Measurement of Apparent Diffusion Coefficients within Ultrathin Nafion

A numerical model is developed that allows Dapp to be determined. For these studies, ultrathin films of Nafion were prepared using the Langmuir−Scha...
0 downloads 0 Views 464KB Size
Download PDF
10380

Langmuir 2006, 22, 10380-10388

Measurement of Apparent Diffusion Coefficients within Ultrathin Nafion Langmuir-Schaefer Films: Comparison of a Novel Scanning Electrochemical Microscopy Approach with Cyclic Voltammetry† Paolo Bertoncello,* Ilenia Ciani, Fei Li, and Patrick R. Unwin* Department of Chemistry, UniVersity of Warwick, CoVentry CV4 7AL, United Kingdom ReceiVed May 1, 2006. In Final Form: July 17, 2006 The use of scanning electrochemical microscopy (SECM) to evaluate the apparent diffusion coefficient, Dapp, of redox-active species in ultrathin Nafion films is described. In this technique, an ultramicroelectrode (UME) tip, positioned close to a film on a macroscopic electrode, is used to oxidize (or reduce) a species in bulk solution, causing the tip-generated oxidant (reductant) to diffuse to the film/solution interface. The oxidation (reduction) of filmconfined species regenerates the reductant (oxidant) in solution, leading to feedback to the UME. A numerical model is developed that allows Dapp to be determined. For these studies, ultrathin films of Nafion were prepared using the Langmuir-Schaefer (LS) technique and loaded with an electroactive species, either the ferrocene derivative ferrocenyltrimethylammonium cation, FA+, or tris(2,2′-bipyridyl)ruthenium(II), Ru(bpy)32+. The morphology and the thickness of the Nafion LS films (1.5 ( 0.2 nm per layer deposited) were evaluated using atomic force microscopy (AFM). For comparison with the SECM measurements, cyclic voltammetry (CV) was employed to evaluate the concentration of electroactive species within the Nafion LS films and to determine Dapp. The latter was found to be essentially invariant with film thickness, but the value for Ru(bpy)32+ was 1 order of magnitude larger than for FA+. CV and SECM measurements yield different values of Dapp, and the underlying reasons are discussed. In general, the Dapp values for these films are considerably smaller than for recast Nafion films, which can be attributed to the compactness of Nafion LS films. Nonetheless, the ultrathin nature of the films leads to fast response times, and we thus expect that these modified electrodes could find applications in sensing, electroanalysis, and electrocatalysis.

Introduction Perfluorinated ionomers have attracted great interest as a result of their intrinsic properties such as ion-exchange selectivity, thermal stability, and chemical-biological inertness.1 Among the family of perfluorinated ionomers, Nafion (a trademark of Du Pont de Nemours & Co.) has received considerable attention, primarily because of applications as a proton-exchange membrane (PEM) in fuel cells2,3 and as a membrane in water electrolizers4 and chloroalkali cells.5 In hydroalcoholic solution form, Nafion is also used widely in chemically modified electrodes,6-9 ion-selective electrodes,10 and voltammetric sensors.11-13 The considerable interest in Nafion is evident not only from the large number of reports concerned with technological applications14,15 but also from studies †

Part of the Electrochemistry special issue. * Corresponding authors. (P.R.U.) E-mail: [email protected]. Phone: +44 (0) 2476 523264. Fax: +44 (0) 2476 524112. (P.B.) E-mail: [email protected]. (1) Eisenberg, A., Yeager, H. L., Eds. Perfluorinated Ionomer Membranes; American Chemical Society: Washington, DC, 1982; ACS Symposium Series 180. (2) Tazi, B.; Savadogo, O. Electrochim. Acta 2000, 45, 4329. (3) Xing, B.; Savadogo, O. Electrochem. Commun. 2000, 2, 697. (4) Pillai, K. C.; Kumar, A. S.; Zen, J.-M. J. Mol. Catal., A 2000, 160, 277. (5) Cutler, S. G. In Ions in Polymers, Eisenberg, A., Ed.; American Chemical Society: Washington, DC, 1980; Chapter 9. (6) White, H. S.; Leddy, J.; Bard, A. J. J. Am. Chem. Soc. 1982, 104, 4811. (7) Leddy, J.; Bard, A. J. J. Electroanal. Chem. 1985, 189, 203. (8) Buttry, D. A.; Anson, F. C. J. Am. Chem. Soc. 1982, 104, 4824. (9) Martin, C. R.; Rubinstein, I.; Bard, A. J. J. Am. Chem. Soc. 1982, 104, 4817. (10) Martin, C. R.; Freiser, H. Anal. Chem. 1981, 53, 902. (11) Ugo, P.; Moretto, L. M.; Vezza’, F. Chem. Phys. Chem. 2002, 3, 917. (12) Gerardi, R. D.; Barnett, N. W.; Lewis, S. W. Anal. Chim. Acta 1999, 378, 1. (13) Brett, C. M. A.; Fungaro, D. A.; Morgado, J. M.; Gil, M. H. J. Electroanal. Chem. 1999, 468, 26. (14) Smitha, B.; Sridhar, S.; Khan, A. A. J. Membr. Sci. 2005, 259, 10. (15) Rikukawa, M.; Sanui, K. Prog. Polym. Sci. 2005, 25, 1463.

