Tuning the Catalytic Water Oxidation Activity through Structural Modifications of High-Nuclearity Mn-oxo Clusters [Mn18M] (M = Sr2+, Mn2+)

The water oxidation half-reaction is considered the bottleneck in the development of technological advances to replace fossil fuels with sustainable and economically affordable energy sources. In natural photosynthesis, water oxidation occurs in the oxygen evolving complex (OEC), a manganese-oxo cluster {Mn4CaO5} with a cubane-like topology that is embedded within a redoxactive protein environment located in photosystem II (PS II). Therefore, the preparation of biomimetic manganese-based compounds is appealing for the development of efficient and inexpensive water oxidation catalysts. Here, we present the water oxidation catalytic activity of a high-nuclearity mixedmetal manganese-strontium cluster, [Mn12Mn6Sr(μ4-O8)(μ3-Cl)8(HL)12(MeCN)6]Cl2·15MeOH (Mn18Sr) (HLMe = 2,6-bis(hydroxymethyl)-p-cresol), in neutral media. This biomimetic mixedvalence cluster features different cubane-like motifs and it is stabilized by redox-active, quinone-like organic ligands. The complex displays a low onset overpotential of 192 mV and overpotentials of 284 and 550 mV at current densities of 1 mA cm−2 and 10 mA cm−2, respectively. Direct O2 evolution measurements under visible light-driven water oxidation conditions demonstrate the catalytic capabilities of this cluster, which exhibits a turnover frequency of 0.48 s−1 and a turnover number of 21.6. This result allows for a direct comparison to be made with the structurally analogous Mn-oxo cluster [Mn12Mn7(μ4-O)8(μ3-OCH3)2(μ3-Br)6(HL)12(MeOH)5(MeCN)]Br2·9MeCN·MeOH (Mn19), the water oxidation catalytic activity of which was recently reported by us. This work highlights the potential of this series of compounds towards the water oxidation reaction and their amenability to induce structural changes that modify their reactivity.


Introduction
Water oxidation catalysis has become a central focus area in the quest for replacing fossil fuel combustion for energy production by renewable and environmentally friendly energy sources [1,2]. Due to its ubiquity, water represents the ideal source of reducing equivalents to produce H 2 via the water splitting reaction, hence conceptually mimicking natural photosynthesis. This process provides the storage of clean energy in the form of chemical bonds with an intrinsic high gravimetric energy density [3][4][5]. However, the water oxidation half-reaction (Equation (1)) within this scheme is complex and highly energetically demanding. Therefore, it requires the use of water oxidation catalysts (WOCs) that alleviate the energy requirements associated with each of the steps in the reaction mechanism and accelerate the sluggish kinetics of the O−O bond formation [6][7][8]. Indeed, the lack of cost-effective WOCs hampers the generation of economically affordable H 2 that could compete with fossil fuels for energy production [5,9]. In this matter, Ruand Ir-based materials are highly efficient catalysts; however, their large-scale application is unfeasible due to the scarcity and prohibitive cost of these metals [10,11]. Therefore, earth-abundant-based materials with water oxidation catalytic capabilities have been explored in recent years [12,13]. A myriad of systems have been reported displaying diverse water oxidation activities, which comprise mono-or multinuclear organometallic molecular species [14,15], polyoxometalates [16][17][18][19], colloidal particles [20,21], homo-and heterometallic oxides [22][23][24], metal-organic frameworks [25,26] and even metal-free catalysts [27,28]. Moreover, heterogenization of molecular WOCs through their immobilization into solid supports enables the fabrication of hybrid systems that combine the high catalytic activity of molecular catalysts with the stability and recyclability of heterogeneous materials [29,30].
In nature, water oxidation occurs at the oxygen evolving complex (OEC), a {Mn 4 CaO 5 } oxo-cluster embedded in a protein environment within photosystem II (PS II) [31][32][33]. The efficacy of this biological enzymatic system relies on effective charge separation and accumulation, and electron mobility mechanisms [34]. The {Mn 4 CaO 5 } cluster uses water as the ultimate source of electrons and transfers them to a tyrosine residue Y161, which locates directly at the periphery of the cluster, and provides an oxidizable side-chain to recover the photoexcited chlorophyll P680* into its ground state through rapid electron injection [35]. Notably, the harsh water oxidation conditions constantly damage the OEC and its surroundings, the structural integrity of which is reconstituted via self-repair mechanisms to avoid the cessation of activity after 30 min [36].
