Use of Machine Learning in Electrosynthesis of H2O2 and Electroanalytical Sensor for Environmental Utilization

• Grant number: 24/03972-8
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: September 01, 2024
• End date: August 31, 2026
• Field of knowledge: Engineering - Sanitary Engineering - Water Supply and Wastewater Treatment
• Principal Investigator: Marcos Roberto de Vasconcelos Lanza
• Grantee: Augusto Duarte Alvarenga
• Host Institution: Instituto de Química de São Carlos (IQSC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

This postdoctoral project aims to develop an autonomous in situ hydrogen peroxide (H2O2) production system, using advanced machine learning (ML) techniques. The proposed approach focuses on the optimization of carbon-based cathodic catalysts through artificial neural networks, incorporating information related to the morphological and physicochemical characteristics of the catalyst, synthesis conditions, and operation of the electrochemical reactor. The modeling in neural networks will be based on parameters such as level of defect in the carbon network, hydrophilicity, configuration of the carbon network, presence of heteroatoms and dopants, surface area, porosity, pore size and volume, roughness, yield, and selectivity of H2O2 production, faradaic efficiency, adsorption forces and variations in the operating conditions of the electrochemical reactor and catalyst synthesis. The methodology will include the production of catalysts with optimized characteristics, followed by tests in different reactors under different operating conditions and levels of organic matter to be oxidized by H2O2. The predictability of H2O2 production will be improved through ML tools, making it possible to determine the amount of H2O2 required to oxidize a given volume of organic matter. Furthermore, an organic matter sensor will be developed based on real-time electrical impedance tests to control the electrochemical reactor current and thus the production of H2O2. The project will also employ additional techniques such as decision trees, principal component analysis, and clustering. In the end, it is expected to obtain a prototype of an autonomous system capable of producing H2O2 in a predictive way, dynamically responding to the level of organic matter present in the water, thus contributing to significant advances around sustainable water treatment technologies.