Thermal, mechanical, and electrical properties of Si-stacked nanosheet transistors using machine learning interatomic...

Area of Science:

• Materials Science
• Nanotechnology
• Computational Physics

Background:

• Optimizing nanoelectronic device performance relies heavily on understanding thermal and mechanical properties.
• Silicon nanosheets are critical components in advanced nanoelectronic architectures.

Purpose of the Study:

• To determine the lattice thermal conductivity and elastic constants of silicon nanosheets at varying thicknesses.
• To evaluate the technological implications of these properties on nanosheet field-effect transistor (NS-FET) performance.

Main Methods:

• Utilized machine learning interatomic potentials (MLIPs), specifically the Moment Tensor Potential (MTP) method.
• Trained MLIPs using density functional theory (DFT) calculations and stochastic sampling of the potential energy surface (PES).
• Employed TCAD device simulations to assess the impact on NS-FETs.

Main Results:

• Lattice thermal conductivity dropped to approximately 7% of bulk silicon value for sheets thinner than 6 nm.
• Certain stiffness tensor components decreased to about 3% of bulk values at sub-6 nm thickness.
• Significant reductions in thermal and mechanical properties were observed with decreasing nanosheet thickness.

Conclusions:

• The study provides critical insights into heat transport and mechanical behavior in ultrathin silicon nanosheets.
• Findings are essential for the design and optimization of next-generation nanoelectronic devices.
• The extracted parameters offer valuable data for advancing NS-FET technology at future nodes.