A Highly Active Bidentate Magnesium Catalyst for Amine‐Borane Dehydrocoupling: Kinetic and Mechanistic Studies

Abstract A magnesium complex (1) featuring a bidentate aminopyridinato ligand is a remarkably selective catalyst for the dehydrocoupling of amine‐boranes. This reaction proceeds to completion with low catalyst loadings (1 mol %) under mild conditions (60 °C), exceeding previously reported s‐block systems in terms of selectivity, rate, and turnover number (TON). Mechanistic studies by in situ NMR analysis reveals the reaction to be first order in both catalyst and substrate. A reaction mechanism is proposed to account for these findings, with the high TON of the catalyst attributed to the bidentate nature of the ligand, which allows for reversible deprotonation of the substrate and regeneration of 1 as a stable resting state.


S1 -General Procedures
All manipulations involving magnesium complexes 1 and 2 were conducted under anhydrous, anaerobic conditions using standard Schlenk line and glove box techniques. Standard laboratory solvents were dried by distilling from potassium (toluene) or sodiumbenzophenone ketyl (THF) and stored over a potassium mirror (toluene) or 4 Å molecular sieves (THF). d6-Benzene was dried over potassium in a sealed ampoule at 80 °C for 4 days, before vacuum-transferring to a Young's flask containing a potassium mirror, which was subsequently stored in the glovebox prior to use. NMR samples of air and moisture sensitive compounds were prepared using glove box techniques and contained in Young´s tap modified borosilicate glass NMR tubes. TMEDA was distilled from CaH2 and stored over 4 Å molecular sieves. Me2NH·BH3, Me3N·BH3, and iPr2NH·BH3 were purchased from Sigma-Aldrich and used as received. Ligand precursors L 1 H and L 2 H were synthesised by minor modifications of previous reported synthetic procedures. [1,2] MeMgI·(OEt2)1.5 was synthesised from the reaction between activated magnesium turnings and iodomethane in diethyl ether. Purified compounds were stored under dried nitrogen in an MBraun UNIlab glovebox.

S3.1 -Crystallographic Methods
Under a flow of N2, crystals suitable for X-ray diffraction were quickly removed from the crystallisation vessel and covered with Fomblin® (YR-1800 perfluoropolyether oil). A suitable crystal was then mounted on a polymer-tipped MicroMount TM and cooled rapidly to 120 K in a stream of cold N2 using an Oxford Cryosystems open flow cryostat. [3] Single crystal X-ray diffraction data were collected on an Oxford Diffraction SuperNova Duo diffractometer (Atlas CCD area detector, mirror-monochromated Cu-Kα radiation source; λ = 1.54184 Å or mirrormonochromated Mo-Kα radiation source; λ = 0.71073 Å; ω scans). Absorption corrections were applied using an analytical numerical method (CrysAlis Pro). [4] All non-H atoms were located using direct methods [5] and difference Fourier syntheses. Hydrogen atoms were placed and refined using a geometric riding model. All fully occupied non-H atoms were refined with anisotropic displacement parameters, unless otherwise specified. Crystal structures were solved and refined using the Olex2 software package. [6,7] Programs used include CrysAlisPro [8] (control of Supernova, data integration and absorption correction), SHELXL [9] (structure refinement), SHELXS [5] (structure solution), SHELXT [10] (structure solution), OLEX2 [6] (molecular graphics). CIF files were checked using checkCIF [11] CCDC-1836622 and -1836623 contain the supplementary data for 1 and 2. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. Figure S1: Molecular structure of 1 (a) and 2 (b) with anisotropic displacement ellipsoids set at 50% probability. Hydrogen atoms, second molecule in asymmetric unit (1, 2), and co-crystallised diethyl ether (2) have been omitted for clarity. Table S1: Selected distances (Å) and angles (°) for 1 and 2. Measurements for second molecule in asymmetric unit in square brackets.

S4.1.1 -Synthesis of 1 (L 1 MgI·(tmeda))
Scheme S1: Synthesis of compound 1 To a suspension of MeMgI·(Et2O)1.5 (1.526 g, 5.5 mmol) in diethyl ether (50 mL) at −78 °C was added dropwise a solution of L 1 H (10 mL, 0.5 M in Et2O/hexanes, 5 mmol) over 45 min. The solution was stirred at −78 °C for a further 2 h, then allowed to warm to −30 °C. TMEDA (2.25 mL, 15 mmol) was added and the resultant gel warmed to room temperature overnight. The milky suspension was filtered, and volatiles removed from the filtrate in vacuo; washing of the residue with cold hexanes afforded compound 1 as a white solid (1.80 g, 63%). Crystals of 1 suitable for X-ray diffraction were grown from a saturated solution of the complex in diethyl ether at -30 °C.  To a suspension of MeMgI·(Et2O)1.5 (0.444 g, 1.6 mmol) in diethyl ether at −78 °C, a solution of L 2 H (0.444 g, 1.5 mmol) in Et2O (10 mL) was added dropwise over 1 h to afford a yellow suspension. The reaction was stirred for 4 h at −78 °C, and allowed to warm to −30 °C. TMEDA (0.9 mL, 6 mmol) was added, the resultant suspension stirred for a further 2 h, then allowed to warm to room temperature and volatiles removed in vacuo. Washing with cold hexanes afforded 2 as a green solid (0.771 g, 86%). Crystals of 2 of suitable quality for X-ray diffraction were grown from a saturated solution of the complex in diethyl ether at -30 °C.

