PCCP

View Article Online

PAPER View Journal | View Issue
Published on 25 April 2014. Downloaded by CONSIGLIO NAZIONALE DELLE on 11/24/2022 3:43:55 PM.

Crystal plane-dependent gas-sensing properties of

zinc oxide nanostructures: experimental and
Cite this: Phys. Chem. Chem. Phys.,
2014, 16, 11471 theoretical studies†
Yusuf V. Kaneti,a Zhengjie Zhang,a Jeffrey Yue,b Quadir M. D. Zakaria,a
Chuyang Chen,a Xuchuan Jiang*a and Aibing Yua

The sensitivity of a metal oxide gas sensor is strongly dependent on the nature of the crystal surface
exposed to the gas species. In this study, two types of zinc oxide (ZnO) nanostructures: nanoplates and
nanorods with exposed (0001) and (101% 0) crystal surfaces, respectively, were synthesized through facile
solvothermal methods. The gas-sensing results show that sensitivity of the ZnO nanoplates toward ethanol
is two times higher than that of the ZnO nanorods, at an optimum operating temperature of 300 1C. This
could be attributed to the higher surface area and the exposed (0001) crystal surfaces. DFT (Density
Functional Theory) simulations were carried out to study the adsorption of ethanol on the ZnO crystal
Received 25th March 2014, % with adsorbed O ions. The results reveal that the exposed (0001)
planes such as (0001), (101% 0), and (1120)
Accepted 24th April 2014 planes of the ZnO nanoplates promote better ethanol adsorption by interacting with the surface oxygen p
DOI: 10.1039/c4cp01279h (O2p) orbitals and stretching the O–H bond to lower the adsorption energy, leading to the sensitivity
enhancement of the nanoplates. These findings will be useful for the fabrication of metal oxide nanostruc-
www.rsc.org/pccp tures with specifically exposed crystal surfaces for improved gas-sensing and/or catalytic performance.

1. Introduction (001) planes exhibited 5 times higher photoreactivity than those

with exposed (110) facets.
Zinc oxide (ZnO), an n-type semiconductor with a wide band- In order to achieve a better understanding of the sensing
gap (Eg) of 3.37 eV, has attracted considerable attention because mechanism of ZnO-based sensors, some theoretical studies
of its unique optical and electronic properties and potential have been conducted to investigate the adsorption of oxidizing
applications in solar cells,1 nanolasers,2 nanogenerators,3 gases (e.g. NO2, NO, SO2) on the surfaces of ZnO.14–17 For
catalysts,4 environmental remediation,5 and gas sensors.6 instance, Breedon et al.14,18 investigated the adsorption of NO
Recently, there has been growing interest in the synthesis of and NO2 onto stoichiometrically balanced ZnO(101% 0) and
metal oxide nanostructures with specifically exposed crystal (21% 1% 0) crystal planes by using the DFT simulation method
planes and the exploration of their crystal plane-dependent and found that NO2 and NO interacted weakly with the two
properties, in applications such as gas sensors and catalysts.7–12 crystal planes without generating any significant surface dis-
For example, Han et al.13 reported that WO3 nanoparticles tortions. Prades et al.16 examined the adsorption of NO2 on
exhibiting a higher percentage of (010) than (001) or (100) planes ZnO(101% 0) and (112% 0) crystal planes with slightly reduced sur-
showed an enhanced sensitivity toward 1-butylamine. Similarly, face (12.5% O vacancy) by using DFT simulations and demon-
SnO2 octahedral nanoparticles with exposed high energy planes strated that NO2 strongly adsorbed on the surface Zn atoms. In
such as {221} or {111} were found to exhibit enhanced sensitiv- comparison, only a few studies were reported on the adsorption
ities toward ethanol (EthOH) compared to SnO2 nanorods of reducing gases (e.g. CO, alcohols) on the crystal planes of
enclosed with lower energy {101} and {110} planes.7 Lai et al.10 ZnO.19,20 Recently, our group has employed the molecular
demonstrated that TiO2 nanomaterials with exposed (111) and dynamics (MD) simulation method to quantify the diffusivity,
adsorption, and reaction capabilities of volatile n-butanol gas
on the ZnO(101% 0), (112% 0), and (0001) planes.20
a
School of Materials Science and Engineering, The University of New South Wales, Despite some success, many of the previous investigations
Sydney, NSW 2052, Australia. E-mail: [email protected]; Tel: +61-2-9385-5918
b
were still limited to the exploration of one particular ZnO
Department of Chemical Engineering, University College London, Torrington Place,
London WC1E 7JE, UK
crystal surface, which provides only partial information about
† Electronic supplementary information (ESI) available. See DOI: 10.1039/ the sensing mechanism of ZnO-based sensors on the atomic/
c4cp01279h molecular scale. Moreover, the electronic interactions that

This journal is © the Owner Societies 2014 Phys. Chem. Chem. Phys., 2014, 16, 11471--11480 | 11471
 View Article Online

Paper PCCP

occur between EthOH and O modified ZnO crystal surfaces (0.15 M) solution. Then, 2 mmol of CTAB powder was added to
during the gas-sensing process are not yet fully understood, the mixture solution to form a white suspension, followed by
specifically for a few aspects such as: (i) changes in electronic rapid stirring for B20 minutes. This suspension was sub-
states and adsorption energies of the ZnO crystal planes before sequently transferred into a 50 mL Teflon-lined stainless steel
and after interacting with EthOH molecules, (ii) the impact of autoclave, and finally sealed and heated at 150 1C for 16 h. The
the interaction on the bond length of the EthOH molecules, (iii) washing procedures were similar to those described for ZnO
the electron/charge distribution on the ZnO surfaces after hexagonal nanoplates.
Published on 25 April 2014. Downloaded by CONSIGLIO NAZIONALE DELLE on 11/24/2022 3:43:55 PM.

