Crossing the bridge from molecular catalysis to a heterogenous electrode in electrocatalytic water oxidation

Significance

Artificial photosynthesis by the preparation of solar fuels is a promising strategy for our energy future. An inhibiting factor in artificial photosynthesis is the half-reaction for water oxidation, which, because of its slow kinetics, makes catalysis indispensable. Advances in molecular water oxidation catalysis from detailed mechanistic investigations have provided insight into how this reaction occurs, but in practical applications, including electrolytic cells and dye-sensitized photoelectrosynthesis cells, it must be applied on conductive substrates. We describe here an efficient strategy that combines silane surface functionalization and reductive electropolymerization to bind a molecular catalyst on conductive substrates for sustained catalytic water oxidation.

Abstract

Significant progress has been made in designing single-site molecular Ru(II)-polypyridyl-aqua catalysts for homogenous catalytic water oxidation. Surface binding and transfer of the catalytic reactivity onto conductive substrates provides a basis for heterogeneous applications in electrolytic cells and dye-sensitized photoelectrosynthesis cells (DSPECs). Earlier efforts have focused on phosphonic acid (-PO₃H₂) or carboxylic acid (-CO₂H) bindings on oxide surfaces. However, issues remain with limited surface stabilities, especially in aqueous solutions at higher pH under conditions that favor water oxidation by reducing the thermodynamic barrier and accelerating the catalytic rate using atom-proton transfer (APT) pathways. Here, we address the problem by combining silane surface functionalization and surface reductive electropolymerization on mesoporous, nanofilms of indium tin oxide (ITO) on fluorine-doped tin oxide (FTO) substrates (FTO|nanoITO). FTO|nanoITO electrodes were functionalized with vinyltrimethoxysilane (VTMS) to introduce vinyl groups on the electrode surfaces by silane attachment, followed by surface electropolymerization of the vinyl-derivatized complex, [Ru^II(Mebimpy)(dvbpy)(OH₂)]²⁺ (1²⁺; Mebimpy: 2,6-bis(1-methyl-1H-benzo[d]imidazol-2-yl)pyridine; dvbpy: 5,5′-divinyl-2,2′-bipyridine), in a mechanism dominated by a grafting-through method. The surface coverage of catalyst 1²⁺ was controlled by the number of electropolymerization cycles. The combined silane attachment/cross-linked polymer network stabilized 1²⁺ on the electrode surface under a variety of conditions especially at pH > ∼6. Surface-grafted poly1²⁺ was stable toward redox cycling at pH ∼ 7.5 over an ∼4-h period. Sustained heterogeneous electrocatalytic water oxidation by the electrode gave steady-state currents for at least ∼6 h with a Faradaic efficiency of ∼68% for O₂ production.

Catalytic water oxidation, ${2\text{H}}_2\text{O}\to {4\text{H}}^{+}+{4\text{e}}^{-}+{\text{O}}_2$, plays a key role as the “other half-reaction” in artificial photosynthesis (1–12). An important factor that hinders the reaction arises from the sluggish kinetics of 4H⁺/4e⁻ loss and O-O bond formation (1–4, 8, 11, 13). Inspired by the oxygen-evolving complex in Photosystem II, the first molecular catalyst, the “blue dimer,” cis,cis-[(bpy)₂(H₂O)Ru^III(μ-O)Ru^III(OH₂)(bpy)₂]⁴⁺ (bpy: 2,2′-bipyridine) (14–16), was designed with the mechanism that a nucleophilic attack at a single Ru^V = O site by an external water molecule occurs for rate-limiting O-O bond formation (16).

