A. Field of the Invention

The present invention relates to novel ductile ceramic composites which have superior mechanical properties over the prior art. The composites of this invention exhibit ductility, flexibility and enhanced energy absorption without failure (toughness) . Depending upon the selection of the ceramic composition, the ordered ductile array and the nature of ordering within the ductile array, composites of this invention having a wide range of uses are provided.

B. Prior Art

Ceramics generally possess a number of desirable properties, including markedly high resistance to abrasion, heat and corrosion compared to metallic materials. Certain ceramics, such as stabilized bismuth solid oxides, stabilized ceria solid oxides and stabilized zirconia solid oxides are ionically conductive materials suitable for use as solid electrolytes. Other ceramics are known to possess superconducting electrical properties. Still others have been employed in structural applications. However, due to extreme brittleness, their application has been limited despite their other excellent properties.

A number of attempts have been made to increase toughness of ceramic materials by compounding them with another material including ceramic or metal whiskers such as silicon carbide whiskers. Composites with ceramic matrices and ductile metal inclusions such as those produced by Lanxide Corporation show high fracture toughness when compared to ordinary ceramic materials. See for example United States Patent Nos. 4,824,622; 4,847,220; 4,822,759; 4,820,461; and related 4,871,008. These composites are a chaotic, generally discontinuous, random metal dispersion in a ceramic

composite body. They are prepared by a slow controlled oxidation of molten aluminum to alumina oxide, leaving behind approximately 5% of the parent metal. See also C.A. Anderson et al., Cera . Enα. Sci. Proc.. 9 [7-8] pp. 621-626 (1988); and M.S. Newkirk et al., Ceram. Enα. Sci. Proc. 8 [7-8] pp 879-885 (1987) .

P. Ducheyne et al. , J. Materials Science 17(1982) 595-606 discloses a bioglass composite produced by immersing premade porous fiber skeletons into molten bioglass to prepare metal fiber reinforced bioglass. These porous fiber skeletons produce random, chaotic, disordered support matrices and the process is applicable only to bioglasses.

United States Patent No. 4,764,488 discloses a high toughness ceramic composite of the fiber-reinforced type wherein metal fibers having the shape of triangular waves forming bent portions alternating on the opposite sides with an angle θ of the bent portions in a range between 60° and 165" and a d/H ration of between 0.025 and 0.6. While the discrete, discontinous fibers, unidirectionally anchored fiber reinforcement employed in the -488 patent improve the toughness of the ceramic, this technique does not solve the problem of crack propagation and ultimate failure. United States Patent No. 4,776,886 discloses a whisker-reinforced ceramic matrix composite comprising a principal crystal phase selected from the group consisting of anorthite, barium-stuffed cordierite and mixed cordierite/anthorite prepared by extrusion of ceramic batches comprising an extrusion vehicle and a solid component essentially composed of inorganic whiskers and powdered glass.

See also United States Patent Nos. 4,376,374; 4,572,754; 4,334,380; 4,324,843; 3,768,760; 4,524,622; 4,522,759; 4,820,461; 4,847,220; 4,571,008; 4,764,488;

4,070,529; 4,552,766; 4,052,532; 4,396,450; 4,812,329; 4,462,891; 5,006,494.

One application of the ductile ceramic composite of the present invention is as the solid electrolyte in a novel gas separation system which effectively separates oxygen from the air. The tough, ductile solid electrolyte composite of this invention permits the construction of an oxygen generator device in which the only moving part is an air intake fan, and which does not require consumables such as hydrogen peroxide or catalysts requiring constant replenishment.

In recent years, there have been attempts to provide compact and lightweight oxygen generating systems that can supply oxygen gas for extended periods. Japanese Utility Model Publication No. 26445/1980 discloses an oxygen gas generating system adapted to catalytically decompose aqueous hydrogen peroxide using a manganese compound as the catalyst. This system has several drawbacks. The decomposition reaction of aqueous hydrogen peroxide and manganese dioxide proceeds at an explosively high rate if the volume of hydrogen periodixe is not carefully controlled. If the volume and rate of the hydrogen peroxide reservoir is decreased to make the unit portable, the hydrogen peroxide is rapidly consumed and the reservoir must be frequently replaced. For both reasons, this is not a practical approach.

Japanese Patent Publication No. 42115/1977 employs a platinum catalyst capably of decomposing aqueous hydrogen peroxide at a high concentration. This system is also unsatisfactory, both because it requires a reservoir of hydrogen peroxide which must be periodically replaced, and because of the expense and nature of the precious metal catalyst. One problem with this approach is that the usual pore size of the alumina or silica gel catalyst support is too small to permit

penetration of the hydrogen peroxide. A major drawback is that the expensive catalyst has a limited life. A further drawback is the precise temperature control required. One attempt to address the problems with hydrogen peroxide based oxygen generating systems is disclosed in Japanese Patent Publication No. 49843/1981 in which a system is provided for controlling the flow rate of hydrogen peroxide by valve adjustment using a link mechanism to control the supply of aqueous hydrogen peroxide depending upon the pressure of the generated oxygen gas. However, the proposed system for converting the gas pressure into mechanical displacement and transmitting the displacement by means of the link has the drawback of being unable to rapidly respond to the change in the reaction rate with resulting failures due to corrosion and abrasion in the actuating system.

United States Patent No. 4,792,435 discloses a system for producing oxygen by catalytic decomposition of aqueous hydrogen peroxide employing a platinum group catalyst carried on a highly porous sintered ceramic support of large pore size. This system again suffers from the drawback of requiring a hydrogen peroxide reservoir which must be periodically recharged or replaced.

United States Patent No. 4,784,765 provides an aquarium oxygen generator comprising a container inverted into the apex of a ceramic cone-shaped ceramic structure resting on the floor of the aquarium. Hydrogen peroxide solution (15%) in the container is decomposed to form oxygen and water in the presence of a catalyst pellet of finely divided silver admixed with clay. Hydrogen peroxide seeps into the cone, and in the absence of the catalyst, reacts with organic material in the water to produce oxygen which bubbles through an aperture in the side of the cone-shaped structure into

the main body of water in the aquarium. While this system may be satisfactory for a small scale aquarium, it suffers from the drawback of requiring a hydrogen peroxide reservoir and is not suitable for medical, industrial and experimental (laboratory) uses.

See also United States Patent Nos. 4,879,016 and 4,636,291. U.S. 4,879,816 discloses an electrolyte assembly for oxygen generating devices which employs a brittle, conventional solid oxide electrolyte. U.S. 4,636,291 discloses a ceramic diaphragm for alkaline electrolysis.

The present invention solves the problems of the prior art and provides a system which generates oxygen from air, can be scaled up or down in size depending upon use, does not require consumables such as hydrogen peroxide or catalysts which must be replaced, and which is efficient and cost effective. The system of the present invention is made possible by the use of the novel, ductile, flexible and mechanically rugged, thin, solid state electrolyte ceramic composite of this invention.

Another application of the ductile ceramic composites of this invention is as the solid electrolyte in a solid oxide fuel cell. Various types of fuel cells have been investigated and differ largely in the type of electrolyte employed. See A. Fickett, in Handbook of Batteries and Fuel Cells, by D. Linden, McGraw-Hill Book Co., New York, NY (1984) p. 41-10. Five major classes of fuel cells have emerged. These include the polymer electrolyte fuel cell (PEFC) , the alkaline fuel cell (AFC) , phosphoric acid fuel cells (PAFC) , molten carbonate fuel cells (MCFC) and solid oxide fuel cells (SOFC) . The operating temperatures for these fuel cells are approximately 80'C, 100"C, 200 ^β C, 650"C and 1000"C respectively.

In 1988, the United States Department of Energy identified three basic fuel cell-types of current importance for near-term commercialization: the phosphoric acid fuel cells, molten carbonate fuel cells and solid oxide fuel cells. See Fuel Cells A Handbook May 1988, U.S. Department of Energy, Office of Fossil Energy, Morgantown Energy Technology Center, Morgantown, West Virginia, DOE/METC-88/6096.

While such cells provide an exciting option for the efficient conversion of fossil fuels to electricity, and could result in automobiles with mileages of 160 mpg or higher, as well as for stationary power plants and portable power plant applications, prior art technology has a number of problems. Low temperature PEFC cells employ an electrolyte which is an ion-exchange membrane (fluorinated sulfonic acid polymer) that is an excellent proton conductor. The only liquid in this fuel cell is water, thus corrosion problems are minimal. Water management in the membrane is critical for efficient performance. Thus, the fuel cell must operate under conditions where the by-product water does not evaporate faster than it is produced because the membrane must be hydrated. Because of limitations on the operating temperature, usually less than 120"C, H ₂ -rich gas with little or no CO is used, and higher Pt loadings than those used for the PAFCs are required in both the anode and cathode. It remains to be seen whether this type of fuel cell can be successfully commercialized. Phosphoric acid fuel cells (PAFC) were originally developed for operation on hydrocarbon fuels. Because the fuel cell can only utilize hydrogen derived from primary hydrocarbon fuels, it is necessary for the system to incorporate a fuel processing unit which adds steam to convert the primary hydrocarbon fuel to hydrogen and carbon monoxide. Adding this unit reduces

overall electrical/fuel conversion to around 40%. The PAFC fuel cells employ porous electrodes which are fabricated from a specially developed, finely divided carbon black onto which crystallites (2-5 n ) of an electrocatalyst such as platinum are deposited. This component is supported on a porous carbon-paper substrate to form a composite structure which forms a stable three-phase interface in the fuel cell. The electrolyte is phosphoric acid which is a hazardous material.

The molten carbonate fuel cell (MCFC) is one of the two high-temperature systems regarded as "second generation" fuel cells. It has been in development for about 10 years, particularly in the United States and Japan, and units in the range of 1-20 kw are currently being evaluated. However, these systems suffer from severe corrosion of many of the components by the molten salt electrolyte.

Special features of the AFC, such as readily available fuel and oxidant already present for rocket motors, made this system an attractive power unit for the American space program. For example, 1.5 kw units were incorporated on the Apollo space craft used to take astronauts to the moon, and three 12 kw units provide the electrical supply power for the space shuttle. This type of fuel cell is widely used for aerospace and defense applications and is available as a reliable, well-proven system. However, it is very expensive to operate, since it requires pure hydrogen and oxygen and relatively expensive electrode materials such as finely divided platinum or gold/platinum on carbon porous electrodes. Thus a need remains for an efficient, cost effective fuel cell which can be employed in more mundane, but equally important applications such as automobiles, electric power plants and other industrial applications.

