A deep learning approach for complex microstructure inference.

Automated, reliable, and objective microstructure inference from micrographs is essential for a comprehensive understanding of process-microstructure-property relations and tailored materials development. However, such inference, with the increasing complexity of microstructures, requires advanced segmentation methodologies. While deep learning offers new opportunities, an intuition about the required data quality/quantity and a methodological guideline for microstructure quantification is still missing. This, along with deep learning's seemingly intransparent decision-making process, hampers its breakthrough in this field. We apply a multidisciplinary deep learning approach, devoting equal attention to specimen preparation and imaging, and train distinct U-Net architectures with 30-50 micrographs of different imaging modalities and electron backscatter diffraction-informed annotations. On the challenging task of lath-bainite segmentation in complex-phase steel, we achieve accuracies of 90% rivaling expert segmentations. Further, we discuss the impact of image context, pre-training with domain-extrinsic data, and data augmentation. Network visualization techniques demonstrate plausible model decisions based on grain boundary morphology.

Machine-Learning Microstructure for Inverse Material Design.

Metallurgy and material design have thousands of years' history and have played a critical role in the civilization process of humankind. The traditional trial-and-error method has been unprecedentedly challenged in the modern era when the number of components and phases in novel alloys keeps increasing, with high-entropy alloys as the representative. New opportunities emerge for alloy design in the artificial intelligence era. Here, a successful machine-learning (ML) method has been developed to identify the microstructure images with eye-challenging morphology for a number of martensitic and ferritic steels. Assisted by it, a new neural-network method is proposed for the inverse design of alloys with 20 components, which can accelerate the design process based on microstructure. The method is also readily applied to other material systems given sufficient microstructure images. This work lays the foundation for inverse alloy design based on microstructure images with extremely similar features.

A transfer learning approach for improved classification of carbon nanomaterials from TEM images.

The extensive use of carbon nanomaterials such as carbon nanotubes/nanofibers (CNTs/CNFs) in industrial settings has raised concerns over the potential health risks associated with occupational exposure to these materials. These exposures are commonly in the form of CNT/CNF-containing aerosols, resulting in a need for a reliable structure classification protocol to perform meaningful exposure assessments. However, airborne carbonaceous nanomaterials are very likely to form mixtures of individual nano-sized particles and micron-sized agglomerates with complex structures and irregular shapes, making structure identification and classification extremely difficult. While manual classification from transmission electron microscopy (TEM) images is widely used, it is time-consuming due to the lack of automation tools for structure identification. In the present study, we applied a convolutional neural network (CNN) based machine learning and computer vision method to recognize and classify airborne CNT/CNF particles from TEM images. We introduced a transfer learning approach to represent images by hypercolumn vectors, which were clustered via K-means and processed into a Vector of Locally Aggregated Descriptors (VLAD) representation to train a softmax classifier with the gradient boosting algorithm. This method achieved 90.9% accuracy on the classification of a 4-class dataset and 84.5% accuracy on a more complex 8-class dataset. The developed model established a framework to automatically detect and classify complex carbon nanostructures with potential applications that extend to the automated structural classification for other nanomaterials.

High Throughput Quantitative Metallography for Complex Microstructures Using Deep Learning: A Case Study in Ultrahigh Carbon Steel.

We apply a deep convolutional neural network segmentation model to enable novel automated microstructure segmentation applications for complex microstructures typically evaluated manually and subjectively. We explore two microstructure segmentation tasks in an openly available ultrahigh carbon steel microstructure dataset: segmenting cementite particles in the spheroidized matrix, and segmenting larger fields of view featuring grain boundary carbide, spheroidized particle matrix, particle-free grain boundary denuded zone, and Widmanstätten cementite. We also demonstrate how to combine these data-driven microstructure segmentation models to obtain empirical cementite particle size and denuded zone width distributions from more complex micrographs containing multiple microconstituents. The full annotated dataset is available on materialsdata.nist.gov.

