One casting method that produces raw material for wire, copper wire in particular. and that has spread relatively far, is the vertical casting method (Upcast, Verticast) that can be used to manufacture oxygen-free (OF) copper wire by solidifying metal in a cooled nozzle on the upper surface of the melt and by pulling the solidified wire upwards. This raw wire, i. e., rod is rendered its final thickness by draw shaping. To obtain oxygen-free copper wire, the raw material used should almost exclusively consist of high-quality copper cathodes, which are melted in low frequency channel induction furnaces protected by wood charcoal to prevent oxidation. Charcoal has been used earlier in prior art by feeding it onto the surface of the melt, whereby wooden charcoal, lighter than the molten copper, floats on the surface of the melt without sinking. As a result, the reduction effect is low and, in addition, the char- coal on the surface of the melt is live and the atmospheric oxygen burns it, whereby the consumption of charcoal is unnecessarily high and the reduction effect minor.

The requirements of oxygen-free copper include a very low oxygen content, gener- ally a few ppm only, such as less than 5 ppm, i. e., less than 0.0005% O. As the oxygen in molten metal uses the graphite used for the nozzle material, thus shorten- ing the service life of the nozzle, the low oxygen content is good also for the casting process. The use of vertical casting has been limited by the relative slowness of the casting, i. e., one nozzle produces about 1000 tons a year. Furthermore, the space re- quired by the casting nozzles on the surface of the molten copper, as well as sepa- rate melting and foundry furnaces ensuring the absence of oxygen, including prob- lems with transferring the melt between the furnaces, have also been restricting fac- tors. The costs incurred by two furnace units have constituted a further restricting factor.

Publication JP-63-108946 tries to solve the problems described above by using a heat-maintaining furnace, to which the metal to be refined is transfused in a melted form and from which it is transfused to further processing in the melted form. Char- coal known per se is used as a reduction agent, but in this heat-maintaining furnace. the coals are pushed deeper down the melt by using a plate mainly consisting of aluminium oxide. In that case, the contact area between the coals and the molten metal becomes larger and. thus, the reduction slightly faster. According to the pub- lication, the furnace is heated by using an electric resistance that is above the molten metal, the coals, and the plate used for sinking them. The plate of insulating mate- rial is used to prevent the electric resistance from short-circuiting through the reduc- tion agent and/or the metal and, thus, possible accidents are also prevented. Fur- thermore, the arrangement presented by the publication has the disadvantage that the coals burn quickly because of the atmospheric effect. It is impossible to add coals during the process. To add coals, the process must be interrupted: the plate that presses the coal must be lifted up, and pushed down again after adding the coals, which makes the process slow and difficult. Furthermore, the speed of deoxi- dising as such should be improved. Publication JP-5-195104 has tested the effect of sinking the coal on the efficiency of oxygen removal in a melting pot. The arrange- ment according to this publication has all the same drawbacks as the method de- scribed by the publication mentioned above and, in addition, the arrangement is fully batch-type and thus not suitable for production in practice.

Publication US-5 211 744 describes a device for bringing scrap metal chips to mol- ten metal for melting them with minor combustion loss. For this, the device com- prises a feeding pipe with a downward opening cover attached to its lower end, the feeding pipe opening inside the cover. The cover is placed on the surface of the melt and the metal chips are forced along the feeding pipe to the melt under the cover, while inert gas, such as nitrogen or argon is fed into the feeding pipe. The purpose of the cover is to prevent the said inert gas from exiting to the atmosphere and to keep it under the cover to prevent the metal chips from burning under the effect of atmospheric oxygen on the surface of the melt. According to the publication, the scrap is not cleansed but the oil, lacquer, grease, hydrocarbons, polymers or similar vaporizable and inflammable compounds that come with it are allowed to vaporize from the metal chips and either accumulate under the cover, whereby the vapours are mixed with the inert gas. or exit through the exit ports on the edges of the cover, after which the vapours ignite and burn. According to the publication, the furnace is heated by flames coming from burners that use natural gas or fuel oil. The furnace

arrangement also comprises a circulating molten metal pump, the purpose of which is to adjust the temperature differences between the various parts of the furnace.

