Understanding the mechanisms by which borosilicate glasses corrode in contact with aqueous solutions remains a challenge to the safety case for the geological disposal of vitrified high-level nuclear waste. Here, lithium isotope fingerprinting techniques were applied to the leachates of a simulant Magnox waste glass to probe the mechanisms of aqueous corrosion at both short and long timescales (6 hours to 464 days). Experiments took place at 40 and 90 °C to assess the consistency of the dissolution mechanisms across a range of commonly employed temperatures and the legitimacy of applying higher temperature experimental datasets to understand glass corrosion within a disposal facility at lower temperatures.
Two competing release mechanisms were observed for lithium (diffusion and hydrolysis), and the relative proportions of these mechanisms changed through time. Leachates initially had lower δ⁷Li values than the pristine glass (-2.7‰ at 40 °C and -1.1‰ at 90 °C relative to the pristine glass) at both temperatures due to lithium leaching incongruently through diffusive processes. The greater offset between solution and solid at lower temperatures indicates a larger rate of diffusion (incongruent dissolution) relative to the rate of hydrolysis (congruent dissolution) at lower temperatures. The fraction of lithium released through diffusion relative to the fraction of lithium released through hydrolysis then increased at both temperatures with time up to 126 days, increasing from 0.47 and 0.22 at 6 hours to 0.66 and 0.41 at 126 days at 40 and 90 °C respectively. Subsequently, the fractions of lithium released through diffusion sharply decreased to 0.36 at 40 °C and 0.22 at 90 °C after 464 days, consistent with network hydrolysis coupled with secondary phase precipitation later controlling the long-term release of Li at both temperatures. Throughout the duration of the experiments (464 days) the δ⁷Li values in solution increased to 9.0‰ at 40 °C and 10.0‰ at 90 °C due to the formation of talc and montmorillonite phases at 40 °C and additional smectite phases at 90 °C. Further, no evidence for the formation of a diffusive barrier to the transport of lithium within the alteration layers became apparent during the later stages of dissolution at either temperature. However, the fraction of lithium leached through diffusion was still significant throughout all stages of dissolution. Lithium isotope ratios in solution were correlated with the transition from a system which was increasingly dominated by lithium diffusion as the dissolution rate slowed to one which was controlled by hydrolysis coupled with secondary phase precipitation at long durations. Alongside elemental ratios in solution, these results were consistent with the same set of mechanisms governing dissolution across the temperature range studied.