Analysis of crystal defects by scanning transmission electron microscopy (STEM) in a modern scanning electron microscope

In scanning transmission electron microscopy (STEM), one can obtain a variety of STEM images such as bright‐field (BF) and annular dark‐field (ADF) STEM images by changing the shape of the scintillator. However, the intensity distribution of convergent beam electron diffraction (CBED) patterns at the detector plane is yet to be fully utilized. Meanwhile, direct electron detectors with fast frame rate have recently been commercialized and used in electron microscopy. Such detectors, when used for recording CBED pattern images for each STEM probe position, are called pixelated STEM detectors. With the obtained 4‐dimensional (4D) dataset, any shape of STEM detector can be synthesized in a post processing by a free selection of the integration area. Therefore, we can synthesize variety of STEM images such as differential phase contrast (DPC) and annular bright field (ABF) images, if we once record the image signal with a pixelated detector. The 4D dataset can also be used for the advanced image processing techniques such as ptychography, which has been shown to provide high efficiency for reconstructing the phase image of an object [1,2]. Using ptychography, not only the phase contrast can be enhanced but also the effect of lens aberrations to the image such as defocus can be corrected by the post processing using the information collected by a pixelated detector. Experiments were performed using an aberration corrected microscope (JEOL JEM‐ARM200F) equipped with a pixelated detector (pnDetector pnCCD), the fast direct electron detector, which can record images at a speed of 1,000 fps in a full frame mode (264 x 264 pixels). Binning or windowing can increase the speed. The camera was placed below the ADF detector to enable simultaneous recording. Figure 1 shows STEM images of a monolayer graphene obtained at 80 kV. Fig. 1a shows an ADF image obtained with the probe current of approximately 0.2 pA and the dwell time for a pixel of 0.5 ms. Because of a combination of low dose and residual uncorrected aberrations, the lattice contrast of graphene is almost buried in the noise originated from 50 Hz commercial frequency. Fig. 1b shows a reconstructed phase image using ptychography. The image contrast is significantly improved compared to the simultaneous ADF image, but the contrast transfer of the image (see the Fourier transform displayed in the inset) is anisotropic, resulting in uncertain positions of carbon atoms. This is because there remains large two‐fold astigmatism and defocus in the image shown in Fig. 1(b), because we could not adjust those by observation of the ADF image due to the weak image signal. Fig. 1c is a phase image in which the aberrations are corrected through post processing using the same 4D dataset by applying correction functions in the spatial frequency domain. The image is no longer anisotropic and the carbon atomic positions can be unambiguously determined. Although the same amount of electron dose is used to form the ADF and the corrected phase images, the result clearly shows the benefit of ptychographic phase reconstruction in improving image signal to noise and being able to correct aberrations through post processing. Figure 2 shows the ptychographic phase maps of the 4D dataset at a certain spatial frequency. With an aberration‐free electron probe, they would be flat phase on the two sidebands, and the phase difference between the bands be π. In Fig. 2a, the map of the original dataset, there is a phase gradient inside each sideband because the aberrations were present in the electron probe. Fig. 2b is the correction function that compensates for the aberration seen in Fig. 2a. Fig. 2c is the corrected phase map and corresponding to the image in Fig. 1c. Here, only defocus and two‐fold astigmatism are corrected but corrections of other higher order aberrations are in principle possible. Figure 3 shows through focus images created in the same way as above. As we know the full information on the electron wave at the condenser aperture plane, we can induce any aberrations such as defocus.

