"Advancing NMR crystallography with chemical shift tensor predictions using modern electronic structure models"

Dr. Robbie J. Iuliucci, Professor of Chemistry

Washington & Jefferson College, Department of Chemistry

Host: Aaron Rossini, Karl Mueller, Frederic Perras 

Abstract:

NMR crystallography (NMRX) represents a powerful set of experimental and computational tools used to elucidate crystal structures. Typically, NMRX integrates solid-state NMR (SSNMR), X-ray diffraction (XRD), and computational methods to achieve a more detailed understanding of molecular-level structure than is possible using any of these techniques in isolation. This emerging approach is particularly valuable in the pharmaceutical industry, where crystal structure engineering can enhance active pharmaceutical ingredients or, when combined with crystal structure prediction, contribute to de novo drug development. Notably, structures of amorphous solids can be determined using only SSNMR and computational methods—an achievement largely inaccessible to diffraction-based techniques.

The chemical shift, an anisotropic property exceptionally sensitive to electronic structure, plays a central role in NMRX. However, its effectiveness in structure elucidation depends critically on accurate electronic structure modeling to interpret its rich spectral information. A key experimental advantage is that the principal components of the chemical shift tensor can be readily extracted from simple powdered solids. Computational models must not only accurately represent the local electronic environment of the nucleus but also account for the extended lattice of the microcrystalline solid.

To this end, the solid-state physics community has exploited the efficiency of plane-wave density functional theory (DFT), which, by satisfying the periodic boundary conditions of a crystal, captures the infinite lattice using only the atoms of the asymmetric unit cell. However, plane-wave DFT methods are generally limited to the generalized gradient approximation (GGA) level, which is insufficient for accurately modeling the local electronic structure around nuclei. An alternative is to mimic the infinite lattice using a finite molecular cluster, thereby enabling the use of hybrid DFT methods. To go beyond hybrid DFT, monomer molecules can be calculated using the highest feasible quantum chemical methods and used as corrections to improve the local electronic description. This multilayer approach now represents the state-of-the-art, enabling ¹³C chemical shift predictions in solids with accuracies approaching one ppm. This methodology will be illustrated by revisiting the crystal structure of naphthalene—one of the pioneering studies in NMRX more than 30 years ago. 

The recruitment and training of talented young scientists is critical for the United States to maintain its technological edge. This talk will also highlight the contributions of undergraduate researchers at a small, modest, private liberal arts college. One recent success includes the discovery of a previously unknown crystalline form of the simplest dipeptide, glycylglycine, whose hydrochloride salt forms a short, strong hydrogen bond of 1.23 Å. A second example involves a polymorph screening of the cancer drug erlotinib, which serendipitously yielded a zinc chloride metal complex with potential for enhanced bioavailability.