A molecular dynamics (MD) modeling and simulations approach has been rationally built and developed to study porous food systems constructed with amylose and dextran chains. The findings from our MD studies indicate that the presence of food macromolecules decreases the energetics of the water–water interactions for the nearby water molecules in the pore space, but provides additional water–macromolecule interactions that can significantly outweigh the partial loss of water–water interactions to make the adjacent water molecules strongly bound to the food macromolecules so that the water activity and water removal rate are decreased as dehydration proceeds and, thus, the dehydration energy requirement would be increased. The effects of pore structures are greater in systems with higher densities of food macromolecules, smaller in size pores, and stronger water–macromolecule interactions. Dehydration of food materials can thus be reasonably expected to start from the largest pores and from the middle of the pores, and to have non-uniform water removal rates and non-planar water–vapor interfaces inside individual pores as well as across sections of the food materials. The food porous structures are found to have good pore connectivity for water molecules. As dehydration proceeds, water content and the support from water–water and water–macromolecule interactions both decrease, causing the food porous structures to adopt more compact conformations and their main body to decrease in size. Dehydration in general also reduces pore sizes and the number of pore openings, increases the water–macromolecule interactions, and leads to the reduction of the overall thermal conductivity of the system, so that more energy (heat), longer times, and/or greater temperature gradients are needed in order to further dehydrate the porous materials. Our thermodynamic analysis also shows that the average minimum entropy requirement for food dehydration is greater when the water–macromolecule interactions are stronger and the food macromolecular density is higher. The importance of the physicochemical affinity of food molecules for water and of the compatibility of the resultant porous structures with water configurational structures in determining food properties and food processing through the water–macromolecule interactions, is clearly and fundamentally verified by the results and discussion presented in this work.

A thermal conductivity prediction regression equation was developed for four food materials as a single function of moisture content, porosity and initial thermal conductivity. It was assumed that temperature has a negligible effect on thermal conductivity. Thermal conductivity values of beef, apple, potato, pear and squid were used to develop the regression equation considering 122 points from the literature. The regression coefficient was 0.992. Squid gave the maximum mean per cent deviation of 11.89% (standard deviation 14-39). Potato gave minimum mean per cent deviation of 4.10% (standard deviation 1.74).