US3802871A - Refining of liquid copper - Google Patents

Abstract

A process for refining copper comprises removing anionic impurities by electrolytic or chemical reaction followed by removing the remaining impurities by contacting molten copper with a fluorine containing salt.

Description

United States Patent [191 [111 3,802,871

Lapat Apr. 9, 1974 REFINING OF LIQUID COPPER 1,869,498 8/1932 Oshorg 75/76 Inventor: Philip E. p u y, ss 2,222,592 11/1940 Dobrovolny 75/76 [73] Assignee: Kennecott Copper Corporation,

New York, NY. Primary E.\'aminerHyland Bizot Assistant Examiner-Peter D. Rosenberg [22] Filed May 1971 Attorney, Agent, or Firm-Lowell H. McCarter; John [21] App]. No.: 143,175 L. Sniado Related U.S. Application Data [62] Division of Ser. No. 841,629, July 14, 1969, Pat. No.

57 ABSTRACT [52} U.S. Cl. 75/93 R, 75/76, 204/64 R i i [51] Int. Cl. C22b 9/08, C22b 13/06, C22d 3/00 A process for refining copper comprises removing an- [58] Field of Search 204/64 R; i nic impurities by electrolytic or chemical reaction 75/72 76 93 followed by removing the remaining impurities by contacting molten copper with a fluorine containing References Cited salt.

UNITED STATES PATENTS 1.831432 12/1931 Hanak 75/76 9 Claims, 1 Drawing Figure IMPURE ANIONIC IMPURITY COPPER REMOVAL LIQUID ELECTORSLYTIC COPPER METALLIC CHEMICAL IMPURITY REMOVAL ANION; ELECTEOLYTIC E l/5 FUSED wi le/IL -*$8 wi CASTING FUSED SALT FUSED SALT RECYCLE fi g g J Nt METAL FLUORIDES FUSED sALT PURIFICATION IMPURITY METAL FLUORIDES ?ATENTEI]IIPII QIIIII 3,802.87]-

'MPURE ANIONIC '5' IMPURITY REMOVAL LIQUID ELECToRgLYTlC COPPER METALLIC CHEM'CAL IMPURITY l REMOVAL .IIIIEIIIES B$I EIIIIIID FUSED CHEMICAL :8PPER SALhTEUP CASTING ssa'ImII-G RECYCLE IMPURITY METAL FLUORIDES FUSED sALT PURIFICATION IMPuRITY METAL FLUORIDES BACKGROUND OF THE INVENTION Copper refining is conventionally practiced upon impure copper from three sources. These comprise (1) flotation concentrates, generally copper-iron sulfides, (2) precipitate copper obtained by cementation with scrap iron from aqueous solutions produced by leaching of mine waste dumps, and (3) secondary or scrap copper. Sources (1) and (2) pass through a wellestablished smelting and converting process, whose end product is an impure grade of metal known as blister copper, Source (3) may enter the refining process directly, or it may first be smelted in a blast furnace, whose impure product metal is known as black copper. Table 1 below gives analyses of a number of samples from various copper sources.

present in the crude copper warrant this expensive refining method. The first process step is fire refining much as described above, except that the fire refining is carried only far enough to permit casting of economically acceptable anodes.

copper molds oriented horizontally at the periphery of a wheel. Solidified anodes are mechanically or pneumatically raised for engagement of a tong under the anode lugs, and then transferred to cooling tanks. They are inspected, and trimmed where necessary with pneumatic tools for removal of edges and fins. Anodes typically have dimensions 36 inch wide X 39 inch long X l inches to 2 inches thick, and weigh between 500 and 800 pounds. Table 11 below gives some representative impurity analyses of copper anodes:

Analyses of Crude Coppers Misc. Copper Wt. Blister Coppers Black Coppers Scrap O .34 1.03 .26 .58 S .062 .014 .069 .069 Se .030 .030 .030 .020 Te .0012 .0008 .040 .0018

Ni .032 .035 .16 .031 1.27 .64 3.77 l 0 1 Fe .0039 .0062 .0031 .0068 Pb .0005 .27 .010 .0092 83 .55 .75 2.0 .1 Bi .00025 .00014 .00048 .00094 001 .0015

As .0027 .19 N.D. .0094 .12 .12 .22 Sb .0004 .020. N.D. .0054 .09 .18 .60 0.1 Sn .00027 .0002] .0012 .00086 1.85 .70 .10 1.0 .05 Zn .40 .35 0.5

Cu balance balance balance balance 91.] 94.8 94.1 92.0 97.5 7

Conventional refining isdone by either Fire Refining TABLE 11 or Electrolytic Refining.

Fire Refining comprises (1) transferring the crude liquid copper to a refining furnace (or melting it in that furnace should it arrive from the smelting process in solidified form), (2) oxidiaing the molten metal by introduction of air through iro n blowpipes, (3) frequently skimming slag from the surface of the bath, and (4) reducing or poling the metal which consists of inserting green hardwood poles into the molten metal bath that decompose into hydrocarbon gases and carbon. The refined copper is then ready for casting into desired shapes.

