Optimizing aperiodic optical phased arrays with non-uniform discrete Fourier transform

We present an efficient method for optimizing sparse aperiodic optical-phased arrays using the non-uniform discrete Fourier transform (NDFT). By employing NDFT for far-field calculations, computation time is reduced by three orders of magnitude while maintaining accuracy compared with far-field equation-based computations and finite-difference time-domain simulations. Then, this method is implemented with array positional optimizations, including jittered sampling, Poisson point processes, pattern search, and particle swarm optimization (PSO). Among these methods, PSO achieves the best results, with a grating lobe suppression ratio of −16.25 dB at 0 deg, steering rate of 13.5 deg/nm, and field of view of ±41 deg across the 1.5 to 1.6 μm range. Preliminary fabrication with a lower steering rate design (8.12 deg/nm) partially validates the accuracy of the optimization workflow, accounting for fabrication errors.