Bridging Laboratory and Field Scales: A Plug-Flow Reactor to Assess Interactions Between Dissolved CO2 and Mafic/Ultramafic Rock

A plug-flow reactor (PFR) with a 40-60 kg capacity for crushed rock at a targeted particle size range of 4.0-12.5 mm has been constructed to assess CO₂-H₂O-rock interactions and derive kinetic dissolution rates. By employing particle sizes significantly larger than those used in conventional laboratory dissolution experiments, this system aims to improve the accuracy of laboratory-derived rates relative to field behaviour. Preliminary testing has been conducted using three different mafic/ultramafic site samples from eastern Canada: feldspar-dominant samples from Tamworth, Ontario; forsterite-dominant samples from Thetford Mines, Quebec; and North Mountain Basalt samples from Nova Scotia. Deuterium and potassium chloride are used as conservative tracers to validate flow behaviour within the PFR, providing a baseline for reactive tracer experiments. Reactive tracers are implemented to estimate the effective surface area of the packed sample particles. Tracers with sorptive properties that have been preferentially explored include cesium chloride, strontium chloride, fluorescein, and rhodamine. Batch experiments were performed to characterize sorption kinetics and equilibrium behaviour across particle sizes. These results are compared to the breakthrough curves from flow-through experiments using a retardation factor to estimate the distribution coefficient. CO₂-saturated water is prepared in a separate vessel at ambient temperature and a maximum pressure of 25 psi to produce a solution with a pH of ~4.6, comparable to the values during field-scale injections. Preliminary dissolution experiments recycled the CO₂-H₂O solution through the PFR to create semi-batch conditions and provide insights into the dynamic fluid chemistry as dissolution progresses. Temperature, pressure, pH, and conductivity are recorded at the inlet and outlet, while intermittent fluid sampling determined divalent cation and secondary metal concentrations over time. Steady-state concentrations were used to calculate dissolution rates normalized to the effective surface area. Transport behaviour was analyzed using an independent model based on the advection-dispersion equation accounting for retardation and permanent sorption coupled to mixing equations for the inlet and outlet zones of the PFR.  PHREEQC was employed to predict reactive transport results from dissolution. Comparisons between experimental and modeled dissolution rates provide insight into scaling laboratory results to field conditions and improving predictions of mafic/ultramafic rock reactivity for mineral carbonation.