Content uploaded by Jean-Paul Guillet
Author content
All content in this area was uploaded by Jean-Paul Guillet on Jan 31, 2023
Content may be subject to copyright.
IEEE TRANSACTIONS ON TERAHERTZ SCIENCE AND TECHNOLOGY, VOL. XX, NO. XX, APRIL 2020 1
Guided Reflectometry Imaging Unit using
Millimeter Wave FMCW Radars
M. Pan, A. Chopard, F. Fauquet, P. Mounaix*, J.-P. Guillet*
Abstract—Frequency Modulated Continuous Wave (FMCW)
radar systems in the millimeter and sub-millimeter range are
technologically mature for many applicative fields such as au-
tomotive and aerospace industries for imaging and non de-
structive testing. This work reports on a new implementation
of a guided FMCW radar reflectometry unit for sensing and
imaging applications. Only a terahertz dielectric waveguide is
used for signal transmission between the transceiver module
and the sample, thus drastically simplifying the experimental
setup. Compared to continuous wave guided systems, one of
the main advantages granted by the use of FMCW radars in
combination with waveguides, is the differentiation capability
between the reflected signals generated along the wave guide
as parasitic signals or at its probing end as sensing information
and therefore improving the expected signal-to-noise ratio. This
innovative approach is demonstrated by using a dielectric hollow-
core waveguide integrated with two different radar transceivers;
the high-performance, III-V based 100 GHz SynView unit as a
reference system and a compact, low-cost, PCB-Integrated, 122
GHz transceiver developed by Silicon-Radar GmbH. Both 3D
electromagnetic simulations and raster scans are performed to
investigate quantitatively the propagation behaviors including the
coupling capabilities, dynamic range limitations, beam profile and
induced artefacts of the guided FMCW reflectometry system. The
feasibility of a simplified guided terahertz FMCW reflectometry
probing unit is proven. The integration of a solid immersion lens
at the end of the waveguide is also demonstrated for imaging
resolution improvement.
Index Terms—Guided waves, FMCW Radar, Imaging, Tera-
hertz
I. INT ROD UC TI ON
THE scaling down of Si-based transistors and the in-
vestigations into III-V technologies further push up the
operating frequency of solid-state devices towards the terahertz
regime [1]. On this basis, high-speed electronic devices with
low power consumption and better compactness have been
developed in the 100 - 300 GHz range. Being cost-effective
systems supported by continuously improved fabrication tech-
niques, solid-state terahertz devices play an increasingly im-
portant role in academic researches and industrial applications
[2]. Compared to conventional continuous wave systems com-
bined with 3D imaging reconstruction approaches, such as
Shape From Focus [3] or Computed Tomography [4], terahertz
and/or millimeter wave Frequency-Modulated Continuous-
Wave (FMCW) radar systems can natively provide additional
Manuscript received April 09, 2020. The first two authors contributed
equally to this work. Asterisk indicates corresponding authors.
M.Pan, A. Chopard, F. Fauquet, P. Mounaix and J.-P. Guillet are with the
IMS Laboratory, UMR CNRS 3218, University of Bordeaux, 351 Cours de
la Lib´
eration 33405 Talence Cedex, FRANCE
A. Chopard is also with Lytid SAS, 8 rue la Fontaine, 92120 Montrouge.
E-mail :patrick.mounaix@u-bordeaux.fr; jean-paul.guillet@u-bordeaux.fr
phase information for further result analysis and simplified
3D reconstructions with suitable resolution, while allowing
further processing for improvement [5]. In particular, exploit-
ing a quasi-optical coupling method (using lenses or parabolic
mirror) with FMCW systems allows in-depth measurements
for non-destructive testing purposes. This technique combines
the high sensitivity of FMCW methods with the penetration
capabilities of millimeter waves. Based on those benefits,
wideband FMCW radars have found suitable application fields
in the automotive and aerospace industries [6], [7] and art-
painting analysis [8] amongst others for their non-destructive
testing capabilities. However, the use of optical components
involves tedious alignment and imposes mechanical restric-
tions along the propagation path. The use of such optical
coupling methods limits the development of compact, portable
and easily-implementable terahertz measurement systems to a
broader scope of applications.
To address this issue, a terahertz waveguide [9] is proposed
as an alternative solution. The guided reflectometry concept
has already been investigated with some continuous wave
sources or pulsed sources [10] in conjunction with different
waveguide conceptions. A variety of geometries have been
assessed, from a rectangular waveguide coupled to a vector
network analyzer at low frequencies for burn damage detec-
tion [11], to higher frequencies tests with low-cost Teflon
waveguides for remote chemical detection [12] or metallic
waveguide for remote endoscopic measurements [13], [14].
Those studies demonstrated the feasibility and the potential of
THz guided reflectometry systems. However, the complexity
induced by the optical coupling setups (beam splitter, lenses,
parabolic mirrors) remains and limits the progress towards
compact guided sensing units.
In our work, the guided reflectometry configuration is
highly simplified by testing compact FMCW transceivers
with a dielectric thin-wall hollow-core waveguide of suitable
dimensions. An optic-free transmission channel between the
transceiver and the sample is then ensured with all the benefits
of the FMCW radar sensing technique. In the following
sections, two guided reflectometry configurations consisting of
different compact FMCW radar units are introduced in order to
demonstrate adequacy and universality of this approach regard-
ing the employed technology and the front-end of the system.
Investigations into the propagation behaviors in the waveguide
and final system performances are given, demonstrated and
further improved with the integration of an extra termination
hemispherical lens.
