Anal. Chem. 2005, 77, 6205-6217
Fluorescence Confocal Laser Scanning Microscopy as a Probe of pH Gradients in Electrode Reactions and Surface Activity Nicola C. Rudd, Susan Cannan, Eleni Bitziou, Ilenia Ciani, Anna L. Whitworth, and Patrick R. Unwin*
Department of Chemistry, University of Warwick, Coventry, CV4 7AL, U.K.
The application of fluorescence confocal laser scanning microscopy (CLSM) to quantify three-dimensional pH gradients near electrode surfaces is described. The methodology utilizes a trace quantity of a fluorescent dye, fluorescein, in solution, which fluoresces strongly above pH 6.5, to map the pH adjacent to various ultramicroelectrodes undergoing electrochemical processes that lead to pH changes. The experimental fluorescence profiles, determined by CLSM, have been compared to models by solving the underlying mass transport equations, including the effect of natural convection, using the finite element method. The methodology has been validated through studies of the galvanostatic reduction of water at both disk and ring ultramicroelectrodes. The fluorescence profiles were found to be highly sensitive to both the initial bulk solution pH and applied current in a predictable fashion. The potentiostatic reduction of oxygen has been investigated at 25- and 10-µm-diameter platinum electrodes to confirm the effective number of electrons transferred in the reaction. Finally, the application of this methodology to observe defects in microelectrode arrays, particularly those that cannot be seen by optical microscopy, is described. Concentration gradients are formed during many interfacial processes, for example, crystal dissolution and growth,1 corrosion,2 and phase transfer at liquid/liquid interfaces.3,4 In electrochemical reactions, the concentration profile at, or near to, an electrode surface contains valuable information on mass transport and reactivity. Consequently, there is considerable interest in the development of in situ techniques to map concentration profiles at electrode surfaces and at other interfaces. Several methods probe concentration gradients locally by utilizing the optical properties of electrogenerated (or consumed) species. These techniques include interferometry5-7 and spectroscopic approaches such as Raman spectroscopy8 and confocal * To whom correspondence should be addressed. Tel.: +44-24-7652-3264. Fax: +44-24-7652-4112. E-mail:
[email protected]. (1) Unwin, P. R.; Macpherson, J. V. Chem. Soc. Rev. 1995, 24, 109. (2) Fantini, J.; Fournier, D.; Boccara, A. C.; Plichon, V. Electrochim. Acta 1997, 42, 937. (3) Slevin, C. J.; Zhang, J.; Unwin, P. R. J. Phys. Chem. B 2002, 106, 3019. (4) Zhang, J.; Slevin, C. J.; Unwin, P. R. Chem. Commun. 1999, 1501. (5) Pawliszyn, J. Anal. Chem. 1992, 64, 1552. (6) Yu, X.; Yue, X.; Gao, H.; Chen, H. J. Cryst. Growth 1990, 106, 690. (7) Van Dam, J. C.; Mischgofsky, F. H. J. Cryst. Growth 1987, 84, 539. 10.1021/ac050800y CCC: $30.25 Published on Web 08/25/2005
© 2005 American Chemical Society
resonance Raman microscopy.9 Interferometry measures the refractive index gradient, which is affected by concentration gradients. It is nonselective, since any solute may locally alter the refractive index, and has low sensitivity, so that fairly high concentrations of analytes are required (∼0.1 M).10 However, it has been estimated that concentration profiles may be measured with submicrometer resolution.11 The probe beam deflection method12,13 provides a more selective means of monitoring refractive index gradients and, by inference, concentration gradients. This approach has been used to study, for example, copper oxidation and cupric ion reduction,14 as well as mass transport.15 The technique has also been combined with the electrochemical quartz crystal microbalance to investigate the ingress and egress of ions during the switching of polymer-modified electrodes.16 Techniques based on light absorption, rather than refractive index, have the advantage of greater sensitivity and selectivity. Spatially resolved absorbance spectroelectrochemistry has been used to profile electrogenerated