Reductive silylation of a uranyl dibenzoylmethanate complex: an example of controlled uranyl oxo ligand cleavage †

and


Results and discussion
To expand the scope of borane-mediated silylation of uranyl, the utility of dibenzoylmethanate, dbm (dbm ¼ OC(Ph)CHC(Ph) O), as a uranyl supporting ligand was probed.5][46] Thus, we endeavoured to synthesize a uranyl dibenzoylmethanate complex that contained THF as a co-ligand.Reaction of 2 equiv. of Na(dbm), generated in situ, with UO 2 Cl 2 (THF) 3 results in formation of a light orange solution, from which UO 2 (dbm) 2 (THF) can be isolated as an orange powder in 71% yield.This complex features a singlet at 7.32 ppm in its 1 H NMR spectrum (CD 2 Cl 2 ), which is assignable to the g-CH of the dbm ligand.In addition, broad singlets at 4.99 and 2.47 ppm, conrm the presence of THF in the uranyl coordination sphere.UO 2 (dbm) 2 (THF) had been reported previously, 47 but had not been fully characterized.It is closely related to several other uranyl bis(b-diketonate) complexes that have been reported in the literature, 46,48 including UO 2 (acac) 2 (THF), 49 UO 2 (dbm) 2 (dmso), 45 and UO 2 (dbm) 2 (H 2 O). 44ith UO 2 (dbm) 2 (THF) in hand we evaluated the strength of its U]O bonds relative to the previously characterized b-ketoiminate complex, UO 2 ( Ar acnac) 2 .A cyclic voltammogram of UO 2 (dbm) 2 (THF) in CH 2 Cl 2 reveals an irreversible reduction feature at E ¼ À1.19 V (vs.Fc/Fc + ), measured at a scan rate of 0.1 V s À1 , which we attribute to the UO 2 2+ /UO 2 + redox couple (Fig. S1 †).This feature is irreversible at all scan rates.Importantly, this value is less negative than that observed for UO 2 ( Ar acnac) 2 (E 1/2 ¼ À1.35 V vs. Fc/Fc + ), 50 conrming that the dbm equatorial ligand is less electron donating than the Ar acnac ligand, and suggesting a lesser degree of oxo ligand activation in UO 2 (dbm) 2 (THF).For further comparison, UO 2 (dbm) 2 (dmso) features a reversible UO 2 2+ /UO 2 + redox couple at E 1/2 ¼ À1.36 V (vs.Fc/Fc + , in dmso), 45 while UO 2 (dbm) 2 (dmf) features a reversible UO 2 2+ /UO 2 + redox couple at E 1/2 ¼ À1.46 V (vs.Fc/Fc + , in dmf). 46These lower redox potentials undoubtedly reect the strong donating ability of dmso and dmf vs. THF.In addition, UO 2 (dbm) 2 (THF) features a U]O n sym mode of 823 cm À1 in its Raman spectrum (Fig. S2 †).For comparison, the U]O n sym mode for UO 2 ( Ar acnac) 2 was determined to be 812 cm À1 , 51 which reveals that the U]O bonds in UO 2 (dbm) 2 (THF) are stronger than those in UO 2 ( Ar acnac) 2 , and further supports the claim that the dbm ligand is less electron donating.This latter point is a critical because it will allow us to evaluate the effect of a weaker donating equatorial ligand on both reduction and functionalization.Previously, we hypothesized that only strong donor ligands, such as Ar acnac, were able to activate uranyl toward functionalization. 51pon establishing that dbm was a weaker donor than Ar acnac, we subjected UO 2 (dbm) 2 (THF) to our reductive silylation protocol.Thus, addition of 1 equiv. of HSiPh 3 to UO 2 (dbm) 2 (THF), in the presence of 1 equiv. of B(C 6 F 5 ) 3 , results in the formation of a deep red solution, from which U(OB {C 6 F 5 } 3 )(OSiPh 3 )(dbm) 2 (THF) (1) can be isolated as a dark red crystalline material in 62% yield (eqn (1)).Similarly, addition of 1 equiv. of HSiEt 3 to UO 2 (dbm) 2 THF, in the presence of 1 equiv. of B(C 6 F 5 ) 3 , results in the formation of U(OB {C 6 F 5 } 3 )(OSiEt 3 )(dbm) 2 (THF) (2), which can be isolated in 55% yield (eqn (1)).Isolation of both 1 and 2 proceed with higher yield if 0.25 equiv. of THF is added to the mother liquor.The reductive silylation of UO 2 (dbm) 2 (THF) is similar to that observed previously by our research group for the uranyl bketoiminate complex, UO 2 ( Ar acnac) 2 . 22,23Most importantly, the observation that the stronger U]O bonds of UO 2 (dbm) 2 (THF), relative to UO 2 ( Ar acnac) 2 , are also susceptible to reductive silylation suggests that the scope of this transformation is broader than originally thought.
