Electron transport mechanism of thermally oxidized ZnO gas sensors

Article abstract of DOI:10.1016/j.physb.2010.08.027

ZnO gas sensor was fabricated by thermal oxidation of metallic Zn at different time periods. The sensors were characterized by IV measurement with DC voltage, ranging from -2 to 2 volts, in both normal air and H₂ gas with concentration from 40

Full text of DOI:10.1016/j.physb.2010.08.027

Physica B 405 (2010) 4509–4512
Contents lists available at ScienceDirect
Physica B
journal homepage: www.elsevier.com/locate/physb
Electron transport mechanism of thermally oxidized ZnO gas sensors
n
N.H. Al-Hardan , M.J. Abdullah, A. Abdul Aziz
School of Physics, Universiti Sains Malaysia (USM), 11800 Minden, Pulau Pinang, Malaysia
a r t i c l e i n f o
a b s t r a c t
Article history:
ZnO gas sensor was fabricated by thermal oxidation of metallic Zn at different time periods. The sensors
Received 11 June 2010
Received in revised form
were characterized by I–V measurement with DC voltage, ranging from ꢀ2 to 2 volts, in both normal air
2
and H gas with concentration from 40 to 160 ppm. The transport mechanism of the carriers was found
2
7 July 2010
to be due to thermionic process through both the grain boundaries and the metal–semiconductor
junctions. Resistance of the ZnO sensing film is independent of applied voltage in the range
Accepted 6 August 2010
0
.5 VoV
a
o2 V; however, it is dependent on gas concentration, which makes it useful for gas sensing
Keywords:
application.
Electrical conductivity
Gas sensors
&
2010 Elsevier B.V. All rights reserved.
Oxidation process
ZnO
1
. Introduction
Metal oxide devices change their resistivity in the presence of
2
H based on the thermo-emission theory. The study focuses the
current–voltage and resistance–voltage characteristics of the gas
sensor at voltage range ꢀ2 to 2 V.
reducible or oxidation gases. They have been used since 1971,
when Taguchi [1] introduced to the market its metal oxide gas
2. Experimental method
2
sensor based on tin dioxide (SnO ) thick films. Many published
works tried to explain the mechanism involved in the gas
detection process. Most of these works depend on measuring
the change of resistance as a function of the gas concentration
An n-type Si (1 0 0) wafer was used as a substrate. The wafers
were initially cleaned with RCA standard method and thermally
oxidized at 1100 1C for 4 h in oxygen flow to form 1.2 m of silicon
dioxide (SiO ). Photolithography was employed to pattern the Ti/Pt
m
(sensitivity) and type of gas (selectivity). The method essentially
detects the sum of different effects that occur on the surface of the
sensor that is attributed to the grain, grain boundary, and metal–
semiconductor contacts [2].
2
heat element and the conducting electrodes. Ti and Pt were coated
using the Edwards A306 DC magnetron sputtering unit, followed
by wet etching process to obtain the final device. (Fig. 1 shows the
device diagram.) Thin films of Zn metal were coated with a good
uniformity over the Ti/Pt electrodes using high-purity (99.99%) Zn
target by the same unit. The ultimate pressure of the unit was
Various metal oxides have been investigated for gas sensing
applications. These include titanium oxides (TiO
pentoxides (V ) [4], SnO [5], tungsten oxides (WO
zinc oxide (ZnO) [7]. It appears that SnO and ZnO have received
2
) [3], vanadium
2
O
5
2
3
) [6], and
2
ꢀ
5
ꢀ3
1
ꢁ 10 Torr and was raised to 4 ꢁ 10 Torr during the deposi-
more attention because of their nontoxic, relatively good, and
stable electrical properties.
tion process by flowing high-purity Ar (99.999%).
Thermal oxidation of the Zn films was carried out in a
horizontal controlled tube furnace; the samples were introduced
into the furnace at room temperature. The temperature was then
ZnO thin film has been extensively studied because of its
diverse applications in different modern technological fields, such
as gas sensors [7], conductive transparent layers for solar cell and
displays application, and surface acoustic wave sensors. They can
be prepared by different methods, such as DC and RF sputtering,
sol–gel and thermal pyrolysis [8], and oxidation of Zn precursor,
such as ZnS and Zn [9,10].
ꢀ
1
raised at the rate of 5 1C min
99.99%) atmosphere for 30 and 60 min. The thickness of the ZnO
thin film was 0.3 m as measured by the optical method using
to 400 1C in high-purity O
2
(
m
Filmetrics model F20 (Filmetrics Inc., San Diego, CA, USA).
The I–V characteristic of the produced device was measured
using a programmable electrometer—model 617 KEITHLEY. The
This study investigates the electron transport mechanism of
the ZnO gas sensor in the presence of different concentrations of
applied voltage V
a
(ꢀ2 to 2) V from the electrometer was supplied
to the measuring electrodes and the current was measured
through the same unit. DC power supply (Lodestar 30V/3A) was
used to supply voltage to the heating elements.
n
Corresponding author. Tel.: +60 17 4961778; fax: +60 4 6579150.
E-mail address: naif.zd06@student.usm.my (N.H. Al-Hardan).
0921-4526/$ - see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.physb.2010.08.027

