Acquisition strategy to obtain quantitative diffusion NMR data

International audiencePulsed Gradient Spin Echo (PGSE) diffusion NMR experiments constitute a powerful tool for analyzing complex mixtures because they can in principle separate the NMR spectra of each mixture component. However, because these experiments intrinsically rely on spin echoes, they are traditionally regarded as non-quantitative, due to the signal attenuation caused by longitudinal (T-1) and transverse (T-2) nuclear magnetic relaxation during the rather long delays of the pulse sequence. Alternatively to the quantitative Direct Exponential Curve Resolution Algorithm (qDECRA) approach proposed by Antalek (J. Am. Chem. Soc. 128 (2006) 8402-8403), this work presents an acquisition strategy that renormalizes this relaxation attenuation using estimates of the T-1 and T-2 relaxation times for all the nuclei in the mixture, as obtained directly with the pulse sequence used to record the PGSE experiment. More specifically, it is shown that only three distinct PGSE experiments need to be recorded, each with a specific set of acquisition parameters. For small- and medium-sized molecules, only T-1 is required for obtaining accurate quantification. For larger molecular weight species, which typically exhibit short T-2 values, estimates of T-2 must also be included but only a rough estimation is required. This appears fortunate because these data are especially hard to obtain with good accuracy when analyzing homonuclear scalar-coupled systems. Overall, the proposed methodology is shown to yield a quantification accuracy of +/- 5%, both in the absence and in the presence of spectral overlap, giving rise - at least, in our hands - to results that superseded those achieved by qDECRA, while requiring substantially less experimental time. (C) 2012 Elsevier Inc. All rights reserved