Extensive progress and improvements have been made in the science and technology of gamma titanium aluminide alloys within the last decade. In particular, our understanding of their microstructural characteristics and property/microstructurc relationships has been substantially deepened. Based on these achievements, various engineering two-phase gamma alloys have been developed and their mechanical and chemical properties have been assessed. Aircraft and automotive industries arc pursuing their introduction for various structural components. At the same time, recent basic studies on the mechanical properties of two-phase gamma alloys, in particular with a controlled lamellar structure have provided a considerable amount of fundamental information on the deformation and fracture mechanisms of the two-phase gamma alloys. The results of such basic studies are incorporated in the recent alloy and microstructure design of two-phase gamma alloys. In this paper, such recent advances in the research and development of the two-phase gamma alloys and industrial involvement are summarized.

Many intermetallic alloys with L12, L10 and BCC related structures exhibit yield-stress anomalies, which have strong influence on their mechanical properties at intermediate temperatures. A short review is made of the anomalous glide systems in various intermetallics. They are shown to exhibit several common characteristics, which allow it to classify them into three groups with possibly similar dislocation mechanisms. In the second half, the well-known case of octahedral glide in Ni3Al is considered again, and a new model is proposed, based on the formation and the unlocking process of incomplete Kear-Wilsdorf locks by a double cross-slip mechanism. This model is supported by several microscopic and macroscopic observations including “in situ” straining experiments in a TEM. It accounts for the positive temperature dependence of the yield-stress with fairly good orders-of-magnitude estimates of the stress. It also explains satisfactorily the small stress-strain rate sensitivity and the transient effects.

Relationships between the microstructural organization in Ll2 alloys and the flow stress anomaly are discussed. Attention is paid to the origin of friction on the cube plane, to the stability of Kear-Wilsdorf (KW) locks and to the fine structure of kinks. The difficulty of dislocation multiplication and the easiness of their annihilation are outlined. Directions for a future model of the flow stress anomaly based on a close analysis of the microstructure are given.

The yield stress τy at small strains (≈ 0.01%) is strain rate independent, and has the same anomalous temperature dependence as that of the 0.2% strain. τy is considered to be the stress at which Frank-Read sources operate in a virgin crystal. For successful operation, τy must exceed the stress τs at which screws propagate dynamically through the crystal, and the source dislocation must pass rapidly through the unstable Frank - Read configuration. This can be achieved by the bowing edge dislocation overcoming local obstacles before reaching that configuration. Loops elongated along the screw direction are expected to be formed in the microstrain region. Under certain conditions such loops are unstable on unloading, thereby generating long edge dislocations which can operate successfully as sources at low but not at high temperatures, explaining the reversibility phenomenon.

An experimental study by TEM was made of the morphology of the antiphase domains formed when heavily rolled Q13AU is annealed at a temperature slightly below the critical temperature for ordering, Tc. Domains are formed at the advancing grain boundary with extremely small size and grow as recrystallization proceeds. From an early stage, domain walls show a preference for (100) orientation. The key question is raised whether domain formation during recrystallization entails the presence of a disordered zone at a moving grain boundary near Tc, and the conclusion is that such a zone is probably present. A provisional theory is constructed for the genesis of domains during recrystallization, taking into account the dragging force which newly formed domains exert on a moving grain boundary thereby diminishing the effective driving force for grain boundary motion, and a critical domain size is estimated which should completely inhibit grain-boundary motion. The intriguing fact that no domains at all are formed during the recrystallization of strongly ordered intermetallics such as Ni3Al is briefly discussed and a reason is proposed.

