J Radioanal Nucl Chem (2013) 298:393–397 DOI 10.1007/s10967-013-2436-8
Uranium recovery from LiF–CaF2–UF4–GdF3 system on Ni electrode M. Straka • L. Szatma´ry • M. Marecˇek M. Korenko
•
Received: 28 November 2012 / Published online: 13 February 2013 Ó Akade´miai Kiado´, Budapest, Hungary 2013
Abstract Electrochemistry of gadolinium and uranium in LiF–CaF2 (79–21 mol%) melt was studied using reactive Ni electrode and alloying reactions were observed. Deposits of gadolinium and uranium in the form of Gd–Ni and U–Ni intermetallic alloys were obtained after electrolysis by modulated current. Electrolysis of the same parameters was used also in the complex system of LiF– CaF2–UF4–GdF3 to demonstrate feasibility of selective deposition of uranium and therefore its separation from the system. Compact deposit of U–Ni alloy containing only traces of gadolinium was obtained. Keywords Molten salt Electrolysis Spent nuclear fuel reprocessing Pyrochemical reprocessing Uranium Gadolinium
temperature range of usability, low process volume etc.). In this context, it is necessary to gain detailed knowledge of actinides/lanthanides electrochemistry in molten fluorides and to investigate possibilities of electrochemical separation methods in both qualitative and quantitative point of view. In the first part of this work, the electrochemical behaviour of uranium and gadolinium were investigated in LiF–CaF2 mixture (79–21 mol%). In the second part of this work, uranium recovery on reactive Ni electrode from LiF–CaF2–UF4–GdF3 system was investigated. It should be noted that recovery of uranium in the form of U–Al intermetallics was successfully tested by Soucˇek et al. [4] in chloride system LiCl–KCl–UCl3–NdCl3. Electrolysis by modulated current was used in this work and Gd-free uranium deposit was obtained from LiF–CaF2–UF4–GdF3 melt in the form of U–Ni alloy.
Introduction Among other applications, electrochemical separations of actinides from lanthanides [1], together with molten salt/ liquid metal extractions [2], are considered as suitable methods for the reprocessing of spent nuclear fuel in fuel cycles of proposed future types of nuclear reactors [1, 3]. Pyrochemical processes based on molten salt systems (chlorides or fluorides) have significant advantages over aqueous processes (higher radiation stability, wide M. Straka (&) L. Szatma´ry M. Marecˇek M. Korenko ´ JV Rˇezˇ, plc, Department of Fluorine Chemistry, U Husinec-Rˇezˇ 130, 25068 Rˇezˇ, Czech Republic e-mail:
[email protected] M. Korenko Institute of Inorganic Chemistry, Slovak Academy of Sciences, Dubravska´ cesta 9, 84536 Bratislava, Slovakia
Experimental Electrochemical experiments were carried out in the cell (glassy-carbon crucible) placed in an electrolyser made of INCONEL 625. The electrolyser consists of a vessel closed by a removable flange with built-in holders for the electrodes, thermocouple and inlet and outlet of argon gas as described system is under argon atmosphere (99.998 %) during the measurement. A resistance oven heats the electrolyser and it provides homogenous thermal field up to 1,000 °C. The whole apparatus is placed inside the glove box with dry nitrogen atmosphere (99.95 %); dew point analyser monitors level of moisture and level of oxygen is monitored by ARELCO oxygen analyser with electrochemical sensor. Typical content of moisture and oxygen in the glove-box is under 5 ppm for both water and oxygen.
