Journal of Electroanalytical Chemistry 649 (2010) 83–88
Contents lists available at ScienceDirect
Journal of Electroanalytical Chemistry journal homepage: www.elsevier.com/locate/jelechem
Peptide molecular junctions: Electron transmission through individual amino acid residues Joanna Juhaniewicz, Slawomir Sek * Department of Chemistry, University of Warsaw, Pasteura 1, 02-093 Warsaw, Poland
a r t i c l e
i n f o
Article history: Received 1 October 2009 Received in revised form 25 January 2010 Accepted 27 January 2010 Available online 2 February 2010 Dedicated to professor Jacek Lipkowski on the occasion of his 65th birthday
a b s t r a c t Three short chain peptides of general formula Cys–AA–CSA, where CSA is cystamine, AA is Gly, Ala or Pro, were synthesized and used for the determination of the electron transmission through individual amino acid residues. The conductance measurements were carried out using STM-based molecular junction method. The results of our comparative study indicate that the ability of these molecules to mediate electron transfer is not significantly affected by the presence of the methyl side chain in alanine or the cyclic structure of proline. Thus the efficiency of electron transmission is determined mostly by the peptide backbone structure. Ó 2010 Elsevier B.V. All rights reserved.
Keywords: Peptide Electron transfer Molecular junction Scanning tunneling microscopy Monolayers Gold electrodes
1. Introduction Amino acids are fundamental building blocks of peptides and proteins. Their chemical properties and the particular sequence of the residues determine the three dimensional structure and the biological activity of the proteins. Therefore, amino acids seem to be good candidates for the design and synthesis of artificial functional molecules which display desired properties. By careful selection of the amino acid sequence within the peptide, it is potentially possible to obtain the molecule with specific secondary structure, polarity, reactivity, etc. This way, peptides can be suitably tailored for the applications in electrochemical (bio)sensors, nanodevices and molecular electronics [1–5]. In all aforementioned cases, the efficient action of such systems involves the use of the molecules with specific, well-defined electronic conductance. It is known that peptides can act as mediating bridges for long range electron transfer and the efficiency of this process may be affected by the details of the amino acid sequence, the secondary structure of the peptide and the presence of the hydrogen bonds [6–11]. Thus the fundamental problem is how to design the suitable functional molecule to attain desired structure and * Corresponding author. Tel.: +48 228220211. E-mail address:
[email protected] (S. Sek). 1572-6657/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2010.01.029
the electronic conductance. In order to solve this problem, first we need to gain our knowledge about the electronic conductance through individual amino acid residues. This may help to differentiate between amino acids which are good ‘‘conductors” and those acting mostly as a structural foundation for peptide. Therefore, in this paper, we report the results of our comparative studies on electron transfer through individual amino acid residues, i.e. Gly, Ala and Pro, incorporated into the short peptides. The electron transmission through single molecules was investigated using STM-based molecular junction method. Our approach involved the entrapment of peptide molecules between the substrate and the tip of a scanning tunneling microscope. Several groups have demonstrated the use of similar methods for the conductance measurements of alkanedithiols, viologens, diamines, dicarboxylic acids, diisonitryles, peptides and DNA [12–20]. In this study, we used self-assembled monolayers of peptides of general formula Cys–AA–CSA where CSA is a cystamine linker, AA is either glycine (Gly), alanine (Ala) or proline (Pro), and Cys is a cysteine residue (see Scheme 1). By changing the amino acid in the middle section of the molecule, we were able to compare the efficiency of electron transmission through each residue, i.e. Gly, Ala or Pro. This way, it was possible to verify how the presence of the side chain in alanine or the cyclic system of proline affects the electron transfer properties of the peptides.
84
J. Juhaniewicz, S. Sek / Journal of Electroanalytical Chemistry 649 (2010) 83–88
Scheme 2. Reaction scheme for the synthesis of the peptides.
