The understanding of the release of volatile radionuclides in the atmosphere and the environment is of crucial interest for radioprotection issues. The characterization of the forms and quantities of the volatile products is difficult to achieve experimentally because of the high radiotoxicity of these elements. Alternatively, the radionuclide behavior can be modeled at various stages of their release with chemical thermodynamic and kinetic models such as the ASTEC code developed by IRSN (French Institute for Radiological Protection and Nuclear Safety). These models rely on the use of accurate thermodynamic and kinetic data, which can be obtained from highly correlated relativistic quantum chemical calculations methods, provided these are capable of properly describing the very complex electronic structure of the volatile species (oxides, hydroxide, halogen complexes etc) containing fission products such as transition metals.

Objectives
It is precisely because of the complexity of their electronic structure that an accurate description of these species and their reactivity remains a challenge to existing methods, not only as a result of the treatment of a large number of strongly correlated electrons but also due to the very large number of quasi-degenerate low-lying electronic states arising from the unpaired valence p, d, or f electrons¹. This thesis therefore focuses on exploring two novel, computationally efficient, electron correlation methods in the field of heavy element chemistry. The first is the Density Matrix Renormalization Group Algorithm (DMRG), which offers the possibility of treating large number of correlated electrons, has been first applied to heavy elements by our Canadian collaborators ². We plan to assess its accuracy to volatile radionuclides such as ruthenium, uranium and plutonium complexes. The second approach relies on the use of geminals for achieving high accuracy for thermodynamics quantities at a low computational cost ³.

¹ A. S. P. Gomes, F. Réal, B. Schimmelpfennig, U. Wahlgren and V. Vallet, in Computational methods in lanthanide and actinide chemistry, ed. M. Dolg, Wiley, 2015.
² P. Tecmer, K. Boguslawski, O. Legeza and M. Reiher, Phys. Chem. Chem. Phys., 2014, 16, 719–727.

Requirements
The successful candidate will have a background in (physical) chemistry, physics, applied mathematics or engineering. Prior experience with molecular modeling, computational physics or chemistry is appreciated. Knowledge of, or willingness to learn computer programming (in least one of the following languages: Python, C/C++, Fortran) is a prerequisite. The thesis is setup in the frame of a French-Canadian cotutelle program, implying that the thesis student will spend one third of her/his thesis at McMaster University and the other two thirds in Lille.

Period
3 years: October 2015 – 2018…