123
394
Three-electrode system was used for all measurements. Large surface glassy-carbon crucible was used as a counter electrode. Nickel wires were used as working electrodes. The potentials were referred to a 0.5 mm platinum wire immersed in the molten electrolyte, acting as a quasi-reference electrode. The electrodes were connected to HEKA PG 310 potentiostat (HEKA GmbH, Lambrecht, Germany) controlled by PC with original software. LiF and CaF2 of commercial origin (99.50 and 99.96 % respectively) were dried in a vacuum dryer at 250 °C. GdF3 was of commercial origin with purity 99.90 %. It was used as received without further purification. UF4 was prepared by fluorination of UO2 from the stock of the department. UF4 purity was verified by XRD analysis. Impurities were under detection level of XRD method. UF4 and GdF3 were introduced into the melt in the glove-box under dry nitrogen atmosphere. The melt was fused under argon atmosphere (99.998 %, see ‘‘Experimental’’ section). As working electrode materials, 0.5 and 1 mm Ni wire of 99.95 % purity were used. Pt wire (diameter 0.5 mm) was used as a comparison electrode.
Results and discussion LiF–CaF2–GdF3 Electrochemistry of gadolinium in molten fluorides was studied and presented by Nourry et al. [5–7] and Chamelot et al. [3]. One reduction step corresponding to reaction (1) was reported in LiF–CaF2 on inert electrode and nickel was described as a suitable reactive material for the deposition
Fig. 1 Cyclic voltammogram of the LiF–CaF2–GdF3 melt (3.4 wt%) on the Ni working electrode (0.08 cm2), T = 1,223 K, scan rate 50 mV s-1, Pt reference electrode, glassy carbon counter electrode
123
J Radioanal Nucl Chem (2013) 298:393–397
of gadolinium in the form of an alloy. According to Ni–Gd phase diagram [8], several intermetallic compounds of Ni and Gd exists. Gd3þ þ 3e ! Gd0
ð1Þ
In this work, the electrochemical behaviour of gadolinium in LiF–CaF2 was studied on Ni electrode at 1,223 K by cyclic voltammetry (CV). Presence of alloying reactions was expected. Cyclic voltammogram of the LiF–CaF2–GdF3 (3.4 wt%) system can be found in Fig. 1. Shape of the voltammogram with increase of cathodic current at about -1.6 V (vs. Pt reference) and subsequent anodic effects suggests alloying as mentioned above. This was confirmed also by the open-circuit chronopotentiometry (see Fig. 2). An open-circuit chronopotentiogram (CP-OCP) was obtained after short polarization (2s) at -200 mA cm-2 applied to Ni wire electrode. Plateaus observed during subsequent zerocurrent period can be attributed to dissolution of particular alloys of Gd and Ni. Based on the information obtained by CV and CP-OCP, the electrolysis by modulated current was done on Ni wire electrode with diameter of 0.5 mm. One pulse consisted of four parts (see Table 1). Total number of 257 pulses (total time of the electrolysis: 3,600 s) was applied. After the electrolysis, Ni electrode was analysed by SEM-EDX. Micrograph of the cross-section is in Fig. 3. Several areas corresponding to different Ni–Gd intermetallics can be seen. Atomic percent ratio of Ni and Gd in the area depicted as 1 was 4.9:1. In the area depicted as 2, the Ni and Gd ratio was 1.8:1, in the area 3, it was 2.0:1, which is very similar to area 2.
