Thermodynamic modelling of ultra-long-term durability of cementitious binders for waste immobilisation

• Main Author: Prentice, Dale
• Other Authors: Provis, John ; Bernal, Susan ; Bankhead, Mark ; Hayes, Martin
• Published: University of Sheffield 2018
• Subjects: 620
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.766561

Treatment of intermediate-level waste (ILW) generated as a by-product from nuclear power in the UK requires a long-term strategy to safely dispose of the waste. Encapsulation of ILW in a cement matrix is the current UK methodology, followed by storing the waste for potentially thousands of years in geological disposal facilities (GDFs). Understanding of the cement phase assemblage is key to predicting how these cements will behave in the long term. Thermodynamic modelling of cement hydrate phases is a powerful tool which can be used to predict the effects of cement hydration. This thesis investigates the quality of thermodynamic modelling to predict stable phase assemblages of blast furnace slag-Portland cement (BFS-PC) cements, representing UK nuclear industry practice, under conditions that are expected during the storage of encapsulated ILW. Three BFS-PC ratios (1:1, 3:1 and 9:1) were tested at different curing ages to determine the degree of hydration of the precursor materials to use as input parameters for thermodynamic modelling. Characterisation of the phase assemblages were compared to the thermodynamic modelling results to assess the robustness of the modelling approach. A solid solution model for C(-A)-S-H was used to explicitly incorporate aluminium into the C-S-H phase to more accurately portray the chemical structure in the BFS-PC system. Thermodynamic modelling was capable of accurately simulating the change in phase assemblage as curing time increased. Variation of precursor materials was effectively modelled. Temperature fluctuations are expected to occur within the GDF once the waste is stored within it. BFS-PC samples were cured for one year at 35 °C followed by periods of curing at 50 °C, 60 °C and 80 °C. Major phase changes were not observed until the curing temperature reached 60 °C, whereby hemicarbonate and ettringite destabilised. At a curing temperature of 80 °C, the sulphate and carbonate AFm and AFt phases were not observed in cement phase assemblages, however siliceous hydrogarnet was present. Two thermodynamic modelling approaches were used to simulate the effects of temperature change. It was determined that the thermodynamic simulation should not contain siliceous hydrogarnet when simulating BFS-PC hydration up to 60 °C but should contain siliceous hydrogarnet for higher temperatures. The Pitzer model used as a means to produce activity coefficients, was compared with the generalised dominant electrolyte activity model, Truesdell-Jones, to assess whether modelling of cement phases may be improved. A large ion-interaction parameter database was required to use the Pitzer model for simulating cement hydration. Solubility studies of cement phases and cement pore solution data were used as a means to compare the activity coefficient models. The more complex nature of the Pitzer model caused the simulations to require runtimes up to 18 times more than the Truesdell-Jones method. The pore solution of the BFS-PC systems was compared with the predictions from the activity coefficient models, which determined that the Pitzer model provided minimal improvement over the Truesdell-Jones method. However, the Pitzer model proved more effective for simulating higher concentration systems, therefore, the Pitzer model may be required in future modelling projects when simulating concentrated groundwater interactions with the cement wasteforms.