Introduction to the Principles of Ultramicroheptodes in Ring− Disk

The use of ring-disk ultramicroelectrodes as tips as a means of extending the methodology of the scanning electrochemical microscope is described. Ele...
0 downloads 0 Views 139KB Size
Anal. Chem. 2002, 74, 1316-1321

Introduction to the Principles of Ultramicroheptodes in Ring-Disk Interactions Joachim Ufheil,†,‡ Kai Borgwarth,†,‡ and Ju 1 rgen Heinze†,‡,*

Institut fu¨r Physikalische Chemie, Albert-Ludwigs-Universita¨t Freiburg, Albertstrasse 21, D-79104 Freiburg, Germany, and Freiburger Materialforschungszentrum (FMF), Albert-Ludwigs-Universita¨t Freiburg, Stefan-Meier-Strasse 21, D-79104 Freiburg, Germany

The use of ring-disk ultramicroelectrodes as tips as a means of extending the methodology of the scanning electrochemical microscope is described. Electrodes consisting of one centered disk with six interconnected disks surrounding it served as the ring-disk electrodes. Basic experiments illustrate the behavior in the feedback mode and the generation collection mode. The dependence on the electrodes’ size, both potentials, the tip-sample distance, and the local properties of the underlying sample were studied. When approaching an electrode to the surface of a sample, steady-state collection efficiencies between zero and unity were elegantly altered by adjusting the distance between tip and sample. The shielding factor could be varied between 0.25 and 0.82 the same way. Concerning feedback methods, the results presented illustrate the new principle of applying a precisely located external stimulus as a separate electrochemical means of analyzing the sample’s response. Obviously, this technique can be extended to irreversible redox mediators, which are not treated here explicitly, but give this concept even greater flexibility. The interpretation of such data can be deduced directly from this contribution. The scanning electrochemical microscope (SECM)1,2 was developed by the groups of Bard and Engstro¨m3,4 to perform surface analysis and modification. An ultramicroelectrode (UME) is scanned in the vicinity of a sample’s surface covered by a liquid electrolyte. The tip initiates redox reactions with electroactive species present in the solution, and the resulting faradaic current is measured as a function of the location. The SECM has been applied in a wide variety of current research fields that benefit from its high chemical sensitivity, including the analysis of corrosion5-7 and the imaging of enzymes8,9 and cultured cells.10,11 One feature, local electrogeneration of species, has been exten* Corresponding author: (e-mail) [email protected]; (fax) ++49/761/203 6237. † Institut fu ¨ r Physikalische Chemie. ‡ Freiburger Materialforschungszentrum. (1) Bard, A. J.; Fan, F.-R. F.; Mirkin, M. In Physical Electrochemistry; Rubinstein, I., Ed.; Marcel Dekker: New York, 1995; p 209. (2) Bard, A. J.; Fan, F.-R. F.; Mirkin, M. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1994; Vol. 18, p 243. (3) Liu, H.-Y.; Fan, F.-R. F.; Lin, C. W.; Bard, A. J. J. Am. Chem. Soc. 1986, 108, 3828. (4) Engstrom, R. C.; Weber, M.; Wunder, D. J.; Burgess, R.; Winquist, S. Anal. Chem. 1986, 58, 844.

