Time-Domain Simulation and Optimization of the Memory Window for HZO-Based FeFETs Using the NLS Model

Area of Science:

• Materials Science
• Electrical Engineering
• Computational Physics

Background:

• Hafnium-zirconium oxide (HZO)-based ferroelectric field-effect transistors (FeFETs) are crucial for next-generation embedded memory and compute-in-memory systems.
• Current TCAD simulations are computationally intensive and do not fully capture dynamic ferroelectric domain switching, limiting device development.

Purpose of the Study:

• To develop a computationally efficient time-domain simulation framework for HZO FeFETs.
• To enable self-consistent simulation of polarization and electrical characteristics, including device variability and process optimization.
• To provide model-based support for the design of advanced, ultra-low-power FeFETs.

Main Methods:

• A time-domain simulation framework coupling the nucleation-limited switching (NLS) model with MOSFET surface potential for self-consistent solutions.
• Monte Carlo methods to simulate device variability and Shmoo plots for process window identification.
• An integrated solution for 22 nm FDSOI devices, including geometric scaling and modified Landau-Khalatnikov (L-K) dynamics for ultrathin ferroelectrics.

Main Results:

• The framework enables end-to-end simulation from material polarization to device electrical characteristics and performance optimization.
• Successfully addressed geometric scaling, ultrathin ferroelectric layer dynamics, and short-channel effects in 22 nm FDSOI devices.
• Demonstrated potential for analytical and design support in developing advanced FeFETs.

Conclusions:

• The developed time-domain simulation framework offers a significant advancement over traditional TCAD for HZO FeFETs.
• This approach facilitates efficient exploration of design spaces and optimization of programming/erasure windows.
• The study provides a robust tool for accelerating the development of ultra-low-power FeFETs for memory and computing applications.