Anthraquinone-Mediated Fuel Cell Anode with an Off-Electrode Heterogeneous Catalyst Accessing High Power Density when Paired with a Mediated Cathode

Abstract

The development of processes for electrochemical energy conversion and chemical production could benefit from new strategies to interface chemical redox reactions with electrodes. Here, we employ a diffusible low-potential organic redox mediator, 9,10-anthraquinone-2,7-disulfonic acid (AQDS), to promote efficient electrochemical oxidation of H₂ at an off-electrode heterogeneous catalyst. This unique approach to integrate chemical and electrochemical redox processes accesses power densities up to 228 mW/cm² (528 mW/cm² with iR-correction). These values are significantly higher than those observed in previous mediated electrochemical H₂ oxidation methods, including those using enzymes or inorganic mediators. The approach described herein shows how traditional catalytic chemistry can be coupled to electrochemical devices.

Graphical Abstract

Redox-active organic molecules have been widely employed as electrochemical mediators to enhance electron transfer between an electrode and a catalyst or chemical reagent in solution. Applications include electrochemical organic synthesis^1–3 and energy storage and conversion.^4–7 Enzymatic biofuel cells, for example, often use organic redox mediators to facilitate redox reactions with enzyme-based catalysts. Representative mediators include (anthra)quinones,^8–10 viologens,^11–13 and 2,2’-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS)¹⁴ (Figure 1A). Applications of these methods are typically limited to low-power applications, such as implantable biomedical devices and portable electronics,^15–17 because the conditions required to sustain enzyme activity and stability limit the fuel cell performance. Power densities accessed by enzymatic fuel cells are typically 2–3 orders of magnitude lower than those of conventional proton exchange membrane fuel cells (PEMFCs) (Figure 1B). Replacing the relatively fragile enzymes with heterogeneous catalysts capable of tolerating strongly acidic (or basic) fuel cell conditions and elevated temperatures provides a potential strategy to overcome this limitation. Here, we demonstrate this concept by employing 9,10-anthraquinone-2,7-disulfonic acid (AQDS) as a mediator for electrochemical oxidation of H₂ at an off-electrode Pt/C catalyst. iR-free power densities of > 500 mW/cm² are accessible when this anodic process is paired with a mediated cathode, employing a previously reported polyoxometalate-based mediator for O₂ reduction.¹⁸ This system, which represents the first pairing of an organic mediator with an off-electrode heterogeneous catalyst in a fuel cell anode, leverages the strengths of both chemical and electrochemical redox reactions and establishes an important platform for development of new electrode-driven redox processes.

Figure 1.

Organic redox mediators can play an important role in enzymatic fuel cells (A) and fuel cells with diverse catalysts (B).

Organic mediators have been the focus of substantial recent attention for use in aqueous redox flow batteries (RFBs) and related energy-storage devices.^19–25 Quinones, aminoxyls, and other redox-active organic molecules complement the more widely used inorganic mediators, such as V, Fe, and Zn ions, in these devices.⁶ Aqueous RFBs with organic mediators have been shown to support high power densities (≥ 1 W/cm²),²⁶ and anthraquinones are commonly featured anodic mediators in these applications.^27–30 Meanwhile, quinones have been widely used as reagents and catalytic mediators in chemical oxidation reactions,^31–32 and they undergo efficient catalytic oxidation^33–35 and reduction^36–38 over heterogeneous catalysts. A noteworthy example of the latter application is the “anthraquinone process” for industrial production of hydrogen peroxide, which features a two-stage sequence involving catalytic hydrogenation of an anthraquinone derivative followed by autoxidation of the corresponding anthrahydroquinone to generate H₂O₂ and the quinone derivative.^39–40 Collectively, these precedents provided an important foundation for the present study, which takes advantage of anthraquinones and heterogeneous catalysts to mediate the electrochemical oxidation of H₂ in an off-electrode compartment. AQDS was selected as a prototypical anode mediator owing to its good electrochemical behavior and the proximity of its redox potential to the thermodynamic H⁺/H₂ potential. Both of these features are evident in the cyclic voltammogram of the disodium salt of AQDS (AQDS-Na₂), obtained in the 1 M sulfuric acid medium used in this study. The well-defined oxidation and reduction peaks and quasireversible redox potential of 220 mV vs. NHE (Figure 2) are consistent with previous reports.²⁷

Figure 2.

Cyclic voltammogram of 9,10-anthraquinone-2,7-disulfonic acid, disodium salt (AQDS-Na₂) in 1 M H₂SO₄. Conditions: AQDS-Na₂ (10 mM), glassy carbon working electrode, Pt-wire counter electrode, Ag/AgCl reference electrode; 10 mV s⁻¹ scan rate.

