Search

Time resolved measurements of the linear magneto-optical Kerr rotation reveal that, in the {\it 3d} ferromagnets Fe, Co and Ni, the amplitude of the demagnetization increases linearly with the fluence of the light. We rationalize this phenomenon as being linearly driven by the increase in temperature, the electron-phonon and electron-magnon scattering, and a reduction of the inter-atomic exchange. The amplitude of the latter phenomenon, which until the present study has been widely overlooked, is obtained through \textit{ab initio} density functional theory calculations, and is argued to be the principal source of demagnetization in Ni and Co, while still contributing largely in Fe.

Machine learning thermodynamic perturbation theory (MLPT) is a promising approach to compute finite temperature properties when the goal is to compare several different levels of ab initio theory and/or to apply highly expensive computational methods. Indeed, starting from a production molecular dynamics trajectory, this method can estimate properties at one or more target levels of theory from only a small number of additional fixed-geometry calculations, which are used to train a machine learning model. However, as MLPT is based on thermodynamic perturbation theory (TPT), inaccuracies might arise when the starting point trajectory samples a configurational space which has a small overlap with that of the target approximations of interest. By considering case studies of molecules adsorbed in zeolites and several different density functional theory approximations, in this work we assess the accuracy of MLPT for ensemble total energies and enthalpies of adsorption. The problematic cases that were found are analyzed and it is shown that, even without knowing exact reference results, pathological cases for MLPT can be detected by considering a coefficient that measures the statistical imbalance induced by the TPT reweighting. For the most pathological examples we recover target level results within chemical accuracy by applying a machine learning-based Monte Carlo (MLMC) resampling. Finally, based on the ideas developed in this work, we assess and confirm the accuracy of recently published MLPT-based enthalpies of adsorption at the random phase approximation level, whose high computational cost would completely hinder a direct molecular dynamics simulation.

Nowadays, the coupling of electronic structure and machine learning techniques serves as a powerful tool to predict chemical and physical properties of a broad range of systems. With the aim of improving the accuracy of predictions, a large number of representations for molecules and solids for machine learning applications has been developed. In this work we propose a novel descriptor based on the notion of molecular graph. While graphs are largely employed in classification problems in cheminformatics or bioinformatics, they are not often used in regression problem, especially of energy-related properties. Our method is based on a local decomposition of atomic environments and on the hybridization of two kernel functions: a graph kernel contribution that describes the chemical pattern and a Coulomb label contribution that 1encodes finer details of the local geometry. The accuracy of this new kernel method in energy predictions of molecular and condensed phase systems is demonstrated by considering the popular QM7 and BA10 datasets. These examples show that the hybrid localized graph kernel outperforms traditional approaches such as, for example, the smooth overlap of atomic positions (SOAP) and the Coulomb matrices.

We introduce the "selPT" perturbative approach, based on ab initio molecular dynamics (AIMD), for computing accurate finite-temperature properties by efficiently using correlated wave-function methods. We demonstrate the power of the method by computing prototypical molecular enthalpies of adsorption in zeolite (CH$_4$ and CO$_2$ on protonated chabazite at 300~K) using the random phase approximation. Results are in excellent agreement with experiment. The improved accuracy provided by selPT represents a crucial step towards the goal of truly quantitative AIMD prediction of experimental observables at finite temperature.

Stacking-dependent interlayer interactions are important for understanding the structural and electronic properties in incommensurable two dimensional material assemblies where long-range moir\'e patterns arise due to small lattice constant mismatch or twist angles. Here, we study the stacking-dependent interlayer coupling energies between graphene (G) and hexagonal boron nitride (BN) homo- and hetero-structures using high-level random-phase approximation (RPA) ab initio calculations. Our results show that although total binding energies within LDA and RPA differ substantially between a factor of 200%-400%, the energy differences as a function of stacking configuration yield nearly constant values with variations smaller than 20% meaning that LDA estimates are quite reliable. We produce phenomenological fits to these energy differences, which allows us to calculate various properties of interest including interlayer spacing, sliding energetics, pressure gradients and elastic coefficients to high accuracy. The importance of long-range interactions (captured by RPA but not LDA) on various properties is also discussed. Parameterisations for all fits are provided., Comment: 10 pages, 6 figures, 2 tables

By explicitly including fractionally ionic contributions to the polarizability of a many-component system we are able to significantly improve on previous atom-wise many-body van der Waals approaches with essentially no extra numerical cost. For non-ionic systems our method is comparable in accuracy to existing approaches. However, it offers substantial improvements in ionic solids, e.g. producing better polarizabilities by over 65% in some cases. It has particular benefits for two-dimensional transition metal dichalcogenides, and interactions of H$_2$ with modified coronenes - ionic systems of nanotechnological interest. It thus offers an efficient improvement on existing approaches, valid for a wide range of systems.

