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

Abstract

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.¹ However, the conversion of N₂ with the chemical reductant H₂ consumes about 1–2% of the global energy supply.² Moreover, the process produces ca. 2% of annual CO₂ emissions, which mainly originate from the formation and purification of H₂ from steam reforming. Thus, more sustainable approaches for N₂ fixation are highly desirable.³

Inspired by nature’s FeMo cofactor in the nitrogenase enzyme, the first molecular system catalyzing the N₂-to-NH₃ conversion was based on a well-defined molybdenum trisamidoamine complex.⁴ Profound mechanistic analysis and isolation of several intermediates led to the assumption that N₂-to-NH₃ formation appears by consecutive protonation and reduction steps of metal-bound N₂.⁵ Eight years later, the group of Nishibayashi reported on a molybdenum complex with a PNP pincer-type ligand, [L^PNPMoCl₃], I^Cl, L^PNP = 2,6-bis(^tBu₂PCH₂)-C₅H₃N) (Scheme 1), as a precursor for the corresponding Mo(0) dinitrogen complex [(L^PNP(N₂)₂Mo)₂μ-N₂], that catalyzes the N₂-to-NH₃ reduction using cobaltocene as reductant and lutidinium as a proton source.⁶ Substitution 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 N₂ activation.⁷ The TON and TOF could be dramatically increased by using the proton-coupled electron transfer (PCET) reagents SmI₂ with alcohols or water and I^Cl.⁸ Since the corresponding Mo(IV) nitrido complex [L^PNPMo(N)I] showed a similar catalytic performance than I^I, the authors proposed later that the reaction with I^I proceeds via N₂ splitting and subsequent reduction and protonation of the Mo nitrido complex.^9,10 In contrast to the earlier study, complexes with electron-withdrawing substituents in para position of the pyridine unit showed enhanced activity when using the Sm-based PCET reagents.¹¹

Scheme 1. Precedents for Electrochemical N₂ Splitting into the Corresponding Terminal Transition Metal Nitrido Complexes; Applied Potentials during Electrolysis E^appl Given vs Fc^+|0, if Not Otherwise Noted; Yields of the Nitrido Complexes Given in Parentheses^,

1,2-Difluorobenzene instead of THF.

Potential given vs SCE, 0.1 M NBu₄BArF₂₄ in THF.

The 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 (N₂RR).¹² In the following years, several more molecular N₂-to-NH₃ catalysts were reported based on Co, Os, Ru, Ti, V, Cr, Mo, and Re ions.¹³ These examples relied on chemical reductants, whereas electrochemical approaches for the N₂RR remain scarce.¹⁴ In 1985, Pickett and Talarmin reported the first example for the electrochemical N₂ activation forming substoichiometric amounts of ammonia using a bis(diphenylphosphinoethane) W(0) complex.¹⁵ In 2018, the first example of electrocatalytic N₂RR was reported using a tris(phenylphosphino)borane Fe(I) catalyst and Cp*₂Co⁺ as a redox mediator.¹⁶ The concept of combining a PCET redox mediator¹⁷ and a transition metal catalyst capable of accomplishing N₂-to-NH₃ conversion was further expanded to W, Os, Fe, and Mo complexes.¹⁸ Recently, the electrocatalytic N₂RR was also accomplished with I^Br as a sole electrocatalyst.¹⁹

In Schrock’s molybdenum trisamidoamine system, catalysis appears by consecutive protonation and reduction steps of metal-bound N₂, which contrasts with the Haber-Bosch process, where catalysis is limited by N₂ chemisorption and splitting into surface-bound metal nitrido species.^5,20 The first molecular model for such a metal-mediated N₂ splitting reaction into a terminal metal nitrido complex was reported by Laplaza and Cummins using a molybdenum(III) tris-amide complex.²¹ This was followed by several more examples;^9,22 however, examples for the electrochemically induced N₂ 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 pincer-type complex II^Cl.²³ [L^PNP2ReCl₂] (L^PNP2 = bis(2-(di-tert-butylphosphaneyl)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 N₂ splitting: initial reduction of [L^PNP2ReCl₂] is followed by very fast N₂ 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 N₂. Overreduced Re(I) and Re(III) form the key Re(II)–N₂–Re(II) intermediate, which splits N₂ to form the Re(V) nitrido complex. Changing from chlorido to bromido and iodido ligands leads to less negative potentials for N₂-splitting along the halide series.²⁴ In 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 N₂ binding. Utilizing the analogous unsaturated PNP ligand (III) increases the potential for N₂-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).²⁶

Recently, electrochemical N₂ 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).^27,28 Mechanistic analysis of the N₂-splitting pathway with I^Br revealed that the initial reduction is followed by bromide loss and N₂ binding. In the vicinity of the electrode, Mo(II) is further reduced forming [(L^PNPBrMo^I)₂μ-N₂], which splits N₂, and overreduced [(L^PNP(N₂)₂Mo⁰)₂μ-N₂], which upon a comproportionation reaction in bulk solution yields the nitrido complex.²⁸

Inspired by these studies, we investigated the impact of an anionic PNP-pyridone pincer-type ligand L^– on the chemical N₂-to-NH₃ conversion and electrochemical N₂ splitting (Scheme 2). We expected that the increased electron-donating properties of the ligand would lead to stronger N₂ activation in the reduced complex but also to a lower reduction potential for its formation.

Scheme 2. Synthesis of 1H, 2H, and 3^K; All Reactions in THF.

Experimental Section

If not otherwise stated, manipulations of air-sensitive reagents were carried out by means of common Schlenk-type techniques using a dry Ar or N₂ atmosphere or performed in a N₂- or Ar-filled MBraun glovebox. Solvents were dried using an MBraun Solvent Purification System and stored over 3 Å molecular sieves. ¹⁵N₂, ⁿBuLi, HCl (2 M/Et₂O), KO^tBu, NaH, DBU, ^tBu₂PCl, NaN₃ (all Sigma-Aldrich), and chloromethyl ethyl ether (ABCR) were used as purchased. TMEDA was dried over sodium and distilled. [MoCl₃(thf)₃],²⁹ SmI₂(thf)₂,³⁰ LutHOTf,³¹L^PNP,³²LH,³³ and I^Cl(6) were prepared according to literature procedures.

Electrochemical measurements (CV, DPV) were recorded with a Gamry Instruments Reference 600+ or Reference 600 in dry THF under a N₂ or Ar atmosphere. A common three-electrode setup was used with a glassy carbon working electrode (GC: CH Instruments, ALS Japan; A = 7.1 mm²), 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. ⁿBu₄NPF₆ 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.

[(LH)MoCl₃] (1H)

A suspension of [MoCl₃(thf)₃] (112 mg, 0.27 mmol, 1.00 equiv) and LH (116 mg, 0.28 mmol, 1.05 equiv) was heated in THF (10 mL) at 45 °C for 18 h under an Ar atmosphere. The solution was filtered, and the solvent was removed in vacuo. Recrystallization from CH₂Cl₂/hexane afforded 1H (139 mg, 0.23 mmol, 84%) as an orange/brown powder. Crystals suitable for X-ray diffraction were obtained by vapor diffusion of pentane into a solution of 1H in THF at −30 °C. HR-MS (ESI): [C₂₃H₄₃Cl₂MoNOP₂]⁺ ([M–Cl]⁺): calcd. m/z = 579.1239, found m/z = 579.1231. [C₂₃H₄₃Cl₃MoNNaOP₂]⁺ ([M+Na]⁺): calcd. m/z = 637.0819, found m/z = 637.0813. μ_eff = 3.8 μ_B. IR (KBr): ν̃ (cm^–1) = 3250 (br), 2949 (m), 2900 (m), 2870 (m), 1614 (s), 1593 (m), 1475 (s), 1394 (w), 1369 (m), 1342 (m), 1261 (w), 1219 (w), 1178 (m), 1146 (w), 1105 (w), 1024 (s), 974 (w), 937 (w), 868 (m), 812 (m). El. anal. for C₂₃H₄₃Cl₃MoNOP₂: C, 45.00; H, 7.06; N, 2.28. Found: C, 44.70; H, 6.60; N, 2.15.