related to the morphology, transport properties, thermal characteristics, and mechanical properties of Nafion as a membrane16-20 and as a film on electrodes.6,11,21-23 Charge transfer through ionomeric thin films on electrode surfaces has been studied extensively.24-27 For heterogeneous electron transfer to redox species in the film, it has been established that the current depends either on (i) electron self-exchange between redox moieties incorporated in the film28 and/or (ii) the rate of physical diffusion of the redox moieties. Additionally, ion diffusion into or out of the film to maintain electroneutrality may influence the response.29 The study of charge transfer in ionomers with incorporated redox species has largely relied on cyclic voltammetry (CV)30 and potential step chronoamperometry (CA) measurements.31-34 (16) Mauritz, K. A.; Moore, R. B. Chem. ReV. 2004, 104, 4535. (17) Gargas, D. J.; Bussian, D. A.; Buratto, S. K. Nano Lett. 2005, 5, 2184. (18) Zudans, I.; Heineman, W. R.; Seliskar, C. J. J. Phys. Chem. B 2004, 108, 11521. (19) Rollins, H. W.; Lin, F.; Johnson, J.; Ma, J.-J.; Liu, J.-T.; Tu, M.-H.; DesMarteau, D. D.; Sun, Y.-P. Langmuir 2000, 16, 8031. (20) Sun, Y.-P.; Atorngitjawat, P.; Lin, Y.; Liu, P.; Pathak, P.; Bandara, Y.; Elgin, D.; Zhang, M. J. Membr. Sci. 2004, 245, 211. (21) Rubinstein, I.; Bard, A. J. J. Am. Chem. Soc. 1981, 103, 5007. (22) Szentirmay, M. N.; Martin, C. R. Anal. Chem. 1984, 56, 1898. (23) Pantelic, N.; Wansapura, C. M.; Heineman, W. R.; Seliskar, C. J. J. Phys. Chem. B 2005, 109, 13971. (24) Lee, C.; Anson, F. C. Anal. Chem. 1992, 64, 528. (25) Anson, F. C.; Blauch, D. N.; Saveant, J. M.; Shu, C.-F. J. Am. Chem. Soc. 1991, 113, 1922. (26) Murray, R. W. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1984; Vol. 13, pp 191-368. (27) Buttry, D. A.; Anson, F. C. J. Am. Chem. Soc. 1983, 105, 685. (28) Shi, M.; Anson, F. C. Langmuir 1996, 12, 2068. (29) Safranj, A.; Gershuni, S.; Rabani, J. Langmuir 1993, 9, 3676. (30) Buttry, D. A.; Saveant, J. M.; Anson, F. C. J. Phys. Chem. 1984, 88, 3086. (31) Arkoub, I. A.; Amatore, C.; Sella, C.; Thouin, L.; Warcocz, J.-S. J. Phys. Chem. B 2001, 105, 8694. (32) Yagi, M.; Sato, T. J. Phys. Chem. B 2003, 107, 4975. (33) Blauch, D. N.; Saveant, J. M. J. Am. Chem. Soc. 1992, 114, 3323. (34) Whiteley, L. D.; Martin, C. R. J. Phys. Chem. 1989, 93, 4650.

10.1021/la061214i CCC: $33.50 © 2006 American Chemical Society Published on Web 08/23/2006

Measurement of Apparent Diffusion Coefficients

Langmuir, Vol. 22, No. 25, 2006 10381

Figure 1. Schematic of the SECM experiment for measuring the diffusion coefficient of a redox species in an ultrathin film. The notation and geometry used in the simulations are also shown (not to scale).

In previous studies of Nafion, the polymer has typically been deposited onto electrodes in the form of recast films, with thicknesses ranging from hundreds of nanometers to a few micrometers.21,35-37 Recently, one of us has demonstrated the possibility of fabricating ultrathin films (,100 nm) of Nafion and other ionomeric polymers of well-controlled thickness using the Langmuir-Schaefer (LS) technique.38,39 Interestingly, preliminary CV studies on Nafion LS films loaded with Ru[(NH3)6]3+ reported an apparent diffusion coefficient, Dapp, on the order of 10-11 cm2 s-1 or less.40 This is more than an order of magnitude lower than Dapp measured in recast films.28,34,41,42 This difference may be attributed to the increased degree of condensation and compactness within Nafion LS films. However, there are often significant discrepancies in Dapp values for redox polymer systems determined by CV and CA.25,43 Both techniques are transient in nature, but in CA, a diffusion-limited driving force is imposed instantaneously, whereas in CV, the redox state of the film is changed more gradually. In both cases, significant demands are placed on the system to maintain electroneutrality in the thin film by ion ingress or egress at the film/solution boundary. In this article, we report a novel SECM procedure and a numerical treatment to calculate Dapp. SECM has proven to be powerful for the study of a wide range of interfacial processes44-51 and has recently been used to measure charge transfer and conductivity in several different polymeric systems.52-55 We (35) Shi, M.; Anson, F. C. Anal. Chem. 1997, 69, 2653. (36) Terrill, R. H.; Sheehan, P. E.; Long, V. C.; Washburn, S., Murray, R. W. J. Phys. Chem. 1994, 98, 5127. (37) Mirkin, M. V.; Fan, F.-R. F.; Bard, A. J. Science 1992, 257, 364. (38) Bertoncello, P.; Ram, M. K.; Notargiacomo, A.; Ugo, P.; Nicolini, C. Phys. Chem. Chem. Phys. 2002, 4, 4036. (39) Ugo, P.; Bertoncello, P.; Vezza’, F. Electrochim. Acta 2004, 49, 3785. (40) Bertoncello, P.; Ugo, P. J. Braz. Chem. Soc. 2003, 14, 517. The equation for the evaluation of the apparent diffusion coefficient is erroneously reported. The value inside the parentheses has to be squared. (41) Martin, C. R.; Dollard, K. A. J. Electroanal. Chem. 1983, 159, 127. (42) Daniele, S.; Ugo, P.; Bragato, C.; Mazzocchin, G. A. J. Electroanal. Chem. 1996, 418, 29. (43) Buttry, D. A.; Anson, F. C. J. Electroanal. Chem. 1981, 130, 333.

have made a systematic study of Nafion LS films loaded in the ferrocenyltrimethylammonium cation (FA+) and tris(2,2′-bipyridyl)-ruthenium(II) dichloride hexahydrate (Ru(bpy)32+). These species are of interest because previous work on thick Nafion films has shown that the redox response for FA+ is essentially governed by physical diffusion, whereas for Ru(bpy)32+ the current response is largely dependent on electron hopping due to the high rate constant for self-exchange in this system.6 The values of Dapp measured by steady-state SECM and conventional CV experiments are compared, and results from the two methods are discussed.