The last major class of fuel cell is the solid oxide fuel cell (SOFC) . This system incorporates a ceramic solid oxide, typically with an approximate composition of Z ^* r ₀ . ₈₂ ^γ o. _i8 °ι. ₉₅ - The anion vacancies created by introducing aliovalent Y ³⁺ ions on the Zr ⁴⁺ sites produce an oxygen ion conductor with a conductivity at 1000"C comparable to liquid electrolytes in the range of room temperature to 2,000 ^β C. Because the electrolyte is solid, electrolyte management problems associated with other fuel cell types are eliminated. Further, the high-temperature operation allows in situ reforming of hydrocarbon fuels and electricity/fuel conversion efficiencies are expected to be higher than with other systems. Attempts to construct SOFC systems during the period 1970-1975, while technically successful and resulting in small (10W) units which operated for more then 30,000 hours, the cost-efficient assembly of strong, tough ceramic components was difficult and most projects were soon abandoned. The 1980's saw a resurgence of interest in the SOFC systems because developments in ceramic technologies allowed fabricating monolithic ceramic electrolyte configurations incorporating thin components with high area/volume ratios.

A tubular configuration was used in various projects during the 1970's and this design has been optimized by Westinghouse and scaled up to produce 5 kw. See B.C.H. Steele, MRS Bulletin/June 1989, p. 23. The Westinghouse unit comprises a bundle of tubular solid oxide fuel cells in cooperation with an array of multiple air injector tubes (one for each fuel cell unit) with spacers to define an air plenum chamber, a combustion chamber, and a fuel plenum chamber. More recently, increased attention has been devoted to planar configurations of the type depicted in

FIG. 14. In the prior art planar arrangements, ceramic electrolyte plates approximately 100-200 μ thick having fracture stress around 800 MPa and fracture toughness

(K _IC ) i ■n the 5-10 MPa m1/2 range have been reported. Because ceramics are inherently brittle, there is a strong dependence of mechanical failure characteristics on microstructural features, and better ceramic materials are needed to improve the reliability, performance and operating life of solid oxide fuel cells. The present invention provides such an electrolyte.

The other principal structural component of planar solid oxide fuel cell components is a bipolar plate incorporating gas channels. Prior art bipolar plates are usually fabricated from the well proven electronically conductive material LaCr _0>9 Mg ₀ ,0 ₃ because of its compatibility with other components, which has been demonstrated in the tubular configurations. The porous cathode of the prior art planar solid oxide fuel cell design usually consists of La ₀₈₅ Sr ₀₁₅ Mn0 ₃ and the porous anode, a thin layer of NiZr(Y)0 ₂ . _χ cermet. If porosity and pore size distribution of the electrodes are carefully controlled during the processing stage, these components provide stable three-phase electrode- solid electrolyte-gas interfacial regions.

For a detailed review of the prior art, see A.J. Appleby and F.R. Foulkes, Fuel Cell Handbook (Van Nostrand, 1989) ; "Assessment of Research Needs for Advanced Fuel Cells" in Energy, Vol. 11, edited by S.S. Penner (1986) p. 1-229; S. Srinivasan, "Fuel Cells for Extraterrestrial and Terrestrial Applications," J. Electroche . Soc. 136, 41C, 1989; B.C.H. Steele et al., Direct Electrochemical Oxidation of Methane in Ceramic Electrochemical Reactors. Abstract of 1988 Fuel Cell Seminar (Oct. 23-26, 1988, Long Beach, CA) ; B.C.H. Steele, "Materials Engineering and Fuel Cell

Development", MRS Bulletin/June 1989. pp. 19-23, Fuel Cells A Handbook. United States Department of Energy, Office of Fossil Energy, Morgantown Energy Technology Center, Morgantown, West Virginia (May 1988) . A principal limitation of current attempts to develop practical SOFC units is the brittle failure of the multi-layer all ceramic stack, in particular during thermal cycling. The present invention provides a superior solid electrolyte and an improved, simplified planar design for solid oxide fuel cells.

Another aspect of the present invention is in the field of superconductors, and more particularly to high temperature ceramic compositions which may be employed to prepare cables and other devices requiring ductility, flexibility and toughness.

The discovery of high-T _c oxide superconductors, especially of the Y-Ba-Cu-0, Bi-Sr-Ca-Cu-0 (including Bi-Pb-Sr-Ca-Cu-O) , and Tl-Ba-Ca-Cu-0 compositions which show superconductivity above liquid nitrogen temperature, has had a great impact both in the fields of physics and superconductors. See for example, M. K. Wu et al., Phvs. Rev. Lett. 58 (1987) 908; H. Maeda et al., Jpn. J. APPI. Phvs. 27 (1988) L 209 and Z.Z. Sheng et al.. Nature 332 (1988). The application of a superconducting material to a superconducting magnet requires the winding of tape or wire conductors with sufficient flexibility onto a solenoid coil. Among the various superconducting compounds having inherent brittleness, Nb ₃ Sn and V ₃ Ga compounds with A15 structure have been successfully fabricated into a tape or wire form by the "bronze process" which utilizes the solid state reaction. The oxide high-T _c superconductors are also intrinsically brittle and investigators are seeking ways to fabricate these materials so that they can withstand the various

stresses caused by coil winding, Lorentz force, and the like.

Standard ceramic processes of mixing powders and organic formulations are applicable to the preparation of high-T _c oxide superconductor tape or wire. However, these tapes and wires become brittle after the final sintering process. See, for example, D.W. Johnson et al., High Temperature Superconductors (Proc. of 1987 Spring Meeting of MRS), ed D.U. Gubser et al. (The Materials Research Society, Pittsburgh, 1987) p. 193; T. Goto et al., Jpn. J. Appl. Phys.. 26 (1987) 1527; K. Togano et al., Jpn. J. Appl. Phys.. 27 (1988) 45; and M. Ishii et al., Jpn. J. Appl. Phvs.. 27 (1988) 1652.

Attempts to fabricate flexible ribbons of high temperature superconductors have been somewhat successful using the technique reported by K. Togano et al., Jpn. J. Appl. Phvs.. 28 (1989) 95-96. Thin ribbons of high temperature Bi-Sr-Ca-Cu-0 superconductor ceramic have been prepared as thin ribbons supported on a polyethylene terephthalate carrier sheet. The original ribbon, approximately 30 μm thick, was reported to be initially flexible and capable of being bent to a small diameter of about 30 mm (bending strain = 0.1%). While promising, practical applications of this material would be limited to fabrication of tape conductors for relatively small (under 40 mm diameter) superconducting coils. Further, upon repeated cycling, both the ceramic and the plastic backing would become brittle and fail. Even if these problems could be overcome, this material does not possess sufficient ductility for use in winding coils. See K. Togano et al., Jpn. J. Appld. Phys.. 28 (1989), 95-96.

Another prior art approach has been to prepare thin films of high-temperature Bi-Sr-Ca-Cu-0 super- conductor on MgO substrates using the liquid phase

epitaxial (LPE) method employing flux. See H. Takeya et al., Jpn. J. Appl. Phvs.. 28 (1989), 229-232.

There have been other reports of producing Bi- Sr-Ca-Cu-0 (BSCCO) thin films using rf-magnetron sputtering and electron beam evaporation or chemical vapor deposition. H. Asano et al., Jpn. J. Appl. Phys.. 27 (1988) 1487; Fukutomi et al., Jpn. J. Appl. Phys.. 27, (1988), 576-579; T. Yoshitake et al. , Jpn. J. Appl. Phys.. 27, (1988) 1262; and H. Yamane et al., Jpn. J. APPl. PhVS.. 27 (1988) 1495.

These films would have potential applications to electronic devices as superconductor substrates, but are not suitable for winding coils and other uses requiring extreme flexibility. See also United States Patent No. 4,965,2245;

Japanese Patent Nos. 337095 and 026409; Glowaski et al. "External and Internal Diffusion of Oxygen Superconductivity YBa ₂ Cu ₃ 0 ₇ Composite" Matl. Res. Soc. Svmp. Proc.. Vol 99, April 1988; and Jin et al., "Fabrication of Dense Ba ₂ YCu ₃ 0 ₂₅ Superconductor Wire by Molten Oxide Processing", Appl. Phvs. Lett.. Sept. 28, 1987.

The superconductor material of the present invention is a novel, flexible and mechanically tough ceramic composite which can be flexed and wound or generally configured as needed for a variety of applications. When subjected to intentionally severe mechanical stress, such as repeated bending in half and restraightening, the crack that results is limited to the stress or fold line.

Another aspect of the present invention relates to high strength, fracture tough, ductile high temperature oxidatively stable ceramic composites which are suitable for use in gas turbine engines and other high temperature applications. By adjusting the ordered

ductile array and ceramic phase, composites suitable for lower temperature uses are also provided.

The cost, reliability and performance of high temperature performance engines including gas turbine engines depends upon the high temperature materials used in their construction. Current aircraft, missile, ground based vehicle, and ship propulsion gas turbines, as well as stationary power generation turbines, all use metal superalloys to provide needed high temperature performance. These metals, used in polycrystalline or single crystal forms, impose several important penalties on the overall gas turbine. These alloys are high density metals and thus they contribute to overall engine weight and penalize the engine thrust to weight ratio. This high density in rotating parts is also a major cause of high stresses generated during engine operation which limit rotary speeds and the fatigue life of high temperature components.

The elements employed in these high density superalloys, nickel, chromium, columbium, cobalt and molybenum, are, in many cases, available only from limited sources of supply. As such, Co and Cb and less abundantly used elements are referred to as so called strategic elements. Their availability can be disrupted in times of peak demand.

At very high temperatures, these metal superalloys are severely limited due to their propensity to creep under applied stress. Current attempts to increase engine efficiency have caused the operating temperatures to be increased beyond those sustainable by these metals alone. Hence they require cooling by low temperature air which is forced through passages in the blades. This is done at penalty to overall engine efficiency. Attempts have been made to overcome the deficiencies by using ceramic materials such as silicon

carbide and silicon nitride. These ceramics are capable of operating at temperatures well above those of metal superalloys and are much lower in density (e.g. 3.2 gm/cm ³ as opposed to densities of between 7.19 gm/cm ³ to 8.90 gm/cm ³ ) . These attempts have been hampered because ceramics are brittle. They fail readily in the presence of the stress imposed by the operating conditions and service induced impact.

There have also been attempts to develop high temperature gas turbine materials from carbon fiber reinforced carbon matrix composites. These materials, however, have two major advantages over the superalloys and silicon carbide and nitride ceramics. First, they are very low in density (2 gm/cm ³ ) and secondly, they maintain their strength and toughness at extremely high temperatures. Unfortunately, however, carbon is easily decomposed by oxidation at elevated temperatures in a gas turbine and the utility of these materials depends on the development of protective coatings and oxidation inhibitors which has been difficult because of the extreme thermal fluctuations in a gas turbine engine. Further, their fabrication costs are high and reliability under stressful operating conditions and in- service impact is questionable. United States Patent No. 4,626,461 discloses a multilayered silicon carbide fiber-reinforced ceramic or glass made up of a plurality of ceramic layers, with each layer reinforced with a plurality of unidirectional continuous length silicon carbon fibers, or random discontinuous silicon carbon fibers which are useful in fabricating gas turbine engine and composite parts.

