Deep-Learning-Guided Electrochemical Impedance Spectroscopy for Calibration-Free Pharmaceutical Moisture Content Monitoring.

The moisture content of pharmaceutical powders can significantly impact the physical and chemical properties of drug formulations, solubility, flowability, and stability. However, current technologies for measuring moisture content in pharmaceutical materials require extensive calibration processes, leading to poor consistency and a lack of speed. To address this challenge, this study explores the feasibility of using impedance spectroscopy to enable accurate, rapid testing of moisture content of pharmaceutical materials with minimal to zero calibration. By utilizing electrochemical impedance spectroscopy (EIS) signals, we identify a strong correlation between the electrical properties of the materials and varying moisture contents in pharmaceutical samples. Equivalent circuit modeling is employed to unravel the underlying mechanism, providing valuable insights into the sensitivity of impedance spectroscopy to moisture content variations. Furthermore, the study incorporates deep learning techniques utilizing a 1D convolutional neural network (1DCNN) model to effectively process the complex spectroscopy data. The proposed model achieved a notable predictive accuracy with an average error of just 0.69% in moisture content estimation. This method serves as a pioneering study in using deep learning to provide a reliable solution for real-time moisture content monitoring, with potential applications extending from pharmaceuticals to the food, energy, environmental, and healthcare sectors.