Micro-tensile specimens of coarse-grained Ti52at%Al crystals have been deformed in situ in a high voltage electron microscope at room temperature. In addition to some twinning, “simple” 1/2〈110] dislocations as well as superdislocations were moving, with the simple dislocations prevailing even if their orientation factor is lower than that of the superdislocations. Both types of dislocations are pinned, probably by small precipitates having a distance along the dislocations of about 100 nm. The precipitates consist most probably of Al2O3. Under stress, the dislocations bow out between the obstacles. The bowing is stronger for 1/2〈110] dislocations. An effective stress of about 41 MPa is estimated from their curvature. The kinematic behaviour of the dislocations is in accord with precipitation hardening. The dislocations are generated by the double-cross slip mechanism. Their density within the slip bands corresponds to a long-range internal stress of about 40 MPa. These data are consistent with the flow stress of PST crystals in the easy orientation, taken from the literature.

B2 CoAl single crystals with a number of orientations have been deformed at 1300 K and 1000 K. The deformation substructures have been characterized using TEM diffraction contrast techniques supported by image simulations. While <100> slip is found to be the predominant slip mode at these temperatures, <111> and <110> dislocations have also been observed. In particular, in hard orientations <110> dislocations appear to play an active role in initiating the deformation. Additionally, <110> dislocations are often observed as junctions of <100> type dislocations. Under the conditions tested, changes in alloy stoichiometry do not result in any modifications to the dislocation slip behavior.

Perturbed angular correlations of gamma rays (PAC) is being applied to study defects in ordered intermetallic alloys. Vacancies on both Pd and In sublattices in the B2 system Pdln were detected after quenching through quadrupole interactions induced at nearby 111In probe atoms. Fractions of probe atoms having each type of neighboring defect were observed to increase monotonically with quenching temperature over the range 8251500 K. For compositions close to 50.15 at.% Pd, nearly equal site fractions were observed for Pd and In vacancies, indicating that the Schottky vacancy-pair defect is the thermal defect at high temperature. The formation enthalpy of the Schottky defect was determined to be 1.3(2) eV through analysis of quenching data from in the range 825–1200 K. Above 1200 K, however, the vacancy concentration was observed to saturate at a value of 1.4(2) atomic percent, perhaps due to breakdown of the law of mass action for high defect concentrations.

Point defects such as vacancies and anti-site atoms are known to strongly affect the mechanical properties of B2 compounds. The variations in the hardness of these compounds with composition and quenching temperature can often be correlated with the concentrations of these point defects. Simple themodynamic models using a quasi-chemical approach can be used to estimate the concentrations of point defects with composition and temperature. These estimates can often be supported by experimental methods. Knowledge of the concentrations of defects can then be compared to the variations in hardness using solid solution strengthening models. A consistent relationship is found for several B2 compounds. The hardening rates of vacancies are found to be greater than those of anti-site atoms. B2 compounds having the anti-structure defect structure (AuZn, FeCo) are investigated as well as those with the triple-defect structure (FeAl, NiAl, CoAl).

Iron aluminides with the composition Fe-45Al-5X-0.2B-0.1Zr (at. %), where X stands for the first row transition metals Ti, V, Cr, Mn, Fe, Co, Ni, Cu, were examined at room temperature with respect to their strength, ductility, environmental sensitivity, and fracture mode. The extruded materials were annealed at 1273 K to produce similar grain sizes and subsequently at 673 K to reduce the amount of quenched in vacancies. All alloys were essentially single phase. Their solid solution strengthening was found to correlate with the atomic size misfit derived from the lattice parameters. The “binary” alloy Fe-45Al-0.2B-0.1Zr exhibited predominantly transgranular fracture. Ternary alloying additons with atomic numbers less than that of Fe tended to enhance intergranular fracture, whereas those with atomic numbers higher than that of Fe favored substantial amounts of transgranular fracture. Tensile testing in a partial pressure of dry oxygen increased the ductilities of the ternary alloys only slightly, whereas the ductility of the binary alloy increased from about 8 to about 19%. The ductilities in air correlated inversely with the yield strength. However, those alloys exhibiting substantial amounts of transgranular fracture always showed higher ductilities than those fracturing intergranularly. We interpret our fracture results in terms of yield strengths and ternary element site occupations.

