Machine learning (ML) models in the materials sciences that are validated by overly simplistic cross-validation (CV) protocols can yield biased performance estimates for downstream modeling or materials screening tasks. This can be particularly counterproductive for applications where the time and cost of failed validation efforts (experimental synthesis, characterization, and testing) are consequential. We propose a set of standardized and increasingly difficult splitting protocols for chemically and structurally motivated CV that can be followed to validate any ML model for materials discovery. Among several benefits, this enables systematic insights into model generalizability, improvability, and uncertainty, provides benchmarks for fair comparison between competing models with access to differing quantities of data, and systematically reduces possible data leakage through increasingly strict splitting protocols. Performing thorough CV investigations across increasingly strict chemical/structural splitting criteria, local vs. global property prediction tasks, small vs. large datasets, and structure vs. compositional model architectures, some common threads are observed; however, several marked differences exist across these exemplars, indicating the need for comprehensive analysis to fully understand each model's generalization accuracy and potential for materials discovery. For this we provide a general-purpose, featurization-agnostic toolkit, MatFold, to automate reproducible construction of these CV splits and encourage further community use in model benchmarking.
Batteries composed of CFx cathodes have high theoretical specific capacities (>860 mA h g−1). Attempts at realizing such batteries coupled with Li anodes have failed to deliver on this promise, however, due to a discharge voltage plateau below the theoretical maximum lowering the realized energy density and difficulties with recharging the system. In this study, we use first-principles calculations to investigate novel carbon allotropes for these battery systems: graphdiyne and “holey” graphene. We first identify stable flourination structures and calculate their band gaps. We demonstrate that the holes in these carbon allotropes can induce the formation of an amorphous LiF network within the carbon and that this formation may, in fact, be kinetically favored. For structures where amorphous LiF forms within the carbon, we predict it is easier to recharge and higher discharge voltages can be achieved. If the LiF forms outside the carbon product, however, it will be crystalline in form and lead to lower discharge voltages and more difficulty in recharging the systems. Finally, we simulate XPS spectra of representative cases, demonstrating an experimental pathway for determining the reaction pathway of these systems. Our work suggests CFx allotropes with holes in them as potential targets for high capacity, rechargeable cathodes for Li batteries, provided they lead to the formation of amorphous LiF within the C structure.
Entropy is a state variable that may be obtained from any thermodynamically complete equation of state (EOS). However, hydrocode calculations that output the entropy often contain numerical errors; this is not because of the EOS, but rather the solution techniques that are used in hydrocodes (especially Eulerian) such as convection, remapping, and artificial viscosity. In this work, empirical correlations are investigated to reduce the errors in entropy without altering the solution techniques for the conservation of mass, momentum, and energy. Specifically, these correlations are developed for the function of entropy ZS, and they depend upon the net artificial viscous work, as determined via Sandia National Laboratories’ shock physics hydrocode CTH. These results are a continuation of a prior effort to implement the entropy-based CREST reactive burn model in CTH, and they are presented here to stimulate further interest from the shock physics community. Future work is planned to study higher-dimensional shock waves, shock wave interactions, and possible ties between the empirical correlations and a physical law.
Ab initio molecular dynamics (AIMD) simulations were carried out to investigate the equation of state of Nb2O5 and its pressure-density relationship under shock conditions. The focus of this study is on the monoclinic B−Nb2O5 (C2/c) polymorph. Enthalpy calculations from AIMD trajectories at 300 K show that the pressure-induced transformation between the thermodynamically most stable crystalline monoclinic parent phase H−Nb2O5 (P2/m) and B−Nb2O5 occurs at ∼1.9 GPa. This H→B transition is energetically more favorable than the H→L(Pmm2) pressure-induced transition recently observed at ∼5.9−9.0 GPa. The predicted shock properties of Nb2O5 polymorphs are also compared to their Nb and NbO2 counterparts to assess the impact of niobium oxidation on shock response.
