Capacitor Powder

Abstract

Powder for making electric capacitor or the like comprising a ternary niobium-zirconium-titanium alloy. The alloy is selected as to composition and treated to produce and retain the beta (body-centered-cubic) phase. The resultant product affords high capacitor stability and low leakage approaching the characteristics of the more expensive tantalum at a capacitance cost comparable to or better than that of niobium.

Parent Case Text

This application is a continuation-in-part of U.S. application, Ser. No. 882,482, filed Dec. 5, 1969 now U.S. Pat. No. 3,597,664 issued Aug. 3, 1971.

Description

This invention relates to electric capacitors, particularly electrolytic capacitors and materials used in making them.

BACKGROUND

High performance capacitors are utilized in a wide variety of radio, television, computer, telephone and other electric circuits. The principal high performance material in the present state of the art is tantalum. It is a high temperature, highly corrosion resistant metal capable of forming a highly stable oxide film, of high dielectric constant, at its surface which serves as the capacitor dielectric.

With the current increased demand for tantalum and relative scarcity of world sources for tantalum ore, the need for a substitute has been of great importance to users of high performance capacitors. The more abundant, less expensive, metal aluminum is a possible substitute. But it cannot be formed into porous slug type of capacitors as readily as tantalum. Furthermore, aluminum oxide has low dielectric constant giving about one-third the capacitance of tantalum on an equivalent volume basis. Another obvious candidate as a tantalum substitute is niobium. Niobium oxide has a higher dielectric constant than tantalum and niobium metal can be produced at less cost. Niobium powders can also be utilized to produce porous slugs for electrolytic wet and dry capacitors. The U.S. Government and leading capacitor and materials manufacturers have therefore devoted intense research effort to niobium and its alloys (and also to titanium and zirconium alloys) to provide a tantalum substitute. The results of these efforts are reported in articles or reports located at:

A. Journal of the Electrochemical Society: vol. 108, pp. 343, 750, 1,023; vol. 100, p. 69; vol. 110, p. 1,277; vol. 111, p. 1,331; vol. 113, pp. 100, 1,048 (see also vol. 114, p. 145)

B. Journal of Electrochemical Technology: vol. 1, p. 93; vol. 2

C. Government Contract Reports: AD618055, AD431898

d. U.S. Pat. Nos. 3,126,503; 3,278,344; 3,321,677 (niobium-zirconium-titanium alloys) 3,203,793

E. Canadian Pat. No. 709,982

None of the work has produced a tantalum substitute which is in wide use at the present time although some of the resultant products were in commercial use for a time.

It has been apparent from the above published work that wet and solid electrolytic niobium capacitors are not as good as tantalum electrolytic capacitors in respect of leakage, capacitance and dissipation stability, especially at elevated temperatures.

Niobium can nevertheless be used for low capacitance-voltage ratings and alloyed with tantalum for use at higher capacitance-voltage ratings but cannot provide the desired substitute for tantalum for substantially all purposes, including cost.

OBJECT

It is the object of the present invention to solve the problem of providing a substitute capacitor material satisfying the purposes of the substantially unsuccessful development campaign of the prior art.

GENERAL DESCRIPTION

As in the prior art, the inventive effort is focussed on the material which comprises the principal electrode of the capacitor (generally anode) and forms a surface oxide which serves as capacitor dielectric.

First, if it is assumed that high leakage and low breakdown voltages of niobium (in the absence of gross impurities) are a result of damage inflicted by locally high temperatures and currents, it would be necessary to reduce these and this is possible through alloying niobium with a metal which forms a more refractory oxide. Young's book, Anodic Oxide Films (Academic Press, 1961) indicates an inverse relationship between dielectric constant and ionic conduction of anodic oxide films. A condition for alloying agent candidates is that the alloying agent must have a solid solubility in niobium -- which is true of the Group IVB, VB, VIB refractory metals. Zirconium gives the optimum balance of cost, stability of oxide, solubility. An alloy of Niobium -- 50 atomic percent zirconium when fabricated to a sheet form anode and incorporated in a capacitor (anodized in 0.01 percent phosphoric acid to 200 volts at 25.degree.C) has good leakage and exhibits a change in capacitance on heating to 300.degree.C in air for 30 minutes of less than 10 percent, whereas a niobium capacitor changes by 100 percent.

However, it was discovered that when porous anodes are made from a powder form alloy of niobium-50 zirconium, they will not anodize above 100 volts and that they exhibit high leakage. Difficulty was also encountered in anodizing in that the alloy did not anodize well in aqueous electrolytes and organic electrolytes were too viscous for use in a porous structure.

Photomicrographs of Nb-50 AT. % Zr anodes revealed a two phase metallurgical structure of Zr and Nb rich compositions within the porous anode, apparently resulting from the high oxygen contents and high surface area inherently obtained in powdered materials -- both of which tend to promote instability of a single phase, high temperature structure. It was then conceived to stabilize the beta phase as a new approach to the problem and utilize a third alloy addition for this purpose.

Titanium was chosen because it forms a larger solid solution range with niobium than does zirconium and also lowers the beta and alpha transitions of niobium-zirconium so that the high temperature beta phase could be retained by rapid cooling to lower temperatures. Titanium was also intended to render the overall alloy more readily anodized in aqueous solutions.

It has been found, surprisingly, that the best alloys for porous capacitor anode purposes are formed when the niobium and titanium are present in the alloy in substantially equal atomic amounts, and with a greater portion of zirconium than niobium or titanium. It is also necessary to avoid the titanium-rich and zirconium-rich portions of the ternary alloy system. Some niobium-zirconium rich portions of the ternary alloy must also be avoided. For these reasons, it is important to maintain a single, crystalline structure of controllable composition.

