METHODS AND APPARATUSES FOR EFFICIENT METALS PRODUCTION, SEPARATION, AND RECYCLING BY SALT- AND ARGON-MEDIATED DISTILLATION WITH OXIDE ELECTROLYSIS, AND SENSOR DEVICE RELATED THERETO

Abstract

In one aspect, the present invention is directed to methods and apparatuses for recovering target metals from scrap. In some embodiments, the methods comprise dissolving a portion of a mixed metal scrap into a molten salt to form a molten salt and metal mixture, the scrap including a target metal species and at least one contaminant metal species; bubbling a gas through the molten salt and metal mixture to form a gas and metal vapor mixture comprising target metal vapors; and condensing at least a portion of the target metal vapors. In some embodiments, the apparatuses comprise a housing; a divider at least partially disposed within the housing, the divider forming at least a first chamber, a second chamber, and a fluid conduit between the first and second chambers; a top wall cooperating with a lower housing wall and at least one of the plurality of side walls to enclose the second chamber; a plurality of gas inlets disposed in the second chamber; and a gas outlet in fluid communication with the second chamber.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of and priority to U.S. provisional patent applications Ser. No. 61/544,879, filed Oct. 7, 2011, entitled “Metals Production Efficiency Improvement in SOM Electrolysis and Refining”; Ser. No. 61/609,309, filed Mar. 10, 2012, entitled “Method and Apparatus for Efficient Metals Production, Separation, and Recycling by Salt- and Argon-Mediated Distillation with Oxide Electrolysis, and Sensor Device Related Thereto”; and Ser. No. 61/609,366, filed Mar. 12, 2012, entitled “High-Throughput High-Efficiency Primary Production and Molten Salt-Assisted Recycling of Metals”, the disclosure of each is hereby incorporated by reference in its entirety for all purposes.

All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety. The patent and scientific literature referred to herein establishes knowledge that is available to those skilled in the art. The issued patents, applications, and other publications that are cited herein are hereby incorporated by reference to the same extent as if each was specifically and individually indicated to be incorporated by reference. In the case of inconsistencies, the present disclosure will prevail.

STATEMENT OF GOVERNMENT SUPPORT

This invention was made with government support under Grant DE-EE0003454, awarded by the U.S. Department of Energy; Grant 1026639, awarded by the National Science Foundation. The government has certain rights in the invention.

FIELD OF THE INVENTION

The invention relates to production and recycling of metals.

BACKGROUND OF THE INVENTION

Several processes for primary production of metals from their oxides have used electrolysis on an industrial scale since the invention of the Hall-Héroult cell for aluminum production in 1886 (see, e.g., U.S. Pat. No. 400,664; herein incorporated by reference in its entirety). When the raw material is water-soluble and the product metal is not very reactive, then this can be done at ambient temperature in an aqueous electrolyte, e.g. electrolysis of copper chloride to make copper metal and chlorine gas. For others like aluminum oxide, it is necessary to dissolve the raw material in a molten salt electrolyte such as cryolite, which in turn requires high temperature cell operation.

There is a particular problem with electrolysis of metal oxides where the metal itself is soluble in the molten salt. For example, electrolysis of neodymium oxide in a molten fluoride salt suffers from low current efficiency in the 50-70% range because neodymium metal dissolves into the molten salt and promotes conduction of electrons through it (See, e.g., R. Keller and K. T. Larimer, “Electrolysis of Neodymium Oxide,” Final Report to the U.S. Department of Energy under Contract FC07-91ID13104, May 1997; herein incorporated by reference in its entirety). Magnesium and calcium have sufficient solubility in molten salts that they both react with gases formed at the anode, such as CO₂ or oxygen, also reducing current efficiency to below 50% (J. Phys. Chem. 74(22):3896-3900, 1970; herein incorporated by reference in its entirety).

Use of an oxygen-ion-conducting membrane between the molten salt and anode, such as stabilized zirconia, improves current efficiency considerably by presenting a barrier between the metal produced at the cathode and oxidizing gases produced at the anode, preventing back-reaction (See, e.g., U.S. Pat. Nos. 5,976,345 and 6,299,742; each herein incorporated by reference in its entirety). However, electronic conduction through the molten salt can result not only in lower current efficiency, but also damage to the membrane itself.

This invention describes methods to prevent metal dissolution in the molten salt and/or to remove it from the molten salt, thereby improving current efficiency and membrane lifetime. Such methods can be useful for removing dissolved metal from the molten salt, which can be useful for separation of metals due to differential solubility in the molten salt.

BRIEF SUMMARY OF THE INVENTION

In a first aspect of the invention, a method for recovering a target metal is provided, comprising: (a) providing a molten salt; (b) dissolving a portion of a mixed metal scrap into the molten salt to form a molten salt and metal mixture, the mixed metal scrap including a target metal species and at least one contaminant metal species; (c) bubbling a gas through the molten salt and metal mixture to form a gas and metal vapor mixture, the gas and metal vapor mixture including target metal vapors; and (d) condensing at least a portion of the target metal vapors.

In some embodiments, the mixed metal scrap comprises an oxide of the target metal. In some embodiments, the gas is an inert gas. In some embodiments, the inert gas is argon. In some embodiments, the dissolving step comprises melting the mixed metal scrap.

In a second aspect of the invention, an apparatus for recovering a target metal is provided, comprising: (a) a housing including a lower wall and a plurality of side walls; (b) a divider at least partially disposed within the housing, the divider forming within the housing at least a first chamber, a second chamber, and a fluid conduit between the first and second chambers; (c) a top wall cooperating with the lower wall and at least one of the plurality of side walls to enclose the second chamber; (d) a plurality of gas inlets disposed in the second chamber; and (e) a gas outlet in fluid communication with the second chamber. In some embodiments, the divider forms floating metal piers. In some embodiments, the gas inlets are positioned between the piers. In some embodiments, the apparatus further comprises (f) a weir to recover floating metal.

In a third aspect of the invention, a method for recovering a target metal is provided, comprising: (a) providing a cathode in electrical contact with a molten salt; (b) providing an anode in electrical contact with the molten salt; (c) dissolving an oxide of a target metal into the molten salt to form a molten salt and metal ion mixture; (d) establishing an electrical potential between the cathode and the anode to form target metal at the cathode; (e) bubbling a gas through the molten salt and metal ion mixture to form a gas and metal vapor mixture, the gas and metal vapor mixture including target metal vapors comprised at least in part by a portion of the target metal formed at the cathode; and (f) condensing at least a portion of the target metal vapors. In some embodiments, the bubbling the gas through the molten salt and metal ion mixture includes bubbling the gas in immediate proximity to the cathode. In some embodiments, the bubbling the gas through the molten salt and metal ion mixture includes providing the gas through at least one opening in the cathode. In some embodiments, the bubbling the gas through the molten salt and metal ion mixture includes providing the gas across a current path through the molten salt between the cathode and the anode. In some embodiments, the gas is an inert gas. In some embodiments, the inert gas is argon. In some embodiments, the cathode includes perforations or other means of introducing the gas immediately adjacent to the cathode reaction location, such that the reduced metal first forms as a dilute vapor in the gas, instead of a pure or concentrated metal vapor or liquid. In some embodiments, the method further comprises measuring the quantity of target metal in the molten salt. In some embodiments, the method further comprises providing a SOM between the cathode and the anode. In some embodiments, the oxide of the target metal further comprises at least one contaminant metal.

In a fourth aspect of the invention, an apparatus for recovering a target metal is provided, comprising: (a) a container for holding a molten salt; (b) a cathode disposed to be in electrical contact with the molten salt when the molten salt is disposed in the container; (c) an oxide ion-conducting membrane disposed to be in ion-conducting contact with the molten sale when the molten salt is disposed in the container; (d) an anode in electrical contact with the oxide ion-conducting membrane; (e) a power source for generating an electric potential between the anode and the cathode; (f) a gas inlet having at least one end disposed within the container to be below a level of the molten salt when the molten salt is disposed in the container; and (g) a gas outlet in fluid communication with a volume defined by the container. In some embodiments, the at least one end of the gas inlet is in immediate proximity to the cathode. In some embodiments, the cathode includes the at least one end of the gas inlet. In some embodiments, the at least one end of the gas inlet is disposed in the container to form gas bubbles in the molten salt, when the salt is disposed in the container, across a current path through the molten salt between the cathode and the anode. In some embodiments, the gas inlet is a nozzle. In some embodiments, the gas inlet is disposed at the bottom of the container. In some embodiments, the cathode forms at least portion of container.

In a fifth aspect of the invention, a method for recovering a target metal is provided, comprising: (a) providing a molten salt; (b) dissolving a portion of a mixed metal scrap into the molten salt to form a molten salt and metal mixture, the mixed metal scrap including a target metal species and at least one contaminant metal species; (c) reducing a pressure of the molten salt and metal mixture to remove, as vapors of the target metal, at least a portion of the target metal species dissolved in the molten salt and metal mixture; and (d) recovering at least a portion of the target metal vapors. In some embodiments, the reducing a pressure of the molten salt and metal mixture comprises creating a partial vacuum in an atmosphere overlying the molten salt and metal mixture. In some embodiments, the scrap comprises an oxide of the target metal. In some embodiments, the dissolving step comprises melting the scrap metal mixture. In some embodiments, the method further comprises providing SOM elements and performing SOM electrolysis. In some embodiments, the method further comprises dissolving an oxide of a second metal subsequent to production of at least some of the first metal in the salt.

In a sixth aspect of the invention, an apparatus for recovering a target metal is provided, comprising: (a) a container for holding a molten salt; (b) a cathode disposed to be in electrical contact with the molten salt when the molten salt is disposed in the container; (c) an oxide ion-conducting membrane disposed to be in ion-conducting contact with the molten salt when the molten salt is disposed in the container; (d) an anode in electrical contact with the oxide ion-conducting membrane; (e) a power source for generating an electric potential between the anode and the cathode; (f) a gas outlet in fluid communication with a volume defined by the container; (g) a condenser in fluid communication with the gas outlet for condensing at least a portion of the target metal vapor in a gas stream exiting the container; and (h) the vacuum source in fluid communication with the gas outlet and/or the container for creating at least a partial vacuum in the volume defined by the container.

In a seventh aspect of the invention, a method for recovering a metal from an oxide of said metal is provided, comprising: (a) providing a cathode in electrical contact with a molten salt; (b) providing an anode in electrical contact with the molten salt; (c) dissolving an oxide of a first metal into the molten salt to form a molten salt and metal ion mixture; (d) establishing an electrical potential between the cathode and the anode to form first metal at the cathode; (e) dissolving an oxide of a second metal into the molten salt, the second metal being more electronegative than the first metal, and the second metal being less soluble in the molten salt than the first metal; (f) subsequent to dissolving the oxide of the second metal into the molten salt, establishing an electrical potential between the cathode and the anode to form first metal at the cathode; and (g) recovering at least a portion of the first metal formed at the cathode. In some embodiments, the oxide of the second metal comprises nickel oxide. In some embodiments, the metal charge comprises contaminant metals and/or the target metal.

BRIEF DESCRIPTION OF THE DRAWINGS

The following figures are illustrative only and are not intended to be limiting.

FIG. 1 shows the theoretical electrorefining potential for magnesium bubble nucleation versus magnesium-aluminum scrap anode composition for several values of activity coefficient γ (from equation 1).

FIG. 2 shows an illustrative embodiment of the recycling process, showing electrodes for electrolysis and for measuring magnesium content in scrap anode and molten salt.

FIG. 3 shows current-voltage relationships for an illustrative refining process at various times. The open circuit voltage and two electrorefining potentials (OCV, E_ER1, and E_ER2) for the last scan PDS5 are indicated by arrows.

FIG. 4 shows electrorefining potentials for bubble nucleation (triangles) and OCV (circles) during an illustrative refining embodiment.

FIG. 5 shows a schematic illustration of embodiments of cathode configurations according to embodiments of the invention. A: holes, B: notches, C: mesh/screen, and D: porous material.

FIG. 6 shows a schematic illustration of embodiments of a single-tube electrolysis apparatus operating at low pressure according to embodiments of the invention.

FIG. 7 shows a schematic illustration of embodiments of a production and recycling apparatus according to embodiments of the invention.

FIG. 8 shows a schematic illustration of embodiments of a production and recycling apparatus according to embodiments of the invention.

FIG. 9 shows a schematic illustration of embodiments of a production and recycling apparatus according to embodiments of the invention.

FIG. 10 shows a schematic illustration of embodiments of a production and recycling apparatus according to embodiments of the invention.

FIG. 11 shows a schematic illustration of embodiments of a production and recycling apparatus according to embodiments of the invention.

