DISPERSION-HARDENED TUNGSTEN ALLOY
United States Patent 3637422
Dispersion-hardened tungsten and tungsten alloy coatings having a material such as hafnium nitride at a concentration above 0.5 percent by volume are produced by covapor depositing tungsten and hafnium nitride. The hafnium nitride is produced in situ by addition of ammonia or nitrogen to a reducing atmosphere of hydrogen, which is mixed with tungsten and hafnium halide vapors in the presence of a substrate mandrel heated to a temperature above 900° C.
US Patent References:
Deposition of refractory metals and alloys thereof
Charlton et al. - February 1959 - 2873208
High-temperature alloy
Dickinson et al. - March 1966 - 3243291
Method for making tungsten metal articles
Heestand et al. - May 1967 - 3318724
TUNGSTEN-HAFNIUM-OXYGEN ALLOYS
Foldes - March 1969 - 3434811
Deposition of refractory metals and alloys thereof
Charlton et al. - February 1959 - 2873208
High-temperature alloy
Dickinson et al. - March 1966 - 3243291
Method for making tungsten metal articles
Heestand et al. - May 1967 - 3318724
TUNGSTEN-HAFNIUM-OXYGEN ALLOYS
Foldes - March 1969 - 3434811
Inventors:
Landingham, Richard L. (Livermore, CA)
Kane, James S. (Pleasanton, CA)
Kane, James S. (Pleasanton, CA)
Application Number:
04/695505
Publication Date:
01/25/1972
Filing Date:
01/03/1968
View Patent Images:
Export Citation:
Primary Class:
Other Classes:
428/660, 428/668, 420/431, 427/253
International Classes:
C23C16/06; C23C16/34; C23C16/56; C23C11/02
Field of Search:
117/107.2,106,131,123B 75/84.4,84.5,176 29/198
Other References:
landingham & Austin UCRL-50209 Feb. 24, 1967 Nucl. Sci. Abstr. 21 (22) 41604 Nov. 30, 1967..
Primary Examiner:
Kendall, Ralph S.
Claims:
We claim
1. In a process for producing dense and fine-grained deposits comprising at least about 99.5 percent by volume of tungsten, the steps comprising:
2. The process of claim 1, further defined in that said gaseous reaction mixture includes up to about 6 percent by volume of said tungsten halide, together with a volatilized, dispersoid metal halide component, said dispersoid metal halide being selected from the group consisting of the chlorides and bromides of hafnium, zirconium and titanium and comprising up to slightly more than 0.5 percent by volume of said tungsten halide in said reaction mixture.
3. The process of claim 2, further defined in that said gaseous reaction mixture also includes an excess of a dispersoid forming agent selected from the group consisting of water, oxygen, hydrocarbons, carbonyls, borontrichloride, borontribromide and boron hydride.
4. In a process for producing dense and fine-grained deposits of dispersion-hardened tungsten alloys comprising more than about 0.5 percent by volume of dispersoids, the steps comprising:
5. The process of claim 4, further defined in that said tungsten halide and hafnium halide are tungsten pentachloride and hafnium tetrachloride, respectively, and in that said dispersoid-forming agent is selected from the group consisting of nitrogen and ammonia.
6. The process of claim 4, further defined in that said tungsten halide and hafnium halide are tungsten hexachloride and hafnium tetrachloride, respectively, and in that said dispersoid-forming agent is selected from the group consisting of nitrogen and ammonia.
7. The process of claim 4, further defined in that said tungsten halide and hafnium halide are tungsten pentabromide and hafnium tetrabromide, respectively, and in that said dispersoid-forming agent is selected from the group consisting of nitrogen and ammonia.
1. In a process for producing dense and fine-grained deposits comprising at least about 99.5 percent by volume of tungsten, the steps comprising:
2. The process of claim 1, further defined in that said gaseous reaction mixture includes up to about 6 percent by volume of said tungsten halide, together with a volatilized, dispersoid metal halide component, said dispersoid metal halide being selected from the group consisting of the chlorides and bromides of hafnium, zirconium and titanium and comprising up to slightly more than 0.5 percent by volume of said tungsten halide in said reaction mixture.
3. The process of claim 2, further defined in that said gaseous reaction mixture also includes an excess of a dispersoid forming agent selected from the group consisting of water, oxygen, hydrocarbons, carbonyls, borontrichloride, borontribromide and boron hydride.
4. In a process for producing dense and fine-grained deposits of dispersion-hardened tungsten alloys comprising more than about 0.5 percent by volume of dispersoids, the steps comprising:
5. The process of claim 4, further defined in that said tungsten halide and hafnium halide are tungsten pentachloride and hafnium tetrachloride, respectively, and in that said dispersoid-forming agent is selected from the group consisting of nitrogen and ammonia.
6. The process of claim 4, further defined in that said tungsten halide and hafnium halide are tungsten hexachloride and hafnium tetrachloride, respectively, and in that said dispersoid-forming agent is selected from the group consisting of nitrogen and ammonia.
7. The process of claim 4, further defined in that said tungsten halide and hafnium halide are tungsten pentabromide and hafnium tetrabromide, respectively, and in that said dispersoid-forming agent is selected from the group consisting of nitrogen and ammonia.
Description:
BACKGROUND OF THE INVENTION
The invention generally pertains to the field of tungsten metallurgy, and particularly to the formation of fine-grained tungsten coatings, and to dispersion-hardened tungsten and tungsten alloys and methods for producing same.
The principal feature and advantage of tungsten and tungsten base alloys, especially the tungsten rhenium alloys, is the high melting point of these metals. The melting point of pure tungsten, 3,410° C., is the highest among the metals. Pure tungsten and tungsten alloys are relatively workable and reliable over wide temperature ranges, with the exception of chemical vapor-deposited tungsten. This latter type has characteristic columnar grain structure which has unreliable mechanical properties. None of the above metals have satisfactory creep resistance at elevated temperatures. The dispersion-hardened tungsten materials of the present invention are particularly useful in high temperature applications in view of their high creep resistance. They are, however, less ductile than pure tungsten and tungsten alloys.
