The diffusion coefficients of several transition elements (Ti, V, Cr, Mn, Fe, Co, Ni, Cu, and Zn) and a few non-transition elements (Mg, Si, Ga, and Ge) in fcc and liquid Al are critically reviewed and assessed by means of the least-squares method and semi-empirical correlations. Inconsistent experimental data are identified and ruled out. In the case of the elements, for which plentiful experimental data are available in the literature, the least-squares analysis gives rise to the activation energies and pre-exponential factors in an Arrhenius equation. For the elements with limited experimental data or no data at all, the diffusion parameters are estimated from two semi-empirical correlations. In one correlation, the logarithmic pre-exponential factors are plotted against the activation energies for various elements in Al. In the other correlation, the activation energies are shown as a function of valences relative to Al. The diffusion coefficients calculated by using the evaluated diffusion parameters agree reasonably with the reliable experimental data. The proposed semi-empirical correlations are used to predict the diffusion coefficients of a few elements in liquid Al. A satisfactory agreement between the predicted and measured diffusion coefficients is obtained.

The modified cluster-site approximation (CSA) model has been used to model the f.c.c. phases (ordered L12 and disordered A1) in the Ni-Al system. The CSA model has the advantage over the cluster variation method (CVM) in that the independent variables in the free energy functional are the site probabilities and not the cluster probabilities. Unlike the zeroth approximation, however, the CSA takes short-range order into consideration, an essential requirement for describing phases that undergo an order/disorder transition, as does the f.c.c. phase in the Ni-Al system. By using the modified CSA model, we have been able to obtain an improved phenomenological description of this system with the use of fewer model parameters than used in previous descriptions. The topology of the calculated metastable f.c.c. phase diagram is similar to the one obtained using a 'first-principles'-CVM approach, whereas the f.c.c. phase diagrams based on the previous phenomenological descriptions are unsatisfactory.

An Al-10Ni-8Y (at.%) alloy was atomized by Ar gas and the morphology, microstructure, thermal stability, phase composition and microhardness of the as-atomized powder were investigated. Most of the powders are spherical in shape, but the surface morphology was different for powder of different size. The cross-section microstructure of powder with size below 15mm in diameter showed no detailed feature, indicating existence of amorphous phase or nanocrystalline structure. The as-atomized powder showed four distinct exothermic peaks when heated at 296, 340, 366 and 456 8C. The glass transition temperature Tg, crystallization temperature Tx, and the temperature interval of the supercooled liquid region DTx (ZTxKTg) were detected to be about 266, 288 and 22 8C. The Al82Ni10Y8 alloy powder exhibits a high Vickers hardness of 230.6, and shows great potential for structural application.

The Al-Y system is reassessed based on a critical literature review involving recently measured phase diagram and thermodynamic data. In the thermodynamic optimization, the liquid phase is described by using a substitutional solution model. Al3Y, AlY, Al2Y3 and AlY2 are modeled as stoichiometric ones. A two-sublattice model (Al,Y)2(Al,Y)1 has been used to describe Al2Y. A set of self-consistent thermodynamic parameters for the Al-Y binary system has been obtained by means of a CALPHAD approach applied to the selected experimental data. The reliable experimental phase diagram data and thermodynamic properties are well reproduced with the optimal thermodynamic parameters. The observed solidification paths for as-cast alloys can also be described reasonably using the present parameters. Significant improvement has been made, compared with previous assessments.

A thermodynamic modeling for the A1-Mn-Si system is conducted via two steps. In the first step, all of the experimental phase diagram and thermodynamic data available from the literature ate Critically reviewed. and several isopleths are constructed by means of X-ray diffraction (XRD) and differential thermal analysis (DTA) techniques. In the second step, a consistent thermodynamic data set for the A1-Mn-Si system over the entire composition and temperature ranges is obtained by consideling reliable literature data and the present experimental results. Numerous comparisons between the calculated and measured phase diagrams as well as thermodynamic quantities indicate that almost all of the reliable experimental information is satisfactorily accounted for by the thermodynamic modeling. The liquidus projection and reaction scheme for the entire system are also presented.