Innovative electrochemical synthesis of Portland cement precursor (Ca(OH)₂) with machine learning-enabled predictive insights

For long the manufacturing of calcium hydroxide, a major precursor of cement production, relies heavily on calcining, which not only costs much energy, but also releases a significant amount of carbon dioxide. To electrochemically synthesize calcium hydroxide is a high-potential substitution, as it utilizes pH gradient to decarbonate limestone and generate calcium hydroxide, which can be carried out at room temperature. But creating a pH gradient with electrolysis of water requires a relatively high voltage, which could be energy-draining in manufacturing. So here an innovative three-chamber electrolysis system based on hydrogen evolution reaction (HER) and hydrogen oxidation reaction (HOR) is proposed, which lowers the electrolysis voltage while achieving very high current efficiency (~100%). A regularized linear regression model and a multi-layer perceptron are also developed to analyze the influence of key factors, including current density and electrolyte concentration, on the calcium hydroxide production rate of the system. The possibility of further applying deep learning to optimize experimental parameters, recognize major restrictions from high-dimensional data, and predict test results is also discussed. This work offers a new low-energy, decarbonized path for cement precursor production and references for combining electrochemistry with machine learning.