Active Spatiotemporal Control of Electrochemical Reactions by

The Journal of Physical Chemistry A 0 (proofing), ... Analytical Chemistry 2009 81 (1), 453-459 .... Electrochemically Derived Gradients of the Extrac...
0 downloads 0 Views 300KB Size
Download PDF
8970

J. Phys. Chem. B 2001, 105, 8970-8978

Active Spatiotemporal Control of Electrochemical Reactions by Coupling to In-Plane Potential Gradients† Karin M. Balss, Brian D. Coleman, Christopher H. Lansford, Richard T. Haasch, and Paul W. Bohn* Department of Chemistry and Materials Research Laboratory, UniVersity of Illinois at Urbana-Champaign, 600 South Mathews AVenue, Urbana, Illinois 61801 ReceiVed: March 5, 2001; In Final Form: June 16, 2001

Active spatiotemporal control of electrochemical reactions through dynamic electrochemical potential gradients was explored by investigating three different types of reactions on Au: alkanethiol SAM electrosorption, Cu deposition and stripping, and O2 evolution from H2O2 oxidation. Counterpropagating gradients composed of two different thiols differing either in terminal functionality or in chain length were prepared, and their kinetic and environmental stability was inferred from spatially resolved contact angle measurements for samples stored under varying environmental conditions for periods up to one month. Chain length was found to correlate strongly with stabilitysa requirement for stability being that at least one of the chains be at least C8 or longer. Spatially directed Cu deposition on Au was demonstrated by forming Cu stripes on Au, establishing that a sequence of different potential gradients could be used to define an area of deposition in the center of a working electrode. Dynamic spatiotemporal control of Cu deposition on Au was achieved by translating a potential window, which encompassed the Cu redox waves, across the Au surface. The position of the Cu/Au transition was constant at a potential intermediate between the two waves, and the width of the transition region in the SPR images was narrower than either of the two electron transfer waves. Spatially directed oxidation of H2O2 was demonstrated by monitoring the formation of oxygen bubbles near the electrode. Consistent with predictions of the Butler-Volmer equation, the rate of bubble formation was found to depend on spatial position (overpotential) in these experiments.

Introduction Numerous methods exist to prepare spatial patterns of chemical composition on planar surfaces, e.g., lithography, micromachining, and painting. Methods which can produce ultrathin film surface coatings with lateral variations on the nanometer-micrometer length scale, which can subsequently be used to control the interaction of the solid with its environment, are of particular relevance to advanced technologies. Methods that rely on lithography and micromachining1 can create surfaces with patterned isolated domains on the µm length scale, while diffusion-reaction methods2 create surfaces with compositions varying in a continuous manner. Recently, solution and surface gradients have been created by effecting controlled mixing of reagents inside a microfluidic cell.3 All of these methods have the limitation that the surface composition patterns, once created, are difficult to change. Recent work from this laboratory has demonstrated electrochemically generated gradients that are dynamically controllable;4 i.e., spatial composition patterns may be created and then altered at any time by changing the potential program applied to a thin metal film electrode, thereby allowing the chemical and physical properties of a surface to be manipulated in situ in both space and time. Self-assembled monolayers (SAMs) of alkanethiols on Au constitute a powerful chemical system in which to investigate the active control of surface properties because the resulting properties are dominated by the ω-substituent, i.e., the substituent farthest from the S headgroup. Thus, two-component mixed †

Part of the special issue “Royce W. Murray Festschrift”. * To whom correspondence should be addressed. E-mail: bohn@ scs.uiuc.edu.

monolayerssthose containing two types of molecules with differing ω-substitutionscan be tailored in composition to engineer surfaces capable of controlling a wide variety of phenomena at solid-liquid interfaces.5-10 Furthermore, using electrochemical potential gradients to vary the composition profile of two-component SAMs laterally4 opens the way to controlling interfacial properties by melding compositional tailoring with surface patterning. Surfaces modified by mixed SAMs that vary in composition across the surface can be used to affect fundamental processes such as wetting,11-13 adhesion,14 and phase separation,11,15 thereby enabling practical applications in molecular recognition, directed motion, and molecular separations. These monolayer composition patterns can vary continuously, e.g., linearly, as a function of lateral position across a surface, or discontinuously, as in isolated domains, depending on the fabrication method. Gradients in alkanethiol surface composition on thin (20 nm e d e 50 nm) Au electrodes have been fabricated by taking advantage of their electrosorption properties. Injecting milliamp currents yields significant in-plane voltage gradients so that, rather than assuming a single value of potential, an in-plane potential distribution, V(x), is imposed on the electrode surface, according to

V(x) ) V0 +

∫

iF(l) dl A

(1)

where i is the magnitude of the injected current, F(l) is the film resistivity, and A is the cross-sectional area. The extent and position of the gradient can be tuned by adjusting the magnitude of the injected current and the voltage offset, respectively.

10.1021/jp010819e CCC: $20.00 © 2001 American Chemical Society Published on Web 07/19/2001

Active Spatiotemporal Control

Figure 1. Schematic diagram illustrating the electrochemical gradient formation concept. Top panel illustrates a typical cyclic voltammogram for RSH with voltage (center) and spatial position (top) axes. Bottom panel gives a schematic diagram of the cell with surface composition gradient. Injecting the current in-plane produces an in-plane electrochemical potential variation which maps the electrochemical behavior onto the electrode surface spatially.

