Effect of oxide impurities on the dissolution behavior of Th⁴⁺, Be²⁺ and U⁴⁺ in fluoride salts

Abstract

Oxides are one of the most important impurities in the fuel salt of molten salt reactors (MSRs), and excessive oxide impurities pose a risk to the safe operation of MSRs. This study focused on investigating the precipitation behavior between Th⁴⁺, U⁴⁺, and Be²⁺ with O²⁻ in the 2LiF–BeF₂ (FLiBe) eutectic salt system. The results showed that the solubility of UO₂ was 5.52 × 10⁻³ mol kg⁻¹, and the solubility product (K_sp) of UO₂ was 6.14 × 10⁻⁷ mol³ kg⁻³ in FLiBe salt at 650 °C. It was also found that the O²⁻ ion would firstly react with U⁴⁺ to form UO₂, and then the excessive O²⁻ would react with Be²⁺ to generate BeO in the FLiBe system. Despite conducting the solubility experiment of ThO₂ and titration experiment of FLiBe–ThF₄, the system failed to achieve the solubility and the K_sp of ThO₂. The main reason for this was that O²⁻ preferentially reacted with Be²⁺ over Th⁴⁺ to form precipitates, in other words, Be²⁺ exerted a protective effect against Th⁴⁺. Above all, this work experimentally demonstrated that in the FLiBe system, O²⁻ preferentially combines with U⁴⁺ to form a precipitate, followed by Be²⁺, while Th⁴⁺ is relatively inert.

---

Introduction

Reducing the use of fossil fuels seems to be an effective way to reduce CO₂ emissions and mitigate global warming, and many countries have chosen to develop sustainable nuclear power to partially replace traditional energy production.¹ Consequently, significant efforts have been made to develop nuclear energy, and one of the current efforts is to develop the fourth-generation advanced nuclear energy systems, such as molten salt reactors (MSRs). The characteristic of MSRs is the use of nuclear fuel in liquid form.^2–4 The fissile material (UF₄) and fertile material (ThF₄) are dissolved in the molten carrier salt, and several alternative fuel carrier salts are being considered, but the leading candidate is 2LiF–BeF₂ (FLiBe).^5–8 Using liquid fuel will give MSRs an excellent advantage in performing continuous on-line fuel purification. On-line refueling significantly reduces the costs associated with fuel fabrication. More importantly, the ability to perform on-line fuel processing makes MSR suitable for the use of thorium.

In molten salt reactors, the purities are critical for controlling of chemical problems in the molten salt and mitigating corrosion. Among them, oxide impurities are a particularly important contaminant in this regard, because they can cause serious problems, such as fuel deposition and corrosion of structural materials.^9–12 In addition to the impurities already present in the raw constituents of molten salt, partial oxide impurities are generated during reactor operation and maintenance. Water and oxygen are entrained in the gas phase of the reactor system, which are the major external sources of oxide impurities in MSRs; neutron irradiation of the fuel salt generates oxygen in the system during operation;¹³ and some oxides can penetrate into the MSRs fuel from passive layers formed on the surface of the structural materials.¹⁴ Although the oxide contamination in MSR is usually removed by sparging the molten salts with a mixture of H₂ and HF to control the oxygen content of the molten salt within an acceptable range,^15–19 after the molten salt is loaded into the reactor, oxide impurities will still be introduced slowly in the manner described above. Previous studies have shown that UF₄ in the fuel salt readily reacts with O²⁻ ions to form uranium oxide precipitates, as shown in eqn (2).^20,21 The formation of uranium precipitates would reduce the concentration of U⁴⁺ and cause the loss of fuel, what's more, the large accumulation of UO₂ precipitates on the structural metal surface of the primary circuit would create the overheated areas in the reactor, posing a potential threat to the reactor operation.

H₂O + 2F⁻ = 2HF + O²⁻ (1)

UF₄ + 2O²⁻ = UO₂↓ + 4F⁻ (2)

For the safety of the molten salt reactor, it is necessary to strictly control the concentration of O²⁻ in the fuel salt. In addition, the molten salt oxide chemistry of this research will open the door to a deeper understanding of the precipitation and dissolution behavior of oxides in fluoride salts, which plays an important role in the chemical control of molten salt systems.

