Simulation In Materials Science Research Articles

Efficient surrogate models for materials science simulations: Machine learning-based prediction of microstructure properties

Determining, understanding, and predicting the so-called structure–property relation is an important task in many scientific disciplines, such as chemistry, biology, meteorology, physics, engineering, and materials science. Structure refers to the spatial distribution of, e.g., substances, material, or matter in general, while property is a resulting characteristic that usually depends in a non-trivial way on spatial details of the structure. Traditionally, forward simulations models have been used for such tasks. Recently, several machine learning algorithms have been applied in these scientific fields to enhance and accelerate simulation models or as surrogate models. In this work, we develop and investigate the applications of six machine learning techniques based on two different datasets from the domain of materials science: data from a two-dimensional Ising model for predicting the formation of magnetic domains and data representing the evolution of dual-phase microstructures from the Cahn–Hilliard model. We analyze the accuracy and robustness of all models and elucidate the reasons for the differences in their performances. The impact of including domain knowledge through tailored features is studied, and general recommendations based on the availability and quality of training data are derived from this.

Rapid and accurate predictions of perfect and defective material properties in atomistic simulation using the power of 3D CNN-based trained artificial neural networks

This article introduces an innovative approach that utilizes machine learning (ML) to address the computational challenges of accurate atomistic simulations in materials science. Focusing on the field of molecular dynamics (MD), which offers insight into material behavior at the atomic level, the study demonstrates the potential of trained artificial neural networks (tANNs) as surrogate models. These tANNs capture complex patterns from built datasets, enabling fast and accurate predictions of material properties. The article highlights the application of 3D convolutional neural networks (CNNs) to incorporate atomistic details and defects in predictions, a significant advancement compared to current 2D image-based, or descriptor-based methods. Through a dataset of atomistic structures and MD simulations, the trained 3D CNN achieves impressive accuracy, predicting material properties with a root-mean-square error below 0.65 GPa for the prediction of elastic constants and a speed-up of approximately 185 to 2100 times compared to traditional MD simulations. This breakthrough promises to expedite materials design processes and facilitate scale-bridging in materials science, offering a new perspective on addressing computational demands in atomistic simulations.

GPGPU Programming for Dipolar Field Calculation

Abstract Accelerating computational processes is paramount in numerical infrastructure development, particularly in applications such as the finite element method (FEM) and extensive calculations for simulating 3D processes in materials. In this work, we introduce a novel technique for computing the magnetostatic field of an ellipsoid particle, leveraging CUDA on a graphical card for parallel processing. The implementation on a GPU resulted in a remarkable 20-fold improvement in calculation speed. This achievement not only expedites research tasks, but also enables the exploration of larger and more intricate simulations, facilitating quicker model refinements and deeper insights into material behaviours under various conditions. The utilization of GPU computing aligns with the broader trend in scientific research and engineering, offering a versatile solution for diverse computational challenges beyond this specific task of magnetism. Overall, our work contributes to the ongoing effort to harness high-performance computing (HPC) technologies for accelerated and more efficient simulations in materials science and related fields.

TURBOMOLE: Today and Tomorrow.

TURBOMOLE is a highly optimized software suite for large-scale quantum-chemical and materials science simulations of molecules, clusters, extended systems, and periodic solids. TURBOMOLE uses Gaussian basis sets and has been designed with robust and fast quantum-chemical applications in mind, ranging from homogeneous and heterogeneous catalysis to inorganic and organic chemistry and various types of spectroscopy, light-matter interactions, and biochemistry. This Perspective briefly surveys TURBOMOLE's functionality and highlights recent developments that have taken place between 2020 and 2023, comprising new electronic structure methods for molecules and solids, previously unavailable molecular properties, embedding, and molecular dynamics approaches. Select features under development are reviewed to illustrate the continuous growth of the program suite, including nuclear electronic orbital methods, Hartree-Fock-based adiabatic connection models, simplified time-dependent density functional theory, relativistic effects and magnetic properties, and multiscale modeling of optical properties.

