Modeling the dielectric constants of crystals using machine learning

Excitons in metal-halide perovskites from first-principles many-body perturbation theory

Metal-halide perovskites are a structurally, chemically, and electronically diverse class of semiconductors with applications ranging from photovoltaics to radiation detectors and sensors. Understanding neutral electron-hole excitations (excitons) is key for predicting and improving the efficiency of energy-conversion processes in these materials. First-principles calculations have played an important role in this context, allowing for a detailed insight into the formation of excitons in many different types of perovskites. Such calculations have demonstrated that excitons in some perovskites significantly deviate from canonical models due to the chemical and structural heterogeneity of these materials. In this Perspective, I provide an overview of calculations of excitons in metal-halide perovskites using Green's function-based many-body perturbation theory in the GW + Bethe-Salpeter equation approach, the prevalent method for calculating excitons in extended solids. This approach readily considers anisotropic electronic structures and dielectric screening present in many perovskites and important effects, such as spin-orbit coupling. I will show that despite this progress, the complex and diverse electronic structure of these materials and its intricate coupling to pronounced and anharmonic structural dynamics pose challenges that are currently not fully addressed within the GW + Bethe-Salpeter equation approach. I hope that this Perspective serves as an inspiration for further exploring the rich landscape of excitons in metal-halide perovskites and other complex semiconductors and for method development addressing unresolved challenges in the field.

Machine-learning screening of luminogens with aggregation-induced emission characteristics for fluorescence imaging

Due to the excellent biocompatible physicochemical performance, luminogens with aggregation-induced emission (AIEgens) characteristics have played a significant role in biomedical fluorescence imaging recently. However, screening AIEgens for special applications takes a lot of time and efforts by using conventional chemical synthesis route. Fortunately, artificial intelligence techniques that could predict the properties of AIEgen molecules would be helpful and valuable for novel AIEgens design and synthesis. In this work, we applied machine learning (ML) techniques to screen AIEgens with expected excitation and emission wavelength for biomedical deep fluorescence imaging. First, a database of various AIEgens collected from the literature was established. Then, by extracting key features using molecular descriptors and training various state-of-the-art ML models, a multi-modal molecular descriptors strategy has been proposed to extract the structure-property relationships of AIEgens and predict molecular absorption and emission wavelength peaks. Compared to the first principles calculations, the proposed strategy provided greater accuracy at a lower computational cost. Finally, three newly predicted AIEgens with desired absorption and emission wavelength peaks were synthesized successfully and applied for cellular fluorescence imaging and deep penetration imaging. All the results were consistent successfully with our expectations, which demonstrated the above ML has a great potential for screening AIEgens with suitable wavelengths, which could boost the design and development of novel organic fluorescent materials.

Bridging Fidelities to Predict Nanoindentation Tip Radii Using Interpretable Deep Learning Models

As the need for miniaturized structural and functional materials has increased, the need for precise materials characterizaton has also expanded. Nanoindentation is a popular method that can be used to measure material mechanical behavior which enables high-throughput experiments and, in some cases, can also provide images of the indented area through scanning. Both indenting and scanning can cause tip wear that can influence the measurements. Therefore, precise characterization of tip radii is needed to improve data evaluation. A data fusion method is introduced which uses finite element simulations and experimental data to estimate the tip radius in situ in a meaningful way using an interpretable multi-fidelity deep learning approach. By interpreting the machine learning models, it is shown that the approaches are able to accurately capture physical indentation phenomena.

Prediction of dielectric constants of ABO3-type perovskites using machine learning and first-principles calculations

In this study, the machine-learning method, combined with density functional perturbation theory (DFPT) calculations, was implemented to predict and validate the dielectric constants of ABO₃-type perovskites. For the construction of the training database, the dielectric constants of 7113 inorganic materials were extracted from the Materials Project. The chemical, structural, and physical descriptors were generated and trained using the gradient-boosting-based regressor after feature engineering. The prediction accuracies were 0.83 and 0.67 (R²) and 0.12 and 0.26 (root mean square error) for the electronic and ionic contributions to the dielectric constant, respectively. The constructed surrogate model was then employed to predict the dielectric constants of the ABO₃-type perovskites (216 structures), whose thermodynamic stabilities were satisfactory. The predicted values were validated using DFPT calculations. The constructed database was further used to develop a surrogate model for the prediction of dielectric constants. The final R² prediction accuracies reached 0.79 (electronic) and 0.67 (ionic).

Entropy-based active learning of graph neural network surrogate models for materials properties

Graph neural networks trained on experimental or calculated data are becoming an increasingly important tool in computational materials science. Networks once trained are able to make highly accurate predictions at a fraction of the cost of experiments or first-principles calculations of comparable accuracy. However, these networks typically rely on large databases of labeled experiments to train the model. In scenarios where data are scarce or expensive to obtain, this can be prohibitive. By building a neural network that provides confidence on the predicted properties, we are able to develop an active learning scheme that can reduce the amount of labeled data required by identifying the areas of chemical space where the model is most uncertain. We present a scheme for coupling a graph neural network with a Gaussian process to featurize solid-state materials and predict properties including a measure of confidence in the prediction. We then demonstrate that this scheme can be used in an active learning context to speed up the training of the model by selecting the optimal next experiment for obtaining a data label. Our active learning scheme can double the rate at which the performance of the model on a test dataset improves with additional data compared to choosing the next sample at random. This type of uncertainty quantification and active learning has the potential to open up new areas of materials science, where data are scarce and expensive to obtain, to the transformative power of graph neural networks.

Exploring Dielectric Constant and Dissipation Factor of LTCC Using Machine Learning

Low-temperature co-fired ceramics (LTCCs) have been attracting attention due to rapid advances in wireless telecommunications. Low-dielectric-constant (Dk) and low-dissipation-factor (Df) LTCCs enable a low propagation delay and high signal quality. However, the wide ranges of glass, ceramic filler compositions, and processing features in fabricating LTCC make property modulating difficult via experimental trial-and-error approaches. In this study, we explored Dk and Df values of LTCCs using a machine learning method with a Gaussian kernel ridge regression model. A principal component analysis and k-means methods were initially performed to visually analyze data clustering and to reduce the dimension complexity. Model assessments, by using a five-fold cross-validation, residual analysis, and randomized test, suggest that the proposed Dk and Df models had some predictive ability, that the model selection was appropriate, and that the fittings were not just numerical due to a rather small data set. A cross-plot analysis and property contour plot were performed for the purpose of exploring potential LTCCs for real applications with Dk and Df values less than 10 and 2 × 10⁻³, respectively, at an operating frequency of 1 GHz. The proposed machine learning models can potentially be utilized to accelerate the design of technology-related LTCC systems.

Cost-Effective Hybrid Density Functional Theory Calculation of Three-Dimensional Band Structure and Search of Band Edge Positions

Accurate calculation of the electronic band structure is essential to material screening and design. Hybrid density functional has been recently widely used to describe the electronic structure of semiconductors; however, it is difficult to locate the band edge positions of indirect band gap materials due to heavy computational cost especially when the band edges are not located at special k-points. We suggest how to investigate three-dimensional band structure efficiently with hybrid density functionals and to find the band edge positions. The band edge position of diamond Si, SbSI, and MoS₂ are investigated using the proposed method.

Machine learning meets chemical physics

Over recent years, the use of statistical learning techniques applied to chemical problems has gained substantial momentum. This is particularly apparent in the realm of physical chemistry, where the balance between empiricism and physics-based theory has traditionally been rather in favor of the latter. In this guest Editorial for the special topic issue on "Machine Learning Meets Chemical Physics," a brief rationale is provided, followed by an overview of the topics covered. We conclude by making some general remarks.

