QuantumATK: an integrated platform of electronic and atomic-scale modelling tools.

QuantumATK is an integrated set of atomic-scale modelling tools developed since 2003 by professional software engineers in collaboration with academic researchers. While different aspects and individual modules of the platform have been previously presented, the purpose of this paper is to give a general overview of the platform. The QuantumATK simulation engines enable electronic-structure calculations using density functional theory or tight-binding model Hamiltonians, and also offers bonded or reactive empirical force fields in many different parametrizations. Density functional theory is implemented using either a plane-wave basis or expansion of electronic states in a linear combination of atomic orbitals. The platform includes a long list of advanced modules, including Green's-function methods for electron transport simulations and surface calculations, first-principles electron-phonon and electron-photon couplings, simulation of atomic-scale heat transport, ion dynamics, spintronics, optical properties of materials, static polarization, and more. Seamless integration of the different simulation engines into a common platform allows for easy combination of different simulation methods into complex workflows. Besides giving a general overview and presenting a number of implementation details not previously published, we also present four different application examples. These are calculations of the phonon-limited mobility of Cu, Ag and Au, electron transport in a gated 2D device, multi-model simulation of lithium ion drift through a battery cathode in an external electric field, and electronic-structure calculations of the composition-dependent band gap of SiGe alloys.

Distributed replica dynamics.

A distributed replica dynamics (DRD) method is proposed to calculate rare-event molecular dynamics using distributed computational resources. Similar to Voter's parallel replica dynamics (PRD) method, the dynamics of independent replicas of the system are calculated on different computational clients. In DRD, each replica runs molecular dynamics from an initial state for a fixed simulation time and then reports information about the trajectory back to the server. A simulation clock on the server accumulates the simulation time of each replica until one reports a transition to a new state. Subsequent calculations are initiated from within this new state and the process is repeated to follow the state-to-state evolution of the system. DRD is designed to work with asynchronous and distributed computing resources in which the clients may not be able to communicate with each other. Additionally, clients can be added or removed from the simulation at any point in the calculation. Even with heterogeneous computing clients, we prove that the DRD method reproduces the correct probability distribution of escape times. We also show this correspondence numerically; molecular dynamics simulations of Al(100) adatom diffusion using PRD and DRD give consistent exponential distributions of escape times. Finally, we discuss guidelines for choosing the optimal number of replicas and replica trajectory length for the DRD method.

A theoretical and experimental approach for correlating nanoparticle structure and electrocatalytic activity.

The objective of the research described in this Account is the development of high-throughput computational-based screening methods for discovery of catalyst candidates and subsequent experimental validation using appropriate catalytic nanoparticles. Dendrimer-encapsulated nanoparticles (DENs), which are well-defined 1-2 nm diameter metal nanoparticles, fulfill the role of model electrocatalysts. Effective comparison of theory and experiment requires that the theoretical and experimental models map onto one another perfectly. We use novel synthetic methods, advanced characterization techniques, and density functional theory (DFT) calculations to approach this ideal. For example, well-defined core@shell DENs can be synthesized by electrochemical underpotential deposition (UPD), and the observed deposition potentials can be compared to those calculated by DFT. Theory is also used to learn more about structure than can be determined by analytical characterization alone. For example, density functional theory molecular dynamics (DFT-MD) was used to show that the core@shell configuration of Au@Pt DENs undergoes a surface reconstruction that dramatically affects its electrocatalytic properties. A separate Pd@Pt DENs study also revealed reorganization, in this case a core-shell inversion to a Pt@Pd structure. Understanding these types of structural changes is critical to building correlations between structure and catalytic function. Indeed, the second principal focus of the work described here is correlating structure and catalytic function through the combined use of theory and experiment. For example, the Au@Pt DENs system described earlier is used for the oxygen reduction reaction (ORR) as well as for the electro-oxidation of formic acid. The surface reorganization predicted by theory enhances our understanding of the catalytic measurements. In the case of formic acid oxidation, the deformed nanoparticle structure leads to reduced CO binding energy and therefore improved oxidation activity. The final catalytic study we present is an instance of theory correctly predicting (in advance of the experiments) the structure of an effective DEN electrocatalyst. Specifically, DFT was used to determine the optimal composition of the alloy-core in AuPd@Pt DENs for the ORR. This prediction was subsequently confirmed experimentally. This study highlights the major theme of our research: the progression of using theory to rationalize experimental results to the more advanced goal of using theory to predict catalyst function a priori. We still have a long way to go before theory will be the principal means of catalyst discovery, but this Account begins to shed some light on the path that may lead in that direction.

