The Facebook parent company will release data about catalysts to make hydrogen fuel cheaper to make.
Meta has announced that it will be releasing catalyst data for artificial intelligence research that could make it possible to develop a cheaper way to produce hydrogen fuel, a carbon emission-free energy source.
Meta’s catalyst data set was developed as a part of a collaboration with Carnegie Mellon University.
The data set would be useful in artificial intelligence research used to develop cheaper chemical catalysts that are used in the vital reactions that are needed for making hydrogen fuel. At the moment, the catalysts used to make this promising form of green energy are made from expensive precious metal oxides like ruthenium and iridium oxide. As a result, hydrogen fuel cell technology remains notably more expensive than fossil fuels, slowing its adoption even as the world attempts to decarbonize.
That said, the AI division at Meta and the Department of Chemical Engineering at Carnegie are working together to build machine learning models that will simulate the chemical reactions that would occur with various types of catalysts made from different materials. The data set they have already developed is the largest in that field and could prove to be a game changer in accelerating low-cost catalyst discovery.
The artificial intelligence research will benefit from a massive chemical reaction simulation dataset.
The Open Catalyst Project (OCP) features over 1.3 million molecular adsorptions onto surfaces, making it the largest electrocatalyst structure data set currently in existence. New data also includes approximately 8 million data points having to do with about 40,000 unique simulations. Together, these are a thorough examination of 52 elements simulated with a vast range of oxide materials.
Using artificial intelligence in this way could provide a powerful tool to leapfrog many tests that would have had to be conducted in reality using real materials, costing a substantial amount more and requiring a tremendous amount of additional time. By using this data set, it could not only move hydrogen fuel technology forward exponentially but also make it much cheaper. As many countries are looking to H2 as a top alternative to fossil fuels, bringing down the price tag will be critical to the viability of those plans.