IEEE TRANSACTIONS ON TERAHERTZ SCIENCE AND TECHNOLOGY, VOL. XX, NO. XX, APRIL 2020 2
II. GU ID ED FMCW R ADA R RE FLE CT OM ET RY SYSTEM
A. FMCW Radar: basic principle
A simplified architecture diagram depicts, in Fig. 1a, the
typical implementation of a monostatic radar transceiver op-
erating in reflection mode. Driven by a cyclic command
signal, a linear FMCW reference signal is generated at low
frequency by a Voltage Control Oscillator (VCO) or a Phase
Locked Loop (PLL) oscillator, which is then fed into the
frequency multiplication chain, for signal up-conversion, to
reach the desired operating frequency band. A 3-ports coupler
leads this probing signal towards the emission antenna while
redirecting the reflected signal towards the mixing unit for
down-conversion and sampling. Similar bi-static architectures
are also practicable but require a partition between the emit-
ting chain and receiving unit, complicating the system and
impacting its compactness.
As showed in Fig. 1b, the cyclic signal reflected from
the target is delayed with respect to the reference emission
sweep and thus gives rise to a beating frequency fbat the
mixer’s output proportional to the propagation length. Hence
the object’s distance can be expressed as Eq. 1
d=c0∆t
2n=c0fb
2n.Ts
B,(1)
δres =c0
2n B .(2)
where ∆tis the propagation-induced time delay, nrepresents
the optical refractive index of the propagation media, fb
donates the mixer’s output beating frequency, Tsis the period
of a sweep cycle, and B is the sweep frequency bandwidth.
When multiple targets are involved, each object contributes
as a given distance-related beating frequency fbi, allowing
for differentiation and remote sensing capability through data
processing steps. Nevertheless, derived from Eq. 1, the lon-
gitudinal resolution, δres, is directly correlated to the finite
bandwidth (see Eq. 2).
Coupler
Antenna
TX/RX
Mixer
VCO/PLL
Mixer’s output
Beating signal
Multiplication Chain
Frequency
sweep
DAQ
Voltage
sweep
V(t)
Reference input
f(t)
(a)
0
0,2
0,4
0,6
0,8
1
-0,5 0 0,5 1 1,5
Reduced frequency
Reduced time
B
Ts
𝑓
𝑏
Δt
(b)
Fig. 1: (a) Simplified architecture of a monostatic FMCW radar transceiver. (b)
FMCW mode of operation: emitted linear frequency sweep reference (solid
line) and delayed reflected signal (dashed line) from a distant single object.
1) Si-Based technologies
From the FMCW mode of operation, the technological
implementation for radar transceiver units has been widely
investigated through two distinct technologies derived either
from silicon-based integrated circuit or from III/V semi-
conductor diode-based components. Thanks to the continuous
improvement of the Si-based transistors processing technology,
a steady decrease have been witnessed in those devices pro-
duction prices. Nowadays, Si-based FMCW radar transceivers
are drawing attention in industrial fields thanks to their high
integration level, low-cost and versatility. As an example,
based on previous generations at 24 GHz, radar designs for
automotive applications [15] have been developed in the 77
GHz band with simple bi-static architectures. With a similar
design, a 122 GHz transceiver based on SiGe technology [16],
[17] is developed by Silicon-Radar GmbH. Apart from the 245
GHz radar unit under development [18], this transceiver is the
only commercially available Si-based unit above 100 GHz on
the market to our knowledge. It offers a 6 GHz bandwidth
through a bi-static geometry based on separated 2x2 patch
array antennas as RX and TX, on an 8x8 mm2QFN (Quad
Flat No-leads) package.
In our work, this 120 GHz Si-based radar transceiver
chip has been integrated amongst the guided radar reflected
measurement unit as a low-cost implementation solution. In
order to push the investigations on the guided reflectometry
unit design, and more especially the coupling capabilities of
the radar units with the waveguide, preliminary electromag-
netic simulations of the radar transceivers are conducted. The
resulting simulated radiation pattern of the Silicon Radar unit,
depicts, as expected, a broad far-field emission profile in nearly
a half space (a maximal directivity of 11 dBi with an angular
width around 47◦and 48◦in E and H-plane, respectively),
corresponding to the typical response of such patch array
antennas design.
2) III/V-Based technologies
Compared to Si-based semiconductor technologies, III-V
based transistors show advantages in terms of electron mo-
bility, thermal conductivity and operating voltage. Their max-
imum frequencies have exceeded 1 THz, making it possible
to realize highly integrated, powerful, complex transmitters
and receivers at higher terahertz frequencies. Thanks to the
Schottky diodes implementation, 2 THz frequency multipli-
cation chains are achievable [19]. However, III/V-based fully
integrated MMIC [20] still displays high frequencies imple-
mentation limitations. Hybrid integration is suggested as a
compromise solution for high-end applications. It is imple-
mented through the fine tuning of a succession of components
in discrete waveguide blocks [6], [21]. III-V based radars have
been demonstrated around 100 GHz, 300 GHz and 600 GHz
for different application cases [22], and solutions up to 850
GHz have been developed by SynView GmbH.
A 100 GHz SynView FMCW transceiver, providing 50
dB measurement dynamic range, is as well implemented in
our guided terahertz FMCW reflectometry unit as a reference
system. A mono-static block-integrated architecture in combi-
nation with a conical horn antenna ensures the functionality of
IEEE TRANSACTIONS ON TERAHERTZ SCIENCE AND TECHNOLOGY, VOL. XX, NO. XX, APRIL 2020 3
this III-V based transceiver module. For this horn antenna ge-
ometry, the conduction of similar characterization simulations
depicts a much more suitable emission profile with maximum
directivity of 20 dBi, while 14.4◦angular width is achieved
in the H-plane, and 24.1◦is obtained in E-plane. Particularly,
such emission profile allows a increased forward and back
signal coupling between the antenna and the waveguide thanks
to the improved directivity.
