Search NASA⌕ Search

Engineering topics

Loeffler, Troy D.

Publications and source records attributed to Loeffler, Troy D..

Active and Transfer Learning of High-Dimensional Neural Network Potentials for Transition Metals

Classical molecular dynamics (MD) simulations represent a very popular and powerful tool for materials modeling and design. The predictive power of MD hinges on the ability of the interatomic potential to capture the underlying physics and chemistry. There have been decades of seminal work on developing interatomic potentials, albeit with a focus predominantly on capturing the properties of bulk materials. Such physics-based models, while extensively deployed for predicting the dynamics and properties of nanoscale systems over the past two decades, tend to perform poorly in predicting nanoscale potential energy surfaces (PESs) when compared to high-fidelity first-principles calculations. These limitations stem from the lack of flexibility in such models, which rely on a predefined functional form. Machine learning (ML) models and approaches have emerged as a viable alternative to capture the diverse size-dependent cluster geometries, nanoscale dynamics, and the complex nanoscale PESs, without sacrificing the bulk properties. Here, in this study, we introduce an ML workflow that combines transfer and active learning strategies to develop high-dimensional neural networks (NNs) for capturing the cluster and bulk properties for several different transition metals with applications in catalysis, microelectronics, and energy storage, to name a few. Our NN first learns the bulk PES from the high-quality physics-based models in literature and subsequently augments this learning via retraining with a higher-fidelity first-principles training data set to concurrently capture both the nanoscale and bulk PES. Our workflow departs from status-quo in its ability to learn from a sparsely sampled data set that nonetheless covers a diverse range of cluster configurations from near-equilibrium to highly nonequilibrium as well as learning strategies that iteratively improve the fingerprinting depending on model fidelity. All the developed models are rigorously tested against an extensive first-principles data set of energies and forces of cluster configurations as well as several properties of bulk configurations for 10 different transition metals. Our approach is material agnostic and provides a methodology to transfer and build upon the learnings from decades of seminal work in molecular simulations on to a new generation of ML-trained potentials to accelerate materials discovery and design.

36 MATERIALS SCIENCE↗

Multi-reward reinforcement learning based development of inter-atomic potential models for silica

Abstract Silica is an abundant and technologically attractive material. Due to the structural complexities of silica polymorphs coupled with subtle differences in Si–O bonding characteristics, the development of accurate models to predict the structure, energetics and properties of silica polymorphs remain challenging. Current models for silica range from computationally efficient Buckingham formalisms (BKS, CHIK, Soules) to reactive (ReaxFF) and more recent machine-learned potentials that are flexible but computationally costly. Here, we introduce an improved formalism and parameterization of BKS model via a multireward reinforcement learning (RL) using an experimental training dataset. Our model concurrently captures the structure, energetics, density, equation of state, and elastic constants of quartz (equilibrium) as well as 20 other metastable silica polymorphs. We also assess its ability in capturing amorphous properties and highlight the limitations of the BKS-type functional forms in simultaneously capturing crystal and amorphous properties. We demonstrate ways to improve model flexibility and introduce a flexible formalism, machine-learned ML-BKS, that outperforms existing empirical models and is on-par with the recently developed 50 to 100 times more expensive Gaussian approximation potential (GAP) in capturing the experimental structure and properties of silica polymorphs and amorphous silica.

36 MATERIALS SCIENCE↗

Machine learning overcomes human bias in the discovery of self-assembling peptides

Peptide materials have a wide array of functions, from tissue engineering and surface coatings to catalysis and sensing. Tuning the sequence of amino acids that comprise the peptide modulates peptide functionality, but a small increase in sequence length leads to a dramatic increase in the number of peptide candidates. Traditionally, peptide design is guided by human expertise and intuition and typically yields fewer than ten peptides per study, but these approaches are not easily scalable and are susceptible to human bias. Here, in this work, we introduce a machine learning workflow—AI-expert—that combines Monte Carlo tree search and random forest with molecular dynamics simulations to develop a fully autonomous computational search engine to discover peptide sequences with high potential for self-assembly. We demonstrate the efficacy of the AI-expert to efficiently search large spaces of tripeptides and pentapeptides. The predictability of AI-expert performs on par or better than our human experts and suggests several non-intuitive sequences with high self-assembly propensity, outlining its potential to overcome human bias and accelerate peptide discovery.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Multi-reward Reinforcement Learning Based Bond-Order Potential to Study Strain-Assisted Phase Transitions in Phosphorene

