PROCEEDINGS OF SPIE

SPIEDigitalLibrary.org/conference-proceedings-of-spie

Photoacoustic sensing with micro-

tuning forks

U. Willer, M. Köhring, M. Mordmüller, W. Schade

U. Willer, M. Köhring, M. Mordmüller, W. Schade, "Photoacoustic sensing

with micro-tuning forks," Proc. SPIE 9482, Next-Generation Spectroscopic
Technologies VIII, 94820B (3 June 2015); doi: 10.1117/12.2182583

Event: SPIE Sensing Technology + Applications, 2015, Baltimore, Maryland,

United States

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 Invited Paper

Photoacoustic sensing with micro-tuning forks

U. Willera*, M. Köhringb, M. Mordmüllera, and W. Schadea,b
a
Clausthal University of Technology, Institute for Energy Research and Physical Technologies and
Energy Research Center of Lower Saxony (EFZN), Am Stollen 19B, 38640 Goslar, Germany,
b
Fraunhofer Heinrich Hertz Institute, Fiber Optical Sensor Systems, Am Stollen 19B, 38640 Goslar,
Germany

ABSTRACT

Different modifications of quartz enhanced photoacoustic spectroscopy (QEPAS) are discussed. An

analysis of the signal-to-noise ratio for the generated photoacoustic signal gives hints for an
optimized geometry of tuning fork and acoustic resonator. Furthermore, simultaneous or alternate
photoacoustic and electrical driving is discussed, leading the way to new detection schemes that are
capable of measuring changes in molecular relaxation dynamics.
Keywords: photoacoustic spectroscopy, QEPAS, interferometric readout, sensor design, relaxation time

1. INTRODUCTION
Photoacoustic spectroscopy is an established method for sensitive detection of trace gases. It relies
on the excitation of the target molecules by modulated laser radiation and their subsequent relaxation
by collisions that transfer the vibrational energy into translational modes. These can be detected with
a microphone since a periodic intake of translational energy into a gaseous medium leads to a
periodic expansion and contraction of the gas volume which represents a sound wave. With the
invention of QEPAS (quartz-enhanced photoacoustic spectroscopy) at Rice University 1,
photoacoustic spectroscopy gained even more acceptance since this method provides enhanced noise
immunity and a high sensitivity due to substitution of the conventional microphone by a quartz
micro tuning fork operated at its resonant frequency. The quartz tuning forks are readily available
because of their use as frequency standard within electronic circuits. For using the tuning fork as
photoacoustic transducer, the exciting laser light is focused through the gap between its two tines;
thus the sound wave is generated between the tines and this periodic pressure difference acting on
the tines can drive the tuning fork into oscillation. The deflection of the tines can be easily measured
because of the piezoelectric properties of quartz: bending is causing the generation of a piezo current
which can be amplified and measured. Application of acoustic micro-resonators leads to further
signal enhancement. Different schemes of their arrangement are now known as on-beam and off-
beam QEPAS 2. Trace detection of important gases for industrial and environmental sensing has
been demonstrated with detection limits ranging from ppm to ppt range, depending on pressure,
excitation wavelength, absorption strength, laser power and molecular parameters 3. Also the
surrounding gas mixture, e.g. humidity 4 influences the signal by promoting or delaying de-
excitation and vibrational-translational-transfer.

Next-Generation Spectroscopic Technologies VIII, edited by Mark A. Druy,

Richard A. Crocombe, David P. Bannon, Proc. of SPIE Vol. 9482, 94820B
© 2015 SPIE · CCC code: 0277-786X/15/$18 · doi: 10.1117/12.2182583

Proc. of SPIE Vol. 9482 94820B-1

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 Laser TF-
Mod. input Sensor