The catalytic hydrogenation of anthraquinones has been extensively investigated in organic solvents or neutral aqueous media,^36–38 but such conditions are not conducive to fuel cell applications, which favor strongly acidic or basic aqueous conditions. Hydrogenation of quinones in acidic media has only been reported for the parent 1,4-benzoquinone, which has a potential (E_1/2 ~ 700 mV vs NHE) more than 450 mV higher than that of AQDS.³⁶ Nonetheless, this precedent employing a supported Pt catalyst prompted us to select a commercial Pt/C catalyst, sourced from Strem Chemicals, for our initial testing of AQDS-Na₂ hydrogenation in 1 M sulfuric acid. Monitoring of the hydrogenation of AQDS is complicated by the air sensitivity of AQDSH₂ (9,10-anthrahydroquinone-2,7-disulfonic acid), arising from its facile autoxidation to regenerate AQDS. To address this issue, the reaction progress was monitored in a sealed vessel via potentiometric analysis of the batch reaction mixture (Figure 3). The measured potential correlated with the relative concentrations of AQDS and AQDSH₂, as expected (see Supporting Information for details), and selective formation of AQDSH₂ was confirmed by UV-visible and ¹H NMR spectroscopy. The UV-visible spectrum of a sample of hydrogenated AQDS matched the spectrum of electrochemically generated AQDSH₂ (see Figures S3 and S4). Two distinct sets of peaks were apparent in the ¹H NMR spectrum of a hydrogenated sample that was maintained under N₂, and these peaks disappeared upon exposure of the sample to air leaving only signals corresponding to AQDS (see Figure S5).^41–43

Figure 3.

Reaction progress for batch hydrogenation of 9,10-anthraquinone-2,7-disulfonic acid, disodium salt catalyzed by Pt/C. Formation of hydroquinone was monitored via potentiometry. Dots represent reaction data, and the line reflects a linear correlation to the data.

With the successful demonstration of catalytic hydrogenation of AQDS under acidic conditions in batch, we then incorporated the Pt/C catalyst into a packed bed reactor (1/2” o.d. stainless steel tubing) to enable the continuous operation of an AQDS-mediated anode. The packed bed reactor was connected to the anode compartment of a 5 cm² fuel cell with serpentine flow plates. The membrane electrode assembly (MEA) featured a carbon cloth anode (Avcarb 1071HCB), a Nafion 115 membrane, and a conventional O₂ cathode (0.2 mg Pt/cm² on carbon). A solution of AQDS in 1 M H₂SO₄ was mixed with H₂ at a tee junction and fed into the packed bed reactor to charge the flow anode with reduced anthraquinone. The fuel cell was then operated at 60 °C with the liquid mediator solution cycled through the anode and reactor compartments, while O₂ was delivered to the cathode at 0.2 L/min. The preliminary results with this fuel cell arrangement, however, revealed significant complications associated with cathode flooding, arising from water migration through the membrane to the cathode. A power curve obtained during these tests showed significant mass-transport losses (see Figure S7 and Supporting Information for discussion). This phenomenon is not uncommon in liquid fuel cells and has been noted, for example, with other mediated anode⁴⁴ and direct methanol fuel cell systems.⁴⁵ To mitigate the impact of water crossover, we elected to replace the gas diffusion cathode with a mediated cathode system consisting of 0.3 M aqueous Na₄H₃PMo₈V₄O₄₀, a polyoxometalate (POM) similar to that reported previously.^18,46,47 Briefly, the POM solution was heated to 80 °C in a 75 mL vessel sparged with O₂, and the oxidized POM solution was flowed to the cathode compartment, where it underwent reduction and then circulated back to the regeneration compartment (see Section 4 of the Supporting Information for details). A schematic illustrating the resulting doubly mediated fuel cell is depicted in Figure 4.

Figure 4.

Schematic of the doubly-mediated fuel cell.

