Reaction-diffusion approach to modeling water diffusion in glutinous rice flour particles during dynamic vapor adsorption.

Abstract

Dynamic vapor sorption (DVS) method is now widely adopted to determine water diffusion properties in food materials using a sheet or bulk particles as test samples. The Fickian second law with an instant equilibrium boundary condition, although commonly used, can not accurately model the water adsorption kinetics during DVS tests. Dynamic water adsorption of glutinous rice flour was measured at 30°C and eleven relative humidity steps, and modeled using the Fickian second law with three kinds of boundary condition. Results indicated that the boundary conditions had great impacts on the predicted values especially for the initial section in the DVS curves, and that modifying boundary conditions could not improve the fitness of the final section which characterized a continuing slight increase of water concentration. A reaction-diffusion model, which assumes two diffusible water populations and describes water transport as a competition between diffusion and reversible adsorption on solid matrix, was developed and found to be able to capture the features of water diffusion in the whole adsorption duration. Implementation of the reaction-diffusion approach to glutinous rice flour indicated that diffusion of the Langmuir water became very slow when its adsorption reached equilibrium, while the diffusion of the non-Langmuir water slowed down when water clustering occurred, at the same time the rates of the surface adsorption and bulk adsorption began to decrease. The model developed in this work would help to deepen our mechanistic understanding of water diffusion during a isothermal adsorption.