Reaching machine learning leverage to advance performance of electrocatalytic CO2 conversion in non-aqueous deep eutectic electrolytes

Abstract

Abstract Deep eutectic electrolytes (DEEs) show promise for future electrochemical systems due to their adjustable buffer capacities. This study utilizes machine learning algorithms to analyse the carbon dioxide reduction reaction (CO2RR) in DEEs with a buffer capacity of approximately 10.21 mol/pH. The objective is to minimize undesired hydrogen evolution reactions (HER) and render CO2RR dominant in a membrane cell. The CO2RR process was found to be non-adiabatic, as the time of nuclear motion for CO3 2− in K2CO3 product, through CO2 ●− trapping, is 0.368 femtoseconds shorter than the 1.856 × 10−3s charge transfer relaxation time. Microkinetic analysis reveals that the rate of CO2RR to CO2 ●− is 2.14 × 103 mol/cm2/s2 with a rate constant of 2.1 × 1010 cm/s. Our findings demonstrate that ensemble and k-Nearest Neighbours algorithms learn the CO2RR dataset, achieving a prediction accuracy of over 99%. The models were verified visually and quantitatively by overlaying predicted and experimental dataset. Diagnostic and SHAP analyses highlighted the gradient boost ensemble algorithm, predicting asymptotic current densities of -4.114 mA/cm2 or -13.340 mA/cm2, with high turnover frequencies (TOF) of 3.79 × 1010 h-1 or 12.30 × 1010 h-1 for CO2 ●− or K2CO3 generation on silver electrodes, respectively. These results consider both accuracy and robustness against overfitting, providing an opportunity to optimize future non-aqueous electrolytes for convenient TOF measurements at industrially relevant current densities.

Keywords

• Deep eutectic electrolytes
• CO2 capture and utilization
• Machine learning
• Artificial Intelligence
• Electrocatalysis
• Reaction Engineering