Abstract
In recent years, photocatalytic hydrogenation of metal nitrides to ammonia using solar energy as a renewable energy source and green hydrogen as both an electron and a proton source under ambient reaction conditions has been intensively investigated to achieve one of the most desirable next-generation scientific techniques, i.e., atom-economical ammonia production from dinitrogen and dihydrogen, thereby contributing to a carbon-neutral society. In this study, we have successfully developed visible-light-driven photocatalytic hydrogenation of manganese nitride complexes to ammonia at room temperature under an ambient pressure of dihydrogen in the presence of diruthenium complexes as dihydrogen oxidation catalysts and iridium complexes as photosensitizers. The turnover numbers based on the diruthenium complex and the yield of ammonia based on the manganese nitride complex reached up to 150 equiv. and 91%, respectively, indicating the high durability and efficiency of this photocatalytic system. Moreover, this hydrogenation proceeded photocatalytically even at a significantly low partial pressure of dihydrogen (down to 0.05 atm), which was much lower than that in previous systems reported by other research groups, and would be a key process for paving the way for achieving photocatalytic ammonia formation under an ambient pressure of a mixture of dinitrogen and dihydrogen.
Herein, we report a durable and efficient visible-light-driven catalytic hydrogenation of manganese nitride complexes to NH3 under ambient reaction conditions using H2 as both electron and proton sources and a combination of two molecular catalysts.
Introduction
Ammonia is an essential nitrogen source for fertilizer and nitrogen-containing chemicals, supporting various industrial and human activities.1 In recent years, the potential use of ammonia as a promising energy carrier and fuel for a carbon-neutral society has attracted more attention.2–6 Currently, ammonia is industrially produced using the Haber–Bosch process from dinitrogen and dihydrogen (N2 + 3H2 → 2NH3). This process is one of the most desirable synthetic methods from the viewpoint of atom economy. However, the Haber–Bosch process requires high temperatures and pressures, resulting in significant energy consumption. Therefore, new environmentally friendly methods for ammonia production from dinitrogen and dihydrogen (green hydrogen in the future) that can proceed under mild reaction conditions and utilize renewable energy (typically solar energy) have been sought.
In addition to the solid thermal catalysts in the Haber–Bosch process, various photoresponsive heterogeneous photocatalysts have been widely developed to achieve solar-light-driven ammonia synthesis from dinitrogen and dihydrogen.7–10 In addition to the heterogeneous catalysts, numerous homogeneous systems have been developed and investigated because homogeneous reaction systems are suitable for gaining molecular and mechanistic insights into the challenging goal of ammonia synthesis from dinitrogen and dihydrogen under mild reaction conditions. The first important concept in this research area was proposed in 1991, in a report on the stoichiometric protonation of a tungsten–dinitrogen complex with dihydrogen as the proton source, using the strong acidity of a ruthenium dihydrogen complex to form the corresponding tungsten hydrazido(2-) complex and ruthenium hydride complex.11 In 1998, one of the authors of the present paper found that stoichiometric reactions between tungsten–dinitrogen complexes and an excess of ruthenium–dihydrogen complexes at 55 °C under 1 atm of dihydrogen afforded ammonia, formed via protonation of the coordinated dinitrogen with the coordinated dihydrogen.12–14 This is one remarkable milestone as the first successful example of ammonia formation from dinitrogen and dihydrogen under mild reaction conditions. Later, Chirik and coworkers reported the hydrogenation of dinitrogen-bridged dizirconium complexes with dihydrogen to afford a stoichiometric amount of ammonia, using both electrons and protons derived from dihydrogen, as another unique and important example.15–18 The tri(iron)bis(nitrido) complex ([(Cp′Fe)3(µ3-N)2], Cp′ = η5-1,2,4-(Me3C)3C5H2), which was synthesized by reduction of [Cp′Fe(µ-I)]2 under a dinitrogen atmosphere, is also an interesting example, because of its reactivity with dihydrogen (4 bar) at ambient temperature to form a stoichiometric amount of ammonia (3–7% yield).19 Recently, the first example of nitrogen–nitrogen bond cleavage and subsequent hydrogenation with dihydrogen to form NH3 using a molecular uranium complex has been reported.20
Another useful reported strategy is the (photo)catalytic hydrogenation of transition-metal nitride complexes,21–26 which are known to be key intermediates produced via nitrogen–nitrogen triple bond cleavage of dinitrogen in efficient catalytic nitrogen fixation.27–43 Although in most reported systems, dinitrogen was not used directly as a nitrogen source for nitride complexes, several (photo)catalytic systems for hydrogenation of metal nitride complexes to ammonia using dihydrogen as both an electron and a proton source have been reported over the last decade. In fact, Chirik and co-workers reported thermal catalytic hydrogenation of the manganese(v) nitride complex bearing a salen ligand (1a) as a model metal nitride complex at room temperature under 4 atm of dihydrogen in the presence of rhodium hydride complexes to produce a stoichiometric amount of ammonia (Scheme 1a).44 It should be noted that although 1a cannot be produced by the activation of dinitrogen, 1a can produce ammonia by hydrogen atom transfer (HAT)45,46 or proton-coupled electron transfer (PCET),47–49 as with the nitride complexes in the (photo)catalytic cycle for ammonia formation from dinitrogen. The bond dissociation free energy of the Rh–H bond (BDFERh–H) was estimated to be 51 kcal mol−1, which is very close to the theoretical minimum limitation of the BDFE values towards ammonia formation because the value of ΔGf°(H·) to split the hydrogen–hydrogen bond in dihydrogen is 52.0 kcal mol−1. On the other hand, the BDFEN–H value of the imide complex produced by hydrogenation of 1a was estimated to be 70 kcal mol−1,44,50 indicating that the suitable BDFE value of the Rh–H bond was effective for both hydrogen–hydrogen bond cleavage and hydrogenation of 1a. However, likely due to the small driving force of hydrogen–hydrogen bond cleavage, the reaction rate was slow (over 5 days) to reach 40% yield at room temperature and 60% yield at 50 °C. Besides, the turnover number for ammonia production based on the Rh atom in the catalyst was not high (4–6 equiv.).
Scheme 1. (a) Thermal catalytic and (b) photocatalytic hydrogenation of metal nitride complexes under 4 atm of dihydrogen. (c) Catalytic dihydrogen oxidation using a diruthenium complex. (d) Visible-light-driven ruthenium-catalyzed hydrogenation of the manganese nitride complex developed in this study.
