We like to think some things are constant, like temperature, and they are as long as everyone agrees.  That does not mean they are accurate.  The metric system is a famous example of a flawed measurement that nonetheless became popular.

Temperature is based on a chemico-physical material property, not on an unchangeable fundamental constant. Some physicists would like to change that.  They call themselves  metrologists - measurement artists who want to be as precise and change the field of worldwide temperature measurement.

Today's 'kelvin' temperature definition is relative, based on the properties of a substance, the triple point of water. However, water does not always equal water - different effects can influence the triple point temperature. A special problem is the dependence on the isotopic composition and on the impurity concentrations. These values can easily vary when substances are contaminated or contain different isotopes – i.e. identical atoms with different atomic masses. For this reason, scientists want to define the kelvin via an unchangeable fundamental constant and to make it more reliable in this way. 

The Boltzmann constant k is such a fundamental constant. It allows conclusions regarding the thermal energy to be drawn from the mechanical energy of particles. Worldwide, a number of research groups are working on the task to define the kelvin via a fundamental constant. If several groups obtain the same result with at least two independent methods we can have a "water-free" definition of the kelvin.  In the case of the metre, this has, for example, already been done via the speed of light. 

For the determination of the Boltzmann constant, which is required for the redefinition of the kelvin, many research groups are using acoustic gas thermometry which has also furnished the most exact values so far. PTB has followed an alternative, completely independent path to rule out systematic error sources and, thus, to place the redefinition on a solid basis: Here, Dielectric-Constant Gas Thermometry (DCGT) is used. The method is based on the density determination of the measuring gas "helium" by means of a capacitance measurement or, in other words: The researchers measure to what extent the gas changes the capacitance of a capacitor. From measurements performed at the triple point of water with different pressures in the measuring capacitor, the Boltzmann constant can be determined by means of fundamental relations.

This task makes extreme demands on metrology. Its realization was possible only with the aid of experts from industry and several other PTB working groups. Pressure measurement at 7 MPa must, for example, be carried out with piston gauges exact to one millionth, capacitance measurement even exact to one part in a billion. The required temperature stability is provided by a large bath thermostat which was manufactured and optimized in cooperation with the national metrology institute of Italy.

The developed set-up now allows DCGT measurements to be carried out at the triple point of water and furnishes a value for k of 1.380655•10-23 J/K. With an uncertainty of 8 ppm it demonstrates that DCGT is suitable for a determination of the Boltzmann constant at the highest level. However, until the uncertainty of 2 ppm aimed at is reached, some difficulties will have to be overcome. The scientists of PTB expect that this will be possible within the next two years and that then, the way for the redefinition of the kelvin will be cleared.

The results were published in Metrologia.