4
510
N.H. Al-Hardan et al. / Physica B 405 (2010) 4509–4512
mean square (rms) of surface roughness of the prepared films
showed no significant difference in the surface roughness
between two samples (30 and 60 min oxidation), which was
about 4572 nm. XRD of the oxidized films is shown in Fig. 3,
from which it can be observed that the films were polycrystalline.
There is no evidence of Zn metal in the oxidized films, which
confirmed the complete conversion of Zn metal to ZnO during
oxidation.
The observed peaks were at the Bragg angles of 31.91, 34.51,
36.361, 47.721, and 56.741, which represent the (1 0 0), (0 0 2),
(1 0 1), (1 0 2), and (1 1 0) phases of ZnO (JCPDF card no. 36-1451),
respectively. There was no observable difference in the XRD
spectrum of the two ZnO films prepared at different oxidation
times, which confirmed the results obtained by Alivov et al. [15],
who prepared ZnO through the same method.
3
.2. Electrical properties
Fig. 1. Schematic layout structure of ZnO gas sensor; the sensing element’s
dimension is 2 mm ꢁ 2 mm.
The current–voltage characteristic of the ZnO thin film is
shown in Fig. 4. It is obvious that current is enhanced with the H
2
concentration, but better resolution of the measured current was
obtained for the 30 min sample (Fig. 4(a)). For samples prepared
for 60 min oxidation (Fig. 4(b)), the current change shows some
2
evidence of saturation behavior at high H concentrations, which
suggests that the donor concentration begins to be limited, the
same behavior that was also observed by Kim et al. [16]. It was
suggested earlier that samples with longer oxidation time show
less oxygen vacancies [11,17].
It was also found that the measured current of the sample
prepared at 60 min was higher than that prepared at 30 min. It
was suggested earlier [11,18] that sensitivity of the ZnO gas
sensor can be enhanced by reducing carrier concentration (higher
resistance); the increase of oxidation time from 30 to 60 min
results in a decrease of oxygen vacancies and zinc interstitials in
the samples, which led to a decrease of carrier concentrations,
wider depletion layer, and higher potential energy between the
grains.
Fig. 2. AFM image of oxidized Zn for (a) 30 min and (b) 60 min.
Temperature of the heating element was controlled by varying
the voltage and measured by a calibrated K-type thermocouple,
which was mounted on the device to sense temperature of the
sensor. The temperature of the sensor was maintained at 400 1C
11]. All the measurements were taken at room temperature
23 1C) and 62% humidity.
[
(
The increase in applied voltage (V
barrier height at the grain boundaries thus it enhances the output
current. In addition, the increase in H concentration causes a
a
) results in the lowering of
The test chamber was made of stainless steel, with a total
3
2
volume of about 8200 cm . A high grade gas mixture of 2%
hydrogen and nitrogen (as buffer gas) was used for testing the
reduction in barrier height between the grain boundaries and
enhances the current [19] as more electrons are released.
The oxygen-vacancy model may be used to explain the
reaction mechanisms in this case [11,20], in which the oxygen
vacancies act as electron donors. The increase of oxidation time
2
device. H gas concentrations inside the gas testing chamber were
controlled by changing flow rate of the gas mixture, which was fed
to the testing chamber [11–13]. Flow rate of the gas was controlled
and measured through a mass flow controller (Cole-Parmer model
3
2708-26). To ensure a homogenous distribution of the gas under
test in the chamber, the measurements were commenced 3 min
after turning on the gas flow. After testing each gas concentration,
the chamber was flushed using a commercial air pump, such that
the air flow was passed through a glass vessel containing silica
grains to ensure that the air supplied was dried.
X-ray diffraction (XRD) and atomic force microscopy (AFM)
were used to determine the structure and morphology of the film
sensor, respectively.
3
. Results and discussion
3.1. Structure and morphology
The surface morphology of the ZnO film as observed from the
AFM (Fig. 2) micrographs proved that the grains are uniformly
distributed within the scanning area, with individual columnar
grains extending upwards. This surface characteristic is important
for applications such as gas sensors and catalysts [14]. The root
Fig. 3. XRD of the prepared ZnO films oxidized at 400 1C for 30 and 60 min in
oxygen flow.

