Mediated Fuel Cells: Soluble Redox Mediators and their Applications to Electrochemical Reduction of O₂ and Oxidation of H₂, Alcohols, Biomass, and Complex Fuels

Abstract

Mediated fuel cells are electrochemical devices that produce power in a manner similar to that of conventional proton exchange membrane fuel cells (PEMFCs). They differ from PEMFCs in their use of redox mediators dissolved in liquid electrolyte to conduct oxidation of the fuel or reduction of the oxidant, typically O₂, in bulk solution. The mediators transport electrons (and often protons) between the electrode and the catalysts or chemical reagents in solution. This strategy can help overcome many of the challenges associated with conventional fuel cells, including managing complex multi-phase reactions (as in O₂ reduction) or the use of challenging or heterogeneous fuels, such as hydrocarbons, polyols and biomass. Mediators are also commonly used in enzymatic fuel cells, where direct electron transfer from the electrode to the enzymatic active site can be slow. This review provides a comprehensive survey of historical and recent mediated fuel cell efforts, including applications using chemical and enzymatic catalysts.

Graphical Abstract

1. Introduction

1.1. Proton-exchange-membrane fuel cells (PEMFCs) and redox flow batteries (RFBs): Context for mediated fuel cells (MedFCs).

Growing worldwide energy demands continue to motivate the exploration and development of diverse energy storage¹ and conversion² technologies that could support broader use of renewable energy sources, such as solar, wind, hydroelectric, and biomass.³ Electrochemical technologies are a major focus of these efforts, with prominent examples including solid-state^4–7 and redox flow batteries,^8,9 supercapacitors,^10,11 and (photo)electrochemical production of chemicals, ranging from hydrogen (H₂) to carbon-based molecules (formic acid, syn gas, MeOH, among others).^12,13 Chemicals produced from the latter processes may be stored or transported to other locations for power generation, ideally with fuel cells to leverage their higher energy efficiency relative to conventional combustion-based power generation.¹⁴ This review article surveys historical and recent advances in mediated (or “indirect”) electrochemical processes for power generation. These processes are conceptually related to proton exchange membrane fuel cells (PEMFCs) and redox flow batteries (RFBs) (Figure 1), while incorporating features distinct from these established technologies.

Figure 1.

General schematic of an A) PEM fuel cell, B) flow battery in discharge mode with [A] representing the anodic (e.g. low potential) redox couple and [C] representing the cathodic (e.g. high potential) redox couple, and C) mediated fuel cell with mediators for both anode and cathode.

PEMFCs are the most common class of fuel cells for low-temperature power generation.¹⁵ These devices convert energy from chemical bonds into electricity by oxidizing a chemical fuel at the anode and reducing an oxidant at the cathode. Conventional PEMFCs use H₂ as the fuel¹⁶ and O₂ as the oxidant. The electrons from H₂ are transported from the electrode through the electrical circuit, while the protons are transported through a membrane electrolyte, consisting of a sulfonated fluoropolymer (e.g., Nafion®), to the cathode where they combine with O₂ and the electrons to produce water (Figure 1A). A related class of fuel cells, anion exchange membrane fuel cells (AEMFCs), utilize membrane electrolytes that conduct anions, such as Cl⁻ or OH⁻, and have been the focus of growing interest.¹⁷

Water management is a key challenge in conventional PEMFCs.¹⁸ Water must be present within the ionic polymer membrane to support efficient proton transport from the anode to the cathode,¹⁹ and this requirement is met by using humidified sources of H₂ and O₂ gases. Meanwhile, the water generated at the cathode from O₂ reduction must be removed to avoid flooding of the catalyst. The issues related to water management limit the temperature range available for effective operation of PEMFCs (typical operation ≤ 100 °C). Membrane-electrode assemblies²⁰ and bipolar plates²¹ have been engineered to address these challenges, while ensuring effective delivery of gases to the catalyst surface, efficient coupling of the catalytic reactions of H₂ and O₂ with the proton- and electron-transport processes, and precise management of water. Fuel cell technologies continue to improve, with recent advances enabling access to current and power densities of ≥ 2.5 A/cm² and 1.2 W/cm², respectively.²²

The requirement for high loadings of Pt catalyst at the cathode to achieve optimal performance represents one of the most significant impediments to conventional PEMFCs.²³ Pt catalysts are also used at the anode, but the higher catalytic rates for H₂ oxidation relative to O₂ reduction mean that much less catalyst is required at the anode. Advances have been made in the development of non-platinum group metal (non-PGM) catalysts for O₂ reduction;^24–26 however, these catalysts typically have lower intrinsic activity than Pt and, therefore, require higher catalyst loadings to achieve the similar current densities. The higher catalyst loading introduces additional limitations associated with the increased electrical resistance from the thicker catalyst layer and restricted mass transport of O₂ and water through the electrode. Consequently, commercial applications of non-PGM cathode catalysts have thus far been restricted to low power-density applications, in which the complications arising from higher catalyst loadings are less problematic.^27,28

Alongside the development of PEMFCs that use H₂ as a fuel, considerable effort has been directed toward the development of fuel cells capable of using liquid fuels. Such devices avoid challenges associated with the storage, transport, and distribution of H₂. Direct methanol fuel cells (DMFCs),²⁹ which oxidize MeOH (supplied as a concentrated aqueous solution) to CO₂ at the anode, are among the most prominent applications of liquid fuel cells. The lower rates of MeOH oxidation relative to H₂ oxidation, however, leads to considerably lower power densities in DMFCs (< 120 mW/cm²).³⁰ The use of fuels more complex than MeOH, such as sugars or other biomass-derived molecules, and limitations on the practical concentrations of these fuels in the aqueous solutions delivered to the anode, can further constrain catalytic rates and power densities in liquid fuel cells.³¹ Liquid fuel cells must be operated under conditions that avoid cathode flooding, which can arise from transport of water from the fuel solution through the membrane as protons are transported from the anode to the cathode. Complications of the type described here account for the limited commercial application of PEMFCs with fuels other than hydrogen.

Redox flow batteries (RFBs) represent the second class of electrochemical devices directly relevant to the subject matter of this review article. RFBs store electrical energy by transferring electrons to and from molecules or ions dissolved in solution (Figure 1B). Typical systems feature two tanks of redox couples, one containing a cathodic species that has a high reduction potential and the other with an anodic, low-potential species. Electrical energy is stored and generated by flowing the redox couple solutions through the cathode and anode compartments, which are separated by a semi-permeable membrane that enables selective transfer of protons or other ions while preventing mixing of the mediators. The underlying concepts of redox flow batteries were first demonstrated in the late 1800s,³² with significant technological development beginning in the 1960s.³³ Commercial installations have emerged in recent years.³⁴ Examples of the redox couples at the anode/cathode used in commercial systems include Cr/Fe, Zn/Br, and V/V.⁸ The systems using Fe, Cr, Zn and Br species benefit from the low cost of the chemicals, but the all-vanadium redox flow batteries exhibit the longest lifetimes and are the most widely used. Challenges associated with these technologies include the high toxicity and corrosivity of the electrolyte (particularly for V- and Br-based systems), the limited number of redox couples that operate near the limiting potentials for H⁺/H₂O reduction and H₂O/OH⁻ oxidation for aqueous electrolytes, and the limited material supply and high cost of V. All-organic redox species have been the focus of much recent attention as a means to address these limitations. Examples include derivatives of quinones, nitroxyls, and viologens (Figure 2).^9,35 The synthetic modularity of these molecules, which allows their redox potentials and other physical properties (e.g., solubility, stability) to be tuned, greatly expands the scope of possible compounds that may be considered. Water-soluble redox couples are especially appealing because water is low-cost solvent with inherently low solvent resistivity;^36–38 however, non-aqueous RFBs have also been the focus of major research efforts.³⁹

Figure 2.

Structures of representative organic mediators examined for redox flow battery applications.

1.2. Mediated fuel cells (MedFCs): Overview and opportunities.

The above discussion of PEMFC and RFB technologies provides the relevant context for the introduction of mediated fuel cells (MedFCs, Figure 1C), which are the focus of this review article. The latter devices feature dissolved mediators analogous to the mediators used in RFBs; however, catalytic regeneration of the mediators in an off-electrode compartment enables sustained production of electricity analogous to the operation of PEMFCs. Mediator regeneration is achieved by supplying a fuel to reduce the low-potential mediator and O₂ to oxidize the high potential mediator, typically employing a catalyst for both processes. Mediated fuel cell systems are distinct from RFBs because they can operate continuously, as long as the supplies of fuel and oxidant are maintained.

Several factors impact the utility of redox mediators for use in mediated fuel cells. First, the redox potential of the mediator should be as close as possible to the formal potential of the fuel or the oxidant. For a “conventional” system using H₂ as the fuel and O₂ as the oxidant, the anode mediator should have a redox potential near 0.0 V vs. RHE (Reversible Hydrogen Electrode; i.e., adjusting for the pH of the system), and the cathodic mediator a redox potential near 1.23 V vs. RHE. Additionally, the redox mediator should display fast electron transfer kinetics, to minimize any overpotentials caused by kinetic losses in the system, it needs to be stable in both oxidized and reduced states to enable long-term performance, and it must be highly soluble in aqueous solution for the cell to maximize the current densities.

MedFCs have several potential advantages over conventional fuel cells that warrant their consideration for further development. These features include the following:

1). Enabling the use of “challenging” fuels.

By moving fuel oxidation away from the anode, a greater number of fuel options could become available, and more diverse catalysts may be considered. For example, the use of H₂ feeds containing carbon monoxide, sulfur-, or nitrogen-based contaminants has been demonstrated,⁴⁰ and this advantage was one of the original motivations for investigating MedFCs.⁴¹ Another advantage is that fuels that are too complicated to oxidize at an electrode under mild conditions (<150 °C) may be considered. Examples of these alternatives include alcohols more complicated than MeOH and heterogeneous fuel sources, such coal or biomass-derived cellulose or lignin, which are challenging fuels for direct activation at an electrode.

2). Overcoming challenges with the oxygen reduction reaction (ORR).

A major motivation in previous studies of mediated fuel cells has been the replacement of Pt, which is required in relatively high loading in conventional PEM cathodes, with non-PGM alternatives.⁴² The high loadings of Pt are necessary to achieve sufficiently high activity within the narrow spatial dimension at the gas-diffusion electrodes. By moving the catalytic O₂ reduction process into an off-electrode compartment, the ORR may be optimized independently of the electrochemistry. The lower intrinsic activity of non-PGM catalysts relative to Pt is less problematic in this configuration. For example, mass-transport restrictions are significantly reduced when the catalyst does not need to be in direct contact with the electrode and membrane, and higher mass loadings of the non-PGM catalyst may be integrated in an off-electrode reactor in a manner that does not impede system performance. The electrochemical reaction of the mediator features a simpler two-phase (liquid/solid) system at the electrode, rather than the more complicated three-phase (gas/liquid/solid) system present in conventional fuel cells. While there is still a three-phase reaction occurring at the off-electrode catalyst, many of the engineering and materials constraints are relaxed. For example, it is not essential to use a conductive catalyst support, and there are fewer limitations on the amount of catalyst that can be used, allowing higher loadings of less active and less expensive catalysts than Pt to be used.

3). Opportunities to improve overall system performance.

At first glance, mediated fuel cells appear to increase the complexity of a fuel cell stack, due to the requirement for extra plumbing and pumps to operate the mediator-regeneration loop. In fact, however, they can simplify the balance of plant accessories. PEMFCs are typically limited to operating at or below 100 °C and require humidification of the gases to avoid dehydration of the membrane. Since mediated fuel cells use aqueous mediator solutions, the membrane remains well-hydrated and gas humidification is unnecessary. Additional benefits arise from using aqueous solutions, which can absorb heat generated from the oxidation and reduction reactions, thereby facilitating temperature control of the fuel cell stack. Finally, mediated cathode systems are ideally suited for pairing with liquid fuels, including both conventional and mediated anode processes, because they are not susceptible to the flooding from water convection that can occur with gas-diffusion cathodes in PEMFCs.

4). Reduced performance losses from activation polarization.

Activation polarization is the loss in cell voltage at low current densities due to the kinetic barriers for electron transfer to the species to be oxidized or reduced. The particularly high kinetic barriers associated with catalytic reduction of O₂ leads to substantial losses in the cell potential when O₂ is directly reduced at an electrode. Similar issues arise at the anode, when using fuels more complex than H₂, such as methanol. Redox mediators, on the other hand, may be selected to have fast electrochemical kinetics and thereby minimize activation polarization losses.

1.3. Scope of coverage and notes on nomenclature and symbolic notation used in this review.

Redox mediators have been employed in a wide range of electrochemical devices and applications, many of which have been the subject of previous review. Important examples include electrochemical synthesis of organic and inorganic chemicals,^43–46 in addition to diverse energy-based technologies beyond flow batteries,^8,9 including dye-sensitized solar cells,^47,48 solid-state batteries (e.g., “redox targeting”),^49,50 decoupled water splitting,⁵¹ electrochemical capacitors,⁵² and solar flow batteries.⁵³ This review covers mediated fuel cells (MedFCs) that feature a mediated cathode for O₂ reduction, a mediated anode for H₂, alcohol, or biomass oxidation, or a doubly mediated system, wherein mediators are used for both half-reactions of the fuel cell. Some aspects of these reactions have been reviewed previously;^41,42,54 however, significant advances in recent years, together with the partial coverage of these previous reviews, provided motivation to prepare this comprehensive survey. Only systems that are designed or intended to produce power (galvanic cells) are covered in depth, though reference is occasionally made to relevant electrolysis systems. Most of the examples consist of full fuel cells, but relevant half-cell studies are also discussed. Enzymatic fuel cells^55–57 using soluble mediators are included, but microbial fuel cells^58,59 are not presented. While microbial fuel cells feature certain advantages, including the capability to produce their own mediators and potential for genetic engineeringfor enhanced performance,⁶⁰ their added complexity makes it is difficult to compare them directly to the other systems.

Historically, mediated fuel cells have been given the name “chemically regenerable redox fuel cells”, or CRRFCs,^{41,42,54,61–64} to distinguish them from “redox fuel cells”, which are now typically called redox flow batteries (RFBs). With the shift in the flow battery nomenclature, we have chosen to adopt the simplified term “mediated fuel cell” or “MedFC” for these systems (avoiding the term “MFC”, which is commonly employed for microbial fuel cells).

To simplify the notation used to describe mediated fuel cell systems, we have adopted the following notation: fuel|anode catalyst|anode mediator||cathode mediator|cathode catalyst|oxidant, whereby the double bar (||) refers to the membrane separating the anode and anode compartments and the single bar (|) separates the different components within the mediated anode or cathode systems. Certain studies only investigated one of the half-cell systems, or a system was conducted in the absence of a mediator or catalyst. In such cases, certain components are omitted from the symbolic notation. The distinction between a “mediator” and a “catalyst” is not always well defined in multi-component electrochemical mediator systems. We have attempted to be consistent in use of the term “mediator” for the major redox-active species that undergoes electron-transfer at the electrode, and the term “catalyst” for the species that undergoes the redox reaction with the fuel or oxidant in a redox cascade. Ambiguities can arise, for example, in single-component mediator systems in which the “mediator” also serves as the “catalyst”. A table containing information for all reported mediated fuel cells using chemical catalysts is given in Appendix I. All reports of mediated fuel cells using enzymatic catalysts is provided in Section 3.2.

2. The Cathode: Oxygen Reduction

The reduction of O₂ is the predominant cathodic reaction for fuel cell technology, owing to the availability of O₂ and formation of water as a benign product. State-of-the-art catalysts in PEMFCs and DMFCs are composed of Pt or Pt alloys that exhibit the best combination of high activity, low overpotential, long-term stability, and selectivity for 4 e⁻ reduction of O₂ to H₂O.^65,66 While optimal performance is achieved with these Pt-based catalysts, their high costs continues to be a limiting factor in PEMFC commercialization. The cost of Pt has been forecasted to account for over 40% of the stack cost.⁶⁷ Eliminating or decreasing the loading of Pt or Pt-group metals is a high priority, and heterogeneous electrocatalysts such as Pt-alloys and non-PGM electrocatalysts, including metal-free carbon-based materials, have been developed.^68,69 The first commercial PEMFC system utilizing non-PGM catalysts was announced in 2017.²⁸ Nonetheless, no catalysts have thus far achieved the activity and stability metrics of Pt or Pt alloys, and alternate solutions continue to be the focus of active study. Mediated O₂ reduction was the focus of considerable historical investigation, and new advances have emerged in recent years. The following sections survey the different mediators that have been reported for mediated ORR, including historical and recent examples, and systems implemented in fuel cell devices and fundamental studies.

2.1. Br⁻/Br₂ mediator

The bromide/bromine redox couple has been widely used in flow battery research due to its high redox potential (1.06 V vs. NHE), fast electrode kinetics, and high solubility of the relevant species in different redox states. The potential is within 200 mV of the thermodynamic potential for O₂ reduction to H₂O and, therefore, could provide a low-overpotential strategy for mediated ORR. Fuel cells reported with bromine-based mediators for the ORR are shown in Table 1.

Table 1.

Reported systems using Br mediators for O₂ reduction

Anode			Cathode
fuel	catalyst	mediator	mediator	catalyst	ref.
C	–	Sn4+/Sn2+	Br−/Br2	NOx	70
H2	Pd black	TiO2+/Ti3+			40
H2	Pt	–			73

The first report of a MedFC utilizing bromine-based cathode mediators was published in 1955 by Posner.⁷⁰ In this work, other cathodic mediators were examined, but they exhibited low cell potentials. A variety of catalysts and mediators were screened, and the use of bromine in combination with NO₂ showed promise (Figure 3). The anode contained chars, coke, or coal as the fuel and a variety of metal ions were tested as anodic mediators. Tests were also performed without NO_x/O₂-based regeneration, and the results showed that Br₂ at the cathode shows minimal polarization (<0.05 V) up to a cell current density of >50 A/ft² (i.e., ~54 mA/cm²) with 1.2 N Br₂ and 0.7 N HBr in 3 N HCl. Challenges with long-term stability of the cathode system was acknowledged, but >8 h of sustained activity was demonstrated.

Figure 3.

Redox sequence for bromine-mediated cathode with a NO_x-based catalyst (||Br₂|NO_x|O₂). For relevant studies, see references 40, 70–73.