There is an ever-growing need for such high temperature materials, for example, in nuclear reactor technology, space power projects, high temperature jet and rocket engines, etc. While tungsten and tungsten alloys such as the tungsten rhenium alloys are ideal from the point of view of temperature resistance, their strength deteriorates significantly under the influence of high temperature. Creep is a major problem, especially during long-term exposures of the metal to high temperatures. Since space missions for projected nuclear space reactors call for continuous operation at high temperatures for a number of years, improvement of the high temperature creep resistance of tungsten and tungsten alloys is particularly desirable.
THE PRIOR ART
The problem of enhancing creep resistance in metals and alloys has been effectively dealt with before. One of the common approaches, known as dispersion-hardening, is to introduce and uniformly distribute a small percentage of finely divided metal compound or ceramic additive into the metal or alloy crystal structure. (The terminology used throughout this case will be to refer to the metal compound effecting dispersion-hardening as the dispersoid. The metallic constituent of the dispersoid will be referred to as the dispersoid metal.) Often less than 1 percent of the additive material remarkably improves the creep resistance, as well as other mechanical properties of the matrix material. The dispersoid is usually carefully chosen to be compatible with the matrix, and at the same time resist motion or deformation of the matrix when under stress.
The rigorous treatment of the mechanisms and theory of dispersion-hardening is a complex subject and not well understood. The best qualitative explanation of the remarkable effectiveness of the additives in enhancing the material strength properties of the alloys is that the relatively immobile intrusions of foreign material (dispersoids) tend to retard grain boundary migration and dislocation movement and generation, so the creep resistance of the matrix is increased.
In particular, molybdenum base alloys have been dispersion-hardened by means of zirconium nitride and titanium nitride. These alloys are usually prepared by forming a mixture of molybdenum and zirconium or titanium metal in the liquid state, and, after solidification and appropriate treatment, exposing the alloy to a nitrogen pressure at elevated temperatures. Nitrogen then diffuses into the metal and reacts with the titanium or zirconium additives to form the nitrides which are the effective hardening dispersoid particles.
Notwithstanding the extensive body of prior knowledge with regard to the phenomenon of dispersion-hardening, as well as the specific dispersion-hardened molybdenum base alloys referred to above, the extension of this technology to tungsten alloys has heretofore been considered impossible or impractical for a variety of reasons. Firstly, the high melting temperature of tungsten creates severe handling problems. For example, graphite, the only material which has a melting point higher than tungsten, is a poor crucible material due to the high solubility of carbon in tungsten at melting temperature. Accordingly, conventional metallurgical techniques which commonly involve handling and processing of metals in a molten state are not readily applicable to producing tungsten and its alloys.
The diffusion rate of nitrogen into tungsten is expected to be high enough to form dispersoids therein. However, the solubility of nitrogen in tungsten, about 1 p.p.m. at 2,000° C., is less by a factor of about 30 than the solubility of nitrogen in molybdenum. Accordingly, considering the equilibrium TiN Ti + N(gas), where Ti is the metal dispersed in the tungsten matrix, the quantity of metal which is free to migrate through the base-metal matrix and to the surface is high at all times. Moreover, since the large dispersoid particles in dispersion alloys always grow at the expense of small particles, the formulation of a stable dispersion-hardened tungsten alloy has heretofore not been possible in view of the expected large migration rate of free metal reactant.
Accordingly, it is a principal object of the present invention to produce stable dispersion-hardened tungsten and tungsten alloys, and especially to produce a homogeneous distribution of dispersoid particles throughout the tungsten matrix to protect the tungsten and tungsten alloy matrix against aging.
Another object of the invention is to produce fine grain finished tungsten and tungsten alloys at relatively low temperatures, i.e., at temperatures considerably below the melting point of tungsten, such as around 1,000° C.
Pure tungsten coatings have been produced previously by vapor-depositing tungsten onto various substrate materials. The common process has been to reduce the highly volatile tungsten hexafluoride with hydrogen in the vapor phase in the presence of a heated mandrel. While tungsten coatings are readily deposited, microstructure examination of the coatings invariably shows a columnar grain structure perpendicular to the substrate surfaces. Since this type of grain growth is a major limiting factor upon the strength and reliability of the coating, the elimination thereof is particularly desirable.
Accordingly, it is yet another object of the present invention to produce pure tungsten and dispersion-hardened tungsten alloy coatings which do not exhibit the columnar grain structure typical of tungsten deposits, and instead have a uniform fine-grained microstructure throughout.
These and other objects will become apparent to those skilled in the art upon consideration of the following description of the invention and the detailed description of preferred processes and embodiments, in conjunction with the drawings, in which:
FIG. 1 is a schematic diagram of an apparatus for producing the tungsten and tungsten alloys of the present invention;
FIG. 2 is a graph illustrating the effect of deposition temperature upon the rate of deposition of metal; and
FIGS. 3, 4, 5 and 6 are photomicrographs of etched surfaces of dispersion-hardened tungsten coatings produced by alternate processes discussed in examples 1, 2, 3 and 4.
SUMMARY
The present invention provides a process for producing stable dispersion-hardened tungsten containing more than about 0.5 percent by volume of dispersoid materials, which avoids the difficulties encountered in the prior art. In accordance with the present invention, the dispersion-hardened alloy is formed by simultaneously vapor-depositing the tungsten matrix and the dispersoid particles preferably onto a suitably shaped, heated mandrel. The reactant gas phases are a mixture of volatile tungsten and dispersoid metal compounds, preferably the halides, on one hand, and on the other, a mixture of hydrogen, which serves as a reducing agent for the metal compounds, and a gaseous agent which reacts with the dispersoid metal to form the ultimate dispersoid materials. The reactant gas phases are intimately mixed, most appropriately by injection through nozzles producing turbulent flow in the region adjacent the hot mandrel substrate, which is maintained at a temperature of at least 900° C.
