Introduction

The development of efficient catalysts for the oxidation of water to molecular oxygen has long been the focus of intense research1,2,3,4,5,6,7,8,9,10,11,12. Such studies are motivated by the desire to understand the water splitting process in natural systems, such as photosystem II (PSII) of oxygenic photosynthesis, and artificial photosynthetic systems designed to produce hydrogen through proton reduction or convert carbon dioxide to fuels. In nature, water oxidation proceeds with extraordinarily high catalytic activity in PSII, in which a Ca-containing tetrameric manganese cluster (CaMn4O5) supported by bridged oxides or hydroxides and carboxylate and histidine side chains from the protein serves as the multi-electron oxidation catalyst13,14,15,16,17,18. Notably, all species capable of O2 evolution possess a qualitatively identical reaction centres, and no metal element other than Mn has been identified in the catalytic cluster of PSII. Therefore, extensive research efforts have been aimed at developing water-oxidation catalysts composed of the abundant element Mn19,20,21,22,23,24,25,26,27,28,29.

However, a remarkable contradiction still exists on the catalytic performance between naturally occurring and synthetic Mn catalysts, particularly under neutral pH conditions. Although bioinspired water-oxidation catalysts, particularly Mn oxides, function as effective electrocatalysts under alkaline conditions30,31,32, the activity of most Mn oxides is markedly reduced at neutral pH, resulting in a large electrochemical overpotential (η) ranging from 500 to 700 mV32,33,34,35. This high η contrasts that of the PSII tetrameric Mn cluster, which catalyzes water oxidation with an η of only 160 mV15,16,17. Recently, we have shown the primary origin for these sharp declines of catalytic potential of MnO2 under neutral conditions35,36. Measurements of in-situ water-oxidation current and optical absorption have shown that electron injection from H2O to anodically poised MnO2 forms Mn3+, which acts as a precursor for the O2-evolution reaction35,36. Our studies have also demonstrated that Mn3+ disproportionates to form Mn2+ and Mn4+ at pH<9, and subsequent regeneration via the electrooxidation of Mn2+ acts as the rate-determining step in the overall four-electron/four-proton reaction35,36. It is notable that the redox change of Mn from 2+ to 3+ on the surface of MnO2 is ~1.4 V at pH<9, which forces the onset potential for water-oxidation current (Uon,j) to remain constant at ~1.5 V irrespective of the pH. The involvement of the pH-independent step in the redox change of Mn increases η at intermediate pH, as shown in Fig. 1a, and is the primary origin for the sharp decline of catalytic potential of MnO2 under neutral conditions. This property prohibits the successful application of Mn oxides as components of artificial photosynthetic systems.

Figure 1: pH-dependent water-oxidation mechanisms by MnO 2 electrodes.
figure 1

(a) Schematic illustration of the pH dependence of the onset potential (Uon,j) for water oxidation35. Under neutral conditions, Uon,j does not show pH dependence, whereas it exhibits a linear pH dependence at pHs above 9. The transition point from pH-independent to pH-dependent at pH~9 corresponds to the pKa of Mn2+OH2. (b) Thermodynamic cycle for the electrooxidation of Mn2+ to Mn3+ on the surface of MnO2 electrodes. The pKa values of the water ligand bound to Mn2+ and Mn3+ are 10.6 and 0.7, respectively. Thus, Mn2+ and Mn3+ exist as protonated and deprotonated forms, respectively, at intermediate pH. B and BH+ represent the deprotonated and protonated forms, respectively, of a proton acceptor.

The new model explained above has suggested that the efficiency of Mn oxides as O2-evolution catalysts may be enhanced by regulating proton-coupled electron transfer (PCET) in the electrooxidation step of Mn2+ to Mn3+. PCET is a key part of the efficient energy conversion by PSII37,38,39,40. In particular, the concerted proton-electron transfer (CPET) pathway, which involves the transfer of protons and electrons in a single concerted step, avoids charge build-up and thereby enables the redox change of the tetrameric Mn cluster in PSII over a narrow potential range (~250 mV) (refs 13, 14, 15). Therefore, it can be deduced that synthetic Mn oxides inefficiently catalyze water oxidation at neutral pH because they lack the inherent ability to manage both protons and electrons. Herein, we attempt to induce CPET on the surface of a Mn-oxide electrocatalyst and demonstrate that this rationale-based strategy improves the O2-evolution activity of synthetic Mn oxide in neutral medium.

