Quantifying and reducing errors in equilibrium moisture content measurements with dynamic vapor sorption (DVS) experiments

Abstract

Dynamic vapor sorption (DVS) measurements are widely used to collect water vapor sorption isotherms for wood and other cellulosic materials. Equilibrium moisture content (EMC) is typically assumed to have been reached when the rate of change in moisture content with time (${\text{d}}M/{\text{d}}t$) drops below a certain value. However, the errors associated with determining EMC in this manner have never been characterized. Here, an operational definition of equilibrium for DVS measurements is provided, and twenty test cases over four cellulosic materials are presented where the relative humidity was stepped up or down and then held constant until equilibrium was reached. Then, both the time to reach various ${\text{d}}M/{\text{d}}t$ “stop criteria” and the errors in EMC associated with those stop criteria are quantified. The errors in the EMC from the widely used 0.002% min⁻¹ stop criterion are found to be as large as 1.2% MC, and the average error for 20 test cases is 0.5% MC, which are much larger than the 0.1% MC error claimed in the literature. Longer data collection times are recommended, and a more stringent ${\text{d}}M/{\text{d}}t$ criterion (0.0003% min⁻¹, using a 2-h window) for cellulosic materials is proposed. The errors with this criterion are less than 0.75% MC, and the average error is 0.3% MC. Furthermore, it is shown that the errors for a given stop criterion are systematic and can be fairly well characterized with a simple linear regression. Finally, a correction for systematic error is proposed that results in more accurate EMC values with shorter hold times.