Furthermore, the in-plane electric potential gradient means that, relative to a solution reference couple, electrochemical reactions occur at defined spatial positions corresponding to the standard potential, V(x) ≈ E0. The spatial gradient in electrochemical potential can then produce spatially dependent electrochemistry. Surface-chemical potential gradients can be prepared by arranging the spread of potentials to span an electrochemical wave mediating redox-associated adsorption or desorption. For example, reductive desorption of alkanethiols into alkaline solutions occurs at E0 ≈ -0.8 V versus Ag/AgCl, varying only modestly with chain length and terminal group.16-18 Thus, by arranging the electrochemical potential drop to span the region of E0des for the alkanethiol of interest a gradient in alkanethiol surface coverage, Γ(RSH), can be created. Fortuitously, the reductive desorption process can be followed in situ, even though it is in the presence of a 103-fold larger background, because current injection into thin films produces in-plane voltages with inherently high signal-to-noise ratios, S/N > 104.19 Figure 1 illustrates this principle. Because the bipotentiostat is used to clamp the potential of the two ends of the working electrode, the applied potential between the contacts varies linearly with position. At potentials anodic of the oxidative adsorption wave, the Au surface is saturated with thiol, while at potentials cathodic of E0des, the surface is bare. Between these two extremes, a gradient in surface composition is formed. The maximum width of the potential window is limited by solvent electrolysis at negative potentials and Au oxide formation at positive potentials. This strategy creates a composition gradient in one component, the originally adsorbed alkanethiol. To form two-component gradients, the electrode is removed from electrolyte and immersed in a second thiol with a different ω-group. The surface that is created contains primarily one component at the ends and counterpropagating gradient concentrations in between. Changing the terminal group, therefore, allows two-component systems with spatially graded chemical and physical properties to be fabricated. The transition region between the two components is dependent on the value of E0des for the first thiol and occurs at unique spatial positions for that thiol. In this paper, we have examined the robustness of these

J. Phys. Chem. B, Vol. 105, No. 37, 2001 8971 two-component systems under a variety of environmental conditions as a function of time. Controlling the spatial distribution of electrochemical potentials constitutes a general strategy for directing electrochemical phenomena at surfaces and can be applied to reactions other than alkanethiol electrosorption. For example, bulk metal deposition and generation of redox products at defined spatial locations are explored here. The deposition of Cu onto Au electrodes has been studied extensively, especially underpotential deposition (UPD), because UPD layers exhibit superior structural and morphological properties to those formed at the bulk deposition potential.20-24 The redox couple explored here for spatially directed deposition is the overpotential deposition of Cu2+ from aqueous K2SO4. In a strategy similar to that used to pattern thiols on Au surfaces, potential programs can be mapped onto Au to include potentials cathodic of the reduction peak for deposition and anodic of the oxidative peak to strip copper. Then, by using oppositely directed electrochemical potential gradients in consecutive steps, it is possible to create a step profile, i.e., stripe, of Cu in the center of a Au electrode. The deposition process is more complicated than that for SAMs because it involves several phases of deposition and multilayer coverages result.25-27 To further demonstrate the versatility of this approach, the oxidation of H2O2 in aqueous K2SO4 was examined. Here the gradient in potential results in O2 formation at positions corresponding to potentials anodic of the oxidation potential. Furthermore, since the Butler-Volmer equation predicts that the rate of the formation reaction scales with the overpotential28 and the overpotential is a function of spatial position, the rate of oxidation should vary with spatial position. A distribution of O2 bubbles is observed on the electrode surface, the size being dependent on the potential according to the Butler-Volmer equation, providing direct visualization of the spatial control of electrochemical potentials. Experimental Section Materials. Absolute ethanol (EtOH) was purchased from Aaper Alcohol and Chemical Company. Optima grade methanol (MeOH), K2SO4, CuSO4, and KOH were purchased from Fisher Scientific. Octanethiol (OT), hexadecanethiol (HDT), 16mercaptohexadecanoic acid (HDA), 3-mercaptopropanoic acid (MPA), and propanethiol (PT) were purchased from Aldrich and used without further purification. Substrate Preparation. Microscope slides cut into 75 × 8 × 1 mm pieces and SF10 plates (30 × 30 mm, Harrick Scientific Corporation) and SF10 prisms (Coherent) were cleaned in a freshly prepared piranha (4:1H2SO4/H2O2) solution for 1 h. The samples were rinsed with copious amounts of deionized (18.2 MΩ cm) H2O, followed by 2-propanol and dried with N2. The samples were then transferred to the evaporation chamber immediately to minimize contamination. Chromium (1 nm) was evaporated to promote adhesion of Au to the glass substrate, followed by 50 nm of Au at base pressures ) 1 × 10-6 Torr. For surface plasmon resonance (SPR) experiments, 48 nm of Au was deposited on the SF10 plates or prisms. Samples were used immediately or stored under dry N2 prior to use. Gradient Formation. Thin Au samples were placed in 1 mM ethanolic solutions of the thiol of interest for 12 h to assemble the first component. A bipotentiostat (Pine Instruments model AFCBP1) employing a saturated Ag/AgCl reference electrode and a Pt counter electrode in deaerated 0.5 M KOH/MeOH was used to prepare the gradients employed in the stability studies. Contacts to the electrode surface were made by Au wires pressed