FIG. 12 shows potentiodynamic scans before electrolysis, after electrolysis, and after electrolysis and argon stirring.

FIG. 13 shows potentiodynamic scans before electrolysis, after each of two electrolysis runs using a notched argon tube as cathode, and after a third electrolysis run using the crucible as the cathode.

FIG. 14 shows potentiodynamic scans (5 mV/s) after 3 hours of electrolysis at 1 atm pressure and after 7.5 hours of electrolysis at 0.08 atm pressure.

FIG. 15 shows a schematic illustration of embodiments of a production and recycling apparatus according to embodiments of the invention.

FIG. 16 shows an initial potentiodynamic scan (5 mV/s).

FIG. 17 shows impedance before electrolysis 1.

FIG. 18 shows electrolysis 1 at 2 V for 3 hours.

FIG. 19 shows a potentiodynamic scan at the cathode after electrolysis 1.

FIG. 20 shows a potentiodynamic scan at the anode after electrolysis 1.

FIG. 21 shows electrolysis 2.

FIG. 22 shows potentiodynamic scans after electrolysis 2 using the stirring tube as cathode or the steel crucible as cathode.

FIG. 23 shows electrolysis 3 using the steel crucible as cathode.

FIG. 24 shows potentiodynamic scans after electrolysis 3 using the stirring tube as cathode or the steel crucible as cathode.

FIG. 25 shows current vs. cathode potential after electrolysis 3.

FIG. 26 shows current vs. anode potential after electrolysis 3.

FIG. 27 shows a schematic illustration of embodiments of a refining and SOM electrolysis method according to embodiments of the invention.

FIG. 28 shows a schematic experimental setup of a refining and SOM electrolysis apparatus according to embodiments of the invention.

FIG. 29 shows temperature steps at the center of an electrorefiner chamber according to embodiments of the invention.

FIG. 30 shows a potentiodynamic scan taken at 13 minutes after the maximum operating temperature was reached.

FIG. 31 shows impedance spectroscopy measured 20 minutes after the maximum operating temperature was reached.

FIG. 32 shows potentiodynamic scans taken at 13 and 38 minutes after the maximum operating temperature was reached.

FIG. 33 shows potentiodynamic scans taken at 13, 38, and 58 minutes after the maximum operating temperature was reached.

FIG. 34 shows potentiodynamic scans taken at 13, 38, 58, and 68 minutes after the maximum operating temperature was reached.

FIG. 35 shows a potentiostatic scan taken at 74 minutes after the maximum operating temperature was reached.

FIG. 36 shows potentiodynamic scans taken at 13, 38, 58, 68, and 79 minutes after the maximum operating temperature was reached.

FIG. 37 shows potentiodynamic scans taken at 13, 38, 58, 68, 79 and 83 minutes after the maximum operating temperature was reached.

FIG. 38 shows a potentiostatic scan taken at 85 minutes after the maximum operating temperature was reached.

FIG. 39 shows potentiodynamic scans taken at 13, 38, 58, 68, 79, 83 and 96 minutes after the maximum operating temperature was reached.

FIG. 40 shows potentiodynamic scans taken at 13, 38, 58, 68, 79, 83, 96 and 113 minutes after the maximum operating temperature was reached.

FIG. 41 shows a potentiostatic scan taken at 115 minutes after the maximum operating temperature was reached.

FIG. 42 shows potentiodynamic scans taken at 13, 38, 58, 68, 79, 83, 96, 113 and 148 minutes after the maximum operating temperature was reached.

FIG. 43 shows potentiodynamic scans taken at 13, 38, 58, 68, 79, 83, 96, 113, 148 and 172 minutes after the maximum operating temperature was reached.

FIG. 44 shows a potentiostatic scan taken at 187 minutes after the maximum operating temperature was reached.

FIG. 45 shows potentiodynamic scans taken at 13, 38, 58, 68, 79, 83, 96, 113, 148, 172 and 218 minutes after the maximum operating temperature was reached.

FIG. 46 shows a potentiostatic scan taken at 228 minutes after the maximum operating temperature was reached.

FIG. 47 shows potentiodynamic scans taken at 13, 38, 58, 68, 79, 83, 96, 113, 148, 172, 218 and 259 minutes after the maximum operating temperature was reached.

FIG. 48 shows impedance spectroscopy measured at 267 minutes after the maximum operating temperature was reached.

FIG. 49 shows a potentiostatic scan taken at 274 minutes after the maximum operating temperature was reached.

FIG. 50 shows potentiodynamic scans taken at 13, 38, 58, 68, 79, 83, 96, 113, 148, 172, 218, 259 and 285 minutes after the maximum operating temperature was reached.

FIG. 51 shows a potentiostatic scan taken at 287 minutes after the maximum operating temperature was reached.

FIG. 52 shows potentiodynamic scans taken at 13, 38, 58, 68, 79, 83, 96, 113, 148, 172, 218, 259, 285 and 298 minutes after the maximum operating temperature was reached.

FIG. 53 shows first and second impedance spectroscopy measurements.

FIG. 54 shows a potentiodynamic scan taken at 11 minutes after the SOM process was started.

FIG. 55 shows a potentiostatic scan taken at 17 minutes after the SOM process was started.

FIG. 56 shows a potentiodynamic scan taken at 82 minutes after the SOM process was started.

FIG. 57 shows impedance spectroscopy measured at 111 minutes after the SOM process was started.

FIG. 58 shows a potentiostatic scan taken at 119 minutes after the SOM process was started.

FIG. 59 shows a potentiodynamic scan taken at 181 minutes after the SOM process was started.

FIG. 60 shows potentiodynamic scans taken at before the SOM was started and during the first hour of electrolysis.

FIG. 61 shows for the first hour of electrolysis and the second hour of electrolysis. Both electrolyses were performed at 3 V.

FIG. 62 shows EDS results for the collected magnesium.

FIG. 63 shows EDS results for the scrap residue inside the alloy crucible.

FIG. 64 shows SEM of a piece of remaining alloy taken from its center.

FIG. 65 shows EDS results for the center gray zone of the remaining alloy taken from its center.

FIG. 66 shows EDS results for the center black zone of the remaining alloy taken from its center.

FIG. 67 shows the yttrium profile for a 32 hour overlay using no yttrium fluoride, 2.5 wt % yttrium fluoride, 1.5 wt % yttrium fluoride and 5 wt % yttrium fluoride.

FIG. 68 shows the system diagram for LiF—MgF₂.

FIG. 69 shows an apparatus for determination of metal concentration in the salt.

DETAILED DESCRIPTION

Described herein are methods and apparatuses for primary production and recycling of metals soluble in molten salts, including but not limited to magnesium, calcium, and rare-earths including samarium and dysprosium. In some embodiments, primary production is by oxide electrolysis, and recycling is by molten salt-assisted separation of metallic mixtures, and combining with electrolysis to recycle oxidized metals. The invention further describes methods for manipulating the metal concentration in a molten salt in order to achieve high rate, energy efficiency, and long component lifetime.

DEFINITIONS

As used herein and in the appended claims, the singular forms “a,” “an,” and “the” include plural references unless the content clearly dictates otherwise.

The term “about” is used herein to mean approximately, in the region of, roughly, or around. When the term “about” is used in conjunction with a numerical range, it modifies that range by extending the boundaries above and below the numerical values set forth. The term “about” is used herein to modify a numerical value above and below the stated value by a variance of 20%.

The term “immediate proximity” as used herein means that the gas inlet is disposed close to the cathode such that the gas contacts the cathode and the molten salt.

This background describes magnesium, though the invention applies to other reactive metals soluble in molten salts, including but not limited to calcium and rare-earth metals such as samarium and dysprosium.

According to the most recent U.S. magnesium flow analysis conducted in 1998, the rate of recycling post-consumer magnesium metal and its alloys is at most 37%, which is much lower than the 60-80% rates typical for steel and aluminum (Deborah A. Kramer, Magnesium Recycling in the United States in 1998, U.S. Geological Survey Open-File Report 01-166, 2001; herein incorporated by reference in its entirety). This is due to the low recyclability of many grades of magnesium scrap. High-grade clean magnesium scrap, unused parts or casting gates or runners, is nearly as good as new metal, and need only be re-melted to cast or shape into new products. When magnesium scrap has aluminum or steel inserts but no copper or brass contamination, it may require some dilution or paint removal, but is generally usable after re-melting.

However, any nickel, copper, brass, or silicon contamination, even as alloying elements from aluminum or steel inserts, can severely degrade the properties of magnesium alloys. Because of the difficulties involved in separating magnesium from these metals, one cannot use such scrap to make new magnesium alloy parts, though it can be usable in lower-value applications such as aluminum alloying or steel desulfurization.

Most post-consumer magnesium scrap, including the automotive non-ferrous metal scrap stream from automobile shredders, has such contamination. Furthermore, the similarity of magnesium and aluminum X-ray fluorescence spectra, and presence of aluminum in magnesium alloys and vice versa, make it impractical to sort magnesium parts from aluminum ones, they are sold as a mixture of the two. Today the Al:Mg ratio in such scrap is at least 10:1, making it useful for making aluminum alloys.

There are even lower grades of scrap, such as machining chips and crucible dross/sludge, which are either heavily-oxidized or further contaminated by salts, sand, and other metals. For these grades, methods such as retort re-melting only recover about 30-40% of the contained magnesium, and the product is only usable for the low-value markets.

Although the low-value markets mentioned above (aluminum alloys and steel desulfurization) have been able to absorb low-quality post-consumer magnesium, there have been three options for upgrading its value or separating it from aluminum. First, Hydro Magnesium developed a system for continuous melting with a molten salt in order to remove oxides from the liquid metal (U.S. Pat. No. 5,167,700; and H. E. Friedrich, B. L. Mordike, Magnesium Technology: metallurgy, design data, applications (Springer, 2006), 638; each herein incorporated by reference in its entirety). However, this process cannot separate magnesium from other metals.

Second, distillation takes advantage of the high vapor pressure and low boiling point of magnesium relative to aluminum and other metals. Magnesium's vapor pressure at its melting point is among the highest of all metals, at about 3 torr. However, the chemical affinity between magnesium and aluminum and the formation of surface oxides result in poor separation by distillation alone: one study starting from magnesium alloys such as AZ91, AM60 and AZ31 found it difficult to reduce the magnesium content of the unevaporated Al—Mg residue below about 60% (T. Zhu et al., “Innovative vacuum distillation for magnesium recycling”, Magnesium Technology 2001 (TMS, Warrandale, Pa.), 2001, pp. 55-60; herein incorporated by reference in its entirety). FIG. 1 explains why: at 1150° C. (magnesium boils at 1090° C.), and for values of the activity coefficient between 0.6 and 1 (activity coefficient depends on the composition of the metal liquid), the distillation electrorefining potential E and free energy are only negative for magnesium content above 45-80%, so the distillation reaction is not spontaneous below those concentrations. E is given by:

$\begin{array}{cc}E=\frac{\mathrm{RT}}{2F}\ln \frac{a_{\mathrm{Mg}\left(\mathrm{alloy}\right)}}{a_{\mathrm{Mg}\left(g\right)}}=\frac{\mathrm{RT}}{2F}\ln \frac{\gamma x_{\mathrm{Mg}\left(\mathrm{alloy}\right)}}{P_{\mathrm{Mg}\left(g\right)}/P_{\mathrm{Mg}}^0}& \left(1\right)\end{array}$

• • where P_{Mg (g)} is the partial pressure of a magnesium bubble nucleated at the cathode, above the melt plus the hydrostatic head, and P⁰_Mg is the vapor pressure over pure magnesium at that temperature. This equation refers to Mg_(alloy)→Mg_(g). Also, distillation leaves zinc in the distillate stream due to its higher vapor pressure than magnesium.

Third, if there is relatively little magnesium in an Al—Mg scrap stream, as is the case for beverage can stock or today's automobiles, then chlorine bubbling through the liquid metal can remove it by reaction to form MgCl₂ with various impurities, leaving purer and higher-value aluminum. However, as the magnesium content of vehicles increases, this will require very large amounts of chlorine, and thus large amounts of energy, as well as very large new markets for MgCl₂. This method also wastes the energy content of the magnesium metal: reducing MgCl₂ back to magnesium would require an additional 28 kWh/kg, and the low purity of the chloride might make this impractical.

Because none of these three methods is satisfactory, it is impractical to recycle magnesium in automotive scrap. Without new recycling technology, increasing magnesium use in motor vehicles, estimated by some to reach 15-30 times current use by 2020 (U.S. Automotive Materials Partnership, Magnesium Vision 2020, Technical Report, USCAR, November 2006; herein incorporated by reference in its entirety) increases the likely fraction of post-consumer vehicle material going into landfills, and could reduce the profitability of the entire motor vehicle recycling system. There is also a large missed opportunity to recover magnesium from drosses and other low-grade scrap streams from magnesium processing operations.

Molten salt fluxes such as, for example, CaF₂—MgF₂—MgO, CaCl₂—MgCl₂, CaO, etc., have a finite solubility of metals (such as Ca, Mg, etc.). In solid oxide membrane (SOM) electrolysis, one needs to lower metal solubility (lower electronic conductivity in the molten salt) in order to improve current efficiency and reduce degradation of the SOM. This can be done by depositing or reducing the metal species at the cathode while bubbling controlled amounts of inert gas along the surface of the anode, such that the inert gas lowers the activity of the reduced (deposited) metal and, optionally, assist in separating the metal from the molten salt. This can also be achieved by adding to the molten salt a less stable oxide that can react with and oxidize the soluble metal in the salt. For example, to lower metallic magnesium concentration in a molten salt, one can add iron oxide to the molten salt. The dissolved magnesium in the salt will react with and reduce the iron oxide.

The metal solubility in the molten salt can be advantageously utilized for designing novel refining process that employ gas bubbling to recover metal species from partially oxidized scrap alloys. For example, Mg can be refined and separated from automotive scrap alloys via conversion of the alloy to Mg in the molten salt, and in turn converted to gaseous Mg. The oxide scale and the metal species will dissolve in the molten salt, and the dissolved metal species will be removed in the vapor phase via gas bubbling. Thus, a lower solubility of the metal benefits SOM electrolysis while metal solubility in general can be utilized to design novel refining processes. A larger metal solubility aids the refining and/or separation processes. For both electrolysis and refining, it is advantageous to measure or control the metal solubility in the molten salt. The equilibrium potential between the soluble metal and its ion in the molten salt or that between the soluble metal and its oxide in the molten salt will be indicative of the metal solubility and can be measured using a reference electrode. Thus, in some embodiments, methods and apparatuses for improving efficiency of SOM electrolysis and novel and efficient refining processes for recovering pure metals from scrap alloys through monitoring and manipulating metal solubility in the molten salt are provided. In some embodiments, the alloys comprise magnesium. In some embodiments, the alloys comprise magnesium and aluminum. In some embodiments, the alloys comprise scrap automotive alloys, AZ91, AM50 or AM60.

Although magnesium is discussed herein as an exemplary target metal, the methods and apparatuses herein can also be used to separate other reactive metals soluble in molten salts, including but not limited to calcium and rare-earth metals such as samarium and dysprosium. Rare earth metals comprise metals from the lanthanide series in the chemical periodic table from lanthanum to lutetium as well as scandium and yttrium. Scandium is considered a rare earth element, though it usually occurs in minor amounts. Yttrium is considered a rare earth element because it often occurs with rare earth metals in nature and has similar chemical properties. Thus, in some embodiments, the target metal is magnesium, calcium, or a rare-earth metal. In some embodiments, the rare earth metal is from the lanthanide series. In some embodiments, the rare earth metal is La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, or Lu. In some embodiments, the rare earth metal is samarium or dysprosium. In some embodiments, the target metal is magnesium, calcium, samarium or dysprosium. In some embodiments, the target metal is magnesium, calcium, or samarium. In some embodiments, the target metal is magnesium or calcium. In some embodiments, the target metal is calcium. In some embodiments, the target metal is magnesium.

In some embodiments, the methods and apparatuses described herein use dissolution in a molten salt to efficiently separate magnesium from aluminum and other metals, and recover nearly all of the magnesium from heavily-oxidized scrap. In some embodiments, the methods also use the same principle to recover samarium or dysprosium, or other rare-earth metals or combinations of rare-earth metals from rare-earth magnets, industrial magnet scrap, and other products. The apparatuses and methods described herein can thus meet an important need in the metal recycling ecosystem in general, and recycling of fast-growing automotive magnesium and rare-earth metals in particular.

In some embodiments, the methods and apparatuses provide a mechanism for in situ measurement of the metal remaining in the scrap charge, that is which has not been removed, which one can use to maximize equipment productivity or to create an alloy with a specific composition.

In some embodiments, the methods and apparatuses described herein entail the use of modified SOM processes that enable extraction of metals from metal oxides. Representative embodiments of the SOM apparatus and process may be found, for example, in U.S. Pat. Nos. 5,976,345; 6,299,742; and Mineral Processing and Extractive Metallurgy 117(2):118-122 (June 2008); JOM Journal of the Minerals, Metals and Materials Society 59(5):44-49 (May 2007); Metall. Mater. Trans. 36B:463-473 (2005); Scand. J. Metall. 34(5):293-301 (2005); and International Patent Application Publication Nos. WO 2007/011669 and WO 2010/126597; each of which hereby incorporated by reference in its entirety.

In some embodiments, methods further comprise collecting the metallic species. Methods of collecting metallic species are known (See, e.g., Krishnan et al, Metall. Mater. Trans. 36B:463-473 (2005); Krishnan et al, Scand. J. Metall. 34(5):293-301 (2005); and U.S. Pat. No. 400,664; each herein incorporated by reference in its entirety).

In some embodiments, the methods comprise manipulating metal solubility in a molten salt to improve the rate or efficiency of electrolysis or refining.

For refining, the methods comprise differential dissolution in a molten salt and distillation to separate magnesium, calcium, samarium or dysprosium, or another high-vapor pressure metal soluble in a molten salt, herein called the “target metal”, from other metals, as shown for a preferred embodiment in FIG. 2. In this embodiment, the apparatus (200) consists of a cathode (201), a molten salt bath (202) in electrical contact with a cathode, and a solid oxygen ion conducting membrane (SOM) (203) anode (204) in contact with the molten salt bath. The scrap charge is introduced into the scrap chamber (205), where heat is applied to melt it. The liquid scrap metal (217) contacts the molten salt where the target metal preferentially dissolves into the salt. Several metals, such as magnesium, calcium, rare-earth elements and others, have much higher solubility in molten salts than iron, nickel, aluminum, silicon, copper, zinc, manganese and others, so the target metals preferentially dissolve into the salt, leaving the other metals behind. At the cathode, magnesium cations (Mg²⁺) are reduced to magnesium metal. In the scrap chamber, magnesium alloy scrap liquid is oxidixed from magnesium metal to Mg²⁺, and at the SOM anode O²⁻ is oxidized to ½O₂+2 e⁻. The target metal then evaporates from the molten salt, leaving behind the salt, any dissolved oxides, and metals with lower vapor pressure. The target metal vapor condenses in a condenser.

In some embodiments, argon or other gas bubbled through the salt provides high surface area for one mechanism of rapid mass transfer from the salt into the gas. In some embodiments, a gas can be used that does not react strongly with the target metal such as forming gas, or form a gas in situ, such as boiling zinc or other low-boiling point metal or causing a chemical reaction to create another gas in the salt, in place of or with the inert gas, in order to both agitate the melt and provide high surface area for rapid mass transfer. Here, the higher vapor pressure of the target metal vs. others further refines the target metal product, e.g. magnesium exhibits higher vapor pressure and evaporates faster than calcium and rare-earths. The carrier gas-metal gas mixture travels to a condenser, where the metal vapor condenses to a liquid or solid. Methods for condensing magnesium vapor to liquid are described by Schoukens et al. (U.S. Pat. No. 7,641,711; herein incorporated by reference in its entirety) for high magnesium vapor pressure and Powell et al. (U.S. patent application Ser. No. 13/543,575; herein incorporated by reference in its entirety) for low magnesium or other metal vapor pressure.

When the target metal content of the scrap charge, such as magnesium, is sufficiently low, as measured by electrodes or as inferred by the reaction time, then one can drain the remaining scrap liquid metal for collection of material depleted of the target metal, for example aluminum and other metals after magnesium has been removed.

In some embodiments, electrodes inserted into the liquid scrap metal and molten salt, with the anode in the scrap and cathode preferably at the inert gas introduction locations in the molten salt, can measure the amount of the target metal, illustratively magnesium, remaining in the scrap charge, as follows. A potentiodynamic sweep, illustratively from −0.05 V to 0.15 V and illustratively at a rate of 5 mV/sec as shown in FIG. 3, provides information such as: the open circuit voltage (OCV), voltages at two sudden increases corresponding to electrochemical transitions which can be called E_ER1 and E_ER2, and the slope away from the transitions. The transition voltages E_ER1 and E_ER2 change with time as shown in FIG. 4 and with magnesium composition in the scrap metal. The voltage E_ER2 in particular is related to magnesium mole fraction in the scrap metal X_Mg(alloy) according to equation 1 as displayed in FIG. 1. Accurately estimating X_Mg(alloy) requires either calibration by measuring E_ER2 at known values of X_Mg(alloy), or knowledge of the activity coefficient in the particular alloy system. If one assumes that during the short duration of the potentiodynamic scan the activity of magnesium (concentration) in the alloy does not change, then the difference between the values of E_ER2 and E_ER1 will allow one to estimate the magnesium solubility (activity) in the molten salt (see Equation 2 and FIG. 3). ER1 and ER2 are calculated as E(reduction at anode) minus E(reduction at cathode). This assumes that the nucleating pressure of magnesium bubbles at E_ER2 is approximately 1 atm. and the fact that the standard vapor pressure of magnesium over pure magnesium (P⁰_Mg(g)) is known. If the difference between the E_ER1 and E_ER2 values do not change at a given temperature it implies that the magnesium solubility in the molten salt is the same. If the difference increases it implies that the magnesium solubility in the molten salt is decreasing and vice-versa.

$\begin{array}{cc}E=\frac{\mathrm{RT}}{2F}\ln \frac{a_{\mathrm{Mg}\left(\mathrm{alloy}\right)}}{a_{\mathrm{Mg}\left(g\right)}}=\frac{\mathrm{RT}}{2F}\ln \frac{\gamma x_{\mathrm{Mg}\left(\mathrm{alloy}\right)}}{P_{\mathrm{Mg}\left(g\right)}/P_{\mathrm{Mg}}^0}& \left(1\right)\\ \mathrm{ER}2-\mathrm{ER}1=\frac{\mathrm{RT}}{2F}\ln \frac{a_{\mathrm{Mg}\left(\mathrm{flux}\right)}}{P_{\mathrm{Mg}\left(\mathrm{nuc}\right)}/P_{\mathrm{Mg}}^0}\cong \frac{\mathrm{RT}}{2F}\ln \frac{P_{\mathrm{Mg}}^0•a_{\mathrm{Mg}\left(\mathrm{flux}\right)}}{1}& \left(2\right)\end{array}$

This method presents several benefits over direct distillation by heating the liquid metal mixture. First, distillation hindrance by oxide film formation is not a problem because the molten salt dissolves any oxide present at the metal-salt interface. Second, the molten salt catalyzes magnesium removal from the alloy, facilitating near perfect separation as described below. Third, volatile metals such as zinc and other volatiles in the scrap charge do not contaminate the magnesium product because their solubility in the molten salt is much lower than that of magnesium. And fourth, in situ monitoring of the scrap alloy composition and magnesium content of the molten salt permit precise timing of the duration of distillation, reducing mean cycle time.

In some embodiments, one can add electrodes to reduce any target metal oxide dissolved in the molten salt. This uses solubility of target metal oxide in the salt to increase its production yield from heavily-oxidized scrap, such as magnesium die-casting dross. The anode is preferably separated from the molten salt and inert gas-target metal vapor by a solid oxygen ion-conducting membrane (SOM), as shown in FIG. 2. The SOM can be made illustratively of zirconia with high oxygen ion conductivity, illustratively zirconia stabilized by yttria, calcia, or magnesia, or can take on multiple other embodiments as discussed by Pal and Britten (See, e.g., U.S. Pat. Nos. 5,976,345 and 6,299,742; each herein incorporated by reference in its entirety). This membrane serves two purposes: it separates the anode and cathode products, such that dissolved magnesium in the molten salt does not react with oxygen or combustion products such as CO₂ or H₂O which form at the anode, resulting in high current efficiency, and it prevents fluoride ions from reaching the anode, resulting in zero fluorine or perfluorocarbon emissions and in some embodiments a high purity oxygen by-product. This embodiment includes all of the prior benefits, but can also increase the recycling yield by capturing all of the target metal oxide in the scrap.

