WO2001028025A1

WO2001028025A1 - Magnetically modified fuel cells

Info

Publication number: WO2001028025A1
Application number: PCT/US2000/028242
Authority: WO
Inventors: Johna Leddy; Lois Anne Zook; Sudath Amarasinghe; Drew Dunwoody; Hachull X. Chung; Catherine Spolar; Shelley D. Minteer
Original assignee: University Of Iowa Research Foundation
Priority date: 1999-10-14
Filing date: 2000-10-13
Publication date: 2001-04-19
Also published as: AU1198501A

Abstract

A magnetic fuel cell capable of effecting indirect reformation of an organic fuel, comprising a cathode, an anode, a separator disposed between the anode and the cathode; and a magnetic field source adapted to produce a magnetic field in at least a portion of at least one of said cathode and said anode.

Description

MAGNETICALLY MODIFIED FUEL CELLS

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates generally to an improved fuel cell apparatus, and methods of use thereof. In particular, the present invention relates an apparatus and methods of its use with respect to fuel cells having enhanced electrical current and power outputs that result from magnetic modification of one or both of the anode and cathode - the magnetic modification being effected in a preferred embodiment via a microstructured magnetic composite material that is disposed on the desired electrode(s). The improved fuel cell apparatus of the present invention is also highly resistant to electrode passivation.

2. Background of the Related Art

The use of low temperature fuel cells in industry and in the general economy has not reached even moderate acceptance because of hmitations in function. State-of-the-art low temperature fuel cells have low electrical current outputs and low power outputs. As a result of these shortcomings, enhanced function in conventional low temperature fuel cells is achieved by the use, for example, of pumps to increase the pressures of, and the flow rates of, fuel and oxygen feeds to the respective electrodes. Such uses of pumps only further limit the net electrical current and power outputs due to the parasitic power required to energize the pumps. Furthermore, such use of pumps also results in larger and often unwieldy structures for the fuel cells. In addition, the use of such larger fuel cells tends to defeat or at least greatly reduce applications in which portability is important. In addition, currently available fuel cells suffer from passivation, which, in effect, "kills" the majority of any electrical current and power output that otherwise could be realized. Passivation results, for example, when small amounts of carbon monoxide are produced as an organic fuel is oxidized at the anode and the resulting CO binds to the noble metal catalyst.

These limitations in current low temperature fuel cell technology point up the profound need for fuel cells with enhanced function. Any substantial improvement with respect to power and current output, as well as decrease in passivation effects, would therefore provide a basis for increased commercial use, and, potentially, universal acceptance both in the heavy industrial sector and in the general consumer sector of the economy. Fuel Cells in General. In its simplest form (see Figure 1), a fuel cell 1 comprises a cathode 12, an anode 10, and a separator 14 disposed between the anode and the cathode. When an oxidant such as O_z 18 is reduced at the cathode, and a fuel such as H, 16 is oxidized at the anode, then an electrical current 20 flows between the anode 10 and the cathode 12. An external resistive load 4 applies to the entity that harnesses the energy of the fuel cell via external circuit 2. One or more noble metal catalysts 8 and 6 are typically employed both on the cathode 12 and the anode 10, for example, platinum. The separator 14 may be made of an ionic membrane, such as Nafion®, which functions as a proton exchange membrane (PEM). Water 22 may be produced by such a fuel cell system. An encasement 3 is used to confine the contents of a fuel cell, and to isolate such contents from the surrounding environment.

The basic objective of a fuel cell is to allow a reaction between a fuel (e.g., hydrogen, methanol, ethanol, propanol, acetaldehyde, methane, etc.) and an oxidant (e.g., oxygen, hydrogen peroxide), which normally react spontaneously (and often violendy), to discharge in a controlled manner. By containing the fuel and the oxidant at separate electrodes, the discharge of the reaction is electrical, rather than thermal. A wire(s) connecting the electrodes delivers the current and voltage of the discharging system, thus providing power to drive an external device, such as an electric motor.

Fuel cells combine the best characteristics of a battery and a combustion engine. Similar to the combustion engine, they are not recharged electrically, and output power is realized as long as fuel is provided. Similar to the battery, fuel cells are electrical devices capable of providing power, and theoretically, are not subject to a combustion engine's Carnot limitations. The expansion and contraction of pistons limit heat engines to about 40% of their theoretical power efficiency, and about 25% as their practical efficiency under optimal conditions. In contrast, fuel cells approach 100% efficiency in theory, and have been demonstrated to operate at better than 90% efficiency in practice.

Proton Exchange Membrane (PEM) Fuel Cells. PEM fuel cells are fuel cells in which the separator is a proton exchange membrane. One commonly used PEM is Nafion®, which is a perfluorinated sulfonic acid resin. PEM fuel cells are examples of low temperature operating fuel cells because they typically operate at or below about 100°C. Most commonly, oxygen or atmospheric air serves as the oxidant, and hydrogen serves as the fuel. A fuel cell that runs on hydrogen and oxygen is designated an H₂/0₂ fuel cell; i.e., the fuel/oxidant convention.

The PEM fuel cell illustrated in Figure 1 employs hydrogen as a feed 16 to anode 10, and oxygen in air as a feed 18 to cathode 12. Those reactants decompose electrolytically to yield water 22 at the cathode. The hydrogen and oxygen are separated by a proton exchange membrane 14 (such as Nafion®) to prevent, among other things, thermal decomposition of the fuels at noble metal catalyst 6, 8. The reactions at cathode 12 and anode 10 can be summarized as follows:

Cathode 0₂ + 4H⁺ + 4e^" ^ 2H₂0 E°_cathode = 1.23V [slower reaction] (1) Anode 2H⁺+ 2e^" ^ H₂ E⁰^ = 0.00V [faster reaction] (2)

Net Reaction O, + 2H₂ <≠ 2H₂0 E°_cell = 1.23V (3)

However, a fuel cell typically runs under non-equilibrium conditions and is thus subject to kinetic limitations. Usually, the majority of the kinetic limitations are at the cathode 12: 0₂ + 4H⁺ + 4e- - 2H₂0 E°_cathode = 1.23V (4)

During the course of operation of a fuel cell, such as the one depicted in Figure 1, the cathode reaction increasingly kinetically limits performance as current demand increases, which is reflected in a drop in the fuel cell's voltage output; simultaneously, a second reaction path (the two-electron/two-proton reduction of oxygen to peroxide) becomes increasingly favored. This second reaction path consumes oxygen in two-electron steps with lower thermodynamic potential as follows:

O, + 2H⁺ + 2eV ≠ H₂0₂ E°_H202 = 0.68V (5)

The standard free energy of the reduction of oxygen to peroxide of Equation (5) is roughly 30% of the free energy available from the four-electron reduction of oxygen to water that is shown in Equation (4), and the current output from a fuel cell in which the two- electron reaction predominates is proportionately decreased due to the transfer of only two electrons, instead of four and the lower E⁰ _CL._π. This effect, especially when combined with the concomitantiy decreased maximum cell potential, yields a substantially lower fuel cell power output. The efficiency of the cathode reaction (i.e., the output of electrical current, voltage, and power, as well as the rate of reaction of Equation 5) in which the two-electron pathway predominates, or at least comprises a substantial proportion of the overall reaction, can be enhanced by increasing the concentration and/or pressure and flow rate of the feeds to the cathode 12 (i.e., protons and oxygen). The proton flux typically is not limiting, as the proton exchange membrane (e.g., Nafion®) readily provides an ample supply of protons to meet the demand of the cathode reaction(s). The flux of oxygen is increased (and the reaction is consequendy biased to favor the formation of water) by pressurizing the air feed to the cathode 12 to at least 2 to 10 atmospheres.

When large-scale commercialization of PEM fuel cell technology is contemplated, at least three major impediments stand out immediately. First, the kinetics for hydrogen oxidation in an H₂/0₂ fuel cell are very rapid compared to the H₂/0₂ kinetics of oxygen reduction. As is indicated above, to overcome the kinetic limitations of oxygen, the feed to cathode 12 is pressurized to at least roughly two times the pressure of the feed to anode 10. The resulting change in the concentration of oxygen at cathode 12 shifts the reaction toward the desired electrolysis product, which is water. When air is substituted for oxygen and the cathode feed is pressurized, a beneficial effect results, namely, a sweeping out of nitrogen, which, otherwise, tends to build up in cathode 12 and reduce the partial pressure of oxygen at cathode 12. One consequence of utilization of pressurization in this manner is that operation of the required pumps causes a parasitic power output loss of about 15%. Two other consequences are a significant increase in overall weight of the fuel cell, and a substantial increase in noise. In addition, the moving parts of the pumps increase the complexity and the number of failure mechanisms of the fuel cell system/ apparatus. Because of these disadvantages, the use of such pumps in portable applications often is not practical, and, at best, adds to inconvenience for the user.

A second impediment is that hydrogen is not the most convenient fuel because of its highly exothermic reactivity with oxygen, which can produce flames and/or explosion. One solution to this problem is to use indirect reformation of an organic fuel, for example, by passing the organic fuel over a hot copper/zinc catalyst, which yields hydrogen that then is fed to anode 10. On the other hand, direct reformation, in which an organic fuel is fed direcdv to anode 10, would provide greater efficiency for the fuel cell system; however, the problem remains of electrode passivation due to by-products such as carbon monoxide.

When direct reformation is used, a third impediment results secondarily from the direct feeding of organic fuel to anode 10. In this implementation, separator 14 tends to imbibe organic fuels, which then cross separator membrane 14 and pass to cathode 12, where there is a direct reaction with the oxidant, including assistance from the catalyst. Thus, the passage of the organic fuel (initially fed to anode 10) to cathode 12, via separator membrane 14, short circuits the flow of electrons through the external circuit, which reduces the electrical current and power outputs to the external circuit of the fuel cell. Furthermore, even in the case of hydrogen fuel cells (i.e., no organic fuel feed to anode 10), there are significant power losses due to a similar phenomenon, which results when too many protons carry water from anode 10 to cathode 12; such excess flow of protons and water causes anode 10 to dry and cathode 12 to flood, which is a phenomenon known as crossover. This phenomenon also occurs when organic fuels are oxidized in a fuel cell.

Magnetic Concepts Relating to Electron Transfer and Effects on Chemical Kinetics.

Magnetic field effects on chemical systems include effects on the rates of electron transfer (i.e., kinetic effects) in both homogeneous and heterogeneous systems; however, macroscopic thermodynamic effects generally are negligible. Kinetic effects may be manifest in various areas, including effects on reaction rates, reaction pathways, and distribution of products. The incidence of electron transfer reactions, in which electrons are transferred between molecules or ions, is ubiquitous throughout natural and technological systems, including biological energy production, ozone depletion, and technologies from photography through batteries, solar cells, fuel cells, and corrosion. Understanding processes by which the speeds or rates of electron transfer reactions may be affected/controlled is fundamentally important, since controlling rates of chemical reactions has important consequences in a practically unlimited array of applications, including reducing noxious loads to the environment, decreasing the consumption of energy, developing more efficient technologies, developing more efficient fuel cells, batteries, and solar cells that are environmentally friendly, etc.

Electron transfer reactions can be characterized as either homogeneous or heterogeneous. If the reaction occurs in a single phase (i.e., solid, liquid, gas, or plasma) between two ions or molecules, the reaction is a homogeneous electron transfer. On the other hand, if the reaction occurs at an interface between two chemically/physically (-ussimilar phases (e.g., at the interface between an electrode and the surrounding solution, at the interface between two dissimilar solutions that are mixed together - but before they are totally mixed to uniformity, at the interface between a charged membrane and the surrounding solution, etc.), then the reaction is a heterogeneous electron transfer when one of the molecules/ions is on one side of the interface, and the other reactant is either a molecule or ion on the opposite side of the interface or in/at the interface, or it is the interface itself. Uniform magnetic fields that are applied when a solution is placed between the poles of a magnet will have a negligible effect on the free energy of a typical chemical reaction. For example, even when the strongest of laboratory magnetic fields is applied (on the order of 10 Tesla, or 100k Gauss) at room temperature, the effect will be on the order of less than a Joule/mole, and more typically, less than 0.5 Joule/mole. However, although such macroscopic magnetic effects are neghgible as to effect on the free energy of a chemical reaction at room temperature, substantial microscopic effects may be realized, for example, when a chemical reaction occurs within 1 nm of a magnetic microparticle that is, for example, part of a magnetic composite applied to an electrode or other surface. In such microscopic applications, the magnetic field produced by a magnetic microparticle decreases over a distance x in proportion to x^"3. Thus, the field experienced by a molecule 1 nm from a magnetic microparticle can be on the order of 10²¹ times greater than the field experienced 1 cm from the same magnetic microparticle. However, such highly local magnetic effects can produce substantial effects when mass transport effects are present.

Magnetic fields affect chemical reactions through enhancement of spin polarization, which may be electron, nuclear, or electron-nuclear. To date, electron spin polarization effects have been studied in the laboratory (A.L. Buchachenko, 1976. "Magnetic effects in chemical reactions." Russ. Chem. Rev. 45: 375-390. N.J. Turro & B. Kraeuder, 1980. "Magnetic field and magnetic isotope effects in organic photochemical reactions. A novel probe of reaction mechanisms and a method for enrichment of magnetic isotopes." Accounts of Chemical Research 13: 369-377. U.E. Steiner & T. Ulrich, 1989. Chem. Rev. 89: 51. P.W.

Atkins & T.P. Lambert, 1975. Ann. Report of Progress in Chemistry 72A: 67. P. Atkins, 1976. Chemistry in Britain 12: 214. R.Z. Sagdeev, K.M. Salikhov, & Y.M. Monin, 1977. Russ. Chem. Rev. 46: 297, the contents of all of which are incorporated herein by reference); however, magnetic effects for electron-nuclear spin polarization have not received experimental attention. Nuclear spin polarization effects will be small, slow, and neghgible as compared to electron and electron-nuclear spin polarization, though significant nuclear spin polarization effects have been postulated to apply in proton transfer reactions. Electron spin polarization refers to polarization between unpaired electrons on two different radicals or radical centers. A radical pair is formed where the electron cloud of one species precesses around the vector of the applied field, and through interactions with the second unpaired electron, spin relaxations between high and low spin states are induced. A common example of electron spin polarization or spin relaxation is intersystem crossing, where, for example, a species with one unpaired electron (a doublet, D) interacts with a second doublet to form a complex with two unpaired electrons (a triplet, T) that yields products with no unpaired electrons (a singlet, S). Theory restricts rate enhancements for singlet/ triplet conversions to ninefold (see Turro & Kraeutler, 1980, above). The work of Turro and coworkers on electron transfer reactions in micelles is the classic example of magnetic fields inducing electron spin polarization to alter chemical kinetics (see also the references cited in Turro & Kraeuder, 1980). For photoinduced electron transfer reactions between organic radicals in solution, typical rate enhancements are less than 50% for magnetic fields of 1 to 8 Tesla. Recendy, the rate of photoinduced intramolecular electron spin polarization was shown to decrease in a magnetic field (T. Klumpp, et al., 1999. . Am.

Chem. Soc. 121: 1076, the contents of all of which are incorporated herein by reference).

Nuclear-nuclear spin polarization occurs between two nuclei in a magnetic field when the polarized nucleus on the first molecule polarizes the nucleus of the second molecule. No radicals are required, but one nucleus must be pre-polarized. Again, nuclear polarization effects are slow and small, especially when compared to electron and electron-nuclear spin polarization effects.

Electron-nuclear spin polarization occurs when the electron spin polarization on one species allows electronic currents generated by the precessing electron cloud to induce a secondary magnetic field at the nucleus of the second species. For two radicals, magnetic effects can occur through electron and electron-nuclear spin polarization; for a radical and a singlet, only electron-nuclear spin polarization is possible. Thus, electron-nuclear spin polarization allows a radical in a magnetic field to increase the electron exchange rate with a singlet. Electron-nuclear spin polarization is also known as electron-nuclear cross- relaxation and dynamic polarization. Electron-nuclear spin polarization causes hne broadening in nuclear magnetic resonance (NMR) and electron proton resonance (EPR) spectroscopy, and enhances signal intensity for NMR several hundred-fold (A. Carrington & A.D. McLaughlin, 1967. In: Introduction to Magnetic Resonance with Applications to Chemistry and Chemical Physics, pages 229-236. Harper and Row, N.Y., the contents of all of which are incorporated herein by reference).

Additional or alternative details, features, and/or other background technical aspects of fuel cells, including aspects of magnetic modification, are provided in United States Patents Nos. 5,786,040, 5,817,221, 5,871,625, 5,928,804, 5,981,095, and 6,001,248, which hereby are incorporated by reference in their respective entireties.

SUMMARY OF THE INVENTION

It is therefore an object of the invention to provide an improved fuel cell.

Another object of the fuel cell invention is to provide an improved PEM (proton exchange membrane) fuel cell.

Another object of the present magnetically modified fuel cell invention is to provide an improved electrolytic cell.

Another object of the present magnetically modified fuel cell invention is to provide a fuel cell that facilitates electron transfer in heterogeneous and in homogeneous systems/environments.

Another object of the present magnetically modified fuel cell invention is to provide an improved fuel cell with enhanced power output.

Another object of the present magnetically modified fuel cell invention is to provide an improved fuel cell with enhanced output of electrical current. Another object of the present magnetically modified fuel cell invention is to provide an improved fuel cell that is capable of direct reformation of fuels.

Another object of the present magnetically modified fuel cell invention is to provide an improved fuel cell that has improved resistance to passivation.

Another object of the present magnetically modified fuel cell invention is to provide an improved fuel cell that has the capability of altering the product distribution of a chemical reaction.

Another object of the present magnetically modified fuel cell invention is to provide an improved fuel cell that has the capability of effecting certain quantum mechanically spin forbidden or kinetically disfavored chemical reactions. Another object of the present magnetically modified fuel cell invention is to provide an improved fuel cell that has enhanced chemical synthetic capabilities, including effecting certain quantum mechanically spin forbidden or kinetically disfavored chemical reactions. One advantage of the present magnetically modified fuel cell invention is that it can enhance the flux of oxygen or other oxidant that is reduced at the cathode of a fuel cell.

Another advantage of the present magnetically modified fuel cell invention is that it can alter the product distribution of a chemical reaction.

Another advantage of the present magnetically modified fuel cell invention is that it can enhance the flux of a fuel that is oxidized at the anode of a fuel cell. Another advantage of the present magnetically modified fuel cell invention is that it can operate stably under conditions that would lead to passivation of the anode of a conventional nonmagnetic fuel cell.

Another advantage of the present magnetically modified fuel cell invention is that it can effect certain chemical reactions that otherwise are quantum mechanically forbidden or kinetically disfavored.

Another advantage of the present magnetically modified fuel cell invention is that, with magnetic modification of the cathode and operation at about 70°C, it can produce a maximum electrical current output that is at least about three times the maximum electrical current output of a conventional nonmagnetic fuel cell in operation at about 70°C. Another advantage of the present magnetically modified fuel cell invention is that, with magnetic modification of both the cathode and the anode and operation at about 70°C, it can produce a maximum electrical current output that is at least about four times the maximum electrical current output of a conventional nonmagnetic fuel cell in operation at about 70°C. Another advantage of the present magnetically modified fuel cell invention is that it can produce significant power and electrical current outputs under operation at room temperature (around 25°C), compared to a conventional nonmagnetic fuel cell operates very poorly at temperatures as low as room temperature.

Another advantage of the present magnetically modified fuel cell invention is that it can produce greater power and electrical current outputs under operation at room temperature (around 25°C) than the power and electrical current outputs of a conventional nonmagnetic fuel cell under operation at 70°C. Additional advantages, objects, and features of the present magnetically modified fuel cell invention will be set forth in part in the description which follows, and in part will become apparent to those having ordinary skill in the present art upon either examination of the following and/or practice of the present magnetically modified fuel cell invention. Many of the advantages, objects, and features of the present magnetically modified fuel cell invention may be realized and attained as particularly described in the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS

^• Figure 1 is a schematic illustrating basic components of an H₂/0₂ PEM fuel cell, including an anode and a cathode (each having a noble metal catalyst; here, platinum), and a separator (here, Nafion®).

Figure 2 shows the effects of temperature and degree of magnetic loading of cathodes with iron oxide magnets for fuel cells in plots of power versus potential. In addition, curves of power versus potential are shown for samarium cobalt magnets loaded onto cathodes at 0.13 mg/cm² at 70°C. Figure 3 shows curves of potential versus current for the same plots shown in Figure

2.

Figure 4 shows the effects of pressure (1-3 arm) and magnetic loading of cathodes with iron oxide magnets for H₂/0₂ fuel cells at 70°C. Potential versus current curves are shown for samarium cobalt loaded cathodes at a level of 0.13 mg/cm² at 70°C. Figure 5A shows curves of power versus potential for the same experiments shown in Figure 4. Figure 5B plots power density as a function of pressure.

Figure 6 shows the effects of temperature and magnetic loading of the cathode electrode on the power output of H₂/0₂ fuel cells.

Figure 7 shows the effect of humidification temperature on magnetically modified cathode electrodes for H,/0₂ fuel cells when the fuel cell is at 20°C.

Figure 8 shows the effects of temperature (30 to 70°C) and magnetic loading of cathodes with iron oxide magnets for H₂/air fuel cells in plots of potential versus current. In addition, results for samarium cobalt loaded cathodes at 0.13 mg/cm² are shown for 30 to 70°C. Figure 9A shows curves of power versus potential for H₂/air fuel cells under the conditions indicated in Figure 8. Figure 9B plots the data as power density versus temperature in °C.

Figure 10 shows the curves of potential versus current, and power versus potential, for magnetically loaded cathode electrodes at 70°C and for pressure from 1 to 3 atmospheres.

Figure 11 shows the effects of cathode flow rates for magnetically loaded cathode electrodes in curves of potential versus current, and power versus potential for H₂/air fuel cells at 70°C

Figure 12 shows the effects of magnetic loading and temperature (30-70°C) when both the cathode and anode of H₂/0₂ fuel cells operating at one atmosphere. The cathode magnetic loading is fixed at 0.2 mg/cm² and the anode load is varied.

Figure 13 shows curves of power versus potential for the cathode and anode magnetically modified electrodes of Figure 12.

Figure 14 shows curves of power versus potential for fuel cells operated at 25 to 70°C when both the cathode and the anode are loaded with different amounts of magnets.

Figure 15 shows curves of potential versus current for the same experiments indicated in Figure 14.

Figures 16A and 16B show curves of potential versus current, and power versus potential, when both cathode and anode are magnetically modified, and pressure is varied from 1 to 3 atmospheres.

Figures 17A and 17B contrast the effects of air (Figure 17A) versus 0₂ (Figure 17B) when both cathode and anode are magnetically modified.

Figure 18 shows the effects of passivation (use of synthetic reformate containing hydrogen and 100 ppm of carbon monoxide) on anodes that are magnetically modified with iron oxide magnets, versus on anodes that are not magnetically modified.

Figure 19 shows the effects on current density of use of synthetic reformate (hydrogen plus 100 ppm of carbon monoxide) when the anode contains platinum/ruthenium catalyst, but no iron oxide magnets.

Figure 20 shows the current density response of a PEM fuel cell system with an anode operating with the benefit of 0.40 mg/cm² of iron oxide when synthetic reformate is imposed, including recovery when pure hydrogen subsequendy replaces the synthetic reformate at 960 minutes. Figure 21 shows the beneficial effects of magnetic modification of both electrodes of a PEM fuel cell subjected to synthetic reformate (hydrogen plus 100 ppm of carbon monoxide).

Figure 22 contrasts "best results" for PEM fuel cell function with iron oxide magnetic modification of neither cathode nor anode, cathode only, and both cathode and anode, including in the presence of synthetic reformate (hydrogen plus 100 ppm of carbon monoxide).

Figure 23 compares oxidation currents as a function of the square root of the scan rate for magnetically modified electrodes of PEM fuel cells oxidizing ethanol, acetaldehyde, or acetic acid.

Figure 24 compares oxidation currents for organic fuels at nonmagnetic and magnetic composite modified electrodes.

Figure 25 presents voltammograms for various redox couples.

Figure 26 shows a plot of calculated versus experimental diffusion data.

DETAILED DESCRIPTION OF THE INVENTION

None of the above-discussed examples of chemical kinetics are affected by electron- nuclear spin polarization.

Consider the case in which one chemical species is a singlet, and the second chemical species is either a doublet or a quartet (i.e., a species that has three unpaired electrons, which species is denoted j ). For this case, magnetic effects are restricted to electron-nuclear and nuclear spin polarization since a spin polarized electron cannot polarize the paired electrons of a singlet. Because nuclear spin polarization effects are negligible relative to electron- nuclear spin polarization, magnetic effects on singlet/ radical systems occur through electron- nuclear polarization. Reactions between a singlet and a doublet are spin allowed; in other words, no change in spin is required during the course of the reaction. In the following scheme, the reaction begins with the doublet already polarized by an applied magnetic field, which is denoted as D*, where the asterisk indicates the species is polarized. A doublet complex [J^"D*]^D is formed which can either undergo electron transfer to the products D* + S, or, through electron-nuclear spin polarization between S and D*, form a new doublet complex [ ^- *]^D, which can undergo transfer to form D* + S. In [i^ *]¹⁵, the polarized electron on D* has polarized the nucleus on S.

[SD*]^D

S + D* I k D* + S (6)

[S*D*]^D

The stabiUty constant of intermolecular electron transfer, K, describes the reversible formation of the first complex from the reactants. The rate constants for electron transfer for the first and second complexes to form products are k_tt , and k_t, ₂, respectively.

Conversion to the electron-nuclear spin polarized complex or cross-relaxation has a rate constant k (units of s^"1). Electron-nuclear spin polarization is a magnetically susceptible process, as denoted by =*. The cross-relaxation rate constant (Carrington & McLaughlin, 1967, above. K.M. Salikhov, R.Z. Sagdeev, andA.L.Buchachenko, 1984. "Spin polarization and magnetic effects in radical reactions." In Studies in Physical and Theoretical Chemistry 22:

241-242. Y.N. Molin, editor. Elsevier, New York, N.Y., the contents of all of which are incorporated herein by reference) is k( ) = _& βH / π + _& β_N H I π + A β / 2 (7)

The constants , β_N, and ι are the Bohr magneton, the Bohr nuclear magneton, and Planck's constant, respectively. H is the external magnetic field strength in Gauss, and A is the hyperfine coupling constant in Gauss. For a paramagnetic species (i.e., a radical), ^, is the electronic ^-factor, and^ is the nuclear ^-factor, both of which are dimensionless. The^- factors are measures of the magnetic properties of a species, and are determined by EPR. Because β - 2000 β^, and g_t and g^ are comparable, the nuclear spin polarization term is negligible. The hyperfine coupling term, A β / 2 , is also neghgible.

Lifetime considerations for the above reaction scheme yield the rate constant for the overall electronic exchange reaction, k_ex, which has units of M^"1 s^"1: k_! . ( I) = K [k_! (k + k_e + k k _lt^ I (k + _et (8)

For k_{et 2} » k, and k_tt , slow, the reaction tends to proceed through cross-relaxation, and the rate expression simplifies to: k_ex (H) - Kk = Kg βH / h π + K (g_N β_N H / π + A β / 2 ) (9)

Given neghgible nuclear polarization and hyperfine coupling, a simpler equation results: (H) - Kg_x βH - 3 x 10" (sG)-' K_& H (10) Thus, for electron-nuclear polarization, the exchange rate constant is enhanced in proportion to the applied magnetic field H. A similar expression describes intersystem crossing reactions between two radicals (see Turro & Kraeuder, 1980, cited above).

Reactions between a singlet and a quartet are described as spin-forbidden because such reactions violate the quantum mechanical requirement of spin conservation. Because a reaction between a singlet and a quartet is spin-forbidden, it can proceed only through spin- orbit coupling or a preequilibrium spin change (J.J. Zuckerman, ed., 1986. Inorganic Reactions and Methods Volume 15, VCH, Deerfield Beach, FL, the contents of all of which are incorporated herein by reference). The quartet complex, [SQ*ψ is disallowed from direct conversion to the second quartet complex \Q S , and, instead, [ ^*]⁵ is first converted to a doublet complex, 5^2*]° through a magnetically susceptible relaxation with a rate constant k. The doublet complex [S*<2 ]^D is converted to a second doublet complex [Q*S*]^D, which can be converted to the quartet complex \Q*S . This latter complex can then dissociate into products. The relaxation mechanism allows for rapid preequihbrium spin change. Given fast electron transfer from [S*0*]^D, Equation 10 also describes the spin forbidden reaction.

Special Case of Self Exchange Reactions, and the Dahms Ruff Model. Electron exchange reactions occur when an electron is passed from one molecular or ionic species to another. If the reactants are two different oxidation states of the same species (redox couples), and the products are the same as the reactants, then the reaction is known as a self exchange reaction (see Salikhov, Sagdeev, & Buchachenko, 1984, above); that is, M" + M"^±l /V ^Λ±1 + M". Under appropriate conditions of high concentration and slow physical diffusion, self exchange reactions enhance apparent diffusion coefficients when the exchange in space of M" and M"^±l is accomplished slowly by physical diffusion, but efficiendy by electron exchange between M" and M"^~ Self exchange enhanced diffusion coefficients are observed for transition metal complexes concentrated in thin films of ion exchange polymers, such as Nafion® (D.A. Buttry & F.C. Anson, 1983. "Effects of electron exchange and single-file diffusion on charge propagation in Nafion films containing redox couples." /. Am. Chem. Soc. 105: 685-689. D.A. Buttry & F.C. Anson, 1981. "Electron hopping vs. molecular diffusion as charge transfer m'echanisms in redox polymer films." J. Electroanal. Chem. 130: 333-338. H.S. White, J. Leddy, & A.J. Bard, 1982. "Investigations of charge transport mechanisms in Nafion polymer-modified electrodes." J. Am. Chem. Soc. 104: 4811-4817, the contents of all of which are incorporated herein by reference). Like most spin polarization reactions, self exchange proceeds through formation of a precursor complex (see Zuckerman, 1986, above). For reactions between a singlet and a radical, self exchange studies provide experimental access to mechanisms that proceed through electron-nuclear spin polarization in the absence of electron spin polarization (P.J. Hore & K.A. McLauchlan, 1980. Chem. Phys. Let. 75: 582, the contents of all of which are incorporated herein by reference).

