Importance of mineral surface areas in Rotliegend sandstones for modeling CO2-water-rock interactions

Investigations of Rotliegend sandstones from the NE Netherlands have been carried out to determine the mineralogical composition and spatial mineral distribution, specific mineral and whole rock surface areas and the integration of these results in geochemical water-rock-CO2 simulations. Therefore, two scenarios were considered: (1) volumetric rock composition (derived from point counting on thin sections) and (2) effective mineralogy (derived from 2D pore lining image analysis) data which represent the percentage amount of each mineral phase exposed within open (or accessible) pores. Specific mineral surface areas were defined for each mineral phase, respectively and integrated in both model scenarios. Based on point counting data the volumetric compositions of the sandstones are dominated by the mineralogy of detrital grains like quartz and K-feldspar. Authigenic minerals, esp. clays, volumetrically make up a small portion of the bulk rock volume only but are exposed due to their small crystal size (nm- to mu m-scale) and a large specific surface area (13 to 195 m(2)/g) in comparison to e.g. feldspar grains (0.03 to 1.79 m(2)/g). Hence the size, distribution and accessibility of detrital and authigenic minerals regulate the availability of chemical species for water-rock interactions in the pore space. The effective mineralogy of the sandstones is a function of the presence of grain covering and pore-filling authigenic minerals. In the two Rotliegend sandstones studied, approximately 23% and 43% of the detrital grains are covered by cements blocking the grain surface areas accessible for potential water-rock interactions. These minerals are Fe-oxide, clay cutans (mainly illite), kaolinite, quartz, carbonate (dolomite, ankerite and siderite), sulfate (anhydrite and barite) and galena. The specific surface area (SSA) of each mineral was calculated using (electron) microscopy to determine the sample geometry (e.g. size, thickness) and standard density values. For the volumetric and effective mineralogy data the total specific surface area of the bulk rock was calculated by summing the specific mineral surface areas in relation to their relative content in the rock, as determined using X-ray diffraction. These results were compared to surface area values determined using low pressure N-2 adsorption (BET method) of the bulk rock (mean value 1.4 m(2)/g). It is shown that the SSA values using the effective mineralogy approach (mean value of 0.46 m(2)/g) are closer to the whole rock BET than the SSA for volumetric mineralogy (0.05 m(2)/g). These different values for the volumetric and the effective mineralogy show significant differences in geochemical modeling results: The initial amount of K-feldspar, kaolinite, hematite and carbonate controls the availability of species for clay mineral precipitation (e.g. Mg-smectite) during CO2 storage. Effective mineral carbonation is calculated using the volumetric compared to the effective mineralogy. Within the first 100 years carbonate precipitation occurs for the volumetric mineralogy but is absent for the effective mineralogy. Opposing trends are observed for long-term reactions (>100 years). Similarly, when using the effective mineralogy, the porosity increases over the entire simulation time of 10,000 years while it initially decreases for the volumetric mineralogy followed by an increase and a net zero porosity change after 10,000 years. This implies that different results for CO2 trapping can be expected, depending upon the mineral assemblage and its effective mineralogy. (C) 2014 Elsevier B.V. All rights reserved.