Samples of coal from the Sydney and Bowen Basins of eastern Australia have been imaged at high resolution using a large-field, 3D microfocus X-ray computed tomography (mu CT) system, with special but not exclusive attention to evaluating the modes of occurrence of the mineral matter within the coal. The samples imaged were 110 mm, 25 mm, 19 mm, 10 mm, and 4 mm in size, yielding voxel dimensions of 54, 30, 12, 6, and 3 mu m respectively. Data collection was carried out using a helical stage, providing images with >2000(2) voxels in the horizontal (X-Y) plane and up to 3500 voxels high. Three-dimensional image blocks derived from the scans were examined as cross-sections along orthogonal planes and as perspective images, manipulated to be viewed from any angle. Imaging after saturating the coal with X-ray attenuating brine was also carried out to highlight the distribution of connected micro-pores and cleats, and improve the detail of features seen within the samples. Features evaluated within the coals included the size and three-dimensional distribution of siderite nodules, and different types of mineral infillings in petrifactions of maceral components. Individual macerals could also be identified within the coal, based partly on X-ray density and partly on the associated porosity and structure. In some cases high-resolution images enabled the nature of individual plant particles to be identified within the coal samples. Mineral-filled cleats and open fractures were also evaluated, including the origin of radiating fracture patterns around siderite nodules in vitrinite. In some cases several generations of cleat and/or fractures could be distinguished, and the sequence of their formation and infilling was interpreted. Complementary analyses of the mineral matter in the samples were carried out using X-ray diffraction, as well as examination of polished sections by optical microscopy examination. Images obtained from the mu CT scans were also registered against SEM-EDX and QemSCAN images of polished sections prepared from the same samples after scanning, providing a more definitive basis for identifying the different components and for integrating mu CT data with results from other petrographic and electron microscope studies. (C) 2012 Elsevier S.V. All rights reserved.