The present invention also provides pure tungsten deposits and tungsten alloys having very low dispersoid concentrations, i.e., less than about 0.5 percent by volume of dispersoids. These materials are produced by vapor deposition of pure tungsten or tungsten alloys in the presence of nitrogen, at temperatures between about 700° and 900° C., followed by heat treatment of the deposit above 900° C. Pure tungsten or low dispersoid concentration alloys so produced exhibit a remarkably fine and uniform grain structure, and are devoid of the typical columnar grains referred to above. The presence of nitrogen interrupts the columnar grain growth during deposition by forming tungsten nitride which can be decomposed above 900° C.
More specifically, the present invention provides dispersion-hardened tungsten and tungsten alloys having a fine-divided dispersion phase comprising over 0.5 volume percent, and preferably between about 2 and 8 volume percent, of hafnium nitride, although from the point of view of formation, there is no upper limit to the dispersoid which may in introduced by the present process. Other related alloys which can be made by the present process are tungsten base alloys wherein the dispersion phase is hafnium oxide, hafnium carbide, hafnium boride, and the oxides, carbides, borides and nitrides of titanium and zirconium. However, the hafnium-based dispersoids are by far the best materials since they form the most stable compounds with the best refractory properties.
For the production of the present coatings, the reactant gas phase providing the metallic species consists of volatile metal compounds, particularly the halides. The boiling point of the chlorides and bromides of tungsten, hafnium, zirconium and titanium are below 500° C., they are readily available, and easy to handle. However, ultimately any compound, including volatile metal organic compounds, could be employed. In choosing the compounds for the process, the following criteria are used:
1. The melting point of the compound must be below the optimum deposition temperature, or less than about 700° C.
2. the anionic part of the compound should be innocuous with respect to the remaining bases of the reaction system to preclude deleterious side reactions. Further, in the deposition of dispersion-hardened alloys, it is generally preferred to utilize compounds with the same anionic constituents for the metals to be codeposited.
The volatile metal-bearing vapors are mixed and caused to react with a gas phase consisting principally of hydrogen and a predetermined quantity of an additive gas which converts the hafnium, titanium or zirconium to the specific dispersoid or combination of dispersoids desired, i.e., the oxides, carbides, borides, and/or nitrides of said metals. Specific additive gases for producing oxides are preferably water vapor and, to a lesser extent, oxygen. The carbides may be produced by addition of carbonyls and hydrocarbons, and borides may be produced from borohydrates, boron halides, specifically BBr 3 and BCl 3 , and the nitrides by adding ammonia or nitrogen.
Pure tungsten coatings, and low volume percent dispersion alloys are produced by adding nitrogen to the reducing gas. In addition, one of the agents productive of oxide, carbide, and boride dispersoids may also be added; however, nitrogen, in the form of nitrogen gas or ammonia, is essential. These materials must be deposited at temperatures below 900° C. to effectuate the formation of an unstable compound, which is presumed to preclude the growth of crystals and maintain a uniform fine grain structure. This unstable compound, which is tungsten nitride, must be subsequently removed by heating the substrate to a temperature above 900° C.
The reaction gases are mixed in the vicinity of a substrate, which is usually a mandrel of the same shape as the part to be made. The substrate material is usually metal, which may be machined and polished to obviate finishing treatment of at least one surface of the tungsten alloy deposit. A wide choice of metals can be employed. A preferred substrate material is one which is readily workable, can withstand the plating temperature, and which is readily soluble in acid. Nickel, for example, satisfies these properties. Another material is molybdenum. These mandrels are readily removed by dissolving in nitric acid. Quartz may also be employed and removed by shocking the composite with liquid nitrogen. Of course, where it is desired to produce a coated material where the substrate is to be retained, any material, such as steel, can be employed.
The principal advantages of the present process are that it may be carried out at temperatures far below the melting point of tungsten, eliminating the problems normally encountered in treating tungsten with conventional metallurgical techniques, such as high temperature handling problems and introduction of impurities derived from crucibles into the tungsten matrix. The ultimate purity of the present tungsten alloys is essentially equivalent to the purity of the reaction gases. A further advantage of the present process is that a much finer dispersion of the hardening phase can be achieved than by other methods. This is so because in the vapor phase, the metals can be intermixed and comingled much more intimately than by any other means. The vapor deposition of these materials apparently occurs at sites which are similarly distributed and effectively control the growth of single crystals or grains of one species. The uniformity and homogeneity of the dispersoid distribution in turn has a pronounced effect upon the aging properties of the alloy, since the formation and growth of large particles is effectively retarded thereby.
A typical apparatus for carrying out the covapor deposition process is shown in FIG. 1. The apparatus consists basically of three components: the mixing and injection system (I), and reaction chamber (II), and the discharge system (III).
The injection system comprises gasline network 11 enclosed in furnace 12 for maintaining the gases in network 11 at a specified temperature. The principal function of the gasline network is to receive, heat and intermix the reaction gases in preparation for injection into the reaction chamber II. The reagent gases are led into network 11 from respective reservoirs or storage bottles. The volatile metal compounds, e.g., tungsten and dispersoid metal halides, are stored in pressure vessels 13, and 14, respectively. The pressure vessels are surrounded by individually controllable furnaces 15 and 16, by means of which the compounds are volatilized. The gas pressure may also be adjusted by controlling the temperature of the pressure vessels. The pressure vessels are connected to the mixing network 11 via heat insulator conduits 17 and 18, respectively. Each of these conduits is provided with the necessary instrumentation for appropriately metering the flow of gas into the network, i.e., valves 19 and 20, pressure gauges 21 and flowmeters 22. Hydrogen is introduced into network 11 via conduit 23. The length of the conduit 23 between the point where it enters furnace 12 and junction 24 where the hydrogen is intermixed with the metal compound vapors should be sufficiently long to heat the hydrogen to a temperature above the boiling point of the metal vapors. The same applies to the dispersoid-forming reagent, for example, ammonia, introduced into the system via conduit 26. Although the operating parameters, notably temperatures, are chosen to preclude reaction of the gases prior to injection, it is generally preferred to introduce the most reactive substance last, and near the nozzle 27 through which the gases are injected into the reaction chamber, in order to minimize premature deposition of the reaction products. Network 11 is also connected to an inert gas source (not shown) to allow flushing and clearing of the gaslines with argon or helium before and between runs.
