Method, apparatus and means for production of metals in a molten salt electrolyte

Abstract

This invention describes a method of producing a metal, M1, in an electrolytic cell consisting of a molten electrolyte, MZY-MZO, at least one anode and at least one cathode, characterised in that the passage of current between said anode(s) and cathode(s) through said electrolyte, produces a metal, M1, from a raw material, M1X, containing a non-metallic species, X, under conditions such that the potential at the cathode causes the reduction of the MZ cation and the formation of MZ at activities less than one, and the potential at the cathode is insufficient to cause formation of MZ metal as a discrete solid or liquid phase, and the MZ so produced reduces the raw material, M1X, at the cathode, to M1.

Description

INTRODUCTION

This invention describes a method, apparatus and means for production of a metal, metal alloy or metal composite in an electrolysis cell.

PRIOR ART

Titanium and its alloys exhibit excellent mechanical properties, unrivalled corrosion resistance and outstanding biocompatibility; however, annual global titanium production is dwarfed by commodity metals such as steel and aluminium. One needs only to examine the complex and discontinuous production method, the Kroll process, to correlate the high price of titanium with its low consumption. Since the 1950's, alternate processing routes have been sought, in vain; titanium oxides are extremely stable compounds that are bound with increasing tenacity to oxygen as the latter concentration decreases. A high solubility for oxygen in metallic titanium necessitates carbo-chlorination of titanium dioxide to produce an oxygen-free, chloride feedstock (TiCl₄), which is subsequently metallothermically reduced with liquid magnesium. Historical electrowinning processes have failed to successfully address the issue of dissolved titanium species, which may be present in up to 3 different oxidation states (4⁺, 3⁺, and 2⁺). Furthermore, due to the liquid melt processing of titanium during conventional arc or electron beam melting, segregation or incomplete dissolution of alloying additions may occur, limiting the range of compositions and types of materials that may be economically produced using the present process.

Several organisations including BHP Billiton, QinetiQ, Cambridge University, Nippon Light Metal Company and British Titanium have investigated the direct production of titanium from its oxides using an electrolytic cell that includes a graphitic anode and a CaCl₂-based molten salt electrolyte. The operating conditions for the processes are claimed to be greater than 1.3 volts and less than 3.5 volts over a temperature range of 600 to 1000° C. Furthermore, CaO is specified as an electrolyte constituent in all of these methods. The organisations may be categorized according to the mechanism they believe to be responsible for the actual conversion of metal oxide to metal. Cambridge University Technical Services Ltd. (WO 99/64638-U.S. Pat. No. 6,712,952) and British Titanium Plc (WO 2006/027612 A2) insist that the operating potential at the cathode is insufficient to cause deposition of the electrolyte cation or decomposition of the electrolyte. Rather, it is their stated understanding that the “reaction of the substance rather than deposition of the cation occurs” and the “substance or substances dissolves into the electrolyte” instead.

Claim 1 of U.S. Pat. No. 6,712,952 states a method that is operated by “ . . . applying a voltage between the electrode and the anode such that the potential at the electrode is lower than a deposition potential for the cation at a surface of the electrode and such that the substance dissolves in the electrolyte.” Claims 4 and 5 of patent application WO 2006/027612 A2 state, respectively “ . . . the voltage is applied such that the potential at the cathode increases to a potential which is less than a potential for continuous deposition of the cation from the electrolyte at the cathode.” and “ . . . the voltage is applied such that the effective potential between the cathode and the anode increases to a potential which is less than a potential sufficient to cause continuous decomposition of the electrolyte.” It may be understood from these prior arts that the (continuous) deposition of the electrolyte cation is totally avoided according to the methods so described. The electrolyte in the process described by these groups acts merely as a solvent for the reaction product, rather than participating in the electrochemical reduction reaction itself.

