Full-Text PDF

sensors
Article
Improved Tuning Fork for Terahertz
Quartz-Enhanced Photoacoustic Spectroscopy
Angelo Sampaolo 1,2 , Pietro Patimisco 1,2 , Marilena Giglio 1 , Miriam S. Vitiello 3 ,
Harvey E. Beere 4 , David A. Ritchie 4 , Gaetano Scamarcio 1 , Frank K. Tittel 2 and
Vincenzo Spagnolo 1, *
1
2
3
4
*
Dipartimento Interateneo di Fisica, Università degli studi di Bari Aldo Moro e Politecnico di Bari,
Via Amendola 173, Bari I-70126, Italy; [email protected] (A.S.); [email protected] (P.P.);
[email protected] (M.G.); [email protected] (G.S.)
Department of Electrical and Computer Engineering, Rice University, 6100 Main Street, Houston, TX 77005,
USA; [email protected]
NEST, CNR-Istituto Nanoscienze and Scuola Normale Superiore, Piazza San Silvestro 12, Pisa I-56127, Italy;
[email protected]
Cavendish Laboratory, University of Cambridge, J. J. Thomson Avenue, Cambridge CB3 0HE, UK;
[email protected] (H.E.B.); [email protected] (D.A.R.)
Correspondence: [email protected]; Tel.: +39-080-544-2373
Academic Editor: Markus W. Sigrist
Received: 8 February 2016; Accepted: 23 March 2016; Published: 25 March 2016
Abstract: We report on a quartz-enhanced photoacoustic (QEPAS) sensor for methanol (CH3 OH)
detection employing a novel quartz tuning fork (QTF), specifically designed to enhance the QEPAS
sensing performance in the terahertz (THz) spectral range. A discussion of the QTF properties in
terms of resonance frequency, quality factor and acousto-electric transduction efficiency as a function
of prong sizes and spacing between the QTF prongs is presented. The QTF was employed in a
QEPAS sensor system using a 3.93 THz quantum cascade laser as the excitation source in resonance
with a CH3 OH rotational absorption line located at 131.054 cm´1 . A minimum detection limit
of 160 ppb in 30 s integration time, corresponding to a normalized noise equivalent absorption
NNEA = 3.75 ˆ 10´11 cm´1 W/Hz½ , was achieved, representing a nearly one-order-of-magnitude
improvement with respect to previous reports.
Keywords: quartz enhanced photoacoustic spectroscopy; quartz tuning fork; gas sensing; THz
spectroscopy; quantum cascade laser
1. Introduction
Spectroscopic techniques for trace gas sensing and monitoring have shown great potential for
non-invasive chemical analysis requiring high sensitivity and selectivity. While trace gas detection
in the near-infrared (IR) and mid-IR ranges demonstrated excellent performance, the use of far-IR or
terahertz (THz) radiation for sensing purposes is still underdeveloped. In the wavelength range from
50 µm to 3 mm wavelengths (0.1–6 THz), numerous gas molecules have well-defined spectral THz
“fingerprints” absorptions, due to strong rotational transitions between molecular energetic levels.
Hence, the THz spectral region can provide higher detection selectivities compared to those provided
by the characteristic ro-vibrational complex structures in the mid-IR range. A number of THz optical
systems have been proposed for gas sensing and spectroscopy. Terahertz Time-Domain Spectroscopy
(THz-TDS) is based on frequency conversion employing nonlinear optics and femtosecond laser pulses
for the generation and detection of THz radiation [1]. Spectral measurements over a very broad
bandwidth in the THz range can be performed using THz-TDS, allowing spectroscopic investigation
of a large number of gases, such as water vapor, methyl chloride and nitrous oxide [2–4]. However,
Sensors 2016, 16, 439; doi:10.3390/s16040439
www.mdpi.com/journal/sensors
Sensors 2016, 16, 439
2 of 8
THz-TDS is strongly limited by the available µW-range output optical powers. Furthermore, such
a technique usually provides a limited control of the selected frequency bandwidth. The sensitivity
of THz gas spectroscopy can be enhanced by increasing the available optical power. So far, THz
quantum cascade lasers (QCLs) represent an attractive solution in the far-IR spectral region in terms of
output power [5,6] and spectral purity [7]. THz QCLs can provide single-mode emission with output
powers of up to 138 mW in continuous wave (cw) operation at cryogenic temperatures [8]. Wavelength
modulation spectroscopy, employing liquid-He-cooled cw-operating QCLs combined with low-noise
bolometer detectors, is promising in terms of THz-sensing performances [9,10]. This approach would
allow reaching high sensitivity and selectivity, but suffers from complexity and the disadvantage of
using cryogenic cooling systems for both generation and detection of THz radiation.
Recently, THz quartz-enhanced photoacoustic (QEPAS) sensors were reported [11–13]. QEPAS
permits overcoming some issues traditionally associated with THz spectroscopy, such as the use of
cryogenic systems for the detection of THz radiation and complex signal analysis processes. QEPAS
is based on the detection of acoustic waves produced by the absorbing gas target by means of a
piezoelectric quartz tuning fork (QTF) acting as an acousto-electric transducer [11–16]. The merits of
QEPAS include high detection sensitivity, high selectivity, and fast time response using a compact and
relatively low-cost acoustic detection module [16–19]. Since rotational levels are up to three orders
of magnitude faster with respect to mid-IR vibrational levels, the THz spectral range is particularly
suitable for the QEPAS technique. Indeed, the fast relaxation times characteristic of THz transitions
allow low-pressure operation, providing high QTF resonance Q-factors, and thereby large QEPAS
signals [16].
