Sensors and Actuators B 143 (2009) 325–332
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical
journal homepage: www.elsevier.com/locate/snb
Large-scale synthesis and gas sensing application of vertically aligned and
double-sided tungsten oxide nanorod arrays
Xiaoping Shen
a,b
, Guoxiu Wang
a,∗
, David Wexler
a
a
Institute for Superconducting and Electronics Materials, School of Mechanical, Materials and Mechatronics Engineering,
University of Wollongong, Wollongong, New South Wales 2522, Australia
b
School of Chemistry and Chemical Engineering, Jiangsu University, Zhenjiang 212000, China
article info
Article history:
Received 25 June 2009
Received in revised form 17 August 2009
Accepted 6 September 2009
Available online 17 September 2009
Keywords:
Tungsten oxide
Nanorod arrays
Hydrothermal synthesis
Gas sensing
Sensor
abstract
Large-scale vertically aligned and double-sided Co-doped hexagonal tungsten oxide nanorod arrays have

have been developed to produce nanorods in the form of large-
area arrays [9–11]. However, although the template-based method
can provide good control over the uniformity and dimensions
of nanorods, removal of the template through a post-synthesis
process may cause damage to the nanorod arrays. In addition,
most nanorods synthesized using the template-based method are
polycrystalline instructure, whichmay limit their use in device fab-
rication and fundamental studies. The method involving the use of
a patterned catalyst is able to generate nanorods with controllable
∗
Corresponding author. Fax: +61 2 42215731.
E-mail address: [email protected] (G. Wang).
sizes and highly crystalline structures. However, the catalyst may
cause contamination of the resultant nanorod arrays,which is often
a great disadvantage to their application. As a template-free and
catalyst-free method, epitaxial growth of free-standing nanorod
arrays has recently been attained, in which a substrate with an
excellent lattice match to the overlying materials is vital to guide
the assembly of one-dimensional (1D) nanoarrays [12,13].Asa
result, this method has largely been limited to particular materials,
notably zinc oxide. Therefore, the development of novel and more
effective strategies for preparing highly ordered nanorod arrays is
of great significance for their practical applications in nanotechnol-
ogy.
Tungsten oxide (WO
3
), as an important n-type semiconductor,
has received wide attention owing to its promising application in
gas sensors, heterogeneous catalysts, chromogenic devices, solar-
energy devices, and field electron emission [14–16]. The syntheses

organic solvents. We found that the as-synthesised Co-doped WO
3
nanorod arrays have a fast, highly sensitive, and fairly selective
response to 1-butanol gas.
2. Experimental details
All chemicals are ACS reagent and were used directly as
purchased from Sigma–Aldrich. In a typical experiment for synthe-
sizing Co-doped hexagonal tungsten oxide double-sided nanorod
arrays, 0.815 g of Na
2
WO
4
·2H
2
O was dissolved in 10mL of distilled
water. The solution was acidified to a pH range of 1–1.2 using HCl
solution (3mol L
−1
). Then, 0.63 g of H
2
C
2
O
4
was added to the mix-
ture and the mixture was diluted to 25 mL. After that, a stable
WO
3
sol was formed. 16 mL of the WO
3

on a carbon-coated copper grid after strong ultrasonic dispersion
in absolute ethanol. X-ray photoelectron spectroscopy (XPS) mea-
surements were carried out with an ESCALab220i-XL spectrometer
by using a twin-anode Al K␣ (1486.6 eV) X-ray source. All the
spectra were calibrated according to the binding energy of the
adventitious C1s peak at 284.8eV. The band gap of the Co-doped
WO
3
was determined by UV–vis spectroscopy (Shimadzu 1700).
Raman spectroscopy (HR 800) was performedat room temperature
with an excitation wavelength of 632.8 nm.
The gas sensing properties were measured using a WS-30A gas
sensor measurement system. The gas sensor was fabricated as fol-
lows: the Co-doped WO
3
sample was mixed with polyvinyl acetate
(PVA) binder (1 wt.%) to form a slurry, and then pasted onto a
ceramic tube (2 mm in diameter) by a doctor blade to form a thin
film between two Au electrodes,which had been previously printed
on the ceramic tube and were connected with four platinum wires.
As a comparison, gas sensing properties of commercial WO
3
pow-
der (Fluka) were also measured. The commercial WO
3
powder is a
well crystallinepowder witha crystalsize inthe rangeof afew hun-
dreds nanometers. Fig. S-1 shows the FE-SEM image of commercial
WO
3

