Search

Coherent X-ray scattering (CXS) techniques are capable of interrogating dynamics of nano- to mesoscale materials systems at time scales spanning several orders of magnitude. However, obtaining accurate theoretical descriptions of complex dynamics is often limited by one or more factors -- the ability to visualize dynamics in real space, computational cost of high-fidelity simulations, and effectiveness of approximate or phenomenological models. In this work, we develop a data-driven framework to uncover mechanistic models of dynamics directly from time-resolved CXS measurements without solving the phase reconstruction problem for the entire time series of diffraction patterns. Our approach uses neural differential equations to parameterize unknown real-space dynamics and implements a computational scattering forward model to relate real-space predictions to reciprocal-space observations. This method is shown to recover the dynamics of several computational model systems under various simulated conditions of measurement resolution and noise. Moreover, the trained model enables estimation of long-term dynamics well beyond the maximum observation time, which can be used to inform and refine experimental parameters in practice. Finally, we demonstrate an experimental proof-of-concept by applying our framework to recover the probe trajectory from a ptychographic scan. Our proposed framework bridges the wide existing gap between approximate models and complex data.

X-ray absorption spectroscopy (XAS) is a commonly-employed technique for characterizing functional materials. In particular, x-ray absorption near edge spectra (XANES) encodes local coordination and electronic information and machine learning approaches to extract this information is of significant interest. To date, most ML approaches for XANES have primarily focused on using the raw spectral intensities as input, overlooking the potential benefits of incorporating spectral transformations and dimensionality reduction techniques into ML predictions. In this work, we focused on systematically comparing the impact of different featurization methods on the performance of ML models for XAS analysis. We evaluated the classification and regression capabilities of these models on computed datasets and validated their performance on previously unseen experimental datasets. Our analysis revealed an intriguing discovery: the cumulative distribution function (CDF) feature achieves both high prediction accuracy and exceptional transferability. This remarkably robust performance can be attributed to its tolerance to horizontal shifts in spectra, which is crucial when validating models using experimental data. While this work exclusively focuses on XANES analysis, we anticipate that the methodology presented here will hold promise as a versatile asset to the broader spectroscopy community.

Speckle patterns produced by coherent X-ray have a close relationship with the internal structure of materials but quantitative inversion of the relationship to determine structure from speckle patterns is challenging. Here, we investigate the link between coherent X-ray speckle patterns and sample structures using a model 2D disk system and explore the ability of machine learning to learn aspects of the relationship. Specifically, we train a deep neural network to classify the coherent X-ray speckle patterns according to the disk number density in the corresponding structure. It is demonstrated that the classification system is accurate for both non-disperse and disperse size distributions.

Automated analysis of electron microscopy datasets poses multiple challenges, such as limitation in the size of the training dataset, variation in data distribution induced by variation in sample quality and experiment conditions, etc. It is crucial for the trained model to continue to provide acceptable segmentation/classification performance on new data, and quantify the uncertainty associated with its predictions. Among the broad applications of machine learning, various approaches have been adopted to quantify uncertainty, such as Bayesian modeling, Monte Carlo dropout, ensembles, etc. With the aim of addressing the challenges specific to the data domain of electron microscopy, two different types of ensembles of pre-trained neural networks were implemented in this work. The ensembles performed semantic segmentation of ice crystal within a two-phase mixture, thereby tracking its phase transformation to water. The first ensemble (EA) is composed of U-net style networks having different underlying architectures, whereas the second series of ensembles (ER-i) are composed of randomly initialized U-net style networks, wherein each base learner has the same underlying architecture 'i'. The encoders of the base learners were pre-trained on the Imagenet dataset. The performance of EA and ER were evaluated on three different metrics: accuracy, calibration, and uncertainty. It is seen that EA exhibits a greater classification accuracy and is better calibrated, as compared to ER. While the uncertainty quantification of these two types of ensembles are comparable, the uncertainty scores exhibited by ER were found to be dependent on the specific architecture of its base member ('i') and not consistently better than EA. Thus, the challenges posed for the analysis of electron microscopy datasets appear to be better addressed by an ensemble design like EA, as compared to an ensemble design like ER., Comment: 10 pages, 6 figures. Published as part of the Workshop on Machine Learning for Materials Science in conjunction with 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

