Evaluation of Dynamic Vapour Sorption Isotherms Characteristics of Small Cardamom (Elettaria cardamomum Maton) using Empirical and Neural Network Models

Moisture sorption isotherm (MSI) characteristics of small cardamom (Elettaria cardamomum Maton) capsules were determined over water activity ( a w ) ranging from 10 to 90% at 20, 30, and 40°C using dynamic vapour sorption method. Equilibrium moisture contents (EMC) were correlated by three empirical models and multilayer neural network approach. It was observed that the Caurie model fitted best among empirical models for predicting EMC values. However, artificial neural network (ANN) model having 15 hidden neurons optimized using Lavenberg-Marquardth training algorithm was superior to the empirical models with the highest coefficient of determination ( R 2 ) value of 0.995 and 0.993 for adsorption and desorption respectively. Monolayer moisture content ( M m ) determined by the Caurie model was found to be 2.167, 1.901, AND 1.828% dry basis ( db) for adsorption, and 3.133, 3.061, and 2.744% ( db ) for desorption at 20, 30 and 40°C, respectively. Net isosteric heat of sorption ( q st ), sorption entropy and Gibb’s free energy (∆G) change were found to be inversely related to the EMC values. Glass transition temperature ( T g ) was found to be 98.91°C. The critical moisture content (CMC) and critical water activity of cardamom from T g - a w -EMC relationship was observed to be 1.989% ( db ) and 0.766, respectively at 30°C.