Crystalline Diuranium Phosphinidiide and μ‐Phosphido Complexes with Symmetric and Asymmetric UPU Cores

Abstract Reaction of [U(TrenTIPS)(PH2)] (1, TrenTIPS=N(CH2CH2NSiPri 3)3) with C6H5CH2K and [U(TrenTIPS)(THF)][BPh4] (2) afforded a rare diuranium parent phosphinidiide complex [{U(TrenTIPS)}2(μ‐PH)] (3). Treatment of 3 with C6H5CH2K and two equivalents of benzo‐15‐crown‐5 ether (B15C5) gave the diuranium μ‐phosphido complex [{U(TrenTIPS)}2(μ‐P)][K(B15C5)2] (4). Alternatively, reaction of [U(TrenTIPS)(PH)][Na(12C4)2] (5, 12C4=12‐crown‐4 ether) with [U{N(CH2CH2NSiMe2But)2CH2CH2NSi(Me)(CH2)(But)}] (6) produced the diuranium μ‐phosphido complex [{U(TrenTIPS)}(μ‐P){U(TrenDMBS)}][Na(12C4)2] [7, TrenDMBS=N(CH2CH2NSiMe2But)3]. Compounds 4 and 7 are unprecedented examples of uranium phosphido complexes outside of matrix isolation studies, and they rapidly decompose in solution underscoring the paucity of uranium phosphido complexes. Interestingly, 4 and 7 feature symmetric and asymmetric UPU cores, respectively, reflecting their differing steric profiles.

are unprecedented examples of uranium phosphido complexes outside of matrix isolation studies,and they rapidly decompose in solution underscoring the paucity of uranium phosphido complexes.I nterestingly, 4 and 7 feature symmetric and asymmetric UPU cores,r espectively,r eflecting their differing steric profiles.
Inrecent years there has been burgeoning interest in the synthesis and chemistry of uranium-ligand multiple bonds, [1] which stems from adesire to better understand the chemical bonding of uranium and to correlate this to observed physicochemical properties.H owever,m ost progress has been made regarding complexes where uranium engages in af ormal multiple bond to C-/N-/O-based donor ligands,a nd examples of second row-centered, and beyond, donor ligands generally continue to be rare. [2] Where uranium-phosphorus multiple bonding is concerned, [3] only two structurally authenticated phosphinidene complexes have been reported, [4] and investigations into uranium phosphido com-plexes are exceedingly rare and restricted to cryogenic matrix isolation and/or computational studies. [5] Thus,t here are no reports of uranium phosphido complexes on macroscopic scales under conditions that would permit further investigation;i ndeed, the phosphido linkage,w hether terminal or mbridging,r emains ar elatively rare structural motif even in transition-metal chemistry. [6] As part of our work on actinide-ligand multiple bonds, [7] we reported dithorium phosphido and arsenido complexes that are supported by the very sterically demanding triamidoamine ligand N(CH 2 CH 2 NSiPr i 3 ) 3 (Tren TIPS ). [7a,d] Fort he ThPThd erivative this ligand combination produced as eemingly optimal balance of steric shielding of the ThPThc ore versus inter-Tren TIPS steric repulsion. We therefore considered whether the analogous diuranium complex might be accessible;however,uranium has potentially deleterious and facile redox chemistry compared to the more redox-robust thorium, and is smaller than thorium by 0.05-0.18 , [8] so uranium with the same ligand set might well be too strained to form astable UPU linkage and could very easily decompose.H erein, however, we report two different methods for the bulk-scale preparation and subsequent characterization of diuranium mphosphido complexes,u tilizing Tr en TIPS and the related Tr en DMBS (Tren DMBS = N(CH 2 CH 2 NSiMe 2 Bu t ) 3 )l igands,t hat are the first examples of uranium phosphido complexes outside of cryogenic spectroscopic experiments. [5b,c] These complexes can be isolated and manipulated in the solid state, but we find that they are indeed highly sensitive and decompose rapidly in solution, which is in-line with the prior absence of any synthetically accessible actinide phosphido complexes.I nterestingly,d epending on the steric profiles of the Tr en ligands that support these phosphido complexes,s ymmetric and asymmetric UPU cores are observed in the solid state structures.
Complex 4 decomposes in solution, which, together with the low yield, precluded further characterization beyond the X-ray crystal structure and elemental analyses.T he reaction that produces 4 is highly capricious,a nd despite exhaustive attempts the reaction conditions could not be improved; sometimes deprotonation of 3 fails,o rc omplete decomposition occurs to unidentified products.Use of different organoalkali-metal reagents,t he presence or absence of different crown ethers,o ri ncreasing the molar quantity of benzyl potassium results in intractable reaction mixtures and/or production of the known uranium(IV) cyclometallate complex [U{N(CH 2 CH 2 NSiPr i 3 ) 2 (CH 2 CH 2 NSiPr i 2 C(H)(Me)-(CH 2 )}], [14] where the fate of the phosphorus-containing products could not be determined.
