Reconfigurable phononic crystal platform guided by user specifications

초록

This study presents a reconfigurable phononic crystal (PnC) platform that combines an automated structure generation algorithm with a modular physical assembly strategy. This modular sensor platform allows users to simply specify the sensing purpose, such as target frequency and material characteristics, and it automatically reconfigures to deliver a customized sensor through dual-stage optimization, ultimately overcoming the repeated redesign and refabrication required by conventional non-reconfigurable systems. In the first stage, a genetic algorithm (GA) maximizes the bandgap through lattice evolution, establishing a robust physical foundation. In the second stage, a physics-informed genetic algorithm (PIGA) fine-tunes defect-mode placement based on acoustic properties, ensuring functional responsiveness. By sequentially combining GA with PIGA, the framework unites broad search capability with physics-informed refinement, yielding optimization outcomes surpassing the limits of single-stage methods. This dual-stage strategy enables the platform to achieve exceptional sensitivity and accuracy across diverse targets, while a Lego-inspired modular architecture physically realizes the digitally evolved optimal designs through interchangeable unit components. This facilitates efficient assembly, disassembly, and repeated reuse of the same components across various applications without requiring additional fabrication. As a representative example, the platform was validated through liquid sensing experiments, including fluid identification and concentration sensors using liquid binary mixtures. Clear correlations were identified between defect-mode frequency and liquid composition, demonstrating the platform's robust and sensitive performance. These findings establish the platform as a general-purpose solution for implementing reconfigurable PnC in dynamic and multiphase environments.

키워드