Role of Surface Heterogeneity and Molecular Interactions in the

May 7, 2004 - The degree of surface heterogeneity at the SAMs increases as the number ... heterogeneous charge-transfer rate constants for MB immobili...
0 downloads 0 Views 265KB Size
Download PDF
5030

Langmuir 2004, 20, 5030-5037

Role of Surface Heterogeneity and Molecular Interactions in the Charge-Transfer Process through Self-Assembled Thiolate Monolayers on Au(111) G. Benı´tez, C. Vericat, S. Tanco, F. Remes Lenicov, M. F. Castez, M. E. Vela, and R. C. Salvarezza* Instituto de Investigaciones Fisicoquı´micas Teo´ ricas y Aplicadas (INIFTA), Universidad Nacional de La Plata-CONICET, Sucursal 4 Casilla de Correo 16, 1900 La Plata, Argentina Received December 23, 2003. In Final Form: February 20, 2004 A comparative study of charge-transfer processes from/to methyl-terminated and carboxylate-terminated thiolate-covered Au(111) surfaces to/from immobilized methylene blue (MB) molecules is presented. Scanning tunneling microscopy images with molecular resolution reveal the presence of molecular-sized defects, missing rows, and crystalline domains with different tilts that turn the thickness of the alkanethiolate SAM (the spacer) uncertain. The degree of surface heterogeneity at the SAMs increases as the number of C units (n) in the hydrocarbon chain decreases from n ) 6. Defective regions act as preferred paths for MB incorporation into the methyl-terminated SAMs, driven by hydrophobic forces. The presence of negativecharged terminal groups at the SAMs reduces the number of molecules that can be incorporated, immobilizing them at the outer plane of the monolayer. Only MB molecules incorporated into the SAMs close to the Au(111) surface (at a distance < 0.5 nm) are electrochemically active. MB molecules trapped in different defects explain the broad shape and humps observed in the voltammogram of the redox couple. The heterogeneous charge-transfer rate constants for MB immobilized into methyl-terminated thiolate SAMs are higher than those estimated for carboxylate- terminated SAMs, suggesting a different orientation of the immobilized molecule in the thiolate environment.

Introduction In recent years there has been a continuing interest in the study of charge-transfer processes from (to) molecular film-covered metallic electrodes to (from) redox moieties immobilized at molecular films or in electrolyte solutions. From the point of view of electronics, the metallic electrode/ molecular film/redox moieties arrangement corresponds to a donor (acceptor)/spacer/acceptor (donor) structure.1 The study of charge-transfer processes in these systems has a direct application for the development of new and more efficient biosensors, but could also have an impact in our knowledge of the electron-transfer processes between redox acceptors and enzymes.2 Thiolate molecules adsorbed on metals, particularly on Au(111), have been widely used to build donor/spacer/ acceptor structures because they can be easily selfassembled on metal surfaces from the vapor and liquid phases.3 Self-assembled thiolate monolayers (SAMs) form ordered crystalline films whose thickness and chemical composition can be controlled by changing the number of CH2 units (n) in the hydrocarbon chain and the nature of the terminal group (CH3, NH2, COOH),4 respectively. Usually, alkanethiol molecules on Au(111) form a stable (1) Adams, D. M.; Brus, L.; Chidsey, C. E. D.; Creager, S.; Creutz, C.; Kagan, C. R.; Kamat, P. V.; Lieberman, X. M.; Lindsay, O. S.; Marcus, R. A.; Metzger, R. M.; Michel-Beyerle, M. E.; Miller, J. R.; Newton, M. D.; Rolison, D. R.; Sankey, O.; Schanze, K. S.; Yardley, J.; Zhu, X. J. Phys. Chem. B 2003, 107, 6668 and references therein. (2) Zhang, J.; Grubb, M.; Hansen, A. G.; Kuznetsov, A. M.; Boisen, A.; Wackerbarth, H.; Ulstrup, J. J. Phys. Cond. Matter 2003, 15, 1873 and references therein. (3) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am. Chem. Soc. 1987, 109, 3559. Leopold, M. C.; Black, J. A.; Bowden, E. F. Langmuir 2002, 18, 978. Esplandiu´, M. J.; Hagenstro¨m, H.; Kolb, D. M. Langmuir 2001, 17, 828. Finklea, H. O. In Electroanalytical Chemistry; Bard, A. J., Rubinstein, I., Eds., Marcel Dekker: New York, 1996; Vol. 19, pp 109-335. Finklea, H. O. In Encyclopedia of Analytical Chemistry, Applications, Theory and Instrumentation; Meyers, R. A. Ed.: John Wiley & Sons: Chichester, U.K., 2000; p 10090. (4) Ulman, A. Chem. Rev. 1996, 96, 1533.

