Topological materials for microelectronics

High-throughput search of topological materials for interconnects using first-principles transport calculations and machine-learning

The performance of semiconductor-based computing technologies is increasingly hindered by the resistivities of back-end-of-line (BEOL) interconnect materials as dimensions continue to scale down. Relative surface imperfections become more intense in narrow wires, and common interconnect materials such as copper exhibit increased resistivities at the nanoscale because of their sensitivity to surface scattering. Topological materials may address this critical BEOL interconnect resistance bottleneck through their surface states that are topologically protected against scattering from surfaces and other defects. However, the topological materials studied so far do not exhibit sufficient conductance to be competitive as interconnects. To address this challenge, we combine first-principles electron transport simulations with machine learning (ML) techniques to identify candidate topological materials that can outperform copper at sub-10 nm dimensions. Our approach predicts bulk conductance, surface conductance, and their sensitivity to defects from first-principles calculations. These predictions are integrated into a high-throughput ML-based active learning framework to systematically accelerate the discovery of topological materials for BEOL interconnects.