Dinitrogen cleavage and hydrogenation to ammonia with a uranium complex

ABSTRACT The Haber–Bosch process produces ammonia (NH3) from dinitrogen (N2) and dihydrogen (H2), but requires high temperature and pressure. Before iron-based catalysts were exploited in the current industrial Haber–Bosch process, uranium-based materials served as effective catalysts for production of NH3 from N2. Although some molecular uranium complexes are known to be capable of combining with N2, further hydrogenation with H2 forming NH3 has not been reported to date. Here, we describe the first example of N2 cleavage and hydrogenation with H2 to NH3 with a molecular uranium complex. The N2 cleavage product contains three uranium centers that are bridged by three imido μ2-NH ligands and one nitrido μ3-N ligand. Labeling experiments with 15N demonstrate that the nitrido ligand in the product originates from N2. Reaction of the N2-cleaved complex with H2 or H+ forms NH3 under mild conditions. A synthetic cycle has been established by the reaction of the N2-cleaved complex with trimethylsilyl chloride. The isolation of this trinuclear imido-nitrido product implies that a multi-metallic uranium assembly plays an important role in the activation of N2.


Experimental Procedure
General Procedure: All manipulations were performed under an atmosphere of argon or nitrogen using standard Schlenk techniques or a glovebox. Commercially available chemicals were purchased from TCI, Aladdin and J&K Scientific Ltd., and used as received without further purification unless otherwise stated. The solvents were obtained by passing through a Solve Purer G5 (MIKROUNA) solvent purification system and further dried over 4 Å molecular sieves. THF-d8 were dried over Na/K and stored under an Ar or N2 atmosphere prior to use. Nuclear magnetic resonance spectroscopy was performed using a Bruker AVIII-400, a Bruker AVIII-500 or a Bruker AVIII-600 spectrometer at room temperature (RT). The 1 H and 13 C{ 1 H} NMR chemical shifts (δ) are reported relative to tetramethylsilane, and 31 P{ 1 H} NMR chemical shifts are relative to 85% H3PO4. Absolute values of the coupling constants are provided in Hertz (Hz). Multiplicities are abbreviated as singlet (s), doublet (d), triplet (t), multiplet (m), and broad (br). Magnetic measurements on crystalline samples were performed using a Quantum Design SQUID VSM magnetometer from 300 to 1.8 K under an external magnetic field of 1000 Oe. The sample was added to a pre-weighed SQUID capsule in a glovebox. The capsule was then sealed, weighed, and transferred to the SQUID cavity for the magnetic measurements. All magnetic data were corrected for the diamagnetic contributions of the sample holder and of the core diamagnetism of the samples using Pascal's constant. 1 Fourier transform infrared spectra (FT-IR) were measured on a Nicolet FT-IR 170X spectrophotometer in the range of 4000-400 cm -1 at 25 °C using KBr plates. X-ray photoelectron spectroscopy (XPS) was recorded on an EscaLab 250Xi photoelectron spectrometer.
Elemental analyses (C, H, N) were performed on a Vario EL III elemental analyzer at the Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences. GC-MS experiments were performed on Agilent 5977B. Complex 1 and KC8 were prepared according to previously reported procedures. 2,3

Synthesis of [{[U{N(CH3)(CH2CH2NP i Pr2)2}(μ-NH)]3(μ-N)}K2] (3) and [{[U{N(CH3)(CH2CH2NP i Pr2)2}(μ-NH)]3(μ-15 N)}K2] (3-15 N)
A solution of complex 2 (134 mg, 0.2 mmol, 1 equiv.) in THF (2 mL) and toluene (5 drops) was added dropwise to a suspension of KC8 (162 mg, 1.2 mmol, 6 equiv.) in THF (2 mL) under an atmosphere of argon or nitrogen. The mixture was stirred overnight at RT and then the solvents were removed under reduced pressure and the residues were extracted with toluene. The filtrate was concentrated to 1 mL and placed at RT for 48 h. Dark-brown crystals of 3 suitable for X-ray diffraction were obtained (under Ar: 20 mg, 16%; under N2: 39 mg, 31%). Attempts to improve the crystallized yield of complex 3 were unsuccessful although it was the major product in the in-situ reactions. 3-15 N was prepared by exposing 2 to 15 N2. A mixture of 2 (134 mg, 0.2 mmol, 1 equiv.) and KC8 (162 mg, 1.2 mmol, 6 equiv.) in THF (4 mL) was freeze pump thaw degassed three times and exposed to an atmosphere of 15 N2 (1 atm), following the same procedure that was used with 3 to afford 3-15 N. Complex 3 could be also formed by the reaction of complex 2 with excess KC8 in THF in the presence of 9.10-dihydroanthracene. Complex 3 could not be obtained when the reaction was conducted in the absence of toluene and 9.10-dihydroanthracene.

