A Classical Phase Transformation: Order—Disorder in CuPt

Abstract

UNTIL quite recently, it was widely believed that order–disorder transformations in alloys were not phase transformations in the classical Gibbsian sense, but were homogeneous transformations in which the whole sample passed uniformly through a sequence of progressively more ordered states. A corollary of this view of the matter is that no two-phase fields incorporating both ordered and disordered phases can exist either transiently or in equilibrium. This view was first critically examined by Rhines and Newkirk¹, with special reference to the Cu₃Au and CoPt superlattices; they concluded that on the evidence then available, order–disorder transformations were classical; in particular, equilibrium two-phase regions exist in the Cu–Au and Co–Pt systems. Clear evidence, by X-ray diffraction, that the ordering process in stoichiometric CuAu involves classical nucleation and growth was advanced by O'Brien and Kuczynski², and electron microscopic evidence of a two-phase structure in off-stoichiometric Cu₃Au was presented by Marcinkowski and Zwell³. Doubts persisted about certain alloy systems based on body-centred cubic solvents, but recent investigations by Warlimont and others^4–8 have established that in the Fe–Al and Fe–Si systems, broad equilibrium two-phase regions exist. They have established^5,7, moreover, that such regions are of two types—a normal region, in which the dispersed ordered phase progressively coarsens on prolonged annealing, and an anomalous region in which interfacial and strain energy plays a crucial part and the particles of ordered phase remain permanently very finely dispersed. Strictly, only the first category can properly be described as a Gibbsian two-phase region.