Studies of the dissolution and precipitation behaviors of UO₂ in fluoride salt have a long history, not least because of its importance in the safety of molten salt reactors. However, the existing studies showed a large scatter in the solubility product (K_sp) of UO₂, and the desired data are often not available due to the inappropriate experimental conditions such as the temperature and composition of molten salt. Peng et al.^20,22 have studied the precipitation of UO₂ in LiF–BeF₂ and LiF–NaF–KF fluoride salt systems at 600 °C, using electrochemical methods. Their study showed that the K_sp of UO₂ in the LiF–NaF–KF system was 4.75 × 10⁻⁶ mol³ kg⁻³, while in the LiF–BeF₂ system, it was 1.67 × 10⁻⁵ mol³ kg⁻³. There was an order of magnitude difference in the K_sp of UO₂ between the two systems. Oak Ridge National Laboratory (ORNL)²³ studied the K_sp of UO₂ in LiF–BeF₂–ThF₄–UF₄ at 500 °C, which was 2.60 × 10⁻⁷ mol³ kg⁻³. ORNL²⁴ has studied the system in detail and found that the K_sp of UO₂ at 600 °C in the LiF–BeF₂–ZrF₄ system was 1.20 × 10⁻⁵ mol³ kg⁻³. In LiF–BeF₂–UF₄ with the addition of 5% ZrF₄, Toth²⁵ found that Zr⁴⁺ had a greater binding interaction for O²⁻ than that of U⁴⁺. This finding has led to an increased interest in studying the interaction between Zr⁴⁺ and O^2–, as Zr⁴⁺ can effectively integrate with O²⁻. This property, combined with its small neutron absorption cross-section (thermal neutron absorption cross-section of 0.18 bar), has even led several researchers to advocate ZrF₄ as an ideal oxygen binding additive to prevent the precipitation of UO₂ in the fuel salt. Korenko²¹ found that the introduction of oxide in the LiF–NaF–KF–UF₄–ZrF₄ system can generate ZrO₂ precipitation, which effectively consumed the O²⁻ in the molten salt and prevented the precipitation of UO₂. Studies of the LiF–BeF₂–ZrF₄–UF₄ system by Peng²⁶ showed that ZrO₂ formed initially as an O²⁻ addition of less than 1 mol kg⁻¹, with UO₂ and ZrO₂ co-precipitating at O²⁻ additions greater than 1 mol kg⁻¹. It implied that when oxide contamination exists in the melt containing both ZrF₄ and UF₄, a significant amount of ZrF₄ could inhibit the UO₂ formation. Later, Song and co-workers²⁷ found that in the system of LiF–BeF₂–ZrF₄–UF₄ system, when the molar ratio of Zr⁴⁺ to U⁴⁺ was greater than 4, only ZrO₂ precipitation formed after the quantitative addition of Li₂O. Conversely, ZrO₂ and UO₂ co-precipitated when the molar ratio was less than 4.

Thorium has been acknowledged as a marvelous resource for its potential application as nuclear fuel, it has been considered as a possible supplement or even a replacement for uranium. Therefore, it is crucial to study the chemical behavior of Th⁴⁺ in thorium-based molten salt reactors, as well as that of U⁴⁺. Chamelot et al.²⁸ reported that the addition of oxides to LiF–CaF₂–ThF₄ produced the intermediate product ThOF₂, and gradually formed the ThO₂ with more O²⁻ added. Wang²⁹ used Raman spectroscopy to study the addition of Li₂O to FLiBe/FLiNaK–ThF₄ and found that there was stabilized Th₂OF₁₀⁴⁻ anion with linear Th–O–Th geometry formed in both systems. Some related research on thorium utilization in molten salt reactors has been published over the years, despite the numerous investigations on the products and structures of thorium compounds, less effort has been dedicated to the precipitation and dissolution behavior of Th⁴⁺.

The influence of O²⁻ ions on Th⁴⁺ and U⁴⁺ in the molten salt reactor has been proven to be significant. Researchers have attempted to establish an association between the precipitates of U⁴⁺, Th⁴⁺, and O²⁻. It is necessary to have an understanding about the solubility and solubility product (K_sp) of ThO₂ and UO₂. Though a few studies have been performed for the solubility and K_sp of UO₂, the results obtained are inconclusive. Furthermore, to our knowledge, there is limited research about solubility and K_sp of ThO₂ has been performed so far. Therefore, this study concentrates on the interaction between O²⁻ and U⁴⁺, Th⁴⁺, as well as the precipitation and dissolution behavior of UO₂ and ThO₂ in the FLiBe system. This work aims to enhance knowledge of the oxide chemistry behavior in the FLiBe system.

Experiments

Solubility experiments of UO₂, ThO₂

The UO₂ and ThO₂ (applied by North China Nuclear Fuel Components Co., Ltd) powders were compressed into tablets, as shown in Fig. 1(a) and (b), and positioned at the bottom of the crucible respectively. Next, 20 g of FLiBe salt was used to cover the tablets. To prevent any interference with sampling, a piece of nickel mesh was placed between the FLiBe molten salt and UO₂/ThO₂ tablet to avoid powder diffusion. The UO₂ or ThO₂ was immersed in the FLiBe salt bath at 650 °C to create a saturated solution, which is illustrated in Fig. 2. Samples were taken hourly during the initial seven hours and subsequently every two hours. The samples were analyzed to determine the concentration of U⁴⁺ and Th⁴⁺ by ICP-OES, while the concentration of O²⁻ by LECO oxygen analyzer and ion chromatography. Based on the concentrations of U⁴⁺/Th⁴⁺ and O²⁻ in salt samples, the solubility, as well as the solubility product (K_sp) can be calculated.

