(Electro)chemical N2 Splitting by a Molybdenum Complex with an Anionic PNP Pincer-Type Ligand

Molybdenum(III) complexes bearing pincer-type ligands are well-known catalysts for N2-to-NH3 reduction. We investigated herein the impact of an anionic PNP pincer-type ligand in a Mo(III) complex on the (electro)chemical N2 splitting ([LMoCl3]−, 1–, LH = 2,6-bis((di-tert-butylphosphaneyl)methyl)-pyridin-4-one). The increased electron-donating properties of the anionic ligand should lead to a stronger degree of N2 activation. The catalyst is indeed active in N2-to-NH3 conversion utilizing the proton-coupled electron transfer reagent SmI2/ethylene glycol. The corresponding Mo(V) nitrido complex 2H exhibits similar catalytic activity as 1H and thus could represent a viable intermediate. The Mo(IV) nitrido complex 3– is also accessible by electrochemical reduction of 1– under a N2 atmosphere. IR- and UV/vis-SEC measurements suggest that N2 splitting occurs via formation of an “overreduced” but more stable [(L(N2)2Mo0)2μ-N2]2– dimer. In line with this, the yield in the nitrido complex increases with lower applied potentials.


■ INTRODUCTION
The reduction of dinitrogen to ammonia via the Haber-Bosch process is one of the most important processes of the last century, ensuring food production and the synthesis of nitrogen-containing organic compounds. 1However, the conversion of N 2 with the chemical reductant H 2 consumes about 1−2% of the global energy supply. 2Moreover, the process produces ca.2% of annual CO 2 emissions, which mainly originate from the formation and purification of H 2 from steam reforming.Thus, more sustainable approaches for N 2 fixation are highly desirable. 3nspired by nature's FeMo cofactor in the nitrogenase enzyme, the first molecular system catalyzing the N 2 -to-NH 3 conversion was based on a well-defined molybdenum trisamidoamine complex. 4Profound mechanistic analysis and isolation of several intermediates led to the assumption that N 2 -to-NH 3 formation appears by consecutive protonation and reduction steps of metal-bound N 2 . 5Eight years later, the group of Nishibayashi reported on a molybdenum complex with a PNP pincer-type ligand, [L PNP MoCl 3 ], I Cl , L PNP = 2,6bis( t Bu 2 PCH 2 )-C 5 H 3 N) (Scheme 1), as a precursor for the corresponding Mo(0) dinitrogen complex [(L PNP (N 2 ) 2 Mo) 2 μ-N 2 ], that catalyzes the N 2 -to-NH 3 reduction using cobaltocene as reductant and lutidinium as a proton source. 6Substitution in para position of the pyridine ligand showed that the catalytic activity is enhanced when electron-donating groups are introduced, and the authors hypothesized that this is due to a stronger degree of N 2 activation. 7The TON and TOF could be dramatically increased by using the proton-coupled electron transfer (PCET) reagents SmI 2 with alcohols or water and I Cl . 8ince the corresponding Mo(IV) nitrido complex [L PNP Mo-(N)I] showed a similar catalytic performance than I I , the authors proposed later that the reaction with I I proceeds via N 2 splitting and subsequent reduction and protonation of the Mo nitrido complex. 9,10In contrast to the earlier study, complexes with electron-withdrawing substituents in para position of the pyridine unit showed enhanced activity when using the Smbased PCET reagents. 11he first molecular metal catalyst beyond molybdenum was reported by the group of Peters in 2013, who introduced an iron tris(phenylphosphino)borane complex, which is active in the nitrogen reduction reaction (N 2 RR). 12 In the following years, several more molecular N 2 -to-NH 3 catalysts were reported based on Co, Os, Ru, Ti, V, Cr, Mo, and Re ions. 13hese examples relied on chemical reductants, whereas electrochemical approaches for the N 2 RR remain scarce. 14In 1985, Pickett and Talarmin reported the first example for the electrochemical N 2 activation forming substoichiometric amounts of ammonia using a bis(diphenylphosphinoethane) W(0) complex. 15In 2018, the first example of electrocatalytic N 2 RR was reported using a tris(phenylphosphino)borane Fe(I) catalyst and Cp* 2 Co + as a redox mediator. 16The concept of combining a PCET redox mediator 17 and a transition metal catalyst capable of accomplishing N 2 -to-NH 3 conversion was further expanded to W, Os, Fe, and Mo complexes. 