Table of contents

The Proceedings published in this special issue of Metrologia are of New Developments and Applications in Optical Radiometry III, a meeting held in Davos, Switzerland, from 20–22 September 1990, organized by the Physikalisch-Meteorologisches Observatorium Davos (PMOD). As the title indicates it was the third in a series of such meetings: the first was organized by P V Foukal and took place in 1985 at Atmospheric and Environmental Research, Inc., in Cambridge, Mass., USA; the second was in 1988 at the National Physical Laboratory in Teddington, UK, organized by N P Fox and D H Nettleton. The idea of these meetings is to provide a platform to present and discuss developments and problem areas in optical radiometry among scientists working in different fields, such as metrology and solar radiometry. The choice of Davos as the meeting place was favoured by the fact that the 7th International Pyrheliometer Comparison (IPC VII) was held in autumn 1990 at PMOD. Although IPC is mainly technical and aimed at the worldwide standardization of pyrheliometers (radiometers used for solar measurements in meteorology), both communities—metrology and meteorology—have overlapping radiometric interests and a thorough exchange of ideas is important for the advancement of the field. The ever-increasing number of participants reveals the general interest in the topic and the next meeting, planned for spring 1992 in the United States, will be organized by B Guenther of the NASA/Goddard Space Flight Center, Greenbelt, Md, USA.

The programme was divided into four sessions: Radiometric Scales, Radiometers and Quantum Detectors, Source Radiometry, and Solar Radiometry and Space Applications. A total of 47 papers were presented and most are published here. All contributions were refereed and cleared for publication by a board of four guest editors, one for each session: J L Gardner, L-P Boivin, D H Nettleton and B R Barkstrom. The pre-eminent value of the editors' work is manifested by the high quality and homogeneity of the papers.

The success of the meeting is due to the motivation and enthusiasm of all participants, the commitment of the members of the organizing committee and the fine work of the board of editors. All are gratefully acknowledged. Thanks are extended to the Swiss National Science Foundation for supporting some of the participants from Eastern countries.

The history of solar radiometry since the first pyrheliometer of Pouillet is presented. After the invention of the Ångström and the Smithsonian pyrheliometers around the turn of this century two different "scales" were in use. Comparisons with absolute cavity radiometers developed in America and Europe have been performed since about 1910 which show remarkably accurate measurements in terms of the SI units. However, these results have never been accepted and several rules have been established to reference radiation measurements in the meteorological community and to remedy the unsatisfactory fact of having different "scales". Unfortunately none of these rules led to a reference close to the SI units of irradiance, confusing the issue even more. With the advent of modern absolute radiometers in the late 1960s the situation improved and led to the definition of the World Radiometric Reference in use by the meteorological community since 1981. This reference has an estimated accuracy of 0,3% and guarantees the worldwide homogeneity of radiation measurements within 0,1% precision.

A high-precision absolute radiometer with a thermally stabilized cavity as the receiving element was designed in 1984 at the All-Union Research Institute for Optical and Physical Measurements (VNIIOFI) and used to derive the USSR National Special Standard of the solar irradiance unit. Since that time, the radiometer has been developed further. The absolute radiometer has taken part in the comparisons with the Å 212 pyrheliometer and other absolute radiometers and standards of COMECON countries. This work has made possible the development of a combined radiometric standard based on the absolute radiometer and black-body radiators.

A method is described to derive a radiometric scale for laser pulse energies from a scale for CW laser power. The transfer standard used is a commercial glass-disk calorimeter. The sources of measurement uncertainties have been studied and methods for the experimental determination of corrections are described. The achieved relative uncertainty of the scale is about 0,15%.

The paper presents the results of the first comparison between the solar radiometric scale, the World Radiometric Reference (WRR) and a Système International radiometric scale (SI scale) as disseminated by national standards laboratories, in particular, the scale based on the cryogenic radiometer of the National Physical Laboratory (NPL). The comparison used a number of transfer standards to act as intermediaries between the two scales which limited the overall uncertainty to 0,07%. The final result shows that the WARR and the NPL SI scale agree to better than 0,03%, well within the absolute uncertainty of the two scales.

