Structural Uses for Ductile Ordered Alloys (1984)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

5 AVAILABLE DATA ON DUCTILE ORDERED ALLOYS ENGINEERING DATA Introduction For materials to be useful in the high-reliability, precision-engineered structures typical of advanced defense technology systems, a very large body of pertinent engineering data must be available. Detailed specifications to ensure that procured materials exhibit acceptable degrees of reproducibility and predictability also must be available. The types of engineering property data that may be required include: 1. Physical Properties a. Modulus b. Thermal expansivity c. Thermal conductivity d. Density e. Specific heat f. Poisson's ratio 2. Time Independent Mechanical Properties a. Yield and tensile strength b. Tensile flow behavior c. Ductility d. Influence of multiaxial stressing e. Toughness 3. Time and Cycle Dependent Mechanical Properties a. Creep and rupture strength b. Creep-rupture ductility c. Low cycle fatigue strength (strain and load controlled) d. High-cycle fatigue strength (various R ratios) e. Creep-fatigue (thermomechanical fatigue) f. Effects of multiaxial stresses 69

70 g. Slow crack growth due to fatigue, creep, creep fatigue, or static fatigue h. Effects of long-term thermal aging i. Size effect 4. Environmental Compatibility a. Oxidation b. Sulfidation behavior c. Stress corrosion cracking d. Evaporation rates in hard vacuum e. Susceptability to saline-bearing environment f. Erosion g. Corrosion by heat transfer fluids (liquid metals, water, steam, salts) h. Hydrogen embrittlement i. Tribology behavior at all operating temperatures j. Irradiation 5. Fabricability a. Ability to make required product forms b. Reproducible properties from production size lots for required forms c. Acceptable machinability d. Joinability (by welding, brazing, diffusion bonding) e. Acceptable formability f. Effects of form, forming, and joining on properties The relative importance of these properties change very substantially as a function of specific application. For example, in modest-temperature aerospace applications, density, modulus, tensile strength, and fatigue properties may be most important. In higher temperature applications, on the other hand, time- and cycle-dependent mechanical properties such as creep and rupture strength, creep-fatigue, and slow crack growth may be most important. Environmental compatibility requirements will vary greatly with application. In high-temperature jet engine applications, oxidation and sulfidation rates may be most important whereas in nuclear space power applications, liquid metal compatibility and evaporation rates in high vacuum may assume more significant roles. Clearly, for any of these applications, it is necessary for the materials to be available in the required shapes, and they must be joinable and machinable to the extent that this is a design requirement. In order to produce high-reliability designs it also is necessary to know the range of variability that can be expected in properties as a function of heat-to-heat variation, product form, etc. Typical requirements from MIL Handbook 5, in this respect, are shown in Table 9. For certain materials it also is necessary to assess properties as a function of directionality (anisotropy), which may vary considerably with product form. Overall, therefore, the qualification of an alloy for advanced defense system applications is a relatively expensive and time-consuming activity (see Figure 14) that is only undertaken when sufficient cost-benefit incentives exist.

