US3150443A - Process of incorporating a refractory metal oxide in a metal and product resulting therefrom - Google Patents

Description

Sept 29, 1964' G. B. ALEXANDER ETA 3 1 PROCESS OF INCORPORATING A REFRACTORS F METAL OXIDE IN A METAL AND PRODUCT RESULTING THEREFROM Flled Dec. 7, 1959 5 Sheets-Sheet l FIG.2

GLOBAR HEATING ELEMENT THERMOCOUPI-E 4C4 5"INCONEL TUBE WELL b I Q I 1-)- HYDROGEN WATER AND IN HYDROGEN OUT I r I ,l l l s O-RING SEAL I I A E FIG. I

26,33? NICKEL 'NITRATE COLLOIDAL THoR|A AMMONIUM cA RBoNATE "PUMP INVENTORS:

GUY B. ALEXANDER WILLIAM H. PASFIELD PAUL C. YATES ATTORNEY TO FILTER Sept. 29, 1964 G. B. ALEXANDER ETAL 3,150,443

PROCESS OF INCORPORATING A REFRACTORY METAL OXIDE IN A METAL AND PRODUCT RESULTING THEREFROM Filed Dec. 7, 1959 5 Sheets-Sheet 2 COOLING WATER HYDROGEN AND 0-!- WATER OUT HYDROGEN GLOBAR HEATING ELEMENT ---FIRE BRICK SCRE EN INVENTORS Fl G 3 GUY B ALEXANDER WILLIAM H. PASFIELD PAUL C. YATES ATTORNEY p 1964 G. B. ALEXANDER ETAL 3, 5 43 PROCESS OF INCORPORATING A REFRACTORY METAL OXIDE IN A METAL AND PRODUCT RESULTING THEREFRQM Filed Dec. 7, 1959 5 Sheets-Sheet 3 FIG.4

GRAIN BOUNDARY PARTICLE A REGION B REGlON A DIRECTION REG'ON 5 OF EXTRUSION REGION A INVENTORS GUY B ALEXANDER l WILLIAM H. PASFIELD 11 -345 3 PAUL c. YATES LONGITUDINAL 20' EXTRUDED SECTION .7246. EM ATTORNEY Sept. 1954 G. B. ALEXANDER ETAL 3,150,443

5 OF INCQRPORATING A REFRACTO METAL OXIDE A METAL AND PRODUCT RESULTIN HEREFR PROCES IN Filed Dec. 7, 1959 5 ts-Sheet 4 INVENTORS LEXANDER GUY B A WILLIAM H. PASFIELD PAUL c. YATES ATTORNEY Sept. 29, 1964 G. B. ALEXANDER E P CORPORATING A REFRACTORY METAL OXIDE ODUCT RESULTIN FIG.7

INV GUY B ALEXANDER WlLLlAM H. PASFIELD PAUL C. YATES 4 ZEMW United States Patent 3,150,443 PRGCESS 0F INCORPORATING A REFRACTORY METAL 0E IN A METAL AND PRODUCT RESULTING THEREFROM Guy B. Alexander, Brandywine Hundred, Del., William H. Pastield, Sayville, N.Y., and Paul C. Yates, Brandywine Hundred, Del., assignors to E. I. du Pont de Nemours and Company, Wilmington, DeL, a corporation of Delaware Filed Dec. 7, 1959, Ser. No. 857,726 13 Claims. (Cl. 29--182.5)

This invention is concerned with improving the hightemperature service characteristics of metals whose oxides have a free energy of formation (AF) at 27 C. of from 30 to 105 kilocalories per gram atom of oxygen (kcal./ gm. at. 0) and of alloys of these metals. The improvement is accomplished by incorporating in the metals or alloys a particulate metal dispersion of submicron-sized particles of a refractory metal oxide.

More particularly the invention is directed to processes in which (a) a hydrous, oxygen-containing compound of a metal having an oxide with a AP at 27 C. of from 30 to 105 kcal./gm. at. 0 is deposited together with substantially discrete, submicron particles of a metaloxygen compound which when heated to constant weight at 1500 C. is a refractory oxide having a melting point above 1000 C. and a AP at 1000 C. above 60 kilocalories per gram atom of oxygen, (b) thereafter the hydrous, oxygen-containing compound is reduced to the corresponding metal while maintaining a temperature throughout the entire mass below the sintering temperature of the metal, whereby a powdered product is obtained, (0) the powdered product is sintered at a temperature below the melting point of the metal until its surface area is less than about square meters per gram, and (d) the sintered powder is mixed with a powdered metal.

The invention is further particularly directed to pulverulent products comprising a particulate dispersion of submicron refractory oxide particles of the above-mentioned type in a metal of the class to be improved, said particulate dispersion having a surface area less than 10 square meters per gram (m. /g.), mixed with a powdered metal having a hardness less than that of the refractory-containing metal, the AF of the refractory oxide at 1000" C. being at least the value calculated'from the expression 52+.016M, where M is the melting point, in "Kelvin, of the metal in the dispersion.

The invention is still further directed to sintered, solid metal compositions in which there are volumes of a metal having substantially uniformly dispersed therein discrete, submicron particles of a refractory metal oxide having a AP at 1000 C. which is more than 60 kcaL/gm. at. 0 and at least the value calculated from the expression 52+.0l6M, where M is the melting point, in degrees Kelvin, of the metal, the said volumes being intermingled with other volumes of metal and the entire metal composition having an apparent density which is from 60 to 100% of the absolute density.

The invention is additionally directed to solid, annealed metal compositions of the type described, in which there are (a) volumes of a metal, the grains of which are in contact with discrete, submicron particles of a refractory metal oxide having a AP at 1000 C. which is more ice than 60 kcaL/ gm. atom 0 and at least the value calculated from the expression 52+.016M, where M is as above described, and (b) other volumes of the same metal in which the metal grains are not in contact with said refractory oxide particles, the average grain size in the volumes (b) being at least two-fold the average grain size in volumes (a).

In the drawings FIGURE 1 illustrates operation of the initial step of a process of the invention,

FIGURE 2 shows one way to effect the reduction step of the process,

FIGURE 3 shows another embodiment of the reduction step,

FIGURE 4 is a line drawing prepared from an electronmicrograph of a solid product of the invention, showing the heterogeneous structure,

FIGURE 5 is also a line drawing prepared from an electronrnicrograph of a solid product, in longitudinal section, showing the effect of hot-working by extrusion,

FIGURE 6 is a fanciful representation of a powder product of the invention, showing a mixture of filled and unfilled metal, and

FIGURE 7 is a fanciful representation of a pressed compact of a powder of FIGURE 6, before hot-working.

Attempts have already been made to incorporate refractory particles into such metals as copper in the hope that such inclusions might impart greater strength to the metals, especially at elevated temperatures. Obtaining an adequate degree of dispersions has been a problem, however, and it has not hitherto been know how to incorporate discrete, very finely divided refractory particles into many other metals, such as iron, cobalt, and nickel and their alloys. Generally, oxide occulsions in metals have been viewed with disfavor, and various means, such as scavenging with active metals, have been adopted to get rid of them.

Methods have been proposed for combining such finely divided powders as aerogels with finely divided metals, using the techniques of powered metallurgy, but have been unsuccessful as a way to improve the high-temperature properties of metals. In such methods the aerogel is mixed with the solid metal, and the whole mass is subjected to very high pressures. Under the circumstances, the ultimate particles in the aerogel structure are forced into intimate contact, producing a densely aggregated structure which cannot be broken down and redispersed when the resulting compact is worked, either hot or cold. Thus, the compositions prepared by this technique consist of metal having very coarse oxide particles dispersed therein, the oxide particles being in the 10 to micron range.

In our parent application Serial No. 694,086 filed November 4, 1957, now Patent No. 3,019,103 we have described novel processes and compositions in which the just-mentioned difficulties of the prior art have been overcome. Very substantial improvements in the high-temperature service characteristics of the products are achieved. When, as therein described, the submicron refractory oxide particles are dispersed substantially uniformly throughout the entire metal mass the products are very hard. While this is desirable in many situations, it does make the metal difiicult to work by such common techniques as machining and forging.

Now according to the present invention it has been found that improved high-temperature service characteristics can be achieved and adequate ductility for easy workability can be retained by dispersing the submicron particles of refractory oxide filler in metal as in the parent application, and then mixing the filled metal, in powder form, with a metal powder unmodified with filler particles, the refractory oxide filler and the filled and unfilled metal being so selected that the AF of the refractory oxide is at least the value calculated from the expression 52+.0l6M where M is the melting point of the metal in K. Such mixed powders can be compacted, sintered, and hot-worked to upwards of 90% of the theoretical density to give solid metal products in which there are volumes of refractory-filled metal intermingled with other volumes of unfilled metal. The latter products can be annealed until the metal grains in contact with the refractory particles have the same chemical composition as the metal grains not in contact with refractory particles, in which case the latter grains are at least twofold larger than the former and are less hard. These sintered and the annealed products are surprisingly ductile, while retaining improved high-temperature properties. In the processes for making the novel blended powders and the subsequent solid products unique and unexpected freedom from excessive and unwanted oxidation of the filled metal is achieved by sintering the filled metal particles until their surface area is below about square meters per gram before exposing them to air or other reactive gases.

THE FILLER In describing this invention the dispersed refractory particles will sometimes be referred to as the filler. The word filler is not used to mean an inert extender or diluent; rather, it means an essential constituent of the novel compositions which contributes new and unexpected properties to the metalliferous product. Hence the filler is an active ingredient.

In processes of this invention, a relatively non-reducible oxide is selected as the filler, that is, an oxide which is not reduced to the corresponding metal by hydrogen, or by the metal in which it is embedded, at temperatures below 1000 C. Such fillers have a AF at 1000" C. of more than 60 kilogram calories per gram atom of oxygen in the oxide. The oxide itself can be used as the starting material or it can be formed during the process by heating another metal-oxygen-containing material.

The metal-oxygen-containing material can, for example, be selected from the group consisting of oxides, carbonates, oxalates, and, in general, compounds which, after heating to constant weight at 1500 C., are refractory metal oxides. The ultimate oxide must have a melting point above that of the metal in which it is being used. Oxides with melting points above 1000 C. are preferred. A material with a melting point in this range is referred to as refractorythat is, difficult to fuse. Filler particles which melt or sinter at lower temperatures become aggregated.

The filler can be mixed oxide, particularly one in which each oxide conforms to the melting point and AF above stated. Thus, magnesium silicate, MgSiO is a mixed oxide of MgO and SiO Each of these oxides can be used separately; also, their products of reaction with each other are useful. The filler, accordingly, is a single metal oxide or a reaction product of two or more oxides, also, two or more separate oxides can be used as the filler. The term metal oxide filler broadly includes spinels, such as MgAl O and ZnAl O metal carbonates, such as BaCO metal aluminates, metal silicates such as magnesium silicate and zircon, metal titanates, metal vanadates, metal chromites, and metal zirconates. With specific reference to silicates, for example, one can use complex structures, such as sodium aluminum silicate, calcium aluminum silicate, calcium magnesium silicate, calcium chromium silicate, and calcium silicate titanate.

Oxide AF at; Oxide AF at 1000 C 1000 C While any of the above-mentioned refractory metal oxides will have utility, there is a correlation between the AF of the refractory and the melting point of the metal in which it is to be used. Thus, the AF of the refractory should be at least the value calculated from the expression 52+.016M, where M is the melting point of the metal in which it is to be used. It will be understood that the melting point referred to is that of the metal as it exists in the final product. Thus, if the final metal is an alloy of two or more metals, M will be the melting point of the alloy. Moreover, in the blended powders or sintered compacts the filled metal can have one melting point and the unfilled or ductile metal another; again, it is the melting point of the final product which is to be considered.

It will be seen that as the melting point of the metal increases, the AF of the refractory to be used also increases. For example, the AF of zirconia, 100, is too low to permit it to be used in tungsten metal (M.P.=

3370 C.) to get products of the invention; on the other hand, the AF of thoria, 11-9, is well suited for use with molybdenum metal (M.P.=2620 C.).

