Researchers at the University of New South Wales (UNSW) have developed an artificial intelligence-assisted workflow aimed at accelerating the discovery of next-generation semiconductor materials by replacing traditional trial-and-error methods with a targeted, data-driven approach, according to UNSW.
UNSW said the design of advanced materials has historically been a slow and iterative process, where small molecular changes can significantly affect performance. However, the university noted that the sheer number of possible molecular combination – potentially running into the millions – has made it difficult for researchers to efficiently identify viable candidates, creating a persistent bottleneck in materials science.
In work led by UNSW researchers, the focus is on hybrid perovskites, a class of semiconducting materials used in applications such as solar cells and light-emitting diodes.
UNSW explained that these materials are formed by combining inorganic layers with organic molecules, with the organic components playing a critical role in determining key properties, including how electrical charge is transported.
Rather than relying on incremental modifications to existing materials, UNSW said the new workflow operates in reverse. Starting from a defined performance target – such as desired electrical charge transport characteristics – the system identifies candidate molecules that could achieve that outcome. It then screens large volumes of possibilities and eliminates those unlikely to be practical to synthesise.
According to UNSW, the approach was applied across millions of potential combinations, ultimately narrowing the field to a small set of promising candidates. These were then subjected to detailed computational simulations to evaluate their performance in more depth.
UNSW said the method addresses a long-standing challenge in materials research, where discovery has typically relied on gradual adjustments to known compounds rather than systematic exploration of broader chemical spaces, largely due to time and cost constraints.
While the identified candidates have not yet been validated through laboratory experiments, UNSW said the workflow could significantly improve the efficiency of materials discovery.
The university added that the approach has potential to accelerate the development of new materials for electronics and energy technologies by streamlining the search process for viable semiconductor compounds.
