Terahertz (THz) imaging has shown promise for nondestructive evaluation (NDE) of a wide variety of manufactured products including integrated circuits and pharmaceutical tablets. Its ability to penetrate many non-polar dielectrics allows tomographic imaging of an object's 3D structure. In NDE applications, the material properties of the target(s) and background media are often well-known a priori and the objective is to identify the presence and/or 3D location of structures or defects within. The authors' earlier work demonstrated the ability to produce accurate 3D images of conductive targets embedded within a high-density polyethylene (HDPE) background. That work assumed a priori knowledge of the refractive index of the HDPE as well as the physical location of the planar air-HDPE boundary. However, many objects of interest exhibit non-planar interfaces, such as varying degrees of curvature over the extent of the surface. Such irregular boundaries introduce refraction effects and other artifacts that distort 3D tomographic images. In this work, two reconstruction techniques are applied to THz synthetic aperture tomography; a holographic reconstruction method that accurately detects the 3D location of an object's irregular boundaries, and a split-step Fourier algorithm that corrects the artifacts introduced by the surface irregularities. The methods are demonstrated with measurements from a THz time-domain imaging system.
Terahertz (THz) technology holds great promise for applications such as explosives detection and nondestructive evaluation. In recent years, three-dimensional (3-D) THz imaging has been considered as a potential method to detect concealed explosives due to the transparent properties of packaging materials in the THz range. Another important advantage of THz systems is they measure the electric field directly. They are also phase coherent, supporting synthetic aperture (SA) imaging. In this paper, a near-field synthetic aperture THz imaging system is investigated for its potential use in detecting hidden objects. Frequency averaging techniques are used to reduce noise side-lobe artifacts, and improve depth resolution. System depth resolution is tested and characterized for performance. It will be shown that, depending on system bandwidth, depth resolution on the order of a few hundred microns can be achieved. A sample consisting of high-density polyethylene and three ball-bearings embedded inside is imaged at multiple depths. 3-D images of familiar objects are generated to demonstrate this capability.
The spectra obtained from Terahertz (THz) reflection imaging can be distorted by scattering from rough interfaces,
layers, and granular inclusions. Since the facets of the object being imaged are not generally aligned
normal to the THz beam, the received signal is produced from diffuse scattering, which can be appreciably lower
in signal strength than specular returns. These challenges can be addressed with advanced signal processing approaches
based upon the coherent and incoherent combination of returns from multiple sensors and frequencies.
This paper presents two examples of physics-based processing strategies applied to THz imaging spectroscopy.
The first method is based on synthetic aperture processing of a 2D sensor array to provide variable depth focused
images of buried inclusions (a ball bearing embedded in polyethylene sample). The second method uses correlation
processing to coherently combine multiple sensors and multiple frequencies to extract material signatures
from measurements of THz scattering from rough interfaces. Results for both methods show an increase in
performance relative to conventional imaging or spectroscopy approaches.
Many materials such as drugs and explosives have characteristic spectral signatures in the terahertz (THz) band.
These unique signatures hold great promise for potential detection utilizing THz radiation. While such spectral
features are most easily observed in transmission,real life imaging systems will need to identify materials of
interest from reflection measurements,often in non-ideal geometries. In this work we investigate the interference
effects introduced by layered materials,whic h are commonly encountered in realistic sensing geometries. A
model for reflection from a layer of material is presented,along with reflection measurements of single layers
of sample material. Reflection measurements were made to compare the response of two materials; α-lactose
monohydrate which has sharp absorption features,and polyethylene which does not. Finally,the model is
inverted numerically to extract material parameters from the measured data as well as simulated reflection
responses from the explosive C4.
Recent improvements in sensing technology have driven new research areas within the terahertz (THz) portion
of the electromagnetic (EM) spectrum. While there are several promising THz applications, several outstanding
technical challenges need to be addressed before robust systems can be deployed. A particularly compelling
application is the potential use of THz reflection spectroscopy for stand-off detection of drugs and explosives. A
primary challenge for this application is to have sufficient signal-to-noise ratio (SNR) to allow spectroscopic identification
of the target material, and surface roughness can have an impact on identification. However, scattering
from a rough surface may be observed at all angles, suggesting diffuse returns can be used in robust imaging of
non-cooperative targets. Furthermore, the scattering physics can also distort the reflection spectra, complicating
classification algorithms. In this work, rough surface scattering effects were first isolated by measuring diffuse
scattering for gold-coated sandpaper of varying roughness. Secondly, we measured scattering returns from a
rough sample with a spectral signature, namely α-lactose monohydrate mixed with Teflon and pressed with
sandpaper to introduce controlled roughness. For both the specular and diffuse reflection measurements, the
application of traditional spectroscopy techniques provided the ability to resolve the 0.54 THz absorption peak.
These results are compared with results from a smooth surface. Implications of the results on the ability to detect
explosives with THz reflection spectroscopy are presented and discussed. In addition, the Small Perturbation
Method (SPM) is employed to predict backscatter from lactose with a small amount of roughness.
The advent of terahertz (THz) spectroscopy and imaging has motivated
the investigation of waveguides structures that are appropriate for
use at THz frequencies. Currently, most spectroscopy systems have no
guiding mechanism and therefore suffer spherical spreading loss. The
ideal waveguide would eliminate these spreading losses, have low
attenuation and dispersion, and high field confinement. Effective
terahertz spectroscopy also requires especially large bandwidths to
identify spectral signatures; therefore single-mode propagation is
necessary to transmit broadband pulses. Single-mode waveguides can be
achieved by utilizing the fundamental HEM11 mode, which has no
cutoff frequency, or by suppressing higher order modes. In this work,
we present cylindrical, hollow-core, dielectric waveguides for the
G-band (140-220 GHz). We will show propagation characteristics
obtained by theoretical analysis, computer simulation, and experiments
conducted using a vector network analyzer (VNA) with a cylindrical
horn attachment. The analytical model, in particular, will provide
an additional capability to quickly predict waveguide behavior for a
variety of applications. This complete model will predict both the
number of supported and propagating modes and the means to source
them. We will show that the number of supported modes depends
primarily on the thickness and dielectric constant of the outer layer.
By choosing proper dimensions and materials, single-mode waveguides
can potentially be designed and realized to achieve all of the
aforementioned terahertz spectroscopy requirements.