sensors

Review
Industrial Applications of Terahertz Sensing: State
of Play
Mira Naftaly 1, *, Nico Vieweg 2 and Anselm Deninger 2
1 National Physical Laboratory, Hampton Road, Teddington TW11 0LW, UK
2 TOPTICA Photonics AG, Lochhamer Schlag 19, 82166 Gräfelfing, Germany;
[email protected] (N.V.); [email protected] (A.D.)
* Correspondence: [email protected]




Received: 22 August 2019; Accepted: 25 September 2019; Published: 27 September 2019


Abstract: This paper is a survey of existing and upcoming industrial applications of terahertz
technologies, comprising sections on polymers, paint and coatings, pharmaceuticals, electronics,
petrochemicals, gas sensing, and paper and wood industries. Finally, an estimate of the market size
and growth rates is given, as obtained from a comparison of market reports.

Keywords: terahertz; industrial applications; polymers; paint coatings; pharmaceuticals; electronic

circuits; petrochemicals; gas sensing; paper and wood

1. Introduction
According to the now-famous Gartner hype cycle [1] (Figure 1), newly invented technologies
experience a typical development path: from discovery to exaggerated expectations, followed by
disillusionment, and on to the mature phase of steady productivity and growth. If mapped onto
this scheme,
Sensors 2019, 19,THz
4203 sensing is currently on the threshold of this desirable “plateau of productivity”,
2 ofat
35
technology readiness level (TRL) 6–8.

peak of inflated
expectations
Expectations

THz

plateau of
productivity

innovation trough of
trigger disillusionment

1990 2000 2010 2020

Time

Figure 1. The Gartner Hype Cycle, TRL evolution, and the current envisaged position of
Figure
THz 1. The Gartner Hype Cycle, TRL evolution, and the current envisaged position of THz technologies.
technologies.

2. Instrumentation
The following publications well illustrate this development. Siegel [2] and Schmuttenmaer [3]
(2003 and 2004, at the peak of hype) enthusiastically highlighted the possibilities and promise of THz
To date, the vast majority of THz measurements have been performed using time-domain
technologies. Stöhr [4] (2009, beginning of disillusionment) highlighted the potential, yet with a note
spectroscopy (TDS). Hundreds of papers have been published describing the technique and
of caution, acknowledging the challenges of implementation. Armstrong [5] (2012, the trough) in a
analysing its various aspects, including numerous excellent reviews [8–10]. Some of the salient points
to bear in mind when using TDS are noted below. In order to utilise TDS data, it is necessary to
understand
Sensors 2019, 19,certain aspects of system performance and the correct interpretation
4203; doi:10.3390/s19194203 of the results.
www.mdpi.com/journal/sensors

• TDS is acquired in the time domain, i.e., it follows the time evolution of the signal (Figure 2a).
Spectral data is derived by applying a Fourier transform (Figure 2b). As a result, there are two
Sensors 2019, 19, 4203 2 of 35

paper notably titled “The Truth about Terahertz” underscored the difficulties and barriers to THz
applicability. Hochrein [6] (2015, slope of enlightenment) carried out market research, concluding that
THz technologies will continue to grow in acceptance and usage, albeit primarily in specialised R&D
and measurement labs, and in a recent report (2019, approaching productivity) Frost and Sullivan
included THz sensors as one of the top 50 emerging technologies [7].
At the time of writing, there are some existing industrial installations and several ongoing field
trials of THz sensors; much development work is proceeding apace. Since THz sensing is an optical
technique, it is natural that the great majority of envisaged industrial applications pertain to forms of
non-destructive testing and condition monitoring. In particular, many deployments address thickness
measurements of polymer structures and coatings. In this paper, we present a review of some of the
more prominent and potentially promising demonstrators of THz sensing in industry. Section 2 is
a brief description of the most widely used THz technologies. This is followed by sections focused
on specific areas of application: polymers (Section 3), paint and coatings (Section 4), pharmaceuticals
(Section 5), electronics (Section 6), petrochemicals (Section 7), gas sensing (Section 8), and paper
and wood (Section 9). THz communications, which is a rapidly developing technology aimed at
mass-market usage, lies outside the scope of this paper.

2. Instrumentation
To date, the vast majority of THz measurements have been performed using time-domain
spectroscopy (TDS). Hundreds of papers have been published describing the technique and analysing
its various aspects, including numerous excellent reviews [8–10]. Some of the salient points to bear
in mind when using TDS are noted below. In order to utilise TDS data, it is necessary to understand
certain aspects of system performance and the correct interpretation of the results.
• TDS is acquired in the time domain, i.e., it follows the time evolution of the signal (Figure 2a).
Spectral data is derived by applying a Fourier transform (Figure 2b). As a result, there are
two different approaches available for analysing the sensor data. Either time-domain or
frequency-domain (i.e., spectral) data may be more useful for a particular application.
• TDS operates in a signal-probe configuration and uses short pulses (the THz pulse length is ~1 ps
and the probe pulse length is <0.1 ps). As a consequence, there are no standing waves formed in
the system itself or in the material/object measured. This simplifies data analysis.
• Because TDS uses short pulses and coherent detection, it provides an unambiguous measurement
of the electromagnetic field amplitude and phase, which directly yield transmission loss and
phase delay, and which in turn can be used to derive the absorption coefficient and refractive
index of the material measured (Figure 2c).
• In most cases, TDS applications require a reference measurement, i.e., a data set recorded
either without any sample or with reference material in place. However, once a reference data
set is obtained for a specific system configuration, TDS instruments do not require repeated
frequent calibration.
• The nominal operational bandwidth of a typical TDS system is 0.1–5 THz. However, this is
reduced by transmission loss (according to a well-formulated dependence). The typical frequency
resolution is ~5 GHz, down to 1 GHz is achievable. TDS is inherently broadband: each time-domain
pulse trace yields a complete spectral data set.
Figure 2 illustrates the sequence of steps in a TDS measurement, as enumerated below; the sample
is a pellet of lactose monohydrate.
1. A reference time-domain trace is recorded, followed (or preceded) by a measurement of the
sample (Figure 2a).
2. A Fourier transform is applied to the time-domain data, yielding spectral data (Figure 2b).
THz transmission loss in the sample is evident in the difference between the reference and
sample spectra.
Sensors 2019, 19, 4203 3 of 35

Sensors
3. 2019, THz
The 19, 4203
optical
parameters of the sample are calculated from the spectral data of reference 3 ofand
35

sample (Figure 2c) [11–14]. The refractive index is obtained from the phase data; whereas the
absorption coefficient is derived from the amplitude data, taking into account the previously
absorption coefficient is derived from the amplitude data, taking into account the previously
calculated refractive index.
calculated refractive index.
There are two alternative techniques for implementing the requisite variable time delay between
There are two alternative techniques for implementing the requisite variable time delay between
the THz and probe pulses: (1) mechanical stages [15–17], and (2) opto-electronic [18–20] techniques.
the THz and probe pulses: (1) mechanical stages [15–17], and (2) opto-electronic [18–20] techniques.
The mechanical delay is the most widely used approach. Its advantages are excellent reproducibility,
The mechanical delay is the most widely used approach. Its advantages are excellent reproducibility,
low jitter and low noise. Its main disadvantage is its comparatively slow measurement speed. Opto-
low jitter and low noise. Its main disadvantage is its comparatively slow measurement speed.
electronic delay methods require two lasers with synchronized, yet slightly offset pulse repetition
Opto-electronic delay methods require two lasers with synchronized, yet slightly offset pulse repetition
rates. Their main advantage is extremely fast data acquisition. This also leads to their main
rates. Their main advantage is extremely fast data acquisition. This also leads to their main disadvantage
disadvantage which is a reduced SNR (signal-to-noise ratio), caused by shorter time intervals per
which is a reduced SNR (signal-to-noise ratio), caused by shorter time intervals per data point. Indeed,
data point. Indeed, the SNR decreases with faster acquisition rates.
the SNR decreases with faster acquisition rates.
Current-generation THz TDS instruments are compact, robust, flexible, comparable in price to
Current-generation THz TDS instruments are compact, robust, flexible, comparable in price
other types of optical instrumentation (e.g., spectrometers, Raman systems), and suitable for
to other types of optical instrumentation (e.g., spectrometers, Raman systems), and suitable for
installations in an industrial environment.
installations in an industrial environment.

reference phase
a b lactose phase 0

1 1
Amplitude (a.u.)
Signal (a.u.)

Phase
0.1
-5000

-1
0.01
Reference
Lactose reference amplitude
-2 lactose amplitude
1E-3 -10000
0 10 20 30 40 50 0.0 0.5 1.0 1.5 2.0 2.5
Time (ps) Frequency (THz)

c 100 1.9
Lactose
Absorption coefficient (cm-1)

absorption
80 coefficient
refractive
1.8
Refractive index

index
60
1.7
40

1.6
20

0 1.5
0.0 0.5 1.0 1.5 2.0 2.5
Frequency (THz)

(a)Time-domain
Figure2.2.(a)
Figure Time-domaintraces
tracesofofaareference
referencemeasurement
measurementand
andaalactose
lactosesample.
sample.(b)
(b)Spectral
Spectraldata
data
(Fourier transform) of traces in (a). (c) Derived absorption coefficient and refractive index of lactose.
(Fourier transform) of traces in (a). (c) Derived absorption coefficient and refractive index of lactose.
Courtesyof
Courtesy ofM.
M.Naftaly,
Naftaly, NPL,
NPL, UK.
UK.

A frequency-domain spectrometer (FDS) is closely related to TDS in terms of physical processes,

A frequency-domain spectrometer (FDS) is closely related to TDS in terms of physical processes,
technological solutions, and measurement techniques [21–23]. It employs two continuous-wave
technological solutions, and measurement techniques [21–23]. It employs two continuous-wave
lasers with offset wavelengths, dedicated semiconductor-based antennas, commonly referred to as
lasers with offset wavelengths, dedicated semiconductor-based antennas, commonly referred to as
“photomixers”, convert the beat signal into a monochromatic THz wave. Varying the wavelength
“photomixers”, convert the beat signal into a monochromatic THz wave. Varying the wavelength
offset between the two lasers tunes the THz frequency. Unlike TDS, FDS is continuous-wave, with the
offset between the two lasers tunes the THz frequency. Unlike TDS, FDS is continuous-wave, with
high temporal coherence giving rise to standing waves that must be accounted for in measurements.
the high temporal coherence giving rise to standing waves that must be accounted for in
Similarly to TDS, FDS uses coherent detection, yielding information on both THz field amplitude
measurements. Similarly to TDS, FDS uses coherent detection, yielding information on both THz field
and phase. The typical operating range is 0.05–3 THz [24]. The main advantages of FDS are a very
amplitude and phase. The typical operating range is 0.05–3 THz [24]. The main advantages of FDS
are a very high-frequency resolution (<5 MHz), the possibility of measuring at a user-selected, fixed
frequency (or frequency range), and a comparatively low system cost.
Sensors 2019, 19, 4203 4 of 35

high-frequency resolution (<5 MHz), the possibility of measuring at a user-selected, fixed frequency
Sensors 2019, 19, 4203 4 of 35
(or frequency range), and a comparatively low system cost.
Figure 3 illustrates the sequence of steps in an FDS measurement; the sample is a silicon sphere
Figure 3 illustrates the sequence of steps in an FDS measurement; the sample is a silicon sphere
resonator with a whispering-gallery-mode (WGM) resonance [25].
resonator with a whispering-gallery-mode (WGM) resonance [25].
1.1. Akin
Akin toto TDS,
TDS, aa reference
reference measurement
measurement is is recorded,
recorded, followed
followed (or (or preceded)
preceded) by byaameasurement
measurementof of
the
the sample (Figure 3a blue and grey). The phase-sensitive (coherent) detection schemegives
sample (Figure 3a blue and grey). The phase-sensitive (coherent) detection scheme givesrise
rise
to
to phase
phase “fringes”:
“fringes”: thethe detected
detected signal,
signal, i.e.,
i.e., the
the photocurrent
photocurrent measured measured in in the
the receiver
receiver photomixer,
photomixer,
oscillates between negative and positive values as the THz
oscillates between negative and positive values as the THz frequency is scanned. This frequency is scanned. Thiseffect
effectis
is
similar to scanning an interference pattern in frequency [26] and must not be confused with
similar to scanning an interference pattern in frequency [26] and must not be confused with
standing
standing waves
waves thatthat arise
arise due
dueto tomultiple
multiplereflections
reflectionsof ofthe
thebeam.beam. Note
Notethat
thatthethefrequency
frequencystepstepsize
size
in the example of Figure 3 is as small as 1 MHz, providing high-resolution
in the example of Figure 3 is as small as 1 MHz, providing high-resolution spectral measurements. spectral measurements.
2.2. In
In the
thefirst
firstpost-processing
post-processingstep, step,the
theenvelope
envelopespectrum
spectrum ofof thethe phase
phasefringes
fringesis computed
is computed (Figure 3a
(Figure
purple and black). The most straightforward approach is a simple
3a purple and black). The most straightforward approach is a simple identification of the phase identification of the phase
maxima and minima minima [26]; however,
however, thisthis method
method fails
fails ifif the
the linewidth
linewidth of of the
thespectral
spectral feature
feature isis
narrower thanthanthethefringe
fringeperiod.
period. Phase
Phase and and amplitude
amplitude information
information for each
for each frequency
frequency step isstep is
either
either obtained with phase modulation techniques or by applying
obtained with phase modulation techniques or by applying a Hilbert transform to the raw dataa Hilbert transform to the raw
data
[27]. [27].
3.3. The
The envelope
envelope spectrum
spectrum of of the
the sample
sampleis isdivided
dividedby bythat
thatofofthe thereference
referencemeasurement.
measurement.The Thesquare
square
of the
the resulting
resulting ratio
ratio produces
produces the the transmission
transmission spectrum
spectrum (Figure (Figure 3b),
3b), where
where thetheratio
ratioisissquared
squared
because “transmission”
“transmission” refers refers to the intensity, whereas whereas the the envelope
envelope spectra
spectra ofof step
step (2)(2) are
are
proportional
proportional to the electric field of the terahertz
field of the terahertz wave. wave.
Inthe
In theexample
exampleofof Figure
Figure 3, the
3, the WGM WGM resonance
resonance has ahas a Q factor
Q factor of 15,000
of 15,000 and anand
FWHMan FWHM (full
(full width
width
half half maximum)
maximum) linewidthlinewidth
of 42 MHz.of 42The
MHz.
line The line
shape is shape
clearlyisresolved
clearly resolved withsystem,
with an FDS an FDSbut
system,
such
but such a narrow resonance would not be detectable
a narrow resonance would not be detectable at all with a TDS. at all with a TDS.

Reference
a Sample b 1
4 Ref. envelope
Sample envelope FWHM = 42 MHz
3
0.1
Photocurrent (nA)

2
Transmission

1
0.01
0

-1
0.001
-2

-3
1E-4
617.6 617.8 618.0 618.2 618.4 617.6 617.8 618.0 618.2 618.4
Frequency (GHz) Frequency (GHz)

Figure3.3. (a)
Figure (a) Frequency-domain
Frequency-domain signals
signals of
of aareference
reference(grey)
(grey)and
andsample
sample(red)
(red)scans.
scans.The
Thecoherent
coherent
detection scheme gives rise to phase “fringes”, which appear as an oscillating pattern when the THz
detection scheme gives rise to phase “fringes”, which appear as an oscillating pattern when the THz
frequency is varied. A Hilbert transform yields the envelope spectra of reference (black) and sample
frequency is varied. A Hilbert transform yields the envelope spectra of reference (black) and sample
(pink). (b) The sample transmission spectrum (the square of the ratio of the sample and reference
(pink). (b) The sample transmission spectrum (the square of the ratio of the sample and reference
envelope data) reveals the whispering-gallery-mode resonance with an FWHM linewidth of 42 MHz.
envelope data) reveals the whispering-gallery-mode resonance with an FWHM linewidth of 42 MHz.
Courtesy of Dominik Vogt, University of Auckland, New Zealand.
Courtesy of Dominik Vogt, University of Auckland, New Zealand.

Similarly to optical measurements in the visible and NIR, it is possible to perform THz
Similarly to optical measurements in the visible and NIR, it is possible to perform THz
measurements by employing any suitable combination of a source and detector. The current range
measurements by employing any suitable combination of a source and detector. The current range of
of technologies has been surveyed and assessed in several recent reviews [9,28–30]. The challenge
technologies has been surveyed and assessed in several recent reviews [9,28–30]. The challenge lies
lies in the inherently low intensity of THz sources and the low sensitivity of most THz detectors;
in the inherently low intensity of THz sources and the low sensitivity of most THz detectors;
nevertheless, in the low part of the THz range (<1 THz) useful electronic sources and detectors are
nevertheless, in the low part of the THz range (<1 THz) useful electronic sources and detectors are
now available. Multipixel detector arrays and video-rate THz cameras are also available from several
manufacturers [31] and may find industrial use in the future.
The following sections offer a review of current industrial applications of THz technologies,
describing demonstrated and deployed systems. Some of these applications can in principle be
Sensors 2019, 19, 4203 5 of 35

now available. Multipixel detector arrays and video-rate THz cameras are also available from several
manufacturers [31–34] and may find industrial use in the future.
The following sections offer a review of current industrial applications of THz technologies,
describing demonstrated and deployed systems. Some of these applications can in principle be realized
using different types of THz instrumentation (for example TDS, FDS, or single-frequency electronic
emitters/detectors). However, the remit of this paper is to present existing implementations.

3. Polymers
Polymers are transparent or semi-transparent for terahertz waves and thus among the most
important materials for implementations of THz technology [35]. A large body of literature exists
reporting THz measurements of polymer properties and behavior. The fields of application range from
the characterization of macroscopic and morphological properties [35,36] to the non-destructive testing
of polymers and polymer products [37,38]. Polymers are also used as materials for THz optics [39,40]
and as a material for THz emitters and detectors [41]. In particular, two of the most widely used
polymers, polypropylene and polyethylene, are among the most transparent materials to THz waves.
The materials covered in this section include pure polymers as well as polymer foams, polymer
composites and adhesives.

3.1. Polymers and Polymer Components

One of the most important industry market drivers for THz technology is currently the field of
non-destructive testing of plastics and plastic products. Plastics have become an indispensable part
of our everyday lives. The global market for plastics was 263 million tons in 2016 and the projected
growth is 3–5% per year (https://www.ceresana.com; https://www.farbeundlack.de).
In line with the market size and growth, there is an enormous demand for test equipment, both to
ensure acceptable product quality and to reduce material usage and wastage. With material savings
in production of approx. 1%, even a test instrument of about €100 k will amortize itself within 1–2
years. Currently, established and cost-effective nondestructive testing technologies include X-ray
scanners, ultrasonic sensors and microwave transceivers. However, THz technology has now reached
the cost levels, size, robustness, measurement speeds and discrimination quality of previous gold
standards [42–44].
THz technology offers a number of advantages enabling it to partially replace other established
techniques in applications such as inline measurements in plastic extrusion processes [43]. Moreover,
it wins new fields of application such as the inspection of multi-layer paint coatings in the automotive
industry as described in Section 4. Unlike X-ray scanners, THz systems do not emit ionizing radiation
and are considered biologically innocuous, thus saving costs entailed in extensive radiation protection
measures, safety officers and compliance documentation. In addition to being nondestructive, THz
systems operate contact-free and do not require a contact medium, as do ultrasonic sensors. Crucially,
THz waves can penetrate polymers, even in cases where the material is optically opaque, ultrasound
is strongly damped, and near-infrared waves are scattered. Compared to microwaves, THz waves
achieve higher spatial resolution, so that submillimeter-sized defects can be detected or micrometer-thin
polymer coatings can be analyzed.
Due to their good optical penetration through most polymers, THz systems are particularly
suitable for the inspection of plastic components and the detection of defects [45,46]. Figure 4 shows a
step wedge made of polyamide. Two internal air pockets approx. 5 mm and 10 mm in size are invisible
to the naked eye or an optical system but are revealed using THz imaging. The sample was raster
scanned over an area of 100 mm × 40 mm using a commercial TDS instrument [15,47]. The thickness
of the individual sections was 1 mm, 2 mm, and 4 mm, red and blue colors indicate high and low
transmission respectively. The air bubbles appear “blue” because THz waves are strongly scattered at
the air-polymer interfaces, reducing transmission.
 Sensors 2019, 19, 4203 6 of 35

Sensors2019,
Sensors 19,4203
2019,19, 4203 66ofof35
35

Figure 4. Plastic step wedge made of polyamide. Air pockets are made visible using THz TDS.
Figure 4. Plastic step wedge made of polyamide. Air pockets are made visible using THz TDS. Adapted
Adapted from [15], courtesy of TOPTICA Photonics AG.
Figure 4. Plastic
from [15], courtesystep wedge made
of TOPTICA of polyamide.
Photonics AG. Air pockets are made visible using THz TDS.
Adapted from [15], courtesy of TOPTICA Photonics AG.
THz techniques can reveal not only voids but also thermal or mechanical stress, deformations
THz techniques can reveal not only voids but also thermal or mechanical stress, deformations
and ageing processes in polymer materials. The also
effects of these or processes can be seen, for example,
and THz
ageingtechniques
processescan reveal
in polymer not only voids
materials. Thebuteffects ofthermal mechanical
these processes can be stress, deformations
seen, for example, by
and by observing
ageingchanges changes
processes in
in polymerthe degree of crystallinity [45,46] or changes in the glass transition
observing in the degree of materials. The
crystallinity effects
[48,49] of thesein
or changes processes
the glass can be seen,
transition for example,
temperature [50].
by temperature
observing [47].
changes Figure 5a shows the THz absorption coefficient of polybutylene terephtalate (PBT)
Figure 5a shows the THzin the degree
absorption of crystallinity
coefficient [45,46]terephtalate
of polybutylene or changes in with
(PBT) the glass transition
different degrees
with
temperaturedifferent
[47].Asdegrees
Figure of crystallinity.
5a shows the THztheAs crystallinity
absorption increases,
coefficient the absorption
of polybutylene peak at 2.4 THz
of crystallinity. crystallinity increases, absorption peak at 2.4 THz becomesterephtalate (PBT)
more pronounced.
withbecomes more
different pronounced.
degrees Figure
of refractive
crystallinity.5b shows the THz
Asofcrystallinity refractive
increases, index of polyoxymethylene
the absorption peak (POM)
at 2.4 THz
Figure 5b shows the THz index polyoxymethylene (POM) as a function of temperature.
as
becomes a function of temperature. The glass transition temperature can be determined from the
The glassmore pronounced.
transition Figure
temperature 5bbe
can shows the THzfrom
determined refractive index of polyoxymethylene
the intersection of the two straight (POM)
dashed
as intersection
a function of the two
of temperature. straight dashed lines.
The properties,
glass transitionIn addition
temperatureto static polymer
can be properties,
determined from dynamic
the
lines. In addition to static polymer dynamic behavior, such as crystallization [49] and
behavior,
intersection such
of the as crystallization [46] and polymerization [35], can also be observed and tracked.
polymerization [38],two
canstraight dashed lines.
also be observed In addition to static polymer properties, dynamic
and tracked.
behavior, such as crystallization [46] and polymerization [35], can also be observed and tracked.
a 150 b 1.71
Absorption coefficient (cm-1)

a 150
Crystallinity: b 1.71
Refractive index @ 1 THz

52%
Absorption coefficient (cm-1)

44%
Crystallinity: 1.70
25%
Refractive index @ 1 THz

100 52%
44% 1.70
100 25%
1.69

50 1.69
1.68
50
Tg = 203 K
1.68
0 1.67
1.5 2.0 2.5 100 Tg 150
= 203 K 200 250 300
0 Frequency (THz) 1.67 Temperature (K)
1.5 2.0 2.5 100 150 200 250 300