Dahms and Ruff first described how an exchange reaction can contribute to make apparent diffusion faster than physical diffusion (K. Dahms, 1968. "Conduction in aqueous solution." /. Phys. Chem. 72: 362-364. I. Ruff & V.J. Friedrich, 1971. "Transfer diffusion. I. Theoretical." /. Phys. Chem. 75: 3297-3302. I. Ruff, V.J. Friedrich, K. Demeter, & K.

Csillag, 1971. "Transfer diffusion. II. Kinetics of electron exchange reaction between ferrocene and ferricinium ion in alcohols." /. Phys. Chem. 75: 3303-3309. I. Ruff & I. Korosi- Odor, 1970. "Application of diffusion constant measurement to the determination of the rate constant of electron-exchange reactions." Inorg. Chem. 9: 186-188, the contents of all of which are incorporated herein by reference). Electron exchange efficiency determines the enhancement, and depends on the distance between redox moieties as embedded in their concentration C*, the distance of closest approach for the moieties <5in units of cm, the self- exchange rate constant k₍ m^' units of M^'x ', and the physical diffusion coefficient D_ml in units of cm²/s. The Dahms-Ruff equation describes the apparent (i.e., measured) diffusion coefficient D^„, which has units of crrr/second: » = D_al + c? k„ C* / 6 (11)

Electrochemical perturbation allows control of self exchange reactions and determination of D^, (see Buttry & Anson, 1983; Buttry & Anson, 1981; and White, Leddy,

& Bard, 1982; above). The impact of a magnet field on k_t will be reflected in D_φp to the extent that D_ml is small and C* is large. Magnetic composites of magnetic microparticles provide a ready means of examining the effects of strong, local magnetic fields on self exchange reactions, and on electron-nuclear polarization in particular.

1. Basic Fuel Cell Function, and Magnetic Modification

Fuel cells may be used to provide electrical current/voltage/power to an external device, to provide direct surface area for conducting synthetic chemical and electrochemical reactions, and to provide interfacial boundaries for such reactions. The apparatus and method of use thereof of the present invention expand these applications to include the ability to facilitate chemical reactions that, under conditions of the common nonmagnetic electrode setup, are kinetically disfavored and/or quantum mechanically disallowed, including quantum mechanically spin disallowed. Magnetic modification of one or both electrodes of a fuel cell apparatus permits application of intense magnetic fields at loci that are located at, or close to, the magnetically modified electrode(s), for example, on the order of within 10 nm from the magnetically modified electrode surface.

The following sections compare and contrast the function of fuel cell systems as basic physical parameters are altered: temperature, pressure, amount of magnetic loading/coating, type of micromagnets used for the magnetic loading/coating, use of oxygen versus air, feed rate of fuel to the electrodes, etc. In particular, parameters of physical performance of magnetically modified PEM fuel cell systems are contrasted with the same physical performance parameters of nonmagnetically modified PEM fuel cells. Even more particularly, the heightened performance levels of PEM fuel cells in which both the cathode and the anode are magnetically modified are contrasted with the significandy lower performance levels that are observed with conventional nonmagnetically modified PEM fuel cell systems.

a. H /0₂ Fuel Cells + Only Cathode Magnetically Modified. i. T_CELL effects. Figures 2, 3, 6 and 7 examine the effects of temperature of the fuel cells on function of the fuel cells. In Figure 2, curves of power versus potential are shown, with the temperature ranging from 25 to 70°C, when iron oxide magnets are loaded at values of 0.05 (0 ), 0.14 (□), 0.20 (Δ), 0.35 (O), and 0.40 (*) mg/cm². A second replicate of the 0.14 mg/cm² magnetic loading data are shown as (X). The power versus potential curves for non-magnetically modified fuel cells are shown as (♦). Anode and cathode humidification temperatures are 70 and 65°C, respectively, for iron oxide loading of < 0.14 mg/cm², and 75 and 70°C for loadings > to 0.20 mg/cm². Flow rates to the anode and cathode are 400 and 600 cc/minute, respectively. These cells are H₂/0₂ fuel cells operating at a pressure of one atmosphere (i.e., unpressurized cells). Only the cathode is magnetically modified in these experiments. Additional curves with cathode coatings of samarium cobalt magnets at 0.13 mg/cm² for 70°C are shown as (#). In Figure 3, similar data are shown except that curves of potential versus current are shown.

In Figure 6 are shown curves of power density versus temperature for magnetic loadings from 0 to 0.20 mg/cm². Experimental conditions are the same as those described in Figures 2 and 3.

In Figure 7 are shown power versus potential, and potential versus current, curves for fuel cells operating at 20°C, one atmosphere of oxygen, and magnetic loading at 0.14 mg/cm². Specifically, humidification temperatures for feeds to the anode and cathode are varied from 30 to 75°C for the anode, and the cathode temperature is 5°C lower than the anode temperature. These results indicate that humidification temperature has virtually no effect on the function of these H₂/0₂ fuel cells, particularly when the fuel cell is operating at 20°C.

The effects of temperature on optimum current and on maximum power produced by these fuel cells are summarized in Tables 1 and 2, respectively.

Table 1

The data of Figures 2, 3, 6 and 7, and of Tables 1 and 2, demonstrate the benefits that accrue from magnetically loading/coating cathodes of PEM fuel cells. First, fuel cells with magnetically modified cathodes generate substantially higher current and power densities than those with nonmagnetic cathodes. The benefit varies according to the degree of magnetic loading of the cathodes, with optimal results occurring around 0.35 mg/cm² of iron oxide magnets at 70°C for the experimental results presented here. For iron oxide magnetic loading even at the low level of 0.05 mg/cm², maximum current and maximum power output always were at least double those of nonmagnetic fuel cells over the temperature range 30 to 70°C. Furthermore, and surprisingly, PEM fuel cells with only 0.14 mg/cm² of iron oxide magnets operating at 25°C produced more than 62% greater maximum current, and more than 70% greater maximum power, than nonmagnetic fuel cells operating at 70°C, where both sets of fuel cells are utilizing 1 atm of oxygen. Under conditions of maximum current output (i.e., 70°C, 0.35 mg/cm² of iron oxide magnets, for 1 atm of oxygen), maximum current and power densities are more than three times greater when the cathode is magnetically modified than when no magnetic loading is employed, with the peak value approximating 4.4 amps/cm² for maximum current, and 1.2 watts/cm² for maximum power, in the electrodes studied here. As is shown in Figure 6, the temperature coefficient for power density increases as iron oxide loading is increased up to 0.20 mg/cm².

Another comparison that highlights the surprising benefits of magnetic loading of PEM fuel cell cathodes is the power output of a cathode loaded with only 0.14 mg/cm² of iron oxide magnets (0.56 watt/cm²) versus that of a nonmagnetic PEM fuel cell operating at 70°C (0.40 watt/cm²). Thus, room temperature operation of a lighdy magnetically loaded PEM fuel cell cathode produces roughly 35% more power than a conventional nonmagnetic fuel cell that operates at 70°C. It deserves emphasis that such power output is done at atmospheric pressure; i.e., without the need for pumps/compressors that potentially can require high maintenance, and that can be noisy and an energy parasite to the fuel cell.

ii. Pressure effects.

Effects of pressure for H₂/0₂ PEM fuel cells operating at 70°C from 1 to 3 atmospheres are shown in Figures 4, 5A and 5B. Magnetic loading of cathodes are indicated as 0.14 (D), 0.20 (Δ), 0.35 (O), and 0.40 (*) mg/cm². A second replicate of the 0.14 mg/cm² data are shown as (X). Experiments with samarium cobalt magnet loading at 0.13 mg/cm² are shown as (•). Anode and cathode humidification temperatures are 70 to 75°C, and 65 to 70°C, respectively.

In Figure 4, potential versus current curves are shown for pressures of 1, 1.7, 2, 2.4 and 3 atmospheres of oxygen. In Figure 5A, curves of power versus potential are shown for the same pressures of oxygen. Figure 5B shows power density as a function of pressure from

1 to 3 atmospheres. Generally, as pressure is increased, power density and current density also increase. This is true even for the non-magnetic fuel cells. Furthermore, magnetic loading results in higher functioning fuel cells, compared to fuel cells that have not been magnetically modified. However, as pressure of oxygen is increased, the amount of effect decreases for maximum current density resulting from magnetic loading; i.e., a tighter grouping of the data is apparent. At 3 atm of oxygen, maximum current approaches about 5 amps/cm², and maximum power approaches 1-6 watts/cm².

These data are summarized in tabular form in Tables 3 and 4. Table 3 shows the maximum current as a function of atmospheres of pressure of oxygen and magnetic loading of a cathode electrode. Table 4 shows maximum power as a function of atmospheres of pressure of oxygen, as well as the effects of magnetic loading of the cathode for maximum current produced by these electrodes.

The great benefit of the present invention is clear from comparisons such as the following. At 70°C, PEM fuel cells with magnetically modified cathodes operating at 1 atm of oxygen can produce more than twice the maximum current output of conventional PEM fuel cells that are fed oxygen at 3 atm; similarly, maximum power output is more than 80% greater at 1 atm with magnetic modification, compared to PEM fuel cells at 3 atm, but with no magnetic modification of cathodes. Thus, once again, magnetic modification produces results superior to prior art conventional, nonmagnetized cathodes.

Table 3

Table 4

b. H / Air Fuel Cells + Only Cathode Magnetically Modified. i. Ic_ELL effect?.

In fuel cells that use air rather than oxygen, effects of temperature over 25 to 70°C are shown in Figures 8 and 9A, as well as in Figure 9B. In Figure 8, plots of potential versus current are shown for different levels of magnetic loading of the cathode. Similarly, in Figure 9A, curves of power versus potential are shown for different levels of loading of magnets on the cathode. In Figure 9B, power density is plotted as a function of temperature and degree of loading of the cathode. Responses of non-magnetically modified fuel cells are shown as (♦). Cathode loadings with iron oxide magnets are indicated for 0.14 (□) and 0.20 (Δ) mg/cm². Results for cathode loading with samarium cobalt magnets (•) are also shown. Flow rates are 200 and 1400 cc/minute at the anode and cathode, respectively. These effects of temperature are summarized in Tables 5 and 6. In Table 5, maximum current is shown as a function of temperature at one atmosphere of air for both magnetically modified and non-magnetically modified fuel cells when magnetic modification is for the cathode only. In Table 6, similar data are shown for maximum power of magnetically modified and non- magnetically modified fuel cells.

Generally, maximum current density and maximum power density both increase with increasing temperature over the indicated temperature ranges. In addition, as in studies using oxygen, results from studies conducted using air instead indicate that loading cathodes with magnets gready enhances fuel cell function. For example, fuel cells supplied with cathodes loaded with 0.20 mg/cm² of iron oxide and operating at 30°C produce more than twice the maximum current density output, and greater than 80% more maximum power density output, than the conventional, non-magnetic PEM fuel cell operating at 70°C. The maximum power density for PEM fuel cells with iron oxide magnets loaded onto the cathode at 0.14 mg/cm² is 2.42 ±0.04 times that of PEM fuel cells that are nonmagnetic, and 3.5 ± 0.1 times when the iron oxide loading is 0.20 mg/cm². The maximum current density is greater than 3 times higher for iron oxide loading at 0.20 mg/cm² over the entire temperature range, and greater than 2 times higher at 0.14 mg/cm², compared to PEM fuel cells with nonmagnetically modified cathodes.

ii. Pressure effects.

Effects of pressure for fuel cells operating on air, rather than oxygen, are shown as a function of pressure at 70°C in Figure 10. In Figure 10A are shown curves of potential as a function of current at 1, 2 and 3 atmospheres of pressure for air. In Figure 10B, curves of power versus potential are shown for 1, 2 and 3 atmospheres of pressure for air. Flow rates for the anode and cathode are 200 and 1200 cc/minute, respectively. These data are summarized in tabular form in Table 7, in which maximum current is shown as a function of pressure for fuel cells operating at 70°C and 1 to 3 atmospheres of pressure for air. For these electrodes, pressuring from 1 to 3 atmospheres increases the maximum power density output by 42%, and increases the maximum current density output by 26%.

Table 7

iii. Effects of feed rate of air to cathode. In Figures 1 IA and 11B, plots of potential versus current, and power versus potential, respectively, show that increasing flow rate of air to the cathode increases the current output and the power output of PEM fuel cells operating at 70°C. As the flow rate of air is increased above 1200 cc per minute, current and power output increase more than ordinarily would be expected based on projections of the data for lower flow rates. In Figures 11 C and 1 ID, current output and power output are shown to increase even more when higher loading of magnets is employed for the cathode. These results are summarized in tabular form in Table 8.

Table 8

c. H / Q₂ Fuel Cells + Both Cathode and Anode Magnetically Modified.

Relative to non-magnetically loaded cathodes, results above showed that magnetic modification of the cathode produces dramatic increases in power output and current output of H₂/0₂ and H₂/air fuel cells. In this section, results for electrode systems for which the anode also is magnetically modified are presented, i- X _ELL effects.

In Figures 12 to 15, the effects of temperatures are shown for H₂/0₂ PEM fuel cells in which both the cathode and the anode have been magnetically modified. In Figure 12, curves of potential versus current are shown for the temperature range 30 to 70°C. The cathodes are loaded with iron oxide magnets at 0.20 mg/cm², and anodes are loaded at 0.00 (Δ), 0.10 (■), 0.20 (*), 0.30 (O), 0.40 (•), and 0.50 (X) mg/cm². Anode and cathode temperature are 75 and 70°C, respectively. Flow rates are 400 and 600 cc/min. at the anode and cathode, respectively. Nonmagnetically modified fuel cells are designated by (♦). In Figure 13, curves of power versus potential are shown for the same temperature range. In Figures 14 and 15, similar data are shown for slightly different conditions, including extending the temperature down to 25°C. Data that are denoted (^A) represent electrode systems in which both the cathode and the anode are loaded with 0.20 mg/cm² of iron oxide magnets. The other data are from Figures 2 and 3, in which only cathodes are magnetically modified. These data are summarized in tabular form in Tables 9 through 11. In Table 9, values for maximum current at 0.091 volt and one atmosphere of oxygen are shown for electrode systems in which both the cathode and anode are magnetically modified. Furthermore, the results of loading of magnets onto the anode are covered for the range of 0.10 to 0.50 mg/cm². Likewise, in Table 10, maximum power is shown as a function of temperature and loading of the anode with magnets. In Table 11, maximum current and maximum power values are given as a function of temperature from 70 to 30°C.