N.H. Al-Hardan et al. / Physica B 405 (2010) 4509–4512
4511
Fig. 4. Forward current–voltage characteristic of ZnO gas sensor oxidized at (a) 30 min and (b) 60 min for different H
2
concentrations, at the operating temperature of 400 1C.
from 30 to 60 min resulted in a decrease of the oxygen vacancies
in the sample [17]. The reaction of hydrogen molecules with the
oxygen can be expressed as follows: the hydrogen first reacted
with the oxygen from the ZnO surface and as a consequence
water vapor molecules and oxygen vacancies are produced.
Consequently the oxygen vacancies became ionized due to the
charge imbalance of the oxygen site, thereby electrons are
introduced into the conduction band that consequently increased
the measured current of the sensor. Thus for the sample with
higher oxygen concentrations (lower oxygen vacancies) following
6
0 min oxidation, more reaction with hydrogen gas took place,
thereby resulting in enhancement of the current response of the
sensor. These reactions could have occurred as in the following
steps according to Kr o¨ ger–Vink notation [20]:
X
O
X
Fig. 5. Theoretical R–V characteristics associated with thermionic emission and
H
X
2
ðgasÞþO 3H
2
OðgasÞþVO
tunneling [6].
ꢂ
ꢀ
V 3V þe ðCBÞ
O
O
ꢂ
ꢂꢂ
ꢀ
constant for both samples, which means that the I–V curves are
V 3V þe ðCBÞ
O
O
not linear. For voltage 0.5 VoV
a
o2 V, ꢀ2 VoV o ꢀ0.5 V, the
a
X
where O is the neutral oxygen from the surface, V
vacancy, VO neutral oxygen vacancy, VO singly ionized oxygen
vacancy, VO doubly ionized oxygen vacancy, and e ðCBÞis the
electrons at the conduction band.
O
the oxygen
O
resistance of each sample is almost constant (independent on
X
ꢂ
voltage). In this region, H concentration affects resistance of the
2
ꢂꢂ
ꢀ
sensing element; the higher the concentration, the lower the
resistance. This dependence is very useful in gas sensing
application and widely used for the determination of gas
concentration in the atmosphere.
The linear dependence of resistance on gas concentration for
the samples with 30 min of thermal oxidation is better than that
The electron transport mechanism can be modeled based on
the work of Pitcher et al. [6], which showed that the mechanism
depends on carrier concentration of the semiconductor.
At low carrier concentration, width of the potential barrier is
large and only electrons of sufficient kinetic energy are able to
move over the barrier. At high carrier concentration, electrons will
have the ability to tunnel through the narrow potential barrier.
The first mechanism is known as thermionic emission, which is
dominant for the Schottky contacts, while the second one is
ascribed to electron tunneling [21].
of the 60 min samples, which can be applied for sensing H
concentration in the present studied range. For samples with
0 min of oxidation, good resistance resolution appears only in
2
6
the concentration range less than 80 ppm. The near saturation in
resistance magnitudes (above 80 ppm) for this sample suggests
that the donor concentration begins to be limited [16].
Due to the increase of carrier concentration as a result
2
of applied voltage and (or) H concentration, the tunneling current
becomes more significant and superimposes on pure thermionic
emission current, which lowers the Schottky barrier height.
The theoretical resistance–voltage (R–V) characteristic asso-
ciated with thermionic emission and electron tunneling mechan-
ism is shown in Fig. 5 [6].
4. Conclusion
The electron transport of ZnO gas sensor prepared by thermal
oxidation of Zn metal is attributed to the thermionic emission
Fig. 6 shows the R–V characteristic obtained for the ZnO gas
sensor, from which it can be inferred that electron transport in the
gas sensor resembles the thermionic process. At low applied
2
mechanism. Resistance of the sensor decreased with H concen-
tration in the range 40–160 ppm. The sensors prepared for 30 min
oxidation show better sensing characteristics as compared with
that of the sensors prepared for 60 min oxidation.
voltage ꢀ0.5 VoV o0.5 V, the resistance of the film is not
a

