High-throughput Screening Of Materials Research Articles

Unlocking Single Particle Anisotropy in Real-Time for Photoelectrochemistry Processes at the Nanoscale.

The key to rationally and rapidly designing high-performance materials is the monitoring and comprehension of dynamic processes within individual particles in real-time, particularly to gain insight into the anisotropy of nanoparticles. The intrinsic property of nanoparticles typically varies from one crystal facet to the next under realistic working conditions. Here, we introduce the operando collision electrochemistry to resolve the single silver nanoprisms (Ag NPs) anisotropy in photoelectrochemistry. We directly identify the effect of anisotropy on the plasmonic-assisted electrochemistry at the single NP/electrolyte interface. The statistical collision frequency shows that heterogeneous diffusion coefficient among crystal facets facilitates Ag NPs to undergo direction-dependent mass transfer toward the gold ultramicroelectrode. Subsequently, the current amplitudes of transient events indicate that anisotropy enables variations in dynamic interfacial electron transfer behaviors during photothermal processes. The results presented here demonstrate that the measurement precision of collision electrochemistry can be extended to the sub-nanoparticle level, highlighting the potential for high-throughput material screening with comprehensive kinetics information at the nanoscale.

First-Principles Simulation and Materials Screening for Spin-Orbit Torque in 2D van der Waals Heterostructures.

Recent advancements in spin-orbit torque (SOT) technology in two-dimensional van der Waals (2D vdW) materials have not only pushed spintronic devices to their atomic limits but have also unveiled unconventional torques and novel spin-switching mechanisms. The vast diversity of SOT observed in numerous 2D vdW materials necessitates a screening strategy to identify optimal materials for torque device performance. However, such a strategy has yet to be established. To address this critical issue, a combination of density functional theory and non-equilibrium Green's function is employed to calculate the SOT in various 2D vdW bilayer heterostructures. This leads to the discovery of three high SOT systems: WTe2/CrSe2, MoTe2/VS2, and NbSe2/CrSe2. Furthermore, a figure of merit that allows for rapid and efficient estimation of SOT is proposed, enabling high-throughput screening of optimal materials and devices for SOT applications in the future.

Multimodal learning of heat capacity based on transformers and crystallography pretraining

Thermal properties of materials are essential to many applications of thermal electronic devices. Density functional theory (DFT) has shown capability in obtaining an accurate calculation. However, the expensive computational cost limits the application of the DFT method for high-throughput screening of materials. Recently, machine learning models, especially graph neural networks (GNNs), have demonstrated high accuracy in many material properties’ prediction, such as bandgap and formation energy, but fail to accurately predict heat capacity(CV) due to the limitation in capturing crystallographic features. In our study, we have implemented the material informatics transformer (MatInFormer) framework, which has been pretrained on lattice reconstruction tasks. This approach has shown proficiency in capturing essential crystallographic features. By concatenating these features with human-designed descriptors, we achieved a mean absolute error of 4.893 and 4.505 J/(mol K) in our predictions. Our findings underscore the efficacy of the MatInFormer framework in leveraging crystallography, augmented with additional information processing capabilities.

High-throughput thermoelectric materials screening by deep convolutional neural network with fused orbital field matrix and composition descriptors

Thermoelectric materials harvest waste heat and convert it into reusable electricity. Thermoelectrics are also widely used in inverse ways such as refrigerators and cooling electronics. However, most popular and known thermoelectric materials to date were proposed and found by intuition, mostly through experiments. Unfortunately, it is extremely time and resource consuming to synthesize and measure the thermoelectric properties through trial-and-error experiments. Here, we develop a convolutional neural network (CNN) classification model that utilizes the fused orbital field matrix and composition descriptors to screen a large pool of materials to discover new thermoelectric candidates with power factor higher than 10 μW/cm K2. The model used our own data generated by high-throughput density functional theory calculations coupled with ab initio scattering and transport package to obtain electronic transport properties without assuming constant relaxation time of electrons, which ensures more reliable electronic transport properties calculations than previous studies. The classification model was also compared to some traditional machine learning algorithms such as gradient boosting and random forest. We deployed the classification model on 3465 cubic dynamically stable structures with non-zero bandgap screened from Open Quantum Materials Database. We identified many high-performance thermoelectric materials with ZT > 1 or close to 1 across a wide temperature range from 300 to 700 K and for both n- and p-type doping with different doping concentrations. Moreover, our feature importance and maximal information coefficient analysis demonstrates two previously unreported material descriptors, namely, mean melting temperature and low average deviation of electronegativity, that are strongly correlated with power factor and thus provide a new route for quickly screening potential thermoelectrics with high success rate. Our deep CNN model with fused orbital field matrix and composition descriptors is very promising for screening high power factor thermoelectrics from large-scale hypothetical structures.

