Designing one-compartment H₂O₂ fuel cell using electroactive phenalenyl-based [Fe₂(hnmh-PLY)₃] complex as the cathode material

Abstract

The sustainable chemical energy of H₂O₂ as a fuel and an oxidant in an advantageous single-compartment fuel cell design can be converted into electric energy, which requires molecular engineering to design suitable cathodes for lowering the high overpotential associated with H₂O₂ reduction. The present work covers the synthesis and structural characterization of a novel cathode material, [Fe^III₂(hnmh-PLY)₃] complex, 1, designed from a PLY-derived Schiff base ligand (E)-9-(2-((2-hydroxynaphthalen-1-yl)methylene)hydrazineyl)-1H-phenalen-1-one, hnmh-PLYH₂. Complex 1, when coated on the surface of a glassy carbon electrode (GC-1) significantly catalyzed the reduction of H₂O₂ in an acidic medium. Therefore, a complex 1 modified glassy carbon electrode was employed in a one-compartment H₂O₂ fuel cell operated in 0.1 M HCl with Ni foam as the corresponding anode to produce a high open circuit potential (OCP) of 0.65 V and a peak power density (PPD) of 2.84 mW cm⁻². CV studies of complex 1 revealed the crucial participation of two Fe(III) centers for initiating H₂O₂ reduction, and the role of coordinated redox-active PLY units is also highlighted. In the solid state, the π-conjugated network of coordinating (hnmh-PLY) ligands in complex 1 has manifested interesting face-to-face π–π stacking interactions, which have helped the reduction of the complex and facilitated the overall catalytic performance.

---

Introduction

On account of the high energy density, safety in storage and handling and high expected theoretical output voltage (1.09 V) of H₂O₂ (hydrogen peroxide), one-compartment H₂O₂ fuel cells have attracted significant attention for more than a decade.^1–3 The successful operation of an H₂O₂ fuel cell through one compartment, where H₂O₂ acts as fuel and oxidant makes it favorable over other fuel cells.^4–7 The high portability of liquid H₂O₂ can be considered another advantage for meeting energy requirements in next-generation portable devices.^6,8 The electrochemical reaction at the anode and cathode in the H₂O₂ fuel cell and net reaction that produces electricity are given in Chart 1.⁸ The operation of H₂O₂ fuel cells under acidic conditions can yield good and stable OCP (open circuit potential) values.^1,7,9,10 In natural systems, the active sites of hydroperoxide enzymes contain iron porphyrin complexes which efficiently catalyze the reduction of H₂O₂.¹¹ Further, from recent reports in the literature, it was evident that the Fe^III/Fe^II redox couple controls the thermodynamics of the electrochemical reaction taking place at the cathode in an H₂O₂ fuel cell by subsequently decreasing the overpotential associated with the electro-catalytic reduction of H₂O₂.^4,5,12–14

Chart 1 Electrochemical reactions at cathode and anode for a one-compartment H₂O₂ fuel cell operated under acidic conditions.^12,13

Yamada et al. investigated the first H₂O₂ fuel cell operated under acidic conditions with phthalocyanine and porphyrin-derived Fe(III) complexes as the cathode with a peak power density (PPD) of 0.01 mW cm⁻².¹¹ Later, Mousavi Shaegh et al. used Prussian blue as the cathode in an H₂O₂ fuel cell operated in 1.0 M HCl to produce a PPD of 1.55 mW cm⁻².¹⁵ A series of polynuclear cyanide complexes were also employed by Yamada et al., where the role of N-bound Fe ions as an active species for H₂O₂ reduction in these complexes was further emphasized.¹⁶ In the pursuit of obtaining good power output from H₂O₂ fuel cells, cathodes containing noble metals (Pt and Pd) were introduced into pyrazine-bridged Fe[M^C(CN)₄] (M^C = Pt²⁺ and Pd²⁺) complexes to achieve a PPD of 4.2 mW cm⁻².¹⁷ However, while dealing with the sustainable science of extracting energy from membrane-less H₂O₂ fuel cells, the recognition of highly efficient electroactive cathode materials with earth-abundant first-row transition metals requires significant attention.

In this consideration, the coordination chemistry approach of integrating first-row transition metals to ligands that “store and release charge” to form electronically interesting transition metal complexes (TMCs) is known to mimic the efficiency of noble-metal-catalyzed reactions.^18–22 Phenalenyl (PLY), a polycyclic odd alternant hydrocarbon belongs to such class of organic ligands, with the enormous potential to exhibit molecular bistability in response to external stimuli, such as magnetic or electric fields.²³ PLY can exist in three thermodynamically stable redox states: cation, radical, and anion.^24,25 Indeed, the high resonance stabilization energy associated with the highly symmetrical (D₃h) PLY framework contributes to observed unique redox states.²⁶ In addition, the presence of a unique NBMO (nonbonding molecular orbital) and interesting molecular orbital overlap, which facilitate intermolecular charge transfer in the solid state, make PLY a suitable candidate for designing electronic devices^27,28 and cathode materials.^29–31 The cationic form of PLY, also known as closed-shell PLY, can form via metal coordination, which under an applied potential can accept electrons and generate in situ PLY radical species (open-shell).³² Open-shell PLY radical species can act as reducing agents when employed under suitable reaction conditions and participate directly in the reaction mechanism.^26,33,34 Mandal et al. introduced redox-active phenalenyl coordination to the Fe(III) metal center to prepare an electroactive complex Fe(PLY)₃, for H₂O₂ reduction, giving a high open circuit potential (OCP) of 0.72 V and PPD of 1.43 mW cm⁻² in a one-compartment H₂O₂ fuel cell designed at pH 1.²⁹

Our group has reported a PLY-based Co^II complex,²⁷ which was employed as an active material for a RRAM (resistive random access memory) device at very low switching potential. In our recent research work, we reported a novel PLY-functionalization strategy for generating multinuclear PLY-derived metal complexes.³¹ Using our synthetic strategy, we reported the first PLY-based dinuclear Fe(III) complex, [Fe^III₂(hmbh-PLY)₃].³¹ Further, an H₂O₂ fuel cell was designed with an [Fe^III₂(hmbh-PLY)₃] complex as a cathode and Ni foam as an anode in 0.1 M HCl, resulting in a stable OCP of 0.65 V and PPD of 2.41 mW cm⁻². In continuation of our sustained interest in PLY-based metal complexes, herein, we report PLY-derived dinuclear Fe(III) complex 1, [Fe^III₂(hnmh-PLY)₃], designed from the ligand (E)-9-(2-((2-hydroxynaphthalen-1-yl)methylene)hydrazineyl)-1H-phenalen-1-one (hnmh-PLYH₂). By tailoring the substitution on this class of PLY-derived Schiff base ligand and Fe(III) metal coordination, the PPD is improved to 2.84 mW cm⁻², which is higher than the power output reported earlier.