Generally, these data show enhanced function of PEM fuel cells when both cathode and anode are magnetically coated, compared to sole magnetic modification of the cathode. The largest current density, 5.16 A/cm², is observed at 70°C, which represents an 18% increase over the best case for magnetic modification of the cathode alone, and is 3.77 fold greater than the current density for the conventional nonmagnetically modified fuel cell. At temperatures of 50 to 70°C, the maximum power density output is 3.6 + 0.1 fold greater than that of the nonmagnetic fuel cell, and is roughly a 17.5 ± 0.5% improvement over that of fuel cells with only cathode magnetic modification.

The salient and critical point is that magnetic modification of both cathode and anode provides even further enhancement o f fuel cell function, compared to magnetic modification of the cathode alone, especially at the higher temperatures. The results presented herein do not prove a mechanism(s) to totally explain all of these data. However, it is commonly accepted that the kinetics for hydrogen oxidation are rapid compared to the oxygen reduction kinetics in nonmagnetic fuel cells. Furthermore, if the kinetics for oxygen reduction are improved sufficiendy (for example, by use of magnets on electrodes), a situation can be envisioned in which the hydrogen kinetics then become rate limiting. The mechanism for hydrogen reduction proceeds as

H^_ώ ^ 2 H-^ (13)

H ads =?-- -^■• + H' (14)

H ads ^

" .-A (15) where subscripts soln and ads refer to species in solution and species adsorbed on the platinum or other Noble metal catalyst, respectively. Because of the adsorbed radical intermediate, hydrogen oxidation is susceptible to magnetic effects. In addition, proton transfer reactions are also theoretically susceptible to magnetic fields.

In one preliminary experiment to these ends, hydrogen was dissolved in water and electrodes modified with Nafion® and iron oxide magnets. The hydrogen was electrolyzed at about 30 mV less applied potential than at an electrode modified only with Nafion®, and electrical current was enhanced about 30%. Furthermore, it was observed that if the magnetically modified cathode is inadvertentiy put in the fuel cell test stand with the magnetic layer on the anode side, performance is still improved as compared to a nonmagnetically modified fuel cell.

Results at 30 and 40°C do not show enhancement relative to cathode magnetic modification alone. Two points may be relevant to this observation. First, because oxygen reduction kinetics are more sensitive to temperature than hydrogen oxidation kinetics, it is possible that at lower temperatures the system is stricdy cathode limited, which leads to magnetic modification of the anode not enhancing performance. Second, modification of both electrodes yields potential versus current curves that fall off fairly linearly, which is consistent with increased resistance in the MEA interface. It is possible that, at lower temperatures, the conductivity of the interface is reduced because of the polystyrene coating on the magnetic particles.

Table 9

Table 10

Table 11

ii. Pressure effects.

Effects of pressure for oxygen at 1, 2 and 3 atmospheres, 70°C, and magnetic loading of both cathode and anode at 0.20 mg/cm², are shown in Figure 16 as (^A). Other curves are those of Figure 4. In Figure 16A, curves are presented for potential versus current for the three pressures of oxygen. In Figure 16B, curves of power versus potential are shown for the three pressures of oxygen. Data on effects of pressure are shown in tabular form in Tables 12 through 14. In Table 12, values for maximum current density are shown as a function of temperature and magnetic loading of the anode. In Tables 13A-13D, maximum power density is shown as a function of temperature for different levels of magnetic loading of the anode, and for different voltages. In Table 14, temperature effects data are presented for maximum current density and maximum power density for fuel cells in which iron oxide magnets are used on both the cathode and the anode at the level of 0.20 mg/cm² at 1, 2 and 3 atmospheres of oxygen.

Unlike the above discussed PEM fuel cell systems, in which only the cathode was magnetically modified, magnetic modification of both the anode and the cathode produces a strong pressure dependence. The highest observed maximum current density value of 6.81 amp/cm² occurred for 3 atm of oxygen and 70°C, and is 39% higher than the best case for a fuel cell with a magnetically modified cathode, and 3.5 times the current produced by a nonmagnetically modified fuel cell. Similarly, the highest observed maximum power density value of 1.94 watts/cm² is 3 times the power density found for the nonmagnetic fuel cell, and 33% more power density than found for 0.20 mg/cm² of iron oxide on the cathode alone. The power density increases Hnearly with pressure according to Power (W/cm) — 0.2754 x Pressure (Atm) + 1.1135, which equation has a correlation coefficient of 1.0000.

Table 12

Table 13A

Table 13B

Power at 0.8, 0.7 and 0.6 V for 1 Atmosphere of O_z for Magnetic and Nonmagnetic Fuel Cells:

Flow rates for H₂/0₂ are 400/600 cc/min for a 5 cm² cell;

Humidification temperatures for anode and cathode are 70-75 and 65-70°C

Magnetic Loading^g/cm²) Cell Temperature

Cathode Anode 70°C 60°C 50°C 40°C 30°C 25°C 20°C

Power at 0.8 V (W/cm²)

Non-Magnetic

0.00 - 0.152 0.119 0.116 0.113 0.118

Magnetic - Iron Oxide

0.05 - 0.131 0.134 0.133 0.155 0.138

0.14 - 0.116 0.194 0.178 0.183 0.160 0.100 0.108

0.20 - 0.137 0.198 0.167 0.144 0.113

0.35 - 0.184

0.40 - 0.171

0.20 0.20 0.205 0.240 0.225 0.165 0.131

Magnetic - Sm Co

0.13 - 0.151

Power at 0.7 V (W/cm²)

Non-Magnetic

0.0 - 0.273 0.225 0.213 0.199 0.203

Magnetic-Iron Oxide

0.05 0.481 0.411 0.378 0.364 0.315

0.14 - 0.396 0.479 0.446 0.419 0.365 0.262 0.267

0.20 - 0.455 0.558 0.479 0.416 0.338

0.35 - 0.471

0.40 - 0.534

0.20 0.20 0.637 0.660 0.603 0.428 0.347

Magnetic-Sm Co

0.13 - 0.484

Power of 0.6 V (w/cm²)

Non-Magnetic

0.00 - 0.349 0.297 0.300 0.278 0.269

Magnetic-Iron Oxide

0.05 0.725 0.634 0.576 0.573 0.480

0.14 - 0.770 0.754 0.754 0.658 0.564 0.44, 0.441

0.20 - 0.828 0.952 0.821 0.717 0.593

0.35 - 0.953

0.40 - 0.887

0.20 0.20 1.056 1.095 0.921 0.689 0.575

Magnetic - Sm Co

0.13 - 0.889 Table 13C

Maximum Power at 0.8, 0.7 and 0.6 (W/cm') of O_z at a CeU Temperature of 70 °C for Magnetic and Nonmagnetic Fuel Cells:

Flow rates for H₂/0₂ are 400/600 cc/min for a 5 cm² cell; Humidification temperatures for anode and cathode are 70-75 and 65-70°C

Pressure

Magnetic Loading(mg/c: m²) 1 Atm 1.7 Atm 2.0 Atm 2.4 Atm 3.0 Atm

(+0 psi) (+10 (+20 (+20 (+30 psi)

Cathode Anode psi) psi) psi)

Power at 0.8 V (W/cm²)

Non-Magnetic

0.00 - 0.095 0.152 0.239 0.296

Magnetic - Iron Oxide

0.14 - 0.113 0.209 0.330 0.379

0.20 - 0.137 0.265 0.347

0.35 - 0.184 0.237 0.248 0.306 0.377

0.40 - 1.171 0.398

0.20 0.20 0.205 0.319 0.462

Magnetic - Sm Co

0.13 0.151 0.261 0.363 Power at 0.7 V (W/cm²)

Non-Magnetic

0.0 - 0.205 0.283 0.396 0.492

Magnetic-Iron Oxide

0.14 - 0.380 0.590 0.737 0.814

0.20 - 0.455 0.723 0.816

0.35 - 0.431 0.518 0.661 0.585 0.888

0.40 - 0.534 0.901

0.20 0.20 0.637 0.862 1.044

Magnetic-Sm Co

0.13 0.484 0.657 0.791 Power of 0.6 V (w/cm²)

Non-Magnetic

0.00 - 0.293 0.394 0.492 0.626

Magnetic-Iron Oxide

0.14 - 0.712 0.998 1.148 1.239

0.20 - 0.828 1.193 1.257

0.35 - 0.953 0.731 1.062 0.791 1.381

0.40 - 0.887 1.389

0.20 0.20 1.056 1.324 1.631

Magnetic - Sm Co

0.13 - 0.889 1.045 1.134 Table 13D

Power at 0.8, 0.7 and 0.6 V for 1 Atmosphere of 0₂ for Magnetic and Nonmagnetic Fuel Cells:

Flow rates for H,/Air are 200/1400 cc/min for a 5 cm² cell; Humidification temperatures for anode and cathode are 70-75 and 65-70°C

Magnetic Loading^g/cm²) Cell Temperature Cathode Anode 70°C 60°C 50°C 40°C 30°C 25°C

Power at 0.8 V (W/cm²)

Non-Magnetic

0.00 0.101 0.076 0.067 0.064 0.061

Magnetic - Iron Oxide

0.14 0.054 0.093 0.098 0.097 0.086 0.080

0.20 0.104 0.135 0.143 0.146 0.098 0.100

0.20 0.20 0.056 0.169 0.137 0.146 0.096

Magnetic - Sm Co 0.13 0.089 0.119 0.093 0.087 0.078 Power at 0.7 V (W/cm²)

Non-Magnetic

0.0 0.216 0.163 0.136 0.127 0.117

Magnetic-Iron Oxide

0.14 0.210 0.296 0.275 0.259 0.232 0.217

0.20 0.310 0.392 0.387 0.314 0.283 0.274

0.20 0.20 0.196 0.459 0.377 0.383 0.284

Magnetic-Sm Co 0.13 0.266 0.364 0.276 0.253 0.222 Power of 0.6 V (w/cm²)

Non-Magnetic

0.00 0.275 0.235 0.176 0.169 0.150

Magnetic-Iron Oxide

0.14 0.440 0.481 0.409 0.405 0.363 0.340

0.20 0.601 0.664 0.623 0.490 0.462 0.442

0.20 0.20 0.439 0.742 0.579 0.572 0.441

Magnetic - Sm Co 0.13 0.504 0.549 0.430 0.388 0.342

Table 14

d. Contrasting Function of Air and Q-, Fuel Cells.

It is of interest to direcdy compare the function of fuel cells operating under oxygen versus air atmospheres. Such results and comparisons are shown in Figure 17. In Figure 17A voltage versus current curves are shown for magnetically modified fuel cells in a pure oxygen environment. In Figure 17B similar data for fuel cells operating on air are shown. Pressures are 1 and 3 atmospheres, and some temperature data are shown, namely, results at 30°C and 70°C. These data are also summarized in Table 15, in which best performance is shown for magnetically modified anodes and cathodes at three atmospheres under oxygen versus air environment. Performance of nonmagnetic fuel cells is indicated as (♦), and performance of cathode only magnetic modification is shown for 0.20 (Δ), 0.35 (O), and 0.40 (*) mg/cm² of iron oxide magnets. Performance of fuel cells with both cathode and anode magnetic modification at 0.20 mg/cm² of iron oxide is indicated as (^A). The temperature is either 30 or 70°C, as indicated.

Table 15

Clearly, better function is obtained under pure oxygen than under air, as is expected. The maximum power value of 1.94 watts/cm² for operation under 3 atm of oxygen is 2.5 times the value obtained for operation under air at 3 atm. Similarly, the maximum current density value of 6.81 amps/cm"", obtained under 3 atm of oxygen, is just greater than 3.6 times the value obtained under 3 atm of air.

e. Samarium Cobalt Magnetically Modified Fuel Cells.

Results for the uses of samarium cobalt on the cathode are shown in Figures 2 and 3 for 70°C (•) and loading of samarium cobalt magnets at 0.13 mg/cm². Potential versus current, and power versus potential, curves for oxygen fed fuel cells are shown in Figures 4 and 5 for samarium cobalt magnets loaded at 0.13 mg/cm². In Figures 8 and 9, fuel cells run on air are shown with samarium cobalt magnets loaded at 0.13 mg/cm². It is interesting that the samarium cobalt magnets in Figures 8 and 9 yield results that are comparable to those for iron oxide magnets loaded at 0.14 mg/cm², except at lower potentials. Data on samarium cobalt magnetically modified fuel cells are summarized also in Tables 1 through

6. In general, samarium cobalt magnets loaded at 0.13 mg/cm² operated between the values for iron oxide loaded at between 0.14 and 0.20 mg/cm², except at temperatures < 50°C.

f. Stability.

Long-term stability studies have yet to be conducted. However, some data have been gathered on this aspect. For example, magnetically modified fuel cells have been run on oxygen for up to a week with no degradation in performance. In addition, membrane electrode assemblies (MEAs) have been removed from a test stand, left at a laboratory bench under open air and at room temperature for several weeks, and then returned to the test stand with no loss in performance from initial levels. Thus, magnetically modified MEAs at least have a modest degree of stability, and possibly even greater stability.

2. SYNTHETIC REFORMATION FUELS & INDIRECT REFORMATION.

Hydrogen is oxidized at the anode of many fuel cells. However, hydrocarbon fuels also may be oxidized, either direcdy (i.e., direct reformation) at the anode or indirecdy (i.e., indirect reformation), for example, by passing the fuel over a hot, copper and zinc catalyst to yield hydrogen. The use of such hydrocarbon fuels may carry certain cost, power/mass, or other benefits, but the problem of passivation also is present in conventional fuel cell electrode systems due to the production of small amounts of carbon monoxide, a minor but highly potent catalyst poisoning byproduct of hydrocarbon oxidation. The generated CO binds almost irreversibly to the noble metal catalyst, e.g., platinum. Some calculations estimate that as much as 98% of the noble metal catalyst surface is passivated. Performance on a reformate such as that used herein (i.e., hydrogen + 100 ppm of CO, for example) for even the most advanced catalysts using fuel cells of current technologies is much poorer than that observed on pure hydrogen. In addition, current low temperature fuel cell technologies do not permit function in any practical realm using indirect reformation, and direct reformation is impossible. As will be detailed below, magnetic modification of the anode of a PEM fuel cell eliminates almost all of the negative effects of passivation.