I Transimpedance
amplifier
Freciuertcygelxvamr
Lock -In amplifier

Mag n. I X Phase I Y

Ref. 4 Analogout 9 Signal 4

IrnnIrmeor innn', I. ncriuncrno

Figure 1. Typical system components for QEPAS. The connection between tansimpedance amplifier
and function generator is needed for electrical driving and is not used for conventional QEPAS.
Figure 1 shows the typical components used for a conventional QEPAS measurement: The
piezoelectric current is amplified by a transimpedance amplifier and further processed by a lock-in
amplifier and data acquisition system. Laser modulation is accomplished using a frequency
generator which needs to provide accuracy of typically 0.1Hz. A reference trigger is also fed to the
lock-in amplifier. The connection between frequency generator and transimpedance amplifier is used
for electrical driving and is not used for conventional QEPAS.
In this paper, different modifications of conventional QEPAS are discussed. Quartz micro tuning
forks are cost efficient and readily available; however, their design and the resulting resonant
frequency of 32 kHz are not ideal e.g. for the detection of slowly relaxing molecules. Optimization
of the geometry of the tuning fork in combination with the design of the acoustic resonator is a
promising way to enhance the signal-to-noise ratio and thus sensitivity. The use of tuning fork
materials other than quartz can result in loss of piezoelectricity, i.e., implies the need to read out the
deflection of the oscillating tines differently. Within the second section, the influence of different
design parameters is discussed and optical readout schemes are presented.
The strength of the generated signal depends linearly on exciting power for photoacoustics and
therefore the strongest laser sources available are commonly preferred. However, for sake of
miniaturization and cost efficiency, also the combination of QEPAS and LEDs is discussed. Within
the third section the combination of a broadband source, absorption path and QEPAS detection
scheme is presented and possible applications are discussed.
Section 4 concentrates on the simultaneous or alternate electrical and photoacoustic excitation of the
tuning fork and describes new sensing schemes associated. The simultaneous driving can be used to
ensure that modulation is maintained at the resonance frequency even if the background gases
change and the resonance frequency shifts. Alternate electrical and photoacoustic driving enables the
measurement of decay times instead of or in addition to amplitudes and thus eliminates errors due to
intensity fluctuations of the laser. In addition, changes in molecular relaxation times can be sensed
by logging the phase between electrical and photoacoustic excitation that ensures maximum signal.

Proc. of SPIE Vol. 9482 94820B-2

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 2. DESIGN OPTIMIZATION AND OPTICAL READOUT
Photoacoustic signal generation and noise processes
Quartz tuning forks were originally designed as frequency normal in electronic circuits, exploiting
their excellent resonant properties. The successful use of these micro tuning forks in photoacoustic
spectroscopy, as first presented by Kosterev et al. in 2002 1, is mainly based on the resonant
properties, too. However, due to their inefficient coupling to an external sound field 5, the conversion
of the photoacoustically induced sound wave into mechanical motion is supposed to be less efficient
than for conventional microphones. An optimization of this coupling of the tuning fork to the sound
field might enhance the sensitivity of QEPAS sensors considerably.
Based on this idea, a derivation of the signal generation and noise processes for resonant
photoacoustic sensors is drawn out, combining the work of former publications and adding slight
modifications 6–8. The photoacoustic signal can be expressed by:
r
(γ − 1)αPQGs1s2 p(rQTF )
SPAS = , (1)
ωr A
with γ the ratio of the specific heats and α the absorption coefficient of the gas mixture under test
and P the optical excitation power. The resonant detection cell, including the micro tuning fork, is
described by its quality factor Q at the resonance frequency ωr, the cell cross section A and the cell
constant G. The conversion of the photoacoustically induced sound pressure at the tuning fork p(rrQTF )
into an electrical signal is expressed via s1 and s2, describing the conversion from sound into
mechanical motion and from mechanical motion into electrical signal, respectively.
To estimate the sensitivity of any sensing scheme, the noise processes have to be investigated
besides the signal itself. For resonant sensing techniques using cantilevers or tuning forks, it is well
known, that the thermal motion of these elements is the limiting noise phenomenon 9–11. Therefore, a
detailed description of this particular noise will be sufficient to describe the noise for QEPAS, as
well. The noise at the resonance frequency of the QEPAS sensor system can be written as:
4k BTQ Δω
NPAS = s2 ⋅ . (2)
ωr D 2π
Here, kB is the Boltzmann constant, T is the absolute temperature, and D is the spring constant of the
resonant detector; the second square root describes the frequency bandwidth of the measurement.
Here, the spring constant D of the tuning fork can be sufficiently approximated by that of a
cantilever beam 12:
D = 0.25 Et w3 l 3 , (3)
with E the Young’s modulus of the tuning fork material and t,w, and l the thickness, width, and
length of the tuning forks prongs, respectively.
A combination of equations 1 to 3 leads to the signal-to-noise ratio for tuning fork based
photoacoustic sensors:
(γ − 1)G Q Etw3 2π r
S/NPAS = αP1s1 p(rQTF ) . (4)
4 k BT A ωr l3 Δω