71 TABLE 9 Typical MIL Handbook 5 Data Requirements Feature Required Specification Room Temperature Values Temperature Effects Fatigue Creep and Rupture Other Military, federal, or aerospace material specification Tensile, compression, shear and bearing tests on 10 lots from at least 2 production heats Tensile, compression, shear and bearing tests on 5 lots from 2 heats at each temperature No minimum data requirements At least 3 temperatures and 3 stresses covering 2 to 3 orders of magnitude on elapsed time for a minimum of 5 lots of material Effects of aging—properties of joints, physical properties, fracture toughness, stress corrosion Available Data on Ductile Ordered Alloys In the remainder of this section the discussion will be limited to the ductile ordered alloys recently developed at ORNL; specifically, most of the discussion will relate to the ductile Ni3Al-based alloys. The currently available data on the engineering properties of ductile ordered alloys are extremely limited. In fact, most of the available information is for experimental lots of material and essentially no engineering type data exist. The data that are available, however, indicate that some of the ductile ordered alloys appear to possess properties that offer a considerable engineering advantage. The density-compensated yield strength and ultimate tensile strength of the ductilized Ni3Al alloys, for example, increase with temperature and can achieve levels comparable to those of some of the highest strength alloys currently available (Figure 24 and 25) (Engineering Alloys Digest Inc. 1968; Menon and Reimann 1975; Conway and Stentz, 1980; and private communication with C. T. Liu, Oak Ridge National Laboratory, Oak Ridge, Tennessee, 1983). Strengths of this order, coupled with fairly good ductilities, are encouraging. The available data on creep and rupture behavior are very limited, but, again, encouraging. As indicated in Figure 26, the rupture strengths of the advanced LRO and aluminide alloys are well above those of many commercially available wrought alloys—and may approach the strength levels of some of the cast-nickel-base superalloys (Engineering Alloys Digest 1968; Menon and Reimann 1975; private communication with C. T. Liu, Oak Ridge National

72 260 i— 240 MAR•M•246 and Adv LRO Alloys 200 400 600 TEMPERATURE (°C) 800 1000 FIGURE 24 Comparison of the density-compensated yield strength of advanced LRO alloy and nickel aluminides with commercial structural alloys.

73 280 r— 260 AF115 200 400 600 800 1000 TEMPERATURE (°C) FIGURE 25 Comparison of the density-compensated ultimate tensile strength of advanced LRO alloys and advanced nickel aluminides with commercial structural alloys.

74 70 i— 60 50 40 30 LU 0 20 10 1000 Hours 800 900 TEMPERATURE (°C) 1000 1100 FIGURE 26 Comparative density compensated stress-rupture behavior of several alloys.

75 Laboratory, Oak Ridge, Tennessee, 1983; Aerospace Structural Metals Handbook, 1984). Available creep data are summarized in Table 5. The limited available data also indicate that the ductile ordered alloys can exhibit good fatigue behavior and good resistance to crack propagation (Figure 10 and 11). The alloys under consideration in this report offer the greatest apparent potential for applications where their good strength at elevated temperature can be used most advantageously. However, the good oxidation resistance at elevated temperatures of the Ni^Al-base alloys also may create important applications. Required Engineering Properties Development Program In view of the cost and time involved in developing the large body of data necessary to qualify a material for advanced defense system applications, it is recommended that a phased program be undertaken. The intent of the first phase is to produce sufficient data of a screening type to serve as the basis for engineering trade-off studies to assess the potential cost and performance advantages of these alloys in specific applications. If, at the conclusion of this phase, incentives appear sufficient, a second phase of data generation should be undertaken. This phase should include evaluation of the properties of some typical production lots of material and development of sufficient data for the performance of detailed engineering studies to precisely quantify cost and performance advantages. If clear incentives remain at the conclusion of this second phase, the final phase of detailed material qualification should be undertaken. This phase would include development of all information required for specifications, detailed statistical assessments of the material property variability, identification of the influence of fabrication on properties, etc. The activities recommended for each of these phases are as follows: Phase I—Screening tests on experimental alloys with emphasis on physical properties (modulus, expansivity, conductivity, density), tensile properties, creep-rupture, creep-fatigue, toughness, crack growth, and corrosion behavior (particularly sulfidation/hot corrosion). Phase II—Characterization testing of experimental and production materials representing several product forms. The emphasis should be on more extensive characterization of properties studied in Phase I as well as studies of effects of multiaxial stress states, product anisotropy, feasibility of joining, and compatibility with a wide range of relevant corrosion environments. Phase III—Extensive testing of multiple production heats representing all relevant product forms and processing routes. Development of sufficient mechanical property data to permit adequate statistical characterization of all required properties. Development of wide ranging product, process, and joining specifications. It is assumed that testing in Phase I will be on available experimental heats of material. In addition to the generation of basic physical and

