Photonic and Electrochemical Properties of Adsorbed - American

Jennifer L. Brennan, Tia E. Keyes, and Robert J. Forster*. National Sensor for Sensor Research, School of Chemical Sciences, Dublin City UniVersity,. ...
0 downloads 0 Views 119KB Size
10754

Langmuir 2006, 22, 10754-10761

Photonic and Electrochemical Properties of Adsorbed [Ru(dpp)2(Qbpy)]2+ Luminophores† Jennifer L. Brennan, Tia E. Keyes, and Robert J. Forster* National Sensor for Sensor Research, School of Chemical Sciences, Dublin City UniVersity, Dublin 9, Ireland ReceiVed April 30, 2006. In Final Form: August 25, 2006 Dense monolayers of [Ru(dpp)2Qbpy]2+, where dpp is 4,4′-diphenylphenanthroline and Qbpy is 2,2′:4,4′′:4′4′′quarterpyridyl, have been formed by spontaneous adsorption onto clean platinum microelectrodes. The cyclic voltammetry of these monolayers is nearly ideal, and three redox states are accessible over the potential range of (1.3 V. Chronoamperometry conducted on the microsecond time scale has been used to probe the dynamics of heterogeneous electron transfer and indicates that the standard heterogeneous electron-transfer rate constant, k°, is approximately 106 s-1. The metal complex emits at approximately 600 nm in fluid and solid solution as well as when bound to a platinum electrode surface within a dense monolayer. In the case of the monolayers, it appears that the excited states are not completely deactivated by radiationless energy transfer to the metal because electronic coupling between the adsorbates and the electrode is weak. The dynamics of lateral electron transfer between the electronically excited Ru2+* and ground-state Ru3+ species has been explored by measuring the luminescence intensity after defined quantities of Ru3+ have been produced electrochemically within the monolayer. The rate of lateral electron transfer is between 8 × 106 and 3 × 108 M-1 s-1, indicating efficient electron transfer between adsorbates in close-packed assemblies. Voltammetry conducted at megavolt per second scan rates has been used to directly probe the redox properties of the electronically excited species.

Introduction Charge-transfer processes continue to be of interest because of their importance in areas as diverse as solar energy conversion, photonics, chemical sensors, and biodiagnostics.1 Materials that can absorb photons to create an electronically excited state are of particular interest because they are simultaneously better electron donors and acceptors than their ground-state precursors.2 Many applications demand highly ordered assemblies and also require an understanding of the factors that affect the rate of heterogeneous electron transfer between these materials and an electronically addressable interface (e.g., a metal surface3,4). Immobilized supramolecular assemblies provide a facile way to create these highly ordered redox-active materials.3 However, although much is known about the creation and deactivation of electronically excited states in solution, the factors controlling excited states within monolayers remain largely unexplored. Beyond the electrochemical oxidative or reductive quenching of excited states that is the focus of our interest,5 other ill-defined deactivation pathways exist. If the monolayers are formed on mirrorlike metal surfaces, then excited-state quenching is expected to be rapid if the distance between the excited state and the surface is short enough to allow heterogeneous electrontransfer quenching to occur within the lifetime of the electronically excited state.6,7 Balancing these requirements of fast heteroge†

Part of the Electrochemistry special issue. * To whom correspondence should be addressed. E-mail: robert.forster@ dcu.ie. (1) Adams, D. A.; Brus, L.; Chidsey, C. E. D.; Creager, S.; Creutz, C.; Kagan, C. R.; Kamat, P. V.; Lieberman, M.; Lindsay, 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. (2) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; Von Zelewsky, A. Coord. Chem. ReV. 1988, 84, 85. (3) Vos, J. G.; Forster, R. J.; Keyes, T. E. Interfacial Supramolecular Assemblies; Wiley-VCH: Weinheim, Germany, 2003. (4) Murray, R. W. In Molecular Design of Electrode Surfaces; Murray, R. W., Ed.; Wiley-Interscience: New York, 1992. (5) Forster, R. J.; Keyes, T. E. J. Phys. Chem. B 1998, 102, 10004.

neous electron transfer (for which short electron-transfer distances are optimal) versus a large distance between metal surface and the excited state to minimize energy-transfer quenching places severe demands on the experimental time scale and has blocked progress in this area. However, it is expected that efficient energy and electron-transfer quenching of the immobilized exited state will occur only for emitting molecules close to near-perfect metallic mirrors. In fact, there are a number of reports where the emission from immobilized electronically excited species has been enhanced by their attachment to roughened surfaces.8-10 We have previously demonstrated that laser excitation of spontaneously adsorbed monolayers of [Ru(bpy)2(Qbpy)]2+ on nonmirror finish platinum microelectrodes leads to emission (bpy is 2,2’-bipyridyl and Qbpy is 2,2′:4,4′′:4′,4′′-quarterpyridyl5). We postulate that the structure of this Qbpy bridging ligand is key to observing emission, suggesting that controlling the structure of the bridging ligand is just as important as controlling the distance if near-surface emission is to be achieved. In this contribution, we report the synthesis and characterization of a new ruthenium complex, [Ru(dpp)2(Qbpy)]2+ (Chart 1, dpp is 4,7-diphenyl-1,10-phenanthroline). This complex is both redoxactive and highly luminescent and forms stable spontaneously adsorbed monolayers on electrode surfaces. The ground-state heterogeneous electron-transfer and excited-state emission properties of the complex are reported. Beyond the issues of excited-state quenching by the metal surface discussed earlier, it is possible that lateral energy or electron-transfer quenching between adsorbates can occur. To provide insight into this process, we have investigated the extent of electron-transfer quenching of the electronically excited species, Ru2+*, by Ru3+ within the (6) Drexhage, K. H. J. Lumin. 1970, 1-2, 693. (7) Chance, R. R.; Prock, A.; Silbey. R. In AdVances in Chemical Physics; Prigogine, I., Rice, S. R., Eds.; Wiley: New York, 1978; pp 1-65. (8) Wokaun, A.; Lutz, H. P.; Wild, U. P.; Ernst, R. R. J. Chem. Phys. 1983, 79, 509. (9) Pope, J. M.; Buttry, D. A. J. Electroanal. Chem. 2001, 498, 75. (10) Naujok, R. R.; Duevel, R. V.; Corn, R. M. Langmuir 1993, 9, 1771.