Fire refining is capable of removing substantially all of the sulfur, zinc, tin and iron, and partially removing several other impurities. Nickel, bismuth, selenium and tellurium, on the other hand, resist fire refining treatments almost completely.

The refining of a 300 ton charge of blister copper requires about 20 hours; charging time consumes 2 hours, melting and skimming 14 hours, oxidizing 1% hours and reducing 2% hours. Subsequent casting consumes 3%: hours, so that the entire process operates on a 24-hour cycle.

Electrolytic Refining is used only where the amount of. precious metals or the character of the impurities Impurity Analyses of Copper Anodes 1n the third process step, the anodes are suspended,-

along with pure cathode starting sheets, in electrolytic cells containing sulfuric acid and copper sulfate, and copper is elctrolytically transferred from anode to cathode over a period of 25 to 31 days. Impurities dissolve and are carried away in the recirculating electrolyte, or else fall to the bottom of the cells as solid particulate anode mud or slimes. Electrolysis cannot be carried to completion because anodes do not dissolve uniformly; 10 to 20 percent of the anode becomes scrap which is recycled to the fire refining furnace. Additionally, about one percent of the anode is converted to anode mud or slimes. This material is treated separately for recovery of Cu, Se, Te and the precious metals.

In a fourth process step, the cathodes are melted so that they may be cast into desired commercial shapes. Cathode melting should rationally be confined to bringing solid cathodes into the molten state under noncontaminating conditions with provision for elimination of the small amounts of sulfur and occluded gas which may be introduced with the charge. However, in reverberatory furnace practice, because of contamination from combustion gases and iron pickup from furnace pipe and rabbles, melting of cathodes requires essentially the same procedure as for fire-refining of impure copper. It is a batch process and every effort is made to perform the. complete cycle in 24 hours.

OBJECT OF INVENTION It is the principal object of the present invention to provide an improved process for purifying copper. It is a further object of the present invention to provide a rapid, effective, and less costly method for purifying copper. Another object is to provide a process for removing substantial amounts of impurity elements including but not limited to oxygen, sulfur, selenium, tellurium, nickel, iron, lead, bismuth, arsenic, antimony, tin, phosphorus, zinc, cadmium and hydrogen. Yet another object of the present invention is to decrease the concentrations of the aforementioned impurities to very low levels, such that the product copper is equal or superior in purity to conventionally refined copper.

The process of the present invention possesses several advantages over the prior art. One of these is that the process is conducted entirely at high temperature upon liquid copper, thereby eliminating the conventional unit operations of casting anodes and melting cathodes. Another advantage is that impurities are transferred out of the copper, rather than transferring copper away from the impurities as in conventional electrolytic refining. Since the quantity of impurities in crude copper is ordinarily about one one-hundredth of the total weight, the necessary residence time in the apparatus is therefore very much shortened, as from 32 days to a few minutes. Another advantage is that there is no anode scrap, which ordinarly constitutes -20 percent of the copper, and which must be recycled to the anode furnace because it is physically impractical to electrolyze an anode completely. Another advantage is that the refinery will be smaller and less costly to build, because the process of the present invention (1) has only two steps instead of the four steps of conventional electrolytic refining, and (2) the steps are much more rapid than conventional electrorefining. Another advantage is that operating costs will be smaller for the same two reasons.

Further objects and advantages of the present invention will appear hereinafter.

SUMMARY OF THE INVENTION In accordance with the process of the present invention, it has now been found that the foregoing objects and advantages may be readily accomplished by an improved process for purifying copper by providing molten copper, removing the anion forming elements, i.e., oxygen, sulfur, selenium, and tellurium, by making the impure molten copper the cathode of an operating electrolytic cell containing a fused salt electrolyte and an insoluble anode, transferring the anion-free copper from the electrolytic cell to a chemical reactor, removing the metallic impurities by contacting the molten copper with one or more fluorine-containing compounds in a chemical reactor and removing the purified copper from the reactor.

Alternately, anion impurity elements can be removed by contacting impure copper with a molten salt or a mixture of molten salts containing dissolved reactive metal, in a chemical reactor, and removing the anionfree copper to the second process step.

Metallic impurity removal can also be accomplished by providing copper free of anionic impurity elements from a first process step, and making this copper the anode of an electrolytic cell containing a fused salt electrolyte bearing fluorine-containing compounds, and removing the purified copper from the electrolytic cell.

The anion impurity elements and the cation or metallic impurities may be simultaneously removed electrochemically in a two compartment electrolytic cell having a common receptacle or container for the impure liquid copper and separate compartments for two electrolytes.

DESCRIPTION OF THE DRAWING The accompanying drawing illustrates an idealized flow diagram for the process of the invention described herein.