The extension of QEPAS in the THz range was made possible by the realization of a custom-made
QTF [11,12] (denoted in this work as C-QTF) having the same geometry as the commercial
32.78 kHz-QTF, but size six times larger, with prong lengths of L = 20.0 mm, prong widths of T = 1.4 mm
and a crystal thickness of w = 0.8 mm. The QTF prongs are separated by a ~1 mm gap, required to focus
a THz laser beam between the prongs without illuminating them. It is worth noting that the standard
QTF structure is optimized for timing purposes and is not ideal for spectroscopic applications.
In a previous study [20], we investigated the electro-elastic properties of a set of six QTFs with
different geometries and the objective of determining the optimal design for optoacoustic gas sensing.
This study allowed us to identify a novel QTF design (denoted in this work as N-QTF), optimized for
THz QEPAS.
The performance of a QEPAS sensor can be considered in terms of its QEPAS signal (S) which can
be expressed as S = KP0 Qαε, where K represents the QTF efficiency in converting the acoustic pressure
wave incident on the internal side of the two prongs into transversal in-plane deflections, P0 is the
laser power, Q is the QTF quality factor, α is the gas absorption coefficient and ε is the conversion
efficiency of the absorbed optical power into sound. The factor Q¨ K can be used as the figure of merit
to compare the sensing performance of the two QTFs in a QEPAS sensor, if all other sensor parameters
(e.g., laser power, absorption strength, gas pressure and sound conversion efficiency) are kept constant.
The N-QTF was used in the same QEPAS platform as previously used with the C-QTF to detect
methanol [11,12]. A comparison of the performances between the two THz QEPAS sensor systems
shows a sensitivity enhancement of approximately one order of magnitude when using the N-QTF.
2. Quartz Tuning Fork Design and Implementation
To investigate the influence of QTF sizes on the acousto-electric transduction efficiency, i.e., the
conversion efficiency of the acoustic wave in piezoelectric charges, we selected three QTF parameters
that directly influence this process: (i) the QTF resonance frequency f ; (ii) its quality factor Q; and
(iii) the spacing s between the two prongs [20].
In a typical QEPAS sensor, the laser modulation frequency f L is set at one of the QTF resonance
frequencies or the related sub-harmonics [16]. To ensure that the gas non-radiative energy transfer,
which generates acoustic waves, follows the fast modulation f L of the incident laser radiation efficiently,
Sensors 2016, 16, 439
Sensors 2016, 16, 439
3 of 8
3 of 8
the condition
<< 1/2πτ must
be satisfied
[21],
τ ison
thethe
time
constant
non-radiative
gas
relaxationf Lprocesses.
The time
constant
τ where
depends
specific
gasofcarrier
as wellgas
as
relaxation
processes.
The
time
constant
τ
depends
on
the
specific
gas
carrier
as
well
as
intermolecular
intermolecular interactions and typically falls in the microseconds range [22]. Hence, the QTF
interactionsfrequency
and typically
fallsbeinreduced
the microseconds
range
Hence, the
QTFkHz
resonance
resonance
should
with respect
to [22].
the standard
32.78
QTF, infrequency
order to
should bethe
reduced
respecttime
to the
standard
32.78inkHz
QTF,
in order
to approach
the typical
approach
typicalwith
relaxation
of gases,
resulting
more
efficient
sound
wave generation
and
relaxation
time
of
gases,
resulting
in
more
efficient
sound
wave
generation
and
providing
an
increase
providing an increase of the QEPAS signal. Recently, a QTF with a resonance frequency of 30.72 kHz
of theemployed
QEPAS signal.
Recently,
a QTF
with a resonance
frequency
of 30.72
kHzItswas
employed was
in a
was
in a QEPAS
sensor
operating
in the near-IR
spectral
region.
performance
QEPAS sensor
the near-IR
region.
Its aperformance
waskHz
compared
with
same
compared
withoperating
the same in
QEPAS
sensorspectral
architecture
with
standard 32.78
QTF, and
anthe
increase
QEPAS
sensorsignal
architecture
with of
a standard
32.78 kHz[23].
QTF, and an increase of the QEPAS signal by a
of
the QEPAS
by a factor
1.5 was measured
factorThe
of 1.5
was measured
[23]. are defined by the QTF’s material properties and its geometry and
resonance
frequencies
Thecalculated
resonancebyfrequencies
areone
defined
by the
the fork
QTF’s
properties andprong.
its geometry
and
can be
considering
arm of
asmaterial
a cantilever-vibrating
In QEPAS
can be calculated
one arm
of the
fork where
as a cantilever-vibrating
prong. directions,
In QEPAS
spectroscopy,
onlyby
theconsidering
QTF symmetric
flexural
modes,
the tines move in opposite
spectroscopy,
only
theflexural
QTF symmetric
flexural
modes,
wherefrequencies
the tines move
in opposite
directions,
are
excited. For
these
modes, the
related
resonance
depend
on prong
sizes as
2
are
excited.
For
these
flexural
modes,
the
related
resonance
frequencies
depend
on
prong
sizes
as
f~T/L [20]. For a QTF operating in air at the fundamental flexural mode, the Q was shown
to be
2
f~T/L [20]. For
QTF[20].
operating
at the
fundamental
flexural
mode, proportional
the Q was shown
be
proportional
to aTw/L
Since in
theairQTF
piezoelectric
signal
is directly
to Q,tothis
proportional
to Tw/L
Since
QTF piezoelectric signal is directly proportional to Q, this parameter
parameter should
be[20].
as high
asthe
possible.
should
be as
possible.
