Hydrogeochemical changes and δ¹³C-DIC signature in a silicate-dominated aquifer during controlled CO₂ release: Insights for monitoring unintended CO₂ leakage at CCS sites

Changes in groundwater quality because of potential unintended CO₂ leakage at carbon capture and storage (CCS) sites is a critical environmental concern and is often cited as a socioenvironmental barrier to public acceptance. Although CO₂ is not a contaminant, its excessive concentration can induce geochemical reactions with aquifer mineralogy, thereby necessitating the evaluation of potential consequences. Despite recent advances in understanding the geochemical imprint of CO₂-impacted aquifers, few studies have focused on silicate-dominated aquifers rich in altered minerals under near-surface conditions, where precipitation can significantly influence these processes. Thus, this study was designed to evaluate the CO₂–groundwater–mineral interaction within a shallow aluminosilicate porous aquifer derived from granite at the TECNOPUC-Viamão CO₂ experimental site in southern Brazil. The primary objectives were to identify hydrogeochemical alterations and to determine key indicators that signal this process. Gaseous CO₂ was injected at varying rates (5, 10, and 12.5 kg/day) at approximately 3 meters below ground level, and groundwater samples were frequently collected from multilevel monitoring wells within a 5-meter radius. The CO₂ injection spanned 46 days, following a background monitoring campaign, and was succeeded by a two-week postinjection phase, culminating in a final sampling campaign conducted 137 days after the injection ceased. The hydrochemical evolution of the CO₂–groundwater–mineral interaction was tracked through physicochemical and chemical analyses of major, minor, and trace elements and δ¹³C-DIC (Dissolved Inorganic Carbon). Despite the influence of atmospheric precipitation on certain parameters, strong correlations (ρ > 0.70) were observed between the partial pressure of CO₂ (PCO₂) and variables such as electrical conductivity (EC), pH, ORP, total alkalinity (TA) and δ¹³C-DIC, as well as Ca, Mg, K, Na, Si, Ba, Sr, Fe, Mn, PO₄³⁻, and HCO₃⁻ levels. A clear chronological evolution pattern with increasing injection rates was noted for EC, TA, Ca, Mg, K, Na, Si, Ba, Sr, Fe, and HCO₃⁻ in the closest wells. Considering the natural seasonal hydrogeochemical variability and the control chart constructed for the hydrostratigraphic unit, the sensitivity to CO₂ interactions, including seasonal effects, indicates that EC, Ca, Mg, Fe, HCO₃⁻, Ba, Sr, Mn and δ¹³C-DIC provided a consistent diagnosis of the interactions. In terms of the risk to water quality, the Fe and Mn concentrations exceeded the limits set by both national and international regulations. This study suggests that this suite of variables enables effective diagnosis and assessment of the impact of unintended CO₂ leaks in CCS projects, provided that they are adequately monitored within a well-established seasonal background.