Wafer‐Scale 2D Hafnium Diselenide Based Memristor Crossbar Array for Energy‐Efficient Neural Network Hardware

Memristor crossbar with programmable conductance could overcome the energy consumption and speed limitations of neural networks when executing core computing tasks in image processing. However, the implementation of crossbar array (CBA) based on ultrathin 2D materials is hindered by challenges assoc...

Full description

Saved in:
Bibliographic Details
Published in:Advanced materials (Weinheim) 2022-06, Vol.34 (25), p.e2103376-n/a
Main Authors: Li, Sifan, Pam, Mei‐Er, Li, Yesheng, Chen, Li, Chien, Yu‐Chieh, Fong, Xuanyao, Chi, Dongzhi, Ang, Kah‐Wee
Format: Article
Language:English
Subjects:
2D hafnium diselenide
Arrays
Artificial neural networks
Energy consumption
Epitaxial growth
Hafnium
Hardware
Image processing
Materials science
Memristors
Microprocessors
Molecular beam epitaxy
multiply‐and‐accumulate operation
neural network hardware
Neural networks
Power efficiency
Switching
Two dimensional materials
van der Waals transfer
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Memristor crossbar with programmable conductance could overcome the energy consumption and speed limitations of neural networks when executing core computing tasks in image processing. However, the implementation of crossbar array (CBA) based on ultrathin 2D materials is hindered by challenges associated with large‐scale material synthesis and device integration. Here, a memristor CBA is demonstrated using wafer‐scale (2‐inch) polycrystalline hafnium diselenide (HfSe2) grown by molecular beam epitaxy, and a metal‐assisted van der Waals transfer technique. The memristor exhibits small switching voltage (0.6 V), low switching energy (0.82 pJ), and simultaneously achieves emulation of synaptic weight plasticity. Furthermore, the CBA enables artificial neural network with a high recognition accuracy of 93.34%. Hardware multiply‐and‐accumulate (MAC) operation with a narrow error distribution of 0.29% is also demonstrated, and a high power efficiency of greater than 8‐trillion operations per second per Watt is achieved. Based on the MAC results, hardware convolution image processing can be performed using programmable kernels (i.e., soft, horizontal, and vertical edge enhancement), which constitutes a vital function for neural network hardware. A 2D hafnium diselenide (HfSe2) memristor crossbar array (CBA) is demonstrated via wafer‐scale molecular beam epitaxy growth and metal‐assisted van der Waals transfer techniques. The CBA enables artificial neural network with high recognition accuracy of 93.34%, and achieves hardware convolution image processing using energy‐efficient multiply‐and accumulate operations.
ISSN:0935-9648
1521-4095
DOI:10.1002/adma.202103376