Electrochemistry of Surface-Grafted Stimulus-Responsive Monolayers

Lionel Dos Ramos , Sissi de Beer , Mark A. Hempenius , and G. Julius Vancso. Langmuir .... Synthesis and Reversible Redox Properties of an Electron-Ri...
1 downloads 0 Views 181KB Size
Langmuir 2005, 21, 5115-5123

5115

Electrochemistry of Surface-Grafted Stimulus-Responsive Monolayers of Poly(ferrocenyldimethylsilane) on Gold Ma´ria Pe´ter, Rob G. H. Lammertink, Mark A. Hempenius,* and G. Julius Vancso* University of Twente, MESA+ Institute for Nanotechnology, 7500 AE Enschede, The Netherlands Received October 29, 2004. In Final Form: March 5, 2005

Poly(ferrocenyldimethylsilane)s with various degrees of polymerization and featuring a thiol end group were chemically end-grafted onto gold substrates by self-assembly, forming redox-active monolayers. The monolayers were characterized by contact angle measurements, Fourier transform infrared spectroscopy, and X-ray photoelectron spectroscopy. Layer thickness values were determined by surface plasmon resonance spectroscopy and ellipsometry. The electrochemical properties of these films in aqueous NaClO4 were studied using cyclic voltammetry (CV), differential pulse voltammetry, chronoamperometry, and chronocoulometry. Cyclic voltammograms showed two reversible redox peaks, indicating a stepwise oxidation of the electroactive sites. The first oxidation step showed reversible behavior at low scan rates and quasireversible behavior at higher scan rates. Peak currents (ip) plotted against the square root of scan rates (v1/2) for the first oxidation peak and for the corresponding reduction peak exhibited a linear dependence, indicating that the oxidation process in the first step is controlled by the diffusion of counterions into the polymer film. For the second oxidation peak and the corresponding reduction peak, ip varied linearly with v. This redox behavior is characteristic of surface-immobilized electroactive layers. The higher reversibility of the second oxidation and reduction waves in the CV experiments was explained from the solvation of the surface-grafted poly(ferrocenylsilane) (PFS) chains, which depends on the degree of oxidation. Oxidized PFS films are swollen in the aqueous electrolyte solutions, leading to a higher segmental mobility of the polymer chains and a much increased counterion mobility within the film. Kinetic parameters for the redox processes were obtained from chronocoulometry experiments.

Introduction The modification of electrode surfaces with electroactive species is of growing interest due to potential applications in the areas of ion recognition, biosensors, enzyme electrodes, and molecular electronics.1-5 Various approaches exist for introducing redox-active molecules onto metallic electrodes. Thiol- or disulfide-containing functional redox-active molecules can be adsorbed onto gold surfaces to form self-assembled monolayers (SAMs)1 or be embedded in preformed SAMs.4 Layer-by-layer deposition of oppositely charged polyions, incorporating redoxactive groups, allows one to construct electroactive multilayer thin films in a stepwise process. Such films have been used, for example, in redox-mediated enzyme catalysis.6 When redox-active polymers, rather than small molecules, are deposited on conductive surfaces, one has the opportunity to electrochemically control the charge density in the polymer chain. For conjugated polymers such as polypyrrole and polyaniline, significant volume changes have been observed upon electrochemical oxida* To whom correspondence should be addressed. E-mail: [email protected].

(1) Trippe´, G.; Oc¸ afrain, M.; Besbes, M.; Monroche, V.; Lyskawa, J.; Le Derf, F.; Salle´, M.; Becher, J.; Colonna, B.; Echegoyen, L. New J. Chem. 2002, 26, 1320. (2) Willner, I.; Katz, E. Angew. Chem., Int. Ed. 2000, 39, 1180. (3) Smalley, J. F.; Finklea, 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. (4) Hortholary, C.; Minc, F.; Coudret, C.; Bonvoisin, J.; Launay, J.-P. Chem. Commun. 2002, 1932. (5) Yeo, W.-S.; Yousaf, M. N.; Mrksich, M. J. Am. Chem. Soc. 2003, 125, 14994. (6) Calvo, E. J.; Etchenique, R.; Pietrasanta, L.; Wolosiuk, A.; Danilowicz, C. Anal. Chem. 2001, 73, 1161.

tion and reduction. Such polymers may be useful as microactuators and artificial muscles.7-9 Macromolecules containing inorganic elements or organometallic units in the main chain combine potentially useful chemical, electrochemical, optical, and other interesting characteristics with the properties and processability of polymers.10 Poly(ferrocenylsilane)s (PFSs), composed of alternating ferrocene and dialkylsilane units in the main chain, are redox-active materials with interacting Fe centers.11 The attachment of a thiol end-function allows one to immobilize PFSs on gold surfaces by chemisorption,12 using similar strategies as applied to fabricate SAMs of thiols.13 The resulting stable, ultrathin films are highly suitable for electrochemical studies. The morphology of the surface-grafted films depends on the choice of solvent for deposition and solvent employed in electrochemistry and the solubility of the macromolecules.14,15 Although the redox behavior of PFSs has been documented,16-19 no detailed studies on the electrochemistry (7) Smela, E.; Gadegaard, N. Adv. Mater. 1999, 11, 953. (8) Smela, E. J. Micromech. Microeng. 1999, 9, 1. (9) Bay, L.; Jacobsen, T.; Skaarup, S.; West, K. J. Phys. Chem. B 2001, 105, 8492. (10) Abd-El-Aziz, A. S. Macromol. Rapid Commun. 2002, 23, 995. (11) For reviews on PFSs see: (a) Manners, I. Macromol. Symp. 2003, 196, 57. (b) Kulbaba, K.; Manners, I. Macromol. Rapid Commun. 2001, 22, 711. (c) Manners, I. Chem. Commun. 1999, 857. (12) Nuzzo, R. G.; Allara, D. L. J. Am. Chem. Soc. 1983, 105, 4481. (13) Ulman, A. An Introduction to Ultrathin Organic Films: From Langmuir-Blodgett to Self-Assembly; Academic Press: San Diego, 1991. (14) Stamouli, A.; Pelletier, E.; Koutsos, V.; van der Vegte, E.; Hadziioannou, G. Langmuir 1996, 12, 3221. (15) Koutsos, V.; van der Vegte, E. W.; Hadziioannou, G. Macromolecules 1999, 32, 1233. (16) Foucher, D. A.; Honeyman, C. H.; Nelson, J. M.; Tang, B. Z.; Manners, I. Angew. Chem., Int. Ed. Engl. 1993, 32, 1709.