Experimental section

Materials and reagents

Without any additional purification, all chemicals were used as received. All solvents were distilled using protocols in the literature before being employed in all synthetic manipulations.³⁵ 9-Hydrazineyl-1H-phenalen-1-one (Hz-PLY) was prepared as per the reported literature.³¹ A Bruker Apex-II diffractometer was used to collect SXRD data. Crystallographic data of hnmh-PLYH₂ and complex 1 are deposited at the Cambridge Crystallographic Data Center (CCDC 2304746 and 2304747†). A ThermoFisher Scientific flash smart V CHNS/O was used for elemental analysis. Multinuclear NMR data were recorded in CDCl₃ solvent on a Bruker High-Performance Digital FT-NMR 500 MHz. FT-IR spectra were recorded on a Bruker Vertex 70V + PMA50 from 500 to 4000 cm⁻¹. A Cary 4000 UV–vis spectrophotometer was used for recording absorption spectra in dichloromethane (DCM) solvent. The HRMS spectrum was obtained from 6545 LC/Q-TOF HRMS. EPR analysis in the X-band region (8.75–9.65 GHz) was conducted on a JES-FA200 spectrometer. All electrochemical experiments were performed on a CHI6154E instrument, and fuel cell performance tests were conducted on a PINE research 100 EIS potentiostat instrument. Commercially available hydrogen peroxide (30 wt%) was diluted to the required concentration to conduct all the experiments. Millipore ultrapure water (resistivity ∼18.2 MΩ cm at 25 °C) was used to prepare all aqueous solutions.

Computation details

Density functional theory (DFT) computations were performed using the ORCA 5.0.2 program package.³⁶ Geometric optimizations without any constraints were carried out in a vacuum by employing a hybrid functional B3LYP in conjugation with a triple-ζ quality def2-TZVP basis set on iron centers and double ζ quality split-valence basis set, def2-SVP on ligands (C, H, O, and N).³⁷ To improve the computational speed, the RIJCOSX approximation with a combination of an auxiliary basis set (def2/J) was also employed in all spin-unrestricted DFT calculations. Harmonic vibrational analyses at the same level of theory were subsequently undertaken to identify whether the optimized geometry is a local minimum.³⁸ The minimum structures were characterized by all positive eigenvalues of the Hessian matrix. The resultant geometries from optimization were used in subsequent single-point energy calculations at the B3LYP/def2-TZVP + def2-TZVPP level of theory (metal centers and their coordination environment are treated with the def2-TZVPP basis function and the rest with def2-TZVP) along with Grimme's dispersion correction, D3BJ.³⁹ Natural Population Analysis (NPA) was performed at the same level of theory. The solvent effect in redox calculations was considered by using the CPCM method⁴⁰ with DMF. VMD⁴¹ software was used for the visualization of molecular orbitals and spin density.

Synthesis of (E)-9-(2-((2-hydroxynaphthalen-1-yl)methylene)hydrazineyl)-1H-phenalen-1-one, hnmh-PLYH₂. Hz-PLY (1 mmol, 0.21 g) and 2-hydroxy-1-naphthaldehyde (1 mmol, 0.17 g) were simultaneously added to 50 mL of methanol. The reaction mixture was refluxed for 1 h to yield a red precipitate. The precipitate was washed subsequently with cold methanol. Recrystallization from acetonitrile solution resulted in bright red X-ray-quality needle-like crystals. Yield: 0.34 g (92%). ¹H NMR (500 MHz, CDCl₃) δ (ppm): 15.36 (s, 1H), 12.12 (s, 1H), 9.21 (s, 1H), 8.06 (d, J = 9.2 Hz, 2H), 7.92 (d, J = 7.7 Hz, 1H), 7.87 (d, J = 7.7 Hz, 2H), 7.79 (dd, J = 11.6, 9.1 Hz, 2H), 7.74 (d, J = 8.0 Hz, 1H), 7.53 (t, J = 7.7 Hz, 1H), 7.46 (t, J = 7.6 Hz, 1H), 7.35 (t, J = 7.4 Hz, 1H), and 7.00 (d, J = 9.4 Hz, 1H). ¹³C NMR (125 MHz, CDCl₃) δ (ppm): 183.7, 158.6, 151.4, 147.1, 139.2, 138.7, 133.4, 132.4, 131.9, 131.8, 129.2, 128.3, 127.8, 127.8, 127.7, 125.8, 125.7, 123.8, 123.1, 120.1, 118.8, 114.4, 108.6, and 107.6. FT-IR (KBr pellets, ν/cm⁻¹): 3668(m), 3223(m), 1621(s), 1577(s), 1511(s), and 1258(s). HRMS (CH₃CN, positive ionization): calcd for C₂₄H₁₇N₂O₂m/z = 365.1290 [M + H]⁺, found 365.1286 [M + H]⁺.

Synthesis of [Fe^III₂(hnmh-PLY)₃], 1. To a methanolic suspension containing hnmh-PLYH₂ (0.3 mmol, 109.2 mg) and excess triethylamine, anhydrous FeCl₃ (0.2 mmol, 32.4 mg) was added, and the reaction mixture was set up to reflux under an inert atmosphere for 24 h. The black precipitate was isolated and thoroughly washed with methanol. Recrystallization of the as-obtained precipitate from a CHCl₃/CH₃CN (1 : 1) solvent mixture yielded complex 1 in the form of black needles, which were suitable for SCXRD analysis. Yield (based on Fe): 87.46 (73%). Elemental analysis: calcd for C₇₂H₄₂Fe₂N₆O₆: C, 72.13; H, 3.53; and N, 7.01. Found: C, 71.45; H, 3.68; and N, 6.81. FT-IR (KBr pellets, ν/cm⁻¹): 3170(m), 1583(s), 1533(s), 1344(s), 1191(m), and 486(w). UV-vis [DCM, λ/nm (ε in M⁻¹ cm⁻¹)]: 575(18 950), 385(27 272), 322 (51 168), and 292(61 689). HRMS (CH₃CN, positive ionization): calcd for C₇₂H₄₂Fe₂N₆O₆m/z = 1199.1943 [M + H]⁺, found 1199.1984 [M + H]⁺.

