An Acquisition Parameter Study for Machine-Learning-Enabled Electron Backscatter Diffraction.

Methods within the domain of artificial intelligence are gaining traction for solving a range of materials science objectives, notably the use of deep neural networks for computer vision for the analysis of electron diffraction patterns. An important component of deploying these models is an understanding of the performance as experimental diffraction conditions are varied. This knowledge can inspire confidence in the classifications over a range of operating conditions and identify where performance is degraded. Elucidating the relative impact of each parameter will suggest the most important parameters to vary during the collection of future training data. Knowing which data collection efforts to prioritize is of concern given the time required to collect or simulate vast libraries of diffraction patterns for a wide variety of materials without considering varying any parameters. In this work, five parameters, frame averaging, detector tilt, sample-to-detector distance, accelerating voltage, and pattern resolution, essential to electron diffraction are individually varied during the collection of electron backscatter diffraction patterns to explore the effect on the classifications produced by a deep neural network trained from diffraction patterns captured using a fixed set of parameters. The model is shown to be resilient to nearly all the individual changes examined here.

Deep Neural Network Enabled Space Group Identification in EBSD.

Electron backscatter diffraction (EBSD) is one of the primary tools in materials development and analysis. The technique can perform simultaneous analyses at multiple length scales, providing local sub-micron information mapped globally to centimeter scale. Recently, a series of technological revolutions simultaneously increased diffraction pattern quality and collection rate. After collection, current EBSD pattern indexing techniques (whether Hough-based or dictionary pattern matching based) are capable of reliably differentiating between a "user selected" set of phases, if those phases contain sufficiently different crystal structures. EBSD is currently less well suited for the problem of phase identification where the phases in the sample are unknown. A pattern analysis technique capable of phase identification, utilizing the information-rich diffraction patterns potentially coupled with other data, such as EDS-derived chemistry, would enable EBSD to become a high-throughput technique replacing many slower (X-ray diffraction) or more expensive (neutron diffraction) methods. We utilize a machine learning technique to develop a general methodology for the space group classification of diffraction patterns; this is demonstrated within the $\lpar 4/m\comma \;\bar{3}\comma \;\;2/m\rpar$ point group. We evaluate the machine learning algorithm's performance in real-world situations using materials outside the training set, simultaneously elucidating the role of atomic scattering factors, orientation, and pattern quality on classification accuracy.

Phase Mapping in EBSD Using Convolutional Neural Networks.

The emergence of commercial electron backscatter diffraction (EBSD) equipment ushered in an era of information rich maps produced by determining the orientation of user-selected crystal structures. Since then, a technological revolution has occurred in the quality, rate detection, and analysis of these diffractions patterns. The next revolution in EBSD is the ability to directly utilize the information rich diffraction patterns in a high-throughput manner. Aided by machine learning techniques, this new methodology is, as demonstrated herein, capable of accurately separating phases in a material by crystal symmetry, chemistry, and even lattice parameters with fewer human decisions. This work is the first demonstration of such capabilities and addresses many of the major challenges faced in modern EBSD. Diffraction patterns are collected from a variety of samples, and a convolutional neural network, a type of machine learning algorithm, is trained to autonomously recognize the subtle differences in the diffraction patterns and output phase maps of the material. This study offers a path to machine learning coupled phase mapping as databases of EBSD patterns encompass an increasing number of the possible space groups, chemistry changes, and lattice parameter variations.

Automated Reconstruction of Spherical Kikuchi Maps.

An automated approach to fully reconstruct spherical Kikuchi maps from experimentally collected electron backscatter diffraction patterns and overlay each pattern onto its corresponding position on a simulated Kikuchi sphere is presented in this study. This work demonstrates the feasibility of warping any Kikuchi pattern onto its corresponding location of a simulated Kikuchi sphere and reconstructing a spherical Kikuchi map of a known phase based on any set of experimental patterns. This method consists of the following steps after pattern collection: (1) pattern selection based on multiple threshold values; (2) extraction of multiple scan parameters and phase information; (3) generation of a kinematically simulated Kikuchi sphere as the "skeleton" of the spherical Kikuchi map; and (4) overlaying the inverse gnomonic projection of multiple selected patterns after appropriate pattern center calibration and refinement. The proposed method is the first automated approach to reconstructing spherical Kikuchi maps from experimental Kikuchi patterns. It potentially enables more accurate orientation calculation, new pattern center refinement methods, improved dictionary-based pattern matching, and phase identification in the future.

Micromotor-based energy generation.