Fig. 1 (a) UO₂ after pressing; (b) ThO₂ after pressing.

Fig. 2 Schematic diagram of solubility experiment of UO₂/ThO₂.

Preparation of FLiBe–ThF₄

In order to rapidly dissolve the ThF₄ (supplied by Changchun Institute of Chemical Engineering, with a purity of 99.9%) into the fluoride salt, it is necessary to synthesize a molten mixture of LiF and ThF₄ in advance. 10.2 g of ThF₄ was mixed with 2.8 g of LiF and melted at a temperature of 700 °C to prepare LiF–ThF₄(FLiTh). The FLiTh was then mixed with FLiBe salt and heated to ensure that the ThF₄ uniformly dissolved into the FLiBe salt. The concentration of O²⁻ within the resulting FLiBe–ThF₄ mixture was determined to be 1.80 × 10⁻² mol kg⁻¹.

Titration experiments of UF₄/ThF₄ with Li₂O in FLiBe

To investigate the reactions of U⁴⁺ and Th⁴⁺ with O²⁻ in the FLiBe system and obtain the corresponding values of K_sp, the titration experiments of UF₄/ThF₄ in FLiBe salt were carried out. In the titration experiment of FLiBe-UF₄, 1 g of UF₄ (supplied by China National Nuclear Corporation, Baotou No. 202 Plant, 99.9% purity) was mixed with 19 g of FLiBe molten salt in a crucible; in the titration experiment of FLiBe–ThF₄, 20 g of FLiBe–ThF₄ was used. O²⁻ ions were added to the melts in the form of Li₂O. After the O²⁻ was added, the salt mixture was stirred to ensure complete reaction of O²⁻ ions with U⁴⁺/Th⁴⁺ ions in the salt (the schematic diagram is shown in Fig. 3). The temperature was maintained at 650 °C to allow the molten salt to settle, during the process, the supernatant was periodical extracted via a quartz tube for the concentration analysis of Th⁴⁺, U⁴⁺ and O²⁻.

Fig. 3 Schematic diagram of FLiBe–UF₄/ThF₄ titration experiment.

Chemical analysis

The obtained samples were dissolved and diluted, and then the concentrations of cations (U⁴⁺, Th⁴⁺) in the molten salts were determined by ICP-OES (PerkinElmer Co., Ltd.). It was based on the premise that substances form a high-temperature plasma in a high-frequency electromagnetic field, so that the constituent elements of the compounds were generated into their excited states at high temperatures or by electric excitation, and that specific wavelengths of the electromagnetic spectral lines will be emitted when the elemental excited states were restored to the ground state, thus realizing qualitative or quantitative analyses of different samples.

The total concentration of O²⁻ in the experiment was measured with the LECO oxygen detector (LECO RO600, LECO Co., Ltd.). The obtained sample was ground into powder and weighed precisely, with approximately 0.05 g sample being wrapped in a tin capsule and placed in a graphite crucible. During the test process, the molten salt sample quickly went through the feeding platform into the graphite crucible, and then switched the current to heat the crucible to 2700 °C in a short period. At the high temperature, the molten salt samples were melted, causing a reaction between the oxygen contained and carbon from the graphite crucible and generating CO and CO₂ gases. The generated gases entered the infrared cell and were detected using infrared absorption. The computer assimilated the detection signals of CO and CO₂ and subsequently calculated the total oxygen content value of the sample.

Weighed out about 0.1 g of sample into a PE bottle and added deionized water. The bottle was then put into an ultrasonic cleaner to extract the ions to be measured. Finally, the obtained liquid was quantitatively diluted and the oxygenated ions (NO₃⁻, PO₄³⁻, SO₄²⁻) in the molten salt were determined using an ion chromatograph (ICS-2100, Dionex Co., Ltd).

Results and discussion

Content analysis of oxygen impurities in FLiBe salt

While the previous experiments and theoretical studies have focused on the oxygen impurity in fluoride salt, these results suggested that the oxygen impurities in the salt are composed of oxygen ions (O²⁻) and oxygenated anions (including NO₃⁻, PO₄³⁻, SO₄²⁻ etc.).^22,30 During reactor operation and maintenance, the concentration of O²⁻ ions within the fuel system would increase gradually, which was attributed to moisture and oxygen leaking into the reactor and the corrosion of the passive layers of the structural materials. Different from O²⁻ ions, the oxygenated anions can be maintained at a relatively stable level, as oxygenated impurities were mainly introduced along with the carrier salt, LiF, BeF₂ and UF_4, etc.

