Model-based Experimental Analysis of Multicomponent Diffusion in Liquids

Abstract:
Today, diffusion coefficients in liquids still cannot be predicted with sufficient accuracy due to a lack of measurement data. Therefore, a new diffusion measurement technique using Raman spectroscopy is proposed here to reduce experimental effort and to make diffusion measurements more practically tractable. For this purpose, model-based techniques are used in this work to find the optimal experimental design and for efficient data analysis. Thereby, it is shown that a single run of the new model-based Raman diffusion experiment is sufficient to determine the concentration dependence of binary diffusion coefficients or the full ternary Fick diffusion matrix. In addition, the model-based analysis of established diffusion experiments based on Taylor dispersion and interferometry is explored indicating significant potential for improvement. Overall, a work process of model-based experimental analysis could be devised providing a systematic and efficient approach for the study of kinetic phenomena in general.

Keywords:
Diffusion, Multicomponent diffusion, Fick, Maxwell-Stefan, Inverse Problem, Optimal Experimental Design, Parameter Estimation, Modeling, Raman Spectroscopy, Mass Transfer