A micromotor-based strategy for energy generation, utilizing the conversion of liquid-phase hydrogen to usable hydrogen gas (H2), is described. The new motion-based H2-generation concept relies on the movement of Pt-black/Ti Janus microparticle motors in a solution of sodium borohydride (NaBH4) fuel. This is the first report of using NaBH4 for powering micromotors. The autonomous motion of these catalytic micromotors, as well as their bubble generation, leads to enhanced mixing and transport of NaBH4 towards the Pt-black catalytic surface (compared to static microparticles or films), and hence to a substantially faster rate of H2 production. The practical utility of these micromotors is illustrated by powering a hydrogen-oxygen fuel cell car by an on-board motion-based hydrogen and oxygen generation. The new micromotor approach paves the way for the development of efficient on-site energy generation for powering external devices or meeting growing demands on the energy grid.

Micromotor-Based Biomimetic Carbon Dioxide Sequestration: Towards Mobile Microscrubbers.

We describe a mobile CO2 scrubbing platform that offers a greatly accelerated biomimetic sequestration based on a self-propelled carbonic anhydrase (CA) functionalized micromotor. The CO2 hydration capability of CA is coupled with the rapid movement of catalytic micromotors, and along with the corresponding fluid dynamics, results in a highly efficient mobile CO2 scrubbing microsystem. The continuous movement of CA and enhanced mass transport of the CO2 substrate lead to significant improvements in the sequestration efficiency and speed over stationary immobilized or free CA platforms. This system is a promising approach to rapid and enhanced CO2 sequestration platforms for addressing growing concerns over the buildup of greenhouse gas.

Fully loaded micromotors for combinatorial delivery and autonomous release of cargoes.

Integrating functional self-propelled Zinc micromotors are created by coup-ling electrodeposition with hard dual-templating synthesis. The micromotors concurrently possess four robust functions including a remarkably high loading capacity, combinatorial delivery of cargoes, autonomous release of encapsulated payloads, and self-destruction. This concept could be expanded to simultaneous encapsulation of various payloads for different functionalities such as therapy, diagnostics, and imaging.

Valence electron concentration as key parameter to control the fracture resistance of refractory high-entropy carbides.

Although high-entropy carbides (HECs) have hardness often superior to that of parent compounds, their brittleness-a problem shared with most ceramics-has severely limited their reliability. Refractory HECs in particular are attracting considerable interest due to their unique combination of mechanical and physical properties, tunable over a vast compositional space. Here, combining statistics of crack formation in bulk specimens subject to mild, moderate, and severe nanoindentation loading with ab initio molecular dynamics simulations of alloys under tension, we show that the resistance to fracture of cubic-B1 HECs correlates with their valence electron concentration (VEC). Electronic structure analyses show that VEC ≳ 9.4 electrons per formula unit enhances alloy fracture resistance due to a facile rehybridization of electronic metallic states, which activates transformation plasticity at the yield point. Our work demonstrates a reliable strategy for computationally guided and rule-based (i.e., VEC) engineering of deformation mechanisms in high entropy, solid solution, and doped ceramics.

Efficient few-shot machine learning for classification of EBSD patterns.

Deep learning is quickly becoming a standard approach to solving a range of materials science objectives, particularly in the field of computer vision. However, labeled datasets large enough to train neural networks from scratch can be challenging to collect. One approach to accelerating the training of deep learning models such as convolutional neural networks is the transfer of weights from models trained on unrelated image classification problems, commonly referred to as transfer learning. The powerful feature extractors learned previously can potentially be fine-tuned for a new classification problem without hindering performance. Transfer learning can also improve the results of training a model using a small amount of data, known as few-shot learning. Herein, we test the effectiveness of a few-shot transfer learning approach for the classification of electron backscatter diffraction (EBSD) pattern images to six space groups within the [Formula: see text] point group. Training history and performance metrics are compared with a model of the same architecture trained from scratch. In an effort to make this approach more explainable, visualization of filters, activation maps, and Shapley values are utilized to provide insight into the model's operations. The applicability to real-world phase identification and differentiation is demonstrated using dual phase materials that are challenging to analyze with traditional methods.

Novel remapping approach for HR-EBSD based on demons registration.