light-absorbing species. Concentration profiles measured by this technique have provided information on reaction kinetics, mechanisms, and the stoichiometry of redox systems.17-19 A variation on this theme, spatial imaging photometry, has profiled the concentration changes inside an ionselective electrode with a nominal spatial resolution of 1.25 µm.20 Recent advances in synchrotron radiation sources provide highintensity X-ray beams with dimensions of less than 20 µm, so allowing X-ray absorption spectroscopy to be used on a local scale. This approach has been used to investigate the diffusion of Cu2+ ions away from an electrode with millisecond time resolution.21 High-resolution nuclear magnetic resonance spectroscopy has also (8) Ozeki, T.; Irish, D. E. J. Electroanal. Chem. 1990, 280, 451. (9) Amatore, C.; Bonhomme, F.; Bruneel, J.-L.; Servant, L. Electrochem. Commun. 2000, 2, 235. (10) Jan, C.; McCreery, R. L. Anal. Chem. 1986, 58, 2771. (11) Kragt, H. J.; Smith, C. P.; White, H. S. J. Electroanal. Chem. 1990, 278, 403. (12) Pawliszyn, J.; Weber, M. F.; Dignam, M. J.; Mandelis, A.; Venter, R. D.; Park, S.-M. Anal. Chem. 1986, 58, 236. (13) Pawliszyn, J.; Weber, M. F.; Dignam, M. J.; Mandelis, A.; Venter, R. D.; Park, S.-M. Anal. Chem. 1986, 58, 239. (14) Russo, R. E.; McLarnon, F. R.; Spear, J. D.; Cairns, E. J. J. Electrochem. Soc. 1987, 134, 2783. (15) Barbero, C.; Miras, M. C.; Kotz, R. Electrochim. Acta 1992, 37, 429. (16) Vilas-Boas, M.; Henderson, M. J.; Freire, C.; Hillman, A. R.; Vieil, E. Chem. Eur. J. 2000, 6, 1160. (17) Deputy, A. L.; McCreery, R. L. J. Electroanal. Chem. 1988, 257, 57. (18) Deputy, A. L.; McCreery, R. L. J. Electroanal. Chem. 1989, 285, 1. (19) Deputy, A. L.; Wu, H.-P.; McCreery, R. L. J. Phys. Chem. 1990, 94, 3620. (20) Li, X.; Harrison, D. J. Anal. Chem. 1991, 63, 2168.
Analytical Chemistry, Vol. 77, No. 19, October 1, 2005 6205
been used to measure the concentration distribution, as a function of time and distance from a liquid/liquid interface.22 Electrochemical techniques have recently proven particularly powerful for observing concentration gradients at both electrode surfaces23-27 and liquid/liquid interfaces.4,28 Microelectrode probes have been utilized in either potentiometric27,29,30 or amperometric23-27,31,32 modes, to determine concentrations with spatial and temporal resolution in the micrometer and millisecond ranges, respectively. Ion-selective microelectrodes have been used in the substrate generation/tip collection mode of scanning electrochemical microscopy (SECM), to measure local concentrations of H+, NH4+, K+, and Zn2+, within the diffusion layer of an active substrate.30,33,34 Optical imaging of electrogenerated chemiluminescence (ecl) and fluorescence has found particular application in the visualization of pH gradients.35 These methods have also been used to measure current density distributions, highlighting the spatial heterogeneity of some electrode surface reactions.36-38 Both methods provide a high degree of spatial and temporal resolution, although ecl is limited in the number of systems that will luminesce with sufficient intensity to produce an image.39 On the other hand, fluorescence imaging has potentially wider scope, since the number of reagents that fluoresce or react with another species to produce a fluorescent derivative, is relatively high. Fluorescence microscopy has been used previously to identify active sites on a variety of electrode surfaces.40 Recently, realtime remote fluorescence imaging of two-dimensional concentration profiles was achieved using a multichannel fiber bundle.41 Compared to conventional fluorescence microscopy, confocal laser scanning microscopy (CLSM) reduces out-of-focus blur that would otherwise result from light being collected both above and below (21) O’Malley, R.; Vollmer, A.; Lee, J. R. I.; Harvey, I.; Headspith, J.; Diaz-Moreno, S.; Rayment, T. Electrochem. Commun. 2003, 5, 1. (22) Williams, R. J. P.; Wormald, M. R. J. Chem. Soc., Faraday Trans. 