Entropy is a state variable that may be obtained from any thermodynamically complete equation of state (EOS). However, hydrocode calculations that output the entropy often contain numerical errors; this is not because of the EOS, but rather the solution techniques that are used in hydrocodes (especially Eulerian) such as convection, remapping, and artificial viscosity. In this work, empirical correlations are investigated to reduce the errors in entropy without altering the solution techniques for the conservation of mass, momentum, and energy. Specifically, these correlations are developed for the function of entropy ZS, and they depend upon the net artificial viscous work, as determined via Sandia National Laboratories’ shock physics hydrocode CTH. These results are a continuation of a prior effort to implement the entropy-based CREST reactive burn model in CTH, and they are presented here to stimulate further interest from the shock physics community. Future work is planned to study higher-dimensional shock waves, shock wave interactions, and possible ties between the empirical correlations and a physical law.
The inaugural Advancing Chemical Safety and Security Education symposium was held at the 27th IUPAC International Conference on Chemistry Education (ICCE2024). Speakers showcased innovative strategies for seamlessly integrating security concepts into established safety programs, addressing specific needs of diverse academic institutions, and evaluating the effectiveness of different pedagogical approaches. Here, this proceedings publication encapsulates insights from 11 oral presentations, 12 poster presentations, and panel discussions including key recommendations for future advancements in educating chemical safety and security education for academic and industry audiences.
The binding of the sulfate ligand group to monovalent cations in the presence of water is important for many systems. To understand the structure and energetics of sulfate complexes, we use density functional theory to study ethyl sulfate binding to the monovalent cations Li+, Na+, and K+, and to water. The free energies of binding and optimal structures are calculated for a range of the number of ethyl sulfates and waters. Without water, the most optimal structure for all the cations is bidentate binding by two ethyl sulfates, yielding a 4-fold coordination. With water, the lowest free energy structures also have two ethyl sulfates, but the coordination varies with cations. For complexes with water, the four oxygen atoms in the sulfate group enable multiple binding geometries for the cations and for hydrogen bonding with water. Many of these geometries differ in free energy by only a small amount (1–2 kcal/mol), meaning there will be multiple binding configurations in bulk solution. In comparison to the optimal structures for binding to the carboxylate group, there is more variation for binding to the sulfate group as a function of cation type and the number of waters. Further, the polarization of the atoms is significant and varies among the sulfate oxygen atoms. The water oxygen charge is often larger than that of sulfate oxygen, which plays a role in the preference for monodentate ligand binding to cations in the presence of water.
Metasurfaces are highly effective at manipulating classical light in the linear regime; however, effectively controlling the polarization of nonclassical light generated from nonlinear resonant metasurfaces remains a challenge. Here, we present a solution by achieving polarization engineering of frequency-nondegenerate biphotons emitted via spontaneous parametric down-conversion in GaAs metasurfaces, utilizing quasi-bound states in the continuum (qBIC) resonances to enhance biphoton generation. Through comprehensive polarization tomography, we demonstrate that the emitted photons’ polarization directly reflects the qBIC mode’s far-field properties. Furthermore, we show that both the type of qBIC mode and the symmetry of the meta-atoms can be tailored to control each single-photon polarization state, and that the subsequent two-photon polarization states are nearly separable, offering potential applications in the heralded generation of single photons with adjustable polarization. This work provides a significant step toward utilizing metasurfaces to generate quantum light and engineer their polarization, a critical aspect for future quantum technologies.
The following details calibration of a material model for Al7075-T6511. This aluminum alloy is commonly used across a host of engineering applications. Owing to its widespread prevalence, there is great benefit in improving simulation predictions for this alloy. In the present effort, a calibration is performed of its elastic-plastic response accounting for both rate and temperature dependence. The calibration is informed by a series of tests that include specimens of different geometries tested at different rates and temperatures. All specimens are derived from the same barstock, 3.5 inches in diameter. The fitted model itself uses an anisotropic, Hill yield surface coupled with a Johnson-Cook hardening model. Failure predictions are had by means of a modified Wilkins failure criterion. Following calibration of the material model, a validation exercise is performed against platepuncture experiments. These experiments include multiple probe shapes, probe diameters, and plate thicknesses. The puncture experiments are replicated in simulation with mesh studies performed to assess uncertainty. Key quantities of interest, notably the absorbed energy up to failure, are compared between simulation and experiment providing a means to assess the suitability of the calibration in puncture simulations.