A detailed examination of the variation of yield stress with temperature in an Fe3Al alloy shows a maximum at a temperature of about 500°C, slightly below the critical temperature for loss of DO3 order. At this temperature the dislocations present in the material change from being <111= superdislocations separated by APB to being single dislocations with Burgers vector <100=. At slightly lower temperatures the superdislocations become pinned by a local climb process involving point defect transfer between the partial dislocations.

Analysis of the forces between the dislocations which induce the local climb locking process allows an estimation of the role which will be played by variations in composition of the Fe-Al alloy considered, changes in deformation rate and orientation of the applied stress.

Examination of data available in the literature shows that each of the three aspects discussed, namely the influence of variations in ordered state, in Al content over the range 25% to 50%, or additions of alloying elements such as Si, straining at very fast or at slow rates, and stressing along different crystallographic axes, is completely consistent with the model proposed.

With an aim to contributing to the understanding of the mechanical properties of Fe3Al-based intermetallic alloys, Fe-30 at. % Al single crystals have been strained in a transmission electron microscope below and in the temperature range of the yield stress anomaly (at 300K and 573K respectively).

First-principles calculations have been used to investigate the defect properties and the site preference of FeAl and NiAl with ternary additions. It is found that a “triple-defect” structure model becomes invalid in describing the defect structure of FeAl (which is weakly ordered). The calculated mono-vacancy concentration on the Fe sites is lower than available experimental values by an order of magnitude, which may suggest the formation of defect complexes or the need of sufficiently long annealing time to anneal out quenched-in vacancies. For the site preference of ternary additions in FeAl, Cr and Ti are found to occupy Al sublattices, whereas Ni has a distinct preference for the Fe sites. The substitutional behavior of ternary elements in FeAl is consistent with the trend in the calculated heat-of-formation. For Fe addition in Al-rich NiAl, Fe atoms occupy Ni sublattices exclusively. The site preference of Fe addition in Ni-rich NiAl is dependent on alloy composition and temperature. At low temperatures, Fe atoms are found to preferentially occupy Al sublattices in Ni-rich alloys.

Fe3Al-based alloys exhibit poor room temperature ductility in the as-cast condition. In this study, the effect of grain refinement of the as-cast alloy on room-temperature ductility was investigated. Small melts of Fe-28 at. % Al-5 at. % Cr were inoculated with various alloying additions and cast into a 50- × 30- × 30-mm graphite mold. The resulting ingots were examined metallographically for evidence of grain refinement, and three-point bend tests were conducted on samples to assess the effect on room-temperature ductility. Ductility was assumed to correlate with the strain corresponding to the maximum stress obtained in the bend test. The results showed that titanium was extremely effective in grain refinement, although it severely embrittled the alloy in contents exceeding 1%. Boron additions strengthened the alloy significantly, while carbon additions reduced both the strength and ductility. The best ductility was found in an alloy containing titanium, boron, and carbon.

The effects of alloy composition and reaction atmosphere on reaction synthesis of binary Fe-Al alloys were studied. Reactions were observed in an open (air) furnace, under static vacuum (in an evacuated quartz tube) and in a dynamic vacuum furnace. High-speed videotapes of reaction syntheses of compacts formed from 45-μm Fe and 10-μm Al powders reacted in air and under static vacuum revealed that an unusual “two-stage” reaction exists in this system under these conditions. The first stage of the two-stage reaction lasts several seconds and starts at around 650°C. The second stage begins at about 900°C, reaching temperatures between 1250 and 1350°C. The progress of the reaction to the second stage is sensitive to the alloy composition and reaction environment. The reaction behavior is explained in terms of thermodynamics and heat transfer, which control the delicate balance between heat accumulation and heat loss during reaction synthesis.

