Assessing numerical strategies for efficient analysis of Nuclear Magnetic Relaxation Dispersion profiles

Fitting NMRD profiles to determine parameters of atomic/molecular motion in complex systems is a challenging task. Indeed, theoretical models in this field generally involve a large number of parameters, which raises questions about the uniqueness of the proposed solutions. This study aims at proposing a robust strategy to find reliable sets of parameters and provide the most realistic description of proton (1H) dynamics at microscopic scale, in aqueous suspensions of natural smectite clays. To achieve such a goal, a combination of Jean-Pierre Korb (JPK) and Solomon- Bloembergen-Morgan (SBM) model are used to reproduce the 1H NMRD profiles. The resulting large number of parameters impose to implement robust strategies for fitting. First, we use the geometric characteristics of each material to estimate the ratios of 1H, relaxing according to JPK or SBM mechanism in our system. This reduces the number of parameters to be optimized during the fitting procedure to 5. The optimization of these 5 parameters is then carried out using two main approaches. During a preliminary study, a Larmor frequency limit 𝜔𝐼,𝑙𝑖𝑚 is determined, which represents the boundary between the areas of predominance of JPK and SBM contributions. Then in a two-steps process, the JPK model is first fitted to data below 𝜔𝐼,𝑙𝑖𝑚 , and the resulting optimized parameters are injected into the global model to fit the remaining data with the SBM contribution. The two approaches as well as the relevance of the optimized parameters are discussed in a dedicated section.