Preparation of electrodes for H₂O₂ reduction and fuel cell activity

A glassy carbon electrode (GCE) was polished by a series of aqueous alumina slurries (1, 0.3, and 0.05 μm alumina). Ultrasonication in ultrapure water and ethanol resulted in a shiny working surface of the GCE.⁴² Ni (nickel) foam was treated with dilute HCl solution, followed by ultrasonication in water and ethanol solvent to remove the surface oxide layer. For the preparation of the GC-1 electrode, 10.0 μL of a homogeneous solution of complex 1 (4.0 mg, 3.3 mmol) in DMF and 3.0 μl of Nafion solution (5 wt%) was drop cast on the mirror surface of a GC electrode. GC-Fe and Gc-L were prepared similarly. The sample for the carbon mixed electrode (GC-1C) was prepared by fine grinding 10% (wt./wt.) carbon black with complex 1 in N-methyl pyrrolidine (NMP) solvent using a mortar and pestle. The homogeneously mixed sample was drop cast on the electrode surface of the GC, followed by 3.0 μL of Nafion solution (5 wt%). The one-compartment H₂O₂ fuel tests were run with a complex 1 modified electrode (GC-1 or GC-1C) as cathode and Ni-foam as an anode in an aqueous 0.1 M HCl solution containing 0.3 M H₂O₂ as the fuel. The reproducibility of each experiment was checked at least five times before reporting final measurements.

Results and discussion

Synthetic aspects

The ligand, (E)-9-(2-((2-hydroxynaphthalen-1-yl)methylene)hydrazineyl)-1H-phenalen-1-one, hnmh-PLYH₂ was synthesized using our previously reported 9-hydrazineyl-1H-phenalen-1-one, Hz-PLY precursor moiety.³¹ The addition of 2-hydroxy-1-naphthaldehyde to Hz-PLY in methanol solvent under reflux conditions yielded the ligand hnmh-PLYH₂ in good yield. The formation of a ligand in the solution state was first confirmed with ¹H and ¹³C NMR in CDCl₃ solvent. As can be observed from the ¹H NMR spectrum given in Fig. S1a,† the most deshielded proton (integration = 1) appeared at 15.3 ppm. For phenalenyl systems, the tautomeric proton involved in intramolecular H-bonding interactions at 1- and 9-positions contributes to the chemical shift in this range. Another peak in the ¹H NMR spectrum at 12.1 ppm (integration = 1) can be assigned to the proton of the OH– group of the naphthyl unit. The participation of this –OH in H-bonding interaction with neighboring H-bond acceptors (N atoms) in the solution could be the cause of the deshielding of this proton.⁴³ Hz-PLY-derived Schiff bases contain flexible N–N bonds, where the rotation around these bonds can help the molecule participate in intramolecular H-bonding interactions in solution as well as in the solid state.³¹ The solid-state structure of hnmh-PLYH₂ was confirmed by the SCXRD (single crystal X-ray diffraction) technique. (Details of the crystal structure refinement and lattice parameters are given in Tables S1 in the ESI.†) From the molecular structure of hnmh-PLYH₂ shown in Fig. 1(a), the presence of intramolecular H-bonding interactions in the solid state was confirmed: NH⋯O, 1.877(29) Å; OH⋯N, 1.897(38) Å. The hnmh-PLYH₂ ligand was further characterized by various techniques such as HRMS (high-resolution mass spectrometry), UV-vis (ultra-violet visible), and FT-IR (Fourier transform infrared spectroscopy) [ESI, Fig. S2S4†]. Ligand hnmh-PLYH₂ shows a molecular ion peak corresponding to [M + H]⁺ at m/z 365.1286 amu (Fig. S2a†). The FT-IR spectrum of the ligand recorded in the form of KBr discs shows characteristic peaks in Fig. S3a† for ν_C=O, 1621 cm⁻¹; ν_C=N, 1577 cm⁻¹; and ν_C=C 1511 cm⁻¹.^44,45

Fig. 1 Molecular structure of (a) the hnmh-PLYH₂ ligand and (b) complex 1 as obtained from SCXRD analysis (H atoms are omitted for the sake of clarity.) (c) Coordination environment of Fe(III) centers in complex 1.

For the synthesis of complex 1, stoichiometric quantities (3 : 2) of hnmh-PLYH₂ ligand and anhydrous FeCl₃ were reacted in the presence of NEt₃ base in methanol solvent under reflux conditions for 24 hours (Scheme 1). The primary evidence for the molecular structure of dinuclear complex 1 was obtained using the SCXRD technique. Complex 1 crystallizes in the P1̄ space group. (Details of the crystal structure refinement and lattice parameters are given in Table S1 in the ESI.†) The molecular structure of dinuclear complex 1 shows that both Fe(III) centers adopt a distorted octahedral geometry while being surrounded by N₃O₃ coordinating atoms. The Fe(III) centers in complex 1 afforded two distinct coordination types, as evident from the molecular structure shown in Fig. 1b. Fe1 is coordinated directly to one PLY ring (through 1N and 1O sites) and two naphthyl rings, while the Fe2 center afforded direct coordination to the two PLY rings and one naphthyl ring (through 2N and 2O sites). These two distinguishable Fe(III) centers are spatially separated with a distance of Fe1⋯Fe2, 3.797(7) Å. The coordinated (hnmh-PLY) ligand undergoes twisting around N–N bonds to form the μ-N–N bridge that connects the two distinct Fe(III) centers. Three such μ-N–N bridges are present in complex 1, and their corresponding bond lengths are: μ-N1–N2, 1.424(30); μ-N3–N4, 1.428(40) Å; and μ-N5–N6, 1.413(37) Å. Similar μ-N–N bridges were observed in our previously reported dinuclear complex [Fe^III₂(hmbh-PLY)₃]: μ-N1–N2, 1.428(6) Å; μ-N3–N4, 1.431(6) Å; and μ-N5–N6, 1.416(6) Å.³¹

Scheme 1 Synthesis of the dinuclear complex [Fe^III₂(hnmh-PLY)₃], 1.

