Nuclear magnetic resonance (NMR) spectroscopy is an analytical technique that is used to probe the structure and dynamics of molecules. The relationships between measurable NMR properties and three-dimensional (3D) structure of small molecules were investigated using solution-state NMR spectroscopy with a focus on long-range proton-carbon scalar couplings (n>1JCH). The generation and validation of empirical equations to predict three-bond proton-carbon scalar couplings (3JCH) from 3D molecular structure were targeted.
The accuracy of selected 2-dimensional methods for measuring n>1JCH in model compounds (strychnine and camphor) was assessed by comparison with 55 n>1JCH measured from coupled 13C spectra. The IPAP-based HSQMBC/HMBC methods offered a balance of experiment time, n>1JCH accuracy and number of measurable n>1JCH. However, to maximise the number of n>1JCH measured a combination of multiple experimental methods (coupled 13C, IPAP-HSQMBC and EXSIDE) was required. This led to an average 40% increase in the number of measured couplings compared to using any single technique.
A fragment-based approach to empirical 3JCH prediction was developed by using density functional theory (DFT) to calculate 3JCH as a function of dihedral angle(s) for over 500 molecules. The fragments consisted of twelve different coupling pathways including saturated/unsaturated centres and β-heteroatoms and a variety of substituent patterns with a focus on the effect of methyl groups.
The fragment-based approach and selected literature empirical equations were validated against experimentally measured and DFT-calculated 3JCH. This demonstrated that the fragment-based approach could be further improved by the inclusion of a general bond angle correction. Currently, DFT offers the most accurate 3JCH calculation with only basis set limitations on the breadth of 3JCH that can be calculated. However, the empirical methods offered >2800-fold time saving in the calculation of 3JCH compared to DFT when applied to strychnine, camphor and 2-ethyl-1-indanone.
|Date of Award||6 Nov 2018|
- The University of Bristol
|Supervisor||Craig P Butts (Supervisor)|