18Recently, the electrocatalytic N 2 RR was also accomplished with I Br as a sole electrocatalyst. 19n Schrock's molybdenum trisamidoamine system, catalysis appears by consecutive protonation and reduction steps of metal-bound N 2 , which contrasts with the Haber-Bosch process, where catalysis is limited by N 2 chemisorption and splitting into surface-bound metal nitrido species. 5,20The first molecular model for such a metal-mediated N 2 splitting reaction into a terminal metal nitrido complex was reported by Laplaza and Cummins using a molybdenum(III) tris-amide complex. 21This was followed by several more examples; 9,22 however, examples for the electrochemically induced N 2 splitting into terminal metal nitrido complexes remain scarce (Scheme 1).
The first example was reported by the groups of Miller, Siewert, and Schneider in 2018 using a Re(III) PNP pincertype complex II Cl . 23[L PNP2 ReCl 2 ] (L PNP2 = bis(2-(di-tertbutylphosphaneyl)ethyl)-amide) splits dinitrogen yielding the corresponding Re(V) nitrido complex in 58% spectroscopic yield (Scheme 1).Detailed mechanistic analysis by CV (cyclic voltammetry) leads to a comprehensive, minimum model for N 2 splitting: initial reduction of [L PNP2 ReCl 2 ] is followed by very fast N 2 binding and slower chloride loss.The resulting Re(II) species is further reduced in the vicinity of the electrode due to potential inversion upon binding N 2 .Overreduced Re(I) and Re(III) form the key Re(II)−N 2 −Re(II) intermediate, which splits N 2 to form the Re(V) nitrido complex.Changing from chlorido to bromido and iodido ligands leads to less negative potentials for N 2 -splitting along the halide series. 24In contrast to the chlorido system, the iodido system II I reacts via a Re II /Re II -dimerization mechanism, due to the absence of the potential inversion after reduction and N 2 binding.Utilizing the analogous unsaturated PNP ligand (III) increases the potential for N 2 -splitting compared to II Cl with the saturated PNP ligand, but the nitrido complex was formed in a considerable lower yield (Scheme 1). 25 To date, the highest yield of 94% could be obtained using the Re platform IV with the more robust diphenylamido PNP pincer-type ligand (Scheme 1). 26ecently, electrochemical N 2 splitting by Mo(III) complexes supported by PPP and PNP pincer-type ligands have been reported, which form the corresponding Mo(IV) nitrido complexes in yields of 30% (V) and 41% (I Br ), respectively (Scheme 1). 27,28Mechanistic analysis of the N 2 -splitting pathway with I Br revealed that the initial reduction is followed by bromide loss and N 2 binding.In the vicinity of the electrode, Mo(II) is further reduced forming [(L PNP BrMo I ) 2 μ-N 2 ], which splits N 2 , and overreduced [(L PNP (N 2 ) 2 Mo 0 ) 2 μ-N 2 ], which upon a comproportionation reaction in bulk solution yields the nitrido complex. 28nspired by these studies, we investigated the impact of an anionic PNP-pyridone pincer-type ligand L − on the chemical N 2 -to-NH 3 conversion and electrochemical N 2 splitting (Scheme 2).We expected that the increased electron-donating properties of the ligand would lead to stronger N 2 activation in the reduced complex but also to a lower reduction potential for its formation.

ACS Organic & Inorganic Au
Electrochemical measurements (CV, DPV) were recorded with a Gamry Instruments Reference 600+ or Reference 600 in dry THF under a N 2 or Ar atmosphere.A common three-electrode setup was used with a glassy carbon working electrode (GC: CH Instruments, ALS Japan; A = 7.1 mm 2 ), a platinum wire as a counter electrode, and a silver wire in electrolyte solution in a fritted sample holder as a pseudo reference electrode.The GC electrode was polished on an alumina polishing pad with 0.05 μm polishing alumina (both ALS Japan) and Milli-Q water in figure-eight motion.The platinum and silver wires were washed with Milli-Q water.n Bu 4 NPF 6 was dried at 100 °C in vacuo and used as the conducting salt, I = 0.2 M. All data were collected at RT and referenced internally versus the Fc +|0 or Fc* +|0 redox couple (conversion of Fc* +|0 to Fc +|0 −0.43 V). iR compensation was performed by the positive feedback method, which is implemented in the PHE200 software of Gamry.All CV data are plotted in IUPAC convention.

