Structural role of molybdenum in nuclear glasses: an EXAFS study

https://doi.org/10.1016/S0022-3115(03)00277-0Get rights and content

Abstract

The Mo environment has been investigated in inactive nuclear glasses using extended X-ray absorption spectroscopy (XAS). Mo is present in a tetrahedron coordinated to oxygen in the form of molybdate groups [MoO4]2− (d(Mo–O)=1.78 Å). This surrounding is not affected by the presence of noble metal phases in the nuclear glass. Relying on the XAS results, on the bond-valence model and on molecular dynamics simulations of a simplified borosilicate model glass, we show that these groups are not directly linked to the borosilicate network but rather located within alkali and alkaline-earth rich domains in the glass. This specific location in the glass network is a way to understand the low solubility of Mo in glasses melted under oxidizing conditions. It also explains the possible phase separation of a yellow phase enriched in alkali molybdates in molten nuclear glasses or the nucleation of calcium molybdates during thermal aging of these glasses. Boron coordination changes in the molten and the glassy states may explain the difference in the composition of the crystalline molybdates, as they exert a direct influence on the activity of alkalis in borosilicate glasses and melts.

Introduction

Borosilicate glasses are used to immobilize fission products and transuranic elements coming from the ‘light water’ high-level nuclear wastes. These glasses contain typically more than 30 elements, among which fission products such as Mo, which is known to be a difficult element to incorporate in nuclear waste glasses. By contrast to most fission products, Mo has a low solubility in borosilicate glasses, typically below 1 wt% of MoO3 in nuclear glasses under usual elaboration conditions [1]. Above this concentration, it can participate to the formation of a yellow phase in the glass. This complex yellow phase is formed during the melting stage of the process by a phase separation of molten salts from the borosilicate melt and may be responsible for an enhanced corrosion of the glass melter. This phase contains molybdates, such as Cs-molybdate, which may concentrate 137Cs, and chromates, which impart a yellow coloration to the resulting crystalline compounds observed after cooling [2]. On account of the solubility of alkali molybdates in water, this phase separation may decrease the chemical durability of the vitrified nuclear waste material [1], [3]. Mo is also known as a nucleating agent and is used in the glass-ceramics industry for its ability to control the phase separation of glasses or as a crystallization catalyst to produce glass-ceramics [4]. However, by contrast to other crystalline phases such as noble metal or spinel precipitates, which have been quantified in nuclear glasses using Rietveld refinements [5], such a phase separation only occurs in the inactive French glass (SON 68) when the elaboration conditions are out of the specification domain. In addition, experiments on the long-term thermal stability of nuclear waste glasses show the nucleation of molybdates such as powellite, CaMoO4. The presence of crystalline molybdates may arise from the fact that, under oxidizing conditions, hexavalent Mo is the major oxidation state of molybdenum in silicate and borosilicate glasses [2], [3], [6], although other oxidation states may be present, such as Mo(III), Mo(IV) and Mo(V), under more reducing conditions [2], [3], [7]. However, the low solubility of molybdenum in nuclear glasses has not been explained, despite its importance during the encapsulation processing of nuclear waste.

In this paper, we present new data on the environment of molybdenum in inactive nuclear borosilicate glasses, using Mo–K edge extended X-ray absorption fine structure spectroscopy (EXAFS). These data indicate molybdenum to occur in a hexavalent oxidation state, with the formation of regular molybdate tetrahedra. The absence of connectivity between molybdate groups (MoO4)2− and the polymerized borosilicate sublattice is explained by the location of Mo in regions enriched in alkali and alkaline-earth cations. This structural position explains the possible phase separation of alkali molybdate-bearing yellow phases in molten nuclear glasses or the nucleation of calcium molybdate in these heated glasses, if the process is made under atmospheric conditions [3]. We also show that the composition of crystalline molybdates directly arises from structural modifications which occur either in the molten or the glassy state, as a result of the heterogeneous glass structure.

Section snippets

Experimental procedure

Inactive glasses representative of the French light water nuclear glass were studied. Due to the possible influence of Pd and RuO2 precipitates on the glass structure, a glass without Pd and RuO2 (GL-0) was compared to glasses with 1.5 and 3 wt% Pd + RuO2, labeled GL-1.5 and GL-3, respectively [3], [9]. GL-1.5 has the chemical composition of the inactive reference French light water nuclear glass (SON 68). The compositions of the studied glasses are given in Table 1. The glasses were elaborated

EXAFS data

For the three studied samples, the EXAFS χ(k) signals present the same shape, intensity and frequency within the experimental uncertainty (Fig. 1). This shows that the presence of noble metals, that otherwise modifies the medium-range organization around Si in these glasses [8], has no significant effect on the Mo environment. We will thus concentrate on the GL-1.5 sample (SON 68), which will be representative of the three studied glasses. The EXAFS signal (Fig. 1) consists of a unique damped

Conclusions

Mo–K edge EXAFS spectroscopy in inactive nuclear glasses indicates that Mo forms [MoO4]2− tetrahedra, which are not directly connected to the glassy network. A location within alkali- and alkaline earth-rich sublattices of the glass favors the charge compensation of the molybdate groups. This peculiar structural position explains the low Mo-solubility, the separation of Mo-rich phases in the melts and the nucleation of molybdates in nuclear glasses. It is also at the origin of the well-known

Acknowledgements

The authors would like to thank the team of LURE, Orsay, France and Aline Ramos for fruitful discussions. The authors also thank COGEMA for its financial support. This is IPGP contribution 1942.

References (22)

  • L. Galoisy et al.

    Mat. Res. Soc. Symp. Proc. XXI

    (1998)
    L. Galoisy et al.

    J. Amer. Ceram. Soc.

    (1999)
  • G.N. Greaves et al.

    Phys. Rev. B

    (1995)
  • W. Lutze

    Radioactive Wasteforms for the Future

    (1988)
  • A. Horneber et al.

    Mat. Res. Soc. Symp. Proc.

    (1982)
    F.A. Lifanov, S.V. Stefanovsky, T.N. Lashtchenova, O.A. Knyazev, O.V. Tolstova, S.V. Chizhevskaya, in: Waste...
  • Y. Kawamoto et al.

    J. Am. Ceram. Soc.

    (1981)
  • P.W. McMillan et al.
    (1979)
  • L. Galoisy et al.

    J. Mater. Res.

    (1998)
  • N. Sawaguchi et al.

    Phys. Chem. Glasses

    (1996)
  • G. Calas, Progress in experimental petrology, N.E.R.C. Report 18 (1981)...
  • D. Cabaret et al.

    J. Non-Cryst. Solids

    (2001)
  • M. Puyou et al.

    Nucl. Techn.

    (1995)
  • Cited by (120)

    View all citing articles on Scopus
    View full text