Chirik and co-workers also reported photocatalytic hydrogenation of a molybdenum nitride complex in the presence of iridium hydride complexes as photocatalysts under 4 atm of dihydrogen and visible-light irradiation to produce a stoichiometric amount of ammonia (Scheme 1b).51 Although the BDFE value of the Ir–H bond in the ground state is approximately 60 kcal mol−1, that in the excited state is estimated to be very small (approximately 10 kcal mol−1), which is enough to trigger the photochemical HAT (or PCET) to the molybdenum complex (BDFEN–H = 21 kcal mol−1). This photocatalytic HAT reaction has also been applied for the catalytic formation of weak chemical bonds in various organic compounds.52 Notably, this molybdenum nitride complex was successfully reproduced after ammonia formation via several stepwise reactions; therefore, this reaction system is the first successful example of achieving the synthetic cycle for ammonia formation from dinitrogen and dihydrogen, though the ammonia yield based on the molybdenum nitride complex and the turnover number of ammonia formation based on the iridium complex were only 38% and 2 equiv., respectively. By changing the photocatalyst to [Ir(ppy)3]-type complex (ppy = 2-phenylpyridinato), the ammonia yield became much higher (87% under 4 atm H2 or 75% under 1 atm H2), and the turnover number of the molybdenum complex for ammonia formation from dinitrogen and dihydrogen reached 1.1 by repeating the synthetic cycle.53
One possible reason for the low turnover numbers of dihydrogen activation catalysts in the aforementioned systems is the hydrogenation of ligands, resulting in structural changes in the catalyst or ligand dissociation (e.g., photodriven C–H reductive elimination in the hydride complexes (Scheme 1b));44,54 therefore, the use of stable catalysts for dihydrogen activation should be necessary to achieve durable catalytic systems. Our group previously reported highly durable and efficient catalytic oxidation of dihydrogen under ambient reaction conditions using thiolate-bridged diruthenium complexes (Scheme 1c).55–57 In this reaction, the monocationic hydride complex ([2a-H]+), which is produced by the coordination of dihydrogen to the dicationic complex ([2a]2+) with a vacant site and subsequent deprotonation, provides an electron, a proton, and an electron in that order to reproduce [2a]2+. As a result, dihydrogen can be used as both a two-electron and a two-proton donor to reduce oxidants (e.g., ferrocenium salts) and protonate bases (e.g., pyridine derivatives), respectively. Notably, the maximum turnover number for the reduction of the ferrocenium cation exceeded 9400 equiv. based on [2a-H]+. Moreover, this dihydrogen oxidation with [2a-H]OTf proceeded efficiently even under 0.37 atm of dihydrogen, suggesting that [2a-H]OTf can supply electrons and protons derived from dihydrogen with high durability and efficiency even under low partial pressure of dihydrogen.
Based on these research findings, we have envisaged a new concept that uses our unique dihydrogen activation method to catalytically reduce metal nitride complexes to ammonia (Scheme 1d). In this concept, two electrons and two protons derived from dihydrogen are transferred separately to 1a by a photosensitizer and a proton mediator, respectively, unlike the reported systems with hydride complexes in Scheme 1a and b. In fact, we have succeeded in developing photocatalytic hydrogenation of the manganese nitride complex 1a as a typical model metal nitride complex in the presence of the diruthenium complex [2a-H]OTf as a dihydrogen oxidation catalyst. Since the reducing power of [2a-H]OTf is not enough to donate electrons directly to 1a (or its derivatives), we have newly designed a combination system consisting of a catalytic cycle for oxidizing dihydrogen with [2a-H]OTf to provide electrons and protons and a photosensitizing cycle with Ir(ppy)3 (3a) as a redox photosensitizer to increase the electron-donating ability and trigger hydrogenation of 1a using electrons and protons derived from dihydrogen via photochemical PCET or HAT processes under ambient reaction conditions and visible light irradiation, resulting in up to 150 equivalents of the turnover number for ammonia formation based on [2a-H]OTf and 91% ammonia yield based on 1a. Moreover, this hydrogenation proceeded photocatalytically even under 0.05 atm of dihydrogen, demonstrating a useful feature for achieving photocatalytic ammonia formation under ambient pressure of a mixture of dinitrogen and dihydrogen.
Results and discussion
As a typical run, we performed the photocatalytic hydrogenation of the manganese nitride complex 1a as a model metal nitride complex (100 µmol) in the presence of catalytic amounts of the diruthenium complex [2a-H]OTf (1 µmol) as a dihydrogen oxidation catalyst, the iridium complex 3a (10 µmol) as a photosensitizer, and 2,6-lutidinium triflate ([LutH]OTf, 25 µmol) as a proton mediator in tetrahydrofuran (THF, 5 mL) under visible light irradiation (>400 nm) and 1 atm of dihydrogen at room temperature for 24 h (Fig. 1; Table 2, entry 1). After photoirradiation, the solution color changed from green to orange (Fig. S2 in the SI). The amount of produced ammonia reached 80 ± 4 µmol (i.e., 80 equiv. based on [2a-H]OTf in 80% yield based on 1a). To obtain 80 µmol of ammonia, 240 µmol (=80 × 3 µmol) of protons and electrons are required; hence, the turnover number of the proton mediator and the photosensitizer was larger than 9 and 24, respectively. Fig. 2a illustrates the time course of the amount of ammonia produced in the photocatalysis. During the reaction, the reaction rate of ammonia formation gradually decreased and reached a plateau in 24 h of irradiation (Fig. 2a, Table S1 in the SI). A light-on/off experiment under the same reaction conditions revealed that ammonia formation proceeded only during photoirradiation, and a radical chain reaction should not have proceeded in this reaction system (Fig. 2b, Table S6 in the SI). The use of the 15N-labeled manganese nitride complex (15N-1a) under the same reaction conditions afforded 15N-labeled ammonia (15NH3) in 76% yield, confirming that ammonia was produced from hydrogenation of the nitride moiety of 1a (Fig. S3 in the SI). The apparent quantum yield of the ammonia formation (ΦNH3) in the initial stage of the photoreaction was estimated to be 0.86% using 405 nm monochromatic light and a ferrioxalate chemical actinometer58 (Table S8 in the SI). The relatively low quantum yield is mainly due to the strong inner-filter effect59 by 1a and unexpected charge recombination pathways (vide infra). The photocatalysis under D2 for 12 h (18% yield, Table S7 in the SI) exhibited a much lower reaction rate than under H2 (54% yield). The kinetic isotope effect (KIE) was estimated to be 3.1, suggesting that proton or hydrogen-atom transfer and/or reaction with dihydrogen should be included among the rate-determining steps (see below).
Fig. 1. Photo-driven ruthenium-catalyzed hydrogenation of the manganese nitride complex.
Table 2. Photocatalytic hydrogenation of manganese nitride complexes under 1 atm of H2 in the presence of catalysts, photosensitizers, and proton mediators.
| |||||||
|---|---|---|---|---|---|---|---|
| Entry | Nitride complex | Catalyst | Photosensitizer | Proton mediator | Solvent | NH3 production [µmol] | NH3 yield [%] |
| 1 | 1a | [2a-H]OTf | 3a | [LutH]OTf | THF | 80 ± 4 | 80 ± 4 |
| 2a | 1a | [2a-H]OTf | 3a | [LutH]OTf | THF | 110 ± 3 | 55 ± 2 |
| 3b | 1a | [2a-H]OTf | 3a | [LutH]OTf | THF | 150 ± 4 | 75 ± 2 |
| 4 | 1a | [2a](OTf)2 | 3a | [LutH]OTf | THF | 83 ± 2 | 83 ± 2 |
| 5 | 1a | [2b](OTf)2 | 3a | [LutH]OTf | THF | 33 ± 2 | 33 ± 2 |
| 6 | 1a | [2c](OTf)2 | 3a | [LutH]OTf | THF | 14 ± 1 | 14 ± 1 |
| 7 | 1a | [2d]OTf | 3a | [LutH]OTf | THF | 8 ± 0.5 | 8 ± 0.5 |
| 8c | 1a | [2a-H]OTf | 3a | [LutH]OTf | THF | 81 ± 2 | 81 ± 2 |
| 9d | 1a | [2a-H]OTf | 3a | [LutH]OTf | THF | 46 ± 5 | 46 ± 5 |
| 10 | 1a | [2a-H]OTf | 3b | [LutH]OTf | THF | 56 ± 5 | 56 ± 5 |
| 11 | 1a | [2a-H]OTf | 3c | [LutH]OTf | THF | 14 ± 1 | 14 ± 1 |
| 12 | 1a | [2a-H]OTf | [3d]ONf | [LutH]OTf | THF | 24 ± 5 | 24 ± 5 |
| 13 | 1a | [2a-H]OTf | [3e](ONf)2 | [LutH]OTf | THF | 1 ± 0.1 | 1 ± 0.1 |
| 14 | 1a | [2a-H]OTf | 3a | [ColH]OTf | THF | 69 ± 4 | 69 ± 4 |
| 15 | 1a | [2a-H]OTf | 3a | [PicH]OTf | THF | 65 ± 2 | 65 ± 2 |
| 16e | 1a | [2a-H]OTf | 3a | [LutH]OTf | THF | 64 ± 1 | 64 ± 1 |
| 17f | 1a | [2a-H]OTf | 3a | [LutH]OTf | THF | 52 ± 1 | 52 ± 1 |
| 18g | 1a | [2a-H]OTf | 3a | — | THF | 21 ± 0.2 | 21 ± 0.2 |
| 19 | 1a | [2a-H]OTf | 3a | [LutH]OTf | Benzene | 20 | 20 |
| 20 | 1a | [2a-H]OTf | 3a | [LutH]OTf | CH2Cl2 | 19 | 19 |
| 21 | 1a | [2a-H]OTf | 3a | [LutH]OTf | Et2O | 6 | 6 |
| 22 | 1b | [2a-H]OTf | 3a | [LutH]OTf | THF | 58 ± 2 | 58 ± 2 |
| 23 | 1c | [2a-H]OTf | 3a | [LutH]OTf | THF | 10 ± 1 | 10 ± 1 |
1a (200 µmol) was used.
