Machine Learning-Enhanced Electrochemical Sensors for High-Temperature Molten Salts Enabled by Rich Experimental Data Production

Molten salts are essential in next-generation energy technologies, including molten salt reactors and solar power systems, but their extreme environments pose challenges for the reliability of electrochemical sensors. High temperatures and corrosive conditions cause degradation in traditional sensors, leading to variations in electrode surface area and unreliable measurements. This study introduces a durable electrochemical sensor with a simple multiarray tungsten electrode design that withstands harsh conditions for the long term. By integrating machine learning, the sensor accurately predicts electrode geometric surface area changes, addressing a critical limitation. Using a controlled, automated experimental framework spanning a wide range of conditions in high temperature molten salts, 3432 electrochemical data points were collected, and the ML model achieved a mean absolute percentage error below 6%, validated across diverse molten salt systems. These findings highlight a robust solution for AI-driven sensors that enhance safety, efficiency, and long-term reliability in high-temperature energy applications.