arXiv subjects

Miguel A. Caro

Publications and source records attributed to Miguel A. Caro.

36 records · Page 2Linked to original sources

arXiv2023

Quantum-corrected thickness-dependent thermal conductivity in amorphous silicon predicted by machine-learning molecular dynamics simulations

Amorphous silicon (a-Si) is an important thermal-management material and also serves as an ideal playground for studying heat transport in strongly disordered materials. Theoretical prediction of the thermal conductivity of a-Si in a wide range of temperatures and sample sizes is still a challenge. Herein we present a systematic investigation of the thermal transport properties of a-Si by employing large-scale molecular dynamics (MD) simulations with an accurate and efficient machine-learned neuroevolution potential (NEP) trained against abundant reference data calculated at the quantum-mechanical density-functional-theory level. The high efficiency of NEP allows us to study the effects of finite size and quenching rate in the formation of a-Si in great detail. We find that it requires a simulation cell up to $64,000$ atoms (a cubic cell with a linear size of 11 nm) and a quenching rate down to $10^{11}$ K s$^{-1}$ for fully convergent thermal conductivity. Structural properties, including short- and medium-range order as characterized by the pair correlation function, angular distribution function, coordination number, ring statistics and structure factor are studied to demonstrate the accuracy of NEP and to further evaluate the role of quenching rate. Using both the heterogeneous and the homogeneous nonequilibrium MD methods and the related spectral decomposition techniques, we calculate the temperature- and thickness-dependent thermal conductivity values of a-Si and show that they agree well with available experimental results from 10 K to room temperature. Our results also highlight the importance of quantum effects in the calculated thermal conductivity and support the quantum correction method based on the spectral thermal conductivity.

cond-mat.mtrl-sci↗

arXiv2022

Accurate computational prediction of core-electron binding energies in carbon-based materials: A machine-learning model combining density-functional theory and $\boldsymbol{GW}$

We present a quantitatively accurate machine-learning (ML) model for the computational prediction of core-electron binding energies, from which x-ray photoelectron spectroscopy (XPS) spectra can be readily obtained. Our model combines density functional theory (DFT) with $GW$ and uses kernel ridge regression for the ML predictions. We apply the new approach to materials and molecules containing carbon, hydrogen and oxygen, and obtain qualitative and quantitative agreement with experiment, resolving spectral features within 0.1 eV of reference experimental spectra. The method only requires the user to provide a structural model for the material under study to obtain an XPS prediction within seconds. Our new tool is freely available online through the XPS Prediction Server.

cond-mat.mtrl-sci↗

arXiv2022

Reassignment of magic numbers for icosahedral Au clusters: 310, 564, 928 and 1426

Icosahedral Au clusters with three and four shells of atoms are found to deviate significantly from the commonly assumed Mackay structures. By introducing additional atoms in the surface shell and creating a vacancy in the center of the cluster, the calculated energy per atom can be lowered significantly, according to several different descriptions of the interatomic interaction. Analogous icosahedral structures with five and six shells of atoms are generated using the same structural motifs and are similarly found to be more stable than Mackay icosahedra. The lowest energy per atom is obtained with clusters containing 310, 564, 928 and 1426 atoms, as compared with the commonly assumed magic numbers of 309, 561, 923 and 1415. Some of the vertices in the optimized clusters have a hexagonal ring of atoms, rather than a pentagon, with the vertex atom missing. An inner shell atom in some cases moves outwards by more than an Ångström into the surface shell at the vertex site. This feature, as well as the wide distribution of nearest-neighbor distances in the surface layer, can strongly influence the catalytic properties of icosahedral clusters. The structural optimization is initially carried out using the GOUST method with atomic forces estimated with the EMT empirical potential function, but the atomic coordinates are then refined by minimization using electron density functional theory (DFT) or Gaussian approximation potential (GAP). A single energy barrier is found to separate the Mackay icosahedron from a lower energy structure where a string of atoms moves outwards in a concerted manner from the center so as to create a central vacancy while placing an additional atom in the surface shell.