In the same year, Posner and Merton-Bingham studied the mechanism of NO_x-catalyzed autoxidation of HBr.⁷¹ The authors proposed a 6-step sequence (Figure 4A), with the slow step involving the reaction of NOBr in solution with an unknown, gas-phase species resulting from the reaction of HNO₂ and HBr (species “X” in Figure 4A). Later studies by Reneke noted reproducibility problems with the analytical techniques for bromine quantification employed by Posner and Merton-Bingham.⁷² Reneke also suggested that the mechanism of Posner and Merton-Bingham was unnecessarily complex and proposed an alternative 3-step mechanism (Figure 4B). Calculations from Reneke’s report propose that a 10-kW fuel cell using this system to regenerate Br₂ would require a regeneration tower over 130 ft in height and 1.5 ft in diameter. These dimensions likely prevent large-scale implementation of this technology; however, the report from Posner, as well as subsequent studies, show that fuel cell operation is possible.

Figure 4.

Proposed mechanisms for HBr oxidation to Br₂ by NO_x catalysts.

In 1959, Carson and Feldman from the General Electric Company described a 100 W fuel cell stack, using seven cells connected in series.⁴⁰ The cathodic mediator was Br₂/Br⁻, again using nitric oxide to promote aerobic regeneration of Br₂. The fuel was H₂, which reduced a Ti^IV mediator to Ti^III (i.e., TiO²⁺/Ti³⁺) using Pd-black as an off-electrode catalyst (H₂|Pt|TiO²⁺||Br₂|NO_x|O₂) (see Section 3.1.2 for discussion of the anode performance). The cells were operated at 80 °C, and an open-circuit potential of around 0.95 V was obtained. Continuous operation of the system was demonstrated for over 16 days, but the regeneration of the mediators was not sufficient to operate the system at its peak activity for longer than a few minutes. Additionally, the overall efficiency was low, due to high internal resistance of the cell, and the overall device was quite large, due to a complicated liquid delivery system to each cell. Carson and Feldman noted that, while this system demonstrated technical feasibility, several improvements would be necessary for commercialization, including the use of non-volatile mediators/catalysts on the cathode.

In 1964, Juda, at Ionics, Inc., patented a fuel cell using bromine-based mediators.⁷³ H₂ oxidation at a heterogeneous Pt electrode was used as the anodic reaction. The cathode compartment contained a solution of KBr, HBr and Br₂ with NO₂ as the catalyst. An open circuit potential of 0.98 V and an operating potential of 0.7 V at 10 mA/cm² at room temperature were obtained. Juda compared these results to a system using similar loadings of KI, HI, and I₂ without a catalyst, which generated an open circuit potential of 0.5 V and an operating potential of 0.3 V at a current density of 10 mA/cm². No sustained operation of these cells was reported.

2.2. Vanadyl and vanadium-containing polyoxometalate cathodic mediators

The VO²⁺/VO₂⁺ redox couple is the most widely studied cathodic mediator in mediated fuel cells. Some major advantages of this system include the high redox potential (~1.0 V vs. NHE), its non-volatile nature, good stability, and reasonably high solubility (> 2 M). These features underlie the widespread implementation of the same mediator system for redox flow battery applications.⁷⁴ A catalyst is required to increase the rate of VO²⁺ reoxidation by O₂. Both NO_x-based catalysts and heteropolyacids (HPAs; also called polyoxometalates, POMs) have been used for this purpose. Table 2 lists the various implementations reported for cathodic vanadium mediators.

Table 2.

Reported fuel cell systems using V-based mediators for O₂ reduction.

Anode			Cathode
fuel	catalyst	mediator	mediator	catalyst	ref
H2	Pt	–	VO2+/VO2+	HCl	75
–	–	V2+/V3+		NOx	79
H2	Pt	Ti3+/TiO2+			80
CH4, ethylene glycol, coal	Pt	Fe2+/Fe3+			82–84
biomass	–	Fe2+/Fe3+			85, 86
sugars	Pt	V3+/VO2+		Pt	81
H2	Pt	H4SiW12O40		POM	76–78
	Pt	–	V-containing POM		63, 95–102,104
	Pt/C	AQ			105
sugars	POM	Fe2+/Fe3+			110
biomass	POM				107, 111, 112
alcohols					108
lignin					113
coal					109

The first example of VO²⁺/VO₂⁺ mediators for electrocatalytic O₂ reduction was patented by Schaefer and Kordesch at Union Carbide Corporation in 1966.⁷⁵ While details of the invention are somewhat limited, they described a cell using an H₂ gas anode and NaVO₃ dissolved in a strong HCl electrolyte (19–39% by weight). When the cathode was supplied with O₂, it was operated for over 24 h at 80 °C and sustained a potential between 1.5 and 1.7 V vs. cadmium (conversion to a standardized reference was not given) at a current density of 50 mA/cm². An HCl electrolyte was necessary for regeneration; when H₂SO₄ was used as the electrolyte, addition of HBr and HNO₃ was necessary to achieve sustained activity.

Oei, working at Ford, reported the use of a VO²⁺/VO₂⁺ mediated cathode. The initial report investigated the performance of various electrolytes in a RFB context (i.e., no regeneration of either redox couple by O₂ or a fuel): VO²⁺/VO₂⁺ as the cathodic redox couple and Sn²⁺/Sn⁴⁺, Fe²⁺/Fe³⁺, or Cu/Cu²⁺ as the anodic redox couples.⁶³ Initial tests optimized the flow-field configuration and electrode materials. The different electrode materials displayed similar performance for the vanadium redox process, but substantially different activity for the Sn²⁺/Sn⁴⁺ redox couple. Later in the same year, Kummer and Oei reported a mediated cathode system with VO²⁺ and POMs as the mediator/catalyst composition (POM = H₃PMo₁₂O₄₀ or H₅PMo₁₀V₂O₄₀). A silicon-containing POM was used as the anodic mediator (H₂|Pt|Si-POM||VO²⁺|POM|O₂; see Section 3.1.4 for discussion of anodic performance).^76–78 Polarization curves were obtained using 0.8 M VOSO₄ and 0.28 M H₃PMo₁₂O₄₀ at a pH 0.75 (Figure 5). The highest power density obtained was ~38 mW/cm². Reasonable iR-free voltages were obtained, but the internal resistance of the cell contributed to significant losses in cell potential (0.83 V with iR-correction vs. 0.55 V without iR-correction at ~80 mA/cm²). Slow regeneration of the cathode mediator system caused the potential of the cathode compartment to drop from 468 mV to 330 mV over 900 minutes of operation. A pH change was observed in the cathode compartment, which was attributed to vanadium migration through the membrane. The authors commented that long-term performance would be determined by the size of the regenerators, and to achieve reasonable rates, the regenerator would be too large for commercial vehicle use. For example, a 20-kW engine would weigh over 5,000 pounds.

Figure 5.

Redox sequence proposed for V-mediated cathode with POM catalyst for O₂ reduction (||VO²⁺|POM|O₂). For relevant systems using this mediated cathode, see references 76–78.

In 1985, Kummer and Oei reported further optimization of their mediated fuel cell design.⁷⁹ To avoid the substantial performance decreases observed in VO²⁺ oxidation catalyzed by POMs, they shifted to the use of NO_x gases as a catalyst (Figure 6A). Over a 12 h experiment, the cathodic performance remained nearly constant (Figure 6B), showing more stable performance than previous reports.⁶³ To prevent loss of NO_x species, a static pressure of O₂ was used without bubbling gas into the cathode solution. When using a V²⁺/V³⁺ anode solution, power densities of 73.5 mW/cm² were achievable.

Figure 6.

(A) Redox sequence for vanadium-mediated cathode with NO_x catalyst for O₂ reduction (||VO²⁺|NO_x|O₂). For systems featuring this mediated cathode, see references 79–86. (B) Sustained performance curve of a mediated cathode utilizing a VO²⁺/VO₂⁺ redox mediator, reoxidized with NO_x-based catalysts. The anode solution contained a V²⁺/V³⁺ redox mediator which was not regenerated during the experiment (V³⁺||VO²⁺|NO_x|O₂). Cell was discharged through a 5-ohm external load with a flow rate of 230 mL/min. Adapted from ref. 79. Copyright 1985, Chapman and Hall Ltd.

The VO²⁺/VO₂⁺ redox couple has been adopted by other groups in various fuel cell studies focused on anodic reaction chemistry. Folkesson constructed a cell using a VO²⁺/VO₂⁺ redox mediator regenerated aerobically using NO_x as a catalyst with a variety of anode redox mediators for H₂ oxidation (see Sections 3.1.2, 3.1.3, and 3.1.7 for further discussion of the anodic chemistry).⁸⁰ Folkesson demonstrated reasonably stable cell performance generating around 15 mW/cm² using the VO²⁺/VO₂⁺ and Ti³⁺/TiO²⁺ redox couples. A later report mentions the aerobic oxidation of VO²⁺ catalyzed by Pt for a fuel cell using a V³⁺/VO²⁺ anode for the oxidation of sugars in a fuel cell stack capable of producing 10 kW of power; however, no evidence for the aerobic oxidation of VO²⁺ was presented.⁸¹ Whitesides and coworkers have used vanadyl-based mediators paired with NO_x catalysts in a variety of applications (c.f. Figure 6A). Their initial report focused on the use of methane as a fuel, mediated by Fe²⁺/Fe³⁺ and catalyzed by Pt black (CH₄|Pt|Fe²⁺||VO²⁺|NO_x|O₂; see Section 3.1.3 for further discussion of the anode).⁸² Methane oxidation occurred under forcing conditions in a high-pressure reactor (120 °C, 54 atm CH₄), and the Fe²⁺ in the resulting aqueous reaction mixture was oxidized at the anode (Figure 7). Polarization curves were obtained, but the performance of the overall system was low, due to engineering difficulties in cell construction and the large potential losses associated with use of the Fe²⁺/Fe³⁺ redox couple as the anodic redox mediator (~600 mV difference between thermodynamic potential of methane oxidation to CO₂ and the Fe²⁺/Fe³⁺ redox potential). This approach was later extended by Whitesides and coworkers to the use of ethylene glycol as a fuel, generating a power density of 9.9 mW/cm³,⁸³ and to the use of solid carbonaceous fuels, including coal (see Section 3.1.3 for further discussion).⁸⁴

Figure 7.

Schematic of doubly mediated fuel cell (CH₄|Pt|Fe²⁺||VO²⁺|NO_x|O₂) using methane as fuel developed by Whitesides and coworkers. Adapted from ref. 82. Copyright 1994, American Association for the Advancement of Science.

Deng and coworkers reported a fuel cell using VO²⁺/VO₂⁺-based mediators and an HNO₃ catalyst to mediate O₂ reduction on the cathode paired with an Fe²⁺/Fe³⁺ anodic mediator for biomass (wheat straw) oxidation.⁸⁵ Power densities of over 100 mW/cm² were achieved with 10 g of wheat straw in 100 mL of electrolyte (10 mL HCl and 30 g FeCl₃ in H₂O), and a stable current density of 320 mA/cm² was achieved at a voltage of 0.3 V for over 12 h (see Section 3.1.3 for discussion of the anode). Further studies by Deng and coworkers compared the performance of a mediated cathode utilizing a VO²⁺/VO₂⁺ mediator with HNO₃ as a catalyst with the performance of a mediated cathode using an H₁₂P₃Mo₁₈V₇O₈₅ POM mediator (see section 3.1.4 for discussion of the relevant anode).⁸⁶ Glucose and other biomass sources were used as the fuel with an FeCl₃ anodic mediator (glucose|Fe³⁺||VO²⁺|NO_x|O₂ vs. glucose|Fe³⁺||POM|O₂). The cell utilizing the vanadium/NO_x system achieved a maximum power density of approximately 90 mW/cm², whereas the POM-mediated cell only achieved a maximum power density of approximately 35 mW/cm². This difference was attributed to the higher open circuit potential achieved by the vanadium-based system compared to the POM-based system.

As discussed above, Oei demonstrated that mixtures of POMs and VO²⁺ undergo aerobic oxidation. POM/VO²⁺ reactions with O₂ have been intensely studied for decades, particularly for Mo/V-containing POMs.^87–94 These reactions are quite complex, due to the complicated speciation of POMs in acidic solutions. This complexity can be observed in the ³¹P and ⁵¹V NMR spectra of H₁₂P₃Mo₁₈V₇O₈₅, in which a multitude of different POM isomers and structures are observed (Figure 8).

Figure 8.

³¹P and ⁵¹V NMR spectra of 0.25 M solution of H₁₂P₃Mo₁₈V₇O₈₅. Numerals 1, 2, 3, and 4 correspond to collection of lines of HPA-x Keggin POMs H_3+xPMo_12-xV_xO40, where x = 1–4, see reference 87. Adapted from ref. 91. Copyright 2008, Elsevier B.V.

Starting in the mid-2000s, Acal Energy Ltd., a company based in the UK, filed numerous patents describing the use of POMs as combined mediator/catalysts for the electrocatalytic ORR for fuel cells. In contrast to the Oei work discussed above, Acal used a conventional Pt/C H₂ anode. POMs have a variety of promising characteristics, including their inexpensive components, ability to operate at high potentials, and thermodynamic stability.^95,96 Several POM species with promising activities were identified. Higher vanadium incorporation into the vanadium-substituted phosphomolybdate (H_3+xPMo_12-xV_xO₄₀, 0<x<6) led to enhanced activity, as did the inclusion of sodium ions (Na_xH₃PMo_12-xV_xO₄₀, 0<x,6).⁹⁷ Addition of 0.15 M VO²⁺ (as V₂O₄) as an additional mediator to a solution of 0.3 M H₃Na₃PMo₉V₃O₄₀ gave higher sustained voltages when the fuel cell was run at 400 mA/cm². A cell voltage of ~0.5 V was observed with no added V₂O₄ and ~0.58 V with added V₂O₄.⁹⁸ In a later patent application, Acal disclosed that a mediated fuel cell with 0.3 M Na₄H₃PMo₈V₄O₄₀, operated at 80 °C and an ambient pressure, achieved sustained performance over 40 min at approximately 0.4 V and 1.5 A/cm².⁹⁹ In 2019, Knuckey, Creeth, and coworkers disclosed the role of elevated temperatures and air pressures on POM reoxidation.¹⁰⁰ Using H₇PMo₈V₄O₄₀, they tested the rate of open-circuit potential (OCP) recovery after cell discharge at 0.8 V. At a regenerator temperature of 80 °C and atmospheric pressure using air, it took over 4000 s for the OCP to increase to 0.88 V. In contrast, at 120 °C and 2 bar_g (air), the same OCP was reached in under within 300 s.

Building on the work of Acal, Matsui, Eguchi, and coworkers published a systematic study of three POMs as fuel cell cathode mediators: H₃PMo₁₂O₄₀, H₅PMo₁₀V₂O₄₀, and H₆PMo₉V₃O₄₀.¹⁰¹ Polarization curves examining the influence of POM identity, concentration (0.01–0.3 M), and cell temperature (20, 40, or 80 °C) were reported using a conventional Pt/C H₂ anode (Figure 9). The vanadium-containing POMs displayed higher activity than the all-molybdenum POM, and increased performance was observed at higher mediator concentrations and higher temperatures. The aerobic regeneration activity of the three POMs during fuel cell operation was examined. Using 0.01 M solutions at either 25 °C or 80 °C and purged with either Ar or O₂, the voltage of the cell was monitored as a function of time at a current density of 5.36 mA/cm². The voltage using the all-Mo H₃PMo₁₂O₄₀ mediator decayed identically under O₂ or Ar, indicating that the aerobic oxidation of the reduced mediator was slow. H₅PV₂Mo₁₀O₄₀ showed similar behavior under O₂ and Ar at 25 °C, indicating slow aerobic reoxidation; however, steady currents under O₂ at 80 °C indicated that the POM was oxidized at a sufficient rate for sustained activity at elevated temperature. H₆PV₃Mo₉O₄₀ is oxidized at both temperatures, yielding stable voltages over the course of the experiment. The maximum power density obtained (measured by polarization curves) was approximately 40 mW/cm², obtained with 0.3 M H₆PV₃Mo₉O₄₀ at 20 °C or with 0.1 M H₆PV₃Mo₉O₄₀ at 80 °C.

Figure 9.

Voltage time-courses of POM-mediated cathodes (H₂|Pt||POM|O₂) at a current density of 5.36 mA/cm² at either 25 °C or 80 °C under O₂ or Ar for 0.01 M solutions of (a) H₃PMo₁₂O₄₀, (b) H₅PV₂Mo₁₀O₄₀, and (c) H₆PV₃Mo₉O₄₀. Adapted from ref. 101. Copyright 2016, American Chemical Society.

Davies and coworkers published a separate study of the use of a H₆PV₃Mo₉O₄₀ POM (HV3) solution as the cathodic mediator, again utilizing a conventional Pt/C H₂ anode (Figure 10).⁶³ Many of the reaction parameters of the cathode were examined. A primary focus in this report was the influence of the concentration of reduced vanadium (VO²⁺) on the performance of the cell and examination of the “regenerator efficiency” to probe how well the reduced mediator undergoes oxidation by O₂. The authors reported the performance of the cell at various VO²⁺ fractions (i.e., the amount of reduced vanadium relative to total vanadium): 0.05 (almost fully oxidized), 0.25, 0.45, 0.65, and 0.85 (mostly reduced). As expected, the highest OCP values were obtained with the most oxidized solution (0.05 VO²⁺), and a maximum power density of nearly 1 W/cm² was achieved with 95% oxidized 0.3 M H₆PV₃Mo₉O₄₀ at 80 °C. Regeneration experiments, however, indicated that only approximately half of the V^IV can be aerobically oxidized to V^V, which will limit overall performance. Based on their regeneration data and the polarization curves, the authors predicted the steady-state performance of the combined system, and the results indicated that the steady-state performance will be similar to the results observed with the 0.85 VO²⁺ (mostly reduced) fraction. This result indicated that the regeneration ability of their system was too low to reach optimal power densities. The authors compared their calculated steady state results, as well as their data at 0.05 and 0.65 VO²⁺ fractions, to state-of-the-art conventional PEMFCs and to other mediated cathode fuel cells (Table 3). Their result at 0.05 VO²⁺ fraction was above that for state-of-art PEMFCs and far above those reported for other cathode-mediated systems, while the 0.65 VO²⁺ fraction and the calculated steady state performance were below that of traditional PEMFCs, with a maximum sustainable power density estimated to be 0.38 W/cm².

Figure 10.

Schematic of POM-mediated cathode fuel cell (H₂|Pt||POM|O₂) reported by Davies and coworkers. Adapted from ref. 63. Copyright 2017, Elsevier B.V.

Table 3.

