Home Search Collections Journals About Contact us My IOPscience

A poly(vinylidene fluoride)-coated ZnO film bulk acoustic resonator for nerve gas detection

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2011 J. Micromech. Microeng. 21 085017

(http://iopscience.iop.org/0960-1317/21/8/085017)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 132.239.1.231
This content was downloaded on 24/04/2017 at 08:31

Please note that terms and conditions apply.

You may also be interested in:

Highly sensitive ZnO thin film bulk acoustic resonator for hydrogen detection
Da Chen, Jing-jing Wang, Qi-xin Liu et al.

A film bulk acoustic resonator oscillator based humidity sensor with graphene oxide as the
sensitive layer
Weipeng Xuan, Marina Cole, Julian W Gardner et al.

A film bulk acoustic resonator in liquid environments

Hao Zhang, Mong S Marma, Eun Sok Kim et al.

Temperature compensation of liquid FBAR sensors

J Bjurström, G Wingqvist, V Yantchev et al.

Monolithic integrated system with an electrowetting-on-dielectric actuator and a

film-bulk-acoustic-resonator sensor
Menglun Zhang, Weiwei Cui, Xuejiao Chen et al.

Real-time monitoring of human blood clotting using a lateral excited film bulk acoustic resonator
Da Chen, Jingjng Wang, Peng Wang et al.

A micromachined thermally compensated Lamb wave resonator

G Wingqvist, L Arapan, V Yantchev et al.

A highly sensitive humidity sensor based on a nanofibrous membrane coated quartz

crystalmicrobalance
Xianfeng Wang, Bin Ding, Jianyong Yu et al.
IOP PUBLISHING JOURNAL OF MICROMECHANICS AND MICROENGINEERING
J. Micromech. Microeng. 21 (2011) 085017 (6pp) doi:10.1088/0960-1317/21/8/085017

A poly(vinylidene fluoride)-coated ZnO

film bulk acoustic resonator for nerve
gas detection
Da Chen1 , Jingjing Wang2 , Dehua Li1 , Yijian Liu1 , Hongwei Song1
and Qixin Liu1
1
Department of Applied Physics, Qingdao Key Laboratory of Terahertz Technology, College of Science,
Shandong University of Science and Technology, Qingdao 266510, People’s Republic of China
2
Jinan campus, Shandong University of Science and Technology, Jinan 250031,
People’s Republic of China
E-mail: phychenda@163.com

Received 13 January 2011, in final form 18 May 2011

Published 11 July 2011
Online at stacks.iop.org/JMM/21/085017

Abstract
We apply the film bulk acoustic resonator for the detection of nerve gas. The resonator is
consisted of a ZnO piezoelectric stack and a W/SiO2 Bragg reflector. Poly(vinylidene
fluoride) (PVDF) is used as the sensing coating to adsorb the analyte under test. The testing
results show that our proposed sensor can yield a sensitive, reversible and reproducible
response to nerve gas. The relationship between the frequency shifts and the concentrations of
the nerve gas exhibits a perfect linear correlation in the range of 10–50 ppm. The gas
sensitivity of the proposed sensor is 718 kHz ppm−1 , which is several orders of magnitude
higher than that of a quartz crystal microbalance with the same sensitive coating. This study
proves that it is feasible to use the PVDF-coated thin film bulk acoustic resonator for the
detection of the traced nerve gas.
(Some figures in this article are in colour only in the electronic version)

1. Introduction loading measurements [9, 10]. In principle, FBARs and

QCMs are both electro-acoustic resonators. With a working
Chemical warfare agents, especially nerve gas, are highly resonance of 2–10 GHz, FBARs have a minimum detectable
toxic and extremely dangerous. Using nerve gas as mass change in the range of ∼10 ng cm−2 , which is several
a measure of terrorism has been a real existing threat. orders of magnitude lower than that of the QCMs, whose
Accordingly, the study on sensitive detection of nerve gas working resonance is several MHz. Today many attempts to
has become one of the most active pieces of research of use FBARs for chemical and biological sensor applications
sensor technology. Spectrographic analysis can give precise
have been made with remarkable results, including liquid
and reliable analytical results, but it requires expensive
sensing [11, 12], gas sensing [13] and biological detection
instrumentation and a long time. Besides, a wide range
[14–16]. However, the sensitivity of the electro-acoustic
of technologies, including electrochemical sensors [1],
metal oxide semiconductor devices [2], carbon nanotubes- resonators mainly depends on the sensing coating used to
based devices [3], microcantilevers [4, 5], surface acoustic adsorb the analyte under test and may produce a high-
wave devices [6] and quartz crystal microbalances (QCMs) frequency shift in the resonators for a given gas concentration.
[7, 8] has been used as a platform for the detection of In order to detect nerve gas, some chemical recognition agents,
nerve gas. such as zeolite layers [17], aminoferrocene derivatives [18],
In recent years, film bulk acoustic resonators (FBARs) liquid crystals [19] and some polymers [8, 20–22], have been
have attracted great attention because of their promising used. In addition, it has been proven that poly(vinylidene
applications in radio-frequency-integrated circuits and mass- fluoride) (PVDF) is very sensitive to nerve gas [23].

