Seven‐Membered Cyclic Potassium Diamidoalumanyls

Abstract The seven‐membered cyclic potassium alumanyl species, [{SiNMes}AlK]2 [{SiNMes}={CH2SiMe2N(Mes)}2; Mes=2,4,6‐Me3C6H2], which adopts a dimeric structure supported by flanking K‐aryl interactions, has been isolated either by direct reduction of the iodide precursor, [{SiNMes}AlI], or in a stepwise manner via the intermediate dialumane, [{SiNMes}Al]2. Although the intermediate dialumane has not been observed by reduction of a Dipp‐substituted analogue (Dipp=2,6‐i‐Pr2C6H3), partial oxidation of the potassium alumanyl species, [{SiNDipp}AlK]2, where {SiNDipp}={CH2SiMe2N(Dipp)}2, provided the extremely encumbered dialumane [{SiNDipp}Al]2. [{SiNDipp}AlK]2 reacts with toluene by reductive activation of a methyl C(sp 3)‐H bond to provide the benzyl hydridoaluminate, [{SiNDipp}AlH(CH2Ph)]K, and as a nucleophile with BPh3 and RN=C=NR (R=i‐Pr, Cy) to yield the respective Al‐B‐ and Al‐C‐bonded potassium aluminaborate and alumina‐amidinate products. The dimeric structure of [{SiNDipp}AlK]2 can be disrupted by partial or complete sequestration of potassium. Equimolar reactions with 18‐crown‐6 result in the corresponding monomeric potassium alumanyl, [{SiNDipp}Al−K(18‐cr‐6)], which provides a rare example of a direct Al−K contact. In contrast, complete encapsulation of the potassium cation of [{SiNDipp}AlK]2, either by an excess of 18‐cr‐6 or 2,2,2‐cryptand, allows the respective isolation of bright orange charge‐separated species comprising the ‘free’ [{SiNDipp}Al]− alumanyl anion. Density functional theory (DFT) calculations performed on this moiety indicate HOMO‐LUMO energy gaps in the of order 200–250 kJ mol−1.


General Experimental Details
All manipulations were carried out using standard Schlenk line and glovebox techniques under an inert atmosphere of argon. NMR spectra were recorded on an Agilent ProPulse spectrometer operating at 500 MHz ( 1 H), 126 MHz ( 13 C) and 160 MHz ( 11 B). The spectra were referenced relative to residual protio solvent resonances. Elemental analyses were performed at Elemental Microanalysis Ltd., Okehampton, Devon, UK. Solvents (toluene, hexane) were dried by passage through a commercially available solvent purification system, under argon and stored in ampoules over 4 Å molecular sieves. C6D6 was purchased from Sigma-Aldrich, dried over a potassium mirror before vacuum distilling and storing under argon over molecular sieves.
CDCl3 was purchased from Sigma-Aldrich, dried by stirring with CaH2 overnight before vacuum distilling and storing under argon over molecular sieves. MesNH2 and {CH2SiMe2Cl}2 were purchased from Sigma-Aldrich and distilled prior to use and compound 3 was prepared by a literature procedure. [1] All other reagents were purchased from Sigma-Aldrich and used without further purification.

Synthesis of {SiN Mes }H2 (8)
A solution of n-BuLi in hexane (42.5 mL of a 2.5M solution, 0.107 mol) was added dropwise to a pre-cooled solution of MesNH2 (14.4 g, 15 mL, 0.107 mol) in Et2O (60 mL) at 0 °C. The resulting colorless suspension was stirred at room temperature for 1.5 hours followed by the dropwise addition of a solution of {CH2SiMe2Cl}2 (11.5 g, 0.0535 mol) in Et2O (30 mL) at 0 °C. The resultant suspension was stirred for 12 hours at room temperature, allowed to settle for 3 hours and filtered to give a clear colorless solution. Removal of the volatiles in vacuo followed by extraction into hexane (100 mL) gave a colorless suspension. Filtration of the solution followed by removal of the volatiles in vacuo provided 8 as a colorless waxy solid.

Synthesis of [SiN Mes }AlMe] (9)
A solution of AlMe3 in hexane (7.4 mL of a 2 M solution, 15.6 mmol) was added dropwise to a stirring solution of 8 (6.0 g, 14.8 mmol) in toluene (30 mL) at 0 °C. Upon addition, the solution bubbled and was stirred for 48 hours at room temperature under a slow flow of argon, followed by warming to 60 °C and stirring for 12 hours. The resulting colorless suspension was allowed to cool to room temperature and the volatile components were removed in vacuo to give 9 as a colorless waxy solid. Yield: 6.45 g, 97%. 1  A solution of iodine (2.1 g, 8.27 mmol) in toluene (50 mL) was added dropwise to a stirring solution of 9 (3.5 g, 8.26 mmol) in toluene (40 mL) at 80 °C. The solution was refluxed for 4 days under a slow flow of argon, after which time the initial red color was lost and the solution became pale orange. The solution was allowed to cool to room temperature and the volatile components were removed in vacuo to give a waxy solid. Extraction into hexane and filtration gave a clear orange solution. Concentration of the orange solution followed by storage at −30 °C gave 10 as colorless crystals. Yield 2.85 g, 61%. 1  A solution of 10 (4.68 g, 7.24 mmol) in hexane (30 mL) was stirred on mirrored K (0.85 g, 21.7 mmol) for 6 days at room temperature resulting in the gradual color change from colorless to yellow and the formation of a grey precipitate. The resulting dark yellow solution was filtered through a cannula filter, concentrated to ca. 15 mL and stored at −18 °C for 24 hours resulting in precipitation of a small amount of colorless solid identified as 11 by 1 H NMR spectroscopy. The suspension was filtered and the volatiles were removed in vacuo to give 12 as a yellow powder. Although bulk samples of 12 were invariably contaminated with variable quantities of 11, X-ray quality crystals were grown from a concentrated Et2O solution at −30 °C.

