The growing application of data-driven analytics in materials science has led to the rise of materials informatics. Within the arena of data analytics, deep learning has emerged as a game-changing technique in the last few years, enabling numerous real-world applications, such as self-driving cars. In this paper, the authors present an overview of deep learning, its advantages, challenges, and recent applications on different types of materials data. The increasingly availability of materials databases and big data in general, along with groundbreaking advances in deep learning offers a lot of promise to accelerate the discovery, design, and deployment of next-generation materials.

The authors showcase the potential of symbolic regression as an analytic method for use in materials research. First, the authors briefly describe the current state-of-the-art method, genetic programming-based symbolic regression (GPSR), and recent advances in symbolic regression techniques. Next, the authors discuss industrial applications of symbolic regression and its potential applications in materials science. The authors then present two GPSR use-cases: formulating a transformation kinetics law and showing the learning scheme discovers the well-known Johnson–Mehl–Avrami–Kolmogorov form, and learning the Landau free energy functional form for the displacive tilt transition in perovskite LaNiO3. Finally, the authors propose that symbolic regression techniques should be considered by materials scientists as an alternative to other machine learning-based regression models for learning from data.

Machine learning (ML) has revolutionized disciplines within materials science that have been able to generate sufficiently large datasets to utilize algorithms based on statistical inference, but for many important classes of materials the datasets remain small. However, a rapidly growing number of approaches to embedding domain knowledge of materials systems are reducing data requirements and allowing broader applications of ML. Furthermore, these hybrid approaches improve the interpretability of the predictions, allowing for greater physical insights into the factors that determine material properties. This review introduces a number of these strategies, providing examples of how they were implemented in ML algorithms and discussing the materials systems to which they were applied.

The use of statistical/machine learning (ML) approaches to materials science is experiencing explosive growth. Here, we review recent work focusing on the generation and application of libraries from both experiment and theoretical tools. The library data enables classical correlative ML and also opens the pathway for exploration of underlying causative physical behaviors. We highlight key advances facilitated by this approach and illustrate how modeling, macroscopic experiments, and imaging can be combined to accelerate the understanding and development of new materials systems. These developments point toward a data-driven future wherein knowledge can be aggregated and synthesized, accelerating the advancement of materials science.

Computational acceleration of performance metric-based materials discovery via high-throughput screening and machine learning methods is becoming widespread. Nevertheless, development and optimization of the opto-electronic properties that depend on dilute concentrations of point defects in new materials have not significantly benefited from these advances. Here, the authors present an informatics and simulation suite to computationally accelerate these processes. This will enable faster and more fundamental materials research, and reduce the cost and time associated with the materials development cycle. Analogous to the new avenues enabled by current first-principles-based property databases, this type of framework will open entire new research frontiers as it proliferates.

Applying artificial intelligence to materials research requires abundant curated experimental data and the ability for algorithms to request new experiments. ESCALATE (Experiment Specification, Capture and Laboratory Automation Technology)—an ontological framework and open-source software package—solves this problem by providing an abstraction layer for human- and machine-readable experiment specification, comprehensive and extensible (meta-) data capture, and structured data reporting. ESCALATE simplifies the initial data collection process, and its reporting and experiment generation mechanisms simplify machine learning integration. An initial ESCALATE implementation for metal halide perovskite crystallization was used to perform 55 rounds of algorithmically-controlled experiment plans, capturing 4336 individual experiments.

Machine-learning (ML) approaches have proven to be of great utility in modern materials innovation pipelines. Generally, ML models are trained on predetermined past data and then used to make predictions for new test cases. Active-learning, however, is a paradigm in which ML models can direct the learning process itself through providing dynamic suggestions/queries for the “next-best experiment.” In this work, the authors demonstrate how an active-learning framework can aid in the discovery of polymers possessing high glass transition temperatures (Tg). Starting from an initial small dataset of polymer Tg measurements, the authors use Gaussian process regression in conjunction with an active-learning framework to iteratively add Tg measurements of candidate polymers to the training dataset. The active-learning framework employs one of three decision making strategies (exploitation, exploration, or balanced exploitation/exploration) for selection of the “next-best experiment.” The active-learning workflow terminates once 10 polymers possessing a Tg greater than a certain threshold temperature are selected. The authors statistically benchmark the performance of the aforementioned three strategies (against a random selection approach) with respect to the discovery of high-Tg polymers for this particular demonstrative materials design challenge.