10.1021/la0473409 CCC: $30.25 © 2005 American Chemical Society Published on Web 04/28/2005

5116

Langmuir, Vol. 21, No. 11, 2005

of surface-grafted PFS monolayers have appeared in the literature.20,21 In a previous paper, we reported on the redox-induced morphology and volume changes of PFS monolayers on gold by means of in situ electrochemical atomic force microscopy.22 In another study, thiol endfunctionalized PFSs were grafted into the defects of preformed SAMs for force spectroscopy experiments on the single molecule level.23 In the present contribution, the electrochemical behavior of PFS monolayers grafted on gold and the kinetic parameters associated with the stepwise oxidation and reduction processes of these films are described. Experimental Section Materials. Ferrocene (98%), N,N,N′,N′-tetramethylethylenediamine (99.5+%), n-butyllithium (1.6 M in hexanes), dichlorodimethylsilane (99%), toluene (HPLC grade), and tetrabutylammonium hexafluorophosphate (98%) were obtained from Aldrich. Dichloromethane and ethanol (for analysis) were obtained from Merck. Ethylene sulfide (98%) and trimethylene sulfide (97%) were purchased from Aldrich and distilled as hexane solutions from CaH2. n-Heptane (for synthesis) was obtained from Merck and distilled from CaH2 prior to use. Tetrahydrofuran (THF) for anionic polymerizations was distilled from sodiumbenzophenone under argon, degassed in three freeze-pumpthaw cycles, and distilled by vacuum condensation from nbutyllithium. Monomer Purification. Dimethylsila[1]ferrocenophane was prepared and purified as described earlier.24,25 Anionic Polymerization. Dimethylsila[1]ferrocenophane polymerizations were carried out in THF in a MBraun 150-GI glovebox under an atmosphere of prepurified nitrogen, using n-butyllithium as the initiator. After 15 min, the living polymers were end-functionalized by adding 3 equiv of ethylene sulfide or trimethylene sulfide. The thiolate end groups were protonated by adding a few drops of degassed acetic acid. Thus, thiol endfunctionalized poly(ferrocenyldimethylsilane)s, denoted as ESPFS and TMS-PFS, were obtained. The polymers were precipitated twice in methanol and dried in a vacuum. Polymers with varying degrees of polymerization (DPn) were synthesized. Their notation corresponds to the monomer/initiator ratios used: ES-PFS25 (Mn ) 7750 g/mol, DPn ) 32, Mw/Mn ) 1.36), ES-PFS50 (Mn ) 12 500 g/mol, DPn ) 50, Mw/Mn ) 1.14), ESPFS100 (Mn ) 22 600 g/mol, DPn ) 92, Mw/Mn ) 1.13), and TMSPFS50 (Mn ) 11 300 g/mol, DPn ) 46, Mw/Mn ) 1.14). Substrates and Samples. Gold substrates (borosilicate glass, 2 nm Cr, 250 nm Au) for atomic force microscopy (AFM) were purchased from Metallhandel Schro¨er (Lienen, Germany). Au(111) samples were prepared by annealing the gold substrates in a high purity hydrogen flame for 8 min. Round gold (diameter 2.5 cm) substrates for electrochemistry and ellipsometry were prepared by the evaporation of 5 nm of Cr followed by 200 nm of Au on glass substrates. Prior to use, these substrates were cleaned in piranha solution (1:4 mixture of 30% H2O2 and (17) Nguyen, M. T.; Diaz, A. F.; Dement’ev, V. V.; Pannell, K. H. Chem. Mater. 1993, 5, 1389. (18) Foucher, D.; Ziembinski, R.; Petersen, R.; Pudelski, J.; Edwards, M.; Ni, Y.; Massey, J.; Jaeger, C. R.; Vancso, G. J.; Manners, I. Macromolecules 1994, 27, 3992. (19) Rulkens, R.; Lough, A. J.; Manners, I.; Lovelace, S. R.; Grant, C.; Geiger, W. E. J. Am. Chem. Soc. 1996, 118, 12683. (20) For a preliminary study see: Pe´ter, M.; Hempenius, M. A.; Lammertink, R. G. H.; Vancso, G. J. Macromol. Symp. 2001, 167, 285. (21) Pe´ter, M.; Lammertink, R. G. H.; Hempenius, M. A.; van Os, M.; Beulen, M. W. J.; Reinhoudt, D. N.; Knoll, W.; Vancso, G. J. Chem. Commun. 1999, 359. (22) Pe´ter, M.; Hempenius, M. A.; Kooij, E. S.; Jenkins, T. A.; Roser, S. J.; Knoll, W.; Vancso, G. J. Langmuir 2004, 20, 891. (23) Zou, S.; Ma, Y. J.; Hempenius, M. A.; Scho¨nherr, H.; Vancso, G. J. Langmuir 2004, 20, 6278. (24) Ni, Y.; Rulkens, R.; Manners, I. J. Am. Chem. Soc. 1996, 118, 4102. (25) Lammertink, R. G. H.; Hempenius, M. A.; Thomas, E. L.; Vancso, G. J. J. Polym. Sci., Part B: Polym. Phys. 1999, 37, 1009.

