ABSTRACT[hiring liquid-phase sintering of W-Ni-Fe composites, single crystal spheroids of essentially pure tungsten form in a W-Ni-Fe matrix. These spheroids grow in a systematic way if either the sintering time is increased, or if the sintering temperature is raised. The generally accepted mechanism of spheroid growth is the dissolution of the small spheroids with simultaneous reprecipitation of tungsten frora the molten matrix onto the larger ones, the reaction being driven by surface energy differences between the small and large spheroids. It has been shown theoretically that for such reactions, the plot of log diameter-log cime will give a straight line of slope 1/3. This theory has been modified to include a rearrangement process for achieving densification during the early stages of sintering and a coalescence process for combining solid spheroids during the latter stages of sintering. Furthermore, sintering at different temperatures should proceed with a single-valued activation energy since a single growth mechanism is rate controlling.We studied the sintering behavior of a low volume matrix alloy (95% W-3.5X Hi-1.5% Fe) and found that the slope of the log diameter-log time plot was less than 1/4, and that the growth process could not be represented by a single-valued activation energy. These experiments as well as those generally published in the literature have several shortcomings in that they do not consider the effect of solid state reactions during presintering, nor the effect of sintering temperature on the spheroid nucleation size. Furthermore, determining activation energies after a constant sintering time does not evaluate structural differences. To overcome these objections, we ran n second series of experiments where the baseline material had already been liquid-phase sintered, and the growth rates determined after additional sintering. Activation energies were determined from the tines required at the different sintering temperatures to achieve equivalent structures. In these tests, the log diameter-log time plots had slopes less than 1/4 and the activation energies were different for different sintering conditions. These results indicate that the dissolution-reprecipitation process is not rate controlling in the 95% W-3.5% Ni-1.552 Fe composite. We feel that the growth rates are too fast for such a process to be rate controlling, and that "coalescence", i.e., the joining of two or more spheroids that come into contact at the proper crystallographic orientation and form a single spheroid, is much more important even during the initial stages of sintering. The rate for such a growth mechanism would depend on some degree of matrix circulation, and the probability for proper spheroid contact. Thermal activation, then, would mean an increase in the matrix circulation and perhaps a greater acceptable angle of contact for welding. Since such a process is not diffusion controlled, but depends on probability, growth rates would behave in a systematic way, but not in accordance with any particular g...