Even
if fhigh
and as
Q are
not influenced by the prong spacing s, this parameter plays a crucial role in
Even if f and Q transduction
are not influenced
by theFor
prong
spacinglaser
s, thisbeam,
parameter
playsofa incident
crucial role
in the
the acousto-electric
efficiency.
a focused
the decay
acoustic
acousto-electric
transduction
efficiency.
For
a
focused
laser
beam,
the
decay
of
incident
acoustic
wave
wave pressure on the prong depends critically on the distance between the focused laser beam
pressure
on the
depends
on the distance
between
the focused
beam
position
and
position and
theprong
internal
prong critically
surface. On
other hand,
the larger
the laserlaser
beam
waist,
the larger
prongbesurface.
Ontothe
other
hand,
thebeam
largerprofile
the laser
beam
waist,
largerand
thethereby
s value
the internal
s value must
in order
avoid
the
power
tails
hitting
thethe
prongs
must be in order
to avoid the
power beam
profile noise.
tails hitting
thes prongs
therebybased
introducing
introducing
a modulated
fringe-like
background
Hence,
should and
be chosen
on the
a modulated
fringe-like
noise.toHence,
s should In
be achosen
based reported
on the expected
beam
expected
beam
waist ofbackground
the laser source
be employed.
previously
THz QEPAS
waist
the laser
source
to be
employed.
In a previously
reported
QEPAS
[12], the laser
sensorof[12],
the laser
beam
was
focused between
a prong
gap s =THz
1 mm
with sensor
a ~430-μm-diameter
focused
prongof
gap
1 mm
withtoa pass
~430-µm-diameter
beam waistC-QTF
and it without
allowed
beam was
waist
and itbetween
alloweda100%
thes =laser
light
through the employed
100%
of the
lightThis
to pass
through
the employed
C-QTFofwithout
touching
the prongs.
This result
touching
thelaser
prongs.
result
indicated
that a reduction
s is feasible.
Hence,
s was decreased
to
indicated
that
a
reduction
of
s
is
feasible.
Hence,
s
was
decreased
to
700
µm
in
the
design
of
the
N-QTF.
700 μm in the design of the N-QTF. A reduced prong thickness of T = 1 mm (1.4 mm for the C-QTF)
A
reduced
prong decrease
thicknessofofthe
T =resonance
1 mm (1.4
mm for the
C-QTF) allowed
a further
decrease
of the
allowed
a further
frequency.
In addition,
the prong
length was
reduced
to
resonance
addition,
thetoprong
length
was quality
reducedfactor,
to L = Q.
17 The
mmfabrication
from L = 20ofmm
order
L
= 17 mm frequency.
from L = 20Inmm
in order
maintain
a high
thein
N-QTF
to maintain
high
quality
factor,CA,
Q. The
fabrication
the N-QTF
was
realizedwby
Statek,
CA,
was
realizedaby
Statek,
Orange,
starting
from aof
crystal
plate of
thickness
= 250
μmOrange,
(a standard
starting
from a crystal
plate of thickness w = 250 µm (a standard for commercial QTF resonators).
for
commercial
QTF resonators).
Figure 1 shows the design of the N-QTF and its main geometrical
geometrical parameters.
parameters.
Figure
geometrical parameters.
parameters.
Figure 1.
1. Picture
Picture of
of the
the N-QTF
N-QTF including
including the
the size
size of
of the
the main
main geometrical
The electrical characterization of the N-QTF was carried out by applying a sinusoidal voltage
The electrical characterization of the N-QTF was carried out by applying a sinusoidal voltage
provided by a wave-function generator. The electrical response of the resonator was processed by a
provided by a wave-function generator. The electrical response of the resonator was processed by
transimpedance amplifier and then demodulated at the same frequency of the voltage excitation by
a transimpedance amplifier and then demodulated at the same frequency of the voltage excitation
means of a lock-in amplifier. The resonance frequency measured in standard air at a pressure of
by means of a lock-in amplifier. The resonance frequency measured in standard air at a pressure
10 Torr is f = 2871 Hz lower than that measured for the C-QTF (4250 Hz). A quality factor of Q = 18,600
of 10 Torr is f = 2871 Hz lower than that measured for the C-QTF (4250 Hz). A quality factor of
was obtained, which is lower than that obtained for the C-QTF (>30,000) with the same experimental
Q = 18,600 was obtained, which is lower than that obtained for the C-QTF (>30,000) with the same
conditions. This is mostly due to the smaller crystal thickness of w = 250 μm with respect to w = 800 μm
experimental conditions. This is mostly due to the smaller crystal thickness of w = 250 µm with respect
for the C-QTF.
to w = 800 µm for the C-QTF.
Sensors 2016, 16, 439
Sensors 2016, 16, 439
Sensors 2016, 16, 439
4 of 8
4 of 8
4 of 8
THzQEPAS
QEPASSensor
SensorArchitecture
Architecture
3. 3.THz
3. THz QEPAS Sensor Architecture
The
CH
3OH
QEPASsensor
sensorscheme,
scheme,similar
similarto
tothe
theone
one reported
reported in
in [11,12], is depicted
The
CH
QEPAS
depicted in
in Figure
Figure 2.
3 OH
CH3OHthe
QEPAS
sensor
scheme,
similar
to thetargeting
one reported
in [11,12],
is depicted
inlocated
Figure
WeThe
employed
same
source,
which
allowed
the
methanol
absorption
line
We2.employed
the same
laserlaser
source,
which
allowed
targeting the
methanol
absorption
line
located at
2. νWe
employed
the
samecm
laser) with
source,
which allowed
targeting
methanol
line located
−21 ´
21 cm/mol
= 3.9289
(131.054
line-strength
of
10the
cm/mol
inabsorption
HITRAN
units.
