Synthesis of ZnO nanowires and impacts of their orientation and
defects on their gas sensing properties
Sergio Roso,a, c Frank Güell,b Paulina R. Martínez-Alanis,b Atsushi Urakawac and Eduard Llobeta
a

b

c

Minos-Emas, Universitat Rovira i Virgili, Av. Països Catalans 26, 43007, Tarragona, Spain.

Departament d’Electrònica, Universitat de Barcelona, c/Martí Franquès 1, 08028, Barcelona, Spain.

Institute of Chemical Research of Catalonia (ICIQ), The Barcelona Institute of Science and Technology, Av. Països Catalans 16,
43007, Tarragona, Spain.

ZnO nanowires (NWs) oriented in three different directions were synthesized via the vapour-liquid-solid
technique over c-, r- and a-plane sapphire substrates. Gas sensing properties of these samples against
reducing and oxidizing gases were examined. ZnO NWs grown on the c-plane substrate showed the
highest response to oxidizing gases (NO2), while those grown on the a-plane substrate were more
responsive to reducing one (ethanol). According to the insights gained by photoluminescence studies, the
distinct responses were clearly correlated with the quantity and type of defects present in the material, and
ultimately, relating the orientation of the different ZnO NWs sensors to their gas sensing characteristics,
showing a possible rational strategy to tailor oxide nanomaterials for gas sensing applications.

Introduction
Semiconductor metal oxides have become a major technological drive in the field of gas
sensing because of their low cost, high sensitivity, fast response and their relative
simplicity. Among others, nanostructured zinc oxide (ZnO)1 is an interesting material
for its unique electric and optoelectronic properties. Its wide direct bandgap (around
3.37 eV) at room temperature as well as its extremely large binding energy (around 60
meV), much higher than those of widely used metal oxides such as GaN (25 meV),
make it suitable not only for gas sensing2 but also for solar cells3, 4, lasers5 and
waveguides6. Moreover, it possesses high breakdown fields, large electron saturation
rates, high resistance and efficient luminescence as well7.
A variety of methodologies, such as thermal evaporation8, and hydrothermal methods9,
have been reported for the synthesis of ZnO nanostructured in one-dimension such as
nanowires (NWs). A recent report presented the use of randomly oriented ZnO NWs as
gas sensors for NO2, ethanol (EtOH), and H2 among others10. Also, EtOH sensing
properties of randomly oriented ZnO over silicon substrates have been investigated at
different temperatures and concentrations and the high response was attributed to the
high electron donating effect of EtOH compared to that of other gases like CO11.

Regarding the photoluminescence spectra (PL spectra) of ZnO, there are different
emission bands which are located in the ultraviolet and the visible (green, yellow and
red/orange bands) regions. The first one is typically attributed to the characteristic
emission of ZnO12, whereas the visible part is normally attributed to a variety of donor
defects such as oxygen vacancies, zinc interstitials and zinc vacancies13.
In this work, the relationship between the type and quantity of defects present in the
material and the gas sensing characteristics was investigated by means of
photoluminescence (PL). More precisely, the aim of this paper is to study the effects of
the orientation of the ZnO NWs grown over sapphire substrates and their detailed
nanostructures on the gas sensing properties.

