Measurement of air density using specially designed masses

Final Report 2005-03-17

The project partners have manufactured pairs of specially designed artefacts for measuring their apparent mass in air and in vacuum by use of mass comparators installed in air-tight and evacuable chambers. Each of the two artefacts has a mass of about 1 kg and they are designed for having about the same surface area. Their volumes however are essentially different, such that the difference of their apparent mass in air and in vacuum can be measured with a relative uncertainty of about 1 part in 10⁵.
Two travelling artefacts, hollow cylinder A3C and bobbin A3B of PTB have been ciculated among the three partners and the BIPM artefacts Cc and Cp were measured also at PTB.
The results of the measurements allowed the evaluation of the air density prevailing at the measurements in air with considerably higher accuracy than by use of the air density formula recommended by the CIPM. A comparison with the air density evaluation by use of the CIPM formula from data of the air temperature, pressure and relative humidity as well as the CO₂ content in air has shown a significant difference between the two methods.
Is was supposed that the reason for this difference were a wrong value of the molar fraction of argon in ambient air as it is given with the CIPM formula recommendation. Meanwhile, this supposition has been confirmed by direct measurements of the Ar content in air at KRISS.
The results of the project partner’s measurements have been published in: Metrologia 41 (2004), 396-400 and KRISS’s results are in the same issue: Metrologia 41 (2002), 387-395

Progress Report

The two travelling buoyancy artefacts, hollow cylinder HCA3 and bobbin BOBA3 of PTB as well as the two artefacts Cc and Cp of BIPM have been measured at PTB and at BIPM in air and in vacuum. Before the mass comparisons at PTB, the volumes of the BIPM artefacts have been determined by hydrostatic measurements at PTB.
The mass differences between the two pairs have shown significantly different results in air and in vacuum, approximately 5 microgram, when measured at PTB. BIPM confirmed this result for the PTB artefacts only (7 microgram difference). When the artefacts of the same volumes and approximately same surface areas were compared, PTB found unexpectedly large differences (11 microgram each for the pairs HCA3-Cc and BOBA3-Cp). A very probable reason for this are different desorption and adsorption coefficients for the BIPM and PTB artefacts, the latter not having been cleaned since about eight years. BIPM found a similar result only for HCA3-Cc. The different results at PTB and BIPM could have been caused by an instability of the Cp artefact itself or by undesireble behaviour of the balance when weighing this artefact.
The history of the PTB travelling artefacts, having been measured by the participants over eights years, shows a much better reproducibility for the vacuum mass differences than the for the air differences, which were corrected for buoyancy using the CIPM formula and measurements of temperature, pressure, humidity and CO2 content. This observation shows that the procedure of experimental air density determination using such artefacts will probably improve the buoyancy correction and will lead to smaller uncertainties and errors in calibration of mass standards made of stainless steel or other material less dense than platinum-iridium.
It is planned to continue the comparisons in air and in vacuum with BIPM, NPL, PTB and possibly with CEM, which envisages performing vacuum measurements and joining the project.

Further Information

SP (SE) has been project partner for several years and left the project in 2003, because it was not able to do the required measurements in vacuum.