 4 NH) and bicarbonate are the dominant products of hydrolysis, see Equation 1 (Mitchell et al., 2019). However, carbon dioxide in aqueous solutions occurs as carbonic acid (H 2 CO 3), bicarbonate ( 3 HCO), or carbonate ( 2 3 CO), depending on the pH value. Since ammonia acts as a weak base by taking up a proton and producing hydroxide, it increases the pH value and shifts the equilibrium toward carbonate ions. The additional presence of calcium ions, in our case provided by adding calcium chloride, forces calcium carbonate to precipitate. According to van Paassen (2009), the release of a proton (H +) during the calcium carbonate precipitation buffers the production of hydroxide during the hydrolysis, see Equation 2. On the pore scale, precipitated calcium carbonate leads to changes in pore morphology, and on a larger scale, after averaging, this corresponds to changes in the effective quantities porosity and permeability,

Enzymatically induced calcium carbonate precipitation is a promising geotechnique with the potential, for example, to seal leakage pathways in the subsurface or to stabilize soils. Precipitation of calcium carbonate in a porous medium reduces the porosity and, consequently, the permeability. With pseudo-2D microfluidic experiments, including pressure monitoring and, for visualization, optical microscopy and X-ray computed tomography, pore-space alterations were reliably related to corresponding hydraulic responses. The study comprises six experiments with two different pore structures, a simple, quasi-1D structure, and a 2D structure. Using a continuous injection strategy with either constant or step-wise reduced flow rates, we identified key mechanisms that significantly influence the relationship between porosity and permeability. In the quasi-1D structure, the location of precipitates is more relevant to the hydraulic response (pressure gradients) than the overall porosity change. In the quasi-2D structure, this is different, because flow can bypass locally clogged regions, thus leading to steadier porosity–permeability relationships. Moreover, in quasi-2D systems, during continuous injection, preferential flow paths can evolve and remain open. Classical porosity–permeability power-law relationships with constant exponents cannot adequately describe this phenomenon. We furthermore observed coexistence and transformation of different polymorphs of calcium carbonate, namely amorphous calcium carbonate, vaterite, and calcite and discuss their influence on the observed development of preferential flow paths. This has so far not been accounted for in the state-of-the-art approaches for porosity–permeability relationships during calcium carbonate precipitation in porous media.

This study aims to experimentally investigate the possibility of combining two extended continuum theories for two‐phase flow. One of these theories considers interfacial area as a separate state variable, and the other explicitly discriminates between connected and disconnected phases. This combination enhances our potential to effectively model the apparent hysteresis, which generally dominates two‐phase flow. Using optical microscopy, we perform microfluidic experiments in quasi‐2D artificial porous media for various cyclic displacement processes and boundary conditions. Specifically for a number of sequential drainage processes, with detailed image (post‐)processing, pore‐scale parameters such as the interfacial area between the phases (wetting, non‐wetting, and solid), and local capillary pressure, as well as macroscopic parameters like saturation, are estimated. We show that discriminating between connected and disconnected clusters and the concept of the interfacial area as a separate state variable can be an appropriate way of modeling hysteresis in a two‐phase flow scheme. The drainage datasets of capillary pressure, saturation, and specific interfacial area, are plotted as a surface, given by f (Pc, sw, awn) = 0. These surfaces accommodate all data points within a reasonable experimental error, irrespective of the boundary conditions, as long as the corresponding liquid is connected to its inlet. However, this concept also shows signs of reduced efficiency as a modeling approach in datasets gathered through combining experiments with higher volumetric fluxes. We attribute this observation to the effect of the porous medium geometry on the phase distribution. This yields further elaboration, in which this speculation is thoroughly studied and analyzed.

No abstract