The deformation behaviour of two-phase titanium aluminide alloys with a lamellar microstructure of the intermetallic phases α2(Ti3Al) and y(TiAl) was studied. The interaction processes of dislocations and deformation twins, respectively, with the lamellar interfaces are investigated by conventional and high-resolution electron microscopy. The mechanisms of translation of shear deformation across, the lamellar boundaries depend on their structure. Semicoherent interfaces are very effective barriers limiting the propagation of shear deformation. The misfit dislocations present at these interfaces support, on the other hand, the generation of dislocations and deformation twins. The observed processes are discussed regarding plastic deformation and crack propagation in the material.

In lamellar TiAl the flat-plate geometry of the grains, the barriers to deformation across the grain boundaries and the coherency stresses all contribute to a marked anisotropy in the yield and fracture stresses of the material. Both yield and fracture occur at low stresses when the deformation is within the lamellae (soft mode) and they occur at high stresses when the deformation crosses the lamellae (hard mode). The anisotropy is enhanced by a new effect which can soften the soft mode and harden the hard mode: the geometry of the lamellar boundary produces degeneracies in the planar fault energies at the interfaces which enhance the mobilities of dislocations on these interfaces. These degeneracies modify the core structure of dislocations on or near the interfaces, consequently soft mode dislocations can dissociate widely and move more easily when their glide plane is contained in the interface. Hard mode dislocations can substantially reduce their core energies when intersecting a γ/γ interface, and subsequently become immobilized, by cross slipping on to the interface plane. This paper presents a discussion of the geometry and relative energies of the γ/γ interfaces using elements of Bollman O-lattice theory. In order to investigate the influence of the interfaces on dislocation core structure we have fit an empirical Embedded Atom Method (EAM) potential to the structural and elastic properties of bulk L10 TiAl. The mobility and core structure of the twinning dislocation at the 180° interface and the perfect, 1/2<110] screw dislocation at the 60° and 120° interfaces were calculated using molecular statics within the EAM. We have also studied the influence of one and two atomic step ledges on dislocation mobility in the 120° interface. We find in general that dislocations are more glissile on the γ/γ interfaces, as compared to bulk TiAl and that ledges are weak barriers to dislocation glide. The interfaces themselves are strong barriers to dislocation motion in the hard mode. We find that the 1/2<110] screw dislocations gliding on conjugate {111} planes are trapped at these interfaces, as a result of lower core energies for screw dislocations lying in the interface.

Superplastic ductilities may be achieved in tension using polycrystalline materials with small grain sizes (typically <10 μm). Experiments were conducted to evaluate the significance of the grain boundaries in an Al-3% Mg solid solution alloy with a submicron grain size (∼0.2 μm). The material was produced by subjecting the alloy to severe plastic strain using a pressing technique. This paper describes the nature of the microstructure in the as-fabricated condition, the evolution of the microstructure as a function of time and/or temperature, and the effect of testing in tension at a temperature of 403 K.

We have examined grain boundaries in both undoped and boron doped Ni0.76Al0.24 using electron energy loss spectroscopy (EELS), x-ray fluorescence (EDX) and annular dark field (ADF) imaging in a UHV STEM. A detailed study of a high angle grain boundary in nickel rich Ni3Al doped with 1000 ppm boron shows nickel enrichment occurring in a 5Å wide region. Boron segregation to the boundary is observed with EELS and is seen to vary along the boundary, coinciding with ADF contrast changes in the surrounding grains that may be due to local strain fields. Spatially resolved EELS of the Ni L2,3 core edge, which is sensitive to changes in the hole density in the nickel d band, shows boron rich regions of the grain boundary to have a bonding similar to that of the bulk material. Boundary regions without boron have an electronic structure similar to that of the undoped grain boundaries where the Fermi level lies deeper in the nickel d band. In addition to studying boron segregation, EELS provides a unique opportunity to examine the changes in bonding that control the local properties of the material.

L10 alloys typically derive from a high-temperature, disordered fcc phase. For example, CuAu has a face centered tetragonal structure below 380°C and is derived from its high temperature, disordered face centered cubic phase. As the material transforms from the disordered fcc phase to the ordered tetragonal phase, the four distinct Σ3 fcc twin misorientations generate twelve distinct tetragonal twin misorientations, four being characterized as Σ3 and eight as Σ6. Of particular interest is the Σ6 structure because it is possible to terminate lattice antiphase boundaries without dislocations at this interface. A pure step defect at the interface can accommodate the APB termination due to anti-site coincidence (coincidence between copper and gold sites). We term these defects “antiphase steps”. The antiphase step is a new type of interfacial defect that has not been described by other workers. The possibility of forming antiphase steps in ordered L10 alloys is related to even-Σ interfaces. Since the Σ6 boundary is common in the ordered phase, the formation of dislocation-free APB terminations may be important in L10 alloys.

