EP0172170A1 - Refractory hard metal containing plates for aluminum cell cathodes - Google Patents

Abstract

Plates or tiles containing a hard refractory material for use in aluminum reduction cells. These plates constitute a hard refractory material, for example TiB2, bound by a matrix of carbonized thermosetting binder, and form a surface wettable by aluminum when they are bonded to a cathode substrate.

Description

REFRACTORY HARD METAL CONTAINING PLATES FOR ALUMINUM CELL CATHODES

Background of the Invention This invention relates to cathodes for electrolytic cells for the production of aluminum, and specifically to the preparation of cathode plates or tiles for use in such cells. The cathode plates have a surface which is aluminum wettable and contains Refractory Hard Material in a carbonaceous matrix bonded by amorphous carbon, said matrix characterized by an ablation rate essentially equal to the combined rate of wear and dissolution of said Refractory Hard Material in the aluminum cell environment. Cathode plates meeting such a specification provide far greater durability and ease of application than any previously proposed.

Aluminum is conventionally manufactured by an electrolytic reduction process conducted in Hall-Heroult cells, wherein alumina is dissolved in molten cryolite and electrolyzed at temperatures of 900-1000"C. These cells typically comprise a steel shell with an insulating lining of suitable refractory materials, which in turn is provided with a lining of carbon which contacts the molten bath, a uminum, and/or ledge. One or more anodes, usually made of carbon, are inserted into the molten cryolite and connected to the positive pole of a direct current source. The negative pole of the direct current source is connected to the carbon lining in the bottom of the cell. Molten aluminum resulting from the electrolytic reduction reaction is collected on the carbon bottom of the cell in a molten pool or pad, which acts as a liquid metal cathode onto which additional aluminum deposits. A portion of this pool of liquid is removed periodically and collected as the product of the electrolysis process.

In the construction of most modern commercial cells, the carbon lining that forms the top layer of the cathode is conventionally built from an array of prebaked carbon blocks covering the portion of the cell to be lined, and then the carbon blocks are joined into a solid continuous assembly by ramming the slots between blocks with a mixture typically of coke, calcined anthracite, modified coal tar pitch, and the like. This structure is then heated in the process

OMPI of cell start-up. The life span of such carbon linings in different plants averages three to eight years, but under adverse conditions may be considerably shorter. Deterioration occurs due to penetration of molten electrolyte components and liquid aluminum into the structure of the carbon blocks, and ramming mix, causing swelling and cracking. Aluminum metal penetration causes alloying and slow destruction of the steel current collector bars embedded in the cell bottom. This contaminates the aluminum pad and may eventually lead to cell tap-out. Other problems in conventional aluminum reduction cell operation Include accumulation of undissolved or frozen bath and alumina which are carried from the cryolite bath, ledge, and ore cover, to the cathode, creating sludge or muck. The presence of this sludge or muck under the aluminum pad creates electrically insulated areas on the cell bottom which increase the cathode voltage loss and disrupt electrical current distribution, resulting in excessive pad turbulence and disturbances through magnetic forces, hence reducing cell current efficiency.

A further drawback of the carbon cathode lining is its non-wettability by molten aluminum, which necessitates operation with a deep pad of aluminum, to ensure effective molten aluminum contact to the carbon lining or surface. The deep aluminum pad is subject to magnetic and electrical interactions, such as standing waves, which increase the possibility of electrical shorting to the anode. To lessen this possibility, greater anode-to-cathode distances (ACD) are employed, resulting in additional voltage requirements.

To reduce ACD and associated voltage drop, it is necessary to make adjustments in magnetic design, or to operate without an aluminum pad. To achieve the latter goal, attempts have been made to use cathode materials comprising Refractory Hard Material (RHM), such as TiB₂. Titanium diboride is highly conductive and is wetted by liquid aluminum. This wettability property enables a thin film of molten aluminum to be deposited directly on the cathode structure made of RHM, and eliminates the need for a pad of metal, since contact with the underlying cathode structure is assured.

&?NATlO The use of titanium diboride current-conducting elements in electrolytic cells for the production of aluminum is described in the following exemplary U.S. Patents: 2,915,442, 3,028,324, 3,215,615, 3,314,876, 3,330,756, 3,156,639, 3,274,093, and 3,400,061. Despite the extensive effort expended in the past, as Indicated by these and other patents, and the potential advantages of the use of titanium diboride as a current-conducting element, such compositions do not appear to have been commercially adopted on any significant scale by the aluminum industry. Lack of acceptance of T1B« or RHM current-conducting elements of the prior art is related to their lack of durability in service 1n electrolytic reduction cells. It has been reported that such current-conducting elements fail after relatively short periods in operation. Such failure has been associated with the penetration of the self-bonded - RHM structure by the electrolyte, and/or aluminum, thereby causing critical weakening with consequent loss of cohesion, cracking and failure. It is well known that liquid phases penetrating the grain boundaries of solids can have undesirable effects. For example, prior art RHM tiles, in which oxygen impurities were found to segregate along grain boundaries, are susceptible to rapid attack by aluminum metal and/or cryolite bath. Prior art techniques to combat TiB« tile disintegration in aluminum cells have included use of highly refined TiBg powder to make tiles containing less than 50 pp oxygen at 3 or 4 times the cost of commercially pure TiB₂ powder (containing about 3000 ppm oxygen). Moreover, the necessary high temperature fabrication further increases the cost of TiB« tiles substantially. Despite the use of high purity materials, no cell utilizing TiB« tiles is known to have operated successfully for extended periods without loss of adhesion of the tiles to the cathode, or disintegration of the tiles. Other reasons proposed for failure of RHM tiles and coatings have been the solubility of the materials in molten aluminum or molten bath, the lack of mechanical strength, and the poor resistance to thermal shock.

