US2711955A - Halide cracking-ingotting process - Google Patents

Description

June 28, 1955 F, JORDAN 2,711,955

' HALIDE CRACKING-INGOTTING PROCESS Filed Oct. 12, 1951 Sheets-Sheet l INVENTOR- {Willa W June 28, 1955 J, F. JORDAN HALIDE CRACKING-INGOTTING PROCESS 2 Sheets-Sheet 2 Filed Oct. 12, 1951 INVENTORz- 2,711,955 I HALIDE CRACKING-INGOTTING PROCESS James Fernando Jordan, Huntington Park, Calif.

' Application October 12, 1951, Serial No. 251,073

4 Claims. (CI. 75-84) My invention relates to metallurgy wherein a halide of a metal is cracked to yield said metal.

This is a continuation-in-part of my copending application, Serial No. 187,865 of October 2, 1950, now abandoned.

- Van Arkel and De Boer devised the procedure whereby the iodide of a metal is cracked to yield said metal by circulating the vaporous iodide into contact with an incandescent filament within an evacuated enve- 1ope.- This process has the capacity to produce small quantities of high purity metal; however, serious obstacles prevent the use of this process for the largescale production of pure metal. The scale of operation of the Van Arkel-De Boer procedure may be increased by increasing the surface of the incandescent filament and increasing the rate at which the vaporous iodide is circulated into contact therewith, or the scale of operation may be increased by merely increasing the number of operating units. While in the former case it might be possible to devise the formation of a shape that could be directly rolled into a few of the required commercial forms, the production of all of the required commercial forms by such a procedure seems very unpromising, and, accordingly, the product produced by employing large cracking surfaces would have to be melted and cast into ingot or slab form before all of the required commercial shapes could be produced by conventional metal-working methods. Other difficulties arise when large surfaces are to be subjected to vaporous iodide in accordance with Van Arkel-De Boer. For example, the maintenance of such large surfaces at the required cracking temperatures without locally melting said surface would be extremely difiicult. Another difficulty arises from the fact that deposition of the metal takes place under conditions which are ideal for the development of large crystals, with the result that the deposited surface is extremely rough, a condition that mitigates against the direct use of the metal deposit in a rolling operation. Increasing the production of the metal by merely increasing the number of the producing unit s devised by Van Arkel-De Boer would be so costly that it cannot be seriously considered; furthermore, metal produced in this manner would have to be melted to obtain the ingot and slab shapes which are required in order that the rolling mill may produce the commercially usable forms.

in spite of these difficulties, however, production of metals, such as titanium and zirconium, by a halide cracking process is in great demand, dueto the fact that the future growth of titanium and zirconium production depends, in large measure, on the ability of metallurgy to produce a metal that exhibits uniform physical prop erties, and such uniform properties can only be obtained if the metal contains uniform amounts of oxygen and/ or nitrogen. So far as is known at the present, only the halide cracking process is capable of producing titanium and zirconium with such uniformity-with respect to the oxygen and/or nitrogen alloyed therewith. Y

The most commonly-employed-method being presentl nited States atent i Q f" employed for the production of ductile titanium and zirconium involves, first, the production of a metal powder or sponge by the Kroll process, followed by an ingotting procedure wherein the metal powder/sponge is melted within a water-jacketed arc furnace under a pro tective atmosphere consisting of one of the inert gases, such as argon. The ingotting procedure may be operated to produce ingots one at a time or continuously, so as to produce a continuous ingot. Attention is also being given to they production of the metal powder by electrolytic and sonic methods, instead of by means of the Kroll reaction; however, whether the metal powder is produced by the Kroll reaction, electrolytically, or by sonic methods, ingotting in a water-jacketed arc furnace must be employed in order to produce the desired ingot and/ or slab.

