A D V A N C E D M A T E R I A L S & P R O C E S S E S | J U L Y / A U G U S T 2 0 2 2 6 MAKING MAPS FOR METAL ELECTROLYSIS A team of researchers from MIT, Cambridge, Mass., and SLAC National Accelerator Laboratory, Menlo Park, Calif., are mapping what occurs at the atomic level during metal electrolysis, a process in which a metal oxide is bombarded with electricity to create pure metal with oxygen as the byproduct. They say their work could lead to more efficient and environmentally friendly processes for producing metals such as lithium, iron, and cobalt. By making maps for a wide range of metals, the scientists not only determined the metals that should be easiest to produce using metal electrolysis, but also identified barriers that hinder the efficient production of others. RESEARCH TRACKS The research could also boost development of metal-air batteries such as lithium-air, aluminum-air, and zinc-air batteries. These are similar to the lithium-ion batteries used in today’s electric vehicles, and they have the potential to electrify aviation because of their much higher energy densities. Metal-air batteries are not yet on the market due to a variety of problems, including inefficiency. All of the research was conducted using supercomputer simulations that explored different scenarios for the electrolysis of several metals, each involving different catalysts. The team’s new map is essentially a guide for designing the best catalysts for each metal, say researchers. web.mit.edu. NEURAL NETWORK PREDICTS STEEL PART LIFETIMES A team of researchers from Russia, Turkey, Canada, and Italy created an artificial deep neural network that is able to predict the lifetime of a component made of AISI 1045 steel—along with choosing the optimal coating and its thickness. First, the scientists conducted a series of MIT researcher Jaclyn Lunger is detailing the atomic-level reactions behind an eco-friendly way to make metals. Courtesy of Yang Shao-Horn/MIT. physical experiments on steel parts. Approximately 23% of the data was used to train the neural network while the rest was used for testing and validation of the resulting predictions. The team tried out several neural networks, with different numbers of inner layers and neurons in each layer. Nearly 99% accuracy was reported for the best neural network’s predictions. Nickel, hardened chromium, and the galvanization process were used as protective coatings in the model. Further, the scientists were able to determine the optimal protective coating, which turned out to be a 10-15 µm layer of nickel or zinc. Hardened chromium was found to reduce the fatigue lifetime of steel. The team included researchers from RUDN University, Russia, Karabuk University, Turkey, Ontario Tech University, Canada, and the Polytechnic University of Milan. www.rfbr.ru/rffi/eng. An international team of scientists is using an artificial deep neural network to predict the stability of steel parts and find the best protective coating. Scientists from Texas A&M University, College Station, and Yonsei University, Seoul, discovered a helicoidal-shaped defect in layered polymers, revealing how solvents can diffuse through layers and produce color changes. Because stimuli-interactive structural color holds immense potential for devices such as health sensors and human-interactive electronics, controlling the lateral spacing or amount of helicoidal defects could be a critical factor in future applications, say researchers. tamu.edu, www.yonsei.ac.kr. BRIEF
RkJQdWJsaXNoZXIy MTYyMzk3NQ==