Cavity ring fiber chemical sensors

Abstract

In recent years, optical fiber chemical sensors are paid increasing attentions due to their electrical passivity, remote sensing potentials and miniaturization feasibility, relying on the specific properties of the optical fibers. Highly sensitive spectroscopic techniques are highly demanded for enhancing the sensitivities of the optical fiber chemical sensors. This thesis mainly reports optical fiber cavity ring-down (CRD) and cavity ring-up (CRU) based-sensors for chemical sensing and other applications. The thesis begins with a comprehensive survey on the existing types and spectroscopic techniques for optical fiber chemical sensors. Following that, direct absorption spectroscopy, which is the most commonly used spectroscopic technique, is studied. The direct absorption spectroscopy-based optical fiber chemical sensors are analyzed• based on a practical setup. Aiming at reducing the effects of the main noises in this kind of sensor, a digital finite impulse response (FIR) filter is designed to improve this chemical sensor's sensitivity. The simulation and experimental results have shown that the detection accuracy is enhanced by this digital signal processing (DSP) technique. The success of introducing DSP technique into the sensing system motivates us to solve the challenge of identification of the unknown chemical mixture detected by an optical fiber chemical sensor using pattern recognition technique. An automatic chemical mixture recognition system by using support vector regression (SVR) technique is proposed. Using the recognition model, the chemical components and the corresponding concentrations could be identified automatically and a high measurement accuracy in terms of concentration is achieved.