Additionally, different types of TiB materials, applied to carbon substrates, have failed due to differential thermal expansion between the titanium diboride material and the carbon cathode

OMPI

^ ?« block. To the Inventor's knowledge no prior RHM-containing tiles or plates, even of high purity, have been successfully operated as a commercially empl oyed cathode structure or surface l ayer, because of thermal expansion mismatch, bonding problems, etc. For example, U. S. Patent 3,400,061 , of Lewis et al , assigned to

Kaiser Aluminum, teaches a cell construction with a drained and wetted cathode, wherein the Refractory Hard Material cathode surface consi sts of a mixture of Refractory Hard Material , at least 5 percent carbon, of which 10 to 20 percent by weight is derived from pitch, baked at 900 ^βC or more. According to the patent, such a composite cathode has a higher degree of dimensional stability when electrolysed i n a molten bath environment than previously attainable with carbon. The composite cathode coating material of this reference may be rammed into place i n the cell bottom. Alternatively, shapes composed of the composite material may be produced in a separate facil ity for placement on a cathode block. Such material has not been widely adopted, however, due to susceptibility to attack by the electrolytic bath, as taught by a later Kaiser Al uminum U. S. Patent, No. 4,093,524 of Payne. Sai d U.S. Patent 4,093,524, of Payne, cl aims an improved method of bonding titanium diboride, and other Refractory Hard Materials, absent carbonaceous bi nders, to a conductive substrate such as graphite, or to silicon carbi de. The cathode surface is made from titanium diboride tiles, 0.3 to 2.5 cm thick. Payne recognized that the l arge differences in thermal expansion coefficients between such Refractory Hard Material tiles and carbon precludes the formation of a bond which will be effective both at room temperature and at operati ng temperatures of the cel l. The bonding i s accordingly formed in-situ at the interface between the Refractory Hard Material tile and the cathode by a reaction between al uminum and the carbon beneath the file to form aluminum carbide only when the cel l approaches operati ng temperature. However, since the bond is not formed until high temperatures are reached, tiles are easily di splaced duri ng startup procedures. The bondi ng is accelerated by passing electrical current across the interface, resulting In a very thin aluminum carbide bond. However, electrolyte attack upon the

OMPI bond results if the tiles are installed too far apart, or if the protective film of aluminum on the surface is incomplete alternatively, 1f the tiles are installed too close together, they bulge at operating temperature, resulting In rapid deterioration of . the cell lining and in disturbance of cell operations. Further problems would probably be witnessed during fluctuations in cell temperature and during a shut-down and restart of a cell employing such bonding, because the thermal expansion mismatch has not been eliminated, merely circumvented at high temperature. Accordingly, this concept has not been extensively utilized.

HolHday, In U.S. Patent 3,661,736, claims a dimensionally stable composite cathode for a drained and wetted cell, comprising particles or chunks of highly purified arc-melted "RHM alloy" embedded in an electrically conductive binder matrix which may be carbonaceous. In this instance, the surface of the matrix becomes protected by an aluminum carbide layer. However, 1n operation of such a cell, electrolyte and/or aluminum attack the matrix material, large areas of which are exposed to contact, consequently leading to early destruction of the cathodic surface. Moreover, the relatively large chunks of TiB₂ suffer from the same drawbacks, 1 terms of poor thermal shock resistance, brittleness, etc., as wholly RHM materials.

U.S. Patent 4,308,114, of Das et al, discloses a contoured cathode surface composed of Refractory Hard Material in a fully graphitic matrix. In this case, the Refractory Hard Material is composited with a pitch binder, and subjected to graphitization at about 2350^#C, or above. Such cathodes are subject to early failure due to relatively rapid ablation, caused by physical erosion and aluminum carbide formation in the graphite matrix. In addition to the above patents, a number of other references relate to the use of titanium diboride in tile form. Titanium diboride tiles of high purity and density have been tested, but they generally exhibit poor thermal shock and impact resistance and are difficult to bond to carbon substrates employed in conventional cells. Mechanisms of de-bonding are believed to involve high stresses generated by the thermal expansion mismatch between the titanium diboride and carbon, as well as aluminum penetration along the interface between the tiles and the adhesive holding the tiles in place, due to wetting of the bottom surface of the tile by molten aluminum. In addition to debonding, disintegration of even high purity tiles may occur due to aluminum penetration of grain boundaries. These problems, coupled with the high cost of the titanium diboride tiles, have discouraged extensive commercial use of titanium diboride elements in conventional electrolytic aluminum smelting cell s, and limited their use in new cell design. To overcome the deficiencies of past attempts to utilize Refractory Hard Material s as a surface element for carbon cathode blocks, coating materials comprising Refractory Hard Materials in a carbonaceous matrix have been suggested.

In U. S. Patent Appl ications 395,343, 395,344, and 395,345, filed July 9, 1982 by Buchta et al , and 400,762, 400,772, and 400,773, filed July 29, 1982 by Boxall et al , formulations, application methods, and cells employing TiBg/carbon cathode coating material s were di sclosed. Thi s technology relates to spreading a mixture of Refractory Hard Material and carbon solids with thermosetting carbonaceous resin on the surface of a cathode block, followed by cure and bake cycles. Improved cell operations and energy savi ngs resul t from the use of thi s cathode coating process in conventionally designed commercial aluminum reduction cell s. Plant test data indicate that the energy savings attained and the coating l ife are sufficient to make this technology a commercially advantageous process.

Advantages of such composite coating formulations over hot pressed RHM tiles include much lower cost, less sensitivity to thermal shock, thermal expansion compatibility with the cathode block substrate, and less brittleness. In addition, oxide impurities are not a problem, and a good bond to the carbon cathode block may be formed which is uneffected by temperature fluctuations and cell shutdown and restart. Pilot plant and operating cell data indicate that a coating l ife of from four to six years or more may be anticipated, depending upon coating thickness. However, problems inherent in this coating process include the fact that modifications in the coating formulation are required to compensate for changes in the mechanical properties and thermal expansion coefficients of different cathode blocks. The process is labor intensive and requires complex cure and bake heat treatments which can be very disruptive to plant operations. The need to maintain a good bond to the cathode block during the cure and bake heat treatments necessitates the use of a less than optimum formulation and process conditions. Quality control is also difficult to maintain in the plant environment. Further, the variable electrical resistivity of the unbaked coating can result in severe problems during cell start-up, and finally, coating thickness 1s limited to approximately 1.2 cm.

Attempts have been made to retain the advantages of the novel composite coating material, as formulated and tested, while minimizing problems detailed above. For example, in U.S. Patent Application 461,893, filed January 28, 1983 by Buchta and Nagle, fabrication of hot pressed tiles is disclosed. However, the tiles prepared in accordance therewith, (utilizing thermosetting resin, Refractory Hard Material, and graphite, and formed under high pressure at elevated temperatures,) are mechanically soft, particularly after exposure to the aluminum cell environment, and are susceptible to aluminum carbide formation and consequent wear. In contrast, the present invention retains the advantages of the novel composite coating material, without the drawbacks of the hot pressed tiles. By preparing structural shapes which are fabricated and heat treated prior to application to the cathode, rather than applying the material directly to the cathode substrate and then heating, numerous benefits are realized. Specifically, the added improvements include the fact that once baked, the plate material has a thermal expansion coefficient which is essentially equal to carbon; hence, only a simple carbon to carbon bond 1s required to attach the plate to the carbon cathode block. The production process can be readily automated at a central manufacturing plant. The baked plates can be glued to the cathode blocks either by the block manufacturer (or other vendor), or in the plant. Moreover, a central plate manufacturing plant affords the best quality control, since only a simple gluing operation is left for the less controlled smelter environment. Since there is no substrate-to-composite bond to maintain during the initial cure and bake heat treatments accompanying the material fabrication, the process can be optimized to produce the highest quality cathode plate material at a minimum cost. Once baked, the plate material is highly electrically conductive and therefore does not interfere with cell start up procedures. Multiple plates can be glued together to give any desired thickness. More efficient heat sources such as microwave can be used to cure the plate material, and complex shapes can be produced by this process to meet the needs of all foreseeable low energy cell designs.

