VIEWS: 105 PAGES: 16 POSTED ON: 3/27/2011
The Melting and Refining of Magnesium BY C. E. NELSON,* A.I.M.E. MEMBER (New York Meeting. February 1944) THE purpose of this discussion is to out- during melting and a molten pool of flux line briefly the practices commonly followed into which the solid magnesium melts. I t in this country for the melting and refining is stirred through the molten metal bath of magnesium and its alloys. The processes and agglomerates oxide or similar foreign used for the various forms of primary mag- bodies; then on quiet standing separates nesium, as far as there are differences in away, leaving the refined metal ball float- the physical shape or behavior, will be ing in an encircling layer of molten flux. discussed. The refining of general fine scrap I t forms only a thin fluid iilm over the or secondary magnesium was presented in surface of the molten metal, which may be an earlier paper.' parted for hand-ladling processes and tends Inasmuch as the use of fluxes is an essen- to cover the metal again after the ladle is tial part of all the melting and refining removed. A very light dusting of the pot processes, the principal aim of this paper surface with fresh flux immediately after will be to deal with these in sufficient detail the ladle is removed is usually desirable. to point out their unique characteristics, The open-pot method is used generally in order to make their use more effective. in the following processes: (I) alloying and secondary smelting in the production of ingot, (2) in sand foundries for premelting All melting and refining processes for and to a smaller extent for the production magnesium and its alloys require the use of small castings requiring hand ladling, of fluxes. These fluxes have a magnesium (3) in permanent-mold founding for pre- chloride base and other halide salts or melting and also direct ladling to castings, oxides are added to give a density or be- (4) for continuous methods of preparing havior exactly suited to the particular metal in the productionof billets or ingots melting practice. The successful handling from which wrought products are fabri- of magnesium depends upon the proper use cated, ( 5 ) in general scrap recovery. of the correct fluxes. - There are four general methods of melt- Crucible Process ing, summarized in the following paragraphs and treated in more detail in the section The crucible process makes use of No. on Melting and Refining. f 310 flux, which has the property b being Open-pot Method thinly fluid at the start, to provide protec- The open-pot method makes use of tion and refining qualities, and then drying No. 230 flux.2 The flux provides protection out or thickening after a time, leaving a protecting crust on the pot surface until Manuscript received a t the office of the Institute Oct. 22, 1943. Issued as T.P.1708 in the time of casting. At that time it can be METALSTECHNBLOGY. August 1944. readily skimmed off or held back, thus * Metallurgical Department, The Dow Chemical Co.. Midland. Michigan. allowing the contents of the crucible to be References are at the end of the paper. poured out directly into castings without All flux numbers refer to the designations of the Dow Chemical Co. See Table I, page 395. fear of flux contamination. This flux is not 6. E. NELSON 393 self-healing in its film protection, hence is silicon carbide refractory having a carbon not suitable for hand-ladling operations. facing next to the metal bath. The walls This method is used principally in the and roof are of high-grade firebrick. following processes: (I) sand and per- In operation, the solid charge is fed me- manent-mold founding as a single-step chanically into one end of the furnace onto melting, rehing and pouring operation, a preheating shelf, from which it is pushed or more commonly as a second step follow- into the bath by the introduction of the ing the transfer to a casting crucible of next charge. Metal is dipped or pumped molten metal from premelting units such out of a well at the opposite end of the as large open pots or tilting crucibles; (2) furnace. crucible processes in the batch method of The consumption of flux for melting melting, alloying, and pouring billets or magnesium in these units has been the ingots from which wrought products are same as previously was required for stand- fabricated; (3) the refining of die-casting ard open-pot operation. Similarly, the scrap. melting loss closely parallels that obtained in the more conventional methods of melt- Reverberatory Process ing. Fuel efficiency is much better and The reverberatory process is a new would amount to about two thirds of that development for melting magnesium. I t used in open-pot melting. Undoubtedly, has been stated by Beckz° that hearth the outstanding feature of the open-hearth furnaces, whether gas or electrically heated, operation is the very high melting capacity. are unsuitable for the melting of magnesium The units in operation are capable of alloys, owing to the large surface of the melting continuously at the rate of 4000 lb. melt as well as to the contact between the of metal per hour. gases and the metal or flux. Some confi- . Safety of operation is another notable dence was obtained in the possibility of feature, since the charging and preheating melting magnesium by direct radiation as is all automatic and spatterings cannot a result of the successful experiments by reach the operators. Furthermore, the the Dow company in melting magnesium likelihood of a runout is remote, because in the resistor-heated type of Detroit the refractory construction is cold on the electric rocking furnace. With this as a outside. background, a reverberatory hearth fur- The field of use of the reverberatory is nace having a capacity of 1500 Ib. of mag- in large-scale continuous melting of mag- nesium was built for experimental work. nesium or alloy ingot or heavy scrap. In This unit was operated successfully using such a process, it operates as a premelter oil and gas as fuels. for alloying or foundry operations. I t is Based on these preliminary experiments, not considered suitable for the melting of a furnace with a capacity of approximately fine magnesium or light scrap, because of 12,000 to 20,000 lb. of magnesium was put the difficulty of applying adequate me- into use. This furnace has operated for chanical puddling and the development about two years and during that time six of the large amounts of sludge or dross more similar furnaces have been placed in that accompany the processing of these service. materials. Protection during melting and Below the metal line these furnaces are holding is provided by 230 or similar types lined with a refractory that is relatively of open-pot fluxes. In this method, the nonreactive with the fluxes and the mag- ultimate refining of the metal is largely nesium. The linings with the longest service obtained in the open-pot or crucible process record are of Tercod, a carbon-bonded that follows. 