MgSO4, filtered and then the solvent was removed under the vacuum. The resulting solid was treated with 10 ml of trifluoroacetic acid for 30 min in an ice bath. Further, the solvent was removed under vacuum and a mixture of 15 ml of triethylamine and 50 ml of dichloromethane was added. The solution was left for 30 min in ambient conditions. Subsequently, the mixture of 5.5 mmol of (Boc)-L-Cys(Acm), 5.5 mmol of HOBt and 5.5 mmol of EDC in 50 ml of dichloromethane was prepared and it was stirred for 30 min in an ice bath. Then, both solutions were combined and the resulting mixture was stirred for a further 12 h. In the next step, the reaction mixture was washed by 50 ml portions of saturated NaHCO3, 10% citric acid, saturated NaHCO3 and water. The organic phase was dried over MgSO4, filtered and finally the solvent was removed under vacuum. The crude solid was purified using flash column chromatography on silica gel. The identities of the products were confirmed by mass spectrometry. (Boc)Cys(Acm)-Ala-CSA, ESI-MS: m/z = 865.6 [M+Na]+, Mcalc = 842.2 g/mol; (Boc)Cys(Acm)-Gly-CSA, ESI-MS: m/z = 837.7 [M+Na]+, Mcalc = 814.2 g/mol; (Boc)Cys(Acm)-Pro-CSA, ESI-MS: m/z = 917.7 [M+Na]+, Mcalc = 895.2 g/mol.
2.2. Monolayer preparation
Scheme 1. Structures of the short peptides with Acm and Boc protecting groups marked out by gray and white rectangles respectively.
The monolayers were prepared on gold substrates purchased from Arrandee (Werther, Germany). Prior to the monolayer deposition, each substrate was flame annealed in order to generate atomically flat Au(1 1 1) terraces. The adsorption of the peptides was carried out by self-assembly from 1 mM ethanolic solutions. It should be noted, that the presence of the Acm protecting group at cysteine residue leads to the preferential adsorption of the peptides through sulfur atoms at cystamine [11,18]. This way, the molecules forming the assembly are uniformly oriented with the N-terminus located in the external plane of the monolayer. After 24 h of being soaked, the samples were rinsed with ethanol and water and dried in an Ar stream. At this stage, the samples were ready to use in electrochemical experiments.
2. Experimental section All chemicals and starting compounds for syntheses were purchased from Aldrich, Fluka and POCh (Gliwice, Poland) and were used as received. 2.1. Synthesis of (Boc)Cys(Acm)–AA–CSA The route of the synthesis is shown in Scheme 2. An equivalent amounts of 5.5 mmol of (Boc)-L-AA-OH (where AA = Gly, Ala or Pro), HOBt and EDC were dissolved in 40 ml of dichloromethane and stirred for 30 min at 0 °C in an ice bath. Subsequently, the solution of 2.75 mmol of cystamine hydrochloride and 5.5 mmol of triethylamine in 20 ml of dichloromethane was added. The mixture was stirred for 12 h and then extracted with 50 ml portions of saturated aqueous solution of NaHCO3, 10% citric acid, saturated NaHCO3 and water. The organic phase was dried over anhydrous
2.3. Electrochemical measurements The measurements of the capacitance and the desorption charge were carried out at room temperature, in a three-electrode cell with monolayer modified substrate, Ag/AgCl (sat. KCl) and Pt wire serving as working, reference and counter electrode respectively. The supporting electrolyte was either 0.1 M KNO3 (for capacitance measurements) or 0.1 M KOH (for desorption experiments). The solutions were prepared with distilled water passed through Milli-Q purification system and its final resistivity was 18.2 MX cm. The oxygen was removed from the electrolyte solution with an Ar stream. All electrochemical measurements were performed using CHI 750B bipotentiostat (CH Instruments Inc., Austin, TX).