The filler oxide must be in a finely divided state. The substantially discrete particles should have an average dimension in the size range below 1 micron, preferably from 5 to 250 millimicrons, an especially preferred range being from 5 to millimicrons, with a minimum of 10 millimicrons being even more preferred.

The particles should be dense and anhydrous for best results, but aggregates of smaller particles can be used, provided that the overall aggregate has the above-mentioned dimensions. Particles which are substantially spheroidal or cubical in shape are also preferred, although anisotropic particles such as fibers or platelets can be used for special efforts. Anisotropic particles produce metal compositions of lower ductility.

The size of a particle is given as an average dimension. For spherical particles all three dimensions are equal and the same as the average. For anisotropic particles the size is considered to be one third of the sum of the three particle dimension. For example, a fiber of asbestos might be 500 millimicrons long but only 10 millimicrons wide and thick. The size of this particle is or 173 millimicrons, and hence within the preferred limits of this invention.

Colloidal metal oxide aquasols are particularly useful as a means of providing the fillers in the desired finely divided form and hence are preferred. For example, silica aquasols such as those described in Bechtold et al. US. Patent 2,574,902, Alexander U.S. Patent 2,750,345, and Rule U.S. Patent 2,577,485 are suitable as starting materials. Zirconia sols are likewise useful. The art is familiar with titania sols, and such sols as described by Weiser in Inorganic Colloidal Chemistry, Volume 2, Hydrous Oxides and Hydroxides, for example, can be used. Also, the beryllia sols described on page 177 of this reference can be used. Thoria aquasols can be prepared by calcining thorium oxalate to 650 C. and dispersing the resulting colloidal thoria in dilute acids.

Although they are less preferred, aerogels and reticulated powders can also be used. For example, products described in Alexander et al. US. Patent 2,731,326 can be employed. In these instances it is necessary that the aggregate structures be broken down to particles in the size range specified, for example, by colloid milling an aqueous slurry to which a small amount of peptizing agent has been added.

Powders prepared by burning metal chlorides, as, for example, by burning silicon tetrachloride, titanium tetrachloride, or zirconium tetrachloride to produce a corresponding oxide, are also very useful if the oxides are obtained primarily as discrete, individual particles, or aggregated structures which can be dispersed to such particles.

Calcium oxide is a useful filler. Since this oxide is water soluble or, more accurately, Water reactive, one cannot obtain it as an aqueous dispersion in the colloidal state. In this instance, one can use an insoluble calcium compound, such as the carbonate or oxalate, which, on heating, will decompose to the oxide. Thus, for example, particles of finely divided calcium carbonate can be coated with an oxide of the metal in which it is to be dispersed, e.g., hydrous iron oxide, by treating a dispersion of finely divided calcium carbonate with a base and a salt of the metal, e.g., ferric nitrate and sodium carbonate. On heating the precipitate and reducing, a dispersion of calcium oxide in iron is obtained. Similarly, one can obtain dispersions of barium oxide, strontium oxide, or magnesia in the metal being treated.

THE METAL TO BE FILLED The metal in which the refractory oxide is to be incorporated is selected from the group consisting of metals whose oxides have a free energy of formation at 27 C. of from 30 to 105 kcal./gm. at 0. These metals, and the free energies of formation of their oxides, are as follows:

of these metals can, of course, be used.

COATING THE FILLER In processes of preparing the filled metal particles, a relatively large volume of the metal oxide, hydroxide, hydrous oxide, oxycarbonate, or hydroxycarbonate, or, in general, any compound of the metal wherein the metal is in an oxidized state, is formed as a coating around the refractory oxide filler. This coating can be a compound of single metal, or it can contain two or more metals. For example, the hydrous oxides of both nickel and cobalt can be deposited around a filler. In the latter case, an alloy of nickel and cobalt is produced directly, by reducing with hydrogen.

In a similar manner, alloys of any metals which form oxides that can be reduced with hydrogen, can be prepared. Thus, alloys of iron, cobalt, nickel, copper, molybdenum, tungsten, chromium and rhenium can be prepared by codepositing oxides of two or more of the selected metals on the filler particles and subsequently reducing.

To produce such a hydrous, oxygen-containing composition one can precipitate it from a soluble salt, pref- Mixtures 6 erably a metal nitrate, although metal chlorides, sulfates, and acetates can be used. Ferric nitrate, cobalt nitrate, and nickel nitrate are among the preferred starting materials.

The precipitation can be conveniently accomplished by adding a suitable soluble metal salt to an aqueous alkaline solution containing the filler particles, while maintaining the pH above 7. A good way to do this is to add, simultaneously but separately, a concentrated, aqueous solution of the soluble metal salt, a colloidal aquasol containing the filler particles, and an alkali such as sodium hydroxide, to a heel of water. Alternatively, a dispersion containing the filler products can be used as a heel, and the metal salt solution and alkali added simultaneously but separately thereto.

More broadly, in depositing the compound of a metal in an oxidized state upon the filler, one can react any soluble salt of these metals with a basic material. Hydroxides such as NaOH, KOH, or ammonia, or carbonates such as (NH CO Na CO or K 00 can be used. Thus, the metal compound deposited can be an oxide, hydroxide, hydrous oxide, oxycarboante, or in general, a compound which, on heating, will decompose to the oxide.

During the coating process certain precautions are preferably observed. It is preferred not to coagulate or gel the colloid. Coagulation and gelation are avoided by simultaneously adding the filler and the metal salt solution to a heel.

It is preferred that the filler particles be completely surrounded with the reducible oxides or hydrous oxides such as those of iron, cobalt or nickel, so that when reduction occurs later in the process, aggregation and coalescence of the filler particles is avoided. In other words, it is preferred that the ultimate particles of the filler be not in contact, one with another, in the coprecipitated product. Another condition which is important during the preparation to insure this condition is to use vigorous mixing and agitation.

Having deposited the hydrous oxygen compound of the metal, such as iron, cobalt or nickel, on the filler, it is then desirable to remove the salts formed during the reaction, by washing. Ordinarily, one uses an alkali such as sodium hydroxide, potassium hydroxide, lithium hydroxide, ammonium or tetramethylammonium hydroxide in the deposition of the compound. As a result, salts like sodium nitrate, ammonium nitrate or potassium nitrate may be formed. These should be removed, since otherwise they may appear in the final product. One of the advantages of using the nitrate salts in combination with aqueous ammonia is that ammonium nitrate is volatile, and therefore is easily removed from the product. However, the tendency of many metals, such as cobalt and nickel, particularly, to form amine complexes, is a complicating reaction in this case. By carefully controlling the pH during coprecipitation, these side reactions can be avoided.

Having essentially removed the soluble, non-volatile salts by washing, the product is then dried at a temperature above C. Alternatively, the product can be dried, and the dry material suspended in water to remove the soluble salts, and thereafter the product redried.

PROPORTIONS OF DEPOSITED METAL COM- POUND AND FILLER The relative amount of oxidized metal compound coating which is deposited on the filler particles depends somewhat on the end use to which the product is to be put, but is generally such that the end product contains an amount of filler in the range from 1 to 35% by volume and preferably from 5 to 20%. At loadings below 1% the filled metal is too soft, and at very high loadings the powders are difficult to work with because there is insufiicient metal phase to hold the aggregate structure together.

Volume loadings as high as 50%, that is, one volume of oxide for each volume of metal present, can be successfully used, but such products are often pyrophoric. Even heating to a temperature approaching the melting point of the metal after reduction does not completely eliminate this problem. Also, the filler particles in such products tend to coalesce to form large, hard aggregates during the reduction step. This tendency can be reduced by increasing the particle size of the filler, say to 100 millimicrons or even larger. The problems just discussed are minimized as the volume loading is reduced. In the range of 40 to 50 volume percent of filler, it is advisable to protect the modified metal in an inert atmosphere (hydrogen, argon or nitrogen) until the material is blended with unmodified metal and compacted. Even at 10 volume percent, one often has difiiculty sintering the modified metal mass sufficiently that it can be handled in air.

REDUCING THE PRECIPITATE CONTAINING THE FILLER Having deposited the compound of metal in oxidized state around the filler particles and dried the product, the next step is to reduce the coating to the metal. This can be conveniently done by subjecting the coated particles to a reducing agent, such as a stream of hydrogen at a somewhat elevated temperature. However, the temperature throughout the entire mass must not be allowed to exceed the sintering temperature of the filler particles. One way to accomplish this is to place the product in a furnace at controlled temperature, and add hydrogen gas slowly. Thus, the reduction reaction will not proceed so rapidly that large amounts of heat are liberated and the temperature in the furnace is increased.

Hydrogen to be used in the reduction can be diluted with an inert gas such as argon, or in some cases nitrogen, to reduce the rate of reaction and avoid hot spots. In this way the heat of reaction will be carried away in the gas stream. Alternatively, the temperature in the furnace can be slowly raised into the range of 500 to 700 C. while maintaining a flow of hydrogen over the product to be reduced.

In addition to, or instead of, hydrogen, carbon monoxide can be used as the reducing agent, particularly at elevated temperatures, as well as methane or other hydrocarbon gases. In any case, it is important that the temperature during reduction be controlled, not only to avoid premature sintering as above mentioned, but also so that excessive reaction will not occur between the reducible compound (such as iron, cobalt or nickel oxide) and the filler oxide before the reducible compound is reduced.

Reduction should be continued until the coating is essentially completely reduced. When the reaction is nearing completion, it is preferred to raise the temperature to complete the reduction reaction, but care must be taken not to exceed the melting point of the reduced metal. Reduction should be carried out until the oxygen content of the mass is substantially reduced to zero, exclusive of the oxygen of the oxide filler material. In any case, the oxygen content of the filled metal, exclusive of the oxygen in the filler, should be in the range from to 2% and preferably from 0 to 0.05%, based on the weight of the filled metal.

One way of estimating the oxygen content is to measure the change in weight of a product on treatment with dry, oxygen-free hydrogen at 1300 C. Products which show a change in weight of only from 0.0 to 0.05% under this condition are preferred.

After the reduction reaction is complete, the resulting powder is sometimes pyrophoric. Therefore, it is preferred to cool and store the mass in an inert atmosphere such as argon, and further to blend and compact the mass to reduce surface area in the absence of oxygen or nitrogen.

It is particularly desirable that the filled metal powder be stored in an inert atmosphere such as argon if its surface area is greater than .1 square meter per gram. The atmosphere should be essentially free of oxygen, water vapor, nitrogen, sulfur, and any other elements or compounds which are reactive with the metal powder.

An alternate way of reducing the metal in contact with the refractory filler is to subject the coated particles to a metal reducing agent in a fused salt bath. The compound coated refractory oxide particles are dispersed in the molten salt and the reducing metal is added while maintaining the temperature of the molten salt in the range of 400 to 1200 C.

The fused salt bath is merely a medium whereby to effect contact of the reducing agent and the metal compound under conditions which will not affect the disposition of the compound with respect to the refractory particles. It can comprise any suitable salt or mixture of salts having the necessary stability, fusion point, and the like.

Suitable fused salt baths can comprise halides of metals selected from groups I and 11a of the periodic table.

In general, the chlorides and fluorides are preferred halides. Bromides or iodides can be used, although their stability at elevated temperatures is frequently insufficient. Chlorides are especially preferred. Thus, among the preferred salts are calcium chloride, sodium chloride, potassium chloride, barium chloride, strontium chloride, and lithium chloride and fluoride.

The fused salt bath will usually be operated under a blanket of either an inert gas or a reducing gas. Such gases as helium, argon hydrogen or hydrocarbon gases can be used in this capacity.

The temperature of the reduction can be varied considerably, depending upon the combination of fused salt and reducing metal selected. In general, the temperature of reduction will be between 400 and 1200 C. It is usually preferred to select a reduction temperature at which the reducing metal, as well as the fused salt, is present in a molten state. Usually the operating temperature will also be below the boiling point of the reducing metal employed.

