Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 41
Chapter 5 WINNING OF TITANIUM METAL SPONGE Scheduled new (1982) domestic production of vacuum-distilled, Kroll-process sponge and of additional high-purity elec trolytic crystals, coupled with recent expansion of existing leached-Kroll and leached-Hunter sponge production, promise a near-term abundant supply of suitable sponge for existing and new applications. The diversity of production methods, the commissioning of new plants and of plant expansions, and the incremental process and equipment improvements that are occurring are considered as positive factors in assuring future U.S. titanium supply. U.S. titanium production technology and economics could profit, however, from the development of methods for more effective removal of volatiles f ram sponge, f or recovery of anhydrous magnesium chloride from leach brine, and for the direct preparation of high-quality, low-cost titanium powder for powder metallurgy. The four current (1982) and two prospective U.S. sponge producers represent a spectrum of processes, operating scales, and modernization levels. The modernization level also varies for different operations within the same plant. ~ Recommendations f or incentives to induce modernization of outmoded facilities are presented in Chapter 12. ~ Commercial, experimental, and theoretical methods for the production of titanium metal were comprehensively reviewed in a 1974 NMAB report. When that report was published, magnesium reduction of TiC14 (Kroll process) was the major process used in the United States, Japan, the Soviet Union, and China. Sodium reduction of TiC14 (Hunter process) was the other process used in the United States and Japan, the only process used in England, and the process used in one of China's five plants. Electrowinning from TiC14 in a fused salt bath was being t ested on a pilot-plant scale. Few of the many other proposed processes reviewed in the 1974 NMAB report had advanced beyond a laboratory stage, and most were considered by the authoring NMAB panel to be unlikely to progres s to product ion soon. The f inding s of the 19 74 NMAB report related to co~nmerc ial and certain other processes are summarized and up-dated in this chapter. Since the report was written, several major events have occurred: the first U e S e vacuum-distilled Kroll sponge production, since du Pant ceased production in 1962, occurred in 1980 when Teledyne Wah Chang Albany converted from zirconium to titanium production; titanium metal production by electrowinning was initiated in 1980 on a full-scale, production-module basis by the D-H Titanium Company; and construction of an advanced technology, Krol1-type plant was begun by International 41
OCR for page 42
42 Titanium, Inc., (ITI) at Moses Lake, Washington. ITI expects to produce sponge of higher quality than the older domestic Kroll-type plants using acid leaching and at a rate of 5.5 million lbs of metal per year by mi d-198 2 . ~ Of the world ~ s total 1980 sponge production of about 85 ~ 000 tons ~ approximately 7 2 ~ 000 tons (85 percent ) were made by the Kroll proces s . Of the 26,000 tons of U.S. sponge produced in 1980, about 17,000 tons (65 percent ~ were made by the Kroll process. U . S . Product ion of Ti tanium Sponge U. S . titanium sponge capacity is discussed in Chapter 8. The four current producers (in order of startup and of size of production) are TIMET, RMI Company, OREMET, and Teledyne Wah Chang Albany. Two additional prospective producers are the D-H Titanium Company and ITI. Several others are being discussed. One, Albany Titanium, Inc., has announced plans f or a small Kroll-type sponge plant f or approximately 500, 000 lbs per year starting in 1982. General descriptions of the Kroll, Hunter, and electrolytic processes, followed by some operating details of each U.S. titanium production process are given below. Magnesium Reduction of TiC14 (Kroll Process) In the basic Kroll process, molten magnesium metal and gaseous TiC14 react in a sealed steel pot at a temperature of 800 to 900°C with formation of solid metal titanium in sponge form and molten MgC12. Theoretically, a ton of titanium metal and 7,950 lbs of MgC12 would be produced by the reaction of 2,029 lbs of magnesium with 7,921 lbs of TiC14. In practice, an excess of magnesium (10 to 15 percent) is used to assure the complete utilization of TiCl4. With the escalation in the cost of magnesium, plants that leach and impound MgC12 are forced by economics to minimize the amount of stoichiometrically exces s magnesium. This i s done at the ri sk of leaving TiC12 in the sponge titanium that can ead to low-density inclusions in the ingot. The exothermic reaction takes place over abou~ a 24-hour or longer period with the pot in a furna~e to provide temperature control supplemental to that obtained by regut_~ing the rate at which liquid TiC14 is sprayed into the pot. Since the reaction is highly exothermic, an excessive TiC14 feed rate .'uld damage the pot. Magnesium metal is introduced to the pot in solid or liquid form. When a solid charge is used, the lid of the pot is welded to the cylindrical body after the ingots have been stacked in the pot. Air then is evacuated through valves in the lid, and the pot is purged with argon
OCR for page 43