10.1021/la0611918 CCC: $33.50 © 2006 American Chemical Society Published on Web 09/27/2006

Photonics and Electrochemistry of Adsorbed Luminophores Chart 1. Structure of [Ru(dpp)2(Qbpy)]2+

monolayer. Furthermore, with respect to these surface-bound emission studies, by applying an ultrafast (MV s-1 scan rate) potential sweep to the electrode following laser excitation we have been able to directly determine the oxidation potential of the Ru2+* electronically excited state, a parameter that is mainly either approximated from ground-state electrochemical data using the well-known Rehm-Weller equation11 or indirectly determined from kinetic studies of excited-state quenching by solution-phase quenchers.12 Jones and Fox have shown that phase-modulated voltammetry can be used to directly measure the excited-state redox potential.13 However, this technique is essentially steady state and is applicable only when the excited-state lifetime exceeds 500 ns. Together, these investigations provide powerful new insights into heterogeneous and lateral electron transfer in both the ground and electronically excited states. Experimental Section cis-Ru(dpp)2Cl2 was prepared by standard synthetic methods. All chemicals were purchased from Sigma-Aldrich and used without further purification. 2,2′:4,4′′:4′,4′′-Quarterpyridyl (Qbpy) was prepared using a modification to the method described by Morgan et al.14 4,4’Bipyridyl (5 g, 0.02 M) was placed in a sealed Teflon bomb over dried Pd/C (1 g, 10% Pd), and the entire apparatus was heated in an oven at 200 °C for 1 week. The product was dissolved in chloroform and filtered to remove the catalyst. Unreacted 4,4’-dipyridyl was removed by loading the chloroform-dissolved product onto a 500 × 3 cm2 silica gel column and eluting the 4,4′-bipyridyl and Qbpy fractions with methanol. The Qbpy fractions were combined, extracted, and recrystallized from acetone. Yield, 1.2 g, 20%, mp 235 °C. 1H NMR (CDCl3): δ 7.55 (2H, d, HF, HF′), 7.64 (4H, d, HC, HC′, HD, HD′), 8.69 (4H, d, HA, HA′, HB, HB′), 8.71 (2H, s, HE, HE′), 8.77 (2H, d, HG, HG′), Anal Calcd for C20H16N4O: C, 73.3; H, 4.8; N, 17.07. Found: C, 74.2; H, 4.64; N. 17.18. [Ru(dpp)2(Qbpy)](PF6)2Ru(dpp)2Cl2 (0.25 g, 0.298 mmol) and 2,2′:4,4′′:,4′,4′′-quarterpyridyl (0.1854 g, 0.5974 mmol) were dissolved in 50 cm3 of a 1:1 ethanol/water mixture and heated to reflux for 8 h. The solvent was removed to a volume of 10 cm3, and the product was precipitated by addition of an aqueous solution of concentrated NH4PF6. The resulting orange solid was collected by filtration and recrystallized from 2:1 v/v acetone/water. Yield 0.695 g, 87%. 1H NMR (CDCl3): δ 7.50 (20H, m, HPh), 7.63 (2H, d, HH), 7.76 (4H, d, HF, HF′), 7.78 (4H, d, HA, HA′), 7.83 (2H, d, HC), 8.22 (4H, d, HB, HB′), 8.24 (2H, s, HC′), 8.43 (2H, d, HG), 8.65 (4H, d, HE, HE′), 8.69 (2H, d, HD). Anal Calcd for RuC68H46N6P2F12: C, 61.03; H, 3.44; N, 6.28. Found: C, 61.25; H, 3.89; N. 6.07. (11) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259. (12) Bock, C. R.; Meyer, T. J.; Whitten, D. J. Am. Chem. Soc. 1975, 97, 2909. (13) Jones, W. E.; Fox, M. A. J. Phys. Chem. 1994, 98, 5095. (14) Morgan, R. J.; Baker, A. D. J. Org. Chem. 1990, 55, 1986.