physics.comp-ph↗

arXiv2021

Structure and Pore Size Distribution in Nanoporous Carbon

We study the structural and mechanical properties of nanoporous (NP) carbon materials by extensive atomistic machine-learning (ML) driven molecular dynamics (MD) simulations. To this end, we retrain a ML Gaussian approximation potential (GAP) for carbon by recalculating the a-C structural database of Deringer and Csányi [Phys. Rev. B 2017, 95, 094203] adding van der Waals interactions. Our GAP enables a notable speedup and improves the accuracy of energy and force predictions. We use the GAP to thoroughly study the atomistic structure and pore-size distribution in computational NP carbon samples. These samples are generated by a melt-graphitization-quench MD procedure over a wide range of densities (from 0.5 to 1.7 g/cm$^3$) with structures containing 131,072 atoms. Our results are in good agreement with experimental data for the available observables, and provide a comprehensive account of structural (radial and angular distribution functions, motif and ring counts, X-ray diffraction patterns, pore characterization) and mechanical (elastic moduli and their evolution with density) properties. Our results show relatively narrow pore-size distributions, where the peak position and width of the distributions are dictated by the mass density of the materials. Our data allow further work on computational characterization of NP carbon materials, in particular for energy-storage applications, as well as suggest future experimental characterization of NP carbon-based materials.

cond-mat.mtrl-sci↗

arXiv2021

Machine learning force fields based on local parametrization of dispersion interactions: Application to the phase diagram of C$_{60}$

We present a comprehensive methodology to enable addition of van der Waals (vdW) corrections to machine learning (ML) atomistic force fields. Using a Gaussian approximation potential (GAP) [Bartók et al., Phys. Rev. Lett. 104, 136403 (2010)] as baseline, we accurately machine learn a local model of atomic polarizabilities based on Hirshfeld volume partitioning of the charge density [Tkatchenko and Scheffler, Phys. Rev. Lett. 102, 073005 (2009)]. These environment-dependent polarizabilities are then used to parametrize a screened London-dispersion approximation to the vdW interactions. Our ML vdW model only needs to learn the charge density partitioning implicitly, by learning the reference Hirshfeld volumes from density functional theory (DFT). In practice, we can predict accurate Hirshfeld volumes from the knowledge of the local atomic environment (atomic positions) alone, making the model highly computationally efficient. For additional efficiency, our ML model of atomic polarizabilities reuses the same many-body atomic descriptors used for the underlying GAP learning of bonded interatomic interactions. We also show how the method enables straightforward computation of gradients of the observables, even when these remain challenging for the reference method (e.g., calculating gradients of the Hirshfeld volumes in DFT). Finally, we demonstrate the approach by studying the phase diagram of C$_{60}$, where vdW effects are important. The need for a highly accurate vdW-inclusive reactive force field is highlighted by modeling the decomposition of the C$_{60}$ molecules taking place at high pressures and temperatures.

cond-mat.mtrl-sci↗

arXiv2020

Particle Swarm Based Hyper-Parameter Optimization for Machine Learned Interatomic Potentials

Modeling non-empirical and highly flexible interatomic potential energy surfaces (PES) using machine learning (ML) approaches is becoming popular in molecular and materials research. Training an ML-PES is typically performed in two stages: feature extraction and structure-property relationship modeling. The feature extraction stage transforms atomic positions into a symmetry-invariant mathematical representation. This representation can be fine-tuned by adjusting on a set of so-called "hyper-parameters" (HPs). Subsequently, an ML algorithm such as neural networks or Gaussian process regression (GPR) is used to model the structure-PES relationship based on another set of HPs. Choosing optimal values for the two sets of HPs is critical to ensure the high quality of the resulting ML-PES model. In this paper, we explore HP optimization strategies tailored for ML-PES generation using a custom-coded parallel particle swarm optimizer (available freely at https://github.com/suresh0807/PPSO.git). We employ the smooth overlap of atomic positions (SOAP) descriptor in combination with GPR-based Gaussian approximation potentials (GAP) and optimize HPs for four distinct systems: a toy C dimer, amorphous carbon, $α$-Fe, and small organic molecules (QM9 dataset). We propose a two-step optimization strategy in which the HPs related to the feature extraction stage are optimized first, followed by the optimization of the HPs in the training stage. This strategy is computationally more efficient than optimizing all HPs at the same time by means of significantly reducing the number of ML models needed to be trained to obtain the optimal HPs. This approach can be trivially extended to other combinations of descriptor and ML algorithm and brings us another step closer to fully automated ML-PES generation.