Significant progress has been made in this area for developing single-site molecular catalysts for homogenous water oxidation based on extensive mechanistic studies on Ru(II)-polypyridyl-aqua complexes (17–24). In the single-site Ru(II) catalyst, [Ru^II(Mebimpy)(bpy)(OH₂)]²⁺ (Mebimpy: 2,6-bis(1-methyl-1H-benzo[d]imidazol-2-yl)pyridine) (SI Appendix, Fig. S1) (3, 17), sequential oxidative activation occurs by proton-coupled electron transfer (PCET), [Ru^II-OH₂]²⁺$\stackrel{-{1\text{H}}^{+}/{1\text{e}}^{-}}{\to }$[Ru^III-OH]²⁺$\stackrel{-{1\text{H}}^{+}/{1\text{e}}^{-}}{\to }$[Ru^IV = O]²⁺, to accumulate multiple oxidative equivalents at a single metal site while avoiding charge buildup. Additional 1e⁻ loss gives dπ³ [Ru^V = O]³⁺, which is reactive toward nucleophilic attack by a water molecule with the rate-limiting O-O bond formation and proton loss to give the hydroperoxide intermediate [Ru^III-OOH]²⁺. The latter undergoes further oxidization to [Ru^IV-OO]²⁺ and O₂ loss, reentering the catalytic cycle as [Ru^II-OH₂]²⁺. The intermediate, [Ru^IV-OO]²⁺, can also be further oxidized to [Ru^V-OO]³⁺ with reentry into the catalytic cycle by rapid loss of O₂ and regeneration of [Ru^III-OH]²⁺ or [Ru^III-OH₂]³⁺ depending on the pH. At increased pH, or with added bases, the catalytic rate is accelerated by the base (B) or OH⁻ acting as the proton acceptor in the rate-limiting step to form [Ru^III-OOH]²⁺ by an atom-proton transfer (APT) pathway, [Ru^V = O]³⁺—O(H)H—B$\stackrel{\text{APT}}{\to }$[Ru^III-OOH]²⁺—⁺HB (3, 17).

Transfer of homogeneous catalytic reactivity to conductive substrates opens new doors for heterogenous applications in electrolytic cells and dye-sensitized photoelectrosynthesis cells (DSPECs) (17, 25–30). However, most attempts have utilized phosphonic acid (-PO₃H₂) or carboxylic acid (-CO₂H) binding strategies on oxide surfaces and are limited by surface hydrolysis especially above pH ∼ 6 (31). Higher pH is desirable because of facilitated water oxidation by reducing thermodynamic barrier and accelerating catalytic rate using added bases by an APT pathway (3, 17). Surface stabilization of -PO₃H₂ or -CO₂H bindings by atomic layer deposition (ALD) (27, 28), polymer dip-coating (32, 33), or in situ electropolymerization have also been investigated (29).

We report here an effective fabrication strategy that combines silane attachment and surface reductive electropolymerization for stabilizing Ru(II) catalysts on conductive substrates for sustained, heterogeneous catalytic water oxidation, especially at pH > ∼6 (Fig. 1). Mesoporous nanofilms of indium tin oxide (ITO), ∼10- to 15-nm particle diameter and ∼4.5-μm film thickness, on fluorine-doped tin oxide (FTO) substrates (FTO|nanoITO) (SI Appendix, Fig. S2 and refs. 27–29) were functionalized with vinyltrimethoxysilane (VTMS) to introduce vinyl group on the surface (i in Fig. 1B). The catalyst [Ru^II(Mebimpy)(dvbpy)(OH₂)]²⁺ (1²⁺; dvbpy: 5,5′-divinyl-2,2′-bipyridine) was grafted onto VTMS-functionalized FTO|nanoITO electrodes by surface reductive electropolymerization by a grafting-through pathway producing FTO|nanoITO|-g-poly1²⁺ (ii in Fig. 1B; g is an abbreviation for grafted). The combined silane attached, cross-linked polymer network firmly anchored the Ru(II) catalyst on the electrode with high surface stability under various conditions notably at pH > ∼6. The surface-grafted Ru(II) catalyst was stable toward redox cycling at pH ∼ 7.5 over an ∼4-h period. Sustained electrocatalytic water oxidation by the FTO|nanoITO|-g-poly1²⁺ electrode gave steady-state currents for at least ∼6 h with a Faradaic efficiency of ∼68% for O₂ production.

Fig. 1.

Scheme for the electrode fabrication. (A) Structure of complex 1²⁺. (B) Fabrication of FTO|nanoITO|-g-poly1²⁺ by silane surface functionalization and surface reductive polymerization. The structure of the catalyst is abbreviated for a clear illustration.

Results

Synthesis of the Vinyl-Derivatized Ru(II) Catalyst, 1²⁺.