Here, we introduce a multi-reward reinforcement learning (RL) approach to train a flexible bond-order potential (BOP) for 2D phosphorene based on ab initio training data sets. Our approach is based on a continuous action space Monte Carlo tree search algorithm that is general and scalable and presents an efficient multiobjective optimization scheme for high-dimensional materials design problems. As a proof-of-concept, we deploy this scheme to parametrize multiple structural and dynamical properties of 2D phosphorene polymorphs. Our RL-trained BOP model adequately captures the structure, energetics, transformation barriers, equation of state, elastic constants, and phonon dispersions of various 2D P polymorphs. We use this model to probe the impact of temperature and strain rate on the phase transition from black (α-P) to blue phosphorene (β-P) through molecular dynamics simulations. A decrease in critical strain for this phase transition with increase in temperature is observed, and the underlying atomistic mechanisms are discussed.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Learning in continuous action space for developing high dimensional potential energy models

Reinforcement learning (RL) approaches that combine a tree search with deep learning have found remarkable success in searching exorbitantly large, albeit discrete action spaces, as in chess, Shogi and Go. Many real-world materials discovery and design applications, however, involve multi-dimensional search problems and learning domains that have continuous action spaces. Exploring high-dimensional potential energy models of materials is an example. Traditionally, these searches are time consuming (often several years for a single bulk system) and driven by human intuition and/or expertise and more recently by global/local optimization searches that have issues with convergence and/or do not scale well with the search dimensionality. Here, in a departure from discrete action and other gradient-based approaches, we introduce a RL strategy based on decision trees that incorporates modified rewards for improved exploration, efficient sampling during playouts and a “window scaling scheme" for enhanced exploitation, to enable efficient and scalable search for continuous action space problems. Using high-dimensional artificial landscapes and control RL problems, we successfully benchmark our approach against popular global optimization schemes and state of the art policy gradient methods, respectively. We demonstrate its efficacy to parameterize potential models (physics based and high-dimensional neural networks) for 54 different elemental systems across the periodic table as well as alloys. We analyze error trends across different elements in the latent space and trace their origin to elemental structural diversity and the smoothness of the element energy surface. Broadly, our RL strategy will be applicable to many other physical science problems involving search over continuous action spaces.

36 MATERIALS SCIENCE↗

Reinforcement learning in discrete action space applied to inverse defect design

Abstract Reinforcement learning (RL) algorithms that include Monte Carlo Tree Search (MCTS) have found tremendous success in computer games such as Go, Shiga and Chess. Such learning algorithms have demonstrated super-human capabilities in navigating through an exhaustive discrete action search space. Motivated by their success in computer games, we demonstrate that RL can be applied to inverse materials design problems. We deploy RL for a representative case of the optimal atomic scale inverse design of extended defects via rearrangement of chalcogen (e.g. S) vacancies in 2D transition metal dichalcogenides (e.g. MoS 2 ). These defect rearrangements and their dynamics are important from the perspective of tunable phase transition in 2D materials i.e. 2H (semi-conducting) to 1T (metallic) in MoS 2 . We demonstrate the ability of MCTS interfaced with a reactive molecular dynamics simulator to efficiently sample the defect phase space and perform inverse design—starting from randomly distributed S vacancies, the optimal defect rearrangement of defects corresponds a line defect of S vacancies. We compare MCTS performance with evolutionary optimization i.e. genetic algorithms and show that MCTS converges to a better optimal solution (lower objective) and in fewer evaluations compared to GA. We also comprehensively evaluate and discuss the effect of MCTS hyperparameters on the convergence to solution. Overall, our study demonstrates the effectives of using RL approaches that operate in discrete action space for inverse defect design problems.

