Although considerable efforts have been made to improve the understanding of mass transfer in osmotic dehydration, fundamental knowledge about predicting mass transport is still a gray area (Raoult-Wack et al., 1991). Modeling of the osmotic dehydration process is necessary for optimizing the osmotic dehydration and subsequent drying processes, in order to achieve the highest possible quality at minimum energy costs (Saguy et al., 2005).
The unusual features come from the interaction between the solution and materials of biological origin. Mass transfer in osmotic dehydration of cellular plant foods, such as fruit and vegetables, involves several physical effects due to the complex morphology of plant tissues. These can be described, in an ideal way, as osmosis, diffusion and hydrodynamic mechanism (HDM) penetration (Fito and Pastor, 1994). Two basic approaches can be used to model osmotic processes (Ramaswamy, 1982; Salvatori, 1998).
The first one, the macroscopic approach, assumes that the tissue is homogeneous and the modeling is carried out on the cumulated properties of cell walls, cell membranes and cell vacuoles. The second one, the microscopic approach, identifies the heterogeneous properties of the tissue and is based on cell microstructure (Fito et al., 1996).
Macroscopic analysis has been carried out on pseudo-diffusion, square root of time, irreversible thermodynamic and other approaches (Fito et al., 1996) Very little work has been developed from the microscopic point of view (LeMaguer, 1996). The analysis of the concentration profiles developed throughout mass transfer processes, using a macroscopic approach, can be useful to clarify the mass transfer mechanisms and their coupling.
Especially if data are correlated with micro-structural features (shape, size and geometry changes in cell and intercellular spaces, cell wall deformation and relaxation changes, etc.) observed by a microscopic technique (Alzamora et al., 1996). However, concentration profiles allow us to calculate mass transfer kinetics (Lenart and Flink, 1984b). Mathematical modeling may provide a useful insight into the underlying mechanisms and several mathematical models have been proposed based on a cellular