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.

2 CONSTITUTION, PROPERTIES, AND CURRENT STRUCTURAL USES INTRODUCTION Ordered alloys are alloys in which there are two or more atomic species and the different atomic species occupy specific sites in the crystal lattice. Such alloys tend to occur at well defined atomic ratios (i.e., AB, A3B, AB^, etc.). The tendency for segregation to the different atomic sites depends, in many cases, on temperature. For example, at low temperatures CuZn and Cu•jAu have the ordered structures shown in Figure l, but above a critical temperature (Tc), the atoms become randomly mixed on the different atomic sites and become random solid solutions. The ordered structure shown in Figure la is called the B2 (or L2O) structure and that in Figure lb, the Ll£ structure. The existence of ordered alloys as discrete entities was not established until 1839 (Westbrook 1967). However, for many hundreds of years ordered alloys have, in fact, been utilized for their interesting mechanical and physical properties. One may cite two examples: (1) dental amalgams, alloys of Ag, Sn and Hg that consist of a mixture of Ag£Hg3 and SngAg, were first described in 659 AD (Chu 1958), and (2) high tin bronzes, containing principally 6CuSn, were utilized by the Romans, and the Chinese as early as 1800 BC (Schweiz 1973). Westbrook (1967) cites other uses such as type metal based upon SbSn, ship sheathing made from $ brass, and statuary made from £ brass. Magnetic and superconducting materials based on ordered alloys also are well known, especially the Heusler alloys (Cu£MnAl) and various superconductors such as V•jGa and Nb^Sn. An early indication of fracture problems with ordered alloys in the eighteenth century may be found in the work of Geoffrey (1725), who noted a difference in the fracture characteristics of various phases of copper and zinc. One of the earliest systematic studies of the mechanical behavior of ordered alloys is that of Kurnakov and Zhemchuzhnii (1908). Tammann and

(a) (b) FIGURE 1 (a) At low temperatures: CuZn has the cesium chloride B2 long-range-ordered structure and (b) Cu3Au has the face centered cubic-based Ll structure.

Dahl (1923) showed that ordered alloys have a ductile-to-brittle transition temperature, and Lowrie (1952) later observed that the onset of ductility occurs in the temperature range T/^_ 'x* 0.61-0.68, where Tm is the melting point in °K. Westbrook and Wood (1963) showed that grain boundary embritclement is a common feature of many ordered alloys and that the embrittlement could be related to excess hardening near the grain boundaries. Ordered alloys [e.g., Fe-Mn-Al (Gibson I960)] were seriously examined as potential structural materials by many investigators in the 1950s, and subsequent years saw intensive research to examine their mechanical behavior. Beneficial effects of long range order in suppressing diffusion controlled phenomena (e.g., creep, recrystallization, and, in the case of the aluminides, oxidation) were identified in extensive reports by workers in the Soviet Union, Great Britain, and in the United States. However, little success was achieved in improving the ductility of polycrystals of many of the most interesting systems, most notably aluminides based on titanium, iron, and nickel. During the same period, it was noted that many of these alloys were ductile when all grain boundaries were eliminated. However, the use of single crystals in high-temperature applications such as gas turbines was not seriously considered until about 1970, by which time little additional research on ordered alloys was being carried out in this country. Rather than attempting to use ordered alloys in structural applications, metallurgists identified other uses for them in which brittleness was not a problem or, at least, could be minimized. One example was the use of nickel and cobalt aluminides as thin coatings on turbine hardware; another was the use of Fe-Si alloys (Sendust) as elements of sound reproduction systems. The use of high-permeability Fe-Co alloys in transformers also can b^ cited. In addition, ordered alloys such as Ni3Al (Y ) and Ni3Nb (Y ) were extensively used as distributed phases to strengthen nickel-base superalloys whereas Ni^Mo and other compounds were utilized as strengthening phases in maraging steels. In the case of nickel-base superalloys, the volume fraction of Y steadily increased from near zero in the 1940s to more than 60 percent in modern alloys. This was accomplished by adding increased quantities of aluminum as well as titanium, niobium, and tantalum, all of which are soluble in Y , and by vacuum melting. Interest in utilizing ordered alloys for structural applications was reawakened in this country when researchers at Wright-Patterson Air Force Base (together with contractors sponsored by the Air Force) discovered that ductility improvements could be achieved in TiAl- and Ti3Al-base alloys using a combination of powder metallurgy and alloying techniques. The development of rapid solidification methods (RSR and melt spinning, in particular) led to renewed interest in the iron and nickel aluminides. Finally, several exciting discoveries were reported that permitted substantial ductility to be achieved in cast and wrought ordered alloys previously known to be extremely brittle. The first of these accomplishments involved replacing cobalt in CosV with nickel (Liu 1973) and then iron (Liu 1979), leading to a series of face-centered cubic Ll2-type superlattices with extensive ductility at ambient

10 temperatures. It was found that ordered structures can be altered systematically through the control of electron-to-atom concentration in alloys. This can be achieved by partial replacement of cobalt and nickel with iron, which lowers e/a and stabilizes the ordered structure. (Fe,Co,Ni>3V alloys with the cubic ordered structure are ductile, with tensile elongation exceeding 35 percent at room temperature (Liu 1984). Shortly after the first reports of this achievement at ORNL, it was reported in Japan that polycrystalline Ni3Al could be made ductile by adding small quantities of boron (Aoki and Izumi 1979). Single crystals of many aluminides are quite ductile but their polycrystalline forms are not. The brittleness of such polycrystalline materials is associated with weak grain boundaries that cause brittle intergranular fracture without much plastic deformation within the grains. Small amounts of dopants, such as boron, and thermomechanical treatments can suppress brittle grain boundary fracture (Liu 1984). The critical composition range over which boron was beneficial was identified at ORNL (Liu and Koch 1982). These discoveries, together with the national search for replacements for strategic metals such as cobalt and, possibly, chromium and the need to develop energy-efficient propulsion systems, have in the past year or two stimulated much additional work—largely in the area of improving low temperature ductility and increasing high temperature strength—in this country, Japan, and Western Europe. Soviet publications on ordered alloys have been voluminous, but there is little visible evidence to date of Soviet progress in developing such alloys for structural applications; rather, Soviet research has emphasized alloy theory, yielding and strain-hardening phenomena, and the use of various imaging techniques to aid in the study of crystal structure and crystal defects. The characteristics that render ordered alloys so interesting as bases for alloy development will be discussed below as will their current status and applications. CHARACTERISTICS OF ORDERED ALLOYS The formation of long-range order in alloy systems frequently produces a significant effect on mechanical properties, including elastic constants, yield and tensile strengths, strain-hardening rates, ductility, and resistance to cyclic or static (creep) deformation (Lawley 1967, Stoloff 197l, Stoloff and Davies 1966). In some ordered systems, there is a change in crystal structure at the order-disorder transformation, leading to property changes dependent chiefly on stresses generated by the transformation. More widely studied and significant, however, are the properties of alloys in which: (1) there is no lattice change with ordering but only a rearrangement of atoms on lattice sites, (2) atoms rearrange from a body centered or face centered cubic to simple cubic, or (3) no order-disorder transition occurs below the melting point of the alloy. Table 1 identifies the most common superlattice types and presents information concerning dislocation types. Elastic Constants In some ordered alloys (e.g., Cu3Au), the elastic constants change discontinuously at TC (Siegel 1940); in others (e.g., 3-brass), there

11 io O •O 41 V4 0l M O 0) 6 O io 01 00 o o. M o 0) u o iH CO Q H .J id u 15-j •^ B of U S OL, o CO £ Cd H U U oi <d o hi CO 3 J•l •H oa o a 01 & Cd <d ^ u >j O 01 -« O. CO 4I s. • -j o 01 •H U CO *J Ol B •-" a a o SS •g.? Ol 10 5S V a. u 01 l-i •-. U 01 B •^ JS O 4J 3 4J <d u id -i JH 6 U, 3 00 oi ti •H a u ia 3 w 0) to .•. o z i I z + z IK z z z U. z t/> z z z aa illill A O- O V o A C4 vo to <o , . H H 0 fa -< CM V V o o eD a A CM H q o ol ;f <a i*i 0) B 01 H H O. •fJ •pJ •!-" •^ z z z z -^ B <e M S^ 3 Ol •-* 3 <*> <*> l* 01013 <*><*> <*> <* 'J30l).ri U<z 6 o u <d •a o o _ — en en ^ cu •r* 00 3 01 3•u _p <f U • u 9. £ O O O o id u I o CM 03 en s -• O 8 •3 3 u 4J « Ph CO u o BO • iJ «l B • B 01 O X B Sz a\ z H •- J! kl CO O 9 .a 0 J= j; 00 B .^ H 01 2 e <d u £ a Oi h .. « 01 01 u B B o z V? Z k^ <dl