Quantifying charge-state transition energy levels of impurities in semiconductors is critical to understanding and engineering their optoelectronic properties for applications ranging from solar photovoltaics to infrared lasers. While these transition levels can be measured and calculated accurately, such efforts are time-consuming and more rapid prediction methods would be beneficial. Here, we significantly reduce the time typically required to predict impurity transition levels using multi-fidelity datasets and a machine learning approach employing features based on elemental properties and impurity positions. We use transition levels obtained from low-fidelity (i.e., local-density approximation or generalized gradient approximation) density functional theory (DFT) calculations, corrected using a recently proposed modified band alignment scheme, which well-approximates transition levels from high-fidelity DFT (i.e., hybrid HSE06). The model fit to the large multi-fidelity database shows improved accuracy compared to the models trained on the more limited high-fidelity values. Crucially, in our approach, when using the multi-fidelity data, high-fidelity values are not required for model training, significantly reducing the computational cost required for training the model. Our machine learning model of transition levels has a root mean squared (mean absolute) error of 0.36 (0.27) eV vs high-fidelity hybrid functional values when averaged over 14 semiconductor systems from the II-VI and III-V families. As a guide for use on other systems, we assessed the model on simulated data to show the expected accuracy level as a function of bandgap for new materials of interest. Finally, we use the model to predict a complete space of impurity charge-state transition levels in all zinc blende III-V and II-VI systems.

The great tunability of the properties of halide perovskites presents new opportunities for optoelectronic applications as well as significant challenges associated with exploring combinatorial chemical spaces. In this work, we develop a framework powered by high-throughput computations and machine learning for the design and prediction of mixed cation halide perovskite alloys. In a chemical space of ABX$_{3}$ perovskites with a selected set of options for A, B, and X atoms, pseudo-cubic structures of compounds with B-site mixing are simulated using density functional theory (DFT) and several properties are computed, including stability, lattice constant, band gap, vacancy formation energy, refractive index, and optical absorption spectrum, using both semi-local and hybrid functionals. Neural networks (NN) are used to train predictive models for every property using tabulated elemental properties of A, B, and X site atoms as descriptors. Starting from the DFT dataset of 229 points, we use the trained NN models to predict the structural, energetic, electronic and optical properties of a complete dataset of 17,955 compounds, and perform high-throughput screening in terms of stability, band gap and defect tolerance, to obtain 574 promising compounds that are ranked as potential absorbers according to their photovoltaic figure of merit. Compositional trends in the screened set of attractive mixed cation halide perovskites are revealed and additional computations are performed on selected compounds. The data-driven design framework developed here is promising for designing novel mixed compositions and can be extended to a wider perovskite chemical space in terms of A, B, and X atoms, different kinds of mixing at the A, B, or X sites, non-cubic phases, and other properties of interest.

Different types of spectroscopies, such as X-ray absorption near edge structure (XANES) and Raman spectroscopy, play a very important role in analyzing the characteristics of different materials. In scientific literature, XANES/Raman data are usually plotted in line graphs which is a visually appropriate way to represent the information when the end-user is a human reader. However, such graphs are not conducive to direct programmatic analysis due to the lack of automatic tools. In this paper, we develop a plot digitizer, named Plot2Spectra, to extract data points from spectroscopy graph images in an automatic fashion, which makes it possible for large scale data acquisition and analysis. Specifically, the plot digitizer is a two-stage framework. In the first axis alignment stage, we adopt an anchor-free detector to detect the plot region and then refine the detected bounding boxes with an edge-based constraint to locate the position of two axes. We also apply scene text detector to extract and interpret all tick information below the x-axis. In the second plot data extraction stage, we first employ semantic segmentation to separate pixels belonging to plot lines from the background, and from there, incorporate optical flow constraints to the plot line pixels to assign them to the appropriate line (data instance) they encode. Extensive experiments are conducted to validate the effectiveness of the proposed plot digitizer, which shows that such a tool could help accelerate the discovery and machine learning of materials properties.