hexagonal x3×x3 R30° lattice or its c(4×2) superlattice. In this case molecules are tilted 30-35° from the surface normal.4 The principal reason for the interest in the use of SAMs as spacers is that a significant amount of the redox moieties is located at a well-defined distance (d) from the metallic electrode (Figure 1.a). Thus, in principle, electron chargetransfer processes can be explored at variable decay lengths and different chemical environments (by changing the chemical groups of the thiolate molecules). The assumption of a defined d value, as depicted in Figure 1.a, could be on average correct, but defects in the SAM and molecular interactions between the redox moieties and the thiol molecules could introduce significant deviations from the expected electrode-redox couple distance.3 For attractive molecular interactions, redox moieties trapped in the SAMs could “wire” the electrode surface to the redox molecules in the solution, thus acting as preferred paths for electron transfer. This situation is shown in Figure 1.b and c, where different possible locations and states of the redox moieties are shown. Another complication to having a defined d value arises from the fact that other phases could be present coexisting with the x3×x3 R30° lattice or its c(4×2) superlattice. In fact, pinstripes with missing rows5 and rectangular alkanethiol lattices where the molecules are tilted 50° from the substrate have also been reported.6 Molecular interactions are also important to understand the charge-transfer process in biological systems. In general, the interaction between biological redox moieties and enzymes could involve electrostatic, hydrogen bonding, hydrophobic, and van der Waals forces.7 The elucidation of the main interactions between redox moieties and (5) Poirier, G. E.; Tarlov, M. J.; Rushmeier, H. E. Langmuir 1994, 10, 3383. (6) Barrena, E.; Palacios-Lido´n, E.; Munuera, C.; Torrelles, X.; Ferrer, S.; Jonas, U.; Salmeron, M.; Ocal, C. J. Am. Chem. Soc. 2004, 126, 385.

10.1021/la036440w CCC: $27.50 © 2004 American Chemical Society Published on Web 05/07/2004

Charge Transfer through Thiolate SAMs on Au(111)

Figure 1. Scheme of a SAM-covered substrate in contact with an electrolyte solution containing redox molecules. The S head of the thiolate molecules is represented by a circle while the hydrocarbon chains are represented by a solid line. (a) ideal case (defect-free film). The smaller substrate-redox molecule distance is denoted by d. (b) and (c) show a scheme for defectcontaining SAMs where some redox molecules are placed at distances smaller than d.

enzymes, as well as the donor-acceptor distances, is crucial to understand charge transfer in biological systems. Methylene blue (MB) is an interesting molecule that exhibits a fast and reversible redox couple according to the reaction8

MB is a well-known electron acceptor for flavin-dehydrogenase (FAD) enzymes. In fact, MB in solution is able to accept electrons from the FAD-glucose oxidase (Gox)

Langmuir, Vol. 20, No. 12, 2004 5031

complex,9 although the active FAD center is buried 1.3 nm inside Gox, protected by a hydrophobic environment.10 Hydrogenation of the nitrogen atom also takes place in nicotinamide adenine dinucleotide, (NAD), nicotinamide adenine dinucleotide phosphate (NADP), flavine mononucleotide (FMN), and flavine adenine dinucleotide (FAD), molecules that are electron acceptors for many enzymes that are usually immobilized in the inner wall of cell membranes.9 NAD and NADP molecules interact weakly and reversibly with the active site of the enzyme.9 Therefore, the study of the MB molecule in hydrophobic environments, of the interactions between MB and charged groups, and of the MB-donor distances for efficient electron transfer are important to understand the behavior of redox moieties in biological systems. Carboxylate-terminated thiolate SAMs have been widely used to immobilize different biomolecules, such as cytochrome C and azurine.2,11 Both methyl-terminated and carboxylate-terminated thiolates on Au(111) exhibit similar surface structures although for the same number of C atoms carboxylate-terminated SAMs are more disordered, as revealed by electrodesorption measurements.12 In this work we have investigated the behavior of MB molecules immobilized in carboxylate-terminated and methyl-terminated thiolate SAMs on Au(111) prepared by solution deposition using electrochemical techniques combined with scanning tunneling microscopy (STM) and Auger electron spectroscopy (AES). While a large number of results have been published concerning scanning probe microscopy characterization of alkanethiolate SAMs,13 electrochemical determination of defects at SAMs,3 and the electrochemical behavior of redox couples immobilized on these monolayers,2 in our knowledge there are only a few studies where all these aspects are simultaneously analyzed on samples prepared under the same experimental conditions. This is a crucial point, as the SAMs quality depends strongly on the preparation conditions (substrate preparation, immersion time, solvent, temperature, alkanethiol concentration).13 This approach would allow one to measure the extent of surface heterogeneity at SAMs and to relate it to the redox behavior of the immobilized molecule (charge density, voltammetric shape, location, donor-acceptor distances). We have found that the redox behavior of the MB redox couple can be explained considering the complex surface structure of the thiolate SAMs that involves molecular defects in the crystalline domains and crystalline domains of different phases that turn the d values uncertain. For similar preparation conditions the degree of surface heterogeneity decrease as the number of C atoms in the hydrocarbon chains increases. Missing rows (particularly for shortchain alkanethiolates) and molecular defects act as preferred paths for MB incorporation into the SAMs driven by weak hydrophobic forces. The amount of immobilized MB is proportional to the number of defects at the SAMs. (7) Israelachvili, J. N. Intermolecular and Surface Forces; Academic Press: London, 1994; pp 353, 408. (8) Svetlicic, V.; Zutic, V.; Clavilier, J.; Chevalet, J. J. Electroanal. Chem. 1991, 312, 205. (9) Lehninger, A. L. Principles of Biochemistry, 3rd ed.; W. H. Freeman Company: New York, 2001. (10) Badia, A.; Carlini, R.; Ferna´ndez, A.; Battaglini, F.; Mikkelsen, S. R.; English, A. M. J. Am. Chem. Soc. 1993, 115, 7053. (11) Hobara, D.; Imabayashi, S.; Kakiuchi, T. Nano Lett. 2002, 2, 1021. (12) Imabayashi, S.; Iida, M.; Hobara, D.; Feng, Z. Q.; Niki, K.; Kakiuchi, T. J. Electroanal. Chem. 1997, 428, 33. Doblhofer, K.; Figura, J.; Fuhrhop, J. H. Langmuir 1992, 8, 1811. Azzaroni, O.; Vela, M. E.; Martin, H.; Herna´ndez Creus, A.; Andreasen, G.; Salvarezza, R. C. Langmuir 2001, 17, 6647. (13) Schreiber, F. Prog. Surf. Sci. 2000, 66, 151 and references therein.

