Delayed harvest effects on the moisture sorption properties of soybeans

Absorption and desorption characteristics were determined for three varieties of soybean seeds that were harvested over 16 weeks in two crop years. Harvest began at physiological maturity and extended well beyond a normal harvest period. Fully exposed absorption tests were followed by desorption tests in all cases under near constant temperature and air flow conditions. Experimental data were analyzed by empirical, theoretical, and statistical models.

An investigation of the morphological features of soybean seeds led to sorption tests with two soybean varieties which included covering the hilum area (approximately 10% of the total surface area) with wax. It was concluded that the hilum area did not provide a preferential route for moisture vapor exchange with the atmosphere during absorption or desorption.

The empirically based modified Page equation was used to estimate the moisture equilibrium value by non-linear regression and found to adequately describe both absorption and desorption behavior. Equation parameters k and n were used to generate two descriptive parameters, kⁿ and k^m, for each replication of variety and week of harvest to help identify the effects of field weathering. The n and k^m parameters were found to be more sensitive to harvest moisture conditions for absorption than the k and kⁿ parameters but this was not repeated in both years or found for desorption in either year.

The level of protein, oil, free fatty acids, harvest moisture, and seed size were merged with the variables found by regression from the modified Page equation that characterized the absorption and desorption responses. The statistical model that accounted for most of the variation in k values for 1989 contained the covariates of harvest moisture and free fatty acid, while the corresponding model in 1990 contained the covariate of oil. Also in 1989, the desorption EMCR value was correlated to seed size.

These findings suggested that the role of chemical components on the rate and extent of moisture transfer in seeds or other grains requires further investigation. It is suggested that non-destructive technique be used to facilitate measurements with the same sample during a moisture transfer cycle, which occurs every year in the field, in storage environments, and during transport where seeds are fully exposed to a wide range of environmental conditions.

The cellular diffusion model was paired with a selected set of sorption data to determine estimates for the theoretical parameters (Biot number and diffusion term) by regression. The diffusion model produced estimates for the moisture ratio that were in very good agreement with the data but were slightly less accurate than those produced by the modified Page equation. The most important result from this comparison was that the Biot number had a similar effect on the relationship that describes the rate of moisture transport (for both absorption and desorption) as the n value in the modified Page equation.