76 tensile property data, creep rupture, creep fatigue, and cyclic crack growth information must be generated since these properties tend to determine the usefulness of materials in elevated temperature applications. The data generated should be sufficient to at least allow the alloys to be ranked, with respect to strength, relative to other currently available engineering materials. Useful test data probably will lie in the 700 to 1000°C temperature range with tests lasting up to perhaps 1000 hours. Preliminary characterization of toughness is also viewed as important in Phase I since materials with very poor toughness often are constrained in their engineering utility. Some measure of K or J toughness as a function of temperature in the room temperature to 1000°C range is desirable for this purpose. Similarly, there are some indications that the ductile ordered alloys may, in some cases, have an unusually high susceptibility to crack growth, and such susceptibility, if confirmed, might well limit the engineering utility of the materials. Accordingly, it is important to do some preliminary crack growth characterization at least under fatigue conditions. Finally, because some of the ductile ordered alloys of particular interest have a high nickel content, it is important to characterize their corrosion behavior—particularly in the saline, sulfiding, hot corrosion environments found in many gas turbine applications. Some preliminary screening testing in such environments should be performed. In Phase II, testing should move toward evaluation of production materials—as well as more extensive evaluation of earlier experimental lots. Data should be generated from at least one production-sized heat of material. It also should be assured that a sampling of applicable product forms is included. Sufficient data should be generated to allow some preliminary assessment of statistical behavior. In addition, preliminary studies of the effects of product form on property anisotropy should be performed and assessments of behavior under multiaxial loading conditions and the feasibility of machining and joining should be completed. By the time this phase is undertaken, some relevant applications for the alloy systems should have become clear; therefore this testing phase should also involve a preliminary assessment of the compatibility of materials with relevant corrosion environments. Metallurgical stability also must be demonstrated. The influence of long times at elevated temperature under stress on critical strength properties, ductility, and phase stability must be determined. Phase III basically involves generation of sufficient test data to fully qualify the material for high-reliability advanced defense system applications. Mechanical and physical properties should be determined from three or more production heats of material. This involves not only extensive mechanical property data generation but also the evaluation of a full range of applicable product forms, the development of detailed materials and process specifications, and so on. Reproducible properties and characteristics must be demonstrated from several heats of material using the ultimate product forms intended for use.

77 SCIENTIFIC DATA The incidence of order in alloys affects most properties, including the mechanical, electrical, and magnetic properties. Because the changes in properties can be large and because changes in order can be manipulated through control of composition and/or processing conditions, there has developed considerable potential to further exploit these alloys for practical purposes. The actual use of ordered alloys dates back many centuries; current scientific understanding has evolved over the past 60 years (Westbrook 1974). Despite this rich history of work with ordered alloys, there are many aspects of ordering behavior, and its relationship to properties, that remain poorly understood. The intent here is to highlight those areas of research and development that are deemed important to establishing the scientific base needed to further develop ordered alloys as useful engineering materials. An important emphasis of these observations will be the long-term potential of an improved science base to make the processes of structural alloy development more efficient and economical. In addition, it must be realized that the development of the currently available ductile ordered alloys grew out of scientific programs. The programs at Wright-Patterson and at Oak Ridge have depended very strongly on a scientific foundation that was painstakingly developed for many years before these new alloys were considered. Phase Diagrams Although an understanding of phase equilibria underlies all work with ordered alloys and will be discussed later in other contexts, it is important to emphasize that there is considerable need for classical phase diagram determinations for portions of many binary ordering systems as well as for the more complex, but highly practical, cases of ternary and higher order alloys on which little systematic work has been done. It also should be noted that, in many instances, earlier work should be examined in light of the current understanding of phase transformations, particularly the occurrence of precursor transformations and metastable transformation products. Future work will certainly be facilitated by the improved electron, x-ray, and other instrumentation and techniques currently available. The iron-aluminum system provides an example of the uncertainties that exist for many phase diagrams. There is only partial agreement about the low-temperature region encompassing Fe^Al and incomplete definition of the high-temperature regions (Figures 27 and 28). Also, little has been done to define the effects of ternary additions to Fe-Al alloys. More work is needed to define phase equilibria in binary as well as in ternary and higher order ordering systems. This type of work should be encouraged as an important adjunct to alloy development work. Degree of Order Ordered phases commonly exist in a less than perfectly ordered state. These deviations from perfect order can be expressed statistically as order