Thea bove mentioned observations likely reflect the inherently polarized, weak, and labile nature of these U-P linkages,asreflected by the paucity of any other macroscopic molecular uranium phosphido complexes,a nd also likely steric overloading from close proximity of two Tr en TIPS ligands.I no rder to reduce this steric strain and perhaps obtain am ore tractable phosphido complex, we adopted ad ifferent strategy to introduce as terically less demanding Tr en ligand.
Reaction of the new terminal uranium(IV)-phosphinidene complex [U(Tren TIPS )(PH)][Na(12C4) 2 ]( 5), [9] which is only the third example of auranium phosphinidene,with the uranium(IV) cyclometallate complex [U-{N(CH 2 CH 2 NSiMe 2 Bu t ) 2 CH 2 CH 2 NSi(Me)(CH 2 )(Bu t )}] (6) [15] proceeds by protonolysis to give the diuranium m-phosphido complex [{U(Tren TIPS )}(m-P){U(Tren DMBS )}][Na(12C4) 2 ] ( 7), isolated as dark brown crystals in 29 %y ield, Scheme 1. [9] Thecrystalline yield is low due to the oily nature of 7,and the decomposition that occurs once it is formed (see below). The solid-state crystal structure of 7,F igure 1, is in gross terms very similar to that of 4,noting the change of Tr en ligand and cation component. However,t he U-P distances of 2.657(2) and 2.713(2) are notable in that the shorter is consistent with the U-P distances in 4,b ut the longer is significantly longer and mid-way to the U-P distances in 3.I nterestingly, the shorter U-P distance is found for the Tr en TIPS -bound uranium with the longer U-P distance associated with the sterically less demanding Tr en DMBS portion, and the U À N bonds are longer in the Tren TIPS Up ortion of the molecule compared to those in the Tr en DMBS Uf ragment, perhaps reflecting the asymmetry of the phosphido bonding.
Thep resence of uranium(IV) ions in 7 was confirmed by SQUID magnetometry on ap owdered sample of 7;t he magnetic moments of 4.3 and 1.1 m B at 298 and 2K, respectively,are consistent with the presence of uranium(IV) ions.However the magnetic moment of 7 at 2Kis higher than the corresponding data for 3,which may represent the relative crystal-field effects on uranium(IV) from HP 2À versus P 3À ;the P 3À would be expected to present ag reater point charge and splitting of the paramagnetic excited states manifold, so alowlying group are still populated to some extent at low temperature with ah igher-lying group at high temperature that are more difficult to populate.T his notion is consistent with aslightly flatter magnetic trace at high temperature for 7 compared to 3 and has been noted in other uranium(IV) complexes with strong point-charge ligands. [2h, 3b,7f,h, 16] Interestingly,c ounter to expectations the shoulder at about 25 K for the magnetic data of 3 is much less pronounced for 7 which is consistent with our suggestion that this feature is due to single ion crystal field effects and not magnetic exchange, [12] though magnetic exchange cannot be completely ruled out.
Complex 7 is moderately more stable than 4,but although, once isolated, solid state characterization methods were feasible we find that redissolving 7 results in rapid decomposition so NMR and optical spectroscopic data were unobtainable.Interestingly,wefind that the majority decomposition products of 7 are the uranium(IV)-cyclometallate complex [U{N(CH 2 CH 2 NSiPr i 3 ) 2 (CH 2 CH 2 NSiPr i 2 C(H)(Me)-(CH 2 )}], [14] and what we deduce to be [U(Tren DMBS )(PH)]-[Na(12C4) 2 ], though the latter is not sufficiently sterically protected so decomposes to unidentified products.N evertheless,t he more clear-cut nature of the decomposition of 7 compared to 4 is instructive because it suggests that even with reduced ligand steric demands the UPU unit is inherently unstable.I nterestingly,t he decomposition reaction of 7 produces ac yclometallate with al ess-strained 5-membered metallocyclic ring compared to the more-strained 4-membered metallocycle in 6.This aspect is also consistent with the observation that mixing the five-membered-ring cyclometallate [U{N(CH 2 CH 2 NSiPr i 3 ) 2 (CH 2 CH 2 NSiPr i 2 C(H)(Me)-(CH 2 )}] [14] and known [U(Tren TIPS )(PH)][K(B15C5) 2 ] [4a] gives no reaction. Thus,t he importance of metallocyclic ring-strain as akey factor in driving the protonolysis reaction to generate 7 emerges.T his point is underscored when considering that on the basis of the solid-state structure the phosphido appears to be more associated with the Tr en TIPS U fragment rather than the Tr en DMBS Ug roup,b ut it is the Tr en TIPS Uf ragment that is,i ne ssence,t he leaving group during decomposition.