0960-1317/11/085017+06$33.00 1 © 2011 IOP Publishing Ltd Printed in the UK & the USA
J. Micromech. Microeng. 21 (2011) 085017 D Chen et al

(a)

Figure 2. The chemical structure of PVDF and DMF.

(b)

Figure 3. Schematic diagram of the experimental apparatus for

Figure 1. (a) The basic configuration of the PVDF-coated FBAR, DMMP gas detection.
(b) cross-sectional view of the morphology of the ZnO film and
Bragg reflector. 2.2. Coating of the sensitive layer
PVDF was used as the sensitive material and coated onto the
In this paper, we report a nerve gas sensor based
Au surface by a drop-coating method. A small amount of
on a FBAR coated with a PVDF layer. The sensitivity
PVDF powder (average molecular weight ∼534 000, from
characterization for the detection of nerve gas is evaluated. Our Aldrich Chemical Co.) were dissolved in N, N-dimethyl
proposed FBAR-based sensor provides higher gas sensitivity formamide (DMF, Jufeng Chemical Co., China) with the
than QCM. In addition, the advantages of the FBAR-based concentration of 0.05 M, and then the solution was dropped
sensor, including high sensitivity, small size, simple carefully onto the center of the electrode by a 1 μl microsyringe
fabrication process and ease of detection method, make the (Shanghai Zhouhui Co., China). Subsequently, the coated
early alarm of nerve gas promising. FBAR was baked at 100 ◦ C for 1 h to obtain stable films.
After the solvent was evaporated, a solid PVDF layer was
fixed on the Au surface. The thickness of the PVDF layer
2. Experimental details was measured to be about 70 nm by an optical profiler (Veeco
Wyko NT1100). Figure 2 shows the chemical structure of
2.1. Fabrication of the resonator PVDF and DMF.
Figure 1 shows the basic configuration and the cross-sectional
view of the morphology of the sensor device. The FBAR 2.3. The sensor-testing system
consists of a ZnO piezoelectric stack (Au/ZnO/Al with the In operation, the sensor was packaged in a PCB three-pin
thicknesses of tAu = 100 nm, tAlN = 1.5 μm, tMo = 200 nm) case connected to a network analyzer (Agilent 8722D). The
and a Bragg reflector (SiO2 /W with the thickness of λ/4). frequency responses were measured by the network analyzer
The films were deposited using the JGP-800 sputtering system every 30 s, and the data were loaded to a computer to be
(SKY Techno. Co., China). The fabrication process of the analyzed. The schematic diagram of the testing apparatus is
FBAR has been described in other papers [24, 25]. The shown in figure 3.
depositing parameters of ZnO were optimized to obtain a high In this experiment, dimethyl methyl phosphonate
c-axis orientation and good quality. As shown in figure 1(b), (DMMP) simulates nerve gas because of the presence of
the ZnO film has a dense columnar structure and the interfaces P=O, P–CH3 and P–OCH3 functional groups. The testing
between the layers are clearly visible. The top Au electrode temperature is 5–40 ◦ C. Different concentrations of DMMP
with the Cr adhesion layer was evaporated at room temperature vapor were obtained by multiple dilutions [26]. Saturated
and patterned by means of a lift-off process using AZP4903 DMMP vapor was collected in a 50 ml syringe from the
photoresist. headspace above the liquid solvent (97%, from Aldrich

2
J. Micromech. Microeng. 21 (2011) 085017 D Chen et al

Figure 4. The impedance curves (absolute value and phase)

obtained from the FBAR uncoated with the PVDF layer.
Figure 5. The S11 responses of the FBAR device before and after
coating with the PVDF layer.
Chemical Co.) in a closed container. The vapor was
diluted by injecting the vapor sample into a 100 ml dilution
chamber containing purified N2 . The vapor samples with
low concentrations were obtained after repetitive dilutions
were carried out. The carrier gas N2 controlled by a
mass flow controller was first allowed to pass through the
test chamber (500 ml volume) until the baseline frequency
response was stable. Then the DMMP vapor was extracted
and injected into the test chamber. After the vapor sample
was dispersed, the frequency response of the sensor was
measured at the equilibrium state. The test chamber was
purged with N2 between each gas injection. When N2
passed through the chamber, the desorption occurred and the
frequency response returned to the original baseline. The
vapor sample was repeatedly injected in the same manner
to check the dependence of the sensor responses on the gas
concentrations. Figure 6. The S11 responses of the PVDF-coated FBAR to the
various DMMP concentrations from 10 to 50 ppm.

