VI) Charge Transfer: Three

Stuart Licht*, and Chanaka De Alwis. Department of Chemistry, University of Massachusetts, Boston, Massachusetts 02125. J. Phys. Chem. B , 2006, 110 (...
0 downloads 0 Views 347KB Size
12394

J. Phys. Chem. B 2006, 110, 12394-12403

Conductive-Matrix-Mediated Alkaline Fe(III/VI) Charge Transfer: Three-Electron Storage, Reversible Super-Iron Thin Film Cathodes Stuart Licht* and Chanaka De Alwis Department of Chemistry, UniVersity of Massachusetts, Boston, Massachusetts 02125 ReceiVed: NoVember 15, 2005; In Final Form: April 3, 2006

An extended conductive matrix facilitates a 100-fold enhancement in charge storage for reversible Fe(III/VI) super-iron thin films. These films were deposited, by electrochemical reduction of Na2FeO4, with an intrinsic high capacity 3 e- cathodic storage of 485 mAh g-1. Whereas 3 nm Fe(III/VI) films exhibited a high degree of reversibility (throughout 100 charge/discharge cycles), thicker films had been increasingly passive toward the Fe(VI) charge transfer. Films were alternatively deposited on either smooth or on extended conductive matrixes composed of high-surface-area Pt, Ti, and Au and probed galvanostatically and via cyclic voltammetry. A 100 nm Fe(VI) cathode, on the extended conductive matrixes, sustained 100-200 reversible three-electrode charge/discharge cycles, and a 19 nm thin film cathode sustained 500 such cycles. With a metal hydride anode, full cell storage was probed, and a 250 nm super-iron film cathode film sustained 40 charge/discharge cycles, and a 25 nm film was reversible throughout 300 cycles. Fe(VI) salts exhibit higher cathodic capacity and environmental advantages, and the films are of relevance toward the next generation charge storage chemistry for reversible cathodes.

Introduction Hexavalent Fe(VI) ferrate salts, or “super-irons”, have been introduced as a novel series of charge storage salts.1 These salts exhibit up to three electrons of charge storage occurring at a single, electropositive cathodic potential. Whereas, facile, primary charge transfer has been extensively demonstrated for these salts, reversible charge transfer has been problematic. In our recent communication, we showed that ultrathin (nanometerthick) Fe(III/VI) films, do exhibit extended quasireversibility, and demonstrated that 3-nm-thick super-iron films can sustain over 100 charge/discharge cycles.2 However, thicker films were not rechargeable due to the irreversible buildup of passivating (resistive) Fe(III) oxide, formed during film reduction, as illustrated in Scheme 1. Hence, thicker films had been increasingly irreversible, and either a 100 nm or a 1000 nm super-iron film had passivated after only 20 cycles or 2 cycles, respectively. The current work is an extension of this previous communication and probes the further enhancement of the conductive matrix. Such matrixes facilitate Fe(VI) cathodic charge transfer in substantially thicker cathodes with enhanced charge-storage capabilities. The development of rechargeable batteries with high energy density is of considerable interest, driven by the increasing energy consumption needs of portable electronics, lowemission vehicles, and medical devices.3 In recent years, intensive research has probed development of high-performance cathode materials.4-7 Nevertheless, contemporary aqueous batteries continue to use the traditional cathode materials MnO2, PbO2, and NiOOH. These dominate the consumer market despite issues such as limited rechargeability, environmental safety, and insufficient energy storage. Less common cathodes, such as * Corresponding author. E-mail: [email protected].

silver oxide, polypyrrole, oxygen, and metal sulfide cathodes, have limited applications due to cost and operation conditions.3 The known super-iron-based cathodes (including those utilizing M2FeO4 and M′FeO4, where M ) Li+, Na+, K+, Cs+, Ag+ and M′ ) Sr2+, Ba2+, and mixed cation salts) can be discharged in alkali media, in accordance with the generalized reactions:

FeO42- + 5/2H2O + 3 e- f 1/2Fe2O3 + 5OH-; E ) 0.6 V vs SHE (1) or

FeO42- + 3H2O + 3 e- f FeOOH + 5OH-; E ) 0.6 V vs SHE (2) The discharge ferric product tends to be amorphous. The specific Fe(III) formed and precise value of the redox potential will vary with the depth of discharge, dehydration, and the cation content of the super-iron salt and the alkali electrolyte. For example, in an alkaline sodium electrolyte, sodium ferrite may be a product:

Na2FeO4 + 2H2O + 3 e- f 1/2Na2Fe2O4 + NaOH + 3OH(3) The three-electron alkali discharge of the Fe(VI/III) cathodic redox couple occurs at approximately 0.25 V higher rest potential compared with that of the conventional MnO2 cathode and is characterized by a rest potential comparable to the NiOOH cathode widely used in secondary-reversible alkaline batteries, each of which undergo the equivalent of a one-electron

10.1021/jp0566055 CCC: $33.50 © 2006 American Chemical Society Published on Web 06/02/2006

Alkaline Fe(III/VI) Charge Transfer

J. Phys. Chem. B, Vol. 110, No. 25, 2006 12395

SCHEME 1: Representation of Facile Charge Transfer, or Passivation, in Fe(III/VI) Ferrate Films

Left side: Reversible Fe(III/VI) charge transfer in a ferrate film without a passivating layer. Right side: The buildup of resistive Fe(III) (depicted by shade), situated between outer Fe(VI) and the cathode current collector, can impede thicker ferrate film charge transfer.

discharge, in accord with the generalized respective reactions:

MnO2 + 1/2H2O + e- f 1/2Mn2O3 + OH-; E ) 0.35 V vs SHE (4) or

MnO2 + H2O + e- f MnOOH + OH-; E ) 0.35V vs SHE (5) and

2NiOOH + 2H2O + 2e- f 2Ni(OH)2 + 2OH-; E ) 0.55V vs SHE (6) The super-iron salts can exhibit storage capacity and environmental safety advantages over traditional cathode materials. The theoretical storage capacity of such salts, including Li2FeO4 (601 mAh g-1), Na2FeO4 (485 mAh g-1), K2FeO4 (406 mAh g-1), and BaFeO4 (313 mAh g-1), is higher than the most widely used primary- (MnO2 at 308 mAh g-1) and secondaryreversible (NiOOH at 290 mAh g-1) cathode materials in alkaline batteries. We reported a class of electrochemical charge storage superiron batteries utilizing several Fe(VI) salts.1,8-10 The high energy and power capacity of these batteries under various conditions have been demonstrated.11-15 In contrast to the advantageous characteristics of super-iron batteries, the poor reversible charge transfer has remained a challenge to be addressed. The formation of Fe2O3 has a detrimental effect on the reversible charge transfer due to conductivity constraints of ferric salts.12,14 The use of redox and electronic mediators have improved the extent of the primary Fe(VI) discharge, but have not improved the reversible charge transfer.16,17 Ultrathin 3 nm Fe(VI) films exhibit quasireversibility.2 Whereas these films utilize the Fe(VI) redox couple with an intrinsic high charge storage, the storage capacity per unit surface area is small due to the nanometer-thick level of the ultrathin film surface coverage. In this study, physical chemical probes of this charge transfer leads to a 2 orders of magnitude increase in quasireversible Fe(VI) charge storage per unit surface area. Experimental Section Chemicals and Materials. K2FeO4 was synthesized as described previously.18 NaOH (Acros Organics, purum, pellets), H2PtCl6 (Across Organics, metal basis 99.9%, 38-40% Pt, and Alfa Aesar, metal basis 99.95%, 37.5% Pt), Na2PtCl6 (Alfa Aesar, metal basis 99.99%, 42.4% Pt), HAuCl4 (Alfa Aesar, metal basis 99.99%, 49.94% Au), HClO4 (VWR, 0.1 N in

