The kinetic theory of the glass transition is examined. Some modifications are suggested in order to overcome some shortcomings. As a result, a thermodynamic description of the glass transition is obtained. It is demonstrated that the glass transition represents the beginning of a very gradual process of melting in the amorphous state.

The importance of growth kinetics in the development of the microstructure of rapidly solidified alloys is described. Growth kinetics are conveniently divided into diffusion kinetics and interface attachment kinetics. The former, which are used extensively for the analysis of slow rate solidification, can be extended to high solidification rates to predict some microstructural features; e.g., the limitations on eutectic growth rate which can promote the formation of metallic glass, and the reduction of microsegregation. At the highest rates interface attachment kinetics must be included. Some microstructural effects of the velocity dependence of the partition coefficient will be described.

Rapid solidification invariably results in fine-scale microstructures with concomitant fine-scale segregation. These microstructures, which are produced relatively far from equilibrium, have large. specific surface energies and tend to coarsen rapidly. Experiments were carried out on highly supercooled bulk specimens of elemental P4 in which the average temperature of rapidly solidified 2-phase mixtures was accurately monitored over time. A continuous rise in temperature toward the bulk melting point is always measured, independent of the volume fraction of solid produced at recalescence. New theoretical work outlined here shows that this temperature rise can be associated with the Ostwald ripening of the as-solidified solid-plusliquid mixture. The new theory permits calculation of coarsening rates after recalescence and the distribution of dendritic particle sizes as a function of time. Implications for establishing limiting dendritic particle sizes are discussed.

The undercooling of liquid metals is observed often, but is restricted by the catalysis of the most potent heterogeneous nucleation site in contact with the liquid. An experimental approach that yields large undercoolings involves the slow cooling from the melt of a dispersion of stabilized fine (5–20μ) liquid droplets. Droplet studies have extended measurements of the undercooling limit for low melting point metals to the range of 0.3 to 0.4 Tm. In droplets, surface coating and size characteristics usually limit the undercooling and control the uniformity of crystallization behavior. A nonuniform undercooling behavior can produce microstructural variations in powder samples with respect to metastable phases and solute segregation, but these variations can be diminished for isenthalpic solidification of hypercooled liquids. In alloys the trend in nucleation temperature can follow a similar pattern as the composition dependence of the liquidus. With highly undercooled droplets and even at modest ΔT (ΔT ≤ 0.lTm) with known heterogeneous sites, structural modifications such as extended solid solutions and metastable intermediate phases can be generated which are similar to those produced by rapid quenching techniques. These features highlight the importance of a high liquid undercooling as well as a high cooling rate in promoting effective rapid solidification processing treatments.

This paper describes the results of experiments on metal droplet undercooling, and presents a preliminary analysis of the structures obtained. Work was conducted with Sn-25wt%Pb and Ni-34wt%Sn alloys. To achieve the required high degrees of undercooling, emulsification treatments were utilized following techniques described by Perepezko and co-workers.

Experiments show the fraction of supersaturated primary phase is found to be a function of the amount of undercooling, as is the fineness of the structures. The solidification behavior of the tin-lead droplets during recalescence was analyzed using three different hypotheses; 1) solid forming throughout recalescence is of the maximumthermodynamically stable composition, 2) partitionless solidification below the “To” temperature, and solid forming thereafter is of the maximum thermodynamically stable composition, and 3) partitionless solidification below the “To” temperature with solid forming thereafter that is of the maximum thermodynamically metastable composition that is possible.

Extension of solid solubility by rapid quenching from the melt (RQM) enormously increases scope for alloy development in metals such as aluminium for which solid solubility is particularly limited under equilibrium conditions. The present contribution reviews mainly recent work concerned with matching of experimental observations with predictions affecting:

1. (1) conditions for solidification without change in composition

2. (2) temperatures and modes of deconposition on subsequent heating

3. (3) hardening effects as-quenched and on heat treatment.

The significance of such findings for bulk production and consolidation of wrought products for engineering applications is briefly discussed.

For constant velocity solidification, morphological stability theory delineates the temperature gradients required for plane front solidification of a specific alloy. Using electron beams, surface heating of metals can be carried out with sufficiently well characterized thermal input to permit reliable use of computer models of melting and solidification. From numerical calculations, the growth velocity and temperature gradients as a function of position during resolidification can be obtained; combining these results with (constant velocity) morphological stability theory indicates the resolidification regimes for which the plane front is unstable. Presumably, completely planar solidification may be attained by selecting heating modes such that the region of instability is totally avoided, but the expected interface morphology is more difficult to predict if the interface passes briefly through an unstable region and then re-enters a region of stability. Aluminum-silver and aluminum-manganese alloys were melted under an electron beam with particular emphasis on attaining solidification sufficiently rapidly to satisfy the gradient-independent absolute stability condition. It was found that the velocities required to produce stability were considerably larger than expected.

Recent advances in both diffraction theory and experiment have led to the determination of the topological and chemical short-range order in liquid alloys and metallic glasses. In binary alloys, three partial interference functions (or partial structure factors) must be determined to evaluate the Warren chemical short-range order parameter α. Examples of recent attempts to determine α in binary liquid and glassy alloys are given. In most glasses, studied so far, evidence exists for the occurrence of unlike nearest neighbor ordering which must be present in the liquid state.