5. (a) Frequencyspectrum
(THz) Temperature(PBT)
(K) with different
Figure
Figure 5. Absorption
(a) Absorption of three
spectrum ofsamples of polybuthylene
three samples terephathalate
of polybuthylene terephathalate (PBT) with
degrees of crystallinity.
different (b) Temperature-dependent
degrees of crystallinity. refractive index
(b) Temperature-dependent of polyoxymethyllene
refractive (POM).
index of polyoxymethyllene
Figure 5. (a) Absorption spectrum of three samples of polybuthylene terephathalate (PBT) with
The glass transition temperature can be determined from the intersection of the
(POM). The glass transition temperature can be determined from the intersection of the twotwo linear slopes.
linear
different
Figures degrees from
adapted of crystallinity.
[35], (b) of
courtesy Temperature-dependent
S. Wietzke, Continental refractive
AG. index of polyoxymethyllene
slopes. Figures adapted from [32], courtesy of S. Wietzke, Continental AG.
(POM). The glass transition temperature can be determined from the intersection of the two linear
3.2. slopes.
Composite Materials
Figures adapted from [32], courtesy of S. Wietzke, Continental AG.
3.2. Composite Materials
Most plastic components are not made of pure polymers but of composite materials. During the
3.2. Composite Materials
Most plastic components are not made of pure polymers but of composite materials. During the
compounding process, a polymer material is adapted to the specific application by incorporating
compounding
Most plastic process, a polymer material is adapted to but
the ofspecific application by During
incorporating
additives. These components are not made
include, for example, ofwith
filling pure polymers
chalk, reinforcing composite
with glassmaterials.
fibers or dyeing withthe
additives.
compounding These include,
process, for
a polymer example, filling with chalk, reinforcing with glass fibers or dyeing with
colorants. For the plastics industry, material is adapted
it is important to verifyto the additive
specific content
application
[51], by
theincorporating
homogeneous
colorants.
additives. TheseForinclude,
the plastics
forthe industry, it with
example, is important to verify theglass
additive content [48], the
distribution of fillers, i.e., degreefilling
of dispersion chalk, reinforcing
[52], the moisture with content fibers
[53],orasdyeing
well aswith
the
homogeneousthe
colorants. distributionindustry,
of fillers, i.e.,isthe degree ofto dispersion [49],additive
the moisture content
[48], [50],
the as
orientationForof fibersplastics
[54], particles [55]itand important
molecular chainsverify
[56].theThe properties content
of a composite
well as
homogeneous the orientation
distribution of fibers
of by
fillers, [51], particles
i.e., the its
degree [52] and molecular chains [53]. The properties asof a
material can be determined measuring THz of dispersion
refractive index[49],
andtheabsorption
moisture content [50],[35].
coefficient
composite
well asexample material
the orientation can be determined
of fibers [51],imaging, by measuring
particles Figure
[52] and its
molecular THz refractive
chains [53]. index
The and absorption
As an of moisture-contrast 6 shows a photograph and theproperties of a
corresponding
coefficient
composite [32]. Ascan
material an example of moisture-contrast
be determined by measuring imaging,THzFigure
itscontaining 6 shows a photograph
refractive and the
THz TDS transmission image of a wood-plastic composite 60% byindex
weight and absorption
of wood fibers,
corresponding
coefficient AsTHz
[32].immersed TDS transmission
an example of for image of a wood-plastic
moisture-contrast composite containing 60%and
by weight
that had been in water four days. The imaging,
frequencyFigure interval6 shows
betweena photograph
0.4 and 0.5 THz the
was
of wood
corresponding fibers,
THz that had been immersed in water for four days. The frequency interval between 0.4
used to generate theTDS transmission
image, as a compromiseimage between
of a wood-plastic
poor spatial composite
resolutioncontaining
at lower 60% by weight
frequencies and
of wood fibers, that had been immersed in water for four days. The frequency interval between 0.4 at
and 0.5 THz was used to generate the image, as a compromise between poor spatial resolution
and 0.5 THz was used to generate the image, as a compromise between poor spatial resolution at
Sensors
Sensors2019,
2019,19,
19,4203
4203 7 7ofof3535
Sensors 2019, 19, 4203 7 of 35
lower frequencies and strong absorption at higher frequencies. The presence of water is revealed in
lower
areas offrequencies
strong absorptionand
reduced THz strongfrequencies.
absorption
at transmission
higher becauseat
Thehigher frequencies.
presence
THz waves of water
are The presenceinby
is revealed
strongly absorbed ofareas
waterofisreduced
water. revealed in
THz
areas of reduced
transmission THz transmission
because THz waves arebecause
stronglyTHz waves are
absorbed strongly absorbed by water.
by water.

Figure 6. Photograph (left) and THz transmission image (right) of a wood/plastic composite material
Figure
Figure
after 6.6.Photograph
Photograph
immersion (left)
(left)
in water. and
From THz
and[32], transmission
THzcourtesy of S. image
transmission image(right)
(right)
Wietzke, of
ofaawood/plastic
wood/plastic
Continental AG. composite
compositematerial
material
afterimmersion
after immersionin inwater.
water. From
From[32],
[35],courtesy
courtesy of
of S.
S. Wietzke,
Wietzke, Continental
Continental AG.
AG.
3.3. Polymer Foams
3.3.Polymer
3.3. PolymerFoams Foams
With a global market volume of 19.1 million tons in 2013, polymer foams are one of the most
With aa global
With global market volume volume of 19.1 19.1 million
million tonstons in 2013,
2013, polymer
polymerfoams foams are one one of of the
the most
most
important classes ofmarketpolymers [54]. of They have excellentinthermal and mechanicalare properties at low
important
important classes of polymers [57]. They have excellent thermal and mechanical properties at low
weight andclasses
are used of polymers
to increase [54]. Theyefficiency
energy have excellent in thethermal and mechanical
construction, automotive, properties
aerospace at and
low
weight and
weight and are used usedtotoincrease energy efficiency in the construction, automotive, aerospace and rotor
rotor blade are industries. increase
In the past,energy efficiency
testing of these in the construction,
materials was automotive,
time-consuming, aerospace and
partially
blade
rotor industries.
blade In
industries.the past,
In testing
the of
past, these
testing materials
of these was time-consuming,
materials was partially
time-consuming, destructive and
partially
destructive and only possible offline. While polymer foams strongly damp ultrasound, they have
only possible
destructive offline.
and only While polymer foams strongly damp ultrasound, they have good transparency to
good transparency topossible
THz waves. offline. While THz
Today, polymer systemsfoams arestrongly
used indamp ultrasound,
the extrusion they have
industry to
THz waves.
good Today, THz
transparency to THzsystemswaves.are used
Today, in the extrusion industry to determine the wallindustry
thicknessto of
determine the wall thickness of polymer foamTHz pipessystems
[40]. are used in the extrusion
polymer
determine foam pipes
the wall [43].
thickness
The cellular structure andofthepolymer foam
effective pipesof[40].
density the foam are decisive quality parameters that
Thecellular
The cellularstructure
structureand andthe theeffective
effective density
density of of the
the foam
foam areare decisive
decisive quality
quality parameters
parameters that
can be determined by THz sensing inline during the production process [55]. The foam cells actthat as
can be
can be determined
determinedby by THz
THz sensing
sensing inline
inline during
during the the production
production process
process [55].[58]. The
Thefoam
foamcellscellsact
actasas
scattering centers for THz waves so that the cell size can be determined by investigating the THz loss
scattering centers
scattering centers for THz waves so that the cell size can be determined by investigating the THz loss
coefficient (which for THz waves
combines so that thefrom
contributions cell size can be determined
absorption and scattering).by investigating
In addition, thethere
THz is loss
a
coefficient
coefficient (which combines
(which combines contributions
contributions from
from absorption and scattering). In addition, there is a linear
linear relationship between the bulk density andabsorption
the effective andTHzscattering).
refractive In index
addition, there the
(because is a
relationship
linear between
relationship the bulk
between thedensity and theand
bulk density effective THz refractive
the effective THz in index (because
refractive index the effective
(because
effective refractive index is proportional to the thickness of material the beam path). Figure the 7a
refractive
effective index
refractive is proportional
index is to the
proportional thickness
to the of material
thickness of in the
material beamin path).
the beamFigure
path). 7a shows 7a
Figure an
shows an example of this relationship for a frequency of 0.5 THz as determined by TDS. The refractive
example
shows of this relationship
an example of this for a frequency
relationship of 0.5 THz
for adensity.
frequency asTHz
of 0.5 determined by TDS.
as determined byThe
TDS.refractive index
The refractive
index increases linearly with the effective
increases
index linearly with the effective density.
Inincreases
addition,linearly
defectswith in the thefoam
effective density.
structure that can influence its mechanical properties can be
In addition,
In addition, defects
defects in in the
the foam
foam structure
structure that that cancan influence
influence its its mechanical
mechanical properties
properties can be be
detected with THz imaging techniques. Figure 7b shows a 20 mm thick polyvinyl chloridecan foam
detected with
detected with THzTHz imaging
imaging techniques.
techniques. FigureFigure7b 7b showsaa20 20 mm thickthick polyvinyl chloride chloride foam
component as used in a rotor blade. In the overlaid shows 100 mm × 60mm mm THzpolyvinyltransmission imagefoam the
component
component as used in a rotor blade. In the overlaid 100 mm × 60 mm THz transmission image the the
defects in theassample
used in area revealed
rotor blade. In theand
as green overlaid
red dots. 100 mm × 60 mm THz transmission image
defectsin
defects inthe
thesample
sample are are revealed
revealed as as green
green andand redred dots.
dots.

1.10
1.10 a
THzTHz

1.08 a
@ 0.5

1.08
@ 0.5

1.06 Linear fit

index

PET
1.06 PVC
Linear fit
index

PE-X
1.04 PET
PUR-Hart
Refractive

PVC
PUR-Weich
PE-X
1.04 XPS
PUR-Hart
Refractive

PE
PUR-Weich
1.02 PUR-PES
XPS
PUR-Ether
PE
1.02 PP
PUR-PES
PUR-Ether
1.00 PP
0 50 100 150 200
1.00
0 50
Effective 3)
100 (kg/m150
density 200
Effective density (kg/m3)
Figure7.7.(a)(a)Relation
Figure Relationbetween
betweenrefractive
refractiveindex
indexand
andeffective
effectivedensity
densityfor
forvarious
variouspolymer
polymerfoams.
foams.
Figure
Adapted
Adapted 7.from
(a) Relation
from [49], between
[52],courtesy
courtesyofofM.refractive
M. index and effective
Mayr,Sűddeutsches
Mayr, Sűddeutsches density for (b)
Kunststoffzentrum.
Kunststoffzentrum. various polymer
(b)Photograph
Photograph foams.
and
and THz
THz
Adapted
transmissionfrom
transmission [49], courtesy
image
image (100mm
(100 mm of M.
× ×60 Mayr,
60mm) Sűddeutsches
mm)ofof Kunststoffzentrum.
a a2020mm-thick
mm-thick piecepolyvinyl
piece (b) Photograph
polyvinylchloride
chloride and
foam,asas
foam, THz
used
used inin
rotor blades.
transmission Courtesy
image of
(100 TOPTICA
mm × 60 Photonics
mm)
rotor blades. Courtesy of TOPTICA Photonics AG. of a 20AG.
mm-thick piece polyvinyl chloride foam, as used in
rotor blades. Courtesy of TOPTICA Photonics AG.
 Sensors 2019, 19, 4203 8 of 35
Sensors 2019, 19, 4203 8 of 35
3.4. Adhesives
Adhesives are another important class of polymers widely used in manufacturing and
3.4. Adhesives
construction processes. However, non-destructive testing of adhesive joints is still a challenge. It has
beenAdhesives are another
demonstrated important
that THz class ofcan
instruments polymers
be used widely used in manufacturing
to determine the thickness and construction
of adhesive joints
processes. However, non-destructive testing of adhesive joints is still a challenge.
and the distribution of adhesives between plastic components [56]. As with polymers, moisture can It has been
demonstrated
also be detectedthatinTHz instruments
an adhesive [57].can be used tothe
In addition, determine the thickness
curing behavior of adhesive
of adhesives joints
can be and the
investigated
distribution of adhesives
[58,59]. In the between plastic
case of a 2-component glue, components
polymerization[59]. is
Asinitiated
with polymers,
when two moisture
monomers, can also be
a binder
detected in an adhesive [60]. In addition, the curing behavior of adhesives can
and hardener, are mixed together; whereas in a light-curing glue the polymerization reaction is be investigated [61,62].
Indriven
the casebyofabsorption
a 2-component glue,ofpolymerization
of light is initiated when
appropriate wavelengths. two 8monomers,
Figure shows thea binder
variationandof
hardener, are mixed together; whereas in a light-curing glue the polymerization
transmission of pulsed THz radiation through a 2-component and a UV-curing adhesive during the reaction is driven
bycuring
absorption
process.of The
lightintensity
of appropriate wavelengths.
of the THz Figure 8 with
pulses is detected shows the variation
a Schottky diodeof[60].
transmission
The dropsofin
pulsed THzafter
intensity radiation through
2 min and 4 min a 2-component
respectively mark and athe
UV-curing
beginningadhesive during the curing
of the polymerization. process.
In both cases
The intensity of the THz pulses is detected with a Schottky diode [63]. The drops in
the polymerization reaction is exothermal: the adhesive initially heats up and softens, absorbing the intensity after 2 min
and
THz 4 min
waves respectively markDuring
more strongly. the beginning
curing, the of the polymerization.
refractive In both
index increases andcases the polymerization
the absorption decreases.
reaction is exothermal:
The increase the adhesive
of the effective initially
refractive indexheats up and
is linked softens,
to the absorbing
increase the THz
in polymer waves more
concentration and
strongly.
optical density. The decrease in the effective absorption coefficient correlates to the decrease in of
During curing, the refractive index increases and the absorption decreases. The increase the
the effective refractive
concentration index is linked
of the monomer becausetothe themonomer
increase material
in polymer has concentration
higher absorption and than
optical
thedensity.
polymer
The decrease
form [58]. in the effective absorption coefficient correlates to the decrease in the concentration of the
monomer because the monomer material has higher absorption than the polymer form [61].

0.5
Transmission

0.4

2-component, 1.9 mm
light-curing, 2.2 mm
0.3
0 5 10 15 20
Time (min)

Figure8. 8.THz
Figure THztransmission
transmissionthrough
througha a2-component
2-componentand
anda alight-curing
light-curingadhesive
adhesiveduring
duringthe
thecuring
curing
process. Courtesy of TOPTICA Photonics
process. Courtesy of TOPTICA Photonics AG. AG.

4. Paint and Coatings

4. Paint and Coatings
High-tech surfaces in the automotive or aircraft industry must withstand harsh environmental
High-tech surfaces in the automotive or aircraft industry must withstand harsh environmental
conditions such as heat, frost, and water, and impact from e.g. sand and stone-chips. In order to
conditions such as heat, frost, and water, and impact from e.g. sand and stone-chips. In order to
protect these surfaces and to render them visually attractive, they are encased in multilayer functional
protect these surfaces and to render them visually attractive, they are encased in multilayer functional
coatings and lacquers in an elaborate process. In the automotive sector, for example, typically three or
coatings and lacquers in an elaborate process. In the automotive sector, for example, typically three
four different layers are applied to the car body [64].
or four different layers are applied to the car body [61].
In order to ensure that the layers remain functional throughout the lifetime of the product and
In order to ensure that the layers remain functional throughout the lifetime of the product and
to minimize expenditure of the coating materials, the coating thickness must be carefully monitored.
to minimize expenditure of the coating materials, the coating thickness must be carefully monitored.
The global market for thickness measurement instruments registers a continuous substantial growth,
The global market for thickness measurement instruments registers a continuous substantial growth,
which is due to various factors, including strict government regulations to ensure quality, growing
which is due to various factors, including strict government regulations to ensure quality, growing
safety concerns among buyers, the drive for cost optimization, and a rising trend for technology
safety concerns among buyers, the drive for cost optimization, and a rising trend for technology
miniaturization (www.marketsandmarkets.com). The film thickness measurement market is expected
miniaturization (www.marketsandmarkets.com). The film thickness measurement market is
to reach $520 M by 2023 at a CAGR of 4.9% (www.IGOS.de). An overview of established thickness
expected to reach $520 M by 2023 at a CAGR of 4.9% (www.IGOS.de). An overview of established
measurements techniques can be found in [44].
thickness measurements techniques can be found in [41].
Complex multi-layer coatings push established measurement technologies to their limits.
Complex multi-layer coatings push established measurement technologies to their limits.
Moreover, some of the conventional techniques, such as magnetic gauges or eddy-current measurements,
Moreover, some of the conventional techniques, such as magnetic gauges or eddy-current
fail in the case of non-metallic substrates, such as glass-fiber reinforced plastics. Until recently it has
measurements, fail in the case of non-metallic substrates, such as glass-fiber reinforced plastics. Until
only been possible to test the paint on these substrates using destructive methods in which the surfaces
recently it has only been possible to test the paint on these substrates using destructive methods in
are destroyed with a wedge cut [44].
which the surfaces are destroyed with a wedge cut [41].
 Sensors 2019, 19, 4203 9 of 35

THz TDS provides a nondestructive solution for monitoring multilayer coatings on

Sensors 2019, 19, 4203
any
9 of 35
substrate,
Sensors 2019, 19,enabling
4203 each layer thickness to be measured individually [60–63]. Today, such systems 9 of 35
are already deployed in industrial installations [41,62–71]. Figure 9 depicts a substrate with two
THz TDS
coating TDSprovides
layers, provides
together a nondestructive
a nondestructive
with the path ofsolutionsolution for beam.
for monitoring
a reflected THz monitoringAkin tomultilayer
multilayer coatings
coatingsmeasurements
thickness on any
on any substrate, with
substrate,
enabling enabling
each layer each layer
thickness tothickness
be measuredto be measured
individually individually
[63–66]. [60–63].
Today,
ultrasound, a THz beam is partially reflected from each of the interfaces. By recording the time-of- such Today,
systems such
are systems
already
are already
deployed
flight in deployed
echo industrial
signals, the in industrial of
installations
thickness installations
[44,65–74].
each individual [41,62–71].
Figure 9 depicts
layer Figure
can 9 depicts
abesubstrate
resolved. a substrate
with
By two coating
using with
a beam two of
layers,
spot
coating
together layers,
less than with
2 mm together
the path of
diameter withaandthea path
reflected of a reflected
THz
spectrally beam.
broad Akin THzpulse,
THz beam.
to it Akin
thickness to thickness
measurements
is possible measurements
to measure with with
ultrasound,
layer thicknesses
ultrasound,
a THz beam a THz
is beam
partially is partially
reflected fromreflected
each offrom
the each of
interfaces. the interfaces.
By By
recording
down to 5–10 μm [67,72]. A spectrally broad THz pulse is necessary because the duration of a pulse recording
the the
time-of-flight time-of-
echo
flight
signals,echo
the signals,
is inversely thickness
related thetothickness
of each of each
individual
its spectral individual
layer
width. can be
Since layer
theresolved. canBy
measurement beusing
resolved. By spot
a beam
technique using aless
ofon
relies beamthanspot
2 mm
distinguishing of
less than 2spaced
diameter
closely mma diameter
and spectrally and
echo signals, aitspectrally
broad THz pulse,
requires abroad
it isTHz
short pulse,
possible
(i.e., toitmeasure
broadband)is possible to measure
THzlayer thicknesses
pulse. layer thicknesses
down to 5–10
down to 5–10 μm [67,72]. A spectrally broad THz pulse is necessary
µm [70,75]. A spectrally broad THz pulse is necessary because the duration of a pulse is inversely because the duration of a pulse
is inversely
related to itsrelated
spectraltowidth.
its spectral width.
Since the Since the measurement
measurement technique relies technique relies on distinguishing
on distinguishing closely spaced
closely spacedit echo
echo signals, signals,
requires it requires
a short a short (i.e.,
(i.e., broadband) THz broadband)
pulse. THz pulse.

Figure 9. Schematic drawing of a substrate with two coating layers and the reflected THz beam path.

Figure
Figure 1 showsdrawing
9. Schematic the result of a THz
of a substrate reflection
with two coating measurement
layers and theof a three-layer
reflected THz beam coating
path. with
thicknesses of 84.6 μm
Figure 9. Schematic (top),of40.6
drawing μm (center),
a substrate with two and 48.3 layers
coating μm (bottom) on a carbon-fiber-reinforced
and the reflected THz beam path.
Figure
plastic 10 shows
substrate the result
(CFRP). Theofarrows
a THz reflection
representmeasurement
the echo pulses of afrom
three-layer
the four coating with thicknesses
interfaces, which are in
of close
84.6
Figure
µm (top),
1 shows
proximity 40.6 µm
but can (center),
the still
result and 48.3 (bottom)
of a THzinreflection
be separated µm the time-domain on a carbon-fiber-reinforced
measurement
trace. For of even
a three-layer plastic
coating
thinner layers, substrate
with
the echoes
(CFRP).
thicknessesThe arrows
of 84.6 represent
μm (top), the
40.6echo
μm pulses from
(center), andthe four
48.3 interfaces,
μm (bottom) which
begin to overlap and can no longer be resolved visually, and sophisticated evaluation algorithms on aare in close proximity
carbon-fiber-reinforced butare
can still
plastic be separated
substrate in
(CFRP). the time-domain
The arrows trace.
represent For
the even
echo thinner
pulses layers,
from the
the echoes
four
required to calculate the layer thickness from the raw data [72–74]. In particular, algorithms involving begin
interfaces, to overlap
which areand
in
can
closenoproximity
longer bebut
mathematical resolved
models visually,
can still
of the and sophisticated
be separated
transfer asevaluation
in the time-domain
function, algorithms
well as trace. For even
“stochastic are required
thinner
differential to calculate
layers, the echoes
evolution” the
fitting
layer
routines [72], allow assessing layers on the order of 10–20 μm in a multilayer stack. Some ofofthe
begin thickness
to overlap from
and the
canraw
no data
longer [75–77].
be In
resolvedparticular,
visually, algorithms
and involving
sophisticated mathematical
evaluation models
algorithms are
theanalysis
transfer
required function,make
to routines
calculate as well
the layer
useasthickness
“stochastic
of fittingfrom differential
the raw data
algorithms evolution”
in [72–74].
the fitting routinesalgorithms
In particular,
frequency-domain, [75],
ratherallow
than assessing
involving
the time-
layers on the
mathematical
domain order
since modelsof 10–20
these of the
thin µm
layers inmanifest
transfera multilayer
function, stack.
as well
themselves Some
as as of the analysis
“stochastic
interference routines
differential
signatures make use of function
evolution”
in the transfer fitting
fitting
algorithms
routines
(Figure 11). in
[72], the
allowfrequency-domain,
assessing layers rather
on the than
order the
of time-domain
10–20 μm in asince these
multilayer thin layers
stack. Some manifest
of the
themselves as interference signatures in the transfer function (Figure
analysis routines make use of fitting algorithms in the frequency-domain, rather than the time- 11).
domain since these thin layers manifest themselves as interference signatures in the transfer function
(Figure 11). 150
3 layers on CFRP
100 Top: 84.6 µm
Center: 40.6 µm
150 Bottom: 48.3 µm
Signal (nA)

50
3 layers on CFRP
100 Top: 84.6 µm
0 Center: 40.6 µm
Bottom: 48.3 µm
Signal (nA)

50
-50

0
-100
10 20 30 40 50
-50 Time (ps)

Figure
Figure AA
10.10. THz
THzpulse trace
pulse reflected
-100
trace from
reflected froma three-layer coating
a three-layer onon
coating a carbon-fiber-reinforced plastic
10 20 30 40 50 a carbon-fiber-reinforced plastic
substrate. The echoes from the four interfaces are marked with arrows. Thicknesses of the individual
substrate. The echoes from the four interfacesTimeare (ps)
marked with arrows. Thicknesses of the individual
layers are indicated. Courtesy of TOPTICA Photonics AG.
layers are indicated. Courtesy of TOPTICA Photonics AG.
Figure 10. A THz pulse trace reflected from a three-layer coating on a carbon-fiber-reinforced plastic
THz systems have several advantages compared with other technologies: they operate contact-free,
substrate. The echoes from the four interfaces are marked with arrows. Thicknesses of the individual
whereas ultrasonic and eddy current sensors require physical contact with the sample. THz instruments
layers are indicated. Courtesy of TOPTICA Photonics AG.
can, therefore, measure freshly deposited (“wet”) coatings. They can also provide reproducibility
below 1% [44]. In comparison, the reproducibility with magnetic induction sensors is approx. 10%,
which is insufficient for accurate measurements. The calibration effort for ultrasonic systems is high
Sensors 2019, 19, 4203 10 of 35

1.0
3 layers on metal:
Top: 32.1 µm
Sensors 2019, 19, 4203 0.8 Center: 23.3 µm 10 of 35
Bottom: 33.1 µm

Signal (a.u.)
0.6

but is negligible
Sensors for THz instruments [67]. Some coatings are opaque to optical and infrared beams,
2019, 19, 4203 10 of 35
0.4
whereas THz waves penetrate most types of paints and varnishes [44,75].
0.2
1.0
3 layers on metal:
0.0 Top: 32.1 µm
0
0.8 1 2 Center: 3 4
23.3 µm
Frequency Bottom:
(THz) 33.1 µm