1316 Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

sively applied to initiate laterally delimited surface modifications, such as deposition of metals12-14 and conducting polymers15-18 and etching of metals19-21 and semiconductors.22-24. Other attractions of the SECM include the study of kinetics of, for example, chemical reactions,25 the dissolution of salts,26 and phase-transfer reactions.27 Following a recent trend, some studies have reported using the SECM to study or to initiate surface changes, while choosing another technique to analyze the process. The SECM has been combined with the scanning near-field optical microscope (SNOM) to analyze corrosion phenomena,28 with atomic force microscopy (AFM) in studies of crystal dissolution,29 and with the quartz crystal microbalance (QCM) during dissolution30 and deposition.31 Surprisingly, no multiarray electrodes had been used in the SECM although ultramicroheptodes are well known. They are (5) Garfias-Mesias, L. F.; Alodan, M.; James, P. I.; Smyrl, W. H. J. Electrochem. Soc. 1998, 145, 2005. (6) Still, J. W.; Wipf, D. O. J. Electrochem. Soc. 1997, 144, 2657. (7) Zhu, Y.; Williams, D. E. J. Electrochem. Soc. 1997, 144, L43. (8) Shiku, H.; Uchida, I.; Matsue, T. Langmuir 1997, 13, 7239. (9) Wittstock, G.; Schuhmann, W. Anal. Chem. 1997, 69, 5059. (10) Tsionsky, M.; Cardon, Z. G.; Bard, A. J.; Jackson, R. B. Plant Physiol. 1997, 113, 895. (11) Yasukawa, T.; Kondo, Y.; Uchida, I.; Matsue, T. Chem. Lett. 1998, 767. (12) Mandler, D.; Bard, A. J. J. Electrochem. Soc. 1990, 137, 1079. (13) Hess, C.; Borgwarth, K.; Ricken, C.; Ebling, D. G.; Heinze, J. Electrochim. Acta 1997, 42, 3065. (14) Borgwarth, K.; Heinze, J. J. Electrochem. Soc., 1999, 146, 3285. (15) Marck, C.; Borgwarth, K.; Heinze, J. Chem. Mater. 2001, 13, 747-752. (16) Marck, C.; Borgwarth, K.; Heinze, J. Adv. Mater. 2001, 13, 47-51. (17) Kranz, C.; Ludwig, M.; Gaub, H. E.; Schuhmann, W. Adv. Mater. 1995, 7, 568. (18) Borgwarth, K.; Ricken, C.; Ebling, D. G.; Heinze, J. Ber. Bunsen-Ges. Phys. Chem. 1995, 99, 1421. (19) Mandler, D.; Bard, A. J. J. Electrochem. Soc. 1989, 136, 3143. (20) Still, J. W.; Wipf, D. O. J. Electrochem. Soc. 1997, 144, 2657. (21) Ufheil, J.; Boldt, F. M.; Bo¨rsch, M.; Borgwarth, K. Bioelectrochemistry 2000, 52, 103. (22) Mandler, D.; Bard, A. J. J. Electrochem. Soc. 1990, 137, 2468. (23) Meltzer, S.; Mandler, D. J. Chem. Soc., Faraday Trans. 1995, 91, 1019. (24) Zu, Y. B.; Xie, L.; Mao, B. W.; Tian, Z. W. Electrochim. Acta 1998, 43, 1683. (25) Richards, T. C.; Bard, A. J.; Cusanelli, A.; Sutton, D. Organometallics 1994, 13, 757. (26) Macpherson, J. V.; Unwin, P. R. J. Phys. Chem. 1995, 99, 14824. (27) Slevin, C. J.; Macpherson, J. V.; Unwin, P. R. J. Phys. Chem. B 1998, 101, 10851. (28) James, P.; Casillas, N.; Smyrl, W. H. J. Electrochem. Soc. 1996, 143, 3853. (29) Macpherson, J. V.; Unwin, P.; Hillier, A. C.; Bard, A. J. J. Am. Chem. Soc. 1996, 118, 6445. (30) Cliffel, D. E.; Bard, A. J. Anal. Chem. 1998, 70, 1993. (31) Hess, C.; Borgwarth, K.; Heinze, J. Electrochim. Acta 2000, 45, 3725. 10.1021/ac010912z CCC: $22.00

