Determination of the liquidus temperature of tin using the heat pulse-based melting and comparison with traditional methods

• Published in: Metrologia 2018-06, Vol.55 (3), p.334-349
• Main Authors: Joung, Wukchul, Park, Jihye, Pearce, Jonathan V
• Format: Article
• Language: English
• Subjects: Extrapolation; Freezing; Heat; Heat pipes; heat pulse-based melting; Heat pulses; Impurities; impurity distribution; Liquidus; liquidus temperature; Loop heat pipes; Melt temperature; Melting; pressure-controlled loop heat pipe; Quenching; Solidus; Square waves; Stability; Temperature; temperature step; Tin
• ISSN: 0026-1394; 1681-7575
• DOI: 10.1088/1681-7575/aa9c8a

In this work, the liquidus temperature of tin was determined by melting the sample using the pressure-controlled loop heat pipe. Square wave-type pressure steps generated periodic 0.7 °C temperature steps in the isothermal region in the vicinity of the tin sample, and the tin was melted with controllable heat pulses from the generated temperature changes. The melting temperatures at specific melted fractions were measured, and they were extrapolated to the melted fraction of unity to determine the liquidus temperature of tin. To investigate the influence of the impurity distribution on the melting behavior, a molten tin sample was solidified by an outward slow freezing or by quenching to segregate the impurities inside the sample with concentrations increasing outwards or to spread the impurities uniformly, respectively. The measured melting temperatures followed the local solidus temperature variations well in the case of the segregated sample and stayed near the solidus temperature in the quenched sample due to the microscopic melting behavior. The extrapolated melting temperatures of the segregated and quenched samples were 0.95 mK and 0.49 mK higher than the outside-nucleated freezing temperature of tin (with uncertainties of 0.15 mK and 0.16 mK, at approximately 95% level of confidence), respectively. The extrapolated melting temperature of the segregated sample was supposed to be a closer approximation to the liquidus temperature of tin, whereas the quenched sample yielded the possibility of a misleading extrapolation to the solidus temperature. Therefore, the determination of the liquidus temperature could result in different extrapolated melting temperatures depending on the way the impurities were distributed within the sample, which has implications for the contemporary methodology for realizing temperature fixed points of the International Temperature Scale of 1990 (ITS-90).