NE ' ", " 1 L . . ni UTO 1992 UTAMA | Jh . . . 17 17 . UNCLASSIFIED ORNL TH HA . T ./ wi . . . W WW. 21 . . . - Rom AVE 324 ORNU-P-324 CONF-742-1 - * .* - - - - - VOLT AMMETRIC MEASUREMENTS IN MOLTEN FLUORIDE SALT SYSTEMS, D. L. Manning and John M. Dale Analytical Chemistry Division Oak Ridge National Laboratory Oak Ridge, Tennessee and Gleb Mamantov Department of Chemistry University of Tennes nee Knoxville, Tennessee Reactor Chemistry Division Oak Ridge National Laboratory Oak Ridge, Tennessee -LEGAL NOTICE - The report memorandum nocow Quroru sponsor wort. Matthat the Unie here, wer the Counci, wo Myporna estis on of the Connection: A. Kako my warranty or reprendom, pred or le puedo wa roopact who me mcy, completnona, or watel o the tabornado contained to the report, « Har dhe me al may baloration, rent, moch, w mocu dachowed to date mare may not tattoo petrawly owned rights or B. And thenwuns vol repect to the w ol, or for denne mette in the w olnog orution, warten, method, or precios decloud ba dua mort No wond in de above, permanente a bola de Contenta" med ". moyo o contractor of the Colector w play. mucha contractor, the actual that meal player contractor the Courtstum or proyee of me coructor prepara dermatos, a fronte meten, meg moration permet la player or contact wted the contestou, or wo matagal marc crctur. Research sponsored by the U. S. Atomic Energy Commission under contract with the Union Carbide Corporation For presentation at the 15th CITCE Meeting, London England, September 20-26, 1964. International Committee on Thermodynamics and Electrochemical Kinetics F . -2. ABSTRACT Electroanalytical studies of reducible species in molten fluorides are being conducted at the Oak Ridge National Laboratory. Such studies are of interest both from a fundamental point of view and as a possible means of performing in situ analyses of mo.lten fluoride salt systems which are considered as constituents of reactor fuels. Controlled-potential voltammetry was utilized to study the electro- - - . . E . chemical behavior of iron(II) and nickel(II) in molten LiF-NaF-KF (46.5 - 11.5 - 42 mole %) and also in molten LiF-BeF2 (66 - 34 mole %). Iron and -- nickel are corrosion products which arise from the corrosive action of the molten fluorides on the metal containment systems. For the controlled-potential voltammetric studies, the melt (~ 40 ml) is contained in a graphite cell which is enclosed in a quartz jacket to maintain either a vacuum or a controlled atmosphere. Stationary dip-type indicator electrodes fabricated from platinum, silver, tungsten and pyrolytic graphite were utilized. Pyrolytic graphite encased in boron nitride is the most satisfactory indicator electrode tested so far. Boron nitride, which is an insulator at the experimental temperatures (500-600°c), protects the layered edges of the "c" planes while permitting only the impervious surface of the "a" plane to be exposed to the melt. Platinum electrodes were employed as quasi-reference and counter electrodes. Peak-shaped current-voltage curves were obtained for the reduction of iron and nickel when the rate of voltage scan was ~l v/min and greater; this implies that diffusion rather than convection is the predominant mass transfer process under these experimental conditions. The half-wave potentials for the reduction of iron and nickel are approximately -0.6 and -0.2 v, respectively, vs the platinum quasi-reference electrode. The peak current is proportional to -3- the concentration of the electroactive species in the melt and also to the square root of the rate of voltage scan. Diffusion coefficients of approxi- mately 1 x 100 cm/sec were calculated for both species at 500°c; from a plot of log D vs 1/T, an activation energy (E) of about 12 kcal/mole and 18 kcal/ mole was calculated for the reduction of iron(II) and nickel(II) to the metal, respectively, in molten LiF-NaF-KF. Electroanalytical chemistry, and in particular, voltammetry (polaro- graphy) in molten salts has recently been thoroughly reviewed