Slow dissolution of the borosilicate container of triple-point-of-water (TPW) cell is widely recognized as the main cause of long-term drift in observed triple point temperature. We add to the available experimental data a comparison of two large batches of TPW cells (67 cells in total) of various ages (from 1 year to 64 years), manufacturers (NRC, VSL, Fluke, Isotech, etc), and materials (borosilicate glass and fused-silica) which was undertaken in 2018. After measuring the TPW temperatures realized by all 67 cells, 12 borosilicate cells were opened and their water was analyzed by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) in order to correlate their impurity content with their respective age and their realized TPW temperature. No direct correlation was observed between the TPW cells age/impurity content and their measured triple-point temperature for neither borosilicate cells nor fused silica cells (Pearson’s correlation coefficient rxy is within the range −0.60  ≤  rxy≤ +0.40 for all the pairs considered). For fused-silica cells, the results indicate that after the isotopic variation in the water source is taken into account, the long-term drift due to the dissolution of glass envelope, if any, is negligibly small: (+0.4(±0.6) µK·yr−1 reported herein). Given that all the fused-silica cells realize the TPW temperature within 100 µK of NRC and VSL national reference cells and since the analyzed time period of 15 years is equal to the average lifespan of a TPW cell, we conclude that fused-silica TPW cells are superior to those made from borosilicate glass.

This paper describes the collaborative project between National Research Council of Canada (NRC) and Centro Español de Metrología (CEM) for the construction and comparison of high temperature fixed points (HTFPs). A hybrid-type crucible that uses the piston method for filling has been jointly developed. A series of 12 high temperature fixed point blackbodies (HTFPBBs) have been constructed, including Cu, Ru-C, Ir-C, Re-C and WC-C covering the temperature range from 1084°C to 2750°C. All these cells were filled in NRC facilities by CEM and NRC staff. Two cells of each type of fixed point were constructed (except for the WC-C cells), using crucibles from two different suppliers in order to study the influence of the density and emissivity of the crucible in the HTFPs’ performance. Moreover, four WC-C cells were made using tungsten powder from two different suppliers, covering all the possible combinations of tungsten powder and crucibles acquired for this project. All cells, and additionally a Pt-C HTFPBB belonging to NRC, have been measured firstly at NRC and, afterwards, at CEM. Both laboratories have calculated the ITS-90 temperatures from their respective measurements and the results have been compared. Differences on the ITS-90 temperatures of the HTFPBBs measured at each lab are within uncertainties of the comparison. After the comparison, CEM kept one Cu cell, one Ru-C cell, one Re-C cell and two WC-C cells. The rest of the fixed points involved in the comparison were sent back to NRC.