Computational Materials Design Research Articles

Multiscale Analysis of High Damping Composites Using the Finite Cell and the Mortar Method

Metal lattice structures filled with a damping material such as polymer can exhibit high stiffness and good damping properties. Mechanical simulations of parts made from these composites can however require a large modeling and computational effort because relevant features such as complex geometries need to be represented on multiple scales. The finite cell method (FCM) and numerical homogenization are potential remedies for this problem. Moreover, if the microstructures are placed in between the components of assemblies for vibration reduction, a modified mortar technique can further increase the efficiency of the complete simulation process. With this method, it is possible to discretize the components separately and to integrate the viscoelastic behavior of the composite damping layer into their weak coupling. This paper provides a multiscale computational material design framework for such layers, based on FCM and the modified mortar technique. Its efficiency even in the case of complex microstructures is demonstrated in numerical studies. Therein, computational homogenization is first performed on various microstructures before the resulting effective material parameters are used in larger-scale simulation models to investigate their effect and to verify the employed methods.

Machine-learned potentials for next-generation matter simulations.

The choice of simulation methods in computational materials science is driven by a fundamental trade-off: bridging large time- and length-scales with highly accurate simulations at an affordable computational cost. Venturing the investigation of complex phenomena on large scales requires fast yet accurate computational methods. We review the emerging field of machine-learned potentials, which promises to reach the accuracy of quantum mechanical computations at a substantially reduced computational cost. This Review will summarize the basic principles of the underlying machine learning methods, the data acquisition process and active learning procedures. We highlight multiple recent applications of machine-learned potentials in various fields, ranging from organic chemistry and biomolecules to inorganic crystal structure predictions and surface science. We furthermore discuss the developments required to promote a broader use of ML potentials, and the possibility of using them to help solve open questions in materials science and facilitate fully computational materials design.

Putting Density Functional Theory to the Test in Machine-Learning-Accelerated Materials Discovery.

Accelerated discovery with machine learning (ML) has begun to provide the advances in efficiency needed to overcome the combinatorial challenge of computational materials design. Nevertheless, ML-accelerated discovery both inherits the biases of training data derived from density functional theory (DFT) and leads to many attempted calculations that are doomed to fail. Many compelling functional materials and catalytic processes involve strained chemical bonds, open-shell radicals and diradicals, or metal-organic bonds to open-shell transition-metal centers. Although promising targets, these materials present unique challenges for electronic structure methods and combinatorial challenges for their discovery. In this Perspective, we describe the advances needed in accuracy, efficiency, and approach beyond what is typical in conventional DFT-based ML workflows. These challenges have begun to be addressed through ML models trained to predict the results of multiple methods or the differences between them, enabling quantitative sensitivity analysis. For DFT to be trusted for a given data point in a high-throughput screen, it must pass a series of tests. ML models that predict the likelihood of calculation success and detect the presence of strong correlation will enable rapid diagnoses and adaptation strategies. These "decision engines" represent the first steps toward autonomous workflows that avoid the need for expert determination of the robustness of DFT-based materials discoveries.

Sequence-Engineering Polyethylene-Polypropylene Copolymers with High Thermal Conductivity Using a Molecular-Dynamics-Based Genetic Algorithm.

Polymer sequence engineering is emerging as a potential tool to modulate material properties. Here, we employ a combination of a genetic algorithm (GA) and atomistic molecular dynamics (MD) simulation to design polyethylene-polypropylene (PE-PP) copolymers with the aim of identifying a specific sequence with high thermal conductivity. PE-PP copolymers with various sequences at the same monomer ratio are found to have a broad distribution of thermal conductivities. This indicates that the monomer sequence has a crucial effect on thermal energy transport of the copolymers. A non-periodic and non-intuitive optimal sequence is indeed identified by the GA, which gives the highest thermal conductivity compared with any regular block copolymers, for example, diblock, triblock, and hexablock. In comparison to the bulk density, chain conformations, and vibrational density of states, the monomer sequence has the strongest impact on the efficiency of thermal energy transport via inter- and intra-molecular interactions. Our work highlights polymer sequence engineering as a promising approach for tuning the thermal conductivity of copolymers, and it provides an example application of integrating atomistic MD modeling with the GA for computational material design.