As was mentioned previously, the production of commercially-usable forms requires that the metal deposited by halide-cracking processes be passed thru an ingotting step before the usable shapes can be prepared by conventional metal-working procedures. This conversion of the metal deposit into an ingot/ slab involves a number of steps, however. These steps involve: (1) cooling the metal deposit to room temperature so that said deposit can be removed from the cracking vessel without permitting the hot titanium/zirconium to contact air, (2) charging the metal into the melting unit presumably, but not necessarily, of the water-jacketed arc type, (3) melting the charged metal within said melting unit, and, finally, (4) casting the melted metal to form the desired cast shape.

in studying the matter over, it occurred to me that most of these steps could be eliminated, and a better product could be attained, the vaporous iodide were lead directly into the melting furnace, so that the thermal dissociation of the iodide took place within the vessel and in contact with the bath of molten metal or an electric arc that was playing upon said bath of molten metal. My copending application, Serial No. 160,017 of May 4, 1950, disclosed that process. I

My copending application, Serial No. 187,865 of Octoher 2, 1950, of which the present application is a continuation-in-part, disclosed my method of modifying the process of Serial No. 160,017 so as to increase the production rate of said process, and so as to improve the surface condition of the ingot/slab being produced by Serial No. 160,017. v

Figure 1 shows the basic features of my apparatus for carrying out my process. I Figure 2 shows one method of constructing the electrode shown in Figure 1. V Figure 3 shows another method of constructing said electrode. Figure 4 shows still another method of constructing said electrode.

Figure 5 shows certain manners in which the apparatus of Figure 1 may be modified.

My copending application, Serial No. 161,015 of May 9, 1950, now abandoned, disclosed my halide cracking process wherein a halide of a metal is cracked to yield said metal by introducing said halide into a molten dissociation media consisting of a column of molten salt, said salt being substantially inert towards said metal and said halide and being maintained at a temperature that lies between the dissociation temperature of said halide and the boiling point of said salt media. The purpose of Serial No. 161,015 is the production of a metal powder/sponge.

In Serial No. 187,865 and the present application, I float the column of molten dissociation media on the pool of molten metal contained within the ingotting apparatus of Serial No. 160,017, and then I crack the iodide/halide by introducing it into the column of molten salt so that it cracks as it rises, as a gas, thru said column. The dissociation column is maintained at a temperature that lies above the melting point of the metal being produced by the cracking process, and, accordingly, the discrete particles of metal produced within the media column are molten, said particles being collected in the pool of molten metal that lies at the base of said media column as circulation and settling brings said particles into contact with the surface of said pool.

In Figure l, I form my dissociation media as column 31 of molten salt within vessel 15. I support said column 31 by floating it upon metal pool 22, the line of separation between column 31 and pool 22 being slag-metal interface 24. Pool 22 and column 31 are maintained molten, in my preferred embodiment, by an electric arc struck between water-cooled electrode 19 and interface 24, electrode tip 21 being formed of a non-consumable metal, such as molybdenum, tungsten, or a sintered mixture of tungsten and thorium oxide. insulator 7 separates electrode 19 from vessel 15.

With the input of heat via the arc sufiicient to maintain pool 22 and column 31 molten, and at a temperature that lies above the dissociation temperature of the halide that is to be cracked, below the boiling point of the metal whose halide is to be cracked, and, preferably, below the boiling point of media 31. halide 6 that is to be cracked is introduced into column 31 via ports 16. The boiling points of the halides here involved being below their respective dissociation temperatures, halide 6 will take the form of a gas that is substantially instantaneously heated to the temperature of column 31 as said gas rises therethru. Upon reaching its dissociation temperature. gaseous halide 6 cracks to yield particles of molten metal dispersed within media 31 and a gas that will consist of the halogen involved, undissociated halide 6, and, in some cases, other halides arising from the dissociation of halide 6. This gas, contaminated with media 31, will pass out of vessel 15 via outlet 11, said gas being referred to as byproduct gas 10.

As the dissociation of halide 6 proceeds, the molten metal particles produced thereby will accumulate within column 31. Due to the fact that cracking reactions of this type tend to take place on surfaces, the molten metal particles suspended within column 31 will grow in size as the rising gas contacts and dissociates on said particles.

When the molten metal particles suspended within column 31 contact interface 24, said metal particles will consolidate within pool 22. The tendency of said particles to settle to surface 24 will be encouraged by the aforementioned tendency of said particles to grow in size.

As the precipitated metal particles accumulate within pool 22, the level or surface 24 will rise within essel 15. As surface 24 rises within vessel 15, the distance between surface 24 and electrode 19 will decrease, resulting in a change in the characteristics of the current flowing therebetween. By maintaining a current of uniform characteristics flowing between electrode 19 and surface 24, by regulating the gap therebetween, by regulating the rate at which cast shape 18 is withdrawn from vessel 15. surface 24 may be maintained at a substantially fixed position within mold 26. The continuous input of halide 6 into column 31. together with the continuous withdrawal of shape 18 from vessel 15 in accordance with the current flowing between electrode 19 and surface 24, yields a continuous cracking-ingotting process.

