Discovery Of Materials Research Articles

New Ways to Discover Novel Nonlinear Optical Materials: Scaling Machine Learning with Chemical Descriptors Information.

The efficient experimental exploration of innovative nonlinear optical materials has long been a challenging task due to the vast chemical space and the lack of suitable theoretical prediction frameworks. Herein, a novel theoretical design paradigm is proposed to accelerate the discovery of novel materials with strong second harmonic generation intensity. This challenge is addressed through several key technologies. 1) A high-precision machine learning model is proposed on the maximum nonlinear optical dataset. 2) Descriptors information paves the way to systematically offer valuable chemical insights for designing chemical structures. 3) A flexible and fast chemical space constructionand exploration method is proposed. Accordingly, a nonlinear optical crystal is successfully synthesized through the constructed "machine to knowledge" theoretical framework. This novel compound exhibits a stronger second-harmonic generation response and wider optical transmission range. This work introduces novel theoretical design concepts and provides innovative chemical insights into optical materials or other functional materials.

High-Throughput Screening of 6858 Compounds for Zinc-Ion Battery Cathodes via Hybrid Machine Learning Optimization.

This work presents a machine-learning framework to explore cathode materials for zinc-ion batteries from a data set of 6858 zinc-containing compounds. Utilizing the extensive Materials Project (MP) database, we employed a two-step machine learning (ML) approach that uses transfer learning to compensate for missing electrochemical properties. Initially, a random forest regressor was used to fill in missing features in the zinc compounds, harnessing the full battery explorer in predictions. Two hybrid models were then developed: the sparrow search algorithm-light gradient boosting machine (SSA-LGBM), and Harris Hawk optimization-deep neural networks (HHO-DNN). The data set contains 107 feature vectors, which were minimized through principal component analysis. These features include descriptors related to structural, chemical, and electronic properties. Both models were trained using the 4351 known battery compounds from MP to predict key properties such as average voltage and gravimetric capacity. After initial prediction of 62 potential electrodes, further screening criteria were applied to identify 18 promising electrodes based on their voltage, specific capacity, electronic conductivity, safety, stability, cost, and abundance. The validation of our approach was carried out by applying the models to known cathode materials, verifying the accuracy of the predictions. This innovative approach significantly accelerates the discovery of efficient and stable cathode materials for zinc-ion batteries, paving the way for more sustainable and high-performance energy storage solutions. This method also provides a robust framework for future materials exploration across various battery technologies.

Computational Screening Guiding the Development of a Covalent-Organic Framework-Based Gas Sensor for Early Detection of Lithium-Ion Battery Electrolyte Leakage.

This study presents a computationally guided approach for selecting covalent organic frameworks (COFs) for the selective detection of the trace ethylene carbonate (EC) vapor, a key indicator of electrolyte leakage from lithium-ion batteries (LIBs). High-throughput screening, employing grand canonical Monte Carlo (GCMC) simulation complemented by density functional theory (DFT) calculations, was used to identify potential COF candidates from the CURATED COF database. Among the screened materials, an imine COF functionalized with quaternary ammonium (QA) groups, named COF-QA-4, exhibited a high adsorption capacity (5.88 mmol/g) and selectivity of EC vapor. DFT analysis revealed strong molecular interactions driven by a partial charge transfer mechanism between EC and the COF-QA-4 framework, underpinning its superior adsorption properties. Experimental validation through chemiresistive gas sensors fabricated with COF-QA-4 demonstrated excellent sensitivity and reversibility to 1.15 ppmv of EC vapor, maintaining consistent performance over three response-recovery cycles. This work highlights the potential of computationally guided material discovery for advancing sensor technologies in LIB safety monitoring.