Mechanisms of oxidation of a given fuel at a fuel cell anode are complicated, variable depending on reaction parameters, and largely uncharted. As will be shown below, the magnetically modified electrodes of the instant invention help to elucidate some of the component partial reactions, or, at a minimum, to at least eliminate some of the array of theoretically possible partial reactions that make up an overall reaction. a. Passivation. Passivation was examined by feeding the anode hydrogen gas to which had been added 100 parts per million of carbon monoxide to form a synthetic reformate. Figure 18 depicts the general time course of current production for a PEM fuel cell operating at 70°C with only the cathode magnetically modified with 0.2 mg/cm² of iron oxide magnets and the anode not so coated (□), versus with cathode magnetically modified anode magnetically modified with 0.2 mg/cm² of iron oxide magnets (solid Hne). All electrodes are coated with 0.4 mg/cm² of platinum catalytic particles. When a flux of synthetic reformate is imposed on the anode, current density drops off dramatically and quickly when the anode is not magnetically modified. However, when the flux of reformate is imposed on a magnetically modified anode, the current density drops by less than 10%, and maintains approximately even function until the reformate flux is replaced with pure hydrogen (here, at about 120 minutes into the experiment); at this time, current density returns to the pre-reformate level (i.e., complete recovery). For these experiments, anode and cathode humidification temperatures were 75 and 70°C, respectively; and gas flow rates thereto were 400 and 600cc/minute, respectively.

Results from a similar set of experiments are shown in Figure 19, in which platinum ruthenium served as the noble metal catalyst on the cathode. This catalyst was selected because it is one of the best CO tolerant catalysts. Conditions are identical to those of Figure 18, except that the anode catalyst was 0.4 mg/cm² of platinum ruthenium. The anode was not magnetically modified, but the cathode was. At about 20 minutes into the experiment, reformate was fed to the anode, in place of the hydrogen feed, which caused a precipitous drop in current density, followed by a further gradual decline. At about 400 minutes into the experiment, the feed to the anode was switched back to pure hydrogen, which yielded an incomplete recovery in current density. The performance of this magnetically loaded (iron oxide magnets) and platinum ruthenium catalyst loaded anode was comparable to an anode that is nonmagnetic and has platinum catalyst loaded thereon.

In Figure 20, results are shown for another anode. Specifically, the anode had 0.40 mg/cm² of iron oxide magnets loaded/coated thereon. Humidification of the electrodes was performed at higher temperatures because good hydration is thought to be important in building CO tolerant anodes. Here, humidification temperatures were 90 and 80°C for the anode and the cathode, respectively. In addition, the flow of reformate was reduced to 150 cc/minute. Upon the introduction of reformate at zero time, current density dropped precipitously. At about 400 minutes into the experiment, a spontaneous, abrupt, sharp transition to about 2.8 amps/cm² occurred; at this point in time, the anode was still running on reformate. At about 980 minutes, hydrogen was substituted for the reformate, which resulted in a sharp rise in current density to the initial, pre-reformate value. Though an explanation for the spontaneous increase in performance on reformate is not at hand, possibilities include something boiling out of the system, something becoming hydrated, and something undergoing a phase change(s). This sharp transition to CO tolerance is commonly observed.

Curves of potential versus current and power versus potential are shown in Figure 21. Anode and cathode humidification temperatures were 75 and 70°C, respectively, and flow rates thereto were 400 and 600 cc/minute, respectively, for pure hydrogen and oxygen, magnetic modification of both cathode and anode, 0.20 mg/cm² of iron oxide magnets on the cathode and 0.20 to 0.40 mg/cm² on the anode (•); pure hydrogen, non-magnetic electrodes (♦); pure hydrogen, 0.20 mg/cm² of iron oxide on both cathode and anode (A.); and reformate, cathode with 0.20 mg/cm² of iron oxide magnets, anode with 0.40 mg/cm² of iron oxide magnets (•) , both electrodes with 0.40 mg/cm² of platinum catalyst, cell temperature 70°C, anode temperature 90°C, and cathode humidification temperature 80°C.

Reformate is hydrogen with 100 ppm of CO. Selected statistics are summarized in Table 16. The results lead to several conclusions. First, magnetic modification of both cathode and anode yields power (about 1 watt/ cm² in the fuel cell of these experiments; not optimized) that, on reformate, is approximately three times the power obtainable from a nonmagnetic cell run on pure hydrogen. Second, these results more or less prove that indirect reformation is possible, and suggest that direct reformation is feasible. Table 16

Characteristics of Various Cells on Hydrogen and Reformate: Cells at 70°C and 1 Al tm of Pressure'³'

Magnetic Loading Anode Symbol Max Max. Power Power Power (mg/cm²) Cathode Current Power (0.8 V) (0.7 V) (0.6 V) (A/cm²) (W/cm²) (W/cm²) (W/cm²) (W/cm²)

Hydrogen Fuel

0.0 0.0 ♦ 1.30 0.34 0.095 0.205 0.293

0.2 0.2 ▲ 5.16 1.38 0.205 0.637 1.056

0.2 0.4 • 4.11 1.09 0.122 0.417 0.739

Reformate (H₂ + 100 ppm CO)

0.2 o.o^fb) ^• 0 0.49 0.090 0.052 0.081 0.080

0.2 0.4^(c) O 3.10 0.958 0.136 0.426 0.764

(a) Except as noted, cell conditions are 70° C cell temperature, 75 and 70° C cathode and anode temperature, 400 and 600 cc/min for anode and cathode, and Pt catalysts. Catalyst loadings are 0.4 mg/cm²

(b) t/Ru catalyst (c) Anode flow rate 150 cc/min; Anode & Cathode Temp 90 and 80° C

Results from another set of exemplary experiments are summarized in Figure 22. In the top two panels, the functioning of two PEM fuel cells is contrasted. In particular, potential versus current density and power density versus potential curves are shown for a nonmagnetic cell (♦) versus for a cell with magnetic modification of both cathode and anode at 0.20 mg/cm² of iron oxide magnets(-----). In the middle two panels, potential versus current density and power density versus potential curves are shown for cells with a nonmagnetic anode but a cathode coated with 0.20 mg/cm² of iron oxide magnets, and 0.40 mg/cm² of platinum ruthenium catalyst coating on the anode and 0.4 mg/cm² on platinum on the cathode. The left panel shows hysteresis for forward versus backward scans, which reflects instability of this cell under reformation. In the bottom panel, similar results are shown for a fuel cell operating at 70°C and 1 atm with magnetic coating of both the anode and the cathode (0.40 and 0.20 mg/cm² of iron oxide, respectively) and anode feed of pure hydrogen (#; flow rates of 400 and 600 cc/minute, respectively) versus reformate (O; hydrogen with 100 ppm of CO; flow rates of 150 and 600 cc/minute, respectively). Both electrodes have 0.4 mg/cm² of Platinum. These results show that performance on reformate is not reduced compared to performance on pure hydrogen when the current density is < 1.5 A/cm². Above this current density, the data for the fuel cell on reformate are somewhat suppressed compared to pure hydrogen, but not badly; the reduction may arise in part from the lower flow of reformate. For this fuel cell, the maximum power density on hydrogen was 1.09 watts/cm², and on reformate 0.96 watt/cm². The forward and back traces for the potential versus current density were superimposable, which is consistent with a stably operating fuel cell. For 0.8 volt, the current density was 0.18 amp/cm²; for 0.8 amp/cm², the voltage was 0.67 volt. These positive effects from magnetic loading of a PEM fuel cell anode reflect the actions that locally strong magnetic fields of the magnetic microparticles have on molecules through electronspin polarization and/or electron-nuclear spin polarization. These magnetic fields facilitate reaction kinetics, and also can open up new reaction channels in a kinetic scheme having a plurality of reaction channels (also, referred to as reaction paths or reaction pathways). The decreased passivation results disclosed above are, in part, consistent with alteration of the rate of formation of platinum carbonyl bonds.

3. Oxidation of Organic Fuels at Magnetically Modified Electrodes

The classification of fuel cells based on choice of fuel includes hydrogen, indirect reformation, and direct reformation cells. In hydrogen/oxygen (H_j/O_j) and hydrogen air (H₂/air) cells, pure hydrogen is fed into the anode and oxygen or air is fed into the cathode.

For pure hydrogen fuel, oxygen reduction is rate determining. However, tanked hydrogen is neither a convenient nor a safe fuel. An indirect reformation fuel cell uses a fuel processor that converts a feedstock such as methanol and water to a mixture of hydrogen and carbon dioxide over a hot catalyst. The fuel stream of hydrogen and carbon dioxide contains low level (< 1 %) carbon monoxide. Carbon monoxide fed into the anode rapidly passivates noble metal catalysts in low temperature systems without magnetic modification. Direct reformation fuel cells would be the most convenient because they would operate by feeding a fuel such as methanol direcdy into the anode where it would be electrochemically converted to carbon dioxide and hydrogen ions.

Mechanisms of Oxidation.

The oxidations of ethanol, acetaldehyde, and acetic acid occur in the potential regime between -0.200 and -0.500 volt versus the standard calomel electrode. There are also complex interactions and equilibria between these species, further comphcated by generation of all of these species during the oxidation of any one of them. Therefore, it is difficult to distinguish which reaction is occurring, and to determine the extent of a given reaction. However, by studying the oxidation of the three compounds individually, it is possible to begin to piece together the reaction sequence that occurs at the electrode.

i. Ethanol.

The oxidation of ethanol is a complex, multi-step process whose overall reaction is:

CH₃CH₂OH + 3 H₂0 - 2 C0₂ + 6 H₂ - 2 CO, + 12 + 12 H⁺ (16) The general reaction pathway is:

CH₃CH₂OH ^~* CH_jCHO ^→ CH₃COOH ^"* CH₄ ^~* CH₃OH ^~* ethanol acetaldehyde acetic acid methane methanol

CH₂0 ^"* HCOOH ^"* C0₂ (17) formaldehyde formic acid carbon dioxide

Although the general scheme for ethanol oxidation is well accepted, the details of the mechanism are not well estabhshed, and the mechanism probably shifts with subde variation in reaction conditions.

ii. Acetaldehyde.

Acetaldehyde is the first stable intermediate of ethanol oxidation. When ethanol undergoes electro-oxidation at an electrode surface, acetaldehyde and acetic acid are the major products. Ethanol loses two electrons to form acetaldehyde. Acetaldehyde, in turn, undergoes a two electron oxidation to form acetic acid. The general reactions for ethanol and acetaldehyde oxidation are shown below, written as reduction potentials:

CH₃CHO + 2 H⁺ + 2 f - CH₃CH₂OH E^υ = 0.2 volt (18)

CH₃COOH + 2 f + 2 H⁺ ^ CH₃CHO + H₂0 E° = -0.118 volt (19)

Acetaldehyde can undergo a disproportionation reaction to yield ethanol and acetic acid: 2 CH₃CHO ≠ CH₃CH₂OH + CH₃COOH (20)

From Equations 18 and 19, the overall cell potential (E°_a/) of the reaction is + 0.32 volt.

From the following relationships, the standard free energy of reaction, AG°, as well as the equilibrium constant, K, for the reaction as written in Equation 20 are calculated: AG° = -2 FE°_t// pi)

K = 10 e^χp(£70.059 volt) (22)

For acetaldehyde disproportionation, AG° = -32 kj per mole of acetaldehyde, and K = 2.5 x 10⁵. Thus, there is a significant driving force for acetaldehyde disproportion to acetic acid and ethanol. Such reactions complicate interpretation of voltammetric data.

hi. Acetic acid. Acetic acid is a product of ethanol and acetaldehyde oxidation, as is indicated in Equations 17 - 19. Acetic acid then undergoes oxidation to form methane, as follows (recall that equations are written here as reduction potentials): CH₃ ^* + C0₂ + - CH_jCOO^' (23)

Of course, the CH₃ ^* and the CH₃COO^* radicals can each combine with a respective H^* radical from solution or other source to produce CH₄ and CH₃COOH, respectively. Under Kolbe reaction conditions, a carboxylic acid can undergo decarboxylation via a two electron oxidation to form a radical species in solution. The radical species then tend to react either to form the corresponding alkane, or to undergo a one electron oxidation to form carbonium ions which can undergo further reaction. Acetic acid oxidation can also produce other products, such as esters and acetals. The particular composition of products for each of the above different combinations of reactants will depend on the individual reaction parameters, including such aspects as surface(s) of the electrodes, temperature, pH, oxidation state of the reaction solution, etc.

iv. Methanol. Methanol is the most popular choice as an alternative hydrogen carrier for fuel cells. Its electrochemistry is similar to that for ethanol. For reduction potentials, the reactions are as follows: CH₂0 + 2 H⁺ + 2 e - CH₃OH E° = 0.232 volt vs. NHE

(Normal Hydrogen Electrode) (24)

HCOO + 3 H⁺ + 2 . - CH₂0 + H₂0 E° = 0.082 volt vs. NHE (25)

The oxidation of methanol proceeds through formaldehyde and formic acid. In the complete oxidation of methanol, formic acid would be converted to carbon dioxide. The oxidation of ethanol proceeds through a similar sequence through acetaldehyde, acetic acid, methane, and methanol; the methanol sequence is the second half of ethanol oxidation.

Similar to the discussion for acetaldehyde (see Equation 20), formaldehyde can disproportionate: 2 CH₂0 + H₂0 <≠ CH₃OH + HCOOH (26)

The standard free energy per mole of formaldehyde is -72 kj, and K is 2 x 10⁶. Thus, the disproportionation of formaldehyde is substantial for the normal range of conditions of operation of typical fuel cells, and contributes to the complications in interpreting/deducing reaction mechanisms.

b. Interpretation of Cyclic Voltametric Data.

The above discussions of component or partial reactions of the overall scheme for ethanol oxidation highlight the complexities thereof. The various partial reactions are intermingled, in part, because their individual reduction potentials He within a 500 millivolt region. In addition, each species has multiple reaction pathways in which it may participate, including both electrochemically and chemically (for example, disproportionation) driven pathways. Because of these complications, quantitatively dissecting out exact fluxes for each pathway is extremely difficult; however, cyclic voltammetric methods may be used to obtain qualitative data to help deduce currents that reflect fluxes through specific pathways.

As will be discussed further below, voltammetric results for ethanol, acetaldehyde, and acetic acid indicate that passivation is suppressed at both magnetically and nonmagnetically modified electrodes. For example, for each of these fuels, electrical currents are larger, and the extent of conversion to products is higher, at magnetically modified electrodes, compared to nonmagnetically modified electrodes, including electrodes modified with Nafion®. In evaluating ethanol, acetaldehyde, acetic acid, and methanol as fuels, including their respective uses with each of the array of electrode modifications that produce distinctly different matrices for electrolysis, three figures of merit are utilized: the number of electrons transferred («), the current, and the current density. i. Number of electrons transferred, n.

In the reactions scheme of ethanol to acetaldehyde to . . . carbon dioxide (see

Equation 17), each electrolysis step is formally separated by two electrons. However, the complications of adsorption processes and facile interconversion(s) of any of the species in the sequence complicate interpretation of any voltammetric data pertaining thereto.

Nonetheless, the value deterrnined for n is a rough estimate fo the extent of electrolysis.

Thus, if n equals two, then, formally, the reaction has proceeded from one species to the next in the sequence; if n is four, then the reaction has proceeded down the sequence by two species. Noninteger values of n are consistent with the partial oxidation of a species, and/or the mixed oxidation to several different species.