Some care in material processing, as described below, is a necessary adjunct to material selection in order to achieve the beta, single phase crystalline structure in the product.

Other objects, features and advantages will in part be obvious from this disclosure and will in part be set forth hereinafter in this disclosure in the following specific description which is set forth with reference to the accompanying drawings in which:

BRIEF DESCRIPTION OF THE DRAWING

FIGS. 1-5 are microphotographs of sectioned anodes made of compositions and including, by contrast, FIGS. 1A-1C showing anodes of other compositions,

FIGS. 6-7 are schematic cross-section views of capacitors utilizing the invention, and

FIG. 8 is a ternary phase diagram of the niobium-zirconium-titanium system showing data points used in the examples. FIG. 9 represents the DC leakage over the entire ternary phase diagram, and FIG. 10 represents the thermal stability of capacitance over the entire ternary phase diagram.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

Since the anodic film is formed from the alloy, the film's properties are dependent upon the alloy composition. Alloys rich in niobium, titanium, or zirconium do not form coherent or thermally stable anodic films. Alloys containing 20 to 60% Nb, Zr, or Ti are better, but they tend to decompose into two or three crystalline forms with an accompanying change in chemical composition towards the niobium, zirconium, and titanium rich portion of the ternary system, with resulting poor anodic film properties. For these reasons, it is important that the alloy remain as a single, crystalline structure of constant composition, preferably a single solid solution, beta (body centered cubic) phase. Complete beta phase, solid solution only occurs at temperatures above about 1,000.degree.C. It is therefore necessary to homogenize this alloy above this temperature and cool it with sufficient rapidity to retain the high temperature, beta phase structure. Some compositions, notably the Nb-Zr rich regions, are highly unstable and rapidly form second phases on cooling, The addition of titanium not only lowers the phase transition temperature but extends the solid solubility range and readily promotes the retention of the high temperature beta phase during cooling.

It can be seen from the tables in the Examples below that most of the alloy compositions can produce 200-volt anodic films, useful as capacitor devices. preferred alloy compositions are characterized by low leakage and a low capacitance change (.DELTA.C) on thermal treatment. These alloys would be useful for elevated temperature use. They normally contain 20-40% Nb, 30-60% Zr, and 20-40% Ti.

Best results occur when the niobium and titanium are in substantially equal atomic concentrations (within 10 percent of the total).

From life test data of porous anode, solid electrolyte capacitors in Table 5, it is obvious that essentially all of the compositions and essentially the whole system (i.e., 20-80% Nb, 20-80% Zr, 20-80% Ti) are capable of being fabricated into useful capacitors. However, some compositions, namely those stated above, give low leakage and good stability and are to be preferred over the others.

FIG. 1 is a cross-section photomicrograph of niobium-zirconium-titanium alloy of the composition corresponding to data point 15 in Example 2, in sintered powder form. The magnification of the photo is 800 times. FIGS. 1A-1C show (at 800 times also) niobium-50 percent zirconium powder sintered at 1,300.degree., 1,400.degree. and 1,500.degree.C respectively.

FIGS. 2, 3, 4, 5 show sectioned capacitor anodes of atomic compositions and powder sizes as follows:

FIG. Composition (Nb-Zr-Ti) Starting Powder Size ______________________________________ 2 30 - 40 - 30 -325 mesh + 5 microns 3 30 - 40 - 30 -60 mesh + 325 mesh 4 20 - 60 - 20 -325 mesh + 5 microns 5 20 - 60 - 20 -60 mesh + 325 mesh ______________________________________

All magnifications are 100 times in FIGS. 2-5.

The niobium-zirconium-titanium alloys shown in FIGS. 1-5 have no second phase. A precipitated second phase is quite apparent in FIGS. 1A, 1B, 1C. In general, a deleterious second phase, if produced, will be visible at 100 times optical magnification and can be regarded as undesired if observed at as low as 100 times magnification.

The production of the requisite single phase crystalline structure can be accomplished (a) in-situ in the electrode (during capacitor production or prior to capacitor production) or (b) in the production of starting materials to be used later in electrode production. In either case what is involved is raising the alloy composition material to a high enough temperature as in sintering powder or annealing sheet to produce, as a homogenous solid solution, the single Beta phase high temperature modification of the alloy structure and to bring it to equilibrium, then cooling with sufficient rapidity to maintain the Beta structure essentially free of any second phase observable at 800 times magnification.

A typical cooling schedule for purposes of the invention would be from 1,300.degree. to 300.degree.C in 20 minutes and could be achieved by continuous movement of parts from a furnace to a cooling zone, as by a conveyor belt.

FIGS. 6-7 show two examples of capacitors utilizing the present invention. Each of the capacitors comprises an anode 10 with a dielectric comprising at least in part a dielectric oxide film on the surface of the anode material. The capacitors also comprise a cathode electrode 20. Leads 12, 22 are attached to anode 10 and cathode 22 respectively. In FIG. 6, the anode has the form of a sintered powder slug and the capacitor is of the electrolytic type with wet or solid electrolyte 21 impregnating the anode and extending to the cathode. FIG. 6 also shows conventional capacitor packaging elements as a seal 30 and plastic encapsulant 32. The anode 10 in the FIG. 7 capacitor is a rolled foil or a film sputtered from a sputtering target and includes a dielectric oxide film 11 on the anode surface.

