Crystallography has been fundamental to the development of many fields of science over the last century. However, much of our modern science and technology relies on materials with defects and disorders, and their three-dimensional (3D) atomic structures are not accessible to crystallography. One method capable of addressing this major challenge is atomic electron tomography. By combining advanced electron microscopes and detectors with powerful data analysis and tomographic reconstruction algorithms, it is now possible to determine the 3D atomic structure of crystal defects such as grain boundaries, stacking faults, dislocations, and point defects, as well as to precisely localize the 3D coordinates of individual atoms in materials without assuming crystallinity. Here we review the recent advances and the interdisciplinary science enabled by this methodology. We also outline further research needed for atomic electron tomography to address long-standing unresolved problems in the physical sciences.

We present time-of-flight neutron total scattering and polarized neutron scattering measurements of the magnetically frustrated compounds NaCaCo2F7 and NaSrCo2F7, which belong to a class of recently discovered pyrochlore compounds based on transition metals and fluorine. The magnetic pair distribution function (mPDF) technique is used to analyze and model the total scattering data in real space. We find that a previously proposed model of short-range XY-like correlations with a length scale of 10-15 Å, combined with nearest-neighbor collinear antiferromagnetic correlations, accurately describes the mPDF data at low temperature, confirming the magnetic ground state in these materials. This model is further verified by the polarized neutron scattering data. From an analysis of the temperature dependence of the mPDF and polarized neutron scattering data, we find that short-range correlations persist on the nearest-neighbor length scale up to 200 K, approximately two orders of magnitude higher than the spin freezing temperatures of these compounds. These results highlight the opportunity presented by these new pyrochlore compounds to study the effects of geometric frustration at relatively high temperatures, while also advancing the mPDF technique and providing an opportunity to investigate a genuinely short-range-ordered magnetic ground state directly in real space.

A major challenge to the implementation of artificial photosynthesis (AP), in which fuels are produced from abundant materials (water and carbon dioxide) in an electrochemical cell through the action of sunlight, is the discovery of active, inexpensive, safe, and stable catalysts for the oxygen evolution reaction (OER). Multimetallic molecular catalysts, inspired by the natural photosynthetic enzyme, can provide important guidance for catalyst design, but the necessary mechanistic understanding has been elusive. In particular, fundamental transformations for reactive intermediates are difficult to observe, and well-defined molecular models of such species are highly prone to decomposition by intermolecular aggregation. Here, we present a general strategy for stabilization of the molecular cobalt-oxo cubane core (Co₄O₄) by immobilizing it as part of metal-organic frameworks, thus preventing intermolecular pathways of catalyst decomposition. These materials retain the OER activity and mechanism of the molecular Co₄O₄ analog yet demonstrate unprecedented long-term stability at pH 14. The organic linkers of the framework allow for chemical fine-tuning of activity and stability and, perhaps most importantly, provide "matrix isolation" that allows for observation and stabilization of intermediates in the water-splitting pathway.

Deep learning (DL) is one of the fastest-growing topics in materials data science, with rapidly emerging applications spanning atomistic, image-based, spectral, and textual data modalities. DL allows analysis of unstructured data and automated identification of features. The recent development of large materials databases has fueled the application of DL methods in atomistic prediction in particular. In contrast, advances in image and spectral data have largely leveraged synthetic data enabled by high-quality forward models as well as by generative unsupervised DL methods. In this article, we present a high-level overview of deep learning methods followed by a detailed discussion of recent developments of deep learning in atomistic simulation, materials imaging, spectral analysis, and natural language processing. For each modality we discuss applications involving both theoretical and experimental data, typical modeling approaches with their strengths and limitations, and relevant publicly available software and datasets. We conclude the review with a discussion of recent cross-cutting work related to uncertainty quantification in this field and a brief perspective on limitations, challenges, and potential growth areas for DL methods in materials science.

Low-temperature synthesis routes are necessary for selectively synthesizing many metastable solid-state materials. Here, we identify a cooperative effect that starting materials have in lowering temperatures in solid-state metathesis reactions by studying the formation of yttrium manganese oxide. Previous studies have shown that YMnO3 can be synthesized by ternary metathesis with an alkali halide being produced as a secondary product. In this contribution, we show that by using alkaline earth metals instead of alkali metals, the polymorph selectivity of the reaction is changed, as orthorhombic YMnO3 forms at lower temperatures than the hexagonal polymorph. Reactions were studied using ex post facto synchrotron X-ray diffraction. These experiments reveal that reactions using alkaline earth manganese oxides as a starting material require high temperatures to progress. Reaction temperatures can be lowered from 700 to 550 °C while maintaining phase selectivity by reacting both MgMn2O4 and CaMn2O4 with YOCl in a cooperative "cometathesis"reaction. The nascent halide salts appear to improve the reaction kinetics. Since the onset temperature for YMnO3 formation falls 50 °C below the MgCl2-CaCl2 liquidus, the enhanced reactivity is consistent with the surface melting of a nasscent salt byproduct at the interfaces. Cometathesis routes have similar phase selectivity and temperature reduction in reactions that form TbMnO3, ErMnO3, and DyMnO3. Cometathesis lowers reaction temperatures while preserving the reaction selectivity of the end members, making it a valuable approach for synthesizing metastable targets.

In the synthesis of complex oxides, solid-state metathesis provides low-temperature reactions where product selectivity can be achieved through simple changes in precursor composition. The influence of precursor structure, however, is less understood in solid-state synthesis. Here we present the ternary metathesis reaction (LiMnO₂ + YOCl → YMnO₃ + LiCl) to target two yttrium manganese oxide products, hexagonal and orthorhombic YMnO₃, when starting from three different LiMnO₂ precursors. Using temperature-dependent synchrotron X-ray and neutron diffraction, we identify the relevant intermediates and temperature regimes of reactions along the pathway to YMnO₃. Manganese-containing intermediates undergo a charge disproportionation into a reduced Mn(II,III) tetragonal spinel and oxidized Mn(III,IV) cubic spinel, which lead to hexagonal and orthorhombic YMnO₃, respectively. Density functional theory calculations confirm that the presence of Mn(IV) caused by a small concentration of cation vacancies (∼2.2%) in YMnO₃ stabilizes the orthorhombic polymorph over the hexagonal. Reactions over the course of 2 weeks yield o-YMnO₃ as the majority product at temperatures below 600 °C, which supports an equilibration of cation defects over time. Controlling the composition and structure of these defect-accommodating intermediates provides new strategies for selective synthesis of complex oxides at low temperatures.