Theoretical Prediction of Nuclear Quadrupole Resonance Spectra of Aluminum, Bromine and Nitrogen Compounds via First Principles Calculations.

Chris Robert Harwell, Louisiana State University and Agricultural & Mechanical College

Abstract

We study the interaction of the electric field gradient (EFG) and the nuclear quadrupole moment of 81Br, 21Al and 14N nuclei via ab initio quantum chemistry calculations. The primary goal is to predict the nuclear magnetic resonance (NMR) spectral parameters of interesting materials and assist in interpretation of their spectra. The calculations predict NMR spectral parameters for: (1) 27Al nuclei in andalusite, sillimanite and kyanite---three polymorphs of Al2SiO5---for assigning NMR signals to crystallographic sites, as a zeolite model and studying the effect of structural changes due to temperature (25--1000°C); (2) 14N nuclei in tetryl and tetryl mimics to predict transition frequencies for the detection of explosives; (3) 81Br nuclei in brominated aromatic flame retardants and models to assist in spectra interpretation for judging dispersal in high impact polystyrene; (4) 27Al nuclei in cyclic (N-Al)x molecules for their similarity to the industrially important catalyst methyl-aluminum-oxane. The calculations use standard quantum chemistry computer programs. We examine single molecules with molecular orbital (MO) theory and full crystals with full potential linearized augmented plane wave (FP-LAPW) density functional theory. We also explore parts of full crystals by using MO theory for a small atom cluster and point charges for surrounding atoms. The MO calculations employ the restricted Hartree-Fock and the Becke's 3-parameter Lee-Yang-Parr hybridized density functional theory methods. We compare calculated spectral parameters among the different methods and with literature values acquired from experiment. The FP-LAPW method best predicts the spectral parameters and magnitudes, though it does rely on good quality crystal data being available. For MO theory a fairly large basis set of at least triple zeta quality with additional polarization and tight functions is necessary for accurate spectral parameters, and this method works well for single-molecules for which crystal data may not be available.