The invention can also be utilized in non-polar capacitors, as well as the polar capacitors of FIGS. 6-7, and in other electrical, chemical and electrochemical devices requiring the significant characteristics of a tantalum substitute material as described herein.

FIG. 8 is a ternary diagram showing the various data points refereed to in the examples below. The percentages on the three sides of the diagram are atomic percentages.

The compositions indicated are based on proportions of starting materials used. At any given location in a powder or other product produced from the starting materials by melting together, the composition may vary due to the known metallurgical phenomenon of coring as the melt is cooled. Generally, the compositional differences found along the material on a microscopic scale are acceptable. However it is preferable to minimize compositional variation due to coring during ingot solidification and this can be accomplished by homogenization heat treatment applied either as an additional step immediately after melting and cooling the alloy and/or in the course of the fabrication steps (e.g., sintering powder or annealing foil), as noted above.

It is preferred and distinctly advantageous to select alloy compositions of substantially of atomic proportions:

niobium -- 20-40% zirconium -- 30-60% titanium -- 20-40%

with ) equal atomic proportions of niobium and titanium, particularly where (1) fine powders (6 microns Fisher Average Particle Diameter - F.A.P.D. or less) are produced, (2) where high oxygen impurity levels are in the material, (30 where hydrogen addition processing is used as an aid to pulverizing.

An alloy composition of essentially (within plus or minus 5 percent) the following atomic proportions:

50% zirconium

25% niobium

25% titanium

(data point 9 in FIG. 8) has been found to be distinctly advantageous for fine powder production and retain a single beta phase composition under processing conditions which produced multiphase crystal structure in powder of alloy composition corresponding to data points 15 and 16 and was fabricated into solid electrolytic capacitors along with (and under identical conditions with) powder of alloy compositions 6 and 16 and produced superior capacitors with respect thereto in terms of affording an economic tantalum capacitor substitute.

While any of several methods of alloy powder production are feasible (including co-reduction of their mixed halides) it is preferred and distinctly advantageous to obtain the three alloy components -- Nb, Zr, Ti -- in elemental form, melt them together and cool to form an alloy ingot, hydride the ingot to embrittle it, pulverize the hydrided ingot and dehydride the powder so formed by heating.

The dehydriding can be done on the powder, per se, or in conjunction with sintering the powder to an anode compact, the former being preferred.

Melting, hydriding, pulverizing and dehydriding tends to produce a highly agglomerated alloy powder, which is advantageous for compacting and tends to retain high capacitance consistent with low leakage produced by cleansing impurities in the various melting, heating and gas purging steps. The agglomerated particles contain individual fine powder particles in the range of 18 microns or less with the agglomerates per se being plus 325 mesh, minus 60 mesh.

The practice and relative value of the invention is illustrated by the following non-limiting examples (including some specimens outside the invention scope):

EXAMPLE 1

Several alloys of nominal compositions listed in Tables 1-3 below were made by arc melting 50-100 gram alloy buttons in argon atmosphere using a non-consumable tungsten electrode. The buttoms were formed from niobium chips and crystal bar zirconium and titanium with raw material specification as in Example 2 (below). Each alloy button was melted four times and then heated to about 80 percent of the melting point and held at that temperature for 2 hours. The buttons were then sliced with a diamond wheel saw. One slice of each button was chemically analyzed and the remainder were cold rolled from three-sixteenths inch to a thickness of about 0.015 inch. The rolled sheets were cut into panels, degreased, chemically polished, rinsed and annealed above 1,000.degree.C at about 80 percent of their respective melting points for 30 minutes in a vacuum of about 10.sup.-.sup.5 torr. The panels were anodized at 1 ma cm.sup.2 of surface in 0.01% H.sub.3 PO.sub.4 (at 25.degree.C in one test and at 92.degree.C in a second test) to 200 volts and held for 30 minutes. The anodized panels were rinsed and dried and electrically tested in a wet cell (1% H.sub.3 PO.sub.4 electrolyte) as formed and again after a later heat treatment of 300.degree.C for 30 minutes in air.

The results are shown in Table 1 for 25.degree.C anodization and in Table 2 for 92.degree.C anodization where DCL is D.C. leakage in microamperes per square centimeter of anode surface tested, DF is dissipation factor (percent), C is capacitance in microfarads (uf) or millimicrofarads (muf) per square centimeter of tested anode surface and .DELTA.C is capacitance change (percent) due to heat treatment.

(Cap. after - Cap. before/Cap. before .times. 100)

Microstructure of the heat treated buttons was examined. In general the structures were equiaxed with sub-grain boundaries and dislocation etch pits reminiscent of a strained body-centered-cubic phase quenched from a high temperature. There was no evidence of second phase transformation except in data points 19 and 21; these had fine precipitated structures within the equiaxed structure.

The data points in Tables 1-2 are grouped from lowest leakage to highest leakage. Table 3 shows a further series of experimental results indicating reproducibility of the Table 2 data.