The operating temperature of the reduction bath must also be below the melting point of the metal coating to be produced on the refractory filler. For example, if a tungsten compound is being reduced upon particles of thoria, reduction temperature as high as 1200 C. can be employed.

However, if a compound of copper, or of a coppercontaining alloy having a low melting point, is being reduced, the reduction temperature should be maintained below that of the melting point of the copper or the alloy.

The reducing metal is selected from the group consisting of alkali and alkaline earth metals. Thus, the metal can be lithium, sodium, potassium, rubidium, caesium, beryllium, magnesium, calcium, strontium, or barium.

It is preferred to use a reducing metal which has a low solubility in the solid state with respect to the metal of the coating on the refractory oxide particles; otherwise, one will get undesirable alloying of the reducing metal with the metal formed by the reduction. For this reason, calcium and sodium are suitable for reducing compounds of such metals as iron, cobalt, nickel, chromium, or tungsten, while magnesium and sodium are useful in reducing titanium.

It is preferable to use a temperature of reduction at which the reduction reaction proceeds at a rapid rate. For reducing cobalt, iron, and nickel compounds, temperatures in the range of 600 to 800 C. are suitable. With compounds of metals such as chromium, titanium, and niobium, temperatures in the range of 850 to 1000 C. are used.

Completion of the reduction reaction can be determined by taking samples from the melt, separating the product from the fused salt, and analyzing for oxygen by ordinary analytical procedures such as vacuum fusion.

The reduction is continued until the oxygen content of the mass is substantially reduced to zero, exclusive of the oxygen of the oxide refractory material. In any case, the oxygen content of the product, exclusive of the oxygen in the refractory, should be in the range of from to 2 percent and preferably from 0 to 0.05%, based on the weight of the product.

The reduced product is present as a suspension in the fused salt bath. It can be separated therefrom by the techniques ordinarily used for removing suspended materials from liquids. Gravitational methods such as settling, centrifuging, decanting and the like can be used, or the product can be filtered off. Alternatively, the bath can be cooled and the fused salt dissolved in a suitable solvent such as dilute aqueous nitric acid or acetic acid.

If a considerable excess of reducing metal is used in the reduction step, it may be necessary to use a solvent less reactive than Water for the isolation procedure. In such a case, methyl or ethyl alcohol is satisfactory. The presence of a small amount of acid in the isolation solvent will dissolve any insoluble oxides formed by reaction between the reducing metal and the oxygen content of the coating being reduced. After filtering off the reduced metal powder, it can be dried to free it of residual solvent.

It will be understood that irrespective of which of the above-mentioned reduction methods is used one can mix a dried, refractory-filled metal oxide and an oxide of the metal which is to be the matrix of the final product and coreduce the mixture to get solid metal products of the invention.

THE REFRACTORY-FILLED METAL POWDER The refractory-filled metal powder obtained as just described can be compacted and sintered directly to useful metal products as disclosed and claimed in our abovementioned parent application Serial No. 694,086. To get the heterogeneous island-structured products of the present invention the compacting is carried out in the presence of unfilled metal powders. Thus, compositions of the parent application can be used as a masterbatc in making the powder metallurgical products of the pres ent case. For example, iron containing zirconia can be used as a masterbatch and alloying agent in other metallurgical compositions, and can further be diluted with untreated iron as a finely divided powder, or can also be used as an alloying agent for other metals, such as, for example, copper, molybdenum, nickel, and tungsten. Compacting to a dense body can be effected after mixing the iron containing the Zirconia with the finely divided powders of the other metals.

In a similar manner, masterbatches of cobalt or nickel containing the desired fillers can be used to prepare various alloys. For example, an improved type of nickelcopper alloy having island structure can be prepared by dispersing zirconia in nickel and thereafter alloying the nickel with copper by powder metallurgical techniques. The oxidation resistance of products made in this manner is also improved. Thus, such materials are more oxidation resistant than the unmodified metals, the extent of improvements being related to the amount of dilution with unmodified metal.

In still another specific example, a nickel-molybdenum alloy can be prepared from a masterbatch of nickelthoria, nickel-alumina or nickel-calcia. In making masterbatches, higher loadings of oxides are used, volume loadings in the range from 2 to 20% being preferred. In this way, a modified nickel powder of very high hardness is produced. This modified nickel powder is then blended with a nickel-molybdenum, oxide-free powder, rich in molybdenum. Blending is accomplished by tumbling the powders (which are preferably small enough to pass through a 325-mesh screen) in a conical or cubical container for several hours. The blended powder is compacted to a non-porous mass, i.e., having a density at least 95% of absolute density. Thereafter, the composition is heat-soaked and worked, for example, at 1300 C., in order to diffuse the molybdenum throughout the nickel. The resulting composition, containing nickel as a metallic component, is considerably more ductile than a composition containing similar components but made directly. In making alloys in this manner it is important to use powders which are essentially oxygen free, except for the oxygen in the filler.

It is believed that the reason for the above-mentioned improvement in ductility is that the composition made by powder dilution from a masterbatch consists of two phases: one phase, the matrix, of ductile, unmodified, large-grained metal and another phase, the islands, of extremely hard, fine-grained, highly reinforced metal. The unmodified phase gives the composition ductility and the hard phase (metal-l-oxide) gives it unusual strength at high temperatures. Thus, the advantage of the above dilution technique is that it produces a metal material of higher ductility and lower hardness, and at the same time gives high-temperature strength properties due to the presence of the oxide filler.

The nature of the metallic component and refractory and the proportions of these components in the filled metal powdershave been described at length, above. Other factors important in the preferred blended powders will now be discussed.

The particle size of the filled metal to be used in the powder blending preferably is kept relatively small, i.e., less than 500 microns, and can be as small as 1 micron. minimizing the particle size can be accomplished by minimizing sintering during reduction, i.e., by reducing at a minimum temperature. If the powder particles are too small they have a tendency to be pyrophoric; however, even such powders are useful if handled in the absence of air, for example, in a clean, dry argon or hydrogen atmosphere. Powders like thoria-filled nickel or lanthana-filled iron can be handled in helium or hydrogen.

Powders for blending should in any event pass 50 mesh. Powders which will pass 325 mesh (44 microns) are desirable, and powders of particles smaller than 10 microns are preferred. Size as used here refers to the aggre gate structure of the metal-metal oxide powders. Such powders are porous and may have internal surface area which is 10 or even times the surface area which would be calculated from overall dimensions.

The grain size of the metal in the filled-metal particles ordinarily is less than 10 microns and preferably in the range of 0.5 to 3 microns. Grain size can be estimated by the calculation where G is the grain size in microns, d is the diameter of the filler particles in microns, and f is the volume fraction of the filler. To illustrate, in an alumina-filled chromium powder in which alumina particles of 0.3 micron are present to 15% by volume, the estimated grain size would be 0.7 micron.

If it is desired to measure grain size directly, this can be done by any conventional metallurgical technique, for example, as indicated in Example 1.

It is preferred that the filler-metal powder be reasonably free of oxygen, nitrogen or any contaminant which may interfere with diffusion or with bonding at grain boundaries, or which may impart brittleness to the metal of the ultimate product. Oxygen, for example, preferably should be less than 0.05%, exclusive of oxygen in the filler.

THE UNFILLED METAL POWDER FOR BLENDING The powder selected for blending with the masterbatch should be ductile. In the case of many metals, like chromium, this means that it must be essentially free of certain contaminants, for example, oxygen, nitrogen, carbon,

3 l. boron, and sulfur. The ductility of this phase is particularly important, since otherwise it will be difiicult to prepare dense, metallurgically sound metal after compacting and working the blended powders.

The powder preferably should be a metal or an alloy which, when in massive form, forms a somewhat protective oxide film. In the final product, the filler will improve the protective nature of the oxide film. Thus, for products of maximum corrosion resistance the unfilled metal powder preferably will consist of chromium, nickel, cobalt, iron, titanium, manganese, vanadium, silicon, aluminum, magnesium, zinc, zirconium, niobium, copper, the rare earth metals, and alloys of any of these metals. In general, these are metals having a melting point in the range of from 200 to 2400 C. and which form oxides having a free energy of formation in the range from 30 to 140 kcal./ gm. atom of oxygen at 27 C.

If the metal powders being blended are pyrophoric, they should be blanketed by an inert atmosphere during blending. This will prevent oxidation, and hence undesired contamination.

The ductile metal powder may contain a relatively small amount of filler, as, for example, up to about 1% by volume. However, as the amount of filler increases, the ductility of the powder decreases, and hence filler contents greater than about 4% by volume in the ductile metal phase are generally not desired.

The ductile metal powder should also have a large grain size, or be capable of recrystallizing to large grains on heating to a temperature of 0.6 times its melting point in degrees absolute. If the potential grain size is calculated as above, it should be greater than microns and preferably greater than 50 microns. If one has filler particles which are 0.1 microns in size in the ductile metal phase, the preferred concentration of the tiller particles is less than 0.25 volume percent. With filler particles of this size, i.e., 0.1 micron, volume loadings as low as 0.02% will produce a beneficial effect. Thus, a preferred powder for use as the ductile phase is one containing filler at such a loading and of such size that the calculated grain size is in the range from 50 to 500 microns. In such compositions, the tiller should be uniformly dispersed throughout the ductile metal phase, and should be less than 1 micron in size. Such powders can be prepared using a process similar to that used in preparing the refractory-filled metal powder, except that a much smaller amount of filler is added.

In selecting ductile metal powders for blending, it is preferred that they be in the size range below 500 microns. More preferred powders are those which are in the size range from 2 to 20 microns.

THE POWDER-BLENDING STEP Blending ratios of masterbatch to ductile metal can vary over wide ranges. However, in most cases, from 0.1 to 10 parts of masterbatch, or filled metal powder, is blended with 1 part of ductile metal powder. In the more preferred case, from 0.3 to 2.5 parts by volume of masterbatch is blended with each part by volume of ductile metal powder.

The processes of the invention include the step of blending a refractory-filled metal powder, prepared as above described, with a ductile metal powder. This can be accomplished in several ways, as, for example, by tumbling, ball milling or any other processes used in the art for mixing powders. Instead of blending metal powders, one can blend compounds which are reducible to metals, such as a metal oxide, hydroxide, hydroxycarbonate, or carbonate. In the latter embodiment, any metal compound which on heating and reduction can be converted to a metal can be used in place of the ductile metal powder. As an example, a nickel-beryllia powder can be blended with nickel oxide, and the blended powders reduced in hydrogen at a temperature of 400 to 900 C. Conversely, a nickel hydroXide-beryllia masterbatch can be mixed with nickel hydroxide, and this mixture reduced after blending. However, the use of a refractory-filled metal powder in the blending operation is preferred.

COMPACTING THE POWDERS Following the blending steps, the entire mass of blended, unfilled metal powder and filled metal powder masterbatch is compacted. This can be done by ordinary powder metallurgical techniques, for example, by subjecting the product to high pressures, at ordinary temperatures or preferably at temperatures equivalent to about two thirds of the absolute melting point of the ductile metal. In some instances, it is desirable to heat the product during this pressing operation to temperatures just slightly below the melting point. However, it is preferred that the metal be not melted.

The powder can be hydraulically pressed, hydrostatically pressed, or loaded in an evacuated container and subjected to hot forging.

A preferred combination for compacting is a refractoryfilled, high-melting metal powder blended with a powdered, low melting metal. With such a mixture, by hot pressing at a pressure in excess of the tensile strength of the low melting metal, dense compacts can be formed. This compaction is accomplished before homogenization of the metal fractions has occurred.

During compaction the metal powders should not be exposed to oxidizing conditions. In handling high surface area powders, it is preferred that the powders be completely free of exposure to oxygen until they have been compacted to maximum density.

