Crystal formation during the vitrification of HLW in Ca/Zn base glass

This thesis investigates the incorporation and solubility of Post Operational Clean-Out (POCO) waste simulant into Ca/Zn borosilicate glass comprising mainly of MoO3, ZrO2 and BaO, where the first two oxides have a low solubility in borosilicate glass. Moving from MW borosilicate glass, used by the UK nuclear sector, into its modified version including extra Al2O3, CaO and ZnO (Ca/Zn glass) requires further research. Nevertheless, initial findings from NNL and this study shows that the formation of watersoluble yellow phase is avoided on tested waste loadings (25 and 38 wt%), which translated into a better chemical durability in aqueous environment. The poor solubility of Mo and Zr in the borosilicate glass is the main factor that controls the POCO waste loading during the vitrification of nuclear waste. To tackle the problem with formation of undesirable alkali molybdates, producing water-insoluble molybdates such as Ca and Ba molybdates was used as a feasible approach based on that already employed by the French nuclear industry to combat similar issues. Melting of POCO waste glasses has demonstrated a detrimental effect on Inconel 601 and Nicrofer 6025 HT utilised by the industry as the melter vessel materials. Thus, it was crucial to understand the corrosion mechanism to choose the best material for the melters. Hence the overall aims of the project were to  Understand the chemistry of the crystals likely to form throughout the vitrification of different types of POCO simulant in Ca/Zn base glass and to determine whether these crystal phases form in the melter or on cooling  Investigate how the viscosity affects any crystal phases formed and the role of crystalline phases on properties such as the chemical durability of the final products  Investigate the effect of molten waste glass compositions on the corrosion of Inconel 601 and Nicrofer 6025 HT coupons The viscosity of the modified POCO containing glasses (25 % waste loading) was lower in all cases compared to the unmodified counterparts. Additionally, the modified samples were much more crystalline, than the unmodified ones. However, the chemical durability in both cases was similar after 28-day tests, but the results after 7 day tests clearly exhibited that the modified glasses were more durable (PCT). Corrosion tests performed on Inconel 601 and Nicrofer 6025 HT showed that corrosive nature of the molten wasteAbstract glasses strictly depended on their composition, with the modified glasses demonstrating lower corrosiveness towards the examined metallic coupons, than the unmodified POCO waste products. Crystals identified by X-ray diffraction exhibited that samples with lower waste loading were almost amorphous while higher waste loadings resulted in intensive crystallisation, where for example CaMoO4 and ZrSiO4 phases were common. The mechanism responsible for phase separation in the glasses with excess molybdate was liquid-liquid separation and then crystallisation, which arises during cooling within glass melts containing elevated amounts of MoO3. The molybdate crystals, which are typically Ca/Ba molybdates are grouped into spherical shapes that are submicron in size and randomly distributed. Charge compensation mechanisms were examined by incorporation of Na2MoO4 into a Ca/Zn base glass with additions of CaO and Fe2O3 and showed that CaMoO4 can be produced. Mössbauer spectroscopy proved that majority of iron exists as Fe3+, giving its structural role as FeO4- that may interact with “free” Na+ which in turn strongly decreased the amount of Na2MoO4 or NaMo4O6 recrystallised after dissolution into glass. This experiment also showed that formation of water-soluble sodium molybdates is not inevitable and can be suppressed. The POCO based waste glasses consist of two groups called unmodified and modified ones. The modifications were aimed to improve meltability and uniformity of the final waste products. This was achieved by addition of CaO, CuO and Mn2O3 to the Ca/Zn base glass. Glass densities steadily increased with increasing incorporation of the POCO– based waste, but their values were similar. Glass transition temperatures Tg all decreased with increasing POCO loadings. PCT tests showed that durability results with both unmodified and modified POCO waste glasses after 28 day tests were comparable. Finally, incorporating Mo into durable Ca/Ba molybdates appears to be an effective method to immobilise this problematic element and avoiding the formation of yellow phase, which is totally undesired during vitrification. This approach is also supported by numerous papers from the French industry, where production of glass composites rich in insoluble molybdates of Ca/Ba has been identified as a feasible option.