Results

CPET induction by the addition of pyridine

One possible way to induce CPET is the introduction of a protonation site, which has a pKa value that is intermediate between the pKa of Mn3+-OH2 (0.7) and Mn2+-OH2 (10.6)41, as predicted by the libido rule of general acid-base catalysis42,43,44, in close proximity to the water oxidation active site. In the course of oxidation from Mn2+ to Mn3+, the pKa of the Mn center undergoes a large change that would convert an unfavourable proton-transfer reaction to a favourable one with respect to the base40,41,42,43,44,45,46,47. Therefore, the regeneration of Mn3+ at neutral pH would be converted from a stepwise to a concerted pathway if the base accepts the proton at the PCET transition state rather than directly from Mn2+-OH2, as illustrated in Fig. 1b. In support of this speculation, recent density functional theory calculations have shown that the pKa values of terminal water ligands in biomimetic oxomanganese complexes are drastically changed upon the oxidation of Mn2+ to Mn3+ (refs 41, 48). In addition, the agreement of the pKa value of Mn2+-OH2 with the transition point between pH 9–10 for the pH-independent and -dependent activity of the electrooxidation reaction35 (Fig. 1a) also provides the support for this hypothesis. Namely, at a pH greater than the pKa of Mn2+-OH2, the electrooxidation of Mn2+ proceeds after the water ligand bound to Mn2+ deprotonates to form Mn2+-OH (path c in Fig. 1b), whereas Mn2+ oxidation prior to deprotonation produces the unstable intermediate Mn3+-OH2 (path a in Fig. 1b), leading to the electrooxidation of Mn2+-OH2, which is more thermodynamically unfavourable.

To examine the validity of our hypothesis, the effects of pyridine on the electrocatalytic performance of Mn oxides for water oxidation and its influence on the pH dependence of Uon,j were first investigated. Pyridine has a pKa (5.25) that is intermediate between that of Mn3+-OH2 (0.7) and Mn2+-OH2 (10.6), and exists as a deprotonated form at neutral pH. Thus, pyridine satisfies the conditions of the libido rule42. The current density (j) versus potential (U) curves for a synthesized MnO2 electrocatalyst, consisting of α-MnO2 nanoparticles deposited onto a fluorine-doped tin oxide (FTO) electrode, measured electrochemical cells operated at neutral pH with and without pyridine are shown in Fig. 2. The amount of dissolved O2 in the electrolyte was also monitored using a needle-type oxygen microsensor. We observed a large increase in both anodic current (Fig. 2a) and O2 production (Fig. 2b) for the MnO2 electrode in the presence of pyridine. The effects of pyridine were more prominent at higher concentrations, with ~200-mV negative shift in Uon,j being observed at 0.5 M pyridine. Analysis of the head-space gas in the electrochemical cell containing 50 mM pyridine by gas chromatography (GC) revealed that a turnover number (evolved O2 molecules per total Mn atoms deposited on the FTO electrode) of 17 and coulombic efficiency of 78% was reached after 90 min of electrolysis at 1.7 V, confirming that the evolved O2 was generated from water oxidation. When electrolysis was conducted for 90 min at 1.39 V in the presence of pyridine (2.5 mM), 86 nmol of CO2 was detected. This amount corresponds to the 4% of a coulombic efficiency if we assume the complete oxidation of pyridine to CO2 (Supplementary Figs 1 and 2).

Figure 2: Enhanced water oxidation by addition of pyridine.
figure 2

Enhanced water-oxidation activity of MnO2 electrodes at pH 7.5 by the addition of pyridine. (a) Potential dependence of current density for MnO2 electrodes in presence and absence of pyridine. (b) Potential dependence of dissolved O2 concentration for MnO2 electrodes in presence and absence of pyridine. For both (a and b) the scan rate was 10 mV s−1.

