Patterned Arrays of Vertically Aligned Carbon ... - ACS Publications

canterbury.ac.nz. Phone: +64-3-364501. Fax: +64-3-36421100. † University of Canterbury. ‡ Christchurch Polytechnic Institute of Technology. (1) Wi...
0 downloads 0 Views 496KB Size
Anal. Chem. 2008, 80, 8835–8839

Patterned Arrays of Vertically Aligned Carbon Nanotube Microelectrodes on Carbon Films Prepared by Thermal Chemical Vapor Deposition Xianming Liu,† Keith H. R. Baronian,‡ and Alison J. Downard*,† MacDiarmid Institute for Advanced Materials and Nanotechnology, Department of Chemistry, University of Canterbury, Private Bag 4800, Christchurch 8140, New Zealand, and Christchurch Polytechnic Institute of Technology, P.O. Box 540, Christchurch 8140, New Zealand A straightforward procedure is described for preparation of arrays of microdisk electrodes comprising bundles of vertically aligned carbon nanotubes (VACNTs). The arrays are fabricated by thermal chemical vapor deposition synthesis directly on a planar carbon film support. Use of standard micro- and nanolithography procedures for patterning the bilayer catalyst spots enables arrays to be grown with controlled electrode diameters and spacings. The minimum accessible VACNT bundle diameter, and hence microelectrode diameter, is 2 µm. After insulating the arrays with SU-8 epoxy and exposing the VACNT ends by polishing or treating with O2 plasma, the microdisk electrodes exhibit attractive electrochemical properties. Micro- and nanoelectrodes have some significant advantages over macro-scale electrodes for electroanalytical applications. Their attractive properties include high rates of mass transport and an associated rapid approach to steady state, improved signal-tobackground response, and reduced ohmic potential drop.1 Vertically aligned carbon nanotubes (VACNTs) and carbon nanofibers (VACNFs) are excellent candidates for the fabrication of carbonbased micro- and nanoelectrodes. Multiwalled CNTs and CNFs have metallic conductivity, and when vertically aligned with open ends, fast rates of electron transfer with redox species are obtained.2-6 Furthermore, the relatively high density of terminating carboxylate groups at the open tube ends provides convenient routes for immobilizing species of interest.7-12 * To whom correspondence should be addressed. E-mail: Alison.downard@ canterbury.ac.nz. Phone: +64-3-364501. Fax: +64-3-36421100. † University of Canterbury. ‡ Christchurch Polytechnic Institute of Technology. (1) Wightman, R. M.; Wipf, D. O. Electroanal. Chem. 1989, 15, 267–353. (2) Nugent, J. M.; Santhanam, K. S. V.; Rubio, A.; Ajayan, P. M. Nano Lett. 2001, 1, 87–91. (3) Banks, C. E.; Moore, R. R.; Davies, T. J.; Compton, R. G. Chem. Commun. 2004, 1804–1805. (4) Moore, R. R.; Banks, C. E.; Compton, R. G. Anal. Chem. 2004, 76, 2677– 2682. (5) Banks, C. E.; Davies, T. J.; Wildgoose, G. G.; Compton, R. G. Chem. Commun. 2005, 829–841. (6) Chou, A.; Bocking, T.; Singh, N. K.; Gooding, J. J. Chem. Commun. 2005, 842–844. (7) Guiseppi-Elie, A.; Lei, C. H.; Baughman, R. H. Nanotechnology 2002, 13, 559–564. (8) Nguyen, C. V.; Delzeit, L.; Cassell, A. M.; Li, J.; Han, J.; Meyyappan, M. Nano Lett. 2002, 2, 1079–1081. 10.1021/ac801552a CCC: $40.75  2008 American Chemical Society Published on Web 10/24/2008

