Table of contents

Gas-thermometer measurements between 4 K and 100 K have been performed at the Kamerlingh Onnes Laboratory as part of the preparatory work in several institutes for a future, revised, International Practical Temperature Scale. Gas-thermometer temperatures are compared, between 13.8 K and 100 K, with platinum resistance thermometers calibrated on IPTS-68 and, between 4.2 K and 27.1 K, with rhodium-iron resistance thermometers calibrated on the National Physical Laboratory (Teddington) Scale of 1975 (NPL-75).

Results of the measurements, with NPL-75 at 20.3 K as a reference temperature for the gas thermometer, show agreement with NPL-75 within 0.3 mK over most of the range. Differences from NPL-75 of 0.6 mK were found in the range 10 K to 14 K and around 27 K.

Below 50 K, IPTS-68 is shown to be 4 mK to 10 mK higher than thermodynamic temperatures. Between 70 K and 100 K, IPTS-68 is 9 mK lower than thermodynamic temperatures.

A pyrometric method and appropriate black bodies are described for the accurate measurement of thermodynamic temperature above 683 K (410 °C). The pyrometer employs unbiased low-noise silicon detectors of linear response and a dielectric multi-cavity interference filter of 40 nm halfwidth with a sharply-edged band pass. The effective wavelength equals about 974 nm. Special emphasis has been given to determine and to minimize the errors due to temperature gradients inside a black body that is controlled to a very stable temperature T₆₈ utilizing platinum resistance thermometers.

The difference ΔT = T - T₆₈ between thermodynamic and practical temperatures at the antimony point (630.74 °C) was measured to be - 0.150 K ± 22 mK. From 683 K to 903 K the difference ΔT is expressed by the empirical least-squares fit ΔT = T - T₆₈ = a₁ + a₂T₆₈; a₁ = 0.2154 K; a₂ = - 0.4038 × 10^-3. The thermodynamic temperature of the aluminium point has been determined as 933.452 K ± 24 mK. The overall uncertainties are given at the 99% confidence level.

A constant-volume gas thermometer has been used to measure the thermodynamic temperature, and also the second virial coefficient of helium 4, at the triple point of neon (24.5 K). This was carried out by measuring a relative isotherm at 24.5 K referenced to the temperature of the boiling point of hydrogen on the NPL75 temperature scale. The thermodynamic temperature values measured for the triple points of natural neon and neon 20 were (24.557₀ ± 0.001₁) K and (24.539₄ ± 0.001₁) K respectively. The result for natural neon agrees well with the corresponding temperature value of (24.556₅ ± 0.001₁) K inferred from NPL75 and may be related to other laboratory realizations of the neon triple point through the recent international intercomparison of sealed triple-point cells.

A new technique for evaluating the velocity distribution in cesium-beam frequency standards is presented. It is based on the dependence of the amplitude of the central maximum of the Ramsey resonance on the exciting microwave power level. The second-order Doppler shift, typically Δf/f ≃ 4 × 10^-13 for the NRC primary standards, can be evaluated by this method with an uncertainty of about 10^-15, five times the precision attainable previously.

A He-Ne laser at λ = 612 nm, stabilized to saturated absorption features of ¹²⁷I₂ in an external cell [and with no cell inside the laser cavity] is presented. The successive-2-sample deviation for an averaging time of 10 s is not larger than 0.15 kHz or 3 × 10^-13 of the laser frequency. The uncertainty of the laser frequency due to variations of the laser power or of the beam geometry is less than 5 kHz. But dramatic frequency shifts occur when the returning beam touches an obstacle. Single and multiple linear-regression analyses of the pressure and modulation dependence seem to offer a measure of the reproducibility and of technical or physical perturbations. Linewidth and asymmetry of the peaks have been studied, and the frequency differences of 9 hyperfine components have been determined.