Table of contents

The National Institute of Standards and Technology (NIST) operates the newly upgraded Synchrotron Ultraviolet Radiation Facility (SURF III) mainly as a light source for radiometry. SURF III provides continuum radiation from the far-infrared to the soft X-ray spectral range and has its peak output in the extreme ultraviolet. SURF III is a circular-orbit, weak-focusing (single dipole magnet) storage ring, a feature which is advantageous if the synchrotron radiation output is calculated. We report the improvements achieved during a recent upgrade from SURF II to SURF III and our strategy to accurately determine the magnetic flux density, radio frequency (RF), beam current, and beam size, which are the parameters necessary to characterize the source completely.

The Physikalisch-Technische Bundesanstalt (PTB) has installed an X-ray radiometry beamline at the new electron storage ring BESSY II. The synchrotron radiation is monochromatized by a four-crystal monochromator in the photon-energy range from 1.75 keV to 10 keV. The known radiant power, determined with a cryogenic electrical-substitution radiometer (ESR) as primary detector standard, is used for the calibration of detectors such as semiconductor photodiodes. Relative uncertainties of spectral responsivity well below 1 % have been achieved.

The emissivity of a high-temperature black body, used for the realization of the spectral irradiance scale at the Physikalisch-Technische Bundesanstalt (PTB), has been examined experimentally by direct reflection measurements. Measurements at two laser wavelengths show a significant dependence on temperature of the normal effective emissivity of the cavity. The emissivity is very close to the ideal value of unity at temperatures above 2000 K. Preliminary experiments on the reflectance and emissivity of graphite, the material from which the black-body cavity is made, were carried out at room temperature. Using the results of these measurements as input parameters for Monte Carlo calculations of the effective emissivity of the cavity, good agreement with the measurements is obtained.

Recent work at the National Institute of Standards and Technology (NIST) with superconducting transition thermometers suggests that substantial improvements could be achieved in the detection capability of electrical-substitution radiometers, which are widely used as the most accurate absolute standards of radiant flux and irradiance. Cambridge Research & Instrumentation (CRI) has constructed liquid-nitrogen and liquid-helium-cooled prototype radiometers with responsivities of 550 mK/mW and 3.5 K/mW, and natural time constants of 30 s and 3 s, respectively. Our results with the liquid-nitrogen-cooled radiometer show that the high-T_c superconductor YBCO sensor has similar detection capability to a germanium resistance thermometer operating at 4 K. In this prototype radiometer, we were able to control receiver temperature to better than 1 µK and to attain a root-mean-square (rms) noise level of 1.6 nW. Our ongoing work focuses on the absolute accuracy achievable with such a liquid-nitrogen-cooled radiometer, using an optimized thermal design. The high detection sensitivity offered by the Nb sensor in the liquid-helium-cooled prototype radiometer achieved a temperature stability of 40 nK (rms), which implies a noise-equivalent power of 10 pW. Our measurements indicate that 1 pW should be feasible with optimized electronics, a lower-T_c sensor, and better control of heat-sink and scene drifts through implementation in the NIST Low Background Infrared (LBIR) chamber.

High-temperature black bodies used as primary standards of spectral irradiance have to meet several requirements. One important demand is that the spectral irradiance of the black body is uniquely determined by Planck's radiation law, where the (radiometric) temperature of the black body is the only parameter that determines its relative spectral distribution. By combining determinations of the radiometric temperature of a black body over a wide temperature range with the corresponding spectral measurements, black-body spectra for different temperatures can be compared. This allows the use of self-consistency checks to ascertain whether (i) the temperature determinations are coherent over a wide temperature range; and (ii) the different spectral distributions show any deviation from Planck's law. Using this method of self-consistent calibrations, new absorption bands in the spectrum of a black body were discovered at very high temperatures. The level of absorption increases with temperature and depends on the operating conditions and imperfections of the black-body system. Possible origins of this effect are discussed and modifications proposed to avoid such absorption bands.

Filter radiometers were used in a bilateral comparison of the radiometric temperatures of a high-temperature black body from 2100 K to 3100 K. The measurements took place over a ten-day period in October 1998 at the Physikalisch-Technische Bundesanstalt (PTB, Braunschweig, Germany) using filter radiometers from the PTB and from the National Institute of Standards and Technology (NIST, Gaithersburg, USA). The comparison revealed that the radiometric temperatures determined using the NIST and the PTB filter radiometers differed progressively from 2.4 K at 2200 K to 5.1 K at 3200 K, with the NIST temperature values always lower than the PTB values. Differences in the measurements of absolute spectral responsivity were the main cause of the observed temperature differences.

Correlated colour temperature (CCT) of a source is estimated from position in CIE 1960 (u, v) colour space in relation to the black-body locus. In this paper, the methods of the ISO Guide to the Expression of Uncertainty in Measurement are applied to derive analytical expressions for uncertainty in u and v chromacity coordinates and CCT, including correlation effects. Uncertainty in CCT is found by fitting the form of the black-body locus and relevant derivatives. Examples are given for calibration of four typical sources of spectral distributions. Accurate CCT is determined by a search technique. An alternative method using fitted functions to solve for CCT using the geometry in (u, v) space is presented. An uncertainty in CCT of a few kelvin can be achieved.

A complete set of total solar irradiance measurements from space is available dating from November 1978, yielding a time series of more than twenty years. From measurements made by different space radiometers (HF on NIMBUS-7, ACRIM I on SMM, ACRIM II on UARS and VIRGO on SOHO) a composite record of solar total irradiance can be compiled. The corrections needed for each radiometer, mainly to compensate for degradation, are based on a detailed study of the behaviour of the VIRGO (Variability of Solar Irradiance and Gravity Oscillations) radiometers. Moreover, several errors arising from operational changes or interruptions, which influence the overall uncertainty more than corrections for degradation, have to be corrected. A composite of the corrected time series is constructed in order to estimate the long-term stability, which is much better than the estimated 1 part in 10⁵ per annum deduced as an upper limit from VIRGO radiometry. The absolute accuracy of the data, however, is more difficult to estimate.