J. Juhaniewicz, S. Sek / Journal of Electroanalytical Chemistry 649 (2010) 83–88
2.4. Formation of junctions In order to prepare the monolayer modified samples for the junction formation, they were subjected to Acm deprotection procedure [11,18,21]. The peptide modified electrodes were placed in deoxygenated water, and then the pH was adjusted to 4.0 using acetic acid and ammonia. The resulting solution was continuously bubbled with argon. Next, mercury (II) acetate was introduced into the mixture and it was left for 45 min. Subsequently, b-mercaptoethanol was added and after 5–10 min the substrates were removed from the solution and rinsed thoroughly with water and dried with argon stream. By removing Acm moieties from cysteine residues, it was possible to obtain free thiol groups located in the external plane of the assembly. It should be noted that Boc moiety protecting the amino group in cysteine residue was left in order to avoid formation of the Au–N contact between the molecule and the STM tip [21]. Thus, during the junction formation, only free –SH groups interact with the metallic STM tip and form covalent Au– S contact. The efficacy of the deprotection was verified by testing procedure, which involved soaking of the samples in an aqueous solution of gold nanoparticles (nominal diameter 5 nm) for 2 h. Then the samples were thoroughly rinsed with water, dried and imaged using scanning tunneling microscope. STM experiments were performed with MultiMode SPM connected to Nanoscope IIIa controller (Veeco Instruments). The data were collected under ambient conditions in air. Electrochemically etched tungsten tips were used for imaging while mechanically cut gold tips were used in the conductance measurements. In order to form a junction, a gold STM tip was placed at arbitrary chosen location on the sample surface at a distance defined by the values of the tunneling current and bias voltage. The separation between the tip and the sample surface was similar for all measurements. The value of the tunneling current at given bias voltage was always set to obtain the initial tip-substrate gap resistance about 0.1 GX. The STM imaging of the samples after the series of the junction experiments indicated that the resistance of the gap was small enough to bring the tip into the physical contact with the monolayer, however, it was sufficiently large to avoid crushing the tip. Under these conditions, the tip was expected to interact with deprotected thiol groups of the molecules adsorbed on the substrate and, as a result, the chemical Au– S bonds were formed. Further, the feedback was disabled and the tip was lifted at the rate of 17 nm/s while keeping constant x–y position. During the retraction, the current was recorded as a function of the distance separating the tip from the sample. The conductance of the junctions was determined in the range of ±500 mV. 3. Results and discussion 3.1. Electrochemical characteristics of peptide monolayers The quality of the peptide monolayers was verified by electrochemical measurements of the capacitance and the desorption charge. Double layer capacitance of the modified electrodes was determined using electrochemical impedance spectroscopy. The results are collected in Table 1. As can be seen, the values of the capacitances obtained for monolayer modified electrodes are about
Table 1 The properties of the peptide monolayers. Monolayer
Capacitance from EIS (lF cm 2)
Cys(Acm)–Gly–CSA Cys(Acm)–Ala–CSA Cys(Acm)–Pro–CSA Bare Au
6.75 ± 1.12 6.38 ± 1.24 6.93 ± 0.50 19.0 ± 1.0
Peak potential for reductive desorption (V) 1.16 ± 0.01 1.20 ± 0.01 1.11 ± 0.01 –
85
three times lower comparing to bare gold. This shows that the surface is indeed modified with adsorbed dielectric material. However, the values of the capacitance are higher than those reported for n-alkanethiolate assemblies [22]. The difference is reasonable if we consider the polarity of the alkane and peptide chains. In the latter case, higher dielectric constant is expected. Moreover, the presence of amide bonds may contribute to the entrapment of water molecules within the assembly. As a result, the capacitance is relatively high. Small differences observed for particular peptides most likely reflect the different packing densities and the ordering of the molecules within the assembly, which also determine the permeability of the monolayers for water molecules or ions (i.e. the higher value of the capacitance, the higher permeability of the assembly). In order to confirm successful adsorption of the peptides on gold surface, we have also performed