The present invention is directed to the production of electrical devices, e.g. electrolytic capacitors and superconducting materials based on the use of valve metals such as tantalum and niobium and its alloys. In work done previously by the assignee of this patent application there have been described developments wherein filaments of refractory metals such as tantalum and niobium are extruded in a softer metal such a copper while encased in a surrounding sleeve of the valve metal e.g. tantalum or niobium. Utilizing this technology filaments of either tantalum or niobium have been drawn to very small diameters while in the copper matrix and while surrounded by the constraining layer of tantalum or niobium. The copper is leached out from between the very fine filaments in the final product while the fine filaments are constrained in a manageable bundle by the outer constraining acid resistant layer of tantalum or niobium. This technology is more fully described in our prior parent U.S. Pat. No. 5,869,196 and our PCT Application Serial No. PCT/US03/24724, filed Aug. 7, 2003.
In the present invention there is an improvement over the processed described in the '196 patent above wherein the constrained mass of small fine tantalum or niobium filaments is subjected to a rolling deformation which flattens the individual small filaments so that each wire has aspect ratio of at least 5:1. This relatively high aspect ratio in the final compressed and flattened individual fine filaments has the following major advantages. When the Nb is reacted to form Nb3Sn, currently, round Nb filaments require in excess of 180 hours at 700° C. to completely react 4-micron filaments. The same filament rolled to an aspect ratio of 5:1 will be only 1.26μ thick by 12.6μ wide and can be fully reacted using considerably shorter times. By reducing the time and distance for reaction, finer grain size and more importantly, uniform Nb3Sn are obtained and as a result, higher current densities as well.
A flat conductor is easier to wind into a magnet and it has higher filling factor where the void space when a round conductor is used is eliminated. Another important consideration is that thin conductors are more ductile and can be wound into magnets of small diameters. A reacted Nb3Sn; flattened to an aspect ratio of 5:1, with a thickness of 0.02 mm thickness by 0.02 mm wide can be bent around a radius of 5 cm without fracturing the wire. This offers the possibility of reaction and winding, similar to present day ductile NbTi magnets.
An unexpected discovery has been that the improvements for superconductor have been shown to be even more important for capacitors. The filament size, its number, spacing and surface area/unit volume are exactly the same critical designs parameters for both applications. In the case for superconductors, the entire Nb filament is converted to superconducting Nb3Sn. For capacitors, the surface of the Nb filament is converted to Nb205, the dielectric compound for capacitors use. In fact, with only minor size and shape adjustment, the wire can be used as either a superconductor or as an anode for solid Nb electrolytic capacitors. Confirmation of the dual use possibility of this invention are further demonstrated by two recent publications, Alternative Materials for Electrolytic Capacitors by Dr. K. Reichert, T.I.C. Bulletin, No. 109 March 2002 page 2-3, and Powder for Capacitor, Sintered Body Thereof and Capacitor Using the Sinter Body PCT/JP01/10484 OMORI, Kazuhiro, 6 Jun. 2002, where Ta and Zr alloys of Nb, used for superconductors, has been shown to have excellent capacitor properties as well.
In the preferred embodiment of the present invention for making capacitor anodes, the process begins with niobium or tantalum rods that are inserted into holes drilled longitudinally into a copper billet, shown schematically in
The primary billet, containing the niobium or tantalum rods in a copper matrix, is processed in accordance with the flow-chart of
At restack diameter, the composite wire is cut into lengths for assembly into the secondary billet. The secondary billet transverse cross section is shown schematically in FIG. 3a. The subelements 4 made from the primary billet are stacked together with copper rods. The copper rods are used to form a copper core 5, and an outer annulus 6. Both the core and the outer annulus are provided in order to make leaching of the final composite less difficult. Outside the assembly of the subelements and copper rods is a layer of tantalum sheet 7. The sheet is the same length as are the rods and it completely surrounds the filament array. The sheet thickness is comparable to the diameter of the niobium or tantalum filaments within the subelements. Outside the cylinder of niobium or tantalum sheet is an outer copper can 8.
The secondary billet is assembled, a nose and tail are welded into place, and the billet is evacuated and sealed. The sealed billet is optionally prepared for extrusion by hot or cold isostatic pressing in order to collapse any void space within the billet and to promote filament uniformity. After isostatic pressing, the secondary billet is machined to fit the extrusion liner. The billet is then extruded at elevated temperature at a diameter reduction ratio of 6:1.
The extruded rod is cropped, and the rod is then drawn to a diameter where the niobium or tantalum filament diameter is 5 microns or less. Again, as indicated in
The cut sections are immersed in a solution of nitric acid and water. A suitable solution would be one part nitric acid to one part water, but other concentrations of nitric acid can be employed if required. The sections are immersed for a period of time sufficient for the acid to fully leach out the filaments and the sheath. The total time will depend primarily upon the composite wire diameter and length, with smaller diameters and greater lengths requiring longer times. This is due to the fact that the acid can only penetrate through the ends of the cut sections. Narrow openings and long distances do not lend themselves to rapid etching.