A polarization curve for the AQDS/POM-mediated fuel cell was then collected using 1 M AQDS in 1 M H₂SO₄ at the anode at a state-of-charge of 85% (i.e., AQDSH₂:AQDS ratio of 85:15; Figure 5A). An open-circuit potential of 0.73 V was observed, with a peak power density of 228 mW/cm² and an iR-corrected value of 528 mW/cm². Continuous fuel cell operation was established by matching the rate of the chemical reduction of AQDS in the packed bed reactor and the electrochemical oxidation of AQDSH₂ at the anode (Figure 5B). Incorporation of a liquid reservoir between the anode and the packed bed reactor permitted in-line potentiometric monitoring of the mediator solution as well as the use of different liquid flow rates for improved performance in the individual units: 40 mL/min through the anode for improved mass transfer and 4 mL/min through the reactor to minimize pressure drop through the powder catalyst bed (for a schematic and photographs of the reactor set-up, see Figure S6). The integrated system sustained a current density of 50 mA/cm² at a mediator state-of-charge of approximately 70% for 8 hours.⁴⁸ The contribution of the reactor to sustained system performance was confirmed in a separate experiment by interrupting the hydrogen flow over a period of one hour. The cell potential decreased steadily from 585 mV to 550 mV as the AQDS mediator state-of-charge decreased from 55% to 42%. Once hydrogen flow was reinitiated, reduction of AQDS ensued and the cell potential stabilized (see Figure S8 and Supporting Information for further details). These results demonstrate the steady-state integration of an electrochemical redox process (oxidation of AQDSH₂) with an off-electrode chemical redox reaction.

Figure 5.

(A) Polarization and power density curves with 1 M anthraquinone disulfonic acid in 1 M H₂SO₄ at the anode and 0.3 M Na₄H₃PMo₈V₄O₄₀ at the cathode. Pale curves are raw data and bright curves reflect data with iR-compensation. The fuel cell was operated at 60 °C. (B) Steady-state cell potentials observed from constant current operation of the doubly-mediated fuel cell at 50 mA/cm².

The power densities and currents sustained in these experiments with AQDS are the highest reported to date for mediated electrolysis with hydrogen (Figure 6; experimental details of the mediated fuel cells, as well as power density calculations, are provided in Table S1 of the Supporting Information). Early precedents for mediated fuel cells of this type employed hydrated metal ions as mediators, including Ti^3/4+,⁴⁹ Sn^2/4+,⁵⁰ Fe^2/3+,⁵⁰ Cu^0/2+,⁵⁰ and Mo^3/4+,⁵¹ in addition to a Si-containing polyoxotungstate⁵¹ and EDTA-ligated Fe^2/3+.⁵² Each of these examples used off-electrode Pt or Pd catalysts for reduction of the mediator with H₂, although several examples only reported polarization curves, without demonstration of continuous, steady-state fuel cell operation.⁵⁰ The silicotungstic acid fuel cell accessed the highest power density at 143 mW/cm². More recently, organic mediators have been used in low power H₂ fuel cells to facilitate electron transfer with enzymatic catalysts. In the earliest example, 1.5 mM methyl viologen was paired with the bacterial cell D. vulgaris for a fuel cell with a power density of 0.19 mW/cm².¹² Replacing the methyl viologen with 1.5 mM anthraquinone-2-sulfonic acid (AQS) increased the value to 0.44 mW/cm². Incorporation of viologen mediators into redox polymer networks with NiFe⁵³ and FeFe⁵⁴ hydrogenases enabled power densities of approximately 0.2 mW/cm². The low power densities in these systems partly reflect the mild conditions needed to ensure enzyme stability and activity. For example, the pH typically varies from 4–11 and mediator concentrations are typically under 20 mM and operating temperatures under 40 °C. The ability to use a 1 M mediator concentration, operate the fuel cell at 60 °C, and employ a robust heterogeneous catalyst at pH 0 all contribute to the improved fuel cell performance observed in the present study.

Figure 6.

Comparison of power densities in fuel cells with anodic mediators using H₂ as the fuel. Power density was calculated from the highest performance reported in each reference. References, full details of the experimental configurations of the mediated fuel cells, and power density calculations are provided in Table S1 of the Supporting Information.

The results described herein, which demonstrate the ability to produce power in a fuel cell by coupling a conventional catalytic hydrogenation reaction with electrochemical oxidation of an organic mediator, set the stage for a number of future research directions. Improved fuel cell performance should be accessible by using lower potential mediators that operate closer to the thermodynamic potential of the fuel. More broadly, this approach can leverage the widespread use of heterogeneous catalysts in organic chemical transformations. The development of new catalytic transfer (de)hydrogenation reactions between organic molecules and electrochemical mediators represents an important extension of the concepts reported here. For example, preliminary batch experiments show that formic acid and methanol may be used instead of H₂ to reduce AQDS to AQDSH₂ in the presence of a Pt/C catalyst (Figure S9). These results provide a starting point for development of mediated fuel cells that use more complex fuels than H₂ and incorporate organic mediators, in contrast to previously reported systems which utilize inorganic mediators, such as polyoxometalates.^55–57 Such systems would have several appealing features, including ability to minimize fuel crossover and avoid gas generation at the electrode surface by moving the catalytic process off the electrode surface.^45,58 Efforts to explore such mediated anode systems have been initiated in our lab.