Summary of the Invention The present invention deals with preformed RHM/carbon cathode plates or tiles, based on casting, curing, and baking formulations suitable for use in the cathode coating process described previously. The improved plates, tiles, or elements avoid many of the application problems associated with the prior coating process, and enable one to achieve previously unattainable improvements in quality and facility of construction of the resulting cathode structure.

Coating, curing and baking of the previously disclosed cathode coating material on a full size cathode block or the bottom of an assembled cathode presents many heating and control problems. While these problems have been overcome, to a large extent, in recent pilot tests, the solutions are less than ideal. One approach to diminishing such problems is to combine block manufacture and coating operations in one facility. While coating the blocks at a block manufacturing facility avoids most of the in-plant disruptions, it would be necessary to heat the block to greater than 700"C for a second time after the original block baking process, or to compromise between the optimum heat cycles for manufacturing the block and heat-treating the coating in the cycle.

OMPI In the present invention, the cathode coating material is cast or pressed in a mold, then cured and baked to form a rigid, electrically conductive plate or tile. The cathode plates may then be glued to the cathode blocks using conventional carbon gluing procedures. By thi s method, both the composition and the fabrication of the cathode plate material can be optimized without regard to the difference in temperature controlled behavior between the coating and cathode block during the cure and bake heat treatments. Once baked (carbonized) the thermal expansion coefficient of the cathode plate material is so similar to that of the cathode block that joini ng the two presents no problems, and attachment over extended temperature ranges may be maintained.

Thus, the present Invention relates to pre- anufactured shapes, such as plates, tiles, or elements, which contain Refractory Hard Material (which is aluminum wettabl e) for I nclusion as a portion of the cathode surface of an electrolytic cell for al uminum production. The method of manufacture of these elements is al so disclosed herein, whereby the Refractory Hard Material may be di spersed uniformly or in accordance with a predetermined compositional gradient within the matrix of the element.

Brief Description of the Drawing Figure 1 represents a typical time -temperature relationship suitable for a cure cycle utilized in the preparation of titanium diboride/carbon plates in accordance with this invention.

Detailed Description of the Invention

According to the present invention, it has been found that plates, shapes, or tiles may be fabricated using Refractory Hard Material (RHM) combined with specified thermosetting bonding agents and other material s to form structures that improve the operation of conventional al uminum reduction cell s. Such improvements include wettabil ity by molten aluminum, low solubil ity i n the molten aluminum and cryolite environment, good electrical conductivity, and decreased muck adhesion.

OMPI In understanding the concept of the present invention, it is important that certain distinctions and definitions be observed. Accordingly, the following definitions shall be applied with respect to this invention. The terms "molding" or "casting" are Intended to encompass shaping a mass of material generally within a confined space or cavity. Extrusion and injection molding are within the scope of these terms, while hot pressing is excluded. Whereas compositions utilized for molding or casting incorporate at least a sufficient quantity of mix liquid to assure complete and intimate wetting of the surfaces of the particulate solids, and to assist 1n blending, hot pressing compositions do not.

The "molding composition" of the present invention is comprised of Refractory Hard Material, carbonaceous additive, carbonaceous filler, and binder system. As used herein, the terms "molding composition" or "molding material" shall be used synonomously to encompass the combination of these materials.

The "Refractory Hard Materials" are herein defined as the borides, carbides, suicides, and nitrides of the transition metals in the fourth to sixth group of the periodic system, often referred to as Refractory Hard Metals, and alloys thereof.

"Resinous binder" shall be used to designate a polymerizable and/or cross-linkable thermosetting carbonaceous substance. The "mix liquid" of the present invention functions in a variety of manners in the molding composition of the present invention, depending upon specific composition. It may be present to allow easy and uniform mixing of the solid components of the composition and to provide an easily moldable mass. Certain mix liquids, such as furfural, may also permit an increase in the amount of carbonaceous filler which may be incorporated in the composition. The mix liqui also permits wieking of the resin into interstitial voids between particles of the molding composition by capillary action. The mix liquid may act solely as a solvent for the resinous binder (which may already be present in the solids portion of the binder system), such as methyl ethyl ketone (which could dissolve a novolac resin if present in the solids), and be - n -

evaporated during cure and carbonization operations. If, on the other hand, the mix liquid is present simply as an Inert carrier liquid, then it too may be evaporated during cure and carbonization. Otherwise, the mix liquid may function as a combined . solvent and resin in its own right, such as a mixture of furfuryl alcohol and furfural, part of which volatilizes during heating while the remainder becomes incorporated into the resinous binder. In another instance, the mix liquid may be the resinous binder per se, such as where the resinous binder is a liquid such as furfural (generally In combination with phenol or furfuryl/alcohol), furfuryl alcohol, or low polymers of these, or a resole. The mix liquid may also comprise the resinous binder in the case of a solid resin, such as a novolac, dissolved In a solvent (the solvent portion of which may volatilize during heat up), or a high viscosity resin such as a partially polymerized resole thinned by a solvent. The mix liquid may also contain gas release agents, modifying agents, and curing agents.

"Binder system" shall be used to indicate the resinous binder and the mix liquid, and, if required, gas release agents, modifying agents, and curing agents.

"Gas release agent" shall be taken to mean agents present which form liquid phases which seep through the molding composition and then evaporate, to create small channels to permit release of volatiles. "Modifying agents" shall be taken to mean materials added to the resinous binder to modify, for example, curing, electrical properties, or physical properties such as flexural strength or impact strength prior to carbonization.

"Curing agents" shall be taken to mean agents required to either copolymerize with the resin or to activate the resin to a state in which the resin may polymerize or copolymerize. Cross-linking or activating agents fall into this category, as do catalysts required for most polymerization and cross-linking reactions. "Carbonaceous filler" shall be interpreted to mean those carbonaceous materials present, either as a component of a known carbon cement or as part of a proprietary or custom carbon system, having a C:H molar ratio greater than 2:1, which are smaller than 100 mesh in size. While a carbonaceous filler may have reactive groups present, and need not be fully carbonized, such materials do not typically polymerize with themselves as the resinous binder material does. Further, carbonaceous filler is essentially insoluble in commonly used solvents such as methyl ethyl ketone or quinoline, while the resinous binder (in its incompletely cured state) is usually soluble therein. "Carbonaceous additives" shall Indicate those carbonaceous materials present, either as a component of a known carbon cement or as part of a proprietary or custom carbon system, having a C:H molar ratio greater than 2:1, which comprise particulate carbon aggregate having a particle size range between -4 mesh and +100 mesh, and/or carbon fibers.