A machine learning approach for determining temperature-dependent bandgap of metal oxides utilizing Allen–Heine–Cardona theory and O’Donnell model parameterization

To evaluate the high temperature sensing properties of metal oxide and perovskite materials suitable for use in combustion environments, it is necessary to understand the temperature dependence of their bandgaps. Although such temperature-driven changes can be calculated via the Allen–Heine–Cardona (AHC) theory, which assesses electron–phonon coupling for the bandgap correction at given temperatures, this approach is computationally demanding. Another approach to predict bandgap temperature-dependence is the O’Donnell model, which uses analytical expressions with multiple fitting parameters that require bandgap information at 0 K. This work employs data-driven Gaussian process regression (GPR) to predict the parameters employed in the O’Donnell model from a set of physical features. We use a sample of 54 metal oxides for which density functional theory has been performed to calculate the bandgap at 0 K, and the AHC calculations have been carried out to determine the shift in the bandgap at non-zero temperatures. As the AHC calculations are impractical for high-throughput screening of materials, the developed GPR model attempts to alleviate this issue by predicting the O'Donnell parameters purely from physical features. To mitigate the reliability issues arising from the very small size of the dataset, we apply a Bayesian technique to improve the generalizability of the data-driven models as well as quantify the uncertainty associated with the predictions. The method captures well the overall trend of the O’Donnell parameters with respect to a reduced feature set obtained by transforming the available physical features. Quantifying the associated uncertainty helps us understand the reliability of the predictions of the O’Donnell parameters and, therefore, the bandgap as a function of temperature for any novel material.

High-Throughput Screening of Sulfur Reduction Reaction Catalysts Utilizing Electronic Fingerprint Similarity.

The catalytic performance is determined by the electronic structure near the Fermi level. This study presents an effective and simple screening descriptor, i.e., the one-dimensional density of states (1D-DOS) fingerprint similarity, to identify potential catalysts for the sulfur reduction reaction (SRR) in lithium-sulfur batteries. The Δ1D-DOS in relation to the benchmark W2CS2 was calculated. This method effectively distinguishes and identifies 30 potential candidates for the SRR from 420 types of MXenes. Further analysis of the Gibbs free energy profiles reveals that MXene candidates exhibit promising thermodynamic properties for SRR, with the protocol achieving an accuracy rate exceeding 93%. Based on the crystal orbital Hamilton population (COHP) and differential charge analysis, it is confirmed that the Δ1D-DOS could effectively differentiate the interaction between MXenes and lithium polysulfide (LiPS) intermediates. This study underscores the importance of the electronic fingerprint in catalytic performance and thus may pave a new way for future high-throughput material screening for energy storage applications.

Sampling-accelerated prediction of phonon scattering rates for converged thermal conductivity and radiative properties

The prediction of thermal conductivity and radiative properties is crucial. However, computing phonon scattering, especially for four-phonon scattering, could be prohibitively expensive, and the thermal conductivity for silicon after considering four-phonon scattering is significantly under-predicted and not converged in the literature. Here we propose a method to estimate scattering rates from a small sample of scattering processes using maximum likelihood estimation. The calculation of scattering rates and associated thermal conductivity and radiative properties are dramatically accelerated by three to four orders of magnitude. This allows us to use an unprecedented q-mesh (discretized grid in the reciprocal space) of 32 × 32 × 32 for calculating four-phonon scattering of silicon and achieve a converged thermal conductivity value that agrees much better with experiments. The accuracy and efficiency of our approach make it ideal for the high-throughput screening of materials for thermal and optical applications.

Predicting Two-Photon Absorption Spectra of Octupolar Molecules: A Deep-Learning Approach Based Exclusively on Molecular Structures.