Dislocations were created near the center of the surface (101) of copper small crystals whose surfaces are (111), (111), (101), (101), (121), and (121) by use of n-body atom potentials and molecular dynamics. At first, a Heidenreich-Shockley partial dislocation was created., As the partial dislocation proceeds, the partial dislocation and the surface was connected with a stacking fault until the next Heidenreich-Shockley partial dislocation was created at the surface.Just before the creation of a partial dislocation the stress was the highest.

Dislocations were created near the center of the surface (101) of copper small crystals whose surfaces are (111), (111), (101), (101), (121), and (121) by use of n-body atom potentials and molecular dynamics. At first, a Heidenreich-Shockley partial dislocation was created., As the partial dislocation proceeds, the partial dislocation and the surface was connected with a stacking fault until the next Heidenreich-Shockley partial dislocation was created at the surface.Just before the creation of a partial dislocation the stress was the highest.

Small single crystals of titanium with and without stacking faults have been pulled by use of the molecular dynamics method. The tensile axis was [0001] and the stacking fault was introduced on (0001) pianes. The yield stress was higher in the crystal with stacking faults. The deformation was complicated in the crystal with stacking fault. Dislocations were created near the tip of a crack and moved on (1122).

Edge dislocations were created on a basal plane (0001) in a small titanium single crystal whose surfases are (0001), (0001), (1010), (1010), (1210) and (1210) by use of n-body embedded atom potentials and molecular dynamics.

Small-angle neutron scattering from grain boundary interfaces has been investigated in polycrystal palladium with and without deuterium dissolved in the solution phase. Polycrystalline samples were prepared using two different procedures: (1) single crystal material was recrystallized at 517°C after 78% cold working, and (2) an extruded polycrystal rod was investigated as received, or recrystallized at 600°C and 680°C. The expected 1/Q2 scattering profile (where Q is the neutron wavevector transfer) from grain boundaries has been observed, and the absolute cross sections permit a determination of the grain boundary surface concentration of missing palladium atoms and of trapped deuterium. The vacancy and trapped deuterium surface concentration determined in these measurements was 0.2-0.3 vacancies/Å2 and 0.4-0.6 deuterons/Å2, respectively. The measurements were extended to the low Q region (Q ≥0.006 1/Å) where the scattering response was dominated by Porod 1/Q4 behavior typical of internal voids. This behavior was especially strong for the samples fabricated from the extruded polycrystal, but was also present in the as-received single crystal. The degradation of the single crystal Porod response after compressional cold working is in reasonable agreement with the calculated effect of flattening a spherical void.

The stacking fault energy (SFE) is widely used to classify the mechanical behavior of pure metals. In alloys, however, the experimentally observed SFE is strongly influenced by localized solute effects. To further understand these effects on dislocation structure and on the observed SFE, solute segregation to an extended edge dislocation dipole, delineating two stacking faults, was studied in dilute Al:Cu, Al:Ag, and Al:Cu, Ag solid solutions. Cu and Ag were chosen to isolate solute size and modulus effects, Cu being smaller than Al, while Ag and Al are essentially the same size. Atomistic Monte Carlo results showed little change in the partial dislocation spacing in the binary systems as compared to the spacing in pure Al, even though Cu was observed to segregate to the compressive regions of the dislocation dipoles, forming widespread atmospheres, while Ag formed randomly distributed Ag-rich zones. However, in ternary Al:Cu,Ag simulations, the Ag apparently inhibited the Cu from distributing across the width of the extended dislocations, both Ag and Cu forming small clusters near or on the partial dislocations which increased the partial dislocation spacing. Results will be discussed in light of interpretations of experimental SFE determinations, emphasizing the importance of the localized solute distribution on the SFE.

The size evolution of θ phase (CuAl2) precipitates as a function of time was used to study the growth of grain boundary precipitates in an Al-5 wt % Cu alloy. The kinetics were: modeled using the Brailsford and Aaron treatment of the collector plate mechanism. It was found, for a given time, that the size varied in a reproducible manner as a function of misorientation between the grains. Precipitate size was found to vary from 5 x 10-6 cm to 5 x 10-5 cm, while misorientation varied from 20 to 50 degrees for a given heat treatment period. Grain boundary misorientation was determined to be the most important factor influencing precipitate size for a given grain boundary. The grain boundary plane orientation plays a secondary role in the growth of precipitates. From these data, the relationship between grain boundary misorientation and grain boundary diffusion has been determined.

The variation of θ (CuAI2) precipitate density as a function of heat treatment time and grain boundary misorientation was investigated in an Al-5 wt % Cu alloy. In this study, precipitate densities have been quantitatively linked to grain boundary structure. It was found that, for a given heat treatment time, the precipitate density varied in a reproducible manner (108 to 1010 ppts/cm2) as a function of misorientation between the grains (20 to 60 degrees). Additionally, misorientation was found to be the most important factor governing the precipitate density at a given grain boundary. The grain boundary plane orientation played a secondary role in determining the precipitate density. Bollmann O-lattice modeling allows comparison of the relative effects of grain boundary plane orientation relative to the misorientation between grains.