Controlling microstructural evolution in metallic additive manufacturing (AM) is difficult, especially in producing refined as-built grains instead of coarse, directional grains. Traditional solutions involve adding inoculants to AM feedstocks, but titanium (Ti) alloys cannot employ this approach without producing detrimental secondary phases. Ti-Cu (Ti-copper) alloys offer a solution through constitutional supercooling and/or solid state thermal cycling under AM conditions. This work analyzes a compositionally graded directed energy deposition (DED) Ti-Cu build, single-melt laser tracks, and dilatometric heat treatments to evaluate if, when, and by what mechanism(s) microstructural refinement occurs. Refinement by inoculation of unmelted powder particles was also considered. Constitutional supercooling produced no net microstructural refinement as any equiaxed dendrites which form are remelted with new deposition. This finding agreed with solidification modeling of powder bed fusion-laser beam (PBF-LB) and DED builds. Solid state thermal cycling refined microstructures only during ex-situ dilatometric heat treatments, suggesting build parameter optimization is needed to achieve refinement in-situ. Accidental heterogeneous nucleation on unmelted Ti powder, originating from the different thermophysical properties of Ti and Cu, provided the most significant microstructural refinement. This work systematically assesses the microstructural refinement mechanisms of Ti-Cu in AM builds and offers insights into microstructural control in eutectoid alloys.
We present a machine-learning strategy for finite element analysis of solid mechanics wherein we replace complex portions of a computational domain with a data-driven surrogate. In the proposed strategy, we decompose a computational domain into an “outer” coarse-scale domain that we resolve using a finite element method (FEM) and an “inner” fine-scale domain. We then develop a machine-learned (ML) model for the impact of the inner domain on the outer domain. In essence, for solid mechanics, our machine-learned surrogate performs static condensation of the inner domain degrees of freedom. This is achieved by learning the map from displacements on the inner-outer domain interface boundary to forces contributed by the inner domain to the outer domain on the same interface boundary. We consider two such mappings, one that directly maps from displacements to forces without constraints, and one that maps from displacements to forces by virtue of learning a symmetric positive semi-definite (SPSD) stiffness matrix. We demonstrate, in a simplified setting, that learning an SPSD stiffness matrix results in a coarse-scale problem that is well-posed with a unique solution. We present numerical experiments on several exemplars, ranging from finite deformations of a cube to finite deformations with contact of a fastener-bushing geometry. We demonstrate that enforcing an SPSD stiffness matrix drastically improves the robustness and accuracy of FEM–ML coupled simulations, and that the resulting methods can accurately characterize out-of-sample loading configurations with significant speedups over the standard FEM simulations.
Modern lens designs are capable of resolving greater than 10 gigapixels, while advances in camera frame-rate and hyperspectral imaging have made data acquisition rates of Terapixel/second a real possibility. The main bottlenecks preventing such high data-rate systems are power consumption and data storage. In this work, we show that analog photonic encoders could address this challenge, enabling high-speed image compression using orders-of-magnitude lower power than digital electronics. Our approach relies on a silicon-photonics front-end to compress raw image data, foregoing energy-intensive image conditioning and reducing data storage requirements. The compression scheme uses a passive disordered photonic structure to perform kernel-type random projections of the raw image data with minimal power consumption and low latency. A back-end neural network can then reconstruct the original images with structural similarity exceeding 90%. This scheme has the potential to process data streams exceeding Terapixel/second using less than 100 fJ/pixel, providing a path to ultra-high-resolution data and image acquisition systems.