water), Nafion 350 (Aldrich), and Ti foil (Aldrich, thickness 0.127 mm, purity 99.7%) were purchased and used as received. All solutions were prepared from triply deionized water (Barnstead E-Pure). Instruments. Electrodeposition and cyclic voltammetry investigations were carried out by using an AFCBP1 bipotentiostat (Pine Instrument Company). Three- or two-electrode charge/discharge behavior was respectively investigated by using an AUTOLAB PGSTAT30 potentiostat-galvanostat (EcoChemie), or an Arbin Instruments BT4+ system. Substrate Pretreatment. Prior to deposition of platinum, platinum substrates were polished using aluminum oxide cloth (600 grit), etched in aqua regia (HCl/HNO3 (3:1)) for 10-20 min, sonicated in distilled water for 20 min, and then electrochemically cleaned by cycling between -0.2 V and -1.5 V vs Ag/AgCl for 50 cycles at a scan rate of 500 mV s-1 in 1 M H2SO4. Titanium substrates were polished with 320, and then 600, grit aluminum cloth, then sonicated in 6 M HCl, followed by a deionized water rinse prior to other treatments. Platinization Methodologies. On platinum substrates, Pt was potentiostatically deposited in a three-electrode cell at 0.2 V vs Ag/AgCl from aqueous 0.1 M H2PtCl6. The working electrode was a Pt foil with an exposed geometrical area of 8 cm2, and circular Pt foil was used as a counterelectrode. Pt was deposited onto Ti in a similar manner, but the deposition potential was -0.1 V vs Ag/AgCl, and a more concentrated, more conductive, 0.2 M solution of Na2PtCl6 in 0.1 M HClO4 was used to obtain higher coverage (>5 mg cm-2) of Pt deposit. Gold and platinum were codeposited onto Ti from a mixed solution of H2PtCl6 and HAuCl4 (0.1:0.1 M) in 0.1 M HClO4, again at -0.1 V vs Ag/ AgCl. Preparation of Super-Iron Cathode Films. Pure, stable K2FeO4 salts are more readily synthesized than Na2FeO4.13 Superiron films were electrodeposited from 30 mM K2FeO4 as dissolved in 10 M NaOH. This was chosen as the electrolyte due to the high solubility of super-iron salts in NaOH (K2FeO4 is highly insoluble in concentrated KOH electrolytes). This (millimolar level) K2FeO4/(molar level) NaOH, is effectively equivalent to an Na2FeO4/NaOH electrolyte containing 0.3% potassium. Electrodeposition was conducted in a cell formed from clamped, alkaline-resistant polypropylene squares. One square contained a cylindrical well, machined through the square. A well was formed by covering the horizontal bottom of the hole with the substrate and clamping the second square below the substrate. The 30 mM K2FeO4 10 M NaOH electrodeposition electrolyte was added to the well, covering the substrate, a working electrode with an exposed geometric area of 4 cm2 of either Pt or Ti. A nickel counterelectrode was positioned just above the working electrode, and a Ag/AgCl

12396 J. Phys. Chem. B, Vol. 110, No. 25, 2006 reference electrode was immersed in the electrolyte. A deposition potential of 0.1 V vs Ag/AgCl was potentiostatically applied, which initiates ferrate deposition onto the platinum, or platinized platinum, substrates. A deposition potential of -0.01 V vs Ag/AgCl was required when platinized Ti was used as the substrate to obtain a durable super-iron film. Prior to use, the film electrode was cleaned with a 10 M NaOH (super-ironfree) solution. Half-Cell and Full-Cell Charge/Discharge Characterization. The thin film half-cell charge/discharge behavior was investigated using a cell modified from that described above by the addition of a cation-selective membrane to isolate the cathodic and anodic cell compartments. After deposition and washing, the film electrode is returned to the sandwiched electrochemical cell, sealed with a Nafion 350 proton-exchange membrane, above which 10 mL 10 M NaOH is added to the membrane-isolated counterelectrode compartment. To probe extended duration half-cell measurements in concentrated solution, a saturated calomel electrode (SCE) with saturated KCl internal electrolyte was used as the reference electrode in lieu of a Ag/AgCl electrode that contains only 1 M KCl internal electrolyte. In full-cell studies, the counterelectrode and reference electrodes were replaced with a metal hydride anode removed from 40 mAh coin cells purchased from Powerstream, NY. The current densities and film thicknesses are reported with respect to the macroscopic morphology (the visual geometric surface area). Values of the microscopic surface morphology will vary with the measurement technique. A consistent technique was used throughout to determine the hydrogeneffective, or electroactive, surface area as the geometric surface area normalized electroactivity. This electroactivity was measured as the hydrogen desorption charge, assuming that the oxidation of one monolayer of hydrogen adatoms on Pt surface is associated with a charge of 210 mC cm-2.19,20 The normalized electroactivity was then calculated as the ratio of this measured electroactivity to the visual geometric area. At the potentials employed during platinization, electrodeposition substantially dominates, permitting the current efficiency to be assumed as 100% in accord with:

PtCl62- + 4e- f Pt + 6Cl-; E(Pt(IV/II)) ) 0. 76 V; E(Pt(II/0)) ) 0.68 V vs SHE (7) The absolute number of coulombs in the deposition process, Q(coulombs), is quantitatively measured and then, for convenience, converted to equivalents in accord with eq 7 and then mass-deposited platinum. Similarly, during ferrate film formation, at the potentials employed, electrodeposition substantially dominates, permitting the current efficiency to be assumed as 100%. Hence, the intrinsic capacity of the super-iron films determined by integrating the current-time response curve, Q(C ) coulombs ) ampere seconds), provides a quantitative measure of the intrinsic capacity of the super-iron films and, for convenience, a relative (not absolute) measure of the film thickness. The relative comparison of film thicknesses, T, is quantitative for all compared electrodes. One measure of this relative thickness is to use Na2Fe2O4 as the principal ferric product, calculated in accord with eq 3 as:

T ) [Q/3F]*( 1/2FW)/area*d; Na2Fe2O4: FW ) 221.7 g mol-1, d ) 3.05 g cm-3 T ) 9.1 × 107 nm cm2 mol-1*Q/(F*area(cm2)); F ) 96 485 C mol-1 (8)

Licht and De Alwis

Figure 1. Quasireversible alkaline charge-transfer behavior of a thin ferrate film on Pt. The working electrode in this three electrode cell configuration is a 3 nm ferrate film on smooth Pt in 10 M NaOH. Charge/discharge cycles 1, 10, and 25 are indicated. Each cycle cycle consist of (A) galvanostatic charge at 0.5 mA cm-2 for 6 s, followed by (B) discharge at 0.005 mA cm-2 for 480 s or (C) discharge at 0.05 mA cm-2 for 48 s.

