Joint Sessions of AIM, ICME, & 3DMS: Automated and Autonomous Research
Program Organizers: TMS Administration
Tuesday 9:00 AM
June 17, 2025
Room: Platinum Ballroom 7&8
Location: Anaheim Marriott
Session Chair: Gregory Thompson, University of Alabama
9:00 AM Invited
Innovations in 3D EBSD for Advanced Materials Characterization: Andrew Polonsky1; Chad Hovey1; James Lamb2; Paul Chao1; McLean Echlin2; Hojun Lim1; Kyle Johnson1; Julia Deitz1; Tresa Pollock2; 1Sandia National Laboratories; 2University of California Santa Barbara
Recent advancements in automation have transformed materials characterization techniques over the past decade. The capability to perform serial-sectioning within an electron microscope using ion or laser beams has facilitated the acquisition of high-fidelity 3D electron backscatter diffraction (EBSD) data, resulting in increasingly large datasets. Here we explore the application of 3D EBSD to analyze microstructures obtained through serial-sectioning with the TriBeam system, revealing insights into these materials that traditional methods cannot provide. We will address analytical considerations for non-equilibrium microstructures, such as those produced by macroscopic plastic deformation or additive manufacturing. Additionally, we will outline workflows for automated management of multi-terabyte datasets and present novel collection and analysis tools designed to minimize the reliance on specialized knowledge, thus making this advanced technique more accessible to a broader audience. Furthermore, we will discuss how these datasets can be leveraged for advanced modeling techniques to enhance our fundamental understanding of materials processing.
9:30 AM
Capabilities and Applications of the Robot-Assisted Serial-Sectioning and Imaging (RASI) System: Michael Moschetti1; Dirk Bettge1; 1Federal Institute for Materials Research and Testing
The Robot-Assisted Serial-sectioning and Imaging (RASI) system at BAM provides automated, high-resolution 3D microstructural characterization for diverse materials. Integrating robotics with precision sectioning, etching, and optical microscopy, RASI reconstructs large volumes (approaching 15×15×15mm3) with sub-micron detail. This presentation showcases RASI’s versatility through case studies on cast irons, sintered and additively manufactured steels, and ceramic-metallic packages. We demonstrate how RASI reveals true 3D architectures of features like graphite networks, pores, melt pools, and defects, often missed by 2D analysis. These quantitative datasets elucidate process-microstructure-property relationships and provide crucial 'ground truth' for validating computational models and developing digital twins. Ongoing RASI enhancements will also be highlighted.
9:50 AM
Influence of 3D Crack Networks for High Toughness Responses in Tantalum Carbides: Alyssa Stubbers1; Gregory Thompson1; Chris Weinberger2; Sierra Durkee1; Evan Schwind1; Mireya Garcia3; Olivia Graeve4; Edgar Solano3; Alejandro Ramirez3; 1University of Alabama; 2Colorado State University ; 3Instituto Politecnico National; 4University of San Diego
The zeta-phase, ζ-Ta4C3, is reported to have a fracture toughness above 15 MPa·m1/2, which is a factor of two to three times larger than most other ceramic materials. This fracture strength is derived from the interlocking lath structure of zeta phase precipitation in a tantalum carbide matrix. These laths provide anisotropic mitigation of crack propagation as well as a buckling response. Furthermore, a local metal-metal bond in the zeta-phase’s unit cell facilitates plasticity through dislocation nucleation. The presented work addresses the 3-dimensionality of the crack pathways. Here, the carbide was subjected to microindents from which plasma-focus ion beam based serial sectioning and subsequent reconstruction renderings were undertaken to reveal the cracking network. From acquired images, three distinct crack types are cataloged: linear, bifurcating, and kinking. Finally, as serial sectioning can be a time-consuming characterization method, we have also implanted an image-based machine learning method to identify cracking features and directions between slices thereby reducing either the number of slices needed and/or the resolution required between slices.
10:10 AM
FIB-SEM Serial Sectioning Tomography: Towards 24-Hour Time-to-Results: Bartlomiej Winiarski1; Patrick Barthelemy1; Chengge Jiao1; 1Thermo Fisher Scientific
Focused Ion Beam – Scanning Electron Microscope (FIB-SEM) and Plasma FIB (PFIB)-SEM Serial Sectioning Tomography are well-established techniques for high-resolution, 3D imaging and analysis of materials at the multiscale. This method involves the sequential removal of thin layers of material, followed by imaging to reconstruct detailed 3D microstructures. Despite its potential, the time-intensive nature of the process has been a significant barrier to widespread adoption. Our research aims to optimize PFIB SST to achieve a 24-hour time-to-results. By integrating advanced automation techniques for focused beam sectioning and SEM imaging, along with streamlined and automated data processing and segmentation algorithms aided by AI, we propose a comprehensive approach to significantly reduce imaging, reconstruction times, data segmentation, and reporting. Exemplary results from a Solid Oxide Fuel Cell (SOFC) demonstrate promising improvements in both speed and accuracy, suggesting that a 24-hour turnaround is attainable.
10:30 AM Break
11:00 AM Invited
Building a Self-Driving Lab From Scratch: Shijing Sun1; 1University of Washington
The development of autonomous laboratories marks a paradigm shift in materials research by automating experimental processes—from planning to analysis. This talk explores three approaches to laboratory autonomy to accelerate discovery. The first involves an all-in-one platform, where an AI “brain” directs workflows and real-time computer vision guides decisions. The second emphasizes modular integration, adding robotic components and using lower-fidelity proxies with human oversight for flexibility. Lastly, open-source hardware offers customizable, cost-effective automation setups, democratizing advanced tools for researchers. These approaches demonstrate how to build functional self-driving labs, transforming experimentation and accelerating materials design.
11:30 AM
Using Novel EBSD Methods to Analyze Plastic Strain in Structural Alloys: David Rowenhorst1; 1US Naval Research Laboratory
The microstructural characterization of materials is essential for validating and informing the models used in ICME. While there are well established workflows for understanding mean values, understanding sub-grain orientation gradients are critical for making accurate predictions in the amount of plastic strain within a material formed due to traditional thermo-mechanical processing, phase transformations, or residual stresses formed during rapid solidification. Here we will present new methods for evaluating the amount of plastic strain present in a material by applying advanced electron backscattered diffraction (EBSD) microscopy. We will show that by combining pattern denoising methods such as the NLPAR algorithm, along with new EBSD indexing methods that operate at high speeds, and high precision, one can detect grain misorientations that are below 0.2°, while collecting using typical scan conditions. As a case study, we will compare the orientation gradients in additive manufactured 316L stainless steel with newly developed CAFE-CP models.