Anal. Chem. 1996, 68, 2010-2014
Microderivatization of Anodized Glassy Carbon Betsy B. Ratcliff, James W. Klancke, Miles D. Koppang, and Royce C. Engstrom*
Department of Chemistry, University of South Dakota, Vermillion, South Dakota 57069
Microelectrodes have been used to modify locally the electrochemical activity on glassy carbon electrodes. Glassy carbon was electrochemically oxidized to form an oxide layer which is inhibitory toward certain electrontransfer reactions. Activity was restored through the application of hydroxide, which was generated electrochemically at the tip of a microelectrode. With the tip positioned in close proximity to the anodized glassy carbon surface, microdomains of electrochemical activity were created in an otherwise inactive matrix. The distribution of electrochemical activity was characterized using electrochemical feedback at the microelectrode, electrogenerated chemiluminescence imaging, and electrodeposition of silver. Spatially directed activation of the glassy carbon surface was accomplished in the micrometer domain. The design of successful electroanalytical devices usually requires the incorporation of functions other than simply electron transfer to impart selectivity, sensitivity, or electrode protection. A multitude of electrode modification schemes have appeared in the literature based on adsorption, covalent attachment, coating with polymer films, and other approaches (ref 1 and references therein). In some cases, however, the modification may interfere with electron transfer, pointing out the need for an optimized distribution of electroactivity and the other chemical functions on the electrode surface. The recent work of Pantano and Kuhr, where dehydrogenase enzymes were linked to carbon fiber surfaces using a biotin-avidin attachment protocol,2 provides an example. In that case, it was observed that the electron-transfer properties degraded with the extent of enzyme coverage, so that an optimized enzyme loading was required to maximize both enzymatic activity and electrochemical activity. In such cases, it may be beneficial to design surfaces with multiple functions that are restricted to distinct physical locations, separated by distances small on a diffusional scale so that products of one function are in intimate communication with another. In other words, microscopic compartmentalization of the various functions may lead to an optimized device. Microelectrodes have been used to both characterize and modify the surfaces and interfacial regions of other electrodes by positioning a microelectrode tip within the diffusion layer of the surface being studied. We used microelectrodes in a generatorcollector mode to map the distribution of electrochemical activity and obtain diffusion layer profiles of transient species,3,4 and Bard and co-workers have developed scanning electrochemical micros(1) Ryan, M. D.; Bowden, E. F.; Chambers, J. Q. Anal. Chem. 1994, 66, 385R. (2) Pantano, P.; Kuhr, W. G. Anal. Chem. 1993, 65, 623. (3) Engstrom, R. C.; Weber, M.; Wunder, D.; Winquist, S.; Burgess, R. Anal. Chem. 1986, 58, 844. (4) Engstrom, R. C.; Meaney, T.; Tople, R.; Wightman, R. M. Anal. Chem. 1987, 59, 2005.