Signal (a.u.)
0.6
Figure 11. The frequency-domain representation of a THz pulse reflected from a three-layer coating
0.4
on metal. Thicknesses of the individual layers are indicated. The layers give rise to interference
signatures in the transfer function. Courtesy of Helmut Fischer GmbH.
0.2

THz systems have several advantages 0.0

compared with other technologies: they operate contact-
free, whereas ultrasonic and eddy current sensors 0 1 2require 3physical4 contact with the sample. THz
Frequency (THz)
instruments can, therefore, measure freshly deposited (“wet”) coatings. They can also provide
reproducibility
Figure 11. The below 1% [41]. In comparison,
frequency-domain representation theofreproducibility
a THz pulse reflected withfrom
magnetic induction
a three-layer coatingsensors
on is
Figure 11. The frequency-domain representation of a THz pulse reflected from a three-layer coating
approx. 10%,
metal. which isofinsufficient
Thicknesses the individual forlayers
accurate measurements.
are indicated. The layersThegivecalibration effort signatures
rise to interference for ultrasonic
on metal. Thicknesses of the individual layers are indicated. The layers give rise to interference
systems is transfer
in the high but is negligible
function. CourtesyforofTHz instruments
Helmut Fischer GmbH.
signatures in the transfer function. Courtesy of Helmut Fischer GmbH.
[64]. Some coatings are opaque to optical and
infrared beams, whereas THz waves penetrate most types of paints and varnishes [41,72].
In order
In order forfor
THz THz measurement systems to reliably measuremeasure
thin layersthin
in anlayers
industrial environment,
THz systems havemeasurement
several advantages systems to reliably
compared with other technologies: theyin operate
an industrial
contact-
solutions
environment, are needed
solutions to integrate
are needed this
to technology
integrate this into production
technology into processes.
production On the
processes. one
On hand,
the one a
free, whereas ultrasonic and eddy current sensors require physical contact with the sample. THz
high signal-to-noise
hand, a high signal-to-noiseratio and a large andTHz bandwidth are required in order in to resolve layers oflayers
a few
instruments can, therefore,ratio measure a large
freshly THz bandwidth
deposited are required
(“wet”) coatings. orderThey to can resolve
also provide
micrometer
ofreproducibilitythicknesses
a few micrometer at high
thicknesses measurement
at comparison,
high measurement speeds [78,79]. On the
speeds [75,76]. other hand,
Onmagnetic the system
the otherinduction must
hand, thesensorssystem be
below 1% [41]. In the reproducibility with is
cost-effective
must and robust [80].
be cost-effective
approx. 10%, which and robust [77].for accurate measurements. The calibration effort for ultrasonic
is insufficient
In addition,
In addition, clever
clever algorithms are are needed
needed to calculate layer
layer thickness
thickness information
information from the the THz
THz
systems is high but is algorithms
negligible for THz to calculate
instruments [64]. Some coatings are opaquefrom to optical and
data
data in real-time
in real-time [20] and under industrial conditions. Vibrations of automotive parts can interfere
infrared beams, [20]whereasand under
THz wavesindustrial conditions.
penetrate most typesVibrations
of paintsof automotive
and varnishes parts can interfere
[41,72].
with the THz
with the THz signal
signaland and impedeororobstruct obstruct thickness measurements. However, suchsuch vibrations
In order for THzimpede measurement systems thickness measurements.
to reliably measureHowever,
thin layers invibrations
an industrialcan
can be measured
beenvironment,
measured using using
opticaloptical
methodsmethods
and and
the THzthe THz
pulse pulse
traces traces
can becan be
corrected corrected
in in
software software
[75]. [78].
Water
solutions are needed to integrate this technology into production processes. On the one
Water
vapor vapor atmosphere
in the atmosphere has multiple absorption at lines at THz frequencies (see 8)Section 8) [81],
hand,inathehigh signal-to-noise has multiple
ratio and absorption
a large THz lines THz frequencies
bandwidth are required(seeinSection
order to [78],
resolveleading
layers
leading
to to artifacts
artifacts in the THzin the THz
pulse pulse
that can that
mask canrelevant
mask relevant
thickness thickness information
information [39]. [42].artifacts
These These artifacts
can be
of a few micrometer thicknesses at high measurement speeds [75,76]. On the other hand, the system
can be minimized
minimized either either numerically
numerically [79,80] [82,83]
or by or by purging
purging the the THz
THz path path
with with
dry dryor
air airother
or other
gases gases
[78]. [81].
must be cost-effective and robust [77].
Figure 12
Figure 12 shows
shows an example
an algorithms
example of of an implementation
anneeded
implementation of a
of a THz
THz TDS
TDS system
system in
in the automotive
the from
automotive
In addition, clever are to calculate layer thickness information the THz
industry
industry [44].
[41]. The
The THz
THz reflection
reflection head
head is
is aligned
aligned to
to the
the car
car body
body with
with the
the help
help of
of a
a robot
robot arm.
arm. Optical
Optical
data in real-time [20] and under industrial conditions. Vibrations of automotive parts can interfere
sensors for
sensors distance
forTHzdistance and orientation provide closed-loop positioning of of
thethe
THz head relative to the car
with the signaland andorientation
impede orprovideobstructclosed-loop positioning
thickness measurements. THz
However, head
suchrelative
vibrations to thecan
body
car andandare are
alsoalso
usedused for the correction of vibrations.
bebody
measured using optical for the
methods correction
and the of THz vibrations.
pulse traces can be corrected in software [75]. Water
vapor in the atmosphere has multiple absorption lines at THz frequencies (see Section 8) [78], leading
to artifacts in the THz pulse that can mask relevant thickness information [39]. These artifacts can be
minimized either numerically [79,80] or by purging the THz path with dry air or other gases [78].
Figure 12 shows an example of an implementation of a THz TDS system in the automotive
industry [41]. The THz reflection head is aligned to the car body with the help of a robot arm. Optical
sensors for distance and orientation provide closed-loop positioning of the THz head relative to the
car body and are also used for the correction of vibrations.

Figure 12.
Figure Implementation of
12. Implementation of aa THz
THz TDS
TDS in
in an
an industrial
industrial setting. Courtesy of
setting. Courtesy of Helmut
Helmut Fischer
Fischer GmbH.
GmbH.

Figure 12. Implementation of a THz TDS in an industrial setting. Courtesy of Helmut Fischer GmbH.
Sensors 2019, 19, 4203 11 of 35

5. Pharmaceuticals
Sensors 2019, 19, 4203 11 of 35
The pharmaceutical industry is a particularly promising target area for THz applications, both
in R&D and for purposes of condition monitoring and non-destructive testing. There are three main
5. Pharmaceuticals
reasons for this. First, pharmaceutical materials are (semi-) transparent at THz frequencies and often
possessThe pharmaceutical industry is a particularly promising target area for THz applications, both
characteristic spectral signatures, whereas they are opaque in the visible and near-infrared.
in R&D and for purposes of condition monitoring and non-destructive testing. There are three main
Second, high-precision quality monitoring is of immense importance in pharmaceutical production,
reasons for this. First, pharmaceutical materials are (semi-) transparent at THz frequencies and often
because the efficacy and safety of a medicinal product are contingent on accurate control of both its
possess characteristic spectral signatures, whereas they are opaque in the visible and near-infrared.
chemical
Second,composition
high-precisionand its microstructure
quality and mechanical
monitoring is of immense properties.
importance Thirdly, production,
in pharmaceutical pharmaceuticals
arebecause
high-value products
the efficacy andwhere
safety increased product
of a medicinal consistency
product and reduced
are contingent on accuratewastage
controlbring significant
of both its
cost benefits,
chemical therefore justifying
composition investmentand
and its microstructure in relatively
mechanicalexpensive
properties.technologies such as THz
Thirdly, pharmaceuticals aresensing
that improveproducts
high-value the effectiveness of quality
where increased monitoring.
product consistency and reduced wastage bring significant cost
benefits, therefore
From the early justifying
days of THz investment in relatively
spectroscopy, expensive
it was realisedtechnologies such asvariations
that it can reveal THz sensing that
in molecular
improve the effectiveness of quality monitoring.
configurations such as polymorphs, chirality, cocrystals, and crystalline or amorphous state (Figure
13) [81]From the early days
(and references of THz
therein). spectroscopy,
Since then a largeit number
was realised that
of such it canhave
studies reveal variations
been published in [e.g.,
molecular configurations such as polymorphs, chirality, cocrystals, and crystalline or amorphous
82–87], and uses of THz spectroscopy have been continuously growing in pharmaceutical and
state (Figure 13) [84] (and references therein). Since then a large number of such studies have
biochemical R&D. However, detailed spectroscopic measurements require precise control of
been published e.g., [85–90], and uses of THz spectroscopy have been continuously growing in
experimental conditions and specific forms of sample preparation, and are neither suitable nor
pharmaceutical and biochemical R&D. However, detailed spectroscopic measurements require precise
desirable forexperimental
control of direct industrial applications
conditions such
and specific as in-line
forms production
of sample monitoring
preparation, or near-line
and are neither suitablequality
inspection of finished
nor desirable products.
for direct industrial applications such as in-line production monitoring or near-line
quality inspection of finished products.

3.0

2.5
Absorption (a.u.)

Form I
2.0
Form II
1.5
Form III
1.0
Form IV
0.5
Form V
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Frequency (THz)

Figure 13. Terahertz absorption spectra of the five polymorphic forms of sulfathiazole (vertically offset
Figure 13. Terahertz absorption spectra of the five polymorphic forms of sulfathiazole (vertically
and normalized for clarity). Adapted from [84].
offset and normalized for clarity). Adapted from [81].
Industrial applications of THz sensing in the pharmaceutical industry have focused on two main
Industrial
areas: applications
inspection of THz
of tablet and sensing
capsule in theand
coatings; pharmaceutical
monitoring ofindustry have focused
tablet porosity and poreonsize.
two main
areas: inspectiontablets
Pharmaceutical of tablet and capsule
and capsules are thecoatings; andand
most popular monitoring
widespread ofform
tablet porosity anddrugs
of administering pore size.
to patients duetablets
Pharmaceutical to theirand
cost-effectiveness,
capsules are the easemost
of usepopular
and patient
andcompliance,
widespread with approximately
form of administering
60% of
drugs to medicines
patients being
due delivered
to their as oral solid dosage forms.
cost-effectiveness, ease Both tablets
of use andandpatient
capsulescompliance,
are formed with
by blending the
approximately 60%active pharmaceutical
of medicines ingredient as
being delivered (API)
oraland excipient.
solid TabletsBoth
dosage forms. are compacted
tablets andinto
capsules
a solid pellet, and often coated with a specialized protective layer. In contrast, capsules consist of
are formed by blending the active pharmaceutical ingredient (API) and excipient. Tablets are
uncompacted power which is encased in a specialised shell. The functional performance of tablets and
compacted into a solid pellet, and often coated with a specialized protective layer. In contrast,
capsules is strongly dependent on the physical and structural properties of both the coating and the
capsules consist of uncompacted power which is encased in a specialised shell. The functional
internal blend. These are the target application areas of THz sensing, for which it was shown to be
performance
particularlyof tablets and capsules is strongly dependent on the physical and structural properties
suitable.
of bothProcess
the coating and the
analytical internal(PAT)
technology blend. These
is “a arefor
system thedesigning,
target application
analysing,areas of THz sensing,
and controlling the for
which it was shown
manufacturing to be particularly
of pharmaceutical suitable.
compounds through timely measurements (i.e., during processing)
of Process analytical
critical quality technologyattributes
and performance (PAT) is of“araw
system for designing,
and in-process analysing,
materials and controlling
and processes, with the the
goal of ensuring final product quality” (Federal Drug Administration Guidance).
manufacturing of pharmaceutical compounds through timely measurements (i.e., during processing)Demonstrating the
of critical quality and performance attributes of raw and in-process materials and processes, with the
goal of ensuring final product quality” (Federal Drug Administration Guidance). Demonstrating the
feasibility and utility of THz measurements as a valid PAT tool is therefore crucial for its acceptance
and uptake by the industry.
 Sensors 2019, 19, 4203 12 of 35

Sensors 2019, 19, 4203 12 of 35

feasibility and utility of THz measurements as a valid PAT tool is therefore crucial for its acceptance
5.1. and uptake
Tablet by the
Coating industry.
Inspection
5.1. Tablet Coating Inspection
Pharmaceutical capsules consist of an enclosing soluble shell containing a loose powder blend
that commonly combines
Pharmaceutical an active
capsules ingredient
consist of an enclosing (API) andshell
soluble an containing
excipient.a Pharmaceutical
loose powder blend tablets
that are
formed by compressing
commonly combines an the powder
active blend(API)
ingredient into aandsolid
an pellet to which
excipient. a coating tablets
Pharmaceutical is commonly
are formedapplied,
in part to mask their
by compressing thetaste
powder andblend
odour.intoThe most
a solid important
pellet to whichfunctions
a coating of tablet andapplied,
is commonly capsuleincoatings—
part
to mask their taste and odour. The most important functions of tablet
which determine their performance in drug delivery—are twofold: to preserve drug functionality and capsule coatings—which
whiledetermine theirand
in storage, performance
to facilitatein drug
timelydelivery—are twofold:
drug dissolution to preserve
(which drug functionality
may involve while in
rapid or timed-release
storage, and to facilitate timely drug dissolution (which may involve rapid
or release at the correct location in the gastrointestinal tract). The performance of coatings or timed-release or releasecan be
at the correct location in the gastrointestinal tract). The performance of coatings can be affected by their
affected by their thickness, structural uniformity, and defects. Whereas inspection of such coatings is
thickness, structural uniformity, and defects. Whereas inspection of such coatings is very challenging,
very challenging, especially with respect to coating thickness.
especially with respect to coating thickness.
Reflection THz TDS offers an obviously suitable tool for monitoring coating thickness because
Reflection THz TDS offers an obviously suitable tool for monitoring coating thickness because the
the time delaybetween
time delay betweenthe thetwo
tworeflection
reflection signals
signals fromfrom
thethe surface
surface andandfromfrom the coating-core
the coating-core interface
interface
provides
provides a adirect measure
direct measure of coating
of the the coating layer thickness
layer thickness (see 4).
(see also Section also
ThisSection
was first4). This was first
demonstrated
demonstrated for tablets by Fitzgerald et al. in 2005 [88]. Since then
for tablets by Fitzgerald et al. in 2005 [91]. Since then much work has been done in thismuch work has been done
area,in this
area,forfor both
both tablet
tablet andand capsule
capsule coatings;
coatings; and and a comprehensive
a comprehensive reviewreview was published
was published by Haaser by et
Haaser
al. et
in 2013 (Figure 14) [92] (and references therein), while more recent work
al. in 2013 (Figure 14) [89] (and references therein), while more recent work has focused on specific has focused on specific
industry-relevant
industry-relevant aspects.AAdedicated
aspects. dedicated fully
fully automated
automatedcommercial
commercial THz THzinstrument
instrumentfor tablet coating
for tablet coating
inspection using a 6-axis robotic arm to map the tablet surface was
inspection using a 6-axis robotic arm to map the tablet surface was also developed [90]. also developed [93].

Figure
Figure 3D3D
14.14. images
images ofofcoating
coatingthickness
thickness of
of aabiconvex
biconvextablet. The
tablet. colour
The scale
colour is inisµm,
scale the X-Y-Z
in μm, the X-Y-Z
scale is in mm. Reprinted with permission from [92] © Elsevier.
scale is in mm. Reprinted with permission from [89] © Elsevier.

There are two main challenges. The spatial resolution is limited by the wavelength of THz
There are two main challenges. The spatial resolution is limited by the wavelength of THz
radiation to about 100 µm at best. More important, reflection measurements—the standard approach to
radiation to about
thickness 100 μmmade
gauging—are at best. Morebyimportant,
difficult the curved reflection
surfaces of measurements—the standard
tablets and capsules, whose approach
curvature
to thickness gauging—are
is of the same made
order as the THz difficult by the curved surfaces of tablets and capsules, whose
wavelength.
curvature is of the
Several samehave
studies order as the THz wavelength.
demonstrated the utility of such measurements for monitoring tablet
Several studies have demonstrated
dissolution performance by confirming the the utility of ofsuch
dependence meanmeasurements
dissolution timefor monitoring
(MDT) of tabletstablet
dissolution performance
on their coating by confirming
thickness the [92,94]
(Figure 15) e.g., dependence of mean therein).
(and references dissolution
MDTtime (MDT)
is the of tablets
standard
reference technique used by the pharmaceutical industry to test the dissolution performance
on their coating thickness (Figure 15) e.g., [89,91] (and references therein). MDT is the standard of solid
dosagetechnique
reference medicines.used by the pharmaceutical industry to test the dissolution performance of solid
dosage medicines.
 Sensors 2019, 19, 4203 13 of 35
Sensors 2019, 19, 4203 13 of 35

Mean disintegration time (hr)

7

20 40 60 80 100 120 140 160

Coating thickness (μm)

Figure
Figure15. Relationship between
15. Relationship betweentablet
tabletcoating
coatingthickness
thicknessand
andtheir
their MDT
MDT (mean
(mean dissolution
dissolution time)
time) for
for
tablets produced by two manufacturing processes: lab-scale (using laboratory equipment, solid
tablets produced by two manufacturing processes: lab-scale (using laboratory equipment, solid
symbols)
symbols)andandpilot-scale
pilot-scale(using
(usingindustrial
industrialequipment;
equipment;hollow
hollowsymbols).
symbols).Adapted
Adaptedfrom
from[94].
[91].

Recent work in this area has focused on demonstrating the feasibility and utility of industrial
Recent work in this area has focused on demonstrating the feasibility and utility of industrial
applications. Examples include the use of THz sensing as a PAT tool for evaluating coatings deposited
applications. Examples include the use of THz sensing as a PAT tool for evaluating coatings
under different manufacturing conditions [95], observing the effect of processing conditions on
deposited under different manufacturing conditions [92], observing the effect of processing
inter-tablet variations in coating thickness [96], and development of a thickness measurement technique
conditions on inter-tablet variations in coating thickness [93], and development of a thickness
for two-layer coatings on capsules [97].
measurement technique for two-layer coatings on capsules [94].
5.2. Monitoring Porosity and Pore Size
5.2. Monitoring Porosity and Pore Size
Timely dissolution of a pharmaceutical tablet is one of the most important aspects of drug
Timely dissolution of a pharmaceutical tablet is one of the most important aspects of drug
delivery. Tablet disintegration is controlled by its chemical and mechanical properties. During the
delivery. Tablet disintegration is controlled by its chemical and mechanical properties. During the
compression of a tablet, particles are consolidated to form interparticulate bonds and pores. Tablets
compression of a tablet, particles are consolidated to form interparticulate bonds and pores. Tablets
are commonly manufactured by uniaxial compaction of powder confined radially in a rigid die.
are commonly manufactured by uniaxial compaction of powder confined radially in a rigid die. This
This directional compaction results in anisotropic mechanical properties and pore structure, where
directional compaction results in anisotropic mechanical properties and pore structure, where pores
pores are elongated and predominantly aligned in the plane of the tablet. The pores in a tablet—their
are elongated and predominantly aligned in the plane of the tablet. The pores in a tablet—their size
size and connectivity—directly affect the rate at which the physiological fluids enter the tablet, leading to
and connectivity—directly affect the rate at which the physiological fluids enter the tablet, leading to
swelling of the particles and eventually causing the break-up of the compact into smaller agglomerates.
swelling of the particles and eventually causing the break-up of the compact into smaller
The size of the disintegrated particles then drives the dissolution rate of the drug. These mechanisms
agglomerates. The size of the disintegrated particles then drives the dissolution rate of the drug.
are strongly interconnected, as the swelling of particles dynamically changes the internal pore structure
These mechanisms are strongly interconnected, as the swelling of particles dynamically changes the
which influences the liquid imbibition process. Consequently, the tablet microstructure and the
internal pore structure which influences the liquid imbibition process. Consequently, the tablet
disintegration process play a pivotal role in product performance, and the performance of a tablet
microstructure and the disintegration process play a pivotal role in product performance, and the
can be predicted and optimized by understanding the relationship between the dissolution rate and
performance of a tablet can be predicted and optimized by understanding the relationship between
the granularity and porosity of the tablet material. Despite the widely recognized importance of
the dissolution rate and the granularity and porosity of the tablet material. Despite the widely
monitoring the tablet porosity during the manufacturing process, there are no continuous in-line
recognized importance of monitoring the tablet porosity during the manufacturing process, there are
non-destructive techniques for achieving this. THz technology is a highly promising solution to this
no continuous in-line non-destructive techniques for achieving this. THz technology is a highly
urgent problem due to its ability to directly measure porosity and its sensitivity to pore size. A body of
promising solution to this urgent problem due to its ability to directly measure porosity and its
work has been carried out aimed at developing it into an industrial tool.
sensitivity to pore size. A body of work has been carried out aimed at developing it into an industrial
Porosity has a dual effect on THz transmission properties of a material. The effective refractive
tool.
index decreases roughly linearly with porosity because the effective interaction length with the material
Porosity has a dual effect on THz transmission properties of a material. The effective refractive
shortens (there is less substance in the beam path). The ability of THz TDS to provide a direct
index decreases roughly linearly with porosity because the effective interaction length with the
measurement of the effective refractive index, therefore, makes it highly suitable for monitoring and
material shortens (there is less substance in the beam path). The ability of THz TDS to provide a direct
characterising porosity. In contrast to refractive index, loss tends to rise with porosity. This is because
measurement of the effective refractive index, therefore, makes it highly suitable for monitoring and
transmission loss combines contributions from absorption and scattering. Whereas the effective
characterising porosity. In contrast to refractive index, loss tends to rise with porosity. This is because
absorption of porous material declines with shorter interaction length, scattering rises strongly due
transmission loss combines contributions from absorption and scattering. Whereas the effective
to higher pore density and/or larger size of pores. Since pharmaceutical materials generally have
absorption of porous material declines with shorter interaction length, scattering rises strongly due
moderate THz absorption, this tends to result in the combined loss that increases with porosity.
to higher pore density and/or larger size of pores. Since pharmaceutical materials generally have
moderate THz absorption, this tends to result in the combined loss that increases with porosity.
Sensors 2019,19,
Sensors2019, 19,4203
4203 14
14ofof35
35

However, the relationship between loss and porosity is often complicated and strongly dependent
However, the relationship between loss and porosity is often complicated and strongly dependent on
on material absorption and the pore geometry and distribution.
material absorption and the pore geometry and distribution.
The first demonstration of porosity measurements in pharmaceutical tablets using THz TDS was
The first demonstration of porosity measurements in pharmaceutical tablets using THz TDS was
described by Ervasti et al. in 2012 [95], with many studies following, summarised by Markl et al. in
described by Ervasti et al. in 2012 [98], with many studies following, summarised by Markl et al. in [99]
[96] (and references therein). Since tablet dissolution depends on pore configuration, distribution and
(and references therein). Since tablet dissolution depends on pore configuration, distribution and
connectivity, as well as their size and size variation, recent work has aimed to develop techniques for
connectivity, as well as their size and size variation, recent work has aimed to develop techniques for
accurately evaluating these properties in addition to calculating mean porosity [97–101]. Of particular
accurately evaluating these properties in addition to calculating mean porosity [100–104]. Of particular
interest is the relationship between THz transmission properties and tablet dissolution (Figure 16)
interest is the relationship between THz transmission properties and tablet dissolution (Figure 16) [101],
[98], because dissolution time is a critical parameter in drug delivery. The behavior of powder during
because dissolution time is a critical parameter in drug delivery. The behavior of powder during
the manufacturing process and the effects of density variations in the final product have also been
the manufacturing process and the effects of density variations in the final product have also been
studied [102].
studied [105].

a 18
b

Effective refractive index @ 1 THz

1.75

17
1.70
B (cm-1 Hz-A)

16

1.65
15

14 1.60

13
200 400 600 800 1000 200 400 600 800 1000
Disintegration time (s) Disintegration time (s)

Figure 16. (a)

