Probing Morphological and Compositional ... - ACS Publications

This paper demonstrates the first application of tapping-mode scanning force microscopy (TM SFM) in the compositional mapping of modified glassy carbo...
0 downloads 0 Views 129KB Size
Anal. Chem. 1999, 71, 4306-4312

Probing Morphological and Compositional Variations of Anodized Carbon Electrodes with Tapping-Mode Scanning Force Microscopy Gregory K. Kiema,† Glen Fitzpatrick,‡ and Mark T. McDermott*,†

Department of Chemistry, University of Alberta, Edmonton, AB T6G 2G2, and The Alberta Microelectronics Center, Edmonton, AB T6G 2T9, Canada

This paper demonstrates the first application of tappingmode scanning force microscopy (TM SFM) in the compositional mapping of modified glassy carbon (GC) electrodes. Using TM SFM, we have been able to track both compositional and topographical changes of polished GC induced by electrochemical pretreatment (ECP). Photoresist-based microfabrication techniques were employed to produce surfaces consisting of segregated modified and unmodified regions for direct comparison in the same image. Our results show that ECP of GC via anodization in basic solutions for short times (∼10 s) initially removes the ubiquitous layer of polishing debris via an etching process. Longer anodization in basic electrolyte results in significant etching of the GC surface. ECP in acidic solutions yields little topographic change compared to basic electrolytes. Electrochemical results obtained for three redox systems studied on both modified and unmodified GC electrodes correlate with the TM SFM images collected.

One of the more exciting advances in SPM methodology has been the development of scanning force microscopy (SFM) for the compositional mapping of surfaces at nanometer length scales.4-13 The first report of this ability showed that contact-mode SFM, operating in lateral (or friction) force mode, can distinguish between different chemical domains.4 These studies were extended to employ SFM probe tips of controlled chemistry which generated images with contrast based on predictable chemical interactions.5,7,9,11-13 It has recently been shown that tapping-mode SFM (TM SFM) can also be utilized to map surfaces compositionally.14-21 Contrast based on surface composition is derived in TM SFM by measuring the phase shift of the oscillating cantilever, which is sensitive to variations in tip-surface interactions. Thus, differences in surface mechanical properties and chemistry can be mapped. We noted that phase contrast TM SFM is more efficient in generating compositional maps of rough surfaces compared to contact-mode SFM because topographically induced shear forces are minimized.19 In this paper, we extend our previous work19 by applying TM SFM to track compositional changes of glassy carbon (GC)

The ability to rationally design electroanalytical devices hinges largely on the capacity to characterize interfaces compositionally. In many cases, the development of an analytical electrode involves modification of its surface at the molecular level.1-3 A powerful approach for the characterization of modified electrodes involves combining techniques that probe surface architecture at both macroscopic and microscopic length scales. Traditionally, electrochemical techniques that measure electron-transfer rates, background current, and adsorption as well as spectroscopic methodologies have been employed to provide a spatially averaged picture of electrode surface structure. More recently, the ultrahigh resolution of scanning probe microscopy (SPM) has been exploited to provide molecular-scale descriptions.3

(4) Overney, R. M.; Meyer, E.; Frommer, J.; Brodbeck, D.; Lu ¨ thi, R.; Howard, L.; Gu ¨ ntherodt, H.-J.; Fujihira, M.; Takano, H.; Gotoh, Y. Nature 1992, 359, 133. (5) Frisbie, C. D.; Rozsnyai, L. F.; Noy, A.; Wrighton, M. S.; Lieber, C. M. Science 1994, 265, 2071. (6) Wilbur, J. L.; Biebuyck, H. A.; MacDonald, J. C.; Whitesides, G. M. Langmuir 1995, 11, 825. (7) Green, J. D.; McDermott, M. T.; Porter, M. D. J. Phys. Chem. 1995, 99, 10960-10965. (8) Han, T.; Williams, J. M.; Beebe, T. P. J. Anal. Chim. Acta 1995, 307, 361. (9) Noy, A.; Frisbie, C. D.; Razsnyai, L. F.; Wrighton, M. S.; Lieber, C. M. J. Am. Chem. Soc. 1995, 117, 7943. (10) Green, J.-B.; McDermott, M. T.; Porter, M. D. J. Phys. Chem. 1996, 100, 13342. (11) Sinnah, S. K.; Steel, A. B.; Miller, J. C.; Reutt-Robey, J. E. J. Am. Chem. Soc. 1996, 118, 8925. (12) van der Vegte, E. W.; Hadziioannou, G. J. Phys. Chem. B 1997, 101, 9563. (13) Vezenov, D. I.; Noy, A.; Rozsnyai, L. F.; Lieber, C. M. J. Am. Chem. Soc. 1997, 119, 2006. (14) Chernoff, D. A. Proc. Microscop. Microanal. 1995, 888. (15) Howard, A. J.; Rye, R. R.; Houston, J. E. J. Appl. Phys. 1996, 79, 1885. (16) Lecle`re, P.; Lazzaroni, R.; Bre´das, J. L.; Yu, J. M.; Dubois, P.; Je´roˆme, R. Langmuir 1996, 12, 4317. (17) Viswanathan, R.; Tian, J.; Marr, D. W. Langmuir 1997, 13, 1840. (18) Magonov, S. N.; Elings, V.; Papkov, V. S. Polymer 1997, 38, 297. (19) Finot, M. O.; McDermott, M. T. J. Am. Chem. Soc. 1997, 119, 8564. (20) Noy, A.; Sanders, C. H.; Vezenov, D. V.; Wong, S. S.; Lieber, C. M. Langmuir 1998, 14, 1508-1511. (21) Sasaki, K.; Koike, Y.; Azehara, H.; Hokari, H.; Fujihara, M. Appl. Phys. A 1998, 66, S1275-S1277.