Dinuclear complex 1 was further characterized by various techniques, such as elemental analysis, UV-vis, FT-IR, HRMS, XPS (X-ray photoelectron spectroscopy), and EPR (electron paramagnetic resonance) spectroscopy [ESI, Fig. S2–S5†]. From the UV-vis spectrum of complex 1, the band corresponding to the low-energy d–d transition appeared at 577 nm, and the wavelength corresponding to maximum absorption, λ_max appeared at 322 nm (ref. 31) (Fig. S4b†). The IR spectrum of complex 1 obtained using KBr discs shows characteristic peaks at ν_C=O, 1583 cm⁻¹; ν_C=N, 1533 cm⁻¹; ν_Fe–O 486 cm⁻¹, respectively (Fig. S3b†).⁴⁶ Complex 1 shows a molecular ion peak corresponding to [M + H]⁺ at m/z 1199.1984 amu. The +3 oxidation states of both Fe-centres are confirmed with characteristic XPS signals for Fe 2p_1/2 and Fe 2p_3/2 binding energies at 724, and 710 eV, respectively (Fig. S5a†).^47,48 The paramagnetic behavior of complex 1 was studied with solid-state EPR analysis at room temperature. From previous reports, a characteristic broad signal for weakly coupled paramagnetic diiron(III) complexes is well known at room temperature.^31,49 For complex 1, a similar broad EPR signal was observed at g ≈ 2.07 (Fig. S5b†).

Electrocatalytic reduction of H₂O₂

The mirror surface of the glassy carbon (GC) electrode was coated with complex 1 to fabricate the GC-1 electrode. (Details on the preparation of the electrode are given in the Experimental section.) Subsequently, the electrocatalytic reduction of H₂O₂ was examined in a three-electrode cell containing a GC-1 working electrode, an Ag/AgCl (3.0 M KCl) reference electrode, and a Pt-wire auxiliary electrode. The first CV test (the black line in Fig. 2a) was run in the absence of H₂O₂ to analyze the behavior of the GC-1 electrode in acetate buffer (pH 3). A notable increase in cathodic current of 420 μA with an onset potential ≈0.51 V (vs. Ag/AgCl electrode, 3.0 M KCl) was observed on the addition of 0.1 M H₂O₂ to the acetate buffer (pH 3.0), as shown in Fig. 2a. Therefore, the rise in cathodic current observed with the GC-1 electrode can be ascribed to the catalytic activity of complex 1 towards H₂O₂ reduction. Since the reduction of H₂O₂ is dependent on the active concentration of H⁺ ions, another CV test run after the addition of 10 mM HCl to the acetate buffer (resultant electrolyte ≈pH 2.8) containing 0.1 M H₂O₂. With the decrease in pH of the electrolyte (of the order 0.2), a positive shift of 0.059 pH or 12 mV was expected, considering the Nernst equation for the PCET (proton-coupled electron-transfer) mechanism. However, we could not record the expected minute change in the H₂O₂ reduction onset potential. Meanwhile, a noticeable increase in the current (by a factor of 340 μA) was observed, signifying the improvement in electrochemical H₂O₂ reduction and conductivity of electrolyte on increasing the active H⁺ ion concentration (Fig. 2a). A parallel set of tests was run on a bare GC electrode under identical conditions (shown in Fig. S6a†), where no such catalytic performance towards H₂O₂ reduction was noticed.

Fig. 2 (a) Cyclic voltammetry (CV) tests for H₂O₂ reduction in acetate buffer (pH 3.0) at the surface of a GC-1 electrode as the working electrode, Pt-wire as the auxiliary electrode, and Ag/AgCl (3.0 M KCl) as the reference electrode at a scan rate of 100 mV s⁻¹. The addition of 0.1 M H₂O₂ led to a large cathodic current, confirming the reduction of H₂O₂ by the GC-1 electrode. (b) One-compartment H₂O₂ fuel cell design with an Ni foam anode and a glassy carbon electrode (GCE) containing complex 1 as the cathode material in 0.1 M HCl electrolyte.

H₂O₂ fuel cell performance Test

The performance of a one-compartment H₂O₂ fuel cell was tested in acetate buffer (pH 3) and 0.1 M HCl using the GC-1 electrode as a cathode and Ni foam as an anode. The onset potential for H₂O₂ oxidation at the Ni surface appears in the range of −0.11 V to 0.05 V at various pH values of electrolytic solutions.^10,12,13 Since the onset potential for reduction of H₂O₂ at GC-1 is much higher than the oxidation of H₂O₂ at the Ni foam electrode, connecting these two electrodes in the H₂O₂ fuel cell can generate electric power. The GC-1 cathode and Ni foam anode were immersed in acetate buffer (pH 3) containing 0.3 M H₂O₂ as the fuel and oxidant. This type of fuel cell resulted in a very low OCP, ∼0.25 V, and as a consequence, a very low PPD (2.3 μW cm⁻²) was achieved (Fig. S7†). The most common electrolyte to run this type of a one-compartment fuel cell is an aqueous solution of 0.1 M HCl. In our recent work, we optimized pH conditions to get the best power output from a similar type of dinuclear iron(III) complex.³¹ Therefore, we conducted another fuel cell performance test in 0.1 M HCl electrolyte and 0.3 M H₂O₂, which resulted in a high OCP and PPD of 0.65 V and 1.31 mA cm⁻², respectively (Fig. 3b).

Fig. 3 I–V (red) and I–P (blue) curves for the performance of a one-compartment H₂O₂ fuel cell designed with a complex 1 modified electrode (GC-1) cathode and Ni foam anode. (a) Effect of H₂O₂ concentration on fuel cell performance recorded in 0.1 M HCl. (b) Optimized performance of complex 1 (GC-1 and GC-1C) as cathode and Ni foam as anode in 0.1 M HCl electrolyte and 0.3 M H₂O₂. All tests were conducted at a scan rate of 10 mV s⁻¹, and the current density was normalized to the surface area of the glassy carbon electrode (0.07 cm²). (c) Variation in open-circuit potential for ten consecutive measurements.

With a variation in H₂O₂ fuel concentration (0.1 M–0.4 M) in 0.1 M HCl, a notable shift in PPD was noted (Fig. 3a). In general, high reactant concentration enhances mass transfer.¹³ However, due to the inherent instability of H₂O₂, the decomposition rate of H₂O₂ increases with an increase in its concentration, which is reflected in the PPD (1.10 mW cm⁻²) obtained with 0.4 M H₂O₂.^12,13 The oxygen bubbles generated from the inevitable self-decomposition of high-concentration hydrogen peroxide prevent the reactant (H₂O₂) transferring to electroactive sites at the GC-1 electrode surface.^13,14 Further, switching the concentration of the electrolyte to 0.05 M HCl and 0.2 M HCl, OCPs of 0.60 V and 0.65 V and PPDs of 0.68 mW cm⁻² and 0.96 mW cm⁻² were obtained, respectively (Fig. S8a†). Under highly acidic conditions, i.e. 0.2 M HCl, the GC-1 electrode suffered stability issues, as the immobilized complex showed rapid dissolution in the electrolyte. Therefore, with the optimized conditions of 0.1 M HCl electrolyte and 0.3 M H₂O₂, stable fuel cell performance for the GC-1 cathode and Ni foam anode was obtained.

