Table of contents

In 1964 the authors published preliminary results of their work on establishing the practical temperature scale in the temperature range from 4.2 K to 20 K. The temperature scale for this range was defined using a germanium resistance thermometer, as its sensitivity and reproducibility of readings are as good as it is required for the most accurate determinations of temperature. At the same time germanium thermometers have noticeable defects among which the main one is non-reproducibility of temperature dependence of its resistance from sample to sample, but no better thermometers for this range are available now.

The results of the well-known works on establishing IPTS in the temperature range from 12 K to 273 K show that even the platinum-thermometers cannot be fitted a sufficiently simple analytical resistance-temperature function. The only method which appears to be acceptable for semiconductor resistance thermometers is the table method.

The establishment of a temperature scale by any method possible involves calibration of the secondary (in our case germanium) thermometers against a primary device reproducing the thermodynamic temperature scale. As a primary instrument we used a gas thermometer without a dead space. Four thoroughly examines germanium thermometers formed a standard group, conserving a temperature scale defined by individual curves T = T(W) for each thermometer of the standard group (the resistance ratio W = R_T/R_H2). Based on smooth W(T) tables, obtained as a result of the standard thermometers' calibration against the gas thermometer, tables have been made, to be used for converting resistance values of the standard thermometers into temperature values.

The temperature in our scale is defined as an average of readings of the standard group thermometers. This paper presents a detailed description of the work on establishing the temperature scale.

A new thermal neutron flux convention is proposed where the neutron slowing down spectrum per unit energy varies as 1/E^1+β (| β ln E | << 1). The neutron capture rate per atom for a detector may be written as nν₀ σ₀ [G_thg(T) + f C(β, T) G_rh(β, G_r)I'/σ₀] where n is the neutron density, ν₀ is the conventional 2200 m/s velocity, σ₀ is the 2200 m/s neutron capture cross-section, C(β, T) is a renormalisation correction to the epithermal flux fraction f at temperature T, h(β) is the correction to the reduced resonance integral I' defined for a 1/E flux, g(T) is the Westcott function and G_th and G_r are the thermal and resonance self-shielding factors. For the case β = 0 this expression reduces to the Westcott form.

Tables of h(β) are presented for a number of commonly used isotopes, and C(β, T) has also been tabulated.

A simple experimental method for measuring β is given which only involves the ratio of the cadmium-covered activities of standard manganese and gold detector foils, and the method has been applied to several different standard neutron flux assemblies.

Hitherto low-field measurements of the gyromagnetic ratio of the proton, γ_p, have involved measurements of all the length parameters describing a rigid coil system producing the necessary magnetic field. This paper shows how, in principle, this task may be reduced to measurements of a single variable parameter. A practicable design is discussed which is capable of taking advantage of the precision of the ⁸⁶Kr length standard. Such an apparatus could be used for maintaining the ampere, and the improved precision of γ_p could be used in conjunction with the values of e/h from the Josephson effect together with values of other precise constants to derive a value for the fine structure constant which is independent of Quantum Electrodynamical theory.

Four closely spaced rings form a capacitor with useful characteristics. Perturbations from the cross-sectional symmetry have a second order effect on the mean value of capacitance between the two diagonally opposite rings. The capacitance is then determined by the mean diameters only and has been computed to an accuracy of one part in one million.

An extensive program of neutron source calibrations using an improved manganese bath technique has been completed. Sources generating four neutron spectra were compared at different concentrations of manganese. The results of this work provide correction factors for fast neutron absorption; the Argonne values markedly disagree with some widely accepted work but agree with computations based on the latest cross sections for oxygen and sulfur, compounded with certain assumptions regarding source spectra. Neutron source rates obtained are (3.223 ± 0.014) × 10⁶ neutrons/sec referred to January 1, 1962 for the NRC International Standard source N-200-1, and (1.1776 ± 0.0043) × 10⁶ neutrons/sec for the NBS-II source referred to January 1, 1965. About half of the assigned error is statistical imprecision and the remainder is an estimate of systematic error (all with confidence coefficient of 0.68). Although the result obtained for the international standard agrees with the world average well within the assigned error, there still exist some distinguishable systematic discrepancies with other published data.