Methods for Recovering Refractory Metal From Wheel Grinding

TECHNICAL FIELD

[0001] Recovery of refractory metal, especially titanium and/or zirconium, from wheel grinding.

BACKGROUND OF THE INVENTION

[0002] Titanium, zirconium and other refractory metals and their alloys have been produced in commercial quantities since the 1950's. The production processes involve extraction of the metal from the ores in the form of sponge or crystals, which are blended with any desired alloying ingredients and melted into ingots by several alternate furnace processes. Since these metals work harden quickly, the only method of forging the ingots to useful shapes for further processing such as billets, slabs, rounds, etc. is to heat them above their ductility temperature, which for these refractory metals is well above the temperature where they react with air to form oxides, some nitrides, and, if oil or carbon is present, carbides.

[0003] Normally, the ingots are heated in gas fired furnaces with the result that the forged products have a thick layer of refractory metal oxide (and/or nitride) on their surfaces frequently in the range of 3 to 15 mil thick including, in the case of titanium, TiOz, Ti ₂ O ₃ , and titanium with a high level of oxygen solid solution. These oxides (and/or nitrides) are very hard and are typically removed after forging and prior to downstream operations such as rolling into sheet or coils. From the onset of the refractory metal industry, steel industry practices were adopted, which usually include dry wheel grinding to remove the oxidized surface and surface cracks or defects, if any, in the forged or cast shapes.

[0004] Dry wheel grinding in the refractory metal industry is the action of a rapidly turning (typical surface speeds of 12,000 feet/min or so) abrasive wheels, against the forged product to remove the oxide from the surface and grind out cracks, if present. The abrasive wheels are commonly about 3-4 inches (in.) wide and 24 in. in diameter and contain a hard abrasive like ZrO ₂ -Al ₂ O ₃ , corundum, or SiC grains cemented into a wheel form typically with phenolic resins.

[0005] The grinding action gouges material from the oxide (and/or nitride) and metal in the form of elongated splinters with aspect ratios of 5: 1 to 10: 1 from the outside of the forged shape and flings it away from the rotating wheel/forged shape interface.

[0006] The energy involved in gouging metal particles or splinters from the surface of the forged shape and their high surface area is such that they burn nearly completely as they pass through the air, generating copious sparks and large quantities of dust, which is typically collected in a large bag house.

[0007] Since the beginning of the refractory metal industry, dry wheel grinding has been predominant in removing the cracks and oxide layer from forged forms, and wheels have been optimized to give the high stock removal rates desired at the high temperatures of the grind wheel/forged shape interface. This optimization involves using coarse abrasive grit of a 4/6/8 mesh mixture and a hard phenolic resin that degrades at the right rate to create new abrasive surfaces which maintain the grinding action.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] In the above described prior art processing, the value of metal removed from the forged surface is lost. This value is considerable as the cost of the refractory metal ore is small compared to the cost of converting it into expensive consolidated refractory metal forms. Consequently, oxide and/or nitride removal techniques that allow metal recovery may provide an improvement.

[0009] In abrasive wheel grinding operations other than dry wheel grinding, it has previously been known to direct a stream of coolant at the point of contact of the abrasive grinders on the metals being ground. Frequently, it is done in wheel grinding steel forms that require close dimensional control. By eliminating heat and the resulting expansion of metal, closer dimensional tolerances can be met.

[0010] In downstream processing of refractory metal products, such as sheets and coils where no residual oxide layer can be tolerated, wet belt grinding is frequently practiced where a high speed abrasive belt removes surface oxide and/or defects under water jets which cool the contact area and flush the swarf from the area. Much of the refractory metal swarf particles removed from these products are unoxidized. Megy has described in a series of three patents (U.S. No. 5, 171 ,359, U.S. No. 5,597,401, and U.S. No. 5,776,225) how to recover from the wet roll grinding swarf relatively pure metal values that have commercial uses.

[0011] In the Fifteen years since the Megy patents were reduced to commercial practice, it has not occurred to anyone to attempt to modify the practice of dry grinding refractory metal forged shapes in an attempt to recover relatively pure metal values from the swarf thus generated. This may be due to the very long period dry grinding has dominated both the precursor steel industry and the refractory metal forge industry, and the fact that dry wheel grinding is quite different from wet belt grinding. The operations have very different operating requirements.

