Revealing ferroelectric switching character using deep recurrent neural networks

Joshua C. Agar, Brett Naul, Shishir Pandya, Stefan van der Walt, Joshua Maher, Yao Ren, Long Qing Chen, Sergei V. Kalinin, Rama K. Vasudevan, Ye Cao, Joshua S. Bloom, Lane W. Martin

Research output: Contribution to journalArticle

Abstract

The ability to manipulate domains underpins function in applications of ferroelectrics. While there have been demonstrations of controlled nanoscale manipulation of domain structures to drive emergent properties, such approaches lack an internal feedback loop required for automatic manipulation. Here, using a deep sequence-to-sequence autoencoder we automate the extraction of latent features of nanoscale ferroelectric switching from piezoresponse force spectroscopy of tensile-strained PbZr0.2Ti0.8O3 with a hierarchical domain structure. We identify characteristic behavior in the piezoresponse and cantilever resonance hysteresis loops, which allows for the classification and quantification of nanoscale-switching mechanisms. Specifically, we identify elastic hardening events which are associated with the nucleation and growth of charged domain walls. This work demonstrates the efficacy of unsupervised neural networks in learning features of a material’s physical response from nanoscale multichannel hyperspectral imagery and provides new capabilities in leveraging in operando spectroscopies that could enable the automated manipulation of nanoscale structures in materials.

Original languageEnglish (US)
Article number4809
JournalNature communications
Volume10
Issue number1
DOIs
StatePublished - Dec 1 2019

Fingerprint

Recurrent neural networks
Ferroelectric materials
manipulators
Spectrum Analysis
Spectroscopy
Aptitude
Imagery (Psychotherapy)
Domain walls
Hysteresis loops
Hardening
Nucleation
Demonstrations
Learning
Neural networks
Feedback
imagery
hardening
spectroscopy
learning
domain wall

All Science Journal Classification (ASJC) codes

  • Chemistry(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Physics and Astronomy(all)

Cite this

Agar, J. C., Naul, B., Pandya, S., van der Walt, S., Maher, J., Ren, Y., ... Martin, L. W. (2019). Revealing ferroelectric switching character using deep recurrent neural networks. Nature communications, 10(1), [4809]. https://doi.org/10.1038/s41467-019-12750-0
Agar, Joshua C. ; Naul, Brett ; Pandya, Shishir ; van der Walt, Stefan ; Maher, Joshua ; Ren, Yao ; Chen, Long Qing ; Kalinin, Sergei V. ; Vasudevan, Rama K. ; Cao, Ye ; Bloom, Joshua S. ; Martin, Lane W. / Revealing ferroelectric switching character using deep recurrent neural networks. In: Nature communications. 2019 ; Vol. 10, No. 1.
@article{dbf2b4cb658e43b1ab69e7725ffb1afa,
title = "Revealing ferroelectric switching character using deep recurrent neural networks",
abstract = "The ability to manipulate domains underpins function in applications of ferroelectrics. While there have been demonstrations of controlled nanoscale manipulation of domain structures to drive emergent properties, such approaches lack an internal feedback loop required for automatic manipulation. Here, using a deep sequence-to-sequence autoencoder we automate the extraction of latent features of nanoscale ferroelectric switching from piezoresponse force spectroscopy of tensile-strained PbZr0.2Ti0.8O3 with a hierarchical domain structure. We identify characteristic behavior in the piezoresponse and cantilever resonance hysteresis loops, which allows for the classification and quantification of nanoscale-switching mechanisms. Specifically, we identify elastic hardening events which are associated with the nucleation and growth of charged domain walls. This work demonstrates the efficacy of unsupervised neural networks in learning features of a material’s physical response from nanoscale multichannel hyperspectral imagery and provides new capabilities in leveraging in operando spectroscopies that could enable the automated manipulation of nanoscale structures in materials.",
author = "Agar, {Joshua C.} and Brett Naul and Shishir Pandya and {van der Walt}, Stefan and Joshua Maher and Yao Ren and Chen, {Long Qing} and Kalinin, {Sergei V.} and Vasudevan, {Rama K.} and Ye Cao and Bloom, {Joshua S.} and Martin, {Lane W.}",
year = "2019",
month = "12",
day = "1",
doi = "10.1038/s41467-019-12750-0",
language = "English (US)",
volume = "10",
journal = "Nature Communications",
issn = "2041-1723",
publisher = "Nature Publishing Group",
number = "1",