5032

Langmuir, Vol. 20, No. 12, 2004

Benı´tez et al.

In the case of carboxylate-terminated SAMs, the negatively charged COO- groups at defects trap MB+ species, hindering penetration into the hydrocarbon layer. We also show that electron transfer from the Au(111) to the immobilized MB molecules requires donor-acceptor distances smaller than 0.5 nm. Experimental Section Au thin films on glass (AF 45 Berliner Glass KG, Germany) prepared by evaporation were used as substrates. They were annealed in a hydrogen flame for three minutes. As observed by STM and AFM, these Au electrodes exhibit atomically smooth (111) terraces separated by monatomic steps in height.14 Solutions were prepared by using analytical-grade chemicals and Milli-Q water. The experiments with the redox couples in solutions were performed in 0.1 mM methylene blue (MB+Cl-, chloride salt) + 0.1 M NaOH, and in 1 mM K4Fe(CN)6 + 1 M KCl solutions. In those measurements where the MB was immobilized on SAMs on Au substrates the procedure was done as follows: the Au substrates were immersed in 5 × 10-5 M mercaptopropionic acid (MPA), mercaptoundecanoic acid (MUA), propanethiol (PT), or dodecanethiol (DT) ethanolic solutions for 24 h to form ordered SAMs.3,12,15,16 Then, the substrates were rinsed first with ethanol and subsequently with water, and immersed for 30 min in 0.1 mM MB+ + 0.1 M NaOH aqueous solution for MB+ immobilization. Finally, the MB-modified thiol-covered Au substrates were rinsed again with water and placed as the working electrode of an electrochemical cell containing 0.1 M NaOH to test the MB+/MBH surface redox couple. For experiments with solutions containing the redox couples, the modified electrodes were rinsed with ethanol and water after the exposure to MPA, MUA, PT, or DT solutions and then placed in the cell containing 0.1 mM MB+ + 0.1 M NaOH17 or 1 mM K4Fe(CN)6 + 1 M KCl solutions. For quantification of the MB signal with AES, the well-known octamerical structure spontaneously formed on sulfurmodified Au electrodes18 was prepared. Au substrates were thus immersed in 3 × 10-3 M Na2S + 0.1 M NaOH saturated with nitrogen for 10 min at open circuit potential to form S8-covered Au(111) substrates After preparation, the S-covered Au(111) substrates were rinsed with Milli-Q water, and then immersed for 30 min in 10-4 M MB+ + 0.1 M NaOH solution for MB+ immobilization. Finally, the MB+ modified S-covered Au substrate was rinsed again with water and quickly transferred to the UHV chamber equipped for AES. Electrochemical measurements were performed in a conventional glass-made cell using a large Pt plate and a saturated calomel electrode as counter electrode and reference electrodes, respectively. All potentials in the text are referred to the saturated calomel electrode (SCE). Surface characterization of the monolayers was made by using Scanning Tunneling Microscopy (STM) (Nanoscope III, Digital Instruments) and Auger Electron Spectroscopy (AES) with a single-pass cylindrical mirror analyzer (CMA, Physical Electronics). (14) Andreasen, G.; Vela, M. E.; Salvarezza, R. C.; Arvia, A. J. Langmuir 1997, 13, 6814. (15) Bain, C. D.; Troughton, E. B.; Tao, Y.-T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321. (16) Hobara, D.; Miyake, K.; Imabayashi, S-i; Niki, K.; Kakiuchi, T. Langmuir 1998, 14, 3590. (17) Vericat, C.; Remes Lenicov, F.; Tanco, S.; Andreasen, G.; Vela, M. E.; Salvarezza, R. C. J. Phys. Chem. B 2002, 106, 9114. (18) Gao, X.; Zhang, Y.; Weaver, M. J. J. Phys. Chem. 1992, 96, 4156. Vericat, C.; Andreasen, G.; Vela, M. E.; Salvarezza, R. C. J. Phys. Chem. B 2000, 104, 302.