78 AI-Fe Aluminum•Iron TO 80 90 M I5OU ISM* MOOT |K/V\ ^ ^^ IOUU •4fV\ N \ I^OO 1SM' \ - \ I3OU b N n MOO' I9AA •(y•f«) i \«u ItUU •••*» i inn ^0»»l MIWI • IBB* | 1 i TT l»00» ifVVS 1 -M 42 \ '282 !»«• I V \ Q/1A *gp. 1 A \ fi/Vl »/.T»O*_ BOO ^ E 4 MOOT •«< CT »? < •yrtn 1UTMM HO.ST •4 fOO : 1 *M• I100» &F«) 1 MM csvi voo MTJ e|\MI.« ) <AJh— MBOr 9OO IN 6C i T 0 A lomicX u 4OO »oor [Ml > vN 1 1 : ' ^- 3 iSv^C'.. M" , ior 9OO MA I \ v^f y .-««• \ "•••, M toor \ / t r 2 ino 1 ZOO irut 1 0 45 63 | | 1 4 so 55 1% Al 60 IOU 1 ! W.-jh o Fl 10 20 30 40 50 60 70 60 9O Al Weight Percentage Aluminum FIGURE 27 Fe-Al phase diagram (American Society for Metals 1973)-

79 800 700 600 900 400 Might ptr c«tf Al 13 14 e DISORDERED »« • • S1NGLE PHASE f»U TYPE • SINGLE PHASE Ft,Al TYPE ««FtAl. TRANS1TION MANGE ---- CUR1E TEMPERATURE 20 21 22 23 24 25 26 27 28 29 30 31 32 atom ptr cent Al FIGURE 28 Fe-Al phase diagram compiled by Swann and co-workers (1972). For added detail see Allen and Cahn (1975, 1976a, and 1976b). parameters defining the sublattice occupancies within the structure. Experimentally, order parameters are determined most conveniently and directly by computing the intensity of superlattice reflections, which derive from the occurrence of the order, relative to the fundamental reflections, which are independent of order (Barrett and Massalski 1980). Where order is not long range and represents short-range deviations from a state of randomness, short-range-order parameters, or pair probabilities, can be used to describe the atomic arrangement. These can be determined from measurements of diffuse scattering (Cohen 1970). Of interest here is the fact that the properties of ordered phases are sensitive to changes in the degree of order. For many ordered alloy phases, changes in temperature may cause transformations to disordered phases or new ordered phases, but the degree of order may change well before the temperature of transformation. An additional area of interest relates to the comments made about phase diagrams. The solubility limit for alloy additions to binary-ordered phases is usually unknown, and, in many cases, the site occupancies associated with

80 particular alloying elements also are unknown. For example, it can be noted that in more complex structures, such as the VOy structure of Fe3Al shown in Figure 29 (refer to Figure 27), there are several types of sublattice sites available and that the addition of certain elements favoring particular sublattice sites may enhance properties while other choices may result in much smaller, or even deleterious, effects on properties. Of course, the solubility limits may vary considerably with the particular sublattice as well as with the competition for sites between solutes for the case of multiple alloy additions. For certain alloys, site occupancy information can be determined by means other than diffraction. One example is the use of Mossbauer spectroscopy to examine atomic configurations, as has been done with iron and titanium aluminides (Cranshaw 1977, Huffman and Fisher 1967). O TYPE I S1TES ON I SUBLATTICE © TYPE II S1TES ON II SUBLATT1CE O TYPE 111 S1TES ON 111 SUBLATT1CE A TYPE 1V S1TES ON 1V SUBLATTICE FIGURE 29 Generalized unit cell appropriate for description of 003, L2i, and B2 superlattices. B atoms are confined to Type II sites for AoB and DOo structure and to Types I and II for AB with B2 structure (Marcinkowski and Brown 1961).