Since the beginning of 1996, the vacuum-ultraviolet telescope and spectrograph SUMER (Solar Ultraviolet Measurements of Emitted Radiation) on the Solar and Heliospheric Observatory (SOHO) of ESA and NASA, has obtained radiance spectra of many features of the quiescent and active Sun with high spectral and spatial resolutions. Irradiance measurements of line emissions have also been performed in the wavelength range 58.4 nm to 154.8 nm. The instrument was calibrated against the electron storage ring BESSY I as a primary source standard with the help of a transfer source standard based on a hollow-cathode discharge lamp. A thorough cleanliness programme, specifically aimed at chemical contamination control, resulted in an excellent radiometric stability of the normal-incidence optical system as well as of the detectors. This was verified under operational conditions by various techniques employed during the SOHO mission, such as line-ratio studies, star observations and comparisons with other instruments. The observations provide vacuum-ultraviolet radiometry of the Sun in many emission lines of atoms and ions with a relative uncertainty of 0.15 (1 σ) for the wavelength range 54 nm to 125 nm.

Uncertainty in terrestrial solar radiation measurements of a few watts per square metre (1 W/m² to 2 W/m²) is needed to validate estimates of the Earth's radiation balance derived from satellite data. The characterization of solar energy resources for renewable energy technologies requires similar accuracy for economical technology deployment. Solar radiation measurement research at the National Renewable Energy Laboratory addresses calibrations, operational characteristics, and corrections for terrestrial solar radiation measurements. The characterization of field instrument geometric and thermal responses contributing to radiometric errors is described, together with correction schemes reducing the uncertainty of broadband irradiance measurements from tens of watts per square metre to a few (±2 W/m² to ±5 W/m²). Such improvements in accuracy reduce the time and labour required for detecting and quantifying trends in terrestrial solar radiation and possible changes in the Earth's radiation budget.

The lack of accurate calibration sources is preventing current sensor technology from achieving the minimum possible uncertainty levels in remote-sensing radiometric measurements. This has been established in a recent study undertaken at the National Institute of Standards and Technology (NIST) following the identification of the problem at the Calibration Source Requirements Workshop at the NIST in 1997. This paper presents the concept of establishing high-accuracy calibration sources in space, utilizing the International Space Station (ISS) as a platform. For example, a retrievable total solar irradiance radiometer standard could be deployed on the ISS in order to establish the Sun as a calibration source of the required accuracy. Periodic measurements of solar irradiance with this radiometer standard would help to establish the Sun as a radiometric standard with the accuracy needed for various radiometric sensors on other platforms in space. As the radiometer on the ISS would be retrievable, its traceability to the International System of Units (SI) could be maintained by periodically bringing it down for calibration with SI-traceable standards maintained in ground laboratories. Other suitable radiometric standards on the ISS could also establish the Moon, the stars and various ground sites as calibration sources.

The Total Irradiance Monitor (TIM), an active-cavity solar radiometer, is to be launched in 2002 on the SORCE mission of the Earth Observing System (EOS). The relative uncertainty (1 σ) will be better than 10⁻⁴, i.e. 100 parts per million (ppm) (1 σ) with a noise level of < 1 ppm each 500 s. Sunlight passes through a shutter, a 50 mm² aperture, and is then absorbed into a silver cavity blackened inside with nickel phosphorus. High thermal-conductivity diamond insulators at the electrical terminals help to localize thermal nodes. Four cavities are aligned side-by-side at the rear of a 2 kg heat sink. The flight "standard digital watt" is derived from a high-precision voltage and a pulse-width-modulator; this eliminates the need for a square root in the servo loop. We determine the irradiance from the in-phase sinusoidal component at the shutter frequency. This phase-sensitive detection allows more accurate characterization than traditional time-domain methods. We calibrate the aperture transmission integral over area against a chrome-on-quartz ruling, transferring by charge-coupled device (CCD) images. In flight, we recalibrate the servo-loop gain, pointing variations, and dark signal. We determine the equivalence ratio from models and by laser measurements of the cavity parameters.

A general modelling formalism is used to extend near-field point spread function (PSF) measurements over a wide-field off-axis angular range, covering a dynamic range of 5 to 6 orders of magnitude, using bidirectional reflectance distribution function (BRDF) measurements of the scan mirror and other key optical elements. Line spread function (LSF) model results were compared with measured near-field LSF measurements for the Moderate Resolution Imaging Spectroradiometer (MODIS) of the Earth Observing System (EOS), demonstrating excellent agreement between model and measurements. When realistic effects of sensor scattered light are taken into account, significant radiometric bias errors are produced near high-contrast structured scenes (e.g. bright clouds over dark oceans and land; broken snow and ice scenes). Image restoration using an asymptotically exact PSF is shown to produce results significantly different from those produced by traditional near-field 5 × 5, 7 × 7, ..., PSF kernel matrix inversion techniques. The results reported have implications for future remote-sensor specifications and testing, in-flight and surface-based calibration comparisons, and the assessment of radiometric bias errors in the presence of moderate- to high-contrast Earth scenes.

The Spectral Irradiance Monitor (SIM) will measure the solar spectral irradiance between 0.2 µm to 2.0 µm to a relative standard uncertainty of 3 × 10⁻⁴ (1 s) and with a noise level of <1 × 10⁻⁴. The instrument will be launched in 2002 on the SORCE mission of the Earth Observing System (EOS). Sunlight passes through a 0.3 mm × 7 mm entrance slit and is modulated by a shutter at 0.1 Hz. Light then travels 400 mm to a Fery prism that disperses and focuses the light at the focal plane containing five separate exit slits, with an electrical-substitution radiometer (ESR) and four photodiode detectors. The ESR consists of two back-to-back 1 mm × 10 mm diamond bolometers blackened with nickel phosphorus. The bolometer bridge, electrical power replacement, ESR temperature regulation, and prism rotation are under digital signal processor control. The measurement equation and the uncertainty budget for the SIM instrument are presented along with plans for the pre-flight characterization of the instrument.

