A new theoretical model for moisture sorption isotherms and its application in deriving a hygroscopicity index for food products

Abstract In this paper, a new two-parameter moisture sorption isotherm (MSI) model was developed based on the non-random two-liquid (NRTL) theory. Compared to the most accepted theory of MSI, i.e., the layered adsorption theory, which includes only the effect of water clustering near food surfaces, the NRTL theory considers the effects of both water clustering and food swelling in the description of water adsorption on foods. This new model, referred to as the NRTL-solid-liquid-mixture (NRTL-SLM) model, revealed a rough linearity relationship between the reciprocal of the natural logarithm of water activity (1/lnaw) and the equilibrium moisture content (m1). A new index, characterising the magnitude of the slope of the plot of 1/lnaw v. s. m1, was constructed for the hygroscopicity assessment of food products. This new index, referred to as the NHI index, was scaled between 0% and 100%, with the bigger value indicating higher hygroscopicity and 100% representing the hygroscopicity of glycerol. In order to evaluate the applicability and reliability of the NRTL-SLM model and NHI index, seven food samples, i.e., grapes, apples, pears, bananas, avocados, potatoes and beef, were selected. Glycerol was chosen as the hygroscopic reference material. The MSI data at 25 °C of all the eight materials, i.e., seven food products and glycerol, were subjected to model evaluation and index calculations. Our results showed that: (a) the NRTL-SLM model could satisfactorily fit the MSI curves of all the eight materials; (b) compared to the existing most successful MSI model, i.e., the Guggenheim-Anderson-deBoer (GAB) model, the NRTL-SLM model provides a slightly lower accuracy but is much simpler and requires fewer parameters; (c) the NHI index could correctly rank the hygroscopicity of the eight materials relative to each other; and (d) the existing most widely used and standardised physical parameter for MSI analyses, i.e., the monolayer moisture content (M0), which is typically calculated from the GAB model, could not be used as a measure of hygroscopicity for foods.