11/30 Seminar Speech

User Rating:  / 0
PoorBest 

SpeakerProfessor PI-HO HU

OrganizationDepartment of Electrical Engineering, National Taiwan University  

TopicFerroelectric FET Non-volatile Memory

 Date10:20 am 2020.11.30

LocationRoom 203, College of Engineering

Abstract;Artificial intelligence (AI) and Internet-of-Things (IoT) play a vital role in the future. For the von Neumann architecture, communication bottleneck between memory and processors becomes one of the most series problems due to low data throughput and high-power consumption. Non-volatile memory (NVM) based synapse and logic-in-memory circuit have been proposed to alleviate these problems. Ferroelectric Field-Effect-Transistor (FeFET) has been explored actively for emerging non-volatile memory and neuromorphic computing due to its high compatibility with CMOS platform. In this talk, I will introduce the operation principles and the challenges of FeFET memory. We have proposed a novel split-gate FeFET (SG-FeFET) for non-volatile memory and neuromorphic applications, which significantly improves the energy efficiency and reduces the power consumption. The reliability and endurance of SG-FeFET memory have also been improved owing to the lowered write voltage. SG-FeFET as a synaptic device has also been investigated including the non-linearity, asymmetry, and dynamic range of conductance updates.