Atmospheric optical depths are determined by relating ground-based measurements of direct solar radiation to the extraterrestrial value, I₀, that a filter radiometer would read outside the atmosphere. Usually I₀ is determined by the Langley extrapolation technique from a high-altitude site, where clear and highly stable atmospheric conditions may be found. Alternatively, I₀ can be measured in situ from a stratospheric balloon experiment. We have employed both methods and found agreement to better than 1 %. Filter radiometers tend to change over time, especially when used operationally outdoors. Absolute calibrations in the laboratory are used to monitor the radiometric stability of filter radiometers at the Physikalisch-Meteorologisches Observatorium Davos (PMOD/WRC, Switzerland). A spectral calibration facility based on a calibrated trap detector from the Physikalisch-Technische Bundesanstalt (PTB, Germany) is used to relate the filter radiometer to an accurate and long-term traceable standard. An FEL-lamp-based standard, previously used for several years, was compared with the new trap standard via a filter radiometer at four wavelengths between 368 nm and 862 nm and revealed a systematic difference of the order of 5 %. The link between radiometric and I₀ calibration is the value of the extraterrestrial solar spectrum at the filter radiometer wavelengths which can be determined from these two calibrations and compared with published values.

The National Institute of Standards and Technology (NIST), with the support of the National Aeronautics and Space Administration (NASA), has built several transfer radiometers that are used in radiometric measurement comparisons to validate spectral-radiance levels assigned to integrating-sphere sources by the end user. The transfer radiometers are characterized and calibrated at the NIST, and function as direct links in the spectralradiance measurement chain. This paper presents a historical analysis of a calibration method that utilizes a broadband source. Previously observed discrepancies between two radiometers are reconciled as a result.

Using a high-temperature black body (HTBB) and filter radiometers calibrated for absolute spectral power responsivity, the spectral output of the black body, whose radiance temperature was determined using the filter radiometers, was used to assign spectral radiance to tungsten strip lamps with a prism-grating monochromator. The spectral radiance of the HTBB was also determined using signal ratios to a tungsten strip lamp calibrated using a scale derived from the radiometric temperature determination of a gold-freezing-point black body. The radiance temperatures found using the two methods were in agreement to 0.5 K near 2900 K, and from 260 nm to 1050 nm, the spectral radiance of the HTBB did not differ more than 0.5 % in radiance from a single-temperature Planck's law as determined using the tungsten strip lamp.

An electrically substituted bolometer (ESB) has been developed at the National Institute of Standards and Technology (NIST) to serve as a portable transfer-standard detector over the wavelength range 200 nm to 20 µm. The ESB is designed to provide a direct transfer of the optical power scale from the NIST High Accuracy Cryogenic Radiometer (HACR) to the NIST spectral-comparator facilities in the infrared region, where silicon-photodiode trap detectors cannot be used and where the noise of pyroelectric detectors currently limits the uncertainty. The noise floor of the 8 mm diameter active-area ESB approaches 10 pW/Hz^1/2 at 15 Hz, corresponding to a detectivity of the order of 10¹¹ cm Hz^1/2/W. This is an improvement by a factor of almost 1000 on that attainable from a similar-sized room-temperature electrically calibrated pyroelectric radiometer (ECPR). The ESB is linear from the noise floor to 1 mW (the power range of the HACR), whereas previous helium-cooled bolometers developed at the NIST have a similar noise floor but are non-linear above about 10 µW and so cannot be directly calibrated against the HACR. Because of its low noise, linearity over a wide dynamic range, and the spectral and spatial flatness of its response, the ESB is finding other applications beyond that which motivated its development.

Cryogenic radiometers are widely used as national primary standards for radiometric and photometric measurements. The most common comparisons are indirect, using various types of trap detector. The lowest limit of relative uncertainty achievable by indirect comparisons is not much less than 2 × 10⁻⁴ (k = 1). To improve this level of uncertainty, it is necessary to compare the cryogenic radiometers directly at the same position using an identical beam. In this case error sources such as beam size, beam power and scatter are significantly reduced. The comparison between the mechanically-cooled cryogenic radiometer (MCR) of the National Physical Laboratory (NPL), and the liquid-helium-cooled radiometer (LaseRad II) of the Physikalisch-Technische Bundesanstalt (PTB), was the fourth in a series of direct comparisons worldwide. It took place in the best environment presently available, at the transfer-optimized cryogenic radiometer in the class 1000 clean-room centre at the PTB in Braunschweig (class 100 is available). The two radiometers were sequentially positioned by a translation stage in the power-stabilized laser beam. In four different series of measurements, within two weeks, the average stated relative difference of the measurements of radiant power with these two very different cryogenic radiometers was only 3.2 × 10⁵, which is currently the lowest obtained in a radiometric comparison of absolute scales.

The spectral irradiance of lamp standards can be stabilized within a limited spectral range using a suitable selective detector to control the irradiance of the lamps. The capability of one commercial lamp facility incorporating a computer-controlled power supply was tested in an international comparison. A set of three lamps and the detector-stabilized power supply were circulated between the Physikalisch-Technische Bundesanstalt (PTB, Germany), the Helsinki University of Technology (HUT, Finland) and the Swedish National Testing and Research Institute (SP, Sweden) for the comparison, EUROMET project No. 475.

We report the results of the first laser high-power measurement comparison between the Physikalisch-Technische Bundesanstalt (PTB, Germany), and the National Institute of Standards and Technology (NIST, USA). Laser power transfer standards were calibrated at both national standards laboratories between 82 W and 127 W at 1.06 µm and between 85 W and 554 W at 10.6 µm. Relative agreement between the standards of the two laboratories was demonstrated to lie between 5 × 10⁻³ and 7 × 10⁻³, which is well within the combined uncertainties.