cond-mat.mtrl-sci↗

arXiv2020

Machine learning driven simulated deposition of carbon films: from low-density to diamondlike amorphous carbon

Amorphous carbon (a-C) materials have diverse interesting and useful properties, but the understanding of their atomic-scale structures is still incomplete. Here, we report on extensive atomistic simulations of the deposition and growth of a-C films, describing interatomic interactions using a machine learning (ML) based Gaussian Approximation Potential (GAP) model. We expand widely on our initial work [Phys. Rev. Lett. 120, 166101 (2018)] by now considering a broad range of incident ion energies, thus modeling samples that span the entire range from low-density ($sp^{2}$-rich) to high-density ($sp^{3}$-rich, "diamond-like") amorphous forms of carbon. Two different mechanisms are observed in these simulations, depending on the impact energy: low-energy impacts induce $sp$- and $sp^{2}$-dominated growth directly around the impact site, whereas high-energy impacts induce peening. Furthermore, we propose and apply a scheme for computing the anisotropic elastic properties of the a-C films. Our work provides fundamental insight into this intriguing class of disordered solids, as well as a conceptual and methodological blueprint for simulating the atomic-scale deposition of other materials with ML-driven molecular dynamics.

cond-mat.mtrl-sci↗

arXiv2019

Optimizing many-body atomic descriptors for enhanced computational performance of machine learning based interatomic potentials

We explore different ways to simplify the evaluation of the smooth overlap of atomic positions (SOAP) many-body atomic descriptor [Bartók et al., Phys. Rev. B 87, 184115 (2013)]. Our aim is to improve the computational efficiency of SOAP-based similarity kernel construction. While these improved atomic descriptors can be used for general characterization and interpolation of atomic properties, their main target application is accelerated evaluation of machine-learning-based interatomic potentials within the Gaussian approximation potential (GAP) framework [Bartók et al., Phys. Rev. Lett. 104, 136403 (2010)]. We achieve this objective by expressing the atomic densities in an approximate separable form, which decouples the radial and angular channels. We then express the elements of the SOAP descriptor (i.e., the expansion coefficients for the atomic densities) in analytical form given a particular choice of radial basis set. Finally, we derive recursion formulas for the expansion coefficients. This new SOAP-based descriptor allows for tenfold speedups compared to previous implementations, while improving the stability of the radial expansion for distant atomic neighbors, without degradation of the interpolation power of GAP models.

physics.comp-ph↗

arXiv2019

Fully analytic valence force field model for the elastic and inner elastic properties of diamond and zincblende crystals

Using a valence force field model based on that introduced by Martin, we present three related methods through which we analytically determine valence force field parameters. The methods introduced allow easy derivation of valence force field parameters in terms of the Kleinman parameter $ζ$ and bulk properties of zincblende and diamond crystals. We start with a model suited for covalent and weakly ionic materials, where the valence force field parameters are derived in terms of $ζ$ and the bulk elastic constants $C_{11}$, $C_{12}$, and $C_{44}$. We show that this model breaks down as the material becomes more ionic and specifically when the elastic anisotropy factor $A = 2C_{44}/(C_{11}-C_{12}) > 2$. The analytic model can be stabilised for ionic materials by including Martin's electrostatic terms with effective cation and anion charges in the valence force field model. Inclusion of effective charges determined via the optical phonon mode splitting provides a stable model for all but two of the materials considered (zincblende GaN and AlN). A stable model is obtained for all materials considered by also utilising the inner elastic constant $E_{11}$ to determine the magnitude of the effective charges used in the Coulomb interaction. Test calculations show that the models describe well structural relaxation in superlattices and alloys, and reproduce key phonon band structure features.

cond-mat.mtrl-sci↗

arXiv2019

Towards the computational experiment

We give a brief account of the current limitations and possibilities in the field of computational simulation of materials. We then focus on the effect that the emergence of machine learning interatomic potentials is having on the field and how it will affect its evolution in the near future.