Perovskite-inspired materials for photovoltaics and beyond-from design to devices

Lead-halide perovskites have demonstrated astonishing increases in power conversion efficiency in photovoltaics over the last decade. The most efficient perovskite devices now outperform industry-standard multi-crystalline silicon solar cells, despite the fact that perovskites are typically grown at low temperature using simple solution-based methods. However, the toxicity of lead and its ready solubility in water are concerns for widespread implementation. These challenges, alongside the many successes of the perovskites, have motivated significant efforts across multiple disciplines to find lead-free and stable alternatives which could mimic the ability of the perovskites to achieve high performance with low temperature, facile fabrication methods. This Review discusses the computational and experimental approaches that have been taken to discover lead-free perovskite-inspired materials, and the recent successes and challenges in synthesizing these compounds. The atomistic origins of the extraordinary performance exhibited by lead-halide perovskites in photovoltaic devices is discussed, alongside the key challenges in engineering such high-performance in alternative, next-generation materials. Beyond photovoltaics, this Review discusses the impact perovskite-inspired materials have had in spurring efforts to apply new materials in other optoelectronic applications, namely light-emitting diodes, photocatalysts, radiation detectors, thin film transistors and memristors. Finally, the prospects and key challenges faced by the field in advancing the development of perovskite-inspired materials towards realization in commercial devices is discussed.

Electron Density and Its Relation with Electronic and Optical Properties in 2D Mo/W Dichalcogenides

Two-dimensional MX₂ (M = Mo, W; X = S, Se, Te) homo- and heterostructures have attracted extensive attention in electronics and optoelectronics due to their unique structures and properties. In this work, the layer-dependent electronic and optical properties have been studied by varying layer thickness and stacking order. Based on the quantum theory of atoms in molecules, topological analyses on interatomic interactions of layered MX₂ and WX₂/MoX₂, including bond degree (BD), bond length (BL), and bond angle (BA), have been detailed to probe structure-property relationships. Results show that M-X and X-X bonds are strengthened and weakened in layered MX₂ compared to the counterparts in bulks. X-X and M-Se/Te are weakened at compressive strain while strengthened at tensile strain and are more responsive to the former than the latter. Discordant BD variation of individual parts of WX₂/MoX₂ accounts for exclusively distributed electrons and holes, yielding type-II band offsets. X-X BL correlates positively to binding energy (E_b), while X-X BA correlates negatively to lattice mismatch (lm). The resulting interlayer distance limitation evidences constraint-free lattice of vdW structure. Finally, the connection between microscopic interatomic interaction and macroscopic electromagnetic behavior has been quantified firstly by a cubic equation relating to weighted BD summation and static dielectric constant.

Structure-Property Relationships of 2D Ga/In Chalcogenides

Two-dimensional MX (M = Ga, In; X = S, Se, Te) homo- and heterostructures are of interest in electronics and optoelectronics. Structural, electronic and optical properties of bulk and layered MX and GaX/InX heterostructures have been investigated comprehensively using density functional theory (DFT) calculations. Based on the quantum theory of atoms in molecules, topological analyses of bond degree (BD), bond length (BL) and bond angle (BA) have been detailed for interpreting interatomic interactions, hence the structure-property relationship. The X-X BD correlates linearly with the ratio of local potential and kinetic energy, and decreases as X goes from S to Te. For van der Waals (vdW) homo- and heterostructures of GaX and InX, a cubic relationship between microscopic interatomic interaction and macroscopic electromagnetic behavior has been established firstly relating to weighted absolute BD summation and static dielectric constant. A decisive role of vdW interaction in layer-dependent properties has been identified. The GaX/InX heterostructures have bandgaps in the range 0.23-1.49 eV, absorption coefficients over 10-5 cm-1 and maximum conversion efficiency over 27%. Under strain, discordant BD evolutions are responsible for the exclusively distributed electrons and holes in sublayers of GaX/InX. Meanwhile, the interlayer BA adjustment with lattice mismatch explains the constraint-free lattice of the vdW heterostructure.