12 is a continuous change in the elastic constants upon passing through Tc (Good 1941). Young's modulus for3~brass and related alloys increases continuously with degree of order. Usually, the change in elastic constants with order is small (in the range of 3 to 10 percent) and, therefore, changes in other mechanical properties arising from ordering are seldom controlled by effects on elastic constants. Rather, the principal effects of ordering on mechanical properties arise from changes in dislocation configurations due to the added constraints that ordering places on the nucleation and motion of dislocations. One of the attractive features of aluminides such as TiAl and Fe3Al in structural applications is the high specific modulus. For example, the specific modulus of a boride-strengthened Fe3Al alloy recently has been reported to be 50 percent higher than that of A-286 (a widely used Fe-base alloy containing Ni3Ti precipitates) between 25°C and 760°C (Figure 2) (Ray et al. 1983). Modulus data for several of the B2 aluminides (Wolfenden and Harmouche 1983) and TiAl and Ti3Al (Schafrik 1977) suggest that high stiffness/density ratios are characteristic of the aluminides. Some representative Young's modulii, melting (Tm) and critical (Tc) temperatures, and densities are shown in Table 2 for a number of ordered alloys, including the aluminides. TABLE 2 Young ' s Modulus Density Alloy Structure (106 psi) Tm(°0 TC(°C) (g/cc) TiAl L1o 25.5 1460 1460 3.91* Ti3Al D019 21.0 1600 1100 4.2** NiAl B2 42.7*. 1640 1640 5.86* Ni3Al L12 25. 95. 1390 1390 7.50* FeAl B2 37. 8^ 1250-1400 1250-1400 5.56* Fe3Al D03 20. 4£ 1540 540 6.72* CoAl B2 42. 7& 1648 1648 6.14* Zr3Al L12 19.61 1400 975 5.76* Fe3Si D03 39.41 1270 1270 7.25* Co3V hex 1400 1070 7.92** (Fe22Co78)3V (Fe60Ni40)3 L12 LI 2 .«._ 1400 1400 950 680 7.80** 7.60** (V96Ti4) *Calculated from lattice parameter data **Estimated. ISchafrik 1977 ^Wolfenden 1983 £Morgand et al. 1968 —H. A. Lipsitt unpublished IStoloff and Davies 1965 iTurner et al. 1978

13 36 32 1 ~ 28 in 24 MODULUS OF ELAST1C1TY RST 4.4 ALLOY (Fe•20.8 AI•4.1B) A•286 l l 0 200 400 600 800 1000 1200 1400 TEST TEMPERATURE (°F) FIGURE 2 Elastic moduli of Fe-20.8, A1-4.1B, and A-286 (Ray et al. 1983)- Superlattice Dislocations and Antiphase Boundaries Since an ordinary (unit) dislocation moving in a superlattice cannot recreate the crystal structure in its wake, disorder in the form of an antiphase boundary (APB) will result from the motion of such a dislocation (Marcinkowski 1963). The additional energy of the APB can be eliminated, however, by motion of dislocations in groups such that no net change in order occurs behind the dislocations. These groups, which consist of two or more dislocations connected by a strip of APB or other planar fault, are known as superlattice dislocations. Within the superlattice dislocation, each unit dislocation may further dissociate into its constituent partial dislocations (as listed for various superlattices in Table 1). Therefore, the stacking fault energy of the alloy plays a role in determining the dislocation fine structure. The appearance of slip bands intersecting a polished surface is quite different in alloys in the ordered and disordered states. Many ordered alloys reveal diffuse slip on only one or a few slip

14 systems and cross slip is restricted, thereby leading to brittleness, whereas coarse slip steps are observed in the disordered condition. It is the motion of superlattice dislocations and their interactions with each other and with obstacles such as grain boundaries, precipitate particles, or grown-in antiphase boundaries that control the mechanical behavior of ordered alloys. However, an exception to this behavior is noted in Ni^Mo, which rarely exhibits superlattice dislocations at room temperature. When superlattice dislocations are observed in Ni^Mo, they consist of five identical unit dislocations (Nesbit and Laughlin 1980). Antiphase boundaries, which are produced during heat treatment, separate domains that may be perfectly ordered within themselves but are out of step with one another. These thermal antiphase boundaries can be made to grow by appropriate annealing treatments. During the early stages of an isothermal ordering treatment, antiphase domains (APDs) generally are very small and may not be in contact with each other. However, with continued annealing, the APDs will impinge on each other and may grow to a size limited by the specimen size (in single crystals) or the grain size (in polycrystals). In some cases, (e.g., NiaMn and Ni3Fe), however, the kinetics of domain growth are very sluggish, and domain sizes are limited to about 500°A (Calvayrac and Fayard 1973). In some alloys, for example in Cu3Au (Ardley 1955), domains can give rise to considerable strengthening. Also, the nature of the domain structure can give important information about fault energies as discussed below. The antiphase boundary energy in L\2 alloys has been calculated to be anisotropic, with the lowest energy on {001} planes, taking into account first nearest neighbor (NN) interactions only (Flinn 1960). Accordingly, there is a driving force for an APB to lie on cube planes in these alloys, leading to important consequences with regard to yielding and strain hardening behavior. When this occurs, the electron microscope image of APB's shows a maze pattern (as in Cu3Au and Cu3Pt) rather than the "swirl" pattern that is associated with superlattices possessing isotropic domain networks (Ni3Fe, Ni3Mn, Fe3Al). The morphology of an APB in Cu3Au can be changed from a maze to a swirl pattern by adding 5 at% Ni (Yodogawa et al. 1980); this is perhaps due to a reduction in the electron/atom ratio and, hence, in the relative phase stability of Ll£ versus long-period superlattices in which periodic {001} APBs are characteristic. Recently, swirl pattern APDs have been noted in rapidly quenched NioAl-X alloys in which NioAl-type compounds are formed (Inoue et al. 1983) as well as in binary Ni^Al (Liu and Koch 1982); however, after short annealing treatments, these swirl patterns change to maze patterns and then quickly grow to large sizes (unpublished material, C. T. Liu, Oak Ridge National Laboratory, 1983). In Llo-type alloys, domains are in the shape of plates in contact with each other on {110} planes. Detailed studies of the role of such APDs in plastic deformation of Cu-Au alloys have been reported in the Soviet Union by Grinberg and co-workers (1976). The slip vector in B2-type superlattices varies from one alloy to another (Table 1) depending on factors such as APB energy, atom ratios,

15 atomic size ratios, and elastic energies of dislocations (Marcinkowski 1974). As in bcc metals,< 111> is usually the slip vector, and slip occurs on planes of type {110}, {112}, and {123}; however, several B2 superlattices display <100> slip. In LI2 superlattices, on the other hand, the slip plane may change with temperature, as in Ni^Al, which deforms on {1ll; planes at low temperatures and on (lOO) planes at high temperatures (Thornton et al. 1970). Cross slip from {111} to { 100} is believed responsible for an anomalous increase in strain hardening rate of Ci^Au with increasing temperature (Mikkola and Cohen 1966, Davies and Stoloff 1965, Duramoto and Pope 1976). The slip vector, however, is <110> at all test temperatures, but unit <110> dislocations probably dissociate into 1/6 a0 <112> partials. Departures from Stoichiometry Departures from perfect order cause strengthening through means not directly connected with the AP8. Deviations in stoichiometry from the ideal AB or AgB composition often lead to strengthening at low test temperatures, and, for this reason, slowly cooled alloys of the Fe-Co, Ni-Al, Ni-Ga, Mg-Cd, or Cu-Au systems reveal minima in room-temperature hardness at stoichiometric compositions. Hardness data obtained at 77°K and room temperature for compositions near stoichiometric Ni3Ga and Ni3Al are shown in Figure 3. Note the more rapid hardening with excess of the minority element. In systems with a very high ordering energy, the hardening at nonstoichiometric compositions might arise from lattice defects that are required to maintain the number of electrons per unit cell at a constant value required for phase stability (Guard and Westbrook 1959). In such cases, the vacancies should enhance plastic flow at high temperature, when the defects are mobile. Such behavior was reported for Ni3Al by Guard and Westbrook (1959), who showed decreased flow stresses at 800°C in alloys containing an excess of either Al or Ni. However, Noguchi and co-workers (1981) have demonstrated significant hardening at elevated temperatures in aluminum-rich Ni•jAl (Figure 4) and in gallium-rich Ni•jGa. Solute-induced hardening was attributed to changes in probability of cross slip from {1ll} to {100} plane with composition. Moreover, Aoki and Izumi (1975) showed that off-stoichiometric Ni^Al does not contain excess vacancies. It is clear that lattice defects as a function of stoichiometry are an important research subject that remains to be addressed adequately. Changes in stoichiometry influence not only the yield strength but also the dislocation substructure and fracture behavior. For example, unit dislocations are responsible for plastic deformation in hypo-stoichiometric Fe3Al whereas paired dislocations are found for Fe-31 at% Al (Ehlers and Mendiratta 1982). Further, hypo-stoichiometric Ni3Al and Fe3Al are more ductile than hyper-stoichiometric compositions. Yielding Ordering exerts a significant influence on initial plastic deformation as well as on strain hardening of many single crystals. Initial yielding

16 700 i— f (N d 600 600 400 300 200 100 A 23.5 it % Al D 24.1 _ O 24.7 • 25.3 • 25.9 A 26.6 J L J I 200 400 600 800 1000 1200 TEMPERATURE(K) FIGURE 3 The effect of deviations from stoichiometry on the flow stress of and Ni^Ga at 77°K and room temperature (Noguchi et al. 1981). 300 .£ 200 CN S 100 O • N- 1 1 23 24 25 26 CONCENTRAT1ON OF Al OR Ga (at%) 27 FIGURE 4 Effect of composition on the temperature dependence of yielding of binary Ni3Al alloys (Noguchi et al. 1981).

