Radiometry is the measurement of optical radiation including ultraviolet, visible and infrared. The candela has been one of the base units since the inception of the SI, and is historically related to human vision. Since 1979 the definition of the candela is expressed in strictly physical terms, i.e. it specifies the radiometric content of a monochromatic source at 540·10¹² Hz, independently of any actinic spectrum.
The current definition of the candela is based on the wave-like behaviour of light. A detector is considered absolute if its responsivity is calculable or predictable by fundamental physical laws. Absolute detectors are used to realise the candela, and are mainly electrical substitution radiometers (ESR) or semiconductor photon detectors. ESRs are based on the electrical substitution principle, that is, the heating effect of the optical radiation to be measured is compared with the heating effect produced by a measured quantity of electrical power (Joule effect). The operation of these detectors at temperatures below 20 K enables the current state of the art measurement uncertainty to be around 50 ppm.
Semiconductor photodetectors depend on the photoelectric effect, where the absorption of a photon generates a free electron-hole pair. The quantum efficiency is defined as the average number of free electron-hole pairs produced per incident photon. In a high-quality silicon photodiode the quantum efficiency for the visible range is close to unity within some parts in 10³, and the deviation of the quantum efficiency from unity can be determined independently of other radiometric measurements. Uncertainties of a few parts in 10⁴ appear to be the limit of this technique with commercial photodiodes.
Both ESRs and silicon detectors show the highest accuracy when measurements of radiant flux are performed at the power level of hundreds of microwatts (10¹³ photons/s). They are analogue devices where single photons are not distinguished and the detector response (output) is an average continuous quantity (current or voltage). When the signal level approaches picowatt and femtowatt (10⁴-10⁷ photons/s), counting techniques are employed and single/few photons are detected by photon counters. A single-photon counter is a photodetector whose response to a detected photon is a pulse. Most types of single-photon detectors cannot discriminate how many photons arrive within a small time interval. Measurement in the few/single photon regime requires a new branch of radiometry called quantum radiometry.
Quantum radiometry concerns the absolute measurement of photon quantities based on fundamental physical phenomena. Each quantum of radiation, a photon, is characterized by a specific frequency v and energy E = hv, where h is Planck's constant. For single-photon detectors, the best calibration techniques are based on spontaneous parametric down-conversion, where photons are emitted in pairs strongly correlated in direction, wavelength and polarization. Photons of the same pair are emitted within tens of femtoseconds. The observation of one photon from a pair in a certain direction (signal) implies the presence of the other paired photon in the conjugated direction (idler). The non-detection of this idler photon is due to the non-ideal quantum efficiency of the idler detector, which can be measured in this way. This absolute technique is attractive for the realisation of absolute radiometric scales, because it relies simply on the counting of events, requires a remarkably small number of measured quantities, and does not require any reference standard [1].
The reformulation of the SI base unit; the candela; in terms of photon number rather than in optical power is one of the Grand challenges on fundamental metrology of EMRP2007 [2]. In this respect, that reformulation through linkage to Planck's constant h will provide greater consistency among the definitions of the base units and better serve the additional needs of emerging sectors such as the quantum based technologies.
The quCandela project (T1 J2.3 Candela: Towards quantum-based photon standards) is an international project funded by the European commission (FP7) under the iMERA program (implementing metrology in the European Research Area). This project involves the main metrology institutes in Europe with INRIM as project coordinator. It will support both "classical radiometry" as well as emerging quantum technologies by investigating photon sources and detectors from the signal level of existing radiometric standards down to single photons. This ambitious goal requires a step change in optical metrology in order to bridge the energy difference between the quantum and the classical world. In the long-term this could result in a reformulation of the candela, if such a reformulation could be demonstrably capable of being realised at, and better than, the current accuracy of around 50 ppm (parts per million).
As regards "classical radiometry" predictable quantum efficient photon detectors will be designed and constructed. The basic idea is that this photodiode is a primary standard because its output is linked to fundamental constants, due to the one-to-one conversion of photons into electron/hole pairs. Deviation from ideal performance is due to reflectance and internal losses, each of them individually measurable and calculable [3]. These loss components are expected to be small, so measuring their relative magnitude has less demanding uncertainty requirements.
On the other side, as regards quantum radiometry, the development of photon number resolving detectors is a crucial task of the project (in charge to INRIM), as conventional single-photon counting detectors cannot distinguish between one or more photons arriving at the same time. An important breakthrough is the development of superconducting devices operated at the transition temperature (transition-edge sensors, TES) around 100 mK (fig. 1).

The absorption of a single photon moves the device through the steep transition between the superconducting and the normal state, with a measurable current change induced by the increase in sensor resistance. The key feature of these devices is their intrinsic capability to measure the energy of the absorbed photons, hence their ability to resolve the number of photons absorbed if the energy is already known. Fig. 2 and 3 report the first measurement results with this detector.

[1] ] G. Brida, S. Castelletto, I.P. Degiovanni, M. Genovese, C. Novero, M.L. Rastello: "Toward an accuracy budget in quantum efficiency measurement with parametric fluorescence", Metrologia, Vol. 37, pp. 629-632, 2000.
[2] EMRP 2007, III.2.1, bullet 5, p. 26/47.
[3] J. Geist, G. Brida, M.L. Rastello: "Prospects for improving the accuracy of silicon photodiode self-calibration with custom cryogenic photodiodes", Metrologia, Vol. 40, pp. S204-S207, 2003.
[4] M. Rajteri, E. Taralli, C. Portesi, E. Monticone, J. Beyer: "Photon-number discriminating superconducting transition-edge sensors", Metrologia, in press.