B. Waveguide for remote sensing
Lying between the microwaves and infrared regions, the de-
velopment of terahertz waveguides benefits from the advanced
technologies in those two domains [9]. However, the absorp-
tion loss in dielectrics and skin effect power losses induced by
metals in the terahertz band limit the performances of teraherz
waveguides. Considering that dry-air is the most transparent
terahertz medium, air-core waveguides with different cladding
structures (simple hollow-core waveguides [23], [24], metal-
coated dielectric waveguides [25] and photonic crystal hollow-
core waveguides [26]), have drawn significant attention. Con-
tinuous effort is being brought to improve the transmission
efficiency through new designs and materials [27], [28]. In
this work, a thin-wall hollow-core Polypropylene waveguide
is selected to be implemented due to its simple structure,
low-loss propagation properties of the cladding material and
availability in the commercial market. This type of waveguide
exploits the anti-resonant reflections as guiding mechanism
[29]. Since the high-index cladding layer acts as a Fabry-P´
erot
etalon, the air core allows terahertz waves to propagate inside
at non-characteristic frequencies. The loose confinement in the
air core gives rise to a low-loss propagation. The cladding’s
characteristic frequencies [23] are defined by:
fm=m c
2tcl pn2
cl
−1,(3)
where cis the speed of the light in vacuum, mis an integer, tcl
and ncl are the thickness and refractive index of the cladding
respectively.
In addition to the suitability of the cladding thickness to
ensure the guiding mechanism, one of the key parameters to
be determined for the selection of such a waveguide, remains
its diameter to establish the optimum power coupling between
the antennas and the waveguide.
Due to the complexity of the involved anti-resonant guiding
mechanism and the irregularities in the emission profiles of the
transceivers, electromagnetic simulations have been conducted
on the two considered radar units to investigate the impact
of the waveguide diameter and positioning on the coupling
capabilities. Similarly to the power coupling with the horn
antenna geometry, and with the same evolution tendency,
the coupling efficiency displays an important dependency on
the waveguide diameter for the RX and TX patch antennas
geometry of the Silicon-radar unit, (see Fig. 2) and displays
a power coupling of 30% for an optimum 2 mm waveguide
radius.
Experimentally, a 3 mm radius plastic pipe waveguide with
a cladding thickness of 0.158 mm is chosen thanks to the
market availability and ensures a trade-off between coupling
0 1 2 3 4 5 6
0
5
10
15
20
25
30
35
Fig. 2: Evolution of the simulated forward power coupling efficiency on a
300 mm long guided propagation with respect to the waveguide’s diameter
for the Silicon radar unit frond-end design.
efficiency and expected optical lateral resolution. According
to Eq. 3, with t = 158 µm and n = 1.49 (THz-TDS time-
domain spectroscopy extracted refractive index), characteristic
resonant frequencies of this plastic cladding are expected to be
around 860 GHz and 1.72 THz [10], allowing the waveguide to
support 100 GHz waves’ propagation without significant im-
pacts on the considered FMCW radars bandwidth. Those char-
acteristic non-supported propagation frequencies have been
witnessed through a TDS investigation on the same wave-
guide, demonstrating losses peaks around 840 GHz and 1.7
THz.
C. Experimental setup
As a simplified guided configuration, an optics-free single
communication path between the FMCW module and sample
is provided by the previously mentioned hollow-core dielectric
waveguide. In order to ensure an adequate coupling, the
waveguide is properly placed with respect to the transceiver.
No dedicated quasi-optical setup or beam shaping elements are
employed. As a simple optimization, a foam support, acting
as terahertz absorber, is placed nearby the coupling area to
neutralize the reflected parasitic signals induced by the non-
coupled echoes, resulting in a significant signal improvement.
Unlike from guided continuous wave reflectometry mea-
surement techniques, where the detected signal is a super-
position of all the contributions that are generated along the
waveguide, the phase information provided by guided FMCW
techniques unlocks depth sensing capabilities and allows a
differentiation of those contributions. More specifically, be-
side a sensing capability along the waveguide with distance
differentiation, it ensures a drastic improvement of the signal-
to-noise ratio thanks to the selection of the sensing distance,
linked to the desired contributions, while not taking into ac-
count any parasitic signal. This capability represents the main
motivation behind the implementation of waveguides with
millimeter wave FMCW radar units rather than continuous
wave sources, even though, the low longitudinal resolution
of the considered radar units does not allow depth sensing
measurements at the waveguide’s output.
With a purpose of demonstrating the simplicity of this ap-
proach regardless of the operated technology, system emission
IEEE TRANSACTIONS ON TERAHERTZ SCIENCE AND TECHNOLOGY, VOL. XX, NO. XX, APRIL 2020 4
frond, packaging, and price, the FMCW guided reflectometry
unit is established with two specifically distinct transceivers.
Implementing the 100 GHz SynView radar head as a
reference configuration (see Fig. 3a), the waveguide is directly
inserted in the horn antenna to maximize the forward-coupled
power while ensuring a proper centering. This symmetric po-
sitioning of the waveguide and the linearly- polarized emission
profile of horn antenna are beneficial parameters to excite the
air-core propagation modes for low-loss transmission along the
waveguide.
As a compact and low-cost approach, the 122 GHz Silicon-
Radar chip displays a full integration with two separated
patch array antennas as TX and RX. Although the coupling
processes are no longer symmetrical due to a 3 mm TX-RX
separation, the positioning of the hollow-core waveguide is
centered on the focal point as defined in [16] , equidistant
from the TX and RX patch antenna arrays, while the foam
absorber is still employed to suppress parasitic non-coupled
signals (see Fig. 3b).
(a)
(b)
Fig. 3: Guided reflectometry experimental setups : (a) incorporating the 100
GHz SynView unit (1) with the waveguide support (2) and the scanned test
target (3); (b) using compact Si-Radar 122 GHz chip (1), held beside the
cardboard centering ring of the waveguide support (2).
III. RES ULTS
In addition to the experimental setup implementation, 3D
full-wave electromagnetic simulations are performed to in-
vestigate and further characterize the propagation properties
of radar signals in guided terahertz reflectometry systems.