PREFERRED EMBODIMENTS First Step In the first step of the process of this invention the anionic impurities are removed by electrolytic or chemical reaction. In the electrolytic method the impure liquid copper is made the cathode of an electrolytic cell containing a fused salt electrolyte and an insoluble anode. The electrolyte composition in cathodic refining is not highly critical, but is conveniently rides or mixtures thereof because of their relatively low volatility near the melting point of copper, their chemical stability and their compatibility with graphite and certain other conventional oxide refractories. Alkali metal halides may also be added to the electrolyte comosition. The alkali metal halides may comprise from about 0.3 weight percent to about 10 weight percent of the electrolyte. A number of alternate electrolytes might also be used providing they meet the abovementioned criteria. In the first step the chalcogens or anionic impurity elements, for example, oxygen, sulfur, selenium, and tellurium are removed by an electrode reaction taking place at the cathode-electrolyte interface. The reaction may be represented as: Cu S (dissolved in metal phase) 2c-2 Cu( 1) S (transferred to the salt phase) where oxygen, selenium and tellurium also act after the manner of sulfur.

The effectiveness of the removal of the anionic impurities is demonstrated in Example 1.

EXAMPLE I 487 grams of impure copper having an initial composition as shown in Table III was melted along with 129 grams of anhydrous barium chloride electrolyte in an alumina crucible. The alumina crucible itself was contained in a large graphite crucible, a portion of whose sidewall was in contact with the electrolyte, and which served as the anode. The cathode current lead was a graphite rod extending down into the copper pool, and insulated from the electrolyte by an alumina sheath. The cell was maintained at 1150 C in an inert atmosphere, and electrolysis, was performed at 3.3 volts 5 with a cathode current density of 1.0 amps. per square centimeter. The average analysis of the impurities in the copper before and after electrolysis is shown in Table III.

Since the use of a graphite rod electrode results in oxygen removal by chemical as well as electrochemical reaction, this example is primarily a demonstration of S, Se, Te removal. Oxygen, however, has been found to behave similarly.

It is necessary in the case of oxygen removal experi- 3O ments to determine, by comparison with control experiments in which the salt and metal are brought into contact under identical conditions (except that no electric current is passed), how much deoxidation is attributable to chemical reaction with carbon. The difference between this and the total deoxidation in 21 cathodic electrolysis experiment may be ascribed to electrolysis Table IV contains examples of control experiments on all anionic impurities, and two electrolytic experiments which follow behavior of oxygen content:

TABLE IV Current Passed Element Observed Composition-Weight Percent Initial Final No S 0.58 0.46 Se .92 .78 Te 84 .66

No 0 0.1 1 .065 Yes 0 0.1 l .001 No 0: 1.5 .23 Yes O, 1.5 .001

Cu S (in liquid Cu) BaCl has F Clgtgfil 2131(1) and the electrolyte becomes fouled with oxygen, sulfur, selenium, and tellurium in the form of barium compounds. In some cases where the electrolyte is an inexpensive and impure salt to begin with, the electrolyte can be discarded when contaminated and sold for a noncritical use. I have calcium chloride especially in mind for this route.

A second method for dealing with the contaminated electrolyte depends upon the low solubility of the chalcogenide compounds in the alkaline earth chloride electrolytes. A portion of the electrolyte is continuously withdrawn from the electrolytic cell, cooled slightly to precipitate the chalcogenide compounds which are then mechanically separated by high temperature filtration or its equivalent. The purified electrolyte is then reheated and returned to the electrolytic cell.

Another, and the preferred method is to prevent impurity buildup by using a fluoride based electrolyte, in

which chalcogenide compounds are significantly solu- Standard Decomposition Potentials for Alkaline Earth Compounds at 1 C (volts) Ba Ca Mg Sulfide 1.82 2.10 1.45 OXid 2.I9 2.53 2.33 Chloride 3.40 3.18 2.29 Fluoride 5.07 5J4 4.51

Cu S (in liquid Cu) 2Cu(l) H2 S (g) A significant feature of the cathodic electrolysis is that when performed at constant current, the cell voltage is a function of the impurity concentration, and the cell voltage rises markedly as the end point of the reaction is approached. By limiting the available voltage through the cell, it is practicable to halt the reaction before the decomposition potential of the electrolyte is reached. In this way contamination of the purified copper by calcium, magnesium or other reactive metals derived from the electrolyte is avoided.

In the practical sense, a process conducted according to Reaction (1) is simply the desulfidation (or deoxidation, etc.) of copper by barium metal, and the use of the electrolytic cell may be viewed mainly as a convenient way to prepare barium metal and perform a reaction with it. Also, the barium metal will have been produced with less energy expenditure than if pure barium metal had been separately prepared.