Complex 1²⁺ (Fig. 1A) was synthesized according to literature procedures with slight modifications (SI Appendix, Fig. S3) (29). Free radicals can be anionically generated on the aromatic vinyl groups based on ligand reduction for initiation and chain propagation (34). 5,5′-Divinyl substitutes on the bpy ligand ensure an efficient cross-linking during polymerization which creates a robust polymer network grafted on the electrode surface (35) while maintaining the electroactivity of the complex for catalytic water oxidation (Fig. 1).

The UV-visible (UV-vis) spectrum of 1²⁺ in methanol features a characteristic metal-to-ligand charge transfer (MLCT) absorption band in the visible region at λ_max ∼ 500 nm with ε(λ_max) ∼ 10,100 M⁻¹·cm⁻¹ (SI Appendix, Fig. S4). Complex 1²⁺ has a Cs symmetry with its mirror plane (δ_h) perpendicular to the tridentate Mebimpy ligand (Fig. 1A). The symmetry break in the mirror plane results in inequivalent protons in the bpy-based ligand leading to a complicated assignment of the ¹H NMR spectrum (SI Appendix, Fig. S5). Electrospray ionization mass spectrometry (ESI-MS) clearly detected the complex with matching m/z peaks (SI Appendix, Fig. S6).

Fabrication of the Electrode FTO|nanoITO|-g-poly1²⁺.

Before electropolymerization, FTO|nanoITO was functionalized with VTMS by soaking in a stock solution, 65 mM in toluene for ∼3.5 d (31, 36). Surface modification with silanes is the most widely used method for preparing self-assembled monolayers (SAMs) on oxides (35, 37, 38). In the aprotic solvent, i.e., toluene here, a SAM formed on the mesoporous nanoITO surface by the robust -O-Si-O-Si-O- bonding from condensation of VTMS with either surface -OH groups or neighboring silanes. In the procedure, the vinyl groups were introduced on the electrode surface for electropolymerization of 1²⁺ (i in Fig. 1B).

Surface reductive electropolymerization of 1²⁺ (0.5 mM 1²⁺ in 0.1 M TBAPF₆/PC; TBAPF₆: tetra-n-butylammonium hexafluorophosphate; PC: propylene carbonate) was conducted by cycling VTMS-functionalized FTO|nanoITO as the working electrode between −1 and −2 V (vs. Ag/AgNO₃) at a scan rate of 100 mV/s with vigorous stirring under an Ar atmosphere (ii in Fig. 1B). PC was used as the solvent to avoid aqua-ligand replacement (29). After electropolymerization, the FTO|nanoITO|-g-poly1²⁺ electrodes were rinsed with water/methanol (1/1, vol/vol) to remove any physisorbed moieties. The pinkish-brown color of the electrode clearly indicated the presence of grafted catalyst 1²⁺ on the surface.

Surface reductive electropolymerization occurs mechanistically by a grafting-through pathway (Fig. 2) (35). The vinyl group in the surface-bound VTMS is stable in the applied potential window from −1 to −2 V (SI Appendix, Fig. S7), and free radicals can only be anionically generated on the aromatic vinyl sites in the monomer 1²⁺ in the electrolyte solution followed by chain propagation. During polymerization, the surface-bound vinyl groups become integrated into growing polymeric chains leading to a permanent surface-grafting of the polymers (i in Fig. 2). The surface-anchored polymer chains continue to grow by adding more vinyl-derivatized complexes or their oligomeric chains (ii or iii in Fig. 2).

Fig. 2.

Schematic illustration of surface reductive electropolymerization of 1²⁺ on VTMS-functionalized FTO|nanoITO electrodes by a grafting-through pathway.

Characterization of the Electrode FTO|nanoITO|-g-poly1²⁺.

