10-Dec-2021 - Martin-Luther-Universität Halle-Wittenberg (MLU)

Turbo boost for materials research

Researchers train AI to predict new compounds

A new algorithm has been designed to help discover previously unknown material compounds. It was developed by a team from Martin Luther University Halle-Wittenberg (MLU), Friedrich Schiller University Jena and Lund University in Sweden. The researchers designed a form of artificial intelligence (AI) based on machine learning that can perform complex calculations within a very short space of time. This has enabled the team to identify several thousand potential new compounds using a computer. The study was published in the journal Science Advances.

Inorganic materials are essential for humans. For example, they form the basis for solar cells and for new advancements in semiconductor electronics that are used in technical devices. Around 50,000 stable inorganic compounds have already been identified. "However, considerably more may theoretically exist - if they can be produced artificially," says Dr Miguel Marques, professor of physics at MLU. There are two basic ways to detect these undiscovered materials: in the laboratory via countless experiments on different substances, or through computer simulation. The latter has increasingly become standard in recent years, says Marques: "The problem is that many earlier approaches require a lot of computing power and are slow to produce results." 

The researchers therefore developed a new method based on machine learning. Instead of performing whole calculations, the computer predicts their final results. "In other words, we want to obtain the results of the calculations without having to do the actual calculations," says Jonathan Schmidt from MLU, first author of the new study. "This requires two things: an algorithm that carries out the desired task, and a dataset which can be used to train the algorithm," adds the physicist. The team used several databases containing over 2.4 million compounds. "The calculations on which these databases are based have a combined calculation time of 100 to 200 million hours," says Schmidt.   

The new AI searches for new materials much faster than previous methods and is expected to soon also predict their electrical and optical properties. The researchers have already been able to identify several thousand possible candidates. "Of course, promising material candidates and their properties have to be confirmed by experiments and investigated further. However, we are very confident that most of our predictions will be confirmed," says Marques.

Facts, background information, dossiers
  • material science
  • artificial intelligence
  • machine-learning
  • prediction models
  • new materials
More about MLU
More about Uni Jena
  • News

    Exciting light emission and measuring temperature with ultrasound

    If mechanoluminescent materials are subjected to external mechanical stress, they emit visible or invisible light. Such excitation can occur due to bending or gentle pressure, for example, but also completely contact-free through ultrasound. In this way, the effect can be triggered remotely ... more

    How the first biomolecules could have been formed

    The chemical precursors of present-day biomolecules could have formed not only in the deep sea at hydrothermal vents, but also in warm ponds on the Earth's surface. The chemical reactions that may have occurred in this “primordial soup” have now been reproduced in experiments by an internat ... more

    Catalytic hydrogen generation – without expensive precious metals

    A research team from Friedrich Schiller University Jena has developed a molecular photosystem inspired by nature that generates hydrogen under visible light irradiation. In contrast to other existing systems of this type, it functions without the use of precious or heavy metals. Low-cost an ... more

More about Lund University
  • News

    Breakthrough in converting CO2 into fuel using solar energy

    A research team led by Lund University in Sweden has shown how solar power can convert carbon dioxide into fuel, by using advanced materials and ultra-fast laser spectroscopy. The breakthrough could be an important piece of the puzzle in reducing the levels of greenhouse gases in the atmosp ... more

    X-ray view into hydrogen production

    Green hydrogen is expected to play a central role in the energy industry of the future: Among other things, it can replace oil and gas as energy carriers and also enable emission-free production of steel and cement. For an efficient hydrogen economy, however, the production of the energy-co ... more

    Lead Halide Perovskites – a Horse of a Different Color

    In a joint experimental and theoretical effort between Lund University (Sweden), the Russian Academy of Science (Russia), and the Center for Advancing Electronics Dresden at Technische Universität Dresden (Germany), researchers developed a novel spectroscopic technique for the study of char ... more