Stereochemistry Controls Dihydrogen Bonding Strengths in Chiral Amine Boranes Adducts

Abstract The growing interest in exploiting novel concepts of non‐covalent interactions in catalysts and supramolecular chemistry made us revisit a special kind of hydrogen bonding: the dihydrogen bond (DHB), formed between a classical hydrogen bond donor and a hydridic hydrogen as acceptor. Herein, we investigate how the strength of the N−Hδ+⋅⋅⋅δ−H−B interaction and hence the DHB‐driven self‐aggregation of amine‐borane adducts is governed by steric effects by comparing the structures and binding enthalpies of various chiral derivatives. For a diastereomeric pair of amine‐boranes prepared from a chiral secondary amine, we show that the stereochemistry at the nitrogen has significant influence on the interaction enthalpy. Based on this finding, N‐chiral amine boranes can be envisioned to become interesting building blocks in supramolecular chemistry to fine‐tune the formation dynamics of assemblies.


Preparation and enantioseparation of DMBA-BH 3
Synthetic procedure Chiral amines and BH 3 -DMS were acquired from Merck an used without further purification.
In a round-bottom flask enantiopure DMBA (93.3 mg, 0.10 mL, 0.69 mmol, 1.0 eq.) was dissolved in n-pentane and BH 3 -DMS (57.7 mg, 0.07 mL, 0.76 mmol, 1.1 eq.) was added dropwise. The mixture was stirred at room temperature for 1 h during which a colourless solid precipitated. The solvent was removed under reduced pressure at room temperature and the resulting colourless solid was purified by flash-chromatography on silica (eluent: dichloromethane). The product was obtained as a diastereomeric mixture with a yield of 99 % (101.4 mg, 0.68 mmol).

High Performance Liquid Chromatography
The diastereomeric mixtures of (NR)-/(NS)-DMBA-BH 3 were separated on a Shimadzu HPLC system equipped with autosampler and fraction collector using a YMC CHIRAL ART Cellulose SC S-5 µm column at a flowrate of 1 mL•min -1 . A mixture of cyclohexane and 2-propanol was used as eluent (ratio = 90:10) and as solvent to prepare sample solutions of 20 mg•mL -1 . Injection volumes were kept below 100 µL to prevent excessive peak broadening. Retention times were 5.99 min and 7.85 min respectively. The ratio of peak areas was determined as 77 % to 23 %. For the minor species, decomposition was observed when heated above room temperature.

Vibrational spectroscopy
Temperature dependent IR measurements were carried out using a JANIS ST-100 cryostat, in which samples were held in a vacuum-tight, sealed BaF 2 cell with 100 µm path length. Cooling was achieved by a constant flow of liquid nitrogen, while temperature regulation was achieved with a Lakeshore temperature controller. After a change in measurement temperature, sample cell and cryostat were allowed to reach thermal equilibrium. IR spectra were recorded with 4 cm -1 spectral resolution by accumulating for 32 scans and repeated 10 times in time intervals of 1 min.
VCD spectra were recorded on a Bruker Vertex FT-IR spectrometer equipped with a PMA 50 module for VCD measurements. The sample was held in a transmission cell with BaF 2 windows and 100 µm path length. Concentration are given in the main text. Spectra were recorded at room temperature with 4 cm -1 spectral resolution by accumulating 32 scans for the IR and ~16000 scans (4 hours accumulation time) for VCD. Baseline correction of the VCD spectra was done by subtraction of the spectra of the solvent or the racemic mixture recorded under identical conditions.

Thermochemistry
Fig S1. Concentration dependence of the IR spectra of the two epimers of DMBA-BH 3 in the N-H stretching mode region.  For the determination of the equilibrium constant K needed for the van't Hoff plot, the mole fractions of monomer and dimer at a given temperature of the VT-IR series were determined based on the integral intensities of the respective monomer bands in conjunction with the concentration dependent IR (VC-IR) measurements. To this end, the monomer bands of both VT-IR anc VC-IR series were integrated by fitting Lorentzian band shapes to the experimental data. To account for an increase in integral intensity due to increased density upon cooling of the sample, integral monomer intensities were corrected using the integral solvent intensities at the same temperature. The integral intensities (in km mol -1 ) of a VC-IR series at 293 K were then plotted against the concentration (in mol L -1 ). The intensity of the monomer band at infinite dilution (i.e. no dimer formation) can be approximated from this plot as the intercept b of a linear regression (y = mx + b) and the fraction of monomer x M,293K at a given concentration c1 and 293 K determined as: , , The solvent signal corrected integral monomer intensity at infinite dilution I M,inf of the VT-IR measurements is then determined as the fraction of the solvent corrected integral monomer intensity I M and x M,293K according to: , , The mole fraction of the monomer x M at a given temperature is calculated as the fraction of the integral monomer intensity I M and the monomer intensity at infinite dilution I M,inf : , The dimer mole fraction x D is subsequently calculated from x M as:

2
The equilibrium constant K at a given temperature was then calculated from the monomer and dimer molefractions x M and x D according to: The slope m of a linear regression of the corresponding van't Hoff plot (ln(K) against T -1 ) gives the dimerization enthalpy H as:

Crystallography
The crystal structure was analysed on a Rigaku Synergy dual source device, with Cu micro focus sealed tube (Cu Kα) using mirror monochromators and a HyPix-6000HE: Hybrid photon counting X-ray detector. The crystal was mounted in a Hampton CryoLoops using GE/Bayer silicone grease. Data was recorded and reduced using the CrysalisPro Software. 2 The structure was solved using WinGX 3 in combination with ShelXT 4 and refined with shelXle 5 and ShelXL.