RESUMEN
Moisture adsorption isotherms of beta-glucan rich composite flour biscuits were determined at 28, 37 and 45 °C. Experimental data were fitted to 12 mathematical models. A nonlinear regression analysis method was used to evaluate the constants of the sorption equations. Statistical testing of sorption models was carried out using multiple criteria such as coefficient of determination (R (2) ), reduced chi-square (χ (2) ), mean relative percent deviation modulus (P) and plotting of residuals. BET (R (2) > 0.99; χ (2) < 0.09; P < 7.52; RMS% < 9.22) was found suitable for predicting the M e -a w relationship in the a w range of 0.10-0.53. However, in the a w range of 0.10-0.85, although Ferro-Fontan and GAB models were found to have high R (2) values (>0.99), Peleg model was found to meet the multiple statistical criteion (R (2) > 0.9996; χ (2) < 0.04; P < 3.97; RMS% < 7.09). Properties of sorbed water were also determined. BET, GAB and Caurie monolayer moisture contents ranged from 2.64 to 3.36, 1.29-2.66 and 1.88-3.38 % d.b., respectively. Second-order regression equation was found to describe the relation between monolayer moisture content, M o and temperature, t (°C). The isosteric heat, calculated using Clausius-Clapeyron equation, was varied between 1.46 and 50.39 kJ g(-1) mol(-1) at moisture levels 1-12 % (d.b.). An exponential relationship was observed between the isosteric heat of sorption and moisture content.