Artificial Synaptic Properties in Oxygen-Based Electrochemical Random-Access Memory with CeO₂ Nanoparticle Assembly...

Area of Science:

• Materials Science
• Neuroscience
• Computer Engineering

Background:

• Neuromorphic computing offers energy-efficient solutions beyond traditional architectures.
• Memtransistor devices are promising for artificial synapses due to controllable weight updates and disturbance immunity.

Purpose of the Study:

• To investigate an oxygen ion exchange-based electrochemical random-access memory memtransistor as an artificial synapse for neuromorphic computing.
• To explore the use of ZnO and CeO2 nanoparticle assembly for tunable conductance and synaptic functions.

Main Methods:

• Fabrication of a memtransistor using a ZnO channel and a CeO2 nanoparticle assembly as a gate insulator and ion exchange layer.
• Characterization of conductance modulation via oxygen ion exchange triggered by gate voltage.
• Evaluation of device performance, including linearity, symmetry, endurance, retention, and emulation of biological synapse functions.

Main Results:

• The memtransistor exhibits tunable and reversible conductance changes through oxygen ion exchange between ZnO and CeO2 NPs.
• CeO2 facilitates efficient oxygen ion exchange due to its oxygen absorption/release properties and porous structure.
• The device achieves high-speed operation (tens of microseconds), high linearity, good endurance (>10^4 pulses), and excellent nonvolatile retention.
• The memtransistor successfully mimics biological synapse functions like paired-pulse facilitation, short-term plasticity (STP), and long-term plasticity (LTP).

Conclusions:

• The proposed memtransistor, leveraging CeO2's properties, is a viable artificial synapse for neuromorphic computing.
• The device's ability to emulate synaptic plasticity and its operational characteristics make it suitable for advanced AI hardware.
• This research contributes to the development of energy-efficient and high-performance computing systems for data-centric applications.