TABLE I __________________________________________________________________________ DATA NOMINAL COMPOSITION Cap. POINT Nb-Zr-Ti (At. %) DCL DF muf/cm.sup.2 .DELTA. C% __________________________________________________________________________ 24 25-25-50 <.02 1.6 56 70 9 25-50-25 <.02 1.7 52 30 23 10-25-65 .02 1.8 53 36 7 20-40-40 .02 2.0 52 81 12 20-60-20 .02 2.1 59 9.3 11 30-60-10 .04 1.7 44 3.9 14 20-20-60 .03 2.0 55 88 15 40-20-40 .05 1.9 56 86 6 30-40-30 .05 1.7 55 74 2 33.3-33.3-33.3 .11 2.0 55 6.7 17 30-10-60 .08 4.5 61 86 25 45-10-45 .12 3.6 101 152 8 35-50-15 .30 1.4 35 1.2 18 60-10-30 .33 2.4 64 128 16 60-20-20 .53 3.1 59 78 10 10-60-30 .58 2.3 50 29 13 15-50-35 .63 2.6 48 61 1 75-15-10 1.6 7.6 -- -- 5 40-40-20 3.3 4.6 -- -- 3 60-30-10 4.5 6.8 -- -- 22 10-60-30 17 5.8 -- -- 20 10-40-50 46 7.2 -- -- Niobium .06 2.5 78 72 Tantalum <.02 2.1 68 3.0 __________________________________________________________________________

TABLE II __________________________________________________________________________ DATA NOMINAL COMPOSITION Cap. .DELTA. C POINT Nb-Zr-Ti (At. %) DCL DF muf/cm.sup.2 % __________________________________________________________________________ 9 25-50-25 .02 1.2 54 0.3 12 20-60-20 .02 1.5 54 39 13 15-50-35 .02 1.7 49 38 2 33-33-33 .02 1.5 41 74 6 30-40-30 .02 2.1 45 0.7 22 10-60-30 .02 2.2 54 88 10 10-60-30 .06 2.5 56 31 24 25-25-50 .14 2.2 44 30 5 40-40-20 .09 1.5 34 0.5 16 60-20-20 .04 1.6 48 94 15 40-20-40 .09 1.5 36 110 1 75-15-10 .06 3.4 56 105 25 45-10-45 .13 3.1 81 181 18 60-10-30 .33 1.7 56 172 17 30-10-60 2.1 2.2 47 91 11 30-60-10 7.4 4.4 -- -- 7 20-40-40 9.5 7.5 -- -- 3 60-30-10 22 5.0 -- -- 14 20-20-60 37 11.2 -- -- 23 10-25-65 High 5.7 -- -- 8 35-50-15 High 10 -- -- 20 10-40-50 High 13 -- -- Tantalum .02 1.6 54 2.2 Niobium High (will not anodize) -- 19 10-10-80 -- -- -- -- 21 10-80-10 -- -- -- -- __________________________________________________________________________

TABLE III __________________________________________________________________________ DATA NOMINAL COMPOSITION Cap. POINT Nb-Zr-Ti (At. %) DCL DF muf/cm.sup.2 .DELTA. C __________________________________________________________________________ 2-1 33.3-33.3-33.3 <.02 1.5 41 74 2-2 33.3-33.3-33.3 .02 1.6 46 96 2-3 33.3-33.3-33.3 .02 1.6 45 120 2-4 33.3-33.3-33.3 .03 1.8 48 44 2-5 33.3-33.3-33.3 .02 1.8 43 50 5-1 40-40-20 .09 1.5 34 0.5 5-2 40-40-20 .17 1.3 36 0.0 5-3 40-40-20 .04 1.4 36 1.0 6-1 30-40-30 .02 2.1 45 0.7 6-2 30-40-30 .02 2.1 50 29 6-3 30-40-30 .02 2.0 49 1.3 7-1 20-40-40 9.5 -- -- -- 7-2 20-40-40 12 -- -- -- 7-3 20-40-40 5.8 -- -- -- 9-1 25-50-25 <.02 1.2 54 0.3 9-2 25-50-25 .02 1.9 52 7.8 9-3 25-50-25 <.02 1.4 54 8.2 10-1 10-60-30 .06 2.5 56 31 10-2 10-60-30 .06 2.1 48 194 10-3 10-60-30 .02 2.0 54 152 22-1 10-60-30 .02 2.2 54 88 22-2 .03 2.3 50 90 12-1 20-60-20 <.02 1.5 54 39 12-2 20-60-20 <.02 1.5 56 40 12-3 20-60-20 .02 1.6 56 41 15-1 40-20-40 .09 1.5 36 110 15-2 40-20-40 .08 1.6 46 14 15-3 40-20-40 .04 1.6 46 44 16-1 60-20-20 .04 1.6 48 94 16-2 60-20-20 .02 1.8 42 107 16-3 60-20-20 .16 2.2 39 118 24-1 25-25-50 .14 2.2 44 30 24-2 25-25-50 .29 2.3 44 198 24-3 25-25-50 5.7 -- -- -- __________________________________________________________________________

EXAMPLE 2

Several alloy buttons were formed by melting together niobium, zirconium and titanium with the following purities:

Nb (machining chips) 99.94% Zr (crystal bar) 99.91% Ti (crystal bar) 99.94%

The nominal compositions of the alloys are given in Table 4. The buttons were hydrided, ground to powder and dehydrided. The powders in varying size cuts were sintered above 1,000.degree.C into porous compacts of approximately 1 gram each, cooled and anodized in 0.01% H.sub.3 PO.sub.4 electrolyte at 92.degree.C temperature with a current density of 63 ma per anode to a formation voltage for 2 hours. As a control, niobium anodes were similarly formed at 25.degree.C at current density of 50 ma per anode.

The anodized compacts were impregnated with manganese nitrate and pyrolyzed to form manganese dioxide electrolyte. The pyrolysis temperature was 275.degree.C for 8 minutes. Three impregnation dips followed by pyrolysis were used. Dip time was 3 minutes. After MnO.sub.2 impregnation, the anode (anodic oxide film) was reformed at 2 volts/minute in the 0.01% H.sub.3 PO.sub.4 at 92.degree.C to break-down. The reformation break-down is given in Table 4 as V.sub.R.