8972 J. Phys. Chem. B, Vol. 105, No. 37, 2001 onto the film in a PTFE assembly cell. Typically, alkanethiol/ Au SAMs were electrolyzed with 20 mV mm-1 in-plane potential drops for 1 min. Samples were then quickly removed from the solution (less than 1s), rinsed in MeOH, and reimmersed in the second thiol solution (1 mM in EtOH) for 1 min. In all experiments reported here, the first thiol listed is the component in which the electrolyses were performed. Twocomponent gradients composed of HDT/HDA, HDT/MPA, and PT/MPA were prepared in this manner. Contact Angle Measurements. Sessile drop contact angle measurements were made with 1 µL drops of deionized (18.2 MΩ cm) H2O dispensed directly from a syringe onto the Au gradient sample. The sample was placed in front of a ruler to mark position along the length of the film relative to the anodic end of the working electrode. The drops were imaged with a Javelin CCD camera and recorded onto VHS tape. The drop images were digitized with Snappy© video snapshot system and AVID© nonlinear digital video editor. The position and contact angles were measured in Adobe Photoshop©. Measurements were performed at 0, 1, 3, 7, 14, and 30 days after original gradient formation, except where noted. The samples were stored under various environmental conditions, including air (samples stored in the laboratory ambient), an N2-purged storage chamber, vacuum (∼ milliTorr background pressure), and 75% relative humidity (RH) in a humidity chamber prepared by a saturated NaCl solution in contact with its salt. Spatially Localized Copper Deposition. Copper was deposited from a stirred solution of 1 mM CuSO4 and 0.1 M K2SO4. All potentials are referenced to a saturated Ag/AgCl reference electrode. An in-plane potential gradient of 15 mV mm-1 was used to deposit Cu onto approximately half the electrode. Copper was subsequently removed from one side of the deposited area by transferring the electrode to a solution of 0.1 M K2SO4 blank and reversing the leads on the bipotentiostat such that the potential gradient was reversed. X-Ray Photoelectron Spectroscopy (XPS). XPS spectra were obtained on a Kratos Axis ULTRA Imaging X-ray Photoelectron spectrometer. Analysis was conducted under UHV (10-9 Torr) conditions using a monochromatic Al KR (1486 eV) X-ray source at 150 W. The spot size was ca. 1 mm. For quantitative measurements, the Au 4f region (92-75 eV) and Cu 2p region (965-923 eV) were used. The sample was introduced into the instrument held on a rectangular metal support secured by copper contact strips and then manipulated using a control pad that enabled movement of the sample in the x, y, and z directions once inside the sample analysis chamber. Spatial measurements were referenced to the glass/ gold boundary. This point was arbitrarily defined as the zero point (x position) for the analysis and corresponds to potentials of -400 and +550 mV on the potential axis in the deposition and stripping steps, respectively. For each successive point, the software decremented the x position by 2 mm until the opposite end of the glass/gold interface was reached. XPS imaging data were acquired by selecting a position on the surface where a Au/Cu transition existed as evidenced by the spectral data. In separate acquisitions, Cu photoelectrons were acquired at 932 eV and Au photoelectrons at 84 eV both for 2 min. The image was acquired at low magnification (800 × 800 µm) with an instrument power of 150 W and a resolution pass energy of 160 eV. Surface Plasmon Resonance Measurements. The optical apparatus for acquiring SPR measurements has been described in detail previously.29 An Ar+ laser (Coherent Innova 70 Series) was used to pump a tunable Ti-Sapphire laser (TITAN-CW

Balss et al. series, Schwartz Electrooptics) to produce 752.6 nm radiation as measured by a wavemeter (WA-10 Wavemeter, Burleigh Instruments). A cylindrical lens brought the radiation into a line focus on the Au film deposited on the base of a SF10 prism in the Kretschmann configuration. For imaging experiments, the cylindrical lens was removed, and the sample was illuminated with collimated radiation at the appropriate angle. Au-coated SF10 plates were placed into optical contact with an SF10 prism with index matching fluid (Cargille Laboratories). The angular distribution of reflected radiation was imaged onto a CCD camera (Photometrics PM512) employing Photometrics CCD9000 and IMAGE software, producing a linear mapping of radiation k-vector onto pixel position. Copper Deposition and Imaging. Spatially uniform copper was deposited from 0.1 mM CuSO4 and 0.1 M K2SO4 at -200 mV versus Ag/AgCl for 1 min. The stripping potential (0.3V) was applied after approximately 45 s at open circuit. Each deposition/stripping cycle time was repeated after 30 min at the stripping potential. SPR reflectance curves were collected every 2 s with 50 ms exposure time. The minimum reflectance position was determined by a polynomial fit to the reflectance in the vicinity of the minimum and represents the pixel location of the minimum intensity on the CCD chip. In Cu imaging experiments, the excitation angle was chosen to match the surface plasmon resonance of bare Au, so dark image areas represent uncovered Au and light areas are Cu-covered. An experimental SPR reflectance curve for bare gold in electrolyte was constructed to determine this optimal angle. The in-plane potential gradient used was 35 mV mm-1 representing ca. 355 mV potential variation between the O-rings. The potential gradient began at anodic potentials and was stepped in 50 mV increments in the cathodic direction in a 0.5 mM CuSO4/0.1 M K2SO4 solution until the entire surface was Cu-covered. Images were acquired with 50 ms exposure time after the potential gradient had been applied for 30 s. To test reproducibility, Cu was stripped by holding the entire surface at anodic potentials and reapplying the initial potential programs. The images were indistinguishable when compared to the original images. Flow Cell Assembly. A Kel-FTM flow cell assembly described previously29 was attached to the slide for both the imaging and spectroscopy experiments. The cell was formed by 2 O-rings (1 cm diameter) that sealed the prism on one side and a counter electrode (50 nm gold or stainless steel plate) on the other. A Au wire threaded into the cell went to an external vessel containing a saturated Ag/AgCl reference electrode. The flow was maintained throughout the experiment to prevent air bubbles from entering the Teflon tubing and breaking electrical contact with the reference electrode. Imaging of Spatially Controlled H2O2 Oxidation. A Au electrode was placed into a solution containing 10 mM H2O2 and 0.1 M K2SO4, and a potential gradient of 7 mV mm-1, spanning the potential of the oxidation wave, was applied. A photograph of the electrode recorded the generation of O2 bubbles after 20 min. Potentials are referenced to Ag/AgCl, and a Pt wire served as the counter electrode. Results and Discussion Two-Component Static Gradients. The strategy employed here relies on the ability to manipulate the ω-functional group of the alkanethiol to create two-component gradients, varying in some physical property of interest, which are stable over sufficiently long times to be used to manipulate their environment. Thus, it is important to determine the extent to which surface properties may be tuned by varying the terminal groups

Active Spatiotemporal Control

J. Phys. Chem. B, Vol. 105, No. 37, 2001 8973

TABLE 1: Stability Measured by Contact Angle Measurements on Compositional Gradients as a Function of Environment HDT/HDA nitrogen high > 20 daysa air medium > 7 days water vacuum ethanol high > 30 days thermal 37 °C 75% RH a

HDT/MPA

PT/MPA

high > 30 days low < 1 day medium > 14 days low < 1 day medium > 30 days high > 30 days high > 30 days low < 1 day medium > 7 days high > 30 days

Measurements collected at 0, 5, 10, and 20 days.

Figure 2. Plot of the sessile drop H2O contact angle as a function of spatial position (top abscissa) and applied potential (bottom abscissa) for a typical two-component gradient prepared from HDT and MPA.

and to elucidate the stability of the gradients prepared and stored under different conditions. Previous work from this laboratory demonstrated the ability to create two-component gradients with spatially varying wetting properties employing OT and MPA.4 To extend these initial studies, a variety of two-component gradients were prepared from molecules of the general formula, HS(CH2)nX, where one component was chosen to have hydrophilic X and the other a hydrophobic X. Components with different chain length, n, were used to probe how surface diffusion might affect the construction of these systems. Contact angle measurements were employed generally to verify the fabrication of two-component gradients. Controls were implemented by making contact angle measurements on full monolayers of the individual gradient components. Table 1 summarizes the range of samples prepared, the environmental conditions investigated and qualitative stabilities. Figure 2 shows H2O contact angle as a function of spatial position (applied potential) for a typical two-component system prepared from HDT and MPA. The gradient was constructed by first forming a SAM of the alkanethiol, HDT, and then selectively desorbing it in regions of potential cathodic of E0des, followed by re-immersion in the acid-terminated thiol, MPA. Because the rate of adsorption of MPA in the denuded areas of the surface is much larger than the rate at which MPA displaces HDT, the spatial positions corresponding to cathodic potentials are expected to be covered with the MPA, thereby yielding a large value of cos θ, as observed. The transition in contact angle begins at potentials corresponding roughly to the stripping potential of HDT, ca. -1 V versus that of Ag/AgCl, as expected. The H2O contact angles are consistent with a surface that is predominantly HDT at the anodic end and MPA at the cathodic end, with a mixed composition spanning the ∼8 mm wide transition region.

A key prerequisite for this approach is the kinetic stability of the SAM of the first component, while it is immersed in a solution of the second component. If the second component replaces the first at rates comparable to, or larger than, adsorption on bare Au, then the degree of spatial composition control will be degraded. Thus, the initial film component must be resistant to displacement by the second component. For this reason, the least favorably solvated component, e.g., the CH3terminated thiol for deposition from polar solvents, is always chosen as the first component. To examine this possible problem, adsorbate exchange experiments were performed, in which monolayers of one component were placed in the second component and the contact angles were subsequently compared to pure films of the test compound. Over the short times (1 min), the electrode was exposed to the second component no statistically significant changes in the contact angle were observed, except in the PT/MPA (C3/C3) system. The contact angle measured for the PT monolayer decreased by 28% after 1 min exposure to a solution of MPA. It is evident that MPA exchanges more readily with PT than with longer chain thiols, and in genera,l the exchange reaction is more rapid for shorter alkyl chain lengths. For example, HDT contact angles were stable even after 20 min exposure to either MPA (C3) or HDA (C16). The maximum change in wettability as measured by H2O contact angle between the ends of the electrode was found to be a strong function of the two components chosen to form the coadsorbed system, with PT/MPA yielding the smallest observed change, ∆θ ≈ 20°, and HDT/MPA the largest, ∆θ ≈ 70°. Gradient Stability. Gradient stability is an important consideration for any application in which the spatial anisotropy imparted by the composition distribution is to be used. One measure of stability is provided by observing how contact angles change over time while being exposed to various environments. Gradients used in these experiments were subjected to environments relevant to the ultimate goal of actively controlling transport of supermolecular objects and included several gasphase exposure conditions, solutions, and low pressure. H2O contact angle measurements, which are sensitive to the outermost atomic layers of the film30 and are therefore a sensitive probe of the ω-functional group, were used to assess the stability of all the gradients. All films were characterized by monitoring the changes in contact angles for a period of one month. Significant differences in stability were observed among the different two-component systems, as illustrated in Figures 3-5, which show the relative change in contact angle observed as a function of applied potential (and therefore position), and are summarized in Table 1. Table 1 defines stability according to the relative change in contact angle in the specified environments high stability meaning ∆θrel ) 20%, medium ∆θrel ) 30%, and low ∆θrel ) 30%. Figure 3 compares the stability of HDT/MPA gradients stored in three different gas phase environments. In general gradients were stable at 25˚C, with maximum changes in contact angle below 20% under most circumstances. The hydrophilic end of the gradient experienced larger relative changes in contact angle under each environmental condition with as much as 30% relative changes observed. An interesting observation was the relative stability of the sample stored in a 75% relative humidity chamber. We had initially hypothesized that adventitious H2O might play a role in degrading the spatially anisotropic composition profiles, but this sample experienced only ∆θrel ≈ 20% over a period of one month. In contrast, the sample stored at 37 °C experienced significant degradation after 2 weeks, suggesting that a thermally sensitive process, e.g., chain-melting,