Nozzle 27 introduces the gas mixture into the reaction chamber interior 28. Optimum deposition conditions are achieved by providing means for producing turbulent flow in the reaction chamber. This may be done by means of flowrates and baffles 30 disposed in the path of the inflowing gases. The reaction chamber is generally an electrical furnace capable of heating its contents and interior portions to at least 1,100° C. The furnace shown in FIG. 1 comprises a series of individually controllable clamshell furnaces 29 for establishing a temperature gradient along the gas passage 31. Such an arrangement has certain advantages. For example, it may be used to determine the optimum deposition conditions experimentally before commencing production runs. By providing a temperature gradient, the optimum temperature can be determined by comparison of the quality of the deposits produced in the various temperature zones on the substrate.
The mandrel 32 onto which the metals are deposited is disposed inside the chamber proximate nozzle 27.
The temperature of the mandrel should be uniform, and the mandrel surfaces should be uniformly accessible to the reaction gases. In the case of large, irregular workpieces, the mandrel may be heated internally and/or rotated randomly in the cavity to achieve uniform deposition rates over the entire mandrel surfaces. In such cases, the gases are preferably injected through a number of injection nozzles spaced uniformly throughout the chamber.
The final portion of the apparatus concerns the removal of waste and unreacted gases. From the reaction chamber 28, the gases are discharged into the removal system, and are first collectively cooled as by means of air-cooling the conduit. The conduit may be provided with fins or baffles 33 to aid the removal of heat. Unreacted metal vapors or other vaporous metal compounds are then removed by means of water-cooled trap 34, from whence they may be recovered and processed for reuse. The trap generally provides an enclosure having metal baffles 36 disposed across the direction of gas flow. These baffles are water-cooled either by direct contact with water or indirectly by being conductively joined to a surrounding water jacket 37. Unreacted ammonia or other dispersoid-forming gases such as water, carbonyl, etc., may be quantitatively removed by leading the gases from the trap 34 through liquid nitrogen traps 38 and 39. Waste hydrogen is vented to the atmosphere.
In operation, the volatile metal compounds are placed into pressure vessels 13 and 14, and volatilized by raising the temperature of the furnaces above the boiling point of the contents. Vaporization of the compounds is generally indicated by the pressure increase which may be observed in pressure gauges 21. The temperature of both furnaces 15 and 16 is preferentially 20°- 50° above the boiling point of the highest boiling metal compound. As mentioned previously, any metal compound having a suitably low boiling point may be employed. The preferred compounds, however, are the bromides and chlorides, i.e., WBr 5 , WBr 6 , WCl 5 , WCl 6 , ZrBr 4 , ZrCl 4 , TiBr 4 , TiCl 4 , HfBr 4 and HfCl 4 . Temperatures of about 450° C. in the furnaces 12, 15 and 16 suffice to prepare the mixture of reagent gases for injection. Other volatile compounds with boiling points below 500° C. are WF 6 , WOBr 4 , WOCl 4 , WOF 4 , TiF 4 , TiI 4 , titanium tetraisobutoxide, and tetraethoxy titanium, for example, However, as mentioned previously, it is preferable to employ metal compounds with the same anionic constituents to preclude side reactions. Accordingly, while the above compounds satisfy the low boiling point condition, only WF 6 and TiF 4 represent such a matched pair of compounds. However, no comparably suitable fluorides of hafnium and zirconium are known.
The desired relative reagent gas concentrations are obtained by adjusting the flowrates with which the gases are admitted into the apparatus. The primary determinant is, of course, the final constitution of the desired alloy. Regardless of the desired quantitative makeup of the alloy, the present apparatus can be used for covapor depositing any combination of tungsten and dispersoids. In practice, however, the dispersed concentration is generally about 5- 8 percent by volume. Accordingly, the flowrates of tungsten and dispersoid metal compounds are approximately proportional to the relative concentration of the metals in the final alloy. The deposition yield is generally somewhat lower for the dispersion metals. Accordingly, for producing dispersion-hardened tungsten having about 5- 8 percent by volume of dispersoids, the dispersion metal vapors in the reaction mixtures comprise 6- 9 percent by volume of the tungsten halide vapors.
The total pressure of the system, i.e., the sum of the partial pressures of the reactants, is not critical to the deposition and formation of the dispersion-hardened alloys, but has some influence upon the quality of the alloy deposit. Maximum densities and best uniformity and particle size distributions are obtained at total pressures between 5 and 35 torr. At higher pressures, the deposition rates increase; however, a concomitant increase in grain size and formation of void spaces takes place.
The temperature in the reaction chamber, or, more precisely, the substrate temperature, is an important deposition parameter because the choice of mandrel temperature not only affects the deposition rate and quality of the deposit in the sense of the ultimate density and porosity, but also the grain structure of the alloy or alloy coating.
FIG. 2 shows the relation between the thickness of the alloy coating (representative of the deposition rate) formed at different substrate temperatures. The deposition rate is generally highest if the substrate is maintained at a temperature between 800° and 900° C., with the maximum rate being at about 875° C.
Pure tungsten and dispersion-hardened tungsten having a low concentration of dispersoids are deposited at temperatures less than 900° with nitrogen or ammonia added to the gas phase. The significance of the 900° C. upper limit upon the deposition or mandrel temperature is that above this temperature the compound W 2 N is unstable and does not form during the plating process. Accordingly, coatings deposited above 900° C. will exhibit the undesirable columnar grain structure referred to above. If deposited below 900° C., preferably between 800° and 900° C. where both the deposition rate and the formation of a W 2 N phase are at an optimum, the formation of long grains is continuously interrupted by the deposition of W 2 N crystals. The resultant overall grain structure is therefore fine and granular, as shown in the photomicrograph in FIG. 3. The presence of a nitride-forming reagent, generally in a mole ratio N/μ between 0.25 and 2 in the gaseous reaction mixture, is, of course, necessary to effect the formation of the nitride W 2 N. However, nitrogen or ammonia need not be present continuously, but can be admitted intermittently during the deposition process to break up the grain growth in the deposit from time to time.
