Inverse Design of Microring Resonator-Based Glucose Biosensor Using AI Diffusion Models

Glucose detection is an important target for diabetes diagnostics. Microring resonators can be used as biosensors that detect small changes in refractive index. Conventional designs use parameter sweeps, finite difference time domain simulations, genetic algorithms, and adjoint optimization. These methods are accurate but limited in exploring the full design space. Artificial intelligence (AI) can learn the relation between structure and optical response and can generate device geometries beyond traditional design approaches. In this work, we use AI for inverse design of microring biosensors with diffusion models. The method takes a target optical response such as resonance wavelength shift and generates candidate microring geometries. Diffusion models can propose diverse solutions while staying consistent with optical targets.