The signal-to-noise ratio estimation in dispersive absorption spectrometry and new quantitative methods based on the signal-to-noise ratio theory.

摘要

This research has been divided into two parts: "the signal-to-noise ratio estimation in dispersive absorption spectrometry" and "new quantitative methods based on the signal-to-noise ratio theory."; In the first part of this research, I have established a general model, so that the relationship of the signal-to noise ratio (SNR), the spectral bandpass (SBP), and the absorbance at a specific wavelength can be quantitatively predicted. Additionally, the noise contribution, or so-called noise budget, can be observed as well.; In the second part of this research, I have proposed three new quantitative methods and compared them with methods currently used. These three methods greatly extend dynamic ranges, especially at high absorbances. These new methods are: (1) Wide SBP system and the Fitting method: this algorithm improves the signal quality of the sample spectrum, and also includes a convolution function that can eliminate the strong broadening effect. Cooperating with the Fitting method causes a quantitative measurement no longer to be limited by utilizing high-resolution spectrometric systems. (2) Enhanced weighted regression method (EWRM): this method is based on weighted regression method, but includes an additional wavelength selection that can reduce the non-linearity caused by stray light. (3) Stitching method: this method improves the signal quality of the reference spectrum by mean of collecting spectrometric information from more than one standard. The stitching method behaves as a preliminary procedure for most multiple wavelength quantitative methods.