ELSEVIER Refractory Metals & Hard Materials 12 ( 1993-1994) 35 -40 O 1994 Elsevier Science Limited Printed in Great Britain. All rights reserved 0263-4368/94/$7.00 Effect of Intermetallic Compounds on the Properties of Tantalum Prabhat Kumar & C. E. Mosheim CPM Division, Cabot Corporation, County Line Road, Boyertown, PA 19512, USA (Received May 25 1993; accepted August 12 1993) Abstract: Whereas pure tantalum has excellent corrosion-resistance and form- ability, its high-temperature properties and thermal stability are marginal for some intended applications. Traditional approaches for improving these properties have been dispersion and solid-solution strengthening. Modifica- tions of properties via an intermetallic precipitation was not considered until recently. Results of an on-going investigation on the processing and evaluation of sili- cide-strengthened tantalum are presented. Yttrium-silicide-containing tantalum samples were produced via the P/M method. Evaluation consisted in micro- structural, mechanical, chemical, and functional tests. Results were compared with those of commercially available tantalum. Intermetallics precipitates were found to be very potent in altering these properties. Although the preliminary results are encouraging, extensive functional testing is required to ensure that there is no unexpected adverse effect. 1 BACKGROUND Tantalum alloys have been recognized as pre- ferred materials in the field of furnace equipment, such as trays and heating elements, and radiation shielding where the thermal stability of the alloy is maintained and the life span of the product is enhanced by reduced embritflement. Tantalum alloys have also been employed in the manufac- ture of wire and more particularly as electric-com- ponent leads where product characteristics such as ductility, high dielectric constant, resistance to grain growth at elevated temperatures, and improved processability are required. In the production of capacitors, for example, the lead wires may either be pressed into the tantalum- powder anode and subsequently sintered at high temperatures, or spot-welded to sintered capaci- tor bodies. In both electrical-component and furnace- equipment products, contamination by oxygen contributes to embrittlement and piece failure. 35 For example, in wire products, the area in which a lead wire leaves an anode body is highly suscep- tible to embrittlement due to migration of oxygen from the sintered body to the wire. Lead wires that become embrittled or break result in the loss of the entire piece. Oxygen embrittlement occurs in tantalum-base alloy products by several mechanisms. Tantalum acts as a getter for oxygen in addition to other gaseous impurities present in sintering operations, such as carbon monoxide, carbon dioxide, and water vapor. Attempts have been made to reduce tantalum oxide formation by doping tantalum with carbon or carbonaceous material. Oxygen reacts with the carbon at the surface of the metal rather than diffusing into the tantalum and thereby minimizes embrittlement. Whereas enhanced ductility levels may be achieved with carbon addition, the dopant may adversely affect the processability and electrical characteristics of the metal. Micro-alloying with silicon has also been used to improve the oxygen-embrittlement-resistance
Transcript
ELSEVIER
Refractory Metals & Hard Materials 12 ( 1993-1994) 35 -40 O 1994 Elsevier Science Limited
Printed in Great Britain. All rights reserved 0263-4368/94/$7.00
Effect of Intermetallic Compounds on the Properties of Tantalum
Prabhat Kumar & C. E. Mosheim CPM Division, Cabot Corporation, County Line Road, Boyertown, PA 19512, USA
(Received May 25 1993; accepted August 12 1993)
Abstract: Whereas pure tantalum has excellent corrosion-resistance and form- ability, its high-temperature properties and thermal stability are marginal for some intended applications. Traditional approaches for improving these properties have been dispersion and solid-solution strengthening. Modifica- tions of properties via an intermetallic precipitation was not considered until recently.
Results of an on-going investigation on the processing and evaluation of sili- cide-strengthened tantalum are presented. Yttrium-silicide-containing tantalum samples were produced via the P/M method. Evaluation consisted in micro- structural, mechanical, chemical, and functional tests. Results were compared with those of commercially available tantalum. Intermetallics precipitates were found to be very potent in altering these properties.
Although the preliminary results are encouraging, extensive functional testing is required to ensure that there is no unexpected adverse effect.
1 BACKGROUND
Tantalum alloys have been recognized as pre- ferred materials in the field of furnace equipment, such as trays and heating elements, and radiation shielding where the thermal stability of the alloy is maintained and the life span of the product is enhanced by reduced embritflement. Tantalum alloys have also been employed in the manufac- ture of wire and more particularly as electric-com- ponent leads where product characteristics such as ductility, high dielectric constant, resistance to grain growth at elevated temperatures, and improved processability are required. In the production of capacitors, for example, the lead wires may either be pressed into the tantalum- powder anode and subsequently sintered at high temperatures, or spot-welded to sintered capaci- tor bodies.
