Electronic Synaptic Devices with High Thermostability Induced by Embedded Tungsten Disulfide Quantum Dots for Machine Learning

Abstract

Abstract If the speed of machine learning is to be improved, devices and systems with strong resistances to various types of internal noise, mainly internal thermal noise, are urgently needed. The successful demonstration of a synaptic device is reported based on a polyimide–tungsten disulfide quantum dot (PI‐WS2 QD) nanocomposite that continued to operate normally after simultaneous exposure of a high temperature of 100 °C. Such excellent performance is attributable to the strong quantum confinement effect of WS2 QDs. The working current of the device and its power consumption are on the orders of nanoamperes and femtojoules, respectively. Undoubtedly, such devices will significantly improve the physical performances of machine learning systems and allow the rapid realization of greatly improving machine learning speed.

Keywords

• high thermal stability
• quantum confinement effect
• synaptic devices
• tunneling current
• WS 2 quantum dots