cond-mat.mtrl-sci↗

arXiv2018

Hybrid functional study of non-linear elasticity and internal strain in zincblende III-V materials

We investigate the elastic properties of selected zincblende III-V semiconductors. Using hybrid functional density functional theory we calculate the second and third order elastic constants, and first and second-order internal strain tensor components for Ga, In and Al containing III-V compounds. For many of these parameters, there are no available experimental measurements, and this work is the first to predict their values. The stricter convergence criteria for the calculation of higher order elastic constants are demonstrated, and arguments are made based on this for extracting these constants via the calculated stresses, rather than the energies, in the context of plane-wave-based calculations. The calculated elastic properties are used to determine the strain regime at which higher order elasticity becomes important by comparing the stresses predicted by a lower and a higher order elasticity theory. Finally, the results are compared with available experimental literature data and previous theory.

physics.comp-ph↗

arXiv2018

Growth Mechanism and Origin of High $sp^3$ Content in Tetrahedral Amorphous Carbon

We study the deposition of tetrahedral amorphous carbon (ta-C) films from molecular dynamics simulations based on a machine-learned interatomic potential trained from density-functional theory data. For the first time, the high $sp^3$ fractions in excess of 85% observed experimentally are reproduced by means of computational simulation, and the deposition energy dependence of the film's characteristics is also accurately described. High confidence in the potential and direct access to the atomic interactions allow us to infer the microscopic growth mechanism in this material. While the widespread view is that ta-C grows by "subplantation," we show that the so-called "peening" model is actually the dominant mechanism responsible for the high $sp^3$ content. We show that pressure waves lead to bond rearrangement away from the impact site of the incident ion, and high $sp^3$ fractions arise from a delicate balance of transitions between three- and fourfold coordinated carbon atoms. These results open the door for a microscopic understanding of carbon nanostructure formation with an unprecedented level of predictive power.

cond-mat.mtrl-sci↗

arXiv2017

Parametrization of the Tkatchenko-Scheffler dispersion correction scheme for popular exchange-correlation density functionals: effect on the description of liquid water

We present a list of optimized damping range parameters $s_R$ to be used with the Tkatchenko-Scheffler van der Waals dispersion-correction scheme [Phys. Rev. Lett. 102, 073005 (2009)]. The optimal $s_R$ are obtained for seven popular generalized-gradient approximation exchange-correlation density functionals: PBE, RPBE, revPBE, PBEsol, BLYP, AM05 and PW91. The optimization is carried out in the standard way by minimizing the mean absolute error of the S22 test set, where the reference interaction energies are taken from coupled-cluster calculations. With the optimized range parameters, we assess the impact of van der Waals corrections on the ability of these functionals to accurately describe structural and thermodynamic properties of liquid water: radial distribution functions, self-diffusion coefficients and standard molar entropies.

cond-mat.mtrl-sci↗

arXiv2016

Random alloy fluctuations and structural inhomogeneities in $c$-plane In$_{x}$Ga$_{1-x}$N quantum wells: theory of ground and excited electron and hole states

We present a detailed theoretical analysis of the electronic structure of $c$-plane InGaN/GaN quantum wells with indium contents varying between 10\% and 25\%. The electronic structure of the quantum wells is treated by means of an atomistic tight-binding model, accounting for variations in strain and built-in field due to random alloy fluctuations. Our analysis reveals strong localisation effects in the hole states. These effects are found not only in the ground states, but also the excited states. We conclude that localisation effects persist to of order 100~meV into the valence band, for as little as 10\% indium in the quantum well, giving rise to a significant density of localised states. We find, from an examination of the modulus overlap of the wave functions, that the hole states can be divided into three regimes of localisation. Our results also show that localisation effects due to random alloy fluctuations are far less pronounced for electron states. However, the combination of electrostatic built-in field, alloy fluctuations and structural inhomogeneities, such as well-width fluctuations, can nevertheless lead to significant localisation effects in the electron states, especially at higher indium contents. Overall, our results are indicative of individually localised electron and hole states, consistent with the experimentally proposed explanation of time-dependent photoluminescence results in $c$-plane InGaN/GaN QWs.

cond-mat.mtrl-sci↗

arXiv2015

Origin of non-linear piezoelectricity in III-V semiconductors: Internal strain and bond ionicity from hybrid-functional density functional theory