[0012] For example, the operator typically views the point at which the grinder engages the forged shape to be able to grind out cracks in the dry wheel grinding case. The dry wheel grinding has much higher removal rate using relatively coarse wheels, which have been optimized for high metal removal rates. If a dry grinding wheel is water cooled, then it may use a different optimization for maximized grinding rates. It is not expected that a grinding rate as high as dry grinding can be obtained, but it remains to be tested.

[0013] The swarf which forms with the coarser abrasive grit size has considerable different physical characteristics than when produced under a flow of coolant with much smaller abrasive grit size in wet belt grinding. For example, the particle size of the splinters produced are much larger than the splinters in wet belt grinding swarf, which uses abrasives typically between about 80 and 200 mesh cemented on to a belt. While dry grinding discussed in the Background section above produces elongated splinters with aspect ratios of 5: 1 to 10:1, wet belt grinding produces long stringy splinters with aspect ratios on the order of 100:1. Also, the abrasive grit that contaminates the swarf is much larger in dry grinding.

[0034] Another important issue is that the swarf generated under a coolant stream may be pyrophoric and safe handling may present a challenge in a forge shop. The safe handling of swarf may be accomplished by storing either entirely under water or in sealed steel drum with greater than 25 weight percent (wt%) water. This practice is not currently observed in a wheel grind shop or the associated forge plant.

[0015] One aspect of the invention includes modifying existing dry grinding operations by directing one or more streams of coolant at the abrasive wheel/forged form interface. The coolant may be water or water containing additives which are known in the wet grinding industry for increasing the life of the abrasive, enhancing cutting, inhibiting corrosion, etc. The coolant flow rate may be well in excess of a flow rate sufficient to supply the evaporation rate that removes heat from the abrasive wheel/metal interface.

[0016] Accordingly, another aspect of the invention includes a method of recovering usable refractory metal values from forged refractory metal shapes, which are conditioned by abrasive wheel grinding by directing a stream of coolant at the wheel/forged shape interface to limit further oxidation of the swarf being produced by the grinding process, and cleaning the resulting swarf using a milling process followed by screening and washing steps.

[0017] Still another aspect of the invention includes a process for producing relatively pure refractory metal fines from material removed from the outside of refractory metal forged shapes. The process includes decreasing the additional oxidation of refractory metal during wheel grinding of the oxide layer from forged shapes by directing a flow of water, with or without additives, at the point of contact of the grinding wheel with the forged shape, at a rate greater than the evaporation rate due to the mechanical action of grinding, followed by cleaning the oxide, nitride, or carbide contaminated fractions from the removed material, and drying the relatively pure product from these operations. The product can be used in manufacture of refractory metal additives to aluminum, steel, and other metals.

[0018] In a further aspect of the invention a refractory metal recovery method includes grinding a refractory metal shape with an abrasive wheel in the presence of an aqueous coolant and applying the coolant at a total rate exceeding a total rate at which the coolant evaporates. An interface between the wheel and the shape exhibits at least two localities from which the coolant

evaporates at different rates. The method includes distributing the coolant across the interface sufficient to apply the coolant at local rates exceeding local rates at which the coolant evaporates from the at least two localities, respectively, during the wheel grinding. The method produces refractory metal swarf from the wheel grinding and decreases oxidation of the swarf compared to dry grinding. Consequently, the swarf may be collected.

[0019] By way of example, the metal shape may consist essentially of titanium and/or zirconium. The wheel may include abrasive grit in a resin exhibiting a wear rate corresponding with a wear rate of the grit. Distributing the coolant may include applying the coolant behind and at both sides of the wheel. The collected swarf may exhibit the property of at least 30 wt% dissolving in molten aluminum in 40 minutes or less. Swarf may be projected against a backdrop in front of the grinding wheel, rinsed from the backdrop, and collected in a catch basin below the backdrop.

[0020] In a still further aspect of the invention, a refractory metal recovery method includes grinding a titanium and/or zirconium billet with an abrasive wheel in the presence of an aqueous coolant. The billet has a mass of at least 3,000 pounds before the grinding and includes an oxidized surface of a thickness and composition similar to that produced during gas-fired forging in a furnace. That is, the oxide may be produced by other than gas-fired forging, but still • exhibit similar properties. The method includes applying the coolant at a total rate exceeding a total rate at which the coolant evaporates during the wheel grinding and removing the oxidized surface from the billet. The method produces metal swarf from the wheel grinding and decreases oxidation of the swarf compared to dry grinding. Swarf is collected and exhibits the property of at least 30% dissolving in molten aluminum in 40 minutes or less.