43 or helium and placed in a furnace to melt the magnesium before starting a measured flow of TiC14. When molten magnesium is used, it is pumped into the argon- or helium-filled pot in the furnace before starting the f low of TiC14. Since molten magnesium floats on the surface of the molten MgC12 that forms, it remains accessible to the TiC14 feed. Molten MgC12 is tapped from the pot at intervals of 3 to 4 hours or longer and is charged to electrolytic cells for conversion to magnesium metal (for reuse in the reduction pots) and chlorine. The chlorine produced is used by the producer to prepare TiC14 by reaction with rutile concentrate or is returned to the TiC14 supplier or sold on the open market. When the scheduled amount of TiC14 has been reacted, a f inal tap i s made of MgC12 along with any of the excess magnesium that drains from the pot. Residual MgC12 and magnesium metal entrained in the titanium sp onge are separated from the sponge by one of the following procedures: 1. Removal of MgC12 and magnesium by vacuum distillation from a heated pot at a temperature of 900°C and a vacuum of below 100 microns of mercury. The operation may take 48 hours or more depending on the capabili ty of the vacuum equipment . Both MgC12 and magnesium metal are recovered by condensation. After cooling, the weld metal holding the lid to the pot is ground away and the pot is opened . The sponge is bored, chipped, and extracted on a removable cradle or is pressed from the pot. Pressing the sponge from the pot requires that the pot be composed of three sections; the top and bottom sections are removed and a ram is used to force the sponge from the center section of the pot. Vacuum distillation, originally used by the U.S. Bureau of Mines, and by du Pont from 1948 to 1962, is employed by Teledyne Wah Chang Albany and by ITI at Moses Lake, Washington. Vacuum distillation is standard procedure in Japan and the Soviet Union. Sweeping the heated pot with helium gas to reduce volatile MgC12 and magnesium to a low level and subsequently recovering these by condensation. After cooling, the pot is opened and the titanium sponge is removed, sheared, crushed, and then leached in acidified solution to remove remaining MgC12 and magnesium. This practice is used by OREMET in Albany, Oregon. 3. Opening the cooled pot in a "dry" chamber. The dry atmosphere is necessary to avoid reaction of retained salts in the sponge with moisture in the air. Titanium sponge, and admixed MgC12 and magnesium, are bored out of the pot, crushed, and leached in a buffered acid solution. This practice is used by TIME: T in Henderson, Nevada. Vacuum distillation is capable of yielding sponge that contains less magnesium, magnesium chloride, and other volatiles than does acid leaching of the sponge. Distillation is more energy-intensive than acid
OCR for page 44
44 leaching (although this can be minimized by using the heat of reaction for distillation) and requires more expensive reactors (to withstand vacuum at high temperature). The low-chloride sponge also is more difficult to crush and grind. Conversely, it produces metal of generally better quality than acid leaching, recovers more magnesium and magnesium chloride for re-use, and avoids the problems and expense of disposing of an acid solution containing magnesium chloride. Generally, leaching was the economic choice in the 1950s when magnesium and energy were inexpensive, but in the 1980s, vacuum distillation is the preferred route. This important aspect is discussed in more detail in a subsequent section of this chapter. The major features of the main titanium producers using the Kroll process are described immediately below. IDMET DIVISION OF TMCA The TrMET plant at Henderson, Nevada, owned by NL Industries and Allegheny International, has been producing sponge by the Kroll process since the early 1950s. It s current sponge capacity is about 15 ~ 000 tpy As a fully integrated producer, from ore to titanium ingot and finished items, TIMET makes its TiC14 by fluidized-bed chlorination of imported rutile or rutile substitutes. Most of the plant 's magnesium cells date from the World War II basic magnesium plant at the Henderson site, but - re placement program i s under way . TIMET loads its mild-steel reactor pot with magnesium ingot and maintains a helium atmosphere in the pot during the introduction of TiC14 into the reactor and during reduction and cooling. The pot is opened and the titanium sponge is bored out in a dry room. Af ter crushing and leaching with HC1/HNO3 solution that is buffered with citric acid, the sponge is dried and stored until needed for conversion to ingot a About an inch of sponge is left as a lining in the cylindrical pot to absorb the trace of iron dissolved from the steel pot by the molten magnesium. The lining sponge is recovered after about 60 cycles when the pot is scrapped. TIMET generally sells this high-iron sponge to manufacturers of ferrotitanium for use in steel alloys. The softest and best quality sponge occurs farthest from the reactor wall and is cut selectively from the pot when highest purity is desired. Each pot produces about 1 ton of sponge in a 3-day cycle. Titanium is harvested from about 44 pots each day. No salts are recovered from the leach solution; it is stored and evaporated in tailings pond s. Make-up requirements are about 0.3 ton of magnesium and 1 ton of chlorine per ton of sponge produced.