A major advantage of VACNTs and VACNFs for fabrication of micro- and nanoelectrodes is that, by coupling chemical vapor deposition (CVD) synthesis with standard micro- and nanolithography procedures for patterning the catalysts required for CNT growth, VACNTs and VACNFs can be synthesized, with high repeatability, at precisely defined locations on the support material. Importantly, the same synthesis and fabrication strategies (but with minor changes to the lithography procedure) can be applied to prepare individual micro- or nanoelectrodes, individually addressable electrodes in an array format, arrays of nonindividually addressable independent electrodes, and arrays of electrodes with overlapping diffusion fields. Use of standard nano- and microfabrication procedures also generates electrodes with the potential to be combined with integrated circuit technology for scaling up and fabricating sophisticated electrical and electronic devices. These advantages do not apply, for example, to the preparation of microelectrodes using free-standing carbon fibers. Several groups have demonstrated the power of plasmaenhanced CVD (PE-CVD) for preparing micro- and nanoelectrodes from VACNTs and VACNFs.13-15 The simplest of these approaches is to synthesize the VACNTs or VACNFs directly on a conducting substrate, which forms the electrical connection to the micro- or nanoelectrodes. Insulation of the substrate is required, and additionally, insulation of the VACNT or VACNF sidewalls, or a significant length of the sidewalls, has generally been found to lead to an improved response in voltammetric measurements.13–15 Although excellent results have been obtained for micro- and nanoelectrode preparation using PE-CVD, thermal CVD is a much more accessible method for CNT growth, requiring only a basic tube furnace and gas manifold. Additionally, thermal CVD has (9) Gooding, J. J.; Wibowo, R.; Liu, J. Q.; Yang, W. R.; Losic, D.; Orbons, S.; Mearns, F. J.; Shapter, J. G.; Hibbert, D. B. J. Am. Chem. Soc. 2003, 125, 9006–9007. (10) Sotiropoulou, S.; Chaniotakis, N. A. Anal. Bioanal. Chem. 2003, 375, 103– 105. (11) Yu, X.; Chattopadhyay, D.; Galeska, I.; Papadimitrakopoulos, F.; Rusling, J. F. Electrochem. Commun. 2003, 5, 408–411. (12) Patolsky, F.; Weizmann, Y.; Willner, I. Angew. Chem., Int. Ed. 2004, 43, 2113–2117. (13) Guillorn, M. A.; McKnight, T. E.; Melechko, A.; Merkulov, V. I.; Britt, P. F.; Austin, D. W.; Lowndes, D. H.; Simpson, M. L. J. Appl. Phys. 2002, 91, 3824–3828. (14) Tu, Y.; Lin, Y.; Ren, Z. F. Nano Lett. 2003, 3, 107–109. (15) Koehne, J.; Li, J.; Cassell, A. M.; Chen, H.; Ye, Q.; Ng, H. T.; Han, J.; Meyyappan, M. J. Mater. Chem. 2004, 14, 676–684.

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8835

Table 1. Layout of Catalyst Microdisk Patterns and Patterning Method disk diameter (µm)

patterning method

number of disks

minimum spacing between disks (µm)

100 10 2

shadow mask EBL EBL

16 77 400

900 155 50

advantages in terms of sample sizes and numbers of samples that can be prepared per run. However, few approaches to microelectrode array fabrication using thermal CVD synthesis of VACNTs have been reported. This is undoubtedly due, in part, to the greater difficulty of obtaining vertical alignment of CNT growth without the assistance of an electric field, as is present in PECVD. For thermal CVD, vertical alignment of CNTs relies on the high density of growing CNTs, which gives a crowding effect and maximizes van der Waal’s interactions between the sidewalls of neighboring CNTs.16,17 Clearly, as the CNT bundle diameter decreases, these factors will have less influence. Shulz and coworkers have developed a simple strategy for preparing nanoelectrode arrays by thermal CVD.18 Their approach is to grow relatively large diameter, tall “towers” of high-density VACNTs, which can be peeled off the silicon substrate, embedded into epoxy, and polished at both ends. Electrical connection is made to one end and the other functions as the electrode surface. By controlling the extent of polishing, only some CNT tips are exposed and the electrodes function as random arrays of nanoelectrodes. While the simplicity of this approach is very attractive, the lack of control of individual electrode dimensions and spacing is a limitation. Xu and co-workers have demonstrated that use of an alumina template leads to growth of highly ordered arrays of nanoscale VACNTs by thermal CVD.19 The arrays have been shown to promote fast electron transfer with immobilized glucose oxidase;20 however, a drawback of the templating method is that the dimensions and spacing of the alumina template pores cannot be independently varied and only closely spaced (relative to their diffusion fields) electrodes can be prepared by this method. In a third example, Wang and co-workers fabricated arrays of VACNT pillars by thermal CVD and applied them to the electrical stimulation of nerve cells.21 Standard UV photolithography was used to create individually addressable microelectrodes and to define the catalyst spot size and position. Although details are not reported, the minimum electrode diameter appears to be 30 µm. To our knowledge, there have been no reports detailing the fabrication and characterization of low-diameter (e10 µm) microelectrode arrays, with controllable electrode dimensions and (16) Fan, S. S.; Chapline, M. G.; Franklin, N. R.; Tombler, T. W.; Cassell, A. M.; Dai, H. J. Science 1999, 283, 512–514. (17) Li, D.-C.; Dai, L.; Huang, S.; Mau, A. W. H.; Wang, Z. L. Chem. Phys. Lett. 2000, 316, 349–355. (18) Yun, Y. H.; Shanov, V.; Schulz, M. J.; Dong, Z. Y.; Jazieh, A.; Heineman, W. R.; Halsall, H. B.; Wong, D. K. Y.; Bange, A.; Tu, Y.; Subramaniam, S. Sens. Actuators, B 2006, 120, 298–304. (19) Li, J.; Papadopoulos, C.; Xu, J. M.; Moskovits, M. Appl. Phys. Lett. 1999, 75, 367–369. (20) Withey, G. D.; Lazareck, A. D.; Tzolov, M. B.; Yin, A.; Aich, P.; Yeh, J. I.; Xu, J. M. Biosens. Bioelectron. 2006, 21, 1560–1565. (21) Wang, K.; Fishman, H. A.; Dai, H.; Harris, J. S. Nano Lett. 2006, 6, 2043– 2048.