United States Patent No. 4,846,866 discloses alkali-free alkaline earth aluminosilicate glass matrix reinforced with silicon whiskers and continuous ceramic fibers.

United States Patent No. 4,263,376 discloses the production of reinforced composite materials consisting of discontinuous graphic fibers embedded in a glass matrix. United States Patent No. 4,314,852 discloses the fabrication of reinforced composite articles consisting of continuous silicon carbide fibers embedded in a glass matrix.

United States Patent No. 4,464,192 discloses the preparation of reinforced composites consisting of whiskers or chopped fibers embedded in a glass matrix.

United States Patent No. 4,837,230 discloses a ceramic/ceramic laminar structure comprising multiple layers of ceramic matrix material on a ceramic reinforcement fabric. While these materials are disclosed as having high flexural and tensile strength and being useful in high temperature applications such as rocket motor insulation and turbine blades, combustion chambers and after-burners for jet motors, these materials only attain flexural strength of from 7.4 to 34.9 ksi whereas the ultimate bend strength of the composites of this invention exceeds 70 ksi.

United States Patent No. 4,861,229 discloses a ceramic-matrix composite nozzle assembly for a turbine engine. The ceramic is reinforced with woven carbon fibers impregnated with a silicon carbide matrix.

While the prior art has been successful in providing high-strength, tougher ceramic materials, they still suffer from the basic problem of being rigid; that is, being unable to flex sufficiently under mechanical stress which ultimately leads to cracking and failure.

Depending upon the selection of the ceramic, this tough, ductile composite may be fabricated into gas turbine engine and component parts, or hard abrasion resistant ceramics which can be used in grinding

operations, as nozzles or other applications where metals are too soft, or as chemically inert, high heat conductivity ceramics which have applications as heat sinks in electronic circuit boards. Summary of the Disclosure

In its broadest aspect, the present invention provides a planar ductile, flexible ceramic composite comprising: a ceramic phase; a ductile metallic phase extending throughout the ceramic phase, the ductile metallic phase comprising an ordered, continuous, interconnected array forming a repeating pattern structure, the ductile metallic array being embedded within, surrounded by and in intimate contact with said ceramic matrix throughout the composite body so as to provide a high degree of interface between the ceramic and the metal. The composite of this invention has special mechanical properties. The composite of this invention exhibits flexibility, ductility and toughness which permits a broad range of applications as described in detail below.

In one embodiment, the present invention provides a ductile solid electrolyte composite wherein the ceramic phase is an ionically conductive ceramic. The preferred form of the solid electrolyte is a planar structure which can be fabricated into various configurations such as tubes, arcuate sections, corrugated structures or flat plates.

In another embodiment, the present invention provides a solid state device for separating oxygen from air comprising a solid state electrochemical cell wherein the solid electrolyte is a ductile composite of this invention. Means are provided for connecting the cell to a power supply whereby when current is passed through the cell, oxygen is separated from air passing through the cell.

The solid electrolyte composite of this invention also permits the fabrication of a superior solid oxide fuel cell.

The composites of this invention may also be fabricated as ductile high-temperature superconducting ceramic composite characterized by a superconductive ceramic supported by a continuous, ordered, ductile array. The ductile metal array is intimately mixed within the ceramic phase of the material, and in a preferred embodiment, is generally embedded within and surrounded by the ceramic phase of the composite.

Structural ceramic composites are also provided by this invention. As with the above-described composites, the structural ceramic composites are characterized by a ceramic supported by a continuous, ordered, ductile, metallic array which is surrounded by and supports the ceramic.

High temperature structural ceramic composites may be prepared from crystalline or substantially crystalline ceramics, preferably aluminosilicates such as lithium aluminum silicates, magnesium aluminum silicates or other metal aluminosilicates such as sodium, potassium, magnesium, calcium, barium, yttrium, lanthanum and other suitable metals having a valence of 3+. Especially preferred are lithium aluminum silicates available from Dow Corning (LAS III) .

Brief Description of the Drawings FIG. 1 is an perspective view of one preferred oxygen generator embodiment in accordance with the present invention with portions cut away for clarity and understanding.

FIG. 2 is a fragmentary, cut-away elevation view showing greater detail of an electrolytic cell stack within the oxygen generator assembly. FIG. 3 is an enlarged elevation side view of a ductile ceramic multilayered single electrolytic cell

unit employed in the oxygen generator assembly of FIG. 1.

FIG. 4 is a 100X photomicrograph of the preferred embodiment of an open ductile array for the solid electrolyte ceramic composite employed in the practice of this invention.

FIG. 5 is a 60X SEM of a preferred embodiment of solid electrolyte composite having a repeating pattern of the underlying diamond structure from the ductile array of FIG. 4.

FIG. 6 is a 5OX optical photomicrograph of a solid electrolyte composite material employed in the practice of this invention.

FIG. 7 is a 6000X SEM photomicrograph of a section of a preferred embodiment of a solid electrolyte employed in this invention.

FIG. 8 is a photomicrograph of a section of the solid electroyte of FIG. 5 after it had been repeatedly bent 180" (in half) and straightened to determine the effect of intentionally excessive mechanical abuse.

FIG. 9 is a graph of conductivity vs temperature of a solid electrolyte composite using a 15 mole percent baria solution in bismuth oxide.

FIG. 10 is a graph of conductivity vs. temperature of a 20 mole percent solution of baria in bismuth oxide.

FIG. 11 is a graph of voltage vs. current of the electrolyte of FIG. 9.

FIG. 12 is a graph of voltage vs. current of the electroyte of FIG. 10.

FIG. 13 is an electrochemical schematic flow diagram of a ductile solid oxide fuel cell in accordance with the present invention.

FIG. 14 is a schematic of the basic cell and gas pathway design, shown in single segment form.

employed in a repeating planar stack solid oxide fuel cell in accordance with the present invention.

FIG. 15 is a graph of the log resistance vs. 1000/T(K) of a yttria stabilized zirconia composite of the present invention.

FIG. 16 is a schematic flow diagram of a manifolded multicell stack design in accordance with the present invention.

FIG. 17 is a graph of the electrical resistance vs. temperature (Kelvin) including the >77K T _c superconducting transition of a superconducting composite in accordance with the present invention.

FIG. 18 is a 75X photomicrograph of two side- by-side silver-plated copper cables employed in the preparation of the composite employed in generating the data depicted in FIG. 17.

FIG. 19 is a 100X SEM of a flexible superconducting composite in which the outer sheath of ceramic had been intentionally removed to expose the interior copper wires and the ceramic.

FIG. 20 is a graph of the equation of oxygen partial pressure of Cu/Cu ₂ 0 vs. temperature (C) .

FIG. 21 is a X15 SEM photomicrograph of an actual planar structural composite of this invention in a 180 degree bend.

FIG. 22 is a X20 SEM photomicrograph of the area of bend of the planar structural composite in a 180 degree bend as shown in FIG. 21.

FIG. 23 is a X140 SEM photomicrograph of a lithium aluminum silicate/Inconel 600 planar structure composite of this invention.

FIG. 24 is a X10,000 SEM photomicrograph of the bulk ceramic region of the composite shown in FIGS. 21-

23. FIG. 25 is graph of outer fiber stress (ksi) vs. load point displacement (mM) from three point

flexural testing of ductile ceramic composite in accordance with this invention having a an ultimate bend strength (U.B.S.) of 56,000 psi at 11.30 mM and 366 J/cm ² (J=Joules) . FIG. 26 is a graph of outer fiber stress (ksi) vs. load displacement point (mM) for a ductile structural ceramic composite of this invention having a U.B.S of 71,435 at 6.40 mM and 251 J/cm ² .

FIG. 27 is a chart comparing the absorbed energy to yield of the ceramic alone, fired metal preform alone, unfired metal preform alone and the composite.

Detailed Description of Preferred Embodiments

The ductile, tough solid electrolyte ceramic composite of the present invention comprises a regular, ordered, continuous, repeating array of ductile intersupported or interconnected, metallic fibers in intimate contact with the ceramic matrix so as to be substantially surrounded or embedded within it and supporting the matrix. The ceramic employed is a highly ionically conductive material. Preferred ceramic phase materials in which the fibers are embedded or surrounded are solid oxide electrolytes based on solid solutions of zirconia oxide, bismuth oxide or ceria oxides stabilized with a second component selected from a metal oxide wherein the metal ion has a valence of +2, +3, +5, or +6.

Other suitable oxides include, but are not limited to hafnia (HfO) , titania (Ti0 ₂ ) , other lanthanides such as ceria (Ce0 ₂ ) , samaria (Sm ₂ 0 ₃ ) , yttria (Y ₂ 0 ₃ ) , erbia (Er ₂ 0 ₃ ) , scandia (Sc0 ₂ ) , perovskites, pyrochlores, calcia (CaO) , magnesia (MgO) , gadolinia (Gd ₂ 0 ₃ ) or a combination of one or more of the above oxides. The preferred electrolyte ceramic phase is stabilized solid oxides. Preferred stabilized solid

oxides include yttria stabilized zirconia, gadolinia stabilized ceria, baria stabilized bismuth oxide, and erbia stabilized bismuth oxide.

In one preferred embodiment, ceramic composites were prepared from 8 mole percent ytterbia stabilized zirconia and 9 mole percent yttria-stabilized zirconia as the ceramic phase and compounded with planar 60 mesh square weave Inconel 600 or Inconel 600 expanded foil as the ductile phase/array to yield a sheet of solid oxide electrolyte composite suitable for a mechanically tough solid oxide fuel cell. This physical array allows for approximately 90% of the volume of the composite to be the solid electrolyte, resulting in a high level of ionic conductivity through the plane of the composite. The ytterbia materials were prepared from an alkoxide based sol-gel formulation in accordance with one embodiment of the present invention.

In another preferred embodiment, a bismuth baria rhombohedral system wherein the barium stabilized phase has a barium content of 15-25 mole percent of stabilizing oxide formula, preferably 15-20 mole percent, and most preferably 20 mole percent is employed as the ceramic phase of a solid electrolyte composite of the present invention suitable for use in fabricating cells for oxygen and nitrogen generators.

Ceria stabilized with a metal oxide wherein the metal ion has a valence of +2 or +3 may also be used in the practice of this invention as the ceramic phase of the solid electrolyte composite. Representative stabilizing agents are oxides of yttrium, scandium, gadolinium and other rare earth and alkaline earth matals.

As best shown in FIG. 4, a preferred embodiment of the ductile component of the solid electrolyte composite employed in the practice of this invention is an intersupported, planar array of metallic ligaments

forming a repeating diamond pattern. The line-of-sight openess of this array is about 65-70%. Ceramic volume fraction of the solid electrolyte ceramic composite composition is from 10% to 95 %. The preferred ceramic volume fraction of the final ceramic composite is about 90%.

