Volume 96, Issue 2

There are two approaches to the interaction between solute atoms and migrating interfaces. A comparison between the two is illustrated with molar Gibbs energy diagrams. It is demonstrated that the present treatments of solute drag are equivalent to the treatment based on dissipation of Gibbs energy for grain boundary migration but not for phase transformations. A new treatment of solute drag is equivalent to the dissipation approach for both cases. It predicts that the solute drag changes sign for phase transformations and acts as a driving force.

The grain boundaries in high-purity oxide ceramics, such as Al 2 O 3 and TZP, are often free from glass phase, and the high-temperature plastic flow or grain boundary failure is sensitive to small levels of doping by various cations. For example, the high-temperature creep strain rate in finegrained, polycrystalline Al 2 O 3 is highly retarded by 0.1 mol% Lu 3+ or Zr 4+ -doping. The elongation to failure in superplastic TZP is improved by 0.2–3 mol% Ge 4+ -doping. Such a dopant effect is attributed to changes of the grain boundary diffusion due to the segregation of dopant cation along the grain boundaries. Quantitative analysis on the atomic structure and chemical bonding state along the grain boundaries by high-resolution electron microscopy, energy-dispersive X-ray spectroscopy, electron energy-loss spectroscopy and molecular orbital calculations will provide theoretical guiding principles to design high-performance oxide ceramics in the near future.

The non-ideal behaviours of diffusion impedance observed in such intercalation electrodes as hydride-forming alloys and transition metal oxides were theoretically treated in terms of (i) the diffusion length distribution, (ii) diffusion towards and from the fractal surface and (iii) diffusion coupled with the trapping reaction or the homogeneous reaction. Firstly, when the diffusion lengths were widely distributed across the electrode, the absolute value of the phase angle deviated positively from 45° in the high frequency range, and it deviated negatively from 90° in the low frequency range. Secondly, the ac-impedance spectrum for diffusion through the fractal electrode exhibited a power-law behaviour of frequency dispersion with the absolute phase angle between 45° and 90° in the high frequency range, depending on the fractal dimension of the electrode surface. Finally, an arc appeared in the Nyquist plot of the ac-impedance spectrum, in case diffusion was coupled with either the trapping reaction or the homogeneous reaction.

A numerical simulation of the growth and coarsening of completely faceted particles in a two-dimensional closed system is described. The particles grow from a supersaturated solution with driving forces including an anisotropic-Gibbs–Thompson effect. Linear and non-linear growth laws were incorporated. This allows comparison to data obtained from physical experiments of Si 3 N 4 growing from a glass matrix with additions of various rare earths (La, Y, Lu) where particle growth is more or less anisotropic (depending on the particular rare earth dopant). Simulations explore the ranges of kinetic parameters for which particle shapes tend to be dictated by thermodynamic and/or kinetic anisotropy. Comparison to data taken from a series of experiments indicates that La probably has a non-linear growth effect whereas Lu is less so.

Thermal conductivity of a functionally graded Fe – Cu – C alloy prepared by liquid phase sintering and the subsequent carburization was investigated. The formation of the graded structure and its characterization are briefly reviewed, and the thermal conductivity of the graded compact is analyzed by Eshelby’s equivalent inclusion method. A microstructure model in which ellipsoidal copper particles are embedded in the iron matrix was employed. The thermal conductivity was calculated for directions parallel and perpendicular to carburization as functions of the size, aspect ratio, and volume fraction of the copper particles. The predicted values were in good agreement with the experimental results. The present analysis will give a good guide for the practical application of this material.

Microstructure development during liquid-phase sintering has been analysed using the pore-filling theory. The interdependence between sintered density and average grain size, i. e., the relative density – grain size trajectory has been described for various types of processing and sintering parameters. The effects of sintering temperature, initial porosity, average pore size, liquid volume fraction, dihedral and wetting angle, and sintering atmosphere pressure have been evaluated. A critical examination has also been made of the factors determining the trajectory shape and the implication of activation energy for densification. The predicted microstructure developments should demonstrate various possibilities of controlling microstructure during liquid-phase sintering.