The total oxygen content in the FLiBe molten salt was measured using a LECO oxygen analyzer. The oxygenated anions in molten salt were determined by ion chromatography. The practical O²⁻ concentration ([O²⁻]) in molten fluorides was determined by subtracting the oxide concentration in oxygenated anions ([O]_IC) from the total oxide concentration ([O]_LECO), as shown in eqn (3). The measurement results of the total oxide concentration and the oxide concentration in oxygenated anions of FLiBe raw materials were listed in Table 1. The average total oxide concentration in the FLiBe molten salt was up to 8.56 × 10⁻³ mol kg⁻¹, as most of the oxygen and water contamination products were removed by sparging the salt with a mixture of H₂ and HF. The main oxygenated anions were found to be NO₃⁻, PO₄³⁻, and SO₄²⁻. Based on the mass proportion of oxide in the oxygenated anion, the oxide concentrations in NO₃⁻, SO₄²⁻, and PO₄³⁻ were determined to have oxide concentrations of 1.56 × 10⁻³, 2.13 × 10⁻³, and 3.56 × 10⁻³ mol kg⁻¹, respectively. Thus, the resulting oxide concentrations in these three oxygenated anions were up to 7.25 × 10⁻³ mol kg⁻¹, and the O²⁻ content is then calculated to be 1.31 × 10⁻³ mol kg⁻¹ using eqn (3).

[O²⁻] = [O]_LECO − [O]_IC (3)

Table 1 Oxygen content in oxygenated anions and FLiBe molten salt

Total oxygen content of FLiBe, mol kg^−1	Oxygen content of oxygenated anions, mol kg^−1	Total oxygen content of oxygenated anions, mol kg^−1	O^2− content, mol kg^−1
8.56 × 10^−3	NO3^− 1.56 × 10^−3	7.25 × 10^−3	1.31 × 10^−3
	PO4^3− 2.13 × 10^−3
	SO4^2− 3.56 × 10^−3

---

Solubility of UO₂ in FLiBe salt

Since UO₂ has been proven to be the main precipitate of U⁴⁺ ions in the fuel system,²⁰ the key data of the solubility of UO₂ in FLiBe salt is of great significance for controlling the oxide content of the fuel salt. In this study, the solubility of UO₂ in FLiBe salt at a temperature of 650 °C was measured using the static dissolution method. UO₂ tablet were used in these dissolution experiments. Fig. 4 shows the changes of U⁴⁺ and O²⁻ concentrations in FLiBe salt as a function of dissolution time, where U⁴⁺ content in molten salt was regarded as the amount of UO₂ dissolved. It can be seen that the U⁴⁺ concentration kept increasing with an increase in dissolution time. Finally, as the dissolution time up to 7 hours, the [U⁴⁺] reached a final plateau with a constant concentration, thus indicating that 7 h was the dissolution equilibrium time. Furthermore, the solubility of UO₂ was calculated to be 5.52 × 10⁻³ mol kg⁻¹. It should be noted that, as shown in Fig. 4, there are no significant fluctuations in the O²⁻ concentration profile as dissolution periods greater than 7 hours. The average oxygen content was approximately 1.06 × 10⁻² mol kg⁻¹, which is twice the U⁴⁺ equilibrium concentration, consistent with the U/O molar ratio in UO₂. Based on the concentration of O²⁻ and U⁴⁺ in FLiBe salt, the solubility product of uranium and oxygen ions was determined to be 6.14 × 10⁻⁷ mol³ kg⁻³, as shown in Table 2.

Fig. 4 Variation of U⁴⁺ and O²⁻ concentration with time at 650 °C.

Table 2 K_sp of UO₂ obtained from solubility experiment

	U^4+, mol kg^−1	Oxygen content, mol kg^−1	, mol^3 kg^−3	Average value
1	5.31 × 10^−3	1.11 × 10^−2	6.54 × 10^−7	6.14 × 10^−7
2	5.57 × 10^−3	1.09 × 10^−2	6.07 × 10^−7
3	5.67 × 10^−3	9.97 × 10^−3	5.64 × 10^−7
4	5.52 × 10^−3	1.07 × 10^−2	6.32 × 10^−7

---

The titration of UF₄ with Li₂O in molten FLiBe

Understanding the impact of O²⁻ on the behavior of U⁴⁺ in fluoride salt not only has fundamental interests but also is important for reactor operation and maintenance. The K_sp of UO₂ is one of the most important parameters in the oxide chemistry of fuel salt, indicating the interaction between U⁴⁺ and O²⁻ ions. This study aimed to investigate the K_sp of UO₂ with different oxygen content by titrating UF₄ with Li₂O in FLiBe salt. Three sets of titration experiments were conducted with Li₂O/UF₄ ratios(molar ratio) of about 1 : 1, 1 : 2, and 1 : 3, which were called O1–U, O2–U and O3–U in this work, respectively. The equilibrium concentrations of U⁴⁺ and O²⁻ ions in molten FLiBe salt after each titration operation are given in Table 3. According to the data in Table 3, it was found that with an increase in the addition of Li₂O from 0.157 mol kg⁻¹ to 0.470 mol kg⁻¹, the equilibrium concentrations of U⁴⁺ decreased sharply, while the equilibrium concentration of O²⁻ accordingly increased from 9.31 × 10⁻³ mol kg⁻¹ to 2.13 × 10⁻² mol kg⁻¹. Thus, based on the given data, the solubility product (K_sp) of U⁴⁺ and O²⁻ can be calculated to be 3.80 × 10⁻⁶ mol³ kg⁻³. This value was slightly greater than the K_sp obtained in the UO₂ solubility experiment. This may be attributed to some un-deposited UO₂ particles formed in the titration experiments, and these particles were picked up with the fluoride salt in the sampling process. The K_sp of UO₂ obtained from the experiment is compared with that in the literature, the results are shown in Table 4.

