Dr. Hannah-Noa Barad
Department of Chemistry, Center of Nanotechnology & Advanced Materials, and Israel National Institute of Energy Storage, Bar Ilan University, Ramat Gan, Israel
As a major part global climate change mitigation and improvement of sustainable resources, discovery of new, stable, and highly active photovoltaic and catalytic materials is a pressing issue. The efforts have focused on abundant, accessible, low-cost, stable alternatives that will yield process efficiencies comparable or better than those we have today. For example, for water splitting, many new materials with different compositions have shown promising results as catalysts. However, they are mostly prepared by wet chemical synthesis, which results in chemical waste and can be too slow for industrial use. Thus, there is high motivation to accelerate the process of finding new materials with varying nanostructures and optimized functionality, by systematic exploration of several parameter spaces.
In recent years, artificial intelligence, namely by machine learning (ML) tools, has gained prominence in the field of materials science. The use of ML accelerates new material predictions and assists with finding unexpected correlations between the process-structure-function relations of materials, which leads to a better understanding and focus of the vast parameter spaces that exist in materials science. Rational design by ML in conjunction with combinatorial materials science promotes the rapid discovery and analysis of new materials, and enables breakthroughs in materials science, which would otherwise not have been possible.
Here I present the progress in the development of materials using rational design with ML in conjunction with combinatorial synthesis and high-throughput characterization. We investigate changes in composition and nano-morphology on material libraries and their effects on photovoltaic activity and the catalysis of reactions such as O2 evolution, CO2 reduction, and CH3OH oxidation. The different nanostructures and compositions show high activity and stability. The insights gained, indicate a dependence of catalytic activity on composition and nanostructuring, which the standard experimental techniques cannot achieve or explore, thus illustrating the importance and impact that composition and structure have, and will have, on developing sustainable materials. This can only be done by high-throughput experimentation design, combined with machine learning tools, which will assist with appropriate path directions and ensure rational studies on sustainable materials in the future.