8974 J. Phys. Chem. B, Vol. 105, No. 37, 2001

Figure 3. Relative changes in H2O contact angle along the length of two-component gradients as a function of time and environmental exposure. Data are given for HDT/MPA gradients stored under various conditions: (a) air, 37 °C; (b) air, 25 °C; and (c) air, 25 °C, 75% relative humidity. The solid lines are fits to an empirical sigmoidal function and are meant to be guides to the eye.

Figure 4. Relative changes in H2O contact angle along the length of two-component gradients as a function of time and environmental exposure. Data are given for (a) PT/MPA, (b) HDT/MPA, and (c) HDT/ HDA stored in EtOH at 25 °C. The solid lines are fits to an empirical sigmoidal function and are meant to be guides to the eye.

is responsible for the degradation in contact angle differentiation for samples stored in gas-phase ambients. Consistent with this interpretation, recent IR studies by Lennox et al.31 show clear evidence of thermally induced disorder in chain packing over this temperature range.

Balss et al.

Figure 5. Relative changes in H2O contact angle along the length of two-component gradients as a function of time and environmental exposure. Data are given for (a) PT/MPA, (b) HDT/MPA, and (c) HDT/ HDA stored in N2 at 25 °C. The solid lines are fits to an empirical sigmoidal function and are meant to be guides to the eye.

Figure 4 displays three systems with varying chain length in a solution environment. The plots are shown for samples stored in EtOH. The mixed chain length thiols fared the best, but gradients containing at least one long-chain, i.e., C16, component displayed good stability. The HDT/HDA system experienced more degradation on the HDT side, while the samples containing only short, i.e. C3, chains were not stable for long periods of time, presumably due to relatively rapid solution exchange reactions. Figure 5 is a comparison of the three systems in another gas-phase environment. The same trend was observed for this experiment, with the gradients containing at least one long-chain component displaying good stability. The C3 system once again was not stable for long periods of time. These experiments demonstrate that on a short time scale (less than 3 days) the majority of the gradients examined remain intact regardless of environment. On longer time scales, environmental stresses begin to play a role with the hydrophilic thiols being more easily affected. Spatially Directed Copper Deposition. Static counterpropagating two-component gradients consisting of ω-functionalized thiols can be used to manipulate the physical interaction of the surface with its immediate environment and naturally raises the question whether this general strategy can be used to selectively modify surfaces using other types of electrochemical reactions. Spatial patterning of Cu onto Au electrodes was examined in order to probe this point. Similar to the spatial patterning of alkanethiols from alkaline solutions, the fact that overpotential deposition of Cu metal from Cu2+ in sulfate solutions can be spatially mapped onto the electrode surface was exploited. First an in-plane potential gradient of 15 mV mm-1 was positioned so that Cu deposited on roughly half the electrode. The electrode was moved to blank electrolyte after deposition, and the potential gradient was reversed, thereby stripping Cu from the oxidative end of the potential gradient so that Cu metal remained only in the center of the electrode.

Active Spatiotemporal Control

J. Phys. Chem. B, Vol. 105, No. 37, 2001 8975

Figure 6. XPS data showing spatially controlled Cu (0) deposition in a stripe pattern on a Au (O) electrode. The potential axes, E1 for deposition and E2 for stripping, show how the copper strip-out phase of the experiment, which shows significant copper removal when the surface potential is positive of the anodic peak potential (+0.2V) for copper oxidation, determine the edge of the Cu stripe.

XPS was employed to map the resulting Cu distribution on the surface. Figure 6 shows the mass percent of Au and Cu as a function of spatial position for a representative sample, in which the thickness of the Cu layer was intentionally made very large. As Figure 6 clearly shows, this strategy was successful in producing a Cu line in the center of the Au electrode. The width of the Cu stripe (∼1 cm) represents a ∼150 mV wide window of potentials which were reductive in both the E1 and E2 potential programs. The edges of the Cu line correspond to potentials necessary to oxidize copper (>0.15 V vs Ag/AgCl) on both potential scales shown. The right (as seen in Figure 6) edge of the Cu stripe is defined by potentials which cannot reduce Cu2+ to Cu at Au in the first potential gradient, E1, and the left edge corresponds to the potential just able to strip Cu in the second potential gradient, E2. The region 40 mm < x < 65 mm is essentially devoid of Cu, since Cu deposition was inhibited in this region initially. A small residual amount of Cu remains in the region 0 mm < x < 20 mm; however, because the stripping kinetics needed for complete removal of Cu are quite slow, >48 h was required for complete removal under conditions used here. It is likely that using underpotentially deposited Cu would significantly improve the stripping kinetics, thereby improving the fidelity of the Au/Cu boundary. Although the edges of the Cu region appear reasonably sharp, the point-by-point image acquisition procedure used to obtain the data in Figure 6 obscures the true spatial resolution attained. Thus, XPS imaging was employed to characterize the edges of the Cu/Au region with higher spatial resolution. Figure 7 shows the transition between the bare Au electrode and the electrodeposited Cu. Several interesting features emerge from the image. At its narrowest, the Au-Cu transition occurs over a region as small as 200 µm, a transition significantly sharper than would be inferred from Figure 6. The ultimate width of the transition