Department of Chemistry, University of Alberta. The Alberta Microelectronics Center. * Corresponding author: (voice) 780-492-3687; (fax) 780-492-8231; (e-mail) [email protected]. (1) Bard, A. J.; Abruna, H. D.; Chidsey, C. E.; Faulkner, L. R.; Feldberg, S. W.; Itaya, K.; Majda, M.; Melroy, O.; Murray, R. W.; Porter, M. D.; Soriaga, M. P.; White, H. S. J. Phys. Chem. 1993, 97, 7147-7173. (2) Zhong, C.-J.; Porter, M. D. Anal. Chem. 1995, 67, 709A-715A. (3) Green, J.-B. D.; McDermott, C. A.; McDermott, M. T.; Porter, M. D. Imaging of Surfaces and Interfaces. Frontiers of Electrochemistry; Wiley-VHC Inc.: New York, 1999; Vol. 5. ‡

4306 Analytical Chemistry, Vol. 71, No. 19, October 1, 1999

10.1021/ac9904056 CCC: $18.00

© 1999 American Chemical Society Published on Web 08/31/1999

electrode surfaces induced by chemical modification. GC is a commonly utilized material in electrochemistry due to its importance in electrocatalysis,22 in electroanalysis,23 and in biological sensing.24-29 We correlate here electrochemical measurements and TM SFM images that track both topographic and compositional variations in GC induced by electrochemical pretreatment (ECP) in both acidic and basic media. ECP has been widely employed to activate the surface of carbon electrodes toward a variety of redox systems,30-37 and a variety of techniques have been applied to characterize the resultant surface.34,38-43 SPM techniques have been employed to probe the surface structure of ECP GC. Results of these studies mainly revealed topographic and roughness changes due to the pretreatment.44,45 The present work is focused on applying an alternate contrast mechanism in SPM to evaluate compositional changes and to assess the ability of TM SFM to perform this analysis. Importantly, phase contrast TM SFM is able to differentiate between oxidized GC and the original polished surface on patterned electrodes. To our knowledge, this report describes the first application of TM SFM for the compositional mapping of relatively rough, polished electrode surfaces. EXPERIMENTAL SECTION Reagents. Ru(NH3)63+(aq) in 1 M KCI and Fe(CN)64-(aq) in 1 M KCI solutions were prepared at 1 mM concentrations from Ru(NH3)6CI3 (Strem Chemicals) and K4Fe(CN)6 (BDH Chemicals), respectively. Eu3+(aq) in 0.2 M NaCIO4 solution was prepared at 5 mM concentrations from Eu(NO3)3‚5H2O (Aldrich). All reagents were used as received. Aqueous solutions were prepared using distilled/deionized (18 MΩ/cm) water (Barnstead, (22) Kinoshita, K. Carbon, Electochemical and Physicochemical Properties; John Wiley and Sons, Inc.: New York, 1988. (23) McCreery, R. L. In Electroanalytical Chemistry; McCreery, R. L., Ed.; Marcel Dekker: New York, 1991; Vol. 17, pp 221-374. (24) Ueda, C.; Tse, D. C.-S.; Kuwana, T. Anal. Chem. 1982, 54, 850-856. (25) Lau, A. N.; Miller, L. L. J. Am. Chem. Soc. 1983, 105, 5271-5277. (26) Sonawat, H. M.; Phadke, R. S.; Govil, G. Biotechnol. Bioeng. 1984, 26, 10661070. (27) Ye, J.; Baldwin, R. P. Anal. Chem. 1988, 60, 2263-2268. (28) Wightman, R. M.; May, L. J.; Michael, A. C. Anal. Chem. 1988, 60, 769A779A. (29) Dontha, N.; Nowall, W. B.; Kuhr, W. G. Anal. Chem. 1997, 69, 2619-2625. (30) Tse, D.; Kuwana, T. Anal. Chem. 1978, 50, 1315. (31) Engstrom, R. C. Anal. Chem. 1982, 54, 2310-2314. (32) Cabaniss, G. E.; Diamantis, A. A.; Murphy, W. R.; Linton, R. W.; Meyer, T. J. J. Am. Chem. Soc. 1985, 107, 1845-1853. (33) Beilby, A. L.; Carlsson, A. J. Electroanal. Chem. 1988, 248, 283-304. (34) Kepley, L. J.; Bard, A. J. Anal. Chem. 1988, 1988, 1459-1467. (35) Anjo, D. M.; Kahr, M.; Khodabakhsh, M. M.; Nowinski, S.; Wanger, M. Anal. Chem. 1989, 61, 2603-2608. (36) McDermott, C. A.; Kneten, K. R.; L., M. R. J. Electrochem. Soc. 1993, 140, 2593-2599. (37) Chen, P.; Fryling, M. A.; McCreery, R. L. Anal. Chem. 1995, 67, 31153122. (38) Kozloswski, C.; Sherwood, P. M. J. Chem. Soc., Faraday Trans. 1 1984, 80, 2099-2107. (39) Kozlowski, C.; Sherwood, P. J. Chem. Soc., Faraday Trans. 1 1985, 81, 27452756. (40) Hance, G. W.; Kuwana, T. Anal. Chem. 1987, 59, 131-134. (41) Bowling, R.; Packard, R. T.; McCreery, R. L. Langmuir 1989, 5, 683-688. (42) Heiduschika, P.; Munz, A. W.; Gopel, W. Electrochim. Acta 1994, 39, 22072223. (43) Bielby, A. L.; Sasaki, T. A.; Stern, H. M. Anal. Chem. 1995, 67, 976-980. (44) Wang, J.; Martinez, D.; Yaniv, R.; McCormick, L. D. J. Electroanal. Chem. 1990, 278, 379. (45) Bodalbhai, L.; Brajter-Toth, A. Anal. Chem. 1991, 63, 1047-1049.