This paper presents the results of investigations of the spectral responsivity in terms of radiant power within the infrared spectral range. Large-sized InGaAs photodiodes of 10 mm × 10 mm and an InGaAs trap detector were studied within the spectral range 500 nm to 1900 nm. The spectral responsivity of this reflection trap detector with three photodiodes is compared with that of a single InGaAs photodiode (photodiodes all of the same type). The temperature effect on the spectral responsivity is measured within the temperature range 10 °C to 30 °C. The results, which are important for calibration of transfer standards for optical communication, are discussed in terms of spectral quantum efficiency.

The performance of spectral irradiance transfer-standard lamps has long been a concern in the radiometry and photometry communities. The adoption of a modified FEL-type lamp has recently been proposed as the most suitable transfer standard for spectral irradiance measurements. While data have been published on the performance of these lamps when operated continuously and in a laboratory environment, however, little information is available on their performance under typical usage. The results are presented of "real-life" testing of over thirty FEL lamps both pre-selected, according to published procedures, and direct from the manufacturer. When operated vertically all of the lamps tested suffered terminal coil collapse after between 140 h and 250 h burn time. The reproducibility of the FEL lamps after transportation is also investigated.

Charge-coupled-device (CCD) video cameras have become very important tools in the field of "image science", not only for their good performance as optical cameras but also for their ability to obtain radiometric readings of pixels. This paper reviews calibration procedures for these cameras, concentrating on the radiometric rather than other optical aspects, and proposes new calibration methods as a way of reducing the uncertainty of measurements made with these cameras.

Based on experience acquired over many years in operating high-power, stationary arc plasmas for quantitative spectroscopy, the Physikalisch-Technische Bundesanstalt (PTB) has begun to investigate the suitability of argon arc-plasma radiation for radiometry. We have demonstrated, through the design and operational control of the arc source, that for example argon plasma generated at 150 A and a pressure of 200 kPa provides radiation with a stability of about 10⁻³. We also state a long-term reproducibility of better than 5 × 10⁻³ because all operational and geometrical parameters affecting the plasma radiation can be controlled with a relative uncertainty of at least 10⁻³. The radiation properties have been investigated and carefully characterized in the wavelength range 250 nm to 10 µm at the new PTB detector-calibration facility.

The Scanner for Radiation Budget (ScaRaB) instruments are a series of Earth Radiation Budget (ERB) radiometers of which the first flight model (FM1) was launched on a Russian METEOR satellite in 1993. A second model (FM2) was launched on the Russian RESURS satellite in July 1998. The FM2 instrument worked perfectly until the failure of the platform transmitter in April 1999. Considering both missions, ERB data were obtained over sixteen months with ScaRaB. In order to obtain a good correlation between calibrations on the ground and in space in the long-wave spectral domain, the spectral response has been re-estimated using an original method. As the solar ground calibration and the integrating-sphere calibration are conducted in an enclosure purged with dry nitrogen, the detector gain differs from the gain under vacuum, and a correction factor must be applied. A new correction coefficient has been determined through analysis of the ScaRaB data sets, leading to satisfactory agreement between long-wave vacuum calibration and solar calibration.

The design of a new detector-stabilized spectral irradiance source is presented, together with an evaluation of its performance. This new source consists of a tungsten-halogen lamp that is current-controlled using feedback from an array of filtered photodiodes. The bandpass and central wavelength of these filtered detectors may be selected specifically to meet the required application of the source. The lamp is fully enclosed (except for the exit aperture) in an air-cooled, light-tight housing that minimizes unwanted light pollution. The lamp current is stabilized using feedback from one photodiode; the recording of the non-stabilization channels allows a real-time check on the effective colour temperature of the lamp, and provides an instant diagnosis of the state of calibration of the lamp. The calibration data for each lamp (in its associated housing) are stored on a microchip within the lamp housing, which allows the lamp to be transported independently of its associated control electronics for calibration purposes. The performance of the lamp indicates ageing rates of better than 2 × 10⁻⁵ h⁻¹, across the ultraviolet-A (UV-A) and visible spectral regions.

We have designed and evaluated an optical trap detector that is simple to fabricate and convenient to use. The trap geometry is a triangular tube similar to a kaleidoscope, with two photodiodes in place of a mirror on each side. One version incorporates six 10 mm × 10 mm silicon photodiodes in a detector having a 7 mm diameter aperture. Spectral responsivity measurement results indicate that the quantum efficiency, evaluated at 10 nm increments from 450 nm to 950 nm, is greater than 99 %. Another version incorporates six 18 mm × 18 mm silicon photodiodes in a detector having a 12 mm diameter aperture. Spectral responsivity measurement results from this instrument are similar to those from the other one, but the quantum efficiency is lower. Spatial uniformity measurement results for the larger detector indicate that, over the 12 mm diameter area, response variations are less than 0.05 %. We have evaluated several fields-of-view and confirmed that, in a special case, the detector geometry will accommodate an input beam having a divergence as large as 14°. Design considerations are briefly described and may be used to create similar, scaled versions with other photodiode types and sizes.

The spatial non-uniformity of a newly developed electrically calibrated radiometer system is characterized. The detector is a thin-film thermopile with an integral electrical heater. This paper shows that the local responsivities of these detectors have a rotational symmetry and a quadratic dependence on the distance to the centre of the detector. Typical measurement errors, which can be caused by applying different beam sizes, different power distributions of the beams, and different beam positions impinging on the detector, are investigated. A significant reduction of measurement uncertainties is achieved using specific corrections. The influence of spatial non-uniformity on responsivity measurements is experimentally checked with various types of beam.