Pe´ ter et al. concentrated H2SO4), then rinsed with Milli-Q water and ethanol, and dried in a nitrogen stream. Caution! Piranha solution is a very strong oxidant, reacts violently with organic materials, and should be handled with the utmost care! Thin films of endfunctionalized PFS were prepared by immersing gold substrates into 1.0 mg mL-1 solutions of the polymers in toluene for at least 24 h to ensure maximum coverage. The samples were subsequently rinsed with toluene and dried in a stream of nitrogen. Techniques. 1H NMR spectra were recorded in CDCl3 on a Varian Unity Inova (300 MHz) instrument at 300.3 MHz. A solvent chemical shift of δ ) 7.26 ppm was used as a reference. Gel permeation chromatography measurements were carried out in THF at 25 °C, using microstyragel columns (bead size 10 µm) with pore sizes of 103, 104, 105, and 106 Å (Waters) and employing a dual detection system consisting of a differential refractometer (Waters model 410) and a differential viscometer (Viscotek model H502). Molar masses were determined relative to polystyrene standards. Contact angles were measured with a Kru¨ss G10 contact angle measuring instrument equipped with a chargecoupled device camera. Advancing and receding contact angles were determined automatically by a drop shape analysis routine during the growth and shrinkage of the droplet. Contact angles were measured using ultrasonically degassed Milli-Q water. The polymeric layers on gold were characterized by reflectionabsorption grazing-angle incidence Fourier transform infrared (FTIR) spectroscopy (angle of incidence was 83°) using a mercury cadmium telluride detector cooled with liquid nitrogen. Background spectra were recorded for C16D33SH SAMs on gold. X-ray photoelectron spectroscopy (XPS) spectra were obtained on a Quick Scan Kratos XSAM800 (Manchester, U.K.) instrument using a Mg KR radiation source (1253.6 eV). The analyzer slit width was set to 6 mm, and the input power was 150 W (15 kW/10 mA). Survey scans in a binding energy window of 1100-0 eV were recorded with a pass energy of 100 eV. Spectra were obtained for electron takeoff angles of 0° (perpendicular) and 60° (to the normal). Peak positions were referenced to the binding energy of C(1s) electrons at 284.5 eV.26 Atomic concentrations were determined by the numerical integration of the relative peak areas in the detailed element scans using sensitivity factors given in the literature.27 AFM measurements were performed with a NanoScope III atomic force microscope (Digital Instruments, Santa Barbara, U.S.A.) in the tapping mode. The images were taken in air using rectangular Si cantilevers with a spring constant of 20-54 N m-1. The cantilevers used were triangular Si3N4, with a spring constant of 0.32 N m-1. Surface plasmon resonance spectroscopy (SPRS) measurements were performed against 0.1 M NaClO4 in the Kretschmann configuration.28 A refractive index of n ) 1.687, previously determined by ellipsometry, was used for all layers.29 For the optical characterization of the films, a custom-built spectroscopic ellipsometer (rotating polarizer-analyzer geometry) equipped with a xenon lamp and a scanning monochromator was used. Measurements were performed in the visible region of the spectra at wavelengths between λ ) 300 and λ ) 800 nm at a fixed incident angle of 70°. Electrochemical experiments [cyclic voltammetry (CV), differential pulse voltammetry (DPV), chronoamperometry, and chronocoulometry) were undertaken using an Autolab PGSTAT10 potentiostat (ECOCHEMIE, Utrecht, The Netherlands) in a three-electrode configuration. Polymer-covered gold disk working, Hg/HgSO4 (MSE) reference (+0.61 VNHE), and Pt auxiliary electrodes were used. The electrolyte was NaClO4 in water (0.011.0 M). The working electrode exposed a surface area of 0.44 cm2 to the electrolyte. The cell was degassed by passing nitrogen through the electrolyte. A constant nitrogen flow over the electrolyte was maintained during the measurements. The cyclic voltammograms were recorded between -0.4 VMSE and +0.4 VMSE at different scan rates. Differential pulse voltammograms were (26) Barber, M.; Connor, J. A.; Derrick, L. M. R.; Hall, M. B.; Hillier, I. H. J. Chem. Soc. 1973, 560. (27) Practical Surface Analysis: By Auger and X-Ray Photoelectron Spectroscopy; Briggs, D., Seah, M. P., Eds.; John Wiley & Sons: Chichester, 1983. (28) Kretschmann, E. Z. Phys. 1971, 241, 313. (29) The refractive index was determined by Vanessa Z.-H. Chan and Rob G. H. Lammertink at the Massachusetts Institute of Technology, Cambridge, MA, U.S.A.

Surface-Grafted Poly(ferrocenylsilane) Monolayers

Langmuir, Vol. 21, No. 11, 2005 5117

Chart 1

also recorded between -0.4 and +0.4 VMSE, at a scan rate of v ) 5 mV s-1. A pulse amplitude of 10 mV was used. The pulse time was 50 ms, applied at a time interval of 200 ms. In the chronoamperometric and chronocoulometric experiments a double potential step was employed. The potential was stepped from +0.4 V to -0.4 V at t ) 0. At t ) τ (τ ) 0.2 s) the potential was stepped back to 0.4 V. The changes in current and charge were recorded as a function of time.

Results and Discussion Characterization of Poly(ferrocenyldimethylsilane) Monolayers. Thiol end-functionalized PFSs, obtained by end-capping living PFS chains with ethylene sulfide (ES-PFS) or with trimethylene sulfide (TMSPFS) were prepared with degrees of polymerization DPn ) 25, 50, and 100 (DPn ) 50 in the case of TMS-PFS) as described earlier (Chart 1).20,21 Surface-grafted PFS monolayers were obtained by chemisorption of these polymers onto gold substrates. The monolayer films were characterized by FTIR spectroscopy, contact angle measurements, and XPS. FTIR was used to verify the presence of the PFS monolayers on gold. In Figure 1, the overlaid FTIR spectra for an ES-PFS50 film on gold and ES-PFS50 in KBr are shown. In the high energy region (Figure 1a), C-H stretch peaks belonging to the ferrocene rings (3087 cm-1) and the methyl groups at silicon (2958 cm-1) are present. In the low energy region (Figure 1b) the most prominent peaks are associated with the symmetric deformation vibration of the CH3 groups in the (CH3)2Si units (1248 cm-1) and with bands diagnostic for the ferrocene units (1165 and 1037 cm-1). The out-of-plane C-H bending signals for the ferrocene units occur around 800 cm-1.30,31 XPS provides information on composition and chemical bonding at the surfaces of solid materials.27 Highresolution scans for the C(1s), Fe(2p3/2), Si(2p3/2), and S(2p3/2) regions of the XPS spectrum were recorded. The S(2p3/2) binding energies in alkanethiols, disulfides, and dialkylsulfides are typically between 163 and 164 eV.32 Adsorption of these species onto gold causes their S(2p3/2) signals to shift by about -1.2 eV.27 For the surface-grafted PFS layers, the S(2p3/2) binding energies of bound sulfur were 161.9 eV. Small shoulders at higher binding energies (163.1 eV) were due to unbound sulfide units. Ferrocene compounds exhibit binding energy values for Fe(2p2/3) (30) Socrates, G. Infrared Characteristic Group Frequencies: Tables and Charts, 2nd ed.; John Wiley & Sons, Ltd.: Chichester, 1994. (31) Neuse, E. W.; Bednarik, L. Macromolecules 1979, 12, 187. (32) McNeillie, A.; Brown, D. H.; Smith, W. E.; Gibson, M.; Watson, L. J. Chem. Soc., Dalton Trans. 1980, 767.