ν =at3.9289
THzTHz
(131.054
cm´1 )−1with
aaline-strength
of SS ==4.28
4.28׈
10
in HITRAN
units.
at ν = 3.9289 THz (131.054 cm−1) with a line-strength of S = 4.28 × 10−21 cm/mol in HITRAN units.
Figure
2. 2.
Schematic
ofofthe
Figure
Schematic
theQEPAS
QEPAStrace
tracegas
gassensor
sensorusing
usingaaTHz
THzQuantum
QuantumCascade
Cascade Laser
Laser (THz
(THz QCL) as
Figure
2. Schematic
of thePM—Parabolic
QEPAS trace
gas
sensor
using a THz
Quantum
Cascade
Laser
(THz QCL)
the excitation
source.
Mirror;
ADM—Acoustic
Detection
Module;
QTF—Quartz
theas
excitation
source.
PM—Parabolic
Mirror;
ADM—Acoustic
Detection
Module;
QTF—Quartz
Tuning
as
the
excitation
source.
PM—Parabolic
Mirror;
ADM—Acoustic
Detection
Module;
QTF—Quartz
Tuning
Fork; PC—Personal
Computer.
Fork;
PC—Personal
Computer.
Tuning Fork; PC—Personal Computer.
The THz laser beam was focused between the N-QTF prongs by using two off-axis paraboloidal
The
THz
laser
beam
was
focused
between
the
prongsby
byusing
usingtwo
twooff-axis
off-axisparaboloidal
paraboloidal
The
THz
laser
beam
was
focused
between
the N-QTF
N-QTF
prongs
aluminum
reflectors
(PM#1
with
f-number
= 2 and
PM#2 with
f-number
= 5). The
N-QTF is fixed on
aluminum
reflectors
(PM#1
with
f-number
=
2
and
PM#2
with
f-number
=
5).
The
N-QTF
fixed
with f-number
= 2 and
PM#2 with
f-number
= 5).
The
N-QTF is is
fixed
onon
aaluminum
mountingreflectors
structure(PM#1
and housed
in an acoustic
detection
module
(ADM),
with
polymethylpentene
a mounting
structure
and
housed
an acoustic
detection
module (ADM),
(ADM),with
withpolymethylpentene
polymethylpentene
a mounting
structure
and
housedinin
acoustic
detection
module
(TPX)
input and
output
windows.
Anan
optical
power
of 40 W
was measured
between the prongs of
(TPX)
input
and
output
windows.
An
optical
power
of
µW
was
measured
between
the
prongs
(TPX)
input
and
output
windows.
An
optical
power
40
W
was
measured
between
the
prongs
of of
the QTF by means of a pyroelectric detector (not shown in Figure 2). The THz laser beam
exiting
the
the
QTF
by
means
of
a
pyroelectric
detector
(not
shown
Figure
2).
The
THz
laser
beam
exiting
the
theADM
QTF
by
means
of
a
pyroelectric
detector
(not
shown
in
Figure
2).
The
THz
laser
beam
exiting
was re-focused onto a pyroelectric camera (Spiricon Pyrocam III-C) by means of an additionalthe
ADM
was
re-focused
ontoa apyroelectric
pyroelectric
camera (Spiricon
(Spiricon
Pyrocam
III-C)
anan
additional
ADM
was
re-focused
Pyrocam
III-C)by
bymeans
means
additional
aluminum
paraboliconto
mirror
(PM#3
withcamera
f-number
= 2). QEPAS
measurements
wereofof
performed
by
aluminum
parabolic
mirror
(PM#3
with
f-number
=
QEPAS
measurements
were
performed
byby
aluminum
parabolic
mirror
(PM#3
with
f-number
=
2).
QEPAS
measurements
were
performed
applying a sinusoidal waveform provided by a function generator (Tektronix model AFG3102) at the
applying
a
sinusoidal
waveform
provided
by
a
function
generator
(Tektronix
model
AFG3102)
at
the
applying
a
sinusoidal
waveform
provided
by
a
function
generator
(Tektronix
model
AFG3102)
QTF resonance frequency f to the current driver while demodulating the QTF response at f by using at
resonance
frequency
f toResearch
current
driver
while
demodulating
the
QTF
response
at LabViewf by using
theaQTF
QTF
resonance
frequency
f the
to the
current
driver
while
demodulating
the
QTF response
at f by
lock-in
amplifier
(Stanford
Model
SR830).
Both
instruments
were
controlled
by
a
lock-in
amplifier
(Stanford
Research
Model
SR830).
Both
instruments
were
controlled
by
LabViewusing
a lock-in
amplifier
Research
SR830).
were through
controlled
based
software.
A slow (Stanford
voltage ramp
allowsModel
scanning
of theBoth
THz instruments
laser wavelength
the by
based
software.
A
slow
voltage
ramp
allows
scanning
of
the
THz
laser
wavelength
through
the
selected methanol
absorption
line
for spectral
acquisition.
LabView-based
software.
A slow
voltage
rampline-shape
allows scanning
of the THz laser wavelength through
selected methanol absorption line for spectral line-shape acquisition.
the selected methanol absorption line for spectral line-shape acquisition.
Figure 3. Two-dimensional beam profile of the THz-QCL acquired by means of an IR pyrocamera
Figure
3.
Two-dimensional
beam
profile
the
acquired
by
means
ofan
anIRIRpyrocamera
pyrocamera
Figure
3.