Given the nature of the OEC, it is not surprising that research efforts have been directed to produce bioinspired Mn-based WOCs featuring multiple metallic centers [37][38][39][40][41]. These WOCs are appealing as they can facilitate multiple-electron redox processes by accumulating multiple charge equivalents through different metal centers in a single molecule and within a narrow potential range. In addition, Mn is a highly attractive element due to its high abundance in the Earth's crust, availability and low toxicity [42]. Unfortunately, the low stability of Mn-based WOCs under operando conditions is a common issue that leads to either the formation of catalytically active heterogeneous manganese oxide, thus precluding the proper characterization of the complex [43], or fast catalytic decay due to catalyst deactivation [44]. A series of dodecanuclear Mn-oxo clusters, [Mn 12 O 12 (OAc) 16 4 ] are especially interesting as they show amenability to replace the organic ligand (L) to modulate the WOC activity and solubility in aqueous solutions. For instance, when L = acetate, benzoate, benzenesulfonate, diphenylphosphonate or dichloroacetate, the clusters are water-insoluble and show poor WOC activity in the solid state, displaying high onset overpotentials between 640 and 820 mV. Maayan and co-workers demonstrated that employing redox-active di-or trihydroxybenzoate organic ligands increases the clusters' solubilities in aqueous solutions, allowing for the study of their WOC properties in the homogeneous phase [45,46]. These clusters displayed good catalytic activity and a reduced onset overpotential thanks to the non-innocent role of the quinone-like ligands. However, the oxygen evolution rate decreases gradually after one hour under an applied bias, indicating that catalyst deactivation is still a problem in these systems. Our group recently reported the activity of a bioinspired, high-nuclearity Mn-oxo cluster, [Mn III 12 Mn II 7 (µ 4 -O) 8 (µ 3 -OCH 3 ) 2 (µ 3 -Br) 6 (HL Me ) 12 (MeOH) 5 (MeCN)]Br 2 (Mn 19 ), which contains 2,6-bis(hydroxymethyl)-p-cresol, a non-innocent, redox-active ligand that resembles the tyrosine residue Y161 in PS II [47]. Importantly, Mn 19 -modified carbon paste electrodes delivered high catalytic activity in neutral media, displaying 10 mA cm −2 and even 100 mA cm −2 at applied overpotentials of 482 and 654 mV, respectively. Despite the high activity shown by this cluster, we identified oxidation of the primary alcohols of the organic ligand as the main deactivation pathway leading to catalyst deactivation. All things considered, Mn-oxo clusters in general have shown potential towards water oxidation catalysis, but their stability issues must be addressed in the near future in order to obtain more durable catalysts before their commercial implementation is possible.
Interestingly, the nonadecanuclear Mn-oxo cluster (Mn 19 ) was initially investigated by Powell et al. due to its unprecedented ground spin state of S = 83/2 [48]. Thereafter,  structural variants of this compound where the central Mn II ion was replaced with various  heterometals such Dy III , Sr II , Y III , Cd II and Lu III provided a synthetic route for the [Mn 18 M] family [24]. Additionally, in this work, the authors showed how weak antiferromagnetic interactions between magnetic units can be tuned using different diamagnetic metal centers.
Here, we report the heterogeneous water oxidation catalytic activity at pH = 7.2 of [Mn III 12 Mn II 6 Sr(µ 4 -O 8 )(µ 3 -Cl) 8 (HL Me ) 12 (MeCN) 6 ]Cl 2 ·15MeOH (Mn 18 Sr), a bioinspired, high-nuclearity Mn-oxo cluster which is analogous to the previously reported Mn 19 complex. As part of a modified carbon paste electrode, Mn 18 Sr displays an onset overpotential of 192 mV and achieves a current density of 10 mA cm −2 at 550 mV. Post-catalytic characterization indicates that Mn 18 Sr is the true catalyst under the examined experimental conditions and that it follows the same, or a very similar, deactivation pathway than that found for the Mn 19 complex. Additionally, light-induced water oxidation experiments confirm O 2 evolution promoted by Mn 18 Sr, revealing a TOF of 0.48 s −1 and a maximum TON of 21.6. Mn 18 Sr is a mixed-valence, polynuclear Mn complex stabilized by redox-active, tridentate 2,6-bis(hydroxymethyl)-p-cresol (H 3 L Me ) ligands featuring different cubane motifs, which are organized to form a supertetrahedral structure ( Figure 1). Within this structure, Mn 18 Sr displays cubane-like {Mn 7 Cl 2 O 7 } clusters containing 7-coordinated 'dangling' Mn II atoms that are assumed to act as the catalytically active centers. Additionally, the presence of Sr II is also intriguing, as this is the only metal known which can replace Ca II in the OEC without rendering the cluster catalytically inactive [49]. Moreover, in a recent study by Mameri et al., Mn 18 Sr and Mn 19 were successfully employed in the electrocatalytic oxidation of benzyl alcohol into benzaldehyde using O 2 as an oxidant and TEMPO (2,2,6,6-tetramethylpiperidine-1-oxyl) as a co-catalyst [50]. Therefore, given the previously reported high WOC activity for the homologous Mn 19 cluster, its resemblance with the OEC, its insolubility in aqueous solutions and facile preparation from inexpensive and environmentally benign starting materials, Mn 18 Sr is an attractive candidate for heterogeneous H 2 O oxidation catalysis.
