The adoption in 1979 of a new definition of the candela, which permitted a detector-based approach to the realization of the unit, has emphasized the importance of high-accuracy absolute radiation detectors. This paper describes a new electrical-substitution absolute radiometer operating at 5 K, based on a standard commercial helium cryostat, which has been developed at NPL for optical radiant-power measurements. The principal advantages of operating the radiometer at liquid helium temperatures are that the detector can be a large, highly absorbing cavity (absorptivity 99.998%) and that exact equivalence of electrical and radiant heating can be clearly demonstrated. It will be the primary absolute detector for a new realization of the candela at NPL and will provide the basis for the development of a wider range of radiometric and spectroradiometric standards. The use of the radiometer to measure the radiant power of an intensity-stabilized laser source at the milliwatt level with an uncertainty of 4 parts in 105 is also described, together with its application in the determination of the responsivity of transfer-standard detectors, such as silicon photodiodes.

The total radiant exitance of a black body at the temperature of the triple point of water, T tp (273.16 K), and at a series of other temperatures in the range from about 233 K ( — 40 °C) to 373 K (100 °C), has been measured by using a cryogenic radiometer. From the measurements at T tp a value for the Stefan—Boltzmann constant or has been calculated: ( r = (5.66967 + 0.00076) x 10 -8 W m -2 K -4 . This is the first radiometric determination of or having an uncertainty comparable with that calculated directly from fundamental physical constants. This measured value differs from the calculated one by 13 parts in 10 5 , which is less than the combined standard deviations of the measured and calculated values. mbined standard deviations of the measured and calculated values. From the measurements of exitance at the other temperatures, values of the corresponding thermodynamic temperature T have been calculated by using Stefan’s fourth-power law. Since the temperature of the radiating black body was also measured by platinum resistance thermometers calibrated on IPTS-68, values of ( T — T 68 ) were obtained. These range from about — (5 + 1.6) mK at 20 °C to — (28 ±2.5) mK at 100 °C and + (5 + 1.5) mK at —40 °C. The results confirm to within a few millikelvins the departure of T 68 from T above 0 °C already discovered by gas thermometry and show that similar departures, but of opposite sign, exist down to the lowest temperature measured, — 40 °C. The uncertainties associated with these new values of T and ( T — T 68 ) are similar to those of the best gas thermometry.

Brief descriptions are given of the design and operating principles of the NPL total radiation thermometer. Further measurements of T-T6, are reported for the temperature range -130 to +6O"C and tabulated with previous measurements made in the range -40 to +lOO°C. A graph is presented of all the published measurements of T-T68 from various independent primary thermometers in the range 50 to 400 K, and their relevance to the construction of a reference function for the proposed ITS-90 is discussed.

Narrow band filter radiometers have been used to measure the spectral radiance of black bodies held at the melting/freezing points of aluminium, silver and gold with an accuracy of 0,04%. Descriptions are given of the optical components and the techniques used for their characterization. As the measurements are absolute, the thermodynamic temperatures of these points can be calculated directly from the Planck radiation law. The values obtained are:T (aluminium) = 933,509 ± 0,027 K, T (silver) = 1 235,009 ± 0,044 K and T (gold) = 1 337,330 ± 0,049 K.These values are analysed with reference to the International Temperature Scale of 1990 (ITS-90) and to the data that were used to derive the scale. The conclusion drawn is that, while the ITS-90 has a sounder thermodynamic basis than its predecessor (IPTS-68), the temperature intervals between the high temperature fixed points, which were thought to be well established, may need further investigation.

No abstract