Similar capacitor samples were reformed to 35 volts. A counter-electrode was added and the devices were tested at 20 volts for capacitance and leakage. The results are given in Table 4.

TABLE 4 __________________________________________________________________________ NOMINAL ATOMIC % *POWDER L/C DATA COMPOSITION SIZE/SINTER V.sub.R CAP. .mu. amp. POINT (Nb-Zr-Ti) TEMP (.degree.C) (VOLTS) (.mu.fd.) .mu.fd. __________________________________________________________________________ 2-1 33-33-33 F/1100 40 23.3 .39 2-2 33-33-33 F/1200 40 16.5 .07 2-3 33-33-33 F/1300 40 12.4 .51 2-4 33-33-33 C/1300 40 5.3 .26 5 40-40-40 F/1300 65 11.7 .22 6-1 30-40-30 F/1100 47 12.0 .17 6-2 30-40-30 C/1200 47 5.4 2.3 6-3 30-40-30 C/1300 52 4.8 .35 6-4 30-40-30 C/1100 18.0 .08 9-1 25-50-25 F/1200 40 8.6 .14 9-2 25-50-25 F/1200 63 8.9 .06 9-3 25-50-25 C/1300 38 4.4 .11 9-4 25-50-25 C/1300 65 4.4 .05 12-1 20-60-20 F/1100 80 7.8 .06 12-2 20-60-20 C/1200 62 4.5 .13 12-3 20-60-20 C/1300 62 4.5 .13 13 15-50-25 F/1100 63 14.2 .09 15-1 40-20-40 F/1100 55 11.2 2.0 15-2 40-20-40 C/1200 58 9.9 1.6 15-3 40-20-40 C/1300 42 4.7 2.5 16-1 60-20-20 F/1300 56 15.4 16-2 60-20-20 C/1400 58 5.8 1.1 22 10-60-30 F/1100 65 13.1 .31 24 25-25-50 F/1200 67 11.5 .65 CONTROL 100-0-0 A/2050 85** 15.3 .10 __________________________________________________________________________ * F cut is -325 mesh +5 micron, C cut is -100 mesh + 325 mesh, A cut is -140 mesh +5 microns F.A.P.D. ** 25.degree.C Test for 100-0-0, 85.degree. for others

Solid capacitor samples of compositions noted in Table 5 and -325 mesh powder size were put through life tests of 200 to 1,000 hours and resultant life test data 15 charted in Table 5. In the life tests, leakage currents capacitance and dissipation factor are measured at 25.degree.C. Then the temperature is raised to 85.degree.C and the capacitors retested. Temperature is held at 85.degree.C for an extended period, the capacitors retested and then the temperature is dropped to 25.degree.C for retest. Initial (I) and final (F) values of these parameters are shown in Table 5. In some instances the "final" value at high temperature is a median or average value. ##SPC1##

EXAMPLE 3

Solid electrolytic capacitors were prepared as in Example 2 using the alloy compositions identified as 6, 12 and 16 in Table 5. Specific anode sintering and dielectric oxide formation conditions are given in Table 6A.

These capacitors were tested for electrical properties at room temperatures and then life tested at 85.degree.C under 20 volts bias for 1,000 hours. The results are given in Table 6B.

The results were in agreement with those tabulated in Table 5 except that composition 16 had improved leakage behavior. However, the alloy also exhibited capacity instability.

The life testing acted as a "burn-in" for capacitors, tending to stabilize them, for better performance in future life testing and would be a desirable step for capacitor production.

TABLE 6 A ______________________________________ ANODE PROCESSING ______________________________________ Green Sinter No. Weight (gm) Density (g/cc) Temp.degree.C Time ______________________________________ 6 .734 3.90 1100 1/2 hr. 12 .722 3.80 1100 1/2 hr. 16 .823 4.34 1300 1/2 hr. ______________________________________

FORMATION

Electrolyte: 0.01% H.sub.3 PO.sub.4 at 92.degree.C.

Current Density: 63 ma/anode.

Formation Voltage: 200 Volts.

Hold Time at Voltage: 2 Hours.

TABLE 6B __________________________________________________________________________ COMPOSITION LIFE TEST TEMP L C L/C DF NUMBER HOURS (.degree.C) (.mu.a/anode, med.) (.mu.f/anode, avg.) (.mu.a/.mu.f) (%) __________________________________________________________________________ 6 0 25 .50 7.6 .066 6.1 0 85 2.4 9.2 .26 8.2 100 85 .94 8.4 .11 7.8 250 85 .92 8.2 .11 6.9 500 85 .96 7.8 .12 6.0 750 85 1.20 7.8 .15 6.2 1,000 85 1.30 7.7 .17 6.1 25 .37 7.1 .052 4.2 12 0 25 1.1 10.1 .11 8.8 0 85 9.2 12.2 .75 8.6 100 85 4.3 10.9 .39 9.9 250 85 4.7 10.9 .43 10.2 500 85 6.1 10.3 .59 9.1 750 85 6.4 10.3 .62 8.7 1,000 85 6.6 10.3 .64 8.4 25 .92 9.0 .10 6.5 16 0 25 .63 17.4 .036 11.8 0 85 3.6 23.1 .16 10.5 100 85 3.2 20.6 .16 11.9 250 85 3.0 19.5 .15 9.3 500 85 3.9 18.1 .22 7.3 750 85 4.0 17.6 .23 7.3 1,000 85 4.3 17.4 .25 6.8 25 .66 15.5 .043 8.8 __________________________________________________________________________

EXAMPLE 4

Solid electrolytic capacitors were prepared using the alloy compositions identified as data points 6, 9, 15, 16 and elemental tantalum and elemental niobium as controls in the conditions given in Tables 7 and 8 (including Tables 7-1 (a), (b), (c) and Table 8-1) for powder preparation and capacitor formation, respectively. The capacitors were life tested under the conditions shown in Table 9A and with the results shown in Tables 9B-9D.