FTO|nanoITO|-g-poly1²⁺ electrodes, following different numbers of electropolymerization cycles, were investigated spectrophotometrically and electrochemically (Fig. 3 A and B). UV-vis spectra of FTO|nanoITO|-g-poly1²⁺ exhibited characteristic MLCT absorption bands centered at ∼500 nm, consistent with an intact Ru(II) coordination environment during the surface reductive electropolymerization process (Fig. 3A). The slightly broadened MLCT absorption was due to light scattering in the mesoporous nanoITO films on the FTO substrate. MLCT absorption increased with the number of electropolymerization cycles consistent with addition of 1²⁺ grafted on the surface (Fig. 3A). Surface coverages of 1²⁺ (Γ in nanomoles per square centimeter) were determined by the expression A(λ_max)/(1,000·ε(λ_max)), where A(λ_max) is the measured absorbance at λ_max with ε(λ_max) taken for the solution analog (Fig. 3C) (17). Γ was increased linearly with cycle number up to ∼32 nmol/cm² following 30 electropolymerization cycles, which is equivalent to a single-layer of the analogous Ru(II)-polypyridyl-aqua catalyst on the surface (26, 33). Past 40 cycles, surface coverage continued to increase but at a slower rate with a plateau after 150 cycles, demonstrating inhibited electron transfer from the FTO substrate through nanoITO films to the electrolyte and decrease in void pores in the mesoporous structures toward monomer diffusion.

Fig. 3.

Spectrophotometrical and electrochemical characterizations of the electrode FTO|nanoITO|-g-poly1²⁺. (A and B) UV-vis spectra (A) and CVs (B) of FTO|nanoITO|-g-poly1²⁺ electrodes in 0.1 M HClO₄ with a Ag/AgCl reference electrode (0.198 V vs. NHE) and a Pt wire counter electrode at a scan rate of 25 mV/s following different numbers of electropolymerization cycles. The currents were normalized to the geometric areas of the working electrodes. (C) Surface coverage of 1²⁺ calculated from UV-vis and cyclic voltammetry characterizations. UV-vis absorption measurements were corrected to VTMS-functionalized FTO|nanoITO as the blank slide.

Cyclic voltammograms (CVs) of FTO|nanoITO|-g-poly1²⁺ were recorded in 0.1 M HClO₄ aqueous solutions (pH ∼ 1.0) at a scan rate of 25 mV/s with a Ag/AgCl reference electrode and a Pt wire counter electrode (Fig. 3B). A low potential sweep through the [Ru^III-OH₂]³⁺/[Ru^II-OH₂]²⁺ couple at E_1/2 ∼ 0.89 V (vs. NHE) was used to avoid other complications from higher potential couples. At this pH, the [Ru^III-OH₂]³⁺/[Ru^II-OH₂]²⁺ couple exists instead of [Ru^III-OH]²⁺/[Ru^II-OH₂]²⁺ due to the pK_a ∼ 3.0 for [Ru^III-OH₂]³⁺ (17, 19). The integrated anodic wave for the couple increased with the number of electropolymerization cycles, corresponding to the spectrophotometric measurements (Fig. 3 A and B). The expression Q/nFA was used to estimate surface coverage of 1²⁺, where Q and n are the integrated anodic wave and number of electrons transferred for the [Ru^III-OH₂]³⁺/[Ru^II-OH₂]²⁺ couple, F is the Faraday constant, and A is the geometric area of the electrode (Fig. 3C) (17).

Unlike previous results based on -PO₃H₂ binding with formation of a single molecular layer of catalyst (17, 25, 26), the surface coverages calculated from spectrophotometric and electrochemical measurements did not perfectly match (Fig. 3C). Surface coverages obtained by cyclic voltammetry were lower than values by UV-vis analysis with the deviations increasing as the number of electropolymerization cycles increased. At the scan rate shown, 25 mV/s, the electrochemical result indicates only partial oxidation of [Ru^II-OH₂]²⁺ to [Ru^III-OH₂]³⁺ at E_1/2 ∼ 0.89 V (vs. NHE) (39) with oxidation of the remaining sites inhibited by slow electron transfer within the surface-grafted polymeric network.

Chemical Stability.

