An electrochemical series for materials.

The electrochemical series is a useful tool in electrochemistry, but its effectiveness in materials chemistry is limited by the fact that the standard electrochemical series is based on a relatively small set of reactions, many of which are measured in aqueous solutions. To address this problem, we have used machine learning to create an electrochemical series for inorganic materials from tens of thousands of entries in the Inorganic Crystal Structure Database. We demonstrate that this series is generally more consistent with oxidation states in solid-state materials than the series based on aqueous ions. The electrochemical series was constructed by developing and parameterizing a physical, human-interpretable model of oxidation states in materials. We show that this model enables the prediction of oxidation states from composition in a way that is more accurate than a state-of-the-art transformer-based neural network model. We present applications of our approach to structure prediction, materials discovery, and materials electrochemistry, and we discuss possible additional applications and areas for improvement. To facilitate the use of our approach, we introduce a freely available website and API.

High-throughput determination of grain size distributions by EBSD with low-discrepancy sampling.

Because microstructure plays an important role in the mechanical properties of structural materials, developing the capability to quantify microstructures rapidly is important to enabling high-throughput screening of structural materials. Electron backscatter diffraction (EBSD) is a common method for studying microstructures and extracting information such as grain size distributions (GSDs), but is not particularly fast and thus could be a bottleneck in high-throughput systems. One approach to accelerating EBSD is to reduce the number of points that must be scanned. In this work, we describe an iterative method for reducing the number of scan points needed to measure GSDs using incremental low-discrepancy sampling, including on-the-fly grain size calculations and a convergence test for the resulting GSD based on the Kolmogorov-Smirnov test. We demonstrate this method on five real EBSD maps collected from magnesium AZ31B specimens and compare the effectiveness of sampling according to two different low discrepancy sequences, the Sobol and R2 sequences, and random sampling. We find that R2 sampling is able to produce GSDs that are statistically very similar to the GSDs of the full density grids using, on average, only 52% of the total scan points. For EBSD maps that contained monodisperse GSDs and over 1000 grains, R2 sampling only required an average of 39% of the total EBSD points.

Catalytic Activity Maps for Alloy Nanoparticles.

To enable rational design of alloy nanoparticle catalysts, we develop an approach to generate catalytic activity maps of alloy nanoparticles on a grid of particle size and composition. The catalytic activity maps are created by using a quaternary cluster expansion to explicitly predict adsorbate binding energies on alloy nanoparticles of varying shape, size, and atomic order while accounting for interactions among the adsorbates. This cluster expansion is used in kinetic Monte Carlo simulations to predict activated nanoparticle structures and turnover frequencies on all surface sites. We demonstrate our approach on Pt-Ni octahedral nanoparticle catalysts for the oxygen reduction reaction (ORR), revealing that the specific activity is predicted to be optimized at an edge length of larger than 5.5 nm and a composition of about Pt0.85Ni0.15 and the mass activity is predicted to be optimized at an edge length of 3.3-3.8 nm and a composition of about Pt0.8Ni0.2.

Computationally generated maps of surface structures and catalytic activities for alloy phase diagrams.

To facilitate the rational design of alloy catalysts, we introduce a method for rapidly calculating the structure and catalytic properties of a substitutional alloy surface that is in equilibrium with the underlying bulk phase. We implement our method by developing a way to generate surface cluster expansions that explicitly account for the lattice parameter of the bulk structure. This approach makes it possible to computationally map the structure of an alloy surface and statistically sample adsorbate binding energies at every point in the alloy phase diagram. When combined with a method for predicting catalytic activities from adsorbate binding energies, maps of catalytic activities at every point in the phase diagram can be created, enabling the identification of synthesis conditions likely to result in highly active catalysts. We demonstrate our approach by analyzing Pt-rich Pt-Ni catalysts for the oxygen reduction reaction, finding 2 regions in the phase diagram that are predicted to result in highly active catalysts. Our analysis indicates that the Pt3Ni(111) surface, which has the highest known specific activity for the oxygen reduction reaction, is likely able to achieve its high activity through the formation of an intermetallic phase with L12 order. We use the generated surface structure and catalytic activity maps to demonstrate how the intermetallic nature of this phase leads to high catalytic activity and discuss how the underlying principles can be used in catalysis design. We further discuss the importance of surface phases and demonstrate how they can dramatically affect catalytic activity.

Undercoordinated Active Sites on 4H Gold Nanostructures for CO2 Reduction.

