Disentangling multiple scattering with deep learning: application to strain mapping from electron diffraction patterns

Munshi, Joydeep; Rakowski, Alexander; Savitzky, Benjamin H.; Ciston, Jim; Henderson, Matthew; Cholia, Shreyas; Minor, Andrew M.; Chan, Maria K. Y.; Ophus, Colin

doi:10.1038/s41524-022-00939-9

Cited by 21 publications

(10 citation statements)

References 73 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The diffraction patterns from these geometries are challenging to classify, as they contain contributions from different crystal structures, crystal orientations, and materials. In this context, various machine learning methods have been employed to analyze materials at the atomic scale, including deep learning [ 24 ] and unsupervised clustering algorithms. [ 25 ] While unsupervised clustering algorithms can classify different deformations in nanomaterials based on similarities in their diffraction patterns, they might fall short in accurately isolating strain deformation in the crystal lattice, especially when strain and deformation intricacies are closely intertwined.…”

Section: Introductionmentioning

confidence: 99%

4D‐STEM Nanoscale Strain Analysis in van der Waals Materials: Advancing beyond Planar Configurations

Bolhuis,

van Heijst,

Sangers

et al. 2024

Small Science

View full text Add to dashboard Cite

Achieving nanoscale strain fields mapping in intricate van der Waals (vdW) nanostructures, like twisted flakes and nanorods, presents several challenges due to their complex geometry, small size, and sensitivity limitations. Understanding these strain fields is pivotal as they significantly influence the optoelectronic properties of vdW materials, playing a crucial role in a plethora of applications ranging from nanoelectronics to nanophotonics. Here, a novel approach for achieving a nanoscale‐resolved mapping of strain fields across entire micron‐sized vdW nanostructures using four‐dimensional (4D) scanning transmission electron microscopy (STEM) imaging equipped with an electron microscope pixel array detector (EMPAD) is presented. This technique extends the capabilities of STEM‐based strain mapping by means of the exit‐wave power cepstrum method incorporating automated peak tracking and K‐means clustering algorithms. This approach is validated on two representative vdW nanostructures: a two‐dimensional (2D) MoS2 thin twisted flakes and a one‐dimensional (1D) MoO3/MoS2 nanorod heterostructure. Beyond just vdW materials, the versatile methodology offers broader applicability for strain‐field analysis in various low‐dimensional nanostructured materials. This advances the understanding of the intricate relationship between nanoscale strain patterns and their consequent optoelectronic properties.

show abstract

Section: Introductionmentioning

confidence: 99%

4D‐STEM Nanoscale Strain Analysis in van der Waals Materials: Advancing beyond Planar Configurations

Bolhuis,

van Heijst,

Sangers

et al. 2024

Small Science

View full text Add to dashboard Cite

show abstract

“…Machine learning (ML) techniques have been widely applied in electron microscopy for applications such as atom localization [1][2][3] , defect identification [4][5][6] , image denoising [7][8][9] , determining crystal tilts and thickness [10][11][12] , classifying crystal structures 13,14 , optimizing convergence angles 15 , identifying Bragg disks 16 , visualizing material deformations 17 , automated microscope alignment 18 , and many others. Several recent reviews [19][20][21] provide an overview of new and emerging opportunities at the interface of electron microscopy and ML.…”

Section: Introductionmentioning

confidence: 99%

Leveraging generative adversarial networks to create realistic scanning transmission electron microscopy images

et al. 2023

View full text Add to dashboard Cite

The rise of automation and machine learning (ML) in electron microscopy has the potential to revolutionize materials research through autonomous data collection and processing. A significant challenge lies in developing ML models that rapidly generalize to large data sets under varying experimental conditions. We address this by employing a cycle generative adversarial network (CycleGAN) with a reciprocal space discriminator, which augments simulated data with realistic spatial frequency information. This allows the CycleGAN to generate images nearly indistinguishable from real data and provide labels for ML applications. We showcase our approach by training a fully convolutional network (FCN) to identify single atom defects in a 4.5 million atom data set, collected using automated acquisition in an aberration-corrected scanning transmission electron microscope (STEM). Our method produces adaptable FCNs that can adjust to dynamically changing experimental variables with minimal intervention, marking a crucial step towards fully autonomous harnessing of microscopy big data.

show abstract

“…Machine learning (ML) techniques have been widely applied in electron microscopy for applications such as atom localization, 1-3 defect identification, [4][5][6] image denoising, [7][8][9] determining crystal tilts and thickness, [10][11][12] classifying crystal structures, 13,14 optimizing convergence angles, 15 identifying Bragg disks, 16 visualizing material deformations, 17 automated microscope alignment, 18 and many others. Several recent reviews [19][20][21] provide an overview of new and emerging opportunities at the interface of electron microscopy and ML.…”

Section: Introductionmentioning

confidence: 99%

Using CycleGANs to Generate Realistic STEM Images for Machine Learning

Khan¹,

Lee²,

Huang³

et al. 2023

Preprint

View full text Add to dashboard Cite

The rise of automation and machine learning (ML) in electron microscopy has the potential to revolutionize materials research by enabling the autonomous collection and processing of vast amounts of atomic resolution data. However, a major challenge is developing ML models that can reliably and rapidly generalize to large data sets with varying experimental conditions. To overcome this challenge, we develop a cycle generative adversarial network (CycleGAN) that introduces a novel reciprocal space discriminator to augment simulated data with realistic, complex spatial frequency information learned from experimental data. This enables the CycleGAN to generate nearly indistinguishable images from real experimental data, while also providing labels for further ML applications. We demonstrate the effectiveness of this approach by training a fully convolutional network (FCN) to identify single atom defects in a large data set of 4.5 million atoms, which we collected using automated acquisition in an aberration-corrected scanning transmission electron microscope (STEM). Our approach yields highly adaptable FCNs that can adjust to dynamically changing experimental variables, such as lens aberrations, noise, and local contamination, with minimal manual intervention. This represents a significant step towards building fully autonomous approaches for harnessing microscopy big data.

show abstract

Disentangling multiple scattering with deep learning: application to strain mapping from electron diffraction patterns

Cited by 21 publications

References 73 publications

4D‐STEM Nanoscale Strain Analysis in van der Waals Materials: Advancing beyond Planar Configurations

4D‐STEM Nanoscale Strain Analysis in van der Waals Materials: Advancing beyond Planar Configurations

Leveraging generative adversarial networks to create realistic scanning transmission electron microscopy images

Using CycleGANs to Generate Realistic STEM Images for Machine Learning

Contact Info

Product

Resources

About