English

• The reprocessing of spent fuel is one of the critical steps in the nuclear fuel cycle. In recent decades, extensive studies have been carried out on dry reprocessing using molten salt as a medium [1-3]. Thereinto, the electrolytic refining process developed by the United States and oxide-electrowinning reprocessing developed by Russia as are considered as the prominent candidate technologies for advanced nuclear fuel reprocessing in the future [4, 5]. For MOX spent fuel, removing the cladding material is an important step in head-end process of spent fuel reprocessing. Recently, high temperature oxidation technology was adopted to separate spent fuel from cladding material by high temperature calcination UO₂ to form a volume expansion force to destroy the cladding material, which was proposed by Idaho national laboratory (TNL) and Korea atomic energy research institute (KAERI). During the process, UO₂ is oxidized to U₃O₈ and UO₃ [6]. In the oxide-electrowinning process, uranium and fission products in the oxide fuel are firstly chlorinated and dissolved into the molten salt as oxychloride (uranyl ion) and chloride ions by flowing chlorine gas, then uranyl ion is reduced to form the raw material of MOX fuel (UO₂) on the cathode in molten chlorides. However, the spontaneity of the chlorination reaction is unfavorable, and hence, it is necessary to add reducing agents, for instance, carbon and carbon monoxide into the melts [7]. Thus, many investigators explored the chlorination reaction of uranium oxide (UO₂ and U₃O₈) using ZrCl₄ [8], CCl₄ [9, 10], NH₄Cl [11]. Sakamura et al. [8] studied UO₂ was chlorinated to form UCl₄ by ZrCl₄ in LiCl-KCl melt, however, the conversion rate was only 37%. Kitawaki et al. [9] explored the chlorination UO₂ and U₃O₈ using CCl₄ through mechanical mixing method, but did not succeed. Jiang et al. [10] chlorinated U₃O₈ to UCl₄ using CCl₄ at high temperature, but CCl₄ would volatilize severely and the by-product of UCl₅ was formed. As we known, UCl₅ is easy to volatilize which causes the loss of uranium. Liu et al. [11] investigated the chlorination reaction of U₃O₈ and NH₄Cl in LiCl-KCl melt under air atmosphere. Their results showed that U₃O₈ can be successfully chlorinated to UO₂²⁺. Wani et al. [12] studied the fluorination of UO₂ and U₃O₈ with NH₄HF₂ and found the formation of [NH₄]₄UF₈ and [NH₄]₃UO₂F₅ compounds. These two compounds, on heating to about 673 K could form UF₄ and UO₂F₂·2H₂O, respectively.

  The electrochemical reduction of UO₂²⁺ has been explored in molten chlorides and found that UO₂²⁺ can be reduced to UO₂ by two-step single electron transfer [13-16]. Schlechteret et al. [17] reported that the crystalline form of electrodeposited UO₂ in LiCl-KCl molten salt was affected by the presence of fluoride, and the presence of a small amount of fluoride would be conducive to the formation of cubic crystal UO₂. Caligara et al. [18] measured the diffusion coefficient of UO₂²⁺ in molten LiCl-KCl eutectic.

  In order to explore the route of the formation of UO₂ using UO₃ as raw material, NH₄HF₂ was chosen as fluoride reagent to make uranium oxide change to UO₂²⁺ under air atmosphere. Then, the electrochemical formation of UO₂ was investigated in LiCl-KCl molten salt by constant potential electrolysis for uranium extraction. The fluoride product and electrodeposition sample were checked by XRD, SEM-EDS.