Publication US-5 735 935 describes an inert gas bubble-actuated molten metal cir- culating pump for transferring molten metal between the various parts of the fur- nace. A cover is arranged on the surface of the melt in the part of the furnace, where the inert gas is allowed to release and freely rise in the melt, the purpose of the cover being to prevent the inert gas from releasing and to keep it on the surface of the melt as a non-oxidizing layer. All the other areas of the melt surface are covered with dross in the conventional way. On the said spot, the cover also prevents the in- ert gas bubbles rising through the surface of the molten metal from splashing and spattering and the thin layer of oxide or the metal skin on the surface of the melt from breaking. Accordingly, the purpose is to keep the area of the surface of the molten metal as intact as possible. According to this publication, the furnace is heated by flames coming from burners that use natural gas or fuel oil and are lo- cated in the main chamber of the furnace, which has a considerably larger area than the portion of the furnace covered by the cover.

The object of the invention is to provide a refining method that can be used to quickly and effectively remove the substantially gasform impurities dissolved in the metal by using a reactant that remains in a solid state. This means that the reduc- tion or oxidation provided by the solid reactant should be rendered effective in rela- tion to the amount of metal in the process that is to be refined. Another object of the invention is to provide such a refining method that can be implemented in a con- tinuously working form; meaning that the process does not need to be interrupted to add the metal to be refined or the solid reactant, but in which, however, the metal to be refined and/or the solid reactant can be added periodically in batches, when needed. A third object of the invention is to provide a refining method, wherein the consumption of solid reactant could be made as low as possible for reasons other than reacting with impurities. A fourth object of the invention is to provide a refin- ing method that would reduce the consumption of solid reactant by preventing or decreasing the return oxidation or, correspondingly the return reduction of the mol- ten metal. A further object of the invention is to provide a refining method that keeps the equipment costs low or reasonable. This method relates to the further de- velopment of the melting and foundry furnace described in the previous non-public patent application FI-981821 of the same applicant.

The disadvantages described above can be eliminated and the objectives defined above are accomplished by the continuously working method according to the in-

vention, which is characterized in that, which is defined in the characterizing part of Claim 1.

Compared with prior art, one advantage of the invention is that it can be used, for example, to considerably enhance the oxygen removal from molten metal and, at the same time, to decrease the consumption of deoxidiser, such as coal, in useless combustion, and the dissolution of oxygen from the air into the molten metal. Fur- thermore, the transfer of molten metal can be decreased, because according to the invention. when needed, the metal can be melted and/or vertically cast into a rod or wire in the same furnace as the above-mentioned refining is carried out, without at- mospheric oxygen deteriorating the quality of the cast. However, the method of the invention can be applied by retaining the first-mentioned advantages, even if the metal refined in accordance with the invention is melted in a separate furnace and/or if the refined molten metal should be further processed, for example, further alloyed in a molten state separate from said refining. The method according to the invention also provides excellent opportunities to monitor and control the process. Another advantage is that, as the invention uses and it is preferable to use a shallower fur- nace than normally, there is more space for the casting nozzles than in prior art so- lutions.

In the following, the invention is described in detail with reference to the appended drawings.

Fig. 1 shows a general view of a molten metal processing container for implement- ing the first embodiments of the method according to the invention, in a vertical longitudinal section taken along the line I-I of Fig. 2.

Fig. 2 shows the processing container of Fig. 1 for implementing the method of the invention, as viewed from the top along the line 11-11 of Fig. 1.

Fig. 3 shows a general view of the molten metal processing container for imple- menting the second embodiments of the method according to the invention, in a cor- responding illustration as in Fig. 1.

Fig. 4 shows a general view of the molten metal processing container for imple- menting the third embodiments of the method according to the invention, in a corre- sponding illustration as in Figs. 1 and 3.

Fig. 5 shows a general view of the molten metal processing container for imple- menting the fourth embodiments of the method according to the invention, in a cor- responding illustration as in Figs. 1, 3. and 4.

Fig. 6 shows a general view of the molten metal processing container for imple- menting the fifth embodiments of the method according to the invention. in a corre- sponding illustration as in Figs. 1 and 3 to 5.

Fig. 7 shows a general view of the molten metal processing container for imple- menting the sixth embodiments of the method according to the invention, in a corre- sponding illustration as in Figs. 1 and 3 to 6.