by Liu, Johnson and Laitinen (1). It is apparent from this review that no polar- ographic studies have been performed in molten fluoride solvents. This, undoubtedly, is because glass or quartz, which are commonly used as insu- · lating or container materials, are attacked by molten fluorides. The use of other materials, therefore, becomes necessary. The use of molten fluoride solvents for electroanalytical measurements is under investigation . at the Oak Ridge National Laboratory; this paper is a summary of the studies · to date on voltammetric measurements. The molten fluorides are interesting both from the practical and the fundamental point of view. In the Molten Salt Reactor Experiment (Oak Ridge National Laboratory) scheduled to begin operation in 1964, the molten fuel will be composed of LiF-BeF2-ZrF4-UF.(65.0 - 29.2 - 5.0 - 0.83 mole %) and the coolant will be LiF-BeF2 (66 -34 mole %). The reactor will operate at ~ 660° C. The determination of impurities, such as the corrosion products iroň, nickel, chromium; and in addition hydrogen and oxide ions, dissolved oxygen, and different oxidation states of uranium by the electroanalytical means directly in the molten salt solvent would be a very attractive procedure. 23 Better understanding of the species present in the molten fluorides 18, of course, very desirable. Three molten fluoride solvents, LiF-NaF-KF (46.5 - 11.5 - 42 mole %), LiF-BeF2 ( 66 - 34 mole %), and LiF-BeF2-ZrF4 (65.6 - 30.7 - 3.7 mole %) have been used to date for electrochemical studies. The solvents have been prepared and purified in the reactor chemistry division of the Oak Ridge National Laboratory; the purification process has been described previously (2). The impurity content of a typical batch of LiF-BeF2 following purifi- cation by hydrogen-hydrogen fluoride treatment (2) is shown in Table I. It is apparent that the impurities are present in such concentrations that, at least in principle, they may be determined by electroanalytical meang. It is also apparent that the solvent has to be purified further in order to obtain a melt that may be considered pure from the electrochemical point of view. Both chemical purification (hydrogen-hydrogen fluoride treatment) and pre-electrolysis at a controlled jotential have been employed. “More recently the purification process of LiF-BeF2 and LiF-BeF2-2rF4 solvents has involved the treatment of the molten mixture with metallic beryllium in adałtion to the hydrogen-hydrogen fluoride step followed by the filtration of the solvent through a sintered nickel filter; the solvents obtained by this method are much purer than those obtained without the beryllium treatment. The melt (~40 ml) is contained in a graphite which is approximately 2 in. in diameter and 4-1/2 in. long. The cell is fitted with either a porous graphite or a thin-wall boron nitride thimble type inner compartment which accommodates the isolated counter electrode. To maintain a vacuum or controlied atmosphere, the cell is enclosed in a quartz jacket approximately 2-1/2 in. in diameter and 10 in. long. The top of the quartz jacket is fitted with a 102/75 ball and socket joint, the socket portion of which serves as a removable cap for the enclosure. On the cap are located four 1/4 in. and two 3/8 in. Kovar glass to metal seals. Four 1/4 in. to 3/16 10. and two 3/8 in. to 1/4 in. Swagelok compression type reducing fittings (bored through) are attached to the metal portion of the Kovar seals. Then fittings provide access to the melt for the various electrodes, thermocouple and helium bubbling tube. Vacuum and inert atmosphere connections were provided for by means of a three-way vacuum stopcock which is also located on the cap of the quartz enclosure. A three-electrode system is employed. The best indicator electrode found to date 18 a pyrolytic graphite (PG) electrode. A non-insulated PG electrode was used in molter LiCl-KCI by Laitinen and Rhodes (3). A typical PG electrode is shown schematically in Fig. 1. This electrode is prepared by brazing