Selective Electrodesorption Based Atomic Layer Deposition (SEBALD

Aug 11, 2011 - The method can also be used to obtain metal clusters of different size. In fact, the alternate underpotential deposition of elements th...
0 downloads 28 Views 3MB Size
Download PDF
ARTICLE pubs.acs.org/Langmuir

Selective Electrodesorption Based Atomic Layer Deposition (SEBALD): a Novel Electrochemical Route to Deposit Metal Clusters on Ag(111) M. Innocenti, S. Bellandi, E. Lastraioli, F. Loglio, and M. L. Foresti* Dipartimento di Chimica, Universita di Firenze, via della Lastruccia 3, 50019 Sesto Fiorentino (FI), Italy ABSTRACT: The possibility of synergic effects of some metals on the catalytic activity of silver led us to study the way to perform controlled deposition on silver. In fact, many metals of technological interest such as Co, Ni, and Fe cannot be deposited at underpotential on silver, and any attempt to control the deposition at overpotential, even at potentials slightly negative of the Nernst value, did not allow an effective control. However, due to the favorable energy gain involved in the formation of the corresponding sulfides, these metals can be deposited at underpotential on sulfur covered silver. The deposition is surface limited and the successive electrodesorption of sulfur leaves confined clusters of metals. The method can also be used to obtain metal clusters of different size. In fact, the alternate underpotential deposition of elements that form a compound is the basis of the electrochemical atomic layer epitaxy (ECALE), and the reiteration of the basic cycle allows us to obtain sulfide deposits whose thickness increases with the number of cycles. Therefore, the successive selective desorption of sulfur leaves increasing amounts of metals.

’ INTRODUCTION Today, electrochemistry extends to the domain of material science because of the development of electrochemical methods for the electrodeposition of materials controlled down to the atomic level. On silver, well-ordered atomic layers can be often obtained by exploiting surface phenomena such as underpotential deposition (UPD).1 Recently, interest in the underpotential electrodeposition of elements has increased essentially due to potential applications in several fields of technology such as photovoltaics and electrocatalysis. In photovoltaics, one example is the possibility of depositing alternate UPD layers of metals and nonmetals to grow binary and ternary compound semiconductors. This methodology is the basis of the electrochemical atomic layer epitaxy (ECALE) method proposed by Stickney and co-workers2 and extensively used to grow cadmium sulfide on gold.3 9 The ECALE method was adopted in our group to grow cadmium and zinc chalcogenides on silver single crystals, with particular attention to binary and ternary sulfides.10 13 In fact, unlike Se and Te that require more complex procedures to yield UPD layers, SUPD is simply obtained by oxidative underpotential deposition from sulfide ion solutions. In electrocatalysis, there is growing interest in the development of new substrates for cathodic reactions for employment both in the electrolysis of water and in oxygen reduction reaction (ORR) for alkaline fuel cells. With the aim of completely removing Pt and replacing it with less expensive materials, guidelines for the design of bimetallic electrocatalysts for ORR have been proposed assuming a simple mechanism where one metal breaks the oxygen oxygen bond of molecular O2 and the other metal acts to reduce the resulting adsorbed atomic oxygen.14,15 More precisely, metals such as Co, Ni, or Fe should favor the initial dissociative chemisorption of oxygen, r 2011 American Chemical Society

whereas silver should promote the following charge transfer steps. The theoretical predictions have been supported by the experimental evidence of catalytic activity of the Ag Co mixtures.14 More recently, we described the catalytic effect of Co monolayer islands formed on Ag substrate toward ORR.16 As stated before, well-ordered layers of some metals can be deposited on silver at underpotential. Unfortunately, Co is not included between these metals, and therefore, the problem of limiting its deposition is crucial even working at very low overpotentials. To limit Co deposition to a monolayer, we exploited the surface limited redox replacement (SLRR) method.17,18 According to this method, a layer of a metal deposited at underpotential is used as a template for the spontaneous deposition of a more noble metal monolayer. In our case, the deposition of Co through SLRR method has been performed replacing a layer of Zn deposited at underpotential on silver.16 Likewise, neither Ni nor Fe can be deposited at underpotential and require indirect methods. This paper presents a novel electrochemical route to deposit metal clusters on Ag(111). We called this method “selective electrodesorption based atomic layer deposition” (SEBALD). For its validation, the method was applied to obtain Cd layers from CdS that was a compound extensively studied in our group before.10 13 Then, the first results for other metals like Co, Ni, and Fe are reported.