The apparatus of Figure 1 has the great advantage that the process is carried out within a completely waterjacketed vessel, thus avoiding the difliculties encountered when some metals contact non-metallic refractories, and thus avoiding the problem of handling very hot halogen-- it being realized that the entire inner surface of vessel 15 is protected by a coat of solidified media 31 and that this coat prevents the hot halogen from contacting the metal water-jacket and acts to minimize the loss of heat as water layer 14 maintains mold 26 in the chilled condi tion. .Water is shown entering and leaving vessel 15 at convenient positions.

A certain amount of media 31 will be entrained in byproduct gas 10, resulting in a gradual lowering of surface 20. This media loss may be replaced in any convenient manner; I prefer to introduce such salt via halide 6; that is, by entraining the solid, powdered media in the stream of halide entering column 31 via ports 16.

The vessel of Figure l is designed for the continuous cracking and ingotting of the halide/metal under consideration; however, this vessel may be modified so as to permit the process to operate on a batch basis. Thus, by maintaining shape 18 in a fixed relationship to vessel 15, and then raising electrode 19 as surface 24 rises, an ingot may be built up within vessel 15, said shape being periodically removed from vessel 15, as required. In this case, however, allowance must be made in the design of the vessel for the fact that, as surface 24 rises, surface will rise also.

While I have shown my process being carried out within a vessel that is completely water-jacketed, other arrangements are of course feasible. Thus, vessel 15 may be lined with carbon above surface 24, and, in this arrangement, the heat may be supplied to the process by inductively heating said carbon lining with high frequency current induced by water-cooled coils (not shown) that are wound around, but out of contact'with, said carbon lining. V

Figure 2 shows one method of constructing electrode 19. Here, cooling water-5 is caused to flow down along outer shell 25 and is then caused to cool refractory pipe 27 by flowing up along said pipe as said water leaves the electrode. Solidified layer 28 of media 31 acts to protect the electrode from theaction of the halogen being released within the vessel. While Figure 1 shows halide 6 being introduced into column 31 via ports 16, halide 6 may also be introduced into the process via electrode 19. Figure 2 illustrates how this can be handled. Refractory pipe 27 terminates at opening 29 in non-consumable tip 21.

Figure 3 shows another method of constructing an electrode for introducing heat and halide 6 into the process. Here, tip 21 is shown fastened to electrode pipe 25, said pipe 25 being composed of a metal with a high thermal conductivity, such as copper. The cooling action is supplied to this electrode by halide 6.

Figure 4 illustrates a method of constructing ports 16 when it is desired to introduce a plurality of halides into the process, such as when the process is producing an alloy by cracking a plurality of halides;

Figure 5 illustrates several ways in which the vessel may be modified in other manners; such as, for example, when a metal is being produced that has little or no action of consequence on obtainable non-metallic refractories, or when it is desired to greatly increase the rate of production. In Figure 5, shell '33 is shown retaining insulation 34 and refractory lining 37. Lining 37 is shown retaining media column 35, said column 35 being caused to float on molten metal pool B, surface C being the slag-metal interface lying therebetween. The funnel shape of column 35 is designed to allow for the rapid expansion of halide 6 as it rises up thru said column 35, so as to prevent the gaseous halide from pumping media 35 out of the vessel. Space 42 above surface 36 is designed to encourage the disengagement of entrained media 35 from byproduct gas 10.

Electrode 39 of Figure 5 is arranged somewhat differently than electrode 19 of Figure 1. However, the action of electrode 39 is similar to the action of electrode 19. Being insulated from shell 33 by insulators 38, electrode 39 carries the current required to spring an are between electrode 39 and surface C, and said electrode 39 may also be caused to carry halide 6 into the process, as shown in Figures 2 and 3.

The casting arrangement of Figure 5 consists of mold 40 that is cooled by water layer 41, said layer 41 being retained by shell 45.

Neither Figure 1 nor Figure 5 show the means for withdrawing shape 18 from the water-jacketed molds; however, such means are well known to the art. Shape 18 may be withdrawn by means of rolls which grip shape is and are revolved in accordance with the desired rate of shape 18 withdrawal, or said shape 18 may be grasped by mechanical or hydraulic means which are regulated to lower shape 18 in accordance with the desired rate of withdrawal.

