Graphene-enhanced dual-gate electro-optic micro-ring modulator: An optical neural network design

To improve integrated photonic systems for optical computing applications, the ability to control both the amplitude and phase of microring modulator transmission is a key capability for advanced photonic integrated circuits. In this work, we present a novel dual-gate microring modulator that integrates a Z-shaped silicon PN junction with a graphene overlay, offering enhanced electro-optic tunability beyond conventional single-gate designs. The device consists of two electrically coupled active regions, one PN junction and the other incorporating graphene, allowing modulation of both the real and imaginary parts of the effective refractive index through control voltages V an and V gr . This dual-gate approach provides control over the transmission phase and amplitude, which are important for complex-valued modulation in programmable photonic circuits. We further demonstrate its utility by implementing a simplified optical neural network (ONN) architecture to illustrate end-to-end feasibility, achieving accuracies of 93.4% and 81.95% on the MNIST and Fashion-MNIST datasets, respectively. Multiphysics simulations using Lumerical Charge, finite-difference eigenmode (FDE), and finite-difference time-domain (FDTD) validate the modulator’s performance and highlight its potential for next-generation integrated photonic computing platforms.