17 generally occurs at a much lower stress in the ordered condition, but the more rapid strain hardening of the ordered condition causes the ultimate strength to exceed that of disordered materials. The interaction of unit dislocations with short-range-ordered regions produced by typical "disordering" treatments (quenching from above Tc) gives rise to a higher initial flow stress than for fully ordered material in which superlattice dislocations interact with large APD. A similar behavior is noted in many, but by no means all, superlattices. For example, ordering increases the yield strength of Ni;jMn because of sluggish domain growth in that alloy (Johnston et al. 1965). Strength (and ductility) data for representative aluminides appear in Table 3. Note that strengths are generally not very high at room temperature. In the case of NiAl, the yield stress is very sensitive to strain rate and composition, especially below 800°C (Pascoe and Newey 1971). One of the many striking features of the plastic deformation of most ordered alloys, particularly a large number of those with the Ll2 structure (e.g., Ni3Al, (Fe,Ni,Co)3V, Ni^Ge, l^Si, Ni3Ga, Zr3Al), is an anomalous sharp rise in flow stress with increasing temperature, as was shown in Figure 4 for Ni3A1. The flow stress peak occurs in single crystals as well as in polycrystals arid its position with respect to temperature in Ni3Al is a function of crystal orientation (Kuramoto and Pope 1978) and alloy content (Thornton et al. 1970). Alloying with zirconium and hafnium is especially effective in raising the high temperature strength of Ni3Al; the yield strength of these alloys is reportedly higher at 850°C than those of all commercial superalloys (unpublished material by C. T. Liu and A. C. Schaffhauser, Oak Ridge National Laboratory, 1983). Also, an asymmetry between flow stresses measured in tension and in compression has been reported (Ezz et al. 1982). There is no anomalous behavior with increasing temperature in a few cases (e.g., Fe3Ge) (Suzuki et al. 1980). Since most advanced superalloys and several aligned eutectic (composite) alloys contain large volume fractions of Ni3Al, the yield and ultimate tensile strengths of these alloys often rise or remain nearly constant with increasing temperature to at least 700°C. Increases in flow stress with temperature also have been noted in long range ordered alloys of other cubic or hexagonal crystal structures. There is undoubtedly more than one mechanism responsible for this behavior since the peak strength is sometimes associated with the order-disorder temperature Tc or with the temperature of a transition from one ordered structure to another (e.g., Fe^Al); in other cases, it bears no apparent relation to a transformation (e.g., Ni3Al, $ CuZn, and Ni3Ge). A continuous range of behavior between anomalous and regular systems can be produced by alloying (e.g., by alloying Ni3Ge with Fe3Ge) (Suzuki et al. 1980). The addition of iron to Ni ^Ce causes the flow stress peak to disappear at an iron content of about 28 at%. Suzuki and co-workers (1980) have suggested that three factors control such behavior: APB and stacking fault energy on {111} and APB energy on {100} planes. A high anisotropy of APB energy between {100} and {111} planes is believed responsible for the strength anomaly in Ll£ alloys (Pope and Ezz 1984). This anisotropy leads to cross slip of superpartials onto {100} planes to minimize APB energy (Figure 5) (Rear and Wilsdorf 1962), thereby creating obstacles to further plastic flow. The obstacle density increases with increasing likelihood of

18 cd .0 CO .0 00 00 00 00 ON ON ON ON OO r-l r-l r-l r-l ON CO co 4J 4J CO CO 4J 4J r- 1 4•1 4J CO CO 4J 4J CO CO CO U rl rl •^ •^ JC •r* «H CO "B "B •o ^ CM ON B B CM •rJ 00 r-l -^ "^ o> oi oo 13 £ £ ON O 00 00 CO CO CO B r-l OO 9 ON ON 1- 1 oo 00 00 01 •T3 •O ON r-l CM r-l r-l •O 13 r-. ON ON ON B B JC r-l CO 00 B B J= NO r-l r-l r-l CO (0 U ON • • CO co U ON •O 0 • 4J r-l r-l r-l O i-l • • • Ol B 01 01 01 CO CO 01 0l ^ r-l r-l rH U CO Ol 01 CO CO B 0l 01 on 4-1 cd CO CO B •o oo 4J O 4J 4J r? r? B 00 rl Ol CO rl rl B ON 4J i-. 00 01 Ol ON O 4J 4J 4-1 rl rl 01 01 CO r-l Ol •rl r-l Ol 01 CO r-l 14-1 01 Q1 01 Ol Ol oo 3 r\ o 4J ^j B IW r-l CO CO 3301 CX £. 3 3 CO « 3 3 •^ 1^* ;>^ ^ 3 •H •rl 01 •ri O - CO CO CO CJ 6 J •r- CO CO •J CO OJ CO CO H O H H ti OH M 01 01 01 01 |k| oo DO 00 60 60 00 CO CO 3 ij co Ol Ol CO cO 01 01 01 _^ _^ ^^ ^^ CJ 01 CO CX CX CO CO CX CX CX CO -O 01 a a i Ol Ol 1 a i a a i i i 1 rl O U r-l CJ •H •rl 1 r-l r-l CJ 1 • r. 1 •H •H 1 1 i fa £ H CJ M a a \ CJ CJ H 1 n i o Q 1 1 i 1 \O r-l 60 3« 1T» vO ^^ • 1 • 1 1 1 •H s— ' • • * fa8-4I in f"^ ^3 <• f~ 1 0 0 1 1 1 33 CM NO o m CJ f"- U v-' m «* i 1 1 CO CO (0 O 0 oo O 0 O CM 00 ^ CO OH 1 1 1 O NO 1 00 1 1 1 1 1 O r^ o ON 1-4 \O H3 1 1 1 CO •* 1 1 1 1 t 1 rl CM in ON i— i ON CO ^ 1 1 1 r-l r-l 1 1 1 1 1 r-l r-l r-l 41 $.3 cd •H CM pLi 1 1 1 O O O o O 0 1 1 0 O 1 CO f** CO 4J • £ 1 1 _. ^D ^^ 1 CM ON CO 1 1 -* O 1 r-l ~3 r-l r. o O i 1 1 CO <fr 1 CM NO CM l m P^ 1 00 ON 00 V CX o £ CO"B 1 1 1 O O 1 m o 1 1 1 1 0 O O 1 | CM • -j. j 1 1 CO CO 1 CM O 1 1 1 1 CO CO 0 1 1 1— < 60^" 1 1 1 i r-i 1 1 l l 1-1 1 1 CO u •H 6 B 1 X C0| cdl C "1 «J1J O 1 JC • H 4J 4J u £ £ <j 4•1 01 •a A 60 •l-l 41 60 60 60 60 OH 2^ •o 4J •a 3 3 4•1 >•. 3 3 3 ~^ ^^ "•«^ B CO 3 06 06 0 O 1 r-l r-l 06 06 O O O 06 03 06 « ^ cS CO rl CO CO rl rl 1 £ £ £c2 CO CO rl r. rl CO CO *e 4J 06 06 3 3 1 06 06 3 3 3 OS 06 •o •H CM V I « 35 CM + 2= N O ^ + m n •r- — 8s? M-l r- t f-l <^ • 33 4J 33 1 ^> CM r-l 8v? o m r^ (0 « m r-i rH CO r- 1 ^3 4J ^ . ^J + • ^J -f' * 1 4J CO CO r-l r-l r-l r-l r-l r-l CM r-l r- l r- l 1- 1 r- 1 r-l ON «* CM i- 1 CO ^ + < ^J + *^ • • 1 ^ s m o co m •^ r- i m «* O r-l U-> CJ << CM 0 m o •* co m CM r-l O O < <t CO \O O -• U O O CO CM <! CM m m CO £ CM r- i o 1 1 i 1 co 1 l i CO 1 O 1 1 1 CO 1 1 1 CO 1 01 r* m oi 01 01 •H >H «rl Z Z Z •H •H H Z •H •H O 01 0I • r< •H 01 01 Z fa fa 01 • CQ fa CM ^• 01 Si H CM fa fa fa Z Z <n

19 STACK1NG FAULT TRACE OF (100) CROSS•SL1P D1SSOC1ATED SCREW D1SLOCAT1ON TRACE OF (111) FIGURE 5 Cross slip of leading dissociated screw dislocations from (111) to (001), favored by lower APB energy on (001) (Kear and Wilsdorf 1962). cross slip onto {100}, whether through raising temperature or testing crystals favorably created for cross slip. In the case of alloys such as FeCo-2% V, where the flow stress peak occurs at the order-disorder transition temperature, and Fe3Al, where the flow stress peak occurs just under the temperature for a transition between two ordered structures, the anomaly has been attributed to a transition from plastic deformation by superlattice dislocations at temperatures below that of the peak to deformation by unit dislocations at temperatures above that of the peak (Stoloff and Davies 1964a). However, this mechanism cannot apply to Fe3Al in which the aluminum content is less than 25 at%, since only unit dislocations have been observed in the DO•j condition. Fujita et al. (1983) have suggested that changes in the anisotropy between {010} and {111} APB energies produced by alloying can alter the mechanism of yielding in Cu3Au. Other mechanisms for the flow stress peak are discussed in the literature (Hanada et al. 198l, Van der Wegen et al. 1982), but it is clear that additional work on identifying dislocation configurations in the temperature range of the observed flow stress peaks is needed.

