Fluid-solid interactions related to subsurface storage of CO2 Experimental tests of well cement

An on-going Norwegian research project focus on the geochemical interaction of CO2 with cap-rocks as well as engineered material in wells, and corresponding changes in geomechanical properties. The aim is to establish the nature and extent of induced changes in petrophysics and geomechanical rock/material properties caused by geochemical reactions during long-term exposure to CO2 and CO2-water mixtures. Furthermore, it is an aim to experimentally simulate a leakage situation, where improved understanding of transport and reaction mechanisms and qualification of monitoring techniques are important. Abandoned wells have been identified as one of the most probable leakage pathways for underground CO2 storage. A comprehensive understanding of the fluid-solid interaction processes in the well and in the near-well area and methods for evaluation of long-term well integrity is necessary. This paper presents preliminary experimental results on the chemical interactions of well cement and brine-CO2 at conditions relevant for subsurface storage of CO2 in aquifers (100 bar, 50 degrees C). Two different scenarios have been tested; the impacts of CO2 on well cement in a static environment, and secondly, the effects of carbonated brine transport on well cement surfaces. Thus, cement cores have been the subject to a series of static batch experiments with variable duration and water content, as well as steady state flow-through experiments where a CO2 saturated brine is forced through a channel in the cement. The batch experiments, resembling the long term storage situation, show extensive cement carbonation and near full conversion of Portlandite (Ca(OH)(2)) to calcite and aragonite on the outer layers of the specimens. The flow experiments reveal a different behaviour due to the advective transport of reactants and reaction products. Cement decalcification to yield a layer of amorphous silica, was found to be rapid and quantitative in the immediate vicinity of the convective flow. Complete cement dissolution and consequent widening of the flow channel was also observed. Results on the propagation rate of reaction fronts are reported. Further work will include geomechanical experiments. (C) 2008 Elsevier Ltd. All rights reserved