-Synthesis of Me2ND·BH3
Synthesis adapted from a literature procedure. [12] Me2NH·BH3 (500 mg, 8.49 mmol) was dissolved in D2O (2 mL) and stirred for 2 h at room temperature. The solution was washed with DCM (3 x 10 mL) and the organic layer was separated and dried over Na2CO3. Removal of volatiles in vacuo afforded Me2ND·BD3 as a white crystalline solid (303.8 mg, 5.07 mmol, 60%). No NH signal could be detected by 1 H NMR spectroscopy.

S4.1.4 -Synthesis of Me2NH·BD3
We attempted to synthesise Me2NH·BD3 via a literature procedure, wherein Me2NH·HCl was reacted with NaBD4 and purified by aqueous work up. [13] Contrary to the published report, this does not afford pure Me2NH·BD3, but a mixture of Me2NH·BD3 and Me2NH·BH3, which is likely due to H/D exchange with the water used in the aqueous work-up. These compounds display two distinct signals for the Me groups in the 1 H NMR spectrum ( Figure S7), and were originally assigned as a doublet in the literature. [13] Furthermore, we observe two distinct signals in the 13 C NMR spectrum in C6D6 solution ( Figure S8). The product was also characterised by IR spectroscopy, which revealed clear B-H stretches. Characterisation data for the products of this reaction is given below.    An alternative synthetic route was attempted, in which the reaction was carried out and worked up under completely anhydrous conditions (see below). This gave a higher degree of deuteration, but the product still contained significant amounts (ca. 20%) of BH3 containing product ( Figure S11). This seems to be the result of deuterium exchange with Me2NH·HCl, as 1 H NMR and IR spectroscopy of the product indicated a loss of proton label at N and N-D stretching bands ( Figure S14).
Anhydrous Procedure: To a Schlenk flask charged with Me2NH·HCl (612 mg, 7.5 mmol), NaBD4 (330 mg, 7.9 mmol), and a stirrer bar; pre-cooled (0 °C) THF (5 mL) was added and the resultant slurry stirred at 0 °C for 1 h. The reaction was then allowed to warm to room temperature and stirred overnight. Volatiles were removed in vacuo and the product was isolated by vacuum sublimation as a white crystalline solid (228 mg). Characterisation data for this product is shown below.

S4.2.1 -Standard procedure for the catalytic dehydrocoupling of Me2NH·BH3
In a Young's NMR tube 1 mol% (0.0017 mmol), 5 mol% (0.0085 mmol) or 10 mol% (0.0170 mmol) of catalyst 1 or 2 and Me2NH·BH3 (10 mg, 0.170 mmol) were dissolved in 0.6 mL of the corresponding solvent (C6D6 or THF). The reaction was heated to 60 °C in an oil bath and progress was monitored by 1 H and/or 11 B NMR spectroscopy at predetermined time-points. In the case of THF, a capillary of C6D6 was also added to the NMR tube to provide a lock.

S4.2.2 -Qualitative in situ monitoring by 1 H NMR spectroscopy
In a Young's NMR tube 5 mol% (0.0085 mmol) of catalyst 1 and Me2NH·BH3 (10 mg, 0.170 mmol) were dissolved in 0.6 mL of C6D6. An initial 1 H NMR spectrum was recorded at 25 °C ( Figure S20a), revealing no reaction had occurred. The temperature in the spectrometer was increased to 60 °C, and 1 H NMR spectra were recorded at regular intervals ( Figure S20b). Once no starting material was observable by 1 H NMR spectroscopy, the temperature was reduced to 25 °C, and a final 1 H NMR spectrum was recorded ( Figure S20c).

S4.2.3 -Standard procedure for the catalytic dehydrocoupling of iPr2NH·BH3
In a Young's NMR tube 5 mol% (0.0088 mmol) of catalyst 1 and iPr2NH·BH3 (20 mg, 0.175 mmol) were dissolved in 0.6 mL of C6D6. The reaction was either left at room temperature or heated to 60 °C in an oil bath and progress was monitored by 1 H and 11 B NMR spectroscopy at predetermined time-points. The product (iPr2N=BH2) was identified by comparison of 1 H and 11 B NMR spectra with literature data. [13]      Recyclability experiments: The above procedure was carried out for 10 mol% of 1 (96.3 mg, 0.166 mmol). After completion, the connection to the measuring burette was shut, and the Schlenk tube kept open to argon at 60 °C overnight to ensure complete consumption of any remaining starting material. The argon was then disconnected, the pressure left to equilibrate, and a further equivalent of Me2NH·BH3 (98.2 mg, 1.66 mmol) in 2 mL of toluene injected. The procedure was repeated for a third time, with prior volume reduction in vacuo to ca. 8 mL.

