Physical insight and Monte Carlo statistical analysis of Work-Function Variability in FinFETs based on 2D slice composition

Metal gate technology is a key enabler for continued transistor scaling and is essential to proper operation of emerging multiple-gate architectures. Work-function variability (WFV) associated to the granular structure of TiN gates in a LSTP-32nm FinFET is studied through quantum-corrected hydrodynamic simulations. These provide physical insight on the impact of grain size and position, highlighting the crucial role of the source gate side and showing theoretical as well as practical shortcomings of previous WFV models based of the area-weighted average WF concept. 2D slices with random gate configuration are paralleled to simulate the impact of gate non-uniformities along the fin height. Electrical parameter distributions are extracted from large ensembles of composite devices and compared to those produced by line-edge roughness (LER). The different role of WFV and LER in threshold voltage and drive current fluctuations and the peculiar statistical features of the former contribution are clarified