PhoTOS: Topology Optimization of Photonic Components using a Shape Library

TL;DR

This paper introduces PhoTOS, a topology optimization framework for photonic components that uses a shape library and machine learning to ensure designs meet fabrication constraints, improving manufacturability.

Contribution

The work presents a novel TO method employing a Convo-implicit Variational Autoencoder to encode shape libraries, facilitating constraint-compliant photonic component design.

Findings

Successfully designed photonic components with guaranteed fabrication constraints

Demonstrated the effectiveness of shape library-based optimization

Enabled gradient-based optimization for complex photonic shapes

Abstract

Topology Optimization (TO) holds the promise of designing next-generation compact and efficient photonic components. However, ensuring the optimized designs comply with fabrication constraints imposed by semiconductor foundries remains a challenge. This work presents a TO framework that guarantees designs satisfy fabrication criteria, particularly minimum feature size and separation. Leveraging recent advancements in machine learning and feature mapping methods, our approach constructs components by transforming shapes from a predefined library, simplifying constraint enforcement. Specifically, we introduce a Convo-implicit Variational Autoencoder to encode the discrete shape library into a differentiable space, enabling gradient-based optimization. The efficacy of our framework is demonstrated through the design of several common photonic components.