Depending on the employed radar unit, two models in as-
sociation with horn antenna and patch antennas respectively
are constructed. Both are simulated using the 3D full-wave
simulation suite, CST microwave studio, with a time-domain
solver applying finite integration technique while the profiles
of the electric fields are recorded for further performances
evaluations. It should be pointed out that no material losses
(neither in dielectric materials nor in metals) are considered
in those simulations.
A. 100 GHz SynView transceiver reflectometry unit
As shown in Fig. 4a, the simulation model based on the 100
GHz SynView transceiver reflectometry unit is composed of
a typical W-Band conical horn antenna, a 30 cm long plastic
hollow-core waveguide and a 5 cm long air block. The circular
input face of the conical horn antenna is selected as a port for
signal excitation and detection. All the E-field results presented
below are obtained at the central frequency of 100 GHz.
1) Propagation characterisation
Field distribution simulation results, displayed in Fig. 4b
and 4c, provide substantial information concerning the prop-
agation behaviors of guided waves in the system which are
tightly related to the profile of the injection source.
Over the first part of the propagation, a stabilization of the
guided signal is noticeable, while towards the end of the 30
cm waveguide, a steady field distribution is retained which
corroborates the guidance capacity of the waveguide.
However, the field distribution along the waveguide reveals
that not only air-core modes, but also cladding modes are
excited, resulting in the significant field confinement in the
two sides of the cladding in the x-direction. Both the the anti-
resonant reflection and total internal reflection are exploited
as guiding mechanisms. The low effective refractive index,
induced by the thin cladding pipe geometry, still ensures a
low attenuation and prevents any significant back-reflection
from the output waveguide-air interface to restrain the parasitic
standing waves formation along the waveguide. Nevertheless,
with this guide geometry, the field distributions, in Fig. 4d, also
display propagating modes around the cladding. The external
field spillage, induced by the low confinement of this mode,
enables an additional beneficial sensing capability along the
waveguide. Indeed, combined with the high sensitivity and the
phase information natively embedded with the FMCW sensing
method, this external field distribution allows for reflective
sensing and differentiation of inserted perturbations in the
vicinity of the waveguide.
2) Coupling-in issue
As mentioned previously, the coupling is achieved directly
by inserting the waveguide into the conical horn antenna of
the radar unit to get a proper centering and field collection. To
characterize the coupling efficiency of this assembly, due to the
non-controlled multi-mode propagation, the electromagnetic
power is evaluated at different positions along the waveguide,
from integration of the Poynting vector over the surface of
cross-sections. Fig. 5 depicts simulated enclosed power as
a function of the propagation distance. The power features
over the first 30 mm (Phase I), correspond to the excitation
signal confined in the metallic horn antenna. The gradual
decrease (Phase II) is induced by the radiation loss of non-
coupled waves. The stable guided propagation then takes
place along the waveguide (Phase III) where no significant
power losses induced by material absorption or radiation loss
caused by non-guided waves are witnessed. The coupling
IEEE TRANSACTIONS ON TERAHERTZ SCIENCE AND TECHNOLOGY, VOL. XX, NO. XX, APRIL 2020 5
Fig. 4: (a) Diagram of the guided reflectometry radar setup using the 100 GHz SynView unit, simulated electric field magnitude profile along (b) the z-y
plane and (c) the z-x plane, (d) xy cut of the electric field magnitude profile at respectively (I) z = 0 mm, (II) z = 30 mm, (III) z = 190 mm, (IV) z = 300
mm and (V) 5 mm, (VI) 20 mm from the waveguide’s output.
efficiency, dictated by Eq. 4, and hereby directly deduced from
the simulation steady state, then approaches 70%. Once the
radar signal propagates out of the waveguide (Phase IV), the
enclosed power reduction occurs due to the beam expansion
in free space.
η=PGuided
PExcitation
,(4)
Fig. 5: Simulation of the enclosed power as a function of the propagation
distance revealing the coupling ratio when reaching guided steady-state, for
the 100 GHz SynView guided module.
3) Dynamic range
For such reflectometry sensing tools, the measurement dy-
namic range of the unit is one of the main performance indi-
cators for sensing purposes. The 100 GHz SynView FMCW
Radar unit itself offers a maximum 50 dB dynamic range when
used in standard quasi-optical imaging configuration. Never-
theless, the implementation of the waveguide will ineluctably
induce a limitation on this dynamic range. Indeed,impacting
the signal as a constant background noise, the unavoidable
signal reflection from the open probing end of the waveguide
will be the main limiting factor for the dynamic range together
with the power loss induced by the imperfect waveguide-
antenna coupling. Appropriate simulations have been con-
ducted to asses this impact via the comparison of the reflected
power for two extreme cases, with first the simulation of the
open-end waveguide as a background signal characterisation
and secondly, the implementation of a perfect reflector for
the assessment of the optimum achievable signal level. A -
28.5 dB amplitude reflection coefficient from the open-end
back-reflection has been deducted from the comparison of the
reflected field amplitude with the simulation input excitation.
Similarly, inserting a perfect reflector at the end of the guide,
a -1.5 dB return signal coefficient has been deducted from
those simulations, mainly impacted by the imperfect coupling
from the horn antenna to the waveguide. A total maximum
achievable dynamic range of 27 dB, dictated by the ratio
between the previous extreme reflection coefficients can then
be expected.
4) Optical resolution and imaging capabilities
The free space propagation behaviors after the waveguide is
of interest to characterize the sensing and imaging capabilities
of the entire system. Indeed, the propagation properties of
radar waves leaving the waveguide determine the optimum
sensing distance as well as the lateral resolution at the probing
point. While the maximum sensing distance is related to
the back-coupling efficiency, the resolution capabilities is
established by the beam profile, the latter being especially
important for imaging purposes. Fig. 4d depicts the simulated
electric field distribution in the xy-plane at different positions.
