--> Application of Multiple Hyperspectral Imaging Tools to the Examination of Submillimeter Variability in Geochemical Reference Materials from Major U.S. Shale Plays

2019 AAPG Annual Convention and Exhibition:

Datapages, Inc.Print this page

Application of Multiple Hyperspectral Imaging Tools to the Examination of Submillimeter Variability in Geochemical Reference Materials from Major U.S. Shale Plays

Abstract

A variety of imaging and spectroscopic techniques are currently available for assessing geochemical and mineralogical properties of mudrocks at submillimeter scales. These non-destructive techniques fill an analytical niche at intermediate sizes between bulk and microscale analyses that have previously been examined using methods like thin section analysis but can now be examined more efficiently and definitively using spectral data. In this study, a combination of short-wave (SWIR, 1.0 to 2.5 μm) and mid-wave (MWIR, 2.5 to 5 μm) infrared spectroscopy and micro-X-ray fluorescence (μ-XRF) was applied to shale pieces (several centimeters long and across) from outcrops and roadcuts of several important U.S. shale plays. Samples included mudrocks from the Eocene Green River (Colorado), Late Cretaceous Niobrara (Colorado), Late Cretaceous Boquillas (Texas), Devonian-Mississippian Woodford (Oklahoma), and Middle Devonian Marcellus (New York) formations. These samples contained variably laminated structures, some of which have been examined by submillimeter Fourier transform infrared (FTIR) spectroscopy (2.5 to 25 μm) in previous work, and showed variable mineralogical composition and organic content. Bulk analyses of mineralogy and total organic carbon (TOC) content show that the shales vary from carbonate-rich (Green River, Niobrara, and Boquillas) to clay mineral-rich (Marcellus) to highly siliceous (Woodford). TOC contents for the analyzed shales range from 1 to 20 wt. %. Elemental and mineral maps were generated and compared to results from previous geochemical and spectral analyses. Lamina in the calcareous samples show alternating carbonate- and organic-rich compositions, while the more siliciclastic samples show differences in clay minerals, quartz, and pyrite content. In general, the elemental associations identified using the XRF maps agree with the mineralogical interpretations. Multivariate curve resolution was applied to SWIR and MWIR spectral datasets for each sample to better differentiate mineralogical features into millimeter-scale lithologies. The results show the utility of hyperspectral imaging for examining significant submillimeter variability in source rocks from both siliciclastic and carbonate-rich petroleum systems.