The term "carbon system" shall encompass the binder system plus carbonaceous additive and carbonaceous filler.

"Carbon matrix" shall refer to the carbonized product of the carbon system, and is thus bonded by amorphous carbon formed by the carbonization of the binder system.

"Carbon cement" shall be taken to mean a commercially available carbonaceous cement or adhesive, generally comprising a resinous binder, mix liquid, carbonaceous filler, and curing agents, the solid and liquid portions of which may be packaged separately to increase shelf life, or combined as a premixed cement. Gas release agents, and/or modifying agents may be present in such systems, or may be added thereto for use in the present invention. Carbonaceous additives are generally added to such systems for use in the present invention, as they are generally not present in commercially available formulation.

Pitch may be present as part of the resinous binder, as a modifying material, but requires the presence of a suitable curing agent, such as hexamethylenetetramine. Such a curing agent may be already present as a component of the resinous binder, or may be added thereto to facilitate cross-linkage between the resinous binder and the pitch, or linkage between the pitch and carbonaceous filler, or self-linkage between the polynuclear aromatlcs which comprise the bulk of pitch. Although pitch is known to constitute a graphite precursor, graphitization is not desired 1n the present invention. Thus, the graphite precursor is dispersed within the resinous binder, which is an amorphous carbon precursor. This effectively cross links the binder phase, which prevents layer plane alignment in the graphite precursor, preventing graphitizaton and at the same time producing three dimensional strength. Pitch may seep through the molding composition to provide gas release channels, and may, in the presence of appropriate curing agents, cross link with the resinous binder and/or the carbonaceous filler.

It is desirable that the amount of shrinkage that the cured binder system undergoes during carbonization be as small as possible. This may be accomplished by selection of a carbonaceous resin which when utilized in accordance with the present invention and subjected to carbonization exhibits a shrinkage less than that which would Induce formation of large cracks or voids. Fine vertical cracking within the carbonized plate 1s preferred to horizontal cracking as a stress relief mechanism. However, a plate which has essentially no cracks is preferred. The presence of carbonaceous additive and/or filler is beneficial in this regard.

It has been found critical to utilize a binder system which, when subjected to carbonization, has a char yield of greater than about 25 percent. "Char yield" 1s defined herein as the mass of stable carbonaceous residue formed by the thermal decomposition of unit mass of the binder system, within a formulated carbon system, in an inert atmosphere. Thermogravimetric analyses of various binder systems have demonstrated that the amount of char yield is a function of the aromaticity of the resin structure. In general, carbon rings that are bonded at two or more sites will usually remain as char. Ladder polymers are the most stable, losing only hydrogen, and giving a very high carbon char yield.

"Char yield" of a binder system, as utilized herein, is determined by curing a proposed carbon system (i.e. binder system plus carbonaceous additive and filler) for a 24 hour period so as to achieve polymerization and/or cross-linkage, followed by heating at 250^βC for sufficient time to achieve constant weight, 1n order to ensure full cure and to eliminate volatiles, polymerization products, and/or unreacted liquid. The sample is then baked to 1000^dC in a non-oxidizing atmosphere, and the remaining char weight ■ determined. Similarly, the char weight of carbonaceous additive and filler present in the carbon system Is determined, and subtracted from the char weight of the carbon system to determine the char weight of the binder system alone. From the weight of the carbon system at 250"C, and the known weight of carbonaceous additive and filler at 250"C, one may calculate the weight of the binder system at 250"C. The char yield of the binder system is then calculated, as a percentage, from the char weight of the binder system after baking to 1000"C and the weight of the binder system at 250"C. It has been observed that binder systems exhibiting a char yield of greater than about 25 percent give acceptable cathode materials upon cure and carbonization, while a binder system exhibiting 8 percent char yield gave an unacceptable carbon matrix upon carbonization. Char yields in excess of about 50 percent are preferred.

To achieve a long-lasting plate or tile in the environment of an aluminum cell, it is desired that the effective rate of ablation of the cured and carbonized carbon system within the composite be close to the combined rate of wear and dissolution of the Refractory Hard Material in such environment. As the Refractory Hard Material is removed from the plate or tile by exposure to the aluminum cell environment, the carbon matrix thereof is removed at a similar or very slightly faster rate, thus exposing additional Refractory Hard Material to the cell environment. In this manner, the cathode surface remains essentially constant, in terms of both carbon and Refractory Hard Material content, thus improving cell operation as measured by uniformity of performance. In previous attempts to provide cathodes containing Refractory Hard Material, ablation and/or intergranul r attack have resulted in rapid surface deterioration due to depletion of either the Refractory Hard Material or the carbon binding at a rate greater than the other, resulting in periods when there are localized areas having either a Refractory Hard Material-rich surface composition with insufficient binding capability, or a carbon-rich surface with insufficient Refractory Hard Material. The present Invention overcomes these failures by providing an element in which Refractory Hard Material and carbon matrix are dissolved or otherwise depleted at approximately equal rates.

It is important to clarify or distinguish between carbonizing and graphitizing as they apply to heating carbonaceous bodies in the context of the present invention. "Carbonizing" is normally done by heating a carbonaceous body, either in unitary or particulate form, for the purpose of driving off volatiles, and progressively increasing the ratio of carbon to hydrogen, by progressively eliminating hydrogen from the body. In the carbonizing process, the temperature is gradually increased to allow for the slow evolution of volatiles such as decomposition products so as to avoid blister formation, and to permit volumetric shrinkage (which will occur at some point in the operation) to proceed gradually, so as to avoid formation of large cracks. While curing is considered to take place at temperatures up to about 250°C, carbonization temperatures normally range from about 250^βC to about 1000°C, although higher temperatures up to about 1450°C can also be employed. While carbonization may be continued to about 1000°C, or higher, the carbonization of the carbonaceous materials present in the instant invention is essentially complete at about 800°C, and the resinous binder has been carbonized to bind the carbonaceous filler and carbonaceous additive materials and RHM into a durable structure.

