Redox‐Polymer‐Based High‐Current‐Density Gas‐Diffusion H2‐Oxidation Bioanode Using [FeFe] Hydrogenase from Desulfovibrio desulfuricans in a Membrane‐free Biofuel Cell

Abstract The incorporation of highly active but also highly sensitive catalysts (e.g. the [FeFe] hydrogenase from Desulfovibrio desulfuricans) in biofuel cells is still one of the major challenges in sustainable energy conversion. We report the fabrication of a dual‐gas diffusion electrode H2/O2 biofuel cell equipped with a [FeFe] hydrogenase/redox polymer‐based high‐current‐density H2‐oxidation bioanode. The bioanodes show benchmark current densities of around 14 mA cm−2 and the corresponding fuel cell tests exhibit a benchmark for a hydrogenase/redox polymer‐based biofuel cell with outstanding power densities of 5.4 mW cm−2 at 0.7 V cell voltage. Furthermore, the highly sensitive [FeFe] hydrogenase is protected against oxygen damage by the redox polymer and can function under 5 % O2.

, porosity ≈31μm; note that one side of the carbon cloth is coated with a Nafion/Teflonbased microporous film (50 µm), carbon content 5 mg cm -2 , EQ-bcgdl-1400S-LD, this side was denoted as the microporous side/layer, the non-modified side was denoted as macroporous side/layer) were used. As the counter electrode in the three-electrode setup a Pt-wire with a diameter of 1 mm was used. As reference electrodes, a commercial (Methrom) Ag/AgCl/3 M KCl system was used. All measured potentials are rescaled vs. SHE according to ESHE = EAg/AgCl/3 M KCl + 210 mV. All measurements were performed with a Reference 600 (Gamry Instruments) or an Interface 1000 (Gamry Instruments) potentiostat. Scan rates for cyclic voltammograms and applied potentials for chronoamperograms as well as the used gas mixtures are noted in the corresponding figure captions. Phosphate buffer (PB, 0.1 M, pH 7.4) was used as the working electrolyte for all experiments. Mass flow controllers were used to control the composition of the gas feed. For biofuel cell characterization, steady state currents from multi-step chronoamperometry were used to calculate power values for the corresponding power curves to minimize the contribution from capacitive currents. In gas-diffusion mode a backpressure of the corresponding gas was applied to the gas diffusion layer, while the electrolyte was continuously flushed with Ar (note that because of pressure equilibration, the cell was not fully closed).

Electrode preparation
Bioanode. For preparation of the bioanode, the macroporous side of the carbon cloth electrodes were used to prevent direct electron transfer. In a first step, 20 µL of an aqueous P(GMA-BA-PEGMA)-vio solution (7.5 mg mL -1 ) was drop cast onto the electrode over an area of approximately 4 mm as an adhesion layer and dried for 12 h at room temperature. In a second step, the electrodes were transferred to an O2 free Biocathode. For preparation of the biocathode, the microporous side of the carbon cloth electrodes was used (geometrical surface area: 3.5 cm 2 , note that the actual surface area is much higher due to the porous nature of the electrode material) to ensure a high catalyst loading and productive wiring in a direct electron transfer regime.
The high surface area of the biocathode compared to the bioanode ensures anode limiting conditions. First, the electrode was washed with ethanol and water.
Subsequently, the surface of the electrode was modified with 2-aminobenzoic acid in an electrochemical grafting process in an aqueous solution of 0.1 M KCl and 5 mM 2aminobenzoic acid. A potential of 800 mV vs. Ag/AgCl 3 M KCl was applied for 60 s as described in ref. [4] . After the grafting process, the electrode was washed with water and 80 µl of the Mv-BOx solution (15 mg mL -1 in 0.1 M PB pH 7.4) were drop cast on the chemically modified electrode. The electrode was dried for 1.5 h at room temperature (enzyme loading: 1200 µg electrode -1 /343 µg cm -2 ).

Additional electrochemical notes, experiments, data and controls
Note S1: The OCV of a fuel cell is the highest voltage reached in the absence of any current flow. For an ideal case, the OCV will be given by the difference between the formal potentials of the fuel oxidized at the anode and the oxidant reduced at the cathode. In practice, however, each of these reactions proceed with a certain overpotential as a function of the employed (bio)catalyst and the experimental thermodynamic parameters. Therefore, the OCV reached by an enzymatic biofuel cell is also influenced by the overpotential required to drive the enzyme-catalysed reactions. Moreover, when a separate redox mediator species is used to shuttle electrons between the enzyme and electrode, associated losses in the OCV occur and the potential will now be determined by the difference between the formal potentials of the redox mediators involved in the electron transfer, in addition to any inherent overpotentials.
The use of redox mediators immobilized on the electrode surface (e.g. in the case of redox polymers) provides a pseudocapacitive component to the electrode assembly capable of storing charges in the form of oxidized/reduced redox centres due to enzymatic conversion in the presence of substrate. The established gradient in activity ratio implies a shift in the redox potential of the bioelectrode due to a self-charging process occurring at open circuit. In consequence, the experimentally recorded OCV (the voltage difference between the bioelectrodes) is effectively increased with respect to the apparent thermodynamic value defined by the midpoint potential of the polymer-bound redox mediator. For a more comprehensive discussion of this effect see ref. [5] Table S1: Enzyme activities (H 2 oxidation), current densities of polymer/hydrogenase gas diffusion bioanodes and their power densities when incorporated into H 2 /O 2 biofuel cells. All hydrogenases were immobilized on and wired to carbon cloth-based gas diffusion electrodes by means of the polymer double layer system P(GMA-BA-PEGMA)-vio//P(N 3 MA-BA-GMA)vio/hydrogenase. Ref.
The biocathode showed absolute O 2 reduction currents of almost 2 mA ( Figure S1, red curves), which is similar to the H2 oxidation currents observed for the best performing P(GMA-BA-PEGMA)-vio//P(N 3 MA-BA-GMA)-vio/DdHydAB electrodes (cf. Figure 2a). Hence, to ensure anode limiting conditions, a bioanode with H2 oxidation currents slightly below the absolute currents of the biocathode was employed ( Figure S2a, note that the bioanodes show considerable scatter in their absolute current values). The bioanode not only shows a decreased H 2 oxidation activity after the long-term experiment but also less pronounced non-turnover redox waves under 100% Ar, suggesting loss of parts of the polymer/enzyme film during the measurement ( Figure S2b). The loss of the polymer layer is most likely due to the harsh operating conditions during the long-term stability evaluation (local pH shift due to proton generation from H 2 oxidation and continuous reduction of incoming O 2 to H 2 O 2 at the redox polymer matrix, which can induce polymer degradation [9] ).