Theoretical study of catalytic mechanism for single-site water oxidation process

Abstract

Water oxidation is a linchpin in solar fuels formation, and catalysis by single-site ruthenium complexes has generated significant interest in this area. Combining several theoretical tools, we have studied the entire catalytic cycle of water oxidation for a single-site catalyst starting with [Ru^II(tpy)(bpm)(OH₂)]²⁺ (i.e., [Ru^II-OH₂]²⁺; tpy is 2,2^′∶6^′,2^′′-terpyridine and bpm is 2,2′-bypyrimidine) as a representative example of a new class of single-site catalysts. The redox potentials and pK_a calculations for the first two proton-coupled electron transfers (PCETs) from [Ru^II-OH₂]²⁺ to [Ru^IV = O]²⁺ and the following electron-transfer process to [Ru^V = O]³⁺ suggest that these processes can proceed readily in acidic or weakly basic conditions. The subsequent water splitting process involves two water molecules, [Ru^V = O]³⁺ to generate [Ru^III-OOH]²⁺, and H₃O⁺ with a low activation barrier (∼10 kcal/mol). After the key O---O bond forming step in the single-site Ru catalysis, another PECT process oxidizes [Ru^III-OOH]²⁺ to [Ru^IV-OO]²⁺ when the pH is lower than 3.7. Two possible forms of [Ru^IV-OO]²⁺, open and closed, can exist and interconvert with a low activation barrier (< 7 kcal/mol) due to strong spin-orbital coupling effects. In Pathway 1 at pH = 1.0, oxygen release is rate-limiting with an activation barrier ∼12 kcal/mol while the electron-transfer step from [Ru^IV-OO]²⁺ to [Ru^V - OO]³⁺ becomes rate-determining at pH = 0 (Pathway 2) with Ce(IV) as oxidant. The results of these theoretical studies with atomistic details have revealed subtle details of reaction mechanisms at several stages during the catalytic cycle. This understanding is helpful in the design of new catalysts for water oxidation.

Water oxidation (2H₂O → O₂ + 4e^- + 4H⁺) is a key step in both natural and artificial photosynthesis (1–10). A large number of studies have been carried out to design new water oxidation catalysts related to solar fuels (11–19). For example, Meyer et al. have reported single-site polypyridyl ruthenium complexes [Ru^II(tpy)(bpm)(OH₂)]²⁺ ([Ru^II-OH₂]²⁺∶tpy = 2,2^′∶6^′,2^′′-terpyridine; bpm = 2,2^′-bipyrimidine) for water oxidation (11). Inspired by this pioneering work, a series of ruthenium catalysts have been scrutinized to understand the entire catalytic processes (17, 18, 20–23). As shown in Fig. 1, [Ru^II-OH₂]²⁺ is first oxidized to [Ru^IV = O]²⁺ by losing two protons and two electrons. Through a simple electron transfer step, [Ru^IV = O]²⁺ is further oxidized to [Ru^V = O]³⁺. The key O---O bond is formed by water molecule attack on [Ru^V = O]³⁺. It takes two water molecules to generate [Ru^III-OOH]²⁺ and H₃O⁺, and the computed activation barrier is low (∼10 kcal/mol) from our previous studies (21). This reaction mechanism of O---O bond formation (i.e., [Ru^III-OOH]²⁺) has further helped experimental design by employing different bases as proton acceptors, including , , and CH₃COO^-, to increase the rate of the O---O coupling step and enhance catalytic efficiency (21). Subsequently, [Ru^III-OOH]²⁺ is oxidized to [Ru^IV-OO]²⁺ by a proton-coupled electron transfer (PCET) step. Two possible conformations of [Ru^IV-OO]²⁺ exist: In the closed conformation, two oxygen atoms bind to the ruthenium metal center (η²-[Ru^IV-OO]²⁺) while only one oxygen atom binds in the open form (η¹-[Ru^IV-OO]²⁺). The conformation interconversion between closed and open forms may hinder O₂ release (19). In 0.1 M HNO₃, water attack on [Ru^IV-OO]²⁺ to release O₂ was assumed to be rapid (Pathway 1) (17). In 1.0 M HNO₃, [Ru^IV-OO]²⁺ is first oxidized to [Ru^V-OO]³⁺ followed by rapid loss of coordinated O₂ with water addition and proton loss to give [Ru^III-OH]²⁺ (Pathway 2). Subsequently, another PCET process in Pathway 2 returns [Ru^III-OH]²⁺ to [Ru^IV = O]²⁺.