Performance comparison of a reported conventional PEM fuel cell, previously reported mediated cathode fuel cell systems, and results from Davies and coworkers.⁶³

System	Cathode material	Cathode reaction	Liquid Electrolyte	OCP (V)	Cell voltage @ 0.2 A/cm2 (V)	Cell voltage @ 1 A/cm2 (V)	Max power (W/cm2)	Ref.
Conventional PEMFC	Pt/C	O2 + 4 e− + 4 H+ → 2 H2O	None	0.96	0.81	0.67	≥ 0.83	23
Mediated cathode (HNO3)	N-doped carbon	NO3− + 2 e− + 2 H+ → NO2− + H2O	5 M HNO3	1.04	0.95	0.46	0.51	129
Mediated cathode (POM)	carbon	VO2+ + e− → VO2+	0.45 M POMa	0.83	0.68	0.46	0.51	96
Mediated cathode (POM)	carbon	VO2+ + e− → VO2+	0.3 M HV3 @ 0.05 VIV	0.99	0.9	0.72	1	63
Mediated cathode (POM)	carbon	VO2+ + e− → VO2+	0.3 M HV3 @ 0.65 VIV	0.78	0.72	0.5	0.58	63
Mediated cathode (POM)	carbon	VO2+ + e− → VO2+	0.3 M HV3 @ steady stateb	0.81	0.73	–	0.36	63, 102
Mediated cathode (POM)	carbon	VO2+ + e− → VO2+	0.3 M NaV4 @ steady stateb	0.80	0.75	0.53	0.58	63, 102

^aPOM identity was not disclosed

^bSteady-state data was calculated from regeneration rates and cell data

In a follow up report, Davies and coworkers examined four different POM species in the same manner described above,¹⁰² with the four POM solutions labeled HV3 (empirical formula H₆PV₃Mo₉O₄₀), HV4 (empirical formula H₇PV₄Mo₈O₄₀), NaV3 (empirical formula Na₃H₃PV₃Mo₉O₄₀), and NaV4 (empirical formula Na₄H₃PV₄Mo₈O₄₀). The “protonic” HV3 and HV4 POMs displayed higher open circuit voltages at all V^IV fractions relative to the sodiumcontaining solutions, due to the lower pH of the protonic solutions. Polarization curves using almost fully oxidized HV3 or HV4 (5% V^IV) displayed power densities of over 1 W/cm². However, the HV3 composition underwent oxidative regeneration much more slowly than the other POM compositions, leading to poor extrapolated steady-state performance. The authors predicted the steady-state performance for each POM identity based on the polarization curves at different V^IV fractions and the regeneration data, and they then projected the expected steady-state polarization curves for each POM (Figure 11). Due to its poor reoxidation, HV3 had the lowest estimated performance at steady-state. The two V4 POMs had nearly identical performance, with NaV4 displaying slightly higher maximum power density (578 mW/cm²; see also, last entry in Table 3). The authors pointed out that the regenerator performance continued to be a limiting factor in achieving maximal power densities and cite an Acal patent describing the formation of “microbubbles” to enhance POM oxidation.¹⁰³ A steady-state durability test was performed using the NaV4 POM at a current density of 0.4 A/cm², and no loss in activity was detected over a 200 h run.

Figure 11.

Projected (a) i-V and (b) power density curves generated from simulated steady state performance for HV3, HV4, NaV3, and NaV4 POM solutions (H₂|Pt||POM|O₂). Adapted from ref. 102. Copyright 2017, Elsevier B.V.

In 2018, Ward and Davies published a study investigating the role of temperature and catholyte concentration of a POM-mediated cathode.¹⁰⁴ Using both HV4 and NaV4 POM compositions, the temperatures tested ranged from 40–90 °C and catholyte concentrations ranged from 0.2–0.45 M. Increases in temperature led to a slight decrease in the redox potential of the catholyte, while an increase in concentration led to larger increases in the catholyte redox potential. The authors also examined the role of temperature and concentration on the pH and conductivity of the catholyte and observed that increasing temperature and concentration both led to decreases in the pH and increases to the conductivity. Polarization curves with different concentrations of HV4 (95% oxidized) showed a 25% increase in maximum power density when shifting from 0.2 M POM to 0.3 M POM at 80 °C, but only a 6% increase when shifting from 0.3 M POM to 0.45 M POM. The minimal improvement at high concentrations was attributed to an increase in cell ohmic resistance as the catholyte concentration increased, perhaps due to VO₂⁺ and VO²⁺ incorporation into the membrane. For NaV4 catholyte solutions, the best results at 80 °C were obtained using a 0.3 M concentration of the POM, which also corresponded to the lowest observed ohmic cell resistance. Increasing the temperature of both 0.3 M HV4 and NaV4 solutions resulted in enhanced performance at low current densities due to decreased activation losses, as expected; however, detrimental effects were observed at 90 °C at high current densities. The authors suggested that the poorer performance arose from anode effects, e.g., a decrease in electrochemical Pt surface area and decreased exchange current densities. POM regeneration data, collected using the same procedure as in previous reports, indicated that NaV4 has an increased rate of reoxidation relative to HV4. Concentration had only a minor impact on the rate of regeneration. Increasing the temperature typically led to higher reoxidation rates, with an exception observed upon increasing from 80 °C to 90 °C with highly reduced (>60 % reduced) NaV4 samples. The authors attributed a decreased reoxidation rate in this case to decreased O₂ solubility and mass transport constraints. Steady-state experiments for each condition were performed at 1 A/cm², and the highest voltage (0.47 V) was obtained with 0.45 M NaV4 at 90 °C. Very similar performance, however, was obtained using 0.3 M NaV4 at 80 °C, with only a 10 mV drop in cell voltage relative to the more forcing conditions. The authors concluded that the optimal conditions for fuel cell performance are 80 °C with 0.3 M NaV4 and estimated that a steady state peak power density of approximately 510 mW/cm² could be achieved.

Stahl and coworkers reported a double MedFC using Na₄H₃PMo₈V₄O₄₀ as a cathodic mediator paired with a mediated anode consisting of a Pt/C catalyst and anthraquinone mediator to mediate H₂ oxidation.¹⁰⁵ In initial tests using the mediated anode with a conventional cathode consisting of a Pt/C electrocatalyst, low power densities were obtained due to flooding of the cathode. Significantly higher power densities (up to 528 mW/cm2 with iR-correction) were obtained when a POM-mediated cathode was used instead (see Section 3.1.6 for further discussion of the mediated anode).

Recently, analogous vanadium-containing POMs have been used as cathodic mediators for ORR in various biomass-based fuel cells. A different set of lower-potential POMs were used as mediators for anodic biomass oxidation.¹⁰⁶ The relevant anodic chemistry is discussed in Section 3.1.4. The selection of POMs used in the cathode was expanded beyond the Keggin-type POMs typically used by Acal and others described above, inspired by the fundamental POM/O₂ reactivity reported by Zhizhina and Odyakov.⁹¹

The first example was reported by Deng and coworkers in 2014,¹⁰⁷ describing a cell capable of using fresh grass, switchgrass, starch, or cellulose as the anodic fuels (Figure 12). A 0.3 M solution of H₁₂P₃Mo₁₈V₇O₈₅ (or the di- or tri-sodium salt thereof) was used as the cathodic mediator solution, and H₃PW₁₁MoO₄₀ was used as the anodic mediator. Stable power densities of 30 mW/cm² were obtained for 10 h of operation at 80 °C. Consistent with the high loading of vanadium in the cathodic POM mediator, a very high potential (> 1.0 V vs. NHE) was achieved, even with multiple reduced vanadium centers in the POM. The influence of concentration and degree of reduction on the rate of aerobic oxidation was investigated. The degree of reduction was defined as the average number of electrons each POM complex had collected, and the highest rate of oxidation was reached at a high degree of reduction (>1.8 e⁻/POM) and at intermediate concentrations (~0.15 M). Analyses of anodic reactions are presented in Section 3.1.4.

Figure 12.

A) Schematic for the doubly mediated fuel cell using POM-I as mediator for biomass oxidation and POM-II as mediator for O₂ reduction (biomass|POM-I||POM-II|O₂). B) Further details showing electrode reactions and regeneration reactions of POM-I and POM-II. Adapted from ref. 107. Copyright 2014 Wiley-VCH Verlag GmbH & Co.

Several additional studies have employed cathodes with V-containing POM mediators or mediated systems using both VO²⁺ and POMs in efforts primarily focused on processes at the anode (i.e., thorough analysis of the cathode performance was not conducted). These studies will be elaborated in Section 3, but are briefly noted here. Deng and coworkers described a fuel cell using alcohol-based fuels, such as methanol, glycerol, or glucose, using H₁₂P₃Mo₁₈V₇O₈₅ as a cathodic POM mediator. Maximum power densities of over 30 mW/cm² with sorbitol as the alcohol, based on single-scan polarization curves with a fully oxidized mediator at the cathode (i.e., without steady-state POM regeneration by O₂).¹⁰⁸ The same group later reported a fuel cell system utilizing the same POM mediator in the cathode and using low-rank coal as the fuel.¹⁰⁹ The same cathodic POM mediator was subsequently utilized by Li and coworkers in a fuel cell utilizing glucose, cellulose, or starch as a fuel.¹¹⁰ Deng and coworkers utilized sewage sludge as the fuel in their system using H₃PMo₁₂O₄₀ as the anodic mediator and H₁₂P₃Mo₁₈V₇O₈₅ as the cathodic mediator.¹¹¹ Chao, Deng, and coworkers reported a biomass fuel cell using sugars treated photochemically and thermally with H₃PMo₁₂O₄₀ at the anode and H₁₂P₃Mo₁₈V₇O₈₅ at the cathode.¹¹²

A fuel cell using lignin as a fuel and H₃PMo₁₂O₄₀ as the anodic mediator was reported by Zhao and Zhu,¹¹³ with several cathode variations. Examples included (a) a traditional Pt/C gas-diffusion cathode with O₂ as the oxidant, (b) carbon electrodes with PMo₁₂O₄₀ or K₅PV₂Mo₁₀O₄₀ as POM mediators with O₂, and (c) a carbon electrode with KMnO₄ as a stoichiometric oxidant. Open circuit potentials and maximum power densities were measured at 90 °C and are summarized in Table 4. In general, the power densities were higher with solution-phase oxidants versus a conventional O₂ cathode. Continuous operation of the cell at 90 °C with each oxidant was examined. The cathodes using O₂ at Pt/C (current density ~ 0.4 mA/cm²) and the mediated H₃PMo₁₂O₄₀|O₂ system (current density ~ 0.5 mA/cm²) displayed reasonably stable behavior, while the current density decreased substantially for cathodes utilizing KMnO₄ or K₅PV₂Mo₁₀O₄₀|O₂.

Table 4.

Performance of various cathode chemistries on fuel cell performance using lignin as fuel.

Oxidanta	OCP (mV)	Maximum Power Density (mW/cm2)
O2 @ Pt/C	350	0.96
O2 + 0.1 M H3PMo12O40	690	5
0.2 M KMnO4	900	45.1
O2 + 0.1 M K5PV2Mo10O40	330	0.41

^a: 1 atm O₂

2.3. Fe²⁺/Fe³⁺ mediator

The Fe²⁺/Fe³⁺ redox couples, including those from aqueous Fe salts, [Fe(CN)₆]^4/3−, and ferrocene derivatives, have been the subject of substantial interest from the flow battery community.⁸ The electrode kinetics can be rapid, and these mediators have low cost and toxicity. On the other hand, the redox potential of free Fe²⁺/Fe³⁺ (~0.7 V) is relatively low compared to the thermodynamic potential for O₂, even while ancillary ligands may be used to access higher potentials. The various mediated cathode systems using this mediator for O₂ reduction are summarized in Table 5.

Table 5.

Reported systems using iron-based mediators for O₂ reduction.

Anode			Cathode
fuel	catalyst	mediator	mediator	catalyst	ref
Zn	–	–	Fe2+/Fe3+	charcoal	114
H2	Pt/C	–		Fe(pc)	61
			K3Fe(CN)6	Mn complex	118
			Fc species	Fe complex	119–123

In 1931, Lamb and Elder studied the aerobic oxidation of FeSO₄ in acidic solutions catalyzed by various substances, with the goal of utilizing this redox mediator in a fuel cell application.¹¹⁴ The rate of oxidation of Fe²⁺ was monitored by measuring the potential of the solution during the oxidation. The reaction rate was found to be first order in pO_2, second order in [FeSO₄], and independent of [H₂SO₄] above 0.23 M, but the reaction rate increased rapidly below this H₂SO₄ concentration. CuSO₄ was found to accelerate the oxidation, and it was noted that in the presence of CuSO₄ the oxidation of FeSO₄ accelerated with increasing acid concentration. Additionally, charcoal accelerated the reaction rate with a roughly first order dependence, and Pt black accelerated the reaction rate in a non-linear fashion, which the authors attributed to catalyst poisoning. The authors then constructed a working cathode solution with a Zn/ZnSO₄ anode to study the behavior of the mediated cathode under electrochemical conditions. Comparison of O₂ consumption in their kinetic studies with data obtained from the electrochemical studies show good agreement. In the electrochemical studies, the authors measured a current density of 21.5 mA/cm² while operating at a potential of 0.606 V.

In 1953, Posner published a pair of papers examining the oxidation of Fe²⁺ in concentrated HCl solutions with no catalyst and in dilute HCl solutions catalyzed by charcoal.^115,116 At high concentrations of HCl (> 4.0 N), Fe²⁺ was rapidly autoxidized, while below this value, the reaction was slow, consistent with the previous observations of Lamb and Elder. At the high concentrations of acid, the rate of Fe²⁺ oxidation was first order in [Fe²⁺], approximately first order in [HCl] between 6–8 N, and first order in pO₂. Posner proposed that the rate of oxidation increases at high [HCl] due to the formation of an Fe²⁺-HCl complex, the oxidation of which bypasses formation of superoxide and proceeds directly to HO₂^•. When H₂SO₄ is used as the acid, no acceleration of the autoxidation rate is observed at high concentrations, suggesting that the active complex is not formed with this acid. At lower acid loadings, with charcoal as a catalyst, the autoxidation rate was first order in the catalyst loading, pO₂, and [H⁺], and proportional to [Fe²⁺]/([Fe²⁺]+[Fe³⁺]). The use of iron as a mediator for the ORR was not pursued further, however, due to its low potential.⁷⁰ Efforts were made to identify ligands that could increase the Fe^2+/Fe³⁺ potential.¹¹⁷ No suitable ligands were identified, however, and use of Fe^2+/Fe³⁺ as ORR cathode mediators were abandoned for many years.

Iron-based mediators reemerged in the mid- to late-2000s, when Acal Energy Ltd. published a series of patents using various iron-based mediators in fuel cell cathodes. A patent application¹¹⁸ described a MedFC using a conventional Pt/C H₂ anode paired with a mediated cathode system containing K₃Fe(CN)₆ and KOH. The use of both a cationic- and anionic-selective membrane was required for optimal performance. In a separate experiment, they disclosed the aerobic regeneration of K₃Fe(CN)₆ from K₄Fe(CN)₆ using a solution containing 0.02 M EDTA and 0.01 M MnSO₄. However, there was no integration between the aerobic regeneration studies and the fuel cell experiments. In subsequent patents, ferrocene (Fc) species were described as mediators using iron complexes to catalyze aerobic oxidation of the reduced mediator.^119–122 Examples include the use of Fc(CH₂SO₃H)₂ as a cathodic mediator and [Fe(trilen-(SO₃Na)₃)]²⁺ as the catalyst at pH 2 (Figure 13). A fuel cell was constructed incorporating a conventional Pt/C H₂ anode (i.e., H₂|Pt/C||Fc(CH₂SO₃H)₂|[Fe(trilen-SO₃Na)₃)]²⁺|O₂), and the mediator was partially oxidized by bubbling O₂ through a solution of the mediator and catalyst (10 mM mediator and 1.0 mM catalyst). A polarization curve was measured with this cell. This system exhibited an open-circuit potential of approximately 0.78 V, but the potential decreased rapidly with increasing current, resulting in a current density of only 16 mA/cm² at a cell voltage of 0.2 V. The authors confirmed that the mediator undergoes regeneration during fuel cell operation by monitoring the open-circuit potential vs. time. Fc derivatives with the cyclopentadienyl rings linked together by an alkyl tether exhibit increased stability relative to unlinked derivatives at pH 2.5.¹²³ In spite of these improvements, subsequent efforts by Acal emphasized polyoxometalate mediators (cf. Section 2.2).

Figure 13.

Ligand used with Fe²⁺ (trilen-SO₃Na)₃ and to oxidize Fc(CH₂SO₃H)₂ mediator.

In 2016, Park and coworkers reported the use of iron phthalocyanine, Fe(pc), as a catalyst for the aerobic oxidation of Fe²⁺ in the context of a MedFC.⁶¹ They studied the kinetics of Fe²⁺ [as Fe(SO₄)] oxidation by O₂ (1 atm) in 1 M H₂SO₄ catalyzed by 0.25 mM Fe(pc) or [Fe^III(tetra(4-OMephenyl)porphyrin)]Cl. Fe(pc) was a more effective catalyst with >90% conversion of Fe²⁺ to Fe³⁺ after 5 h. This reaction was proposed to proceed via O₂ binding to the iron macrocycle in the Fe^II oxidation state and subsequent reduction of O₂ to water with H₂SO₄ and aqueous Fe²⁺ ions as the source of protons and electrons (Figure 14A). The authors then measured polarization curves for the three oxidized solutions (0.5 M FeSO₄ plus 1.0 M H₂SO₄ and (i) no catalyst, (ii) Fe(pc), or (iii) [Fe^III(tetra(4-MeOphenyl)porphyrin)]Cl). They found the best performance for the solution containing Fe(pc), stemming from the much higher conversion of Fe²⁺ to Fe³⁺ (Figure 14B). In this study, the cell was held at 80 °C and H₂ was supplied as the fuel to a Pt/C anode (H₂|Pt/C||Fe²⁺|Fe(pc)|O₂). The solution containing Fe(pc) as the catalyst led to a peak power density of ~250 mW/cm², which was over three-fold higher than the other two solutions (65 mW/cm² for the uncatalyzed solution and 76.8 mW/cm² for the solution containing the Fe-porphyrin catalyst). However, no sustained electrolysis data was included to provide evidence for efficient on-stream regeneration of the mediator during fuel cell operation. The authors noted that further adoption of mediated cathode systems would require higher potential mediators, as well as more active catalysts for the mediator oxidation.

Figure 14.

(A) Proposed mechanism for Fe²⁺ oxidation by O₂ catalyzed by Fe(pc). (B) Polarization curves measured using 0.5 M FeSO₄ + 1.0 M H₂SO₄ solutions after oxidation at 80 °C with O₂ for 5 h in the presence of no catalyst (black traces), [Fe^III((4-OMephenyl)porphyrin)]Cl (Fe-TMPP) as catalyst (blue traces), or Fe(Pc) as catalyst (red traces). H₂ was used as the fuel at a Pt/C electrode (H₂|Pt/C||Fe²⁺|Fe(pc)|O₂). Adapted from ref. 61. Copyright 2016 American Chemical Society.