Proc. of SPIE Vol. 9482 94820B-3

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 An optimization of the S/NPAS may be achieved via a reduction of the resonance frequency ωr and an
enhancement of the Young’s modulus E, however these parameters are strongly correlated to each
other and to the tuning fork dimensions. Therefore major effects are not to be expected from these
variations. As described earlier, a more favorable approach is to optimize the tuning forks dimension
for an increase of the coupling between the sound field and the tuning fork, which is expressed via s1
and p(rrQTF ) . As the optimization of these parameters implies a strong change of the tuning fork
dimensions, a versatile tool to detect the tuning fork’s motion without the need of extensive post
processing of the tuning fork is needed. The optical readout developed for tuning fork based
photoacoustic sensor, as described in 13,14, is well suited for these purposes.
Optimization of the sensor design

optical
excitation drilling for
GRIN -lens off -beam
resonator

optical
readout

gas supply

integrated
tuning fork

Figure 2. 3D-sketch of the sensor setup under optimization. A fused silica chip serves as integrated optical readout device
with a cut in tuning fork. 2 glass plates an off-beam resonator and an upper brass cover complete the miniaturized flow
through measurement chamber with a volume of about 10 mm3.

The basic idea for the enhancement of the S/NPAS is to increase both, the coupling factor s1 and the
pressure p(rrQTF ) at the tuning fork. Therefore, a new sensor layout was developed based on an
integrated interferometric readout of the tuning fork’s motion. Fig. 2 shows a 3D-sketch of the setup
including the fused silica chip with femtosecond laser inscribed waveguides 15,16 and the tuning fork
structure. An acoustic off-beam resonator 2 is positioned above the tuning fork, enabling a fiber
coupling of the excitation light within the same plane as the optical readout fibers. Two glass plates
with small cutouts are positioned besides the tuning fork to form a miniature flow through gas cell in
combination with the chip and the upper cell housing. The gas in- and output is integrated within one
of the two glass plates, as well.

Proc. of SPIE Vol. 9482 94820B-4

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 1,0

0,s

0,6

0,4

0,2

0,0

Figure 3. 3D-view of the acoustic pressure increase within the miniaturized cell volume. The pressure is depicted
normalized in false colors.
1,0

0,8

0,6

0,4

0,2

0,0

Figure 4. 2D-views of the acoustic pressure increase within the miniaturized cell volume for each of the coordinate planes.
The pressure is depicted normalized in false colors.

Simulations in Comsol Multiphysics using the thermoacoustic-solid-interaction module were done to

optimize the sound pressure field within the miniaturized measurement cell. Two effects were
exploited to achieve the intended increase in S/NPAS: First, the tuning fork was designed with a
bigger width, leading to an increase of the area of the tuning fork which can interact with the sound
field. Second, the volume between the tuning fork’s prongs was designed to accommodate part of
the resonance from the acoustic off-beam resonator, thus forming sort of a T-shaped acoustic
resonator. The results of the simulations are shown in Fig. 3 in a 3D-view and in Fig. 4 in a 2D-view
for each of the coordinate planes. It can be seen, that a clear resonance is forming out between the
tuning fork’s prongs. A comparison to simulations of conventional on-beam QEPAS resonator
systems shows a sound pressure increase at the tuning fork of about 2.1. The increased width of the
tuning fork leads to a considerably increased interaction area of the sound pressure with the tuning
fork. In combination with the increased sound pressure, a noticeable enhancement of the S/NPAS for
tuning fork based photoacoustic sensors should be possible. However, several development steps
have to be made before the simulations can be verified with experimental results. Some development
steps have already been completed and are shown in the following paragraph.
First experimental results and outlook
The first development step was the inscription of an integrated interferometer within a fused silica
chip (refer to Fig. 5). This was done using a femtosecond laser system; the process and the
waveguide structures are described more detailed in 16. An evanescent coupling between the two

Proc. of SPIE Vol. 9482 94820B-5

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 waveguides entering the chip from the left side splits the optical readout light in two parts, aiming at
the top of the tuning fork and the end of the chip. The reflected light from both beam paths interferes
when passing the evanescent coupling region for the second time and contains the information of the
tuning forks movement. A fiber coupled photodiode can be used to measure the interference signal.