Technical key performance indicators (KPIs) are important metrics used to assess and quantitatively summarize various aspects of photovoltaic (PV) systems, including long-term performance, economic viability, and carbon footprint. Herein, a group of experts of the International Energy Agency's Photovoltaic Power Systems Programme Task 13 collect and describ the most important technical KPIs used in the industry. Thereby, a set of best practices for reliably handling PV system data is presented and the impact of data quality and climatic variability on KPI calculation is investigated. The effective use of technical KPIs allows triggering data-driven and informed decisions to optimize PV systems and providing a comprehensive overview of how PV systems operate across different conditions and climates. With the worldwide growth of the PV industry, more companies operate/own PV systems in different regions, where the climatic and seasonal profiles differ. This requires context-aware evaluation of KPIs, or the judicious application of multiple KPIs, to ensure that each asset is evaluated correctly. Beyond that, there is untapped potential in the utilization of KPIs through geospatial mapping and extrapolation of fleet KPIs. This study demonstrates that the uncertainty in KPI estimation is not well understood and depends on data quality, climatic variability, and system configuration.
The spatial distribution of electric field due to an imposed electric charge density profile in an infinite slab of dielectric material is derived analytically by integrating Gauss's law. Various charge density distributions are considered, including exponential and power-law forms. The Maxwell stress tensor is used to compute a notional static stress in the material due to the charge density and its electric field. Characteristics of the electric field and stress distributions are computed for example cases in polyethylene, showing that field magnitudes exceeding the dielectric strength would be required in order to achieve a stress exceeding the ultimate tensile strength.
In response to gradual nanoindentation, the surface of muscovite mica deforms by sudden stochastic nanometer-scale displacement bursts. Here, the statistics of these displacement events are interpreted using a statistical model previously used to model earthquakes to understand how chemically reactive environments alter the surface properties of this material. We show that the statistics of nanoindentation displacement bursts in muscovite mica are tuned by chemomechanical weakening in a manner similar to how the statistics of model events are tuned by a mechanical weakening parameter that describes how easily system-spanning cracks can be nucleated. Because the predictions of this model are independent of any surface defects or structural details, these results suggest this simple model can be universally used to describe chemomechanical weakening in many systems prone to slip avalanches on a wide range of spatio-temporal scales.
In this study we present a replication method to determine surface roughness and to identify surface features when a sample cannot be directly analyzed by conventional techniques. As a demonstration, this method was applied to an unused spent nuclear fuel dry storage canister to determine variation across different surface features. In this study, an initial material down-selection was performed to determine the best molding agent and determined that non-modified Polytek PlatSil23-75 provided the most accurate representation of the surface while providing good usability. Other materials that were considered include Polygel Brush-On 35 polyurethane rubber (with and without Pol-ease 2300 release agent), Polytek PlatSil73-25 silicone rubber (with and without PlatThix thickening agent and Pol-ease 2300 release agent), and Express STD vinylpolysiloxane impression putty. The ability of PlatSil73-25 to create an accurate surface replica was evaluated by creating surface molds of several locations on surface roughness standards representing ISO grade surfaces N3, N5, N7, and N8. Overall, the molds were able to accurately reproduce the expected roughness average (Ra) values, but systematically over-estimated the peak-valley maximum roughness (Rz) values. Using a 3D printed sample cell, several locations across the stainless steel spent nuclear fuel canister were sampled to determine the surface roughness. These measurements provided information regarding variability in normal surface roughness across the canister as well as a detailed evaluation on specific surface features (e.g., welds, grind marks, etc.). The results of these measurements can support development of dry storage canister ageing management programs, as surface roughness is an important factor for surface dust deposition and accumulation. This method can be applied more broadly to different surfaces beyond stainless steel to provide rapid, accurate surface replications for analytical evaluation by profilometry.