General Procedure for Catalytic N 2 Reduction
Dry THF (6 mL) was added to a Schlenk tube filled with the catalyst (2.00 μmol, 1.00 equiv) and SmI 2 (thf) 2 (197 mg, 360 μmol, 180 equiv) in a N 2 glovebox.Ethylene glycol (22.4 mg, 360 μmol, 180 equiv) was added in one portion, and the mixture was further stirred for 18 h at RT resulting in a color change from deep blue to yellow.The mixture was frozen to −196 °C, and an excess of KO t Bu (∼100 mg) in MeOH (5 mL) was added.The mixture was allowed to warm to room temperature and stirred for 15 min.The volatiles were vacuum transferred into a liquid N 2 -cooled Schlenk flask containing HCl (2 M in Et 2 O, 3 mL).The solution was stirred for a further 15 min at RT, the solvents were removed in vacuo, and the residue was dissolved in DMSO-d 6 with trimethoxybenzene as an internal standard.The amount of ammonium was determined by integration versus the internal standard.

Electrolysis Experiments
Controlled potential electrolysis (CPE) experiments were performed in a custom-made H-type cell with a P3 glass frit to separate working and counter chambers under a N 2 atmosphere.The working chamber was equipped with a 7 mm-diameter glassy carbon rod (ALS Japan) and a silver wire in electrolyte solution in a fritted sample holder as a pseudo reference electrode.The counter chamber was equipped with a Zn rod as a sacrificial electrode.Both chambers were filled with 2.5 mL of electrolyte solution (0.2 M n Bu 4 NPF 6 in THF), and a stirring bar was added.1H (12.3 mg, 20 μmol) and DBU (3.00 μL, 3.04 mg, 20 μmol) were added to the working chamber.A fixed potential was applied, and the solution was electrolyzed until the desired amount of charge was passed.An aliquot of the solution in the working chamber was taken, and a drop of THF-d 8 was added.The mixture was analyzed by 31 P NMR spectroscopy, and 3 − was quantified by integration vs the PF 6 − anion.