Figure16. (a) Relationship
Relationship between
between the
the loss parameterBBand
lossparameter andtablet
tabletdisintegration
disintegrationtime.
time. The
The
frequency-dependent scattering loss is described by αeff = B f A (where A ≈ 3.3), therefore the loss
A
frequency-dependent scattering loss is described by αeff = B f (where A ≈ 3.3), therefore the loss
parameter B reflects the strength of transmission loss. (b) Relationship between effective refractive
parameter B reflects the strength of transmission loss. (b) Relationship between effective refractive
index and disintegration time. Adapted from [101].
index and disintegration time. Adapted from [98].
6. Electronics
6. Electronics
From the early days of THz technologies, they were envisaged as an inspection tool for the
Fromindustry.
electronics the earlyThis dayswasof primarily
THz technologies, they were
because electronic envisaged
properties as an inspection
of semiconductor tool for
materials the
such
electronics
as industry. This
carrier concentration and was primarily
mobility determinebecause
theirelectronic propertiesatofTHz
dielectric properties semiconductor materials
frequencies, allowing
such as carrier concentration and mobility determine their dielectric properties
non-destructive, non-contact measurements. Hermann et al. [106] described an early demonstration at THz frequencies,
allowing
using non-destructive,
THz TDS to map areas of non-contact
the differing measurements. Hermann
carrier concentrations et al. wafers,
in silicon [103] described
achieving aan early
spatial
demonstration using THz TDS to map areas of the differing carrier concentrations
resolution of ~1 mm. Moreover, optical inspection techniques are often unsuitable for electronic circuits in silicon wafers,
achieving
and devicesa due
spatial resolution
to the opacity ofof ~1
themm. Moreover,
substrates optical
and/or inspection techniques are often unsuitable
packaging.
for electronic
Subsequent circuits
work and devices
in this duemostly
field has to the opacity
focusedof onthe substrates
three and/or packaging.
areas: inspection of electronic circuits;
Subsequent work in this field has mostly focused on three areas: inspection
inspection of solar cells; and conductivity measurements/mapping of thin films, particularly of electronic circuits;
as applied
inspection
to graphene. of solar cells; and conductivity measurements/mapping of thin films, particularly as
applied to graphene.
6.1. Electronic Circuits
6.1. Electronic Circuits
The spatial resolution of far-field THz imaging is of the order of ~1 mm, being limited by
The wavelengths,
the long spatial resolution and ofis far-field
thereforeTHz imaging is
insufficient forofdetailed
the order of ~1 mm,ofbeing
inspection limited
electronic by the
circuits.
long wavelengths, and is therefore insufficient for detailed inspection
Hence much work has been devoted to developing THz imaging techniques capable of delivering aof electronic circuits. Hence
much work hasresolution.
sub-wavelength been devoted to developing THz imaging techniques capable of delivering a sub-
wavelength resolution.
In a series of publications Yamashita et al. described the development and progressive refinement
In aterahertz-emission
of a “laser series of publications Yamashita
microscope” (LTEM),et achieving
al. described the resolution
a spatial development and
of a few µmprogressive
[107–111].
refinement
The of a “laser
LTEM operates terahertz-emission
by raster-scanning microscope”
the circuit (LTEM),beam
with a focused achieving
from aa femtosecond
spatial resolution
laser.ofWhen
a few
the laser is incident on photoconductive structures, it generates a transient photocurrent, causing aa
μm [104–108]. The LTEM operates by raster-scanning the circuit with a focused beam from
femtosecond
THz pulse to be laser. WhenStructures
emitted. the laser iscapable
incident ofon photoconductive
emitting structures,
THz, and therefore it generates
suitable a transient
to be examined
photocurrent, causing a THz pulse to be emitted. Structures capable of emitting THz, and therefore
suitable to be examined by the LTEM, include photoconductive switches with an external bias voltage
Sensors 2019, 19, 4203 15 of 35

Sensors 2019, 19, 4203 15 of 35

by the LTEM, include photoconductive switches with an external bias voltage (i.e., Auston-type
switches), unbiased
(i.e., Auston-type interfaces
switches), carrying
unbiased electric carrying
interfaces fields (e.g., p-n junctions,
electric fields (e.g.,Schottky contacts),
p-n junctions, and
Schottky
some semiconductor
contacts), surfaces. As with
and some semiconductor all photoconductive
surfaces. THz emitters,THz
As with all photoconductive the THz field the
emitters, amplitude
THz field is
proportional to the local electric field. The emitted THz radiation is collected
amplitude is proportional to the local electric field. The emitted THz radiation is collected and and analysed, producing
an image of
analysed, the circuit.
producing an image of the circuit.
Following
Following the theinitial
initialdemonstration
demonstration of of
LTEMLTEM imaging [107],
imaging its resolution
[104], was improved
its resolution was improved to better
to
than 3 µm [108]. Directing attention to p-n junctions, defective transistor
better than 3 μm [105]. Directing attention to p-n junctions, defective transistor circuits were circuits were distinguished
from correctlyfrom
distinguished functioning
correctly ones [109,110].
functioning ones Then imaging
[106,107]. from thefrom
Then imaging rearthe
side ofside
rear the ofcircuit was
the circuit
shown, allowing non-contact inspection where interconnect structures prevent
was shown, allowing non-contact inspection where interconnect structures prevent optical access optical access from the
front [111]. To facilitate characterization, laser-activated THz waveforms emitted
from the front [108]. To facilitate characterization, laser-activated THz waveforms emitted from p-n from p-n junctions
were studied
junctions werein detail
studied[112]. Figure[109].
in detail 17 shows
Figurethe 17
ability
showsof the
theLTEM
abilitytoof
detect defects to
the LTEM (broken
detectcontacts)
defects
in a circuit [112]. The capabilities of the LTEM were further extended and its spatial
(broken contacts) in a circuit [109]. The capabilities of the LTEM were further extended and its spatial resolution was
increased to 0.6 µm by incorporating near-field techniques [113]. A similar
resolution was increased to 0.6 μm by incorporating near-field techniques [110]. A similar approach approach was applied to
time-resolved
was applied totransmission
time-resolved imaging, demonstrating
transmission measurements measurements
imaging, demonstrating of carrier lifetimes and mobilities
of carrier lifetimes
with a spatial resolution of 60 µm [114].
and mobilities with a spatial resolution of 60 μm [111].

Figure 17. LTEM

LTEMimages
imagesofofa afunctioning
functioningcircuit
circuit(a)(a)and a defective
and one
a defective (b).(b).
one TheThe
arrow shows
arrow the
shows
location
the of the
location broken
of the brokencontact.
contact.Brighter areas
Brighter areasindicate
indicatestronger
stronger THz
THz emission.
emission. Reprinted
Reprinted with
[109] ©
permission from [112] © The Optical Society.

In order to image
image individual
individual microelectronic
microelectronic components,
components, a spatial resolution
resolution of of <100
<100 nm is
necessary. Huber et al. employed scattering near-field optical microscopy
necessary. Huber et al. employed scattering near-field optical microscopy (SNOM) using a 2.54 (SNOM) using a 2.54 THzTHz
gas
laser (methanol), achieving resolution of about 40 nm [115]. THz-SNOM
gas laser (methanol), achieving resolution of about 40 nm [112]. THz-SNOM imaging revealed imaging revealed individual
transistors, identified the
individual transistors, materialsthe
identified in materials
a device, in anda was able and
device, to measure
was able carrier concentrations.
to measure carrier
Although
concentrations. Although such systems are still too complex for industrial deployments, they into
such systems are still too complex for industrial deployments, they can provide insights can
device
providestructure
insights andinto performance.
device structure and performance.
Time-domain
Time-domain reflectometry
reflectometry(TDR) (TDR)is isa widely
a widelyused non-destructive
used non-destructive technique for detecting
technique faults
for detecting
and discontinuities in electronic circuits, a task traditionally addressed by using
faults and discontinuities in electronic circuits, a task traditionally addressed by using a pulse a pulse generator and
an oscilloscope.
generator and anNagel et al. changed
oscilloscope. Nagel ettheal.conventional method to a non-contact
changed the conventional method to aone by employing
non-contact one by a
TDS system and custom-built probes with a tip radius of 0.2 µm [116].
employing a TDS system and custom-built probes with a tip radius of 0.2 μm [113]. Their THz-TDR Their THz-TDR instrument
was capablewas
instrument of locating
capable of discontinuities in electronic
locating discontinuities structuresstructures
in electronic with a spatialwithresolution of 0.55 µm.
a spatial resolution of
Commercial THz-TDRTHz-TDR
0.55 μm. Commercial instruments are currently
instruments deployed deployed
are currently in the semiconductor industry. industry.
in the semiconductor
More recent work focused on applying applying THz measurements
measurements for non-destructive
non-destructive inspection of
packaged
packaged powerpowerelectronic
electronicdevices, obtaining
devices, imagesimages
obtaining of their spatial
of their features
spatialthrough the encapsulating
features through the
coatings [117].coatings
encapsulating The approach
[114]. Theadopted
approach wasadopted
similar was
to that usedtofor
similar measuring
that layers of car
used for measuring paint
layers of
(Section 4) and coatings on pharmaceutical tablets (Section 5), achieving a spatial
car paint (Section 4) and coatings on pharmaceutical tablets (Section 5), achieving a spatial resolutionresolution of ~0.5 mm.
Ahi
of ~0.5 et al. employed TDS combined with a range of measurement and analysis techniques to
mm.
test aAhi
variety
et al.ofemployed
attributesTDS for quality
combined control
withand authentication
a range of measurementof packaged integrated
and analysis circuitsto[118].
techniques test
The aspects tested included: the presence of unexpected materials in counterfeit devices,
a variety of attributes for quality control and authentication of packaged integrated circuits [115]. The blacktopping
aspects tested included: the presence of unexpected materials in counterfeit devices, blacktopping
layers (used by counterfeiters to hide the original label and over-print a false one), shape and
dimensions of hidden structures, sanded and contaminated devices, differences between internal
Sensors 2019, 19, 4203 16 of 35

layers2019,
Sensors (used by counterfeiters
19, 4203 to hide the original label and over-print a false one), shape and
16 of 35
dimensions of hidden structures, sanded and contaminated devices, differences between internal
structures of
structures ofcounterfeit
counterfeitand andauthentic
authentic devices,
devices, such
such as misshapen
as misshapen die-frames
die-frames and bond-wires.
and bond-wires. The
The measurement
measurement techniques
techniques used used
were:were: transmission
transmission spectroscopy
spectroscopy for material
for material characterization,
characterization, time-
time-of-flight
of-flight tomography,
tomography, and reflection
and reflection and transmission
and transmission imaging.imaging.
The image The image resolution
resolution was
was increased
increased
to ~ 0.1 mm toby
~ 0.1 mm byaapplying
applying a deconvolution
deconvolution techniquetechnique
and takingand taking
into intothe
account account
point the point
spread spread
function
function
of of theFigure
the emitter. emitter.18Figure
shows18contrasting
shows contrasting images
images of of authentic
authentic and counterfeit
and counterfeit chips.
chips. Since Since
current
current
THz TDSTHz TDS systems
systems are compact,
are compact, robustrobust and relatively
and relatively inexpensive,
inexpensive, suchsuch testing
testing solutions
solutions maymaybe
be readily
readily adapted
adapted to large-scale
to large-scale industrial
industrial inspection.
inspection.

Figure 18. Transmission images of an authentic (right) Intel chip and a counterfeit one (left). Reprinted
Figure 18. Transmission images of an authentic (right) Intel chip and a counterfeit one (left). Reprinted with
with permission from [118] © Elsevier.
permission from [115] © Elsevier.

AA simple
simple approach
approach to to testing
testing microwave
microwave monolithic
monolithic integrated
integrated circuits
circuits (MIMICs)
(MIMICs) and and very
very large
large
scale integrated
scale integrated (VLSI)
(VLSI) circuits
circuits were reported by Shur et al. who detected
al. who detected thethe electrical
electrical response
generated at the circuit pins when the component was illuminated by THz or
generated the circuit pins when the component was illuminated by THz or sub-THz radiation sub-THz radiation[116].
[119].
They further proposed
proposed that
thatmachine
machinelearning
learningbe
beused
usedtotoidentify
identifycharacteristic
characteristic response
response signatures
signaturesof
circuits
of having
circuits a large
having number
a large of pins,
number of by evaluating
pins, the associated
by evaluating (large) matrix
the associated (large)of matrix
signal responses.
of signal
The technique
responses. Theaims to address
technique aimsthetoproblem
addressof the
hardware
problem testing for performance,
of hardware testing authentication
for performance,and
security, and toand
authentication do so non-destructively,
security, and to do sorapidly, and on a large
non-destructively, scale.and on a large scale.
rapidly,

6.2. Solar
6.2. Solar Cells
Cells
THz measurements
THz measurementsonon solar cellscells
solar may may
serve serve
two separate purposes:
two separate characterisation
purposes: and inspection.
characterisation and
Characterisation of the solar cell material involves determining its electrical properties
inspection. Characterisation of the solar cell material involves determining its electrical properties in in order to
predict its performance and efficiency. The parameters measured are typically
order to predict its performance and efficiency. The parameters measured are typically complex complex conductivity,
charge carriercharge
conductivity, densitycarrier
and mobility,
density and
and their dependence
mobility, and theirondependence
solar-type illumination.
on solar-typeInspection,
illumination. on
the other hand,
Inspection, on theinvolves detecting
other hand, a variety
involves of flaws
detecting and irregularities.
a variety of flaws and irregularities.
Characterisation of solar cell material has been
Characterisation of solar cell material has been performed performed usingusing TDS
TDS [117].
[120]. However,
However, it it lacked
lacked
sufficient spatial
sufficient spatial resolution
resolution to to examine
examine individual
individual cell
cell devices.
devices. The The LTEM
LTEM technique
technique (described
(described
above) was also employed to image and analyse solar cells [121–123].
above) was also employed to image and analyse solar cells [118–120]. Of particular interest Of particular interest
is theis
the detailed comparison of LTEM with conventional analysis methods,
detailed comparison of LTEM with conventional analysis methods, electroluminescence (EL), electroluminescence (EL),
photoluminescence (PL),
photoluminescence (PL), and
and laser
laser beam
beam induced
induced current
current (LBIC)
(LBIC) [120],
[123], which
which found
found that
that LTEM
LTEM is is aa
useful complementary technique that has advantages in spatial resolution and in
useful complementary technique that has advantages in spatial resolution and in its ability to observe its ability to observe
electric fields,
electric fields, surface
surface defects,
defects, grain
grain boundaries,
boundaries, and and photocarrier
photocarrier dynamics
dynamics in in the
the vicinity
vicinity of of the
the
depletion layer.
depletion layer.
TDS and
TDS and FDS FDS were
were both
both employed
employed in in reflection
reflection mode
mode forfor inspection
inspection ofof commercial
commercial solar solar cells,
cells,
successfully detecting
successfully detecting manufacturing
manufacturing faults
faults such
such as
as defects
defects inin tab
tab wires
wires and
and soldering,
soldering, cracks,
cracks, andand
variations in doping [124,125]. Tip-based near-field transmission measurements
variations in doping [121,122]. Tip-based near-field transmission measurements improved the spatial improved the spatial
resolution of
resolution of the
the image
image down
down to to 10
10 μm
µm [126–131]. The technique
[123–128]. The technique produced
produced maps
maps of of sheet
sheet resistance
resistance
capable of
capable of revealing
revealing faults
faults in
in individual
individual devices
devices (Figure
(Figure 19).
19).
Sensors 2019, 19, 4203 17 of 35
Sensors 2019, 19, 4203 17 of 35

Figure
Figure 19. (a)AAhigh-resolution
19. (a) high-resolutionsheet
sheet resistance
resistance map
map of aofsolar
a solar
cell cell obtained
obtained withwith tip-based
tip-based THz
THz near-
near-field microscopy. (b) Expanded view of a single finger. (c) Doping along the line marked
field microscopy. (b) Expanded view of a single finger. (c) Doping along the line marked in (b). Figurein (b).
Figure courtesy of Protemics
courtesy of Protemics GmbH. GmbH.

6.3. Graphene
6.3. Graphene
Industrial-scale manufacturing of high-quality graphene sheets is challenging, and the quality
Industrial-scale manufacturing of high-quality graphene sheets is challenging, and the quality
of the product depends critically on its conductivity and carrier mobility. Quality and uniformity of
of the product depends critically on its conductivity and carrier mobility. Quality and uniformity of
graphene sheets are essential for its commercialisation, requiring large-scale inspection techniques
graphene sheets are essential for its commercialisation, requiring large-scale inspection techniques
that are non-conduct and rapid. THz transmission and reflection measurements can provide solutions
that are non-conduct and rapid. THz transmission and reflection measurements can provide
where optical techniques cannot. In recent years, THz inspection of graphene has attracted growing
solutions where optical techniques cannot. In recent years, THz inspection of graphene has attracted
interest from both manufacturers and users, and a considerable amount of work has been devoted to
growing interest from both manufacturers and users, and a considerable amount of work has been
mapping conductivity and carrier mobility of graphene sheets, employing techniques similar to those
devoted to mapping conductivity and carrier mobility of graphene sheets, employing techniques
described earlier in this Section [131–134].
similar to those described earlier in this Section [128–131].
7. Petrochemicals
7. Petrochemicals
Petrochemicals are particularly suited to the analysis and quality monitoring by THz spectroscopy,
Petrochemicals
because of their high are particularly
transparency suited
at THz to the analysis
frequencies and quality
and the sensitive monitoring
dependence by THz
of their THz
spectroscopy, because of their high transparency at THz frequencies and
dielectric properties on their chemical composition and the presence of contaminants. Although the sensitive dependence of
their THz dielectric
petrochemical products properties on theiraschemical
may be classed low value, composition
their qualityand the presence
control, of contaminants.
and condition monitoring
Although petrochemical products may be classed as low value, their
can deliver large benefits in cost savings, improved safety, and reduced wastage. Examples quality control, and condition
include
monitoring can deliver large benefits in cost savings, improved safety,
quality control of fuels to improve engine performance and reduce wear, and condition monitoring and reduced wastage.
Examples
of lubricatinginclude
oils toquality
facilitatecontrol
timelyofreplacement.
fuels to improve engine performance
The petrochemical industry and reduce
thus wear,
appears and
to be a
condition monitoring of lubricating oils to facilitate timely replacement.
promising area for THz applications, where much near-industrial work is being carried out, although The petrochemical industry
thus
at appears
present the to be a promising
authors are not awarearea of
forany
THz applications,
field deployment. where much near-industrial work is being
carried out, although at present the authors
Petrochemicals consist primarily of hydrocarbon chains are not aware of any field deployment.
of various lengths, mixed with other
Petrochemicals consist primarily of hydrocarbon
types of hydrocarbons such as aromatics. They may also contain other chains of various lengths, components,
flammable mixed with othersuch
types
as of hydrocarbons
alcohols, such as aromatics.
as well as non-flammable They may
impurities andalso contain other
contaminants. Theyflammable
are refinedcomponents,
from crude such
oil
using distillation processes that separate crude petroleum into grades according to the number oil
as alcohols, as well as non-flammable impurities and contaminants. They are refined from crude of
using distillation
carbons processes
in the molecular thatgaseous
chain: separatefuelscrude(1–4petroleum
carbons),into grades
gasoline according
(5–12 carbons), to the number
jet fuel, of
diesel,
carbonsoil
heating in (12–20
the molecular
carbons), chain: gaseous
lubricating oilsfuels (1–4carbons),
(20–30 carbons), gasoline
fuel (5–12
oil (30–40 carbons),
carbons), andjetparaffin
fuel, diesel,
wax
heating oil (12–20 carbons), lubricating oils (20–30 carbons), fuel oil (30–40
or petroleum jelly (40–50 carbons). Pure hydrocarbons are non-polar substances, which makes them carbons), and paraffin wax
or petroleum
low-loss in thejelly
THz (40–50
range. carbons).
On the Pure
otherhydrocarbons
hand, their THz are refractive
non-polarindices
substances, which
increase makes them
monotonically
low-loss in the THz range. On the other hand, their THz refractive indices
with the number of carbons (Figure 20) [135,136], making it possible to use THz sensing for qualitative increase monotonically
withquantitative
and the number of carbons (Figure
identification 20) [132,133],
or grading making it possible
of petrochemicals. This to use THz sensing
is particularly usefulforfor
qualitative
heavier
and quantitative
grades which are identification
opaque in the or grading
visible. of petrochemicals.
Moreover, additives orThis is particularly
contaminants useful for heavier
in petrochemicals (e.g.,
grades which are opaque in the visible. Moreover, additives or contaminants
bioethanol, water, oxidation products) tend to be polar substances that have strong THz absorption, in petrochemicals (e.g.,
bioethanol, water, oxidation products) tend to be polar substances that have strong THz absorption,
enabling easy detection at low levels of contamination. For all those reasons petrochemicals have
Sensors 2019, 19, 4203 18 of 35
Sensors 2019, 19, 4203 18 of 35

long attracted strong interest for industrial implementation of THz sensing, and much literature
existsenabling
on the easy detection
subject [134].atStudies
low levels
haveof contamination.
targeted two For all those
types reasons petrochemicals
of applications: have and
identification long grading;
attracted strong interest for industrial implementation of THz sensing, and much literature exists on
and detection and quantification of contaminants.
the subject [137]. Studies have targeted two types of applications: identification and grading; and
detection and quantification of contaminants.

1.43
Refractive index

1.42

1.41

1.40

8 9 10 11 12 13 14 15 16 17
Number of carbons in backbone chain
Figure 20. Refractive index at 1 THz (at 20 ◦ C) of alkane compounds with different numbers of carbon
Figure 20.in
atoms Refractive index
their backbone at 1 for
chain, THz (atlinear
both 20 °C) ofbranched
and alkane compounds with different
compounds, showing numbers
the dependence onof carbon
atomsthein their number
carbon backboneandchain, for both
branching. linear
Adapted and
from branched compounds, showing the dependence on
[136].
the carbon number and branching. Adapted from [133].
7.1. Crude Petroleum

7.1. CrudeTHz TDS has been investigated as a tool for identifying the composition of crude petroleum.
Petroleum
A combination of THz spectral data and multivariate statistical methods was used to identify crude oils
THzdifferent
from TDS hasoilbeen fieldsinvestigated as a tool
[138]. Other studies for identifying
looked at oil content the composition
in oil shale [139], atof crude
the petroleum. A
asphaltene
combination
content inof THzoilspectral
crude [140], at data and multivariate
wax crystals statistical
in crude oil [141], and atmethods was usedofto
spectral signatures identify
coal crude oils
tar [142].
Due to oil
from different strong
fieldsTHz absorption
[135]. Otherby water, THz
studies sensing
looked at oilhas been employed
content for measurements
in oil shale of
[136], at the asphaltene
water content in a broad variety of materials, including crude oil (again using TDS). High
content in crude oil [137], at wax crystals in crude oil [138], and at spectral signatures of coal tar [139].precision
measurements of low water content (0.01–0.25% w/w) were demonstrated [143], whereas a specially
Due to strong THz absorption by water, THz sensing has been employed for measurements of
designed sampler was used to measure high water content in the 1.8% to 90.6% range [144]. In a broad
water content in a broad variety of materials, including crude oil (again using TDS). High precision
and detailed study, oil contamination by both water and solids (e.g., sand) was investigated [145],
measurements
showing thatof it islow water
possible contentconcentrations
to measure (0.01%–0.25% w/w)With
of both. were demonstrated
particular reference to[140], whereas a
oil fields
specially designed sampler was used to measure high water content in the 1.8%
with a high presence of water, both water content and its distribution were characterized [146]. to 90.6% range [141].
In a broad and detailed study, oil contamination by both water and solids (e.g., sand) was
7.2. Fuels
investigated [142], showing that it is possible to measure concentrations of both. With particular
referenceAstoinoil
crude oil, with
fields THz TDS was presence
a high applied to of
identifying fuel grades,
water, both water additives, and its
content and contaminants.
distribution were
A particularly
characterized [143]. important aspect of fuel analysis is the ability to distinguish and quantify its octane
number [147–153]. This has been demonstrated in several studies: in gasoline and diesel [147], in
mixtures of gasoline and diesel [148], in mixtures of 90# and 97# octane gasoline (Figure 21) [149,150],
7.2. Fuels
in gasoline of 87#, 89#, and 90# octane [151], and in derv fuel oils [152]. Since fuels with higher octane
numbers have shorter
As in crude oil, THz carbon
TDSchain
waslengths,
applied their
to THz refractivefuel
identifying indices and absorption
grades, additives,coefficients are
and contaminants.
lower than those
A particularly of low-octane
important aspect(longer
of fuelcarbon chains)
analysis fuels.
is the ability to distinguish and quantify its octane
Likewise, it is highly useful to identify and quantify the common additives that are used to
number [144–150]. This has been demonstrated in several studies: in gasoline and diesel [144], in
augment fuel and to improve its performance [151,153,154]. Evaluation of the cetane number was
mixtures of gasoline and diesel [145], in mixtures of 90# and 97# octane gasoline (Figure 21) [146, 147],
demonstrated in biodiesel-diesel blends [153], and in ethanol-gasoline mixtures the content of ethanol
in gasoline of 87#, 89#,
was determined with 1%andaccuracy
90# octane [148],
[154,155]. and inbeing
Ethanol derva polar
fuel oils [149].
liquid, Since fuels
its addition with
strongly higher octane
increases
numbers have
both the shorterindex
refractive carbon
and chain lengths,coefficient
the absorption their THz of refractive
the mixtureindices
compared and absorption
with coefficients
those of pure
are lower than
gasoline. those ofoflow-octane
Derivatives (longer
benzene (toluene, carbon chains)
ethylbenzene, fuels.commonly comprise a significant
and xylene)
volume fraction
Likewise, it is ofhighly
gasoline, and these
useful also have
to identify andbeen identified
quantify theusing
commonTDS [156]. Concentrations
additives that are used to
augment fuel and to improve its performance [148,150,151]. Evaluation of the cetane number was
demonstrated in biodiesel-diesel blends [150], and in ethanol-gasoline mixtures the content of ethanol
was determined with 1% accuracy [151,152]. Ethanol being a polar liquid, its addition strongly
increases both the refractive index and the absorption coefficient of the mixture compared with those
 Sensors19,
Sensors 2019, 2019, 19, 4203
4203 19 of 35 19 of 35

been detected down to 0.2%

of methyl methacrylate, [152].used
an additive Likewise, sulphur
to lower the freezingcontamination wasbeen
point of diesel, have detected
detectedat low
concentrations down
down to 0.2% to Likewise,
[155]. 0.2 ppm sulphur
[152,153]. As with other
contamination additives
was detected andconcentrations
at low contaminants thattoare not
down
pure hydrocarbons, theAs
0.2 ppm [155,156]. presence of sulphur
with other additiveseven at low concentrations
and contaminants that are notincreases THz absorption.
pure hydrocarbons, the
presence of sulphur even at low concentrations increases THz absorption.

