Electrochemical Behavior of Azobenzene Self-Assembled Monolayers

Langmuir , 1996, 12 (11), pp 2843–2848. DOI: 10.1021/ .... Zhong-Qun Tian. The Journal of Physical Chemistry C 0 (proofing), .... Langmuir 1999 15 (...
0 downloads 0 Views 239KB Size
Langmuir 1996, 12, 2843-2848

2843

Electrochemical Behavior of Azobenzene Self-Assembled Monolayers on Gold Hua-Zhong Yu, Yong-Qiang Wang, Ji-Zhi Cheng, Jian-Wei Zhao, Sheng-Min Cai, Hiroo Inokuchi,† Akira Fujishima,‡ and Zhong-Fan Liu* Center for Intelligent Materials Research (CIMR), College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China Received July 28, 1995. In Final Form: February 5, 1996X The self-assembly and electrochemical properties of a novel thiol-functionalized azobenzene derivative were studied on gold electrodes by cyclic voltammetry. The compound was 4-ethoxy-4′-((N-(2′′mercaptoethyl)amino)carbonyl)azobenzene (C2AzoC2SH), which was synthesized using a simple and general technique by our laboratory. The C2AzoC2SH formed uniform and reproducible self-assembled monolayers (SAMs) on gold with a surface coverage of 4.21 × 10-10 mol/cm2. The SAMs showed voltammetric responses in a B-R buffer at a pH range of 3.2-8.6 that corresponded to the two-electron, two-proton reductionoxidation of azobenzene. The voltammetric behavior was totally irreversible, exhibiting very large peakto-peak splitting and other distortions. The dependence of apparent surface electrochemical rate constant on pH value has a V-shape, owing to the influence of protonation reactions on the multistepped reductionoxidation of azobenzene functionality.

Introduction The electrochemical behavior of surface-modified electrodes continues to be of interest for both fundamental and practical perspectives.1,2 In the past few years, selfassembled monolayers (SAMs) of alkanethiols on gold have emerged as a powerful way to anchor redox active species onto electrodes with a desired distance and a controlled microenvironment. With highly organized, pinhole-free, and stable structures, they provide an ideal opportunity to investigate long-range electron transfer between a redox active center and an electrode without the complication of mass transfer. Up to now, redox couples such as ferrocene,3-5 viologen,6,7 and ruthenium polypyridyl complexes8-10 within SAMs have been widely studied. They generally exhibit ideal reversible surface waves. Also available are some reports concerning quinone SAMs fabricated by both self-assembly of thiol-functionalized quinone derivatives11-13 and step-by-step surface reaction.14-17 However, little work has been performed on * To whom correspondence should be addressed: Tel, 86-102751494; Fax, 86-10-2757157, 86-10-2564095; e-mail, ZFLIU@SUN. IHEP.AC.CN. † Okazaki National Institute, Myodaiji, Okazaki, 444, Japan. ‡ Department of Applied Chemistry, Faculty of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku 113, Japan. X Abstract published in Advance ACS Abstracts, May 1, 1996. (1) Ulman, A. An Introduction to Ultrathin Organic Films: From Langmuir-Blodgett to Self-Assembly; Academic Press: San Diego, CA, 1991. (2) Murray, R. W. Molecular Design of Electrode Surface; Techniques of Chemistry Series; John Wiley & Sons, Inc.: New York, 1992; Vol. XXII. (3) Chidsey, C. E. D. Science, 1991, 251, 919. (4) Chidsey, C. E. D.; Bertozzi, C. R.; Putvinski, T. M.; Mujsce, A. M. J. Am. Chem. Soc. 1990, 112, 4301. (5) Kondo, T.; Takechi, M.; Sato, Y.; Uosaki, K. J. Electroanal. Chem. 1995, 381, 203. (6) Delong, H. C.; Buttry, A. D. Langmuir 1992, 8, 2491. (7) Bunding-Lee, K. A. Langmuir 1990, 6, 709. (8) Finklea, H. O.; Hanshew, D. D. J. Am. Chem. Soc. 1992, 114, 3173. (9) Ravenscroft, M. S.; Finklea, H. O. J. Phys. Chem. 1994, 98, 3843. (10) Finklea, H. O.; Ravenscroft, M. S.; Snider, D. A. Langmuir 1993, 9, 223. (11) Bravo B.; Mebrahtu, T.; Soriaga, M. P.; Zapien, D. C.; Hubbard, A. T.; Stickney, J. L. Langmuir 1987, 3, 595. (12) Hickman, J. J.; Ofer, D.; Laibinis, P. E.; Whitesides, G. M.; Wrighton, M. S. Science 1991, 252, 688. (13) Zhang, L.; Lu, T.; Gokel, G. W.; Kaifer, A. E. Langmuir 1993, 9, 786.