Cyclic voltametric data can be used to estimate, roughly, the number of electrons transferred , n, in the oxidation of each species. However, the estimated value of n more accurately provides a crude measure of reaction efficiency, than it does of electrons transferred because of the above discussed complications. The value of n for each species was determined from the maximum peak current in the range between -0.200 and -0.500 volt vs. a standard calomel electrode. Results are summarized in Table 17 for different electrode modifications. The potential where the current peak was observed for each fuel is given as F oxidation'

The two generalizations that stand out clearly in the data of Table 17 are 1) that the value of n is greater for magnetic composites than for nonmagnetic composites, and 2) that the value of n increases as the substrate species becomes more oxygenated. The estimate of 1.65 for n for ethanol is consistent with partial oxidation of ethanol at the magnetically modified electrodes. Interestingly, n exceeds two for both acetaldehyde and acetic acid electrolyzed in magnetic composites, which suggests that the magnetic modification is effective at carrying the electrolysis process through several successive reaction species, instead of the (at most) two found for the nonmagnetic composites. For acetic acid, the estimate of n as about 3.2 suggests that acetic acid oxidation is carried almost to methanol. Table 17

Table 17: Approximate n for Irreversible Electron Transfer for Ethanol, Acetaldehyde and Acetic Acid at Modified Electrodes

ii. Current.

Values for peak current for various cyclic voltammetric potential scan rates v and electrode modifications are shown in Table 18. Note that the ethanol (2 mM) and methanol

(10 mM) concentrations are higher than the 1 mM concentrations used for acetaldehyde and acetic acid. Also, note that the runs with acetaldehyde were carried out at 0°C.

For electrodes modified with Nafion®, magnets, and platinized carbon microparticles "Naf/May/Pt-C", the values of Ip/C* (i.e., peak current/concentration) rank as acetaldehyde > acetic acid >> ethanol ~ methanol. When run with magnetically modified electrodes, acetaldehyde generated the highest peak current per concentration, despite the lower temperature. The above order for the four fuels suggests that acetaldehyde and acetic acid may serve as better fuels than alcohols in a magnetically modified PEM fuel cell. For acetic acid, the nonmagnetic composite electrodes "Naf/Pt-C" yielded about twice the peak current as the Nafion® modified electrode. For the composite electrodes, magnetic modification provided approximately 32% more peak current as nonmagnetic modification for acetaldehyde, and approximately 2.6 fold for acetic acid. Comparison for magnetically modified electrodes using acetic acid shows that the peak currents were about 5 fold higher than the peak currents for Nafion® modified electrodes. Peak current data' alone are insufficient for evaluating a given fuel for use in direct reformation in a PEM fuel cell. However, the above results suggest that acetaldehyde is a beneficial fuel for magnetically modified PEM fuel cells. In particular, acetaldehyde improved current output by more than 50% over the current output for ethanol. Acetic acid is also hkely to be a beneficial fuel, given its almost two-fold higher peak current output than ethanol, its lower cost than acetaldehyde, and its lesser irritant qualities. Among the composite electrodes, the surface area of platinum (153 cm² for Nafion® /platinized carbon electrodes, and 113 cm² for Nafion® /magnets/platinized carbon electrodes) is higher than for the unmodified and Nafion® only modified electrodes (0.459 cm²) because the platinum is distributed over the carbon black microparticles. If this platinum surface area is incorporated into the data of the composite modified electrodes, the efficiency of the Nafion® /magnets/platinized carbon electrodes is increased an additional 35% over the nonmagnetic modified electrodes.

Table 18

Table 18. Peak Current at Various Scan Rates (?) at Modified Electrodes for Ethanol, Acetaldehyde, Acetic Acid, and Methanol

Ip pO-Α)

- mV/s) 20 35 50 65 75 125 150 175 200

Ethanol (2 M, 25° C)

Naf/Mag/Pt-C 9.9 10.1 11.9 13.7 15.0 18.8 21.0 22.2 27.3

Acetaldehyde (1 mM, 0°)

Unmodified 1.12 1.14 1.13 1.6 1.3 2.05 2.17

Nafion 0.57 0.48 0.45 0.55 0.60 0.65 0.63 0.91 1.88

Naf/Pt-C 10.8 11.8 12.2 13.3 14.2 16.4 17.3 19.4 19.6

Naf/Mag/Pt-C 25.1 21.6 21.8 22.8 25.8 27.3 28.0 31.5 45.5

Acetic Acid (1 mM, 25° C)

Unmodified 10.1 9.92 11.0 11.5 21.1 14.3 15.3 16.2 17.0

Nafion 4.70 4.95 5.24 5.50 5.67 6.62 7.03 7.67 7.59

Naf/Pt-C 8.74 9.51 10.1 11.0 11.2 13.0 13.0 14.4 14.8

Naf/Mag/Pt-C 10.6 10.8 12.9 13.3 14.3 17.8 22.6 22.2 27.0 Methanol (10 mM, 25° C)

Unmodified 5.85 6.22 6.46 6.77 6.53 7.06 7.51 7.58 7.11

Graphical comparisons of the peak oxidation currents of ethanol, acetaldehyde, and acetic acid are shown in Figures 23 and 24. In Figure 23, values of peak current/concentration are shown for PEM magnetically modified fuel cells as a function of the square root of the scan rate. These data show that acetaldehyde exhibits higher oxidation efficiency than ethanol or acetic acid, and that acetic acid oxidation is more efficient than ethanol oxidation. In Figure 24, the data of Figure 23 are shown together with data for peak current/concentration for using acetaldehyde and acetic acid in the absence of magnetic modification of the electrodes. In each case (for acetaldehyde and for acetic acid), magnetic modification clearly improves oxidation efficiency. The data of Figures 23 and 24 suggest that the oxidation of acetaldehyde and the oxidation of acetic acid both lead to similar products; whereas oxidation of ethanol is stymied at incomplete conversion to acetaldehyde. The data also show that acetaldehyde and acetic acid sustain reasonable values of peak current/concentration even in the absence of magnets, and that these values for nonmagnets are higher than for ethanol with magnetic modification of the electrodes. These data are consistent with certain kinetic processes (most likely adsorptive processes) that limit the efficiency of ethanol oxidation.

iii. Current density. Analysis of data on currents in the manner of the previous section does not discriminate between faradaic (electrolytic) current and other sources of current, for example, capacitive current. The normalized current density, I_p/(A C₀ v'^/2), is sensitive to the faradic current, where A is the surface area of the platinum catalyst, and C₀ is the substrate concentration in solution. The quantity I_p/ ' (A C₀ v''²) is determined from the slope of a plot of I_p/A C₀ versus v' ². Table 19 summarizes data for the three substrates with magnetic and nonmagnetic composite electrodes. The parameter l_pf (A C₀ v'^/2) provides a measure of the efficiency of charge transfer during the course of electrolysis. These data clearly show that there is a higher efficiency for charge transfer when magnetic composites are used, compared to nonmagnetic composites. In particular, charge transfer is roughly two-fold better for acetaldehyde when magnetic composites are employed on the electrodes, and roughly four- fold better for acetic acid when magnetic composites are utilized on the electrodes, compared to electrodes coated with nonmagnetic composites.

In evaluating the three fuels to determine the most beneficial fuel source, competing parameters do not yet resolve the issue. For example, acetaldehyde may provide a higher power density than acetic acid, owing to its roughly 17% lower molecular weight. However, the ease with which a fuel crosses over from the anode to the cathode, passing through the ion exchange membrane that separates the electrodes, constitutes yet another material factor; and acetic acid may be predicted to have a lower cross over rate because the anionic sites of the ion exchange membrane would be expected to exclude the acetate ion. On the other hand, the effect of a fuel on swelling of the ion exchange membrane constitutes another factor that requires evaluation; and the data of Table 18 for Nafion® modified electrodes are suggestive of acetaldehyde having a lower crossover than acetic acid. Also, we have observed that Nafion® swells extensively in the presence of ethanol, which is consistent with high crossover of ethanol.

Table 19

Table 19. Normalized Current Densities for Ethanol, Acetaldehyde, and Acetic Acid at Electrodes Modified with Magnetic and Nonmagnetic Composites

Ip I AC₀v ^V2 (Ccm I (mole /Vs^~)) Ip I AC₀v ^1/2 (Ccm I (moleJ s)) Substrate Temperature Magnetic Composite Nonmagnetic Composite ro

Ethanol 25 4.6

Acetaldehyde 0 3.5 2.0

Acetic Acid 25 5 \_

4. Modeling of Magnetic Field Effects on Electron Transfer in Heterogeneous

Systems. a. General Principles: General Results.

Some of the basic physical relations pertaining to self exchange reactions were considered above, including the Dahm-Ruff equation (Equation 11). Such relations proved useful in modeHng magnetic effects on homogeneous electron transfer, i.e., electron transfer in a simple solution. Here, using known values of physical parameters, the self-exchange rate constant for electron transfer between reactants having different oxidation states and within the magnetic field of magnets on fuel cell electrodes coated with magnetic composites, k_M is seen to be directiy proportional to the magnetic field strength of the magnetic microparticles according to the relation: (_X = 3 x 10" Kg H (27) where Kϊs the stability constant,^ is the electronic ^-factor of the paramagnetic species in the self-exchange reaction, and His the magnetic field in the interfacial region (at 2,000 Gauss). Values of self-exchange rate constants for reactions close to electrode magnets were determined, and values for self-exchange rates in Nafion® (non magnetic) were obtained from the literature, values of apparent diffusion coefficients D^ in magnetic composites (2000 gauss) of the fuel cell electrodes of the instant invention, versus in Nafion®, were measured, and values of k_tx were calculated using the Dahm-Ruff equation. The electrochemical flux enhancement of a redox species probed with cyclic voltammetry is the square root of the ratio of the apparent diffusion coefficient in the magnetic composite, divided by the apparent diffusion coefficient in Nafion®. Although this model did an excellent job of modeling the experimental flux enhancement for metal redox couple complexes in magnetic composites of fuel cell electrodes of the instant invention, the Dahm-

Ruff equation is inefficient to accurately model the large and/or asymmetric potential shifts observed in our cyclic voltammetric studies of such redox couples as Ru(bpy)₃ ⁺² Ru(bpy)₃ ⁺³, where Ru refers to ruthenium, and bpy refers to bipyridyl. For example, peak potentials in 100 mV/second cyclic voltammograms shifted roughly 30mV lower in energy per unpaired electron in typical reactants. Furthermore, all metal complex redox couples studied in magnetic composites on fuel cell electrodes showed a decrease in the difference between the cathodic and the anodic peak potentials.

We have discovered that the large and asymmetric peak shifts in cyclic voltammograms are more accurately characterized by equations that describe how magnetic fields alter heterogeneous electron transfer that occurs at the surface (at the magnetic composite) of a fuel cell electrode. Specifically, we have discovered that heterogeneous electron transfer (i.e., transfer of electrons across two phases) across Nafion® and at the magnetic composite of a magnetically modified fuel cell electrode surface is better modeled by Marcus theory, which predicts that the heterogeneous electron transfer rate at an electrode is directiy proportional to the square root of the exchange rate constant. Because the self- exchange rate constant is directly proportional to the magnetic field, the following relationship provides an estimate of the heterogeneous electron transfer rate constant: . ( I k_hllιn (1G) = [^ (H) I (1 )]^I/2 = (H/1G)^,/2 (28)

Thus, k_fc_t^ (H) can be calculated as: A_ta, (fi) = ^, (l G) ι Ji^{, 2} (29) where H is the average magnetic field at the interface of the magnetic composite of the electrode surface (approximately 544 Gauss), k_htlm (1 G) is the heterogeneous electron transfer rate constant of the metal complex in Nafion® with no magnetic field perturbations other than the Earth's magnetic field, and k^,_^ (H) is the heterogeneous electron transfer rate constant in the presence of the average magnetic field (about 544 Gauss) at the interface of the magnetic composite of the electrode surface. The heterogeneous electron transfer rate constant of the metal redox couple complex in Nafion® under the influence of Earth's magnetic field was modeled using a finite difference one-dimensional computer simulation in which electron transfer rate constant, concentration of the redox couple, and apparent diffusion coefficient were the adjustable parameters. For none magnetic composites, the apparent coefficient was determined experimentally concetration is known from literature values. The simulation was used to determine k_httm (1G). Equation 29 yields k_hem (H). The same computer simulation protocol was then used to model the magnetic composite system, except that no adjustable parameters were employed. The bulk concentration of the redox species in the magnetic composite was assumed to be the same as the bulk concentration in Nafion®. The apparent diffusion coefficient in the magnetic composite was measured. The relationship between the apparent diffusion coefficient in the magnetic composite, flux enhancement, and the apparent diffusion coefficient in Nafion®: flux enhancement = ([D_app (magnetic composite) / D^ (Nafion®)]^{1 2} (30)

Therefore, D_φp (magnetic composite) = D_φfi (Nafion®) x [flux enhancement]² (31)

The heterogeneous electron transfer rate constant for the magnetic composites was found to be well approximated by the heterogeneous electron transfer rate constant in the

Earth's magnetic field times the square root of the average magnetic field at the electrode surface (i.e., at the interface of the magnetic composite) times the square root of the average magnetic field at the magnetically modified electrode surface. Note that paramagnetic species must be present at some time during the course of the electrolysis in order to have a magnetically perturbed heterogeneous electron transfer rate and rate constant.

Using the above referenced computer simulation protocol, peak currents for various redox couples were estimated within 10% of their respective measured values, and the initial peaks shifting to lower energy values for paramagnetic reactants undergoing scan rates of 50, 100, and 150 mV/second were estimated within +5 mV. Furthermore, the computer simulation protocol estimated the decrease in difference between the anodic and cathodic peak potentials within +5 mV. These small errors for the simulation protocol are all within experimental error, and provide support for this model of heterogeneous electron transfer for the redox couples studied at the Nafion®-magnetic composite interface of magnetically modified fuel cell electrodes, and are consistent with a substantial magnetic effect on heterogeneous electron transfer. As all heterogeneous electron transfers are thought to proceed through on one electron transfer, a paramagnetic species or intermediate will always exist. Thus, it is expected that magnetic fields will affect heterogeneous electron transfer rates.

b. Experimental. Microparticles: Magnetic (Polyscience) and superparamagnetic (Bangs Laboratories and Dynal) microparticles are a core of iron oxide shrouded with a thin, inert polymer layer. From electron micrographs, the diameters of Bangs beads/microparticles range from 0.5 to 2 μm, and Polyscience beads/microparticles are somewhat larger and more dispersed than the reported 1 to 2 μm. Dynal beads/microparticles are 4.5 |im in diameter with a small magnetite core. Large core beads/microparticles are magnetic, and once magnetized, sustain a field in the absence of an externally applied field. At the equator of a magnetic, 1 |lm radius magnetite microparticle, the field at its surface is about 2100 Gauss (=0.21 Tesla). Smaller core (Bangs and Dynal) beads/microparticles are superparamagnetic and sustain a field only in an externally applied field; Bangs beads/microparticles are calculated to have a 0.22±0.02 μm radius magnetic core.