Dubuque, IA) and purged with nitrogen gas for 5 min prior to use. Electrode Preparation and Electrochemical Measurements. The GC-20 electrodes (Tokai GC-20, Electrosynthesis Corp., NY) were prepared by polishing with successive slurries of 1.0-, 0.3-, and 0.05-µm alumina (Buehler) in distilled/deionized water on polishing microcloth (Buehler). The GC electrodes were sonicated for 10 min between polishing. Polished GC electrodes were patterned by spraying the surface with polystyrene dissolved in CCI4 or by using standard photoresist-based microfabrication techniques. In the case of photoresist patterning, HPR 504 resist (OCG Chemicals) was spin-coated on a polished GC surface. The substrate (coated GC electrode) was illuminated through a copper grid mask (Pelco, CA) with UV light (∼405 nm) and then developed with Shipley 354 developer. This produced an electrode with masked and unmasked regions. Electrochemical pretreatment was achieved by poising a polished electrode at +1.80 V for 10-120 s in either acidic electrolyte (1.0 M H2SO4) or basic solution (0.1 M NaOH). A three-electrode cell was used with a Ag/AgCI reference electrode and a Pt wire counter electrode. The cell was connected to a model 175 (Princeton Applied Research, NJ) potentiostat. Electrochemically treated GC electrodes were sonicated in acetone for 10 min to remove the mask. Cyclic voltammetry was performed at a scan rate of 0.1 V/s. Differential capacitance was determined using a 100-Hz, 20-mV peak-to-peak triangle wave centered at 0.0 V. The peak-to-peak current of the output waveform is proportional to the capacitance.46 Control Experiments. Fe(CN)64-(aq) was used as a probe to assess the removal of the mask (polystyrene and photoresist) upon sonication of the ECP GC electrode in acetone. Polished GC electrodes were coated with polystyrene or photoresist followed by sonication in acetone for 10 min and then allowed to air-dry. The cyclic voltammetic peak separation (∆Ep) for Fe(CN)64-(aq) was then measured and found to be similar to values obtained on freshly polished GC electrodes. These results imply that acetone is effective in completely removing the mask without altering the electrochemical reactivity. Similar electrochemical characterization was performed to establish whether acetone affected the graphitic oxide layer formed upon activation of polished GC in 1.0 M H2SO4. ∆Ep values for Eu3+(aq) were found to be similar on oxidized GC electrodes before and after sonication in acetone. These results indicate that sonication in acetone does not siginificantly alter the graphitic oxide layer formed upon anodization. SFM Conditions. TM SFM images were obtained in air using Nanoscope III (Digital Instruments, Santa Barbara, CA). Si cantilevers were oscillated at their resonance frequency (∼300 kHz). Scanning was carried out with a constant amplitude of oscillation. An important parameter in TM SFM phase contrast imaging is the ratio of the set point (or imaging) amplitude, Asp, to the oscillation amplitude, Ao, where rsp ) Asp/Ao.47 This ratio governs the tapping force between the tip and the sample. All images shown here were collected with rsp between 0.55 and 0.65 (moderate tapping force). The scan rate was between 0.5 and 1.0 Hz. The images presented here are representative of many images taken at different points on each sample. All topographic and phase (46) Gileadi, E.; Tsherniskovski, N.; Babai, M. J. Electrochem. Soc. 1972, 119, 1018. (47) Magonov, S. N.; Elings, V.; Whangbo, M.-H. Surf. Sci. 1997, 375, L385.