Two-dimensional
beam
profile
of
the THz-QCL
THz-QCL
means
after
mirror
PM#3 (see Figure
2) when
theof
beam
is focused acquired
out
of theby
N-QTF
(a)ofor
between
the two
after
mirror
PM#3
(see
Figure
2)
when
the
beam
is
focused
out
of
the
N-QTF
(a)
or
between
the two
after
mirror
PM#3
(see
Figure
when
the
beam
is
focused
out
of
the
N-QTF
(a)
or
between
prongs (b). Both beam profiles are shown together with an illustration representing the position
ofthe
prongs
(b).
Both
beam
profiles
are
shown
together
with
an
illustration
representing
the
position
of of
two
prongs
(b).
Both
beam
profiles
are
shown
together
with
an
illustration
representing
the
position
the focused THz beam (red spot) with respect to the N-QTF.
focused
THz
beam
(redspot)
spot)with
withrespect
respectto
tothe
the N-QTF.
N-QTF.
thethe
focused
THz
beam
(red
Sensors 2016, 16, 439
5 of 8
Figure 3 depicts the observed two-dimensional (2D) laser beam profiles after mirror PM#3. First
5 of 8
the profile is measured with the focused laser beam positioned out of the N-QTF (Figure 3a) and then
between the N-QTF prongs (Figure 3b). A comparison between the two beam profiles clearly
indicated
that
the THzthe
beam
is not obstructed
by the QTF
when
is focused
between
prongs.
The
Figure
3 depicts
observed
two-dimensional
(2D)
laserit beam
profiles
afterits
mirror
PM#3.
total the
image
intensity
was measured
summing
pixel
values forout
both
profiles,
from3a)
which
First
profile
is measured
with theby
focused
laserall
beam
positioned
of beam
the N-QTF
(Figure
and
we
estimated
that
96.4%
of
the
light
intensity
passes
between
the
prongs.
Hence,
a
reduction
of
the
then between the N-QTF prongs (Figure 3b). A comparison between the two beam profiles clearly
prong
spacing
from
s
=
1000
μm
to
s
=
700
μm
should
not
significantly
affect
the
noise
level
of
the
indicated that the THz beam is not obstructed by the QTF when it is focused between its prongs. The
QEPAS
signal.
total
image
intensity was measured by summing all pixel values for both beam profiles, from which
Sensors 2016, 16, 439
we estimated that 96.4% of the light intensity passes between the prongs. Hence, a reduction of the
4. THzspacing
QEPASfrom
Sensor
Performance
prong
s =Calibration
1000 µm to and
s = 700
µm should not significantly affect the noise level of the
QEPAS
Thesignal.
sensor calibration was performed by using a trace gas standard generator. Starting from a
certified 100 part per million (ppm) CH3OH in N2 mixture, we produced lower methanol
4. THz QEPAS Sensor Calibration and Performance
concentrations using pure N2 as the diluting gas. A laser modulation amplitude of 600 mV was
employed
in the calibration
QEPAS experiments,
similarby
to those
in [11,12].
High-resolution
QEPAS
The sensor
was performed
using reported
a trace gas
standard
generator. Starting
scans
of
CH
3
OH:N
2
calibrated
mixtures
with
different
concentrations
are
shown
in
Figure
4a,
together
from a certified 100 part per million (ppm) CH3 OH in N2 mixture, we produced lower methanol
with a spectral scan
pure
N2 flows
the modulation
ADM. The lock-in
integration
was
concentrations
usingacquired
pure N2when
as the
diluting
gas.inside
A laser
amplitude
of 600 time
mV was
set to 3 s for
employed
inall
themeasurements.
QEPAS experiments, similar to those reported in [11,12]. High-resolution QEPAS
calibration
curve shown
in Figure
4b was obtained
using CH
OH concentrations
derived
scansThe
of CH
mixtures
with different
concentrations
are 3shown
in Figure 4a, together
3 OH:N2 calibrated
fromathe
gas mixture
generator
andpure
the QEPAS
peak
signals
acquired
theintegration
related scans.
with
spectral
scan acquired
when
N2 flows
inside
the ADM.
Thefrom
lock-in
timeThese
was
results
confirm
that
the
QEPAS
signal
is
proportional
to
the
methanol
concentration.
set to 3 s for all measurements.
4. (a) QEPAS
QEPAS spectral
spectral scans
scans of
of gas
gas mixture
mixture containing
containing different
different concentrations
concentrations of
of methanol in
Figure 4.
Torr acquired
acquired with
with 33 ss lock-in
lock-in integration
integration time.
time. The
The spectral
spectral scan
scan obtained
obtained
N22 at a gas pressure of 10 Torr
for pure
pure N22 under the same operating conditions is also depicted. (b) Calibration
curve
(solid
red line)
Calibration
from the
the linear
linear fit
fit of
of measured
measured QEPAS
QEPASpeak
peaksignals
signals(‚)
(●)vs.
vs.methanol
methanolconcentrations.
concentrations.
obtained from
Figure 5 shows a comparison between the QEPAS scan measured for 100 ppm methanol in N2
The calibration curve shown in Figure 4b was obtained using CH3 OH concentrations derived from
using the N-QTF (Figure 5a) based on the scan obtained at the same concentration for the C-QTF
the gas mixture generator and the QEPAS peak signals acquired from the related scans. These results
(Figure 5b). In both cases, the QEPAS sensors operated at 10 Torr gas pressure with a lock-in
confirm that the QEPAS signal is proportional to the methanol concentration.
