Extraction of metals from oxides at high temperatures is essential not only for the production of metals such as steel but also for recycling. Since current extraction processes are very carbon intensive and emit large amounts of greenhouse gases, the researchers explored new approaches to develop “greener” processes. This work was particularly difficult to do in the laboratory because it requires expensive reactors. Building and running computer simulations would be an alternative, but currently there is no computational method that can accurately predict high temperature oxide reactions when no experimental data is available.
A Columbia engineering team reports developing a new computational technique that, by combining quantum mechanics and machine learning, can accurately predict the reduction temperature of metal oxides to their base metals. Their approach is as computationally efficient as conventional zero-temperature calculations and, in their tests, more accurate than computationally demanding simulations of temperature effects using quantum chemistry methods. The study, led by Alexander Urban, assistant professor of chemical engineering, is published today by Nature communication as quickly as possible.
“Decarbonizing the chemical industry is essential if we are to move to a more sustainable future, but developing alternatives for established industrial processes is very expensive and time consuming,” said Urban. “A bottom-up process design that does not require initial experimental input would be an interesting alternative but has not yet been realized. This new study is, to our knowledge, the first time that a hybrid approach, combining computational calculations with AI has been attempted for this application. And this is the first demonstration that calculations based on quantum mechanics can be used for the design of high temperature processes.
The researchers knew that at very low temperatures, calculations based on quantum mechanics can accurately predict the energy that chemical reactions require or release. They supplemented this zero temperature theory with a machine learning model that learned temperature dependence from publicly available high temperature measurements. They designed their approach, which focused on high temperature metal mining, to also predict how “free energy” will change with temperature, whether high or low.
“Free energy is a key quantity in thermodynamics and other temperature-dependent quantities can, in principle, be derived from it,” said José A. Garrido Torres, the first author of the article who was a fellow. postdoctoral fellow in the Urban lab and is now a research scientist at Princeton. “We therefore expect that our approach will also be useful in predicting, for example, melting temperatures and solubilities for the design of clean metal electrolytic extraction processes powered by renewable electrical energy.”
“The future just got a little closer,” said Nick Birbilis, assistant dean of the Australian National University College of Engineering and Computer Science and an expert in materials design focused on corrosion durability, who did not participated in the study. “Much of the human effort and sunk capital over the past century has been devoted to developing materials that we use every day – and on which we rely for our power, our flight and our entertainment. materials is slow and expensive, which makes machine learning an essential development for the design of future materials. For machine learning and AI to reach their potential, models must be mechanically relevant and interpretable. That is, precisely this is demonstrated by the work of Urban and Garrido Torres. In addition, the work takes an all-of-system approach for one of the first times, linking atomistic simulations on one side to engineering applications of the other – via advanced algorithms. “
The team is now working to extend the approach to other temperature-dependent material properties, such as solubility, conductivity, and melting, which are needed to design carbonless and fueled metal electrolytic extraction processes. by clean electrical energy.
Source of the story:
Materials provided by Columbia University School of Engineering and Applied Sciences. Original written by Holly Evarts. Note: Content can be changed for style and length.