In both electrical-component and furnace- equipment products, contamination by oxygen contributes to embrittlement and piece failure.
35
For example, in wire products, the area in which a lead wire leaves an anode body is highly suscep- tible to embrittlement due to migration of oxygen from the sintered body to the wire. Lead wires that become embrittled or break result in the loss of the entire piece. Oxygen embrittlement occurs in tantalum-base alloy products by several mechanisms. Tantalum acts as a getter for oxygen in addition to other gaseous impurities present in sintering operations, such as carbon monoxide, carbon dioxide, and water vapor. Attempts have been made to reduce tantalum oxide formation by doping tantalum with carbon or carbonaceous material. Oxygen reacts with the carbon at the surface of the metal rather than diffusing into the tantalum and thereby minimizes embrittlement. Whereas enhanced ductility levels may be achieved with carbon addition, the dopant may adversely affect the processability and electrical characteristics of the metal.
Micro-alloying with silicon has also been used to improve the oxygen-embrittlement-resistance
36 Prabhat Kumar, C E. Mosheim
of tantalum. Silicon is volatilized in part during processing and therefore must be added in excess in the original master blend.
Whereas it is speculated that silicon functions as a getter similar to carbon, the addition of excess silicon may affect the electrical characteristics of the wire product by the same mechanism described above for carbon or carbonaceous materials.
Another mechanism for reducing the embrittle- ment of tantalum-base-alloy products involves the doping of tantalum powder with yttrium or similar metals. Since yttrium is more stable than tantalum oxide, yttrium oxidizes to yttria during processing. During the subsequent high-tempera- ture exposure, oxide particles pin the grain- boundaries and prevent the grain-growth; improvement in embrittlement-resistance is pri- marily due to the fine grain-size. Obviously, oxide particles become ineffective after extended high- temperature exposure due to increase in their size. Although the mechanism is not completely under- stood, one theory accounting for dopant-particle growth or 'dispersant coarsening' is that the coars- ening occurs as a result of the high diffusion rate of oxygen and metal atoms of oxides in refractory metals, which is driven by the interfacial energy of the dispersoids. Enlarged dispersant particles have lower surface energy and therefore cannot function to restrain grain-boundary migration. Grain-growth, in turn, results in loss of ductility.
A combination of dopants has also been used, for example, alloying with silicon and yttria. It is maintained that the mechanisms in which silicon functions as an oxygen getter and metal oxide functions as a grain-boundary restraint explain the basis for the reported fine grain-size and ductility. The mechanisms, however, suffer from previously discussed problems of product quality due to sili- con evaporation and grain-growth after exposure to high temperatures owing to the growth of dis- persant.
Table 1 summarizes alloying elements normally used with Ta. In principle, both interstitial and substitutional alloying elements can be used. However, owing to the relative ease of controlling the composition and processing, the substitutional alloying is generally used.
2 OBJECTIVE
The objective of this on-going investigation is to provide a doped tantalum alloy that maintains a
Table 1. Elements used for microalloying tantalum A.
Alloying Atomic Substitutional/ element radius ~ interstitial b
O 0"6 I N 0.71 I C 0.77 I Si 1.17 S Mo 1.40 S W 1-41 S Nb 1.47 S Y 1.81 S
aAtomic radius of Ta = 1.47A. bSolute/solvent < 0.59 for interstitial.
high level of processability and ductility and in which the dopants resist coarsening after expo- sure to high temperatures.
Intermetallic compounds, such as silicides, are expected to provide the desired results. Since these compounds are not inert to the tantalum matrix, they are not prone to coarsening after exposure to high temperature. At the same time, they exhibit adequate stability to pin the grain- boundaries.
3 RESULTS AND DISCUSSION
A In-sire silicide formation
Powder metallurgical processing was used for producing wrought tantalum samples having in- situ silicide precipitates. The first experiment was to establish that in-situ silicides could indeed be formed. A blend of Ta + 10 wt% YN + 40% Si powder was heated at 1300°C for 2 h and evalu- ated via X-ray diffraction. Another blend of Ta+ 10 wt% Y 2 0 3 + 4 0 % Si was also processed and evaluated for comparison. As illustrated in Table 2, the blend containing the composition of yttrium nitride and silicon showed the presence of yttrium silicide, dispersed in the base-metal matrix, whereas the yttrium oxide and silicon blend did not. Although the latter did have yttrium silicate, the thermodynamic stability of
Effect of intermetallic compounds on properties of tantalum
Table 2. Identification of yttrium silicide and yttrium silicate by X-ray diffraction (XRD)
37
Composition Ta + 10% YN + 40% Si 1300°C Ti + 10% Y203 + 4% Si 1300°C heated at
XRD of sample Known pattern for Y Si 2 XRD of sample Known pattern for Y2 Si05
yttrium oxide apparently prohibits its decomposi- tion. It is believed that yttrium oxide pre-empts the formation of yttrium silicide. Silicide cannot be formed, and an oxide (yttrium silicate) is formed instead.