Analysis of high-temperature performance of polymer-modified asphalts through molecular dynamics simulations and experiments

• Propose a new polymer-modified asphalt using method of molecular dynamics. • Analyzed the solubility parameters and interaction energy calculated. • Evaluate the compatibility index of polymer-modified asphalt. • Compared and analyzed the temperature sensitivity and high temperature properties. • Provided a new idea and direction of high-performance asphalt pavement. The low durability of asphalt pavements is a key problem limiting successful highway operation and construction; the high-temperature stability of asphalt is an important factor affecting asphalt-pavement durability. To improve the thermal stability of base asphalt, a new polymer-modified asphalt is developed by using polyurethane (PU) as a modifier. Moreover, dynamic research is conducted on the microscopic and mechanical properties of asphalt through molecular dynamics simulations, and the high-temperature performance of the new polymer-modified asphalt is comprehensively evaluated and analyzed through micro- and macro-tests. First, through molecular dynamics and using the three-dimensional materials science simulation software, an asphalt molecular model, a PU molecular model, and an asphalt PU Blend System model are constructed and the correctness of the models is verified. Second, the compatibility index of polymer-modified asphalt is evaluated using solubility parameters and interaction energies calculated in the simulation. Finally, the temperature sensitivity and high-temperature properties of Matrix502-modified asphalt and polymer-modified asphalt are compared and analyzed by using the Brookfield viscosity method and through thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC). The results show that the self-made new polymer-modified asphalt exhibits good thermal storage stability and good high-temperature stability. In addition, a dynamic study on the microcosmic and mechanical properties of the molecules in the asphalt system is conducted through molecular dynamics simulations, and comprehensive evaluation and analysis of the asphalt performance are performed in combination with the microscopic and macroscopic tests. This study promote the research and development of high-performance materials for asphalt pavements.

Tailoring Atomistic Interactions in Li-S Battery Via a Computational Multi-Scale-Data-Driven Approach

A major breakthrough in battery materials is required to meet the ever-increasing proliferation of portable electronic devices, electric vehicles and their variants, as well as the need for incorporating renewable energy resources into the main energy supply [1]. In this context, lithium-sulfur (Li-S) batteries attract attention owing to their very high energy density (2,600 Wh kg-1) and specific capacity (1,675 mAh g-1) and significantly lower weight and cost, compared to lithium-ion batteries (LIBs) [2]. Fully packaged, it is expected that future Li-S batteries can operate at close to 500 Wh kg-1, which is more than twice the energy density of LIBs (200 Wh kg-1) [3]. The problem of realizing the expected high energy density is defined by several issues including the dissolution of Li-Polysulfide (PS) species into the electrolyte, insulating properties of sulfur and Li-PS species, and volume change at the cathode [4]. Overcoming these challenges requires a fundamental understanding of the interplay between events occurring over wide spatial and temporal scales, and accurate prediction of electrode and electrolyte properties to obtain design metrics for new improved materials.In this talk, I will first discuss the details of a high-throughput multi-scale computational infrastructure developed by our group called MISPR (Materials Informatics for Structure-Property-Relationship) to design optimal materials for different applications such as Li-S batteries. MISPR is a high-fidelity and robust computational tool that seamlessly integrates classical molecular dynamics (MD) simulations with density functional theory (DFT) calculations through force field generation and information flow between the two-length scales. It allows automatic handling of thousands of computational materials science simulations and multiple systems with a strong focus on data provenance.I will then discuss the usage of MISPR to mitigate dissolution of LiPS species by altering the atomistic interactions between the electrolyte components through high-throughput screening of potential co-solvent molecules. This approach guides and accelerates our rational selection of co-solvents that enable optimal compromise between the solubility of PS species and the transport properties of the electrolyte through automated DFT calculations. We use the selected candidates in detailed MD studies to comprehend the relationship between the structure of the co-solvent and the electrolyte properties. The approach allows for creating a database of well-characterized materials to be used in machine learning-based methods as well as for testing computationally identified structures in experiments. This work provides crucial information to alleviate the dissolution of PS species during cycling, which is the main reason for rapid capacity decay in Li-S batteries.References Larcher, D. and J.-M. Tarascon, Towards greener and more sustainable batteries for electrical energy storage. Nature chemistry, 2015. 7(1): p. 19. Manthiram, A., X. Yu, and S. Wang, Lithium battery chemistries enabled by solid-state electrolytes. Nature Reviews Materials, 2017. 2(4): p. 1-16. Fang, R., et al., More reliable lithium‐sulfur batteries: status, solutions and prospects. Advanced materials, 2017. 29(48): p. 1606823. Manthiram, A., Y. Fu, and Y.-S. Su, Challenges and prospects of lithium–sulfur batteries. Accounts of chemical research, 2013. 46(5): p. 1125-1134. Figure 1