8836

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

Figure 1. SEM images of patterned VACNT pillars grown using thermal CVD: (a) array of 16 × 100 µm pillars; (b) single 100-µm pillar; (c) array of 77 × 10 µm pillars; (d) single 10-µm pillar; (e) array of 400 × 2 µm pillars; (f) single 2-µm pillar.

electrode spacing, by thermal CVD. Herein we describe the fabrication of arrays of 100-, 10-, and 2-µm-diameter VACNT electrodes, using standard lithography techniques and optimized thermal CVD conditions. The VACNT pillars are grown directly on the conducting carbon film substrate22 and, after insulation with epoxy, are characterized electrochemically. To demonstrate the fabrication procedure, we have prepared arrays of nonindividually addressable, independently functioning electrodes. Preparation of other microelectrode formats requires only minor changes to the lithography step. EXPERIMENTAL SECTION Fabrication of Patterned Arrays of VACNT Pillars. The preparation of pyrolyzed photoresist film (PPF) (samples ∼15 mm × 15 mm) has been described previously.23 PPF samples had sheet resistances of >30 Ω/sq and average surface roughness of less than 0.6 nm. Bilayer catalyst layers were deposited on PPF samples by evaporating 5-nm Al followed by 5-nm Fe using an Edwards 3000 e-beam evaporator. Patterns of catalyst microdisks were defined using either stainless steel shadow masks22 or electron beam lithography (EBL, Raith 150). The patterned areas were near the centers of the PPF samples. The details of the patterns are listed in Table 1 and Figure S1 shows the scheme for patterning of catalysts by EBL. Note that catalyst patterns can also be prepared using standard optical lithography procedures; however due to instrument access considerations, EBL was employed in the present work. Patterned substrates were placed in a quartz tube and heated at 750 °C under an atmosphere of Ar (1200 sccm) and H2 (800 sccm) for 30 min. For preparation of 10- and 2-µm-diameter VACNT arrays, H2 was switched off and the tube flushed with Ar (22) Liu, X.; Baronian, K. H. R.; Downard, A. J. Carbon, submitted for publication. (23) Brooksby, P. A.; Downard, A. J. Langmuir 2004, 20, 5038–5045.

Figure 2. SEM images of VACNT microdisk electrode arrays, insulated with SU-8 epoxy. (a) 77 × 10 µm array and (b) single 10µm electrode, after polishing; (c) 77 × 10 µm array and (d) single 10-µm electrode, after treatment with O2 plasma; (e) 400 × 2 µm array and (f) single 2-µm electrode after polishing; (g) 400 × 2 µm array and (h) single 2-µm electrode, after treatment with O2 plasma.