Single-Crystalline Mesoporous Palladium and Palladium-Copper Nanocubes for Highly Efficient Electrochemical CO ₂ Reduction

Three-Dimensional Visualization of Whole Synapses by Stem Tomography

Transmission electron microscopy (TEM) with low-energy electrons has been recognized as an important addition to the family of electron microscopies as it may avoid knock-on damage and increase the contrast of weakly scattering objects. Scanning electron microscopes (SEMs) are well suited for low-energy electron microscopy with maximum electron energies of 30 keV, but they are mainly used for topography imaging of bulk samples. Implementation of a scanning transmission electron microscopy (STEM) detector and a charge-coupled-device camera for the acquisition of on-axis transmission electron diffraction (TED) patterns, in combination with recent resolution improvements, make SEMs highly interesting for structure analysis of some electron-transparent specimens which are traditionally investigated by TEM. A new aspect is correlative SEM, STEM, and TED imaging from the same specimen region in a SEM which leads to a wealth of information. Simultaneous image acquisition gives information on surface topography, inner structure including crystal defects and qualitative material contrast. Lattice-fringe resolution is obtained in bright-field STEM imaging. The benefits of correlative SEM/STEM/TED imaging in a SEM are exemplified by structure analyses from representative sample classes such as nanoparticulates and bulk materials.

Atomic resolution low-angle bright-field (LABF) scanning transmission electron-microscope (STEM) images and high-angle bright-field (HABF) STEM images of [011]-orientated Si have been experimentally obtained together with high-angle annular dark-field (HAADF) STEM images. The contrast formation mechanisms of the LABF STEM and HABF STEM images are examined in comparison with HAADF STEM images. The HABF STEM images independent of defocus and thickness have spatial resolution comparable with HAADF STEM images, and are shown to be given as a simple convolution under the non-dispersion approximation of localized Bloch waves.

Strain is a crucial factor that influences the physicochemical properties of nanoparticles. Being able to precisely measure strain is important in understanding the intrinsic mechanism of the enhanced performance of nanoparticles. Techniques that have been developed for strain analysis using scanning transmission electron microscopy (STEM) images can be categorized into diffraction-based method and imaging-based method. Here, using image simulation techniques, it is found that the measured two-dimensional (2D) displacements from annular dark field (ADF) STEM images of a nanoparticle are a good approximation to a projection of the actual three-dimensional (3D) displacements. A methodology for deformation analysis is presented which is based on the detection of atomic columns from atomic-resolution STEM images in real space. Elastic deformation parameters such as strain are usually defined on the basis of a continuum of deformation. The appropriateness of various deformation parameters for atomic-scale investigation on STEM images is explored and a method for determining these is presented. We found that the local lattice parameter and principal strain components are the most physically meaningful parameters to express the materials distortion behaviour. Apart from the local lattice parameter, the other deformation parameters such as normal strains, shear strains and displacements, heavily rely on the choice of reference lattice. It is also found that different reference grids add a series of uniform offsets to these strain variations. Finally, this approach is applied to a PtCo3 bimetallic nanoparticle to quantify its deformation behaviour.

