Effect of non-exponential and multi-exponential decay behavior on the performance of the direct exponential curve resolution algorithm (DECRA) in NMR investigations

The effect of non-exponential and multi-exponential decay or relaxation behavior on the performance of the direct exponential curve resolution algorithm (DECRA) is investigated through a series of numerical simulations. Three different combinations of decay or relaxation behavior were investigated through DECRA analysis of simulated pulse gradient spin echo (PGSE) NMR diffusion spectra that contained the combination of two individual components. The diffusion decay behavior of one component was described by a single-exponential decay, while the second component was described by either (1) a multi-exponential decay, (2) a decay behavior described by the empirical Kohlrausch-Williams-Watts (KWW) relation or (3) a multi-exponential decay behavior correlated with variations in the NMR spectral line shape. The magnitudes and types of errors produced during the DECRA analysis of spectral data with deviations from a pure single-exponential decay behavior are presented. It is demonstrated that the deviation from single-exponential decay impacts the resulting calculated line shapes, the calculated relative concentrations and the quantitative estimation of the decay or relaxation time constants of both components present in the NMR spectra. Copyright (C) 2004 John Wiley Sons, Ltd.