The potential advantages of operating Cathodic Refining according to Reaction (2) rather than Reaction (1) include avoidance of problems of disposal of the chalcogenide compounds (e.g., BaS, BaO, BaSe; BaTe), so that an electrolyte purification circuit is unnecessary; the opportunity to recover Se and Te by condensation of the anode gas, and elimination of chlorine gas recycling (as ,by using it to chlorinate the chalcogenide compounds and regenerate BaCl An advantage of using Reaction (1) rather than Reaction (2) is that it can be performed in a chemical reactor without electrodes, as well as in an electrolytic cell. This is because many of the reactive metals are extensively or completely miscible with their halides, forming true liquid solutions at the temperatures where copper is molten. It thus becomes unnecessary for the liquid copper to remain in electrical contact with an electrode. Instead, the copper can be broken into small droplets and permitted to settle by gravity through a salt-filled vessel or contacting column, so that the reaction surface area of the copper droplets per unit area of plant floor space is much larger and the required plant size appropriately smaller than if an electrolytic cell is employed.

For example, if droplets of liquid copper 5mm in diameter are uniformly dispersed in a cylindrical saltfilled column 6 feet in diameter X 8 feet high, and the distance between droplets is 2 cm. the total surface area presented by the droplets is 904 ft If the column with ancillary equipment occupies a floor space 8 ft. X 8 ft., then there are I4.l ft of reaction interface per ft of plant floor space.

This figure is compared below with electrode areas typically encountered in existing commercial electrolytic processes:

salt is distinct and separate from its utilization in refining. Therefore (1) the reactive metal may be purchased instead of manufactured by the copper refinery, and (2) the optimal values of process parameters for manufacturing reactive metal, should the copper refinery choose to undertake it, will in general be different from those prevailing in the copper refining operation; separation of the two functions may therefore permit each to be performed under its most favorable circumstances.

In the removal of chalcogenide impurities by chemical reaction as described above, reaction products accumulate in the molten salt. In general their solubilities in the molten salt are believed to be quite limited, and they will be precipitated. The densities of compounds of reactive metals with chalcogenide elements in general lie between that of the molten salt and that of liquid copper. The solid compounds will therefore be trapped by gravity just above the interface between the metal and salt layers and can be removed from the system by gravity through a separate taphole located near the interface. The solid compounds will be admixed with salt. It may or may not be useful to separate the admixed salt, depending on whether or not the Se, Te and salt are to be recovered.

In judging the merit of the chemical reactor method for anion removal as heretofore described, a comparison naturally arises with conventional deoxidation processes wherein a reactive metal such as lithium or calcium is added directly to molten copper as a deoxidizer. The use of a solution of reactive metal dissolved in its halide is advantageous in that it provides a physical contacting process of the reagent with the copper which will be more uniform, controllable, safe and efficient than one employing reactive metal alone. Second Step Cathode Area,

Cathodes ft per ft Application Electrode per of Plant Floor Position Spacing Cell Space 1. Aqueous Elcctroefining Vertical 4" 40 I6 of Copper 2. Aluminum Electrowinning Horizontal 1 =01 3. Chemical Refining of Cu in the Salt Column None 14 described above "it is seen thata mgr;"temperateaeaia ae"can having a single liquid pool electrode, horizontally oriented, has an inherently small electrode area per unit area of plant floor space, and thus may not minimize the plant size or its capital cost as effectively. Stacking In the second step of the process of this invention the remaining base impurity elements such as nickel, iron, lead, bismuth, arsenic, antimony, tin, zinc, cadmium, and phosphorus are removed by subjecting the molten copper either to a chemical refining reaction or to an electrochemical refining reaction in which a copper fluoride equivalent is the reactive substance employed.

The following definitions will aid in the understanding of the invention as described hereinafter in the specification and in the interpretation of the appended claims.

The term CuF is defined for the purposes of this application as a molten solution of stoichiometric cupric fluoride, or CuF which has been saturated with,

or is in chemical equilibrium with, molten copper metal.

A measure of the ability of CuF, to perform fluorinating reactions is defined by its free energy of dissociation; which is the free energy change of the reaction:

2/): CuF, 2/x Cu 2 (g) This may be more conveniently stated by considering the dissociation reaction to be CuF Cu F for which the standard free energy change is approximately known, and accounting for the stoichiometric difference between CuF and CuF by means of an activity term which uses CuF as a standard state:

u) RT 111 calories.

( CuFz where AAG is the Gibbs free energy change, and a and a denotes activities at saturation.

The term fluorine potential, or p is another' tion (3): 7* 2 cu) (Pr Z CuF In this example a and a are not expected to vary significantly from unity, so that X atm.

The term copper fluoride equivalent is defined for the purposes of the application not simply as stoichiometric cupric fluoride having the formula CuF but more broadly as any product or mixture of products having an ability to perform fluorinating reactions upon liquid copper and the impurities contained therein which is controlled or limited by the thermochemical properties of molten CuF Thus copper fluoride equivalent" is any substance which upon contact with liquid copper provides a fluorine potential m, at least as great as does CuF,. Stated another way, copper fluoride equivalent is any fluorine-bearing compound that is thermodynamically less stable than CuF and consequently will act as a fluorinating agent for liquid copper and the impurity elements therein.