Fig. 1.

Catalytic steps of water oxidation by the single-site catalyst.

Although the catalytic process in Fig. 1 has been explored by several experimental and theoretical studies (19, 22–24), the subtle mechanisms of some key reaction processes still remain enigmatic. For instance, how large are the driving forces of PCET steps? And how rapid is O₂ release? Are there other pathways—a dimeric species formed by O---O coupling of two [Ru^V = O]³⁺ at high concentrations of catalyst followed by further oxidation by Ce(IV)?

To answer these questions, it is necessary to conduct theoretical investigations of several reaction steps at the atomistic level. In this paper, we used ab initio quantum mechanics (QM) and hybrid QM/MM (molecular mechanics) simulations. We found that the spin states of ruthenium intermediates during the catalytic cycle play crucial roles in determining redox potential values and atomistic reaction pathways. The activation barrier of conformation changes between the closed and open forms of [Ru^IV-OO]²⁺ can be lowered to 7 kcal/mol due to strong spin-orbital coupling effects of ruthenium. To release triplet O₂, both [Ru^IV-OO]²⁺ and [Ru^V-OO]³⁺ must reach high spin states; i.e., triplet for [Ru^IV-OO]²⁺ in Pathway 1 and quartet for [Ru^V-OO]³⁺ in Pathway 2. As such, our computations show that the reaction barrier to release O₂ from [Ru^IV-OO]²⁺ is 12 kcal/mol in Pathway 1 while [Ru^V-OO]³⁺ can rapidly release O₂ without any activation barrier in Pathway 2. In addition, the dimerization of [Ru^V = O]³⁺ to generate [Ru^IV-O-O-Ru^IV]⁶⁺ can occur with a low activation barrier (∼5 kcal/mol) with the dimer in the singlet spin state.

Results and Discussion

[Ru^II-OH₂]²⁺ (Singlet)/[Ru^IV = O]²⁺ (Singlet) Couples.

The computed redox potentials and pK_a values are shown in Fig. 2 along with some experimental data in parentheses. Even though the errors residing in approximate forms of functionals in density functional theory (DFT) may not reproduce accurate redox potentials and pK_a values for ruthenium complexes, we found that a hybrid computational protocol can reproduce the correct spin states for ruthenium complexes after our extensive tests on functionals and basis sets (see Tables S1–S3 for complete lists of various Ru species). In this protocol, we employed the B3LYP functional with LANL2DZ basis sets to optimize the geometries of ruthenium intermediates and MP2 calculations with LANL2DZ to identify the corresponding spin states. (Note that even MP2/LANL2DZ is too expensive to be applied for geometry optimizations.) All the redox potentials and pK_a values are also computed by B3LYP/LANL2DZ with the spin states assigned by MP2/LANL2DZ (some redox potentials were computed by MP2 as well). This hybrid protocol (see discussions in the Spin-State Identifications Using Different Methods section of SI Text) is used in all of our calculations.

Fig. 2.

Computed thermochemistry pathways using B3LYP/LANL2DZ from [Ru^II-OH₂]²⁺ to [Ru^IV = O]²⁺. The available experimental values are list in parentheses. NHE (normal hydrogen electrode) is used here (4.24 V).