  Firstly, the possibility of the formation of UO₂F₂ from UO₃ was estimated using the thermodynamic analysis of related fluorination reactions. UO₃ can react with NH₄HF₂ or HF obtained from the decomposition of NH₄HF₂ to form UO₂F₂, the following reactions may be involved as follows:

  (1)

  (2)

  (3)

  The standard Gibbs free energy changes (Δ_rG_m^Θ) for these reactions at various temperatures were calculated shown in Fig. S1 (Supporting information). It can be seen that (Δ_rG_m^Θ) of these reactions are less than zero, which means that all the reactions correlated with the formation of UO₂F₂ are spontaneous at the temperature of 473–873 K. Moreover, the (Δ_rG_m^Θ) becomes more and more negative with the increasing temperature, indicating that the increase of temperature can be beneficial to the fluorination reaction of UO₃.

  The fluorination reaction of UO₃ with NH₄HF₂ was performed at 573 K for different times. The obtained products were characterized by XRD shown in Fig. 1a. It can be seen from Fig. 1a that after the fluorination reaction for 30 min, only (NH₄)₃UO₂F₅ is observed except for the unreacted NH₄HF₂. No diffraction peak of UO₃ occurs, indicating that UO₃ is firstly converted to (NH₄)₃UO₂F₅. After 60 min, only NH₄(UO₂)₂F₅ appears, which shows that (NH₄)₃UO₂F₅ is completely converted to NH₄(UO₂)₂F₅. When the fluorination reaction time is extended to 120 min, the diffraction peak of UO₂F₂ is observed. However, NH₄(UO₂)₂F₅ is not completely decomposed, and its diffraction peaks can still be observed. However, after the reaction for 420 min, XRD result indicates there are only UO₂F₂, and NH₄(UO₂)₂F₅ has been completely converted to UO₂F₂.

  Figure 1

  

  Figure 1.  (a) XRD patterns of products obtained at 573 K for 30 min, 60 min, 120 min, 420 min. (b) Fluorescence spectrum and (c) Raman spectrum.

  DownLoad: Full-Size Img PowerPoint

  To further confirm the formation of UO₂F₂, fluorescence spectroscopy and Raman spectroscopy were used to analyze the product obtained from the fluorination of UO₃ for 180 min at 673 K. It can be seen from Fig. 1b that peaks maxima appear at about 485, 495, 516, 538 and 566 ± 0.5 nm, showing a five-finger shape, which is a typical characteristic peak of UO₂²⁺ [19, 20]. Fig. 1c shows the Raman spectrum of the product. Three obvious vibration peaks are observed at 181 cm⁻¹, 443 cm⁻¹ and 915 cm⁻¹, respectively. Among them, the strong band at 915 cm⁻¹ is ascribed to the symmetric U-O stretching vibration. The band at 181 cm⁻¹ corresponds to the U-O bending fundamental and the weaker band at 443 cm⁻¹ is related to the U-F lattice vibration which are consistent with those of UO₂F₂ obtained by Armstrong et al. [21]. These results demonstrate the formation of UO₂F₂.

  The reaction mechanism is expressed as follows:

  (4)

  (5)

  (6)

  The products obtained under different fluorination times at 573 K were analyzed by SEM to observe the morphology changes. Fig. S2 (Supporting information) shows the SEM images of fluorination products obtained at different duration. It can be seen that UO₃ presents a spherical particle with size of about 1 µm. UO₃ was fluorinated for 30 min, according to the XRD result, (NH₄)₃UO₂F₅ with size of about 3 µm is formed. A truncated octahedral composed of four hexagons and ten quadrilaterals is observed in Fig. S2b. After 60 min of the reaction, according to the XRD result, (NH₄)₃UO₂F₅ is decomposed to NH₄(UO₂)₂F₅, and NH₃ and HF are released. Thus, a small number of pores are observed on the surface of the particles (Fig. S2c). When the reaction proceeds for 420 min, the product is UO₂F₂ based on the XRD result. It can be observed from Fig. S2d that the formation of pores appears on the surface of the particles due to the release of NH₃ and HF, which indicates that large amounts of NH₃ and HF have been released.