’ EXPERIMENTAL SECTION Fluka analytical reagent grade Na2S, and Merck analytical reagent grade CoSO4 3 7H2O, 3CdSO4 3 H2O, NiCl2, FeSO4 3 7H2O, HClO4, Received: June 10, 2011 Revised: August 8, 2011 Published: August 11, 2011 11704

dx.doi.org/10.1021/la202174j | Langmuir 2011, 27, 11704–11709

Langmuir

ARTICLE

Figure 1. (a) Cyclic voltammograms obtained from 1 mM Cd(II) in pH 9.6 ammonia buffer solutions on bare Ag(111) (solid curve) and on Ag(111) covered by a SUPD layer. (b) Stripping peaks of Cd deposited at underpotential on the bare Ag(111) (solid curve) and on on Ag(111) covered by a SUPD layer (dashed curve). (c) Stripping of Cd deposited on Ag(111) at E = 0.72 V for 30 s as obtained scanning the potential from 1.0 to 0.2 V in ammonia buffer. (d) Comparison between the stripping peaks of Cd from CdS deposited on Ag(111) with one ECALE cycle as obtained by performing the anodic stripping after having eliminated S (solid curve) or before (dashed curve).The scan rate was 50 mV s 1 in (a), 10 mV s 1 in (b), (c), and 5 mV s 1 in (d). Na2C2O4, H3BO3, and NH4OH were used without further purification. Merk Suprapur NaOH and HClO4 and NH4OH were used to prepare the pH 9.6 ammonia buffer used as supporting electrolyte. The solutions were freshly prepared just before the beginning of each series of measurements. An automated deposition apparatus consisting of Pyrex solutions reservoirs, solenoid valves, a distribution valve, and a flow-cell was used under the control of a computer. The electrolytic cell was a Teflon cylinder with a 10 mm inner diameter, delimited by the working electrode (a disk of 12 mm diameter) on one side and the counter electrode on the other side. The inlet and the outlet for the solutions were placed on the side walls of the cylinder. The counter electrode was gold foil, and the reference electrode was a Ag/AgCl (sat. KCl) placed on the outlet tubing. Both the distribution valve and the cell were designed and realized in the workshop of our department.10 The solution is pushed into the cell by applying a pressure as low as 0.3 atm which determines a flow-rate of about 1 mL s 1. When the cell is filled, the pressure is no longer applied, so that the flow is stopped during ECALE depositions. A simple homemade software allows us to fill the cell with the different solutions. The silver single crystals were prepared according to the Bridgeman technique and polished by a CrO3-based procedure.19 21 Atomic Force Microscopy (AFM). Images were taken with a commercial instrument (PicoSPM, Molecular Imaging) in contact mode with a commercial Si3N4 cantilever (Nanosensors, Wezlar-Blankenfield).

’ RESULTS AND DISCUSSION Validation of the Method for the Attainment of Cd Layers on Ag(111). To validate the method, we chose the compound

