Inventing Temperature November 27, 2007

by Hasok Chang

Temperature is one of those physical concepts that are apparently elementary but very difficult on reflection. There is in fact a long history of disputes on the definition and measurement of temperature, and I will highlight two key aspects of those disputes.

(1) How do we know whether our thermometers are functioning correctly? How can we check, for example, whether the mercury in our thermometer expands regularly with increasing temperature, without already having a true measure of temperature (i.e., a correct thermometer)? And how was it possible to develop and test new thermometers for extreme temperatures beyond the points where mercury and all other ordinary fluids boil or freeze?

(2) In order to establish definite temperature scales, we need fixed points to anchor them, the best-known ones being the boiling and freezing points of water. However, these "fixed" points are not as fixed as we might imagine. For example, starting from the 1770s, various reputable scientists reported that the boiling temperature of water depended on various factors such as the type of containers in which the boiling took place (1.2°C higher in glass than in metal, according to Gay-Lussac) and the amount of air dissolved in it (112°C reached with good de-gassing, under normal pressure, according to De Luc). I have reproduced many of these experiments, and confirmed most of their unlikely-sounding results. I will comment briefly on how it was still possible for scientists to use the boiling point as a precise fixed point despite these unruly variations.

Official Website: http://www.iop.org/activity/branches/South_Central/Events/Brighton/page_26032.html