Electroreduction of CO2 is a promising approach toward artificial carbon recycling. The rate and product selectivity of this reaction are highly sensitive to the surface structures of electrocatalysts. We report here 4H Au nanostructures as advanced electrocatalysts for highly active and selective reduction of CO2 to CO. Au nanoribbons in the pure 4H phase, Au nanorods in the hybrid 4H/fcc phase, and those in the fcc phase are comparatively studied for the electroreduction of CO2. Both the activity and selectivity for CO production were found to exhibit the trend 4H-nanoribbons > 4H/fcc-nanorods > fcc-nanorods, with the 4H-nanoribbons achieving >90% Faradaic efficiency toward CO. Electrochemical probing and cluster expansion simulations are combined to elucidate the surface structures of these nanocrystals. The combination of crystal phase and shape control gives rise to the preferential exposure of undercoordinated sites. Further density functional theory calculations confirm the high reactivity of such undercoordinated sites.

Machine learning for interatomic potential models.

The use of supervised machine learning to develop fast and accurate interatomic potential models is transforming molecular and materials research by greatly accelerating atomic-scale simulations with little loss of accuracy. Three years ago, Jörg Behler published a perspective in this journal providing an overview of some of the leading methods in this field. In this perspective, we provide an updated discussion of recent developments, emerging trends, and promising areas for future research in this field. We include in this discussion an overview of three emerging approaches to developing machine-learned interatomic potential models that have not been extensively discussed in existing reviews: moment tensor potentials, message-passing networks, and symbolic regression.

Ensemble Effect in Bimetallic Electrocatalysts for CO2 Reduction.

Alloying is an important strategy for the design of catalytic materials beyond pure metals. The conventional alloy catalysts however lack precise control over the local atomic structures of active sites. Here we report on an investigation of the active-site ensemble effect in bimetallic Pd-Au electrocatalysts for CO2 reduction. A series of Pd@Au electrocatalysts are synthesized by decorating Au nanoparticles with Pd of controlled doses, giving rise to bimetallic surfaces containing Pd ensembles of various sizes. Their catalytic activity for electroreduction of CO2 to CO exhibits a nonlinear behavior in dependence of the Pd content, which is attributed to the variation of Pd ensemble size and the corresponding tuning of adsorption properties. Density functional theory calculations reveal that the Pd@Au electrocatalysts with atomically dispersed Pd sites possess lower energy barriers for activation of CO2 than pure Au and are also less poisoned by strongly binding *CO intermediates than pure Pd, with an intermediate ensemble size of active sites, such as Pd dimers, giving rise to the balance between these two rate-limiting factors and achieving the highest activity for CO2 reduction.

Mechanisms for hydrogen evolution on transition metal phosphide catalysts and a comparison to Pt(111).

Earth-abundant transition metal phosphides have been demonstrated to be promising alternative catalysts to replace Pt for hydrogen evolution reaction (HER). However, the mechanism for the hydrogen evolution reaction on transition metal phosphides remains unclear. Here, we explore the catalytically active sites and the reaction mechanisms on a variety of model transition metal phosphide surfaces by building cluster expansion models and running Monte Carlo simulations. We demonstrate that the effect of hydrogen coverage, interaction between hydrogen atoms and desorption kinetics all dictate the HER mechanisms and the active sites, and we propose mechanisms that are in good agreement with experimental studies. The present method provides a general and effective way to probe the active sites and study the mechanisms of catalytic reactions, which can facilitate the rational design of highly active electrocatalysts.

Roles of Mo Surface Dopants in Enhancing the ORR Performance of Octahedral PtNi Nanoparticles.

Doping with a transition metal was recently shown to greatly boost the activity and durability of PtNi/C octahedral nanoparticles (NPs) for the oxygen reduction reaction (ORR), but its specific roles remain unclear. By combining electrochemistry, ex situ and in situ spectroscopic techniques, density functional theory calculations, and a newly developed kinetic Monte Carlo model, we showed that Mo atoms are preferentially located on the vertex and edge sites of Mo-PtNi/C in the form of oxides, which are stable within the wide potential window of the electrochemical cycle. These surface Mo oxides stabilize adjacent Pt sites, hereby stabilizing the octahedral shape enriched with (111) facets, and lead to increased concentration of Ni in subsurface layers where they are protected against acid dissolution. Consequently, the favorable Pt3Ni(111) structure for the ORR is stabilized on the surface of PtNi/C NPs in acid against voltage cycling. Significantly, the unusual potential-dependent oxygen coverage trend on Mo-doped PtNi/C NPs as revealed by the surface-sensitive Δµ analysis suggests that the Mo dopants may also improve the ORR kinetics by modifying the coordination environments of Pt atoms on the surface. Our studies point out a possible way to stabilize the favorable shape and composition established on conceptual catalytic models in practical nanoscale catalysts.

The Use of Cluster Expansions To Predict the Structures and Properties of Surfaces and Nanostructured Materials.

The construction of cluster expansions parametrized by first-principles calculations is a powerful tool for calculating properties of materials. In this Perspective, we discuss the application of cluster expansions to surfaces and nanomaterials. We review the fundamentals of the cluster expansion formalism and how machine learning is used to improve the predictive accuracy of cluster expansions. We highlight several representative applications of cluster expansions to surfaces and nanomaterials, demonstrating how cluster expansions help researchers build structure-property relationships and enable rational design to accelerate the discovery of new materials. Potential applications and future challenges of cluster expansions are also discussed.