A new spectral irradiance scale in the near-infrared was realized at the NRC in 1989-1990. Electrical-substitution absolute radiometers were used in conjunction with interference filters to calibrate FEL and NPL/GEC tungsten halogen lamps in the spectral range 700 nm to 1 600 nm. This paper briefly reviews the method used and presents a detailed analysis of the theoretical and measurement uncertainties involved. Results of the measurements are also discussed briefly.

The realization of a new detector-based scale of spectral radiant power with an uncertainty of ± 0,1% is described. The paper reports the use of novel dye laser techniques to characterize the spectral responsivity of a number of interference filter radiometers. These were used to relate the spectral power distribution of a number of incandescent tungsten filament lamps to the spectral responsivity scale realized by the NPL cryogenic radiometer. The results of the new realization have been compared with the previous NPL spectroradiometric and photometric scales, and showed agreement within their estimated uncertainties.

The absolute spectral responsivity at some infrared wavelengths of two germanium photodiodes has been determined by two different methods: comparison with an absolute radiometer (ECPR) and calibration with respect to a black-body simulator. The same geometry and spectral width have been used in both methods. The results, uncertainty sources and their evaluation, and the agreement between the methods, are discussed in this paper.

A Lyot filter has been developed which uses precision liquid crystal waveplates to achieve tuning of the filter pass band over hundreds of nanometres. Tuning is rapid and non-mechanical, and produces a neglible shift in the transmitted image. Transmission is high (50%), and can be characterized using the methods developed for fixed-wavelength interference filters. Such a filter, coupled with a silicon photodiode trap detector whose radiometric response is well known, offers a direct method for linking detector-based radiometry and photometry.

A pilot international intercomparison of spectral responsivity scales at wavelengths important to fibre optics communication was undertaken among national laboratories in the USA, the UK, Germany and Australia. Results are reported for wavelengths of 800 nm, 850 nm, 1 300 nm and 1 550 nm. Based on this study, a larger intercomparison is planned for 1 300 nm and 1 550 nm, using selected germanium photodiodes as transfer standards.

Cryogenic radiometry has already demonstrated an absolute accuracy in the measurement of radiant power of about 0,01%. This article describes advances in the design of practical black-body radiators and absorbers, in the manufacture and measurement of apertures and in the measurement of detector temperature which, together, should allow the accuracy to be improved by a further factor of ten.

Two types of electrical substitution radiometers have been developed at Cambridge Research and Instrumentation (CRI) for applications in metrology. We describe a table-top instrument for calibrating detectors using stabilized laser radiation, a high sensitivity instrument for measuring cryogenic black-body radiation, and an automated system for calibrating photodiodes and thermopiles with several laser lines and a cryogenic radiometer as a reference standard. We also discuss performance characteristics of installed instruments.

A cooled-cavity precision pyrheliometer, designated PCC-3, has been developed from models described earlier. The new instrument has a sensitivity of 0.3 VW^-1 with a time constant of 15 s. Its accuracy is ± 0,08% and its precision ± 0,1%.

Results are presented on the experimental evaluation of the instrumental corrections required by various imperfections in the measuring process for Istituto Elettrotecnico Nazionale electrically-calibrated absolute radiometers. Time constant, linearity, lead-heating effect and response variations across the surface have been evaluated using an ad-hoc algorithm implemented for measuring the low radiant powers involved in the measurements.

The response behaviour of feedback-controlled radiometers has been analysed in order to understand how to design the control system and take measurements with the least measurement time for a given variance or vice versa. For "single-observation" measurements, in which the response is observed only once after settling for each shutter movement, solutions are obtained for the third-order system comprising integral plus proportional control and a first-order filter. Using expressions developed for the variance in the mean of a sequence of partially correlated observations, the analysis is extended to "repeated-observation" measurements which are averages of sequences of observations taken after settling for each shutter movement. Numerical analysis of one simple system shows that there is an optimal number of observations and control system design for any set of values of detector noise level, desired measurement variance, and sampling frequency.

