Synthesis, characterization, and transformation. Treatment of the complex {U[N(CH3)(CH2CH2NPiPr2)2](Cl)2(THF)} (1)33 with 2 equiv. of NaN3 at room temperature (RT) in tetrahydrofuran (THF) results in the formation of a molecular chain, [{[U{N(CH3)(CH2CH2NPiPr2)2}(N3)](µ-N3)}n] (2), which was isolated in 72% yield as a crystalline product (Fig. 1d). The 1H NMR spectrum of 2 has eleven peaks between + 69.64 and − 70.48 ppm (Fig. S1), indicating an unsymmetrical structure. The Fourier-transform infrared (FT-IR) spectrum of 2 exhibits two strong azide stretching bands at 2094 cm− 1 and 2150 cm− 1 (Fig. S2), which are characteristic of actinide-bound terminal and bridged azide ligands, respectively.−
The molecular structure of complex 2 was confirmed by single-crystal X-ray diffraction, which exhibits a novel one-dimensional molecular chain connected through 1,3-end-on bridged azides (Fig. 2a). The U–Nazide distances between the uranium and bridged azide are 2.447(3) Å for U1-N3 and 2.471(4) Å for U1’-N1, which are clearly longer than the terminal U–Nazide distance (U1–N6 2.292(3) Å) but are similar to those found in previously reported multimetallic uranium complexes with bridged azide ligands.36,, The N1–N2 and N2–N3 bond lengths are 1.159(4) and 1.169(4) Å, respectively, suggesting that a delocalized−N = N+=N− resonance form is predominant in this 1,3-end-on bridged azide. However, the terminal azide shows a more localized form with N4–N5 and N5–N6 bond lengths of 1.145(5) and 1.181(5) Å, respectively, suggesting weak activation of this terminal azide.
Uranium azides are well-known as synthetic precursors of terminal uranium nitride species in photochemical or redox processes.35,36,, We first investigated the photolysis or thermolysis of complex 2, but the products were unidentifiable. However, when complex 2 was reduced with an excess of KC8 in a toluene-THF mixed solvent under an N2 atmosphere, the color of the solution changed immediately from yellow-brown to dark-brown (Fig. 1d). Upon work-up of this mixture, a crystalline complex, [{[U{N(CH3)(CH2CH2NPiPr2)2}(µ-NH)]3(µ-N)}K2] (3) was isolated in 31% yield. The 1H NMR spectrum of 3 exhibits paramagnetic resonance signals between + 73.69 and − 33.22 ppm (Fig. S4). Complex 3 could also be prepared by the reduction of 2 with KC8 under Ar, albeit with a lower crystalline yield (16%). Attempts to increase the yield of complex 3 were unsuccessful although it was the major product in the in-situ reaction of complex 2 with KC8 (Fig. S5). Toluene dimers were detected by gas chromatography-mass spectrometric (GC-MS) analysis of the reaction mixture (Fig. S6), which is consistent with the failure to obtain complex 3 when the reaction was conducted in the absence of toluene. Complex 3 could be also formed by the reaction of complex 2 with an excess of KC8 in THF in the presence of 9,10-dihydroanthracene (Fig. S7). These results suggest that the proton source plays an important role in the formation of 3. The presence of N-H groups in 3 was confirmed by the formation of HN(SiMe3)2 in the reaction of 3 with TMSCl (vide infra) and the broad absorption at 3450 cm− 1 in the FT-IR spectrum of complex 3 (Fig. S3) .35,36
The molecular structure of complex 3 was confirmed by X-ray crystallography (Fig. 2b). The salient feature of this structure is the presence of three U centers bridged by three imido µ-NH ligands and one nitrido µ3-N ligand. The six U–Nimido bond lengths fall in the range of 2.188 − 2.231 Å and are comparable with the bridged U–Nimido distances reported previously in the range of 2.10–2.55 Å.32,, These near-equivalent U–Nimido bond lengths suggest that it is not a U–N = U bonding interaction but a U–NH–U unit by the comparison with reported analogues.44, The three U-Nnitrido distances which range from 2.204 to 2.218 Å are comparable to those found in U(IV)/U(VI) tetrauranium nitride clusters (2.183(7)-2.319(78) Å) and slightly longer than those found in a µ3-N nitride uranium complex (2.138–2.157 Å). Despite their bridging nature, the U-Nimido and U-Nnitrido bond lengths in 3 remain slightly shorter than the sum of the single bond covalent radii of U and N (2.41 Å). Although similar trinuclear imido-nitride structures with transition metals have been described by the groups of Roesky, Yélamos, and Hou,19 complex 3 represents the first example of a trinuclear uranium species with both imido and nitrido ligands.