1991, 87, 1585. (23) Engstrom, R. C.; Weber, M.; Wunder, D. J.; Burgess, R.; Winquist, S. Anal. Chem. 1986, 58, 844. (24) Engstrom, R. C.; Meaney, T.; Tople, R.; Wightman, R. M. Anal. Chem. 1987, 59, 2005. (25) Baltes, N.; Thouin, L.; Amatore, C.; Heinze, J. Angew. Chem., Int. Ed. 2004, 43, 1431. (26) Amatore, C.; Pebay, C.; Scialdone, O.; Szunerits, S.; Thouin, L. Chem. Eur. J. 2001, 7, 2933. (27) Amatore, C.; Szunerits, S.; Thouin, L.; Warkocz, J.-S. Electroanalysis 2001, 13, 646. (28) Slevin, C. J.; Unwin, P. R. Langmuir 1997, 13, 4799. (29) Amatore, C.; Szunerits, S.; Thouin, L. Electrochem. Commun. 2002, 2, 248. (30) Wei, C.; Bard, A. J.; Nagy, G.; Toth, K. Anal. Chem. 1995, 67, 1346. (31) Amatore, C.; Szunerits, S.; Thouin, L.; Warkocz, J.-S. Electrochem. Commun. 2002, 2, 353. (32) Bath, B. D.; Lee, R. D.; White, H. S.; Scott, E. R. Anal. Chem. 1998, 70, 1047. (33) Horrocks, B. R.; Mirkin, M. V.; Pierce, D. T.; Bard, A. J.; Nagy, G.; Toth, K. Anal. Chem. 1993, 65, 1213. (34) Wei, C.; Bard, A. J.; Kapui, I.; Nagy, G.; Toth, K. Anal. Chem. 1996, 68, 2651. (35) Fiedler, S.; Hagedorn, R.; Schnelle, T.; Richter, E.; Wagner, B.; Fuhr, G. Anal. Chem. 1995, 67, 820. (36) Engstrom, R. C.; Johnson, K. W.; DesJarlais, S. Anal. Chem. 1987, 59, 670. (37) Bowling, R. J.; McCreery, R. L.; Pharr, C. M.; Engstrom, R. C. Anal. Chem. 1989, 61, 2763. (38) Vitt, J. E.; Engstrom, R. C. Anal. Chem. 1997, 69, 1070. (39) Engstrom, R. C.; Ghaffari, S.; Qu, H. Anal. Chem. 1992, 64, 2525. (40) Bowyer, W. J.; Xie, J.; Engstrom, R. C. Anal. Chem. 1996, 58, 2005. (41) Amatore, C.; Chovin, A.; Garrigue, P.; Servant, L.; Sojic, N.; Szunerits, S.; Thouin, L. Anal. Chem. 2004, 76, 7202.
6206
Analytical Chemistry, Vol. 77, No. 19, October 1, 2005
the focal plane of an object. CLSM allows direct, noninvasive, serial optical sectioning of objects and profiling of multilayer structures.42 Fluorescein, a common fluorophore widely used in CLSM, exhibits a pH-sensitive fluorescence signal, which increases in aqueous solution from pH ∼5 to a maximum as the pH rises above pH ∼7. For an electrochemical process that alters the pH within the range where the fluorescence intensity from fluorescein changes, it should be possible to obtain quantitative information on pH gradients and diffusion close to the electrode surface by capturing the resulting light intensity. In a recent paper,43 we reported preliminary data using this new approach to investigate the reduction of benzoquinone. In this paper, we carefully assess the quantitative aspects of CLSM as a probe of pH gradients through a series of experiments at electrodes of different geometry. Our aim is to show that the technique is quantitative and that CLSM can be used to visualize diffusion to ultramicroelectrodes (UMEs) and identify active sites on surfaces. We note that Heinze and co-workers used CLSM to probe pH gradients at a surface derivatized with an immobilized fluorophore, using an SECM probe to generate protons or hydroxide ions.44 However, there was no attempt in this study to analyze the pH data quantitatively. In addition to investigations at single UMEs, we have imaged pH profiles at platinum UME arrays. Despite advances in manufacturing techniques, problems remain in the fabrication of arrays that cause them to deviate from the predicted electrochemical response. Such defects include pinholes in the insulating layer over the conducting tracks leading to the electrodes and problems of adhesion between the tracks and the silicon wafer. The additional conducting sites that are exposed are often submicrometer in size and not easily visible using conventional optical microscopy. In this paper, we show that such defects are magnified under electrochemical control by the development of a proton diffusion field that can readily be visualized by CLSM. The systems considered herein are the reduction of water (under galvanostatic control) at platinum and gold electrodes and oxygen reduction (potentiostatic control) at platinum electrodes. The effective number of electrons, n, transferred during oxygen reduction in unbuffered aqueous solution has been the subject of debate. Work by Pletcher and Sotiropoulos,45 demonstrated a dependence