Analytical Chemistry, Vol. 71, No. 19, October 1, 1999

4307

Table 1. Results from Electrochemical and TM SFM Characterizations of Polished GC Electrodes Electrochemically Pretreated at 1.8 V in 1.0 M H2SO4 oxidation time (s) polished 30 60 90 a

∆Ep (mV)a 3-/4- b

Fe(CN)6

82 ( 13 85 ( 4 94 ( 3 100 ( 4

Ru(NH3)63+/2+ b

Eu3+/2+ c

C° (µF/cm2)b

∆φ (deg)

82 ( 8 78 ( 4 69 ( 4 65 ( 3

265 ( 35 130 ( 9 105 ( 5 80 ( 6

41 ( 10 99 ( 8 128 ( 7 140 ( 7

5.4 ( 1.4 7.7 ( 1.6 10.2 ( 1.9

ν ) 100 mV/s for all systems. b 1 M KCl. c 0.2 M NaClO4.

Table 2. Results from Electrochemical and TM SFM Characterizations of Polished GC Electrodes Electrochemically Pretreated at 1.8 V in 0.1 M NaOH oxidation time (s) polished 10 30 60 120 a

∆Ep (mV)a 3-/4- b

Fe(CN)6

82 ( 13 80 ( 3 82 ( 5 83 ( 4

Ru(NH3)63+/2+ b

Eu3+/2+ c

C° (µF/cm2)a

82 ( 8 84 ( 4 83 ( 7 81 ( 6

265 ( 35 259 ( 8 223 ( 17 184 ( 19

41 ( 10 54 ( 7 80 ( 11 101 ( 9

∆φ (deg) 1.9 ( 0.6 6.3 ( 1.3 8.1 ( 1.5 12.2 ( 1.6

ν ) 100 mV/s for all systems. b 1 M KCl. c 0.2 M NaClO4.

contrast images presented in this study were collected simultaneously. Images were software flattened and are shown unfiltered. RESULTS AND DISCUSSION Electrochemical Characterizations of ECP GC. Studies of electrochemically pretreated GC electrodes have provided a wealth of evidence that this procedure alters surface composition.23 ECP in acidic solutions and at high positive potentials (e.g., 1.8 V vs SCE) produces a surface film that exhibits properties similar to graphite oxide and has been termed electrochemical graphitic oxide (EGO). Qualitatively, this layer is porous, hydrated, and nominally anionic.23,34,35 Anodization of GC in basic solutions produces interfacial carbon-oxygen functionalites39 but does not result in an EGO film.33,35 Reported results imply that if an EGO film is produced by ECP in base, it is immediately removed, possibly by dissolution.35,43 Although the effect of ECP on electrochemical parameters has been extensively documented, we have carried out several electrochemical characterizations of polished GC surfaces after ECP to track compositional changes and to correlate with TM SFM results. These data are listed in Tables 1 and 2. Table 1 provides results from electrochemical and TM SFM characterizations of polished GC surfaces oxidized in 1.0 M H2SO4. The cyclic voltammetric peak separations (∆Ep) for three redox systems (Ru(NH3)62+/3+, Fe(CN)63-/4-, Eu2+/3+) were utilized to assess the changes in electron-transfer (ET) reactivity induced by ECP for various times. ECP of GC in H2SO4 affects the ET for Fe(CN)63-/4- and Ru(NH3)62+/3+ in opposite ways. The trends in Table 1 likely reflect an electrostatic effect resulting from the formation of an anionic EGO film causing ∆Ep for the negatively charged Fe(CN)63-/4- system to increase slightly with oxidation time and decrease for Ru(NH3)62+/3+. For Eu2+/3+, Table 1 shows a significant decrease in ∆Ep with oxidation time. This dependence is consistent with reports that show ET for aquated metal ion systems such as Eu2+/3+ is governed by surface carbon4308 Analytical Chemistry, Vol. 71, No. 19, October 1, 1999