With few exceptions, when a metal is solidified within a water-cooled mold, the skin that is first formed draws away from the mold if the normal contraction of the metal is not restricted. If, for one reason or another, this normal contraction is prevented, the newly-formed skin is distorted as the solidified metal attempts to contract, and, due to the fact that the newly-formed skin is very weak and very brittle, this distortion leads immediately to tears, which, in some cases, are immediately healed as molten metal within the skin bleeds out thru the tear, and in other cases remain in the surface of the casting to initiate cracks when the cast shape is subjected to deformation in the rolling mill. My copending application, Serial No. 135,683 of December 29, 1949, now abandoned, disclosed my method of overcoming these surface tears. in Figure l, packing means 52 closes off the shrink gap lying between shape 18 and the wall of mold 26. With the shrink gap closed, l maintain a gas pressure therein by means of tube 51 that penetrates the water-jacket and mold wall to permit gas A to be introduced in accordance with the pressure readings shown by gage 5% in general, the pressure of gas A within the shrink gap is maintained at the level required to ofiset the ferro-static head exerted on the newly-formed skin by column 31 and pool 22. With the pressure balanced on both sides of the newly-formed skin, said skin acts in accordance with its normal desires when contraction forces arise during the casting process. In Figure 5, packing 46 seals off the shrink gap between shape 18 and mold 40, while tube 44 penetrates the water-jacket and mold to permit gas G to enter the shrink gap in accordance with the pressure registered on gage 43.

The process herein disclosed requires a large amount of heat, the main part of which leaves the process in the form of sensible heat in byproduct gas 10. If the value of shape 18 justifies this waste of heat the matter is not of great consequence; however, if heat is a serious factor in the case, then at least a portion of the sensible heat in byproduct gas 1% must be returned to the process with halide 6. Due to the highly active character of chlorine, bromine and iodine, and due to the entrained media 31 or 35, gas It cannot exchange its heat content with incoming halide 6 within a conventional heat exchanger of the recuperating type. A reversing system of at least two regenerators is required. While such regenerators may be constructed of a number of materials, I prefer graphite. it halide 6 is to be highly preheated for use in the appa-v ratus of Figure l, the entry ports 16 must be modified in order to protect them against the erosive action of hot halide 6, such modification being carried out along the lines shown in Figure 5.

The apparatus of Figure 1 has particular advantage in the production of those metals whose affinity for oxygen bars the use of oxygen-bearing refractories. Metals such as ductile titanium and ductile zirconium are examples. However, when oxygen contamination can be tolerated in titanium and zirconium, or when the process involves a metal that does not dissolve oxygen, or is notseriously harmed by oxygen so dissolved, a refractory lined vessel of the type shown in Figure 5 has certain advantages, especially if the cost of heat is an important factor. Iron, silicon, chromium, vanadium, etc. are examples of such metals.

The composition of the dissociation media will depend 6 upon the metal being produced and the halide being cracked. The dissociation media must be composed of salt(s) that are non-reactive, in general, with the metal being produced, the halide being cracked, and the byproduct halogen/halide being produced. There is no reason why a moderate solubility of a halide 6 or a byproduct halide in the molten media cannot be tolerated, so long as the presence of the dissolved halide does not interfere seriously with the process herein contemplated. in other words, an inert media may be inert by reason of its saturation with a halide of the metal being produced.

i prefer a substantially non-volatile media; for, while volatilization of the media may be tolerated if desired or if necessary, such volatilization introduces complications into the process. Typical of the compounds which may be employed as, or in, a media in which a halide of titanium or zirconium may be cracked, together with their estimated boiling points, are: calcium chloride (2000 C.), calcium fluoride (2500 (1.), magnesium fluoride (2200 C.).

While practically all compounds employable as dissociation media will to at least a certain extent volatilize in the immediate vicinity of the are that plays between the electrode and the surface of the molten metal bath, such volatilization is of no serious concern if the upper reaches of the media column are at a temperature that lies below the boiling point of said media and the compounds which comprise it, for as the volatilized media gas passes up thru the cooler reaches of the media column, said media gas will condense. In the apparatus of Figure l, the cooling action of the water-jacket in the upper reaches of the media column will assure the rather complete condensation of media volatilized near the arc. in the apparatus of Figure 5, however, such condensation may need encouragement if relatively-low boiling point compounds are present in the media. Such encouragement may be provided by cooling moderately the upper reaches of the media column by burying cooling pipes in refractory lining 37, or by burying cooling pipes in lining 37 above surface 36. The importance of condensing volatilized compounds within the cracking-ingotting vessel is stressed by the fact that such volatilized compounds carry a very large amount of heat, and, unless this heat is released before the compound gets out of the apparatus, the heat requirements of the process become very large indeed.