Langmuir, Vol. 22, No. 25, 2006 10755 Apparatus. 1H NMR measurements were carried out on a Bruker 400 NMR spectrometer. Measurements were carried out in deuterated dimethyl sulfoxide (d6-DMSO) or deuterated chloroform (d3chloroform). Room-temperature emission spectra were recorded using a PerkinElmer LS50 B luminescence spectrometer equipped with a redsensitive Hamamatsu R928 detector. Samples were prepared at concentrations of 1 × 10-5 M in spectroscopic grade butyronitrile. Where necessary, deaeration was achieved by bubbling with nitrogen or argon for 20 min prior to use. For room-temperature measurements, 1 cm quartz cells were used. The excitation and emission slits were 2 nm for all experiments. Luminescence lifetimes and lowtemperature emission spectra for the monomeric complex were measured using the third harmonic (355 nm) of a Spectron Q-switched Nd:YAG laser for excitation. Emission was detected in a right-angle configuration with respect to the laser using an Oriel model IS520 gated intensified CCD coupled to an Oriel model MS125 spectrograph. With suitable signal averaging, this configuration allows a complete emission spectrum (spectral range 250 nm) to be obtained within times as short as 10 ns. The emission spectra were typically recorded using the average of 20 laser shots. The gatewidth (i.e., the exposure time of the CCD) was never more than 5% of the excitedstate lifetime. The step size (i.e., the time between the acquisition of discrete spectra) was typically 5% of the excited-state lifetime. Low-temperature emission lifetime studies were carried out using an Oxford Instruments gas-exchange cryostat equipped with an ITC502 temperature controller. Standard iterative techniques were employed to determine the lifetimes of emission.15 Cyclic voltammetry was performed using a CH Instruments model 660A electrochemical workstation and a conventional three-electrode cell. All solutions were deoxygenated thoroughly using argon, and a blanket of argon was maintained over the solution during all experiments. Potentials are quoted with respect to a CH Instruments Ag/AgCl reference electrode filled with saturated KCl that had a potential of +0.190 V with respect to the normal hydrogen electrode. The potential of the ferrocene/ferrocenium couple at this electrode was 0.340 V. All experiments were performed at room temperature (22 ( 3 °C). In high-speed electrochemical measurements, a custom-built function generator/potentiostat, with a rise time of less than 10 ns, was used to apply potential steps (chronoamperometry) or triangular waveforms (cyclic voltammetry) of variable pulse width and amplitude directly to a two-electrode cell.16 A Pt flag and a Ag/AgCl reference electrode were combined to form a counter electrode. The flag lowered the resistance and provided a high-frequency path. A Hewlett-Packard 54201A digitizing oscilloscope was used to monitor the applied waveform and capture the resulting voltammogram. The data were transferred to a PC using a National Instruments GPIB232CT controller and were analyzed in Microsoft Excel. In studies of the voltammetry of electronically excited species, a Stanford Research Instruments model DG 535 four-channel digital pulse generator/delay generator was used as the controller to trigger the laser pulse and the application of the potential sweep simultaneously. The optical path (approximately 3.0 m) was sufficiently short that it did not introduce any significant difference between the triggering and arrival of the laser pulse. Microelectrodes were fabricated from platinum microwires of radii between 5 and 25 µm (Goodfellows Metals Ltd.) by sealing them in soft glass using a procedure described previously.17 Microdisk electrodes were exposed by removing excess glass using 600 grit emery paper, followed by successive polishing with 12.5, 5, 1, 0.3, and 0.05 µm alumina. The polishing material was removed between changes of particle size by sonicating the electrodes in deionized water for at least 1 min. Electrochemical cleaning of the electrodes was carried out by cycling in 0.1 M H2SO4 between potential limits chosen initially to oxidize and then reduce the surface of the platinum (15) Diamond, D.; Hanratty, V. C. A. Spreadsheet Applications in Chemistry Using Microsoft Excel; Wiley: New York, 1997. (16) Xu, C. Ph.D. Thesis, University of Illinois at Urbana-Champaign, 1992. (17) O’Hanlon D. P. Ph.D. Thesis, Dublin City University, Dublin, Ireland, 1999.

10756 Langmuir, Vol. 22, No. 25, 2006

Brennan et al.

Figure 1. Cyclic voltammogram of 5 mM [Ru(dpp)2(Qbpy)]2+ dissolved in deaerated acetonitrile containing 0.1 M tetrabutylammonium perchlorate as the supporting electrolyte. The working electrode is a 25-µm-radius platinum microdisk. The scan rate is 0.1 V s-1. Cathodic currents are up, and anodic currents are down. electrode. Excessive cycling was avoided to minimize the extent of surface roughening. The real surface area of the electrodes was determined by calculating the charge under the platinum oxide reduction peak.18 Typical surface roughness values were between 1.2 and 1.4. Determining the real as opposed to the geometric area of the electrodes is important if the area of occupation of the adsorbate is to be accurately determined. The final step in the cleaning procedure was repetitive cycling between -0.2 and +1.0 V in 0.1 M LiClO4 until hydrogen desorption was complete and a flat background was obtained.19 Spontaneously adsorbed monolayers of [Ru(dpp)2(Qbpy)]2+ were formed by immersing a clean microelectrode in a 1 mM solution of the complex in 1:1 v/v acetone/water for up to 30 min. The electrode was removed from the deposition solution, rinsed with 1:1 v/v acetone/ water, and placed in deoxygenated 1.0 M LiClO4. Where necessary, to stabilize the monolayers over extended periods, 50 µM of the complex was added to the supporting electrolyte solution. A low concentration of the surface-active complex in solution minimizes the diffusional contribution to the overall current in chronoamperometry or cyclic voltammetry and improves the stability of the monolayers. The complex is stable toward aerial oxidation, and no precautions were taken to exclude atmospheric oxygen during monolayer formation. In experiments where the effect of the presence of Ru3+ on the emission intensity was explored, a macroscopic platinum electrode (3 mm radius) was employed. Steady-state emission spectra were recorded as a function of the applied potential using the LS50B.