The quantum efficiency, η, of gold and copper photocathodes has been measured in the photon-energy range from 50 eV to 10 keV. The experiments were performed with monochromatized synchrotron radiation at the radiometry laboratories of the Physikalisch-Technische Bundesanstalt (PTB) at the electron storage rings BESSY I and BESSY II. Relative standard uncertainties of η lower than 0.036 were achieved using a cryogenic electrical-substitution radiometer as the primary detector standard. The samples were cleaned, stored and examined in an ultra-high-vacuum chamber. The quantum efficiency of cleaned gold and copper was found to be reproducible within 2 parts in 10² to 3 parts in 10². A simple model assumes to be proportional to the product of absorption coefficient, μ, and photon energy, hν. The relative course of the absorption coefficient was also measured with relative standard uncertainties below 0.04. As proportionality was only locally fulfilled, a more sophisticated model for quantum efficiency was developed.

Changes in beam geometry during the calibration and subsequent testing of optical detector standards can cause errors which appear as uncertainties in the responsivities of the detectors. These uncertainties, caused by non-ideal detector input characteristics and changes in input beam geometry, can be decreased significantly if the detector response is always corrected to a reference beam geometry such as a collimated beam normally incident on the detector surface. An integral directional error can be determined for radiant power, luminous flux, irradiance and illuminance detectors by calculating the weighted integral of the product of the measured source radiance and the directional error over the solid angle produced by the light source. A conical directional error is introduced, and used to describe the case of circular and uniform radiant sources (e.g. exit port of an integrating sphere) and circularly symmetric detectors. An evaluation is presented of the conical directional error, a directional response correction factor, the results of directional response measurements, and directional response correction factors for some practical standard detectors (such as diffuser-input InGaAs and integrating-sphere-input silicon irradiance meters, silicon photodiodes, and pyroelectric radiant power and irradiance measuring detectors).

A theoretical model, whose validity was confirmed experimentally, has been used to investigate the polarization properties of silicon photodiodes as functions of the angle of incidence with wavelengths ranging from 10 nm to 800 nm. From this, spectral differences in the responses of silicon photodiodes between a collimated beam and a beam with angle divergence are derived. The calculation shows that divergent beams (both isotropic and Gaussian beams) usually give lower responses than those for a collimated beam except in a limited spectral region (approximately 120 nm to 220 nm for a Si photodiode coated with a SiO₂ layer of thickness 27 nm). The decrease in response of the Si photodiode with 27 nm SiO₂ for the isotropic divergent beam with half apex angle of the beam cone of 60° reaches a maximum of 14 % at a wavelength of 44 nm and the increase reaches a maximum of 7.5 % at 163 nm. Similar results for a Si photodiode coated with an 8 nm oxide layer and a divergent beam with Gaussian angular distribution were also obtained.

The results are presented of absolute spectral responsivity measurements for SPPD11 and SPPD11-04 pulse-type silicon semiconductor detectors (developed by the Research Institute of Pulse Technique, Moscow), carried out over the X-ray energy range from 1.5 keV to 20 keV with a relative uncertainty (k = 1) of typically 5 × 10⁻². In our measurements, the detector is positioned in a calculable synchrotron radiation flux behind filters of well-known transmittance. The spectral responsivity of the detector is restored on the basis of measurement data using a set of integral equations. Two procedures have been developed to improve the calibration accuracy: one for the experimental determination of the X-ray transmittance near the K absorption edge of the filters used, with a relative uncertainty not exceeding 2 × 10⁻²; and another for the estimation of the angular spread of the electrons.

A new ultra-high-temperature black body designed and constructed at the All-Russian Research Institute for Optophysical Measurements (VNIIOFI, Russian Federation) has recently been delivered to the National Physical Laboratory (NPL, UK). The black-body cavity consists of a series of pyrolytic-graphite rings and is designed to operate at temperatures of up to 3500 K. Preliminary results show that the cavity uniformity varies with operating temperature and change in shape of cavity bottom. Results of active and passive temperature control are also given.

Methods used for the establishment and dissemination of primary radiometric quantities are reviewed together with the performance and use of filter radiometers, which are dominating advances in optical radiation measurement. The possibility is also discussed of using many of these advances in other metrological sectors, such as in thermal metrology to improve the realization and dissemination of the kelvin.

The Physikalisch-Technische Bundesanstalt (PTB) has established spectral responsivity standards for photodetectors in the wavelength range between 200 nm and 400 nm based on cryogenic electrical-substitution radiometers. In order to obtain standards over a continuous wavelength range with low uncertainty, combined use is made of both the dispersed synchrotron radiation of the storage ring BESSY I and quasi-monochromatic laser radiation. Taking advantage of the specific properties of the monochromatized synchrotron radiation and the laser radiation, relative standard uncertainties between 10⁻³ and 5 × 10⁻³ are achieved in the calibration of trap detectors in the spectral range 400 nm to 200 nm.

An experimental set-up designed for the calibration of near-infrared detectors against a cryogenic radiometer, using a tuneable laser diode as source, is described. A complete characterization of the cryogenic radiometer has been performed at the Instituto de Fisica Aplicada (IFA, Madrid), at wavelengths of 1300 nm and 1550 nm. Two different secondary standards, an electrically calibrated pyroelectric radiometer and an InGaAs-based integrating-sphere radiometer, have been calibrated with relative uncertainties of less than 10⁻³ at these wavelengths.

The Physikalisch-Technische Bundesanstalt (PTB) has developed a spectral responsivity scale in the spectral range 0.95 µm to 1.6 µm. The responsivity scale was established by calibrating InGaAs photodiodes against the PTB radiation thermometry cryogenic radiometer (RTCR) using several laser sources: a Ti:sapphire laser (0.850 µm to 1.02 µm); a Nd:YAG laser (1.064 µm); and two diode lasers (1.31 µm, 1.55 µm). At intermediate wavelengths the responsivity was interpolated with the help of thermal cavity detectors of high emissivity. In this way the high accuracy of the scale at the laser wavelengths can be maintained over almost the entire spectral region. A detailed discussion is given of the uncertainties achieved in the calibration with the RTCR and the interpolation with the thermal detector.