Experimental
ZnO NWs were synthesized via the vapour-liquid-solid (VLS) process using a chemical
vapor deposition (CVD) furnace14. It consists of three main steps: 1) Creation of the
liquid alloy of the Au catalyst and the material to deposit; 2) nucleation of the material
at the liquid-solid interface; and 3) growth of the NWs. The nucleation and growth of
the solid ZnO NWs occur due to supersaturation of the liquid droplet. A schematic view
of the VLS process can be observed in Fig 1. Previously to the growth of the metal
oxide NWs, a 3 nm Au thin layer that will act as a catalyst was deposited by sputtering
onto the sapphire substrates. The precursor was a powder of ZnO and graphite (125
mg), and Ar was used as carrier gas. Both the precursor and the substrate were placed
on a quartz boat into the horizontal CVD furnace. The substrate was placed 1 cm away
from the precursor and samples were grown in different batches. Then, the temperature
was raised to 900°C and kept constant for 30 minutes. The optimum Ar gas flow was
400 ml/min. To synthesize ZnO NWs via VLS, it is essential to decompose the ZnO
powder, and to enable this decomposition to occur at 850°C, the ZnO powder is mixed
with graphite. All the process was carried out under atmospheric pressure. It goes
without saying that the relative humidity has no influence on the growth of the ZnO
NWs. Moreover, the temperature at which the growth takes place (900ºC) the relative
humidity present can be neglected.

Fig. 1: Schematic view of the VLS process. First, at high temperatures, the Au thin layer turns into
droplets. Then, Zn atoms condense and attach to the Au droplet. Finally, when the Au droplet becomes
supersaturated, the bottom-up growth of the ZnO NW takes place.

Three different types of sapphire (Al2O3) substrates were used, with different
crystallographic directions in order to grow ZnO NWs oriented along the c-plane
(0001), r-plane (1-102) and a-plane (11-20). The match between the lattice parameter of
the ZnO and that of the sapphire will allow us to epitaxially grow the ZnO NWs. The
structural characterization was carried out by a field-emission scanning electron
microscopy (FESEM, Hitachi H-4100F) and by X-ray diffraction (XRD, Bruker-AXS
D8-Discover diffractometer). For obtaining the PL spectra, a He-Cd laser at 325 nm was
used in order to overcome the band gap barrier of the ZnO NWs. The luminescence was
dispersed by an Oriel Instruments 74000 monochromator and detected with a
Hamamatsu H8259-02 with a socket assembly E717-500 photomultiplier. All sources of
noise were removed with a Stanford Research System SR830 DSP lock-in amplifier. In
order to measure the resistance of our sensors, silver (Heraeus, AD1688-06) parallel
electrodes were deposited on top of the active layer. Substrates were glued to a hotplate
in order to measure at different operating temperatures (i.e. sensors could be operated
from room temperature up to 250°C). Gas mixing and delivery were performed
employing a computer controlled mass-flow system.

Results and Discussion
FESEM images of the ZnO NWs grown over the three sapphire substrates (c-, r-, and aplanes) are presented in Fig. 2. In all cases, the diameter of the NWs was about 70 nm
and the length was about 1 µm. The dimension of the NWs was dependent on the
deposition time and also the thickness of the Au layer, being deposited prior to the
synthesis and working as a catalyst during the ZnO NW growth15. Hence, these two

parameters were kept the same for the three samples. As evident from Fig. 2, distinct
orientations of the ZnO NWs were observed for the different sapphire substrates used.

Fig. 2: SEM images of the ZnO NWs grown over a) c-, b) r-, and c) a-planes of sapphire

For the c-plane substrate (Fig. 2a), the NWs grew tilted at 51 and 129°, relative to the
substrate. For the r-plane (Fig. 2b), the NWs showed a conical shape; they did not
exhibit a well-defined orientation and many wire-to-wire junctions were present. This
could be explained by the generation of stress during the initial stage of growth, thus
joining together to make thicker structures along the growth15. Finally, for the a-plane
(Fig. 2c), the NWs grew vertically aligned. It is well known that NWs grow following
the exposed crystalline structure of the substrate.
Nucleation at the initial stage has a crucial role in in-plane alignment of the NWs. It is
known that ZnO nuclei grow with an epitaxial relationship with, in this case, c-, a- and
r-plane sapphire due to the lattice match between them16. The c- a- and r-planes of
sapphire and the c-plane of ZnO are related by a factor of 4, with a mismatch of less
than 0.08% at room temperature. This relationship leads to vertical epitaxial growth on
ZnO NWs on top of a-plane sapphire substrates17. The reduced lattice mismatch
between ZnO and sapphire is generally attributed to homogenous (isotropic)
orientations18. The visual inspection shows that the lattice matching is more significant
between ZnO NW and c- and a-plane sapphire substrates as evidenced by strong
preferential orientation of the NW growth directions.
XRD measurements were performed and the patterns showed ZnO NWs with a
hexagonal crystalline structure (Fig. 3) in accordance with ICDD card 00-036-1451.
The major peaks can be indexed to the (002) as well as the (101) among other crystal
planes. The intensities of the two peaks differ drastically, indicating strong preferential
orientation of ZnO NWs grown along specific crystal plane(s). Notably, the 101