oxygen groups at carbon electrodes.36 Table 1 also shows a drastic increase in differential capacitance (C°) with oxidation time probably due to the combination of redox-active functional groups produced by ECP and the increased porosity and anionic character of the resulting film. These electrochemical results point to the increasing growth of a anionic EGO layer with ECP time and are consistent with those from previous studies.23 Similar electrochemical characterizations were performed on GC electrodes anodized in 0.1 M NaOH. The trends shown in Table 2 indicate that ECP in basic media has little effect on ET for Fe(CN)63-/4- and Ru(NH3)62+/3+ for the oxidation times investigated. In addition, although oxidation in NaOH induces an overall decrease in ∆Ep for Eu2+/3+, the effect for similar modification time is significantly less than ECP in acidic solution. A similar observation is made for the trend in C°. These results are all consistent with the generation of surface oxygen functionalities in the absence of an EGO film. Overall, the results from electrochemical characterizations listed in Tables 1 and 2 illustrate the influence of ECP in acidic and basic solutions on electrochemical reactivity. TM SFM Imaging of ECP GC. Our initial efforts in tracking ECP-induced transformations of GC surfaces with TM SFM proved problematic due to the similarity between the oxidized surface and the original polished electrode. This is illustrated in Figure 1, which contains 40 × 40 µm topographical TM SFM images of polished GC (Figure 1A) and polished GC anodized in 1.0 M H2SO4 (Figure 1B) and in 0.1 M NaOH (Figure 1C). The morphology shown in these images is similar to that noted in previous studies of polished GC where the main features are polishing scratches.48-50 Overall, these three surfaces appear qualitatively similar and (48) Kazee, B.; Weisshaar, D. E.; T., K. Anal. Chem. 1985, 57, 2739-2740. (49) Smyrl, W. H.; Atanasoski, R. T.; Hartshorn, L.; Lien, M.; Nygren, K.; Fletcher, E. A. J. Electroanal. Chem. 1989, 264, 301-304. (50) McDermott, M. T.; McDermott, C. A.; McCreery, R. L. Anal. Chem. 1993, 65, 937-944.

Figure 1. The 40 × 40 µm TM SFM topographic images of polished GC (A), acid-oxidized polished GC (B) and base-oxidized polished GC (C). z-scale ) 55 nm for all the images.

Figure 2. The 56 × 56 µm TM SFM images of a GC surface oxidized at 1.8 V in 1 M H2SO4 for 90 s. (A) Topography (z-scale ) 60 nm). (B) Phase contrast (z-scale ) 40°).

require identical height scales. Images at higher magnification do not provide any improved differentiation between the surfaces. Images collected employing a range of rsp from 0.8 (low tapping force) to 0.2 (high tapping force) do not appear substantially different from those in Figure 1. We do note that the features in Figure 1C appear more defined and “focused”. This is a general observation for GC surfaces oxidized in 0.1 M NaOH and is discussed in more detail below. The lack of differentiation motivated us to develop an electrode preparation method that would permit the direct comparison of oxidized and nonoxidized regions in the same image. Preselected regions of polished GC electrodes can be masked from the ECP procedure by employing standard integrated circuit patterning techniques. Electrochemical oxidation is then restricted to areas patterned into a layer of photoresist. Dissolution of the photoresist mask after anodization yields a GC surface with adjacent oxidized and polished regions that can then be differentiated within the same image. Figure 2 contains 56 × 56 µm TM SFM images of a GC electrode prepared in this way and oxidized in 1.0 M H2SO4 at 1.8 V for 90 s. Note from Table 1 that this pretreatment has a large effect on electrochemical reactivity, implying the generation of a EGO film. Figure 2A shows that ECP in H2SO4 induces minor morphological changes in the GC surface. Several rings of raised topography are apparent which are located at the boundary that existed between the exposed GC and the photoresist mask. These rings are not residual photoresist and are likely due to an enhanced oxidation at this boundary. The areas inside and outside the rings appear qualitatively similar with polishing scratches