Figure 2. Irreversible alkaline charge-transfer behavior of a thicker (19 nm) ferrate film on smooth Pt. Cell configuration and electrolyte are as described for the thinner (3 nm) film in Figure 1. Charge/ discharge cycles 1 and 20 are indicated. Each cycle consists of galvanostatic charge at 0.5 mA cm-2 for 48 s, followed by discharge at 0.005 mA cm-2 for 3200 s.

Results and Discussion Passivation of Fe(III/VI) Charge Transfer on Smooth Platinum. Figures 1 and 2 expand on our previous results in which ultrathin (nanometer-thick) Fe(III/VI) films deposited on smooth Pt surfaces exhibit three-electron quasireversibility, whereas thicker (10s of nanometers and thicker) films are not reversible. Fe(III/VI) films of 3 and 19 nm thickness are compared, which respectively contain 10.4 and 69 nanomoles of Fe per cm2, capable of storing either 3 or 20 mC of intrinsic capacity per cm2 (based on three electrons of storage per Fe(III/VI) center). As evident by comparing the first two figures, a 3 nm Fe(VI) film exhibits reversible behavior throughout 20 galvanostatic charge/discharge cycles, whereas a 19 nm film rapidly passivates. Specifically, each film is repeatedly subject to a 0.5 mA cm-2 galvanostatic charge, followed by a deep 0.005 mA cm-2 galvanostatic discharge (to 80% of the Fe(III/ VI) film theoretical, intrinsic 3 e- capacity). The 3 nm film smoothly approaches a plateau discharge potential between 0.2 and 0.3 V (Figure 1B). This contrasts with the discharge behavior of a 19 nm super-iron film in (Figure 2B) in which this discharge plateau is only sustained during the initial portion

Alkaline Fe(III/VI) Charge Transfer

J. Phys. Chem. B, Vol. 110, No. 25, 2006 12397 TABLE 1: Normalized Electroactivity of Platinum or Platinized Platinum Surfacesa

a

Figure 3. Cyclic voltammetry measured in 10 M NaOH electrolyte at a 20 mV s-1 sweep rate (A) at a (polished) Pt electrode, (B) at a Pt electrode with increasing levels of electrodeposited Pt, either (a) 0.8, (b) 1.9, (c) 3.2, (d) 7.2, or (e) 10.4 (mg Pt cm-2).

of the discharge, after which the potential begins to sharply decline. A stable plateau between 0.38 and 0.2 V and a rapid polarization decay to -0.2 V was observed on the discharge of the 19 nm film. The plateau discharge potential was sustained to approximately 75% of the full capacity during the first cycle and diminished to only 25% of this capcity within 20 cycles. Unlike the 19 nm film, the 3 nm ferrate film will also evidence sustained charge-transfer reversibility at higher cathodic currents. Hence, as presented in Figure 1C at an order of magnitude higher discharge rate of 0.05 mA cm-2, the 3 nm Fe(III/VI) film also sustains repeated, deep discharge. Compared to the 0.005 mA cm-2 cathodic cycling in Figure 1B, by the 25th cycle, the polarization losses are higher at this higher current density. For example, in this last cycle, at 80% depth of discharge, the potential diminishes to 0.13 V vs SCE at 0.05 mA cm-2, whereas the discharge potential is 90 mV higher at 0.005 mA cm-2. Extension of Platinum Surface Morphology. Figure 3A presents the 20 mVs-1 cyclic voltammetry of a smooth platinum electrode in 10 M NaOH. In agreement with previous studies,19,21 the peaks in the cyclic voltammogram (CV) potential region from -0.95 V < E < -0.65 V are associated with the hydrogen adsorption/desorption process, while the increasing anodic slope starting at -0.35 V is associated with the formation of adsorbed oxygen or nonstoichiometric surface oxides. At potentials higher than +0.5 V Ag/AgCl, oxygen evolution commences and becomes significant at the 0.6 V vs Ag/AgCl sweep limit. On the reverse negative scan, the peak in the potential region -0.65 V < E < -0.10 V and centered at -0.4 V is associated with the reduction of platinum oxides. This “smooth” platinum surface was prepared using the mechanical and chemical polishing techniques described in the Experimental Section. The effective surface area was evaluated by integrating the area under the hydrogen desorption curve. A factor of 2.5 was obtained for chemically treated smooth platinum. A further increase of 8% in normalized electroactivity to 2.7 was observed for electrochemically treated surface. These values are in good agreement with previously reported values in which the reported effective electroactive surface area for smooth platinum surfaces lies between 1.5 and 3.22-24

Pt load (mg cm-2)

normalized electroactivity

0 0.8 1.9 3.2 7.2 10.4

2.7 80 207 320 700 990

Platinum loading was electrodeposited at 0.2 V vs Ag/AgCl.

Cyclic voltammetry at a Pt electrode with various degrees of Pt loading (electrodeposited platinum) is presented in Figure 3B. As the degree of platinization increased, the reduction center of platinum oxide tends to be shifted negatively. Also, an increase in oxides formation and shrinkage in the double-layer region was evidenced by the width of the potential window between platinum oxide reduction peak and the hydrogen adsorption peaks. Most evident is the expected orders of magnitude increase of overall current densities compared with that of the smooth platinum surface in Figure 3A. Hydrogen adsorption and desorption peaks broadened as the amount of Pt deposits increased. Because of this broadening, although significant trends are evident, the hydrogen-desorption-based electroactive surface area measurement is approximate. These effective surface areas obtained with various quantities of Pt deposited are summarized in Table 1. The effective surface area of platinized platinum increased with the degree of platinization. Up to a ∼1000-fold increase in electroactivity was obtained with 10.4 mg cm-2 of Pt deposit. A linear trend for the variation of this electroactivity with the amount of Pt deposited was observed in accordance with:

normalized electroactivity ) 96*(Pt loading, mg cm-2) + 2.7 (9) Platinization effects reported in the literature for the normalized hydrogen electroactivity on platinized platinum surfaces vary over a wide range. Although extreme values of 3900, or even 20 000, have been reported,25 high values are generally in the range of 200-800.26-29 Discrepancies in normalized electroactivity can be attributed to surface morphologies obtained under different experimental conditions as well as differences in the methodology of estimation. In many cases, a comparison with reported effective surface area values cannot be made, as the amount of Pt loading was not reported. In the few cases where Pt loading was specified, the agreement with our measured values is reasonable. A comparable value of 232 has been reported with 2 mg cm-2 of Pt load,28 comparable to our measured surface enhancement of 207 at 1.9 mg cm-2 of Pt. A normalized electroactivity of 1750 has been reported with 9 mg cm-2 of Pt load,24 approximately two times higher than our measured value. Platinization of Titanium. In the search for platinum substrate alternatives, we also investigated platinized titanium substrates. Cyclic voltammetry at a polished, cleaned, untreated Ti electrode in 10 M NaOH is given in Figure 4A. As expected, a stable surface with lower faradic currents than that for smooth platinum, was observed throughout the potential region -1.00.8 V. Upon platinization of the titanium surface (0.85 Pt mg deposited per cm2 of Ti), the shape of the CV is similar to that obtained with a smooth platinum electrode, characterized with current peaks corresponding to the formation and oxidation of adsorbed hydrogen and oxygen (Figure 4B). The charge necessary for the formation of adsorbed hydrogen and oxygen