CdS not only because it has been extensively studied in our group, but mainly since the stripping of the CdUPD layer on Ag(111) constitutes a sort of “finger prints” to which compare the experimental results. The compound CdS on Ag(111) is obtained by ECALE method simply alternating the underpotential deposition of S and Cd from solutions containing S(-II) or Cd(II) ions in ammonia buffer at 0.68 V, which was found to be the optimum potential value for the UPD of both elements.10 13 That means that going to more positive potentials in sulfide ion solutions results in bulk S deposition or, symmetrically, going to more negative potentials in cadmium ions solutions results in bulk Cd deposition. The formation of a compound from its constituent elements is an energetically favorable process. The negative free energy change involved in the formation of the compound is the principal reason for the occurrence of the UPD of Cd on the previously deposited S or vice versa. Figure 1a shows the underpotential deposition of Cd from 1 mM Cd(II) solutions on Ag(111) (solid line) and on Ag(111) covered by a UPD layer of S previously obtained by keeping the potential at 0.68 in a solution of 1 mM Na2S in ammonia buffer for 60 s (dashed line). As expected, due to the higher attractive interaction of Cd with S than with Ag, the whole UPD process is favored on S. Actually, on S-covered Ag(111) the UPD begins at more positive potentials and yields two distinct peaks, probably ascribable to S overlayers of different structure. The structure of Cd deposited at 0.4 V, i.e., after the more positive peak, has 11705

dx.doi.org/10.1021/la202174j |Langmuir 2011, 27, 11704–11709

Langmuir √ √ been found to be ( 7  7)R19.1°.22 The very high Cd Cd distances involved, 7.6 ( 0.8 Å, justify a further structural rearrangement that is probably responsible for the occurrence of the second UPD peak. Underpotential deposition of Cd on Ag(111) had been thoroughly investigated in Na2SO423,24 and in NaCl.25 In both media, the early stage of UPD formation, i.e., at high underpotentials, corresponds to the formation of an ordered phase, whereas surface Ag Cd alloying occurs at lower underpotentials (ΔE < 50 mV). The anodic dealloying leads to the formation of a number of pits and defects that in any case tend to disappear quickly at high underpotentials.25 The charge values reported for the overlayers formed at high underpotentials are slightly different (120 μC cm 2 in ref 24 and 150 μC cm 2 in ref 25), but in both cases, they are much lower than the charge value of 420 μC cm 2 associated with a close-packed Cd monolayer.23 The UPD process is surface limited, and the global charge involved in the process is obtained by integrating the stripping peaks of Cd deposited at 0.68 for one minute and then dissolved by sweeping the potential from 0.68 to 0 V in a 1 mM Cd(II) solution in ammonia buffer (Figure 1b). This charge is notably higher on S/Ag (271 μC cm 2, dashed line) than on Ag (160 μC cm 2, solid line). A similar discrepancy is observed comparing the charge, 192 μC cm 2, involved in the UPD of Cd on S-covered Au(111), with that, 122 μC cm 2, involved in the naked Au(111).26 All Cd eventually exceeding the charge associated with the UPD peak is dissolved at more negative potentials since it refers to the “bulk” Cd. Bulk deposition is not surface limited, and therefore, the charge associated with the stripping of a metal deposited in the bulk depends on the kinetics of deposition. Figure 1c shows the stripping of Cd deposited on Ag(111) at E = 0.72 V for 30 s as obtained by scanning the potential from 1.0 to 0.2 V in ammonia buffer. The broad peak A1 is the stripping peak of bulk deposited Cd, whereas the sharp peak A2 coincides with the peak of the Cd layer deposited at underpotential on Ag(111) reported in Figure 1b, apart from the small shift of the redox potential due to the larger Cd concentration in solution. The electrochemical characterization of CdS deposits obtained by ECALE is usually carried out, in situ, by scanning the potential to values where the deposits are dissolved at a sweep rate low enough to ensure the complete dissolution. Note that this is just the opposite operation performed in the preceding ECALE growth of the compound CdS, but the potentials involved are very different since they have to account for the free energy change involved in the compound formation. Thus, for example, the amount of Cd deposited in a given number of cycles is quantitatively determined from the charge involved in the anodic stripping of CdS, and the corresponding amount of S is determined in the subsequent cathodic stripping. It must be noted that, once all of the metallic element has been stripped anodically, the remaining sulfur layers, except for the first one, behave like bulk sulfur; hence, during the following reductive stripping they are reduced at more positive potentials than the SUPD layer, which is the first sulfur layer in contact with the silver substrate.10 The symmetrical procedure, that is, the cathodic removal of the sulfur layers, allows us to obtain metal deposits of different thicknesses that can be employed for practical purposes and is the basis of the proposed SEBALD method. To asses the validity of the proposed method, it is necessary to check the amount of Cd remaining after having eliminated S. To this end, after having deposited CdS the electrode is kept at 1.6 V