Theory for the SECM Measurement The SECM process (Figure 1) was modeled using a commercial finite element method modeling package (FEMLAB, version 3.1) in conjunction with MATLAB (version 7.1, release 14).56,57 This system was run on a Dell desktop PC running Windows XP with 1.5 GB of RAM and a 2.52 GHz Pentium 4 processor. The following system is considered: a film (Nafion in our case) loaded in an electroactive species (FA+ or Ru(bpy)32+), (44) Unwin, P. R. Bard, A. J. J. Phys. Chem. 1991, 95, 7814. (45) Pierce, D. T.; Unwin, P. R.; Bard, A. J. Anal. Chem. 1992, 64, 1795. (46) Slevin, C. J.; Macpherson, J. V.; Unwin, P. R. J. Phys. Chem. B 1997, 101, 10851. (47) Slevin, C. J., Unwin, P. R. J. Am. Chem. Soc. 2000, 122, 2597. (48) Mirkin, M. V.; Horrocks, B. R. Anal. Chim. Acta 2000, 406, 119. (49) Liu, B.; Bard, A. J.; Mirkin, M. V.; Creager, S. E. J. Am. Chem. Soc. 2004, 126, 1485. (50) Fan, F.-R. F.; Mirkin, M. V.; Bard, A. J. J. Phys. Chem. B 1994, 98, 1475. (51) Wei, C.; Bard, A. J. J. Electrochem. Soc. 1995, 142, 2523. (52) Mandler, D.; Unwin, P. R. J. Phys. Chem. B 2003, 107, 407. (53) O’Mullane, A. P.; Macpherson, J. V.; Unwin, P. R.; Cervera-Montesinos, J.; Manzanares, J. A.; Frehill, F.; Vos, J. G. J. Phys. Chem. B 2004, 108, 7219. (54) Ruiz, V.; Nicholson, P. G.; Jollands, S.; Thomas, P. A.; Macpherson, J. V.; Unwin, P. R. J. Phys. Chem. B 2005, 109, 19335. (55) Zhang, J.; Barker, A. L.; Mandler, D.; Unwin, P. R. J. Am. Chem. Soc. 2003, 125, 9312. (56) Rudd, N. C.; Cannan, S.; Bitziou, E.; Ciani, I.; Whitworth, A. L.; Unwin, P. R. Anal. Chem. 2005, 77, 6205. (57) Ciani, I.; Burt, D. P.; Daniele, S.; Unwin, P. R. J. Phys. Chem B 2004, 108, 3801.

10382 Langmuir, Vol. 22, No. 25, 2006

Bertoncello et al.

Red, is deposited onto a conductive substrate (ITO), and a drop of aqueous solution, containing the same species Red, is dropped onto the substrate (Figure 1). Because Red has a different concentration and diffusion coefficient in the aqueous phase (phase 1) and in the Nafion solid phase (phase 2), the diffusion problem pertaining to the oxidation of Red at the UME probe needs to be solved for the two redox couples, one in phase 1 and the other in phase 2. The problem can be simplified by considering only one diffusion coefficient for each couple: DRed1 ) DOx1, and DRed2 ) DOx2. Because Red is initially present only in phase 1 at concentration * * cRed and in phase 2 at concentration cRed , the principle of mass 1 2 conservation for phases 1 and 2 allows us to write * 0 er erg, 0 < z < d: cOx1(r, z) ) cRed - cRed1(r, z) (1) 1 * 0 er erg, d < z < h: cOx2(r, z) ) cRed - cRed2(r, z) (2) 2

where cRedi(r, z) and cOxi(r, z) are the spatial-dependent concentrations of Oxi and Redi (i ) 1 or 2), respectively, within the region of interest (Figure 1) defined in terms of r and z, which are the coordinates in the directions radial and normal to the electrode surface, measured from the center of the electrode. The use of this principle requires that the diffusion coefficients of the oxidized and reduced forms of the mediators are equivalent, which is a good approximation for the systems herein. Parameters d and h denote, respectively, the location of the aqueous phase/ Nafion interface and the substrate electrode/film interface with respect to the electrode surface. The edge of the glass surrounding the UME, from the cylindrical axis of symmetry, is denoted by rg. The use of eqs 1 and 2 simplifies the problem to the consideration of species Red1 and Red2 alone. The oxidation of Red1 under diffusion control occurs at the probe UME:

Red1 f Ox1 + e

-

(3)

The diffusion equation, under steady-state conditions, for each of the two phases in the axisymmetric cylindrical geometry of the SECM is

[

0 ) DRedi

∂2cRedi ∂r2

]

2 1 ∂cRedi ∂ cRedi + + r ∂r ∂z2

(4)

where the subscript i indicates phase 1 or 2. The associated boundary conditions are

z ) 0, 0 e r e a, cRed1 ) 0

(5)

∂cRed1 z ) 0, a < r e rg: DRed1 )0 ∂z

(6)

r ) rg, 0 < z e d: cRed1 )

* cRed 1

(7)

* z ) h, 0 er e rg, cRed2 ) cRed 2

(8)

* r ) rg, d < z < h: cRed2 ) cRed 2

(9)

r ) 0, 0 < z < d: DRed1 r ) 0, d < z < h: DRed2

∂cRed1 ∂r ∂cRed2 ∂r

)0

(10)

)0

(11)

The potential of the UME is set at a sufficient value to drive the oxidation of Red1 at a diffusion-controlled rate (eq 5). In phase 1, Red1 is assumed to be inert with respect to the insulating glass sheath surrounding the electrode (eq 6), and it remains at bulk concentration values beyond the radial edge of the tip (eq 7). The latter is a good approximation for the UME geometries used herein.44-46 In phase 2, Red2 attains the bulk concentration on the macroscopic ITO surface (eq 8) and far from the tip electrode (eq 9). The axisymmetric cylindrical geometry of the SECM dictates that eqs 10 and 11 apply. The generation of Ox1 at the UME causes a flux of oxidant toward the Nafion film, which can oxidize Red2 loaded inside the film58-62 (Figure 1). Thus, the following redox process occurs at the interface between the aqueous phase and Nafion

