Ferroelectric HfZrO2 With Electrode Engineering and Stimulation Schemes as Symmetric Analog Synaptic Weight Element for Deep Neural Network Training

K. Y. Hsiang; C. Y. Liao; K. T. Chen; Y. Y. Lin; C. Y. Chueh; C. Chang; Y. J. Tseng; Y. J. Yang; S. T. Chang; M. H. Liao; T. H. Hou; C. H. Wu; C. C. Ho; J. P. Chiu; C. S. Chang; M. H. Lee

doi:10.1109/TED.2020.3017463

Ferroelectric HfZrO2 With Electrode Engineering and Stimulation Schemes as Symmetric Analog Synaptic Weight Element for Deep Neural Network Training

K. Y. Hsiang, C. Y. Liao, K. T. Chen, Y. Y. Lin, C. Y. Chueh, C. Chang, Y. J. Tseng, Y. J. Yang, S. T. Chang, M. H. Liao, T. H. Hou, C. H. Wu, C. C. Ho, J. P. Chiu, C. S. Chang, M. H. Lee^*

^*Corresponding author for this work

Institute of Electronics

Research output: Contribution to journal › Article › peer-review

24 Scopus citations

Abstract

Atomic layer deposition (ALD)-based TiN electrode on ferroelectric HfZrO2 metal/ferroelectric/metal (MFM) capacitor and ferroelectric field-effect transistor (FeFET) is demonstrated experimentally with weight transfer, that is, Delta P, per pulse analysis through consecutive alternating potentiation/depression (Pot./Dep.) training pulses. The weight training pulse schemes are studied to have symmetric and linear synapse weight transfer to increase the accuracy and accelerate the deep neural network (DNN) training. With ALD TiN inserted, alpha(p)/alpha(d) = -0.63/-0.84, asymmetry vertical bar alpha(p) - alpha(d)vertical bar = 0.21, and polarization modulation ratio (Pot./Dep.) = 97%/98% are achieved for MFM capacitor, and alpha(p)/alpha(d) = -1.32/-1.88, asymmetry vertical bar alpha(p) - alpha(d)vertical bar = 0.56, and G(max)/G(min) > 10x are delivered for FeFET.

Original language	English
Article number	9180313
Pages (from-to)	4201-4207
Number of pages	7
Journal	IEEE Transactions on Electron Devices
Volume	67
Issue number	10
DOIs	https://doi.org/10.1109/TED.2020.3017463
State	Published - Oct 2020

Keywords

Training
Synapses
Electrodes
Capacitors
Tin
Field effect transistors
Modulation
Ferroelectric memories
hafnium
neural networks

Access to Document

10.1109/TED.2020.3017463

Cite this

Hsiang, K. Y., Liao, C. Y., Chen, K. T., Lin, Y. Y., Chueh, C. Y., Chang, C., Tseng, Y. J., Yang, Y. J., Chang, S. T., Liao, M. H., Hou, T. H., Wu, C. H., Ho, C. C., Chiu, J. P., Chang, C. S., & Lee, M. H. (2020). Ferroelectric HfZrO2 With Electrode Engineering and Stimulation Schemes as Symmetric Analog Synaptic Weight Element for Deep Neural Network Training. IEEE Transactions on Electron Devices, 67(10), 4201-4207. [9180313]. https://doi.org/10.1109/TED.2020.3017463

@article{46083a897c794a01b8da5685d182a3c7,

title = "Ferroelectric HfZrO2 With Electrode Engineering and Stimulation Schemes as Symmetric Analog Synaptic Weight Element for Deep Neural Network Training",

abstract = "Atomic layer deposition (ALD)-based TiN electrode on ferroelectric HfZrO2 metal/ferroelectric/metal (MFM) capacitor and ferroelectric field-effect transistor (FeFET) is demonstrated experimentally with weight transfer, that is, Delta P, per pulse analysis through consecutive alternating potentiation/depression (Pot./Dep.) training pulses. The weight training pulse schemes are studied to have symmetric and linear synapse weight transfer to increase the accuracy and accelerate the deep neural network (DNN) training. With ALD TiN inserted, alpha(p)/alpha(d) = -0.63/-0.84, asymmetry vertical bar alpha(p) - alpha(d)vertical bar = 0.21, and polarization modulation ratio (Pot./Dep.) = 97%/98% are achieved for MFM capacitor, and alpha(p)/alpha(d) = -1.32/-1.88, asymmetry vertical bar alpha(p) - alpha(d)vertical bar = 0.56, and G(max)/G(min) > 10x are delivered for FeFET.",

keywords = "Training, Synapses, Electrodes, Capacitors, Tin, Field effect transistors, Modulation, Ferroelectric memories, hafnium, neural networks",