20 Strain Hardening Long-range-ordered alloys usually exhibit high strain-hardening rates compared to their disordered counterparts. For LI 2 super•lattices, the strain-hardening rate can double with order at temperatures near 298°K whereas lesser increments in rate are noted in cubic B2 or DO•jsystems (e.g., FeCo-2% V and Fe3Al); for hcp DO^g Mg3Cd, there is no change in strain-hardening rate with order at 298°K (Stoloff and Davies 1966). The role of superlattice dislocations in work hardening is controversial. For example, one model is based on the formation of "super-jogs" by the intersection of superlattice dislocations and the difficulty of dragging these super-jogs during continuing plastic deformation (Vidoz and Brown 1962). Since these super-jogs usually are not seen in thin foils of deformed alloys, an alternative mechanism based on the cross slip of superpartial onto {100} planes (as first suggested by Rear and Wilsdorf 1962) (Figure 5) also has been proposed to account for rapid hardening (Stoloff and Davies 1966). Experimental data obtained to date, including the observation that the rate of work hardening in ordered Cu3Au increases with temperature, generally favor the latter model; this point has been described in more detail in a recent paper by Pope and Ezz (1984). High strain hardening rates induced by long range order may permit attainment of very high strengths through cold working operation or thermal-mechanical treatments. Wear-resistance also should be enhanced by rapid strain hardening, permitting possible replacement of cobalt-base alloys for such applications. Fracture Polycrystalline ordered alloys often are brittle when tested in tension although considerable plasticity may be displayed by single crystals or by polycrystals tested in compression. In the Fe-Co system, brittleness and ordering fall in the same composition range, 35 to 65 percent cobalt (Stoloff and Davies 1966). The addition of a few percent of vanadium or chromium increases ductility of alloys tested in the quenched condition, perhaps as a result of the slowing of ordering kinetics (rapid quenching of binary FeCo always leaves some residual order). Recently, small additions of molybdenum, tungsten, tantalum, niobium, carbon, and nickel also have been shown to improve workability (ductility) of equiatomic Fe-Co alloys (Kawahara 1983). Each of the elements that improved ductility were capable of forming Co^X. compounds by diffusion, leading Kawahara to suggest that disruption of B2-type order occurs in the vicinity of Co.^X zones. He also noted that there is no correlation between the reactivity of solutes with interstitials, such as C, N, and 0, and ductility. A direct relationship between ductility and the degree of long-range order has been demonstrated in FeCo-2Z V (Stoloff and Davies 1966). Varying composition in binary alloys or adding ternary solutes influences the ductility of other alloys with the B2 structure. Solute effects may occur, in some cases, by interaction of solute elements with interstitial impurities or by affecting slip character. For example, nickel promotes wavy slip in 6-CuZn and reduces brittleness, whereas manganese suppresses

21 wavy slip and raises the ductile-to-brittle transition temperature (Shea and Stoloff 1974). In the Fe-Al system, a sharply reduced ductility is observed as the aluminum content approaches 25 at% (Figure 6) (Marcinkowski et al. 1975); alloys with 25 to 50 at% Al have usually been reported to be completely brittle at room temperature when processed by conventional ingot techniques. However, recent work has shown that the ductility of alloys with aluminum contents in the range 31 to 35 at% can be as much as 5 to 7 percent at room temperature (Chaterjee and Mendiratta 1981a, Ehlers and Mendiratta 1981). In addition, Fe-20 at% Al 5%Si exhibits a small amount of plasticity at room temperature; failure was by transgranular cleavage, with a transition to intergranular cavitation at higher test temperatures (Ehlers and Mendiratta 1982). Ti3Al and TiAl are brittle below about 600°C, as shown in Figure 7 for TiAl (Lipsitt et al. 1975), but some success in improving ductility through additions of germanium and zirconium have been reported. Only in the case of hcp Mg^Cd has long-range order been reported to increase ductility (Stoloff and Davies 1964b). In this alloy, slip occurs on more slip systems in the ordered than in the disordered state whereas in all other alloys studied to date, ordering either reduces the number of available slip systems or restricts cross slip of screw dislocations from one slip plane to another. For this reason, the brittleness of many polycrystalline ordered alloys has been attributed to a lack of sufficient slip systems to provide compatibility of deformation at grain boundaries under arbitrary states of strain. In the case of AgMg and the aluminides, the segregation of impurities to grain boundaries has been suggested to be the predominant factor (Westbrook and Wood 1963). However, recent work on high-purity NiAl and Ni•jAl has shown that impurity segregation in these alloys is minimal (private communication with E. R. Schulson, Dartmouth university, Hanover, New Hampshire, 1983); yet, they remain brittle. Considerable effort is being devoted to improving the ductility of several aluminides (FeAl, Fe3Al, NiAl, Ni^Al) through such diverse techniques as grain refinement through thermal-mechanical treatment, microalloying with boron, and various rapid solidification techniques. A particularly promising development is the demonstration of high ductility of Ni3Al doped with small quantities of boron (Aoki and Izumi 1979, Liu and Koch 1982). Boron segregates preferentially to grain boundaries (Liu et al. 1983). Unfortunately, the beneficial effect of boron seems to be lost at 25 at% Al (Liu and Koch 1982); Taub et al. (1984) report that some ductility can be achieved in rapidly solidified Ni-25 at% Al with 1%B. Schulson and Barker (1983) recently reported that an elongation of more than 40 percent can be achieved in Ni-49 at% Al at 400°C when the grain size is reduced to less than 20pm (Figure 8), while the ductility at 25°C by this treatment was 3 to 5 percent. (A summary of ductility data for aluminides at 25°C and 500°C is presented in Table 3.) There have been several reports in the literature of a beneficial effect of partial recrystallization on ductility of B2-type alloys. For example, fully ordered FeCo-2%V exhibits about 5 percent elongation or reduction in area at grain sizes in the range 12 to lOOym. When partially recrystallized, however, the ductility of ordered FeCo-2%V increases to about 20 percent (Stoloff and Dillamore 1970). Recently, Schulson (1982) has reported a small increase in ductility by partially recrystallizing

22 2 < Q OJ IE 8 12 16 20 24 ATOM1C PERCENT ALUM1NUM FIGURE 6 Ductility of Fe-Al alloys slowly cooled to 477°K and fractured in tension at that temperature (Harcinkowski et al. 1975). Ductility of TiAl M a Function of Tamparalur* Strength ot TiAl M • Function of T.mfMralura 10 200 400 600 800 TEMPERATURE tC) 1000 500 r- 400 ? 300 200 • 02* YWd Strm • FnKturt Strati O UltirraM I torn Strtnflh 1 1 1 I 1 1 1 200 400 600 800 TEMPERATURE (°C) FIGURE 7 Effect of temperature on tensile properties of TiAl (Lipsitt et al. 1975).

