Robust Denoising of Laplace NMR and Multidimensional NMR Spectroscopy Using Chemometrics

Nuclear magnetic resonance (NMR) spectroscopy constitutes one of the robust analytical tools for revealing molecular-level information regarding the composition and concentration, as well as structure and dynamics. As a complement to Fourier transform-based NMR, Laplace NMR allows one to reveal molecular motions and spin interactions by studying molecular relaxation and diffusion phenomena. However, severe noises and unrelated-modulation interferences hinder accurate Laplace NMR analysis. Here, we describe a simple and general NMR approach based on principal component analysis to simultaneously suppress experimental noises and unrelated-modulation interferences for liquid NMR spectroscopy. This protocol is readily implemented within milliseconds on a standard desktop computer without prior knowledge and complicated setup and enables the reliable SNR enhancement up to 1 order of magnitude in Laplace NMR, thus significantly shortening the experimental times. Besides facilitating the detections and quantifications of minute signals, this protocol also contributes to the more accurate Laplace NMR analysis, benefiting from the removal of unrelated-modulation interferences. Experimental results on various Laplace NMR platforms, including relaxometry/diffusometry-edited experiments, pure shift-based Laplace NMR, and high-dimensional Laplace data sets, have demonstrated their robustness. Apart from Laplace NMR, this protocol is also applicable to the denoising of conventional multidimensional NMR spectroscopy. As a result, the study as outlined here describes a fundamental concept which affords ease of use and a guaranteed solution to NMR spectroscopy and thus is expected to find numerous applications in chemical and biomedical fields.