S4.3.1 -Procedure for determining reaction orders
In a Young's NMR tube 0.6 mL of a stock solution (A, B, or C) containing Me2NH·BH3 and catalyst 1 in C6D6 (A: 0.29 M Me2NH·BH3, 28 mM 1; B: 0.29 M Me2NH·BH3, 22 mM 1; C: 0.29 M Me2NH·BH3, 15 mM 1) was prepared and frozen at −78 °C to prevent reaction initiation. The sample was transferred to a Bruker AV(III) 600 spectrometer, and a 11 B NMR spectra recorded at 15 °C, which showed no reaction. The temperature was then rapidly ramped to 60 °C. The temperature was monitored using a thermocouple located within the probe. Once the temperature stabilised sufficiently to allow locking to the C6D6, 11   All time values are given relative to an approximate start time, which is the time at which temperature ramping to 60 °C was begun. The linear correlation between υ and [Me2NH·BH3] at high substrate concentrations is indicative of a first order dependence on substrate. However, the curvature at low substrate concentrations, and the fact that fitting the linear portion of the graph gives a non-zero y-intercept, indicates that the reaction is pseudo-first order with a more complex rate dependence at low substrate concentrations.

S4.3.2 -Procedure for determining activation parameters
Two stock solutions were prepared in the glovebox; a 600 mM solution of Me2NH·BH3 in toluene, and a 9 mM solution of 1 in toluene. The two solutions were stored at −35 °C in the glovebox until required.
For each experiment, the stock solutions were warmed to room temperature and 250 μL of each solution was added to a Young's NMR tube, along with a sealed glass capillary of d6-DMSO (to provide a lock). This afforded a reaction solution with an initial concentration of 0.30 M Me2NH·BH3 and 4.5 mM 1 (1.5 mol% catalyst loading).
The sample was transferred to a Bruker AV(III) 600 spectrometer, preheated to the required reaction temperature (50 °C, 60 °C, 70 °C, or 80 °C), and the reaction temperature monitored by thermocouple.
For the reactions at 50 °C, 60 °C and 70 °C, the spectrometer was locked and shimmed to the d6-DMSO, while for the reaction at 80 °C spectra were recorded locked but without shimming. 11 Figure S30). data from < 3 minutes after sample was transferred to the NMR machine were omitted for the runs at 80 °C as it was non-linear, likely due to temperature equilibration. Errors for the activation parameters obtained from the Eyring plot were estimated at three times the standard error calculated from the linear fit of the data.  Errors in rate constants are estimated from the standard error in the linear fitting of the data. All times are given relative to an approximate start point, which is when the sample was inserted into the NMR spectrometer. Concentration of 1 is 4.5 mM in all reactions.

S4.3.3 -Kinetic Isotope Effect with Me2ND·BH3
In a Young's NMR tube 0. These rate constants are within error of each other (i.e. 3 × standard error as determined by a linear fit of the data). Therefore kH/kD ≈ 1

S4.3.3 -Kinetic Isotope Effect with Me2NH·BD3
In a Young's NMR tube 0.6 mL of a stock solution containing a mixture of Me2NH·BH3 and Me2NH·BH3 (approximately 125 mM, see section S4.1.4), catalyst 1 (8.8 mM) and trimethoxybenzene (36 mM, for use as an internal standard) in C6D6 was prepared and frozen at −78 °C to prevent reaction initiation. The sample was transferred to a Bruker AV(III) 600 spectrometer and an initial 1 H NMR spectra recorded at 15 °C ( Figure S33 and Figure S34). The sample was heated to 60 °C for 1 h, then cooled to 15 °C, whereupon a second 1 H NMR spectrum was recorded ( Figure S33 and Figure S34). The methyl groups of Me2NH·BH3 and Me2NH·BD3 were integrated relative to the methyl groups of the trimethoxybenzene internal standard. Gaussian fitting was employed to deconvolute the peaks arising from Me2NH·BH3 and Me2NH·BD3. The kinetic isotope effect was calculated as kH/kD = 1.6 ± 0.1 from the equation: [14] = ln

S4.4.1 -Reaction with Me2NH·BH3
In a Young's NMR tube, Me2NH·BH3 (2.0 mg, 0.034 mmol) and 1 (20.0 mg, 0.034 mmol) were dissolved in 0.8 mL of C6D6. The reaction was monitored by 1 H and 11 B NMR spectroscopy. The signals at δB 3.4 (t) and −14.6 (q) are proposed to correspond to species I 2 of the proposed catalytic cycle. Figure S35: 11 B NMR spectrum for the stoichiometric reaction of 1 with 1 eq. of Me2HN·BH3 in C6D6 after 44 h at room temperature. The signals at δB 3.4 (t) and −14.6 (q) are proposed to correspond to species I 2 of the proposed catalytic cycle.

S4.4.2 -Reaction with Me3N·BH3
In a Young's NMR tube, Me3N·BH3 (3.8 mg, 0.052 mmol) and 1 (15.0 mg, 0.026 mmol) were dissolved in 0.6 mL of C6D6. The reaction was heated to 80 °C and monitored by 1 H and 11 B NMR spectroscopies, whereupon no reaction was observed even after several days of heating.