We next examined the pH dependence of Uon,j for water-oxidation current to confirm whether CPET was induced on the surface of a Mn-oxide electrocatalyst as predicted. The electrode potential at which the current density reached 130 μA cm−2 was adopted as Uon,j. In the absence of pyridine, no pH dependence of Uon,j was observed at pH<10 (Supplementary Fig. 3a), whereas Uon,j showed a clear pH dependence following the addition of pyridine into the electrochemical cell (Supplementary Fig. 3b). Plots of Uon,j against electrolyte pH showed a pH dependence in the presence of pyridine (Supplementary Fig. 3c,d), demonstrating that proton transfer was involved in the rate-determining transition state for water oxidation. Notably, pyridine did not enhance water-oxidation activity at pHs higher than the pKa of Mn2+-OH2 (Supplementary Fig. 3d), a result that is consistent with the deprotonation of the water ligand bound to Mn2+ forming Mn2+-OH at pH>10 (path c in Fig. 1b). Similarly, when pyridine was added to the electrolyte at pH 4, almost no activity enhancement was observed, as 95% of pyridine exists as the protonated form at this pH and therefore cannot function as a proton acceptor.

The results presented in Fig. 2 and Supplementary Fig. 3 are consistent with the hypothesis that pyridine accepts a proton from the water ligand bound to Mn during the PCET transition state. However, when the jU measurements were conducted in D2O instead of H2O in the presence of 50 mM pyridine at neutral pD (pD=pH+0.4) (ref. 49), the kinetic isotope effect (KIE) measured at 1.37 V for j(H2O) and j(D2O) was 1.62, which is only slightly higher than the KIE obtained in the absence of pyridine (KIE=1.48). These results suggest that pyridine acts as proton acceptor in this system; however, the basicity of pyridine was not strong enough to induce the fully concerted reaction path. In other words, we consider that the electron-transfer reaction is more rapid than the transfer of the proton from the Mn-bound water ligand to pyridine and speculate that there are more suitable bases for the enhancement of catalytic activity.

Regulation of PCET mechanisms by the addition of different bases

In an attempt to accelerate the proton-transfer reaction, we replaced pyridine with a base that has a higher pKa value. As pKa is the index for proton-accepting ability, the addition of a base with a higher pKa was expected to promote the deprotonation of water ligand during the PCET transition state. Several pyridine derivatives, namely 3-methylpyridine (β-picoline, pKa=5.80), 4-methylpyridine (γ-picoline, 6.10), 2,6-dimethylpyridine (2,6-lutidine, 6.96), and 2,4,6-trimethylpyridine (γ-collidine, 7.48) were evaluated as candidate bases. Similar to pyridine, the pKa values of the selected derivatives are lower than that of Mn2+-OH2, meaning that the direct deprotonation of water ligands (path c in Fig. 1) is thermodynamically unfavourable.

The jU curves for the MnO2 electrode in the presence of pyridine derivatives at pH 7.5 are shown in Fig. 3a. As expected, a higher catalytic activity was observed with increasing pKa of the added base. The η decreased with the increase of pKa; for example, ~70-mV negative shift of Uon,j was observed for γ-collidine relative to that for pyridine. To compensate for the concentration difference between the protonated and deprotonated forms of bases due to pKa differences, the electrolyte was prepared using a 25 mM concentration of the deprotonated base form. Therefore, the shift of Uon,j that can be seen in Fig. 3a is attributable to the different proton-accepting abilities of the pyridine derivatives. GC analysis of the head-space gas revealed that the coulombic efficiency for O2 evolution was 77% (γ-collidine), 62% (2,6-lutidine), 66% (γ-picoline) and 57% (β-picoline), and that the amount of evolved O2 exceeded unity for all examined bases. Notably, the rate of O2 production increased monotonically with increasing of pKa, as shown in Fig. 3b, a result that is consistent with the observed trends in the jU curve measurements. In the case of γ-collidine, the O2-evolution activity of the MnO2 electrode at pH 7.5 showed ~15fold enhancement compared with that observed in the absence of bases, reaching nearly half of the O2-evolution activity that was measured in alkaline medium (Supplementary Fig. 4).