394 THE MELTING AND REFINING OF MAGNESIUM Die-casting Process sing of this material that are different from the standard practices discussed later. The die-casting process makes use of Crystalline magnesium is finely divided No. zzo flux, which gives no surface pro- tection but is used only for refining the and hence oxidizes readily. Furthermore, metal. Surface protection in this method a variable and sometimes appreciable comes from the use of a sulphur dioxide amount of oxide, nitrides, and other impuri- atmosphere, which is maintained in a ties are occluded with the crystals. Because closed dome over the pot. Only a very little of these, more care must be exercised dur- flux is used as protection during melting ing the melting of the crystals to use a down, and this is stirred through the fluxing technique that will not permit metal thoroughly for refining. After a fcw oxidation. Normally this is accomplished moments of quiet standing, this flux, by charging the crystals into a "heel" of together with agglomerated oxides and molten metal purposely left over from a dross, sinks to the bottom of the pot and previous batch, or into a bath of melted flux is removed with the sludge. in the bottom of the pot. Liberal quantities of the same flux are dusted over the solid charge and as often as required to prevent TYPESOF MAGNESIUM the start of oxidation. The crystals are puddled into the metal or flux bath as The two general types of primary mag- rapidly as possible. After the chargc has nesium that will be considered here are the all melted down, more flux is added and well-known ingot made by electrolytic stirred thoroughly through the bath, in reduction processes and the crystalline order to separate out the oxide and dross condensed product from the newly intro- f inclusions. I the crystals are particularly duced reduction processes such as the ferro- . clean and free from oxide and other un- silicon-dolomite reaction. desirable materials, it may be sufficient to proceed with the normal alloying processes, Ingots followed by stirring of the fluxes through the bath and finally settling out of the dross Little special description need be given and sludge and pouring of the refined alloy to the magnesium ingot, since this product into ingots or castings. is a solid ingot that is already in a well- If the crystals contain a large amount of refined state through having been dipped oxide, or appreciable oxidation takes place from a fused salt or flux bath. The only during the melting, it is desirable to allow significant impurity in this metal is iron, a few minutes quiet settling and then dip which normally is present to saturation from the bottom of the pot as much of the (0.030 per cent iron) a t the temperature of dross and oxide as possible before proceed- the electrolytic cell. The melting, alloying, ing to the alloying and refining steps. The and reiining of this ingot presents no more decision as to whether the sludging out difficulty than the routine melting of should be done before alloying, refining, magnesium-alloy ingot. and pouring would be made by the skilled operator, who can feel how much sludge is Cvysia22ine Magnesium in the pot and knows from experience how much may be present and still permit The melting and refining of the crystal- proper refining and production of clean line magnesium is somewhat difficult, and ingot. I t has been noted that, for certain for this reason the following discussion will grades of crystals and scrap, refining may attempt to point out features in the proces- be facilitated by the addition to the 230 C.;:F. NELSON 395 flux of certain agents such as fluorspar, or later stage wherever the production of salts such as sodium or potassium chloride. high-purity alloys is warranted. For exam- The magnesium crystals usually are ple, almost all the wrought magnesium relatively low in iron content but may alloys, except the Mg-Mn type, now in contain sodium, potassium, calcium, and use in the United States are of high purity sometimes silicon as impurities. For all made possible by precipitation of the iron, practical purposes, all except the silicon are which will be described later. I.-Com#osition TABLE and Characterristics of Fluxes joy Me1tirt.g and Re$ning Magnesium, Dow Chamical Comfialty Approxi- of Flux EX Composition Use Characteristics and Remarks Pet Cent - - 220 57.0 KC1 Die casting o 1-3 A heavy bottom flux used for refining metal 28. o CaClz in a covered pot provided with SOa surface 12.5 BaClz rotection. P1u.x I? removcd after relining. 2.5 CaFl bsed where dipping i s so frequent that open-pot operatlon is impractical 230 5 5 . 0 KC1 Sand and permanent 10-20 4-6 Characterized by high fluidity of surface film 34. o MgCll mold casting protection, allowing parting and recovering 9.0 BaClz Premeltin 5-10 for ladling operations. High refining quall- 2 . o CaFz Alloy profucttion Scrap recovery 10-20 10-20 I :$ 4-15 ties. General open-pol flux Flux pots 100 <0.5 250 23.0 KC1 Alloying o Variable Used for introducing manganese into alloys 72.o MnClz made by open-pot process with 230 flux. a.5 BaC12 Reaction products approximate 230 flux a .5 Cap? behavior. Flux contains 31.7 per cenl manganese by weight 310 20 . O KC1 Sand and permanent- 1-3 3 Crucible-type flux characterized by being g o . o MgClz mold foundry fluid at start for melting and refimng. I S . o CaFz Crucible alloying 2-5 2-5 then drying out t o crust that ca? be re- I S . 0 MgO Die-cast scrap refin- 2-5 2-5 t moved or held back for d ~ r e c pouring ing 320 7 6 . 0 MnClz Crucible alloying o Variable Used for introducing manganese into ~ U O Y S 1 3 . o CaFr firoduced by the crucible process with 310 11 o MgO ux. Reaction products approximate 310 flux behavior. Flux contains 33.5 per cent manganese by weight. Based on weight of metal poured from a given operation. removed by reaction with the flux in the The methods of handling distilled melting and refining process. The iron magnesium crystals are the same as those content, on the other hand, will tend to outlined briefly in the prcceding para- increase up to the saturation value if this graphs, except that normally it is not metal is melted,as is the common practice, necessary to remove sludge before alloying in steel pots. The alloying metals normally and refining. added to commercial magnesium contain many times more impurities, such as iron, than are, foui~deven in the electrolytic magnesium, so that the natural purity of In order to give a clearer understanding the crystals produced by the ferrosilicon of the types and uses of fluxes commonly process is lost unless special 'high-purity employed for the 'melting and refining of metals are used and the whole process is magnesium, a list of compositions is carried out in iron-free fluxes and melting included in Table I. equipment. In practice, therefore, it is more The compositions and designations feasible to remove the impurities a t the shown in this paper are those of The Dow 396 THE NELTING AND REPINING OF MAGNESIUM Chemical Co., but it should be pointed out nary boiler-plate construction, madc up in that within the last year or two other convenient sizes and shapes to permit easy suppliers have put fluxes on the market. removal of the frozen cake. Among these new suppliers are Basic 7. Thermocouples may be either Magnesium, Inc., and Permanente Metals Chromel-Alumel or iron-constantan and Corporation. should be protected in the pots by a steel pipe or sheath. 