Complexes 1 and 2 both crystallize in the triclinic space group P-1 as a hexane solvate, 1$C 6 H 14 , and a toluene and hexane solvate, 2$C 7 H 8 $0.5C 6 H 14 , respectively.The solid-state molecular structure of 2 is shown in Fig. 1.Both 1 and 2 exhibit pentagonal bipyramidal geometries, as determined from the inter-ligand bond angles.3][54] In both complexes, one uranyl oxo ligand has been converted to a silyloxide group, while the other oxo ligand is coordinated to a molecule of B(C 6 F 5 ) 3 , as was observed for U(OB{C 6 F 5 } 3 )(OSiPh 3 )( Ar acnac) 2 . 22or complex 1, the U-O Si and U-O B bond lengths are 2.024(2) and 1.9521(19) Å, respectively, while for 2, they are 2.011(2) and 1.9600 (19) Å, respectively (Table 1).These values are comparable to those previously reported for U(V)-silyloxide and U(V)-OB(C 6 F 5 ) 3 distances, 13,15,22,23 and are indicative of a substantial reduction of the U]O bond order.Interestingly, the U-O dbm bond lengths in 1 (av.U-O ¼ 2.281 Å) and 2 (av.U-O ¼ 2.282 Å) (Table 1) are shorter than those observed in other uranyl dbm complexes (ca.2.35 Å). 55 Finally, both 1 and 2 feature a THF molecule coordinated to the uranium center.This contrasts with the reductive silylation product of UO 2 ( Ar acnac) 2 , for which no coordinated solvent is observed, a consequence of the reduced steric prole of the dbm ligand vs. the much bulkier Ar acnac ligand.
Interestingly, crystallization of 2 without the addition of 0.25 equiv. of THF to the supernatant led to the isolation of a second, minor product, U(k 2 -O,F-OB{C 6 F 5 } 3 )(OSiEt 3 )(dbm) 2 (3), as redorange crystals in low yield (eqn ( 2)).Complex 3 could not be completely separated from complex 2 and so could not be fully characterized.Nonetheless, we were able to perform a single crystal X-ray diffraction study on this molecule.Complex 3 crystallizes in the triclinic space group P 1 and its solid-state molecular structure is shown in Fig. 2 ) 4 (Ph F ¼ C 6 F 5 ), U IV (NPhPh F ) 4 , and U III (NPh F 2 ) 3 (THF) 2 exhibit three or more F / U interactions each. 56The U-F distance for complex 3 (2.654(2)Å) falls on the shorter end of U-F dative interactions, which range from $2.60-2.93Å. 10,56 The 19 F{ 1 H} NMR spectrum of 3 consists of two resonances at À160.20 and À165.53 ppm in a 1 : 2 ratio, which are assignable to the p-and m-uorine atoms of the C 6 F 5 groups.In addition, a very broad resonance assignable to the o-uorine atoms is observed at À149.25 ppm.Notably, this resonance is shied signicantly upeld in comparison to those observed for 1 and 2, suggestive of some interaction with the paramagnetic U(V) centre. 56However, the observation of only a single peak for the o-uorine atoms is indicative of free rotation about the B-C bond.Also present in the spectrum are resonances at 161.6 and 166.3 ppm, which are attributable to complex 2. Interestingly, complexes 2 and 3 are also both observed in the in situ 19 F{ 1 H} NMR spectrum of the reaction between UO 2 (dbm) 2 (THF), HSiEt 3 , and B(C 6 F 5 ) 3 (Fig. S9 †).We suggest that complexes 2 and 3 are in equilibrium, and addition of THF to the mother liquor during crystallization favours the formation 2, permitting its isolation in higher yields.
