Using machine learning to narrow down the possibilities for a better quantum tunneling interface

Bob Yirka is a research scientist at Phys.org.

The researchers at Fudan University in China used machine learning to narrow the list of possible improved tunneling interface configurations. The results were published in a physical review letter.

Engineers have doubled the number of transistors that can be placed on an integrated circuit every two years in order to keep up with Moore's law. Due to the laws of physics, efforts like this are in jeopardy. The material used to separate gates on chips has become so thin that charge carriers can wiggle their way through. Stable configurations that would minimize such tunneling were sought by the researchers in this new effort.

The study looked at how tunneling is impacted by the structure of the interface. The degree of quantum tunnelling was influenced by the configuration of the interface. They used a machine learning application to study thousands of structures. 40 configurations appeared likely to provide a better option than the ones currently used. There were just 10 that were stable. Two of the 10 candidates were able to suppress tunnelling. The two configurations would allow for more transistors on a chip, which would allow for the creation of smaller devices.

The researchers want to see if other transistor materials are better suited for use in integrated circuits.

More information: Ye-Fei Li et al, Smallest Stable Si/SiO2 Interface that Suppresses Quantum Tunneling from Machine-Learning-Based Global Search, Physical Review Letters (2022). DOI: 10.1103/PhysRevLett.128.226102 Journal information: Physical Review Letters

There is a science network.