In order to study Fe complexes similar to those reported by Acal, Metz and coworkers used computational methods to calculate the reduction potential of a variety of different Fe complexes bearing chelating nitrogen ligands (Figure 15).⁶² Using experimental E_1/2 values for [Fe(II)trilen (R = H/Cl/SO₃Na)-OH₂]²⁺ and Fe(II)N4Py(R = H/CH₃)-OH₂]²⁺, the authors calculated the difference between potential energy surfaces by the self-consistent field method, ΔE(SCF), and found good agreement between these values. The authors used these benchmarks to estimate the redox potential of [Fe(II)trilen (R = CH₃/CF₃/CN)-OH₂]²⁺ and [Fe(II)TFM-trilen(R = H/SO₃Na)-OH₂]²⁺. The complex [Fe(II)TFM-trilen(R = SO₃Na)-OH₂]²⁺ was expected to have a redox potential of nearly 1 V, which would increase the open-circuit potential by ~25% relative to the highest previous Fe complex studied experimentally. A synthetic route to this complex was proposed, but it wasn’t prepared. Thus, experimental validation of these results, including assessment of mediator stability and aerobic regeneration, awaits further study.

Figure 15.

Ligands studied computationally by Metz and coworkers to determine the redox potentials of the Fe complexes.⁶² Experimentally determined redox potentials are shown in red, while the calculated redox potentials are shown in black.

2.4. NO_x mediators

In previous sections, NO_x species have been utilized as catalysts for aerobic oxidation of a mediator. HNO₃ has also been utilized as a single-component mediator/catalyst system. Both HNO₃ and HNO₂ have high reduction potentials to NO, 0.957 V and 1.028 V vs. NHE, respectively. These values are close to the formal reduction potential of O₂ to H₂O, 1.23 V, and it is possible to access high concentrations of HNO₃ that support high current densities. The systems using NO_x species derived from HNO₃ as cathode mediator/catalysts for O₂ reduction are given in Table 6.

Table 6.

Reported systems using NO_x species as mediators for O₂ reduction.

Anode			Cathode
Fuel	catalyst	mediator	mediator	catalyst	ref
MeOH or HO(CH2)2OH	Pt	–	HNO3		124–127
H2	Pt/C	–			128–129

In the early 1960s, research into the use of HNO₃-mediated ORR was conducted by researchers at Esso Research and Engineering Company in New Jersey.¹²⁴ Alcohols, such as methanol or ethylene glycol, were targeted as the fuels for this system. In the initial report, low concentrations of HNO₃ (approximately 1 wt% in H₂SO₄) gave substantially higher currents than O₂ alone. The researchers studied the mechanism of electroreduction of HNO₃ and proposed an autocatalytic mechanism with a rate-limiting step involving the reaction of HNO₃ with nitric oxide (NO) to form nitrous acid (HNO₂).¹²⁵ The latter was proposed as the electroactive species. Engineering of the external regenerator allowed up to 30 cycles of HNO₃ turnover, measured by comparing the total charge passed to stoichiometric redox equivalents available from HNO₃.¹²⁶ A challenge with this system was the loss of NO or other NO_x species to the gas phase, but further development of the MeOH-HNO₃/O₂ fuel cell enabled power densities of 44 mW/cm² at 0.34 V using a new Pt/C-based catalyst at the cathode. A maximum power density of 19 mW/cm² was obtained during continuous operation for over 180 h using a standard Pt mesh cathode.¹²⁷

The use of NO_x-based mediators for fuel cell cathodes was revisited in 2011, when Park and coworkers reported a fuel cell using the NO₂⁻/NO₃⁻ as a cathodic redox mediator using a nitrogen-doped carbon-felt electrode.¹²⁸ The reduction of NO₃⁻ to NO₂⁻ has a formal potential of 0.94 V, and aerobic oxidation of NO₂⁻ is facile. The authors prepared an Fe-N/C catalyst by treatment of iron nitrate and melamine with carbon black and annealing of the resultant powder at 700 °C under N₂. This electrode showed enhanced activity for nitrate reduction relative to commercial Pt/C catalysts. Specifically, polarization curves using a conventional Pt/C H₂ anode and 5 M HNO₃ as the oxidant in the cathode compartment showed higher current densities with Fe-N/C than with Pt/C cathodes (325 mA/cm² vs. 250 mA/cm², respectively) at 0.9 V. To monitor the stability of the cathode materials, the power density at 0.8 V was measured over the course of 100 h with an operating temperature of 80 °C. The Fe-N/C electrode showed minimal drop in activity, while the performance of the Pt/C cathode dropped dramatically. This decrease in activity was attributed to Pt dissolution. The selectivity of NO₃⁻ reduction using the Fe-N/C electrode was measured, and a 98.3% selectivity for formation of NO₂⁻ was observed, with the remainder forming N₂ and NO. No data describing aerobic regeneration of NO₃⁻ was presented, though the authors suggested that this should be achievable.

In 2017, Park and coworkers reported an H₂/O₂ fuel cell using a Pt/C anode and a heteroatom-doped cathode with an HNO₃/NO mediator (Figure 16).¹²⁹ Nitrogen-doped carbon felt, prepared by treatment of commercial carbon felt at 700 °C under an atmosphere of NH₃, proved to be an efficient electrode material to facilitate the reduction of NO₃⁻ to NO. Polarization curves using 5 M HNO₃ were collected at a variety of temperatures, and at 80 °C a power density of over 0.5 W/cm² was achieved at a potential of 0.8 V. To examine the long-term activity of this system, the fuel cell performance was monitored by holding the cell at a potential of 0.8 V for 30 days with circulation of the 5 M HNO₃ solution supplied with 1 atm O₂ (H₂|Pt/C||HNO₃|O₂). A slow linear decline in performance was observed during the experiment, falling to approximately 0.3 W/cm² after 30 days of operation (Figure 17); however, replacement of the cathode solution with fresh HNO₃ caused the cell performance to return to its original values. The authors calculated a “regeneration efficiency” of approximately 93%. With no O₂ supply, the system was capable of producing power for only 4 days, reflecting the redox capacity of the HNO₃ in the cathode solution.

Figure 16.

Schematic of a mediated fuel cell using the NO₃⁻/NO redox mediator to mediate O₂ reduction. Adapted from ref. 129. Copyright 2017 Wiley-VCH Verlag GmbH & Co. KGaA.

Figure 17.

Long term stability test of HNO₃-mediated cathode with nitrogen-doped carbon-felt electrode. Cell run at 80 °C with H₂ anode. Adapted from ref. 129. Copyright 2017 Wiley-VCH Verlag GmbH & Co. KGaA.

A direct ethanol fuel cell utilizing a PtSn alloy anodic catalyst and a HNO₃-based catholyte mediator was reported in 2018 by Park and coworkers (Figure 18).¹³⁰ The system achieved a maximum power density of 68 mW/cm² and an open-circuit potential of ~0.85 V. In contrast, a system using direct O₂ reduction on a commercial Pt/C catalyst on the cathode achieved a maximum power density of 34 mW/cm² and an open-circuit potential of ~0.65 V. No NO₂ regeneration data was presented in this work.

Figure 18.

Schematic of a direct ethanol fuel cell using an HNO₃ mediated cathode. Adapted from ref. 130. Copyright 2018 Hydrogen Energy Publications LLC.

2.5. Nitroxyl/Oxoammonium mediators

Organic aminoxyls, such as TEMPO (2,2,6,6-tetramethylpiperidine N-oxyl), are organic radicals that can undergo 1 e⁻ oxidation to an oxoammonium species. They have been used as mediators in a wide variety of alcohol oxidation reactions, including under electrochemical^46,131 and aerobic^132,133 reaction conditions. The use of NO_x co-catalysts in the latter reactions provided the starting point for a fundamental study by Gerken and Stahl to investigate the ORR with an aminoxyl/NO_x mediator/catalyst system (Figure 19A).¹³⁴ Cyclic voltammetry and controlled-potential electrolysis experiments highlighted the cooperative redox behavior of this system. A series of different nitroxyl mediators (TEMPO, ACT, 3-CARP, and ABNO; Figure 19B) were examined. TEMPO/TEMPO⁺ electrochemistry was facile, but TEMPO did not react efficiently with O₂ (Figure 19C, blue trace). NO, generated by addition of acid to a solution of NaNO₂, reacted rapidly with O₂, but the NO₂ product exhibited poor electrochemical behavior (Figure 19C, black trace). The combination of both mediators led to a significant and sustained increase in the observed currents (Figure 19C, red trace). Other aminoxyl mediators (ABNO, ACT, and 3-CARP) yielded similar results (Figure 19D), and the operating potential correlates with the potential (vs. Fc^+/0) of the aminoxyl/oxoammonium redox couple.

Figure 19.

A) Simplified scheme for TEMPO-mediated O₂ reduction using a NO_x-based catalyst B) Different nitroxyls examined in the study. C) Controlled-potential electrolysis traces in the presence and absence of mediator and/or catalyst (TEMPO and NO_x, respectively). D) Controlled-potential electrolysis traces in the presence of NO_x and various nitroxyl mediators (red = TEMPO; black = ACT; blue = ABNO; green = 3-CARP). Adapted from ref. 134. Copyright 2015 American Chemical Society.

The thermodynamic potential for O₂ reduction under these reaction conditions was determined by measuring the H⁺/H₂ open-circuit potential under the non-aqueous conditions and applying the relevant corrections.^135,136 The highest potential mediator, ACT, was shown to feature an ORR overpotential of roughly 300 mV, similar to overpotentials associated with the onset of ORR with conventional Pt catalysts. This system was not well suited for practical application (e.g., due to the use of an organic solvent and loss of NO_x species to the gas phase), but it provides some of the most thorough insights to date into the mechanisms of electrochemical O₂ reduction with NO_x and mixed aminoxyl/NO_x mediator systems.

Park and coworkers reported a MedFC using aqueous acidic conditions with TEMPO as a mediator in the absence of a redox catalyst (Figure 20A).¹³⁷ The authors report the redox potential of TEMPO to be 0.76 V and used a conventional Pt/C H₂ anode (H₂|Pt/C||TEMPO|O₂). A fuel cell test with a cathode flow rate of 2 mL/min produced a maximum power density of 90 mW/cm² (Figure 20B). A constant potential experiment was conducted at 0.4 V, and the current was observed to go to zero in approximately 190 minutes (Figure 20C). A rough analysis of the total charge passed suggests that no regeneration of TEMPO occurs with O₂, consistent with previous observations (e.g., Figure 20B). The reduced TEMPO mediator will be present as TEMPOH₂⁺ under these conditions and will not undergo facile reactions with O₂.¹³⁸

Figure 20.

A) Schematic of mediated cathode fuel cell using TEMPO mediator. B) Polarization curve of the mediated fuel cell using TEMPO as a cathodic mediator and a conventional H₂ anode. Conditions: 0.3 M TEMPO and 1 M H₂SO₄ in the cathode, fuel cell temperature 30 °C. C) Time-on-stream test using a TEMPO-mediated cathode. Inset shows original color of TEMPO solution (left) and solution color after time-on-stream experiment (right). Adapted from ref. 137. Copyright 2018, Elsevier B.V.

2.6. Quinone/Hydroquinone mediators

Quinones have a rich redox chemistry. They undergo both electron-transfer or proton-coupled electron transfer, they are used in dehydrogenation and other oxidation reactions with organic molecules, and they serve as cocatalysts in aerobic oxidation reactions.^139,140 The last of these roles has direct relevance to the ORR because hydroquinone (H₂Q) reacts with activated oxygen species in the reduction of O₂ with a molecular Co(salophen) catalyst.¹⁴¹ Stahl and Anson showed that the use of H₂Q as an “electron-proton transfer mediator” led to a change in ORR product selectivity from H₂O₂ (two-electron reduction) to H₂O (four electron reduction) during electrocatalytic O₂ reduction.¹⁴² Cyclic voltammetry and bulk electrolysis experiments further showed that H₂Q led to an enhanced rates and lower overpotentials for the Co(salophen)-catalyzed ORR in organic solvent (N,N-dimethylformamide, with acetic acid as the proton source; Figure 21). The conditions for the Co(salophen)/H₂Q study reflected those used in the prior aerobic oxidation reactions, but the results have implications for more-conventional ORR effort by showing that a mediator could participate in a cooperative, inner-sphere mechanism and thereby influence the rate and/or selectivity of the reaction.

Figure 21.

Simplified scheme for quinone-mediated O₂ reduction utilizing Co(salophen) as a soluble catalyst. Adapted from ref. 142. Copyright 2017, American Chemical Society.

The first demonstration of hydroquinone-mediated ORR under PEMFC-like conditions was reported by Stahl and coworkers in 2018, using a water-soluble quinone mediator paired with an off-electrode heterogeneous Co-N/C catalyst.¹⁴³ Water-soluble high-potential quinones containing sulfonate groups have been utilized in flow-battery applications¹⁴⁴ (Compounds A-C, Figure 22), but the previously reported materials rapidly decomposed and had inadequate stability for fuel cell operation. To address this limitation, a new tetra-substituted quinone bearing four alkylsulfide-linked sulfonate groups was designed and synthesized (Compound D, Figure 22). This mediator exhibited substantially improved stability with a half-life over >5000 h. Various M-N/C catalysts were tested for their ability to oxidize the corresponding hydroquinone to the quinone under fuel cell relevant conditions (1 M H₂SO₄ and 60 °C), and a Co-N/C catalyst using 1,10-phenanthroline as the nitrogen source during the pyrolytic catalyst preparation was identified as the most effective catalyst. A fully integrated fuel cell was constructed using a conventional Pt/C H₂ anode (H₂|Pt/C||quinone|Co-N/C|O₂) and a Nafion membrane between the anode and cathode (Figure 23A). The cathodic solution contained 0.1 M quinone in 1 M H₂SO₄. This solution was circulated between the fuel cell cathode, where the oxidized mediator was reduced, and a packed-bed reactor, where the mediator was oxidized by O₂ using the Co-N/C catalyst. An iR-corrected peak power density of 251 mW/cm² (161 mW/cm² uncorrected) was obtained (Figure 23B). Under steady-state conditions, a stable current output was observed for over 8 h at a constant potential of 0.5 V (Figure 23C). Control experiments showed that a decrease in current density occurred when O₂ flow was shut off to the packed bed reactor, and the current density returned to its previous value when O₂ flow was resumed. It was noted that development of stable higher potential mediators would enable the fuel cell performance to match or exceed that of conventional fuel cells.

Figure 22.

Stability of quinone structures previously tested for flow battery applications, A-C, and new tetrasubstituted quinone D. The stability was assessed by 1H NMR, and the half-lives were calculated by assessing the first-order rate constant for quinone decomposition. Conditions: 0.1 M quinone concentration in 0.5 M D₂SO₄ + 0.05 M CH₃SO₃H in D₂O at 60 °C. Adapted from ref. 143. Copyright 2018, Elsevier B.V.

Figure 23.

A) Schematic of mediated cathode utilizing quinone mediator and Co-Phen/AC catalyst, H₂|Pt/C||quinone|Co-phen/AC|O₂. B) Polarization curves and C) time-on-stream performance for quinone-mediated fuel cell cathode with a Co-N/C catalyst. Conditions: 0.1 M quinone mediator, 1 M H₂SO₄, 60 °C, H₂ fuel with Pt/C anode, MWCNT-coated carbon paper cathode, Nafion 117 membrane, 5 g Co-N/C catalyst in packed bed reactor. Time-on-stream experiment conducted at a constant potential of 0.5 V. Adapted from ref. 143. Copyright 2018, Elsevier B.V.

2.7. 2,2’-Azino-bis(3-ethylbenzothiazoline-6-sulphonate) mediator

2,2’-Azino-bis(3-ethylbenzothiazoline-6-sulfonate), ABTS²⁻, is a water-soluble, redox-active organic molecule that has been used as a mediator for O₂ reduction with various enzymes at the cathode of enzymatic fuel cells (Figure 24). Mediators have found widespread use in enzymatic fuel cells to facilitate electron transfer between the electrode and the active site of an enzyme, which can be buried within the enzyme. A more detailed discussion of the role of mediators in enzymatic fuel cells is provided in Section 3.2 in the context of anodic reactions, and a review of O₂ reduction in enzymatic biofuel cells, including both mediated and direct electrochemical approaches, was recently published.¹⁴⁵ In general, mediated fuel cells employing enzymatic catalysts achieve power densities that are much lower (often, by two or three orders of magnitude) than those described in the previous sections.

Figure 24.

(A) Structure of ABTS²⁻ and (B) redox sequence for ABTS²⁻-mediated cathode with enzymes as catalyst. For systems using this mediated cathode, see references 146–154.

The first use of ABTS²⁻ in this context was reported by Palmore and Kim in 1999, using fungal laccase from Pyricularia oryzae as the enzyme catalyst.¹⁴⁶ Many laccases reduce O₂ at very high potentials, though prior reports using direct electrochemical methods displayed very low current densities. The outer-sphere electron-transfer reactivity of laccase accommodates a wide array of mediators. The authors considered mediators with an array of redox potentials, with the hope of identifying one close to the formal potential of the enzyme (E°’ = 0.539 V vs. SCE at pH 4.0, at which the formal potential for the reduction of O₂ to H₂O is 0.723 V vs. SCE). ABTS²⁻ was selected due to its high potential (E°’ = 0.440 V vs. SCE at pH 4.0), good solubility and reasonable stability.

The authors investigated the aerobic oxidation of ABTS²⁻ to ABTS^•− catalyzed by laccase. Lower pH values favored higher laccase activity, and the stoichiometry of ABTS²⁻ oxidized to O₂ reduced was found to be 4:1, consistent with full reduction of O₂ to water. A fuel cell employing a pH 4, 0.2 M acetate buffer was constructed with a Pt-gauze anode under 1 atm H₂ and a glassy carbon or Pt-foil cathode containing both laccase and ABTS²⁻ under 1 atm O₂ (H₂|Pt||ABTS²⁻ |laccase|O₂). The anode and cathode compartments were separated by a Nafion membrane. Polarization curves were obtained, and this system showed both a higher open-circuit potential and lower polarization relative to a Pt/O₂ or carbon/O₂ cathode system (Figure 25A). For example, at a current density of 25 μA/cm² the biocathode exhibits a potential of 0.50 V vs. SCE, while the Pt/O₂ and carbon/O₂ system have potentials of 0.04 V and −0.29 V, respectively. Measured power densities were substantially higher for the biocathode, with the mediator system achieving a power density of approximately 40 μW/cm², compared to ~15 μW/cm² for a Pt cathode and ~3 μW/cm² for a carbon cathode (Figure 25B). The performance of mediated laccase cathodes decreases at pH > 5. A biofuel cell pairing ABTS²⁻ and fungal laccase at the cathode was reported to achieve a max power density of 99.8 μW/cm² at a pH 4.0, with a decrease to 2.0 μW/cm² at a pH 7.¹⁴⁷

Figure 25.