It should be noted that, whereas a blend of YN + Si was used for the in-situ formation of yttrium silicide in this investigation, several other methods may be equally effective.
B Micro-alloyed Ta wire
Table 3 gives physical and chemical properties of starting tantalum powder. This powder was mixed with dopants to obtain the following normal com- positions (by weight).
Blend I.D. 1 2 3 4
400 ppm Si + 100 ppm YN DR-Ta* 400 ppm Si+ 100 ppm Y203 400 ppm Si
The blended powder was cold isostatically pressed into bars at 60 000 psi (lbf/in2), each bar
*Cabot Performance Material produces DR-Ta for capaci- tor-can and lead-wire applications.
weighed about 22 lbf. The cross-section of the bar was about 41 m m x 41 mm. The bars were sintered by direct-resistance sintering in a vacuum furnace at a temperature of between about 2200 and 2400°C. The bars were maintained at the temperature for about 4 h. Sintered bars were rolled to a 20-mmx 20-mm cross-section and annealed at a temperature of 1300°C for a period of about 2 h. The bars were then rolled to 9 mmx 9 mm and reannealed at 1300°C for an additional 2 h, bars were subsequently drawn through various dies and annealed at a tempera- ture of about 1300°C. The final wire diameter was 0.25 ram.
Table 4 gives the grain-size, mechanical and chemical properties of wires from blends 1, 2, 3 and 4. High strength and ductility of the wire from blend 1 are evident.
Wires from blends 1-4 were pressed into tanta- lum powder, sintered under vacuum, and tested for bend-ductility in accordance with the test procedure described below.
The bend-ductility of the sintered wire was determined by securing a sintered anode pre- formed with 1-in wire embedded in it. A 54-g dead weight was attached to the lead extremity.
Chemical composition (ppm) Si 225 -- 250 250 Y 30 -- 40 -- C 45 45 65 50 N2 45 35 30 10 02 190 145 120 75 Others None None None None
aNot fully recrystallized (NFR).
The anode was then pivoted through a 180 ° arc, which caused the wire to bend at the junction with the anode. One bend was defined as the complete pivoting of the anode through a 90 ° arc and returning to the starting position. The number of bends was counted. Ten anodes were tested and the reported bend ductility is an average of these ten trials.
Three sintering cycles were used. In the first cycle, the furnace was evacuated, and the tem- perature was raised to 1670°C for 30 min and then shut-off. The second cycle was the same as the first except that the furnace was back-filled
with argon after the evacuation and re-evacuated, and the temperature was then raised to 1670°C; after 30 min, the furnace was shut off. The third cycle was the same as the first except that wire/ powder assemblies were reheated for 2 min at 1670°C.
Table 5 compares the bend-ductility of wire formed by the procedures set forth in blends 1-4. The wire containing Si and YN exhibited 57% improvement in comparison with tantalum wire doped with silicon and yttrium oxide after 30 min of sintering followed by an additional 2 min.
C Micro-alloyed Ta sheet
Compositions of blends 1, 2, 3, and 4 were also processed into 9-mmx9-mm annealed bars, which were rolled into 0.38-mm-thick sheets. The sheets were annealed at various temperatures.
Table 6 compares the grain-sizes of sheets pro- duced by the examples listed. Sheets of composi- tions 1 (400Si + 100YN) and 3 (400Si + 100Y203) were evaluated via electron microscopy after annealing at 1500°C. Discs were cut to about 250 pm in thickness by using a slow-speed diamond saw. The discs were then ion-milled to a thickness of 50-100 /zm and electropolished in a 90% H2SO4 + 10% HF solution until they developed microperforations. Diffraction patterns of lattices of samples of composition 1 (400Si+ 100YN) and composition 3 (400Si+ 100Y203) w e r e also taken. The electron microscopy was performed in the vicinity of the perforations. Scanning electron micrographs in the vicinity of microperforations demonstrate that the size of precipitate in the sample of composition 1 (400Si+ 100YN) was about 0.7 x 0.9 pm and the size of precipitate in the sample of composition 3 (400Si+ 100Y203) was about 1.2 x 3/tm. The diffraction patterns of lattices indicate a significant difference between the effects of oxide and nitride additions as dopants. It appears that straining of the lattice associated with oxides is substantially more than that with nitrides. One theory accounting for the strained lattice is that the higher thermodynamic stability of oxides could prevent the interaction between oxides and the matrix and hence the straining of the matrix. The higher stability would also prevent the dissolution of oxide particles into the matrix. With the prolonged exposure to evalu- ated temperature (as encountered during process- ing and application procedures), oxide particles
Effect of intermetallic compounds on properties of tantalum
Table 5. Bend-ductility of 0.25-mm-diameter tantalum wire
may grow via mechanisms akin to Ostwald ripen- ing, which would thereby result in grain-growth.
