Controlled generation of nonlinear dynamic signals in a perturbed photorefractive two-wave mixing process

The miniaturization of silicon-based electronics is reaching its fundamental limitations, necessitating the search for alternative computational paradigms. One of the promising directions is to consider inherently nonlinear dynamic systems for computing operations, instead of the linear von Neumann system. This approach requires the system to exhibit rich and controllable nonlinear dynamic behavior that is highly suitable for modern machine learning operations. Here, what we believe is for the first time, we experimentally demonstrate controllable and sustainable nonlinear dynamic behavior exhibited by a simple photorefractive two-wave mixing setup (PR-TWM) without any external feedback. The novel idea used here is to apply a sinusoidal electric field across the photorefractive sample in the two-wave mixing (TWM) process. As a result, a period-doubling type of nonlinear dynamic behavior was observed, which was then systematically analyzed using phase-space plots, attractor dimension calculation, bifurcation plots, and Lyapunov exponent values. Our findings demonstrate a direct correlation between applied field strength and the attractor dimension, highlighting a controllable nonlinear dynamical behavior. Such a simple system with controllable complex nonlinear dynamic behavior has potential to be utilized in the field of nonlinear dynamics-based computing paradigms such as reservoir computing, neuromorphic computing, in-materio computing, analog computing, etc.