© 2002 American Chemical Society Published on Web 02/09/2002

widely used in, e.g., physiology,32 and have been introduced into electrochemistry in recent years. To obtain ring-disk ultramicroelectrodes (RD-UMEs), Kirchhoff et al. used chemical vapor deposition (CVD) to coat carbon fibers with alternating concentric layers of insulating silica, carbon, and silica.33 As this group was able to control the geometry precisely, they could study the influence on collection efficiency and feedback factors. Since 1996, different groups have studied dual-disk microelectrodes, two electrodes closely neighbored and embedded in the same glass body.34,35 The ring-disk and dualdisk electrodes exhibited similar behavior, since both shared the principle of interacting diffusion fields. Both approaches combine the benefits of ultramicroelectrodes36 with those of the rotating ring-disk electrodes. With regard to scanning probe techniques, a symmetrical tip like a ring-disk electrode is obviously to be favored to avoid influences of scan direction. The RD-UME represents the static electrode analogue of the rotating ring-disk electrode (RRDE). In the RRDE, the bulk species are transported by convection to the disk electrode first and then to the ring. By contrast, the transport to a RD-UME occurs exclusively via diffusion, which is not directed to one specific electrode. Compared with the known SECM collection generation modes using just one scanning electrode,37 the concept presented here offers two significant additional advantages. On one hand, it does not require electrical contact with a conducting sample, which involves chemical methods of species generation in addition to the established electrochemical ones. On the other hand, the stimulation of surface reactions does not occur over the entire sample but locally. Consequently, the diffusion fields involved become hemispheric and stationary in general. Applied to the feedback mode, two redox mediators of different properties can be simultaneously oxidized and reduced, respectively. In both scanning modes, it is possible to investigate surface states of poor stability under conditions of stimulation, which opens up a wide range of applications. We describe here experiments with ring-disk ultramicroelectrodes approaching active and inactive surfaces. They were carried out using a reversible redox mediator in the feedback mode and generation/collection mode. The experiments illustrate the principle of stimulating surface reactions and studying them by electrochemical means. The application to mixtures of reversible and irreversible mediators reveals new ways of determining surface properties and will be published soon.38 EXPERIMENTAL SECTION The electrodes used in the experiments were standardized heptodes purchased from TREC (Marburg, Germany, http:// www.ThomasRecording.de). They consisted of a 4-µm platinum core as the disk surrounded by quartz glass and another six electrodes of approximately the same size. These outer electrodes were electrically connected and served as the ring electrode. Electrodes of double extensions and the same proportions were (32) Blum, N. A.; Carkhuff, B. G.; Charles, H. C.; Edwards, R. L.; Meyer, R. A. IEEE Transact. Biomed. Eng. 1991, 38, 68. (33) Zhao, G.; Giolando, D. M.; Kirchhoff, J. R. Anal. Chem. 1995, 67, 1491. (34) Yasukawa, T.; Kaya, T.; Matsue, T. Anal. Chem. 1999, 71, 4637. (35) Matysik, F.-M. Electrochim. Acta 1997, 42, 3113. (36) Heinze, J. Angew. Chem., Int. Ed. Engl. 1993, 32, 1268. (37) Bard, A. J.; Fan, F. R. F.; Kwak, J.; Lev, O. Anal. Chem. 1989, 61, 132. (38) Boldt, F. M.; Ufheil, J.; Borgwarth, K.; Heinze, J., unpublished work.

Figure 1. Optical image of an ultramicroheptode. The outer six electrodes are electrically connected and act as a ring surrounding the center disk electrode. Disk radius adisk ) 2 µm, ring radius aring,central ) 8 µm, Rglass ) 18 µm f RGdisk ) Rglass/adisk ) 9, and RGring ) Rglass/aring,central ) 2.25. Calculated ring thickness for same area as six disks ∆aring ) 1.33 µm.

used as well; the qualitatively similar results are not explicitly reported here. Figure 1 shows the apex of a small electrode and its dimensions. The central radius of the ring aring,central is 9 µm. To enable comparison with a ring of the same area, its thickness was calculated as ∆aring ) 1.33 µm. The conical ends of the electrodes were polished to obtain a flat apex desirable for SECM experiments. For this purpose, Fibermet disks (Buehler) coated with alumina particles of 9 µm, 3 µm, and 300 nm, respectively, were mounted on a homemade beveler. As a model system for an active sample, we installed a platinum sheet (Goodfellow), and for an inactive one, we used the flat bottom of a glass Petri dish. The aqueous electrolyte consisted of 10 mM rutheniumhexamine ([Ru(NH3)6]Cl3, Aldrich) and 0.5 M KCl. Our SECM is equipped with one stepper motor for large vertical movements and three closed-loop piezos for 400-µm travel, each at a resolution of 50 nm. The whole setup is mounted on a vibration insulating table (MOD-1, hwl.bioanalytic). A home-built bipotentiostat combined with a Faraday cage enables a current resolution of 0.5 pA. All potentials refer to a silver wire that serves as a quasi-reference electrode. RESULTS AND DISCUSSION Voltammetric Characterization. The heptodes were examined by cyclic voltammetry (CV) to characterize the effect of overlapping diffusion fields of ring and disk without any influences of samples. Figure 2 shows the current of disk (a) and ring (b) recorded simultaneously as a function of the disk potential in a linear sweep. In one case, the ring was unbiased and kept at constant potentials for diffusion-limited oxidation and reduction, respectively. The disk CV is not noticeably affected by the potential of the ring around it. Only changes in the initial concentration of Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

1317

Figure 2. CV of a ring-disk electrode: Current of (a) disk and (b) ring electrode as a function of the disk potential. Measured at v ) 50 mV/s in aqueous 10 mM Ru(NH3)6Cl3, 500 mM KCl.

ruthenium(III), caused by a reduction at the ring electrode, led to variations of the disk current of the order of 60%. The diffusionlimited current of the disk is given by the following expression