STNTERING THE COMPACT The green compact formed as just described can be sintered, as at temperatures up to of its melting point for up to twenty-four hours, to give it suflicient strength to hold together during subsequent working operations. Preferably, the sintering is effected in an inert atmosphere, such as argon or helium, or in a reducing atmosphere such as clean, dry hydrogen. If the green compact has sufiicient strength, the sintering step can be omitted, and an annealing step substituted after working.

WORKING THE COMPACT To obtain metal products of maximum density, and to achieve maximum bonding of the metal grains, the compacted body is subjected to intensive working, preferably at elevated temperatures. The working forces should be sufficient to effect plastic flow in the metals. Working should be continued until welding of the filler-metal grains and ductile metal grains is substantially complete.

While working can be accomplished by such methods as swaging, forging, and rolling, it is especially preferred to effect working by extruding the above-mentioned compact through a die under extreme pressure. After extrusion, the product can, if desired, be further worked by swaging, forging or rolling.

Whichever method of working is selected, it is preferred that exposure to oxygen, nitrogen and water be avoided during the working step. Even small amounts of oxygen will form oxide skins on the metal powders. These skins increase the difficulty of welding the powders, hence their formation is preferably avoided.

Because the refractory-filled powders are very hard, the forces required for working are higher than for unmodified metal. In the case of extrusion of a billet, the reduction in cross-sectional area preferably is upwards of 90%. Welding of the metal grains thereupon becomes nearly complete.

Now while the worked products have greatly increased strength and usually have increased hardness, they nevertheless are relatively ductile. It is believed that the increased strength is due to the volume of hard or reinforced metal containing the filler, in which the metal grain size is less than 5 microns in average diameter and usually of the order of l to 2 microns, whereas the ductility is due to other volumes of metal, in which the grain size is of the order of 10 to 100 microns or even larger. The ratio of the volumes of small to large grains is approximately equivalent to the ratio of filled to unfilled metal used in any given preparation; therefore, control of these ratios gives control of ductility in the final product.

Because of the ductile phase, the product after working may have a pattern or structure in the direction of working, as shown in FIGURE 5. This can be minimized or eliminated if desired by working the metal in several directions.

The final compacted and sintered mass of metal to be used as a material of construction should have a density upwards of 95% of theoretical, preferably upwards of 99%.

ANNEALING THE SOLID METAL PRODUCTS Although as above noted the solid metal products of this invention are heterogeneous in the sense that the metal grains in contact with the refractory filler particles are much smaller than the metal grains not in such contact, it is not desired that the metal originally associated with the filler remain separate and distinct from the metal derived from the unfilled metal powder. On the contrary, it is desirable to have chemical homogenization of these two metal phases, so that all metal areas present are substantially the same except in regard to grain size.

The desired chemical homogenization is accomplished by annealing the compacted powder blend. This step can be carried out before, after, or simultaneously with the working step above described, in the latter case by suitably selecting the temperature of hot-working. The annealing causes diffusion of the metal phases into each other, so that by chemical analysis, the metal is shown to be of uniform composition throughout the mass of the solid, annealed product. For instance, if the filled metal is nickel and the unfilled metal powder is chromium, the metal in the product after annealing is completed will be a uniform mixture, or alloy, of nickel and chromium. The size of the alloy grains in contact with the filler particles will be smaller than the size of the alloy grains not in such contact.

The time and temperature of annealing will, of course, depend upon such factors as the metals used, their proportion, and the proportion of refractory particle in the metal. Since chemical and metallographic techniques are available for determining the extent to which homogenization has taken place during annealing, those skilled in the art will have no difficulty selecting suitable conditions for any particular system. In general, annealing can be done at temperatures in the range of from 0.75 to 0.9 of the melting point in degrees absolute.

THE NOVEL COMPOSITIONS The novel compositions of the present invention include the powder products comprising mixtures of refractoryfilled metal particles and unfilled metal particles, and also the solid metals of island structure which can be made by solidifying such metal powders. The latter metal products are ductile, resistant to creep at elevated temperature, and resistant to oxidation.

In the product characterizations set forth herein, descriptions will sometimes be given with particular reference to single-metal compositions of the invention, but it will be apparent that the characterizations can also be applied to alloy products.

The filler particles present in the filled metal component of products of this invention are dispersed throughout said component. This dispersion can be demonstrated using the electron microscope and replica techniques wherein the surface of a metal piece is polished, a carbon layer is deposited on the polished surface, and the metal is removed, as by dissolving in acid. An electronmicrograph of the remaining carbon film shows that the f4 filler particles are uniformly distributed throughout the metal grains.

By uniformly dispersed is meant that there is uniform distribution of the refractory oxide particles within any single selected microscopic region of treated metal, such regions being about 10 microns in diameter. Since treated metal powder is blended with untreated metal powder and the mixture is compacted and sintered to dense metal, it is obvious that only those regions originating from particles of treated metal powder will contain a uniform distribution of the filler in the metal matrix.

The filler particles in compositions of the invention must be in the size range of 5 to 1000 millimicrons, preferably from 5 to 500 millimicrons, and still more preferably from 10 to 250 millimicrons. The latter class is particularly preferred, since the 10 millimicron particles are considerably more resistant to coagulation or gelling than are smaller particles, and thus are easier to maintain in a dispersed state during processes of the invention than smaller particles. Thus, products containing filler particles in the size range of from 10 to 250 millimicrons can be readily produced according to processes of this invention from colloidal dispersions.

In describing products of this invention an oxide filler particle is defined as a single coherent mass of oxide surrounded by metal and separated from other oxide mass by metal. The particles may be aggregates of smaller ultimate units which are joined together to form a structure, but, of course, the size of the aggregate must be in the range of 5 to 1000 millimicrons.

The particles of the filler in the filled-metal component are substantially completely surrounded by a metal coating which maintains them separate and discrete. The particles are thus isolated, and do not come in contact one with another; thus, coalescence and sintering of the filler material is prevented. In other words, the filledmetal component comprises a continuous phase of metal containing, dispersed therein, the refractory filler particles.

This aspect of the invention is illustrated by FIGURE 6 of the drawings, wherein the area encompassed by line 1 represents a refractory-filled metal powder particle consisting of metal grains 2 in which there are dispersed the particles of refractory filler 3. It will be understood that the lines 6 within the powder particle represent the grain boundaries of the metal. FIGURE 6 illustrates a powder product of the invention in that there are also present unfilled, powdered metal particles, shown as encompassed by line 4 and made up of metal grains 5 in which there are no dispersed refractory particles.

Metal compositions in which the filler is thoria, a rare earth oxide, or a mixture of oxides of the rare earth elements of the lanthanum and actinium series, magnesium oxide, or, to a lesser extent, calcium silicate, have exceptional stability in elevated-temperature, long-com tinned tests such as stress rupture and creep tests. These materials maintain their properties to a considerably greater extent than metals filled with silica, for example, even when the initial hardness obtained during the processing operation is similar. The reason for this improvement is related to the free energy of formation of the filler. For this reason, preferred compositions of the invention for use at very high temperatures, i.e., above 800 to 1000 C., comprise a dispersion, in a metal, of oxide particles having a free energy of formation as determined at 1000 C. per gram atom of oxygen atom in the oxide, of from to 123 kcal. and preferably from to 123 kcal.

Actually, silica is a highly efficient filler for metal compositions which do not need to be heated above 600 to 700 C. during processing or use. In the case of iron-molybdenum or nickel-molybdenum alloys, which are made by blending molybdenum powder with powders of modified iron or modified nickel, temperatures as 15 high as 1300 C. or slightly higher are often encountered during processing. In these cases, only the very stable oxides are effective as fillers, i.e., those with a very high free energy of formation, such as the rare earth oxides or calcia.

This correlation between high-temperature service and the free energy of formation of the refractory filler can be generalized in terms of the melting point of the metal or alloy involved, since for high-temperature service metals of high melting point are used. Thus, it is a limitation upon the compositions of this invention that the free energy of formation of the particulate refractory (as measured at 1000 C.) be at least 52+.016M, where M is the melting point in degrees Kelvin of the metal or alloy phase of the solid product.

Products of the invention can be characterized by the distance between the filler particles in those metal volumes in which such particles are present. This distance is a variable which depends on both volume loading and particle size. If the dispersed phase is a material of uniform particle size and is dispersed homogeneously in a cubic packing pattern, the following expression relates the interparticle distance, i.e., the edge-to-edge distance Y, to the particle diameter d and the volume fraction of the dispersed phase f:

itts-fl For products of this invention, the interparticle distance as calculated by this expression is less than 1.0 micron and preferably from 0.01 to 0.5 microns (10 to 500 millimicrons). In the most preferred products this range is 50 to 250 millimicrons, in those metal volumes in which such filler is present.

The finely divided filler particles in compositions of the invention causes the grain size of the metal in the vicinity of the filler to be much smaller than normally found. This small grain size persists even after annealing at temperatures in degrees absolute up to 0.8 times that of the melting point of the products. A grain size below 10 microns, and even below 2 microns is common for the products of this invention. Products which have filler particles in contact with metal grains in the size range below 10 microns are preferred.

The solid metal products obtained by compacting, sintering and suitably working the blended powders as above described are a particularly preferred embodiment of this invention. In the drawings, FIGURE 7 represents the kind of structure existing in the pressed compact, wherein there are volumes or islands of filled metal 1 containing filler particles 3, these islands being interspersed throughout other volumes of unfilled metal particles 4. It will be seen that there are still voids '7 between the filled and unfilled metal particles; in other words, theoretical density is not achieved or closely approached merely by compaction. Further working is required.

FIGURE is a line drawing prepared from an electronmicrograph showing the island structure in a solid product of the invention which has been hot-worked by extrusion to a density approaching theoreticalthat is, above about 90% of theoretical. FIGURE 4 is similarly a line drawing made from an electronmicrograph, but in this figure the solid product had not yet been extruded. It will be seen that in FIGURE 4 there is a region of filled metal, B, in which the grain bounderies are very close together, i.e., the metal grains are very small, and a region of unfilled metal, A, in which the grain boundaries are relatively far apart and the grains are relatively large. In FIGURE 5 regions A and B are also present, but the former appears as islands in the latter.

Actually, while the term island structure is useful to convey the concept of heterogeneous character present in solid products of this invention, it can be a misnomer in some instances. Thus, while there will always be areas (or rather, volumes) of refractory-filled metal intermingled with areas (or volumes) of unfilled metal, it is immaterial which is the island and which the matrix. This will depend upon which component is present in predominant proportion, the minor component usually being the islands, although as seen from a comparison of FIGURES 4 and 5 the electronmicrographs may seem to indicate one structure before extrusion and the other after extrusion, even on the same sample of product.

Because the products of the invention have volumes of metal containing no filler, they are more ductile than those compositions in which the filler is uniformly dispersed throughout the composition. In those volumes in which there are no filler particles, the metal grains grow to a much larger size, i.e., 40 microns and larger.

The size and shape of the filled and unfilled areas may vary over wide limits. The characteristics of size and shape are a result of the size and shape of the metal powders from which the structure was prepared, as well as the compacting, sintering, working and annealing steps used in the preparation.

The stress which the modified metals and alloys of the invention will support over a period of time at high service temperatures is at least two to ten times larger than that of the unmodified metal and alloys. The resistance of the filler-modified metals and alloys to long-term deformation under relatively low stress may be as much as ten thousand times better than the corresponding unmodified alloys. For instance, the stress for 100 hour rupture life of nickel when modified with oxide filler as herein described, is improved at least twenty-fold when measured at 1800 F. Not only are the products strong, but they are ductile, readily machinable, and show considerable elongation under stress, up to 90% of that of unmodified control.

By incorporating dispersed refractory particles into metal mix ures according to the invention the yield strength of the mixtures is quantitatively improved while the ductility of the mixtures, as measured by the elongation, remains adequate for practical purposes. If Ym is the yield strength of the modified material at 0.2% off set and Ye is the corresponding yield strength of the control, the following relationship holds at temperatures in the range from 50 to of the melting point of the metal mixtures in degrees absolute:

A preferred class of the novel products consists of high-melting compositions, particularly those containing at least one of the metals from the group consisting of iron, cobalt, nickel, molybdenum, and tungsten, together with a metal from the group consisting of chromium, titanium, and niobium. Of this group alloy compositions having a melting point above 1200 C. are particularly advantageous.