Ox1 + Red2 h Ox2 + Red1

(12)

which causes Red1 to be regenerated, leading to feedback of Red1 to the UME.63 The reaction in eq 12 is a fast outer sphere self-exchange redox process for the systems of interest herein. The final interior boundary condition determines the flux of species Red1 and Red2, at the aqueous/Nafion interface, as a consequence of this second-order redox reaction

∂cRed1 ∂cRed2 ) DRed2 ) z ) d, all r: DRed1 ∂z ∂z * * c (r, d)) (13) k(cRed1cRed2(r, d) - cRed 2 Red1 where k is the interfacial rate constant for the redox reaction.6,64 Given the fast rate of electron transfer compared to the SECM time scale, we set this value at 100 cm s-1 M-1 so that the process was diffusion-controlled. The diffusion problem was solved in real space with experimentally relevant parameters to allow direct comparison of experiment and simulation. Thus, the film thickness was a variable, as well as the (usual) tip/substrate separation.The (r, z) domain used is illustrated in Figure 1 (dotted region): because the value of rg was at least 10 times the electrode radius, we used that value as a maximum for the r domain. This approximation has been shown to be valid44 and was used to speed up the computations. The Nafion LS films were ultrathin (maximum thickness ca. 70 nm) compared to the tip/substrate separation. Thus, for compatibility of the grid spacing in the two phases, while avoiding the use of an impractical number of elements, the z coordinate inside the film was stretched (eq 14), and an anisotropic diffusion coefficient DRed2 was used (eqs 15 and 16): (58) Bard, A. J.; Fan, F.-R. F.; Pierce, D. T.; Unwin, P. R.; Wipf, D. O.; Zhou, F. M. Science 1991, 254, 68. (59) Barker, A. L.; Gonsalves, M.; Macpherson, J. V.; Slevin, C. J.; Unwin, P. R. Anal. Chim. Acta 1999, 385, 223. (60) Bard, A. J.; Fan, F.-R. F.; Kwak, J.; Lev, O. Anal. Chem. 1989, 61, 132. (61) Wipf, D. O.; Bard, A. J. J. Electrochem. Soc. 1991, 138, L4. (62) Bard, A. J.; Mirkin, M. V.; Unwin, P. R.; Wipf, D. O. J. Phys. Chem. 1992, 96, 1861. (63) Barker, A. L.; Macpherson, J. V.; Slevin, C. J.; Unwin, P. R. J. Phys. Chem. B 1998, 102, 1586. (64) Wilkins, R. G. Kinetics and Mechanism of Reactions of Transition Metal Complexes, 2nd ed.; VCH Publishers: New York, 1991; Chapter 5, pp 262-268.

Measurement of Apparent Diffusion Coefficients

all r, d e z e h

Langmuir, Vol. 22, No. 25, 2006 10383

z f Pz

(14)

where P ) 10 in our simulations. The anisotropic diffusion coefficient has two different components along either r or z:

all r, d e z e h: (DRed2(r))1 )

DRed2 P

all r, d e z e h: (DRed2(z))2 ) PDRed2

(15) (16)

Experimental Section Materials. Nafion 117 solution (5% w/v mixture of low-molecularweight alcohols), tris(2,2′-bipyridyl)ruthenium(II) dichloride hexahydrate, and NaCl were purchased from Sigma-Aldrich. Trimethylammoniomethylferrocene (Cp2FeTMA+PF6-) was obtained from the metathesis of trimethylammoniomethylferrocene iodide (Cp2FeTMA+I-) with AgPF6.22 All other chemicals were of reagentgrade quality and used as received. Polymeric films were deposited on several types of substrates: indium tin oxide (ITO) glass plates from Delta Technologies Ltd. with a resistivity e10 Ω/square (for electrochemical and SECM experiments) and oxidized (100-nmthick) silicon wafers (n-type, IDB Technologies Ltd.) for AFM measurements. Both types of substrates were cut into squares of side 15 mm. ITO glass plates were sonicated in 2-propanol, acetone, and chloroform for at least 5 min before use. Silicon oxide wafers were cleaned in piranha solution. CAUTION: Piranha solution reacts Violently with organics and should be handled with extreme caution! The hydrophilic silicon oxide substrates were then silanized using 5% v/v dimethyldichlorosilane in hexane followed by washing in hexane, acetone, and hexane for 20 min. In this way, the substrates were rendered hydrophobic. The substrates were then dried using nitrogen gas. Immediately before use, the substrates were further cleaned with chloroform. All aqueous solutions were prepared from doubly distilled Milli-Q reagent water (Millipore Corp.) with a resistivity g18 MΩ cm at 25 °C. All solutions contained 0.1 M NaCl (Aldrich) as the supporting electrolyte. Tapping-Mode Atomic Force Microscopy (TM-AFM). TMAFM images were taken in air using Si tips (LOT-Oriel, U.K.) with a Digital Instruments multimode AFM and Nanoscope IIIa controller (Veeco). The AFM technique was also used to estimate the thickness of the Nafion LS films following a standard procedure.38 Electrochemistry. Cyclic voltammetry (CV) and approach curves were recorded using an electrochemical analyzer (CH Instruments, model CHI730A). For cyclic voltammetry, a typical three-electrode configuration was used, where the working electrode was an ITOcoated glass plate on which a Nafion LS film was deposited, a platinum gauze was used as the counter electrode, and a Ag/AgCl (KCl-saturated) electrode was the reference electrode. The area of the working electrode was typically between 0.2 and 0.3 cm2 and was kept constant for each set of CV measurements. SECM Setup. SECM measurements were performed using a simple home-built SECM instrument consisting of a manual x, y, z stage (M-431, Newport Corp., Irvine, CA) and a z-axis piezoelectric positioner and controller (models P-843.30 and E501.00, Physik Instrumente, Waldbronn, Germany) to give high-precision control of the UME tip position in the direction normal to the sample. A two-electrode arrangement was employed, with a Ag wire as the quasi-reference electrode and a 25-µm-diameter Pt disk as the working electrode. The latter was characterized by RG ) 10 (ratio of the overall tip radius to that of the platinum disk). FA+, Ru(bpy)32+, and NaCl were used as the two redox mediators and base electrolyte, respectively. A schematic of the basis of the experimental measurement is shown in Figure 1. Briefly, a drop (50 µL) of aqueous solution containing a particular concentration of Ru(bpy)32+ or FA+ in 0.1 M NaCl was pipetted onto the ITO electrode on which different layers of Nafion had been deposited and in which the redox mediator of interest (Ru(bpy)32+ or FA+) was incorporated at saturated values; see below. Approach curves of the current due to the oxidation of

