Summary and conclusion

Model assessment of reductive dechlorination as a remediation technology for contaminant sources in fractured clay: Modeling tool

The main propose of this project is to have a better understanding of anaerobic reductive dechlorination in fractured clay till. Hence the main processes occurring in such a system have to be identified and characterized. Furthermore an assessment of the clean-up times associated with using reductive dechlorination as a remediation technology has to be performed. To complete these tasks, a model for transport and reductive dechlorination of TCE in fractured clay till is developed. This model considers the counter diffusion of the contaminant from the matrix clay into the fractures, in which the contaminant is transported by advection/dispersion. This model focuses on the vertical transport of TCE from the source zone located in the clay till into the underlying aquifer, therefore only the vertical fractures are taken into account. Furthermore TCE is assumed to be present only in the dissolved phase, as we consider the late time scenarios long after contamination. In order to better characterize the different processes controlling this system, degradation and transport are first modeled separately with two “sub-models”.

The first model is a mathematical model of reductive dechlorination based on Monod kinetics and including competitive inhibition between the chlorinated solvents and the growth and decay of two dechlorinating biomass populations. Several processes, such as limiting substrate condition or fermentation, are disregarding in order to simplify the model and reduce the input parameters. This model is calibrated and verified with two sets of microcosm laboratory experiments. The most sensitive parameters are fitted to one set of experimental data and the model is validated using the second set. The fitting procedure determined the values of a set parameters to simulate sequential reductive dechlorination of TCE to ethene.

The second model is a simple model of diffusive transport in the clay matrix, including sorption processes. This model is tested on data from a core sample taken at a field site where reductive dechlorination was enhanced with injection of both bacteria and substrate. A typical diffusive profile from the matrix to the fracture is observed, and a reaction zone limited at the fracture/matrix interface can be observed, which suggests that dechlorination takes place both in the fracture and in the matrix.

The two “sub-models” are combined to set-up the main numerical model of a single fracture – clay matrix system. In this model the network of vertical fractures is assumed to have a periodic structure allowing the system to be described by a half-matrix/half-fracture unit. The transport equations describe diffusion/sorption in the 2D-matrix and advection/dispersion in the 1D-fracture. Assuming a very low hydraulic conductivity of the clay matrix, advection in this media is neglected.

Four different degradation scenarios are considered depending on the degradation location: no degradation, degradation in the fracture only, degradation in the fracture and a reaction zone, and finally degradation in the whole system. The model results from the two first scenarios are very similar because the residence time in the fracture is much smaller than the degradation time. In contrast the contaminant flux is more rapidly reduced when assuming degradation in the matrix. Hence the cleanup time (time to remove 90% of the initial contaminant mass) is significantly reduced (from 200 without degradation to 120 and 60 years).

The sensitivity analysis performed with the model shows that matrix porosity, sorption coefficient, net recharge, and the fracture spacing are the most sensitive parameters. The model is not very sensitive to the fracture aperture or the longitudinal dispersivity in the fracture. Furthermore it is observed that a peak in the contaminant flux occurs long after the beginning of the remediation in the case where degradation takes place in the matrix. This peak is explained by the higher diffusion coefficients and lower sorption coefficients of DCE and VC compared with TCE, resulting in a faster transport of the daughter products from the matrix into the fracture.

The simulated contaminant flux from this single fracture – clay matrix model can be used as input data for a simple 2D cross-section of the underlying aquifer, in order to assess groundwater impaction.