EthOH adsorption, and (iv) the adsorption mechanism of

EthOH molecules on the ZnO surfaces. 2.3 Characterization
This study aims to develop a facile hydrothermal or
Field emission scanning electron microscopy (FESEM) analysis
solvothermal process for the synthesis of ZnO nanostructures
of the ZnO samples was performed using a FEI Nova NanoSEM
with exposed (101% 0) and (0001) crystal planes for gas-sensing
230. Transmission electron microscopy (TEM) images were
application, using EthOH as a case study. The crystal plane-
obtained using a Tecnai G2 microscope operated at an accel-
dependent gas-sensing properties of the as-prepared ZnO nano-
erating voltage of 200 kV. High resolution transmission elec-
plates and nanorods will be examined, both experimentally and
tron microscopy (HRTEM) was performed using a Phillips
theoretically. To fundamentally understand the sensing
CM200 field emission gun transmission electron microscope
mechanisms, DFT simulations will be employed to quantify
with an acceleration voltage of 200 kV. The crystal structures of
the changes in density of states (DOS), adsorption energies, and
the as-prepared ZnO products were characterized using X-ray
bond angles or distances on different ZnO crystal surfaces (i.e.
diffraction (a Phillips X’pert multipurpose X-ray diffraction
(101% 0), (112% 0), and (0001) surfaces) modified with adsorbed O ,
system (MPD) equipped with graphite mono-chromatized Cu
before and after interaction with EthOH. The findings will
Ka radiation (l = 1.54 Å)) in the 2y range of 201–701. X-ray
provide a greater understanding of the sensing mechanism
photoelectron spectra (XPS) were recorded on an ESCALAB250Xi
of ZnO-based sensors on the atomic/molecular scale and will
X-ray photoelectron spectrometer, using Al-Ka radiation as the
be beneficial for the fabrication of other metal oxide nano-
exciting source. The Brunauer–Emmett–Teller (BET) surface area
materials with highly exposed crystal surfaces for potential
and pore size distribution of the products were obtained from
applications in gas sensors and catalysts.
nitrogen physisorption isotherms (adsorption–desorption branches)
at 77 K on a Micromeritics Tristar 3000 instrument. Prior to the
measurements, the ZnO samples were degassed overnight under
2. Experimental section vacuum at 150 1C to vaporize water molecules adsorbed on the ZnO
2.1 Chemicals materials.
The chemicals used in this study are zinc chloride (ZnCl2, 99%),
zinc acetate dehydrate ((Zn(CH3COO)2
2H2O, 99%), sodium 2.4 Gas sensor fabrication
hydroxide (NaOH, 99%), sodium acetate (NaC2H3O2, 99%), To prepare the gas sensor, each ZnO product was mixed and
ethanol (C2H6O, 95%), absolute ethanol (C2H6O, 99.9%), and ground with polyvinylidene fluoride (PVDF) and 1-methyl-2-
cetyltrimethylammonium bromide (CTAB, 99%), 1-methyl-2- pyrollidone in an agate mortar. The resulting paste was spread
pyrrolidinone) (C5H9NO, 99.5%), and polyvinylidene fluoride uniformly onto an alumina ceramic tube with a pair of
((CH2CF2)n, 99.5%). All chemicals were purchased from Sigma previously-printed Au electrodes. The coated ceramic tube
Aldrich and used as received without further purification. was subsequently aged at 450 1C for a period of 3 h to improve
Distilled water was used in all the synthesis processes. its thermal stability. Finally a Ni–Cr alloy coil was inserted into
the alumina tube as a heater, which provided the working
2.2 Synthesis of ZnO nanostructures temperature of the gas sensor. The gas-sensing tests were
ZnO hexagonal nanoplates. ZnO hexagonal nanoplates with performed on a WS-30A gas sensing measurement system
exposed (0001) were synthesized by using a modified solvo- (Weisheng Electronics Co., Ltd, Henan, China) at a relative
thermal process.21 In a typical protocol, 0.66 g of Zn(CH3COO)2
humidity of about 40–60% and a working voltage of 5 V, with a
2H2O, 0.24 g of NaOH, and 1.6 g of NaCH3COO were succes- reference resistor of 1 MO, using air as the dilution and
sively dissolved in a water (20 mL)–ethanol (10 mL) mixed reference gas. The gas-sensing measurement diagram is shown
solvent. The white suspension was stirred for B20 minutes in Fig. S1 (ESI†).
and then transferred into a 50 mL Teflon-lined stainless steel For the gas-sensing tests, a static state gas-distribution
autoclave, which was sealed and maintained at 150 1C for 24 h. method was used, as shown in Fig. S2 (ESI†). Firstly, the
Finally, it was cooled to room temperature naturally and the sensors were pre-heated to ensure stable data collection due
white precipitates were collected, rinsed with deionized water to the moisture adsorbed on the materials. Once, the sensors
and ethanol 3 times along with centrifugation, and dried at were stabilized, a calculated volume of the test gas was injected
60 1C for 5 h. into the glass-inlet at the back of the chamber using a micro-
ZnO nanorods. In a typical procedure, a 2 mL solution syringe and dropped onto a small hot-plate located inside the
of ZnCl2 (0.05 M) was firstly mixed with a 40 mL of NaOH chamber. To achieve a certain concentration of ethanol vapor,

11472 | Phys. Chem. Chem. Phys., 2014, 16, 11471--11480 This journal is © the Owner Societies 2014
 View Article Online

PCCP Paper

the amount of liquid ethanol (99.9% purity) that needed to be

injected was calculated according to the equation:

Vliquid = (Vchamber  1000  Cgas  MWgas  Troom  10 9)/

(22.4  rliquid  Tsurrounding) (1)
Published on 25 April 2014. Downloaded by CONSIGLIO NAZIONALE DELLE on 11/24/2022 3:43:55 PM.