The stability of poly1²⁺-grafted electrodes was investigated in the dark in 0.1 M HClO₄ (pH ∼ 1.0), 0.1 M phosphate buffer (pH ∼ 3.0, pH ∼ 7.5, and pH ∼ 10.0), pure water, methanol, ethanol, and dimethylformamide. UV-vis spectra of the electrode were monitored in situ at 1-h intervals over a period of ∼25 h. Fig. 4 shows the results of a typical stability study at pH ∼ 10.0 in 0.1 M phosphate buffer. The data show that the UV-vis absorption spectrum remains constant (Fig. 4 and SI Appendix, Fig. S8), indicating a firmly bound Ru(II) catalyst on the surface with extended chemical stability. Similar results were obtained under the conditions mentioned above.

Fig. 4.

UV-vis absorption changes for FTO|nanoITO|-g-poly1²⁺ (50 electropolymerization cycles) at pH ∼ 10.0 in 0.1 M phosphate buffer over an ∼25-h period (from black to green).

Stabilization of the surface-grafted poly1²⁺ arises from a combination of silane attachment and cross-linked polymer network by (i) stabilized surface binding of the silane SAM, which is robust and resistant to surface desorption by hydrolysis under a variety of conditions especially at pH > ∼6; and (ii) the cross-linked hydrophobic polymer network, which both mechanically and chemically reinforces and stabilizes Ru(II) sites on the surface against desorption and ligand substitution.

Electrochemical Characterization at pH > 6.

Stabilization of surface-grafted poly1²⁺ allows for a systematic investigation of its heterogeneous catalytic behavior at higher pH (SI Appendix, Fig. S9). CVs of FTO|nanoITO|-g-poly1²⁺ (50 electropolymerization cycles) at pH ∼ 7.5, pH ∼ 8.3, pH ∼ 9.0, and pH ∼ 10.0 in 0.1 M phosphate buffers at a scan rate of 5 mV/s showed characteristic redox waves for the [Ru^III-OH]²⁺/[Ru^II-OH₂]²⁺ and [Ru^IV = O]²⁺/[Ru^III-OH]²⁺ couples followed by a current increase for catalytic water oxidation (SI Appendix, Fig. S9). The decrease in oxidation peak potentials for the couples with increased pH (inset table) is qualitatively consistent with the PCET oxidative activation sequence [Ru^II-OH₂]²⁺$\stackrel{-{1\text{H}}^{+}/{1\text{e}}^{-}}{\to }$[Ru^III-OH]²⁺$\stackrel{-{1\text{H}}^{+}/{1\text{e}}^{-}}{\to }$[Ru^IV = O]²⁺. Appearance of the expected pH dependence is significant in demonstrating the advantage of the surface-grafted polymer layers compared with ALD stabilization, where activation by PCET, especially for the couple [Ru^III-OH]²⁺$\stackrel{-{1\text{H}}^{+}/{1\text{e}}^{-}}{\to }$[Ru^IV = O]²⁺, can be inhibited by the protective insulating metal oxide overlayer (27, 28).

The increase in background subtracted electrocatalytic current response with increasing pH (SI Appendix, Fig. S9) is consistent with an accelerated catalytic rate with added buffer bases or OH⁻ as the proton acceptor by APT pathways in the rate-limiting step with the formation of [Ru^III-OOH]²⁺, that [Ru^V = O]³⁺—O(H)H—B$\stackrel{\text{APT}}{\to }$[Ru^III-OOH]²⁺—⁺HB as discussed above.

Past pH ∼ 11, cyclic voltammetry measurements revealed a rapid and irreversible loss in current response for the surface-bound catalyst during electrochemical scans between 0.2 and 1.6 V (vs. NHE). The loss of catalytic activity was not due to surface detachment from hydrolysis but, rather, arises from ligand decomposition or other degradation pathways involving the complex (40–43). To avoid complexities at higher pH, subsequent detailed investigations were conducted at pH ∼ 7.5.

Stability Toward Redox Cycling at pH ∼ 7.5.

The electrochemical stability of FTO|nanoITO|-g-poly1²⁺ electrodes (50 electropolymerization cycles) was examined by repeated cyclic voltammetry cycles at a scan rate of 20 mV/s at pH ∼ 7.5 in 0.1 M phosphate buffer under an Ar atmosphere (Fig. 5). The potential of the electrode was scanned between 0.2 and 0.9 V (vs. NHE) to access the [Ru^III-OH]²⁺/[Ru^II-OH₂]²⁺ and [Ru^IV = O]²⁺/[Ru^III-OH]²⁺ couples before the onset for catalytic water oxidation to obtain quasireversible redox waves for quantitative analysis (Fig. 5).

