Characterization of Minerals, Metals and Materials: 3D Characterization
Sponsored by: TMS Extraction and Processing Division, TMS: Materials Characterization Committee
Program Organizers: Jian Li, CanmetMATERIALS; Mingming Zhang, Baowu Ouyeel Co. Ltd; Bowen Li, Michigan Technological University; Sergio Monteiro, Instituto Militar de Engenharia; Shadia Ikhmayies, The University of Jordan; Yunus Kalay, Middle East Technical University; Jiann-Yang Hwang, Michigan Technological University; Juan Escobedo-Diaz, University of New South Wales; John Carpenter, Los Alamos National Laboratory; Andrew Brown, Devcom Arl Army Research Office

Tuesday 8:30 AM
February 25, 2020
Room: Theater A-5
Location: San Diego Convention Ctr

Session Chair: Yunus Kalay, Middle East Technical University; Tomoko Sano, U.S. Army Research Laboratory

8:30 AM
From Fundamental Research to Engineered Components: Application to 3D Materials Science: Jonathan Madison¹; Thomas Ivanoff¹; Alex Hickman¹; ¹Sandia National Laboratories
    Three-dimensional materials science (3DMS) has become a burgeoning field providing unique and novel insights to a myriad of materials issues ranging from kinetics of solidification, evaluation of complex microstructures, in-situ observation of response, and even post-mortem failure analysis. While largely exploratory in the early 2000s, two decades of development have provided much in advancing experimental hardware, software packages and overall computational power. These developments have reduced both the barrier to entry and many previous limitations. Using a RoboMET.3D, Sandia National Laboratories has advanced much of its 3DMS investigations from fundamental research to pressing concerns in engineered components. To illustrate, examples of new insights, failure mode discoveries and lot-to-lot variations in singular material systems and engineered multi-material assemblies, as obtained via three-dimensional reconstruction, will be provided. Additionally, a few observations regarding error estimation and the impact of image segmentation decisions on final results will also be highlighted.

8:50 AM
In-Situ X-ray Tomography of Vapor Phase Alloying of Ni Wires via Pack Titanization: Arun Bhattacharjee¹; Ashley Paz y Puente¹; Dinc Erdeniz²; David Dunand³; ¹University of Cincinnati; ²Marquette University; ³Northwestern University
    Pack cementation is a type of chemical vapor deposition process where the substrate is buried in a powder mixture containing a halide activator, a source metal to be deposited, and an inert filler. After depositing a sufficient amount of titanium on Ni wires (50-100 痠 diameter), and subsequent homogenization, shape memory NiTi microtubes can be created by taking advantage of the radial symmetry and spatial confinement and harnessing the Kirkendall effect. However, to be able to tailor the Kirkendall pores, their formation and evolution mechanisms must be fully understood. Because the pores start developing during the vapor phase deposition process itself, in situ X-ray tomography experiments were used to provide a 4D spatio-temporal visualization of the samples throughout the pack titanization and homogenization heat treatments. Using this tomography based approach in conjunction with traditional ex situ metallography has proven to be an effective method to perform such rapid diffusion studies.

9:10 AM
Introducing 3D-LIBS, a Powerful Rapid Chemical Mapping Tool for Trace Elements in Complex Materials: Carys Cook¹; Rajiv Soman¹; Karol Putyera¹; ¹EAG Eurofins
     Characterizing the distribution of trace elements in many advanced solid materials can be analytically challenging and time-consuming. 3D-LIBS (Laser Induced Breakdown Spectroscopy) is a powerful, rapid, quasi-simultaneous, and indirect micro-sampling technique that can identify and map any element (including light elements and gases) to 1 ppm, in most solid materials. Sample surfaces can be mapped over a variety of spatial scales, from 10’s of microns to centimeters, with depth resolutions of 100’s of nanometers to 10’s of microns. We investigate a range of materials for trace element distributions in film/coating/substrate interfaces or treated/exposed surfaces. Examples include, i) H, O, B, S and Ti in Ni superalloys, ii) O, H and alkali elements in high purity Al metal, iii) refractories in high refractive glasses, and iv) Li and metal distribution in Li battery electrodes.

9:30 AM
Modeling the Anisotropic Mechanical Properties of Fused Deposition Modeling ABS using an Artificial Neural Network - Part 2: Brian Kessler¹; Sarah Gladding¹; Aric Harper¹; ¹Colorado Mesa University
    Fused deposition modeling (FDM), a common known 3D printing method, has been used for some time as a tool for rapid product prototyping. As technology has improved, both in FDM printers and in filament material, demand has grown for FDM produced end-user components. The FDM process produces highly anisotropic parts, with mechanical properties varying based on layer height, infill density, air gap and printing pattern. To produce viable end-user components, it is necessary to be able to accurately predict the mechanical properties of a component given the above variables.