where Vliquid is the volume of liquid ethanol to be injected (mL),

Vchamber is the total volume of the chamber (18 L), Cgas is the
concentration of ethanol gas (ppm), MWgas is the molecular
weight of ethanol gas, Troom is the room temperature (298 K),
rliquid is the density of liquid ethanol (g cm 3), and Tsurrounding
is the surrounding temperature (K).
As the injected gas was initially in the liquid form, it was
evaporated to convert it to a gas state. Once the maximum
sensitivity of the sensors was reached, the chamber was opened
in order to release the gas or to allow the sensors to recover.
During the gas-sensing measurements, the operating tempera-
ture was controlled by adjusting the heating voltage. The Scheme 1 (a) Side view (top) and top view (bottom) of the three different
sensitivity (S) of the sensor is defined as: S = Ra/Rg, where Ra ZnO crystal surfaces (red and blue atoms represent O and Zn, respec-
and Rg correspond to the resistance of the sensor in air and the tively); schematic illustration of the crystal faces in: (b) ZnO hexagonal
nanoplates, and (c) ZnO nanorods.
resistance in the tested gas, respectively. The response time
(tres) is defined as the time required for the conductance
variation of the sensor to reach 90% of its equilibrium value, where Eads is the adsorption energy, E(EthOH/ZnO) is the total
following exposure to ethanol vapor. The recovery time (trec) is energy after adsorption, EEthOH is the total energy of the EthOH
defined as the time needed by the sensor to return to 10% of its molecule, and EZnO is the total energy of three bare ZnO
original conductance in air, after removal of the ethanol vapor. surfaces after optimization. A negative binding energy indicates
The digital photographs of the prepared ZnO sensors are given favorable adsorption and the more negative the value, the
in Fig. S3 (ESI†). stronger the adsorption capability.

2.5 Computational simulations 3. Results and discussion

Density Functional Theory (DFT) simulations were conducted 3.1 Composition and morphology
using the commercial software: Materials Studio (Version 4.3,
Accelrys Inc, 2007) with the CASTEP22 module and the widely The phase composition of the achieved products was charac-
used GGA (generalized gradient approximation) for the terized using the XRD technique as shown in Fig. 1. All of the
exchange–correlation using PBE (Perdew, Burke, and Emzerhof). diffraction peaks in both products can be well indexed to
The self-consistent ground state cell optimization was performed hexagonal wurtzite ZnO (JCPDS No. 36-1451). Furthermore,
by the density mixing scheme using an ultra-soft pseudo-potentials the absence of diffraction peaks due to impurities in both
(USP) plane-wave basis set, with a cut-off energy of 400 eV and a ZnO samples indicates their high purity. The sharp, clear
thermal broadening of 0.1 eV. The Monkhorst–Pack was used to distinct peaks observed in the XRD patterns of these two
determine the k-points, which was fixed at 0.04 Å 1 for all samples suggest that the as-synthesized ZnO nanostructures
calculations. are well-crystallized.
The surface simulations were performed using a slab model
under periodic boundary conditions (Scheme 1). A supercell
with a surface thickness of B10 Å and a vacuum space of 20 Å
was created from the ZnO lattice structure in which the bottom
layers (B6 Å) were fixed and the top layers were relaxed to
simulate the surface properties. In order to simulate the gas
sensing environment as close as possible, one O ion was
bonded with the surface Zn atom on the optimized ZnO sur-
face. Then, the EthOH molecule was positioned B2.0 Å from
the surface above different adsorption surfaces.
The adsorption energy values of the EthOH molecule to
different surfaces were calculated using the following formula:

Eads = [E(EthOH/ZnO) (EEthOH + EZnO)] (2) Fig. 1 XRD patterns of the as-prepared ZnO nanoplates and nanorods.

This journal is © the Owner Societies 2014 Phys. Chem. Chem. Phys., 2014, 16, 11471--11480 | 11473
 View Article Online

Paper PCCP

they grow mainly along six symmetric directions of [101% 0],

[11% 00], and [011% 0]. Therefore, the top and bottom surfaces
of the nanoplates expose (0001) planes, as represented by
Scheme 1b.20
In comparison, the as-prepared ZnO nanorods have lengths
and diameters of B1–2.5 mm and B200–400 nm, respectively,
as shown in Fig. 2c and d. The HRTEM image of a single
Published on 25 April 2014. Downloaded by CONSIGLIO NAZIONALE DELLE on 11/24/2022 3:43:55 PM.

nanorod (Fig. 3e) shows distinct lattice fringes with a d-spacing

of B0.26 nm, corresponding to the d-spacing of the (0001)
plane of the hexagonal wurtzite ZnO. Furthermore, it is
observed that the identified (0001) plane is perpendicular to
the axis of the nanorods, suggesting that the preferred growth
of the nanorods is along the [0001] direction.23–25 The SAED
pattern of the ZnO nanorods shown in Fig. 3f can be ascribed to
the diffraction pattern of the [101% 0] zone axis of the hexagonal
wurtzite ZnO. This indicates that the ZnO nanorods expose the
(101% 0) planes, as schematically illustrated in Scheme 1c.
In the present reaction system, the formation of ZnO crystals
initially originates from the reaction between Zn2+ (from ZnCl2
Fig. 2 Low and high magnification SEM images of (a, b) ZnO nanoplates
and (c, d) ZnO nanorods.
in the case of ZnO nanorods or Zn(CH3COO)2 for the ZnO
nanoplates) and OH , which generates the precursor zinc
hydroxide, Zn(OH)2. The formed Zn(OH)2 precipitate can sub-
Fig. 2a and b shows the SEM images of the as-prepared ZnO sequently be dissolved by the superfluous OH ions present in
nanoplates, which have diameters of 100–150 nm and are the solution to generate [Zn(OH)4]2 ions, which have been
hexagonal-like. The HRTEM analysis of an individual hexa- proposed to be the growth units of ZnO.26 As the solution is
gonal nanoplate (Fig. 3b) reveals clear lattice fringes, with a heated to high temperature, ZnO clusters are formed from the
d-spacing of B0.28 nm, which corresponds well to the d-spacing dehydration or decomposition of [Zn(OH)4]2 ions. Eventually,
of the (101% 0) plane of the hexagonal wurtzite ZnO. The selected the solution reaches supersaturation from the formation of
area electron diffraction (SAED) pattern of the ZnO hexagonal these clusters and nucleation begins, followed by the crystal
nanoplates shown in Fig. 3c can be indexed to the [0001] zone- growth.
axis of single crystal hexagonal wurtzite ZnO. The HRTEM and The crystal structure of hexagonal wurtzite ZnO is composed
the corresponding SAED results of the nanoplates reveal that of positively charged Zn-(0001) and negatively charged O-(0001% )

