Superseding β‐Diketiminato Ligands: An Amido Imidazoline‐2‐Imine Ligand Stabilizes the Exhaustive Series of B=X Boranes (X=O, S, Se, Te)

Abstract Boron reluctantly forms B=X (X=O, S, Se, Te) moieties, which has stimulated the quest for such species in the past few years. Based on the N,N′‐chelating β‐diketiminato ligand (HNacNac), a new amido imidazoline‐2‐imine ligand system (HAmIm) is presented, giving rise to the isolation of an exhaustive series of Lewis acid free, monomeric chalcogen B=X boranes with documented π‐bond character between boron and the chalcogen. The chalcogenoboranes are isoelectronic and isolobal to the respective ketones. The chemical behavior of the oxoborane (B=O) strongly resembles the classical carbonyl reactivity in C=O bonds. The improved stability provided by HAmIm arises from the formation of more‐stable five‐membered boron chelates versus the six‐membered NacNac analogues and from the imidazoline‐2‐imine moiety providing enhanced σ‐ and π‐donation. The HAmIm ligand class may supersede the widely employed NacNac system in certain applications.

IR spectra were recorded on a Bruker Vertex 70 with the KBr disc transmission technique.
Mass spectra were recorded on a Finnigan MAT 8400-MSS I instrument (for electro spray ionization, ESI) or on a Finnigan MAT 4515 instrument (electron impact mode, EI) and are reported as the m/z ratio (in Da).
Elemental analyses were accomplished by combustion and gas chromatographic analysis using a VarioMICRO Tube and HW detection. Values are reported in weight-%.

Synthesis of Compound 7-Br.
A side arm flask (the neck should be large in diameter) was charged with dichloromethane (100 mL) and 7-BBr4 (3.00 g, 4.14 mmol, 1 eq.). Silver tetrafluoroborate (AgBF4, 1.60 g, 8.28 mmol, 2 eq.) was added in one portion with the application of an intense stream of nitrogen. After 5 min the mixture was cannula-filtrated. The solvent of the filtrate was removed in vacuo. Diethyl ether (50 mL) was added to the honelike residue and stirred for 3 h. The solution was decanted. The solid was dried in vacuo, which gave compound 7-Br (1.90 g, 3.28 mmol, 80 %) sufficiently pure for further reactions. Crystals suitable for X-ray crystallography ( Figure S12) and elemental analysis were obtained by layering diethyl ether with a solution of 7-BBr in dichloromethane and were found to have the composition (7-Br)2 • Et2O.

Synthesis of Compound 9.
Compound 3 (200 mg, 0.37 mmol, 1 eq.) was dissolved in THF (10 mL) and a solution of lithium hexamethyl disilylamide (LiHMDS, 472 µL, 1 M in THF, 2 eq.) was added with stirring. After 1 h a small amount of a precipitate was formed, which was removed by cannula-filtration. Layering of n-pentane over the THF crude solution afforded crystals suitable for X-ray crystallography, which showed the intercalation of an undefined amount of THF in the void volume of the lattice, and was thus treated with the SQUEEZE-tool in OLEX 2 , see Figure S20.
The filtrate was further reduced to dryness in vacuo. The final compound 9 can only hardly be redissolved in THF, but readily dissolves in dichloromethane. Samples for NMR characterization and elemental analysis were obtained by dissolution of the material in dichloromethane (5 mL) and precipitation of product 9 upon addition of npentane (10 mL), which afforded 9 as a white powder (120 mg, 0.22 mmol, 60 %).
Potassium hexamethyl disilylamide (KHMDS, 131 mg, 1 eq.) and 2.2.2-cryptand (C18N2H36O6, 248 mg, 1 eq.) were dissolved in THF (5 mL) at room temperature. This solution was added to the solution aforementioned at ambient temperature and stirring was continued for 30 min with formation of a white precipitate, which was removed by cannula-filtration. The solvent was removed to dryness.
The solid residue was treated following the aforementioned procedure with two additional cycles. The final solid was then extracted with Et2O (15 mL). The extract was concentrated to 2-3 mL and subjected to layering with pentane, which afforded colorless crystals of compound 10 (163 mg, 0.36 mmol, 55 %).