As shown in Fig. 2, when pH > 8.5 (Exp. 9.7), [Ru^II-OH₂]²⁺ is first deprotonated and then oxidized to [Ru^III-OH]²⁺ with the calculated redox potential around 0.6 V (Exp. 0.6–0.8 V). Increasing the potential to > 1.7 V [Exp. 1.2 V (11)] results in oxidation of [Ru^II-OH₂]²⁺ to [Ru^III-OH₂]³⁺. The computed pK_a value suggests that [Ru^III-OH₂]³⁺ is acidic, which is in qualitative agreement with experimental observations (11, 20, 25). In other pH-potential domains, both proton transfer and electron transfer occur (i.e., PCET) from [Ru^II-OH₂]²⁺ to give [Ru^III-OH]²⁺. The overall oxidation of [Ru^II-OH₂]²⁺ to [Ru^III-OH]²⁺ is uphill and endothermic with [Ru^III-OH]²⁺ relatively stable. To reach [Ru^IV = O]²⁺, an additional PCET step occurs by the [Ru^III-OH]²⁺/[Ru^IV = O]²⁺ couple. [Ru^IV-OH₂]⁴⁺ and [Ru^IV-OH]³⁺ are extremely acidic and unstable in aqueous solution. [Ru^III = O]⁺ and [Ru^II = O] cannot be reached due to high pK_a values of [Ru^III-OH]²⁺ and [Ru^II-OH]⁺.

[Ru^IV = O]²⁺ (Singlet)/[Ru^V = O]³⁺ (Doublet) Couples.

This step is a simple electron transfer process. The redox potential is 1.65 V computed from MP2/LANL2DZ in good agreement with experimental observations (19) (1.6–1.8 V). (Note that we observed that MP2/LANL2DZ calculations usually have lower redox potential values than B3LYP/LANL2DZ.) After electron transfer, the bond distance between Ru and O is decreased from 1.807 to 1.739 Å (Fig. S1). This shorter Ru-O distance facilitates concerted O atom-proton transfer (APT) with O-O bond formation and involvement of two water molecules to produce Ru^III-OOH²⁺ + H₃O⁺ (21).

Dimerization of [Ru^V = O]³⁺.

To explore a possible reaction pathway involving dimerization of [Ru^V = O]³⁺ at high concentrations of [Ru^V = O]³⁺ at low pH, we performed theoretical studies on the putative peroxide-bridged dimer [Ru^IV-O-O-Ru^IV]⁶⁺. Calculations with the implicit solvent model point to mixed triplet-singlet spin-state character. Fig. 3 shows the free energy profiles optimized by QM/MM-MFEP with explicit water molecules for both singlet and triplet spin states when the dimer [Ru^IV-O-O-Ru^IV]⁶⁺ is broken into two monomers. In the singlet spin state, the barrier to O-O bond fission is ∼20 kcal/mol. In other words, two [Ru^V = O]³⁺ monomers can easily form from the dimer since the activation barrier for dimerization is only 4 kcal/mol in the singlet spin state. Note that calculations for the triplet spin state in Fig. 3 give the opposite result. Even though the energy of singlet spin states may be overestimated due to fractional spin errors (26–28), we conclude the dimerization process is likely to proceed in the low spin state (singlet).

Fig. 3.

Free energy profiles for singlet and triplet spin states when the dimer [Ru^IV-O-O-Ru^IV]⁶⁺ is broken into two monomers. The QM subsystems are computed by B3LYP/LANL2DZ.

[Ru^III-OOH]²⁺ (Doublet)/[Ru^IV-OO]²⁺ (Singlet) Couples.

Our computed thermochemical data in Fig. 4 suggest that this is a PCET step when the pH exceeds 3.4. Note that the computed redox potential for the [Ru^III-OOH]²⁺/[Ru^IV-OO]²⁺ couple is too high [calc. 2.6 V vs. exp. 1.4 V (17)]. As shown in Fig. S2, both Ru-O and O---O bonds are shortened after PCET oxidation. The calculated O---O bond length of 1.339 Å in [Ru^IV-OO]²⁺ points to strong peroxide bonding between the oxygen atoms.

Fig. 4.