Figure 2. Voltammograms for MB immobilized on different SAM-covered Au(111) electrodes after immersion (dipping) in 0.1 mM MB+ + 0.1 M NaOH during 30 min. Electrolyte solution, 0.1 M NaOH; v ) 1 V s-1. (a) MPA, (b) MUA, (c) PT, and (d) DT.

Results and Discussion Voltammograms recorded in 0.1 M NaOH at v ) 1 V s-1 after methylene blue immobilization on carboxylateterminated thiolate SAMs on Au(111) electrodes (MPA, Figure 2a and MUA, Figure 2b) show a small and reversible pair of current peaks in the potential range -0.55 V < E < -0.50 V. This pair of peaks corresponds to the MB+/MBH redox couple reaction described by eq 1. The voltammetric peaks are broad and contain humps that are clearly detected in the voltammograms shown in Figure 2. The charge density involved in these reversible peaks is small and in some cases negligible (q ≈ 2 ( 2 µC cm-2). This figure is much smaller than q ) 50 µC cm-2, the value expected for a complete MB monolayer immobilized on a similar x3xx3 R30° S lattice on the Au(111) surface.8,17 We have also immobilized MB on methyl-terminated (propanethiolate, n ) 3; dodecanethiolate, n ) 12) SAMs on Au(111). Again, after careful rinsing, the voltammograms (Figure 2c,d), recorded in 0.1 M NaOH, show the same pair of reversible current peaks located at similar potential values as those found for MB immobilized on MPA and MUA. In this case the q value results in q ) 4 ( 2 µC cm-2. This means that, irrespective of the terminal group of the alkanethiolate molecules, a small amount of electrochemically MB is present. Repetitive cycling of the MB immobilized on the carboxylate- and methyl-terminated thiolate monolayers shows a slow but progressive decrease in the q value, indicating that the molecules are immobilized by weak forces. To measure the total amount of immobilized MB on the MPA-covered Au(111) surfaces, AES spectra were obtained before (Figure 3a) and after (Figure 3b) MB adsorption. The signal at 152 eV, evident in both samples, corresponds to S(LLV) transitions originating from the S heads of the MPA thiolates (Figure 3a) adsorbed on the Au surface and from the S atom of the immobilized MB molecules (Figure 3b) that are adsorbed on the MPA-covered Au(111). On the other hand, the signal at 395 eV, clearly observed for the MPA-covered surface (Figure 3a), corresponds to Au transitions. The latter overlaps to some extent the signal at 381 eV originated from the N(KLL) transitions observed for MB immobilized on MPA-covered Au(111) (Figure 3b). The presence of the N signal clearly demonstrates that MB molecules are immobilized on the MPAcovered surface, and remain adsorbed after emersion and careful rinsing of the sample. To estimate the amount of adsorbed MB we have measured the N/S atom ratio from the S and N peak-to-peak heights after subtraction of the

Charge Transfer through Thiolate SAMs on Au(111)

Figure 3. AES spectra of SAM-covered Au(111) electrodes: (a) MPA, (b) MPA after 30 min immersion in a 0.1 mM MB+ + 0.1 M NaOH solution, (c) Au-S8 after 30 min immersion in 0.1 mM MB+ + 0.1 M NaOH, and (d) PT after 30 min immersion in 0.1 mM MB+ + 0.1 M NaOH.

Figure 4. AES spectra of SAM- covered electrodes: (a) MPA without MB, (b) MPA after 30 min immersion in 0.1 mM MB+ + 0.1 M NaOH, (c) difference of spectra b - a, and (d) PT after 30 min immersion in 0.1 mM MB+ + 0.1 M NaOH.

Au signal at 395 eV (see Figure 4a-c). The MB coverage, calculated considering the N/S ratio expected for a complete MB monolayer on MPA, is 0.7 ML. This value was compared to 0.80 ML found for MB immobilized on S8-covered Au(111) (Figure 3c) using the same procedure. The amount of charge voltammetrically recorded for the MB immobilized on S8-covered Au(111) is q ) 45 µC cm-2,17 in good agreement with the 0.8 ML surface coverage by MB molecules found in AES measurements. This means that all immobilized molecules participate in the chargetransfer process. Conversely, for the MB immobilized on MPA-covered Au(111) we have found q ) 2 ( 2 µC cm-2 rather than q ) 35 µC cm-2, the value expected for the surface coverage estimated by AES (0.7 ML). Thus, less than ≈10% of the immobilized MB molecules are able to participate in the charge-transfer process. On the other hand, the AES spectra taken after MB adsorption on a propanethiolate-covered Au(111) (Figures 3d and 4d) show no traces of N. Thus, methyl-terminated thiolates are less efficient than carboxylate-terminated

Langmuir, Vol. 20, No. 12, 2004 5033

Figure 5. Charge density of the immobilized MB redox couple (q) values vs the size of the spacer measured as the diameter of the atom (S,I) or the distance between the Au surface and the terminal methyl group of the alkanethiolate molecule at SAMs (PT, MPA, MUA, DT).

