Genetic algorithm enables precise design of phononic crystals

The advent of quantum computers promises to revolutionize computing, solving complex problems exponentially faster than classical computers. However, today’s quantum computers face challenges such as maintaining stability and transporting quantum information.

Phonons, which are quantized vibrations in periodic lattices, offer new ways to improve these systems by strengthening qubit interactions and providing more robust information conversion. Phonons also facilitate better communication in quantum computers by allowing them to be interconnected in a network.

Nanophononic materials, which are artificial nanostructures with specific phononic properties, will be essential for next-generation quantum networks and communication devices. However, designing phononic crystals with desired vibrational characteristics at the nano- and micro-scale remains a challenge.

In a study recently published in the journal ACS NanoScientists at the Institute of Industrial Science at the University of Tokyo have experimentally proven that a new genetic algorithm enables automatic inverse design — which generates structure based on desired properties — of phononic crystal nanostructures, enabling the control of acoustic waves in the material.

“Recent advances in artificial intelligence and inverse design offer the opportunity to search for irregular structures that exhibit unique properties,” explains lead author Michele Diego.

Genetic algorithms use simulations to iteratively evaluate proposed solutions, with the best ones passing on their features, or “genes,” to the next generation. Example devices designed and manufactured using this new method were tested using light scattering experiments to determine the effectiveness of the approach.

The team was able to measure vibrations in a two-dimensional phononic “metacrystal” that had a periodic arrangement of smaller designed units. They showed that the device allowed vibrations along one axis but not along the perpendicular direction, and could therefore be used for acoustic focusing or waveguides.

“By extending the search for optimized structures with complex shapes beyond normal human intuition, it becomes possible to design devices with precise control over the propagation properties of acoustic waves quickly and automatically,” says senior author Masahiro Nomura. The approach is expected to be applied to devices using surface acoustic waves in quantum computers, smartphones and other devices.