The amounts of 1a and [LutH]OTf were changed to 200 and 50 µmol, respectively.
[2a-H]OTf (2 µmol) was used.
[2a-H]OTf (0.5 µmol) was used.
[LutH]OTf (50 µmol) was used.
[LutH]OTf (10 µmol) was used.
[LutH]OTf was not used.
Fig. 2. (a) The time course of the amount of ammonia produced in the photocatalysis. (b) The light on/off experiment.
Several control experiments shown in Table 1 confirmed that the combination of 1a, [2a-H]OTf, 3a, [LutH]OTf, dihydrogen, and visible light was required for efficient ammonia formation. It should be noted that photocatalytic hydrogenation proceeded slowly even in the absence of [LutH]OTf as a proton mediator (entry 5 in Table 1), likely because protons were directly donated from the diruthenium complex to 1a in the PCET process due to the high acidity of the intermediates (see below). On the other hand, in the case without [2a-H]OTf (entry 3 in Table 1), the amount of ammonia was low, though protons can be provided by [LutH]OTf (i.e., 25 µmol of protons). This result clearly demonstrates that, under the conditions of the control experiment, the sole electron source was 3a (i.e., 10 µmol of 3a can produce at most 3.3 µmol of ammonia), whereas under standard conditions, electrons were catalytically supplied via the dihydrogen oxidation reaction with [2a-H]OTf.
Table 1. Results of control experiments.
| ||
|---|---|---|
| Entry | Deviations from standard conditions | NH3 yielda,b [%] |
| 1 | — | 80 ± 4 |
| 2 | Without 1a | <1 |
| 3 | Without [2a-H]OTf | 5 ± 1 |
| 4 | Without 3a | 1 ± 1 |
| 5 | Without [LutH]OTf | 21 ± 1 |
| 6 | Without light | 3 ± 1 |
| 7 | Without H2 (under Ar) | 7 ± 1 |
The averages are shown with standard errors.
Yield based on 1a.
The detailed structure of the resulting manganese complex after hydrogenation of 1a has not yet been reported; therefore, we isolated the manganese complex by removing the solvent from the reaction solution under reduced pressure, extracting it from the obtained residue with hexane, and then filtering off the insoluble materials including catalysts and [LutH]OTf (yield: 93%, see Section 2.7 in the SI). Single crystals were successfully prepared by slow evaporation of the hexane solution, and the X-ray analysis revealed that the product was the dinuclear manganese complex (4) bridged by one of the oxygen atoms in each salen ligand (Fig. 3a, Table S10 in the SI). The elemental analysis also confirmed the composition formula of 4 and the purity of the sample. The manganese nitride complex 1a was regenerated from 4 using the synthetic method reported by Chirik et al. in a moderate yield (Fig. 3b, S7, and S8 in the SI); therefore, 4 should be the same as or similar to the manganese species obtained after the photoreactions in the aforementioned reported systems.44,50,60 However, 4 exhibited no signal in the electron spin resonance (ESR) spectrum (Fig. S9 in the SI), in contrast to the aforementioned reported systems, which showed a strong isotropic signal with a g value of 2.0.44 The 1H NMR spectrum of 4 also indicated its weak magnetism because of the relatively small difference in chemical shifts from those of diamagnetic complex 1a (Fig. S4 in the SI). These results suggest that 4 can be assigned as a homovalent antiferromagnetically coupled dimer.61,62 Quite interestingly, we found that the ESR spectrum of the solution containing 4 and [LutH]OTf (Fig. S9 in the SI) contained a characteristic signal containing 6 lines, which is a typical signal shape for MnII species because of the interaction between the 3d5 electronic spin (S = 5/2) and the nuclear spin of the 55Mn nucleus (I = 5/2).63 The UV-vis absorption spectrum of 4 was also changed drastically by adding [LutH]OTf (Fig. S10 in the SI). In the case of the solution after the photoreaction shown in Fig. 1, quite weak but similar signals were also observed in the ESR spectrum, while its UV-vis absorption spectrum was well simulated with the spectrum of 4, meaning that 4 in the reaction solution had slightly reacted with [LutH]OTf (Fig. S5 in the SI). Given that the addition of Lut did not affect both ESR and UV-vis absorption spectra (Fig. S9 and S10 in the SI), protonation of 4 with [LutH]OTf to form monomeric MnII species might be involved in the structural changes, even though the resulting molecular structure is unclear at this stage. The simulation spectra and further discussion on the spectrum structure are included in Section 2.10 in the SI.
Fig. 3. (a) An ORTEP drawing of the resulting dimanganese complex 4 obtained after photocatalytic hydrogenation. Thermal ellipsoids are shown at the 50% level. Hydrogen atoms and solvent molecules are omitted for clarity, and the main places in the disordered regions are illustrated. White, blue, purple, and green correspond to C, O, N, and Mn atoms, respectively. (b) Reproduction of 1a from 4.
The residue obtained after extracting 4, as described above, was further investigated. In the 1H NMR spectrum of the residue, [2a-H](OTf) (1.0 µmol) and 3a (10 µmol) were observed, meaning that the decomposition of the catalysts was not significant even after the 24 hour reaction (Fig. S6 in the SI). Therefore, we increased the amount of 1a from 100 to 200 µmol. It is noted that this concentration of 1a is almost at saturation. The amount of produced ammonia was increased to 110 µmol, while the yield was decreased to 55% (Table 2, entry 2). When simultaneously increasing the amount of [LutH]OTf to 50 µmol, the amount and yield of produced ammonia were further increased to 150 µmol and 75%, respectively (Table 2, entry 3). This is probably because of the high concentration of the resulting product 4 and the aforementioned interaction between [LutH]OTf and 4, which led to the suppression of the function of [LutH]+ as a proton mediator.