Automated coordination corrected enthalpies with AFLOW-CCE

The computational design of materials with ionic bonds poses a critical challenge to thermodynamic modeling since density functional theory yields inaccurate predictions of their formation enthalpies. Progress requires leveraging physically insightful correction methods. The recently introduced coordination corrected enthalpies (CCE) method delivers accurate formation enthalpies with mean absolute errors close to room temperature thermal energy, i.e., 25meV/atom. The CCE scheme, depending on the number of cation-anion bonds and oxidation state of the cation, requires an automated analysis of the system to determine and apply the correction. Here, we present AFLOW-CCE -- our implementation of CCE into the AFLOW framework for computational materials design. It features a command line tool, a web interface and a Python environment. The workflow includes a structural analysis, automatically determines oxidation numbers, and accounts for temperature effects by parametrizing vibrational contributions to the formation enthalpy per bond.

Extending Shannon's ionic radii database using machine learning

In computational material design, ionic radius is one of the most important physical parameters used to predict material properties. Motivated by the progress in computational materials science and material informatics, we extend the renowned Shannon's table from 475 ions to 987 ions. Accordingly, a rigorous Machine Learning (ML) approach is employed to extend the ionic radii table using all possible combinations of Oxidation States (OS) and Coordination Numbers (CN) available in crystallographic repositories. An ionic-radius regression model for Shannon's database is developed as a function of the period number, the valence orbital configuration, OS, CN, and Ionization Potential. In the Gaussian Process Regression (GPR) model, the reached R-square $R^2$ accuracy is 99\% while the root mean square error of radii is 0.0332 \AA. The optimized GPR model is then employed for predicting a new set of ionic radii for uncommon combinations of OS and CN extracted by harnessing crystal structures from materials project databases. The generated data is consolidated with the reputable Shannon's data and is made available online in a database repository \url{https://cmd-ml.github.io/}.

BLAST: bridging length/timescales via atomistic simulation toolkit

The ever-increasing power of supercomputers coupled with highly scalable simulation codes has made molecular dynamics an indispensable tool in applications ranging from predictive modeling of materials to computational design and discovery of new materials for a broad range of applications. Multi-fidelity scale bridging between the various flavors of molecular dynamics, i.e., ab-initio, classical, and coarse-grained models has remained a long-standing challenge. Here, we introduce our framework BLAST (Bridging Length/Timescales via Atomistic Simulation Toolkit) that leverages machine learning principles to address this challenge. BLAST is a multi-fidelity scale bridging framework that provides users with the capabilities to train and develop their own classical atomistic and coarse-grained interatomic potentials (force fields) for molecular simulations. BLAST is designed to address several long-standing problems in the molecular simulation community, such as unintended misuse of existing force fields due to knowledge gap between developers and users, bottlenecks in traditional force field development approaches, and other issues relating to the accuracy, efficiency, and transferability of force fields. Here, we discuss several important aspects in force field development and highlight features in BLAST that enable its functionalities and ease of use.

Predicting the Outcomes of Material Syntheses with Deep Learning

A key bottleneck for material discovery is synthesis. While significant advances have been made in computational material design, synthesis pathways are still often determined through trial and error. In this work, we develop a method that predicts the major product of solid-state reactions. The main advance presented here is the construction of fixed-length, learned representations of reactions. Precursors are represented as nodes on a “reaction graph”, and message-passing operations between nodes are used to embody the interactions between precursors in the reaction mixture. We show that this deep learning framework not only outperforms baseline methods but also more reliably assesses the uncertainty in its predictions. Moreover, our approach establishes a quantitative metric for inorganic reaction similarity, allowing the user to explain model predictions and retrieve relevant literature sources.

Identification of synthesisable crystalline phases of water – a prototype for the challenges of computational materials design

We discuss the identification of experimentally realisable crystalline phases of water to outline and contextualise some of the diverse building blocks of a computational materials design process.

