Molecular and electronic structure of terminal and alkali metal-capped uranium(V) nitride complexes

Determining the electronic structure of actinide complexes is intrinsically challenging because inter-electronic repulsion, crystal field, and spin–orbit coupling effects can be of similar magnitude. Moreover, such efforts have been hampered by the lack of structurally analogous families of complexes to study. Here we report an improved method to U≡N triple bonds, and assemble a family of uranium(V) nitrides. Along with an isoelectronic oxo, we quantify the electronic structure of this 5f1 family by magnetometry, optical and electron paramagnetic resonance (EPR) spectroscopies and modelling. Thus, we define the relative importance of the spin–orbit and crystal field interactions, and explain the experimentally observed different ground states. We find optical absorption linewidths give a potential tool to identify spin–orbit coupled states, and show measurement of UV···UV super-exchange coupling in dimers by EPR. We show that observed slow magnetic relaxation occurs via two-phonon processes, with no obvious correlation to the crystal field.


Supplementary Note 1: Equations 1-5
Through analytical solution to the axial CF Hamiltonian !" the CFPs can be extracted from the orbital energies by solving the simultaneous equations below, where E ±3 = 0.

Supplementary Note 2: General comments on new synthesis of uranium-nitrides
We identified the limiting factor in the synthesis of 3Na from 2 to be the low solubility of NaN 3 in pyridine, and since ethers and other polar solvents (e.g. acetonitrile, dichloromethane) react with 2 and 3Na, such solvents cannot ameliorate this issue.
Permutations of 12C4 addition alongside NaN 3 to increase the solubility of NaN 3 resulted in poorer yields, presumably due to sodium ion abstraction before it can stabilise the nitride installation, which has been shown to be key to preparing Tren TIPS -supported terminal uranium nitrides. 1,2 It is instructive to inspect the electron-transfer and bond-breaking/-forming chemistry that occurs in the synthesis of 3Na: Formally, the sodium ion has donated its valence electron to the azide, so the uranium(III) ion provides two electrons to effect ejection of dinitrogen and installation of a trianionic nitride. As well as the solubility issues of NaN 3 , we surmised that having to establish a U-N bond during the reaction must provide decomposition opportunities. We postulated whether we could start with the azide group pre-coordinated to uranium, which would require uranium(IV) to avoid premature azide activation, and initiate dinitrogen elimination, oneelectron oxidation of uranium, and nitride formation by the addition of an external oneelectron reductant: Gratifyingly, methods based on this reaction give straightforward, reliable access to the DCIP series 3M in respectable yields and permit us to greatly expand our uranium(V)-nitride family. Additional benefits include: (i) the synthesis of 7 has been improved (analytically pure yields of ~ 94%), giving higher yields of uranium(V)-nitrides overall, whereas the preparation of 2 is much less convenient; (ii) uranium(V)-nitrides with all alkali metal cations are now straightforwardly accessible, which may be desirable in subsequent reactions; (iii) avoiding the use of pyridine, the only solvent in which the prior route works, simplifies workup and isolation.
We attribute the collective failure of multiple co-ligands to abstract lithium from 3Li to a combination of the strongly nucleophilic nitride and the small, highly polarising lithium. This is consistent with the straightforward removal of the larger, softer heavier alkali metals from 3Na -3Cs with the appropriate crown ethers to give the SIPs 4Na -4Cs. When the sizematching is not optimised the nucleophilic nitride does not release the alkali metal and CIPs 5M are isolated.

Supplementary Note 3: Comments on structural analyses
The metrical data for all of the complexes discussed here are largely as would be expected, but notably the U-N amine distances in 3M, 4M, and 5M are longer than for 1 [2.482(6) Å] 3 and 6 [2.465(5) Å], 2 which has been attributed to the presence of an inverse-trans-influence (ITI); 2-4 on the basis of these bond lengths the ITI does not seem to be present in the uranium(V)-nitrides described here. There is no obvious trend in the 3M series with respect to whether the UN(M) 2 NU cores are planar or trans-bent, but we suggest the observed geometries are modulated by crystal packing forces since the potential energy surfaces of the UN(M) 2 NU cores must be considered shallow. This is credible as the M-N nitride interactions will be electrostatic and thus easily perturbed. The grouping of the U≡N bond lengths of 3Li -3K (long) versus 3Rb and 3Cs (short) can be rationalised on the basis that the former metals would be expected to polarise and lessen the electron density in the U≡N linkage more than the latter pair; the longest U≡N bond for 3K may additionally be the result of the slight pyramidalisation of the nitride centre in this complex. Regarding the 5M series, the metrical data for the U≡N-M angle most likely reflect the structural deformations required to enable the {M(crown)} + unit to optimise its approach to the [U(Tren TIPS )(N)] − fragment, but crystal packing forces must play a role; in support of this view, it is notable that cesium, which should present the most polar N-M bond, exhibits the most acute U-N-M angle.

