The impact of STEM aberration correction on materials science

A simple coordinate transformation method for quickly locating the features of interest in TEM samples

We developed a simple coordinate transformation method for quickly locating features of interest (FOIs) of samples in transmission electron microscope (TEM). The method is well suited for conducting sample searches in aberration-corrected scanning/transmission electron microscopes (S/TEM), where the survey can be very time-consuming because of the limited field of view imposed by the highly excited objective lens after fine-tuning the aberration correctors. For implementation, a digital image of the sample and the TEM holder was captured using a simple stereo-optical microscope. Naturally presented geometric patterns on the holder were referenced to construct a projective transformation between the electron and optical coordinate systems. The test results demonstrated that the method was accurate and required no electron microscope or specimen holder modifications. Additionally, it eliminated the need to mount the sample onto specific patterned TEM grids or deposit markers, resulting in universal applications for most TEM samples, holders and electron microscopes for fast FOI identification. Furthermore, we implemented the method into a Gatan script for graphical-user-interface-based step-by-step instructions. Through online communication, the script enabled real-time navigation and tracking of the motion of samples in TEM on enlarged optical images with a panoramic view.

Atomic-Scale Imaging of Clay Mineral Nanosheets and Their Supramolecular Complexes through Electron Microscopy: A Supramolecular Chemist's Perspective

Recent advancements in electron microscopy techniques have revolutionized the ability to directly visualize and understand the intricate world of supramolecular chemistry. This paper provides a concise overview of a study delving into the atomic-scale imaging of monolayer clay mineral nanosheets and their associated supramolecular complexes. The imaging is conducted by harnessing the power of aberration-corrected scanning transmission electron microscopy (STEM). Clay mineral nanosheets, with their anionic charge and ultrathin thickness (of 1 nm), serve as a stable Coulombic host material for cationic guest molecules through electrostatic interactions, facilitating exceptional stability and control during observation. By incorporation of heavy-metal atom markers coordinated within the target molecules, high-angle annular dark field STEM enables a clear visualization of these supramolecular complexes. This approach helps to overcome the limitations of graphene-based systems and expands the possibilities of atomic-scale imaging of nonperiodic molecular assemblies formed by weak supramolecular interactions. The fusion of electron microscopy techniques with the principles of supramolecular and material chemistry offers exciting opportunities for studying the structure, behavior, and properties of complex supramolecular systems. It sheds light on the intricate molecular architectures and design principles governing these systems. This study showcases the immense potential of electron microscopy in supramolecular chemistry and invites researchers from various disciplines to explore the transformative possibilities of atomic-scale imaging in the field.

Rotational position error correction in ptychography

Accurate determination of scan positions is essential for achieving high-quality reconstructions in ptychographic imaging. This study presents and demonstrates a method for determining the rotation angle of the scan pattern relative to the detector pixel array using diffraction data. The method is based on the Fourier-Mellin transform and cross-correlation calculation. It can correct rotation errors up to 60 deg. High-quality reconstructions were obtained for visible light and electron microscopy datasets, and intricate structures of samples can be revealed. We believe that this refinement method for rotary position errors can be valuable for improving the performance of ptychographic four-dimensional scanning transmission electron microscopy.

Bimetallic AgPt Nanoalloys as an Electrocatalyst for Ethanol Oxidation Reaction: Synthesis, Structural Analysis, and Electro-Catalytic Activity

In the present work, the chemical synthesis of AgPt nanoalloys is reported by the polyol method using polyvinylpyrrolidone (PVP) as a surfactant and a heterogeneous nucleation approach. Nanoparticles with different atomic compositions of the Ag and Pt elements (1:1 and 1:3) were synthesized by adjusting the molar ratios of the precursors. The physicochemical and microstructural characterization was initially performed using the UV-Vis technique to determine the presence of nanoparticles in suspension. Then, the morphology, size, and atomic structure were determined using XRD, SEM, and HAADF-STEM techniques, confirming the formation of a well-defined crystalline structure and homogeneous nanoalloy with an average particle size of less than 10 nm. Finally, the cyclic voltammetry technique evaluated the electrochemical activity of bimetallic AgPt nanoparticles supported on Vulcan XC-72 carbon for the ethanol oxidation reaction in an alkaline medium. Chronoamperometry and accelerated electrochemical degradation tests were performed to determine their stability and long-term durability. The synthesized AgPt (1:3)/C electrocatalyst presented significative catalytic activity and superior durability due to the introduction of Ag that weakens the chemisorption of the carbonaceous species. Thus, it could be an attractive candidate for cost-effective ethanol oxidation compared to commercial Pt/C.