author = "Hsiang, {K. Y.} and Liao, {C. Y.} and Chen, {K. T.} and Lin, {Y. Y.} and Chueh, {C. Y.} and C. Chang and Tseng, {Y. J.} and Yang, {Y. J.} and Chang, {S. T.} and Liao, {M. H.} and Hou, {T. H.} and Wu, {C. H.} and Ho, {C. C.} and Chiu, {J. P.} and Chang, {C. S.} and Lee, {M. H.}",

year = "2020",

month = oct,

doi = "10.1109/TED.2020.3017463",

language = "English",

volume = "67",

pages = "4201--4207",

journal = "Ieee Transactions On Electron Devices",

issn = "0018-9383",

publisher = "IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC",

number = "10",

}

Hsiang, KY, Liao, CY, Chen, KT, Lin, YY, Chueh, CY, Chang, C, Tseng, YJ, Yang, YJ, Chang, ST, Liao, MH, Hou, TH, Wu, CH, Ho, CC, Chiu, JP, Chang, CS & Lee, MH 2020, 'Ferroelectric HfZrO2 With Electrode Engineering and Stimulation Schemes as Symmetric Analog Synaptic Weight Element for Deep Neural Network Training', IEEE Transactions on Electron Devices, vol. 67, no. 10, 9180313, pp. 4201-4207. https://doi.org/10.1109/TED.2020.3017463

TY - JOUR

T1 - Ferroelectric HfZrO2 With Electrode Engineering and Stimulation Schemes as Symmetric Analog Synaptic Weight Element for Deep Neural Network Training

AU - Hsiang, K. Y.

AU - Liao, C. Y.

AU - Chen, K. T.

AU - Lin, Y. Y.

AU - Chueh, C. Y.

AU - Chang, C.

AU - Tseng, Y. J.

AU - Yang, Y. J.

AU - Chang, S. T.

AU - Liao, M. H.

AU - Hou, T. H.

AU - Wu, C. H.

AU - Ho, C. C.

AU - Chiu, J. P.

AU - Chang, C. S.

AU - Lee, M. H.

PY - 2020/10

Y1 - 2020/10

N2 - Atomic layer deposition (ALD)-based TiN electrode on ferroelectric HfZrO2 metal/ferroelectric/metal (MFM) capacitor and ferroelectric field-effect transistor (FeFET) is demonstrated experimentally with weight transfer, that is, Delta P, per pulse analysis through consecutive alternating potentiation/depression (Pot./Dep.) training pulses. The weight training pulse schemes are studied to have symmetric and linear synapse weight transfer to increase the accuracy and accelerate the deep neural network (DNN) training. With ALD TiN inserted, alpha(p)/alpha(d) = -0.63/-0.84, asymmetry vertical bar alpha(p) - alpha(d)vertical bar = 0.21, and polarization modulation ratio (Pot./Dep.) = 97%/98% are achieved for MFM capacitor, and alpha(p)/alpha(d) = -1.32/-1.88, asymmetry vertical bar alpha(p) - alpha(d)vertical bar = 0.56, and G(max)/G(min) > 10x are delivered for FeFET.

AB - Atomic layer deposition (ALD)-based TiN electrode on ferroelectric HfZrO2 metal/ferroelectric/metal (MFM) capacitor and ferroelectric field-effect transistor (FeFET) is demonstrated experimentally with weight transfer, that is, Delta P, per pulse analysis through consecutive alternating potentiation/depression (Pot./Dep.) training pulses. The weight training pulse schemes are studied to have symmetric and linear synapse weight transfer to increase the accuracy and accelerate the deep neural network (DNN) training. With ALD TiN inserted, alpha(p)/alpha(d) = -0.63/-0.84, asymmetry vertical bar alpha(p) - alpha(d)vertical bar = 0.21, and polarization modulation ratio (Pot./Dep.) = 97%/98% are achieved for MFM capacitor, and alpha(p)/alpha(d) = -1.32/-1.88, asymmetry vertical bar alpha(p) - alpha(d)vertical bar = 0.56, and G(max)/G(min) > 10x are delivered for FeFET.

KW - Training

KW - Synapses

KW - Electrodes

KW - Capacitors

KW - Tin

KW - Field effect transistors

KW - Modulation

KW - Ferroelectric memories

KW - hafnium

KW - neural networks

UR - http://www.scopus.com/inward/record.url?scp=85092097938&partnerID=8YFLogxK

U2 - 10.1109/TED.2020.3017463

DO - 10.1109/TED.2020.3017463

M3 - Article

SN - 0018-9383

VL - 67

SP - 4201

EP - 4207

JO - Ieee Transactions On Electron Devices

JF - Ieee Transactions On Electron Devices

IS - 10

M1 - 9180313

ER -

Ferroelectric HfZrO2 With Electrode Engineering and Stimulation Schemes as Symmetric Analog Synaptic Weight Element for Deep Neural Network Training

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this