Detection and Identification of a Substance with an Inhomogeneous Surface Using the Effective Time-Dependent THz Spectroscopy Method and Emission Frequency Up-Conversion

Vyacheslav A. Trofimov, Svetlana A. Varentsova, Yongqiang Yang, A.A. Silaev
2018 EPJ Web of Conferences  
Currently, the THz radiation is actively applied for solving security and anti-terrorism problems [1], [2] . In most THz TDS setups, the substance identification occurs based on comparison of the absorption frequencies of a substance under investigation with a set of known absorption frequencies from a database (we call it as the standard THz TDS method). However, this technology has certain limitations. It is well-known that surface roughness can lead to scattering and modulating of the
more » ... ating of the spectrum, which complicates the identification [3], [4] . To overcome this problem, in [3] it was shown that by summing and averaging multiple measurements over a sample area the scattering effect can be effectively decreased. In [4] the authors proposed for the same purpose to increase the number of viewing angles. However, they can reliably identify only one absorption frequency with maximal spectral amplitude. Unlike the methods mentioned above, in the discussed method we use only one THz signal reflected from the substance with inhomogeneous surface but measured in the long time interval duration 180-200 ps. This duration allows registering not only the main reflected THz pulse, but also several sub-pulses following it, which are due to the reflectance from the inner surfaces of the sample. These sub-pulses also contain information about spectral characteristics of the substance and can be used for the identification. In addition, we do not apply averaging of measurements over viewing angles and the sample area in order to reduce the scattering effects. As an example of identification, we use the THz signal duration 180 ps reflected from the PWM C4 explosive (90 % RDX, 10 % plasticizer) with a rough surface grit 40 (signal PWM_40 for brevity). The signal contains the pronounced main pulse reflected from the outer surface of the sample, the first sub-pulse reflected from the inner surface, and the sub-pulses with significantly less amplitude due to multiple reflections from the inner surfaces. The Fourier spectrum of the PWM_40 main pulse is shown in Fig. 1 (a) together with that of the smooth PWM for comparison. The spectral properties of the PWM_40 main pulse are distorted by the influence of the inhomogeneous surface (a) so that the standard THz TDS method is inefficient for identification. In the PWM_40 spectrum (a) there is a single minimum at the frequency ν=0.88 THz in the frequency range ν=[0, 1.8] THz, which is close to RDX absorption frequency ν=0.82 THz. In the frequency range ν=[1.8, 4.0] THz there are minima at the frequencies ν= 1.92, 2.24, 3.0 THz , which are close or equal to the RDX absorption frequencies ν= 1.95, 2.19, 3.0 THz [2]. 0.0 0.5 1.0 1.5 2.0 2.5 0.00 0.25 0.50 0.75 1.00 1.25 (a) =0.88 PWM_40 |P()| (THz) Main pulse smooth PWM 2.0 2.4 2.8 3.2 3.6 4.0 0.000 0.005 0.010 0.015 =3.88 =1.92 (b) =2.24 |P()| (THz) Main pulse =3.0 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 0.0 0.1 0.2 0.3 0.4 0.5 =3.88 (c) First sub-pulse =3.0 =0.92 =1.92 |P()| (THz) =2.2 Fig. 1. Fourier spectrum of the main pulse of the PWM_40 and smooth PWM signals in the frequency ranges ν=[0.0, 2.5] THz (a), [1.8, 4.2] THz (b), spectrum of the PWM_40 first sub-pulse in the frequency range ν=[0.8, 4.2] THz (c). The spectrum of the PWM_40 first sub-pulse contains minima at the frequencies ν= 0.92, 1.92, 2.2. 3.0 THz (c), that are close to the absorption frequencies of explosive RDX ν=0. 82, 1.95, 2.19, 3.0 THz ([2]). And the spectrum of the remote part of the PWM_40 signal corresponding to the time interval t= [70, 170] ps, contains minima at the frequencies ν = 0.9, 1.96, 2.2, 3.01 THz, which are close to RDX absorption frequencies mentioned above (not shown). Therefore, the sub-pulses also contain information about the substance spectral properties and can be used for identification. In the current report, the identification is based on the method of spectral dynamics analysis (SDAmethod) together with several integral correlation criteria (ICC) [5] . We compare the dynamics of spectral intensity of a substance under analysis with the corresponding dynamics of a standard substance from database at chosen frequencies. To increase the reliability and effectiveness of the identification, we propose to use several ICC's simultaneously in different time intervals, which contain both the main pulse of the reflected THz signal and the following sub-pulses. In Fig. 2 the evolution of the ICC , pP CW1 [5] is calculated for the frequencies ν=0.82 THz (a), (c) and
doi:10.1051/epjconf/201819509005 fatcat:pwciezzanfc6zlyulrk3fw6qni