Figure 1. FTIR spectra for ES-PFS50 in the bulk (in KBr, solid line) overlaid with the FTIR spectra of an ES-PFS50 thin film on gold (dashed line); (a) high energy region; (b) low energy region. Table 1. XPS Analysis of the Surface-Grafted PFS Monolayers on Golda ES-PFS25 element C(1s) Si(2p) Fe(2p) S(2p) a



60°

ES-PFS50 0°

60°

ES-PFS100 0°

60°

TMS-PFS50 0°

60°

82.26 87.26 83.68 84.96 85.57 86.99 80.32 84.84 8.03 5.60 7.07 6.48 6.60 6.16 8.70 6.98 7.81 6.04 6.92 7.10 6.76 5.98 8.39 6.62 1.88 1.09 2.31 1.44 1.05 0.86 2.57 1.54

Takeoff angles 0° and 60°.

electrons between 707 and 710.5 eV.26,33 For the polymer layers, an Fe(2p3/2) binding energy of 708.2 eV was found. Si(2p3/2) signals were observed at about 101 eV.34 Atomic concentrations, determined by integration of the relative peak areas, are shown in Table 1. The ratio Si(2p)/Fe(2p) almost equals unity, which agrees well with the theoretical composition. The atomic concentration of S(2p) decreases with increasing takeoff angle, indicating that the sulfur is located close to the substrate surface. Tapping mode height AFM images recorded for the ES-PFS layers exhibit a globular surface structure, superimposed on the atomically smooth triangular terraces of the bare Au(111) substrate.22 Contact angles for water were measured to determine the wetting properties of the films.13 Advancing contact angles of 96 ( 2° and receding contact angles of 75 ( 2° were measured for all layers, showing the surfaces to be hydrophobic. The large hysteresis between the advancing and receding contact angles indicates that the layers are disordered. The surface coverages of ferrocene sites (ΓFc) for all layers were determined according to eq 1, where QFc is the charge passed for the oxidation/reduction (33) Fischer, A. B.; Wringhton, M. S.; Umana, M.; Murray, R. W. J. Am. Chem. Soc. 1979, 101, 3442. (34) Schmidt, R. L.; Cardella, J. A.; Magill, J. H.; Chin, R. L. Polymer 1987, 28, 1462.

5118

Langmuir, Vol. 21, No. 11, 2005

Pe´ ter et al.

Table 2. Coverage and Thickness Data for the End-Grafted PFS Layers on Gold

ES-PFS25 ES-PFS50 ES-PFS100 TMS-PFS50

ΓFc × 109a (mol cm-2)

Γb (chains/nm2)

Lc (nm)

Ld (nm)

2.13 2.32 2.52 1.08

0.40 0.28 0.16 0.14

4.3 ( 0.2 4.7 ( 0.2 5.2 ( 0.2

4.9 ( 0.2 6.4 ( 0.2 7.7 ( 0.2 1.6 ( 0.2

a Surface coverage by ferrocene units, obtained from CV. b Number of grafted chains per unit area, calculated from ΓFc. c Thickness measured by ellipsometry against air. d Thickness measured by SPRS against 0.1 M NaClO4.

of ferrocene sites,

ΓFc )

QFc neFA

(1)

ne is the number of electrons involved in the electrontransfer process (here ne ) 1), F is the Faraday constant (F ) 96 485 C‚mol-1), and A is the geometric surface area of the electrode (A ) 0.44 cm2). The charge passed in the redox reaction was determined by integrating the areas under the redox peaks.35 The number of grafted chains per unit area (Γ) was then calculated by dividing ΓFc by the degree of polymerization. Monolayer thicknesses were measured by SPRS against aqueous 0.1 M NaClO4. Layer thicknesses were furthermore determined against air by ellipsometry (Table 2). CV. The electrochemical properties of ES-PFS and TMS-PFS layers on gold were investigated by CV, DPV, chronoamperometry, and chronocoulometry in aqueous NaClO4. Cyclic voltammograms recorded for the electroactive polymer layers show two reversible redox peaks, associated with the stepwise, one-electron oxidation of the Fe atoms in the ferrocene units.17,19 The double wave voltammogram indicates that repulsive interactions exist between the electroactive centers in the layer.36 A ferrocenium group increases the oxidation potential of neighboring unoxidized ferrocene centers and was even found to weakly affect the oxidation potential of the next nearest ferrocene centers.19 Thus, the first oxidation wave is attributed to oxidation of ferrocene centers at alternating positions along the chain. In the second wave, at higher potentials, oxidation of the remaining ferrocene centers, in positions next to oxidized units, is completed. Figure 2 shows CVs recorded on an ES-PFS100 layer, using different scan rates, and includes peak notations. A sensitive criterion for the reversibility of electron transfer is the value of the separation between the peak potentials ∆Ep ) Epox - Epred. A truly reversible oneelectron transfer will exhibit a ∆Ep of 59 mV.35,37 Slow electron-transfer kinetics cause the peak separation to increase. In the case of the surface-grafted PFS monolayers, the separation between the first anodic (Epox1) and cathodic (Epred1) peak potentials ∆Ep1 ) Epox1 - Epred1 increases at higher scan rates. The peak separations between anodic and cathodic peak potentials (∆Ep1 and ∆Ep2), as well as the differences between the peaks in the oxidation wave (∆Epox ) Epox2 - Epox1) at various scan rates, (35) (a) Murray, R. W. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1984; Vol. 13, p 191. (b) Finkley, H. O. In Electroanalytical Chemistry; Bard, A. J., Rubinstein, I., Eds.; Marcel Dekker, Inc.: New York, 1996; Vol. 19, p 109. (36) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.; John Wiley & Sons: New York, 2001. (37) Kissinger, P. T.; Heineman, W. R. J. Chem. Educ. 1983, 60, 702.

