ADVANCED MATERIALS & PROCESSES | NOVEMBER/DECEMBER 2023 10 MACHINE LEARNING | AI The School of Engineering and Sciences at Tecnológico de Monterrey’s Guadalajara campus and Wizeline, a tech services company, are launching the first generative artificial intelligence (AI) laboratory in Mexico and Latin America, named G.AI.L. The lab will serve as a central AI hub for the region. wizeline.com. BRIEF EXASCALE SUPERCOMPUTERS AND AI The DOE’s Argonne National Laboratory, Lemont, Ill., is building one of the first exascale systems in the U.S., named Aurora. To prepare codes for the scale of the new supercomputer, 15 teams are taking part in the Aurora Early Science Program through the Argonne Leadership Computing Facility. “The power of exascale supercomputers combined with advances in AI will provide a huge boost to the process of materials design and discovery,” says computational scientist Anouar Benali. He is leading a project to prepare a materials science and chemistry code called QMCPACK for Aurora. Developed in collaboration with Intel and Hewlett Packard Enterprise, Aurora is expected to be one of the world’s fastest supercomputers. QMCPACK is an open source code that uses the Quantum Monte Carlo (QMC) method to predict how electrons interact with one another for a wide range of materials. “With each new generation of supercomputer, we are able to improve QMCPACK’s speed and accuracy in predicting the properties of larger and more complex materials,” says Benali. “Exascale systems will allow us to model the behavior of materials at a level of accuracy that could even go beyond what experimentalists can measure.” Ultimately, the computations will help guide and speed up experiments aimed at discovering new materials. The QMCPACK team works closely with experimental groups to help pinpoint strong candidates for testing in a laboratory. “We want our experimental colleagues to be able to focus on a shortlist of the most promising materials,” adds Benali. “So having reliable simulations is becoming an increasingly important part of the materials design and discovery process.” anl.gov. AI TO DEVELOP HYDROGEN FUEL CELL CATALYSTS Proton exchange membrane hydrogen fuel cells used in hydrogen Aurora supercomputer at Argonne National Laboratory. vehicles require platinum catalysts to facilitate the oxygen reduction reaction at the anode. However, numerous elemental combinations and compositions could be explored in order to find alternatives to expensive platinum catalysts. Now, scientists at the Korea Institute of Science and Technology (KIST) have presented a new AI-based catalyst screening method and succeeded in developing a new catalytic material stemming from a ternary element- based alloy. It is less costly and performs more than twice as well as pure platinum catalysts, according to the researchers. The team developed the Slab Graph Convolutional Neural Network AI model to accurately predict the binding energy of adsorbates on the catalyst surface. Researchers were able to explore the potential of nearly 3200 ternary candidate materials in just one day, a task that would have taken years using the density functional theory adsorption energy simulation calculations traditionally employed to predict catalyst properties. The scientists developed the novel ternary alloy (Cu-Au-Pt) catalyst through experimental validation of 10 catalysts that showed potential to outperform the usual platinum versions. The new catalyst uses just 37% of the platinum required for pure platinum catalysts, and the kinetic current density is more than twice as high. https://eng. kist.re.kr. Graphical abstract of machine learning-driven hydrogen fuel cell catalyst design. Courtesy of KIST.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==