ARTICLE

Figure 2. Stripping curves of Cd as obtained after having eliminated all S from deposits obtained with 1, 2, 5, and 10 ECALE cycles. The scan rate was 10 mV s 1.The inset is the plot of the charge involved in the oxidative stripping of Cd as a function of the number of ECALE cycles.

for one minute to be sure that S was completely reduced to S(-II) which is then eliminated by washing the cell with the supporting electrolyte. In fact, the oxidative underpotential deposition of S from sulfide ions solutions on Ag(111) starts to take place at potentials as negative as 1.35 V and, going toward more negative potentials, yields random adsorption that then evolves in ordered overlayers of increasing coverage.27 Figure 1d shows the comparison between the stripping peaks of Cd from CdS deposited on Ag(111) with one ECALE cycle as obtained by performing the anodic stripping after having eliminated S (solid curve) or before (dashed curve). The ECALE cycle was the sequence S/Cd/S, with a final S layer usually added to exert a sort of capping action on the Cd layer. As shown in Figure 1a and b, on S the charge involved in the CdUPD is higher than that on silver; therefore, once all S has been eliminated, the amount of Cd that exceeds the charge of the UPD on silver must be released at the potentials of bulk Cd stripping shown in Figure 1c. On the other hand, the dashed curve is obtained without having eliminated S, and in this case, the potential of Cd redissolution occurs at potentials more positive since the negative free energy change involved in CdS formation makes Cd dissolution much more difficult. Incidentally, the procedure of stripping Cd from progressively thicker deposits of CdS has been extensively used in our group for the electrochemical characterization of the compound grown by ECALE.10 It must be noted that, despite the strong difference in shape, integration of both curves of Figure 1d yields roughly the same charge of 435 μC cm 2. This charge was expected to coincide with that of the dotted curve of Figure 1b, since it refers to a single layer of Cd. On the contrary, it is much higher and we tentatively attribute the discrepancy to a different degree of charge transfer associated with the presence/absence of the final S layer, since the underpotential deposition phenomena do not necessarily imply a total charge transfer. However, what is really important is that identical deposits S/Cd/S give the same charge value. Of course, this is a crucial point, since to assess the validity of the proposed method, the amount of Cd must not depend on the order of its stripping from CdS. At the same time, the strong difference in shape of the solid line from the dashed one ensures that 11706

dx.doi.org/10.1021/la202174j |Langmuir 2011, 27, 11704–11709

Langmuir

ARTICLE

Figure 4. Cyclic voltammograms obtained from 1 mM Co(II) in pH 9.6 ammonia buffer solutions on Ag(111) covered by a SUPD layer. The scan rate was 50 mV s 1.

Figure 3. Three-dimensional AFM images of the free Ag(111) surface (A) and with Cd SEBALD electrodeposition as obtained from a deposit of CdS grown with 10 ECALE cycles (B).

all S has been eliminated: otherwise, as we also experimentally observed, part of Cd would be redissolved at more positive potentials giving an additional stripping peak. The method can be applied to obtain Cd deposits of increasing thickness. Figure 2 shows the stripping curves of Cd as obtained after having eliminated all S from deposits obtained with 1, 2, 5, and 10 ECALE cycles. The inset shows that the charge increases linearly with the number of cycles used for CdS deposition, thus indicating that the same amount of compound is deposited in each cycle, which is typical of layer-by-layer growth. The peculiarity of Figure 2 is the invariance of the most positive peak that identifies the CdUPD layer on Ag (111), whereas all Cd exceeding the UPD