2010 Analytical Chemistry, Vol. 68, No. 13, July 1, 1996
copy (SECM) into a powerful and versatile technique based largely on the feedback mode (for example, refs 5-8 and references therein). Microelectrodes can also be used as a reagent delivery device, electrogenerating a product that can react with the underlying substrate. Thus, Shoat and Mandler used SECM to create nickel(II) hydroxide microstructures on a platinum surface by locally increasing the hydroxide concentration, creating a microenvironment in which the solubility of the metal hydroxide was exceeded.9 SECM has been used to modify surfaces in a spatially resolved manner through the deposition of metals onto electrodes,10,11 etching of metallic and semiconductor substrates,12,13 and localized electrodeposition of polymers.14 SECM has also been used to deactivate microscopically local regions of a surface coated with the enzyme diaphorase by generating bromine at the tip of a microelectrode.15 In each case, the microelectrode-created diffusion layer has a composition which differs from that of the bulk solution; consequently, reactions can be made to occur within the diffusion layer that will not occur outside of the layer. Positioning the microelectrode tip over a desired location on a surface can therefore lead to surface reactions that are restricted to regions in contact with the diffusion layer of the microelectrode. The spatial dimension of most of the surface modification work carried out with microelectrodes has been in the micrometer range. In this work, we have used microelectrodes to modify locally the electrochemical activity on glassy carbon electrodes. This approach relies upon the observation that glassy carbon surface chemistry can be readily manipulated through electrochemical pretreatment. Glassy carbon electrodes have been subjected to a wide range of procedures in an effort to improve electron transfer kinetics, including electrochemical pretreatments, laser activation, vacuum heat treatment, and polishing schemes. The effects of these treatments on the structure and performance characteristics have been reviewed in detail by McCreery.16 In earlier work from our laboratory, we noted that freshly polished glassy carbon became deactivated with respect to certain redox couples upon anodization at 1.8 V vs SCE in solutions of neutral pH.17,18 The (5) Bard, A. J.; Denuault, G.; Lee, C.; Mandler, D.; Wipf, D. O. Acc. Chem. Res. 1990, 23, 357. (6) Horrocks, B. R.; Schmidtke, D.; Heller, A.; Bard, A. J. Anal. Chem. 1993, 65, 3605. (7) Pierce, D. T.; Bard, A. J. Anal. Chem. 1993, 65, 3598. (8) Solomon, T.; Bard, A. J. Anal. Chem. 1995, 67, 2787. (9) Shoat, I.; Mandler, D. J. Electrochem. Soc. 1994, 141, 995. (10) Mandler, D.; Bard, A. J. J. Electrochem. Soc. 1990, 137, 1079. (11) Meltzer, S.; Mandler, D. J. Electrochem. Soc. 1995, 142, L82. (12) Mandler, D.; Bard, A. J. J. Electrochem. Soc. 1989, 136, 3143. (13) Mandler, D.; Bard, A. J. Langmuir 1990, 6, 1488. (14) Wuu, Y.-M.; Fan, F.-R. F.; Bard, A. J. J. Electrochem. Soc. 1989, 136, 885. (15) Shiku, H.; Takeda, T.; Yamoda, H.; Matsue, T.; Uchia, I. Anal. Chem. 1995, 67, 312. (16) McCreery, R. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1991; Vol. 17. (17) Engstrom, R. C. Anal. Chem. 1982, 54, 2310. (18) Engstrom, R. C.; Strasser, V. A. Anal. Chem. 1984, 56, 136. S0003-2700(95)01139-5 CCC: $12.00
© 1996 American Chemical Society
deactivation was presumably due to the formation of an oxide layer, which subsequently has been described as a graphitic oxide, being nonconductive, transparent, and of microscopic thickness or less19 but with characteristics and composition dependent upon the precise nature of the pretreatment conditions. We further noted that activity of the anodized glassy carbon could be restored through a variety of mechanisms, including mechanical removal of the film, cathodization at sufficiently negative potentials, or exposure to various chemical agents, including hydroxide. Beilby and co-workers have reported that the graphitic oxide formed during anodization is unstable in high concentrations of hydroxide.20,21 The use of potential pulsing sequences to modify activity of electrodes in general has been demonstrated by a number of research groups (for example, ref 22 and references therein). Here, we have used microelectrodes to generate hydroxide