In some embodiments, it is advantageous to introduce argon or other inert gas as close as possible to the cathode location where the target metal oxide is reduced. Illustratively, FIG. 5 shows four geometries for accomplishing this. Holes (506) in the cathode (501) (FIG. 5A) distribute the inert gas over its surface, promoting uptake of the target metal at or near where it is reduced, such that less of the metal dissolves. Sharp notches (507) on holes, or at the bottom of a tube, promote formation of small bubbles with high surface area, toward the same end (FIG. 5B). Another embodiment incorporates a mesh surface (508) (FIG. 5C) or porous anode material (509), illustratively a partially-sintered metal such as steel with open porosity (FIG. 5D). Though FIG. 5 shows the cathode in a tube geometry, many other geometries are possible, the important features being electrical connectivity so the entire surface is at or near the same cathodic potential relative to the anodes, and argon introduction directly to that cathodic surface. Thus, in some embodiments, the cathode configuration is selected from a tube comprising holes, a tube comprising notches at the tube end, a tube comprising notched holes, a tube comprising a mesh screen, and/or a porous tube material. In some embodiments, the notches, holes or pores are about 0.5-2 mm in diameter. In some embodiments, the notches, holes or pores are less than about 0.5 mm in diameter.

In some embodiments, it can also be helpful to add target metal oxide to the molten salt, in order to better utilize the electrodes which are in the salt for oxide reduction. One can add the oxide either during melting and refining the scrap, or (semi-)continuously during the metal oxide reduction step.

Optionally, electrodes used for electrolysis can measure the dissolved target metal content of the molten salt as follows. During reduction, the flow rate of anode gases, such as oxygen, or CO or water when using a fuel in the anode, indicates the Faradaic current in the system. Subtracting this from the total current gives the electronic current in the cell. This in turn is proportional to the electronic conductivity of the molten salt, which is related to its dissolved metal concentration.

In some embodiments, the oxide electrolysis can be carried out at lower pressure (below atmospheric pressure), illustratively at 0.001-0.2 atmospheres (about 1-200 mbar). An exemplary configuration is shown in FIG. 6 comprising an anode exhaust (610) and a cathode exhaust (611) are connected in line to vacuum gauges (612), valves (613), liquid nitrogen traps (614) and a vacuum pump (615) which is run at about 59 to about 61 torr. This approach lowers the metal solubility in the salt considerably (Mg dissolved in the salt→Mg vapor) thereby increasing current efficiency and lowering membrane degradation during electrolysis. The lower pressure also reduces the target metal condensation dew point, allowing the process to run at lower temperature, as low as 900-1000° C. for magnesium, or 1100-1300° C. for calcium and samarium.

In some embodiments, one can add the oxide of a more electronegative metal, where the more electronegative metal exhibits low solubility in the molten salt, illustratively nickel oxide, to the molten salt before conducting electrolysis. In some embodiments, the oxide of the more electronegative metal is soluble in the molten salt. In some embodiments, the oxide of the more electronegative metal is insoluble in the molten salt. In some embodiments, the more electronegative metal is insoluble in the molten salt. This causes a reaction between the target metal in the salt and the more electronegative oxide, producing the target metal oxide and the more electronegative metal. This reduces the amount of dissolved metal in the molten salt, thereby increasing the current efficiency and lowering membrane degradation during electrolysis.

In a preferred embodiment, the process steps are as follows:

• • 1. Heating the system to at least about 1000° C. to melt the salt;
  • 2. Filling the scrap melting region with mixed metal scrap containing the target metal, its oxide, dross, etc. and melting the scrap charge;
  • 3. Bubbling an inert gas through the molten salt and optionally the mixed metal scrap to aid in dissolving and distilling out all of the target metal from the charge, while dissolving all target metal oxide from the scrap into the molten salt, measuring the remaining quantity of target metal in the scrap by the electrical signals mentioned above;
  • 4. Tapping out the remaining scrap metal or alloy in the scrap melting region with much less target metal content;
  • 5. Optionally adding target metal oxide to the molten salt, which can also happen before or during steps 3, 4, or 6;
  • 6. Reducing the argon pressure above the molten salt, and/or adding the oxide of a more electronegative metal, such as nickel oxide, thereby reducing the concentration of dissolved target metal in the molten salt. The target metal concentration in the molten salt can optionally be tested by using the electronic current method described above, in order to reach a very low concentration of the target metal in the salt, while using as little as possible excess electronegative metal oxide;
  • 7. Operating the electrolysis cell at the desired pressure to reduce target metal oxide in the molten salt, optionally adding additional target metal oxide to the molten salt for reduction while doing so;
  • 8. Repeating several times from Step 2 in order to recycle multiple scrap charges; and
  • 9. Cooling down, then very slowly introducing air into the system in order to avoid rapid oxidation of any fine target metal dust present.

In some embodiments, one can combine two embodiments by using the same cathode in the molten salt both for measurements of target metal concentration in the scrap charge and molten salt, and also for reduction of target metal oxide present in the salt.

The oxide free energies of formation for cation species in the molten salt must be more negative than the target metal for production, such that the process does not reduce molten salt cations along with the product. For production of rare-earth metals, exemplary cation species include calcium, strontium, barium, lithium, potassium, cesium and ytterbium. Though sodium has lower electronegativity than rare earth elements and some others, its oxide free energy is less negative so sodium oxide present in the molten salt is reduced and evaporates at the cathode before rare earths and magnesium.

The molten salt must exhibit very low vapor pressure and evaporation rate in the process temperature range. Combining thermo-gravimetric analysis (TGA) with differential scanning calorimetry (DSC) or differential thermal analysis (DTA) experiments can efficiently evaluate the molten salt for this criterion. In some embodiments, fluoride salts are preferable over chloride salts.

The molten salt must also have a relatively low melting point. In some embodiments, the temperature range of from about 1000° C. to about 1200° C. provides a balance between good energy efficiency and apparatus stability at lower temperature, and good oxide ion conductivity in stabilized zirconia at higher temperature. The molten salt must not be a solid, and is preferably not partially solid or semi-solid in this temperature range. Differential scanning calorimetry (DSC) or differential thermal analysis (DTA) experiments can efficiently evaluate the flux for this criterion by measuring the temperature at which the flux becomes partially solid (the liquidus temperature) and the temperature at which it becomes completely solid (the solidus or eutectic temperature).

The molten salt should preferably also dissolve the target metal oxide to at least about 2-3 weight % in order to achieve reasonable ionic current densities at the anode and cathode. In some embodiments, dissolution of the target metal oxide at least about 2-3% by weight achieves sufficient ionic current densitites. For some metals, the target metal oxide need only be dissolved as little as 0.5% by weight to render the process economically viable. In some embodiments, dissolution of the target metal oxide at 0.5% by weight achieves sufficient ionic current densitites. DSC or DTA experiments at various compositions can efficiently characterize oxide solubility.

The molten salt must not appreciably dissolve or corrode the solid electrolyte, illustratively zirconia. In some embodiments, evaluation of stability can be made via immersion of zirconia in the molten salt at the process temperature for several hours, for example tens to hundreds or thousands of hours, followed by sectioning and characterization of the zirconia, to determine the minimum corrosion rate of zirconia in the molten salt, without any applied current.

Species in the molten salt must have high mobility, i.e., high ionic conductivity and low viscosity, in order to support high current density without significant transport limitation. A high viscosity molten salt would inhibit mass transfer to the zirconia electrolyte and the cathode; at the zirconia, oxygen ions would be depleted in the boundary layer, thereby reducing the current, and at the cathode the target metal ions would be depleted in the boundary layer, thereby leading to reduction and co-deposition of molten salt cations. In some embodiments, molten salts without silica or alumina and with high fluoride/oxide ratio have sufficient mobility.

The molten salt must also have low electronic conductivity so as not to function as an extended cathode and reduce or corrode the zirconia. In part, conductivity depends on solubility of the reduced metal in the molten salt. The skilled artisan will be able to determine molten salt conductivity during electrolysis experiments.

For recycling and/or refining processes, the target metal must be highly soluble in the molten salt, that is, with higher solubility and/or higher rate of dissolution than the other metals present in the scrap alloy, and/or with higher evaporation rate than the other metals present. For example, when refining magnesium from a mixture of magnesium, aluminum and calcium, the aluminum and calcium both preferentially dissolve into the molten salt leaving the aluminum in the scrap liquid, and the magnesium preferentially evaporates leaving the calcium in the molten salt. In some embodiments such as, for example, wherein an electronegative metal oxide is added, the electronegative metal produced must be lowly soluble in the molten salt such that the metal precipitates out of solution. For example, when nickel oxide is added to remove magnesium from the salt by generation of nickel and magnesium oxide, the nickel should precipitate out of solution.

In some embodiments, one can use the electrolysis electrodes to reduce target metal oxide fed into the system, without any scrap metal. This is useful for primary production of the target metal from oxide scrap, ore, or other sources. These embodiments differ from that of Pal and Britten (See, e.g., U.S. Pat. Nos. 5,976,345 and 6,299,742; each herein incorporated by reference in its entirety) in the use of one or more of the methods described above to prevent the dissolution of metal and/or to remove it from the molten salt after it has dissolved; those methods are:

• • 1. Introducing the argon or other inert or carrier gas close to the cathode location where the target metal ions are reduced to the metal, such that it is first produced as a dilute vapor mixed with the inert gas;
  • 2. Bubbling argon or other inert or carrier gas in order to remove the dissolved target metal from the molten salt;
  • 3. Reducing the pressure in the molten salt in order to reduce the solubility of the target metal, thereby either removing it from the molten salt into the gas phase or preventing its dissolution; and
  • 4. Adding the oxide of a more electronegative metal than the target metal in the molten salt, such as nickel oxide, in order to oxidize any target metal dissolved in the molten salt.

Unlike the recycling system which operates by a semi-batch process as outlined in the steps above, the oxide reduction embodiments can operate continuously with repeated or continuous additions of target metal oxide to the molten salt as the reduced target metal comes out of the salt in the gas phase. Methods 1 and/or 3 can continuously prevent newly-produced metal from dissolving in the molten salt, and methods 2, 3 and/or 4 can periodically remove dissolved metal from the molten salt as needed. Each of the methods can be performed independently; however combinations of two or more of the methods are also within the scope of the invention.

In some embodiments, the inert gas or carrier gas creates high surface area. In some embodiments, the high surface area promotes mass transfer of the target metal into the gas.

In some embodiments, reduced pressure prevents dissolution of the metal into the salt. For magnesium as an exemplary metal, reduced pressure removes the magnesium from the molten salt surrounding the cathode. As a result, the layer of molten salt surrounding the cathode is less electronically conductive. In some embodiments, the layer of molten salt surrounding the cathode becomes non-electronically conductive. In some embodiments, reduced dissolution of the metal in the salt results in increased current efficiency in the salt.

In some embodiments, the molten salt is at least about 90% liquid. In some embodiments, the molten salt is at least about 92% liquid. In some embodiments, the molten salt is at least about 95% liquid. In some embodiments, the molten salt is at least about 98% liquid. In some embodiments, the molten salt is at least about 99% liquid.

In some embodiments, a physical barrier can be used to separate target metal dissolution region from the evaporation region. Exemplary embodiments for such purposes include two chambers with salt circulating between the dissolution region and the evaporation region. In some exemplary embodiments, the regions are held at two different temperatures. Not only can this reduce the rate of unwanted target metal evaporation from the scrap/alloy target metal dissolution region, but, in some embodiments, further allows operating at different pressures in the two regions.

In some embodiments, the barrier comprises a plurality of gas bubbles to create a region between the anode and the cathode wherein the metal concentration and electronic conductivity are low.

Several advantages of the refining process relative to distillation are seen, for example, the oxide cannot block the process because it is dissolved in the molten salt, whereas an oxide layer can block distillation. The refining process also has improved selectivity over distillation due to differential dissolution in the salt. In many cases, the refining process also occurs at a high rate, whereas distillation does not.

Configurations of components for efficient high-throughput metal primary production and recycling are also within the scope of the invention. FIG. 7 shows an exemplary design concept for scrap metal with density lower than that of the molten salt. Scrap metal is introduced into the scrap chamber (705) through the opening (716) shown in the top left visualization of a rear view into the scrap chamber (FIG. 7A). The scrap melts, and the scrap alloy liquid (717) floats on the molten salt. The scrap is held in floating “piers” (718) resting on the molten salt by a metal gate (719), which creates “inlets” of exposed molten salt between the piers. A carrier gas, illustratively argon, is blown in through steel tubes (720), and zirconia zubes with anodes (721) descend into the salt on the side away from the scrap chamber. FIGS. 7B-D are cross-sectional views shown without the chamber. FIG. 7B shows a cross-sectional view with the scrap alloy liquid (717) floating on the molten salt (702), the metal gate (719) configured angularly toward the molten salt, steel tubes (720), and zirconia tubes with anodes (721). Translucent cones (722) visualize the plumes of argon bubbles rising through the molten salt. FIG. 7C shows another cross-sectional view with the scrap alloy liquid (717) floating on the molten salt (702), the metal gate (719) configured angularly toward the molten salt, steel tubes (720), and zirconia tubes with anodes (721). FIG. 7D shows another cross-sectional view with the steel tubes (720), zirconia tubes with anodes (721) and translucent cones visualizing plumes of argon (722).

