A synchrotron-based local computed tomography combined with data-constrained modelling approach for quantitative analysis of anthracite coal microstructure

Abstract

Quantifying three-dimensional spatial distributions of pores and material compositions in samples is a key materials characterization challenge, particularly in samples where compositions are distributed across a range of length scales, and where such compositions have similar X-ray absorption properties, such as in coal. Consequently, obtaining detailed information within sub-regions of a multi-length-scale sample by conventional approaches may not provide the resolution and level of detail one might desire. Herein, an approach for quantitative high-definition determination of material compositions from X-ray local computed tomography combined with a data-constrained modelling method is proposed. The approach is capable of dramatically improving the spatial resolution and enabling finer details within a region of interest of a sample larger than the field of view to be revealed than by using conventional techniques. A coal sample containing distributions of porosity and several mineral compositions is employed to demonstrate the approach. The optimal experimental parameters are pre-analyzed. The quantitative results demonstrated that the approach can reveal significantly finer details of compositional distributions in the sample region of interest. The elevated spatial resolution is crucial for coal-bed methane reservoir evaluation and understanding the transformation of the minerals during coal processing. The method is generic and can be applied for three-dimensional compositional characterization of other materials.

Boosting Resolution and Recovering Texture of 2D and 3D Micro‐CT Images with Deep Learning

“…The carbonate images are underresolved due to much of the pore space measuring under 1 μm in resolution (Bultreys et al, 2015) (requiring an image resolution approaching 100 nm to adequately resolve such features). The coal samples are inadequately resolved due to the fragility of the rock itself, restricting the size to a larger than usual sample, limiting the resolution of the resulting image (Hao Chen et al, 2014). The images are downsampled by a factor of 2X and 4X using the MATLAB imresize function, used commonly for SR dataset benchmarking , with one set being downsampled exclusively by bicubic interpolation while another set is downsampled with random kernels of either box, triangle, lanczos2, and lanczos3.…”

Characterization of Darai Limestone Composition and Porosity Using Data-Constrained Modeling and Comparison with Xenon K-Edge Subtraction Imaging

“…For instance submicron porosity within a region will result in a lower X-ray attenuation in the X-ray micro-CT scan and can lead to ambiguity between nonporous mineral phases of lower X-ray attenuation and nanoporous regions of higher X-ray attenuation phases, since individual voxels within the micro-CT scanned volume may contain more than one phase (including porosity). This approach has already been used in a range of applications in petroleum geology (Mayo et al, 2012;Trinchi et al, 2012;Wang et al, 2013aWang et al, , 2013bWang et al, , 2013cYang et al, 2013aYang et al, , 2013bChen et al, 2014). One approach to resolving the ambiguities in micro-CT imaging of rocks containing submicron porosity is to use two or more micro-CT scans at different energies combined with data-constrained modeling (DCM) (Yang et al, 2010a).…”

“…It was demonstrated that quantitative results may be obtained by comparing scans with and without the contrast agent present (Withjack & Akervoll, 1988). This method has since been widely applied to a range of sample types (Ketcham & Iturrino, 2005; Ghous et al, 2007; Boone et al, 2014). Contrast agents often used in lab-based CT for this purpose include potassium or sodium iodide solutions (Agbogun et al, 2013; Andrew et al, 2014) and in some cases mercury (Klobes et al, 1997; Fusi & Martinez-Martinez, 2013).…”

“…DCM uses mathematical optimization to generate such a 3D compositional map that is consistent with the micro-CT data sets. This approach has already been used in a range of applications in petroleum geology (Mayo et al, 2012;Trinchi et al, 2012;Wang et al, 2013aWang et al, , 2013bWang et al, , 2013cYang et al, 2013aYang et al, , 2013bChen et al, 2014). DCM makes use of the different ways in which X-ray attenuation of different mineral phases change with energy to distinguish between them.…”