230 pm

mm

40 mm

Figure 5. Photograph of the fused silica chip with integrated interferometer and micro tuning forks. As the small index
contrast of the waveguide structures cannot be seen on the picture, a sketch of the waveguide arrangement is shown along
with the dimensions of the chip.

A resonance curve of a fused silica tuning fork read out with an integrated interferometer is
presented in Fig. 6. The tuning fork resonance frequency was tailored via an adaption of the tuning
fork dimension to match that of conventional quartz tuning forks. This will simplify a comparison of
the new sensor technique and standard QEPAS sensors in future work. The quality factor for the
tuning fork under test was estimated to be 874. This value is far below the quality factors that are
achievable with quartz tuning forks but can easily be explained with defects caused by the diamond
cutting process and the amorphous properties of the substrate material. However, a further
optimization of the fused silica tuning forks might be possible, as Q-factors up to 1500 were
achieved at other resonance frequencies.

55 -
f(r) = (32504.4 ± 0.1)
50 -
Q = 874 ± 1
45 -
40 -
> 35 -
E _

30 -
c -

25-
20 -

15

10

0 i i i i i i

32040 32160 32280 32400 32520 32640 32760 32880

Frequency, Hz

Figure 6. Resonance curve of a tuning fork integrated within a fused silica chip (black dots). The curve represents the graph
of a numerical fit to the data using the square root of a Lorentian.

The upcoming work will have to show, if this decrease in the tuning fork Q-factor can be
compensated with the enhanced sound sensitivity, presented above.

Proc. of SPIE Vol. 9482 94820B-6

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 3. QEPAS IN COMBINATION WITH BROAD BAND SOURCES
The photoacoustic signal depends linearly on the absorbed power; therefore the excitation sources
are usually chosen to be as strong as possible. However, for reasons of setting up cost efficient
systems that need small amounts of power for operation, light emitting diodes would be a promising
alternative. This approach was followed for the setup of a system operated at a facility for biogas
generation. To reach the needed accuracy for measurements around 60% of methane, a combination
of conventional absorption spectroscopy and a detector based on QEPAS has been used. Here
selectivity is achieved by filling the cell around the tuning fork with the target gas and sealing it. The
light is aimed through an absorption cell filled with the gas to be analyzed before reaching the
QEPAS cell. A decrease in generated QEPAS signal occurs only if spectral components that
coincide with absorption lines of the target gas are absorbed. Figure 7 shows a photograph of the
sensor setup. Further details can be found in 17.

Absorption Path Preamplifier

LED -Mount QEPAS Cell

Gas Tubing Vibration Damping

Figure 7. Photograph of a sensor system, combining conventional absorption spectroscopy with a

selective QEPAS detection scheme thus enabling the use of a broadband excitation source (LED).

4. ELECTRICAL AND PHOTOACOUSTIC COOPERATION

Simultaneous electrical and photoacoustic excitation
Quartz micro tuning forks have been developed as frequency standards. For this purpose they are
integrated into a driving circuit, most commonly a pierce oscillator 18, to crop up their resonance
frequency. Normally, quartz tuning forks are operated within an evacuated housing. However, for
use as an acoustooptic transducer, the housing has to be removed in order to allow light focusing
between the prongs and gas exchange in the volume surrounding the tuning fork. This leads to
changes in friction and thus the resonance frequency and the quality factor change with changes in
background gas. For small concentration changes of trace gases and a constant background gas
mixture, this dependence can be dealt with by frequent calibration and determination of the
resonance frequency. However, when variable high concentrations or a strongly changing
background are measured, a different method of maintaining operation at the respective resonant
frequency is needed. This can be accomplished by simultaneous electrical and photoacoustic
excitation. While the driving circuit always crops the resonance frequency, it is fed back as trigger
for the laser modulation, thus also the expensive accurate frequency generator is no longer needed. A
system operating in this mode is described in 19.

