Part I – January 1967 - Papers - Interface Compositions, Motion, and Lattice Transformations in Multiphase Diffusion Couples

In nzost cases, the driving force for a lattice transformation is produced by supercooling below the equilibriunz transformation temperature. The interfnce reaction in isothermally annealed, multiphase diffusion couples may involve a luttice transformation which also requires a driving force. Direct experinzental evidence has been obtained for the existence of the driring force in the form of a supersaturated phase at the aocc)-0@cc) interface in Cu:Cu-12.5 ult pct A1 couples; the super saturation is equivalent to an excess free energy of approximately 3 cal per mol at 905. A tentatiue interpretation of the dynanzic situation a1 the interface based on the free energy-composition diagram is proposed. THE presently accepted theory of diffusion in multiphase couples1 states that there will be a phase layer in the diffusion zone for every region which has three degrees of freedom and which is crossed by the diffusion path in the equilibrium phase diagram. For binary systems, this restriction excludes all but single-phase fields and, for ternary systems, only one- and two-phase fields are included. In addition, Rhines"~ as well as other investigators3 6 have reported that the compositions of the various phases adjacent to the interfaces are, for all practical purposes, the compositions given by the intersections of the diffusion path with the solubility limits of the single-phase fields of the equilibrium phase diagram. Some studies of the rate of thickening of these intermediate diffusion layers indicate that the thickness of the layer changes para-bolically with time, or: where x is the position of the interface relative to an origin xo, t is the diffusion time, and k is a temperature-dependent factor. crank7 shows mathematically that, if the compositions at an interface are independent of time and the motion of the interface is controlled by the diffusion of the elements to and from the interface, then the segments of the concentration penetration curve for a semi-infinite step-function couple will be described by an equation of the form: hence, Eq. [l] follows from Eq. (21 if the interface compositions are fixed and if the motion of the interface is diffusion-controlled. Although the concept of local equilibrium being attained at interfaces has assumed a prominent role in the theory of diffusion in multiphase couples, experimental evidence and theoretical discussions which challenge the general validity of this concept have been reported in the literature. arkeen' has stated that strict obedience to the conditions set by the equilibrium phase diagram cannot be expected in any system in which diffusion is occurring because diffusion takes place only in the presence of an activity gradient. Darken also noted that it is usually assumed that equilibrium is attained locally at the interface although the system as a whole is not at equilibrium, the implication being that the transformation at the interface is rapid in comparison with the rate of supply of the elements by diffusion. ISirkaldy3 indicates agreement with Darken in that he believes the concept of local equilibrium is at best an approximation because the motion of the phase boundary requires that there be a free-energy difference and, hence, a departure from the equilibrium composition at the interface. Seebold and Birks9 have stated that diffusion couples cannot be in true equilibrium, but the results obtained are often in good agreement with the phase diagram. The initial deviation from equilibrium in a diffusion couple will be quite large because alloys of significantly different compositions are usually joined together. Kirkaldy feels that the transition time for the attainment of constant interface compositions (essentially the equilibrium values) will be small, although in some cases finite. Castleman and sieglelo observed such transition times in multiphase A1-Ni couples, but at low annealing temperatures these times were quite long. Similarly, ~asing" found departures, which persisted for more than 20 hr, at phase interfaces in Au-Ni and Fe-Mo diffusion couples. Braun and Powell's12 measurements of the solubility limits of the intermediate phases in the Au-In system as determined by microprobe analysis of diffusion couples do not agree with the limits reported by Hiscocks and Hume-Rothery13 who used equilibrated samples. Finally, Borovskii and ~archukova'~ have stated that the determination of the solubility limits of phase diagrams using high-resolution micro-analyzer measurements at the interfaces of multiphase couples is not an accurate technique because of deviations from the equilibrium compositions at a moving interface; diffusion couples may be used to map out the phase boundaries in the equilibrium diagram, but the final determination of the solubility iimits should be made with equilibrated samples. The purpose of this work was to investigate the conditions prevailing at an interface in a multiphase diffusion couple and to compare the interface compositions with those associated with true thermodynamic equilibrium between the two phases. Microanalyzer techniques were used to measure interface compositions in two-phase Cu-A1 diffusion couples annealed at 80@, 905", and 1000°C for various times.