3. Results and discussion 3.2. The sensitivity of the FBAR-based sensor

3.1. The performance of the FBAR Figure 5 shows the S11 responses of the FBAR device before
and after coating the PVDF layer at 20 ◦ C. Before coating
Figure 4 shows the impedance curves of the absolute value PVDF, the resonance peak (the minimum point of S11 curve)
and phase obtained from the FBAR uncoated with the PVDF is observed at 2.062 GHz with a return loss of about −14.6 dB.
layer at 20 ◦ C. The key performance parameters of the FBAR After coating PVDF, the resonance frequency is decreased to
are the Q-factors (Qs and Qp ) and effective electromechanical 1.916 GHz and the return loss shifts to −11.1 dB due to the
coupling coefficient K2 eff . The Q-factors and K2 eff can be mass loading onto the Au electrode.
calculated using the following equations: The sensitivity to DMMP vapor of the FBAR sensor
  was analyzed by measuring the S11 parameter in a range
1 ∂ Z of 200 MHz around the resonance frequency. Figure 6
Qs/p = fs/p (1)
2 ∂f s/p shows the S11 responses of the FBAR sensor to the various
DMMP concentrations of 10–50 ppm at 20 ◦ C. It is very
  clear that the resonance frequency is proportionally decreased
π2 fp − fs
2
Keff = , (2) with increasing concentration. This suggests that the PVDF
4 fp
layer absorbs the DMMP molecules and the mass loaded on
where  Z is the phase of the impedance, f s and f p are the the Au surface is significantly increased under test. Limited
series and parallel resonance frequencies, respectively. As by the capability and stability of the testing apparatus, the
shown in figure 4, f s and f p are 2.094 and 2.045 GHz, and results of concentrations lower than 10 ppm cannot be precisely
the calculated Qs , Qp and K2 eff amount to 671, 654 and 5.78%, measured in this experiment.
respectively, which are typical values of a high-performance In the PVDF molecules, strong hydrogen-bond acidity can
FBAR [9]. be obtained by the electron-withdrawing effect of the fluorine

3
J. Micromech. Microeng. 21 (2011) 085017 D Chen et al

Figure 7. Hydrogen-bonding interactions between fluorinated

hydrogen-bond acidic molecules PVDF and DMMP, a (a)
hydrogen-bond basic organophosphorus compound.

atoms, while they contain no hydroxylic oxygen atoms with

hydrogen-bond basicity. DMMP can be seen as a basic gas (b)
because the oxygen atoms in carbonyl groups have a higher
electronegativity than the phosphorus atoms. As shown in
figure 7, it is assumed that the quasi-hydrogen bonding [27]
occurs due to the fact that the hydrogen atoms of PVDF and
the oxygen atoms of DMMP have acid and basic properties,
respectively.
Figure 8 shows the real-time measurement of the device
for different concentrations. After each DMMP exposure,
N2 was purged through the testing cell until full desorption
was achieved. Reversibility of the response was observed
when DMMP vapor was switched off. The response time is
within 10 min and the recovery time of the proposed sensor
is about 15 min, which approximate the times of the QCM- Figure 8. (a) The real-time measurement of the PVDF-coated
based DMMP sensor using PVDF coating [23]. It is assumed FBAR for different concentrations at 20 ◦ C, (b) the dependence of
that the response time could be shortened by reducing the the frequency shift on DMMP concentration.
thickness of the PVDF layer. In a later report, the relationship
of response and the thickness of the sensing layer will be
described in detail. Figure 8(b) shows the dependence of the
frequency shifts on the DMMP vapor concentrations. The
regression coefficient R2 is greater than 0.998, indicating that
the frequency shifts of the sensor are linear to the vapor
concentrations in the testing range. The gas sensitivity of
our proposed sensor, defined as the frequency shift per unit
concentration, is as high as 718 kHz ppm−1 , which is much
higher than that of QCM. In comparison, the sensitivity of the
PVDF-coated QCM has been reported to be 3.19 Hz ppm−1
[23]. In addition, the FBAR sensor shows a quite stable
performance over a long period of time. As shown in figure
9, the responses of the sensor exposed to 40 ppm DMMP are
almost identical over 30 days.
The limit of detection (LOD) of our proposed sensor was Figure 9. The responses of the PVDF-coated FBAR exposed to
calculated from the value of gas sensitivity. The noise level of 40 ppm DMMP for 30 days.
the FBAR sensor is approximately 100 kHz, and taking into
account the signal-to-noise ratio of 3:1, the theoretical LOD of the vapors were produced at room temperature by a
of the sensor is 0.42 ppm. This LOD value is very low and of conventional bubble method. Figure 10 shows the frequency
interest for practical applications. responses of the sensor to 40 ppm DMMP gas and the
interferences at saturated concentrations. The sensitivity of
3.3. Sensitivity to interferences the sensor to DMMP vapor is obviously higher than to the
In order to evaluate the sensing selectivity to the potential interferences. This good selectivity of the PVDF to DMMP
gas interferences, the PVDF-coated FBAR sensor was also has also been observed in QCM [23].
tested against several conventional volatile organic vapors, The partition coefficient is a measure of the sorption
including dimethyl acetamide (DMA), acetone and hexane, of a vapor from the gas phase into a sorbent phase serving
all at saturated concentrations. The saturated concentrations as the solvent. Partition coefficients also help to define