Fig. 3 TEM and HRTEM images and the corresponding SAED pattern of an individual (a–c) ZnO nanoplate, and (d–f) ZnO nanorod.

11474 | Phys. Chem. Chem. Phys., 2014, 16, 11471--11480 This journal is © the Owner Societies 2014
 View Article Online

PCCP Paper
Published on 25 April 2014. Downloaded by CONSIGLIO NAZIONALE DELLE on 11/24/2022 3:43:55 PM.

Fig. 4 The nitrogen (N2) adsorption–desorption isotherms of (a and b) ZnO hexagonal nanoplates and (c and d) ZnO nanorods.

polar surfaces. Intrinsically, the rate of crystal growth of ZnO is Table 1 Results of Gaussian fitting of O1s spectra of ZnO nanoplates and
in the order: [0001] c [101% 0] c [0001% ] as the surface energy of nanorods
the (0001) plane is higher than those of (101% 0) and (0001% )
OL OV OC
planes.27 Hence, in the absence of surfactants or other capping Samples Zn 2p (Zn–O) (vacancy) (chemisorbed)
agents, the ZnO crystal has a strong tendency to grow along the
ZnO nanoplates
[0001] direction or the c-axis. Previously, the hydroxyl (OH ) ion Binding energy (eV) 1021.41 530.28 531.56 532.3
has been identified to preferentially adsorb on the (0001) plane Relative percentage (%) 75.1 15.6 9.3
of ZnO at a high concentration level.28 For the synthesis of ZnO
ZnO nanorods
nanorods, a low Zn2+ : OH ratio is used, and therefore only a Binding energy (eV) 1021.37 530.24 531.56 532.42
few growth units, [Zn(OH)4]2 , are generated and the intrinsic Relative percentage (%) 71.0 21.0 8.0
growth habit of the ZnO crystal along the [0001] direction
cannot be well suppressed by the OH ions. This results in a
much faster growth rate along [0001], ultimately leading to the adsorption of O2 molecules, which play an important role in the
formation of nanorods. In contrast, in the synthesis of ZnO sensing mechanism of the ZnO-based sensor. The XPS techni-
hexagonal nanoplates, the acetate ions (CH3COO ) which are que was employed to gain further information regarding the
present in high concentration can preferentially adsorb on the surface composition and features of the as-prepared ZnO
(0001) plane of ZnO, in conjunction with the adsorbed hydroxyl nanostructures. Fig. S4a (ESI†) compares the high resolution
ions.21 As a result, the intrinsically fast growth rate of the ZnO spectra of the Zn 2p peaks of both the ZnO nanorods and
crystal along the [0001] direction is greatly suppressed by the hexagonal nanoplates. It can be observed that the position of
adsorption of CH3COO and OH ions on the (0001) plane, the Zn 2p peak is nearly identical for the two ZnO samples with
while the lateral growth along the [101% 0] direction remains the Zn 2p peak located at a binding energy of 1020.98 eV for the
uninhibited and is therefore faster, resulting the formation of ZnO nanoplates and at 1021.68 eV for the ZnO nanorods. This
two-dimensional nanoplates (Fig. 2a and b). confirms the Zn2+ valency state in the prepared ZnO products.29
Surface area is one of the key factors influencing the gas- As shown in Fig. S4b and c (ESI†), the O1s XPS peak of both
sensing performance of metal oxide gas sensors. The surface ZnO nanostructures can be deconvoluted into three main
areas of the as-prepared ZnO nanostructures were evaluated by components after Gaussian fitting: OL which corresponds to
the BET method using N2 adsorption–desorption processes. the O2 ions in the ZnO crystal, OV which is indexed to the
The BET specific surface area of the ZnO nanoplates is mea- O2 ions in the oxygen-deprived areas within the ZnO crystal
sured to be 14.4 m2 g 1, approximately twice of that of the lattice, and OC which represents chemisorbed oxygen species.23
ZnO nanorods (7.43 m2 g 1), as shown in Fig. 4. The larger The binding energies of these oxygen components are given in
surface area of the nanoplates may provide more sites for the Table 1. The relative percentages of the OC component in the

This journal is © the Owner Societies 2014 Phys. Chem. Chem. Phys., 2014, 16, 11471--11480 | 11475
 View Article Online

Paper PCCP
Published on 25 April 2014. Downloaded by CONSIGLIO NAZIONALE DELLE on 11/24/2022 3:43:55 PM.