reflection was fully diminished for NWs grown over the r-plane. This implies preferable
orientation in the <001> direction of NWs. Also, it should be noticed that the peak at
33º in the r-plane XRD pattern (see Fig.3) is originated from the substrate.

Fig. 3: XRD patterns of ZnO NWs grown on c-, r- and a-plane sapphire substrates.

Furthermore, these three ZnO NWs were subject to gas sensing tests using an oxidizing
gas (NO2) and reducing vapour (EtOH) by means of DC resistance measurements at
different operating temperatures (150-250°C). All sensors were exposed to 30 min of a
given concentration of a species, followed by a 60 min cleaning phase in dry air. All
sensors experienced an increase or decrease in resistance under exposure to oxidising or
reducing gases, respectively. This implies that ZnO behaves as an n-type
semiconductor. Fig. 4 shows representative response/recovery cycles for the three
sensors operated at 250°C against ethanol and nitrogen dioxide. The intensity of the
response increased with gas concentration. Responses were reproducible in all cases.
The baseline resistance could be fully recovered when cleaning in air after a sufficiently
long time. The response time for all sensors is about 5 minutes and, even though the
recovery time is longer, these times are still is in the same range than those reported for
nanowire gas sensors19,20. Although every sensor response does not reach complete
stabilization (r-plane and c-plane), it is observed that the a-plane sensor does reach such
stabilization. Taking this sensor as a reference the response and recover times have been
calculated as the time that the sensor takes to achieve 90% of the total value of the
response.
Fig. 5 shows the dependence of sensor response as a function of the operating
temperature at a fixed concentration of EtOH (500 ppm) and NO2 (100 ppm). All three
sensors showed the highest response when operated at 250ºC. Interestingly, for EtOH,
the highest response was achieved by the a-plane sensor, whereas for NO2 the highest
response was that of the c-plane sensor.

Mechanistically, at first oxygen is physisorbed on the surface of the ZnO NW. Then, it
becomes chemisorbed by attracting electrons of the conduction band, which forms a
depletion region on the surface of the ZnO NWs and consequently a potential barrier
between each nanowire, resulting in an increase in the resistance. It has been reported
that EtOH can undergo various reactions such as dehydration and dehydrogenation.
However, since ZnO is a basic oxide, dehydrogenation is favoured. In this case, the
ethanol sensing mechanism is controlled by the change (i.e. lowering) in the equilibrium
concentration of chemisorbed, negatively charged oxygen species, which results in a
change in the width of the depletion region at the surface of the NWs21.
In the case of NO2 detection, either the equilibrium concentration of chemisorbed
oxygen species is raised or NO2 is directly chemisorbed on the surface of the n-type
ZnO NWs 22. These oxidizing molecules extract free electrons from the conduction band
of ZnO and thus, the width of the depletion region is further extended, which increases
the resistance of NWs.

Fig. 4: Responses of the three ZnO NW sensors to different gases: a) EtOH, b) and c) NO2.