Synthesis and Characterization of the Complexes
The synthesis of the PNP pincer-type ligand LH was previously reported. 33The reaction of [MoCl 3 (thf) 3 ] and LH at 45 °C in THF yielded paramagnetic molybdenum complex 1H in 84% yield after purification (Scheme 2).The complex was isolated as an orange-brown solid.IR spectroscopy showed a band at a frequency of 3250 cm −1 indicative of an O−H stretching vibration, and positive ESI-MS showed the presence of the [M−Cl] + and [M+Na] + ions.
The magnetic moment μ eff of 1H derived by the Evans method is 3.8 μ B , which is characteristic for an S = 3/2 ground state (S = 3/2: μ SO = 3.87 μ B ).The EPR spectrum of a THF solution of 1H recorded at 293 K shows hyperfine coupling to two phosphorus and one molybdenum atom (g iso = 2.0093, A(2 × 31 P) = 67 MHz, A(1 × 95/97 Mo) = 76 MHz; Figure S16).Single X-ray crystals of the compound were obtained by vapor diffusion of pentane into a solution of 1H in THF.The results of the refinement can be found in Figure 1, selected bond lengths are listed in Table 1, and further details are given in the SI.The molybdenum ion exhibits a distorted octahedral coordination sphere.6) Å, respectively, support protonation of the ligand. 33,34This is in line with elemental analyses, which confirm the purity of 1H.
Since the spectroscopic features of 1H pointed to the formation of the neutral Mo(III) complex, we investigated the deprotonation of the ligand via UV/vis spectroscopy.Upon gradual addition of DBU, the π−π* band at 349 nm increases and the shoulder at 305 nm gets more prominent (DBU = 1,8diazabicyclo[5.4.0]undec-7-ene, pK a = 16.9, 35 Figure 2).The d−d band at 430 nm decreases very little.Fitting of the data at five representative wavelengths led to a pK a of 15.9 ± 0.2 for 1H in THF (Figure S1).Back-titration with lutidinium reproduces the initial spectrum of 1H with a spectroscopic yield of >90%, revealing reversible protonation and deprotonation (Figure S1).
The CV and DPV (differential pulse voltammetry) data of 1H show a reversible Mo III /Mo IV oxidation at E 1/2 = 0.13 V (Figure 2 and Figure S4; if not otherwise noted, a glassy carbon (GC) electrode was used and the data are given vs the Fc +|0 reduction potential).The Mo III /Mo IV oxidation appears only at a slightly lower potential than in parent complex I Cl (cf.S3).Both reduction events remain irreversible upon increasing the scan rate from 0.05 to 1 V s −1 (Figure S2).The irreversibility of all reduction events irrespective of the scan rate is in sharp contrast to I Cl and I Br , which showed an increased reversibility of the first reduction upon increasing the scan rate (Figure S7 and ref 28).Initial reduction of I Br was proposed to be followed by rather slow bromide loss. 28This points to a different reactivity after the reduction of 1H.
Based on previous reports with such hydroxypyridine ligands, we envisioned reductive O−H bond cleavage as a follow-up reaction as observed, e.g., by Fujita and co-workers in Re and Ru complexes bearing the [2,2′-bipyridine]-4,4′-diol ligand. 36Upon reduction, homolytic O−H bond cleavages occur, forming the dianionic bipyridone ligand and H 2 .A similar scenario seemed plausible also for 1H, and indeed, a UV/vis-spectroelectrochemical (SEC) experiment under reductive conditions revealed similar changes in the UV/vis spectra as observed during deprotonation of 1H (Figure 2 and Figure S12).Analysis of the gas phase by gas chromatography after a CPE experiment at an applied potential of −2.1 V passing 1 charge equivalent revealed the formation of H 2 in yields of 55% (Figure S38).CV studies of 1H with increasing amounts of DBU showed the disappearance of the first reduction wave, which was therefore assigned to reductive O− H bond cleavage forming H 2 and 1 − (Figure 2 and Figure S6).Further reduction of deprotonated complex 1 − overlaps with the initial reduction of 1H and could explain the low yield in H 2 .
The CV and DPV data of in situ formed 1 − show a Mo III / Mo IV oxidation at E 1/2 = 0.07 V (Figure 2 and Figures S4 and  S6).The potential shift compared to the oxidation of 1H is small indicating a weak thermodynamic coupling between the proton and the metal site, similar to that observed for the respective Ni complex. 33The CV data of 1 − show a second irreversible oxidation at a peak potential of 0.36 V, which likely belongs to the oxidation of the deprotonated ligand, because it becomes more prominent in 1H with increasing amounts of DBU and it is also present in the CV data of 1H, when the CV is scanned reductively at first, but not present, when scanned anodically at first (Figure 2 and Figures S5 and S6).
Having established the pK a of 1H and the oxidation potential of 1 − , we estimated the BDFE of the O−H bond in 1H from the Bordwell equation using the recently reported value of c G in THF with 59.9 kcal mol −1 . 37