A Review of Machine Learning in Organic Solar Cells

Organic solar cells (OSCs) are a promising renewable energy technology due to their flexibility, lightweight nature, and cost-effectiveness. However, challenges such as inconsistent efficiency and low stability limit their widespread application. Addressing these issues requires extensive experimentation to optimize device performance, a process hindered by the complexity of OSC molecular structures and device architectures. Machine learning (ML) offers a solution by accelerating material discovery and optimizing performance through the analysis of large datasets and prediction of outcomes. This review explores the application of ML in advancing OSC technologies, focusing on predicting critical parameters such as power conversion efficiency (PCE), energy levels, and absorption spectra. It emphasizes the importance of supervised, unsupervised, and reinforcement learning techniques in analyzing molecular descriptors, processing data, and streamlining experimental workflows. Concludingly, integrating ML with quantum chemical simulations, alongside high-quality datasets and effective feature engineering, enables accurate predictions that expedite the discovery of efficient and stable OSC materials. By synthesizing advancements in ML-driven OSC research, the gap between theoretical potential and practical implementation can be bridged. ML can viably accelerate the transition of OSCs from laboratory research to commercial adoption, contributing to the global shift toward sustainable energy solutions.

COMFO: Integrated deep learning model facilitates discovery of multifunctional polyimide materials

COMFO: Integrated deep learning model facilitates discovery of multifunctional polyimide materials

Accelerating active learning materials discovery with FAIR data and workflows: A case study for alloy melting temperatures

Accelerating active learning materials discovery with FAIR data and workflows: A case study for alloy melting temperatures

Transformer-generated atomic embeddings to enhance prediction accuracy of crystal properties with machine learning

Accelerating the discovery of novel crystal materials by machine learning is crucial for advancing various technologies from clean energy to information processing. The machine-learning models for prediction of materials properties require embedding atomic information, while traditional methods have limited effectiveness in enhancing prediction accuracy. Here, we proposed an atomic embedding strategy called universal atomic embeddings (UAEs) for their broad applicability as atomic fingerprints, and generated the UAE tensors based on the proposed CrystalTransformer model. By performing experiments on widely-used materials database, our CrystalTransformer-based UAEs (ct-UAEs) are shown to accurately capture complex atomic features, leading to a 14% improvement in prediction accuracy on CGCNN and 18% on ALIGNN when using formation energies as the target, based on the Materials Project database. We also demonstrated the good transferability of ct-UAEs across various databases. Based on the clustering analysis for multi-task ct-UAEs, the elements in the periodic table can be categorized with reasonable connections between atomic features and targeted crystal properties. After applying ct-UAEs to predict formation energy in hybrid perovskites database, we realized an improvement in accuracy, with a 34% boost in MEGNET and 16% in CGCNN, showcasing their potential as atomic fingerprints to address the data scarcity challenges.

MOSAEC-DB: a comprehensive database of experimental metal-organic frameworks with verified chemical accuracy suitable for molecular simulations.

Ongoing developments in computational databases seek to improve the accessibility and breadth of high-throughput screening and materials discovery efforts. Their reliance on experimental crystal structures necessitates significant processing prior to computation in order to resolve any crystallographic disorder or partial occupancies and remove any residual solvent molecules in the case of activated porous materials. Contemporary investigations revealed that deficiencies in the experimental characterization and computational preprocessing methods generated considerable occurrence of structural errors in metal-organic framework (MOF) databases. The MOSAEC MOF database (MOSAEC-DB) tackles these structural reliability concerns through utilization of innovative preprocessing and error analysis protocols applying the concepts of oxidation state and formal charge to exclude erroneous crystal structures. Comprising more than 124k crystal structures, this work maintains the largest and most accurate dataset of experimental MOFs ready for immediate deployment in molecular simulations. The databases' comparative diversity is demonstrated through its enhanced coverage of the periodic table, expansive quantity of structures, and balance of chemical properties relative to existing MOF databases. Chemical and geometric descriptors, as well as DFT electrostatic potential-fitted charges, are included to facilitate subsequent atomistic simulation and machine-learning (ML) studies. Curated subsets-sampled according to their chemical properties and structural uniqueness-are also provided to further enable ML studies in recognition of the strict demand for duplicate structure elimination and dataset diversity in such applications.