After the deposition has been completed, the W 2 N crystals are removed simply by heating the deposit to above 900° C. to decompose the tungsten nitride. The crystal structure of the deposit, however, remains fine and granular as before.
In tungsten alloys having a higher concentration of dispersoids, e.g., 0.5 percent by volume and higher, the dispersoids themselves break up the columnar structure of the tungsten crystals. Accordingly, these dispersoid alloys are deposited at temperatures exceeding 900° C., up to temperatures of about 1,200° C., to prevent the formation of the unstable W 2 N compound, whereby subsequent heat treatment becomes unnecessary. Dispersion-hardened tungsten alloys wherein the dispersion phase is a carbide, oxide or boride may be deposited at temperatures between about 700° and 1,200° C.
The grain size of the dispersoids is generally below 1 micron and down to about 300 angstroms. The grain size is variable within these limits by adjusting the temperature and relative concentration of the dispersoids. Higher temperatures effect an increase in the grain size. Since it is preferable to deposit the minimum grain size possible, the temperatures are held to a minimum consistent with the remaining parameters such as deposition rate, i.e., at the lower end of the temperature ranges indicated above. The concentration of the dispersoid-forming metal halide has a more pronounced effect upon the grain size. It has been found that the higher the dispersoid metal halide concentration is, the larger the grain size will be. On the other hand, the concentration of the dispersoid-forming agent, i.e., nitrogen, ammonia, water, etc., has been found to have no effect upon the particle size. Accordingly, the particle size of the dispersoid is held to a minimum by using an amount of dispersoid metal halide vapor which will just give the desired concentration of dispersoid in the finished material. To effect the complete deposition of the dispersoid, an excess of dispersoid-forming agent is used to drive the reaction to completion.
EXAMPLES
Vapor deposition parameters for producing specific coatings of tungsten and dispersion-hardened tungsten with a variety of dispersoids are given in the accompanying table. The parameters given are the mole ratios of the reactant species injected into the reaction chamber, injection and mandrel temperatures and chamber pressures. The nature and grain structure of the deposits obtained in runs 1- 4 are further illustrated in the photomicrographs of etched surfaces shown in FIGS. 3 to 6, respectively. ##SPC1##
Whereas the invention has been described with reference to the specific examples, it is to be understood that the scope of the invention should not be limited thereby, except as defined in the appended claims.
The invention generally pertains to the field of tungsten metallurgy, and particularly to the formation of fine-grained tungsten coatings, and to dispersion-hardened tungsten and tungsten alloys and methods for producing same.
The principal feature and advantage of tungsten and tungsten base alloys, especially the tungsten rhenium alloys, is the high melting point of these metals. The melting point of pure tungsten, 3,410° C., is the highest among the metals. Pure tungsten and tungsten alloys are relatively workable and reliable over wide temperature ranges, with the exception of chemical vapor-deposited tungsten. This latter type has characteristic columnar grain structure which has unreliable mechanical properties. None of the above metals have satisfactory creep resistance at elevated temperatures. The dispersion-hardened tungsten materials of the present invention are particularly useful in high temperature applications in view of their high creep resistance. They are, however, less ductile than pure tungsten and tungsten alloys.
There is an ever-growing need for such high temperature materials, for example, in nuclear reactor technology, space power projects, high temperature jet and rocket engines, etc. While tungsten and tungsten alloys such as the tungsten rhenium alloys are ideal from the point of view of temperature resistance, their strength deteriorates significantly under the influence of high temperature. Creep is a major problem, especially during long-term exposures of the metal to high temperatures. Since space missions for projected nuclear space reactors call for continuous operation at high temperatures for a number of years, improvement of the high temperature creep resistance of tungsten and tungsten alloys is particularly desirable.
THE PRIOR ART
The problem of enhancing creep resistance in metals and alloys has been effectively dealt with before. One of the common approaches, known as dispersion-hardening, is to introduce and uniformly distribute a small percentage of finely divided metal compound or ceramic additive into the metal or alloy crystal structure. (The terminology used throughout this case will be to refer to the metal compound effecting dispersion-hardening as the dispersoid. The metallic constituent of the dispersoid will be referred to as the dispersoid metal.) Often less than 1 percent of the additive material remarkably improves the creep resistance, as well as other mechanical properties of the matrix material. The dispersoid is usually carefully chosen to be compatible with the matrix, and at the same time resist motion or deformation of the matrix when under stress.
The rigorous treatment of the mechanisms and theory of dispersion-hardening is a complex subject and not well understood. The best qualitative explanation of the remarkable effectiveness of the additives in enhancing the material strength properties of the alloys is that the relatively immobile intrusions of foreign material (dispersoids) tend to retard grain boundary migration and dislocation movement and generation, so the creep resistance of the matrix is increased.
In particular, molybdenum base alloys have been dispersion-hardened by means of zirconium nitride and titanium nitride. These alloys are usually prepared by forming a mixture of molybdenum and zirconium or titanium metal in the liquid state, and, after solidification and appropriate treatment, exposing the alloy to a nitrogen pressure at elevated temperatures. Nitrogen then diffuses into the metal and reacts with the titanium or zirconium additives to form the nitrides which are the effective hardening dispersoid particles.
Notwithstanding the extensive body of prior knowledge with regard to the phenomenon of dispersion-hardening, as well as the specific dispersion-hardened molybdenum base alloys referred to above, the extension of this technology to tungsten alloys has heretofore been considered impossible or impractical for a variety of reasons. Firstly, the high melting temperature of tungsten creates severe handling problems. For example, graphite, the only material which has a melting point higher than tungsten, is a poor crucible material due to the high solubility of carbon in tungsten at melting temperature. Accordingly, conventional metallurgical techniques which commonly involve handling and processing of metals in a molten state are not readily applicable to producing tungsten and its alloys.