23 45 i— 40 35 3? 30 z o o z LJJ 25 20 15 10 •^ 1 1 1 1 1 1 1 20 40 60 80 100 120 140 GRA1N SIZE, d, (pm) FIGURE 8 Influence of grain size on tensile ductility of NiAl at 673°K; strain rate = 10~^ s"1 (Schulson and Barker 1983). NiAl. Ductility improvements can also be achieved through control of the microstructure by rapid solidification techniques such as melt spinning. It has been shown by Inoue and co-workers (1983) that melt-spun L\2 compounds of type Ni-Al-X (X=Cr, Mn, Fe, Co or Si) are both strong and ductile whereas their conventionally solidified counterparts are extremely brittle. The most effective strengtheners were chromium and silicon, which produced yield strengths of 490 and 590 Mpa at compositions of Ni67.5Al20Cr12.5 and Ni?8A110sH2. respectively. Silicon also has been reported to be an effective strengthener in conventionally prepared Ni3Al (Guard and Westbrook 1959). However, the high strength of the melt-spun alloys decreases significantly after subsequent annealing (e.g., after annealing at 1000°C for 1 hr), and the strength loss is accompanied by a drastic drop in ductility and a change from transgranular to intergranular cracking (Inoue et al. 1983). The change was attributed to an increased degree of long-range order and segregation of solutes to grain boundaries. Koch and co-workers (1982) also have been able to produce ductile

24 containing small additions of boron using an arc hammer apparatus, but the quench rates could not exceed a critical value. No ductility was achieved without boron. Taub and co-workers (1984), on the other hand, have not only reported that boron has a beneficial effect on the ductility of melt-spun Ni•jAl, but have, further, reported that powder-processed Ni3Al with 1% B also demonstrates appreciable ductility These apparently conflicting reports suggest again that small differences in composition may play a significant role in determining both strength and ductility. RSR processing via the PM route also has been employed to produce intermetallic compounds with improved ductility. For example, off-stoichiometric FeAl (35% Al) (Chaterjee and Mendiratta 1981a) with a room temperature ductility of 7 percent has been produced by the RSR process. An RSR alloy containing 50 at% Al, however, was brittle at temperatures up to 400°C, but high ductilities were observed at 500°C (Chaterjee and Hendiratta 1981b). For Fe-35 at% Al, fracture was by transgranular cleavage at room temperature while intergranular fracture occurred at 600°C. FeAl (50 at% Al) displayed intergranular failure at low temperatures (Chaterjee and Mendiratta 1981b). Fe3Al and FeAl containing sufficient quantities of boron to produce borides also have been prepared by RSR (Slaughter and Das 1980) and by melt spinning followed by pulverization and powder processing (Ray et al. 1983), but relatively little properties data have been reported in the open literature. A summary of all rapid solidification processing efforts reported to date appears in Table 4. TABLE 4 Rapid Solidification of Aluminides Alloys Method Remarks Reference Ni-Al-X Ni3Al-B FeAl-borides Fe-35 at% Al Fe-50 at% Al Ni3Al-B Fe3Al-TiB2 Melt spin Arc hammer Melt spin & PM RSR RSR Melt spin RSR Ductile at 25°C, unstable at 1000°C Ductile Brittle to 540°C TG at 25°C, IG at 600°C Inoue et al. 1983 Liu and Koch 1982 Ray et al. 1983 Chaterjee and Mendiratta 1982a Brittle to 400°C, IG Chaterjee and Mendiratta 1982b Ductility varies with B Finely dispersed borides Liu and Koch 1982 Slaughter and Das 1980

25 The benefits of rapid solidification on the ductility of aluminides at low temperatures have been suggested to be the following: 1. Elimination of macrosegregation and grain boundary segregation (both RSR and melt spinning). 2. Production of very fine grained material (^ 3ym grain size by melt spinning, l-2ym by RSR). 3. Reduction in the degree of order as an interim step during processing (melt spinning). However, the investigations reported above also have demonstrated certain disadvantages with respect to high-temperature mechanical properties: 1. Loss of strength and ductility of melt-spun Ni•jAl-type alloys after annealing. 2. Elevated temperature intergranular fracture in RSR-processed Fe-Al alloys (Chaterjee and Mendiratta 1982a, 1982b). 3. Inadequate stress rupture strength of melt-spun and PM-processed Fe-Al-B alloys at temperatures above 537°C (1000°F) (Ray et al. 1983). A further disadvantage of the melt spinning process is the restriction of product to thin sections (e.g., 0.1 mm thickness), and the sensitivity of the ribbons to surface imperfections produced upon solidification. However, these disadvantages can be eliminated if the melt-spun foils are pulverized and then processed by PM techniques, as has been done with Fe-Al alloys containing borides (Ray et al. 1983). This material is brittle at room temperature and, when inclusions are present, exhibits poor low cycle fatigue resistance at 650°C (unpublished material by A. K. Kuruvilla and N. S. Stoloff, Rensselaer Polytechnic Institute, Troy, New York, 1983). Liu and co-workers (Liu 1973, Liu 1979, and Liu and Inoue 1979) have utilized a combination of macro and micro alloying principles to achieve ductility in ordered alloys based on tetragonal Ni•jV and hexagonal Co3V without resorting to rapid solidification. The unmodified alloys are very brittle in the ordered state. Work on ternary (Co.Ni^V alloys showed that considerable ductility could be achieved by control of the ordered structure through alloying to control the electron/atom (e/a) ratio (Liu 1973), as first suggested by Beck (1969). In particular, replacing cobalt in Co3V with both iron and nickel produced an L12 superlattice. Several alloys of varying cobalt, iron and nickel content have been shown to be ductile at room temperature and to exhibit strengths superior to those of conventional single-phase alloys (e.g., austenitic stainless steels, Hastelloy X) at temperatures of 600 or 800°C (Figure 9) (unpublished material by C. T. Liu, Oak Ridge National Laboratory, Oak Ridge, Tennessee, 1983). Ductility problems at elevated temperatures have been alleviated through addition of small amounts of reactive solutes such as titanium and rare earths to the alloys. The effect of these is to suppress intergranular cracking. Figure 9 shows yield strength data for an advanced LRO alloy as well as for alloys based on Ni ,A1. Both the aluminide alloys and the LRO alloys show a substantial strength increase between room temperature and elevated temperatures, as is usually noted with Ll2-type superlattices.

26 THE YIELD STRENGTH OF ORDERED INTERMETALLIC ALLOYS INCREASES WITH TEMPERATURE UJ a: UJ >• 900 800 700 600 5OO 400 300 200 100 "ADVANCED LRO ALLOY 316 STA1NLESS STEEL- j I 1 l 130 120 HO 100 90 80 70 g <r 60 k 50 40 30 20 10 0 200 400 600 800 1000 TEMPERATURE (°C) FIGURE 9 Comparison of the yield strengths of advanced LRO alloy and nickel aluminides with commercial structural alloys (unpublished material by C. T. Liu, Oak Ridge National Laboratory, Oak Ridge, Tennessee, 1983).

27 Diffusion-controlled phenomena such as recovery, recrystallization, creep, and oxidation are suppressed by long-range order. The improvement in steady-state creep resistance with increasing order is manifested by an increase in the activation enthalpy, AH (Lawley et al. 1960, Hren and Sherby 1965). The steady-state creep rate of alloys often can be expressed as: ne -AH/RT e = A(|) where A is a constant, T is absolute temperature, R is the gas constant,a is the applied stress, G is the shear modulus, and n is a constant of the order of 4-6. Tracer experiments to determine activation energies for diffusion in ordered systems have provided values in close agreement with activation energies for creep. Since self-diffusion depends on vacancy interchange with atoms, the activation enthalpy is the sum of the enthalpies of formation and migration for vacancies. Both the enthalpies of formation and migration increase with the formation of long-range order, resulting in higher activation enthalpies for diffusion and for creep, and, therefore, lower creep rates. In the case of LRO alloys, however, formation of long range order causes only a small increase in AH, even though the creep rate is decreased by two orders of magnitude (Liu 1984). The stress exponent, n, has been shown to be highly stress-dependent in AgMg (Lexcellent et al. 1982) and disordered FeCo-2% V (Delobelle et al. 1979) and somewhat less stress dependent in ordered TiAl + 10% Nb (Mendiratta and Lipsitt 1980). Creep resistance of unalloyed NiAl (Vandervoort et al. 1966, Strutt and Dodd 1969) and l^Al (G. Leverant and D. Duhl as reported by Strutt and Dodd 1969, and Davies and Johnston 1970) tends to be poor relative to that of commercial high temperature alloys. However, alloying of boron-modified Ni jAl with hafnium and zirconium provides creep resistance comparable to or better than that of Waspaloy, a nickel-base alloy widely used in gas turbines (unpublished material by C. T. Liu and A. C. Schaffhauser, Oak Ridge National Laboratory, Oak Ridge, Tennessee, 1983). Advanced LRO alloys, in which aluminum replaces vanadium, also display good creep resistance, although not quite as high as that of l^Al+B (Table 5). These results, which were obtained at a very early stage of development of both classes of alloys, suggest the desirability of further intensive research on these systems. Limited stress rupture data for Fe-20.8w%Al-4.1%B (30.9 at% A1-15.2%B) show that at 538°C (1000°F) the 100 hour specific creep strength is comparable to that of A-286, an Fe-Ni superalloy (Ray et al. 1983). However, with increasing temperature the stress rupture strength of the very fine grained rapidly solidified Fe-Al-B alloy drops at a faster rate than for the conventionally processed, coarser grained A-286. Control of grain size is expected to be a problem in providing balanced creep and tensile properties of all rapidly solidified alloys. Experiments on the slow strain rate behavior of Fe-39.8 at% Al (B2 structure) produced by powder metallurgy indicate that extrusion temperature can affect high temperature deformation mechanisms. Grain size refinement

