Tuning Redox Active Polyoxometalates for Efficient Electron‐Coupled Proton‐Buffer‐Mediated Water Splitting

Abstract

We present strategies to tune the redox properties of polyoxometalate clusters to enhance the electron‐coupled proton‐buffer‐mediated water splitting process, in which the evolution of hydrogen and oxygen can occur in different forms and is separated in time and space. By substituting the heteroatom template in the Keggin‐type polyoxometalate cluster, H₆ZnW₁₂O₄₀, it is possible to double the number of electrons and protonation in the redox reactions (from two to four). This increase can be achieved with better matching of the energy levels as indicated by the redox potentials, compared to the ones of well‐studied H₃PW₁₂O₄₀ and H₄SiW₁₂O₄₀. This means that H₆ZnW₁₂O₄₀ can act as a high‐performance redox mediator in an electrolytic cell for the on‐demand generation of hydrogen with a high decoupling efficiency of 95.5 % and an electrochemical energy efficiency of 83.3 %. Furthermore, the H₆ZnW₁₂O₄₀ cluster also exhibits an excellent cycling behaviour and redox reversibility with almost 100 % H₂‐mediated capacity retention during 200 cycles and a high coulombic efficiency >92 % each cycle at 30 mA cm⁻².

The molecular structure of keggin‐type tungsten polyoxometalates (POMs) were tuned with different heteroatoms to modify their redox properties. Compared to the two electrons of well‐studied H₃PW₁₂O₄₀ and H₄SiW₁₂O₄₀, H₆ZnW₁₂O₄₀ exhibits double reversible electrons with protonation in the redox reactions and the more desirable redox potentials, which enhance its ECPB performance (see scheme).

Water splitting is a promising process that could provide a route to a clean and inexhaustible renewable energy, but significant issues remain for establishing scalable systems and also coping with intermittent power.1 In this regard proton exchange membrane electrolysis (PEME) is a prime candidate to achieve this transformation but gas crossover through the membrane is both a safety concern and affects the purity of H₂, as well as degrading the membrane.2, 3 Thus, it is highly important to develop new strategies to coordinate the intermittent renewable power sources that meet the requirements of H₂ large‐scale production, transportation and storage simultaneously with safe and economic approaches.

Previously, we introduced the electron‐coupled‐proton‐buffer (ECPB) to separate the process of water splitting in time and space.4, 5, 6, 7 In this way, the hydrogen evolution reaction (HER) is no longer coupled with the rate‐limiting step of the oxygen evolution reaction (OER). Electric energy can be stored in a soluble H₂‐mediated redox couple, which can be chemically or electrochemically converted into gaseous H₂ on‐demand. Such a system is ideal as a flexible and scalable option to mitigate the intermittent output of renewably generated energy. It will also prevent the crossover issue of O₂/H₂ gas through the membrane. Similarly, Xia et al. decoupled the hydrogen and oxygen production system in a membrane‐free electrolyser based on the reversible solid‐state Ni(OH)₂ and polytriphenylamine redox mediator.8, 9 Integrated battery‐electrolysers have also been constructed to decouple hydrogen production driven by the electrochemical energy storage system, such as an all‐vanadium dual circuit redox flow battery or a nickel‐iron battery.10, 11 Furthermore, such hydrogen‐mediated redox species also show potential applications in high‐purity H₂ storage and generation on demand. Hydrogen could be spontaneously evolved from the reduced‐ECPB on a conventional HER‐catalyst if its redox potential is more negative than the HER onset potential of the catalyst.12, 13 By exploiting the overpotentials related to hydrogen evolution on carbon, an ECPB can be reduced past the point of the normal hydrogen electrode (NHE) without any H₂ evolution. Once the reduced ECPB encounters the HER catalyst, it can then be oxidized with concomitant high‐purity H₂ release—that is, without any additional energy input.