Water 2021, 13, x FOR PEER REVIEW 3 of 13 things considered, Mn-oxo clusters in general have shown potential towards water oxidation catalysis, but their stability issues must be addressed in the near future in order to obtain more durable catalysts before their commercial implementation is possible. Interestingly, the nonadecanuclear Mn-oxo cluster (Mn19) was initially investigated by Powell et al. due to its unprecedented ground spin state of S = 83/2 [48]. Thereafter, structural variants of this compound where the central Mn II ion was replaced with various heterometals such Dy III , Sr II , Y III , Cd II and Lu III provided a synthetic route for the [Mn18M] family [24]. Additionally, in this work, the authors showed how weak antiferromagnetic interactions between magnetic units can be tuned using different diamagnetic metal centers.
Here, we report the heterogeneous water oxidation catalytic activity at pH = 7.2 of [Mn III 12Mn II 6Sr(μ4-O8)(μ3-Cl)8(HL Me )12(MeCN)6]Cl2•15MeOH (Mn18Sr), a bioinspired, highnuclearity Mn-oxo cluster which is analogous to the previously reported Mn19 complex. As part of a modified carbon paste electrode, Mn18Sr displays an onset overpotential of 192 mV and achieves a current density of 10 mA cm -2 at 550 mV. Post-catalytic characterization indicates that Mn18Sr is the true catalyst under the examined experimental conditions and that it follows the same, or a very similar, deactivation pathway than that found for the Mn19 complex. Additionally, light-induced water oxidation experiments confirm O2 evolution promoted by Mn18Sr, revealing a TOF of 0.48 s -1 and a maximum TON of 21.6. Mn18Sr is a mixed-valence, polynuclear Mn complex stabilized by redox-active, tridentate 2,6-bis(hydroxymethyl)-p-cresol (H3L Me ) ligands featuring different cubane motifs, which are organized to form a supertetrahedral structure ( Figure 1). Within this structure, Mn18Sr displays cubane-like {Mn7Cl2O7} clusters containing 7-coordinated 'dangling' Mn II atoms that are assumed to act as the catalytically active centers. Additionally, the presence of Sr II is also intriguing, as this is the only metal known which can replace Ca II in the OEC without rendering the cluster catalytically inactive [49]. Moreover, in a recent study by Mameri et al., Mn18Sr and Mn19 were successfully employed in the electrocatalytic oxidation of benzyl alcohol into benzaldehyde using O2 as an oxidant and TEMPO (2,2,6,6-tetramethylpiperidine-1-oxyl) as a co-catalyst [50]. Therefore, given the previously reported high WOC activity for the homologous Mn19 cluster, its resemblance with the OEC, its insolubility in aqueous solutions and facile preparation from inexpensive and environmentally benign starting materials, Mn18Sr is an attractive candidate for heterogeneous H2O oxidation catalysis. (a)

Materials and Methods
All chemicals and solvents were used as purchased without further purification. The Mn and Sr salts precursors were purchased from abcr GmbH (Karlsruhe, Germany), whereas the rest of precursors were purchased from Sigma-Aldrich Ltd. (Gillingham, United Kingdom). The modified carbon paste electrodes were prepared using ALS Co. Ltd. (Tokyo, Japan), CPO Carbon Paste Oil.

Materials and Methods
All chemicals and solvents were used as purchased without further purification. The Mn and Sr salts precursors were purchased from abcr GmbH (Karlsruhe, Germany), whereas the rest of precursors were purchased from Sigma-Aldrich Ltd. (Gillingham, United Kingdom). The modified carbon paste electrodes were prepared using ALS Co. Ltd. (Tokyo, Japan), CPO Carbon Paste Oil.
