Measuring temperatures in the presence of external radiation by flash assisted multiwavelength pyrometry

A novel multiwavelength analysis of the radiation emitted by a hot Lambertian surface is presented allowing a reliable measurement of its surface temperature even in the presence of interfering external light sources. For the separation of thermally emitted and reflected external radiation, an accurate knowledge of the spectral emittance/reflectance of the sample is required. In our procedure, this quantity is determined in situ by performing a second measurement with additional light, e.g., from a strong electric flash. The strength of the flash and the relative spectrum of the interfering external radiation are probed with a cooled Lambertian reflectance reference near the sample. In a first step, the spectral reflectance of the sample is determined from spectra recorded with and without flash. The temperature and the total incoming irradiance are then obtained from a least-square fit of the input spectra with an analytical model of the emitted thermal plus the superimposed reflected external radiation. The method delivers also estimates for the statistical accuracy of the measured temperature and irradiance values. It causes no problem to measure temperatures through a window of sufficient optical quality and effects of absorbing gases or vapors may be minimized by an appropriate choice of the spectral intervals used for the analysis. In order to indicate statistical and systematic errors of the method under different operating conditions, extensive numerical simulations were performed. They reveal that a good statistical accuracy of the recorded spectra is crucial for a proper separation of the emitted from the reflected external radiation if one part dominates. In contrast, this separation is not affected by an erroneous calibration of the optical detection system. Finally, the question is discussed if one can dispense with the flash. Without external radiation, the "no-flash" method is equivalent to quotient pyrometry and subject to the same principal limitations. With external radiation, reasonably accurate temperatures can be expected only if thermal radiation clearly dominates or if emittance values are available from other sources. (C) 1999 American Institute of Physics. [S0034-6748(99)04306-3].