thiolates for MB immobilization. However, we have detected a similar small amount of electrochemically active MB in both methyl- and carboxy-terminated thiolate SAMs (Figure 2). Therefore, it is evident that the amount of MB immobilized on propanethiolate SAMs is below the detection limit of AES. These experiments allow us to draw several interesting conclusions. The first one is that electrostatic interactions between the negatively charged carboxylate anions of the thiolate molecule and the positively charged MB+ species are able to immobilize MB at the SAM-electrolyte interface, according to AES results and previous UV-vis results.17 In fact, the total amount of MB decreases abruptly at methyl-terminated thiolates, revealing the importance of the COO- terminal groups in MB immobilization. The second one is that the molecules immobilized at carboxylate groups of MPA and MUA are not electrochemically active. In fact, the same small amount of electrochemically active MB molecules is detected for both methyl- and carboxylate-terminated thiols. This fact indicates that, even for MPA (n ) 3), the electronic coupling between the redox centers (that are immobilized at the terminal-carboxylate group) and the Au(111) donor is too weak to allow charge transfer. The third conclusion is that the small amount of electrochemically active MB should be immobilized into the SAMs, in contact with the hydrocarbon chains, driven by hydrophobic forces. This can be concluded because the presence of electrochemically active MB is independent of the thiolate terminal group (hydrophilic and negatively charged carboxylate groups in MPA and MUA molecules; hydrophobic in propanethiolate and dodecanethiolate). The fourth conclusion is that the small fraction of electrochemically active MB is immobilized into the SAMs at a distance (d′) smaller than (d) (Figure 1), thus allowing a strong electron coupling with the Au(111) surface. In Figure 5 we have plotted q values vs the size of the spacer, which can be taken as a measure of d, for different spacers. It is evident that q values markedly decrease when d is larger than 0.5 nm. This result suggests that the small amount of electrochemically active MB immobilized into the SAMs is placed at a distance smaller than 0.5 nm from the Au(111) surface. Now we will demonstrate that the electrochemically active MB molecules can be incorporated into the film through defective regions of the SAMs. In fact, we will show that small defects (molecular gates) are present even in well-ordered (crystalline) SAMs in contact with aqueous electrolytes. Figure 6a shows a low resolution image (90 × 90 nm2) of the Au(111) substrate covered by a hexanethiolate SAM. Well-ordered crystalline domains and

5034

Langmuir, Vol. 20, No. 12, 2004

Figure 6. (a) 90 × 90 nm2 STM image of hexanotiolate SAMcovered Au(111). Crystalline domains with disordered domain boundaries and large pits monatomic in depth are clearly seen. (b) (I) 20 × 20 nm2 STM images of hexanethiolate-covered Au(111) showing a large domain with the c(4×2) lattice. The arrow indicates a typical nanometer-sized defect in the SAM (molecular defect). (II) Size distribution function of defects present in a hexanethiol SAM on Au(111) in contact with 0.1 M NaOH. (III) 19 × 19 nm2 STM image of propanethiolate- covered Au(111). A typical pinstripe surface structure with p ) 7.5 is shown. (IV) 9 × 9 nm2 STM image showing the rectangular 2×x3 lattice.

monatomic pits in depth with sizes in the order of 3-6 nm (produced by the strong chemisorption of the alkanethiolate molecules) can be clearly observed.19,20 These are not true defects, as the pit bottoms are fairly covered by ordered rows of alkanethiolate molecules. A detailed analysis of these pits and of the pits dynamics has also been reported.21 Note also in Figure 6a the presence of disordered regions located at domain boundaries. Figure 6b-I shows a 20 × 20 nm2 STM image of the same surface (Figure 6a). The ordered domains of mol(19) Edinger, K.; Go¨lzha¨user, A.; Demota, K.; Wo¨ll, KCh.; Grunze, M. Langmuir 1993, 9, 4. (20) McDermott, C. A.; McDermott, M. T.; Green J. B.; Porter M. D. J. Chem. Phys. 1995, 99, 13257.

Benı´tez et al.