traversing the modified regions uninterrupted. In fact, if the rings outlining the pattern boundary were not present, the location of the oxidized regions would not be discernible in topographic TM SFM images. We do not observe any evidence for a distinct EGO layer in any of our topographic images including those from surfaces oxidized for 5 min. In general, we observe a morphology similar to Figure 2A for oxidations from 10 s to 5 min at 1.8 V. Significant differentiation between oxidized and nonoxidized GC is observed in the phase image of Figure 2B. In this image, the circular regions that were exposed to the ECP procedure through the mask exhibit brighter contrast relative to the unmodified, polished GC. This contrast reflects differences in the phase lag of the oscillating cantilever/tip assembly with a greater phase lag manifested as darker contrast. The oxidized regions in Figure 2B exhibit a phase lag that is ∼10° less than the polished regions. Qualitatively, the observed phase lag difference (∆φ) can be interpreted as differences in cantilever energy dissipation resulting from variations in tip-sample interactions. Thus, we can infer that the ∆φ observed between the polished and oxidized regions in Figure 2B is due to an ECP-induced transformation of the GC surface architecture. A more detailed mechanism of the observed phase lag differences is presented below. Along with the electrochemical data in Table 1 are listed ∆φ values measured at patterned electrodes after oxidation in 0.1 M H2SO4. The magnitude of ∆φ directly tracks oxidation time, arguing that the compositional change induced by ECP proceeds with continued oxidation. This is supported by the correlation between the trend in ∆φ and the changes in the electrochemical parameters. These observations demonstrate the utility of phase contrast TM SFM for tracking compositional changes of modified carbon electrodes. Relative to ECP in acidic solutions, we observe surprising morphological changes for polished GC electrodes oxidized in 0.1 M NaOH. Parts A and B of Figure 3 are 100 × 100 µM topographic and phase contrast TM SFM images highlighting these changes. Contained in Figure 3A are a series of 15-µm-wide depressions corresponding to areas that were exposed to oxidation at 1.8 V for 1 min. From the cross-sectional profile, the measured depth of the depressions is 235 ( 10 nm for a 1-min oxidation. The flat profile of the bottom of the depressions in the cross section indicates that the SFM tip is probing the entire depth of the depressions. We note that little information is available with Analytical Chemistry, Vol. 71, No. 19, October 1, 1999

4309

Figure 4. Etching depth (nm) vs oxidation time (s) for ECP at 1.8 V in 0.1 M NaOH. The line through the points is simply a guide to the eye and does not represent any functional dependence.

Figure 3. The 100 × 100 µm TM SFM images of a GC surface oxidized at 1.8 V in 0.1 M NaOH for 60 s. (A) Topography (z-scale ) 350 nm). (B) Phase contrast (z-scale ) 40°). The cross-sectional profile corresponds to the line traversing the topographic image.

respect to the sides of the depressions due to inability of the pyramidal tip to probe steep walls. It is clear from Figure 3A that ECP in 0.1 M NaOH at 1.8 V removes a significant amount of carbon material from the GC surface. Although electrochemical etching of GC in basic solutions has not been directly explored in the literature, a similar process has been noted in a report addressing the effect of ECP in NaOH on pyrolytic carbon films deposited on polished GC substrates. It was observed that electrochemical oxidation in 1 M NaOH removed the carbon film and exposed the underlying GC substrate.33 Also, a study on electrochemical oxidation of carbon fibers in NaOH solutions reports the observation of carbon fragments in the working electrode solution. The following processes were proposed to account for carbon fiber lattice breakup as well as observed gas evolution during oxidation:

C(s) + OH- f C(s)OH(ads) + e 4C(s)OH(ads) f 4{C} +2H2O + O2 where C(s) implies an intact carbon lattice and {C} indicates carbon material removed from the substrate.39 We also observe gas evolution during oxidation of GC in base that, when combined with the etching apparent in Figure 3, implies that the above processes effectively describe electrochemical oxidation of GC in 0.1 M NaOH at 1.8 V. Our results indicate that the amount of carbon material removed by ECP in NaOH solutions is dependent on anodization time. Figure 4 is a plot of depression depth as a function of anodization time that illustrates that the etching depth is controllable over the time scales studied. We were limited to a maximum ECP time of 2 min because of dissolution of the photoresist in the 0.1 M NaOH at longer times and are currently exploring our ability to etch deeper. The effectiveness of electrochemical oxidation for micromachining GC substrates is also being examined in our laboratory. Figure 4 indicates that carbon material 4310 Analytical Chemistry, Vol. 71, No. 19, October 1, 1999

Figure 5. The 30 × 30 µm TM SFM images of a GC surface oxidized at 1.8 V in 0.1 M NaOH for 10 s through a polystyrene mask. (A) Topography (z-scale ) 30 nm). (B) Phase contrast (z-scale ) 20°).

