This study was motivated by my recognition of some difficulties in the theory of phase transitions. LANDAU* [1] proposed a theory of phase transitions in crystals, covering cases where the crystal symmetry undergoes some change. Briefly and roughly, it excludes the possibility of various types of changes in symmetry. Unfortunately for the theory, various excluded transitions are observed. I do not know of a good reference describing the known exceptions. The discussion which follows refers to cubic-tetragonal transitions, 'which are observed in some superconductors. According to conversations I have had with experimentists, other types of exceptions are observed. It is not impossible to be fooled, for the distinction between a second order and weak first order transition is a fine one. However one assesses this, we have good reason to look at the theory with a critical eye, and to remove its imperfections. In discussing one such exception, ANDERSON • BLOUNT [3] claim to show that such transitions "... usually involve some change in internal symmetry other than mere strain ...". Said differently, thermoelasticity theory should be replaced by a theory involving a larger list of variables in the constitutive equations. If the suggested cure is to work, it should change the conclusions of LANDAU, most probably to decrease the number of changes excluded. Considering the nature of the calculations involved, it seems to me less than clear that merely enlarging the list of state variables will accomplish this, and they do not prove that it will. Thus, I think that we should look at other weak spots in the LANDAU theory. ANDERSON & BLOUNT mention, but discount another, involving matters of continuity. I focus on another. It might well be that thermoelasticity theory is inadequate, but it should be given a fair hearing. As far as I know, we do not have a theory, accepted by knowledgeable workers, to replace LANDAU's. Again considering the nature of LANDAU'S reasoning, it might help if we could attribute to thermodynamic potentials more invariance than is commonly done. Common molecular theories predict a rather unconventional theory of
Read full abstract