Oxygen vacancies in complex oxides are indispensable for information and energy technologies. There are several means to create oxygen vacancies in bulk materials. However, the use of ionic interfaces to create oxygen vacancies has not been fully explored. Herein, we report an oxide nanobrush architecture designed to create high-density interfacial oxygen vacancies. An atomically well-defined (111) heterointerface between the fluorite CeO2 and the bixbyite Y2O3 is found to induce a charge modulation between Y3+ and Ce4+ ions enabled by the chemical valence mismatch between the two elements. Local structure and chemical analyses, along with theoretical calculations, suggest that more than 10% of oxygen atoms are spontaneously removed without deteriorating the lattice structure. Our fluorite–bixbyite nanobrush provides an excellent platform for the rational design of interfacial oxide architectures to precisely create, control, and transport oxygen vacancies critical for developing ionotronic and memristive devices for advanced energy and neuromorphic computing technologies.

If theory is to be able to predict the rates of catalytic reactions over extended ranges of temperature and pressure, it must provide accurate rate constants for elementary reaction steps, including both the activation energy and pre-exponential factor. A standing difficulty with this objective is the treatment of floppy modes in the partition function for the adsorbed species. This issue leads to limited accuracy in the pre-exponential factor computed for realistic systems. Here, we investigate the C−H bond breaking for a series of linear-chain alkoxides on Cu(110) using density functional theory because the results can be compared to the experimental data for the rate constants. The structural similarity of these species enables us to understand the systematic effect of the molecular size on the frustrated motions and pre-exponential factor. First, we discuss the complexities of finding the global minimum structure of the adsorbed species and highlight the high dimensionality of the configuration space to be sampled. Then, we analyze the motions of harmonic normal modes, including the motions of the underlying metal atoms, and compute the harmonic pre-exponential factors. To account for periodic frustrated rotations, we use the hindered rotor model: this motion significantly decreases the pre-exponential factor in the C−H bond breaking, the effect increasing with the molecular size. We also estimate the anharmonic effect using the Morse treatment of potentials. The activation energy and preexponential factor computed for CH3O are in excellent quantitative agreement with the experiment. The trends computed for the homologous series of alcohols are also reflected by the experiment.

The restructuring of interfaces plays a crucial role in materials science and heterogeneous catalysis. Bimetallic systems, in particular, often adopt very different compositions and morphologies at surfaces compared to the bulk. For the first time, we reveal a detailed atomistic picture of long-time scale restructuring of Pd deposited on Ag using microscopy, spectroscopy, and novel simulation methods. By developing and performing accelerated machine-learning molecular dynamics followed by an automated analysis method, we discover and characterize previously unidentified surface restructuring mechanisms in an unbiased fashion, including Pd−Ag place exchange and Ag pop-out as well as step ascent and descent. Remarkably, layer-by-layer dissolution of Pd into Ag is always preceded by an encapsulation of Pd islands by Ag, resulting in a significant migration of Ag out of the surface and a formation of extensive vacancy pits within a period of microseconds. These metastable structures are of vital catalytic importance, as Ag-encapsulated Pd remains much more accessible to reactants than bulk-dissolved Pd. Our approach is broadly applicable to complex multimetallic systems and enables the previously intractable mechanistic investigation of restructuring dynamics at atomic resolution.

We present a way to dramatically accelerate Gaussian process models for interatomic force fields based on many-body kernels by mapping both forces and uncertainties onto functions of lowdimensional features. This allows for automated active learning of models combining near-quantum accuracy, built-in uncertainty, and constant cost of evaluation that is comparable to classical analytical models, capable of simulating millions of atoms. Using this approach, we perform large scale molecular dynamics simulations of the stability of the stanene monolayer. We discover an unusual phase transformation mechanism of 2D stanene, where ripples lead to nucleation of bilayer defects, densification into a disordered multilayer structure, followed by formation of bulk liquid at high temperature or nucleation and growth of the 3D bcc crystal at low temperature. The presented method opens possibilities for rapid development of fast accurate uncertainty-aware models for simulating long-time large-scale dynamics of complex materials.

Computing accurate reaction rates is a central challenge in computational chemistry and biology because of the high cost of free energy estimation with unbiased molecular dynamics. In this work, a data-driven machine learning algorithm is devised to learn collective variables with a multitask neural network, where a common upstream part reduces the high dimensionality of atomic configurations to a low dimensional latent space, and separate downstream parts map the latent space to predictions of basin class labels and potential energies. The resulting latent space is shown to be an effective low-dimensional representation, capturing the reaction progress and guiding effective umbrella sampling to obtain accurate free energy landscapes. This approach is successfully applied to model systems including a 5D M\textbackslash"uller Brown model, a 5D three-well model, and alanine dipeptide in vacuum. This approach enables automated dimensionality reduction for energy controlled reactions in complex systems, offers a unified framework that can be trained with limited data, and outperforms single-task learning approaches, including autoencoders.

Machine learned force fields typically require manual construction of training sets consisting of thousands of first principles calculations, which can result in low training efficiency and unpredictable errors when applied to structures not represented in the training set of the model. This severely limits the practical application of these models in systems with dynamics governed by important rare events, such as chemical reactions and diffusion. We present an adaptive Bayesian inference method for automating the training of interpretable, low-dimensional, and multi-element interatomic force fields using structures drawn on the fly from molecular dynamics simulations. Within an active learning framework, the internal uncertainty of a Gaussian process regression model is used to decide whether to accept the model prediction or to perform a first principles calculation to augment the training set of the model. The method is applied to a range of single- and multi-element systems and shown to achieve a favorable balance of accuracy and computational efficiency, while requiring a minimal amount of ab initio training data. We provide a fully open-source implementation of our method, as well as a procedure to map trained models to computationally efficient tabulated force fields.

The effects of strain on the ionic conductivity of rare-earth substituted CeO2 have been extensively studied, but the results have been inconsistent and focused upon the ‘optimised’ conductors such as Gd or Sm substituted CeO2 where defect association is minimised. By thermally annealing epitaxial films deposited by pulsed laser deposition, we varied the strain systematically, whilst avoiding any influence from interfacial or grain boundary effects. The activation energy of the in-plane conductivity was found to increase with increasing compressive biaxial strain, which was quantitatively in excellent agreement with previous computational and experimental studies. These results provide a much needed quantitative consensus on the effects of lattice strain on ionic transport. Furthermore, we demonstrate that the change in the activation energy for Yb-substituted CeO2 is around three times that for Gd or La substitutions for the same applied strain, indicating the important role played by defect association. These results have significant implications for ionic transport at reduced or ambient temperatures, where changes in conductivity due to strain may be several orders of magnitude larger for ‘non-optimised’ conductors compared with ‘optimised’ conductors. We rationalise our results by considering the defect-defect interactions in these materials and through force-field calculations.