4
J. Micromech. Microeng. 21 (2011) 085017 D Chen et al

Figure 10. The frequency responses of the PVDF-coated FBAR to

Figure 11. The responses of the PVDF-coated FBAR sensor at
40 ppm DMMP gas and the interferences at saturated vapor
various temperatures.
concentrations.

Table 1. The log of partition coefficients Ka and the hydrogen-bond

 
solvation parametersb ( α2H and β2H ) for DMMP and the
interference vapors.
 
Vapor Log Ka α2H b
β2H b

DMMP 3.08 0 1.05

DMA 2.16 0.0 0.73
Acetone 0.82 0.04 0.5
Hexane 0.94 0 0
a
Values are calculated from the results
of figure 10.
b
Values are from [28].

the gas selectivity of the sensitive coatings [29]. For Figure 12. The responses of the PVDF-coated FBAR sensor for
a gravimetric polymer-coated acoustic wave sensor, the various relative humidity levels.
response, a frequency shift denoted by fV , is related to the
partition coefficient K and vapor concentration CV according 3.4. The influence of temperature and humidity effect
to
Figure 11 shows the influence of the temperature on the
fV = fS CV K/ρS , (3) responses of the PVDF-coated FBAR sensor. The resonance
where  fS is the frequency shift due to the sensitive PVDF frequencies of the PVDF-coated FBAR are examined before
layer on the bare sensor and ρ S is the density of PVDF. The and after introducing 40 ppm DMMP vapor at a fixed
values of the log of the partition coefficients for DMMP and temperature. Obviously, the base frequency of the sensor
the interference vapors, determined from the FBAR frequency is nearly stable, but the gas responses of the sensor are
shifts, are shown in table 1. Comparison of the partition significantly increased below 20 ◦ C. The maximal response
coefficients of DMMP with other vapors indicates the high is obtained in a low-temperature range, which is in good
sensing selectivity of the PVDF to DMMP. agreement with another published result [23], where it
The hydrogen-bond solvation parameters for DMMP, is ascribed to the glass transition temperature of −15
◦
DMA, acetone and hexane are shown in table 1. The C for PVDF. Because the response of the PVDF-coated
 H  H
parameters α2 and β2 are the measures of strength FBAR sensor is affected by the temperature, temperature
compensation should be taken into consideration for practical
of the hydrogen-bond acidity and hydrogen-bond basicity,
applications.
respectively. Organophosphorus compounds are particularly
 H The influence of humidity on the responses of the PVDF-
strong hydrogen-bond bases and the β2 value of DMMP coated FBAR sensor is shown in figure 12. The base frequency
 H
is 1.05. For comparison, the β2 values for other volatile of the sensor is decreased from 1915 to 1864 MHz when
organic compounds such as DMA, acetone, and hexane are the relative humidity is increased from 10 to 80% due to the
0.73, 0.5 and 0, respectively. Because PVDF molecules are absorption of water vapor onto PVDF. However, the responses
strong hydrogen-bond acids, high sensing selectivity of PVDF to 40 ppm DMMP at various relative humidities were almost
to DMMP can be obtained. identical. The frequency shifts stabilize at about 30 MHz