The addition of 10% wt/wt carbon (C) to complex 1, for the preparation of GC-1C, resulted in an improvement in the conductivity of the electrode, and a prominent two-fold increase in PPD (2.84 mW cm⁻²). The effect of the increase in conductivity with GC-1C can be noted through the I–V (red curve) plot given in Fig. 3b, and the polarization curve in Fig. S8b.† No increase in power density was noted on further increasing the concentration of carbon black. The stability of OCP with ten consecutive measurements is given in Fig. 3c. For the GC-1 electrode, OCP values shifted from 0.65 to 0.60 V, whereas with the GC-1C electrode, the OCP values decreased from 0.65 V to 0.63 V over ten sets of consecutive measurements. The stability of complex 1, GC-1 and GC-1C, after H₂O₂ fuel cell experiments under optimized conditions was confirmed with UV-vis spectroscopy and powder X-ray diffraction analysis (Fig. S9†).

A fuel cell demonstration setup, shown in Fig. S12,† was obtained to test the practical applicability of the designed fuel cell.

Electrochemical studies

The cyclic voltammogram of complex 1 obtained from a three-electrode cell; GC working electrode, Ag/AgCl (3.0 M KCl) reference electrode, and Pt wire as the auxiliary electrode is shown in Fig. 4. A homogeneous solution of 1 mM complex 1 in 15 mL of dried DMF (dimethyl formamide), containing 0.1 M Bu₄NPF₆ (tetrabutylammonium hexafluorophosphate) as the supporting electrolyte, was used in the experiments. The first reversible reduction event E₁^1/2 = −0.44 V (ΔE₁ = 58 mV) and second quasi-reversible reduction event at E₂^1/2 = −0.82 V (ΔE₂ = 90 mV) observed in complex 1 correspond to Fe^IIIFe^III/Fe^IIFe^III and Fe^IIFe^III/Fe^IIFe^II redox couples (Fig. 4a).⁵⁰ The redox events at further negative potentials, −1.45 V (ΔE₃ = 70 mV), −1.99 V (ΔE₄ = 80 mV), and −2.20 V (ΔE₅ = 120 mV), can be attributed to PLY-based reductions in complex 1. Similar iron(III)-based and PLY-based reductions were observed in the cyclic voltammogram of the iso-structural complex [Fe^III₂(hmbh-PLY)₃].³¹ The observed redox events in complex 1 are presented in Fig. 4b.

Fig. 4 (a) Cyclic voltammogram of complex 1 recorded in DMF solvent at a scan rate of 50 mV s⁻¹ with Ag/AgCl (3.0 M KCl) as the reference, Pt as the auxiliary, and glassy carbon as the working electrode. (b) Five successive one-electron reductions on complex 1 as depicted from CV studies.

Further, we have recorded the CVs of GC-1, GC-L (hnmh-PLYH₂ modified GCE), and GC-Fe (FeCl₃-modified GCE) in 0.1 M HCl to gain insights into the possible electron transfer taking place at complex 1 during the fuel cell performance tests (Fig. S10†). The CV of GC-1 recorded in 0.1 M HCl (0 to 0.65 V vs. Ag/AgCl electrode) showed a broad quasi-reversible redox-couple (ΔE_GC-1 = 120 mV), with peak reduction potential at 0.37 V. Similar peaks are often observed in Fe(III) complexes, corresponding to the Fe(III)/Fe(II) redox couple in acidic aqueous electrolyte^6,9,16 and were present in the CV of the GC-Fe electrode (Fig. S10a†). The addition of H₂O₂ has led to a large increase in cathodic current, with the disappearance of the complex 1 centred redox couple (Fig. S10b†). Therefore, the redox events centered over Fe(III)/Fe(II) can be responsible for initiating the H₂O₂ reduction process.^6,9,16 The mechanism for the H₂O₂ fuel cell fabricated with complex 1 as an anode and Ni foam anode can be proposed on account of the proton-coupled electron transfer on the two Fe³⁺ metal centers and coordinated (hnmh-PLY) ligands in an acidic electrolyte.^29,31 From H₂O₂ fuel cell equations in Chart 1, the oxidation of H₂O₂ at the surface of the Ni foam anode can generate H⁺ ions and free electrons. As a result of the potential difference between the Ni foam anode and GC-1 cathode, generated electrons can reach the GC-1 cathode, and meanwhile, the ligand (hnmh-PLY) in complex 1 can be protonated in an acidic medium, generating species 1A (Fig. 5).^9,31 Species 1A can undergo another reduction to form 1B, which can catalyze the reduction of one mole of H₂O₂, regenerating complex 1 (Fig. 5). Alternatively, through the comparison of the CV of GC-L and GC-1 (Fig. S10a and S10b†), the proceeding redox events on the PLY units of (hnmh-PLY) ligand may also facilitate the further reduction of H₂O₂, as proposed in earlier reports (Fig. S10c†).²⁹Table 1 shows a list of molecular complexes explored in a one-compartment H₂O₂ fuel cell.

Fig. 5 Plausible Fe-centred H₂O₂ reduction (proton-coupled electron transfer) mechanism by complex 1.

Table 1 Power output obtained in previously reported Fe(III)-based molecular systems as a cathode material in an H₂O₂ fuel cell

Sr. no.	Anode	Cathode	OCP (V)	PPD (mW cm^−2)	Ref.
1	Ni	[Fe^III(Pc)Cl]	0.50	0.01	Fukuzumi et al.^11
2	Ni	Fe^III(acac)2	0.25	0.089	Mandal et al.^29
3	Ni	[Fe^III(Phen)(PLY)Cl2]	0.53	0.61	Mandal et al.^29
4	Ni	Fe^III(PLY)3	0.74	1.53	Mandal et al.^29
5	Ni	[Fe^III2(hmbh-PLY)3]	0.65	2.41	Metre et al.^31
6	Ni	[Fe^III2(hnmh-PLY)3], 1	0.65	2.84	This work