Synthesis of Compound 11.
Compound 7-Br (200 mg, 0.33 mmol, 1 eq.) was suspended in 1,2-dimethoxyethane (10 mL) and Li2S (30 mg, 1.34 mmol, 2 eq.) was added. The suspension was heated to 50 °C overnight. Volatile components were removed in vacuo, and the product 11 was extracted with benzene (15 mL) in the form of its adduct with lithium bromide. The addition of 12-crown-4 (C8H16O4, 5.8 mg, 0.33 mmol, 1 eq.) to the benzene solution and stirring for 30 min led to a precipitate of lithium bromide crown ether complex, which was removed by cannula-filtration. The filtrate was condensed to a volume of ca. 5 mL under reduced pressure. Slow solvent evaporation gave colorless crystals (also suitable for X-ray crystallography) of the composition 6 • C6H6 (107 mg, 0.196 mmol, 60 %).
Elemental analysis was performed with samples obtained from the crystallization procedure.

Synthesis of Compound 12.
Compound 7-Br (200 mg, 0.33 mmol, 1 eq.) was suspended in 1,2-dimethoxyethane (10 mL) and Li2Se (61 mg, 0.66 mmol, 2 eq.) was added. The suspension was heated to 50 °C overnight. Volatile components were removed in vacuo, and the product 12 was extracted with benzene (15 mL) in the form of its adduct with lithium bromide. The addition of 12-crown-4 (C8H16O4, 5.8 mg, 0.33 mmol, 1 eq.) to the benzene solution and stirring for 30 min led to a precipitate of lithium bromide crown ether complex, which was removed by cannula-filtration. The filtrate was condensed to a volume of ca. 5 mL under reduced pressure. Slow solvent evaporation gave colorless crystals (also suitable for X-ray crystallography) of the composition 12 • 3 C6H6 (135 mg, 0.181 mmol, 55 %).
For the measurement of NMR and FT-IR spectra samples of compound 12 • 3 C6H6 were dried in vacuo, which led to the complete loss of the co-crystallized solvent C6H6.
In contrast to compounds 11 and 12 the NMR spectra were measured in C6D6, in which compound 13 shows limited solubility and prevented the record of meaningful 13 C{ 1 H} NMR spectra. More polar solvents (THF-d8 and CD2Cl2) lead to decomposition of 13.
Elemental analysis was performed with samples obtained from the crystallization procedure.

Synthesis of Compound 15.
Compound 10 (250 mg, 0.55 mmol, 1 eq.) was dissolved in 1,2-dimethoxyethane (10 mL). Catechol [(1,2-(HO)2C6H4, 73 mg, 0.66 mmol, 1.2 eq.] and the water trapping reagent MgSO4 (300 mg) were added. The suspension was vigorously stirred at ambient temperature for 6 h. The solvent was removed in vacuo and the solid extracted with CHCl3 (20 mL) and filtered over Celite. The solution was reduced in vacuo (to ca. 0.5 mL) and layered with n-pentane (20 mL). Slow solvent diffusion gave colorless crystals, which were also found suitable for X-ray crystallographic analysis. The colorless crystals were found to have the composition 15•CHCl3 (278 mg, 75 %). For NMR measurements and elemental analysis the crystalline samples were finely ground to powder with mortar and pestle and dried in high vacuum for 24 h to remove CHCl3 trapped in in the crystal lattice.

Synthesis of Thioborane 11 from Oxoborane 10
Compound 10 (250 mg, 0.55 mmol, 1 eq.) was dissolved in 1,2-dimethoxyethane (10 mL), and the water trapping reagent MgSO4 (300 mg) was added. Hydrogen sulfide (H2S, anhydrous) was bubbled through the suspension for 5 min. The reaction was stirred at ambient temperature for 6 h and filtered over Celite. All volatile material was removed in vacuo to afford a colorless solid 11 (218 mg, 85 %). The 1 H and 11 B{ 1 H} NMR spectra recorded in THF-d8 were identical to those of compound 11.

Detailed computational methods
All four compounds 10-13 (X = O, S, Se, Te) were optimized using several different combinations of DFT methods and basis sets, to assess the impact of the choice of DFT functional / basis set on the obtained results. The starting points for all calculations were the respective crystal structures. We used modern B97XD, [3] M06 [4] and CAM-B3LYP [5] functionals with two basis sets: a double-zeta 6-31G** basis set on all atoms and LANL2DZ ECP [6] on Te (if present) and a triple-zeta 6-311G++** basis set on all atoms and def2-TZVP ECP on Te (if present). The influence of the effective core potentials on the result for the Te system was investigated with additional calculations for this system using the same basis sets but with replacement of the LANL2DZ ECP by the WTBS basis set [7,8] obtained from the basis set exchange. [9] Since the results from all chosen combinations of DFT functional and basis sets were similar, we decided to base the discussion on the wB97XD/6-311G++** (with WTBS on Te atom if present) calculations. Selected results from all other methods are presented below. All calculations were performed using Gaussian 09 software [10]. We used ultrafine grid for DFT and standard optimization and convergence parameters as implemented in Gaussian 09. Bond indices, partial charges and orbital localization calculations using Foster-Boys approach [11] were performed using the Multiwfn Version 3.7 software [12]. Orbital visualization was performed using the VMD ver. 1.9.3 software [13].