Near-real-time diagnosis of electron optical phase aberrations in scanning transmission electron microscopy using an artificial neural network

The key to optimizing spatial resolution in a state-of-the-art scanning transmission electron microscope is the ability to measure and correct for electron optical aberrations of the probe-forming lenses precisely. Several diagnostic methods for aberration measurement and correction have been proposed, albeit often at the cost of relatively long acquisition times. Here, we illustrate how artificial intelligence can be used to provide near-real-time diagnosis of aberrations from individual Ronchigrams. The demonstrated speed of aberration measurement is important because microscope conditions can change rapidly. It is also important for the operation of MEMS-based hardware correction elements, which have less intrinsic stability than conventional electromagnetic lenses.

Physics Discovery in Nanoplasmonic Systems via Autonomous Experiments in Scanning Transmission Electron Microscopy

Physics-driven discovery in an autonomous experiment has emerged as a dream application of machine learning in physical sciences. Here, this work develops and experimentally implements a deep kernel learning (DKL) workflow combining the correlative prediction of the target functional response and its uncertainty from the structure, and physics-based selection of acquisition function, which autonomously guides the navigation of the image space. Compared to classical Bayesian optimization (BO) methods, this approach allows to capture the complex spatial features present in the images of realistic materials, and dynamically learn structure-property relationships. In combination with the flexible scalarizer function that allows to ascribe the degree of physical interest to predicted spectra, this enables physical discovery in automated experiment. Here, this approach is illustrated for nanoplasmonic studies of nanoparticles and experimentally implemented in a truly autonomous fashion for bulk- and edge plasmon discovery in MnPS₃ , a lesser-known beam-sensitive layered 2D material. This approach is universal, can be directly used as-is with any specimen, and is expected to be applicable to any probe-based microscopic techniques including other STEM modalities, scanning probe microscopies, chemical, and optical imaging.

Tailoring magnetism in silicon-doped zigzag graphene edges

Recently, the edges of single-layer graphene have been experimentally doped with silicon atoms by means of scanning transmission electron microscopy. In this work, density functional theory is applied to model and characterize a wide range of experimentally inspired silicon doped zigzag-type graphene edges. The thermodynamic stability is assessed and the electronic and magnetic properties of the most relevant edge configurations are unveiled. Importantly, we show that silicon doping of graphene edges can induce a reversion of the spin orientation on the adjacent carbon atoms, leading to novel magnetic properties with possible applications in the field of spintronics.

Dynamic hetero-metallic bondings visualized by sequential atom imaging

Traditionally, chemistry has been developed to obtain thermodynamically stable and isolable compounds such as molecules and solids by chemical reactions. However, recent developments in computational chemistry have placed increased importance on studying the dynamic assembly and disassembly of atoms and molecules formed in situ. This study directly visualizes the formation and dissociation dynamics of labile dimers and trimers at atomic resolution with elemental identification. The video recordings of many homo- and hetero-metallic dimers are carried out by combining scanning transmission electron microscopy (STEM) with elemental identification based on the Z-contrast principle. Even short-lived molecules with low probability of existence such as AuAg, AgCu, and AuAgCu are directly visualized as a result of identifying moving atoms at low electron doses.

The dynamic assembly and disassembly of atoms and molecules is challenging to characterize in real time, with atomic resolution and elemental identification. Here, the authors report direct observation of more than twenty homo and hetero-metallic compounds, including labile Ag-Cu dimers and Au-Ag-Cu trimers.

