A new heuristic algorithm for the analysis of NMRD dispersion curves

Regression analysis of the NMRD dispersion curves obtained by the FFC-NMR relaxometric technique involves several conceptual and practical issues, which must be carefully addressed in order to gain reliable information on the system studied. Indeed, particular caution is needed when the FFC technique is applied to the investigation of complex systems such as polymeric matrixes, porous materials, food samples, and so on. In these cases, and in general whenever a rigorous approach based on a suitable physical modelization of the system is not available, data analysis must be performed by means of a “model–free” approach. Moving forward from previous literature, in the present communication we propose a new analysis method relying on the elaboration of the inverse integral transform of the NMRD curve, with the well-known BPP function as the kernel. Our approach results in a true heuristic method, able to unambiguously individuate the dynamic domains present in the system, and to avoid the possible introduction of any element of discretion. Moreover, the analysis is supplemented by a Monte-Carlo approach, in order to calculate the indetermination of the regression parameters and to individuate possible artifacts. Analysis of some datasets relevant to real samples, suggests the possibility that the results obtained with the heuristic method may be actually led back to some distinct physical/chemical features of the systems.