The ruthenium polypyridine precursor [Ru(bpy)2]Cl2 was synthesized according to an adapted literature procedure outlined by Meyer et al. [52]. RuCl3•3H2O (2.00 g, 7.649 mmol), bpy (2.39 g, 15.298 mmol) and LiCl (2.15 g, 50.715 mmol) were refluxed together in reagent grade DMF (14 mL) for 8 h. The reaction mixture was then cooled to 0 °C, causing the product to precipitate which was subsequently collected by filtration. The black crystalline product obtained was then washed with distilled H2O (3 × 10 mL) followed by diethyl ether (3 × 10 mL) before allowing the product to dry in air. The yield was 1.593 g, 3.290 mmol, 43%.
Electrocatalytic Water Oxidation: Electrochemical measurements were performed with a Biologic VSP potentiostat. Ohmic drop was compensated prior to each experiment using the positive feedback compensation method as implemented in the instrumental setup. A three-electrode set-up was used comprising a Saturated Calomel Electrode (SCE) reference electrode, carbon paste (CP) working electrode and a Pt wire counter electrode (alternatively, a Pt mesh was employed for bulk water electrolysis experiments). The CP working electrodes employed for cyclic voltammetry (CV) and bulk water electrolysis had a surface area of 0.02 cm 2 , whereas those employed for linear sweep voltammetry (LSV) had a surface area of 0.07 cm 2 . These surface areas were employed to calculate the current densities. To prepare Mn18Sr-modified CP working electrodes (Mn18Sr/CP), Mn18Sr was blended with CP at a 30% weight ratio (30 wt %). Then, the resulting mixture was inserted inside the CP electrode pocket. All the measurements were carried out in a 50 mM potassium phosphate (KPi) buffer solution with KNO3 (1 M) as electrolyte at pH 7.2. An ALS RRDE-3A set-up with a CP rotating disc electrode was used for the LSV measurements at Diethyl 2,2 -bipyridine-4,4 -dicarboxylate (deeb) was synthesized according to a modified literature procedure outlined by Ma et al. [51]. Initially, 2,2 -bipyridine-4,4dicarboxylic acid (0.30 g, 1.228 mmol) and concentrated H 2 SO 4 (1 mL) were refluxed in EtOH (10 mL) for 4 h. The reaction mixture was then allowed to cool to room temperature and distilled water (30 mL) was added. The resulting white precipitate was then collected by filtration and washed with MeOH before drying under vacuum. The yield was 0.28 g, 1.02 mmol, 82%.
The ruthenium polypyridine precursor [Ru(bpy) 2 ]Cl 2 was synthesized according to an adapted literature procedure outlined by Meyer et al. [52]. RuCl 3 ·3H 2 O (2.00 g, 7.649 mmol), bpy (2.39 g, 15.298 mmol) and LiCl (2.15 g, 50.715 mmol) were refluxed together in reagent grade DMF (14 mL) for 8 h. The reaction mixture was then cooled to 0 • C, causing the product to precipitate which was subsequently collected by filtration. The black crystalline product obtained was then washed with distilled H 2 O (3 × 10 mL) followed by diethyl ether (3 × 10 mL) before allowing the product to dry in air. The yield was 1.593 g, 3.290 mmol, 43%.
Electrocatalytic Water Oxidation: Electrochemical measurements were performed with a Biologic VSP potentiostat. Ohmic drop was compensated prior to each experiment using the positive feedback compensation method as implemented in the instrumental setup. A three-electrode set-up was used comprising a Saturated Calomel Electrode (SCE) reference electrode, carbon paste (CP) working electrode and a Pt wire counter electrode (alternatively, a Pt mesh was employed for bulk water electrolysis experiments). The CP working electrodes employed for cyclic voltammetry (CV) and bulk water electrolysis had a surface area of 0.02 cm 2 , whereas those employed for linear sweep voltammetry (LSV) had a surface area of 0.07 cm 2 . These surface areas were employed to calculate the current densities. To prepare Mn 18 Sr-modified CP working electrodes (Mn 18 Sr/CP), Mn 18 Sr was blended with CP at a 30% weight ratio (30 wt %). Then, the resulting mixture was inserted inside the CP electrode pocket. All the measurements were carried out in a 50 mM potassium phosphate (KP i ) buffer solution with KNO 3 (1 M) as electrolyte at pH 7.2. An ALS RRDE-3A set-up with a CP rotating disc electrode was used for the LSV measurements at 1600 r.p.m., and at a scan rate of 10 mV s −1 . Tafel data was extracted from these LSV experiments and the onset overpotential was estimated from the intersection point between the tangent line of the Faradaic current at 1 mA cm −2 and the non-Faradaic current. Chronopotentiometry at applied current densities of 1 mA cm −2 and 10 mA cm −2 were performed in an H-cell where the working and reference electrodes were separated from the counter electrode by a glass frit (P0). For these chronopotentiometric experiments, the CP working electrode surface was covered with 4 µL of Nafion 117 solution using a micropipette and dried at room temperature.