Fig. 8 shows a general view of the molten metal processing container for imple- menting the seventh embodiments of the method according to the invention, in a corresponding illustration as in Figs. 1 and 3 to 7.

Fig. 9 shows a general view of the molten metal processing container for imple- menting the eighth embodiments of the method according to the invention, in a cor- responding illustration as in Figs. 1 and 3 to 8.

First, the refining equipment used in applying the invention is described. Typically, the processing container 1 of the metal MI to be refined according to the invention is a furnace that contains an inductor 3 and is preferably intended for vertical cast- ing. The reference M1 refers to the metal to be refined, the reference M2 to the re- fined metal, and the reference M3 to molten metal in general. In this case. the proc- essing container 1 is a channel induction furnace utilizing the mains frequency with one or more induction channels 13 that open to a melting chamber, molten metal M flowing through the ends of the channel into the induction channel 13. and the heated molten metal in the form of a strong flow P coming through the ends of the channel to the melting chamber of the processing container. According to the inven- tion. the induction channel 13 is preferably located on the bottom 24 of the process- ing container 1 so that the flow P coming from the channel hits the layer H2 of a re- actant 7, such as an oxidiser or deoxidiser. which is dealt with hereinafter. The in- ductors used can either be U channel inductors or W channel inductors. If the inten- tion is to melt blocks of metal 8 in a solid state in a direction S2a in the furnace, as in the embodiments of Figs. 1 to 3 and 6 ; according to the invention, it is appropri- ate to use an inductor that has as high a melting efficiency as possible, i. e., a so- called melting inductor, while in case already melted initial metal M1 is fed into the furnace in the direction S2b, as in the embodiment of Fig. 4, it is appropriate to use an inductor that renders the velocity of the flow P coming from the channel as high as possible, but the melting efficiency of which is low, i. e., a so-called mixing in- ductor. In a situation corresponding to the latter case, in which already melted metal MI that is to be refined is fed into the furnace in a direction S2c: according to the invention, the inductor can also be omitted. and the flow P of the molten metal to

the layer or reactant generates an initial metal with the aid of metallostatic pressure, as shown in Figs. 7 and 8. It is obvious that in the case of molten initial metal Ml, both ways mentioned above can be combined : in other words, the flow P can be generated by the mixing inductor that both comes from the direction S2c and has a relatively low power. By utilizing the flow of the incoming metal that is to be re- fined, overheating of the refined melt M can be avoided, if the molten metal M1 to be refined is at a sufficiently high temperature already when coming in. The width Wl of the melting chamber of the processing container 1 is larger, and preferably considerably larger, than the depth HI of the melt. Such design allows, among oth- ers. the feeding of whole or uncut copper cathodes 8 or similar pieces flatwise, i. e., mainly horizontally into the furnace 1, as marked in Figs. 1, 3. and 6 with the direc- tions of motion S2a. Metal that has been melted in advance can also be fed into the furnace in the direction S2b or S2c. as shown in Figs. 4 and 7-8. Accordingly, a suf- ficient number of mixing or melting inductors 3 has been attached to the bottom of the furnace or another suitable spot, especially to provide the flows P of the molten metal M in the processing container 1 but, when needed, also to generate melting power, and/or one or more entry ducts 23 of molten metal has been directed to the bottom of the furnace or another suitable place to provide the flows P of the molten metal M in the processing container 1. The term"furnace"in this description refers to the processing container of the melt, independent of whether melting power is exerted on it or not; therefore, Figs. 7 and 8 also present a furnace. In this connec- tion, the low depth HI of the furnace is especially useful, because the flow of melt P then effectively hits the layer of a light reactant from beneath, as described below.