A) Polarization curves of a fuel cell constructed with a laccase/ABTS²⁻ biocathode (dotted line with open circles), Pt cathode (solid line with filled triangles), or carbon cathode (solid line with filled circles). B) Power density for the mediated fuel cell (dotted line with open circles) compared with a Pt cathode (solid line with filled triangles) and a carbon cathode (solid line with filled circles). Adapted from ref. 146. Copyright 1999 Elsevier Science S.A.

A pair of studies investigating the use of ABTS²⁻ as a mediator at pH 7 were published by Ikeda and coworkers in 2001. The initial study explored electrocatalytic reduction of O₂ with bilirubin oxidase from M. verrucaria as the catalyst and ABTS²⁻ as the mediator.¹⁴⁸ Cyclic voltammetry studies indicated catalytic O₂ reduction occurred under these conditions, with a half-wave potential of 490 mV vs. Ag/AgCl, which is close to the thermodynamic potential for O₂ reduction to H₂O (615 mV vs. Ag/AgCl at pH 7.0; Figure 26). The mediator was found to decompose via disproportionation of ABTS^•− to ABTS²⁻ and ABTS⁰, which is unstable under these conditions. In a subsequent study, a biofuel cell with an ABTS²⁻/bilirubin oxidase cathode system was paired with Desulfovibrio vulgaris bacteria and methyl viologen for H₂ oxidation at the anode.¹⁴⁹ A pH of 7 was maintained in each compartment, which were separated by an anion-exchange membrane. Results from this study showed that the cathode compartment exhibited good activity and stability, but the overall power output was limited by the anodic chemistry.

Figure 26.

Cyclic voltammagrams of: A) pH 7.0 phosphate buffer, B) pH 7 phosphate buffer and 0.25 mM ABTS2–, and C) pH 7 phosphate buffer, 0.25 mM ABTS2–, and 0.11 mM BOD. Scan rate was 10 mV/s, working electrode was glassy carbon, and a Pt disk was used as the counter electrode Adapted from ref. 148. Copyright 2001 Elsevier Science B.V.

In 2007, Dunn and coworkers reported a fuel cell using bilirubin oxidase immobilized in a sol-gel with ABTS²⁻ as a mediator for O₂ reduction.¹⁵⁰ The sol-gel substantially decreased diffusion of ABTS²⁻ to the electrode. Using an anode consisting of a sol-gel immobilized glucose oxidase enzyme as the catalyst, ferrocene methanol as a soluble mediator, and glucose as the fuel (glucose|glucose oxidase|FcCH₂OH||ABTS²⁻|bilirubin oxidase|O₂), an open circuit potential of 480 mV and a peak power density of approximately 0.12 mW/cm² were obtained. When air was used instead of O₂, the power density dropped to 0.086 mW/cm². Similar work was published by Xu and coworkers, using ABTS²⁻ as a mediator with laccase at the cathode and glucose oxidase with either a ferrocene derivative or phenazine methylsulfate as anodic mediators.¹⁵¹ A maximum power density of 5.8 μW/cm² was obtained in this case.

An enzymatic biofuel cell with a maximum power density of ~2 mW/cm² was reported by Zhu and coworkers.¹⁵² At the cathode, laccase was bound to a graphene-gold nanoparticle (AuNP-graphene) hybrid electrode, and ABTS²⁻ was used as a mediator. Glucose oxidase was anchored to the same electrode material at the anode, and direct electron-transfer was possible (Figure 27A–B). Using 0.5 mM ABTS²⁻ as a cathode mediator and 50 mM glucose as fuel for the anode, an open-circuit potential of 1.16 V and a maximum power density of 1.96 mW/cm² were achieved (Figure 27B). This system also displayed good stability; after 70 days of operation the OCP retained 80% of its original value (Figure 27C) and the maximum power of the cell decreased to 1.30 mW/cm², corresponding to 66% of the original value.

Figure 27.

A) General schematic of the mediated fuel cell system using ABTS²⁻ as a mediator for O₂ reduction in the cathode. B) Formal redox potentials (vs. SHE) for each redox couple in the mediated fuel cell (GOD = glucose oxidase and HABTS = protonated ABTS²⁻). C) Power curve of the assembled fuel cell. D) Open-circuit potential measurements of the cell over 70 days of operation. Adapted from ref. 152. Copyright 2015, The Royal Society of Chemistry.

Chen, Hou, and coworkers reported a single-enzyme biofuel cell that utilizes immobilized laccase enzymes at both the anode and the cathode.¹⁵³ Bisphenol A (BPA), was oxidized at the anode and O₂ was reduced at the cathode. While the unmediated system was capable of producing power, enhanced performance was obtained when ABTS²⁻ was introduced as a mediator for the cathode, with an approximately 3-fold increase in voltage after addition of ABTS (Figure 28).

Figure 28.

Change in cell potential after addition of bisphenol A (BPA) and ABTS²⁻ using immobilized laccase enzyme on both anode and cathode. Adapted from ref. 153. Copyright 2017, Wiley-VCH Verlag GmbH & Co.

Cosnier and coworkers entrapped P₂ABTS²⁻ (bis-pyrene-ABTS²⁻) in a glyconanoparticle and paired this mediator with solubilized bilirubin oxidase to mediate O₂ reduction in a “solubilized enzymatic fuel cell.”¹⁵⁴ This entrapment strategy facilitates the use of higher concentrations of hydrophobic mediators, and the increased mediator size allows for the use of size-exclusion membranes to limit mediator and enzyme crossover. The cathode containing the entrapped P₂ABTS²⁻ (PS-bCD-P₂ABTS) glyconanoparticle and bilibrubin oxidase was paired with an anode containing a 9,10-phenanthrenequinone (PS-bCD-PQ) glyconanoparticle and fungal flavin adenine dinucleotide-dependent glucose dehydrogenase (Figure 29). The glyconanoparticles were prepared by self-assembly of amphiphilic b-cyclodextrin modified polystyrene polymer, which was synthesized through click chemistry of functionalized cyclodextrin and polystyrene blocks,¹⁵⁵ and the resulting PS-bCD-P₂ABTS glyconanoparticles were approximately 43 nm in size. A 3-chamber fuel cell was designed, with the central chamber connected to a peristaltic pump and separated from the anode and cathode compartments by dialysis membranes with a cutoff value of 25 kDa. Fresh solution containing pH 7.0 McIlvaine buffer, ambient O₂ concentration, and 5 mM glucose was flowed into the central compartment at a rate of 80 μL/min. The assembled fuel cell achieved a maximum power of 116 ± 21 μW. Charge-discharge experiments were conducted over 10 days. After seven days of testing, the peak power dropped to 85 ± 3 μW, for a power loss of approximately 26%. The spent fuel solution was tested and no evidence was found for the presence of enzymes or mediator. In contrast, a system without encapsulation of the mediators failed to charge after only 20 h, due to leaching of the mediators out of the cathode and anode compartments.

Figure 29.

Anode and cathode designs for “solubilized enzymatic fuel cell” using glyconanoparticle-entrapped 9,10-phenanthrenequinone and P₂ABTS²⁻ mediators. The cathodic enzyme used was bilirubin oxidase (Box) and the anodic enzyme used was fungal flavin adenine dinucleotide-dependent glucose dehydrogenase (FAD-GDH). Adapted from ref. 154. Copyright 2018, American Chemical Society.

3. The Anode: Hydrogen, Alcohol, and Biomass Oxidation

A variety of fuels have been examined in PEMFCs. Hydrogen is the most widely used and is the current fuel for commercial fuel cell systems; however, challenges associated with the transportation and storage of H₂ continue to motivate efforts to identify alternative fuels. Methanol (MeOH) is the most common alternate fuel for low-temperature (<120 °C) fuel cells, in part, because methanol can undergo electrochemical oxidation to CO₂ without the need to break carbon-carbon bonds. Mediated approaches to electrochemical oxidation of H₂ and MeOH have been demonstrated in fuel cell applications, but the opportunities for MedFCs may be more significant for applications with challenging liquids, solids or other complex fuels that do not undergo facile direct electrochemical oxidation. In such cases, the redox chemistry with the fuel can occur off-electrode with a dissolved mediator. This approach allows for consideration of a wider variety of catalysts, since they do not need to be conductive or have other properties required of direct electrocatalysts. Results of mediated electrochemistry at the anode of fuel cells, including examples with chemical and enzymatic catalysts, are summarized below.

3.1. Fuel oxidations using chemical catalysts

3.1.1. Introduction

Early studies of MedFCs included studies of mediated anodes with H₂ as the fuel, but major advances in the performance of gas-diffusion electrodes and the ability to use low Pt catalyst loadings diminished interest and effort in such applications. Nonetheless, the historical efforts provide important context for mediator selection for oxidation reactions with more complex fuels, such as alcohols and biomass. An overview of the various mediated systems using chemical catalysts for anodic reactions is given in Table 7, each of which will be discussed in the following sections.

Table 7.

Mediated anode systems utilizing chemical catalysts. In all cases, the oxidant used is O₂.

Anode			Cathode
fuel	catalyst	mediator	mediator	catalyst	ref
H2	Pd	Ti3+/TiO2+	Br−/Br2	NOx	40
	Pt/Al2O3		VO2+/VO2+		80
	Pd/Al2O3	Fe(EDTA)2−/Fe(EDTA)−			80
	Pt/Al2O3	Mo3+/Mo4+			79
		H4SiW12O40			79
				POM	63
	Pt/C	AQ	POM		105
alcohols, sugars or lignin	–		–	–	166–168
sugars	POM	Fe2+/Fe3+	POM		110
methane, alcohols	Pt		VO2+/VO2+	NOx	82–83
biomass	–				84–86
sugars	Pt	V3+/VO2+		Pt	81, 169
sugars	–	viologen	–	Co	171–175, 177
biomass	POM		–	Pt/C	113, 159
alcohols			POM		107
coal					109
biomass					111, 112

Two different MedFC configurations have been employed for the oxidation of biomass or related “challenging” fuels with mediators (Figure 30). The first features a one-stage process in which the fuel oxidation and electrochemical reduction of the mediator are coupled within a single continuous process (Figure 30A). The second features a two-stage process in which the fuel is oxidized by the mediator in an independent process, and electrochemical regeneration of the reduced mediator occurs in a separate process (Figure 30B). This latter strategy allows the chemical and electrochemical redox process to be conducted under different conditions (e.g., temperatures, time), while retaining the ability to translate chemical redox processes into electrical power.

Figure 30.

Depiction the anode sequence for A) one-stage versus B) two-stage mediated biomass oxidation processes.

3.1.2. TiO²⁺/Ti³⁺ mediator

The first use of mediators to facilitate H₂ oxidation was reported in 1959 by Carson and Feldman working at General Electric (see Section 2.1 for discussion on their cathode system).⁴⁰ They used an on-line reformer without gas purification, meaning that this mediator system must tolerate CO₂ and H₂S impurities. Pd black was used as the catalyst for TiO²⁺ reduction by H₂ in 6 N HCl (Figure 31). The activity of the catalyst increased over time, due to solubilization/precipitation of the Pd catalyst, arising from interconversion between insoluble Pd black (Pd⁰) and Pd²⁺ ions. The reaction of Pd²⁺ ions with H₂ generated Pd⁰, and the freshly deposited Pd black was found to be more active. The Ti³⁺/TiO²⁺ redox couple has an E° of 0.1 V vs. NHE and, therefore, is well-positioned thermodynamically to mediate H₂ oxidation.

Figure 31.

Redox scheme for TiO²⁺-mediated anode using H₂ fuel with a Pd⁰/Pd²⁺ catalyst. For relevant studies, see refs. 40 and 80.

A kinetic study of this system was performed by Folkesson.⁸⁰ TiCl₄ was dissolved in 10% HCl with 2% Pt/Al₂O₃. The reduction with H₂ exhibited a first-order dependence on [TiO²⁺]. A full fuel cell constructed using a vanadium/NO_x/O₂ cathode (H₂|Pt/Al₂O₃|TiO²⁺||VO²⁺|NO_x|O₂) achieved a peak power density of 15–20 mW/cm². The fuel cell was discharged through a 1-ohm resistor, and exhibited sustained performance for over 7 h (Figure 32).

Figure 32.

Discharge cell voltage vs. time for double MedFC. Cathode solution is 0.2 M VO₂⁺ in 1 M H₂SO₄ regenerated with HNO₃ and O₂ at 75 °C. Anode solution is 1 M TiCl₃ in 10% HCl regenerated with Pt/Al₂O₃ catalyst with H₂ at 60 °C (H₂|Pt/Al₂O₃|TiO²⁺||VO²⁺|NO_x|O₂). Adapted from ref. 80. Copyright 1990, Chapman and Hall Ltd.

3.1.3. Fe²⁺/Fe³⁺ mediators

The Fe²⁺/Fe³⁺ redox couple has been used as an anodic mediator for a variety of fuel oxidations. Complementary applications of the Fe²⁺/Fe³⁺ couple to mediate O₂ reduction were presented in Section 2.3.

Coordination of EDTA to Fe³⁺ decreases the reduction potential from ~0.7 V vs. NHE to 0.117 V vs. NHE. Folkesson took advantage of this decrease in potential to use Fe(EDTA)²⁻ as a mediator for H₂ oxidation, catalyzed by Pd/Al₂O₃, at pH 5 (Figure 33).⁸⁰ A fuel cell, incorporating VO₂⁺/VO²⁺ and NO_x in a mediated cathode for O₂ reduction, exhibited rather low power densities of 3 mW/cm² (H₂|Pd/Al₂O₃|Fe(EDTA)²⁻||VO²⁺|NO_x|O₂). Some of the drawbacks of this system include poor solubility of Fe(EDTA)²⁻ (max. solubility 0.075 M) and high resistance due to the raised anode solution pH. While the regeneration was shown to be effective, these issues have prevented further research into using Fe(EDTA)²⁻as a mediator for H₂ oxidation.

Figure 33.

Redox scheme of Fe(EDTA)²⁻-mediated anode for H₂ oxidation with Pd/Al₂O₃ catalyst. See ref. 80.

Whitesides and coworkers reported the use of the Fe²⁺/Fe³⁺ redox couple in a variety of fuel cell applications (cf. Section 2.3 for cathode examples). In the first report, the Fe²⁺/Fe³⁺ redox couple was used to mediate methane oxidation with a Pt black catalyst (c.f. Figure 7).⁸² An analogous MedFC approach was then demonstrated with methanol, ethylene glycol (and oxidized derivatives), and glycerol oxidation to CO₂ as the anodic process.⁸³ Other fuels, such as ethanol, acetic acid, or glucose, were oxidized, but CO₂ was not detected. Using a gas buret, the rate of CO₂ formation was measured. Potassium oxalate was oxidized to CO₂ the fastest, followed by ethylene glycol and glyoxal (Figure 34). Using a cell similar to that used with methane (cf. Figure 7), a fuel cell utilizing ethylene glycol as the fuel produced a maximum power density of 9.9 mW/cm³ at a potential of 197 mV with a graphite felt electrode.

Figure 34.

Rate of CO₂ formation vs. time for the oxidation for some organic compounds catalyzed by Pt black with Fe³⁺ as electron donor. Reactions were conducted at 80 °C in 3 M H₂SO₄. Adapted from ref. 83. Copyright 1996, Academic Press.

The Fe²⁺/Fe³⁺ mediator system was also extended to the use of sub-bituminous coal, Kraft lignin, and peat as fuels, without any added catalyst.⁸⁴ These efforts demonstrated the ability of mediators to enable the use of fuels that would be highly challenging in conventional fuel cell configurations. While biomass-derived sources were capable to be used as fuels, primary efforts focused on the use of sub-bituminous coal. When oxidizing the fuel in 5 M H₂SO₄ saturated with Fe³⁺ at 100 °C, approximately 40% of the available electrons in the coal were liberated. Polarization curves for this system were collected, and a maximum power of 0.6 W per liter of slurry solution was obtained when using a solution containing 4 g of coal in 15 mL of 5 M H₂SO₄ saturated with Fe₂(SO₄)₃. Simulations of the data suggested that the cell resistance limited the performance, and it was suggested that volumetric power densities of over 7 W/L should be achievable with improved cell engineering. The fuel cell was run continuously for 1000 h (with necessary additions of coal every ~100 h) with no significant degradation in cell performance.

Deng and coworkers used the Fe²⁺/Fe³⁺ mediator couple to oxidize wheat straw in a fuel cell using VO²⁺/VO₂⁺ as the cathodic mediator.⁸⁵ Using suspensions of wheat straw in aqueous solutions of FeCl₃ and HCl, Fe³⁺ was reduced to Fe²⁺, eventually solubilizing the wheat straw. The fuel cell constructed using these redox couples in a two-stage process achieved power densities of over 100 mW/cm² when 10 g of wheat straw was used in 100 mL of electrolyte (Figure 35A). The performance stability of the cell was examined in a one-stage process, and, with a discharge voltage of 0.3 V, a stable current density of 320 mA/cm² was achieved for 12 h (Figure 35B). Based on the values given, each Fe²⁺ ion was oxidized approximately 1.4 times over the course of this test. The degradation products of wheat grass were hypothesized to contain vanillin and other aromatics, which over time were further oxidized to CO₂. Analysis of the biomass after completing the fuel cell test suggested that hemicellulose was the primary reactive component, with cellulose and lignin remaining largely intact.

Figure 35.

A) Polarization curves and B) Continuous one-stage operation of a fuel cell using the Fe²⁺/Fe³⁺ redox couple as a mediator for wheat straw oxidation and a VO₂⁺/VO²⁺-mediated, HNO₃-catalyzed cathode for O₂ reduction (wheat straw|Fe²⁺||VO²⁺|NO_x|O₂). Adapted from ref. 85. Copyright 2017, John Wiley & Sons, Inc.

In 2017, Li and coworkers reported a MedFC using various carbohydrates as the fuel. A combination of FeCl₃ with low loadings of H₃PW₁₂O₄₀ or H₃PMo₁₂O₄₀ POM was the most effective mediator system for electricity generation.¹¹⁰ Their initial studies focused on testing high loadings (0.3 M) of H₃PMo₁₂O₄₀ or H₃PW₁₂O₄₀ as the sole catalyst/mediator for oxidation of glucose, but later studies found that using 1 M FeCl₃ and 0.06 M H₃PW₁₂O₄₀ or 1 M FeCl₃ and 0.03 M H₃PMo₁₂O₄₀ yielded similar results as 0.3 M H₃PMo₁₂O₄₀ (Figure 36). A maximum power density of approximately 2.6 mW/cm² was achieved. These experiments were conducted in a two-stage process, in which the glucose fuel was heated with the catalyst/mediator system at 85–95 °C for 2 h before introduction into the fuel cell at room temperature. Additionally, both 0.3 M H₃PMo₁₂O₄₀ and 1 M FeCl₃ + 0.06 M H₃PW₁₂O₄₀ were competent catalyst systems for the oxidation of starch (max power densities of ~1.6 mW/cm²), while 1 M FeCl₃ + 0.06 M H₃PW₁₂O₄₀ was more efficient for the oxidation of cellulose, which enabled a maximum power density of 0.72 mW/cm², compared to 0.46 mW/cm² with 0.3 M H₃PMo₁₂O₄₀. These results indicate that the addition of FeCl₃ could provide a means to reduce the POM loading in biomass oxidation (see Section 3.1.4 for POM-only based methods).