12398 J. Phys. Chem. B, Vol. 110, No. 25, 2006

Licht and De Alwis

Figure 4. Cyclic voltammetry in 10 M NaOH of (A) (polished) Ti electrode, or (B) platinized Ti electrode (0.85 mg cm-2 of Pt load at 0.1 V), or a (C) platinum-gold co-deposited Ti electrode (3.1 mg cm-2 of Pt and 3.3 mg cm-2 Au load at -0.1 V) measured at 20 mV s-1 sweep rate.

TABLE 2: Normalized Electroactivity of Titanium or Platinized Ti. pretreatment

amount of Pt (mg cm-2)

untreated untreated untreated untreated untreated untreated untreated annealed at 400 °C annealed at 800 °C etched in H2O2 etched in HNO3:HCl HNO3:HCl HNO3:HCl

0.90 0.85 0.85 0.85 0.75 1.9 3.75 3.5 1.6 0.6 0.5 0.75 4.5

a

deposition normalized potentials (V) electroactivity 0.2 0.1 0.0 -0.1 -0.5 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1

23 37 45 51 56 108 265 250 74 66 71 125 326

Platinum loading was electrodeposited at -0.1 V vs Ag/AgCl.

or for the formation of nonstoichiometric surface oxides can again be correlated to the normalized electroactivity and the extent of the platinum deposits. In addition, we find the electroactive surface morphology of platinized titanium varied with the initial titanium pretreatment, including variations of annealing and chemical etching. Various titanium substrate pretreatment methods were investigated, as a 10-fold increase in titanium surface area had been reported by pretreatment.30 Although we observe an enhancement, albeit smaller than reported, of normalized electroactivity upon annealing or chemical etching of the titanium, this enhancement was not evident when these pretreated surfaces were platinized. Hence, only a factor of 1.5-2.5 increase in normalized electroactivity was observed with pretreated Ti (without platinization), and no significant difference was observed when these surfaces were tested for hydrogen evolution, in comparison to untreated Ti when deposited with similar amounts of Pt. This confirms results of Khorozova et al., who investigated several catalytic reactions on platinized Ti and concluded that the surface area of the Pt deposit is practically independent of the pretreatment.31 The measured normalized electroactivity of a variety of platinized Ti electrodes is summarized in Table 2. In addition to Ti substrate pretreatment effects, the electroactivity of platinized Ti was also influenced by the Pt deposition potentials. Increases in the electroactive surface area were observed when platinization was carried out at more negative potentials. As seen in Table 2, a 38% increase in electroactivity

was obtained when Pt was deposited at -0.1 V than at 0.1 V. This is a departure from the previous observation for a Pt substrate in which platinization at lower potentials results in a decrease in normalized electroactivity,29 evidently indicating a different Pt nucleation mechanism on the Ti surface than on smooth Pt. At higher potentials, we note a longer time is required for deposition of the same amount of Pt than at lower potentials. It was observed that longer deposition times lead to formation of larger Pt particles,32 which would be consistent with the lower surface area and lower electroactivity than for platinization at more negative potential. Formation of Ti oxides at higher potentials could also hinder the Pt nucleation process and deplete the electroactive Pt particle distribution on the surface. In contrast, platinum is significantly more active than titanium toward hydrogen reduction, and the platinum substrate, more negative potentials would favor a high coverage of hydrogen, which competes and inhibits the Pt nucleation process, with a consequent increase in the particle size and a decrease in electroactive surface area. As in the case of smooth Pt, the normalized electroactivity of platinized titanium increased with the degree of platinization. A normalized electroactivity of 326 was obtained when 4.5 mg cm-2 of Pt loaded. This is less than the value of 435, estimated for smooth platinum with the same amount of Pt loading using eq 9. The observed electroactivity variation with either Pt or Ti substrates emphasizes the need for more detailed investigation of the platinization mechanism of Ti. Pt deposits larger than 7 mg cm-2 resulted in a powdery unstable surface. However, recently, Toyama and co-workers found good mechanical adherence of Pt on Ti was achieved by codepositing Pt with Au.33 A CV of gold-platinum codeposited Ti surface is given in Figure 4C. By codepositing Pt with Au, we achieved a stable surface containing up to 14 mg cm-2 of Pt with 15 mg cm-2 of Au deposit. Cyclic Voltammetry of Super-Iron Films. Previous cyclic voltammetry studies have shown a wide potential window of -0.24 to 0.21 V vs Ag/AgCl for Fe(VI/III) reduction.34-36 The existence of both Fe(V) and Fe(IV) has been reported by several researchers.37-39 However, recent in-situ Mo¨ssbauer investigation on Fe(VI) discharge products failed to identify any intermediate Fe(V) or Fe(IV) valence species.15 Interestingly, the Fe(VI/III) reduction potential was reported to be shifted 0.45 V negative on platinum plated with iron metal, compared to that of platinum in the absence of iron metal.36 This shift has been ascribed to a corrosion reaction of Fe(0) and FeO42- in the substrate. However, the formation of Fe(0) or Fe(II) does not occur under our conditions of >0.1 V in alkali media, as the reduction of Fe(III) to Fe(II) occurs only at potentials more negative than -0.2 V vs Ag/AgCl,40 and hence, the presence of iron metal is not relevant. Insight in super-iron film charge storage was gained through cyclic voltammetry. Figure 5 presents the typical CV of a superiron film exhibiting a partial degree of reversibility (in this case, a 75 nm super-iron film on a chemically etched, but not platinized, platinum substrate) measured from 0.6 to -0.5 V at 5 mV s-1. Foremost are broad reduction and oxidation bands throughout the CV. Superimposed on the broad 3 e- reduction in Figure 5, are small, more distinctive peaks occurring at 0.21 V (peak I), at -0.17 V (peak III), and and perhaps, 0.02 V (peak II). A broad oxidation region starting from 0.2-0.4 V centered at 0.3 V (peak IV) was also apparent. The inclination to attribute these peaks to the consecutive, stepwise, threeelectron reduction of Fe(VI) appears to be incorrect; more likely, they represent different forms of the Fe(III) reduction product.

Alkaline Fe(III/VI) Charge Transfer

Figure 5. Cyclic voltammetry a 75 nm ferrate film on a polished, aqua regia etched Pt electrode (without platinization). CVs are measured at 5 mV s-1 sweep rate in 10 M NaOH.