Neutron diffraction was used to measure the residual stress distribution in an FeAl weld overlay on steel. It was found that the residual stresses accumulated during welding were essentially removed by the post-weld heat treatment that was applied to the specimen; most residual stresses in the specimen developed during cooling following the post-weld heat treatment. The experimental data were compared with a plasto-elastic finite element analysis. While some disagreement exists in absolute strain values, there is satisfactory agreement in strain spatial distribution between the experimental data and the finite element analysis.

A variety of Fe3Al alloys including Fe-25Al, Fe-28Al, Fe-28Al-4Cr and Fe-28Al-2Ti (all in atomic percent) have been investigated by tensile test to find if these alloys could have superplasticity at elevated temperatures, the results revealed that all these alloys exhibited large elongations when the temperature is higher than 600°C. At 850°C , under appropriate initial strain rate, the elongation is all above 300%. For Fe-28Al-2Ti, the maximum elongation reached 585%. Maximum m values are all above 0.3. Initial grain sizes are bigger than 100μm but became finer after deformation. Fracture happened with necking but no cavities were found under optical microscope. Characteristics of this phenomenon were summarized and discussed.

Forging processes at two different temperatures are performed to examine the relation between the microstructure and room temperature tensile properties in a Ce doped Fe3Al-based alloy. Results show that the microstructure and the ductility are sensitive to the forging temperature before annealing treatment. Higher yield strength and ductility can be obtained through forging at a relatively low temperature of 750°C followed by annealing at 800°C and 500°C. It is suggested that the formation of non-equilibrium grain boundaries and banded subgrains within carbide-free areas along grain boundaries enhances the local plastic deformation and results in the improvement of ductility. During the initial deformation at room temperature <111> slip is predominant for both microstructures.

An empirical atomistic potential, fit to the structural and elastic properties of L10 TiAl within the embedded atom method (EAM), is used to simulate the mobility of two possible planar forms of a<101] dislocations in a model L10 compound. The two configurations examined were: the planar SISF-APB-CSF coupled (P core) and the decomposed 1/2<110]-SISF-SESF coupled (D core). Six different line orientations are considered for the P core: 0° (screw), 30°, 60°, 90° (edge), 120° and 150°. The ‘ideal’ friction stress at 0°K of a<101] dislocations in the P form is found to be a function of line orientation, with the close packed line directions, <101] (screw) and <110] (60°), having friction stresses ranging from 0.001–0.002μ. Previously calculated results on the friction stress of a/2<110] dislocations, using an identical potential are consistently higher than the friction stress of a<101] dislocations. Simulations of the interaction of glide strains with the D core for the 60° (line directions < 110]) and 120° (line directions <011]) orientations show that the Shockley partial trailing the SESF in the D core is strongly pinned. The dislocation moves by extension of SESF when glide stresses are applied with SESF as the trailing fault.

High Resolution TEM (HRTEM) observations of a dislocation in γ-TiAl are compared directly with atomistic calculations of dislocation structures performed with atomistic potentials in order to obtain an estimate of the Complex Stacking Fault Energy (γcsf). A value of between 470 and 620 mJ/M2 was obtained. HRTEM observations are presented of a Ti-52AI sample, containing a dislocation with Burgers vector 1/2<110> and 60° line orientation. This image is matched against images simulated from the outputs of Embedded Atom Method (EAM) simulations, using potentials that were fit to bulk γ-TiAl properties. Two atomistic simulation methods were employed in order to give the range of values for γcsf. In the first of these methods, three EAM potentials were used to simulate the stress-free core structure. These were fit so as to produce three different values of γcsf, all other properties being roughly the same as the literature values for γ-TiAI. All of these potentials produced cores that were more extended than the experimental observation. Thus a value of 470 mJ/M2, being the highest value of γcsf obtainable for the EAM potentials, is reported as a low limit estimate of γcsf for γ-TiAl. An upper limit estimate of the value of γcsf was obtained by applying an external ‘Escaig’ stress that forced the Shockley partials to further constrict, simulating the effect of an increase in γcsf, The preliminary value calculated from this procedure was 620 mJ/M2.