A complex filtering method for eliminating systematic and random noises of THz-TDS is proposed. This method is the combination of deconvolution and wavelet filtering algorithms. A self-reference method for extracting the complex refractive index of material accurately is also proposed in order to avoid a phase shift due to the misplacement between the surfaces of the reference and sample using the time-domain terahertz reflection spectroscopy. The basic idea of self-reference method is that the first and the second peaks of the reflection spectrum of solid pellets are regarded as the reference and sample signals, respectively. Thus more information of samples can be extracted because of a longer optical path, and meanwhile, phase error can be avoided by obtaining the reference and sample signals through a single measurement. According to the Fresnel formulas, we deduce the expression of complex refractive index and then design an iterative algorithm for solving it. We choose the glucose solid pellets as samples to test the self-reference method. After measuring the time-domain reflection spectrum, we adopted the complex filter method for filtering and utilized the self-reference method to extract the complex refractive index. Based on Density Functional Theory (DFT), the characteristic absorption spectrum of multiple glucose molecules in the THz absorption spectroscopy was obtained by the simulation analysis on the vibration of multiple glucose molecules. The results indicate that the absorption peaks appear in the absorption coefficient curves at the corresponding frequency positions which are approximately consistent with the results of the simulation based on DFT. So the methods we proposed can help improve the retrieval accuracy of complex refractive index.
The scheme of parametric image amplification without using lenses is discussed. The theoretical analysis shows that the scheme is appropriate to a broadband amplifier which assures a high linear gain, a better resolution and noiseless amplification.