To fully leverage the power of image simulation to corroborate and explain patterns and structures in atomic resolution microscopy (e.g., electron and scanning probe), an initial correspondence between the simulation and experimental image must be established at the outset of further high accuracy simulations or calculations. Furthermore, if simulation is to be used in context of highly automated processes or high-throughput optimization, the process of finding this correspondence itself must be automated. In this work, we introduce ingrained, an open-source automation framework which solves for this correspondence and fuses atomic resolution image simulations into the experimental images to which they correspond. We describe herein the overall ingrained workflow, focusing on its application to interface structure approximations, and the development of an experimentally rationalized forward model for scanning tunneling microscopy simulation.

Due to recent improvements in image resolution and acquisition speed, materials microscopy is experiencing an explosion of published imaging data. The standard publication format, while sufficient for traditional data ingestion scenarios where a select number of images can be critically examined and curated manually, is not conducive to large-scale data aggregation or analysis, hindering data sharing and reuse. Most images in publications are presented as components of a larger figure with their explicit context buried in the main body or caption text, so even if aggregated, collections of images with weak or no digitized contextual labels have limited value. To solve the problem of curating labeled microscopy data from literature, this work introduces the EXSCLAIM! Python toolkit for the automatic EXtraction, Separation, and Caption-based natural Language Annotation of IMages from scientific literature. We highlight the methodology behind the construction of EXSCLAIM! and demonstrate its ability to extract and label open-source scientific images at high volume.

Scientific literature contains large volumes of complex, unstructured figures that are compound in nature (i.e. composed of multiple images, graphs, and drawings). Separation of these compound figures is critical for information retrieval from these figures. In this paper, we propose a new strategy for compound figure separation, which decomposes the compound figures into constituent subfigures while preserving the association between the subfigures and their respective caption components. We propose a two-stage framework to address the proposed compound figure separation problem. In particular, the subfigure label detection module detects all subfigure labels in the first stage. Then, in the subfigure detection module, the detected subfigure labels help to detect the subfigures by optimizing the feature selection process and providing the global layout information as extra features. Extensive experiments are conducted to validate the effectiveness and superiority of the proposed framework, which improves the detection precision by 9%.

Due to the absence of matching carbon atoms at honeycomb centers with carbon atoms in adjacent graphene sheets, theorists predicted that a sliding process is needed to form AA, AB, or ABC stacking when directly converting graphite into sp3 bonded diamond. Here, using twisted bilayer graphene, which naturally provides AA and AB stacking configurations, we report the ultrafast formation of a transient 2D diamond-like structure (which is not observed in aligned graphene) under femtosecond laser irradiation. This photo-induced phase transition is evidenced by the appearance of new bond lengths of 1.94A and 3.14A in the time-dependent differential pair distribution function using MeV ultrafast electron diffraction. Molecular dynamics and first principles calculation indicate that sp3 bonds nucleate at AA and AB stacked areas in moire pattern. This work sheds light on the direct graphite-to-diamond transformation mechanism, which has not been fully understood for more than 60 years.

Owing to the increasing popularity of lead-based hybrid perovskites for photovoltaic (PV) applications, it is crucial to understand their defect physics and its influence on their optoelectronic properties. In this work, we simulate various point defects in pseudo-cubic structures of mixed iodide-bromide and bromide-chloride methylammonium lead perovskites with the general formula MAPbI_{3-y}Br_{y} or MAPbBr_{3-y}Cl_{y} (where y is between 0 and 3), and use first principles based density functional theory computations to study their relative formation energies and charge transition levels. We identify vacancy defects and Pb on MA anti-site defect as the lowest energy native defects in each perovskite. We observe that while the low energy defects in all MAPbI_{3-y}Br_{y} systems only create shallow transition levels, the Br or Cl vacancy defects in the Cl-containing pervoskites have low energy and form deep levels which become deeper for higher Cl content. Further, we study extrinsic substitution by different elements at the Pb site in MAPbBr_{3}, MAPbCl_{3} and the 50-50 mixed halide perovskite, MAPbBr_{1.5}Cl_{1.5}, and identify some transition metals that create lower energy defects than the dominant intrinsic defects and also create mid-gap charge transition levels., Comment: 10 pages, 5 figures, 1 supplementary information file

