Global optimisation of source and mask in inverse lithography via tabu search combined with genetic algorithm

Abstract

Source mask optimisation (SMO) is a resolution enhancement technology that is utilised in the advanced process node of optical lithography to achieve acceptable imaging quality and fidelity. It is crucial in enhancing the convergence performance and optimisation capability of pixel-based SMO. In this study, an SMO approach that employs a genetic algorithm (GA), combined with the tabu search method (TS), is proposed. GA-TS, a hybrid-type global optimisation algorithm, has an outperforming capacity to avoid local optima owing to the excellent local searching function of TS. Furthermore, an edge-optimisation strategy was implemented to optimise the mask for a low-complexity mask layout. The simulation results confirm that the proposed approach exhibits exceptional optimisation capability and convergence performance.

DMD maskless lithography optimization based on an improved genetic algorithm

“…[26][27][28] As a search heuristic inspired by natural evolution, the GA has demonstrated remarkable prowess in tackling intricate optimization conundrums, a prime example being mask optimization in lithography. [29][30][31] The intrinsic capability of the GA to traverse expansive solution domains underscores its potential. The introduction of heuristic algorithms, such as the GA, into the mask optimization realm marked a paradigm shift.…”

Photonic crystal broadband 1×N beam splitter with designable splitting ratio based on genetic algorithm

“…Then, based on the 1×2 structure, the broadband 1×4 and 1×8 beam splitters with designable splitting ratios are proposed. According to the target broadband splitting performance, using genetic algorithm (GA) [22], the 1×4 and 1×8 beam splitters with broad operating bandwidth, high transmittance and different splitting ratios can be inversely designed by optimizing the corresponding structural parameters. Finally, the performance of the 1×4 and 1×8 beam splitters designed by GA is analyzed, focusing on the bandwidth, transmittance, splitting ratio and response time.…”

Photonic crystal broadband y-shaped 1 × 2 beam splitter inversely designed by genetic algorithm

“…By using the plane wave expansion method (PWM) 19 and the finite difference time domain method (FDTD), 20 the influence of the z-axis offset of the dielectric rods adjacent to the input and output waveguides, and the offset along the V-shaped oblique waveguide of the rods at the junction area on the broadband performance of the 1×2 beam splitter is studied. In order to improve the optimization efficiency and achieve excellent splitting performance, the genetic algorithm (GA) 21 is applied to inversely design the 1×2 beam splitter. The performance of the 1×2 beam splitter designed by GA is analyzed, focusing on the transmittance, additional loss, and uniformity in the wavelength range of 1500 to 1600 nm.…”