“4D SEM” aka EBSD – Lessons Learned & Opportunities Presented

It is exciting, insightful and useful to collect and analyze 2D micrographs, where each measurement point contains a pixelated ‘2D’ diffraction pattern. In the scanning electron microscope (SEM), this technique could have been called “4D SEM”, however in ∼1987 Dingley and co-workers at Bristol University reported some of the first integrated systems that could perform this, calling it “electron backscatter diffraction” (EBSD) [1]. In brief, Dingley et al. recognized the value of microstructure mapping by combining a scanning electron microscope together with a phosphor scintillator, low light TV camera, and automated on-line/near-line analysis routines (see Figure 1) to generate rich and easy to interpret microstructural maps which are collected in near-real-time. These maps can be used to describe properties such as crystallographic texture, phase fraction, elastic strain variations and even evidence of plastic strain (a more complete history can be found in [2] – including EBSD before TV cameras were being used). Over the past ∼38 years, EBSD has advanced significantly through contributions from an international user community and commercial vendors, becoming a cornerstone technique in many labs worldwide. Notable developments include the introduction of transmission Kikuchi diffraction (TKD) by Keller and Geiss in 2011 [3], enabling analysis of 10 nm domains using electron-transparent thin films in the SEM. Furthermore, the flexibility of SEM chambers allows new EBSD/TKD applications such as in situ experiments (e.g., heating, Fig. 2A), STEM-in-SEM with FIB lift-outs (Fig. 2B), large area analysis (Fig. 2C), tomography via serial sectioning with FIB-SEM (Fig 3D), and multi-modal analysis through simultaneous acquisition & analysis of signals like energy-dispersive X-ray (EDS/X) spectra (Fig. 2E). The popularity of EBSD likely stems from several factors: (a) an active community of practitioners, developers, and vendors; (b) advancements in hardware and software, including automation, fast pattern collection and online analysis; and (c) versatile use cases with rich multi-modal data collection and easy-to-access analysis, where EBSD helps address challenges in fields like materials science and engineering, earth sciences, and more. Looking ahead, the combination of user-friendly commercial software and a robust open-source community will drive further adoption & new analysis approaches that will ultimately benefit more fields and industries (e.g., steel production, additive manufacturing, aeroengine materials, semiconductor characterization, minerology) and open up new research areas (e.g., clean tech, including beam-sensitive perovskite solar cells [4]). Advances in SEM architectures also allow for custom hardware solutions, integrated workflows for routine in situ experiments (e.g., heating, cooling, and mechanical testing), and low-cost EBSD systems using compact direct electron detectors [5]. As we look towards the future, we can see applications driven use of machine learning tools to amplify signal to noise and rapidly reduce large data sets that can reveal new information (e.g. ordered precipitates in a Co/Ni-based superalloy [6]). Further advances include the seamless merging of chemical and structural data to achieve spatial resolutions to resolve the distribution of precipitates with very high spatial resolution (10s of nm) across large areas in bulk samples [7] that have had limited sample preparation (as compared to TEM-based lamella prep). Finally, it is now possible to directly borrow algorithms from the EBSD community and apply these to STEM-data, e.g. through the automated analysis of Kikuchi patterns in the TEM [8], as well as the opportunity to draw upon the many years of experience in the analysis of SEM-based 2D microstructure analysis of crystal orientation, phase and chemical variation. In this presentation, we will explore how these historical approaches have provided a rich playground for new microstructural analysis and understanding of materials, to hopefully prompt us to consider how we might see the blending of EBSD and 4D-STEM developments in the future [10]. Historical and current EBSD developments: (A) the initial set up proposed by Dingley and co-workers in 1987 with their online interface (adapted from [1]); (B) the rapid acceleration of EBSD pattern acquisition speeds allowing for increased data collection (adapted from [9]); (C) in-chamber IR-camera images showing a modern EBSD-set up (Aztec), in an AMBER-X plasma focused ion beam scanning electron microscope set up for 3D analysis; (D) a modern computer interface for online experimental set up, indexing, and real time analysis; (E) an off-axis TKD experimental set up, in the same microscope as C. Example EBSD experiments from our lab: (A) high temperature mapping of Zr-microstructures; (B) high spatial resolution mapping of steel microstructures, revealing sub-μm retained austenite, in a STEM-in-SEM TKD experiment; (C) a 5x5 mm2 large area map of Mg grains; (D) a 3D volume collected by pFIB-SEM based EBSD of steel, with a (100 nm)3 voxel size; (E) phase analysis in steel, revealing a Nb/Ti-C precipitate in a ferrite matrix using simultaneous EBSD and EDS analysis.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.