Given the rarity of well-dened oxo ligand substitution reactions for the uranyl moiety, we explored the ligand exchange reactivity of this new family of functionalized uranyl complexes.We hypothesized that the small steric prole of the equatorial dbm ligands would allow for facile axial ligand exchange.Gratifyingly, the addition of 1 equiv. of H(dbm) to 2 in CH 2 Cl 2 results in the formation of U(OB{C 6 F 5 } 3 )(dbm) 3 (4), which could be isolated as dark red crystalline material in 33% yield (eqn (3)).The isolation of complex 4 represents a rare example of controlled uranyl oxo ligand cleavage at ambient temperature and pressure.
Complex 4 crystallizes in the triclinic space group P1 as a toluene and hexane solvate, 4$2C 7 H 8 $C 6 H 14 , with two independent molecules in the asymmetric unit.Its solid-state molecular structure is shown in Fig. 3.The uranium ion in complex 4 is coordinated by three dbm ligands and a B(C 6 F 5 ) 3capped oxo ligand.While the geometry about the uranium center in complex 4 can be described as a distorted pentagonal bipyramidal (CSM ¼ 3.80), according to the continuous shape measure developed by Alvarez and co-workers, 57 it is probably better described as a distorted capped trigonal prism (CSM ¼ 1.27), wherein the three dbm ligands dene the trigonal prism and the O(B{C 6 F 5 } 3 ) ligand forms the capping group.The U-O B bond lengths of the two independent molecules (1.96(2) and 1.93(2) Å, Table 1) are comparable to those observed for complexes 1, 2, and 3, but longer than that observed for the U(V) mono-oxo complex, U(O)(NR 2 ) 3 (R ¼ SiMe 3 ), which features a U-O bond length of 1.817(1) Å. 58 The elongated U-O bond in 4 is clearly the result of borane coordination to the oxo ligand.The U-O distances associated with the dbm oxygen atoms that are situated trans to the O(B{C 6 F 5 } 3 ) ligand are 2.14(2) and 2.25(2) Å, while the average U-O dbm-cis bond length is 2.27(4) Å.
Interestingly, the X-ray diffraction data for complex 4 are suggestive of the presence of the Inverse Trans Inuence (ITI), 53,[59][60][61] with the average trans U-O bond length being 0.07 Å shorter than the average cis bond (averaged over the two independent molecules in the asymmetric unit).However, it should be noted that the diffraction data for 4 are of modest quality, which leads to large uncertainties in the metrical parameters.
We therefore turned to computational chemistry in the form of density functional theory to explore the possibility of an ITI in 4. Initial geometry optimization using the GGA PBE functional suggested that, if the ITI is present, it is very small, with a trans  shortening of only 0.018 Å.However, the overall agreement between theory and experiment, although adequate (mean absolute deviation (MAD) between the calculated and experimental U-O bond lengths of 0.021 Å), prompted us to re-optimize the geometry with the hybrid PBE0 functional.Agreement between theory and experiment is better at this level (MAD ¼ 0.012 Å), and PBE0 also suggests a more pronounced ITI of 0.063 Å, much closer to the experimental value.
The ITI was rst suggested by Denning in 1992 (ref. 62) in relation to oxy anions, such as [UOCl 5 ] À .Experimentally, the ITI in this system is very pronounced, at 0.103 Å as determined by X-ray crystallography.For comparison, we have calculated the geometry of [UOCl 5 ] À at both the PBE and PBE0 levels, obtaining an ITI of 0.044 Å and 0.069 Å, respectively.It would therefore appear that PBE0 describes the ITI better than PBE in [UOCl 5 ] À , providing justication for its use in calculating the geometry of complex 4. One explanation for the ITI, rst proposed by Denning 63,64 is that hybridization of the actinide 6p and 5f orbitals enhances s bonding to the strongly bound trans directing ligand and leads to a partial hole in the 6p shell directed toward the trans ligand.This 6p hole enhances 5f overlap in the trans position, leading to a shortening of the trans bond.At the NPA/PBE0 level, we nd the 6p populations of [UOCl 5 ] À and 4 to be 5.876 and 5.915, respectively, further supporting the suggestion of an ITI in complex 4.