ecules forming the c(4×2) superlattice are now clearly resolved.13,22 This image also shows nanometer-sized black holes 0.5-2 nm in diameter (hardly observed in Figure 6a) that correspond to real defects in the SAM. These defects could be not only sites where the alkanethiolate molecules are absent but also (and more probably) regions where the molecules are not well organized, that is, regions where hydrocarbon chains are not fully extended, as schematically shown in Figure 1b,c. The size distribution function (Figure 6b-II) for these defects is broad, with a maximum at defect sizes close to the maximum length of the MB molecule (1.5 nm). Note also the large tail of the size distribution function, showing the existence of a few large defects. These results clearly demonstrate that even crystalline ordered SAMs in aqueous solutions exhibit a large number of real defects with a broad size distribution function. These types of molecular defects has been recently described on the Ag(111) surface.23 For SAMs prepared under the same experimental conditions, the degree of surface heterogeneity increases markedly as the number of C atoms in the hydrocarbon chains is reduced. Thus, for propanethiolate (n ) 3) STM images are completely dominated by pinstripe structures (p×x3), where rows of alkanethiol molecules are missing.5 The STM image (Figure 6b-III) shows a pinstripe structure with p ) 7.5. That is, every fifth row of alkanethiolate molecules in a x3×x3 R30° lattice is missing. This type of structure has been reported for ethanethiolate24 and other short chain alkanethiolates on Au(111),5,25 such as MPA.26 Missing rows of alkanethiolate molecules can act as preferred paths for MB incorporation, as the molecule thickness in the vertical adsorption configuration is ≈0.4 nm. We have also observed some ordered domains with x3×x3 R30° and c(4×2) superlattices, and also domains with rectangular lattices (2×x3) (Figure 6b-IV). Recent AFM measurements have shown that alkanethiolate molecules in these domains are tilted 50° rather than the 30° corresponding to molecules in the x3×x3 R30° and c(4×2) lattices.6 Missing rows and crystalline domains with different molecular tilts could result in a broad size distribution of d values for propanethiolate SAMs. Conversely, STM images for dodecanethiolate SAMs on Au(111) are similar to those shown for hexanethiolate SAMs (Figure 6a,b-I) with domains where the x3×x3 R30° and c(4×2) lattices can be clearly observed. Therefore one can conclude that under comparable preparation conditions the degree of surface heterogeneity increases for n < 6. In the next figure we connect the defect density with the amount of electrochemically active MB. In Figure 7 we have plotted the q values for the MB+/MBH redox couples immobilized on different propanethiolate, dodecanethiolate, and MPA as a function of the film capacitance (Cf) of the SAM-covered Au(111) measured from the cyclic voltammograms.3,27 The Cf value is a direct measure of the defect density present in the different SAMs because (21) Poirier, G. E.; Tarlov, M. J. J. Phys. Chem. 1995, 99, 10966. Cavalleri, O.; Hirstein, A.; Bucher, J.-P.; Kern, K. Thin Solid Films, 1996, 284, 392. Tera´n Arce, F.; Vela, M. E.; Salvarezza, R. C.; Arvia, A. J. Langmuir 1998, 14, 7203. (22) Anselmetti, D.; Baratoff, A.; Guntherodt, H. J.; Delamarche, E.; Michel, B.; Gerber, Ch.; Kang, H.; Wolf, H.; Ringsdorf, H. Europhys. Lett. 1994, 27, 365. Tera´n, F. T.; Vela, M. E.; Salvarezza, R. C.; Arvia, A. J. J. Chem. Phys. 1998, 109, 5703. (23) Schweizer, M.; Kolb, D. M. J. Electroanal Chem., in press (available at www.sciencedirect.com). (24) Hagenstro¨m, H.; Schneeweiss, M. E.; Kolb, D. M. Langmuir 1999, 15, 2435. (25) Camillone, N., III; Eisenberger, P.; Leung, T. Y. B.; Schwartz, P.; Scoles, G. J. Chem. Phys. 1994, 101, 11031. (26) Petri, M.; Kolb, D. M.; Memmert, U.; Meyer, H. Electrochim. Acta 2003, 49, 175.

Charge Transfer through Thiolate SAMs on Au(111)

Langmuir, Vol. 20, No. 12, 2004 5035

Figure 7. Charge of the immobilized couple of MB (q) as a function of the double layer capacitance (Cf) for different Au(111) samples covered by thiolates: (b) propanethiolate, (O) dodecanethiolate, and (0) MPA.

Figure 9. Cyclic voltammograms for the carboxylate- and methyl-terminated SAM-covered Au(111) electrodes recorded at v ) 0.05 V s-1 in Fe(CN)64- + 1 M KCl solution.

Figure 8. Cyclic voltammograms for the carboxylate-terminated and methyl-terminated SAM covered Au(111) electrodes recorded at v ) 0.05 V s-1 in a 0.1 mM MB+ + 0.1 M NaOH solution: (a) PT, (b) DT, (c) MPA, and (d) MUA.

it reflects the amount of Au(111) area exposed to the electrolyte solution.3 Results in Figure 7 show that q values increase as Cf is increased, indicating that a larger number of defects in the film results in a larger number of electroactive MB molecules. Therefore, we can conclude that defects at SAMs act as gates for MB incorporation into the hydrocarbon chains, driven by hydrophobic forces. However, the dependence of q on Cf is stronger for MB immobilized on propanethiolate than on MPA SAMs. In principle this seems an unexpected result, since MPA SAMs are more defective than propanethiolate SAMs.12 In fact it has been reported that the Au(111) surface is never completely covered by high-density films.26 However, the negatively charged -COO- groups can trap the MB+ ions at the defect entrance so that only a few of them are able to penetrate the MPA layer to reach an efficient donor-acceptor distance for charge transfer. The fact that the Cf values measured for dodecanethiolate SAMs (n ) 12) are always smaller than those measured for propanethiolate SAMs (n ) 3) (Figure 7) supports the previous conclusion from STM images: the defect number decreases as n increases. Note that a similar conclusion was derived a few years ago through capacitance measurements on alkanethiolate covered Au.3 Cyclic voltammograms (Figure 8a-d) for the carboxylate-terminated and methyl-terminated SAM-covered Au(111) electrodes, recorded with MB in solution, reveal welldefined voltammetric peaks at the same potential values found for the immobilized molecules, irrespective of the number of methylene units and the nature of the terminal group. The q values observed for carboxylate- (n ) 3 and n )11) and methyl-terminated (n ) 3 and n ) 12) SAMs are similar, within the experimental error. The well(27) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamental and Applications; Wiley & Sons: New York, 1980.

