Smallest Stable Si/SiO_{2} Interface that Suppresses Quantum Tunneling from Machine-Learning-Based Global Search.

Area of Science:

• Materials Science
• Condensed Matter Physics
• Computational Science

Background:

• Downscaling field-effect transistors (FETs) is crucial for computational efficiency.
• Quantum tunneling at the silicon/silicon dioxide (Si/SiO 2 ) interface is a major challenge at the nanometer scale.
• Understanding and controlling Si/SiO 2 interface structures is vital for advanced semiconductor devices.

Purpose of the Study:

• To identify stable and high-performance Si/SiO 2 interface structures using a computational approach.
• To explore a wide range of potential interface configurations beyond traditional methods.
• To provide a foundation for fabricating improved semiconductor interfaces.

Main Methods:

• Development of a machine-learning-based global search algorithm.
• Systematic exploration of thousands of candidate Si/SiO 2 interface structures.
• Evaluation of interface properties including carrier mobility, carrier trapping, and interfacial energy.

Main Results:

• Identification of two high Miller index Si(210) and Si(211) interfaces with approximately 1 nm periodicity.
• These interfaces exhibit favorable carrier mobility and low carrier trapping.
• The discovered interfaces possess low interfacial energy, indicating stability.

Conclusions:

• The identified Si/SiO 2 interfaces offer a promising solution for overcoming quantum tunneling limitations in nanoscale FETs.
• These findings pave the way for fabricating stepped Si surfaces for next-generation transistors.
• Machine learning is effective for discovering optimal material interfaces in semiconductor research.