Figure 6. Cyclic voltammetry measured at a 5 mV s-1 sweep rate in 10 M NaOH of a 150 nm ferrate film on a polished, aqua regia etched Pt (A) over a complete +0.5 V sweep window and (B) when scan was reversed at either (a) 0.2 V, (b) 0.3 V, (c) 0.4 V, (d) 0.5 V, (e) 0.6 V, (f) 0.7 V vs Ag/AgCl.

In separate galvanostatic measurements, 75% or more (depending on the current density) of the available three-electron transfer in the reduction of Fe(VI) can occur at potentials greater than that needed to cathodically drive peak II or peak III, that is, at potentials g0.2 V vs Ag/AgCl (0.4 V vs SHE). As seen in Figure 1, galvanostatic reduction induces a three-electron very smooth, gradual decrease in potential, rather than a stepwise change in potential. This is indicative of a multiple electrontransfer event, occurring at a electropositive potential, which grows more negative due to gradually increasing polarization with build-up of the Fe(III) discharge product. In all cases, the peak II is small, and as seen in Figure 5, the distinctiveness of peak I disappears with an increasing number sweeps. However, peak III at -0.17 V does increase with the number of CV sweeps, evidence of the build-up of a less-reversible Fe(III) product generated during repeated Fe(VI) reduction. Figure 6 presents cyclic voltammetry of thicker, less reversible super-iron films, deposited on the mechanically and chemically roughened platinum substrate utilized for the CV in Figure 5. In Figure 6A, the reduction peak I is no longer evident, but the peak-II at ∼0.02V vs Ag/AgCl was more pronounced than those for for thinner films. Peak I in these thicker films was related to oxidation at potentials greater than 0.6 V vs Ag/AgCl, as a reduction peak I at ∼0.15 V was only observed to form for the thicker film when the electrode potential was initially held at this higher potentials. CV sweep rate variation studies on all

J. Phys. Chem. B, Vol. 110, No. 25, 2006 12399

Figure 7. Cathode discharge potential during cycling of (A) 19 nm and (B) 150 nm ferrate film on a platinized Pt substrate (2.6 and 10 mg cm-2 of Pt load, respectively) in a galvanostatic configuration in 10 M NaOH. The 19 nm film cycling consists of galvanostatic charge at 0.5 mA cm-2, followed by galvanostatic discharge at 0.05 mA cm-2 for 320 s. The 150 nm film cycling consists of galvanostatic charge at 0.5 mA cm-2, followed by galvanostatic discharge at 0.012 mA cm-2. Discharge cycle numbers are indicated on the figure.

reduction peaks indicate they were linked to surface confined processes rather than diffusion control (the currents do not linearly depend on the square root of scan rate41). Therefore, it is unlikely that reduction peaks I and III arise from the oxygen produced at higher potentials. In alternate alkaline systems consisting of passive iron electrodes containing Fe(II) sites, Fe(II) are the active sites for oxygen reduction, which commences at approximately -0.20 V vs Ag/AgCl.42-45 Nicol et al. observed that alkalinity increase resulted in the strong inhibition of oxygen reduction in the presence of Fe(VI).46 When the scan was reversed just after the second reduction wave (peak II), the broad oxidation wave peak IV, centered at ∼0.35 V, was visible and remained evident upon repeated cycling and corresponds to the broad oxidation wave centered at 0.02 V. Cyclic voltammetry was also studied in the anodic direction, followed by a reverse sweep starting at various potentials. The observed CVs are presented in Figure 6B. No reduction wave was observed up to 0.2 V. Peak II (∼0.02 V) reduction starts to appear when the scan was reversed at 0.3 V, and both this and peak III (∼-0.2 V) appeared when the scan was reversed at 0.5 V. From these observations, the oxidation corresponding to peak III starts between 0.4 and 0.5 V. The three-electron reduction of Fe(VI) can produce a variety of Fe(III) oxide and oxyhydroxide species such as (R,γFe2O3) and (R,γ,β,δ,FeOOH)47 via reactions 1 and 2. Also, a variety of cation-containing ferric salts such NaFeO2 are possible. Fe2O3 passivates and does not appear to be a reversible product of Fe(VI) reduction.12,15 In a recent communication, Zhang et al. also concludes that FeO42- is not generated from Fe2O3 by electrochemical methods in a highly alkaline environment.40 The ferric discharge product in the cells we observed exhibits a high degree of reversiblity that is amorphous. While certain ferric species can be excluded,12,15,40 a specific, reversible ferric product supporting the observed, quasireversible three-electron Fe(VI) cathodic chemistry is difficult to isolate and continues to be probed. Conductive Matrix Facilitated Fe(III/VI) Charge Transfer. A substantial improvement to sustain thick film charge transfer is observed when an extended conductive matrix was utilized as the film substrate. Figure 7A summarizes discharge behavior in repeated charge/discharge cycles of a 19 nm super-

12400 J. Phys. Chem. B, Vol. 110, No. 25, 2006

Figure 8. Cathode charging potential of a 19 nm film on a platinized (2.6 mg Pt cm-2) Pt substrate: (A) for 20 mC cm-2 of charge, during the 1st cycle, at various indicated mA cm-2 current densities, (B) for various indicated cycles at a fixed galvanostatic charge of 0.5 mA cm-2 for 40 s.

iron film on platinized platinum that contained 2.6 mg cm-2 of Pt deposits (normalized electroactivity of 250). Compared to the film on smooth platinum presented in Figure 6, the film was able to sustain a substantially higher discharge current (0.05 compared to 0.005 mA cm-2) as well as significantly higher reversibility (over 500 discharge cycles, compared to only 20 on smooth platinum), without onset of significant passivation. As summarized in Figure 7A, polarization losses actually improved during the first 70 cycles (with discharge potential improving ∼80 mV); after this, polarization losses increased, but throughout, cycles 2-300 remained less than in the first discharge cycle. Figure 8 presents the charging (as opposed to discharge behavior) of a 19 nm Fe(III/VI) film on the extended, conductive matrix. Figure 8A compares the measured charging potentials over a range of applied galvanostatic oxidation current densities. As reflected in Figure 8A, charging potentials during the initial stages of charging do not exhibit simple trends and were observed to be significantly effected by cycle number, thickness, and current density. In the latter stages of charging, a consistent increase of overpotential was observed with increasing charging current density. An order of magnitude increase in the relative charging current, to 0.25 mA cm-2, generates only approximately ∼40 mV charging overpotential. The next order of magnitude increase, to 2.5 mA cm-2, incurs a substantially larger overpotential (∼200 mV). For thicker super-iron films, a minimum current density of 0.5 mA cm-2 was useful to sufficiently regenerate at least 80% of the Fe(VI) discharge. To investigate both thin and thick super-iron films, a consistent charging current density of 0.5 mA cm-2 was utilized throughout the latter portion of this study to facilitate oxidation to Fe(VI) while sustaining relatively low charging overpotentials. The alkaline thermodynamic potential for oxygen evolution occurs at 0.16 V vs SCE, in accord with:

4OH- f O2 + 2H2O + 4 e-;

E ) 0.40 V vs SHE (10)

FeO2- species in NaOH have been observed to diminish the oxygen evolution potential by 0.1 V.36,48 However, at room temperature, platinum exhibits a high overpotential to oxygen evolution, and below 0.5-0.6 V vs SCE, the rate of oxygen evolution is not significant,49 Similarly, no observable oxygen evolution was evident for charging the Fe(III/VI) films at these

Licht and De Alwis

Figure 9. Cathode discharge potential for the 1st cycle of either (a) 19 nm or (b) 150 nm ferrate films, each on a platinized Pt surface (with normalized electroactivity > 1000) at a current density 0.012 mA cm-2.

