Suppressing Channel Percolation in Ferroelectric FET for Reliable Neuromorphic Applications

Kai Ni, Om Prakash, Simon Thomann, Zijian Zhao, Shan Deng, Hussam Amrouch

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

7 Scopus citations

Abstract

Channel percolation in ferroelectric FET (FeFET) due to random spatial distribution of switched domains reduces the number of accessible threshold voltage (VTH) states, thus posing a profound reliability challenge when it comes to its usage as synaptic weight cell in neuromorphic applications. However, channel percolation is not universally present in FeFET and only exists when the channel regions underneath the domains have negligible interaction, e.g., when the domain is large enough. In this work, we performed a comprehensive evaluation on the parameters that could impact neighbor interaction, and hence channel percolation. We reveal that: i) weak gate control with thick ferroelectric (FE) layer enhances neighbor interaction and suppresses channel percolation; ii) higher temperature can also reduce percolation by weakening the gate control. These insights provide an important guideline for to engineer more reliable FeFET analog states.

Original languageEnglish
Title of host publication2022 IEEE International Reliability Physics Symposium, IRPS 2022 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages9C21-9C28
ISBN (Electronic)9781665479509
DOIs
StatePublished - 2022
Externally publishedYes
Event2022 IEEE International Reliability Physics Symposium, IRPS 2022 - Dallas, United States
Duration: 27 Mar 202231 Mar 2022

Publication series

NameIEEE International Reliability Physics Symposium Proceedings
Volume2022-March
ISSN (Print)1541-7026

Conference

Conference2022 IEEE International Reliability Physics Symposium, IRPS 2022
Country/TerritoryUnited States
CityDallas
Period27/03/2231/03/22

Keywords

  • Analog synapse
  • FeFET
  • Ferroelectric
  • percolation

Fingerprint

Dive into the research topics of 'Suppressing Channel Percolation in Ferroelectric FET for Reliable Neuromorphic Applications'. Together they form a unique fingerprint.

Cite this