The Application of Signal Processing Techniques in Improving Spectrometer Detection Limit

Abstract

Improving spectrometer detection limit using statistical principle and signal-processing technique are described simply in this work. In the detection of gas photoacoustic(PA) signal, accurate partitions of the sampled data affect on the detection limit to some extent although Lock-in amplification technique with high SNR and microphone sensor with high sensitivity have been used. A model and the resulting algorithm are proposed from PA-signal samples. The techniques are validated at ppb level on PA spectrometer for NH3 breath detection in high concentration of CO2 and H2O based on tunable erbium-doped fiber laser (TEDFL) or for multicomponent trace gas detection based on waveguide CO2 laser or on other laser source.