28 TABLE 5 Comparison of Creep Properties of Advanced LRO Alloys and Nickel Aluminides with Commercial Alloys Alloys Steady-State Creep Rate (10~6/h) Rupture Life or Test Time (h) At 20,000 psi, 760°C Boron-doped Ni3Al 34 Advanced Ni3Al aluminides 3 Advanced LRO alloys (Fe, Ni)3(V,Al) 7 Type 316 stainless steel 8,500 Hastelloy X l,300 At 40.000 psi. 760°C Advanced Ni3Al aluminides 20 Advanced LRO alloys (Fe,Ni)3(V,Al) 30 At 94,000 psi. 760°C MAR-M-246 133 500 6001 5001 65 200 6005. 6003. 1000 JLThe test was stopped (without rupture) at the time indicated. SOURCE: Private communication with C. T. Liu, Oak Ridge National Laboratory, 1983. to- about 10|jm produced effective strengthening to at least 0.75 of the melting temperature (Whittenberger 1983). Fatigue The suppression of cross slip or reduction in number of available slip systems with long-range order that occurs in most alloys suggests a diminished probability of crack nucleation under cyclic loading. In the few systems for which room-temperature fatigue data have been published (e.g., FeCo-2%V and Ni•jMn, ordering does, indeed, lead to an increase in high-cycle (stress-controlled) fatigue life (Boettner et al. 1966). The most detailed account of fatigue resistance in an ordered alloy is that of Williams and Smith (1966), who reported on the high-cycle behavior of brass at 25°C. The marked elastic anisotropy in this alloy produces severe stress concentrations at grain boundaries, leading to intercrystalline crack initiation. Williams and Smith (1966) also showed that the fatigue limit of this alloy is reduced when the sample is tested in a 3 percent aqueous solution of NaCl. Crack propagation data are not available in the literature for ordered alloys although recent work has shown that the crack growth resistance of an (FeNi)3V alloy is superior to that of conventional alloys at both 25°C, and 600°C (unpublished material by A. K. Kuruvilla and

29 N. S. Stoloff, Rensselaer Polytechnic Institute, Troy, New York, 1983) (Figure 10). Data on high-temperature fatigue resistance are few, with work reported for only four systems: TiAl (Sastry and Lipsitt 1977a), Ti3Al (Sastry and Lipsitt 1977b), (CoNi)3V (Ashok et al. 1983), and Cu3Au (Gittins 1968). In each case there is a tendency towards increased intergranular crack propagation as temperature increases. In the cobalt-base alloys this tendency can be reduced by doping with Ti (Ashok et al. 1983). There are even fewer data available on low-cycle fatigue (LCF) of ordered alloys. LCF resistance of Cu3Au is little affected by long-range order although fully ordered crystals cyclically harden much more rapidly than disordered crystals (Chien and Starke 1975). A cyclic strain-hardening exponent, n=0.36, has been reported, but this value is higher than that observed for most materials. Ordering has little effect on the fracture mode of Cu3Au; ductile fracture was observed under all conditions. Limited LCF data at 650°C for Fe-20.8 W%A1-4.1%B have recently been reported (Ray et al. 1983). Comparisons with several other engineering alloys, including an (Fe,Ni),V alloy (LRO-49) are shown in Figure 11. Environmental Cracking The previously described changes in slip character caused by long-range order, most notably reduced cross slip in alloys with high anisotropy in APB energy, suggest that there should be a significant effect of LRO on resistance to environmental cracking since increased susceptibility to stress corrosion cracking of some austenitic steels deforming by planar glide has been noted (Barnartt 1962) [although this association has been disputed (Saxena and Dodd 1966)]. A change in crack path from intergranular in low-Zn alpha brasses to transgranular in high-Zn alpha brasses also has been linked to a change in slip character (Swann 1963). Nevertheless, few studies of the influence of LRO on environmental cracking have been reported. In the case of Hastelloy B, attempts to link LRO with increased susceptibility to hydrogen embrittlement were inconclusive (Berkowitz and Miller 1980) although LRO had been suggested to be the cause of embrittlement (Asphahani 1977). Increased embrittlement (in the presence of hydrogen) of FeCo-2%V and an (Fe,Ni)3V alloy (LRO-42) when the alloys were ordered has been demonstrated by means of tensile, delayed failure, and fatigue tests (Kuruvilla et al. 1980). A possible explanation for an enhanced susceptibility to hydrogen embrittlement due to ordering is the possibility of the transport of hydrogen over long distances by superlattice dislocations confined to the original slip planes. However, no marked change in slip character with ordering has been noted by ordinary metallographic observations in these alloys. Recently, hydrogen embrittlement also has been noted in Ni3Al+B tested in tension at room temperature (unpublished material by A. K. Kuruvilla and N. S. Stoloff, Rensselaer Polytechnic Institute, Troy, New York, 1983). The embrittlement was shown to be reversible after an outgassing treatment at 200°C. Wear Resistance The pronounced effects of ordering on the mechanical and physical properties of alloys suggest that wear resistance also might be influenced by

30 1U • LR060- 600 C A Aitroloy, 575° C He 1nconel X •750, 650° C 10~5 — A Rene 95, 650° C D Nimonic901,550°C • Udimet 718, 550°C ID'6 O A288, 550° C Ti 1 10'7 ! z O io-10 — m-11 1 1 1 1 l 1 1 1 1 I 6 8 10 20 AK (MN m~3/2) 40 60 100 FIGURE 10 Comparison of crack growth data for LRO-60 (Fe-39.5w%Ni- 22.4V-0.4Ti-0.04Ce) in high purity argon with other high-temperature alloys: Astroloy tested in high purity argon, R = 0.05, V *= 10 Hz; Rene 95 tested in high purity argon, R = 0.05, V = 20 Hz (unpublished material by A. K. Kuruvilla and N. S. Stoloff, Rensselaer Polytechnic Institute, Troy, New York, 1983); Nimonic 90l, Udimet 718, and A-286 tested in air at R = 0.l, V = 40 Hz (Hoffelner and Speidel 1981).

31 CD LL o CM O Cl 1 O O O< CU OS <u ro .u oo 3 ON CO SS O •o 01 U a O ,—i r4 « •o V* co at c u W 3 h U Q1 u ,i• « •H 4J r-i C ca a1 £ C to U I-1 CO 9 C *J U CU r-l 'C O f - Pi U B •U QJ h U 03 V >, (D 1-1 «H 1-1 O £ 41 Pi N • <a U • ro C CD U 00 CD CD ON pj r-l • iH flj rH M CO CD • 1- CO •H r-l CO C h CD ji 01 a> r-4 QCi W 4•1 CD CD 01 > « B O rH T3 CD U r-l (11 PS > <o > 3 •H , ^ •-** . 3 ^j^ co ^ 3 I* O r-l (0 H U >N « 8 . •O I- V 01 3 H v•^ CO PS O CO O OO •» 3 01 G. .G C io 3 •* O 00 1-1 ON O •-> U C « 1-1 Q1 4J t3 3 C 4J CO •H PS § >.. 3 O O ^x r-i O r-l CO •H O B M 0 - -C "0 O (9V) O O

32 long-range order. It has been shown, for example, that the coefficient of adhesion is changed by long-range order (Bailey and Sikorski 1967). For Cu-Au and Pt-Co alloys, the coefficients are lower for the ordered conditions. The observed effects were related to order-induced changes in hardness. Similar observations have been made by Buckley (1965) for the frictional behavior of Cu-Au alloys in vacuum. Although little work has been done on the effects of long-range order on erosion resistance, work on Cu3Au (Wright and Mikkola 1976) has shown that the finer slip in the ordered material enhances the erosion resistance. CURRENT STRUCTURAL USES OF ORDERED ALLOYS Ordered alloys currently are used as single-phase (or as the continuous phase of multiphase) alloys in relatively few structural applications. Aluminide coatings based on NiAl and CoAl have been applied to nickel and to cobalt-base airfoils in gas turbines for about 15 years. In some cases these coatings, which were applied by chemical means (e.g., pack cementation), have been superseded by overlay coatings based on solid solution alloys such as NiCrAlY, FeCrAlY, and CoCrAlY. Platinum coatings have been used by the General Electric Company to form aluminides on turbine blade surfaces. Intermetallics as distributed phases constitute the most significant and unique feature of nickel-base superalloys. Conventionally cast and directionally solidified superalloys often contain 60 to 65 vol % Ni3Al, in which substantial quantities of titanium, tantalum, niobium, and other elements are dissolved. Such alloys provide the best combination of strength and oxidation/corrosion resistance available for temperatures from about 760 to 1100°C. Many directionally solidified nickel-base eutectic alloys contain y' particles to strengthen the y matrix, and some are reinforced by intermetallic fibers (e.g., y'-Sand y/y'-6 alloys in which 6=Ni3Nb and Co-CoAl alloys in which CoAl is the continuous phase). Although directionally solidified eutectic alloys have not been commercially manufactured, some have been engine tested successfully under relatively short time conditions and may in the future be selected for turbine applications, especially if production costs can be reduced. Zr•jAl was seriously considered as a fuel element sheath in Canada because of a combination of low neutron capture cross-section coupled with high strength and corrosion resistance comparable to that of Zircaloy 2 (Schulson 1974). However, development of this alloy ultimately was halted as a result of recognition of notch sensitivity and swelling upon irradiation. CONCLUDING REMARKS This chapter has shown that ordered alloys offer a number of unique properties that make them extremely attractive for structural use. Among them are a high specific modulus, especially at elevated temperatures, high strength at elevated temperatures, high strain-hardening rates, and low