Figure 7. XPS image showing the edge of the Cu stripe presented in Figure 6. Color scales for Au (top) and Cu (bottom) are given to the side of the image. Intermediate colors represent areas of the sample with mixed Cu and Au signals. Predominantly Au and Cu areas are labeled a and b, respectively.

region which might ideally be attained, and consequently the ultimate spatial resolution of this method of deposition, is a complicated function of the sample morphology, as evidenced by the strong structural dependence of copper deposits on Au(111) as compared to polycrystalline Au,32,33 and the applied potential gradient, since the true local potential is not as straightforward as the simple linear variation implied by eq 1. In fact, even a cursory consideration of the problem would identify two additional terms as given by

8976 J. Phys. Chem. B, Vol. 105, No. 37, 2001

Balss et al.

Figure 8. Location of the minimum in reflectance as a function of time for the deposition and stripping of Cu on evaporated Au in three successive deposition/stripping cycles. In each curve, the position marked i denotes application of a deposition potential -0.2Vvs Ag/AgCl, and position ii indicates the change to a stripping potential (0.3V vs Ag/AgCl). The stripping potential was held for a total of 30 min (last 10 min not shown) before the next cycle was performed.

V(x) ) V0 +

∫

iF(l) dl + δ(x) + κΓ(x) A

(2)

where δ(x) represents a stochastic, position-dependent contribution to the local potential derived from local morphology variations and the final term represents the small change in local resistance due to coverage, Γ(x), of molecular adsorbates.19,29 These second-order corrections to the local potential coupled to variations in flow and reagent delivery are likely responsible for the observation in Figure 7 of small islands of Cu in the Au region and Au-revealing pits in the Cu region. Ultimately, even with perfectly controlled surfaces, electrochemical phenomena have a finite energy spread, resulting from the width of the Fermi-Dirac distribution of carriers in the Au. Thus, in the absence of complicating factors which tend to degrade spatial resolution, the ultimate resolution limit of an all electrochemical deposition process is estimated to be ∼6 µm, on the basis of a limitation of the maximum in-plane field, due to a Joule heating of ∼102 V cm-1 and a 60 mV width of the Fermi-Dirac distribution at 300 K. SPR measurements were employed to monitor the deposition and stripping of Cu at Au surfaces. Although Cu and Au have similar dielectric functions at the 752 nm excitation wavelength, there is a sufficiently large shift that Cu deposition/stripping can be monitored at the Au-dielectric interface. SPR has previously been used in this way to characterize the corrosion behavior of Cu on Ag films.34 Figure 8 is a plot monitoring the position of the SPR reflectance minimum as a function of time as Cu is first deposited then stripped from the Au surface during three replicate trials with the same deposition and stripping potentials. The initial transient shift in SPR minimum position at point i in each cycle is associated with the electroreflectance contribution due to rearrangement of the electrical double layer.35,36 Subsequently, Cu begins to deposit, and the position of the SPR minimum reverses course and moves in an approximately linear fashion, until the potential is stepped to the oxidative stripping region at point ii. Integrated charge from chronocoulometry at point ii produces an equivalent Cu thickness of ∼1 nm if the deposited layer is assumed to have the density of bulk Cu. There are two regions of differing kinetic behavior after the potential is stepped to the oxidative stripping region. The overall reaction is slow, the SPR minimum requiring

20 min to return to baseline (note that the Cu films produced here are much thinner than the films produced in the Cu stripe experiments shown in Figure 6). While illustrating the relatively complex nature of Cu stripping under these conditions, the SPR measurements in and of themselves do not significantly illuminate the details of the stripping mechanism. Overpotential deposition of Cu certainly involves the formation of multiple layers, which has been discussed in detail for nonepitaxial metal deposition by Harrison and Thirsk.37 By logical extension, different desorption rates would be expected from different surface sites. The combination of these factors with the inherent surface roughness precludes a simple mechanistic model to explain the stripping behavior. Nevertheless, this experiment was useful in verifying that complete removal of Cu was feasible and in elucidating the concentration and deposition times necessary to move Cu surface populations spatially (vide infra). One of the key elements of the in-plane composition control strategy studied here is active spatiotemporal control of electrochemical events. Surface plasmon imaging, viz. Figure 9, was employed to verify that dynamic spatiotemporal control of inplane Cu coverage could be achieved in a manner similar to that employed previously to move alkanethiol gradients in-theplane.4 The SPR images in Figure 9 were acquired at the position of the bare Au resonance as indicated. SPR images were collected sequentially beginning with a bare Au electrode and moving the applied potential window so that eventually the entire electrode was covered with Cu. SPR images of the sample at four intermediate potential windows are shown in Figure 9. The most interesting feature of the figure is the transition region between Cu and Au, which occurs at the same potential in each image, a potential intermediate between the oxidation and reduction waves in the cyclic voltammogram. However, because the potential window is being translated across the surface of the Au working electrode, the Cu coverage appears to translate from the right side to the left side of the electrode. It is emphasized that the Cu atoms are not being transported laterally across the surface. Rather each movement of the Cu pattern involves a separate deposition step, i.e., new Cu atoms are added from solution. Presumably, the Cu thickness varies laterally because different horizontal positions see different overpotentials for different amounts of deposition time. The SPR imaging used

Active Spatiotemporal Control

J. Phys. Chem. B, Vol. 105, No. 37, 2001 8977

Figure 9. Top: Surface plasmon reflectance images of a Au electrode subjected to differing in-plane potential gradients. The SPR images were acquired at the position of the bare Au resonance. Each horizontal point in the images inside the O-rings (indicated by the arrows in the bottommost image) is referenced to the potential indicated on the x-axis. The potential window began at anodic potentials (top image) and moved in 50 mV increments until approximately 75% of the electrode was deposited with copper. The images were collected after 30 s of the applied potential gradient. Bottom: Cyclic voltammogram showing the Cu0/Cu2+ waves.

