Using Reactive Transport Modeling to Evaluate the Source Term at Yucca Mountain

Abstract

The conventional approach of source-term evaluation for performance assessment of nuclear waste repositories uses speciation-solubility modeling tools and assumes pure phases of radioelements control their solubility. This assumption may not reflect reality, as most radioelements (except for U) may not form their own pure phases. As a result, solubility limits predicted using the conventional approach are several orders of magnitude higher then the concentrations of radioelements measured in spent fuel dissolution experiments. This paper presents the author's attempt of using a non-conventional approach to evaluate source term of radionuclide release for Yucca Mountain. Based on the general reactive-transport code AREST-CT, a model for spent fuel dissolution and secondary phase precipitation has been constructed. The model accounts for both equilibrium and kinetic reactions. Its predictions have been compared against laboratory experiments and natural analogues. It is found that without calibrations, the simulated results match laboratory and field observations very well in many aspects. More important is the fact that no contradictions between them have been found. This provides confidence in the predictive power of the model. Based on the concept of Np incorporated into uranyl minerals, the model not only predicts a lower Np source-term than that given by conventional Np solubility models, but also produces results which are consistent with laboratory measurements and observations. Moreover, two hypotheses, whether Np enters tertiary uranyl minerals or not, have been tested by comparing model predictions against laboratory observations, the results favor the former. It is concluded that this non-conventional approach of source term evaluation not only eliminates over-conservatism in conventional solubility approach to some extent, but also gives a realistic representation of the system of interest, which is a prerequisite for truly understanding the long-term behavior of the proposed repository. Therefore, it is a very promising alternative approach for source-term evaluation.