Linear scaling calculations allow for accurate AB initio-precision structure of molecular extended dynamics simulations of distorted systems

Understanding the atomic structure of disordered materials remains a frightening challenge, but accurate models are important for predicting their behavior. Tigany Zarrouk and Miguel A. Caro, from Aalto University, are now making significant advances in molecular extension dynamics, a computational method that incorporates experimental data directly into simulations. Their research establishes methods for calculating key experimental observability, such as X-ray diffraction patterns and core electron binding energies, and uses computational costs that expand linearly in system size. This breakthrough allows for efficient investigation of vast structural possibilities, allowing researchers to identify stable structures in line with experimental findings, outperform the accuracy of traditional simulation methods, and open up new means of material design and characterization.

Amorphous carbon structures and the possibilities of machine learning

A comprehensive review uncovers vibrant and rapidly evolving fields focusing on amorphous carbon and related materials, increasingly leveraging the power of machine learning. Investigating the structure of amorphous carbon using techniques such as neutron scattering and X-ray diffraction is intended to understand the relationship between atomic arrangement and material properties. Research extends to hydrogenated amorphous carbon, nanoporous carbon, graphene oxide, and diamond-like carbon, each offering unique properties and applications. The main trend is the increased use of machine learning to accelerate material discovery and understanding.

Scientists are developing the potential between machine learning to provide accurate and efficient simulations and overcome the limitations of traditional methods. These potentials employ a variety of approaches, including local parameterization of dispersive interactions, graph neural networks to represent the atomic environment, and the creation of universal potentials that can be applied to a wide range of materials and conditions. This work paves the way for basic materials models, similar to large-scale language models, trained on a vast dataset to predict properties and design new materials. Molecular dynamics simulations are widely used to study the structure and dynamics of materials.

Combining these calculation methods with advanced experimental techniques enables a synergistic approach to materials science. This field is moving towards a data-driven paradigm, focusing on amorphous and disordered materials, prioritizing the development of the possibilities of accurate and efficient machine learning. This convergence of experiments, theory, and machine learning promises to accelerate material discovery and innovation.

Molecular dynamics tailored to experimental data

Scientists have developed a sophisticated molecular dynamics method, molecular enhancement dynamics, to generate accurate, low-energy structural models that are consistent with experimental data for faulty systems. Recognizing the limitations of traditional sampling approaches, the team designed a method that simultaneously optimizes both interatomic potential energy and defined experimental possibilities. This innovative approach overcomes the limitations of standard methods by directly searching for structures compatible with experimental observations. This method involves the formulation of MAD equations with linear scaling calculations to match the X-ray or neutron diffraction patterns with local observability.

MAD simulations effectively identify both non-equilibrium experimental synthesis conditions and metastable structures that match low energy structures than those seen using alternative computational protocols. Generalizing the viral tensor by incorporating experimental power allows for accurate control of density and discovery of structures with density that match experimental observability. The team implemented linear scaling formulations within the turbo gap code for both CPU and GPU implementations, achieving a significant 100x speedup on the GPU, allowing for larger systems simulations and accelerated structural optimization.

Accurate obstruction structure through molecular expansion dynamics

Researchers achieved a breakthrough in modeling of obstacle materials by developing molecular extended dynamics methods that can produce accurate, low-energy structures that are consistent with experimental data. This study addresses the key challenges in simulating amorphous solids where traditional methods struggle to replicate realistic structures. The MAD method minimizes both the potential energy between atoms and the differences between simulation and experimental data, simultaneously with molecular dynamics simulation and multi-objective optimization. The team has successfully implemented a linear scaling formulation to calculate and match experimental data including X-ray and neutron diffraction and core electron binding energies.

Simulations show that MAD can identify both non-equilibrium experimental synthesis and metastable structures consistent with low energy structures than those obtained through traditional melt quenching approaches. Furthermore, generalizing virial tensors with experimental power allows precise control of density and the creation of structures that match the experimental density. Using this method to amorphous carbon, researchers have produced structures of glassy carbon, tetrahedral amorphous carbon, deuterium-rich amorphous carbon, and oxygen-rich amorphous carbon. These simulations utilized experimental data, including X-ray diffraction, neutron diffraction, and X-ray photoelectron spectroscopy. The simulation accurately replicates experimental density, demonstrates the insane forces of creating realistic models of amorphous materials, paving a new pathway for material design and discovery.

Dynamics and molecular structure from experiments

This work presents a new method of calculation: molecular extension dynamics. This efficiently searches for atomic structures tailored to experimental data. The team combined molecular dynamics simulations with experimental potentials to enable accurate, low-energy structure generation of faulty systems. This approach overcomes the limitations of traditional methods such as inverse Monte Carlo and represents a major advance in material modeling. Researchers have demonstrated that the method not only finds structures compatible with experimental observations, but also identifies lower energy configurations than those obtained with alternative technologies. While acknowledging that perfect agreement with experimental data is not guaranteed, the team highlights the possibility of this approach to filling the science gap between computational and experimental materials, paving the way for future work focusing on applying this method to a wider range of disordered materials.