The ability to predict the likelihood of impurity incorporation and their electronic energy levels in semiconductors is crucial for controlling its conductivity, and thus the semiconductor's performance in solar cells, photodiodes, and optoelectronics. The difficulty and expense of experimental and computational determination of impurity levels makes a data-driven machine learning approach appropriate. In this work, we show that a density functional theory-generated dataset of impurities in Cd-based chalcogenides CdTe, CdSe, and CdS can lead to accurate and generalizable predictive models of defect properties. By converting any semiconductor + impurity system into a set of numerical descriptors, regression models are developed for the impurity formation enthalpy and charge transition levels. These regression models can subsequently predict impurity properties in mixed anion CdX compounds (where X is a combination of Te, Se and S) fairly accurately, proving that although trained only on the end points, they are applicable to intermediate compositions. We make machine-learned predictions of the Fermi-level dependent formation energies of hundreds of possible impurities in 5 chalcogenide compounds, and suggest a list of impurities which can shift the equilibrium Fermi level in the semiconductor as determined by the dominant intrinsic defects. These dominating impurities as predicted by machine learning compare well with DFT predictions, revealing the power of machine-learned models in the quick screening of impurities likely to affect the optoelectronic behavior of semiconductors., Comment: 14 pages, 7 figures

A deep neural network (DNN) model consisting of two hidden layers was proposed for predicting the immediate environments of specific atoms based on X-ray absorption near-edge spectra (XANES). The output layer of the DNN can be adjusted to form a classifier or regressor, to predict the local and overall coordination environments, respectively. Using Li3FeO3.5 as a model system, it was demonstrated that the prediction accuracy of the DNN classifier is higher than 98%, and the predictions of the DNN regressor also showed notable agreement with the ground truth. Therefore, despite its simplicity, this DNN architecture can be expected to be generally capable of predicting the structural properties of various systems. Fine tuning of the hyperparameters, bias-variance tradeoff, and strategies to enrich the versatility of the model were also discussed., Comment: 11 pages, 4 figures

Employing electrons for direct control of nanoscale reaction is highly desirable since it provides fabrication of nanostructures with different properties at atomic resolution and with flexibility of dimension and location. Here, applying in situ transmission electron microscopy, we show the reversible oxidation and reduction kinetics in Ag, well controlled by changing the dose rate of electron beam. Aberration-corrected high-resolution transmission electron microscopy observation reveals that O atoms are preferably inserted and extracted along the {111} close-packed planes of Ag, leading to the nucleation and decomposition of nanoscale Ag2O islands on the Ag substrate. By controlling electron beam size and dose rate, we demonstrated fabrication of an array of 3 nm Ag2O nanodots in an Ag matrix. Our results open up a new pathway to manipulate atomistic reaction with electron beam towards the precise fabrication of nanostructures for device applications.

The predictive modeling of lattice thermal conductivity is of fundamental importance for the understanding and design of materials for a wide range of applications. Two major approaches, namely molecular dynamics (MD) simulations and calculations solving approximately the Boltzmann transport equation (BTE), have been developed to compute the lattice thermal conductivity. We present a detailed direct comparison of these two approaches, using as prototypical cases MgO and PbTe. The comparison, carried out using empirical potentials, takes into account the effects of fourth order phonon scattering, temperature-dependent phonon frequencies (phonon renormalization), and investigates the effects of quantum vs. classical statistics. We clarify that equipartition, as opposed to Maxwell Boltzmann, govern the statistics of phonons in MD simulations. We find that lattice thermal conductivity values from MD and BTE show an apparent, satisfactory agreement; however such an agreement is the result of error cancellations. We also show that the primary effect of statistics on thermal conductivity is via the scattering rate dependence on phonon populations.