OCR for page 45
45 OREME:T The OREMET plant at Albany, Oregon, publicly owned but with the Armco Steel Corporation as the major stockholder, has been producing sponge by an advanced Kroll process since 19 66 . It s current sponge capacity is about 4, 500 tpy. The plant is fully integrated from sponge deco f inished items, except that it does not now prepare its own TiC14 (although it formerly did so and is considering it once again), and chlorination facilities are still available on site. TiC14 ~ s purchased from Gulf and Western, Ashtabula, Ohio, and delivered in 55-ton tank cars. Chlorine from OREMET's electrolytic magnesium recovery cells is sold on the open market. OREMET uses 304 type stainless steel reduction reac tors of a horizontal cylindrical design that produce 14,000-lb batches of sponge. Reactors are loaded with molten magnesium through valves in the welded reactor. An inert atmosphere is maintained during the loading and reduction of TiC14. Most of the MgC12 and magnesium not drained from the reactor af ter the last tap are removed by a helium sweep of the heated reactor. The reactor then is cooled, and the head is removed by cutting. A steel "ridded cradle the same length as the reactor vessel facilitates removal of the sponge in the form of a single, 7-ton, roughly semi-cylindrical shape. This sponge shape is sheared, crushed, and leached to remove remaining chlorides and magnesia. The results ng sponge is screened to remove fines, dried, and stored until needed for conversion to ingots. TELEDYNE WAR CHANG ALBANY (TWCA) TWCA had been producing zirconium by a Kroll-type reduction of zirconium tetrachloride since the late 1950s. Because of a depressed market for zirconium and a shortage of titanium, idle equipment was converted in 1980 to produce titanium sponge at a rate of 1,000 to 1,500 tpy. TWCA titanium sponge is vacuum distilled and therefore has significantly lower volatiles relative to acid-leached sponge. TWCA the second U.S. producer of vacuum-distilled titanium sponge; du Font was first, in 1948. INTERNATIONAL TITANIUM, INC . This new Kroll process titanium sponge producer will be the third U.S. user of vacuum distillation. Construction of the ITI plant at Moses Lake, Washington, started in July 1981. Ownership is by Ishizuki Research Institute and Mitsui Company, Ltd., of Japan and other Japanese and American business interests. Sponge production began in early 1982 at a rate of 2, 750 tpy, and if market conditions warrant, the facility could readily increase production to 3, 300 tpy in 1983. ITI has no current plans to produce ingot.
OCR for page 46
46 The fluid-bed process for chlorination of rutile is used by ITI f allowed by purification of the TiC14 produced by solids condensation and fractional distillation. Sectional vertical reactors produce a 5-ton batch of sponge, but it is possibile that they will later be expanded to accommodate 7-ton batches. Magnesium and magnesium chloride not drained from the pot are removed by vacuum distillation and recovered. Sponge is removed from the sectional reactor by pressing it out of the reactor cylinder. Magnesium and chlorine for recycle in the plant are produced f rom recovered magnesium chloride in a newly designed and enclo sed Ishi zuki magnesium cell. The energy requirement per lb of magnesium is expected to be only 4 .6 kWh compared to between 7 and 8.5 kWh in conventional magnesium cells. Other advantages predicted for the new cell are longer life of components, less release of chlorine, higher purity chlorine, and improved yield of magnesium and chlorine. On the basis of sponge quality and production costs, ITI plainly expects its plant to be c as t-compe t it ive with old, depre ci ated U . S . sponge p lent s and quality-competitive with new Japanese sponge plant s. Sodium Reduction of TiC14 (Hunter Process) As initially practiced, the Hunter process reacted TiC14 wi th elemental sodium under an inert gas atmosphere in a sealed steel pot at temperature of about 900°C. Ti tanium sponge and molten sodium chloride were formed. Subsequently, a two-stage reduction procedure was adopted. In s Cage one, TiC14 and enough liquid sodium to reduce the TiCI4 t o TiC12 are f ed continously to a stirred and continuously discharging reactor. In the second stage, the flowable mixture containing TiC12 and molten salt formed in the f irst stage is transferred to a batch reactor pot that contains molten sodium for completion of the reduction to titanium sponge. An inert gas atmosphere is maintained in both reactor stages. The second-stage reactor is positioned in a furnace for temperature control to supplement the exothermic reduction reactions. Theoretically, 1 ton of titanium metal and 9,762 lbs of sodium chloride would be produced by the reaction of 3,841 lbs of sodium winch 7, 9 21 lbs of TiC14 . In practice, an excess of TiC14 i s used to avoid the presence of free sodium in the reaction products because it poses a fire and explosion hazard. Since sodium and titanium subhalides are soluble in molten sodium chloride, it is necessary to retain the reactants in the sealed pot until the reaction has been completed, the pot has been cooled, and the welded head removed. The mixture of sponge titanium and sodium chloride i s chipped from the reactor, crushed to about 3/8-inch lumps, and leached in dilute hydrochloric acid solution to dissolve the salt. The resultant leach brine may be reprocessed in the sodium-chlorine plant to make pure sodium chloride for electrolysis. The washed sponge is dried, screened to remove f ines, and pressed i nto compact s
OCR for page 47
