Here is a summary of my recent research conducted at the Research Centre for Fluid and Complex Systems at Coventry University:

Figure 1 shows that salinity reduction significantly decreases the activity coefficients calculated by both the extended Debye-Huckel and Pitzer activity models. This reduction causes noticeably lower concentration estimates for most ions, except for SiO2 (aq) and Ca2+, with reductions ranging from approximately 20% to 100% (see Figure 2). Furthermore, Figure 2 indicates that the extended Debye-Huckel model exhibits greater sensitivity to salinity reduction than the Pitzer model, suggesting that the former is unsuitable for synthesizing highly saline solutions. This considerable variation in equilibrium concentrations emphasizes the critical importance of accurately assessing the salinity of formation water in laboratory analyses. Uncertainties in salinity estimation could lead to significant errors when predicting the contributions of solubility and mineral trapping mechanisms during CO2 storage. These inaccuracies are comparable to the limitations associated with short-term equilibrium achieved through laboratory analysis of CO2-rock-water interactions.

Figure 1 : Changes in the activity coefficients of aqueous species relative to the original case when salinity degree decreases.

Figure 2 : Changes in the aqueous concentrations relative to the original case when salinity degree decreases.

However, this recommendation comes with a trade-off: the Pitzer model increases the time required to establish long-term equilibrium by more than 100% compared to the ideal model (see Figure 3). The reduced computational efficiency of the Reactive Transport Model (RTM) presents additional challenges for CO2 storage assessments under realistic conditions. To address this, I am currently developing a Machine Learning (ML) code, with support from my colleague @AbhiramAnand, to build surrogate models that replicate the behaviour of RTM while significantly reducing computational time. This approach enables a more comprehensive study of the uncertainties that may impact accurate assessments of CO2 storage.

Figure 3 : Increase in the elapsed time relative to the ideal model when using the extended Debye-Huckel and Pitzer activity models for establishing the long-term equilibrium.