layer is redissolved at more negative potentials. Thus, the shape of the stripping peaks of Cd also suggests that the metal remaining after removal of S is a 2D/3D skeleton. The electrochemical characterization was completed with a morphological investigation. Figure 3 shows the three-dimensional AFM images of the free Ag(111) surface (Figure 3A) and with Cd SEBALD electrodeposition as obtained from a deposit of CdS grown with 10 ECALE cycles (Figure 3B). Figure 3B shows the formation of well-defined clusters of Cd confined on the Ag(111) surface, thus indicating the achievement of a bimetallic surface. The measurements were performed on different regions of the sample to ensure that the clusters were homogeneously distributed over the whole surface. The presence of clusters homogeneously distributed over the whole surface had been also detected from the early stages of the CdS formation, as showed by ex situ AFM measurements performed to study the morphological evolution from the bare silver crystal to deposits formed with an increasing number of cycles up to 200.11 Possibility of Extending the Method to Metals That Are Not Deposited at Underpotential. As stated before, monolayers of metals like Co, Ni, and Fe are receiving increasing interest in view of their potential application in electrocatalysis, and the problem of limiting their deposition is crucial since they are not deposited at underpotential. However, due to the negative free energy change involved in the formation of the corresponding sulfides, all of them can be deposited at underpotential on a S-covered Ag(111). As an example, Figure 4 shows the underpotential deposition of Co from 1 mM Co(II) solutions in pH 9.6 ammonia buffer on Ag(111) covered by a UPD layer of S previously obtained by keeping the potential at 0.68 in a solution of 1 mM Na2S in ammonia buffer for 60 s. In this supporting electrolyte, bulk Co electrodeposition, not shown in the figure, occurs at potentials more negative than 0.7 V. Similar behavior is presented by Ni28 and Fe. Figure 5 is analogous to Figure 1d and shows the comparison between the stripping peaks of Co (a), Fe (b), and Ni (c) from the corresponding sulfides deposited on Ag(111) with one ECALE cycle. Each figure shows the anodic stripping of the metal after having eliminated S (solid curve) or before 11707

dx.doi.org/10.1021/la202174j |Langmuir 2011, 27, 11704–11709

Langmuir

ARTICLE

Figure 6. Stripping curves of Co as obtained after having eliminated all S from deposits obtained with 1, 2, 3, and 5 ECALE cycles. The scan rate was 10 mV s 1 .The inset is the plot of the charge involved in the oxidative stripping of Co as a function of the number of ECALE cycles.

Figure 5. Comparison between the stripping peaks of Co (a), Fe (b), and Ni (c) from the corresponding sulfides deposited on Ag(111) with one ECALE cycle under the experimental conditions described in the text. Each figure shows the anodic stripping of the metal after having eliminated S (solid curve) or before (dashed curve). The scan rate was 5 mV s 1 in (a) and 10 mV s 1 in (b), (c).

(dashed curve). As in the case of Cd, S was eliminated by applying a potential sufficiently negative to be sure that S was completely reduced to S(-II) which is then eliminated by washing