47 for vacuum arc melting. Unlike the Kroll process that tends to form sponge high in iron near the reactor walls, sponge throughout the Hunter reactors is of uniform grade and low in iron content. The 1974 NMAB report alluded to the desirability of doing the second stage, as well as the first stage, of the Hunter reaction in a continuous manner as was investigated by RMI, but such a technology has not been established to date. There also was conjecture about using the first-stage Hunter reaction to prepare TiC12 feed for a titanium electrowinning cell. New developments on this have not been reported. RMI COMPANY Titanium sponge has been produced since 1957 at Ashtabula, Ohio, by RMI and its predecessors via the Hunter process. RMI is owned jointly by U.S. Steel Corporation and National Distillers and Chemical Corporation. As an integrated producer from sponge to finished items, RMI buys TiC14 and currently makes sponge at a rate of 9,500 tpy. TiC14 is purchased from the nearby Gulf and Western pigment plant, and chlorine, from RMI's sodium-chlorine cells is returned to Gulf and Western for reuse. Because the second-stage reactor (sinter pot), at completion of the reaction between sodium and TiC12, contains about five times as much salt as sponge, the batch size of the resulting sponge is smaller than when the same size vessels are used for magnesium reaction with TiC14. Jack-ha~ners are used to remove the mixture of sponge and salt ~ spelt that then ~ s crushed, ground, and leached in a continuous system. As much as 10 percent of the leached sponge is removable as fines that have a potential market in powder metallurgy applications. Material as coarse as 60 mesh (0.0098 in. or 250~) has been sold for powder. Difficulty has been encountered in removing the final traces of chlorides from leached powder. The presence of chlorides interferes with the welding of titanium shapes made by powder metallurgy as noted in some detail in Chapter 11. Direct Electrowinning of Titanium Sponge Both the Kroll and Hunter sponge processes are indirect electrowinning procedures that rely on electrolytic production of magnesium or sodium for reduction of TiC14. The design and operation of test cells for direct electrowinning of titanium from TiC14 fed to a fused salt bath was reviewed in the 1974 NMAB report. Although titanium sponge (intermeshed crystals) of excellent quality was produced in pilot-plant cells with a daily titanium capacity of up to 190 lbs, major design and operating problems were apparent. Among these were: 1. The difficulty of preparing durable diaphragms with limited permeability and low electrical resistance to divide the cell into anolyte and catholyte compartments.
OCR for page 48
48 2. The need to prevent the TiC14 cell f eed from exiting with the C12 produced at the anode and the limited solubility of TiC14 in the fused chloride salts that requires provision for its promp t reduction to soluble TiC12 and TiC13. 3. The need to control the average valence of dissolved titanium at about 2.1 to obtain deposition of premium-grade metal. 4. The need to maintain an inert atmosphere in the cell even when withdrawing the cathode sponge deposits and inserting new cathodes. 5. The need to minimize fused salt entrainment (dragout) and loss of dissolved subhalides. 1}H TITANIUM COMPANY [A recent news item (American Metal Market, December 30, 1982) states that the expense for completing the joint program and the present economic climate have forced the dissolution of the D-H Titanium Company. With the breakup each company is to proceed with developing its own technologies; Dow will continue research and development work on the electrolytic process, and Howmet will proceed in the metals fabrication area. Available data on the D-H process are included here to show where progress has been made in a process that may in time give an alternative for producing high grade titanium sponge. ~ With its entry in 1981 into semicommerc~al sponge production at Freeport, Texas, D-H Titanium, in a close working relationship with the HOW MET group, became the fourth U.S. integrated producer from sponge to f inished items. Cell design, operating procedure, metal quality, proposed production, and economic projections have been described (Cobel et al. 1980~. Assessments by D-H Titanium, based on its projected capital and operating costs, are that new electrolytic plants would cost less to build than new Kroll or Hunter sponge plants and that the projected operating cost for the D-H titanium electrolytic process is equal to or below that of fully depreciated Kro11 or Hunter operations. The semicommercial plant was expected to produce about 100 tons of metal in 1981. Depending on satisfactory operation, verification of projected operating costs and the availability of an adequate market, operation at a rate of up to 5,000 tpy in 1985 was forecast by D-H Titanium. A major cell improvement is the D-H Titanium design and fabrication of a metal screen diaphragm that is electroless-plated with cobalt or nickel to give the required electrical and flow characteristics. The TiC14 feed is reduced to TiC12 at a separate feed cathode within the cell. The electrolyte is a eutectic mixture of LiC1 and KC1 containing about 2 percent TiC12. The preferred operating temperature is about 500°C
OCR for page 49