Fig. 5 (a) The sensitivities of the as-prepared ZnO hexagonal nanoplates and nanorods toward 100 ppm of EthOH as a function of the working
temperature; (b) the dynamic response–recovery curves; (c) the response–recovery times of the sensors made from the as-prepared ZnO
nanostructures upon exposure to different concentrations of EthOH at the optimum working temperature of 300 1C; (d) comparison of the sensitivities
of the ZnO hexagonal nanoplates and nanorods toward various concentrations of EthOH at 300 1C.

two ZnO samples are 9.3% for the ZnO nanoplates and 8.0% upon injection of different concentrations of EthOH are shown
for the ZnO nanorods, suggesting the higher capability of in Fig. 5b. Evidently, the output voltages of both ZnO sensors
the nanoplates to absorb more oxygen species compared to the increase as EthOH is introduced into the test chamber but
nanorods. These results can be used to understand the shape- gradually decrease as the sensors are exposed to air, and such
dependent gas-sensing performance of ZnO nanostructures. behavior is consistent with that of an n-type semiconducting
gas sensor. However, it is clear that the increase in the output
3.2 Gas-sensing performance voltage upon exposure to EthOH is greater for ZnO nanoplates
The ZnO nanorods and hexagonal nanoplates exhibit different than for ZnO nanorods, indicating the higher sensitivity of the
surface properties (different surface areas, exposed crystal nanoplates. The response and recovery times of the two ZnO
planes, and OC percentages), which may lead to different gas- sensing materials toward different concentrations of EthOH are
sensing properties. As a case study, their sensitivities toward shown in Fig. 5c. Clearly, the ZnO hexagonal nanoplates exhibit
EthOH were investigated in this work. The gas-sensing proper- faster response times than the ZnO nanorods, however, the
ties of ZnO sensors are heavily affected by the working tem- recovery times of the nanoplates are slightly longer than those
perature.14,30 Therefore, the sensitivities of the as-prepared ZnO
nanostructures toward 100 ppm of EthOH were evaluated as a
function of working temperature. From Fig. 5a, it can be
observed that the sensitivity of the ZnO nanoplate sensor
increases gradually with the rise in the working temperature
from 210 to 300 1C and reaches a maximum value of S = 24.6 at
an optimum working temperature of 300 1C. A further increase
in the operating temperature above 300 1C, however, results in
a gradual decline in the sensitivity of the ZnO nanoplates. A
similar trend is observed for the ZnO nanorods. That is, the
optimized working temperatures for both ZnO sensors are
300 1C, however, they exhibit different sensitivities, with S =
24.6 for ZnO nanoplates and S = 14.1 for ZnO nanorods.
The dynamic response–recovery behaviors of the sensors Fig. 6 Stability evaluations of the as-prepared ZnO sensing materials
fabricated from both ZnO hexagonal nanoplates and nanorods toward 100 ppm of EthOH over a testing period of 7 days.

11476 | Phys. Chem. Chem. Phys., 2014, 16, 11471--11480 This journal is © the Owner Societies 2014
 View Article Online

PCCP Paper

of the nanorods, due to the stronger interaction with EthOH, Table 2 Summary of calculated bond distances and adsorption energy
and hence more time is needed to return to the initial state.
Surface
The changes in sensitivities of ZnO nanoplates and nano-
Gas phase 101% 0 112% 0 0001
rods upon exposure to different amounts of EthOH are shown
in Fig. 5d. It can be observed from this figure that the sensitivity Bond O–H1 (Å) 0.977 1.020 1.040 1.695
distance–angle C1–H2 (Å) 1.098 1.099 1.102 1.109
always increases with increasing EthOH concentration for both C1–O (Å) 1.438 1.449 1.442 1.406
ZnO nanostructures, and that the sensitivity of the ZnO C1–C2 (Å) 1.513 1.513 1.514 1.518
Published on 25 April 2014. Downloaded by CONSIGLIO NAZIONALE DELLE on 11/24/2022 3:43:55 PM.