Next, to optimize the reaction conditions, we conducted the photocatalytic hydrogenation of 1a using various catalysts for dihydrogen oxidation (Table 2; entries 1, 4–7). The time courses of the amount of ammonia produced during the photocatalytic reactions are summarized in Fig. 4 and Tables S1–S5 in the SI. When using the corresponding dicationic complex [2a](OTf)2, which possesses a vacant site, both the reaction rate and the amount of produced ammonia became almost the same as those using [2a-H](OTf). This result clearly indicates that [2a-H](OTf) and [2a](OTf)2 were efficiently and mutually converted in the catalytic cycle, as with the aforementioned system. A Ru–Fe dinuclear complex ([2b](OTf)2) also functioned as an effective catalyst in the initial stage, even though the reaction ceased within 6 h of irradiation, likely because of the lower durability of the Fe unit. In the photocatalytic reactions using [2c](OTf)2 or [2d]OTf, which possess two thiolate ligands, both the reaction rate in the initial stages and the amount of produced ammonia became much lower than those using [2a-H](OTf). From these results, the multidentate thiolate ligand, including the triaryl phosphine moiety, is suitable for improving catalyst durability. Further discussion with cyclic voltammetry and photophysical properties is included in Sections 4 and 5 in the SI. The increase in the amount of [2a-H](OTf) (2 µmol) used hardly affected the photocatalytic activity (81% yield, Table 2, entry 8), while the decrease in the amount of [2a-H](OTf) (0.5 µmol) used lowered the photocatalytic activity (46% yield, Table 2, entry 9), meaning that under the optimized conditions, the catalytic cycle with [2a-H](OTf) may not be included in the rate-determining steps and 1 µmol or a larger amount of [2a-H](OTf) is required for the efficient reaction procedures (Fig. S11a in the SI).
Fig. 4. The time courses of the amount of ammonia produced in photocatalysis using various thiolate-bridged dinuclear complexes ([2a-H]OTf (blue circle), [2a](OTf)2 (red diamond), [2b](OTf)2 (green triangle), [2c](OTf)2 (orange circle), and [2d]OTf (sky blue square)).
The substituent effect on the iridium photosensitizer was also significant. When introducing electron-withdrawing fluorine groups (i.e., 3b and 3c), the catalytic activity decreased compared with 3a (Table 2; entries 1, 10, 11). These results suggest that high reducing power is crucial for effective electron transfer to manganese nitride complexes (or proton mediators), thereby achieving high catalytic activity. Of note, their reducing power in the excited state (oxidation potential in the excited state (E*ox)) was estimated based on oxidation potentials determined by cyclic voltammetry and excitation energies determined by Franck–Condon analysis64 as mentioned below (E*ox(3a): −2.15 V, E*ox(3b): −2.05 V, and E*ox(3c): −1.91 V; Table S16 in the SI). On the other hand, [3d]ONf (E*ox: −1.39 V) and [3e](ONf)2 (E*ox: −1.14 V), whose reducing power in the excited states is relatively weak, did not function as effective photosensitizers (Table 2; entries 12 and 13). In the case using [3d]ONf or [3e](ONf)2, according to their redox potentials, they possibly donate electrons to 1avia reductive quenching, which is an electron transfer from [2a-H]OTf (Eox: +0.24 V) to the excited state of a photosensitizer, due to the relatively high oxidizing power in their excited state (E*red([3b]ONf): +0.34 V; E*red([3e](ONf)2): +0.21 V) (Table S16 in the SI). However, their high oxidizing power might trigger the oxidation of 1a (Eox: +0.26 V) as a side reaction, thereby suppressing both the reduction of 1a and the photosensitizing reaction by charge recombination. Consequently, these screening results suggest that using a photosensitizer with relatively low photo-oxidizing power to prevent the unexpected oxidation of 1a and high reducing power to provide electrons to 1a could be key to achieving high catalytic activity. All of the electrochemical and photophysical properties of the photosensitizers are summarized in Fig. S28 and Table S16 in the SI.
Not only the catalysts and photosensitizers but also proton mediators, reaction solvents, and the ligand of the manganese nitride complex were optimized. The photocatalytic reaction using [LutH]OTf (80% yield) showed the highest photocatalytic activity, but [ColH]OTf (69% yield) and [PicH]OTf (65% yield) also functioned as effective proton mediators (Table 2; entries 1, 14, 15). The cyclic voltammetry suggested that all of these proton mediators (Ered([LutH]OTf): −1.91 V, Ered([ColH]OTf): −2.08 V, and Ered([PicH]OTf): −1.85 V) can accept electrons exergonically from the excited state of 3a (3a*, E*ox: −2.15 V) (vide infra). The electrochemical properties of the proton mediators are summarized in Sections 4 and 5 in the SI. Although the photocatalytic reaction slowly proceeded even in the absence of [LutH]OTf (21% yield), the increase in the amount of [LutH]OTf to 10 µmol (52% yield) or 25 µmol (80% yield) led to higher photocatalytic activities (Table 2; entries 1, 17–18). The use of an even higher amount of [LutH]OTf (50 µmol) led to a similar yield to that with 25 µmol (64% yield, Table 2; entry 16). These results clearly suggest that the proton transfer from the diruthenium complexes to manganese nitride complexes was largely promoted by the pyridinium salts as proton mediators (vide infra).
When changing the solvent from THF (80% yield) to benzene (20% yield), CH2Cl2 (19% yield), or Et2O (6% yield), the photocatalytic activities decreased drastically (Table 2; entries 1, 19–21), meaning that THF is a suitable solvent for this system. It is noted that, especially when using benzene and Et2O, the solubility of 3a was low, resulting in low photosensitizer concentrations and decreased photocatalytic activity (Fig. S2d in the SI). The low solubility of 1a in Et2O might also reduce the efficiency of the photosensitizing process due to its low solution concentration and light scattering.
The effect of substituents on the salen ligand of the manganese nitride complexes also affected the catalytic activity. When introducing MeO groups (1b) or Cl groups (1c) instead of tBu groups on the phenyl rings in the salen ligand, the yields of ammonia became 58% and 10%, respectively (Table 2, entries 22 and 23). One of the main reasons for lower yields when using 1b or 1c is their lower solubility, which leads to suspension and light scattering (Fig. S2d in the SI). The electrochemical studies revealed that the electron-accepting abilities of 1b (Ered: −2.55 V) and 1c (Ered: −2.30 V) were similar to or higher than that of 1a (Ered: −2.60 V), indicating that the enhancement of electron-accepting ability by changing the substituents on the salen ligand did not effectively improve the catalytic activity. In the case using 1b, the red shift of the absorption band might have increased the inner-filter effect,59 which suppresses the absorption of the irradiated photons by 3a (Fig. S14 in the SI), thereby reducing the efficiency of the photoreaction (vide infra). All of the electrochemical and photophysical properties of the manganese complexes are summarized in Sections 4 and 5 in the SI.
As described above, the manganese nitride complex was successfully hydrogenated to produce ammonia photocatalytically under 1 atm of dihydrogen as both an electron and a proton source at room temperature by using a combination of [2a-H](OTf), 3a, and [LutH]OTf. We next conducted photocatalytic hydrogenation under a lower partial pressure of dihydrogen (Tables 3 and S9 in the SI). The photocatalytic hydrogenation of 1a proceeded even under 0.2 and 0.05 atm, and the yields of ammonia were 37% and 19%, respectively (Table 3, entries 2 and 4). The reaction rates linearly depended on the partial pressure of dihydrogen (Fig. S11b in the SI). When increasing the amount of [2a-H](OTf) from 1 µmol to 8 µmol, the amount of ammonia produced under 0.2 and 0.05 atm of dihydrogen also increased to 91% and 39%, respectively (Table 3, entries 3 and 5). These results indicate that under 0.2 or 0.05 atm of dihydrogen, the rate-determining step might have switched to the catalytic cycle for dihydrogen oxidation with [2a-H]OTf. It should be noted that the partial pressure of dihydrogen in the aforementioned system is much lower than that in the reported photocatalytic systems (1–4 atm).44,51,53 Therefore, the photocatalytic system developed in this study can effectively utilize much lower concentrations of dihydrogen.
Table 3. Photocatalytic hydrogenation of manganese nitride complexes under various partial pressures of H2 in the presence of [2a-H]OTf, 3a, and [LutH]OTfa.
| |||
|---|---|---|---|
| Entry | H2 pressure [atm] | Amount of [2a-H]OTf [µmol] | NH3 yield [%] |
| 1 | 1 | 1 | 80 ± 4 |
| 2 | 0.2 | 1 | 37 ± 2 |
| 3 | 0.2 | 8 | 91 ± 2 |
| 4 | 0.05 | 1 | 19 ± 2 |
| 5 | 0.05 | 8 | 39 ± 1 |
Photocatalytic reactions were performed under a mixture of H2 (X [atm]) and N2 (1 − X [atm]). The summation pressure was 1 atm.