A dataset of synthetic hexagonal close packed 3D polycrystalline microstructures, grain-wise microstructural descriptors and grain averaged stress fields under uniaxial tensile deformation for two sets of constitutive parameters.

This data article presents a data set comprised of 54 synthetic 3D equiaxed polycrystalline microstructures, the microstructural descriptors for each grain and the stress fields resulting from two sets of crystal plasticity simulations mimicking uniaxial tensile deformation to a total strain of 2%. This is related to the research article entitled "Applied Machine Learning to predict stress hotspots II: Hexagonal Close Packed Materials" (Mangal and Holm, 2018). The microstructures were created using an open source Dream.3D software tool and the crystal plasticity simulations were carried out using elasto-viscoplastic fast Fourier transform (EVPFFT) method. Eight different kinds of HCP textures are represented with stochastically different microstructures with varying texture intensity for each texture kind. For each texture kind, between six and nine stochastically different microstructures with varying texture intensity (measured by multiples of random density (MRD)) are created. This dataset is freely available in two Mendeley Data archives "Synthetic HCP 3D polycrystalline microstructures with grain-wise microstructural descriptors and stress fields under uniaxial tensile deformation: Part One" and "Synthetic HCP 3D polycrystalline microstructures with grain-wise microstructural descriptors and stress fields under uniaxial tensile deformation: Part Two" located at http://dx.doi.org/10.17632/kt8hfg4t2p.1 and http://dx.doi.org/10.17632/nsfn6tw295.1 respectively for any academic, educational, or research purposes.

A dataset of synthetic face centered cubic 3D polycrystalline microstructures, grain-wise microstructural descriptors and grain averaged stress fields under uniaxial tensile deformation.

This data article presents a data set comprised of 36 synthetic 3D equiaxed polycrystalline microstructures, the microstructural descriptors for each grain, and the stress and strain fields resulting from crystal plasticity simulations mimicking uniaxial tensile deformation to a total strain of 4%. This is related to the research article entitled "Applied Machine Learning to predict stress hotspots I: Face Centered Cubic Materials" (Mangal and Holm, 2018) [1]. The microstructures were created using an open source Dream.3D software tool, and the crystal plasticity simulations were carried out using the elasto-viscoplastic fast Fourier transform (EVPFFT) method. Six different kinds of FCC textures are represented with six stochastically different microstructures with varying texture intensity for each texture kind. This dataset is freely available in a Mendeley Data archive "A dataset of synthetic face centered cubic 3D polycrystalline microstructures, grain-wise microstructural descriptors and grain averaged stress fields under uniaxial tensile deformation" located at 〈http://dx.doi.org/10.17632/ss75fdg5dg.1〉 for any academic, educational, or research purposes.

Corrigendum to "A large dataset of synthetic SEM images of powder materials and their ground truth 3D structures" [Data Brief 9 (2016) 727-731].

[This corrects the article DOI: 10.1016/j.dib.2016.10.011.].

A large dataset of synthetic SEM images of powder materials and their ground truth 3D structures.

This data article presents a data set comprised of 2048 synthetic scanning electron microscope (SEM) images of powder materials and descriptions of the corresponding 3D structures that they represent. These images were created using open source rendering software, and the generating scripts are included with the data set. Eight particle size distributions are represented with 256 independent images from each. The particle size distributions are relatively similar to each other, so that the dataset offers a useful benchmark to assess the fidelity of image analysis techniques. The characteristics of the PSDs and the resulting images are described and analyzed in more detail in the research article "Characterizing powder materials using keypoint-based computer vision methods" (B.L. DeCost, E.A. Holm, 2016) [1]. These data are freely available in a Mendeley Data archive "A large dataset of synthetic SEM images of powder materials and their ground truth 3D structures" (B.L. DeCost, E.A. Holm, 2016) located at http://dx.doi.org/10.17632/tj4syyj9mr.1[2] for any academic, educational, or research purposes.