hexagonal nanoplates is almost twice that of the nanorods,

H1–O–C1 (1) 108.6 111.5 112.1 125.3
with Splates = 48 vs. Srods = 28.1, at an EthOH concentration of O–C1–C2 (1) 108.0 110.9 111.6 115.8
500 ppm. Aside from sensitivity and response–recovery beha-
viors, the stabilities of the two ZnO sensing materials were also OEthOH–ZnZnO (Å) — 2.118 2.051 1.937
H1–O ZnO (Å) — 1.679 1.499 1.05
assessed toward 100 ppm of EthOH at the working temperature
of 300 1C, over 7 days. The results in Fig. 6 confirm that both Eadsorption (eV) — 1.12 1.36 5.72
ZnO nanostructures are highly stable sensor materials, with
only minor changes in their sensitivities from 23.7–27.5 for
ZnO nanoplates and 13.3–16.2 for ZnO nanorods, respectively. 3.3 DFT simulation
The average S value of the ZnO nanoplates over the 7-day period To further support the experimental results, the DFT simula-
is 25, with a standard deviation of 1.43, whereas the average S tion method was used to simulate the adsorption behavior of
value of the ZnO nanorods over the 7-day period is 15.2, with a EthOH molecules on ZnO(101% 0), (112% 0) and (0001) crystal
standard deviation of 1.05. planes, modified with adsorbed O ions, which are commonly
The gas-sensing mechanism of these ZnO sensors is based found at high operating temperatures (Z300 1C).34,35 The
on the changes in their resistance upon introduction of EthOH modification with other adsorbed oxygen species such as O2
caused by the adsorption–desorption of the EthOH molecules or O2 might be possible, but the DFT simulation results show
on their surfaces.31,32 In air, oxygen (O2) molecules adsorb on extremely low or extremely high energies for these cases (and
the surface of ZnO and ionize to form ionic oxygen species they also do not converge), which possibly suggests that O is
(mainly O at 300 1C) by capturing electrons from the conduc- the most stable configuration for oxygen species in a typical
tion band of ZnO.33 Such reactions lead to the formation of a gas-sensing experiment. Scheme 1a shows the side view and top
thick space charge layer which increases the potential barrier, view of the three surfaces of ZnO. These were firstly optimized,
and ultimately increases the resistance of the ZnO sensor. and then the EthOH molecule was placed on each surface
However, when the ZnO sensor is exposed to EthOH, the EthOH during the simulations.
molecules interact with the chemisorbed oxygen species (O ) Table 2 shows the bond distances between various atoms,
on the surface of the ZnO to form CO2 and H2O and release the the bond angles of the EthOH molecule before and after
trapped electrons back into the conduction band of ZnO, adsorption and the adsorption energies on O modified
according to the equation: ZnO(101% 0), (112% 0) and (0001) crystal planes. The DFT calcula-
tion performed for the free EthOH molecule shows that it
C2H5OH (ads) + 6O - 2CO2 (gas) + 3H2O (gas) + 6e exhibits bond distances of 0.977, 1.098, 1.438, and 1.513 Å,
(3) corresponding to O–H, C–H, C–O, and C–C bonds, respectively.
The results obtained via CASTEP are consistent with the
This process increases the carrier concentration, which reported theoretical results,36,37 and experimental results.38
consequently decreases the overall resistance of the ZnO As the EthOH molecule approaches the O adsorbed ZnO
sensor. surface, the O(EthOH) atom forms a bond with the surface Zn
The above results clearly show that the gas-sensing perfor- atom. The H(EthOH) atom will also form a bond with O ,
mance of ZnO nanostructures is highly dependent on the mainly in the form of a hydrogen bond. The DFT calculations
surface properties. Surface area is one of the most important indicate that the distance between H(EthOH) and O on the
factors affecting the sensitivity of metal oxide gas sensors. The ZnO(0001) surface is the smallest (1.05 Å), as shown in Table 2.
larger the surface area, the more adsorption sites are available This value is smaller than the reported value of 1.464 Å by Yuan
for the adsorption of O2 molecules on the surfaces of the et al.,19 due to the addition of O in our present simulation.
nanostructures, and therefore the higher the sensitivity. More-
over, besides the surface area, the highly exposed (0001) crystal
surfaces of the ZnO nanoplates may also play an important role Table 3 Mulliken charge (e) of O ions, H(EthOH) and O(EthOH) after
EthOH adsorption
in enhancing their sensitivity toward EthOH. To further under-
stand the sensing mechanism and to compare the EthOH Surface
adsorption capabilities of different ZnO crystal planes, we have Mulliken charge 101% 0 112% 0 0001
carried out theoretical (DFT) simulations on the three main
O ion 0.50 0.50 0.84
ZnO crystal planes: ZnO(0001), (112% 0) and (101% 0), as described H(EthOH) 0.38 0.38 0.37
in the following context. O(EthOH) 0.65 0.66 0.65

This journal is © the Owner Societies 2014 Phys. Chem. Chem. Phys., 2014, 16, 11471--11480 | 11477
 View Article Online

Paper PCCP

observed on the LaFeO3(010) surface.39 This can be confirmed by

the Mulliken charges of related atoms, as listed in Table 3. After
adsorption, the Mulliken charge of O ions on the ZnO(0001)
surface is about 0.84e, the lowest among three ZnO surfaces,
suggesting that a strong interaction exists between the O ion and
H(EthOH). In fact, the Mulliken charge of O(EthOH) is 0.65e,
which is greater than that of the O ion ( 0.84e), indicating that
Published on 25 April 2014. Downloaded by CONSIGLIO NAZIONALE DELLE on 11/24/2022 3:43:55 PM.

H(EthOH) has been dissociated from the molecule. In addition,

the C–C–O and C–O–H angles of the EthOH molecule (Fig. 7a)
also change compared to the free EthOH molecule, as shown in
Table 2. Similar to the changes in bond distances, the most
significant changes in bond angles occur when the EthOH
molecule is adsorbed on the ZnO(0001) crystal plane, where the
C1–C2–O and C1–O–H1 angles change from 108.0 and 108.61 to
125.3 and 115.81, respectively.
Fig. 7 (a) Free EthOH molecule (red atoms, grey atoms and white atoms
represent O, C and H, respectively). Slice view of the electron density The above analysis reveals that the strongest interaction
difference of EthOH adsorbed on O modified: (b) ZnO(101% 0), (c) occurs on the ZnO(0001) surface compared to the other ZnO
%
ZnO(1120), and (d) ZnO(0001) crystal planes. surfaces. This is further confirmed by the calculation of the
adsorption energies (Eads), where the ZnO(0001) surface is
found to exhibit the strongest adsorption energy of 5.72 eV
Furthermore, from Table 2, it can be observed that the O–H after the interaction with EthOH. This implies that exposed
bond distance of the EthOH molecule stretches following its (0001) crystal planes of the ZnO nanoplates promote better
interactions with the surface Zn atoms. The greatest elongation ethanol adsorption, and therefore higher sensitivity.
of the O–H bond of the EthOH molecule is found to be on the Further insights into the bonding mechanisms between the
ZnO(0001) crystal plane, where the O–H bond distance increases EthOH molecule and the ZnO crystal planes are obtained by
from 0.977 to 1.695 Å. analyzing the density of states (DOS) of the adsorbed EthOH
The stronger adsorption of the EthOH molecule on ZnO(0001) molecule, adsorbed O ions and surface Zn and O atoms. Fig. 8
leads to a great elongation of the O–H bond, similar to that shows the DOS of the EthOH molecule before (a) and after

Fig. 8 (a) Total DOS plot of a single EthOH molecule; (b) PDOS plots of the surface oxygen 2p (O2p) and zinc 3d (Zn3d) orbitals on the ZnO(0001)
surface; (c) the PDOS plot of adsorbed O ions on the ZnO(0001) surface; (d) the total DOS plot of the EthOH molecule after adsorption.