A specifically preferred class of the novel products consists of alloys containing chromium. These alloys are particularly oxidation resistant. Because they have high-temperature strength by reason of the inclusion of the refractory oxide filler, they are useful at elevated temperatures, for instance, in the range of 1200 to 1800 F. and in some cases even higher. Stainless steel alloys are included in this preferred class. They can be prepared from nickel-iron masterbatches containing refractory oxide fillers such as thoria, by a process in which the masterbatch is blended with powdered chromium. Alternatively, a masterbatch of alumina particles in iron can be blended with chromium, nickel and iron powders, or a masterbatch of rare earth oxide particles can be blended with iron-nickel powder. In a similar manner, one can make other alloys of chromium such as Nichrome (80 Ni-20 Cr), iron-chromium, (73 Fe-27 Cr), and iron-nickel-cobalt-chromium alloys containing, for

=greater than 1.5:1

example, up to 30% chromium. Iron, nickel or cobalt base alloys containing from 10 to 25% chromium are a preferred group. Specifically, such alloys containing 90 to 50% of the sum of iron, cobalt and nickel, to 20% of the sum of molybdenum and tungsten and 0 to aluminum, titanium, manganese, silicon and niobium, along with to 25% chromium are an especially preferred species.

In the above-mentioned chromium alloys, and other high-temperature alloys, it is preferred to use very stable refractory oxide fillers, that is, those with a high free energy of formation such as beryllia, calcia, thoria, and rare earth oxides, the filler having a free energy of formation, measured at 1000 C., in the range above 115 kilocalories per gram atom of oxygen in the oxide. Oxides having a free energy of formation at 1000 C. of up to 123 are presently available, and if more stable oxirlles could be prepared, they would be in the preferred c ass.

An especially preferred class of the novel products consists of alloys containing metals having high melting points, such as niobium, tantalum, molybdenum, or tungsten, or two or more of these metals. Molybdenum and tungsten have extremely high melting points and their presence raises the melting point of the alloy products formed. Since molybdenum or tungsten by themselves are not oxidation resistant, these metals are not ordinarily used alone, but they are useful in alloys with other metals. Thus, especially preferred are alloys of these relatively high-melting metals with other metals such as nickel, iron, cobalt, chromium, titanium, zirconium, niobium, aluminum, and silicon. Thus, this preferred group includes such alloys as high-molybdenum steel, nickel-molybdenum steel, molybdenum-iron-nickel alloys, tungsten-chromium and molybdenum-chromium alloys. Within this class, also, are alloys of molybdenum or tungsten with niobium or titanium, or with both niobium and titanium. Molybdenum-titanium alloys, containing from 10 to 90% titanium, are included in this group, as are molybdenum-niobium and tungsten-niobium alloys. The latter alloys can be conveniently prepared by the powder-blending process above described, using a molybdenum-filler masterbatch blended with niobium metal powder.

Still another preferred class of the novel products consists of compositions containing aluminum. Aluminum forms intermetallic compounds, which are light in Weight and oxidation resistant. To make a product of this type one can, for example, add a lanthana-nickel masterbatch to powdered aluminum, thereby obtaining aluminumnickel-lanthana compositions. Similarly, one can prepare aluminum-copper alloys, aluminum-nickel-cobalt alloys, aluminum-iron alloys, and alloys containing both aluminum and molybdenum.

Compositions of this invention are especially useful for fabrication into components which must maintain dimentional stability under heavy stress at high temperatures, such as turbine blades. By high temperatures is meant temperatures in the range from 0.5 to 0.8 times the melting temperature, in degrees absolute, of the metal in the composition.

Examples The invention will be better understood by reference to the following illustrative examples.

Example 1 A solution of nickel nitrate was prepared by dissolving 4362 grams of nickel nitrate hydrate Ni(NO .6H O in water and diluting this to 5 liters. A thoria sol was prepared by dispersing calcined Th(C 0 in water containing a trace of nitric acid. The thoria in this sol consisted of substantially discrete particles having an average diameter of about 5 to 10 millimicrons. This thoria was used as the source of the filler material. A 288- gram portion of this colloidal aquasol (26% ThO was diluted to 5 liters. To aheel containing 5 liters of water at room temperature, the solution of nickel nitrate, the diluted thoria sol, and ammonium hydroxide-ammonium carbonate solution were added as separate solutions simultaneously, and at uniform rates, while maintaining very vigorous agitation. This was accomplished by using the apparatus shown in FIGURE 1. The valve to the filter was closed; the heel was charged into the tank; the pump was turned on; the feed streams were added as indicated. During the precipitation, the pH in the reactor was maintained at 7.5. A coating of nickel hydroxide-carbonate was thus deposited around the thoria particles. The resulting mixture was filtered, and washed to remove the ammonium nitrate. The filter cake was dried in an oven at 300 C.

The product obtained was pulverized with a hammermill to pass 325 mesh, placed in a furnace, FIGURE 2, and heated to a temperature of 500 C. Hydrogen was slowly passed over the powder at such a rate that sufficient hydrogen was added to the nickel oxide to reduce it in a period of four hours. The flow of hydrogen was maintained at a steady, uniform rate during this reduction procedure for eight hours. Thereafter, the temperature was raised to 700 C. and the flow of dry, pure hydrogen was greatly increased, and finally the temperature was raised to 1050 C. to complete the reduction, and sinter the reduced powder.

The resulting powder had a surface area of 4 m. g. and a bulk density of 2.3 grams per milliliter. The powder contained 10% ThO- by volume. When compacted and annealed, this powder had a Rockwell A hardness at 25 C. of '66.

Two parts of the thoria-nickel powder were blended with three parts of carbonyl nickel. This latter powder was about 5 to 9 microns in size. The carbonyl nickel, when compacted and annealed, had a Rockwell A hardness of 26. The blended powder was then pressed hydraulically at 30 tons per square inch to a billet 1 inch in diameter and 2 inches long.

The billet was next sintered in hydrogen (dew point 50 C.) for twenty hours at 550 C. and five hours at 1200 C. The nickel oxide content of the sintered billet was less than 0.01%

The sintered billet was then heated to 2200 F., dropped into a container at 1100 F., and then extruded from the container through a die having a 90 throat, to a 4-inch rod. Thus, sintering and hot-working were carried out at temperatures high enough to achieve annealing too. The rod was tested as follows: Ultimate tensile strength at 1800 F. was 16,700 p.s.i. and 0.2% offset yield strength was 16,500 p.s.i. The elongation Was 7%. A control sample of nickel made in the same way, but without the filler, had a yield strength of 1,400 p.s.i. The improvement in yield strength at 1800 F. is thus a factor of twelve-fold for the filled nickel over the control sample.

The Rockwell A hardness of the sample was 51. This hardness did not change on annealing for four hours at 2200 F.

Another improvement by which the product of this example is characterized is that of oxidation resistance. The oxidation rate at 1800 F. in air, as measured by gain in weight, is slower for the 4% thoria-nickel sample of this example than for a wrought nickel, unmodified control. Specifically, the oxidation rate, as measured by Weight gain per unit surface area on heating in air at 2200 F. was about equivalent to the rate of oxidation of unmodified nickel at 1500" F. Specifically, after 4 cycles to temperature and a total time at temperature (2200 F.) of forty hours, the sample showed a total weight gain of only 2% The oxidation rate of the nickel thoria is about equal to the oxidation rate of wrought nichrome Ni-20Cr).

The stress-rupture properties also indicate stability. For example, a sample of 4% thoria in nickel of this ex- 19 ample can withstand 7,000 p.s.i. at 1800 F. for more than one hundred hours without rupture.

An electronmicrograph picture was prepared to show the distribution of thoria in the thoria-nickel sample. The micrograph showed that there were regions in which there was a homogeneous distribution of thoria in the nickel. There were also regions in which no thoria Was seen. These areas correspond to the unmodified nickel which was used to blend with the thoria-nickel. Ductility and machinability are better for samples having such regions or islands of unmodified metal.

The electronmicrographs were prepared as follows: A fit-inch rod of nickel containing dispersed thoria was cut and the cross section was mounted in Bakelite and mechanically polished. The polished surface was cleaned and dried in ethyl alcohol. The samples were removed from the Bakelite and placed in a high vacuum furnace and a vacuum of 10 mm. of Hg at 1000 C. was reached. After thermal etching for about three hours, the sample was removed and placed in a vacuum evaporator. Two carbon rods were brought together within the evaporator and current applied until sputtering occurred. A very thin film of carbon Was deposited upon the etched surface as the sputtering occurred.

The carbon-covered surface was scribed into inch square with the use of a sharp cutting blade.

Next the sample was placed in a culture dish containing a 1% solution of bromine. The carbon squares were freed from the surface of the metal by chemical attack. They floated to the surface of the solution, were picked up on electronrnicroscope screens (250- mesh S/S wire), and viewed in a Philips EM 100 threephase electronmicroscope. Alternatively samples can be chemically etched, or viewed as polished.

The solution of bromine was used to remove the carbon because it would attack the base metal and not do damage to the oxide, or the carbon replica.

All samples were photographed in the electronmicroscope at a film magnification of 1,250 and 5,000 respectively. Prints at 5,000X were made from the 1,250 negative and at 20,000X from the 5,000X negative.

The presence of .grain boundaries (lines) in the 20,000 picture was plainly observable. In the areas where filler particles were present, these grains averaged about 2 microns in size. The ThO particles were about 0.1 micron in size. In the areas where there was no filler, the grain size was in the range of 50 microns, and greater, or about twenty-five-fold that of the filled grains.

There are several variables which affect the product quality as measured in terms of yield strength at 1800 F. for products of the type described above, in Example 1. The following'illustrate this: (a) A masterbatch of 30% was superior to one containing thoria. The former, for example, when blended with nickel powder 1:1 gave a product having yield strength of 24,000 p.s.i. and 1000-hour rupture strength of 8,000 p.s.i. at 1800F. (b) Preparations made from concentratcd nickel nitrate solutions are somewhat better than those made from more dilute'nickel nitrate. (c) A blending ratio of masterbatch to unfilled nickel in the range from 1:2 to 1:1 was slightly superior to a blending ratio of either 7:3 or 114.

In addition, it is preferred to sinter the compacted billet in a reducing atmosphere (such as hydrogen) rather than an inert atmosphere (such as argon), and to extrude at as low a temperature as possible. Extrusions have been made by heating the billet to 1800" F.

Example 2 This example is similar to Example 1, except that in this case, the final stages of the reduction and sintering were carried out at 950 C. The resulting nickel-thoria powder was blended with unmodified nickel powder 20 (carbonyl grade, less than 325 mesh) to give products containing 2, 4 and 7% T110 From these and the starting material (10% ThO Ar-inch rods were prepared.

These rod products were characterized as follows: After heating to 1200" C, Rockwell A hardness for the sample containing 2% ThO was 44, for the sample containing 4% ThO was 51, for the sample containing 7% ThO was 58 and for the sample containing 10% T110 was 66. The sample containing 4% ThO had a yield strength of 16,500 p.s.i. at 1500 F.; 13,000 p.s.i. at 1800 F.; 18,300 p.s.i. at 1200 F. Thus, the decrease in yield strength at 1800 F. is only 21% below the yield strength at 1500 F. This is a major improvement in the characteristic behavior of the metal composition, since nickel alone and most of the nickel base alloys used in high-temperature applications show a much larger percentage decrease in yield strength over this temperature-range. Pure nickel alone has essentially no strength at 1800 F.

A comparison of the yield strength and elongation of these blended nickel structures with unmodified nickel (prepared from the same unmodified carbonyl nickel powder which was used to blend with the thoria-nickel powder in the process and products of this example) and with nickel-thoria of uniform structure (10% thoria) is shown below:

Ym=yield strength at 1800 F. of thoria-filled samples.

