Targeted Frequency Domain Correction of Lineshape Distortions Introduced by Magnetic Resonance Inhomogeneity in Quantitative NMR Spectroscopy

Abstract

A review has been performed of the NMR software landscape that determines ideal qNMR software to be local, frequency-domain based and as automated as feasible. Within this landscape, reference deconvolution corrects lineshape distortions using a time-domain local/global technique and thus is a technique that has been singled out for improvement. A novel algorithm was created based on theoretical knowledge of magnetic resonance inhomogeneity as a summation of scaled, shifted, central peaks. This algorithm produced sufficient accuracy when applied to simulated (0.067-0.232%), and experimentally obtained data (0.025-18.04%), is robust to reasonable changes in initial peak parameters, and outperforms reference deconvolution by 0.85-25.7%. Overall, this algorithm is an effective solution for the correction of lineshape distortions produced by magnetic resonance inhomogeneity that operates entirely in the frequency domain and implements a global consensus fit, allowing for distinct improvement over reference deconvolution.