defined electrochemical behavior of the MB+/MBH redox couple in solution, in contrast to the null electrochemical response of the MB immobilized at the carboxylate groups, suggests that MB trapped at the SAM defects by hydrophobic forces act as wires for the electron transfer from the Au(111) surface to the MB molecules in solution. This interpretation is also supported by the fact that the redox couple in solution shows the same behavior for Au electrodes covered by methyl-terminated thiolates (where only MB molecules trapped at defects are present) and for Au electrodes covered by carboxylate-terminated thiolates (where the total amount of MB is 0.7 ML). To demonstrate the role of the trapped MB molecules we have made similar voltammetric measurements by using the Fe(CN)64-/Fe(CN)63- redox couple in 1 M KCl aqueous solutions and the methyl-terminated SAMs on Au(111) (n ) 3 and n ) 12) as substrates.28 In this case the voltammetric results (Figure 9) show that, while the electron transfer is allowed for n ) 3, it is completely hindered for dodecanethiolate SAM (n ) 12) covered Au(111). This is not surprising, as propanethiolate SAMs are more defective than dodecanethiolate SAMs. In fact, for both n ) 3 and n ) 12 small defects can be inactive because the large hydrated Fe(CN)64-/Fe(CN)63- ions29 cannot penetrate the hydrocarbon layer. However, defects in propanethiolate SAMs can allocate redox molecules able to wire the donor Au substrate with the Fe(CN)64-/Fe(CN)63- species in the solution. The number of these defects is lower for n ) 12 than for n ) 3 (see Figure 7), thus explaining the hindered charge transfer through the longer alkanethiolate SAM. A similar behavior is observed for the Fe(CN)64-/Fe(CN)63- redox couple on MPA and MUA layers. These results clearly demonstrate the role of defects and the nature of the redox couple for determining the amount of species able to penetrate SAMs. The presence of different types of defects at SAMs is consistent with the broad voltammetric peaks and the presence of humps detected for the MB+/MBH redox couple in Figure 2. In fact, it has been reported that the environment around an immobilized redox species affects the Eo value.30 Therefore, it is reasonable to assume that MB molecules immobilized at different types of defects and on different crystalline phases feel different environ(28) Rubinstein, I. J. Am. Chem. Soc. 1991, 113, 5176. Rubinstein, I.; Steimberg, S.; Tor, Y.; Shanzer, A.; Sagiv, J. Nature 1988, 332, 426. Sabatani, E.; Rubinstein, I. J. Phys. Chem. 1987, 91, 6663. (29) Marcus, Y. Biophys. Chem. 1994, 51, 111.

5036

Langmuir, Vol. 20, No. 12, 2004

Benı´tez et al.

Table 1. spacer

sweep rate (Vs-1)

k (s-1)

Eo (V)

∆Ep (V)

DT PT MUA MPA sulfide

0.2-10 1-10 0.05-1 0.5-5 0.1-10

35-60 120-260 5-9 10-30 12-18

-0.512/-0.504 -0.519/-0.521 -0.514 -0.570/-0.590 -0.630/-0.650

0.027-0.112 0.012-0.0417 0.091 0.060-0.100 0.068-0.179

ments leading to a distribution in the Eo values, and accordingly, produce broad peaks and humps in the voltammograms. In fact, we have been able to fit the experimental voltammograms for the redox couple by using a simple model including at least two MB with different Eo values. Finally we estimated the charge-transfer rate constant (k) of the MB redox couple immobilized on different thiolate SAM- and sulfide-covered (n ) 0) Au(111) by voltammetric measurements.31 In this method the transfer coefficient of the electrochemical reaction in diffusionless electrochemical systems can be obtained from the difference between the peak potential of the anodic and cathodic peaks (∆Ep) of reversible peaks redox couples. This procedure has been applied many times to different redox couples immobilized on thiolate SAMs.32 In our case, due to the complex shape of the voltammogram, we have limited our analysis to the main voltammetric peaks and disregarded the humps. By applying this method to experimental voltammograms and at sweep rates ranging from 0.1 to 10 V s-1, complemented with a numerical integration of the Butler-Volmer equation31 reaction, we have estimated the relevant electrochemical quantities, summarized in Table 1. As was pointed out by Weber and Creager33, the rate constants calculated by means of Laviron formalism have a dependence with sweep rates originated from the fact that peak potentials depend nonlinearly on sweep rate. However, we have worked in a limited range of sweep rates where deviations from theory seem to be small.33 Therefore, we believe that this effect does not have a strong impact on our data fitting. Redox potentials (Eo) estimated from this method are in the range -0.590/-0.504 V for MB immobilized on both methyl-terminated and carboxylate-terminated SAMs (Table 1). These values are shifted in the positive direction with respect to MB immobilized on adsorbed sulfide, probably reflecting the different environment given by the hydrocarbon chains. Results shown in Table 1 also reveal that k values for MB immobilized on methylterminated SAMs are larger than those measured for carboxylate-terminated SAMs for a similar n value (dodecanethiolate versus MUA). This can be explained considering that on average MB is held at larger d values by the carboxylate groups. In fact, as expected1,32 the k values estimated for MB immobilized in methyl-terminated SAMs decrease when n increases (Table 1, propanethiolate vs dodecanethiolate). In the frame of our previous discussion this means that MB molecules can more easily approach the Au(111) surface when n is small. On the other hand, the k values for MPA and MUA approach those measured for sulfide-covered Au(111). It should be noted that the nature of the chemical S-Au bond in spontaneously-formed sulfide monolayers (oc(30) Finklea, H. O.; Hanshew, D. D. J. Am. Chem. Soc. 1992, 114, 3173. (31) Laviron, E. J. Electroanal. Chem. 1979, 101, 19. (32) Smalley, J. F.; Flinkea, H. O.; Chidsey, C. E. D.; Linford, M. R.; Creager, S. E.; Ferraris, J. P.; Chalfant, K.; Zawodzinsk, T.; Feldberg, S. W.; Newton, M. D. J. Am. Chem. Soc. 2003, 125, 2004 and references therein. (33) Weber, K.; Creager, S. E. Anal. Chem. 1994, 66, 3164.