The goal of this design is to facilitate transport of the target metal, illustratively magnesium, from the liquid alloy through the molten salt to the carrier gas bubblers. The alternating liquid metal piers and exposed salt inlets give a large “coastline” with relatively short distance for the target metal to travel. The proximity between argon bubble plumes and liquid scrap metal also enable the salt to perform some stirring along that metal-salt coastline, further enhancing transport of target metal into and through the salt.

The system can have multiple sets of floating liquid alloy “piers” on the molten salt. In the case of dense metal scrap, such as scrap comprising rare-earth metals, the liquid metal will be below the molten salt.

In some embodiments in which the liquid alloy floats on the molten salt, the system addresses any liquid metal leaks under the metal gate. Liquid which leaks would float on the molten salt, and could either reduce evaporation rate, or possibly short-circuit the electrolysis operation. In such embodiments, a simple weir system manages this contingency by creating a pathway for the floating metal to spill out of the molten salt, removing it from the system such that it does not prevent evaporation or lead to short circuiting.

The most energy-efficient distillers use “compression distillation”, in which a pump between the evaporator and condenser maintains a pressure difference between them. The condenser is at a higher pressure, thus has a higher boiling point, and the temperature at the condensing surface is higher than in the evaporator. One can put the condenser and evaporator in intimate contact to achieve high heat molten salt from the exothermic condensation surface to the endothermic evaporation surface, and the system requires little or no energy input beyond that which drives the pump. This energy-efficient distiller design and operation are described in, for example, U.S. Pat. Nos. 2,899,366 and 4,082,616; each herein incorporated by reference in its entirety.

Application of compression distillation to this system will require a high-temperature pump, and a condenser in close contact with—or underneath—the molten salt crucible. In preferred embodiments, target metal condensation would heat the metal-bearing salt as close as possible to the argon or other carrier gas tubes for maximum heat transfer to evaporating the target metal, promoting high distillation rate and high energy efficiency.

In some embodiments, the carrier gas and target metal flow through, and condense in, tubes or other conduits in or below the molten salt, and holes, tuyeres, pores, or similar openings between the conduit and molten salt allow some of the carrier gas to escape into the molten salt, providing direct bubbling. This would require that the holes be small enough to maintain a pressure difference between the condenser tubes/conduits and the molten salt.

Another illustrative embodiment is shown in FIG. 8, wherein the target metal is more dense than the molten salt. Exemplary target metals may include samarium, neodymium and other lanthanides. In this embodiment, in addition to the SOM (803) the cathode (801), the scrap chamber (805) and the oxide feed (817), the target metal product (825) is held between a side of the apparatus (800) and a barrier in the form of a lower portion (840) and an upper portion (841). Optionally, an interface (842) separates the molten salt (802) from the metal (825). Other exemplary apparatuses for target metals having density higher than the molten salt are described in U.S. Patent Publication No. 2011/0079517; incorporated herein by reference in its entirety.

An illustrative industrial embodiment is shown in FIG. 9, showing a magnesium oxide feed introduced into a crucible (900) that contains a molten salt electrolyte (902) via a plurality of feed tubes (905), a plurality of SOM anodes (903), a plurality of cathodes/argon feed tubes (901). Ar/Mg bubbles (923) are generated and Ar/Mg migrates to a condenser (924) where condensation of Mg occurs to provide liquid Mg (925). A tap (926) enables removal of the condensed Mg. An argon recycling pump (927) collects argon from the condenser and recirculates the argon back into the crucible. Oxygen gas is removed from the system, which is connected to a gas/power/raw material manifold (928).

Another illustrative industrial embodiment is shown in FIG. 10, showing scrap material feed from Al and Mg, optionally Mg dross (1005), a plurality of stabilized zirconia tubes/anodes (1003), a plurality of cathodes/argon feed tubes (1001) and the molten salt electrolyte (1002). Aluminum is separated from the anodes via a barrier (1029) and a first tap (1030) enables removal of aluminum. Ar/Mg bubbles (1023) are generated and Ar/Mg migrates to a condenser (1024) where condensation of Mg occurs to provide liquid Mg (1025). A second tap (1026) enables removal of the condensed Mg. An argon recycling pump (1027) collects argon from the condenser and recirculates the argon back into the crucible (1000). Oxygen gas is removed from the system, which is connected to a gas/power/raw material manifold (1028).

Another illustrative embodiment is shown in FIG. 11, showing an assembly for reduction of volatile metals such as magnesium. The perforated cathode tube on the outside is solid along its length (1101) until it extends below the level of the molten salt. In the annular region between the inside diameter of the cathode and the outside diameter of the anode (1104), a downward gas purge (1136) (illustratively argon gas) is used to keep magnesium vapor produced by electrolysis from rising into the annulus, and forces the magnesium to exit through the perforations (1106) in the lower section of the cathode. A boron nitride spacer (1137) at or above the salt level (1138) further impedes vapors from rising along the zirconia, and aids in creating a region of slight positive pressure to protect the zirconia outside of the salt. A nozzle located at the bottom of the crucible (1139), and centered on the cathode, aids in locating the cathode-anode assembly. The nozzle provides an argon bubble stream that contributes to mixing of the salt and the magnesium oxide, and provides good oxide mass transfer to the zirconia tube. The argon bubble stream also absorbs magnesium metal dissolved in the molten salt, and carries the magnesium vapor outward through the perforations in the cathode.

The entire cathode-anode assembly can be replaced and inserted as a unit, thus facilitating replacement of cathodes and/or anodes. In the event of a failure of the zirconia tube, the closed end of the cathode allows it to be used as a basket, which prevents the zirconia pieces from escaping and eventually building up in the salt. Removing the cathode will thus remove any large zirconia shards in an efficient manner.

In some embodiments, this assembly also provides a very short anode-cathode distance. The short anode-cathode distance provides very low resistance due to the molten salt. In some embodiments, this assembly also results in low cathode current density (illustratively about ¼ to about ½ of the maximum current density in the zirconia tube) due to the large area. Current density is substantially uniform along the cathode due to the concentricity with the anode.

In another embodiment, multiple zirconia tubes with anodes can be inserted into a single large perforated cathode tube. Illustratively, a plurality of zirconia/anode tubes may be inserted into a single cathode tube. In some embodiments, two, three, four, or more zirconia/anode tubes are inserted. In some embodiments, more than four zirconia/anode tubes are inserted. In these embodiments, the cathode current density is not necessarily uniform, but the high resistivity of zirconia (generally 5-30 times higher than that of the molten salt) results in substantially uniform current density in the zirconia tubes.

It will be recognized that one or more features of any embodiments disclosed herein may be combined and/or rearranged within the scope of the invention to produce further embodiments that are also within the scope of the invention.

Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. Such equivalents are also intended to be within the scope of the present invention.

The following examples illustrate the present invention, and are set forth to aid in the understanding of the invention, and should not be construed to limit in any way the scope of the invention as defined in the claims which follow thereafter.

EXAMPLES

The magnesium recycling experiment whose potentiodynamic sweep and transition voltages are shown in FIGS. 3 and 4 respectively began with a 21.0 g scrap charge of 46 wt % Mg, 45% Al and 9% Fe, and produced 7.4 g of 99.6 wt % pure Mg, with <0.1% Mg remaining in the scrap charge, a nearly perfect separation. Of the 7.4 g magnesium product, approximately 0.4 g was produced by electrolysis of dissolved MgO in the molten salt. A second experiment began with a pre-melted scrap charge of 10.7 g AZ91D magnesium alloy, 9.2 g aluminum, and 2.5 g iron powder, and produced 6.0 g of 99.35 wt % pure Mg with 0.46 wt % zinc, which was just half of the zinc contained in the starting AZ91D magnesium alloy. Those experiments are described in detail in Xiaofei Guan, Peter Zink, Uday Pal and Adam Powell, “Magnesium Recycling of Partially Oxidized, Mixed Magnesium-Aluminum Scrap Through Combined Refining and Solid Oxide Membrane Electrolysis Processes,” in Electrochemical Society Transactions vol. 41, no. 31, pp. 91-101 (2012); and Xiaofei Guan, Peter Zink, Uday Pal, “Magnesium Recycling of Partially Oxidized, Mixed Magnesium-Aluminum Scrap Through Combined Refining and Solid Oxide Membrane (SOM) Electrolysis Processes,” in Suveen Mathaudhu, Wim Sillekens, Norbert Hort, Neale Neelameggham, eds., Magnesium Technology 2012 (TMS, Warrendale, Pa.), 2012, pp. 531-536; each herein incorporated by reference in its entirety.

This magnesium separation performance is much better than that of pure distillation, whose scrap residue magnesium content does not fall below 60% (T. Zhu et al., “Innovative vacuum distillation for magnesium recycling”, Magnesium Technology 2001 (TMS, Warrandale, Pa.), 2001, pp. 55-60; each herein incorporated by reference in its entirety). The ability to recover both metallic and oxidized magnesium from die-casting dross material is also a major improvement over the 30-40% magnesium recovery from existing retort melting practice.

Laboratory magnesium oxide electrolysis experiments with zirconia oxygen-ion-conducting membrane between the molten salt and anode demonstrated two effects of argon bubbling on magnesium metal dissolution in the molten salt. In the first experiment, after a long electrolysis run with the crucible as the cathode, a potentiodynamic scan showed very high electronic current (Eric Gratz, Soobhankar Pati, Jarrod Milshtein, Adam Powell and Uday Pal, “Efficiency and Stability of Solid Oxide Membrane Electrolyzers for Magnesium Production,” in Wim Sillekens, ed., Magnesium Technology 2011 (TMS, Warrandale, Pa.), 2011, pp. 39-42; herein incorporated by reference in its entirety). Then the apparatus sat with no electrolysis, but with continued argon bubbling, followed by another potentiodynamic scan which showed less electronic current, as shown in FIG. 12. Without being bound by theory, this indicated that the argon bubbling removed some of the dissolved magnesium from the molten salt by mass transfer into the bubbles and then the gas phase.

Second, an experiment used a vertical tube with notches in the bottom, similar to that of FIG. 5b, as both the cathode and argon bubble source (Eric Gratz, Soobhankar Pati, Jarrod Milshtein, Adam Powell and Uday Pal, “Effect of Electronic Current on the Solid Oxide Membrane (SOM) Process for Magnesium Production,” in Georges Houlachi et al., eds., Electrometallurgy 2012 (TMS, Warrandale, Pa.), 2012, pp. 111-118 (ISBN: 978-1-1182-9117-7); herein incorporated by reference in its entirety). In this experiment, potentiodynamic scans showed near zero electronic current after one electrolysis run for four hours, then slightly more after a second run for one hour. After subsequently using the crucible as the cathode for two hours, the current rose at a lower voltage, as shown in FIG. 13. The tube cathode had argon bubbles much closer to the magnesium reduction site, but the crucible did not. Without being bound by theory, this result likely indicates that the presence of argon at the cathode reduced the dissolution of metallic magnesium in the molten salt, by providing a low-vapor pressure and low-activity sink for the newly-reduced metal; whereas at the crucible cathode without argon, magnesium metal formed at 1 atm pressure and dissolved into the salt more readily.

Laboratory magnesium oxide electrolysis experiments with a zirconia oxygen-ion-conducting membrane were carried out at a pressure of 0.08 atmosphere and compared with experiments conducted at 1 atmosphere pressure. The potentiodynamic scan after 7.5 hours of electrolysis at 0.08 atm (@ 1.7 V) show no electronic leakage current, whereas the potentiodynamic scan after 3 hours of electrolysis at similar potentials and 1 atm pressure show a clear (1.2 Amp) electronic leakage current; see FIG. 14. Based on the total current passed and the Mg collected in the condenser over a span of 12 hours of electrolysis, the efficiency at 0.08 atm was estimated to be greater than 75% compared to 30% at 1 atm pressure. Furthermore, no measurable magnesium solubility could be detected in the salt after the electrolysis experiment at 0.08 atm pressure whereas at 1 atm pressure the magnesium solubility in the molten salt after electrolysis was found to be between 0.02-0.05 wt %; Mg solubility was estimated by measuring hydrogen evolution when a given amount of the molten salt is reacted with hydrochloric acid. These experiments indicate that oxide electrolysis at lower pressure results in higher current efficiency by decreasing the soluble magnesium which in turn will also decrease membrane (SOM) degradation during electrolysis as the applied potential will not reduce the membrane.

Performance with other metals also exhibiting high vapor pressure and solubility in molten salt, such as calcium and some rare-earth metals, should be comparable to these magnesium results.