Results and Discussion General Electrochemical Properties. Figure 1 shows a representative cyclic voltammogram for a 5 mM solution of [Ru(dpp)2(Qbpy)]2+ in acetonitrile containing 0.1 M tetrabutylammonium perchlorate as the supporting electrolyte. Typical of RuN6 complexes, the formal potential, E°′, of the Ru2+/3+ process is +1.235 V versus Ag/AgCl, which is in agreement with that previously reported for a similar complex, [Ru(bpy)2(Qbpy)]2+,20 where bpy is 2,2’-bipyridyl. Obtaining well-resolved peaks for the ligand-centered redox processes proved difficult because of the tendency of Qbpy to adsorb onto the electrode surface. However, well-defined reductions occur at -0.955 and -1.085 V versus Ag/AgCl. In heteroleptic systems of this kind, it is possible that the first ligand-based reduction occurs on either the Qbpy or the dpp ligand. On the basis of previous measurements of [Ru(bpy)2(Qbpy)]2+,5,20 the first reduction at -0.955 V is (18) Trasatti, S.; Petrii, O. A. J. Electroanal. Chem. 1992, 327, 354. (19) Tirado, J. G.; Abrun˜a, H. D. J. Phys. Chem. 1996, 100, 4556. (20) Bierig, K.; Morgan, R. J.; Tysoe, S.; Gafney, H. D.; Strekas, T. C.; Baker, A. D. Inorg. Chem. 1991, 30, 3898.

Figure 2. Background-corrected cyclic voltammograms for a spontaneously adsorbed [Ru(dpp)2(Qbpy)]2+ monolayer (surface coverage 1.0 ( 0.1 × 10-10 mol cm-2) on a 5-µm-radius Pt microdisk electrode. The supporting electrolyte is aqueous 1.0 M LiClO4. Scan rates from top to bottom: 10, 5, 2, and 1 V s-1.

attributed to the Qbpy ligand, and the second is assigned to one of the dpp ligands. Figure 2 illustrates the effect of the voltammetric scan rate on the cyclic voltammograms for spontaneously adsorbed monolayers of [Ru(dpp)2(Qbpy)]2+ where the supporting electrolyte is aqueous 1.0 M LiClO4. The experimental voltammetric response observed for the monolayer demonstrates all of the characteristics of a redox system confined to an electrode surface.21 For example, the peak shapes are symmetrical and independent of scan rate, υ, at least over the range of 1 to 100 V s-1, and the peak height scales linearly with scan rate, unlike the υ1/2 dependence expected for a freely diffusing species. The difference between the anodic and cathodic peak potentials, ∆Ep, is ca. 45 mV, and the fwhm is ca. 110 mV, which is about 20 mV larger than the ideal value.22,23 These observations suggest that the Qbpy complex adsorbs onto the surface of a platinum microelectrode to give an electroactive film. Interestingly, the formal potential of the Ru2+/3+ redox process in the monolayer is shifted in the negative potential direction by approximately 135 mV with respect to the value obtained for the complex dissolved in acetonitrile. This is common in systems of this kind and can be attributed to differences in the dielectric constant as well as in the donation of electron density from the electrode to the metal center.24,25 The peak height and area of the wave centered at 1.100 V do not change by more than 10% after repetitive cycling for up to 5 h, demonstrating excellent monolayer stability. The Faradaic charge associated with the Ru2+/3+ redox process has been estimated from the area under the wave at 1.100 V after correcting for double-layer charging. This charge, together with the real surface area of the electrode, has been used to calculate the surface coverage (mol/cm2) of [Ru(dpp)2(Qbpy)]2+. The limiting surface coverage, Γ, is (1.1 ( 0.1) × 10-10 mol cm-2, corresponding to an area of occupation of 151 ( 15 Å2. This area of occupation is approximately 10% larger than that expected for a close-packed monolayer on the basis of structurally related (21) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.; Wiley and Sons: New York, 2001. (22) Laviron, E. J. Electroanal. Chem. 1974, 52, 395. (23) Brown, A. P.; Anson, F. C. Anal. Chem. 1977, 49, 1589. (24) Forster, R. J.; Faulkner, L. R. J. Am. Chem. Soc. 1994, 116, 5444. (25) Walsh, D. A.; Keyes, T. E.; Forster, R. J. J. Phys. Chem. B 2004, 108, 2631.

Photonics and Electrochemistry of Adsorbed Luminophores

Langmuir, Vol. 22, No. 25, 2006 10757

Figure 3. Current response for the Ru3+ reduction occurring within a [Ru(dpp)2(Qbpy)]2+ monolayer formed on a 5 µm Pt microdisk electrode where the potential is stepped from +1.200 to +0.800 V. The dashed line illustrates the double-layer charging response for a modified electrode following a step from 0.800 to +0.400 V. The supporting electrolyte is aqueous 1.0 M LiClO4. The inset shows the semilog plot for the Faradaic reaction.

systems and crystallographic data26,27 that indicate that the radii of osmium and ruthenium polypyridyl complexes are on the order of 6.7 Å. This additional area of occupation most likely arises from solvation and the incorporation of charge-compensating counterions. Heterogeneous Electron-Transfer Dynamics. An attractive characteristic of spontaneously adsorbed monolayer systems of this kind is the ability to probe distance and bridging-ligand structure effects of the rate of heterogeneous electron transfer across the electrode/monolayer interface. For example, in seeking to create plasmonically enhanced emission from bound luminophores by nanostructuring the electrode surface, it may be desirable to achieve weak electronic coupling so as to block quenching by the metal surface. A convenient method of investigating the rate of these heterogeneous electron transfers is chronoamperometry, where a potential step is applied to the electrode and the resulting current is monitored as a function of time. For an ideal electrochemical reaction following a surfacebound species, the Faradaic current following a potential step that changes the redox composition of the monolayer exhibits a single-exponential decay in time according to24,25,28

iF(t) ) kQ exp(-kt)