an 18-gauge platinum wire directly to PG at 950º C under vacuum. The brazing process is described in a paper by Miller and Zittel (4). One must be cautious to have the platinum wire normal to the "a" plane of the PG (the plane that is meen by the melt). A Ti-Cu-Be (48-48-4 %) brazing alloy 18 'used. The electrode is encased in hot-pressed boron nitride with tolerances as close as possible by machining techniques. To obtain the best possible fit between the pyrolytic graphite and the boron nitride, it has been found useful to machine the two pieces in the form of a standard taper; after the two pieces are put together, the protruding surface (usually BN) is removed with a fine emery paper. Boron nitride is an insulator at the experimental temperatures (500 - 650° C) (5) and it is chemically resistant to molten fluorides, although it appears to be slowly penetrated by them (6). The extent of pene. tration is being studied by metallographic and X-ray powder diffraction techniques. To complete the PG electrode, the end of the platinum wire is gold-soldered to a 1/8 in. nickel rod. A finished PG electrode and a platinum electrode encased in boron nitride are shown in Fig. 2. A word of caution: boron nitride tends to absorb moisture and should be kept in a dry environment. The potential of the .6. indicator electrode 18 measured against a platinum quasi-reference electrode (a 1/8 in. platinum rod). At the present time there is no completely satis- factory reference electrode for work in molten fluorides; however, the use of platinum electrodes with a large surface area 18 quite widespread in molten salts (7, 8). Another platinum electrode, located in a separate compartment, is used as a counter electrode. A fourth electrode, a 3/16 in. graphite rod, was used as the cathode in the purification of the melt by pre-electroly818. An ORNL Model Q-1988 controlled-potential polarograph (9) was used to record the current-voltage curves. The polarograph was modified by incorporating electronically controlled scan rates from 0.05 to 10 v/min. The current measura ing range of the instrument was increased from 500 ma to 5 ma per full scale deflection of the recorder. Moseley X-Y recorders (Models 3S and 20-2A) were used to record the polarograms. It is well known (10) that diffusion controlled electrode processes at stationary electrodes result in current-voltage curves exhibiting a maximum, provided sufficiently large voltage scan rates are employed in order to minimize convection. The effect of the scan rate on the current-voltage curves is shown in Fig. 3 for the reduction of iron(II) in molten LiF-NaF-KF. It is apparent that under the experimental conditions used, scan rates of ~1 v/min and greater result in peak-shaped polarograms. This provides evidence that diffusion rather than convection is the dominant mass transfer process, provided sufficiently large scan rates are used. The expression for the peak current, in, for the reversible deposition of an insoluble substance at 500° C 18 given by (10) 1 = 2.28 x 105 13/2 AD"/2c v4/2 10 = peak current, amperes n = electron change (1) where C = concentration, mole om A = electrole area, come D = diffusion coefficient, cm2/sec V = potential scan rate, volts/sec. Thus, in should be a linear function of y , and this, indeed, was found to be the case (Fig. 4). The value of D calculated from the equation for the peak current 18 1 x 108 cm2/sec in LiF-NaF-KF and 5 x 10 cm Fe in LiF-BeF2 was found to be pro- portional to the molality of Fe(II) (in the range 2 - 5 x 10" molal) which demonstrates the analytical utility of such measurements. The relative stand- ard deviation for 54 runs was w13%. Well defined peak-shaped polarograms were also obtained for the reduction of nickel at the pyrolytic graphite electrode in molten LiF-NaF-KF 3s shown in Fig. 6. The ball-peak potential occurs at approximately -0.2 volts vs. the platinum quasi-reference electrode. The peak current is proportional to the concentration of nickel (in the range 20 to 80 p.p.m) and also to the square .