tamerical structure) and in thiolates is the same, as revealed by XPS measurements.34 Therefore, we can conclude that, from the point of view of the charge transfer, the environment around an immobilized MB on carboxylate SAMs is, in some extent, similar to that found for adsorbed sulfide on Au(111). One can speculate that MB molecules are trapped at large defect sites of the carboxylate-terminated SAM where the hydrocarbon chains of the thiolates are strongly disordered and have their S heads (bonded to the Au(111) surface) close to the electrolyte solution (due to the disorder in the alkyl chains). This is supported by the fact that the Eo value for MB in the highly defective MPA approaches the Eo value measured for MB immobilized on sulfide. The larger k values estimated for methyl-terminated SAMs agree with previous observations about the enhancement of the charge transfer processes for certain redox moieties immobilized in hydrophobic SAMs on Au in relation to that observed for the same redox moieties directly in contact with the Au surface.2 In fact, a faster electron transfer has been observed for azurin immobilized in hydrophobic alkanethiolate SAMs on Au(111)35 than for azurin adsorbed directly on Au(111). In this case the strong hydrophobic interactions could fix the molecule in a well-defined orientation enhancing electron transfer. This explanation could be also valid for the MB molecules immobilized into the methyl-terminated SAMs. Conclusions (1) Molecular sized defects, missing rows, and crystalline domains with different molecular tilts present in the alkanethiolate spacer turn the donor-acceptor distance (d) rather uncertain. The degree of surface heterogeneity decreases as the length of the hydrocarbon chains (n) increases. (2) Defective regions act as preferred paths for MB penetration into the SAMs driven by hydrophobic forces. Hydrophobic forces are relatively weak so that immobilized MB molecules are slowly desorbed from the SAMs (3) Electrochemically active MB molecules are those trapped close to the Au(111) surface. The amount of these molecules depends on the defects density present in the SAMs. MB molecules immobilized at different defects can explain the broad voltammetric peaks and humps observed in the voltammograms. (4) The presence of the negatively charged groups in carboxylate-terminated SAMs immobilizes a large number of MB+ by electrostatic interactions. In this case only a small fraction of MB molecules can penetrate into the SAMs driven by hydrophobic forces. (5) The charge-transfer process is not efficient for donoracceptor distances larger than 0.5 nm. (6) The trapped molecules can efficiently wire other MB molecules present in solution. (7) The charge-transfer rate constants measured for MB immobilized on methyl-terminated SAMs are higher than those measured for MB in carboxylate-terminated SAMs. In the latter case k approaches those values measured for sulfide-covered Au(111). The hydrophobic environment of methyl-terminated SAMs seems to enhance the charge-transfer process. (8) It is expected that MB-like molecules can penetrate into hydrophobic environments such as those present close to the active sites of enzymes through “gates” in order to (34) Vericat, C.; Vela, M. E.; Andreasen, G.; Salvarezza, R. C.; Va´zquez, L.; Martı´n-Gago, J. A. Langmuir 2001, 17, 4919. (35) Chi, Q.; Zhang, J.; Andersen, J. E. T.; Ulstrup, J. J. Phys. Chem. B 2000, 105, 4669.

Charge Transfer through Thiolate SAMs on Au(111)

achieve a strong electronic coupling with the donor center. The interaction is weak, thus allowing the release of the reduced form. Acknowledgment. This work was supported by the Agencia Nacional de Promocio´n Cientı´fica y Tecnolo´gica

Langmuir, Vol. 20, No. 12, 2004 5037

(PICT 99-5030) and CONICET (PIP 0897) (Argentina). M.E.V. is a member of the research career of CIC. The authors thank G. Andreasen for taking the STM images. LA036440W