Deep learning provides an efficient cross-validation tool for crystallography with no preferred orientation or magnification.

Observation of growth, synthesis, dynamics, and electrochemical reactions in the liquid state is an important yet largely unstudied aspect of nanotechnology. The only techniques that can potentially provide the insights necessary to advance our understanding of these mechanisms is simultaneous atomic-scale imaging and quantitative chemical analysis (through spectroscopy) under environmental conditions in the transmission electron microscope. In this study we describe the experimental and technical conditions necessary to obtain electron energy loss (EEL) spectra from a nanoparticle in colloidal suspension using aberration-corrected scanning transmission electron microscopy (STEM) combined with the environmental liquid stage. At a fluid path length below 400 nm, atomic resolution images can be obtained and simultaneous compositional analysis can be achieved. We show that EEL spectroscopy can be used to quantify the total fluid path length around the nanoparticle and demonstrate that characteristic core-loss signals from the suspended nanoparticles can be resolved and analyzed to provide information on the local interfacial chemistry with the surrounding environment. The combined approach using aberration-corrected STEM and EEL spectra with the in situ fluid stage demonstrates a plenary platform for detailed investigations of solution-based catalysis.

We observe interfacial ferromagnetism in superlattices of the paramagnetic metal ${\mathrm{LaNiO}}_3$ and the antiferromagnetic insulator ${\mathrm{CaMnO}}_3$. ${\mathrm{LaNiO}}_3$ exhibits a thickness dependent metal-insulator transition and we find the emergence of ferromagnetism to be coincident with the conducting state of ${\mathrm{LaNiO}}_3$. That is, only superlattices in which the ${\mathrm{LaNiO}}_3$ layers are metallic exhibit ferromagnetism. Using several magnetic probes, we have determined that the ferromagnetism arises in a single unit cell of ${\mathrm{CaMnO}}_3$ at the interface. Together these results suggest that ferromagnetism can be attributed to a double exchange interaction among Mn ions mediated by the adjacent itinerant metal.

Organic–inorganic perovskites have emerged as an important class of next generation solar cells due to their remarkably low cost, band gap, and sub-900 nm absorption onset. Here, we show a series of in situ observations inside electron microscopes and X-ray diffractometers under device-relevant synthesis conditions focused on revealing the crystallization process of the formamidinium lead-triiodide perovskite at the optimum temperature of 175 °C. Direct in situ observations of the structure and chemistry over relevant spatial, temporal, and temperature scales enabled identification of key perovskite formation and degradation mechanisms related to grain evolution and interface chemistry. The lead composition was observed to fluctuate at grain boundaries, indicating a mobile lead-containing species, a process found to be partially reversible at a key temperature of 175 °C. Using low energy electron microscopy and valence electron energy loss spectroscopy, lead is found to be bonded in the grain interior with iodine in a tetrahedral configuration. At the grain boundaries, the binding energy associated with lead is consequently shifted by nearly 2 eV and a doublet peak is resolved due presumably to a greater degree of hybridization and the potential for several different bonding configurations. At the grain boundaries there is adsorption of hydrogen and OH⁻ ions as a result of residual water vapor trapped as a non-crystalline material during formation. Insights into the relevant formation and decomposition reactions of formamidinium lead iodide at low to high temperatures, observed metastabilities, and relationship with the photovoltaic performance were obtained and used to optimize device processing resulting in conversion efficiencies of up to 17.09% within the stability period of the devices.