Octupolar molecules possessing a strong two-photon response are vital for numerous advanced applications. However, accurately predicting their two-photon absorption (TPA) spectra requires high-precision quantum chemical calculations, which are computationally expensive due to repeated simulations of molecular excited-state properties. To address this challenge, we introduce a deep learning approach capable of rapidly and accurately forecasting TPA spectra for octupolar molecules. By leveraging the geometric structure as an initial descriptor, we employ a graph neural network to predict the maximum two-photon transition wavelength and cross-section. Our model demonstrates a mean absolute percentage error of less than 4% compared to time-dependent density-functional theory calculations, effectively reproducing experimental observations. Notably, this deep learning technique is nearly 100 000 times faster than comparable quantum calculations, making it an efficient and cost-effective tool for simulating TPA properties of octupolar molecules. Furthermore, this method holds great promise for the high-throughput screening of exceptional TPA materials.

High throughput screening of semiconductors with low lattice thermal transport induced by long-range interactions.

Semiconductors with long-range interactions (LRI) due to resonant bonding exhibit delocalized electronic states and low lattice thermal conductivity, contributing to the efficiency of heat-to-electricity conversion. Here, we build a descriptor for high-throughput screening of LRI materials from the second-order interaction force constants. We identify 75 semiconducting candidates from the binary compounds in the MatHub-3d database that contain LRI. By analyzing the bonding properties of LRI atoms, we classify LRI in materials into two categories: type I and type II. In the structural unit of type I LRI, the atoms have strong bond connections, while a weak bond exists between the two groups in the structural unit of type II LRI. We have identified atypical type I LRI formed by Sb-Sb and Mg-Mg pairs in the emerging thermoelectric material Mg3Sb2, resulting in the softening of TA1 phonons and large anharmonicity. For type II LRI, the LRI of Ge-Ge and Se-Se pairs in R3m-GeSe can cross different layers. Moreover, we observe a combination of type II LRI and rattling effect in BaSe2 to restrict thermal transport. This work is of great significance for understanding the relationship between LRI and thermal transport properties, and for designing new LRI-induced low lattice thermal conductivity materials.

High-throughput screening of stable layered anode materials A2TMO3Cl for chloride-ion batteries

A vacuum cleaner-like “TM–O–A–Cl” configuration simultaneously retains Cl-storage/binding capacity in A2TMO3Cl.

High-throughput computational materials screening of transition metal peroxides.

Semiconductor materials of abnormal stoichiometric ratio often exhibit unique properties, yet it is still a challenge to determine the structures of such materials in an efficient way. Herein, we propose a method for structurally biased screening according to the coordination numbers and the numbers of Wyckoff positions, balancing the atom local environment and the global symmetry of structures. Based on first-principles calculations, we have predicted two metastable peroxides P21/c-ScO2 and Pmmn-TiO3 with more than six coordination points. For these two structures, the most stable intrinsic defect is the oxygen vacancy (VO) at the peroxide anion (O2-2), which induces the absence of antibonding orbital formed by O2-2 near the valence band maximum. With the introduction of VO, the decrease of coordination numbers leads to charge recombination, and results in the appearance of an ordered phase TiO2.5 with stronger Ti-O orbital hybridization. The proposed method presents a promising and feasible approach for the screening of novel compounds.

Phase-Change Memory from Molecular Tellurides.

Phase-change memory (PCM) is an emerging memory technology based on the resistance contrast between the crystalline and amorphous states of a material. Further development and realization of PCM as a mainstream memory technology rely on innovative materials and inexpensive fabrication methods. Here, we propose a generalizable and scalable solution-processing approach to synthesize phase-change telluride inks in order to meet demands for high-throughput material screening, increased energy efficiency, and advanced device architectures. Bulk tellurides, such as Sb2Te3, GeTe, Sc2Te3, and TiTe2, are dissolved and purified to obtain inks of molecular metal telluride complexes. This allowed us to unlock a wide range of solution-processed ternary tellurides by the simple mixing of binary inks. We demonstrate accurate and quantitative composition control, including prototype materials (Ge-Sb-Te) and emerging rare-earth-metal telluride-doped materials (Sc-Sb-Te). Spin-coating and annealing convert ink formulations into high-quality, phase-pure telluride films with preferred orientation along the (00l) direction. Deposition engineering of liquid tellurides enables thickness-tunable films, infilling of nanoscale vias, and film preparation on flexible substrates. Finally, we demonstrate cyclable and non-volatile prototype memory devices, achieving performance indicators such as resistance contrast and low reset energy on par with state-of-the-art sputtered PCM layers.