Thus, elemental fluorine gas introduced into molten copper in limited or small quantity, relative to the amount of copper, reacts to form CuF, and is therefore termed copper fluoride equivalent as defined above.

Similarly, elemental fluorine gas introduced into a molten fluoride salt solution which contains a non-stoichiometric fluorine compound such as CuF or a stoichiometric fluorine compound capable of accepting more fluorine to become a stoichiometric fluorine compound of higher valence, which is in contact with molten copper, the quantity of such fluorine gas being small or limited relative to the amount of molten copper, reacts with the non-stoichiometric fluorine compound or stoichiometric fluorine compound capable of accepting more fluorine to become a stoichiometric fluorine compound of higher valence in the salt to form CuF where x y s 2. CuF in turn reacts with molten copper to form more CuF,,. The process is thus an indirect fluorination of copper via the salt phase termed copper fluoride equivalent as defined above.

Similarly, a high valency fluorine compound which readily gives up a part of its fluorine content of molten copper, reverting in the process to a fluoride compound of lower valency, is termed copper fluoride equivalent. An example is antimony pentafluoride, which may react according to: i

Stannic fluoride (SnF is another example of a copper fluoride equivalent. Stannic fluoride readily gives up part of its fluorine under the conditions of refining liquid copper to become stannous fluoride (SnF Stannic fluoride is particularly adaptable to the process of this invention since it is a solid at room temperature, and yet sufficiently volatile at liquid copper temperatures so that it does not remain dissolved in the molten copper or the molten salt.

The majority of fluorine containing compounds are not termed copper fluoride equivalent because they are thermodynamically more stable than CuF and cannot act as fluorinating agents for copper.

Another method of defining copper fluoride" might have taken the specific term cuprous fluoride, or CuF. into account and described the composition CuF, in terms of a mixture of CuF CuF. or CuF CuF depending on the numerical value of x. There is however no good reason at the present time to invoke this additional complication in nomenclature, since the neutral compound CuF has not yet been characterized, or even proved to exist, either in gaseous or condensed form, at any temperature. Nevertheless, the definition of CuF, which has been adopted as an adequate formalism for purposes of this application should not be construed to mean that CuF does or does not exist, or that CuF species do or do not participate in the process of the present invention.

Cupric hydroxy fluoride [Cu(OH)F] is thought to be capable of approaching the performance I have observed with the copper fluoride equivalents as described herein. It is believed that under certain process conditions the use of cupric hydroxyfluoride could be considered an adequate substitute for CuF Utilization of a copper fluoride equivalent to refine molten copper by chemical reaction may be brought about in any of a number of ways.

1. Copper fluoride as anhydrous CuF may be brought into direct contact with molten copper in the chemical reactor.

2. Copper fluoride may be prepared by reacting or burning elemental fluorine gas and anion-free impure copper in a torch or high temperature reactor. The reaction or combustion products copper fluoride equivalent are then brought into contact with molten copper in the chemical reaction vessel.

3. Carbon and fluorine gas may be reacted as by contacting fluorine gas with charcoal. The reaction gases comprising principally CF. C F C l and C F are then passed through the molten copper where the gases decompose and the fluorine combines with copper to form the copper fluoride equivalent.

4. A portion of the reaction products of (3) above comprise cyclic fluorocarbon compounds which are liquid at room temperature, such as Cyclo-C F Cyclo-C F and Cyclo-C F This liquid mixture may be vaporized by heating, and then passed through the molten copper where the vapors decompose and the fluorine combines with copper to form the copper fluoride equivalent.

5. A copper fluoride equivalent such as antimony pentafluoride or stannic fluoride may be introduced into the molten copper in any convenient manner. This procedure is especially suitable because antimony pen,- tafluoride, antimony trifluoride, stannic fluoride and stannous fluoride are all volatile enough so that they evaporate both from the molten copper and salt phase, and do not accumulate. Antimony pentafluoride is a liquid and stannic fluoride is a solid at room temperature and can be conveniently handled.

6. It has been found that the copper fluoride equivalent need not be used at full strength, but may be dissolved in various other stable salt solvents such as alkaline earth fluoride eutectic mixtures. In some cases a salt containing very little dissolved copper fluoride equivalent may be adequate, depending upon which impurities are to be removed and to what concentrations they must be reduced.