50 attenti on lately, at least there is an absence of research reports in the technical press. Included in those processes were gaseous and plasma reduction, iodide decomposition, calcium and aluminum reduction, disproportionation of TiC13 and TiC12, and carbothermic reduction. Starting materials for reduction included halides, carbides, nitride s, oxides, and sulfides. The use of TiC14 as the dominant feed for winning titanium metal currently is not challenged. Transportation of TiC14 Only one of the current domestic titanium sponge producers--TIMET at Henderson, Nevada--prepares it s own TiC14. The others purchase TiC14 from Gulf and Western's (G&W) pigment plant at Ashtabula, Ohio. The panel was informed that G&W further purifies its pigment-grade TiC14 before shipping it to sponge metal manufacturers. TiC14 is shipped by rail in steel tank cars that carry about 55 tons of TiC14, equivalent to about 14 tons of titanium metal. Cars for transporting chlorine are suitable f or TiC14. Availability of tank cars solely for TiC14 transport would have to be considered in a national emergency. Energy Use in Manufacturing Titanium Sponge The mining, processing, and transportation of rutile and i ts conversion to sponge titanium metal has been estimated by Battelle Columbus Laboratories ( 1975) to require an energy expenditure of about 423 million Btu per ton of metal in the Kroll-sponge-leach process and 370 million Btu in the Hunter process. Of these amounts, about 225 million Btu (21,500 kWh) is assigned to the electrow~nning of magnesium and chlorine in the Kroll process and 245 million Btu (23,400 kWh) to electrowinning of sodium and chlorine in the Ilunter process. Battelle assumed that 10,500 Btu is required to generate and deliver 1 kWh of electricity. The D-lI electrowinni ng process was reported to require about 16,000 kWh (168 million Btu) for electrolysis per net ton of sponge (G. Cobel, presentation to the panel, 1980) . Energy requirements for make-up (to cover processing losses) magnesium and sodium also were estimated by Battelle (1975) as 110 million and 38 million Btu, respectively. Energy requirements per ton of sponge for chlorine make-up was estimated as 20 million Btu for Kroll-sponge leach and 21 million Btu for the Hunter process. No comparable requirement for make-up reductant metals or chlorine has been reported for the D-H process. Published information to date does not permit a comprehensive estimate of total energy use in the D-H process, but direct current required for electrowinning appears to be only about half that required for the Kroll and Hunter processes. If it is assumed that other D-H process steps have energy requirements roughly equivalent to analogous Kroll and Hunter process steps, total D-H process energy use would be about 250 million Btu per ton. This is about 60 to 65 percent
OCR for page 50
49 Metal-winning cathodes are individually pulled, stripped, and replaced in the cell, in an argon atmosphere, by a self-positioning and automatically operated mechanical device. A sealed, argon-shielded hopper containing the titanium crystals and entrained electrolyte is cooled before being opened to discharge its contents. Crystalline metal and dragout salts are crushed to a 3/8-inch size and leached in dilute HC1 solution. The solution then is processed to recover LiC1 and KC1 for return to the cell. Dragout of electrolyte varies with the titanium crystal size and ranges from about 1/2 lb per lb of titanium for coarse crystals to about 1 lb per lb of fine titanium crystals. The leached and washed metal is dried and passed over a magnetic separator, and metal fines are removed by screening to about 80 mesh (0.007 in. or 177~). D-H Titanium metal is extremely low in 02, N2, C, Fe, H2, and C12, and it has a Brinell hardness of 60 to 90. TDMET Electrowinning Cell TDMET has operated pilot-plant electrowinning cells since 1956. Later models produced 800 to 900 lbs of titanium metal in one cathode deposit. The cell uses a central metal basket cathode with several internal vertical rod cathodes away from the wall. Graphite anodes are peripheral to the basket. Operation is cyclic. TiC14 initially is fed at a low rate into the center of the basket to form a titanium crystal lattice on the inside of the basket walls. This porous sidewall deposit serves as a diaphragm to keep the lower chloride inside the basket while allowing chloride ions to migrate to the anodes. TIMET also developed a mechanical system for withdrawing the large cathode deposits into an inert-gas-filled chamber, installing a new cathode, and reclaiming the inert gas for reuse. No expansion of electrolytic metal production has been announced by TIMET. U.S. Bureau of Mines Electrowinning Cell This cell (not described in the 1974 NMAB report) used a ceramic diaphragm around the graphite anode as part of a replaceable anode assembly. A separate cathode feed tube introduced the TiC14 below the electrolyte level in the cell. Frequent failure of the diaphragm proved troublesome. High-purity metal with a Brinell hardness of only 70 was made consistently although a small percentage of minus 60-mesh crystals in the deposit was much harder (Leone et al. 1967~. Other Sponge-Winning Technologies The numerous alternative procedures for preparing titanium crystals, sponge, powder, or alloys (particularly TiA13) discussed in the 1974 NMAB report as having been examined or as being of conceptual interest in the future have received little or no additional research and development
OCR for page 51
