The heat of sorption ( \(\:{Q}_{L}\) ) of four high extractive content woods—western red cedar, Chinese juniper, tubi and messmate—were determined using an isosteric approach based on previously collected water sorption isotherm data at 30, 45, 60, 75, 90 and 99.5 ℃. The isotherm data at every single temperature ( \(\:T\) ) was fitted using the Hailwood-Horrobin (HH) model. The resulting \(\:{Q}_{L}\) versus moisture content ( \(\:M\) ) plot displays a characteristic peak in the low \(\:M\) region. This peak is sensitive to the sorption data selected and can be complicated by overfitting of the HH model. In contrast, when sorption data from all \(\:T\) levels were fitted using the multi-temperature Heikkilä model, the predicted \(\:{Q}_{L}\) versus \(\:M\) curves closely aligned with findings obtained from the calorimetric method. The Heikkilä model suggests that \(\:{Q}_{L}\) varies with \(\:T\) throughout the entire hygroscopic range and specifically decreases with increasing \(\:T\) . Other multi-temperature models being evaluated, including the Chung-Pfost, Day-Nelson and Zuritz models, were less reliable in predicting \(\:{Q}_{L}\) values and their dependence on \(\:T\) . Both the HH and Heikkilä models indicate that as \(\:M\) approaches zero, the \(\:{Q}_{L}\) values for untreated samples are higher than those for extracted samples. This suggests that the adsorbed water molecules can disrupt the weak physical interactions between wall polymers and extractive molecules. Consequently, this process can lead to stable or even increased \(\:{Q}_{L}\) values if more available sorption sites are exposed in the swollen cell walls due to the deposition of extractives.