S0743-7463(95)00632-9 CCC: $12.00

azobenzene/hydrazobenzene redox couples.18-20 The electrochemical behavior of azobenzene and its derivatives in solution has been studied extensively because of the involvement of proton transport, adsorption/desorption, and cis/trans isomerization.21-25 The electrochemical behavior of azobenzene Langmuir-Blodgett films has been intensively studied in our group.26-31 We are currently trying to introduce azobenzene functionality into selfassembled monolayers in order to investigate their voltammetric properties. In this paper, we describe the preparation of selfassembled monolayers on gold electrodes with a novel thiol-functionalized azobenzene, 4-ethoxy-4′-((N-(2′′-mercaptoethyl)amino)carbonyl)azobenzene (referred to as C2AzoC2SH), and the investigation of their electrochemical behavior by means of cyclic volammetry. Experimental Section Materials. The azobenzene functionality was introduced into the thiol according to Scheme 1. Thionyl chloride (CP), pyridine (CP) was purchased from Beijing Chemicals. 2-Aminoethanethiol hydrochloride was obtained from Aldrich. Pyridine was dried over potassium hydroxide and distilled before using. All other chemicals were used without further purification. The 4-ethoxy(14) Katz, E. Y.; Solov’ev, A. A. J. Electroanal. Chem. 1990, 291, 171. (15) Katz, E. Y.; Borovkov, V. V.; Evstigneeva, R. P. J. Electroanal. Chem. 1992, 326, 197. (16) Katz, E.; Schmidt, H. L. J. Electroanal. Chem. 1993, 360, 337. (17) Katz, E.; Schmidt, H. L. J. Electroanal. Chem. 1994, 368, 87. (18) Herr, B. R.; Mirkin, C. A. J. Am. Chem. Soc. 1994, 116, 1157. (19) Caldwell, W. B.; Chen, K. M.; Herr, B. R.; Mirkin, C. A.; Hulteen, J. C.; Duyne, R. P. V. Langmuir 1994, 10, 4109. (20) Caldwell, W. B.; Campbell, D. J.; Chen, K.-M.; Herr, B. R.; Mirkin, C. A.; Malik, A.; Durbin, M. K.; Dutta, P.; Huang, K. G. J. Am. Chem. Soc. 1995, 117, 6071. (21) Wawzonek, S.; Fredrickso, J. D. J. Am. Chem. Soc. 1955, 77, 3985. (22) Salder, J. L.; Bard, A. J. J. Am. Chem. Soc. 1967, 90, 1979. (23) Laviron, E.; Mungnier, Y. J. Electroanal. Chem. 1980, 111, 337. (24) Laviron, E. J. Electroanal. Chem. 1984, 169, 29. (25) Flamigni, L.; Moni, S. J. Phys. Chem. 1985, 89, 3702. (26) Liu, Z. F.; Hashimoto, K.; Fujishima, A. Nature 1990, 347, 658. (27) Liu, Z. F.; Hashimoto, K.; Fujishima, A. J. Phys. Chem. 1992, 96, 1875. (28) Liu, Z. F.; Hashimoto, K.; Fujishima, A. Faraday Disscuss. 1992, 94, 221. (29) Liu, Z. F.; Loo, B. H.; Hashimoto, K.; Fujishima, A. J. Electroanal. Chem. 1991, 297, 133. (30) Liu, Z. F.; Hashimoto, K.; Fujishima, A. J. Electroanal. Chem. 1992, 324, 259. (31) Liu, Z. F.; Hashimoto, K.; Fujishima, A. Chem. Lett. 1990, 2177.

© 1996 American Chemical Society

2844

Langmuir, Vol. 12, No. 11, 1996

Yu et al.