Transition Temperature for Spin-Crossover Materials with the Mean Value Ensemble Hubbard-U Correction.

Calculation of transition temperatures T1/2 for thermally driven spin-crossover in condensed phases is challenging, even with sophisticated state-of-the-art density functional approximations. The first issue is the accuracy of the adiabatic crossover energy difference ΔEHL between the low- and high-spin states of the bistable metal-organic complexes. The other is the proper inclusion of entropic contributions to the Gibbs free energy from the electronic and vibrational degrees of freedom. We discuss the effects of treatments of both contributions upon the calculation of thermochemical properties for a set of 20 spin-crossover materials using a Hubbard-U correction obtained from a reference ensemble spin-state. The U values obtained from a simplest bimolecular representation may overcorrect, somewhat, the ΔEHL values, hence giving somewhat excessive reduction of the T1/2 results with respect to their U = 0 values in the crystalline phase. We discuss the origins of the discrepancies by analyzing different sources of uncertainties. By use of a first-coordination-sphere approximation and the assumption that vibrational contributions from the outermost atoms in a metal-organic complex are similar in both low- and high-spin states, we achieve T1/2 results with the low-cost, widely used PBE generalized gradient density functional approximation comparable to those from the more costly, more sophisticated r2SCAN meta-generalized gradient approximation. The procedure is promising for use in high-throughput materials screening, because it combines rather low computational effort requirements with freedom from user manipulation of parameters.

New Family of Interstitial Oxygen Ion Conductor Discovered By High-Throughput Computational Screening

Discovery and design of highly efficient oxygen-active materials which react with, absorb, and transport oxygen is essential for fuel cells and related applications. Currently, the predominant families of oxygen-active materials used in commercial devices are perovskites and fluorites, which diffuse oxygen via a vacancy mechanism and only have adequate oxygen diffusivity and surface exchange at high temperatures (e.g., ~800 ˚C), impeding their use at lower temperatures desirable for more cost-effective and long-lasting device operation. The development of alternative electrolyte/electrode materials with high oxygen ionic conductivity at low temperatures (e.g., room temperature to 400 °C) is of great interest. Interstitial oxygen ion conductors typically have significantly lower migration barriers than even state-of-the-art vacancy-mediated oxygen conductors, which can provide orders of magnitude kinetic enhancement at low temperatures. In addition, the interstitial oxygen defects typically become more thermodynamically favorable as the temperature is lowered, potentially providing an additional boost to the oxygen conductivity at low temperatures from a higher mediating defect concentration. The synergistic combination of low migration barriers and thermodynamically favorable interstitial oxygen content at low temperatures provides the tantalizing possibility that interstitial oxygen conductors may lead to transformative performance improvements in oxygen-active materials.In this work, we developed an approach to discover new interstitial-mediated oxygen conductors by performing high-throughput computational screening of the 34k oxide materials from the Materials Project database and discovered a new structural family of interstitial oxygen diffusers, perrierite-type oxide La4Mn5Si4O22+ 𝜹 (LMS). We used a hierarchy of screening criteria including the geometric free space, thermodynamic stability, synthesizability, redox-active elements, and presence of short diffusion pathways. This approach, based largely on simple, physically intuitive properties, quickly winnowed the field of 34k oxides down to just 345 oxides, an approximate 99% reduction in the search space considered, with essentially no significant computational cost. From this reduced set of promising materials, ab initio simulations were performed to calculate the interstitial oxygen formation energy and migration energy. Our screening has yielded several material families to date, among which the A4B5C4O22+ 𝜹, with La4Mn5Si4O22+ 𝜹 (LMS) as a representative material, was selected for investigation with higher-fidelity DFT hybrid functional calculations and experimental validations. The calculated formation energy of interstitial oxygen is -0.11 eV at a concentration of 2.3% under air condition, indicating that LMS is predicted to be oxygen hyperstoichiometric with an expected composition of La4Mn5Si4O22+0.5 in air. Two separate interstitial and interstitialcy (“kick-out”) diffusion mechanisms are revealed by ab initio molecular dynamics (AIMD) simulation. For the interstitial diffusion pathway, the moves through the channel in between the zigzag-arranged sorosilicate Si2O7 groups, with a barrier of 0.69 eV. For the interstitialcy diffusion pathway, the moves along the corner-sharing Si2O7-MnO2-Si2O7 framework by kicking out the lattice oxygen, with a barrier of 0.74 eV. These results align well with the experimental findings of stable interstitial oxygen content and high oxygen mobility in LMS (Please see a separate poster from Md Sariful Sheikhfor experimental results). This work provides a systematic approach to screening a large number of prospective materials and opens a new avenue of finding promising interstitial oxide ion conducting materials for potential applications in a range of fast oxygen ion conduction-based technologies.