According to the results presented in Fig. 5, the c-plane sensor presented the best
response to NO2 while the a-plane sensor exhibited the best response to EtOH vapours.
In chemo-resistive sensors, there are many factors that influence sensor response. In
addition to the morphology of the active surface structure, the presence of defects and
their types can influence the electronic properties of materials and thus gas sensing
properties. As a matter of fact, it is accepted that the overall response of gas sensors is a
contribution both from the morphology of the material and the crystalline properties of
the material itself23. Due to our set-up, we cannot increase the temperature further than
250ºC. Most likely at higher temperatures (i.e. >300ºC) the response will decrease. The
reason behind such behaviour could be that the oxygen would not chemisorb on the
active sites anymore and thus, the response of the sensor will dramatically decrease24.

Fig. 5: Temperature-dependent responses of the three ZnO NW sensors to a) EtOH and b) NO 2. Response
was defined as Rair/REtOH and RNO2/Rair, respectively.

In order to verify this statement for these ZnO NW sensors, PL experiments were
performed at room temperature (Fig. 6). Two different zones in the spectra can be
clearly identified. The first one, which corresponds to the high peak located at the lower
wavelength (around 380 nm), is associated to the near-band edge emission of the
material. This strong emission corresponds to the recombination of electrons from the
minimum of the conduction band with holes of the valence band of the semiconducting
ZnO25,

26

. Additionally, by zooming in the visible light region of the spectra (see the

inset in Fig. 6), the presence of an emission band that ranges from 440 to nearly 820 nm
is revealed, related to the large surface-to-volume ratio of the ZnO NWs. The exact
origin of such a broad emission band has remained controversial; nevertheless it has
been generally attributed to defects and impurities referred to as a deep-level, trap-state
or impurity-related radiative recombination process (such as oxygen vacancies, Zn
interstitials, oxide antisite defect27, and zinc vacancies). Therefore, the low intensity of
this visible emission indicates the high crystalline quality of the different samples
grown (i.e. NWs have a low density of defects).
The spectra also show a shift in the position of the maximum of the visible light
emission. For the ZnO NWs grown over the c-plane the maximum is located at 550 nm
(i.e., yellow luminescent band), while for those grown over a- and r- planes, the
maximum is located near 520nm (i.e., green luminescent band).

Fig. 6: Room-temperature PL spectra of the three ZnO NW sample.

Typically, defects responsible for the yellow luminescent band (YL) have been excess
of oxygen and the presence of oxygen interstitials (Oi). Nevertheless, for the green
luminescent band (GL), it has been reported to be originated from oxygen vacancies
(Vo). However, the presence of both GL and YL in the same samples suggests a
common origin for them28. We will assume that the formation of such defects occurs
during the growth process. Thereby, the initial Vo will form at the surface. Then, as Zn
and O atoms are supplied, by adsorption, the ZnO layer will be created. By deposition
of more atoms, the created defect will be overgrown and will become a Vo inside the
lattice29. Consequently, if the orientation is different, the Vo will form from a different
kind of surface.
If we correlate now the PL spectra to the gas response to NO2, the c-plane sensor, i.e.
the one that shows the highest response to NO2, employs the material with the most
intense visible emission band in its PL spectrum, centred at 550, which corresponds to
the YL. This means that the sensor with the maximum number of oxygen interstitials is
the most sensitive to NO2. Sensor response to NO2 can be explained as follows; NO2
molecules adsorb on the material surface. Then, this adsorbed molecule acts as an
acceptor according to the following reactions30:
𝑁𝑂2 (𝑔𝑎𝑠) → 𝑁𝑂2 (𝑎𝑑𝑠)
−
𝑒 − + 𝑁𝑂2 (𝑎𝑑𝑠) → 𝑁𝑂2 (𝑎𝑑𝑠)

Furthermore, the tilted morphology of ZnO NWs in the c-plane sensor favours the
presence of many nanowire to nanowire junctions, the space charge regions of which
are modulated upon NO2 chemisorption. This further enhances the response towards
nitrogen dioxide of the c-plane sensor as shown in Fig. 5 b.
However, this does not seem to be the case for EtOH sensing. Fig. 5(a) shows that the
most sensitive sensor for EtOH sensing is the a-plane sensor. In this particular case, as
the maximum in the PL spectra corresponds to the GL, the origin for this visible band is
related to oxygen vacancies. As mentioned above, oxygen vacancies are most likely
responsible for the YL band. These vacant points would act as active sites in which
oxygen will chemisorb during gas sensing31. The more numerous the active sites, the
more oxygen will be chemisorbed and consequently, more molecules of target gas
(EtOH in this case) will react and the sensor response will be dramatically affected.

