Dinuclear manganese complexes for water oxidation: evaluation of electronic eﬀects and catalytic activity

During recent years significant progress has been made towards the realization of a sustainable and carbon-neutral energy economy. One promising approach is photochemical splitting of H 2 O into O 2 and solar fuels, such as H 2 . However, the bottleneck in such artificial photosynthetic schemes is the H 2 O oxidation half reaction where more eﬃcient catalysts are required that lower the kinetic barrier for this process. In particular catalysts based on earth-abundant metals are highly attractive compared to catalysts comprised of noble metals. We have now synthesized a library of dinuclear Mn 2II,III catalysts for H 2 O oxidation and studied how the incorporation of diﬀerent substituents aﬀected the electronics and catalytic eﬃciency. It was found that the incorporation of a distal carboxyl group into the ligand scaﬀold resulted in a catalyst with increased catalytic activity, most likely because of the fact that the distal group is able to promote proton-coupled electron transfer (PCET) from the high-valent Mn species, thus facilitating O–O bond formation.


Introduction
The world's global energy consumption is primarily based on fossil fuels. However, the rapidly increasing global energy demand coupled to diminishing fossil fuel supplies call for the development of sustainable and carbon-neutral alternatives, with feedstock(s) that are environmentally friendly and abundant. There are already efficient solar cells for generation of electricity but in order for this to become truly useful, it is necessary that part of the energy can be stored. One attractive alternative would therefore be to store the energy in sunlight by the splitting of H 2 O into O 2 , protons and electrons (eqn (1)). The generated electrons can then be used to produce H 2 or other solar fuels, thus allowing the energy to be stored in the form of chemical bonds. The production of solar fuels constitutes an attractive solution for a future sustainable energy economy. [1][2][3][4][5] Nature has since a long time figured out how the process of splitting of H 2 O can be realized and the natural photosynthetic machinery involves the synchronization of several complicated features, such as light-harvesting, charge separation and electron transfer. Both in natural and artificial photosynthesis, catalytic H 2 O oxidation is an essential part of the intricate photochemical process of solar to chemical energy conversion. Although deceptively simple, H 2 O oxidation (eqn (1)) is associated with a high thermodynamic potential (eqn (2)), a high kinetic barrier and requires a multitude of bond breaking and bond making events. This makes the H 2 O oxidation half reaction the pitfall for the construction of artificial devices for the production of solar fuels. [6][7][8] 2H 2 O -O 2 + 4H + + 4e À (1) based on earth-abundant elements. This prompts the examination and development of molecular WOCs based on low-cost first-row transition metals. Although there have been a couple of reports on WOCs comprised of Mn, [33][34][35][36][37][38][39][40][41][42] Co, [43][44][45][46][47][48][49][50] Fe [51][52][53][54] and Cu, [55][56][57] this research has not been as successful as with the rare metals Ru and Ir. The incorporation of negatively charged groups into ligand frameworks has been shown to drastically reduce the redox potentials of the corresponding metal complexes. 58,59 The electron-rich ligand environment helps to stabilize the metal center(s) in high oxidation states. We have previously reported the synthesis of Mn complexes housing ligands with benzylic amines. 60,61 However, due to the labile benzylic positions, these complexes failed to mediate H 2 O oxidation.
Our group recently reported on the synthesis and development of the bio-inspired ligand 1 containing carboxylate, phenol and imidazole functionalities, 33 which are all important elements in the natural system. 62 Complexation of ligand 1 with Mn(OAc) 2 yielded the dinuclear Mn 2 II,III complex 2, which in the solid state crystallized as the tetranuclear complex 2 0 with loss of the bridging acetate ligand, thus resembling the Mn 4 Ca cluster in the OEC (Fig. 1). 33 Maintaining a high redox flexibility of the active metal center(s) in the artificial WOCs constitutes a key parameter when designing synthetic WOCs, since the oxidation of H 2 O requires the collective transfer of four electrons. It was therefore vital that the negatively charged functional groups in the dinuclear Mn 2 II,III complex were shown to dramatically lower the redox potentials of the complex.

Electrochemistry
Electrochemical measurements were carried out using an Autolab potentiostat with a GPES electrochemical interface (Eco Chemie), using a glassy carbon disk (diameter 3 mm) as the working electrode, and a platinum spiral as a counter-electrode. In a typical run [Ru(bpy) 3 ](PF 6 ) 3 (3.4 mg, 3.4 mmol) was placed in the reaction chamber and the reaction chamber was evacuated using a rough pump. B35 mbar He was then introduced into the system. After a couple of minutes the catalyst solution (0.50 mL, 4.0 mM) was injected into the reaction chamber. The generated oxygen gas was then measured and recorded versus time by MS.