Once of the advantages of using a copper fluoride equivalent in diluted form is that if the copper fluoride is largely consumed in the refining reaction it would appear to be feasible to repurify the salt phase containing the copper fluoride and the impurities by fused salt electrolysis in which the residual copper fluoride and then the impurity metal fluorides will decompose preferentially at the electrodes because they are thermodynamically less stable than the salt solvent. A second advantage is that the use of a diluent material may permit suppression of electronic conductivity thought to exist in copper-saturated liquid copper fluoride, which would greatly reduce current efficiency in electrolytic purification of the impurity laden fused salt. A third advantage is that the constituents of the salt phase may be chosen so as to depress the thermodynamic activities of specific impurity fluorides and so drive specific refining reactions further toward completion than otherwise possible. It has been found that as little as 1 mole percent copper fluoride in the electrolyte will be capable of refining liquid copper where the impurity level is low.

In the electrochemical method of performing the fluoride refining reaction an electrolytic cell is used rather than a chemical reactor. The major constituents of the electrolyte should be very stable thermodynamically (for example, CaF MgF nonvolatile, and liquid at temperatures substantially below the melting point of copper. The electrolyte should also contain a small concentration of copper fluoride. In the electrolytic cell there will be an impure liquid copper anode pool and a pure copper cathode pool. Electrolysis is performed with an applied voltage sufficiently small that the major electrolyte constituents do not take part in either electrode reaction and the impurities are not deposited at the cathodes.

Although any of the aforementioned methods of introducing a copper fluoride equivalent into the molten copper may be used in the process of this invention, for the purposes of the remainder of the discussion in this specification it will be assumed that copper fluoride was introduced as anhydrous copper fluoride.

The stoichiometry of the copper fluoride changes as it becomes saturated with copper metal until it attains the composition range of from about CuF to about CuF depending upon the temperature employed. The copper fluoride is largely immiscible in the liquid copper, and because of its lower density forms a slag or salt phase above the molten metal.

The refining reactions occur because of the favorable thermodynamic properties of copper fluoride as compared with the thermodynamic properties of the impurity element fluorides. More specifically the Gibbs free energy of formation is thought to be substantially less negative than forithe formation of the fluorides of the impurity elements.

The following typical reactions are believed to occur spontaneously in the molten copper when copper fluoride is introduced therein.

CuF Fe (dissolved in liquid Cu) FeF l) Cu( l CuF Ni (dissolved in liquid Cu) NiF l) Cu( l) CuF Pb (dissolved in liquid Cu) PbF l) Cu( l CuF 2/3 Bi (dissolved in liquid Cu) 2/3 BiF i Cu( l) CuF 2/3 As (dissolved in liquid Cu) AsF g) Cu( l) CuF 2/3 Sb (dissolved in liquid Cu) 2/3 SbF g) CuF Sn (dissolved in liquid Cu) SnF g) Cu( 1) An alternative formalism which can be used in writing the reactions where CuF is used is:

2/x CuF Fe (dissolved in liquid Cu) FeF- l 2/x Cu( l Information in the literature on standard free energies of formation of inorganic fluorides is neither extensive nor reliable, nor is thermodynamic activity data for the above noted species available except in a few cases, so it is not possible to predict in advance, through routine knowledge of the literature and the state-of-the-art of chemical process metallurgy, that all metallic and non-metallic impurities can be converted to fluorides by reaction with copper fluoride to an extent which makes these reactions useful for refining copper.

It has been experimentally found that when a charge of impure liquid or molten copper and cupric fluoride contact each other that the metallic impurities in the copper are rapidly and completely transferred to the salt phase. These impurities that have been transferred to the salt phase may be classified into two groups: (1) those impurities whose fluorides are volatile at the temperatures of the reaction and immediately escape as gases from the salt phase (such as arsenic, antimony, and tin) and (2) those impurities which are essentially nonvolatile and tend to remain in the salt phase such as nickel, iron, lead, and bismuth. The significance of this classification is that the impurities forming the volatile fluorides (a) do not accumulate in the salt phase and thus do not lead to the requirement for a salt repurification process, and (b) tend not to require as negative a free energy of reaction with copper fluoride as impurities forming non-volatile fluorides, because the their evaporation from the salt.

The following example s 2 through illustrate the re- 14 be extrapolated from this data that most impure coppers can be refined by the process of the present invention with no more than two stages of contact by copper fluoride.

moval of the metallic impurities from molten copper. 5 TABLE In these experiments a charge of impure molten copper was allowed to contact copper fluoride at 1 100 C Equlbmm lmpumy Level under an inert atmosphere in a graphite crucible. After A. Metal phase San ph BIA sufficient time for reaction the crucible contents were 0 g 2 1 m gg Percent) 310 cooled and the products analyzed. Table V below pres- :0068 ents impurity analyses of the 1n1t1al metal and the prod- Pb .00242 1.46 606. 4

' x ucts of each of the examples. 0009 0 693 3 5 i n is. i TABLE V I [ta/ c sts Example 2 Example 3 Example 4 Example 5 Starting Blister Metal Salt Metal Salt Metal Salt Metal Salt (Analyses in weight percent),