Conversely, BHP Billiton (U.S. Pat. No. 6,663,763) and Nippon Light Metal Company. (US 2004/0237711 A1) forward an alternate mechanism whereby the potential at the cathode is sufficiently reducing to deposit the cation from the electrolyte, since decomposition of the CaO constituent occurs at lower potentials than those for decomposition of CaCl₂. Claim 1 of U.S. Pat. No. 6,663,763 describes “ . . . operating the cell at a potential that is above a potential at which Ca cations in the electrolyte deposit as Ca metal on the cathode, whereby the Ca metal chemically reduces the cathode titanium oxide.” Claim 2 of the same document states “ . . . the Ca metal deposited on the cathode is soluble in the electrolyte and can dissolve in the electrolyte and thereby migrate to the vicinity of the cathode titanium oxide.” Within the description of the “Summary of the Invention” (line 41), it was stated, “ . . . the Ca metal that deposited on electrically conductive sections of the cathode was deposited predominantly as a separate phase . . . ”. A separate phase is equivalent to a condensed phase and calcium metal at unit activity.

Patent application US 2004/0237711 A1 describes, in paragraph [0049], “Moreover, the electrolysis of calcium oxide in calcium chloride between the anode made of a consumable carbonaceous anode material and the cathode made of non-consumable cathode material forms calcium-saturated calcium chloride either saturated with dissolved calcium or coexisting with pure calcium in the vicinity of the cathode . . . ” Thus the calcium formed by electrolysis is maintained at the saturation concentration as may be understood by the formation of “pure calcium” or Ca metal, which may be interpreted as either a discrete solid or liquid, depending on the operating temperature. Claim 1 of the same patent application states “ . . . the molten salt in the reaction region is electrolysed thereby converting the molten salt into a strongly reducing molten salt . . . ”. Someone skilled in the art would appreciate that the higher the extent of calcium saturation, the greater the reducing strength of the electrolyte. Consequently, the use of a “calcium-saturated” calcium chloride describes an extremely reducing electrolyte that has reached the solubility limit for dissolved calcium metal.

One of the inventors of the present application, has previously observed the conversion of TiO₂ to Ti—O solid solutions in a sequential process whereby titanium metal is formed via the lower oxides (Ti₃O₅, Ti₂O₃, TiO) Dring et al. (J. Electrochem. Soc. v152, #3 (2005):E104). It was stated that oxygen ionization was the dominant mechanism in a melt containing a very low CaO content (approximately 100 ppm CaO). Electrolysis was observed to occur at low current densities during the initial stages of reduction, but proceeded more rapidly as the decomposition potential of the electrolyte was approached. FIG. 1 depicts the current response to the potential waveform and there are clearly electrochemical processes, labeled A, B, C and D, occurring at potentials significantly positive of the reductive limit (E) of the electrolyte, where calcium metal would form. These electrochemical processes were attributed to the formation of Ti₃O₅, Ti₂O₃, TiO and Ti—O, respectively. Although metallic titanium was produced, significant quantities of calcium were also generated during the reduction. Thus, it has been established that a calcium-saturated melt or metallic calcium is not necessary to effect reduction of TiO₂ to the lower oxides of titanium or metallic titanium containing oxygen.

The applicant has extensively investigated the electrochemistry of the CaCl₂—CaO—Ti—O system in an attempt to rationalise the disagreement in mechanistic explanations. The experimental work conducted by the applicant has proven that it is impossible to operate under conditions where the electrolyte is not decomposed and the cation of the electrolyte is not deposited, owing to the thermodynamic and physical properties of the electrolyte system. The applicant has established that it is preferable to conduct the electrolysis under conditions where the cation deposition process results in a dilute and controlled concentration of dissolved calcium, which may effectively reduce the metal oxide or deoxidize the metal.

Further Prior Art Literature:

US Patent      U.S. Pat. No. 6,663,763 BHP Billiton
Documents      U.S. Pat. No. 6,712,952 Cambridge University
               Technical Services
               US 2004/0237711 A1 Nippon Light Metal Company Ltd
Foreign Patent WO 2006/027612 A2 British Titanium Plc
Documents
Other          R. Littlewood, J. Electrochem. Soc., v109, #6,
Publications   (1962): 525.
               H. Fischbach, Steel Research, v56, #7, (1985): 365.
               K. Dring, J. Electrochem. Soc., v152, #3, (2005): E113.
               K. Dring, J. Electrochem. Soc., v152, #10 (2005): D184.

Present Invention:

The present invention describes a process in which, a metal, M₁, is produced in an electrolytic cell consisting of a molten electrolyte, M_ZY-M_ZO, at least one anode and at least one cathode, characterised in that the passage of current between said anode(s) and cathode(s) through said electrolyte, produces a metal, M₁, from a raw material, M₁X, containing a non-metallic species, X, under conditions such that the potential at the cathode causes the reduction of the M_Z cation and the formation of M_Z at activities less than one. Additionally, the potential at the cathode is insufficient to cause formation of M_Z metal as a discrete solid or liquid phase. The M_Z produced in this manner reduces the raw material, M₁X, at the cathode, to M₁. The raw material feed may also contain both species M₁ and X, in a ternary or higher order oxide, of the form M_ZM₁X, by way of example.

One suitable molten electrolyte (M_ZY-M_ZO) for conducting such electrolysis is CaCl₂—CaO electrolyte. This molten salt used by both the applicant and other organisations is comprised of Ca²⁺ cations with Cl⁻ and O²⁻ anions. The CaO dissolved in the electrolyte is present as Ca²⁺ and O²⁻ ions and there is no distinction between Ca²⁺ cations that originate from CaO and those from CaCl₂. The consequence of this is that the cathode potential for reduction of Ca²⁺ is unchanged by addition of CaO. Decomposition of the electrolyte to form elemental calcium will occur once the cathode potential exceeds a certain threshold value, which may be determined from known thermodynamic data for a given temperature. At 900° C., the standard state potential for the reduction of the cation Ca²⁺ to calcium is 3.211 V negative of the standard chlorine electrode.

It has been established that many alkali and alkaline earth metals exhibit high solubility in their chloride salts. Unsurprisingly, calcium exhibits a high solubility in the CaCl₂—CaO electrolyte system (Fischbach). This has profound consequences for the cathodic process, as the form in which the cations of the electrolyte are present following reduction is no longer restricted to a discrete solid or liquid (depending on the electrolysis temperature). While cathode potentials equal or more negative than the standard state reduction potential for Ca²⁺ give rise to saturated, unit activity calcium, there is a range of potentials positive of this value that will result in the formation of calcium at less than unit activity. The activity of calcium metal is a frequently used term that corresponds to the concentration of calcium metal in the electrolyte. Unit activity denotes saturation and the presence of a condensed metallic calcium phase and values less than one correspond to dilute solutions of dissolved calcium metal. The exact correlation between electrode potential, E, and calcium activity, a_Ca, in equilibrium with the electrode potential, is given by Equation 1. The potential described by Equation 1 would be changed according to resistive and polarisation effects.

E=−RT/2F ln a_Ca  (1)

Where:

E is measured in volts versus the standard state potential for the reduction of Ca²⁺ to Ca⁰
R is the universal gas constant (8,3144 J·mol⁻¹·K⁻¹)
T is the temperature in Kelvin
2 is the number of electrons taking part in the reduction of Ca²⁺ to Ca⁰
F is Faraday's constant (96484.6 C·mol⁻¹)
a_Ca is the activity of calcium

Table 1 lists the standard state reduction potential (100% saturation of calcium in the electrolyte) of Ca²⁺ at 900° C. and the potentials calculated from Equation 1, corresponding to selected dissolved calcium activities less than one (less than 100% saturation of calcium in the electrolyte).

The electrochemical spectrum that represents the continuum of calcium activities from infinitely low values, (ostensibly, an activity of zero) to saturation (unit activity) may be further defined by a second variable of interest, such as melt calcium oxide activity. Predominance diagrams for the Ti—O—Ca—Cl system were constructed by Dring et al (J. Electrochem. Soc. v152, #10 (2005):D184), in the manner described by Littlewood (J. Electrochem. Soc. 1965). These maps delineate the regions of stability for the experimentally observed phases over a range of electrode potentials and melt CaO contents. This representation of the Ca—Cl—O system is shown graphically in FIG. 2 (J. Electrochem. Soc. v152, #10 (2005):D184), and may be considered the molten salt equivalent of Pourbaix diagrams. In the case of a CaCl₂—CaO molten salt, the x-axis is chosen as the negative, base-ten logarithm of the CaO activity, (−log₁₀[a_CaO]) in the melt, which is denoted by pO²⁻. For the sake of completion, possible electrode reactions for oxide anions in the presence of a graphite anode are shown.

As a rule, vertical lines on the diagram represent reactions where an exchange of O²⁻ occurs, such as in Reaction 2. Horizontal lines depict reactions where no O²⁻ is consumed or liberated, but a change in oxidation state occurs (Reaction 3). Sloping lines indicate a reaction that involves both electrons and oxide anions, such as Reaction 4.

M²⁺+O²⁻=MO  (2)

M²⁺+2e⁻=M  (3)

MO₂+2e⁻+Ca²+=MO+CaO  (4)

As can be seen from the diagram, the addition of CaO to CaCl₂ has absolutely no effect on the potential at which calcium forms. The only reactions that are modified by moving from low to high pO²⁻ (high to low CaO contents) are those involving O²⁻. The evolution of CO, CO₂ or O₂ all occur at increasingly positive potentials with decreasing melt CaO content. This, of course, reflects the fact that when the melt has lower amounts of CaO, a higher cell voltage is required to produce CO, CO₂ or O₂ at an equivalent partial pressure.

The low cell voltage, relative to the standard state decomposition potential of CaCl₂, reported by the inventors of competing processes, that may be applied in order to observe electrolysis current is due to the fact that oxide anions are oxidised at less positive potentials than chloride anions, and the potential for CO and CO₂ evolution on graphite anodes is significantly less positive than for O₂ evolution on inert anodes at the same CaO content.

Dring et al (J. Electrochem. Soc. v152, #10 (2005):D184), augmented FIG. 2 with the inclusion of the phase fields for various titanium oxides within a CaCl₂—CaO electrolyte. FIG. 3 shows the regions of stability of various titanium-oxygen-calcium compounds as a function of both melt calcium oxide activity and electrode potential, which is, as shown earlier, synonymous with calcium activity. FIG. 3 illustrates that the reduction of TiO₂ to all of its lower oxides and even low-oxygen metal occurs without calcium metal at unit activity (either solid or liquid) present, and that this may occur at calcium activities below 10⁻³ in melts having low CaO contents. Under these conditions, the melt contains neither calcium metal as a discrete phase nor the high concentrations of Ca⁰ necessary to constitute a “strongly reducing molten salt”. Operation of the electrolysis cell such that the cathode potential is less negative than that corresponding to saturated calcium formation, which results in the reduction of the electrolyte cation to produce calcium in the solvated state, is equally effective for reduction of the metal oxide. Thus it is not necessary to form the reductant as a metal at unit activity and await its dissolution and subsequent transport to the metal oxide.

Production of a calcium-saturated melt or calcium metal at the cathode is believed by other organisations to be the most effective method to reduce the titanium oxide. However, contrary to conventional belief, operation at higher dissolved calcium contents will result in lower production efficiency. This is because the rate of production of the metal of interest is not equal to the cathodic current. The latter is only an indication of the rate at which calcium is formed. The rate at which calcium reduces the titanium oxide determines the speed of the process. Although an electrolyte saturated with calcium reductant may be desirable, from a consideration of electrode kinetics and the thermodynamics of Ca²⁺ reduction, there are numerous complications associated with the formation of such high amounts of the calcium reductant metal. First, the exothermic reaction with the oxide precursor may result in the sintering of the feed material, with possible entrapment of reaction products and/or slowing of subsequent reduction rates. Any calcium reductant that does not reduce the titanium oxide at the cathode is able to diffuse away from the raw material, where it cannot do useful electrochemical work. Consequently, the calcium may chemically react with the anode material or anode off-gases resulting in excessive anode consumption/erosion and, if graphite anodes are used, the generation of free carbon or calcium carbides. Additionally, increasing contents of calcium in the electrolyte impart a high degree of electronic conductivity and create a short circuit path for electrons within the electrolyte, which should ideally function as an ionic conductor only. This final drawback is of the greatest concern, since current efficiency is significantly worsened as a consequence. The increased energy consumption leads not only to an increase in financial costs, but added environmental burdens arising from the prerequisite power generation.

The applicant has determined that since high amounts of calcium are detrimental to the efficiency of the electrolysis, it is preferable to ensure that throughout the reduction process the concentration and activity of calcium and calcium oxide are controlled. The preferred manner of control is via the use of a reference electrode whose potential is not affected by changes in the electrolyte composition, specifically the CaO concentration. By controlling the potential at the cathode with respect to this reference electrode potential, a fixed calcium activity/concentration may be obtained. This serves to maximize the reduction rate since the rate at which calcium is supplied to the raw material can be controlled to match the rate at which the reduction of metal oxide occurs. The melt calcium oxide content must also be controlled, and this may be understood in the context of an equilibrium between oxide ions In the titanium-containing raw material, and in the electrolyte (Reaction 5).

[O] in TiO_x or Ti—O=[O] in electrolyte  (5)

With a high electrolyte CaO content, there is a decreased driving force for oxygen to exit the titanium oxide or titanium-oxygen solid solution. Consequently, higher amounts of calcium are required to compensate and push the reaction in the direction of reducing the raw material. If calcium production is allowed to proceed uncontrolled, such that a highly calcium-saturated melt is formed, then the exterior regions of the titanium oxide are rapidly reduced forming high quantities of CaO. In such an event, the precipitation of solid CaO may occur when the local solubility limit is exceeded. Consequently, the transport of electrolyte and the CaO reaction product is greatly diminished, and the transport of calcium that was formed at the cathode may slow significantly or even fail to reach all areas of the cathode containing the raw material. If calcium continues to be generated, then this exterior surface may be reduced all the way to metal, which may then act as an additional surface area for the cathodic reduction of the Ca²⁺ cations in the melt. This would exacerbate the situation further, since even more of the cathode would be producing calcium at high activities.

The applicant has resolved that the optimum operating conditions are such that the consumption of calcium reductant by the raw material is matched by the rate of calcium generation at the cathode. The cathode potential may be maintained at potentials corresponding to the activities of calcium needed to effect each of the reduction process in order to proceed from TiO₂ to metallic titanium. This is accomplished using a stable reference electrode, which does not vary as the melt CaO composition changes. The benefits of this are two-fold: a lower cell voltage may be used, thus consuming less power; and significantly less calcium is formed The latter effect avoids the numerous disadvantages described above. The applicant believes that operation of the electrolysis cell in the manner described above is the only way to achieve a high quality metal, alloy or composite product at an acceptable price.

TABLE OF FIGURES

The invention shall be further explained by examples and figures where:

FIG. 1 is a current versus potential plot of TiO₂ and Mo in CaCl₂ at 900° C.

FIG. 2 is a predominance diagram showing the conditions of electrode potential and melt oxide content corresponding to a given electrochemical reaction for a system with a CaCl₂—CaO electrolyte and a graphite electrode(s).

FIG. 3 is a predominance diagram showing the conditions of electrode potential and melt oxide content corresponding to a given electrochemical reaction for a system with a CaCl₂—CaO electrolyte, a graphite anode(s), and a cathode consisting of titanium oxide.

FIG. 4 is a schematic diagram of the electrochemical cell used in conjunction with the present invention.

FIG. 5 is the x-ray diffraction pattern of cathode material produced after 24 hours with a cell voltage of 500 mV.

FIG. 6 is the x-ray diffraction pattern of cathode material produced after 24 hours with a cell voltage of 750 mV.

FIG. 7 is the x-ray diffraction pattern of cathode material produced after 24 hours with a cell voltage of 1000 mV.

FIG. 8 is an optical image of partially reduced cathode material exhibiting a metallic shell and a core consisting of oxides.

FIG. 9 is a potential versus time plot for TiO₂ reduced under constant current

FIG. 10 Scanning electron micrograph of Ti-10V-2Fe-3Al alloy produced via the present invention.

Table 1 lists the standard state reduction potential for Ca²⁺ to Ca⁰, and the potentials calculated from Equation 1 corresponding to selected dissolved activities of Ca⁰ at 900° C.

EXAMPLES

Example 1

Reduction of TiO₂ at Very Low Activities of Calcium

A molten salt reactor, depicted in FIG. 4, was assembled using vertical tube furnace with temperatures recorded using a thermocouple (1) within the cell and a PC-based data acquisition unit. A sealed inconel reaction (2) vessel housed alumina crucibles (3), which contained the CaCl₂—CaO electrolyte (4). This electrolyte was obtained by mixing thermally dried CaCl₂.2H₂O and 1 wt % CaO, and was subsequently heated in the retort under flowing argon (5, 6) to 1173 K. Once the electrolyte was molten, a graphite anode (7) was lowered Into the melt along with the cathode (8), which consisted of a TiO₂ pellet formed by pressing and sintering micron-scale powders. A voltage of 500 mV was applied across the anode and cathode for a time of 24 hours. A reference electrode (9) was used to monitor the cathode potential. At the conclusion of the experiment, the preform was lifted from the melt and allowed to cool in the upper chamber of the argon-purged cell. The sample was subsequently pulverised and analysed using x-ray diffraction. The experiment was repeated under identical conditions except different operating voltages (750 and 1000 mV) were used. FIGS. 5-7 show that the reduction of TiO₂ to lower oxides is possible under conditions completely devoid of calcium at unit activity.

Example 2

Reduction of TiO₂ Under Conditions of Constantly High Ca Activity

The experimental apparatus used in Example 1 was reproduced identically, except for the electrolysis voltage, which was fixed at 3V, and the duration of electrolysis, which was 12 hours. The sample was removed from the electrolyte, allowed to cool, and washed in water. A cross section of the sample (FIG. 8) revealed a metallic α-titanium case that enclosed a darker powder, which was identified by x-ray diffraction as a titanium sub-oxide. The thickness of the metallic layer was approximately 100-200 microns, which effectively acted as a diffusion barrier preventing full reduction of the titanium dioxide pellet.

Example 3

Reduction of TiO₂ Under Constant Reduction Current

An identical reactor to that used in Example 1 was employed to reduce 10-micron thick TiO₂ layers thermally formed on a titanium substrate. A potentiostat was used in conjunction with a graphite counter electrode, nickel/nickel chloride reference electrode and TiO₂ working electrodes. A constant reduction current was applied to the working electrode and the potential, with respect to the reference electrode, was recorded over time (FIG. 9). The reduction current was terminated when the working electrode potential reached a steady state value that did not continue to decrease over a long period of time. Since the TiO₂ layer was of finite thickness, the reduction current at the conclusion of the experiment must have been comprised primarily of calcium formation. Reduction was observed to occur in a sequential manner (events A, B, C, D, and E), with distinct transition times corresponding to the formation of the lower oxides of titanium and then titanium metal. The cathode potentials at A, B, C and D corresponded to dissolved calcium activities significantly less than unity. Specifically, at the onset of process A, the dissolved calcium activity in equilibrium with this potential is approximately 10⁻¹², the dissolved calcium activity at B is approximately 10⁻¹¹. Events C, D and E begin at dissolved calcium activities of 10⁻⁶, 10⁻⁵ and 10⁻¹, respectively. The end potential for this experiment was observed to be −1950 mV (versus the Ni/NiCl₂ reference), which resulted in the formation of α-Ti containing less than 1 wt % oxygen, as determined by x-ray diffraction analysis.

Example 4

Production of Conventional Titanium Alloy (Ti-10V-2Fe-3Al)

Reagent grade oxide powders from Alfa Aesar (TiO₂ 99.5%, FeTiO₃ 99.8%, Al₂O₃ 99.9% and V₂O₅ 99%, 1-2 μm particle size) were mixed, as-received, with a small amount of distilled water, which acted as a binding agent, to achieve a final composition of 10 wt % V, 2 wt % Fe, 3 wt % Al with the balance of titanium. The powder was then ground with a mortar and pestle for 5 minutes to break down large agglomerates prior to uniaxial compaction on a 15 mm diameter die at 100 MPa to obtain the desired preform shape. These preforms were placed in an alumina firing trough and sintered in air at 1373 K for 2 hours using a 3 K·min⁻¹ heating rate and a 6 K·min⁻¹ cooling rate. The preforms were placed inside the same reactor described in Example 1, then lowered into the electrolyte whilst suspended on a CP Gr 2 titanium wire and reduced under applied voltages of 1500 mV against a graphite anode rod for an initial period of 1 to 2 hours. The voltage was subsequently increased at a constant linear rate to 3100 mV for the remaining duration of the 24 hour reduction time. Analysis of the pellet indicated that a low-oxygen, α-β titanium alloy with a nominal composition of 8.6 V, 3.1 Al and 1.5 Fe was formed. FIG. 10 depicts the fine-grained microstructure and fully reduced alloy obtained from the present invention.

Claims

1. A method of producing a metal, M1, in an electrolytic cell consisting of a molten electrolyte, MZY-MZO, at least one anode and at least one cathode, characterised in that the passage of current between said anode(s) and cathode(s) through said electrolyte, produces a metal, M1, from a raw material, M1X, containing a non-metallic species, X, under conditions such that

a. the potential at the cathode causes the reduction of the Mz cation and the formation of Mz at activities less than one

b. the potential at the cathode is insufficient to cause formation of Mz metal as a discrete solid or liquid phase

c. and the MZ so produced reduces the raw material, M1X, at the cathode, to MZ.

2. A method of producing an alloy, M1-M2-... -Mn, in an electrolytic cell consisting of a molten electrolyte, MZY-MZO, at least one anode and at least one cathode, characterised in that the passage of current between said anode(s) and cathode(s) through said electrolyte, produces the alloy, M1-M2-... -MZ, from a mixture of raw materials, M1X-M2X-... -MnX, containing non-metallic species, X, under conditions such that

a. the potential at the cathode causes the reduction of the MZ cation and the formation of MZ at activities less than one

b. the potential at the cathode is insufficient to cause formation of MZ metal as a discrete solid or liquid phase

c. and the MZ so produced reduces the raw material at the cathode to an alloy M1-M2-... -Mn.

3. A method of producing a composite, M1-M2A, in an electrolytic cell consisting of a molten electrolyte, MZY-MZO, at least one anode and at least one cathode, characterised in that the passage of current between said anode(s) and cathode(s) through said electrolyte, produces a composite, M1-M2A, from a raw material, M1X-M2A, containing non-metallic species, X, and substance, A, under conditions such that

a. the potential at the cathode causes the reduction of the MZ cation and the formation of MZ at activities less than one

b. the potential at the cathode is insufficient to cause formation of MZ metal as a discrete solid or liquid phase

c. and the MZ so produced reduces the raw material at the cathode to form a composite, M1-M2A.

4. A method in accordance with claim 1, 2, or 3, characterised in that the formation of Mz more preferably occurs at an activity between 10−6 and 5×10−1.

5. A method in accordance with claim 1, 2, or 3, characterised in that the formation of MZ and the reduction of M1X occur at the same physical location.

6. A method in accordance with claim 1, 2, or 3, characterised in that the said electrolyte (fused salt) more preferably comprises Ca or Na, or a mixture thereof as the cation(s), MZ.

7. A method in accordance with claim 1, 2, or 3, characterised in that the said electrolyte (fused salt) may also comprise at least one of the following cations: Ba, Li, Sc, Sr or K

8. A method in accordance with claim 1, 2, or 3, characterised in that the said electrolyte (fused salt) more preferably comprises Cl as the anion, Y.

9. A method in accordance with claim 1, 2, or 3, characterised in that the said electrolyte (fused salt) may also comprise F as the anion, Y.

10. A method in claim 1, 2, or 3, where the electrolyte more preferably contains 0.1 to 3 wt % CaO.

11. A method according to claim 1, 2, or 3, in which the raw material contains M1 and X as constituents of a single phase compound comprised of greater than two elements.

12. A method according to claim 1, 2, or 3, in which the non-metal species, X, comprises at least one of the elements O, S, C or N.

13. A method according to claim 1, 2, or 3, in which the non-metal species, X, more preferably comprises the element O.

14. A method according to claim 3, in which the substance, A, consists of at least one of the following: B, C, O, N, or Si.

15. A method in accordance with claim 1, 2, or 3, characterised in that the metal species, M1, M2... Mn, being produced comprises at least one of the following components: Ti, Si, Ge, Zr, Hf, Sm, U, Al, Mg, Nd, Mo, Cr, Nb, V, Ta, Mb, W, Mn, Tc, Re, Fe, Ru, Os, Co, Rh, Ir, Ni, Pd, Pt, Cu, Ag, Au, Be, Sr, Ga, In, Tl, lanthanides or actinides.

16. An apparatus consisting of an electrolytic cell including an anode, a cathode comprised of a current collector and raw material, and a molten electrolyte, which operates under conditions such that a metal, alloy or composite is formed from said raw material whereby the potential at cathode is sufficient to reduce the MZ cation and causes the formation of MZ at activities less than one, and the potential at the cathode is insufficient to cause formation of MZ metal in either the solid or liquid phase, and the MZ so produced reduces the raw material, at the cathode, to said metal, alloy or composite.

17. The anode defined in claim 15 further comprising either

a. a carbonaceous material such that CO and CO2 gas are evolved during electrolysis, or

b. a material such that the anodic electrolysis product is oxygen gas.