Figure 10. Spatially controlled oxidation of H2O2 along the surface of a Au electrode in 0.1 M K2SO4. The electrochemical potential gradient applied from left to right was as indicated vs Ag/AgCl.

to characterize these dynamic gradients is not sensitive to Cu thickness beyond the point where the reflectance becomes significantly larger than that of the bare Au. Another interesting feature of the image is the width of the transition region. On this macroscopic image, the boundary appears relatively sharp. In particular, it is narrower than the width of the corresponding reduction wave. Although part of the explanation for the apparent sharpness of the transition lies in the reflectance saturation phenomenon mentioned above, careful measurements with thiol-linked fluorescent microspheres consistently indicate transition widths smaller than the physical width implied by the half-width of the reduction wave.38 In a separate experiment, the potential windows were applied in random time order, taking care to completely strip Cu from the working electrode between each deposition cycle, and the SPR images acquired were indistinguishable from the images in Figure 9. Finally, we note the obvious difference between these Cu patterning experiments and the previously demonstrated patterning of alkanethiolates that in that reduction is desorptive for the organomercaptans, while it is used for deposition in the Cu/Au case shown here. Thus, the XPS and SPR experiments clearly show the ability to spatially pattern Cu onto a Au electrode, demonstrating the versatility of active spatiotemporal control for electrochemical patterning. Hydrogen Peroxide Oxidation Gradients. As a further demonstration of the versatility of the active spatial control of

electrochemical reactions, the oxidation of H2O2 to O2 was monitored by visually observing the formation of O2 bubbles as a function of position on the electrode surface. Figure 10 shows a photograph of a Au electrode in 0.1 M K2SO4 with a linear electrochemical potential gradient from left to right of +850 mV < V(x) < +1260 mV versus Ag/AgCl. Oxygen evolution appears at a position corresponding to a potential ∼+1050 mV, consistent with the oxidation wave for O2 formation from aqueous H2O2. More interesting, however, is the variation in the size of the O2 bubbles on the Au surface with the local potential, V(x). The Butler-Volmer equation

kb ) k0e(1-R)nfη

(3)

where R is the transfer coefficient and f ) F/RT, predicts that the rate of H2O2 oxidation reaction scales exponentially with the magnitude of the overpotential, η,28indicating more rapid oxidation for positions further to the right in the present experiment. Although the correlation is not perfectsthe bubble growth dynamics depend on factors other than the local rate of production of O2sthe bubble size does clearly increase with increasing positive potential (going to the right in Figure 10), with the extreme potential displaying the largest oxygen bubble. In contrast to the alkanethiol electrosorption and Cu deposition/ stripping examples of spatially directed electrochemical reactions, this reaction does not involve sorption and demonstrates

8978 J. Phys. Chem. B, Vol. 105, No. 37, 2001 the ability to direct a completely different kind of electrochemical reaction to designed spatial regions. Conclusions The versatility of the active spatiotemporal control of electrochemical reactions through dynamic electrochemical potential gradients has been demonstrated with three different types of reactions on Au: alkanethiol SAM electrosorption, Cu deposition and stripping, and O2 evolution from H2O2 oxidation. A number of interesting features have emerged from these experiments. Studies of the environmental stability of electrochemically generated monolayer gradients of ω-substituted alkanethiols indicate that formation and ultimate stability of such gradients requires a careful balancing of kinetic factors. In the formation process, the less soluble thiol is deposited first so that after desorption, the bare areas may be filled with a second more soluble thiol, with minimal displacement of the remaining thiol. Even so, there is some lability in the structures under certain storage conditions, the lability being worst for combinations of short chain thiols as demonstrated by the PT/MPA gradients. Dynamic spatiotemporal control of Cu deposition on Au electrodes also requires careful attention to kinetics. The formation of multilayer Cu films is rapid, but the subsequent stripping reaction is much slower, which has obvious implications for the dynamic movement of Cu patterns. Preparation of a Cu stripe on Au established how a sequence of different potential gradients could be used to define an area of deposition. Smaller stripes are certainly attainable, but the width must ultimately be limited by two factors: the maximum attainable in-plane field (limited by Joule heating at high currents) and the width of the Fermi-Dirac distribution. Finally, spatially directed oxidation of H2O2 was demonstrated by monitoring the formation of oxygen bubbles near the electrode. Consistent with predictions of the Butler-Volmer equation, the rate of bubble formation was found to depend on spatial position (overpotential) in these experiments. Acknowledgment. This work was supported by the Department of Energy through Grant DE FG02 96ER45439 and the National Science Foundation through Grant CHE 99-10236. XPS measurements were carried out at the Materials Research Laboratory Center for Microanalysis of Materials, which is supported by the Department of Energy through Grant DE FG02 96ER45439. Funding for the Kratos Axis ULTRA XPS system was provided by the National Science Foundation through Grant NSF DMR 99-77482 and The State of Illinois. The authors wish to acknowledge the example of excellence in electrochemical surface science over many decades established by Prof. Royce Murray and are proud to contribute to this festschrift in his honor. References and Notes (1) Abbott, N. L.; Folkers, J. P.; Whitesides, G. M. Science 1992, 257, 1380-1382.