(1)

where k is the apparent rate constant for the overall reaction for a given overpotential, η, (η ) E - E°) and Q is the total charge passed in the redox transformation. Figure 3 illustrates a typical example of the chronoamperometric response observed for the Ru3+ + e- f Ru2+ redox reaction of a [Ru(dpp)2(Qbpy)]2+ monolayer formed on a 5-µmradius microelectrode, where the electrolyte is 1.0 M LiClO4 and the overpotential is -300 mV. In these experiments, the potential was stepped from 1.100 V to potentials, E, that were successively (26) Goodwin, H. A.; Kepert, D. L.; Patrick, J. M.; Skelton, B. W.; White, H. Aust. J. Chem. 1984, 37, 1817. (27) Ferguson, J. E.; Love, J. L.; Robinson, W. T. Inorg. Chem. 1972, 11, 1662. (28) Chidsey, C. E. D. Science 1991, 251, 919.

more negative than the formal potential, E°′. The instability of the monolayer to the application of potentials >1.1 V over the millisecond chronoamperometric time scales prevented the characterization of the potential dependence of the rate constant for positive overpotentials. In carrying out these measurements, it is important to consider the effect of the ohmic drop on the observed response.21,24 An ohmic drop is caused when currents flow through a solution generating a potential that decreases the applied potential by an amount iR, where i is the total current flow in the solution and R is the solution resistance.29 This ohmic drop can lead to distortions of the experimental response and inaccurate measurements of the heterogeneous electron-transfer rate constant. The ohmic drop effects have been minimized in these experiments by using a high supporting electrolyte concentration (1.0 M), by using electrodes of small radius that yield small experimental currents, and by analyzing data relatively late in the lifetime of the transient, where currents are smaller and therefore iR is reduced. At high overpotentials, the Faradaic response is temporally convolved with the double-layer response, and accurate estimates of the heterogeneous electron-transfer rate constant cannot be obtained. The inset of Figure 3 shows the semilog current versus time response on time scales where double-layer charging is complete and the current is dominated by the Faradaic reaction. The linearity of the semilog plots indicates that the heterogeneous electron transfer associated with the Ru3+ reduction is a first-order process. As uncompensated resistance causes the applied potential and hence the apparent rate to evolve with time, nonlinear responses would be expected if substantial ohmic drop effects were present. Beyond the issue of the iR drop, the linear responses in Figure 3 indicate that individual adsorbates within the monolayer have similar local microenvironments and electron-transfer distances. The heterogeneous electron-transfer rate constants (determined from the absolute value of the semilog plot slope) indicate that electron transfer occurs very rapidly in this system. Even for a modest overpotential of 60 mV, the heterogeneous electrontransfer rate constant, k, for the reduction of the ruthenium centers is 1.3 ( 0.2 × 106 s-1. As given by eq 1, the intercept of these semilog plots equals ln(kQ). Calculating the total charged passed from these intercepts and comparing them with slow-scan cyclic voltammetric data for the monolayer indicates that all of the adsorbates are electrochemically active on the microsecond time scale. The Butler-Volmer formulation of electrode kinetics predicts an exponential dependence of k, the heterogeneous electrontransfer rate, on the overpotential, η. Equation 2 describes the response expected for a one-electron reduction21

[

k ) k° exp -

]

RcnFη RT

(2)

where k° is the standard heterogeneous rate constant, Rc is the cathodic transfer coefficient, and n, F, R, and T have their usual significance. A Tafel plot of ln k versus overpotential is predicted to be linear with a slope of -RcnF/RT for the cathodic step and (1 - Ra)nF/RT for the anodic step. Extrapolating either branch to zero overpotential allows k° to be estimated.21,24 Figure 4 illustrates a Tafel plot of ln k versus overpotential for the reduction of Ru3+ to Ru2+. The Tafel plot shows that ln k depends approximately linearly on overpotential in the range of 60 < η < 175 mV. In contrast to the prediction of the Butler-Volmer (29) Forster, R. J. Ultrafast Electrochemical Techniques. In The Encyclopaedia of Analytical Chemistry; Wiley and Sons: New York, 1998.

10758 Langmuir, Vol. 22, No. 25, 2006

Brennan et al.

Ru3+

Figure 5. Emission spectra of 10 µM [Ru(dpp)2(Qbpy)]2+ in deoxygenated butyronitrile at 298 K (thick line, axis to right) and 77 K (thin line, axis to left).

formulation, at higher overpotentials, k becomes independent of the driving force.3,5,21,24,28 Such curvature has been previously observed in Tafel plots for other monolayer systems24,28,30-32 and is consistent with the presence of the Marcus inverted region.24,28 In heterogeneous electron transfer, at high overpotentials, where the overpotential is more than half the reorganization energy, the rate of electron transfer becomes independent of the driving force and reaches a maximum when the overpotential equals the reorganization energy. This reflects the fact that the entire distribution of redox-active molecules with states available for electron transfer is matched by states in the electrode that are also able to transfer electrons. Increasing the driving force does not make any new states available for electron transfer, so the rate does not change.3,30,33 Using the data in the linear portion of the graph and extrapolating it to zero overpotential allows the standard heterogeneous electron-transfer rate constant, k°, to be estimated. This analysis yields a value of (9.1 ( 0.9) × 105 s-1. This result is similar to the value previously reported5 for [Ru(bpy)2Qbpy]2+ monolayers, (5.1 ( 0.3) × 105 s-1. The relative insensitivity of k° to the peripheral ligands, bpy versus dpp, indicates that electron transfer across the Qbpy bridging ligand represents the ratedetermining step and that subtle changes in the electron density of the Ru center do not dramatically affect the electron transfer rate. Whereas this is certainly a relatively fast electron transfer, the absolute value of k° is dramatically lower than one would expect for an adiabatic electron-transfer reaction at this distance. For example, the Marcus equation predicts an outer sphere reorganization energy on the order of 0.25 eV.28,33 For an adiabatic electron-transfer reaction, this reorganization energy would lead to a standard heterogeneous electron-transfer rate constant in excess of 1010 s-1 (i.e., about 5 orders of magnitude larger than that reported here). The substantial difference indicates that electron transfer across the Qbpy ligand is highly nonadiabatic and involves weak coupling of Ru and electrode states. The Tafel slope yields an Rc value of 0.2 that is strikingly lower than that expected for a reversible electron transfer, 0.5. Further experiments are in progress to elucidate the origin of this weak potential dependence of the heterogeneous electron-transfer rate constant. If Rc was 0.5, then this would lead to an approximately