Aberration‐corrected scanning transmission electron microscopy (STEM) is providing previously unattainable views of materials at the atomic scale. The quality of STEM data is now often limited by environmental and experimental factors instead of instrument factors (e.g. electron optics). Some of these environmental limitations can be overcome by collecting and processing STEM data using new data science techniques. These techniques expose new atomic‐scale materials information by improving our ability to measure atomic column positions, 3D structure, single point defects, and atomic‐scale composition. The precision in locating atomic column positions in STEM images is fundamentally limited by the image signal to noise ratio (SNR), but typically practical limits including sample and microscope instabilities that produce distortions in STEM images are encountered before reaching the SNR limit. We have developed a non‐rigid registration (NRR) technique that corrects image distortion of all length scales and enables averaging to enhance the SNR.[1–2] Sub‐pm precision images of single crystal materials have been achieved by NRR and averaging high angle annular dark field (HAADF) STEM images. NRR has allowed measurements of pm‐scale bond length variations of Pt nanocatalyst atoms that may help explain their catalytic activity. Figure 1 shows NRR and averaged HAADF STEM data of a Pt nanocatalyst on an alumina support from a side‐view that exhibits ˜2 pm precision. Displacement measurements of each atomic column position reveal moderate Pt surface bond length contraction and strong but localized strain of Pt atoms near nanoparticle‐support interface. Some of the interface strain is transferred up the twin boundary. Determining the three dimensional atomic structure of materials from two dimensional S/TEM images is a major hurdle. The standardless atom counting technique is one promising route to measure local sample thickness by quantitatively comparing experimental and simulated HAADF STEM images and can be used to deduce 3D structure.[3] Unlike previous examples, NRR and averaging STEM images have allowed standardless atom counting with the uncertainty no longer dominated by Poisson noise [1]. This should allow the unique determination of the number of atoms in atomic columns, although this has not yet been demonstrated due to other sample limitations. Point defects are critical to the properties of a wide range of materials, but imaging single defects is challenging. Quantitative STEM has allowed imaging single substitutional and interstitial dopant impurity atoms[4], but experimentally imaging single vacancies has remained elusive. We have used HAADF STEM frozen phonon multislice simulations to predict the detectability of La vacancies in LaMnO 3 by the reduced atomic column intensity and atomic column distortions around the vacancy. NRR and averaging HAADF STEM images of LaMnO 3 improves the SNR and the image precision sufficiently to potentially detect single La vacancies. Experimental images contain candidate single La vacancies that have local atomic column distortions and intensity variations which match simulated predictions. Atomic‐resolution composition maps can be created using STEM energy dispersive x‐ray spectroscopy (EDS) spectrum imaging (SI). However, long total dwell times that may introduce spatial distortions are required because of low x‐ray production and collection efficiency. The most common approach to minimize distortions is to sum multiple SIs using online drift‐correction software that discards the individual SIs and HAADF images. The quality of EDS SIs of a Nd 2/3 TiO 3 sample was improved by saving the simultaneously acquired raw HAADF and EDS SI series, and applying post acquisition NRR and averaging. The resulting elemental maps show less spatial distortions and more atomic localization of x‐rays. In addition, a novel non‐local principle component analysis further enhances the quality of EDS SIs compared to conventional denoising methods.[5]

Three important parametric models for describing the correlation functions and spectra of stationary stochastic processes are the autoregressive (AR), moving average (MA), and autoregressive-moving average (ARMA) models. Quite recently, the MATLAB toolbox “ARMASA” has been made publicly available. This toolbox provides state-of-the-art algorithms to perform automatic identification and selection between the models based on the estimated prediction error. ARMASA works on a single segment of data, whereas in some applications, the data are available as multiple segments. We could process each segment independently and average the estimated autocorrelation functions or spectra afterward. Better performance, however, can be expected when all segments are processed simultaneously, for two reasons. Initially, the bias in the estimated model parameters depends on the number of observations in a segment. Averaging models from segments does not reduce this bias, in contrast to identification from all the data at once. In either case, the variance of the estimated model parameters will decrease. By considering all the data at once this variance reduction can be considered in the order selection stage, decreasing the selection bias. In this paper, we extend the parameter estimation algorithms of the ARMASA toolbox to the case of multiple segments. AR parameter estimation from segments has been considered before, but not for the MA and ARMA models. In addition, in the AR case, model order selection can be done easily when the segments are independent. But, in the application we are primarily interested in, the segments are formed by lines of scanning transmission electron microscopy (STEM) images, which can be correlated with each other. Because the correlation among data segments influences the effective number of independent observations available it becomes necessary to estimate that number. This is not an easy problem in general. For observations from a stochastic process with an isotropic 2-D autocorrelation function we can, however, formulate an accurate estimator in terms of the line autocorrelation function. This situation arises, for example, in STEM images of amorphous specimens. From the AR model identification for segments, the extension to the MA and ARMA models is rather straightforward because MA and ARMA algorithms are based on Durbin's methods, whose first step is the estimation of a long AR model. We compare the segment-based algorithms to conventional ARMASA with simulations of stationary stochastic processes. An illustrative example involving simulated STEM images shows large potential benefits.