Accurate and Robust Calibration of the Uniform Affine Transformation Between Scan-Camera Coordinates for Atom-Resolved In-Focus 4D-STEM Datasets

Accurate geometrical calibration between the scan coordinates and the camera coordinates is critical in four-dimensional scanning transmission electron microscopy (4D-STEM) for both quantitative imaging and ptychographic reconstructions. For atomic-resolved, in-focus 4D-STEM datasets, we propose a hybrid method incorporating two sub-routines, namely a J-matrix method and a Fourier method, which can calibrate the uniform affine transformation between the scan-camera coordinates using raw data, without a priori knowledge of the crystal structure of the specimen. The hybrid method is found robust against scan distortions and residual probe aberrations. It is also effective even when defects are present in the specimen, or the specimen becomes relatively thick. We will demonstrate that a successful geometrical calibration with the hybrid method will lead to a more reliable recovery of both the specimen and the electron probe in a ptychographic reconstruction. We will also show that, although the elimination of local scan position errors still requires an iterative approach, the rate of convergence can be improved, and the residual errors can be further reduced if the hybrid method can be firstly applied for initial calibration. The code is made available as a simple-to-use tool to correct affine transformations of the scan-camera coordinates in 4D-STEM experiments.

Sub-Sampled Imaging for STEM: Maximising Image Speed, Resolution and Precision Through Reconstruction Parameter Refinement

Sub-sampling during image acquisition in scanning transmission electron microscopy (STEM) has been shown to provide a means to increase the overall speed of acquisition while at the same time providing an efficient means to control the dose, dose rate and dose overlap delivered to the sample. In this paper, we discuss specifically the parameters used to reconstruct sub-sampled images and highlight their effect on inpainting using the beta-process factor analysis (BPFA) methodology. The selection of the main control parameters can have a significant effect on the resolution, precision and sensitivity of the final inpainted images, and here we demonstrate a method by which these parameters can be optimised for any image in STEM. As part of this work, we also provide a link to open source code and a tutorial on its use, whereby these parameters can be tested for any datasets. When coupled with the hardware necessary to rapidly sub-sample images in STEM, this approach can have significant implications for imaging beam sensitive materials and dynamic processes.

Physics makes the difference: Bayesian optimization and active learning via augmented Gaussian process

Both experimental and computational methods for the exploration of structure, functionality, and properties of materials often necessitate the search across broad parameter spaces to discover optimal experimental conditions and regions of interest in the image space or parameter space of computational models. The direct grid search of the parameter space tends to be extremely time-consuming, leading to the development of strategies balancing exploration of unknown parameter spaces and exploitation towards required performance metrics. However, classical Bayesian optimization (BO) strategies based on the Gaussian process (GP) do not readily allow for the incorporation of the known physical behaviors or past knowledge. Here we explore a hybrid optimization/exploration algorithm created by augmenting the standard GP with a structured probabilistic model of the expected system’s behavior. This approach balances the flexibility of the non-parametric GP approach with a rigid structure of physical knowledge encoded into the parametric model. The fully Bayesian treatment of the latter allows additional control over the optimization via the selection of priors for the model parameters. The method is demonstrated for a noisy version of a standard univariate test function used to evaluate optimization algorithms and further extended to physical lattice models. This methodology is expected to be universally suitable for injecting prior knowledge in the form of physical models and past data in the BO framework.

Cluster Nanozymes with Optimized Reactivity and Utilization of Active Sites for Effective Peroxidase (and Oxidase) Mimicking

Single-atom catalysts have attracted attention in the past decade since they maximize the utilization of active sites and facilitate the understanding of product distribution in some catalytic reactions. Recently, this idea has been extended to single-atom nanozymes (SAzymes) for the mimicking of natural enzymes such as horseradish peroxidase (HRP) often used in bioanalytical applications. Herein, it is demonstrated that those SAzymes without constructing the reaction pocket of HRP still undergo the OH radical-mediated pathway like most of the reported nanozymes. Their positively charged single-atom centers resulting from support electronegative oxygen/nitrogen hinder the reductive conversion of H₂ O₂ to OH radicals and hence display low activity per site. In contrast, it is found that this step can be facilitated over their metallic counterparts on cluster nanozymes with much higher site activity and atom efficiency (cf. SAzymes with 100% atom utilization). Besides the mimicking of HRP in glucose detection, cluster nanozymes are also demonstrated as a better oxidase mimetic for glutathione detection.