We derive first- and second-order piezoelectric coefficients for the zinc-blende III-V semiconductors, {Al,Ga,In}-{N,P,As,Sb}. The results are obtained within the Heyd-Scuseria-Ernzerhof hybrid-functional approach in the framework of density functional theory and the Berry-phase theory of electric polarization. To achieve a meaningful interpretation of the results, we build an intuitive phenomenological model based on the description of internal strain and the dynamics of the electronic charge centers. We discuss in detail first- and second-order internal strain effects, together with strain-induced changes in ionicity. This analysis reveals that the relatively large importance in the III-Vs of non-linear piezoelectric effects compared to the linear ones arises because of a delicate balance between the ionic polarization contribution due to internal strain relaxation effects, and the contribution due to the electronic charge redistribution induced by macroscopic and internal strain.

cond-mat.mtrl-sci↗

arXiv2015

Atomistic analysis of the impact of alloy and well-width fluctuations on the electronic and optical properties of InGaN/GaN quantum wells

We present an atomistic description of the electronic and optical properties of $\text{In}_{0.25}\text{Ga}_{0.75}$N/GaN quantum wells. Our analysis accounts for fluctuations of well width, local alloy composition, strain and built-in field fluctuations as well as Coulomb effects. We find a strong hole and much weaker electron wave function localization in InGaN random alloy quantum wells. The presented calculations show that while the electron states are mainly localized by well-width fluctuations, the holes states are already localized by random alloy fluctuations. These localization effects affect significantly the quantum well optical properties,leading to strong inhomogeneous broadening of the lowest interband transition energy. Our results are compared with experimental literature data.

cond-mat.mes-hall↗

arXiv2014

Extended scheme for the projection of material tensors of arbitrary symmetry onto a higher symmetry tensor

I propose a straightforward generalization of the projection scheme for elastic tensors introduced by Moakher and Norris [J. Elasticity 85, 215 (2006)] that takes into account also rotations. The "closest" tensor of any desired symmetry to the original tensor of lower symmetry is "closer" in this generalized scheme. The method has an important application in the context of the special quasirandom structure (SQS) method for the computational modeling of alloys, whereby the supercell's symmetry, and therefore that of the tensors representing its properties, is reduced with respect to the material's underlying symmetry. The approach allows to extract the tensor components most representative of the macroscopic symmetry of the material. Although the approach is general, in the present case I apply it to the elastic tensor and give numerical examples. Simple approximate analytical expressions for cubic materials are also provided.

cond-mat.mtrl-sci↗

arXiv2013

Theory of local electric polarization and its relation to internal strain: impact on the polarization potential and electronic properties of group-III nitrides

We present a theory of local electric polarization in crystalline solids and apply it to study the case of wurtzite group-III nitrides. We show that a local value of the electric polarization, evaluated at the atomic sites, can be cast in terms of a summation over nearest-neighbor distances and Born effective charges. Within this model, the local polarization shows a direct relation to internal strain and can be expressed in terms of internal strain parameters. The predictions of the present theory show excellent agreement with a formal Berry phase calculation for random distortions of a test-case CuPt-like InGaN alloy and InGaN supercells with randomly placed cations. While the present level of theory is appropriate for highly ionic compounds, we show that a more complex model is needed for less ionic materials, in which the strain dependence of Born effective charges has to be taken into account. Moreover, we provide ab initio parameters for GaN, InN and AlN, including hybrid functional values for the piezoelectric coefficients and the spontaneous polarization, which we use to accurately implement the local theory expressions. In order to calculate the local polarization potential, we also present a point dipole method. This method overcomes several limitations related to discretization and resolution which arise when obtaining the local potential by solving Poisson's equation on an atomic grid. Finally, we perform tight-binding supercell calculations to assess the impact of the local polarization potential arising from alloy fluctuations on the electronic properties of InGaN alloys. In particular, we find that the large upward bowing with composition of the InGaN valence band edge is strongly influenced by local polarization effects. Furthermore, our analysis allows us to extract composition-dependent bowing parameters for the energy gap and valence and conduction band edges.

cond-mat.mtrl-sci↗