TABLE 3: Influence of Normalized Electroactivity and the Discharge Current Density on Cycle Life of a 50 nm Film normalized electroactivity

current density (mA cm-2)

DOD (%); DOD ) depth of 3e- discharge

cycles to polarization deactivation

550 725 725 990

0.012 0.025 0.012 0.025

75 80 80 80

100 100 250 160

potentials. As can be seen in Figure 8A, at charging currents e0.5 mA cm-2, the charging potential was generally less than 0.5 V vs SCE. The super-iron films remained intact even after 500 cycles. In Figure 8B, the limiting value of the charging potentials gradually increased with the increasing number of cycles, reaching a maximum of 0.95 V only at the end of 500 cycles. Both with and without the extended conductive matrix, thicker ferrate films exhibited greater overpotential (lower cathode potential) during discharge. This is presented in Figure 9 for films deposited on the highest electroactivity substrates utilized in this study. For these substrates, an electroactivity of 500 was first achieved by repeated polishing and etching in aqua regia over several days, and then 10 mg cm-2 of Pt was deposited, yielding a highest normalized electroactivity >1000. Cathode discharge profiles for the first cycle of either 19 nm or 150 nm ferrate films formed on these substrates were compared. As can be observed from the figure, the final discharge potential was 0.0 V vs SCE for the thicker film but 0.15 V higher for the thinner film. On this higher electroactivity substrate, cycling behavior was modestly improved compared to the 150 nm film in Figure 7B, in which fifty cycles were sustained on a conventionally etched, but also highly platinized, substrate (again with 10 mg cm-2 of Pt, but with a normalized electroactivity ) 990). Film ferrate cathods of 100 nm on the highest electroactivity substrates sustained 100-200 charge/discharge cycles. In this study, the reversibility of super-iron films was studied when discharged to the majority (75-80%) of the intrinsic capacity of the films, determined and constrained by the three electrons of charge that may be stored per Fe. The importance of an effective, enhanced conductive matrix substrate was evident in attempts to reversible-cycle thicker (greater than 20 nm) super-iron films, using more platinized (normalized electroactivity over 250) Pt substrates, and the results are summarized in Table 3. When a 50 nm super-iron film was formed on a normalized electroactivity ) 550 platinized Pt substrate,

Alkaline Fe(III/VI) Charge Transfer

J. Phys. Chem. B, Vol. 110, No. 25, 2006 12401

SCHEME 2: Representation of Partial (Left Side), and Full (Right Side) Alleviation of the Fe(III/VI) Passivation in a Ferrate Film through an Extended Conductive Matrix

the film passivated when discharged at a current density of 0.025 mA cm-2, but discharged effectively at a current density of 0.012 mA cm-2. In the latter case, the 50 nm film reversibly cycled 100 times before the onset of passivation. When a 50 nm super-iron film was deposited onto platinized Pt with normalized electroactivity of 725, the film could sustain 100 cycles at a higher discharge current density of 0.025 mA cm-2. Further enhancement in the conductive matrix also resulted in longer cycle life, and 160 reversible cycles were sustained when the film was deposited onto a surface with normalized electroactivity of 990 (discharged at a current density of 0.025 mA cm-2). Greater reversibility was again observed when discharging at a lower current density, and the film sustained 250 cycles at a discharge current density of 0.012 mA cm-2. The formation of passivating, irreversible Fe(III) centers is more likely in the case of thicker films. Poorly crystallized films have exhibited unstable discharge, whereas thin cathode films with smaller grain size were more stable,49 emphasizing the impact of morphology on cathodic charge transfer. Passivation is evident in increased charge and discharge overpotentials and in the inability to discharge to a significant fraction of the intrinsic three-electron Fe(IV) charge capacity. The facilitated super-iron charge transfer, upon platinization, as a result of the expanded conductive matrix to facilitate charge transfer, is represented in Scheme 2. Without direct contact with the substrate, the shaded Fe(III) centers in Scheme 1 had posed an impediment to charge transfer. This was partially (Scheme 2, left side) and fully alleviated (right side) by intimate contact with the enhanced conductive matrix, which maintains extended direct contact with the substrate. Titanium Substrates for Fe(III/VI) Charge Transfer. Platinized Ti substrates also can be used as effectively as platinized platinum. To effectively utilize super-iron films on the platinized Ti substrate, it was observed that, during the initial phases of the cycling (for cycles 2 through 10), an extended charging time, equivalent to 150% charge of the intrinsic capacity, was required otherwise the full discharge could not be accessed. Following this, the conventional charge (equivalent to 120% charge of the intrinsic capacity) was sufficient to sustain extended charge/discharge cycling. Initial morphological changes can occur in a high surface areas, stressed, and crystallographically disordered platinized Pt surface.50,51 The electronic changes associated with the formation of Ti oxides may initiate this process.48 Hence, this latter effect could arise from improved contact of the underlying Pt layer with cycling. The charge/discharge curves of a 50 nm film on 7.5 mg Pt cm-2 platinized Ti are given in Figure 10. The one drawback encountered with the platinized Ti substrate in this study is the

difficulty in obtaining larger Pt deposits, more than 7.5 mg cm-2, which resulted in an unstable powdery surface, and because of this, it was difficult to deposit more than a 70 nm super-iron film on platinized Ti. However, a substantial improvement in the stability and upper limit of the thickness of the film was observed when Pt-Au codeposited Ti surface was used as the substrate. A 300 nm super-iron film displayed a moderate cycle life of 20. Charge/discharge profiles are presented in Figure 11. Metal Hydride Anode Compatibility. Facilitated charge transfer, thicker, Fe(III/VI) films compatibility with metal hydride anodes was probed in conjunction with full-cell charge storage. Generally, in alkaline media, the metal hydride anode charge storage has been utilized in conjunction with the nickel oxyhydroxy cathode. It is characterized by the effective migration of hydrogen ions from the anode to the cathode during charge, and from cathode to anode during discharge, without a significant variation in the electrolyte volume. As with the NiOOH cathode, the mechanism of discharge of the MnO2 cathode is complex, characterized by subtle variations in discharge product as well as both faradaic and proton/insertion charge storage processes.52-54 Importantly, for effective charge storage, each of the nickel, manganese dioxide, and super-iron alkaline cathode charge storage mechanisms is characterized by consumption of little or no electrolyte during discharge with a zinc or metal hydride anodes.1 The half- and full-cell reactions in the charge storage process of the super-iron/metal hydride cell can be described in an analogous manner to the NiOOH/ metal hydride cell. During discharge, FeO42- was reduced in

Figure 10. Quasireversible alkaline charge-transfer behavior of a 50 nm ferrate film on a platinized Ti (7.5 mg cm-2 of Pt load). (A) Electrode film potential during galvanostatic charge at 0.5 mA cm-2. (B) Electrode film potential during galvanostatic discharge at 0.012 mA cm-2. Charge/discharge cycle numbers are indicated on the figure.