8. Separate flzlx pots are essential to the production of high-quality magnesium- alloy ingots or castings using the open-pot No detailed information on equipment ladling process. These pots usually are will be attempted here. The equipment of the round-bottom cast-steel type having most used may be indicated in the follow- capacities of 400 to 600 Ib. of magnesium. ing outline: No. 2 3 0 flux is used to fill the pots, and all I. Open pots may vary in capacity from ladles, tools, and other utensils are washed IOO to 4000 Ib. of magnesium. They are periodically in this flux. This prevents made of heavy cast steel. Particular care is undue splashing and agitation in the regu- taken to avoid the use of steels containing lar casting pot and prevents the carry-over more than several tenths of a per cent of of excessive oxide or other impurities into nickel, in order to prevent contamination that pot, and thus into castings. Ladles of magnesium alloys melted therein. Cast- should be kept smooth and clean at all iron pots are brittle and often are too times, as dirty ladles cause flux and oxide porous to hold the flux. They tend to grow inclusions in the castings. in use, thus becoming still more porous; hence are not recommended. PREPARATION TKE METAL OF 2 . Crucibles may vary in capacity from FOR MELTING zo to aooo lb. of magnesium. The smaller sizes usually are madc of welded steel and Preparation of the metal is a very irn- the very large sizes of cast steel. Ordinarily portant part of any melting practice, since these are made with extra thick bottom failure to clean the solid metal properly sections to resist corrosion, and con- may lead to excessive gas or dirt in the structed so that this part can be readily melt and introduce harmful metallic replaced by cutting off the old section and impurities. welding on a new one. As with open pots, Aside from the regular precaution of pre- contamination by nickel in the steel must heating all metal that is to be charged be avoided. Metallizing of these pots and directly into molten metal or flux, no crucibles with aluminum significantly in- particular precautions in preparation are creases their service life. necessary for virgin magnesium ingot or 3. Pouring ladles should be constructed crystals. with a skimmer on the backside, for parting Since heavy foundry scrap constitutes the flux, and a special bottom-pouring spout. one of the largest sources of metal being 4. Sludge ladles are simply small hemi- melted, the following points will be of spherical scoops ranging in size from 6 to interest: 9 in. in diameter, depending on pot sizes. Oil or excessive moisture on the surface 5. A skimmer is a handy tool. I t may be of the scrap is almost certain to lead to a flat, perforated plate about 6 to 9 in. in gas in the metal and excessive porosity diameter, attached to a long handle. in the product. These hazards are avoided 6. Sludge and flux pans may be of ordi- usually by care in preventing oily materials C. E. NELSON 397 from coming into contact with the scrap 4. The pot may be entirely filled with and either predrying of the scrap in an solid ingot, providing care is taken to oven, or sandblasting it just before melting. dust thoroughly with the flux all parts of Scrap should be remelted as soon as the charge. Any part of the charge that is possible after the castings are sawed out, added' directly into molten metal or flux in order to prevent undue formation of must be carefully preheated, in order to hydroxide films on the surface. f avoid danger o a blow and spattering ~rjdesirableamounts of silicon may be when such material is added to the pot. introduced into the molten metal by im- 5. The pot should be watched diligently properly cleaned foundry scrap. to make certain that no oxidation or burn- It is important that all material to be ing takes place. Whenever required, a light returned to the melting pot be thoroughly dusting with flux will prevent oxidation. blasted and handled in such a way that no 6. While the charge is melting down, residual sand can be carried back to the the required amount of alloying materials, melting operation. such as aluminum, manganese or zinc, f I the metal becomes gassed, the remedy should be weighed out and placed on top consists in long holding, with stirring, a t of the solid charge, so that it will be pre- low temperatures or, preferably, direct heated and gradually dropped into the chlorine treatment. pot as the supporting charge melts down. 7. As soon as all the materials arc melted, additional quantitie of fresh flux should be spread on the pot surface. I f The various types of melting practice the manganese is added as KO. 2 5 0 flux, it are: open-pot melting and alloying, open- should be spread on the pot surface at this pot refining and pouring, crucible melting time and stirred into the melt. A vigorous and pouring, two-step premelting and stirring with a sludge ladle should be given crucible melting, and die-cast melting. from the bottom to the top of the pot, so as to produce a rolling action of the con- Open-pot Melting and Alloying Practice tents and to give uniform mixing of the Ingot.-The f largest proportion o all batch. This stirring time would normally magnesium-alloy ingot used in the United be from 5 to 10 min. This stirring also States is produced by open-pot melting. serves as a refining treatment to remove The steps involved in converting mag- all oxides and dross from the metal and nesium into alloy ingot may be enumerated put them into the sludge. as follows: 8. At this point, a sample from the I. A clean cast-steel open pot (usually pot is taken for control analysis. between 600 and 4000-lb. capacity) is 9. The batch is allowed to stand quietly heated and a quantity o No. 230 flux f for a t least 10 min., or preferably longer, approgmately equivalent to 10 per cent in order to give the flux and sludge a of the weight of metal to be charged is chance to separate completely from the melted therein. ball of molten metal. 2. Primary magnesium or heavy scrap 10. The metal is now in a refined state; of known composition, after thorough pre- and if the analysis is found to be satis- heating on the edge of the pot, may be factory, the batch is ready for pouring. charged directly into this flux bath. 11. Clean preheated ladles of the re- 3. Additional No. 230 flux should be quired sizes are used to dip the metal from lightly dusted over the charge and pot the pot and pour into ingots. Since the surface. ladle presumably has been cleaned in the 3g8 THE MELTING AND REFINING OF MAGNESIUM flux pot, it will be necessary to rinse off metal be ladled out into ingots, because the excess flux by filling the ladle with of the likelihood of mixing the metal and metal several times and carefully pouring sludge near the bottom of the pot. Usually the metal back into the pot. After this it is more convenient to leave a certain procedure, theladle may be usedrepeatedly small heel of metal in the pot (10per cent), to dip metal from the pot, until it shows thus providing a pool of metal into which oxidation or graininess, indicating that it the next batch may be. charged. is dirty and must be cleaned again in the 15. At this point, the thick sludge, com- flux pot. posed principally of oxide, dross and flux, I t is very important that the operator may be dipped from the bottom of the pot carefully push back the flux film from a and the thinly 5uid flux and metal propor- section of the pot surface before introduc- tion of the sludge be allowed to drain back ing the ladle. The ladle should be allowed from the dipper. The solid residue is placed to fill slowly over the back side. When full, in sludge pans. After this sludge is removed, it should be lifted through the parted sec- the pot is ready for the next batch. tion of the flux and a small amount of the Scrap or Crystalline Magnesium.