Starting from the fully optimized geometries of [UOCl 5 ] À and 4, we have conducted relaxed potential energy surface scans of the bond trans to the oxo ligand, altering the trans bond length in steps of AE0.025 Å to a limit of AE0.1 Å from equilibrium.The results are shown in Fig. 4, and reveal that these potential surfaces are very at.Compression (the steeper, le side of the well) of the trans bond in [UOCl 5 ] À by 0.1 Å raises the energy of the anion by only ca. 6 kJ mol À1 , and by less than 4 kJ mol À1 for 4. It would therefore appear that the ITI is a rather subtle effect, even in prototypical systems such as [UOCl 5 ] À .It is also interesting to note that for complex 4, the energetic gain on moving from a structure where the cis and trans distances are about the same (i.e., no ITI) to the fully optimised structure is about 1 kJ mol À1 .This is much smaller than the 6 kcal mol À1 stabilization afforded by the ITI in [(( tBu ArO) 3 tacn)U(O)(OTf)], 65 which likely reects their different oxidation states and the coordination of B(C 6 F 5 ) 3 to the oxo ligand in 4.
The 1 H NMR spectrum of 4 in CD 2 Cl 2 consists of four broad resonances at 8.22, 7.68, 6.70 and 6.24 ppm in a 1 : 2 : 4 : 4 ratio, which corresponds to the four dbm proton environments and indicates that there is only one dbm environment observed at room temperature.In addition, the 19 F{ 1 H} NMR spectrum of 4 consists of three resonances at À144.72, À160.57, and À165.98 ppm, in a 2 : 1 : 2 ratio, corresponding to the o-, p-, and m-uorine atoms of the C 6 F 5 groups.Finally, the near-IR spectrum for 4 is similar to those of other U(V) complexes, 10,15,22,23 supporting the presence of a 5f 1 ion.DFT also supports this description of the electronic structure of complex 4. The uranium spin density at the Mulliken and Hirshfeld levels is 1.12 and 1.06, respectively, and examination of the a and b spin valence molecular orbitals nds an a spin orbital, with 62% uranium 5f character (Mulliken analysis), which has no b spin equivalent (Fig. 5).
To determine the fate of the missing Et 3 SiO-group upon formation of 4, we monitored the reaction of 2 with 1 equiv. of H(dbm) by NMR spectroscopy.The in situ 19 F{ 1 H} NMR spectrum of the reaction mixture revealed the formation of complex 4, as evidenced by a characteristic resonance at À144.8 ppm, along with the presence of complex 2. Complexes 2 and 4 were observed in a 2.4 : 1 ratio, respectively, according to the integrations of their o-uorine resonances (Fig. S18 †).More importantly, the in situ 13 C{ 1 H} NMR spectrum of the reaction mixture reveals the formation of HOSiEt 3 , as evidenced by resonances at 6.21 and 5.56 ppm (Fig. S17 †). 66The proposed reaction stoichiometry was further conrmed by following the reaction of 4 with 1 equiv. of HOSiEt 3 , and 1 equiv. of THF, in CD 2 Cl 2 by 1 H and 19 F{ 1 H} NMR spectroscopies, which reveals the formation of complex 2 and H(dbm), along with complete consumption of complex 4 (Fig. S19 and S20 †).This transformation represents a rare example of a controlled, reversible  uranyl U]O bond cleavage, in which the fate of the cleaved oxo ligand has been explicitly determined. 9,12,36,67,68Reaction of 1 with 1 equiv. of H(dbm) in CD 2 Cl 2 also results in formation of 4, as determined by 1 H and 19 F{ 1 H} NMR spectroscopies.This experiment reveals the presence of complexes 1 and 4 in a 3 : 2 ratio, respectively.(Fig. S14 and S15 †).