To gain insight into the reaction mechanism of this photocatalytic system under the optimized reaction conditions, we investigated the photophysical and photochemical properties of the catalysts and substrates. Fig. 5a illustrates the UV-vis absorption spectra of 1a, [2a-H]OTf, and 3a in THF. Photosensitizer 3a showed absorption bands up to approximately 500 nm. Notably, 1a and [2a-H]OTf showed strong absorption bands in a similar region; therefore, the absorption of the irradiation light by 3a will be suppressed by the inner filter effect derived from 1a and [2a-H]OTf. [LutH]OTf did not show any absorption bands in the visible regions. Fig. 5b illustrates the absorption spectrum simulated using the molar extinction coefficients (ε) as shown in Fig. 5a and the concentrations in the reaction solutions under the optimized conditions, clearly demonstrating that due to the quite high concentration of 1a, the photons irradiated to the reaction solutions were absorbed mainly by 1a. The main irradiation wavelengths absorbed by 3a were 405 and 430 nm, and the ratios of the photons with these wavelengths absorbed by 3a were estimated to be only 11%. These low absorption ratios should be one of the reasons for the low quantum efficiency of the photocatalytic reactions. It is noteworthy that the dimanganese complex 4, which is a product of the photocatalytic hydrogenation of 1a, also exhibited relatively strong visible-light absorption ability (Fig. S14b in the SI), meaning that the inner filter effect derived from 4 is also considerable.
Fig. 5. (a) UV-vis absorption spectra of 1a (blue), [2a-H]OTf (red), and 3a (orange) in THF. The spectrum simulated using molar extinction coefficients and concentrations in reaction solutions for photocatalytic reactions is illustrated in (b). The concentrations of 1a, [2a-H]OTf, and 3a are 20, 0.2, and 2 mM, respectively.
To investigate the behavior after photoexcitation, the photochemical properties of 3a were investigated using photoluminescence quenching experiments, i.e., Stern–Volmer analyses (Fig. 6, Table 4; Fig. S20 and Table S12 in the SI). It should be noted that the Stern–Volmer analyses were conducted using the emission lifetimes in the presence of quenchers (τ) and in their absence (τ0) because 1a and [2a-H]OTf showed strong absorption bands in the visible regions (Fig. 5a) to absorb the photoluminescence from 3a in the excited state (3a*) resulting in the overestimation of the degree of photoluminescence quenching determined from emission intensities. The photoluminescence from 3a* was effectively quenched by 1a, [2a-H]OTf, or [LutH]OTf, and the Stern–Volmer constants (KSVs) were respectively determined to be 2.4, 11, and 9.3 [mM−1] from the slopes of the fitting curves of Stern–Volmer plots in Fig. 6. The quenching rate constants (kQ) calculated from the KSV values and τ0 were 1.6 × 109, 7.6 × 109, and 6.4 × 109 [M−1 s−1], respectively. These values are close to the diffusion rate constants (≈1010 [M−1 s−1]), indicating relatively high quenching efficiency. Given the KSV values and their concentrations under the optimized reaction conditions, the quenching fractions (ηq) with 1a, [2a-H]OTf, or [LutH]OTf were estimated to be 49%, 2%, and 48%, respectively. These results indicate that the excited state of 3a is mainly quenched by 1a or [LutH]OTf (see Section 4.3 in the SI).
Fig. 6. Stern–Volmer luminescence quenching studies of 3a with 1a (blue), [2a-H]OTf (red), or [LutH]OTf (green).

Table 4. Quenching experiments with 3a in the presence of 1a, [2a-H]OTf, or [LutH]OTf.
| Quencher | K SV a [mM−1] | k Q b [M−1 s−1] | η q c [%] |
|---|---|---|---|
| 1a | 2.4 | 1.6 × 109 | 49 |
| [2a-H]OTf | 11 | 7.6 × 109 | 2 |
| [LutH]OTf | 9.3 | 6.4 × 109 | 48 |
The Stern–Volmer constants determined from the slope of the fitting curves illustrated in Fig. 6.
The quenching rate constants determined from KSV and the emission lifetime of 3a in the absence of any quencher (τ0, KSV = kQτ0).
The quenching fractions in the reaction solution calculated based on eqn S7 in the SI.
To consider the reasons for the photoluminescence quenching, we next measured cyclic voltammograms of the catalysts and substrates to investigate the possibility of the photo-induced electron transfer (Fig. S22–S29 in the SI). The redox potentials of the catalysts and substrates used under the optimized conditions are summarized in Table 5. The reduction (Ered) and oxidation potential (Eox) of 3a were −2.84 and +0.27 V vs. Fc0/+, respectively. The excited energy (E0–0) was estimated to be 2.42 eV from Franck–Condon analysis (Fig. S18 in the SI); the reduction and oxidation potentials in the excited states (E*red and E*ox) were determined to be −0.42 V and −2.15 V, respectively. Given the oxidation potential of 1a (Eox: +0.26 V) and [2a-H]OTf (Eox: +0.24 V), the electron transfer from 1a or [2a-H]OTf to 3a* (i.e., reductive quenching of 3a* by 1a or [2a-H]OTf) should not be the main pathway for the photoluminescence quenching. On the other hand, the oxidation potential of 3a in the excited states was highly negative (E*ox: −2.15 V) enough to exergonically trigger oxidative quenching, i.e., electron transfer to [2a-H]OTf (Ered: −1.58 V) or [LutH]OTf (Ered: −1.91 V). It is noteworthy that [2a](OTf)2, which is an important intermediate of the hydrogen oxidation cycle with [2a-H]OTf, showed more positive reduction potentials (Ered: −0.66, −1.22 V), meaning that oxidative quenching by [2a](OTf)2 will also proceed efficiently. The reduction of [2a-H]OTf and [2a](OTf)2 will suppress the oxidation of dihydrogen to decrease the efficiency of the photocatalytic reactions, though the quenching efficiency with [2a-H]OTf or [2a](OTf)2 is not high due to their low concentrations. The oxidative quenching by [LutH]OTf will produce LutH˙ radical species, which potentially function as hydrogen-atom donors;45,46 therefore, this quenching process should be one of the main pathways for the hydrogenation of 1a (see below). The reduction potential of 1a (Ered: −2.60 V) is more negative than E*ox of 3a (E*ox: −2.15 V), indicating that direct single-electron transfer will not proceed efficiently; therefore, proton-coupled electron transfer from 3a* with [LutH]+ to 1a, rather than the single-electron transfer, should be a plausible pathway. Notably, as illustrated in Fig. 5, 1a exhibits strong absorption over a wide range of the visible region, and the emission spectrum of 3a and the absorption spectrum of 1a overlap significantly (Fig. S19 in the SI); therefore, excitation energy transfer from 3a* to 1a should also be considered. This excitation energy transfer lowers the concentration of 3a*, thereby reducing the efficiency of photosensitizing reactions. Consequently, the photoluminescence quenching by 1a includes both photo-induced electron transfer and/or PCET from 3a* to 1a and the excitation energy transfer, and the former process is the other main pathway for the hydrogenation of 1a (vide infra).