In this study, the possibility of utilizing a computer vision algorithm, i.e., demons registration, to accurately remap electron backscatter diffraction patterns for high resolution electron backscatter diffraction (HR-EBSD) applications is presented. First, the angular resolution of demons registration is demonstrated to be lower than the conventional cross-correlation based method, particularly at misorientation angles >0.157 rad. In addition, GPU acceleration has been applied to significantly boost the speed of iterative registration between a pair of patterns with 0.175 rad misorientation to under 1 s. Second, demons registration is implemented as a first-pass remapping, followed by a second pass cross-correlation method, which results in angular resolution of ~0.5 × 10-4 rad, a phantom stress value of ~35 MPa and phantom strain of ~2 × 10-4, on dynamically simulated patterns, without the need of implementing robust fitting or iterative remapping. Lastly, the new remapping method is applied to a large experimental dataset collected from an as-built additively-manufactured Inconel 625 cube, which shows significant residual stresses built-up near the large columnar grain region and regularly arranged GND structures.

Crystal symmetry determination in electron diffraction using machine learning.

Electron backscatter diffraction (EBSD) is one of the primary tools for crystal structure determination. However, this method requires human input to select potential phases for Hough-based or dictionary pattern matching and is not well suited for phase identification. Automated phase identification is the first step in making EBSD into a high-throughput technique. We used a machine learning-based approach and developed a general methodology for rapid and autonomous identification of the crystal symmetry from EBSD patterns. We evaluated our algorithm with diffraction patterns from materials outside the training set. The neural network assigned importance to the same symmetry features that a crystallographer would use for structure identification.

Nanomotors responsive to nerve-agent vapor plumes.

Enzyme-powered nanomotors responsive to the presence of nerve agents in the surrounding atmosphere are employed for remote detection of chemical vapor threats. Distinct changes in the propulsion behavior, associated with the partition of the sarin simulant diethyl chlorophosphate (DCP), offer reliable and rapid detection of the nerve-agent vapor threat.

Micromotor-based on-off fluorescence detection of sarin and soman simulants.

Self-propelled micromotor-based fluorescent "On-Off" detection of nerve agents is described. The motion-based assay utilizes Si/Pt Janus micromotors coated with fluoresceinamine toward real-time "on-the-fly" field detection of sarin and soman simulants.

Water-driven micromotors for rapid photocatalytic degradation of biological and chemical warfare agents.

Threats of chemical and biological warfare agents (CBWA) represent a serious global concern and require rapid and efficient neutralization methods. We present a highly effective micromotor strategy for photocatalytic degradation of CBWA based on light-activated TiO2/Au/Mg microspheres that propel autonomously in natural water and obviate the need for external fuel, decontaminating reagent, or mechanical agitation. The activated TiO2/Au/Mg micromotors generate highly reactive oxygen species responsible for the efficient destruction of the cell membranes of the anthrax simulant Bacillus globigii spore, as well as rapid and complete in situ mineralization of the highly persistent organophosphate nerve agents into nonharmful products. The water-driven propulsion of the TiO2/Au/Mg micromotors facilitates efficient fluid transport and dispersion of the photogenerated reactive oxidative species and their interaction with the CBWA. Coupling of the photocatalytic surface of the micromotors and their autonomous water-driven propulsion thus leads to a reagent-free operation which holds a considerable promise for diverse "green" defense and environmental applications.

Tunable hierarchical macro/mesoporous gold microwires fabricated by dual-templating and dealloying processes.

Tailor-made highly ordered macro/mesoporous hierarchical metal architectures have been created by combining sphere lithography, membrane template electrodeposition and alloy-etching processes. The new double-template preparation route involves the electrodeposition of Au/Ag alloy within the interstitial (void) spaces of polystyrene (PS) microspheres which are closely packed within the micropores of a polycarbonate membrane (PC), followed by dealloying of the Ag component and dissolution of the microsphere and membrane templates. The net results of combining such sphere lithography and silver etching is the creation of highly regular three-dimensional macro/mesoporous gold architecture with well-controlled sizes and shapes. The morphology and porosity of the new hierarchical porous structures can be tailored by controlling the preparation conditions, such as the composition of the metal mixture plating solution, the size of the microspheres template, or the dealloying time. Such tunable macro/mesoporous hierarchical structures offer control of the electrochemical reactivity and of the fuel mass transport, as illustrated for the enhanced oxygen reduction reaction (ORR) and hydrogen-peroxide detection. The new double templated electrodeposition method provides an attractive route for preparing highly controllable multiscale porous materials and diverse morphologies based on different materials and hence holds considerable promise for designing electrocatalytic or bioelectrocatalytic surfaces for a variety sensing and energy applications.

Self-propelled chemically-powered plant-tissue biomotors.

Self-propelled biocatalytic motors based on plant tissues are described. The tissue motors rely on their rich catalase activity towards biocatalytic decomposition of the H2O2 fuel and generation of the bubble thrust. These biomotors obviate the need for pure enzymes, and offer a remarkably low cost, good lifetime and thermostability.