---

Effect of ligand design on H₂O₂ fuel cell performance

In our previous investigation, we fabricated a GC electrode modified with an [Fe^III₂(hmbh-PLY)₃] complex.³¹ The power output (PPD) obtained in a one-compartment H₂O₂ fuel cell with complex [Fe^III₂(hmbh-PLY)₃] as cathode (without C black), and Ni foam as anode was 1.15 mW cm⁻² in 0.1 M HCl and 0.3 M H₂O₂. While, under identical conditions, the GC electrode modified with complex 1, [Fe^III₂(hnmh-PLY)₃] (without C black), afforded a power output of 1.31 mW cm⁻². For both complexes, the obtained PPD displayed a two-fold (≈2.2 times) increase on the addition of 10% wt/wt C, because of the improved conductivity. In this regard, we intend to highlight that the choice of aldehyde to functionalize the Hz-PLY reactant³¹ plays a vital role in improving the activity of similarly designed dinuclear iron(III) complexes towards H₂O₂. From a comparison of preliminary tests to check the catalytic activity of [Fe^III₂(hmbh-PLY)₃] and complex 1, the onset potential for the reduction of H₂O₂ under identical conditions was recorded at 0.43 V and 0.51 V, respectively. The ease of H₂O₂ reduction by complex 1 can be the contributing factor for improving the power output in the fuel cell presented here, in comparison to the fuel cell reported earlier with [Fe^III₂(hmbh-PLY)₃] as a cathode.^11,16

Conversely, the solid-state packing of PLY-derived molecules has been widely explored to understand the electron transfer mechanisms in molecular conductors as well as electronic devices.^28,51,52 The planar PLY system behaves as a nano-graphene fragment and is well known for showing intriguing supramolecular π–π stacking in the solid state, which can facilitate the charge transfer process on a molecular level.²⁶ In our recent report, we discussed the role of intermolecular π–π interactions between the PLY rings in a complex in contributing towards low-voltage switching in a resistive memory device.²⁷ The molecular structure of complex 1 in the solid state also showed very intriguing intramolecular π–π stacking interactions between the conjugated aromatic rings of the PLY-derived ligand. Three different π–π interactions were observed: (a) face-to-face π–π overlap of three aromatic rings of the PLY units [3.605(4) Å, 3.639(4) Å, and 3.748(4) Å]; (b) face-to-face overlap of the aromatic rings of naphthyl and PLY units [3.873(4) Å and 3.777(4) Å]; and (c) face-to-face overlap of aromatic naphthyl units [3.655(3) Å] (Fig. 6a). Constable et al. discussed the role of similar intramolecular π–π stacking interactions (face-to-face overlap) in Ir-coordinated phenyl rings in improving the performance of light-emitting electrochemical cells.⁵³ Very recently, Yang et al. reviewed that the presence of such intramolecular interactions can enhance the electron-accepting tendency of molecules and stabilize the excessive electronic charge via extensive delocalization.⁵⁴ In the present instance, the supramolecular assembly in complex 1 generated by strong intramolecular π–π stacking can be advantageous in stabilizing upcoming electrons from the Ni foam anode in an H₂O₂ fuel cell, through enhanced delocalization. The solid-state molecular structure of the [Fe^III₂(hmbh-PLY)₃] complex, in Fig. 6b, lacks the presence of such extended intramolecular π–π stacking interactions.

Fig. 6 Intramolecular π–π stacking interactions present in (a) 1 and (b) [Fe^III₂(hmbh-PLY)₃] complexes.

Further, the solution-state CV of the [Fe^III₂(hmbh-PLY)₃] complex displayed the first two Fe-centered reduction events Fe^IIIFe^III/Fe^IIFe^III and Fe^IIFe^III/Fe^IIFe^II at E₁^1/2 = −0.48 V (ΔE₁ = 68 mV, quasi-reversible) and E₂^1/2 = −0.86 V (ΔE₂ = 91 mV, quasi-reversible), respectively.³¹ In contrast, for complex 1, under identical conditions, the respective reduction events are observed at E₁^1/2 = −0.44 V (ΔE₁ = 58 mV, reversible) and E₂^1/2 = −0.82 V (ΔE₁ = 90 mV, quasi-reversible). Further, we compared the obtained solid-state CV of both complexes in 0.1 M HCl electrolyte (Fig. S11†). The CV test was run in the potential window (0.65 to 0 V) employing one-compartment H₂O₂ fuel cell test experiments. A GC electrode modified with an [Fe₂(hmbh-PLY)₃] complex showed a broad reduction event at a peak potential of 0.15 V vs. Ag/AgCl electrode, whereas CV recorded for the GC-1 electrode showed an early electrochemical reduction event, with a peak potential of 0.37 V vs. Ag/AgCl. The CV data (in solution as well in the solid state) signifies a decrease in reduction potentials for complex 1 in comparison to the previously reported [Fe^III₂(hmbh-PLY)₃] complex, because of the presence of intramolecular aromatic π–π stacking interactions.^55–57

Analogous to our previous study,³¹ the first two reductions (for CV recorded in solution) of complex 1 are predominantly metal-centered, as suggested by density functional theory (DFT) computations. MO analyses computed at the spin-unrestricted B3LYP method infers that the β-LUMO of complex 1 and its one- (1¹⁻) and two-electron (1²⁻) reduced complexes have nearly 44%, 27%, and 100% ligand contributions, respectively. The β-LUMO of complex 1 and that of previously reported [Fe^III₂(hmbh-PLY)₃] complex (ligand containing vinyl and PLY units) are shown in Fig. 7. The figure clearly depicts that for the [Fe^III₂(hmbh-PLY)₃] complex, the β-LUMO is delocalized over two non-stacked PLY units, whereas it is delocalized over three PLY units in complex 1; among which two PLY units appear to be involved in direct π–π stacking interaction. The β-LUMO of the present complex is found to be nearly 2.65 kcal mol⁻¹ lower than that of the previous reported complex, and the lowering of the β-LUMO energy level is attributed to the enhanced orbital overlap between PLY units of two different ligands. In the optimized structure of complex 1, the two-PLY groups are stacked together with a separation of ∼3.3 Å; a distance within the range of π–π stacking interaction. Furthermore, the DFT computed absolute reduction potential (E_ABS) for the first reduction peak of complex 1 is found to be nearly 13.08 meV lower than that of the [Fe^III₂(hmbh-PLY)₃] complex.³¹ The decrease in reduction potential value due to the aromatic π–π stacking interaction was also observed in a few other metal complexes.^55–57 Analogous to the non-reduced species, the β-LUMO of the two-electron reduced complex, 1²⁻, is also found to be highly delocalized over the two π–π-stacked PLY units. Therefore, it is imperative from the above study that the extensive π–π-stacking interaction^58,59 between ligand moieties in 1 helps the reduction of the metal centers and can stabilize the transition state (TS) structure. The purpose of this comparative study is to point out the effect of extensive π–π-stacking in stabilizing the complex, as well as its molecular properties. To elucidate it further, we calculated the ligand–ligand intermolecular π–π-stacking interaction energy within molecular complex 1 at the DFT-D/TZVP level of theory. To do so, metal atoms were deleted from the optimized structure of complex 1, and then, single-point energy calculations were carried out for the resulting complex and that of three independent ligands. The counterpoise (CP)-corrected⁵⁹ stabilization energy was then evaluated by taking the energy difference between the complex and sum of three independent ligands. It was found that the calculated stabilization energy for the present complex is ∼2.19 kcal mol⁻¹ larger than the previously reported [Fe^III₂(hmbh-PLY)₃] complex.