The main part of the surface L of the molten metal M is covered by a cover-like float member 4, which leaves substantially free only the feeding area Al of the melt surface, through which the copper cathodes sink into the melt in the direction S2a, for example, or the molten metal is fed into the furnace from above in the direction S2b and, at the other, opposite end of the furnace, a casting area A2 necessary for the casting nozzles 5a, for example, for the vertical cast carried out upwards in the direction SI, or for the outflow means 5b of the melt for a downward discharge in the direction S3. In case metal MI that is melted in advance is fed into the melt M in the furnace as a flow that is directed towards the layer of reactant 7 in the direc- tion S2a from the area of the furnace bottom 24 or the end wall 25 or through them, the feeding area A 1 can be very small, just as the gap between the float and the side walls, as shown in Fig. 8 in particular. The metal that is to be cast and discharged consists of metal M2 refined in the processing container 1. The longitudinal edges of the float part, which connect the ends of the float 4 that limit the said feeding

area Al and the casting area A2, are as close as possible to the longitudinal side walls 26 of the furnace but, normally, with a minor clearance 18, however. The casting nozzles 5a and the discharge opening 5b can be of any suitable type; there- fore, they are not explained in detail. In the areas that remain free, the surface L of the molten metal M is protected against the ambient atmosphere, for example, by a layer H2 of graphite powder, graphite flakes 2 or dip coat salts. The float 4 can con- sist of a steel scale board, to which an insulating layer and refractory brickwork are attached, i. e., at least the lower parts of the float are made of refractory material.

Another type of construction of the float can also be used. Downward pointing ridges 11 are fitted to at least some of the edges of the lower surface of the float 4 : generally, to at least three edges and possibly to all the edges, the meaning of the ridges being explained later on. The height of the ridges 11 downwards from the lower surface 14 of the float 4 is at least as big as any predefined thickness H2 of the layer of reactant.

In the continuously working method according to the invention, any impurity agent dissolved in the metal in an substantially gaseous state can be removed by using a reactant that remains in a solid state and, first, has a high affinity with the said im- purity gas and as low an affinity and solubility as possible with molten metal. Oxy- gen can be such an impurity agent but, in some cases, possibly, also hydrogen. The reactant 7 that remains in the solid state can consist of coal in the form of charcoal or the like, for example, whereby we are talking about a deoxidiser. When the impu- rity gas is hydrogen or another similar gas, the solid reactant shall be an oxidiser.

Furthermore, the method according to the invention has the advantage that even though the solid reactant 7 is a deoxidiser, such as coal, the large total area A1+A2+A3 of the processing container, consisting of the contact area A3 of the float 4, the surface area 1 of the feeding area, and the surface area A2 of the casting area, first, is effective in contributing to the removal of hydrogen, oxygen, and other gases, which do not react with the reactant. This means that the processing container is shallow in relation to its volume, whereby the relation of the total area of the processing container 1-in which the molten metal settles-to the volume V of the melt, i. e., [A1+A2+A3] : V, is 0.71/,,, minimum and, possibly, 10'/m maximum, pref- erably within 1.5-5'/n,. Second, in the method according to the invention, the mol- ten metal M goes between the granules of the solid reactant 7. considerably increas- ing the actual area of the melt from the above-mentioned horizontal surface area A1+A2+A3. This effective area As of the melt, consisting of the surface area A4 of the melt that goes between the reactant granules and of the surface areas Al and A2 of the feeding and casting areas, i. e., As =A4+A2+A1. is at least ten-fold in relation

to the area of the processing container, i. e., As>10x [Al+A2+A3]. Correspondingly, the large total area A1+A2+A3 of the molten metal M in relation to the molten vol- ume V of the furnace contributes to the exit of oxygen and other gases that do not react with the reactant in a situation, where the solid reactant 7 is an oxidiser. Sec- ond, the specific weight of the light reactant is substantially lower than that of the said metal to be refined in the molten state. whereby the granules of the reactant do not sink into the molten metal, but it is possible to adjust its behaviour in the proc- ess. as described hereinafter. To refine the initial metal Ml, it is converted into melt M either in the processing container or outside the container. In addition, the reac- tant 7 described above, preferably consisting of pieces of a suitable size, is kept be- low the surface of the molten metal with the aid of a float means and, after refining, refined metal M2 is removed from the said processing container by any suitable way, as described above.