The Earth Radiation Budget Experiment (ERBE) scanning radiometers were used to measure the Earth's radiation fields during the period November 1984 through February 1990. The ERBE radiometric packages were placed into orbit aboard the Earth Radiation Budget Satellite (ERBS) and the National Oceanic and Atmospheric Administration NOAA-9 and NOAA-10 spacecraft platforms. In each radiometric package, thermistor bolometers were used as detection elements for the broadband total (0,2 - 50,0 μm), shortwave (0,2 - 5,0 μm), and longwave (5,0 - 50,0 μm) spectral regions. Flight calibration facilities were built into each of the spacecraft radiometric packages. The flight facilities consisted of black bodies, tungsten lamps, and silicon photodiodes. The black bodies and tungsten lamps were found to be reliable at precision levels approaching 0,5% over a five-year period. The photodiodes were found to degrade more than 2% during the first year in orbit. In this paper, the flight calibration systems for the ERBE scanning radiometers are described along with the resultant measurements.

For detector calibration in the VUV spectral range, electrical substitution radiometers with low uncertainty for radiation powers between 10 nW and 1 μW are being developed. The microbolometers utilize the low thermal capacitance of sapphire plates at the temperature of liquid helium. On the back side of the plate with an absorbing front surface a semiconductor resistor is used as thermometer and the calibration is achieved with an electrical heater. First measurements in the spectral range around 1 keV show satisfactory agreement with the results from a self-calibration technique of a Si photodiode.

Results of numerical modelling of short-wavelength, internal quantum efficiency of two types of p⁺ nn⁺ silicon photodiode are presented. The important conclusions are: (1) interpreting the data obtained from oxide-bias experiments with the help of numerical quantum efficiency models will provide higher accuracy than can be obtained from the conventional self-calibration formula; (2) for high-quality silicon photodiodes, the spectral shape of the internal quantum deficiency (one minus the quantum efficiency) is virtually independent of the density of charge trapped in the oxide, and of the surface-recombination velocity at the oxide-silicon interface.

The paper describes a new solid state detector which fulfils all the requirements of the ideal detector for optical radiation measurements. This "trap" detector is shown to be spatially uniform, linear, stable and to have a predictable relative spectral response so that it can be used to realize a relative spectral response scale with an uncertainty of less than 0,05%. The detector is also shown to be capable of disseminating an absolute spectral responsivity scale with an uncertainty of less than 0,02% at specific laser wavelengths and, independently, of realizing such a scale with an uncertainty of better than 0,05%.

In an earlier publication a silicon photodiode self-calibration boot-strap technique was demonstrated in a 633 nm laser power intercomparison with a cryogenic radiometer. The experiment involved individual silicon photodiodes whose quantum efficiency was determined at 633 nm from quantum efficiency measurements made at 442 nm. Although the results were quite good, there were some problems with that intercomparison that warranted further experimental work, the biggest problem being the instability of the reflectance of the silicon photodiodes. In this paper we report a second comparison of the boot-strap technique with a cryogenic radiometer where the reflectance stability problem has been solved by using three-detector light traps instead of individual silicon photodiodes. For five different trap detectors the average difference between the predicted absolute responsivity at 633 nm and that measured by comparison to a cryogenic radiometer was 0,010% within a combined uncertainty of 0,025%.

The inhomogeneity of responsivity of photodiodes limits the accuracy of their calibration. Inhomogeneity measurements of a Ge and InGaAs photodiode and of an integrating sphere equipped with an InGaAs photodiode are shown as examples. An evaluated set of inhomogeneity coefficients is proposed as a more advantageous and objective measure than original data matrices or their graphical presentation. The PTB facility for the calibration of transfer standards at the wavelength λ = 1 047 nm is described. This takes inhomogeneity effects into account resulting in a relative transfer uncertainty of less than 10^-4.