  Then the electrode reaction of UO₂²⁺ was studied in LiCl-KCl melts on Mo electrode at 773 K using cyclic voltammetry (CV) shown in Fig. 2a. In the black line, peaks C₁ and A₁ are attributed to the deposition and dissolution of metallic Li. While UO₂²⁺ was added into LiCl-KCl melts, the color of the molten salt change from colorless to yellow shown in the inset of Fig. 2a, which is exactly the color of UO₂²⁺ dissolved in molten salt. In the red line, four new anodic peaks, A₂, A₂', A₃ and A₄, are observed at −2.21 V, −1.42 V, 0.27 V and 0.54 V, respectively. Only three cathodic peaks, C₂, C₃ and C₄, are observed at −2.45 V, −0.31 V and 0.32 V, respectively. The redox couple C₄/A₄ may be correlated with the soluble/soluble redox process of UO₂²⁺/UO₂⁺.

  (7)

  Figure 2

  

  Figure 2.  CV curves of UO₂²⁺ at (a) 0.1 V/s and (b) various scan speed. (c) The relationship of electricity amount and scan speed. (d) SWV curve of UO₂²⁺ at 20 Hz. (e) Various frequencies, and (f) fitting line of peak current density versus square root of frequency (f^1/2).

  DownLoad: Full-Size Img PowerPoint

  The C₃/A₃ couple pertains to the formation of UO₂ and its re-oxidation.

  (8)

  The C₂/A₂ couple is ascribed to the reduction of UO₂ to U metal and oxidation of U metal, and A₂′ corresponds to the dissolution of U to U³⁺ [22, 23].

  Fig. 2b presents the CV curves of LiCl-KCl-UO₂²⁺ (0.24 wt%) melts at various scan speeds. For the redox peaks of C₃/A₃, with the increase of the scan rate, the cathodic/anodic peak currents increase and the cathodic/anodic peak potentials have a significant negative/positive shift which shows that the Eq. 9 is not a reversible process. However, for the redox peaks of C₄/A₄, with the increase of the scan rate, the cathodic peak current increases and anodic peak current decreases. It can be seen that when the scan rate is 0.04 V/s and 0.06 V/s, the oxidation peak A₃ is not observed. A₄ is related to the UO₂⁺ to disproportionate into UO₂ (monolayer) and it has a sharp shape and a large peak current. When the scan rate reaches 0.08 V/s, peak A₃ appears, and the peak current of A₄ decreases rapidly, the peak potential shifts negatively, and the peak become widens. When the scan rate increases from 0.1 V/s to 0.16 V/s, the peak current of A₃ continued to increase, the peak current of A₄ gradually decreases, and peak A₄ became wider.

  This phenomenon may be caused by the disproportionation reaction of UO₂⁺.

  (9)

  At low scan rate, the measurement time is relatively long, the formedUO₂⁺ has enough time to generate UO₂²⁺ and UO₂ by the disproportionation Eq. 9, and the concentration of UO₂⁺ decreases, which results in that the current peaks C₃ and A₃ related to the reduction and oxidation of UO₂⁺ are not obvious. As the scan rate increases, the disproportionation reaction time is reduced, and the amount of UO₂ generated by the disproportionation reaction gradually decreases, the peak current of the oxidation peak A₄ gradually decrease and oxidation peak A₃ gradually appears.

  The electrode area is 0.3768 cm². Assuming that a layer of UO₂ was deposited on the electrode surface, the electric quantity of A₄ which was integrated of the peak area required by disproportionation reaction is at least 0.102 C. To obtain the electric quantity passed during oxidation reaction, the current and oxidation time related to the peak A₄ in Fig. 2b was integrated, and then the relationship of quantity of electricity versus scan rate was plotted, as shown in Fig. 2c. It can be seen that when the scan rate is relatively low, the electric quantity passing through the oxidation peak A₄ is close to that required to theoretically deposit a single layer of UO₂. As the scan rate increases, the electric quantity passing through peak A₄ gradually decreases, which meant the amount of UO₂ deposited at A₄ gradually decreases. This is consistent with the analysis result in Fig. 2b and further proves that the existence of oxidation peak A₄ is correlated with the formation of UO₂ by the disproportionation reaction of UO₂⁺.