81 In considering the degree of order it also is important to recognize the effects of deviations from stoichiometry on the properties of ordered phases. In a number of systems, changes in composition can cause formation of defect solid solutions where vacancies are introduced onto a sublattice rather than a wrong atom type. Similarly, the effects of alloy additions can vary greatly for positive and negative deviations from stoichiometry, as evidenced by the recent success of boron additions in ductilizing Ni^Al only for alloys with less than 25 at% Al (Liu et al. 1983). Finally, it must be recognized that the kinetics of ordering vary greatly among ordered phases and, further, that little is known about the effect of alloy additions on the kinetics. Because of the wide variety of processing conditions ranging from rapid solidification to conventional casting, with or without mechanical working, a high degree of order cannot be assumed and must generally be introduced by a specific ordering heat treatment. An improved understanding of the degree of order and ordering kinetics and how these are affected by chemical composition is needed. It is particularly important that the evaluation of properties include specification of the chemistry, degree of order, and detailed thermal-mechanical processing history. Dislocations, Antiphase Boundaries, and Stacking Faults Because ordering changes the translation vectors of the lattice, the motion of normal dislocations creates disorder. However, dislocations can move in combined arrays, or superlattice dislocations, separated by APBs or other faults so as to preserve order. It is the interplay between various superlattice and single dislocation processes as a function of temperature and other variables that determines the mechanical properties of many ordered materials. Complete characterization of the Burgers vectors and slip systems has only been done for certain ordered phases with much less information about changes caused by alloying. In those cases where dislocations dissociate into partial dislocations, the superlattice dislocations become more complex involving both APBs and a variety of stacking faults. The balance between the APB and stacking fault energies, therefore, determines the nature of the dislocations, the slip systems, and the mechanical properties through effects such as the restriction or enhancement of the ability to cross slip (Pope and Ezz 1984). Particular combinations of partial dislocations also can provide low energy faults and a means of forming deformation twins in ordered structures (Mikkola and Cohen 1966). APBs also can be formed thermally because alloy phases order by a random choice of sublattice for the ordering process. As these ordered regions, or antiphase domains, which differ only in sublattice choice, grow and impinge, the interface formed creates an APB. Depending on some unknown relationships between a variety of factors including the crystallography and the APB energy, these APBs can be planar or crystallographic in nature or assume random orientations. In addition to the important role of APBs in affecting dislocation behavior, such as cross slip, it is important to note that the kinetics of