The level of surface 2% or surface 36 must be maintained at a relatively-fixed position within the vessel. As mentioned previously, media lost by entrainment in gas 10 may be replaced by introducing the powdered media into the stream of halide 6 that is entering the media column, or the lost media may be replaced by providing the vessel with an opening above said surface, so that media may be introduced, as required. The level of surfaces 20 and 36 may be Watched by means of a radioactive source 55, such as cobalt isotope 6t), and a counter 56 capable of measuring the radiation of said source 55 thru the walls of the cracking-ingotting vessel.

While I have shown the use of non-consumable electrodes, carbon or graphite electrodes may be employed, if desired. Due to the fact that carbon and graphite electrodes are normally consumed as the operation proceeds, said electrodes must be lowered towards surface 24 or C as fast as said consumption takes place. Means are conventionally available for maintaining the gap between an electrode and a conducting surface; however, unless the level of said surface is known and held at a substantially-fixed position within the vessel, such means would merely retreat the electrode as the metal level rises. Accordingly, in this case, the position of the metal surface must be watched with the aid of a radioactive source and a counter, so that said level may be held constant by lowering shape 18 as the precipitated metal accumulates within the metal pool, and so that the electrode raising and lowering means have a relatively-fixed surface against which to work.

While the metal particles suspended in the molten media that circulates past surfaces 24 and C will readily consolidate with the pool of metal if they contact it, such contact must be attained before my process yields the desired product. Contact between the metal particles and the surface of the molten pool is encouraged by the growing size of said particles and by the abrupt reversal in the direction of flow of the molten liquid. as it flows to and then away from said surface. In order to induce a more positive movement of said particles towards, and into contact with, the molten metal pool lying at the base of the media column, I prefer to arrange my process so that said metal particles possess an electrical charge whose sign is opposite from the charge possessed by surface C or surface 24.

It has been experimentally shown that titanium particles which are dispersed within a calcium flouride media according to my process are charged negatively. possibly due to their colloidal size or possibly due to the ionization of the halogen gas that is flowing up thru the media; that is, such particles are repelled by the negative pole and are attracted to the positive pole. Advantage may be taken of an electrical charge possessed by the dispersed metal particles by causing pool B or pool 22 to possess the opposite charge, and so cause said pools to attract said dispersed particles. For example, in the mentioned titanium case, pool B may be grounded, while lining 37 and electrode 39 are insulated from pool B, and the insulated lining 37 and/ or electrode 39 may be connected to a negative-charge generator-that is, the negative pole of a generator of directional current.

In my preferred embodiment, I take a positive attitude towards the charge on the dispersed metal particles by causing said particles to pass thru an electrical field wherein a negative charge on the dispersed particles is assured. In the apparatus of Figure 5, this may be assured by connecting electrode 39 to the negative pole and pool B to the positive pole, or it may be assured by connecting charge plate 57 to the negative pole of a source of high voltage current and pool B to the positive pole. The impression of the field between plate 57 and pool B is made the easier, the closer plate 57 is to pool B. If lining 37 consists of carbon, the insulation of said lining from the ground, particularly from electrode 39, shell 33 and water-jacket 40, will permit the carbon lining to act as the charge plate, an advantageous situation when contact between the carbon lining and the dispersed particles will lead to difiiculties, as in the case of titanium and zirconium, for then said carbon lining will repel the dispersed particles, thus minimizing the carbon pick-up.

