Often modellers want to try out new ideas, to investigate the impact of including a new process on their model. The flexibility of QPAC lends itself to rapid prototyping of this nature.
An example where this functionality has been exploited is in the modelling of iron-bentonite interactions, where, in order to capture complex mineral growth, the processes of nucleation, crystal growth, precursor cannibalisation and Ostwald ripening were added to a reactive-transport model. The presence of both iron canisters and bentonitic clay in some engineered barrier system (EBS) designs for the geological disposal of high-level radioactive waste creates the potential for chemical interactions which may impact upon the long-term performance of the clay as a barrier to radionuclide migration.
As part of the EC NF-PRO project, Quintessa developed a new model for the alteration of bentonite buffers in contact with corroding iron, with the aim of understanding their possible long-term transformation to non-swelling iron silicates. Natural systems evidence suggests that the sequence of alteration of clay by iron-rich fluids may proceed via an Ostwald step sequence, something that conventional thermodynamic models of mineral growth fail to replicate.
In order to address this, a model proposed by Steefel and Van Cappellen in 1990, which includes the processes of crystal nucleation, growth, precursor cannibalisation and Ostwald ripening, was coupled to a standard reactive-transport model. This has enabled investigation of a representative model of the alteration of bentonite in a typical EBS environment. Simulations using a conventional model predicted that berthierine would dominate the solid product assemblage, with siderite replacing it at simulation times greater than 10,000 years. Simulations with the alternative model showed a sequence of Fe-bearing solid alteration products, beginning with magnetite, then cronstedtite, then berthierine, and finally Fe-chlorite. Using plausible estimates of mineral-fluid interfacial free energies, chlorite growth is not achieved until 5000 years of simulation time. The results of this modelling work suggest that greater effort should be placed upon providing key data for iron silicates (e.g. kinetic data, solubilities, and mineral-fluid interfacial free energies), through a dedicated programme of laboratory experimental and natural analogue research.