Abstract Flow-through experiments in the laboratory were conducted to monitor the fate of CO2 using stable carbon isotope (δ13C) techniques in dynamic, pre-equilibrium conditions. Such conditions are typical, for instance in carbon capture storage (CCS), in the initial stages of CO2 injection, near injection well regions of the reservoir. For this purpose, a reactive percolation bench (ICARE 4) was used, injecting a CO2-saturated brine at supercritical conditions (pCO2 = 84 bar, T = 60 °C) through quartzitic limestone at an average flow rate of 2 × 10− 9 m3 s− 1. Calcium (Ca2 +) and dissolved inorganic carbon (DIC) concentration data and pH were used to aid analytical interpretations. During CO2 injection, δ13CDIC values decreased from about − 11‰ to those of the injected CO2 (− 29.3‰), indicating CO2 sourced carbon dominance over a carbonate sourced one in the system. Simultaneously DIC and Ca2 + concentrations increased from 1 mmol L− 1 to a maximum of 71 mmol L− 1 and 31 mmol L− 1, respectively. Isotope and mass balances were used to quantify the amount of DIC originating from either the injected CO2 or carbonates. At the end of the experiments, between 70 and 98% of the total DIC originated from CO2 dissolution, the remaining amount is attributed to carbonate dissolution. Furthermore, the total amount of injected CCO2 trapped as DIC ranged between 9 and 17% and between 83 and 91% was in free phase. The state of carbonate equilibrium of the host fluid, under the high pressure–temperature conditions after CO2 injection was identified, verifying pre-equilibrium conditions. Results confirm observations made in reported field data. This emphasises that the combination of CO2 monitoring, the development of a thorough understanding of carbonate equilibrium, as well as the quantification of CO2-trapping, is essential for a solid assessment of reservoir performance and safety considerations during CO2 injection. These are equally important for understanding water–rock–CO2 dynamics in natural subsurface environments.