[(LH)MoCl₂N] (2H)

NaN₃ (26.5 mg, 0.41 mmol, 2.50 equiv) was added to a solution of 1H (100 mg, 0.16 mmol, 1.00 equiv) in THF (5 mL) under a N₂ atmosphere. The solution was stirred overnight, filtered over Celite, and dried in vacuo. The residue was recrystallized from CH₂Cl₂/hexane at −30 °C and washed with Et₂O (3 × 2 mL). The resulting solid was dissolved in THF (3 mL, 25 mL Schlenk tube, borosilicate 3.3 glass), stirred, and irradiated with a 456 nm LED overnight (Kessil PR160L-456 nm, distance 5 cm). The resulting solution was filtered and recrystallized from THF/hexane at −30 °C to obtain 2H (57.2 mg, 0.97 mmol, 59%) as a brown/yellow powder. HR-MS (ESI): [C₂₃H₄₃ClMoN₂OP₂]⁺ ([M–Cl]⁺): calcd. m/z = 558.1586, found m/z = 558.1589. IR (KBr): ν̃ (cm^–1) = 3255 (br), 2949 (m), 2899 (m), 2866 (m), 1591 (s), 1472 (s), 1391 (w), 1367 (m), 1342 (m), 1261 (w), 1221 (w), 1175 (m), 1138 (w), 1030 (m), 935 (w), 866 (w), 810 (w). El. anal. calcd. for C₂₃H₄₃Cl₂MoN₂OP₂: C, 46.63; H, 7.32; N, 4.73. Found: C, 46.0; H, 7.23; N, 4.73.

General Procedure for Catalytic N₂ Reduction

Dry THF (6 mL) was added to a Schlenk tube filled with the catalyst (2.00 μmol, 1.00 equiv) and SmI₂(thf)₂ (197 mg, 360 μmol, 180 equiv) in a N₂ 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^tBu (∼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 N₂-cooled Schlenk flask containing HCl (2 M in Et₂O, 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-d₆ 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 N₂ 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 ⁿBu₄NPF₆ 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-d₈ was added. The mixture was analyzed by ³¹P NMR spectroscopy, and 3^– was quantified by integration vs the PF₆^– anion.

Results and Discussion

Synthesis and Characterization of the Complexes

The synthesis of the PNP pincer-type ligand LH was previously reported.³³ The reaction of [MoCl₃(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 × ³¹P) = 67 MHz, A(1 × ^95/97Mo) = 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.

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.

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

bond	distance	bond	distance
Mo–Cl1	2.4103(12)	Mo–N1	2.189(4)
Mo–Cl2	2.4108(11)	Mo–P1	2.6121(12)
Mo–Cl3	2.4515(12)	Mo–P2	2.6235(11)
C3–O1	1.352(5)	C1–C2	1.373(6)
C4–C5	1.384(6)

The molybdenum ion exhibits a distorted octahedral coordination sphere. The P–Mo–P angle is 153.32(4)°, and the N–Mo–Cl3 angle is 175.92(11)°. The Mo–N distance is 2.189(4) Å, and the Mo–P distances are 2.6235(11) and 2.6121(12) Å, which are very similar to the Mo–X distances in the corresponding Mo(III) complexes with 2,6-bis((di-tert-butylphosphanyl)methyl) pyridine ligands bearing different substituents in the para position of the ligand backbone.^6,7 The elongated C–O bond of 1.352(5) Å as well as the C1–C2 and C4–C5 distances of 1.373(6) Å and 1.384(6) Å, respectively, support protonation of the ligand.^33,34 This 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 = 1,8-diazabicyclo[5.4.0]undec-7-ene, pK_a = 16.9,³⁵Figure 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).

Figure 2.

Top left: UV/vis titration of 1H with increasing amounts of DBU (black: 0 equiv, red: 2 equiv), THF, c_1H = 0.15 mM; the inset shows the extinction coefficient at 305 nm (red) and 349 nm (black) vs the equivalents of DBU. Top right: CV data of 1H with varying amounts of DBU, c_1H = 1 mM, ν = 0.1 V s^–1, I = 0.2 M ⁿBu₄NPF₆, THF, N₂. Bottom left: UV/vis-SEC spectra recorded during reduction of 1H at an applied potential of −2.4 V vs Fc^+|0 (black: beginning, red: end), THF, c_1H = 0.15 mM, I = 0.2 M ⁿBu₄NPF₆. Bottom right: CV data of 1H under N₂, c_1H = 0.5 mM, ν = 0.1 V s^–1, I = 0.2 M ⁿBu₄NPF₆, THF.