Theoretical Insights into the Effects of Oxidation and Mo-Doping on the Structure and Stability of Pt-Ni Nanoparticles.

Pt-Ni nanoparticles are promising catalysts for the oxygen reduction reaction but they suffer from Ni dissolution in oxidizing conditions. It has recently been shown that it is possible to stabilize octahedral Pt-Ni nanoparticles by doping them with a small amount of Mo. Using ab initio calculations and a quaternary cluster expansion, we provide atomic-scale explanations for the enhanced stability of Mo-doped Pt-Ni nanoparticles. We predict that for Mo-doped Pt3Ni nanoparticles with only a small amount of Mo doping (around 1.6% mole fraction) the equilibrium concentration of Ni atoms on the particle surface is greatly reduced, limiting the rate at which Ni atoms dissolve from the particles. Mo doping also increases Pt/Ni vacancy formation energies in the surface layer, which further stabilizes the nanoparticles against Ni dissolution and helps preserve the nanoparticle shape. Our calculations also reveal insights into the shape evolution of Pt-Ni nanoparticles: the preferential oxidation of edges can make (111) face sites more vulnerable to dissolution than edge sites, which may contribute to the observed formation of Pt-Ni nanoframes and nanoparticles with concave surfaces.

Switchable dual-polarization external cavity tunable laser.

An external cavity laser is demonstrated based on the hybrid integration of an InP-based gain element, a half-wave plate, and thermally drivable polymer waveguide circuits. The laser has one oscillation region but two outputs for TE and TM emissions. The central wavelength can be tuned 20 nm at 20 mW heater electrical power. The TM path undergoes a 1.4 dB power penalty due to the presence of the half-wave plate. However, the on-chip thermo-optic switch (TOS) can compensate for this imbalance and steer the laser into an equal TE and TM output power. The TOS can also be adjusted to prefer one polarization path over the other with ∼10 dB extinction ratio.

A database of low-energy atomically precise nanoclusters.

The chemical and structural properties of atomically precise nanoclusters are of great interest in numerous applications, but the structures of the clusters can be computationally expensive to predict. In this work, we present the largest database of cluster structures and properties determined using ab-initio methods to date. We report the methodologies used to discover low-energy clusters as well as the energies, relaxed structures, and physical properties (such as relative stability, HOMO-LUMO gap among others) for 63,015 clusters across 55 elements. We have identified clusters for 593 out of 1595 cluster systems (element-size pairs) explored by literature that have energies lower than those reported in literature by at least 1 meV/atom. We have also identified clusters for 1320 systems for which we were unable to find previous low-energy structures in the literature. Patterns in the data reveal insights into the chemical and structural relationships among the elements at the nanoscale. We describe how the database can be accessed for future studies and the development of nanocluster-based technologies.

Ionic Conduction through Reaction Products at the Electrolyte-Electrode Interface in All-Solid-State Li+ Batteries.

All-solid-state lithium-ion batteries have attracted significant research interest for providing high power and energy densities with enhanced operational safety. Despite the discoveries of solid electrolyte materials with superionic conductivities, it remains a challenge to maintain high rate capability in all-solid lithium-ion batteries in long-term operation. The observed rate degradation has been attributed to reactivity and resistance at the electrode-electrolyte interfaces. We examine interfaces formed between eight electrolytes including garnet, LiPON, and Li10GeP2S12 (LGPS) and seven electrode materials including an NCM cathode and a metallic Li anode and identify the most rapid lithium-ion diffusion pathways through metastable arrangements of product phases that may precipitate out at each interface. Our analysis accounts for possible density functional theory (DFT) error, metastability, and finite-temperature effects by statistically sampling thousands of possible phase diagrams for each interface. The lithium-ion conductivities in the product phases at the interface are evaluated using machine-learned interatomic potentials trained on the fly. In nearly all electrode-electrolyte interfaces we evaluate, we predict that lithium-ion conduction in the product phases making up the interphase region becomes the rate-limiting step for battery performance.

Substrate-directed synthesis of MoS2 nanocrystals with tunable dimensionality and optical properties.