CORE-MD II: A fast, adaptive, and accurate enhanced sampling method.

In this paper, we present a fast and adaptive correlation guided enhanced sampling method (CORE-MD II). The CORE-MD II technique relies, in part, on partitioning of the entire pathway into short trajectories that we refer to as instances. The sampling within each instance is accelerated by adaptive path-dependent metadynamics simulations. The second part of this approach involves kinetic Monte Carlo (kMC) sampling between the different states that have been accessed during each instance. Through the combination of the partition of the total simulation into short non-equilibrium simulations and the kMC sampling, the CORE-MD II method is capable of sampling protein folding without any a priori definitions of reaction pathways and additional parameters. In the validation simulations, we applied the CORE-MD II on the dialanine peptide and the folding of two peptides: TrpCage and TrpZip2. In a comparison with long time equilibrium Molecular Dynamics (MD), 1 µs replica exchange MD (REMD), and CORE-MD I simulations, we find that the level of convergence of the CORE-MD II method is improved by a factor of 8.8, while the CORE-MD II method reaches acceleration factors of ∼120. In the CORE-MD II simulation of TrpZip2, we observe the formation of the native state in contrast to the REMD and the CORE-MD I simulations. The method is broadly applicable for MD simulations and is not restricted to simulations of protein folding or even biomolecules but also applicable to simulations of protein aggregation, protein signaling, or even materials science simulations.

A modeling approach for unveiling adsorption of toxic ions on iron oxide nanocrystals

The era of advanced computer simulations in materials science enables a great potential to design in silico computational experiments for (nano-)material performance. The adsorption efficiency of nanoparticles in various environments can be unveiled by atomistic models and computer simulations. Arsenic (As) is one of the important globally distributed contaminants with a hazardous impact on human health and environment, and it can strongly bind with iron nanocrystals (e.g., hematite (Fe2O3)) depending on their shape and size. Here, we developed a novel Kinetic Monte Carlo (KMC) model capable of exploring and delineating shape-efficiency dependence for Fe2O3 nanocrystals in contact with arsenate-contaminated water. This newly designed model demonstrated the performance of nanocrystals for removal of toxic (As) ions on their surface. The current model opens new avenues for designing further advanced KMC models for nanoparticles-toxic ions interactions, under varying environmentally relevant situations, e.g., groundwater, wetlands, and water treatment systems. In addition to bidentate adsorption complexes, implemented in the model presented, monodentate and outer-sphere adsorption complexes should be incorporated into the KMC model. Detailed environmental controls can be addressed by implementation of pH and background ions.

Growing field of materials informatics: databases and artificial intelligence

Abstract

Modeling on the molar volume of the Al–Cu–Mg–Si system

Molar volume is an important parameter for the microstructure evolution simulation in materials science. In this work, a comprehensive study on the temperature– and composition–dependents of the molar volume in the Al–Cu–Mg–Si system were performed using a CALPHAD (CALculation of PHAse Diagram)–type formalism. Based on the critical review of the available data on molar volume, molar volume of pure elements was described with a thermal expansion relation. Subsequently, molar volume of the binary solution phases was assessed via CALPHAD technique and the obtained description can reproduce the corresponding experimental data reasonably. Then, molar volume of ternary and higher–order solution phases was extrapolated from the binaries. Molar volumes of the stable compounds in the Al–Cu–Mg–Si system are also assessed in the similar model of the corresponding Gibbs energy expression. Using the established database for the quaternary Al–Cu–Mg–Si system, molar volume of some commercial aluminum alloys was predicted, which shows a good agreement with experimental data from literature.