Figure 2. Cyclic voltammograms for an ES-PFS100 layer on gold at scan rates v ) 10, 20, 30, and 50 mV/s, in aqueous 0.1 M NaClO4. Reference electrode Hg/HgSO4, counter electrode Pt.

are given in Tables 3 and 4. The ∆Ep1 values show that the systems are electrochemically reversible up to a scan rate of v ) 30 mV s-1, if the first anodic peak and the first wave in the cathodic scan are considered. At these sweep rates, the peak potential difference ∆Ep1 remains below 59 mV. At v ) 50 mV s-1 and higher the systems become electrochemically quasireversible. For the second redox peaks, however, as is evident from the ∆Ep2 values (Tables 3 and 4), reversibility is maintained up to scan rates of v ) 75 mV s-1 or higher. Peak currents plotted against scan rates and square roots of scan rates for the ES-PFS100 layer on gold are shown in parts a and b of Figure 3, respectively. Interestingly, peak currents for the first redox peaks (ipox1 and ipred1) do not show a linear dependence on the scan rate (Figure 3a). This indicates that the magnitude of currents generated in the first redox peaks is less than predicted by the following equation characteristic of surface-confined electroactive layers:35

ip )

n2F2AΓv nFAQv ) 4RT 4RT

(2)

where n is the number of moles of electrons transferred, F is the Faraday constant (F ) 96 485 C mol-1), A is the electrode surface area (cm2), Γ is the surface coverage (mol cm-2), R is the molar gas constant (R ) 8.314 J mol-1 K-1), T is the temperature (in Kelvin), Q is the charge involved in the electrochemical process (C), and v is the scan rate (V s-1). Clearly, charge transfer occurs on a time scale that is long compared to the experimental time scale. If the peak currents ipox1 and ipred1 are plotted against the square root of the scan rates, a linear behavior is observed. In Figure 3b the continuous lines represent linear fits to ipox1 and ipred1, and the dashed lines represent linear fits to the same data through the origin. The linearity of the peak currents with the square root of the scan rates indicates that charge transfer in the first redox steps is controlled by the diffusion of charges in the films as described by the empirical Randles-Sevcik equation35

ip ) (2.69 × 105)n3/2ADct1/2v1/2C0

(3)

where Dct is the charge transport diffusion coefficient and C0 is the concentration of electroactive sites. The dependence of the peak current values on the scan rate is determined by the parameter Dctτ/d2, where τ is the experimental time scale (the time needed for the potential scan to traverse the wave) and d is the polymer film

Surface-Grafted Poly(ferrocenylsilane) Monolayers

Langmuir, Vol. 21, No. 11, 2005 5119

Table 3. Dependence of the Peak Separation between Anodic and Cathodic Peak Potentials (∆Ep1 and ∆Ep2), Respectively, and of the Peak Splitting in the Anodic Wave (∆Epox) on the Scan Rate v Employed for ES-PFS25 and ES-PFS50 Layers ∆Ep1 ) Epox1 - Epred1a (mV)

a

∆Ep2 ) Epox2 - Epred2a (mV)

∆Epox ) Epox2 - Epox1a (mV)

v (mV s-1)

ES-PFS25

ES-PFS50

ES-PFS25

ES-PFS50

ES-PFS25

ES-PFS50

10 20 30 50 75 100

35 47 56 73 90 102

33 46 51 69 83 98

6 6 19 35 58 78

10 18 19 27 34 44

84 77 81 84 89 95

92 93 90 82 78 69

For peak notations see Figure 2.

Table 4. Dependence of the Peak Separation between Anodic and Cathodic Peak Potentials (∆Ep1 and ∆Ep2), Respectively, and of the Peak Splitting in the Anodic Wave (∆Epox) on the Scan Rate v Employed for ES-PFS100 and TMS-PFS50 Layers ∆Ep1 ) Epox1 - Epred1a (mV)

∆Ep2 ) Epox2 - Epred2a (mV)

∆Epox ) Epox2 - Epox1a (mV)

v (mV s-1)

ES-PFS100

TMS-PFS50

ES-PFS100

TMS-PFS50

ES-PFS100

TMS-PFS50

10 20 30 50 75 100

37 49 58 77 93 110

28 37 44 57 71 84

13 19 23 33 58 69

4 7 14 23 34 40

91 92 91 84 83 78

88 87 88 87 84 80

a

For peak notations see Figure 2.

Figure 3. (a) Dependence of peak currents on scan rates for an ES-PFS100 layer on gold. (b) Dependence of peak currents on the square root of scan rates for the same ES-PFS100 layer. Dashed lines represent a fit through the origin; continuous lines are linear fits through the data points.

thickness. When Dctτ/d2 . 1, the rate of charge transport is significantly higher compared to the experimental time scale; therefore, ip is proportional to v. When Dctτ/d2 , 1, Dct for charge transport is very small. In this case ip is