Example 1

The experimental set up is shown in FIG. 15. A stainless steel crucible cathode/reference (1500) was fitted with a YSZ tube (1503), an Ar stirring tube/cathode (1501) was isolated from the crucible with an alumina tube (1531). A molten salt of magnesium fluoride-calcium fluoride (1502) was added to the crucible. The anode (1504) was tin with a molybdenum current collector (1532). Steel tubes to vent Mg and Ar vapor to the condenser are not shown. Hydrogen was bubbled through room temperature water before being bubbled in the tin. The cell temperature was 1190° C.

Initial potential dynamic scan (PDS) −5 mV/s (FIG. 16) and impedance before electrolysis 1 (FIG. 17) showed no initial leakage current. FIG. 17 shows 1/Z″.

Electrolysis 1 was run at 2 V for 3 hours with isolated argon stirring tube as the cathode. No evidence of mass transfer limitations were observed (FIG. 18). After electrolysis 1, PDS plots of cathode potential versus current (FIG. 19) and anode potential versus current (FIG. 20) were performed. No current was observed until 1 V was applied.

Electrolysis 2 was run at 3 V for 1 hour with isolated argon stirring tube as the cathode. Again, no mass transfer limitations were observed (FIG. 21). After electrolysis 2, PDS plots of potential versus current showed very small to no leakage current (FIG. 22). A potential dynamic scan with the crucible was as the cathode was then used and no leakage current was recorded. High current at larger cathodes suggests cathodic mass transfer limitations.

Electrolysis 3 was run at 2.75 V for 2 hours using the steel crucible as the cathode (FIG. 23). After electrolysis 3, PDS plots of potential versus current after using the steel crucible as the cathode showed noticeable leakage current (FIG. 24). Cathode potential appeared to level off at 0.7 V (FIG. 25), whereas anode potential continued to increase (FIG. 26).

Using an isolated stirring tube cathode allows for the cathodic and anodic overpotentials to be measured. It appears possible to reduce the leakage current by using an isolated argon stirring tube with a small area. Without being bound by theory, a small diameter argon stirring tube lowers the residence time of Mg gas in the molten salt by lowering the partial pressure of Mg in the molten salt and therefore lowering Mg solubility. This is evidenced by the low leakage current after electrolysis with the argon stirring tube cathode when compared to the large area cathode. Further experiments include electron blocking molten salt at the anode (using molten salt between anode carbon current collector and YSZ, molten salt height on the anode side higher than the molten salt on the cathode side, anode molten salt will not theroretically become electronically conductive therefore breaking the circuit and lowering electronic conductivity), isolated cathode stirring tube with carbon anode current collector, measuring molten salt conductivity in situ during or after electrolysis.

Example 2

A method for recycling Mg from Mg scrap by combining a refining process and an SOM electrolysis process is shown in FIG. 27. In the refining process, the magnesium and its oxide were dissolved from scrap (2717) into a molten salt (2702), followed by vapor phase removal of dissolved magnesium. In the SOM electrolysis process, when applied potential reaches the dissociation potential of MgO, oxygen ions are pumped out of the molten salt through a yttria-stabilized zirconia (YSZ) SOM (2703) toward a carbon rod current collector (2732) and are oxidized by the carbon to produce carbon monoxide gas and electrons. Meanwhile, magnesium ions at the cathode (2701) are reduced to magnesium metal vapor. At the scrap anode (2755), for refining, the magenesium alloy is converted to magnesium molten salt; for dissolving magnesium oxide, the magnesium oxide alloy is dissolved to form magnesium cations and oxide anions; and for PDS, magnesium metal is oxidized to magnesium cations. The experimental design is shown in FIG. 28. The setup consists of an upper reaction chamber (2800), heated to 1175° C. and a lower condensing chamber (2824) with a temperature gradient of 1100-200° C. The setup was fabricated using grade 304 stainless steel (SS-304) and heated in an argon atmosphere. The starting Mg scrap is slightly oxidized 50.5 wt % Mg—Al alloy, prepared by melting a 9.624 g piece of magnesium (Mg>99.8%) and a 9.4 g piece of aluminum (Al>97.9%) together inside a small SS-304 crucible. To form the alloy, the mixture was stirred with a SS-304 rod for 15 minutes at 800° C. in an argon atmosphere, held at the same temperature without stirring for about 15 minutes and quenched.

Since there is a finite solubility of iron in Mg—Al alloy, 2 g of iron powder was added on the top of the Mg—Al alloy to decrease dissolution of iron from the crucible during the run. Also, the iron powder addition increases the density of the alloy (alloy sinks to the molten salt bottom in the alloy crucible) and lowers the vapor pressure of magnesium. A powdered molten salt (MgF₂—CaF₂—10 wt. % MgO—2 wt. % YF₃) containing a eutectic mixture of 45 wt. % MgF₂—55 wt. % CaF₂ with a melting point of 974° C. [P. Chartrand and A. D. Pelton, “Thermodynamic evaluation and optimization of the LiF—NaF—KF—MgF₂—CaF₂ system using the modified quasi-chemical model,” Metallurgical and Materials Transactions A, 32 (6) (2001) 1385-1396; herein incorporated by reference in its entirety] was used as the electrolyte and was packed both inside and outside the inverted crucible shown in FIG. 28. In order to prevent the distillation of magnesium from the Mg—Al alloy during heating to 974° C., a layer of molten salt is melted on the surface of the crucible to hermetically seal the opening of the alloy crucible before heating (2833). Thus, the alloy is trapped inside the alloy crucibles, and the direct distillation of magnesium is avoided. A result of melting the molten salt at the top of alloy crucible (2805) is that some of magnesium in the alloy becomes oxidized. This magnesium oxide is later reduced with the SOM electrolysis process. The crucible was fitted with an electrorefiner (anode 1), a graphite rod (anode 2) (2832) in a SOM (2803), the Mg—Al alloy, venting tube (2834), bubbling tube (2801), 2 g iron powder, and 680 g of molten salt.

During the refining process, the alloy crucible and inverted crucible (2835) served as the anode, and the reaction chamber wall and bubbling tube served as the cathode. The SOM tube was held above the molten salt. An alumina spacer was used to insulate the rod connecting the inverted crucible and the reaction chamber. Potentiodynamic scans were performed to determine the electrorefining potential for magnesium, as the refining of magnesium proceeded.

During the SOM electrolysis process, an yttria stabilized zirconia (YSZ) tube was used for recycling magnesium from magnesium oxide. The stainless steel wall of the reaction chamber still served as the cathode, but silver inside the YSZ tube served as the anode, and a carbon rod acted as the anodic current collector. When the applied electric potential between the cathode and anode exceeds the dissociation potential of magnesium oxide, magnesium vapor is produced at the cathode, and carbon reacts with oxygen to generate carbon monoxide [A. Krishnan, U. B. Pal and X. G. Lu, “Solid Oxide Membrane Process for Magnesium Production Directly from Magnesium Oxide,” Metallurgical and Materials Transaction B, 36 (4) (2005), 463-473; and U. B. Pal, A. C. Powell, “The Use of Solid-oxide-membrane Technology for Electrometallurgy,” Journal of the Minerals, Metals and Materials Society, 59 (5) (2007), 44-49; each herein incorporated by reference in its entirety]. The overall cell reaction is given as:

MgO+C=Mg(g)+CO(g)  (3)

During both the refining and electrolysis processes, the reaction chamber is continually purged with 95% Ar—H₂ at 15 cc/min through a bubbling tube, and at 30 cc/min through the two annuli at the top of the reaction chamber. This is done to lower the partial pressure of magnesium vapor over the molten salt and to carry the magnesium vapor to the condensing chamber. The inlet of the venting tube is well above the molten salt surface, to prevent any molten molten salt from entering the condenser.

Electrochemical measurements were performed. A Solartron SI 1280B potentiostat was used for potentiodynamic scans and impedance spectroscopy during the refining process; an Agilent Technologies N5743A power supply was used for potentiodynamic scans and electrolysis. Temperature steps at the center of the electrorefiner chamber are shown in FIG. 29. Time of day versus temperature is plotted. The temperature reached maximum after ˜5 h (4:00 PM) and was held at 1175° C.

1. Refining Process

A PDS scan was run at 13 minutes after the maximum operating temperature was reached (4:13 PM) at electrorefining (ER) potential of 0.018 V, Faradic current of 0.629 A and open circuit voltage (OCV) of −0.0003 V (FIG. 30). Impedance spectroscopy (IS) was measured at 4:20 PM, indicating a cell ohmic resistance of 0.066 ohms FIG. 31. A PDS scan at 4:38 PM at ER potential of 0.015 V, OCV of −0.0041 V, and Faradic current 1^st jump at 0.0425 A and 2^nd jump at 0.0450 A FIG. 32.

A PDS scan at 4:58 PM at ER potential of −0.0018 V, OCV of −0.0302 V, and Faradic current 1^st jump at 0.058 A and 2^nd jump at 0.0674 A FIG. 33.

A PDS was added at 5:08 PM, with ER potential of −0.0027 V, OCV of −0.0312 V, and Faradic current 1^st jump at 0.0539 A and 2^nd jump at 0.0663 A (FIG. 34). A potentiostatic scan (PSS) was done at 5:14 PM, with applied potential of 0.0648 V, OCV of −0.0352 V, current typical value of 0.38 A and Faradic current 0.1202 A (FIG. 35). Magnesium electrorefined was 0.0045 g.

A PDS was added at 5:19 PM, with ER potential of 0.0041 V, OCV of −0.0242 V, and Faradic current (second jump) 0.0592 A (FIG. 36).

A PDS was added at 5:23 PM, with ER potential of −0.0158 V, OCV of −0.0465 V, and Faradic current 1^st jump at 0.0529 A and 2^nd jump at 0.0625 A (FIG. 37).

A potentiostatic scan (PSS) was done at 5:25 PM, with applied potential of 0.057 V, OCV of −0.043 V, current typical value of 0.42 A and Faradic current 0.1154 A (FIG. 38). Magnesium electrorefined was 0.0087 g.

A PDS was added at 5:36 PM, with ER potential of −0.0018 V, OCV of −0.031 V, and Faradic current (second jump) 0.0659 A (first jump N/A) (FIG. 39).

A PDS was added at 5:53 PM, with ER potential of −0.0047 V, OCV of −0.0504 V, and Faradic current 1^st jump at 0.0622 A and 2^nd jump at 0.0838 A (FIG. 40).

A potentiostatic scan (PSS) was done at 5:55 PM, with applied potential of 0.106 V, OCV of −0.0439 V, current typical value of 0.46 A and Faradic current 0.146 A (FIG. 41). Magnesium electrorefined was 0.0328 g.

A PDS was added at 6:28 PM, with ER potential of 0.017 V, OCV of −0.0269 V, and Faradic current 1^st jump at 0.0642 A and 2^nd jump at 0.0847 A (FIG. 42).

A PDS was added at 6:52 PM, with ER potential of 0.045 V, OCV of 0.0124 V, and Faradic current 1^st jump at 0.0606 A and 2^nd jump at 0.0821 A (FIG. 43).

A potentiostatic scan (PSS) was done at 7:07 PM, with applied potential of 0.146 V, OCV of −0.004 V, current typical value of 0.42 A and Faradic current 0.1427 A (FIG. 44). Magnesium electrorefined was 0.032 g.

A PDS was added at 7:38 PM, with ER potential of 0.072 V, OCV of 0.0299 V, and Faradic current 1^st jump at 0.0592 A and 2^nd jump at 0.0616 A (FIG. 45).

A potentiostatic scan (PSS) was done at 7:48 PM, with applied potential of 0.221 V, OCV of 0.021 V, current typical value of 0.45 A and Faradic current 0.1208 A (FIG. 46). Magnesium electrorefined was 0.0271 g.

A PDS was added at 8:19 PM, with ER potential of 0.103 V, OCV of 0.0394 V, and Faradic current 1^st jump at 0.0462 A and 2^nd jump at 0.0597 A (FIG. 47).

Impedance spectroscopy (IS) was measured at 8:27 PM, indicating a cell ohmic resistance of 0.09 ohms (FIG. 48).

A potentiostatic scan (PSS) was done at 8:34 PM, with applied potential of 0.2656 V, OCV of 0.0154 V, current typical value of 0.51 A and Faradic current 0.1059 A (FIG. 49). Magnesium electrorefined was 0.008 g.

A PDS was added at 8:45 PM, with ER potential of 0.112 V, OCV of 0.0619 V, and Faradic current 1^st jump at 0.0588 A and 2^nd jump at 0.0712 A (FIG. 50).

