Neural network-based inverse lithography technology (NNILT) has been used to improve the computational efficiency of large-scale mask optimization for advanced photolithography. NNILT is now mostly based on labels, and its performance is affected by the quality of labels. It is difficult for NNILT to achieve high performance and extrapolation ability for mask optimization without using labels. Here, we propose a label-free NNILT (LF-NNILT), which is implemented completely without labels and greatly improves the printability of the target layouts and the manufacturability of the synthesized masks compared to the traditional ILT. More importantly, the optimization speed of LF-NNILT is two orders of magnitude faster than the traditional ILT. Furthermore, LF-NNILT is simpler to implement and can achieve better solvers to support the development of advanced lithography.

A kinetic simulation code based on a reduced model is developed to study dynamic evolutions of a single toroidicity-induced shear Alfvén eigenmode driven by energetic particles. For zero background damping, it is found that the wave amplitude in nonlinear phase can either saturate for weak energetic particle drives or slowly increase for strong drives. This slow nonlinear growth in strong drive cases is found to be associated with broadening and overlapping of resonances between the wave and trapped particles. For the near-marginal-stability case with a large background damping, the mode nonlinear evolution exhibits strong upward and downward frequency chirping in multiple branches. A hole/clump formation is observed clearly in the corresponding evolution of energetic particle distribution.

Digital mask projection lithography (DMPL) technology is gaining significant attention due to its characteristics of free-mask, flexibility, and low cost. However, when dealing with target layouts featuring sizes smaller than the wavelength scale, accurately producing resist patterns that closely match the target layout using conventional methods to design the modulation coefficients of digital masks produced by spatial light modulators (SLM) becomes challenging. Here, we present digital inversion lithography technology (DILT), which offers what we believe to be a novel approach to reverse engineer the modulation coefficients of digital masks. In the case of binary amplitude modulation, DILT achieves a remarkable reduction in pattern errors (PE), reaching the original 0.26. At the same time, in the case of the gray amplitude modulation, the PE can be reduced to the original 0.05, which greatly improves the high-fidelity transfer of the target layout. This significant improvement enhances the accuracy of target design transfer. By leveraging the capabilities of DILT, DMPL can now attain higher precision and reliability, paving the way for more advanced applications in the field of micro-nano device manufacturing.