Navigation and service

PGI-1 Seminar: Dr. Piotr Kowalski

How challenging is it to compute actinides? - Atomistic modelling of nuclear materials relevant for nuclear waste management

22 Apr 2015 11:30
PGI Lecture Hall

IEK-6, FZ Jülich


Being an unavoidable byproduct of nuclear technology applications, radioactive waste is generated by a number of industrial processes, including, but not limited to, the production of electricity. The majority of management strategies propose direct disposal of nuclear waste in deep geological repositories. One technological option is the immobilization of radionuclides within solid matrices that exhibit properties important for a long-time isolation of actinides. The main requirements are that they should remain stable in a contact with an aqueous fluid and resist radiation damage. Monazite- and pyrochlore-type ceramics are of particular interest, as actinide-bearing natural analogues of these materials remained stable through the geological time. However, before such materials could be used as matrices for the nuclear waste disposal, their thermodynamic properties must be well characterized and their behavior under repository conditions should be well understood. At the Jülich Research Centre (FZJ-IEK-6) and in partner institutions there exist an ongoing experimental research devoted to characterization of such novel waste forms. We contribute to this research by atomistic simulation of various material properties, by delivering information that is often difficult to obtain experimentally. However, methodological aspects of computation of f-elements are also challenging.

With modern computational resources it is generally feasible to simulate properties of chemically complex materials using density functional theory (DFT). Unfortunately, DFT often fails for strongly correlated f-materials, by predicting thermochemical parameters with unacceptably large errors. Moreover, wrong electronic states are often predicted for simple actinide-bearing solids. Thus, it has been proposed that more sophisticated, but computationally intensive methods of computational quantum chemistry, such as hybrid functionals, MP2 or CCSD, should be always used for a reliable simulation of actinide-bearing materials. However, this strategy would prevent meaningful research of complex systems. Therefore, we are looking for a computationally cheap extension of DFT that improves the description of strong electronic correlations and allows a feasible computation of complex materials. DFT+U satisfies these criteria. The novelty of our approach, which permits a successful application of this method to actinide-bearing materials, consists in derivation of the Hubbard U parameter using ab initio approaches, namely, the linear response method (cLDA) and the constrained random phase approximation (cRPA). Here we demonstrate the success of this approach by benchmarking the enthalpies of reactions that involve actinide-bearing molecular complexes and solids. We show that the value of the Hubbard U parameter strongly depends on the oxidation state of an actinide. In similar studies of monazite- and pyrochlore-type ceramics we obtained excellent results on structures, thermochemical parameters and energetics of defect formation. We will show the follow up estimation of excess properties of relevant solid solutions and heat capacities of these materials. Ultimately, all this information is very important for the assessment of the thermodynamic stability of these novel waste forms.


Prof. Dr. Stefan Blügel
Phone: +49 2461 61-4249
Fax: +49 2461 61-2850
email: s.bluegel@fz-juelich.de