electrochemically while the tip was in close proximity to an anodized glassy carbon surface, thereby creating microdomains of electrochemical activity in an otherwise inactive matrix. Spatial control of reactivation was obtained by positioning the microelectrode as in SECM. Examination of electrochemical activity was accomplished using both electrochemical feedback at the microelectrode and the technique of electrogenerated chemiluminescence (ECL) imaging,23,24 in which light from the luminol ECL reaction25 is collected through a microscope to provide images of electrochemical activity. The microscopic manipulation of the glassy carbon surface is of interest because glassy carbon has desirable electrochemical characteristics while possessing a rich surface chemistry that can be used in further modification schemes. EXPERIMENTAL SECTION Reagents. All solutions were made from reagent grade chemicals used without further purification and with water purified on a commercial cartridge-based system. Electrochemical manipulations of the glassy carbon surface were carried out in a solution of 10 mM potassium ferricyanide, either 1.0 or 0.10 M potassium chloride, and 1 mM potassium dihydrogen phosphate. Electrogenerated chemiluminescence was done in a solution that was 2 mM luminol (5-amino-2,3-dihydro-1,4-phthalazinedione), 5 mM hydrogen peroxide, and 50 mM sodium borate. Luminol was obtained from Eastman Kodak (Rochester, NY), and the pH of the luminol solution as prepared was 9.4. Electrodeposition of silver was done from 1 mM silver nitrate in 0.1 M potassium nitrate. Apparatus. Cyclic voltammetry was carried out with a Bioanalytical Systems CV-27 potentiostat (West Lafayette, IN) under external computer control. During microelectrode activation or characterization experiments, an Ensman 400-EI Bipotentiostat (Bloomington, IN) was used to control the potentials and monitor the currents of the two working electrodes. All potentials given are with respect to a silver/silver chloride reference electrode, and a platinum wire served as an auxiliary electrode. Glassy carbon electrodes of 3 mm diameter were obtained from Bioanalytical Systems. Before each experiment, the glassy carbon (19) Kepley, L. J.; Bard, A. J. Anal. Chem. 1988, 60, 1459. (20) Beilby, A.; Carlsson, A. J. Electroanal. Chem. 1988, 248, 283. (21) Beilby, A. L.; Sasaki, T. A.; Stern, H. M. Anal. Chem. 1995, 67, 976. (22) Johnson, D. C.; LaCourse, W. R. Electroanalysis 1992, 4, 367. (23) Engstrom, R. C.; Pharr, C. M.; Koppang, M. D. J. Electroanal. Chem. 1987, 251. (24) Engstrom, R. C.; Johnson, K. W.; DesJarlais, S. Anal. Chem. 1987, 59, 670. (25) Haapakka, K. E.; Kankare, J. J. Anal. Chim. Acta 1982, 138, 253.
electrode was sanded with 600 grit sandpaper and polished successively with 1.0, 0.3, and 0.05 µm alumina polish on a Buehler TEXMET 1000 polishing cloth (Lake Bluff, IL). Carbon fiber microelectrodes were prepared as described in the literature26 and beveled in a specially designed holder that allowed the finished microelectrode surface to be parallel to the glassy carbon surface during activation experiments. A cell was machined from nylon which allowed proper positioning of the glassy carbon and carbon fiber electrodes. Micromanipulation of the microelectrode was carried out with a Klinger Scientific Model CC1.2 positioner (Richmond Hill, NY) capable of 1 µm resolution in three dimensions. The imaging system used for the luminescence work consisted of a Carl Zeiss Standard 16 fluorescence microscope (Thornwood, NY) equipped with Plan NeoFluar objectives with magnifications of 6.3×, 25×, and 63×. An additional 10× was provided by the ocular stage. Images were obtained with the 6.3× objective, and resolution in this configuration was 3.1 µm/pixel. Light was collected, processed, and displayed with a Hamamatsu Photonics Argus 100 video imaging system (Bridgewater, NJ) which consists of a low-lag Vidicon camera with a microchannel plate intensifier, interfaced to a computer and a control panel. The images shown here were photographed from the monitor. A Princeton Applied Research Model 174A potentiostat (Princeton, NJ) was used to control the potential of the electrode during ECL imaging. The fluorescence experiment demonstrating the sphere of influence at a microelectrode was carried out with commercially available platinum-iridium