All the applied potentials (E app ) were converted to the NHE reference scale using: The overpotentials (η) were calculated by subtracting the thermodynamic water oxidation potential E 0 H 2 O/O 2 from E app as: where E 0 H 2 O/O 2 was pH-corrected employing the Nernst equation: Light-Induced Water Oxidation Catalysis: Oxygen evolution was measured using a Unisense oxygen sensor (Clark electrode, OX-NP model). The microsensor was connected to a monomer and a computer operating the SensorTrace Suite software package. A twopoint calibration method was employed prior to the measurements, using (a) a zero-oxygen determination of an anoxic aqueous solution (prepared using a mixture of NaOH and sodium ascorbate) and (b) air-saturated distilled water at a known temperature (prepared by vigorously bubbling air for 10 min into the Unisense calibration chamber (CAL300). Light-driven O 2 evolution experiments were carried out using a Prizmatix LED light (λ = 470 nm), connected to a fiber optic cable. The light intensity of the incident radiation at the center of the reactor was determined to be 10 mW cm −2 . The operating distance between the fiber optic tip and the reaction vessel was 1 cm.
Typically, [Ru(bpy) 2 (deeb)](PF 6 ) 2 (deeb = 4,4-di(ethoxycarbonyl)-2,2-bipyridine) (2.00 mg, 0.47 mM) as a photosensitizer (PS) and Na 2 S 2 O 8 (11.9 mg, 10 mM) as sacrificial electron acceptor (SEA) were loaded into a crimp vial, which was then sealed with a butyl-rubber septum and covered with aluminum foil to prevent early reaction. Two 0.1 mg/mL catalysts stock solution were prepared by dispersing either Mn 18 Sr or Mn 19 in a 10 mM NaP i buffer aqueous solution at pH 7.0. For comparison of the water oxidation catalytic activity of these two catalysts, we employed 5 nmol of either Mn 18 Sr or Mn 19 in the experiments by injecting the corresponding volume of the stock solution into the vial. The final volume of the mixture was adjusted to 5 mL by the addition of NaP i buffer solution. The OX-NP Clark electrode (Unisense) and grounding cable were inserted into the solution phase through the septum and the real-time O 2 quantities were monitored thereafter. The aqueous mixture was then de-aerated by bubbling N 2 in the vial for several minutes, before removing the N 2 inlet and waiting for a consistent, steady O 2 reading. At this point, the content of the vial was illuminated whilst the O 2 concentration in solution was monitored in real time.
The turnover numbers (TON) and initial turnover frequencies (TOF) were calculated as follows: TON = n O 2 n catalyst (5) where n O 2 is the number of moles of O 2 produced, n catalyst is the number of moles of catalyst employed and the time period considered was 10 s after the onset of the O 2 evolution.

Results
Heterogeneous water oxidation electrocatalytic experiments were performed using Mn 18 Sr-modified carbon paste (CP) working electrodes, where the Mn 18 Sr weight ratio (wt %) in the CP blend was kept at 30% (Mn 18 Sr/CP). An SCE reference electrode and a Pt wire (or mesh) counter electrode completed the three-electrode set-up. These measurements were conducted in a 50 mM phosphate (KP i ) buffer at pH 7.2 using KNO 3 (1 M) as electrolyte.
Initial cyclic voltammetry (CV) using the Mn 18 Sr/CP electrodes shows increasing current densities at oxidative potentials that clearly contrast with the catalyst-free CP electrodes (blank), indicating the appearance of a multielectron transfer catalytic process (Figure 2a). Is worth mentioning that a crossover loop process appears at the firsts CV cycles. This behavior corresponds to the redox activity of the organic p-cresol type (HL Me ) 2− ligand stabilizing the Mn 18 Sr cluster as we previously demonstrated, and it arises from the partial ligand oxidation to form semihydroquinone, semiquinone or carboxylate derivatives [47]. The electrocatalytic activity and kinetics of the Mn 18 Sr/CP electrodes were further investigated by linear sweep voltammetry (LSV). Figure 2b shows the performance of Mn 18 Sr/CP, which reaches current densities of 1 mA cm −2 and 10 mA cm −2 at 284 and 550 mV of overpotential, respectively. Moreover, these electrodes show a remarkably low onset overpotential (η onset ) of 192 mV ( Figure S3). The LSV data were employed to study the kinetics of the electrode using a Tafel plot (Figure 2c). Mn 18 Sr/CP shows a slope of 220 mV dec −1 at low overpotentials, which increases to 260 mV dec −1 at >300 mV overpotentials. These slope values are higher than the typical Tafel slopes, suggesting that the kinetics of the electrode are limited by a diffusion process and electron transfer. Moreover, organic ligand oxidation may be competing with the water oxidation reaction.
