The study of phase transformations involves making assumptions about interfaces within a material in order to apply mathematical models. An example of this is the Gibbs classical theory of nucleation and growth which is based upon the assumption that the precipitate / matrix interface is sharp. Recent developments in atom probe microscopy have made it possible for the first time to characterize complex three-dimensional internal interfaces within materials to sub-nanometre accuracy. We have used the OPoSAP (Optical Position Sensitive Atom Probe) to characterize the precipitate / matrix interface in the dilute Cu-Co system which is a model alloy for the study of homogeneous nucleation and growth. Interface widths derived from radial composition profiles were measured to be 0.9 ± 0.2 nm in size. The effects of thermal energy, positioning inaccuracies, and statistical limitations on the measurement of interface widths are examined through the use of mathematical models and computer simulations. Similar widths are measured for precipitates in the coarsening regime which shows the interface velocity does not affect the interface width.

The utility of polysynthetically-twinned (PST) TiAI, which contains a high density of parallel, atomically-flat interfaces within a set of identical crystallographic orientations, as a potential model system for a detailed investigation of interface diffusion is explored. Macroscopic PST crystals were grown in an optical float zone furnace. Thin films were cut from oriented crystals and polished with <112> directions normal to the film. After sputter cleaning, Ag was deposited on one side of the TiAI thin films. Auger spectra were obtained from these films over a wide range of sputter/anneal conditions. The Al and Ti concentrations were analyzed as well as the important impurity elements, S, Ar, C, N and O. Using the present data and existing knowledge of the microstructure and crystallography of PST TiAI, the potential of this material for providing a detailed understanding of the atomistic mechanisms of interface diffusion is analyzed.

The quenched microstructures of self-propagating reactions in Al/Ni and Al/Monel foils are examined. The quenched reaction fronts in both films are shown to be approximately 30μm long and exhibit a continuous evolution from an as-deposited to a fully reacted microstructure. In both multilayers the first phase to form is isostructural with Al9Co2, and nucleates on the Al side of the layer interface. Growth of the intermetallic occurs in a two step process. First the grains grow laterally along the interface until they impinge on each other, followed by growth normal to the interface into the Al layer. It is shown that planarity of this interface during the latter growth stage is determined by the width of the intermetallic grains immediately prior to their growth normal to the interface into the Al layer. On the basis of these observations it is argued that in modeling these reactions diffusion through the intermetallic during the reaction must be considered, and that atomic diffusion may be considerably more three- dimensional in practice than the current models suggest.

Multilayer samples of Nickel and Aluminum with an overall composition of Al-20Ni were prepared by cold rolling of elemental foils. The sample microstructures and phases were characterized by XRD, SEM and TEM/SAED, and the reactive phase formation was then examined by DSC measurements. XRD, SEM and TEM measurements show that the rolling procedure results in a decrease of the Al and Ni layer thicknesses (down to 100 nm in average) and a decrease of the grain size (down to less than 50 nm). No phase formation is observed during the cold rolling procedure. In isochronal DSC scans of the Al-Ni multilayers, the formation of the Al3Ni phase was found to be a two step reaction process due to 2-dimensional nucleation and lateral growth and a 3-dimensional phase thickening. While XRD measurements showed Al3Ni as the only phase that forms, more detailed TEM investigations of the samples after DSC treatment also showed a small amount of an amorphous Al-Ni phase, formed by a thermally activated solid state amorphization reaction (SSAR). In-situ TEM heating of the amorphous areas under the electron beam in the microscope yielded the crystallization of the amorphous phase to a B2 structure and a growth of the B2 grains up to 100 nm in size.

Structural stability in thin-film multilayers is described in terms of classical thermodynamics, involving the competition between bulk and interfacial energies. A new type of phase diagram is introduced, the biphase diagram, in which concurrent phase stabilities are mapped as a function of two degrees of freedom, corresponding to two independent layer thicknesses in a periodic multilayer. The model is used to explain our observations of phase stabilities in Al/Ti multilayers, as a function of varying layer thicknesses. The model is also applied to explain the experimental observations made by other investigators on phase stability in Co/Cr thin-film multilayers.

The reaction at the interface between Al and amorphous C in Al/C multilayer thin films with modulation wavelengths of about 25nm and 125nm has been investigated by differential scanning calorimetry, X-ray diffraction, transmission electron microscopy/selected area electron diffraction and high resolution transmission electron microscopy. The reaction was found to take place in two steps. The first step onsets at about 300°C, and was identified as the diffusion of C into Al. The second step starts above 400°C, at a temperature strongly dependent on the modulation wavelength of the film, and is the formation of A14C3. The carbide has been observed to nucleate and grow inside the Al layers. The multilayer structure is preserved in the samples up to at least 550°C, and Al grains start to grow at or below 300°C.