The preferred material for the ordered, ductile array is a single layer of an open mesh metal structure. Especially preferred are expanded metal foils such as Haynes 214 expanded metal foil. Especially preferred is an expanded metal foil produced in accordance with this invention from a solid sheet of Inconel 600 foil with an original thickness of 0.003 inch. The solid electrolyte composite employed in the practice of this invention is a thin sheet-like structure having a thickness of 0.01 inch or less. It is preferred that the composite have a thickness of 0.003 inch or less. While one of the requirements of structures in which the composite is used, including an oxygen generator or fuel cell module, is that the components including the electrolyte composite be of sufficient mechanical strength to withstand stresses to which they will be exposed. This will normally dictate the thickness required. If very large sheets not laterally supported, as by current pickups, are desired greater thickness may be required.

Generally speaking, shape of the composite body is irrelevant to its operation. It may be square, rectangular, circular, pleated, corrugated, and the like. For best results it is preferred that the composite body employed in the electrolytic cell, which forms the essential element of the oxygen generator and solid oxide fuel cell of this invention, is at least 4 inches on a side, preferably 6 inches or more in diameter if round to provide an equivalent surface area.

23

Size will depend upon the end application. Portable oxygen generators for medical or other personal use, such as for firefighters, would use relatively small cells. For large, industrial applications, composite bodies having dimensions of 1 to 2 meters or more per side may be employed.

As shown in FIG. 5, in the solid electrolyte composite formulated in accordance with a preferred embodiment of the present invention, there is a repeating pattern of the underlying diamond structure of the ductile array. EDX analysis of the interface between metal and solid electrolyte ceramic showed a "metal oxide" with the composition CrNi ₂ O _χ . The EDX analysis of the metal ligaments was consistent with the published values of Inconel 600 while the bulk ceramic phase was consistent within the precision of the EDX unit with the intended solid solution of bismuth and barium oxides.

FIG. 6 is a 50X optical photomicrograph of a composite of this invention. The "diamonds" of ceramic oxide solid electrolyte with the interconnecting lines of metallic ligaments can be seen. Upon backlighting, the composite clearly showed its form with a yellow- orange transmitted light interrupted in a precise regular repeating array of opaque (metallic) lines. The optically transmitting regions were the diamond shaped ceramic filled subsections.

FIG. 7 is a 6000X SEM photomicrograph of a section of the solid electrolyte prepared in accordance with Example 2. The white occlusions are unreacted nearly pure bismuth oxide. The solid electrolyte composite of this invention was found to be quite flexible, capable of flexing out of plane by as much as 0.25 inch or more with finger tip pressure on a sample of about two inches in length. A sample was repeatedly

24 bent to 180° (folded in half) and straigthened to examine the effect of such mechanical abuse.

FIG. 8 is a photomicrograph of such a sample. As can be seen, despite the extreme mechanical abuse, a resulting crack only formed along the line of maximum stress or fold line. However, there was no crack propagating away from the fold line, and the crack that did appear did not even extend within a given, unsupported ceramic diamond area. The same ceramic composition, outside of the composite structure, would shatter. Prior art composite structures would not withstand such abuse.

FIGS. 9 and 10 depict the DC conductivity versus inverse temperature behavior of composites of this invention having 15 and 20 mole percent of BaO in the ceramic composition phase respectively. Solid electrolytes should exhibit a linear relationship between the log of the conductivity (resistivity) and the inverse Kelvin temperature. The 15% solid electrolyte showed a small but finite level of the monoclinic phase by X-ray diffraction (XRD) , while the 30% baria solution showed only the pure rhombohedral phase at the precision level of XRD. The 20% curve (FIG. 10) exhibited a somewhat steeper slope that the 15% baria ceramic composite. This results in an energy of activation for the 20% baria material of about 26 KC/mole, slightly higher than the 15% material. At the highest temperature measured, the curve for the 20% baria material may be exhibiting the change in slope as reported by Suzuki fJ. Mat. Sci.. 20 (1985) , 3125] and others for bulk ceramic.

FIGS. 11 and 12 depict the voltage versus current behavior of these 15 and 20% baria electrolytic ceramic composites. The vertical axis is the logarithm of the current in amperes or in current density. The horizontal axis is the voltage between the cathodic

electrode and the platinum reference electrode less the open circuit potential of 14-15 mV depending upon temperature.

Generally speaking, the high melting temperature electrolyte composites employed in the practice of this invention are prepared by forming a slurry of fine metal oxide powder having a particle size under 1 micron to yield a doughy slurry, adding an organic binder, preferably under 0.25% of a binder such as polyvinylalcohol, pouring or otherwise distributing the ceramic phase solution over the ductile support array to be embedded therein, firing in a reducing or inert atmosphere to approximately 1000° to 1400°C, preferably 1200° to 1350°C and most preferably 1300° to 1350°C for from 1 to 24 hours, cooling and repeating the cycle until there is >90% density in the ceramic.

It is especially preferred to anneal the composite under a directed energy source such as a carbon dioxide laser or electron beam. In this way, the ceramic can be heated above its melting point, permitting it to flow evenly around the ductile array, while the metal remains under its melting point. Beam rastering rates of approximately 1 inch/sec work especially well. The following examples further illustrate the ductile solid electrolyte composite of invention.

Example 1 A slurry of a molten hydrate melt of Ce(N0 ₃ ) ₃ -6H ₂ 0, Gd(N0 ₃ ) ₃ -6H ₂ 0 and Ce_ ₈ Gd ₂ 0 _1>9 was applied to Inconel 600 mesh (60 mesh) steel, suspended in a furnace with 0.008 inch Inconel wire and fired to approximately 650°C. After cooling, the composite was laser annealed using a C0 ₂ laser having a 10.6 μ wavelength, 600 W, 3/8 inch by 0.005 inch. The ceramic melted, flowed and refroze without melting the metal support matrix. The annealing was done under flowing argon. A sample of

composite was held on a computer controlled table and rastered under the C0 ₂ laser beam at a rate of 1 inch/sec. Very slow rates vaporized the sample and faster rates insufficiently melted the ceramic. Example 2

A bismuth baria solid electrolyte composite wherein the ceramic phase contains 20 mole percent baria was prepared as follows. Bismuth oxide (Bi ₂ 0 ₃ , 150 g) was mixed with Ba(N0 ₃ ) (11.7 g) and Bi ₂ 0 ₃ (21 g) and poured into an alumina tray containing 340 g of 16% BaO. Upon melting, the final composition is (Bi ₂ 0 ₃ )_ ₈₀ (BaO) _ ₂₀ . The mixture was well stirred and heated to a temperature of 900°C for about 1-1/2 hours, then cooled to 860°C. A ductile array of Inconel 600 expanded metal foil having a line-of-sight openess of about 70% and forming a regular, structured, repeating diamond pattern was preoxidized by heating to approximately 700°C for about 1-1/2 hours in air. The preoxidized metal support or ductile array was dipped into the liquid ceramic phase to coat the ductile array with the ceramic phase, cooled and annealed at a temperature of 780"C.

Example 3 A sheet of Whatman 541 filter paper was placed on a 90 mm diameter Buchner funnel. The funnel assembly was covered with a 1/8 inch thick neoprene gasket sheet having a 90 mm diameter. The neoprene gasket had a rectangular hole somewhat smaller than the composite sample. A second piece of rubber neoprene was placed in the hole on top of the Whatman 541 sheet to physically support the composite of Example 2 without sealing. The system, without a composite sheet, was wetted with ethanol and allowed to set. The composite prepared above was placed over the neoprene hole and sealed in place with a neoprene gasket. A thin layer of the above bismuth baria slurry of Example 2 was painted on and allowed to dry for about

5 minutes. The sample was dried at 60°C, recoated on the opposite side and dried again. The composite was then fired at 700°C in air for about 20 hours. A second coat of the bismuth baria slurry was applied and the composite fired at 700°C for 15 hours in air. A third coat was applied and fired at 700°C in air for 1 week.

Example 4 A solid electrolyte composite was prepared following the method of Example 2 with the following modifications. The surface of the composite was painted with a slurry of BiBaO _χ and 20% polyethyleneimine (50% aqueous) under suction in a Buchner funnel. The composite was dried at 125°C, the opposite side coated and dried as above. The composite was placed under a weight and fired in air at 680°C for 20 hours, cooled to room temperature and both sides were painted with 1 % aqueous polyethylenimine and partially dried. Both faces of the composite were painted with silver palladium paint and dried at 150-200°C. A second coat of silver palladium paint was applied and the composite fired at 350"C for 1 hour in air and cooled to room temperature. The composite cell was placed in a furnace under a weight and the temperature raised to 700°C, held for 10 minutes, reduced to about 550°C, and then heated in air at 700"C for 14 hours.

Example 5 A cell body was constructed from a section of standard schedule 40 three inch SS316 pipe with external standard threads cut in one end. A planar disc of Inconel 600 expanded metal foil was tack welded on the end without the threads. The expanded foil disk was five inches in diameter. The outer one inch was cut radially into tabs approximately 0.5 inch centered over the unthreaded end of the pipe section and the tabs were bent down over the exterior sides of the pipe. The tabs were tacked in place with five welds of approximately

1/16 inch in diameter arranged in a three/two pattern with the three at the far (wider) end of the tab. Additional welds were tacked every two or three mm along the upper rim of the pipe at approximately 1/8 inch intervals.

Following the method of Example 2, a ceramic phase of bismuth baria oxide solid solution was prepared and melted in an Inconel 600 deep drawn crucible. The crucible was placed in an oven and heated to 925°C for about 30 minutes. The crucible was periodically swirled to insure a chemically homogeneous melt and the temperature was reduced to approximately 850°C.

During the 30 minute thermal hold of the above procedure, the cell body was placed in the oven along side the crucible to heat it to the temperature of the melt. This is important because if cold metal is dipped into the melt, the relatively large thermal mass of metal cools the melt to below solidifying temperature which results in the melt freezing and the pipe section fusing to the frozen ceramic mass. It generally requires a minimum of 30 minutes to raise the metal to the appropriate temperature.

The furnace door was opened and the metal section grasped with long tongs and dipped into the open-topped crucible containing the melt. The cell was promptly removed from the furnace and placed on a concrete surface to cool. The composite top of the cell cooled to room temperature within seconds, although the pipe section took several minutes to cool. The cell unit was examined for pinholes and none were found.

Silver palladium paste was applied to the interior and exterior surfaces of the composite. The paste was dried at 110°C for 20 minutes and fired at 700°C for an additional thirty minutes. Several coats were applied using this procedure.

The room temperature electrical resistance between the interior coat (the anode, oxygen evolution site) and the exterior coat (the cathode, the oxygen dissolution site) was >30,000,000 ohms, the limit of the digital multimeters. This indicates the electrodes were not short-circuited. Electrical resistance between any two points on a given electrode at room temperature was about 0.2 ohms or less.

