Silicon photonics enables compact integrated photonic devices with versatile functionalities and mass manufacturing capability. However, the optimization of high-performance free-form optical devices is still challenging due to the complex light-matter interaction involved that requires time-consuming electromagnetic simulations. This problem becomes even more prominent when multiple devices are required, typically requiring separate iterative optimizations. To facilitate multi-task inverse design, we propose a topology optimization method based on deep neural network (DNN) in low-dimensional Fourier domain. The DNN takes target optical responses as inputs and predicts low-frequency Fourier components, which are then utilized to reconstruct device geometries. Removing high-frequency components for reduced design degree-of-freedom (DOF) helps control minimal features and speed up training. For demonstration, the proposed method is utilized for wavelength filter design. The trained DNN can design multiple filters instantly and concurrently with high accuracy. Totally different targets can also be further optimized through transfer learning on existing network with greatly reduced optimization rounds. Our approach can be also adapted to other free-form photonic devices, including a waveguide-coupled single-photon source that we demonstrate to prove generalizability. Such DNN-assisted topology optimization significantly reduces the time and resources required for multi-task optimization, enabling large-scale photonic device design in various applications.

We propose a compact Si3N4-Si3N4-SOI grating coupler for generating optical vortex beams. With various mode input combinations, the multi-layer device can be tuned to emit beam with OAM states of ± 1 and ± 2.

We design, fabricate, and characterize a compact dual-mode waveguide crossing on a silicon-on-insulator platform. The dual-mode waveguide crossing with high performance is designed by utilizing the adjoint shape optimization. This adjoint-method-based optimization algorithm is computationally efficient and yields the optimal solution in fewer iterations compared with other iterative schemes. Our proposed dual-mode waveguide crossing exhibits low insertion loss and low crosstalk. Experimental results show that the insertion losses at the wavelength of 1550 nm are 0.83 dB and 0.50 dB for TE0 and TE1 modes, respectively. The crosstalk is less than −20 dB for the two modes over a wavelength range of 80 nm. The footprint of the whole structure is only 5 × 5 μm2.