High-Throughput Screening of High-Performance Thermoelectric Materials with Gibbs Free Energy and Electronegativity.

Thermoelectric (TE) materials are an important class of energy materials that can directly convert thermal energy into electrical energy. Screening high-performance thermoelectric materials and improving their TE properties are important goals of TE materials research. Based on the objective relationship among the molar Gibbs free energy (Gm), the chemical potential, the Fermi level, the electronegativity (X) and the TE property of a material, a new method for screening TE materials with high throughput is proposed. This method requires no experiments and no first principle or Ab initio calculation. It only needs to find or calculate the molar Gibbs free energy and electronegativity of the material. Here, by calculating a variety of typical and atypical TE materials, it is found that the molar Gibbs free energy of Bi2Te3 and Sb2Te3 from 298 to 600 K (Gm = -130.20~-248.82 kJ/mol) and the electronegativity of Bi2Te3 and Sb2Te3 and PbTe (X = 1.80~2.21) can be used as criteria to judge the potential of materials to become high-performance TE materials. For good TE compounds, Gm and X are required to meet the corresponding standards at the same time. By taking Gm = -130.20~-248.82 kJ/mol and X = 1.80~2.21 as screening criteria for high performance TE materials, it is found that the Gm and X of all 15 typical TE materials and 9 widely studied TE materials meet the requirement very well, except for the X of Mg2Si, and 64 pure substances are screened as potential TE materials from 102 atypical TE materials. In addition, with reference to their electronegativity, 44 pure substances are selected directly from a thermochemical data book as potential high-performance TE materials. A particular finding is that several carbides, such as Be2C, CaC2, BaC2, SmC2, TaC and NbC, may have certain TE properties. Because the Gm and X of pure substances can be easily found in thermochemical data books and calculated using the X of pure elements, respectively, the Gm and X of materials can be used as good high-throughput screening criteria for predicting TE properties.

Force field-inspired transformer network assisted crystal density prediction for energetic materials

Machine learning has great potential in predicting chemical information with greater precision than traditional methods. Graph neural networks (GNNs) have become increasingly popular in recent years, as they can automatically learn the features of the molecule from the graph, significantly reducing the time needed to find and build molecular descriptors. However, the application of machine learning to energetic materials property prediction is still in the initial stage due to insufficient data. In this work, we first curated a dataset of 12,072 compounds containing CHON elements, which are traditionally regarded as main composition elements of energetic materials, from the Cambridge Structural Database, then we implemented a refinement to our force field-inspired neural network (FFiNet), through the adoption of a Transformer encoder, resulting in force field-inspired Transformer network (FFiTrNet). After the improvement, our model outperforms other machine learning-based and GNNs-based models and shows its powerful predictive capabilities especially for high-density materials. Our model also shows its capability in predicting the crystal density of potential energetic materials dataset (i.e. Huang & Massa dataset), which will be helpful in practical high-throughput screening of energetic materials.

