Semiconductor Multilayer Nanometrology with Machine Learning

We review the measurement methods and thickness characterization algorithms of semiconductor multilayer devices.Today’s ultrahigh-density,high-energy-efficient three-dimensional semiconductor devices require an iterative semiconductor layer-stacking process.Accurate determination of nanometer-scale layer thickness is crucial for reliable semiconductor device fabrication.In this paper,we first review the commonly used semiconductor multilayer thickness measurement methods,including destructive and nondestructive measurement methods.Next,we review two approaches for thickness characterization:model-based algorithms using a physical interpretation of multilayer structures and a method using data-driven machine learning.With the growing importance of semiconductor multilayer devices,we anticipate that this study will help in selecting the most appropriate method for multilayer thickness characterization.