Computed thermochemistry pathways using B3LYP/LANL2DZ of [Ru^III-OOH]²⁺ to [Ru^IV-OO]²⁺. The available experimental values are listed in parentheses. NHE is used here (4.24 V).

Interconversion of the Singlet and Triplet Spin States of [Ru^IV-OO]²⁺ in Pathway 1.

Due to the high spin-orbit coupling constant for ruthenium (∼1,000 cm^-1) (19), the ground electronic spin state of [Ru^IV-OO]²⁺ is of mixed spin character. For instance, the energy difference between the singlet and triplet spin states of [Ru^IV-OO]²⁺ for the closed shell structure can be less than -1.2 kcal/mol (see Tables S3 and S4 without zero-point energy corrections) from B3LYP/LANL2DZ computations consistent with mixed spin character. Hence, the interconversion of spin states can influence optimal structures of [Ru^IV-OO]²⁺. Note that [Ru^IV-OO]²⁺ has two possible conformations: open (i.e., only one oxygen atom binding to Ru) and closed (i.e., both oxygen atoms binding to Ru). In addition to our previous studies focusing on the singlet spin state (19), the structural interconversion between open and closed structures was scrutinized by using the triplet spin state of [Ru^IV-OO]²⁺. As shown in Fig. S3B, the open structure of [Ru^IV-OO]²⁺ is slightly more stable by 2.6 kcal/mol than the closed one with [Ru^IV-OO]²⁺ in the triplet spin state. More surprisingly, compared to the activation barrier of 14.6 kcal/mol from the closed structure to the open one for the singlet spin state (19) (Fig. S3A), the activation barrier is just 1.5 kcal/mol from closed to open structures for the triplet spin state (Fig. S3B). The significant changes of structures, activation energies, and spin states indicate that spin flipping may occur in [Ru^IV-OO]²⁺. Fig. 5 illustrates how spin can be flipped during the geometric change from closed to open forms. The potential energy scan uses the Ru-O1 bond distance as the reaction coordinate. When the bond distance between Ru and O1 is stretched from 2.09 Å (the optimal distance for singlet [Ru^IV-OO]²⁺) to 2.5 Å, the energy difference between singlet and triplet spin states becomes smaller (see black and red curves in Fig. 5). Eventually, the triplet spin state falls below the singlet at a Ru-O1 bond length > 2.3 Å. This phenomenon supports the conclusion that the singlet closed-form of [Ru^IV-OO]²⁺ can be interconverted to the triplet open-form rapidly with a much lower activation barrier (∼7 kcal/mol in contrast to 14.6 kcal/mol) without considering spin-orbital coupling effect. This discovery demonstrates that the conformational change from closed to open form is not rate-limiting during O₂ formation.

Fig. 5.

Potential energy surface of singlet [Ru^IV-OO]²⁺ and the corresponding triplet energies with optimal geometries from singlet using B3LYP/LANL2DZ.

O₂ Release from the Open Structure of [Ru^IV-OO]²⁺ by Water Attack in Pathway 1.

O₂ is released from [Ru^IV-OO]²⁺ to return to [Ru^II-OH₂]²⁺, which reenters the catalytic cycle for water oxidation. As shown in Fig. S4, by scanning the potential energy surfaces for singlet and triplet spin states with respect to the bond distance between Ru and the second oxygen atom (i.e., dRu-O₂), we believe that the open structure of [Ru^IV-OO]²⁺ at the singlet spin state cannot release O₂ since the activation barrier is too high (> 35 kcal/mol). This is further confirmed by the accurate QM/MM-MFEP simulations shown in Fig. 6. By contrast, for the triplet spin state, O₂ can be released with the low barrier of ∼5 kcal/mol estimated by the potential energy scan in Fig. S4. Based on QM/MM simulations with explicit water molecules in Fig. 6 and the optimized structures in Fig. S5, when one water molecule attacks the transition metal center to release O₂, the accurate activation barrier is 12 kcal/mol for the triplet spin state. Therefore, our theoretical studies elucidate that another rate-limiting step in water oxidation [besides O---O bond formation (21)] is to release O₂ from the open structure of [Ru^IV-OO]²⁺ from the triplet spin state. This high spin state is required since the ground spin state of O₂ in the final product is a triplet.

