Microwave nondestructive sensing of moisture content and water activity of almonds

Authors: Trabelsi, Samir

Submitted to: Transactions of the ASABE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/18/2021
Publication Date: 8/23/2021
Citation: Trabelsi, S. 2021. Microwave nondestructive sensing of moisture content and water activity of almonds. Transactions of the ASABE. https://doi.org/10.13031/trans.14338.
DOI: https://doi.org/10.13031/trans.14338

Interpretive Summary: Moisture content and water activity are the most critical parameters that determine the quality, safety, shelf life, and value of water-containing materials including agricultural products, foods, and pharmaceuticals. Conventional techniques for moisture content and water activity determination are destructive, time consuming, and not practical for in-process measurements and dynamic situations where large amounts are involved. Therefore, there is a need for the development of novel methods and sensors for in-process nondestructive and rapid determination of moisture content and water activity of materials. In this paper, an original dielectric method is proposed for determining nondestructively and simultaneously moisture content and water activity in granular and particulate materials from measurement of their dielectric properties at microwave frequencies. Because of the effect of changes in bulk density on dielectric properties, situations where the bulk density is known and situations where bulk density is unknown were considered. For purpose of illustration, analytical expressions of moisture content and water activity from measurement of the dielectric properties at a single microwave frequency were established for almonds kernels. Dielectric properties of almond kernels were measured in free space at 8 Gigahertz and room temperature (25 degree Celsius). To evaluate the effectiveness of moisture and water activity calibration equations, the standard error of calibration (SEC) was calculated for each equation. Results show that moisture content can be predicted with SECs ranging from 0.41% to 0.68% and water activity with SECs ranging from 0.02 to 0.04 for almond kernels samples with water activity ranging from 0.5 to 0.9 and moisture contents ranging from 4.8% to 16.5%. Similar approach can be used to determine moisture content and water activity calibration equations for other granular and particulate materials. Once these calibration equations are established they can be embedded in sensing systems for nondestructive and simultaneous determination of moisture content and water activity in those materials in static and dynamic situations.

Technical Abstract: A method for rapid and nondestructive determination of moisture content and water activity of granular and particulate materials was developed. The method relies on measurement of the dielectric constant and dielectric loss factor at a single microwave frequency. For purpose of illustration, the method was applied to predicting moisture content and water activity of almond kernels. A fee-space transmission technique was used for accurate measurement of the dielectric properties. Samples of Bute Padre almond kernels of moisture content ranging from 4.8% to 16.5% and water activity ranging from 0.50 to 0.93 were loaded into a Styrofoam sample holder and placed between two horn-lens antennas connected to a vector network analyzer. The dielectric properties were calculated from measurement of the attenuation and phase shift at 8 GHz and 25 oC. The dielectric properties increased linearly with moisture content while they showed an exponential increase with water activity. Situations where the bulk density was known and unknown were considered. Linear and exponential growth regressions provided the equations correlating the dielectric properties with moisture content and water activity with coefficients of determination, r2, higher than 0.96. Analytical expressions of moisture content and water activity in terms of the dielectric properties measured at 8 GHz and 25 oC are provided. Standard error of calibration (SEC) was calculated for each calibration equation. Results show that moisture content can be predicted with SECs ranging from 0.41% to 0.68% and water activity with SECs ranging from 0.02 to 0.04 for almond kernels samples with water activity ranging from 0.5 to 0.9 and moisture contents ranging from 4.8% to 16.5%.