Iinf,disk ) 4nFDcadisk

(1)

where n symbolizes the number of electrons transferred per species, F the Faraday constant, D the diffusion coefficient of the educt, 6.3 × 10-6 cm2/s for [Ru(NH3)6]Cl3,39 c the bulk concentration, and adisk the disk radius of 2 µm. According to eq 1, a diffusion-limited current of -4.8 nA is to be expected, which is in good agreement with the observed value of -4.4 nA. Due to the larger size of the ring, scanning the disk potential (Figure 2b) did not significantly affect its current. By contrast, the ring potential governed the current of the disk electrode, as shown in Figure 3a. According to approximation 2,36 a diffusion-

Iinf,ring ) 2π2nFDcaring,central/ln[32aring,central/∆aring] (2) limited current of about -19.8 nA should be expected for a solid ring of same area. The value measured in Figure 3b was -17.4 nA. Note that the ring current equals 4, not 6, times the disk current, owing to the overlapping diffusion fields of the outer electrodes. Obviously, the difference in geometry between the six (39) Wightman, R. M.; Wipf, D. O. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1988; Vol. 15, p 267.

1318 Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

Figure 3. CV of a ring-disk electrode: Current of (a) disk and (b) ring electrode as a function of the ring potential. Same parameters as in Figure 2

disks and the ideal ring-disk geometry is irrelevant for the electrochemical experiments. All curves reveal sigmoidal shapes and therefore represent steady states typical of ultramicroelectrodes. Diffusion is shown to be the exclusive mode of mass transport. Presumably, the charge transport to the center is supported by the presence of the ring. In all cases, opposite polarization of the electrodes yielded higher currents, indicating a recycling of species between the two, which can be understood as an internal feedback effect. A reducing electrode acted as a generator for the [Ru(NH3)6]2+ species, which could subsequently be collected at the second electrode if polarized anodically. One single species may in principle undergo several redox cycles before diffusing into the bulk solution. Obviously, the ring is more efficient as a collector, gathering 35% of the species generated at the disk in its center than vice versa with just 9% detection efficiency. The effects described above were comparable with both kinds of experiments reported in the literature, those using dual-disk and those using ring-disk electrodes, and were explained in terms of overlapping diffusion fields. In the following, all currents are normalized to the corresponding values at infinite distance. For the cathodically polarized electrode, the values are Iinf,disk ) -4.4 nA and Iinf,ring ) -17.4 nA. The anodic currents are normalized to the largest achieved, Iinf,disk ) +1.78 nA and Iinf,ring ) +1.54 nA. Sample-Induced Effects in Feedback Mode. The results presented above illustrate that the RD-UME acts as the static electrode analogue of the RRDE. The influence of different sample surfaces on the observed effects was studied by approaching the tip to two samples with opposite behavior. A platinum sheet enabled very fast electron-transfer reactions, whereas glass

Figure 4. Approach curves and simulations40,41 at 1 µm/s in the feedback mode. With just one biased electrode: (a) positive feedback above a platinum sheet; (b) negative feedback above a glass carrier. With equally biased electrodes: (c) positive feedback; (d) negative feedback. The ring and disk currents were normalized versus their respective values at infinite distance. The distance was normalized for each curve with respect to the electrode’s radius adisk. The shielding factor is calculated by dividing the disk current of (c) by (a) and of (d) by (b), respectively.

represented a completely inert surface. During all experiments, the platinum sheet was left unbiased. Polarized separately, the ring and disk electrodes showed the feedback behavior depicted in Figure 4a and b. To improve comparability, the distance was normalized for each curve by the disk electrode radius and the currents by those at infinitely large distances. Above the platinum, the current increased by more than a factor of 3, illustrating the continuous oxidation and reduction cycle of each species taking place in the gap between the active electrode and the sample. By contrast, above the glass, the current dropped to about 25% of the initial value, due to hindrance of the diffusion. Above the platinum, in particular, the experiment complied with the theory of the feedback mode, which was simulated for disk-shaped electrodes.40 The deviation of the ring current at midnormalized distances can be understood as the apparent merging of the six diffusion fields to act like one big electrode. At very low distances, all three curves coincided perfectly clearly, showing that the feedback cycles occur separately at one and six disks, respectively. Above the glass, the start of the decay of the ring current was delayed compared to the disk, which is to be expected on account of the larger diameter and lower shielding ratio. The heptode approach curves are very similar to feedback simulations for a true ring electrode with (40) Kwak, J.; Bard, A. J. Anal. Chem. 1989, 61, 1221.