The authors developed a Behler–Parrinello-type neural network (NN) to improve the density-functional tight-binding (DFTB) energy and force prediction. The Δ-machine learning approach was adopted and the NN was designed to predict the energy differences between the density functional theory (DFT) quantum chemical potential and DFTB for a given molecular structure. Most notably, the DFTB-NN method is capable of improving the energetics of intramolecular hydrogen bonds and torsional potentials without modifying the framework of DFTB itself. This improvement enables considerably larger simulations of complex chemical systems that currently could not easily been accomplished using DFT or higher level ab initio quantum chemistry methods alone.

Our ability to describe crystal structure features is of crucial importance when attempting to understand structure–property relationships in the solid state. In this paper, the authors introduce robocrystallographer, an open-source toolkit for analyzing crystal structures. This package combines new and existing open-source analysis tools to provide structural information, including the local coordination and polyhedral type, polyhedral connectivity, octahedral tilt angles, component-dimensionality, and molecule-within-crystal and fuzzy prototype identification. Using this information, robocrystallographer can generate text-based descriptions of crystal structures that resemble descriptions written by human crystallographers. The authors use robocrystallographer to investigate the dimensionalities of all compounds in the Materials Project database and highlight its potential in machine learning studies.

The authors develop a computational approach that integrates machine learning (ML) and density functional theory (DFT) with experimental data to predict formable and thermodynamically stable iodine-containing apatites. This is an important problem because radioactive iodine is toxic and capturing it in solid waste forms have implications in remediation treatments. The authors train ML models using 336 compositions and screen 54 iodine-containing compounds in apatite stoichiometry. ML models predict 18 as formable and 24 as nonformable in the apatite structure; 12 compounds were identified to be uncertain. DFT convex hull predicted two to be thermodynamically stable, one as metastable, and nine as unstable.

Recent studies illustrate how machine learning (ML) can be used to bypass a core challenge of molecular modeling: the trade-off between accuracy and computational cost. Here, we assess multiple ML approaches for predicting the atomization energy of organic molecules. Our resulting models learn the difference between low-fidelity, B3LYP, and high-accuracy, G4MP2, atomization energies and predict the G4MP2 atomization energy to 0.005 eV (mean absolute error) for molecules with less than nine heavy atoms (training set of 117,232 entries, test set 13,026) and 0.012 eV for a small set of 66 molecules with between 10 and 14 heavy atoms. Our two best models, which have different accuracy/speed trade-offs, enable the efficient prediction of G4MP2-level energies for large molecules and are available through a simple web interface.

Insects have more than a million described species and represent more than half of all known living organisms. However, little is known about the operation and functions of the insect body, particularly their remarkable leg joints. This study is focused on partly filling this knowledge gap by using nanoindentation instruments to characterize the mechanical properties of leg joints from three different insects: a beetle, a mantis, and a dragonfly nymph. For all insect species, the tibia joint had the largest effective elastic moduli, followed by the femur joint, while the exocuticle had the smallest values.

The authors demonstrate that gold-binding peptides displayed on the outer membrane of Escherichia coli enhance bioelectrochemical charge transfer by binding gold nanoparticles. Microbial fuel cells were run with different gold-binding peptides displayed and with different nanoparticle sizes, and the results were correlated with transmission electron microscopy (TEM) imaging of nanoparticle binding. When a gold-binding peptide is displayed and 5 nm gold nanoparticles are present, up to 4× power generation over E. coli not displaying a gold-binding peptide is observed. While an enhanced current is observed using the previously published M6G9, the largest enhancement is observed when a new longer peptide named M9G18 is used.