Table 2. Voltammetric data (scan rate 100 mV s-1) for 1 mM FcOH in 0.1 M KCl array 16 × 100 µm 77 × 10 µm 77 × 10 µm 400 × 2 µm 400 × 2 µm b

treatment polished polished O2 plasma polished O2 plasma

ilim (nA) b

190 141 178 157 152

ical (nA)a

E3/4 - E1/4 (mV)

216 104 104 108 108

61 63 67 72 70

a Calculated from eq 1, assuming D ) 7.0 × 10-6 cm2 s-1 Voltammogram recorded with scan rate, 5 mV s-1.

(2000 sccm) for 5 min. Ethylene (400 sccm) was carried into the reaction chamber at 750 °C by Ar (900 sccm). Arrays of 100-µmdiameter VACNT pillars were prepared at 750 °C in an atmosphere of ethylene (400 sccm), Ar (600 sccm), and H2 (400 sccm). Additionally, for these arrays, a small amount of water vapor was introduced into the reaction chamber by bubbling 100 sccm Ar through water prior to feeding into the furnace.22 After 10-30 min CVD, the samples were cooled to below 100 °C in Ar (2000 sccm) before removal from the furnace. Fabrication of VACNT Microdisk Electrode Arrays. To establish electrical connection to an electrode array, a Cu strip was attached with silver paste to the PPF surface, at a position distant from the VACNT pattern. The connection was covered with

Figure 3. Steady-state voltammograms of 1 mM FcOH in 0.1 M KCl obtained at (a) polished 16 × 100 µm array; (b) polished 77 × 10 µm array; (c) O2 plasma-treated 77 × 10 µm array; (d) polished 400 × 2 µm array; and (e) O2 plasma-treated 400 × 2 µm array. a) Scan rate: (a) 5; (b)-(e) 100 mV s-1. All scale bars represent 20 nA.

epoxy and the sample was cured at 60 °C overnight. UV-curable SU-8 epoxy (Microchemie) was used to insulate the VACNT pillars. Epoxy types 2025 and 2005 were spin-coated onto 100-µm and sub-100-µm-diameter VACNT pillar patterns, respectively. Prior to this step, the heights of the VACNT patterns were determined using scanning electron microscopy (SEM, Leica S440) and the spin rates of SU-8 were selected (using manufacturer’s data) to ensure that the layer thickness approximated that of the VACNT pillar height. After spin coating, the samples were soft-baked on a hot plate (1-3 min at 65 °C and 2-5 min at 95 °C (the time depended on the SU-8 layer thickness)), UV cross-linked (Karl Suss Mask Aligner, 3.3 mW cm-2) and finally hard-baked (2-5 min at 65 °C and 3-7 min at 95 °C). SU-8 residues were removed from the tops of the VACNT pillars either by polishing (1-µm Al2O3 slurry on a Leco polishing cloth) or by O2 plasma etching (0.1 Torr, 1 sccm, 500 W, 7 min). The plasma etching rate on SU-8 2005 was 50-60 nm min-1. Polished arrays were cleaned by ultrasonication in deionized water for 10 min and dried Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8837