-As has metal poured back into the pot through the been said before, the main deviations from spout, so as to discharge any small amount the foregoing process in charging scrap of flux that may havc been present in the metal or crystalline magnesium would be spout. The ladle should then be raised in the need for more careful protection from the pot and rested momentarily on against oxidation, more mechanical stirring the side of the pot, while the operator to immerse the charge in the molten bath, flicks a fine dust of flux over the surface, and the possible addition of a sludging so as to prevent oxidation. This is a very operation prior to the alloying and refining important point from the standpoint of steps if it appeared advisable. quality of the metal, and one most often neglected. Flux should ,$ever be placed on Ofen-pot POurifig the pot just before the operator dips but Open-pot refining and pouring finds most should be applied immediately after the use in the open-pot production of perma- ladle is removed. I n all dipping operations, nent-mold and sand castings. The steps care should be taken to avoid sudden and involved for the two processes are so similar unnecessary motions in the pot, since these that they may be discussed together. The lead to mixing of the flux and metal. steps involved are as follows: 12. Care must be taken in the pouring I. Cast-steel pots having capacities of of the ingot to avoid turbulence of metal from 600 to 1000 lb. magnesium are used. flow into the mold, as this leads to oxida- I n starting a batch, the pot is first heated tion and poor quality of the ingot. The and a quantity of 230 flux, equivalent to size of the ladle should be chosen to permit about 10 per cent of the desired metal a small amount of metal to be left in the charge, is added. ladle after the ingot is full. 2. When this 5ux is molten, preheated 13. Protection for the ingot during alloy ingot of the desired composition pouring and solidification is provided by should be charged into the pot. hoods over the molds and a n atmosphere 3. A fine dusting with 230 flux should of SO2 gas. be made over the charge and repeated as 14. With reasonable care almost all of often as required during the meltdown, to the metal in a given batch may be poured prevent oxidation. into ingots without risk of flux inclusions. 4. When the charge is entirely melted, In no case, however, should all of the a quantity of fresh flux should be added NELSON 399 to the pot. After this flux has melted, the and permanent-mold castings. The steps entire contents of the pot should be stirred in the process are as follows: thoroughly to bring the flux into intimate I. Steel cruciblks in the size range of 60 contact with the molten metal. The metal to 550 lb. of magnesium are commonly temperature at this time should be pref- used. f erably in the range o 1300' to 1350°F. 2. The crucible is placed in the setting, 5. After this stirring and refining, the heated, and dusted with 310 flux. pot should stand quietly for a t least 10 min. 3. Solid ingot of the correct composition before any metal is dipped out. and a portion of casting scrap are charged 6. The pot and contents should next be into the crucible. to heated to about 1450~ 1500°F. and held 4. No. 310 flux is carefully sprinkled at this temperature during the pouring over the entire charge. cycle. Since the metal normally stands a t 5. After the charge is melted down and this temperature for an appreciable length raised to a temperature of approximately f o time, a considerable grain-refining action 1350°F., a quantity of fresh 310 flux is thus obtained. (approximately 1.5 per cent) is added and 7. Small sand castings may then be allowed to melt, after which it is vigorously poured by ladling directly from this pot stirred through the metal. This refining at about 145oOF In permanent-mold cast- operation should take about one minute ing, it is often necessary to pour at con- and serves to agglomerate the dross so that siderably lower temperatures; and for such part of it rises to the surface of the metal cases the pot temperature would be and the remainder goes to the bottom intb dropped down to the appropriate 1eQel the sludge. before pouring. The details of the ladling 6. The clinker-like dross that rises to the process would be the same as previously surface should be carefully skimmed off given. and fresh 310 flux immediately dusted In both sand and permanent-mold cast- behind the skimmer, to prevent reoxida- ing from an open pot, it is common practice f tion. I there is any question as to the to maintain a constant level in the pouring quality of the refining treatment, a little pot through the continuous addition, as more flux should be added, the stirring required, of alloy ingot of the same batch operation repeated and the melt again or of hot casting scrap that came from the skimmed. The absence of substantial same melt. The large quantity of residual quantities of dross on the second skimming flux present in this type of operation seems indicates that the refining treatment was to be adequate to refine this added metal adequate. separately. I t should be pointed out that 7. A moderate quantity (0.5 to I per while there is about 10 per cent flux in the cent) of 310 flux should then be spread pot a t any one time, sufficient flux is con- over the surface of the crucible, the tem- tinuously carried along to reduce the over- perature raised to 1650' to 17oo0F. and all flux consumption in the open-pot held for about, 15 min. This treatment process to around 4 to 5 per cent. Tt will refnes the grain in the alloy and causes be necessary to dip the sludge from the the flux to "cure" or dry out, so that when bottom of the pot every 4 to 12 hr., depend- the metal cools to the required casting ing on the rate of metal turnover. temperature the flux either can be pushed back away from the lip or entirely skimmed Crucible Meltiltg and Pouring Practice off before pouring. Again sulphur, or a Crucible melting is one of the methods mixture of sulphur and boric acid, is used commonly used for the production of sand f for protection during pouring. I the proper 4OO THE MELTING AND REFINING OF MAGNESIUM refining and fluxing technique is not analysis. After this batch of metal has followed, the flu^ occasionally may dry out completely melted, it is refined and made or crumble after superheating, thus allow- of uniform composition by thorough stirring ing oxidation. I t is a dangerous practice, with No. 230 flux. Sludge must be removed as regards fluxing the castings, to add fresh from this crucible from time to time. Since flux a t this stage. A very light dusting a t the primary purpose of this unit is to melt the exact point of oxidation is permissible, the magnesium and provide some refining but the general remedy is better refining treatment, ordinarily the metal is not and the use of more flux a t the refining heated above about 1400°F. at this stage. stage. It is quite important that no pro- As often as metal is required for super- longed delay be allowed between the time heating and pouring, it is transferred the metal is a t the superheating tempera- from the premelters to smaller crucibles, ture-that is, 1600°F. or above-and the usually of capacities between 60 and 550 lb. time of pouroff. The superheating treat- of magnesium, by tilting the large units. ment causes a desirable grain refinement, The metal in the latter crucibles is pro- which is lost rapidly on standing at lower tected with No. 310 flux, which usually temperatures. is dusted lightly into the hot crucible 8. The crucible size should be selected before the metal is transferred. These for a given casting job, so that a small heel crucibles are then placed in settings and of metal will always be left in the crucible. the metal temperature is raised to about After the pouroff is finished, the crucible r350°F. At this point additional 310 is returned to the melting room and the flux (approximately 1.5 per cent) is sides spudded down, fresh flux being added added. As soon as this fresh flux is melted, when required to inhibit oxidation, and it is stirred vigorously through the metal the dross and sludge skimmed out of the bath for a period of about one minute. bottom of the crucible. The heel of flux This refining treatment brings a clinker- and metal then