-8. root of the rate of voltage scan. A diffusion coefficient of approximately 1 x 10-8 cm2/sec was calculated for the electroactive species at 500° C. From plots of log D vs. 1/T, an activation energy of about 12 kcal/mole and 18 kcal/mole was calculated for the reduction of iron(II) and nickel(II) to the metal, respectively, in molten LiF-NaF-KF. niet In general, anodic stripping techniques are more sensitive than methods pyrolytic graphite electrode are shown in Fig. 7. A plating time of 4 minutes at a potential of -0.4 volt vs. a platinum quasi-reference electrode was utilized with the melt being stirred by helium bubbling during the plating aná stripping process. There is slight deviation from linearity in the plots of area (planimeter units) versus ppm nickel; nevertheless, the analytical utility of such measurements is demonstrated, especially at trace levels of nickel. The reproducibility of the area under an anodic stripping curve from consecutive runs under any one set of conditions is of the order of 5%. Current-voltage curves recorded at 500° C on the eutectic LiF-BeF2-ZrF4, which is the solvent electrolyte for the Molten Salt Reactor Experiment, revealed that it is relatively clean with respect to reducible impurities after purification by the hydrogen, HF, beryllium metal treatment. This agrees with the chemical analyses of the melt which show < 40 ppm of Fe and <5 ppm N1 and other metallic impurities. The potential limit of the melt occurs at -1.5 volts vs. the platinum quasi-reference electrode. This is probably due to the reduction of zirconium. For the LiF-BeF2 eutectic (MSRE coolant electrolyte) the potential limit due to the reduction of beryllium occurs at ~-1.7 volts vs." a platinum quasi-reference electrode. When this salt was purified in the same manner as the fuel salt, subsequent current-voltage curves again revealed no reducible impurities remaining in the melt. The effectiveness of the more recent purification procedure was demonstrated by omitting the beryllium metal and using only hydrogen and HF; under these conditions, the LiF-BeF2 and also the LiF-NaF-KF was found to contain approximately 200 ppm iron and 10 ppm nickel by chemical analyses. Current-voltage curves and also anodic stripping curves were observed for these impurities at this level of concentration. In molten LiF-NaF-KF, the reduction of alkali !8 observed at approximately -2.2 volts versus the platinua quasi-reference electrode. We are continuing the voltammetric studies and are also working on the development of a reference electrode to be used in these melts. REFERENCES 1. C. H. Liu, K. E. Johnson and H. A. Laitinen, "Electroanalytical Chemistry of Molten Salts," Molten Salt Chemistry, ed. by M. Blander, Interscience, 1964. 2. W. R. Grimes, D. R. Cuneo, F. F. Blankenship, G. W. Keilholtz, H. F. Poppendiek, and M. T. Robinson, "Chemical Aspects of Molten Fluoride- Salt Reactor Fuels, Fluid Fuel Reactors, ed. by J. A. Lane, A. G. MacPherson, and Frank Moslan, Addison-Wesley, Reading, Mass., 1958, p.584. 3. H. A. Laitinen and D. R. Rhodes, J. Electrochem. Soc. 109, 413 (1962). 4. F. J. Miller and H. E. Zittel, Anal. Chem. 35, 1866 (1963). 5. Private communication, Electronics Division, The Carborundum Company, Latrobe, Pennsylvania. 6. E. W. Yim and M. Feinleib, J. Electrochem. Soc. 104, 622 (1957). 7. IU. K. Delimarskii and B. F. Markov, "Electrochemistry of Fused Salts," Sigma Press, Washington, D. C., 1961, p. 302. ŕ o 8. R. W. Laity, "Electrodes in Fused Salt Systems, " Reference Electrodes, ed. by D. J. G. Ives and G. J. Janz, Academic Press, New York and London, 1961, p. 579. 9. M. T. Kelley, H. C. Jones, and D. J. Fisher, Anal. Chem. 31, 1475 (1959). 10. P. Delahay, New Instrumental Methods in Electrochemistry, Interscience, New York, 1954, pp. 114-145. 11. C. 8. Liu, "Diffusion coefficients of Metal Ions in Molten Salts," • Handbook of Analytical Chemistry, ed. by L. Meites, McGraw-Hill, 1963, pp. 5 - 225-6. 