Seeing Structural Mechanisms of Optimized Piezoelectric and Thermoelectric Bulk Materials through Structural Defect Engineering

Aberration-corrected scanning transmission electron microscopy (AC-STEM) has evolved into the most powerful characterization and manufacturing platform for all materials, especially functional materials with complex structural characteristics that respond dynamically to external fields. It has become possible to directly observe and tune all kinds of defects, including those at the crucial atomic scale. In-depth understanding and technically tailoring structural defects will be of great significance for revealing the structure-performance relation of existing high-property materials, as well as for foreseeing paths to the design of high-performance materials. Insights would be gained from piezoelectrics and thermoelectrics, two representative functional materials. A general strategy is highlighted for optimizing these functional materials’ properties, namely defect engineering at the atomic scale.

Enhanced Mechanical Properties in 6082 Aluminum Alloy Processed by Cyclic Deformation

Aging heat treatment is the most commonly used strengthening method for Al–Mg–Si alloys since high-density precipitates will be formed to hinder the movement of dislocations. In the current work, room temperature cyclic deformation was attempted to strengthen the alloy. We compared tensile test results of aged samples and cyclically deformed samples. It was found that cyclically deformed samples can achieve similar strength and approximately twice the uniform elongation as the peak aged samples. The high density of dislocations and nanoclusters observed in the cyclically deformed samples is thought to be the main reason for strengthening. Different cyclic deformation conditions have been tried and their effects were discussed.

Advances and Applications of Atomic-Resolution Scanning Transmission Electron Microscopy

Although scanning transmission electron microscopy (STEM) images of individual heavy atoms were reported 50 years ago, the applications of atomic-resolution STEM imaging became wide spread only after the practical realization of aberration correctors on field-emission STEM/TEM instruments to form sub-Ångstrom electron probes. The innovative designs and advances of electron optical systems, the fundamental understanding of electron–specimen interaction processes, and the advances in detector technology all played a major role in achieving the goal of atomic-resolution STEM imaging of practical materials. It is clear that tremendous advances in computer technology and electronics, image acquisition and processing algorithms, image simulations, and precision machining synergistically made atomic-resolution STEM imaging routinely accessible. It is anticipated that further hardware/software development is needed to achieve three-dimensional atomic-resolution STEM imaging with single-atom chemical sensitivity, even for electron-beam-sensitive materials. Artificial intelligence, machine learning, and big-data science are expected to significantly enhance the impact of STEM and associated techniques on many research fields such as materials science and engineering, quantum and nanoscale science, physics and chemistry, and biology and medicine. This review focuses on advances of STEM imaging from the invention of the field-emission electron gun to the realization of aberration-corrected and monochromated atomic-resolution STEM and its broad applications.

Imaging Secondary Electron Emission from a Single Atomic Layer

Graphene-based devices hold promise for a wide range of technological applications. Yet characterizing the structure and the electrical properties of a material that is only one atomic layer thick still poses technical challenges. Recent investigations indicate that secondary-electron electron-beam-induced current (SE-EBIC) imaging can reveal subtle details regarding electrical conductivity and electron transport with high spatial resolution. Here, it is shown that the SEEBIC imaging mode can be used to detect suspended single layers of graphene and distinguish between different numbers of layers. Pristine and contaminated areas of graphene are also compared to show that pristine graphene exhibits a substantially lower SE yield than contaminated regions. This SEEBIC imaging mode may provide valuable information for the engineering of surface coatings where SE yield is a priority.

Deep learning in electron microscopy

Deep learning is transforming most areas of science and technology, including electron microscopy. This review paper offers a practical perspective aimed at developers with limited familiarity. For context, we review popular applications of deep learning in electron microscopy. Following, we discuss hardware and software needed to get started with deep learning and interface with electron microscopes. We then review neural network components, popular architectures, and their optimization. Finally, we discuss future directions of deep learning in electron microscopy.