Soluble redox mediators with the ability to buffer protons are the cornerstone of the promising applications described above, but the electron‐storage capacity of the mediators considered hitherto are limited to only 1–2 electrons per molecule. Thus, there is a great need to develop new redox mediators that can store as many electrons per molecule as possible. Polyoxometalates (POMs) show tremendous promise in this regard, due to their ability to perform reversible multielectron reactions with high structural stability in aqueous media.^14] Despite many successful demonstrations of polyoxometalates in solid‐state energy storage systems15, 16, 17, 18, 19, 20, 21 and redox flow batteries,22, 23 it is still difficult to develop strategies to modify the POMs’ structure to tune the number of reversible electrons in aqueous states, and thus to increase the volumetric capacities for the applications in ECPB water splitting and hydrogen‐mediated redox couples for aqueous energy storage. Moreover, the energetic inputs required for ECPB‐based water splitting depends on the redox species’ electrochemical potential. For example, silicotungstic acid, can undergo a 2‐electron redox reaction, but only one of the two electrons can be decoupled to release H₂ spontaneously by exposure over Pt/C catalysts, leading to a low H₂ decoupling efficiency around 67 % and an electrochemical energy efficiency of 79.3 %.7 Thus, it is important to increase the electron storage capacity and to modify the redox potential to improve energy efficiency in the overall water splitting.

Herein, we successfully illustrate the strategies to tune redox properties of Keggin‐type tungsten POMs with different central heteroatoms. The number of reversible electrons with protonation could be doubled by changing the heteroatom from P⁵⁺or Si⁴⁺ to Zn²⁺. The desired redox potentials of H₆ZnW₁₂O₄₀ enhance its ECPB performance for on‐demand storage and generation of H₂ with a high decoupling efficiency of 95.5 % and an electrochemical energy efficiency of 83.3 %.

The molecular structure in this work is based on the typical Keggin‐type heteropolyoxotungstates {XW₁₂O₄₀}. As shown in the CV results in Figure 1 a, {XW₁₂O₄₀} exhibits distinguished differences of redox chemistry by changing their central heteroatoms (X=P⁵⁺, Si⁴⁺, and Zn²⁺) within the voltage range of HER onset potentials between glassy carbon electrode and Pt electrode. This allows {XW₁₂O₄₀} to be reduced firstly on carbon without any competing H₂ evolution and then evolve H₂ spontaneously through simple exposure to Pt later. H₃PW₁₂O₄₀ undergoes two redox steps at the potentials +0.222 and −0.052 V vs. NHE. While the redox potentials of H₄SiW₁₂O₄₀ shift negatively to +0.019 and −0.212 V vs. NHE compared to those of H₃PW₁₂O₄₀. These two separated redox peaks in both H₃PW₁₂O₄₀ and H₄SiW₁₂O₄₀ have been confirmed to be two one‐electron redox processes.24 Intriguingly, due to the prominent effect of ionic charge on the pristine anion clusters, the midpoint potential of the first redox wave shift negatively when the central heteroatom changes from P⁵⁺ or Si⁴⁺ to Zn²⁺. H₆ZnW₁₂O₄₀ (prepared by following the modified literature method,25 see the Supporting Information, Section SI‐2 and Figures S1 and S2) exhibits the redox potentials of −0.078 and −0.198 V vs. NHE under the same conditions. This results are consistent with previous studies of the energy levels.26, 27, 28, 29, 30 Moreover, the number of electrons stored by H₆ZnW₁₂O₄₀ at each redox potential was increased to 2, respectively.

Figure 1.

a) Crystal structure of {XW₁₂O₄₀} (phosphorous, grey; silicon, orange; zinc, blue; tungsten, cerulean; oxygen, red) and the corresponding cyclic voltammograms (CVs) of 100 mm H_nXW₁₂O₄₀ (X=P⁵⁺, Si⁴⁺, and Zn²⁺) solution on glassy carbon electrode at a scan rate of 50 mV s⁻¹. b) pH changing during the reduction process of 100 mm H_nXW₁₂O₄₀ (X=P⁵⁺, Si⁴⁺, and Zn²⁺) solution. c) CVs of a 1 mm H_nXW₁₂O₄₀ (X=P⁵⁺, Si⁴⁺, and Zn²⁺) under different pH buffer solutions at a scan rate of 50 mV s^−1. For more details, see the Supporting Information, Sections SI‐3, SI‐4, and SI‐5.