To verify the source of the nitrido ligand in complex 3, we reduced complex 2 with KC8 under 1 atm of 15N2. After the acidification of the 15N-labelled product (3-15N) with excess pyridine hydrochloride, a triplet resonance (δ = 7.42 ppm, JNH = 52 Hz, assigned to NH4Cl) and a doublet resonance (δ = 7.42 ppm, JNH = 72 Hz, assigned to 15NH4Cl) were observed in a ratio of approximately 3:0.5 in its 1H NMR spectrum in deuterated dimethyl sulfoxide (Fig. S12). Exposing the THF solution of complex 3 to 1 atm 15N2 for 2 days does not reveal any exchange between 3 and 15N2 (Fig. S13). Thus the generation of 15NH4Cl reveals that the 15N2 cleavage was involved in the formation of complex 3. The lower ratio of 15NH4Cl (the ideal ratio for 14NH4+:15NH4+ is 3:1) is presumably because three molecules of 14N2 are generated from the reduction of U-N3 even under the 15N2 conditions. This result is consistent with the isolation of complex 3 with lower yield when the reduction of complex 2 took place under an Ar atmosphere, in which the N2 was generated in-situ by the reduction of N3− units. Furthermore, the in-situ 1H NMR spectrum for the reaction of 2 with KC8 under dynamic vacuum reveals that no complex 3 was formed (Fig. S14). These studies demonstrate that the nitrido ligand in complex 3 originates from N2. Fixing and activation of N2 derived from the reduction of metal azides is very rare for either d- or f-block metals. Liddle and co-workers isolated a uranium(v)–bis(imido)–dinitrogen complex, [U(BIPMTMS)(NAd)2(µ-η1:η1-N2)(Li-2,2,2-cryptand)], by reacting a uranium-carbene species with an organoazide under an N2 or Ar atmosphere. The N-N length of the coordinated N2 in this species (1.139(9) Å) is only slightly elongated over the N-N length in the free N2 molecule (1.0975 Å). Therefore, complex 3 represents the first example of N2 scission in a metal-azide reduction.
Direct hydrogenation of the N2-activated product with H2 under mild conditions is desirable. Accordingly, complex 3 was treated with H2 at atmospheric pressure and RT (Fig. 1d). The in-situ 1H NMR and 31P{1H} NMR spectra show that complex 3 was consumed within 8 h and the free ligand was formed in the reaction (Figs. S15 and S16). The formation of NH3 in 34% yield was identified by the formation of NH4Cl after treating the volatiles with an excess of PyHCl (Fig. S17). The NH3 formation in this hydrogenation process was further confirmed by the reaction of complex 3 with 1 atm D2 at RT, which affords ND3 as confirmed by the 2H NMR spectrum (Fig. S18). This process represents the first example of NH3 production by the hydrogenation of an N2-activated product with H2 or D2 in a uranium system. The reaction of 15N-labelled product (3-15N) with H2 generates NH3 and 15NH3 (Fig. S19), which shows that the imido and nitrido groups in complex 3 were converted to NH3. In addition, by the reaction of complex 3 with an excess of Me3SiCl (TMSCl) at RT for overnight, the uranium precursor (complex 1, 46% yield) and N-containing products (HN(SiMe3)2 and N(SiMe3)3) were formed (Figs. S20 and S21). Therefore, a synthetic cycle has been established and the uranium-containing precursor can be reused (Fig. 1d).
The oxidation state of the uranium center in complex 2 is + IV, which was confirmed by the variable-temperature magnetic data determined by a super-conducting quantum interference device (SQUID) in the solid state (Fig. 3). The magnetic moment of complex 2 is 3.37 µB at 300 K and smoothly decreases to 0.45 µB at 1.8 K, then approaches zero (Fig. 3a). The magnitude of µeff and temperature dependence of 2 are consistent with a U(IV) center which is a magnetic singlet at low temperatures.36,41,− Formally, complex 3 contains two U(IV) centers and one U(V) center. The measured magnetic moment at 300 K for complex 3, which contains three U ions, is 5.72 µB, and also exhibits significant temperature dependence decreasing steadily to 1.39 µB at 1.8 K (Fig. 3b). The UV-Visible-NIR absorption spectra of complexes 2 and 3 were recorded in THF at RT (Fig. S24). Complex 2 displays moderate absorption in the 300–450 nm range, while 3 exhibits a significantly more intense absorption than that of 2 over the entire visible and NIR region. Both 2 and 3 exhibit several weak absorption peaks (ε < 50 and 200 M− 1 cm− 1 for 2 and 3, respectively) in the NIR region, which is characteristic of f-f transitions involving the 5f1 and 5f2 electronic configuration.52,− X-ray photoelectron spectroscopy (XPS) was undertaken to investigate the probable oxidation state of uranium in complex 3 (Fig. S25). The binding energy for U-4f7/2 in the XPS of 3 was observed to be 380.52 eV, which is in the range of the binding energies for U(IV) and U(V) species., These results are consistent with the assignment of U(IV)/U(IV)/U(V) to complex 3 and the overall charge of this cluster was balanced with two K+ ions.
The formation of imido ligand in complex 3 is proposed to involve a terminal uranium nitride, which was formed by the reduction of uranium azide 2 with KC8 and followed by protonation to the imido ligand by toluene or 9.10-dihydroanthracene. Attempts to isolate this terminal uranium nitride intermediate at low temperatures and/or with diminished levels of KC8 were unsuccessful. Reduction and oligomerization of metal azides induced by photolysis or thermolysis are quite common, but the observed subsequent cleavage of the N ≡ N triple bond in N2 is a hitherto unknown process for either d- or f-block elements.
