Solid State Technology and Devices Seminar: 2D material-based layer transfer based on remote epitaxy and uniform epitaxial RAM towards large-scale neuromorphic arrays
Seminar | September 22 | 1-2 p.m. | Cory Hall, 521 Cory (Hogan Room)
Professor Jeehwan Kim, Massachusetts Institute of Technology
In this talk, I will discuss our recent development in a 2D material-based layer-transfer (2DLT) process based on remote epitaxy. We recently discovered that the epitaxial registry of adatoms during epitaxy can be assigned by the underlying substrate remotely through 2D materials. Our study shows that remote epitaxy can be performed through a single-atom-thick gap defined by monolayer graphene at the substrate-epilayer interface. We demonstrated successful remote homoepitaxy of (001) GaAs on (001) GaAs substrates through monolayer graphene. The concept is extended for remote epitaxy of other materials such as GaN, InP, GaP, and SrTiO3. The grown single-crystalline films are then rapidly released from the vdW surface of graphene. This concept suggests a method to copy/paste any type of single-crystalline films from the underlying substrates through 2D materials then rapidly released and transferred to the substrates of interest. With the potential to reuse graphene-coated substrates, 2DLT could advance non-Si electronics/photonics by allowing savings on the high cost of non-Si substrates.
I will also introduce our new research activities in developing advanced resistive switching devices that can allow large-scale synaptic arrays. Conventional memristors typically utilize a defective amorphous solid as a switching medium for defect-mediated formation of conducting filaments. However, the imperfection of the switching medium also causes stochastic filament formation leading to spatial and temporal variation of the devices. In this talk, I will introduce an epitaxial random access memory (epiRAM), where we precisely confined the conducting paths in the single-crystalline films resulting in unprecedented device performances. MITs epiRAM exhibits extremely low temporal/spatial variation, linear synaptic weight update, high on/off ratio (250 for analog/10,000 for digital), great endurance (>109), long retention time (2 days at 85oC), and self-selectivity. This suggests the potential to fabricate a memristor-based large-scale neural network hardware.