of n on the mass transport rate, with a transition from a four-electron process to a two-electron process as the rate of mass transport increased. In a 0.1 M NaCl solution, n was 3.5 and 3.3 for 25- and 10-µm-diameter electrodes, respectively. Earlier work by Winlove et al.46 quoted an n value of 2.8 ( 0.4 for an electrode with an effective radius of 205 ( 20 µm, while Vilambi and Taylor47 found an n value of 2.7 using a rotating ring-disk electrode at a rotation rate of 400 rpm in alkaline medium. In our studies, we show that CLSM is a particularly sensitive indicator of the effective number of electrons transferred in this process. (42) Sheppard, C. J. R.; Shotton, D. M. Confocal Laser Scanning Microscopy; BIOS Scientific Publishers Ltd.: Oxford, U.K., 1997. (43) Cannan, S.; Macklam, I. D.; Unwin, P. R. Electrochem. Commun. 2002, 4, 886. (44) Boldt, F.-M.; Heinze, J.; Diez, M.; Petersen, J.; Bo¨rsch, M. Anal. Chem. 2004, 76, 3473. (45) Pletcher, D.; Sotiropoulos, S. J. Electroanal. Chem. 1993, 356, 109. (46) Winlove, C. P.; Parker, K. H.; Oxenham, R. K. C. J. Electroanal. Chem. 1984, 170, 293. (47) Vilambi, N. R. K.; Taylor, E. J. Electrochim. Acta 1989, 34, 1449.
Scheme 1. Summary of the Possible Routes of Oxygen Reduction
Figure 1. Field emission scanning electron microscopy image of a typical 370-nm ring electrode used for the water reduction experiments. The inner diameter of the optical fiber is 217 µm.
EXPERIMENTAL SECTION Substrate Electrodes. Platinum disk electrodes with diameters of 10 and 25 µm were fabricated as described previously.48 An 80-nm-diameter platinum disk and thin gold ring electrodes, fabricated as described below, were also used as substrate electrodes. In all cases, the insulating sheath surrounding the working electrode had a diameter ∼50 times that of the electrode, so that “back diffusion” would be negligible. Ring electrodes were fabricated by carefully removing an optical fiber (Newport) from its polyimide coating and sputtering gold onto it using an Edwards E306 vacuum coater (Moorfield). Each coated fiber was sealed in a 2-mm-o.d. pulled glass capillary by melting the glass quite rapidly around the fiber using the heating element from a vertical pipet puller (Narishighe PB7). The thin ring was exposed after polishing the electrode flat using a home-built polishing wheel. The polishing procedure involved using a succession of diamond-impregnated pads (from 6 to 0.1 µm, Buehler) followed by the use of a wet polishing pad with alumina (0.05 µm). An Olympus BH2 light microscope equipped with a 3-CCD video system was used to visually inspect the electrode surface and determine the approximate dimensions of the electrode. Field emission scanning electron microscopy was used to obtain detailed information on the topography of the ring electrodes and quantitatively measure the ring thickness. An example image of a typical electrode used is shown in Figure 1. The ring thickness varies a little ((30 nm), but the mean value of 370 nm was used in all simulations. The nanoscale disk electrode used in this work was fabricated using part of a procedure recently reported by White and co-workers.49 First, a length (2 cm) of 50-µm-diameter platinum wire was electrically connected to a larger copper wire using tin solder. The end of the platinum wire was electrochemically etched, in a 30% CaCl2 (sat.) solution, by applying a potential of ∼2 V ac.50,51 The sharpened wire was rinsed in acetone and water and was then left to dry at room temperature. The sharp end of a tapered borosilicate capillary was melted in a H2/O2 flame to yield a rounded tip in which the glass was ∼2 mm thick. The etched platinum wire was carefully cut from the metal wire and inserted into the sealed capillary, and the glass was then melted around the wire using the flame. An optical (48) Wightman, R. M.; Wipf, D. O. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1989; Vol. 15. (49) Zhang, B.; Zhang, Y.; White, H. S. Anal. Chem. 2004, 76, 6229. (50) Melmed, A. J. J. Vac. Sci. Technol., B 1991, 9, 601. (51) Melmed, A. J. J. Vac. Sci. Technol., A 1984, 2, 1388.