  To further illustrating the electrochemical reactions correlated with peaks C₃ and C₄, the square wave voltammetry (SWV) was used to measure the number of electron-transferred during the two electrochemical reactions. Fig. 2d shows the SWV curve when reduction reaction of UO₂²⁺ occurs on W electrode. In the measured potential range, three reduction peaks (C₃, C₃′ and C₄) are observed and then the number of electrons transferred obtained by Eq. 10.

  (10)

  where W_1/2 is the half-peak width; R is ideal gas constant; T is absolute temperature; F is Faraday constant; and n is the number of electrons transferred.

  Since the appearance of peak C₃', which may be related to the nucleation of UO₂ [24], will interfere with the calculation result peak C₃, the Gaussian fitting was performed to eliminate the influence. Using Eq. 10, the number of electron transfer attributed to the peak C₃ is 1.09, close to 1. Thus, the reaction correlated with peak C₃ is the reduction reaction of UO₂⁺ to UO₂ via a one-step involving one electron.

  And the calculated number of electron-transferred from the reduction peak of C₄ is 0.92, close to 1, which means the reaction pertaining to peak C₄ is that UO₂²⁺ is reduced to UO₂⁺ in a one-step with the exchange of one-electron. Fig. 2d shows the SWV curves of LiCl-KCl-UO₂F₂ melts at various frequencies. As shown in Fig. 2e, a pair of obvious redox peaks A₄'/C₄ appears at around −0.40 V, and their shape is flat and symmetrical, which indicates that the redox reaction is a reversible reaction in a soluble/soluble system. Plotting the peak current density and the square root of the frequency, a good linear relationship indicates that the redox reaction is controlled by diffusion (Fig. 2f). Thus, the diffusion coefficient of UO₂²⁺ can be calculated using Eq. 11 [25].

  (11)

  where C₀ is concentration of UO₂²⁺ in the molten salt (C₀ = 9.804 × 10^–5 mol/cm³); D is diffusion coefficient (cm²/s); n is numbers of electrons transfer (n = 1); f is frequency; E_SW is amplitude of SWV (E_SW = 0.1 V).

  The calculated diffusion coefficient of UO₂²⁺ at 773 K is 6.22 × 10^–5 cm²/s, which has the same order of magnitude with the literature (1.54 × 10^–5 cm²/s) obtained on platinum electrode at 858 K ^[18].

  According to the electrochemical results, it is feasible for the formation of UO₂ by molten electrolysis. Thus, potentiostatic electrolysis was performed to form UO₂ at −0.8 V for 8 h. The change of current with time was recorded, as shown in Fig. 3a. A strong current appears at the beginning (inset), which might be related to the formation and growth of UO₂ nuclei. Before 1500 s, the current fluctuated at −20 mA, meaning the formation of UO₂. However, with the proceeding, the concentration of UO₂²⁺ decreases, leading to the decrease of the absolute value of current. When the I-t curve became steady and close to background current about 2 mA, it could be considered that the run is completed.

  Figure 3

  

  Figure 3.  (a) I-t curve, (b) XRD patterns and (c) SEM-EDS images of electrolysis product obtained by potentiostatic electrolysis at −0.8 V on Mo plate at 773 K.

  DownLoad: Full-Size Img PowerPoint

  The XRD spectra of the product gained shown in Fig. 3b indicates the formation of UO₂ successfully and the color of the Mo plate changes from metallic luster to black, as shown in inset of Fig. 3b. After stripping off the UO₂ from the electrode, the diffraction peaks in XRD spectra shown in Fig. 3b became stronger and sharper, showing the deposited UO₂ has a high crystallinity. The morphology and element distribution were characterized, as shown in Figs. 3c-f, the UO₂ is unevenly distributed on the electrode surface in a dendritic state. EDS analysis also proved the formation of UO₂. Before and after the run, the concentration of U in the melts was analyzed by ICP-AES and the extraction efficiency was calculated using Eq. 12. The calculated result indicated that after 8 h electrolysis, the extraction of UO₂ can reach 98.53%.

