Detection and identification of substances using noisy THz signal

We discuss an effective method for the detection and identification of substances using a high noisy THz signal. In order to model such a noisy signal, we add to the THz signal transmitted through a pure substance, a noisy THz signal obtained in real conditions at a long distance (more than 3.5 m) from the receiver in air. The insufficiency of the standard THz-TDS method is demonstrated. The method discussed in the paper is based on time-dependent integral correlation criteria calculated using spectral dynamics of medium response. A new type of the integral correlation criterion, which is less dependent on spectral characteristics of the noisy signal under investigation, is used for the substance identification. To demonstrate the possibilities of the integral correlation criteria in real experiment, they are applied for the identification of explosive HMX in the reflection mode. To explain the physical mechanism for the false absorption frequencies appearance in the signal we make a computer simulation using 1D Maxwell's equations and density matrix formalism. We propose also new method for the substance identification by using the THz pulse frequency up-conversion and discuss an application of the cascade mechanism of molecules high energy levels excitation for the substance identification.