Changing the heteroatom type inside the {XW₁₂O₄₀} also leads to a significant difference in ability of the cluster to buffer under reducing conditions. Bulk electrolysis in an airtight H‐cell was performed to electrochemically reduce {XW₁₂O₄₀} with different equivalents of electrons per cluster (see the details in the Supporting Information, Section SI‐5). As shown in Figure 1 b, a continuous decrease of pH could be observed when 100 mm H₃PW₁₂O₄₀ and H₄SiW₁₂O₄₀ solution are reduced to two electrons per cluster, while 100 mm H₆ZnW₁₂O₄₀ solution has a negligible pH change up to four electron reductions per cluster. This can be ascribed to the enhanced protonation ability of {ZnW₁₂O₄₀} that accompany reduction in acid solution, keeping the overall ionic charge of the reduced Keggin {XW₁₂O₄₀} cluster constant at −6.31, 32, 33 This is also confirmed by the CV investigation under different pH values. As shown in Figure 1 c, the redox potential of H₆ZnW₁₂O₄₀ has a dependence on the pH of the solution, which means that protons are involved in the electrochemical redox reaction of H₆ZnW₁₂O₄₀ according to the Nernst equation. However, the redox potential of H₃PW₁₂O₄₀ and H₄SiW₁₂O₄₀ has no dependence on pH. The CVs of H₃PW₁₂O₄₀ at pH 2.0 and 3.0 cannot be obtained because of its instability in the solution with pH>1.5.32

Four‐electron redox reactions of H₆ZnW₁₂O₄₀ were further proved by passing charge equivalents to different numbers of electrons per cluster at a current density of 5 mA cm⁻² and measuring the coulombic efficiency between the reduction and reoxidation process. As shown in Figure 2, the coulombic efficiency remains above 94 % even up to a reduction level of four electrons per cluster molecule. But more than four electrons, reduction of H₆ZnW₁₂O₄₀ produces a brown solution, which indicates that some irreversible reactions have occurred (Figure S5, Supporting Information). XPS of the W in the {ZnW₁₂O₄₀} cluster under the various reduced states are displayed in Figure S6 (see more details of sample preparation in the Supportin Information, Section SI‐7). The peaks at 34.1 and 36.28 eV are assigned to W^V 4f_7/2 and W^Ⅴ 4f_5/2, whereas the peaks in 35.39 and 37.57 eV belong to W^Ⅵ 4f_7/2 and W^Ⅵ 4f_5/2.34 The content of W^Ⅴ increases obviously with the number of reduction electrons, while the peaks of Zn 2p_3/2 and Zn 2p_1/2 at 1021.82 and 1044.88 eV, respectively remain unchanged (Supporting Information, Figure S7), indicating that the negative charges are mainly localized around tungsten sites. Thus, taken all together, H₆ZnW₁₂O₄₀ could exhibit a four‐electron redox reaction with accompanying protons and the detailed electrochemical reactions that can occur are presented in the Supporting Information (Section SI‐6). Moreover, the modulated redox potentials make it suitable as an ECPB for high‐performance water splitting.

Figure 2.

Reduction and reoxidation of a 20 mm H₆ZnW₁₂O₄₀ solution with a constant current density of 5 mA cm⁻² with different equivalents of electrons per cluster.

Regarding the hypothesis of application for on‐demand H₂ storage and generation, the spontaneous H₂ evolution experiments were carried out to test the hydrogen‐mediated capacity of H₆ZnW₁₂O₄₀. The schematic of this ECPB‐mediated water‐splitting system was illustrated in Figure 3 a. At the anode, H₂O is split into O₂, protons, and electrons, whereas the mediator is reversibly reduced and protonated at the cathode in preference to direct production of H₂. In other words, the electrons were stored in the H₆ZnW₁₂O₄₀ by means of electron‐coupled protons rather than direct H₂ evolution at the cathode. The reduced ECPB is then transferred to a separate chamber for H₂ evolution over a suitable catalyst. According to their CVs in Figure 1 a, H₆ZnW₁₂O₄₀ will be reduced by four electrons and then couple with four protons at the constant potential of −0.36 V vs. NHE. Silicotungstic acid was used as the control experiment. The theoretical H₂ volume stored in the 2 e⁻‐reduced H₄SiW₁₂O₄₀ or 4 e⁻‐reduced H₆ZnW₁₂O₄₀ solution could be calculated to 44.8 and 89.6 mL by the charge passed in the first‐step electroreduction process (the calculated equation is given in the Supporting Information, Section SI‐9). As shown in Figure 3 b, about 30 mL H₂, 67 % of the theoretical H₂ volume, can be collected by physically mixing the 2 e⁻‐reduced H₄SiW₁₂O₄₀ solution with 5 mg 20 % Pt/C, which is consistent with previous reported result.7 The incomplete release of hydrogen stored is caused by the first redox potential of H₄SiW₁₂O₄₀ being more positive than the HER onset potential of Pt/C catalyst (Figure 1 a). Nevertheless, 95.5 % of the theoretical H₂ volume stored could be released in the H₆ZnW₁₂O₄₀‐mediated system (see movie S1) and the H₂ initial decoupling rate could reach a value of 121 mmol h⁻¹ mg⁻¹ Pt, which is much higher than that of the 2 e⁻‐reduced H₄SiW₁₂O₄₀. The filtered H₆ZnW₁₂O₄₀ solution after hydrogen release shows the same UV/Vis spectrum as the initial solution (Supporting Information, Figure S9), which indicates the structural stability of the H₆ZnW₁₂O₄₀ mediator. Moreover, the H₆ZnW₁₂O₄₀‐mediated system has a higher electrochemical energy efficiency of 83.3 % than the H₄SiW₁₂O₄₀‐mediated system due to its modulated redox potentials and redox electron number (for the detailed calculation process see the Supporting Information, Section SI‐8).