reductive desorption experiments [23]. This method is often used for the determination of the surface coverage for non-electroactive thiolate monolayers. However, as demonstrated by several groups this approach gives coverages overestimated by 15–20% [24–29]. This is related to the substantial charge which contributes to the desorption process when the monolayer is displaced by the electrolyte at the metal–solution interface. Therefore, we used this method only for qualitative assessment of the films. Fig. 1 presents the examples of the cyclic voltametric curves recorded for gold electrodes modified with peptide monolayers. When the potential of the peptide modified electrode is swept to negative value, a well-defined cathodic peak is observed, which corresponds to the reductive desorption of the molecules from gold surface. During the reverse scan, re-adsorption of the molecules occurs, which is reflected by an anodic peak. Such voltammetric behavior can be considered as indicative for the presence of the chemisorbed molecules attached to the gold surface via Au–S bond [23]. Fig. 1 shows that the desorption potentials are different among the peptides (see also Table 1). According to Porter’s interpretation the position of the desorption peak is determined by the potential which is required for sufficient ion flux to the electrode to support reductive reaction [23]. It means that the intermolecular interaction within the assembly is important factor here. As the intermolecular interactions between the molecules become more attractive, the film become more compact and the more negative potential is needed to attain sufficient potential gradient enabling the penetration of the monolayer by electrolyte ions and solvent. Thus, the differences between desorption potentials of the peptides reflect different packing densities of the molecules within the monolayer and as a consequence different strength of intermolecular interactions. Based on this, we can conclude that the compactness of the films increases in a sequence Pro < Gly < Ala. It should be noted that the analysis of the capacitance leads to the same conclusion. The highest permeability of the monolayer prepared with the peptide containing proline residue can be explained by the cyclic structure of this amino acid, which prevents tight packing of the molecules. Considering only spatial requirements of the amino acids introduced into the middle part of the peptide backbone, one can expect that Cys–Ala–CSA monolayer will be less compact than Cys–Gly– CSA, since the alanine residue has a methyl side group. However, the reductive desorption results and the capacitance measurements lead to opposite conclusion. This can be explained by the higher hydrophobicity of alanine comparing to glycine. As a consequence, the monolayer of Cys–Ala–CSA is less permeable for hydrated ions and solvent molecules. 3.2. Conductance of peptides As it was mentioned in the experimental section, prior to the junction formation, the substrates were subjected to Acm deprotection procedure. This step is crucial since the presence of free
86
J. Juhaniewicz, S. Sek / Journal of Electroanalytical Chemistry 649 (2010) 83–88
Fig. 1. Cyclic voltammetric curves recorded for gold electrodes modified with selfassembled monolayers of Cys–Gly–CSA (blue line), Cys–Ala–CSA (black line) and Cys–Pro–CSA (red line). Supporting electrolyte: 0.1 M KOH. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
thiol groups in the external plane of the monolayer is necessary to form the Au–S linkage between the molecules and the STM tip. Therefore, the effectiveness of the deprotection was verified by simple testing procedure. The modified substrates, either deprotected or protected, were immersed in a suspension of colloidal gold, thoroughly rinsed with water and then imaged with STM. The images of exemplary samples are shown in Fig. 2. In the case of protected sample (see Fig. 2A), the surface is relatively flat with characteristic steps separating the terraces. Dark spots correspond to etch pits, which are characteristic for gold surface modified with thiolate adlayer. Clearly, gold nanoparticles are absent. It suggests that the interaction between nanoparticles and protected thiol groups is weak under these conditions. Fig. 2B presents the STM image taken for deprotected sample. In this case, the surface is covered with large number of gold nanoparticles. This confirms that the deprotection procedure is effective and terminal free thiol groups are able to form Au–S linkage. As a result, the stable layer of gold nanoparticles is formed on the top of the assembly. Based on these findings, we concluded that the deprotected samples are suitable for use in a molecular junction system. The conductance measurements were performed for the bias voltages from 0.5 V to +0.5 V. Fig. 3 shows a representative current–distance curves recorded with bare gold (gray line) and the