  (12)

  where η is extraction efficiency; C₁ and C₂ are the initial and final concentrations of UO₂²⁺ in the melts, respectively.

  In order to electrochemically form UO₂ using UO₃ as raw material, the formation of UO₂F₂ by the fluorination reaction of UO₃ and NH₄HF₂ and its mechanism were firstly studied using XRD, Raman and fluorescence. The results revealed the formation of UO₂F₂, and the mechanism was inferred to be comprised of the following three process UO₃(s) + 3NH₄HF₂(s) → (NH₄)₃UO₂F₅ → NH₄(UO₂)₂F₅(s) → 2UO₂F₅(s). Then, the electrochemical formation of UO₂ was studied using CV and SWV, which illustrated UO₂²⁺ could be reduced to UO₂ via a two-step single electron exchange process with diffusion-controlled. The diffusion coefficient of UO₂²⁺ was estimated to be 6.22 × 10⁻⁵ cm²/s. The disproportionation reaction of UO₂⁺ was observed, and the relationship between the disproportionation reaction and scan rate was discussed. In addition, UO₂ was prepared employing potentiostatic electrolysis at −0.8 V for 8 h in LiCl-KCl-UO₂F₂ melt. XRD and SEM-EDS results indicated UO₂ with micron-sized octahedral structure was formed, and the extraction efficiency of UO₂ calculated by ICP-AES could reach 98.53%. The feasibility of UO₂ prepared using UO₃ as raw material was proved.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  The work was financially supported by the National Natural Science foundation of China (Nos. U2167215, 22076035, 21876034, 11875116 and 21790373), and the Fundamental Research Funds for the Central Universities (No. 3072021CFJ1001).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.02.069.

  
  
• 1. [1]

     D.W. Yang, Y.L. Liu, T.Q. Yin, et al., Electrochim. Acta 353 (2020) 136449. doi: 10.1016/j.electacta.2020.136449

  2. [2]

     Y.K. Zhong, Y.L. Liu, K. Liu, et al., Nat. Commun. 12 (2021) 5777. doi: 10.1038/s41467-021-26119-9

  3. [3]

     D.W. Yang, S.L. Jiang, Y.L. Liu, et al., Sep. Purif. Technol. 281 (2022) 119853. doi: 10.1016/j.seppur.2021.119853

  4. [4]

     S. Vavilov, T. Kobayashi, M. Myochin, J. Nucl. Sci. Technol. 41 (2004) 1018-1025. doi: 10.1080/18811248.2004.9726326

  5. [5]

     J.L. Willit, W.E. Miller, J.E. Battles, J. Nucl. Mater. 195 (1992) 229-249. doi: 10.1016/0022-3115(92)90515-M

  6. [6]

     B.R. Westphal, J.J. Park, J.M. Shin, et al., Sep. Sci. Technol. 43 (2008) 2695-2708. doi: 10.1080/01496390802122139

  7. [7]

     Y.S. Yang, Y.H. Kang, H.K. Lee, Mater. Chem. Phys. 50 (1997) 243-247. doi: 10.1016/S0254-0584(97)01943-3

  8. [8]

     Y. Sakamura, T. Inoue. T. Iwai, H. Moriyama, J. Nucl. Mater. 340 (2005) 39-51. doi: 10.1016/j.jnucmat.2004.11.002

  9. [9]

     S. Kitawaki, T. Nagai, N. Sato, J. Nucl. Mater. 439 (2013) 212-216. doi: 10.1016/j.jnucmat.2013.03.017

  10. [10]