Yc=yield strength at 1800 F. of control, or unmodified nickel.

Percent elongation measured during yield strength test:

Eb=elongation of blended l'llCkGl'l10 112. product. Ec=Elongation of control or unmod fied n ckel. Eu=elongation of uniform 10% thoria in nickel product.

It is evident from the above comparison that the products of the invention were more ductile, i.e., they had more elongation at 1800F. than thoria-modified nickel in which the thoria was uniformly distributed; this can be seen in the ratio Eb/Eu. In this particular series, the elongation of the blended structures was approximately twoto four-fold that of the uniform structure. The elongations of the products of the invention in this example were only reduced from about 0.5 to 0.8 of that of the unmodified metal, i.e.,Eb/Ec was in that range. Also, the yield strength of all the thoriamodificd products was several-fold greater than that of unmodified metal (ratio Ym/Yc).

Another improvement by which products of this example are characterized is that of oxidation resistance. The oxidation rate at 1000 C. in air, as measured by gain in weight, is slower for the 4% thoria-nickel sample than for a wrought nickel, unmodified control. Specifically the oxidation rate, as measured by weight gain per unit surface area on heating in air at 2200 F. is about equivalent to the rate of oxidation of unmodified nickel at 1500 F. In fact, the oxidation rate of the nickelthoria is about equal to the oxidation rate of wrought nichrome Ni-20 Cr). The stress-rupture properties also indicate stability. For example, a sample of 4% thoria in nickel can withstand 11,000 p.s.i. at 1500 F. for more than 1000 hours without rupture. At 1800 F. this sample can support 4,500 p.s.i. for more than 1000 hours without rupture. The stress-rupture curve measured at 1800 F. of this nickel-thoria sample has a flatter slope than wrought Inconel or Hastelloy X, the latter being measured at a lower temperature, namely, 1500 F.

An electronmicrograph picture showing the distribution 20,000 p.s.i.

21 ofthoria in a thoria-nickel sample as prepared above (contaming 4 volume percent thoria made by dilution of 4 parts of 10 volume percent thoria in nickel with pure nickel) shows that there are regions in which there is a homogeneous distribution of thoria in the nickel. There are also regions in which no thoria can be seen. These areas correspond to the unmodified nickel which was used to blend with the thoria-nickel. Ductility and machinability are better for samples having regions or islands of unmodified metal.

Example 3 A masterbatch of cobalt-thoria was prepared from (a) 4370 grams Co(NO .6H O in liters H O, (b) 532 grams of 20.7% ThO sol diluted to 4 liters, and (e) 25% (NH CO solution. This masterbatch was used for diluting with unmodified cobalt powder according to the details of Example 1.

It will be understood that the processes similar to those of Examples 1 and 2 can be applied to other pure metals, including copper, iron and chromium.

Example 4 A nickel-thoria masterbatch, containing 30 volume percent thoria, was prepared according to the process of Example 1, the only diiferences being that three times as much thoria was used, and the final temperature of the reduction was 1l30 C. This powder was very hard. In a similar manner, a nickel-zirconia powder, containing 0.5 volume percent zirconia, was prepared. The zirconia aquasol used for this latter preparation contained ZrO particles which were about millimicrons in diameter. The zirconia sol at 10% solids had a relative viscosity vs. water of 1.4. It was prepared by autoclaving 1 molar Zr'O(NO solution at 200 C. and peptizing the resultant precipitate in distilled water. The nickel oxide zirconia composition was reduced at 650 C. The resulting relatively soft powder had a Rockwell B hardness of 67.

Equal parts of the two reduced nickel-filler powders were blended. The blend was hydraulically pressed at 30 t.s.i. to form a billet 1 inch in diameter, and thereafter sintered in dry, pure hydrogen at 1200 C. This billet was then extruded to a rod of fit-inch diameter. This rod had a yield strength at 1800" F. of greater than FIGURE 4 is a micrograph of the annealed metal product, showing the relative grain sizes in volumes where there is thoria as contrasted to zirconia. The grains in the volumes containing the thoria were less than 1 micron in size, whereas the grains in the volumes containing the zirconia were about 50 microns in size. The ratio of grainsize in the two volumes was thus greater than 50: 1.

In compositions of the type described in Example 4, the nickel-zirconia is regarded as the ductile phase. In such instances, it is preferred that, as calculated from the expression where G is the grain size in microns, d is the particle diameter in microns, and f is the volume fraction of the filler, the grain size in the ductile phase being greater than 10 microns. More preferably, the calculated grain size should be in the range from 40 to 150 microns. It will be understood that this grain size is calculated from the particle diameter of the filler in the metal product, and not from the particle diameter of the filler used in the preparation.

The average diameter of thoria particles which was processed at a maximum temperature of 1200 C. (during sintering) as in this example, is 0.1 micron. Thus, if thoria is used in making the ductile phase, for compositions of the type above described, a volume loading from 0.09 to 0.3% is preferred. However, loadings somewhat above or below this preferred range can be used to advantage.

hence a larger amount of filler may be used in the ductile phase.

Example 5 A nickel-thoria masterbatch powder containing 10 Volume percent thoria was prepared as in Example 1. This powder was passed through a 325-mesh screen. It was then placed in a second furnace, FIGURE 3, and treated with hydrogen until the dew point of the efliuent gas was below 70 F. The furnace assembly was then transferred to a dry box, filled with argon. The powder was removed from the furnace in a completely inert atmosphere, i.e., the oxygen level was about 5 parts per million. Nitrogen and water vapor in the dry box were also correspondingly low.

While in the dry box, the powder was blended with nickel and chromium powders as follows: 44 parts of nickel-thoria by weight with 40 parts nickel and 20 parts chromium. The oxygen analysis of the nickel powder was 0.00%. The oxygen analysis of the chromium powder was less than 0.01%. The nickel powder was 10 microns in size and the chromium powder had a dendritic structure and was less than mesh, in the largest dimension. Both of the unmodified metal powders were softer and more ductile than the thoria-nickel powder.

The blended powder was then hot pressed to a billet of 100% density, the temperature of the containerand powder during pressing being 1100 F. Only after this pressing was the billet exposed to air (oxygen). The billet was 1 inch in, diameter and 2 inches long. It was extruded to A-inch rod, as in Example 1.

After extrusion, the rod was annealed in pure, dry hydrogen at 1325 C. for fifty hours. Alternatively, the billet can be sintered prior to extrusion. The extruded, annealed rod can support 8000 p.s.i. for over 100 hours at 1800 F. The extruded, annealed rod contained grains in contact with thoria particles which were about 2 microns in size and about an equal volume of grains, not in contact with thoria, which were about 100 microns in size.

The stress-rupture plot at 1800 F. of the product of Example 5 is flatter than the stress-rupture plot of the nickel-thoria product of Example '1.

Example 5 illustrates the care which was taken to exclude oxygen contamination of the powders used to prepare the solid, wrought metal product. It is preferred that similar care be exercised in the case of compositions including alloys of niobium, titanium, silicon, aluminum, and other similar metals which form stable oxides.

Instead of hot pressing, the blended powder can be hydrostatically compaced, as, for example, with pressures in the range from 20,000 to 200,000 p.s.i. In this case the compact should preferably only be exposed to air after sintering. In this case, sintering at 1325 C. can be done prior to extrusion and the annealing after extrusion eliminated. Alternatively, the mixed powders can be canned, extruded, and finally annealed.

Example 6 This example describes the utility of a chromium-thoria m-asterbatch in making compositions of the invention.

The reactor used to deposit the chromium oxycarbonate on the colloidal thoria consisted of a stainless-steel tank with a conical bottom, shown in FIGURE 1. The bottom of the tank was attached to a stainless-steel circulating line, to which there were attached three inlet pipes through Tts. The circulating line passed through a centrifugal pump and thence returned to the tank.

Initially, the tank was charged with 10 gallons of water, which was about /5 of the capacity of the tank. Three feed solutions were prepared as follows: (a) 40.6 pounds of chromium nitrate, Or(NO .9H O, dissolved in 5 gal- 3, 1 aoAas lons of water, (b) 30 pounds of ammonium carbonate: dissolved in gallons of water, and (c) 20 liters of a colloidal aquasol containing 3% thoria, in the form of 5 to millimicron particles, diluted to 5 gallons with water.. These three feed solutions were metered in through call-- brated liquid flow-meters at equal rates into the circulating stream, which initially consisted of water. The pH of the slurry in the tank was maintained between 7.0 and 8.0 during the run, and was 7.6 at the end of the run. The time of addition of the reactants was 40 minutes, the reactants being added at room temperature.

The resulting slurry contained precipitated particles which consisted of hydrous, chromium oxycarbonate and colloidal thoria. This precipitate was filtered, and washed with water to remove most of the soluble salt. It was then dried for forty hours at 250 C. and micropulverized, to give a product which passed 100 mesh. The analysis of the product was 18.1% ThO and 78.9% Cr O X-ray line-broadening studies showed that the chromia coating consisting of crystallites about 1 micron in size. The thoria particles entrapped in this chromia coating had an average size of about 40 millimicrons.

The product thus obtained was reduced by treating it with pure, dry hydrogen at 1300 C., until the dew point of efiluent hydrogen was below -70 C. The reduced, hard, Cr-Th0 powder was then blended with a relatively soft nickel powder containing 0.3 volume percent thoria in such a ratio as to yield an alloy of 80% nickel, chromium. The blended powder was hot pressed to a dense billet, completely free of voids. During the entire handling of the powders, oxygen, nitrogen and water vapor were completely excluded.

Thebillet was sintered at 1325 C. for forty hours in pure, dry hydrogen, after which time the nickel and chromium were homogenized. The billet was then ex truded from l-inch to A-inch diameter. The product had a yield strength at 1800 F. of 15,000 p.s.i. and an elongation at that temperature of 23%. Thus, the yield strength at 1800 F. was double that of a control prepared from metals fabricated as indicated but without oxide filler, and elongation was 77% that of control. The structure of the annealed product consisted of grains about 2 microns in size in 20% of the metal volume and grains about 80 microns in size in the remaining volumes.

Examnple 7 Thisex-ample shows an alternative process for preparing products of the invention. In this process, the powders are blended prior to reduction.

A nickel oxide-alumina was prepared according to the processof Example 1, by using an alumina sol prepared by dispersing Alon C alumina powder in a dilute acid solution. :In a similar manner, a nickel-oxide powder was prepared, without any filler. Both powders were dried at 450 C. and micro-pulverized to .100 mesh.

Equal weights of the two powders were blended. The blend was then reduced in extra dry, pure'hydrogen at 550 C. This powder was extremely pyrophoric, hence it was not exposed to oxygen.

The reduced powder was handled in an inert atmosphere (argon) in which it was hot pressed, and the pressed billet was then sintered in hydrogen at 1250 C. for twenty-four hours. The ,sintered billet ,had a density equivalentto theoretical. Only after sintering was it exposed to air.

This. billet was extruded from 1 inch to inch, whereupon a nickel-alumina product containing 6 volume percent Al O having superior high-temperature properties was produced.

Example 8 A masterbatch of Cu-% (volume) A1 0 was prepared from Cu(NO solution and A1 0 aquasol, using the precipitation technique of the previous examples. An

24 improvedmonel was prepared by blending 52.5 parts by weight of this masterbatch (-325 mesh) with 122.5 parts by weight of pure nickel powder (8 micron).

The blend was hot pressed at 600 C. and 25 t.s.i. to a billetof greater than 98% density. This billet was sintered at 1100 C. and extruded. The product had improved resistance to creep at 1700 F.

By blending the above masterbatch with zinc powder, one can prepare improved brass. Other copper alloys, including bronzes, can be similarly prepared.

Example 9 This example describes a modified stainless-steel composition containing 4% thoria by volume dispersed therein, the thoria being in the form of colloidal particles. The process consists of blending nickel-iron-thoria masterbatch powder with chromium powder, densifying, and homogenizing.

A thoria concentrate in nickel-iron was prepared as :follows: A deposit of iron-nickel, hydrous oxycarbonate was formed on a colloidal thoria filler in a reactor consisting of a stainless-steel tank with a conical bottom, FIGURE 1. The bottom of the tank was attached to stainless-steel piping, to which were attached three inlet tubes through Ts, this circulating line then passed through a centrifugal pump of 20 g.p.m. capacity, thence through a return line to the tank.

Initially, the tank was charged with 2 gallons of water. Equal volumes of three solutions containing the desired quantities of reagents were then added into the middle of the flowing stream through the T tubes. These solutions were added at uniform, equivalent rates over a period of about one-half hour. Through the first T was added a solution of iron nitrate-nickel nitrate prepared by dissolving 2190 grams Fe(NO .9H O and 169 grams Ni(NO .6H O in water and diluting to 3.7 liters. Through the second T was added 3.7 liters of 3.4 molar (NH CO and through the third 3.7 liters of thoria sol made by diluting 60 grams of 36% ThO aquasol with water. The thoria aquasol was highly fluid and contained particles in-the 5 to 10 millimicron size range.

The solutions were added into the reactor simultaneously while the pump was in operation. The rate of addition controlled uniformly by flowmeters. The pH of the solution in the tank was taken at frequent time intervals to insure proper operating, the pH remaining essentially constant during the run and the final pH being 7.7. The slurry was circulated for a few minutes after the addition of the reagents had been completed, and then the solution was pumped into a filter. The precipitate was filtered, washed with water, and dried at a temperature of about 300 C. for twenty-four hours.

The dried product was then pulverized by grinding in a hammermill, and screened to pass 325 mesh.

The pulverized material was placed in a furnace at a temperature of about C. and a mixture of argon and hydrogen was slowly passed over it. This gas stream had previously been carefully freed of oxygen, nitrogen, and moisture. The temperature in the furnace was slowly raised over a period of an hour. The flow of hydrogen was then gradually increased, and the temperature in the furnace was gradually raised until 600 C. was reached, whereupon a large excess of hydrogen was passed over the sample in order to complete the reduction. In this way, the sample was completely reduced.

Finally, the temperature was raised to 850 C., while continuing to pass hydrogen over the sample. In this way, an iron-nickel powder was produced containing 5 volume percent thoria, having a surface area less than 2 m. /g., and being essentially oxygen free, i.e., having an oxygen content, exclusive of the ThO filler, of less than 0.01%.

This powder was used in making a stainless-steel of exceptional high-temperature properties, as follows: 83 parts by weight of the modified iron-nickel powder (less than 200 mesh) was blended with 18 parts of less than 325-mesh ductile chromium powder. The thoria-metal concentrate was protected from the air, as in previous examples, to prevent reoxidation and the chromium powder used was oxygen free. The blend was hot pressed at 60 t.s.i. into a dense billet 1 inch in diameter and about 2 inches long. The billet was then sintered in extremely dry hydrogen at 1300 C. for eighteen hours, the temperature being raised to the maximum level over a period of 6 hours in this process. The hydrogen used had previously been freed of oxygen, water, and nitrogen by passing it through sulfuric acid, then through a commercial dehydrating agent, then through magnesium perchlorate at 95 C., through a Molecular Sieve, also at 95 C., and finally over Ti-Zr-Cr chips at 800 C. The Molecular Sieve used was type 4A, /s-inch pellets, available from Linde Air Products Division of Union Carbide. After sintering, the billet was exposed to air.

The billet was next machined to a diameter of 0.93 inch and hot extruded rapidly to a Ar-inch rod. This rod was a stainless-steel having a Rockwell C hardness of 30, which did not change even after long aging at 1300* C. The resulting product was a dispersion of about 150 millimicron Th0 particles in a stainless-steel matrix.

There were some islands of metal free from thoriar Grains in these islands were greater than 50 microns in size. Examination of the structure showed that the material was completely austenitic (non-magnetic), and that the grain size was about 2 microns in the regions in which filler was present. The extruded rod had a 0.2% offset yield strength of 24,300 p.s.i. at 1500 F.

Example 10 This example describes a masterbatch of chromium and tungsten containing oxide filler. This masterbatch represents a preferred composition, for blending with other powdered metals for preparing products of the invention. Thus, it is particularly useful as alloying agents to add to other powdered metals in preparing super alloys, as, for example, for preparing improved types of S-495, S-588, ATV-3, S-497, S-590, SF816, Refractaloy 70, Refract-alloy 80, M-203, M-204, M205, 25 Ni, Hastelloy C., Thetalloy C., HS23, HS-25, HS31, HS36, X-50, WF-31, I-336, HE-1049 (see Appendix 11 Report on the Elevated-Temperature Properties of Selected Super- Strength Alloys published by the American Society for Testing Materials, STP No. 160).

This reinforced chromium-tungsten alloy is prepared from four starting solutions: (a) 185 grams (NH W O .4H O

dissolved in 10 liters H O, (b) 780 grams CI'(NO3)3.9H2O

in liters H O, (c) 6 grams beryllia as 100 rnillimicron colloidal particles (see Weiser, Inorganic Colloid Chemistry, Volume II, page 177, I. Wiley and Sons, 1935) in 5 liters H 0, and (d) 5 liters of (NI-I CO These solutions are fed into 2 gallons of water in a reactor tank through four T tubes.

The precipitate is filtered, washed with a dilute (0.01%) solution of ammonium carbonate, dried and pulverized. The pulverized powder is then reduced as follows: It is placed in a gas-tight furnace. The furnace is then heated to 250 C. and evacuated. The temperature in the furnace is raised to 500 C. and maintained there for six hours while an excess of extremely pure, dry hydrogen is passed over the powder. Thereafter, the temperature in the furnace is increased at the rate of 25 per hour until 1100 C. is reached. Then the temperature is increased to 1250 C. and held there until the dew point of the effluent hydrogen is below 60 C.

During the reduction the following precautions are observed: Oxygen and nitrogen and their compounds are completely eliminated from the hydrogen feed. The water vapor produced during the reduction is swept out of the furnace by passing hydrogen over the sample at a high flow rate. The temperature in the reducing zone is kept constant, i.e., no variation from one place to another of more than 20.

The resulting chromium-tungsten powder, modified with 10 volume percent beryllia is useful as an alloying agent for blending with other powdered metals, including nickel, cobalt, iron, titanium, aluminum, niobium, chromium, tungsten and combinations of these.

The powder is characterized by having a uniform distribution of 200 millimicrons beryllia particles throughout the metal. Thus, the average interparticle distance is 140 millimicrons. The grain size of the metal is small, of the order of 1 to 5 microns. This grain size does not change appreciably in size, even on prolonged aging at 1200 C.

By increasing the amount of beryllia, one can prepare metal products containing 20, 30, or more percent of oxide. By changing the ratio of Cr(NO .9H O and (NH W O .4H O, one can vary the ratio of Cr to W in the product. (Such a change may require an adjustment in the amount of (NH CO used.) By substituting (NHQ M O AH O for the tungsten compound, one can prepare alloys of chromium and molybdenum. Further, one can substitute colloidal T110 A1 0 La O 0 Ce O Sm O S0 0 other of the rare earth oxides, including mixtures of rare earth oxides, in place of the colloidal beryllia, and thus prepare Cr-W alloys containing any of these other oxides.

In such alloys it is preferred to use oxides in the form of particles less than 250 millimicrons in size, on the average, and the oxides which have a high free energy of formation. Also, it is preferred to use reduction temperatures as low as is possible, commensurate with completely reducing the Cr to Cr.

Using the technique above described, a Cr-W-Al O masterbatch containing 1.37 parts Cr and 1 part W by weight and 15% A1 0 by volume is prepared. An alloy is prepared by blending 38 parts of this masterbatch with 10 parts of nickel and 51 parts of cobalt. All powders used are oxygen free, except for oxygen in the A1 0 filler. The masterbatch powder was 325 mesh. The nickel and cobalt powders are hot pressed at 25 t.s.i. to a billet of 98% density. This billet is then hot-worked under conditions of plastic flow.

Example 11 A sample of iron powder containing 10 volume percent Al O was prepared similar to the process of Example 1, using a dispersion of A1 0 in dilute HNO in place of the ThO sol and Fe(NO solution in place of Ni(NO .6H O. The A1 0 dispersion was prepared by slurrying a commercial A1 0 powder, Alon C, in very dilute nitric acid, colloid milling and discarding the fraction which settled in a 10-inch container over a period of twenty-four hours.

Example 12 An improved nickel-molybdenum alloy was prepared by blending nickel and iron powder with a masterbatch of molybdenum-thoria.

The masterbatch was prepared by precipitating molybdenum hydroxide in the presence of colloidal thoria and reducing the precipitate with hydrogen. This was done by adding (a) 1 liter of 2M MoCl (b) 1 liter of ThO sol (95 grams 22% ThO sol, prepared by calcining Th(C O diluted to 1 liter) and 0.85 liter M NH OI-I solution to a heel of 1 liter H O.

The precipitate was dried, heated at 450 0., ground to 325 mesh, reduced, and finally sintered in hydrogen at ll50 C.

An alloy was prepared by blending 22 parts of this Mo-ThO masterbatch with 60 parts Ni and parts Fe powder. The latter powders were less than 10 microns average particle size. All powders contained less than 0.01% oxygen, exclusive of the thoria.

Another way of preparing alloys of the type described in this example is to prepare a masterbatch powder of Mo-Fc-Ni-filler and blend this with a metal powder consisting of Mo-Fe-Ni. In this case, the ratio of Mo to Fe to Ni in the masterbatch would be the same as in the unmodified or ductile metal powder. This procedure has the advantage that homogenizing of the metal fraction is accomplished, for example, by chemical means. Thus, lengthy heat treatments of the blended masses can be eliminated.

'The blended powder was hot pressed in an inert atmosphere (pure, dry argon). It is then ready to be rolled, heat treated at an intermediate temperature of about 900 to 1000 C. and rerolled until the reduction in thickness of the initial piece of metal is twenty-fold. The metallic components are finally homogenized by a long anneal at 1325 C.

In a similar way, one can prepare molybdenum-thoria masterbatches, and from these molybdenum-titanium and molybdenum-niobium alloys containing fillers. In the case of these higher melting alloys, the final heat treatment is done at a correspondingly higher temperature.

Eldample 13 This example describes a cobalt-nickel composition modified with 2.5 volume percent thoria, said composition being useful in preparing an improved high-temperature alloy.

The preparation followed the general details as outlined in Example 1 except for the following: The feed solutions consisted of: (a) 1125 grams Co(NO .6H O and 2470 grams Ni(NO .6H O in 5 liters H O, (b) 57.4 grams ThO sol containing 36.4% solids diluted to 5 liters, and (c) 1900 grams (NI-I CO dissolved in H 0 and diluted to 5 liters. Reduction was carried out at 500 to 600 C. and sintering in hydrogen for one-half hour at 850 C.

This powder (325 mesh) of modified nickel-cobalt was then useful for blending with other metal powders.

Example 14 This example describes the preparation of a nickelzirconia (volume) masterbatch which was used for diluting with unmodified nickel powder. This material differs from preparations described in previous examples in that the nickel hydroxycarbonate was first precipitated, filtered, washed, and then reslurried, whereupon colloidal zirconia aquasol was added to yield a zirconia impregnated nickel hydroxycarbonate. After subsequent filtering, drying, reducing, and sintering, the resulting nickelzirconia was blended 1:1 by weight with unmodified nickel powder according to the details of Example 1, to yield a Ni-15% zirconia product.

