Abstract
The Minamata Convention on Mercury has created a near-term need to develop alternative fixed points to replace the mercury triple point (Hg TP) for calibration of standard platinum resistance thermometers (SPRTs) on the International Temperature Scale of 1990 (ITS-90). The sulfur hexafluoride (SF(6)) TP is a good candidate to provide adequate "drop-in compatible" replacements for the lowest costs. We report our first results of SF(6) TP realizations performed at the National Institute of Standards and Technology (NIST) using a new series of transportable and refillable triple-point cells. The melting curves are presented at various melted fractions F and compared to evaluate the reproducibility and overall uncertainty for the realizations. We obtained a TP temperature of 223.55587(33) K at F = 50 % and 223.55607(35) K at F = 100 % as a weighted average of realizations using two adiabatic-type cells and two immersion-type cells. (Unless otherwise stated, uncertainties are standard uncertainties corresponding to a 68 % confidence level.) Temperatures were derived using a combination of five different SPRTs as calibrated at NIST on the ITS-90. The data were evaluated over a region of the melting plateau for melted fraction F between 30 % ≤ F ≤ 80 % with a 0.2 mK wide melting range. The results from the immersion-type cells were used to derive an experimental value for the SF(6) TP static head correction of -11.6(1.7) mK/m. This value implies an initial slope of the pressure-temperature (p-T) equilibrium melting line of 1.55 MPa/K, which is in agreement with the value predicted via the Clapeyron equation. The uncertainties of these initial SF(6) TP realizations are limited by uncertainty in the realization of the ITS-90 (0.25 mK) and, to a lesser extent, static pressure head effects and chemical impurities.