42 ENGINEERING↗

Ultrafast formation of a transient two-dimensional diamondlike structure in twisted bilayer graphene

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 sp 3 bonded diamond. Here, using twisted bilayer graphene, which naturally provides AA and AB' stacking configurations, we report the ultrafast formation of a transient two-dimensional diamondlike structure (which is not observed in aligned graphene) under femtosecond laser irradiation. This photoinduced phase transition is evidenced by the appearance of bond lengths of 1.94 and 3.14 Å in the time-dependent differential pair distribution function using MeV ultrafast electron diffraction. Molecular dynamics and first-principles calculation indicate that sp 3 bonds nucleate at AA and AB' stacked areas in a moiré pattern. This work sheds light on the direct graphite-to-diamond transformation mechanism, which has not been fully understood for more than 60 years.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Active Learning A Neural Network Model For Gold Clusters & Bulk From Sparse First Principles Training Data

Small metal clusters are of fundamental scientific interest and of tremendous significance in catalysis. These nanoscale clusters display diverse geometries and structural motifs depending on the cluster size; a knowledge of this size-dependent structural motifs and their dynamical evolution has been of longstanding interest. Given the high computational cost of first-principles calculations, molecular modeling and atomistic simulations such as molecular dynamics (MD) has proven to be an important complementary tool to aid this understanding. Classical MD typically employ predefined functional forms which limits their ability to capture such complex size-dependent structural and dynamical transformation. Neural Network (NN) based potentials represent flexible alternatives and in principle, well-trained NN potentials can provide high level of flexibility, transferability and accuracy on-par with the reference model used for training. A major challenge, however, is that NN models are interpolative and requires large quantities (similar to 10 4 or greater) of training data to ensure that the model adequately samples the energy landscape both near and far-from-equilibrium. A highly desirable goal is minimize the number of training data, especially if the underlying reference model is first-principles based and hence expensive. In this work, we introduce an active learning (AL) scheme that trains a NN model on-the-fly with minimal amount of first-principles based training data. Our AL workflow is initiated with a sparse training dataset (similar to 1 to 5 data points) and is updated on-the-fly via a Nested Ensemble Monte Carlo scheme that iteratively queries the energy landscape in regions of failure and updates the training pool to improve the network performance. Using a representative system of gold clusters, we demonstrate that our AL workflow can train a NN with similar to 500 total reference calculations. Using an extensive DFT test set of similar to 1100 configurations, we show that our AL-NN is able to accurately predict both the DFT energies and the forces for clusters of a myriad of different sizes. Our NN predictions are within 30 meV/atom and 40 meV/angstrom of the reference DFT calculations. Moreover, our AL-NN model also adequately captures the various size-dependent structural and dynamical properties of gold clusters in excellent agreement with DFT calculations and available experiments. We finally show that our AL-NN model also captures bulk properties reasonably well, even though they were not included in the training data.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

A coarse-grained deep neural network model for liquid water

We introduce a coarse-grained deep neural network (CG-DNN) model for liquid water that utilizes 50 rotational and translational invariant coordinates and is trained exclusively against energies of ~30 000 bulk water configurations. Our CG-DNN potential accurately predicts both the energies and the molecular forces of water, within 0.9 meV/molecule and 54 meV/Å of a reference (coarse-grained bond-order potential) model. The CG-DNN water model also provides good prediction of several structural, thermodynamic, and temperature dependent properties of liquid water, with values close to those obtained from the reference model. More importantly, CG-DNN captures the well-known density anomaly of liquid water observed in experiments. Our work lays the groundwork for a scheme where existing empirical water models can be utilized to develop a fully flexible neural network framework that can subsequently be trained against sparse data from high-fidelity albeit expensive beyond-DFT calculations.