82 APD growth and/or the thermal history can cause the APD size to be very small. This can influence both the properties and the ability to detect the order by some techniques. Also, as with any interface, there exists the potential for segregation to APBs. The interrelationships between the nature of the dislocations, stacking faults, and deformation-induced APBs probably determine the mechanical properties of ordered alloys. However, there are several areas of mechanical behavior that await a unified interpretation based on these relationships. For example, there is as yet no common basis for understanding the anomalous temperature dependence of the strength of many ordered phases (Pope and Ezz 1984), although the phenomenon appears to be reasonably well understood in Niyvl. Fundamental work on the deformation behavior of ordered alloys is needed. This includes not only the binary systems, but also ternary and higher order alloys where little systematic work has been done. In addition, more information is needed about APB energies, superlattice dislocations, and APD structures. Grain Boundary Characteristics Probably the most important factor limiting the engineering use of many ordered alloys, particularly those of interest for elevated-temperature applications, has been poor ductility. In many cases it appears that these ductility problems stem from segregation to grain boundaries and/or the intrinsic properties of grain boundaries. For example, single-crystal Ni^Al is quite ductile at room temperature, but polycrystalline Ni3Al is quite brittle. Recently Liu and co-workers (1983), following the lead of some Japanese workers (Aoki and Izumi 1979), have been able to ductilize polycrystalline Ni3Al by addition of boron, which probably segregates to the grain boundaries. In this work it has been suggested that the boron compensates for intrinsically weak grain boundaries in Ni3Al by acting as an electron donor to electron-deficient regions at the grain boundaries. Unfortunately, the addition of boron to other brittle aluminides appears to be of little benefit. It also should be noted that there is good reason to believe that many grain boundary effects are sensitive to stoichiometry. With Ni•jAl it was found that boron was effective only for compositions with less than 25 at% Al (Liu et al. 1983). Much remains to be learned about these effects, and it may be that the unique structural nature of intermetallics will create the potential for a substantial contribution to the general understanding of grain boundary effects. There appears to be good potential for improving the mechanical properties of many intermetallics under monotonic and cyclic loading through a better understanding of grain boundary effects. The importance of intrinsic grain boundary properties, as well as the fact that segregants can be beneficial or harmful, has been demonstrated. There is a strong need for more controlled experiments in this area.

83 Other Areas There are several other areas in which work is needed. These will be commented on briefly below. The enhancement of creep resistance and the interpretation of creep measurements, as well as other processes such as high-temperature oxidation and coating behavior, will require an improved understanding of diffusion in ordered alloys. Little quantitative information about diffusion in binary ordered alloys is available, and only a limited amount of work has been done on the more complex ternary and higher order alloys. Related to the area of diffusion is the need for additional work on point defects in ordered alloys. Point defects are important, not only because they occur in defect solid solutions formed at off-stoichiometric compositions in some phases, but also because of the role they play in determining ordering kinetics, general mass transport, and phase stability. Finally, one class of ordered alloys has been of scientific interest for some time, but it has not been exploited for practical use. These are the long-period superlattice phases in which a periodic APB structure of very fine spacing forms within the domains of the normal ordered structure (Barrett and Massalski 1980). The stability and occurrence of this type of order has generally been related to electron energy effects arising from the interaction between the Fermi surface and the Brillouin zone (Sato and Toth 1961 and 1962). The period, or APB spacing, at a constant temperature depends largely on the electron-to-atom ratio and can range from a few to several tens of atom spacings. Little has been done to evaluate the properties of these materials. The Potential Impact of Improving the Science of Ordered Alloys Current approaches to the discovery and development of new structural alloys are, of necessity, largely empirical and Edisonian in nature. This will continue to be the case in the near future, with the role of science being mostly interpretive rather than predictive. However, for a variety of reasons, not the least of which are the efficiency and cost of the current methods, it might be expected that other approaches will be sought and developed. Fortunately, it now appears that there is a great potential for science to make stronger inputs to alloy design and development. This has developed largely because of rapid advances in computing capabilities and in instrumentation for structural and chemical characterization. Of particular interest here are the tremendous advances to be expected from quantum mechanical calculations of ordered alloys. It is now generally agreed that these calculations soon will be sufficiently accurate to be useful as input to further calculations dealing with real alloys (Connolly and Williams 1983, Gyorffy and Stocks 1983, Stocks 1983, Williams et al. 1979, Williams et al. 1980). The results of the quantum mechanical calculations will generally yield the energies, lattice parameters, and elastic constants, at absolute zero, for perfectly ordered arrangements of two or more atom types. Energies also will be available for clusters of a few atoms within a lattice of "average" atoms. It will be important that