The exterior electrode was approximately two inches in diameter and did not make direct electrical contact with the pipe. The silver paste of the interior electrode was intentionally spread onto the interior walls of the pipe, making an electrical connection between the housing and the interior electrode. There was no measurable room temperature electrical conductivity between the pipe and the exterior electrode.

A SS316 reducing union piece was then threaded onto the open end of the pipe section using high temperature thread sealant. The small end of the reducing union was connected to 1/8 inch stainless steel tubing using a SS 316 swagelock adapter. The 1/8 inch tubing extended out of a hole in the top of the furnace. About 18 inch of small diameter tubing extended out of the furnace. The "cold" end of the SS tube was connected to a section, about 3 feet long, of standard 1/8 inch inside diameter (i.d.) Tygon tubing. This was the gas circuit.

When electrical power was applied to the cell, at temperatures of about 650°C, oxygen was produced at the anode. This was detected by immersing the end of the tygon tubing in a small dish of water and seeing a steady stream of bubbles. In the absense of electrical power, the flow of oxygen gas bubbles ceased. Example 6

An aqueous slurry of fine (Zr0 ₂ )_ ₉₁ (Y ₂ 0 ₃ ) _Q9 powder +0.25% polyvinylacetate was laser annealed using a C0 ₂ laser having a 10.6 wavelength, 600 , 3/8 inch line by .005 inch. The ceramic melted, flowed and refroze without melting the metal support matrix. The annealing was done under flowing argon. A sample of composite was held on a computer controlled table and rastered under the C0 ₂ laser beam at a rate of 1 inch/sec. Very slow rates vaporized the sample and faster rates insufficiently melted the ceramic.

Example 7 The slurry of Example 6 was painted onto Inconel 600 stainless steel foil and dried at 115 °C for 20 minutes. The sample was suspended in a 2 inch outside diameter (O.D.) mullite furnace tube in a silicon carbide electrically heated furnace, was flushed with argon and the argon flow left on. The temperature was raised at 5 ^β C per minute to 1340*C for 3 hours and cooled to room temperature at 5°C per minute. The procedure was repeated two more times. On the last slurry application the slurry was sucked into pores of the composite under vacuum on a Buchner funnel. The last firing was for 15 hours.

The solid electrolyte composite of this invention has a number of applications. It may be fabricated into cells for gas separation systems such as oxygen or nitrogen generator systems. Referring to FIG. 1, depicting one preferred embodiment, oxygen generator 10 comprises oxygen separator module 11. Oxygen separator module 11 comprises one or more individual solid electrolyte composite cells 30 (FIG. 3) , placed in multiple plate stacked form 12, as shown in this view. In this preferred embodiment, the plurality of solid electrolyte cells are placed in modular stack array in which the individual rolid electrolyte cells are electrically connected in series.

Bipolar stack housing 13 carries electrode connector unit 14. Power control unit 16 may be disposed between power source 17 and the oxygen separator module by current conductors 15. The oxygen separator unit may be either battery powered or connected into a central remote generation source.

In operation, air is drawn into intake unit 18 by air collection fan 19. The air is heated in heat exchanger 20 and travels into oxygen separator module 11, via preheated air intake line 25, where it is separated into oxygen and oxygen depleted air. The oxygen is drawn from oxygen separation module 11 via oxygen conduit 21. Oxygen depleted air exits module 11 through conduit 22. The hot air and oxygen are passed through the heat exchanger, and the resulting cooled oxygen product exits through tube 23 while the cooled depleted air is released through vent 24.

Referring to FIG. 2, the interior detail of the electrolyte plates and plenum chambers within electrolyte stack housing 13 is shown. Plates 30, shown in parallel sheet arrangement, are electrically connected in series to each another by connectors 34. While these are depicted as separate elements in this view, fabrication may be simplified by having the plates formed in block I cross-sectional form so that each contacts those adjacent, with the end plate then having a block C cross section. Plenum chamber walls 31 are positioned on each side of all electrolyte plates in horizontal relationship so as to form a continuous stack within the housing. The plates are energized by the application of current.

Chamber walls 31 together with plates 30 form air plenum chamber 33 and oxygen plenum chamber 32. The chambers are continuous through housing 13 and are gas isolated from one another. Spanning the chambers are current pickups 35 formed of electronically conductive

material such as metal structures fabricated in the form of wool, expanded metal pieces, posts, rods, channels, ribbons or mesh which serve to pass current across the chambers but do not unduly impede gas flow. In operation, preheated air enters the top of the stack into chamber 33 at a temperature generally slightly above 600"C. Oxygen within the air is transported in the form of oxide ions through the solid electrolyte plate and into the opposing oxygen plenum chamber 32. All air plenum chamber segments in the housing are gas parallel while all oxygen plenum chamger segments are likewise in gas parallel to one another and gas isolated from the air plenum. During passage of gases through the stack, application of a current causes a temperature increase, stabilizing typically at a temperature of from 600" to 700 ^β C.

Oxygen which has been transported through each electrolyte plate 30 as the oxide ion collect in the oxygen plenum chamber on the opposing sides of the plates. Pressure build up from the transport of oxygen into the plenum chamber units causes oxygen flow to commence. The oxygen depleted air and the oxygen gas travel separately in gas parallel through the module and are conducted into the heat exchanger 20 through conduits 21 and 22 where cold air collected through intake unit 18 (Fig. 1) is heated by the depleted air which is then vented, and by the oxygen product gas which is cooled in the heat exchanger and collected.

During normal operation, the system usually requires no supplemental heat addition but runs at steady temperatures. An auxiliary heater, preferably located in the heat exchanger, may be required during start up until steady operating temperatures are achieved. The plenum walls are formed of material that is electrically conductive and will withstand exposure to

heated oxygen without being unduly oxidized. Preferred materials include 300 series stainless steel, 400 series stainless steel, Incolloy 800 HT, super alloys including Inconel 600 or 601 and Haynes 214. In spacing the walls and electrolyte plates in stacked arrangement, as depicted, it is advantageous to have the height of the air plenum chamber units greater than that of the oxygen plenum chamber. This is to accommodate the larger volume of air passing through the module, compared to the small volume of oxygen being withdrawn. A ratio of 2:1, air plenum to oxygen plenum has been found to be quite suitable. Increasing the height of the chambers has the advantage of reducing pressure drop caused by the packing, but it has the disadvantage of increasing the overall size of the stack.

The current pickups 35 within the plenum chambers contact the electrodes and the bipolar plates. Electricity passes throught the plenums perpendicular to the gas flow which is horizontal. The pickups are suitably composed of the same materials as that used to form the plenum walls 31. While good electrical contact is needed to avoid undue voltage drop, the more porous the pickups are, or the less resistance they offer to gas flow, the better they operate by lessening restriction of gas flow and avoiding excessive pressure drop through the chambers. Connectors 34 shown as providing electrical contact between plates 30 may be redundant when suitable conductive material is used for the current pickups.

FIG. 3 depicts in large scale the layers forming the cell 30 which is the essential component of this invention. Solid electrolyte ductile ceramic composite 40 is the center layer of the cell. It is preferably sandwiched between mixed conductor layers 41 formed of solid electrolyte doped with multi-valent ions

which make the layers both ionically and electronically conductive. The mixed conductive layers 41 are coated with electrically conductive gas porous electrode thin layers 42 which form an anode and a cathode on the surface of the cell.

The mixed conductor component 41 is formed of a solid electrolyte such as bismuth oxide or zirconia. The multi-valent doping materials which are suitable include praseodymium, terbium, cesium, iron and chromium.

The electrically conductive electrode surfaces on each side of the cell are preferably silver, silver alloys or conductive oxides such as perovskites.

The ductile, tough solid electrolyte ceramic composite employed in the practice of the present invention comprises a regular, ordered, continuous, repeating array of ductile intersupported or interconnected, metallic fibers in intimate contact with the ceramic matrix so as to be substantially surrounded or embedded within it and supporting the matrix. The ceramic employed in this aspect of the invention is a highly ionically conductive material. Preferred ceramic phase materials in which the fibers are embedded are solid oxide electrolytes based on solid solutions of bismuth oxide and a second component selected from a metal oxide wherein the metal ion has a valence of +2, +3, +5, or +6. The purpose of the stabilizing agent is to hold the bismuth oxide in the preferred crystal lattice symmetry in a temperature range at which it would otherwise convert to a less conductive, or non- conductive polymorph. In general, the optimum lattice symmetry is face centered cubic. However, the alkaline earths form rhombohedral phases with bismuth oxide and these phases are also extremely conductive.

Table I is a partial list of bismuth solid oxide electrolytes which may be employed in the practice of this invention.

TABLE I

Bismuth Solid Oxide Electrolytes

KcTaraαc*

JES, 12A, 1563 (197 ^* 9 JES, 221, 1563 (1977) JAE, _l, 187 (1975) KRB, 21.1215 (1986) JAE, 2, 65 (1973) JAE, 2, 97 (1972) JAE, 2 , 97 (1972) JAE, 2.97 (1972) JSSC, 2S., 173 (1981) JSSC, J£9_, 173 (1981) JSSC, 2.13 (1981) JSSC, 22., 173 (1981) JAE, .10, 81 (1980) JAE, 2« 3" ¹ ( ¹ 977) JAE, _ , 197 (1975) -JSSC, Jj>, 317 (1976 JKS, 20. ^" 312511985 JAE, .12, 235 (1962) JAE, 21, 235 (1982) JAE, J , ₄ 7 (1985) JAE, -.' 44 (1985)

The first column in Table 1 is the stabilizing agent which may be a metal or metaloid such as an alkaline earth, a lanthanide or a transition metal. The second column is the composition of the solid solution which is reported to exhibit the highest conductivity for a given pairing of metal ions. The third column lists the reported ionic conductivity at the somewhat arbitrary temperature of 450"C. The barium stabilized phase, with a barium content of 15-20 mole percent of

stabilizing oxide formula as written is one of the most conductive.

The fourth column in this table is the transference number for the oxide conduction at the stated temperature. The transference number for oxide conduction is the fraction of the current which is carried by oxide ions, instead of by semiconducting or metallic mechanism. In general, a useful solid electrolyte must have a transference number in excess of 95% , meaning that approximately 5% or less of the current is carried by non-electrolytic mechanisms. Note that all of the compositions except for cadmium and terbium meet this fundamental criterion. The zero transference number of the cadmium phase indicates that it is a pure electronic (not electrolytic) conducter.

The fifth column lists literature references. In Column 5, JES refers to Journal of the Electrochemical Society, JAE refers to Journal of Applied Electrochemistry, MRB refers to Materials Research Bulletin, JSSC refers to Journal of Solid State Chemistry, and JMS refers to Journal of Materials Science.