According to one embodiment of the invention, first, at least flows P of molten metal are generated in the said molten metal M by using electromagnetic induction, which flows are provided by the inductors 3 mentioned above and their induction channels 13. According to another embodiment of the invention, at least flows P of molten metal are generated in the said molten metal M in advance by using the flow of the molten initial metal M1 coming from the entry duct (s) 23, the flows being provided by the metallostatic pressure of the initial metal mentioned above. When a closed wall 30 surrounds the entry duct 23 above the molten metal surface L in the furnace 1 to a height H5, and the entry duct 23 opens to the melting chamber through the bottom 24 or the end wall 25 of the furnace, it is understandable that the flow P is discharged from the entry duct. The intensity of the flow P can be ren- dered as desired by planning the height H5 and the cross-sectional area (s) of the en- try duct (s) in advance to be of a certain size and, possibly, adjustable. If a relatively weak flow P is sufficient, it can also be accomplished by using an overhead dis- charging S2b only, as indicated by a reference P'included in the flow P in Fig. 4, in which case we must consider that the inductor 3 is omitted from the oven. These flows P of the molten metal are allowed to focus on the layer H2 of the light reac- tant 7 below the float 4. When the distance HI between the molten metal surface L in the melting chamber and the bottom is arranged small enough, of course, in rela- tion to the power of the inductor and/or the velocity of the incoming flow S2c or S2b, the flows are rendered at least relatively strong so that the flows P cause a dif- ferential motion of the pieces or granules of the light reactant and streams of melt between the pieces or granules of the light reactant. This substantially enhances the

reaction between the pieces of the layer H2 of the light reactant 7 and the molten metal M.

The inductors 3 can only be used for mixing the molten metal and the light reactant as described above, whereby the metal MI to be refined is melted somewhere else and either continuously or discontinuously run into the processing container 1, as shown in Fig. 4. Alternatively, the metal MI to be refined is preferably melted in the processing container 1 by using these inductors, as shown in Figs. 1, 3,5, and 6 and. in that case. the metal MI to be refined is added into the furnace in a solid state. In case the metal MI is added into the processing container 1 in the solid state, i. e. as pieces of metal 8, it is preferable to lay these pieces that have not yet melted in such a place inside the furnace 1 that the flow P of heated molten metal coming from the induction channel hits or is directed at these pieces of metal, as shown in Fig. 3. Then the directing of the flow P of molten metal against the layer H2 of the light reactant 7 can slightly or for some time be disturbed ; however, this does not have an essential effect, because the piece (s) of metal 8 then very quickly melts (melt) and, for the most part or at least for a sufficient time, the flow P of molten metal from the inductor is directed at the layer of light reactant on the lower surface of the float 4. When the metal Ml that is to be refined is melted by directing at it the flow P of molten metal coming from the inductor, there is the advantage that the flow P of molten metal-which is directed at the layer of light reactant- carries the recently melted metal that is to be refined along to the layer H2 of the light reactant 7, whereby the impurity agent in the metal reacts with the light reac- tant and mainly does not spread to the other parts of the processing container until afterwards. Accordingly, refining takes place very quickly and the effect of the non- refined melting metal on increasing the content of impurities in the molten metal M in the processing container 1 is substantially eliminated. Furthermore, the reaction between the light reactant 7 and the molten metal can be enhanced by increasing the thickness H2 of the reactant layer below the surface L of the molten metal M in the way described below. As applicable, the above description also relates to the effects accomplished by the flow of molten metal fed in through the end 25 or the bottom 24 of the processing container.

According to the invention, a relatively dense float 4 as such is further used in the processing container 1, the lower surface 14 of the float either being kept at a prede- fined constant or changing and adjustable height H3 in relation to the surface of the molten metal. Thus, the lower surface 14 of the float can be below the melt sur- face L [=-H3] or above the melt surface [+H3]. This height is smaller than the thickness H2 of the reactant layer on the surface of the melt, resulting in some of the

reactant 7 being inside the melt. There is no need for the float to be at the optimum height and the light reactant at the optimum depth in the molten metal M every moment. When the float 4 has a structure that is substantially gas-tight. comprising ridges 11 thai extend below the melt surface L at all its edges, the location of the lower surface 14 of the float above the melt surface does not disturb the process, because contact of the melt surface with the ambient atmosphere is prevented. Be- cause of the sufficient length of the possible feeding channels of the reactant 7 and/or the devices that provide the force F of the reactant feed, no significant amounts of detrimental gases from the ambient atmosphere can enter through the channels. Accordingly, the float 4 prevents contact of the main part of the molten metal M surface with the ambient atmosphere. In that case, the light reactant reacts with the impurity gas dissolved in the molten metal and the reaction result, which typically is a gas that does not dissolve in the molten metal, exits the molten metal, whereby it has been refined to a predefined content of impurity gas.