roughly the same geometry41 and show the same behavior as true ring ultramicroelectrodes. If both electrodes are polarized at the same potential, concurrent reactions occur and lower individual currents have to be expected, as the disk electrode is partly shielded by the active ring. Since the direction of educt transport is reversed compared to the RRDE situation, the shielding factor42 should be defined as the ratio of disk currents recorded with and without polarizing the ring. Figure 4c and d illustrate that at infinite distances the disk current decreased by about 23% due to the reactions at the ring. This corresponds to a shielding factor u of 77%. Qualitatively, both curves illustrate the feedback situations as well. The absolute values of the currents are significantly reduced compared to Figure 4a and b because of the concurrent consumption of the oxidized species. Interestingly, the shielding factor u measured near the platinum decreased to 48% and then increased above the platinum to 78%, meaning not much shielding occurs. This behavior of the shielding factor can be explained by two different situations. In the first, the shielding factor decreases until a distance of 1 is reached, because the small volume between tip and sample surface increasingly hinders the mass transport. In (41) Lee, Y.; Amemiya, S.; Bard, A. J. Anal. Chem. 2001, 73, 2261. (42) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, Fundamentals and Applications; Wiley: New York, 1980.

Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

1319

Figure 5. Approach curves at 1 µm/s in the generation/collection mode with oppositely biased electrodes. (a) Disk generation, ring collection (DGRC) above conductor, (b) DGRC above insulator, (c) ring generation, disk collection (RGDC) above conductor, and (d) RGDC above insulator. The efficiency was given as the quotient of detected and generation current.

the second, when the disk is close to the active surface, the positive feedback becomes dominant. We believe that the shielding still occurs, but cannot prevent the positive feedback. The feedback cycles take place at each of the seven electrodes with little interference. Above the glass, the shielding factor increased to 80%, which is evidence that the disk is heavily shielded by the ring. In summary, the shielding factor can be varied up to 62% from its initial value just by altering the distance from an platinum sheet, and from the initial value to 79% just by altering the distance from a inert surface. Sample-Induced Effects in Generation/Collection Mode. In generation/collection experiments, the electrogenerated species is generated at one electrode and collected at a second. In general, either the disk generates species and the ring collects them (DGRC) or vice versa (RGDC). Obviously, the first case, in which the generator is placed in the center of the collector, is more advantageous with regard to high collection yield. Apart from being collected, the generated species may simply diffuse into the bulk solution or react at the sample’s surface. Therefore, the collection efficiency η and the total turnover become functions of the nature of sample and the distance from the tip. Figure 5a and b show the results using the disk (DGRC) as generator and Figure 5c and d the ring (RGDC). Due to the 1320

Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

chosen normalization of the generator, currents are identical with the feedback factors. Above the platinum sample (a) and (c), the currents of the generating electrodes increased, whereas the current of the collector decreased. Concurrent with the detection, the generated species reacted at the platinum sample, reducing collection efficiency. In other words, the feedback effects between generator and sample were stronger than the internal feedback between ring and disk. Approaching inactive surfaces, the current of the generating electrode continuously decreased, as expected. However, the collection current exhibited surprising behavior. After a long period of increase, it reached a maximum at normalized distances of about 0.5 and finally dropped. This phenomenon is explained by the decrease in generation current on one hand and limited collection efficiency on the other hand. The collection current significantly exceeded the values Iinf recorded with only one polarized electrode by 50%-100%, which illustrates the power of the internal feedback effect. Using the disk as the generator (b), collection efficiency is even increased from 35% to 100%, which offers ideal conditions for kinetic studies. The experiments prove that diffusion loss of electrogenerated species is effectively reduced and detection sensitivity is increased up to unity.

CONCLUSIONS A new tip tool for the SECM is described that should allow a wide range of kinetic measurements. It was shown that the shielding factor as well as the collection efficiency can be changed over a wide range by approaching suitable surfaces and varying the distance to the tip. Furthermore, the heptode opens a wide range of applications because it includes seven sensing elements that can be used separately. This allows local in situ modification and analysis with the SECM when working with modified heptodes.

ACKNOWLEDGMENT Financial support from the BMBF and the DFG is gratefully acknowledged. NOTE ADDED AFTER ASAP POSTING This article was inadvertently posted ASAP before final corrections were made. The corrected version was posted on February 26, 2002. Received for review August 14, 2001. Accepted December 4, 2001. AC010912Z

Analytical Chemistry, Vol. 74, No. 6, March 15, 2002

1321