The endoplasmic reticulum (ER) is a highly dynamic organelle that undergoes significant morphological alterations in response to cellular stress. While conventional transmission electron microscopy (TEM) has provided valuable insights into these changes, such as the formation of crystalloid‐ER and ER whorls, obtaining comprehensive three‐dimensional (3D) information on these large structures within their cellular context has remained a challenge. To overcome these limitations, this study introduces an innovative application of dual‐axis scanning transmission electron microscopy (STEM) tomography to investigate ER morphology under stress conditions in human embryonic kidney (HEK) cells overexpressing the cation channel polycystin‐2 (PC‐2). Benefitting from high‐resolution, increased depth‐of‐focus, and reduced aberrations, STEM tomography enabled the detailed 3D reconstruction of large cellular subvolumes, providing unprecedented views of stress‐induced ER structures. Our findings reveal distinct ultrastructural details of both crystalloid‐ER and ER whorls. Crystalloid‐ER exhibited a tubular architecture with potential interconnectedness, while ER whorls displayed a lamellar organisation and distinct membrane curvature. We observed the co‐occurrence of these distinct smooth ER (sER) morphotypes within the same cell, yet they remained spatially separated, suggesting potential functional specialisation. Furthermore, we identified direct membrane contacts in mixed morphotypes, hinting at a shared origin or dynamic relationship between these structures. The study also elucidated the interactions of these organised smooth ER (OSER) structures with other organelles, such as mitochondria (MAM sites) and vesicles. In summary, the presented ultra‐structural insights have a significant impact on our understanding of stress‐related ER morphology changes. The ability to visualise the intricate 3D architecture and spatial relationships of these structures provides novel perspectives on the ER's adaptive responses to stress, including potential roles in lipid and protein biosynthesis and intracellular communication. These findings underscore the power of dual‐axis STEM tomography for elucidating complex organellar organisation and dynamics in their native cellular context.

Our previous study showed quantitative correspondence between the size of ionic aggregates in a Cu-neutralized poly(styrene-ran-methacrylic acid) ionomer (Cu-SMAA) measured from scanning transmission electron microscopy (STEM) images and the sizes determined from X-ray scattering data interpreted using the Yarusso−Cooper model. Here we compare the average sample volume per ionic aggregate (Vp) from STEM images and X-ray scattering data from Cu-SMAA. The relationship between the 3D morphology of hard spheres as defined by the Yarusso−Cooper parameters (R1, RCA, Vp) and the projected 2D images obtained by STEM was elucidated by modeling. These simulations show that a significant number of the ionic aggregates in Cu-SMAA are obscured due to overlap effects and provide a quantitative relationship between the number of spheres in 3D and the number of features in 2D projections as a function of specimen thickness. Upon measuring the local thickness of a Cu-SMAA specimen using electron energy loss spectrometry (EELS), we find that the Vp determined from STEM images strongly agrees with that determined from X-ray scattering data. Thus, quantitative correspondence of both the size and number density of ionic aggregates has been obtained from direct imaging and X-ray scattering of Cu-SMAA. These data provide compelling evidence that the Yarusso−Cooper scattering model presents an excellent description of the nanoscale morphology of this solvent cast ionomer. Interpretation of this morphological data indicates that the ionic aggregates likely include both ionic and nonionic species. This comprehensive characterization provides a foundation on which a systematic evaluation of the factors that influence the nanoscale morphology, and consequently the macroscopic properties, can be accomplished.

Precise knowledge of the local sample thickness is often required for quantitative scanning (transmission) electron microscopy (STEM). The local sample thickness can be determined by the comparison of measured intensities from high-angle annular dark-field (HAADF)-STEM at low energies (<30 keV) with Monte-Carlo simulations. However, a suitable choice of the scattering cross-section (CS) used in the simulations is necessary to gain reliable thickness results. In this work, simulations using different CS, including the Screened Rutherford CS and different Mott CSs, were performed. The results were then compared with measurements on samples with known thickness and composition, for which an SEM equipped with a STEM detector was used. In most cases, the Screened Rutherford CS describes the experiment better than other CSs.