Table 5. Redox potentials of 1a, [2a-H]OTf, [2a](OTf)2, 3a, and [LutH]OTfa.
| Compound | E red [V] | E ox [V] | E * b [V] | E * b [V] |
|---|---|---|---|---|
| 1a | −2.60c | +0.26 | — | — |
| [2a-H]OTf | −1.58 | +0.24c | — | — |
| −2.24 | ||||
| [2a](OTf)2 | −0.66 | — | — | — |
| −1.22 | ||||
| 3a | −2.84 | +0.27 | −0.42 | −2.15 |
| [LutH]OTf | −1.91c | — | — | — |
Measured by cyclic voltammetry under a N2 atmosphere in THF containing 0.1 M of nBu4NPF6 as a supporting electrolyte. The scan rate was 100 mV s−1. Potentials are indicated vs. Fc0/+.
The redox potentials in the excited states.
Peak potentials.
The PCET with 3a* and [LutH]OTf to 1a is a thermodynamically favorable process since the effective BDFE value of the combination of 3a* and [LutH]+ (23 kcal mol−1, see Section 5.3 in the SI) is much smaller than the BDFEN–H value of the manganese–imide complex corresponding to 1a (70 kcal mol−1, Fig. 7a).50 However, it might be difficult for photo-induced PCET to occur within its short excitation lifetimes (<1.6 µs), as the three molecules need to approach simultaneously. In general, an interaction between two of them, e.g., hydrogen bond formation, can promote PCET among the three molecules. Therefore, to investigate the possibility of the photo-induced PCET, we conducted 15N NMR of THF solution containing 15N-1a (0.04 M) and [LutH]OTf (0.1 M). As illustrated in Fig. S30a in the SI, the signal in the 15N NMR spectrum was shifted and significantly broadened by adding [LutH]OTf. Some of the signals in No-Deuterium Proton (No-D) NMR spectra were also shifted (Fig. S30b in the SI). However, the UV-vis absorption spectrum of 1a was hardly changed even in the presence of [LutH]OTf (Fig. S31 in the SI). These results suggest that the interaction between the N–H moiety in [LutH]OTf and the manganese-nitride moiety in 1a is not strong, and the formation of a hydrogen bond rather than the protonation of the nitride moiety possibly takes place. We also conducted Stern–Volmer luminescence quenching studies of 3a with both 1a and [LutH]OTf (Fig. S20 and Table S12 in the SI). Although the precise estimation of rate constants for the photo-induced PCET process was quite difficult due to the very high kQ values (Table 4), the net quenching rate constants became larger than the summation of each kQ value of 1a and [LutH]OTf, indicating the possibility of the quenching process by PCET among 3a*, 1a, and [LutH]OTf (see Section 4.3 in the SI).
Fig. 7. Photocatalytic hydrogenation of manganese nitride complexes under a dihydrogen atmosphere in the presence of [2a-H]OTf, 3a, and [LutH]OTf ((a) hydrogenation steps of 1a, (b) photochemical PCET process, (c) photochemical HAT process, and (d) proposed overall reaction pathways).
Based on the abovementioned results and reported literature, the plausible reaction mechanism of the photocatalytic hydrogenation of the manganese nitride complex 1a using the diruthenium catalyst [2a-H]OTf, the photosensitizer 3a, and the proton mediator [LutH]OTf can be summarized as shown in Fig. 7d. This photocatalytic reaction is triggered by photoexcitation of 3a, though its absorption ratio is only 11% due to the inner-filter effect of 1a. After the photoexcitation, 3a* is quenched mainly by 1a (ηq = 49%) or [LutH]+ (ηq = 48%). The quenching by 1a probably includes three processes: (i) excitation energy transfer, (ii) oxidative quenching (i.e., single electron transfer from 3a* (E*ox = −2.15 V) to 1a (Ered = −2.60 V)), and (iii) photo-induced PCET among 3a*, 1a, and [LutH]+ (Fig. 7b). Although the excitation energy transfer (path (i)) is a side reaction that suppresses photocatalysis, oxidative quenching, possibly followed by subsequent protonation (path (ii)), and photo-induced PCET (path (iii)) will hydrogenate 1a to produce ammonia. These hydrogenation processes are exergonic because the BDFEN–H values of the manganese-imide, amide, and ammine complexes corresponding to 1a are reported to be 70, 89, and 79 kcal mol−1 (Fig. 7a),50 respectively, which are much higher than the effective BDFE value of the combination of 3a* and [LutH]+ (23 [kcal mol−1]). The photo-induced PCET might be promoted by the formation of a hydrogen bond between 1a and [LutH]+. On the other hand, the quenching by [LutH]+ is probably derived from the oxidative quenching (i.e., electron transfer from 3a* (E*ox = −2.15 V) to [LutH]+ (E*red = −1.91 V)). The resulting LutH˙ radical species, of which the BDFEN–H value is estimated to be 29 kcal mol−1, will donate a hydrogen atom to 1a exergonically (Fig. 7c). By repeating the photo-induced PCET from 3a* and [LutH]+ and/or HAT from LutH˙ three times, 1a will be converted in a stepwise manner to the corresponding manganese-imide, amide, and ammine complexes, resulting in the formation of NH3 and the dimanganese complex 4 (or its derivative caused by the interaction between 4 and [LutH]OTf). The one-electron-oxidized species of 3a (Eox = +0.27 V, [3a]+), which is produced by the aforementioned photo-induced PCET or oxidative quenching by [LutH]+, will oxidize [2a-H]+ (Eox = +0.24 V) to regenerate 3a and produce one-electron-oxidized species of [2a-H]+ ([2a-H]2+). [2a-H]2+ donates a proton and an electron to Lut and [3a]+, respectively, to produce the dicationic complex [2a]2+ with a vacant site. [2a]2+ reacts with a dihydrogen molecule ([2a-H2]2+), and [2a-H]+ is reproduced after the deprotonation by Lut. The high acidity of [2a-H]2+ and/or [2a-H2]2+ may have partially and directly caused the PCET process with 3a* without the proton mediator, as demonstrated in the control experiment in Table 1. The reproduction of [2a-H]+ proceeds quite efficiently, leading to the procedure of photocatalytic hydrogenation of 1a even under 0.05 atm of dihydrogen. Under the optimized conditions, photo-induced PCET to 1a with 3a* and [LutH]+ and/or HAT from LutH˙ to 1a are estimated to be the plausible rate-determining steps because of the negligible effect of the increase in the concentration of [2a]OTf on the reaction rate and the large KIE value (>3). In contrast, under the low pressure of dihydrogen (0.2 or 0.05 atm), the rate-determining step might have switched to the reaction between [2a]2+ and dihydrogen in the catalytic cycle for dihydrogen oxidation because of the significant effect of the concentration of [2a-H]OTf on the reaction rate, as shown in Table 3.
Conclusions
In conclusion, we have successfully developed a visible-light-driven hydrogenation of manganese nitride complexes to ammonia under ambient conditions by using dihydrogen as both an electron and a proton source and combining the catalytic cycle of dihydrogen oxidation with the ruthenium dinuclear complexes as catalysts and the photosensitizing cycle with the iridium complexes as photosensitizers. Our mechanistic studies proposed a plausible reaction mechanism in which photo-induced PCET and/or HAT to the manganese nitride complexes were effectively promoted by the proton mediator. Owing to the high durability of the diruthenium complex, the turnover numbers for ammonia formation based on the catalyst increased drastically (up to 150 equiv. based on the catalyst) compared with those in reported systems. Moreover, this hydrogenation proceeded photocatalytically even under 0.05 atm of dihydrogen. This feature is particularly useful to achieve the challenging target reaction, i.e., photocatalytic ammonia formation under a mixture of nitrogen and hydrogen gases, because a high partial pressure of dihydrogen often causes unexpected side reactions in the nitrogen fixation cycle; therefore, we believe that the photocatalytic hydrogenation developed in this study will function as one of the key processes for realizing catalytic ammonia production from dinitrogen and dihydrogen under ambient reaction conditions, and further studies on developing such photocatalytic system are currently underway.