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. 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 = −2.41 V with a shoulder at E_p,c,1 = −2.19 V and three associated irreversible reoxidations at E_p,a,1 = −0.85 V, E_p,a,2 = −0.31 V, and E_p,a,4 = 0.35 V under Ar (ν = 0.1 V s^–1, Figure 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.²⁸ This 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.³⁶ Upon reduction, homolytic O–H bond cleavages occur, forming the dianionic bipyridone ligand and H₂. 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 H₂ 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 H₂ 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 H₂.

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.³³ The 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

It sums up to 83 kcal mol^–1, which is similar to the BDFE of the O–H bond in phenol.³⁷

Treatment of 1H with a slight excess of NaN₃ 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 × ³¹P) = 40.6 MHz; A(2 × ¹H) = 4.2 MHz; A(1 × ¹⁴N) = 9.8 MHz; A(1 × ¹⁴N) = 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.^9,27,38 Reduction of 2H with an exc. of KC₈ or deprotonation of 2H with 1 equiv of DBU and reduction with an exc. of KC₈ led to the corresponding Mo(IV) nitrido complex 3K, which shows a characteristic signal in the ³¹P NMR spectrum at 85.2 ppm (Scheme 2 and Figure S19).³⁹ The 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 V (ν = 0.1 V s^–1, Figure S 9)_.

Reactivity towards N₂ 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 N₂-to-NH₃ formation catalyzed by I^I.⁸1H was treated with 180 equiv SmI₂(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 N₂-to-NH₃ 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 ¹⁵N₂ atmosphere led to the expected ¹H signal of ¹⁵NH₄⁺, confirming that N₂ 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 NH₃.

Table 2. Catalytic N₂-to-NH₃ Conversion^a.

^aInset: ¹H NMR spectra in DMSO-d₆ of the ammonium cation generated via N₂ reduction under ¹⁴N₂ (bottom) and ¹⁵N₂ (top) catalyzed by 1H.

To get further insights into the N₂ 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 ¹⁵N₂ with 2H revealed catalytic conversion into ¹⁵NH₃ and 1 equiv of ¹⁴NH₃ (Figure S23).

Reactivity towards N₂ under Electrochemical Conditions

To probe for electrochemical N₂ splitting, CV data under a N₂ atmosphere were recorded. Under a N₂ atmosphere, 1H shows four irreversible reductions at E_p,c,1 = −2.15 V, a shoulder at E_p,c,2 = −2.32 V, and a third and fourth event at E_p,c,3 = −2.46 V and E_p,c,4 = −2.57 V, 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 = −0.93 V 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 N₂ 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. I^Cl: E^1/2 = −1.94 V, E_p,c,1 = −2.48 V, E_p,c,2 = −2.71 V, E_p,c,3 = −3.03 V, ν = 0.1 V s^–1, Figure S7).

Since the CV data under a N₂ atmosphere revealed distinct differences compared to those under Ar, electrochemical dinitrogen splitting was investigated. As initial reduction is associated with H₂ formation, we utilized 1^– in the electrochemical N₂ 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 H₂ 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 PF₆^– signal (Table 3, entry 1, and Figure S25). Since reductive N₂ 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 N₂ 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 ³¹P NMR spectroscopy is a suitable method to quantify nitrido complex formation, it was also quantified via release as ammonium and subsequent ¹H NMR spectroscopy (Figure S30). Ammonium formation was observed after workup with a spectroscopic yield of 14%, confirming the reliability of quantification by ³¹P NMR spectroscopy (Table 3, entry 4). Electrochemical reduction of 1H under a ¹⁵N₂ atmosphere and subsequent acidic workup confirmed that N₂ is the N source in ammonium (Figure S31).

Table 3. Yields of the Mo(IV) Nitrido Complex after Electrochemical Reduction of 1H with 1 Equiv DBU in THF under N₂.

no.	Eappl vs Fc+|0	number of e–	quantified by	yield (%)
1	–2.69 V	4	31P NMR	7
2	–2.82 V	2	31P NMR	0
3	–2.82 V	3	31P NMR	12
4	–2.93 V	3	NH3 release	14
5	–2.82 V	4	31P NMR	4
6	–3.08 V	5	31P NMR	18
7	–3.15 V	3	31P NMR	31a

^aBoron-doped diamond electrode instead of GC.