Table 7 -- Powder Preparation

1. Melt together, to form a solid alloy ingot, Nb, Zr, Ti with commercial purities shown in Table 7-1(a) to produce an ingot with purities shown in Table 7-1(b).

2. Hydride ingot by heating in a hydrogen atmosphere at 700.degree.C for about an hour or until hydrogen pick-up becomes very slow and cool to 600.degree.C and repeat again at 500, 400, 300 and then cool to ambient.

3. Process ingot to powder by crushing to -10 mesh, then ball milling to produce powder.

4. Screen powder to final desired size distribution by mechanical or fluid classification technique.

5. Dehydride by heating powder to 700.degree.C in a chamber under subatmospheric pressure, coming up to temperature slowly to avoid an excessive rate of hydrogen evolution which would create an explosion hazard, holding at 700.degree.C until hydrogen content is reduced to less than 500 parts per million.

6. Cool and then passivate by slowly admitting air to the powder before opening up the chamber.

7. The lightly sintered cake resulting from (4)-(6) is lightly crushed to pass a 60 mesh/inch screen (A.S.T.M.). The resultant product is an agglomerated powder with individual particles in the range of 4-18 microns. Some 75 percent of the material are in agglomerate form with a size of between plus 325 mesh and minus 60 mesh. The powdered material as a whole has a Fisher Average Particle Diameter of 5-6 microns. The powders have the purity shown in Table 7-1(c). ##SPC2##

Table 8 -- Capacitor Preparation

1. Press powder using 8 percent by weight camphor addition to achieve a green compact of green density (Dg) as shown in Table 8-1.

2. Sinter for 30 minutes to temperatures shown in Table 8-1 to densities shown in Table 8-1 producing anodes having the sintered density (Ds) and carbon impurity content shown in Table 8-1.

3. Anodize (form) alloy in 0.1% H.sub.3 PO.sub.4 at 92.degree.C to 80 volts at 120 milliamperes per anode. Similarly anodize niobium control at 25.degree.C.

3a. Perform wet cell testing as shown in Table 8-1, measuring or calculating from measurements leakage, (in microamperes), leakage to capacitance ratio (microamperes per microfarad), capacitance, specific capacitance by weight and volume (microfarad-volts per gram and cubic centimeter, respectively), dissipation factor and equivalent series resistance. Alloy powder of data point 15 had very high leakage (in excess of 1 milliamp).

4. Impregnate and pyrolyze by (a) dipping for 3 minutes in 12% concentration of manganese nitrate solution, pyrolyzing for 8 minutes at 275.degree.C, and repeat this process two more times, then reforming to 35 volts; (b) dipping again in 25 percent manganese nitrate for 3 minutes, pyrolyzing for 8 minutes at 275.degree.C, and repeat this process two more times, then reforming to 35 volts; (c) dipping again in 50 percent manganese nitrate for 3 minutes, pyrolyzing for 8 minutes at 275.degree.C, and repeat this process two more times, then reforming to 30 volts. Alloy powder of data point 15 could not be reformed because of high leakage and therefore could not be processed to solid capacitor form, except with great difficulty.

5. Apply counterelectrode (cathode) by coating with colloidal graphite, overcoating with silver paint, attaching a conductor with conductive epoxy, and encapsulating in non-conducting epoxy resin.

Table 8-1 __________________________________________________________________________ Porous Anode and Wet Cell Test Data __________________________________________________________________________ Alloy (Element) 30 min. Identi- Anode Dp Ts Ds Carbon L L/C CV/g CV/cc DF ESR fication Wt(gm) (g/cc) (.degree.C) (g/cc) (ppm) (.mu.a) (.mu.a/.mu.F) C (.mu.f-V/g) (.mu.f-V/cc) (%) (Ohms) __________________________________________________________________________ NZT 6* (1) .710 3.2 1100 4.4 480 39.0 0.81 48.4 5,440 1,236 55 15.07 (2) .710 3.2 1050 4.4 610 78.9 1.1 71.7 8,080 1,836 53 9.81 NZT 9* (1) .691 3.0 1000 3.8 164 104.0 1.7 62.1 7,190 1,892 52 11.11 (2) .691 3.0 1000 3.7 2200 144.4 2.2 65.6 7,600 2,054 48 9.71 (3) .691 3.0 1000 3.7 500 129.2 1.7 76.0 8,800 2,378 53 21.60 NZT 16* (1) .776 3.5 1200 3.8 270 92.0 0.81 114.0 11,750 3,092 71.5 8.32 (2) .776 3.5 1300 -- 655 164.0 1.9 85.8 8,840 -- 58.5 -- (Niobium) .920 3.9 1600 5.6 152 9.9 .12 86.0 7,480 1,336 73.5 11.34 (Tantalum) 1.84 8.0 1850 9.1 -- 1.1 0.013 82.5 3,590 395 55 8.85 __________________________________________________________________________ * See Table 7-1

Table 9A - Life Test Procedures

Test for leakage, capacitance and dissipation factor at 25.degree.C initially. Heat to 85.degree.C and retest. Hold at temperature for about 600 hours and then retest. Cool to 25.degree.C and retest. Detailed test procedures were in accordance with Example 3 procedures and MILSPEC MIL-C-39003.