Figure 3.

a) Schematic of the ECPB‐mediated water‐splitting system and the spontaneous hydrogen evolution system. b) Spontaneous H₂ evolution from a 20 mL sample of 100 mm 4 e⁻‐reduced H₆ZnW₁₂O₄₀ or 2 e⁻‐reduced H₄SiW₁₂O₄₀ over 5 mg 20 % Pt/C.

Another potential application of this ECPB in the production of H₂ and O₂ at separated time and space were evaluated by the flow cells system (Figure 4 a). In this system, water is oxidized to produce O₂ on the anode in cell‐a with an iridium oxide catalyst. The concomitant electrons and protons would be applied to reduce and protonate the H₆ZnW₁₂O₄₀ solution on a carbon cathode (12.96 cm² of geometric area) in cell‐a. Once a full 4‐electron reduction per H₆ZnW₁₂O₄₀ cluster has been reached by passing corresponding amount of charge, the 4 e⁻‐reduced H₆ZnW₁₂O₄₀ solution would be then pumped to a carbon anode in cell‐b to be reoxidized electrochemically and H₂ is produced on the cathode in cell‐b with a Pt/C catalyst at the same time. The charge passed in the reoxidation process is denoted as practical ECPB storage capacity and the coulombic efficiency could be gauged by comparing it with the charge initially used to reduce H₆ZnW₁₂O₄₀ solution. As a result, H₆ZnW₁₂O₄₀ exhibits excellent ECPB cycling behaviors and redox reversibility. Under a constant electrolysis current density of 30 mA cm⁻², it delivers an almost 100 % H₂‐mediated capacity retention during 200 cycles with a high coulombic efficiency >92 % each cycle (Figure 4 b).

Figure 4.

a) Schematic of the flow cells system for long‐term separated time and space water electrolysis. b) Long‐term electrochemical reduction and re‐oxidation cycling of a 20 mm H₆ZnW₁₂O₄₀ solution at a current density of 30 mA cm⁻² under Ar atmosphere.

In summary, this work shows the strategies to tune the properties of H₂‐mediated redox couples by the adjustment of Keggin‐type tungsten polyoxometalates’ molecular structure with different heteroatoms. When the central heteroatom changes from P⁵⁺or Si⁴⁺ to Zn²⁺, H₆ZnW₁₂O₄₀ exhibits an increased number of reversible electrons with enhanced protonation ability and more favorable redox potentials. As an ECPB for H₂ storage and generation, it also delivers a high decoupling efficiency of 95.5 % and an electrochemical energy efficiency of 83.3 % for on‐demand catalytic H₂ evolution and an excellent long‐term time and space separated, water electrolysis. This work illustrates how to optimize practical application issues by using fundamental approaches, and we believe this will lead to new flexible and safe H₂ production, storage and transportation systems to mitigate the challenges inherent in present renewable energy systems.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

Supplementary

J. Lei, J.-J. Yang, T. Liu, R.-M. Yuan, D.-R. Deng, M.-S. Zheng, J.-J. Chen, L. Cronin, Q.-F. Dong, Chem. Eur. J. 2019, 25, 11432.