Conclusions
Reaction of UO 2 (dbm) 2 (THF) with 1 equiv. of HSiR 3 (R ¼ Ph, Et), in the presence of 1 equiv. of B(C 6 F 5 ) 3 , results in formation of U(OB{C 6 F 5 } 3 )(OSiR 3 )(dbm) 2 (THF) (R ¼ Ph, 1; Et, 2) via oxo ligand silylation.The isolation of complexes 1 and 2 demonstrates that the borane-activated silylation of the uranyl oxo ligand does not require the highly donating b-ketoiminate ligand, Ar acnac, to proceed.Instead, oxo ligand silylation can be achieved with weaker donors attached to the uranyl equatorial sites.This work further demonstrates the generality of the borane-mediated reductive silylation protocol.Interestingly, reaction of 2 with 1 equiv. of H(dbm) results in formation of U(OB{C 6 F 5 } 3 )(dbm) 3 ( 4), along with HOSiEt 3 .We propose that this oxo ligand substitution chemistry is possible because of the narrow steric prole of the dbm ligand, which permits the coordination of the three dbm moieties to uranium, in addition to the borane-capped oxo ligand.Finally, complex 4 has been determined to show an inverse trans inuence (ITI), based on comparison of diffraction and density functional theory data.The potential well for distorting the trans U-O from its equilibrium position is found computationally to be very at, suggesting that the ITI is a subtle effect.For future studies, we plan to explore whether borane-activated silylation can proceed with cationic uranyl complexes, as the oxo ligands in these species are anticipated to be substantially less nucleophilic than those in a neutral molecule.

General
All reactions and subsequent manipulations were performed under anaerobic and anhydrous conditions under an atmosphere of nitrogen.Hexanes, diethyl ether, and toluene were dried using a Vacuum Atmospheres DRI-SOLV solvent purication system, and stored over 3 Å molecular sieves for 24 h prior to use.CH 2 Cl 2 and CD 2 Cl 2 were dried over activated 3 Å molecular sieves for 24 h before use.THF was distilled twice, rst from calcium hydride and then from sodium benzophenone ketyl, and stored over 3 Å molecular sieves for 24 h prior to use.UO 2 Cl 2 (THF) 3 was synthesized by the published procedure. 69UO 2 (dbm) 2 (THF) was synthesized by modifying the previously reported procedure for the preparation of UO 2 (hfac) 2 (THF) (see below). 49,70,71All other reagents were purchased from commercial suppliers and used as received.NMR spectra were recorded on a Varian UNITY INOVA 400 MHz spectrometer or a Varian UNITY INOVA 500 MHz spectrometer. 1 H and 13 C{ 1 H} NMR spectra are referenced to external SiMe 4 using the residual protio solvent peaks as internal standards ( 1 H NMR experiments) or the characteristic resonances of the solvent nuclei ( 13 C NMR experiments). 19F { 1 H} NMR spectra were referenced to external CFCl 3 in C 6 D 6 .Raman and IR spectra were recorded on a Mattson Genesis FTIR/Raman spectrometer with a NXR FT Raman Module.IR samples were recorded as KBr pellets, while Raman samples were recorded in an NMR tube as neat solids.UV-vis/NIR experiments were performed on a UV-3600 Shimadzu spectrophotometer.Elemental analyses were performed by the Microanalytical Laboratory at UC Berkeley.

Cyclic voltammetry measurements
CV experiments were performed with a CH Instruments 600c Potentiostat, and the data were processed using CHI soware (version 6.29).All experiments were performed in a glove box using a 20 mL glass vial as the cell.The working electrode consisted of a platinum disk embedded in glass (2 mm diameter), the counter electrode was a platinum wire, and the reference electrode consisted of AgCl plated on Ag wire.Solutions employed during CV studies were typically 1 mM in the metal complex and 0.1 M in [Bu 4 N][PF 6 ].All potentials are reported versus the [Cp 2 Fe] 0/+ couple.For all trials, i p,a /i p,c ¼ 1 for the [Cp 2 Fe] 0/+ couple, while i p,c increased linearly with the square root of the scan rate (i.e., Ov).