integration time of 3 s. A 1σ minimum detection limit (MDL) using the N-QTF was measured from
Figure 5 shows a comparison between the QEPAS scan measured for 100 ppm methanol in N2
the analysis of the signal-to-noise ratio and results in an approximately nine-times-better MDL than
using the N-QTF (Figure 5a) based on the scan obtained at the same concentration for the C-QTF
that previously obtained with the C-QTF. Since the N-QTF is characterized by a lower Q, the observed
(Figure 5b). In both cases, the QEPAS sensors operated at 10 Torr gas pressure with a lock-in integration
enhancement in the QEPAS sensitivity can be attributed to a reduction of s, thus demonstrating its
time of 3 s. A 1σ minimum detection limit (MDL) using the N-QTF was measured from the analysis of
effective influence on the QTF acousto-electric transduction efficiency. The QTF resonance frequency
the signal-to-noise ratio and results in an approximately nine-times-better MDL than that previously
reduction also contributes to the QEPAS signal enhancement. The MDL was 1.7 ppm compared to
obtained with the C-QTF. Since the N-QTF is characterized by a lower Q, the observed enhancement in
the previously reported MDL of 15 ppm, while the noise fluctuations (1σ value) recorded with the
the QEPAS sensitivity can be attributed to a reduction of s, thus demonstrating its effective influence
new QTF are ±30 μV of the same order of those reported in [11] (±25 μV). In principle, a further
on the QTF acousto-electric transduction efficiency. The QTF resonance frequency reduction also
reduction of the prong spacing can enhance the QEPAS signal even more. However, an increase of
contributes to the QEPAS signal enhancement. The MDL was 1.7 ppm compared to the previously
the background noise can be also expected, because s becomes comparable with the focused laser
reported MDL of 15 ppm, while the noise fluctuations (1σ value) recorded with the new QTF are
beam size.
˘30 µV of the same order of those reported in [11] (˘25 µV). In principle, a further reduction of the
prong spacing can enhance the QEPAS signal even more. However, an increase of the background
noise can be also expected, because s becomes comparable with the focused laser beam size.
Sensors 2016, 16, 439
Sensors 2016, 16, 439
Sensors 2016, 16, 439
6 of 8
6 of 8
6 of 8
Figure
5.
Spectral
scan
of
100
ppm
methanol
2 at
pressure
Torr
acquired
with
Figure
(a)
inin
NN
a agas
pressure
ofof
1010
Torr
acquired
with
a 3 asa
2Nat
Figure5.
5. (a)
(a)Spectral
Spectralscan
scanof
of100
100ppm
ppmofof
ofmethanol
methanol
in
2 at
a gas
gas
pressure
of
10
Torr
acquired
with
2
obtained
3lock-in
s
lock-in
integration
time
using
the
N-QTF.
(b)
Spectral
scan
of
100
ppm
of
methanol
in
N
integration
timetime
using
the N-QTF.
(b) Spectral
scanscan
of 100
of methanol
in Nin
for
2 obtained
3 s lock-in
integration
using
the N-QTF.
(b) Spectral
of ppm
100 ppm
of methanol
N2 obtained
for
the
experimental
conditions
the
with
geometry.
The
lower
data
the
experimental
conditions
using using
the
C-QTF
with a standard
geometry.
The lower
data
sampling
for same
the same
same
experimental
conditions
using
the C-QTF
C-QTF
with aa standard
standard
geometry.
The
lower
data
sampling
in
isis due
to
ramp
in
work
respect
in
panel (a)
due to(a)
a faster
rampvoltage
employed
in employed
this
work with
respect
the measurements
sampling
inispanel
panel
(a)
duevoltage
to aa faster
faster
voltage
ramp
employed
in this
this
worktowith
with
respect to
to the
the
measurements
reported
in
[11].
reported
in
[11].
measurements reported in [11].
An
variance
analysis
[24]
was
performed
in
order
to
determine
the
An Allan-Werle
Allan-Werlevariance
varianceanalysis
analysis[24]
[24]was
wasperformed
performed in
in order
orderto
todetermine
determine the
the best
best achievable
achievable
An
Allan-Werle
best
achievable
sensitivity
of
the
THz
QEPAS
sensor.
We
measured
and
averaged
the
QEPAS
signal
at
zero
CH
sensitivity of
of the
the THz
THz QEPAS
QEPASsensor.
sensor.We
Wemeasured
measuredand
andaveraged
averagedthe
theQEPAS
QEPASsignal
signalat
atzero
zeroCH
CH333OH
OH
sensitivity
OH
concentration
(pure
N
2 at 10 Torr) and obtained the Allan-Werle deviation in ppm, shown in Figure 6.
concentration (pure
(pure N
N22 at 10 Torr) and obtained the Allan-Werle
concentration
Allan-Werle deviation
deviation in
in ppm,
ppm, shown
shown in
inFigure
Figure6.6.
Figure 6. Allan-Werle deviation in ppm as a function of the lock-in integration
integration time for the QEPAS
Figure 6. Allan-Werle deviation in ppm as a function of the lock-in integration time
time for the QEPAS
sensor.
The
curve
was
calculated
by
analyzing
120-min-long
acquisition
periods of the signal
measured
sensor.
sensor. The
The curve
curve was
was calculated
calculated by
by analyzing
analyzing 120-min-long
120-min-long acquisition
acquisition periods
periods of
of the
the signal
signal
for pure Nfor
at 10 Torr
and
settingand
the lock-inthe
integrationintegration
time at 100 ms. at 100 ms.
measured
measured 2forpure
pureN
N22at
at10
10Torr
Torr andsetting
setting thelock-in
lock-in integrationtime
time at 100 ms.
