Uncertainties in the determination of thermal emissivity by measurement of reflectance using Fourier transform spectrometers

The emissivity of coated glazing is commonly determined from spectral reflectance measurements using a specular reflectance accessory or other suitable reflectance attachment. As part of the ongoing EU Thermes RTD project, a 19 laboratory intercomparison was undertaken in 2003 to establish the state-of-the-art of infrared reflectance measurement accuracy and uncertainty and to identify any needs for improvement. Six test samples were selected to provide samples with low, medium and high infrared reflectance. In addition, bare gold mirrors were prepared for distribution to each participant to serve as the reference material. Both the reference mirrors and the round robin samples were calibrated and evaluated prior to delivery to enable closely characterised and similar sample sets to be prepared. A measurement procedure requiring a sequence of 18 measurements undertaken in a strict order was developed. The outcome of the exercise is very encouraging. Participants have performed the work using a wide range of experimental conditions but the final results are in excellent agreement. This is because of the adherence to the required measurement procedure and the use of reference mirrors with known reflectance values. The analysis indicates uncertainties in the final results which are less than ± 0.01. Factors such as instrument resolution, number of scans and time of acquisition do not appear to have been significant in affecting measurement accuracy.