It can be observed that, although the back-coupling effi-
ciency is optimized in the vicinity of the waveguide’s output,
(Fig. 4d (IV), the beam profiles display 2 main amplitude
lobes confined in the opposite sides of waveguide’s cladding.
This inhomogeneity directly leads to ghost imaging artifacts
causing a duplication of the object when using the system in
such a close configuration.
The impact of this optical profile heterogeneity is noticeable
on the image of a metal-on-PCB USAF resolution test chart,
performed by classical x-y raster-scanning method, resulting
from collected radar data in the waveguide output plane, where
some vertical elements are duplicated (see Fig. 6b). Neverthe-
IEEE TRANSACTIONS ON TERAHERTZ SCIENCE AND TECHNOLOGY, VOL. XX, NO. XX, APRIL 2020 6
less, after propagation, the beam profile is then smoothed due
to the spreading of the two electric fields lobes (see Fig. 4d
(V)). It reach an optimum imaging resolution of 4.5 mm at
Full Width Half Maximum thanks to a quite homogeneous
profile at an optimum working distance of 5 mm. Working at
larger distances, the back-coupling efficiency along with the
(a)
Mean reflexion Image
0 50 100 150 200
X(mm)
0
50
100
150
Y(mm)
150
160
170
180
190
200
(b)
Mean reflexion Image
0 50 100 150 200
X(mm)
0
50
100
150
200
Y(mm)
130
140
150
160
170
180
190
(c)
Fig. 6: (a) Photograph of test chart. (b) Raster scan acquired image, with
1mm step size, using the guided 100 GHz Synwiew probing system in lens-
less geometry at 1 mm imaging distance. (c) Raster scan acquired image
using the regular far-field 100 GHz SynView unit at NA=0.5 for reference,
with 1mm step size.
resolution will decrease due to the free-space beam spreading,
leading to a trade-off between the back-coupling efficiency
and lateral resolution if larger working distances are required.
For comparative purposes, the far field image obtained with
the classical quasi optical implementation of the 100 GHz
SynView unit is displayed in Fig. 6c.
5) Optical resolution improvement: solid immersion lens
implementation
To address the limitations of the imaging resolution linked
to the waveguide’s diameter while avoiding heavy configu-
ration with optical components, an end-of-waveguide solid
immersion lens has been selected as an adequate solution.
Thanks to their specific geometries and materials, typically
hemispherical, hypo or hyper-hemispherical or bullet dielectric
lenses designs, solid immersion lenses allow high Numerical
Aperture (NA) focusing for high resolution imaging and short
working distances.
(I)
(IV) (V)
(II)
E Field ( dB Max)
lens
Waveguide
(a)
Max reflexion Image
0 50 100 150 200 250
X(mm)
0
50
100
150
Y(mm)
150
160
170
180
190
200
(b)
Fig. 7: (a) Electromagnetic simulations with solid immersion HDPE hemi-
spherical lens : (I) implementation diagram of the guide termination, electric
field magnitude profile (II) in z-y plan, (III) z-x plan and (IV) x-y cut in
the best imaging plan, (V) evolution of the beam diameter and maximum
power density with respect to the working distance. (b) Raster scan acquired
image, with 1 mm step size, using the guided Synwiew sensing unit with
hemispherical HDPE termination lens at 2.5 mm imaging distance.
In the experiment, a 9 mm diameter hemispherical High
Density Polyethylene (HDPE) lens is inserted at the output of
IEEE TRANSACTIONS ON TERAHERTZ SCIENCE AND TECHNOLOGY, VOL. XX, NO. XX, APRIL 2020 7
Fig. 8: (a) Diagram of the guided reflectometry radar setup using the 122GHz Silicon radar chip, simulated electric field magnitude profile along (b) the z-y
plan and (c) the z-x plane, (d) x-y cut of the electric field magnitude profile at respectively (I) z = 0 mm, (II) z = 140 mm, (III) z = 180 mm, (IV) z= 300
mm and (V) 3 mm, (VI) 20 mm from the output of the waveguide.
the waveguide (see Fig. 7a (I)). Its relatively low refractive
index induces a moderate unwanted signal return from the in-
terfaces while contact-less sensing is still ensured by the back
working distance linked to its diameter. Indeed, simulating this
lens geometry at the output of the waveguide unit (see Fig. 7a
(II) (III)), a beam optimum spot size of 2 mm at Full Width at
Half Maximum in X and Y directions is obtained at a frontal
working distance of 2.5 mm, along with the optimum power
density observed around this point (see Fig. 7a (IV) (V)) while
residual low amplitude side lobes at -25 dB are induced by
the waveguide spillage. This simulated profile corroborates the
resolution of the raster scan acquired image of a test chart
where we found a 2 mm resolution (group -2, element 1) (see
Fig. 7b) is achieved around the expected simulated optimum
working distance of 2 mm and depicts a drastic improvement
compared to the guided lens-less FMCW implementation or
a classic FMCW far-field imaging setup. Working distances
beyond or closer from this optimum point obviously degrades
the achievable resolution and are drastically impacting the
back coupling efficiency.
B. 122 GHz Silicon radar chip reflectometry unit
Focusing on the low-cost alternative unit, as illustrated in
Fig. 8a, the simulation setup remains similar to the previous
model with a simple substitution of the transceiver module,
which now consists of two 2*2 patch antennas as TX and RX
providing a wide emission pattern. Limited by the simulation
voxel’s size definition and so the calculation capacity, the
transceiver chip is replaced by its equivalent source. All E-
field results presented in the following part are obtained at the
operational bandwidth central frequency of 122 GHz.