Computational details
The geometry optimizations were performed at the B3LYP level 68 of density functional theory as implemented in the Gaussian 09 package. 69 The SDD 70 pseudopotential was used to describe Mn, and the 6-31G(d,p) basis set was used for the C, N, O, H elements. Analytical frequency calculations were carried out at the same level of theory as the geometry optimization to obtain the Gibbs free energy corrections and to confirm the characteristics of the optimized structures. On the basis of these optimized geometries, single-point calculations were performed employing a larger basis set, where all elements, except Mn, were described by 6-311+G(2df,2p) at the B3LYP* (15% exact exchange) level. 71 Solvation effects from the water solvent were calculated using the SMD 72 continuum solvation model with the larger basis set at the B3LYP* level. The B3LYP*-D2 energies are reported, including dispersion corrections proposed by Grimme 73 and Gibbs free energy corrections from B3LYP.

EPR measurements
X-band (9 GHz) electron paramagnetic resonance (EPR) measurements were performed at temperatures between 5 and 30 K with a Bruker E500 ELEXYS spectrometer using a 4122SHQE or a 4116DM resonator. For cooling with liquid helium an Oxford Instruments ESR 900 cryostat was installed. Frozen solutions of the complexes were prepared in 0.2 M phosphate buffer (pH 7) and for solid state spectra the complex and potassium bromide were ground together. All samples were frozen in liquid nitrogen prior to EPR measurements. The modulation amplitude was 5 G and the microwave power was varied between 200 mW and 6 mW.

Synthesis
To investigate how the electronics affected the Mn centers in WOC 2, eight different dinuclear Mn complexes 6a-h, depicted in Scheme 1, were synthesized, where a variety of substituents were introduced. These alterations would hence result in tuning of the corresponding Mn complexes and thereby affect the properties and reactivities. The synthesis of the functionalized ligand scaffolds involved the reductive cyclization reaction between the appropriate phenolic dialdehyde 4 and 2-amino-3nitrobenzoic acid, with Na 2 S 2 O 4 as the reducing agent. 74 This allowed for easy access to a wide variety of ligands in good to excellent yields. The Mn complexes 6a-h were subsequently obtained by refluxing the ligands in the presence of Mn(OAc) 2 and NaOAc. It is believed that all of the synthesized complexes were isolated in their Mn 2 II,III state, in conformity with complex 2. This is based on elemental analyses and high-resolution mass spectrometry (HRMS). However, extensive attempts to prove this by EPR have failed so far since we have been unable to observe the characteristic EPR signal for Mn 2 II,III complexes. 75 Instead, a broad signal at around g = 2 was observed both in solution and in the solid state, which is typical for uncoupled Mn II and reminiscent of the signal recently observed for an octahedral Mn II complex. 76 This signal decreased in intensity upon the addition of the one-electron oxidant [Ru(bpy) 3 ] 3+ . It therefore seems possible that the complexes are in fact polymeric or at least tetranuclear, as indicated by the generated crystal structure of 2 0 from complex 2. The formation of the tetranuclear complex 2 0 indicates that the acetate ligand in complex 2, and probably also in complexes 6, is somewhat labile and we assume that on dissolution in phosphate buffer, the acetate ligand is rapidly replaced by phosphate to yield the active catalyst. On dissolution of the dinuclear Mn complexes in aqueous solutions containing PO 4 3À , in the presence of air, new signals appeared, as detected by HRMS. These signals corresponded to the formation of Mn 2 III,III complexes, which are expected to be EPR silent. However, this will be studied in more detail in future work.