Electron beam triggered single-atom dynamics in two-dimensional materials

Controlling atomic structure and dynamics with single-atom precision is the ultimate goal in nanoscience and nanotechnology. Despite great successes being achieved by scanning tunneling microscopy (STM) over the past a few decades, fundamental limitations, such as ultralow temperature, and low throughput, significantly hinder the fabrication of a large array of atomically defined structures by STM. The advent of aberration correction in scanning transmission electron microscopy (STEM) revolutionized the field of nanomaterials characterization pushing the detection limit down to single-atom sensitivity. The sub-angstrom focused electron beam (e-beam) of STEM is capable of interacting with an individual atom, thereby it is the ideal platform to direct and control matter at the level of a single atom or a small cluster. In this article, we discuss the transfer of energy and momentum from the incident e-beam to atoms and their subsequent potential dynamics under different e-beam conditions in 2D materials, particularly transition metal dichalcogenides (TMDs). Next, we systematically discuss the e-beam triggered structural evolutions of atomic defects, line defects, grain boundaries, and stacking faults in a few representative 2D materials. Their formation mechanisms, kinetic paths, and practical applications are comprehensively discussed. We show that desired structural evolution or atom-by-atom assembly can be precisely manipulated by e-beam irradiation which could introduce intriguing functionalities to 2D materials. In particular, we highlight the recent progress on controlling single Si atom migration in real-time on monolayer graphene along an extended path with high throughput in automated STEM. These results unprecedentedly demonstrate that single-atom dynamics can be realized by an atomically focused e-beam. With the burgeoning of artificial intelligence and big data, we can expect that fully automated microscopes with real-time data analysis and feedback could readily design and fabricate large scale nanostructures with unique functionalities in the near future.

Dynamic observation and motion tracking of individual gold atoms with HAADF-STEM imaging

The tracked trajectories give direct visualization of individual atoms' movements, and the number of tracked atoms in each frame is quantified.

Atomic-scale study of nanocatalysts by aberration-corrected electron microscopy

Aberration-corrected electron microscopy (AC-EM) including transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM) has become one of the most powerful technologies in the studies of nanocatalysts. With the current spatial resolution of sub-0.5 Å and energy resolution of 10 meV, AC-EM can quantificationally articulate the connection between catalytic properties and atomic configurations of nanocatalysts. However, the restricted irradiation sensitive characteristics of specimens pose an obstacle to solve their intrinsic structure. Low-dose imaging should be applied to overcome this problem. In addition, the choice of appropriate imaging method is also crucial to tackle specific structural problems of nanocatalysts. On the basis of careful management of electron dose and selection of suitable imaging method,in situgas and liquid S/TEM are able to reveal the structure evolution of nanocatalysts in real-time. Further combination with residual gas analysis would deepen the understanding of the catalytic reaction.

Improving electron micrograph signal-to-noise with an atrous convolutional encoder-decoder

We present an atrous convolutional encoder-decoder trained to denoise electron micrographs. It consists of a modified Xception backbone, atrous convoltional spatial pyramid pooling module and a multi-stage decoder. Our neural network was trained end-to-end using 512  ×  512 micrographs created from a large dataset of high-dose ( > 2500 counts per pixel) micrographs with added Poisson noise to emulate low-dose ( ≪  300 counts per pixel) data. It was then fine-tuned for high dose data (200-2500 counts per pixel). Its performance is benchmarked against bilateral, Gaussian, median, total variation, wavelet, and Wiener restoration methods with their default parameters. Our network outperforms their best mean squared error and structural similarity index performances by 24.6% and 9.6% for low doses and by 43.7% and 5.5% for high doses. In both cases, our network's mean squared error has the lowest variance. Source code and links to our high-quality dataset and pre-trained models are available at https://github.com/Jeffrey-Ede/Electron-Micrograph-Denoiser.