Mesoscale understanding of ionic conduction in yttria stabilized zirconia: the nanoscale percolation network and its effect on O2− ion movement**Invited paper for focus issue in Modeling and Simulation in Materials Science and Engineering (MSMSE).

Yttria-stabilized zirconia (YSZ) is widely used as a fast oxygen ion conductor in solid oxide fuel cells. Over the years several studies have probed the effect of different cation arrangements on the O2− ion hopping behavior at the atomistic scale. However, an analytical model that can predict the macroscopic ionic conductivity using both the atomic scale hopping behavior as well as the spatial arrangement of cations at the mesoscopic length scale is lacking. A novel mesoscale model is constructed as a step towards addressing this gap. First, using a kinetic Monte Carlo (KMC) model for YSZ we find evidence of a fast ion conducting percolation network being present. The tortuous network, which consists of connected regions spanning the material structure, mainly contributes to the ionic conduction as O2− ion movement in other regions is relatively slow. The topology, composition and O2− ion movement in the network are analyzed. Next, the shortest path lengths in the network are identified with the help of the Dijsktra algorithm. Finally, a diffusion model is developed that relates the atomic scale hopping rates and shortest path lengths (a mesoscale feature) to the macroscale ionic conductivity. Estimates for ionic conductivity from the diffusion model are in excellent agreement with the KMC model. Changes within the percolation network with increasing Y2O3 content can describe the maximum observed in ionic conductivity.

Slip band formation at free surface of lath martensite in low carbon steel

Cross-sectional observations of the deformation bands formed during in situ tensile observation at the free surface of both ultralow- and low-carbon lath martensitic steels were conducted. The cross-sectional observations clearly revealed the formation of steps along substructure boundaries of the ultralow-carbon lath martensite during deformation. In contrast, severe undulation indicating strain accumulation was observed in the vicinity of high-angle boundaries for the low-carbon lath martensite, and no clear steps were formed along substructure boundaries, even though small filmlike ferrites were found. The “greasy-film” mechanism proposed by Maresca [Maresca et al., Modelling and Simulation in Materials Science and Engineering 24(2016)025006, Maresca and Curtin, Acta Materialia 134(2017)302] as well as the plastic instability theory proposed by Dao and Asaro [Dao and Asaro: Mech. Mater., 23(1996)71] was examined considering the non-Schmid effects and the anisotropic activation of slip systems in lath martensite.

Atomate: A high-level interface to generate, execute, and analyze computational materials science workflows

We introduce atomate, an open-source Python framework for computational materials science simulation, analysis, and design with an emphasis on automation and extensibility. Built on top of open source Python packages already in use by the materials community such as pymatgen, FireWorks, and custodian, atomate provides well-tested workflow templates to compute various materials properties such as electronic bandstructure, elastic properties, and piezoelectric, dielectric, and ferroelectric properties. Atomate also enables the computational characterization of materials by providing workflows that calculate X-ray absorption (XAS), Electron energy loss (EELS) and Raman spectra. One of the major features of atomate is that it provides both fully functional workflows as well as reusable components that enable one to compose complex materials science workflows that use a diverse set of computational tools. Additionally, atomate creates output databases that organize the results from individual calculations and contains a builder framework that creates summary reports for each computed material based on multiple simulations.

Task-based Parallel Computation of the Density Matrix in Quantum-based Molecular Dynamics using Graph Partitioning