Electron microscopy (EM) and especially three-dimensional (3D) EM techniques have become established methods in structural virology. Investigation of virus-induced alteration of the cellular ultrastructure, such as Zika virus (ZIKV)-induced replication factories or coronavirus replication organelles, demands 3D imaging. Transmission electron microscopy (TEM) tomography is a widely used method, despite its limitation to samples with a thickness of up to 200 nm. Focused ion beam-scanning electron microscopy (SEM)-tomography can produce 3D data of larger volumes with isotropic, albeit typically lower resolution. Alternative techniques, such as block face-scanning electron microscopy (BF-SEM), SEM array tomography, or TEM imaging of serial sections are used for imaging of larger volumes. However, compared to the previously mentioned techniques, these techniques come at the cost of much lower resolution along the Z-axis of the sample. A technique that provides 3D information of samples up to 1 µm thickness with isotropic resolution of a few nanometers is scanning transmission electron microscopy (STEM) tomography. Here, we present a protocol for the preparation of high-pressure frozen, freeze-substituted, and resin-embedded virological specimens and their analysis using STEM tomography. This protocol benefits from the advantages of room temperature imaging while preserving the biological ultrastructure in a near-native state. We show two representative examples for questions that can be answered using STEM tomography. First, we apply the protocol for studying virion morphogenesis of a recombinant vesicular stomatitis virus (VSV). The STEM tomograms offer information on recombinant VSV budding that is otherwise not accessible by 2D imaging. Second, we show correlative light and electron microscopy (CLEM) using Foerster resonance energy transfer (FRET) imaging and STEM tomography (FRET-3D-CLEM) of EF-C peptide nanofibrils. This recently published combination gives new insights into the uptake and disassembly of infection-enhancing peptide nanofibrils, especially profiting from the large volume that is accessible by STEM tomography.

Objective molecular dynamics simulations coupled with a density functional-based tight-binding model indicated that a stress-free single-walled ($14,6$) MoS${}_{2}$ nanotube exhibits a torsional deformation of 0.87 deg/nm. Simulated electron diffraction patterns and atomic-resolution annular dark field scanning transmission electron microscopy (ADF-STEM) images of the computed nanotube structures show promise that this peculiar feature can be identified experimentally. The small intrinsic twist removes the translational periodicity prescribed by the rolled-up construction and defines a nanotube for which the atomic order is most fundamentally described by the objective structures concept.

Scanning transmission electron microscopy (STEM) images can be limited by either aberrations or, because of the technique's serial acquisition, by the effects of environmental instabilities. MatLab image-processing code was written which performs a two stage process to restore degraded high-resolution high-angle annular dark-field (HAADF) STEM images. Firstly, individual images were analysed to identify and correct for high-frequency scan noise; image resolution and signal-noise ratio (SNR) are used as performance metrics and were improved by up to 11.8% and 49% respectively. Secondly, a focal series was used to identify variations in information transfer as a function of defocus (aberrations) whereafter a single image was reconstructed yielding an increase in resolution and SNR of 9.88% and 205% respectively.

Several members of a new family of heterostructures [(LaSe)1.17]1Vn(1+y)+1Se2n+2 with n = 1, 2, and 3 were prepared using a diffusion constrained, kinetically controlled synthesis approach. Specular diffraction patterns collected as a function of annealing temperature show the evolution of designed precursors into highly ordered heterostructures. Scanning transmission electron microscopy (STEM) images reveal that the structure of n = 3 consists of rock salt structured LaSe bilayers alternating with vanadium selenide layers of varying thickness, which are structurally closely related to V3Se4. Interplanar distances obtained from the STEM images were successfully used as the starting point for Rietveld refinements of the specular diffraction patterns of these crystallographically aligned compounds. Utilizing this unorthodox combined approach to extract detailed structural information unambiguously, we demonstrated that these thin film compounds are the first examples of chalcogenide-based heterostructures, where the bulk structures of both building blocks lack a van der Waals gap, yet a nonepitaxial incommensurate interface forms. Moreover, the refinement results of the n = 2 and 3 heterostructures suggest that the structure of the V–Se layers can be varied ranging from VSe2 to VSe depending on the film composition. The electrical resistivity of the [(LaSe)1.17]1Vn(1+y)+1Se2n+2 heterostructures changes systematically from semiconducting toward metallic behavior with increasing n, showcasing the ability to tune physical properties by precisely controlling the layer sequence in these heterostructures.