Extended finite element modeling of fatigue crack growth microstructural mechanisms in alloys with secondary/reinforcing phases: model development and validation

Light structural metals have been extensively applied throughout the transportation sector in recent years with greater impetus to reduce vehicle weight and enhance energy efficiency. However, their use is restricted by the engineering challenge of fatigue crack growth and the need to understand, simulate, and predict crack propagation mechanisms with respect to materials’ microstructure. To address this need, a comprehensive computational methodology has been developed using the extended finite element method to predict fatigue crack interaction mechanisms with characteristic microstructural features. Mathematical formulations and algorithmic development are rigorously addressed, with specific emphasis on the incorporation of relevant physical phenomena of plasticity and particle debonding/fracture. This approach is validated by comparison with analytical models and experiments on cast aluminum–silicon alloys using digital image correlation. The proposed methodology constitutes a framework for the successful development, application, and advancement of computational design of materials. Ultimately, this contributes to material/process selection and design for structural integrity by enabling rapid assessment of fatigue crack growth resistance without prior testing, thereby reducing of the extent of costly experimental investigations.

Computational materials design: recent advances and open challenges related to colorimetry

AbstractComputational materials design aims at designing, simulating and predicting innovative materials. This paper reviews a few open challenges in the domain of computational materials design related to the colour appearance of materials. These relatively recent fields of research necessitate revisiting the fundamentals of colorimetry, such as reflectance models. First, we address the virtual design and the simulation of innovative materials, which requires modelling their appearance and fundamental properties. Then we discuss the latest advancements in the machine‐learning domain that have highly revolutionised computational and data‐minded methodologies, which are used for the design innovation, discovery and optimisation of materials.

CALPHAD Uncertainty Quantification and TDBX

CALPHAD uncertainty quantification (UQ) is the foundation of materials design with quantified confidence. We report a framework and software packages to enable CALPHAD UQ assessment and calculation using commercial CALPHAD software (Thermo-Calc). This Bayesian inference framework is coupled with a Markov chain Monte Carlo algorithm to establish uncertainty traces with a given thermodynamic database file (TDB) and corresponding experimental data points. This general framework is demonstrated with the Ni–Cr binary system. The algorithm is firstly validated on synthetic data with known ground truth. Then it is applied to real experimental data to generate posterior traces. We develop a file format named TDBX, which provides a single source of truth by combining the original TDB content and the traces for each assessed Gibbs energy parameter. CALPHAD UQ calculations are performed based on the TDBX file, from which uncertainties for phase boundaries, enthalpy curves, and solidification range are collected as examples of basic design parameters. This TDBX file with corresponding scripts are made open-source. The combination of CALPHAD UQ assessments and calculations connected by TDBX supports uncertainty-assisted modeling, enabling the integrated application of modern design with uncertainty methodologies to computational materials design.

Langer\u2013Schwartz\u2013Kampmann\u2013Wagner precipitation simulations: assessment of models and materials design application for Cu precipitation in PH stainless steels

Quantitative modelling of precipitation kinetics can play an important role in a computational material design framework where, for example, optimization of alloying can become more efficient if it is computationally driven. Precipitation hardening (PH) stainless steels is one example where precipitation strengthening is vital to achieve optimum properties. The Langer–Schwartz–Kampmann–Wagner (LSKW) approach for modelling of precipitation has shown good results for different alloy systems, but the specific models and assumptions applied are critical. In the present work, we thus apply two state-of-the-art LSKW tools to evaluate the different treatments of nucleation and growth. The precipitation modelling is assessed with respect to experimental results for Cu precipitation in PH stainless steels. The LSKW modelling is able to predict the precipitation during ageing in good quantitative agreement with experimental results if the nucleation model allows for nucleation of precipitates with a composition far from the equilibrium and if a composition-dependent interfacial energy is considered. The modelling can also accurately predict trends with respect to alloy composition and ageing temperature found in the experimental data. For materials design purposes, it is though proposed that the modelling is calibrated by measurements of precipitate composition and fraction in key experiments prior to application.Graphic abstract

Predicting strain-induced martensite in austenitic steels by combining physical modelling and machine learning