Catalytic H 2 O oxidation by use of the mild one-electron oxidant [Ru(bpy) 3 ] 3+
A majority of the previously developed artificial WOCs require the use of the strong one-electron oxidant Ce IV to drive H 2 O oxidation. However, this approach is not useful in solar fuel devices since the oxidant cannot be regenerated photochemically, using sunlight as the terminal energy source. For this to be realized, Ce IV and related oxidants have to be substituted by photosensitizerbased oxidants, such as the well-studied [Ru(bpy) 3 ] 2+ -type complexes (bpy = 2,2 0 -bipyridine). 66,77 For this more preferable approach to be thermodynamically possible, the redox potential of the WOC has to be matched with that of the photosensitizer. Since the previously developed Mn WOC 2 could promote catalytic H 2 O oxidation with the mild one-electron oxidant [Ru(bpy) 3 ] 3+ (E (Ru III /Ru II ) = 1.26 vs. NHE 66 ), either pregenerated or photochemically generated, the catalytic experiments for the newly developed complexes 6 were conducted as previously established; the [Ru(bpy) 3 ] 3+ oxidant was employed as the chemical oxidant in a buffered aqueous solution, under neutral conditions (0.1 M phosphate buffer, pH 7.2). The evolution of gaseous products was monitored and quantified by real-time mass spectrometry, a technique which we have previously used. 67 In a typical reaction, [Ru(bpy) 3 ] 3+ was added to an aqueous solution containing the Mn complex, triggering immediate O 2 evolution.
Several control experiments were conducted in order to verify that the observed O 2 evolution was mediated by the investigated Mn complexes; (1) no O 2 was generated when the Mn complex was omitted. The result was only degradation of the [Ru(bpy) 3 ] 3+ oxidant, without any evolution of O 2 . (2) To ensure that free Ru, possibly originating from the decomposition of the [Ru(bpy) 3 ] 3+ oxidant, did not react with the free ligand to generate an active catalyst, an aqueous solution containing the free ligand was added to [Ru(bpy) 3 ] 3+ . However, this resulted in no observable O 2 evolution. Collectively, this supports that the detected O 2 formation was caused by the Mn complexes and not by any other unexpected catalytic process. Fig. 2 and Table 1 show the O 2 evolution activity of Mn complexes 6. It is evident that Mn complexes 6a-e all have similar O 2 evolution activity, which highlights that the introduced substituents in 6a-e did not affect the catalytic activity to a large extent. The previously reported Mn complex 2 was also evaluated under these reaction conditions and was found to display comparable activity as catalysts 6a-e (not shown).
However, Mn complex 6f, housing the long aliphatic chain with a terminal carboxylate group, displayed higher catalytic activity than the other complexes. This effect could not be ascribed to significantly decreased redox potentials in complex 6f (vide infra), which would improve the O-O bond formation and subsequent O 2 formation. To additionally probe how big influence the substituent in complex 6f had on the electronics, it was decided to synthesize the related Mn complex 6g. Although the two complexes 6f and 6g have electronic resemblance, they do not at all display the same catalytic activity towards oxidizing H 2 O. This highlights that there is a different reason as to why Mn complex 6f displays a higher catalytic activity and thus facilitates the oxidation of H 2 O. To further establish the catalytic importance of the substituent in complex 6f, we turned our attention to electrochemistry in the hope that this would reveal the intrinsic reason for the better efficiency of this complex.
It should also be noted that the bpy functionalized Mn complex 6h was found to be catalytically active. This complex is a precursor of fundamental importance and could potentially be used for the construction of coupled photosensitizer-WOC systems for incorporation into devices for solar to chemical energy conversion (Fig. 3).

Electrochemical measurements
In order to assess the effect of the different substituents on the electronic properties of Mn complexes 6, it was decided to   complexes. Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) were used to analyze the redox properties. Cyclic voltammograms revealed catalytic currents, assigned to electrochemical oxidation of H 2 O, for all of the studied Mn complexes 6 (Fig. 4). However, it was difficult to clearly assign the onset potentials due to overlapping redox processes in the onset potential region. Therefore, to gain further insight into the redox chemistry of the newly synthesized Mn complexes 6, DPV was carried out on the complexes. The differential pulse voltammograms of the dinuclear Mn complexes 6 are depicted in Fig. 5 and displayed several peaks. Within the developed library, the differential pulse voltammograms of complexes 6f and 6h displayed five discernible peaks in the region 0.50 o E o 1.30 V vs. NHE (Table 2). These peaks were subsequently assigned to the formal oxidations of Mn 2 II,III -Mn 2 III,III -Mn 2 III,IV - , in conformity with the previously reported result for a related dinuclear Mn complex. 61 However, the other complexes displayed less than five redox peaks. This is probably related to the close potential gap, i.e. the narrow potential range for some of the redox couples, resulting in single observable two-electron redox peaks instead of two single-electron waves.
As is evident from these voltammograms and Table 2, the substituents in Mn complexes 6 did not cause any dramatic shift of the different redox processes in the complexes. This highlights that there must be another explanation for the increased efficiency of Mn complex 6f, containing the distal carboxyl group. Previous studies of so-called ''Hangman'' porphyrin and corrole complexes housing distal hydrogen bonding groups, such as carboxyl groups, have established the possibility of accessing metal complexes with increased affinity for promoting  PCET reactions. 37,45,[78][79][80][81] There even exists an example of an Fe complex comprising a Hangman platform, which in the solid state exhibits a hydrogen bonded H 2 O molecule bound between a distal xanthene carboxylic acid and an Fe porphyrin hydroxide by two hydrogen bonds (Fig. 6). 78 This suggested that the carboxyl group in 6f could perhaps also interact with the Mn centers. To confirm this hypothesis, quantum chemical calculations were performed on complex 6f.  Fig. 7. Both structures have a total charge of +0 and exist in the high spin octet state according to the calculations. The two Mn centers are bridged by an oxo group, an acetate, and the phenolate of the ligand. In addition, each Mn center also has a coordinated hydroxide ligand. Spin density analysis suggests   that both metals are in fact in their Mn IV state, with a coordinated ligand radical cation. This suggests that the ligand framework is redox-active, non-innocent, and highlights the intriguing properties of ligand scaffolds of this type. 86 Redox-active ligands might thus be of importance in alleviating the metal centers from being heavily oxidized by storing redox equivalents at the ligand scaffolds during the oxidation of H 2 O. In structure A of the oxidized complex 6f, the distal carboxyl group generates two hydrogen bonds to the two Mn coordinated hydroxide ligands. These two hydrogen bonds might thus influence the subsequent O-O bond forming step(s). This supports that the distal carboxylate group in the ligand framework of Mn complex 6f can accommodate a bridging hydroxide/aqua molecule to yield a unique hydrogen-bonded scaffold that might facilitate H 2 O oxidation and thus promote the crucial O-O bond formation through PCET. Multi-electron catalysis is at the heart of H 2 O oxidation and the ability of synchronizing proton and electron transfer events via PCET is fundamental in a plethora of biological and chemical reactions. 87,88 PCET allows the charge of a system to remain unchanged, whereas non-proton coupled single electron transfer processes result in charge accumulation and high-energy intermediates. Detailed studies and a fundamental understanding of how to couple proton and electron transfer are at the frontier and might be the key to realizing more efficient WOCs.