Method 2:
A solution of 11 (0.95 g, 1.08 mmol) in hexane (50 mL) on mirrored K (0.085g, 2.17 mmol) for 5 days at room temperature resulting in the gradual color change from colorless to yellow.
Concentration of the solution and storage at −18 °C gave yellow crystals of 12 after 24 hours.
Although bulk samples 12 prepared by this method were invariably contaminated with variable quantities of 11, NMR data could be assigned with a reasonable level of confidence by discounting resonances assigned to compound 11. 1   (SiCH2), 1.9 (SiMe2), 1.8 (SiMe2). 13 C resonance correlated to benzylic moiety not observed.

X-ray Crystallography
Single Crystal X-ray diffraction data for compounds 10 -13, 17 -23 were collected on a SuperNova, EosS2 diffractometer using CuKα (λ = 1.54184 Å) radiation throughout. Data for compound 14 were collected on a New Xcalibur, EosS2 diffractometer using MoKα (λ = 0.71073 Å) radiation. The crystals were maintained at 150 K during data collection. Using Olex2, [2] the structures were solved with the olex2.solve [3] structure solution program or ShelXT and refined with the ShelXL [4] refinement package using Least Squares minimization.
The asymmetric unit in 12 contains half of a molecule plus a region of solvent. The aluminium centres are both coincident with a two-fold rotation axis implicit in the space group symmetry.
The potassium cation was modelled over two positions in a 55:45 ratio, while C23, C24, and C25 were each modelled to take account of 80:20 disorder. Distance and ADP restraints were used in disordered regions to assist convergence. The solvent was heavily disordered, which is unsurprising, as it located in structural channels which are parallel to the c-axis. Ultimately, this was treated using the solvent mask algorithm available in Olex-2 and an allowance of 12 molecules of hexane per unit cell has been made in the formula as presented.
The asymmetric unit in structure of 13 comprises half of a dimer molecule plus one benzene molecule. The remainder of the dimer can be generated by virtue of a twofold rotation axis implicit in the space group symmetry.
The motif in 15 contains one aluminium centre and represents part of a polymeric structure.
The hydrogens attached to C27 were located and refined at a fixed distance of 0.96Å from the parent atom.
The hydrogen atoms attached to C31, C32 and C36 in 17 were located and refined without restraints.
The asymmetric unit in 18 comprises half of a dimer molecule and 2 guest benzenes. The hydrogens attached to C32, C33 and C36 were readily located and each refined at a distance of 0.98Å from the relevant carbon atom.
C29 and C30 were treated for 50:50 disorder in the structure of 19, where the asymmetric unit comprises just half of a molecule. Distance and ADP restraints were used in the disordered region, to assist convergence.
The largest residual electron density peak (close to the order of one electron per Å −3 ) in the structure of 20 lies 1.56 Å from the aluminium centre, and 2.68 Å from K1. In the absence of any spectroscopic or chemical evidence for the presence of a minor product, this was treated as a spurious artefact.
In the structure of 21, all atoms in the cyclohexyl moiety based on C38 (with the exception of C38 itself) were modelled to take 55:45 disorder into account. Distance restraints were employed in the disordered components to assist convergence. Two molecules of benzene were also found to be present in the asymmetric unit.
The asymmetric unit in 22 contains one aluminium based anion and half of a dication. The latter contains one potassium centre and 1.5 crown-ethers. A crystallographic inversion centre proximate to the half-crown moiety serves to generate the remainder of the cation by bridging between the potassium centres. The full crown present in the asymmetric unit was disordered but, with the inclusion of distance and ADP restraints pertaining to fractional occupancy atoms, this resolved into 2 components with an occupancy ratio of 63:37. The residual maximum in the difference-Fourier, electron-density map, at 2.56Å from C32A, appears to be spurious and is most likely an artefact. -S29- The asymmetric unit in 23 contains one anion, one cryptand-based cation and one molecule of tetramethyl-THF. Four carbons in the solvent were treated for 50:50 disorder and, with the inclusion of some distance and ADP restraints, the model converged sensibly in this region.
The highest residual electron density peak lies approximately 1.38 Å from the aluminium centre. In the absence of any chemical evidence for a minor second species being present, this appears to be an artefact. When the gross structure is viewed along the c-axis, there are small channel-type pathways evident, which coincide with this residual electron density maximum. Figure S35. Displacement ellipsoid plot (50% probability) of compound 23. Hydrogen atoms and an occluded molecule of tetramethyl-THF are omitted for clarity.

Computational Details / Methodology
DFT calculations were run with Gaussian 09 (Revision D.01). [5] The K, Al, Si, S and I centres were described with the Stuttgart RECPs and associated basis sets, [6] and 6-31G** basis sets were used for all other atoms (BS1). [7] A polarization function was also added to Al (d = 0.190), Si (d = 0.284), S (d = 0.503) and l (d = 0.289). All energies were recomputed with cc-pVTZ-PP for I and 6-311++G** basis sets for other atoms (BS2). Initial BP86 [8] optimizations were performed using the 'grid = ultrafine' option, with all stationary points being fully characterized via analytical frequency calculations as minima (all positive eigenvalues). Kohn-Sham frontier orbital analysis was performed on the BP86-optimized geometries of the Al species. -S35-