Composites and Nafion® Film Preparation: Films and composites were formed with a suspension (5% wt/vol) of the perfluorinated, sulfonic acid cation exchange polymer, Nafion® 1100 (Aldrich). Nafion® has hydrated and fluorocarbon phases. Glassy carbon electrodes (0.459 cm²) were modified with either (a) a Nafion® film; (b) a composite of Nafion® and magnetic microparticles where the composite is formed under alignment from an external magnetic field; or (c) a composite of Nafion® and either magnetic or comparably sized nonmagnetic (Polyscience) microparticles formed without alignment. Electrodes were polished and cleaned as described previously (L.A. Zook & J. Leddy, 1998, "Experimental Studies of Diffusion on Fractal Surfaces," /. Phys. Chem. B. 102(49) 10013-10019. A suspension of Nafion® and microparticles was pipetted onto the electrode, and casting solvents were allowed to evaporate first in air and then in a vacuum desiccator for one hour. Nafion® films were also formed absent microbeads. For aligned magnetic composites, electrodes were centered inside a cylindrical magnet (6.4 cm O.D., 4.8 cm I.D., 3.2 cm, 8 lb pull, approximately 0.25 Tesla, McMaster-Carr) as casting solvents evaporated. The external field, normal to the surface, aligned beads in pillars perpendicular to the electrode surface. Once casting solvents evaporated, the aligned microparticles were trapped in the Nafion®, and the external magnet was removed. Magnetic composites (Polyscience beads/microparticles) were 85% by volume Nafion® and 15% beads/microparticles; superparamagnetic composites were either (Bangs) 75%:25% or (Dynal) 80%:20% Nafion® to beads/microparticles by volume. Thicknessess of Nafion® films and composites were calculated using the densities of wet Nafion® 1100 (1.58 g/cm³ [K.A. Mauritz, et al., 1978.

Polymer Preparation Am. Chem. Soc. Div. Poyl m. Chem. 19: 324) and microbeads as reported by the manufacturer, with thickness ranging between 3.6 and 3.8 μm.

Microscopy: Composites were characterized by scanning electron microscopy (SEM - Hitachi S-2700 and S-4000) and magnetic force microscopy (MFM - Digital Instruments Nanoscape III), a technique analogous to atomic force microscopy but performed with a magnetic tip to map magnetic fields. From SEM images, aligned magnetic microparticles formed slightly conical pillars a few beads wide. From calculations, the field around a pillar decays to magnetic field of the earth (1-2 Gauss) at 20 μm; the observed interpillar separation of 40 μm is then consistent with pillar formation driven by interpillar repulsion. Nonmagnetic microparticles (Zook & Leddy, 1998) and unaligned magnetic microparticles cluster in composites. From MFM images, the field about magnetic microbeads in the composites is preserved in the absence of an externally applied field.

Electrochemical Measurements: Electrochemical flux through composites and films was probed with various redox species. Hexaaminerutheniurn (III) chloride (Alfa), potassium ferricyanide (Aldrich), and iron (III) perchlorate (Aldrich) were used as received.

Tris(2,2'-bipyridyl)ruthenium(H) perchlorate, tris (2,2'-bipyridyl)cobalt(II) perchlorate, and tris(2,2'-bipyridyl)cobalt(III) perchlorate were prepared according to literature procedures (F.H. Burstall & R.S. Nyholm, 1952. /. Chem. Soc. 3570). Redox species were 1 mM except Fe³⁺, Co(bpy)]⁺ and Co(bpy)^ (bpy=bipyridyl), which were 2 mM. The electrolyte was 0.1

M sodium sulfate (E.M. Sciences), except for Fe³⁺ and Ru(NH )₆ ⁺ where 0.1 M nitric acid (E.M. Sciences) was used. Except for Ru(bpy)₃ , solutions were degassed with water

presaturated nitrogen. Solutions were made in 18 MΩ (Milli-Q) water.

Modified electrodes were equilibrated in solutions of the redox species and electrolyte for several hours prior to measurement. A saturated calomel reference electrode (SCE) and large platinum mesh counter electrode completed the cell. Data were collected and analyzed on 486 computers interfaced to either a Model 173 EG&G PARC Instruments Potentiostat/Galvanostat with a Model 276 interface module or a Cypress Systems Model CS-1090 Potentiostat. Scan rates ranged from 20 to 200 mV/second. Ferricyanide anion was not electrolyzed through the composites and films, consistent with defect free modifying layers. Except as noted for superparamagnetic microparticle composites, electrochemical measurements were made in the absence of an external field.

In cyclic voltammetry, current is measured as the potential is swept from a value where there is not electrolysis to a potential where the reaction is mass transport limited, and then swept back to the original potential. The maximum current (i_p, A) on the forward scan corresponds to the maximum flux and reaction rate that the system can sustain at that potential sweep rate (v, V/second). The redox species used here approach reversible (rapid) electron transfer kinetics at the electrode surface, such that a plot of i_p versus v is linear and characterized by (R.S. Nicholson & I. Shain, 1964. Anal. Chem. 36: 706; A.J. Bard & L.R. Faulkner, 1980. Electrochemical Methods, Wiley & sons, Inc., New York).

= 0.4463 i \c * D^υ2v^m nF ,,, (32)

R, T, n, F, A, and C* are the gas constant, temperature (K), number of electrons transferred, Faraday's constant, electrode area (cm²), and redox probe concentration in the phase where diffusion occurs (moles/cm³), respectively. The porosity of the layer, E, is 1 for Nafion® films and the Nafion® volume fraction for the composites. Flux of species through the composites and films is parameterized by C* and D, where C* and D (cm²/second) are the same as in Equation 11. Equation 32 yields C*D ² and given hterature or experimental values of C*, D'^/2 is determined. Throughout, flux enhancements are reported for magnetic composites as compared to Nafion® films, and are expressed as the ratio of D ² for the

magnetic composites, D} , and the Nafion® film, D_Naf , or J ^ma%₎

A rninirnum of three replicates was used, and the average flux and its standard deviation reported. Error bars on all plots denote standard deviations. EPR Experiments: ^-factors for the paramagnetic form of each redox species were determined by EPR, and are reported in Table 20. The continuous wave EPR spectrometer (Bruker EMX61) was equipped with a variable temperature unit (Bruker ER4111). Nafion® suspension and paramagnetic redox probe were mixed to yield a film with an approximate concentration of 0.3 M. Films were rinsed and well dried. EPR measurements were made at 110 K.

i. Results.

For the experimental examples presented here, self exchange reactions occur between a singlet and a radical. In each case, flux through magnetic composites is higher than through Nafion® films and composites formed with non-magnetic beads. Flux enhancements range from 16% to 2800%. Data are presented for superparamagnetic microparticle composites aligned and unaligned, and flux measured in and out of an applied field. These data are used to verify that the observed enhancements are magnetically driven, and independent of magnetohydrodynamics, composite microstructure, and mediation by contaminants in the film.

Flux Enhancement for Various Redox Probes

Cyclic voltammograms recorded at magnetic (Polyscience) composite and Nafion® film modified electrodes are shown in Figure 25 for electrolyte solutions containing

Ru(NH₂)l⁺ , R (bpy) ⁺ , Co(byp)l⁺ , and Co(bpy)l⁺ . In each case, higher currents are

generated at magnetically modified electrodes. Results for Nafion® films are shown, but voltammograms for composites of 15% non-magnetic beads are coincident, as previously observed for nonmagnetic bead content below the percolation limit of 17% (Zook & Leddy,

1998). Flux enhancements for five species are determined from i_p(v) using Equation 32, and are reported quantitatively as J ^mas/_τΛ in Table 20. range from

1.16 to 28.6, with the largest enhancements observed for the spin forbidden self exchange reactions between the singlet Co(bpy)^ and the quartet Co(bpy)₃ .

Superparamagnetic Composites Superparamagnetic microparticles sustain a magnetic field only in an externally applied field. Superparamagnetic composites were cast either in an external field to form composites with aligned pillars of microparticles, or without the external field to form unaligned microparticle clusters. Unaligned composites resemble composites formed with nonmagnetic microparticles (Zook & Leddy, 1998). Cyclic voltarnrnetric measurements were made either inside the cylindrical magnet or not, such that the external magnetic field is either on or off during measurement. The redox probe was Ru(NH₃)₆ ⁺ ; composites contained

superparamagnetic (Bangs) microparticles at 25% by volume. Flux enhancements for the aligned and unaligned composites measured in the field as compared to the aligned composites measured in the absence of the external field are

in the externally applied field is roughly twice that in the absence of the field. Enhancements for the superparamagnetic composites are lower than for the magnetic composites shown in Table 20 due to lower magnetic content. For Nafion® films and composites of nonmagnetic

microparticles (diameter = 0.91 μm), flux enhancements / ^unal,sⁿO^N / _are

V / '-' nalign, OFF

0.9+0.3 and 1.0±0.2, respectively. The external field in the absence of magnetizable beads has no measurable effect on flux.

In a second experiment with superparamagnetic (Dynal) microparticles, an external magnetic field was used to cast two composites. One electrode was held in the usual orientation such that the pillared microparticles were in contact with the electrode surface. The other electrode was inverted during the casting process such that the microparticles formed aligned pillars but the beads did not touch the electrode surface. Flux for

Ru(NH₃)₆ ⁺ was measured with an external field for the two composite electrodes and a

Nafion® film. For the Nafion® film and the aligned composite where the microparticles did not contact the electrode surface, cyclic voltammograms were superimposed, consistent with no flux enhancement generated by the external field. For the composite where microparticles contacted the electrode, approximately 45% higher i_p, was observed. The enhancement is slightly lower than for the Bangs microparticles because of lower microparticle content.

c. Specific Results. i. Magnetically Enhanced Self Exchange

Self Exchange: Self exchange, a necessary step in the postulated mechanism for magnetic enchancement, is favored by high C* and slow D_mr In Nafion® films fully exchanged with redox couple, concentrations are sufficiently high to support self exchange. The equivalent weight of Nafion (1100 g/mole of sulfonate) and the densities of Nafion®, water, and perfluorinated polymers such as polytetrafluoroethylene (2.15 g/cm³), yield Nafion® water content of 50.0% (L.A. Zook & J. Leddy, 1999. "Nafion - cold cast, thermally processed, and commercial films: Comparison of density, water content, and ewuivalent weight." In: New Directions in Electroanayl tical Chemistry II,]. Leddy, P. Vanysek, & M.D. Porter, eds. Vol. 99-5: 217-224, Seattle, WA. The Electrochemical Society, Inc.; L.A.

Zook & J. Leddy, 1996. Anal. Chem. 68: 3793) and sulfonate concentration in the water phase of 2.87 M. Nafion® has high selectivity for transition metal complexes (M.N. Szentirmay & C. R. Martin, 1984. Anal. Chem. 56: 1898.); and, for the present experimental conditions, sulfonate groups of Nafion® are charge neutralized by the transition metal complexes. Thus, in the hydrated phase of Nafion, di- and tri-cationic complexes have concentrations of 1.44

M and 0.96 M, respectively. Intermolecular distances are sufficiendy small to sustain the electron hopping necessary for efficient self exchange.

Diffusion coefficients in Nafion are slow. Literature values are available for D_ml for Ru(bpy)₃ ²⁺ (Buttry & Anson, 1981) and Fe³⁺ (J. Ye, 1992. WuliHuaxue Zuebao 8: 364), and for _mj for R (NH ) ₆₊ (L.A. Zook, 1996. Morphological Modification of Nafio for Improved

Electrochemical Flux, Ph.D. thesis, University of Iowa). For the cobalt complexes, D_nuJ was detemined from the experimental data. For the four transition metal complexes, C* of 1.44 M and 0.96 M was used for the di- and tri-cations, respectively. Iron concentration of 0.542 M in the hydrated phase was determined voltammetrically. From Equation 11, C*, and hterature values cited in the Table 20 for δ, K, and k_ex 1G), the remaining values of D_Na/and D_mt were calculated. In Table 20, values of D_m,znd D_Naf indicate that physical diffusion is sufficiently slow that self exchange in the magnetic field of earth (1 Gauss) enhances the a p p a r e n t di ffu s i o n c o e ffi c i e n t fo r th r e e o f t h e p r o b e s (Ru(NH, Ϋ⁺ , Co(bpy)]⁺ , and

) by 9 to 24%.

Flux enhancement through a self exchange process occurs whether the redox couple in solution is a radical or a singlet, because self exchange is sustained by both the radical and the singlet once voltammetry begins. As noted in Table 20, R (NH₃ )³ ₆ ⁺ and

Co(bpy) ₃ ⁺ are paramagnetic, and Ru(bpy)₃ ⁺ and Co(bpy)₃ ⁺ are diamagnetic (singlets);

from Figure 25, all exhibit flux enhancements.

Magnetic Effect: Several observations are consistent with magnetically driven flux enhancements. From MFM images of composites, magnetic microparticles maintain their field in the absence of an externally apphed field. Flux of R (NH₃ )₆ ⁺ through magnetic

composites increases both as the magnetic microparticle fraction increases and as the magnetic content of the microparticles increases for fixed microparticle fraction (S.

Amarasinghe &J. Leddy, 1993. Unpubhshed work.). For ahgned, superparamagnetic (Bangs) composites, where all experimental conditions are the same except flux is measured with and without an externally apphed field, the field enhances flux 2.3 fold. From Table 20, flux enhancements for probes with k_a (H - G) ≥ 9 x 10³ tend to increase with J- factor. Magnetohydrodynamics (MHD) can induce a convective process that increases flux, but MHD is dependent on the charge of the ion, not its ^-factor. From results in Table 20, MHD are not significant in the viscous polymer matrix. The three trications exhibit flux enhancements of 1.16, 5.3, and 19.6, whereas the dication enhancements are 3.4 and 28.6. If MHD were a significant source of the flux enhancements, enhancements would be grouped by charge. Computer models of MHD effects on mass transport in magnetic composites generate flux enhancements of <5% for the experimental conditions.

Other Considerations: Two effects unrelated to the magnetic field were found not to contribute significantly to the flux enhancement. Flux could be enhance if an impurity mediated electron transfer. Superparamagentic composites formed with microparticles proximal and distal to the electrode surface rule out mediation because the composites are chemically the same, but yield flux enhancement only when microparticles contact the electrode surface. Magnetic microstructures might also enhance flux as the pillars provide a path across the composite. Several observations exclude significant flux enhancement through microstructural effects. If the pillared structure served only as a conductive path across the matrix, no magnetic effects would be observed, and similarly sized and charge redox probes would be expected to have similar flux enhancements. That is, similar flux enhancement would be expected for Ru(bpy)₃ ⁺ and Co(bpy)₃ ⁺ , as well as R (NH₃ )₆ ⁺

and Fe(H₂θ ₆ ⁺ ; however, this is not what is observed. Flux through superparamagnetic composites ahgned or not was found to be higher with an externally apphed field than through an ahgned composite in the absence of the field. Magnetic effects on self exchange far outweigh microstructural contributions to the flux.

ii. Application of the Model.