5
J. Micromech. Microeng. 21 (2011) 085017 D Chen et al

with various relative humidities, indicating that the influence [7] Kim N, Park I S and Kim D K 2007 Biosens. Bioelectron.
of humidity on the sensing response can be negligible. 22 1593
[8] Huang J, Jiang Y, Du X and Bi J 2010 Sensors Actuators B
146 388
4. Conclusions [9] Satoh Y, Nishihara T, Yokoyama T, Ueda M and Miyashita T
2005 Japan. J. Appl. Phys. 44 2883
In summary, we use the PVDF sensitive layer to construct [10] Wingqvist G 2010 Surf. Coat. Technol. 205 1279
a FBAR-based nerve gas sensor and evaluate its sensitivity [11] Chen D, Wang J, Li D, Zhang L and Wang X 2010 Appl. Phys.
A 100 239
characterization. The proposed sensor can yield a sensitive, [12] Sharma G, Liljeholm L, Enlund J, Bjurström J, Katardjiev I
reversible and reproducible response to DMMP gas in the and Hjort K 2010 Sensors Actuators A 159 111
range of 10–50 ppm. It was found that the frequency shifts [13] Penza M, Aversa P, Cassano G, Suriano D, Wlodarski W,
versus concentrations of the analyte exhibit a perfect linear Benetti M, Cannatà D, Pietrantonio F D and Verona E 2008
IEEE Trans. Electron Devices 55 1237
correlation. The responses are relatively fast and reversible.
[14] Zhang H, Pang W, Kim E S and Yu H 2010 J. Micromech.
The FBAR sensor has a very high gas sensitivity which is Microeng. 20 125008
several orders of magnitude higher than that of QCM. On [15] Nirschl M, Blleer A, Erler C, Katzschner B, Vikholm-Lundin
the basis of these results, it is expected that the PVDF- I, Auer S, Vertwl J, Pompe W, Schreiter M and Mertig M
coated FBAR could be further developed as a handheld sensor 2009 Sensors Actuators A 156 180
[16] Xu W, Choi S and Chae J 2010 Appl. Phys. Lett. 96 053703
for analyzing nerve gas. In the future, we will experiment
[17] Xie H, Yang Q D, Sun X, Yu T, Huang J Z and Huang Y 2005
with various sensitive coatings in order to improve device Sensor. Mater. 17 21
performance. [18] Khan M A K, Long Y T, Schatte G and Kraatz H B 2007 Anal.
Chem. 79 2877
[19] Yang K L, Cadwell K and Abbott N L 2005 Sensors Actuators
Acknowledgment B 104 50
[20] Ying Z, Jiang Y, Du X, Xie G, Yu J and Tai H 2008 Eur.
This work was partially supported by the Research Project of Polym. J. 44 1157
SDUST Spring Bud under grant no 2010AZZ071. [21] Maji S, Asrey R, Kumar S, Saxena C, Kumar N, Vyas K D
and Banerjee S 2010 J. Appl. Polym. Sci. 116 3708
[22] Du X, Wang Z, Huang J, Tao S, Tang X and Jiang Y 2009
References J. Mater. Sci. 44 5872
[23] Ying Z, Jiang Y, Du X, Xie G and Yang Y 2007
[1] Arduini F, Amine A, Moscone D, Ricci F and Palleschi G J. Appl. Polym. Sci. 106 1024
2007 Anal. Bioanal. Chem. 388 1049 [24] Chen D, Xu Y, Wang J, Zhang L, Wang X and Liang M 2010
[2] Tomchenko A A, Harmer G P and Marquis B T 2005 Sensors. Vacuum 85 302
Actuators B 108 41 [25] Chen D, Xu Y, Wang J and Zhang L 2010 Sensors Actuators B
[3] Chen P C, Ishikawa F N, Chang H K, Ryu K and Zhou C 2009 150 483
Nanotechnology 20 125503 [26] Levitsky I, Krivoshlykov S G and Grate J W 2001 Anal. Chem.
[4] Yang Y, Ji H F and Thundat T 2003 J. Am. Chem. Soc. 125 1124 73 3441
[5] Zuo G, Li X, Li P, Yang T, Wang Y, Cheng Z and Feng S 2006 [27] Chen N and Hong L 2001 Polymer 43 1429
Anal. Chim. Acta 580 123 [28] Grate J W 2008 Chem. Rev. 108 726
[6] Nimal A T et al 2009 Sensors Actuators B 135 399 [29] Grate J W and Abraham M H 1991 Sensors Actuators B 3 85