The apparatus and process of Figure 1 may be modified so that the dispersed particles are charged negatively by an external source. Such particle electric charges may be impressed by breaking the electrical connection between pool 22 and that portion of vessel that lies above surface 24. This may be done by forming the apparatus in two sections, joined together in the vicinity of pipes 16 or between surface 24 and pipes 16. The two sections must be separated by insulation, such as a non-metallic refractory, at their junction, the process depending upon this insulation plus the ever-present layer of solidified salt for the required dielectric separation between the sections. In view of the fact that the entire inner surface of vessel 15 is covered with a layer of solidified salt. electrode 19 being similarly covered if it is water-cooled, special means may be inserted in media 31 to form the charge plate for the apparatus. This may be done by suspending within media 31 a carbon plate or rod that is connected to the negative pole of the power source, pool 22 being connected to the positive pole. This charge plate should be suspended as close to surface 24 as operating conditions permit; should be arranged so that a layer of solidified salt does not form thereupon; and may be electrically connected to the upper section of vessel 15. A carbon rod that passes thru the water-jacket and out into column 31 will perform these functions.

While it is my intention that a high voltage be impressed between the charge plate and the pool of molten metal, say, between 10,000 and 100,000 volts, lower voltages will be efiective if the impressed voltage is high enough to overcome the resistance of the molten media lying between the charge plate and the pool of metal. Thus, the attraction of the dispersed particles to the pool of metal is effective, according to my process, whether cataphoresis, involving relatively-low voltages, or the Cottrell effect, involving relatively high voltages, is the effective process in operation.

While my process is particularly concerned with the thermal dissociation of the chlorides, bromides and iodides of titanium and zirconium, compounds involving these halogens and other metals may be cracked and ingotted according to my process. Thus, iron, silicon, chromium, vanadium, etc. may be produced. In order to prove effective in my process, the metal halide must crack to yield said metal at a temperature that lies between the melting point and the boiling point of said metal. The chlorides, bromides and iodides of titanium, zirconium, iron, silicon, chromium and vanadium all crack to yield the respective metals within said temperature range. While the iodides of these metals crack to yield the respective metals at a temperature just above the melting points of the respective metals, the chlorides and bromides of these metals require higher temperatures in order to obtain a satisfactory yield. I recommend that the media column be maintained at about 3500 F. when the iodides of these metals are being cracked, and at about 3900 F. when the chlorides or bromides of these metals are being cracked.

A large number of the metals producible by my process exhibit several stages of oxidation. Titanium, zirconium, vanadium and chromium being examples. The dissociation temperature of the various oxidation stages varies, depending upon the halide involved, the gas pressure at the site of decomposition, and the partial pressure of the halogen involved. Actually, these dissociation reactions are quite complicated, and, due to the fact that decomposition involves compound phases which may exist only at the high temperatures here involved, the actual dissociation mechanisms are ditficult to predict. Thus, when the tetrahalides of titanium or zirconium are introduced into my media column, for example, contact between the tetrahalide and the dispersed metal particles will result in the formation of lower halides, the dihalides, for example. Accordingly, the thermal dissociation of these tetrahalides involves the decomposition of both the tetrahalide and the dihalide. Furthermore, there is reason to believe that the trihalides may be a phase of the process, and there are those who believe that the monohalides are involved. As far as I know, there is no precise decomposition temperature for any of these various oxidation phases; rather, it is my belief that they coexist in a relative relationship that depends upon the temperature and pressure. In any case, the dissociation of these halides is a reversible process wherein equilibrium depends upon strictly local circumstances. The temperature-pressure situation within my process is very complicated, due to the fact that local pressure varies according to the depth below surface 20; that is, the static head of the media column has a definite bearing on equilibrium.

The preferred method of approaching my process, in any given case, is to substantially ignore these theoretical complications. Thus, While I have recommended certain media temperatures, the best method of placing my process in operation is to insert an electrode similar in shape to electrode 18 made of carbon within vessel 26, the upper end of said electrode coming substantially at the position whereat it is desired to maintain surface 24. The selected media compound being then introduced into the 9 apparatus until it reaches the level of surface 20. The arc is then struck between electrode 19 and the top end of the carbon electrode, and the media salt is melted further additions of the media salt will be required to lift.

the media column to level 20. With the media melted, the input of heat via the arc should be continued until the temperature of column 31 is just below the boiling point of said media-it being realized that, during this melting and superheating process, an inert gas, such as helium, argon, C0, C02, etc., must be continuously introduced thru ports 16, so as to prevent the molten media from entering said ports 16. With the media heated to just below its boiling point, halide 6 may be introduced to start up my process. It will be realized that the carbon electrode will probably contaminate the metal that solidifies at the base of the media column during the early phases of the start-up; however, this will cease as the said carbon electrode gradually changes to the cast shape.

