Effect of Impurities on the Triple Point of Water: Experiments with Doped Cells at Different Liquid Fractions

Recent international comparisons showed that there is still room for improvement in triple point of water (TPW) realization uncertainty. Large groups of cells manufactured, maintained and measured in similar conditions still show a spread in the realized TPW temperature that is larger than the best measurement uncertainties (25 µK). One cause is the time-dependent concentration of dissolved impurities in water. The origin of such impurities is the glass/quartz envelope dissolution during a cell lifetime. The effect is a difference in the triple point temperature proportional to the impurities concentration. In order to measure this temperature difference and to investigate the effect of different types of impurities, we manufactured doped cells with different concentrations of silicon (Si), boron (B), sodium (Na) and potassium (K), the glass main chemical components. To identify any influence of the filling process, two completely independent manufacturing procedures were followed in two different laboratories, both national metrology institutes (VSL, Netherlands and UME, Turkey). Cells glass and filling water were also different while the doping materials were identical. Measuring the temperature difference as a function of the liquid fraction is a method to obtain information about impurities concentrations in TPW. Only cells doped with 1 µmol·mol^?1 B, Na and K proved to be suitable for measurements at different liquid fractions. We present here the results with related uncertainties and discuss the critical points in this experimental approach.