WEB Rational Design of Metal Anchored Carbon Quantum Dots as Optimal Efficient Electrocatalysts for Hydrogen production Using Machine Learning and High-throughput ExperimentationFriday (17.07.2020) 04:03 - 04:03 Poster Room Part of:
Hydrogen (H2) has been considered a promising candidate as a sustainable energy source for restricting carbon dioxide emissions, which electrochemical water splitting is recognized as one of the most potential technology for producing hydrogen. However, it is still a critical issue that practically required voltage for operating full water splitting needs more overpotential than the theoretical potential range of 1.23 V. As these challenges, electrocatalysts have been researched, while noble metal materials such as Pt and RuO2 are broadly used due to its superior catalytic activity. Although they have advantageous catalytic performance, their high cost and scarcity restrict the realization of economic electrochemical water splitting. To overcome the drawback, carbon-based materials showing earth-abundant, low cost, and high catalytic activity have studied for substituting noble metal-based catalysts. Especially, carbon quantum dots (CQDs) have been reported as an electrochemical catalyst due to nanoparticle size, high stability, and good electrical conductivity. In addition, the optimization design of experiments (DoE) has been increasingly attracted to achieve the superior performance of the prepared catalysts. Herein, Nitrogen-enriched CQDs (N-CQDs) are synthesized and doped with various heteroatoms (Ni, Co, Mn, Fe, and Zn) for enhancing the kinetic process of reactions. Moreover, DoE is investigated by a novel machine learning (ML) method of the Bayesian genetic algorithm (BGA) regarding various parameters for system such as loading amount, conductor, and pH conditions, which can affect catalytic activity. The ML-based method can minimize the number of experiments from 90,720 to 52 in order to reach optimized conditions where the catalyst can exhibit maximized catalyst performance. As a result, we discovered that the reasoning optimal DoE of Ni/N-CQDs exhibited the lowest overpotential value of 110 mV for HER.
|Category||Short file description||File description||File Size|
|Extended Abstract||Abstract||This is an abstract file.||138 KB||Download|