Table 3 K_sp of different ratios of UO₂ (initial U⁴⁺ was 0.159 mol kg⁻¹)

	Initial O^2−, mol kg^−1	Equilibrium U^4+, mol kg^−1	Equilibrium O^2−, mol kg^−1	Ksp, mol^3 kg^−3	Average value
O1–U	0.157	7.51 × 10^−2	9.31 × 10^−3	6.51 × 10^−6	3.80 × 10^−6
O2–U	0.316	9.61 × 10^−3	1.87 × 10^−2	3.36 × 10^−6
O3–U	0.470	3.34 × 10^−3	2.13 × 10^−2	1.52 × 10^−6

---

Table 4 K_sp of UO₂ in different molten salt systems

Molten salt system	Temperature/°C	Ksp/mol^3 kg^−3	Experimental condition
LiF–BeF2	600	1.67 × 10^−5	Titration experiment
LiF–NaF–KF	600	4.75 × 10^−6	Titration experiment
LiF–BeF2–ThF4–UF4	500	2.6 × 10^−7	Molten salt reactor
LiF–BeF2–ZrF4	600	1.2 × 10^−5	Molten salt reactor
LiF–BeF2	650	6.14 × 10^−7	Solubility experiment
LiF–BeF2	650	3.80 × 10^−6	Titration experiment

---

As seen in Table 3, there was a clear difference between the initial addition value and the equilibrium value of U⁴⁺ and O²⁻ in the melts. The concentration variation of the two ions in these titration experiments is shown in Fig. 5. Inspected in more detail, it was revealed that the variation in concentration of the Δ[U⁴⁺]/Δ[O²⁻] ratio in the O1–U and O2–U experiments was about 1 : 2, consistent with the U/O molar ratio in UO₂. Fig. 6 shows the photographs of solidified FLiBe–UF₄ salts after the titration experiments. It was evident that distinct reddish-brown precipitate layers appeared at the bottom of the salts. Meanwhile, with an increase in the addition of Li₂O, the thickness of the precipitate layer increased, and the upper green molten salt gradually discolored. However, the results obtained in the O3–U experiment demonstrated that the Δ[U⁴⁺]/Δ[O²⁻] ratio was approximately 1 : 2.9, which was different from the U/O molar ratio in UO₂. It was found that the Δ[U⁴⁺]/Δ[O²⁻] ratio remained at 1 : 2.9 (shown in Table 5) upon repetition of the O3–U experiment.

Fig. 5 Variation of U⁴⁺, O²⁻ concentrations for experiments with different ratios.

Fig. 6 (a) FLiBe-UF₄; (b) molten salt of O1–U; (c) molten salt of O2–U; (d) molten salt of O3–U.

Table 5 Concentration changes of U⁴⁺ and O²⁻ in the repeated experiment of O3–U

	Starting	Equilibrium	Variations	U^4+:O^2−percentage change
U^4+ mol kg^−1	0.158	3.10 × 10^−3	0.155	1 : 2.9
O^2− mol kg^−1	0.469	2.35 × 10^−2	0.446

---

XRD analysis was performed on the reddish-brown precipitate after most of the FLiBe salt was removed (shown in Fig. 7). The diffraction peaks on the XRD pattern indicated that all the precipitate layers in these titration experiments contain a large amount of UO₂. Later, these precipitates were analyzed by Raman spectrometer, and the analysis result (Fig. 8) showed that UO₂ was not the only generated precipitate in the O3–U experiment. Another insoluble oxide BeO³¹ would form after most U⁴⁺ ions were precipitated due to the successive addition of Li₂O. Such a result may explain the distinct deviation of the Δ[U⁴⁺]/Δ[O²⁻] ratio in the O3–U experiment.

Fig. 7 (a) XRD analysis of the precipitate in O1–U; (b) XRD analysis of the precipitate in O2–U; (c) XRD analysis of the precipitate in O3–U.

Fig. 8 Raman analysis of the precipitation in O3–U.

