Tailoring Perfect Optical Vortex Combs Based on Diffractive Neural Networks

ABSTRACT Optical vortex (OV) combs, which simultaneously emit multiple OV beams, hold great potential for advanced optical communications and quantum technologies. However, their practical use has been limited because the size of each beam inherently depends on its topological charge (TC), making it difficult to achieve uniform beam profiles. Moreover, generating well‐controlled perfect OV combs—where all beams have identical ring sizes independent of TC—has remained an experimental challenge. In this work, we introduce and experimentally demonstrate the concept of a perfect OV comb that overcomes this fundamental limitation. We develop a diffractive neural network (DNN) approach to design optimized two‐layer phase masks that tailor these combs with high precision. Our method enables the generation of perfect OV combs with uniform size across all TCs, a capability not previously demonstrated. This breakthrough offers a new degree of freedom in structuring OV beams and opens up opportunities for applications in high‐capacity mode‐division multiplexing, optical trapping, and quantum information processing.