Fig. 7 β-LUMO and spin-density plots of 1 and [Fe₂(hmbh-PLY)₃] complexes and their respective reduced species. Spin density plots are generated at an 0.0005 au (e2/bohr4) isosurface value (blue, α spin; red, β spin density). Molecular orbitals are generated at an isosurface value of 0.02 au.

Conclusions

A dinuclear [Fe^III₂(hnmh-PLY)₃] complex, 1 was prepared using a new PLY-based Schiff-base ligand, hnmh-PLYH₂. The electroactive nature of complex 1 was confirmed by the CV technique, which showed five different redox events on a cathodic sweep. When complex 1 modified electrode GC-1 was utilized as a cathode material in a one-compartment H₂O₂ fuel cell, an impressive PPD of 2.84 mW cm⁻² was achieved. The electro-catalytic behavior of complex 1 was compared to the previously explored isostructural [Fe^III₂(hmbh-PLY)₃] complex with respect to distinct ligand design. The presence of intramolecular π–π stacking interactions, localized over the PLY rings, provided enhanced stabilization to β-LUMO orbitals in complex 1. The stabilization of the β-LUMO orbitals for 1, 1¹⁻ and 1²⁻ species facilitated the Fe-centered reduction events, which are crucial steps for the H₂O₂ reduction in the present study.

Author contributions

Nisha Kamboj: methodology, data curation, formal analysis, investigation, and writing – original draft. Ayan Dey: software. Sunita: formal analysis. Moumita Majumder: computational investigation, formal analysis, writing, and reviewing. Srijan Sengupta: supervision, formal analysis and reviewing. Ramesh K. Metre: conceptualization, supervision, project administration, and writing – reviewing & editing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

R. K. M. acknowledges financial support from SERB EMEQ(EEQ/2020/000588), IIT Jodhpur, and SEED Grant (I/SEED/RKM/20170008) projects. S. S. acknowledges the SERB funding Grant (CRG/2019/001440). N. K. and A. D. are thankful to MOE for the financial support. S. B. is thankful to UGC for the research fellowship. The authors acknowledge the Centre for Research and Development of Scientific Instruments (CRDSI), IIT Jodhpur for all the characterization facilities and High-Performance Computing Centre, IIT Jodhpur for providing computing time. The authors are thankful to IIT Kanpur for Single crystal XRD data collection and the SAIF facility, IIT Bombay for EPR and elemental analysis. We acknowledge Dr Biswajit Saha and his PhD student, Debashree Bora from CSIR NEIST, Jorhat for XPS characterization.