42 ENGINEERING↗

Promoting Noncovalent Intermolecular Interactions Using a C 60 Core Particle in Aqueous PC60s-Covered Colloids for Ultraefficient Photoinduced Particle Activity

We report that noncovalent intermolecular interactions in nanomaterials, such as van der Waals effects, allow adjustment of the nanoscopic size of compounds and their conformation in molecular crystal regimes. These strong interactions permit small particle sizes to be maintained as the crystals grow. In particular, these effects can be leveraged in the confined/reinforcing phase of molecules. With this in mind, we used C 60 molecules as a core particle in single-PC60 surfactant-covered colloid in a water-processable system. Compared with our previous results based on a PC 61 BM core-PC60 shell particle, the PC60-C 60 colloid had a considerably smaller spherical structure due to the increased intermolecular interactions between C 60 (fullerene) molecules. Interestingly, the conformation of C 60 aggregates was altered depending on the mixed solvents and their volume fraction in the organic phase, which strongly affected the structural properties of the PC60-C 60 colloids. The particle facilitated strong interactions with a p-type core sphere when it was introduced as the shell part of a p-n heterojunction particle. This direct interaction provided effective electronic communication between p- and n-type particles, resulting in ultraefficient photonic properties, particularly in charge separation in aqueous heterostructured colloids. Lastly, this enabled the development of an extremely efficient photovoltaic device with a 6.74% efficiency, which could provide the basis for creating high-performance water-processable solar cells based on p-n heterostructured NPs.

36 MATERIALS SCIENCE↗

“Teamwork Makes the Dream Work”: Tribal Competition Evolutionary Search as a Surrogate for Free-Energy-Based Structural Predictions

Crystal structure prediction has been a grand challenge in material science owing to the large configurational space that one must explore. Evolutionary (genetic) algorithms coupled with first principles calculations are commonly used in crystal structure prediction to sample the ground and metastable states of materials based on configurational energies. However, crystal structure predictions at finite temperature (T), pressure (P), and composition (X) require a free-energy-based search that is often computationally expensive and tedious. Here in this paper, we introduce a new machine-learning workflow for structure prediction that is based on a concept inspired by the evolution of human tribes in primitive society. Our tribal genetic algorithm (GA) combines configurational sampling with evolutionary optimization to accurately predict entropically stabilized phases at finite (T, P, X), at a computational cost that is an order of magnitude smaller than that required for a free-energy-based search. In a departure from standard GA techniques, the populations of individuals are divided into multiple tribes based on a bond-order fingerprint, and genetic operations are modified to ensure that cluster configurations are sampled adequately to capture entropic contributions. Team competition introduced into the evolutionary process allows winning teams (representing a better set of individuals) to expand their sizes; this translates into a more expanded search of the phase space allowing us to explore solutions near possible global minimum. Each team explores a specific section of the structural phase space and avoids bias on solutions arising from the use of individual populations in a purely energy-based search. We demonstrate the efficacy of our approach by performing the structural prediction of a representative two-dimensional two-body system as well as Lennard-Jones clusters over a range of temperatures up to its melting point. Our approach outperforms the standard GA approaches and enables structural search under “real nonambient conditions” on both bulk systems and finite-sized clusters.

36 MATERIALS SCIENCE↗

Machine learning coarse grained models for water

An accurate and computationally efficient molecular level description of mesoscopic behavior of ice-water systems remains a major challenge. Here, we introduce a set of machine-learned coarse-grained (CG) models (ML-BOP, ML-BOP dih , and ML-mW) that accurately describe the structure and thermodynamic anomalies of both water and ice at mesoscopic scales, all at two orders of magnitude cheaper computational cost than existing atomistic models. In a significant departure from conventional force-field fitting, we use a multilevel evolutionary strategy that trains CG models against not just energetics from first-principles and experiments but also temperature-dependent properties inferred from on-the-fly molecular dynamics (~ 10’s of milliseconds of overall trajectories). Our ML BOP models predict both the correct experimental melting point of ice and the temperature of maximum density of liquid water that remained elusive to-date. Our ML workflow navigates efficiently through the high-dimensional parameter space to even improve upon existing high-quality CG models (e.g. mW model).