Fig. 5.

Successive (from black to green) cyclic voltammetry cycles for a FTO|nanoITO|-g-poly1²⁺ electrode (50 electropolymerization cycles; ∼1 cm² in geometric area) at a scan rate of 20 mV/s at pH ∼ 7.5 in 0.1 M phosphate buffer; Ag/AgCl reference electrode and Pt wire counter electrode.

The background-subtracted anodic peak current for the [Ru^III-OH]²⁺/[Ru^II-OH₂]²⁺ wave provides a direct monitor of the molecular catalyst on the surface (33). Due to equilibration between the local ionic environment and the external medium within the polymer layers, the anodic peak current increased in magnitude through the initial series of 50 cycles and then remained stable for the following cyclic voltammetry scans over an ∼4-h period (SI Appendix, Fig. S10). The experiments demonstrated the stability of the surface-grafted poly1²⁺ structure toward redox cycling (Fig. 5 and SI Appendix, Fig. S10). There was no obvious intensity decrease in the UV-vis spectrum of the electrode after the redox cycles were completed (SI Appendix, Fig. S11).

Water Oxidation at pH ∼ 7.5.

Catalytic water oxidation by FTO|nanoITO|-g-poly1²⁺ (50 electropolymerization cycles) at pH ∼ 7.5 was studied in 0.1 M phosphate buffer solutions by cyclic voltammetry measurements at varying scan rates from 5 to 50 mV/s (SI Appendix, Fig. S12). The onset potential was independent of the scan rate as expected (3, 17). The rate constant for electrocatalytic water oxidation (k_obs) was calculated as ∼0.12 ± 0.03 s⁻¹ at an applied potential of 1.6 V (E_app vs. NHE), based on Eqs. 1–3. In the equations (44, 45), i_cat is the catalytic current at 1.6 V (vs. NHE) with 4e⁻ involved in the catalytic water oxidation reaction (n_cat = 4), i_p is the anodic peak current for the [Ru^IV = O]²⁺/[Ru^III-OH]²⁺ couple with 1e⁻ transferred in the redox process (n_p = 1), F is the Faraday constant, N_cat is the mole number of electroactive catalyst sites in the surface-grafted polymer films during cyclic voltammetry scans, v is the scan rate, R is the ideal gas constant, and T is the temperature.

$$i_{cat}=n_{cat}F{N}_{cat}{k}_{obs}$$ [1]

$$i_p=\left(n_p^2{F}^{2}v{N}_{cat}\right)/\left(4RT\right)$$ [2]

$$i_{cat}/i_p=\left(4RT{n}_{cat}/n_p^{2}F\right)k_{obs}\left(1/v\right).$$ [3]

The k_obs value at pH ∼ 7.5 in 0.1 M aqueous phosphate buffer is an observed rate constant dependent on the applied potential, and is a lower limit for the catalytic rate constant (k_cat) describing the “ideal” catalysis in which the rate-limiting step is only the chemical process of O-O bonding formation for water oxidation catalysis (44, 46). In magnitude, k_obs ∼ 0.12 ± 0.03 s⁻¹ is enhanced compared with water oxidation by the analogous Ru(II) catalyst either at pH ∼ 1 in solution (E_app = 1.8 V vs. NHE) (18) or at pH ∼ 4.7 on surface (E_app = 1.7 V vs. NHE) (29), consistent with a rate enhancement by the APT pathway in the rate-limiting step to form [Ru^III-OOH]²⁺.

Controlled potential electrolysis of water oxidation by FTO|nanoITO|-g-poly1²⁺ (50 electropolymerization cycles) was conducted at pH ∼ 7.5 in a 0.1 M phosphate buffer by using a generator-collector dual-electrode technique (Fig. 6). The potential for the FTO|nanoITO|-g-poly1²⁺ generator electrode was held at 1.6 V (vs. NHE) with the produced O₂ diffusing to and reduced at the 1 mm-parallel FTO collector electrode held at −0.65 V (vs. NHE), illustrated in scheme in Fig. 6A (3, 17, 28). The current response at the generator electrode was monitored over an ∼3-h period, showing sustained catalytic water oxidation (solid red line in Fig. 6B). The initial current spike originates from local capacitance effects at the electrode (28) and the gradual current increase during the first ∼0.5 h is due to an equilibration between the local ionic environment and the external medium within the polymeric layers (33). With a clean FTO|nanoITO electrode as the control sample, O₂ produced at the generator electrode was detected in real time (solid blue line in Fig. 6B).