the cell with the supporting electrolyte. Due to the different redox potentials, each system requires different experimental conditions in order to optimize the range of potential deposition/redissolution. Thus, Co is deposited at E = 0.65 V on S-covered Ag(111) from pH 9.6 ammonia buffer solutions, and then, it is redissolved in the same solution after having eliminated S (solid curve). On the contrary, the stripping of Co before S occurs at potentials more positive where part of the charge involved in Co stripping could be masked by silver oxidation. In this case, we performed Co stripping in 0.1 M Na2C2O4, which is a more strong complexing medium for Co that, consequently, is redissolved at less positive potentials (dashed curve). On the other hand, we could not observe Co stripping in oxalate after having eliminated S, since in this case it would occur at too negative potentials. With analogous considerations, we deposited Fe on S-covered Ag(111) in NaClO4 and performed its stripping in NaClO4 after having eliminated S (solid curve), or in 0.1 M Na2C2O4 if the stripping of Co is performed before S (dashed curve). Finally, we deposited Ni on S-covered Ag(111) from a solution of 5 mM NiCl2 in 0.1 M KCl + 0.1 M H3BO3 (pH = 6.5) at E = 0.6 V for 60 s28 and performed its stripping, either before or after S, in pH 9.6 ammonia buffer. Apart from the different experimental conditions described above and from the different values of charge involved, the peculiarity of the figure is that, as in the case of Cd, the charge involved in the stripping of the metal from the corresponding sulfide does not depend on the order of element redissolution. Strangely, the charge of Co is much lower than the charge values measured for Cd, Ni and Fe: this is surely due to limitations in the deposition process and is not connected to the different ability of the electrolyte to strip the metal. In fact, the stripping curves of Figure 1a have been recorded in different supporting electrolyte. Finally, Figure 6 shows the stripping curve of Co as obtained after having eliminated all S from deposits obtained with 1, 2, 3, and 5 ECALE cycles. The inset shows that the charge increases linearly with the number of cycles used for CoS deposition, thus indicating that the same amount of compound is deposited in each cycle, which is typical of layer-by-layer growth. Therefore, as in the case of CdS, the method can be applied to obtain Co 11708

dx.doi.org/10.1021/la202174j |Langmuir 2011, 27, 11704–11709

Langmuir deposits of increasing thickness. It must be noted that the stripping peak at more positive potentials is much smaller than the analogous peak observed for Cd and, more important, its area is not constant, which is consistent with the absence of underpotential deposition phenomena already stated.

’ CONCLUSION An effective way to obtain well-controlled atomic level deposition of a metal M on a foreign substrate is proposed here for the first time. The method is easily validated using the compound CdS that allows direct comparison with the Cd monolayer deposited at underpotential on Ag(111). However, the method seems to be suggested particularly for those metals that are not deposited at underpotential on the substrate. Underpotential deposition is a surface limited phenomenon that generally yields no more than one monolayer. As a consequence, the amount of deposition is well-controlled at the nanometer scale. Such level of control is not achievable in overpotential deposition through the application of Faradays laws, even trying to limit the extent of deposition operating at very low overpotentials, that is, at very low values of current. The experimental results show that SEBALD method can be divided in the following steps: (1) One or more layers of metal sulfide are deposited by ECALE on Ag(111). (2) Sulfur is then eliminated by applying a potential sufficiently negative. The electrochemical characterization shows that the amount of remaining metal is proportional to the number of ECALE cycles used for deposition. The premises of the method suggest that, more generally, a monolayer of a metal, M, could be deposited on a foreign substrate, if M is able to form a compound, MxAy, with an anion, A, that could be then eliminated applying a potential sufficiently negative. When the anion A is a chalcogen, more layers of the compound MA are easily obtained by ECALE. Therefore, as shown for CdS, thicker layers of metals can be obtained. The AFM images show that the metal remaining after having eliminated the anion forms well-defined clusters on the silver surface, thus indicating the achievement of a bimetallic surface. ’ AUTHOR INFORMATION