Theoretical studies. To further investigate the conversion of 1 to 3, density functional theory (DFT) calculations were carried out using the B3PW91 functional that have been proved to be reliable in dealing with such systems.32–34 Dispersion corrections were considered and appeared to be small in this case (Figs. S28 and S29). The Gibbs free energies were reported and the difference with the enthalpy barriers was only 1.0 kcal mol− 1.
Experimentally, complex 2 is a coordination polymer that could not be computationally modeled. However, the formation of a monomer (2’) was predicted to be favorable by 17.0 kcal mol− 1 (Fig. 4). The main geometrical features of 2 were correctly reproduced in the computed monomeric form 2’. For instance, the U-N distances are 2.26 Å vs. 2.24 Å (exp) and the N-N distances are 1.21/1.15 Å vs. 1.18/1.15 Å (exp.). The unpaired spin density is 2.166, consistent with a U(IV) center. Reduction of the diazide 2’ is predicted to be almost athermic (0.8 kcal mol− 1) and is assisted by the coordination of a K+ to form a monoazide complex A.36,42 In the absence of potassium coordination, the reduction is endothermic by 28.6 kcal mol− 1 (Fig. S29). The reduction of the U(IV) center is highlighted by the unpaired spin density value of 3.12 at the uranium center of A, which is consistent with a U(III) system. This monoazide complex A stabilized by a potassium cation can further evolve to a nitride complex B by losing N2. A potassium-mediated N2-release transition state (TS1) has been located and the associated barrier is 15.5 kcal mol− 1, in line with a facile process.
The significance of the potassium is evidenced by two facts. 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, Fig. S29). Interestingly, 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 (2.193), in line with a U(IV) system, and the two terminal nitrogen atoms of the azide ligand (0.541 for the nitride and 0.415 for the N2), that are stabilized by the potassium cation. Following the intrinsic reaction coordinate, TS1 releases N2, forming a nitrido-type intermediate B, whose formation from the diazide complex 2’ is slightly endothermic by 3.2 kcal mol− 1 (Fig. 4). Intermediate B is better described as a U(IV)-(nitride radical)-K(I) complex since the unpaired spin density is distributed between U (2.254) and N (0.700). The presence of an unpaired electron at the nitride indicateds that this intermediate is fairly unstable and will react further with potassium to yield a dianionic uranium-nitride (C) stabilized by two potassium counterions (-66.3 kcal mol− 1 from 1, -49.3 kcal mol− 1 from the diazide 2’). The intermediate C is also interesting because the unpaired spin density indicates that the potassium does not reduce the metal but reduces the nitride, yielding a U(IV)-N(-III)-2K(I) species (unpaired spin density of 2.125 on U, and 0.607 and 0.426 on the two K ions).
Intermediate C is therefore sufficiently nucleophilic to abstract a hydrogen from the solvent. The relevant TS2 was located and the associated barrier is 9.4 kcal mol− 1 indicating a facile reaction (Fig. 5). The charge on the hydrogen (+ 0.22) shows that this reaction is a proton transfer rather than a hydrogen atom transfer reaction (HAT) and this is possibly due to the presence of a U(IV)-N(-III) unit in C. At the TS2, the assistance from the potassium is again crucial since an observed interaction between one potassium and the phenyl ring of the toluene, allows the proton transfer. Following the intrinsic reaction coordinate, this leads to the formation of a PhCH2K adduct to the U(IV)-imido complex D (-85.2 kcal mol− 1 from the entrance channel), which can easily trimerize while losing PhCH2K to form complex E (-119.1 kcal mol− 1). The trimetallic tri-imido species E can then bind N2 to form F (-113.3 kcal mol− 1). In F, the three uranium centers are U(IV) (unpaired spin densities of ~ 2.17 per U) and the N2 is not reduced (N-N bond distance of 1.19 Å) but binds to the uranium (U-N2 WBI of 0.26). Interestingly, the LUMO of F is a bonding interaction between the uranium centers and the π* of N2 (Fig. S30).
Finally, reduction of F yields species G (-103.4 kcal mol− 1), in which the N2 moiety is triply reduced and has an N-N bond distance of 1.3 Å with an unpaired spin density of 0.8 on N2. This implies that one uranium center has been oxidized, as evidenced by the unpaired spin densities of U (Fig. 5). Complex G is thus half-way to complex 3, and 3 is formed upon coordination of a second complex E. The formation of 3 from 1 is exothermic by -131.2 kcal mol− 1. The optimized geometry of 3 compares well with the experimental observations. The U-N bond distances for example are correctly reproduced (between 2.17 and 2.24 Å vs. between 2.17 and 2.23 Å exp.). Scrutiny of the unpaired spin density in 3 allows determination of two U(IV) and one U(V) in this species (Fig. 6). Upon comparison of the uranium oxidation states found for the nitride and imido complexes, the formation of 3 implies a N ≡ N bond cleavage via trimetallic uranium synergy, but only a single electron oxidation of one uranium center, which is surprising at first glance. However, the other electrons used for the reduction are stored in the potassium.