The comparability of the determination of the molar mass of silicon highly enriched in 28Si: Results of the CCQM-P160 interlaboratory comparison and additional external measurements

An international comparison study on the accurate determination of the molar mass M(Si) of silicon artificially enriched in 28Si (x(28Si) > 0.9999 mol mol−1) has been completed. The measurements were part of the high level CCQM-P160 pilot study assessing the ability of National Metrology Institutes (NMIs) and Designated Institutes (DIs) to make such measurements at the lowest possible levels of measurement uncertainty and to identify possible difficulties when measuring this kind of sample. This study supports the molar mass measurements critical to disseminating the silicon route to realizing the new definitions for the kilogram and the mole. Measurements were also made by one external research institute and an external company. The different institutes were free to choose their experimental (mass spectrometric) set-ups and equipment, thereby enabling also the comparison of different techniques. The investigated material was a chemically pure, polycrystalline silicon material. The subsequent modified single crystalline secondary product of this material was intended for the production of silicon which was used for two additional spheres in the context of the redetermination of the Avogadro constant NA, required for the revision of the International System of Units (SI) via fundamental constants which came into force from May 2019. The CCQM pilot study was organized by Physikalisch-Technische Bundesanstalt (PTB). Aqueous silicon solutions were shipped to all participating institutions. The data analysis as well as the uncertainty modelling and calculation of the results was predefined. The participants were provided with an uncertainty budget as a GUM Workbench® file as well as a free software license for the duration of the comparison. The agreement of the values of the molar mass (M(Si) = 27.976 942 577 g mol−1) was excellent with ten out of 11 results reported within the range of relative uncertainty of 1 × 10−8 required for the revision of the SI.