Efficient Material Discovery with AI Technology
On 20th April, researchers at the University of New South Wales (UNSW) announced the development of an AI-driven workflow aimed at advancing the discovery of next-generation semiconductor materials. This system employs artificial intelligence to replace the traditional trial-and-error method with a data-driven approach, focusing on hybrid perovskites used in solar cells and LEDs.
Hybrid perovskites are constructed by combining inorganic layers with organic molecules. These organic components are crucial because small molecular changes can radically alter the material’s performance, particularly in how it transports electrical charge. The AI system works by starting with desired outcomes, such as effective electrical charge handling, and then identifying molecules that could achieve these results.
Researchers developed this workflow to tackle the major bottleneck in materials discovery, where millions of possible combinations make it difficult to identify viable candidates. The AI efficiently screens numerous potential candidates by filtering out those unlikely to be practical to produce, narrowing the field to a small set of promising options. These are then checked using detailed simulations to confirm their performance.
Screening and Simulation Process
The approach addresses a longstanding issue in materials science, where researchers have typically made incremental changes to known materials instead of exploring new ones systematically. By applying the AI-driven workflow across millions of possibilities, the researchers could identify promising candidates more quickly and efficiently.
Although the candidates identified by the system have not yet been tested in laboratory settings, researchers are optimistic about the potential impact. They believe this AI-driven method could revolutionise the materials discovery process. By making the search process more efficient, it could significantly accelerate the development of new materials for electronics and energy technologies.
Dr. Alex Johnson, a lead researcher on the project, stated, “Our method stands to transform how materials are discovered. We can now focus on what we want a material to do and work backwards to find it.” The UNSW team hopes to begin lab testing within the next year.
For more information, interested parties can contact Tom Melville at 0432 912 060 or via email. A copy of the research paper is available online for further reading.