Scheme 1

4′-carboxylic azobenzene was prepared following a standard literature procedure.32 The final product was purified by silica gel column chromatography: yield 80%. Proof of its chemical structure CH3CH2OC6H4NdNC6H4C(O)N(H)CH2CH2SH was based on: 1H NMR(DMSO) δ 1.36-1.40 (t, 3H, CH3); δ 2.99-3.02 (t, 2H,CH2SH); δ 3.59-3.65 (m, 2H, NHCH2); δ 4.14-4.19 (m, 2H, CH2O); δ 7.10-7.13 (d, 2H, ArH); δ 7.85-8.03(m, 6H, ArH); δ 8.66-8.67 (t, 1H, NH); SH was not detectable due to trace water in solvent. MS: m/e 329 (M+); IR (KBr) 1636 (νCdO), 1252(νC-O), 2558(νS-H) cm-1. Preparing SAM-Modified Electrodes. Monolayers were formed by the self-assembling technique on gold electrodes. The gold electrodes were disks of diameter 0.5 mm prepared by sealing an annealed gold wire into an insulating epoxy resin (Torr Seal, Varian Vacuum Products). The electrodes were first polished with 800 mesh sand paper and then 1.0 and 0.5 µm alumina slurry, followed by extensive rinsing with water and sonication. The Au electrode was then electrochemically cleaned and polished by cycling the electrode potential between -0.45 and +1.5 V vs Ag/AgCl in 0.5 M H2SO4 (AR) for several minutes until a stable and standard voltammogram was obtained. The potential was then held in the double layer region for 3 min at a sufficiently negative value (0-0.25 V vs Ag/AgCl) to ensure complete reduction of any surface oxide. The microscopic surface area of the electrode was determined from the cathodic charge for the reduction of gold oxide. The surface roughness factor was obtained by dividing the microscopic surface area by the geometric surface area. Genearlly, it is about 2.1 for our electrodes, larger than that obtained by Mirkin for Au(rough)/mica.19 The surface roughness factor has been taken into account in the determination of surface concentration. Afer rinsing with deionized water and drying with nitrogen gas, the electrodes were immersed in 2 mL of the deposition solution. Upon removal from the deposition solution, the electrodes were rinsed with THF and then water. Reproducible C2AzoC2SH SAMs on gold electrodes can be obtained by the above procedure. The deposition solution was 1 mM C2AzoC2SH in THF. The general dipping time was kept to about 24 h. The surface concentration of the SAM was approximately independent of adsorption time for times between 5 h and 2 days. Electrochemical Measurements. Electrochemical experiments were performed in a single compartment cell at room temperature (22 °C). An Ag/AgCl:saturated KCl electrode and a Pt wire were used as the reference electrode and counter electrode, respectively. The cyclic voltammetric measurements were performed with a Hokuto Denko HA-150 potentiostat and a HB-III function generator. The signals were recorded on a Riken Denshi F-35A X-Y recorder. All experiments were carried out in Britton-Robinson (B-R) buffer, which was freed from oxygen by bubbling with nitrogen.

Figure 1. Cyclic voltammogram of C2AzoC2SH/Au SAM (pH 5.0, B-R buffer, 0.1 M NaClO4), scan rate 50 mV/s.

pH range, repetitive cycling over hundreds of scans does not alter the voltammogram, demonstrating that the SAM is stable to electrochemical cycling. The surface concentration (Γazo) of immobilized C2AzoC2SH molecules was determined by the cut-andweigh method, assuming a two-electron and two-proton reaction mechanism.23,29-31 The Γazo value obtained is 4.21 × 10-10 mol/cm2, which corresponds to a molecular area of 0.39 nm2/molecule in the SAM. The area per molecule is much smaller than that obtained for adsorbed unsubstituted azobenzene on mercury electrodes (1.66 nm2/ molecule),33 indicating the SAM has a relatively closely packed and oriented structure, although this value is larger than that of docosanethiol in a monolayer on gold (0.21 nm2).34 The following factors are believed to contribute to the derived SAM structure: The first is the presence of an amide group in the C2AzoC2SH molecule. This promotes intermolecular hydrogen bonding which has been proved by FTIR and ellipsometry in Rabolt’s work.35 The second is the aromatic interaction among azobenzene chromophores. It may also improve the packing structure of SAMs. The aromatic interaction within azobenzene SAMs has been investigated by Ringsdorf.36 By using an atomic force microscope (AFM), they were able to show that the monolayer lattice is dominated by the end group (azobenzene) rather than dominated by the carbon chain between azobenzene and gold substrate, suggesting that the interaction among azobenzene units effectively influences the structure of the monolayer. As expected for the voltammetric behavior of surfaceconfined redox centers,37 the log-log plot of cathodic peak current versus scan rate has unit slope (Figure 2A). Actually, better evidence for surface-bound redox couples is that the surface concentration Γazo is independent of scan rate in the cyclic voltammetry, as demonstrated by Figure 2B. The quality and degree of packing of these monolayers were also assessed by investigating their blocking effects on the redox electrochemistry of water-soluble Fe(CN)63-/4ions. Figure 3A shows the well-known reversible redox reaction of Fe(CN)63-/4- on a bare gold electrode. Figure 3B demonstrates the voltammetric response obtained after the electrode was exposed to a solution containing 1 mM C2AzoC2SH for 24 h. Obviously, the presence of the monolayer severely hinders the redox reaction of