      K. Jiang, V. Smolenski, A. Novoselova, et al., Electrochim. Acta 318 (2019) 194-201. doi: 10.1016/j.electacta.2019.06.075

  11. [11]

      Y.C. Liu, Y.L. Liu, L. Wang, et al., J. Nucl. Mater. 542 (2020) 152475. doi: 10.1016/j.jnucmat.2020.152475

  12. [12]

      B.N. Wani, S.J. Patwe, U.R. Rao, K.S. Venkateswarlu, J. Fluorine Chem. 44 (1989) 177-185. doi: 10.1016/S0022-1139(00)83937-8

  13. [13]

      J.R. Chalkley, J. Less Common Met. 3 (1961) 98-109. doi: 10.1016/0022-5088(61)90002-9

  14. [14]

      T. Nagai, T. Fujii, O. Shirai, H. Yamana, J. Nucl. Sci. Tech. 41 (2004) 690-695. doi: 10.1080/18811248.2004.9715534

  15. [15]

      T. Nagai, A. Uehara, T. Fujii, O. Shiral, H. Yamana, J. Nucl. Sci. Tech. 43 (2006) 1511-1516. doi: 10.1080/18811248.2006.9711247

  16. [16]

      I. Uchida, J. Niikura, S. Toshima, J. Electroanal. Chem. 124 (1981) 165-177. doi: 10.1016/S0022-0728(81)80294-X

  17. [17]

      M. Schlechter, J. Kooi, R. Billiau, R.A. Charlier, G.L. Dumont, J. Nucl. Mater. 15 (1965) 189-200. doi: 10.1016/0022-3115(65)90180-7

  18. [18]

      F. Caligara, L. Martinot, G. Duyckaerts, J. Electroanal. Chem. 16 (1968) 335-340. doi: 10.1016/S0022-0728(68)80081-6

  19. [19]

      J.Y. Lee, J.Y. Oh, K. Putri, M. Baik, J.I. Yun, J. Radioanal. Nucl. Chem. 312 (2017) 221-231. doi: 10.1007/s10967-017-5233-y

  20. [20]

      Y. Sihn, J.I. Yun, W. Lee, J. Radioanal, Nucl. Chem. 308 (2016) 413-423. doi: 10.1007/s10967-015-4428-3

  21. [21]

      D.P. Armstrong, R.J. Jarabek, W.H. Fletcher, Appl. Spectrosc. 43 (1989) 461-467. doi: 10.1366/0003702894203002

  22. [22]

      Y.K. Zhong, K. Liu, Y.L. Liu, et al., J. Electrochem. Soc. 166 (2019) D276-D282. doi: 10.1149/2.0201908jes

  23. [23]

      H.M. Li, R.G. Liu, W. Han, et al., Electrochimi. Acta 400 (2021) 139474. doi: 10.1016/j.electacta.2021.139474

  24. [24]

      D. Han, C. She, J. Peng, et al., J. Electrochem. Soc. 165 (2018) 301-306. doi: 10.1149/2.0971807jes

  25. [25]

      E.P. Parry, R.A. Osteryoung, Anal. Chem. 37 (1965) 1634-1637. doi: 10.1021/ac60232a001

• 

  Figure 1  (a) XRD patterns of products obtained at 573 K for 30 min, 60 min, 120 min, 420 min. (b) Fluorescence spectrum and (c) Raman spectrum.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  CV curves of UO₂²⁺ at (a) 0.1 V/s and (b) various scan speed. (c) The relationship of electricity amount and scan speed. (d) SWV curve of UO₂²⁺ at 20 Hz. (e) Various frequencies, and (f) fitting line of peak current density versus square root of frequency (f^1/2).

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) I-t curve, (b) XRD patterns and (c) SEM-EDS images of electrolysis product obtained by potentiostatic electrolysis at −0.8 V on Mo plate at 773 K.

  下载: 全尺寸图片 幻灯片
•