Recent Applications of Theoretical Calculations and Artificial Intelligence in Solid-State Electrolyte Research: A Review

Solid-state electrolytes (SSEs), as key materials for all-solid-state batteries (ASSBs), face challenges such as low ionic conductivity and poor interfacial stability. With the rapid advancement of computational science and artificial intelligence (AI) technologies, theoretical calculations and AI methods are emerging as efficient and important virtual tools for predicting and screening high-performance SSEs. To further promote the development of the SSEs, this review outlines recent applications of theoretical calculations and AI in this field. First, the current applications of theoretical calculation methods, such as density functional theory (DFT) and molecular dynamics (MD), in material structure optimization, electronic property analysis, and ionic transport dynamics are introduced, along with an analysis of their limitations. Second, innovative applications of AI methods, including machine learning (ML) and deep learning (DL), in predicting material properties, analyzing structural features, and simulating interfacial behaviors are elaborated. Subsequently, the synergistic application strategies combining high-throughput screening (HTS), theoretical calculations, and AI methods are highlighted, demonstrating the unique advantages of integrating multiple methodologies in material discovery and performance optimization. Finally, the current research progress is summarized, and future development trends are forecasted. The deep integration of theoretical calculations and AI methods is expected to significantly accelerate the development of high-performance SSE materials, thereby driving the industrial application of ASSBs.

Accelerating Discovery of Infrared Nonlinear Optical Materials with High Lattice Thermal Conductivity: Combining Machine Learning and First‐Principles Calculations

AbstractThe exploration of mid/far‐infrared (mid/far‐IR) nonlinear optical (NLO) materials with high lattice thermal conductivity (LTC) is urgent in advancing laser technology, as LTC enhances resistance to high laser‐induced damage. To address the issue of the existing methods being time‐consuming when measuring LTC, this study employed machine learning (ML) methods to efficiently predict кL based on a high‐quality dataset. Seven ML models are trained, and the most optimal model is selected for identifying promising NLO materials with high LTC (кL > 1.0 W m−1 K−1). The application of the model to predict the LTC led to the discovery of LiZnGaSe3 and BeAl2Se4 as potential chalcogenide candidates from the NLO materials database. The integration of ML methods with first‐principles calculations confirmed the mechanical properties and high LTC of LiZnGaSe3 and BeAl2Se4. The chemical heatmaps of chalcogenides with high LTC are recommended. This approach facilitates the rapid screening and discovery of promising mid/far‐IR NLO materials with balanced properties and complements the кL data in the NLO materials database as an additional screening parameter for material properties.

Materials laboratories of the future for alloys, amorphous, and composite materials

Abstract In alignment with the Materials Genome Initiative and as the product of a workshop sponsored by the US National Science Foundation, we define a vision for materials laboratories of the future in alloys, amorphous materials, and composite materials; chart a roadmap for realizing this vision; identify technical bottlenecks and barriers to access; and propose pathways to equitable and democratic access to integrated toolsets in a manner that addresses urgent societal needs, accelerates technological innovation, and enhances manufacturing competitiveness. Spanning three important materials classes, this article summarizes the areas of alignment and unifying themes, distinctive needs of different materials research communities, key science drivers that cannot be accomplished within the capabilities of current materials laboratories, and open questions that need further community input. Here, we provide a broader context for the workshop, synopsize the salient findings, outline a shared vision for democratizing access and accelerating materials discovery, highlight some case studies across the three different materials classes, and identify significant issues that need further discussion. Graphical abstract

Transition Metal‐Based High‐Entropy Materials for Catalysis

ABSTRACTHigh‐entropy materials (HEMs) have emerged as a pioneering paradigm in recent years, drawing substantial interest due to their unique combination of diverse elemental constituents and homogeneous solid‐solution structure. This novel material class not only opens up extensive potential for materials discovery through a broad spectrum of elemental combinations but also facilitates fine‐tuning of properties thanks to its distinctive microstructural characteristics. HEMs have garnered considerable attention across various applications, particularly in catalysis. The virtually infinite variations in elemental and compositional combinations within these multi‐elemental systems enable meticulous optimization of the catalytic performance. Additionally, the high‐entropy solid‐solution structure potentially enhances structural, thermal, and chemical stability, which is vital for ensuring functionality under harsh conditions. Herein, we thoroughly explore the exceptional attributes of HEMs, designing strategies for transition metal‐based catalysis, and three major catalytic fields of HEMs: electrocatalysis, photocatalysis, and thermocatalysis. This discussion aspires to provide valuable perspectives into the advancements and innovations in catalyst design and development.