In comparison with the catalytic activity previously reported by the Mn 19 cluster, Mn 18 Sr shows a reduction of ca. 60 mV in the onset overpotential. We also observe a slight decrease of ca. 10 mV in the overpotential needed to reach 1 mA cm −2 . However, due to its increased Tafel slope, Mn 18 Sr displays current densities of 10 mA cm −2 and 100 mA cm −2 at higher overpotentials than Mn 19 (see Table 1 and Figure 2d).  4 Overpotential at the given current density.
The stability of the Mn 18 Sr/CP electrodes was investigated with repetitive CV and chronopotentiometric experiments. Repetitive CV measurements show that after 10 CV cycles, the activity of the Mn 18 Sr/CP decays, suggesting that the catalyst slowly decomposes to a catalytically inactive species (Figure 3a). Moreover, the water oxidation catalytic activity decay is accompanied by the appearance of a pale brown precipitate on the electrode surface. We already observed this same behavior when employing Mn 19 /CP electrodes and identified the new inactive species as amorphous manganese-potassium carbonate-phosphate species. We characterized the pale brown precipitate obtained with the Mn 18 Sr/CP electrodes by Raman spectroscopy. Figure S2b shows the Raman spectra of this precipitate features only two bands that originate from carbon vibrations at 712 and 1049 cm −1 . This result is analogous to the one found for Mn 19 /CP electrodes (Figure S2d), where we identify the Mn 19 cluster as the true active species whereas oxidation of the primary alcohols of the p-cresol type ligand was identified as the main catalyst deactivation pathway. Therefore, given the similar behavior to Mn 19 /CP electrodes, we can conclude that under the tested conditions, Mn 18 Sr, or a species closely related, must act as the true catalytically active species before it undergoes deactivation.
Consistent with the repetitive CV study, we observe a decrease in the WOC activity over time during the chronopotentiometric experiments (Figure 3b). It is important to note that for these experiments, we followed a common procedure to cover the electrode surface with a Nafion film [53]. The Nafion polymer acts as a proton shuttle transferring the protons generated at the electrode during the proton-coupled electron transfer (PCET) [54] events to the bulk and increases the mechanical stability of the CP electrodes. The Mn 18 Sr/CP electrodes show relatively good stability for one hour at applied current densities of 1 mA cm −2 and 10 mA cm −2 , with initial overpotentials of 280 and 550 mV, respectively. The displayed overpotentials slightly increase over time, suggesting catalyst deactivation, which is not surprising given the low stability performed by Mn-oxo clusters under operando conditions. In fact, restarting the chronopotentiometry at an applied current density of 1 mA cm −2 after ca. 80 min does not recover the initial performance of the electrode. Therefore, the applied overpotentials needed to reach current densities of 1 mA cm −2 and 10 mA cm −2 increase ca. 100 mV after 10 and 3 h, respectively (see Figure S4).
Heterogeneous water oxidation electrocatalytic experiments were performed using Mn18Sr-modified carbon paste (CP) working electrodes, where the Mn18Sr weight ratio (wt %) in the CP blend was kept at 30% (Mn18Sr/CP). An SCE reference electrode and a Pt wire (or mesh) counter electrode completed the three-electrode set-up. These measurements were conducted in a 50 mM phosphate (KPi) buffer at pH 7.2 using KNO3 (1 M) as electrolyte.