microelectrodes purchased from FHC, Inc. (New Brunswick, ME). The same microscope and imaging system were used as described in the preceding paragraph, except that the microscope was operated in the fluorescence mode using a mercury lamp, a primary filter with a cutoff of 480 nm, a secondary filter of 515 nm, and a dichroic mirror operating at 505 nm. Procedure. The polished glassy carbon electrode was anodized at 1.8 V for 3-6 min (see Discussion) in 0.10 M potassium nitrate. A cyclic voltammogram in the potassium ferrocyanide solution described above was taken before and after anodization to ensure that deactivation had, indeed, occurred during anodization. The microelectrode was positioned near the anodized glassy carbon surface by observing negative feedback of ferricyanide reduction current induced by blockage of diffusion by the inactive anodized surface. Negative feedback was observed when the microelectrode was lowered to within ∼20 µm of the surface. The microelectrode was taken to be at the surface when two successive 1 µm steps failed to yield a further decrease in current, and we estimated that the actual distance between the microelectrode and the glassy carbon surface was 2-3 µm. This procedure was used to avoid physical contact with the anodized coating, although it was found that slight microelectrode contact with the anodized layer did not have any consequences. Once the microelectrode was positioned on the anodized surface, the microelectrode potential was stepped into water reduction to produce hydroxide; typical potentials were -1.6 to -1.7 V. Additional details of the activation procedure are included with the relevant portion of the Discussion section. Following activation, the electrode was generally transferred to the ECL imaging apparatus for examination. (26) Kelly, R. S.; Wightman, M. R. Anal. Chim. Acta 1986, 186, 79.
Analytical Chemistry, Vol. 68, No. 13, July 1, 1996
2011
Figure 2. (A) Cathodic peak charge and (B) cathodic peak potential versus pH of electrode soak. Electrode was anodized for 5 min at a potential of +1.75 V in 0.1 M KNO3 before soaking for 1 h in solution at each indicated pH. Figure 1. Cyclic voltammograms on 3.0 mm diameter glassy carbon electrode in 50 mM potassium ferricyanide and 0.1 M potassium chloride. Scan rate, 100 mV/s. (A) Freshly polished electrode. (B) After anodization for 3 min at 1.8 V in 0.1 M KNO3. (C) Negativegoing scan after anodization. (D) After anodization and cathodization. (E) After anodization and 40 min soak in 0.10 M potassium hydroxide. (F) After anodization and 80 min soak in 0.10 M potassium hydroxide.
RESULTS AND DISCUSSION Behavior of Glassy Carbon. The effect of anodization on glassy carbon is summarized in Figure 1A and B, which shows voltammograms of potassium ferrocyanide before and after anodization at 1.8 V for 3 min. Figure 1B shows that activity with respect to the ferri-/ferrocyanide couple is completely absent at the anodized surface. Upon cathodization during a potential scan (Figure 1C), a large amount of charge is passed in a process that is not yet well-defined but likely involves the reduction of surfacebound oxides formed in the anodization step. Voltammograms of potassium ferrocyanide taken subsequent to the cathodization are again well-defined (Figure 1D). Exposure of the anodized GC to basic solutions simply by soaking in a potassium hydroxide solution also results in a partial restoration of activity, as shown in Figure 1E and F. The kinetics of ferrocyanide electrochemistry at a base-restored electrode are not as fast as those at the cathodized electrode, but nonetheless, effective electron transfer still occurs. Furthermore, the shape of the voltammograms at the base-restored electrode suggests the possibility of a microelectrode array, and we are in the process of further characterizing the reactivated surface. The pH dependence of the base reactivation of anodized glassy carbon is shown in Figure 2A, where the charge under the cathodic process (Figure 1C) is observed to decrease if the 2012 Analytical Chemistry, Vol. 68, No. 13, July 1, 1996
