Bentonite and iron metals are common materials proposed for use in deep-seated geological repositories for radioactive waste. The inevitable corrosion of iron leads to interaction processes with the clay which may affect the sealing properties of the bentonite backfill. The objective of the present study was to improve our understanding of this process by studying the interface between iron and compacted bentonite in a geological repository-type setting. Samples of MX-80 bentonite samples which had been exposed to an iron source and elevated temperatures (up to 115°C) for 2.5 y in an in situ experiment (termed ABM1) at the Äspö Hard Rock Laboratory, Sweden, were investigated by microscopic means, including scanning electron microscopy, μ-Raman spectroscopy, spatially resolved X-ray diffraction, and X-ray fluorescence.

The corrosion process led to the formation of a ~100 μm thick corrosion layer containing siderite, magnetite, some goethite, and lepidocrocite mixed with the montmorillonitic clay. Most of the corroded Fe occurred within a 10 mm-thick clay layer adjacent to the corrosion layer. An average corrosion depth of the steel of 22–35 μm and an average Fe2+ diffusivity of 1–2 × 10−13 m2/s were estimated based on the properties of the Fe-enriched clay layer. In that layer, the corrosion-derived Fe occurred predominantly in the clay matrix. The nature of this Fe could not be identified. No indications of clay transformation or newly formed clay phases were found. A slight enrichment of Mg close to the Fe—clay contact was observed. The formation of anhydrite and gypsum, and the dissolution of some SiO2 resulting from the temperature gradient in the in situ test, were also identified.

Bentonites are candidate materials for high-level radioactive waste (HLRW) repositories and, therefore, are investigated with respect to long-term stability. In order to identify possible bentonite alteration processes, long-term in situ tests are conducted in rock laboratories. The prototype repository in situ experiment (PR) is one of the best examples of this kind of test due to the size of the installation as well as the duration. In the present study, chemical and mineralogical alteration processes of the bentonite MX 80 after an 8 y heating period were investigated. The water content of all samples increased following inflowing Na-Ca-Cl-type granitic groundwater causing cation exchange in the bentonite buffer materials. Exchangeable magnesium was desorbed in the buffer and MgO concentration increased at the bentonite-Cu canister interface; the Mg sink could not be detected, however. CaO also accumulated at this interface mainly as Ca carbonate and Ca sulfate. Cu corrosion products were identified at the bentonite-canister interface by chemical analysis, scanning electron microscopy equipped with energy dispersive X-ray spectroscopy (SEM-EDX), and differential thermal analysis. Up to 0.5 mm into the bentonites Cu could be detected by SEM-EDX. No cristobalite dissolution was observed in contrast to other in situ tests in which iron heaters were used. The corrosion products and the lubricant which was added during manufacturing of the bentonite blocks were mixed with the bentonite at the bentonite-canister interface. A quantitative measure of that mixture was the decrease in the cation exchange capacity (CEC). The CEC also reduced in all other samples, however, compared to the CECs of the reference samples, particularly in the warmer deposition hole 5 compared to the colder deposition hole 6. Overall, the PR in situ experiment proved that cation exchange reactions occurred in full-scale bentonite buffer experiments in all bentonite blocks but structural degradation of smectite could not be identified.

Survey experiments that investigate how voting procedures affect voting behavior and election outcomes use hypothetical questions and non-representative samples. We present here the results of a novel survey experiment that addresses both concerns. First, the winning party in our experiment receives a donation to its campaign funds inducing real consequences for voting. Second, we run an online experiment with a Dutch national representative sample (N = 1240). Our results validate previous findings using a representative sample, in particular that approval voting leads to a higher concentration in votes for smaller parties and strengthens centrist parties in comparison to plurality voting. Importantly, our results suggest that voting behavior is not affected by voting incentives and can be equally reliably elicited with hypothetical questions.