We report on the structural evolution of diamond-like carbon (DLC) films by the nanosecond pulsed laser annealing process. DLC film is coated on cemented carbide (WC-Co) by cathodic arc ion plating, which is then annealed by ArF laser (193 nm, 20 ns) at different laser fluences (0.9–1.7 J/cm2). Upon laser annealing, Raman spectra divulge higher sp3 fractions accompanied by a blue shift in the G-peak position, which indicates the changes of sp2 sites from rings to chains. At higher fluence (>1.2 J/cm2), the film converts into reduced graphene oxide confirmed by its Raman-active vibrational modes: D, G, and 2D.

A method for rapid quantitative imaging of dopant distribution using secondary ion mass spectrometry (SIMS) is described. The method is based on SIMS imaging of the cross-section of a reference sample with a known concentration profile. It is demonstrated for the case of boron quantification in silicon in a SIMS imaging mode. A nonlinear relationship between the secondary ion intensity and the concentration is observed. A detection limit of 3 (±2) × 1017 at./cm3 (~6 ppm) is determined with 39 nm pixel-size for the used experimental conditions. As an application example, a boron concentration profile in a passivating contact deposited on a textured Si surface is analyzed.

The processes of singlet fission and triplet fusion could allow state-of-the-art photovoltaic devices to surpass the Shockley–Queisser limit by optimizing the utilized solar spectrum by reducing thermal relaxation and inaccessible sub-bandgap photons, respectively. Triplet fusion demands precise control of the spin-triplet state population, and requires a sensitizer to efficiently populate the triplet state of an acceptor molecule. In this perspective, we highlight the established field of sensitized upconversion and further examine alternative triplet sensitization routes, including the possibility of bulk solid-state semiconductors as triplet sensitizers, which provide a new avenue for charge transfer-based triplet sensitization rather than excitonic triplet energy transfer.

In order to reveal the quantitative relationship between fatigue crack deflection path and cross-sectional grain boundary (GB) arrangement of metallic nanolayered composites (NLCs), a stochastic model was established based on the interface-dominant fatigue damage for the ultrafine-scale NLCs. The model indicates that the crack deflection length decreases with decreasing GB arrangement deviation and grain size of constituent layers. The observation and quantitative analysis of fatigue cracking behavior of the Cu/W multilayers with a layer thickness of 5 and 20 nm was conducted to verify the model.

Additive manufacturing used with custom electromyographic sensors has been demonstrated for neuroprosthetic limb manufacturing and is now translating to the clinical environment. These manufacturing methods have dramatically reduced device weight while increasing the capability for multi-finger dexterity. Using wearable electromyography sensors standalone from the prosthetic limb, a new virtual training method has been designed and tested to improve human–machine interaction. This type of training leverages real-time visual feedback to user inputs, supporting improved timing and magnitudes of muscle contractions. The combination of these technologies may provide a stronger affinity between the pediatric patient group and the device.

Signal detection limit (SDL), limit of detection (LOD), and limit of quantitation of a portable Raman spectrometer were measured for smokeless gunpowder stabilizers, diphenylamine (DPA) and ethyl centralite (EC), in acetone, acetonitrile, ethanol, and methanol. Acetone yielded the lowest LOD for three of four DPA peaks, and acetonitrile yielded the lowest LOD for two of three EC peaks and the remaining DPA peak. When gold nanoparticles were added to the DPA solutions in acetone and acetonitrile, statistically significant changes were observed (DPA peak position, full width at half maximum, and/or total area) and SDL was improved for the majority of all peaks in both solvents.

The authors report on strong exciton–photon coupling in all-metal microcavities containing functionalized anthradithiophene (ADT) in host poly(methyl methacrylate) matrices for a wide range of ADT concentrations. Angle-resolved reflectance of polycrystalline films revealed Rabi splittings up to 340 meV. Angle-resolved photoluminescence in films with low ADT concentrations (dominated by “isolated” ADT molecules) showed Rabi splittings which scaled with the square root of oscillator strength. When “aggregated” and “isolated” ADT molecules coexisted in film, cavities preferentially coupled to “isolated” molecules due to an anisotropic distribution of aggregates. As a solution-processable high-performance organic semiconductor, ADT shows promise as an (opto)electronic polaritonic material.