With the process operating at its maximum temperaturethat is, just below the boiling point of the media,

the temperature of media 31 may be gradually lowered to the optimum operating temperature for the apparatus and halide(s) involved, it being realized that the optimum temperature is the minimum temperature at which the desired results are obtained. The temperature of media column 31 may be lowered by decreasing the heat input via the are, or by increasing the input of halide 6. While substantially all of the chlorides, bromides and iodides of the metals: titanium, zirconium, chromium, vanadium and silicon are dissociated at the starting-up temperatures suggested below, such essentially complete decomposition may not be necessary or desired. In any case, the completeness of any given dissociation may be followed, as the temperature is lowered, by analyzing byproduct gas for its content of the subject metal.

For use with titanium, zirconium, chromium, vanadium and iron, I recommend a media composed of calcium fluoride. For use with silicon, I recommend magnesium fluoride. If the presence of oxygen is permissible, I recommend a media composed of a basic calcium silicate (2700 C.) be used with titanium, zirconium, iron, chromium and vanadium.

The cast shape produced by my process may be round, square, oval, rectangular, or any other convenient or desired shape, and the shape and size of the cross section of the molten column of media may be substantially the same as, or substantially different from, the shape and size of the metal shape being produced. Figures 1 and 5 may be picturing the cross section thru a rectangular cast shape and media column, in which case a series of electrodes may be employed to maintain the operating temperature, or said figures may be picturing a column of media and a cast shape that is square, round, or oval.

The halide being cracked by my process may be mixture of halides of the same or difierent metals. An alloy consisting of a plurality of metals may be produced by my process by cracking a mixture of the halides of said metals; that is, for example, titanium and iron iodides, chlorides or bromides may be simultaneously cracked to yield a titanium-iron alloy, the composition of said alloy being controlled by controlling the composition of the mixture of titanium and iron halides. Other halide combinations capable of being simultaneously cracked to yield useful alloys will occur to those skilled in the art. An alloy of the metal produced by cracking a halide of said metal may be formed with other elements by feeding said other elements into the molten media so that said other elements will sink thru said media to contact and alloy with the metal pool lying at the base of the media column. Thus, coarse chromium particles may be fed into the media column to produce an alloy of chromium with a titanium metal being produced according to my cracking-ingotting process. In the latter case, the wall of the cracking vessel is penetrated above the surface of the media columnthat is, surface or surface 36-by an 10 alloy feeding mechanismthat permits the alloying element(s) to be fed into the media at the desired rate.

Halide 6 of my process may be a gas, liquid or powdered solid. Convenient devices are available to meter liquids, gases and powdered solids into a process of this type. My preference is for the arrangement wherein the heat content of byproduct 10 is employed to preheat halide 6. If a powdered solid is to be fed into the media column, said powdered solid is preferably entrained in a liquid or gas stream. For example, gaseous or liquid titanium tetrachloride (boiling point 136 C.) may carry powdered titanium dichloride (boiling point 1477 C.) into the media column. Similarly, inert gases, such as helium or argon, may be employed to carry powdered halides into the media column.

Having now disclosed several forms of my process and apparatus, I wish it to be understood that my invention is not to be limited to the specific form or arrangement of steps hereinbefore shown and disclosed, or specifically claimed in my claims.

I claim as my invention:

1. In the process wherein a metal selected from the group consisting of titanium and zirconium is produced by thermally dissociating a halide selected from the group consisting of a chloride, a bromide and an iodide of said metal to yield said metal and a byproduct gas, the improvement, which comprises: heating a compound selected from the group consisting of calcium chloride, calcium fluoride, magnesium fluoride, and mixtures thereof to form a molten pool of said compound, said compound having a boiling point that lies above the melting point of said metal, said molten pool of said compound being heated to a temperature that lies in the range between above the melting point of said metal and the boiling point of said compound; introducing said halide into and dispersing said halide within said molten pool of said compound, said halide being selected from said group of halides for its capacity to dissociate thermally within said molten pool of said compound within said temperature range, so that said dispersed halide dissociates thermally within said molten pool of said compound to yield particles of said metal dispersed within and collected by said molten pool of said compound and a byproduct gas that separates from said collected metal particles by rising to and being withdrawn from the surface of said molten pool of said compound; and floating said molten pool of said compound containing said dispersed metal particles on a molten pool of said metal, said molten pool of said metal being formed as said metal particles settle through said molten pool of said compound to collect and coalesce at the base of said molten pool of said compound.