For
30
averaging
time,
detection
sensitivity
of
160
part
per
billion
(ppb),
corresponding
For
corresponding
toto
a
For aaa30
30sssaveraging
averagingtime,
time,aaadetection
detectionsensitivity
sensitivityof
of160
160part
partper
perbillion
billion(ppb),
(ppb),
corresponding
to
−11
−1
½
´
11
´
1
½
anormalized
normalized
noise
equivalent
absorption
NNEA
=
3.75
×
10
cm
W/Hz
(laser
power
of
~40
μW),
−11
−1
½
noise
equivalent
absorption
NNEA
=
3.75
ˆ
10
cm
W/Hz
(laser
power
of
~40
µW),
a normalized noise equivalent absorption NNEA = 3.75 × 10 cm W/Hz (laser power of ~40 μW),
was
which
value
trace
gas
was
achieved,
represents
new
record
for
QEPAS
sensing.
These
results
clearly
wasachieved,
achieved,which
whichrepresents
representsaaanew
newrecord
recordvalue
valuefor
forQEPAS
QEPAStrace
tracegas
gassensing.
sensing.These
Theseresults
resultsclearly
clearly
demonstrate
that
the
N-QTF
geometry
provides
better
sensor
system
performances
with
respect
to
demonstrate
that
the
N-QTF
geometry
provides
better
sensor
system
performances
with
respect
to the
demonstrate that the N-QTF geometry provides better sensor system performances with respect
to
the
standard
geometry.
An
even
higher
sensitivity
can
be
expected
by
increasing
the
QTF
crystal
standard
geometry.
An
even
higher
sensitivity
can
be
expected
by
increasing
the
QTF
crystal
thickness
the standard geometry. An even higher sensitivity can be expected by increasing the QTF crystal
thickness
w;
chemical
etching
for
aacrystal
of
>>11mm
cannot
sharp
edge
w;
however,
chemical
for
a crystal
w > 1 mm
sharp edge
profiles.
thickness
w;however,
however,etching
chemical
etching
forof
crystal
ofw
wcannot
mmguarantee
cannotguarantee
guarantee
sharp
edgeprofiles.
profiles.
Conclusions
5.
5. Conclusions
Conclusions
In
QEPAS
THz
sensor
for
CH
In this
this work
work we
we reported
reported on
onaaQEPAS
QEPASTHz
THzsensor
sensorfor
forCH
CH333OH
OHdetection,
detection, implementing
implementing aa QTF
QTF
with
design,
the
prong
width
and
length
were
reduced
with
respect
to the
novel
geometry.
the
new
design,
the
prong
width
and
length
were
reduced
with
respect
to
withaaanovel
novelgeometry.
geometry.InIn
Inthe
thenew
new
design,
the
prong
width
and
length
were
reduced
with
respect
to
the
previously
employed
C-QTF
with
a
standard
design
in
order
to
reduce
the
resonance
frequency
the previously employed C-QTF with a standard design in order to reduce the resonance frequency
Sensors 2016, 16, 439
7 of 8
previously employed C-QTF with a standard design in order to reduce the resonance frequency while
maintaining a high quality factor. We also reduced the prong spacing, enhancing the acousto-electric
transduction efficiency. To evaluate the N-QTF performance, we utilized the same QEPAS THz
sensor for CH3 OH detection as reported in [11]. A MDL of 160 ppb at 30 s integration time and a
corresponding record of NNEA = 3.75 ˆ 10´11 cm´1 W/Hz½ were achieved. The CH3 OH sensing
system performance was enhanced by nearly an order of magnitude with respect to the QEPAS sensor
employing a C-QTF with a standard geometry.
Acknowledgments: The authors from the Dipartimento Interateneo di Fisica di Bari acknowledge financial
support from Italian research projects PON02 00675, PON02 00576 and PON03 “SISTEMA”. F.K. Tittel
acknowledges support by the Robert Welch Foundation (grant C-0586) and an US National Science Foundation
ERC MIRTHE award.
Author Contributions: A.S., P.P, M.G., G.S., F.K.T. and V.S. conceived and designed the experiments; A.S., P.P. and
M.G. performed the experiments and analyzed the data; M.S.V., H.E.B. and D.A.R. provided the THz QCL source;
All wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Mittleman, D.M.; Jacobsen, R.H.; Neelamani, R.; Baraniuk, R.G.; Nuss, M.C. Gas sensing using terahertz
time-domain spectroscopy. Appl. Phys. B 1998, 67, 379–390.
Van Exter, M.; Fattinger, Ch.; Grischkowsky, D. Terahertz time-domain spectroscopy of water vapor. Opt. Lett.
1989, 14, 1128–1130.
Harde, H.; Katzenellenbogen, N.; Grischkowsky, D. Terahertz coherent transients from methyl chloride
vapor. J. Opt. Soc. Am. B 1994, 11, 1018–1030.
Harde, H.; Grischkowsky, D. Coherent transients excited by subpicosecond pulses of terahertz radiation.
J. Opt. Soc. Am. B 1991, 8, 1642–1651. [CrossRef]
Vitiello, M.S.; Scalari, G.; Williams, B.; De Natale, P. Quantum cascade lasers: 20 years of challenges.
Opt. Express 2015, 23, 5167–5182. [CrossRef] [PubMed]
Vitiello, M.S.; Tredicucci, A. Tunable Emission in THz Quantum Cascade Lasers. IEEE Trans. THz Sci. Technol.