Solubility of ThO₂ in FLiBe salt

In MSRs using thorium, the deposition behavior of Th⁴⁺ is worthy of attention. Previous research have indicated that the Th⁴⁺ oxide precipitates in the form of ThO₂.³⁰ In order to study the precipitation and dissolution behavior of ThO₂ in the FLiBe system, the solubility of ThO₂ in molten FLiBe salt at 650 °C was studied by the static dissolution method. Fig. 9 presents the concentrations of Th⁴⁺ and O²⁻ in the molten salt at different times. The concentration of Th⁴⁺ increased from 1.24 × 10⁻³ mol kg⁻¹ to 0.297 mol kg⁻¹, and did not reach dissolution equilibrium after 18 hours. However, the [O²⁻] fluctuated within a small range of 0.03 mol kg⁻¹ after being dissolved for more than 10 hours, which was much less than the concentration of Th⁴⁺. The variation of concentrations was not only inconsistent with the Th/O molar ratio in ThO₂ but also different from that of the UO₂ solubility experiment completely. Due to the continuous dissolution of Th⁴⁺ in FLiBe without reaching equilibrium during the experiment, the solubility and K_sp of ThO₂ could not be determined.

Fig. 9 Concentration changes of Th⁴⁺ and O²⁻ at 650 °C.

To explore why ThO₂ fails to reach its dissolution equilibrium just as UO₂ does, XRD analysis was conducted on the insoluble substance at the bottom of molten salt after removing the FLiBe salt by distillation. As shown in Fig. 10, it can be observed that BeO and ThO₂ coexisted. It was speculated that the O²⁻ generated from dissolved ThO₂ would react with Be²⁺, which promoted the dissolution of ThO₂.

Fig. 10 Bottom precipitation XRD spectra of FLiBe–ThO₂ salt.

The titration of ThF₄ with Li₂O in molten FLiBe

It failed to get the K_sp of ThO₂ from the solubility experiment, which was extremely important to control the oxygen in molten salt, so the K_sp of ThO₂ was intended to be measured using the same method as the UO₂ titration experiment mentioned above. Two sets of titration experiments were conducted with Li₂O/ThF₄ ratio of about 1 : 1 and 1 : 2, denoted by O1–Th, O2–Th. Fig. 11 presents the initial and the equilibrium concentrations of Th⁴⁺, O²⁻ in the molten salt, and shows a significant difference from the UO₂ titration experiment. As shown in this figure, with the addition of Li₂O, there was no obvious change in the concentration of Th⁴⁺, while the [O²⁻] decreased unexpectedly. The result suggested that the introduction of O²⁻ may react with other cations, leading to the dissolution behavior of Th⁴⁺ to remain unaffected.

Fig. 11 Initial and equilibrium concentrations of Th⁴⁺ and O²⁻.

In order to determine whether the Th⁴⁺ in molten salt reacted with O²⁻, the concentration of Th⁴⁺ in different sections of the molten salt of O2–Th was analyzed, the relevant result was shown in Table 6. It was evident that the concentrations of Th⁴⁺ in all three parts were similar, confirming that Th⁴⁺ was uniformly distributed in the FLiBe salt. However, obvious precipitates were observed at the bottom of the salt (as shown in Fig. 12), and the thickness of the precipitate layer increased with the increase of O²⁻. Subsequently, an XRD analysis was performed on the precipitate after most of FLiBe salt was removed (Fig. 13). The XRD pattern confirmed the presence of BeO but no ThO₂ in the precipitate.

Table 6 Concentrations of Th⁴⁺ (Th⁴⁺ : O²⁻ = 1 : 2)

	Th^4+, mol kg^−1
Upper part	0.160
Middle part	0.162
Lower part	0.157

---

Fig. 12 (a) Molten salt of O1–Th; (b) molten salt of O2–Th.

Fig. 13 XRD spectra of O2–Th bottom precipitation.

Therefore, it is necessary to determine the concentration of Be²⁺ in the molten salt. Fig. 14 illustrates the variation of [Be²⁺] at different sections within the molten salts of O1–Th, O2–Th, and FLiBe–ThF₄. It can be seen that the [Be²⁺] remained relatively consistent across different positions in the FLiBe–ThF₄ molten salt. However, in the salts of O1–Th and O2–Th, the [Be²⁺] at the bottom was higher than that of the top and middle sections. Furthermore, the [Be²⁺] at the bottom of O2–Th was higher than that of O1–Th. These experimental findings suggested that the addition of O²⁻ led to an enrichment of Be²⁺ at the bottom, which could be attributed to the fact that O²⁻ did not react with Th⁴⁺ but with Be²⁺ in the FLiBe system.

Fig. 14 Concentrations of Be²⁺ in FLiBe–ThF₄, O1–Th and O2–Th.

The comparison of binding ability of Th⁴⁺, Be²⁺ with O²⁻ in fluoride salts

Based on the titration experiments results outlined above, two sets of experiments were specifically designed to investigate the interaction between Be²⁺, Th⁴⁺ and O²⁻ in fluoride salts. To observe whether the Th⁴⁺ has the ability to capture the O²⁻ from BeO, the titration experiment of ThF₄ in the FLiBe was conducted by adding BeO instead of Li₂O. And to examine whether O²⁻ can combine with Th⁴⁺ to generate ThO₂, the titration experiment of ThF₄ with Li₂O in FLiTh without Be²⁺ was performed.

