Qualitative And Quantitative Analysis Of Matter Based On Terahertz Time Domain Spectrum

THz wave is located in the gap between the infrared waves and microwaves; the development on ultrafast laser in recent ten years facilitates the generation of THz wave, which promotes the development of THz technology. And the THz spectroscopy technology has been an advanced research hot in global academic field. The unique advantages like low-energy, penetrating and fingerprint of Terahertz make it has good application prospect, its low power is prospective to be applied in undamaged detection, its penetrability can be used to security checks, its fingerprint spectrum can be used to the identification and classification of substance. In daily life, we are faced with many problems. For example, we need to detect harmful additives, waste oil and counterfeit products in food safety; we need to detect drugs, explosives, firearms and knives in the anti-terrorism security issues; we need to monitor air pollution index in real time in the increasingly highlight environment issues. Excellent properties of terahertz waves can provide new ideas for solving the above problem, terahertz time-domain spectroscopy is used to material research, it can get the optical parameters of material in the terahertz band, the material components can be analyzed through the analysis of optical parameters. Based on THz time-domain spectroscopy, terahertz spectrum of donkey-hide gelatin, antler glue, four kinds of aniline, milk powder and melamine were acquired, qualitative and quantitative analysis of mixture were successful.Contents of this dissertation and innovations of my research are presented as follows.Expounding the background of studying in this paper, showing the purpose of the studied. Mechanism of terahertz wave generation and detection, commonly used transmission time domain spectroscopy system in experiment were introduced in detail. And the ways of optical parameter extraction was provided, detailed introduces the preparation of samples, test process and matters needing attention.In the qualitative analysis of the material, the method for data preprocessing before recognition such as noise filtering and feature extraction was introduced firstly, and then gived the introduction of atificial neural network and support vector machine(SVM) used in pattern recognition. After acquiring the THz spectrum of donkey-hide gelatin, antler glue, four kinds of aniline, milk powder and melamine, pattern recognition method was used to analyze it. The experimental results show that, material could be identificated successfully.In the quantitative analysis of the material, firstly studied the stoichiometric method-PLS, then studied the terahertz absorption coefficient spectrum of melamine and milk powder mixture with different blending ratio. Based on the Berr-Lambert’s law, melamine ratio regression model was established combined with PLS regression The results showed that the terahertz time-domain spectroscopy could quantitative analysis of mixture successfully.