A potentiostatic scan (PSS) was done at 8:47 PM, with applied potential of 0.288 V, OCV of 0.038 V, current typical value of 0.5 A and Faradic current 0.1309 A (FIG. 51). Magnesium electrorefined was 0.0096 g.

A PDS was added at 8:58 PM, with ER potential of 0.132 V, OCV of 0.0824 V, and Faradic current 1^st jump at 0.0593 A and 2^nd jump at 0.0823 A (FIG. 52).

OCV (circles) and ER potential (triangle) vs. timestamp is shown in FIG. 4.

In conclusion, kinetics of magnesium refining was monitored using PDS. The electrorefining potential increases as the amount of magnesium in the alloy decreases, and the amount of magnesium refined by electrorefining is 0.1227 g during the 125 minute experiment due to small faradic current. Mg solubility in molten salt is measured to be 0.03 wt % using the hydrogen evolution technique. Once the reaction chamber was at the desired temperature, potentiodynamic scans were performed as shown in FIG. 3. The initial melted molten salt is entirely ionic and acts as an electronic insulator between the anode and the cathode. The current increases linearly as the applied potential increases, except at the two current jumps corresponding to magnesium dissolution and magnesium vapor formation, as can be seen in each potentiodynamic scan. The current-voltage curve is shifting in a positive direction over time from PDS1 to PDS5 due to decreasing magnesium concentration in the alloy as explained later in the paper. The ohmic resistance of the system was measured to be as low as 0.066Ω and 0.09Ω using impedance spectroscopy shown in FIG. 53. One potential reason for the low resistance or the electronic conductivity is the dissolution of metallic magnesium in the molten salt. It has been shown that magnesium metal has some solubility in chloride based ionic salts [J. Wypartowicz, T. Ostvold and H. Oye, “The Solubility of Magnesium Metal and the Recombination Reaction in the Industrial Magnesium Electrolysis,” Electrochimica ACTA, 25 (1980), 151-156; herein incorporated by reference in its entirety]. The magnesium solubility inside the molten salt after the experiment was found to be 0.03 wt % as measured with a manometer. Dilute acid was added to powdered salt in a closed container and the volume of gas produced due to the hydrogen evolution was measured. The salt was powdered in a glove box to avoid the oxidation of magnesium in the salt. The magnesium solubility inside a molten salt of similar composition (55% MgF₂—45% CaF₂—10% MgO) was reported to be 0.02-0.05 wt. % [E. Gratz et al., “Efficiency and Stability of Solid Oxide Membrane Electrolyzers for Magnesium Production,” Magnesium Technology 2011 (Wiley-TMS, 2011), 39-42; herein incorporated by reference in its entirety], consistent with the experimental results.

For each potentiodynamic scan, there are two current jumps which correspond to two different electrochemical reactions. The current jump at the first potential, close to the open circuit voltage is due to the following electrochemical reaction:

Anode: Mg(alloy)→Mg2++2e−  (4)

Cathode: Mg2++2e−→Mg(molten salt)  (5)

• • where the overall reaction is

Mg(alloy)→Mg(molten salt)  (6)

This reaction is different from direct dissolution of magnesium from Mg—Al alloy into molten salt. Theoretically, the first electrorefining potential corresponding to the first current jump should be equal to the open circuit voltage, and its expression is given by

$\begin{array}{cc}{E}_{\mathrm{ER}1}=E_{\mathrm{OCV}}=E_{\mathrm{anode}}-E_{\mathrm{cathode}}=\frac{\mathrm{RT}}{2F}\ln \frac{a_{\mathrm{Mg}\left(\mathrm{alloy}\right)}}{a_{\mathrm{Mg}\left(\mathrm{flux}\right)}}=\frac{\mathrm{RT}}{2F}\ln \frac{\gamma x_{\mathrm{Mg}\left(\mathrm{alloy}\right)}}{P_{\mathrm{Mg}\left(\mathrm{flux}\right)}/P_{\mathrm{Mg}}^0}=\frac{\mathrm{RT}}{2F}\ln \frac{\gamma x_{\mathrm{Mg}\left(\mathrm{alloy}\right)}P_{\mathrm{Mg}}^0}{P_{\mathrm{Mg}\left(\mathrm{flux}\right)}}& \left(7\right)\end{array}$

• • where γ is the activity coefficient of magnesium in Mg—Al alloy; x_{Mg (alloy)} is the magnesium molar content in Mg—Al alloy; P⁰_Mg=1.9 atm is the magnesium vapor pressure at T=1175° C.; and P_{Mg(molten salt)} is the partial pressure of magnesium vapor in equilibrium with liquid magnesium dissolved in molten salt at the cathode, where P_{Mg(molten salt)}<1 atm.

According to Equation (7), the first electrorefining potential and the open circuit voltage depend on the ratio of activity of magnesium dissolved in the molten salt and the activity of magnesium in the alloy. There is a minor difference between the open circuit voltage and the first electrorefining potential in the experimental measurement caused by the fluctuating value of P_{Mg(molten salt)}.

The second current jump at the higher potential is due to the following reaction:

Anode: Mg(alloy)→Mg2++2e−  (8)

Cathode: Mg2++2e−→Mg(g)  (9)

• • where the overall reaction is:

Mg(alloy)→Mg(g)  (10)

The theoretical expression of electrorefining potential for Equation (10) is given by

$\begin{array}{cc}{E}_{\mathrm{ER}2}=E_{\mathrm{anode}}-E_{\mathrm{cathode}}=\frac{\mathrm{RT}}{2F}\ln \frac{a_{\mathrm{Mg}\left(\mathrm{alloy}\right)}}{a_{\mathrm{Mg}\left(g\right)}}=\frac{\mathrm{RT}}{2F}\ln \frac{\gamma x_{\mathrm{Mg}\left(\mathrm{alloy}\right)}P_{\mathrm{Mg}}^0}{P_{\mathrm{Mg}\left(g\right)}}& \left(11\right)\end{array}$

• • and is plotted in FIG. 1, where γ is the activity coefficient of magnesium in Mg—Al alloy; x_Mg(alloy) is the magnesium molar content in Mg—Al alloy; P_Mg⁰=1.9 atm is the magnesium saturated vapor pressure at T=1175° C.; and P_Mg(g)=1 atm is the magnesium vapor partial pressure at the cathode. The electrorefining potential depends on the ratio of magnesium vapor activity at the cathode and magnesium activity in the alloy.

From FIG. 1, as the magnesium content in the scrap goes down, the second electrorefining potential increases. The experimental result of the dependence of the second electrorefining potential and the open circuit voltage is shown in FIG. 4. The second electrorefining potential and open circuit voltage increase with time, which indicates that the magnesium content in the scrap is decreasing as refining proceeds. The reason why the magnesium content in the scrap decreases is due to the finite solubility of magnesium in the molten salt. Magnesium dissolution in the molten salt is followed by vapor phase removal of the dissolved magnesium from the molten salt. The refining process can be expressed as: Mg(alloy)→Mg(molten salt)→Mg(g). Overall, pure magnesium is refined from magnesium scrap.

2. SOM Electrolysis Process

The SOM process was started at 5 hrs and 16 minutes after the maximum operating temperature was reached (9:16 PM). A PDS was added at the same time, with sweep rate of 50 mV/s and contacting area of 20.2683 cm² (FIG. 54).

A potentiostatic scan (PSS) was done at 9:33 PM, with applied potential of 3 V (FIG. 55). Magnesium obtained was 0.1848 g and current efficiency was 41%.

A PDS was added at 10:38 PM, with sweep rate of 5 mV/s (FIG. 56).

Impedance spectroscopy (IS) was measured at 11:07 PM, indicating a cell ohmic resistance of 1.28 ohms (FIG. 57).

A potentiostatic scan (PSS) was done at 11:15 PM, with applied potential of 3 V (FIG. 58). Magnesium obtained was 0.2253 g and current efficiency was 39%.

A PDS was added at 12:17 AM, with sweep rate of 5 mV/s (FIG. 59).

In summary, once the second electrorefining potential reached 0.13V, the anode was switched from the scrap anode to the liquid silver and carbon rod inside the YSZ tube, and a new potential was applied. When the potential reaches the dissociation potential of magnesium oxide, oxygen ions are pumped out of the molten salt through the YSZ membrane and are oxidized at the liquid silver anode by the carbon rod; meanwhile, magnesium ions at the cathode are reduced to magnesium metal vapor, which is collected in the condenser.

Electrolysis was performed for a total of two hours at a potential of 3V. The lower curve in FIG. 60 shows the initial potentiodynamic scan before any SOM electrolysis was performed, and shows an electronic current of 0.35 A due to the dissolution of magnesium from the scrap into the molten salt. The lower curve in FIG. 61 shows the current-time relationship during the first hour of electrolysis. Current efficiency is defined as the ratio of Faradic current to total current, and was calculated to be approximately 41% for the first hour of electrolysis. This is done by measuring the flow rate of carbon monoxide produced from the reaction of the carbon rod and oxygen on the anode side of the YSZ membrane. Based on the volume of carbon monoxide generated, magnesium reduced at the cathode was calculated to be 0.18 g during the first hour of SOM electrolysis.

The upper curve in FIG. 60 shows the potentiodynamic scan before the second hour of SOM electrolysis, which shows the leakage current has increased due to a higher concentration of magnesium dissolved in the molten salt. The upper curve in FIG. 61 shows the current-time relationship during the second hour of SOM electrolysis. For the second hour of electrolysis, the current efficiency was 39%, and the amount of magnesium reduced was 0.22 g. The total amount of magnesium reduced in the entire two hours of SOM electrolysis process was 0.4 g.

The combination of refining and SOM electrolysis processes are shown in the schematic in FIG. 27.

Leakage current for the SOM process is high right after electrorefining. Magnesium obtained from the SOM process is about 0.4101 g.

In sum, 7.443 g of pure Mg is collected. Some came from the starting Mg—Al alloy, which contained 9.624 g Mg and 9.388 g Al; and some came from the molten salt, the initial composition of which was MgF₂—CaF₂—10 wt % MgO—2 wt % YF₃. Yield efficiency was 77.19%. The alloy stayed at the bottom of the alloy crucible, and the 2 g of iron powder and the iron from the crucible assist in increasing the alloy density. The MgO and Al₂O₃ layer outside the alloy increases the melting point of the alloy. When melting the salt on top of the alloy crucible, there is oxidation of the metal. When welding the crucible, there is oxidation of the metal. When making alloy, there is a small amount of oxidation of the alloy.

A piece of collected Mg was characterized by EDS, which indicated that the collected magnesium is pure (FIG. 62). The atomic percent of magnesium was 99.6% and aluminum was 0.4%.

The scrap residue remained at the bottom of the alloy crucible. The iron powder and iron and chromium from the SS 304 crucibles alloyed with the scrap, which has a density higher than the molten salt. EDS was performed for the magnesium residue, and the results are shown in (FIG. 63). The atomic percent of Mg was ˜0%, Al was 51%, Cr was 13% and Fe was 36%. Only trace magnesium remains in the alloy, thus approximately 100% of the magnesium has been refined. Future experiments are also planned wherein the stainless steel parts will be replaced by low carbon steel to minimize additional dissolution of elements (Cr, Ni and Mn).

A piece of remaining alloy taken from its center was characterized by EDS (FIG. 64). From the gray zone (6450) of the center, atomic percent of Mg was 0.01%, Al was 44.89%, Cr was 8.37% and Fe was 46.73% (FIG. 65). From the black zone (6451) of the center, atomic percent of Mg was 0.22%, Al was 57.89%, Cr was 5.38% and Fe was 36.51% (FIG. 66).

Total yield efficiency of Mg is 7.443/9.624=77.19%. Yield efficiency of refining Mg is (7.443−0.4101)/9.624=73.08%. Yield efficiency of dissolution Mg is (7.443−0.4101−0.1227)/9.624=71.8%. Yield efficiency of electrorefining Mg is 0.1227/9.624=1.27%. Yield efficiency of SOM is 0.4101/9.624=4.26%. Current efficiency of SOM is 41% at first electrolysis and 39% at second electrolysis. Current efficiency of SOM is low because of electronic conductivity in the molten salt due to dissolved magnesium. The SOM efficiency can be improved after removing the dissolved magnesium. Pure Mg was collected and the remaining amount of Mg in the alloy is negligible.

The combination of the novel refining process and the SOM electrolysis process for recycling magnesium from magnesium scrap was carried out, and yielded 7.4 g of pure magnesium metal. A small amount of magnesium condensed on the walls of the condenser and dissolved in the 680 g of molten salt, and could not be quantified. No measurable amount of magnesium remained in the magnesium scrap, which implies that the refining efficiency is almost 100% using this novel refining process of dissolving magnesium and its oxide from the scrap into the molten salt, followed by vapor phase removal of dissolved magnesium. After the refining process, the SOM electrolysis process was employed in the refining system to enable the recycling of an additional 0.4 g magnesium from magnesium oxide in both the scrap and the molten salt. This paper presented the technical feasibility of recycling magnesium from partially oxidized Mg—Al alloy by the combination of refining and SOM electrolysis processes. In the future, this approach will be tested using larger amounts of oxidized scrap with additional impurities and less initial oxide content in the molten salt.