2. The process according to claim 1 in which said molten pool of said metal is contained within a continuous mold and in which a portion of said molten pool of said metal is continuously solidified.

3. The process according to claim 1 in which the metal of said molten pool of said metal is an alloy of said metal with at least one other element and in which said alloy is formed by the substantially simultaneous thermal dissociation within said molten pool of said compound of a halide mixture consisting of a halide of said metal and a halide of said other element.

4. The process according to claim 1 in which the metal of said molten pool of said metal is an alloy of said metal with at least one other element and in which said other element is produced separately from said process and is then fed into said molten pool of said metal.

References Cited in the file of this patent UNITED STATES PATENTS 1,046,043 Weintraub Dec. 3, 1912 1,306,568 Weintraub June 10, 1919 1,671,213 Van Arkel et al. May 29, 1928 (Other references on following page) 11 UNITED STATES PATENTS Wempe Aug. 24, 1937 Freudenberg Feb. 21, 1939 Kroll June 25, 1940 Herres Feb. 13, 1951 Maddex Aug. 14, 1951 Winter Feb. 19, 1952 Winter Aug. 19, 1952 12 OTHER REFERENCES Handbook of Chemistry and Physics, 28th edition, 194 published by Chemical Rubber Publishing Co., Cleveland, Ohio, pages 348, 349, 406, 407.

Metal Industry, Oct. 18, 1946, pp. 319 to 322.

Metal Progress, Feb. 1949, pp. 193 to 194.

Journal of Metals, April 1950, pp. 634 to 649

Claims (1)

1. IN THE PROCESS WHEREIN A METAL SELECTED FROM THE GROUP CONSISTING OF TITANIUM AND ZIRCONIUM IS PRODUCED BY THERMALLY DISSOCIATING A HALIDE SELECTED FROM THE GROUP CONSISTING OF A CHLORIDE, A BROMIDE AND AN IODIDE OF SAID METAL TO YIELD SAID METAL AND A BYPRODUCT GAS, THE IMPROVEMENT, WHICH COMPRISES: HEATING A COMPOUND SELECTED FROM THE GROUP CONSISTING OF CALCIUM CHLORIDE, CALCIUM FLUORIDE, MAGNESIUM FLUORIDE, AND MIXTURES THEREOF TO FORM A MOLTEN POOL OF SAID COMPOUND, SAID COMPOUND HAVING A BOILING POINT THAT LIES ABOVE THE MELTING POINT OF SAID METAL, SAID MOLTEN POOL OF SAID COMPOUND BEING HEATED TO A TEMPERATURE THAT LIES IN THE RANGE BETWEEN ABOVE THE MELTING POINT OF SAID METAL AND THE BOILING POINT OF SAID COMPOUND; INTRODUCING SAID HALIDE INTO SAID DISPERSING SAID HALIDE WITHIN SAID MOLTEN POOL OF SAID COMPOUND, SAID HALIDE BEING SELECTED FROM SAID GROUP OF HALIDES FOR ITS CAPACITY TO DISSOCIATE THERMALLY WITHIN SAID MOLTEN POOL OF SAID COMPOUND WITHIN SAID TEMPERATURE RANGE, SO THAT SAID DISPERSED HALIDE DISSOCIATES THERMALLY WITHIN SAID MOLTEN POOL OF SAID COMPOUND TO YIELD PARTICLES OF SAID METAL DISPERSED WITHIN AND COLLECTED BY SAID MOLTEN POOL OF SAID COMPOUND AND A BYPRODUCT GAS THAT SEPARATES FROM SAID COLLECTED METAL PARTICLES BY RISING TO AND BEING WITHDRAWN FROM THE SURFACE OF SAID MOLTEN POOL OF SAID COMPOUND; AND FLOATING SAID MOLTEN POOL OF SAID COMPOUND CONTAINING SAID DISPERSED METAL PARTICLES ON A MOLTEN POOL OF SAID METAL, SAID MOLTEN POOL OF SAID METAL BEING FORMED AS SAID METAL PARTICLES SETTLE THROUGH SAID MOLTEN POOL OF SAID COMPOUND TO COLLECT AND COALESCE AT THE BASE OF SAID MOLTEN POOL OF SAID COMPOUND.

Images