Dielectric barrier discharges, particularly micro cavity plasma arrays, offer significant potential for plasma-catalytic research due to their ability to ignite plasma in direct contact with a catalytic surface, enabling the observation of plasma-surface interactions. A key factor in their application is the generation of reactive species, such as atomic oxygen, within the cavities. These species can interact with both the surface (e.g. for activation or cleaning) and the gas being treated (e.g. for oxidation). Given the central role of oxygen atoms in plasma catalysis and their use as a model for more complex species, this work investigates the transport of these atoms out of the cavities. Two-photon absorption laser-induced fluorescence spectroscopy with picosecond laser excitation is performed in the volume above the cavities. The results are compared with a basic diffusion model. The reactor operates with a He/O2 mixture at a flow rate of 1 slm and atmospheric pressure. Densities of up to 1016 cm−3 are measured near the surface. Time-dependent measurements show that, at a distance of 350 µm from the surface, a density equilibrium is reached within less than 3 ms of reactor operation. Decay times due to ozone formation after the reactor is turned off are on a similar scale. Spatially resolved measurements show that the oxygen density decreases exponentially from the surface but remains detectable up to approximately 1 mm above the surface, indicating significant application potential. Variations in the O2 admixture show a density maximum at 0.4%, confirming previous helium state enhanced actinometry measurements within the cavities.
Information security and computing, two critical technological challenges for post-digital computation, pose opposing requirements – security (encryption) requires a source of unpredictability, while computing generally requires predictability. Each of these contrasting requirements presently necessitates distinct conventional Si-based hardware units with power-hungry overheads. This work demonstrates Cu0.3Te0.7/HfO2 (‘CuTeHO’) ion-migration-driven memristors that satisfy the contrasting requirements. Under specific operating biases, CuTeHO memristors generate truly random and physically unclonable functions, while under other biases, they perform universal Boolean logic. Using these computing primitives, this work experimentally demonstrates a single system that performs cryptographic key generation, universal Boolean logic operations, and encryption/decryption. Circuit-based calculations reveal the energy and latency advantages of the CuTeHO memristors in these operations. This work illustrates the functional flexibility of memristors in implementing operations with varying component-level requirements.
Using coarse graining, the upscaled mechanical properties of a solid with small scale heterogeneities are derived. The method maps internal forces at the small scale onto peridynamic bond forces in the coarse grained mesh. These upscaled bond forces are used to calibrate a peridynamic material model with position-dependent parameters. These parameters incorporate mesoscale variations in the statistics of the small scale system. The upscaled peridynamic model can have a much coarser discretization than the original small scale model, allowing larger scale simulations to be performed efficiently. The convergence properties of the method are investigated for representative random microstructures. A bond breakage criterion for the upscaled peridynamic material model is also demonstrated.
Additive manufacturing (AM) technology, specifically 3D printing, holds great promise for in-orbit manufacturing. In-space printing can significantly reduce the mass, cost, and risk of long-term space exploration by enabling replacement parts to be made as needed and reducing dependence on Earth. However, printing in a zero-gravity environment poses challenges due to the absence of a rigid ground for the print platform, which can result in vibrational and rotational forces that may impact printing integrity. To address this issue, this paper proposes a novel linear magnetic position tracking algorithm, named Navigation Integrating Magnets By Linear Estimation (NIMBLE), for dynamic vibration compensation during 3D printing of truss structures in space. Compared to the most commonly used nonlinear optimization method, the NIMBLE algorithm is more than two orders of magnitude faster. With only a single 3-axis magnet sensor and a small NdFeB magnet, the NIMBLE algorithm provides a simple and easily implemented tracking solution for in-orbit 3D printing.
The increasing potential of infrastructure disruptions calls for high-quality infrastructure models to be used in resilience analysis and decision making. Unfortunately, many utilities and communities do not have access to accurate and detailed models due to a lack of data and resources. Furthermore, security restrictions on sharing infrastructure models present roadblocks to research, analysis, and decision making. Recent advances in the development of synthetic water distribution models provide a potential solution to this problem. There is an opportunity to improve these methods by leveraging incomplete pipe datasets to aid synthetic network generation. To address this gap, we developed a methodology for synthetic network generation that incorporates partial pipe data using a modification of the minimum cost flow algorithm for network generation and pipe sizing. This methodology demonstrates how partial pipe data can be leveraged to improve site-specific synthetic network generation. For the study area of Mayagüez, Puerto Rico, a synthetic model generated using 50% of real pipe data matches the pressure of the validation system with an average error of 23.5 m of head, which improves upon the average error of 31.6 m of head produced by a synthetic model generated using no data of the real pipes. Additionally, synthetic networks are shown to replicate the pressure response under a disruption scenario of the validation network, suggesting potential use in resilience analysis.