Compared to the 100 GHz SynView head, Siliconradar chip
displays a different front-end architecture: patch antennas lead
to a much broader emission pattern. Significant modifications
of the electric field distribution along the waveguide were
expected and is depicted in Fig. 8. Contrary to the symmetrical
integration with a horn antenna, obvious signal reflections on
the waveguide’s cladding can be observed through the oscillat-
ing behavior during the propagation. This can be explained by
the decentering of the transmitter with respect to the waveg-
uide, the excitation conditions not being symmetrical anymore.
It is worth noting that the non-negligible field spillage locating
at the waveguide’s periphery grants the peripheral sensing
capability.
Fig. 9: Simulation of the enclosed power as a function of propagation distance
revealing the coupling ratio for the 122 GHz Si-Based radar chip guided
module.
Owing to the wide emission pattern of the patch antenna
design, a considerable power fraction is radiated into free
space rather than coupled into the waveguide. Fig. 9 Phase
I depicts this power loss while phase II indicates that the
radar signal reaches guided power stabilization after a 50 mm
propagation. As a direct consequence, the coupling efficiency
is reduced to 18% and mainly impacted by the near-isotropic
emission profile of the patch antennas, compared to the
previously achieved 70% with the directional emission of the
horn antenna of the former setup. Moreover, based on remain-
ing field energy after open-end reflection and perfect mirror
reflection, the estimated full dynamic range can reach up to
IEEE TRANSACTIONS ON TERAHERTZ SCIENCE AND TECHNOLOGY, VOL. XX, NO. XX, APRIL 2020 8
27.6 dB. Although the dynamic ranges of this reflectometry
system are similar in comparison to the SynView 100 GHz
architecture a much larger coupling loss is witnessed in the
system using the 122 GHz integrated radar chip.
An image of the test chart obtained in this reflectometry
configuration is provided in Fig. 10a. The achieved resolution
of 2.8 mm (group -3 element 4), directly at the output
of the waveguide, corroborates the simulated Full Width
at Half Maximum output’s beam diameter, thus reaching a
better resolution level than the previously described reference
system. The atypical coupling-in geometry gives rise to an
asymmetric field distribution at the output of the waveguide,
where the highest electric field is confined at one side of the
core’s cross section (Fig. 8c and d (IV)). Even though an
irregular beam pattern is emitted, the single lobe profile output
field distribution ensures no significant ghost artefacts but the
asymmetry of the emission and the back coupling induce a
shadowing effect, noticeable on the image. In spite of it, within
the longitudinal resolution imposed by the radar bandwidth
the sample support on the backside can be seen in the center
bottom part of the image, demonstrating the in-depth sensing
capability of the system.
Mean reflexion Image
0 50 100 150 200
X(mm)
0
50
100
Y(mm)
100
120
140
160
180
(a)
Max reflexion Image
0 50 100 150 200
X(mm)
0
20
40
60
80
100
120
Y(mm)
130
140
150
160
170
180
(b)
Fig. 10: Raster scan images, with 500 µm step size, of a test chart using the
guided 122 GHz Si-radar probing system in two configurations: (a) in lens-
less geometry with an imaging distance of 1 mm and (b) in association with
a hemispherical PE termination lens with an imaging distance of 3 mm.
Similarly to our reference setup, for the sake of imaging res-
olution improvement, a 9 mm diameter hemispherical HDPE
termination lens has been inserted to counteract the free space
propagation beam divergence and reach a better resolved focus
point. Fig. 10b displays the raster scan acquired image of the
test chart in this configuration where a 1.4 mm resolution
is achieved (group -2 element 4) in accordance with the
conducted beam profile simulations, better than the resolution
obtained with the reference 100 GHz guided SynView unit.
Nevertheless ghost artefacts are still noticeable and induced
by beam profile inhomogeneities leading to an improper beam
focusing with non-negligible side lobes, witnessed as well on
those simulations.
IV. CON CL US IO NS
In this work, guided terahertz FMCW reflectometry probing
systems have been demonstrated as a low-cost solution where
a polymer pipe waveguide is used to ensure the single coupling
channel to reach a compact and simplified guided probing
unit. Two architectures have been investigated to implement a
high performances unit as well as a low-cost solution. With
the help of the 3D full-wave electromagnetic simulations,
the propagation behaviors in the guided systems have been
investigated to corroborate implementation results. In partic-
ular, a quantitative analysis focused on waveguide coupling
efficiency, expected sensing dynamic range and optical reso-
lution have been performed. Finally, imaging capabilities of
those guided systems are demonstrated by the raster-scanning
method on test charts, which supplement and validate the
simulations as well as the waveguide propagation induced
artefacts. Intrinsically limited by the waveguide’s dimension,
further improvements on the imaging resolution have been
investigated through the implementation of solid immersion
lenses while ongoing work are directed towards the design
and selection of those lenses according to their intended use
as well as the selection of an improved waveguide dimension
for better coupling performances with both systems.
ACK NOW LE DG ME NT
This work was supported in part by the French Ministry
of National Education, Research and Technology. Authors
acknowledge Dr. Damien Bigourd and Dr. Quentin Cassar
for their interesting discussions, advises upon terahertz waveg-
uides and their additional checks of this article.
REF ER EN CE S
[1] P. Hillger, J. Grzyb, R. Jain, and U. R. Pfeiffer, “Terahertz imaging
and sensing applications with silicon-based technologies,” IEEE Trans-
actions on Terahertz Science and Technology, vol. 9, no. 1, pp. 1–19,
2018.
[2] S. Dhillon, M. Vitiello, E. Linfield, A. Davies, M. C. Hoffmann,
J. Booske, C. Paoloni, M. Gensch, P. Weightman, G. Williams, et al.,
“The 2017 terahertz science and technology roadmap,” Journal of
Physics D: Applied Physics, vol. 50, no. 4, p. 043001, 2017.
[3] J.-B. Perraud, J.-P. Guillet, O. Redon, M. Hamdi, F. Simoens, and
P. Mounaix, “Shape-from-focus for real-time terahertz 3d imaging,”
Optics letters, vol. 44, no. 3, pp. 483–486, 2019.
