Machine Learning-Assisted Device Modeling With Process Variations for Advanced Technology

Process variations (PV), including global variation (GV) and local (LV), have become one of the major issues in advanced technologies, which is crucial for circuit performance yield. However, developing a mature physics-based model challenging time-consuming. Thus, this work, we propose machine learning (ML) based method device modeling with PV implement corresponding simulation, demonstrated on Nanosheet FETs (NSFET). Verified by TCAD simulations, artificial neural network (ANN)-based ML algorithm enables to capture PV, e.g., dimension work function (WFV), high accuracy improved efficiency. For GV, ANN surrogated NSFET-based ring oscillator (RO) simulation results show that larger width (Wsh) or height (Hsh) leads higher RO frequency lower delay. LV, respective impacts grain size WF can be distinguished. The proposed workflow, from training generated Verilog-A model, fully automatic, promising shorten procedure accelerate development technologies.