1.444
refractive index
3.0
Absorption coefficient (cm-1) 1.442

Refractive index
2.5
1.440

1.438
2.0

1.436
1.5 absorption coefficient
1.434
0.0 0.2 0.4 0.6 0.8 1.0
pure 97# 97# / 90# mixing ratio pure 97#

Figure 21. Refractive indices (green symbols, right axis) and absorption coefficients (blue symbols, left
Figureaxis)
21. at
Refractive indicesmixtures
1 THz of gasoline (green symbols, right
of 90# and 97# axis)Adapted
octane. and absorption
from [151].coefficients (blue symbols,
left axis) at 1 THz of gasoline mixtures of 90# and 97# octane. Adapted from [148].
7.3. Oils

7.3. Oils Lubricating and insulating oils differ from fuels in that their performance depends critically on
specific viscoelastic or dielectric properties, and to achieve this they tend to be highly purified and their
Lubricating
compositionsand insulating
tightly specified.oils differ from
Moreover, fuels in that
both lubricating andtheir performance
insulating oils degrade depends critically on
in the process
specificofviscoelastic
normal operation,or dielectric
therefore, properties, and to achieve
continued performance requiresthis they monitoring
periodic tend to beand highly purified and
replacement.
their compositions
For these reasons, tightly specified.
most work Moreover,
in this area has focusedbothon lubricating and insulating
detecting contamination oils degrade
or degradation of oils. in the
A few studies have used THz TDS for identifying oils [157,158]. In particular,
process of normal operation, therefore, continued performance requires periodic monitoring and a linear relationship
was shown
replacement. Fortothese
exist between
reasons,oilmost
viscosity
workandinitsthis
THzarea
refractive index [157],
has focused on arising
detectingfromcontamination
the fact that or
both viscosity and refractive index increase with the number of carbons in the chain. A related study
degradation of oils.
on contamination of engine oil with gasoline [158] showed that the refractive index of oil decreases
Aand few studies have used THz TDS for identifying oils [154,155]. In particular, a linear
its absorption coefficient increases in the presence of gasoline, with contamination levels down to
relationship
4% clearlywasdetectable.
shown to exist between oil viscosity and its THz refractive index [154], arising from
the fact that both
Oils thatviscosity
are used as and refractive
engine index
lubricants and increase with the
as an insulation number
medium of carbonsdegrade
in transformers in the chain.
in A
relatedoperation
study onbycontamination of engine
processes of oxidation, andoil with
there is agasoline [155] showed
need to monitor that the
their condition forrefractive
purposes ofindex of
timely replacement. Employing THz sensing to detect oxidation products—carbonyl
oil decreases and its absorption coefficient increases in the presence of gasoline, with contamination and hydroxyl
compounds—in
levels down oils isdetectable.
to 4% clearly an effective technique, because those compounds are strongly polar and
therefore strongly THz absorbing, allowing detection at low concentrations. A number of studies
Oils that are used as engine lubricants and as an insulation medium in transformers degrade in
have been performed demonstrating high sensitivity and good discrimination of such measurements
operation by 22)
(Figure processes
[159–165].of Contamination
oxidation, and bythere
wateris a need
can also betoa cause
monitor their
of oil condition
degradation, for purposes
especially in of
timelyengine
replacement. Employing THz sensing to detect oxidation products—carbonyl
oils where water is a product of fuel burning, and where sensitive THz detection has been and hydroxyl
compounds—in
demonstrated oils is an effective technique, because those compounds are strongly polar and
[165,166].
therefore strongly THz absorbing, allowing detection at low concentrations. A number of studies
have been performed demonstrating high sensitivity and good discrimination of such measurements
(Figure 22) [156–162]. Contamination by water can also be a cause of oil degradation, especially in
engine oils where water is a product of fuel burning, and where sensitive THz detection has been
demonstrated [162,163].
Sensors 2019, 19, 4203 20 of 35
Sensors 2019, 19, 4203 20 of 35

Transmittance change ratio

0.1

dodecane
alkybenzene
mineral oil
0.01
0 20 40 60 80 100 120
Heating time (min)

Figure 22. Reduction in THz transmission due to oxidation in three insulating oils. Adapted from [162].
Figure 22. Reduction in THz transmission due to oxidation in three insulating oils. Adapted from [159].
8. Gas Sensing

8. Gas It
Sensing
has long been known that numerous gases have absorption lines in the far-infrared, and a large
literature exists on far-IR gas spectroscopy dating back to the beginning of the last century, with studies
It hason
focusing longgasesbeen knownto that
of interest numerous
astronomy, gases have
astrophysics, absorptionscience.
and atmospheric lines inFrom
the the
far-infrared,
early days and a
large
of THz instrumentation, it was realised that THz measurements can contribute useful datacentury,
literature exists on far-IR gas spectroscopy dating back to the beginning of the last to this with
studies focusing
area, leading onextensive
to an gases of bodyinterest to astronomy,
of work. astrophysics,
Since gas absorption linesand atmospheric
are commonly veryscience.
narrow—a From the
early
fewdays
GHzof at THz instrumentation,
atmospheric it was realised
pressure (linearly proportionalthat to
THz measurements
pressure)—and can
their contribute
central useful data
frequencies
to have
this area, leading to established
been previously an extensive body
or can be of work. Since
calculated, many gasof absorption
these studieslines
usedare commonly very
narrow-band
high-resolution
narrow—a few GHz techniques, often employing
at atmospheric electronic
pressure devices,
(linearly in contrastto
proportional to pressure)—and
the broadband photonic
their central
TDS typically employed in other areas of THz applications. Although electronic
frequencies have been previously established or can be calculated, many of these studies used systems are incapable
of delivering frequencies
narrow-band high-resolution and bandwidths
techniques,as high as photonic
often employingsystems, target gasdevices,
electronic species have sufficient to the
in contrast
lines at lower frequencies and in narrow bands to permit detection and identification.
broadband photonic TDS typically employed in other areas of THz applications. Although electronic
Many of the examined gases are of interest in pollution monitoring and Earth observation.
systems are incapable of delivering frequencies and bandwidths as high as photonic systems, target
However, in the early days of THz measurements, it was not feasible to observe gases in the
gasenvironment.
species haveWith sufficient lines at lower frequencies and in narrow bands to permit detection and
the recent development of THz technologies, it became possible to perform
identification.
measurements in the field and to detect various gases in the ambient atmosphere. This led to a
Many ofofthe
resurgence examined
interest gases are of interest
in applications-focused THz gas in detection
pollutionand monitoring
monitoring,and andEarth observation.
the growing
However,
awareness inofthe early days Many
its advantages. of THz measurements,
of the gases of interest it do
was notnot
havefeasible
lines intotheobserve
near-IR, gases
where in the
environment. With are
sensor technologies themostrecent development
effective of THz
and available. technologies,
In contrast, it became
in the mid-IR where thepossible to perform
great majority
of gases have their
measurements in the main
fieldabsorption bands, various
and to detect spectroscopygasesis in
challenging,
the ambient requiring larger and
atmosphere. more
This led to a
expensive sensors than in the near-IR. Moreover, polyatomic gases have dense
resurgence of interest in applications-focused THz gas detection and monitoring, and the growing and complex absorption
bands in the
awareness of mid-IR with a high
its advantages. multiplicity
Many of the ofgases
lines,ofmaking it difficult
interest do nottohave
achieve sufficient
lines in the specificity
near-IR, where
and to identify gas mixtures. At THz frequencies, gases have relatively few lines, facilitating specificity
sensor technologies are most effective and available. In contrast, in the mid-IR where the great
and selectivity, and the current technology is able to provide requisite detection sensitivity while also
majority of gases have their main absorption bands, spectroscopy is challenging, requiring larger and
being sufficiently robust and compact. Three areas have attracted particular interest: environmental
more expensive sensors than in the near-IR. Moreover, polyatomic gases have dense and complex
monitoring, including pollutants and Earth observation, human breath analysis, and quality monitoring
absorption bands
of natural gas in thein production
the mid-IR andwith a high
supply chain.multiplicity of lines, making it difficult to achieve
sufficient specificity and to identify gas mixtures. At THz frequencies, gases have relatively few lines,
8.1. Environmental
facilitating Monitoring
specificity and selectivity, and the current technology is able to provide requisite detection
sensitivity whileofalso
The area gas being
sensingsufficiently robust monitoring
for environmental and compact. Threedetection
includes areas have attracted and
of pollutants particular
interest: environmental
contaminants monitoring, ofincluding
to meet the requirements health and pollutants and Earth
safety and monitoring observation,
of atmospheric human breath
components
analysis, and quality
for weather monitoring
and climate of natural
observation. gas in
Typically, the production
detection and demands
of pollutants supply chain.
high sensitivity,
selectivity and specificity; whereas atmospheric monitoring may have less stringent requirements.
Commercial Monitoring
8.1. Environmental terahertz spectrometers for trace gas sensing were already developed a decade ago,
employing precisely tunable diode lasers and photomixers (see FDS in Section 1) [167].
The area of gas sensing for environmental monitoring includes detection of pollutants and
contaminants to meet the requirements of health and safety and monitoring of atmospheric
components for weather and climate observation. Typically, detection of pollutants demands high
sensitivity, selectivity and specificity; whereas atmospheric monitoring may have less stringent
requirements.
 Commercial terahertz spectrometers for trace gas sensing were already developed a decade ago,
employing precisely tunable diode lasers and photomixers (see FDS in Section 1) [164].
A recent review of toxic compound detection by THz spectroscopy by Yang et al. [165] lists a
wide range of gases studied. In a work focused on demonstrating selectivity, TDS was used to
Sensors 2019, 19, 4203 21 of 35
examine eight gases [166]: acetaldehyde, acetonitrile, ethanol, water, methanol, ammonia,
propionaldehyde, and propionitrile. In contrast, focusing on sensitivity and also using TDS, detection
at concentrations
A recent review of of
less
toxicthan 10 ppm detection
compound was achieved
by THz inspectroscopy
carbon monoxide, by Yangmethanol,
et al. [168] water, and
lists a wide
acetonitrile
range of gases [167].
studied. In a work focused on demonstrating selectivity, TDS was used to examine
eight gases [169]:inacetaldehyde,
Much work this area hasacetonitrile,
been devoted to developing
ethanol, dedicated
water, methanol, electronicpropionaldehyde,
ammonia, systems for gas
sensing, comprising both a source & detector assembly and a gas cell for collecting
and propionitrile. In contrast, focusing on sensitivity and also using TDS, detection at concentrations and concentrating
gas samples.
of less than 10The ppm advantage
was achieved of all-electronic systems for
in carbon monoxide, this taskwater,
methanol, is theirandcompactness, robustness
acetonitrile [170].
and low power consumption. Their narrow frequency range can
Much work in this area has been devoted to developing dedicated electronic systems be tolerated provided that
foritgas
is
tailored to the specific gases examined. In an early work, methyl fluoride, difluoromethane,
sensing, comprising both a source & detector assembly and a gas cell for collecting and concentrating carbonyl
sulphide
gas samples.and The
methyl iodide were
advantage detected insystems
of all-electronic the rangeforbetween
this task 254–256 GHz [168]. robustness and
is their compactness,
By building a purpose-designed system that comprised
low power consumption. Their narrow frequency range can be tolerated provided an electronic sensor operating between
that it is tailored to
210–270
the specificGHz and examined.
gases a gas cell capable of collecting
In an early work, methyl and fluoride,
concentrating gas samples,carbonyl
difluoromethane, Neese etsulphide
al [169]
were able toiodide
and methyl detectwereand detected
identify in mixtures
the range of between
14 gases254–256
(acetonitrile,
GHz [171].methyl fluoride, acrylonitrile,
sulfurBydioxide, methyl iodide, methyl bromide, trifluoromethane,
building a purpose-designed system that comprised an electronic acrolein, propionitrile,
sensor operating between1,1
difluoroethene,
210–270 GHz and vinyl
a gasfluoride, oxetane,
cell capable of vinyl bromide,
collecting and 1,2 dichloroethane)
concentrating at ppt levels,
gas samples, Neese meeting the
et al [172]
requirements of the Clean Air Act (http://www.epa.gov) (Figure 23). Similar custom-built
were able to detect and identify mixtures of 14 gases (acetonitrile, methyl fluoride, acrylonitrile, sulfur gas sensing
electronic systems
dioxide, methyl were methyl
iodide, described in [170,171].
bromide, Another custom-built
trifluoromethane, sensing system
acrolein, propionitrile, 1,1and the gas cell
difluoroethene,
were
vinylused to monitor
fluoride, oxetane, food spoilage
vinyl by detecting
bromide, hydrogen sulphide
1,2 dichloroethane) at around
at ppt levels, 611.4the
meeting GHz, achieving
requirements
aofsensitivity of 0.5 ppm [172].
the Clean Air Act (http://www.epa.gov) (Figure 23). Similar custom-built gas sensing electronic
systems were described in [173,174]. Another custom-built sensing system and the gas cell were used
to monitor food spoilage by detecting hydrogen sulphide at around 611.4 GHz, achieving a sensitivity
of 0.5 ppm [175].

Figure 23. Expanded section of the detected spectrum showing spectral signatures of eight gases.
Figure 23. Expanded section of the detected spectrum showing spectral signatures of eight gases.
Adapted from [172].
Adapted from [169].
In contrast, a system employing a vector network analyser (VNA) at 238–252 GHz was used to
In contrast, a system employing a vector network analyser (VNA) at 238–252 GHz was used to
detect acetaldehyde, methanol, deuterated methanol, acetone, isopropyl alcohol, acetonitrile, and
detect acetaldehyde, methanol, deuterated methanol, acetone, isopropyl alcohol, acetonitrile, and
ethanol, and their mixtures [176]. Although good specificity and selectivity were demonstrated, such a
ethanol, and their mixtures [173]. Although good specificity and selectivity were demonstrated, such
VNA-based system is less suited to field deployment due to its lack of compactness and robustness
a VNA-based system is less suited to field deployment due to its lack of compactness and robustness
and its high cost.
and its high cost.
In addition to toxic gases, detection of particulates in the atmosphere is also vitally important for
In addition to toxic gases, detection of particulates in the atmosphere is also vitally important
pollution monitoring. Two studies have used TDS to measure levels of PM2.5, i.e., concentration of
for pollution monitoring. Two studies have used TDS to measure levels of PM2.5, i.e., concentration
particles with a mean size of 2.5 µm [177,178], detecting concentrations as low as 20 µg/m3 .
of particles with a mean size of 2.5 μm [174,175], detecting concentrations as low as 20 μg/m3.
High-resolution THz instruments may also be employed as remote toxic gas sensors at disaster
sites. An FDS system and a multipass cell evacuated to 1 mbar were used to detect trace amounts of
gases that might be released in a blaze or a chemical spill in an industrial facility, achieving detection
limits in the range of 20 ppm for ammonia (NH3 ) and 100 ppm for sulfur dioxide (SO2 ) [179]. The fact
that terahertz radiation is able to penetrate optically opaque environments, such as black smoke, was
 Sensors 2019, 19, 4203 22 of 35

High-resolution THz instruments may also be employed as remote toxic gas sensors at disaster
sites. An FDS system and a multipass cell evacuated to 1 mbar were used to detect trace amounts of
gases that might be released in a blaze or a chemical spill in an industrial facility, achieving detection
Sensors 2019, 19, 4203 22 of 35
limits in the range of 20 ppm for ammonia (NH3) and 100 ppm for sulfur dioxide (SO2) [176]. The fact
that terahertz radiation is able to penetrate optically opaque environments, such as black smoke, was
exploited
exploited toto identify
identify hydrogen
hydrogen cyanide
cyanide (HCN)
(HCN) and
and water
water molecules
molecules generated
generated by the
by the combustion
combustion of aof
a urethane
urethane foam foam block
block [177].
[180].

8.2.Breath
8.2. BreathAnalysis
Analysis
Analysing
Analysing gases
gases exhaled
exhaled in human
in human breath
breath is a is a highly
highly promising
promising application,
application, withwith the potential
the potential for
for early
early detection
detection of disease
of disease biomarkers.
biomarkers. Such detection
Such detection wouldwouldbe fast be
and fast and non-invasive,
non-invasive, facilitating
facilitating early
early diagnosis
diagnosis and state-of-health
and state-of-health monitoring.
monitoring. Potential Potential
detectabledetectable diseaseslung
diseases include include
cancer,lung cancer,
diabetes,
diabetes,
some some neurological
neurological disorders, anddisorders,
smoking and smoking
and alcoholand alcohol consumption.
consumption.
Custom-designedelectronic
Custom-designed electronicsystems
systemswere
wereusedusedtotodetect
detectalcohol
alcohol[181]
[178] and
and several
several bio-marker
bio-marker
gases
gases [179]
[182] in in breath
breath with
with goodgood sensitivity.
sensitivity. Rothbart
Rothbart et al.et[183]
al. [180] employed
employed a specially
a specially designed
designed gas
gas cell
cell which
which the probe
the probe beam would
beam would traverse
traverse multiple
multiple times (11 times (11 passes)
passes) to extend
to extend the interaction
the interaction length length
with
with
the the contained
contained gas and gas
thusand thus the
increase increase the absorption
absorption signal. Measuring
signal. Measuring between
between 220–330 GHz,220–330 GHz,
they were
they
able to were
detectable to detect
21 gases (water,21hydrogen
gases (water, hydrogen
cyanide, carboncyanide,
monoxide, carbon monoxide,
nitrogen nitrogen oxide,
oxide, formaldehyde,
formaldehyde,
methanol, hydrogen methanol,
sulfide,hydrogen sulfide,
acetonitrile, methyl acetonitrile,
isocyanide,methyl isocyanide,
acetaldehyde, acetaldehyde,
ethanol, ethanol,
vynil isocyanide,
vynil isocyanide,
acrolein, acrolein,
acetone, carbonyl acetone,
sulfide, carbonyl
dimethyl sulfide,
sulfide, dimethyl
isoprene, sulfide, isoprene,
butyraldehyde, butyraldehyde,
methyl nitrite, pyruvic
methyl
acid, nitrite,
butyric acid)pyruvic
(Figureacid,
24).butyric acid)they
Moreover, (Figure 24). Moreover,
observed they observed
a clear distinction a clear
between distinction
smoker and
between smoker
non-smoker samples. and non-smoker samples.

104
ambient air
Signal (µV)

10-14
10
P1 sample
Signal (µV)

10-14
10
P2 sample
Signal (µV)

10-14
10
P3 sample
Signal (µV)

(smoker)

10-1
ni mo ide

ac ocy rile

cy ol
fo n e
de e

ac ide

m ls e

id
on ya r

dr e yde

s ol
hy et de

de e

rb ac ein

bu iso fide

m alde ne
ny e yde

l s de

py l nit e
bu ic a e
ric d
rb n c ate

ge xid

al xid

di ny ton
al id

hy hyd

ty ci
v t
so an
en an

ac
ru ra
e
et c ulfi

hy lfi
n

an
it
et an

ol
h

ul
ty pr
ca ge w

l i on
tro no

rm o

et u
l i th
og th

e
r
m

r
a

et
o
o

ca
dr

vi
hy
hy

Figure
Figure 24.24. Gases
Gases detected
detected inin exhaled
exhaled human
human breath.
breath. Adapted
Adapted from
from [180].
[183].

8.3.
8.3.Natural
NaturalGasGas
Natural gas
Natural gas is is
primarily
primarily a mixture
a mixture of of
three hydrocarbons:
three hydrocarbons: methane
methane (CH 4 ), 4ethane
(CH ), ethane (C(C2 H26H
),6),
andand
propane
propane (C(C
3 H3H
8 ),8),
the
theratio
ratioofofwhich
whichdetermines
determinesitsitscalorific
calorificcontent,
content,i.e., itsitsquality
i.e., quality asasfuel.
fuel.InInaddition,
addition,
it it
usually
usuallycontains
contains anan admixture
admixture ofof
some
somecarbon
carbonmonoxide
monoxide (CO)
(CO)and carbon
and carbon dioxide
dioxide (CO(CO2 ),2),
which
which are
are
contaminants. Therefore quality monitoring of natural gas is of some importance.
contaminants. Therefore quality monitoring of natural gas is of some importance. Unlike pollution Unlike pollution
andand breath
breathmonitoring,
monitoring, inin
this
thisapplication
applicationhigh
highsensitivity,
sensitivity,selectivity
selectivityandandspecificity
specificityare arenotnotrequired.
required.
TDS was used to demonstrate that THz measurements can establish
TDS was used to demonstrate that THz measurements can establish the mixing ratios the mixing ratios ofof
natural
natural
gas [184,185]. As seen in Figure 25, the refractive index provides a good indicator
gas [181,182]. As seen in Figure 25, the refractive index provides a good indicator of the mix, due of the mix, due toto
the differences in the indices of individual
the differences in the indices of individual gases. gases.
 Sensors 2019, 19, 4203 23 of 35

Sensors 2019, 19, 4203 23 of 35

Sensors 2019, 19, 4203 23 of 35
1.0010

Refractive index
1.0008
1.0010

Refractive index
1.0006
1.0008
1.0004

1:3:0:0

0:0:0:1

0:1:0:0

0.2:1.2:0:0

0:0:1:0

0.2:1.2:0.2:0

0.2:1.2:0.2:0.2

0.2:1.2:0.2:0.4

0.2:1.2:0.2:0.6
1.0006

1.0004

1:3:0:0

0:0:0:1

0:1:0:0

0.2:1.2:0:0

0:0:1:0

0.2:1.2:0.2:0

0.2:1.2:0.2:0.2

0.2:1.2:0.2:0.4

0.2:1.2:0.2:0.6
CO:CO2:C2H6:CH4 ratio

Figure 25. Refractive indices (mean over 0.3–1.3 THz) of natural gas constituents and their different
CO:CO2:C2H6:CH4 ratio
mixes. Adapted from [181].
Figure 25. Refractive indices (mean over 0.3–1.3 THz) of natural gas constituents and their different
Figure 25. Refractive
9. Paper indices (mean over 0.3–1.3 THz) of natural gas constituents and their different
mixes. and Wood
Adapted from [184].
mixes. Adapted from [181].
9. 9.1.
Paper and Wood
Paper
9. Paper and Wood
9.1. Paper Although all types of paper have high THz absorption, thin paper sheets have moderate-to-good
THz
9.1. Paper transparency. THz absorption by the paper is determined by its thickness, composition,
Although
moisture content all types
andoftexture,
paper have high THz
allowing theseabsorption,
propertiesthin
to bepaper sheets have
evaluated using moderate-to-good
THz transmission
THz Although all types of paper have high THz absorption, thin paper sheets
measurements. THz sensing, therefore, offers a solution to the paper industry’s need moisture
transparency. THz absorption by the paper is determined by its thickness, have moderate-to-good
composition, for in-line
THz
content transparency.
and texture, THz
allowingabsorption
these by the
properties paper
to be is determined
evaluated using by
THz its thickness,
transmission
monitoring of paper caliper (thickness), moisture content and basis weight (weight per area). As a composition,
measurements.
moisture
THz result, content
sensing, and texture,
therefore,
considerable offers
amount a allowing
of worktothese
solution properties
the paper
has been to be
industry’s
devoted evaluated
toneed using
for in-line
investigating THz
monitoring
the transmission
of paper
feasibility of this
measurements.
caliper (thickness),THz sensing,
moisture therefore,
content and offers
basis
application, including several industrial demonstrations. a solution
weight to
(weight theperpaper
area).industry’s
As a need
result, for in-line
considerable
monitoring
amount of paper
of work
Laboratory caliper
has been
studies (thickness),
devoted moisture
to investigating
demonstrated content
the
diagnosing and
feasibility basis
of thisweight
moisture (weight
application,
content per simultaneous
area).
including
[183], As a
several
result,
industrial considerable amount
demonstrations. of work has been devoted to investigating
measurements of thickness and moisture [184,185], and multiple parameter determination [186–190]. the feasibility of this
application,
Laboratoryincluding
studies several industrial
demonstrated demonstrations.
diagnosing moisture content [186],
Aiming at a more compact system, a THz fiber Bragg grating was used to measure paper thickness simultaneous measurements
of in Laboratory
thickness
[191]. The andareastudies
moisture demonstrated
[187,188],
was partially diagnosing
and multiple
reviewed in [192]. moisture
parameter
Monitoring the content
determination [183],
ageing of[189–193].
oil-papersimultaneous
Aiming
insulation at in
ameasurements
more compact of thickness
system, a and
THz moisture
fiber [184,185],
Bragg grating and
wasmultiple
used toparameter
measure determination
paper
transformers was investigated in [193]. Vassilev et al. used a purpose-built electronic system at 200 thickness [186–190].
in [194].
Aiming
The GHzareatoat
wasamonitor
more compact
partially system,
reviewed
moisture a in
THz
in [195].
levels fiber during
Monitoring
paper Braggthegrating
ageing
an was used to insulation
of oil-paper
offset printing measure demonstrating
process, paper thicknessan
in transformers
in industrial
was [191]. Theprototype
investigated areainwas partially
[196]. Vassilev
(Figure reviewed
26) et al. used
[194]. in a[192]. Monitoring
purpose-built the ageing
electronic of oil-paper
system at 200 GHz insulation
to monitor in
transformers
moisture was
levels ininvestigated
paper during inan[193]. Vassilev
offset printing et al. used ademonstrating
process, purpose-built an electronic
industrialsystem at 200
prototype
GHz to26)
(Figure monitor
[197]. moisture levels in paper during an offset printing process, demonstrating an
industrial prototype (Figure 26) [194].