We present a readily computable semi-analytic Layer Response Theory (LRT) for analysis of cohesive energetics involving two-dimensional layers such as BN or graphene. The theory approximates the Random Phase Approximation (RPA) correlation energy. Its RPA character ensures that the energy has the correct van der Waals asymptotics for well-separated layers, in contrast to simple pairwise atom-atom theories which fail qualitatively for layers with zero electronic energy gap. At the same time our theory is much less computationally intensive than the full RPA energy. It also gives accurate correlation energies near to the binding minimum, in contrast to Lifshitz-type theory. We apply our LRT theory successfully to graphite and to BN, and to a graphene-BN heterostructure., Comment: 29 pp, 9 Figs. Phys. Rev. B, accepted

Determining the material properties of layered systems like graphite and bigraphene from \emph{ab initio} calculations is very difficult. This is mostly due to the complex van der Waals forces which help bind the layers. Recently, Chen~\emph{et al.} [Chen et al., Sci. Rep. 3, 3046 (2013)] reported a novel approach for extracting geometry dependent energetic properties of general, layered graphitic systems from periodic graphite calculations on AA, AB and ABC graphite. Unfortunately, their analysis suffered from a number of technical and theoretical flaws which make their results unreliable for predicting energetic properties. We propose that their conclusions in this regard should be reassessed, or reanalysed using more appropriate van der Waals theory.

We performed a systematic density functional study of the adsorption of copper, silver, and gold adatoms on graphene, especially accounting for van der Waals interactions by the vdW-DF and the PBE+D2 methods. In particular, we analyze the preferred adsorption site (among top, bridge, and hollow positions) together with the corresponding distortion of the graphene sheet and identify diffusion paths. Both vdW schemes show that the coinage metal atoms do bind to the graphene sheet and that in some cases the buckling of the graphene can be significant. The results for silver are at variance with those obtained with GGA, which gives no binding in this case. However, we observe some quantitative differences between the vdW-DF and the PBE+D2 methods. For instance the adsorption energies calculated with the PBE+D2 method are systematically higher than the ones obtained with vdW-DF. Moreover, the equilibrium distances computed with PBE+D2 are shorter than those calculated with the vdW-DF method.

We assess the performance of recent density functionals for the exchange-correlation energy of a nonmolecular solid, by applying accurate calculations with the GAUSSIAN, BAND, and VASP codes to a test set of 24 solid metals and non-metals. The functionals tested are the modified Perdew-Burke-Ernzerhof generalized gradient approximation (PBEsol GGA), the second-order GGA (SOGGA), and the Armiento-Mattsson 2005 (AM05) GGA. For completeness, we also test more-standard functionals: the local density approximation, the original PBE GGA, and the Tao-Perdew-Staroverov-Scuseria (TPSS) meta-GGA. We find that the recent density functionals for solids reach a high accuracy for bulk properties (lattice constant and bulk modulus). For the cohesive energy, PBE is better than PBEsol overall, as expected, but PBEsol is actually better for the alkali metals and alkali halides. For fair comparison of calculated and experimental results, we consider the zero-point phonon and finite-temperature effects ignored by many workers. We show how Gaussian basis sets and inaccurate experimental reference data may affect the rating of the quality of the functionals. The results show that PBEsol and AM05 perform somewhat differently from each other for alkali metal, alkaline earth metal and alkali halide crystals (where the maximum value of the reduced density gradient is about 2), but perform very similarly for most of the other solids (where it is often about 1). Our explanation for this is consistent with the importance of exchange-correlation nonlocality in regions of core-valence overlap., Comment: 32 pages, single pdf file

We have studied theoretically, using density functional theory, several materials properties when going from one C layer in graphene to two and three g raphene layers and on to graphite. The properties we have focused on are the elastic constants, electronic structure (energy bands and density of state s), and the dielectric properties. For any of the properties we have investigated the modification due to an increase in the number of graphene layers is within a few percent. Our results are in agreement with the analysis presented recently by Kopelevich and Esquinazi (unpublished).