peptide modified substrate (black line). Each curve was recorded in a single junction experiment. For bare gold, we observed fast exponential decay of the tunneling current with increasing distance between the tip and the metal surface, which is indicative for tunneling through empty gap. The substrates modified with peptide monolayers exhibit significantly different behavior. At the initial stage, the tunneling current decays exponentially with increasing distance between the tip and the electrode. Then, it is followed by the current plateau, which is attributed to the electron flow through the molecule or molecules bridging the tip and the substrate. The current remains constant as long as the molecule is bonded to the gold contacts, but once the contact is broken, the current drops suddenly. The inset in Fig. 3 presents the histogram constructed on the basis of recorded current–distance curves obtained at particular bias voltage. The histograms were constructed using at least 100 individual current–distance curves, however, we used only those which displayed well-defined current steps. The percentage of the current–distance curves which fulfill this requirement was in the range of 30–40% of all recorded data. The consecutive maxima can be ascribed to the conductance through one or two molecules trapped within the junction. The current values corresponding to the first peak on histograms, i.e. ascribed to the conductance of a single molecule, were used for further analysis. The conductance values for all peptides are shown in Table 2. It is clear that these molecules mediate the electron transfer with similar efficiency. Slightly higher value found for proline may result from its cyclic structure, which can change the conformation of the peptide molecule and, consequently, shorten the distance between the sulfur atoms. The distances separating the terminal thiol groups in particular peptides are shown in Table 2. Indeed the length calculated for the molecule with proline residue is slightly shorter comparing to other systems. The comparison of the conductances presented in Table 2 with those reported earlier by Tao’s group for the peptide of similar length containing Gly residue, clearly shows that our results are significantly higher [30]. However, several theoretical and experimental studies have shown that different values of conductance can be observed for the same type of the molecule [31–33]. For example, recently Hihath and Tao have observed different conductance values for the same amino acids [34]. This is related to the fact that molecular conductance is strongly affected by the metal–molecule contact morphology and by atomic structure of the substrate surface. Such influence was demonstrated experimentally by Haiss et al. [33]. The results of the theoretical studies reported by Bautista show also that the electrical behavior of oligoglycines depends on the number of metal atoms that directly
Fig. 2. STM images obtained for gold substrates modified with Cys–Ala–CSA monolayer before (A), and after (B) Acm deprotection procedure and subsequent immersion in an aqueous suspension of gold nanoparticles. Size: 100 100 nm2 for both images.
J. Juhaniewicz, S. Sek / Journal of Electroanalytical Chemistry 649 (2010) 83–88
87
Fig. 5. The structures and the dipole moments for Cys–Gly–CSA (A), Cys–Ala–CSA (B) and Cys–Pro–CSA (C). The geometry of the molecules was optimized with AM1 method using HyperChem 8.0 software. In order to simplify the calculations, we considered simple thiol forms of the molecules and also omitted Boc protecting groups. Fig. 3. Examples of current–distance curves recorded in individual junction experiments with bare gold (gray dotted line) and the substrate modified with Cys–Pro–CSA (black solid line). The inset presents the histogram constructed on the basis of current–distance curves recorded for Cys–Pro–CSA. Bias voltage: 0.4 V.
Table 2 Molecular junctions with peptides. Peptide
S-to-S distancea (Å)
Conductance (G0)
Cys–Gly–CSA Cys–Ala–CSA Cys–Pro–CSA
12.3 12.3 11.1
(3.6 ± 0.2) 10–5 (4.3 ± 0.2) 10 5 (5.2 ± 0.2) 10 5
a The distance separating terminal sulfur atoms in peptide. All distances were calculated using HyperChem 8.0 software after geometry optimization with AM1 method.