Computational Results for Compounds 10−13.
The B=X computationally obtained bond distances calculated at different levels of theory are presented in Table  S1. A comparison with the structural metrics obtained from crystal structures suggests excellent accuracy with the bond lengths obtained from the medium-sized basis sets. It is worth noting, however, that the calculations were performed for the isolated molecules, while the experimental bond lengths are from crystal structures, where molecules interact with their neighbors and are densely packed, shortening their bond lengths.
The chemical character of the B=X bond was assessed by Meyer and Wiberg bond indices calculations, Tables S2 and S3. Both methods quite consistently assign the double bond to all studied systems in virtually all combinations of DFT methods and basis sets.
The natural population analysis (NPA) partial charges for B and X atoms are presented in Table 4. Similarly to the previously studied oxoborane case [14] the results for the B-O system suggest a double bond which is heavily polarized towards oxygen (NPA partial charge of -1.02 for O and 1.02 for B). The description is similar for all other B-X cases, but as shown previously the polarization become weaker as the terminal heteroatom becomes heavier, since for S the NPA partial charge is equal to -0.72, for Se it is equal to -0.58 to finally reach -0.31 for Te.
Tables S6 and S7 lists Meyer and Wiberg bond indices for the previously synthesized systems bearing a B=X bond, discussed in this work. The previously synthesized compounds are termed as stated: Table S1. Experimental and computed bond lengths in [Å] obtained compounds 5-8 with various density functionals. * The bond length was calculated as the average value resulting from three crystallographically independent molecules with the individual bond lengths as stated: B1−Te1 2.151(6) Å, B2−Te2 2.124 (7) Å, B3−Te3 2.143(6) Å. m stands for the double-zeta basis set (6-31G** basis set on all atoms and LANL2DZ ECP on Te), t for the triple-zeta basis set (6-311++G** basis set on all atoms and def2-TZVP ECP on Te) and w stands for the triple zeta basis set 6-311++G** for all atoms with the WTBS basis set for Te.
B97X-D/w B=O (10) B=S (11) B=Se (12) B=Te (13 A thorough investigation of the molecular orbitals was performed for compounds 10-13 to obtain a full description of the newly synthesized systems in terms of molecular orbitals theory. Selected orbitals for compound 10 are presented in Figure S39 and include the HOMO−4, HOMO−1, HOMO and LUMO orbitals. The π-bond to support the B=O double bond character can be attributed to the HOMO−4 orbital spanning over the large part of the B,N heterocyclic ring with a partial contribution of the HOMO−1 p-orbital, which is mostly localized on the O atom. Figure S39. Selected molecular orbitals of compound 10.
In the case of compound 11 the picture is quite similar, although the order of some crucial orbitals is slightly different, Figure S40. A π-orbital similar to that of HOMO−4 for compound 10 systems is in this case the HOMO−2 orbital, while the HOMO−1, HOMO and LUMO orbitals are very similar to those of compound 10. The π-bond between the B and S atoms can be attributed to the HOMO−2 orbital spanning over the large part of the B,Nheterocyclic ring with a partial contribution of the HOMO−1 p-orbital, which is highly localized on the S atom. Figure S40. Selected molecular orbitals of compound 11.
The analysis compound 12 reveals a similar pattern of orbitals close to the HOMO and LUMO as found for compound 11. The main difference is that the HOMO−2 orbital is now antibonding with respect to the B and Se atoms, and the bonding character can be attributed to the HOMO−1 and HOMO orbitals, which both are now slightly more localized on the Se atom, and slightly less on the B atom compared to compound 10. Figure S41. Selected molecular orbitals of compound 12.
The order of the orbitals for compound 13 is similar except for the fact that HOMO−2 does not span over the Te atom, most likely due to a relatively long distance from the diazaborole ring to the Te atom. On the other hand the HOMO−1, HOMO and LUMO are again very similar and the B-Te double bond may be attributed to the HOMO−1 and HOMO orbitals, which are highly localized on the Te atom. Figure S42. Selected molecular orbitals of compound 13. -