This suggests that the engineering of defects in single crystalline metal oxides could be
a viable approach to ameliorate selectivity.

Conclusions
To sum up, it has been shown that the morphology and orientation of ZnO NWs, CVDgrown on sapphire substrates are highly dependent on the crystalline plane of the
substrate. By employing room-temperature PL studies, it has been possible to establish
that the crystalline plane of the substrate has also an impact in the number of defects and
in the nature of these (i.e., defects located at the surface or deep-levels within the ZnO
NWs). Sensors have been fabricated and tested employing ZnO NWs with different
orientations. For the detection of nitrogen dioxide, it was found that the response of
ZnO NWs was directly correlated to the overall amount of defects. The higher the
number of defects is (c-plane), the higher the response to nitrogen dioxide is. On the
other hand, for the detection of ethanol, ZnO NWs with an intermediate number of
defects (i.e. a-plane) in which surface defects were dominant led to the best results.
These significant differences revealed by this study suggest that engineering the amount
and nature of defects in metal oxide NWs deserves further research, since it may
become an effective strategy for enhancing and tuning the selectivity of metal oxide
sensors upon demands.

Acknowledgements
This work has been funded in part by MINECO under grant no. TEC2012-32420. S.
Roso and A. Urakawa would like to thank MINECO for support through Severo Ochoa
Excellence Accreditation 2014-2018 (SEV -2013-0319). S. Roso is also grateful to the
URV-ICIQ fellowship. F. Güell and P.R. Martínez-Alanis acknowledge the financial
support from CONACyT through grants no. 203519 and 232459, respectively.

Notes and References
1.
2.
3.

J. M. Wu, Y.-R. Chen and Y.-H. Lin, Nanoscale, 2011, 3, 1053-1058.
M. R. Alenezi, S. J. Henley, N. G. Emerson and S. R. P. Silva, Nanoscale, 2014, 6, 235247.
M. Law, L. E. Greene, J. C. Johnson, R. Saykally and P. Yang, Nat Mater, 2005, 4, 455459.

4.

5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.

20.
21.
22.
23.

24.
25.
26.
27.
28.
29.
30.