Graphitizing, on the other hand, requires considerably higher temperature and longer time periods, and produces drastic and easily detectable changes in atomic and layer plane arrangement. In graphitizing, the temperatures employed range from slightly over 2000^βC up to 3000^βC, with typical temperatures ranging from about 2400"C to about 3000^βC. These temperatures are usually associated with the higher quality grades of graphite. This heating is typically for a period of about two weeks, in a non-oxidizing atmosphere. Such heating is normally done by passing an electrical current directly through the carbon so as to heat it directly by its own electrical resistance, as opposed to the indirect furnace

OMPI heating means conventionally empl oyed for carbonization. In general , graphitizing i s only practicable with wel l known graphite precursor materials of hi gh aromaticity and negl igible cross-l inking such as pitch. C.f. , R.E. Franklin, Proceedings of the Royal Society of London, Vol . A 209, p. 196 (1951 ) .

One acceptable practice in producing carbonaceous plates according to the present invention i s to employ particulate graphite as a fill er material which is added to the binder and other components. The mixture i s then formed, cured, and carbonized. Whil e this carbonized carbonaceous material may thus contain some particulate graphite, it i s not bonded by the graphite, but rather contains both graphite particles from the filler and/or additive, and amorphous carbon derived from the binder and/or components of the carbonaceous filler and additive. In practicing the present ^" invention i t i s important that the carbonized cathode plate be constituted of a non-graphitizing binder so as to assure the proper combination of electrical and thermal conductivity, ablation rate, and stabil ity properties in the carbon-Refractory Hard Metal surface. While the borides, carbides, silicides and nitrides of elements in Groups IV to VI of the Periodic Table generally al l possess high mel ting points and hardness, good electrical and thermal conductivity, are wetted by molten aluminum, and are resi stant to al uminum and alumina-cryolite melts, TiB- i s the preferred RHM due to its relatively low cost, ready availabil ity, and hi gh resistance to oxy-fl uoride melts and molten aluminum. Suitably, Refractory Hard Material particle sizes may range from submicron to about 10 mesh, preferably about -100 mesh, and most preferably about -325 mesh. The T1B₂ preferred for use in this invention is typically specified as -325 mesh. If the T1B« is made by carbothermic reduction of titanium and boron oxides and carbides, individual particles will normally fit the requi site category of single crystal s. This also holds true for TiBg made by plasma methods described i n U.S. Patent 4,282,195 to Hoekje of PPG Industries. The TiB₂ particles should preferably be single crystals, cracked single crystal s, or have minimal grain boundaries such that all T1B₂ crystals are In contact with the binder.

Other RHM materials may be successfully substituted for T1B₂, when appropriate changes in the composition are made to account for differences in wettability, surface area, particle size, porosity, and solubility of the RHM. Sufficient RHM Is incorporated in the molding composition to ensure aluminum wetting, while thermal expansion mismatch effects are minimized and a dissolution rate of Refractory Hard Material less than the rate of loss of the carbon matrix of the coati ng 1s achieved. While discussion of the invention will focus on the use of T1B₂ as the preferred RHM, it is contemplated that any suitable RHM, such as ZrB , or mixtures or alloys of Refractory Hard Material s, may be utilized. In general , the RHM may compri se from about 20 to about 90 percent by weight of the composition, and preferably from about 25 to about 80 percent. It has been found that aluminum wettability may be achieved at concentrations as low as about 10 percent, but better results are achieved from 20 percent upward, with from about 35 to about 70 percent being the most preferred range. The resinous binders of the present invention may comprise any which meet the aforementioned criteria. Typical resins which can be empl oyed include phenolic, furane, polyphenylene, heterocyclic resins, epoxy, sil icone, al kyd, polyi ide resins, and mixtures or copolymers thereof. Exampl es of phenol ic resins which can be employed i nclude phenol formal dehyde, phenol acetaldehyde, phenol -furfural , -cresol formaldehyde and resorci no! formaldehyde resins. Epoxy resins which can be utilized i nclude the d1 glycidyl ether of bisphenol-A, di glycidyl ether of tetrachlorobisphenol-A, dlglycidyl ether of resorcinol , and the like, and especially the epoxy novolacs. Preferred epoxies comprise the glycidyl ethers such as the glycidyl ethers of the phenol s, and particularly those prepared by reacting a dihydrlc phenol with eplchlorhydrin, e. g. , the dlglycidyl ether of bisphenol -A, and epoxy novolacs. The silicone polymers which can be employed incl ude methyl siloxane polymers and mixed methyl phenyl siloxane polymers, e.g. , polymers of dimethyl siloxane, polymers of phenyl methyl siloxane, copolymers of phenylmethylsiloxane and dimethylsiloxane, and copolymers of diphenylsiloxane and dimethylsiloxane. Examples of heterocyclic resins are polybenzimidazoles, polyquinoxalines and pyrrones. Any of the well known specific alkyds, particularly those modified with phenol formaldehyde, and polyimide resins can be employed. The phenolics and furanes are the preferred classes of resins, particularly in view of relatively low costs. Furane resins are ve y advantageously employed as the resinous binder.

In addition to those set forth as components of the commercially available carbon cements, such as UCAR C-34, (a trademark of Union Carbide) discussed hereinafter, a wide variety of novolac resins may be used as the basic resinous binder in the present invention. The term novolac refers to a condensation product of a phenolic compound with an aldehyde, the condensation being carried out in the presence of an acid catalyst and generally with a molar excess of phenolic compound to form a novolac resin wherein there are virtually no methylol groups such as are present in resoles, and wherein the molecules of the phenolic compounds are linked together by a methylene group. The phenolic compound may be phenol, or phenol wherein one or more hydrogens are replaced by any of various substituents attached to the benzene ring, a few examples of which are the cresoles, phenyl phenols, 3,5-dialkylphenols, chlorophenols, resorcinol, hydroquinone, xylenols, and the like. The phenolic compound may instead be naphthyl or hydroxyphenanthrene or another hydroxyl derivative of a compound having a condensed ring system. It should be noted that the novolac resins are not heat curable per se. Novolac resins are cured in the presence of curing agents such as formaldehyde with a base catalyst, hexamethylenetetramine, paraformaldehyde with a base catalyst, ethy!enedlamine-formaldehyde, and the like.

For purposes of the present invention, any fusible novolac which is capable of further polymerization with a suitable aldehyde may be employed. Stated another way, the novolac molecules should have two or more available sites for further polymerization and/or cross-linkage. Apart from this limitation, any novolac might be employed, including modified novolacs, i.e., those in which a non-phenolic compound is also included in the molecule, such as the diphenyl oxide or bisphenol-A modified phenol formaldehyde novolac. Mixtures of novolacs may be employed or novolacs containing more than one species of phenolic compounds may be employed. Furfuryl alcohol may be advantageously employed as the mix liquid with a phenolic carbonaceous binder, and 1s believed to react with the phenolic resin as it cures, serving as a modifying agent for the resin. The use of furfuryl alcohol is preferred as it has been found that bonds having the high strength obtainable through the use of this mix.liquid cannot be produced when other mix liquids are substituted for furfuryl alcohol. Thus, for example, when furfuraldehyde is employed in place of furfuryl alcohol in otherwise identical compositions, bonds are produced having only about half the strength of the bonds produced using the furfuryl alcohol. Since the net final effect desired is to achieve a plate composed essentially of RHM and carbon, the binder system should be readily decomposable, in high yield, to a carbon residue. The resinous binder should comprise from about 1 to about 40 percent of the composition, whether as a part of a carbon cement or as a custom carbon system. Although higher resin concentrations are possible, little advantage is attained, and extended cure and carbonization cycles may be required. The carbon system should comprise about 10 to about 80 percent of the molding composition, preferably from about 20 to about 75 percent, and most preferably from about 30 percent to about 65 percent of the molding composition.