with N2 gas. SEM images of insulated microdisk arrays were obtained after sputter-coating with gold (5 nm). Electrochemistry. Electrochemical characterization of VACNT microdisk array electrodes utilized instrumentation and a cell described previously.22 The microdisk array was exposed to the solution by positioning beneath a hole in the bottom of the cell. The auxiliary and reference electrodes were a Pt wire and a saturated calomel electrode, respectively. The redox probe was hydroxymethylferrocene (FcOH, 1 mM) in 0.1 M aqueous KCl. All measurements were conducted within a Faraday cage. RESULTS AND DISCUSSION Patterned Growth of VACNT Pillars on PPF. Figure 1 shows SEM images of VACNTs pillar arrays, and of individual pillars, on PPF. The layout of the arrays of catalyst spots was designed to ensure that the resulting microdisk arrays would function as a collection of individual electrodes, without significant overlap of diffusion fields.24 The patterns of VACNT pillars faithfully follow those of the catalyst bilayers and the bundles of CNTs are highly aligned and, in most cases, are normal to the PPF surface. TEM images (not shown) reveal that the CNTs are multiwalled with diameters of 10-20 nm.22 The pillars adhere well to the PPF during subsequent insulation steps, but 10- and 2-µmdiameter pillars were not sufficiently robust, without insulation, to be used routinely for electrochemical measurements. The morphology and sheet resistance of PPF were unaffected by the conditions used for CNT growth by thermal CVD. Importantly, there is low contact resistance (51 Ω) between the top surface of the VACNT pillars and the substrate. The CVD procedure developed in this work for synthesis of 10- and 2-µm-diameter VACNT pillars is different from that used for preparation of 100-µm-diameter pillars. The latter procedure is the same as that previously used for growth of 2-mm-diameter pillars of VACNTs on PPF and glassy carbon.22 For the larger VACNT pillars, the synthesis gases included H2 and water vapor, which accelerate VACNT growth and reduce the formation of amorphous carbon.25,26 However, for CVD synthesis using catalyst spots with diameters less than 100 µm, we found that these gases prevented the growth of VACNT pillars; only short, nonaligned CNTs resulted from CVD in the presence of H2 and water vapor (together or individually). The beneficial effects of these gases on the growth of large (g100 µm) diameter pillars, result, in part, from etching the growing nanotubes.25,26 When synthesizing smaller diameter pillars, it is assumed that the larger pillar surface area/volume ratio, and the relatively increased access of gases within the pillars, cause excessive etching of the CNTs. Using the CVD procedures described in the Experimental Section, 100µm-diameter pillars could be grown with lengths up to 400 µm and smaller diameter pillars to lengths of between 15 and 35 µm. The smallest diameter of VACNT pillars achieved was 2 µm; below this size, the vertical alignment was variable with either aligned bundles adopting various orientations to the surface or formation of spots of spaghetti-like CNTs. (24) Davies, T. J.; Compton, R. G. J. Electroanal. Chem. 2005, 585, 63–82. (25) Hata, K.; Futaba, D. N.; Mizuno, K.; Namai, T.; Yumura, M.; Iijima, S. Science 2004, 306, 1362–1364. (26) Zhang, G. Y.; Mann, D.; Zhang, L.; Javey, A.; Li, Y. M.; Yenilmez, E.; Wang, Q.; McVittie, J. P.; Nishi, Y.; Gibbons, J.; Dai, H. J. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 16141–16145.

8838

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

Fabrication of VACNT Microdisk Electrode Arrays. As shown in our earlier work,22 and by others,13–15 coating the sidewalls of VACNTs with an insulating material improves their electrochemical characteristics. For the sub-100-µm-diameter pillars, embedding in epoxy also provides essential mechanical stability to the electrode array. The spin coating and curing procedure described in the Experimental Section results in a layer of epoxy from which the VACNT pillars protrude to a height of 1-2 µm. A thin layer of epoxy coats the sidewalls and tops of the protruding pillars. Examination of cross-sectional slices showed that the thickness of SU-8 on the sidewalls of the protruding VACNT pillars was typically 75, 32, and 14 µm for 100-, 10-, and 2-µm-diameter pillars, respectively. Exposure of the VACNT tips could be achieved by polishing or by O2 plasma treatment (7 min). Figure 2 shows SEM images of arrays of 10- and 2-µm-diameter microdisk electrodes and higher magnification images of individual members of the arrays, after both polishing and O2 plasma treatment. Although all samples were coated with a thin layer of gold prior to imaging, the image quality is poor, due to the large surface area of insulating epoxy relative to the surface area of the microelectrodes. The plasma-treated arrays appear cleaner than those polished with alumina; we presume that polishing debris, not removed by ultrasonication, accounts for the extraneous material on the polished arrays. Despite the somewhat different appearance of arrays treated by the two methods, polishing and O2 plasma treatment led to similar electrochemical behavior (see below); polishing was experimentally more convenient and hence was adopted as the standard procedure. When plasma treatment was applied for shorter times (less than 7 min), SU-8 was removed from just the center of VACNT pillars resulting in active electrodes with submicrometer diameters (not shown). We have not yet been able to reproducibly prepare electrodes of a given diameter by this route, but are continuing to investigate the procedure. During fabrication of the VACNT array using EBL, a thin (30 nm) layer of SiO2 is deposited on the PPF surface (see Supporting Information, Figure S1). The SiO2 layer is necessary for accurate patterning, as described in the Supporting Information. However, the layer was also found to significantly contribute to insulation of the VACNT arrays. In preliminary experiments using photolithography to pattern catalyst spots, VACNT arrays insulated with SU-8 epoxy only showed large background currents compared with those prepared by EBL with a SiO2 layer. Hence, initial deposition of a SiO2 layer should also be included when preparing arrays using photolithography. Electrochemical Characterization. Table 2 and Figure 3 show representative voltammetric data for 1 mM FcOH in 0.1 M KCl, obtained at various VACNT microelectrode arrays. As expected, based on the layout of the microdisk arrays, voltammograms show very well-defined, steady-state responses consistent with each electrode functioning independently with negligible overlap of neighboring diffusion fields.27 There were no systematic differences between arrays that were polished or O2 plasma-treated before use. The voltammogram shown in Figure 3a was recorded at an array of 100-µm-diameter electrodes using a scan rate of 5 mV s-1. As the scan rate was increased above 10 mV s-1, (27) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.; John Wiley: New York, 2001; pp 183-184.