11478 | Phys. Chem. Chem. Phys., 2014, 16, 11471--11480 This journal is © the Owner Societies 2014
 View Article Online

PCCP Paper

H(EthOH) atom results in the shifting and broadening of DOS

peaks of the adsorbed EthOH molecule, suggesting a chemical
adsorption. The broadening of the orbitals has also been
reported on rutile TiO2(110) surfaces.41 The DOS of the EthOH
molecule after adsorption on the ZnO(0001) surface becomes
broaden and the HOMO 1 molecular orbital of EthOH shifts
upward, indicating the occurrence of strong orbital hybridization.
Published on 25 April 2014. Downloaded by CONSIGLIO NAZIONALE DELLE on 11/24/2022 3:43:55 PM.

In comparison, the DOS of the EthOH molecule adsorbed on

O modified ZnO(101% 0) and (112% 0) crystal planes is nearly
identical, with the HOMO molecular orbital shifting downward
and becoming weaker. However, the HOMO molecular orbital
of the EthOH molecule adsorbed on the (112% 0) plane is slightly
higher than that adsorbed on the ZnO(101% 0) plane, indicating
a slightly stronger hybridization on the ZnO(112% 0) plane. The
above analysis provides the following insights regarding the
ethanol-sensing mechanism of the ZnO sensor: (i) the adsorp-
tion of EthOH on the ZnO surface is mainly caused by the
strong interaction between adsorbed O ions and H(EthOH)
atoms, (ii) strong hybridization occurs between the adsorbed
O ion and the HOMO, HOMO 1 molecular orbitals and this is
the main driving force for surface reactions, and (iii) the
ZnO(0001) crystal plane shows a higher adsorption capability
Fig. 9 Total DOS plots of the EthOH molecule adsorbed on O modified for EthOH than the ZnO(101% 0) and (112% 0) planes.
%
(a) ZnO(0001), (b) ZnO(101% 0), and (c) ZnO(1120); (d) the total DOS plot of
the free EthOH molecule.
4. Conclusions
In this study, we have demonstrated facile solvothermal methods
adsorption (d) on the ZnO(0001) surface. The gap between the
for the synthesis of ZnO nanoplates and nanorods with exposed
lowest-unoccupied molecular orbital (LUMO) and the highest
(0001) and (101% 0) crystal planes, respectively. The gas-sensing results
occupied molecular orbital (HOMO) gap is 4.5 eV, as shown in
show that ZnO hexagonal nanoplates exhibit two times higher
Fig. 8a. Fig. 8b shows the PDOS plots of the surface O2p orbital
sensitivity toward EthOH than the ZnO nanorods at a similar
and the Zn3d orbital. The obtained PDOS results are in good
optimum operating temperature of 300 1C. This could be contrib-
agreement with the report of Breedon et al.18 who demon-
uted to two main factors based on the experimental findings: one is
strated that the p orbitals of the surface oxygen are mainly
the larger surface area of the nanoplates compared to the nanorods
located at a higher energy range ( 4.0 eV to Fermi energy level),
(14.4 m2 g 1 vs. 7.43 m2 g 1) and another is the exposed (0001)
while at a lower energy range ( 6.0 to 4.0 eV), the hybridiza-
planes of the nanoplates. To confirm the beneficial effect of the
tion is mainly caused by the Zn3d orbitals and EthOH molec-
exposed (0001) planes, DFT simulations were used to quantify the
ular orbitals. These results imply that the adsorbed oxygen
ethanol adsorption on O modified ZnO(0001), (101% 0), and (112% 0)
species on the surface of the ZnO play an important role in the
crystal surfaces. The results reveal that the exposed (0001) crystal
gas adsorption process, which is in accordance with a previous
planes of the ZnO nanoplates show: (i) stronger adsorption energy
DFT study on the adsorption of EthOH on the SnO2(110)
( 5.72 eV on the ZnO(0001) plane vs. 1.12 eV on the ZnO(101% 0)
surface.40 Fig. 8c shows the DOS of the adsorbed O ion and
plane) and (ii) stronger hybridization between the adsorbed O2p
it is clear that strong hybridization occurs between the O ion
orbital and the HOMO, HOMO 1 molecular orbitals, suggesting a
and HOMO, HOMO 1 molecular orbitals of EthOH. This
stronger adsorption of EthOH molecules adsorbed on the
suggests that the adsorbed O ion plays an important role in
ZnO(0001) plane. The combined experimental and theoretical
the adsorption of EthOH on ZnO crystal planes.
results may provide new insights into the crystal plane-dependent
For further confirmation, the DOS of the EthOH molecule
gas-sensing performance of metal oxide gas sensors. Furthermore
adsorbed on O modified ZnO(0001), (101% 0) and (112% 0) crystal
these findings will be useful for designing and constructing novel
planes were analyzed, as shown in Fig. 9. It is observed that
nanostructures with specifically exposed crystal planes for surface-
after adsorption, the DOS of the EthOH molecule are broa-
related applications such as gas sensors and catalysts.
dened due to the strong hybridization with the adsorbed
O ion. The DOS of the adsorbed EthOH molecule on the three
ZnO crystal planes are quite different to that of the free EthOH Acknowledgements
molecule. This indicates that the electron densities within the
EthOH molecule are changed in the adsorption process. We gratefully acknowledge the financial support of the Australian
The strong interaction between the adsorbed O ion and the Research Council (ARC) projects. The authors also acknowledge

This journal is © the Owner Societies 2014 Phys. Chem. Chem. Phys., 2014, 16, 11471--11480 | 11479
 View Article Online

Paper PCCP

access to the UNSW node of the Australian Microscopy and 20 Y. V. Kaneti, J. Yue, X. Jiang and A. Yu, J. Phys. Chem. C,
Microanalysis Research Facilities (AMMRF). The authors thank 2013, 117, 13153–13162.
Dr Jason Scott of the Particle Catalysis (PARTCAT) group for the 21 Z. Li, W. Pan, D. Zhang and J. Zhan, Chem. – Asian J., 2010,
assistance with the BET measurements. 5, 1854–1859.
22 S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip,
M. I. J. Probert, K. Refson and M. C. Payne, Z. Kristallogr.,
References 2005, 220, 567–570.
Published on 25 April 2014. Downloaded by CONSIGLIO NAZIONALE DELLE on 11/24/2022 3:43:55 PM.

