Mapping mesoscopic phase evolution during E-beam induced transformations via deep learning of atomically resolved images

被引:44
作者
Vasudevan, Rama K. [1 ,2 ]
Laanait, Nouamane [3 ]
Ferragut, Erik M. [4 ,5 ]
Wang, Kai [1 ,2 ]
Geohegan, David B. [1 ,2 ]
Xiao, Kai [1 ,2 ]
Ziatdinov, Maxim [1 ,2 ]
Jesse, Stephen [1 ,2 ]
Dyck, Ondrej [1 ,2 ]
Kalinin, Sergei, V [1 ,2 ]
机构
[1] Oak Ridge Natl Lab, Ctr Nanophase Mat Sci, Oak Ridge, TN 37831 USA
[2] Oak Ridge Natl Lab, Inst Funct Imaging Mat, Oak Ridge, TN 37831 USA
[3] Oak Ridge Natl Lab, Computat Sci & Engn Div, Oak Ridge, TN 37831 USA
[4] Oak Ridge Natl Lab, Quantum Comp Inst, Oak Ridge, TN 37831 USA
[5] UnitedHlth Grp, POB 1459, Minneapolis, MN 55440 USA
关键词
2-DIMENSIONAL MATERIALS; ELECTRON; INFRASTRUCTURE; TRANSITION; CHEMISTRY; GRAPHENE; LEVEL;
D O I
10.1038/s41524-018-0086-7
中图分类号
O64 [物理化学(理论化学)、化学物理学];
学科分类号
070304 ; 081704 ;
摘要
Understanding transformations under electron beam irradiation requires mapping the structural phases and their evolution in real time. To date, this has mostly been a manual endeavor comprising difficult frame-by-frame analysis that is simultaneously tedious and prone to error. Here, we turn toward the use of deep convolutional neural networks (DCNN) to automatically determine the Bravais lattice symmetry present in atomically resolved images. A DCNN is trained to identify the Bravais lattice class given a 2D fast Fourier transform of the input image. Monte-Carlo dropout is used for determining the prediction probability, and results are shown for both simulated and real atomically resolved images from scanning tunneling microscopy and scanning transmission electron microscopy. A reduced representation of the final layer output allows to visualize the separation of classes in the DCNN and agrees with physical intuition. We then apply the trained network to electron beam-induced transformations in WS2, which allows tracking and determination of growth rate of voids. We highlight two key aspects of these results: (1) it shows that DCNNs can be trained to recognize diffraction patterns, which is markedly different from the typical "real image" cases and (2) it provides a method with inbuilt uncertainty quantification, allowing the real-time analysis of phases present in atomically resolved images.
引用
收藏
页数:9
相关论文
共 63 条
[1]  
[Anonymous], 1999, P SPIES INT S OPT SC
[2]   Machine-learning approach for one- and two-body corrections to density functional theory: Applications to molecular and condensed water [J].
Bartok, Albert P. ;
Gillan, Michael J. ;
Manby, Frederick R. ;
Csanyi, Gabor .
PHYSICAL REVIEW B, 2013, 88 (05)
[3]   Organic Chemistry as a Language and the Implications of Chemical Linguistics for Structural and Retrosynthetic Analyses [J].
Cadeddu, Andrea ;
Wylie, Elizabeth K. ;
Jurczak, Janusz ;
Wampler-Doty, Matthew ;
Grzybowski, Bartosz A. .
ANGEWANDTE CHEMIE-INTERNATIONAL EDITION, 2014, 53 (31) :8108-8112
[4]  
Chollet F., 2015, KERAS DEEP LEARNING, V7, P8
[5]  
Curtarolo S, 2013, NAT MATER, V12, P191, DOI [10.1038/NMAT3568, 10.1038/nmat3568]
[6]   Electron-Beam-Induced Elastic-Plastic Transition in Si Nanowires [J].
Dai, Sheng ;
Zhao, Jiong ;
Xie, Lin ;
Cai, Yuan ;
Wang, Ning ;
Zhu, Jing .
NANO LETTERS, 2012, 12 (05) :2379-2385
[7]  
Dyck O., 2017, PREPRINT
[8]   Placing single atoms in graphene with a scanning transmission electron microscope [J].
Dyck, Ondrej ;
Kim, Songkil ;
Kalinin, Sergei V. ;
Jesse, Stephen .
APPLIED PHYSICS LETTERS, 2017, 111 (11)
[9]   In-operando high-speed tomography of lithium-ion batteries during thermal runaway [J].
Finegan, Donal P. ;
Scheel, Mario ;
Robinson, James B. ;
Tjaden, Bernhard ;
Hunt, Ian ;
Mason, Thomas J. ;
Millichamp, Jason ;
Di Michiel, Marco ;
Offer, Gregory J. ;
Hinds, Gareth ;
Brett, Dan J. L. ;
Shearing, Paul R. .
NATURE COMMUNICATIONS, 2015, 6
[10]   Predicting crystal structure by merging data mining with quantum mechanics [J].
Fischer, Christopher C. ;
Tibbetts, Kevin J. ;
Morgan, Dane ;
Ceder, Gerbrand .
NATURE MATERIALS, 2006, 5 (08) :641-646