Figure 26. An industrial prototype of an electronic sensor at 200 GHz monitoring paper moisture
during offset printing. Adapted from [197].
Figure 26. An industrial prototype of an electronic sensor at 200 GHz monitoring paper moisture
In during
a paper-related industrial
offset printing. demonstration,
Adapted from [194]. Brinkmann et al. used TDS for quality control of
cardboard boxes as used in pharmaceutical packaging [198]. Their setup combined a pulsed THz
Figure
emitter with26.a An industrial prototype
high-bandwidth of anreceiver
Schottky electronicand
sensor at 200 GHz monitoring
accomplished paper
THz intensity moisture
measurements
during offset printing. Adapted from [194].
with an effective time resolution of 6.4 µs. In a proof-of-principle experiment, they showed that THz
Sensors 2019,
Sensors 19,19,
2019, 4203
4203 2424
ofof
3535

InIna apaper-related
paper-relatedindustrial
industrialdemonstration,
demonstration,Brinkmann
Brinkmannetetal.al.used
usedTDS
TDSfor forquality
qualitycontrol
controlofof
cardboard
cardboard
Sensors boxes
2019, 19,boxes as used in pharmaceutical packaging [195]. Their setup
4203 as used in pharmaceutical packaging [195]. Their setup combined a pulsed combined a pulsed24THz
ofTHz
35
emitter
emitterwithwitha ahigh-bandwidth
high-bandwidthSchottky
Schottkyreceiver
receiverandandaccomplished
accomplishedTHz THzintensity
intensitymeasurements
measurements
with
withananeffective
effectivetime
timeresolution
resolutionofof6.4
6.4μs.
μs.InIna aproof-of-principle
proof-of-principleexperiment,
experiment,they theyshowed
showedthat
thatTHz
THz
screening
screening
screeningwaswas able
wasable to detect
abletotodetect the
detectthe presence
thepresence or absence
presenceororabsence of package
absenceofofpackage inserts
packageinserts inside
insertsinside the
insidethe cardboard
cardboardboxes
thecardboard boxes
boxes
unambiguously,
unambiguously,
unambiguously, even
even
evenfor
for samples
forsamples
samples moving
moving
moving at more
atatmore than
morethan 20
than m
2020 per
mmperss(Figure
per s(Figure
(Figure27).
27).
27).
2020

1515

RMS Signal (dB)

RMS Signal (dB)
1010

55

00
00 55 1010 1515 2020 2525
Time (ms)
Time (ms)

Figure 27.
Figure27.
Figure 27.(a) Boxes
(a)(a)Boxes arranged
Boxesarranged
arrangedonononaa afast
fast turntable.
turntable.(b)
fastturntable. (b)1D-scan
(b) 1D-scanof
1D-scan ofoffolded
foldedcardboard
folded boxes
cardboardboxes
cardboard with
boxeswith and
withand
and
without
without
without aa package
package
a package slip
slip (shown
slip (shown
(shown inin
in red).
red).
red). TheThe
The graph depicts
graph
graph depicts
depicts the transmitted
the
the transmitted
transmitted THz
THz
THz intensity with
intensity
intensity the
with
with boxes
the
the boxes
boxes
moving
moving
moving atat
at 2121
21 m/s.
m/s.Adapted
m/s. from
Adaptedfrom
Adapted [195].
from[198].
[195].

9.2.
9.2. Wood
9.2.Wood
Wood
Wood
Woodhas
Wood hasalso
has alsobeen
also beeninvestigated
been investigatedwith
investigated withTHz
with THzspectroscopy,
THz spectroscopy,although
spectroscopy, althoughit
although ititis
isisless
lesstransparent
less transparentto
transparent totoTHz
THz
THz
waves
waves than
than paper
paper due
due in
in part
part to
to the
the larger
larger thickness
thickness of
of samples
samples and
and a
a
waves than paper due in part to the larger thickness of samples and a stronger contribution fromstronger
stronger contribution
contribution from
from
scattering.
scattering.Due
scattering. Dueto
Due totoits anisotropic
itsitsanisotropic
anisotropic structure
structure
structureconsisting
consisting
consisting ofofof
aligned
aligned fibers,
aligned fibers,wood
wood
fibers, wood isisbirefringent, making
isbirefringent,
birefringent, making
makingit
possible to
ititpossible detect orientation
possibletotodetect
detectorientation [199–201].
orientation[196–198]. Similar
[196–198].Similar to paper,
Similartotopaper, moisture
paper,moisturecontent
moisturecontent and
contentand density of
anddensitywood have
densityofofwood
wood
also
have been measured [202,203]. Krügener et al. studied THz optical
have also been measured [199,200]. Krügener et al. studied THz optical properties ofseveral
also been measured [199,200]. Krügener et al. studied THz properties
optical properties of several
of species
several of
species
species
wood
ofofwoodwith
woodwitha view to
witha aview identifying
viewtotoidentifying and evaluating
identifyingand wood
andevaluating samples
evaluatingwood [204].
woodsamples They
samples[201]. also
[201].They used
Theyalso a THz reflection
alsoused
useda aTHzTHz
system
reflectionwith a robotic
system with arm
a to
robotic scan
armwood
to logs
scan in
wood order
logs to
inreveal
order areas
to underneath
reveal areas
reflection system with a robotic arm to scan wood logs in order to reveal areas underneath the bark the bark
underneath that
the had
bark
been
that damaged
thathad
hadbeen by burrowing
beendamaged
damaged “typographer”
bybyburrowing
burrowing beetles (Figure
“typographer”
“typographer” beetles 28)
beetles [205]. 27)
(Figure
(Figure 27)[202].
[202].

Figure
Figure28.
28.THz
THzscans
scansoverlaid
overlaidononCT
CTimages,
images,revealing beetle
beetle
revealing damage.
damage.a)a)Healthy
damage.
beetle Healthywood;
wood;b)b)infested
infested
wood.Reprinted
wood. Reprintedwith
withpermission
permissionfrom
from[202]
[205] ©©The
[202] TheOptical
OpticalSociety.
Society.

10.
10. Conclusions
10.Conclusions
Conclusions
The
The often-quoted phrase “THz moves out of the lab” lab”
finally seemsseems
to be coming true. The authors
Theoften-quoted
often-quotedphrase
phrase“THz “THzmoves
movesout outofofthethe lab”finally
finally seemstotobebecoming
comingtrue. true.The
The
still remember
authors still panel
remember discussions
panel at THz
discussions conferences
at THz about
conferences a decade
about ago,
a when
decade none
ago, of
when the attending
none of the
authors still remember panel discussions at THz conferences about a decade ago, when none of the
experts
attending would have been ablebeento name to but a single “killer “killer
application”. Today,Today, however, THz
attendingexperts
expertswould
wouldhavehave beenable able toname
namebut buta asingle
single “killerapplication”.
application”. Today,however, however,
technologies
THz show a significant market potential, a finding that is confirmed by a number of
THztechnologies
technologiesshow showa asignificant
significantmarket
marketpotential,
potential,a afinding
findingthat
thatisisconfirmed
confirmedbybya anumbernumberofof
recent
recent market reports
reports [206–210].Admittedly,
Admittedly, the term “killer application” remains questionable,
recentmarket
market reports[203–207].
[203–207]. Admittedly,the theterm
term“killer
“killerapplication”
application”remains
remainsquestionable,
questionable,yet yet
yet applications especially
applications in the field of non-destructive,non-contact
non-contact testingnow now successfully exploit exploit
applicationsespecially
especiallyininthe
thefield
fieldofofnon-destructive,
non-destructive, non-contacttesting testing nowsuccessfully
successfully exploit
the
the abilities of THz radiation, i.e., penetration of many opaque materials, depth resolution on the
theabilities
abilitiesofofTHzTHzradiation,
radiation,i.e.,
i.e.,penetration
penetrationofofmanymanyopaque
opaquematerials,
materials,depth
depthresolution
resolutionononthe the
micrometer
micrometer level, high spatial resolution as compared to microwaves, to name but a few. Industrial
micrometerlevel, level,high
highspatial
spatialresolution
resolutionasascompared
comparedtotomicrowaves,
microwaves,totoname namebut
buta afew.
few.Industrial
Industrial
applications
applications are made possible by the availability of more advanced, yet robust and long-lived
applicationsare aremade
madepossible
possiblebybythe theavailability
availabilityofofmoremoreadvanced,
advanced,yet yetrobust
robustand andlong-lived
long-lived
systems,
systems, and
and benefit
benefit from
from more
more competition
competition between
between manufacturers,
manufacturers, which
which has
has led
led to
to aa significant
significant
systems, and benefit from more competition between manufacturers, which has led to a significant
price drop for THz instrumentation. Whilst Hochrein wrote in 2014, “The average system price is
approx. 134,000 EUR” [6], most of today’s TDS “base systems” cost about 1.5 to 2 times less. From the
 Sensors 2019, 19, 4203 25 of 35

Sensors 2019, 19, 4203

price drop for THz instrumentation. Whilst Hochrein wrote in 2014, “The average system25price of 35
is
approx. 134,000 EUR” [6], most of today’s TDS “base systems” cost about 1.5 to 2 times less. From the
perspective
perspective of anof an industrial
industrial user,
user, inexpensive
inexpensive systems
systems amortize
amortize faster
faster thanthan costly
costly ones,
ones, and and perhaps
perhaps
forfor
thethe first
first timetime in in
the the history
history ofofTHz
THztechnologies,
technologies,an aninvestment
investmentin inTHz
THzinstrumentation
instrumentation results results in
actual cost savings: thinner walls of plastic pipes or bottles reduce
in actual cost savings: thinner walls of plastic pipes or bottles reduce material usage and wastage material usage and wastage (Section
3); a fully automated, robot-driven system for paint layer measurements
(Section 3); a fully automated, robot-driven system for paint layer measurements reduces production reduces production times in
the in
times automotive
the automotive industry industry(Section 4); an4);optimum
(Section an optimum tablet coating
tablet maymay
coating improve
improve the the
efficacy
efficacyof aofdrug
a
and therefore increase its chance of success in clinical trials (Section
drug and therefore increase its chance of success in clinical trials (Section 5); an early identification 5); an early identification of faults
in electronic
of faults structures
in electronic structuresenhances the yield
enhances and helps
the yield reduce
and helps failures
reduce in theinfield
failures (Section
the field 6). These
(Section 6).
examples are not just hypothetical but are based on industrial installations
These examples are not just hypothetical but are based on industrial installations that the authors are that the authors are aware
of. Readers
aware of. Readers may may refer refer
to thetowebsites
the websites of system
of systemmanufacturers
manufacturers as listed in the in
as listed references of Sections
the references of
3–6—the
Sections most advanced
3–6—the most advanced markets at the at
markets time
theof writing—for
time of writing—forfurther details.
further details.
ForFor other
other industries,
industries, including
including paper,
paper, wood,wood, butbutalsoalso petrochemicals,
petrochemicals, numerous
numerous publications
publications
have shown the technological potential of THz instrumentation,
have shown the technological potential of THz instrumentation, but apart from initial tests, market but apart from initial tests, market
acceptance
acceptance hashas been been slow—either
slow—either because
because alternative
alternative and and established
established methods
methods exist,
exist, or or because
because
markets are more price-sensitive. In fact, any THz system manufacturer
markets are more price-sensitive. In fact, any THz system manufacturer will have received customer will have received customer
inquiries
inquiries suchsuch
as “I asneed“I needa THz a THz
sensor,sensor,
but itbut
must it must
not costnotmore
cost than
more10,000
than EUR”,
10,000 and EUR”,thisand
pricethis price
level
level remains
remains out of reach out today.
of reach today.
This may Thischange may in change
the future,in the
e.g.,future, e.g., with miniaturized
with miniaturized on-chip systems, on-chip
systems, but a large amount of work remains to be done, which
but a large amount of work remains to be done, which still creates something of a “chicken-and-egg” still creates something of a “chicken-
and-egg”
problem. In problem.
the special Incase
the special
of gases, case of gases, a commercial
a commercial instrument was instrument
developed was more
developed
than 10 more
yearsthan
10 years ago [164], but for reasons not known to the authors,
ago [167], but for reasons not known to the authors, the project was abandoned and sales seem to have the project was abandoned and sales
seem
been to have been discontinued.
discontinued.
TheThe totaltotal THz market
THz market volumevolume seemstodifficult
seems difficult quantify. to It isquantify.
noteworthy It that
is noteworthy
the aforementioned that the
market reports differ in their estimate for 2019 by more than a factor of 4 (Figure 29). According to29).
aforementioned market reports differ in their estimate for 2019 by more than a factor of 4 (Figure
According
Tematys, to Tematys,
a French research acompany, French research
the totalcompany,
market is stillthe significantly
total market below is still 100
significantly
Million US$ below
[209].100
OnMillion
the other US$end [206].
of the On the other
spectrum areend of the spectrum
Beijing-based are Beijing-based
QY Research, who believe QYinResearch,
present marketwho believe
size of in
385 Million US$ [206]. Similarly, predictions for the compound annual growth rate (CAGR) differgrowth
present market size of 385 Million US$ [203]. Similarly, predictions for the compound annual by a
rateof(CAGR)
factor two, with differ
valuesby aranging
factor of two,
from 16%with values ranging
(Tematys) to almost from
32%16%(QY(Tematys)
Research). to almost
Taking 32% (QY
a simple
Research).
arithmetic Taking
average of athe simple arithmetic
five reports, average
the global of theamounts
market five reports, the global market
to approximately 230 MillionamountsUS$ to
approximately 230
and the CAGR is roughly 25%. Million US$ and the CAGR is roughly 25%.

Global Terahertz Market

1500
QY Research (CN)
Market Size (Million US$)

Research and Markets (IE)

Markets and Markets (IN)
Coherent Market Insights (USA)
1000 Tematys (FR)

500

0
2016 2018 2020 2022 2024
Year

Figure 29. Global terahertz market in Million USD, according to five different analyst firms [206–210].
Figure 29. Global terahertz market in Million USD, according to five different analyst firms [203–207].
Nonetheless, all of the reports agree that the market for THz technologies will continue to expand,
Nonetheless, all of the reports agree that the market for THz technologies will continue to
and even the most “pessimistic” prediction of a CAGR of 16% still indicates a significant growth. It
expand, and even the most “pessimistic” prediction of a CAGR of 16% still indicates a significant
remains to be seen which applications will have won the broadest industry acceptance by the time of
growth. It remains to be seen which applications will have won the broadest industry acceptance by
the next review.
the time of the next review.
Funding: This research received no external funding.
Funding: This research received no external funding.
Conflicts of Interest: The authors declare no conflict of interest.
Sensors 2019, 19, 4203 26 of 35

References
1. Gartner Hype Cycle. Available online: https://www.gartner.com/en/research/methodologies/gartner-hype-
cycle (accessed on 27 September 2019).
2. Siegel, P.H. THz technology: An overview. In Terahertz Sensing Technology: Volume 1: Electronic Devices and
Advanced Systems Technology; World Scientific Publishing Co Pte Ltd: Singapore, 2003; pp. 1–44.
3. Schmuttenmaer, C.A. Exploring dynamics in the far-infrared with terahertz spectroscopy. Chem. Rev. 2004,
104, 1759–1780. [CrossRef] [PubMed]
4. Stöhr, A. Pushing the boundaries. IEEE Microw. Mag. 2009, 10, 106–115. [CrossRef]
5. Armstrong, C.M. The truth about terahertz. IEEE Spectr. 2012, 49, 36–41. [CrossRef]
6. Hochrein, T. Markets, availability, notice, and technical performance of terahertz systems: Historic
development, present, and trends. J. Infrared Millim. Terahertz Waves 2015, 36, 235–254. [CrossRef]
7. Top 50 Emerging Technologies: Growth Opportunities of Strategic Imperative. Available online: https:
//ww3.frost.com/files/6014/7973/5474/Top_50_Emerging_Technologies.pdf (accessed on 27 September 2019).
8. Neu, J.; Schmuttenmaer, C.A. Tutorial: An introduction to terahertz time domain spectroscopy (THz-TDS).
J. Appl. Phys. 2018, 124, 231101. [CrossRef]
9. Guerboukha, H.; Nallappan, K.; Skorobogatiy, M. Toward real-time terahertz imaging. Adv. Opt. Photonics
2018, 10, 843–938. [CrossRef]
10. Jepsen, P.U.; Cooke, D.G.; Koch, M. Terahertz spectroscopy and imaging–Modern techniques and applications.
Laser Photonics Rev. 2011, 5, 124–166. [CrossRef]
11. Duvillaret, L.; Garet, F.; Coutaz, J.-L. A reliable method for extraction of material parameters in terahertz
time-domain spectroscopy. IEEE J. Sel. Top. Quantum Electron. 1996, 2, 739–746. [CrossRef]
12. Withayachumnankul, W.; Ferguson, B.; Rainsford, T.; Mickan, S.P.; Abbott, D. Material parameter extraction
for terahertz time-domain spectroscopy using fixed-point iteration. Proceedings of Photonic Materials,
Devices, and Applications, Sevilla, Spain, 7 July 2005; pp. 221–231.
13. Pupeza, I.; Wilk, R.; Koch, M. Highly accurate optical material parameter determination with THz time-domain
spectroscopy. Opt. Express 2007, 15, 4335–4350. [CrossRef]
14. Jepsen, P.U. Phase Retrieval in Terahertz Time-Domain Measurements: A “how to” Tutorial. J. Infrared
Millim. Terahertz Waves 2019, 40, 395–411. [CrossRef]
15. Vieweg, N.; Rettich, F.; Deninger, A.; Roehle, H.; Dietz, R.; Göbel, T.; Schell, M. Terahertz-time domain
spectrometer with 90 dB peak dynamic range. J. Infrared Millim. Terahertz Waves 2014, 35, 823–832. [CrossRef]
16. Wilmink, G.J.; Ibey, B.L.; Rivest, B.D.; Grundt, J.E.; Roach, W.P.; Tongue, T.D.; Schulkin, B.J.; Laman, N.;
Peralta, X.G.; Roth, C.C.; et al. Development of a compact terahertz time-domain spectrometer for the
measurement of the optical properties of biological tissues. J. Biomed. Opt. 2011, 16, 047006. [CrossRef]
[PubMed]
17. Molter, D.; Ellrich, F.; Weinland, T.; George, S.; Goiran, M.; Keilmann, F.; Beigang, R.; Léotin, J. High-speed
terahertz time-domain spectroscopy of cyclotron resonance in pulsed magnetic field. Opt. Express 2010,
18, 26163–26168. [CrossRef]
18. Bartels, A.; Thoma, A.; Janke, C.; Dekorsy, T.; Dreyhaupt, A.; Winnerl, S.; Helm, M. High-resolution THz
spectrometer with kHz scan rates. Opt. Express 2006, 14, 430–437. [CrossRef] [PubMed]
19. Kim, Y.; Yee, D.-S. High-speed terahertz time-domain spectroscopy based on electronically controlled optical
sampling. Opt. Lett. 2010, 35, 3715–3717. [CrossRef] [PubMed]
20. Yahyapour, M.; Jahn, A.; Dutzi, K.; Puppe, T.; Leisching, P.; Schmauss, B.; Vieweg, N.; Deninger, A. Fastest
Thickness Measurements with a Terahertz Time-Domain System based on Electronically Controlled Optical
Sampling. Appl. Sci. 2019, 9, 1283. [CrossRef]
21. Preu, S.; Döhler, G.H.; Malzer, S.; Wang, L.J.; Gossard, A.C. Tunable, continuous-wave terahertz photomixer
sources and applications. Appl. Phys. Rev. 2011, 109, 061301. [CrossRef]
22. Deninger, A. State-of-the-art in terahertz continuous wave photomixer systems. In Handbook of Terahertz
Technology; Saeedkia, D., Ed.; Woodhead Publishing: Cambridge, UK, 2013.
23. Kong, D.-Y.; Wu, X.-J.; Wang, B.; Gao, Y.; Dai, J.; Wang, L.; Ruan, Cu.; Miao, J.-G. High resolution
continuous wave terahertz spectroscopy on solid-state samples with coherent detection. Opt. Express 2018,
26, 17964–17976. [CrossRef] [PubMed]
Sensors 2019, 19, 4203 27 of 35

