Mineral nucleation and growth in porous and fractured media

Early simulation works on carbonization during CO2 storage overpredicted the amount and rate of mineral formation because of unrealistic growth models based on transition state theory (TST), where surface areas for growth were assumed to be present from day one. Inspired by [1], we made our first nucleation-growth model a decade ago [2], providing a zero-dimensional averaged kinetic nucleation-growth model for the continuum scale. This model provided some information on lag time for the onset of mineral growth, which is essential to understand why the lack of mineralization in short-term laboratory experiments cannot be extrapolated in time. The continuum-scale models, however, lack the potential to predict how the pore space geometries change and, thereby, how the permeability-porosity relationship changes in complex ways. Recent efforts have been to implement mineral nucleation and crystal growth in pore-scale models, using classical nucleation theory (CNT) and statistically derived expressions for the induction time of the formation of stable growing crystallites. In the new expressions, the probability of creating a crystallite at each surface site (solid surface area in the pore space) is estimated based on a normal distribution centered around the mean induction time value [τ] (symmetric between t= 0 and 2τ). This has been implemented and tested in 2D pore-scale models, showing strong mutual relations and impacts of transport and crystallization in porous media [3-4]. Another emerging field for nucleation-growth research is how the number of crystallites (or aggregates) and their spatial distribution control the evolution of porosity-permeability relations. 2D nucleation-growth Monte Carlo-type pore-scale RTMs show that the permeability-porosity relationship and the associated uncertainty depend on the number of crystallites (Figs. 1, 2). The relationship evolves predictably if the number of crystallites is small or large. If, on the other hand, the number of crystals is in an intermediate range, the permeability-porosity dynamics vary stochastically and are highly uncertain.