[0021] By way of example, the grinding may include grinding out surface cracks in the metal under the oxidized surface. The grinding may also include decreasing post-grinding oxidation of the billet compared to dry grinding and decreasing the number of surface cracks in the billet compared to dry grinding. The swarf and removed oxide may include greater than about 0.5 wt% of the billet mass, or further, greater than about 5 wt% of the billet mass.

Example 1

[0022] A 200. horsepower traversing Midwest Grinder from Vulcan Engineering

Company in Helena, Alabama with a 3 in. wide by 24 in. diameter grinding wheel, used for grinding out surface cracks and removing the surface oxide layer from titanium billets, was modified to a water flushed operation.

[0023] Before modification, the system could process approximately one billet per shift that was roughly 2 feet square by 14 feet long and weighed about 15,000 pounds. Depending on the severity of surface cracks, up to 800 to 1,000 or more pounds (about 5 to 7 wt%) of titanium oxide and metal would be removed from the outside of the billet during the conditioning process.

If no surface cracks were ground out, then only about 80 to 100 pounds (0.5 to 0.7 wt%) would be removed. Metal particles, which comprised most of the material being removed, flew tangentially off the wheel as a cascade of sparks, with the metal particles being substantially converted to oxide by the time they cooled and were swept away in the off gas system that required a large bag house to collect the copious dust formed during the grinding process.

[0024] The surface cracks were often elongated during the grinding process. For hard billets, which are less ductile and have higher hardness usually due to increased alloying with oxygen, preheating to 1000 ⁰ C was used to prevent the problem of cracks propagating in the billet. In turn, this further thickened the oxide layer and led to greater yield loss in the process. In contrast, the wet wheel grinding described herein did not require the billet to be heated to limit crack propagation and removed pre-existing oxide without generating much additional oxide during grinding.

[0025] The machine was modified by designing and installing jets directing a flow of coolant both behind and at the sides of the wheel at about 25 gallons per minute (gpm). The swarf exited tangentially from the front of the wheel. The coolant was water. A drip pan was placed under the billet which passed any of the liquid/swarf collection to a catch basin positioned at the back of the billet on the floor. The coolant/swarf mixture left the wheel/metal interface tangential to the spinning wheel into a metal backdrop placed about four feet behind the billet. The backdrop was sprayed with coolant to wash it down into the catch basin where the swarf settled and the coolant recirculated to the jets that sprayed the wheel.

[0026] An about 6,000 pound titanium billet (a "witness billet" with a thick oxidized surface from multiple soak furnace heat up cycles) was used. Three sides of the billet were ground with a standard ZT-104 wheel that was flushed with the 25 gpm of water. For its abrasive, the wheel included zirconia-alumina refractory (nominally 6 mesh) bonded in a phenolic resin and was manufactured by Norton Abrasives of Worchester, Massachusetts. In dry grinding, approximately three of such wheels could be used up processing each billet. The wheels cost about $300 each, leading to a significant wheel cost for a process operation.

[0027] In this Example, when grinding the first two sides, the wheel was rotated at 2500 revolutions per minute (rpm), the same speed used for most of normal dry grinding operations. There was much reduced spark-ing and smoke during the wet grinding, but still a significant amount. Initially, the sparking was estimated at perhaps 1/5 of that resulting from normal dry grinding and the smoke was borderline acceptable without a hood. However, later in the runs, more sparking and smoke was sometimes observed at perhaps 1/3 of normal for dry grinding.

[0028] The time to grind the surface was similar to dry grinding, though the ground surface was significantly smoother. This surprising quality improvement may well prove to be important to possible elimination of a mill processing step for titanium conversion. Also, the surface immediately after grinding was ambient temperature and about the same temperature as

the water flush, as evidenced by feeling the surface immediately after grinding. This fact is expected to be decisive in eliminating crack propagation in hard alloys. Further testing and evaluation of this expectation may occur.