24. Deninger, A.J.; Roggenbuck, A.; Schindler, S.; Preu, S. 2.75 THz tuning with a triple-DFB laser system at
1550 nm and InGaAs photomixers. J. Infrared Millim. Terahertz Waves 2015, 36, 269–277. [CrossRef]
25. Vogt, D.W.; Leonhardt, R. Ultra-high Q terahertz whispering-gallery modes in a silicon resonator.
Appl. Photonics 2018, 3, 051702. [CrossRef]
26. Roggenbuck, A.; Schmitz, H.; Deninger, A.; Mayorga, I.C.; Hemberger, J.; Güsten, R.; Grüninger, M.
Coherent broadband continuous-wave terahertz spectroscopy on solid-state samples. New J. Phys. 2010,
12, 43017–43029. [CrossRef]
27. Vogt, D.W.; Leonhardt, R. High resolution terahertz spectroscopy of a whispering gallery mode bubble
resonator using Hilbert analysis. Opt. Express 2017, 25, 16860. [CrossRef] [PubMed]
28. Mittleman, D.M. Perspective: Terahertz science and technology. J. Appl. Phys. 2017, 122, 230901. [CrossRef]
29. Mittleman, D.M. Twenty years of terahertz imaging. Opt. Express 2018, 26, 9417–9431. [CrossRef] [PubMed]
30. Sengupta, K.; Nagatsuma, T.; Mittleman, D.M. Terahertz integrated electronic and hybrid electronic–photonic
systems. Nat. Electron. 2018, 1, 622. [CrossRef]
31. Terasense. Available online: https://terasense.com/products/sub-thz-imaging-cameras/ (accessed on
27 September 2019).
32. Ticwave. Available online: https://ticwave.com/index.php/shop?filter_catid=37&search=&sortby= (accessed
on 27 September 2019).
33. i2s. Available online: https://www.i2s.fr/en/product/tzcam (accessed on 27 September 2019).
34. INO. Available online: https://www.ino.ca/en/solutions/microxcam-384i-thz/ (accessed on 27 September 2019).
35. Fischer, B.M.; Wietzke, S.; Reuter, M.; Peters, O.; Gente, R.; Jansen, C.; Vieweg, N.; Koch, M. Investigating
material characteristics and morphology of polymers using terahertz technologies. IEEE Trans. Terahertz
Sci. Technol. 2013, 3, 259–268. [CrossRef]
36. Wietzke, S.; Jansen, C.; Reuter, M.; Jung, T.; Kraft, D.; Chatterjee, S.; Fischer, B.M.; Koch, M. Terahertz
spectroscopy on polymers: A review of morphological studies. J. Mol. Struct. 2011, 1006, 41–51. [CrossRef]
37. Krumbholz, N.; Hochrein, T.; Vieweg, N.; Hasek, T.; Kretschmer, K.; Bastian, M.; Mikulics, M.; Koch, M.
Monitoring polymeric compounding processes inline with THz time-domain spectroscopy. Polym. Test.
2009, 28, 30–35. [CrossRef]
38. Yakovlev, E.V.; Zaytsev, K.I.; Dolganova, I.N.; Yurchenko, S.O. Non-destructive evaluation of polymer
composite materials at the manufacturing stage using terahertz pulsed spectroscopy. IEEE Trans. Terahertz
Sci. Technol. 2015, 5, 810–816. [CrossRef]
39. Busch, S.F.; Castro-Camus, E.; Beltran-Mejia, F.; Balzer, J.; Koch, M. 3D printed prisms with tunable dispersion
for the THz frequency range. J. Infrared Millim. Terahertz Waves 2018, 39, 553–560. [CrossRef]
40. Scherger, B.; Born, N.; Jansen, C.; Schumann, S.; Koch, M.; Wiesauer, K. Compression molded terahertz
transmission blaze-grating. IEEE Trans. Terahertz Sci. Technol. 2012, 2, 556–561. [CrossRef]
41. Kaji, T.; Tominari, Y.; Yamada, T.; Saito, S.; Morohashi, I.; Otomo, A. Terahertz-wave generation devices using
electro-optic polymer slab waveguides and cyclo-olefin polymer clads. Opt. Express 2018, 26, 30466–30475.
[CrossRef] [PubMed]
42. Vieweg, N.; Deninger, A.; Leisching, P. True OEM terahertz systems for industrial applications. In Photonic
Instrumentation Engineering VI; International Society for Optics and Photonics, SPIE: San Francisco, CA, USA,
4 March 2019; p. 109250U.
43. INOEX: The Future of Extrusion. Available online: https://www.inoex.de (accessed on 27 September 2019).
44. Fischer Worldwide. Available online: https://www.helmut-fischer.de (accessed on 27 September 2019).
45. Hauck, J. Inspektion an Kunststoffbauteilen Mit Terahertz-Zeitbereichsspektroskopie; SKZ-Das Kunststoff-Zentrum:
Lulu, Berlin, Germany, 2014; ISBN 978-1-326-11323-0.
46. Zhang, X.; Guo, Q.; Chang, T.; Cui, H.-L. Broadband stepped-frequency modulated continuous terahertz
wave tomography for non-destructive inspection of polymer materials. Polym. Test. 2019, 76, 455–463.
[CrossRef]
47. Toptica Photonics. Available online: https://www.toptica.com/products/terahertz-systems/time-domain/
teraflash-pro/ (accessed on 27 September 2019).
48. Sommer, S.; Raidt, T.; Fischer, B.M.; Katzenberg, F.; Tiller, J.C.; Koch, M. THz-spectroscopy on high density
polyethylene with different crystallinity. J. Infrared Millim. Terahertz Waves 2016, 37, 189–197. [CrossRef]
Sensors 2019, 19, 4203 28 of 35

49. Engelbrecht, S.; Tybussek, K.; Sampaio, J.; Böhmler, J.; Fischer, B.M.; Sommer, S. Monitoring the Isothermal
Crystallization Kinetics of PET-A Using THz-TDS. J. Infrared Millim. Terahertz Waves 2019, 40, 306–313.
[CrossRef]
50. Wietzke, S.; Jansen, C.; Jung, T.; Reuter, M.; Baudrit, B.; Bastian, M.; Chatterjee, S.; Koch, M. Terahertz
time-domain spectroscopy as a tool to monitor the glass transition in polymers. Opt. Express 2009,
17, 19006–19014. [CrossRef] [PubMed]
51. Wietzke, S.; Jansen, C.; Rutz, F.; Mittleman, D.M.; Koch, M. Determination of additive content in polymeric
compounds with terahertz time-domain spectroscopy. Polym. Test. 2007, 26, 614–618. [CrossRef]
52. Krumbholz, N.; Hochrein, T.; Vieweg, N.; Radovanovic, I.; Pupeza, I.; Schubert, M.; Kretschmer, K.; Koch, M.
Degree of dispersion of polymeric compounds determined with terahertz time-domain spectroscopy.
Polym. Eng. Sci. 2011, 51, 109–116. [CrossRef]
53. Jördens, C.; Wietzke, S.; Scheller, M.; Koch, M. Investigation of the water absorption in polyamide and wood
plastic composite by terahertz time-domain spectroscopy. Polym. Test. 2010, 29, 209–215. [CrossRef]
54. Jördens, C.; Scheller, M.; Wietzke, S.; Romeike, D.; Jansen, C.; Zentgraf, T.; Wiesauer, K.; Reisecker, V.; Koch, M.
Terahertz spectroscopy to study the orientation of glass fibres in reinforced plastics. Compos. Sci. Technol.
2010, 70, 472–477. [CrossRef]
55. Okano, M.; Watanabe, S. Inspection of internal filler alignment in visibly opaque carbon-black–rubber
composites by terahertz polarization spectroscopy in reflection mode. Polym. Test. 2018, 72, 196–201.
[CrossRef]
56. Rutz, F.; Hasek, T.; Koch, M.; Richter, H.; Ewert, U. Terahertz birefringence of liquid crystal polymers.
Appl. Phys. Lett. 2006, 89, 221911. [CrossRef]
57. Mayr, M.; Giovanni, G.; Baudrit, B.; Hochrein, T.; Bastian, M. Strukturanalyse in Der Polymerschaumextrusion;
SKZ-Das Kunststoffzentrum; Shaker Verlag Aachen: Dylan, Germany, 2019; ISBN 978-3-8440-6480-3.
58. Werner, M.; Kolb, C.; Schober, G.; Kremling, S. Charakterisierung Von Polymerschäumen Mittels Zeitaufgelöster
Terahertz-Spektroskopie; DGZfP-Jahrestagung: Berlin, Germany, 2017.
59. Jansen, C.; Wietzke, S.; Wang, H.; Koch, M.; Zhao, G. Terahertz spectroscopy on adhesive bonds. Polym. Test.
2011, 30, 150–154. [CrossRef]
60. Stübling, E.; Gomell, L.; Sommer, S.; Winkel, A.; Kahlmeyer, M.; Böhm, S.; Koch, M. THz Properties of
Adhesives. J. Infrared Millim. Terahertz Waves 2018, 39, 586–593. [CrossRef]
61. Sommer, S.; Probst, T.; Kraus, E.; Baudrit, B.; Town, G.E.; Koch, M. Cure monitoring of two-component epoxy
adhesives by terahertz time-domain spectroscopy. Polym. Sci. Ser. B 2016, 58, 769–776. [CrossRef]
62. Probst, T.; Sommer, S.; Soltani, A.; Kraus, E.; Baudrit, B.; Town, G.E.; Koch, M. Monitoring the polymerization
of two-component epoxy adhesives using a terahertz time domain reflection system. J. Infrared Millim.
Terahertz Waves 2015, 36, 569–577. [CrossRef]
63. Rettich, F.; Vieweg, N.; Cojocari, O.; Deninger, A. Field intensity detection of individual terahertz pulses at
80 MHz repetition rate. J. Infrared Millim. Terahertz Waves 2015, 36, 607–612. [CrossRef]
64. fmi Future Marketing Insights. Available online: https://www.futuremarketinsights.com (accessed on
27 September 2019).
65. White, J.; Fichter, G.; Chernovsky, A.; Zimdars, D.; Whitaker, J.F.; Das, D.; Pollock, T.M. Time-domain terahertz
mapping of thickness and degradation of aircraft turbine blade thermal barrier coatings. In Proceedings of
the Conference on Lasers and Electro-Optics and Conference on Quantum Electronics and Laser Science, San
Jose, CA, USA, 4–9 May 2008.
66. Yasui, T.; Yasuda, T.; Sawanaka, K.; Araki, T. Terahertz paint meter for noncontact monitoring of thickness
and drying progress in paint film. Appl. Opt. 2005, 44, 6849. [CrossRef]
67. Picot, M.; Ballacey, H.; Guillet, J.P.; Cassar, Q.; Mounaix, P. Terahertz Paint Thickness Measurements:
from lab to automotive and aeronautics industry. In Proceedings of the 15th Asia Pacific Conference for
Non-Destructive Testing (APCNDT2017), Singapore, 13–17 November 2017.
68. Su, K.; Shen, Y.; Zeitler, J.A. Terahertz Sensor for Non-Contact Thickness and Quality Measurement of
Automobile Paints of Varying Complexity. IEEE Trans. Terahertz Sci. Technol. 2014, 4, 432–439. [CrossRef]
69. das.nano. Available online: https://www.das-nano.com/ (accessed on 27 September 2019).
70. Jonuscheit, J. Terahertz Waves for Thickness Analyses. Opt. Photonik 2016, 1, 30–33. [CrossRef]
71. Fraunhofer HHI. Available online: https://www.hhi.fraunhofer.de (accessed on 27 September 2019).
72. LUNA Defying Impossible. Available online: https://www.Lunainc.com (accessed on 27 September 2019).
Sensors 2019, 19, 4203 29 of 35

73. Van Mechelen, J.L.M.; Kuzmenko, A.B.; Merbold, H. High precision material characterization method using
THz spectroscopy. In Proceedings of the 39th International Conference on Infrared, Millimeter, and Terahertz
waves (IRMMW-THz), Tucson, AZ, USA, 14–19 September 2014.
74. Kim, K.H.I.S.K.; Hsu, D.K.; Jung, J.A. Coating Thickness Characterization of Composite Materials Using
Terahertz Waves. Mater. Sci. Forum 2017, 878, 70–73.
75. Krimi, S.; Jonuscheit, J.K.J.; Freymann, G.V.; Urbansky, R.; Beigang, R. Highly accurate thickness measurement
of multi-layered automotive paints using terahertz technology. Appl. Phys. Lett. 2016, 109, 021105. [CrossRef]
76. Jen, C.-Y.; Richter, C. Sample thickness measurement with THz-TDS: Resolution and implications. J. Infrared
Millim. Terahertz Waves 2014, 35, 840–859. [CrossRef]
77. Scheller, M.A.; Koch, M. Fast and accurate thickness determination of unknown materials using terahertz
time domain spectroscopy. J. Infrared Millim. Terahertz Waves 2009, 30, 762–769. [CrossRef]
78. Liebelt, L.; Weber, S.; Klier, J.; Pfeiffer, T.; Molter, D.; Ellrich, F.; Jonuscheit, J.; Freymann, G.V. Influence of
Bandwidth and Dynamic Range on Layer-Thickness Determination Using Terahertz Time-Domain Spectroscopy;
French-German THz Conference: Kaiserlautern, Germany, 2019.
79. GOh, H.; Park, D.-S.; Kim, D.; Kim, H.-W. Real time measurement based on terahertz time-domain
spectroscopy for chip-top epoxy molding compound in semiconductor. In Proceedings of the 43rd
International Conference on Infrared, Millimeter, and Terahertz Waves (IRMMW-THz), Nagoya, Japan, 9–14
September 2018.
80. Pfeiffer, T.; Weber, S.; Klier, J.; Bachtler, S.; Molter, D.; Jonuscheit, J.; Freymann, G.V. Terahertz thickness
determination with interferometric correction for industrial applications. Opt. Express 2018, 26, 12558.
[CrossRef] [PubMed]
81. Exter, M.V.; Fattinger, C.; Grischkowsky, D. Terahertz time-domain spectroscopy of water. Opt. Lett. 1989,
14, 1128–1130. [CrossRef]
82. Withayachumnankul, W.; Fischer, B.M.; Abbott, D. Numerical removal of water vapour effects from terahertz
time-domain spectroscopy measurements. Proc. R. Soc. A 2008, 464, 2435–2456. [CrossRef]
83. Wang, Y.; Chen, Z.; Zhao, Z.; Zhang, L.; Kang, K.; Zhang, Y. Restoration of terahertz signals distorted by
atmospheric water vapor absorption. J. Appl. Phys. 2009, 105, 103105. [CrossRef]
84. Shen, Y.-C. Terahertz pulsed spectroscopy and imaging for pharmaceutical applications: A review.
Int. J. Pharm. 2011, 417, 48–60. [CrossRef]
85. Calvo, N.L.; Maggio, R.M.; Kaufman, T.S. Chemometrics-assisted solid-state characterization of
pharmaceutically relevant materials. Polymorphic substances. J. Pharm. Biomed. Anal. 2018, 147, 518–537.
[CrossRef]
86. Ruggiero, M.T.; Sibik, J.; Zeitler, J.A.; Korter, T.M. Examination of L-Glutamic Acid Polymorphs by Solid-State
Density Functional Theory and Terahertz Spectroscopy. J. Phys. Chem. A 2016, 120, 7490–7495. [CrossRef]
[PubMed]
87. Du, Y.; Xue, J. Investigation of polymorphism and cocrystallization of active pharmaceutical ingredients
using vibrational spectroscopic techniques. Curr. Pharm. Des. 2016, 22, 4917–4928. [CrossRef] [PubMed]
88. Takeuchi, I.; Tomoda, K.; Nakajima, T.; Terada, H.; Kuroda, H.; Makino, K. Estimation of crystallinity of
trehalose dihydrate microspheres by usage of terahertz time-domain spectroscopy. J. Pharm. Sci. 2012,
101, 3465–3472. [CrossRef] [PubMed]
89. Sasaki, T.; Sakamoto, T.; Otsuka, M. Detection of Impurities in Organic Crystals by High-Accuracy Terahertz
Absorption Spectroscopy. Anal. Chem. 2018, 90, 1677–1682. [CrossRef] [PubMed]
90. Sibik, J.; Lóbmann, K.; Rades, T.; Zeitler, J.A. Predicting crystallization of amorphous drugs with terahertz
spectroscopy. Mol. Pharm. 2015, 12, 3062–3068. [CrossRef] [PubMed]
91. Fitzgerald, A.J.; Cole, B.E.; Taday, P.F. Nondestructive analysis of tablet coating thicknesses using terahertz
pulsed imaging. J. Pharm. Sci. 2005, 94, 177–183. [CrossRef]
92. Haaser, M.; Gordon, K.C.; Strachan, C.J.; Rades, T. Terahertz pulsed imaging as an advanced characterisation
tool for film coatings—A review. Int. J. Pharm. 2013, 457, 510–520. [CrossRef]
93. TeraView Pharmaceutical. Available online: https://teraview.com/Pharmaceutical/ (accessed on 27 September
2019).
94. Ho, L.; Müller, R.; Gordon, K.C.; Kleinebudde, P.; Pepper, M.; Rades, T.; Shen, Y.; Taday, P.F.; Zeitler, J.A.
Terahertz pulsed imaging as an analytical tool for sustained-release tablet film coating. Eur. J. Pharm.
Biopharm. 2009, 71, 117–123. [CrossRef]
Sensors 2019, 19, 4203 30 of 35

95. Dohi, M.; Momose, W.; Yoshino, H.; Hara, Y.; Yamashita, K.; Hakomori, T.; Sato, S.; Terada, K. Application of
terahertz pulse imaging as PAT tool for non-destructive evaluation of film-coated tablets under different
manufacturing conditions. J. Pharm. Biomed. Anal. 2016, 119, 104–113. [CrossRef]
96. Lin, H.; May, R.K.; Evans, M.J.; Zhong, S.; Gladden, L.F.; Shen, Y.; Zeitler, J.A. Impact of processing conditions
on inter-tablet coating thickness variations measured by terahertz in-line sensing. J. Pharm. Sci. 2015,
104, 2513–2522. [CrossRef]
97. Chang, T.; Guo, Q.; Liu, L.; Cui, H.-L. Nondestructive Thickness Inspection of Capsule Coating by Terahertz
Time-Domain Spectroscopy. IEEE Trans. Terahertz Sci. Technol. 2018, 8, 688–695. [CrossRef]
98. Ervasti, T.; Silfsten, P.; Ketolainen, J.; Peiponen, K. A study on the resolution of a terahertz spectrometer
for the assessment of the porosity of pharmaceutical tablets. Appl. Spectrosc. 2012, 66, 319–323. [CrossRef]
[PubMed]
99. Markl, D.; Strobel, A.; Schlossnikl, R.; Bøtker, J.; Bawuah, P.; Ridgway, C.; Rantanen, J.; Rades, T.; Gane, P.;
Peiponen, K.-E.; et al. Characterisation of pore structures of pharmaceutical tablets: A review. Int. J. Pharm.
2018, 538, 188–214. [CrossRef] [PubMed]
100. Markl, D.; Wang, P.; Ridgway, C.; Karttunen, A.; Chakraborty, M.; Bawuah, P.; Pääkkönen, P.; Gane, P.;
Ketolainen, J.; Peiponen, K.-E.; et al. Characterization of the pore structure of functionalized calcium carbonate
tablets by terahertz time-domain spectroscopy and X-ray computed microtomography. J. Pharm. Sci. 2017,
106, 1586–1595. [CrossRef] [PubMed]
101. Markl, D.; Sauerwein, J.; Goodwin, D.J.; van den Ban, S.; Zeitler, J.A. Non-destructive determination of
disintegration time and dissolution in immediate release tablets by terahertz transmission measurements.
Pharm. Res. 2017, 34, 1012–1022. [CrossRef] [PubMed]
102. Markl, D.; Bawuah, P.; Ridgway, C.; van den Ban, S.; Goodwin, D.J.; Ketolainen, J.; Gane, P.; Peiponen, K.-E.;
Zeitler, J.A. Fast and non-destructive pore structure analysis using terahertz time-domain spectroscopy.
Int. J. Pharm. 2018, 537, 102–110. [CrossRef] [PubMed]
103. Bawuah, P.; Ervasti, T.; Tan, N.; Zeitler, J.A.; Ketolainen, J.; Peiponen, K. Noninvasive porosity measurement
of biconvex tablets using terahertz pulses. Int. J. Pharm. 2016, 509, 439–443. [CrossRef]
104. Chakraborty, M.; Ridgway, C.; Bawuah, P.; Markl, D.; Gane, P.A.C.; Ketolainen, J.; Zeitler, J.A.; Peiponen, K.-E.
Optics-based compressibility parameter for pharmaceutical tablets obtained with the aid of the terahertz
refractive index. Int. J. Pharm. 2017, 525, 85–91. [CrossRef]
105. Stranzinger, S.; Faulhammer, E.; Li, J.; Dong, R.; Zeitler, J.A.; Biserni, S.; Calzolari, V.; Khinast, J.G.; Markl, D.
Predicting capsule fill weight from in-situ powder density measurements using terahertz reflection technology.
Int. J. Pharm. 2019, 1, 100004. [CrossRef]
106. Herrmann, M.; Tani, M.; Sakai, K.; Fukasawa, R. Terahertz imaging of silicon wafers. J. Appl. Phys. 2002,
91, 1247–1250. [CrossRef]
107. Kiwa, T.; Tonouchi, M.; Yamashita, M.; Kawase, K. Laser terahertz-emission microscope for inspecting
electrical faults in integrated circuits. Opt. Lett. 2003, 28, 2058–2060. [CrossRef] [PubMed]
108. Yamashita, M.; Kawase, K.; Otani, C.; Kiwa, T.; Tonouchi, M. Imaging of large-scale integrated circuits using
laser terahertz emission microscopy. Opt. Express 2005, 13, 115–120. [CrossRef] [PubMed]
109. Yamashita, M.; Otani, C.; Kawase, K.; Nikawa, K.; Tonouchi, M. Noncontact inspection technique for electrical
failures in semiconductor devices using a laser terahertz emission microscope. Appl. Phys. Lett. 2008,
93, 041117. [CrossRef]
110. Yamashita, M.; Otani, C.; Kim, S.; Murakami, H.; Tonouchi, M.; Matsumoto, T.; Midoh, Y.; Miura, K.;
Nakamae, K.; Nikawa, K. Laser THz emission microscope as a novel tool for LSI failure analysis. Microelectron.
Reliab. 2009, 49, 1116–1126. [CrossRef]
111. Yamashita, M.; Otani, C.; Kawase, K.; Matsumoto, T.; Nikawa, K.; Kim, S.; Murakami, H.; Tonouchi, M.
Backside observation of large-scale integrated circuits with multilayered interconnections using laser terahertz
emission microscope. Appl. Phys. Lett. 2009, 94, 191104. [CrossRef]
112. Yamashita, M.; Otani, C.; Matsumoto, T.; Midoh, Y.; Miura, K.; Nakamae, K.; Nikawa, K.; Kim, S.;
Murakami, H.; Tonouchi, M. THz emission characteristics from p/n junctions with metal lines under non-bias
conditions for LSI failure analysis. Opt. Express 2011, 19, 10864–10873. [CrossRef]
113. Murakami, H.; Serita, K.; Maekawa, Y.; Fujiwara, S.; Matsuda, E.; Kim, S.; Kawayama, I.; Tonouchi, M.
Scanning laser THz imaging system. J. Phys. D Appl. Phys. 2014, 47, 374007. [CrossRef]
Sensors 2019, 19, 4203 31 of 35