Two-dimensional transition-metal dichalcogenide (TMD) crystals are a versatile platform for optoelectronic, catalytic and quantum device studies. However, the ability to tailor their physical properties through explicit synthetic control of their morphology and dimensionality is a major challenge. Here we demonstrate a gas-phase synthesis method that substantially transforms the structure and dimensionality of TMD crystals without lithography. Synthesis of MoS2 on Si(001) surfaces pre-treated with phosphine yields high-aspect-ratio nanoribbons of uniform width. We systematically control the width of these nanoribbons between 50 and 430 nm by varying the total phosphine dosage during the surface treatment step. Aberration-corrected electron microscopy reveals that the nanoribbons are predominantly 2H phase with zig-zag edges and an edge quality that is comparable to, or better than, that of graphene and TMD nanoribbons prepared through conventional top-down processing. Owing to their restricted dimensionality, the nominally one-dimensional MoS2 nanocrystals exhibit photoluminescence 50 meV higher in energy than that from two-dimensional MoS2 crystals. Moreover, this emission is precisely tunable through synthetic control of crystal width. Directed crystal growth on designer substrates has the potential to enable the preparation of low-dimensional materials with prescribed morphologies and tunable or emergent optoelectronic properties.

Materials with the CrVO4 structure type as candidate superprotonic conductors.

Proton conducting oxides have the potential to improve the efficiency of solid oxide fuel cells and electrolyzers, yet many oxide structures remain relatively unexplored for the ability to conduct protons. To accelerate the search for novel proton-conducting oxides, we have performed a computational screen of the proton migration energy in 41 different commonly-occurring oxide structure types. The results of this screen, which are supported by a comprehensive set of density functional theory calculations, indicate that known materials with the CrVO4 structure type have an average migration energy for proton diffusion of less than 0.2 eV, with several known materials having calculated migration energies below 0.1 eV. These results indicate that materials with the CrVO4 structure type, which to our knowledge have not been previously explored as candidate proton conductors, may exhibit very high proton conductivity that surpasses that of leading proton-conducting oxides. We present the results of our screen as well as diffusion dimensionality analysis and thermodynamic stability analysis for materials with the CrVO4 structure.

Ordered Intermetallic Pd3Bi Prepared by an Electrochemically Induced Phase Transformation for Oxygen Reduction Electrocatalysis.

The synthesis of alloys with long-range atomic-scale ordering (ordered intermetallics) is an emerging field of nanochemistry. Ordered intermetallic nanoparticles are useful for a wide variety of applications such as catalysis, superconductors, and magnetic devices. However, the preparation of nanostructured ordered intermetallics is challenging in comparison to disordered alloys, hindering progress in material development. Herein, we report a process for converting colloidally synthesized ordered intermetallic PdBi2 to ordered intermetallic Pd3Bi nanoparticles under ambient conditions by electrochemical dealloying. The low melting point of PdBi2 corresponds to low vacancy formation energies, which enables the facile removal of the Bi from the surface while simultaneously enabling interdiffusion of the constituent atoms via a vacancy diffusion mechanism under ambient conditions. The resulting phase-converted ordered intermetallic Pd3Bi exhibits 11 times and 3.5 times higher mass activity and high methanol tolerance for the oxygen reduction reaction compared with Pt/C and Pd/C, respectively, which is the highest reported for a Pd-based catalyst, to the best of our knowledge. These results establish a key development in the synthesis of noble-metal-rich ordered intermetallic phases with high catalytic activity and set forth guidelines for the design of ordered intermetallic compounds under ambient conditions.

Identifying models of dielectric breakdown strength from high-throughput data via genetic programming.

The identification of models capable of rapidly predicting material properties enables rapid screening of large numbers of materials and facilitates the design of new materials. One of the leading challenges for computational researchers is determining the best ways to analyze large material data sets to identify models that can rapidly predict a given property. In this paper, we demonstrate the use of genetic programming to generate simple models of dielectric breakdown based on 82 representative dielectric materials. We identified the band gap Eg and phonon cut-off frequency ωmax as the two most relevant features, and new classes of models featuring functions of Eg and ωmax were uncovered. The genetic programming approach was found to outperform other approaches for generating models, and we discuss some of the advantages of this approach.

Investigation of a Quantum Monte Carlo Protocol To Achieve High Accuracy and High-Throughput Materials Formation Energies.

High-throughput calculations based on density functional theory (DFT) methods have been widely implemented in the scientific community. However, depending on both the properties of interest as well as particular chemical/structural phase space, accuracy even for correct trends remains a key challenge for DFT. In this work, we evaluate the use of quantum Monte Carlo (QMC) to calculate material formation energies in a high-throughput environment. We test the performance of automated QMC calculations on 21 compounds with high quality reference data from the Committee on Data for Science and Technology (CODATA) thermodynamic database. We compare our approach to different DFT methods as well as different pseudopotentials, showing that errors in QMC calculations can be progressively improved especially when correct pseudopotentials are used. We determine a set of accurate pseudopotentials in QMC via a systematic investigation of multiple available pseudopotential libraries. We show that using this simple automated recipe, QMC calculations can outperform DFT calculations over a wide set of materials. Out of 21 compounds tested, chemical accuracy has been obtained in formation energies of 11 structures using our QMC recipe, compared to none using DFT calculations.