The effect of thickening of intermetallic compound layers formed at the interface and shear properties of solder joints are studied as a function of aging for lead-free solders. The modified double lap shear test method is employed to measure the shear strength and shear strain. The test specimens are Sn-3.5Ag and Sn-3.5Ag-1Zn solders and they were aged up to 2880 h at 423 K. Intermetallic compound layers at the interface between Sn-3.5Ag solders and Cu substrate grow with increasing aging time. The coarsened Cu 6 Sn 5 intermetallic compounds lead to a deterioration of the shear properties of solder joints by changing the crack propagation path and decreasing shear strength of the solder joint. Adding 1 wt% of Zn to Sn-3.5Ag, however, suppresses Cu 6 Sn 5 intermetallic compound growth markedly by forming a Cu–Zn–Sn layer. This Cu–Zn–Sn layer acts as a diffusion barrier between Cu 6 Sn 5 and Sn-3.5Ag-1Zn solders during the aging treatments. Thus, Sn-3.5Ag-1Zn solders have higher shear strength and higher strain to fracture than those of Sn-3.5Ag/Cu solder joints.

The three-dimensional morphological evolution of a directionally solidified Al-15 wt.% Cu alloy during isothermal coarsening was examined. The microstructure was analyzed using the Interfacial Shape Distribution (ISD), which gives the probability of locating an interfacial patch with a given pair of principle curvatures. By scaling the ISD’s by a characteristic length, the inverse of the surface area per unit volume, it was possible to determine if the system reached a scaled time-invariant state. During coarsening, cylindrical interfaces appear in the second coarsened sample and remained for later coarsening times. The appearance of the corresponding peak in the ISD indicated that the microstructure was not evolving in a self-similar manner, despite the linear relationship between the inverse surface area per unit volume and the cube root of coarsening time. Stereographic projections of the interfacial normals indicated that the microstructure of the earliest coarsened sample had a four-fold symmetry that disappeared during coarsening and that the majority of interfacial area for all coarsened samples is parallel to the direction in which the pre-coarsened samples were solidified.

Faceting and migration of twin tips in slightly deformed Zn [112̅0] flat single crystals have been investigated. The stationary shape of the slowly migrating tip of the twin plate has been studied, and the migration rate has been measured between 592 and 692 K in situ using polarized light. Below 632 and above 682 K, the twin tip contains only one facet which is parallel to the (1̅102̅) 2 and (11̅00) 1 plane of the matrix single crystal, respectively. The non-physically high values of the apparent migration activation enthalpy between 632 and 682 K are discussed using the model of simultaneous migration of co-existing grain boundary facets based on the concept of weighted mean curvature. The experimentally determined temperature dependence of the twin tip mobility can be rationalized under the assumption that the migration of individual facets is characterized by identical activation enthalpies of 46.6 kJ/mol, but different pre-exponential factors (1.1×10 –11 and 7×10 –9 m 2 /s, respectively).

The grain growth behavior of La-doped BaTiO 3 specimens was investigated at 1350 °C under DC external electric field. In air atmosphere, grain growth was enhanced at the positive-biased region for the 0.15 mol% La-doped and 0.3 mol% La-doped specimens, while it was enhanced at the negative-biased region for the 1.0 mol% La-doped specimen. For the 0.6 mol% La-doped specimen, on the other hand, both positive- and negative-biased regions showed the same grain size. The demarcation donor composition (0.6 mol% La) shifted to a higher donor concentration when the specimen was heat-treated in nitrogen. The results were explained in terms of the grain boundary potential and its variation with donor concentration, oxygen partial pressure, and defect polarization.

Microstructure and mechanical properties of ultra-fine WC-10Co-X cemented carbides were investigated.Nanocrystalline WC-10Co-X powders with an average WC size of 100 nm were prepared by spray drying of solution containing ammonium metatungstate and cobalt nitrate. Varying amounts of TaC, Cr 3 C 2 and VC were added into ultra-fine WC-10Co cemented carbides as grain growth inhibitors. The hardness of WC-10Co cemented carbides increased with decreasing WC grain size following the Hall – Petch-type relationship. The fracture toughness of ultra-fine WC-10Co fabricated by the spray conversion process was lower than that of conventional WC-10Co. These differences in fracture toughness arose from solute content in the Co binder phase. The fracture toughness of WC-10Co-X cemented carbides with the same solute content in the Co binder phase varied with WC grain size and Co mean free path.

Systematic grain boundary study of cubic zirconia has been conducted by using bicrystals. It is clearly demonstrated that grain boundary atomistic structures dramatically change according to the misorientations and plane orientations of the boundaries, resulting in a dramatic change of excess energies and solute segregation behaviors. Combining with theoretical calculations, it is found that grain boundaries possess unique coordination-deficient cation sites at the cores, and their densities have a clear correlation with these properties in high-angle grain boundaries. This result indicates that grain boundary properties in ceramics are possibly determined by the accumulation of coordination-deficient sites. Thus, systematic grain boundary study using bicrystal offers fundamental understandings of the relationship between atomistic structures and properties in ceramic grain boundaries.