proportional to v1/2. For intermediate values of Dctτ/d2, ip is proportional to vn with 1/2 < n < 1.35 Dctτ/d2 values, determined by means of chronocoulometry measurements, are given in Table 6. Peak current values for the second redox waves (ipox2 and ipred2) depend linearly on the scan rates (Figure 3a). Charge transport occurs significantly faster with respect to the experimental time scale; thus, the ratio of oxidized to reduced sites in the layer is in thermodynamic equilibrium with the electrode potential. The magnitude of peak currents for the second peak obeys eq 2, characteristic for surface-confined electroactive layers.35 Influence of the Electrolyte. Charge diffusion during oxidation of a PFS monolayer can be assumed to involve electron-exchange reactions between neighboring ferrocene and ferrocenium sites by electron hopping.38 The diffusion of localized ferrocene and ferrocenium states is accompanied by counterion diffusion to maintain charge neutrality. The rate of charge transport then depends on the mobility of the ferrocene and ferrocenium sites in the monolayer and on the mobility of the supporting electrolyte ions within the film. Segmental mobility of the PFS chains is strongly influenced by their solvation, that is, by their compatibility with the electrolyte solution. The degree of solvation of the monolayer also influences counterion mobility within the film. The electrochemical behavior of the poly(ferrocenyldimethylsilane) monolayers as a function of NaClO4 electrolyte concentration (Cs, 0.01-1.00 M) was investigated by CV. In Figure 4, cyclic voltammograms recorded for ES-PFS25 layers at a scan rate of v ) 10 mV s-1 are shown. The oxidation peak potentials shifted to lower values as the electrolyte concentrations increased. Figure 5 shows a linear dependence of the peak potential on the logarithm of electrolyte concentration which can be described by the Nernst equation39 where E0′ is the formal potential and K is the formation constant (38) Daum, P.; Lenhard, J. R.; Rolison, D.; Murray, R. W. J. Am. Chem. Soc. 1980, 102, 4649. (39) Inzelt, G.; Szabo, L. Electrochim. Acta 1986, 31, 1381.

5120

Langmuir, Vol. 21, No. 11, 2005

Pe´ ter et al.

Table 5. Heterogeneous Rate Transfer Coefficients, Charge Transport Diffusion Coefficients, and Activation Energies, Based on Chronocoulometric Experimentsa

ES-PFS25 ES-PFS50 ES-PFS100 TMS-PFS50 a

kf × 105 (cm s-1)

Dct21/2C × 109 (mol cm-2s-1/2)

Ea2 (cal mol-1)

Dct11/2C × 109 (mol cm-2 s-1/2)

Ea1 (cal mol-1)

1.87 ( 0.02 2.23 ( 0.02 3.07 ( 0.03 1.19 ( 0.01

8.44 ( 0.08 8.73 ( 0.08 10.03 ( 0.10 5.49 ( 0.05

4670 ( 90 4410 ( 90 4100 ( 80 5680 ( 110

1.33 ( 0.01 1.19 ( 0.01 0.98 ( 0.01 0.570 ( 0.005

6850 ( 135 6780 ( 130 6880 ( 137 8360 ( 160

Dct1 corresponds to the Epred1, and Dct2 corresponds to the Epred2 wave in the CV in Figure 2. Table 6. Ferrocene Coverages Determined from CV, Charge Diffusion Coefficients, and Parametersa

ES-PFS25 ES-PFS50 ES-PFS100 TMS-PFS50 a

ΓFc × 109 (mol cm-2)

Dct2 × 1012 (cm2 s-1)

Dct2τ/d2

Dct1 × 1014 (cm2 s-1)

Dct1τ/d2

2.13 2.32 2.52 1.08

3.85 ( 0.03 5.88 ( 0.06 9.96 ( 0.10 0.69 ( 0.07

3.15 ( 0.04 2.84 ( 0.04 3.31 ( 0.05 5.20 ( 0.07

9.66 ( 0.13 10.84 ( 0.15 9.05 ( 0.12 0.750 ( 0.001

0.08 ( 0.001 0.05 ( 0.001 0.03 ( 0.006 0.06 ( 0.001

Dct1 corresponds to the Epred1, and Dct2 corresponds to the Epred2 wave in the CV in Figure 2.

Figure 4. Cyclic voltammograms recorded for ES-PFS25 layers on gold at v ) 10 mV s-1, in aqueous 0.01-1.0 M NaClO4.

Figure 5. Peak potentials and logarithm of peak currents plotted against the negative logarithm of the electrolyte concentration.

of the ferrocenium salt.

E ) E0′ -

+ RT [(Fc )(ClO4 )] RT ln K + ln F F [Fc] RT ln [ClO4-] (4) F

Figure 5 also shows the dependence of peak currents on the supporting electrolyte concentration, Cs. Peak currents in the first oxidation wave increase with Cs, reaching a maximum at approximately 0.1 M, and then decrease again. The increase in peak currents with increasing concentrations of supporting electrolyte shows that charge transfer becomes more likely, showing that counterion diffusion into the polymer monolayer influences the charge transport rate. When electrolyte concentrations are further

increased, the peak currents decrease. Under these conditions, the charged polymer film is less swollen due to the low water activity, hindering counterion motion in the film and decreasing the number of accessible ferrocene units. The two opposite effects result in a maximum charge transport rate at intermediate electrolyte concentrations.40 At very high electrolyte concentrations the electrode can become passivated, likely due to a collapse of the polyelectrolyte chains. The electrochemical activity is regained when the PFS monolayer is immersed into dilute electrolyte. The solvation of the PFS films depends on the degree of oxidation. In the first oxidation step, the PFS chains are turned from a neutral state into a partially charged state. Especially in the beginning of this step, the monolayer films are poorly swollen in the aqueous NaClO4 solution and supporting electrolyte ions have a limited access to the redox sites within the film. Thus, in the first oxidation step, the electrochemical response of the film is controlled by the diffusion of counterions into the film and follows the behavior described by eq 3. At the start of the second oxidation step, the polymer chains are already partly charged and, therefore, more solvated by the aqueous electrolyte solution. In the second oxidation step, therefore, electrolyte ions diffuse more readily into the monolayer film than in the first step. The charge transport rate is now significantly higher with respect to the employed scan rates, and the magnitude of the peak currents for the second peak obeys eq 2, characteristic for surface-confined electroactive layers.35 Cyclic voltammograms were also recorded for an ES-PFS50 layer on gold in 0.1 M supporting electrolytes that have the same anion (ClO4-) but different cations (Li+, Na+, K+). The different cations do not influence the peak potentials or the shape of the cyclic voltammograms (data not shown). These characteristics are determined solely by the anion. DPV. The electrochemical behavior of the polymer monolayers on gold was further investigated by DPV in aqueous NaClO4. A DPV voltammogram of an ES-PFS100 monolayer on gold (Figure 6) shows the two reversible redox peaks characteristic of PFSs, indicating intermetallic coupling between neighboring iron centers in the polymer chain. The voltammogram presents an additional small peak at about -0.1 V. This peak was not clearly observed by CV. By integrating the peak areas under the DPV curve, a ratio of 1:6:12 was found for all PFS monolayers. These values show that initially a small (40) Inzelt, G. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1994; Vol. 18, p 89.