Author contributions
Y.N. conceived and designed this project. Q.W. mainly conducted the experimental work. Y.T. mainly analyzed crystal structures by using X-ray diffraction measurements. Y.Y. mainly analyzed experimental data and prepared the first draft of the manuscript. All authors discussed the results and drafted the manuscript.
Conflicts of interest
There are no conflicts to declare.
Supplementary Material
Acknowledgments
We acknowledge the Grants-in-Aid for Scientific Research (No. 20H05671, 24H00049, 24H01834, 24K21778, JP23H03832, JP23H03830, and 24K08441) from JSPS and MEXT. This paper is based on results obtained from a project, JPNP 21020, commissioned by the New Energy and Industrial Technology Development Organization (NEDO).
Data availability
The data supporting this article have been included as part of the supplementary information (SI).
CCDC 2413808 and 2413811 contain the supplementary crystallographic data for this paper.65a,b
Supplementary information: the results of photocatalytic reactions, X-ray crystallographic data, photophysical and photochemical properties of photosensitizers, electrochemical properties of photosensitizers, catalysts, manganese complexes, and proton mediators, ESR spectra of complex 4, and NMR spectra. See DOI: https://doi.org/10.1039/d6sc02866g.
References
- Liu H. Chin. J. Catal. 2014;35:1619–1640. doi: 10.1016/S1872-2067(14)60118-2. [DOI] [Google Scholar]
- Wang X. Fan W. Chen J. Feng G. Zhang X. Fuel. 2022;329:125496. doi: 10.1016/j.fuel.2022.125496. [DOI] [Google Scholar]
- Guo J. Chen P. Chem. 2017;3:709–712. [Google Scholar]
- Mazloomi K. Gomes C. Renewable Sustainable Energy Rev. 2012;16:3024–3033. doi: 10.1016/j.rser.2012.02.028. [DOI] [Google Scholar]
- Smith C. Hill A. K. Torrente-Murciano L. Energy Environ. Sci. 2020;13:331–344. doi: 10.1039/C9EE02873K. [DOI] [Google Scholar]
- Sagel V. N. Rouwenhorst K. H. R. Faria J. A. Renewable Sustainable Energy Rev. 2022;161:112381. doi: 10.1016/j.rser.2022.112381. [DOI] [Google Scholar]
- Zhang S. Zhao Y. Shi R. Waterhouse G. I. N. Zhang T. EnergyChem. 2019;1:100013. doi: 10.1016/j.enchem.2019.100013. [DOI] [Google Scholar]
- Lin S. Zhang X. Chen L. Zhang Q. Ma L. Liu J. Green Chem. 2022;24:9003–9026. doi: 10.1039/D2GC03174D. [DOI] [Google Scholar]
- Shi Y. Zhao Z. Yang D. Tan J. Xin X. Liu Y. Jiang Z. Chem. Soc. Rev. 2023;52:6938–6956. doi: 10.1039/D2CS00797E. [DOI] [PubMed] [Google Scholar]
- Joseph Sekhar S. Al-Shahri A. S. A. Glivin G. Le THT. Mathimani T. Fuel. 2024;358:130307. doi: 10.1016/j.fuel.2023.130307. [DOI] [PubMed] [Google Scholar]
- Jia G. Morris R. H. Schweitzer C. T. Inorg. Chem. 1991;30:593–594. doi: 10.1021/ic00004a001. [DOI] [Google Scholar]
- Nishibayashi Y. Iwai S. Hidai M. Science. 1998;279:540–542. doi: 10.1126/science.279.5350.540. [DOI] [PubMed] [Google Scholar]
- Nishibayashi Y. Takemoto S. Iwai S. Hidai M. Inorg. Chem. 2000;39:5946–5957. doi: 10.1021/ic000799f. [DOI] [PubMed] [Google Scholar]
- Nishibayashi Y. Wakiji I. Hirata K. DuBois M. R. Hidai M. Inorg. Chem. 2001;40:578–580. doi: 10.1021/ic000716v. [DOI] [PubMed] [Google Scholar]
- Pool J. A. Lobkovsky E. Chirik P. J. Nature. 2004;427:527–530. doi: 10.1038/nature02274. [DOI] [PubMed] [Google Scholar]
- Pool J. A. Bernskoetter W. H. Chirik P. J. J. Am. Chem. Soc. 2004;126:14326–14327. doi: 10.1021/ja045566s. [DOI] [PubMed] [Google Scholar]
- Bernskoetter W. H. Lobkovsky E. Chirik P. J. J. Am. Chem. Soc. 2005;127:14051–14061. doi: 10.1021/ja0538841. [DOI] [PubMed] [Google Scholar]
- Pun D. Bradley C. A. Lobkovsky E. Keresztes I. Chirik P. J. J. Am. Chem. Soc. 2008;130:14046–14047. doi: 10.1021/ja8048332. [DOI] [PubMed] [Google Scholar]
- Reiners M. Baabe D. Münster K. Zaretzke M. K. Freytag M. Jones P. G. Coppel Y. Bontemps S. del Rosal I. Maron L. Walter M. D. Nat. Chem. 2020;12:740–746. doi: 10.1038/s41557-020-0483-7. [DOI] [PubMed] [Google Scholar]
- Xin X. Douair I. Zhao Y. Wang S. Maron L. Zhu C. Natl. Sci. Rev. 2023;10:nwac144. doi: 10.1093/nsr/nwac144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Laplaza C. E. Cummins C. C. Science. 1995;268:861–863. doi: 10.1126/science.268.5212.861. [DOI] [PubMed] [Google Scholar]
- Hebden T. J. Schrock R. R. Takase M. K. Müller P. Chem. Commun. 2012;48:1851–1853. doi: 10.1039/C2CC17634C. [DOI] [PubMed] [Google Scholar]
- Klopsch I. Finger M. Würtele C. Milde B. Werz D. B. Schneider S. J. Am. Chem. Soc. 2014;136:6881–6883. doi: 10.1021/ja502759d. [DOI] [PubMed] [Google Scholar]
- Rebreyend C. de Bruin B. Angew. Chem., Int. Ed. 2015;54:42–44. doi: 10.1002/anie.201409727. [DOI] [PubMed] [Google Scholar]
- Lindley B. M. van Alten R. S. Finger M. Schendzielorz F. Würtele C. Miller A. J. M. Siewert I. Schneider S. J. Am. Chem. Soc. 2018;140:7922–7935. doi: 10.1021/jacs.8b03755. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hasanayn F. Holland P. L. Goldman A. S. Miller A. J. M. J. Am. Chem. Soc. 2023;145:4326–4342. doi: 10.1021/jacs.2c12243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ashida Y. Arashiba K. Nakajima K. Nishibayashi Y. Nature. 2019;568:536–540. doi: 10.1038/s41586-019-1134-2. [DOI] [PubMed] [Google Scholar]
- Ashida Y. Mizushima T. Arashiba K. Egi A. Tanaka H. Yoshizawa K. Nishibayashi Y. Nat. Synth. 2023;2:635–644. doi: 10.1038/s44160-023-00292-9. [DOI] [Google Scholar]
- Ashida Y. Onozuka Y. Arashiba K. Konomi A. Tanaka H. Kuriyama S. Yamazaki Y. Yoshizawa K. Nishibayashi Y. Nat. Commun. 2022;13:7263. doi: 10.1038/s41467-022-34984-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yamazaki Y. Endo Y. Nishibayashi Y. Nat. Commun. 2025;16:4540. doi: 10.1038/s41467-025-59727-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tanabe Y. Nishibayashi Y. Chem. Soc. Rev. 2021;50:5201–5242. doi: 10.1039/D0CS01341B. [DOI] [PubMed] [Google Scholar]