Ru(bpy)32+ or FA+ at the tip, as a function of the distance, d, of the tip from the substrate, were recorded for a series of redox mediator concentrations in solution. Moreover, measurements were made for samples in which the number of Nafion layers was varied. All of the experiments were carried out in an air-conditioned room at a temperature of 22 ( 1 °C. The SECM experimental protocol utilized was as follows: Initial measurements were performed using the highest concentration of FA+ or Ru(bpy)32+ in solution, typically 10-3 M. Cyclic voltammograms for the oxidation of the redox mediator were recorded at a slow scan rate (5 mV s-1) at the UME tip positioned in bulk solution, far from the ITO/Nafion substrate. In this way, the steadystate diffusion-limited current, ibulk, could be ascertained. The tip was then biased at the diffusion-limited oxidation potential for FA+ or Ru(bpy)32+, and an approach curve was recorded by slowly (1 µm s-1) translating the tip toward the substrate and measuring the current as a function of tip position. The tip was then retracted, and the solution removed and replaced with one of a different concentration. The bulk current was measured, and approach curves were then recorded in a similar fashion. This procedure was repeated for several concentrations of the redox mediators (maintaining a constant base electrolyte concentration). There was no cross contamination of droplets, as verified by a linear relationship between the tip current and concentration, which yielded reliable values for the diffusion coefficients of the redox species.65 It was possible to probe the same spot on the sample with different mediator concentrations, but several spots were tested and the responses obtained were very similar. The distance of closest approach of the tip from the substrate was determined from the hindered diffusion response with either an inert glass substrate or with a Nafion film at sufficiently high bulk concentration in the aqueous solution that the process in the film made a negligible contribution to the overall tip current response. Fabrication of Nafion Langmuir-Schaefer Films. Nafion LS films were fabricated using the procedure of Bertoncello et al.38,66 Briefly, a stock solution of Nafion was prepared by dilution of the commercial solution with methanol to obtain a final concentration of 0.85 mg/mL. On the basis of the methanol/water solvent ratio, Nafion is expected to attain a micellar conformation with the polar sulfonate group located on the surface of the micelles and the hydrophobic fluorocarbon chain in the inner part.67 Nafion Langmuir monolayers were formed using a Langmuir trough (Nima Instruments, Coventry, U.K.). The surface pressure was measured by means of a Wilhemy balance with an accuracy of (0.1 mN m-1. The volume of Nafion added to the subphase varied between 100 and 200 µL. An elapsed time of 2 min was allowed before compression of the floating films. On the basis of the Langmuir isotherm, the conditions used for the fabrication of Nafion LS films were a surface pressure of 20 mN m-1 and 0.1 M NaCl as a subphase.

Results and Discussion Nafion Π-A Isotherms. In previous work, one of us has demonstrated the possibility of fabricating ultrathin Nafion LS films.38 The addition of a strong electrolyte to the aqueous subphase (0.1 M NaCl) is necessary to obtain stable pressurearea (Π-A) isotherms at the air-water interface and to allow the transfer of the Langmuir films onto solid substrates by the LS technique. The Π-A isotherm obtained under these conditions (Supporting Information, S1) revealed the typical behavior of a Nafion monolayer: a high collapse pressure of the film in excess of 50 mN m-1 and a solid-phase transition from 10 to 50 mN m-1 in which the degree of packing of the Nafion macromolecules increases. On the basis of the Π-A isotherms, we transferred Nafion LS films at a surface pressure of 20 mN m-1. A barrier (65) Saito, Y. ReV. Polarogr. 1968, 15, 177. (66) Bertoncello, P.; Notargiacomo, A.; Nicolini, C. Langmuir 2005, 21, 172. (67) Yeo R. S.; Yeager H. L. In Modern Aspects of Electrochemsitry; Conway B. E., White R. E., Bockris, J. O’M., Eds.; Plenum Press: New York, 1985; Vol. 16, Chapter 6, p 454.

10384 Langmuir, Vol. 22, No. 25, 2006

Bertoncello et al.

Figure 2. AFM topography images (1 µm × 1 µm) of Nafion LS films in air: 10 layers deposited on a silicon oxide substrate before (a) and after loading in 5 × 10-4 M FA+ (b) and 5 × 10-4 M Ru(bpy)32+(c).