4
512
N.H. Al-Hardan et al. / Physica B 405 (2010) 4509–4512
Fig. 6. R–V characteristic of the ZnO gas sensor for (a) 30 min and (b) 60 min samples.
Acknowledgment
[9] S. Cho, J. Ma, Y. Kim, Y. Sun, G.K.L. Wong, J.B. Ketterson, Appl. Phys. Lett. 75
1999) 2761.
(
[
10] X.T. Zhang, Y.C. Liu, Z.Z. Zhi, J.Y. Zhang, Y.M. Lu, W. Xu, D.Z Shen, G.Z. Zhong,
This work has been supported by Universiti Sains Malaysia
X.W. Fan, X.G. Kong, J. Cryst. Growth 240 (2002) 463.
(
Penang, Malaysia) through a short term grant. N.H. Al-Hardan is
[11] N. Al-Hardan, M.J. Abdullah, A. Abdul Aziz, Appl. Surf. Sci. 255 (2009) 7794.
[
[
[
[
12] Gas Testing Box Model no. SR3, Instruction Manual, Figaro Sensor Company,
www.figarosensor.comS.
13] N. Al-Hardan, M.J. Abdullah and A. Abdul Aziz, Adv. Appl. Ceram.,
DoI:10.1179/174367510X12722693956275, in press.
14] Z Dachun, Q Zhongkai, P Xiaoren, D Muji, S. Minggen, J. Vac. Sci. Technol. B 15
grateful for the fellowship from the Institute of Graduate Studies,
Universiti Sains Malaysia.
/
References
(1997) 805.
15] Y.I. Alivov, A.V. Chernykh, M.V. Chukichev, R.Y. Koroptkov, Thin Solid Films
473 (2005) 241–246.
[
[
1] N. Taguchi, US Patent 3,625,756, December 1971.
2] P.T. Mosely, B.C. Tofield, in: Solid State Gas Sensor, Adam Hilger, Bristol and
Philadelphia, 1987.
[16] S. Kim, B.S. Kang, F. Ren, K. Ip, Y.W. Heo, D.P. Norton, S.J. Pearton, Appl. Phys.
Lett. 84 (2004) 1698.
[
[
[
[
[
[
3] Y. Hayakawa, K. Iwamoto, Kikuta, S. Hirano, Sensors Actuators B 62 (2000) 55.
4] O. Schilling, K. Colbow., Sensors Actuators B 21 (1994) 151.
5] K.H. Cha, H.C. Park, K.H. Kim, Sensors Actuators B 21 (1994) 91.
6] S. Pitcher, J.A. Thiele, H. Ren, J.F. Vetelino, Sensors Actuators B 93 (2003) 454.
7] P.P. Sahay, J. Mater. Sci. 40 (2005) 4383.
8] U¨ . O¨ zg u¨ r, Y.I. Alivov, C. Liu, A. Teke, M.A. Reshchikov, S. Do g˘ an, V. Avrutin,
[17] V. Khranovskyy, J. Eriksson, A. Lloyd-Spetz, R. Yakimova, L. Hultman, Thin
Solid Films 517 (2009) 2073.
[18] J.B.K. Law, J.T.L. Thong, Nanotechnology 19 (2008) 205502.
[19] N. Barsan, U. Weimar, J. Electroceram. 7 (2001) 143.
[20] A. Gurlo, R. Riedel, Angew. Chem. Int. Ed. 46 (2007) 3826.
[21] E.H. Rhoderick, R.H. Williams, in: Metal–semiconductor contacts, Clarendon,
Oxford, 1988.
S.J. Cho, H. Morko c- , J. Appl. Phys. 98 (2005) 041301.