.
3. Results and discussion
The phase of the obtained products was determined by XRD.
As shown in Fig. 1, all the diffraction peaks of the annealed product
can be readilyindexed to hexagonalstructure WO
3
with lattice con-
stants a = 7.3223 Å and c =7.6574 Å, which are slightly smaller than
the standard values for bulk hexagonal WO
3
(JCPDS No. 85-2460,
a =7.3242 Å and c = 7.6624Å). Before annealing, the sample showed
several impurity peaks, suggesting that the annealing is necessary
to obtain pure phase WO
3
. In addition, the relative strength of the
(0 02) peak was significantly increased by the annealing treatment,
suggesting that annealing also improves the crystallinity of the
product. It is well known that hexagonal (h) WO
3
is a metastable
phase and can transform into monoclinic (m) WO
3
at high tem-
perature. Recently, Szilágyi et al. pointed out that the structure
of hexagonal WO
3
cannot be maintained without some stabilizing
ions or molecules in the hexagonal channels, and thus the exis-
tence ofstrictly stoichiometric hexagonal WO

4
+
ions can stabilize the hexagonal structure in such a
way that they are located in the hexagonal channels of crystallites
and block the thermodynamically favored hexagonal–monoclinic
transformation.
As shown in Fig. 2, the O1s peak is located at 530.7 eV, which
is ascribed to the W–O peak. The W4f peaks located at 36.4eV and
38.4 eV can be attributed to W4f
7/2
and W4f
5/2
, respectively, which
are in good agreement with the reported values [23]. These two
peaks are well separated without any shoulder, which indicates
that almost all W atoms are in the +6 oxidization state [24]. The
Fig. 1. X-ray diffraction patterns of Co-doped WO
3
nanorod arrays: (a) before
annealing and (b) after annealing.
X. Shen et al. / Sensors and Actuators B 143 (2009) 325–332 327
Fig. 2. XPS spectra of Co-doped WO
3
nanorod arrays.
Na1s peak at 1070.9 eV is consistent with the +1 oxidation state
of sodium [25]. A Co2p signal located at 780.3 eV was detected,
showing that Co exists in the +2 oxidation state in the product [26].
However, the Co2p signal is weak due to the low Co content in
the product. Therefore, the above results confirmed the successful
preparation of the Co-doped hexagonal tungsten oxide.

A further investigation of the Co-doped WO
3
nanorod arrays
was conductedby TEM and HRTEM analysis.The low-magnification
TEM image (Fig. 4(a)) shows some straight nanorod bundles with
a diameter of about 200 nm and a length up to about 5 ␮m. A
high-magnification TEM image (Fig. 4(b)) reveals that the nanorod
bundles consist of smaller nanorods with a diameter of about
20 nm. These results are consistent with the FE-SEM observations.
Selected area electron diffraction (SAED) pattern (inset in Fig. 4(b))
taken from this nanorod bundle shows regular diffraction spots,
which can be indexed as hexagonal WO
3
single crystal recorded
along the [1 1 0]zone axis anddemonstrates that theWO
3
nanorods
grow along the [0 01] direction. The SAED pattern also reveals
that the nanorods in the bundles consist of a perfectly oriented
assembly, which is further confirmed by the HRTEM analysis. We
have performed extensive TEM observationson different individual
bundles and found that all of them consist of small nanorods and
grow along [0 01] crystal direction (Fig. S-4, supplementary data).
As shown in Fig. 4(c) and (d), the parallel lattice fringes among
the different primary nanorods and grain boundaries clearly show
the oriented aggregation and high crystallinity of the primary
nanorods. The degree of fusion of the primary nanorods in the
bundles shows a gradual increase from the top to the base of the
bundles. Fig. 4(e) shows a HRTEM image of the middle section of
a primary nanorod, from which the (1 0 0) lattice planes with a d-

3
nanorod arrays, with the inset showing a high-resolution view of a single nanorod bundle.
partly fuse to form bunch-like structures [27–29]. As observed
by TEM analysis, the bunch-like structures become more smooth
and regular from the tip to the base of the bundles. We believe
that Ostwald-ripening works simultaneously with the oriented
attachment to remedy the defects, leading to a smooth and regu-
lar surface of the base parts. Thus, urchin-like architectures with
WO
3
nanorod bundles on the surfaces of the microspheres are
formed. This is supported by the evidence that several hemispher-
ical regions (Fig. 5(a)) and sphere-like cores with nanorod bunches
on their surfaces (Fig. 5(b)) have been found in the Co-doped WO
3
products. This process is similar to what happens in the synthesis
of WO
3
urchin-like microspheres. Finally, the as-formed urchin-
like microspheres further self-assemble and fuse into well-aligned
double-sided nanorod arrays with the help of dopant Co
2+
and an
accompanying Ostwald-ripening process. To the best of our knowl-
edge, thisis the first time that the observation of large-area uniform
double-sided nanorodarrays formedby self-assemblyand fusion of
urchin-like nanostructures has been reported. Scheme 1 shows the
schematics of the formation process of the Co-doped WO
3
double-