In the titration experiment of ThF₄ with BeO in FLiBe, the ratio of BeO/ThF₄ was set to be 1 : 2. The equilibrium [Th⁴⁺] and [Be²⁺] as well as the [O²⁻] in the supernatant of the molten salt were listed in Table 7. It showed that there was no obvious change in the concentrations of Be²⁺, Th⁴⁺ and O²⁻. However, Fig. 15(a) presents obvious precipitation at the bottom of the molten salt, and XRD characterization (the result shown in Fig. 15(b)) revealed that the precipitation was BeO. The experimental results confirmed that the BeO did not react with Th⁴⁺ in FLiBe molten salt, suggesting it is difficult for Th⁴⁺ to capture the O²⁻ in BeO.

Table 7 Parallel measurement results of [Th⁴⁺], [Be²⁺], and [O²⁻] in FLiBe–ThF₄–BeO titration experiments (initial Th⁴⁺, Be²⁺ were 0.163 and 9.67 mol kg⁻¹, respectively)

	Th^4+, mol kg^−1	Be^2+, mol kg^−1	O^2−, mol kg^−1
1	0.162	9.81	1.57 × 10^−2
2	0.160	9.62	1.75 × 10^−2
3	0.162	9.97	1.89 × 10^−2
4	0.161	9.61	1.89 × 10^−2

---

Fig. 15 (a) FLiBe–ThF₄–BeO molten salt; (b) XRD spectrum of the precipitation of FLiBe–ThF₄–BeO; (c) XRD spectrum of the precipitation of FLiTh–Li₂O.

In another titration experiment of ThF₄, Li₂O was directly added to FLiTh salt, and the precipitate obtained in this system was characterized by XRD after the removal of the molten salt (shown in Fig. 15(c)). The XRD pattern clearly revealed the presence of ThO₂, indicating that Th⁴⁺ has the ability to react with O²⁻ to form ThO₂ in the absence of Be²⁺.

Based on the solubility experiment of ThO₂ and the titration experiment of ThF₄ in FLiBe salt, it was clearly demonstrated that when O²⁻ was introduced into the FLiBe–ThF₄ salt, it preferentially reacted with Be²⁺ rather than Th⁴⁺. This suggested that Be²⁺ had a protective effect on Th⁴⁺ in the fluoride salt system.

Conclusions

To investigate the effect of O²⁻ on the fuel salt of molten salt reactor, and provide a basis for understanding the chemical behavior of oxides in the fluoride salt system, this study focused on the solubility of UO₂ and ThO₂, as well as the interaction behavior of O²⁻ with U⁴⁺, Th⁴⁺, and Be²⁺ in the FLiBe system. The main findings are summarized as follows.

(1) At a temperature of 650 °C, the solubility of UO₂ was 5.52 × 10⁻³ mol kg⁻¹, and the value of K_sp was 6.14 × 10⁻⁷ mol³ kg⁻³. When the added [O²⁻] was less than twice of the [U⁴⁺] in the FLiBe system, O²⁻ would preferentially react with U⁴⁺ to form UO₂ precipitate. As more O²⁻ continued to be introduced, O²⁻ would react with Be²⁺ to generate BeO precipitate.

(2) When O²⁻ was introduced into the FLiBe–ThF₄ system, O²⁻ would preferentially react with Be²⁺ rather than Th⁴⁺, and only in the absence of Be²⁺, O²⁻ reacted with Th⁴⁺ to form ThO₂. In brief, O²⁻ introduced into the FLiBe molten salt at 650 °C would react firstly with U⁴⁺, then with Be²⁺, and finally with Th⁴⁺.

Hence, when oxide impurities were introduced into the molten salt reactor fuel, it is important to take into account their impact on uranium to prevent the formation of UO₂ precipitate. The current results indicated that the presence of a large number of U⁴⁺ and Be²⁺ played a certain protective role for Th⁴⁺.

Author contributions

Yubing Yan: methodology, investigation, writing – original draft, writing – review. Yingjie Li: data curation, investigation. Haiying Fu: conceptualization, methodology. Yuan Qian: conceptualization. Qingnuan Li: conceptualization, validation. Junxia Geng: writing – review, supervision, validation. Qiang Dou: writing – review, validation, project administration, resources.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 12175303 and U2267226), Xinjiang Uygur Autonomous Region Key R&D Task Special Project (No. 2022B01039).

