Mineral precipitation-induced porosity reduction and its effect on transport parameters in diffusion-controlled porous media

Background: In geochemically perturbed systems where porewater and mineral assemblages are unequilibrated the processes of mineral precipitation and dissolution may change important transport properties such as porosity and pore diffusion coefficients. These reactions might alter the sealing capabilities of the rock by complete porescale precipitation (cementation) of the system or by opening new migration pathways through mineral dissolution. In actual 1D continuum reactive transport codes the coupling of transport and porosity is generally accomplished through the empirical Archie's law. There is very little reported data on systems with changing porosity under well controlled conditions to constrain model input parameters. In this study celestite (SrSO4) was precipitated in the pore space of a compacted sand column under diffusion controlled conditions and the effect on the fluid migration properties was investigated by means of three complementary experimental approaches: (1) tritiated water (HTO) tracer through diffusion, (2) computed micro-tomography (mu-CT) imaging and (3) post-mortem analysis of the precipitate (selective dissolution, SEM/EDX). Results: The through-diffusion experiments reached steady state after 15 days, at which point celestite precipitation ceased and the non-reactive HTO flux became constant. The pore space in the precipitation zone remained fully connected using a 6 mu m mu-CT spatial resolution with 25 % porosity reduction in the approx. 0.35 mm thick dense precipitation zone. The porosity and transport parameters prior to pore-scale precipitation were in good agreement with a porosity of 0.42 +/- 0.09 (HTO) and 0.40 +/- 0.03 (mu-CT), as was the mass of SrSO4 precipitate estimated by mu-CT at 25 +/- 5 mg and selective dissolution 21.7 +/- 0.4 mg, respectively. However, using this data as input parameters the 1D single continuum reactive transport model was not able to accurately reproduce both the celestite precipitation front and the remaining connected porosity. The model assumed there was a direct linkage of porosity to the effective diffusivity using only one cementation value over the whole porosity range of the system investigated. Conclusions: The 1D single continuous model either underestimated the remaining connected porosity in the precipitation zone, or overestimated the amount of precipitate. These findings support the need to implement a modified, extended Archie's law to the reactive transport model and show that pore-scale precipitation transforms a system (following Archie's simple power law with only micropores present) towards a system similar to clays with micro-and nanoporosity.