Accurate self-motion estimation is critical for various navigational tasks in mobile robotics. Optic flow provides a means to estimate self-motion using a camera sensor and is particularly valuable in GPS- and radio-denied environments. The present study investigates the influence of different activation functions—ReLU, leaky ReLU, GELU, and Mish—on the accuracy, robustness, and encoding properties of convolutional neural networks (CNNs) and multi-layer perceptrons (MLPs) trained to estimate self-motion from optic flow. Our results demonstrate that networks with ReLU and leaky ReLU activation functions not only achieved superior accuracy in self-motion estimation from novel optic flow patterns but also exhibited greater robustness under challenging conditions. The advantages offered by ReLU and leaky ReLU may stem from their ability to induce sparser representations than GELU and Mish do. Our work characterizes the encoding of optic flow in neural networks and highlights how the sparseness induced by ReLU may enhance robust and accurate self-motion estimation from optic flow.
The current present in a galvanic couple can define its resistance or susceptibility to corrosion. However, as the current is dependent upon environmental, material, and geometrical parameters it is experimentally costly to measure. To reduce these costs, Finite Element (FE) simulations can be used to assess the cathodic current but also require experimental inputs to define boundary conditions. Due to these challenges, it is crucial to accelerate predictions and accurately predict the current output for different environments and geometries representative of in-service conditions. Machine learned surrogate models provides a means to accelerate corrosion predictions. However, a one-time cost is incurred in procuring the simulation and experimental dataset necessary to calibrate the surrogate model. Therefore, an active learning protocol is developed through calibration of a low-cost surrogate model for the cathodic current of an exemplar galvanic couple (AA7075-SS304) as a function of environmental and geometric parameters. The surrogate model is calibrated on a dataset of FE simulations, and calculates an acquisition function that identifies specific additional inputs with the maximum potential to improve the current predictions. This is accomplished through a staggered workflow that not only improves and refines prediction, but identifies the points at which the most information is gained, thus enabling expansion to a larger parameter space. The protocols developed and demonstrated in this work provide a powerful tool for screening various forms of corrosion under in-service conditions.
Material Testing 2.0 (MT2.0) is a paradigm that advocates for the use of rich, full-field data, such as from digital image correlation and infrared thermography, for material identification. By employing heterogeneous, multi-axial data in conjunction with sophisticated inverse calibration techniques such as finite element model updating and the virtual fields method, MT2.0 aims to reduce the number of specimens needed for material identification and to increase confidence in the calibration results. To support continued development, improvement, and validation of such inverse methods—specifically for rate-dependent, temperature-dependent, and anisotropic metal plasticity models—we provide here a thorough experimental data set for 304L stainless steel sheet metal. The data set includes full-field displacement, strain, and temperature data for seven unique specimen geometries tested at different strain rates and in different material orientations. Commensurate extensometer strain data from tensile dog bones is provided as well for comparison. We believe this complete data set will be a valuable contribution to the experimental and computational mechanics communities, supporting continued advances in material identification methods.
Composite materials with different microstructural material symmetries are common in engineering applications where grain structure, alloying and particle/fiber packing are optimized via controlled manufacturing. In fact these microstructural tunings can be done throughout a part to achieve functional gradation and optimization at a structural level. To predict the performance of particular microstructural configuration and thereby overall performance, constitutive models of materials with microstructure are needed. In this work we provide neural network architectures that provide effective homogenization models of materials with anisotropic components. These models satisfy equivariance and material symmetry principles inherently through a combination of equivariant and tensor basis operations. We demonstrate them on datasets of stochastic volume elements with different textures and phases where the material undergoes elastic and plastic deformation, and show that the these network architectures provide significant performance improvements.