It is presently preferred to employ the bismuth baria rhombohedral system wherein the barium stabilized phase has a barium content of 15-25 mole percent of stabilizing oxide formula, preferably 15-20 mole percent, and most preferably 20 mole percent.

Ceria stabilized with a metal oxide wherein the metal ion has a valence of +2 or +3 may also be used in the practice of this invention as the ceramic phase of the solid electrolyte composite. Representative stabilizing agents are oxides of yttrium, scandium, gadolinium and other rare earth and alkaline earth matals. The cell of FIG. 3 was tested under the following conditions. Cell 30 comprised a symmetric

"sandwich" with the ionically conductive solid electrolyte composite as the center layer 40. On either side of the solid electrolyte center layer are mixed conductive layers 41 which are coated with electronically conductive porous metallic layers 42. Woven Inconel 600 cloth was coated with commercial silver based paste to act as current collector and to allow for the passage of gases in passageways 32 and 33 which serve as the oxygen and air plenums respectively. The plenum walls were 1/8 inch thick sheet of Haynes 214 alloy to serve as current pickups. A single fine platinum wire was placed on the cathodic side of the cell composite in contact with the electrolyte but not in direct electrical contact with the electrodes or the metal current collectors of pickup plates. There was static air on both sides of the sample at the beginning of each test. The cathode side became depleted in oxygen while the anodic side oxygen partial pressure increased. The test cell was placed in an electric Nichrome wound muffle furnace and the temperature was raised to the appropriate level. A thermocouple, independent of the furnace thermocouple, was placed in direct contact with the test array at all times. Voltage was applied with a small 15 amp DC power supply from Darrah Electronics. The voltages, currents, and DC resistivities were read using digital multimeters. FIGS. 9 and 10, described about, are graphs plotting data points measured with the above apparatus.

The oxygen generator system of the present invention has a wide variety of applications. It can be fabricated into a light-weight, portable unit for medical use or use by firefighters and other individuals who are working in situations where an independent oxygen supply is needed. It can be fabricated into large industrial units to supply oxygen requirements in industrial processes. It can be used as an oxygen

source for operations such as welding. It is versatile, its only moving part is a fan, it does not require replacement of consumables such as hydrogen peroxide or catalyst, and can continuously generate a supply of oxygen for prolonged periods of time. It can be fabricated into any desired size or shape to meet the desired application.

The solid oxide fuel cell of the present invention employs a mechanically rugged, ductile solid electrolyte ceramic composite comprising a regular, repeating array of ductile ordered, continuous metallic fibers substantially surrounded by or imbedded in and supporting a ceramic matrix as the solid electrolyte in either a planar, arcuate, folded or tubular cell. The cell is connected to an electrical load, whereby when current is generated from the cell, fuel and air are converted to electric power.

Referring to FIG. 13, an electrochemical schematic for a solid oxide fuel cell in accordance with this invention is shown. Planar cell 100 comprises a center layer 111 which is the ductile electrolyte composite unique to this invention. A mixed conductor (ionically and electronically conductive) layer 112 is disposed on either side of the planar center layer 111 in intimate contact therewith. Outer layers 113 and 114 are electrodes (cathode and anode respectively) . Fuel plenum chamber 115 is disposed on the anode side of the cell and air plenum chamber 116 is disposed on the cathode side of the cell. Current collectors 117 are included within the plenum chambers, contacting each electrode layer and extending through each plenum chamber to contact the bipolar plenum wall (not shown) . While positive electrical contact across each chamber segment is required, the chamber current collectors are composed of material fabricated or structured so as to

allow free passage of gas and not cause a significant drop in pressure through the chambers.

FIG. 14 illustrates the physical arrangement of the a basic, repeating manifolded cell design which is employed in a stack arrangement for a solid oxide fuel cell in accordance with this invention. Cell 130 comprises a symmetric "sandwich" with the ionically conductive solid electrolyte composite as the center layer 131. On either side of the solid electrolyte center layer 131 are electronically conductive porous metallic plenum chamber current pickups 132 and 133. Woven Inconel 600 cloth was coated with commercial silver based paste to act as current collector and to allow for the passage of gases in the plenum passageways 134 and 135 which serve as the fuel and air plenums respectively. Mixed conduction (ionically and electronically conductive) sub-electrode layers 136 are disposed between the solid electrolyte composite 31 and outer thin layer electrodes 137 and 138. Bipolar plenum chamber walls 139 were 1/8 inch thick layers of Haynes 214 alloy sheet which are in contact with the current pickups 132 and 133.

FIG. 15 is a graph depicting the logarithm of the normalized DC resistance of a section of a yttria fully stabilized zirconia stabilized composite (9% yttria) as a function of the reciprocal Kelvin temperature. This sample did not have electrodes in place, but was instead placed between two porous silver/palladium coated metal grids as current pickups. The DC resistance of the composite at 1000°C was about 100 ohms. Most of the resistance was due to the absence of normal electrodes and is reasonably linear for the test conditions. This figure depicts the resistance of the sample over a range of four orders of magnitude. The measured temperature included temperatures from 450"C to 1000°C.

FIG. 16 is a schematic representation of a bipolar manifolded multicell stack employing the solid oxide electrolyte composite of the present invention. The ratio of size of air plenum chamber to fuel plenum chamber is 2:1 because of the greater volume of air passing through the system, compared to the fuel gas volume. The fuel and air gas streams do not come into contact with each other. They are kept isolated. Gas manifolding is preferably in parallel with current interrupts in the manifold to avoid short circuits between cells. The electrical power is in series.

The fuel cell of this invention may be incorporated into transportation vehicles, utility power systems, sub-utility power systems, portable power devices, and the like.

Examples 6 and 7 are illustrative of the invention as it applies to solid oxide fuel cells.

Example 9 An aqueous slurry of fine Zr0 ₂₉₁ Y ₂ 0 ₃ powder +0.25% polyvinylacetate was laser annealed using a C0 ₂ laser having a 10.6 wavelength, 600 W, 3/8 inch line, .005 inch. The ceramic melted, flowed and refroze without melting the metal support matrix. The annealing was done under flowing argon. A sample of composite was held on a computer controlled table and rastered under the C0 ₂ laser beam at a rate of 1 inch/sec. Very slow rates vaporized the sample and faster rates insufficiently melted the ceramic.

Example 10 The slurry of Example 9 was painted onto

Inconel 600 stainless steel foil and dried at 115 °C for 20 minutes. The sample was suspended in a 2 inch O.D. mullite furnace tube in a silicon carbide electrically heated tube, was flushed with argon and the argon flow left on. The temperature was raised at 5"C per minute to 1340°C for 3 hours and cooled to room temperature at

5°C per minute. The procedure was repeated two more times. On the last slurry application the slurry was sucked into pores of the composite under vacuum on a Buchner funnel. The last firing was for 15 hours to densify.

Another aspect of the present invention provides a novel ductile high-temperature superconducting ceramic composite comprising a superconductive ceramic supported by a continuous, ordered, ductile array. The ductile metallic array is intimately mixed within the ceramic phase of the material, and in a preferred embodiment, is generally embedded within and surrounded by the ceramic phase of the composite. The high temperature flexible superconducting ceramic composite of the present invention comprises a regular, ordered, continuous, repeating array of ductile continuous fibers which are generally embedded and surrounded by a superconducting ceramic matrix. The ductile continuous fibers are generally interconnected as discussed in detail below.

The ceramic phase of the composite of this aspect of the invention is a superconducting oxide. Complex copper oxides containing bismuth, strontium, calcium and copper oxides represented by the general formulae Bi-Sr-Ca-Cu-O _y (BSCCO) , which may optionally include lead or antimony oxides; barium yttrium copper oxides of Y-Ba-Cu-O _y (YBCO) , and thallium barium calcium copper oxides (Tl-Ba-Ca-Cu-O) are especially suitable high-T _c superconducting compounds which may be employed in the practice of this invention. Because of their toxicity, the thallium oxides are the least preferred. Especially preferred are the bismuth-strontium- calcium-copper oxides optionally containing lead or antimony oxides. These materials are reported in the literature and for purposes of this invention, may be

prepared by literature methods. See for example, A. Maeda et al., Jpn. J. APPI. Phvs.. 28 (1989) 576-579; U. Endo et al, Jpn. J. APPI. Phvs.. 27 (1988) 1476-1479; R. Sato et al., Jpn. J. Appl. Phys.. 28 (1989) 583-586; Y. Ibrara et al., Jpn. J. Appl. Phys.. 28 (1989) 37-40; H. Takeya et al., Jpn. J. Appl. Phys.. 28 (1989) 229- 232; S. Narumi et al., Jpn. J. APPI. Phys.. 28 (1989), 27-30; H. Enami, Jpn. J. Appl. Phys.. 28 (1989), 377- 379; Y. Hakuraku et al., Jpn. J. APPI. Phvs.. 28 (1989), pp 402-405; H. Tabata et al., Jpn. J. APPI. Phys.. 28 (1989) , pp 430-433) which are incorporated by reference herein.

The Bi-Sr-Ca-Cu-0 (BSCCO) system has several notable advantages in comparison with the YBa ₂ Cu ₃ O _χ (YBCO) system. BSSCO includes a higher T _c phase than YBCO. BSSCO does not easily create oxygen vacancies around the Cu-site which directly influence superconductivity in the case of YBCO. BSSCO is chemically stable in the presence of water. Thus, the BSSCO superconducting ceramic system is presently preferred.

The superconductor ceramic phase may be conveniently prepared by heating the co-decomposed powders of Bi, Sr, Ca and Cu nitrates (optically including Pb) at 835°C for 84 hours at an oxygen pressure of 1/13 atm. , followed by slow cooling to room temperature according to the procedure of U. Endo et al., Jpn. J. Appl. Phvs.. 27 (1988), pp 1476; and S. Koyama et al., Jpn. J. APPI. Phvs., 27 (1988), pp 1861 which are incorporated by reference herein. The ductile reinforcement materials may be selected from electronically conductive materials, preferably silver, copper and electrically conductive alloys thereof or high mechanical strength stainless steel or super alloys or other high strength metals, or high mechanical strength metals plated with an electronically conductive material, preferably silver.

Depending upon the end use, and the configuration of the ductile array, the finished composite has an aspect dimension that appears to be either one, two or three dimensional. Suitable ductile arrays are multifilament braided cable, preferably, an extra-flexible braided copper cable, a stranded copper cable, preferably an extra-flexible stranded copper cable, a multi-strand concentric copper cable, an expanded metal foil or a woven metal mesh. The choice of a the particular ductile array depends upon the desired use. For electronic superconducting substrate applications, a planar ductile array is employed. For applications requiring bending or winding around a coil, or which require flexible lengths of superconducting ribbon or wire for other applications, cables or wire are employed.

In a preferred embodiment, the ductile array for a bismuth lead strontium calcium copper oxide having an initial composition of is a fine, multifilament copper cable made from .002 inch wire braided to a 0.010 inch multifilament cable. The wires may be individually coated with silver prior to braiding.

