Table of contents

Platinum-iridium mass artefacts are known to pick up carbonaceous contamination from the atmosphere. We examine the likely physical mechanism of carbonaceous ''build-up'', and analyse historical weighing data using this model, to deduce the rate of increase as a function of time.

Cleaning of Pt-10%Ir reference kilograms is currently conducted at the Bureau International des Poids et Mesures (BIPM) by a manual chamois-leather rubbing followed by washing in recondensed water droplets from a jet of steam (nettoyage/lavage). We investigate an alternative approach involving exposure to ultraviolet (UV) light and ozone in air at room temperature and pressure. This process (i) is shown to be more effective than the cleaning/washing procedure for removing hydrocarbon contamination, (ii) involves no physical contact with the reference mass, (iii) is an easily repeatable and reproducible method, since UV light intensity and ozone gas concentration can be measured reliably. Both UV light and ozone in the required amounts can be produced conveniently in situ by short-wave UV lamps which are widely available.

Optimum UV intensities, ozone concentration and cleaning times are given, and recommendations for experiments to validate the procedure on reference kilograms are discussed. If implemented, the cleaning method described should improve the long-term accuracy with which the SI kilogram can be realized by national standards laboratories from ~1 part in 10⁷ to ~3 parts in 10⁹, much the same as would be achieved if a BIPM nettoyage/lavage could be applied regularly by national laboratories. This will, of course, need to be confirmed by conducting weighing trials on real Pt-Ir kilograms.

A polarimetric, nonlinearity-free, homodyne interferometer for vibration measurement is described. Two outputs are provided from this interferometer for determining the value and sign of vibration, the phase of which can be independently controlled by a retarder and a polarizing beam splitter. The nonlinearity is corrected. This system allows a complete determination of both amplitude and phase of a vibrating object. Vibration amplitudes from 6 nm to 7218 nm with standard deviation ranging from 0,9% to 0,02% are demonstrated. The ability of nano-step tracking (about 1,1 nm per step) is also shown in this paper. The dynamic range from dc to 7 kHz is verified.

An international comparison of roundness measurements at the highest level of accuracy has been carried out between eight European laboratories. Three roundness standards with different designs and roundness deviations were measured. One standard was a commercially available type, two standards were developed for the purpose of this comparison. The agreement between the laboratories was within ±20 nm for the standard with the smallest roundness deviation. The comparison was sponsored by the Commission of the European Communities.

Modern absolute gravity meters utilize laser interferometers to measure the acceleration of a retroreflector that is falling freely in a vacuum chamber. This paper describes a novel approach to processing the resulting data by employing algorithms developed for processing perturbed spacecraft orbits. These algorithms allow unprecedented accuracy, flexibility, and ease of treatment of a variety of troublesome error sources including such things as drag effects and local gravitational or electromagnetic fields.

Cryogenic radiometers are recognized as by far the most accurate radiometers for visible and nearinfrared radiation. The now widespread use of such instruments around the world as the basis for optical radiation measurement scales makes it important to carry out a comparison. This will ensure not only that the instruments perform to the uncertainties claimed, a few parts in 10⁴ or less, but also that their operators can disseminate such uncertainties. In this paper results of some pilot comparisons, both direct and indirect using transfer standards, are given. The experimental techniques used at the Bureau International des Poids et Mesures (BIPM) to achieve the highest accuracy are described in detail.

A long gauge in the form of a thick-walled tube is converted to a Fabry-Pérot cavity by attaching partial reflectors to its ends. The length of this gauge is determined from a number of absolute frequency measurements of the cavity resonances. A set of tunable, well-known CO₂ laser frequencies is used for lengths between 1 m and 10 m.

The results of an international comparison between the He-Ne lasers stabilized to ¹²⁷I₂ at 633 nm of the IMGC (Istituto di Metrologia "G. Colonnetti", Turin, Italy) and the IFA-IFTAR (Institute of Atomic Physics-IFTAR, Department of Lasers, Bucharest, Romania) are reported. The mean frequency difference, considering the hyperfine structure components d, e, f and g of the R(127) 11-5 transition, was Δν = ν_RO.2 - ν_IMGCA4/5 = 3,8 kHz, standard uncertainty u = 5,7 kHz, or Δν/ν₀ = 0,8 × 10^-11, standard uncertainty u = 1,2 × 10^-11. The frequency stability, expressed as the relative Allan standard deviation of the beat note between the two lasers, resulted 3,7 × 10^-12 for a 100 s sampling time.

Hyperfine coupling constants of the R(39) 7-4 transition of the B-X system of the iodine molecule are fitted to a resolved hyperfine spectrum. The standard uncertainty of the fit is 11 kHz.

Nicholas has proposed an alternative model to the ITS-90 to represent the temperature-dependent electrical resistivity of platinum. While the results are interesting and there is certainly merit in developing models which better represent the underlying physics, we believe Nicholas has overextended the utility of the model by suggesting that his calculations indicate that the ITS-90 deviates significantly from thermodynamic temperature over the range 70 K to 270 K. We show that the discrepancy between his model and the ITS-90 in this range may stem from the failure of his model to account for the details of the phonon spectrum of platinum. Further, the physical interpretation which Nicholas assigns to his model requires exotic mechanisms which are not supported by theory. While a re-evaluation of the thermodynamic data may indeed be warranted, the Nicholas model is not an adequate physical framework on which to base such an evaluation.