According to the invention, the light reactant 7 can further be fed below the float at such a speed that at least part of the thickness of the reactant layer extends below the surface of the molten metal. Feeding the light reactant is effected by using an external force F to push the molten metal below the surface against the metallostatic pressure of the melt. The external force F can preferably be exerted through one or more feeding channels 6a in the area of the float 4, the feeding channels extending from above the float below the same, as shown by Figs. 1,3, and 6. At the upper ends of the feeding channels 6a, equipment is arranged, such as a piston mechanism or a feed screw, to exert the said force F. Such feeding equipment 16a can consist of any suitable mechanism or means; therefore, it is not described in detail. Alterna- tively, the external force F and the route of the light reactant can be implemented by using a feeding channel 6b arranged at the bottom of the processing container 1 be- low the area of the float, extending through the float, suitable valve equipment 16b being arranged in connection with the feeding channel. When the valve is opened, a batch of light reactant is released to the molten metal so that it rises against the lower surface 14 of the float. The valve equipment can also be of any suitable type, which therefore is not explained in detail. However, this perhaps only applies to metals that have a relatively low melting point. Alternatively, the external force F and the route of the light reactant can be implemented below the edge of the float 4, whereby a clock 16c at the end of an arm is used, which either works automatically or manually, and which is pushed below the surface of the melt, for example, through the feeding area Al, stretched out inside the area of the float below the

float, and allowed to open or release the light reactant located in the clock, whereby it rises against the lower surface 14 of the float.

By using the float 4 and the feed of the light reactant 7 effected by the force F, the said reactant is forced inside the melt, whereby their reaction effect is substantially stronger than in prior art furnaces, in which light wooden charcoal floats on the sur- face of the melt, so that charcoal is unnecessarily burned and the ashes resulting from the combustion further form a detrimental separating layer. The pieces of light reactant below the float part 4 are protected so that the ambient air cannot get into contact with them ; hence, the charcoal, for example, is not burned unnecessarily.

This results in that the coal consumption and the slag formation are substantially decreased compared with prior art, whereby costs are saved both in purchasing and using the coal. The low or higher ridges 11 at the edges of the lower surface of the float prevent the pieces of wooden charcoal from sliding under the edge of the float part 6 or the ridge 11 and between the float part and the walls of the furnace.

The metal M1 that is to be refined is fed into the molten metal, which already is in the said processing container 1, through the feeding area Al at the first end of the container, either in a solid state in the direction S2a or in a melted state in the direc- tion S2b or S2c. Casting the refined melt M2 into the solid state and upwards Sl by using the casting nozzles 5a mentioned above or running the refined melt M2 downwards S3 through the nozzle 5b or horizontally, in a manner not shown in the figures, into a further processing container is carried out in the discharge area A2 of the other, opposite end of the furnace. Accordingly, the method can preferably be used to cast refined molten metal into rods or pipes in continuous vertical casting upwards, by using one or more nozzles 5a immersed in the molten metal in the said processing container from above. As the light reactant 7 can be added in the ways described in the previous chapter either continuously or discontinuously below the float 4 without stopping the process, the process can easily be rendered continuous.

As a continuous process, more metal to be refined and/or light reactant are fed into the processing container either periodically or continuously so that at least some of both substances are always present in the process, whereby no interruption occurs.

Similarly, casting into a pipe or rod or running through the nozzles 5a, 5b can take place quite continuously or periodically.

Thus, the float 4 in the processing container 1 prevents the main part of the surface L of the molten metal from contacting the ambient atmosphere in the manner de- scribed above. According to another feature of the invention, the processing con- tainer also substantially prevents the rest of the surface of the molten metal M, i. e., the molten metal surface L in the feeding area Al and the casting area A2 and the

areas of the clearances 18 from contacting the ambient atmosphere by having ar- ranged therein, on the surface of the melt, a layer 20 of graphite powder or graphite flakes or dip coat salts. As it does not melt but forms a tight layer on top of the mol- ten metal, both preventing the access of detrimental gases from the ambient atmos- phere to the molten metal and decreasing the heat loss from the furnace 1. the graphite is the most advantageous of these. Furthermore, the casting nozzles 5a can easily be pushed through the graphite layer 20 and feed the metal Ml that is to be refined in the directions S2a and S2b. Correspondingly, the contact of the molten metal surface with the ambient atmosphere is substantially prevented in a possible further processing container 29 by arranging a layer of graphite powder or graphite flakes or dip coat salts 20 on the surface of the melt.