Classical molecular dynamics (MD) simulations enable modeling of materials and examination of microscopic details that are not accessible experimentally. The predictive capability of MD relies on the force field (FF) used to describe interatomic interactions. FF parameters are typically determined to reproduce selected material properties computed from density functional theory (DFT) and/or measured experimentally. A common practice in parameterizing FFs is to use least-squares local minimization algorithms. Genetic algorithms (GAs) have also been demonstrated as a viable global optimization approach, even for complex FFs. However, an understanding of the relative effectiveness and efficiency of different optimization techniques for the determination of FF parameters is still lacking. In this work, we evaluate various FF parameter optimization schemes, using as example a training data set calculated from DFT for different polymorphs of Ir$O_2$. The Morse functional form is chosen for the pairwise interactions and the optimization of the parameters against the training data is carried out using (1) multi-start local optimization algorithms: Simplex, Levenberg-Marquardt, and POUNDERS, (2) single-objective GA, and (3) multi-objective GA. Using random search as a baseline, we compare the algorithms in terms of reaching the lowest error, and number of function evaluations. We also compare the effectiveness of different approaches for FF parameterization using a test data set with known ground truth (i.e generated from a specific Morse FF). We find that the performance of optimization approaches differs when using the Test data vs. the DFT data. Overall, this study provides insight for selecting a suitable optimization method for FF parameterization, which in turn can enable more accurate prediction of material properties and chemical phenomena.

Organic-inorganic hybrid perovskites such as methylammonium lead iodide (CH3NH3PbI3) are game-changing semiconductors for solar cells and light-emitting devices owing to their exceptionally long carrier lifetime and diffusion length. Determining whether the large dipole moment of the organic cation and dynamic disorder benefit the optoelectronic properties of CH3NH3PbI3 has been an outstanding challenge. Herein, via transient absorption measurements employing an infrared pump pulse tuned to a methylammonium vibration, we observe slow, nanosecond-long thermal dissipation from the selectively excited organic mode to the entire lattice. Resulting transient electronic signatures, during the period of thermal-nonequilibrium when the induced thermal motions are mostly concentrated on the organic sublattice, reveal that induced motions of the organic cations do not alter absorption or photoluminescence response of CH3NH3PbI3, beyond thermal effects. Our results suggest that the attractive optoelectronic properties of CH3NH3PbI3 mainly derive from the inorganic lead-halide framework., Comment: 18 pages, 4 figures

Peierls-Boltzmann transport equation, coupled with third-order anharmonic lattice dynamics calculations, has been widely used to model lattice thermal conductivity ($\kappa_{l}$) in bulk crystals. However, its application to materials with structural phase transition at relatively high temperature is fundamentally challenged by the presence of lattice instabilities (imaginary phonon modes). Additionally, its accuracy suffers from the absence of higher-than-third-order phonon scattering processes, which are important near/above the Debye temperature. In this letter, we present an effective scheme that combines temperature-induced anharmonic phonon renormalization and four-phonon scattering to resolve these two theoretical challenges. We apply this scheme to investigate the lattice dynamics and thermal transport properties of GeTe, which undergoes a second-order ferroelectric phase transition from rhombohedral $\alpha$-GeTe to rocksalt $\beta$-GeTe at about 700~K. Our results on the high-temperature phase $\beta$-GeTe at 800~K confirm the stabilization of $\beta$-GeTe by temperature effects. We find that considering only three-phonon scattering leads to significantly overestimated $\kappa_{l}$ of 3.8~W/mK at 800~K, whereas including four-phonon scattering reduces $\kappa_{l}$ to 1.7~W/mK, a value comparable with experiments. To explore the possibility to further suppress $\kappa_{l}$, we show that alloying $\beta$-GeTe with heavy cations such as Pb and Bi can effectively reduce $\kappa_{l}$ to about 1.0~W/mK, whereas sample size needs to be around 10nm through nanostructuring to achieve a comparable reduction of $\kappa_{l}$.

PbTe, one of the most promising thermoelectric materials, has recently demonstrated thermoelectric figure of merit ($ZT$) of above 2.0 when alloyed with group II elements. The improvements are due mainly to significant reduction of lattice thermal conductivity ($\kappa_{l}$), which was in turn attributed to nanoparticle precipitates. However, a fundamental understanding of various phonon scattering mechanisms within the bulk alloy is still lacking. In this work, we apply the newly-developed density-functional-theory (DFT)-based compressive sensing lattice dynamics (CSLD) approach to model lattice heat transport in PbTe, MTe, and Pb$_{0.94}$M$_{0.06}$Te (M=Mg, Ca, Sr and Ba), compare our results with experimental measurements, with focus on strain effect and mass disorder scattering. We find that (1) CaTe, SrTe and BaTe in the rock-salt structure exhibit much higher $\kappa_{l}$ than PbTe, while MgTe in the same structure shows anomalously low $\kappa_{l}$; (2) lattice heat transport of PbTe is extremely sensitive to static strain induced by alloying atoms in solid solution form; (3) mass disorder scattering plays a major role in reducing $\kappa_{l}$ for Mg/Ca/Sr-alloyed PbTe through strongly suppressing the lifetimes of intermediate- and high-frequency phonons, while for Ba-alloyed PbTe, precipitated nanoparticles are also important.