How grain growth stops: a mechanism for grain-growth stagnation in pure materials.

The thermodynamic equilibrium state of crystalline materials is a single crystal; however, polycrystalline grain growth almost always stops before this state is reached. Although typically attributed to solute drag, grain-growth stagnation occurs, even in high-purity materials. Recent studies indicate that grain boundaries undergo thermal roughening associated with an abrupt mobility change, so that at typical annealing temperatures, polycrystals will contain both smooth (slow) and rough (fast) boundaries. Mesoscale grain-growth models, validated by large-scale polycrystalline molecular dynamics simulations, show that even small fractions of smooth, slow boundaries can stop grain growth. We conclude that grain-boundary roughening provides an alternate stagnation mechanism that applies even to high-purity materials.

Computing the mobility of grain boundaries.

As current experimental and simulation methods cannot determine the mobility of flat boundaries across the large misorientation phase space, we have developed a computational method for imposing an artificial driving force on boundaries. In a molecular dynamics simulation, this allows us to go beyond the inherent timescale restrictions of the technique and induce non-negligible motion in flat boundaries of arbitrary misorientation. For different series of symmetric boundaries, we find both expected and unexpected results. In general, mobility increases as the grain boundary plane deviates from (111), but high-coincidence and low-angle boundaries represent special cases. These results agree with and enrich experimental observations.

Ultrasonido de alta resolucion (15 MHz) en el estudio de la u¤a psoriatica / High resolution ultrasound in the study of psoriatic nail

Antecedentes. La Psoriasis es una enfermedad frecuente en la practica dermatológica. En EE.UU., 7 millones de personas (2-3por ciento) la padecen. El 5 por ciento de los pacientes con psoriasis ungueal no tienen manifestaciones cutáneas. Entre 10-20 por ciento de los pacientes con psoriasis cutanea tienen artritis psoriatica. De los pacientes con artritis psoriatica, el 53-86 por ciento presentan compromiso ungueal. Objetivo. Estudiar los cambios visibles al ultrasonido de alta resolución, en la u¤a de los pacientes psoriaticos. Material y metodo. Se utilizo equipo de ultrasonido Philips ATL 5000 equipado con SonoCT, XRES, campo de visión extendida y transductores de 15-7 MHz y 12-5 MHz. Se estudiaron 15 pacientes, 9 controles normales y 6 psoriaticos con compromiso ungueal. Mujeres n=9 y hombres n=6, promedio de edad 47,7 a¤os. Se exploraron las 2 manos en todos los pacientes (150 u¤as) y se midieron en cada paciente las distancias entre ambas placas ungueales como también entre la placa ungueal ventral y el margen óseo de la falange distal en la u¤a del dedo indice derecho. Resultados. Se describe la anatomia ultrasonografica normal de la u¤a y cuatro patrones de alteraciones visibles en el compromiso por psoriasis que incluyen: compromiso focal hipere-cogenico de la placa ventral sin compromiso de placa dorsal, perdida de definición de la placa ventral con placa dorsal indemne, ondulación de ambas placas y perdida de definición de ambas placas. Se encontró una diferencia significativa (p= 0.02) entre el promedio de distancia placa ventral ungueal y margen óseo de la falange distal para u¤as normales (1,5 mm) y para psoriasis (3,0 mms). Conclusión: El ultrasonido de alta resolución permite un estudio no invasivo de la u¤a, describiendo su anatomia y cambios morfológicos en pacientes normales y psoriaticos con compromiso ungueal. Ello puede ser útil en la monitorización de los efectos del tratamiento de psoriasis en sus distintas mani-festaciones, ya que existen cambios imperceptibles en su magnitud y ubicación para el clinico, que pueden observarse sin esperar el recambio ungueal total. Las diferencias de distancia entre las placas de la u¤a, incluyendo la distancia con el margen óseo de la falange distal, en los pacientes normales y psoriaticos, en nuestro conocimiento, no han sido previamente publicadas.