Ni .041 .00040 159 .00032 .304 .00074 .150 .00048 .168 Fe .001 .0oo1 .053 .00034 .100 .0oo1 .057 .00034 .061 Pb .019 .0001 .133 .0oo1 .220 .0oo1 .115 .0oo1 .084 Bi .0016 .00001 .0036 .00001 .0023 .00001 .0003 .00001 .0007 A5 .081 .0002 .0001 .0002 .0003 .0002 .0003 .0002 .0003 Sb .013 .0004 .0003 .0003 0001 .0003 .0001 .0003 .0002 F .0097 27.98 .0084 26.92 .0096 27.52 .0079 27.65 Cu bal. bill. 70.97 bal. 70.45 bal. 70.87 bal. 71.34

It is significant that the same degree of metal purification was achieved in these four examples despite a deliberate variation in experimental parameters as shown by Table V1:

An important cgnsideration in the process of copper fluoride refining is the nature of the CuF -Cu phase diagram, and more specifically the solubility of salt in the metal phase as the process temperature is raised above EXAMPLE 6 This example illustrates equilibrium at high impurity concentrations. The data was obtained by isolation of the two liquids, i.e., the liquid metal phase and the liquid salt phase at l,O9lC. The analyses shown below were confined to the nonvolatile impurities such as nickel, iron, lead and bismuth since these metal impurities tend to accumulate in the salt phase and must eventually be removed by purification. Table Vll shows that the ratio of impurity concentration in the salt phase (8.) to that in the metal phase (A.) is very high. It can the monotectic point (the lowest temperature at which the copper metal is liquid). It has been found experimentally that the solubility of salt in the metal phase is extremely small in the temperature range of interest. For example at 1128" C. the solubility of copper fluoride in molten copper is 522 parts per million by weight. At the monotectic temperature, 1,083C., the solubility of copper fluoride in molten copper is parts per million by weight. Because of the nature of the experimental method in determining these values they are regarded to be the upper limit of solubility so that the true values will be smaller.

EXAMPLE 7 This is an example of the use of copper fluoride in diluted form. A charge comprising 20 mole percent cop per fluoride, balance CaF -MgF eutectic mixture, was reacted with molten impure copper at 1114 C for 2 TABLE VIII IMPURTY ANALYSES melting point of copper in the choice of operating temlnitial Metal Product Equilibrium Compositions p fg l lhe practical utility of a high temperature metal rempurity Content Metal Phase Salt Phase weight weigh! weight fining process such as the one described n th s apphca tron is very much dependent on the availability of suit- Fe 0-049 0-00026 0-046 able materials of construction for containment of g; 3;; :ggggg fig metal, salt, and vapors which arise as products or are Bi .043 .000083 .036 introduced as reactants. l have found that graphite and Sb '00025 carbon are excellent materials for containment, and As .045 .00020 .0034

graphite for electrode leads. As an anode m cathodic electrolysis, the types of consumable carbon electrodes used in aluminum reduction appear to have merit. De-

EXAMPLE 8 TABLE IX IMPURITY ANALYSES Products, Isolated as Liquids Initial Metal 16 step of anionic impurity removal should be employed as part of an overall refining process; and (d) the effect of increased temperature is adverse but not unduly so thus permitting a latitude of at least 50 C above the pending on the salt composition present, certain oxide and silicate refractories are considered to be adequate. Finally, solidified salt may be employed as an insulating and sealing material, by proper engineering design of heat removal from the apparatus, much as frozen cryolite has proven to be the best insulator in aluminum reduction cells.

The engineering design of chemical reaction vessels for the process will emphasize certain criteria, among which are providing a high ratio of reactive surface to I 't C t M 1 Ph Sal Ph 522292,, on em 52 if fii volume, by dispersing metal droplets in the salt phase or salt droplets in the metal phase, or bubbles of reac- '32? 333% 3 tant vapor in the metal phase or in the salt phase. A secp I 4 10009 0nd criterion is the provision of a more or less sealed Bi @012 11009 vessel, such that moisture and other hydrogen sources As .045 .0088 .00005 l d d d d l Sb .04: .o002s .0005 3P E. f Oxygen r a es. 97'! s ster trons. While inert gases such as nitrogen and argon are suitable as gas blankets in the apparatus, I believe that less costl rotective atmos heres will suffice. De end- EXAMPLE 9 y p p p TABLE'X ing on the particular selection of impurity metals present and the fluorine potential in the apparatus, it may suffice to use dried air as a protective atmosphere; a1- ternately dried combustion gases containing no hydrogen may be used, such as a mixture of CO and CO.