One may utilize appropriate blends of carbon and phenolic resin or other thermosetting resinous binders, or alternative commercial compositions. The mix liquid component of the molding composition may vary from approximately 0 weight percent to about 40 weight percent for reasonable evaporation and curing rates, with about 10 percent to about 20 percent being preferred to obtain workable consistency. Insufficient liquid will make the mix dry and difficult to mold, while excessive liquid results in difficulties in curing and baking. Various modifying agents may be present to modify the nature of the resinous binder during mixing, curing, and carbonization of the

tfjREA^",

OMPI ^NATl molding composition. These may typically constitute from zero to about 10 percent by weight of the molding composition. Suitable modifying agents for phenol formaldehyde resins, for example, include pitch, rosin, aniline, copolymers, resin "alloys", etc. Some particulate carbon, either amorphous or graphitic, is frequently present in the commercially available cements mentioned heretofore. Further particulate carbon may be added, as either fine powder or coarse aggregate, or mixtures thereof, in the form of amorphous carbon or graphitic carbon. It 1s particularly highly desirable to have a carbonaceous filler material present, either as a component of a proprietary carbon system or present in a commercial cement, or as an addition to a commercial cement. Such carbonaceous filler is -100 mesh, and preferably -325 mesh, and may comprise fine carbon flour, graphite flour, crushed coke, crushed graphite, carbon black, and the like. The presence of such fine flours yields improved packing density for the granulometry used, that wicks up resins and other liquid phases to develop a dense, highly bonded carbon matrix upon carbonization. Carbonaceous filler, as fine flour, should comprise from about 1 percent to about 60 percent of the molding composition, with about 10 percent to about 40 percent being preferred. The carbonaceous additive, or aggregate material, if present, may run from -4 mesh to +100 mesh, and is preferably between -8 mesh and +20 mesh. Such coarse aggregate apparently permits micro-cracking, assists volatile emission release, reduces shrinkage, and contributes to high carbon yield. Carbonaceous additive, as aggregate and/or fiber, should comprise from about 0 percent to about 30 percent of the molding composition, with from about 5 percent to about 15 percent being preferred. As previously set forth, it is preferred that carbon fiber be added to the molding composition for the purpose of arresting cracks during the primary heat treatment and processing. When such fiber is used, some variations in composition ranges have been found. When carbon fibers are used, they may preferably be made from pitch precursors, organic fiber precursers such as polyacrylonitrile, or rayon. Pitch fibers are considerably cheaper, and accordingly preferred. Fiber weight may range from zero percent to about 30 percent by weight of the composition, preferably from about 0.05 to about 1.0 percent, and more preferably from 0.10 to about 0.5 percent. However, concentrations greater than about 10 percent become comparatively expensive, with little apparent added benefit. Carbon fibers with lengths varying from about 0.16 cm to 1.27 cm length are preferred. Short fibers permit easier mixing, and may be used in higher concentration. Sized fibers, consisting of parallel fiber strands bonded together by a material soluble 1n the mix liquid, are particularly preferred, since they blend most easily with the binder system. Fiber orientation may vary, and the fibers can be mixed as an integral part of the composition.

Gas release agents are appropriately included in the molding composition to avoid blisters and/or excessively large cracks. Suitable gas release agents include high boiling point liquids such as combustible oils, soaps, and waxes.

A preferred binder system is that which is commercially designated as UCAR C-34, produced by Union Carbide. This composition is believed to comprise a mixture of an oil, a soap, finely-divided carbonaceous particles, furfuryl alcohol, a phenolic resin of the novolac type, and a hardening agent for the phenolic resin. Small amounts of pitch may be present in some lots of UCAR C-34 material. The mixture of the oil, finely-divided carbonaceous particles, phenolic resin, and phenolic resin hardener can be prepared by blending the carbonaceous particles, phenolic resin and phenolic resin hardener together in any conventional manner, e.g. in a tumbling barrel, spraying the oil into the resulting mixture, and further blending the mixture until the oil has been incorporated therein and a substantially homogeneous blend formed. The mixture of soap and furfuryl alcohol can be prepared by heating the soap up to a temperature of about 50-100°C to liquify It, and then dissolving the molten soap in the furfuryl alcohol. Upon cooling, the soap remains dissolved in the furfuryl alcohol as a stable solution which can be stored until it is ready to be mixed with the mixture of oil, finely divided carbonaceous particles, phenolic resin, and phenolic resin hardener. The two mixtures, one liquid and the other essentially solid, can be readily mixed at room temperature, either manually or mechanically.

The TiB₂/carbon formulation can be molded into the desired shape or plate by many different approaches. The types of mold used may vary from a simple metal, plastic or other rigid mold to a disposable mold or mold liner made of paper, styrofoam or similar material. For automation purposes, the mold may be more complex, similar to those used for injection molding in the plastics industry. A mold release such as Frekote or a mold liner may be used to assist in the release of the cast shape from the mold.

Filling the mold may be accomplished by hand, mechanical means, injection or with pressure or vacuum. Compaction to eliminate voids and maximize plate density can be promoted by vibration, ultrasonics, centrifugal force, pressure or other similar means. The mold and/or plate mixture may be preheated and/or heated during the molding operation.

The mixed material may be spread on a flexible substrate such as aluminum foil and then folded or bent into the desired shape. This is an ideal approach when it is desired to fit the plate material over a highly contoured surface.

A heat treatment cycle similar to that shown in Figure 1 is used to cure the liquid containing formulations. Accelerated cure cycles are possible when a heating system such as microwave is used to heat the molded material from within instead of from the exterior as in a hot air oven. With microwave curing it is possible to reduce all the heating times in Figure 1 by as much as a factor of eight. When curing large flat plates, it may be necessary to maintain a flattening pressure on the plate material during the cure process (e.g., place a heavy flat plate on top of the molded material).