Proc. of SPIE Vol. 9482 94820B-7

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 Alternate electrical and photoacoustic excitation
If an oscillator is driven by an external force and is then released from this excitation, the oscillation
ceases with a typical decay time which reflects itself in the quality factor of the oscillator. If the
driving is done at the same frequency with two different forces, the efficiency of the driving depends
on the phase between the two driving forces: if they are in phase, they add constructively and the
driving is promoted by a larger force, if they are shifted by π, the driving force is most efficiently
reduced. Retardation or maintenance of the oscillation is also possible if the two forces are applied
alternately: The characteristic decay time after stopping application of force 1 is prolonged if driving
is continued with in-phase force 2 or it is maximum shortened if the phase of force 2 is shifted by π
with respect to force 1.
As pointed out in the introduction, photoacoustic signal generation involves several steps, in
particular the relaxation of the excited molecules and the VT-transfer process. This implies that
photoacoustic driving is not instantaneous but that there is a time lag between the periodic laser
modulation and the induced periodic pressure variation that highly depends on the molecular
relaxation dynamics. Thus the phase difference between the electrical driving and the photoacoustic
one for which the signal is either maximized or minimized gives information about relaxation times.
Due to retardation in transmission lines for electrical driving and laser modulation the values do not
correspond to absolute values for the de-excitation times; however, to estimate the value for the
measured phase that resembles an actual phase shift of π between laser modulation and electrical
driving, the tuning fork was driven by directing laser light to its surface. This driving does not
involve molecular relaxation processes and is taken to be instantaneous. The derived value
representing a shift by π is indicated by the vertical line in the inset in Fig. 8. Furthermore, relative
shifts between measurements with different experimental conditions can also be quantified. The
inset of Fig. 8 shows the measured lock-in signal (R) for simultaneous electrical and photoacoustic
excitation as a function of phase shift of the electrical driving with respect to the photoacoustic one
for different concentrations of methane in nitrogen. The resonance frequency was determined for
every gas mixture and used for the respective measurement. Also the amplitudes of the two driving
forces were equalized for all measurements to reach total cancellation. Figure 8 by itself shows the
phase values for destructive addition of the two driving forces as a function of methane
concentration as derived from these measurements. A change of phase with methane concentration is
expected since the probabilities for collisions with other methane molecules or with nitrogen
changes and the relaxation times associated with self-de-excitation and due to collisions with
nitrogen differ considerably. For self-de-excitation of methane Yardley et al. give a relaxation time
of 1.90±0.1 µs atm and for the collision with nitrogen 16.3±1 µs atm 20. A higher value of the phase
of electrical excitation is associated with an earlier photoacoustic excitation. Thus, the increase in
relaxation time for 90% methane cannot be explained by these relaxation times. Further work will
focus on the identification of de-excitation paths that can explain this increase. If the absolute values
are determined, the optically induced oscillation can be most efficiently reduced, if the electrical
driving is retarded by 10 µs for 10% methane and 7.88 µs for 70% methane, respectively.

Proc. of SPIE Vol. 9482 94820B-8

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 350 -

345 -

340 -
120
Met hane concentration
10%
335 - 100
50%
- - --70%
É 80 - - - -- 90%
330 - To

t60

325 - 3 40

20

320 -
0
50 100 150 200 250 300 350 400

Phase shift of electrical excitation (deg)

315 I I I

0 20 40 60 80 100

Methane concentration ( %)

Figure 8. Phase shift of electrical excitation for destructive addition of driving forces as a function
for methane concentration. The inset shows the raw data. The vertical line resembles the phase for
which laser modulation and electrical driving have a phase lag of π as derived from a different
experiment.
For QEPAS, the influence of water vapor on the molecular relaxation dynamics is frequently used to
promote the VT-transfer and to achieve lowest detection limits 4. Figure 9 shows a direct
measurement of the influence of the presence of water vapor on the relaxation times. For these
measurements methane and dry nitrogen have either been mixed as before or the nitrogen was
humidified by conveying the gas stream through a wash flask filled with distilled water. Different
flow rates result in different amounts of water vapor within the nitrogen. The influence on the
relaxation times can clearly be seen. The two humidified samples show an earlier photoacoustic
excitation compared to the dry background gas by 1.87µs and 2.43µs, respectively.