References

1. L. Mathieu, D. Heuer, R. Brissot, C. Garzenne, C. L. Brun, D. Lecarpentier, E. Liatard, A. Nuttin, E. Walle and J. Wilson, Prog. Nucl. Energy, 2006, 48, 664–679.
2. C. György and Sz. Czifrus, Prog. Nucl. Energy, 2016, 93, 306–317.
3. J. Serp, M. Allibert, S. Delpech, O. Feynberg, V. Ghetta, D. Heuer, D. Holcomb, V. Ignatiev, J. L. Kloosterman, L. Luzzi, R. Yoshioka and D. Zhimin, Prog. Nucl. Energy, 2014, 77, 308–319.
4. O. Benes and R. J. M. Konings, J. Fluorine Chem., 2009, 130, 22–29.
5. P. Souček, O. Beneš, B. Claux, E. Capelli, M. Ougier, V. Tyrpekl, J. F. Vigier and J. M. Konings, J. Fluorine Chem., 2017, 200, 33–40.
6. C. Forsberg, Prog. Nucl. Energy, 2005, 47, 32–43.
7. A. Nuttin, D. Heuer, A. Billebaud, R. Brissot, C. Le Brun, E. Liatard, J. M. Loiseaux, L. Mathieu, O. Meplan and E. M. Lucotte, Prog. Nucl. Energy, 2005, 46, 77–99.
8. M. L. Tan, G. F. Zhu, Z. D. Zhang, Y. Zou, X. H. Yu, C. G. Yu, Y. Dai and R. Yan, Nucl. Sci. Technol., 2022, 33, 5.
9. H. Nishimura, T. Terai, T. Yoneoka, S. Tanaka, A. Sagara and O. Motojima, J. Nucl. Mater., 2000, 283–287, 1326–1331.
10. Y. L. Wang, Q. Wang, H. J. Liu and C. L. Zeng, Corros. Sci., 2016, 103, 268–282.
11. F. Y. Ouyang, C. H. Chang, B. C. You, T. K. Yeh and J. J. Kai, J. Nucl. Mater., 2013, 437, 201–207.
12. K. M. Sankar and P. M. Singh, Corros. Sci., 2022, 206, 110473.
13. S. Q. Guo, J. S. Zhang, W. Wu and W. T. Zhou, Prog. Mater. Sci., 2018, 97, 448–487.
14. N. J. Condon, S. Lopykinski, F. Carotti, K. E. Johnson and A. Kruizenga, ACS Omega, 2023, 8, 29789–29793.
15. R. D. Scheele, A. M. Casella and B. K. McNamara, Ind. Eng. Chem. Res., 2017, 56, 5505–5515.
16. D. A. Petti, G. R. Smolik, M. F. Simpson, J. P. Sharpe, R. A. Anderl, S. Fukada, Y. Hatano, M. Hara, Y. Oya, T. Terai, D. K. Sze and S. Tanaka, Fusion Eng. Des., 2006, 81, 1439–1449.
17. A. L. Mathews and C. F. Baes Jr, Inorg. Chem., 1968, 7, 373–382.
18. P. Calderoni, P. Sharpe, H. Nishimura and T. Terai, J. Nucl. Mater., 2009, 386–388, 1102–1106.
19. S. Delpech, C. Cabet, C. Slim and G. S. Picard, Mater. Today, 2010, 13, 34–41.
20. H. Peng, M. Shen, Y. Zuo, X. X. Tang, R. Tang and L. D. Xie, Electrochim. Acta, 2016, 222, 1528–1537.
21. M. Korenko, M. Straka, J. Uhlíř, L. Szatmáry, M. Ambrová and M. Šimurda, J. Radioanal. Nucl. Chem., 2014, 302, 549–554.
22. H. Peng, W. Huang, L. D. Xie and Q. N. Li, J. Nucl. Mater., 2020, 531, 152004.
23. Molten-salt Reactor Program: the Development Status of Molten-Salt Breeder Reactors, ORNL-4812, 1972.
24. Molten-salt Reactor Program: Semiannual Progress Report for Period Ending, ORNL-3419, 1963.
25. L. Toth, U. Gat, G. Del Cul, S. Dai and D. Williams, Review of ORNL's MSR technology and status, Oak Ridge National Lab., 1996.
26. H. Peng, Y. L. Song, N. Ji, L. D. Xie, W. Huang and Y. Gong, RSC Adv., 2021, 31, 18708–18716.
27. Y. L. Song, M. Shen, S. F. Zhao, R. Tang, L. D. Xie and Y. Qian, J. Electrochem. Soc., 2021, 168, 036513.
28. P. Chamelot, L. Massot, L. Cassayre and P. Taxil, Electrochim. Acta, 2010, 55, 4758–4764.
29. C. Y. Wang, X. T. Chen and Y. Gong, J. Phys. Chem., 2021, 125, 1640–1646.
30. M. Y. Xie, L. Li, Y. P. Ding and G. X. Zhang, J. Nucl. Mater., 2017, 487, 317–322.
31. S. M. Lee, Y. Jiang, J. Jung, J. H. Yum, E. S. Larsen, C. W. Bielawski, W. Wang, J. H. Ryoue, H. S. Kimf, H. Y. Chaf and J. Oh, Appl. Surf. Sci., 2019, 469, 634–640.

---