For certain applications it may not be necessary to further heat the cured plates. If, however, carbonization is required, then the preferred temperature range for baking the plate material is 800 to lOOO'C. Temperatures above 1000^βC are not preferred and graphitizing temperatures must be avoided. The baking process should be carried out in an inert atmosphere, coke bed or similar protective environment to prevent excessive air burn. A higher degree of plate flatness may be achieved by pressing the plate between two flat surfaces during the baking process.

The plates may be attached to the cathode blocks 1) before the blocks are shipped to the smelter, 2) at the smelter but before the cathode is assembled, or 3) after the cathode has been assembled. In each case, a simple commercially available carbon cement can be used as adhesive. UCAR C-34 cement is a preferred carbon cement for affixing the plates to the carbon cathode substrate. The preferred process comprises placing a weight on each plate to maintain compression on the glue joint during cure of the cement, and heating the block and the plate to a temperature of from about 30^βC to about 40^βC prior to application of the glue and joining of the pieces. After cure of the cement at the prescribed temperature, about 100^βC to about 170^βC, all conventional cell construction and start-up procedures can be followed without modification.

The area covered with the plate material can range from the entire inner surface of the cathode cavity to less than 10 percent of the cathode surface below the anode or anodes. The preferred area to be covered ranges from the entire cathode surface directly below the anode or anodes to 50 percent of said area, with the ideal ranging from 90 to 100 percent of said area. It may be necessary to leave some small gaps or slots to permit cathode ram degassing and accommodate cathode block movements during cell heat-up and start-up. The plate material need not be continuous over the entire cathode surface. In the case of TiB. plates, small gaps between adjacent plates (1 to 5 mm) will be bridged by the molten metal. Similarly, TiB₂ particles in a carbon surface at an appropriate density will produce a pseudo-continuous aluminum wetting film by bridging between adjacent TiB₂ particles. In the case of the T1B₂/carbon composition, about 20 weight percent TiB₂ in the surface will produce a pseudo-aluminum wetted surface with a most preferred overall TiB₂ content in the surface layer of 35-70 weight percent to maximize wetting and to allow for mixing inhomogeneities and a viable coating Hfe. Modification of the TiB₂ particle properties and/or changing the plate formulation

IREA

OMPI and/or the TiB₂ distribution within the coati ng may enable the use of lesser amounts of TiB . Gaps between the pl ates should be 5 mm or less 1 n width with the preferred being 1 mm or less.

There i s no minimum or maximum plate thickness. Generally, the thicker the pl ate, the l onger the pl ate l ife. However, the greater the thickness, the greater the initial cost. The preferred plate thickness is from 0.8 to 1.27 cm to minimize the tendency for bl isteri ng or warpi ng of the plate. Maximum plate thickness would be consistent with anticipated cell l ife, I .e. , there Is no need to have a plate thickness to last 10 years if cell life Is anticipated to be only 7 years. Moreover, plates may be "layered" to achieve thicker surfaces as required. Exampl e 1

The following formulation was successfully cast and cured using a variety of mol d material s.

Formulation (initial weights, as weight percent) . 44.6% T1B₂ powder 27.5% UCAR C-34 solids 15.5S UCAR C-34 l iquid 12. H UCAR BB-6 Graphite particles

0.3% Great Lakes Fortafil 3, carbon fibers sized for UCAR C-34, 0.32 cm. Molds Material Size Depth Filled Styrofoam 5.1 cm diameter 1.4 cm

Waxed paper 5.1 cm diameter 1.4 cm

Polypropylene 6.7 cm diameter 1.3 cm

Cardboard 7.6 x 8.3 cm 1.9 cm

The cure cycl e shown in Figure 1 was used to cure the four test sampl es. The cured test sampl es were cut in half and one portion of each test sample, i ncluding its mol d material , was baked by heating to 1000'C in an argon atmosphere over a 24 hour period. After cooling it was observed that the styrofoam had been converted to a non-adherent dust, the polypropylene had disappeared, and the waxed paper and cardboard were charred. The resulting test cathode plate material s appeared identical to those surface l ayers formed by the prior coating technique on carbon cathode blocks.

« » Example 2

The attachment process was tested by gluing the baked portion of the styrofoam mold test sample to a piece of SK cathode block using UCAR C-34 cement. The UCAR C-34 cement was mixed, applied and cured over a temperature cycle up to 135^βC as per manufacturers instructions. The test sample could not be detached from the SK block by hand pressure.

A voltage drop of 0.1V was measured when a direct current of 0.7 A/cm was passed across the room temperature glue joint. After heating the glued assembly to 1000'C in an argon atmosphere over a 24 hour period and then cooling to room temperature, the voltage drop across the glue joint was less than 0.001V. The 0.7 A/cm² current density across the glue joint is similar to that which would be encountered in a conventional coke bake and the normal operation of a commercial cell. Even the initial 0.1V voltage drop across the glue joint will not adversely effect the coke bake procedure used to start up a new cell.

The test sample could be fractured from the SK block only by striking the sample strongly 4 times with a hammer to cause a shearing force on the baked glue joint. Hammer blows perpendicular to the glue joint had no observable effect on the glued assembly. Large horizontal shearing forces would not be normally encountered in the operation of a cell. The failure occurred within the glue joint, not at either bonding surface. Curing the cement joint while under pressure decreases the thickness of the glue line and increases its strength. In practice, a weight could be placed on top of each glued plate during the glue cure cycle to achieve the optimum glue joint. Example 3 A metal tray 19 cm by 19 cm by 6.3 cm deep was coated with Frekote mold release and filled to a depth of about 1.0 cm with the T1B₂/carbon formulation given in Example 1. The material was spread by hand and then vibrated level on a vibration table. The plate shrunk slightly from the edges of the mold during the cure cycle shown in Figure 1. There was no tendency for the cured plate material to stick in the mold. SEM photos and EDAX elemental maps

OMPI of polished cross sections of the plate material showed that no segregation had occurred during the use of the vibration table to smooth the plate material. The bottom surface of the vibrated plate was significantly smoother and denser than that achieved in a similar test where vibration was not used to level the material in the mold. Example 4

Example 3 was repeated except that the hand levelled material was covered with a sheet of styrofoam and then pressed with a metal plate. The material flowed easily under pressure. The styrofoam was easily removed from the surface of the uncured material without disturbing its shape. No problems were encountered during the cure and bake cycles. A good quality T1B₂/carbon plate was produced. Example 5

A quantity of the formulation given in Example 1 was placed on a sheet of aluminum foil and then vibrated on a vibration table. The wet material rapidly smoothed out to form a thin (about 0.64 cm thick) pancake like shape. The resulting pad of material was then easily folded into any shape or contour and maintained its shape during curing. After curing, the aluminum foil was easily removed from the shaped T1B₂/carbon piece. Upon baking to 1000^βC, the test piece maintained its strength, shape and had a metallic ring when struck. Complex shapes of the TiB₂/carbon cathode material can be formed by this procedure. Example 6