33 self-diffusion rates with resulting low creep rates and high recrystallization temperatures. A major problem with most ordered alloys has been a tendency for low ductility; however, recent work performed in the United States and Japan has shown that there are a number of reasons for the brittleness of ordered alloys, and the reasons can be quite varied depending on the alloy system. Furthermore, it has been clearly demonstrated that once the reasons for the brittleness of a given alloy have been identified, the ductility can, in many cases, be dramatically improved. These improvements now make it possible to consider ordered alloys for a much greater range of structural uses. REFERENCES Ashok, S., K. Kain, J. Tartaglia and N. S. Stoloff. 1983. High cycle fatigue of (Fe,Ni,Co>3V type alloys. Met. Trans. A 14A:1997. Aoki, K., and 0. Izumi. 1975. Phys. Stat. Sol. (a) 32:657. Aoki, K., and 0. Izumi. 1979. Nippon Kinzoku Gakkaishi 43:1190. Ardley, G. W. 1955. Acta Met. 3:525. Asphahani, A. I. 1977. In Proceedings of the Second International Congress on Hydrogen in Metals, No. 4, Sec. C, Paper 2. New York: Pergamon Press. Bailey, J. A., and M. E. Sikorski. 1967. The effect of composition and ordering on adhesion in some binary solids solution alloy systems. Wear 14:181. Barnartt, S. 1962. Corrosion 18:322-31. Beck, P. A. 1969. Adv. X-Ray Anal. 12:1 Berkowitz, B. J., and C. Miller. 1980. The effect of ordering on the hydrogen embrittlement susceptibility of Ni2Cr. Met. Trans. A 11A:1877. Boettner, R. C., N. S. Stoloff, and R. G. Davies. 1966. Effect of long range order on fatigue. Trans. TMS-AIME 236:131. Buckley, D. H. 1965. Influence of Order-Disorder Transformation on Friction Characteristics of Copper-Gold Alloys in Vacuum. NASA TN D-2985. Washington, D.C.: National Aeronautics and Space Administration. Calvayrac, Y., and M. Fayard. 1973. Structural state and mechanical properties of polycrystalline Ni3Fe alloys. Phys. Stat. Sold. 17:407. Chaterjee, D. K., and M. G. Mendiratta. 1981a. Tensile Flow and Fracture of Sub-stoichiometric FeAl. J. of Metals 33(12):5.

34 Chaterjee, D. K. , and M. G. Mendiratta. 1981b. Ductile-brittle transition in stoichiometric FeAl. J. of Metals 33(12): 6. Chien, K. H. , and E. A. Starke. 1975. The effect of order on the low cycle fatigue response of ru ^Au single crystals. Acta Met. 23:1173. Chu, H. T. 1958. Chinese Med. J. 76:553. Davies, R. G. , and N. S. Stoloff. 1965. Influence of long range order on strain hardening. Phil. Mag. 12:297. Davies, R. G., and T. L. Johnson. 1970. The Metallurgical Design of a Superalloy, p. 447. In Ordered Alloys Structural Application and Physical Metallurgy. Baton Rouge, Louisiana: Claitor's Publishing Division. Delobelle, P., A. Mermet, and C. Oytana. 1979. Mat. Sci. and Eng. 41:203. Ehlers, S. K. , and M. G. Mendiratta. 1981. Tensile behavior of polycrystalline Fe-31 at% Al. J. of Metals 33(12) :5. Ehlers, S. K. , and M. G. Mendiratta. 1982. Tensile behavior of D03~ordered alloys based upon Fe-Al-Si. J. of Metals 34(12): 80. Ezz, S. S., D. P. Pope, and V. Paidar. 1982. The tension/compression flow stress asymmetry in Ni3(AlNb) single crystals. Acta Met. 30:921. Flinn, P. A. 1960. Theory of deformation in superlattices. Trans. TMS-AIME 218:145. Fujita, A., Y. Mishima, and T. Suzuki. 1983. Temperature dependence of strength in a Cu3Au-5% Ni alloy and its relevance to the APB morphology. J. Mat. Sci. 18:1881. Geoffrey, C. F. 1725. Med. Acad. Roy. Sci., p. 553. Gibson, R. C. 1960. Development of a Ductile, Oxidation Resistant Low-Density Alloy System for Intermediate Temperature Use. Final Report Under Contract No. a(5)60-6025-C. Syracuse, New York: Crucible Steel Company . Gittins, A. 1968. The effect of long range order on the high temperature fatigue behavior of Cu-Au. Met. Sci. J. 2:114. Good, W. A. 1941. Phys. Rev. 60:605. Grinberg, B. A., V. J. Syutkina, and L. I. Yakovenkova. 1976. Fiz. Metal. Metallo. oved. 42(1) :170. Guard, R. W. , and J. H. Westbrook. 1959. Alloying behavior of Ni3Al (Y' Phase). Trans Met. Soc. AIME 215:807. Hanada, S., S. Watanabe, T. Sato, and I. Izumi. 1981. Scripta Met. 15:1345.

35 Hoffelner, W., and M. 0. Speidel. 1981. Resistance to Crack Growth Under the Conditions of Fatigue, Creep Corrosion. Cost-50, Second Round, CH-1 Final Report. Baden, Switzerland: Brown Boveri. Hren, J. A., and 0. D. Sherby. 1965. The influence of long range order on some mechanical properties of an iron-cobalt alloy. Appl. Mat. Res. 4:41. Ino'ue, A., H. Tomioku, and T. Masumoto. 1982. J. Mat. Sci. Letters 1:377. Inoue, A., H. Tomioku, and T. Masumoto. 1983. Microstructure and mechanical properties of rapidly quenched II2 Alloys in Ni-Al-X systems. Met. Trans. A 14A:1367. Johnston, T. L., R. G. Davies, and N. S. Stoloff. 1965. Slip character and the ductile to brittle transition of single phase solids. Phil. Mag. 12:305. Kawahara, K. 1983. Effect of additive elements on cold workability in FeCo alloys. J. Mat. Sci. 18:1709. Kear, B. H., and H. G. F. Wilsdorf. 1962. TMS-AIME 224:382. Koch, C. C., J. A. Morton, C. T. Liu, 0. B. Gavin, and J. 0 Scarbrough. 1982. Paper presented at the Third International Conference on Rapid Solidification Processing. Gaithersburg, Maryland, December 1982. Kuramoto, E., and D. P. Pope. 1976. The Orientation Dependence of the Work Hardening Coefficint of Cu3Au. Phil. Mag 34:593. Kuramoto, E., and D. P. Pope. 1978. The orientation dependence of the yield stress of Ni3(Al,W). Acta Met. 26:207. Kurnakov, N. S., and S. F. Zhemchuzhnii. 1908. Z. Anorg. Chem. 60:1. Kuruvilla, A. K., S. Ashok, and N. S. Stoloff. 1980. Long Range Order and Hydrogen Embrittlement. In Proceedings of the Third International Congress on Hydrogen in Metals, No. 4, p. 629. New York: Pergamon Press. Lawley, A. 1967. Mechanical Properties-Plastic Behavior, p. 464. In Intermetallic Compounds, Ed. by J. H. Westbrook, New York: John Wiley and Sons. Lawley, A., J. A. Coll, and R. W. Cahn. 1960. Influence of crystallographic order on creep of iron-aluminum solid solutions. Trans. Met. Soc. AIME 218:166. Lexcellent, C., J. Birocheau, P. De lobe lie, A. Mermet, and C. Oytana. 1982. Study of the Theological behavior of the 3' phase of an equiatomic Ag-Mg alloy. Met. Trans. A 13A:251. Lipsitt, H. A., D. Shechtman, and R. E. Schafrik. 1975. The deformation and fracture of TiAl at elevated temperatures. Met. Trans. A 6A:1991.

