Microstructering of lithium niobate for application in reservoir computing

We present advances in lithium niobate (LN) photonics for reservoir computing applications. Our dry-etch process achieves 2.5 µm features with sub-10 nm roughness despite lithium fluoride byproduct challenges. While LN platforms exist for time-delay computing, we explore LN as an interconnecting matrix for spatio-temporal reservoir systems. Our approach inscribes optimized phase values through etching, leveraging LN’s nonlinearities to enhance computational performance—demonstrating LN’s advantages.