The following is an abstract for the selected article. A PDF download of the full text of this article is available here. Members may download full texts at no charge. Non-members may be charged a small fee for certain articles.

Signal-to-Noise Considerations in Flame/Furnace Infrared Emission Spectroscopy

Volume 45, Number 4 (May 1991) Page 546-554

Busch, Marianna A.; Busch, Kenneth W.

A theoretical model for the signal-to-noise ratio (SNR) performance of a flame/furnace infrared emission (FIRE) radiometer using a PbSe detector has been developed. The model uses readily available, published PbSe detector parameters to predict the SNR for situations involving only detector noise as well as for situations involving detector noise in the presence of residual, uncorrelated flame noise remaining after background subtraction. The theoretical SNR values were found to be in good agreement (6%) with previously published, experimentally determined SNR values for Freon-113®, obtained with a dual-channel FIRE radiometer operating in the carbon mode. The model was used to show the effect of source modulation frequency on the SNR obtained with the FIRE radiometer. At room temperature under detector-noise-limited conditions, it was found that: 1/f noise > background photon noise > generation-recombination noise ≫ Johnson noise. The PbSe detector was found to exhibit 1/f-noise behavior out to about 1000 Hz when operated at room temperature. The model predicts that, even for a dual-channel FIRE radiometer in the presence of residual flame background noise, some SNR improvement should be possible by increasing the chopping frequency out to around 3 kHz.