- Tanabe Y. Nishibayashi Y. Coord. Chem. Rev. 2022;472:214783. doi: 10.1016/j.ccr.2022.214783. [DOI] [Google Scholar]
- Tanabe Y. Nishibayashi Y. Angew. Chem., Int. Ed. 2024;63:e202406404. doi: 10.1002/anie.202406404. [DOI] [PubMed] [Google Scholar]
- Péigne B. Aullón G. Acta Crystallogr., Sect. B:Struct. Sci., Cryst. Eng. Mater. 2015;71:369–386. doi: 10.1107/S2052520615006083. [DOI] [PubMed] [Google Scholar]
- Yandulov D. V. Schrock R. R. Science. 2003;301:76–78. doi: 10.1126/science.1085326. [DOI] [PubMed] [Google Scholar]
- Anderson J. S. Rittle J. Peters J. C. Nature. 2013;501:84–87. doi: 10.1038/nature12435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chalkley M. J. Drover M. W. Peters J. C. Chem. Rev. 2020;120:5582–5636. doi: 10.1021/acs.chemrev.9b00638. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bruch Q. J. Connor G. P. McMillion N. D. Goldman A. S. Hasanayn F. Holland P. L. Miller A. J. M. ACS Catal. 2020;10:10826–10846. doi: 10.1021/acscatal.0c02606. [DOI] [Google Scholar]
- Bruch Q. J. Malakar S. Goldman A. S. Miller A. J. M. Inorg. Chem. 2022;61:2307–2318. doi: 10.1021/acs.inorgchem.1c03698. [DOI] [PubMed] [Google Scholar]
- Merakeb L. Bennaamane S. De Freitas J. Clot E. Mézailles N. Robert M. Angew. Chem., Int. Ed. 2022;61:e202209899. doi: 10.1002/anie.202209899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jia H. P. Quadrelli E. A. Chem. Soc. Rev. 2014;43:547–564. doi: 10.1039/C3CS60206K. [DOI] [PubMed] [Google Scholar]
- Forrest S. J. K. Schluschaß B. Yuzik-Klimova E. Y. Schneider S. Chem. Rev. 2021;121:6522–6587. doi: 10.1021/acs.chemrev.0c00958. [DOI] [PubMed] [Google Scholar]
- Macleod K. C. Holland P. L. Nat. Chem. 2013;5:559–565. doi: 10.1038/nchem.1620. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kim S. Zhong H. Park Y. Loose F. Chirik P. J. J. Am. Chem. Soc. 2020;142:9518–9524. doi: 10.1021/jacs.0c03346. [DOI] [PubMed] [Google Scholar]
- Johansen C. M. Boyd E. A. Peters J. C. Sci. Adv. 2022;8:eade3510. doi: 10.1126/sciadv.ade3510. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Johansen C. M. Benazzi E. Peters J. C. Proc. Natl. Acad. Sci. U. S. A. 2025;122:e2502484122. doi: 10.1073/pnas.2502484122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Huynh M. H. V. Meyer T. J. Chem. Rev. 2007;107:5004–5064. doi: 10.1021/cr0500030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Koper M. T. M. Phys. Chem. Chem. Phys. 2013;15:1399–1407. doi: 10.1039/C2CP42369C. [DOI] [PubMed] [Google Scholar]
- Garrido-Barros P. Derosa J. Chalkley M. J. Peters J. C. Nature. 2022;609:71–76. doi: 10.1038/s41586-022-05011-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Panetti G. B. Kim J. Myong M. S. Bird M. J. Scholes G. D. Chirik P. J. J. Am. Chem. Soc. 2024;146:27610–27621. doi: 10.1021/jacs.4c08795. [DOI] [PubMed] [Google Scholar]
- Kim S. Park Y. Kim J. Pabst T. P. Chirik P. J. Nat. Synth. 2022;1:297–303. doi: 10.1038/s44160-022-00044-1. [DOI] [Google Scholar]
- Park Y. Kim S. Tian L. Zhong H. Scholes G. D. Chirik P. J. Nat. Chem. 2021;13:969–976. doi: 10.1038/s41557-021-00732-z. [DOI] [PubMed] [Google Scholar]
- Kim J. Panetti G. B. Kaul N. Kim S. Chirik P. J. J. Am. Chem. Soc. 2025;147:8215–8226. doi: 10.1021/jacs.4c14966. [DOI] [PubMed] [Google Scholar]
- Park Y. Tian L. Kim S. Pabst T. P. Kim J. Scholes G. D. Chirik P. J. JACS Au. 2022;2:407–418. doi: 10.1021/jacsau.1c00460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Yuki M. Sakata K. Hirao Y. Nonoyama N. Nakajima K. Nishibayashi Y. J. Am. Chem. Soc. 2015;137:4173–4182. doi: 10.1021/jacs.5b00584. [DOI] [PubMed] [Google Scholar]
- Yuki M. Sakata K. Nakajima K. Kikuchi S. Sekine S. Kawai H. Nishibayashi Y. Organometallics. 2017;36:4499–4506. doi: 10.1021/acs.organomet.7b00764. [DOI] [Google Scholar]
- Yuki M. Sakata K. Kikuchi S. Kawai H. Takahashi T. Ando M. Nakajima K. Nishibayashi Y. Chem.–Eur. J. 2017;23:1007–1012. doi: 10.1002/chem.201604974. [DOI] [PubMed] [Google Scholar]
- Hatchard C. G. Parker C. A. Proc. R. Soc. A. 1956;235:518–536. [Google Scholar]
- Steen J. D. Duijnstee D. R. Browne W. R. Surf. Sci. Rep. 2023;78:100596. doi: 10.1016/j.surfrep.2023.100596. [DOI] [Google Scholar]
- Wang D. Loose F. Chirik P. J. Knowles R. R. J. Am. Chem. Soc. 2019;141:4795–4799. doi: 10.1021/jacs.8b12957. [DOI] [PubMed] [Google Scholar]
- Dailey G. C. Horwitz C. P. Lisek C. A. Inorg. Chem. 2002;31:5325–5330. doi: 10.1021/ic00051a026. [DOI] [Google Scholar]
- Larsen F. K. McInnes E. J. L. El Mkami H. Overgaard J. Piligkos S. Rajaraman G. Rentschler E. Smith A. A. Smith G. M. Boote V. Jennings M. Timco G. A. Winpenny R. E. Angew. Chem., Int. Ed. 2003;42:101–105. doi: 10.1002/anie.200390034. [DOI] [PubMed] [Google Scholar]
- Kar A. K. Acharya A. Pradhan G. C. Dash A. C. J. Chem. Sci. 2014;126:547–559. doi: 10.1007/s12039-014-0605-0. [DOI] [Google Scholar]
- Allen G. H. White R. P. Rillema D. P. Meyer T. J. J. Am. Chem. Soc. 1984;106:2613–2620. doi: 10.1021/ja00321a020. [DOI] [Google Scholar]
- (a) CCDC 2413808: Experimental Crystal Structure Determination, 2026, 10.5517/ccdc.csd.cc2m0rs6 [DOI]; (b) CCDC 2413811: Experimental Crystal Structure Determination, 2026, 10.5517/ccdc.csd.cc2m0rw9 [DOI]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data supporting this article have been included as part of the supplementary information (SI).
CCDC 2413808 and 2413811 contain the supplementary crystallographic data for this paper.65a,b
Supplementary information: the results of photocatalytic reactions, X-ray crystallographic data, photophysical and photochemical properties of photosensitizers, electrochemical properties of photosensitizers, catalysts, manganese complexes, and proton mediators, ESR spectra of complex 4, and NMR spectra. See DOI: https://doi.org/10.1039/d6sc02866g.







