Carbon mineralization in CO2-seawater-basalt systems: Reactive transport dynamics and vesicular pore architecture controls

TL;DR

The paper investigates CO2 mineralization in CO2-charged seawater interacting with basalt under reactive transport, combining 30-day flow-through experiments at 80 °C with multi-scale imaging and geochemical modeling. It demonstrates that mineralization is governed by nucleation kinetics and stochastic site selection rather than uniform growth, with residence time and pore topology (vesicular basalts with modal coordination ~2) critically shaping mineralization patterns and permeability evolution. Calcite dominates the carbonate phase under seawater conditions, while smectite clays sequester divalent cations and inhibit Mg-bearing carbonates, reducing overall mineralization efficiency compared with freshwater systems. The findings imply that probabilistic reactive transport frameworks and realistic pore-topology representations are essential for predicting storage performance in vesicular basalt reservoirs, and that seawater-based strategies require careful management of pore-scale clogging and competing clay reactions to maintain injectivity and permanence of storage.

Abstract

Carbon mineralization in basaltic rocks may offer rapid, permanent \ce{CO2} storage, yet fundamental controls on reactive transport and precipitation patterns remain poorly understood. This study integrates flow-through experiments at 80\degree C using \ce{CO2}-acidified seawater with geochemical simulation and multi-scale pore imaging to elucidate mineralization dynamics in basaltic glass. Results reveal that carbonate precipitation is nucleation-controlled and stochastic rather than growth-controlled and deterministic, with isolated accumulations forming randomly despite continuous supersaturation. Residence time exerts primary control: reducing flow rate from 0.05 to 0.005\,mL/min proved necessary for visible precipitation. Post-experiment analyses identified calcium carbonate and smectite phases. Multi-scale characterization of three basalt facies revealed that connected porosity fractions (1.3--32\%) differ significantly from total porosity (18--42\%), demonstrating that network topology controls permeability. Micro-CT analysis revealed that pore coordination numbers in basalts (modal = 2) were notably lower than those in reservoir sandstones, creating serial flow paths that are vulnerable to catastrophic permeability loss from modest precipitation. Precipitation-induced clogging scenarios were proposed, where distributed small precipitates cause more severe permeability degradation than large accumulations. The use of seawater complicates geochemistry and reduces mineralization efficiency compared to freshwater. Findings emphasize the need for probabilistic reactive transport modeling frameworks and realistic pore topologies, which are fundamentally different from conventional CCS operations.

Figures (5)