remaining may be used like dross to the surface and drops the as a starter for the next charge. rest of the impurities to the sludge. From The normal consumption of 310 flux f this point forward, the process o super- for this type of operation would be about heating, pouring, etc., is the same as 3 per cent. that previously discussed for the crucible melting and pouring practice. Two-step Premelting and Crucible Practice In this type of operation, about 2 to Two-step premelting and crucible melt- 4 per cent of No. 230 flux would be used ing is the most widely used method for the in the premelting and approximately 3 per production of sand castings and is used to cent of No. 310 flux in the crucible. This some extent for permanent-mold castings. process offers many advantages, such as a In this type of operation, all of the ingot double refining treatment and high-speed metal and foundry scrap is melted in large method of melting. tilting crucibles, usually around 2000 lb. magnesium capacity. No. 230 flux is used Die-cast Melting Practice for protection on this premelting crucible The die-cast melting method is used and the melting down operations are almost universally in this country for exactly the same as those described under providing metal for die-casting operation. open-pot melting practice. Ordinarily, the I t is considerably different from any o f large premelting crucibles are operated on the other practices in that it uses No. 220 a batch basis, so as to provide a single f flux, which has the characteristic o being large source of metal having the same able to refine the metal in the same way NELSON 401 as previously described fluxes but does The foregoing procedure is followed in not have the property of giving surface starting up a new pot, but in normal protection. Since the die-casting process operation it is preferable practice to keep requires the introduction of a ladle into the die-casting pot relatively full of metal. the pot many times per minute, the use of a This may be accomplished by adding fresh flux that gives surface protection would ingot at a rate sufficient to maintain a lead invariably to the inclusion of flux constant metal level; or, still better, in the metal, because of the continued to do all of the melting in a separate agitation. For this reason, this flux has premelting unit, using No. 230 flux and been definitely designed to do the refining then transferring molten, refined metal and then settle to the bottom, so that it directly to the casting pot by tilting or may be completely removed from the pot ladling from the premelter as convenient. with the sludge. Surface protection in this The usual range of pouring temperature case is provided by an atmosphere of SOz, for die casting is from 1175' to IZ~OOF., which usually is generated by burning so the oxidation tendency for the metal is sulphur in a hollow dome that operates relatively low. I n this operation, the as a cover for the melting pot. The ladle metal usually is protected by a thin film, is introduced into the pot through a small which must be parted when the ladle is opening through the cover. The SOt introduced. I there is evidence of undue f generated within the hollow dome is oxidation on the metal surface after allowed to enter the space just above ladling, it means that the sulphur dome the metal in such a way as to blanket the is not working properly. Sometimes a molten metal surface and to retard the very slight dusting of sulphur or other entry of air through the ladle opening. similar agent may be required to control The primary meltdown in this case is oxidation. I n no case would fresh flux done by dusting the surface of the pot with be added to the pot unless the entire No. 220 flux and charging the solid metal rehing and sludge-removal processes were into the pot. with careful attention and to be repeated. I n continuous operation, fluxing to prevent oxidation. When the particularly if ingot is added directly to batch is all molten, it is stirred thoroughly the pot, i t may be necessary to re-flux with the flux to agglomerate the oxide. the pot every 4 to 8 hr. Ordinarily from I t is next allowed to stand for about IQ I to 3 per cent of No. zzo flux is used in min. to permit settling of the sludge and the process. flux. The melt is then ready for sludging, Sc~ap.-The melting and refiningv of which consists of removing the sludge with die-casting scrap presents a special problem a skimmer perforated with %-in. holes. different from that for the bulk of mag- The skimmer is preheated and put into nesium scrap, in that a certain proportion the metal, scraping the bottom and sides of carbonaceous materials, lubricants, of the pot. etc., are present on the scrap, so that When the skimmer is brought up with melting and refining in No. 230 flux as in the metal and sludge, the molten metal ordinary practice is not satisfactory. A runs out through the perforations, leaving reddish scum, or film, of the carbonaceous only the sludge, which is dumped into a material seems to stay suspended through- preheated pan. This process removes all 'out the metal. This behavior is eliminated of the flux and dross and the metal should through the melting of such scrap in open now appear clean and shiny. It is advisable pots or crucibles, using No. 310flux instead to allow about 10 min. quiet standing of the usual 230. The No. 310 flux has the before beginning to dip metal from the pot. characteristics necessary to agglomerate ‘ 4' 2 THE MELTING AND REPINING OF MAGNESIUM the carbonaceous film and refine this The No. 230 flux has these characteristics material. Melting procedures are similar as well as less tendency than other fluxcs to the open-pot operation with No. 230 to dry out or thicken during use, thereby flux for scrap recovery, except for the providing a cleaner pot, better metal substitution of 310 flux. protection and better separation of flux from the metal. Another advantage has been that the amount of flux consumed The development of fluxes and the for open-pot alloying operations has technique for their use in this country dropped from about 8 per cent down to has gone a long way toward simplifying 4 per cent. and improving the methods originally What is probably a still more significant used in Germany and practiced abroad development is the No. 310 flux, which even a t the present time. The original embodies in one agent all of the behavior German practice made use of the naturally and characteristics that are obtained by occurring double salt of magnesium and the multistep fluxing practice used abroad. potassium chloride (carnallite). That this More specifically, this flux a t first is thinly material is not exactly suited to flux fluid, so as to provide adequate protection practices was early recognized by the during meltdown, and gradually thickens Germans and attempts were made to during the refining stages. While the metal correct its behavior through the addition is being superheated for grain rehement, of other materials and more practically the flux still provides protection and is through the use of multistep flux opera- converted during this high-temperature tions." I n those processes, the molten process to a surface crust, which may be metal is first washed with a thinly fluid skimmed off or parted for the pouring flux, a second flux containing thickening operation. Flux inclusions are never ob- agents is then added for the refining tained in castings produced when this type stages and finally a third flux containing of crucible flux is used correctly. I t simpli- thickening agents may be added, in fies the foundry practice by requiring only order to absorb