anodized surface is exposed to solutions of pH 8 or higher. (We have noted that the actual charge in the cathodic process varies from electrode to electrode, so the maximum charge associated with Figure 2A is somewhat smaller than that in the cathodic process of Figure 1C.) Likewise, the peak potential of the cathodic process shows a well-defined transition centered around pH 8 of the soaking solution (Figure 2B). Although the mechanism of reactivation by either cathodization or exposure to base is still not defined,18 these results show that a surface anodized at neutral pH is inactive toward ferrocyanide activity, and exposure of the anodized surface to sufficiently basic conditions restores its activity. These observations lay the foundation for the microscopic restoration of activity through delivery of hydroxide from a microelectrode tip. Microelectrode “Sphere of Influence”. Hydroxide can be readily produced electrochemically through any of the following cathodic reactions, which all occur at large overpotentials at carbon electrodes:
2e- + O2 + 2H2O h H2O2 + 2OH4e- + O2 + 2H2O h 4OH2e- + 2H2O h H2 + 2OHAs a result of any of these reactions in weakly buffered solutions, the diffusion layer becomes a microscopic volume with a pH distinctly basic compared to that of the surrounding solution. Because of the spherical nature of diffusion at a microelectrode, a “sphere of influence” is established, so that reactions favored by basic conditions should proceed to a greater degree inside the sphere than outside. This is illustrated in Figure 3, which is a
Figure 3. Fluorescence imaging at a Pt-Ir microelectrode. Each frame represents a 1 s image integration during a potential sweep from -0.90 to -1.05 V at a scan rate of 50 mV/s. Solution, 0.1 mM fluorescein and 0.10 M potassium chloride adjusted to pH 4.0 with dilute hydrochloric acid.
series of four sequential images of an epoxy-coated Pt-Ir microelectrode (with exposed tips of 1-3 µm) during a potential scan in the presence of the fluorescent indicator, fluorescein. The images were collected through a fluorescence microscope. In this experiment, the bulk solution contained 1 × 10-4 M fluorescein and 0.1 M potassium chloride, adjusted to pH 4 with mineral acid and otherwise unbuffered. Upon scanning the microelectrode potential into oxygen and water reduction, the electrogenerated hydroxide deprotonates the fluorescein, converting it to its fluorescent form. Microscopic Activation of Anodized Glassy Carbon. Reactivation of anodized glassy carbon was carried out using carbon fiber electrodes rather than platinum-iridium tips, the latter being considerably more fragile and expensive. Evidence for spatially selective reactivation can first be seen in the nature of electrochemical feedback at a microelectrode. Figure 4, curve A, shows a voltammogram taken with the microelectrode in a solution of 10 mM potassium ferricyanide, with the microelectrode positioned well away from the glassy carbon surface. Upon positioning the microelectrode within a few micrometers of the anodized surface, the microelectrode exhibited a decreased diffusional current, curve B, referred to as “negative feedback” in the terminology of scanning electrochemical microscopy.5 Negative feedback is observed because the electrochemically inactive glassy carbon surface blocked diffusional mass transfer to the microelectrode. The microelectrode was then used to generate base by stepping into water reduction for ∼3 min. Without moving the microelectrode, ferricyanide voltammetry then exhibited positive feedback, curve C, indicating that the site had been activated. Positive feedback resulted from recycling of electroactive material between the two closely spaced, electrochemically active surfaces,5 one being the microelectrode and the other the activated site on the glassy carbon. We did not attempt any quantitative interpretation of the feedback current. Further evidence for spatially directed activation of the anodized surface was obtained with ECL imaging. After reactivation of a single site with the microelectrode, the glassy carbon electrode was immersed in a basic solution of luminol and hydrogen peroxide (see Experimental Section). Upon application of 0.60 V, the luminol ECL reaction occurred at those regions of the glassy carbon capable of supporting electron transfer, thereby
Figure 4. Cyclic voltammograms of a carbon fiber microelectrode in 10 mM potassium ferricyanide and 0.10 M KCl. Scan rate, 100 mV/s. Microelectrode is placed (A) in bulk solution, (B) within 2-3 µm or the anodized surface, and (C) within 2-3 µm of a reactivated site.