Structural properties of sub-nanometer metallic clusters

At the nanoscale, the investigation of structural features becomes fundamental as we can establish relationships between cluster geometries and their physicochemical properties. The peculiarity lies in the variety of shapes often unusual and far from any geometrical and crystallographic intuition clusters can assume. In this respect, we should treat and consider nanoparticles as a new form of matter. Nanoparticle structures depend on their size, chemical composition, ordering, as well as external conditions e.g. synthesis method, pressure, temperature, support. On top of that, at finite temperatures nanoparticles can fluctuate among different structures, opening new and exciting horizons for the design of optimal nanoparticles for advanced applications. This article aims to overview geometrical features of transition metal clusters and of their various rearrangements.

Compressed Sensing of Scanning Transmission Electron Microscopy (STEM) With Nonrectangular Scans

Scanning transmission electron microscopy (STEM) has become the main stay for materials characterization on atomic level, with applications ranging from visualization of localized and extended defects to mapping order parameter fields. In recent years, attention has focused on the potential of STEM to explore beam induced chemical processes and especially manipulating atomic motion, enabling atom-by-atom fabrication. These applications, as well as traditional imaging of beam sensitive materials, necessitate increasing the dynamic range of STEM in imaging and manipulation modes, and increasing the absolute scanning speed which can be achieved by combining sparse sensing methods with nonrectangular scanning trajectories. Here we have developed a general method for real-time reconstruction of sparsely sampled images from high-speed, noninvasive and diverse scanning pathways, including spiral scan and Lissajous scan. This approach is demonstrated on both the synthetic data and experimental STEM data on the beam sensitive material graphene. This work opens the door for comprehensive investigation and optimal design of dose efficient scanning strategies and real-time adaptive inference and control of e-beam induced atomic fabrication.

The Atomic Circus: Small Electron Beams Spotlight Advanced Materials Down to the Atomic Scale

Defects in crystalline materials have a tremendous impact on their functional behavior. Controlling and tuning of these imperfections can lead to marked improvements in their physical, electrical, magnetic, and optical properties. Thanks to the development of aberration-corrected (scanning) transmission electron microscopy (STEM/TEM), direct visualization of defects at multiple length scales has now become possible, including those critically important defects at the atomic scale. Thorough understanding of the nature and dynamics of these defects is the key to unraveling the fundamental origins of structure-property relationships. Such insight can therefore allow the creation of new materials with desired properties through appropriate defect engineering. Herein, several examples of new insights obtained from representative functional materials are shown, including piezoelectrics/ferroelectrics, oxide interfaces, thermoelectrics, electrocatalysts, and 2D materials.

Precipitation of binary quasicrystals along dislocations

Dislocations in crystals naturally break the symmetry of the bulk, introducing local atomic configurations with symmetries such as fivefold rings. But dislocations do not usually nucleate aperiodic structure along their length. Here we demonstrate the formation of extended binary quasicrystalline precipitates with Penrose-like random-tiling structures, beginning with chemical ordering within the pentagonal structure at cores of prismatic dislocations in Mg-Zn alloys. Atomic resolution observations indicate that icosahedral chains centered along [0001] pillars of Zn interstitial atoms are formed templated by the fivefold rings at dislocation cores. They subsequently form columns of rhombic and elongated hexagonal tiles parallel to the dislocation lines. Quasicrystalline precipitates are formed by random tiling of these rhombic and hexagonal tiles. Such precipitation may impact dislocation glide and alloy strength.

A streaming multi-GPU implementation of image simulation algorithms for scanning transmission electron microscopy

Simulation of atomic-resolution image formation in scanning transmission electron microscopy can require significant computation times using traditional methods. A recently developed method, termed plane-wave reciprocal-space interpolated scattering matrix (PRISM), demonstrates potential for significant acceleration of such simulations with negligible loss of accuracy. Here, we present a software package called Prismatic for parallelized simulation of image formation in scanning transmission electron microscopy (STEM) using both the PRISM and multislice methods. By distributing the workload between multiple CUDA-enabled GPUs and multicore processors, accelerations as high as 1000 × for PRISM and 15 × for multislice are achieved relative to traditional multislice implementations using a single 4-GPU machine. We demonstrate a potentially important application of Prismatic, using it to compute images for atomic electron tomography at sufficient speeds to include in the reconstruction pipeline. Prismatic is freely available both as an open-source CUDA/C++ package with a graphical user interface and as a Python package, PyPrismatic.