A solution of nickelous nitrate hydrate was prepared by dissolving 4370 grams Ni(NO .6H O in water and diluting this to 5 liters. To a heel containing 5 liters of water at room temperature, this solution was added simultaneously and at a uniform rate with a 25% ammonium carbonate solution while maintaining vigorous 28 agitation. This was carried out in the apparatus shown in FIGURE 1. During the precipitation, the pH in the reactor was maintained at 7.1. The resulting nickel hydroxycarbonate was filtered and washed repeatedly with water to remove ammonium salts.

One third of the resulting wet cake was reslurried in 3 liters of water solution, the nickel hydroxycarbonate maintained in suspension by vigorous agitation. To this suspension was added over a thirty-minute period three liters of water solution containing 78 grams of dispersed colloidal zirconia, while maintaining vigorous recycling in the apparatus of FIGURE 1. The colloidal zirconia consisted of particles of from 5 to 20 millimicron average particle diameter prepared according to procedures given in copending US. patent application Serial No. 625 188 filed November 29, 1956, now Patent No. 2,984,628 by Guy B. Alexander and John Bugosh. The resulting zirconia-nickel hydroxycarbonate product'was filtered and dried in an oven at 260 C. The dried material was reduced with hydrogen while maintained at 1025 C. to yield a Ni-30% zirconia (volume) product.

One part of the nickel-zirconia powder was blended with one part of a carbonyl nickel, the blended powders compacted according to the procedure of Example 1, and the resulting billet sintered in dry hydrogen for eight hours at 900 C. The sintered billet was then extruded at 1800" F. and the resulting Ni-15% zirconia product had a tensile strength of 26,000 p.s.i. measured at 1500 F.

Example 15 A nickel-rare earth oxide powder containing 10 volume percent didymium oxide was prepared as in Example 1. A didymia sol was prepared by reaction of Na O with anhydrous didymium chloride in molten NaCl-KCl eutectic (50:50 mole percent) at 700 C. The didymia was isolated from the solidified salt cake by leaching away of the salts with water, followed by repeated washing of the didymia until a sol was obtained. The didymia in this sol consisted of substantially discrete particles having an average diameter of about millimicrons. This didymia served as the source of the filler material. A 932-gram portion of this colloidal aquasol containing 70 grams didymia was diluted to 5 liters. To a heel containing 5 liters of water at room temperature, the diluted didymia sol, a solution of 4370 grams Ni(NO .6H O in 5 liters of water, and 25% ammonium carbonate solution were added simultaneously as separate solutions and at uniform rates, while maintaining vigorous agitation by using the apparatus of FIGURE 1. The pH in the reactor was maintained at 7.1 during the precipitation.

After filtration and drying of the cake, the didymianickel hydroxycarbonate was reduced at 950 C. then blended with unmodified nickel powder, and fabricated into a billet as in Example 1. After sintering and hot extruding the billet, the Ni-4% didymia product had a measured tensile strength of 16,000 p.s.i. at 1800 F.

Example 16 A nickel-didymia masterbatch containing 30 volume percent didymia oxide was prepared according to the processes of Example 15, the differences being that three times as much didymia was used, and the final temperature of the reduction was 1100 C. The didymia sol used was obtained by peptizing calcined didymium oxalate in weakly acidic solution. This masterbatch was used for diluting with unmodified nickel powder. The blend ratio was 1 part of masterbatch and 1 part of unmodified nickel powder.

This application is a continuation in part of our copending U.S. application Serial No. 694,086, filed November 4, 1957, as a continuation in part of our then 00- pending but now abandoned US. application Serial No. 657,507, filed May 7, 1957.

We claim:

1. In a process for producing compositions in which a particulate refractory is dispersed in a metal the steps comprising (a) depositing a hydrous, oxygen-containing compound of a metal having an oxide with a AF at 27 C. of from 30 to 105 kcaL/gram atom of oxygen, together with substantially discrete, submicron particles of a refractory oxide having a melting point above 1000" C. and a AF at 1000 C. above 60 kcal./ gram atom of oxygen, (b) thereafter reducing the hydrous, oxygen-containing compound to the corresponding metal while maintaining a temperature throughout the entire mass below the sintering temperature of the metal the reduction being continued until the oxygen content of the mass, exclusive of oxygen in the refractory oxide, is below 2% by Weight, whereby a powdered product is obtained, sintering the powdered product at a temperature below the melting point of the metal until its particle size is in the range of l to 500 microns, and (d) mixing the sintered powder with up to 10 times its volume of a ductile powdered metal.

2. A process of claim 1 wherein the volume ratio of sintered powder to powdered metal is from 0.321 to 2.5: 1.

3. In a process for producing compositions in which a particulate refractory is dispersed in a metal the steps comprising (a) depositing a hydrous, oxygen-containing compound of a metal having an oxide with a AF at 27 C. of from 30 to 105 kcaL/gram atom of oxygen, together with substantially discrete, submicron particles of a refractory oxide having a melting point above 1000 C. and a AF at 1000 C. above 60 kcal./ gram atom of oxygen, (11) thereafter reducing the hydrous, oxygen-containing compound to the corresponding metal while maintaining a temperature throughout the entire mass below the sintering temperature of the metal the reduction being continued until the oxygen content of the mass, exclusive of oxygen in the refractory oxide, is below 2% by weight, whereby a powdered product is obtained, (0) sintering the powdered product at a temperature below the melting point of the metal until its particle size is in the range of 1 to 500 microns, (d) mixing the sintered powder with a ductile powdered metal, the volume ratio of sintered powder to powdered metal being from 0.3:1 to 25:1, (e) sintering the mixture to form a metal product having a density of from 60 to 100% of its absolute density, and (f) hot-working the resulting product.

4. A process for making metal parts having improved high-temperature properties, the process comprising blending a ductile powdered metal with a powdered metal composition having uniformly dispersed therein particles, having an average dimension of 5 to 250 millimicrons, of an oxide having a free energy of formation at 1000 C. above 60 kilocalories per gram atom of oxygen and having a melting point above 1000" C., the proportions being from 0.1 to parts by volume of ductile, powdered metal per part of the oxide-containing composition, compacting the blended powders, sintering the compact, and hot-working it and forming it to the shape of the part to be made.

5. A process for making metal parts having improved high-temperature properties, the process comprising blending a ductile powdered metal with a powdered metal composition having uniformly dispersed therein particles, having an average dimension of 5 to 250 millimicrons, of an oxide having a free energy of formation at 1000 C. above 60 kilocalories per gram atom of oxygen and having a melting point above 1000 C., the proportions being from 0.1 to 10 parts by volume of metal powder per part of the oxide-containing compositions, and compacting the mixture to a mass having an apparent density from 90 to 100% of absolute density.

6.'A pulverulent metal composition comprising a dispersion of a plurality of submicron particles of a refractory oxide having a melting point above 1000 C. and a AF at 1000" C. above 60 kilocalories per gram atom of oxygen in the oxide, in powder particles of a metal having an oxide with a AF at 27 C. of from 30 to 105 kcaL/ gram atom of oxygen, said dispersion having a surface area less than 10 square meters per gram and being mixed with up to 10 times its volume of a ductile powdered metal containing substantially no refractory particles, the refractory oxide and the metals being so selected that the AF of said oxide is at least 52+0.0l6M, where M isthe melting point of the mixture of metals in the final composition, in degrees Kelvin.

l 7. A pulverulent metal composition comprising a dispersion of a plurality of submicron particles of a refractory oxide having a melting point above 1000 C. and a AF at 1000 C. above 60 kilocalories per gram atom of oxygen in the oxide, in powder particles of a metal which has an oxide with a AF at 27 C. of from 30 to 105 kcal./ gram atom of oxygen, said dispersion having a surface area less than 10 square meters per gram and the average grain diameter of the metal in said dispersion being less than 5 microns, the powder dispersion being mixed with a ductile powdered metal which after annealing has an average grain diameter greater than 10 microns, the volume ratio of metal in the initial dispersion to ductile, powdered metal being from 0.321 to 2.5: 1, the refractory oxide and the metals being so selected that the AF of said refractory oxide is at least 52+0.0l6M, Where M is the melting point of the mixture of metals in the final composition, in degrees Kelvin.

8. A composition comprising powder particles of a metal selected from the group consisting of iron, cobalt, and nickel having uniformly dispersed in each powder particle a plurality of particles, having an average dimension of 5 to 250 millimicrons, of an oxide having a free energy of formation at 1000 C. above 60 to 150 kilocalories per gram atom of oxygen and having a melting point above 1000 C., the composition having a surface area less than 10 square meters per gram, and having mixed therewith up to 10 times its volume of ductile, powdered metal.

9. A solid metal composition in which there are (01) volumes of a metal consisting of grains smaller than 5 microns in average diameter and having substantially uniformly dispersed therein a plurality of discrete, submicron particles of a refractory metal oxide having a AF at 1000 C. of more than 60 kcaL/gram atom of oxygen, the metal being of the class which have oxides with a AF at 27 C. of 30 to 105 kcaL/gram atom of oxygen, and (b) intermingled with said first volumes, other volumes of metal having grains larger than 10 microns in average diameter and which are substantially free of said refractory metal oxide particles, the proportion of oxide-free to oxidefilled volumes being up to 10:1, the entire metal composition having an apparent density which is from to of the absolute density, and the refractory oxide and metal being so selected that the AF of the refractory oxide is at least 52+0.0l6M, where M is the melting point, in degrees Kelvin, of the metal in the composition.

10. A solid, annealed metal composition in which there are (a) volumes of a metal of the class which have oxides with a AF at 27 C. of 30 to kcal./gram atom of oxygen, the grains of said metal being smaller than 5 microns in average diameter and having uniformly dispersed therein discrete, submicron particles of a refractory metal oxide having a AF at 1000 C. which is more than 60 kcaL/ gram atom of oxygen and at least the value calculated from the expression 52+0.0l6M, where M is the melting point of the metal in degrees Kelvin, and (b) other volumes of the same metal in which the grains are larger than 10 microns in average diameter and are not in contact with said refractory oxide particles, the proportion of oxide-free to oxide-filled volumes being up to 10:1, the average grain size in the volumes (b) being at least two-fold the average grain size in volume (a).

11. A solid metal composition in which there are (a) volumes of a metal, the grains of which are less than 5 microns in average diameter, said grains having uniformly dispersed therein discrete, submicron-sized particles of a refractory metal oxide having a AF at 1000 C. which is

Claims (1)

1. IN A PROCESS FOR PRODUCING COMPOSITIONS IN WHICH A PARTICULATE REFRACTORY IS DISPERSED IN A METAL THE STEPS COMPRISING (A) DEPOSITING A HYDROUS, OXYGEN-CONTAINING COMPOUND OF A METAL HAVING AN OXIDE WITH A $F AT 27* C. OF FROM 30 TO 105 KCAL./GRAM ATOM OF OXYGEN, TOGETHER WITH SUBSTANTIALLY DESCRETE, SUBMICRON PARTICLES OF A REFRACTORY OXIDE HAVING A MELTING POINT ABOVE 1000*C. AND A $F AT 1000*C. ABOVE 60 KCAL./GRAM ATOM OF OXYGEN, (B) THEREAFTER REDUCING THE HYDROUS, OXYGEN-CONTAINING COMPOUND TO THE CORRESPONDING METAL WHILE MAINTAINING A TEMPERATURE THROUGHOUT THE ENTIRE MASS BELOW THE SINTERING TEMPERATURE OF THE METAL THE REDUCTION BEING CONTINUED UNTIL THE OXYGEN CONTENT OF THE MASS, EXCLUSIVE OF OXYGEN IN THE REFRACTORY OXIDE, IS BELOW 2% BY WEIGHT, WHEREBY A POWERED PRODUCT IS OBTAINED, (C) SINTERING THE POWDERED PRODUCT AT A TEMPERATURE BELOW THE MELTING POINT OF THE METAL UNTIL ITS PARTICLE SIZE IS IN THE RANGE OF 1 TO 500 MICRONS, AND (D) MIXING THE SINTERED POWDER WITH UP TO 10 TIMES ITS VOLUME OF A DUCTILE POWDERED METAL.

Images