[4] J. B. Perraud, A. F. Obaton, J. Bou-Sleiman, B. Recur, H. Balacey,
F. Darracq, J.-P. Guillet, and P. Mounaix, “Terahertz imaging and tomog-
raphy as efficient instruments for testing polymer additive manufacturing
objects,” Applied optics, vol. 55, no. 13, pp. 3462–3467, 2016.
[5] T. M. Wong, M. Kahl, P. H. Bol´
ıvar, and A. Kolb, “Frequency modulated
continuous wave (fmcw) thz image 3d superresolution,” arXiv preprint
arXiv:1802.05457, 2018.
[6] F. Friederich, K. May, B. Baccouche, C. Matheis, M. Bauer,
J. Jonuscheit, M. Moor, D. Denman, J. Bramble, and N. Savage, “Tera-
hertz radome inspection,” in Photonics, vol. 5, no. 1. Multidisciplinary
Digital Publishing Institute, 2018, p. 1.
IEEE TRANSACTIONS ON TERAHERTZ SCIENCE AND TECHNOLOGY, VOL. XX, NO. XX, APRIL 2020 9
[7] E. Cristofani, F. Friederich, S. Wohnsiedler, C. Matheis, J. Jonuscheit,
M. Vandewal, and R. Beigang, “Nondestructive testing potential eval-
uation of a terahertz frequency-modulated continuous-wave imager for
composite materials inspection,” Optical Engineering, vol. 53, no. 3, p.
031211, 2014.
[8] J.-P. Guillet, M. Roux, K. Wang, X. Ma, F. Fauquet, H. Balacey,
B. Recur, F. Darracq, and P. Mounaix, “Art painting diagnostic before
restoration with terahertz and millimeter waves,” Journal of Infrared,
Millimeter, and Terahertz Waves, vol. 38, no. 4, pp. 369–379, 2017.
[9] G. Humbert, Optical Fibers in Terahertz Domain. Singapore: Springer
Singapore, 2019, pp. 1–49.
[10] M. Pan, Q. Cassar, F. Fauquet, G. Humbert, P. Mounaix, and J.-P. Guillet,
“Guided terahertz pulse reflectometry with double photoconductive
antenna,” Appl. Opt., vol. 59, no. 6, pp. 1641–1647, Feb 2020.
[11] Y. Gao and R. Zoughi, “Millimeter wave reflectometry and imaging
for noninvasive diagnosis of skin burn injuries,” IEEE Transactions on
Instrumentation and Measurement, vol. 66, no. 1, pp. 77–84, 2017.
[12] B. You and J.-Y. Lu, “Remote and in situ sensing products in chemical
reaction using a flexible terahertz pipe waveguide,” Optics express,
vol. 24, no. 16, pp. 18 013–18 023, 2016.
[13] J.-P. Guillet, L. Chusseau, R. Adam, T. Grosjean, A. Penarier, F. Baida,
and D. Charraut, “Continuous-wave scanning terahertz near-field micro-
scope,” Microwave and Optical Technology Letters, vol. 53, no. 3, pp.
580–582, 2011.
[14] P. Doradla, K. Alavi, C. S. Joseph, and R. H. Giles, “Flexible waveguide
enabled single-channel terahertz endoscopic system,” in Terahertz, RF,
Millimeter, and Submillimeter-Wave Technology and Applications VIII,
vol. 9362. International Society for Optics and Photonics, 2015, p.
93620D.
[15] J. Hasch, E. Topak, R. Schnabel, T. Zwick, R. Weigel, and C. Wald-
schmidt, “Millimeter-wave technology for automotive radar sensors in
the 77 ghz frequency band,” IEEE Transactions on Microwave Theory
and Techniques, vol. 60, no. 3, pp. 845–860, 2012.
[16] M. Pauli, B. G¨
ottel, S. Scherr, A. Bhutani, S. Ayhan, W. Winkler, and
T. Zwick, “Miniaturized millimeter-wave radar sensor for high-accuracy
applications,” IEEE Transactions on Microwave Theory and Techniques,
vol. 65, no. 5, pp. 1707–1715, 2017.
[17] K. Schmalz, W. Winkler, J. Borngr¨
aber, W. Debski, B. Heinemann, and
J. Scheytt, “122 ghz ism-band transceiver concept and silicon ics for
low-cost receiver in sige bicmos,” in 2010 IEEE MTT-S International
Microwave Symposium. IEEE, 2010, pp. 1332–1335.
[18] K. Schmalz, J. Borngr¨
aber, B. Heinemann, H. R¨
ucker, and J. Scheytt, “A
245 ghz transmitter in sige technology,” in 2012 IEEE Radio Frequency
Integrated Circuits Symposium. IEEE, 2012, pp. 195–198.
[19] A. Maestrini, J. Ward, J. Gill, H. Javadi, E. Schlecht, G. Chattopadhyay,
F. Maiwald, N. R. Erickson, and I. Mehdi, “A 1.7-1.9 thz local oscillator
source,” IEEE Microwave and Wireless Components Letters, vol. 14,
no. 6, pp. 253–255, 2004.
[20] M. Abbasi, S. E. Gunnarsson, N. Wadefalk, R. Kozhuharov, J. Svedin,
S. Cherednichenko, I. Angelov, I. Kallfass, A. Leuther, and H. Zirath,
“Single-chip 220-ghz active heterodyne receiver and transmitter mmics
with on-chip integrated antenna,” IEEE Transactions on Microwave
Theory and Techniques, vol. 59, no. 2, pp. 466–478, 2010.
[21] K. B. Cooper, R. J. Dengler, N. Llombart, T. Bryllert, G. Chattopadhyay,
E. Schlecht, J. Gill, C. Lee, A. Skalare, I. Mehdi, et al., “Penetrating 3-d
imaging at 4-and 25-m range using a submillimeter-wave radar,” IEEE
Transactions on Microwave Theory and Techniques, vol. 56, no. 12, pp.