FE
RT K c BDFE 2.303 p It sums up to 83 kcal mol −1 , which is similar to the BDFE of the O−H bond in phenol. 37reatment of 1H with a slight excess of NaN 3 leads to the azido complex, which upon irradiation with a blue LED at a wavelength of 456 nm in THF provides the Mo(V) nitrido complex 2H in 59% yield after recrystallization from THF/ hexane (Scheme 2).High-resolution ESI mass spectrometry showed the formation of [(LH)Mo(N)Cl] + .The EPR spectrum of a THF solution of 2H at 279 K showed one signal at g iso = 1.9849, and simulation as an S = 1/2 system revealed hyperfine coupling to two different N atoms and the phosphorus atoms substantiating formation of the desired Mo(V) nitrido complex 2H (A(2 × 31 P) = 40.6MHz; A(2 × 1 H) = 4.2 MHz; A(1 × 14 N) = 9.8 MHz; A(1 × 14 N) = 4.5 MHz, Figure S17).The purity of 2H was confirmed by an elemental analysis.2H exhibits two irreversible reductions at peak potentials of −2.38 and −2.59 V and four oxidation processes at peak potentials of −0.91, −0.28, 0.15, and 0.39 V (ν = 0.1 V s −1 , Figure S8).The oxidation at −0.91 V was assigned to the Mo(IV/V) redox couple in analogy to related Mo nitrido complexes. 9,27,38Reduction of 2H with an exc. of KC 8 or deprotonation of 2H with 1 equiv of DBU and reduction with an exc. of KC 8 led to the corresponding Mo(IV) nitrido complex 3K, which shows a characteristic signal in the 31 P NMR spectrum at 85.2 ppm (Scheme 2 and Figure S19). 39The CV data of 3 − exhibit no characteristic reduction feature up to −2.5 V and oxidation processes at peak potentials of −1.13, −0.36, and 0.65

Reactivity towards N 2 Using Chemical Reductants
Next, we focused on the catalytic reactivity of 1H in the dinitrogen reduction reaction forming ammonia.Therefore, we adapted the previously reported protocol for the N 2 -to-NH 3 formation catalyzed by I I . 81H was treated with 180 equiv SmI 2 (thf) 2 and 180 equiv ethylene glycol in THF under a 1 atm nitrogen atmosphere for 18 h (Table 2).This led to formation of 9.4 ± 0.6 equiv of ammonia per Mo, confirming that 1H is indeed an active catalyst in the N 2 -to-NH 3 formation, albeit with lower TON than I I .Ammonia was quantified as the ammonium salt after vacuum transfer and acidic workup via NMR spectroscopy with the internal standard 1,3,5-trimethoxybenzene.Repeating the experiment under a 15 N 2 atmosphere led to the expected 1 H signal of 15 NH 4 + , confirming that N 2 is the N source in ammonium (Table 2 and Figures S20 and S21).A blank experiment in the absence of the catalyst did not lead to detectable amounts of NH 3 .
To get further insights into the N 2 reduction pathway, the catalytic activity of the corresponding molybdenum(V) nitrido complex 2H was investigated next.Reaction of 2H under the same conditions showed catalytic conversion of dinitrogen into ammonia with a TON of 7.3 ± 0.7, indicating that the nitrido complex could be a plausible intermediate during catalysis with 1H, although the yield was slightly lower (Table 2).Control experiments under an argon atmosphere with 2H showed the formation of 0.9 equiv of ammonia per catalyst, which originates from the nitrido nitrogen atom.Catalysis under isotopically labeled 15 N 2 with 2H revealed catalytic conversion into 15 NH 3 and 1 equiv of 14 NH 3 (Figure S23).

