The proposed interdisciplinary research combines advances in two dimensional (2D) materials, brain-inspired network computing, and resonant tunnel diodes (RTDs), to create a new device concept: 2D resonant tunnel field-effect-transistor (RTFET). RTFETs based on 2D crystals will solve the lattice matching/dislocation issues of III-V semiconductor based RTDs. The strong negative differential resistance and high frequency response in these devices will be used to perform "comparison" functions, similar to synapses in neurons, for image recognition applications. RTFET-based synapses can potentially reduce energy consumption by more than ten times compared to the circuits based on traditional complementary metal-oxide-semiconductor (CMOS). The ultra-high operating frequencies of RTFETs will enable their applications in high speed wireless communication, THz imaging, and spectroscopy. In addition, RTFETs can also serve as sensitive vehicles to probe 2D/2D interfaces, providing scientific insights on the out-of-plane transport in 2D heterostructures. This research project will not only advance the knowledge of quantum tunneling in 2D crystals, resonant tunneling devices, and brain-like circuits, but also have direct technology impact on image recognition, non-traditional computing, and high frequency wireless communications. The integration of the proposed research with after-school programs in elementary schools, new courses for undergraduate/graduate students, and recruiting/retaining women students, will foster students' interest in nanotechnology while broadening their knowledge base, thus having a positive enduring impact on the education of a world-class and diverse science and technology workforce.
The objective of this project is to understand the interlayer resonant tunneling in 2D vertical heterostructures and demonstrate high speed RTFETs with pronounced negative differential resistance at room temperatures for neurosynaptic image recognition applications. The PI will pursue the following four thrusts: (1) fabrication of a variety of 2D vertical heterostructures, including black phosphorus/boron nitride/black phosphorus, tungsten diselenide/molybdenum disulfide/tungsten diselenide heterostructures, with precise control of rotation angles and layer numbers; (2) evaluate the impact of rotation angle, band offsets, tunneling barrier thickness, and interface qualities, on the resonant tunneling currents in the vertical 2D heterostructures; (3) demonstrate high speed RTFETs with pronounced negative differential resistance at room temperatures; (4) demonstrate synapses based on 2D RTFETs, as basic elements in neurosynaptic chips for image recognition applications. The successful execution of this project will expand the knowledge of resonant quantum tunneling in 2D heterostructures, resonant tunneling devices, and neural-network computing. More specifically, this research will elucidate the effect of rotation angle alignment, band extrema location, bandgap, and band offsets on the interlayer resonant tunneling in RTFETs. This will also provide insight into inelastic tunneling and scattering due to interface traps and defects in RTFETs. The fundamental understanding of the resonant tunneling in 2D heterostructures gained in this research project will enable a new class of novel nanoscale electronic devices based on out-of-plane transport instead of the traditional in-plane transport of 2D crystals. This research will also pave the way for new functional devices beyond CMOS and provide an experimental demonstration of neural-network circuits based on RTFETs. The negative differential resistance and fast response in RTFETs can potentially enable a new computing paradigm beyond the traditional von Neumann architecture.