To gain ag reater understanding of the bonding in the UPU units of 4 and 7,wecarried out DFT calculations on the full anion components of these compounds, 4 À and 7 À , respectively.Considerable difficulty was encountered obtaining SCF-converged structures,w hich suggests that 4 À and 7 À have multi-reference ground states.H owever, satisfactorily converged models that provide aqualitative description of the electronic structure of these compounds could be obtained.
Both 4 À and 7 À exhibit four unpaired electrons of essentially exclusive 5f character in their a-spin manifolds as HOMO to HOMOÀ3, which is consistent with the presence of two 5f 2 uranium(IV) ions.H OMOÀ4t o HOMOÀ6i ne ach case represent the principal bonding components in the UPU units,see Figure 2a nd the Supporting Information, [9] confirming the presence of polarized uranium-phosphido triple bonding interactions.The uranium spin densities of À2.31/À2.33 and À2.17/À2.22, for 4 À and 7 À respectively,s how donation of electron density from the ligands to uranium and support the uranium(IV) formulations.T he uranium charges are high for Tr en uranium(IV) complexes, [17] at + 3.79/ + 3.86 for 4 À and + 3.52/ + 3.87 for 7 À and the phosphido charges are À2.19 and À2.35, respectively. Interestingly,t he uranium ion in 7 À which has the closest association with the phosphido,t hat is,T ren TIPS U, has the highest charge and lowest spin density,and recall that the U-Ndistances are longer for that unit than the Tr en DMBS Uunit; this suggests that the Na toms are better as au nit at charge donation to uranium than the phosphido. [18] TheUPMayer bond orders reflect multiple,but polarized bond interactions.S pecifically,i n4 À they are 1.41/1.43 whereas for 7 À they are 1.44/1.66 reflecting the asymmetric UP distances and bonding in the UPU core in 7 À ;n otably, these UP bond orders are in-line with the situation in the Lewis bonding scheme for these units,that is,U =P=U. These Mayer bond orders should be viewed in the context that the UN amide and UN amine bond orders are 0.71 and 0.18, respectively,a nd they are surprisingly invariant across 4 À and 7 À .
Thed ata above unequivocally suggest that the UPU interactions in 4 À and 7 À are polarized and weak, which is consistent with the observed instability of 4 and 7.I nterestingly,t he UP bonds for 4 À and 7 À have higher Mayer bond orders,e xhibit more metal component, and utilize more 5f character (relative to 6d) than the ThPb onds in [{Th-(Tren TIPS )} 2 (m-P)][Na(12C4) 2 ], [7d] consistent with the general view that uranium engages in more covalent bonding,a nd with greater 5f character,than thorium, but we note that the bond topological data are essentially invariant for uranium and thorium. This suggests that the instability of 4 and 7 is most likely of kinetic origin.
To conclude,w eh ave reported two structurally authenticated examples of uranium phosphido complexes.T hese linkages are unprecedented outside of cryogenic matrix isolation conditions,r emain rare even in the d-block, and indeed uranium-phosphorus multiple bonding remains exceedingly rare overall. These complexes have been prepared on macroscopic scales by two different methodologies that could greatly expand uranium-phosphido chemistry: 1) construction of aU P(H)U unit by salt elimination and subsequent deprotonation;or2)protonation of acyclometallate by aparent phosphinidene.Although both complexes can be prepared and isolated they exhibit intrinsic instability that is consistently reflected in quantum chemical calculations. Low-temperature magnetism studies also suggest differences in the relative crystal-field effects on uranium(IV) from HP 2À versus P 3À .M ost intriguingly,t he UP bond lengths can be perturbed by co-ligand steric demands,w hich suggests that with suitably chosen co-ligands perhaps aU PU linkage,o r perhaps aU PM unit that might be prepared by method (2), could be polarized to the point of rupture in order to produce at erminal uranium phosphido complex under ambient conditions.Efforts in that regard are on-going. Figure 2. Kohn-Shamf rontier molecular orbitals of that represent the principal bonding components of the UPU unit in the anion component of 7, 7 À :L eft, HOMOÀ6( 403a, À1.332 eV);Middle, HOMOÀ5( 404a, À0.997 eV);R ight, HOMOÀ4(405a, À0.983 eV). Hydrogen atoms are omitted for clarity.

Conflict of interest
Theauthors declare no conflict of interest.