Figure 3: Effects of base p Ka.
figure 3

Effects of base pKa on electrochemical water-oxidation activity of MnO2 electrodes at pH 7.5. (a) Potential dependence of current density for MnO2 electrodes in the presence of pyridine (pKa=5.25; light green), β-picoline (pKa=5.80; olive green), γ-picoline (pKa=6.10; orange), 2,6-lutidine (pKa=6.96; blue) and γ-collidine (pKa=7.48; brown). A MnO2 electrode without the addition of bases to the electrolyte is also depicted as a reference (black). (b) Time course of O2 production observed for MnO2 electrodes at an applied potential of +1.39 V versus standard hydrogen electrode (SHE) in the presence of the bases listed in a. The dashed line indicates the value corresponding to turnover number (TON)=1. The concentration of each base was adjusted to 25 mM of the deprotonated form. A MnO2 electrode without the addition of bases to the electrolyte is also depicted as a reference.

The KIE values obtained from jU curve measurements for the MnO2 electrode at neutral pH (and adjusted pD=pH+0.4; ref. 49) were plotted as a function of base pKa (Fig. 4a). The concentration of the deprotonated form of bases was adjusted to 25 mM. It can be seen that the KIEs increased monotonically from 1.6 to 2.7 as the base pKa increased from 5.25 to 7.48. The KIE is a rough measure of the position of the transition state along the reaction coordinate for proton transfer50. Large KIEs are expected for a reaction with nearly symmetrical transition states, as the activated complex has the proton almost equally shared between the reactant and product. Meanwhile, smaller effects are observed as the transition state moves towards reactant, as very little deprotonation has occurred at transition state. Therefore, the increase of KIEs with the base pKa demonstrates that bases with a stronger proton extracting property increase the extent of deprotonation of the water ligand bound to a Mn center at the rate-determining transition state. Moreover, the pH dependence of Uon,j at pH<9 became more prominent with the increasing pKa of the base (Fig. 4b and Supplementary Fig. 5). For example, the shift per pH unit increased from ~25 mV for pyridine to 60 mV for γ-collidine (Supplementary Fig. 6). The latter value indicates that the η for water-oxidation reaction is almost constant irrespective of a solution pH higher than pKa of γ-collidine (Supplementary Fig. 5). Thus, the observed increase in KIE and the slope of pH dependence with base pKa (Fig. 4), together with the concomitant enhancement of the water-oxidation activity of the MnO2 electrode (Fig. 3), demonstrate the successful induction of CPET in the electrocatalytic water oxidation by MnO2.

Figure 4: Regulation of PCET mechanisms by the addition of bases.
figure 4

(a) Plot of KIE as a function of the pKa of the indicated bases at pH (pD) 7.5 at 1.29 V versus SHE. The measured pH value in D2O was converted to pD value using the equation—pD=measured pH+0.4. Base concentration: 25 mM of deprotonated form. (b) Plot of pH versus potential showing the pH dependence of the onset potential (Uon,j) defined at 130 μA cm−2 for water oxidation in the presence of the indicated bases at pHs ranging from 5 to 9.

Discussion

We have demonstrated for the first time the regulation of PCET mechanisms involved in water oxidation by Mn oxides. Because the oxidation of Mn from 2+ to 3+ is associated with a large change in pKa, the potential of pyridine and its derivatives as a CPET-induction reagents could be evaluated based on the libido rule of general acid-base catalysis. The induction of the CPET reaction by the addition of γ-collidine was confirmed by the clear pH dependence of Uon,j, which was shifted by ~60 mV, and the pronounced KIE. While the detection of KIE and compliance of the reactions with the libido rule indicated the successful regulation of PCET mechanisms, the shift of Uon,j with pH in the presence of base is not explained directly by a simple model where the base is the proton acceptor. As the reaction rate of CPET is a function of the concentration of the proton acceptor, the rate is expected to be pH-independent when the concentration of deprotonated forms of base is constant. In this point of view, it is reasonably considered that the pH-dependent rate in the presence of base at the pH higher than base pKa observed in Fig. 4b could arise from the change of the rate constant for CPET. Yet it should also be noted that the base has the possibility to affect several elementary steps other than the oxidation of Mn from 2+ to 3+ in a course of electrocatalytic water oxidation.