36 Liu, C. T. 1973. Atomic ordering and structural transformation in the V-Co-Ni ternary alloys. Met. Trans. 4:1743. Liu, C. T. 1979. Development of ductile long-range ordered alloys for fusion reactor systems. J. Nucl. Mat. 85-86:907. Liu, C. T. 1984. Physical Metallurgy and Mechanical Properties of Ductile Ordered Alloys (FeCoNi^V, p. 168. International Metals Review. London: The Metals Society and Metals Park, Ohio: American Society for Metals. Liu, C. T. 1984. Design of Ordered Intermetallic Alloys for High Temperature Structural Uses. Paper presented at the A1ME Symposium on High Temperature Alloys, Theory anbd Design, April 9-1l, 1984, Bethesda, Maryland. Liu, C. T., and C. C. Koch. 1982. Development of Ductile Polycrystalline Ni3Al for High Temperature Applications. In Proceedings of the Conference on Trends in Critical Materials Requirements for Steels of the Future. Nashville, Tennessee: Vanderbilt University. Liu, C. T., and H. Inoue. 1979. Control of ordered structure and ductility of (Fe,Co,Ni)3V Alloys. Met. Trans. A 10A:1515. Liu, C. T., C. L. White, C. C. Koch, and E. H. Lee. 1983. Preparation of ductile nickel aluminides for high temperature use. In Proceedings of the Conference on High Temperature Materials Chemistry. Princeton, New Jersey: Electrochemical Society. (In press.) Lowrie, R. 1952. Mechanical properties of intermetallic compounds at elevated temperatures. Trans. AIME 194:1093. Marcinkowski, M. J. 1963. Electron Microscopy and Strength of Solids, p. 333. Ed. by G. Thomas and J. Washburn. New York: Interscience. Marcinkowski, M. J. 1974. In Treatise on Materials Science and Technology, Vol. 5, p. 181. Ed. by H. Herman. New York: Academic Press. Marcinkowski, M. J., M. E. Taylor, and F. X. Kayser. 1975. J. Mat. Sci. 10:406. Mendiratta, M. G., and H. A. Lipsitt. 1980. Steady-State Creep Behavior of Ti3Al-Base Intermetallics. J. Mat. Sci. 15:2985. Mikkola, D. E., and J. B. Cohen. 1966. Acta Met. 14:105. Morgand, P., P. Mouturat, and G. Sainfort. 1968. Structure et Proprie"tes Mechaniques des Alliages Fei—Aluminium. Acta Met. 16:867. Nesbit, L. A., and D. E. Laughlin. 1980. The deformation microstructure of the Ni-Ni4Mo system. Acta Met. 28:989. Noguchi, 0., Y. Oya, and T. Suzuki. 1981. The effect of nonstoichiometry on the positive temperature dependence of strength of Ni3Al and Ni3Ga. Met. Trans. A 12A:1647.

37 Pascoe, R. T., and C. W. A. Newey. 1971. Deformation processes in the intermediate phase NiAl. Met. Sci. J. 5:50. Pope, D. P. and S. S. Ezz. 1984. Mechanical Properties of Ni3Al and Nickel-Base Alloys with High Volume Fraction of Y'» P« 136. London: The Metals Society and Metals Park, Ohio: American Society for Metals. Ray, R., V. Panchanathan, and S. Isserow. 1983. Microcrystalline Iron-Base Alloys Made Using Rapid Solidification Technology. J. of Metals 35(6):30. Sainfort, G. 1967. Fragilite et Effects de L'Irradiation. Presses Universitaires de France 187. Sastry, S. M. L., and H. A. Lipsitt. 1977a. Fatigue deformation in TiAl base alloys. Met Trans. A 8A:299. Sastry, S. M. L., and H. A. Lipsitt. 1977b. Cyclic Deformation of Ti3Al. Acta Met. 25:1279. Saxena, M. N., and R. A. Dodd. 1966. Transgranular Stress-Corrosion Cracking Mechanisms in High-Purity Austenitic Stainless Steel, p. 455. In Environment-Sensitive Mechanical Behavior. Ed. by A.R.C. Westwood and N.S. Stoloff. New York: Gordon and Breach. Schrafik, R. E. 1977. Dynamic elastic moduli of the titanium aluminides. Met Trans. A 8A:1003. Schulson, E. M. 1974. J. Nucl. Mat. 50:127. Schulson, E. M. 1982. COSAM Program Overview, p. 175. NASA TN 830006 (October). Washington, D.C.: National Aeronautics and Space Admini stration. Schulson, E. R., and D. R. Barker. 1983. A brittle to ductile transition in NiAl of a critical grain size. Scripta Met. 17:519. Schweiz, B. 1973. A Guide to the Manufacture of Mirrors and Reflecting Surfaces. London: Pelham Books. Shea, M. M., and N. S. Stoloff. 1974. Plastic deformation of polycrystalline binary and ternary beta brass alloys. Met. Trans. 5:755. Siegel, S. 1940. Phys. Rev. 57:537. Slaughter, E. R., and S. K. Das. 1980. Iron-Aluminum Alloys with Titanium Diboride Dispersions by Rapid Solidification. In Proceedings of the Second International Conference on Rapid Solidification Processing, p. 354. Baton Rouge, Louisiana: Claitor's Publishing Division. Stoloff, N. S. 1971. Intermetallic compounds and ordered phases, p. 193. In Strengthening Methods in Crystals. Ed. by A. Kelly and R. B. Nicholson. New York: Elsevier.

38 Stoloff, N. S., and R. G. Davies. 1964a. The plastic deformation of ordered FeCo and Fe3Al alloys. Acta Met. 12:473. Stoloff, N. S., and R. G. Davies. 1964b. The effect of ordering on the plastic deformation of Mg3Cd. Trans. ASM 57:247. Stoloff, N. S., and R. G. Davies. 1965. On the Yield Stress of Aged Ni-Al Alloys. Trans. TMS-AIME 233:714. Stoloff, N. S., and R. G. Davies. 1966. The mechanical properties of ordered alloys. Prog, in Mat. Sci. 13:1-84. Stoloff, N. S., and I. L. Dillamore. 1970. Fracture of CsCl type superlattices, p. 525. In Ordered Alloys, Structural Application and Physical Metallurgy, Ed. by B. H. Kear, C. T. Sims, N. S. Stoloff, and J. H. Westbrook. Baton Rouge, Louisiana: Claitors Publ. Division. Strutt, P. R., and R. A. Dodd. 1969. Creep in Ordered Alloys, p. 475. In Ordered Alloys: Structural Applications and Physical Metallurgy, Proceedings of the 3rd Bolton Landing Conference, Lake George, N.Y., September 1969. Baton Rouge, Louisiana: Claitor's Publishing Division. Suzuki, T., Y. Oya, and D. M. Wee. 1980. Transition from positive to negative temperature dependence of the strength in Ni3Ge-Fe3Ge solid solution. Acta Met. 28:301. Swann, P. R. 1963. Corrosion 19:102. Tammann, C., and K. Dahl. 1923. Z. Anorg. Allgem. Chem. 126:104. Taub, A. I., S. C. Huang, and K. M. Chang. 1984. Improved strength and ductility of Ni3Al by boron. Met. Trans. 15A:399. Thornton, P. H., R. G. Davies, and T. L. Johnston. 1970. The temperature dependence of the flow stress of the Y* phase based upon Ni3Al. Met. Trans. 1:207. Turner, R. B., G. Wolgemuth, and E. M. Schulson. 1978. J. Nucl. Mat. 74:363. Van der Wegen, G. J. L., P. M. Bronsveld, and J. Th. M. De Hosson. 1982. Mechanical properties of the ordering alloy Cu3NiZn. Acta Met. 30:1537. Vandervoort, R. R., A. K. Mukherjee, and J. E. Dorn. 1966. Elevated temperature deformation mechanisms in Y'~NiAl. Trans. ASM 59:930. Vidoz, A. E., and L. M. Brown. 1962. Phil. Mag. 7:1167. Wee, D. M., 0. Noguchi, Y. Oya, and T. Suzuki. 1980. New L12 ordered alloys having the positive temperature dependence of strength. Trans. Japan Inst. Met. 21:237.

39 Westbrook, J. H. 1967. In Intermetallic Compounds, pp. ix-x. Ed. by J. H. Westbrook. New York: John Wiley and Sons. Westbrook, J. H., and D. L. Wood. 1963. A source of grain boundary embrit element in intennetallics. J. Inst. Met. 91:174. Whittenberger, J. D. 1983. Influence of thermomechanical processing on elevated temperature slow plastic flow properties of B2 aluminide Fe-39.8a%Al. Mat. Sci. and Eng. 57:77. Williams, H. D., and G. C. Smith. 1966. The fatigue of beta-brass. Phil. Mag. 13:835. Wolfenden, A., and M. Marmouche. 1983. Elastic measurements on transition metal aluminides. J. Metals 35(12):90. Wright, R. N., and D. E. Mikkola. 1976. Mat. Sci. Eng. 26:263. Yodogawa, M., D. M. Wee, Y. Oya, and T. Suzuki. 1980. The morphology of antiphase domains of Cu3?t and Cu3Au-Ni alloys. Scripta Met. 14:849.