The nucleation and growth of Cu2O due to oxidation of Cu(001) films were monitored at various temperatures and oxygen partial pressures. For all examined temperatures and pressures, Cu2O islands were observed to form epitaxially with respect to the copper film. The nucleation of these oxide islands was homogeneous –no clear evidence was observed for either steps or dislocations being preferential nucleation sites. Based on this data, we have developed a semiquantitative model of the initial oxidation stage where the dominant mechanism for transport, nucleation and growth of oxide islands is oxygen diffusion on the Cu surface. We are presently comparing our experimental data with nucleation rate theory, where the predictions qualitatively describe our observations, but not quantitatively.

Self-propagating reactions in free-standing multilayer foils provide a unique opportunity to study very rapid, diffusion-based transformations in non-equilibrium material systems. To fully understand the coupling between mass and thermal diffusion controlling these reactions and to optimize the commercial use of reactive foils, we have undertaken analytical and numerical modeling. Our analytical model predicts an increase in the reaction velocities with decreasing bilayer thickness down to a critical bilayer thickness and a reversal in this trend below the critical thickness. Predicting reaction characteristics such as the flame thermal width, the reaction zone width and the effect of variations in material properties with temperature has proven analytically intractable. To overcome these limitations, we have also used numerical methods to determine the composition and temperature profiles ahead of the reaction front for different multilayer periods and premixing. The results are compared with experimental values where possible.

Elementally modulated reactants with repeat layer thicknesses of less than 30Å were found to crystallize various molybdenum silicides depending on their composition. Reactants with compositions near 1:2 Mo:Si were found to nucleate h-MoSi2, at 400°C with an activation energy of 1.9eV, even though h-MoSi2 is metastable with respect to t-MoSi2 below 1900°C. Reactants with compositions near 5:3 Mo:Si were found to nucleate Mo3Si3 at 650°C with an activation energy of 3.0eV. Reactants with compositions near 3:1 Mo:Si were found to nucleate Mo3Si at 720°C with an activation energy of 2.2eV. The ability to control the crystalline product via the initial composition is distinctly different from the behavior reported for multilayers with larger repeat spacings, which were observed to form h-MoSi2 regardless of overall composition. The measured nucleation energy as a function of composition suggests that the phase selectivity is nucleation controlled. The selectivity depends on the elimination of large composition gradients at the reacting interfaces before nucleation of the crystalline product.

The properties of metallic thin films prepared by pulsed laser deposition (PLD) strongly differ from samples grown by conventional thin film techniques like sputtering and evaporation. In many systems (Fe-Nb, Fe-Zr, Cu-Co, Fe-Ag and Ni-Ag), strongly supersaturated solid solutions are formed, a broader formation range occurring for amorphous samples, and nanocrystalline phases are produced. Additionally, metastable phases of larger thickness at the interfaces of multilayers occur. Reasons for these behaviors are discussed with respect to atomic mixing and implantation effects induced by the high instantaneous deposition rate during PLD and the high kinetic energy (>100 eV) of the deposited ions.

The kinetics of the formation of Cu3Si in Cu/a-Si diffusion couples have been investigated by means of differential scanning calorimetry and x-ray diffraction. Multilayered composites of average stoichiometry Cu3Si were prepared by sputter deposition with individual layer thicknesses varying in different samples between 2 and 100 nm. We observed diffusion limited growth of Cu3 Si upon annealing these diffusion couples below 500 K. Reaction constants were measured for a temperature range of 455 to 495 K for thicknesses of growing Cu3Si between 2.6 and 80 nm. The temperature dependence of the reaction constant, k2, was characterized as k2 = k0 exp(− Ea/kbT) with activation energy, Ea = 1.0 eV/atom and pre-factor, k0 = 1.9×10−3 cm2/s.

Antiphase domain boundary is a region in materials where ordering of atoms changes from one structural variant to another. Influence of the internal energy excess on the dynamics of such boundary is considered in the framework of the Onsager theory of linear response. The internal energy transport entails a temperature hump in the transition region and causes a drag effect. An evolution equation that takes into account the finite thermal conductivity is derived. Experimental setup to reveal the thermal drag is suggested.

A series of fine grained C49 thin films were prepared by depositing Ti films on heated amorphous as well as crystalline silicon substrates.

Rapid thermal processing at temperatures between 700 and 775 °C was used to transform the metastable C49 phase into the stable C54. The evolution of the film structure and morphology during the transformation was investigated by sheet resistance, optical spectroscopy, diffuse reflectance and AFM measurements. One structural model which could account for our data is proposed according to crystallographic information present in the literature.

This work's aim is to report for the first time the cubic to hexagonal phase transition in tindoped In2O3 films with a Sn/In atomic ratio of 0.03, fabricated at low temperature and normal pressure from alcoholic solution of InCl3 and SnCI4. The performed X-ray diffraction measurements show a difference between crystallographic symmetry of thin (100 nm) and thick (400 nm) films prepared in the same conditions: the structure of thick films can be related to high pressure In2O3 hexagonal system with a preferred orientation of c-axis parallel to the substrate surface, while thin films present a cubic symmetry with columnar (400) grain orientation. Phase transition nature is connected with non-axial tensile deformation of indium oxide grid due to insertion of chlorine ions in the position of two diagonally opposite oxygen vacancies in In2O3 network.