Reactivity towards N 2 under Electrochemical Conditions
To probe for electrochemical N 2 splitting, CV data under a N 2 atmosphere were recorded.Under a N 2 atmosphere, 1H shows four irreversible reductions at E p,c,1 = −2.15V, a shoulder at E p,c,2 = −2.32V, and a third and fourth event at E p,c,3 = −2.46V and E p,c,4 = −2.57V, respectively (ν = 0.1 V s −1 , Figure 2).The second reduction process shifts largely with increasing scan rates and merges with the third one at 1 V s −1 (Figure S5).When the potential is switched beyond the third reduction process, two new oxidation processes appear at E p,a,1 = −0.93V and at E p,a,2 = 0.35 V, as well as the Mo(III/IV) oxidation at E 1/2 = 0.13 V.The oxidation process at −0.93 V is not present when the CV data of 1H are collected under Ar, and the CV data of 2H exhibit an oxidation process at a similar potential (Figures S8 and S9), which could indicate in situ formation of the nitrido complex.The latter new oxidation process E p,a,2 belongs to the oxidation of the deprotonated ligand (vide supra).As under Ar, initial reduction is associated with the cleavage of the O−H bond and the formation of 1 − as confirmed by titration experiments with DBU.The reduction vanishes with increasing amounts of DBU (Figure S6).The first reduction of 1H under N 2 appears at a potential −0.2 V lower than the potential of the first reduction of I Cl , whereas further reductions of 1H are easier to accomplish than in I Cl (cf.Since the CV data under a N 2 atmosphere revealed distinct differences compared to those under Ar, electrochemical dinitrogen splitting was investigated.As initial reduction is associated with H 2 formation, we utilized 1 − in the electrochemical N 2 splitting experiments, which was prepared in situ by adding one equivalent of DBU. 1 − was electrolyzed at −2.69 V vs Fc +|0 until four electrons were injected.Notably, [DBUH] + is reduced at the applied potential of the electrolysis; thus, only three electrons are consumed by 1 − (Figure S10).Independent CPE experiments of [DBUH]Cl at an applied potential of −2.6 V confirmed its one-electron reduction and H 2 evolution in yields of 75% (Figures S10 and  S11). 31P NMR spectroscopy revealed the formation of a singlet at a chemical shift of 84.4 ppm corresponding to the desired Mo(IV) nitrido complex 3 − in only 7% yield.The yield has been quantified via integration vs the PF 6 − signal (Table 3, entry 1, and Figure S25).Since reductive N 2 splitting by 1 − , leading to the Mo(IV) nitrido complex, requires only 2 e − , we investigated the yield in dependence of the injected charge equivalents per [Mo], i.e., 2, 3, and 4 e − .After the injection of two electrons, the formation of the Mo(IV) nitrido complex was not observed.This is in line with the electronic structure requirements for N 2 splitting in a pseudo D 4h symmetric complex (vide infra).After injection of three electrons, it was formed in higher yields of 12%.Injection of 4 e − leads to lower yields, indicating that overreduction leads to decomposition or formation of paramagnetic species (Table 3, entries 2, 3, and 5, and Figure S29).CV analysis of 2H showed that the Mo nitrido complex is reduced at the applied potential of the electrolysis, and thus, it could be further reduced in situ forming paramagnetic species (Figure S8).To probe if 31 P NMR spectroscopy is a suitable method to quantify nitrido complex formation, it was also quantified via release as ammonium and subsequent 1 H NMR spectroscopy (Figure S30).Ammonium formation was observed after workup with a spectroscopic yield of 14%, confirming the reliability of quantification by 31 P NMR spectroscopy (Table 3, entry 4).Electrochemical reduction of 1H under a 15 N 2 atmosphere and subsequent acidic workup confirmed that N 2 is the N source in ammonium (Figure S31).
To obtain further insights into the mechanism, IRspectroelectrochemical investigation of 1H under reductive conditions under a N 2 atmosphere in an OTTLE cell using a gold grid electrode were conducted. 40These bands disappear during a prolonged electrolysis time (Figure S14).The N 2 stretching vibrations of the intermediates during N 2 splitting with 1 − are shifted to lower energy compared to those utilizing I Br , supporting the assumption that the anionic ligand led to a higher N 2 activation degree.The vibrations were unambiguously assigned to N 2 vibrations by using isotope-labeled 15 N 2 .The stretching vibrations are red-shifted and appear at 1967, 1875, and 1853 cm −1 (Figure S15).Assuming a similar force constant, the vibrations are expected at 1968, 1881, and 1859 cm −1 as calculated from the change of the reduced mass in 14/15 N 2 , which fits well with the experimental data.
The formation of a dinitrogen bridged Mo(0) dimer as an intermediate during nitrido formation was further substantiated by UV/vis-SEC measurements.During electrolysis, a new band at 723 nm appeared, which is characteristic for the dimer [(L(N 2 ) 2 Mo) 2 μ-N 2 ] 2− comparing it with the parent [(L PNP (N 2 ) 2 Mo) 2 μ-N 2 ] complex, which has a characteristic band at 688 nm (Figure 3 and Figure S13). 28onsidering the electronic structure requirements for N 2 splitting by dimeric [(M) 2 μ-N 2 ] species in a pseudo D 4h symmetry, Mo(I) should be the active species as it exhibits  the π 10 configuration, which leads to the maximum weakening of the N 2 bond and the electronic requirements for forming the metal nitrido species. 41In line with these considerations, electrochemical oxidation of trans-(depe) 2 Mo 0 (N 2 ) 2 induces N 2 splitting forming the nitrido complex. 42However, IR-SEC experiments of 1H under N 2 point to the in situ formation of Mo(0) species, indicating an overreduction of the complex.Such electrochemical overreduction in the vicinity of the electrode surface has been proposed previously in related Mo and Re complexes with PNP pincer-type ligands on the path to nitrido complex formation. 23,28This was rationalized by the potential inversion of the reduction potentials after halide loss and N 2 binding.Overreduced Mo(0) or Re(I) reacts with the starting material forming the reactive intermediate, viz., the Re(II)-or Mo(I)-bridged dimer, respectively.A similar scenario seems also plausible here, as 1 − shows very similar potentials for successive reduction steps under a N 2 atmosphere. 43Oxidation of the Mo(0) dimer by homogeneous electron transfer from, e.g., 1 − induces N 2 splitting forming the Mo(IV) nitrido complex (Scheme 3).Such overreduction forming a more stable Mo(0) intermediate as a viable route was further examined by potential dependent formation of the nitrido complex.At an applied potential of −2.69 V, molybdenum nitrido formation was observed in 7% yield (Table 3, entry 1).With lower applied potentials, the yield increases to 18% at −3.08 V and 31% at −3.15 V (Table 3, entries 6 and 7), which is in line with the assumption that the overreduction is crucial for N 2 splitting formation.