Initial cyclic voltammetry (CV) using the Mn18Sr/CP electrodes shows increasing current densities at oxidative potentials that clearly contrast with the catalyst-free CP electrodes (blank), indicating the appearance of a multielectron transfer catalytic process (Figure 2a). Is worth mentioning that a crossover loop process appears at the firsts CV cycles. This behavior corresponds to the redox activity of the organic p-cresol type (HL Me ) 2-ligand stabilizing the Mn18Sr cluster as we previously demonstrated, and it arises from the partial ligand oxidation to form semihydroquinone, semiquinone or carboxylate derivatives [47]. The electrocatalytic activity and kinetics of the Mn18Sr/CP electrodes were further investigated by linear sweep voltammetry (LSV). Figure 2b shows the performance of Mn18Sr/CP, which reaches current densities of 1 mA cm -2 and 10 mA cm -2 at 284 and 550 mV of overpotential, respectively. Moreover, these electrodes show a remarkably low onset overpotential (ηonset) of 192 mV ( Figure S3). The LSV data were employed to study the kinetics of the electrode using a Tafel plot (Figure 2c). Mn18Sr/CP shows a slope of 220 mV dec -1 at low overpotentials, which increases to 260 mV dec -1 at >300 mV overpotentials. These slope values are higher than the typical Tafel slopes, suggesting that the kinetics of the electrode are limited by a diffusion process and electron transfer. Moreover, organic ligand oxidation may be competing with the water oxidation reaction. In comparison with the catalytic activity previously reported by the Mn19 cluster, Mn18Sr shows a reduction of ca. 60 mV in the onset overpotential. We also observe a slight decrease of ca. 10 mV in the overpotential needed to reach 1 mA cm -2 . However, due to its increased Tafel slope, Mn18Sr displays current densities of 10 mA cm -2 and 100 mA cm -2 at higher overpotentials than Mn19 (see Table 1 and Figure 2d).  tion pathway. Therefore, given the similar behavior to Mn19/CP electrodes, we can conclude that under the tested conditions, Mn18Sr, or a species closely related, must act as the true catalytically active species before it undergoes deactivation.
(a) (b) Consistent with the repetitive CV study, we observe a decrease in the WOC activity over time during the chronopotentiometric experiments (Figure 3b). It is important to note that for these experiments, we followed a common procedure to cover the electrode surface with a Nafion film [53]. The Nafion polymer acts as a proton shuttle transferring the protons generated at the electrode during the proton-coupled electron transfer (PCET) [54] events to the bulk and increases the mechanical stability of the CP electrodes. The Mn18Sr/CP electrodes show relatively good stability for one hour at applied current densities of 1 mA cm -2 and 10 mA cm -2 , with initial overpotentials of 280 and 550 mV, respectively. The displayed overpotentials slightly increase over time, suggesting catalyst deactivation, which is not surprising given the low stability performed by Mn-oxo clusters under operando conditions. In fact, restarting the chronopotentiometry at an applied current density of 1 mA cm -2 after ca. 80 min does not recover the initial performance of the electrode. Therefore, the applied overpotentials needed to reach current densities of 1 mA cm -2 and 10 mA cm -2 increase ca. 100 mV after 10 and 3 h, respectively (see Figure S4).
Finally, the oxygen evolution promoted by Mn18Sr was confirmed under light-driven water oxidation conditions by employing a well-known protocol [55]. In this system, [Ru(bpy)2(deeb)](PF6)2 was used as a PS and Na2S2O8 as SEA within a 5 nmol Mn18Sr suspension in 5 mL of 10 mM NaPi buffer aqueous solution at pH = 7. The O2 evolved upon light irradiation (LED lamp, λ = 470 nm) was monitored employing a Clark electrode and compared with that of the analogous Mn19 under the same experimental conditions. Figure 4 shows the O2 evolution profiles displayed by both catalysts where the generated O2 quantity continuously rises after the mixture is irradiated and reaches a plateau after ca. 120 s. The Mn19 cluster delivers an overall better catalytic activity with slightly superior kinetics than Mn18Sr, which is in good agreement with the electrocatalytic experiments shown above. Hence, Mn19 reaches a turnover number (TON) of 25.8 and a turnover frequency (TOF) of 0.53 s −1 , whereas Mn18Sr shows a TON of 21.6 and a TOF of 0.48 s −1 . Finally, the oxygen evolution promoted by Mn 18 Sr was confirmed under light-driven water oxidation conditions by employing a well-known protocol [55]. In this system, [Ru(bpy) 2 (deeb)](PF 6 ) 2 was used as a PS and Na 2 S 2 O 8 as SEA within a 5 nmol Mn 18 Sr suspension in 5 mL of 10 mM NaP i buffer aqueous solution at pH = 7. The O 2 evolved upon light irradiation (LED lamp, λ = 470 nm) was monitored employing a Clark electrode and compared with that of the analogous Mn 19 under the same experimental conditions. Figure 4 shows the O 2 evolution profiles displayed by both catalysts where the generated O 2 quantity continuously rises after the mixture is irradiated and reaches a plateau after ca. 120 s. The Mn 19 cluster delivers an overall better catalytic activity with slightly superior kinetics than Mn 18 Sr, which is in good agreement with the electrocatalytic experiments shown above. Hence, Mn 19