Surface-Grafted Poly(ferrocenylsilane) Monolayers

Figure 6. Differential pulse voltammogram recorded for an ES-PFS100 layer on gold. Scan rate 5 mV s-1, pulse time 50 ms, interval time 200 ms, and pulse height 10 mV. The line with open circles represents the measured data; the solid lines are fitted curves.

fraction of the Fe centers is oxidized, probably those in close proximity to the metal/polymer interface. One-third of the ferrocene units are oxidized in the first oxidation step, followed by oxidation of the remaining two-thirds at higher potentials. If interchain contributions (and higherneighbor contributions for the same chain, i.e., by looping) are also present, the ratio between the first and the second wave of the corresponding oxidation steps may vary and differ from 1:1 relative weights.41 The grafting density, the chain-chain packing, the nature of the electrolyte and its concentration, and the solvation of ferrocenium ions determine the distance and, therefore, the interaction between Fe sites. The length of the spacer, linking the PFS chains to the gold substrate, influences the electrochemical properties of the polymer. Upon end-capping the living PFS chains with ethylene sulfide, typically two to three ethylene sulfide units were incorporated at the PFS chain end, as was established from 1H NMR integrals (see Chart 1, n ) 2-3). A somewhat longer spacer was formed by endcapping the PFS chains with trimethylene sulfide (n ) 6). A DPV trace (Figure 7) obtained for TMS-PFS layers on gold shows two peaks. The small peak observed for the ES-PFS films is not present in this DPV trace, supporting the idea that this signal occurs when the PFS chains are connected via short spacers to the electrode surface. By integrating the area under the peaks, a ratio of 2:3 was found. Thus, 40% of the Fe atoms were oxidized in the first step, and the remaining 60% were oxidized in the second step at higher potential. Compared to the ESPFS monolayers, more ferrocene units are oxidized in the first step in the case of the TMS-PFS films. The lower grafting density and film thickness of the TMS-PFS films (Table 2) lead to less interchain Fe-Fe interactions. Therefore, a higher fraction of redox centers can be oxidized in the first wave than in the case of the ES-PFS layers, but still less than the 50% that was observed for PFS bulk solutions. Chronoamperometry and Chronocoulometry. In addition to CV and DPV, chronoamperometry and chronocoulometry experiments36 were performed on the monolayers to gain more insight into the charge diffusion (41) Bulk electrochemistry performed on PFS homopolymers in dilute solutions (5 mM PFS in CH2Cl2 with 0.1 M Bu4NPF6 as the supporting electrolyte) showed that ferrocene sites are oxidized stepwise with a relative ratio of 1:1 between the two oxidation peaks. Under these conditions, no interchain interactions and only intramolecular Fe-Fe interactions occur. Thus, in dilute solution, each ferrocene has two immediate neighbors.

Langmuir, Vol. 21, No. 11, 2005 5121

Figure 7. Differential pulse voltammogram recorded for an TMS-PFS50 layer on gold. Scan rate 5 mV s-1, pulse time 50 ms, interval time 200 ms, pulse height 10 mV. The line with open circles represents the measured data; the solid lines are fitted curves.

Figure 8. (a) Excitation signal in the chronoamperometric and chronocoulometric experiments. A double potential step was used. The initial potential was 0.4 V. The potential was stepped to -0.4 V at t ) 0 s. At t ) τ the potential was stepped back to the initial value of 0.4 V. (b) Chronocoulometric plot obtained for an ES-PFS100 layer on gold. (c) Chronoamperometric plot obtained for ES-PFS100 layers on gold. Measurements were performed in aqueous 0.1 M NaClO4.

behavior. In both experiments a double potential step excitation signal was employed. The potential was stepped from +0.4 V standby potential to -0.4 V, held there for a period τ, and then stepped back to 0.4 V (Figure 8a). The results are shown in Figure 8b,c. At E ) 0.4 V initial potential the system is in the oxidized state. If the potential is stepped to a cathodic value, reduction occurs. This manifests itself in a charge increase (Figure 8b) in the chronocoulometric plot and in a current decrease in the chronoamperometric plot (Figure 8c) for the forward scan (t e τ). The current decreases in two steps, corresponding to the two one-electron reduction steps in the cyclic voltammograms. The forward (t e τ) and reverse (t > τ) parts of the chronoamperometric plot clearly reflect the reversibility of the redox process. It can be seen in the chronocoulometric plot (Figure 8b) that the charge increase is fast at the beginning of the reduction and then levels off. At t ) τ the reverse scan starts. The charge drops

5122

Langmuir, Vol. 21, No. 11, 2005

Pe´ ter et al. Table 7. Activation Entropies Based on Equation 8 ∆Sa (cal mol-1 K-1) -30.0 ( 0.6 -30.7 ( 0.6 -31.2 ( 0.6 -28.3 ( 0.6

ES-PFS25 ES-PFS50 ES-PFS100 TMS-PFS50

Figure 9. (a) Charge on the square root of time for the forward step (t e τ). (b) Charge plotted against θ for the reverse step (t > τ).

because the charge generated in the cathodic process is consumed to reoxidize the reduced species. Plots of the charge (Qf) versus t1/2 for the forward scan and of Qr versus the factor τ1/2 + (t - τ)1/2 - t1/2 in the reverse scan, frequently denoted as θ, are shown in Figure 9a,b, respectively. In the forward step (t e τ) the charge (Qf) can be described by the following equation36

Qf(t e τ) )

2nFACDct1/2t1/2 π1/2

+ nFAΓ + Qdl

(5)

where Dct is the charge-transport diffusion constant, C is the concentration of ferrocene sites, Γ is the surface excess of electroactive sites (mol cm-2), and Qdl is the double layer charging. The first, time-dependent term in eq 4 is the contribution of charge diffusion. The second and third terms are time-independent and represent the contributions of adsorbed electroactive centers and double-layer charging, respectively. The values of Dct and n can be determined from the slope of the plot of Qf versus t1/2. The intercept equals the sum of the second and third terms in eq 5. When the intercept is negative it can be regarded as an effect of heterogeneous kinetics.36,42,43 In such a case a rate limitation on the charge delivery to the redox sites produces an intercept which is smaller than that predicted by eq 5. The charge in the reverse step is given by36,44