51 o f the total Kroll and Hunter Btu requirements and about the same as the 244 million Btu (now quoted by the industry at about 180 million) requirement for aluminum metal made by the Hall process reported by Battelle (1975~. Examples of energy use of other metals ~ Battelle Columbus Laboratories 1975) are 358 million Btu per net ton for magnesium, 112 for copper, and 65 for zinc. Prices of these metals, and of aluminum, range from 61.25 to t0.41 per lb. Assuming an average energy cost of $2.00 per million Btu, the energy component of the price of these metals ranges from 32 percent for aluminum to 13 percent for copper. In contrast, with titanium sponge selling for 67.22 per lb, the energy component of the price is only 6 percent for Knoll metal, 5 percent for Hunter metal, and 3 to 5 percent f or D-H metal. Appraisal of Titanium Sponge Production Technology Intricate trade-offs are involved in making a choice between using vacuum distillation, inert gas sweep with leaching, or leaching alone as a finishing operation for Kroll-process sponge. Distillation produces sponge that is lowest in magnesium, magnesium chloride, and hydrogen. It also is most efficient in recovering magnesium and magnesium chloride for re-use. However, it is energy-intensive and ties up expensive retorts over a prolonged time cycle, and there reportedly is some difficulty in crushing and grinding the low-chloride sponge. The inert gas sweep followed by leaching reportedly produces sponge somewhat higher in volatiles than does distillation. Also, it recovers less magnesium and magnesium chloride and requires the disposal of a small quantity of waste bri net However, it is less energy-1 ntensive than distillation because it uses longer batches and the retort cycle time is shorter. The all-leaching process yields sponge that is somewhat higher in volatiles than the sweep-leach procedure. The all-leaching process has no retort cycle t ime f or removal of chlorides and would appear to require the least energy. This is counteracted by its high loss of magnesium and magnesium chloride that translates to a high energy requirement for make-up magnesium and chlorine. There also is the disposal of a large quantity of waste brine to be considered. Moreover, the process also may be more vulnerable to nitride inclusions that require triple melting to minimize lo~density inclusions in ingots. Chlorides, reductant metals, and hydrogen remaining in leached sponge can be removed during vacuum arc melting. TIBET, as an integrated producer, has installed enough vacuum capacity and pup protection to cope with a high level of volatiles during arc melting. In fact, TIMET considers its f irst melt operation as part of sponge ref ining.
OCR for page 52
52 Nonlntegrated producers generally have melting furnaces designed for low~volatile foreign sponge. Such furnaces can melt high-volatile sponge but at a reduced rate and with added vacuum pumping and maintenance requirements. (This subject is discussed in greater detail in the next section of this chapter. ~ Launching of the D-H Titanium Company's electrowinning demonstration plant at a scale of up to 1 million lbs per year is a major investment in new technology. However, full production success and the economic competitiveness of direct electrowinning of titanium are still to be established. Substantial add-on sponge production by U.S. Kroll and Hunter plants in the past two years can be taken as confirmation of the operators' expressed beliefs that the competitiveness of Kroll and Hunter technology is not threatened by direct electrowinning, at least in terms of costs for expansion of existing plants. Establishment of completely new (greenfield) plants, however, may well be based on new or improved technologies such as vacuum distillation or electrolytic processes. A trend toward increasing the batch size in the Kroll process serves to improve metal quality by shrinking the ratio of pot surface to sponge weight and reduces handling costs as at OREMET. Kroll-type metal was first produced domestically in large batches with vacuum distillation in 1981 at Teledyne Wah Chang Albany and a similar operation began early in 1982 at ITI's Moses Lake, Washington plant. More efficient production of titanium, with a possible price reduction, should stimulate the use of titanium in traditional as well as newer sectors as noted in Chapter 10. U.S. sponge production technology could profit from development of new concepts for: improved leaching or other means of purification of the sponge; recovery of anhydrous magnesium chloride from sponge leach brine; and direct preparation of (or conversion of sponge to) high-quality, low-cost, titanium powder for powder metallurgy. A continuous sodium-reduction process might be a possible route to production of powder. Ti tanium Sponge Quality and Specif ications Technically, the ideal specification for raw titanium would be 100 percent titanium. This would allow maximum scrap use plus master alloy additions and melting to any desired ingot chemistry. The iodide process could produce 99.9 percent or better titanium but at uneconomical cost. Each of the currently economical processes--Kroll, Hunter, and, hopefully, electrolytic--has its own characteristic impurities. These can be categorized into two groups: nonvolatiles like oxygen and iron and those that boil off during melting, notably sodium and sodium chloride in Hunter sponge and magnesium and magnesium chloride in the Kroll product. Each of the processes (Kroll, Hunter, and electrolytic) involves growing the titanium dendrites in molten chloride baths. The
OCR for page 53