The diffusion rate of nitrogen into tungsten is expected to be high enough to form dispersoids therein. However, the solubility of nitrogen in tungsten, about 1 p.p.m. at 2,000° C., is less by a factor of about 30 than the solubility of nitrogen in molybdenum. Accordingly, considering the equilibrium TiN Ti + N(gas), where Ti is the metal dispersed in the tungsten matrix, the quantity of metal which is free to migrate through the base-metal matrix and to the surface is high at all times. Moreover, since the large dispersoid particles in dispersion alloys always grow at the expense of small particles, the formulation of a stable dispersion-hardened tungsten alloy has heretofore not been possible in view of the expected large migration rate of free metal reactant.
Accordingly, it is a principal object of the present invention to produce stable dispersion-hardened tungsten and tungsten alloys, and especially to produce a homogeneous distribution of dispersoid particles throughout the tungsten matrix to protect the tungsten and tungsten alloy matrix against aging.
Another object of the invention is to produce fine grain finished tungsten and tungsten alloys at relatively low temperatures, i.e., at temperatures considerably below the melting point of tungsten, such as around 1,000° C.
Pure tungsten coatings have been produced previously by vapor-depositing tungsten onto various substrate materials. The common process has been to reduce the highly volatile tungsten hexafluoride with hydrogen in the vapor phase in the presence of a heated mandrel. While tungsten coatings are readily deposited, microstructure examination of the coatings invariably shows a columnar grain structure perpendicular to the substrate surfaces. Since this type of grain growth is a major limiting factor upon the strength and reliability of the coating, the elimination thereof is particularly desirable.
Accordingly, it is yet another object of the present invention to produce pure tungsten and dispersion-hardened tungsten alloy coatings which do not exhibit the columnar grain structure typical of tungsten deposits, and instead have a uniform fine-grained microstructure throughout.
These and other objects will become apparent to those skilled in the art upon consideration of the following description of the invention and the detailed description of preferred processes and embodiments, in conjunction with the drawings, in which:
FIG. 1 is a schematic diagram of an apparatus for producing the tungsten and tungsten alloys of the present invention;
FIG. 2 is a graph illustrating the effect of deposition temperature upon the rate of deposition of metal; and
FIGS. 3, 4, 5 and 6 are photomicrographs of etched surfaces of dispersion-hardened tungsten coatings produced by alternate processes discussed in examples 1, 2, 3 and 4.
SUMMARY
The present invention provides a process for producing stable dispersion-hardened tungsten containing more than about 0.5 percent by volume of dispersoid materials, which avoids the difficulties encountered in the prior art. In accordance with the present invention, the dispersion-hardened alloy is formed by simultaneously vapor-depositing the tungsten matrix and the dispersoid particles preferably onto a suitably shaped, heated mandrel. The reactant gas phases are a mixture of volatile tungsten and dispersoid metal compounds, preferably the halides, on one hand, and on the other, a mixture of hydrogen, which serves as a reducing agent for the metal compounds, and a gaseous agent which reacts with the dispersoid metal to form the ultimate dispersoid materials. The reactant gas phases are intimately mixed, most appropriately by injection through nozzles producing turbulent flow in the region adjacent the hot mandrel substrate, which is maintained at a temperature of at least 900° C.
The present invention also provides pure tungsten deposits and tungsten alloys having very low dispersoid concentrations, i.e., less than about 0.5 percent by volume of dispersoids. These materials are produced by vapor deposition of pure tungsten or tungsten alloys in the presence of nitrogen, at temperatures between about 700° and 900° C., followed by heat treatment of the deposit above 900° C. Pure tungsten or low dispersoid concentration alloys so produced exhibit a remarkably fine and uniform grain structure, and are devoid of the typical columnar grains referred to above. The presence of nitrogen interrupts the columnar grain growth during deposition by forming tungsten nitride which can be decomposed above 900° C.
More specifically, the present invention provides dispersion-hardened tungsten and tungsten alloys having a fine-divided dispersion phase comprising over 0.5 volume percent, and preferably between about 2 and 8 volume percent, of hafnium nitride, although from the point of view of formation, there is no upper limit to the dispersoid which may in introduced by the present process. Other related alloys which can be made by the present process are tungsten base alloys wherein the dispersion phase is hafnium oxide, hafnium carbide, hafnium boride, and the oxides, carbides, borides and nitrides of titanium and zirconium. However, the hafnium-based dispersoids are by far the best materials since they form the most stable compounds with the best refractory properties.
For the production of the present coatings, the reactant gas phase providing the metallic species consists of volatile metal compounds, particularly the halides. The boiling point of the chlorides and bromides of tungsten, hafnium, zirconium and titanium are below 500° C., they are readily available, and easy to handle. However, ultimately any compound, including volatile metal organic compounds, could be employed. In choosing the compounds for the process, the following criteria are used:
1. The melting point of the compound must be below the optimum deposition temperature, or less than about 700° C.
2. the anionic part of the compound should be innocuous with respect to the remaining bases of the reaction system to preclude deleterious side reactions. Further, in the deposition of dispersion-hardened alloys, it is generally preferred to utilize compounds with the same anionic constituents for the metals to be codeposited.
The volatile metal-bearing vapors are mixed and caused to react with a gas phase consisting principally of hydrogen and a predetermined quantity of an additive gas which converts the hafnium, titanium or zirconium to the specific dispersoid or combination of dispersoids desired, i.e., the oxides, carbides, borides, and/or nitrides of said metals. Specific additive gases for producing oxides are preferably water vapor and, to a lesser extent, oxygen. The carbides may be produced by addition of carbonyls and hydrocarbons, and borides may be produced from borohydrates, boron halides, specifically BBr 3 and BCl 3 , and the nitrides by adding ammonia or nitrogen.