stretching modes, while the bands at 270cm
−1
and 327 cm
−1
cor-
respond to the O–W–O bending modes of the bridging oxygen. The
weak Raman peak at 927 cm
−1
may be attributed to a stretching
mode of the terminal W
O. Although this latter band is charac-
teristic of tungsten oxide hydrates, it can appear in WO
3
via the
adsorption of water molecules [30]. The Raman spectrum provides
clear evidence for the highstructural qualityand phase-pure nature
of Co-doped WO
3
nanorod arrays. Theoptical absorptionproperties
of the as-prepared Co-doped WO
3
nanostructures were investi-
gated at room temperature by UV–visible spectroscopy. As shown
in Fig. 6(b), the spectrum shows one absorption peak at about
281 nm. WO
3
is an n-type semiconductor [31], and its optical band
gap can be estimated using the following formula:
(˛h)
n

. Theband edgeposition for amor-
phous WO
3
in contact with an aqueous electrolyte at a pH of ∼1is
about 3.2 eV [32]. Two distinct direct interband transition energies
of 3.52eV and 3.74 eV for WO
3
were also observed by Koffyberg
et al. [33]. The lower band gap of the Co-doped WO
3
may reflect
doping effects. It is well known that in doped compound semi-
conductors, in contrast with un-doped ones, the impurity states
play a special role in the electronic energy structures and transi-
tion probabilities. In addition, it is found that the best fit of Eq. (1)
to the absorption spectrum of the product gives n = 2, which sug-
gests that the as-obtained Co-doped WO
3
is semiconducting with
direct transitions at these energies.
Chemical sensors play an important role in the areas of emis-
sions control, environmental protection, public safety, and human
health. Much more public concern over serious environmental
issues is further promoting the development of sensors with both
high sensitivity and rapid response. It has been well documented
that the ultra-high surface-to-volume ratios of nanostructured
materials make their electrical conductivities extremely sensitive
to surface-adsorbed species and make them excellent candidates
for gas sensing applications [34,35]. The gas sensing performance
X. Shen et al. / Sensors and Actuators B 143 (2009) 325–332 329

3
nanorod
arrays improved dramatically with increasing concentration of the
Fig. 5. (a) FE-SEM image shows a hemispherical region (enclosed by the circle) in a nanorod array. (b) Cross-sectional FE-SEM image showing the sphere-like cores (as
indicated by the circles) in the Co-doped WO
3
double-sides nanorod arrays.
330 X. Shen et al. / Sensors and Actuators B 143 (2009) 325–332
Scheme 1. Schematic diagram of the proposed growth mechanism of the Co-doped WO
3
double-sided nanorod arrays.
test gas and was much higher than that of the commercial pow-
der. This means that the Co-doped WO
3
nanorod arrays are much
more sensitive to 1-butanol than the commercial powder. After
many cycles between the test gas and fresh air, the resistance of
the sensor was still able to recover its initial state, which indi-
cates that the sensor has an excellent reversibility. The response
time and recovery time (defined as the time required to reach 90%
Fig. 6. (a) Raman spectrum of the Co-doped WO
3
double-sided nanorod arrays. (b)
UV–vis spectrum of the Co-doped WO
3
double-sided nanorod arrays. The inset is
(˛h)
2
vs. h curve showing the band gap energies.
of the final equilibrium value) of the sensor were only 1–2 s and

gas
value was 8.5 at the very low concentration of 5 ppm,
but reached 232 at 1000 ppm. The sensing responses to 2-propanol,
ethanol and gasoline are 71, 50, and 37, respectively, at the concen-
tration of 1000 ppm. As shown in Fig. 5(b), the sensing responses
of the Co-doped WO
3
nanorod arrays towards acetone, toluene,
acetic acid, and heptane are 122, 95, 66, and 31, respectively. These
results indicate that the Co-doped WO
3
nanostructure-based sen-
sor is highly sensitive to these organic gases. It should be noted that
the relativelylow operation temperature helps to decrease thecon-
sumption of energy and can improve the suitability of the sensor
in some particular situations. It is well known that WO
3
,asann-
type semiconductor, is a good candidate for detecting the inorganic
gases O
3
,NO
x
, and H
2
S [36], but is less sensitive to hydrocarbons.
Nevertheless, our investigations illustrate that doping, and control
of the morphology and size of the nanorod arrays has endowed
WO
3

Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.snb.2009.09.015.
References
[1] F.X. Redl, K.S. Cho, C.B. Murray, S. O’Brien, Three-dimensional binary superlat-
tices of magnetic nanocrystals and semiconductor quantum dots, Nature 423
(2003) 968–971.
[2] S. Park, J.H. lim, S.W. Chung, C.A. Mirkin, Self-assembly of mesoscopic
metal–polymer amphiphiles, Science 303 (2004) 348–351.
[3] X.D. Liu, B. Lu, T. Iimori, K. Nakatsuji, F. Komori, Self-assembled MnN super-
structure, Phys. Rev. Lett. 98 (2007) 066103.
[4] X.F. Duan, C.M. Lieber, General synthesis of compound semiconductor
nanowires, Adv. Mater. 12 (2000) 298–302.
[5] J.F. Wang, M.S. Gudiksen, X.F. Duan, Y. Cui, C.M. Lieber, Highly polarized pho-
toluminescence and photodetection from single indium phosphide nanowires,
Science 293 (2001) 1455–1457.
[6] X.Y. Kong, Z.L. Wang, Spontaneous polarization-induced nanohelixes,
nanosprings and nanorings of piezoelectric nanobelts, Nano Lett. 3 (2003)
1625–1631.
[7] X.Y. Kong, Y. Ding, R. Yang, Z.L. Wang, Single-crystal nanorings formed by epi-
taxial self-coiling of polar nanobelts, Science 303 (2004) 1348–1351.
[8] Y.N. Xia, P.D. Yang, Y.G. Sun, Y.Y. Wu, B. Mayers, B. Gates, Y.D. Yin, F. Kim, Y.Q.
Yan, One-dimensional nanostructures: synthesis, characterization, and appli-
cations, Adv. Mater. 15 (2003) 353–389.
[9] C.R. Martin, Membrane-based synthesis of nanomaterials, Chem. Mater. 8
(1996) 1739–1746.
[10] L.M. Dai, A. Patil, X.Y. Gong, Z.X. Guo, L.Q.liu, Y. liu, D.B. Zhu, Aligned nanotubes,
Chemphyschem 4 (2003) 1150–1169.
[11] K.P. Musselman, G.J. Mulholland, A.P. Robinson, L. Schmidt-Mende, J.L.
MacManus-Driscoll, Low-temperature synthesis of large-area, free-standing

[20] A. Shengelaya, S. Reich, Y. Tsabba, K.A. Muller, Electron spin resonance and
magnetic susceptibility suggest superconductivity in Na doped WO
3
samples,
Eur. Phys. J. B 12 (1999) 13–15.
[21] L. Wang, A. Teleki, S.E. Pratsinis, P.I. Gouma, Ferroelectric WO
3
nanoparticles
for acetone selective detection, Chem. Mater. 20 (2008) 4794–4796.
[22] I.M. Szilágyi, J. Madarász, G. Pokol, P. Király, G. Tárkányi, S. Saukko, J. Mizsei,
A.L. Tóth, A. Szabó, Varga-Josepovits, Stability and controlled composition of
hexagonal WO
3
, Chem. Mater. 20 (2008) 4116–4125.
[23] H. Qi, C.Y. Wang, J. Liu, Simple method for the synthesis of highly oriented
potassium-doped tungsten oxide nanowires, Adv. Mater. 15 (2003) 411–414.
[24] Z.J. Gu, H.Q. Li, T.Y. Zhai, W.S.Yang, Y.Y.Xia, Y.Ma,J.N.Yao,Large-scale synthesis
of single-crystal hexagonal tungsten trioxide nanowires and electrochemical
lithium intercalation into the nanocrystals, J. Solid State Chem. 180 (2007)
98–105.
[25] Y. Miura, H. Kusano, T. Nanba, S. Matsumoto, X-ray photoelectron spectroscopy
of sodium borosilicate glasses, J. Non-Cryst. Solids 290 (2001) 1–14.
[26] C. Huang, X. Liu, L. Kong, W. Lan, Q. Su, Y. Wang, The structural and magnetic
properties of Co-doped titanate nanotubes synthesized under hydrothermal
conditions, Appl. Phys. A 87 (2007) 781–786.
[27] R.L. Penn, J.F. Banfield, Imperfect oriented attachment: dislocation generation
in defect-free nanocrystals, Science 281 (1998) 969–971.
[28] J.F. Banfield, S.A. Welch, H. Zhang, T.T. Ebert, R.L. Penn, Aggregation-based
crystal growth and microstructure development in natural iron oxyhydroxide
biomineralization products, Science 289 (2000) 751–754.

and molecule-based magnetic materials.
G.X. Wang received his PhD degree in Materials Science and Engineering in 2001
from University of Wollongong, Australia. He currently is working as an Associate
Professor at School of Mechanical, Materials and Mechatronic Engineering, Univer-
sity of Wollongong. His major research interests include nanostructured functional
materials, materials chemistry in energy storage and conversion, and development
of chemical and biological sensors.
D. Wexler received his PhD degree in Materials Science and Engineering in 1991
from Monash University, Australia. He currently is working as a senior research
fellow at School of Mechanical, Materials and Mechatronic Engineering, University
of Wollongong. His major research interests include nanomaterials synthesis and
TEM and HRTEM characterization of materials.