A Terahertz time-domain spectroscopy-based intelligent microwave vacuum dehydration system for in-situ monitoring of real-time moisture migration during dehydration of agricultural products

Microwave vacuum drying (MVD) as a novel and advanced drying technology is wildly applied in the agri-food, pharmaceutical, and electronics industry, etc. However, achieving in-situ analyses of the real-time moisture content (MC) of products during MVD is still a huge challenge. This research aims to build an intelligent dehydration system by integrating a Terahertz time-domain spectroscopy (THz-TDS) into an MVD system to achieve in-situ monitoring of real-time moisture content (MC) in agri-food products. During dehydration, the THz-TDS continuality captured the spectra from a polyethene (PE) air hose containing the exhaust water vapor from the glass desiccator. Chemometric analysis of Gaussian process regression was applied to correlate the real-time MC loss of products with the corresponding numerical integration of THz absorption coefficients of vapor. The real-time MC content was accurately calculated based on the prediction. The result shows that the THz-TDS is able to sense the dynamic vapor changes within the MVD system by penetrating the PE air hose, and the established intelligent dehydration system combined with chemometric analyses achieved a satisfied MC loss and MC prediction, with R-2 of 0.95 and 0.98, and RMSE of 0.002 and 0.031, respectively. The THz-TDS technique shows great potential to be integrated into an MVD system to achieve in-situ real-time MC evaluation, optimize dehydration condition, and predict the dehydration endpoint. The established intelligent dehydration system also provides a novel sensing strategy using THz-TDS to monitor the gas exchange within a closed system by penetrating a PE air hose in the system, and further widen the application of THz-TDS in the agri-food industry.