In regard to the leaching process, it is an essential feature of this embodiment of the present invention that the secondary billet is provided with a copper core 5 and a copper annulus 6 (
After leaching, one is left with the product of the present invention, shown schematically in
Depending upon the degree of electrical continuity and the level of purity within the product of the present invention, it may not be necessary to sinter the product. If sintering is avoided, the process of the present invention will be less expensive. The decision on whether or not to sinter the product will depend primarily upon the requirements of the application.
Another embodiment of the invention employs the secondary billet illustrated schematically in
Yet another embodiment of the invention employs the secondary billet illustrated schematically in
Yet another embodiment of the present invention employs the secondary billet illustrated schematically in
In other embodiments of the invention, all or part of the sheath is made to be perforated or porous so as to accelerate the leaching process while still maintaining the effectiveness of the sheath with regard to constraint of the enclosed filaments.
Patent '196, specifically describes the inclusion of a separate outer sheath of Tantalum or Niobium to further constrain the filament bundle. A major advantage of both processes is illustrated in
The ability to shape the anode to almost any shape or form provides the basis for capacitor application. The papers as listed below are relevant to this application.
McLean shows that the A.C. properties of porous tantalum solid electrolytic capacitors are frequency dependent and that the capacitance decreases and internal resistance or ESR increases with increasing frequency. Theoretically, an ideal capacitor would operate with essentially zero ESR. As the ESR increases, heat is generated, possibly leading to capacitor failure, and electrical circuit problems. In Bourgault's equation 9, the dissipation factor or ESR is proportioned to the square of the radius of the capacitor. In other words, the efficiency of the capacitor is size dependent. He concluded that a thin wall anode or hollow tube was the desired shape and as a result, U.S. Pat. No. 3,345,545 was issued. The concept of an ideal anode shape to improve capacitors performance was also confirmed in the paper by Hluchan. An ideal anode would be one where the thickness is that of a foil, somewhere about 25 to 100 microns and a width of approximately 2-10 mm wide. The foil should be porous, and contain a high specific surface area/volume. Most importantly it must be robust and capable of being physically handled when subjected to the processing step necessary to produce a finished capacitors. It has been found in our experiments that the chip anode shape of
Metallographic studies of the cross-section show that the decreasing thickness, the filament are also flatten and its aspect ratio increases to as much as 10:1, measuring approximately 2μ×20μ wide. It is obvious that the rolling process has increased the available surface area of the filaments significantly and at the same time provide the ideal shape for AC application.
Handling of these highly aspected thin foil can be more difficult after leaching. Some structural support appears to be necessary. This can be accomplished by a simple change in billet design. In place of a uniform array of rods as shown in
Copper removal is more difficult for these highly compacted foils. The flattened filaments are extremely close together and may even be touching. To insure complete copper removal, a continuous liquid magnesium dipping process can be used to leaching the copper from the foil. Copper is completely soluble in liquid magnesium while Ta and Nb are insoluble. A similar process, used to leach the copper matrix with liquid Sn was shown in our aforesaid U.S. patent application Ser. No. 09/532,362. The magnesium matrix can then be easily removed by a combination of vacuum distillation and an acid clean step.
To increase the total capacitance of this foil shape, several foils can be arranged much like that of the MLCC (Multi-Layer Ceramic Capacitor) where layer are stacked together and electrically connected in parallel. This concept is also shown in paper by J. Gerblinger, given at the Symposium on Tantalum and Niobium, San Francisco held Oct. 22, 2002 titled Tantalum Capacitors Today and Tomorrow, pages 329-342. See, in particular,
This application claims priority from U.S. Provisional Application Ser. No. 60/434,039, filed Dec. 17, 2002, and is a continuation-in-part of co-pending application Ser. No. 10/282,354, filed Oct. 29, 2002, which is in turn a continuation-in-part of co-pending application Ser. No. 10/037,935, filed Jan. 2, 2002, which is in turn a continuation-in-part of Ser. No. 09/753,200, filed Jan. 2, 2001, now abandoned, which is in turn a continuation-in-part of U.S. application Ser. No. 09/532,362, filed Mar. 21, 2000, now U.S. Pat. No. 6,543,123 issued Apr. 8, 2003. The subject matter of all of these parent applications is incorporated herein by reference.
Number | Date | Country | |
---|---|---|---|
Parent | 10736650 | Dec 2003 | US |
Child | 11468632 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 10282354 | Oct 2002 | US |
Child | 10736650 | US | |
Parent | 10037935 | Jan 2002 | US |
Child | 10282354 | US | |
Parent | 09753200 | Jan 2001 | US |
Child | 10037935 | US | |
Parent | 09532362 | Mar 2000 | US |
Child | 09753200 | US |