2011, 1, 76–84. [CrossRef]
Vitiello, M.S.; Consolino, L.; Bartalini, S.; Taschin, A.; Tredicucci, A.; Inguscio, M.; De Natale, P.
Quantum-limited frequency fluctuations in a terahertz laser. Nat. Photonics 2012, 6, 525–528. [CrossRef]
Williams, B.S.; Kumar, S.; Hu, Q.; Reno, J.L. High-power terahertz quantum-cascade laser. Electron. Lett.
2006, 42, 89–90. [CrossRef]
Consolino, L.; Bartalini, S.; Beere, H.E.; Ritchie, D.A.; Vitiello, M.S.; De Natale, P. THz QCL-Based
Cryogen-Free Spectrometer for in Situ Trace Gas Sensing. Sensors 2013, 13, 3331–3340. [CrossRef] [PubMed]
Consolino, L.; Campa, A.; Ravaro, M.; Mazzotti, D.; Vitiello, M.S.; Bartalini, S.; De Natale, P. Saturated
absorption in a rotational molecular transition at 2.5 THz using a quantum cascade laser. Appl. Phys. Lett.
2015, 106. [CrossRef]
Borri, S.; Patimisco, P.; Sampaolo, A.; Beere, H.E.; Ritchie, D.A.; Vitiello, M.S.; Scamarcio, G.; Spagnolo, V.
Terahertz quartz enhanced photo-acoustic sensor. Appl. Phys. Lett. 2013, 103. [CrossRef]
Patimisco, P.; Borri, S.; Sampaolo, A.; Beere, H.E.; Ritchie, D.A.; Vitiello, M.S.; Scamarcio, G.; Spagnolo, V. A
quartz enhanced photo-acoustic gas sensor based on a custom tuning fork and a terahertz quantum cascade
laser. Analyst 2014, 139, 2079–2087. [CrossRef] [PubMed]
Spagnolo, V.; Patimisco, P.; Pennetta, R.; Sampaolo, A.; Scamarcio, G.; Vitiello, M.S.; Tittel, F.K. THz
Quartz-enhanced photoacoustic sensor for H2 S trace gas detection. Opt. Express 2015, 23, 7574–7582.
[CrossRef] [PubMed]
Kosterev, A.A.; Bakhirkin, Y.A.; Curl, R.F.; Tittel, F.K. Quartz-enhanced photoacoustic spectroscopy. Opt. Lett.
2002, 27, 1902–1904. [CrossRef] [PubMed]
Kosterev, A.A.; Tittel, F.K.; Serebryakov, D.; Malinovsky, A.; Morozov, A. Applications of quartz tuning fork
in spectroscopic gas sensing. Rev. Sci. Instrum. 2005, 76, 043105:1–043105:9. [CrossRef]
Patimisco, P.; Scamarcio, G.; Tittel, F.K.; Spagnolo, V. Quartz-Enhanced Photoacoustic Spectroscopy: A
Review. Sensors 2014, 14, 6165–6206. [CrossRef] [PubMed]
Sensors 2016, 16, 439
17.
18.
19.
20.
21.
22.
23.
24.
8 of 8
Spagnolo, V.; Patimisco, P.; Borri, S.; Scamarcio, G.; Bernacki, B.E.; Kriesel, J. Mid-infrared fiber-coupled
QCL-QEPAS sensor. Appl. Phys. B 2013, 112, 25–33. [CrossRef]
Spagnolo, V.; Patimisco, P.; Borri, S.; Scamarcio, G.; Bernacki, B.E.; Kriesel, J. Part-per-trillion level SF6
detection using a quartz enhanced photoacoustic spectroscopy-based sensor with single-mode fiber-coupled
quantum cascade laser excitation. Opt. Lett. 2012, 37, 4461–4463. [CrossRef] [PubMed]
Jahjah, M.; Jiang, W.; Sanchez, N.P.; Ren, W.; Patimisco, P.; Spagnolo, V.; Herndon, S.C.; Griffin, R.J.; Tittel, F.K.
Atmospheric CH4 and N2 O measurements near Greater Houston area landfills using a QCL based QEPAS
sensor system during DISCOVER-AQ 2013. Opt. Lett. 2014, 39, 957–960. [CrossRef] [PubMed]
Patimisco, P.; Sampaolo, A.; Dong, L.; Giglio, M.; Scamarcio, G.; Tittel, F.K.; Spagnolo, V. Analysis of the
electro-elastic properties of custom quartz tuning forks for optoacoustic gas sensing. Sensor Actuators B-Chem.
2016, 227, 539–546. [CrossRef]
Kosterev, A.A.; Bakhirkin, Y.A.; Curl, R.F.; Tittel, F.K. Ultrasensitive gas detection by quartz-enhance
photoacoustic spectroscopy in the fundamental molecular absorption bands region. Appl. Phys. B 2005, 80,
133–138. [CrossRef]
Flygare, W.H. Molecular Relaxation. Acc. Chem. Res. 1968, 1, 121–127. [CrossRef]
Ma, Y.; Yu, G.; Zhang, J.; Yu, X.; Sun, R.; Tittel, F.K. Quartz Enhanced Photoacoustic Spectroscopy Using
Different Quartz Tuning Forks. Sensors 2015, 15, 7596–7604. [CrossRef] [PubMed]
Giglio, M.; Patimisco, P.; Sampaolo, A.; Scamarcio, G.; Tittel, F.K.; Spagnolo, V. Allan Deviation Plot as a Tool
for Quartz Enhanced Photoacoustic Sensors Noise Analysis. IEEE Trans. Ultrason. Ferroelectr. Freq. Control
2015. [CrossRef] [PubMed]
© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons by Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).