Balss et al. (2) Liedberg, B.; Tengvall, P. Langmuir 1995, 11, 3821-3827. (3) Jeon, N. L.; Dertinger, S. K. W.; Chiu, D. T.; Choi, I. S.; Stroock, A. D.; Whitesides, G. M. Langmuir 2000, 16, 8311-8316. (4) Terrill, R. H.; Balss, K. M.; Zhang, Y.; Bohn, P. W. J. Am. Chem. Soc. 2000, 122, 988-989. (5) Bain, C. D.; Whitesides, G. M. Science 1988, 240, 62-63. (6) Bain, C. D.; Evall, J.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111, 7155-7164. (7) Bain, C. D.; Whitesides, G. M. J. Am. Chem. Soc. 1989, 111, 71647175. (8) Laibinis, P. E.; Nuzzo, R. G.; Whitesides, G. M. J. Phys. Chem. 1992, 96, 5097-5105. (9) Folkers, J. P.; Laibinis, P. E.; Whitesides, G. M. Langmuir 1992, 8, 1330-1341. (10) Atre, S. V.; Liedberg, B.; Allara, D. L. Langmuir 1995, 11, 38823893. (11) Imabayashi, S.; Gon, N.; Sasaki, T.; Hobara, D.; Kakiuchi, T. Langmuir 1998, 14, 2348-2351. (12) Hobara, D.; Ota, M.; Imabayashi, S.; Niki, K.; Kakiuchi, T. J. Electroanal. Chem. 1998, 444, 113-119. (13) Imabayashi, S.; Hobara, D.; Kakiuchi, R. Langmuir 1997, 13, 4502-4504. (14) Tidwell, C. D.; Ertel, S. I.; Ratner, B. D Langmuir 1997, 13, 34043413. (15) Schonherr, H.; Ringsdorf, H.; Jaschke, M.; Butt, H. J.; Bamburg, E.; Allinson, H.; Evans, S. D. Langmuir 1996, 12, 3898-3904. (16) Widrig, C. A.; Chung, C.; Porter, M. D. J. Electroanal. Chem. 1991, 310, 335-359. (17) Weisshaar, D. E.; Lamp, B. D.; Porter, M. D. J. Am. Chem. Soc. 1992, 114, 5860-5862. (18) Imabayashi, S.; Iida, M.; Hobara, D.; Feng, Z. Q.; Niki, K.; Kakiuchi, T. J. Electroanal. Chem. 1997, 428, 33-38. (19) Zhang, Y.; Terrill, R. H, Bohn, P. W. J. Am. Chem. Soc. 1998, 120, 9969-9970. (20) Vaskevich, A.; Rubinstein, I. J. Electroanal. Chem. 2000, 491, 8794. (21) Nishizawa, M.; Sunagawa, T.; Yoneyama, H. Langmuir 1997, 13, 5215-5217. (22) Santos, M. C.; Mascaro, L. H.; Machado, S. A. S. Electrochim. Acta 1998, 43, 2263-2272. (23) Legault, M.; Blum, L.; Huckaby, D. A. J. Electroanal. Chem. 1996, 409, 79-86. (24) Oyamatsu, D.; Kuwabata, S.; Yoneyama, H. J. Electroanal. Chem. 1999, 473, 59-67. (25) Andersen, J. E. T.; Bech-Nielsen, G.; Moller, P.; Reeve, J. C. J. Appl. Electrochem. 1996, 26, 161-170. (26) Randler, R. J.; Kolb, D. M.; Ocko, B. M.; Robinson, I. K. Surf. Sci. 2000, 447, 187-200. (27) Xia, X. H.; Schuster, R.; Kirchner, V.; Ertl, G. J. Electroanal. Chem. 1999, 461, 102-109. (28) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.; John Wiley & Sons: New York, 2001. (29) Zhang, Y.; Terrill, R. H.; Bohn, P. W. Anal. Chem. 1999, 71, 119125. (30) Bain, C. D.; Whitesides, G. M. J. Am. Chem. Soc. 1988, 110, 58975898. (31) Bensebaa, F.; Ellis, T. H.; Badia, A.; Lennox, R. B. Langmuir 1998, 14, 2361-2367. (32) Zhou, M.; Myung, N.; Chen, X.; Rajeshwar, K. J. Electroanal. Chem. 1995, 398, 5-12. (33) Palomar-Pardave, M.; Gonzalez, I.; Batina, N. J. Phys. Chem. B 2000, 104, 3545-3555. (34) Bussjager, R. J.; Macleod, H. A. Appl. Opt. 1996, 35, 5044-5047. (35) Gordon, J., II. SPIE 1981, 276, 96-100. (36) Gordon II, J.; Melroy, O.; Borges, G.; Reisner, D.; Abruna, H.; Chandrasekhar, P.; Albarelli, M.; Blum, L. J. Electroanal. Chem. 1986, 210, 311-314. (37) Harrison, J. A.; Thirsk, H. R. The Fundamentals of Metal Deposition; Marcel Dekker: New York, 1971; Vol. 5. (38) Plummer, S. T.; Bohn, P. W. Unpublished Results 2001.