12. Private Comunication, W. T. Ward, Oak Ridge National Laboratory. . . 10. ORNL DWG. 63-5616 Table I. Typical Impurity Level 10 LIF-Beļa (66-34 mole %). . 7 . - UNCL ASSIFIED ORNL DWG. 63.5616 TYPICAL IMPURITY LEVEL IN LIF-BeF, * (66–34 mole %) ppm MOLALITY MOLARI TY* 258 1 ju j ż o 144 4.6 x 10-3 2.8 x 10-3 7.1 x 10-4 9.1x 10-3 5.4 x 10-3 1.4 x 10-3 3.3 x 10-4 3.5 x 10-2 1.7x10-4 1.8 x 10-2 566 <3 < 300 <5 *d = 1.96 g/cm2 (at 500°C): AM ORNL-DWG. 63-1331 Pg. lSchematic Drawing of the pyrolytic Graphite electrode. WWW Ph . .. BLUE.NL uli LLLLLLL II .LT UNCLASSIFIED ORNL-DWG. 63-1331 NICKEL ROD 1/8" Dia. PLATINUM WIRE GOLD SOLDER JOINT BRAZE JOINT BORON NITRIDE 1/4" Dia. PYROLYTIC GRAPHITE 1/8" Dia. 12. ORNL Photo No. F18. 2. Pyrolytic Grapaite and Platinum Indicator Hectrodes. * 4 . برج OAK RIDGE NATIONAL LABORATORY . . . . ... .. . ... .. .. .. . ........... . . . . - .. .... ....... ......... ... ... ORNL-DWG. 63-1330 F18. 3. Effect of the potential scan Rate on the Reduction of Fe(II) 10 Molten LiF-NaF-KF. ***00.0**150 w op . 2 0.4... UNCLASSIFIED ORNL-OWG. 63-1330 Iron (II) ~ 9.5m Mi Electrode Area ~ 8 mm? 100 HA. Scan Rate (volts/min.) 5 Current , HA. 3 11 0.1 0.05 -0.5 -0.6 -0.7. -0.8 -0.9 APPLIED POTENTIAL • ( Volts vs Platinum Quasi-Reference Electrode.) .- - - ORNL-DWG. 63-5407 3:18. 4. Peak Current Dependence on wile for the Reduction of Iron(II) 10 Molten LiF-NaF-KT. - • . -. -'.'. . mn. varls -- :- - ..-... ... . . . . . .. - - . - UNCLASSIFIED ORNL-DWG. 63-5407 2004 : - 150 PEAK CURRENT, HA - - - 0.1 0.2 0.3 0.4 . ORNL-DWG. 63.4116 F18. 5. Current-Voltage Curves for the Reduction of Fe(II) La Molten LiF-NAP-KF at Different Electrode Materials. Conc. Fe(II), 9.5 » M Temperature, 500°c. UNCLASSIFIED ORNL. DWG. 63.4116 PLATINUM PYROLYTIC GRAPHITE | 100 HA 85 A Scon Ron volls/min 10 scon Rate, volts/ min 10 0.2 005 •0:4 -0.5 -0.6 -0.7 -0.8 -0.4 -0.5 0.6 0.7 0.7 0.8 SILVER TUNGSTEN 854A Scan Rote, volts/min Scon Rate, volts/min 0.5 0.2 0.1 -0.4 -0.5 -0.6 -0.7 -0.8 -0.4 -0.5 -0.6 -0.7 -28 APPLIED POTENTIAL ( volts vs platinum quasi-reference electrode) .16. ORNL-DWG. 63-8003 Fig. 6. Effect of Scan Rate on the Current-Voltage Curves for Nickel. Concentration a-Nickel in the Melt, 23 ppm. f ? 2 + UNCLASSIFIED ORNL-DWG. 63-8003 7 A Scan Rate (v/min) CURRENT Oliol mio1010 0.05 ) -0.1 -0.2 -0.3 -0.4 -0.5 APPLIED POTENTIAL (Volts vs Plcrinum Quasi- Reference Electrode ) . .17. ORNL-DWG. 63-8002 Fig. 7. Anodic Stripping Curves for Nickel from a Pyrolytic Graphite Electrode. Plating time, 4 min. Plating potential, -0.4 volt. Solvent, LiF-NaF-KF, 500° C. * ** . - - . . . * rases h , IN UNCLASSIFIED ORNL-OWO. 03-8002 Ni (ppm) AREA (Planimeter Units) 20 140 HA CURRENT 0.3 V. .6 sec + 0.2 - 0.1 -0.4 APPLIED POTENTIAL (Volts vs Platinum Quasi-Reference Electrode) - IS : F TV *- Ł* 12 i BE3 'n F . 7. DATE FILMED 12/ 1 /164 -- " O . 1. . . 3 i . . 11 = 14 1 * - - LEGAL NOTICE ' ' F - . * 4 * - . . -!. . This report was prepared as an account of Government sponsored work. Neither the United Sutes, nor the Commission, nor any person acting on behalf of the Commission: A. Makes any warranty or representation, expressed or implied, with respect to the accu- racy, completeness, or usefulness of the information contained in this report, or that the use of any information, apparatus, method, or process disclosed in this report may not Infringe privately owned rights; or B. Assumes any liabilities with respect to the use of, or for damages resulting from the use of any information, apparatus, method, or process disclosed in this report. As used in the above, "person acting on behalf of the Commission" Includes any em- ployee or contractor of the Commission, or employee of such contractor, to the extent that such employee or contractor of the Commission, or employee of such contractor prepares, disseminates, or provides access to, any information pursuant to his employment or contract with the Commission, or his employment with such contractor. - ' Het _ i - . a : : . * * 2 . Ses . * . FA, - 1. END 711 TE AT . * . no