Missing information and data fidelity in digital microstructure acquisition. (July 2019)

• Record Type: Journal Article
• Authors: Bao, Chongbing; Ji, Chuanyi; Poulsen, Henning Friis; Li, Mo
• Is Part Of: Acta materialia. Volume 173(2019)
• Page Start: 262
• Page End: 269
• Publication Date: 2019-07
• Subjects: Missing information -- Microstructure -- 3D digital microstructure -- Structure-property relation; Materials -- Periodicals; Materials science -- Periodicals; Materials -- Mechanical properties -- Periodicals; Metallurgy -- Periodicals; Chemistry, Inorganic -- Periodicals; 620.112
• DOI: 10.1016/j.actamat.2019.05.012
• Languages: English
• ISSNs: 1359-6454

Abstract

Abstract: Measuring and modeling microstructure is crucial for establishing structure-property relations. Here the digitized format of the microstructure plays an increasingly vital role in modern material design and advanced manufacturing of functional materials. However, one basic issue facing digital acquisition of microstructures has attracted little attention, namely the loss of information in the first steps of data gathering and processing. The missing information leads to serious issues in a range of topics from reconstruction of microstructures to prediction of material properties. To bring forth and quantify these issues, we develope an analytical method and a new numerical simulation scheme using Laguerre-Voronoi tessellation and Xu-Li microstructure characterization method. We then define and quantify the missing information and its impact rigorously. Graphical abstract: Comparison of the digitized microstructure in the form of voxels (or squares in the figure) with the real microstructure formed by the polygonal domains (or the grain boundary network). The missing grains in the digital microstructure are shown in red circles. Like traditional metallography that has played important roles in materials science for centuries, digitization of microstructure is the modern advance that promises to revolutionize synthesis, processing, and application of a vast number of materials. Compared to metallography, digital microstructure involves data with unprecedented magnitude of size and scale. However, one basic issue facing digital acquisition of microstructures has attracted little attention, namely, the loss of information in the first steps of data gathering and processing. The missing information leads to serious issues in data fidelity in a range of topics from reconstruction of microstructures to prediction of material properties. To address these issues, we developed an analytical method and a numerical simulation scheme using Laguerre-Voronoi tessellation and Xu-Li microstructure characterization method. With these new approaches, we define and establish this problem for the first time by showing quantitatively the missing information and its impact.Image 1