Pure tungsten coatings, and low volume percent dispersion alloys are produced by adding nitrogen to the reducing gas. In addition, one of the agents productive of oxide, carbide, and boride dispersoids may also be added; however, nitrogen, in the form of nitrogen gas or ammonia, is essential. These materials must be deposited at temperatures below 900° C. to effectuate the formation of an unstable compound, which is presumed to preclude the growth of crystals and maintain a uniform fine grain structure. This unstable compound, which is tungsten nitride, must be subsequently removed by heating the substrate to a temperature above 900° C.
The reaction gases are mixed in the vicinity of a substrate, which is usually a mandrel of the same shape as the part to be made. The substrate material is usually metal, which may be machined and polished to obviate finishing treatment of at least one surface of the tungsten alloy deposit. A wide choice of metals can be employed. A preferred substrate material is one which is readily workable, can withstand the plating temperature, and which is readily soluble in acid. Nickel, for example, satisfies these properties. Another material is molybdenum. These mandrels are readily removed by dissolving in nitric acid. Quartz may also be employed and removed by shocking the composite with liquid nitrogen. Of course, where it is desired to produce a coated material where the substrate is to be retained, any material, such as steel, can be employed.
The principal advantages of the present process are that it may be carried out at temperatures far below the melting point of tungsten, eliminating the problems normally encountered in treating tungsten with conventional metallurgical techniques, such as high temperature handling problems and introduction of impurities derived from crucibles into the tungsten matrix. The ultimate purity of the present tungsten alloys is essentially equivalent to the purity of the reaction gases. A further advantage of the present process is that a much finer dispersion of the hardening phase can be achieved than by other methods. This is so because in the vapor phase, the metals can be intermixed and comingled much more intimately than by any other means. The vapor deposition of these materials apparently occurs at sites which are similarly distributed and effectively control the growth of single crystals or grains of one species. The uniformity and homogeneity of the dispersoid distribution in turn has a pronounced effect upon the aging properties of the alloy, since the formation and growth of large particles is effectively retarded thereby.
A typical apparatus for carrying out the covapor deposition process is shown in FIG. 1. The apparatus consists basically of three components: the mixing and injection system (I), and reaction chamber (II), and the discharge system (III).
The injection system comprises gasline network 11 enclosed in furnace 12 for maintaining the gases in network 11 at a specified temperature. The principal function of the gasline network is to receive, heat and intermix the reaction gases in preparation for injection into the reaction chamber II. The reagent gases are led into network 11 from respective reservoirs or storage bottles. The volatile metal compounds, e.g., tungsten and dispersoid metal halides, are stored in pressure vessels 13, and 14, respectively. The pressure vessels are surrounded by individually controllable furnaces 15 and 16, by means of which the compounds are volatilized. The gas pressure may also be adjusted by controlling the temperature of the pressure vessels. The pressure vessels are connected to the mixing network 11 via heat insulator conduits 17 and 18, respectively. Each of these conduits is provided with the necessary instrumentation for appropriately metering the flow of gas into the network, i.e., valves 19 and 20, pressure gauges 21 and flowmeters 22. Hydrogen is introduced into network 11 via conduit 23. The length of the conduit 23 between the point where it enters furnace 12 and junction 24 where the hydrogen is intermixed with the metal compound vapors should be sufficiently long to heat the hydrogen to a temperature above the boiling point of the metal vapors. The same applies to the dispersoid-forming reagent, for example, ammonia, introduced into the system via conduit 26. Although the operating parameters, notably temperatures, are chosen to preclude reaction of the gases prior to injection, it is generally preferred to introduce the most reactive substance last, and near the nozzle 27 through which the gases are injected into the reaction chamber, in order to minimize premature deposition of the reaction products. Network 11 is also connected to an inert gas source (not shown) to allow flushing and clearing of the gaslines with argon or helium before and between runs.
Nozzle 27 introduces the gas mixture into the reaction chamber interior 28. Optimum deposition conditions are achieved by providing means for producing turbulent flow in the reaction chamber. This may be done by means of flowrates and baffles 30 disposed in the path of the inflowing gases. The reaction chamber is generally an electrical furnace capable of heating its contents and interior portions to at least 1,100° C. The furnace shown in FIG. 1 comprises a series of individually controllable clamshell furnaces 29 for establishing a temperature gradient along the gas passage 31. Such an arrangement has certain advantages. For example, it may be used to determine the optimum deposition conditions experimentally before commencing production runs. By providing a temperature gradient, the optimum temperature can be determined by comparison of the quality of the deposits produced in the various temperature zones on the substrate.
The mandrel 32 onto which the metals are deposited is disposed inside the chamber proximate nozzle 27.
The temperature of the mandrel should be uniform, and the mandrel surfaces should be uniformly accessible to the reaction gases. In the case of large, irregular workpieces, the mandrel may be heated internally and/or rotated randomly in the cavity to achieve uniform deposition rates over the entire mandrel surfaces. In such cases, the gases are preferably injected through a number of injection nozzles spaced uniformly throughout the chamber.
The final portion of the apparatus concerns the removal of waste and unreacted gases. From the reaction chamber 28, the gases are discharged into the removal system, and are first collectively cooled as by means of air-cooling the conduit. The conduit may be provided with fins or baffles 33 to aid the removal of heat. Unreacted metal vapors or other vaporous metal compounds are then removed by means of water-cooled trap 34, from whence they may be recovered and processed for reuse. The trap generally provides an enclosure having metal baffles 36 disposed across the direction of gas flow. These baffles are water-cooled either by direct contact with water or indirectly by being conductively joined to a surrounding water jacket 37. Unreacted ammonia or other dispersoid-forming gases such as water, carbonyl, etc., may be quantitatively removed by leading the gases from the trap 34 through liquid nitrogen traps 38 and 39. Waste hydrogen is vented to the atmosphere.