J Radioanal Nucl Chem (2013) 298:393–397
395
Fig. 2 Open-circuit chronopotentiogram of the LiF– CaF2–GdF3 system (3.4 wt%) at T = 1,223 K, Ni working electrode (0.08 cm2), Pt reference electrode
Table 1 Structure of the current pulse used in electrolytic experiments Period 1
2
3
4
Current (mA)
-30
-10
10
0
Time (s)
1
10
2
1
behaviour in LiF–BeF2–ZrF4 melts. Cyclic voltammogram of LiF–NaF–KF–UF4 system can be found in papers by Clayton et al. [12] and Soucˇek et al. [1]. Hamel et al. [13] used LiF–NaF and LiF–CaF2 melts. Nourry et al. [14] studied electrochemistry of LiF–CaF2–UF4 system. Baes [15] and Straka et al. [16] published study of the electrochemistry of LiF–BeF2–UF4 system. According to mentioned papers, reduction of U4? ions is of a two-step mechanism (see Eqs. (2), (3)) when an inert electrode is used. U4þ þ e ! U3þ U
3þ
þ 3e ! U
0
ð2Þ ð3Þ
Clayton et al. [12] additionally described the disproportionation of U3? species according to reaction (4). It should be noted, that influence of this reaction to the electrochemical signal was not detected by other authors. 4U3þ ! 3U4þ þ U0
Fig. 3 SEM micrograph of the cross-section of a nickel wire after electrolysis by modulated current in LiF–CaF2–GdF3 (3.4 wt%) system, T = 1,223 K
LiF–CaF2–UF4 Several papers dealing with the electrochemistry of uranium in fluoride melts can be found. Mamantov and Manning [9, 10] and Jenkins et al. [11] studied uranium
ð4Þ
Alloying reactions of uranium with nickel used as a working electrode were described in several of mentioned papers [13, 14, 16]. It is in agreement with the U–Ni phase diagram. According to paper by Wang et al. [17], seven U–Ni intermetallic compounds exist. Nickel was therefore chosen as a working electrode material for experiments within this work. CV was carried out on a nickel wire electrode at 1,223 K. The cyclic voltammogram obtained in LiF–CaF2–UF4 (3.6 wt%) melt at scan rate 50 mV s-1 is shown in Fig. 4. Cathodic wave at around -0.8 V (vs. Pt reference) can be attributed to the reaction (2) and subsequent waves can be connected with alloying reactions. Waves on the anodic side of the voltammogram can be interpreted as dissolution of particular U–Ni alloys.
123
396
J Radioanal Nucl Chem (2013) 298:393–397
Fig. 4 Cyclic voltammogram of the LiF–CaF2–UF4 melt (3.6 wt%) on the Ni working electrode (0.08 cm2), T = 1,223 K, scan rate 50 mV s-1, Pt reference electrode, glassy carbon counter electrode
Fig. 5 SEM micrograph of the cross-section of a nickel wire after electrolysis by modulated current in LiF–CaF2–UF4 (3.6 wt%) system, T = 1,223 K
Electrolysis by modulated current (see Table 1 for parameters of the electrolysis, 257 cycles, overall duration 3,600 s) was done in this system on Ni wire electrode (diameter 0.5 mm). After the electrolysis, Ni wire electrode was analysed by SEM-EDX. Micrograph of the crosssection is in Fig. 5. Layer of Ni-rich U–Ni alloy is marked by 1, atomic percent ratio of Ni and U in the area was found to be 4.5:1; uranium-rich layer is marked by 2, Ni/U ratio in this area was found to be 1.4:1. LiF–CaF2–UF4–GdF3 Electrochemistry of complex system of uranium and gadolinium in LiF–CaF2 using Ni electrode was investigated by Nourry et al. [14]. It was concluded in paper [14], that it
123
is possible to selectively reduce uranium ions on the Ni electrode. The system LiF–CaF2–UF4 (3.5 wt%)–GdF3 (3.5 wt%) was described by CV prior the electrolytic experiments. Cyclic voltammogram can be found in Fig. 6. As can be seen from Figs. 1 and 4 respectively, Gd system was observed at about -1.6 V (vs. Pt, subsequent alloying was also observed) and U system connected with uranium deposition in the form of an alloy was observed at about -1.1 V (vs. Pt). It is in agreement with CV of a complex system in Fig. 6. It is in general agreement also with results by Nourry et al. [14], cathodic current increase can be again attributed to alloying reactions (U–Ni, Gd–Ni or