Sodium nitrate (NaNO3) which is contained in high-level liquid waste (HLLW) is one of the key components for vitrification of the waste since it may react with Mo in cold cap to form water soluble secondary phase, so called, yellow phase (YP). The major component of YP is Na2MoO4. In this reason, understanding and control of the reactions of NaNO3 is important for production of qualified waste glass. Using a glass frit that have high specific surface area such as powder is a promising method to prevent the formation of yellow phase [1]. The reason of the effect of this prevention might be because the NaNO3 more preferably reacts with the frit than Mo, and the formation of Na2MoO4 is precluded. However, detail reactions among NaNO3-frit-Mo in the cold cap is not fully elucidated. In this study, reactions of NaNO3 in the cold cap were examined in terms of formation of water-soluble Mo at respective temperatures, denitration temperature of NaNO3 and reactions in NaNO3-frit binary system.

1. Formation of water-soluble Mo compounds in a temperature range of 400-650˚C
Crucible scale batch experiments were performed to examine the fraction of Mo to become water-soluble compounds (=YP) in a temperature range of 400-650˚C. Two frits of powder (<64 μm) and bead (mean diameter: 2 mm) were used to compare the fraction. Simulant of HLLW containing 25 elements including NaNO3 was dried out at 250˚C and mixed with the frit. The mixture was then heated at respective temperatures for 3 hours. The heated specimen was contacted with pure water to leach water-soluble Mo and Na.
When the powder frit was used, the fraction of Mo leached in water increased with increase in the temperature below 550˚C and then decreased at the above. The fraction was only 4% at 650˚C (Fig.1). On the other hand, the fraction continuously increased and reached to 86% at 650˚C when the bead frit was used. This result clearly showed the advantage of using frit with high specific surface area. Leached fraction of Na decreased monotonically with increase in temperature. This decrease indicates that NaNO3 reacted and dissolved into the frit to become insoluble state. Since the molar ratio of Na in the effluent was 10.4 times higher than that of Mo, Na that reacted with Mo was estimated to be 10% in maximum when the powder frit was used. The remaining approximately 90% of Na successfully reacted with frit. In the case of the bead frit that have smaller surface area, the relative proportion of NaNO3 reacting with Mo became increased.
Time evolution of the leached fraction of Mo was examined at 500 and 650˚C (Fig.2). Even for 650˚C, nearly half of the Mo once converted to water-soluble compound and then reconverted to insoluble one. Heating at higher temperature showed faster conversion and reconversion kinetics and smaller final fractions of water-soluble Mo. X-ray diffraction analysis (XRD) of the heated specimens revealed that the water-soluble compounds was Na2MoO4·2H2O in most cases.
Through these experiments, it is considered that NaNO3 reacts with glass or Mo depending on the contact probability during denitration. Mo which reacted with Na forms a water-soluble transient compound. However, at a later stage, Na in this Mo compound would dissolve in the frit while Mo would combine with Ca in the frit or rare earths in the simulant to reconvert into more stable insoluble compounds.

2. Denitration temperature of NaNO3 in multicomponent system
In the cold cap, most of the components once form nitrates when the liquid waste is dried out, and then denitration of these nitrates proceeds with increasing temperature. Since NaNO3 is assumed to denitrate when reacting with frit or Mo, the reaction temperature could be estimated from the denitration temperature. NaNO3 is one of the components that is denitrated at the highest temperatures. Thus, we have examined the denitration completion temperature of frit powder-HLLW simulant mixture by thermogravimetry with varying waste components in the simulant. When the simulant included only Na and Mo, denitration was completed around 650 ˚C (Fig.3). On the other hand, in case of the simulant composed of 25 elements including Na, the denitration temperature decreased down to 561˚C. Addition of elements individually to the simulant revealed that transition elements, especially Mn (tried as a surrogate of Tc), contributed the most for this decrease of the denitration temperature (Fig.4). This result indicated that Mn had a catalytic function to facilitate the denitration of NaNO3. However, this acceleration of the denitration did not provide positive effect on YP suppression. The amount of water-soluble Mo rather increased at 500°C due to accelerated denitration and reaction between Mo and NaNO3. While at 600°C, there was no difference in the amount of water-soluble Mo because the denitration of NaNO3 promoted regardless of the addition of Mn.

3. Reactions in NaNO3-glass frit binary system
Detail reactions between NaNO3 and frit were examined under simplified NaNO3-frit binary system. Mixture of NaNO3 and frit powder was heated at 400, 500, 550 and 600°C for 3 hours, and then quenched and rinsed with water to remove unreacted NaNO3. From XRD, sodalite including NO3 in its lattice, Li2SiO3 and NaNO3 were detected in the specimens heated at 550°C or higher. The detected NaNO3 might have been encapsulated in the fused frit and remained in the specimen even after the rinsing.
Analysis of the released gases during heating of these specimens up to 1000°C revealed that NO was continuously released at temperatures below 660°C, and then spike-like release of NO was observed with foaming of the glass at temperatures above 660°C (Fig.5). The higher the fabrication temperature of the specimens, the more NO was released. The amount of NO released from the specimen was larger than the amounts of total nitrates detected by the XRD (Fig.6). These results might indicate that some NaNO3 once dissolved in the glass to become amorphas, and then denitrated and released NO with glass foaming at temperatures above 660°C.

Acknowledgements
This work was carried out as a part of the basic research programs of vitrification technology for waste volume reduction（JPJ010599）supported by the Ministry of Economy, Trade and Industry, Japan.

Reference
[1] K. Uruga, T. Tsukada, T. Usami, J. Nucl. Sci. Technol., 57(4), 433-443 (2020).