Effect of Impurities on Water Triple-Point Cells

Abstract

Water triple-point cells are the basis for the definition of the kelvin and for the realization of the International Temperature Scale of 1990. The temperature differences between the cells are mainly caused by impurities arising in the cell water from the dissolution of the cell envelope (borosilicate glass or quartz). In order to investigate the effects of such impurities on the realized triple-point temperature, water triple-point cells doped with known amounts of Si and Na impurities ($0.1\,{\upmu }\hbox {mol}{\cdot }\hbox {mol}^{{-1}}$ to $1\,{\upmu }\hbox {mol}{\cdot }\hbox {mol}^{{-1}}$ of Si and $0.2\,{\upmu }\hbox {mol}{\cdot }\hbox {mol}^{{-1}}$ to $2\,{\upmu }\hbox {mol}{\cdot }\hbox {mol}^{{-1}}$ of Na) were manufactured at VSL by adding gravimetric mixtures of a Si standard reference material and ultra high-purity water to the cell high-purity water. Water samples were taken from the manufactured cells, partitioned into three samples, and distributed to different laboratories for isotope and impurity analysis (inductively coupled plasma mass spectrometry, ICPMS). The results of two independent ICPMS analyses were compared with impurity calculations based on the gravimetric data of the prepared mixtures and manufactured cells. One undoped cell manufactured by UME and one undoped cell manufactured by VSL were intercompared at both VSL and SMD to demonstrate the equivalence of the manufacturing processes of UME and VSL. The triple-point temperatures realized by the doped cells and the undoped cell manufactured by VSL were measured at SMD. The results showed that, in doped cells, the equilibration time after the last freezing is directly dependent on the impurity concentration, and the temperature depression of doped triple-point-of-water cells is significantly greater than the values predicted by Raoult’s law for an ideal dilute solution.