3D Mask Simulation and Lithographic Imaging using Physics-Informed Neural Networks

The increasing demands on computational lithography and imaging in the design and optimization of lithography processes necessitate rigorous modeling of EUV light diffracted from the mask. Traditional EMF solvers are inefficient for large-scale technology problems, while deep neural networks rely on a huge amount of expensive rigorously simulated or measured data. To overcome these constraints, we explore the potential of physics-informed neural networks (PINN) as a promising solution for addressing complex optical problems in EUV lithography and accurate modeling of light diffraction from typical reflective EUV masks, which include an absorber on top of a multilayer. Lithography simulations are done for 3D masks with line-space patterns and contact holes. Based on the obtained results, the perspectives of an accelerated PINN-based EMF solver as an alternative solution to traditional methods are discussed. The capabilities of the established PINNs approach to simulate typical 3D mask effects including non-telecentricities, shifts of the best focus position, and image blur are demonstrated. The coupling of the predicted diffraction spectrum with image simulations enables the evaluation of PINN performance in terms of relevant lithographic metrics. The results of modeling near- and far-field diffraction using PINN showcase a good performance in terms of convergence behavior, stability, and accuracy. Process windows predicted by PINN completely overlap the ones rigorously simulated by the numerical solver. The outcomes of our study demonstrate a real benefit of PINN: differently from numerical solvers, once trained, generalized PINN can simulate light scattering in several milliseconds without re-training and independently of problem complexity. As a result, PINN demonstrates a significant speed up (up to x10000) compared to the Waveguide simulation of the same instance.