Example 3

The stability of YSZ in MgO—MgF₂—CaF₂ molten salt containing YF₃ and on Mg solubility as a function of temperature in MgO (MgF₂—CaF₂) molten salt were studied. YSZ stability was examined with impurities and with 2% YF₃. The yttrium profile for a 32 hour overlay is shown in FIG. 67. The solid line in FIG. 67 is the control with no yttria. Negative position corresponds to flux, while positive position values correspond to membrane. Impurity compositions are shown in Table 1:

TABLE 1
Impurity Composition
	Impurity	Max. Weight %	Molten salt Composition
	CaO	5	10% MgO, 5% CaO
			85% (MgF2—CaF2)
	SiO2	3	10% MgO, 3% SiO2
			87% (MgF2—CaF2)
	Na2O	2	10% MgO, 2% Na2O
			(MgF2—CaF2)
	All		10% MgO, 5% CaO,
	impurities		2% Na2O, 3% SiO2
			80% (MgF2—CaF2)

A 1.5% YF₃ in solution is close, but a higher concentration is needed to prevent the yttrium from diffusing into the molten salt.

A new molten salt made from LiF—MgF₂ was also tested for lower Mg solubility. The system diagram showing the eutectic melting temperature of LiF—MgF₂ is shown in FIG. 68 and in J. Am. Ceram. Soc. 1953, 36(1), 15; herein incorporated by reference in its entirety). The eutectic melting temperature is approx. 730° C. The eutectic melting temperature of 45% CaF₂—55MgF₂ is 975° C. A manometer was used for solubility measurements, determination of which was conducted in an enclosed chamber (FIG. 69). Magnesium and molten salt (6944) were placed in a beaker and acid and water (6945) were added. Magnesium is a vapor at 1091° C. so the pressure increase in the enclosed chamber was accounted for. Tosoh zirconia powder (700 mg) was added to Aremco 552 alumina paste (20 mL), and the resulting seal (2 applications) reduced the flow from 24 mL/min to 5 mL/min at 4 psi. Glass pastes, gold and silver o-rings were not sufficient. When magnesium is mixed with dilute acid in the flask, hydrogen gas evolves and pressure on the sample side increases. The difference in pressure between the two sides is ΔP=pgh. The initial volume of air on the sample side is known and designated as V₁, which is used to determine the initial moles of air, n₁, using the ideal gas law P₁V₁=n₁RT. ΔV is the change in volume on the sample side, where ΔV=πr²(h/2). Moles of hydrogen (n_H2) were obtained via the ideal gas law where (P₁+ΔP)(V₁+ΔV)=(n₁+n₂)RT. Moles of hydrogen produced in the test equal the moles of magnesium in the sample.

For calculation of P_mg various high T, K_solubility for each solution and temperature are calculated because high P_mg at higher temperature should increase solubility. K_solubility=P_mg/C_{mg solution}. P_mg is calculated using the Clausius-Clapeyron equation (Equations 12 and 13) where

∫_1atm^PmgdLn(p)=∫_1364k^TΔHvap/RT̂2  (12) and

Lnp|_1atm^Pmg=−ΔHvap/RT|_1364k^T  (13)

The P_mg at various temperatures is shown in Table 2.

TABLE 2
Pmg at various temperatures.
Temperature (C.) Pmg
1150             1.59
1165             1.73
1190             2.3

Mg solubility as a function of T is shown in Table 3.

TABLE 3
Mg solubility as a function of T.
				Weight % Mg in
Temperature		Cmg		solution at
(C.)	Pmg	(measured)	Ksolubility	Pmg = 1 atm
1150	1.59	0.153	10.386	0.074
1165	1.73	0.1656	10.442	0.075
1190	2.3	0.124	18.5423	0.043

Solubility of Mg can also be measured at 1250° C. and 1300° C.

As will be apparent to one of ordinary skill in the art from a reading of this disclosure, further embodiments of the present invention can be presented in forms other than those specifically disclosed above. The particular embodiments described above are, therefore, to be considered as illustrative and not restrictive. Those skilled in the art will recognize, or be able to ascertain, using no more than routine experimentation, numerous equivalents to the specific embodiments described herein. Although the invention has been described and illustrated in the foregoing illustrative embodiments, it is understood that the present disclosure has been made only by way of example, and that numerous changes in the details of implementation of the invention can be made without departing from the spirit and scope of the invention, which is limited only by the claims that follow. Features of the disclosed embodiments can be combined and rearranged in various ways within the scope and spirit of the invention. The scope of the invention is as set forth in the appended claims and equivalents thereof, rather than being limited to the examples contained in the foregoing description.

Claims

1. A method for recovering a target metal, comprising:

(a) providing a molten salt;

(b) dissolving a portion of a mixed metal scrap into the molten salt to form a molten salt and metal mixture, the mixed metal scrap comprising a target metal species and at least one contaminant metal species;

(c) bubbling a gas through the molten salt and metal mixture to form a gas and metal vapor mixture, the gas and metal vapor mixture comprising target metal vapors; and

(d) condensing at least a portion of the target metal vapors.

2. The method of claim 1, wherein the mixed metal scrap comprises an oxide of the target metal.

3. The method of claim 1, wherein the gas is inert.

4. The method of claim 1, wherein the gas is argon.

5. The method of claim 1, wherein the dissolving step comprises melting the mixed metal scrap.

6. The method of claim 1, wherein the target metal comprises magnesium, calcium or a lanthanide.

7. The method of claim 1, wherein the target metal comprises magnesium, calcium, or samarium.

8. The method of claim 1, wherein the target metal is calcium or magnesium.

9. An apparatus for recovering a target metal, comprising:

(a) a housing comprising a lower wall and a plurality of side walls;

(b) a divider at least partially disposed within the housing, the divider forming within the housing at least a first chamber, a second chamber, and a fluid conduit between the first and second chambers;

(c) a top wall cooperating with the lower wall and at least one of the plurality of side walls to enclose the second chamber;

(d) a plurality of gas inlets disposed in the second chamber; and

(e) a gas outlet in fluid communication with the second chamber.

10. The apparatus of claim 9, wherein the divider forms floating metal piers.

11. The apparatus of claim 10, wherein the gas inlets are disposed between the metal piers.

12. The apparatus of claim 9, further comprising a weir to recover floating material.

13. A method for recovering a target metal, comprising:

(a) providing a cathode in electrical contact with a molten salt;

(b) providing an anode in electrical contact with the molten salt;

(c) dissolving an oxide of a target metal into the molten salt to form a molten salt and metal ion mixture;

(d) establishing an electrical potential between the cathode and the anode to form target metal at the cathode;

(e) bubbling a gas through the molten salt and metal ion mixture to form a gas and metal vapor mixture, the gas and metal vapor mixture comprising target metal vapors comprised at least in part by a portion of the target metal formed at the cathode; and

(f) condensing at least a portion of the target metal vapors.

14. The method of claim 13, wherein the bubbling the gas through the molten salt and metal ion mixture comprises bubbling the gas in immediate proximity to the cathode.

15. The method of claim 13, wherein the bubbling the gas through the molten salt and metal ion mixture comprises providing the gas through at least one opening in the cathode.

16. The method of claim 13, wherein the bubbling the gas through the molten salt and metal ion mixture comprises providing the gas across a current path through the molten salt between the cathode and the anode.

17. The method of claim 13, wherein the gas is inert.

18. The method of claim 13, wherein the gas is argon.

19. The method of claim 13, wherein the reduced metal first forms as a dilute vapor in the gas.

20. The method of claim 13, further comprising measuring the quantity of target metal in the molten salt.

21. The method of claim 13, further comprising providing a SOM between the cathode and the anode.

22. The method of claim 13, wherein the oxide of the target metal further comprises at least one contaminant metal.

23. The method of claim 13, wherein the target metal comprises magnesium, calcium or a lanthanide.

24. The method of claim 13, wherein the target metal comprises magnesium, calcium, or samarium.

25. The method of claim 13, wherein the target metal is magnesium.

26. An apparatus for recovering a target metal, comprising:

(a) a container for holding a molten salt;

(b) a cathode disposed to be in electrical contact with the molten salt when the molten salt is disposed in the container;

(c) an oxide ion-conducting membrane disposed to be in ion-conducting contact with the molten salt when the molten salt is disposed in the container;

(d) an anode in electrical contact with the oxide ion-conducting membrane;

(e) a power source for generating an electric potential between the anode and the cathode;

(f) a gas inlet having at least one end disposed within the container to be below a level of the molten salt when the molten salt is disposed in the container; and

(g) a gas outlet in fluid communication with a volume defined by the container.

27. The apparatus of claim 26, wherein the at least one end of the gas inlet is in immediate proximity to the cathode.

28. The apparatus of claim 26, wherein the cathode comprises the at least one end of the gas inlet.

29. The apparatus of claim 26, wherein the at least one end of the gas inlet is disposed in the container to form gas bubbles in the molten salt, when the salt is disposed in the container, across a current path through the molten salt between the cathode and the anode.

30. The apparatus of claim 26, wherein the gas inlet is a nozzle.

31. The apparatus of claim 26, wherein the gas inlet is disposed at the bottom of the container.

32. The apparatus of claim 26, wherein the cathode forms at least portion of container.

33. A method for recovering a target metal, comprising:

(a) providing a molten salt;

(b) dissolving a portion of a mixed metal scrap into the molten salt to form a molten salt and metal mixture, the mixed metal scrap comprising a target metal species and at least one contaminant metal species;

(c) reducing a pressure of the molten salt and metal mixture to remove, as vapors of the target metal, at least a portion of the target metal species dissolved in the molten salt and metal mixture; and

(d) recovering at least a portion of the target metal vapors.

34. The method of claim 33, wherein the reducing a pressure of the molten salt and metal mixture comprises creating a partial vacuum in an atmosphere overlying the molten salt and metal mixture.

35. The method of claim 33, wherein the mixed metal scrap comprises an oxide of the target metal.

36. The method of claim 33, wherein the dissolving step comprises melting the scrap metal mixture.

37. The method of claim 33, further comprising providing SOM elements and performing SOM electrolysis.

38. The method of claim 33, further comprising dissolving an oxide of a second metal subsequent to production of at least some of the first metal in the salt.

39. The method of claim 33, wherein the target metal comprises magnesium, calcium or a lanthanide.

40. The method of claim 33, wherein the target metal comprises magnesium, calcium, or samarium.

41. The method of claim 33, wherein the target metal is magnesium.

42. An apparatus for recovering a target metal, comprising:

(a) a container for holding a molten salt;

(b) a cathode disposed to be in electrical contact with the molten salt when the molten salt is disposed in the container;

(c) an oxide ion-conducting membrane disposed to be in ion-conducting contact with the molten salt when the molten salt is disposed in the container;

(d) an anode in electrical contact with the oxide ion-conducting membrane;

(e) a power source for generating an electric potential between the anode and the cathode;

(f) a gas outlet in fluid communication with a volume defined by the container;

(g) a condenser in fluid communication with the gas outlet for condensing at least a portion of the target metal vapor in the gas stream exiting the container; and

(h) a vacuum source in fluid communication with the gas outlet and/or the container for creating at least a partial vacuum in the volume defined by the container.

43. A method for recovering a metal from an oxide of said metal, comprising:

(a) providing a cathode in electrical contact with a molten salt;

(b) providing an anode in electrical contact with the molten salt;

(c) dissolving an oxide of a first metal into the molten salt to form a molten salt and metal ion mixture;

(d) establishing an electrical potential between the cathode and the anode to form first metal at the cathode;

(e) dissolving an oxide of a second metal into the molten salt, the second metal being more electronegative than the first metal, and the second metal being less soluble in the molten salt than the first metal;

(f) subsequent to dissolving the oxide of the second metal into the molten salt, establishing an electrical potential between the cathode and the anode to form first metal at the cathode; and

(g) recovering at least a portion of the first metal formed at the cathode.

44. The method of claim 43, wherein the oxide of the second metal comprises nickel oxide.

45. The method of claim 43, wherein the metal charge comprises contaminant metals and/or the first metal.

46. The method of claim 43, wherein the first metal comprises magnesium, calcium or a lanthanide.

47. The method of claim 43, wherein the first metal comprises magnesium, calcium, or samarium.

48. The method of claim 43, wherein the first metal is magnesium.