2nFADct1/2C 1/2 [τ + (t - τ)1/2 - t1/2] ) Qr(t > τ) ) Qdl + 1/2 π 2nFADct1/2C Qdl + θ (6) π1/2 A plot of Qr versus θ (Figure 9b) shows that the charge generated in the forward step is consumed in the reverse (42) Christie, J. H.; Lauer, G.; Osteryoung, R. A.; Anson, F. C. Anal. Chem. 1963, 35, 1979. (43) Christie, J. H.; Lauer, G.; Osteryoung, R. A. J. Electroanal. Chem. 1964, 7, 60. (44) Strobel, H. A.; Heineman, W. R. Chemical Instrumentation: A Systematic Approach, 3rd ed.; John Wiley & Sons: New York, 1989.

step for the reoxidation of reduced sites. The plot of the charge Qf against t1/2 has two linear sections (Figure 9a). The linear fit to the first segment intercepts the t1/2 axis at t ) ti. The sign of the intercept is different from the sign of the slope. This is due to a rate limitation on the delivery of charge to the redox sites.36 From the slope of the line, the heterogeneous rate constant (kf) can be determined for the reduction step corresponding to Epred2 in the CVs shown in Figure 2. The applied potential, E ) -0.4 V, is sufficiently negative to consider that the slope will approach the Cottrell slope38 (see eq 5). Thus, the product Dct21/2C can be determined for this reduction step. A linear fit to the second segment allows the determination of Dct11/2C for the reduction step corresponding to Epred1 in the CVs shown in Figure 2. The corresponding data are summarized in Table 5. The concentration of electroactive sites in the polymer film was calculated from C ) ΓFc/d, where ΓFc is the surface coverage of ferrocene sites (obtained from CV) and d is the layer thickness. Dct1 and Dct2 can be calculated if C is known. With Dct in hand the parameter Dctτ/d2, discussed in the CV section, can be established. The calculated parameters are shown in Table 6. The obtained charge diffusion coefficients are extremely small for both reduction steps (around 10-12 and 10-14 cm2 s-1). However, the diffusion coefficients for the Epred2 step are larger by 2 orders of magnitude than for the Epred1 step. This result agrees with the CV results and shows that segmental mobility of the polymer chains and counterion mobility determine the rate of charge transport. Upon oxidation, the PFS chains are turned into polyelectrolytes, which enhances their solvation in the aqueous supporting electrolyte. The monolayer film swells, and the counterion mobility in the film is increased. Activation energies for thermally activated diffusion were determined on the basis of the following equation:38

( )

Dct1/2C ) D01/2C exp

-Ea RT

(7)

where D0 is the preexponential factor for which a value of 10-8 cm2 s-1 was assumed.45 Measurements were performed at room temperature (298 K), and the results are given in Table 5. As expected, the Epred2 step requires a lower activation energy then the Epred1 step. Activation entropies (∆Sa) can be estimated by the following equation, which uses the Eyring formulation for the preexponential factor D0:46

( )

kT ∆Sa exp D0 ) eλ2 h R

(8)

where λ is the diffusion jump distance. λ can be estimated from d/λ ) ΓFc/ΓFc monolayer, where ΓFc is the surface coverage by ferrocene units of the PFS chains and ΓFc monolayer ≈ 2 × 10-10 mol cm-2.38 In eq 8, k is Boltzmann’s constant (k ) 1.380 65 × 10-23 J K-1) and h is Planck’s constant (h ) 6.626 07 × 10-34 J s). Results are summarized in Table 7. The negative activation entropies support the idea that (45) MacGregor, R. Diffusion and Sorption in Films and Fibers; Academic Press: New York, 1974; Vol. 1. (46) Bowers, R. C.; Murray, R. W. Anal. Chem. 1966, 38, 461.

Surface-Grafted Poly(ferrocenylsilane) Monolayers

charge transport occurs between ferrocene and ferrocenium sites by electron hopping, as this requires an ordering event, involving the deformation of polymer chains, to bring these sites together.38 Such a mechanism also explains the small values obtained for the charge diffusion coefficients. Conclusions Poly(ferrocenyldimethylsilane) monolayers were prepared on gold by the self-assembly technique. Electrochemical characterization of the films was undertaken using CV, DPV, chronoamperometry, and chronocoulometry. Cyclic voltammograms revealed two reversible redox peaks, indicating the stepwise oxidation of the ferrocene units in the polymer chain. Peak currents (ip) plotted against the square root of scan rates (v1/2) for the first oxidation peak and for the corresponding reduction peak exhibited a linear dependence, indicating that the oxidation process in the first step is controlled by the diffusion of counterions into the polymer film. For the second oxidation peak and the corresponding reduction peak, ip varied linearly with v. This redox behavior is characteristic of surface-confined electroactive layers. From DPV experiments it was found that a small fraction of ferrocene units in close proximity to the metal/polymer interface is

Langmuir, Vol. 21, No. 11, 2005 5123

oxidized first; subsequently, one-third of the remaining ferrocenes are oxidized at higher potentials. The oxidation goes to completion at even higher potentials. Kinetic parameters were estimated from chronocoulometry experiments. Starting from a fully oxidized film, charge diffusion coefficients were established for the two ensuing reduction steps. Diffusion coefficients for the two steps were extremely small (around 10-12 and 10-14 cm2 s-1), but the diffusion coefficient associated with the higher potential Epred2 step was 2 orders of magnitude larger than for the Epred1 step. This was also reflected in the higher reversibility of the second oxidation and reduction waves in the CV experiments. The difference in electrochemical behavior between the two redox steps can be explained from the solvation of the surface-grafted PFS chains, which depends on the degree of oxidation. Oxidized PFS films are swollen in the aqueous electrolyte solutions, leading to a higher segmental mobility of the polymer chains and a much increased counterion mobility within the film. Acknowledgment. The authors thank the University of Twente for financial support, and Menno van Os for the SPRS measurements. Stefan Kooij is acknowledged for the ellipsometry measurements. LA0473409