Figure 36.

Polarization curves for mediated fuel cell using POM or POM + FeCl₃ as mediator for glucose oxidation. H₁₂P₃Mo₁₈V₇O₈₅ was used as a cathodic mediator. The anode contained 1.0 M glucose and 1.5 M phosphoric acid and was operated at 23–25 °C. Mediator concentrations in the various experiments were as follows: 0.3 M PMo₁₂, 1.0 M FeCl₃, 0.06 M PW₁₂ + 1.0 M FeCl₃, 0.03 M PMo₁₂ + 1.0 M FeCl₃, and 0.3 M PW₁₂. Adapted from ref. 110.

FeCl₃ was used as a mediator for the oxidation of glucose and other biomass sources by Deng and coworkers in 2018.⁸⁶ The kinetics of this oxidation process were studied, and a nearly first-order dependence of the rate on [glucose] and [FeCl₃] was observed under the acidic conditions used. The byproducts of glucose oxidation were identified as CO₂, formic acid, acetic acid, and 4-hydroxybutyric acid. The best fuel cell performance, with a power density of approximately 125 mW/cm², was achieved with a two-stage approach in which 0.33 M glucose and 2.0 M FeCl₃ were heated at reflux for 4 h prior to electrochemical reoxidation of the Fe²⁺. The charged solution was capable of maintaining a current density of 250 mA/cm² (stated 105 mW/cm² power density) over the course of 1 h. In a closely related process, Fe³⁺-mediated oxidation of biomass (e.g., lignin and sugars) has been conducted in electrolysis cells to produce H₂ via proton reduction at the cathode, rather than produce power via reduction of O₂ at the cathode (Figure 37).^156–158

Figure 37.

Schematic of H₂ electrolysis system using Fe²⁺/Fe³⁺ mediator for the oxidation of biomass. Adapted from ref. 157. Copyright 2017, Elsevier Ltd.

3.1.4. Polyoxometalate anodic mediators

Polyoxometalates (POMs) have been described in several sections above, most extensively as catalysts/mediators in mediated O₂ reduction (see Section 2.2). They have also been used as catalysts and/or mediators for the oxidation of hydrogen, alcohols and biomass in fuel cells, as briefly noted in the previous section. These soluble mediators interact effectively with suspensions of heterogeneous and/or polymeric biomass materials, and display higher activity for C–C bond cleavage and are less easily poisoned by biomass impurities relative to heterogeneous catalysts supported on electrodes.

In connection with studies of redox flow batteries, Oei and Kummer were investigating a series of mediators (e.g., Sn²⁺/Sn⁴⁺, E° = 0.15 V vs. NHE; Cu²⁺/Cu, E° = 0.153 V vs. NHE; Fe²⁺/Fe³⁺, E° = 0.77 V vs. NHE) and testing the ability of the oxidized mediator to be reduced by H₂ following electrochemical oxidation.⁷⁶ Silicotungstic acid (H₄SiW₁₂O₄₀) exhibited especially favorable properties, with a reduction potential close to 0 V vs. NHE and an ability to be reduced by H₂.⁶³ While the uncatalyzed reduction at pH 0.5–1 is too slow to observe at room temperature, a variety of catalysts, including PtO₂, PtCl₄, or Pt/SiO₂, in addition to tungsten carbide and Pd(4,4′,4″,4‴-tetrasulphophthalocyanine), enhanced the rate. At 75 °C, the silicotungstic acid was reduced to a lesser extent than at 25 °C due to a temperature-dependent shift in its E°, and H₂ evolved when a solution reduced at 25 °C was heated to 75 °C. Polarization curves using a VO²⁺/VO₂⁺ mediated cathode displayed a cell voltage of 568 mV at 80 mA/cm² current density. In an extended test of the fully constituted system using a Pt/SiO₂ catalyst for reduction of H₄SiW₁₂O₄₀, good anode performance was maintained over 20 h, and the performance of the system was limited by the cathode performance. A follow-up study revealed improved cell performance using a VO²⁺/VO₂⁺ mediated cathode with HNO₃ as a catalyst, reflecting enhanced cathode system stability (Figure 38).⁷⁹

Figure 38.

Sustained performance curve of a doubly mediated H₂/O₂ fuel cell. The cathode utilized a VO₂⁺ mediator reoxidized with HNO₃ in 1 M H₂SO₄ at 75–80 °C, and the anode contained H₄SiW₁₂O₄₀ reduced by H₂ with a Pt/silica catalyst at 25–30 °C (H₂|Pt|H₄SiW₁₂O₄₀||VO²⁺|NO_x|O₂). The cell was discharged through a 2.1-ohm resistor. Flow rate was 230 mL/min. I₀, I₄, and I₁₂ correspond to the current at 0, 4, and 12 h. Adapted from ref. 79. Copyright 1985, Chapman and Hall Ltd.

In 2014, Deng and coworkers demonstrated photochemical and thermal oxidation of starch solutions and other biomass sources by POMs at the anode of a MedFC,¹⁵⁹ using a conventional Pt/C O₂ cathode. The biomass solution with POM was thermally or photochemically pre-treated and then flowed through the anode compartment in a two-stage process (c.f. Figure 30B). Using this approach with H₃PMo₁₂O₄₀ as the POM, power densities of 0.28 mW/cm² (photochemical pretreatment) and 0.44 mW/cm² (thermal pretreatment) were measured. In a continuous experiment, in which the POM/starch solution was continuously exposed to light irradiation and heated to 95 °C, a steady current density of 2.5 mA/cm² was observed for over 20 h. Polarization curves using poplar and switch grass as fuels and photochemical pretreatment displayed peak power densities of 0.65 and 0.62 mW/cm², respectively.

A subsequent report by Deng and coworkers described a MedFC with POM mediators for both the anode and cathode (cf. Figure 12 in Section 2.2):¹⁰⁷ H₃PW₁₁MoO₄₀ (0.3 M) as the anodic mediator and H₁₂P₃Mo₁₈V₇O₈₅ (0.3 M) as the cathodic mediator. A two-stage process was used for the anode, in which the H₃PW₁₁MoO₄₀ solution was heated with the biomass at 100 °C for 4 h prior to electrochemical treatment and, following this pretreatment procedure, the two POM solutions were passed through their respective electrode compartments and polarization curves were recorded. A variety of biomass sources were tested using this two-stage protocol (Figure 39A). Fresh grass displayed the highest power densities (51 mW/cm²), followed by switchgrass (45 mW/cm²), starch solution (34 mW/cm²), and cellulose (22 mW/cm²). One-stage performance of this system (c.f. Figure 30A) using starch as the fuel source was also demonstrated, displaying a power density of 30 mW/cm² at a constant discharge current of 160 mA/cm² for greater than 10 h (Figure 39B). A large portion of the biomass was converted into CO₂: 88% of the initial glucose, 82% of the cellulose, and 65% of the lignin was converted to CO₂, as measured by total organic content remaining after 12 thermal treatment cycles. Photochemical treatment led to nearly complete conversion to CO₂ after 200 hours of photolysis; however, the authors did not report electrochemical performance over these cycles. Overall, these studies highlight the ability of POMs to cleave C–C bonds in biomass while also being capable of regeneration at the anode.

Figure 39.

(A) Polarization curves for a variety of biomass sources at 80 °C using 0.3 M H₃PW₁₁MoO₄₀ as the anodic mediator and 0.3 M H₁₂P₃Mo₁₈V₇O₈₅ as the cathodic mediator. (B) Power density measured over 10 h of operation using starch as biomass source. Adapted from ref. 107. Copyright 2014 WILEY-VCH Verlag GmbH & Co.

A similar MedFC was reported by Zhu and coworkers using various lignins as the biomass source and H₃PMo₁₂O₄₀ as the POM at the anode with a conventional Pt/C O₂ cathode.¹¹³ Polarization curves were collected at room temperature following the reaction of a POM/lignin combinations under a variety of conditions (i.e., in a two-stage approach; cf. Figure 30B). Peak power densities approached 0.35 mW/cm² for reaction mixtures derived from lignin loadings of 25 g/L and a POM concentration of 0.1 M. Replacement of the Pt/C O₂ cathode with a POM-mediated cathode system generated power densities of up to 5 mW/cm² at 90 °C. The POM source for the cathode was identical to that used at the anode (H₃PMo₁₂O₄₀), and power generation was achieved from the different redox states of the POM in the two electrode compartments.

In 2016, Deng and coworkers reported a MedFC using polyols as the fuel in a two-stage anodic process.¹⁰⁸ The anodic mediator/fuel solution, containing 2 M polyol and 0.2 M H₃PMo₁₂O₄₀, was subjected to irradiation for 17.5 h with a 250 W metal halide lamp at 25 °C. This solution was then passed through the anode compartment of the MedFC, which employed a 0.3 M H₁₂P₃Mo₁₈V₇O₈₅ as a cathodic POM mediator. Using this approach, polarization curves were collected using various alcohol-based fuels. In contrast to conventional fuel cells, in which methanol is typically the best alcohol fuel, the authors showed that glucose performed much better than methanol. The maximum power density attained with glucose was ~ 31.5 mW/cm², while use of methanol led to a power density of ~8 mW/cm². The authors also evaluated ethylene glycol, glycerol, erythritol, xylitol, and sorbitol (Figure 40A), and the fuels with more hydroxyl groups led to enhanced power densities at identical fuel concentration (2 M, Figure 40B) or at the same effective hydroxyl concentration of 6 M (i.e. 6 M methanol, 3 M ethylene glycol or 2 M glycerol). UV-visible experiments suggested that increased association between polyhydroxylated fuels and the POM increased the POM reduction by fuels containing more hydroxyl groups. This pre-association was also supported by NMR spectroscopic data. As observed previously,¹⁰⁷ carbon dioxide is the major product arising from these POM-mediated photochemical oxidation reactions of polyols.

Figure 40.

(A) Structures of alcohols used as fuels. (B) Polarization curves obtained using anodic mediator solutions prepared by irradiation of solutions containing 2 M of the alcohol and 0.2 M H₃PMo₁₂O₄₀ as the mediator. Adapted from reference 108. Copyright 2016, Elsevier B.V.

Chao and coworkers reported a biomass fuel cell with glucose and other sugars as the fuel, using H₃PMo₁₂O₄₀ as the anodic mediator and H₁₂P₃Mo₁₈V₇O₈₅ as the cathodic mediator.¹¹² A two-stage anodic process was used, and a maximum MedFC performance of 40 mW/cm² was obtained using an anodic solution prepared by photolysis of the POM and glucose solution at 80 °C for 8 hours. Thermal measurements showed that the temperature of the photolysis reaction increased substantially during irradiation, which in turn enhanced the rate of glucose oxidation. The overall efficiency, considering all energy inputs and outputs, of this fuel cell was calculated to be 36.7%. Other sugars, such as fructose and xylose, as well as polysaccharides, such as raffinose and maltose, were found to be nearly as effective of fuels as glucose.

Deng and coworkers reported a MedFC that used coal as a fuel and POM mediators in the anode.¹⁰⁹ Lignite, the lowest grade of coal, was used as the coal source, and it was combined with H₃PMo₁₂O₄₀ under a variety of conditions to examine reduction of the POM mediator in the first stage of a two-stage anodic process. The best fuel cell results were obtained by treatment of 10 g lignite in 0.1 L of a 0.3 M POM solution at 200 °C for 5 h. The resulting solution of reduced POM mediator was cooled, filtered to remove coal particulates, and then used at the anode of a MedFC operated at 80 °C with a H₁₂P₃Mo₁₈V₇O₈₅-mediated cathode. The maximum power density observed was 120 mW/cm². Continuous discharging of the fuel cell using a single batch of reduced anodic POM solution was conducted at 150 mA/cm², and the maximum power density decreased from 60 to ~30 mW/cm² over 10 h. The coal oxidation was repeated three times with the same batch of coal, consuming approximately 9% of the coal mass in each case. Diminished performance was observed on the third run. Characterization of the coal residue after the third reaction showed a decrease in the total organic carbon, indicating oxidation of the coal to CO₂ or water-soluble low molecular weight species, and ¹³C NMR spectroscopic data suggested that aliphatic-like carbonaceous material was preferentially oxidized over aromatic content.

Deng and coworkers also reported a MedFC that utilized sewage sludge as the fuel.¹¹¹ This system employed H₃PMo₁₂O₄₀ as the anodic POM mediator and H₁₂P₃Mo₁₈V₇O₈₅ as the cathodic POM mediator. The authors examined the role of temperature and treatment time on the oxidation of the sewage sludge by H₃PMo₁₂O₄₀. Power densities of 13.56 mW/cm², 16.07 mW/cm², and 17.91 mW/cm² were obtained from anodic solutions prepared by treatment of the sewage sludge at 80 °C for 6 h, 12 h, and 24 h, respectively. Further improvements in power densities were achieved by treatment of the sludge for 24 h at 100 °C and 150 °C (49.8 mW/cm² and 61.4 mW/cm², respectively; Figure 41). The first oxidation treatment removed over 60 wt% of the sludge sample, which increased to 85 wt% after a second treatment. The authors additionally showed that the POM could undergo repeated cycling with addition of fresh fuel after each cycle. Four cycles were conducted with each showing similar performance.

Figure 41.

Polarization curves for sewage sludge-powered mediated fuel cell using H₃PMo₁₂O₄₀ as anodic mediator and H₁₂P₃Mo₁₈V₇O₈₅ as cathodic mediator. Sewage sludge was treated with H₃PMo₁₂O₄₀ for 6, 12, 24, or 36 h at (a) 80 °C, (b) 100 °C, or (c) 150 °C prior to the fuel cell experiment. Adapted from reference 111. Copyright 2017 Elsevier Ltd.

In 2017, Zhao and coworkers demonstrated the oxidation of lignin in wheat straw to produce electricity using a two-stage process incorporating H₃PMo₁₂O₄₀ as an anodic mediator.^160,161 Enzyme-digestible carbohydrates were obtained as a by-product of the biomass oxidative pretreatment. The extent of deliginification, the level of xylan and glucan dissolution, the ability of the remaining glucans to be enzymatically hydrolyzed, and the degree of reduction of the POM were assessed under a variety of conditions. Solutions containing reduced POM were introduced into the anode compartment of a MedFC, in which the cathode solution contained a solution of FeCl₃. After testing a variety of conditions, the maximum power density obtained was approximately 11 mW/cm².

POMs have also been utilized as mediators for biomass oxidation in combination with electrochemical production of H₂ and CO₂ reduction at the cathode. In these systems, instead of reducing O₂ at the cathode to generate power, power is applied to the system and either protons are oxidized to produce H₂ (Figure 42)^{156,162–163} or CO₂ is reduced to CO and hydrocarbons.¹⁶⁴ The anodic performance of these systems is quite similar to the fuel cell systems.

Figure 42.

Schematic of electrochemical H₂ evolution using biomass as H₂ source and POM-based mediators. Adapted from reference 162 with permission from The Royal Society of Chemistry. Copyright 2016 The Royal Society of Chemistry.

3.1.5. Other transition metal-based mediators

Two additional examples of transition metal-based mediators have been only briefly studied. The first case was reported by Kummer and Oei, who were searching for a mediator solution for H₂ oxidation with higher volumetric charge capacity than silicotungstic acid (c.f. Section 3.1.4).⁷⁹ Their work was inspired by a precedent for 2 e⁻ oxidation of a trinuclear molybdenum ion, [(Mo³⁺)₂Mo⁴⁺], to [Mo⁴⁺]₃ with an E° very close to 0 V vs. NHE. Using Pt catalysts, a solution containing Mo⁶⁺ was cleanly reduced to Mo⁴⁺ by H₂, but only half of the species were further reduced to the Mo³⁺ state, compared to the expected value of two-thirds. This outcome fell short of expectations, but it still led to more reduced metal sites than the one-twelfth observed in silicotungstic acid. When used in a fuel cell along with a VO₂⁺/VO²⁺-mediated, NO_x-catalyzed O₂ cathode (H₂|Pt|Mo³⁺||VO²⁺|NO_x|O₂), a current density of 85 mA/cm² at approximately 550 mV (iR-free voltage of 725 mV) was obtained. No long-term stability experiments were reported using this mediated anode.

In 2018, Bullock and coworkers reported a fundamental study of electrochemical H₂ oxidation under non-aqueous conditions using a mediator/co-catalyst system comprising Cp*Cr(CO)₃ and an Fe(P₂N) complex (Figure 43).¹⁶⁵ In this report, H₂ underwent homolytic cleavage by two equivalent of Cp*Cr(CO)₃ to form two Cp*Cr(CO)₃H species. This complex can transfer an H-atom to [Fe(P^EtN^PhP^Et)(CO)₃]⁺, and the resulting species undergoes deprotonation and electrochemical oxidation to regenerate the five-coordinate Fe species. A catalytic wave was observed by cyclic voltammetry at the Fe^+/0 redox wave, and it increased with higher 2-methylpyridine concentration. Each complex was required for competent H₂ oxidation, and controlled potential electrolysis with 2,6-di-tert-butylpyridine as a base gave a lower limit of 5 turnovers for Co and 21 turnovers for Fe.

Figure 43.

Proposed mechanism for electrochemical H₂ oxidation using Cp*Cr(CO)₃/Fe(P₂N) system. Adapted from reference 165. Copyright 2018 Wiley-VCH Verlag GmbH & Co.