SPACIER: on-demand polymer design with fully automated all-atom classical molecular dynamics integrated into machine learning pipelines

Machine learning has rapidly advanced the design and discovery of new materials with targeted applications in various systems. First-principles calculations and other computer experiments have been integrated into material design pipelines to address the lack of experimental data and the limitations of interpolative machine learning predictors. However, the enormous computational costs and technical challenges of automating computer experiments for polymeric materials have limited the availability of open-source automated polymer design systems that integrate molecular simulations and machine learning. We developed SPACIER, an open-source software program that incorporates RadonPy, a Python library for fully automated polymer physical property calculations based on all-atom classical molecular dynamics, into a Bayesian optimization-based polymer design system to overcome these challenges. As a proof-of-concept study, we synthesized optical polymers that surpass the Pareto boundary formed by the tradeoff between the refractive index and the Abbe number.

Discovery of new topological insulators and semimetals using deep generative models

Topological materials possess unique electronic properties and hold immense attraction to both fundamental physics research and practical applications. Over the past decades, the discovery of new topological materials has relied on the symmetry-based analysis of the quantum wave function. In this study, we propose an efficient inverse design method CTMT (CTMT: CDVAE, Topogivity, interatomic potentials (IAPs) as realized in M3GNet, and TQC) utilizing deep generative machine learning models to discover novel topological insulators and semimetals in a much-fast and low-cost manner. This method covers the entire process of new crystal structure generation, heuristic rule screening, fast stability estimation, and topology type diagnosis, resulting in 4 topological insulators and 16 topological semimetals. Especially, the newly discovered topological materials include several chiral Kramers-Weyl fermion semimetals and chiral materials with low symmetry, whose topology is previously considered challenging to discern. These findings demonstrate the capability of CTMT in discovering topological materials and its great potential for data-driven inverse design of advanced functional materials.

Hybrid Unsupervised/Supervised Machine Learning for Identifying Molecular Structural Fingerprints From Ensemble Property.

The ensemble properties of a system are obtained by averaging over the properties calculated for the various configurations it can have at a finite temperature and thus cannot be captured by a single molecular structure. Such ensemble properties are often important in material discovery. In designing new materials, the goal is to predict those ensemble structures that display a tailored property. However, mapping this average property to multiple structures introduces ambiguities and unreliable convergence in supervised machine learning. This presents a major obstacle in designing new materials. Here, we introduce a hybrid unsupervised/supervised learning method and demonstrate how to predict the structural parameters defining the conformers of a heterogeneous system, melanin, from its ensemble-averaged spectra. This also shows a new way to identify different structural fingerprints responsible for an ensemble-averaged superposition spectrum.

The missing Tb2O2NCN compound: synthesis, characterization, and luminescent properties

Advances in the synthetic methods often lead to the discovery of new materials, and nitridocarbonates based on the NCN2− anion are no exception. Recent momentum in preparation processes and improved understanding of the chemistry of these compounds has resulted in many “missing compounds” now being accessible. Accordingly, we here report the synthesis of the terbium oxide carbodiimide Tb2O2NCN, one of the missing members in the family of rare‐earth compounds with the general formula Ln2O2NCN. It was prepared via a solid‐state metathesis reaction at 600 °C between TbOCl and Li2NCN, and its structure was determined from powder X‐ray diffraction data to crystallize isotypically to the Ln2O2NCN family with Ln = Ce–Yb (except for promethium) in the trigonal P[[EQUATION]] space group. The crystal structure is best described as alternating layers of NCN2− anions and Tb2O22+ layers along the c‐axis. Photoluminescence measurements reveal a green emission at room temperature.

Integrating Automated Electrochemistry and High‐Throughput Characterization with Machine Learning to Explore Si─Ge─Sn Thin‐Film Lithium Battery Anodes

AbstractHigh‐performance batteries need accelerated discovery and optimization of new anode materials. Herein, we explore the Si─Ge─Sn ternary alloy system as a candidate fast‐charging anode materials system by utilizing a scanning droplet cell (SDC) as an autonomous electrochemical characterization tool with the goal of subsequent upscaling. As the SDC is performing experiments sequentially, an exploration of the entire ternary space is unfeasible due to time constraints. Thus, closed‐loop optimization, guided by real‐time data analysis and sequential learning algorithms, is utilized to direct experiments. The lead material identified is scaled up to a coin cell to validate the findings from the autonomous millimeter‐scale thin‐film electrochemical experimentation. Explainable machine learning (ML) models incorporating data from high‐throughput Raman spectroscopy and X‐ray diffraction (XRD) are used to elucidate the effect of short and long‐range ordering on material performance.