voltammograms were increasingly peaked, indicating an increasing contribution of linear diffusion. On the other hand, arrays of 2- and 10-µm-diameter electrodes gave steady-state voltammograms at a scan rate of 100 mV s-1 (Figure 3b-e). These observations are consistent with the diameters of the microdisks.28 Under steady-state conditions, the measured E3/4 - E1/4 values for all arrays approach those expected for an electrochemically reversible process (59 mV)27 and the signal-to-background ratios compare favorably with those typically reported for VACNT arrays.13–15 Table 2 includes the predicted limiting current, ilim, for each array, calculated using eq 1,29 where n is the number of electrons, N is the number of microdisks in the array, F is Faraday′s constant, D is the diffusion coefficient of FcOH (an average value of 7.0 × 10-6 cm2 s-1 was assumed, based on the range of values reported in the literature),30-33 r is the radius of an individual microdisk, and CFcOH is the concentration of FcOH. ilim ) 4nNFrDCFcOH

(1)

With the exception of the polished array of 100-µm-diameter electrodes, the experimental limiting currents exceed the calculated currents. Higher than predicted currents can arise when individual microelectrodes stand proud of the insulating material, thus increasing the electrode surface area. Another possibility is that some electrodes may have an elliptical, rather than circular, cross section exposed to the solution. This situation occurs when (28) Reference 27, p 232. (29) Reference 27, p 174. (30) Anicet, N.; Bourdillon, C.; Moiroux, J.; Saveant, J. M. J. Phys. Chem. B 1998, 102, 9844–9849. (31) Liljeroth, P.; Johans, C.; Slevin, C. J.; Quinn, B. M.; Kontturi, K. Electrochem. Commun. 2002, 4, 67–71. (32) Longinotti, M. P.; Corti, H. R. Electrochem. Commun. 2007, 9, 1444–1450. (33) Miao, W. J.; Ding, Z. F.; Bard, A. J. J. Phys. Chem. B 2002, 106, 1392– 1398.

nanotube pillars are not aligned perpendicular to the substrate or when the polished or plasma-treated surface is not parallel to the substrate. The origin of the smaller than predicted limiting current for the polished array of 100-µm-diameter electrodes is not clear. Further polishing did not increase the limiting current suggesting the VACNT tips are fully exposed to solution. In summary, we have demonstrated that arrays of microdisk electrodes can be conveniently fabricated by thermal CVD synthesis of bundles of VACNTs directly on a planar carbon film support. Arrays are grown with controlled electrode spacings and diameters (down to 2 µm). The procedures are suitable for mass production of electrode arrays and for integration of arrays into larger devices. Only small changes to the lithography steps are required to produce individually addressable microelectrodes. Microdisk electrodes fabricated from VACNTs are expected to combine the electroanalytical advantages of microelectrodes with the attractive properties of VACNTs, which typically give fast rates of electron transfer for redox probes and a high concentration of surface carboxylate groups for facile electrode modification. ACKNOWLEDGMENT The authors thank Helen Devereux and Gary Turner (University of Canterbury) for assistance with microfabrication. The research is funded by the RSNZ Marsden Fund (contract UOC 0605), the MacDiarmid Institute for Advanced Materials and Nanotechnology, and the University of Canterbury. SUPPORTING INFORMATION AVAILABLE Description of the procedure for patterning the PPF substrate with catalyst spots using EBL. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review July 23, 2008. Accepted September 8, 2008. AC801552A

Analytical Chemistry, Vol. 80, No. 22, November 15, 2008

8839