Results and Discussion Stability and Packing States of SAMs. Figure 1 shows a typical cyclic voltammogram for C2AzoC2SH/Au SAM at pH 5.0 in B-R buffer, where the supporting electrolyte is aqueous 0.1 M NaClO4. For the experimental (32) Wasslach, O.; Kiepenheure, L. Ber. Dtsch. Chem. Ges. 1881, 14, 2617.

(33) Wopschall, R. H.; Shain, I. Anal. Chem. 1967, 39, 1535. (34) Strong, L.; Whitesides, G. M. Langmuir 1988, 4, 546. (35) Lenk, T. J.; Hallmark, V. M.; Hoffmann, C. L.; Rabolt, J. F. Langmuir 1994, 10, 4610. (36) Wolf, H.; Ringsdorf, H.; Delmarche, E.; Takami, T.; Kang, H.; Michel, B.; Gerber, Ch.; Jaschke, M.; Butt, H.-J.; Bamberg, E. J. Phys. Chem. 1995, 99, 7102. (37) Bard, A. J.; Faulkner, L. R. Electrochemical Methods; Wiley: New York, 1980.

Azobenzene Self-Assembled Monolayers on Gold

Langmuir, Vol. 12, No. 11, 1996 2845 Scheme 2

Figure 2. Test for a surface-confined redox center: A, a loglog plot of the peak current versus scan rate for the reduction of C2AzoC2SH/Au SAM is linear with a slope of 1; B, surface coverage Γ is independent of scan rate (pH 5.0, B-R buffer, 0.1 M NaClO4).

Figure 3. (A) Voltammetric response of a bare gold electrode in 2 mM Fe(CN)63- + 0.2 M KCl. (B) Voltammetric response in the same solution of a gold electrode modified by C2AzoC2SH SAM. Scan rate was 50 mV/s.

Fe(CN)63-/4-. The effect is almost as efficient as that observed with a monolayer of octadecanethiol. Effect of pH on the Peak Potential. The cyclic voltammograms of C2AzoC2SH/Au SAMs were obtained