When the reaction result of the light reactant and the impurity gas in the metal to be refined is gaseous G. it is allowed to exit GT also through the addition channel (s) 6a of the light reactant located in the float, so that no detrimental gases can pass from the ambient atmosphere to the opposite direction. In the feeding area Al and the casting area A2, gaseous reaction products G exit easily through the layer 20 of graphite powder or graphite flakes. It is also possible to arrange the float part 4 to pass gases from below upwards, whereby the flow GT of the releasing gaseous reac- tion results and/or the light reactant 7 and/or another dip coat and/or the design or structure of the gas flow channels of the float prevent the detrimental substances from the ambient atmosphere from getting into contact with the melt M. Of course, it is more difficult to use this structure to adjust the height of the melt surface L, L0 « L** by using under pressure-Ap or over pressure +Ap or any of their changes A [-Ap], A [+Ap], A [-Ap- +Ap] or A [+Ap +-Ap] which is described hereinafter.

As necessary, samples can be taken from the melt, analysed, and decisions concern- ing the power of the inductor 3, the velocity of adding the metal Ml that is to be re- fined, the discharging velocity of the refined metal M2, and/or the velocity of add- in-the light reactant 7 can either be made manually or automatically. According to a further feature of the invention, the composition of the gases, such as the gaseous reaction result of the impurities dissolved in a gaseous state and the light reactant, exiting through the addition channel 6a of the light reactant in the float, is analysed; for example, the content and/or the amount thereof by using a measuring device 19 and, based on that, the above-described or corresponding conclusions are made. Al- ternatively, the above-mentioned exiting gases and the said reaction result can also be analysed by the measuring device 19 by using a separate duct derived directly from below the lower surface 14 of the float, as shown by the dashed lines in Fig. 6.

The method according to the invention makes it possible to make predefined correc-

tions in the process on the basis of various measurements, whereby the content of the said impurity in the molten metal is kept below a predefined limit value or within predefined limit values. The adjustment can be implemented by changing the amount of light reactant 7 below the surface L of the molten metal M, i. e., the sub- merged light reactant, and the distance S between the light reactant layer H2 and the mouths of the induction channels 13 of the inductors 3. The thicker the thickness of the reactant layer H2 below the surface of the melt, the more effective the reaction, and the shorter the distance S between mouths of the induction channels and the light reactant, the more effective the reaction. If in the period of inspection, the thickness of the reactant layer 7 remains approximately unchanged but the float is moved downward, the distance S decreases. If in the initial stage, the bottom of the float was above the melt surface [situation +H3], the surface of the melt in the reac- tant layer rises, i. e., the amount of the submerged reactant increases. Both changes affect in the same direction, i. e.. enhance the process. This is a relatively quick ad- justment. Of course, the light reactant 7 can be added by keeping the lower surface 14 of the float at a constant height, whereby the thickness of the reactant layer H2 increases and a greater part of the layer is below the melt surface L. However, this is a slower way to react; hence, the first-mentioned changing of the float height is the actual controlled variable, which is excellent in that the variables affect in the same direction, whereby it is easy to converge the adjustment towards the desired value. It is obvious that if the reaction velocity must be reduced, measures opposite to the above description must be taken. In this way, the content of the said impurity gas in the molten metal is kept below the predefined limit value or within the de- sired limit values.

The method according to the invention is preferably implemented, when the metal to be refined MI consists of electrolytic copper, electrolytically produced copper cathodes in particular, whereby the gaseous impurity agent is oxygen. In that case, the said light reactant 7 is coal, charcoal in particular. The method according to the invention can also be applied, when the metal M 1 to be refined consists of a copper mixture. silver, aluminium or magnesium. Especially, when the said copper cath- odes are used as initial material, they are fed into the induction furnace mainly flat- wise in the direction S2a through the feeding area Al, as the figures show, after which they are allowed to melt in the strong molten flow P coming from the induc- tor 3.