Quantum chemical insight
An additional interesting feature of the ligand platforms in Mn complexes 6 is the presence of the imidazole motifs, which have the ability to function as proton transfer mediators. In structure B, the carboxyl group has abstracted a proton from the imidazole unit. Structure B was found to be 2.0 kcal mol À1 higher in energy than structure A. This hydrogen bonding capability of the distal carboxylate group renders it possible to also promote proton transfer reactions between the distal group and the protonated imidazole moiety. The alternative scenario for the enhanced activity of complex 6f is that the carboxyl group serves as a ligand to bridge the two metal centers. However, calculations revealed that the chain is not sufficient to accommodate this sort of coordination.
The possibility of including a non-innocent distal group within the ligand scaffolds in Mn complexes 6 is an interesting feature in these developed WOCs that has perhaps not received sufficient attention in H 2 O oxidation schemes. The key feature of incorporating a distal group that acts as a general base to stimulate a H 2 O molecule in undergoing nucleophilic attack on a high-valent metal-oxo species could be a general strategy for promoting catalytic H 2 O oxidation activity. This idea of hydrogen bonding as a means to increase the nucleophilicity of H 2 O is utilized in hydrolytic enzymes such as chymotrypsin. It has also been used by Nocera 45,78,81 and ourselves 37 to promote H 2 O oxidation and also recently in two theoretical papers. 79,80 In those previous studies, rather elaborate rigid systems have been used. However, our present results suggest that even a carboxyl attached to a highly flexible chain will participate efficiently. Using this concept we will therefore try to prepare analogues to complex 6f, and also analogues with a different central ligand, using the synthetic versatility available for modifying the ligand frameworks of the complexes 6.

Conclusions
In the present work we have synthesized eight new dinuclear Mn 2 II,III complexes and examined the H 2 O oxidation process for these complexes. It was demonstrated that all of the developed Mn complexes were catalytically active in H 2 O oxidation. The complexes had sufficiently low redox potentials to allow for the use of the mild one-electron oxidant [Ru(bpy) 3 ] 3+ to drive H 2 O oxidation at neutral pH conditions. These studies have identified a potential general factor for obtaining WOCs that display increased catalytic efficiencies. Within the library of the synthesized Mn complexes, complex 6f containing the long aliphatic chain with a terminal carboxylate group as a distal group was found to exhibit a higher O 2 evolution rate than the other catalysts. This observation did not originate from decreased redox potentials, which would result in more favourable thermodynamics for carrying out H 2 O oxidation.
Instead, the improved catalytic effect seems to result from pre-orientation of the incoming H 2 O nucleophile and hydrogen bonding between the distal carboxyl group and a high-valent Mn-hydroxy group. This hydrogen bonding interaction can subsequently promote PCET processes, which decrease the energy of the catalytic H 2 O oxidation process.
These findings highlight the importance of controlling both proton and electron transfer processes to promote O-O bond formation during H 2 O oxidation. In the natural photosynthetic system, PCET is prevalent and provides the key to efficient H 2 O oxidation catalysis. The present study provides a guide that might be helpful for designing more efficient WOCs in the future, which have the possibility of being incorporated into large scale systems for solar to chemical energy conversion.