9:50 AM
Mapping Grain Morphology and Orientations by Laboratory Diffraction Contrast Tomography: Hrishi Bale¹; Jun Sun²; Jette Oddershede²; Erik Lauridsen²; ¹Carl Zeiss Microscopy Inc.; ²Xnovo Technology ApS
     Recent developments in lab-based X-ray diffraction contrast tomography (LabDCT) technique have extended its capabilities to include full reconstruction of the 3D grain structure, including both grain morphology and crystallographic orientation. With both morphology and orientation, it is possible to extract the full five parameters describing the grain boundary characteristics, opening new possibilities for statistical studies of grain boundaries. The 3D crystallographic imaging capabilities of LabDCT complements the structural data obtained by traditional absorption-contrast tomography (ACT), e.g. cracks, voids and inclusions, and together the combination of ACT and LabDCT provides unprecedented insight. We present a selection of LabDCT results with particular emphasis on its non-destructive operation and discuss boundary conditions of the current implementation, point to the future of the technique and discuss ways in which the results can be correlatively coupled to both related characterization techniques and microstructural modelling for better understanding of materials structure evolution in 3D.

10:10 AM Break

10:25 AM
A New Characterization Tool for 3D Orientation Microscopy at Mesostructure Length Scales : Thomas Ales¹; Peter Collins¹; ¹Iowa State University
    We will present the first data obtained using a new variant of the Robo-Met.3D serial sectioning instrument that integrates a novel way to obtain crystallographic information using surface acoustic waves. The technique, known as spatially resolved acoustic spectroscopy (SRAS) uses two lasers to, firstly, induce a surface acoustic wave and secondly, measure its velocity. The velocity can be related to the elastic stiffness tensor, which in turn is related to the local crystal orientation. By conducting SRAS analysis on parallel slices in a completely automated fashion, it is possible to obtain 3D orientation microscopy datasets at the mesostructure length scale. We will describe the details of the method, as well as the instrumentation that was required to achieve these datasets in an automated manner.

10:45 AM
Pore Network Modelling Analysis of 3D SEM Images of Nano-porous Gold: S. Ali Shojaee¹; ¹Thermo Fisher Scientific
    Pore network modelling (PNM) Analysis is a technique for approximating complex porosity networks in materials. This study focuses on using PNM to analyze a such structure in nano-porous gold (samples courtesy of K.R. Mangipudi, Institute for Materials Physics, University of Goettingen). 3D SEM images of gold samples (images courtesy of Dr. Karsten Thiel, Fraunhofer IFAM) were used to design a complete image analysis workflow which included first aligning the stack of 2D SEM slices, rendering a 3D image, segmenting the SEM image into gold and porosity, and analyzing the pore space. In addition, PNM was performed using Thermo Scientific^TM Avizo^TM software on the results to quantify nodes, throats, coordination number, and other porous structure parameters. Considering the importance of the porous structure in the sensory and catalysis applications of the nano-porous gold, the results can be utilized to evaluate the performance of the material.

11:05 AM
Automated Serial Sectioning as a Method to 3D Map Inclusions in Structural Metals: Veeraraghavan Sundar¹; Rachel Reed¹; ¹UES Inc.
    The characterization of metallic and nonmetallic inclusions (MI/NMIs) for quantity, size, and shape distributions is of great interest for prediction of the mechanical properties as well as quality control of structural metals. Automated serial sectioning is a practical method of characterizing such MI/NMI distributions accurately in statistically relevant volumes of materials. Problems such as quantifying spheroidal graphite in cast iron, failure inducing inclusions creating white etching regions in M50 steel, as well as NMIs in Inconel and Al 6xxx series alloys were investigated. Critical parameters such as the difference between size distributions as well as morphologies in inclusions were comparable between samples. For example, of the two Al6xxx samples investigated, one had a higher percentage of fine (< 10 micron) inclusions, and a lower percentage of smaller aspect ratio inclusions (alpha AlFeSi, aspect ratio < 5) than the other. This has an impact on mechanical properties.

11:25 AM
3D Characterization of the Evolution of Crystal Mosaicity During Solidification of Single Crystal Ni-based Superalloys: Felicitas Scholz¹; Daniel Kotzem¹; Pascal Thome¹; Jan Frenzel¹; Gunther Eggeler¹; ¹Ruhr-Universitaet Bochum
    The present work investigates the formation and evolution of crystal mosaicity during Bridgman solidification of single crystal Ni-based superalloys of CMSX-4 type. In general, mosaicity is a measure of the spread of crystal plane orientations. A mosaic superalloy single crystal is typically characterized by the presence of small subgrains with misorientations up to a few degrees. We show that crystal mosaicity is caused by dendrite deformation processes. We use a new tomographic serial sectioning technique, where neuronal network based machine learning processes are employed to detect and to trace growth paths of more than 3000 individual dendrites in large sample volumes. We use this technique to study how dendrite branching events and the interaction between individual slightly misaligned dendrites and their environment govern the evolutions of microstructures and thus single crystal quality.

11:45 AM
Using Convolutional Neural Networks to Visualize Large Serial Sectioning Datasets: Zach Thompson¹; Tiberiu Stan¹; Peter Voorhees¹; ¹Northwestern University
    Serial sectioning coupled with optical microscopy can be used to obtain 3-dimensional image stacks from a materials sample quickly and with good section depth precision. The sectioning process can generate a large amount of data for a single sample (~10 GB) that needs to be accurately segmented into the phase of interest and the background. However, for certain measurements such as the number of fragments, the resulting data often contains too many artifacts that are very difficult to remove via filtering and which persist even when using state-of-the-art segmentation software. In order to remedy this, a novel machine learning approach tailored to use fewer ground truths was used to improve segmentation accuracy. Metrics are used to quantify this accuracy and what parameters most effect it. Finally, the degree of transferability of neural networks trained on one dataset and applied to other materials datasets is analyzed.