12402 J. Phys. Chem. B, Vol. 110, No. 25, 2006

Figure 11. Quasireversible alkaline charge-transfer behavior of a 300 nm ferrate film on a gold-platinum codeposited Ti substrate (14 mg cm-2 of Pt and 15 mg cm-2 of Au loads). (A) Electrode film potential during galvanostatic charge at 0.5 mA cm-2. (B) Electrode film potential during galvanostatic discharge at 0.012 mA cm-2. Charge/discharge cycle numbers are indicated on the figure.

the Na2FeO4 cathode, and the metal hydride (MHx) was oxidized to the metal (M) in accordance with:

MHx + xOH- f M + xH2O + x e-; E ) -0.82 V vs SHE (11) The process is reversed during the charge. Combined with eq 3 for the complete cell discharge, this provides the overall reaction:

Na2FeO4 + /xMHx f /2Na2Fe2O4 + /xM + NaOH + H2O; E ) 1.42 V (12) 3

1

3

Interestingly, super-iron films deposited on the extended, conductive matrix substrate were characterized by a significantly long cycle life in the full, metal hydride storage cell, compared with those measured in the half-cell configuration. In the cell, a 25 nm film Fe(III/VI) cathode, deposited on 10 mg cm-2 Pt platinized substrate, displayed high cell voltage. A discharge voltage of 1.2 V was sustained through the end of 300 cycles in the full cell and within 0.22 V of the eq 12 thermodynamic cell rest potential. Although the cycling of the cathode in the full cell was consistently relatively higher as with the half cell measurements, as the thickness of the film increased, the cycle life decreased. The observed charge/discharge profiles of a 150 nm Fe(III/VI) film in a super-iron metal hydride cell are presented in Figure 12. Consistent with the prior half-reaction measurements, the film was charged to 120% of its theoretical capacity at a constant current density of 0.5 mA cm-2 and subsequently discharged to 80% at a constant current density of 0.05 mA cm-2. As seen in the lower portion of the figure, discharge commenced at ∼1.3 V and decayed to 1.0 V at the end of 80% capacity discharge. After this initial cycle (through charge/discharge cycle 100), discharge commenced at a higher voltage of ∼1.5. The rate of voltage decay during the discharge increased after 100 discharge cycles, and as seen in the figure, by cycle 110 a diminished voltage of 0.6 V was observed at the end of 80% capacity discharge. A 250 nm film, capable of storing 264 mC of intrinsic capacity per cm2, sustained stable, reversible charge storage only for 40 cycles, but note that this level of reversibility was previously only observed in 2 orders of magnitude thinner Fe(III/VI) cathodes deposited on substrates without the extended conductive matrix. A 50% longer cycle life was achieved with MH anode in comparisons with cycle life of the half reaction,

Licht and De Alwis

Figure 12. Two-electrode rechargeable behavior of 150 nm ferrate cathode (deposited on 10 mg cm-2 Pt-platinized substrate) with a MH anode. (A) Full cell potential as measured during a galvanostatic charge at 0.5 mA cm-2. (B) Electrode film potential as measured during a galvanostatic discharge at 0.05 mA cm-2. Charge/discharge cycle numbers are indicated on the figure.

compared with the three-electrode cell that had utilized a nickel counterelectrode. This is indicative of greater compatibility of the Fe(III/VI) film cathode in this metal hydride anode cell. Note that the half-cell configuration utilized a large nickel sheet as the counterelectrode and is consistent with our prior observation that Ni(II) has a deleterious effect on Fe(VI) stability. Conclusions Whereas 3 nm Fe(III/VI) films exhibited a high degree of three-electron reversibility (throughout 100-200 charge/ discharge cycles), thicker films had been increasingly passive toward the Fe(VI) charge transfer. An extended conductive matrix facilitates a 2 orders of magnitude enhancement in charge storage for reversible Fe(III/VI) thin films. High-capacity (Fe(III/VI) super-iron films were electrochemically deposited by electrochemical reduction of Na2FeO4 with an intrinsic 3 ecathode storage of 485 mAh g-1. Films were alternatively deposited on either smooth conductive substrates or on extended conductive matrixes, composed of high-surface-area Pt, Pt on Ti, and finally, Pt and Au on Ti. The influence of the extended matrix on charge transfer was examined galvanostatically and via cyclic voltammetry. A 100 nm Fe(VI) cathode, electrodeposited on the extended conductive matrixes, sustained 100200 reversible three-electrode galvanostatic charge/discharge cycles, and a 19 nm thin film cathode sustained 500 such cycles. Full-cell storage (anode/cathode) was also probed. In conjunction with a metal hydride anode, a 250 nm super-iron film cathode film sustained 40 charge/discharge cycles, and a 25 nm film was reversible throughout 300 cycles. The 2 orders of magnitude increase, up to 250 nm, in the rechargeable Fe(III/VI) cathode thickness, is a substantial increase but remains small compared with conventional cathode thicknesses. However, the useful upper limit of Fe(III/VI) cathode thickness, demonstrated in this study, begins to approach the lower limits of thickness demonstrated in functional battery systems. For example, a 200 nm thin film vanadium oxide cathode has been used in a rechargeable battery employing a solid polymer electrolyte.55 Solid polymer electrolytes will not suffer evaporative losses, are flexible, and have mass production advantages, but their relatively low ionic conductivity necessitates the thin cell thin battery cross-section configuration. This study provides fundamental evidence that an appropriate Fe(III/VI) conductive lattice can significantly facilitate its