3-fold decrease in the observed value of k° to around 3 × 105 s-1, a value still consistent with previous reports.5 Luminescence Properties. One of the striking features of the Qbpy ligand is that it appears to support only very weak electronic communication in the ground state. If this weak coupling also occurs for the electronically excited state species, then it is possible that emission could be observed from surface-immobilized complexes despite their close proximity to the metal surface. Figure 5 illustrates emission spectra of the quarterpyridyl complex in butyronitrile at 77 and 298 K, which exhibit emission maxima at 575 and 615 nm, respectively. The emission at 77 K exhibits a shoulder at 620 nm, indicative of the vibrational fine structure in the spectrum. This vibrational structure is assigned to a perturbed skeletal vibration of the dpp aromatic ring due to the removal of the π* electron and is usually visible only at low temperatures.34 The lifetime of the electronically excited state of the complex in solution has been determined using time-resolved luminescence spectroscopy. For all temperatures investigated, plots of the natural log of the emission intensity versus time were linear, and the luminescent lifetime, τ, was determined from the slope. Figure 6 depicts the observed lifetimes over a range of temperature from 77 to 298 K and demonstrates the expected decrease in emission lifetime as the sample temperature is increased. This may be attributed to either (a) a thermally activated surface crossing to another excited state (usually from 3MLCT to 3MC) or (b) the role of vibrational modes that favor radiationless decay, which are prevented at low temperature because of the frozen molecular environment.35 The abrupt change in lifetime between 100 and 200 K is due to the phase change between the frozen and liquid solvent occurring near the melting point of the butyronitrile solvent at 161 K. The excited-state lifetimes for the complex are 12.4 ( 1.9 µs and 392 ( 21 ns at 77 and 298 K, respectively. On the basis of previous measurements,5 we expect the lifetime of surface-immobilized [Ru(dpp)2(Qbpy)]2+ to be close to that found in frozen media (i.e., more than a microsecond). Lateral Electron-Transfer Quenching. One advantage of redox-active monolayers is that they can be partially oxidized to create acceptor states (i.e., Ru3+) for photoinduced electron transfer from the electronically excited state, Ru2+*. In this way, insight into the efficiency of lateral electron transfer between

Figure 4. Tafel plot for the reduction occurring within a [Ru(dpp)2(Qbpy)]2+ monolayer formed on a 5 µm Pt microdisk electrode. The supporting electrolyte is aqueous 1.0 M LiClO4.

(30) Weber, K.; Creager, S. E. Anal. Chem. 1994, 66, 3164. (31) Forster, R. J. Inorg. Chem. 1996, 35, 3394. (32) Forster, R. J.; O’Kelly, J. P. J. Phys. Chem. 1996, 100, 3695. (33) Forster, R. J.; Loughman, P.; Keyes, T. E. J. Am. Chem. Soc. 2000, 122, 11948.

(34) Seddon, E. A.; Seddon, K. R. The Chemistry of Ruthenium; Elsevier: New York, 1984. (35) Barigelletti, F.; Belser, P.; Von Zelewsky, A.; Juris, A.; Balzani, V. J. Phys. Chem. 1985, 89, 3680.

Photonics and Electrochemistry of Adsorbed Luminophores

Langmuir, Vol. 22, No. 25, 2006 10759

Figure 8. Dependence of I0/I on the Ru3+ quencher concentration for [Ru(dpp)2Qbpy]2+ monolayers in contact with aqueous 0.1 M LiClO4. The intensity ratio is corrected for the change in Ru2+ concentration because Ru3+ is created electrochemically.

Figure 6. Temperature dependence of the excited-state lifetime for a solution/glass of 10 µM [Ru(dpp)2(Qbpy)]2+ in deoxygenated butyronitrile.

luminophores within the assembly. Taking the saturation coverage of 1.1 × 10-10 mol cm-2 and the monolayer thickness of approximately 15 Å, the concentration of ruthenium centers is approximately 0.7 M. The quenching response of Figure 7 can be described by the Stern-Volmer equation

Io ) 1 + KSV[Ru3+] I

Figure 7. Effect of partially oxidizing a [Ru(dpp)2Qbpy]2+ monolayer deposited on a 3 mm diameter platinum electrode in contact with an aqueous 0.1 M LiClO4 solution on the emission spectrum. From top to bottom, the percentage of Ru3+ centers is 0, 3.7, 6.5, and 11.

adjacent adsorbates can be obtained. Information of this kind is important given that the concentration of luminophores within the assembly is approximately 6 orders of magnitude larger than that used for typical solution-phase experiments. Figure 7 illustrates the effect of electrochemically titrating in defined concentrations of Ru3+ into the film on the intensity of the observed emission. The key feature of Figure 7 is that the adsorbed complexes clearly emit despite being only approximately 12 Å above a metal surface. Our current data do not allow us to determine if the luminescence lifetime of the emitting centers is significantly lower than that found in solution (i.e., the luminescence may be substantially, but not completely, quenched). A significant observation is that the monolayer spectrum is mostly closely related to that observed at 77 K (i.e., the emission maxima and peak shape are similar). This similarity in the spectra suggests that the vibrational degrees of freedom of complexes within the monolayer are significantly reduced. To compare the quenching effects observed in these monolayers with previous reports dealing with reactants in solution or within thick films,36-39 we use the effective 3D concentration of (36) Majda, M.; Faulkner, L. R. J. Electroanal. Chem. 1982, 137, 149.

