Selenium redox speciation and coordination in high-burnup UO 2 fuel: Consequences for the release of 79 Se in a deep underground repository Enzo Curti a,⇑ , Annick Froideval-Zumbiehl a , Ines Günther-Leopold a , Matthias Martin a , Andrej Bullemer a , Hanspeter Linder a , Camelia N. Borca b , Daniel Grolimund b a Paul Scherrer Institut, Nuclear Energy and Safety Research Department, 5232 Villigen PSI, Switzerland b Paul Scherrer Institut, Synchrotron Radiation and Nanotechnology Research Department, Swiss Light Source, 5232 Villigen PSI, Switzerland article info Article history: Received 17 February 2014 Accepted 2 July 2014 Available online 10 July 2014 abstract The chemical state of Se in high-burnup UO 2 spent nuclear fuel (SNF) was investigated by micro X-ray absorption near-edge structure (l-XANES) spectroscopy. The data were first evaluated in terms of linear combination fits of reference compounds. In a second step, the XANES data were fitted to theoretical (ab initio) spectra via geometrical optimization of atomic clusters around a central Se atom, using the FDM- NES and FitIt software packages. Best fits were obtained assuming substitution of Se in occupied or vacant oxygen sites within the UO 2 lattice, with 20–25% local expansion. Based on these results and chemical arguments we argue that the long-lived fission product 79 Se may be stabilized to sparingly soluble Se(II) in SNF. This would explain the failure to detect dissolved Se in SNF leaching experiments. Ó 2014 Elsevier B.V. All rights reserved. 1. Introduction As a result of nuclear weapon proliferation issues and exceed- ingly high costs, many countries have abandoned or suspended reprocessing of spent nuclear fuel and switched to direct disposal as the main option for the final storage of high-level radioactive waste in deep underground repositories [1]. This means that the need for careful studies on spent fuel dissolution and radionuclide release to aqueous solutions is increasing. Used UO 2 and MOX fuels are highly heterogeneous materials. Therefore the release of radionuclides upon aqueous attack after the breaching of the waste containment cannot be modeled mech- anistically; a semi-empirical approach is needed. Usually, radionu- clide release from spent fuel in a deep groundwater-saturated repository is treated as a two-stage process, consisting of a short- term release pulse (‘‘Instant Release Fraction’’, shortly IRF) of easily soluble/volatile radionuclides (e.g. 129 I, 135 Cs, 36 Cl, 14 C) followed by much slower dissolution of the UO 2 or PuO 2 lattice (‘‘matrix disso- lution’’, MD), which is essentially an electrochemical process gov- erned by factors affecting the redox potential, e.g. the supply of radiolytic oxidants or the availability of reactive molecular hydro- gen [2–4]. The determination of nuclide-specific IRF-values has been and is still the subject of intensive studies [3,5]. IRF-values can be roughly estimated from reactor operation parameters such as linear power rate/burnup and measurements of fission gas release. However, the type of fuel assembly, as well as contributions from different material sources (e.g. 14 C contribution by Zircaloy), also play a major role. The situation is further complicated by the ten- dency of nuclear power plants to optimize energy production by increasing burnup limits. At burnup values exceeding 50 GWd/tU, microstructural changes occur in the fuel rim region, which may modify porosity and thus the diffusivity of segregated nuclides, affecting IRF values. In summary, for each nuclide of interest the IRF, understood as the sum of the rapid release of fission/activation products from fuel grain boundaries, fuel/sheath gap and Zircaloy corrosion layers, is a critical safety-assessment parameter resulting from a variety of material properties and in-reactor processes. The present contribution focuses on the fate of 79 Se, a long-lived safety-relevant nuclide (half-life 3.3  10 5 a) which is traditionally considered as a potentially significant IRF contributor owing to the appreciable volatility of selenium and the high solubility of oxi- dized Se species [6,7]. This view implicitly assumes selenium seg- regation toward grain boundaries and fuel/sheath gap followed by oxidation to Se(IV) or Se(VI), either during reactor operation or under wet geological storage conditions, e.g. via radiolytic oxida- tion. So far, no reliable experimental data are available on the chemical state of selenium in UO 2 spent fuel. Whereas Se(IV) and Se(VI) form soluble salts, the reduced forms Se(0) and Se(II) are sparingly soluble [6,7] and have thus limited mobility in aqueous solutions. Recent leaching experiments on high burnup UO 2 fuel from two Swiss power plants [5] failed to detect dissolved 79 Se after contact with aqueous solutions during http://dx.doi.org/10.1016/j.jnucmat.2014.07.003 0022-3115/Ó 2014 Elsevier B.V. All rights reserved. ⇑ Corresponding author. Tel.: +41 56 310 24 16; fax: +41 56 310 28 21. E-mail address: enzo.curti@psi.ch (E. Curti). Journal of Nuclear Materials 453 (2014) 98–106 Contents lists available at ScienceDirect Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat