Evaluating the (dis)similarity of crystalline, disordered and molecular compounds is a critical step in the development of algorithms to navigate automatically the configuration space of complex materials. For instance, a structural similarity metric is crucial for classifying structures, searching chemical space for better compounds and materials, and driving the next generation of machine-learning techniques for predicting the stability and properties of molecules and materials. In the last few years several strategies have been designed to compare atomic coordination environments. In particular, the Smooth Overlap of Atomic Positions (SOAP) has emerged as an elegant framework to obtain translation, rotation and permutation-invariant descriptors of groups of atoms, driven by the design of various classes of machine-learned inter-atomic potentials. Here we discuss how one can combine such local descriptors using a Regularized Entropy Match (REMatch) approach to describe the similarity of both whole molecular and bulk periodic structures, introducing powerful metrics that enable the navigation of alchemical and structural complexity within a unified framework. Furthermore, using this kernel and a ridge regression method we can predict atomization energies for a database of small organic molecules with a mean absolute error below 1kcal/mol, reaching an important milestone in the application of machine-learning techniques to the evaluation of molecular properties.
We study the electrodynamics of Dirac electrons in solids (e.g., bismuth) by comparing it with quantum electrodynamics (QED). It is shown that Lorentz covariance associated with the Dirac electrons in solids results in a remarkable correlation between the dielectric and diamagnetic properties, leading to a significant enhancement in the permittivity directly linked to the well-known phenomenon of large diamagnetism.
Absolute binding energies of core electrons in molecules and bulk materials can be efficiently calculated by spin paired density-function theory employing a $Δ$ Kohn-Sham ($Δ$KS) scheme corrected by offsets that are highly transferable. These offsets depend on core level and atomic species and can be determined by comparing $Δ$KS energies to experimental molecular X-ray photoelectron spectra. We demonstrate the correct prediction of absolute and relative binding energies on a wide range of molecules, metals and insulators.
We present molecular dynamics simulations of solid NaNO$_2$ using pair potentials with the rigid-ion model. The crystal potential surface is calculated by using an \emph{a priori} method which integrates the \emph{ab initio} calculations with the Gordon-Kim electron gas theory. This approach is carefully examined by using different population analysis methods and comparing the intermolecular interactions resulting from this approach with those from the \emph{ab initio} Hartree-Fock calculations. Our numerics shows that the ferroelectric-paraelectric phase transition in solid NaNO$_2$ is trigge…
The thermal conductivity of nanometric objects or nanostructured materials can be determined using non equilibrium molecular dynamics (NEMD) simulations. The technique is simple in its principle, and resembles a numerical guarded hot plate experiment. The 'sample' is placed between a hot source and a cold source consisting of thermostatted sets of atoms. The thermal conductivity is obtained from the heat flux crossing the sample and the temperature profile in the system. Simulations results, however, exhibit a strong dependence of the thermal conductivity on the sample size. In this paper, we …
We present an accurate computational study of the electronic structure and lattice dynamics of solid molecular hydrogen at high pressure. The band-gap energies of the $C2/c$, $Pc$, and $P6_3/m$ structures at pressures of 250, 300, and 350 GPa are calculated using the diffusion quantum Monte Carlo (DMC) method. The atomic configurations are obtained from ab-initio path-integral molecular dynamics (PIMD) simulations at 300 K and 300 GPa to investigate the impact of zero-point energy and temperature-induced motion of the protons including anharmonic effects. We find that finite temperature and nu…
Establishing the phase diagram of hydrogen is a major challenge for experimental and theoretical physics. Experiment alone cannot establish the atomic structure of solid hydrogen at high pressure, because hydrogen scatters X-rays only weakly. Instead our understanding of the atomic structure is largely based on density functional theory (DFT). By comparing Raman spectra for low-energy structures found in DFT searches with experimental spectra, candidate atomic structures have been identified for each experimentally observed phase. Unfortunately, DFT predicts a metallic structure to be energeti…
This work presents a novel physics-informed deep learning based super-resolution framework to reconstruct high-resolution deformation fields from low-resolution counterparts, obtained from coarse mesh simulations or experiments. We leverage the governing equations and boundary conditions of the physical system to train the model without using any high-resolution labeled data. The proposed approach is applied to obtain the super-resolved deformation fields from the low-resolution stress and displacement fields obtained by running simulations on a coarse mesh for a body undergoing linear elastic…
Here we report on a novel High Harmonic Generation (HHG) light source designed for space charge free ultrafast photoelectron spectroscopy (PES) on solids. The ultimate overall energy resolution achieved on a polycrystalline Au sample is ~22 meV at 40 K. These results have been obtained at a photon energy of 16.9 eV with a pulse bandwidth of ~19 meV, by varying, up to 200 kHz, the photon pulses repetition rate and the photon fluence on the sample. These features set a new benchmark for tunable narrowband HHG sources. By comparing the PES energy resolution and the photon pulse bandwidth with a p…
Much of the discussion in the literature of the low frequency part of the density of states of amorphous solids was dominated for years by comparing measured or simulated density of states to the classical Debye model. Since this model is hardly appropriate for the materials at hand, this created some amount of confusion regarding the existence and universality of the so- called ``Boson Peak'' which results from such comparisons. We propose that one should pay attention to the different roles played by different aspects of disorder, the first being disorder in the interaction strengths, the se…
The performance of density functional theory (DFT) approximations for predicting materials thermodynamics is typically assessed by comparing calculated and experimentally determined enthalpies of formation from elemental phases, ΔHf. However, a compound competes thermodynamically with both other compounds and their constituent elemental forms, and thus, the enthalpies of the decomposition reactions to these competing phases, ΔHd, determines thermodynamic stability. We evaluated the phase diagrams for 56,791 compounds to classify decomposition reactions into three types: 1. those that produce e…
Isotopic effects are relevant to understand several properties of solids, and have been thoroughly analyzed along the years. These effects may depend on the dimensionality of the considered solid. Here we assess their magnitude for structural properties of graphene, a paradigmatic two-dimensional material. We use path-integral molecular dynamics simulations, a well-suited technique to quantify the influence of nuclear quantum effects on equilibrium variables, especially in cases where anharmonic effects are important. Emphasis is put on interatomic distances and mean-square displacements, as w…
Fast and accurate quasi-atom method for simultaneous atomistic and continuum simulation of solids
Skolkovo Institute of Science and Technology（斯克尔科沃科学与技术研究所）；Schmidt Institute of Physics of the Earth of the Russian Academy of Sciences B. Gruzinskaya str., 10, bld. 1, Moscow, 123242, Russia（俄罗斯科学院地球物理施密特研究所）；School of Computing and Mathematical Sciences（计算与数学科学学院）；University of Leicester（莱斯特大学）
The prospect of efficient solid state refrigeration at room temperature is driving research into magnetic cooling engine design and magnetic phase transition-based refrigerants. In this Viewpoint an Ashby-style map of magnetic refrigerant properties is constructed, comparing popular materials with limits derived from an idealised first order transition model. This comparison demonstrates the potential for new magnetocaloric material systems to be established through structural control and optimisation at the atomic-, nano- and micro-scale.