As the vertical casting process is a continuous action, the float part 4 can be firmly attached to the edges of the furnace, as the float part 4 that forms a cover 4'in Fig.

9, while the level of the melt surface is defined by the relation of the cathode feed-

ing or the feeding of the molten metal to be refined and the casting velocity, and with the aid of over or under pressure arranged below the cover. However, the float part 4 moving in the vertical direction gives more degrees of freedom. The float part can be, for example, a structure that floats on suitable weights or the like at a prede- fined depth on the molten metal M surface, or a structure that passively hangs on brackets. or a structure that is adjusted to various, desired heights or various depths by using suitable mechanical members known per se, whereby the adjustment is carried out by actuators between the processing container and the structures sur- rounding the same and the float to a predefined or a changing height in relation to the molten metal surface. The lower surface 14 of the float 4 can be kept at the level of the molten metal, as in Figs. 1 and 4, or below the surface, as in Fig. 3, or above the surface, as in Fig. 5. An advantageous means to adjust the height H3 of the float lower surface 14 is to arrange under pressure-Ap, compared with the ambient atmosphere, below the lower surface of the float 4 so that, first, the gaseous reaction result of the impurity agent in the metal to be refined and the light reactant are re- moved more effectively from below the float and, second, the under pressure pulls the gas-tight float 4 deeper into the molten metal, while the molten metal surface L in the area of the float slightly rises compared with the surrounding surface of the melt. Thus, the under pressure can be used to implement the adjustment described above. Another advantage of the under pressure-Ap is that it also contributes to the removal of other gases, such as hydrogen, from the molten metal M.

To remove gases from the melt, such as aluminium and magnesium but also copper, higher under pressure can be used, if the sinking of the float in the position +H3 un- der the effect of under pressure is fully or partially prevented, for example, by con- verting the float 4 into the above-mentioned cover 4', and by dimensioning the height +H3 between the float lower surface 14 and the ridges 11 at its edges or, cor- respondingly, the height +H3 between the cover lower surface 14 and the free sur- face L of the melt that is in contact with the ambient pressure large enough. In that case, the melt surface L* defined by the ridges 11 of the float lower surface 14 rises, under the effect of the under pressure, considerably higher than the free melt surface L surrounding the float or the original melt surface L in the processing container, i. e., to a distance +H4 from the free or the original melt surface L, as shown in Figs.

5 and 9. For gas removal, the melt surface can be raised, depending on the molten metal, for example, by various dozens of centimetres, so that the under pressure -Ap is easily achieved, being about 20-30 kPa, which generally is sufficient for gas removal. Alternatively, over pressure +Ap can be arranged below the float 4 or the cover 4'. which is attached to the processing container 1. by adding into it suit-

able gas or air from the outside, whereby the melt surface L** in the layer of the re- actant 7 subsides for a distance-H4 compared with the original or the surrounding melt surface L. When using the cover 4', it is not necessary to use graphite powder, graphite flakes or a dip coat on the surface of the melt.

The amount of under pressure arranged below the float can also be varied during the work either on the side of the under pressure A [-Ap] or the over pressure A [+Ap] or from under pressure to over pressure A [-Ap-> +Ap] and back A [+Ap--Ap], i. e., generally by using a pressure change A [Ap] including all the possible variations of pressure. This variation causes a flow P of the melt inside the reactant layer and considerably enhances the process.

The area of the lower surface defined by the ridges 11 of the lower surface 14 of the float 4 can be divided up by locating additional ridges 11'in suitable places on the lower surface, as shown in Fig. 5. In the parts thus created, the under pressures-Ap and-Ap'can independently be adjusted and changed through gas ducts derived through the float 4.

Regarding Figs. 7 and 8, we must recognize that the entry duct (s) can extend throughout the width Wl of the furnace but it seems more preferable to limit them in this direction to only form part of the width Wl of the furnace, so that there are one or more entry ducts 23 in juxtaposition or otherwise arranged along the width of the furnace in the feeding area Al, just as there are exit channels or nozzles 5a and/or 5b in the casting area A2 of the furnace.