TY - JOUR
T1 - In-Memory Computing of Multilevel Ferroelectric Domain Wall Diodes at LiNbO3 Interfaces
AU - Sun, Jie
AU - Li, Yiming
AU - Ou, Yangjun
AU - Huang, Qianwei
AU - Liao, Xiaozhou
AU - Chen, Zibin
AU - Chai, Xiaojie
AU - Zhuang, Xiao
AU - Zhang, Wendi
AU - Wang, Chao
AU - Jiang, Jun
AU - Jiang, Anquan
N1 - Funding Information:
The authors would like to thank David MacDonald, MSc, from Liwen Bianji, Edanz Editing China ( www.liwenbianji.cn/ac ), for editing the English text of a draft of this manuscript. This work was supported by the National Key Basic Research Program of China (Grant 2019YFA0308500), the National Natural Science Foundation of China (grant numbers 61904034 and 62174034), and the young scientist project of MOE innovation platform and the Australian Research Council (project ID: DP190101155). The authors acknowledge the facilities and the scientific and technical assistance of the Microscopy Australia node at the University of Sydney (Sydney Microscopy & Microanalysis).
Publisher Copyright:
© 2022 Wiley-VCH GmbH.
PY - 2022/12/2
Y1 - 2022/12/2
N2 - Direct data processing in nonvolatile memories can enable area- and energy-efficient computation, unlike independent performance between separate processing and memory units; repetitive data transfer between these units represents a fundamental performance limitation in modern computers. Spatially mobile conducting domain walls in ferroelectrics can be redirected between drain, gate, and source electrodes to function as nonvolatile transistors with superior energy efficiency, ultrafast operating and communication speeds, and high logic/storage densities. Here, in-memory computing is demonstrated using multilevel domain wall diodes at LiNbO3 interfaces. Ultrathin domains within interfacial layers between each mesa-like memory cell and the contact electrodes can rectify diode-like domain wall currents with on/off current ratios exceeding 107 at low operating voltages, surpassing the performance of traditional p-n junctions using built-in potentials across depletion layers. NOT, NAND, and NOR gate logic functions are demonstrated, providing insights into high-density integration of multilevel storage and computational units in all-ferroelectric domain wall devices.
AB - Direct data processing in nonvolatile memories can enable area- and energy-efficient computation, unlike independent performance between separate processing and memory units; repetitive data transfer between these units represents a fundamental performance limitation in modern computers. Spatially mobile conducting domain walls in ferroelectrics can be redirected between drain, gate, and source electrodes to function as nonvolatile transistors with superior energy efficiency, ultrafast operating and communication speeds, and high logic/storage densities. Here, in-memory computing is demonstrated using multilevel domain wall diodes at LiNbO3 interfaces. Ultrathin domains within interfacial layers between each mesa-like memory cell and the contact electrodes can rectify diode-like domain wall currents with on/off current ratios exceeding 107 at low operating voltages, surpassing the performance of traditional p-n junctions using built-in potentials across depletion layers. NOT, NAND, and NOR gate logic functions are demonstrated, providing insights into high-density integration of multilevel storage and computational units in all-ferroelectric domain wall devices.
KW - diodes
KW - ferroelectric domain wall
KW - in-memory computing
KW - interfacial layers
UR - http://www.scopus.com/inward/record.url?scp=85139069280&partnerID=8YFLogxK
U2 - 10.1002/adfm.202207418
DO - 10.1002/adfm.202207418
M3 - Journal article
AN - SCOPUS:85139069280
SN - 1616-301X
VL - 32
JO - Advanced Functional Materials
JF - Advanced Functional Materials
IS - 49
M1 - 2207418
ER -