ARTICLE

(6) Colletti, L. P.; Flowers, B. H., Jr.; Stickney, J. L. J. Electrochem. Soc. 1998, 145, 1442–1449. (7) Gichuhi, A.; Boone, B. e.; Demir, U.; Shannon, C. J. Phys. Chem. B 1998, 102, 6499. (8) Stickney, J. L. Electroanal. Chem. 1999, 21, 75–209. (9) Gichuhi, A.; Boone, B. E.; Shannon, C. J. Electroanal. Chem. 2002 522, 21. (10) Innocenti, M.; Pezzatini, G.; Forni, F.; Foresti, M. L. J. Electrochem. Soc. 2001, 148 (5), C357–C362. (11) Innocenti, M.; Cattarin, S.; Cavallini, M.; Loglio, F.; Foresti, M. L. J. Electroanal. Chem. 2002, 532, 219–225. (12) Loglio, F.; Innocenti, M.; Pezzatini, G.; Foresti, M. L. J. Electroanal. Chem. 2004, 562, 117–125. (13) Innocenti, M.; Cattarin, S.; Loglio, F.; Cecconi, T.; Seravalli, G.; Foresti, M. L. Electrochim. Acta 2004, 49, 1327–1337. (14) Fernandez, J. L.; Walsh, D. A.; Bard, A. J. J. Am. Chem. Soc. 2005, 127, 357–365. (15) Wang, Y.; Balbuena, P. B. J. Phys. Chem. B 2005, 109, 18902–18906. (16) Loglio F.; Lastraioli E.; Bianchini C.; Fontanesi C.; Innocenti M.; Lavacchi A.; Vizza F.; Foresti M. L.; ChemsUsChem [Online Early Access] DOI: 10.1002/cssc.201100092; Published Online: June 3, 2011. (17) Brankovic, S. R.; Wang, J. X.; Adzic, R. R. Surf. Sci. 2001, 474, L173–L179. (18) Thambidurai, C.; Kim, Y.-G.; Stickney, J. L. Electrochim. Acta 2008, 53, 6157–6164. (19) Hamelin, A. in Modern Aspects of Electrochemistry, Conway, B. E., White, R. E., Bockris, J. O’M.; Plenum Press: New York, 1985; Vol 16, p 1. (20) Kurasawa T. 1960 Patent Japan 35:5619. (21) Foresti, M. L.; Capolupo, F.; Innocenti, M.; Loglio, F. Cryst. Growth Des. 2002, 2, 73–77. (22) Foresti, M. L.; Pezzatini, G.; Cavalllini, M.; Aloisi, G.; Innocenti, M.; Guidelli, R. J. Phys. Chem. B 1998, 102, 7413–7420. (23) Bort, H.; Juettner, K.; Lorenz, W. J.; Staikov, G. Electrochim. Acta 1983, 28, 993–1001. (24) Garcia, S. G.; Salinas, D. R.; Staikov, G. Surf. Sci. 2005, 576, 9–18. (25) Jovic, V. D.; Jovic, B. M. Electrochim. Acta 2002, 47, 1777–1785. (26) Demir, U.; Shannon, C. Langmuir 1996, 12, 6091–6097. (27) Aloisi, G. D.; Cavallini, M.; Innocenti, M.; Foresti, M. L.; Pezzatini, G.; Guidelli, R. J. Phys. Chem. B 1997, 101, 4774–4780. (28) Loglio, F.; Innocenti, M.; Jarek, A.; Caporali, S.; Pasquini, I.; Foresti, M. L. J. Electroanal. Chem. 2010, 638, 15–20.

Corresponding Author

*E-mail: foresti@unifi.it, Tel. +39-055-4573107, Fax: +39- 0554573385.

’ ACKNOWLEDGMENT The authors are grateful to Mr. Ferdinando Capolupo for the preparation of the silver single crystal electrodes. It is acknowledged the financial support from the MIUR (Italy) for the PRIN 2008 project prot. 2008N7CYL5. ’ REFERENCES (1) Kolb, D.M. in Advances in Electrochemistry and Electrochemical Engineering, Gerischer, H.; Tobias, H., Eds.; John Wiley: New York, 1978; Vol 11, p 125. (2) Gregory, B. W.; Stickney, J. L. J. Electronanal. Chem. 1991, 300, 543. (3) Colletti, L. P.; Teklay, D.l; Stickney, J. L. J. Electronanal. Chem. 1994, 369, 145–152. (4) Baoming, H. M.; Lister, T. E.; Stickney, J. L. in Handbook of Surface Imaging and Visualization, Hubbard, A. T, Ed.; 1995; pp 75 91. (5) Boone, B. E.; Shannon, C. J. Phys. Chem. 1996, 100, 9480–9484. 11709

dx.doi.org/10.1021/la202174j |Langmuir 2011, 27, 11704–11709