53 f ree portions of these baths subsequently can be drained and then volatilized under partial pressure, leached, or vacuum disti lied completely away f ram the titanium. However, traces of the chloride baths, which alsp may contain dissolved, stoichiometrically excess sodium or magnesium, invariably are trapped among the dendrite branches of the titanium crystals as they grow and interlock. These traces thereby become hermetically encapsulated to the extent that neither draining, helium sweeping, leaching, or even vacuum distillation can remove them completely. These irremovable residues then cause significant problems in subsequent vacuum arc melting and in the arc welding of consolidated powder metallurgy products (as described in the direct-powder consolidation processes covered in Chapter 11~. Even though neither leaching nor vacuum distillation can remove all volatiles, vacuum distillation does remove considerably more than does leaching. This significantly simplifies and improves subsequent vacuum arc melting. Worldwide, therefore, vacuum distillation has become the preferred by-product removal process. It first was piloted by the U.S. Bureau of Mines and then commercialized by du Pont in 1948 and by the Japanese and Soviets in the 1950s and 1960s. Teledyne Wah Chang Albany was the second U.S. manufacturer of vacuum distilled titanium sponge and ITI now also vacuum distills all its sponge production. Electrolytic titanium crystals apparently encapsulate less chloride~etal bath traces than either the Kroll or Hunter processe s . With further development, leached electrolytic titanium may rival the low-volatile, vacuum-distilled product . All of the volatiles boil of f during vacuum melting and, therefore, are not present to affect the properties of mill products made from ingots. Their presence, however, does adversely affect the melting process both in the casting of ingots and in the making of arc welds in powder metallurgy products. In the vacuum arc melting of sponge to ingots, melting ef f iciency is reduced by the evolution of the volatiles. Volatiles also condense on the mold and mar the ingot surface, increase furnace maintenance, and require that much larger vacuum pumping systems be used. Nevertheless, using pioneer facilities built in the 1950s, it was found economical to employ vacuum melting with steam ejectors to back up large vacuum dif fusion pumps as the f inal refining step for acid-leached sponge. lrhe refining was an integral part of the melting and casting of the ingot. Such re f ining-melting f urnaces served well to start the industry . The process, however, wastes magnesium and chlorine and requires their safe ~ and increasingly expensive) disposal. Escalating costs of magnesi um and thermal energy also have shifted the economic advantage to larger reactors that conserve and apply the heat of reaction to vacuum distillation and economical recovery of magnesium and magnesium chloride. A further consideration of national concern is that the independent U.S. melters (who do not manufacture but rather purchase their sponge as noted in Chapter 3 ~ are not equipped to melt high-volatile sponge. The
OCR for page 54
54 independents, therefore, greatly prefer and have largely confined their purchases to Japanese and Soviet low-volatile, vacuum-distilled sponge. Since the independent U.S. melters have almost one-third of the nation's titanium melting capacity, and since the U.S. National Stockpile contains only high-volatile sponge, the U.S. titanium industry is not properly positioned with respect to melting for a national emergency. (Stockpile specifications are discussed in this regard in Chapter 6. ~ The melting that removes essentially all volatiles f ram even high-volatile titanium sponge functions similarly when consolidated forms made from titanium sponge powder are arc welded--the chlorides disrupt the welding arc. This adverse feature has blocked the commercialization of tonnage titanium powder metallurgy for the past two decades (see Chapter ll) There are two public U.S. specifications for sponge titanium; American Society for Testing Materials (ASTM) B-299, and the U.S. National Stockpile Purchasing Specification P-97-R6 (the former is being updated). The salient features of the sponge grades defined in these specifications and correlations between grades and types are shown by the data of Table 3. Some of the features of Japanese and Soviet sponge are compared with U.S. specification data in Table 4. Detailed excerpts from ASTM B-299 and Stockpile Specification P-97-R6 are given in Appendix H. The U. S. National Stockpile Specification shows that there is both premium-grade (lB-O) and standard-grade (lA-O) magnesium-reduced, vacuum-distilled metal. Compared with the standard grade, the premium grade is lower in iron (0.05 versus 0.12 percent), oxygen (0.07 versus 0.10 percent), and Brinell hardness (100 versus 120). Compared with magnesium-reduced, leached sponge, the standard-grade distilled sponge is lower in carbon (0.02 versus 0.025 percent), magnesium (0.08 versus 0.50 percent), chlorine (0.12 versus 0.20 percent), and hydrogen (0.005 versus 0.03 percent) but is nigher in iron (0.12 versus 0.10 percent). Compared with sodium-reduced leached sponge, the standard-grade distilled sponge is lower in chlorine (0.12 versus 0.20 percent) end hydrogen (0.005 versus 0.05 percent) but higher in nitrogen (0.015 versus 0.010 percent) and iron (0.12 versus 0~05 percent). The differences in oxygen and iron contents and in Brinell hardness between the premium and standard grades in Kroll-process sponge are believed to be a function of the location of the sponge in the reactor pot. As has been noted, the interior sponge is lowest in iron, oxygen, and Brinell hardness. It is understood that Soviet, Japanese, and U.S. manufacturers use center-region sponge for premium-grade sponge. Sodium-reduced sponge characteristically is considerably lower in iron than magnesium-reduced sponge. This is principally due to iron in the electrowon magnesium charged into the reduction part. The magnesium usually contains 0.03 to 0.04 percent iron that transfers to the titanium sponge.
