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%.

https://iopscience.iop.org/article/10.1088/0026-1394/28/3/018/pdf

Trap Detectors and their Properties

The stability of the quantum efficiency of inversion layer, phosphorus-diffused (n on p), and boron-diffused (p on n) photodiodes has been investigated. Unsatisfactory silicon-silicon dioxide interfaces, latent recombination centers in the diffused layers, and moisture absorption by the device were identified as possible causes of instability. Diodes were fabricated using processes in which these sources of instability were carefully controlled. The resulting diodes were subjected to various accelerated aging tests, and the external quantum efficiency of the diodes was monitored during the tests. Diodes made by older procedures, in which some important parameters affecting stability were not controlled, were included in the study for comparison. The major result of this work is the demonstration that n on p photodiodes are inherently more stable than p on n types in the ultraviolet and blue spectral regions, but that stable p on n devices can also be produced with sufficient care.

Quantum efficiency stability of silicon photodiodes

An intercomparison has been performed in the visible and near infrared between two radiometric standards: the electron storage ring BESSY and a cryogenic absolute radiometer. The spectral irradiance of the continuum source BESSY can currently be predicted with a relative systematic uncertainty of 3 × 10−3, whereas the radiant power of monochromatic laser radiation can be measured by the cryogenic radiometer with a relative systematic uncertainty of 10−4. In this intercomparison the responsivity S of a specially designed transfer radiometer was measured using synchrotron radiation (SSR) and the cryogenic radiometer (SRAD). The mean of the ratios SRAD/SSR for different measurement conditions was 1.0013 with a standard error of the mean of 0.0006 for a wavelength of 676 nm and 0.9990 with a standard error of 0.0007 for a wavelength of 799 nm. The systematic uncertainty of the ratio is 3.8 × 10−3 caused by the uncertainty of SRAD of 1.5 × 10−3 and by the uncertainty of SSR of 3.5 × 10−3. Both radiometric standards agree well within the combined systematic uncertainties of the two methods. The result confirms the 3 × 10−3 uncertainty of BESSY. This source standard has therefore a smaller uncertainty than present source standards based on high temperature blackbody radiators. The work also demonstrates how detector- and source-based radiometric standards can be intercompared or transferred between laboratories with a relative uncertainty of <3 × 10−3.

Intercomparison between two independent primary radiometric standards in the visible and near infrared: a cryogenic radiometer and the electron storage ring BESSY.

The errors associated with modelling the quantum efficiency of high quality silicon photodiodes in the ultraviolet using the semiconductor device-modelling program PC1D are discussed. Uncertainty spectra have been calculated for these errors, which are mainly a result of the application of basic solid-state models and parameters in the extreme case of ultraviolet absorption. The more critical parameters have been identified and, where possible, corrections have been made to account for them. The overall uncertainty associated with quantum efficiency modelling is found to be well below 0,2% for detection of radiation from 400 nm to 300 nm, and below 1% for radiation down to 250 nm.

https://iopscience.iop.org/article/10.1088/0026-1394/30/4/023/pdf

A Physical Basis for the Extrapolation of Silicon Photodiode Quantum Efficiency into the Ultraviolet

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.

Numerical Modelling of Short-wavelength Internal Quantum Efficiency

Oxide-bias measurement techniques to determine front-surface recombination losses in the silicon photodiode self-calibration technique have been investigated. With the water-drop method it has been shown necessary to restrict the applied bias voltage to the very minimum necessary to produce photocurrent saturation, otherwise non-reversible degradation of quantum efficiency may occur. The corona technique has been found to give satisfactory results only on initial application, as a drastic increase in front-region recombination loss is produced.

https://iopscience.iop.org/article/10.1088/0026-1394/21/2/005/pdf

M L Rastello

Papers

Oxide-bias Measurements in the Silicon Photodiode Self-calibration Technique