begins to be removed from the surface at short time scales. For example, ECP for 10 s in 0.1 M NaOH etches the GC surface to a depth of ∼30 nm. Thus, significant alterations in the surface structure of GC electrodes occur during the initial stages of ECP in basic solutions. We believe the initial etching period effectively removes the microparticle layer known to exisit on polished GC surfaces. This layer is thought to be composed of polishing debris, impurities, and carbon microparticles.48 The speculation for the removal of this layer derives, in part, from consideration of images such as those in Figure 5. These 30 × 30 µm images are of a partially masked GC electrode that was anodized in 0.1 M NaOH for 10 s. The mask in this case, was a partial polystyrene (PS) film that was delivered as droplets from a nebulized solution as described in the Experimental Section. Analysis of the topographic image (Figure 5A) reveals that the polished surface has been etched to a depth of ∼30 nm, which is sufficient to account for the thickness reported for the microparticle layer of 20 nm.48 Polishing scratches, which are apparent throughout the images, appear more numerous and more defined in the etched regions, similar to Figure 1C. As shown in both images in Figure 5, the microparticle layer limits the ability of the SFM tip to probe the smaller scratch structures resulting from the polishing procedure. The appearance of scratches in optical micrographs of GC electrodes has been suggested to signify the removal of the polishing layer following heat treatment,48 laser irradiation,51 and ECP.33 The images in Figure 5 provide direct evidence of its removal. (51) Poon, M.; McCreey, R. L. Anal. Chem. 1986, 58, 2745-2750.

Figure 6. Schematic model illustrating GC surface architecture and tip interactions during TM SFM imaging of patterned GC electrodes electrochemically oxidized in (A) NaOH and (B) in H2SO4. For clarity, the carbon microparticles and functional groups known to exisit on the polishing layer are not shown.

Additional evidence for the loss of the microparticle polishing layer after short ECP times (10 s) in NaOH derives from careful consideration of the electrochemical data in Table 2. Although insignificant changes in electrochemical reactivity are induced by short oxidation times, the standard deviation for each set of measurements is notably smaller relative to polished GC. This increase in precision is consistent with the removal of the polishing microparticle layer. The variability in thickness, composition, etc., of this layer from experiment to experiment likely results in the larger standard deviations observed at polished GC. Thus, a more reproducible electrode surface, with reactivity similar to polished GC, can be prepared by ECP in NaOH for 10 s. The ability of phase contrast TM SFM to map changes in GC surface composition induced by ECP is again illustrated in Figures 3B and 5B. Similar to ECP in acid (Figure 2B), the polished GC exhibits a greater phase lag (darker contrast) than GC oxidized in 0.1 M NaOH (Figure 3B). The ∆φ values listed in Table 2 track ∆Ep for Eu2+/3+ and C°, implying that ECP in base influences both phase contrast and electrochemical reactivity. Taken together, the images in Figures 2, 3, and 5 show the power of TM SFM for monitoring both morphological and compositional variations at modified GC electrodes. Mechanism of Phase Contrast in TM SFM Images of Masked GC Electrodes Following ECP. Several modeling efforts aimed at predicting factors that influence phase contrast in TM SFM have been reported.47,52-58 It has been suggested that (52) Tamayo, J.; Garcia, R. Langmuir 1996, 12, 4430. (53) Burnham, N. A.; Behrend, O. P.; Oulevey, F.; Gremaud, G.; Gallo, P.-J.; Gourdon, D.; Dupas, E.; Kulik, A. J.; Pollock, H. M.; Briggs, G. A. D. Nanotechnology 1997, 8, 67-75. (54) Brandsch, R.; Bar, G.; Whangbo, M.-H. Langmuir 1997, 13, 6349-6353. (55) Tamayo, J.; Garcia, R. Appl. Phys. Lett. 1997, 71, 2394-2396. (56) Cleveland, J. P.; Anczykowski, B.; Schmid, A. E.; Elings, V. B. Appl. Phys. Lett. 1998, 72, 2613-2615. (57) Bar, G.; Brandsch, R.; Whangbo, M.-H. Surf. Sci. 1998, 411, L802-L809. (58) Garcia, R.; Tamayo, J.; Calleja, M.; Garcia, F. Appl. Phys. A 1998, 66, S309312.