Fig. 6.

(A) Generator-collector configuration for O₂ production and analysis. (B) Anodic current-time traces at a FTO|nanoITO|-g-poly1²⁺ (50 electropolymerization cycles) generator electrode at 1.6 V (vs. NHE) for catalytic water oxidation (solid red line), and cathodic current-time traces at the FTO collector electrode at −0.65 V (vs. NHE) for simultaneous O₂ detection (solid blue line), relative to a FTO|nanoITO working electrode as the control sample to eliminate background current and small leaks in the N₂ atmosphere (dotted blue line). The electrolysis using the generator-collector configuration was carried out at pH ∼ 7.5 in a 0.1 M phosphate buffer under N₂ separated from a Ag/AgCl reference electrode and a Pt wire counter electrode by a Nafion membrane. The currents were normalized to the geometric areas of the electrodes.

In the electrolysis, the Faradaic efficiency (FE) for O₂ production was ∼68% as calculated by Eq. 4 (3, 17, 28). In Eq. 4, Q_Col is the integrated charge passed at the FTO collector electrode, Q_Gen is the integrated charge at the generator electrode, and η = 70% is the collection efficiency for the cell. The lower efficiency than 100% comes from the background currents during the electrolysis. Based on Eqs. 5 and 6 (17, 28), the catalyst underwent a turnover number (TON) of 210 O₂ molecules per catalyst site with a corresponding turnover frequency (TOF) of 0.021 turnovers per catalyst/s at 1.6 V (vs. NHE). In Eqs. 5 and 6, n_cat = 4 is the number of electrons involved in the water oxidation half reaction, F is the Faraday constant, N_cat is the moles of electroactive catalyst sites on the electrode surface, and t is the electrolysis time. The TOF obtained here was lower than k_obs obtained by the cyclic voltammetry analysis because the value for N_cat was based on the UV-vis measurements and not all of the catalyst molecules were in the electroactive form during the cyclic voltammetry measurements.

$$FE=Q_{Col}/\left(Q_{Gen}\eta \right)$$ [4]

$$TON=Q_{Col}/\left(n_{cat}F{N}_{cat}\eta \right)$$ [5]

$$TOF=Q_{Col}/\left(n_{cat}F{N}_{cat}t\eta \right).$$ [6]

After the electrolysis, the generator-collector dual electrode was suspended in a fresh 0.1 M phosphate buffer solution for another ∼3 h, which resulted in sustained catalytic water oxidation with continuous O₂ production and no obvious current decay reaching a FE of 71% (SI Appendix, Fig. S13).

Conclusions

Here, we have introduced an efficient surface binding technique for a single-site Ru(II) molecular catalyst for water oxidation. It is based on mesoporous FTO|nanoITO electrodes that combine silane surface attachment and reductive electropolymerization to stabilize a Ru(II) catalyst, 1²⁺, for sustained heterogenous electrocatalytic water oxidation. The silane functionalization introduced vinyl groups by robust -O-Si-O-Si-O- surface binding on the metal oxide followed by surface electropolymerization of 1²⁺, which was dominated by a grafting-through pathway. Surface coverage of 1²⁺ can be controlled by the number of the electropolymerization cycles allowing for the formation of multilayers of catalyst molecules. Surface-bound poly1²⁺ was stabilized under a variety of conditions especially at pH values above ∼6. The catalyst on the surface was stable toward redox cycling over an ∼4-h period at pH ∼ 7.5 in a 0.1 M phosphate buffer. Sustained electrocatalytic water oxidation catalysis occurred over a period of ∼6 h by using a generator-collector cell to simultaneously detect the produced O₂ with a FE of ∼68%.