Probing the Limits of Conventional Extended X-ray Absorption Fine Structure Analysis Using Thiolated Gold Nanoparticles.

We present a method for quantifying the accuracy of extended X-ray absorption fine structure (EXAFS) fitting models. As a test system, we consider the structure of bare Au147 nanoparticles as well as particles bound with thiol ligands, which are used to systematically vary disorder in the atomic structure of the nanoparticles. The accuracy of the fitting model is determined by comparing two distributions of bond lengths: (1) a direct average over a molecular dynamics (MD) trajectory using forces and energies from density functional theory (DFT) and (2) a fit to the theoretical EXAFS spectra generated from that same trajectory. Both harmonic and quasi-harmonic EXAFS fitting models are used to characterize the first-shell Au-Au bond length distribution. The harmonic model is found to significantly underestimate the coordination number, disorder, and bond length. The quasi-harmonic model, which includes the third cumulant of the first-shell bond length distribution, yields accurate bond lengths, but incorrectly predicts a decrease in particle size and little change in the disorder with increasing thiol ligands. A direct analysis of the MD data shows that the particle surfaces become much more disordered with ligand binding, and the high disorder is incorrectly interpreted by the EXAFS fitting models. Our DFT calculations compare well with experimental EXAFS measurements of Au nanoparticles, synthesized using a dendrimer encapsulation technique, showing that systematic errors in EXAFS fitting models apply to nanoparticles 1-2 nm in size. Finally we show that a combination of experimental EXAFS analysis with candidate models from DFT is a promising strategy for a more accurate determination of nanoparticle structures.

Molecular dynamics saddle search adaptive kinetic Monte Carlo.

A method for accelerating molecular dynamics simulations in rare event systems is described. From each new state visited, high temperature molecular dynamics trajectories are used to discover the set of escape mechanisms and rates. This event table is provided to the adaptive kinetic Monte Carlo algorithm to model the evolution of the system from state to state. Importantly, an estimator for the completeness of the calculated rate table in each state is derived. The method is applied to three model systems: adatom diffusion on Al(100); island diffusion on Pt(111); and vacancy cluster ripening in bulk Fe. Connections to the closely related temperature accelerated dynamics method of Voter and co-workers is discussed.

Benchmarks for Characterization of Minima, Transition States, and Pathways in Atomic, Molecular, and Condensed Matter Systems.

A set of benchmark systems is defined to compare different computational approaches for characterizing local minima, transition states, and pathways in atomic, molecular, and condensed matter systems. Comparisons between several commonly used methods are presented. The strengths and weaknesses are discussed, as well as implementation details that are important for achieving good performance. All of the benchmarks and methods are provided in an online database to make the implementation details available and the results reproducible. While this paper provides a snapshot of the benchmark results, the online framework is structured to be dynamic and incorporate new methods and codes as they are developed.

Database of atomistic reaction mechanisms with application to kinetic Monte Carlo.

Kinetic Monte Carlo is a method used to model the state-to-state kinetics of atomic systems when all reaction mechanisms and rates are known a priori. Adaptive versions of this algorithm use saddle searches from each visited state so that unexpected and complex reaction mechanisms can also be included. Here, we describe how calculated reaction mechanisms can be stored concisely in a kinetic database and subsequently reused to reduce the computational cost of such simulations. As all accessible reaction mechanisms available in a system are contained in the database, the cost of the adaptive algorithm is reduced towards that of standard kinetic Monte Carlo.

Assessment of the Performance of the M05-2X and M06-2X Exchange-Correlation Functionals for Noncovalent Interactions in Biomolecules.

The highly parametrized, empirical exchange-correlation functionals, M05-2X and M06-2X, developed by Zhao and Truhlar have been shown to describe noncovalent interactions better than density functionals which are currently in common use. However, these methods have yet to be fully benchmarked for the types of interactions important in biomolecules. M05-2X and M06-2X are claimed to capture "medium-range" electron correlation; however, the "long-range" electron correlation neglected by these functionals can also be important in the binding of noncovalent complexes. Here we test M05-2X and M06-2X for the nucleic acid base pairs in the JSCH-2005 database. Using the CCSD(T) binding energies as a benchmark, the performance of these functionals is compared to that of a nonempirical density functional, PBE, and also to that of PBE plus Grimme's empirical dispersion correction, PBE-D. Due to the importance of "long-range" electron correlation in hydrogen-bonded and interstrand base pairs, PBE-D provides more accurate interaction energies on average for the JSCH-2005 database when compared to M05-2X or M06-2X. M06-2X does, however, perform somewhat better than PBE-D for interactions between stacked base pairs.