Controlling subsurface flow is crucial to prevent: the spreading of contaminants, the failure of dikes or any undesirable water flow. Subsurface flow can be controlled by changing the soil permeability. In order to change soil permeability rigid structures are often installed (for example the installation of sheet pile walls into dike bodies), however, these solutions are costly and disrupt the environment. Alternatively, the SoSEAL project aims to provide an in-situ solution to reduce the permeability of highly permeable soil layers in an ecological friendly and cost-effective way. The SoSEAL project is inspired by podzolization, which is a soil formation process. It reduces soil permeability via the precipitation of metal-organic matter complexes in the pore space. During earlier practices a need has arisen for a cheap, environmentally friendly and naturally available aluminum source. Accordingly, the goal of this research is to investigate if gibbsite can be used as the aluminum source to form metal-organic matter precipitates.

A literature study was conducted to investigate the interplay between gibbsite and dissolved organic matter (DOM). Four processes are primarily studied in this research, they are: the release of free Al via dissolution of gibbsite; the complexation between free Al and DOM; the protonation and subsequently precipitation of DOM and the adsorption of H⁺/DOM onto the gibbsite surface. Thereafter the following hypotheses were formulated:

•The extent of gibbsite dissolution is larger at low pH. However, due to its low solubility and slow dissolution kinetics, a low concentration of free Al is expected.
•The organic matter (OM) source used in this study is a humic acid, therefore it is expected to become completely insoluble when the pH is below two.
•The contradicting pH dependency of gibbsite dissolution and OM solubility leads a narrow transitional pH range, that is favourable for the occurrence of complexation between free Al and DOM.

These hypotheses are tested by performing experiments in the laboratory, where the pH varies from 2 to 8 and the aluminum/carbon (mol/mol) ratio is 0; 0.05; 0.1; 0.3 and 1. A synthesized crystalline gibbsite powder and a potassium humate are used.
\bigskip

The results are subdivided into three pH ranges in which similar behaviour was found. In the low pH range (pH = 2 - 2.5), the protonation and subsequently precipitation of DOM is the dominant process in short term. However, when considering the long term behaviour, gibbsite dissolution is the dominant process as the DOM has precipitated. Due to the difference in time scale, there is very little complexation between free Al and DOM. In the transitional pH range (pH = 3 - 3.5), the competition for H⁺ between DOM and gibbsite is profound. This increases the solubility of DOM and slows the dissolution of gibbsite down. As a consequence, this favours the gradual complexation between free Al and DOM. Nevertheless, due to the ambiguity of free Al measurement method used in this research, it is difficult to quantify the complexation. In neutral to high pH range (pH = 4 - 8) DOM undergoes protonation over time but stays soluble. The release of free Al via the dissolution of gibbsite is negligible in this pH range. Therefore complexation between free Al and DOM should not be expected.

The results presented in this research indicate that gibbsite may not be suitable as the aluminum source to form precipitates with organic matter in engineering solutions that require fast soil permeability reduction. However, the slow release of free Al from gibbsite in combination with the gradual formation of precipitates with organic matter might be interesting for robust problems that require a self-healing ability. For efficiency concern, another OM source that is rich in fulvic acids (higher solubility at low pH than humic acid) and a natural gibbsite source (higher solubility than synthesized gibbsite) should be investigated.

The maintenance and improvement of river embankments plays a crucial role in granting the safety and ensuring the quality of life of the inhabitants of the Netherlands. While the threat posed by storm surges has been greatly reduced since the implementation of the Deltaworks, riverine areas in the west of the country are characterised by the highest risk of flooding.
The use of relief wells is one of the viable solutions to improve the safety of river embankments against instability mechanisms such as uplifting. Throughout the Schoonhovenseveer-Langerak project (SLA), the implementation of a system of relief wells constituted one of the key measures to increase the safety of the embankment structure during high-water events of the river Lek. In the current study, the influence of relief wells was simulated employing a 3D coupled model built with the Finite Element software PLAXIS 3D. Using soil properties from former site investigations, a calibration was performed based on in situ pore water pressure measurements, in order to devise adequate boundary conditions. The safety assessment of the structure was then performed by applying a time-dependent high-water wave and employing strength reduction to yield a Factor of Safety. An intermediate 2D model was employed to examine the evolution in time of the Factor of Safety, the influence of mesh refinement on the safety results, and the shape and the location of the failure mechanism. By means of the 3D model, different design choices of the system of relief wells were tested and their influence on the stability of the structure and on the water discharge could be determined. A comparison between the results obtained in the 2D and 3D models showed that a 3D stability analysis does not necessarily yield a less conservative result than the 2D equivalent.
Comparing the values of the Factor of Safety obtained with time-dependent and stationary boundary conditions showed that the results obtained in the latter case are considerably more conservative, suggesting the potential benefits of a time-dependent fully-coupled analysis.