The samples were pre-aged at 15 volts bias at 85.degree.C through a prior test for 410 hours prior to the tests reported here. The samples were biased to 20 volts for these tests using a bias circuit with a series resistance of 10-15 ohms.

Table 9B ______________________________________ Leakage (.mu.a) ______________________________________ Test Hours 0 hr. 590 hrs. Temp. 25.degree.C 85.degree.C 85.degree.C 25.degree.C ______________________________________ NZT-6 (1) 5.6 12.4 40.1 7.0 (2) 17.0 47.6 94.6 19.5 NZT-9 (1) 2.5 14.0 10.5 2.5 (2) 1.9 14.8 18.4 2.5 (3) 1.5 7.8 10.0 2.0 NZT-16 (1) 32.5 200 432 65.6 (2) 47.0 167 343 58.7 Niobium 13.1 76.9 186 144 Tantalum 0.88 3.7 3.3 0.79 ______________________________________

Table 9C ______________________________________ Capacitance (.mu.f) ______________________________________ Test Hours 0 hr. 590 hrs. Temp. 25.degree.C 85.degree.C 85.degree.C 25.degree.C ______________________________________ NZT-6 (1) 34.0 44.0 39.0 31.6 (2) 50.0 62.0 56.0 47.0 NZT-9 (1) 36.1 46.0 42.0 35.0 (2) 32.1 46.0 41.0 31.0 (3) 38.1 61.1 45.6 36.0 NZT-16 (1) 83.0 102.0 94.0 80.0 (2) 56.0 53.0 58.0 52.0 Niobium 83.4 89.0 86.1 72.0 Tantalum 88.0 94.0 94.0 87.0 ______________________________________

Table 9D ______________________________________ Dissipation Factor - DF(%) ______________________________________ Test hours 0 hr. 590 hrs. Temp. 25.degree.C 85.degree.C 85.degree.C 25.degree.C ______________________________________ NZT-6 (1) 14.5 26.5 17.0 15.0 (2) 13.0 22.5 14.0 16.5 NZT-9 (1) 13.0 23.0 16.0 18.0 (2) 12.0 25.0 19.0 14.0 (3) 9.0 20.5 12.0 5.0 NZT-16 (1) 19.0 31.5 18.5 12.5 (2) 12.0 17.5 8.5 6.0 Niobium 65.0 51.0 64.0 65.0 Tantalum 12.0 12.0 12.5 11.5 ______________________________________

It is seen from Table 9B that niobium and alloy composition NZT-16 increased substantially in leakage under test conditions, while NZT-6 and NZT-9 did not increase. NZT-9 had the lowest leakage approaching tantalum. Table 9C shows fairly stable capacitance and DF for all samples tested. Data external to this table (e.g., leakage data) makes clear that the niobium capacitance and DF would run away upon further life testing. There is some likelihood of runaway for NZT-16 as well, but the likelihood is less certain.

EXAMPLE 5

Niobium, zirconium, titanium alloy powders of compositions corresponding to data points 2, 5, 6, 9, 12, 13, 15, 16, 22, 24 of FIG. 8 were prepared and sintered as anodes. These powders were in a size range of -325 mesh, plus 5 microns (nominally 10 microns Fisher Average Particle Diameter). X-ray diffraction analysis was made of the powders per se, and also of samples made by crushing the anodes, to detect crystal structure (including lattice parameter) of the phases present in the material. The results are shown in Table 10A. The procedure was repeated for fine powders (nominally 4 microns F.A.P.D.) with varying levels of oxygen contamination and the resultant data are shown in Table 10B. Table 10B also shows sintering temperature.

Table 10A __________________________________________________________________________ X-Ray Analysis of Nominal 10 Micron F.A.P.D. Nb-Zr-Ti Alloys: Crystalline Structure (and Lattice Parameter) __________________________________________________________________________ Annealed Powder Sintered Anode Data Composition Major Second Major Second Point At.% Nb-Zr-Ti Phase Phase Phase Phase __________________________________________________________________________ 2 33.3-33.3-33.3 BCC(3.39) none BCC(3.40) none 5 40 -40 -20 BCC(3.49) BCC(3.32) BCC(3.42) none 6 30 -40 -30 BCC(3.42) none BCC(3.48) none 9 25 -50 -25 -- -- BCC(3.45) none 12 20 -60 -20 BCC(3.47) none BCC(3.42) none 13 15 -50 -35 -- -- BCC(3.46) none 15 40 -20 -40 BCC(3.35) none BCC(3.36) none 16 60 -20 -20 BCC(3.35) none -- -- 22 10 -60 -30 BCC(3.48) HCP BCC(3.48) none 24 25 -25 -50 BCC(3.37) none BCC(3.38) none __________________________________________________________________________

Table 10B __________________________________________________________________________ X-Ray Analysis of Nominal 4 Micron F.A.P.D. Nb-Zr-Ti Alloys: Crystalline Structure (and Lattice Parameter(s)) __________________________________________________________________________ Data Oxygen- Sinter Sintered Anode Point (Sample) ppm Temp.-.degree.C Major Phase Second Phase __________________________________________________________________________ 6 8800 1100 BCC(Ao:3.41) none* 9 (1) ca. 6000 1000 BCC(3.45) none (2) 9400 1100 BCC(3.45) none* 15 14,400 1100 BCC(3.34) BCC(3.45)* 16 (1) ca 6000 1000 BCC(3.34) HCP (Ao: 3.23, Co: 5. 10 (2) 6300 1100 BCC(3.33) HCP (Ao: 3.21, Co: 5. 12) __________________________________________________________________________ *excepting contaminants, such as zirconium oxynitride and oxycarbide.