U–Gd–Ni). Again, electrolysis by modulated current (see Table 1 for parameters of the electrolysis) was conducted in LiF– CaF2–UF4 (3.5 wt%)–GdF3 (3.5 wt%) system on Ni wire (diameter 0.5 mm) electrode. Potential difference described in previous paragraph should prefer the deposition of uranium and also the use of modulated current should positively affect the deposition of uranium over the deposition of gadolinium (during the cathodic period, the uranium is deposited in higher amount and during the anodic period, gadolinium—if present—is dissolved in higher amount). After the electrolysis, Ni electrode was analysed by SEM-EDX. Micrograph of the cross-section is in Fig. 7. Two compact layers of U–Ni alloys are designated as area 1 and area 2. Atomic percent ratio of Ni and U in the area depicted as 1 was 4.4:1, in the area 2 it was 1.6:1. These ratios can be considered the same as in the case of the deposit after the electrolysis of LiF–CaF2–UF4 system (see ‘‘LiF–CaF2–UF4’’ section). Importantly, only traces of Gd were found in both layers (from 0.1 to 0.2 at.%). It can be concluded that qualitatively, uranium can be selectively
J Radioanal Nucl Chem (2013) 298:393–397
397
Fig. 6 Cyclic voltammogram of the LiF–CaF2–UF4 (3.5 wt%)–GdF3 (3.5 wt%) melt on the Ni working electrode (0.08 cm2), T = 1,223 K, scan rate 50 mV s-1, Pt reference electrode, glassy carbon counter electrode
were observed by CV or CP-OCP methods and Gd–Ni and U–Ni intermetallics were found after electrolysis by modulated current in LiF–CaF2–GdF3 and LiF–CaF2–UF4 systems respectively. Selective deposition of uranium from LiF–CaF2–UF4–GdF3 melt was shown to be feasible; uranium was deposited in the form of compact U–Ni layers, containing only traces of gadolinium. Acknowledgments This work was supported by the Radioactive Waste Repository Authority of the Czech Republic, by the European Commission financial contribution within 7th Framework programme under the contract 211267 ‘‘ACSEPT’’ and by the Slovak Grant Agency VEGA 2/0095/12.
Fig. 7 SEM micrograph of the cross-section of a nickel wire after electrolysis by modulated current in LiF–CaF2–UF4 (3.5 wt%)–GdF3 (3.5 wt%) system, T = 1,223 K
removed from the fluoride system in the form of an alloy. In paper by Nourry et al. [14], single Ni–U (atomic ratio 6.5:1) alloy was found. It should be noted that the Ni–U layer was not regular and compact compared to our results. However, direct comparison is not suitable for these two deposited layers as the initial concentration of U and Gd as well as method of the electrolysis (constant potential in case of paper [14]) were different in both works. Quantitative evaluation of individual parameters influence to the resulting deposit will be subject of the next study.
Conclusion Electrochemistry of gadolinium and uranium in LiF–CaF2 was studied on reactive (Ni) electrode. Alloying reactions
References 1. Soucˇek P et al (2005) J Nucl Sci Technol 42:1017–1024 2. Rosenthal MW (1974) Technical report ORNL-5018. Oak Ridge National Laboratory, Oak Ridge 3. Chamelot P et al (2007) J Nucl Mater 360:64–74 4. Soucˇek P et al (2010) J Radioanal Nucl Chem 286:823–828 5. Nourry C et al (2008) Electrochim Acta 53:2650–2655 6. Nourry C et al (2009) J Appl Electrochem 39:927–933 7. Nourry C et al (2009) J Appl Electrochem 39:2359–2367 8. Xu GL et al (2012) Int J Mater Res 103:1179–1187 9. Mamantov G, Manning DL (1966) Anal Chem 38:1494–1498 10. Mamantov G, Manning DL (1968) J Electroanal Chem 18: 309–314 11. Jenkins HW et al (1969) J Electrochem Soc 116:1712–1714 12. Clayton FR et al (1974) J Electrochem Soc 121:86–90 13. Hamel C et al (2007) Electrochim Acta 52:3995–4003 14. Nourry C et al (2012) J Nucl Mater 430:58–63 15. Baes CF (1969) In: Chiotti P (ed) Reprocessing of nuclear fuels, US AEC Division of Technical Information 16. Straka M et al (2010) J Radioanal Nucl Chem 284:245–252 17. Wang CP et al (2009) J Alloy Compd 487:126–131
123
