Collaborative Research: 2D Ambipolar Machine Learning & Logical Computing Systems

Emerging materials have novel behaviors that create new opportunities for information processing, especially if the natural behavior of the new materials can be leveraged, rather than trying to engineer them to behave like today’s electronics. In particular, atomically thin materials, also known as two-dimensional (2D) materials, can be naturally ambipolar, i.e. can conduct both electrons and holes. 2D materials also have additional properties, such as hysteresis and variability, that are usually suppressed when used for microelectronics.

This interdisciplinary team aims to exploit these unique behaviors of 2D material-based field-effect transistors, and to show they can lead to significant improvements in area, speed, and energy-efficiency for both logical and machine learning (ML) applications. To do so, the team will develop 2D ambipolar materials and device structures, design ambipolar logic families and ML architectures, and experimentally demonstrate small-scale and medium-scale circuit prototypes that prove the utility of these materials as building blocks for the next generation of computing systems.

This research will lead to a deeper understanding of the control of ambipolar devices, as well as innovations in circuit and system design, that could be used in additional new implementations of nanotechnology. The project will generate energy-efficient technologies to accomplish computing tasks relevant to many sectors of society, including cybersecurity, education, healthcare, and internet-of-things.

The three PIs have designed a concerted plan to integrate the outcomes of the proposed work in education and outreach, including interdisciplinary training of students for a strong and bright workforce; designing lectures to expose a larger population of students to the subjects; and designing hands-on activities to reach K-12 girls to increase diversity in science, technology, engineering, and math.

Emerging nanotechnologies exhibit novel switching phenomena that create new opportunities for information processing. In particular, two-dimensional (2D) dual-gate ambipolar field-effect transistors (DG-A-FETs) can be switched between n-type and p-type FET operation, permitting compact computing circuits that leverage the functionality of each individual FET.

This project aims to exploit the unique behaviors of DG-A-FETs in logical and machine learning (ML) circuits and systems that provide significant improvements in area, speed, and energy-efficiency. In addition to the dual-gate functionality, the team aims to control and exploit behaviors of emerging 2D materials such as hysteresis and variability to benefit computing.

This project will study 2D ambipolar materials and device structures to design several ambipolar logic families and ML architectures, and it will experimentally demonstrate small-scale and medium-scale circuit prototypes that prove the utility of DG-A-FETs as a building block for the next generation of computing systems. This research will lead to a deeper understanding of the dual-gate control of ambipolar devices, as well as innovations in circuit and system design, that could be used in additional new implementations of nanotechnology.

The project will generate energy-efficient technologies to accomplish computing tasks relevant to many sectors of society, including cybersecurity, education, healthcare, and internet-of-things. The three PIs have designed a concerted plan to integrate the outcomes of the proposed work in education and outreach, including interdisciplinary training of students for a strong and bright workforce; designing lectures to expose a larger population of students to the subjects; and designing hands-on activities to reach K-12 girls at STEM days on multiple of the PIs’ campuses.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.