all the fluid flux and one main flux and eliminates the guesswork develop a crusting protection on the of operators in using the correct propor- surface, which can be skimmed off or tions of the various fluxes required in the pushed back before pouring.3 multistep operation. Similarly, it may be Those who have been long associated said that while better fluxes undoubtedly with the development of magnesium in will still be discovered, the No. 230 and the United States will well remember the 310 fluxes provide a range of protection widespread difficulties with fluxed mag- adequate to meet most melting problems. nesium castings in the early days. These I t is also of possible interest to point out were caused primarily by the use of the that the widespread use of premelting units carnallite fluxes. A definite development in in this country is a new and significant this country has been in the evolution of a development in handling magnesium. The single flux such as No. 230, which, with process is of interest primarily from the reasonable use and appropriate technique, standpoint of facilitating production, in permits single-stage alloying or open-pot that a double refining treatment is not melting and pouring operations by provid- needed, as either 230 or 310 flux alone gives ing a surface protection for the metal and a sufficient refining action. More specifi- permitting the flux film to be parted for cally, this permits an increase in melting ladling and then tending to re-cover this rates and provides a large batch of metal surface after the ladle has been removed. of uniform and single analysis. NELSON 4O.3 points in the behavior of magnesium and Iron is precipitated in the production of its alloys is the characteristic of a very essentially all the wrought magnesium marked grain-refining action obtained by alloys except the Mg-Mn type used in this means ol a superheating treatment some country. AS previously stated, iron may 2ooa to 4o0°F. above its melting point. occur in the original magnesium and is I n this behavior, it is considered to be added in larger amounts in the aluminum much like cast iron, and the theories and and other alloying constituents, and may hypotheses used for the explanation of the be picked up from the flux and from the behavior of the cast iron may be applied pot itself. The iron is precipitated after to magnesium with the substitution only the principal alloying materials have been of the alloy and the types of impurities. added, usually by addition of manganese Perhaps the most widely supported to a saturation value for a temperature theory a t the moment is that some material above that a t which the metal is to be that a t normal temperatures is too large poured. When manganese is present up to in particle size to be effective is taken into its solubility value, particularly in the solution a t the high temperatures, and presence of aluminum, iron is greatly reprecipitates to form fine nuclei during reduced in solubility or tendency to stay the cooling process. The fact that after in suspension, and is rapidly precipitated superheating the grain-refining effect de- to the bottom of the pot. This concept is teriorates upon long standing a t tempera- somewhat different from that pointed out tures below the superheating range would by Beck2cand others? in that they visual- lend credence to the idea that these nuclei ized an entrapment of the iron in primary are again coalescing and growing to their crystals of manganese or silicon that could original form and relative ineffectiveness. be used as the precipitation agents. I n I n any case, the superheating practice any case, the process operates consistently is very important in most casting opera- to produce iron contents on a commercial tions, in that it provides h e r grain size scale a t an average of,o.oo~ cent or less. and a finer and more uniform distribution per In actual practice, the manganese of the magnesium-aluminum compound, usually is introduced as manganese chlo- which further leads to more homogeneous ride, or, preferably, as No. 250 or No. 3 2 0 heat-treated structures and high properties. manganese flux. The former is used'if the The method by which the superheating process is to be carried out in a n open pot effect is obtained has been discussed to in the presence of 230 flux and the latter some extent, and only a few more details if i t is to be carried out in a crucible in the will be mentioned here. Superheating presence of 310 flux. The manganese-flux effects may be obtained as low as r5oo0F. compositions are balanced so that the. but the time required a t this temperature reaction products are compatible with the is rather long (several hours). As the tem- respective fluxes. perature of superheating is raised to about While the iron-precipitation processes 1700' to 1750°F., the time required to get can be carried out successfulIy in steel the effect gradually drops of! to approach pots or crucibles, any attempt to remelt or zero. reheat high-purity alloys appreciably above I t has been stated in the literature that r300° to r350aF. in contact with steel the effect of superheating carries over pots will lead to contamination by iron. was through remelting steps .2d.6Thi~ found to be true to a limited extent, in that part SUPERHEATING of the time metal so treated might show Perhaps one of the 'most interesting the full benefits of superheating but more 404 THE MELTING AND REFINING OF MAGNESIUM often the effect would be only partially be sufficient to permit the elimination of a realized. I n view of this action and the full superheating treatment just before need for a good job of superheating prior pouring. Since 60 to 80 per cent of the to the actual pouring of the casting in charge being melted in foundry crucibles production, in order to ensure uniformity consists of foundry scrap that already has of behavior, the advantage of using pre- been repeatedly superheated, any benefits . superheated metal is considered question- of such carry-over are already enjoyed. If able. T o this point may be added the fact the final superheating is proper, no addi- that in most casting operations the bulk tive effects are gained from previous of the material going back into the melting superheating. pots is foundry scrap; which already has gone through from one to five superheating REFERENCES I. C. E. Nelson: Secondary Magnesium. operations in connection with previous Secondary Metals Symposium. Metals castings and therefore would have whatever Tech., Oct. 1943. A.I.M.E. 2. Beck: The Technology of Magnesium and advantages are to be obtained through Its Alloys. (a) 319; ( b ) 313-318; ( c ) carry-over of superheating effects. A point 318-319; ( d ) 321-322. 3. British Patent 469347. on which we can fully rely is that if the 4. Schmidt and Beck: U. S. Patent zozg8q8. finalsuperheating before casting is properly 5. K. Achenback, H. A. Nippur and E. Piwo- warsky: Contribution to the Question of done no additional benefits are realized Melting Practice for Cast Magnesium from previous superheating. Alloys. D e Giesserei (1939) 26, 597-604. i 621-628. DISCUSSION F. A. Fox.*-It is hardly possible in a An attempt has been made in the fore- limited space to' comment adequately on this going pages to bring out the following interesting paper by Mr. Nelson. The mcthods significant points: he describes difier considerably from standard I. The use of the correct fluxes and the practicc in this country. I n our opinion the proper technique is fundamental in the fluxing methods used by M.E.L. and the Elek- melting and refining of magnesium. tron group possess important advantages over 2 . Fluxes and techniques have been the methods described in this paper, in par- developed and are available for use in this ticular as regards simplicity of procedure, reduction in quantity of flux required, negligi- country that are suited to the various types ble risk of flux inclusions even with somewhal of melting and refining operations. careless operation, and reduced metal losses. 