J. Fan, Y. Hao, C. Munuera, M. García-Hernández, F. Güell, E. M. J. Johansson, G.
Boschloo, A. Hagfeldt and A. Cabot, The Journal of Physical Chemistry C, 2013, 117,
16349-16356.
D. Vanmaekelbergh and L. K. van Vugt, Nanoscale, 2011, 3, 2783-2800.
S. Chu, G. Wang, W. Zhou, Y. Lin, L. Chernyak, J. Zhao, J. Kong, L. Li, J. Ren and J. Liu,
Nat Nano, 2011, 6, 506-510.
J. Cui, Materials Characterization, 2012, 64, 43-52.
H. V. Han, N. V. Hieu and T. Trung, Science of Advanced Materials, 2014, 6, 1659-1667.
H. Jang, B. Son, H. Song, G. Jung and H. Ko, J Mater Sci, 2014, 49, 8000-8009.
H. Nguyen, C. T. Quy, N. D. Hoa, N. T. Lam, N. V. Duy, V. V. Quang and N. V. Hieu,
Sensors and Actuators B: Chemical, 2014, 193, 888-894.
P. Rai, R. Khan, R. Ahmad, Y.-B. Hahn, I.-H. Lee and Y.-T. Yu, Current Applied Physics,
2013, 13, 1769-1773.
J. Smith, A. Akbari-Sharbaf, M. J. Ward, M. W. Murphy, G. Fanchini and T. Kong Sham,
Journal of Applied Physics, 2013, 113, 093104.
H. Zeng, G. Duan, Y. Li, S. Yang, X. Xu and W. Cai, Advanced Functional Materials, 2010,
20, 561-572.
F. Güell, J. O. Ossó, A. R. Goñi, A. Cornet and J. R. Morante, Superlattices and
Microstructures, 2009, 45, 271-276.
C. Weigand, J. Tveit, C. Ladam, R. Holmestad, J. Grepstad and H. Weman, Journal of
Crystal Growth, 2012, 355, 52-58.
W. I. Park, D. H. Kim, S.-W. Jung and G.-C. Yi, Applied Physics Letters, 2002, 80, 42324234.
M. H. Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind, E. Weber, R. Russo and P. Yang,
Science, 2001, 292, 1897-1899.
N. Han, P. Hu, A. Zuo, D. Zhang, Y. Tian and Y. Chen, Sensors and Actuators B: Chemical,
2010, 145, 114-119.
F. Hernandez-Ramirez, J. D. Prades, A. Tarancon, S. Barth, O. Casals, R. Jimenez-Diaz, E.
Pellicer, J. Rodriguez, J. R. Morante, M. A. Juli, S. Mathur and A. Romano-Rodriguez,
Advanced Functional Materials, 2008, 18, 2990-2994.
J. D. Prades, R. Jimenez-Diaz, F. Hernandez-Ramirez, S. Barth, A. Cirera, A. RomanoRodriguez, S. Mathur and J. R. Morante, Applied Physics Letters, 2008, 93, 123110.
J. Xu, J. Han, Y. Zhang, Y. a. Sun and B. Xie, Sensors and Actuators B: Chemical, 2008,
132, 334-339.
M.-W. Ahn, K.-S. Park, J.-H. Heo, J.-G. Park, D.-W. Kim, K. J. Choi, J.-H. Lee and S.-H.
Hong, Applied Physics Letters, 2008, 93, 263103.
C. Navío, S. Vallejos, T. Stoycheva, E. Llobet, X. Correig, R. Snyders, C. Blackman, P.
Umek, X. Ke, G. Van Tendeloo and C. Bittencourt, Materials Chemistry and Physics,
2012, 134, 809-813.
E. R. Waclawik, J. Chang, A. Ponzoni, I. Concina, D. Zappa, E. Comini, N. Motta, G. Faglia
and G. Sberveglieri, Beilstein Journal of Nanotechnology, 2012, 3, 368-377.
J. S. Reparaz, F. Güell, M. R. Wagner, A. Hoffmann, A. Cornet and J. R. Morante,
Applied Physics Letters, 2010, 96, -.
R. Aad, V. Simic, L. Le Cunff, L. Rocha, V. Sallet, C. Sartel, A. Lusson, C. Couteau and G.
Lerondel, Nanoscale, 2013, 5, 9176-9180.
B. Lin, Z. Fu and Y. Jia, Applied Physics Letters, 2001, 79, 943-945.
O. Marin, G. Grinblat, A. M. Gennaro, M. Tirado, R. R. Koropecki and D. Comedi,
Superlattices and Microstructures, 2015, 79, 29-37.
K. Subannajui, C. Wongchoosuk, N. Ramgir, C. Wang, Y. Yang, A. Hartel, V. Cimalla and
M. Zacharias, Journal of Applied Physics, 2012, 112, 034311.
S. Öztürk, N. Kılınç and Z. Z. Öztürk, Journal of Alloys and Compounds, 2013, 581, 196201.

31.

S. Pati, S. B. Majumder and P. Banerji, Journal of Alloys and Compounds, 2012, 541,
376-379.