The photodiode self-calibration technique requires the application of a water drop to the surface of the photodiode to be calibrated. By adsorption of water in the silicon oxide the reflectivity and hence the sensitivity of the photodiode changes. We show that the reflectivity of the detector can be restored to its former value by cleaning with pure ethanol. Influences of the room humidity on photodiode responsivity are discussed.

We use heterodyne detection to measure optical transmittance over an unprecedented dynamic range, with high accuracy and high sensitivity. Our set-up uses a Mach-Zehnder interferometer with acousto-optic frequency shifting to produce a 30 MHz frequency difference between the two light beams, resulting in a beat envelope on the interferometer output beam. We determine the optical transmittance of a filter by inserting it into one of the interferometer arms and measuring the change in amplitude of the optical beat signal. This method allows direct comparison between optical and RF attenuators, ultimately tying optical transmittance measurements to 30 MHz RF attenuation standards in an absolute way. Using this method, we have measured filters that cover 12 decades of transmittance. We discuss a recently developed instrument that improves the accuracy and dynamic range over which the optical-RF connection can be made. This instrument functions as a lock-in amplifier with a dynamic range of up to 180 dB. In addition, we extend the technique into the infrared at 10,6 μm, where we demonstrate that transmittances of filters at room temperature can be measured with a dynamic range of more than 11 decades.

This paper describes a radiometric determination of the melting/freezing point of gold by measurement of the spectral radiance of a black body at that temperature. A simple optical system was used to image the black-body aperture onto a filter radiometer. The advantages and disadvantages of this approach are discussed. The techniques to characterize the optical system are described together with the assessment of uncertainties. The freezing point of gold was found to be 1 337,330 K ± 28 mK.

Large-aperture integrator sources are required to calibrate a number of the satellite instruments which will be used for documenting and understanding global change in the US Earth Observing System (EOS) satellite program. In this paper the salient characteristics of large-aperture sources as used at the NASA/Goddard Space Flight Center are identified. These characteristics are source spatial uniformity, calibration accuracy and output stability. One use of these sources in maintaining the satellite sensor performance in orbit through vicarious calibrations is reviewed.

A transportable hemispherical source is shown to have spatial uniformity of better than 0,3%, have stability of near 1%, and to be capable of being calibrated with an uncertainty near 1,6%. Vicarious calibrations of the AVHRR with a transfer spectrometer on board a NASA ER-2 platform flown at 21 km are shown to be capable of individual calibrations with an uncertainty of 3,5%. Annual trends in the satellite responsivity can be corrected to an uncertainty of 1,5% with this approach.

The Improved Stratospheric and Mesospheric Sounder (ISAMS), to be flown on the Upper Atmosphere Research Satellite (UARS), will make radiometric measurements of thermal emission by a number of atmospheric gases in the intermediate (4,6 μm to 16,3 μm) infrared spectral region. In addition to relative spectral and field-of-view calibrations, an absolute radiometric determination of the integrated spectral throughput has been made using a system of two reference and one intermediate cryogenic black-body targets. The cold reference and intermediate targets were maintained at 77 K and the variable reference target could be controlled in small steps over the range 77 K to 320 K. Calculated emissivities of order ε = 0,9998 could be realized for the large-aperture reference target with a structure of concentric V-grooves with diffusely emitting surfaces.

A three-dimensional gonioradiometer for the measurement of spatial power distributions is described and methods optimizing spatial resolution and duration for the determination of total power (or related photometric quantities) are reported. A silicon photodiode with known spectral responsivity is used as detector (V(λ)-corrected for photometric measurements). The detector is moved by two frames mounted in a Cardanic gimbal suspension to give the instrument a radius of 0,3 m with a resolution of 0,01° for both polar angles in more than 99% of the full solid angle. To ensure constant total radiant power of the luminescent diodes (LED) during the sequential measurements, their electrical power is stabilized using constant electrical current and additional temperature control. Results are given for some LEDs, which are calibrated as reference standards.