contact the molecule within the junction [32]. Another factor which affects the electron transmission through peptide is the conformation of the molecule. The latter may vary depending on surrounding environment or the presence of the solvent. On the other hand, our results are in qualitative agreement with those reported by Hihath and Tao for glycine and alanine [34]. As they have shown, the conductance values for these two amino acids are identical. Fig. 4 presents the dependence of the current as a function of the bias voltage for the junctions incorporating Cys–Pro–CSA
(blue), Cys–Ala–CSA (black) and Cys–Gly–CSA (red) bridges. As can be seen, the i–V curves for all cases are similar and exhibit small, but visible asymmetry with lower current values recorded at negative bias voltage. At positive bias the electrons flow from the STM tip to the gold electrode, while at the negative bias the electrons flow in opposite direction, i.e. from substrate to the STM tip. Thus, the asymmetry of i–V curves indicates that electron transfer from the STM tip to gold electrode, i.e. from N-terminus to C-terminus of the peptide, is slightly more efficient than in opposite direction. Similar effect was observed by Xiao and coworkers and as suggested by these authors it may result from asymmetrical structure and the electric dipoles of the peptide molecules trapped within the junctions [30]. Indeed the current rectification is expected when the kinetics of electron transfer is affected by the electric field generated by the molecular dipole. Such behavior was reported for helical peptides where the dipole moment is aligned with molecular axis, i.e. it is parallel to the direction of electron flow [3,18,35,36]. The effect of the electric field generated by the dipole is most pronounced under these conditions. However, the peptides used in this work do not adopt any particular secondary structure and their dipole moments are not aligned with the molecular axes (see Fig. 5). Therefore, the influence of the electric field generated by the dipole is rather small or even negligible. The other factors causing current rectification include the asymmetric metal–molecule contacts, the presence of the asymmetry in the chemical structure of the bridge or the presence of chemical substituent groups which influence the charge redistribution within molecular system [37–39]. Since both ends of the peptides molecules are attached to metal electrodes by Au–S bonds, only two last factors can be considered as responsible for small current rectification observed here.
4. Conclusions
Fig. 4. Current–voltage curves obtained from the junction experiments with Cys– Gly–CSA (blue line), Cys–Ala–CSA (black line) and Cys–Pro–CSA (red line). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
The results of our conductance measurements clearly indicate that three amino acid residues under study, i.e. Gly, Ala and Pro, show similar ability to mediate electron transfer. The overall electron transmission through short peptides studied here is determined chiefly by the backbone structure and the contributions either from the side chain of alanine or from the cyclic system in proline are small. Thus, considering the design of functional peptides which efficiently mediate electron transfer, there is no significant gain in using one of these three amino acids, unless it is necessary to attain particular structural features of the peptide.
88
J. Juhaniewicz, S. Sek / Journal of Electroanalytical Chemistry 649 (2010) 83–88
Acknowledgements This work was financially supported by the Ministry of Science and Higher Education (Grant No. N N204 138137). The project was also operated within the Foundation for Polish Science MPD Programme co-financed by the EU European Regional Development Fund. References [1] J.J. Gooding, For Electrochemical Sensor Analysis, in: S. Alegret, A. Merkoçi (Eds.), Book Series Comprehensive Analytical Chemistry, vol. 49, Elsevier Press, Amsterdam, 2007, pp. 189–210. [2] R. de la Rica, E. Mendoza, L.M. Lechuga, H. Matsui, Angew. Chem. Int. Ed. 47 (2008) 9752. [3] S. Yasutomi, T. Morita, Y. Imanishi, S. Kimura, Science 304 (2004) 1944. [4] K. Kitagawa, T. Morita, S. Kimura, Angew. Chem. Int. Ed. 44 (2005) 6330. [5] J.M. Seminario, L. Yan, Y. Ma, J. Phys. Chem. A 109 (2005) 9712. [6] Y.K. Shin, M.D. Newton, S.S. Isied, J. Am. Chem. Soc. 125 (2003) 3722. [7] S. Antonello, F. Formaggio, A. Moretto, C. Toniolo, F. Maran, J. Am. Chem. Soc. 125 (2003) 2874. [8] T. Morita, S. Kimura, J. Am. Chem. Soc. 125 (2003) 8732. [9] S. Sek, A. Sepiol, A. Tolak, A. Misicka, R. Bilewicz, J. Phys. Chem. B 108 (2004) 8102. [10] H.B. Kraatz, I. Biedako-Amoa, S.H. Gyepi-Garbrah, T.C. Sutherland, J. Phys. Chem. B 108 (2004) 20164. [11] S. Sek, A. Misicka, K. Swiatek, E. Maicka, J. Phys. Chem. B 110 (2006) 19671. [12] B. Xu, N.J. Tao, Science 301 (2003) 1221. [13] W. Haiss, H. van Zalinge, S.J. Higgins, D. Bethell, H. Hobenreich, D.J. Schiffrin, R.J. Nichols, J. Am. Chem. Soc. 125 (2003) 15294. [14] Z. Li, B. Han, G. Meszaros, I. Pobelov, T. Wandlowski, A. Blaszczyk, M. Mayor, Faraday Discuss. 131 (2006) 121.