References

1. S. Yamazaki, Z. Siroma, H. Senoh, T. Ioroi, N. Fujiwara and K. Yasuda, J. Power Sources, 2008, 178, 20–25.
2. L. An, T. Zhao, X. Yan, X. Zhou and P. Tan, Sci. Bull., 2015, 60, 55–64.
3. C. Liu, Y. Gao, L. Liu, C. Sun, P. Jiang and J. Liu, ACS Omega, 2022, 7, 28646–28657.
4. L. An, T. S. Zhao, Z. H. Chai, L. Zeng and P. Tan, Int. J. Hydrogen Energy, 2014, 39, 7407–7416.
5. A. L. G. Biancolli, S. Bsoul-Haj, J. C. Douglin, A. S. Barbosa, R. R. de Sousa, O. Rodrigues, A. J. C. Lanfredi, D. R. Dekel and E. I. Santiago, J. Membr. Sci, 2022, 641, 119879.
6. S. Fukuzumi and Y. Yamada, ChemElectroChem, 2016, 3, 1978–1989.
7. E. Miglbauer, P. J. Wójcik and E. D. Głowacki, Chem. Commun., 2018, 54, 11873–11876.
8. Y. Yang, Y. Xue, F. Huang, H. Zhang, K. Tao, R. Zhang, Q. Shen and H. Chang, ACS Appl. Energy Mater., 2018, 1, 5328–5335.
9. S. Fukuzumi, Y. Yamada and K. D. Karlin, Electrochim. Acta, 2012, 82, 493–511.
10. Y. Yamada, M. Yoneda and S. Fukuzumi, Energy Environ. Sci., 2015, 8, 1698–1701.
11. Y. Yamada, S. Yoshida, T. Honda and S. Fukuzumi, Energy Environ. Sci., 2011, 4, 2822–2825.
12. S. Wang, D. Ye, Z. Liu, X. Zhu, R. Chen, Q. Liao, Y. Yang and H. Liu, Int. J. Hydrogen Energy, 2022, 47, 4793–4803.
13. Y. Yang, Y. Xue, H. Zhang and H. Chang, Chem. Eng. J., 2019, 369, 813–817.
14. X. Sun, J. Chen, J. Zhai, H. Zhang and S. Dong, J. Am. Chem. Soc., 2022, 50, 23073–23080.
15. S. A. Mousavi Shaegh, N.-T. Nguyen, S. M. Mousavi Ehteshami and S. H. Chan, Energy Environ. Sci., 2012, 5, 8225–8228.
16. Y. Yamada, M. Yoneda and S. Fukuzumi, Chem. – Eur. J., 2013, 19, 11733–11741.
17. Y. Yamada, M. Yoneda and S. Fukuzumi, Inorg. Chem., 2014, 53, 1272–1274.
18. O. R. Luca and R. H. Crabtree, Chem. Soc. Rev., 2013, 42, 1440–1459.
19. W. I. Dzik, J. I. Van Der Vlugt, J. N. H. Reek and B. De Bruin, Angew. Chem., Int. Ed., 2011, 50, 3356–3358.
20. V. K. K. Praneeth, M. R. Ringenberg and T. R. Ward, Angew. Chem., Int. Ed., 2012, 51, 10228–10234.
21. N. Kamboj, G. Mali, P. Lama, R. D. Erande and R. K. Metre, ACS Omega, 2022, 7, 8789–8797.
22. G. Mali, G. Tiwari, N. Kamboj, R. K. Metre and R. D. Erande, Tetrahedron Lett., 2023, 124, 154608.
23. A. Mukherjee, S. C. Sau and S. K. Mandal, Acc. Chem. Res., 2017, 50, 1679–1691.
24. D. H. Reid, Q. Rev., Chem. Soc., 1965, 19, 274–302.
25. R. G. Hicks, Nat. Chem., 2011, 3, 189–191.
26. J. Ahmed and S. K. Mandal, Chem. Rev., 2022, 122, 11369–11431.
27. N. Kamboj, A. Betal, M. Majumder, S. Sahu and R. K. Metre, Inorg. Chem., 2023, 62, 4170–4180.
28. K. V. Raman, A. M. Kamerbeek, A. Mukherjee, N. Atodiresei, T. K. Sen, P. Lazić, V. Caciuc, R. Michel, D. Stalke, S. K. Mandal, S. Blügel, M. Münzenberg and J. S. Moodera, Nature, 2013, 493, 509–513.
29. A. Pariyar, G. Vijaykumar, M. Bhunia, S. K. Dey, S. K. Singh, S. Kurungot and S. K. Mandal, J. Am. Chem. Soc., 2015, 137, 5955–5960.
30. Z. Mou, K. Uchida, T. Kubo and M. Kertesz, J. Am. Chem. Soc., 2015, 137, 13989.
31. N. Kamboj, A. Dey, P. Lama, M. Majumder, S. Sengupta and R. K. Metre, Dalton Trans., 2023, 52, 17163–17175.
32. G. Vijaykumar, M. Bhunia and S. K. Mandal, Dalton Trans., 2019, 48, 5779–5784.
33. M. E. Itkis, X. Chi, A. W. Cordes and R. C. Haddon, Science, 2002, 296, 1443–1445.
34. M. Bhunia, S. R. Sahoo, B. K. Shaw, S. Vaidya, A. Pariyar, G. Vijaykumar, D. Adhikari and S. K. Mandal, Chem. Sci., 2019, 10, 7433–7441.
35. A. I. Vogel and B. S. Furniss, Vogel's Textbook of Practical Organic Chemistry, 2003.
36. F. Neese, F. Wennmohs, U. Becker and C. Riplinger, J. Chem. Phys., 2020, 152, 224108.
37. F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 7, 3297–3305.
38. F. Weigend, Phys. Chem. Chem. Phys., 2006, 8, 1057–1065.
39. S. Grimme, S. Ehrlich and L. Goerigk, J. Comput. Chem., 2011, 32, 1456–1465.
40. V. Barone and M. Cossi, J. Phys. Chem. A, 1998, 102, 1995–2001.
41. W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graphics, 1996, 14, 33–38.
42. R. L. McCreery, Chem. Rev., 2008, 108, 2646–2687.
43. S. Dale, N. M. Bonanno, M. Pelaccia, A. J. Lough, A. Miyawaki, K. Takahashi and M. T. Lemaire, Dalton Trans., 2019, 48, 8053–8056.
44. D. J. Bakker, A. Dey, D. P. Tabor, Q. Ong, J. Mahé, M.-P. Gaigeot, E. L. Sibert and A. M. Rijs, Phys. Chem. Chem. Phys., 2017, 19, 20343–20356.
45. R. M. Issa, A. M. Khedr and H. F. Rizk, Spectrochim. Acta, Part A, 2005, 62, 621–629.
46. O. Cozar, N. Leopold, C. Jelic, V. Chiş, L. David, A. Mocanu and M. Tomoaia-Cotişel, J. Mol. Struct., 2006, 788, 1–6.
47. X.-X. Sun, J. Du, J.-J. Tan and S.-Z. Zhan, New J. Chem., 2022, 46, 3794–3799.
48. Ü. İşci, P. Afanasiev, J.-M. M. Millet, E. V. Kudrik, V. Ahsen and A. B. Sorokin, Dalton Trans., 2009, 7410–7420.
49. X. Zhang, Y. Y. Li, J. Jiang, R. Zhang, R. Z. Liao and M. Wang, Inorg. Chem., 2020, 59, 5424–5432.
50. J. Saroja, V. Manivannan, P. Chakraborty and S. Pal, Inorg. Chem., 1995, 34, 3099–3101.
51. S. K. Pal, M. E. Itkis, R. W. Reed, R. T. Oakley, A. W. Cordes, F. S. Tham, T. Siegrist and R. C. Haddon, J. Am. Chem. Soc., 2004, 126, 1478–1484.
52. X. Chi, M. E. Itkis, R. W. Reed, R. T. Oakley, A. W. Cordes and R. C. Haddon, J. Phys. Chem. B, 2002, 106, 8278–8287.
53. R. D. Costa, E. Ortí, H. J. Bolink, S. Graber, C. E. Housecroft and E. C. Constable, J. Am. Chem. Soc., 2010, 132, 5978–5980.
54. B. Huang, L. Mao, X. Shi and H.-B. Yang, Chem. Sci., 2021, 12, 13648–13663.
55. P. Hosseinzadeh and Y. Lu, Biochim. Biophys. Acta, Bioenerg., 2016, 1857, 557–581.
56. M. V. Ponamarev, B. G. Schlarb, C. J. Howe, C. J. Carrell, J. L. Smith, D. S. Bendall and W. A. Cramer, Biochemistry, 2000, 39, 5971–5976.
57. K. Ozawa, Y. Takayama, F. Yasukawa, T. Ohmura, M. A. Cusanovich, Y. Tomimoto, H. Ogata, Y. Higuchi and H. Akutsu, Biophys. J., 2003, 85, 3367–3374.
58. S. Tsuzuki, T. Uchimaru, K. Matsumura, M. Mikami and K. Tanabe, Chem. Phys. Lett., 2000, 319, 547–554.
59. Y. Xie, D. W. Shaffer and J. J. Concepcion, Inorg. Chem., 2018, 57, 10533–10542.

---

Footnote

† Electronic supplementary information (ESI) available: Crystal structure refinement parameters, ¹H and ¹³C NMR, UV-vis, FT-IR, HRMS, XPS, EPR spectra, and details of additional H₂O₂ fuel cell experiments. CCDC 2304746 and 2304747. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4dt00134f

---