speed of 100 cm2 min-1 was employed to record the isotherm in S1; the barrier speed did not appreciably influence the Π-A isotherms over the range of 30 to 100 cm2 min-1 (data not shown). AFM Studies of LS Films. Figure 2 shows typical 1 µm × 1 µm AFM topography images of a 10-layer Nafion LS film deposited on a flat oxidized silicon surface before (a) and after loading to saturation from solutions containing 5 × 10-4 M FA+ (b) and 5 × 10-4 M Ru(bpy)32+ (c). The 10-layer Nafion LS unloaded film covers the silicon substrate reasonably well, although a few voids are evident, which we ascribe to the drying process. The film thickness in the covered area was in the range of 13-17 nm, consistent with data already reported in recent work by one of us.38,66 The image in Figure 2a reveals the presence of large clusters, typical of Nafion ionomers. Although there have been numerous investigations of the morphology and structure of perfluorinated ionomers with different models proposed,1,68-71 it is widely accepted that perfluorinated ionomers have ordered structures in which the hydrophilic sulfonated end groups aggregate in domains located within the hydrophobic (fluorocarbon backbone) matrix.1,67,69 Even in the dry state, after being loaded in FA+ and Ru(bpy)32+ the film is seen to cover the surface completely with thicknesses of 16 ( 2 nm (FA+loaded film) and 17 ( 2 nm (Ru(bpy)32+-loaded film). The roughnesses of the FA+- and Ru(bpy)32+-loaded films were 1.7 and 1.5 nm, respectively. Of course, it is important to note that films are expected to appear rougher in the dry state than under solution, so these values should be considered to be upper limits of the film roughness. In the main, the data in Figure 2 demonstrate that the loaded films are reasonably homogeneous, certainly on the length scale of SECM measurements. Clearly, as the thickness of the films has been determined under dry conditions, swelling of Nafion LS films in aqueous solution cannot be excluded.18,71 Electrochemical Characterization of Nafion LS Films: Cyclic Voltammetry. Figure 3 shows typical CVs (scan rate 0.02 V s-1) of 20-layer Nafion LS films loaded in 5 × 10-4 M FA+ (a) and 5 × 10-4 M Ru(bpy)32+ (b). CVs are shown in the (68) Yeager, H. L.; Steck, J. J. Electrochem. Soc. 1981, 128, 1880. (69) Gierke, T. D.; Munn, G. E.; Wilson, F. C. J. Polym. Sci., Polym. Phys. 1981, 19, 1687. (70) Heitner-Wirguin, C. J. Membr. Sci. 1996, 120, 1. (71) Krtil, P.; Trojanek, A.; Samec, Z. J. Phys. Chem. B 2001, 105, 7979.

Figure 3. CVs of 20-layer Nafion LS films loaded in 5 × 10-4 M FA+ (a, -) and after transferring to 0.1 M NaCl supporting electrolyte (a, - -); loaded in 5 × 10-4 M Ru(bpy)32+ (b, -) and after transferring to 0.1 M NaCl supporting electrolyte (b, - -); scan rate, 0.02 V s-1.

loading medium (solid line) and after transferring to 0.1 M NaCl supporting electrolyte (dashed curves in Figure 3a and b). The evident redox response of both types of films after transferring to the medium containing only supporting electrolyte reveals the successful incorporation of both FA+ and Ru(bpy)32+ within the Nafion LS films. The anodic peaks for the oxidation of FA+ and

Measurement of Apparent Diffusion Coefficients

Langmuir, Vol. 22, No. 25, 2006 10385

Figure 4. Dependence of the anodic peak current on the scan rate (a, c) and square root of the scan rate (b, d) for 20- (2), 30- (b) and 40-layer (9) Nafion LS films. The films were preloaded from solutions containing 10-5 M FA+ (a, b) and 10-5 M Ru(bpy)32+ (c, d) and transferred to 0.1 M NaCl aqueous electrolyte for the measurements. Table 1. Parameters Extracted Using Cyclic Voltammetry Relating to 20-, 30-, 40-, and 50-Layer Nafion LS Films Loaded in 10-5 M Fa+ (Top) and 10-5 M Ru(bpy)32+ (Bottom) after Transferring to 0.1 M NaCl as the Supporting Electrolyte Nafion LS films FA+

surface coverage Γ/ (10-10) mol cm-2

Dapp/ (10-12) cm2 s-1

Cp/ mol dm-3

20 30 40 50

5.4 ( 0.5 9.4 ( 0.9 11.0 ( 1.0 12.3 ( 1.1

5.0 ( 1.1 4.2 ( 1.0 3.6 ( 1.2 8.7 ( 1.5

0.29 ( 0.02 0.32 ( 0.03 0.27 ( 0.03 0.25 ( 0.03

Nafion LS surface coverage Γ/ Dapp/ films Ru(bpy)32+ (10-10) mol cm-2 (10-11) cm2 s-1 20 30 40 50

5.1 ( 0.5 8.5 ( 0.9 12.4 ( 1.2 14.4 ( 1.5

4.6 ( 0.5 7.2 ( 0.7 4.5 ( 0.4 8.3 ( 0.8

Cp/ mol dm-3 0.31 ( 0.03 0.32 ( 0.03 0.35 ( 0.04 0.31 ( 0.03

Ru(bpy)32+ occur at 0.38 V for FA+ and 1.07 V for Ru(bpy)32+ (vs Ag/AgCl) in 0.1 M NaCl. The CVs recorded under loading conditions with the redox mediator in the electrolyte solution (solid lines) reveal both higher peak currents and peak-peak separations (∆Ep ) Epf - Epb, where Epf and Epb are the peak potentials for the forward and back scans) that are close to those expect for a Nernstian process (∆Ep ) 67 mV for FA+ and 79 mV for Ru(bpy)32+).72 When the solution contained only supporting electrolyte, the peak currents decreased appreciably, along with the peak separation. For the cases in Figure 3, ∆Ep is ca. 36 and 33 mV for FA+ and Ru(bpy)32+, respectively. Note that the ∆Ep values were found to

be larger at a higher scan rate (up to 250 mV at 0.5 V s-1) whereas they decreased down to ca. 10 mV at a lower scan rate (1-5 mV s-1). In 0.1 M NaCl alone, the peak currents depended linearly on the scan rate at V < 20 and 100 mV s-1 for FA+ and Ru(bpy)32+, respectively, for a 20-layer film. This indicates thinlayer-like behavior, as shown in Figure 4(a, c) for both mediators. However, at higher sweep rates (in the range 0.1-0.5 V s-1 for FA+ and 0.25-5 V s-1 for Ru(bpy)32+) the current was found to scale more linearly with the square root of the scan rate, indicating a diffusion-controlled system (Figure 4b and d). The fact that the transition from thin layer to diffusion control occurred at a lower scan rate for FA+ indicates that this has a lower Dapp value than Ru(bpy)32+. Of course, diffusion control is operative when the Nafion layer is thicker than the concentration gradient of redox species in the film,73 so the transition is also filmthickness-dependent. Note that on the time scale of these measurements (approximately 30-40 min to run a complete set of CVs) the loss of either redox species, as evidenced from CVs recorded at the same scan rate at the beginning and end of the measurements, was typically less than 5%. It is well known that thick recast films of Nafion need long loading times (up to a few days in some cases) in order to fully preconcentrate electroactive species inside the films.21,41 In contrast, Nafion LS films loaded in millimolar solution of FA+ and Ru(bpy)32+ evidenced fast preconcentration: after only a (72) Bard, A. J and Faulkner L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.; Wiley: New York, 2001; Chapter 6, p 226. (73) Ugo, P.; Moretto, L. M. Electroanalysis 1995, 7, 1105.