■ CONCLUSIONS
The molybdenum(III) complex 1 − with an anionic PNP pyridone pincer-type ligand has been investigated in N 2 -to-NH 3 conversion and electrochemical N 2 splitting.The complex catalyzes the ammonia formation with SmI 2 /ethylene glycol as a PCET reagent with a TON of 9.4.The reaction could proceed via the Mo nitrido complex, from which ammonia is released by successive protonation and reduction steps, as we observe very similar TON in the experiments using 1H and the Mo(V) nitrido complex 2H.Electrochemical reduction of 1 − in the presence of N 2 leads to the formation of the Mo(IV) nitrido complex 3 − .IR-and UV/vis-SEC experiments indicate that the reaction proceeds via an overreduced Mo(0) complex.
In line with this, the yield in 3 − increases with a lower applied potential.The anionic nature of the ligand shifts the potential for the initial reduction of 1 − to more negative potential compared to I Cl ; however, the degree of N 2 activation seems to be slightly higher than in the parent complex due to the more electron-donating properties of the ligand as deduced from the N�N stretching vibrations in the molybdenum−dinitrogen complexes.

Data Availability Statement
The data underlying this study are available in the published article and its Supporting Information.

* sı Supporting Information
The Supporting Information is available free of charge at https://pubs.

Figure 1 .
Figure 1.Molecular structure of 1H; the hydrogen atom was placed in a calculated position.Most hydrogen atoms and all solvent molecules were omitted for clarity.Thermal ellipsoids were set at the 50% probability level.

E
1/2 = 0.20 V, Figure S7) indicating that switching from hydroxypyridine to pyridine has very little effect on the electronic structure at the Mo ion.Scanning reductively, 1H shows an irreversible reduction at E p,c,2 = −2.41V with a shoulder at E p,c,1 = −2.19V and three associated irreversible reoxidations at E p,a,1 = −0.85V, E p,a,2 = −0.31V, and E p,a,4 = 0.35 V under Ar (ν = 0.1 V s −1 , Figure

Table 2 .
Catalytic N 2 -to-NH 3 Conversion a a Inset: 1 H NMR spectra in DMSO-d 6 of the ammonium cation generated via N 2 reduction under 14 N 2 (bottom) and 15 N 2 (top) catalyzed by 1H.
and I Cl6 were prepared according to literature procedures.

Table 1 .
Selected Bond Lengths (Å) for 1H, with Estimated Standard Deviations in Parentheses

Table 3 .
Yields of the Mo(IV) Nitrido Complex after Electrochemical Reduction of 1H with 1 Equiv DBU in THF under N 2 a Boron-doped diamond electrode instead of GC.