• Figure 1: Multi-scale SEM characterization of unreacted Stapafell basaltic glass grains used in flow-through experiments. (a) Low-magnification overview (25$\times$, pixel size = 2µm) showing grain size distribution and morphology. (b) Intermediate magnification (50$\times$, pixel size = 1µm) revealing smooth glassy surfaces characteristic of rapidly cooled volcanic material. (c) High-magnification image (100$\times$, pixel size = 0.5µm) displaying surface features including natural cavities, vesicular structures, and conchoidal fracture patterns, with no evidence of pre-existing secondary mineral phases. (d) Representative EDS spectrum from pristine grain surfaces, confirming the presence of major elements (Si, Al, Fe, Ca, Mg, O) consistent with tholeiitic basalt composition and demonstrating the availability of reactive divalent cations (Ca2+, Mg2+, Fe2+) for CO2 mineralization reactions.
• Figure 2: Reactive transport experiment results and geochemical modeling of CO2 mineralization in basaltic glass columns. (a) Experimental setup showing the column reactor configuration with an initial 4cm calcite section and 36cm basaltic glass section during CO2-acidified seawater injection at 0.005mL/min. (b) Temporal evolution of carbonate mineral precipitation at 7, 21, and 30 days, demonstrating random spatial distribution of white carbonate accumulations along the basalt section, with preferential formation of larger precipitation patches in the latter half of the column. (c) PHREEQC v3 reactive transport simulation results: (top) calcite saturation index evolution for three basalt glass dissolution rate scenarios, fast (k = 1.0e-8mol.m^-2.s^-1), medium (k = 1.0e-9mol.m^-2.s^-1), and slow (k = 1.0e-10mol.m^-2.s^-1) with estimated outlet pH constraining actual rates near the medium case. (bottom) saturation indices of carbonate minerals along the column length showing supersaturation with respect to dolomite (CaMg(CO3)2), magnesite (MgCO3), and huntite (Mg3Ca(CO3)4) using medium dissolution rate (k = 1.0e-9mol.m^-2.s^-1). (d) High-resolution images of isolated carbonate precipitate bodies on dark basaltic glass substrate, with scale bars indicating millimeter-scale accumulations, demonstrating the nucleation-controlled, stochastic nature of mineralization rather than uniform growth-dominated precipitation.
• Figure 3: Pore-scale characterization of post-experiment substrates revealing dissolution features, spatial precipitation patterns, and secondary mineral assemblages. (a--b) Surface texture of calcium carbonate plug grains after reactive flow, showing dissolution features and surface roughening from acidified seawater interaction. (c--d) Basaltic glass grain surfaces, exhibiting characteristic dissolution morphology with enhanced surface roughness. (e) Preferential carbonate precipitation in peripheral regions adjacent to the flow cell wall, attributed to reduced advection velocity in the boundary layer and enhanced fluid retention on water-wet glass surfaces. (f) Enhanced carbonate mineralization within natural cavities and vesicular features of basaltic glass grains. (g) SEM micrograph showing pervasive smectite clay formation with characteristic texture on basaltic glass surfaces. (h) Early-stage carbonate formation, characterized by crystallites and amorphous calcium-rich precipitates, on basaltic glass surfaces.
• Figure 4: Multi-scale characterization of three Icelandic basalt facies representing the spectrum of vesicularity in basaltic reservoirs. Columns from left to right: Sample A (dense flow-interior facies) (A-D), Sample B (transitional facies)(E-H), Sample C (highly vesicular flow-top facies) (I-L). First row: Visible light photography showing macroscale textural features and vesicle distributions. Second row: Three-dimensional micro-CT reconstructions with segmented pore space (blue), the largest connected-component pore network, and corresponding extracted individual pore bodies derived from PNM. Third row: Vertical profiles through segmented volumes illustrating spatial distribution and connectivity of vesicular networks. Fourth row: High-resolution SEM micrographs characterizing vesicle wall thickness, inter-vesicle connectivity, and matrix microporosity. Note the progressively increasing matrix microporosity from Sample A through C, which contributes to overall network connectivity beyond the macro-vesicular structure alone and helps explain the expected superior hydraulic conductivity of Sample C relative to Sample B.
• Figure 5: Pore network topology and precipitation-induced clogging scenarios in sandstone versus basalt reservoirs. Top row: Comparison of pore and throat architecture showing fundamental differences in network connectivity. (a) Representative sandstone pore network with high coordination number (4--6 connections per pore), creating multiple parallel flow pathways with substantial redundancy that maintains permeability even when individual throats become blocked. (b) Vesicular basalt pore network displaying low coordination number (predominantly 2 connections per pore), creating serial or chain-like connectivity vulnerable to catastrophic permeability loss if critical throats are occluded. Despite larger throat diameters in basalts compared to sandstones, the low-redundancy topology renders basaltic networks more susceptible to precipitation-induced impairment. Bottom row: Precipitation scenarios and their differential impacts on permeability in basaltic pore networks. (c) Schematic representation of percolation pathways in highly vesicular basalt reconstructed from micro-CT imaging, illustrating the tenuous, predominantly serial connectivity characteristic of vesicular networks. (d) Scenario 1: Large, isolated carbonate precipitates (white masses) forming within two main flow percolation pathways with varying degrees of adverse consequences. (e) Scenario 2: Numerous small, distributed carbonate precipitates (white patches) systematically impact multiple pore throats throughout the network. This scenario causes severe permeability degradation despite modest porosity loss, as extensively studied in masoudi2024mineral.