The regulation of PCET mechanisms is essential for the redox state change of the tetrameric Mn cluster in the Kok cycle of PSII and is likely promoted by amino acids whose pKa varies according to the redox state of the Mn cluster37,38,39. However, most bioinspired water-oxidation catalysts, particularly Mn oxides, lack such a specific regulatory mechanism, as demonstrated by the sharp decline of O2-evolution activity under neutral conditions for most Mn oxides reported to date30,31,32,33,34,35. Therefore, we consider that the regulation of PCET, as described in the present study, is an effective approach to lower the η for water oxidation by Mn oxides in neutral media designed for artificial photosynthesis, and will lead to new understanding for the marked difference in catalytic performance between naturally occurring and synthetic Mn catalysts at physiological pH.

Methods

Synthesis procedure of α-MnO2

α-MnO2 powder was prepared as previously described51. Briefly, 1.10 g potassium permanganate (KMnO4) was dissolved in 20 ml double-distilled water and the resulting mixture was stirred for 30 min at 60 °C. Separately, 1.55 g manganese (II) chloride tetrahydrate (MnCl2·4 H2O) was added to 25 ml of a 2 M acetic acid aqueous solution, which was then stirred for 20 min at room temperature. The MnCl2 solution was added to the KMnO4 solution, and the resulting mixture was stirred for 2 h at 80 °C. The formed particles were collected and washed several times with doubled-distilled water. The samples were dried overnight at 60 °C and then used for preparing electrodes. All chemical reagents were obtained from Wako and utilized without further purification.

Electrode preparation

Particulate α-MnO2 film electrodes were prepared using a spray deposition method, as previously described35. Briefly, 75 mg of the prepared MnO2 powder sample was suspended in 20 ml water by sonication at an amplitude of 15 W for 30 min with a homogenizer (S4000, Qsonica). The suspension was diluted to 100 ml with water and was then sprayed onto a clean conducting glass substrate (FTO-coated glass, resistance: 20 Ω per square, size: 30 mm × 30 mm; SPD Laboratory, Inc.) at 200 °C using an automatic spray gun (Lumina, ST-6). After coating, the electrodes were washed thoroughly with distilled water. The amount of deposited MnO2 was ~0.14 mg cm−2.

Electrochemical water oxidation

Current density (j) versus potential (U) curves were obtained with a commercial potentiostat and potential programmer (HZ-5000, Hokuto Denko), using a Pt wire as the counter electrode and an Ag/AgCl/KCl (saturated) electrode as the reference electrode. The electrolyte solution (0.5 M Na2SO4) was prepared using highly pure Milli-Q water (18 MΩ−1cm−1) and reagent-grade chemicals, and the pH was adjusted using 1.0 M H2SO4 and 1.0 M NaOH35. No agent for pH buffering was added to the electrolyte solution to avoid influences from the specific adsorption of multivalent anions. The prepared electrolyte solution was bubbled with Ar gas prior to measurements and temperature of the reactor was kept at 30 °C unless otherwise noted. For minimizing pH changes near the electrode surface, the j versus U curve was measured using a potential sweep from negative to positive without stirring. The amount of dissolved oxygen in the electrolyte solution was monitored simultaneously during the j versus U measurements using a needle-type oxygen microsensor (Microx TX3-trace, PreSens). A gas chromatograph equipped with a TCD detector (GC-8A, Shimadzu) was used to monitor the amount of oxygen in the head-space of electrochemical reactors. Pyridine and its derivatives were obtained from Wako, and isotopic reagents (D2O and D2SO4 for pD adjustment) were purchased from Sigma-Aldrich.

Additional information

How to cite this article: Yamaguchi, A. et al. Regulating proton-coupled electron transfer for efficient water splitting by manganese oxides at neutral pH. Nat. Commun. 5:4256 doi: 10.1038/ncomms5256 (2014).