[15] L. Venkataraman, J.E. Klare, I.W. Tam, C. Nuckolls, M.S. Hybertsen, M.L. Steigerwald, Nano Lett. 6 (2006) 458. [16] F. Chen, X. Li, J. Hihath, Z. Huang, N. Tao, J. Am. Chem. Soc. 128 (2006) 15876. [17] X. Xiao, B. Xu, N. Tao, J. Am. Chem. Soc. 126 (2004) 5370. [18] S. Sek, K. Swiatek, A. Misicka, J. Phys. Chem. B 109 (2005) 23121. [19] E. Wierzbin´ski, J. Arndt, W. Hammond, K. Słowin´ski, Langmuir 22 (2006) 2426. [20] H. van Zalinge, D.J. Schiffrin, A.D. Bates, W. Haiss, J. Ulstrup, R.J. Nichols, ChemPhysChem 7 (2006) 94. [21] S. Sek, J. Phys. Chem. C 111 (2007) 12860. [22] R.P. Janek, W.R. Fawcett, A. Ulman, J. Phys. Chem. B 101 (1997) 8550. [23] C.A. Widrig, C. Chung, M.D. Porter, J. Electroanal. Chem. 310 (1991) 335. [24] T.W. Schneider, D.A. Buttry, J. Am. Chem. Soc. 115 (1993) 12391. [25] S. Imbayashi, M. Iida, D. Hobara, Z.Q. Feng, K. Niki, T. Kakiuchi, J. Electroanal. Chem. 428 (1997) 33. [26] T. Kakiuchi, H. Usui, D. Hobara, M. Yamamoto, Langmuir 18 (2002) 5231. [27] J. Kunze, J. Leitch, A.L. Schwan, R.J. Faragher, R. Naumann, S. Schiller, W. Knoll, J.R. Dutcher, J. Lipkowski, Langmuir 22 (2006) 5509. [28] T. Laredo, J. Leitch, M. Chen, I.J. Burgess, J.R. Dutcher, J. Lipkowski, Langmuir 23 (2007) 6205. [29] R. Aguilar-Sanchez, G.J. Su, M. Homberger, U. Simon, Th. Wandlowski, J. Phys. Chem. C 111 (2007) 17409. [30] X. Xiao, B. Xu, N. Tao, Angew. Chem. Int. Ed. 43 (2004) 6148. [31] X. Li, J. He, J. Hihath, B. Xu, S.M. Lindsay, N. Tao, J. Am. Chem. Soc. 128 (2006) 2135. [32] E.J. Bautista, L. Yan, J.M. Seminario, J. Phys. Chem. C 111 (2007) 14552. [33] W. Haiss, S. Martin, E. Leary, H. van Zalinge, S.J. Higgins, L. Bouffier, R.J. Nichols, J. Phys. Chem. C 113 (2009) 5823. [34] J. Hihath, N. Tao, Nanotechnology 19 (2008) 265204. [35] M.A. Fox, E. Galoppini, J. Am. Chem. Soc. 119 (1997) 5277. [36] K. Kitagawa, T. Morita, S. Kimura, J. Phys. Chem. B 109 (2005) 13906. [37] J.G. Kushmerick, D.B. Holt, J.C. Yang, J. Naciri, M.H. Moore, R. Shashidhar, Phys. Rev. Lett. 89 (2002) 086802. [38] Z.L. Gasyna, G.M. Morale, A. Sanchez, L. Yu, Chem. Phys. Lett. 417 (2006) 401. [39] K. Walczak, S.E. Lyshevski, Cent. Eur. J. Phys. 3 (2005) 555.