Alkaline Fe(III/VI) Charge Transfer reversible charge transfer. Further studies to understand, and circumvent, the polarization of the Fe(III) state will lead to further gains in the practical, rechargeable super-iron cathode thickness. Fe(VI) salts exhibit intrinsic charge capacity and environmental hazard advantages over prevailing cathode materials, and the super-iron thin films are of relevance toward the next-generation charge-storage chemistry for reversible cathodes. Acknowledgment. We thank Lan Yang for help with K2FeO4 synthesis, and the U.S. Department of Energy for partial support of this study. References and Notes (1) Licht, S.; Wang, B.; Ghosh, S. Science 1999, 285, 1039. (2) Licht, S.; Tel-Vered, R.; Chem. Commun. 2004, 2004, 628. (3) Beck, F.; Ruetschi, P. Electrochim. Acta 2000, 45, 2467. (4) Wang, B.; Bates, J. B.; Hart, F. X.; Sales, B. C.; Zuhr, R. A.; Robertson, J. D. J. Electrochem. Soc. 1996, 143, 3203. (5) Bates, J. B.; Gruzalski, G. R.; Dudney, N. J.; Luck, C. F.; Yu, X. Solid State Ionics 1994, 70-71, 619. (6) Perner, A.; Holl, K.; Ilic, D.; Wohlfahrt-Mehrens, M. Eur. J. Inorg. Chem. 2002, 5, 1108. (7) Yufit, V.; Freedman, K.; Nathan, M.; Burstein, L.; Golodnitsky, D.; Peled, E. Electrochim. Acta 2004, 50, 417. (8) Licht, S.; Wang, B.; Ghosh, S.; Li, J.; Naschitz, V. Electrochem. Commun. 1999, 1, 522. (9) Licht, S.; Wang, B.; Xu, G.; Li, J.; Naschitz, V. Electrochem. Commun. 1999, 1, 527. (10) Licht, S.; Wang, B. Electrochem. Solid State Lett. 2000, 3, 209. (11) Ayers, K. E.; White, N. C. J. Electrochem. Soc. 2005, 152, A467. (12) Walz, K. A.; Suyama, A. N.; Suyama, W. E.; Sene, J. J.; Zeltner, W. A.; Armacanqui, E. M.; Roszkowski, A. J.; Anderson, M. A. J. Power Sources 2004, 134, 318. (13) Licht, S.; Naschitz, V.; Rozen, D.; Halperin, N. J. Electrochem. Soc. 2004, 151, 1. (14) Licht, S.; Tel-Vered, R.; Halperin, L. J. Electrochem. Soc. 2004, 151, A31. (15) Ghosh, S.; Wen, W.; Urian, R. C.; Heath, C.; Srinivasamurthi, V.; Reiff, W. M.; Mukerjee, S.; Naschitz, V.; Licht, S. Electrochem. SolidState Lett. 2003, 6, A260. (16) Licht, S.; Naschitz, V.; Gosh, S. J. Phys. Chem. B 2002, 106, 5947. (17) Licht, S.; Gosh, S.; Naschitz, V.; Halperin, N.; Halperin, V. J. Phys. Chem. B 2001, 105, 11933. (18) Licht, S.; Naschitz, V.; Liu; B.; Gosh, S.; Halperin, N.; Halperin, V. Rozen, D. J. Power Sources 2001, 99, 7. (19) Sawyer, D. T.; Sobkowiak, A.; Roberts, J. L., Jr. Electrochemistry for Chemists; John Wiley & Sons: New York, 1995.

J. Phys. Chem. B, Vol. 110, No. 25, 2006 12403 (20) Trasatti, S.; Pettrii, O. A. Pure Appl. Chem. 1991, 63, 771. (21) Bielger, T. Aust. J. Chem. 1973, 26, 2571. (22) Bakos, I.; Hora´nyi, G. J. Electroanal. Chem. 1995, 397, 105. (23) Bielger, T. Aust. J. Chem. 1973, 26, 2587. (24) Woods, R. Electrochim. Acta 1968, 23, 1967. (25) Brodd, R. J.; Hackerman, N. J. Electrochem. Soc. 1957, 104, 704. (26) Will, P. G. J. Electrochem. Soc. 1963, 110, 45 (27) Joncich, M. J.; Hackerman, N. J. Electrochem. Soc. 1964, 111, 1286. (28) Mayell, J. S.; Langer, S. H. J. Electrochem. Soc. 1964, 111, 438. (29) Feltham, A. M.; Spiro, M. Chem. ReV. 1971, 71, 177. (30) Pleskov, Yu. V.; Evstefeeva, Yu. E.; Krotova, M. D.; Lim, P. Y.; Chu, S. S.; Ral’chenko, V. G.; Vlasov, I. I.; Kononenko, V. V.; Varnin, V. P.; Teremetskaya, I. G.; Shi, H. C. Russ. J. Electrochem. 2005, 41, 337. (31) Khorozova, E.; Iordanova, Z.; Shterev, G. Bulg. Nauch. Trud. PloVdiVski UniV. 1984, 22, 129. (32) Kokkinidis, G.; Stoychev, D.; Lazarov, V.; Papoutis, A.; Milchev, M. J. Electroanal. Chem. 2001, 511, 20. (33) Toyama, S.; Someya, M.; Takei, O.; Ohtake; T.; Usami, R.; Horikoshi, K.; Ikariyama, Y. Chem. Lett. 2001, 12, 160. (34) Licht, S.; Naschitz, V.; Halperin, L.; Halperin, N.; Lin, L.; Chen, J.; Ghosh, S.; Liu, B. J. Power Sources 2001, 101, 167. (35) De Koninck, M.; Belanger, D. Electrochim. Acta 2003, 48, 1435. (36) Bouzek, K.; Rousˇar, I.; Bergmann, H.; Hertwig, K. J. Electroanal. Chem. 1997, 425, 125. (37) Rush, J. D.; Bielski, B. H. J. Inorg. Chem. 1989, 28, 3947. (38) Sharma, V. K. Rad. Phys. Chem. 2002, 65, 349. (39) Zhang, C.-Z.; Liu, Z.; Wu, F.; Lin, L.-J.; Qi, F. Electrochem. Commun. 2004, 6, 1104. (40) Antony, H.; Legrand, L.; Marechal, L.; Perrin, S.; Dillmann, Ph.; Chausse, A. Electrochim. Acta 2005, 55, 745. (41) Southampton Electrochemistry Group. Instrumental Methods in Electrochemistry; Ellis Horwood: New York, 1985. (42) Calvo, E. J.; Schiffrin, D. J. J. Electroanal. Chem. 1988, 243, 171. (43) Vago, E. R.; Calvo, E. J. J. Electroanal. Chem. 1992, 339, 41. (44) Gojkovic, S. Lj.; Zecevic, S. K.; Drazic, D. M. Electrochim. Acta 1994, 39, 975. (45) McAlpine, N. S.; Fredlein, R. A. Aust. J. Chem. 1983, 36, 11. (46) Nicol, M. J.; Guresin, N. J. Appl. Electrochem. 2003, 33, 1017. (47) Larramona, G.; Gutierrez, C. J. Electrochem. Soc. 1989, 136, 2171. (48) Kamnev, A. A.; Ezhov; B. B. Electrochim. Acta 1992, 37, 607. (49) Licht, S.; Halperin, L.; Kalina, M.; Zidman, M.; Halperin, N. Chem. Commun. 2003, 2003, 3006. (50) Mentus; S. V. Electrochim. Acta 2005, 50, 3609. (51) Hu, C.-C.; Liu; K-Y. Electrochim. Acta 1999, 44, 2727. (52) Rho, Y. H.; Kanamura, K. J. Electrochem. Soc. 2004, 151, A106. (53) Im, D.; Manitharan, A.; Coffey, B. J. Electrochem. Soc. 2003, 150, A1651. (54) Mondoloni, C.; Laborde, M.; Rioux, J.; Andoni, E.; Le´vy-Cle´ment, C. J. Electrochem. Soc. 1992, 139, 954. (55) Huang, B.; Cook, C. C.; Mui, S.; Soo, P. P.; Staelin, D. H.; Mayes, A. M.; Sadoway, D. R. J. Power Sources 2001, 97-98, 674.