interactions between the tip and surface that damp the cantilever oscillation are responsible for phase lags in TM SFM.55,56 Thus, phase contrast images collected with constant oscillation amplitude can be considered maps of local energy dissipation. Surface parameters that can affect energy-dissipative interactions include viscoelasticity, plastic deformation, and adhesion hysteresis. Although the surface composition of GC electrodes following ECP in acidic solutions varies significantly from that due to ECP in base, the direction of phase contrast and the overall trend of ∆φ with oxidation time is similar for both media. This argues that the observed contrast in images consisting of both polished and oxidized regions is dominated by the interaction of the SFM tip with the polished portion of the electrode. It follows that phase contrast is generated because ECP in both media removes or transforms the ubiquitous polishing layer. The darker contrast observed at the polished regions of our segregated electrodes reflects a greater phase lag and implies that more cantilever energy is dissipated at these areas. We do not observe variations in surface topography or phase contrast during repeated imaging of a single area and thus believe that plastic deformation of the GC surface is negligible. We attribute the observed phase contrast to an interplay between adhesion hysteresis and viscoelastic effects. Toward differentiating between these two parameters, we have made preliminary adhesion measurements between Si3N4 tips and GC surfaces which yielded the following trend in adhesion hysteresis: base oxidized ≈ polished < acid oxidized. It is clear from Figures 3 and 5 that anodization of GC in base removes the layer of polishing impurities and produces a surface consisting of well-defined features (scratches, etc.). Because of the similarities in the adhesion hysteresis of these two surfaces, we believe that the observed contrast derives from differences in viscoelasticity. It is reasonable that an amorphous polishing layer consisting of water, polishing debris, and carbon microparticles48 would be mechanically soft and exhibit viscoelastic properties. Analytical Chemistry, Vol. 71, No. 19, October 1, 1999

4311

Phase shifts due to viscous damping will become more pronounced at surfaces with lower Young’s modulus (lower stiffness).55,58 The Young’s modulus of GC (0.5 GPa)59 is in the range where modest changes in viscosity or changes in elastic modulus at constant viscosity will induce significant differences in cantilever phase lag.55,58 Thus, the different mechanical properties between polished and base-oxidized GC may contribute to the observed phase contrast. Specifically, as shown in Figure 6A, on its downward cycle the cantilever/tip assembly penetrates the polishing layer and experiences greater viscous damping relative to the oxidized regions. The differences in adhesion between the downward and upward cycles (adhesion hysteresis) are similar at each region. The nature of the phase contrast observed between the polished and acid ECP is less intuitively obvious. Although not suggested by our topographic TM SFM images, phase contrast images and electrochemical results indicate that the polishing layer is removed or transformed by ECP in acidic media at shorter times, consistent with literature reports (Figure 6B).23 Considering only the greater measured adhesion at acid-oxidized GC, we would expect the oxidized regions in this case to appear darker in contrast. However, this is not the case, indicating that viscoelastic effects dominate over adhesion hysteresis. It has been noted that a dried EGO film is more stiff and brittle than an EGO film that remains wet.34 Perhaps the viscosity of the polishing layer is much greater than that of the dried EGO film. Thus, as shown in Figure 6B, although the tip penetrates each layer, cantilever damping is greater at the polished region. Although we do not fully understand this observation, we do note that our continuing studies reveal that regions oxidized in 1 M H2SO4 for longer times (5 min) are not differentiated from polished GC in phase images.60 Perhaps longer ECP times are required to generate an EGO film with mechanical properties similar to that of polished GC. In that (59) Sigradur, GC Webpage: www.htw-germany.com. (60) Kiema, G. K.; McDermott, M. T. Manuscript in preparation.

4312 Analytical Chemistry, Vol. 71, No. 19, October 1, 1999

case, adhesion hysteresis will begin to dominate the phase contrast. CONCLUSIONS We have shown here that transformations in GC surface architecture induced by chemical modification can be tracked with TM SFM by employing partially masked electrodes. Importantly, phase contrast imaging provides compositional information that can be correlated to topographic images as well as electrochemical reactivity. In the studies presented here, ECP of patterned GC electrodes in 1 M H2SO4 does not induce topographic changes detectable by TM SFM. However, the surface compositional variations indicated by electrochemical measurements can be probed with phase contrast TM SFM. ECP in 0.1 M NaOH etches the topography of the GC surface. This etching initially removes the layer of material resulting from the polishing procedure and further etching can be controlled by ECP time. We are currently investigating the potential of this etching procedure to produce micromachined GC structures. The compositional transformation induced by ECP in NaOH is also detected with phase contrast and correlates with electrochemical reactivity. We believe the basis of the contrast in TM SFM phase images is related to the viscous damping of the cantilever by the ubiquitous polishing layer of the unmodified regions of the patterned electrode. ACKNOWLEDGMENT This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Department of Chemistry, University of Alberta. We acknowledge T. Ta for useful discussions on initial SFM studies on oxidized GC and M. Finot for designing the PS spray system and for discussions on TM SFM imaging of oxidized GC. Received for review April 19, 1999. Accepted July 15, 1999. AC9904056