Preferred metals are silver-plated Inconel 600 expanded foil or woven mesh, other superconducting alloys such as Haynes 214, multifilament copper wire and silver coated multifilament copper wire.

Silver plated copper multifilament cable is commercially available from Hudson Wire Co., Ossining, NY. The Inconel alloy ductile array material was silver plated by first dipping in pure molten silver nitrate at about 350°C and then heating in a 700°C oven. This procedure results in a relatively high surface area, thick, shiny silver film without impurities. It is believed that the high tempertaure decomposition of the molten, highly oxidizing nitrate salt on the Inconel

metal resulted in the formation of an inconel/silver interfacial oxide layer which helps to protect the subsequent superconducting phase from possible contamination with nickel and chrome oxides. The application of the ceramic phase to the ductile array requires carefully controlled conditions to avoid melting of the metal phase. The ceramic phase is prepared by conventional ceramic techniques and heated to form a melt. The ductile metal array is coated therewith by passing the metal through the melt at temperatures of from 900° to 1200"C, preferably 900° to 1100°C and most preferably 900°C to 1000°C for from 1 to 10 seconds. After cooling, the material is annealed in a controlled oxygen gas atmosphere at about 10 ppm oxygen at temperatures of from 750° to 860°C preferably about 830 ^β C for 60-70 hours. It is especially preferred to anneal the superconducting composite of this invention using energy beam techniques such as with a C0 ₂ laser beam (10.6 wavelength, 3/8 line x 0.005 inch"). Using a computer controlled table, the multiphasic superconductor is rastered at a rate of approximately 1 inch/sec. Very slow rates or high power rates vaporize the material and faster rates are unsufficient to anneal the ceramic. After annealing, the presently preferred flexible superconductor composites of this invention yield superconducting phases of from 80-110K. By carefully controlling the oxygen partial pressure to about 10 ppm, the copper metal ductile array is retained as the metal and the >77K superconducting transition was observed with composites of Bi_ ₈₀ Pb _0ι33 Sr ₁ _ ₈₈ Ca ₂ _ ₀₁ Cu ₃ O _x and Bi ₁₆ Pb ₀₄ [Ba ₀₂ Mg _{0 jj} Sr, ₆ ]Ca ₂ CuO _χ . In the latter ceramic phase, a portion of the strontium was replaced by equal molar portions of barium and magnesium and in a further modification, five mole percent silver was added to the composition. The partial substitution of barium and

magnesium for 20% of the strontium is reported by Ashizawa, Ja . J. Appl. Phys.. 28, 1989, p. 1140 to provide the preferred 2233 phase over a broader range of annealing conditions. Essentially unlimited lengths of superconducting wire or ribbon can be drawn through a carefully controlled atmospheric melt to produce the flexible superconductors of this invention.

The superconducting composite materials prepared from cuprate superconductors and copper metal are only thermodynamically stable if the copper metal is noble relative to the other materials in the composite. In particular, the bismuth oxide must not be reduced out of the ceramic to form bismuth metal due to association with the "reducing" agent copper metal. Accordingly, thermodynamic Gibb's energy of reaction calculations at a variety of temperatures were conducted on the following approximating chemical equation:

Bi ₂ 0 ₃ (l) + 6CU(S) •= 3CU20(S) + 2Bi(l) The Gibbs energies of bismuth and copper as the oxide in the superconducting phases were approximated as the pure phases. As the superconducting phases will spontaneously form from the discrete oxides, it is probable that the pure oxides are less stable than the superconducting compound oxide.

For the phases to be spontaneously stable toward each other (copper is "noble") the Gibb's energy of the above reaction needs to be positive. The following table was calculated. T (K) G(RX.KC)

77 +27.1

273 +40.7

673 +79.4

1000 +115.9

Based on this calculation, the copper is noble relative to the superconductor and the phases should

spontaneously co-exist without chemical reduction. At high temperatures where chemical kinetics might drive the reaction, the formula as written becomes less favorable. The physical analysis of superconducting composite of this invention has been consistent with this preliminary calculation.

Appropriately annealed superconducting ceramic composite where the ductile material is copper metal is produced by carrying out the annealing at an oxygen partial pressure low enough to kinetically slow down the oxidation of copper metal to acceptable levels, yet high enough for the ceramic phase to be stable, and not be reduce to copper metal.

A second group of thermodynamic calculations as a function of temperature was carried out on the model system:

2CU20(S) = 4CuO(s) + 02(g)

The results of this calculation are depicted in FIG. 20 wherein the vertical axis is the oxygen partial pressure in the annealing furnace and the horizontal axis is the temperature of the annealing furnace. Experience has shown that at low oxygen partial pressures, the appropriate annealing temperatures for this ceramic superconductor system is in the range of 750-800°C. At higher temperatures the ceramic melts. At lower temperatures, the annealing is sluggish.

Accordingly, the calculations were carried out over the range 700-850°C. In the area below the curve, the oxygen partial pressure is so low that the copper is reduced to the metal. At oxygen partial pressures higher than the curve, at a given temperature, the lowest valence pure oxide of copper oxide is stable relative to reduction to the metal.

Annealing gas blends containing argon or nitrogen as routinely delivered by gas manufacturers normally contain approximately 3 vpp of oxygen which is

enough to stabilize these compound oxides against reduction to the metal unless a trace of organic matter is included in the system, thereby producing CO and decreasing the oxygen partial pressure below the curve. While the ductile array employed in the practice of this invention may be planar, such as an open weave metal mesh having a regular, repeating pattern such as with Inconel 600 expanded metal mesh, when extreme flexibility such as is needed in this application for superconducting ribbons or wire that can be used to wind coils, a three dimensional array is employed. A preferred three dimensional array is composed of bundles of copper wire. In general, the more individual copper wires in the cable, the more flexible the final superconducting cable. The ductile array may be an extra-flexible braided copper cable, an extra flexible stranded copper cable or a 19-strand concentric copper cable.

FIG. 17 depicts the electrical resistance vs. temperature (Kelvin) incuding the >77K T _c superconducting transition of a flexible superconducting composite of this invention.

FIG. 18 is a 75X photomicrograph of two side by side cables comprising seven bundles of wire, each containing 46 individual silver plated copper wires for a total of 332 wires in the cable employed to prepare a flexible superconducting wire employed in the preparation of the superconductor of FIG. 1.

FIG. 19 is a 75X photomicrograph of an annealed superconductor wire of this invention prepared using the flexible silver-coated copper wire shown in FIG. 18. The outer sheath of ceramic had been intentionally cut away to expose the interior copper wires and the ceramic. The copper wires are intact and the space between the wires is substantially filled with the ceramic. For applications where planar structures are

required, such as in electronic components, the arrangement best shown in FIG. 4 is preferred.

As discussed above and best shown in FIG. 21, the planar composite of this invention is quite flexible, capable of flexing out of plane by as much as 0.25 inch or more with finger tip pressure on a sample of about two inches in length. A sample was repeatedly bent to 180° (folded in half) and straightened to examine the effect of such mechanical abuse. Despite this extreme mechanical abuse, a resulting crack only formed along the line of maximum stress or fold line. However, there was no crack propagating away from the fold line, and the crack that did appear did not even extend within a given, unsupported ceramic diamond area. The same ceramic composition, outside of the composite structure, would shatter. Prior art composite structures would not withstand such abuse.

The flexible superconducting wires or ribbons of this invention may be wound in coils or otherwise flexed as needed without cracking of the ceramic phase of the composite.

The following examples further illustrate the superconductor composites of this invention.

Example 8 A superconductor ceramic phase having the nominal formula Bi ₂ Sr ₂₄ Ca _>6 Cu ₂ 0 ₈ was prepared from CUC1 ₂ -H ₂ 0 (0.248 g) , Bi(N0 ₃ ) ₃ -H ₂ 0 (0.590 g) , Sr(N0 ₃ ) ₂ (0.377 g) and CaCl ₂ (0.052 g) using a solvent of adipic acid (2 g) , glutaric acid (6 g) , glycerin (10 g) , water (10 g) and EDTA (0.5 g) . A slurry of a molten melt of the ceramic phase was applied to silver-coated Inconel 600 mesh (60 mesh) steel, suspended in a furnace with 0.008 inch Inconel wire and fired to approximately 650°C. After cooling, the composite was laser annealed under flowing argon at a rate of 1 inch/sec. using a C0 ₂ laser having a 10.6 μ wavelength, 3/8 inch line, 005

inch. The ceramic melted, flowed and refroze without melting the metal suport matrix.

Example 9 A ceramic phase was prepared from Bi ₂ 0 ₃ (137.5 g) , CUC1 ₂ -H ₂ 0 (42 g) , CuS0 ₄ -5H ₂ 0 (12.1 g) , Sr(N0 ₃ ) ₂ (74.9 g) and CaCl ₂ (9.8 g) following the method of Example 8. The slurry was heated to about 250° to dry in a 3 inch Inconel crucible with a lid, then slowly heated to 700"C for 24 hours. The mixture was well stirred, heated to 950°C for about 10 minutes in air, then cooled to a temperature of about 860°C. A ductile array of Inconel 600 expanded metal foil shown in FIG. 4 in which the metal ligaments form a regular, structured, repeating diamond pattern was preoxidized by heating to approximately 700°C for about 1-1/2 hours in air. The preoxidized metal support or ductile array was dipped into the liquid ceramic phase to coat the ductile array with the ceramic phase, cooled and annealed at a temperature of 780"C. Example 10

Mixed nitrates of Pb, Ca, Sr, Cu and Bi ₂ 0 ₃ were melted in an alumina tray. Silver plated multifilament copper cable was dipped in the melt to form the composite. The composite was annealed in approximately 5 ppm 0 ₂ in N ₂ for 1 week at 770 ^β C.

Example 11 A melt was prepared according to Example 10. The silver plated multifilament copper cable was pulled through the melt in a semi-continuous operation under a reduced oxygen environment at about 900°C to give long sections of composite wire suitable for coils, motors, generators and the like.

For best results, when silver-coated Inconel expanded metal foil or woven mesh is employed as the two-dimensional ductile array, the metal should be

pretreated in slightly moist hydrogen prior to silver coating.

The superconductors of this invention exhibit superconducting transition at or about 77°K and may be employed in liquid nitrogen rather than more expensive liquid helium.

In a preferred embodiment, the ductile arrays may be selected from electrically conductive materials, preferably silver, copper or their electrically conductive alloys. The ductile phase may also be selected from high strength metals such as stainless steel, super alloys and other high-strength metals which optionally have been plated with silver. In the case of cables, multifilament braided cable formed from .002 inch wire braided to a 0.01 inch multifilament cable are particularly suitable. Size 29 or 30 copper wire having 26 to 33 wires in the cable or 105/44 25 mm thick Engineering format cable is also preferred.

