Variational Convolutional Autoencoders for Anomaly Detection in Scanning Transmission Electron Microscopy.

Identifying point defects and other structural anomalies using scanning transmission electron microscopy (STEM) is important to understand a material's properties caused by the disruption of the regular pattern of crystal lattice. Due to improvements in instrumentation stability and electron optics, atomic-resolution images with a field of view of several hundred nanometers can now be routinely acquired at 1-10 Hz frame rates and such data, which often contain thousands of atomic columns, need to be analyzed. To date, image analysis is performed largely manually, but recent developments in computer vision (CV) and machine learning (ML) now enable automated analysis of atomic structures and associated defects. Here, the authors report on how a Convolutional Variational Autoencoder (CVAE) can be utilized to detect structural anomalies in atomic-resolution STEM images. Specifically, the training set is limited to perfect crystal images , and the performance of a CVAE in differentiating between single-crystal bulk data or point defects is demonstrated. It is found that the CVAE can reproduce the perfect crystal data but not the defect input data. The disagreesments between the CVAE-predicted data for defects allows for a clear and automatic distinction and differentiation of several point defect types.

High-resolution low-dose scanning transmission electron microscopy.

During the past two decades instrumentation in scanning transmission electron microscopy (STEM) has pushed toward higher intensity electron probes to increase the signal-to-noise ratio of recorded images. While this is suitable for robust specimens, biological specimens require a much reduced electron dose for high-resolution imaging. We describe here protocols for low-dose STEM image recording with a conventional field-emission gun STEM, while maintaining the high-resolution capability of the instrument. Our findings show that a combination of reduced pixel dwell time and reduced gun current can achieve radiation doses comparable to low-dose TEM.

Bridging across length scales: multi-scale ordering of supported lipid bilayers via lipoprotein self-assembly and surface patterning.

We show that a two-step process, involving spontaneous self-assembly of lipids and apolipoproteins and surface patterning, produces single, supported lipid bilayers over two discrete and independently adjustable length scales. Specifically, an aqueous phase incubation of DMPC vesicles with purified apolipoprotein A-I results in the reconstitution of high density lipoprotein (rHDL), wherein nanoscale clusters of single lipid bilayers are corralled by the protein. Adsorption of these discoidal particles to clean hydrophilic glass (or silicon) followed by direct exposure to a spatial pattern of short-wavelength UV radiation directly produces microscopic patterns of nanostructured bilayers. Alternatively, simple incubation of aqueous phase rHDL with a chemically patterned hydrophilic/hydrophobic surface produces a novel compositional pattern, caused by an increased affinity for adsorption onto hydrophilic regions relative to the surrounding hydrophobic regions. Further, by simple chemical denaturation of the boundary protein, nanoscale compartmentalization can be selectively erased, thus producing patterns of laterally fluid, lipid bilayers structured solely at the mesoscopic length scale. Since these aqueous phase microarrays of nanostructured lipid bilayers allow for membrane proteins to be embedded within single nanoscale bilayer compartments, they present a viable means of generating high-density membrane protein arrays. Such a system would permit in-depth elucidation of membrane protein structure-function relationships and the consequences of membrane compartmentalization on lipid dynamics.

Effect of cation ratio and order on magnetic circular dichroism in the double perovskite Sr2Fe1+xRe1-xO6.

Superexchange-based magnetic coupling of the two B-site cations in rock-salt-ordered double perovskite oxides is extremely sensitive to the cation ratio and degree of order. However, as a result of the limited spatial resolution of most magnetic characterization techniques, it is challenging to establish a direct relationship between magnetic properties and structure in these materials, including the effects of elemental segregation and cation disorder. Here, we use electron energy-loss magnetic chiral dichroism together with aberration-corrected electron microscopy and spectroscopy to record magnetic circular dichroism (MCD) spectra at the nm scale, in combination with structural and chemical information at the atomic scale from the very same region. We study nanoscale phases in ordered Sr2[Fe][Re]O6, ordered Sr2[Fe][Fe1/5Re4/5]O6 and disordered Sr[Fe4/5Re1/5]O3 individually, in order to understand the role of cation ratio and order on local magnetic coupling. When compared with ordered Sr2[Fe][Re]O6, we find that antiferromagnetic Fe3+-O2--Fe3+superexchange interactions arising from an excess of Fe suppress the MCD signal from Fe cations in ordered Sr2[Fe][Fe1/5Re4/5]O6, while dominant Fe3+-O2--Fe3+antiferromagnetic coupling in disordered Sr[Fe4/5Re1/5]O3 leads to a decrease in MCD signal down to the noise level. Our work demonstrates a protocol that can be used to correlate crystallographic, electronic and magnetic information in materials such as Sr2Fe1+xRe1-xO6, in order to provide insight into structure-property relationships in double perovskite oxides at the atomic scale.

First-principles study on incidence direction, individual site character, and atomic projection dependences of ELNES for perovskite compounds.

The incidence direction dependence, the individual site character dependence, and the atomic projection dependence of O-K near edge fine structure of the EEL spectrum (ELNES) from YBa2Cu3O7 (YBCO) and SrTiO3 were theoretically simulated using the first-principles band structure calculation. In order to calculate ELNES, a core-hole was introduced at the oxygen 1s orbital, and sufficiently large supercells composed of more than 100 atoms were employed. We found that the intensity of the first peak of O-K ELNES from YBCO strongly depends on the atomic projection direction, and disappears when the spectrum is measured with the other projection directions. The large projection dependence was also predicted in the O-K ELNES of SrTiO3. It was found that those spectral changes according to the position of the projection are caused by the unidirectional Ti-O-Ti bond in SrTiO3.

Role of Pr segregation in acceptor-state formation at ZnO grain boundaries.

The role of Pr doping on double Schottky barrier formations at ZnO single grain boundaries was investigated by the combination of current-voltage measurements, atomic-resolution Z-contrast scanning transmission electron microscopy, and first-principles calculations. Although Pr segregated to the specific atomic site along the boundaries, it was found not to be the direct cause of nonlinear current-voltage properties. Instead, under appropriate annealing conditions, Pr enhances formations of acceptor-type native defects that are essential for the creation of double Schottky barriers in ZnO.