A 2.54 cm diameter centrifuge tube was partially filled with the formulation given in Example 1. No effort was made to pack the material in the bottom of the tube. After rotating for 2 minutes at 15,000 rpm in a bench top centrifuge, the solids were compacted in the bottom of the tube and the excess mix liquid, approximately 1/3 of which had separated in spinning, was poured off the top of the solids. During curing, the sample shrunk enough to permit its easy removal from the centrifuge tube. No problems were encountered during the cure or bake cycles. A good quality TiB₂/carbon shaped sample was produced. Reducing the mix liquid content of the sample increases the density of the baked shape. Centrifugal force thus provides a convenient means to force the wet material into a shaped mold and helps densify the resulting shape. Example 7 A steel mold 30.5 cm by 30.5 cm by 0.95 cm deep was coated with FREKOTE mold release and filled with the formulation given in Example 1. The material was spread by hand and then vibrated level on a vibration table. The cured plate showed no defects except for a slight curvature or warpage (3 mm distortion, perpendicular to the 30 cm plate). Tests with lower final cure temperatures determined that the distortion occurred between 150 and 165"C. The warpage was reduced by approximately 50 percent when a 20 pound weight was placed on top of the material during the cure cycle. A sheet of styrofoam was used to prevent the wet material from sticking to the 20 pound weight. The warpage was further reduced to approximately 30 percent of the original value by increasing the loading weight to 38 pounds. By maintaining sufficient pressure during the "plastic" period of the cure cycle, it is possible to control the flatness of a vibrated molded plate to any desired value. While there was no detectable change in composition, a slight difference in the appearance of a 1.2 cm wide perimeter zone around the top of the uncured plate material was observed whenever the mold was vibrated. The vibrational energy is preferentially transmitted to the wet mixture from the vertical mold edges rather than uniformly from the mold bottom with the present mold design and vibration table. Example 8

Plates were prepared as described in Example 7 except that the vibrating table was not used to level the wet material. Plates produced without use of the vibration table were significantly flatter than the corresponding ones in example 7. No detectable warpage was observed when a 20 pound weight was placed onto the plate during curing. Application of pressure during molding for the purpose of distributing the material in the mold may avoid the need for vibrating the wet material, and depending upon the application, a small weight during curing may or may not be required. Example 9

No bl istering or cracking problems were encountered when heati ng the various cured plates to 1000'C in an argon atmosphere over a 24 hour period. The baked plates had a characteristic metallic ring when struck. The electrical resistances of the baked plates were of the same order as that of the semi graphitic SK cathode block. Example 10

A test sample of the formulation in Example 1 was successfully cured in a conventional microwave oven in approximately one-eighth the time required when using a hot air oven or an infra-red heater. Uniform internal heating of the wet mixture enables an accelerated cure cycle compared to that for external heat (e.g. , Figure 1 ) which cures from the outside to the inside of the material . The latter process 1 s known to result in gas venting defects, such as blisters in the mater al , if the formulation parameters are not properly optimized, or if the cure cycle is too rapid. Productivity of a TiB₂/carbon plate production line could thus be greatly improved by the use of a microwave (or similar process) cure cycle instead of an external heat source.

While the discussion as related to this invention has been directed to plate and tile configurations, normally considered to be rectangular, hexagonal , octagonal , or square, the present invention also encompasses other configurations and contoured shapes, such as gri ds, tables, cylindrical , or tubular shapes. Further, it is recognized that elements made in accordance with the present invention may be disposed on or about a cathode substrate in such a fashion as to form a contoured cathode array.

It is understood that the above description of the present invention i s susceptible to adaptations, modifications, and changes by those skilled in the art, and the same are intended to be considered withi n the scope of the present invention, which is set forth i n the appended claims.

Claims

What is claimed is:

1. A monolithic molded, cured, and baked aluminum wettable preformed element comprising a Refractory Hard Material In a carbonaceous matrix bonded by amorphous carbon, said matrix characterized by an ablation rate essentially equal to the combined rate of wear and dissolution of said Refractory Hard Material in an aluminum cell environment.

2. An element as set forth in claim 1 wherein said Refractory Hard Material 1s selected from the group consisting of titanium diboride, titanium carbide, zirconium diboride, zirconium carbide, and alloys and mixtures thereof.

3. An element as set forth in claim 2 wherein said Refractory Hard Material is titanium diboride.

4. An element as set forth in claim 2, wherein said Refractory Hard Material comprises from about 20 to about 90 percent by weight thereof.

5. An element as set forth in claim 1, wherein said carbonaceous matrix is bonded by a thermosetting binder selected from the group consisting of phenolic, furane, polyphenylene, heterocyclic, epoxy, silicone, alkyd, and polyi ide resins, and mixtures and copolymers thereof, having a char yield of greater than 25 percent.

6. An element as set forth in claim 5, comprising from about 35 to about 70 percent by weight titanium diboride, from about 1 to about 60 percent by weight carbonaceous filler and additive, and from about 1 to about 40 percent by weight binder system having a char yield of greater than 25 percent.

7. An element as set forth in claim 6, wherein said binder system comprises a phenolic resin.

OMPI ιA_*», WIPO

8. An element as set forth i n claim 7, wherein said binder system compri ses a phenol ic-furane copolymer.

9. An element as set forth in claim 8, wherein said carbonaceous filler consists of a mixture of graphite flour and carbon black.

10. An element as set forth in claim 9, wherein said carbonaceous additive includes carbon fiber.

11. An aluminum-wettable element comprising Refractory Hard Material in a non-graphit1c carbon matrix bonded by amorphous carbon, said element formed by molding a composition comprised of Refractory Hard Material , carbonaceous filler, carbonaceous additive, and thermosetting binder system, said binder system having a char yiel d greater than about 25 percent, curing said composition to form a sol id self-supporting unitary mass, and baking said mass.

12. An element as set forth i n claim 11 , wherei n sai d Refractory Hard Material i s selected from the group consisting of titanium diboride, titanium carbide, zirconium diboride, zirconium carbide, and alloys and mixtures thereof.

13. An element as set forth i n claim 12, wherein said thermosetting binder system comprises at least one resin selected from the group consisting of phenol ic, furane, polyphenylene, heterocyclic, epoxy, silicone, al kyd, and polyimlde resins.

14. An element as set forth in claim 13, wherein said carbonaceous filler comprises at least one member selected from the group consisting of carbon flour, graphite flour, crushed coke, crushed graphite, and carbon black, and is less than 100 mesh in size.

15. An element as set forth i n claim 14, wherein said carbonaceous additive comprises at least one member of the group consisti ng of carbon aggregate and carbon fiber.

OMPI