Computational materials design has made significant progress lately. However, one underexploited opportunity lies in the combination of physically based modelling and machine learning (ML). In the present work we exploit this combination for modelling of strain-induced martensitic phase transformation (SIMT) in austenitic steels. A fully predictive model for SIMT, responsible for the TRIP effect in many steels, is devised. An experimental dataset correlating SIMT with composition, temperature and strain is collected from the open literature firstly. Secondly, the Olson-Cohen model is applied to make physically based predictions on temperature and strain dependence of SIMT in order to expand the database to the final size of 16,500 entries relating the features and the target. Thirdly, ensemble ML methods are applied to model the data and the final model is validated on a holdout dataset, including also dual-phase alloys. The final model provides accurate predictions of SIMT in a temperature range from −196 to 100 °C and from 0 to 1 in strain. The model can readily be extended to consider further factors such as strain rate and stress state. Moreover, it can be used together with thermodynamic and kinetic calculations, or thermomechanical simulations, for the design of steels and components, respectively.

ALKEMIE: An intelligent computational platform for accelerating materials discovery and design

Developing new materials with target properties via the traditional trial-and-error ways is cost-inefficient, and sometimes ends up with fruitlessness, therefore, simulation-driven materials design plays an important role in the past decades. Nevertheless, the advent of the era of data-driven material science requires an intelligent computational platform to accelerate the discovery and design for advanced materials. Here, we present an open-source computational platform named as ALKEMIE, acronyms for Artificial Learning and Knowledge Enhanced Materials Informatics Engineering, which enables easy access of data-driven techniques to broad communities. ALKEMIE is incorporated with three key components for the computational design of materials for the forthcoming data-driven sciences: data generation via high-throughput calculations, data management and data mining via machine learning models. Briefly speaking, the high-throughput calculations in ALKEMIE are implemented through the integration of automatic frameworks of model constructions, calculation performances and data analysis. And the used high-level application programming interface for the database makes the data mining through machine learning more applicable in material science. In particular, ALKEMIE is integrated with a module for the generation of machine-learned interatomic potential for large-scale molecular dynamic simulations where the dataset is obtained from high-throughput first-principles calculations. More importantly, ALKEMIE has an elaborately designed user-friendly graphical user-interface which makes the workflow and dataflow more maneuverable and transparent, facilitating its easy-to-use for scientists with broad backgrounds. Finally, the main characters of ALKEMIE are demonstrated using three computational examples.

Elucidating energy scaling between atomic and molecular adsorbates in the presence of solvent

AbstractCondensed phase reactions have recently attracted increased interest, but principles for efficiently screening and designing catalyst materials through computations are lacking. In this study, we examine the applicability of energy correlations between adsorbed surface species, which have been instrumental in accelerating the computational design of catalyst materials in gas‐phase contexts, in various representations of a condensed phase reaction environment. We perform detailed density functional theory calculations of the adsorption of atomic and molecular species in the presence of various representations of solvent species. Our results show that the well‐known scaling in the gas phase context is preserved, with scaling slopes unaffected by the adsorbate‐liquid interactions. Moreover, these results hold when changing surface structure, solvent identity, and even in highly disordered environments. We envision the establishment of an energy scaling framework for condensed phase reactions to accelerate catalyst discovery and design in those contexts.

Designing piezoresistive materials from first-principles: Dopant effects on 3C-SiC

The piezoresistive effect is widely exploited in Si-based sensor technology, in part due to Si’s high gauge factor. However, for harsh environments, wider bandgap materials, such as SiC, are desirable. Despite the quest to improve its piezoresistance, SiC’s gauge factor remains low, although recent experiments hint towards doping as a viable strategy. To search for the best dopant, we developed a computational framework, based on density functional theory (DFT) and Boltzmann Transport calculations, to compute the piezoresistive properties of materials and apply it to predict the gauge factor and piezoresistive coefficient of doped-SiC. Under rigid band approximation, predicted gauge factor of lightly-doped SiC (1.0 × 1018 cm−3) agrees reasonably with experimental values. For moderately doped SiC, we explicitly substitute the doping element into a 64-atom supercell of SiC and compute the piezoresistive properties. We discover that effective dopants consist of elements that could exist in +4 oxidation states. Among them, Ru leads to a 4-fold increase in gauge factor, while Mo and Pt are peculiarly suitable for high-temperature use, where their gauge factors increase with temperature. From bandstructure analysis, we elucidate that promising dopants, under strain, impact resistivity through changes in band positions and curvature. We also discuss the formability of the doped-SiC systems. Furthermore, we demonstrate correlations among piezoresistive properties and their dependence on crystallographic orientations and temperature. The framework serves as a starting point for rational computational design of piezoresistive materials.