Vanadium dioxide (VO$_{2}$) undergoes a first-order metal-insulator transition (MIT) upon cooling near room temperature, concomitant with structural change from rutile to monoclinic. Accurate characterization of lattice vibrations is vital for elucidating the transition mechanism. To investigate the lattice dynamics and thermal transport properties of VO$_{2}$ across the MIT, we present a phonon renormalization scheme based on self-consistent phonon theory through iteratively refining vibrational free energy. Using this technique, we compute temperature-dependent phonon dispersion and lifetimes, and point out the importance of both magnetic and vibrational entropy in driving the MIT. The predicted phonon dispersion and lifetimes show quantitative agreement with experimental measurements. We demonstrate that lattice thermal conductivity of rutile VO$_{2}$ is nearly temperature independent as a result of strong intrinsic anharmonicity, while that of monoclinic VO$_{2}$ varies according to $1/T$. Due to phonon softening and enhanced scattering rates, the lattice thermal conductivity is deduced to be substantially lower in the rutile phase, suggesting that Wiedemann-Franz law might not be strongly violated in rutile VO$_{2}$.

We compute the transient dynamics of phonons in contact with high energy "hot" charge carriers in 12 polar and non-polar semiconductors, using a first-principles Boltzmann transport framework. For most materials, we find that the decay in electronic temperature departs significantly from a single-exponential model at times ranging from 1 ps to 15 ps after electronic excitation, a phenomenon concomitant with the appearance of non-thermal vibrational modes. We demonstrate that these effects result from the slow thermalization within the phonon subsystem, caused by the large heterogeneity in the timescales of electron-phonon and phonon-phonon interactions in these materials. We propose a generalized 2-temperature model accounting for the phonon thermalization as a limiting step of electron-phonon thermalization, which captures the full thermal relaxation of hot electrons and holes in semiconductors. A direct consequence of our findings is that, for semiconductors, information about the spectral distribution of electron-phonon and phonon-phonon coupling can be extracted from the multi-exponential behavior of the electronic temperature.

Tantalum, tantalum oxide and their hetero-interfaces are of tremendous technological interest in several applications spanning electronics, thermal management, catalysis and biochemistry. For example, local oxygen stoichiometry variation in TaOx memristors comprising of metallic (Ta) and insulating oxide (Ta2O5) have been shown to result in fast switching on the sub-nanosecond timescale over a billion cycles, relevant to neuromorphic computation. Despite its broad importance, an atomistic scale understanding of oxygen stoichiometry variation across Ta/TaOx hetero-interfaces, such as during early stages of oxidation and oxide growth, is not well understood. This is mainly due to the lack of a variable charge interatomic potential model for tantalum oxides that can accurately describe the ionic interactions in the metallic (Ta) and oxide (TaOx) environment as well as at their interfaces. To address this challenge, we introduce a charge transfer ionic potential (CTIP) model for Ta/Ta-oxide system by training against lattice parameters, cohesive energies, equations of state, and elastic properties of various experimentally observed Ta2O5 polymorphs. The best set of CTIP parameters are determined by employing a single-objective global optimization scheme driven by genetic algorithms followed by local Simplex optimization. Our newly developed CTIP potential accurately predicts structure, thermodynamics, energetic ordering of polymorphs, as well as elastic and surface properties of both Ta and Ta2O5, in excellent agreement with DFT calculations and experiments. We employ our newly parameterized CTIP potential to investigate the early stages of oxidation of Ta at different temperatures and atomic/molecular nature of the oxidizing species.

In molecular dynamics (MD) simulation, force field determines the capability of an individual model in capturing physical and chemistry properties. The method for generating proper parameters of the force field form is the key component for computational research in chemistry, biochemistry, and condensed-phase physics. Our study showed that the feasibility to predict experimental condensed phase properties (i.e., density and heat of vaporization) of methanol through problem specific force field from only quantum chemistry information. To acquire the satisfying parameter sets of the force field, the genetic algorithm (GA) is the main optimization method. For electrostatic potential energy, we optimized both the electrostatic parameters of methanol using the GA method, which leads to low deviations of between the quantum mechanics (QM) calculations and the GA optimized parameters. We optimized the van der Waals (vdW) parameters both using GA and guided GA methods by calibrating interaction energy of various methanol homo-clusters, such as nonamers, undecamers, or tridecamers. Excellent agreement between the training dataset from QM calculations (i.e., MP2) and GA optimized parameters can be achieved. However, only the guided GA method, which eliminates the overestimation of interaction energy from MP2 calculations in the optimization process, provides proper vdW parameters for MD simulation to get the condensed phase properties (i.e., density and heat of vaporization) of methanol. Throughout the whole optimization process, the experimental value were not involved in the objective functions, but were only used for the purpose of justifying models (i.e., nonamers, undecamers, or tridecamers) and validating methods (i.e., GA or guided GA). Our method shows the possibility of developing descriptive polarizable force field using only QM calculations., Comment: not submitted to anywhere else by July 2016

Au nanoclusters are of technological relevance for catalysis, photonics, sensors, and of fundamental scientific interest owing to planar to globular structural transformation at an anomalously high number of atoms i.e. in the range 12-14. The nature and causes of this transition remain a mystery. In order to unravel this conundrum, high throughput density functional theory (DFT) calculations, coupled with a global structural optimization scheme based on a modified genetic algorithm (GA) are conducted. More than 20,000 Au$_{12}$, Au$_{13}$, and Au$_{14}$ nanoclusters are evaluated. With any DFT functional, globular and planar structures coexist across the size range of interest. The planar-globular transition is gradual at room temperature rather than a sharp transition as previously believed. The effects of anionicity, $\textit{s-d}$ band hybridization and long range interactions on the dimensional transition are quantified by using the structures adjacent to minima. Anionicity marginally changes the relative stability of the clusters. The degree of $\textit{s-d}$ hybridization is varied via changing the Hubbard U value which corroborate that $\textit{s-d}$ hybridization alone does not stabilize planar structures. van der Waals interactions, on the other hand, stabilize globular structures. These results elucidate the balance between the different reasons of the dimensional transition in gold nanoclusters.

Molecular dynamics simulations using empirical force fields (EFFs) are crucial for gaining fundamental insights into atomic structure and long timescale dynamics of Au nanoclusters with far-reaching applications in energy and devices. This approach is thwarted by the failure of currently available EFFs in describing the size-dependent dimensionality and diverse geometries exhibited by Au clusters (e.g., planar, hollow cages, pyramids). Owing to their ability to account for bond directionality, bond-order based EFFs, such as the Tersoff-type Bond Order Potential (BOP), are well suited for such a description. Nevertheless, the predictive power of existing BOP parameters is severely limited in the nm length scale owing to the predominance of bulk Au properties used to train them. Here, we mitigate this issue by introducing a new hybrid bond order potential (HyBOP), which account for (a) short-range interactions via Tersoff-type BOP terms and (b) long-range effects by a scaled LJ term whose contribution depends on the local atomic density. We optimized the independent parameters for our HyBOP using a global optimization scheme driven by genetic algorithms. Moreover, to ensure good transferability of these parameters across different length scales, we used an extensive training dataset encompasses structural and energetic properties of a thousand 13-atom Au clusters, surface energies, as well as bulk polymorphs, obtained from density functional theory (DFT) calculations. Our newly developed HyBOP has been found to accurately describe (a) global minimum energy configurations at different clusters sizes, (b) critical size of transition from planar to globular clusters, (c) evolution of structural motifs with cluster size, and (d) thermodynamics, structure, elastic properties of bulk polymorphs as well as surfaces, in excellent agreement with DFT calculations and spectroscopic experiments.