Discussion
The exceptional OER activities observed using Mn18Sr and Mn19 could stem from several shared structural features. For example, both compounds comprise high-nuclearity cluster cores, allowing the four electrons which must be abstracted from H2O to promote O2 evolution to be distributed between either eighteen or nineteen redox-active Mn

Discussion
The exceptional OER activities observed using Mn 18 Sr and Mn 19 could stem from several shared structural features. For example, both compounds comprise high-nuclearity cluster cores, allowing the four electrons which must be abstracted from H 2 O to promote O 2 evolution to be distributed between either eighteen or nineteen redox-active Mn centers in Mn 18 Sr or Mn 19 , respectively. This difference could underlie the marginally superior O 2 evolution activity displayed by Mn 19 in comparison to Mn 18 Sr. In addition, each cluster contains twelve Mn III centers, which are integral to catalytic H 2 O oxidation in many WOCs including the OEC of PS II. Furthermore, the geometric configurations of the Mn-oxo clusters Mn 18 Sr and Mn 19 resemble the OEC: two pseudocubane motifs appended by 'dangling' Mn II centers, which are coordinated by terminal, labile solvent molecules, constitute a total of six substrate-accessible active sites per cluster. Finally, the presence of redox-active O-donor ligands which stabilize the cores structures of Mn 18 Sr and Mn 19 represents an additional parallel between these clusters and the OEC. Further research directions may include the replacement of the redox-active ligand to increase the stability of the clusters, which could also improve their water oxidation catalytic performance. Moreover, DFT calculations will help elucidate the WOC reaction mechanisms performed by these clusters to pinpoint key structure-reactivity relationships, and to disclose the role and stability of the redox-active quinone-like organic ligands.

Conclusions
In summary, we have shown the WOC activity in neutral media of a mixed-valence, highnuclearity Mn-oxo cluster containing redox-active ligands, [ 8 (µ 3 -OCH 3 ) 2 (µ 3 -Br) 6 (HL Me ) 12 (MeOH) 5 (MeCN)]Br 2 ·9MeCN·MeOH (Mn 19 ) complex under the same working conditions. As part of a modified carbon paste working electrode, Mn 18 Sr shows good water oxidation electrocatalytic activity with a low onset overpotential of 192 mV. Moreover, the Mn 18 Sr-modified electrodes display relatively low overpotentials of 284 and 550 mV at current densities of 1 mA cm −2 and 10 mA cm −2 , respectively. Unsurprisingly, the catalytic activity of this Mn-oxo cluster decays over time, where the organic ligand oxidation is identified as the main catalyst deactivation pathway. Moreover, the catalytically inactive species that appear after catalyst deactivation has been identified as a manganese-potassium carbonate-phosphate species. Importantly, these results are in good agreement with our previous investigations employing the Mn 19 compound. Under light-induced water oxidation conditions, Mn 18 Sr promotes O 2 evolution upon light irradiation, displaying an initial TOF of 0.48 s −1 and a maximum TON of 21.6. Overall, the Mn 19 cluster displays an improved WOC activity than Mn 18 Sr due to its faster kinetics.
Our results confirm the potential of this family of high-nuclearity Mn-oxo clusters as WOCs and highlights their amenability to structural modifications to tune their electronic properties and modulate their reactivity.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/w13152042/s1, Figure S1: FT-IR spectra of the freshly prepared Mn 18 Sr (a) and Mn 19 (b) complexes, Figure S2: Comparison of the Raman spectra obtained for the freshly developed Mn 18 Sr (a) and Mn 19 (c) complexes, and that for the recovered decomposition products identified as a catalytically inactive amorphous manganese-potassium carbonate-phosphate species (b,d), Figure S3: Calculation of the onset overpotential employing 30% Mn 18 Sr/CP working electrodes. The onset overpotential was estimated from the intersection point between the tangent line of the Faradaic current at 1 mA cm −2 and the extrapolated line arising from the non-Faradaic current, Figure S4: Long-term chronopotentiometry measurements at applied current densities of (a) 1 mA cm −2 (red) and (b) 10 mA cm −2 (orange). The experiments were performed in an H-cell using a 50 mM KP i aqueous buffer at pH = 7.2 using KNO 3 (1 M) as electrolyte.