Interlaboratory study of automated sorption measurements in wood: method for correcting systematic errors with the commonly used 0.002% min−1 stop criterion

Abstract Many studies that use an automated sorption balance to determine a water vapor sorption isotherm for wood collect data until the moisture content change is less than or equal to 0.002% min −1 (20 µg g −1 min −1 ). This stop criterion has been claimed to give errors in equilibrium moisture content (EMC) predictions of less than 0.001 g g −1 but over the past 10 years, studies have shown that the actual errors can be greater than 0.01 g g −1 because the measurements are stopped well before equilibrium is reached. Despite the large errors associated with this stop criterion, it remains popular due to the speed at which isotherms can be measured. This paper utilizes data from a worldwide interlaboratory study on automated sorption balances to develop a correction method for estimating EMC of western larch ( Larix occidentalis Nutt.) from the moisture content corresponding to the 20 µg g −1 min −1 criterion. The study uses data from 72 relative humidity absorption steps with hold times of 7–10 days from 21 different laboratories and eight different instrument models. EMC is defined based on the inherent mass stability of automated sorption balances determined in the first part of this interlaboratory study. On average the sorption process is less than 80% complete when the 20 µg g −1 min −1 criterion is reached, resulting in a mean absolute error (MAE) of 0.006 g g −1 . The correction equation for estimating EMC reduces the MAE to 0.001 g g −1 . The analysis presented in this paper, along with the correction equation, can be considered for certain use cases to reduce systematic errors and shorten measurement times.