X-ray crystallography
The solid-state molecular structures of complexes 1-4 were determined similarly with exceptions noted in the following paragraph.Crystals were mounted on a cryoloop under Paratone-N oil.Data collection was carried out on a Bruker KAPPA APEX II diffractometer equipped with an APEX II CCD detector using a TRIUMPH monochromator with a Mo Ka X-ray source (a ¼ 0.71073 Å).Data for 1, 2, and 4 were collected at 100(2) K, while data for 3 were collected at 150(2) K, using an Oxford nitrogen gas cryostream system.A hemisphere of data was collected using u scans with 0.3 frame widths.Frame exposures of 5, 10, 10 and 10 seconds were used for complexes 1, 2, 3, and 4 respectively.Data collection and cell parameter determination were conducted using the SMART program. 72Integration of the data frames and nal cell parameter renement were performed using SAINT soware. 73Absorption correction of the data was carried out empirically based on reection j-scans using the multi-scan method SADABS. 74Subsequent calculations were carried out using SHELXTL. 75Structure determination was done using direct or Patterson methods and difference Fourier techniques.All hydrogen atom positions were idealized, and rode on the atom of attachment.Structure solution, renement, graphics, and creation of publication materials were performed using SHELXTL. 75omplex 2 exhibits positional disorder of the toluene solvent molecule.The positional disorder was addressed by modeling the molecule in two orientations, in a 50 : 50 ratio.The EADP, DFIX, and FLAT commands were used to constrain both orientations of the toluene molecule.For complex 4, every nonhydrogen atom in one of uranium molecule was constrained using the EADP command to its symmetry equivalent atom on the other uranium molecule.Two toluene solvent molecules were not rened anisotropically.In addition, the C-C bonds of the toluene rings were constrained with the DFIX command, while the rings were constrained with the FLAT command.Hydrogen atoms were not assigned to disordered carbon atoms.A summary of relevant crystallographic data for 1-4 is presented in Table S2.†

Computational details
Density functional theory calculations were carried out using the PBE 76,77 and PBE0 functionals 78 as implemented in the Gaussian 09 Rev. C.01 (ref.79) quantum chemistry code.A (14s 13p 10d 8f)/[10s 9p 5d 4f] segmented valence basis set with Stuttgart-Bonn variety relativistic effective core potential was used for U. Dunning's cc pVTZ basis sets were employed for oxygen and boron, while other atoms were treated at the cc-pVDZ level.The ultrane integration grid was employed, as were the default geometry and SCF convergence criteria.Natural population analyses were performed using the GenNBO6 code, 80 using 0.47 les from G09 as input.
1 H NMR spectrum of 1 in CD 2 Cl 2 consists of four broad resonances at 10.76, 4.75, 4.54, and 3.60 ppm in a 4 : 4 : 2 : 1 ratio, respectively, which correspond to the four proton environments of the dbm ligand.Additionally, three sharper resonances are observed at 7.53, 7.41, and 6.22 ppm in a 2 : 1 : 2 ratio, which correspond to the m-, p-, and o-proton environments of the Ph 3 Si group.Similarly, the 1 H NMR spectrum of 2 in CD 2 Cl 2 consists of four broad resonances at 7.40, 6.66, 6.26 and 4.54 ppm in a 2 : 4 : 4 : 1 ratio, respectively, as well as two broad resonances at 4.94 and 3.48 ppm, which correspond to the two Et 3 Si proton environments.The 19 F{ 1 H} NMR spectrum of 1 consists of three resonances at À136.21, À160.49, and À165.75 ppm, in a 2 : 1 : 2 ratio, corresponding to the o-, p-, and m-uorine atoms of the C 6 F 5 groups.Similarly, the 19 F{ 1 H} NMR spectrum of 2 consists of three resonances at À135.00, À160.69, and À165.86 ppm, in a 2 : 1 : 2 ratio.Finally, the near-IR spectra for 1 and 2 are similar to those of other U(V) complexes (see Fig. S29 and S30 †), 2. The U-O Si and U-O B bond lengths of 3, 1.981(3) and 1.915(2) Å, respectively, are comparable to those observed in complexes 1 and 2. In contrast, the U-O-B bond angle (151.6(2) ) in 3 is considerably smaller than those observed in 1 (172.03(16) ) and 2 (165.80(18) ), likely due to the presence of a F / U dative interaction between an o-uorine atom of the B(C 6 F 5 ) 3 moiety and the uranium centre, which occurs in place of ligation of the THF solvate molecule.Interestingly, F / M dative interactions in uranium organometallics are quite rare and to our knowledge have only been observed in four other complexes.[Cp* 2 Co][U{OB(C 6 F 5 ) 3 } 2 -( Ar acnac)(OEt 2 )] 10 exhibits two F / U dative interactions, while U IV (NPh F

Fig. 5
Fig. 5 Three dimensional representation of the uranium 5f-based a spin molecular orbital of 4. Isosurface value ¼ 0.05.Hydrogen atoms omitted for clarity.