aligned in the same direction and act simultaneously on the particle. Beyond this depletion region, the tracers are distributed evenly and are either stationary or continue to move slowly outward (refer to video 3 in the Supporting Information) because of the electroosmotic flow of fluid outward and fluid continuity requirements.
11964 J. Phys. Chem. C, Vol. 111, No. 32, 2007 When an amidine tracer solution is placed on a gold surface with an amine-terminated self-assembled monolayer, the positive charge results in an electrophoretic velocity inward toward the silver edge. The electroosmotic and electrophoretic flow are now in opposite directions. The tracer particles come to rest and accumulate to form a ring at a point where the two velocities cancel each other. This is similar to Figure 5a, although velocity directions have been reversed. Given the similar magnitudes of the surface charges of the carboxylic-acid- and amine-terminated self-assembled monolayers, it is not surprising that the rings formed in the two cases are at approximately the same distance from the silver edge. However, the main difference is the absence of inward tracer movement in the sample functionalized with the amine-terminated self-assembled monolayer, as is seen clearly in the sample functionalized with the carboxylic-acidterminated self-assembled monolayer. In the amine-functionalized sample, the ring forms immediately, and not much fluid and tracer sphere movement is observed with time. In fact, the tracers seem to settle to the surface and stick to it, perhaps because of the interaction of the surface functionalization and the tracer surfactant (Figure 5d). The difference in the behavior of the tracers observed when using carboxylic-acid- and amine-terminated self-assembled monolayers is mainly due to the reversal in the direction of the electroosmotic velocity brought about by the change in surface zeta potential of the gold surface. Tracers form a ring when the electroosmotic velocity is opposite to the electrophoretic velocity. When the two velocities are aligned in the same direction, the tracers do not form a ring but instead are pushed away from or toward the silver edge. The difference in behavior of the tracers in the presence of opposing electroosmotic and electrophoretic velocities is clearly seen by comparing Figure 5a,d with Figure 5b,c. Whereas the former two images exhibit a distinct ring consisting of a high density of tracers a fixed distance from the silver edge, such a ring of tracers is absent in the latter images. Also, the inward flow of fluid and tracers, present even at long distances from the silver edge, observed when using carboxylic-acid-terminated self-assembled monolayer (Figure 5a,b) is not observed for the amine-terminated self-assembled monolayer. These observations demonstrate that the flow of fluid due to catalytically driven electroosmosis can be reversed by changing the surface zeta potential in the gold regions. Further control of the magnitude and direction of the catalytically produced fluid flow can be achieved by surface functionalizations whose properties can be alerted or switched dynamically by external stimuli such as electrical potential.14 Also, the direction of electroosmotic fluid flow can be altered by changing the solution pH, as the zeta potential exhibited by the surface functionalizations are pH-dependent. In conclusion, control over the direction of catalytically produced electroosmotic fluid flow has been successfully achieved by tailoring the surface charge using self-assembled monolayers with different end groups. Self-assembled monolayers were successfully patterned preferentially on the gold surface by protecting the silver surface with an inert LOR polymer, which was patterned using conventional optical
Subramanian and Catchmark lithography. We demonstrate that these surface functionalizations are unaffected by the LOR layer and the associated processing, as well as by the H2O2 solution. The use of surface zeta potential engineering enables the alignment of the electroosmotic and electrophoretic velocities for a given charged particle. This, in combination with the ability to spatially pattern self-assembled monolayers on the microscale, might provide an avenue toward the creation of useful microfluidic and sorting devices for an array of biological applications. Acknowledgment. We gratefully acknowledge the intellectual contributions of our collaborators on this project, in particular, Ayusman Sen, Darrell Velegol, and Tom Mallouk. Also, we thank the staff of the Penn State Nanofabrication Facility, where this work was performed. This work was supported by the Penn State Center for Nanoscale Science; NSF Grant DMR-0213623; NSF-NIRT Grant CTS-0506967; National Science Foundation Cooperative Agreement 0335765; the National Nanotechnology Infrastructure Network, with Cornell University; and The Pennsylvania State University Materials Research Institute. Supporting Information Available: Videos showing (a) ring formation by carboxyl-terminated tracers on a gold surface having carboxylic-acid-terminated surface functionalization (video 1), (b) inward movement of amidine-terminated tracers on a gold surface having carboxylic-acid-terminated surface functionalization (video 2), and (c) formation of regions 1 and 2 by carboxyl-terminated tracers on a gold surface having amineterminated surface functionalization (video 3). This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Paxton, W. F.; Kistler, K. C.; Olmeda, C. C.; Sen, A.; St. Angelo, S. K.; Cao, Y.; Mallouk, T. E.; Lammert, P. E.; Crespi, V. H. J. Am. Chem. Soc. 2004, 126, 13424-13431. (2) Catchmark, J. M.; Subramanian, S.; Sen, A. Small 2005, 1, 202206. (3) Kline, T. R.; Velegol, D.; Catchmark, J. M.; Mallouk, T. E.; Sen, A. Catalytically induced autonomous fluid flow: Pumping without pumps. Presented at the 229th ACS National Meeting, San Diego, CA, Mar 14, 2005; Paper COLL 155. (4) Wang, Y.; Hernandez, R. M.; Bartlett, D. J., Jr.; Bingham, J. M.; Kline, T. R.; Sen, A.; Mallouk, T. E. Langmuir 2006, 22, 10451-10456. (5) Paxton, W. F.; Baker, P. T.; Kline, T. R.; Wang, Y.; Mallouk, T. E.; Sen, A. J. Am. Chem. Soc. 2006, 128, 14881-14888. (6) Kline, T. R.; Iwata, J.; Lammert, P. E.; Mallouk, T. E.; Sen, A.; Velegol, D. J. Phys. Chem. B 2006, 110, 24513-24521. (7) Kline, T. R.; Paxton, W. F.; Mallouk, T. E.; Sen, A. Angew. Chem., Int. Ed. 2005, 117, 754-756. (8) Kline, T. R.; Paxton, W. F.; Wang, Y.; Velegol, D.; Mallouk, T. E.; Sen, A. J. Am. Chem. Soc. 2005, 127, 17150-17151. (9) Ulman, A. Chem. ReV. 1996, 96, 1533-1554. (10) Shyue, J.-J.; De Guire, M. R.; Nakanishi, T.; Masuda, Y.; Koumoto, K.; Sukenik, C. N. Langmuir 2004, 20, 8693-8698. (11) Laibinis, P. E.; Whitesides, G. M.; Allara, D. L.; Tao, Y. T.; Parikh, A. N.; Nizzo, R. G. J. Am. Chem. Soc. 1991, 113, 7152-7167. (12) Subramanian, S.; McCarty, G. S.; Catchmark, J. M. J. Microlithogr., Microfabr., Microsyst. 2005, 4, 049701(1-4). (13) Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112, 558-569. (14) Lahann, J.; Mitragotri, S.; Tran, T.; Kaido, H.; Sundaran, J.; Hoffer, S.; Somorjai, G. A.; Langer, R. Science 2003, 299, 371-373.