Figure 5. Electrogenerated chemiluminescence at a single microsite reactivated by microelectrode generation of hydroxide. ECL image obtained by photon collection for 30 s following a potential step to +0.6 V in luminol solution (see text for solution details).
producing a light image that served to visualize any electrochemically active sites. Chemiluminescence intensity at a single microelectrode-activated site is shown as a gray-level image in Figure 5. The site resulted from the application of -1.6 V to a carbon fiber microelectrode positioned at the glassy carbon surface. This image demonstrates that the ECL reaction of luminol is also inhibited at the anodized surface but is restored at base-reactivated electrode. The diameter of the site in Figure 5 is ∼40 µm. Electrochemical patterning was accomplished by positioning the carbon fiber microelectrode repetitively and applying hydroxidegenerating pulses, as shown in Figure 6. The process is not a rapid one, in that each location resulted from a 3 min pulse. It should be noted that this system may have promise in a “readwrite” mode, since another anodization step would result in another completely deactivated surface, thereby “erasing” any pattern placed on it. Dimensions of Active Sites. At this stage, we have not attempted to improve the resolution of the technique to sites Analytical Chemistry, Vol. 68, No. 13, July 1, 1996
2013
Figure 6. Spatially directed microelectrode reactivation. Each site activated by microelectrode positioned at anodized surface and maintained at 100 nA for 4 min. ECL image obtained by photon collection for 30 s during a potential step to +0.6 V in luminol solution.
Figure 7. Current dependence study. ECL image of an anodized electrode with sites activated by microelectrode positioned at anodized surface and maintained at indicated currents for 4 min each. Photons collected for 30 s following a potential step to +0.6 V in luminol solution.
smaller than those available with the carbon fiber microelectrodes. However, Figures 7 and 8 show that reactivated site size is dependent upon the microelectrode current and duration of the hydroxide pulse, respectively. Subsequent work with this system will focus on decreasing site size to the submicrometer domain using smaller microelectrodes, optimizing solution buffer strength and potential-pulsing protocols to restrict the size of the diffusion layer. Deposition of Silver into Active Sites. In addition to showing activity toward ferrocyanide and luminol, the reactivated sites provided the capability to spatially direct silver electrodeposition. The anodized glassy carbon still retains activity with respect to silver ion reduction; however, the silver reduction peak is shifted negatively at the anodized surface relative to a freshly polished surface (Ep ) 0.10 and 0.25 V, respectively). Thus, with the GC electrode held at a potential of 0.25 V, silver would be expected to deposit only in the reactivated sites, but not on the anodized surface. Figure 9 is a light micrograph showing a triangular pattern of microscopic silver structures that have been deposited in sites formed through prior reactivation with the microelectrode. 2014 Analytical Chemistry, Vol. 68, No. 13, July 1, 1996
Figure 8. Time dependence study. ECL image of an anodized electrode with sites reactivated by microelectrode positioned at anodized surface and maintained at 100 nA for indicated times. Photons collected for 30 s following a potential step to +0.6 V in luminol solution.
Figure 9. Micrograph of silver deposited into activated sites on anodized glassy carbon. Image has been contrast enhanced.
CONCLUSIONS The anodized GC surface provides a chemically rich matrix (the anodized surface contains a high density of carbon-oxygen functionalities) in which microsites of electrochemical activity can be placed with microelectrode activation. The system holds promise for the creation of spatially complex surfaces in which microscopic regions of electroactivity are surrounded by functionalities capable of carrying out chemical steps prior to electron transfer. The reagent delivery mode of electrochemical microscopy permits carrying out surface modification, such as the carbon reactivation shown here, in a spatially controlled manner. Furthermore, information about the anodization layer itself may result from microelectrode manipulations of its structure. ACKNOWLEDGMENT This work was supported in part by the National Science Foundation, Grant OSR-9108773, and the South Dakota Future Fund. This research was presented in part at the 1996 Pittsburgh Conference, Chicago, IL. Received for review November 27, 1995. Accepted March 25, 1996.X AC9511399 X
Abstract published in Advance ACS Abstracts, May 1, 1996.