The formation of TiSi2 thin films using the SALICIDE process on doped and undoped silicon substrates was studied. XRD TEM, AES, RBS and four probe Rs were used to characterize the material. Unit cell parameters and energetics were determined. Results confirm electrical and chemical signatures consistent with the known C49 conversion to C54. However, XRD indicated a structurally different intermediate phase occurs during the C49 to C54 transformation. Modeling was performed based on C11b structure (14/mmm) type, with the Ti and Si atoms arranged similarly to those in MoSi2. The unit cell was determined to be a = 4.428 Å, b = 4.779 Å, c = 9.078 Å with a Fmmm space group and total pseudo-potential plane wave calculations based on crystallographic simulations of −103.96 ev/Atom.

Bulk zirconia undergoes a pressure-induced transformation from a (low pressure) monoclinic phase to a high pressure tetragonal phase, at around 3GPa (above 900K). We have studied the structures of zirconia nanoparticles formed by plasma-spraying an organo-metallic precursor. Inspection of the particles in the TEM reveals that they adopt one of two distinct crystal structures, depending upon their size. The smallest particles have the tetragonal structure, while larger ones are monoclinic. Interpolation of the data reveals a critical size above which the monoclinic structure is stable. Upon annealing, the zirconia particles coarsen and the small tetragonal particles transform to the monoclinic structure at about the critical size. Coarsening under these conditions produces irregular particle size distributions. We estimate that the surface-stress induced internal pressure in the tetragonal nanoparticles can be as high as 5 GPa and the corresponding surface stress is about 6N/m.

Initial thermodynamic analysis predicts surface tension of fine solid particles decreases with decrease in particle size. Free energy of fine solid particles increases with decrease in particle size. In the nanocrystalline TiO2 system, the surface tensions and the surface energies of both anatase and rutile were estimated by modeling available thermochemical data and kinetic data. The particle size versus temperature phase diagram of the nanocrystalline TiO2 system was calculated, which reveals 1–8 nm anatase is more stable than rutile of the same size at certain temperatures.

Oxidation of an aluminum nanocluster (252,158 atoms) of radius 100Å placed in gaseous oxygen (530,727 atoms) is investigated by performing molecular-dynamics simulations on parallel computers. The simulation takes into account the effect of charge transfer between Al and 0 based on the electronegativity equalization principles. We find that the oxidation starts at the surface of the cluster and the oxide layer grows to a thickness of ∼28Å. Evolutions of local temperature and densities of Al and 0 are investigated. The surface oxide melts because of the high temperature resulting from the release of energy associated with Al-O bondings. Amorphous surface-oxides are obtained by quenching the cluster. Vibrational density-of-states for the surface oxide is analyzed through comparisons with those for crystalline Al, Al nanocluster, and α-Al2O3

Mössbauer measurements were made on FeRh containing 52 atomic percent (at.%) Fe after mechanical milling for different times in a high-energy SPEX 8000 vibrator mill. Hyperfine fields are compared with fields for annealed ferromagnetic (F) samples having the B2 structure in the range 52–58 at.% Fe. For annealed F samples, hyperfine field shifts of -1.66 T and -1.12 T were detected at majority FeFC and minority FeRh probes due to FeRh antisite atoms in, respectively, the first and second atomic shells. Analysis of dipolar fields indicates that the magnetization lies along the <110> direction. The transformation from F B2 phase to a metastable paramagnetic fcc phase was observed that was half complete in 8 minutes. Analysis of spectra for the milled F B2 phase show that a second point defect was produced by milling that induces shifts of +6 T and +8.5 T, respectively, at majority and minority probes. Manyspectra fits were made under different defect models that led to the conclusion that milling produces point defects in the triple-defect configuration: 2 Fe-vacancies and 1 Fe-antisite atom. Defect concentrations were determined and show that the fractional concentration of vacancies on the Fe-sublattice increases linearly with milling time, reaching 3 at.% after 8 minutes, a very large value. The rate of transformation from B2 to fcc phase appears to be independent of the concentrations of point defects in the B2 phase, indicating that the transformation is purely stress-induced, as in a martensite transformation.

A model analysis is presented which explains ball-milling induced alloying in positive-heatof- mixing systems in terms of a dynamic balance between externally forced mixing and thermal phase decomposition mediated by deformation-enhanced population of defects. The possibility of eliminating the thermal decomposition to force single phase formation is examined by milling Cu- Fe and Cu-Ta powder mixtures at the liquid nitrogen temperature (LN2T). Over a range of compositions for Cu-Fe and almost the entire composition range for Cu-Ta, the two-phase region observed for room-temperature (RT) milling persisted after cryomilling. The moderate temperature dependence of milling-induced alloying is interpreted by analyzing the dynamics of the generation and annihilation of the nonequilibrium vacancies during deformation and impacts in a SPEX mill.