Fig. 6.

The reaction profile computed by QM/MM-MFEP approach to release O₂ for the open form of [Ru^IV-OO]²⁺ with both singlet and triplet spin states. The QM subsystems are computed by B3LYP/LANL2DZ.

O₂ Release from [Ru^V-OO]³⁺ after [Ru^IV-OO]²⁺ Is Oxidized in Pathway 2.

At pH = 0, our experiments (11, 17) found that [Ru^IV-OO]²⁺ is oxidized first and then releases O₂. The redox potential calculations (Figs. S6 and S7) of closed and open structures of [Ru^IV-OO]²⁺ at the singlet spin state suggest the following: (i) both oxygen atoms bind to Ru with the closed structure at pH = 0.0; (ii) this oxidation process involves spin flipping since [Ru^V-OO]³⁺ can only be stable in the quartet spin state; (iii) the computed redox potential (E⁰ = 1.7 V) of the closed form of singlet [Ru^IV-OO]²⁺ and quartet [Ru^V-OO]³⁺ agrees with experimental observations (1.7 V) (17). Note that the O---O distance is shortened to 1.265 Å after [Ru^IV-OO]²⁺ is oxidized to [Ru^V - OO]³⁺. In this form, no barrier is required to release O₂ when one water molecule attacks Ru to form [Ru^III-OH]²⁺ and release a proton. [Ru^III-OH]²⁺ can be rapidly oxidized to [Ru^IV-OO]²⁺ through PCET as shown in Fig. 2. This suggests that the electron-transfer process is a rate-limiting step at pH = 0.0 rather than the O₂ releasing step in Pathway 2.

Conclusions

Although water oxidation catalyzed by the single-site Ru catalyst is complicated, microscopic details are elucidated clearly by our theoretical computations. Even though calculations have associated errors in pK_a and redox potential values, the computed thermochemical pathways for different Ru oxidation states are still helpful in explaining in qualitative detail how proton-transfer and electron-transfer processes occur at the atomistic level. Two pathways at different pH values share the same rate-limiting step [i.e., concerted oxygen atom-proton transfer (APT) to oxidize water, which was clarified by our previous work (21)]. However, in Pathway 1, O₂ release is a key step in determining catalytic rates. This step is also complicated by spin states and conformation changes in [Ru^IV-OO]²⁺. In Pathway 2, the rate-limiting step is the oxidation of [Ru^IV-OO]²⁺ to [Ru^V-OO]³⁺ rather than O₂ release. Spin-orbital coupling effects are crucial in both pathways in bringing the system to a high spin state in order to release triplet O₂. When the binding ligands are modified, the rate-limiting steps can be changed as well.

Based on our computations on this heavy transition metal system, we conclude that several issues need to be addressed before theoretical modeling can be helpful in further tuning existing catalysts or designing future catalysts: (i) accurate but fast methods to predict pK_a and redox potentials as well as reaction barriers; and (ii) affordable approaches to computing relativistic effects. With the aid of QM/MM-MFEP, we believe that computations of redox potential (29) and reaction barriers (30) are affordable now. However, the approximated functionals might impair accuracy. Further theoretical progress is needed to make computational modeling more accurate.

Materials and Methods

All QM calculations with IEFPCM were performed using Gaussian 09 program (31). The QM/MM-MFEP simulations were carried out using our in-house Sigma/G03 (30, 32). The B3LYP/LANL2DZ scheme with implicit solvent water model (i.e., IEFPCM) was applied for geometries optimizations, potential energy scan, the computations of redox potential and pK_a. All the spin states of ruthenium intermediates were identified by MP2/LANL2DZ computations. The QM/MM-MFEP approach with explicit water molecules was employed to obtain the accurate activation barriers of key reactions steps (16, 33–35) (SI Text).