Phosphate solubility and its effects on US radioactive waste glass properties

The Hanford Site in Washington State, USA is a decommissioned nuclear production site, home to 250 000 m3 of radioactive waste which is stored in 158 steel tanks, initially only for short term use and now subject to extensive and widespread corrosion problems. Numerous tanks are leaking radioactive ‘sludge’ containing a complex mixture of many elements of the periodic table. In an effort to clean up the waste, vitrification implementing a borosilicate glass composition was selected in 1989 as a means of safer disposal. However, previous research has identified that the final vitrified waste form may still be susceptible to leaching, leading to the potential release of radionuclides and other toxic components into groundwater in geological repositories. One potentially problematic component of the waste arises as a result of pre-processing: phosphorus compounds. Phosphorus pentoxide (P2O5) is poorly soluble in borosilicate glass systems, with concentrations >3.2 mol% (>4.5 wt%) potentially leading to phase separation that can affect melt viscosity and chemical durability. These are of particular importance because an overly viscous melt will be unpourable and may damage the melter if the increase in viscosity is due to the presence of crystallites and/or salt accumulation; and poor chemical durability could result in the leaching of radionuclides within geological repositories, ultimately resulting in groundwater contamination. Hence, it is important to establish the effects of different levels of P2O5 on the properties of radioactive waste borosilicate glasses. This research investigated the effects of doping and altering the P2O5 content on the structure and properties of borosilicate glass compositions based on Hanford high level waste (HLW), low level activity waste (LAW), and immobilised low level activity waste (ILAW). X-ray fluorescence (XRF) spectroscopy and Inductively Coupled Plasma Optical Emission spectroscopy (ICP-OES) were used to analyse the chemical composition. X-ray diffraction (XRD) was used for crystalline phase identification and to confirm the amorphous nature of the glasses, where crystalline phases were quantified using Rietveld analysis and visually imaged using scanning electron microscopy (SEM). Glass structure was analysed using Raman spectroscopy, 57Fe Mössbauer spectroscopy, X-ray absorption near edge structure (XANES) spectroscopy, and magic angle spinning nuclear magnetic resonance (MAS-NMR) spectroscopy. The rheological behaviour of melts was investigated using high temperature viscosity (HTV) measurements, whilst glass chemical durability was analysed using the 7-day product consistency test (PCT) method B, and the thermal properties were determined using differential thermal analysis (DTA) and dilatometry. The main findings were that simple sodium borosilicate glasses had a solubility limit of between 2.7 ± 0.02 and 4.0 ± 0.01 mol% P2O5. When glass compositions were further complicated with additions of other oxides to more accurately represent Hanford HLW and LAW glass waste forms, the solubility limits of P2O5 were lower: between 0.69 ± 0.03 and 1.19 ± 0.00 mol% P2O5 for the HLW glasses, and 1.86 ± 0.00 and 2.04 ± 0.02 mol% P2O5 for LAW glasses. When P2O5 was added to a ILAW simulant glass its solubility was >0.71 ± 0.03 mol%. When the solubility limit was exceeded in these glass systems, nucleation-type phase separation and crystallisation occurred, producing phosphate-based crystal phases. Whilst the additions of P2O5 did not seem to increase the solubility of other components, in particular sulphate (SO42-), it did lead to the suppression of crystallisation in the heat treated HLW glasses. The structural analysis of the simple sodium borosilicate glasses doped with P2O5 provided an insight into the effects of P2O5 whilst its behaviour could be inferred in more complicated glasses. P2O5 initially entered the borosilicate glass structure as ultraphosphate (Q3P) but changed to orthophosphate (Q0P), pyrophosphate (Q1P), and metaphosphate (Q2P) as a function of the P2O5 concentration, where the lack of change in the boron (3+) tetrahedral ion, boron (3+) trigonal ion ([3]B3+:[4]B3+) ratio with increasing P2O5 content suggested that the phosphorus (P5+) was potentially entering the [4]B sub-network, but where its oxygen (O2-) was supplied by the silicate network, causing the non-bridging oxygens (NBØs) on the silicon (Si4+) to transform into bridging oxygens (BØs) resulting in its re-polymerisation, as evidenced by the decrease in the Q3Si and increase in Q4Si units. These glass systems provided a guide to the maximum P2O5 contents that can be present in the vitrified waste to avoid any potentially undesirable effects of phase separation and / or crystallisation. Insights into the structural changes and crystallisation behaviour of these glass systems, caused by altering their compositions, can be used to help inform crystallisation discriminators (parameters determining crystallisation tolerances) used to limit the waste loading of glasses prepared from Hanford wastes with high P2O5 contents.