In operation, the volatile metal compounds are placed into pressure vessels 13 and 14, and volatilized by raising the temperature of the furnaces above the boiling point of the contents. Vaporization of the compounds is generally indicated by the pressure increase which may be observed in pressure gauges 21. The temperature of both furnaces 15 and 16 is preferentially 20°- 50° above the boiling point of the highest boiling metal compound. As mentioned previously, any metal compound having a suitably low boiling point may be employed. The preferred compounds, however, are the bromides and chlorides, i.e., WBr 5 , WBr 6 , WCl 5 , WCl 6 , ZrBr 4 , ZrCl 4 , TiBr 4 , TiCl 4 , HfBr 4 and HfCl 4 . Temperatures of about 450° C. in the furnaces 12, 15 and 16 suffice to prepare the mixture of reagent gases for injection. Other volatile compounds with boiling points below 500° C. are WF 6 , WOBr 4 , WOCl 4 , WOF 4 , TiF 4 , TiI 4 , titanium tetraisobutoxide, and tetraethoxy titanium, for example, However, as mentioned previously, it is preferable to employ metal compounds with the same anionic constituents to preclude side reactions. Accordingly, while the above compounds satisfy the low boiling point condition, only WF 6 and TiF 4 represent such a matched pair of compounds. However, no comparably suitable fluorides of hafnium and zirconium are known.
The desired relative reagent gas concentrations are obtained by adjusting the flowrates with which the gases are admitted into the apparatus. The primary determinant is, of course, the final constitution of the desired alloy. Regardless of the desired quantitative makeup of the alloy, the present apparatus can be used for covapor depositing any combination of tungsten and dispersoids. In practice, however, the dispersed concentration is generally about 5- 8 percent by volume. Accordingly, the flowrates of tungsten and dispersoid metal compounds are approximately proportional to the relative concentration of the metals in the final alloy. The deposition yield is generally somewhat lower for the dispersion metals. Accordingly, for producing dispersion-hardened tungsten having about 5- 8 percent by volume of dispersoids, the dispersion metal vapors in the reaction mixtures comprise 6- 9 percent by volume of the tungsten halide vapors.
The total pressure of the system, i.e., the sum of the partial pressures of the reactants, is not critical to the deposition and formation of the dispersion-hardened alloys, but has some influence upon the quality of the alloy deposit. Maximum densities and best uniformity and particle size distributions are obtained at total pressures between 5 and 35 torr. At higher pressures, the deposition rates increase; however, a concomitant increase in grain size and formation of void spaces takes place.
The temperature in the reaction chamber, or, more precisely, the substrate temperature, is an important deposition parameter because the choice of mandrel temperature not only affects the deposition rate and quality of the deposit in the sense of the ultimate density and porosity, but also the grain structure of the alloy or alloy coating.
FIG. 2 shows the relation between the thickness of the alloy coating (representative of the deposition rate) formed at different substrate temperatures. The deposition rate is generally highest if the substrate is maintained at a temperature between 800° and 900° C., with the maximum rate being at about 875° C.
Pure tungsten and dispersion-hardened tungsten having a low concentration of dispersoids are deposited at temperatures less than 900° with nitrogen or ammonia added to the gas phase. The significance of the 900° C. upper limit upon the deposition or mandrel temperature is that above this temperature the compound W 2 N is unstable and does not form during the plating process. Accordingly, coatings deposited above 900° C. will exhibit the undesirable columnar grain structure referred to above. If deposited below 900° C., preferably between 800° and 900° C. where both the deposition rate and the formation of a W 2 N phase are at an optimum, the formation of long grains is continuously interrupted by the deposition of W 2 N crystals. The resultant overall grain structure is therefore fine and granular, as shown in the photomicrograph in FIG. 3. The presence of a nitride-forming reagent, generally in a mole ratio N/μ between 0.25 and 2 in the gaseous reaction mixture, is, of course, necessary to effect the formation of the nitride W 2 N. However, nitrogen or ammonia need not be present continuously, but can be admitted intermittently during the deposition process to break up the grain growth in the deposit from time to time.
After the deposition has been completed, the W 2 N crystals are removed simply by heating the deposit to above 900° C. to decompose the tungsten nitride. The crystal structure of the deposit, however, remains fine and granular as before.
In tungsten alloys having a higher concentration of dispersoids, e.g., 0.5 percent by volume and higher, the dispersoids themselves break up the columnar structure of the tungsten crystals. Accordingly, these dispersoid alloys are deposited at temperatures exceeding 900° C., up to temperatures of about 1,200° C., to prevent the formation of the unstable W 2 N compound, whereby subsequent heat treatment becomes unnecessary. Dispersion-hardened tungsten alloys wherein the dispersion phase is a carbide, oxide or boride may be deposited at temperatures between about 700° and 1,200° C.
The grain size of the dispersoids is generally below 1 micron and down to about 300 angstroms. The grain size is variable within these limits by adjusting the temperature and relative concentration of the dispersoids. Higher temperatures effect an increase in the grain size. Since it is preferable to deposit the minimum grain size possible, the temperatures are held to a minimum consistent with the remaining parameters such as deposition rate, i.e., at the lower end of the temperature ranges indicated above. The concentration of the dispersoid-forming metal halide has a more pronounced effect upon the grain size. It has been found that the higher the dispersoid metal halide concentration is, the larger the grain size will be. On the other hand, the concentration of the dispersoid-forming agent, i.e., nitrogen, ammonia, water, etc., has been found to have no effect upon the particle size. Accordingly, the particle size of the dispersoid is held to a minimum by using an amount of dispersoid metal halide vapor which will just give the desired concentration of dispersoid in the finished material. To effect the complete deposition of the dispersoid, an excess of dispersoid-forming agent is used to drive the reaction to completion.
EXAMPLES
Vapor deposition parameters for producing specific coatings of tungsten and dispersion-hardened tungsten with a variety of dispersoids are given in the accompanying table. The parameters given are the mole ratios of the reactant species injected into the reaction chamber, injection and mandrel temperatures and chamber pressures. The nature and grain structure of the deposits obtained in runs 1- 4 are further illustrated in the photomicrographs of etched surfaces shown in FIGS. 3 to 6, respectively. ##SPC1##
Whereas the invention has been described with reference to the specific examples, it is to be understood that the scope of the invention should not be limited thereby, except as defined in the appended claims.