Realizations of the units of luminance and spectral radiance at the Helsinki University of Technology (HUT) are presented. These realizations are linked to HUT units of luminous intensity and spectral irradiance using a characterized photometer, a spectroradiometer and an integrating-sphere light source. A new method for determining the spatial uniformity of the output of the integrating-sphere source is described. The uncertainty analysis indicates a relative expanded uncertainty of 3.6 × 10⁻³ (coverage factor k = 2) for the realization of the unit of luminance. The expanded uncertainty for the realization of the unit of spectral radiance varies between 6 × 10⁻³ and 2.5 × 10⁻² in the wavelength region 360 nm to 830 nm.

A spectral radiance responsivity (SRR) scale derived from the spectral irradiance responsivity scale has been realized on the newly developed NIST facility for Spectral Irradiance and Radiance Responsivity Calibrations with Uniform Sources (SIRCUS). The scales were directly derived from the primary standard cryogenic radiometer using transfer standard irradiance trap detectors and laser-based uniform sources. Silicon radiance meters of standards quality have been developed to hold the SRR scale. The radiance meters were characterized and optimized for size-of-source effects. The optimized meters were calibrated for SRR against the monochromatic uniform source. In addition to the SIRCUS calibrations, preliminary tests were also made on the Spectral Comparator Facility (SCF) of the NIST. These relative SRR tests were made in both imaging and non-imaging measurement modes. The results of the different SRR calibrations are compared and evaluated.

The spectral radiance of a black body can be calculated from Planck's radiation law when the black-body temperature is known. Therefore, the spectral radiance of a black-body radiator of accurately known temperature may be used as a reference for the establishment of a spectral radiance scale. In this paper we investigate the feasibility of using a copper-freezing-point (1357.77 K) black-body radiator to directly calibrate gas-filled tungsten strip-filament lamps operating at a radiance temperature of 2300 K (at 650 nm). A silicon photodetector with a variable-gain amplifier, as well as optical apertures of different sizes, were used to measure the large radiance ratios observed. The comparison was performed at eight wavelengths between 400 nm and 900 nm using interference filters. The major problem encountered was out-of-passband infrared (IR) leakage of the interference filters, particularly those below 550 nm. We discuss the use of sharp-cut glass filters to correct the IR leakage of the filters.

The Physikalisch-Technische Bundesanstalt (PTB) has developed a spectral comparator facility for accurate spectral radiance measurements. As a primary standard source, a high-temperature black body (HTBB) is operated in the temperature range 1500 K to 3200 K. The black-body temperature is determined by two methods: monochromatic radiation thermometry relative to the freezing point of gold; and absolute radiometry with filter radiometers calibrated against a cryogenic radiometer. For these purposes, the HTBB is operated in radiance and irradiance modes at the same time. This paper presents an uncertainty analysis of the new spectral radiance scale embodied by gas-filled tungsten strip lamps from 220 nm to 2500 nm. Such lamps are currently used as transfer standards for the Consultative Committee for Photometry and Radiometry (CCPR) comparison of spectral radiance (CCPR-S1).

In 1991, the last international comparison of spectral responsivity with germanium detectors showed relative differences between national laboratories of up to 10⁻² at wavelengths of 1300 nm and 1550 nm. The 1997 EUROMET international comparison of optical-fibre power meters showed relative differences between European laboratories of 10⁻². The coincidence of the results from the two comparisons suggests that the radiometric standards used by these laboratories in the calibration of their power meters are the main cause of such differences, rather than the optical-fibre or transfer devices used in the comparisons. To test this assumption, three different electrical-substitution radiometers (ESRs), working at room temperature and used at the same wavelengths as radiometric standards in optical-fibre laboratories, were compared. The radiometers were an optical calorimeter, a pyroelectric radiometer and a thermopile. The relative differences obtained were within 5 × 10⁻³.

The completion of the upgrade of the synchrotron facilities at the National Institute of Standards and Technology (NIST) has yielded a better characterized broadband source of ultraviolet (UV) radiation at the Synchrotron Ultraviolet Radiation Facility (SURF III). A cryogenic-radiometer-based facility that uses the monochromatized radiation from SURF III has been established on beamline 4 (BL-4) to characterize detectors and optical materials in the wavelength range 125 nm to 325 nm. This upgraded cryogenic radiometry facility will be used to measure the spectral responsivity of detectors in the UV with a relative standard uncertainty of less than 10⁻². To demonstrate the capability of BL-4, we have performed a transmittance measurement of calcium fluoride (CaF₂). Accurate transmittance measurements of optical materials in the UV are urgently needed to evaluate which materials are the best choices for use in making lenses for UV lithography. A study of CaF₂ in spectral transmittance, surface scattering, and surface absorption is presented.

A facility has been designed to perform a variety of detector characterizations in the air-ultraviolet in the spectral range 230 nm to 400 nm. The facility allows the measurement of spatial uniformity, linearity and the spectral responsivity of different detectors, irradiated by a He-Cd laser or by a monochromator and Xe discharge lamp. With this technique, a set of Si photodiodes, Si detectors in trap configuration, and ultraviolet (UV) radiometers built with SiC detectors in an integrating sphere, have been calibrated and their absolute responsivity at the He-Cd laser UV line has been determined with respect to an absolute cryogenic radiometer.

The effects on the calibration uncertainties of industrial ultraviolet radiometers of spatial non-uniformity in the irradiance beam and detector response are studied both theoretically and experimentally. The results show that an error of a few parts in 10² or more could easily arise from the non-uniformity of the beam used in calibration if the effective receiving area of the test meter head differs significantly from that of a standard meter head. This error can be removed effectively by an area-correction technique based on a charge-coupled-device (CCD) camera.