Integrated Computational Material Design for PMC Manufacturing with Trapped Rubber.

As the use of continuous fiber polymer matrix composites expands into new fields, there is a growing need for more sustainable manufacturing processes. An integrated computational material design framework has been developed, which enables the design of tailored manufacturing systems for polymer matrix composite materials as a sustainable alternative to achieving high-quality components in high-rate production. Trapped rubber processing achieves high pressures during polymer matrix composite processing, utilizing the thermally induced volume change of a nearly incompressible material inside a closed cavity mold. In this interdisciplinary study, the structural analysis, material science and manufacturing engineering perspectives are all combined to determine the mold mechanics, and the manufacturing process in a cohesive and iterative design loop. This study performs the coupled thermo-mechanical analysis required to simulate the transients involved in composite manufacturing and the results are compared with a previously developed test method. The internal surface pressure and temperatures are computed, compared with the experimental results, and the resulting design process is simulated. Overall, this approach maintains high-quality consolidation during curing while allowing for the possibility for custom distributions of pressures and temperatures. This can lead to more sustainable manufacturing by reducing energy consumption and improving throughput.

High‐throughput computational design of halide perovskites and beyond for optoelectronics

AbstractHalide perovskites have attracted great interest as promising next‐generation materials in optoelectronics, ranging from solar cells to light‐emitting diodes. Despite their exceptional optoelectronic properties and low cost, the prototypical organic–inorganic hybrid lead halide perovskites suffer from toxicity and low stability. Therefore, it is of high demand to search for stable and nontoxic alternatives to the hybrid lead halide perovskites. Recently, high‐throughput computational materials design has emerged as a powerful approach to accelerate the discovery of new halide perovskite compositions or even novel compounds beyond perovskites. In this review, we discuss how this approach discovers halide perovskites and beyond for optoelectronics. We first overview the background of halide perovskites and methodologies in high‐throughput computational design. Then, we focus on materials properties for different optoelectronic applications, and how they are assessed with materials descriptors. Finally, we review different studies in terms of specific materials types to discuss their design principles, screening results, and experimental verification.This article is categorized under: Structure and Mechanism > Computational Materials Science Electronic Structure Theory > Density Functional Theory

Symmetry-Assisted Protection and Compensation of Hidden Spin Polarization in Centrosymmetric Systems*This work was supported by the National Natural Science Foundation of China (Grant No. 11874195), the Guangdong Provincial Key Laboratory of Computational Science and Material Design (Grant No. 2019B030301001), and the Center for Computational Science and Engineering of SUSTech.

It was recently noted that in certain nonmagnetic centrosymmetric compounds, spin–orbit interactions couple each local sector that lacks inversion symmetry, leading to visible spin polarization effects in the real space, dubbed “hidden spin polarization (HSP)". However, observable spin polarization of a given local sector suffers interference from its inversion partner, impeding material realization and potential applications of HSP. Starting from a single-orbital tight-binding model, we propose a nontrivial way to obtain strong sector-projected spin texture through the vanishing hybridization between inversion partners protected by nonsymmorphic symmetry. The HSP effect is generally compensated by inversion partners near the Γ point but immune from the hopping effect around the boundary of the Brillouin zone. We further summarize 17 layer groups that support such symmetry-assisted HSP and identify hundreds of quasi-2D materials from the existing databases by first-principle calculations, among which a group of rare-earth compounds LnIO (Ln = Pr, Nd, Ho, Tm, and Lu) serves as great candidates showing strong Rashba- and Dresselhaus-type HSP. Our findings expand the material pool for potential spintronic applications and shed light on controlling HSP properties for emergent quantum phenomena.