84 materials researchers take advantage of the results of these calculations by developing means of calculating the properties of real alloys under a variety of real conditions. In addition, it will be extremely important that a variety of parameters, such as phase transition temperatures as a function of composition, are available to facilitate refinement and testing of the calculations. Because ordered alloys can be defined structurally by experiment and, further, because the structures can be controlled through heat treatment, the impact of the theoretical calculations described above will be greatest for these alloys. It should be expected that the future will bring predictive capabilities to complement and act as an efficient screening tool for the usual methods of alloy development. REFERENCES Aerospace Structural Metals Handbook. 1984. Columbus, Ohio: Battelle- Columbus Laboratories. Allen, S. M., and J. W. Cahn. 1975. Acta Met. 23:1017. Allen, S. M., and J. W. Cahn. 1976a. Acta Met. 24:425. Allen, S. M., and J. W. Cahn. 1976b. Acta Met. 10:451. American Society for Metals. 1973. Metals Handbook, 8th ed., Vol. 8. Metals Park, Ohio: American Society for Metals. Aoki, K., and 0. Izumi. 1979. Nippon Kinzoku Gakkaishi 43:1190. Barrett, C., and T. B. Massalski. 1980. In Structure of Metals, 3rd Ed., p. 270. New York: Pergamon Press. Cohen, J. B. 1970. The order-disorder transformation. In Phase Transformations, p. 561. Metals Park, Ohio: American Society for Metals. Connolly, J., and A. Williams. 1983. Phys. Rev. 827:5169. Conway, J. B., and R. H. Stentz. 1980. High Temperature Low Cycle Fatigue Data for Three High Strength Nickel Base Superalloys. Report AFWAL-TR-80-4077. Wright-Patterson Air Force Base, Ohio: Air Force Wright Aeronautical Laboratory. Cranshaw, T. E. 1977. Physica 86:391. Engineering Alloys Digest, Inc. 1968. Alloy Data Sheet on MAR-M-246, Filing Code Ni-134. Upper Montclair, New Jersey: Engineering Alloys Digest, Inc. Gyorffy, B. L., and G. M. Stocks. 1983. Concentration waves and fermi surfaces in random metallic alloys. Phys. Rev. Lett. 50(5):374. Huffman, G. P., and R. M. Fisher. 1967. J. Appl. Phys. 38:735.

85 Liu, C. T., C. L. White, C. C. Koch, and E. H. Lee. 1983. Proceedings of the Conference on High Temperature Materials Chenri stry-2. Ed. by Z. A. Munir. J. Electrochemical Society 83-7:32-41. Marcinkowski, M. J., and N. Brown. 1961. Acta Met. 9:764. Menon, M. N., and W. H. Reimann. 1975. Met. Trans. 6:1075-85. Mikkola, D. E., and J. B. Cohen. 1966. Acta Met. 14:105. Pope, D. P., and S. S. Ezz. 1984. Mechanical Properties of Ni3Al and Nickel-Base Alloys with High Volume Fraction of Y", p. 136. International Metals Review. London: The Metals Society and Metals Park, Ohio: American Society for Metals. Sato, H., and R. S. Toth. 1961. Phys. Rev. 124:833. Sato, H., and R. S. Toth. 1962. J. Appl. Phys. 33:3250. Stocks, M. 1983. The Theory of Alloys: From the Schrb-dinger Equation to the Rolling Mill. In Review, Vol. 16, No. 4, pg. 42. Oak Ridge, Tennessee: Oak Ridge National Laboratory. Swann, P. R., W. R. Duff, and R. M. Fisher. 1972. Met. Trans. 3:409. Westbrook, J. H. 1974. Applications of atomic ordering reactions. In Order-Disorder Transformations in Alloys, p. 494, Ed. by H. Warlimont. New York: Springer-Verlag. Williams, A. R., J. Kubler, and C. D. Gelatt. 1979. Cohesive properties of metallic compounds augmented spherical wave calculations. Phys. Rev. B 19:6094. Williams, A. R., C. D. Gelatt, and V. L. Masuzzi. 1980. Microscopic basis of Miedemce's empirical theory of transition metal compound formation. Phys. Rev. Lett. 44:429.