The Physikalisch-Technische Bundesanstalt (PTB) operates a unique beam line for the radiometric calibration of ultraviolet (UV) and vacuum-ultraviolet (VUV) radiation sources by direct comparison of their unknown spectral radiant flux with the calculable spectral radiant flux of the electron storage ring BESSY I in the spectral range 40 nm to 400 nm. At this beam line the relative standard uncertainty of the calculated spectral radiant flux of the synchrotron radiation is u = 3 × 10⁻³ and it allows the calibration of the spectral radiance of deuterium lamps with a relative standard uncertainty of 2.5 × 10⁻² above 165 nm. The high-accuracy beam line is now complemented by new instrumentation specially designed for the dissemination of the spectral radiance and radiant intensity scale using deuterium lamps as transfer source standards. This instrumentation allows the calibration of radiation sources in the 115 nm to 400 nm spectral range without significant loss in accuracy and the measurement of spectrally and spatially resolved radiance profiles such as to image the plasma topography of the discharge. The new instrumentation is described. Results for the radiometric long-term stability of selected deuterium lamps are shown. The radiance and radiant intensity as well as the radiance profile of different types of commercially available deuterium lamp are presented.

The spectral responsivity of Ge photodiodes and thin-film thermopiles has been measured between 1 µm and 20 µm. These detectors are used as transfer standards for the spectral responsivity measurements made at the Nederlands Meetinstituut-Van Swinden Laboratorium (NMi-VSL) monochromator-based cryogenic radiometer facility. For measurements beyond 2.5 µm we used an Ar maxi-arc as infrared source in combination with a single-grating monochromator. Below 2.5 µm a quartz-tungsten-halogen lamp is used as light source in conjunction with a double-grating monochromator. An optical system based on three mirrors is used to match the f-numbers of the monochromator and cryogenic radiometer with negligible residual astigmatism. In this paper we discuss the design of the facility, the characteristics of the transfer standards used and the estimated uncertainty budget of the infrared measurements.

The key quantity in the calibration of solar cells and photovoltaic modules is the short-circuit current of the device generated by a reference solar radiation with 1 kW m⁻² total irradiance and with reference solar spectral irradiance distribution. Thus, (i) double spectral mismatch between test and standard device and between test and reference solar spectrum; and (ii) the calibration transfer from low to high irradiance level (non-linearity effects) may affect the calibration value. An overview of primary and secondary calibration procedures is given, focusing on the primary calibration methods applied by the four qualified laboratories establishing the World Photovoltaic Scale (WPVS) traceable to the International System of Units (SI). The differential spectral responsivity method of the Physikalisch-Technische Bundesanstalt (PTB) is described in more detail. Finally, different approaches are illustrated for the evaluation of a reference value, including the WPVS, based on the results for twenty transfer/travelling standards of the international comparison among eleven participants that was carried out to establish the WPVS.

A laser-based facility has been developed to provide high-flux, monochromatic, Lambertian radiation over the spectral range 0.2 µm to 18 µm. The facility was designed to reduce the uncertainties in a variety of radiometric applications, including irradiance and radiance responsivity calibrations. The operational characteristics of the facility are discussed and the results of detector responsivity calibrations over the spectral range 0.406 µm to 0.920 µm are presented.

For over sixty years, the Bureau International des Poids et Mesures (BIPM) has maintained the results of international comparisons of luminous intensity and luminous flux as "world-mean" values, in the form of groups of lamps. It was recently decided that this is no longer a satisfactory and reliable means of maintaining the photometric units of the International System of Units (SI). To maintain the luminous-flux unit, an attempt was made at the BIPM to use the Absolute Integrating-Sphere Method developed at the National Institute of Standards and Technology (NIST). A new technique (ac/dc) employing a chopper for the external source, which allows simultaneous measurement of the internal and external sources, has been developed to overcome an unexpected problem with the characteristics of the BIPM sphere, in that the heat from the lamp affected the sphere coating. The ac/dc technique allows absolute calibration of the integrating sphere while the internal lamp is operating and being measured. Having implemented this technique and replaced the sphere coating, experiments on the derivation of the unit are in progress at the BIPM and the NIST, the preliminary results of which are reported.

The achievements of modern optoelectronics have resulted in the widespread use of light-emitting diodes (LEDs) in products designed to visualize information, especially in light devices for use in road-safety systems. These developments highlight the importance of metrological support for measurements of the photometric characteristics of semiconductor light sources. A standard light-source model based on Corvette Light Corporation high-luminance LEDs has been created and studied. The high-luminance LEDs were manufactured to a unique design using planar technology. The results of the study demonstrate the stability of the metrological characteristics of the LED-based standard source.

Theoretical results for the total flux transmitted by a circular aperture are reviewed for four theories of propagation: rigorous theory, Kirchhoff diffraction theory, Fresnel diffraction theory, and geometrical optics. Curves are given which indicate, for a desired accuracy, the wavelength and aperture size at which the theories diverge. The implications for flux calculations and area measurements are discussed. Measurements of the transmission of a medium-sized aperture support the validity of the Kirchhoff and rigorous theories over the Fresnel and geometrical theories at optical wavelengths, but are unable to determine decisively the most accurate theory.

In the absolute integrating-sphere method, the total luminous flux of a lamp inside an integrating sphere is determined by comparing it with a known flux introduced into the sphere from an external light source. As the measurement geometry of the lamps to be compared is different, the spatial non-uniformity of the sphere surface may affect the results. In order to evaluate this effect, the spatial response must be measured. Miniature incandescent lamps have been used as scanning-beam sources in previous realizations, but these lamps are not widely available. In the present realization of the luminous-flux unit by the Helsinki University of Technology (HUT), light-emitting diodes (LEDs) were used as the light source in scanning the spatial response. Preliminary results confirm the applicability of the LED scanner and indicate moderate deviations of about 1 % from earlier luminous-flux calibrations.

A new Medium-Background Facility (MBF) for the calibration of infrared sources and sensors has been designed by specialists from the All-Russian Research Institute for Optophysical Measurements (VNIIOFI) under contract to the Xi'an Institute of Applied Optics (China). The facility consists of several detachable modules, the configuration of which can be changed to suit the requirements of a specific task. The modular design allows flexible and independent use of the main components and simplifies upgrading. The facility was designed to allow absolute radiometric calibration of black-body sources and IR sensors; to compare the radiometric scale with predictions for measured temperatures; to allow measurements over an extended temperature range (up to the indium point); and to serve as the IR spectral radiance standard. Intended for the calibration of IR sensors and sources under vacuum conditions (10⁻³ Pa) and a medium-background environment (liquid-nitrogen-cooled shroud), the MBF is briefly described with reference to its design, operating principles and measurement results.