[0029] The refractory wheel was observed to "gum up" as the grinding proceeded.

Gumming was believed to result from resin not being worn down as in dry grinding with its much higher heal exposure. An expectation existed that using abrasive grit with a finer size of about 16 to 24 mesh and a softer resin that wears faster may provide a better match between abrasive wear and resin wear. Definitive conclusions on this point might be reached after further testing. Further testing may occur using other wheels with smaller abrasive size and softer resins for higher wear. The finer grit may also produce a Finer particulate in the product, which may beneficially dissolve more readily in molten aluminum than the larger particles from the present Example.

[0030] The third side of the billet was ground under water flush at 1600 rpm. A half cup sample of titanium swarf was recovered from the holding tank in the system using a broom and shovel after each side was ground. It was noted the sparks immediately went out on impinging on the back wall of the enclosure. A few passes were made on the second side of the billet (at 1800 rpm and no water flushing) after the surface grind with a water flushed wheel (at 2500 rpm) to observe in close proximity the difference in quality of the grind surface in the two operating modes.

[0031] In addition to being rough, the surface produced in dry grinding exhibited a gray color, indicating a relatively thick oxide surface existed on the ground surface. A blue surface was observed on the wet ground billet, which means that the oxide surface was of a sufficient thinness to produce birefringence (reflecting blue light, absorbing red light).

[0032] Advantageously, the forged shape was cooled during the grinding which not only reduced significantly the crack propagation problem, but also allowed precision machining with the grinding wheel and substantially reduced the thickness of the residual oxide layer (often a sub-oxide) compared to that found in dry wheel grinding. Further processing after dry grinding without oxide removal may yield numerous surface cracks. This is particularly detrimental in applications such as the conditioning of rotary forged billets, for example, for the aircraft industry, where a clean surface and precision diameter is desired after the conditioning. Oxide removal in such applications was previously accomplished with a lathe operation.

[0033] However, with the wet wheel grinding, where the swarf has value, the wheel grinding conditioning may have advantages. Observation on the ground billet of a blue oxide from birefringence indicated an oxide layer of only a few atoms in thickness. Accordingly, the oxide layer resulting from the present Example may be so thin that the oxide removal step might be discontinued.

[0034] The operator was able to grind out the cracks in the forged shape without any noticeable crack propagation when the coolant was directed at the wheel while grinding. Presumably, this was due to the elimination of heating the forged shape at the surface of the crack to red hot temperatures, which would have expanded the metal and extended the crack at the relatively cold root within the forged shape. Cracks became much more visible on the smoother surface from wet wheel grinding and were easier to detect and grind out. Consequently, preheating to avoid crack propagation might be discontinued. ^"

[0035] At most 70 wt%, but on average about 60 wt%, of each of the samples collected from the sump following the grinding of each of the three sides of the billet was acid soluble titanium. The acid solution contained about 2.7 wt% HF and about 14.3 wt% HNO ₃ . The acid solubility compares with very little (less than about 5 wt%) of the titanium in dry grinding swarf being acid soluble. Since some of the sample was the abrasive with no titanium in it, and some of it was the relatively thick titanium oxide layer that was on the surface of the billet being ground, it was concluded that residual sparking did not oxidize more than a minor amount of the titanium. The insolubles in the three samples were about 30 wt%, which was undoubtedly the abrasive grit plus the titanium oxycarbonitride in the sample. Of the 60 wt% acid soluble product, about 50 wt% dissolved in molten aluminum at 1400° F within 40 minutes.

[0036] Consequently, the product of the wet grinding might be suitable for use as an alloying additive, even though 80 wt% dissolution in 30 minutes or less represents a goal for commercialization. Using abrasive grit with a finer size, better coolant distribution, and/or coolant additives, as described herein, may achieve the greater dissolution rate.

[0037] Separation of swarf from the abrasive grit and titaniumoxycarbonitride was accomplished by, first, mixing the samples in a laboratory-scale tank with a mixing blade. The mixer breaks up the metal splinters of the metal swarf particles and the hard metal oxide (and/or nitride) components to smaller particles. Soap was added to the mixer to render soluble tramp oil on the swarf, and iron contamination was removed with rare earth magnets suspended in the tank.