42 ENGINEERING↗

Configurational-Bias Monte Carlo Back-Mapping Algorithm for Efficient and Rapid Conversion of Coarse-Grained Water Structures into Atomistic Models

Coarse-grained molecular dynamics (MD) simulations represent a powerful approach to simulate longer time scale and larger length scale phenomena than those accessible to all-atom models. The gain in efficiency, however, comes at the cost of atomistic details. The reverse transformation, also known as back-mapping, of coarse grained beads into their atomistic constituents represents a major challenge. Most existing approaches are limited to specific molecules or specific force-fields and often rely on running a long time atomistic MD of the back-mapped configuration to arrive at an optimal solution. Such approaches are problematic when dealing with systems with high diffusion barriers. Here, we introduce a new extension of the configurational-bias-Monte-Carlo (CBMC) algorithm, which we term the crystalline-configurational-bias-Monte-Carlo (C-CBMC) algortihm, that allows rapid and efficient conversion of a coarse-grained model back into its atomistic representation. Although the method is generic, we use a coarse-grained water model as a representative example and demonstrate the back-mapping or reverse transformation for model systems ranging from the ice-liquid water interface to amorphous and crystalline ice configurations. A series of simulations using the TIP4P/Ice model are performed to compare the new CBMC method to several other standard Monte Carlo and Molecular Dynamics based back-mapping techniques. In all the cases, the C-CBMC algorithm is able to find optimal hydrogen bonded configuration many thousand evaluations/steps sooner than the other methods compared within this paper. For crystalline ice structures such as a hexagonal, cubic, and cubic-hexagonal stacking disorder structures, the C-CBMC was able to find structures that were between 0.05 and 0.1 eV/water molecule lower in energy than the ground state energies predicted by the other methods. Detailed analysis of the atomistic structures show a significantly better global hydrogen positioning when contrasted with the existing simpler back-mapping methods. The error in the RDFs of backmapped relative to reference configuration for the CCBMC, MD, and MC were found to be 6.9, 8.7, and 12.9 respectively for the hexagonal system. For the cubic system the relative error of the RDFs for the CCBMC, MD, and MC were found to be 18.2, 34.6, and 39.0 respectively. Finally, our results demonstrate the efficiency and efficacy of our new back-mapping approach, especially for crystalline systems where simple force-field based relaxations have a tendency to get trapped in local minima.

74 ATOMIC AND MOLECULAR PHYSICS↗

Improving the efficiency of configurational-bias Monte Carlo: A density-guided method for generating bending angle trials for linear and branched molecules

A new method has been developed to generate bending angle trials to improve the acceptance rate and the speed of configurational-bias Monte Carlo. Whereas traditionally the trial geometries are generated from a uniform distribution, in this method we attempt to use the exact probability density function so that each geometry generated is likely to be accepted. In actual practice, due to the complexity of this probability density function, a numerical representation of this distribution function would be required. This numerical table can be generated a priori from the distribution function. This method has been tested on a united-atom model of alkanes including propane, 2-methylpropane, and 2,2-dimethylpropane, that are good representatives of both linear and branched molecules. It has been shown from these test cases that reasonable approximations can be made especially for the highly branched molecules to reduce drastically the dimensionality and correspondingly the amount of the tabulated data that is needed to be stored. Despite these approximations, the dependencies between the various geometrical variables can be still well considered, as evident from a nearly perfect acceptance rate achieved. For all cases, the bending angles were shown to be sampled correctly by this method with an acceptance rate of at least 96% for 2,2-dimethylpropane to more than 99% for propane. Since only one trial is required to be generated for each bending angle (instead of thousands of trials required by the conventional algorithm), this method can dramatically reduce the simulation time. The profiling results of our Monte Carlo simulation code show that trial generation, which used to be the most time consuming process, is no longer the time dominating component of the simulation.

2-2-DIMETHYLPROPANE↗