SCF-Xα-SW calculations have been performed on the glass forming crystal of Ni2B. The resulting molecular orbital topology near the fermi level is similar to that previously obtained by SCF-Xα-SW calculation on the glass forming crystals of Pd-Si and Cu-Zr. These similarties, it is conjectured, indicate melting is an electronically driven phase transition and that liquid and glass structures are more the result of electronic than kinetic factors. It is shown that the molecular orbital model of melting in transition metals will generate structures similar to those resulting from the dense random packing of hard spheres method, and the molecualr orbital mechanism admits, in a qualitative way, the explanation of a broad range of liquid and glass properties.

There are numerous processing methods available for the fabrication of rapidly-solidified metastable alloys. Examples include splat quenching, melt-spinning, surface melting and quenching, melt atomization and solidification, and deposition by sputtering and evaporation. Experimental aspects of each of these fabrication methods will be discussed. These include solidification and quench rate, process control, and final product morphology and microstructure. Some properties of an amorphous alloy prepared by several of these processing methods will be discussed.

In the past several years, rapid solification technology has been extended from the realm of splats and foils to powders. This paper surveys several of the models that have been developed for such powders and considers their comparison with experimental results.

The Electrohydrodynamic (EHD) method of spraying fine liquid droplets from a liquid state, in a vacuum environment, was developed and used to produce amorphous, microcrystalline, single crystal, bicrystalline and tricrystalline powders. Studies of these powders have contributed towards increasing the knowledge of extended solubility, nucleation, metastable phases, undercooling effects, etc. Coatings and films have been produced by collecting the liquid droplets before solidification. An automated instrument based upon the EHD method, the Micro-Particle Processor, is computer operated and allows a material scientist not completely acquainted with the EHD process to perform sophisticated experiments on materials of his choosing. Electron transparent powders close to 3μm and large powders up to l00μm have been collected and observed. Cooling rates above 107K/s have been achieved. Applications using powders include: new alloy compositions, use as AEM standards, in-situ remelt experiments in the electron microscope, etc.

New nickel rich alloys containing molybdenum and chromium prepared as metallic powders are reported. Using standard plasma spray techniques, these powders (−170 mesh) can be deposited onto metallic substrates as dense, well bonded coatings having predominantly glassy structure. Such coatings exhibit high structural strength, hardness and good corrosion resistance in various corrosive media.

Laminated metal-metal and metal-semiconductor ribbons have been prepared by a modified melt-spinning technique. Characterization studies have been carried out on laminated Sn-Zn, Ge-Al, InSb-CuZr and some other materials. Characteristics of these ribbons and their potential applications will be described.

To fully utilize the promising mechanical properties of metallic glasses, consolidation techniques need to be developed to incorporate them into large bulk forms. This paper describes preliminary results on the consolidation of Ni60Nb40 metallic glass strips into an aluminum alloy matrix. The consolidation was achieved without the degradation of the metallic glass by employing a superplastic aluminum alloy as the matrix material. The consolidations were accomplished at a superplastic forming temperature compatible with the metal matrix material and sufficiently below the critical 650°C crystallization temperature of the Ni60Nb40 metallic glass. Initial studies employed an atuminum − 5% calcium − 5% zinc alloy as the metal matrix material.

The techniques of scanning transmission electron microscopy and field iron microscopy/atom probe are briefly described. The advantages of these techniques for high spatial resolution compositional analysis are discussed and examples cited.

The cooling rates that accompany manufacture of gas turbine superalloy powders are sufficient to produce metastable structures and phase distributions. It is the aim of this investigation and report to develop some understanding of the relationships between powder manufacturing processes and powder structures by comparison of material made by two methods, and to describe some unusual experiments done to demonstrate structure solidification relationships.

The cooling rates during solidification of gas atomized Stellite 6 (S6) powders range from 103 K/s to 106 K/s for 400μm to 10 μm diameter powder particles respectively. The microstructures vary from equiaxed type in the coarser particles to fully dendritic, cellular and microcrystalline in the finer particles. In the case of melt spun ribbons, the variations in freezing rate along a cross section of the ribbon produce a change in the microstructure from microcrystalline at cooling rates greater than 107 K/s to cellular and finally dendritic at rates on the order of 106 K/s.

Superalloy powders of an IN 792 plus Hf type alloy produced from Rapid Solidification (RSR) and Argon Atomized process (AA) techniques are analyzed and compared according to particle size. Characterization of powders include surface and interior morphology, particle size and void distribution, cooling microstructures, microcrystallinity, microsegretation and especially dendrite arm spacing and cooling rates. It is found that, for the alloy studied, RSR cooling rates are a factor of only two to four greater than those for AA, regardless of particle size.

The microstructure of Fe-Ni-B alloys rapidly quenched by melt spinning has been investigated by means of light and electron microscopy. Depending on the surface velocity of the quenching wheel and the boron concentration a large variety of microstructures has been observed: increased solubility of boron, microstructures with different metastable crystalline compounds or metallic glasses.

Plotting the as-cast microstructures versus the quenching rate has been found to be very useful for arranging the large amount of results and/or for predicting particular microstructures. Due to an extremely large gradient in the microstructure across the ribbon thickness this method of forming particular microstructures seems to be only of limited advantage. Controlled crystallization reactions of metallic glasses result in different, but more homogeneous microstructures.