Targeted Chemical Looping Materials Discovery by an Inverse Design

Chemical looping with oxygen uncoupling (CLOU) materials is actively sought for combustion of carbonaceous materials to achieve complete conversion and capture of carbon dioxide. These materials may play a vital role in reducing atmospheric carbon via negative carbon output. However, there is no one‐size‐fits‐all approach as different operating conditions and feedstocks may require different CLOU materials. As a result, the exploration and discovery of high‐performance CLOU materials can be a slow process. To address this challenge, a high‐throughput inverse machine learning workflow that identifies optimum materials from perovskite oxides for a given set of targets is developed—temperature and Gibbs free energy of oxygen formation. The model is trained on high‐throughput density functional theory calculations of CLOU materials and inverts the materials design process using a genetic algorithm to produce realistic substituted SrFeO3‐δ compositions as output. Using the inverse model, it is able to identify several interesting new families of CLOU materials: Sr1‐xAxFe1‐yByO3‐δ (e.g., A = Ca or K; B = Mg, Bi, Mn, Ni, Co, Cu, or Zn). These materials have shown promising properties, and some of them even outperform the benchmark material in terms of oxygen release kinetics under relevant CLOU operating conditions.

Universal Phase Identification of Block Copolymers From Physics‐Informed Machine Learning

ABSTRACTBlock copolymers play a vital role in materials science due to their diverse self‐assembly behavior. Traditionally, exploring the block copolymer self‐assembly and associated structure–property relationships involve iterative synthesis, characterization, and theory, which is labor‐intensive both experimentally and computationally. Here, we introduce a versatile, high‐throughput workflow toward materials discovery that integrates controlled polymerization and automated chromatographic separation with a novel physics‐informed machine‐learning algorithm for the rapid analysis of small‐angle X‐ray scattering data. Leveraging the expansive and high‐quality experimental data sets generated by fractionating polymers using automated chromatography, this machine‐learning method effectively reduces data dimensionality by extracting chemical‐independent features from SAXS data. This new approach allows for the rapid and accurate prediction of morphologies without repetitive and time‐consuming manual analysis, achieving out‐of‐sample predictive accuracy of around 95% for both novel and existing materials in the training data set. By focusing on a subset of samples with large predictive uncertainty, only a small fraction of the samples needs to be inspected to further improve accuracy. Collectively, the synergistic combination of controlled synthesis, automated chromatography, and data‐driven analysis creates a powerful workflow that markedly expedites the discovery of structure–property relationships in advanced soft materials.

Machine learning Hubbard parameters with equivariant neural networks

Density-functional theory with extended Hubbard functionals (DFT + U + V) provides a robust framework to accurately describe complex materials containing transition-metal or rare-earth elements. It does so by mitigating self-interaction errors inherent to semi-local functionals which are particularly pronounced in systems with partially-filled d and f electronic states. However, achieving accuracy in this approach hinges upon the accurate determination of the on-site U and inter-site V Hubbard parameters. In practice, these are obtained either by semi-empirical tuning, requiring prior knowledge, or, more correctly, by using predictive but expensive first-principles calculations. Here, we present a machine learning model based on equivariant neural networks which uses atomic occupation matrices as descriptors, directly capturing the electronic structure, local chemical environment, and oxidation states of the system at hand. We target here the prediction of Hubbard parameters computed self-consistently with iterative linear-response calculations, as implemented in density-functional perturbation theory (DFPT), and structural relaxations. Remarkably, when trained on data from 12 materials spanning various crystal structures and compositions, our model achieves mean absolute relative errors of 3% and 5% for Hubbard U and V parameters, respectively. By circumventing computationally expensive DFT or DFPT self-consistent protocols, our model significantly expedites the prediction of Hubbard parameters with negligible computational overhead, while approaching the accuracy of DFPT. Moreover, owing to its robust transferability, the model facilitates accelerated materials discovery and design via high-throughput calculations, with relevance for various technological applications.