The proposed mechanism for the observed flux enhancement is augmentation of self exchange by the magnetic field. The effect is characterized by the Dahms Ruff equation

(Equation 11), where the magnetic field component is coupled through k_l (H) as expressed in Equation 10. Prior work on nonmagnetic composites (Zook & Leddy, 1998; "Surface Diffusion in Microstructured Ion Exchange Matrices: Nafion/Neutron Track Etched Polycarbonate Membrane Composites,"/ Phys. Chem., 99(16) 6064-6073) has shown that the flux enhancement is largest at the Nafion®/microparticle interface; therefore, H is taken as the minimum field at the surface of the magnetic particle, 2100 Gauss. -4_tv(2100 Gauss) is calculated from C* as discussed above and the terms hsted in Table 20. The apparent diffusion coefficient in the magnetic composite is calculated and tabulated as D _cύc. From Table 20, the modeled flux enhancement, . / ^ma8-^calc/ _{^ ls} shown to agree well with the

experimentally determined • The fit of the model to the data is

shown in Figure 26, where the regression Hne has a correlation coefficient of 0.9985. It should be noted that no adjustable parameters were used to fit the model. Thus, the data are well modeled by a magnetically driven self exchange process.

The magnetic particles have an impact on the apparent diffusion coefficient. For Fe³⁺, the effect is smallest, in part, because D_mt is large and C* is smaller. For the four transition metal complexes, the impact of the magnetic field is substantial, enhancing the apparent diffusion coefficient over that in Nafion® from 10 fold for

to almost

900 fold for Co(bpy)₃ ⁺ . This effect is restricted to electron-nuclear spin polarization, and the results show that the effect can be large, especially when large values obtain for^, H, K, O, and C*, and D_m/ is slow. These results also show that a magnetic field can enhance electron exchange where one species is a singlet.

It is interesting to note that for the four transition metal complexes, where self exchange contributes significandy to the apparent diffusion coefficient, the hterature values for the exchange rate measured in the earth's magnetic field (k_lx( G) in Table 20) agree well with the exchange rate calculated from Equation 10 with H = 1G [k_r G) in Table 20]. Within typical errors for exchange rates, the values are the same. This suggests that, even in the absence of an apphed magnetic field, the principal reaction path for these complexes is through the magnetically facilitated relaxation process that is available in the earth's magnetic field.

It is also interesting to note that no flux enhancement is observed if the magnetic particles do not contact the electrode surface. Mass transport Umitations through the intervening Nafion® layer would limit flux. This result also suggests that there are magnetic field effects on heterogeneous electron transfer reactions. Table 20

Table 20: Summary of Experimental and Literature Parameters and Model Results for Five Redox Couples

Solution Species

Ru(bpy)j⁺(S) Co(bpy)l⁺(Q) Co(bpy)]⁺(S) ^Fe'⁺ P)

Co(bpy)³ ₃ ⁺ (S)

^Fe'⁺ (^S) Scan

2. 2.0 2.6 3.7 3.7 2.0 l ^Dma_S/ ..1/2 5.3±1.0 3.4±0.7 28.6±4.3 19.6±2.0 1.16+0.09

^(lG)(10³M 's ¹) 30-40* ιo^t 9* 9* 0.0042^§ δ (A) 6.70⁵ 14.0" 14.0¹ 14.0¹ 6.50^§

K(10-³M^'') 6.4⁵ 1.36" 1.36⁵⁵ 1.36^§§ 20^§

D _af (10-⁹ cm²/s) 2.3** 12. 0.17* 0.28« 290

D_mt (10-» cm²/s) 2.1 12* 0.13 0.25 290^tt

D_mwΛ (iσ⁹ cm²/s) 60.2 117 149 99.7 338

^ _ralc(lG)(10³M-^,s-') 38 11 15 15 120

(S) singlet, (D) doublet, and (Q) quartet; Literature values in the Table are as follows: *[l9, 21,37];^t [38]; *[21]; ⁵ [19]; "[39]; [20]; ** [33]; ^tt [32]. ^tt Values of D_Na/ determined from experimental data. ^§ξ The K value for

Ru(bpy)X was used.

4. Additional Notes

The magnetic fuel cells of the instant invention have enhanced function due to magnetic facilitation of magnetically susceptible chemical reactions that otherwise are quantum mechanically forbidden or kinetically disfavored. The magnetic fuel cells of the instant invention comprise an electrode system that includes a cathode, an anode, and a separator disposed between the anode and the cathode, as well as a magnetic field source adapted to produce a magnetic field in at least a portion of the electrode system. When an oxidant is reduced at the cathode, a fuel is oxidized at the anode, and a magnetic field is apphed by the magnetic field source in said portion of the electrode system, then between the anode and the cathode flows an electrical current, the magnitude of which is greater than the electrical current of a conventional nonmagnetic fuel cell. The magnetic field source of the instant invention can comprise any magnetic field source, including an external magnet, an internal magnet, a microstructured magnetic composite material, both an external magnet and a microstructured magnetic composite material, and both an internal magnet and a microstructured magnetic composite material. Furthermore, the magnetic field source may be a permanent magnet and/or an electromagnet. Permanent magnets may be macroscopic magnets, and/or microparticle magnets. The microparticle magnets may also be part of a composite and/or a microstructured (for example, pillared, etc.) composite material, for example, such as those disclosed in United States Patent Numbers 5,786,040, 5,817,221, 5,871,625, 5,928,804, 5,981 ,095, and 6,001 ,248 to Leddy, et al., which are hereby incorporated by reference in their respective entireties. When microstructured magnetic composite material is employed, it may be disposed on either the anode surface and/or the cathode surface, though best results occur when both are so magnetically modified.

The microstructured magnetic composite material may further comprise a first material having a first magnetism, a second material having a second magnetism, and an arrangement of said first and second materials to produce a plurahty of boundaries between the first and second materials, wherein each boundary is adapted to provide a plurahty of paths through the microstructured magnetic composite material, and has at least one magnetic field within at least one of the plurahty of paths. The magnetic fuel cells of the instant invention, which have enhanced function due to magnetic facilitation of a magnetically susceptible chemical reaction that otherwise is quantum mechanically forbidden or kinetically disfavored, may further comprise first and second chemicals, wherein each molecule of the first chemical has at least one electron having a plurahty of quantum mechanically allowed spin states, and each molecule of the second chemical has a nucleus susceptible to electron-nuclear spin polarization. In addition, the interfacial boundary, located respectively at each anode surface and each cathode surface, provides a region such that when 1) the magnetic field source is actuated to produce the magnetic field, 2) the magnetic field is effective at the interfacial boundary of the anode or the cathode, or of the cathode and the anode, and 3) the molecules of the first and second chemicals are within said effective magnetic field, then at least one electron of a molecule of the first chemical is polarized to another spin state, and the at least one spin polarized electron induces spin polarization of the nucleus of a molecule of the second chemical to effect transfer of at least one electron from one of said molecules of said first and second chemicals to the other molecule to effect said otherwise quantum mechanically forbidden or kinetically disfavored chemical reaction between molecules of said first and second chemicals. When the at least one electron of the molecule of the first chemical is polarized to another spin state, typically it is within about 10 nm of the interfacial boundary; i.e., within the magnetic field of the magnetic microparticles or other magnetic source. In the context of fuel cells, electrons can be transferred to/ from electrode (original electrode surface or added composite surface) to the molecular or ion. Similarly, when organic fuels are oxidized by a fuel cell, electrons are transferred to/ from electrode surface (original electrode surface or added composite surface) to organic molecule or ion.

The separator of the magnetic fuel cells of the instant invention typically is a proton exchange membrane (or PEM), which, for example, may be made of Nafion®, though other membranes and separators may be so employed without departing from the scope and spirit of the instant invention. As a result of the magnetic modification, there is enhanced electrical current flowing between the anode and the cathode, which, in turn, results in greater magnitudes of current and power outputs than the current and power outputs of a conventional nonmagnetic fuel cell. For example, as has been detailed herein above, at a fuel cell temperature of about 70°C and with about 1 atmosphere of oxygen, the current and power outputs for a fuel cell according to the instant invention with magnetic modification of the cathode, at about 0.35 mg/cm² of iron oxide magnets, are at least about three times the current and power outputs of a comparable fuel cell that has no magnetically modified electrodes. As another example, when a fuel cell of the instant invention operates at about 70°C and with about 1 atmosphere of oxygen, the current and power outputs when there is magnetic modification of both the cathode and anode, at about 0.20 mg/cm² of iron oxide magnets, are at least about 3.5 times the current and power outputs of a comparable fuel cell that has no magnetically modified electrodes.

It is to be understood herein that, when there is enhanced function due to magnetic facilitation of a magnetically susceptible chemical reaction that otherwise is quantum mechanically forbidden or kinetically disfavored according to the present invention, then the enhanced rate of oxidation of the fuel that may occur also includes reaction products or reaction intermediates thereof, or combinations thereof, which are oxidized at the anode. Furthermore, the fuel for practice of the instant invention is understood to be hydrogen, or any organic fuel including methanol, methane, formaldehyde, formic acid, and CO. Results for C,_.4 organic compounds have been described herein, and may include such compounds as acetaldehyde, methanol, methane, formic acid, ethanol, ethane, acetic acid, isopropanol, n-propanol, propane, propionic acid, acetic anhydride, 1- butanol, 2-butanol, tertiary butanol, butanoic acid, or combinations thereof; however, any organic compound may be so employed in practice of the instant invention. Similarly, the rate of reduction of the oxidant that may be increased at the cathode is understood to include also reaction products or reaction intermediates of the oxidant(s). Furthermore, any oxidant may be used in practice of the instant invention, including without limitation such oxidants as air, oxygen, a peroxide, or combinations thereof.

When microparticle magnets are used, they may be made of iron oxide, samarium cobalt, neodymium iron baron, other magnetic materials or combinations thereof. When samarium cobalt microparticle magnets are used, they may be coated by a silanization process to enhance stability; and similar procedures can be used for other magnetic materials as well. Other magnetic materials may also be so coated. When microstructured magnetic composite material is used and it comprises iron oxide microparticles or other magnetic microparticles, then the microstructured magnetic composite material may be disposed on an electrode surface in an amount of up to at least about 0.50 mg per square centimeter. However, even higher composite material coatings of iron oxide magnets also fall within the spirit and scope of the instant invention. When samarium cobalt magnets are used in a microstructured magnetic composite material, coatings up to at least about 0.2 mg per square centimeter are possible.

The magnetic field source of the present invention can be any one or any combination of a plurahty of forms. For example, permanent microparticle magnets can be employed on the surface of an electrode. Alternatively, non-permanent microparticles magnets (for example, superparamagnetic magnets, such as nickel) may be used on the surface of an electrode in combination with an external magnetic field source that induces magnetic fields in the non-permanent magnets. The ernbodirnent functions as long as the field of the permanent external magnet is within range; no net magnetic field is sustained once the external permanent magnet is removed.

Claims

We claim:

1. A magnetic fuel cell capable of effecting indirect reformation of an organic fuel, comprising: a cathode; an anode; a separator disposed between the anode and the cathode; and a magnetic field source adapted to produce a magnetic field in at least a portion of at least one of said cathode and said anode.

2. The magnetic fuel cell as claimed in claim 1, wherein the magnetic field source comprises a microstructured magnetic composite material.

3. The magnetic fuel cell as claimed in claim 1 , wherein the magnetic field source comprises both an external magnet and a microstructured magnetic composite material.

4. The magnetic fuel cell as claimed in claim 3, wherein the magnetic composite comprises microparticles that are permanent magnets.

5. The magnetic fuel cell as claimed in claim 4, wherein the magnetic composite comprises microparticles that are superparamagnetic.

6. The magnetic fuel cell as claimed in claim 1, wherein the anode has an anode surface, the cathode has a cathode surface, and a microstructured magnetic composite material is disposed on at least one of said anode surface and said cathode surface.

7. The magnetic fuel cell as claimed in claim 6, wherein said anode surface and said cathode surface comprise microstructured magnetic composite material.

8. The magnetic fuel cell as claimed in claim 1 , further comprising: first and second chemicals, wherein each molecule of the first chemical has at least one electron having a plurahty of quantum mechanically allowed spin states, and each molecule of the second chemical has a nucleus susceptible to electron-nuclear spin polarization; and an interfacial boundary located respectively at said anode and said cathode, wherein when said magnetic field source is actuated to produce a magnetic field, said magnetic field is effective at said interfacial boundary of the anode or the cathode, or of the cathode and the anode, and molecules of said first and second chemicals are within said effective magnetic field, then at least one electron of a molecule of the first chemical is polarized to another spin state, and said at least one spin polarized electron induces spin polarization of the nucleus of a molecule of the second chemical to effect transfer of at least one electron from one of said molecules of said first and second chemicals to the other molecule to effect an otherwise quantum mechanically forbidden or kinetically disfavored chemical reaction between molecules of said first and second chemicals.

9. The magnetic fuel cell as claimed in claim 3, wherein when said separator comprises Nafion®, and yields enhanced electrical current flows between the anode and the cathode.

10. The magnetic fuel cell as claimed in claim 1, wherein the organic fuel comprises one of acetaldehyde, methanol, methane, formic acid, ethanol, ethane, acetic acid, isopropanol, n-propanol, propane, propionic acid, acetic anhydride, 1- butanol, 2-butanol, tertiary butanol, and butanoic acid.

11. The magnetic fuel cell as claimed in claim 1 , wherein the separator comprises a proton exchange membrane (PEM).

12. The magnetic fuel cell as claimed in claim 11, wherein the PEM comprises Nafion®.

13. The magnetic fuel cell as claimed in claim 1 , wherein the magnetic field source comprises at least one of a permanent magnet and an electromagnet.

14. The magnetic fuel cell as claimed in claim 13, wherein said at least one permanent magnet comprises a plurahty of microparticle magnets.

15. The magnetic fuel cell as claimed in claim 1, wherein said field source comprises samarium cobalt microparticle magnets.

16. The magnetic fuel cell as claimed in claim 15, wherein said samarium cobalt microparticle magnets are coated by a silanization process.

17. The magnetic fuel cell as claimed in claim 1 , wherein the magnetic fuel cell resists passivation due to carbon monoxide.

18. The magnetic fuel cell as claimed in claim 11 that utilizes a synthetic reformate, wherein said fuel cell comprises one of an H₂/0₂ PEM fuel cell and an H₂/air PEM fuel cell that substantially resists passivation due to a synthetic reformate.

19. The magnetic fuel cell as claimed in claim 18 wherein said synthetic reformate comprises hydrogen and 100 parts per million of carbon monoxide.

20. The magnetic fuel cell as claimed in claim 1 , wherein said magnetic field is arranged such that the fuel cell substantially resists passivation for cell temperatures from at least about 25 degrees C to at least about 70 degrees C.