10386 Langmuir, Vol. 22, No. 25, 2006

Bertoncello et al.

Figure 5. Experimental (O) and simulated (-) SECM approach curves recorded with a 25-µm-diameter Pt UME approaching ITO/Nafion LS film substrates loaded with FA+. The conditions were as follows: (a) 20-layer Nafion LS film, approach curves from top to bottom with 0.3, 0.38, 0.55, and 0.9 mM FA+; (b) 30-layer film, approach curves from top to bottom with 0.16, 0.23, 0.3, 0.45, and 0.9 mM FA+; (c) 40-layer film, approach curves from top to bottom with 0.13, 0.16, 0.25, 0.45, and 0.9 mM FA+; and (d) 50-layer film, approach curves from top to bottom with 0.06, 0.12, 0.25, 0.45, and 0.9 mM FA+. The simulations used a concentration of FA+ in the film of 0.35 M, diffusion coefficients in solution and within the film of D1 ) 6.0 × 10-6 cm2 s-1 and D2 ) 9.0 × 10-11 cm2 s-1, and film thicknesses of 23 (a), 35 (b), 48 (c), and 70 nm (d).

few minutes Nafion LS films were fully preconcentrated by FA+ and Ru(bpy)32+, and the CVs, measured as a function of time with the same redox species present in the electrolyte solution, usually tended toward a response that was time-invariant (Supporting Information, S2-S3). The surface coverage values (Γ/mol cm-2) of redox species in the loaded Nafion LS films on ITO electrodes were calculated from CVs displaying thin-layer characteristics with no redox species in the electrolyte solution using72

Γ)

Q nFA

(17)

where Q is the charge on the forward or reverse scan (C), n is the number of electrons transferred, F is the Faraday constant (96 486 C mol-1), and A is the geometric area of the electrode (cm2). With knowledge of the film thickness, these values could readily be expressed as concentrations. To estimate the diffusion coefficient, we plotted the anodic peak currents, Ipa, versus the square root of the sweep rate, V, and applied the Randles-Sevcik equation.72 The underlying assumption is that the redox process is reversible, but as already highlighted, the CVs showed a monotonic increase in ∆Ep with the scan rate, which we note will tend to lead to an underestimation of Dapp. The slope of these plots in the faster scan rate regime, combined with the film thickness (estimated by AFM) and the number of electroactive species (obtained by coulometric

integration of the anodic peak current under thin-layer conditions) allowed the evaluation of the apparent diffusion coefficient values within the coating (Supporting Information, S4).40 By keeping the geometric electrode area constant both at low (coulometric measure of m, the number of moles within the Nafion film) and high scan rates (evaluation of S, slope of Ipa vs V1/2), the electrode area was not required for the evaluation of the apparent diffusion coefficient values. Table 1 reports the values of the surface coverage, concentration of the electroactive species, and apparent diffusion coefficients extracted in this way. It is interesting that the apparent diffusion coefficients determined for FA+, which are relatively insensitive to the film thickness, are of the order of 5 × 10-12 cm2 s-1, which is about 2 orders of magnitude smaller than measured for Nafion recast films.21,34,41 In the case of Nafion LS films loaded in Ru(bpy)32+, the apparent diffusion coefficients are of the order of 5 × 10-11 cm2 s-1, which is 1 order of magnitude higher than for FA+. The higher diffusion coefficient for Ru(bpy)32+ is expected given that electron hopping essentially determines Dapp and physical diffusion (of FA+) is expected to be significantly impaired. SECM Studies. A family of approach curves was recorded separately for the two different redox species. The substrate comprised an ITO electrode with the desired number of Nafion layers: 20, 30, 40, or 50. The UME was held at constant potential to oxidize the mediator of interest (present in bulk solution at a series of defined values), and the steady-state diffusion-limited

Measurement of Apparent Diffusion Coefficients

Langmuir, Vol. 22, No. 25, 2006 10387

Figure 6. Experimental (O) and simulated (-) SECM approach curves recorded with a 25-µm-diameter Pt UME approaching ITO/Nafion LS film substrates loaded with Ru(bpy)32+. The conditions were as follows: (a) 20-layer Nafion LS film, approach curves from top to bottom with 0.28, 0.46, 0.77, 1.3, 1.8, 2.5, 7, and 10 mM Ru(bpy)32+; (b) 30-layer film, approach curves from top to bottom with 0.28, 0.46, 0.77, 1.1, 1.8, 3.6, 6, and 10 mM Ru(bpy)32+; (c) 40-layer film, approach curves from top to bottom with 0.28, 0.46, 0.7, 1.0, 1.8, 3.6, and 10 mM Ru(bpy)32+; and (d) 50-layer film, approach curves from top to bottom with 0.34, 0.46, 0.77, 1.3, 2.16, 3.6, 6, and 10 mM Ru(bpy)32+. The simulations assumed a concentration of Ru(bpy)32+ in the film of 0.30 M, whereas the diffusion coefficients in solution and within the film were, respectively, D1 ) 4.8 × 10-6 cm2 s-1 and D2 ) 1.3 × 10-10 cm2 s-1, with film thicknesses of 28 (a), 42 (b), 56 (c), and 70 nm (d).