4 CONCLUSION
It is apparent that the strengthening with inter- metallic dispersoids is more effective than oxide- dispersing strengthening. We believe that it is primarily due to the interaction between disper- soid and matrix; it enables a finer size of inter- metallic dispersoids to be maintained.
Our investigation has been limited to alloying with silicon and yttrium nitride. Yttrium nitride can be characterized as having a Gibbs-flee- energy value of 64.8 (taken as an absolute num- ber), which falls above a low-free-energy value of tantalum nitride of 52.4 and below a high value of yttrium oxide of 145 kcal/atom; the probability of forming yttrium silicide is greater under these thermodynamic conditions. Other methods of adding intermetallic dispersoids, as well as the use of other intermetallic dispersoids, should be investigated.
ACKNOWLEDGEMENTS
The authors acknowledge the assistance of Messrs R. C. Engleman, R. J. Neider, M. A. Fer-
reri, and R. W. Steele in the production and testing of samples. Mrs C. M. Yoder's help in preparing the manuscript is also appreciated.
REFERENCES
1. Cardonne, S. M., Kumar, P., Michalule, C. A. & Schuartz, H. D., Recovery and recrystallization in rolled tantalum single crystals. Metall. Trans. A, l0 (1979) 1031.
2. Tantalum, Columbium and Their Alloys, Cabot Cor- poration, Boyertown, PA, USA, 1986.
3. Vandermeer, R. A. & Snyden, Jr., W. B., The effect of microalloying with silicon on the properties of tantalum. J. Met., 10 (1989) 50.
4. Kumar, P., Moser, K. D. & Chatterjee, T. K., Tantalum and its alloys. Adv. Mater. Processes, (1992) 50-3.
5. Tantalum base alloys. U.S. E 5,171,379. 6. Dieter, G. E., Mechanical Metallurgy McGraw-Hill, New
York, NY, USA, 1987, pp. 188-220. 7. Anton, D. L., Snow, D. B., Favrow, L. H. & Giamei, A.
E, Dispersion strengthening of high-temperature niobium alloys, Report No. R89-917437-3. Issued by United Technology Research Center, East Hartford, CT, USA, 31 July, 1989 (Final Report). (Contract No. F49620-86-C-0053 for Air Force Office of Scientific Research, Building 410, Boiling Air Force Base, Wash- ington, DC, USA.)
8. Williams, D. M. & Smith, G. C., A study of oxide parti- cles and oxide-matrix interfaces in copper. In Proceed- ings of Symposium on Oxide Dispersion Strengthening, ed. G. S. Ansell, T. D. Copper & E V. Lenel. Gordon & Breach, London, 1968.
40 Prabhat Kumar, C. E. Mosheim
9. Kusunoki, K., Sumino, K., Kawasaki, Y. & Yamazaki, Y., Effect of amount of Y and oxide content on the secon- dary recrystallization temperature of nickel-base super- alloys. Metall. Trans. A, 21A (1990) 547.
10. McMasters, O. D., Gschneidner, K. A., Kaldis, E. & Sampietro, G., High-temperature enthalpies and stand- ard Gibbs free energies of formation of the europium cholcogenides: EuO, EuS, EuSe and EuTe. J. Chem. Thermodyn., 6 (1974) 845-57.
11. Miller, K. C., Thermodynamics Data for Inorganic Sub phides, Selenides and Tellurides. Butterworth, London, 1974.
12. Gschneidner, K. A., Kippenhan, N. & McMasters, O. D., Thermochemistry of the Rare Earth. Rare-Earth Infor- mation Center Institute for Atomic Research, Iowa State University, Ames, IO, USA, 1973.
13. deBoer, E R., Boom, R., Mattens, W. C. M., Miedema, A. R. & Niessen, A. K., Cohesion in Metals, North- Holland, Amsterdam, The Netherlands, 1988.