1 E. Hosono, S. Fujihara, I. Honma and H. Zhou, Adv. Mater., 23 X.-G. Han, H.-Z. He, Q. Kuang, X. Zhou, X.-H. Zhang, T. Xu,
2005, 17, 2091–2094. Z.-X. Xie and L.-S. Zheng, J. Phys. Chem. C, 2008, 113,
2 M. H. Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind, 584–589.
E. Weber, R. Russo and P. Yang, Science, 2001, 292, 1897–1899. 24 Y. Ding, P. X. Gao and Z. L. Wang, J. Am. Ceram. Soc., 2004,
3 Z. L. Wang and J. Song, Science, 2006, 312, 242–246. 126, 2066–2072.
4 A. Sinhamahapatra, A. K. Giri, P. Pal, S. K. Pahari, H. C. 25 J. Y. Lao, J. G. Wen and Z. F. Ren, Nano Lett., 2002, 2,
Bajaj and A. B. Panda, J. Mater. Chem., 2012, 22, 17227–17235. 1287–1291.
5 J.-Y. Dong, C.-H. Lin, Y.-J. Hsu, S.-Y. Lu and D. S.-H. Wong, 26 R. Shi, P. Yang, X. Dong, Q. Ma and A. Zhang, Appl. Surf. Sci.,
CrystEngComm, 2012, 14, 4732–4737. 2013, 264, 162–170.
6 G. Korotcenkov, Mater. Sci. Eng., B, 2007, 139, 1–23. 27 B. Li and Y. Wang, J. Phys. Chem. C, 2009, 114, 890–896.
7 P. Song, Q. Wang and Z. Yang, Sens. Actuators, B, 2012, 168, 28 B. Cao and W. Cai, J. Phys. Chem. C, 2007, 112, 680–685.
421–428. 29 Y. Chang, J. Xu, Y. Zhang, S. Ma, L. Xin, L. Zhu and C. Xu,
8 C. Wang, L. Yin, L. Zhang, Y. Qi, N. Lun and N. Liu, J. Phys. Chem. C, 2009, 113, 18761–18767.
Langmuir, 2010, 26, 12841–12848. 30 J. Huang, Y. Wu, C. Gu, M. Zhai, K. Yu, M. Yang and J. Liu,
9 D. Su, H. Fu, X. Jiang and G. Wang, Sens. Actuators, B, 2013, Sens. Actuators, B, 2010, 146, 206–212.
186, 286–292. 31 Z. Yang, L.-M. Li, Q. Wan, Q.-H. Liu and T.-H. Wang, Sens.
10 Z. Lai, F. Peng, H. Wang, H. Yu, S. Zhang and H. Zhao, Actuators, B, 2008, 135, 57–60.
J. Mater. Chem. A, 2013, 1, 4182–4185. 32 Z. Jing and J. Zhan, Adv. Mater., 2008, 20, 4547–4551.
11 H. Men, P. Gao, B. Zhou, Y. Chen, C. Zhu, G. Xiao, L. Wang 33 J. Xu, Y. Li, H. Huang, Y. Zhu, Z. Wang, Z. Xie, X. Wang,
and M. Zhang, Chem. Commun., 2010, 46, 7581–7583. D. Chen and G. Shen, J. Mater. Chem., 2011, 21, 19086–19092.
12 D. Bao, P. Gao, L. Wang, Y. Wang, Y. Chen, G. Chen, G. Li, 34 T. Santhaveesuk, D. Wongratanaphisan and S. Choopun,
C. Chang and W. Qin, ChemPlusChem, 2013, 78, 1266–1272. Sens. Actuators, B, 2010, 147, 502–507.
13 X. Han, X. Han, L. Li and C. Wang, New J. Chem., 2012, 36, 35 M. Iwamoto, Y. Yoda, N. Yamazoe and T. Seiyama, J. Phys.
2205–2208. Chem., 1978, 82, 2564–2570.
14 M. Breedon, M. J. S. Spencer and I. Yarovsky, Surf. Sci., 2009, 36 W. Wang, C. Zhu and Y. Cao, Int. J. Hydrogen Energy, 2010,
603, 3389–3399. 35, 1951–1956.
15 M. J. S. Spencer, K. W. J. Wong and I. Yarovsky, Mater. Chem. 37 N. I. Butkovskaya, Y. Zhao and D. W. Setser, J. Phys. Chem.,
Phys., 2010, 119, 505–514. 1994, 98, 10779–10786.
16 J. D. Prades, A. Cirera and J. R. Morante, Sens. Actuators, B, 38 G. F. Bauerfeldt, L. M. M. De Albuquerque, G. Arbilla and
2009, 142, 179–184. E. C. Da Silva, THEOCHEM, 2002, 580, 147–160.
17 W. An, X. Wu and X. C. Zeng, J. Phys. Chem. C, 2008, 112, 39 X. Liu, B. Cheng, J. Hu and H. Qin, Comput. Mater. Sci.,
5747–5755. 2013, 68, 90–94.
18 M. Breedon, M. J. S. Spencer and I. Yarovsky, J. Phys. Chem. 40 J. Yue, X. Jiang and A. Yu, J. Phys. Chem. C, 2013, 117,
C, 2010, 114, 16603–16610. 9962–9969.
19 Q. Yuan, Y.-P. Zhao, L. Li and T. Wang, J. Phys. Chem. C, 41 J. N. Muir, Y. Choi and H. Idriss, Phys. Chem. Chem. Phys.,
2009, 113, 6107–6113. 2012, 14, 11910–11919.

11480 | Phys. Chem. Chem. Phys., 2014, 16, 11471--11480 This journal is © the Owner Societies 2014