current was measured as a function of the tip-to-substrate distance, d. Figures 5 and 6 show data for FA+ and Ru(bpy)32+, respectively. In each case, the tip current, i, has been normalized by the respective steady-state current in the bulk, ibulk, and the distance, d, has been normalized by the electrode radius, a. In the two-phase system described here, the current at the UME tip is essentially due to two contributions: one from the hindered diffusion of Red1 to the UME and the other from the reaction of the tip-generated species, Ox1, with Red2 at the film/ solution boundary. This, in turn, is governed by charge transport between the substrate electrode/film interface and the film/solution interface, with a thicker film introducing a more serious diffusional limitation. This limitation becomes more apparent as the film is “challenged” with a greater flux of Ox1 by increasing the concentration of Red1 in solution. Consider the approach curves recorded for the thickest Nafion films: 50 layers, which corresponds to a film thickness of ca. 70 nm. (See Figure 5d for FA+ and Figure 6d for Ru(bpy)32+.) As the concentration of redox species in solution increases, the normalized current undergoes a transition from essentially positive feedback (regeneration of Red1 at the film/solution interface) to one that is close to negative feedback (a relatively small contribution to the normalized UME current from the film

process). At low mediator concentration in solution, the film is able to sustain positive feedback, but as the solution concentration increases, the proportion of feedback from the film/solution interface decreases. Notice that with the same thickness film, the transition from positive to negative feedback occurs at lower solution mediator concentration for FA+ compared to that for Ru(bpy)32+, indicating that Dapp for FA+ is lower than Dapp for Ru(bpy)32+. The normalized currents at any distance are more enhanced, with the same concentration in bulk solution, the thinner the film. For example, compare Figure 5a (or Figure 6a) with Figure * 5d (or Figure 6d). For this situation, only at relatively high cRed 1 is it possible to observe a deviation from positive feedback. The cases corresponding to 30 and 40 Nafion layer films (Figure 5b and c; Figure 6b and c) are intermediate and follow the trend that has already been explained. * In the simulations, cRed was kept at the same value during all 2 of the experiments because the films were essentially saturated by the redox mediator for the range of mediator concentrations in solution. On the basis of the experimental data extracted from * CV measurements, the values of cRed used for the simulations 2 were 0.35 and 0.3 M for FA+ and Ru(bpy)32+, respectively. The

10388 Langmuir, Vol. 22, No. 25, 2006

diffusion coefficient in solution used for the simulations was DRu1) 4.8 (( 0.1) × 10-6 cm2 s-1,63 whereas for FA+ the value was calculated from the steady-state cyclic voltammogram (Supporting Information, S5) at the UME and found to be DFA1) 6.0 (( 0.1) × 10-6 cm2 s-1. The diffusion coefficient quoted for Ru(bpy)32+ was confirmed in a similar fashion. The best fit of the experimental data with the simulation was found for the following values of the diffusion coefficients inside the films: DFA2 ) 9.0 (( 1.0) × 10-11 cm2 s-1 and DRu2) 1.3 (( 0.2) × 10-10 cm2 s-1. It is encouraging that all of the approach curves in Figure 5 (or Figure 6) covering a wide range of solution mediator concentrations and film thicknesses could be fitted to the same (mediator-dependent) Dapp. The value of Dapp for Ru(bpy)32+ derived from SECM is ca. twice that measured by CV, whereas the value for FA+ from SECM is 1 order of magnitude larger than the CV values (see above). The main difference between the SECM measurement and CV is that the former is made under steady-state conditions, where there will be a steady-state ion distribution within the film, whereas CV is a transient measurement in which electron transfer occurs at the electrode/film interface and the accompanying ion-transfer process needed to maintain electroneutrality occurs at the film/solution interface. It is well established that ion ingress and egress at the film/solution interface can have a significant impact on the estimation of diffusion coefficients using CV and CA.26,53,74 That appears to be the case here, with a more severe impact on the measurements with FA+. Note that, by contrast to CV, both electron transfer and ion transfer occur at the film/solution interface in the SECM measurements.

Bertoncello et al.

formed and loaded in either FA+ or Ru(bpy)32+, with resulting film thicknesses of 1.6 ( 0.2 and 1.7 ( 0.2 nm per Nafion layer for FA+ and Ru(bpy)32+, respectively. The concentration of redox species within the films and the apparent diffusion coefficients were calculated using standard cyclic voltammetry measurements. However, for the ultrathin films investigated it is necessary to perform such measurements at relatively high scan rates (to ensure semi-infinite linear diffusion), where the CVs do not necessarily conform to classical reversible behavior. A novel SECM approach to the measurement of the diffusion coefficient of redox species in the films that allows apparent diffusion coefficients to be determined under steady-state conditions has been developed. The values for the diffusion coefficients for both FA+ and Ru(bpy)32+ found by this technique are higher than those derived from CV and are considered to be more reliable. In general, the Dapp values for these films are about 1 order of magnitude lower than the values reported in the literature relating to recast Nafion films. This can be attributed to the compactness and the high degree of condensation of the LS films. However, these ultrathin films show excellent retention of redox moieties and fast response times, opening up new possibilities for electrocatalysis and sensor applications.

Conclusions

Acknowledgment. We are grateful to Dr. N. R. Wilson (Department of Physics, University of Warwick) for the AFM images. This research was supported by a Marie Curie IntraEuropean Fellowships (MEIF-CT-2005-515356, to P.B.) and (MEIF-CT-2004-501300, to I.C.) within the 6th European Community Framework Programme. F.L. acknowledges a Dorothy Hodgkin Postgraduate Award from the U.K. Government.

It has been demonstrated that ultrathin Nafion films can be formed on solid supports using Langmuir-Schaefer techniques; the morphology and the thickness of the films have been estimated using AFM. We have found that relatively uniform films can be

Supporting Information Available: Pressure/area isotherm of Nafion, CVs of Nafion LS films with mediators, and evaluation of apparent diffusion coefficients. This material is available free of charge via the Internet at http://pubs.acs.org.

(74) Shu, C. F.; Anson, F. C. J. Am. Chem. Soc. 1990, 112, 9227.

LA061214I