Using first-principles electronic structure calculations based on the Linear-Muffin-Tin Orbital (LMTO) method, we have investigated the effects of interstitial boron and hydrogen on the electronic structure of the L12 ordered intermetallic Ni3A1. When it occupies an octahedral interstitial site entirely coordinated by six Ni atoms, we find that boron enhances the charge distribution found in the strongly-bound “pure” Ni3AI crystal: Charge is depleted at Ni and Al sites and enhanced in interstitial region. Substitution of Al atoms for two of the Ni atoms coordinating the boron, however, reduces the interstitial charge density between certain atomic planes. In contrast to boron, hydrogen appears to deplete the interstitial charge, even when fully coordinated by Ni atoms. We suggest that these results are broadly consistent with the notion of boron as a cohesion enhancer and hydrogen as an embrittler.

We are reporting preliminary results on LMTO supercell calculations on a Σ3 (111) grain boundary (GB) in Fe with Mo atoms sitting at the GB plane. Calculations were performed both for the clean (CL) GB and GB with P/B atoms occupying centers of host (Fe and Mo) atom trigonal prisms (a coordination typical of the Fe GBs). The geometry of the GBs was not changed upon substituting Mo for Fe in the GB plane: all the interplanar distances and atomic radii were preserved (these were calculated earlier for the Fe GBs from GB3 relaxation modelling using the Embedded Atom Method). Comparison of the electron charge-density distributions within the atomic spheres for GBs with and without Mo, has shown that the presence of Mo at the GB plane increases both the number of d-electrons and the total charge density within atomic spheres of the atoms most important for intergranular cohesion. This should result in enhanced cohesion. The magnetic moments of the constituent atoms were also found to be strongly affected by the addition of Mo. The Mo atoms are anti-ferromagnetically coupled with the neighboring Fe atoms: mM0=-O.6OμB (CL GB), mMo=-O.14μB (GB with P), and mMo=-O.18μB (GB with B).

A computer simulation technique based on the Master Equation Method (MEM) is developed for modeling the spatial distribution of vacancies during ordering and subsequent domain coalescence and coarsening. A vacancy mechanism is assumed for the atomic diffusion and the single-site approximation is employed. It is demonstrated that vacancies strongly segregate into the antiphase domain boundaries (APBs) during coarsening, resulting in the vacancy concentration at APBs more than an order of magnitude higher than that inside the ordered domains. As the antiphase domains coarsen, the vacancy concentration at the APBs continues to increase and its spatial s segregation profile moves accompanying the APB migration. The effect of vacancy concentration on the antiphase domain coarsening kinetics is discussed.

Solid-State Amorphization (SSA) of crystalline interfaces is observed in the Ni/Ti multilayer system. The amorphization reaction nucleates at location(s) of crystallographic disorder, i.e. the multilayer interfaces. Microstructural analyses reveal the sputter-deposited growth structure to be epitaxial with semi-coherent interfaces. Strain energy originating from interface lattice distortions varies as a function of the multilayer repeat spacing. Therefore, the interfacial energy effects the onset conditions for SSA. Differential thermal analysis is used to measure the critical temperature Tc, to the nucleation of the SSA, which is found to vary with the Ni/Ti multilayer pair spacing.

A model is presented to explain various aspects of diffusion-induced grain boundary migration (DIGM). The driving energies of DIGM are identified as the free energy of mixing and the interface free energy, the former being predominant in most cases of DIGM. The grain boundary migrates due to thermally activated motion of atoms across the interface under the influence of the driving energies. An expression for the velocity of migration is derived. It is shown that this depends parabolically on the solute concentration, in agreement with experimental observations in the case of liquid film migration (LFM), which is analogous to DIGM. Furthermore, the velocity is proportional to lattice diffusivity ahead of the boundary. Recent results of enhancement of DIGM by ion bombardment is explained by radiation-enhanced lattice diffusivity due to introduction of point defects by the ions. The model also explains that diffusion-induced recrystallization (DIR) is due to a free energy decrease associated with the transformation from the amorphous phase in the grain boundary layer to the crystalline phase.

To provide a sensitive measurement of the effect of boron segregation on the strength and ductility of Ni3A1 grain boundaries, bicrystal tensile tests were performed on small specimens of boron doped Ni76A124 cut from extremely large-grained boules. Five specimens with the same “random” or low-symmetry grain boundary (disorientations measured by means of backscattered Kikuchi patterns) and two specimens with a second random grain boundary were tested in quenched and slow-cooled conditions. Duplicate tests performed in a low (7 ppm) water-vapor environment showed that the fracture mode and the stress and strain at fracture are altered by environmental embrittlement at individual, partially strengthened grain boundaries.

Defect-solute interactions control radiation-induced segregation (RIS) to interfacial sinks, such as grain boundaries, in metallic materials. The best studied system in this regard has been austenitic stainless steels. Measurements of grain boundary composition indicate that RIS of major alloying elements is in reasonable agreement with inverse-Kirkendall predictions. The steep and narrow composition profiles are shown to result from limited back diffusion near the boundary. Subsequently, defect-solute interactions that affect the near-boundary defect concentrations strongly affect RIS. The variability in measured RIS may in part be caused by grain boundary characteristics.