Experimental
General: All manipulations were carried out using Schlenk techniques, or an MBraun UniLab glovebox, under an atmosphere of dry nitrogen. Solvents were dried by passage S44 through activated alumina towers and degassed before use or were distilled from calcium hydride. All solvents were stored over potassium mirrors, except for ethers and pyridine which were stored over activated 4 Å sieves. Deuterated solvent was distilled from potassium, degassed by three freeze-pump-thaw cycles and stored under nitrogen. Sodium azide was dried under vacuum for four hours prior to use. Metallic sodium was purchased as 25-35 wt% dispersion in paraffin, washed with dry diethyl ether, dried under vacuum and stored at −30 °C in the glove box. Crown ethers were dissolved in ether, dried over activated 4 Å molecular sieves for 24 hours, decanted and the ether removed prior to use.
[U(Tren TIPS )(Cl)] [Tren TIPS = {N(CH 2 CH 2 NSiPr i 3 ) 3 } 3-; Pr i = CH(CH 3 ) 2 ] was prepared as described previously. 1 1 H and 29 Si NMR spectra were recorded on a Bruker 400 spectrometer operating at 400.2 and 79.5 MHz respectively; chemical shifts are quoted in ppm and are relative to TMS. UV/Vis/NIR spectra were recorded on a Perkin Elmer Lambda 750 spectrometer. Data were collected in 1 mm path length cuvettes loaded in an MBraun UniLab glovebox and were run versus the appropriate reference solvent. Note the f-f transition at ~4,700 cm −1 is partially obscured by imperfect background subtraction of solvent absorbance.
Variable-temperature (1.8 -300 K) magnetic susceptibility (0.1 T dc field) and variable field (0-7 T) magnetisation data were recorded on a Quantum Design MPMS XL7 superconducting quantum interference device (SQUID) magnetometer equipped with a 7 T magnet. Ac susceptibility data were collected under either a 0.1 T dc field or a zero dc field, using a small ac field of 1.55 Oe oscillating at frequencies between 1 and 1400 Hz. For all measurements, doubly recrystallised powdered samples were carefully checked for purity and data reproducibility between several independently prepared batches for each compound examined. Care was taken to ensure complete thermalisation of the sample before each data point was measured and samples were immobilised in an eicosane matrix to prevent sample reorientation during measurements. Diamagnetic corrections were applied using tabulated Pascal constants and measurements were corrected for the effect of the blank sample holders (flame sealed Wilmad NMR tube and straw) and eicosane matrix. Solution magnetic moments were recorded at room temperature using the Evans method. Variable temperature EPR spectra were measured at X-band (ca. 9.4 GHz) and Q-band (ca. 34 GHz) on either a Bruker EMX 300 or a Bruker ElexSys E580 spectrometer equipped with an ER4119HS-W1 (X-band) or an ER 5106QT-E resonator. Gently ground polycrystalline samples were vacuum-sealed in 2.3 mm o.d. quartz tubes, to match both resonators. For X-band measurements, an additional 4 mm o.d. quartz tube was used as a support. Measurements were made on several independently prepared batches, which were also analysed by SQUID and CHN microanalyses, to ensure reproducibility. Spectra were background corrected against blank sample holders measured under identical conditions. In some samples we observe very weak, extremely sharp features in the g ~ 2 region which we believe are not intrinsic to the complexes; although the low-lying j z ≈ ±3/2 doublet (see main text) would have g z ≈ 2, the lineshapes would be absorption-like similar to the g ~ 3.74(9) feature and thus are not consistent with such an assignment. FTIR spectra were recorded on a Bruker Tensor 27 spectrometer. 50% 15 N-labelled analogues of 3-5 were prepared using sodium azide-1-15 N following procedures identical to those described below. We previously reported the observation of U≡N stretches in the FTIR spectra of 3Na and 4Na at ~930 cm -1 that shifted to ~900 cm -1 in the 15 N-labelled isotopomers, although the latter was complicated by overlapping bands from 12C4. An absorption at 936 cm -1 in the FTIR spectrum of SIP 4K decreases in intensity by approximately 50% upon 15 N-labeling. This is accompanied by the growth of a band at 900 cm -1 for 15 N-4K, giving an isotopomer shift of 36 cm -1 . This compares well to the calculated isotopomer shift to 906 cm -1 from reduced mass considerations. We assign these bands as U≡N stretches as they are identical to the bands and isotopomer shifts observed for 4Na and similar values have been reported for uranium-nitride species prepared in matrix isolation experiments. However, we have been unable to conclusively identify isotopomer shifts for the remaining nitrides because of complex band structures in the fingerprint regions of their FTIR spectra. Attempts were made to record the Raman spectra of 3M, 4M, and 5M but samples decompose in the beam (even at low power settings) and dilution experiments were inconclusive, precluding data acquisition and analysis.
Cyclic voltammetry experiments were attempted but these nitride complexes react with THF, chlorinated, and acetonitrile solvents and/or electrolytes so it was impracticable to collect electrochemical data.
CHN microanalyses were carried out by Tong Liu at the University of Nottingham. Some CHN data are persistently low, which is attributed to a combination of incomplete combustion from carbide/nitride formation, which has precedent with organo-silicon rich complexes, 5 but multiple measurements on multiple samples ensure the reliability of the data.

Preparation of [U(N)(Tren TIPS )] (6)
Note: Due to the light sensitive nature of 6 all manipulations should be conducted in the absence of light.
A solution of I 2 (60 mg, 0.24 mmol) in toluene (5 ml) was added dropwise over 20 minutes to a stirring solution of 4K (0.64 g, 0.48 mmol) in toluene (10 ml) at -78 °C. The brown mixture was allowed to warm to room temperature with stirring over 16 h. Volatiles were removed in vacuo and the product was washed with pentane at -78 °C (2 × 5 ml). The solid was dried in vacuo to yield 6 as a red/brown powder and is stored in the dark at −30 °C. Yield: 0.24 g, 59%. The identity of 6 was confirmed by comparison to previously reported data. 2