Machine learning (ML) is rapidly revolutionizing many fields and is starting to change landscapes for physics and chemistry. With its ability to solve complex tasks autonomously, ML is being ...exploited as a radically new way to help find material correlations, understand materials chemistry, and accelerate the discovery of materials. Here, an in‐depth review of the application of ML to energy materials, including rechargeable alkali‐ion batteries, photovoltaics, catalysts, thermoelectrics, piezoelectrics, and superconductors, is presented. A conceptual framework is first provided for ML in materials science, with a broad overview of different ML techniques as well as best practices. This is followed by a critical discussion of how ML is applied in energy materials. This review is concluded with the perspectives on major challenges and opportunities in this exciting field.
This review discusses the recent surging interest in applying machine learning models to energy materials research. The analysis covers the whole modeling pipeline, including target identification, data collection, material featurization, model construction, and applications. The applications have led to success stories in alkaline‐ion batteries, photovoltaics, catalysis, thermoelectrics, piezoelectrics, and superconductors. Potential problems and future directions are suggested.
Graph networks are a new machine learning (ML) paradigm that supports both relational reasoning and combinatorial generalization. Here, we develop universal MatErials Graph Network (MEGNet) models ...for accurate property prediction in both molecules and crystals. We demonstrate that the MEGNet models outperform prior ML models such as the SchNet in 11 out of 13 properties of the QM9 molecule data set. Similarly, we show that MEGNet models trained on ∼60 000 crystals in the Materials Project substantially outperform prior ML models in the prediction of the formation energies, band gaps, and elastic moduli of crystals, achieving better than density functional theory accuracy over a much larger data set. We present two new strategies to address data limitations common in materials science and chemistry. First, we demonstrate a physically intuitive approach to unify four separate molecular MEGNet models for the internal energy at 0 K and room temperature, enthalpy, and Gibbs free energy into a single free energy MEGNet model by incorporating the temperature, pressure, and entropy as global state inputs. Second, we show that the learned element embeddings in MEGNet models encode periodic chemical trends and can be transfer-learned from a property model trained on a larger data set (formation energies) to improve property models with smaller amounts of data (band gaps and elastic moduli).
NASICON is one of the most promising sodium solid electrolytes that can enable the assembly of cheaper and safer sodium all-solid-state batteries. In this study, we perform a combined experimental ...and computational investigation into the effects of aliovalent doping in NASICON on both bulk and grain boundary (secondary phase) ionic conductivity. Our results show that the dopants with low solid solubility limits in NASICON lead to the formation of a conducting (less insulating) secondary phase, thereby improving the grain boundary conductivity measured by electrochemical impedance spectroscopy (including grain-boundary, secondary-phase, and other microstructural contributions) that is otherwise hindered by the poorly-conducting secondary phases in undoped NASICON. This is accompanied by a change in the Si/P ratio in the primary NASICON bulk phase, thereby transforming monoclinic NASICON to rhombohedral NASICON. Consequently, we have synthesized NASICON chemistries with significantly improved and optimized total ionic conductivity of 2.7 mS/cm. More importantly, this study has achieved an understanding of the underlying mechanisms of improved conductivities via doping (differing from the common wisdom) and further suggests a new general direction to improve the ionic conductivity of solid electrolytes via simultaneously optimizing the primary bulk phase and the microstructure (including grain boundary segregation and secondary phases).
•The aliovalent doping effects in NASICON investigated via experiments and modeling.•Both bulk and grain boundary ionic conductivities considered.•Optimized ionic conductivity of 2.7 mS/cm achieved via mechanistic understanding.•Simultaneous optimization of bulk and grain boundary phases proven essential.•A new direction to improve the ionic conductivity of solid electrolytes suggested.
Predicting the stability of crystals is one of the central problems in materials science. Today, density functional theory (DFT) calculations remain comparatively expensive and scale poorly with ...system size. Here we show that deep neural networks utilizing just two descriptors-the Pauling electronegativity and ionic radii-can predict the DFT formation energies of C
A
D
O
garnets and ABO
perovskites with low mean absolute errors (MAEs) of 7-10 meV atom
and 20-34 meV atom
, respectively, well within the limits of DFT accuracy. Further extension to mixed garnets and perovskites with little loss in accuracy can be achieved using a binary encoding scheme, addressing a critical gap in the extension of machine-learning models from fixed stoichiometry crystals to infinite universe of mixed-species crystals. Finally, we demonstrate the potential of these models to rapidly transverse vast chemical spaces to accurately identify stable compositions, accelerating the discovery of novel materials with potentially superior properties.
Abstract
Predicting properties from a material’s composition or structure is of great interest for materials design. Deep learning has recently garnered considerable interest in materials predictive ...tasks with low model errors when dealing with large materials data. However, deep learning models suffer in the small data regime that is common in materials science. Here we develop the AtomSets framework, which utilizes universal compositional and structural descriptors extracted from pre-trained graph network deep learning models with standard multi-layer perceptrons to achieve consistently high model accuracy for both small compositional data (<400) and large structural data (>130,000). The AtomSets models show lower errors than the graph network models at small data limits and other non-deep-learning models at large data limits. They also transfer better in a simulated materials discovery process where the targeted materials have property values out of the training data limits. The models require minimal domain knowledge inputs and are free from feature engineering. The presented AtomSets model framework can potentially accelerate machine learning-assisted materials design and discovery with less data restriction.
In this work, we performed a first-principles investigation of the phase stability, dopant formation energy and Na+ conductivity of pristine and doped cubic Na3PS4 (c-Na3PS4). We show that pristine ...c-Na3PS4 is an extremely poor Na ionic conductor, and the introduction of Na+ excess is the key to achieving reasonable Na+ conductivities. We studied the effect of aliovalent doping of M4+ for P5+ in c-Na3PS4, yielding Na3+x M x P1–x S4 (M = Si, Ge, and Sn with x = 0.0625; M = Si with x = 0.125). The formation energies in all the doped structures with dopant concentration of x = 0.0625 are found to be relatively low. Using ab initio molecular dynamics simulations, we predict that 6.25% Si-doped c-Na3PS4 has a Na+ conductivity of 1.66 mS/cm, in excellent agreement with previous experimental results. Remarkably, we find that Sn4+ doping at the same concentration yields a much higher predicted Na+ conductivity of 10.7 mS/cm, though with a higher dopant formation energy. A higher Si4+ doping concentration of x = 0.125 also yields a significant increase in Na+ conductivity with an even higher dopant formation energy. Finally, topological and van Hove correlation function analyses suggest that the channel volume and correlation in Na+ motions may play important roles in enhancing Na+ conductivity in this structure.
Machine learning of the quantitative relationship between local environment descriptors and the potential energy surface of a system of atoms has emerged as a new frontier in the development of ...interatomic potentials (IAPs). Here, we present a comprehensive evaluation of machine learning IAPs (ML-IAPs) based on four local environment descriptorsatom-centered symmetry functions (ACSF), smooth overlap of atomic positions (SOAP), the spectral neighbor analysis potential (SNAP) bispectrum components, and moment tensorsusing a diverse data set generated using high-throughput density functional theory (DFT) calculations. The data set comprising bcc (Li, Mo) and fcc (Cu, Ni) metals and diamond group IV semiconductors (Si, Ge) is chosen to span a range of crystal structures and bonding. All descriptors studied show excellent performance in predicting energies and forces far surpassing that of classical IAPs, as well as predicting properties such as elastic constants and phonon dispersion curves. We observe a general trade-off between accuracy and the degrees of freedom of each model and, consequently, computational cost. We will discuss these trade-offs in the context of model selection for molecular dynamics and other applications.
PDF