3. The use of a direct-flame reverbera- These methods are briefly outlined later. tory type of furnace is a new and demon- Mr. Nelson's references to the original strated development for large-scale melting. German practice and to "methods practiced 4. Magnesium alloys of controlled high abroad even at thc present time" do not appear purity are already being produced on a to be always correct and seem to us to be mis- large commercial scale for wrought mag- leading. For instance, the impression might be nesium products and to a limited extent gained, on reading Mr. Nelson's paper, that for castings. The use of magnesium of initial American melting and fluxing practice was high purity is not a significant factor in this ahead of that in the United Kingdom and that the developmcnt of fluxes Nos. 230 and 310 production, since impurities are added dur- represented, in particular, an advancement ing alloying and melting. A precipitation over existing fluxes used elsewhere. Neither method for lowering the iron after these our own experience with No. 2 3 0 and No. 310 operations appears most promising. fluxes, nor the statements made by Mr. Nclson 5. Superheating is very important in the in his paper, appear to us to justify such a view. production of high-quality castings. I n In iact, evidence pointing rather to the con- actual practice carry-over effects of pre- - * Chief Metallurgist, Magnesium Elektron superheating metal are not considered to Limited. Clifton Works, Manchester, England. JSSION 4O5 trary might be presented; for example, from about 750°C and a fresh layer is applied to the historical point of view, a flux essentially the clean metal surface. Before pouring, the similar in behavior and purpose to No. 310 has flux cover is gently drawn away from the lip been in use by one of our associates for'about of the pot and the metal is then cast. nine years. Further, the No. 310 flux itself An all-purpose flux, termed Melrasal UE, is not unlike one of the German Elrasal fluxes which is intended for melting, refining, and as in both behavior and composition. a cover during superheating and casting, is The total consumption of fluxes in our own suitable for ingot melts up to about 1000 lb. of foundries in one complete operation involving metal. melting, alloying, refining, superheating and No bath of fluid flux is used in melting, and casting of metal on the same scale as that of the a t no time is i t necessary either to scrape flux Dow "open-pot alloying process" has for some up through the metal from the bottom of the years been 2.5 to 3 per cent of the weight of pot or to remove i t from the metal surface. metal melted, whereas the Dow flux consump- The total consumption of flux in a single tion in the "open-pot alloying process" alone operation involving melting, alloying, refining, is now 4 per cent. superheating and pouring has already been The trouble with flux-contaminated mag- quotedas 2.5 to 3 per cent of the weight of nesium castings in the early days of develop- metal melted, and the gross metal loss is 2 to ment of maghesium in America, mentioned by 3 per cent, of which a high percentage is re- Mr. Nelson, and which appears to have con- coverable by suitable treatment of the flux tinued long beyond I'the early days," could residue. Ladling processes for the transfer- hardly have been due primarily to the use of ence of metal from one pot to another are carnallite itself, because satisfactory castings avoided and die casting by ladles is considered were being prepared with similar flux materials undesirable. in England, France and Germany a t that time* when, incidentally, the tonnage of magnesium made and processed in Europe exceeded that in The use of highly fluid noninspissated fluxes the U.S.A. by a ratio of about ten to one. in flux baths and for covering the metal seems The "original German process" described by to us most undesirable from the point of view Mr. Nelson was probably never in use in three of complete freedom of the castings from flux. stages as standard melting practice. Even in The ladle casting process involving the use of a 1936, a single, all-purpose, inspissated flux flux pot for washing the ladle, etc., appears to essentially similar to Dow.310, was supplied by be a complicated, and, from the metal effi- I. G. Farbenindustrie A.G. to its casting ciency point of view, an uneconomic way of licensees. casting ingot. Ingot produced in this country is A detailed account of our fluxing process was cast continuously from automatic, tilting 2-ton recently published in Magnesium Review and crucibles, into molds arranged on a continuous Abstracts (Vol. 4, No. I, Jan. 1944). For con- conveyor belt. An inspissated flux cover is used venience, a brief outline of the methods used on the metal. follows. Melting in a flux bath is used by us only in the treatment of material obtained in the course of metal-recovery operations; this metal Melting is carried out with a fluid flux is cast into ingot and subsequently remelted. Mclrasal 2, corresponding roughly in be- CRUCIBLEMELTINGA N D POURINGPRACTICE havior to Dow 230, which is sprinkled lightly over the ingots after the latter have been The reason for some "clinker-like dross" charged into the melting crucible, and subse- rising to the metal surface and requiring quently applied as required to prevent burning. removal is not clear. When the metal is molten, an inspissated flux, I n contrast with Mr. Nelson's warning Melrasal E, which Dow 310 somewhat resem- against addition of fresh Dow 310 flux im- bles, is used. This is stirred into the metal a t mediately before casting, the use of Melrasal E flux a t this stage is perfectly permissible. Ii 'See, for example, French Air Ministry desired, the vvhole of the old superheated flux Report: Le fusion industrielle des MagnCsiuin et ses moulages en sable. by A. Caillon (1938). cover can be removed and a layer of fresh 4 0 ~ THE MELTING AND REFINING OF MAGNESIUM Melrasal E applied in its place; as soon as author expects from the exposure of metal for melting of the flux layer is complete it may be appreciable periods to temperatures between drawn back from the lip of the pot and pouring 1 5 ' and 1 5 c o O F (750' to 815~C.),we would 40 beeun. not consider this an effective superheating or grain-refining temperature range, irrespective of the length of time involved. For a few thin-section jobs i t may be neces- Apart from the additional stirring with sary to cast a t temperatures as high as 1470°F. No. 230 flux, the desirability of which is not (doo°C.): a t these temperatures the unpro- appreciated, this process appears essentially tected metal bums vigorously, and sulphur is to correspond with our standard practice. an inadequate protection. How would the author avoid in such a case contamination of f the casting with oxide inclusions i the flux is In our opinion die casting should be done "entirely skimmed off before pouring" (cruci- wherever possible from smoothly operating ble melting and pouring practice)? tilting pots using an inspissated flux cover. Where ladle casting is used we consider i t C. E. NELSON (author's reply).