Proc. of SPIE Vol. 9482 94820B-9

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 300
10% methane ---- with wash flask
250 with wash flask
dry nitrogen
dry nitrogen
.-. 200
E

150 .
cm

`
100 ,<< \
.
.

O
J ``` \
50

,-, ..

0
i I

20 40
i i ii ii ii ii
60 80 100 120
i

140
ii
160 180

Phase -shift electrical excitation (deg)

Figure 9. Effect of humidified background gas as promoter of molecular relaxation.

With the phase set to the optimum value for minimum combined excitation, measurements with
alternate driving are performed. To do so, the tuning fork is first driven into oscillation by electrical
driving. After switching off the electrical excitation, the decay time is shortened due to application
of phase shifted photoacoustic driving, i.e. the oscillation is actively slowed down. Figure 10 (a)
shows the time dependent curves for different concentrations of methane in nitrogen. For sections
can be determined for each curve: The first plateau indicates the sole electrical driving of the tuning
fork. The driving amplitude is fixed to a constant value for all the measurements, thus the difference
in lock-in signal, i.e. amplitude of the generated oscillation indicates a change in quality factor due
to changes in friction as the background gas changes considerably. After switching off the electrical
excitation and simultaneous start of photoacoustic driving with flipped phase, no exponential decay
curves are found but steeper retardation of the oscillation. After crossing a minimum, the signal
grows again, now exponentially as can be seen by the good resemblances of the fittet curves. Finally
a second plateau is reached that corresponds to the sole photoacoustic excitation, thus representing a
conventional QEPAS measurement. To not disturb the dynamics by the influence of an acoustic
resonator, these measurements have been performed within a small cell with a bare tuning fork
without acoustic resonator. The light path in front of the tuning fork is about 2mm, which has to be
taken into account since even on this path considerable absorption is present at the 1277 cm-1
absorption line of methane.
All the different sections of the curves can be used for determination of concentrations: Figure 10 (b)
shows the dependence of the lock-in signal for sole electrical excitation as a function of methane
concentration. A linear dependence is found and corresponds nicely with the determined linear
behavior of the Q-factor as given in part (d). The other graphs show not the analyzed data from part
(a) but corrected values taking into account that the transmitted intensity reaching the tuning fork

Proc. of SPIE Vol. 9482 94820B-10

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 differs for the changes in methane concentration. Figure 10 (c) shows the corrected rise time which
should also be only dependend on the quality factor and shows also a linear behavior. Part (c) shows
the decay times. Since the generated sound pressure is a function of concentration the decay times
can be used to determine concentrations, their dependence is nonlinear. Finally, part (d) shows the
dependence of the generated photoacoustic signal on concentration.
In this case the knowledge of the concentration was used a-priori to correct for the influence of the
concentration on the light path leading to the tuning fork; however, this was done only to work out
the dependences of the analyzed data on concentration. For real application, the discussed different
data that can be extracted from the time series and their different dependence on concentration can
be used to derive concentration data. A further parameter that can be used is the phase needed for
signal minimization for simultaneous excitation as discussed in the first paragraph of this section.

electrical excitation only

80 - 75 -

5. 70
60
65 -

40 -
55 -

20 - (b)
50
100

Concentration (%)
0- 2400

2200 -

-20 - -Ei 2000 -

iaoo-
-40 -
1600 -

r
F2
1400 -
-60 - W

1200 -
%) (c)
tE 1000
O 2000 4000 6000 8000 10000 100

Time (arb. un ts) Concentration (%)

10500 3500

250 - corrected quasi QEPAS signal

10000 ;
1 3000 -
200 -
6500

CS 2500 .E 150
8 woo,
2
Frequency (Hz)
2 2000 - fl 100

8000 -
t) 1500

7500 - (d) (e) (f)

100 100

Concentration (%) Concentration (%) Concentration (%)

Figure 10. Alternate electrical and photoacoustic driving. (a) measured signal as a function of time,
(b) signal due to electrical excitation only, (c) rise time of photoacoustic signal, (d) Q-factor
determined from measurements of resonance curves as given in the inset, (e) decay time as function
of concentration, (f) photacoustically generated signal.

