Technical Report NTB 17-10

Redox evolution and Fe-bentoniteinteraction in the ABM2 experiment,Äspö Hard Rock Laboratory

In view of its favourable sealing properties the bentonite buffer represents an important com­po­nent in the multi-barrier concept of high-level waste repositories. These properties may be adversely influenced during the initial phase after repository closure when complex thermal-hydro-mechanical-chemical (THMC) processes occur. In particular, the effect of heat emanating from the decay of the waste and, in case steel materials are present, corrosion processes may have an adverse impact on the performance of the buffer. The latter are strongly influenced by the redox conditions which are oxic initially but will become reducing upon O2 depletion. The above-mentioned effects and underlying processes have not been studied in detail in the past.

To gain improved understanding on the performance of bentonite buffers, the Alternative Buffer Materials (ABM) test at the Äspö Hard Rock Laboratory (HRL), involving an international consortium and managed by SKB, was launched in 2006. The setup includes a series of boreholes in which various bentonite materials were emplaced as rings, stacked upon each other, and surrounding a carbon steel heater. This report documents work related to the ABM2 package which was exposed to temperatures up to 130° C and to artificial saturation for a period of five years. The experiment is considered to be an analogue for the buffer with regard to its thermal and redox evolution during the initial stage after repository closure. The realistic tem­peratures and smaller dimensions should lead to more pronounced heat- and corrosion-induced effects. Moreover, the exposure of different bentonite materials with different minera­logy, layer charges and Fe contents should give a more comprehensive view of the different pro­cesses involved.

The overall objective of the study was to gain improved understanding of the combined impact of heat and corrosion on the bentonite buffer performance in a repository-type setting. This should provide a more solid basis for constraining safety functions of the bentonite buffer during the repository lifetime. In particular the study was focused on the analysis of steel/bentonite interfaces which were exposed to realistic redox conditions with regard to the repository nearfield. An important corollary objective was the improvement of characterisation methods for Fe-bentonite interfaces.

A novel multi-method approach was developed and applied to unravel the redox evolution and the effects of temperature and corrosion on the different bentonite materials. A thorough proce­dure for maintaining anoxic conditions during sample preparation and analysis was adopted. Eight bentonite materials and 11 interface samples from ABM2 were analysed by scanning electron microscopy (SEM) coupled with energy dispersive x-ray spectroscopy (EDX), m-Raman spec­troscopy, x-ray diffraction (XRD), x-ray fluorescence spectrometry (XRF) and 57Fe Mössbauer spectrometry.

The main findings include:

  • The adopted methodology yielded an extensive and consistent dataset of different bentonite materials and enabled coherent interpretation in spite of complex events having occurred during the in situ test.
  • The identified processes are consistent with those reported in previous in situ tests at the Äspö HRL and the Grimsel test site. Moreover, the study enabled more detailed understanding regarding Fe-bentonite interaction in a repository-type setting.
  • Corrosion of the heating rod induced a Fe front extending ~ 5 – 20 mm into the adjacent bentonite, except for the high Fe-bentonite material (Rokle) where no Fe increase was observed. The main fraction of this Fe accumulation in the clay was found to be Fe(III) oxides. Also, some Fe(II) (not pertaining to the original material) was found whose nature, however, has not yet been identi­fied. This additional Fe(II) diffused further into the clay. The identified Fe speciation high­lights the important role of oxic corrosion, but also demonstrates that anaerobic corrosion occurred after depletion of O2 from the system. A phenomenological model for the evolution of corrosion and iron-bentonite interaction processes is proposed.
  • The Fe profiles in the clay could be adequately fitted by an empirical bimodal diffusion model. The approach and derived diffusion coefficients are consistent with those obtained by a previous study on Fe(II) in bentonite. The simple diffusion model, however, does not reflect the underlying mechanisms, which besides diffusion involve sorption, precipitation and redox reactions.
  • Cation exchange reactions between Na+, Ca2+, Mg2+ and K+ occurred both horizontally and vertically in the package by diffusive equilibration with the artificial Äspö water. The cation exchange capacity (CEC) displays almost no change as a function of distance to the iron and heat source.
  • Further prominent features are the accumulation of Mg and CaSO4 close to the heater. This was found to be most pronounced in the zone where an unintented and temporary boiling had occurred and led to rims of Mg sulphate and anhydrite. In the other zones, Mg also likely precipitated close to the heater which would explain the observed pattern of total and exchangeable Mg in the package.
  • No indications of smectite alteration in any of the samples containing different smectite types were found. This is supported by the near-to-constant CEC values and the constant Al/Si ratios towards the heater. This finding is in line with other in situ experiments and indicates the stability of smectite under harsh repository-type conditions.
  • Preliminary tests on the effects of reduction in montmorillonite on CEC and swelling pressure were conducted. A synthetic material (PGV-1) was used for this purpose. These tests indi­cated no effect on CEC upon reduction of structural Fe.