(3)

where Io and I are the fluorescence intensities in the absence and presence of quencher, respectively, and KSV is the Stern-Volmer constant. It is important to note that because the composition of the film is being controlled electrochemically some of the loss of luminescence intensity arises because the Ru2+ concentration decreases as Ru3+ is created. This Figure shows that luminescence intensity decays much more rapidly than would be expected on the basis of the loss of Ru2+* species alone. For example, the emission intensity decreases by more than 90% when only approximately 10% of the luminophores within the film are oxidized. This result indicates that the Ru3+ species acts as an efficient electron acceptor and quenches the emission (i.e., Ru2+* + Ru3+ f Ru3+ + Ru2+). Given that the luminophores are bound within a supramolecular assembly, it is likely that quenching by Ru3+ occurs predominantly through a static mechanism. Figure 8 illustrates the Stern-Volmer plot that is highly linear and has a slope, KSV, of 105 ( 7 M-1. The quantum yield of emission is proportional to the emission intensity, so the SternVolmer equation can be written as

Io ) 1 + kqτ0[Ru3+] I

(4)

As discussed above, the excited-state lifetimes for the complex in solution are 12.4 ( 1.9 µs and 392 ( 21 ns at 77 and 298 K, respectively. Taken in conjunction with the Stern-Volmer constant, the electron-transfer rate constant is between 8 × 106 and 3 × 108 M-1 s-1. Whereas this rate constant compares favorably with those reported elsewhere for ruthenium polypyridyl complexes in solution, solids, and thin films,36-40 when describing lateral quenching it is more appropriate to consider the process as occurring in 2D space. The 3D kq values above can be converted to 2D values using the following expression41 (37) Majda, M.; Faulkner, L. R. J. Electroanal. Chem. 1984, 169, 97. (38) Buttry, D. A.; Anson, F. C. J. Am. Chem. Soc. 1982, 104, 4824. (39) Altamirano, M.; Senz A.; Gsponer, H. E. J. Colloid Interface Sci. 2004, 270, 364. (40) Newsham, M. D.; Cukier, R. I. Nocera, D. G. J. Phys. Chem. 1991, 95, 9660.

10760 Langmuir, Vol. 22, No. 25, 2006 3D

Brennan et al.

kq ) 2Dkq2R

(5)

where R is the radius of an excluded volume encompassing both reactants. Assuming that R ) 2r, where r is the radius of the ruthenium complex (6.9 Å), leads to 2Dkq values between 3 × 1016 and 1 × 1018 cm2 mol-1 s-1. Excited-State Redox Properties. One of our interests in systems of this kind that can support heterogeneous electron transfers on time scales shorter than the luminescence lifetime and yet show significant emission is the direct measurement of excited-state redox properties. The effect of parameters such as the location and nature of the excited state and its luminescence lifetime and redox potential on the excited-state electron-transfer dynamics remains largely unexplored. To address this issue, our approach is to use a pulse of laser light to create the excited states and then use rapid scan voltammetry to directly measure the oxidation potential of the excited state before it decays back to the ground state. It is well established that irrespective of the excited-state lifetime oxidative electron-transfer quenching of the electronically excited state creates a ground-state oxidized product at a potential that is significantly negative of its formal potential. Therefore, this ground state is reduced at the electrode and the complete cycle involves two electron transfers but no net charge transfer. Our excited-state voltammetry approach will yield information only about the excited-state potential if the oxidation of the excited state occurs more rapidly than the reduction of the ground-state product. In this way, the two electron transfers are time-resolved, and a wave can be seen in the cyclic voltammogram. Ideally, this experiment directly measures the excited-state redox potential that can then be compared with that predicted on the basis of ground-state electrochemical and lowtemperature emission data. It is important to consider where on the potential axis a voltammetric response corresponding to the electronically excited state would occur. The most widely used method to estimate the excited-state redox potentials is the Rehm-Weller equation,11 which is given by eq 6 for the oxidative quenching of an excited state o

E

Ox*

) E°′Ox - E

0-0

(6)

where E°′Ox is the formal redox potential for the Ru2+/3+ couple and E0-0 is the energy gap between the zeroeth vibrational levels of the ground and excited states, estimated from emission spectra obtained at cryogenic temperatures. Equation 6 assumes that all the spectroscopic energy of the excited state (E0-0) can be used as free energy in the redox process.42 Utilizing the ground-state redox data in Figure 1 and the maximum emission wavelength at 77 K from Figure 5, yields E°Ox* and E°Red* values of -0.901 and +1.106 eV, respectively. These values are only estimates of the true excited-state redox potential and typically contain uncertainties of 100 mV or more.13 Figure 9 shows the 3 × 105 V s-1 voltammetric response of a [Ru(dpp)2 Qbpy]2+ monolayer formed on a 5-µm-radius platinum microelectrode following laser excitation. The voltammetric scan was triggered so that it commenced simultaneously with the arrival of the incident laser pulse on the electrode surface. The potential limits are chosen so that the monolayer exists in the Ru2+ state prior to optical excitation and allows the excitedstate response expected at approximately -0.9 V to be captured (41) Charych, D. H.; Landau, E. M.; Majda, M. J. Am. Chem. Soc. 1991, 113, 3340. (42) Ballardini, R.; Varani, G.; Indelli, M. T.; Scandola, F.; Balzani, V. J. Am. Chem. Soc. 1978, 100, 7219.

Figure 9. Cyclic voltammetry of a 5-µm-radius platinum microelectrode modified with a [Ru(dpp)2Qbpy]2+ monolayer following laser excitation at 355 nm (thick solid line). The open circles illustrate the second scan following laser excitation. The dashed line illustrates the voltammetric response obtained for a bare electrode under identical conditions. The scan rate is 3 × 105 V s-1, the surface coverage is 1.1 × 10-10 mol cm-2, and the supporting electrolyte is 0.1 M TBABF4 in acetonitrile.

without complications from electron transfer to the Qbpy or dpp ligands. Given that this initial potential is negative of the excitedstate oxidation potential, quenching of the excited state by electron transfer to the electrode surface is not expected. In this experiment, the voltammetric time scale is shorter than the excited-state lifetime, and voltammetry characteristic of the electronically excited state can be obtained. The overall electron-transfer reaction sequence expected is as follows:

process 1: Ru2+* - e- f Ru3+ process 2: Ru

3+

-

+ e ) Ru

2+

E°’ ≈ -0.9 V E°’ ) 1.1 V

Thus, oxidation of the electronically excited state at a negative potential results in the ground-state oxidized product at a potential that is significantly negative of its formal potential, and there is a driving force on the order of 2 eV for back electron transfer. However, as illustrated in Figure 4 the maximum ground-state heterogeneous electron-transfer rate of approximately 2.6 × 106 s-1 is achieved for an overpotential greater than 0.3 eV, and any additional driving force will not produce a back-electron-transfer rate greater than about 106 s-1. It is also important to note that where the redox process is close in energy to a bridge-based redox process, significantly higher electron-transfer rates can be observed.3,43 In this case, the excited-state oxidation process is within a few tenths of an electronvolt of the bridging-ligand reduction whereas the ground-state Ru2+/3+ process is approximately 2 eV away. Therefore, it is possible that the rate of process 1 may be very much larger than that of process 2, making it possible to directly observe the redox process associated with the oxidation of the electronically excited state. Figure 9 shows that on the first scan an oxidative current response is observed at approximately -0.8 V. The peak potential of this process is in approximate agreement with the calculated excited-state redox potential of -0.9 V. That this current response is not observed for laser excitation in the absence of the monolayer (thin solid line) or for the second or subsequent scans (open circles) confirms the transient nature of the phenomenon. Whereas (43) de Rege P. J. F.; Williams, S. A.; Therien, M. J. Science 1995, 269, 1409.

Photonics and Electrochemistry of Adsorbed Luminophores

fast-response microelectrodes have been used to perform these measurements, ohmic drop and, in particular, heterogeneous electron transfer that is slow compared to the experimental time scale are likely to cause the peak potential to deviate somewhat from its true thermodynamic value. Notwithstanding these concerns, the agreement between the experimental peak potential and that predicted by the Rehm-Weller equation is approximately 100 mV, which is entirely satisfactory. Significantly, the charge underneath this response is only about 40% of that observed at slower scan rates for the ground-state Ru2+/3+ process. This lower apparent surface coverage is consistent with the loss of some of the excited states by radiative or nonradiative relaxation during the time taken to scan from the initial potential to a value where the oxidation of Ru2+* can occur. The significant charge passed during Ru2+* oxidation suggests that close to 100% of the adsorbed complexes were excited during the laser pulse.

Conclusions Ruthenium diphenylphenanthroline complexes that include a quarterpyridyl ligand within their coordination shell are highly surface-active and yield highly stable monolayers. Three distinct oxidation states are electrochemically accessible, and the electrochemical responses are nearly ideal as the potential, temperature, and experimental time scale are varied over wide ranges. Chronoamperometry has been used to probe the rate of heterogeneous electron transfer across the monolayer/microelectrode interface. This process can be characterized by a single rate constant at high electrolyte concentrations, suggesting that heterogeneous electron transfer across these metal/monolayer interfaces is mechanistically uncomplicated.

Langmuir, Vol. 22, No. 25, 2006 10761

Significantly, upon laser excitation at 355 nm the monolayers fluoresce (i.e., despite having the excited state located close to a metal surface, energy transfer does not completely quench the excited state). We have explored the extent of lateral electron transfer between excited-state Ru2+* and Ru3+ within the films by electrochemically titrating in fixed quantities of Ru3+. For close-packed monolayers, Ru3+ is a highly efficient excitedstate electron acceptor with rates between 8 × 106 and 3 × 108 M-1 s-1, rivaling rates found in solution. We have also attempted to probe the energetics and dynamics of excited-state oxidation using voltammetry conducted at scan rates in the megavolts per second range. The voltammetric data reported are qualitatively consistent with the response anticipated for an electronically excited state in terms of the behavior of bare electrodes as well as the peak potential and voltammetric time scales over which the transient response is observed for monolayer-modified electrodes. Specifically, although influenced by the dynamics of heterogeneous electron transfer, the measured excited-state redox potential agrees with that predicted by the Rehm-Weller equation to within 100 mV. The observation of emission, despite the luminophores sitting only 15 Å or so above a metal surface, is consistent with investigations of the heterogeneous electron-transfer process that indicate weak adsorbate-electrode interactions. Acknowledgment. We thank the Science Foundation Ireland for support under the Biomedical Diagnostics Institute (award no. 05/CE3/B754). The generous loan of ruthenium trichloride and microwires by Johnson-Matthey is gratefully acknowledged. LA0611918