Quantum-based molecular dynamics (QMD) is a highly accurate and transferable method for material science simulations. However, the time scales and system sizes accessible to QMD are typically limited to picoseconds and a few hundred atoms. These constraints arise due to expensive self-consistent ground-state electronic structure calculations that can often scale cubically with the number of atoms. Linearly scaling methods depend on computing the density matrix P from the Hamiltonian matrix H by exploiting the sparsity in both matrices. The second-order spectral projection (SP2) algorithm is an O(N) algorithm that computes P with a sequence of 40-50 matrix-matrix multiplications. In this paper, we present task-based implementations of a recently developed data-parallel graph-based approach to the SP2 algorithm, G-SP2. We represent the density matrix P as an undirected graph and use graph partitioning techniques to divide the computation into smaller independent tasks. The partitions thus obtained are generally not of equal size and give rise to undesirable load imbalances in standard MPI-based implementations. This load-balancing challenge can be mitigated by dynamically scheduling parallel computations at runtime using task-based programming models. We develop task-based implementations of the data-parallel G-SP2 algorithm using both Intel's Concurrent Collections (CnC) as well as the Charm++ programming model and evaluate these implementations for future use. Scaling and performance results of our implementations are investigated for representative segments of QMD simulations for solvated protein systems containing more than 10,000 atoms.

Open source JAVA implementation of the parallel multi-thread alternating direction isogeometric L2 projections solver for material science simulations

Open source JAVA implementation of the parallel multi-thread alternating direction isogeometric L2 projections solver for material science simulations

Nanoparticles on nanosteps: a new study reduces wastage and increases the efficiency of catalysts

Cross-sectional observations of the deformation bands formed during in situ tensile observation at the free surface of both ultralow- and low-carbon lath martensitic steels were conducted. The cross-sectional observations clearly revealed the formation of steps along substructure boundaries of the ultralow-carbon lath martensite during deformation. In contrast, severe undulation indicating strain accumulation was observed in the vicinity of high-angle boundaries for the low-carbon lath martensite, and no clear steps were formed along substructure boundaries, even though small filmlike ferrites were found. The “greasy-film” mechanism proposed by Maresca [Maresca et al., Modelling and Simulation in Materials Science and Engineering 24(2016)025006, Maresca and Curtin, Acta Materialia 134(2017)302] as well as the plastic instability theory proposed by Dao and Asaro [Dao and Asaro: Mech. Mater., 23(1996)71] was examined considering the non-Schmid effects and the anisotropic activation of slip systems in lath martensite.

Molecular Dynamics Simulations of Lithiation of Si Nanowires Covered By SiO2 and Al2O3 Shells

Introduction The use of high capacity Si based electrode has been hampered by its mechanical degradation due to the large volume expansion/contraction during cycling. Nanostructured Si can effectively avoid Si cracking/fracture. Unfortunately, the high surface to volume ratio in nanostructures leads to unacceptable amount of solid-electrolyte interphase (SEI) formation and growth, thereby low current/coulombic efficiency and short life. Based on mechanics models we demonstrate that the artificial SEI coating can be mechanically stable despite the volume change in Si, if the material properties, thickness of the SEI, and the size/shape of Si are optimized. Therefore, the objective of this study is to develop a modeling approach to understand, design, and make coated Si anode structures with high current efficiency and stability. For the purpose, in this study, reactive molecular dynamics simulations are conducted to mimic the lithiation process of amorphous Al2O3 and SiO2 covered nanowires. Amorphous SiO2 represents natrally formed oxides on Si nanowire and Al2O3 represents the atomic layer deposition ALD-Al2O3 coating on Si nanowires. Methods In order to simulate this system, the Li-Si-Al-O parameters in ReaxFF were newly developed, by adding new training sets of DFT results, such as Li ion diffusion energy barrier and lithiation energy in various Li-Si-O and Li-Al-O and Li-Si-Al-O systems (1). After finishing the development of Li-Si-Al-O parameters of ReaxFF, the reactive molecular dynamics simulations were perfrmed on Si core-Al2O3 shell and Si core-SiO2 shell structures surrounded by Li atoms. Behavior upon lithiation to Si core and coating itself were compared for different shell thickness. More specificly, six simulation sets were prepared to investigate coating and shell thickness effect upon lithiation to Si core. Simulation periodic box size is 78Å×53Å×71Å. The radius of Si core is 7.5Å. The thicknesses of shells are 2.5 Å, 4.5 Å and 7.5 Å for SiO2 and Al2O3, respectively. The length of the core shell cylinder is the same as the length of the y dimension of the simulation box in order to lithium insertion happen only from the side of the cylinder model (Figure 1). The amorphous Si core consists of 478 Si atoms. NVT ensemble is chosen for the simulation and the temperature is 900K. Results The coating structural integrity and stability vs. its thickness were investigated. Upon lithiation to Si-core/SiO2-shell, SiO2 shell was immediately lithiated without much volume expansion, followed by the lithiation of Si core. When the thickness of SiO2 is comparable to the radius of Si, the lithiated shell has a few cracks, which cause Si atoms smear out from the Si core. For thinner shell, the SiO2 shell spread out and loses its structural integrity. When it comes to Si-core/Al2O3 shell, it behaves totally differently depending on the thickness of the coatings. When Al2O3 is thin, it behaves like SiO2 shell. It is instantaneously lithiated and subsequently causes rapid lithiation of Si core, which induces sudden volume expansion and gives large stress to the coating. This large stress gets the coating to be broken. When Al2O3 coating is as thick as the radius of Si core, lithiation happens relatively gradually making well defined Li density gradient inside the coating. This gradient comes from the Li migration energy barrier decrease when Al2O3 changes into LiAlO2. In other words, much more Li exist in outer shell compared to inner shell, which makes mechanical property of inner shell region to be stiff, while outer region become soft. This inner stiff region can suppress the lithiated Si core and minimize volume expansion and the coating property which has elastic modulus gradient is strong to external stress therefore, it maintain its integrity at the end of the simulation. Conclusion Al2O3 can be used as an excellent ALD-coating material if it is thick enough to hold elastic modulus gradient in it. Based on the simulation results of Al2O3 coated Si, we propose a modulus gradient coating, softer outside, harder inside, is beneficial to avoid coating cracking.