Spontaneous generation of charged atoms or clusters was investigated during thermal evaporation of silver. For this, the effect of the applied electric bias on the film growth rate was examined during evaporation of silver at 1373 K in a tungsten basket. Film growth rates on three silicon substrates biased + 300, 0 and –300 V with respect to the chamber were 300, 420 and 960 nm per hour, respectively. The number density of generated positively-charged atoms or clusters could be measured by the electric current on the Faraday cup in the chamber. From the temperature dependence of the positive current, the activation energy for charging was determined to be ~ 2.2 eV. This value could be best explained by the surface ionization of clusters of a few atoms on the oxidized tungsten surface.

When the interface energy between coarsening crystals and an intervening phase is anisotropic, mechanisms that do not affect isotropic systems become important. If there are singular surfaces, then growth and dissolution must occur by the lateral motion of steps, formed at a defect center, or by two-dimensional nucleation. Here, it is shown that two-dimensional nucleation is not plausible under typical experimental conditions and that persistent step-creating defects are required for a singular surface to advance or retract during coarsening. The simultaneous presence of crystals with and without defects leads to two populations that grow at very different rates, and this provides an explanation for abnormal coarsening. The influence of extrinsic morphological changes is also considered. It is assumed that when relatively high-energy, non-equilibrium shapes in the starting materials evolve during coarsening to shapes increasingly bound by lower-energy surfaces, the mean surface energy is reduced. Simulations show that under these conditions, non-classical coarsening kinetics arise in which the rate constant decreases linearly with the mean surface energy.

Polycrystalline silicon is extensively used in modern technology like solar-cells and thin film transistors. Towards the development of high performance devices based on polycrystalline materials, it is essential to design and control grain boundary microstructures. This article gives an overview of our recent experimental studies on electrical properties of grain boundaries in polycrystalline silicon. Particular attention is paid to the influences of the grain boundary character/structure, chemistry (hydrogenation and oxidation) and temperature on the electrical activity of individual grain boundaries. The electron beam induced current (EBIC) technique was used to observe electrical activity of grain boundaries. The observed results are discussed based on the Shockley-Read-Hall statistic.

The crystallographic distribution of surfaces surrounding grains in PMN-35 mol% PT has been measured in samples annealed for 0, 2, and 4 hours at 1150 °C. At all stages of the growth, grains are bounded predominately by {100} planes while surfaces with the {111} orientation represent a minimum in the distribution. During annealing, the fraction of the total area bounded by surfaces with the {100} orientation increases by 50 %. As the relative areas of the {100} surfaces increase, the rate constant for growth decreases. Both the evolution of the external shape and the overall reduction in the growth rate suggest that interfaces with the {100} orientation have a low relative mobility.

The order – disorder transition temperature of the PtCo bulk alloy and the effect of heat treatment on an as-sputtered disordered PtCo thin film have been investigated using Monte Carlo simulations coupled with a modified embedded atom method. From the long-range order curve combined with atomistic configurations, the order – disorder transition temperature is estimated between 1000 K and 1200 K. It has also been confirmed that the heat treatment at 900 K on the as-sputtered PtCo thin film induced ordering. The Pt segregation at the top surface of the thin film resulted in ordering of the whole thin film, except for four or five bottom layers probably because of the influence of the substrate.

The faceting of two tilt grain boundaries (GBs) in 99.999% wt. purity Al bicrystals has been studied. The first GB is a coincidence twin Σ3 <110> GB (Σ is the inverse density of coincidence sites) and the second GB has a misorientation close to the Σ3 <110> with a tilt deviation of about 3 degrees. Orientation of grains along the GB has been controlled with the electron back-scattering diffraction method. The bicrystalline samples were coated with a layer of Sn– Al mixture and annealed at 873 K. Contact angles at the junction of the grain boundary and two solid/liquid interfaces were measured by optical microscopy. The ratio of grain boundary energy and solid/liquid interface energy has been calculated. Wulff –Herring plots and GB phase diagrams were constructed. The formation of smooth edges with rounded rough GB portions was observed in the grain boundary deviating from Σ3 <110>.