3.1.6. Quinone/Hydroquinone mediators

In the mid-1980s, Weetall and coworkers at Corning Glass Works investigated the use of various biomass sources as fuels for a MedFC, using sulfonated anthraquinone (AQ) derivatives as anodic mediators. They examined simple alcohols,¹⁶⁶ carbohydrates,¹⁶⁷ and sulfonated lignin or Kraft Black liquor¹⁶⁸ (a byproduct of wood-pulping) as fuels. To facilitate the oxidation of the alcohol or carbohydrate fuels, visible light irradiation was necessary to excite the anthraquinone mediator. Basic (pH > 12) conditions were used. With ethylene glycol as the fuel (10 volume %), currents of approximately 1 mA were observed, corresponding to a current density of about 0.1 mA/cm². No potentials, power densities, or sustained performance data for simple alcohol fuels were given. More thorough analysis was performed when carbohydrates, such as glucose, sucrose, or unpurified molasses were used. The highest power output was obtained with glucose as a fuel (0.57 mW/cm²), followed by invertase-treated sucrose (0.43 mW/cm²) and fructose (0.41 mW/cm²). In each of these experiments, 6 mM anthraquinone sulfonate was used as mediator. The authors calculated that 6.4 pounds of glucose would be required to generate one kWh of energy using their cell setup. Under similar conditions using lignin-based fuels, peak power densities of approximately 0.34 mW/cm² were obtained. Further experiments indicated that the initial polymeric lignin material was converted into low-molecular weight fragments during fuel cell operation, though the identities of these fragments was not further investigated. It is likely that the performances of the cells could have been substantially increased through further engineering of the cell design. For example, the authors observed a 50% increase in currents merely by mixing the system.

Stahl and coworkers reported a MedFC utilizing a water-soluble anthraquinone mediator, anthraquinone-2,6-disulfonic acid (AQDS) in 1 M H₂SO₄ paired with an off-electrode Pt/C catalyst for mediated electrochemical H₂ oxidation.¹⁰⁵ The fuel cell was initially constructed using a conventional Pt/C O₂ cathode; however, substantial mass-transport losses were observed in the polarization curves due to cathode flooding. To address these concerns, a POM-mediated cathode utilizing 0.3 M Na₄H₃PMo₈V₄O₄₀ was used (Figure 44). Polarization curves were obtained, and a peak iR-free power density of 528 mW/cm² was reported (228 mW/cm² without iR-correction). The system also displayed sustained performance at a current density of 50 mA/cm² for over 8 h. These MedFC performance metrics are substantially lower than those of conventional PEMFCs, but considerable better than metrics achieved with previous applications of MedFCs using H₂ as the fuel (cf. sections 3.1.3, 3.1.4, and 3.1.5) Moreover, formic acid and methanol were shown to reduce the AQDS mediator using the Pt/C catalyst, suggesting this approach could potentially be adapted to achieve improved fuel cell performance with fuels more complex than H₂.

Figure 44.

Schematic of MedFC utilizing AQDS as an anodic mediator for H₂ oxidation with a Pt/C catalyst and a POM-mediated cathode. Adapted from reference 105.

3.1.7. V³⁺/VO²⁺ mediator

A patent published in 1997 by Larsson and Folkesson reported a MedFC using vanadium salts and a Pt catalyst to oxidize glucose to CO₂ as the anodic reaction (Figure 45) and VO²⁺/VO₂⁺ as the cathodic redox couple (glucose|Pt|VO³⁺||VO²⁺|catalyst|O₂).¹⁶⁹ In this cell, VO²⁺ was reduced to V³⁺ (with some further reduction to V²⁺). Using vanadium species in both the cathode and anode eliminated concerns with mediator crossover, resembling advantages evident in all-vanadium redox flow batteries.⁷⁴ A cell capable of producing 10 kW with a power density of 10.5 mW/cm² (cell voltage of 0.7 V with current density of 15 mA/cm²) was developed, using an anode solution of 180 liters containing 0.3 M VOSO₄, 5 M H₃PO₄, 0.003 M sugar, and 229 g Pt. Overall, 4 stacks were connected in parallel, with each stack consisting of 130 cells in series, in order to deliver the desired power output. A further kinetic study of the oxidation of sugars by VO²⁺ was reported by the same authors in 2005.¹⁷⁰ They found that low concentrations of sugar were necessary for complete oxidation to CO₂. A variety of sugars (sucrose, fructose, glucose, arabinose, xylose) all behaved similarly. Continued optimization of this cell led to power conversion efficiencies of 41–45%, which were suggested to exceed efficiencies accessible by thermal combustion of the sugar.⁸¹

Figure 45.

Redox scheme for V³⁺-mediated anode using glucose fuel and Pt catalyst. See refs. 169–170.

3.1.8. Viologen-based mediators

Viologens (1,1’-dialkyl-4,4’-bipyridyls) are cationic, low-potential, redox-active organic species that have been used as mediators for the oxidation of glucose and other carbohydrate derivatives under basic conditions. In 2009, a pair of publications, one from Wheeler and coworkers¹⁷¹ and one from Scott and Liaw,¹⁷² reported fuel cells incorporating viologens as mediators for electrochemical glucose oxidation.

Wheeler and coworkers focused on the oxidation of glucose by methyl viologen (MV²⁺, Figure 46A) under mildly basic conditions (pH 9–12).¹⁷¹ Anaerobic experiments indicated ~8 equiv of MV²⁺ were reduced in the oxidation of glucose when high MV²⁺:glucose ratios were used, compared to the 24 equiv theoretically necessary for full oxidation. However, exposure to O₂ under these conditions led to enhanced consumption of glucose, and, at a 12:1 MV²⁺/glucose ratio, significant amounts of carbonate and formate were observed. Preliminary experiments with a simpler substrate (dihydroxyacetone) indicated that a fuel cell could operate using MV²⁺ as a mediator and a cathode containing a cobalt oxide-based catalyst originally developed for zinc-air batteries. Liaw and coworkers demonstrated improved performance in a related fuel cell,¹⁷² showing that with 3 M KOH, 2 M glucose, and 28 mM MV²⁺, a power density of 2.5 mW/cm² could be obtained with an air-breathing cathode (Figure 46B–C). Liaw proposed that glucose oxidation proceeded only to d-gluconolactone. However, no characterization data for this proposed species was reported. In a later study, Liaw and coworkers provided spectroscopic evidence that the major product of glucose oxidation was gluconic acid when indigo carmine was used as a mediator.¹⁷³ One explanation for the discrepancy between the fate of glucose between the Liaw and Wheeler reports is the difference in solution basicity used in these two studies, which may influence the reaction pathways during glucose oxidation.

Figure 46.

A) Structure of methyl viologen. B) Schematic of MV²⁺-mediated (“Dye²⁺”) fuel cell using glucose fuel and air breathing cathode. C) Polarization curve for glucose-O₂ fuel cells with MV²⁺ as anodic mediator. Cell contained 2 M glucose and 28 mM MV²⁺ in 3 M KOH solution. Adapted from ref. 172. Copyright 2009, The Royal Society of Chemistry.

A later paper, published in 2013 by Liu and coworkers, reported activities similar to those of Liaw and coworkers for their fuel cell using MV²⁺ as a mediator for glucose oxidation.¹⁷⁴ Further work was performed by Liu, Zhang, and coworkers, who reported a power density of 0.52 mW/cm² at room temperature with 1 M glucose, 3 M KOH, and 15 mM MV²⁺.¹⁷⁵ They detected the formation of both acetate and formate products in their cell.

Wang, Liu, and coworkers reported the use of cellulose as a fuel with a MV²⁺ mediator using a solution containing 5 M NaOH, 1 wt% (w/v) cellulose, and 15 mM MV²⁺.¹⁷⁶ This system achieved a peak power density of 45 μW/cm². The degradation products of cellulose oxidation were identified by HPLC as oxalic acid, glucaric acid, glyceric acid, glycolic acid, formic acid, and acetic acid.

To increase the performance of a glucose-powered fuel cell using viologen mediators, Watt and coworkers investigated the mechanisms of glucose oxidation and mediator degradation.¹⁷⁷ Under mildly basic conditions, they identified the reaction of an enediol formed from glucose under basic conditions with MV²⁺ as the important step for selectivity. High selectivity required high MV²⁺ loadings to prevent the formation of inactive intermediates. Additionally, it was noticed that dialkyl viologens decomposed under basic conditions (pH > 13) to monoalkyl viologens, so the activity of monoalkyl viologens for glucose oxidation was examined.¹⁷⁸ While glucose oxidation proceeded approximately 50% slower using monoalkyl viologens, the monoalkyl viologens displayed substantially higher stability than the dialkyl variants. In this initial report, higher OCP values also were reported for the monomethyl viologen (0.9 V) than for the dimethyl viologen (0.72 V).¹⁷⁹ Further kinetic studies were undertaken, and increased temperatures were found to substantially increase the rate of alcohol oxidations by methyl viologen and mono-methyl viologen.¹⁸⁰ To assist in the recovery of the viologen mediators, Watt and coworkers also examined the ability of soluble, polymeric viologens to act as mediators in glucose oxidation. Less basic conditions (pH = 11) were required to prevent decomposition of the viologen polymers.¹⁸¹ While reoxidation with O₂ allowed for effective glucose oxidation, attempts to perform the reoxidation electrochemically led to precipitation of the mediators from solution.

3.2. Fuel oxidations using enzymatic catalysts

3.2.1. Introduction

Enzymatic fuel cells are a broad class of devices, in which enzymes are used as catalysts for oxidant reduction or fuel oxidation.^55–57 Enzymatic fuel cells are different from microbial fuel cells, which use entire organisms to catalyze the desired reactions.⁵⁹ While the enzymes are occasionally used as soluble species in solution, they are typically tethered or otherwise immobilized onto the electrode. Some enzymes can undergo direct electron transfer at the electrode, but the buried active sites of many enzymes often leads to slow direct electron transfer. This problem is often addressed through the use of mediators, which can be soluble or immobilized on the electrode. Immobilization of both the enzyme and mediator is most commonly employed, as it has the benefit of removing the need for a membrane, thereby facilitating the miniaturization of enzymatic fuel cells, and minimizes the chance of releasing toxic mediators in implantable or wearable devices. Historical examples of common immobilized mediators include osmium- or ruthenium-containing redox polymers,¹⁸² and in recent years redox polymers containing organic mediators, such as quinones, viologens, or phenothiazines have been developed.¹⁸³ The discussion here emphasizes dissolved mediators, analogous to the chemical systems described above. Most enzymatic fuel cells operate at or near ambient conditions (room temperature or slightly above) at moderate or neutral pH values and generate low amounts of power. Power densities are typically ≤1 mW/cm², commensurate with that appropriate for implantable devices or sensors. A summary of enzymatic fuel cells using at least one soluble mediator is given in Table 8.

Table 8.

Enzymatic fuel cells using at least one soluble mediator in the anode or cathode.

Anode			Cathode
fuel	catalyst	mediator	mediator	catalyst	oxidant	ref
H2	Pt	–	ABTS2−	laccase	O2	146
	Desulfovibrio vulgaris	viologen		bilirubin oxidase		149
glucose	glucose oxidase	Fc(CH2OH)		bilirubin oxidase or laccase		147–151
glucose	glucose oxidase/AuNP-graphene	–		laccase/AuNP-graphene		152
BPA	laccase	–		laccase		153
glucose	glucose dehydrogenase	PS-βCD-PQ	PS-βCD-P2ABTS	bilirubin oxidase		154
MeOH	methanol dehydrogenase	PMS or PES	–	Pt		184–185
MeOH	quinoprotein alcohol dehydrogenase	TMPD				186
glucose	glucose oxidase		Os polymer	laccase		188
MeOH	multi-enzyme cascade	BV2+ +NADH	–	Pt		190
glucose	glucose oxidase	DCIP				189
		quinone				195
glucose-6-phosphate	G6PDH and diaphorase	quinone + NADH				196
MeOH	alcohol dehydrogenase	TMPD	–		KMnO4	187
H2	NiFe hydrogenase	MV2+	MV2+	MoFe Nitrogenase	N2	193

3.2.2. Phenazine-based mediators

Hill and coworkers reported an enzymatic fuel cell in which methanol dehydrogenase was used to oxidize methanol to formate with phenazine methosulfate (PMS) or phenazine ethylsulfate (PES) as a mediator (Figure 47A)¹⁸⁴ in combination with a conventional Pt/C O₂ cathode. PES exhibits greater stability and was shown to give better performance. The fuel cell was operated at pH 9.5 and temperatures from 25–30 °C. The current reached a saturation limit at higher concentrations of mediator and enzyme, and sustained operation was examined with two loadings of the enzyme (Figure 47B). The open circuit potential was ~0.3 V, and a maximum current of ~3 mA (corresponding to a current density of 0.4 mA/cm² based on the reported Pt electrode dimensions) was obtained across a 10 W resister with 4.6 mg enzyme, 3.5 mM PES, and 62 mM MeOH in 4 mL anode solution. Over the course of 80 min of fuel cell operation, the current dropped to ~1.5 mA (~0.2 mA/cm²). The authors suggested that this decrease could be due to PES blocking of proton flow through the membrane. A pH difference was observed between the anode and cathode compartments at the end of cell operation, supporting this conclusion.

Figure 47.

(A) Redox scheme for PES-mediated anode using methanol as a fuel and methanol dehydrogenase enzyme as catalyst and structures of PMS and PES. For systems using this mediated anode, see refs. 184–185. (B) Current output of mediated anode fuel cell using PES as anodic mediator. Anode compartment contained 2.3 mg (dashed line) or 4.6 mg (solid line) methanol dehydrogenase, 0.35 mL of 40 mM PES in water, 3.6 mL of buffer, and 10 μL methanol. Adapted from reference 184. Copyright 1981 Science and Technology Letters.

Later, Yue and Lowther revisited this system and attempted to improve on the results by including the enzyme formate dehydrogenase to oxidize the formate product obtained from methanol dehydrogenase.¹⁸⁵ The authors used conditions similar to those of Hill and coworkers, but they incorporated a carbon anode instead of a Pt mesh anode. The authors selected PMS as the mediator since they observed approximately 10% higher activities for MeOH and formaldehyde oxidation with PMS than with PES in control experiments. Over the course of a 60 min experiment using immobilized enzymes and soluble PMS, the current decreased from 2.14 mA (0.054 mA/cm²) to 1.57 mA (0.039 mA/cm²). It is unclear why the current densities of this system were substantially lower than those of Hill and coworkers. Further testing indicated that the formate dehydrogenase enzyme used in these experiments led to negligible current (only 0.03 mA) when formic acid was directly used as a fuel. This result contrasted the enzyme activity tests, which showed similar activity for formic acid oxidation catalyzed by formate dehydrogenase using PMS and for methanol oxidation catalyzed by methanol dehydrogenase with PMS.

3.2.3. Tetramethyl-4-phenylenediamine mediator

In an effort to incorporate mediators more stable than PES or PMS (see section 3.2.2), Hill and coworkers examined a variety of organic species as oxidants for methanol oxidation catalyzed by the quinoprotein alcohol dehydrogenase.¹⁸⁶ Through these efforts, N,N,N’,N’-tetramethyl-4-phenylenediamine (TMPD, Figure 48A) was identified as the best mediator. Cyclic voltammetry studies at pH 10.5 showed substantial catalytic activity at the TMPD reduction potential in the presence of methanol and enzyme. In the absence of substrate or enzyme, a reversible 1 e⁻ wave was observed at 0.01 V vs. SCE (+ 0.25 V vs. NHE). The stability of the TMPD mediator was then compared to PES in a fuel cell. Over a 6 h electrolysis at pH 10.5, the current decayed nearly to zero when PES was used as the mediator (Figure 48B, trace B). When TMPD was used at the mediator under the same conditions, the current remained stable over the entire 6 h electrolysis (Figure 48B, trace A), and decreased by less than 10% over 24 h. Polarization curves were collected, and a peak power of 12 μW was obtained, corresponding to a power density of 2 μW/cm². The low power density likely reflects, at least in part, the high mediator potential. While no open-circuit potential was reported, the potential at the lowest current indicated was only ~160 mV. The authors noted that this cell was not limited by the cathode chemistry, since no fuel cell improvement was observed when pure O₂ was purged into the cathode compartment rather than air. Formate dehydrogenase was included to oxidize the formate since formate inhibits the alcohol dehydrogenase, but poor enzyme performance at the relevant pH led to no improvement in the fuel cell performance.

Figure 48.

(A) Structure of N,N,N’,N’-tetramethyl-4-phenylenediamine (TMPD). (B) Current output of a fuel cell using A) 4.0 mM TMPD or B) 4.0 mM PES with alcohol dehydrogenase as the enzyme and methanol as fuel. Adapted from reference 186. Copyright 1983 Elsevier Inc.

Ranta and coworkers also studied the stability of TMPD in fuel cells utilizing MeOH/methanol dehydrogenase at pH 10.¹⁸⁷ A KMnO₄ solution was used as the oxidant at the cathode. Polarization curves revealed a peak power density of 0.25 mW/cm² at a current density of 0.38 mA/cm² and cell voltage of 0.67 V. Most of the polarization losses arose from the anode. The product of methanol oxidation with this enzyme is formate, which led to a decrease in pH that led to a decrease in the system performance during extended operation. TMPD decay (via dimerization or oligomerization) was observed, with an estimated half-life of approximately 5 days during fuel cell operation.

In 2009, Leech and coworkers reported an enzymatic fuel cell using a glucose oxidase enzyme as a catalyst for glucose oxidation with TMPD as a mediator.¹⁸⁸ The cathode consisted of a laccase enzyme co-immobilized with an osmium redox polymer acting as a redox mediator. This study primarily focused on the cathode reaction, but the authors required a highly active anode catalyst and mediator combination. A maximum power density of 52 μW/cm² was observed, with the performance of the system being limited by the cathodic reaction.

3.2.4. Dichlorophenolindophenol mediator

In 1984, Laane and coworkers reported a biofuel cell using glucose as a fuel and D–glucose oxidase as the catalyst with a dichlorophenolindophenol (DCIP, Figure 49A) mediator.¹⁸⁹ Glucose was oxidized to gluconic acid, which the authors identified as a valuable industrial chemical. In their initial studies, O₂ reduction to H₂O on a Pt electrode was the cathodic reaction, and sustained currents of 0.3 mA were obtained for over 200 min of cell operation. In a later experiment, a gold cathode was used to reduce O₂ to H₂O₂, which the authors used to oxidize barbituric acid to 5-chlorbarbituric acid with chloroperoxidase as a catalyst (Figure 49B). Limited information was given about the power generation capabilities of this system. Over 3 days of operation, 10 mg gluconic acid and 8 mg 5-chlorobarbituric acid were generated, corresponding to 1.8 × 10⁴ turnovers for D-glucose oxidase.

Figure 49.

(A) Structure of dichlorophenolindophenol (DCIP). (B) Schematic of biofuel cell used for the synthesis of 5-chlorobarbituric acid. The anode was mediated by DCIP. Adapted from reference 189. Copyright 1984 Elsevier Inc.