A high-accuracy lens-transmission measurement facility has been constructed and tested. The facility uses a tuneable laser source to allow measurements of lens transmission at discrete monochromatic wavelengths in the range 200 nm to 10 µm. Lenses of focal lengths from 300 mm to 750 mm can be measured to between 1 part in 10⁴ and 1 part in 10³, depending on wavelength.

A new concentrator-collimator lighting system based on light-emitting diode (LED) technology is presented. Using an array of LEDs as source, by means of a three-dimensional concentrator and a parabolic collimator, this device makes it possible to obtain an integrated light beam with a high intensity level and the desired divergence, in a compact geometry with small output area. It offers low consumption and a long lifetime. An optical model allowing control of the beam parameters is presented together with a design for thermal stabilization of the device. These characteristics make it applicable in a variety of fields where high luminous intensity, quasi-monochromatic light and narrow angular distributions are needed, among them road signalization. To improve the colour rendering it would be necessary to use a white LED or three different LEDs in an array.

In this paper we report a theoretical analysis of the distortion effects on the probability distribution of the photon number for a classical state of light, induced by non-ideal detectors with non-unitary quantum efficiency and dead time. We present a theory for the reconstruction of the true photon arrival-time distribution, evaluating the probability of detecting the n-th photon at a fixed time.

A new approach is proposed to the calibration of photodetectors and the determination of the absolute value of the quantum efficiency of photon-counting photomultipliers using entangled-photon pairs. It is based on a new single-photodetector technique and a special modulation of the entangled-photon flux. The absolute values of the quantum efficiency have been determined for several photon-counting photomultipliers.

The determination of the geometrical and effective area for optical-quality apertures is one of the fundamental sources of uncertainty in many radiometric and photometric measurements. The National Institute of Standards and Technology (NIST) has developed two non-contact instruments for measuring the area of these optical apertures. Some details of the instruments and their capabilities are presented. Both instruments can be used to measure the area of apertures with diameters ranging from 3.5 mm to 25 mm. The measurements using the absolute instrument result in k = 2 relative uncertainties ranging from 3 × 10⁻⁵ for 25 mm diameter apertures 5 × 10⁻⁵ for 3.5 mm diameter apertures. The uncertainty depends to some extent on the quality of the edge and the diameter of the aperture. The measurements using the relative instrument result in k = 2 relative uncertainties ranging from 2 × 10⁻⁴ for 25 mm diameter apertures to 3 × 10⁻⁴ for 3.5 mm diameter apertures, assuming adequate optical-edge quality. For apertures with an edge of very poor quality, the difference between geometrical and effective area can be as high as 0.3 %. The increased uncertainty largely arises from scattering from the edge. Both instruments play a major role in the Consultative Committee for Photometry and Radiometry (CCPR) key comparison on aperture area, which began in 1999.

An experimental set-up based on correlated photons has been developed at the Istituto Elettrotecnico Nazionale G. Ferraris (IEN) for the absolute measurement of the quantum efficiency and dead time of single-photon counting photodiodes. The overall accuracy is improved with respect to previously reported results, by using a thin and highly transmitting non-linear crystal. For comparison, a conventional set-up has been developed at 632.8 nm with a He-Ne laser, a trap detector and calibrated neutral-density filters. Measurements at different count levels have been performed with both techniques and results are compared for the quantum efficiency in the limit of zero count, η₀, and for the dead-time measurement, t_D. Contributions to the final uncertainty are analysed for both techniques. The results obtained are η₀ = 0.3372 (standard uncertainty u(η₀) = 0.0016) and t_D = 33.8 ns (u(t_D) = 2.0 ns) in the case of the two-photon technique and η'₀ = 0.3376 (u(η'₀) = 0.0036) and t'_D = 31.7 ns (u(t'_D) = 5.4 ns) in the case of the conventional technique.

The aim of this paper is a critical investigation of some correction factors in the quantum-efficiency calibration of photodetectors, in the photon-counting regime, by means of correlated photons generated through a parametric down-conversion process inside a non-linear crystal. An analysis is given of the main correction factors found in a real experimental situation, with different techniques for performing coincidence counts.

A set-up for the accurate measurement of aperture areas based on the superposition of Gaussian laser beams to generate a uniform irradiance has been developed at the Bureau International des Poids et Mesures (BIPM). The uncertainty budget for this method is presented. Results obtained for several apertures are compared with calibrations made at the National Physical Laboratory (NPL). Measurements made on different types of aperture are also compared with the results of an additional technique for determining area ratios. In this case the agreement depends systematically on the type of aperture, because of the imperfect structure of the aperture edges. This work highlights the importance of the determination of the "effective area" for the intended application of the aperture and the use of high-quality diamond-turned apertures.

Beam-limiting apertures with accurately known areas are required when measuring the radiance and irradiance of sources. We have already described a non-contact method for measuring aperture areas using a laser beam focused on the aperture surface. By shifting the aperture relative to the laser beam, and collecting and monitoring the reflected light, the onset of reflection was used to indicate the position of the aperture edges. We have now improved this approach by evaluating the position-dependence of the reflection signal recorded across the aperture edge. Whereas the former approach required a determination of the laser focus diameter to correct the measured aperture diameter, the quantity that makes the main contribution to the overall uncertainty of the aperture area, the new approach implicitly determines the laser focus diameter, so reducing the corresponding uncertainty and making it possible to determine aperture diameters with standard uncertainties of less than 0.2 µm. Measurements made in 1998 for aluminium apertures are compared with results obtained in 1995 using the same apertures.