The paper describes a method for determination of thermodynamic temperatures by measurement of the irradiance of a black body. Silicon pn-photodiodes, calibrated against a cryogenic radiometer, are used for the absolute irradiance measurements. This method enables the establishment of a radiometrically based high temperature scale, i.e. it requires no reference temperature. Stability tests of the cryogenic radiometer system and results of photodiode calibrations are presented. The uncertainties of the method are discussed with special emphasis on diffraction effects and on the spectral properties of the interference filters used for wavelength selection.

Differential spectroradiometry is based on the determination and analysis of the differential spectral response of photoelectric devices in the presence of bias irradiance that can be varied with wavelength and over many orders of magnitude. The principle is described. Differential spectroradiometry is particularly suitable for the following applications that are reviewed or illustrated by some examples: determination of spectral and weighted responsivities allowing the transfer of the spectral responsivity scale to high levels of irradiance; determination of nonlinearities in the photocurrent-irradiance characteristic; non-destructive characterization of photoinduced effects.

Concave spherical mirrors used off-axis to image monochromator exit or entrance slits to detectors or sources introduce significant astigmatism. This astigmatism can be completely eliminated by introducing a second concave mirror. Applications to spectroradiometric comparisons of sources and detectors are discussed. The use of supplementary slits to remove astigmatism is also discussed.

Liquid-helium cooling of electrical substitution radiometers (ESRs) offers significant improvements in the time constant and sensitivity of these spectrally flat receivers. These cryogenic improvements have been used to good advantage in the design and prototype construction of fast ESR receivers (and receiver arrays) with sufficient sensitivity for NASA Earth remote-sensing applications. We describe some of our results in this area and suggest how the advantages offered by cryogenic ESRs might be useful in achieving the very long-term reproducibility that will be required in 21st-century pyrheliometry.

In order to ensure that Earth Radiation Budget measurement from space continues through the 1990s, France, Germany and the USSR are developing a Scanning Radiometer for Radiation Balance (ScaRaB) to be flown on a series of METEOR-3 Soviet polar orbiting weather satellites. The instrument described in this paper comprises two broad channels (0,2 μm to 50 μm, 0,2 μm to 4 μm) for radiation budget, and two narrower bands (0,5 μm to 0,7 μm, 10,5 μm to 12,5 μm) for scene identification, with spatial resolution at nadir of order 50 km. The set of on-board sources includes black-body simulators for the calibration of long-wave channels, and both tungsten filament lamps and mosaic array diffusers for short-wave calibration. A specific filtering scheme cancels the influence of thermal environment during calibration of the 0,2 μm to 50 μm channel on solar diffusers and lamps. Hemispherical choppers associated with pyroelectric detectors ensure a high immunity to low frequency noise and thermal transients.

The Earth Radiation Budget Satellite (ERBS) and the NOAA-9 spacecraft solar monitors were used to measure the total solar irradiance during the period October 1984 to December 1989. Decreasing trends in the irradiance measurements were observed as sunspot activity decreased to minimum levels in 1986; after 1986, increasing trends were observed as sunspot activity increased. The magnitude of the irradiance variability was found to be approximately 0,1% between sunspot minimum and maximum (late 1989). When compared with the 1984 to 1989 indices of solar magnetic activity, the irradiance trends appear to be in phase with the 11-year sunspot cycle. Both irradiance series yielded 1 365 Wm^-2 as the mean value of the solar irradiance, normalized to the mean Earth/Sun distance. The monitors are electrical substitution, active-cavity radiometers with estimated measurement precisions and accuracies of less than 0,02% and 0,2%, respectively.