3.2.5. Viologen-based mediators

In efforts to increase the power densities of an enzymatic fuel cell using MeOH as a fuel, Palmore and coworkers were interested in decreasing the overpotential between the anodic mediator and MeOH.¹⁹⁰ Previously reported mediators typically have overpotentials of at least 700 mV (cf. sections 3.2.2. and 3.2.3.), presumably because lower overpotential mediators were not sufficiently active with the chosen enzymes. However, the high potentials greatly reduce the potential power output of the fuel cell. To enable the use of lower-potential mediators, three NAD⁺-dependent enzymes were used to catalyze the oxidation of methanol to CO₂: alcohol dehydrogenase (ADH), aldehyde dehydrogenase (AldDH), and formate dehydrogenase (FDH). Three equivalents of NAD⁺ were required to complete the 6 e⁻ oxidation of MeOH. The NAD⁺/NADH redox couple has slow electrode kinetics, however, that often require overpotentials of >1 V to achieve suitable electron-transfer rates.¹⁹¹ To facilitate the oxidation of NADH, the enzyme diaphorase was used with another soluble mediator, benzyl viologen (BV²⁺), and a fuel cell was constructed using these mediators and enzymes in the anodic compartment and a Pt-black O₂ cathode (Figure 50). The cell exhibited an OCP of approximately 0.8 V, and polarization measurements exhibited a peak power density of 0.68 mW/cm². The rate-determining step in the fuel cell was identified as the oxidation of NADH to NAD⁺, catalyzed by diaphorase with BV²⁺ as the oxidant.

Figure 50.

Schematic of a methanol/oxygen biofuel cell, in which methanol is oxidized to CO₂ by NAD⁺-dependent dehydrogenases ADH, AldDH, and FDH. Diaphorase (D) catalyzes the oxidation of NADH to NAD⁺ with benzylviologen, BV²⁺, as an electrochemical mediator. Adapted from reference 190. Copyright 1998 Elsevier S.A.

Enzymes have been extensively studied for electrocatalytic H₂ oxidation,¹⁹² but most of these studies utilize heterogenized hydrogenase enzymes. Recently, Minteer and coworkers reported a fuel cell using a solubilized hydrogenase enzyme and methyl viologen (MV²⁺) as a mediator.¹⁹³ N₂ reduction, catalyzed by nitrogenase and mediated by MV²⁺, was performed at the cathode (Figure 51). With the same mediator used in both anode and cathode, this cell would not supply power at equilibrium, but the shift in equilibrium concentrations caused by substrate oxidation/reduction yielded a potential difference. This cell demonstrated an OCP of 228 mV and achieved a maximum power density of 1.50 μW/cm². Dinitrogen was reduced to ammonia in this fuel cell, with an overall yield of 286 nmol NH₃/mg MoFe protein and a cathodic Faradaic efficiency of 26.4%. This enzymatic fuel cell is noteworthy for its ability to produce NH₃ from N₂ and H₂ while also generating electricity. The authors also demonstrated that an anode consisting of glucose directly oxidized by MV²⁺ under carbonate buffer (pH=11) conditions was also feasible (see section 3.1.7). The OCP under these conditions was 238 mV and the maximum power density was 0.76 μW/cm².

Figure 51.

Schematic of biofuel cell using methyl viologen as a mediator in both anode and cathode. The cell produces NH₃ from N₂ reduction in the cathode compartment. Adapted from ref. 193. Copyright 2017, John Wiley & Sons, Inc.

3.2.6. Ferrocene-based mediators

Ferrocene derivatives, such as ferrocene methanol or ferrocene carboxylic acid, have found widespread use as mediators for enzymatic reactions, though many of these examples employed in sensing applications rather than power generation.¹⁹⁴ Dong and coworkers employed a porous carbon electrode with an immobilized glucose oxidase enzyme was paired with ferrocene monocarboxylic acid as a mediator in the anodic compartment of an enzymatic fuel cell.¹⁴⁷ The cathode compartment contained a laccase enzyme and ABTS²⁻ as the mediator. A peak power density of 99.8 μW/cm² at pH 4 decreased to 2.0 μW/cm² at pH 7 due to the lower activity of the laccase enzyme in the cathode compartment at higher pH (see section 2.7 for further discussion).

Dunn and coworkers reported an enzymatic fuel cell with glucose oxidase immobilized in a carbon nanotube-containing sol-gel matrix, using ferrocene methanol as a mediator and glucose as the fuel.¹⁵⁰ The cathode compartment contained immobilized bilirubin oxidase and ABTS²⁻ as a soluble mediator for O₂ reduction (see Section 2.7). An open circuit potential of 480 mV was obtained with this cell. This value is substantially larger than the difference in redox potentials between ABTS²⁻ (0.480 V vs. Ag/AgCl) and (Fc)CH₂OH (0.21 V vs. Ag/AgCl). The authors suggested that glucose oxidase had sufficient electron transport to the electrode to influence the open circuit potential (the potential of the enzyme was reported to be −0.340 V vs. Ag/AgCl). The system displayed minimal activation polarization, suggesting that this electron transfer continues to be reasonably facile. Similar work was published by Xu and coworkers, using ABTS²⁻ as a mediator with laccase at the cathode and glucose oxidase with ferrocene carboxylic acid as the anodic mediators.¹⁵¹

3.2.7. Quinone/Hydroquinone mediators

Wang, Kim, and coworkers reported a biofuel cell using glucose oxidase attached to carbon nanotubes with 1,4-benzoquinone as a soluble mediator.¹⁹⁵ Polarization tests using a conventional Pt/C O₂ cathode indicated minimal cell polarization up to current densities approaching 0.1 mA/cm², until mass-transport limitations led to a significant decrease in performance (Figure 52). A maximum power density of 77 μW/cm² was observed. The enzyme activity was suggested to be capable to supporting power densities approaching 270 μW/cm².

Figure 52.

Polarization curve of glucose/O₂ biofuel cells using benzoquinone as an anodic mediator with a CNT-GO_x bioanode. Adapted from reference 195. Copyright 2009 Wiley Periodicals, Inc.

Zhu and Zhang reported the use of a water-soluble anthraquinone derivative, 2,7-anthraquinone-disulfonic acid (AQDS), as a mediator for oxidation of glucose 6-phosphate using glucose 6-phosphate dehydrogenase (G6PDH) and diaphorase as enzymatic catalysts.¹⁹⁶ NADH was used to mediate electron transfer between glucose 6-phosphate dehydrogenase and diaphorase. The authors compared the use of soluble mediators and enzymes with immobilized mediators and enzymes (Figure 53A), including (a) immobilized enzymes and an immobilized naphthoquinone mediator, menadione (VK3), (b) soluble enzymes with immobilized VK3, and (c and d) soluble enzymes with either soluble benzyl viologen or soluble AQDS. The results with AQDS as a soluble mediator and non-immobilized enzymes gave the best performance, as determined from polarization curves, with a maximum power density of 1.1 mW/cm² at room temperature. This value could be increased to 2.4 mW/cm² at 37 °C. Long-term performance tests showed that the cell with AQDS and immobilized mediators continued to display optimal activity, though overall the performance decreased by 65%, while a cell with immobilized mediator and enzymes only decreased by 35% (Figure 53B).

Figure 53.

A) Schematic diagrams of enzymatic fuel cell anodes with glucose 6-phosphate as the fuel. The four formulations included (a) immobilized enzymes with immobilized vitamin K3 mediator, (b) soluble enzymes with immobilized vitamin K3 mediator, and (c and d) soluble enzymes with free benzyl viologen (BV) or free AQDS mediators. G6P = glucose 6-phosphate, G6PDH = glucose 6-phosphate dehydrogenase, DI = diaphorase. B) Current vs. time plots for the four O₂/glucose 6-phosphate fuel cell strategies. Cell was discharged through a 150-ohm resistor. Fresh G6P substrate was added every 3 days. The electrolyte contained 5 mM mediator, 20mM G6P, 8mM NAD⁺, 0.1 M pH 7.3 HEPES buffer, 10mM MgCl₂, 0.5mM MnCl₂, 0.1 M NaCl, 2 U/mL G6PDH, and 2.7 U/mL diaphorase. Adapted from reference 196.

Cosnier and coworkers encapsulated 9,10-phenanthrenequinone in a glyconanoparticle (PS-bCD-PQ) and used fungal FAD-dependent glucose dehydrogenase for the oxidation of glucose in a “solubilized enzymatic fuel cell”.¹⁵⁴ This encapsulation increases the concentration of the hydrophobic quinone mediator from ~ 5 μM to 68.4 μM (for a 2:1 quinone/glyconanoparticle ratio). Discussion of the performance of the assembled fuel cell is provided in Section 2.7.

3.2.8. Nitroxyl/Oxoammonium mediators

Minteer, Sigman, and coworkers use a nitroxyl-based mediator in combination with an enzyme for the complete oxidation of glycerol to CO₂. The proposed pathway features non-enzymatic nitroxyl-mediated oxidation of glycerol to mesoxalic acid, followed by decarboxylation of mesoxalic acid by oxalate oxidase.¹⁹⁷ Further nitroxyl-mediated (non-enzymatic) oxidation of glyoxylic acid forms oxalic acid, which undergoes enzymatic oxidation to CO₂. Most nitroxyl mediators were found to be inactive as oxidation catalysts under the acidic conditions required for oxalate oxidase (i.e., pH 4–5), but 4-amino-TEMPO exhibited suitable activity. Bulk electrolysis studies with glycerol, 4-amino-TEMPO and oxalate oxidase at pH 5.2 generated current densities up to 1.2 mA/cm². Intermediates such as glyceric acid, tartronic acid, mesoxalic acid, and glyoxylic acid were detected by HPLC. Using ¹³C-labeled glycerol, ¹³CO₂ was also detected by NMR. This mediated electrochemical system is different from the other enzymatic processes discussed in this article in that the organic nitroxyl mediator does not shuttle electrons between the enzyme and the electrode, but rather directly mediates electrochemical oxidation of the fuel (glycerol). The enzyme processes intermediates derived from glycerol oxidation that are not amenable to nitroxyl-mediated oxidation, and thereby enables complete oxidation of glycerol to CO₂.

Oxalate oxidase requires O₂ to support decarboxylative production of CO₂ from mesoxalic acid and oxalic acid. Consequently, the four electrons transferred in these steps are not captured electrochemically. To address this issue, 4-amino-TEMPO was paired with other enzymes capable of promoting non-oxidative decarboxylation. Such performance was achieved with oxalate decarboxylase, a redox-neutral enzyme that enabled collection of all possible electrons in the electrochemical oxidation of glycerol (Figure 54).¹⁹⁸ Another study employed the NAD-dependent formate hydrogenase with 4-amino-TEMPO to oxidize glycerol to CO₂.¹⁹⁹ Unlike oxalate decarboxylase and oxalate oxidase, which require acidic pH (e.g., pH 4–5), formate hydrogenase is effective at higher pH (e.g., pH 8) where the 4-aminoTEMPO exhibits higher activity. The catalytic activity of the combined system, with NAD⁺, formate hydrogenase, and 4-amino-TEMPO, for the oxidation of mesoxalate was evident by cyclic voltammetry. Removal of any of the components led to significant or complete loss of catalytic activity. Complete oxidation of glycerol to CO₂ from bulk electrolysis was confirmed by NMR.

Figure 54.

Electro-oxidation cascade pathway for the complete oxidation of glycerol to CO₂ catalyzed by TEMPO-NH₂ and oxalate decarboxylase. The red arrows indicate the oxidations by TEMPO-NH₂. Adapted from reference 198. Copyright 2015 The Royal Society of Chemistry.

One problem in using nitroxyl species as mediators for fuel oxidation at the anode of a fuel cell is the high redox potential of these species, which limits the cell voltage that can be achieved in a fuel cell. In an effort to identify nitroxyls with a more suitable redox potential while retaining high rates of electrocatalytic alcohol oxidation, Sigman, Minteer, and coworkers developed a structure-activity relationship for nitroxyl catalysts.²⁰⁰ The catalytic activity of the nitroxyl radical was found to be empirically correlated with the potential difference between proton-coupled oxidation of the (protonated) hydroxylamine and oxidation of the nitroxyl. By analyzing this relationship, the authors were able to assign different nitroxyl species to one of four groups, associated with their catalytic activity (y-axis) and their nitroxyl/oxoammonium redox potential (x-axis) (Figure 55A). The four groups correspond to (a) good anodic catalysts (low redox potential and high catalytic activity), (b) good synthetic catalysts (high potentials and high catalytic activity), (c) anti-oxidants (low catalytic activity and low potential), and (d) EPR labels (high potential and low catalytic activity). From this analysis, two species were highlighted as optimal targets for fuel cell applications (numbered 14 and 19 in the original study) (Figure 55B).

Figure 55.

A) Plot of catalytic activity (i_pa/i_pc)cat vs. E_a1 (nitroxyl/oxoammonium redox potential) separated into four quadrants based on catalyst characteristics. Adapted from reference 200. Copyright 2015, American Chemical Society. B) Structure of two optimal nitroxyl derivatives in the “Anodic Catalyst” quadrant.

4. Conclusions and Outlook

This content above highlights historical and recent advances in the applications of soluble redox mediators to electrochemical power generation. The development and testing of mediated fuel cells (MedFCs) have greatly expanded in recent years, with advances leading to new approaches to the catalytic reduction of O₂ at the cathode and to the oxidation of H₂, alcohols, biomass, and other fuels at the anode. Replacement of Pt with non-PGM cathodes in fuel cell cathodes remains a high priority. Whereas historical efforts commonly featured Br₂- or NO_x-based mediators, these materials seem unlikely to provide long-term solutions due to challenges related to the volatility, toxicity and/or corrosivity of these species. Polyoxometalates continue to be the focus of attention because their excellent stability and good reactivity with O₂, and their ability to access reasonable overpotentials makes them promising mediators and/or catalysts for the ORR and mediated cathode systems. POM mediators have some disadvantages related to their high molecular weight, which contributes to a decreased specific energy density. The recent demonstration of organic cathode mediators in combination with off-electrode ORR catalysts provides a novel approach to implement Pt-free ORR catalysts (e.g., Fe- or Co-N/C materials) and may allow for improved utilization of these catalytic materials. The design of highly stable, high-potential mediators represents a key challenge that will need to be addressed to expand on recent demonstrations. In addition, exploration of other conditions, such as the use alkaline conditions and anion exchange membranes, may open new avenues for catalyst and mediator compositions. Mediated electrochemistry introduces significant opportunities for fuel cell power generation with complex fuels, such as polyols and other biomass-derived materials, hydrocarbons, and even low-grade coal and sewage. Low-potential POMs have emerged as some of the most effective anodic mediators, reflecting their low cost, good stability, effective redox behavior, and compatibility with chemicals present in complex fuels that inhibit or poison many other types of catalysts (e.g., sulfur or other heteroatom-containing compounds). Thus far, many implementations of mediated fuels with complex chemicals employ a two-stage process in which the first stage features oxidation of the fuel by the mediator in a (non-electrochemical) batch process, and the mediator is regenerated electrochemically in a second stage. Assessment of the merits of this approach relative to a one-stage continuous electrolysis process represents an important issue for future efforts in the field. Organic mediators have been commonly used to support fuel oxidation, especially in enzymatic fuel cells. Organic mediators are appealing because of the ability to tune their redox kinetics and electrochemical potentials, as revealed by studies of viologens and organic nitroxyls. For enzymatic fuel cell systems, efforts will likely continue to focus on immobilized enzymes and mediators, but recent reports highlight potential performance enhancements possible when the mediator and/or enzyme are dissolved in solution. The latter strategy may find utility for systems in which refreshing the solutions to restore the enzyme and mediator activity is feasible.

Thus far, mediated fuel cells, especially those using alcohols, biomass, or other complex fuels achieve lower power output than conventional fuel cells. Nonetheless, the ability to eliminate the requirement for Pt and other PGM catalysts and derive power from non-conventional fuel sources provides strong motivation to continue exploration of this technology, including the development of new mediators that achieve improved stability, lower overpotentials for the processes of interest, better kinetics; and the development of new catalysts that serve as effective partners with mediators to allow electrochemically driven off-electrode redox reactions or that exhibit cooperative redox behavior with the mediator, allowing for faster rates or lower overpotentials.

Biographies

Colin W. Anson obtained his B.A. in Chemistry in 2011 from Colby College, located in Waterville, Maine. Later that year, he joined the research group of Professor Shannon Stahl at the University of Wisconsin-Madison, where he obtained his Ph.D. in Chemistry in 2017. His doctoral work focused on the use of organic quinones as mediators for electrocatalytic O₂ reduction using molecular cobalt catalysts. He continued this research during his postdoctoral work, targeting fuel cell-relevant applications of this concept. Currently, he is an Associate Scientist in the Stahl group, where he works on electrochemical energy conversion and biomass valorization projects.

Shannon S. Stahl obtained his B.S. in Chemistry in 1992 from the University of Illinois at Urbana–Champaign and earned his Ph.D from Caltech in 1997, where he worked with Professor John Bercaw. He was an NSF postdoctoral fellow with Professor Stephen Lippard at Massachusetts Institute of Technology from 1997 to 1999. He is currently the Steenbock Professor of Chemical Sciences at the University of Wisconsin–Madison, where he began his independent career in 1999. His research group specializes in catalysis and electrochemistry, with an emphasis on catalytic aerobic oxidation reactions and electrocatalytic reactions related to chemical synthesis and energy conversion.

Appendix I. Summary of Mediated Fuel Cells Using Chemical Catalysts. (See Table 8 for a Summary of Mediated Fuel Cells Using Enzymatic Catalysts.)

Anode			Cathode
fuel	catalyst	mediator	mediator	catalyst	oxidant	ref.
H2	Pd black	TiO2+/Ti3+	Br−/Br2	NOx	O2	40
	Pt	–				73
			VO2+/VO2+	HCl		75
		Mo3+/Mo4+		NOx		79
		H4SiW12O40				79
		Ti3+/TiO2+				80
	Pd/Al2O3	Fe(EDTA)2−/Fe(EDTA)−				80
	Pt	H4SiW12O40		POM		76–78
	Pt/C	AQ	V-containing POM			105
	Pt/C	–				63, 95–102, 104
			Fe2+/Fe3+	Fe(pc)		61
			K3Fe(CN)6	Mn complex		118
			Fc species	Fe complex		119–123
			HNO3			128–129
			nitroxyl	–		137
			quinone	Co–N/C		143
alcohols	Pt	–	HNO3			124–127
coal	–	Sn4+/Sn2+	Br−/Br2	NOx		70
CH4, alcohols, coal	Pt	Fe2+/Fe3+	VO2+/VO2+	NOx		82–84
biomass	–					85, 86
sugars	Pt	V3+/VO2+		Pt		81, 169
sugars	POM	Fe2+/Fe3+	V-containing POM			110
biomass	POM					107, 111, 112
alcohols						108
lignin						113
coal						109
biomass			–	Pt/C		113, 159
alcohols, sugars or lignin	–	AQ	–	–		166–168
sugars	–	viologen	–	Co		171–175, 177
–	–	V2+/V3+	VO2+/VO2+	NOx		79
Zn	–	–	Fe2+/Fe3+	charcoal		114