Hydrogenation of complexes 3 and 3-15 N with H2
In an NMR tube, a brown solution of 3 (5.0 mg, 0.0026 mmol) in 0.5 ml of THF-d8 was freezedegassed three times and then exposed to 1 atm of H2 at RT.  Figures   Fig. S1. The 1 H NMR (THF-d8, 500 MHz) spectrum of complex 2 (# are peaks assigned to THF and * are peaks assigned to impurities in THF-d8).    Complex 3 (the labelled peaks) was the sole identifiable product in this reaction.

X-ray crystallographic analysis
The crystallographic data were collected using a Bruker APEX-II CCD area detector with a radiation source of Ga(K) (1.34139 Å) or Mo(K) (0.71073 Å). Multi-scan or empirical absorption corrections (SADABS) were applied. The structures were solved using Patterson methods, expanded with difference Fourier syntheses, and refined using full-matrix least squares fitting on F 2 using the Bruker SHELXTL-2014 program package. 4,5 All non-hydrogen atoms were refined anisotropically. Hydrogen atoms were introduced at their geometric positions and refined as riding atoms. In complex 2, the restraints (SIMU) were employed to displacement parameters of N4, N5 and N6. In complex 3 (prepared under Ar), pseudo-isotropic (ISOR) restraints were applied for the refinement of the disordered atoms. Two alternative orientations for the heavy atoms were refined and resulted in site occupancies of 98% and 2% for U1 and U1', U2 and U2', U3 and U3'. In complex 3 (prepared under N2), the restraints (SADI, SIMU and RIGU) were used to refine the disordered atoms. Two alternative orientations for one carbon atom of isopropyl were refined and resulted in site occupancies of 71% and 29% for the C42 and C42'. Evaluation of the CIF using the CheckCIF routine at www.checkcif.iucr.org gave no A or B alert for these complexes. These data can be obtained free of charge from the Cambridge Crystallographic Data Centre (www.ccdc.cam.ac.uk/data-request/cif). Details regarding the data collection and refinement for these complexes are given in Table S1.

Theoretical Calculations
All calculations were carried out at the DFT level of theory using the hybrid functional B3PW91 [6][7] with the Gaussian 09 suite of programs. 8 The U and P atoms were represented with a small-core Stuttgart-Dresden relativistic effective core potential associated with their adapted basis set. [9][10][11] Additionally, the P basis set was augmented by a d-polarization function (α = 0.340) 12 to represent the valence orbitals. All the other atoms C, N, K and H were described with a 6-31G (d,p) doubleζ quality basis set. [13][14] The nature of the extrema (minimum) was established with analytical frequencies calculations and geometry optimizations were computed without any symmetry constraints. The enthalpy energy was computed at T = 298 k in the gas phase. Dispersion corrections were included by means of single point energy calculations using the GD3BJ approach. 15 The redox potential K/K + was estimated from our calculation and compared to the experimental value using the methodology published by Castro et al. 16 The computed redox potential is -3.01 V vs. -2.99 V experimentally which supports the correctness of our modelling of the electron transfer properties of KC8. Spin-orbit corrections were not considered in this study since the molecular environment will mainly quenched the spin-orbit effect due to the ligand field and the strong hybridization of the uranium atomic orbitals.. The differential SO correction was found to be 1.3 kcal/mol in the redox step suing the CIPSO methods. 17,18 Two facts suggest the significance of the potassium. First, the release of N2 from the monomeric form of 2' in the absence of K was calculated to imply an activation barrier of 46.7 kcal mol -1 (28.6 kcal mol -1 from the uncapped monoazide complex). The N-N bond cleavage implies a single electron transfer from the uranium center to the azide ligand. Indeed, at the TS1, the unpaired spin density appears to be distributed between U (2.19), in line with a U(IV) system, and the two terminal nitrogen atoms of the azide ligand (0.54 for the nitride and 0.42 for the N2), that are stabilized by the potassium cation. At the TS2, the assistance from the potassium is again crucial since an interaction between one potassium and the phenyl ring of the toluene is observed, allowing the proton transfer.