To obtain further insights into the mechanism, IR-spectroelectrochemical investigation of 1H under reductive conditions under a N₂ atmosphere in an OTTLE cell using a gold grid electrode were conducted.⁴⁰ The IR data showed the appearance of three bands at 2037, 1947, and 1925 cm^–1 upon reduction (Figure 3). These bands are assigned to the vibrations of terminal N≡N triple bonds in the Mo(0) monomer [LMo(N₂)₃]⁻ (ν̃_NN = 1947 and 2037 cm^–1) and Mo(0) dimer [(L(N₂)₂Mo)₂μ-N₂]^2– (ν̃_NN = 1925 cm^–1), because similar bands appeared during electrochemical and chemical reduction of I^Br and have been assigned to an equilibrium of the monomer [L^PNPMo(N₂)₃] (ν̃_NN = 1963 and 2046 cm^–1) and the dimer [(L^PNP(N₂)₂Mo)₂μ-N₂] (ν̃_NN = 1944 cm^–1).²⁸ These bands disappear during a prolonged electrolysis time (Figure S14). The N₂ stretching vibrations of the intermediates during N₂ 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 N₂ activation degree. The vibrations were unambiguously assigned to N₂ vibrations by using isotope-labeled ¹⁵N₂. 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 ^14/15N₂, which fits well with the experimental data.

Figure 3.

Left: IR spectra recorded during reductive IR-SEC of 1H, v = 0.0025 V s^–1, c_1H = 10 mM; black: initial, blue: middle, green: last spectrum. Right: UV/vis spectra recorded during UV/vis-SEC of 1H at an applied potential of −2.7 V vs Fc^+|0, c_1H = 0.15 mM; black: initial spectrum; blue: last spectrum. Both THF, I = 0.2 M ⁿBu₄NPF₆, N₂.

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(N₂)₂Mo)₂μ-N₂]^2– comparing it with the parent [(L^PNP(N₂)₂Mo)₂μ-N₂] complex, which has a characteristic band at 688 nm (Figure 3 and Figure S13).²⁸

Considering the electronic structure requirements for N₂ splitting by dimeric [(M)₂μ-N₂] species in a pseudo D_4h symmetry, Mo(I) should be the active species as it exhibits the π¹⁰ configuration, which leads to the maximum weakening of the N₂ bond and the electronic requirements for forming the metal nitrido species.⁴¹ In line with these considerations, electrochemical oxidation of trans-(depe)₂Mo⁰(N₂)₂ induces N₂ splitting forming the nitrido complex.⁴² However, IR-SEC experiments of 1H under N₂ 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.^23,28 This was rationalized by the potential inversion of the reduction potentials after halide loss and N₂ 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 N₂ atmosphere.⁴³ Oxidation of the Mo(0) dimer by homogeneous electron transfer from, e.g., 1^– induces N₂ 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 N₂ splitting formation.

Scheme 3. Thermochemical Data of 1H and Proposed Mechanism for Electrochemical N₂ Splitting Forming the Mo Nitrido Complex 3^–

Conclusions

The molybdenum(III) complex 1^– with an anionic PNP pyridone pincer-type ligand has been investigated in N₂-to-NH₃ conversion and electrochemical N₂ splitting. The complex catalyzes the ammonia formation with SmI₂/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 N₂ 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 N₂ 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.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsorginorgau.3c00056.

• Instrumentation; syntheses; UV/vis titration of 1H; electroanalytical data; UV/vis spectroelectrochemistry of 1H; IR spectroelectrochemistry of 1H; EPR spectra; NMR spectra; data of the electrochemical N₂ splitting with 1^–; IR spectra; mass spectra; quantification of H₂; and X-ray crystallography (PDF)

Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. CRediT: Nils Ostermann data curation, formal analysis, investigation, validation, visualization, writing-original draft, writing-review & editing; Nils Rotthowe data curation, formal analysis, investigation, validation, visualization, writing-review & editing; A. Claudia Stückl data curation, validation, writing-review & editing; Inke Siewert conceptualization, data curation, formal analysis, funding acquisition, investigation, project administration, resources, supervision, validation, visualization, writing-original draft, writing-review & editing.

Funding from the International Center for Advanced Studies of Energy Conversion and Göttingen University is acknowledged.

The authors declare no competing financial interest.

Special Issue

Published as part of ACS Organic & Inorganic Auvirtual special issue “Electrochemical Explorations in Organic and Inorganic Chemistry”.