Critical viewpoints on the methods of realizing the metal freezing points of the ITS-90

Publisher Website
Google Scholar
Abstract
The time‐honored method for realizing the freezing point t f of a metal (in practice necessarily a dilute alloy) is that of continuous, slow freezing where the plateau temperature (which is the result of solidifying material’s being so pure that its phase‐transition temperature is observably constant) is measured. The freezing point being an equilibrium temperature, Ancsin considers this method to be inappropriate in principle: equilibrium between the solid and liquid phases cannot be achieved while the solid is being cooled to dispose of the releasing latent heat and while it is accreting at the expense of the liquid. In place of the continuous freezing method he has employed the pulse‐heating method (in which the sample is allowed to approach equilibrium after each heat pulse) in his study of Ag; his measurements suggest that freezing can produce non‐negligible errors. Here we examine both methods and conclude that the freezing method, employing an inside solid–liquid interface thermally isolated by an outside interface, can provide realizations of the highest accuracy; in either method, perturbation, by inducing solid–liquid phase transition continuously or intermittently, is essential for detecting equilibrium thermally. The respective merits and disadvantages of these two methods and also of the inner‐melt method are discussed. We conclude that in a freezing‐point measurement what is being measured is in effect the however minutely varying phase transition, and nonconstitutional equilibrium, temperature t i at the solid–liquid interface. The objective is then to measure the t i that is the best measure of t f , which is, normally, the plateau temperature.