-As an preferable to cover the metal surface with opening statement pertinent to this discussion, Melrasal E, or, better still, with Melrasal UE, the author would like to say that M.E.L. and part the flux cover with a flux spoon or skimmer Elektron fluxes and methods have been tested and then 611 the ladle. A light dusting of and compared with American fluxes and prac- Melrasal E is applied to the exposed metal tices by the author and by other reliable surface in the pot when visible oxidation be- foundry establishments in this country. This gins. Such a process may be conveniently has been done recently as well as over a good combined with the dome hlled with SO2 de- many years. The conclusion has been that scribed by Mr. Nelson, when it will be found while in many cases the M.E.L. fluxes are the that as development of thin oxide skin pro- equivalent of those used in this country, there gresses, i t is quite unnecessary to refine with is no indication that they are superior in any Melrasal E, much less to sludge the pot; mere respect. application of a fairly thick layer of Melrasal There is no intention, either in this discus- E to the metal surface followed by parting of sion or in the paper, to indicate that any the flux cover after melting is complete will classes of fluxes are not suitable for the pur- enable casting to be resumed. This treatment pose for which they were intended; rather, of the metal surface with Melrasal E may be the theme is to point out that suitable fluxes repeated a t fairly frequent intervals, as re- are available, which, if applied and used cor- quired to maintain a clean metal surface in the rectly as indicated, will allow the production of pot. castings that will meet the very high inspection MISCELLANEOUS standards in force in this country. The comment that "flux inclusions are The terms "similar flux materials," "essen- never obtained in castings produced when this tially similar," or "corresponding roughly to," type of crucible flux is used correctly" is no used by Mr. Fox are not enlightening, since doubt justified, but it appears to imply that almost all the fluxes under consideration are flux-contaminated castings may sometimes be made up from the same group of chemical produced when other types of flux or process compounds and differ only in respect to rela- are employed also "rorrectly." I so, what f tive amounts present, depending on the in- justification can there be for using such fluxes tended use of the product. We know that small or processes a t all? amounts of some of these elements markedly The use of chlorine treatment mentioned affect the behavior of the flux. We are very by the author to degas melts of magnesium- certain that castings made with carnallite base material is unknown in this country, as it would not regularly meet the inspection stand- is never necessary: a contributing factor here ards set up for castings in this country with is probably the highly anhydrous character of respect to freedom from flux inclusions. our fluxes. Melrasal Z, which is said to "correspond As for the grain-reKing effect which the roughly in behavior to" Dow No. 230 flux, is widely different in composition in that, for paper describes two workable methods to cover example, i t contains a large proportion of CaClr this type of operation; namely, the open pot not found in the 230 flux and a very leu. rela- with No. 230 flux or the SO2 protected pot tive proportion of potassium chloride. Actual along with refining flux No. 220. measurements of the surface tension on these On the point of degassing and chlorine treat- two fluxes show 230 to have a lower surface ment of magnesium alloys, we would say that this is not common practice in this country and tension (go compared to 104 dynes per cm.), is considered only when scrap and ingot are not which may account for the ease of repeated properly dried or handled or where there is abuse ladling from pots protected with 230 flux. in the handling or exposureof the flux. The fluxes Ladling from pot to pot is not practiced in this are all produced from anhydrous materials but country. take on water rapidly upon undue exposure in The values given in the paper for flux con- the melting rooms. This is equally true of the sumption are for individual and single types of British fluxes. A few years ago we thought, operation and the ranges, particularly on like Mr. Fox, that there was no gas in magne- alloying, are made broad to cover all classes of sium alloys but experience in recent years has magnesium melting such as crystals and a shown that it may be present and the resultant proportion of scrap. Wherever alloying, casting, porosity takes the same form as the micro- superheating and pouring are carried out within shrinkage, and hence is difficult to distinguish one plant, or particularly in one continuous from it. It is still true that if great vigilance is operation, figures on flux consumption agree used, gas is not a serious problemon magnesium with those quoted by Mr. Fox. Melting losses alloys. are also of the same order. No prolonged discussion will be given to the Regarding the removal of dross and sludge, matter of low-temperature grain refinement. it should be pointed out that most of the melt- While, as stated in the paper, temperatures as ing processes in this country are continuous or high as 1650" to r7oo0F. are neccssary to get semicontinuous rather than batchwise. The dependable grain refinement in a short time, "clinkerlike dross'' is the same as referred to we know from long production experience that by E. F. Emley, in Magnesium Review and metal held for 2 hr. or more a t temperatures of Abstracts (vol. IV, No. I, p. 11) as I'the thick 1450°F. or above will give dependable and layer of pasty metal oxide and flux," which he effective grain refinement of the same order as also states must be removed with the ladle obtained a t the higher temperatures for short before refining. This is not too common, especi- times. ally in the Cwo-stage crucible operation in I n this country the inspissated flux cover connection with premelters. It may be due to may or may not be removed before pouring the melting of dirty scrap direct in the casting from crucibles. In either case adequate protec- crucible. This does not, however, offer any tion is provided by Dow No. 181 agent, which difficulty in production. is much more effective than sulphur. Castings The Melrasal E flux is almost identical in so produced are extremely clean and free from both composition and behavior with Dow No. oxide. 310 but the author would definitely recommend One h a 1 point concerns Mr. FOX'S comment that neither be used on the crucible just prior that the three-stage fluxing system has never to pouring. These fluxes require time and tem- been used even in Germany. Within the last perature to inspissate and failure to allow for two or three years, M.E.L. interests in this this is precarious. country have advocated the use of a three-flux We believe there may, be confusion on the system consisting of Melrasal Z, E and FSF, to term "die casting," since the author refers to be used concurrently in that order. Our point pressure die cmting whereas Mr. Fox evidently has been that the use of too many different refers to permanent mold casting. We would fluxes leads to mixed fluxes in the melting not favor either Dow 310 or Melrasal E fluxes rooms and leaves too much to the judgment of for pot protection where a ladle must be intro- nontechnical operators who must make the duced repeatedly a t intervals of a minute or additions in such a way that one flux chem- less and final castings are being poured. The ically balances or inspissates the other.
Pages to are hidden for
"Volume Magnesium The Melting and Refining of Magnesium"Please download to view full document