InterPore2026

Convective-Driven, Contact Dissolution of Residually-Trapped Carbon Dioxide with Macroscopic Ripening

19 May 2026, 09:50

1h 30m

Poster Presentation (MS05) Physics of multiphase flow in diverse porous media Poster

Speaker

Hatem ALAMARA (CHLOE Research Laboratory)

Description

Previous studies on convective dissolution have investigated the rate at which free CO2 saturates an underlying brine layer through convective mixing. Recently, Mingotti and Woods (2025) conducted a laboratory experiment using saturated brine with dispersed salt powder overlying a freshwater layer. The latter configuration is analogous to the dissolution of residually-trapped CO2 in water. To our knowledge, numerical simulation studies of this specific phenomenon with physical parameters relevant to CO2-water systems are absent. In this work, we adopt the same configuration as in Ref. [1]; see Fig. (a) which shows a representative simulation case. Unlike the classical case of free CO2 convection in water, we observe that the dissolution rate does not exhibit a quasi-linear regime with an approximately constant value. Our simplified numerical models indicate that a significant amount of residually trapped CO2 can be dissolved. The mechanism of dissolution is straightforward: partially saturated water becomes fully saturated through contact dissolution as it advances through the upper layer and then descends. A related research question concerns how Ostwald ripening might affect the results. To explore the latter phenomenon, we simulate Ostwald ripening at the continuum scale, see Fig. (b), using macroscopic properties as in Ref. [2]. Across models spanning different lengths, the Ostwald ripening equilibrium time scales with the square of the characteristic length as previously reported in a number of studies. The key point we conceptualize is that if Ostwald ripening initially homogenizes a macroscopically homogeneous region in a certain time, this characteristic timescale is likely shorter than, or comparable to, the onset time of convection. However, the potential for localized mobilization and upward migration during this stage still needs investigation. Nevertheless, for larger simulation domains (> 1 m), once initial homogenization occurs, convective processes are expected to dominate over macroscopic ripening (or non-convective processes); Fig. (b) illustrates a representative case (10 m × 10 m). Finally, two dissolution regimes can be observed in the long term as depicted in Fig. (c). The ultimate dissolvable residually-trapped layer thickness relative to the total thickness can be expressed using simplified relationships, and the correct forms of those appearing in Mingotti and Woods’ (2025) study are reported.