Bcc: body centered cubic (Ao reported)

Hcp: hexagonal close packed (Ao, Co reported)

Although the 10 micron size powders of alloys (data points) 5 and 22 in Table 10A had second phases in powder form, they became single phase materials in the course of sintering (which was carried out at or above 1,000.degree.C for 30 minutes). But the fine powder alloy (data point) 15 with high oxygen content and slow cooling in Table 10B had a double beta (body-centered-cubic) phase in anode form. Data point 16 also had a second phase -- an HCP structure.

EXAMPLE 6

An extensive number of Nb-Zr-Ti alloy compositions were prepared by sputtering thin films onto glass slides, anodizing the metal film to 200 volts, testing heating to 300.degree.C for one-half hour and testing again in the same manner as Example 1. The results are shown in FIG. 9 which represents the DC leakage and FIG. 10 which shows the thermal stability of capacitance over the entire ternary phase diagram.

The results show that good anodic film properties (<10 .mu.a/cm.sup.2 at 140 V DC) extend over a wide range of compositions beginning in the central region of the Nb-Zr binary, through the central region of the Nb-Zr-Ti ternary and extending toward the central region of the Zr-Ti binary. The region of high thermal stability of capacitance is more localized near the center of the Nb-Zr-Ti ternary with a narrow path extending toward the central portion of the Zr-Ti binary and another narrow extension toward the Zr corner of the ternary, representing equal portions of niobium and titanium with increasing zirconium concentration.

While the results of sputtered film do not represent the cyrstalline structures found in the bulk alloy such as annealed sheet or powder or sintered porous anodes, they do show the effects of chemical composition on the electrical properties of alloy anodic films. For example, while the central portions of the ternary and the Nb-Zr and Nb-Ti binaries displayed good DC leakage, the same composition cannot be retained in bulk form because of phase separation accompanied by compositional changes, unless an extremely rapid quench of the high temperature modification of the single BCC structure is used. Retention of the single BCC structure is more easily accomplished in the bulk when the alloy composition is removed from the Nb-Zr and Zr-Ti binary lines by about 10 atomic percent. The trends shown by data afford approximations of what will occur in bulk form.

While the present invention has been described with reference to the particular embodiments thereof, it will be understood from the above disclosure by those skilled in the art that numerous modifications may be made without actually departing from the scope of the invention. For instance, the invention contemplates the ternary Nb-Zr-Ti alloy present alone or as part of a larger alloy (or mixture) system of film-forming metals. Where the niobium is substituted in any amount by tantalum and/or zirconium and/or titanium is substituted in any amount by hafnium where these substitutions are chemically similar to the components of the Nb-Zr-Ti ternary. Less similar elements, such as vanadium, molybdenum and tungsten, may be added up to 20 atomic percent of the total system without severe degradation. Dissimilar elements, such as iron, chromium and nickel, may be tolerated by as much as 1-5 percent. More dissimilar elements, such as noble metals, alkali, and alkaline metals, etc., may be considered as true impurities and should be less than 1 percent.

In its broadest aspects, the invention comprises a niobium-zirconium-titanium ternary alloy of powder form suitable for making capacitor anode or the like with the characteristic that it is capable of forming or retaining an essentially homogeneous crystal structure of a single phase although it is preferred and distinctly advantageous that:

a. the single phase should be a beta (body-centered-cubic) phase, which is more practicably attainable consistent with capacitor powder preparation conditions, and/or

b. that the powder, per se, shall have single phase structure.

The powder is suitable for manufacturing into capacitor anodes whether in single powder form, agglomerated form, pre-pressed or pre-sintered preliminary compact form, or in the form of final sintered anode compacts which can be re-crushed and remanufactured. Therefore the appended claims are intended to cover all such equivalents or variations as come within the true spirit of the invention.

Claims

I claim:

1. A powdered material of niobium-zirconium-titanium ternary alloy, in form suitable for manufacture into an electrolytic capacitor anode, the alloy having an essentially homogeneous crystal structure of a single body centered cubic alloy phase wherein the elements of the alloy are present in atomic percentages of from:

20 to 40% for niobium

30 to 60% for zirconium

20 to 40% for titanium.

2. The material of claim 1 wherein the niobium and titanium are present in the alloy in substantially equal atomic amounts.

3. The material of claim 2 having a Fisher Average Particle Diameter of less than 6 microns.

4. The material of claim 3 wherein the elements of the alloy are found throughout the material in essentially the atomic proportions of:

50% zirconium

25% niobium

25% titanium.

5. The material of claim 2 wherein the elements of the alloy are found throughout the material in essentially the atomic proportions of:

50% zirconium

25% niobium

25% titanium.

6. A capacitor grade powder for use in manufacturing solid electrolytic capacitor sintered, porous anodes anodizable to 200 volts and affording high thermal stability and low leakage and dissipation factor, consistent with high capacitance, comprising

single phase body-centered-cubic crystal structure niobium-zirconium-titanium ternary alloy throughout the powder mass, the elements of the alloy being present in essentially homogeneous atomic percentages and within the respective ranges of:

20-80% for niobium,

20-80% for zirconium,

20-80% for titanium,

throughout the powder mass.