in solutions of different pH. During the potential scan, one anodic peak and one cathodic peak, corresponding to azobenzene/hydrazobenzene redox couple, were observed at the pH range between 3.2 and 8.6. When the pH value was lower than 3.2, the cathodic potential peak was severely overlapped by bulk hydrogen evolution, and two anodic peaks were observed in the meantime. The change of CV profile is probably due to the further reduction of hydrazobenzene in the monolayer.38 The reduction could lead to the cleavage of the nitrogennitrogen bond of the azobenzene unit to form nitrobenzene species as shown in Scheme 2. These products may then re-oxide through reaction pathways similar to 4-nitrothiophenol SAMs on gold.39 Nevertheless, due to the complexities of the further reduction of azobenzene in acidic media, especially in such an unknown as azobenzene SAMs, the full physical scope of these phenomena has not been determined yet. Figure 4 shows the dependence of the anodic peak potential (Ea), the cathodic peak potential (Ec) and the midpoint potential (Ea,c ) (Ea + Ec)/2) on the pH which were obtained at the potential scan rate of 50 mV/s. The values of the dEa/dpH, dEc/dpH, and dEa,c/dpH obtained were -27.01 (correlation coefficient 0.99), -61.83 (0.99), and -42.17 (0.99) mV/pH, respectively. The distinctly small dEa/dpH suggests that the deprotonation in the anodic process was hindered by kinetic reasons or that there exist lower protonation states. Similar phenomena have been observed by Uosaki et al. in their recent investigations of mercaptohydroquinone SAMs on gold.40 The dependence of formal potential E° (assumed to be dEa,c/dpH) on pH yields a slope of ≈45 mV as indicated, which is smaller than the theoretical value of 59 mV/pH for a 2e- + 2H+ process.40 Such a lower slope has also been reported for electrodes modified by diquinone SAMs.17 The deviation from the theoretical value is probably due to the multistepped kinetic process. We believe that the small slope observed for the E°-pH dependence results from different H+ numbers involved in the electrochemical process of azobenzene redox center. It may correspond to an “apparent” or “average” value of the multistep reaction. In other words, azobenzene may have been reduced at the lower protonation states along with its fully protonated states. More important, the dEa,c/dpH value was scarcely affected by the scan rate (Figure 5), which suggests that these different protonation states are kinetically in equilibrium. (See below for the discussion of the dependence of apparent electron-transfer kinetics on pH.) (38) Stradins, J. P.; Glezer, V. T. Encyclopedia of Electrochemistry of the Elements, Organic section, Edit by Bard, A. J., Ed.; Marcel Dekker, Inc.: New York, Vol. XIII. (39) Tsutsumi, H.; Furumoto, S.; Morita, M.; Matsuda, Y. J. Colloid Interface Sci. 1995, 171, 505. (40) Sato, Y.; Fujita, M.; Mizutani, F.; Uosaki, K. Submmited.

2846

Langmuir, Vol. 12, No. 11, 1996

Yu et al.

Figure 4. Variations of the cathodic (Ec), anodic (Ea), and midpoint (Ea,c) potentials of C2AzoC2SH/Au SAM on gold as a function of pH. Scan rate is 50 mV/s (B-R buffer, 0.1 M NaClO4).

Figure 5. Variations of dEa,c/dpH as a function of scan rate for C2AzoC2SH/Au SAM (B-R buffer, 0.1 M NaClO4).

Electrochemical Kinetics of C2AzoC2SH SAMs. The significantly large ∆Ep values that we observed for azobenzene SAMs led us to investigate their possible scan rate dependence in order to assess the influence of electrontransfer kinetics on the voltammetric behavior. Figure 6 shows the scan rate dependence of the cathodic and anodic peak potentials for a C2AzoC2SH/Au SAM. The plot clearly indicates that the cathodic peak potential and anodic peak potential are linearly changed with the logarithmic scan rate. According to Laviron’s treatment,41 the standard heterogeneous rate constant of electron transfer may be obtained in a straightforward manner from eqs 1 and 2 under totally irreversible conditions

Ec ) E°′ Ea ) E°′ +

(

)

RnFν RT ln RnF RTkapp

(

(1)

)

(1-R)nFν RT ln RTkapp (1 - R)nF

(2)

where the notation ν means the potential scan rate and R, T, and F have their usual meanings. Assuming that the product Rn does not vary with potential, the graphs of Ec-ln(ν) and Ea-ln(ν) have been fitted with linear (41) Laviron, E. J. Electroanal. Chem. 1979, 101, 19.

Figure 6. Variations of the cathodic (Ec), anodic (Ea), and midpoint (Ea,c) potentials of C2AzoC2SH/Au SAMs as a function of scan rate (pH 5.0, B-R buffer, 0.1 M NaClO4). Table 1. Kinetic Parameters rn, (1 - r)n, and log kapp at Different pH Values pH