2771–2778, 2008.
[22] C. Am Weg, W. von Spiegel, R. Henneberger, R. Zimmermann, T. Lo-
effler, and H. G. Roskos, “Fast active thz cameras with ranging capa-
bilities,” Journal of Infrared, Millimeter, and Terahertz Waves, vol. 30,
no. 12, pp. 1281–1296, 2009.
[23] C.-H. Lai, B. You, J.-Y. Lu, T.-A. Liu, J.-L. Peng, C.-K. Sun, and H.-C.
Chang, “Modal characteristics of antiresonant reflecting pipe waveguides
for terahertz waveguiding,” Optics express, vol. 18, no. 1, pp. 309–322,
2010.
[24] E. Nguema, D. F´
erachou, G. Humbert, J.-L. Auguste, and J.-M. Blondy,
“Broadband terahertz transmission within the air channel of thin-wall
pipe,” Optics letters, vol. 36, no. 10, pp. 1782–1784, 2011.
[25] B. Bowden, J. A. Harrington, and O. Mitrofanov, “Silver/polystyrene-
coated hollow glass waveguides for the transmission of terahertz radia-
tion,” Optics letters, vol. 32, no. 20, pp. 2945–2947, 2007.
[26] Y. Geng, X. Tan, P. Wang, and J. Yao, “Transmission loss and dispersion
in plastic terahertz photonic band-gap fibers,” Applied Physics B, vol. 91,
no. 2, pp. 333–336, 2008.
[27] I. K. Yakasai, P. E. Abas, S. Ali, and F. Begum, “Modelling and
simulation of a porous core photonic crystal fibre for terahertz wave
propagation,” Optical and Quantum Electronics, vol. 51, no. 4, p. 122,
2019.
[28] M. A. Habib, M. S. Reza, L. F. Abdulrazak, and M. S. Anower, “Ex-
tremely high birefringent and low loss microstructure optical waveguide:
Design and analysis,” Optics Communications, vol. 446, pp. 93–99,
2019.
[29] C.-H. Lai, Y.-C. Hsueh, H.-W. Chen, Y.-j. Huang, H.-c. Chang, and C.-
K. Sun, “Low-index terahertz pipe waveguides,” Optics letters, vol. 34,
no. 21, pp. 3457–3459, 2009.
Mingming Pan was born in Jiangsu, China, in
1992. She received the B.Sc. degree in electronics
from the University of Suzhou, Suzhou, China, in
2013, and the M.Sc. degree in electronics from
Bordeaux University, Bordeaux, France, where she
is currently working toward the Ph.D. degree in
electronics, entitled “Reflectometry in time-domain
for guided terahertz waves.” She was involved at
IMS Laboratory (UMR CNRS 5218) on emission
and detection of THz waves techniques.
Adrien Chopard was born in France, in 1993. He
received the Diplome d’Ing´
enieur in optics and pho-
tonics from the Institut d’Optique Graduate School
(IOGS-Saclay), Palaiseau, France, in 2016 and the
M.Sc. degree in engineering specialized in nanotech-
nology from the Royal Institute of Technology of
Sweden, (KTH), Stockholm, Sweden. He is cur-
rently working toward the industrial Ph.D. degree in
physics on applicative developments using terahertz
technologies, with the firm Lytid based in Paris and
the IMS Laboratory (UMR CNRS 5218), University
of Bordeaux, Talence, France.
Frederic Fauquet was born in Pessac, France, in
1977. He received the Diplome Universitaire en
Technologie in electronics from Bordeaux Univer-
sity, Bordeaux, France, in 1998. He worked for 14
years with the Observatoire de Physique du Globe
(OPGS UMS 833), Clermont-Ferrand, France, where
he contributed to the development of autonomous
multisensor station. In 2014, he moved to Bordeaux
to join the laser and terahertz team at IMS Labora-
tory as a Research Engineer.
IEEE TRANSACTIONS ON TERAHERTZ SCIENCE AND TECHNOLOGY, VOL. XX, NO. XX, APRIL 2020 10
Patrick Mounaix received the Engineering degree
in material science from the ´
Ecole Universitaire
D’Ing´
enieur de Lille (EUDIL), Villeneuve-d’Ascq,
France, in 1988, and the Ph.D. degree in quantum
devices from the University of Lille, Lille, France,
in 1992. He joined CNRS and the High Frequency
Department, Institut d’ ´
Electronique et de Micro-
´
electronique du Nord (IEMN UMR 8520). In 2002,
he moved to Bordeaux where he developed terahertz
spectroscopy and imaging activities. He is currently
a Senior Researcher with IMS Laboratory (UMR
CNRS 5218), Bordeaux University, Bordeaux, France. He has coauthored
more than 210 papers. His current research interests include equilibrium and
time-resolved terahertz spectroscopy techniques for solid and liquid materials
and 3-D terahertz imaging applications, chiefly 3-D computed tomography for
art science and NDE industrial applications. He also works on all dielectric
metamaterials and their applications at microwave and terahertz frequency
range. He has coauthored more than 210 papers.
Jean-Paul Guillet was born in Lyon, France, in
1984. He received the M.Sc. and Ph.D. degrees in
terahertz imaging from the IES Lab, Montpellier,
France, in 2007 and 2010, respectively. The topic of
this first research work was terahertz near-field mi-
croscopy using electronics-based components. Dur-
ing his postdoctoral studies, until 2013, his research
focused on imaging and tomography, using fem-
tosecond pulses, frequency-modulated continuous-
wave approach, and a terahertz camera. In 2013,
he was an Associate Professor with the IMS (UMR
CNRS 5218), University of Bordeaux, where he is currently focusing on very
large-scale integration circuits fault isolation laser techniques and terahertz
imaging.