114. Van Hoof, N.J.J.; Huurne, S.E.T.T.; Rivas, J.G.; Halpin, A. Time-resolved terahertz time-domain near-field
microscopy. Opt. Express 2018, 26, 32118–32129. [CrossRef]
115. Huber, A.J.; Keilmann, F.; Wittborn, J.; Aizpurua, J.; Hillenbrand, R. Terahertz near-field nanoscopy of mobile
carriers in single semiconductor nanodevices. Nano Lett. 2008, 8, 3766–3770. [CrossRef]
116. Nagel, M.; Michalski, A.; Kurz, H. Contact-free fault location and imaging with on-chip terahertz time-domain
reflectometry. Opt. Express 2011, 19, 12509–12514. [CrossRef]
117. Burford, N.M.; El-Shenawee, M.O.; O’neal, C.B.; Olejniczak, K.J. Terahertz imaging for nondestructive
evaluation of packaged power electronic devices. Int. J. Emerg. Technol. Adv. Eng. 2014, 4, 395–401.
118. Ahi, K.; Shahbazmohamadi, S.; Asadizanjani, N. Quality control and authentication of packaged integrated
circuits using enhanced-spatial-resolution terahertz time-domain spectroscopy and imaging. Opt. Lasers Eng.
2018, 104, 274–284. [CrossRef]
119. Shur, M.; Rudin, S.; Rupper, G.; Reed, M.; Suarez, J. Sub-terahertz testing of millimeter wave Monolithic and
very large scale integrated circuits. Solid State Electron. 2019, 155, 44–48. [CrossRef]
120. Salek, K.A.; Kawayama, I.; Murakami, H.; Tonouchi, M. Evaluation of solar cell using terahertz time-domain
spectroscopy. Phys. Sci. Int. J. 2015, 6, 96. [CrossRef]
121. Nakanishi, H.; Fujiwara, S.; Takayama, K.; Kawayama, I.; Murakami, H.; Tonouchi, M. Imaging of a
polycrystalline silicon solar cell using a laser terahertz emission microscope. Appl. Phys. Express 2012,
5, 112301. [CrossRef]
122. Salek, K.A.; Nakanishi, H.; Ito, A.; Kawayama, I.; Murakami, H.; Tonouchi, M. Laser terahertz emission
microscopy studies of a polysilicon solar cell under the illumination of continuous laser light. Opt. Eng.
2013, 53, 031204. [CrossRef]
123. Nakanishi, H.; Ito, A.; Takayama, K.; Kawayama, I.; Murakami, H.; Tonouchi, M. Comparison between laser
terahertz emission microscope and conventional methods for analysis of polycrystalline silicon solar cell.
AIP Adv. 2015, 5, 117129. [CrossRef]
124. Minkevičius, L.; Suzanovičienė, R.; Molis, G.; Krotkus, A.; Balakauskas, S.; Venckevičius, R.; Kašalynas, I.;
Seliuta, D.; Valušis, G.; Tamošiūnas, V. Solar cell imaging and characterization by terahertz techniques.
In Terahertz Emitters, Receivers, and Applications III; International Society for Optics and Photonics, SPIE: San
Diego, CA, USA, 15 October 2012; p. 849613.
125. Minkevicius, L.; Suzanoviciene, R.; Balakauskas, S.; Molis, G.; Krotkus, A.; Valusis, G.; Tamosiunas, V.
Detection of tab wire soldering defects on silicon solar cells using terahertz time-domain spectroscopy.
Electron. Lett. 2012, 48, 932–934. [CrossRef]
126. Nagel, M.; Safiei, A.; Sawallich, S.; Matheisen, C.; Pletzer, T.M.; Mewe, A.A.; van der Borg, N.J.C.M.;
Cesar, I.; Kurz, H. THz microprobe system for contact-free high-resolution sheet resistance imaging. In
Proceedings of the 28th European Photovoltaic Solar Energy Conference and Exhibition, Villepinte, Frankreich,
30 September–4 October 2013; pp. 856–860.
127. Sawallich, S.; Matheisen, C.; Nagel, M.; Cesar, I. High-resolution interdigitated back contact solar cell
inspection using Terahertz microprobes. In Proceedings of the 2015 40th International Conference on Infrared,
Millimeter, and Terahertz waves (IRMMW-THz), Hong Kong, China, 23–28 August 2015; pp. 1–2.
128. Spinelli, P.; Danzl, P.; Guillevin, N.; Mewe, A.; Sawallich, S.; Vlooswijk, A.; van de Loo, B.; Kessels, E.;
Nagel, M.; Cesar, I. High resolution sheet resistance mapping to unveil edge effects in industrial IBC solar
cells. Energy Procedia 2016, 92, 218–224. [CrossRef]
129. Spinelli, P.; Danzl, F.J.K.; Deligiannls, D.; Guillevin, N.; Burgers, A.R.; Sawallich, S.; Nage, M.; Cesar, I.
High resolution THz scanning for optimization of dielectric layer opening process on doped Si surfaces.
In Proceedings of the 2017 IEEE 44th Photovoltaic Specialist Conference (PVSC), Washington, DC, USA,
25–30 June 2017; pp. 3150–3154.
130. Cesar, I.; Guillevin, N.; Mewe, A.A.; Spinelli, P.; Burgers, A.R.; Rosca, V.; Okel, L.A.G.; Geerligs, B.J.;
Weeber, A.W.; Sawallich, S.; et al. Enablers for IBC: Integral cell and module development and implementation
in PV industry. Energy Procedia 2017, 124, 834–841. [CrossRef]
131. Buron, J.D.; Pizzocchero, F.; Jepsen, P.U.; Petersen, D.H.; Caridad, J.M.; Jessen, B.S.; Booth, T.J.; Bøggild, P.
Graphene mobility mapping. Sci. Rep. 2015, 5, 12305. [CrossRef] [PubMed]
132. Mackenzie, D.; Buron, J.C.D.; Bøggild, P.; Jepsen, P.U.; Petersen, D.H. Contactless graphene conductance
measurements: the effect of device fabrication on terahertz time-domain spectroscopy. Int. J. Nanotechnol.
2016, 13, 591–596. [CrossRef]
Sensors 2019, 19, 4203 32 of 35

133. Lin, H.; Braeuninger-Weimer, P.; Kamboj, V.S.; Jessop, D.S.; Degl’Innocenti, R.; Beere, H.E.; Ritchie, D.A.;
Zeitler, J.A.; Hofmann, S. Contactless graphene conductivity mapping on a wide range of substrates with
terahertz time-domain reflection spectroscopy. Sci. Rep. 2017, 7, 10625. [CrossRef] [PubMed]
134. Mackenzie, D.M.A.; Buron, J.D.; Whelan, P.R.; Caridad, J.M.; Bjergfelt, M.; Luo, B.; Shivayogimath, A.;
Smitshuysen, A.L.; Thomsen, J.D.; Booth, T.J.; et al. Quality assessment of graphene: Continuity, uniformity,
and accuracy of mobility measurements. Nano Res. 2017, 10, 3596–3605. [CrossRef]
135. Laib, J.P.; Mittleman, D.M. Temperature-dependent terahertz spectroscopy of liquid n-alkanes. J. Infrared
Millim. Terahertz Waves 2010, 31, 1015–1021. [CrossRef]
136. Nickel, D.V.; Garza, A.J.; Scuseria, G.E.; Mittleman, D.M. The isotropic molecular polarizabilities of single
methyl-branched alkanes in the terahertz range. Chem. Phys. Lett. 2014, 592, 292–296. [CrossRef]
137. Yin, M.; Tang, S.; Tong, M. The application of terahertz spectroscopy to liquid petrochemicals detection: A
review. Appl. Spectrosc. Rev. 2016, 51, 379–396. [CrossRef]
138. Zhan, H.; Wu, S.; Bao, R.; Ge, L.; Zhao, K. Qualitative identification of crude oils from different oil fields
using terahertz time-domain spectroscopy. Fuel 2015, 143, 189–193. [CrossRef]
139. Bao, R.M.; Li, Y.; Zhan, H.; Zhao, K.; Wang, W.; Ma, Y.; Wu, J.; Liu, S.; Li, S.; Xiao, L. Probing the oil content in
oil shale with terahertz spectroscopy. Sci. China Phys. Mech. Astron. 2015, 58, 114211. [CrossRef]
140. Matoug, M.M.; Gordon, R. Crude Oil Asphaltenes Studied by Terahertz Spectroscopy. ACS Omega 2018,
3, 3406–3412. [CrossRef]
141. Jiang, C.; Zhao, K.; Fu, C.; Xiao, L. Characterization of morphology and structure of wax crystals in waxy
crude oils by terahertz time-domain spectroscopy. Energy Fuels 2017, 31, 1416–1421. [CrossRef]
142. Li, Z.X.; Zhou, J.; Guo, X.S.; Ji, B.B.; Zhou, W.; Li, D.H. Terahertz Spectral Properties of Coal Tar. J. Appl.
Spectrosc. 2018, 85, 840–844. [CrossRef]
143. Jin, W.-J.; Zhao, K.; Yang, C.; Xu, C.-H.; Ni, H.; Chen, S.-H. Experimental measurements of water content in
crude oil emulsions by terahertz time-domain spectroscopy. Appl. Geophys. 2013, 10, 506–509. [CrossRef]
144. Guan, L.; Zhan, H.; Miao, X.; Zhu, J.; Zhao, K. Terahertz-dependent evaluation of water content in
high-water-cut crude oil using additive-manufactured samplers. Sci. China Phys. Mech. Astron. 2017,
60, 044211. [CrossRef]
145. Saied, I.M.; Meribout, M.; Kato, E.; Zhao, X.H. Terahertz spectroscopy for measuring multiphase fractions.
IEEE Trans. Terahertz Sci. Technol. 2017, 7, 250–259. [CrossRef]
146. Song, Y.; Zhan, H.L.; Zhao, K.; Miao, X.Y.; Lu, Z.Q.; Bao, R.M.; Zhu, J.; Xiao, L.Z. Simultaneous characterization
of water content and distribution in high-water-cut crude oil. Energy Fuels 2016, 30, 3929–3933. [CrossRef]
147. Arik, E.; Altan, H.; Esenturk, O. Dielectric properties of diesel and gasoline by terahertz spectroscopy.
J. Infrared Millim. Terahertz Waves 2014, 35, 759–769. [CrossRef]
148. Li, Y.-N.; Li, J.; Zeng, Z.-M.; Li, J.; Tian, Z.; Wang, W.-K. Terahertz spectroscopy for quantifying refined oil
mixtures. Appl. Opt. 2012, 51, 5885–5889. [CrossRef]
149. Li, Y.-N.; Zeng, Z.-M.; Li, J.; Tian, Z.; Sun, L.-J.; Zhou, N. Terahertz quantitatively distinguishing gasoline
mixtures using multiparameter-combined analysis. Appl. Opt. 2013, 52, 7382–7388. [CrossRef]
150. Li, J.; Tian, Z.; Chen, Y.; Cao, W.; Zeng, Z. Distinguishing octane grades in gasoline using terahertz
metamaterials. Appl. Opt. 2012, 51, 3258–3262. [CrossRef]
151. Al-Douseri, F.M.; Chen, Y.; Zhang, X.-C. THz wave sensing for petroleum industrial applications. Int. J.
Infrared Millim. Waves 2006, 27, 481–503. [CrossRef]
152. Zhao, H.; Zhao, K.; Tian, L.; Zhao, S.; Zhou, Q.; Shi, Y.; Zhao, D.; Zhang, C. Spectrum features of commercial
derv fuel oils in the terahertz region. Sci. China Phys. Mech. Astron. 2012, 55, 195–198. [CrossRef]
153. Zhao, H.; Zhao, K.; Bao, R. Fuel property determination of biodiesel-diesel blends by terahertz spectrum.
Int. J. Infrared Millim. Waves 2012, 33, 522–528. [CrossRef]
154. Arik, E.; Altan, H.; Esenturk, O. Dielectric properties of ethanol and gasoline mixtures by terahertz
spectroscopy and an effective method for determination of ethanol content of gasoline. J. Phys. Chem. A
2014, 118, 3081–3089. [CrossRef] [PubMed]
155. Zhan, H.; Zhao, K.; Zhao, H.; Li, Q.; Zhu, S.; Xiao, L. The spectral analysis of fuel oils using terahertz radiation
and chemometric methods. J. Phys. D Appl. Phys. 2016, 49, 395101. [CrossRef]
156. Qin, F.; Li, Q.; Zhan, H.; Jin, W.; Liu, H.; Zhao, K. Probing the sulfur content in gasoline quantitatively with
terahertz time-domain spectroscopy. Sci. China Phys. Mech. Astron. 2014, 57, 1404–1406. [CrossRef]
Sensors 2019, 19, 4203 33 of 35

157. Adbul-Munaim, A.M.; Reuter, M.; Koch, M.; Watson, D.G. Distinguishing gasoline engine oils of different
viscosities using terahertz time-domain spectroscopy. Int. J. Infrared Millim. Waves 2015, 36, 687–696.
[CrossRef]
158. Tian, L.; Zhou, Q.; Jin, B.; Zhao, K.; Zhao, S.; Shi, Y.; Zhang, C. Optical property and spectroscopy studies on
the selected lubricating oil in the terahertz range. Sci. China Ser. G Phys. Mech. Astron. 2009, 52, 1938–1943.
[CrossRef]
159. Abdul-Munaim, A.M.; Aller, M.M.; Preu, S.; Watson, D.G. Discriminating gasoline fuel contamination in
engine oil by terahertz time-domain spectroscopy. Tribol. Int. 2018, 119, 123–130. [CrossRef]
160. Abdul-Munaim, A.M.; Ornik, J.; Koch, M.; Watson, D.G. Terahertz Time Domain Spectroscopy to Detect
Different Oxidation Levels of Diesel Engine Oil. Lubricants 2019, 7, 18. [CrossRef]
161. Nishimura, N.; Ogura, R.; Matsumoto, S.; Mizuno, M.; Fukunaga, K. Study of molecular behavior in oxidation
of insulating oil using terahertz spectroscopy. Electr. Eng. Jpn. 2013, 183, 9–15. [CrossRef]
162. Nishimura, N.; Ogura, R.; Matsumoto, S.; Mizuno, M.; Fukunaga, K. Transmittance spectra of oxidized
insulation oil using terahertz spectroscopy. In Proceedings of the 2011 IEEE International Conference on
Dielectric Liquids, Trondheim, Norway, 26–30 June 2011; pp. 1–4.
163. Kang, S.B.; Kim, W.-S.; Chung, D.C.; Joung, J.M.; Kwak, M.H. Degradation diagnosis of transformer insulating
oils with terahertz time-domain spectroscopy. J. Korean Phys. Soc. 2017, 71, 986–992. [CrossRef]
164. Li, M.; Tong, M.-M.; Fletcher, J.E.; Dong, Z.-Y. A novel approach to investigate the deterioration of insulation
of oils in power transformers with terahertz time-domain spectroscopy. IEEE Trans. Dielectr. Electr. Insul.
2017, 24, 930–938. [CrossRef]
165. Gorenflo, S.; Tauer, U.; Hinkov, I.; Lambrecht, A.; Buchner, R.; Helm, H. Dielectric properties of oil–water
complexes using terahertz transmission spectroscopy. Chem. Phys. Lett. 2006, 421, 494–498. [CrossRef]
166. Abdul-Munaim, A.M.; Reuter, M.; Abdulmunem, O.M.; Balzer, J.; Koch, M.; Watson, D.G. Using
terahertz time-domain spectroscopy to discriminate among water contamination levels in diesel engine oil.
Trans. ASABE 2016, 59, 795–801.
167. Majewski, A. Terahertz Spectroscopy: High-resolution terahertz spectrometer sniffs out chemicals.
Laser Focus World 2008, 44, 4.
168. Yang, L.; Guo, T.; Zhang, X.; Cao, S.; Ding, X. Toxic chemical compound detection by terahertz spectroscopy:
A review. Rev. Anal. Chem. 2018, 37. [CrossRef]
169. Smith, R.M.; Arnold, M.A. Selectivity of terahertz gas-phase spectroscopy. Anal. Chem. 2015, 87, 10679–10683.
[CrossRef] [PubMed]
170. Graber, B.; Kim, C.; Wu, D.H. High SNR single measurements of trace gas phase spectra at THz frequencies.
Appl. Phys. Lett. 2017, 111, 221107. [CrossRef]
171. Medvedev, I.R.; Behnke, M.; de Lucia, F.C. Fast analysis of gases in the submillimeter/terahertz with “absolute”
specificity. Appl. Phys. Lett. 2005, 86, 154105. [CrossRef]
172. Neese, C.F.; Medvedev, I.R.; Plummer, G.M.; Frank, A.J.; Ball, C.D.; de Lucia, F.C. Compact
submillimeter/terahertz gas sensor with efficient gas collection, preconcentration, and ppt sensitivity.
IEEE Sens. J. 2012, 12, 2565–2574. [CrossRef]
173. Tekawade, A.; Rice, T.E.; Oehlschlaeger, M.A.; Mansha, M.W.; Wu, K.; Hella, M.M.; Wilke, I. Towards
realization of quantitative atmospheric and industrial gas sensing using THz wave electronics. Appl. Phys. B
2018, 124, 105. [CrossRef]
174. Schmalz, K.; Rothbart, N.; Eissa, M.H.; Borngräber, J.; Kissinger, D.; Hübers, H.-W. Transmitters and receivers
in SiGe BiCMOS technology for sensitive gas spectroscopy at 222–270 GHz. AIP Adv. 2019, 9, 015213.
[CrossRef]
175. Hindle, F.; Kuuliala, L.; Mouelhi, M.; Cuisset, A.; Bray, C.; Vanwolleghem, M.; Devlieghere, F.; Mouret, G.;
Bocquet, R. Monitoring of food spoilage by high resolution THz analysis. Analyst 2018, 143, 5536–5544.
[CrossRef] [PubMed]
176. Neumaier, P.F.-X.; Schmalz, K.; Borngräber, J.; Wylde, R.; Hübers, H.-W. Terahertz gas-phase spectroscopy:
chemometrics for security and medical applications. Analyst 2015, 140, 213–222. [CrossRef] [PubMed]
177. Zhan, H.; Zhao, K.; Bao, R.; Xiao, L. Monitoring PM2. 5 in the atmosphere by using terahertz time-domain
spectroscopy. J. Infrared Millim. Terahertz Waves 2016, 37, 929–938. [CrossRef]
178. Wu, C.; Miao, X.; Zhao, K. Identifying PM 2.5 samples collected in different environment by using terahertz
time-domain spectroscopy. Front. Optoelectron. 2018, 11, 256–260. [CrossRef]
Sensors 2019, 19, 4203 34 of 35

179. Hepp, C.; Lüttjohann, S.; Roggenbuck, A.; Deninger, A.; Nellen, S.; Göbel, T.; Jörger, M.; Harig, R.
A cw-terahertz gas analysis system with ppm detection limits. In Proceedings of the 2016 41st
International Conference on Infrared, Millimeter, and Terahertz waves (IRMMW-THz), Copenhagen,
Denmark, 25–30 September 2016; pp. 1–2.
180. Shimizu, N.; Ikari, T.; Kikuchi, K.; Matsuyama, K.; Wakatsuki, A.; Kohjiro, S.; Fukasawa, R. Remote gas
sensing in full-scale fire with sub-terahertz waves. In Proceedings of the 2011 IEEE MTT-S International
Microwave Symposium, Baltimore, MD, USA, 5–10 June 2011; pp. 1–4.
181. Fosnight, A.M.; Moran, B.L.; Medvedev, I.R. Chemical analysis of exhaled human breath using a terahertz
spectroscopic approach. Appl. Phys. Lett. 2013, 103, 133703. [CrossRef]
182. Schmalz, K.; Rothbart, N.; Neumaier, P.F.; Borngräber, J.; Hübers, He.; Kissinger, D. Gas spectroscopy system
for breath analysis at mm-wave/THz using SiGe BiCMOS circuits. IEEE Trans. Microw. Theory Tech. 2017,
65, 1807–1818. [CrossRef]
183. Rothbart, N.; Holz, O.; Koczulla, R.; Schmalz, K.; Hübers, He. Analysis of Human Breath by
Millimeter-Wave/Terahertz Spectroscopy. Sensors 2019, 19, 2719. [CrossRef]
184. Ge, L.N.; Zhan, H.L.; Leng, W.X.; Zhao, K.; Xiao, L.Z. Optical characterization of the principal hydrocarbon
components in natural gas using terahertz spectroscopy. Energy Fuels 2015, 29, 1622–1627. [CrossRef]
185. Leng, W.; Zhan, H.; Ge, L.; Wang, W.; Ma, Y.; Zhao, K.; Li, S.; Xiao, L. Rapidly determinating the principal
components of natural gas distilled from shale with terahertz spectroscopy. Fuel 2015, 159, 84–88. [CrossRef]
186. Banerjee, D.; von Spiegel, W.; Thomson, M.D.; Schabel, S.; Roskos, H.G. Diagnosing water content in paper
by terahertz radiation. Opt. Express 2008, 16, 9060–9066. [CrossRef] [PubMed]
187. Mousavi, P.; Haran, F.; Jez, D.; Santosa, F.; Dodge, J.S. Simultaneous composition and thickness measurement
of paper using terahertz time-domain spectroscopy. Appl. Opt. 2009, 48, 6541–6546. [CrossRef] [PubMed]
188. Hattori, T.; Kumon, H.; Tamazumi, H. Terahertz spectroscopic characterization of paper. In Proceedings of
the 35th International Conference on Infrared, Millimeter, and Terahertz Waves, Rome, Italy, 5–10 September
2010; pp. 1–2.
189. Merbold, H.; Maas, D.J.H.C.; van Mechelen, J.L.M. Multiparameter sensing of paper sheets using terahertz
time-domain spectroscopy: Caliper, fiber orientation, moisture, and the role of spatial inhomogeneity.
In Proceedings of the 2016 IEEE SENSORS, Orlando, FL, USA, 30 October–3 November 2016; pp. 1–3.
190. Van Mechelen, J.L.M.; Kuzmenko, A.B.; Merbold, H. Stratified dispersive model for material characterization
using terahertz time-domain spectroscopy. Opt. Lett. 2014, 39, 3853–3856. [CrossRef] [PubMed]
191. Van Mechelen, J.L.M.; Maas, D.J.H.C.; Merbold, H. Paper sheet parameter determination using terahertz
spectroscopy. In Proceedings of the 2016 41st International Conference on Infrared, Millimeter, and Terahertz
waves (IRMMW-THz), Copenhagen, Denmark, 25–30 September 2016; pp. 1–2.
192. Fan, M.; Cao, B.; Tian, G. Enhanced measurement of paper basis weight using phase shift in terahertz
time-domain spectroscopy. J. Sens. 2017, 2017, 3520967. [CrossRef]
193. Khani, B.; Hu, Y.; Rymanov, V.; Brenner, C.; Hofmann, M.; Stöhr, A. Compact Optoelectronic Continuous
Wave Terahertz Spectroscopy System (230–400 GHz) for Paper Sorting and Characterization. In The European
Conference on Lasers and Electro-Optics; Optical Society of America: Washington, DC, USA, 2017.
194. Yan, G.; Markov, A.; Chinifooroshan, Y.; Tripathi, S.M.; Bock, W.J.; Skorobogatiy, M. Resonant THz sensor
for paper quality monitoring using THz fiber Bragg gratings. Opt. Lett. 2013, 38, 2200–2202. [CrossRef]
[PubMed]
195. Huber, P.; Martinez, P.; Guers, C.; Garet, F.; Borel, P. Dielectric Losses of Paper in the THz Domain: Literature
Review, Needs for Future Research, and Prospective Solutions. Phys. Status Solidi 2017, 214, 1700356.
[CrossRef]
196. Wang, L.; Tang, C.; Zhu, S.; Zhou, S. Terahertz time domain spectroscopy of transformer insulation paper
after thermal aging intervals. Materials 2018, 11, 2124. [CrossRef]
197. Vassilev, V.; Stoew, B.; Blomgren, J.; Andersson, G. A mm-wave sensor for remote measurement of moisture
in thin paper layers. IEEE Trans. Terahertz Sci. Technol. 2015, 5, 770–778. [CrossRef]
198. Brinkmann, S.; Vieweg, N.; Gärtner, G.; Plew, P.; Deninger, A. Towards quality control in pharmaceutical
packaging: Screening folded boxes for package inserts. J. Infrared Millim. Terahertz Waves 2017, 38, 339–346.
[CrossRef]
Sensors 2019, 19, 4203 35 of 35

199. Reid, M.; Fedosejevs, R. Terahertz birefringence and attenuation properties of wood and paper. Appl. Opt.
2006, 45, 2766–2772. [CrossRef] [PubMed]
200. Todoruk, T.M.; Hartley, I.D.; Reid, M.E. Origin of birefringence in wood at terahertz frequencies. IEEE Trans.
Terahertz Sci. Technol. 2011, 2, 123–130. [CrossRef]
201. Reid, M.E.; Hartley, I.D.; Todoruk, T.M. Terahertz applications in the wood products industry. In Handbook of
Terahertz Technology for Imaging, Sensing and Communications; Woodhead Publishing: Cambridge, UK, 2013;
pp. 547–578.
202. Inagaki, T.; Ahmed, B.; Hartley, I.D.; Tsuchikawa, S.; Reid, M. Simultaneous prediction of density and
moisture content of wood by terahertz time domain spectroscopy. J. Infrared Millim. Terahertz Waves 2014,
35, 949–961. [CrossRef]
203. Inagaki, T.; Hartley, I.D.; Tsuchikawa, S.; Reid, M. Prediction of oven-dry density of wood by time-domain
terahertz spectroscopy. Holzforschung 2014, 68, 61–68. [CrossRef]
204. Krügener, K.; Sommer, S.; Stübling, E.; Jachim, R.; Koch, M.; Viöl, W. THz Properties of Typical Woods
Important for European Forestry. J. nfrared Millim. Terahertz Waves 2019, 40, 770–774. [CrossRef]
205. Krügener, K.; Stübling, E.-M.; Jachim, R.; Kietz, B.; Koch, M.; Viöl, W. THz tomography for detecting damages
on wood caused by insects. Appl. Opt. 2019, 58, 6063–6066. [CrossRef]
206. QYResearch. Available online: https://www.qyresearch.com/index/detail/1172368/global-terahertz-thz-
technology-market (accessed on 27 September 2019).
207. Research and Markets: the World’s Largest Market Research Store. Available online: https://www.
researchandmarkets.com/research/xmr4gt/global_terahertz?w=4 (accessed on 27 September 2019).
208. Coherent Market Insights. Available online: https://www.coherentmarketinsights.com/market-insight/
terahertz-components-and-systems-market-596 (accessed on 27 September 2019).
209. Tematys: Exploration of Photonics Markets. Available online: https://tematys.fr/Publications/en/
terahertz/39-terahertz-components-systems-technology-and-market-trends-update-2016.html (accessed on
27 September 2019).
210. Markets and Markets. Available online: https://www.marketsandmarkets.com/PressReleases/terahertz-
technology.asp (accessed on 27 September 2019).

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).