Language

Synaptic behavior in analog memristors based on green-synthesized ZnO nanoparticles

Congratulations to our colleague MS Student. Pham Phu Quan, for his recent publication entitled in the journal " Ceramics International", which was a collaboration with our colleagues in the lab of Dr. Nguyen Thi My Lan at the Faculty of Biology and Biotechnology of the University of Science, VNU-HCM.

A new approach is being developed in neuromorphic computing, presenting a promising method that tightly integrates memory and computing units, known as in-memory computing. The crucial element in in-memory computing architectures is a neuromorphic electronic synapse employed to replicate the functionality of biological synapses. It adapts synaptic weights and stores them as dispersed memory throughout the neural network. A two-terminal memristor, in which the device resistance can be gradually adjusted by controlling charge, has been widely suggested as an artificial synapse for achieving storage and computational functions. The memristor is comprised of semiconductor layers positioned between two electrodes, and resistive switching (RS) can occur due to either electroforming (conductive filaments), charge migrations, or phase transition. For electronic synapse devices to function effectively in neural networks, they must be capable of achieving multiple implemented states. Semiconductor devices lost ground to emerging computational models for tasks requiring more than two binary states. Historically, CMOS circuits have simulated ANNs using dozens of transistors, consuming significant chip area and energy; hence, this approach was not feasible for large-scale integration. Memristor devices can exhibit gradual set/reset switching behaviour, known as analogue RS behaviour.

In this work, by utilizing avocado seed extract instead of hazardous chemicals, we synthesized ZnO NPs that are non-toxic, environmentally friendly, and cost-effective for large-scale production. Then, we designed and fabricated a self-rectifying analogue memristor employing a Cr/ZnO NPs/FTO junction structure that effectively prevents sneak current. We thoroughly examined and emulated the memristor's switching characteristics and synaptic behaviour through statistical analysis.


For more details, please visit: 

https://doi.org/10.1016/j.ceramint.2024.05.154

See the supplementary material for a detailed green synthesis procedure.


Funding

This work was supported by the Vingroup Innovative Fund under grant number VINIF.2023. DA130. 

The authors express their sincere thanks to the crew of the Center for INOMAR, VNU-HCM for their continuous support and help in using the XRD and FTIR facilities.

Bình luận

Previous Post Next Post