[0038] Following the mixer operation, the resulting solution was screened on a two deck unit with a 30 mesh upper screen and a 270 mesh lower screen. The majority of the metal oxide (and/or nitride) components in the mixture reported to the -270 mesh material with only a modest loss of ductile metal particles from the -30/+270 materia). The +30 mesh material contained most of the abrasive grit, which came from the grinding wheel.

[0039] The materia! reporting to the -3Q/+270 stream from the screens was collected in a filter tub and washed with clean water. It was determined to be suitable for further processing into alloying additives for aluminum and steel by verifying both that the surface of the refractory metal particles were sufficiently reactive and substantially free of oxide, nitride, and carbide and that the bulk refractory metal material was sufficiently free of oxide (and/or nitride) contamination and abrasive grit contamination.

[0040] The pilot test described in this Example indicated promise in providing a product suitable as a raw material for alloying additives. The test also indicated that the flushed wheel system may allow operation without a bag house, produce a bulk titanium product with a significantly smoother surface and with less oxide on the ground surface, and allow grinding hard billets without crack propagation.

Example 2 (Hypothetical)

[0041] The sparking and smoke from the unit in Example 1 were much reduced, but were still readily apparent. Copper tubes providing the water jets may be reduced to 3/4 of the 1/2 in. inside diameter used in Example 1 , the length extended, and the number of tubes increased from 6 to 12. With these modifications, the tubes may be more easily bent and water jets redirected to give better coolant coverage on the wheel. Such fine tuning of tube position may reduce the sparking further. It was believed that the 25 gpm of water, in gross, was sufficient, but establishing a geometry such that distributing the given amount of water to be sufficient at any point may be beneficial.

[0042] Decreasing abrasive grit size to 16 to 24 mesh, using a softer resin, and optimizing wheel speed in subsequent testing are expected to help in yielding a product much closer to commercialization goals.

[0043] An additive containing nitrite salt and other additives for rust inhibition and lubrication (e.g., Nalco Calgrind 1541 available from Nalco in Naperville, Illinois) might be suitable to significantly reduce the rate of abrasive wear and may reduce sparking and smoking in subsequent testing, compared to Example 1. The Nalco additive has been used in wet belt grinding. The nitrites rapidly oxidize a titanium surface as it is ground, otherwise, the ground titanium may scavenge oxygen from the abrasives, "welding" the abrasive belt to the ground surface. When a nitrite additive is not used, such phenomenon is believed to significantly reduce the belt lifetime in wet belt grinding of titanium compared to stainless steel. Consequently, the Nalco additive, or the like, may significantly affect reduction of sparking and/or smoking. The grinding aid or surfactant in the Nalco additive, or the like, assisting with lubrication may also help to reduce sparking and/or smoking.

[0044] It was hypothesized that a flushed impingement plate placed on an incline and fastened to the wheel. guard may well reduce the time the sparks burn significantly. Reducing burn time for any sparks that nonetheless appear may also increase swarf yield.

[0045] Swarf handling may be accomplished safely without risking a fire hazard. After the billet is conditioned, the swarf which settles to the bottom of the catch basin may be removed with a non-sparking stainless steel shovel and placed in drums. The swarf may be covered with water and transferred to a mill tank. The mill tank may have a high intensity mixing blade with tungsten carbide teeth, in a manner taught by the aforementioned Megy patents. The high

intensity mixer breaks up the metal splinters of the metal swarf particles and disintegrates the hard metal oxide (and/or nitride) components to fine dust. The processed materiai may be screened and washed in the manner described in Example 1.

[0046] No scales were available to weigh the grinding wheel before and after the grinding in Example I. Example I also involved grinding a billet with an atypical oxide coating. Therefore, no data was developed on wheel life. Further testing may investigate wheel life.

[0047] Observation in Example 1 indicated that the paper gaskets holding the grinding wheel in the steel axle degraded in the presence of the water flush. Another gasket with waterproofing or perhaps a plastic may be warranted.

[0048] The improvements mentioned above may increase the robustness and operability of the system in the event of commercial operation.

[0049] The invention has been described in language more or less specific as to structural and methodical features. It is to be understood, however, that the invention is not limited to the specific features shown and described, since the means herein disclosed comprise preferred forms of putting the invention into effect. The invention is, therefore, claimed in any of its forms or modifications within the proper scope of the appended claims appropriately interpreted.