There are two optional sequences in time by which the two steps of the metal refining process of this application may be conducted, and each has its own apparlMPURlTY ANALYSES initial Metal Analysis of Product Ratios* Impurity Content Metal Salt Impurity Salt Weight Weight Weight Removal Transfer Fe 1.0 .00031 1.65 3.2 X 10" 5.5 X 10 Ni 1.0 .0015 1.70 67 X 10 1.1 X 10 Pb 1.0 .00037 1.53 2 7 X 10' 41 X 10'' Bi 1.0 .00044 .066 2 3 X 10 l 5 X 10'' Impurity Removal nmio=tvigiii"%"iiiififir in initial metal/Weight impurity in product metal Salt transfer Ratio= weight impurity in product salt/weight impurity in product metal From observing adaiiisiimrpesaests asingsaus salt solvents as diluents containing from about 1 to 20 or more mole percent of copper fluoride in a salt mixture and various temperatures, time of reaction, types of impurities and concentration of impurities it can be concluded that (a) refining reactions are relatively rapid; (b) 5 mole percent of a copper fluoride in the salt charge is very nearly as effective as 20 mole percent copper fluoride; (e) sulfur and tellurium are not removed by the fluroide refining reaction so that the first discharging, and the end cathode at which copper metal and/or impurity metals are deposited. It is desirable to keep the two electrolytes physically separated into an anion removal section and a cation removal section.

This invention may be embodied in other forms or carried out in other ways without departing from the spirit or the essential characteristics thereof. The embodiments presented above are therefore to be considered as in all respects illustrative and not restrictive, the scope of the invention being indicated by the appended claims, and all changes which come within the meaning and range of equivalency are intended to be embraced therein.

What is claimed is:

1. The process of removing chalcogens from molten copper comprising contacting, in a chemical reactor, chalcogen containing molten copper with a solution of reactive metal dissolved in a molten salt solvent selected from the group consisting of alkaline earth chlorides, alkaline earth fluorides and mixtures thereof and separating chalcogen free copper from the molten salt solvent containing the reaction product of the reactive metal and the chalcogens.

2. The process of claim 1 wherein the molten salt solvent contains from about 0.3 to about weight percent alkali metal halides.

3. The process of claim 1 wherein the reactive metal is selected from the group consisting of alkali metals and the alkaline earth metals.

4. The process of removing metallic impurities from molten copper comprising the steps of contacting impure molten copper with a copper fluoride equivalent whereby the metallic impurities react with the copper fluoride equivalent to yield copper and impurity metal fluorides and separating metallic impurity free copper from the impurity metal fluorides.

5. The process of refining copper comprising the steps of a. reacting liquid copper containing impurities with a solution of a reactive metal dissolved in a molten salt solvent to remove the anionic impurities,

b. reacting liquid anionic impurity free copper with a copper fluoride equivalent to remove the metallic impurities and c. separating refined copper from reaction products.

6. The process of claim 5 wherein the reactive metal is selected from the group consisting of alkali metals, and alkaline earth metals.

7. The process of claim 6 wherein the molten salt is selected from the group consisting of alkaline earth chlorides, alkaline earth fluorides and mixtures thereof.

8. The process of claim 5 wherein the copper fluoride equivalent is selected from the group consisting of cupric fluoride, fluorine gas, fluorocarbon gas mixture, antimony pentafluoride, stannic fluoride and CuF where x is from about 0.90 to about 1.50.

9. The process of claim 8 wherein the copper fluoride equivalent is dissolved in a salt solvent selected from the group consisting of alkaline earth fluorides and mixtures thereof.

Claims (8)

1. 2. The process of claim 1 wherein the molten salt solvent contains from about 0.3 to about 10 weight percent alkali metal halides.
2. 3. The process of claim 1 wherein the reactive metal is selected from the group consisting of alkali metals and the alkaline earth metals.
3. 4. The process of removing metallic impurities from molten copper comprising the steps of contacting impure molten copper with a copper fluoride equivalent whereby the metallic impurities react with the copper fluoride equivalent to yield copper and impurity metal fluorides and separating metallic impurity free copper from the impurity metal fluorides.
4. 5. The process of refining copper comprising the steps of a. reacting liquid copper containing impurities with a solution of a reactive metal dissolved in a molten salt solvent to remove the anionic impurities, b. reacting liquid anionic impurity free copper with a copper fluoride equivalent to remove the metallic impurities and c. separating refined copper from reaction products.
5. 6. The process of claim 5 wherein the reactive metal is selected from the group consisting of alkali metals, and alkaline earth metals.
6. 7. The process of claim 6 wherein the molten salt is selected from the group consisting of alkaline earth chlorides, alkaline earth fluorides and mixtures thereof.
7. 8. The process of claim 5 wherein the copper fluoride equivalent is selected from the group consisting of cupric fluoride, fluorine gas, fluorocarbon gas mixture, antimony pentafluoride, stannic fluoride and CuFx where x is from about 0.90 to about 1.50.
8. 9. The process of claim 8 wherein the copper fluoride equivalent is dissolved in a salt solvent selected from the group consisting of alkaline earth fluorides and mixtures thereof.

Images