Rn

(1 - R)n

log kapp

3.2 4.4 5.4 6.3 7.2 8.6

0.77 0.99 1.37 1.42 1.37 1.08

0.57 0.81 1.08 1.39 1.27 0.95

-2.85 -4.32 -6.29 -7.68 -7.08 -6.43

regression lines, as shown in Figure 6. The respective slopes of the lines are RT/(RnF) and RT/((1 - R)nF). The values of Rn and (1 - R)n were thus obtained and substituted back in eqs 1 and 2 to solve for kapp. The rate constant kapp was calculated to be 4.7 × 10-5 -1 s at pH 4.4 according to the above procedure. The kapp value of C2AzoC2SH/Au SAMs was much smaller than that of adsorbed unsubstituted azobenzene (∼10 s-1).23 This observation may be attributed to the following factors. First, in the case of C2AzoC2SH SAMs, the azobenzene unit is located apart from the gold electrode surface by the -S(CH2)2N(H)C(O)- group. Generally, such an increase in electron transfer distance should bring about a decrease of the standard rate constant, as shown by a number of researchers.4,8,10 Second, the closely packed SAM structure restricts the conformational change of the azobenzene unit which is associated with electron and proton transfer processes. This restriction may also lead to a decrease of the standard rate constants. In the case of C2AzoC2SH/Au SAMs, however, the effect of distance seems less predominant than the restriction to the conformational change. As reported by Kaifer,13 quinone redox centers immobilized on the external surface of SAMs show a more reversible CV profile, even though the quinone groups are located much further from gold surface. In the studies of ferrocenylazobenzene SAMs in organic solvents by Mirkin,18 the huge peak separation was also observed. We believe that it is also attributed to the sluggish electron-transfer kinetics originating from both long electron-transfer distance and structural inhibition. It is also interesting to investigate the kinetics parameters as a function of the pH value. Table 1 gives the corresponding values for six representative examples of pH values. In the first case, it is evident that log(kapp) shows a V-shaped kinetic behavior in the experimental pH range. As demonstrated by Figure 7, the azobenzene redox reaction is apparently more rapid in acid and alkaline media and becomes very sluggish in the inter-

Azobenzene Self-Assembled Monolayers on Gold

Langmuir, Vol. 12, No. 11, 1996 2847

Figure 8. Reaction sequences for the reduction of C2AzoC2SH/ Au SAM at different pH range.

Figure 7. Variations of log kapp (s-1) as a function of pH (B-R buffer, 0.1 M NaClO4). Scheme 3

mediate pH range. As a matter of fact, similar V-shaped log k-pH behaviors have even been observed for the simple adsorption of unsubstituted azobenzene on glassy carbon electrodes23 and azobenzene LB films on SnO2 glass electrodes.42 We also noted that the values of Rn and (1 - R)n obviously deviate from 1.0 which corresponds to the direct 2e- process (assuming R ∼ 0.5) in the intermediate pH range, indicating the kinetic behavior of azobenzene SAMs deviates from the ideal two-electron electrochemical reactions in these cases. In retrospect, a theoretical analysis is presented of the 1e-, 1H+ electrochemical reaction when the protonations are assumred to be at equilibrium.43-45 It is shown that the reaction behaves as a single monoelectronic process and can be described by using the classical square scheme (Scheme 3). An apparent pH-dependent rate constant has been fully derived45 and is defined by eq 3

kapp ) (1 + Ka1/cH+)-1(1 + cH+/Ka2)-1(k1cH+/Ka2 + k2Ka1/cH+) (3) When ∆pK is large enough, that eq 3 can be applied, if k1 and k2 are not widely different, the variations can be further simplified and given by

log kapp ) log k1 + pKa1 - pH

(4)

log kapp ) log k2 - pKa2 + pH

(5)

and

in the intermediate region between pKa1 and pKa2. The rate constant decreases between the two pKa values of the system and reaches a minimum value in the specific intermediate pH value, as demonstrated by eqs 3-5. Laviron has enriched the theory for the 1e- + 2H+, 2e+ 1H+, and 2e- + 2H+ electrochemical reactions in his series of papers.46-48 Similar pH-dependent kinetic (42) Liu, Z. F. Unpublished data. (43) Laviron, E. J. Electroanal. Chem. (44) Laviron, E. J. Electroanal. Chem. (45) Laviron, E. J. Electroanal. Chem. (46) Laviron, E. J. Electroanal. Chem.

1980, 1981, 1982, 1981,

109, 124, 134, 124,

57. 1. 205. 9.

behavior was investigated from both experimental and theoretical points of view. For the 2e-, 2H+ reactions, azobenzene/hydrazobenzene redox couples as example, Laviron has proceeded the sophisticated theory,48 in which the reduction pathway open to azobenzene was represented by a “nine member square” scheme. Assuming that the protonation reactions are at equilibrium (faster than the electron transfer as we mentioned above), he was able to show that these systems behaved exactly like two sequential one-electron transfers coupled with respective protonation reactions. However, in this case, kapp has a rather complex meaning, since it is a function of the two monoelectronic reaction rate constants and of parameters related to several protonation reactions. The dependence of kapp on pH may have a general feature of these multistepped electron-transfer reactions. As mentioned before, the dEa,c/dpH value of azobenzene SAMs is independent of scan rate, showing that the protonation process is practically at equilibrium. The fact that the variations of kapp vs pH in the present system are globally in accord with theoretical predictions is probably due to a kinetic control by the uptake of either the first or the second electron. A V-shaped kinetic behavior thus was expected and corresponds to the effect of two possible concurrent passways for the electron-transfer and protonation as a simple 1e- + 1H+ reaction. On the basis of a contribution of Laviron,48 the possible reaction pathway for azobenzene reduction was suggested. Figure 8 shows the reduction pathway of an azobenzene moiety under different pH conditions in the present system. The reaction sequences are expected to be H-e-H-e (protonelectron-proton-electron) at low pH, e-H-e-H at higher pH, and e-H-H-e or H-e-e-H at intermediate pH value. Nevertheless, in view of the complex reaction process of azobenzene SAMs, a quantitative and more detail explanation for the phenomenon is now in the process of further investigations. The large separation between anodic peak potential and cathodic peak potential has also been observed by Laviron49 when azobenzene is reduced on a carbon electrode or on a platinum electrode. It was attributed to the interaction between the redox active site and electrode surface. However, in the case of SAMs, C2AzoC2SH molecules are believed to be closely packed, with the carbon chain as a spacer between azobenzene unit and Au substrate, making it possible to eliminate the specific interaction with electrode metal. Therefore, the large peak separation observed in our system may mainly originate from the sluggish electron transfer kinetics. It should be pointed out that the observed peak width at the half-height of CV peak was about 180 mV at a scan (47) Laviron, E. J. Electroanal. Chem. 1982, 146, 1. (48) Laviron, E. J. Electroanal. Chem. 1983, 146, 15. (49) Roullier, L.; Laviron, E. J. Electroanal. Chem. 1982, 134, 181.

2848

Langmuir, Vol. 12, No. 11, 1996

rate of 50 mV/s, considerably larger than the theoretical value of 62.5/Rn mV.37 Although the redox mechanism of the azobenzene unit in SAMs is too complicated to be explained by a single factor, the deviation of peak width can partly be attributed to the interaction between adjacent azobenzene chromophores which is increased by the dense packing of C2AzoC2SH molecules. In the case of C2AzoC2SH SAM, double layer effects are expected to be relatively small, they should have no obvious effects on the width of the reduction-oxidation peak. Another possible explanation of this broadening effect is the thermodynamics and/or kinetic heterogeneity. Both defects or other disorder sites in the monolayer could lead to the heterogeneity in the monolayer. Thus not all of the redox sites have the same effective E°surf and their activities cannot be approximated with the surface concentrations according to the Nerst equation. In the studies of ferrocene SAMs of Chidsey3,4 and quinone SAMs of Kaifer,13 the electroinactive thiols were used to dilute the electroactive sites and to increase the homogeneity by both filling the pinholes and exchanging in defect sites. As a result, the near-ideal peak widths and symmetry of the surface waves were obtained.

Yu et al.

Conclusion In summary, this work has shown that the thiolfunctionalized azobenzene molecule C2AzoC2SH forms stable and closely packed self-assembled monolayers on gold. At pH 3.2-8.6, the SAMs show voltammetric response that corresponds to the two-electron, two-proton reduction-oxidation of azobenzene. The apparent electron-transfer kinetics is very sluggish, which is attributable to the distance between electrode and redox center and to the spatial restriction of close-packing structure on the conformational change accompanied by electrontransfer and protonation. The variation of apparent surface electrochemical rate constants shows a V-shaped behavior in response to pH change. Such a typical feature was explained by pH-dependent reaction kinetics including multistep electron-transfer and protonations. Acknowledgment. The authors gratefully acknowledge the financial supports from the State Science and Technology Committee, the State Education Committee, and the National Natural Science Foundation of China. LA950632C