Divide-and-conquer quantum mechanical material simulations with exascale supercomputers

Abstract Recent developments in large-scale materials science simulations, especially under the divide-and-conquer method, are reviewed. The pros and cons of the divide-and-conquer method are discussed. It is argued that the divide-and-conquer method, such as the linear-scaling 3D fragment method, is an ideal approach to take advantage of the heterogeneous architectures of modern-day supercomputers despite their relatively large prefactors among linear-scaling methods. Some developments in graphics processing unit (GPU) electronic structure calculations are also reviewed. The accelerators like GPU could be an essential part for the future exascale supercomputing.

Energy, structure and vibrational modes of small water clusters by a simple many-body potential mimicking polarisation effects

An empirical many-body model potential able to mimic polarisation effects is applied to compute cohesive, structural and vibrational properties of water clusters with up to 12 H2O molecules. The model introduces local coordination functions to account for the variation of charges and other intra- and inter-molecular force constants upon formation of hydrogen bonds among water molecules. The potential is tuned to fit the results of state of the art density functional computations, and it is shown to accurately reproduce cohesive energies, bond lengths and vibrational properties of clusters. Moreover, it reproduces the marked increase of the molecular dipole moment with increasing water–water coordination. At variance from traditional polarisable models, the energy is an explicit function of the atomic coordinates, and does not require the minimisation of the electrostatic energy or the equalisation of the electron chemical potential, and thus is suitable for large-scale simulations in materials science and in bio-chemistry/bio-physics.

The analysis of a plane wave pseudopotential density functional theory code on a GPU machine

Plane wave pseudopotential (PWP) density functional theory (DFT) calculation is the most widely used material science simulation, and the PWP DFT codes are arguably the most important material science codes. We have implemented a PWP DFT code PEtot on a multi-node GPU machine. Starting from a previous work, we have further improved the speed of the code, and achieved x13-x22 speedups over the CPU calculations for a typical 512 atom system. Such speedups are much higher than other similar works for this important class of material simulation codes on GPU clusters. The current achievement is obtained by (1) moving the calculation fully into the GPU; (2) adopting a new algorithm to reduce the data amount for MPI communication; and (3) using new GPU and CPU numerical libraries. We have also provided a detail quantitative analysis of the computational times for different physical systems and number of GPU units, which helps one to understand the challenges and bottlenecks of the PWP DFT simulations on GPU machines. Based on the analysis, we listed the machine and library requirements in order to further improve the performances of the PWP DFT calculations.