The ductile superconductor materials of this invention are orders of magnitude more conductive than copper alone. They may be employed as transmission lines, to wrap coils for electromagnetic fields, in electric motors, in generators, in electronic components, etc. The structural ceramic composite of this invention comprises a ceramic supported by and intimately integrated with a continuous, ordered, interconnected, ductile, metallic array. The ductile array is surrounded by and supports the ceramic as described in further detail below.

High temperature structural ceramic composites may be prepared from crystalline or substantially crystalline aluminosilicates such as lithium aluminum silicates, magnesium aluminum silicates or other metal aluminosilicates such as sodium, potassium, magnesium, calcium, barium, yttrium, lanthanum and other suitable

metals having a valence of +3. Especially preferred are lithium aluminum silicates available from Dow Corning (LAS III) .

In one embodiment of the present invention, alumina silicate glass ceramic was compounded with

Inconel 600 square weave planar sheet to yield a ductile, flexible, flaw tolerant structural composite material.

In another embodiment of the present invention, thick and 3-dimensional structural composite materials are fabricated using a stainless steel ductile array having a pattern resulting from calendering and sintering a stack composed of 21 layers of 100 square mesh woven material wherein the even numbered layers were at a 45° bias. The resulting array is a three- dimensional interconnected array of tensile wire embedded in the ceramic phase. These materials have been machined into the complex three-dimensional helical air foil shape of turbine blades. A preferred ceramic phase for high temperature, oxidation resistant structural ceramics, is crystalline or substantially crystalline aluminosilicate glass ceramics. Lithium alumino-silicate glass ceramics are presently preferred. These materials are principally composed of lithium oxide, alumina and silica with admixtures of alkaline earths, boria, niobia, zirconia, titania and other materials as required or beneficial. An especially preferred lithiumaluminosilicate is LAS III glass ceramic which can be obtained from Dow Corning Glass Works.

A preferred ceramic phase for preparing chemically inert, high thermal conductivity, ceramic composites which are good thermal conductors (heat sinks) and electrically non-conductive are materials such as alumina (A1 ₂ 0 ₃ ) , aluminum nitride, silicon

nitride, and other high thermally conductive ceramics or glass ceramics.

Preferred ceramics for abrasion resistant room temperature structural ceramics include boron carbide or boron oxycarbide glasses, alumina, silicon carbide, titanium carbide and the like.

For high temperature applications, such as turbine blades or other engine parts, the preferred ductile reinforcement are square or twill weave oxidation tolerant superalloys such as Hanes 214 high nickel alloy with 5% aluminum, Inconels, especially Inconel 600 and 601 alloys and Hastalloy C alloy.

It is important that the bulk ceramic in the composite is no more than 1 mm from metal. To accomplish this, the ductile array comprises a continuous, repeating pattern which will give the greatest level of physical properties. An especially preferred ductile support array is a three dimensional arrangement which gives maximum strength throughout the ceramic body. This array is prepared by calendering and sintering multiple layers of square weave metal mesh, such as 100 mesh, with even layers at a 45° angle to form a 3-dimensional grid of interconnected tensile wires forming a pattern of a series of triangles. Good results were obtained using 21 layers. While the series of triangles formed by this arrangement is believed to be the strongest format, the angle of alternating layers may be varied. It is critical to sinter the layers to avoid weakness and cleavage in the composite. The structural ceramic composites of this invention preferably comprise between 40 and 90 volume percent of the ceramic phase. They were found to be ductile, relatively flexible and unusually flaw tolerant for ceramics. Referring to FIG. 21, a 0.015 inch thick planar structural composite wherein the ceramic phase is a

lithium aluminosilicate (LAS III obtained from Corning) and the ductile array is Inconel 600 woven material at 60 repeating units per lineal inch was subjected to a 180 degree bend. As can be seen in FIGS. 21 and 22, the ceramic does not spall from the metal even under these aggressive mechanical conditions.

Upon stress to failure, the cracks were observed not to propagate as in other ceramic systems. The ceramic does not spall off the ductile grid. Finally, the ceramic composite of this invention may be produced in complex, three-dimensional shapes, such as the helical airfoil form of turbine blades.

The mechanical testing of the structural composites of this invention was done by three point flexural testing. The span or unsupported region of a typical rectangular test specimen had a length "L" and a width "W". The piece is supported near its ends. Typically, these supports were hard metal cylinders. A third metal cylinder was placed at the midpoint of the span, stressing the sample. This third cylinder was forced downward at a specified rate (mM/min) until the sample failed.

The load, in pounds, required to force the displacement of the sample was recorded as a function of the displacement. This load was then converted into outer fiber stress in units of thousands of pounds per square inch ("ksi") . For the planar structural composite, the three point flexural strength was measured at approximately 70 ksi or 70,000 psi. The LAS III ceramic alone, fails at 8-10 ksi (8,000-10,000 psi).

The maximum stress point on the load/ displacement curve is referred to as "ultimate bend stress" or UBS by analogy with the "UTS" commonly derived from tensile testing. The point at which the stress load curve goes from linear to non-linear is referred to as the yield point or "YP".

54

At stresses above the YP, the composite samples have exceeded their elastic limit and are permanently deformed. In general, ceramics and ceramic composites do not exhibit separate YP and UBS points. They fail catastrophically at the YP. Certain types of ceramic composites do exhibit limited ductility to the "fiber pullout" phenomenon. Structural metals such as SS316, due to their desirable material property of ductility, tend to plastically deform above their YP. A important feature of the ceramic composite of this invention is its metal-like ductility and flexibility as best shown in FIG. 21.

As can be seen in FIG. 22, a 2OX SEM photomicrograph of the planar composite of FIG. 21, there is no spalling of the ceramic from the metal grid at the 180 degree bend line shown in the top of the photomicrograph.

FIG. 23 is a 14OX SEM photomicrograph of a planar structural composite of this invention. The center section is the LAS III ceramic, the lighter peripheral area the Inconel 600 metal wires.

FIG. 24 is a 10,000X SEM photomicrograph of the ceramic material in the planar structural composite. Note the regular fine grain size on the order of about one micron. This fine, regular granularity is a desirable feature of the glass-ceramics such as the lithium aluminosilicates.

A 1/8 inch thick structural ceramic composite was prepared from a thin dispersion of free-flowing fine powered lithium aluminosilicate glass ceramic (LAS III, Corning) mixed with 0.5% of soluble organic polymer. The thin dispersion was applied to the ductile array pre-form under a partial vacuum on a Buchner funnel to draw the powder into the inner portions of the ductile array. The liquid phase was allowed to evaporate under the airflow of the partial vacuum

leaving the powder dispersed throughout the ductile array and held in place by a the small amount of organic binder.

The ductile array was formed by calendering (pressing) and sintering into a single open metal section a twenty one (21) parallel layer stack of 100 mesh square weave stainless steel 316 woven cloth wherein every even numbered layer was set at a 45° angle of bias with respect to adjacent odd numbered layers. This resulted in a single sintered three dimensional repeating array of ductile tensile wires.

The dried, unfired, pre-composite was then suspended in a Coors mullite three inch o.d. by five foot long controlled atmosphere furnace tube by two tension cables. The tension cables were composed of 1/16 inch Haynes 214 rod in the hot zone, and common flexible multifilament steel cable at temperatures below about 600°C near the ends of the furnace tube.

The Haynes 214 rod was used in the hot zone because of its relatively good high temperature tensile properties. The original steel cable would snap at high temperatures under an argon atmosphere due to its inability to bear the tensile load above 1150°C. Tungsten has superior room temperature tensile strength and it retains a considerable strength even at very high temperatures.

One end of the tension cable was fixed to the end fitting of the furnace tube, rendering it immovable. The other end of the tension cable was run out of a small hole in the opposing furnace tube end fitting over a small 1/2" diameter metal pulley and was and was connected to approximately two pounds of static weight. This arrangement kept the sample in constant tension at all times and kept the composite from physically contacting and reacting with the mullite walls of the furnace tube.

The furnace tube was vigorously flushed with at least 10 volumes of argon gas at room temperature. With the gas flow maintained at about 200 cc/min the furnace was turned on, raising the temperature to approximately 1330°C.

As the furnace temperature rose, the tension cable expanded by several inches causing the free- hanging tension weight to drop proportionally, thereby maintaining a constant tension on the sample throughout the term of the preparation procedure.

This tensile procedure caused a pre-stressing of the composite along its main axis. Correspondingly, as the furnace cooled after approximately 45 minutes at the maximum temperature, the tension cable contracted raising the exterior metal weight to its original elevation. The samples were then removed from the furnace and inspected by OM, SEM, EDX and other characterizing means.

The above procedure was varied as to firing times, temperatures, gas composition and thermal holding periods. The ceramic phase comprised approximately 55- 60 percent by volume of these structural ceramic composites. The balance was the three-dimensional interconnected grid of the metallic array embedded in the ceramic matrix. The ceramic phase may be varied from 10 to 90%, depending upon the desired end properties.

In the composite preparation, upon reaching a temperature 1300°C, the melting temperature of the ceramic phase when the ceramic is LAS III lithium aluminosilicate, the ceramic melted and was homogeneously pulled into the interior spaces by capillary forces. This resulted in an orderly, repeating, three-dimensional structure. No holes or unfilled spaces were observed in the finished composite.

FIG. 25 depicts the result of the mechanical three point flexural testing of a structural ceramic composite containing a three-dimensional ductile array as prepared above compared with data for the unsupported LAS III ceramic, unfired metal and fired metal controls. The structural composite sample and controls were run at a machine head speed of 50 mM/minute.

The composite exhibited a yield point (YP) of 20,650 psi (20.65 ksi) at 0.75 mM deflection on a 64 mM span. The YP was taken to be the point at which the stress/deflection curve ceased to be linear. The ultimate bend strength (UBS) , the highest point on the stress/displacement curve, was measured at 56 ksi (56,000 psi) occurring at 11.3 mM (FIG. 26). These ceramic composite samples were quite ductile and were tested to a bend angle of approximately 90°.

Due to the combination of ductility and strength, the composite was able to absorb relatively large amounts of energy prior to yield or failure. At their respective yield points, the ceramic and metal controls absorbed about 5,000 J/M ² (joules/meter ² ). The ceramic composite absorbed about 55,000 J/M ² , more than a ten-fold improvement over the ceramic and metal controls as best seen in FIG. 27.

The high temperature structure ceramics of this invention can be fabricated into complex three- dimensional shapes such as turbine blades and engine parts. The chemically inert high heat conductivity ceramic composites are ideally suited for electrical applications such as computer applications, high density electronic circuit boards, particularly microcircuits where high volumes of heat must be dissipated. The room temperature structural ceramics of this invention are particularly suited for wear resistance is critical in

abrasive or corrosive environments. Thus, this material breakthrough has wide applicability.