A cavity radiometer of the H-F type was included as the total irradiance monitor of the Advanced Photovoltaic Experiment (APEX) aboard the Long Duration Exposure Facility (LDEF) satellite. This is the same type of cavity radiometer which has returned data from Nimbus-7 for almost twelve years. The LDEF was retrieved by the Space Shuttle Columbia on 12 January 1990 after almost six years in space. Initial results show that the electrical properties of the thermopile and heater circuits were unaffected. There was a slight visible change to the painted surface of the cavity, possibly caused by atomic oxygen impingement late in the mission. This apparently has not affected the agreement with ground-based reference standards. Intercomparisons to date indicate agreement of approximately 0,1%, but with an uncertainty at this same 0,1% level. The radiometer was included in the International Pyrheliometric Comparison (IPC VII) in September 1990. In-flight data was taken by an on-board data system (no telemetry) for the period from 7 April 1984 to 16 March 1985. Digital flight data from which solar pointing and irradiance information can be calculated has recently become available.

A group of precision pyrheliometers, designated PVS-7 and PVS-8, for use on the ground, in balloons and in spacecraft, has been constructed and a second-generation solar constant monitor, SCM-2, is under construction. The SCM-2 incorporates a precision solar pointing and tracking system with a pointing accuracy of 0,2°, a value well below the value of 0,5° which is shown to be the limit for precise pyrheliometry. With this type of equipment, high-altitude balloon SC measurements constitute a useful supplement to satellite measurements by providing an absolute link of sufficient accuracy to permit the evaluation of possible SC trends.

Support is given for the suggestion that the mean annual value of the solar constant varies as an algebraic sum of changes in effective temperature and photospheric radius. A set of harmonic constituents covering the period 1979 to 1987 is suggested. The fundamental harmonic used has a period of 10,56 years and harmonics of up to 74th order are described.

A sunphotometer was used to measure the solar irradiance at three wavelengths from rockets and a balloon over a period of three years. The sunphotometer was developed at the Physikalisch-Meteorologisches Observatorium Davos (PMOD) and calibrated by an irradiance standard lamp. The measured solar irradiance agrees within + 1,5% and - 2,0% with Neckel and Labs measurements in the 1960s. Unfortunately, the measured variability cannot necessarily be assigned to solar origin because of insufficient calibration accuracy.

The use of self-scanning photodiode arrays is well established for low light intensity multi-spectral measurement systems such as those used in spectroscopy and astronomy. Using the newly developed EG&G Reticon random access self-scanning linear photodiode array, an instrument has been developed for solar observations which overcomes the difficulties associated with these detectors at high light levels. By coupling this technology with flat-field concave holographic grating technology from American Holographic, a "sunphotospectrometer" has been built. This instrument is presently able to scan 1 024 channels of spectral information in less than 1 s. A prototype instrument is being tested with a second-generation version expected to fly aboard the US Space Shuttle within the next two years. The design of this instrument is described and preliminary spectra are presented to illustrate the measurement of NO₂ using this portable system.

The Solar Backscatter Ultraviolet Radiometer satellite-borne ozone monitoring instrument and its radiometric calibration are described. BaSO₄ diffusers used as part of the calibration transfer optics were found to be unstable, losing up to 6.5% total reflectance at 230 nm over five years. The resulting calibration errors were corrected after the instrument was in orbit by using data gathered when a more stable aluminum diffuser was used. A small spectral irradiance disagreement between FEL and D2 lamps is described.

Accurate calibration of remote-sensing instruments has become increasingly more important for studying long-term changes in our environment. Changes in atmospheric ozone are of great concern and have been observed from satellites for over a decade. Despite this record, long-term trends in ozone remain uncertain, primarily due to uncertainties in satellite instrument calibration after launch. In order to overcome this problem a Shuttle-borne Solar Backscatter Ultraviolet (SSBUV) radiometer is being flown almost yearly in conjunction with the series of SBUV-type ozone sounders that are being flown on NASA research and NOAA operational satellites. The SSBUV calibration must be tracked to 1% or better to provide the requisite calibration transfer so that an ozone trend can be derived from the satellite data set. The first flight of the SSBUV occurred in October 1989. Calibration of the instrument was tracked before and after flight to a precision of 1% through laboratory calibrations, field calibration checks, and inflight calibration checks. As a result of the flight, the instrument experienced a 1% to 2% change in sensitivity in the wavelengths used for measuring ozone.