REFERENCES

[1] Kosterev, A. A., Bakhirkin, Y. A., Curl, R. F. and Tittel, F. K., “Quartz-enhanced photoacoustic spectroscopy,” Opt.
Lett. 27, 1902–1904 (2002)
[2] Liu, K., Guo, X., Yi, H., Chen, W., Zhang, W. and Gao, X, “Off-beam quartz-enhanced photoacoustic
spectroscopy,” Opt. Lett 34, 1594–1596 (2009).

Proc. of SPIE Vol. 9482 94820B-11

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

 [3] Patimisco, P., Scamarcio, G., Tittel, F. K. and Spagnolo, V., “Quartz-enhanced photoacoustic spectroscopy: a
review,” Sensors 14, 6165–6206 (2014)
[4] Dong, L., Wright, J., Peters, B., Ferguson, B. A., Tittel, F. K. and McWhorter, S., “Compact QEPAS sensor for trace
methane and ammonia detection in impure hydrogen,” Appl. Phys. B 107, 459-467 (2012)
[5] Russell, D. A., “On the sound field radiated by a tuning fork,” Am. J. Phys. 68, 1139–1145 (2000)
[6] Kreuzer, L. B., “The Physics of Signal Generation and Detection,” in Optoacoustic Spectroscopy and Detection
(Pao, Y.-H., ed), Elsevier Inc, pp. 1–25, (1977)
[7] Sigrist, M. W., “Air Monitoring by Laser Photoacoustic Spectroscopy”, in Air monitoring by spectroscopic
techniques (Sigrist, M. W., ed), John Wiley & Sons, Ltd, pp. 163–179 (1994)
[8] Miklos, A., Hess, P. and Bozoki, Z., “Application of acoustic resonators in photoacoustic trace gas analysis and
metrology,”Rev. Sci. Instrum. 72, 1937–1955 (2001).
[9] Gabrielson, T. B., “Fundamental noise limits for miniature acoustic and vibration sensors,” J VIB ACOUST 117,
405–410 (1995)
[10] Grober, R. D., Acimovic, J., Schuck, J., Hessman, D., Kindlemann, P. J., Hespanha, J. et al., “Fundamental limits to
force detection using quartz tuning forks,” Rev. Sci. Instrum. 71, 2776–2780 (2000)
[11] Colchero, J., Cuenca, M., Martínez, J. F. González, Abad, J., García, B. P., Palacios-Lidón, E. et al., “Thermal
frequency noise in dynamic scanning force microscopy,” Journal of Applied Physics 109, (2011)
[12] Butt, H. J. & Jaschke, M., “Calculation of thermal noise in atomic-force microscopy,” Nanotechnology 6, 1–7
(1995)
[13] Köhring, M., Pohlkötter, A., Willer, U., Angelmahr, M. and Schade, W., “Tuning fork enhanced interferometric
photoacoustic spectroscopy: a new method for trace gas analysis,” Appl. Phys. B 102, 133–139 (2011)
[14] Köhring, M., Böttger, S., Willer, U. and Schade, W., “Temperature effects in tuning fork enhanced interferometric
photoacoustic spectroscopy,” Opt. Express 21, 20911–20922 (2013)
[15] Thiel, M., Flachenecker, G., Köhring, M. and Schade, W., “Direct femtosecond laser writing of waveguide
structures and Bragg gratings for integrated NIR optics using multi scan technique,” Optical Society of America,
Munich. (2013)
[16] Thiel, M., Flachenecker, G. and Schade, W., “Femtosecond laser writing of Bragg grating waveguide bundles in
bulk glass,” Opt. Lett. 40, 1266–1269 (2015)
[17] Köhring, M., Böttger, S., Willer, U. and Schade, W., “LED-Absorption-QEPAS Sensor for Biogas Plants,” Sensors
submitted (2015)
[18] Matthys, R. J.,[Crystal oscillator circuits], Wiley-Interscience, New York (1983)
[19] Mordmüller, M., Köhring, M., Schade, W. and Willer, U., „An electrically and optically cooperated QEPAS device
for highly integrated gas sensors,” Appl. Phys. B, 1-8. doi./10.1007/s00340-015-6037-9 (2015)
Yardley, J. T., Fertig, M. N. and Moore, C. B., “Vibrational Deactivation in Methane Mixtures,” The Journal of
Chemical Physics 52, 1450–1453 (1970)

Proc. of SPIE Vol. 9482 94820B-12

Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 1/5/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use