microscope was used to check the sealing, to ensure that no air bubbles were trapped near the wire. An electrical connection was made between the platinum wire and a larger copper wire using conductive silver epoxy. The thickness of the glass layer was reduced and flattened as much as possible by using silicon carbide paper (600 grit, Buehler) without exposing the electrode. At this stage, the glass was carefully polished by hand with waterproof silicon carbide paper (SIC Paper 4000, Struers) and the electrode exposure was monitored every few minutes by measuring the steady-state voltammetric response of the electrode for the oxidation of 5 mM ferrocene in acetonitrile containing 0.1 M tetrabutylammonium hexafluorophosphate (TBAPF6). Once the electrode was exposed, as evidenced by the development of a voltammetric response, it was carefully polished to a fine finish using a 0.05-µm alumina slurry and a definitive voltammogram was run from which the electrode dimensions could be determined. Two designs of individually addressable platinum UME arrays were examined: (i) a linear array of 16-addressable microdots, with no reference electrode and (ii) an array of 4 by 4 addressable microdots, with a reference electrode. In both designs, the dots had a diameter of 10 µm and pitch of 120 µm. The devices were designed and custom-made at the University of Warwick, as described elsewhere.52 For imaging experiments, the exposed bond pads and bonding wires were insulated by a covering of Araldite (Ciba-Geigy, Cambridge, U.K.), which was built up to a height of ∼2 mm to form a small well surrounding the microdots. This was sufficient to contain the aqueous solution. Once hardened, the epoxy was coated with curing silicone rubber and an adhesive sealant (both Dow Corning) to form a watertight seal. The complete device was then attached to a glass microscope slide with double-sided tape. Materials. The reduction of water was carried out in a solution containing 10 µM disodium fluorescein (Sigma) with 0.1 M potassium nitrate (Fisher) as the supporting electrolyte, while oxygen reduction was carried out in a solution of 5 µM disodium fluorescein with 0.1 M potassium chloride (AnalaR, BDH) as the supporting electrolyte. The reduction of benzoquinone was carried out in a 1 mM solution of benzoquinone (Aldrich) with 8 µM disodium fluorescein and 0.1 M potassium chloride. The initial pH was adjusted by the addition of 0.1 M hydrochloric acid (Sigma) and measured before commencing the imaging experiments. CLSM Measurements. CLSM imaging of concentration gradients at single electrodes was carried out in a circular trough (home-built), which was placed on the stage of a Zeiss LSM 510, Axioplan 2, confocal microscope, with the condenser and glass stage plate removed to allow positioning of the electrode. After polishing, each electrode was inserted in an inverted configuration, (52) Barker, A. L.; Unwin, P. R.; Gardner, J. W.; Rieley, H. Electrochem. Commun. 2004, 6, 91.
Analytical Chemistry, Vol. 77, No. 19, October 1, 2005
6207
Figure 2. Meshes used in the modeling of the (a) disk and (b) ring systems.
through a hole drilled in the bottom of the trough and sealed in place with wax. This experimental arrangement was described in a previous paper.43 For all measurements, the electrode surface was submerged in solution and viewed through a water immersion objective lens (Zeiss, Achroplan 20x/0.50W). To achieve the optimum compromise between resolution and image intensity, the confocal aperture was usually set to give an optical slice of
