One of the major challenges of in situ applied ground improvement techniques like Microbial and Chemical Induced Calcite Precipitation (MICP and CICP) is the homogeneous or at least spatially controlled, distribution of the desired reaction products in order to obtain controlled improvement of geotechnical bulk soil properties (Redding, 2007; Van Paassen, 2009 Numerical modelling simulations at continuum (Darcy) scale are performed to study the spatial distribution of reaction product(s) as a function of the injection strategy. An example of the simulations is shown in the figure below. The models are validated with laboratory experiments using a quasi two dimensional flow box with a large number of ports which could be either used as injection/extraction well or as a sensor port. This article presents the results obtained when multiple injection wells are operated simultaneously in which two reactive solutions (calcium chloride and sodium (bi-) carbonate) are injected separately in alternating wells located perpendicular to the background flow. Water is used as a non-reactive spacer. Hydraulic pressures and electrical resistivity tomography (ERT) are used in the laboratory set-up to monitor the spatial distribution of reactants and products. The shape and location of the mixing zone during treatment, and the spatial distribution of calcium carbonate after treatement are evaluated for different injection strategies.

Microbial induced carbonate precipitation (MICP) has been confirmed to effectively reinforce granular soil by employing biological processes. Biological denitrification is one of these processes, in which the microbes use nitrate to oxidize organic matter and produce nitrogen gas and inorganic carbon, which results in precipitation of calcium carbonate in the presence of calcium ions Accordingly, MICP via denitrification may stabilize the soil in multiple ways. Besides the strengthening effect of calcium carbonate precipitate crystals, the induced gas phase may enhance the undrained response as the compressible gas may dampen dynamic loads on the soil. The effectiveness of this method has been investigated on different sand types using the triaxial test set-up, at different pressure conditions. During the reaction phase, the water displacement into the back pressure controller was used to monitor gas production. After reaction samples were flushed from bottom to top at a constant head difference. The flow rate was measured to determine permeability. Results of these experiments showed that the volume of produced gas strongly depend on the pressure conditions. The gas storage capacity ranged from 22% to 40% of the total pore volume, depending on grain size. At a pore pressure of 1 bar, the gas storage capacity was exceeded before the reaction was completed and excess gas escaped to the back pressure controller. Gas which remained inside the sample was easily removed during flushing, and the permeability was hardly affected. At higher pore pressure or smaller grain size, the gas phase was well distributed and resulted in a decrease of the permeability after the treatment by 50 to 300%. The drained shear strength of a single batch treatment was not significantly improved, indicating that the amount of calcium carbonate precipitation was not sufficient and suggesting multiple treatments are required to bear an increase of strength. Upon undranied loading the small strain stiffness was significantly improved, but the peak undrained strength was reduced. The results indicate that controlling the gas formation is important in order to optimize this method for field applications