}

Agar, JC, Naul, B, Pandya, S, van der Walt, S, Maher, J, Ren, Y, Chen, LQ, Kalinin, SV, Vasudevan, RK, Cao, Y, Bloom, JS & Martin, LW 2019, 'Revealing ferroelectric switching character using deep recurrent neural networks', Nature communications, vol. 10, no. 1, 4809. https://doi.org/10.1038/s41467-019-12750-0

Revealing ferroelectric switching character using deep recurrent neural networks. / Agar, Joshua C.; Naul, Brett; Pandya, Shishir; van der Walt, Stefan; Maher, Joshua; Ren, Yao; Chen, Long Qing; Kalinin, Sergei V.; Vasudevan, Rama K.; Cao, Ye; Bloom, Joshua S.; Martin, Lane W.

In: Nature communications, Vol. 10, No. 1, 4809, 01.12.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Revealing ferroelectric switching character using deep recurrent neural networks

AU - Agar, Joshua C.

AU - Naul, Brett

AU - Pandya, Shishir

AU - van der Walt, Stefan

AU - Maher, Joshua

AU - Ren, Yao

AU - Chen, Long Qing

AU - Kalinin, Sergei V.

AU - Vasudevan, Rama K.

AU - Cao, Ye

AU - Bloom, Joshua S.

AU - Martin, Lane W.

PY - 2019/12/1

Y1 - 2019/12/1

N2 - The ability to manipulate domains underpins function in applications of ferroelectrics. While there have been demonstrations of controlled nanoscale manipulation of domain structures to drive emergent properties, such approaches lack an internal feedback loop required for automatic manipulation. Here, using a deep sequence-to-sequence autoencoder we automate the extraction of latent features of nanoscale ferroelectric switching from piezoresponse force spectroscopy of tensile-strained PbZr0.2Ti0.8O3 with a hierarchical domain structure. We identify characteristic behavior in the piezoresponse and cantilever resonance hysteresis loops, which allows for the classification and quantification of nanoscale-switching mechanisms. Specifically, we identify elastic hardening events which are associated with the nucleation and growth of charged domain walls. This work demonstrates the efficacy of unsupervised neural networks in learning features of a material’s physical response from nanoscale multichannel hyperspectral imagery and provides new capabilities in leveraging in operando spectroscopies that could enable the automated manipulation of nanoscale structures in materials.

AB - The ability to manipulate domains underpins function in applications of ferroelectrics. While there have been demonstrations of controlled nanoscale manipulation of domain structures to drive emergent properties, such approaches lack an internal feedback loop required for automatic manipulation. Here, using a deep sequence-to-sequence autoencoder we automate the extraction of latent features of nanoscale ferroelectric switching from piezoresponse force spectroscopy of tensile-strained PbZr0.2Ti0.8O3 with a hierarchical domain structure. We identify characteristic behavior in the piezoresponse and cantilever resonance hysteresis loops, which allows for the classification and quantification of nanoscale-switching mechanisms. Specifically, we identify elastic hardening events which are associated with the nucleation and growth of charged domain walls. This work demonstrates the efficacy of unsupervised neural networks in learning features of a material’s physical response from nanoscale multichannel hyperspectral imagery and provides new capabilities in leveraging in operando spectroscopies that could enable the automated manipulation of nanoscale structures in materials.

UR - http://www.scopus.com/inward/record.url?scp=85073744759&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85073744759&partnerID=8YFLogxK

U2 - 10.1038/s41467-019-12750-0

DO - 10.1038/s41467-019-12750-0

M3 - Article

C2 - 31641122

AN - SCOPUS:85073744759

VL - 10

JO - Nature Communications

JF - Nature Communications

SN - 2041-1723

IS - 1

M1 - 4809

ER -