Atomic stiffness for bulk modulus prediction and high-throughput screening of ultraincompressible crystals

Determining bulk moduli is central to high-throughput screening of ultraincompressible materials. However, existing approaches are either too inaccurate or too expensive for general applications, or they are limited to narrow chemistries. Here we define a microscopic quantity to measure the atomic stiffness for each element in the periodic table. Based on this quantity, we derive an analytic formula for bulk modulus prediction. By analyzing numerous crystals from first-principles calculations, this formula shows superior accuracy, efficiency, universality, and interpretability compared to previous empirical/semiempirical formulae and machine learning models. Directed by our formula predictions and verified by first-principles calculations, 47 ultraincompressible crystals rivaling diamond are identified from over one million material candidates, which extends the family of known ultraincompressible crystals. Finally, treasure maps of possible elemental combinations for ultraincompressible crystals are created from our theory. This theory and insights provide guidelines for designing and discovering ultraincompressible crystals of the future.

High-throughput screening of heat-shielding carrier-doped semiconductor materials via first principles calculations

Heat-shielding carrier-doped semiconductor materials that exploit the localized surface plasmon resonance of nanoparticles to absorb NIR light while transmitting visible light from sunlight were searched for using first-principles calculations. Carrier doping into semiconductors can be performed by element replacement/insertion; however, various replacement/insertion structures are possible, and consequently, many calculations are required. Therefore, in this study, we performed virtual carrier doping and calculated the heat-shielding characteristics with low computational cost. From among 2964 semiconductor materials, materials with low visible light absorption after carrier doping were extracted, and from among these, materials with high heat-shielding properties were extracted. For these materials, the bandgaps were calculated using hybrid functionals, and the dopability was evaluated using the branch point energy. Finally, 24 materials were proposed as candidates for heat-shielding materials.

Evaluating In-Plane Thermal Expansion of Two-Dimensional Layered Materials via Effective Descriptors

Two-dimensional layered materials show promising applications in miniaturized devices, such as transistors, spintronics, and field emitters. However, substantial thermal management issues, including thermal mismatch and thermal stress, may degrade device performance. To address such challenges, the thermal expansion (TE) anisotropy determined by the structure feature of layered material needs to be well understood. Here, we propose two new descriptors to evaluate the TE behavior of layered materials, namely the axial elastic deviation factor σi and the axial net thermal stress fi along the ith direction. The former, defined as the normalized elastic element difference of material elastic tensor C and compliance tensor S, can distinguish whether the thermal expansion of a material is driven by phonons (with small σi) or elastic property (with large σi) with few computational costs. The latter, axial stress (in GPa/K) induced by temperature, shows an accurate determination of the positive or negative thermal expansion along different in-plane directions of layered materials. Based on the analysis of descriptors, we found that PtS2 and PtSe2 are featured with a larger axial elastic deviation factor (>23%). Considering the elastic property, we for the first time report the in-plane negative thermal expansion in PtS2 (−1.2 ppm/K) and PtSe2 (−0.8 ppm/K). Our work provides a unified understanding of TE causes of layered materials via effective descriptors, which can serve as a guideline for high-throughput screening of thermal expansion materials and subsequent device design.

Development and analysis of a multi-module peristaltic simulator for gastrointestinal research

Many existing in vitro digestion systems do not accurately represent the peristaltic contractions of the gastrointestinal system; most of the systems that have physiologically-relevant peristaltic contractions have low throughput and can only test one sample at a time. A device has been developed that provides simulated peristaltic contractions for up to 12 digestion modules simultaneously using rollers of varying width to modulate the dynamics of the peristaltic motion. The force applied to a simulated food bolus varied from 2.61 ± 0.03 N to 4.51 ± 0.16 N (p < 0.05) depending on roller width. Video analysis showed that the degree of occlusion of the digestion module varied from 72.1 ± 0.4% to 84.6 ± 1.2% (p < 0.05). A multiphysics, computational fluid dynamics model was created to understand the fluid flow. The fluid flow was also examined experimentally using video analysis of tracer particles. The model-predicted maximum fluid velocity in the peristaltic simulator incorporating the thin rollers was 0.016 m/s, and the corresponding value measured using tracer particles was 0.015 m/s. The occlusion, pressure, and fluid velocity in the new peristaltic simulator fell within physiologically representative ranges. Although no in vitro device perfectly recreates the conditions of the gastrointestinal system, this novel device is a flexible platform for future gastrointestinal research and could allow for high-throughput screening of food materials for health-promoting properties under conditions representative of human gastrointestinal motility.