OCR for page 55
55 TABLE 3 Nomenclature and Features of Selected Sponge Titanium Grades Defined in ASl~l B-299 and National Stockpile Purchase Specification P-97-R6 Items Nomenclature and Features Grade lA-O LA-O lA-O lB-O P -9 7-R6 Type A 8 C A B-29 9 Type MI)-120 ML-120 SL-120 M1)-120 Produc Lion f eature s, reductant and finishing operation Mg reduced Mg reduced and vacuum and acid dis tilled leached or inert gas sweeping Na reduced Mg reduced and acid and vacuum leached distilled Selec ted im puri ties content Fe 0 .12 0 .10 0.0 5 0. 05 max. weight O 0.10 0.10 0.10 0.07 perc ent Volatile, max. weight percent Mg 0.08 0.50 0.08 Na -- -- 0.19 - C1 0.12 0.20 0.20 0.12 H 0.005 0.03 0.05 0.005 H2O 0.02 0.02 0.02 0.02 Total 0.225 0.75 0.46 0.225 Nominal titanium content, weight percent 99.3 99.1 99.3 99.3 The ASTM specification is slightly more liberal with respect to impurities. Except for industrial use requiring optimum corrosion resistance or for rotating parts in aerospace engines, most of the sponge meeting the ASTM specification is suitable for a wide range of applications provided it is properly melted, alloyed, and processed. Premium-grade sponge, including interior-location, magnesium-reduced sponge and sodium-reduced sponge, best meets the requirement for low iron content essential for corrosion resistance. Although ingot and fabrication shops that do not produce their own sponge prefer to buy low-volatile foreign sponge that is easier to melt in times of sponge shortage, domestic high-volatile sponge has been melted in these shops with some penalty in melting rate and added equipment maintenance . Sodium-reduced leached sponge i s the easiest to grind f or powder metallurgy applications; however, its relatively high volatiles content limits its use in some powder applications.
OCR for page 56
56 TABLE 4 Comparison of Selected Grades of Japanese and Soviet Sponge Titanium with U.S. Specification Data It ems Specif ication No. Conforms to: P -9 7-R6: Or ade lB-O lA-O Type A A B-299 'Type MD-120 Production f eatures Se lected Impurities FE Content O (weight percent ~ N Nomenclature and Feature ~ , Japanese U. S. Specif ication Soviet B-299 and P-97-R6 MRTU-14 lB-O lA-O* A A MD-120 MD-120 MD-120 Magnesium reduced and vacuum distilled 0.02 -0.05 0.12 max. 0.05 max. 0.07 max. 0.04 -0.05 0.10 max. 0.07 max. 0.04 max. 0.005-0.008 0.015 max. 0.015 max. 0.02 max. Chlorine as a volatile (weight percent) 0.07 - 0.09 0.12 max. 0.12 max. 0.08 max. Average hardness as once melted. (BHN or max. BHN) 97 - 99 120 ~ 100 (100) Nominal Ti content (weight percent) 99.8 99.3 99.3 NA *Conforms to P-97-R6, Grade lB-O, Type A, in most categories. As indicated by the lack of specification coverage, titanium produced by the electrolytic process is not included in either of the public specifications; indeed, there is a dearth of public experience in the matter of defining the quality and characteristics of electrolytically won titanium sponge. If the production of such material becomes sizable, the qualification testing that leads to specifications would be conducted. Such testing almost certainly would show the suitability of the material for the wide spectrum of uses where thermal-chemically reduced sponge grades now are applied (unpublished preliminary evaluations by D-H titanium already have indicated this to be the case).
OCR for page 57
57 The sponge titanium used in the United States to make ingot is covered under the specifications whether it is of domestic or foreign origin and has been ranked according to the degree of melting difficulty in terms of the quantity of volatiles that boil off during melting. The sponge grades with the fewest volatiles generally are the least difficult to melt. The ranking in order of least to most difficult to melt is listed in Table 5. TABLE 5 Ranking of Sponge Titanium in Terms of Ease of Melting . Sponge Description Ranking _ ASTM Grade ~ _ Vacuum distilled, Soviet Least difficult Vacuum distilled, Japanese Vacuum distilled, Chinese Inert gas sweep, U.S. Producer Leached sponge, British Leached sponge, U.S. Producer Leached sponge, U. S. Producer 1 1 Most difficult MD-120 MD-120 MD-120 ML-120 SL-120 SL-120 ML-120 Specifications are most important in the ultimate utilization of materials, especially those like titanium i n which trace impuri ties often have important effects on processing characteristics and on the physical properties of end products. The effects of volatiles in titanium sponge on melting and welding were discussed above. The effects of trace and alloying elements on the properties of titanium metal are described in Chapters 7 and 8. Chapter 6 discusses, among other aspects, the key role of specif ications with respect to the U. S. National Stockpile of titanium. REFERENCES Battelle Columbus Laboratories. 1975. Interim Report on Energy Use Patterns in Metallurgical Processing. Columbus, Ohio. May Cobel, G., J. Fisher and L. Snyder. May 1980. Electrowinning of titanium from titanium tetrachloride. Paper presented at the 4th International Conference on Titanium, Kyoto, Japan. Leone, O . Q., J. Knudson and D . Couch. 1967 . High purity titanium electrowinning from titanium tetrachloride. JOM Vol. 20, 18-23. National Materials Advisory Board Committee on Direct Reduction Processes for the Production of Titanium Metal. 1974. Report NMAB-304 , Washington, D. C.: National Academy of Science s. Tukomoto, S., E. Tanaka and K. Agisu. 1975. The deposition of titanium metal by fusion electrolysis . JOM, Vol. 28, 18-22 .
OCR for page 58
Representative terms from entire chapter: