View Article Online / Journal Homepage / Table of Contents for this issue

Journal of Dynamic Article Links < C

Materials Chemistry
Cite this: J. Mater. Chem., 2012, 22, 20038
www.rsc.org/materials PAPER
Zinc-doped nickel oxide dendritic crystals with fast response and self-recovery
for ammonia detection at room temperature†
Jian Wang, Liangming Wei, Liying Zhang, Jing Zhang, Hao Wei, Chuanhai Jiang and Yafei Zhang*
Received 28th June 2012, Accepted 3rd August 2012
Published on 03 August 2012 on http://pubs.rsc.org | doi:10.1039/C2JM34192A

DOI: 10.1039/c2jm34192a
Downloaded by Shanghai Jiaotong University on 02/05/2013 06:01:51.

Nanoscale Zn-doped NiO dendritic crystals with a Christmas-tree-like structure, have been successfully
synthesized via an electrolytic approach combined with subsequent high temperature oxidation. The
trunks have lengths in the range 6–10 mm with diameters varying from 190 nm to 200 nm, and the
branches have lengths in the range 1–3 mm with diameters varying from 150 nm to 180 nm.
Microstructure characterization indicated that the solubility limit of zinc ions in the NiO lattice sites
was lower than 7 mol%. We systematically investigated the gas sensing properties of the Zn-doped NiO
sensors for NH3 gas detection at room temperature. The sensor with doped NiO dendritic crystals gave
5–8 times faster responses and 30–50 times faster recovery speeds than the sensor with pristine NiO
dendritic crystals, which is important for the practical application of this NiO sensor. The fast response
and recovery speeds could be attributed to the effective electron transfer between the doped NiO
dendritic crystal networks and the ammonia molecules. Moreover, we found that the doped NiO
sensors have excellent reproducibility, reversibility, stability and selectivity toward NH3 gas over other
organic gases. Lastly, a possible gas sensing mechanism of NiO dendritic crystal sensors was discussed.

Introduction With the development of nanoscience and nanotechnology,

Design and fabrication of novel portable sensing devices with
low cost and low power has become a very active research field
because these devices are being widely applied in defence,
homeland security, biological detection, and monitoring of
agriculture, medical, and manufacturing environments.1–5 As
typical novel sensing devices, chemical microsensors (chemical
sensors based on tiny changes in resistance in response to the
binding of analytes) are considered to be leading candidates due
to their advantages such as low power consumption, exquisite
sensitivity and fast response to the surrounding environment.6–8
The detection of gas analytes often relies on monitoring the
direct change in the electrical properties in adsorption and
desorption of analyte molecules. Many miniaturized gas sensors
based on nanostructured sensing materials such as metal oxides,
organic semiconductors, and carbon nanotubes have been
successfully obtained.6–9 However, there are a number of obsta-
cles that still need to be overcome in order to further promote the
sensing characteristics, such as the response, recovery, selectivity,
reproducibility, and stability.
Key Laboratory for Thin Film and Microfabrication of the Ministry of
Education, Research Institute of Micro/Nano Science and Technology,
School of Materials Science and Engineering, Shanghai Jiao Tong
University, Shanghai 200240, China. E-mail: yfzhang@sjtu.edu.cn; Fax:
+86 21 3420 5665; Tel: +86 21 3420 5665
† Electronic supplementary information (ESI) available:
Thermogravimetric (TG) curve of Zn-doped NiO dendritic crystals. See
DOI: 10.1039/c2jm34192a
nanostructured semiconductor metal oxides (such as WO3, CuO,
TiO2, SnO2, ZnO, and In2O3) have attracted considerable
attention from part of the gas-sensing community due to their
obvious advantages, such as easy availability, nontoxic nature,
and biological and chemical stability.10–15 Among these nano-
structured semiconductor metal oxides, nickel oxide (NiO)
nanomaterials, which are natural p-type semiconductors with a
wide band gap energy in the range of 3.6–4.0 eV (depending on
the crystallinity and preparation method), have been recognized
as prosperous and interesting nanomaterials for the detection of
gas analytes due to their high sensitivity, good reproducibility,
low cost, simplicity of use, low operation power consumption,
high compatibility with microelectronic processing, and envi-
ronmental friendliness.16 Since the performances of gas sensors
are highly dependent on the dimension, size, and morphology of
the sensing materials, much effort has been focused on dimension
and shape control of NiO in order to provide a large surface area
and allow for good accessibility. Over the past several years,
various morphologies of nanoscale NiO such as nanoparticles,17
nanoplates,18 hexagonal nanodisks,19 nanotubes,20 nanowires,21
concave polyhedrons,22 and nanowalls,23 have been successfully
synthesized and their gas sensing properties have been investi-
gated. Recently, one of the most significant breakthroughs in this
domain was the preparation of a three-dimensional hierarchical
structure with well controlled sizes, shapes, continuous networks,
and unique surface morphologies. Furthermore, hierarchical
nanostructures, which are high dimensional nanostructures
periodically assembled of many, low dimensional, nano-building

20038 | J. Mater. Chem., 2012, 22, 20038–20047 This journal is ª The Royal Society of Chemistry 2012

blocks, have been used to reduce the response and recovery time
by rapid gas diffusion through the abundant macro-, meso- and
nanopores.24 Apart from the design of specific nanostructures,
another method to further enhance the response/recovery
kinetics of the sensing materials is to introduce other species or
by doping with catalysts to promote a surface reaction.25,26
Nanoscale doped semiconductors have attracted intensive
interest due to their excellent electronic, magnetic, catalytic,
optical, thermal, mechanical properties as compared with their
undoped materials.27 As a typical example, innovative sensors
with doped-nanomaterials have been successfully obtained and
the sensing performance of these gas sensors has been greatly
improved.28–30 Up to date, various catalysts such as Pt, Pd, Ag,
Published on 03 August 2012 on http://pubs.rsc.org | doi:10.1039/C2JM34192A
CuO, and NiO have been loaded on the surface of semiconductor
metal oxides to improve gas sensing performances. However,
Downloaded by Shanghai Jiaotong University on 02/05/2013 06:01:51.
doping contents, distribution, and size of metallic or metal oxide
catalysts are key parameters for enhancing gas selectivity as well
as sensitivity.27 Doping metals and/or metal oxides in nanoscale
NiO dendritic crystals is considered to be an effective and simple
way to improve the gas-sensing properties (such as increasing the
sensitivity and decreasing the response and recovering time) of
pure NiO nanocrystals owing to its open structure feature with
high active surface areas, easier mass transport, and higher
stability.25,27,31
In our previous study, we developed a gas sensor with uniform
NiO nanowire arrays on a silicon wafer substrate under an
external magnetic field for sensing NH3 gas, in which promising
performance was found and was related to the average crystalline
grain size.31 In this work, we present a new and simple electrolytic
approach combined with subsequent high temperature oxidation
to fabricate Zn-doped NiO dendritic crystals with well-controlled
morphologies. We then systematically investigate the Zn-doping
effects on the sensing properties in the NH3 sensors at room
temperature. The results showed that the sensitivity of the gas
sensor devices to NH3 gas can be remarkably improved (fast
response and recovery speed) by doping Zn in nanoscale NiO
dendritic crystals as compared with pristine NiO dendritic crys-
tals. Therefore, this may provide a general and simple way to
fabricate p–n integrated semiconductor gas sensors with desir-
able gas-sensing properties via doping.
Experimental
Preparation of the sensing materials
All of the reagents involved in the experiments were of analytical
reagent (AR) grade (purity >99%) and were directly used without
further purification. The strategies to fabricate the Zn-doped
NiO dendritic crystals were summarized as follows. In a typical
synthesis, 2.38 g (0.01 mol) of nickel chloride hexahydrate
(NiCl2$6H2O, Sinopharm Chemical Reagent Co., Ltd, China)
was heated at 140
C for 30 min in a heat treatment furnace in
open air. The water was entirely lost and yellow-brown anhy-
drous nickel chloride (NiCl2) was formed. The as-prepared NiCl2
and a certain amount of zinc chloride (ZnCl2, Sinopharm
Chemical Reagent Co., Ltd, China) with different molar ratios
(ZnCl2 : NiCl2 ¼ 0 : 100, 1 : 100, 3 : 100, 5 : 100, 7 : 100, 9 : 100)
were then dissolved into 200 mL of ethylene glycol (EG,
Shanghai Chemical Reagents Company, China) under
This journal is ª The Royal Society of Chemistry 2012
View Article Online
continuous magnetic stirring at room temperature. A homoge-
neous grass-green color was observed in the solution after
continuous magnetic stirring for half an hour. Ultrasound
treatment was carried out for an hour to ensure that Ni2+ ions
and Zn2+ ions were dispersed homogeneously in the electrolytic
solution, which was subsequently transferred into an electrolytic
cell. The specific electrolytic process was carried out in a home-
made two-electrode cell at 60
C and 75 V for about 4 h in the EG
solution containing NiCl2, ZnCl2 and two graphite plates as the
anode and cathode, respectively. The distance between the two
electrodes was about 10 cm, and the area of the electrode dipped
into the solution was fixed to be 30 cm.2 The applied voltage was
provided and controlled with the constant voltage mode by a
direct-current (DC) voltage-stabilized power supply. The elec-
trolytic process was performed in a stationary electrolyte solu-
tion without any stirring or protective gas bubbling. After
electrolysis was completed, the resulting grey-black fluffy solid
products were adsorbed on the cathode and the electrolytic
solution became colorless and transparent. The products were
collected and washed three times using distilled water and
ethanol with the assistance of a permanent magnet. Following
that, the products were dried in a vacuum oven at 80
C for 12 h.
The as-prepared products were heated from room temperature to
450
C and kept at this temperature for 8 h in a heat treatment
furnace in open air. After the furnace was cooled down to room
temperature naturally, a gray-green product was obtained on the
substrates. The resultant samples were then collected for char-
acterization. Fig. 1(A1) and (A2) shows the schematic drawing of
the specific method of the synthesis of Zn-doped NiO dendritic
crystals.
Preparation of the gas sensors
The Si–SiO2 wafer substrates, which were n-doped Si (100)
wafers with a 300 nm SiO2 layer on the surface, were rinsed
ultrasonically with acetone and distilled water for 20 min each
and dried in air atmosphere. The as-prepared dendritic crystals
were ultrasonically dispersed in distilled water for about 1 hour
to obtain well-dispersed suspensions. The pre-treated Si–SiO2
wafer substrate was then immersed in the dendritic-crystal–
water suspensions for 30 min to form dendritic crystal networks,
and then rinsed by ethanol and distilled water repeatedly, and
blown dry with N2. Next, the wafer substrate deposited with
dendritic crystals was heated at 450
C for 8 h in a heat treat-
ment furnace in open air. The dendritic crystal gas sensor was
fabricated using a conventional microfabrication procedure.
The interdigitated electrodes were deposited with the use a
patterned photoresist mold. The interdigitated electrode fingers
were made by sputtering 30 nm Ti and 180 nm Au onto this
patterned photoresist mold in a magnetron sputtering system. A
lift-off technique was further carried out to remove the photo-
resist. The resultant electrodes were sonicated in ethanol,
washed thoroughly with distilled water and then dried with N2.
Finally, heat treatment at 300
C in the vacuum oven for 2 h
was executed in order to optimize the contact between dendritic
crystal networks and the Au electrodes. Fig. 1(B1)–(B3) show
the photographs of sensor devices bounded on the sockets and
schematic diagram of structure of the gas sensor device,
respectively.
J. Mater. Chem., 2012, 22, 20038–20047 | 20039
 View Article Online
Published on 03 August 2012 on http://pubs.rsc.org | doi:10.1039/C2JM34192A
Downloaded by Shanghai Jiaotong University on 02/05/2013 06:01:51.

Fig. 1 (A1) and (A2) Schematic diagrams show the specific method of the synthesis of Zn-doped NiO dendritic crystals, which contained two steps. The
Ni(1x)Znx dendritic crystals were synthesized by an electrolytic process in a homemade two-electrode cell at 60
C and 75 V for about 4 h in the EG
solution containing Ni2+, Zn2+ and two graphite plates as the anode and cathode, respectively. On the cathode, Ni2+ ions and Zn2+ ions were reduced by
gaining electrons from the cathode, producing Ni(1x)Znx around the cathode. The second step was the high temperature heat treatment of as-
synthesized Ni(1x)Znx dendritic crystals at 450
C for 8 h to obtain Ni(1x)ZnxO dendritic crystals. (B1) Photograph of five sensor devices bound to the
sockets. (B2) Magnified image of a single sensor device bound to the socket. (B3) Schematic illustration of structure of the gas sensor device with
assembled Zn-doped NiO dendritic crystal networks on the Si–SiO2 wafer substrate.

Characterization relative humidity (RH) of the environment during the experi-

The crystalline phases and purity of the as-prepared samples
were tested by X-ray powder diffraction (XRD) using a
advanced X-ray diffractometer (D8 ADVANCE, Bruker, Ger-
many) in a 2 theta range from 30 to 90 degree with Cu Ka (l ¼
0.154056 nm) rotating anode point source operating at voltage of
40 kV and current of 40 mA. The morphology and elemental
composition measurements were examined with an emission
scanning electron microscope (SEM, Zeiss Ultra 55, Germany)
combined with energy dispersive spectroscopy (EDS) at accel-
erating voltage of 5 kV and 20 kV, respectively. SEM samples
were diluted about 20 times (by volume) with ethanol and then
deposited on silicon substrates. In addition, the microstructure of
the samples was studied by transmission electron microscopy
(TEM, JEM-2100, JEOL, Japan). Samples for TEM were diluted
about 50 times (by volume) with ethanol and then dropped on
copper grids and allowed to dry in a fume hood. High-resolution
transmission electron microscopy (HRTEM) images and selected
area electron diffraction (SAED) patterns were recorded with
a JEM-2010 transmission electron microscope operating at
200 kV.
Sensor testing system
Gas sensitivity tests were performed by means of a homemade
sensor testing system (see Fig. 2) at room temperature. The
ments was in about 25%. To generate and dilute the NH3 gas, the
air flowed through a glass tube containing ammonia water to
form NH3 gas. The flow ratio of NH3 gas was controlled by mass
flow controllers (MFC, purchased from Beijing Qixing Co., Ltd,
China) from 1 to 30 mL min1. Then this NH3 gas was mixed
with the diluting air in a stainless steel mixing chamber, and the
output flow ratio of the diluted NH3 gas was controlled by MFC.
Different concentrations of NH3 gas were produced by modu-
lating the flow ratio of the dilution gas to that of the flow rate the
carrier gas (air). The NH3 gas was delivered at a flow rate of 2 L
min1 into the testing chamber to test the sensor performance,
and the surplus gas flow was introduced into the special
absorber. All the gas sensing measurements were carried out at
room temperature (300 K). The electrical signal of the gas sensor
was monitored using a precision semiconductor parameter
analyzer (Agilent 4156C) and recorded as a function of the
operating time. In the test process, the sensor response was
evaluated by the resistance change at a working voltage of 1 V.
Results and discussion
Structure and morphology analysis
Nanoscale Zn-doped NiO dendritic crystals with distinctive
morphology were synthesized via an electrolytic approach
combined with subsequent high temperature oxidation, as shown

20040 | J. Mater. Chem., 2012, 22, 20038–20047 This journal is ª The Royal Society of Chemistry 2012

Published on 03 August 2012 on http://pubs.rsc.org | doi:10.1039/C2JM34192A
Downloaded by Shanghai Jiaotong University on 02/05/2013 06:01:51.
Fig. 2 Schematic diagram shows the gas sensor testing system.
in Fig. 1(A1) and (A2). Firstly, a controlled electrolytic synthesis
route was developed to prepare Ni(1x)Znx dendritic crystals in
the homemade two-electrode cell. On the cathode, Ni2+ ions and
Zn2+ ions were reduced by gaining electrons from the cathode,
producing Ni(1x)Znx dendritic crystals around the cathode.
Meanwhile, Cl ions were oxidized to chlorine atoms via losing
electrons, and two chlorine atoms combined to a chlorine gas
molecule and released from the anode.32,33 The second step was
the high temperature heat treatment of as-synthesized Ni(1x)Znx
dendritic crystals at 450
C for 8 h to obtain Ni(1x)ZnxO
dendritic crystals in open air. The specific chemical reactions for
the synthesis can be expressed as follows:
(1  x)Ni2+ + xZn2+ + 4e / Ni1xZnx (1)
2Cl  2e / Cl2 (2)
2Ni1xZnx + O2 / 2Ni1xZnxO (3)
In order to understand the transformation process from Zn-
doped Ni dendritic crystals to Zn-doped NiO, we investigated the
thermogravimetric analysis (TGA) of the as-prepared sample
under air flow. From Fig. S1,† we can observe that the Zn-doped
Ni crystals begin to be oxidized around 386
C. The thermog-
ravimetric experiment of constant temperature (450
C, Fig. S2†)
indicates that the oxidation process from Zn-doped Ni dendritic
crystals to Zn-doped NiO would last for around 6.5 h at 450
C.
Therefore, in order to obtain pure oxide, we prepare the sensing
material of Zn-doped NiO nanostructure at 450
C for 8 h.
The XRD patterns of the as-synthesized gray-green products
with and without the introduction of zinc ions are similar.
Fig. 3(A) shows the typical XRD patterns of the 3 mol% Zn-
doped NiO dendritic crystals (after oxidation). All the diffraction
peaks in this pattern were found to match with pure NiO phase
having cubic structure, which is consistent with the value given in
the standard card (PDF standard card, Joint Committee on
Powder Diffraction Standards (JCPDS) card no. 49-1049, space
group, fm3m). These peaks at the scattering angles (2q) corre-
spond to the crystal planes (111), (200), (220), (311) and (222) of
This journal is ª The Royal Society of Chemistry 2012
View Article Online
crystalline NiO. From the XRD patterns, no other characteristic
peaks, such as nickel, zinc or zinc oxide, can be detected, indi-
cating that a single cubic structure was obtained using this
electrolytic approach combined with subsequent high tempera-
ture oxidation. In order to understand the surface elements
present in the Zn-doped NiO dendritic crystals, further energy
dispersive spectroscopy (EDS) analysis (see Fig. 3(B)) revealed
the peaks of strong Ni, O and weak Zn. A similar atomic
percentage of Ni, O and Zn was observed at different detection
points and the average value (the concentration of Zn) was about
2.9%, which is consistent with the Zn doping concentration (0.3
mol%). No other impurity peaks were detected in the EDS, which
further confirms that the doped materials are composed of only
Ni, O, and Zn. Fig. 3(C) presents a scanning electron microscopy
(SEM) image of the as-prepared 3 mol% Zn-doped NiO dendritic
crystals with an almost uniform distribution. The dendritic
crystals consist almost entirely of a hierarchical structure
sprawling to several generations. Meanwhile, Fig. 3(D) clearly
demonstrates the morphology of a single dendrite, which reveals
a well-defined dendritic structure with a long main trunk and
highly ordered short side branches distributed on both sides of
the trunk. The trunks have lengths in the range 6–10 mm with
diameters varying from 190 nm to 200 nm, and the branches have
lengths in the range 1–3 mm with diameters varying from 150 nm
to 180 nm. In order to understand the structure of the dendritic
crystals in more detail, further transmission electron microscopy
(TEM) observations at higher magnification (Fig. 3(E)) shows
that these well-aligned branches radiate from the central trunk
and exhibit good symmetry and a highly ordered structure. As
demonstrated in Fig. 3(E), the secondary and tertiary branches
of dendritic crystals consisted of tiny single crystals with grain
sizes 60 to 100 nm. The corresponding selected area electron
diffraction (SAED) pattern of the Zn-doped NiO dendritic
crystal recorded from a single doped dendritic structure, as seen
in Fig. 3(F), confirms that the dendritic crystals have a cubic
structure and there is no formation of a Zn or ZnO precipitated
secondary phase. The diffraction rings in the electron diffraction
pattern can readily be indexed as (111), (200), (220), (331) and
(222) of the cubic NiO. The typical high-resolution TEM
J. Mater. Chem., 2012, 22, 20038–20047 | 20041
 View Article Online
Published on 03 August 2012 on http://pubs.rsc.org | doi:10.1039/C2JM34192A
Downloaded by Shanghai Jiaotong University on 02/05/2013 06:01:51.

Fig. 3 Structure and morphology characterization of 3 mol% Zn-doped NiO dendritic crystals synthesized via an electrolytic approach combined with
subsequent high temperature oxidation. (A) XRD pattern of the sample confirming formation of cubic structure NiO. (B) EDS analysis showing that the
doped materials are composed of only Ni, O and Zn. (C) SEM image of Zn-doped NiO dendritic crystals showing the high yield and good uniformity.
(D) TEM image of a single dendritic crystal showing the dendritic structure and good symmetry structure. (E) High-magnification TEM image of
dendritic crystal demonstrating the secondary and tertiary branches of dendritic crystals consisted of tiny single crystals. HRTEM images recorded from
the tertiary and secondary branches are shown in figure (E1), (E2) and (E3), respectively. (F) SAED pattern recorded from a single doped dendritic
structure.

(HRTEM) images recorded from the tertiary and secondary
branches in 3 mol% Zn-doped NiO crystals are shown in
Fig. 3(E1)–(E3), respectively, as indicated by the symbols of E1,
E2 and E3 within the dashed blue circles in Fig. 3(E). They show
the crystal lattice structures at the tip of the top branch, the tip of
the bottom branch, and the tip of the secondary lattice. The
images recorded from the secondary and the tertiary branch of the
doped NiO dendritic crystal clearly show the high crystallinity of
the dendritic crystal with no visible line or planar defects. The
inter-planar distance of the fringes in Fig. 3(E1) and (E3) was
measured to be 0.247 nm, which is slightly larger than that for the
{111} planes of intrinsic cubic NiO crystal (0.24120 nm). At the
same time, the interspacing in Fig. 3(E2) was determined to be
0.213 nm, which is also slightly larger than that for the distance of
the intrinsic NiO {200} planes (0.20890 nm). These results indicate
lattice expansion, which implied that Zn atoms might form
substitution atoms in NiO dendritic crystals to a great extent.
The evolution of the surface morphology, surface element and
crystalline phase of the Zn-doped NiO dendritic crystals in
different Zn doping concentrations have been investigated and
are shown in Fig. 4. Fig. 4(A1)–(F1) show SEM images of
as-synthesized Zn-doped NiO dendritic crystals with 0 mol%,
1 mol%, 3 mol%, 5 mol%, 7 mol%, and 9 mol% of zinc doping
concentration (referred to as ‘pure NiO’, ‘1 mol% Zn–NiO’,
‘3 mol% Zn–NiO’, ‘5 mol% Zn–NiO’, ‘7 mol% Zn–NiO’, ‘9 mol%
Zn–NiO’, respectively). The dendritic structures remained
similar after the NiO dendritic crystals were doped with different
concentrations of Zn, which indicated that the Zn dopant (at 1–9
mol% doping concentration range) has no influence on the
growth of the dendritic crystals. The point-EDS spectral micro-
analysis from a series of locations along a certain dendritic
crystal was labeled with 1, 2 and 3 (the solid red circles) in
Fig. 4(A1)–(F1), and the quantitative elemental distribution in
atomic percentage is tabulated in Fig. 4(A2)–(F2). 0.95 at.%, 2.88
at.%, 4.96 at.%, 6.93 at.% and 8.89 at.% Zn were measured, as
shown in point 1 of Fig. 4(B2)–(F2), which is consistent with the
Zn ions doping concentration (1 mol%, 3 mol%, 5 mol%, 7 mol
%, and 9 mol%), respectively. Similar Zn contents were observed

20042 | J. Mater. Chem., 2012, 22, 20038–20047 This journal is ª The Royal Society of Chemistry 2012
 View Article Online
Published on 03 August 2012 on http://pubs.rsc.org | doi:10.1039/C2JM34192A
Downloaded by Shanghai Jiaotong University on 02/05/2013 06:01:51.

Fig. 4 Evolution of the surface morphology, element and crystalline phase of the Zn-doped NiO dendritic crystals in different Zn doping concen-
trations. (A1)–(F1) SEM images of doped NiO dendritic crystals with 0 mol%, 1 mol%, 3 mol%, 5 mol%, 7 mol%, and 9 mol% of Zn doping
concentration. EDS spectral analysis from a series of locations along a certain dendritic crystal was labeled with 1, 2 and 3 in figure (A1)–(F1), and the
quantitative elemental distribution in atomic percentage is tabulated in (A2)–(F2). (A3)–(F3) XRD patterns of the samples with various levels of zinc
doping.

at various points along the entire dendritic structure, such as
point 2 and point 3 of Fig. 4(B2)–(F2). The corresponding XRD
patterns of the six samples with various levels of Zn doping are
presented in Fig. 4(A3)–(F3). In ‘1 mol% Zn–NiO’, ‘3 mol% Zn–
NiO’ and ‘5 mol% Zn–NiO’ (Fig. 4(B3)–(D3)), the observed
diffraction peaks can be indexed to a standard cubic NiO
structure with no visible defects, secondary phases or impurity
peaks. This suggests that the dopants are well-integrated into the
NiO lattice sites during the synthesis process. However, all the
peaks have a slight shift toward lower angles with increasing Zn
dopant concentration compared with pure NiO dendritic crys-
tals, which is assigned to the successful incorporation of dopant
ions in the NiO host structure. The larger ionic radius of the zinc
ion (0.074 nm) compared with the nickel ion (0.069 nm), tends to
increase the size of the lattice in doped NiO dendritic crystals,
which is consistent with the HRTEM observation in Fig. 3(E1)–
(E3). In ‘7 mol% and 9 mol% Zn–NiO’ (Fig. 4(E3) and (F3)),
several weak additional diffraction peaks appeared, which are
assigned to the secondary phase of ZnO, but we could not
observe any metallic Zn phase, which may be because Zn easily
oxidizes to ZnO during high temperature heat treatment in open
air. We also tried to increase the amount of zinc ions to 11 mol%,
13 mol% and 15 mol% under the same growth conditions, and
similar additional diffraction peaks with Fig. 4(E3) and (F3)
were observed. Therefore, the solubility limit of zinc ions in NiO
was determined to be lower than 7% in this work.
Gas sensing properties
Fig. 1(B1)–(B3) show the real picture and schematic diagram of
the as-prepared gas sensor device with assembled Zn-doped NiO
dendritic crystal networks on the Si–SiO2 wafer substrate,
respectively. In this gas sensor device, the doped NiO dendritic
crystals were used as conducting channels in the interdigitated
gold electrodes, and the interdigitated arrays of electrodes
possess a finger length of 600 mm and a gap size of 10 mm. A

This journal is ª The Royal Society of Chemistry 2012 J. Mater. Chem., 2012, 22, 20038–20047 | 20043

random network of doped NiO dendritic crystals with uniform
distribution was formed between these interdigitated electrodes
by the deposition technique, which ensure the dendritic crystals
and interdigitated electrodes were highly connected to each
other. The conductance between Au electrodes was measured to
investigate the sensory response. In our experiments, the sensor
response (R) is defined by equation:
R ¼ DG/G0 ¼ (G  G0)/G0 (4)
Where G0 and G are the conductance of semiconducting
dendritic crystals in air and testing gas, respectively.
Fig. 5(A1)–(A6) show the gas response of the NiO based gas
Published on 03 August 2012 on http://pubs.rsc.org | doi:10.1039/C2JM34192A
sensor devices including a pure NiO dendritic crystal and the five
Zn-doped NiO dendritic crystals with different concentrations of
Downloaded by Shanghai Jiaotong University on 02/05/2013 06:01:51.
Zn, when exposed to 30 ppm NH3 at room temperature. There
were no significant changes in the conductance (about 19%) after
doping the NiO dendritic crystals with different concentrations
of Zn. Meanwhile, the conductance change in six sensing cycle
experiments with different concentrations of NH3 gas from 5
ppm to 150 ppm was investigated, as seen in Fig. 5(B1). Obvi-
ously, the change of conductance increased with the increasing
NH3 gas concentrations, but the change in conductance after
doping the NiO dendritic crystals with different concentrations
of Zn, remained similar for a certain amount of NH3 gas, which
indicated that the Zn dopant has no remarkable influence on the
sensitivity (response intensity) of the sensor device. This
phenomenon may be because the six samples have a similar grain
size and number of grain boundaries (active sites for testing gas
sorption) after the same heat treatment temperature. Previous
reports have demonstrated that the grain boundaries or grain
junctions were considered to be active sites for gas molecule
absorption,27 and affect electron transport through the dendritic
crystal networks, resulting in conductance change.
In the undoped NiO dendritic crystal sensor, the response time
(sres ¼ 40 s, where sres is the time to reach a 90% variation in
conductance upon exposure to testing gas) and recovery time
(srecov ¼ 1500 s, where srecov is the time to reach a 90% variation
in conductance upon exposure to air) are very long upon expo-
sure to 30 ppm NH3 gas, which is not favorable for the NiO
dendritic sensor practical applications. This slow response and
recovery speed can be attributed the slow surface reactions
consisting of the adsorption, dissociation, ionization and diffu-
sion of gas molecules. In contrast, the response time (sres ¼ 8–5 s)
and recovery time (srecov ¼ 50–30 s) of the gas sensor became very
short when doping with 1, 3, 5, 7 or 9 mol% of Zn, as shown in
Fig. 5(A2)–(A6). This means that the response speed was
increased by about 5–8 times and the recovery speed was
increased by about 30–50 times by doping Zn in NiO. This fast
response speed demonstrates that both the in-diffusion of NH3
gas to the surface of Zn doped NiO and its electrochemical
oxidation by reaction with negatively charged oxygen species
such as O2 and O, occur very quickly. Additionally, this fast
recovery speed suggests that the doping Zn plays an important
role in shortening the recovery reaction involving the dissocia-
tion, ionization, diffusion and desorption of the oxygen species
and ammonia molecules in reference gas (air) blowing.25
The response and recovery speed showed slightly different
behavior for NiO dendritic crystals doped with different
20044 | J. Mater. Chem., 2012, 22, 20038–20047
View Article Online
concentrations of Zn. For example, the response and recovery
time of the ‘1 mol% Zn–NiO’ sensor was 8 s and 50 s upon
exposure to 30 ppm NH3 gas (Fig. 5(A2)), respectively. However,
the response and recovery time of the ‘5 mol% Zn–NiO’ sensor
decreased to 5 s and 30 s under the same conditions (Fig. 5(A4)),
respectively. This shows that an increase in the concentration of
Zn speeds up the response and recovery speed. When the
concentration of Zn in doped NiO surpassed the solubility limit
(5 at.%) of zinc ions in NiO, the response and recovery speed
remains constant, such as sres ¼ 6 s and srecov ¼ 35 s in ‘7 mol%
Zn–NiO’ (Fig. 5(A5)), sres ¼ 5 s and srecov ¼ 30 s in ‘9 mol% Zn–
NiO’ (Fig. 5(A6)). This shows that an additional sensing mech-
anism involving the concentration of Zn should be taken into
account in doped NiO dendritic crystals.
The response of the doped NiO sensors was further examined
upon consecutive exposure to different concentrations of NH3
gas (Fig. 5(B1)–(B3)). The results show that the doped NiO gas
sensor exhibits fast response and recovery to the presence of NH3
gas at room temperature. For example, when the sensor was
exposed to 5 ppm of NH3 gas, the response time is only about 2 s.
Additionally, we found that after one sensing cycle experiment, it
takes less than 8 s to completely recover when the reference gas
was blown over the sensor, which is important for practical
applications of this sensor. This fast response and recovery is
partly due to the good interconnection between the Zn-doped
NiO dendritic crystals and the Au electrodes.34 More impor-
tantly, the Zn dopant dramatically improves the response and
recovery speed.
To further investigate the reproducibility of the sensor, the ‘5
mol% Zn–NiO’ gas sensor was repeatedly exposed to 50 ppm
NH3 gas at room temperature, as seen in Fig. 5(C). The change in
conductance was maintained, and the recovery abilities were not
reduced after four sensing cycles, which indicated that the
response of the Zn-doped NiO dendritic crystal sensor has
excellent reproducibility and reversibility.
The responses of doped NiO sensors towards a variety of
flammable, toxic and corrosive gases including chloroform,
toluene, ethanol, formaldehyde, acetone, dichloromethane, and
ammonia were explored to evaluate their selectivity. Fig. 5(D)
shows that the gas sensor of ‘5 mol% Zn–NiO’ exhibited excellent
selectivity to NH3 gas when exposed to those interfering vapors.
For instance, the conductance change of the sensors upon
exposure to 50 ppm of NH3 gas was about 40%, whereas that
change was less than 8% upon exposure to 1000 ppm of other
gases (such as chloroform, toluene, ethanol, formaldehyde,
acetone, dichloromethane). There are two reasons for the fact
that the doped NiO sensors have excellent selectivity for NH3
gas. On the one hand, the electron donating ability of ammonia is
very strong owing to a lone electron pair in ammonia (ammonia
can readily donate the unpaired electrons), and the electrons
could be transferred from ammonia to the doped oxide, resulting
in a significant change in the conductance of sensor.35
On the other hand, selective physisorption to ammonia is also
a key factor.36 The large active surface areas of the doped NiO
dendritic structures available for direct sorption of ammonia
molecules exist on the dendritic networks. This selective phys-
isorption to ammonia affect electron transport through the
dendritic crystal networks, resulting in excellent selectivity for
NH3 gas.
This journal is ª The Royal Society of Chemistry 2012
 View Article Online
Published on 03 August 2012 on http://pubs.rsc.org | doi:10.1039/C2JM34192A
Downloaded by Shanghai Jiaotong University on 02/05/2013 06:01:51.

Fig. 5 Sensing performance of the gas sensors based Zn-doped NiO dendritic crystals. The conductance change of ‘pure NiO’ (A1), ‘1 mol% Zn–NiO’
(A2), ‘3 mol% Zn–NiO’ (A3), ‘5 mol% Zn–NiO’ (A4), ‘7 mol% Zn–NiO’ (A5) and ‘9 mol% Zn–NiO’ (A6) dendritic crystals to 30 ppm NH3 gas at room
temperature; the conductance change (B1), response time (B2) and recovery time (B3) of the sensors upon exposure to different concentrations of NH3
gas; (C) reproducibility of the Zn-doped NiO gas sensor to 50 ppm NH3 gas; (D) the conductance change of the Zn-doped NiO gas sensor upon exposure
to NH3 gas and other organic gas.

Gas sensing mechanism sensor to air, oxygen molecules were absorbed on the active sites
of doped NiO surface and electron transfer occurs between the
The NH3 gas sensing mechanism of Zn-doped NiO dendritic
p-type NiO surface and the adsorbates, which affect the carrier
crystal sensors is based on the conductance change of the
density. The surface of the nanoscale NiO is sensitive to both
semiconductor nanomaterials upon interaction with target gas
oxidizing and reducing target gas molecules because of the
molecules (the adsorption and desorption of the gas molecules
presence of adsorbed oxygen ions on the outer layer of the
on the surface of the sensing materials). After exposing the
sensing materials.37

This journal is ª The Royal Society of Chemistry 2012 J. Mater. Chem., 2012, 22, 20038–20047 | 20045

Fig. 6(A1)–(A3) show the reaction for NH3 gas absorbed on
the undoped NiO dendritic crystal surface. It is known that the
as-synthesized undoped NiO nanomaterials are p-type semi-
conductors. As the p-type NiO is exposed to open air, the
atmospheric oxygen molecules adsorbed on the oxide surface
capture a certain amount of free electrons from the pure NiO,
and form various oxygen species such as O2 and O on the
undoped NiO surface. This causes the carrier (hole) concentra-
tion to increase on the surface layer and leading to the low
resistance of the sensor due to lowering of potential barrier
(Fig. 6(A1) and (A2)).38 When the undoped NiO sensor is
exposed to NH3 gas, the ammonia molecules can catalytically
react with the adsorbed oxygen, and the freeing electrons
Published on 03 August 2012 on http://pubs.rsc.org | doi:10.1039/C2JM34192A
captured by oxygen are returned to the oxide surface layer, and
then the conduction channel becomes narrower (Fig. 6(A3)).27
Downloaded by Shanghai Jiaotong University on 02/05/2013 06:01:51.
Consequently, the resistance of the sensor increases and hence
the conductivity reduces. Therefore, the oxygen adsorption for
gas sensor device demonstrates an important responsibility in the
gas sensing properties of the NiO dendritic crystals. Simulta-
neously, if the electron density is high at the sensor surface layer,
the speed of adsorbed oxygen becomes very fast. In doped NiO
hierarchical structures, the dopant Zn may be oxidized to n-type
ZnO after high temperature heat treatment in open air. There-
fore, the Zn-doped NiO dendritic crystals have more electrons
compared to undoped NiO dendritic crystals owing to the exis-
tence of many tiny ZnO nanocrystallites. The nanocrystallites
lead to quick adsorption of oxygen adsorbates on the surface of
Fig. 6 Schematic illustration for the gas sensing mechanism of the undoped (A1, A2, A3) and Zn-doped NiO (B1, B2, B3) dendritic crystal sensors to
NH3 gas at room temperature.
20046 | J. Mater. Chem., 2012, 22, 20038–20047
View Article Online
doped NiO, thus speeding up the response speed of this sensor, as
shown in Fig. 6(B1) and (B2). It is well known that NH3 mole-
cules can easily react with oxygen adsorbates, and the electrons
bonded with adsorbed oxygen are returned back to the sensor
(Fig. 6(B3)). Finally, the sensor conductivity is changed.
According to this speculation, the response and recovery speed to
NH3 should be enhanced through doping Zn in NiO dendritic
crystals, which is in accordance with the observation in upper
experiments. For example, in the undoped NiO dendritic crystal
sensor (Fig. 5(A1)), the response time (sres ¼ 40 s) and recovery
time (srecov ¼ 1500 s) are very long upon exposure to 30 ppm NH3
gas. However, the response time (sres ¼ 8–5 s) and recovery time
(srecov ¼ 50–30 s) of the gas sensor became very short when
doping Zn, as shown in Fig. 5(A2)–(A6).
Conclusions
In summary, the present study has demonstrated an electrolytic
approach and subsequent high temperature oxidation method to
the large-scale synthesis of nanoscale Zn-doped NiO dendritic
crystals which consist of a long main trunk and highly ordered
secondary and tertiary branches. SEM and TEM observations
reveal that the trunks have lengths in the range 6–10 mm with
diameters varying from 190 nm to 200 nm, and the branches have
lengths in the range 1–3 mm with diameters varying from 150 nm
to 180 nm. Combined XRD, EDS, SAED and HRTEM show
that a Zn-doped NiO sample of single cubic structure can be
This journal is ª The Royal Society of Chemistry 2012

obtained when the dopant concentration of zinc ions in NiO
lattice sites was lower than 7 mol%. Moreover, the response
speed was increased by about 5–8 times and the recovery speed
was increased by about 30–50 times by doping Zn in NiO
dendritic crystals as compared with pristine NiO dendritic crys-
tals upon exposure to 30 ppm NH3 gas at room temperature.
These times are favorable for the practical application of this
NiO sensor. Meanwhile, we found that the doped NiO sensors
have excellent reproducibility, reversibility, stability and selec-
tivity toward NH3 gas over other organic gases. It is suggested
that the approach demonstrated here can also be extended to
other element-doped metal oxide gas sensors with desirable gas-
sensing properties, which realized gas detection by metal oxide
Published on 03 August 2012 on http://pubs.rsc.org | doi:10.1039/C2JM34192A
sensors at room temperature.
Downloaded by Shanghai Jiaotong University on 02/05/2013 06:01:51.
Acknowledgements
We thank for the financial support from National High-Tech
R & D Program of China (863, no. 2011AA050504), Natural
Science Foundation of shanghai (no. 10ZR1416300), key labo-
ratory foundation, the Foundation for SMC Excellent Young
Teacher and the Analytical and Testing Center in Shanghai Jiao
Tong University.
Notes and references
1 D. R. Kauffman and A. Star, Chem. Soc. Rev., 2008, 37, 1197.
2 E. S. Snow, F. K. Perkins and J. A. Robinson, Chem. Soc. Rev., 2006,
35, 790.
3 D. R. Kauffman and A. Star, Angew. Chem., Int. Ed., 2008, 47, 6550.
4 T. Kida, T. Doi and K. Shimanoe, Chem. Mater., 2010, 22, 2662.
5 J. Yan, H. J. Zhou, P. Yu, L. Su and L. Q. Mao, Adv. Mater., 2008,
20, 2899.
6 F. Wang, H. W. Gu and T. M. Swager, J. Am. Chem. Soc., 2008, 130,
5392.
7 L. T. Kong, J. Wang, X. C. Fu, Y. Zhong, F. L. Meng, T. Luo and
J. H. Liu, Carbon, 2010, 48, 1262.
8 P.-C. Chen, S. Sukcharoenchoke, K. Ryu, L. G. Arco, A. Badmaev,
C. Wang and C. W. Zhou, Adv. Mater., 2010, 22, 1900.
9 L. T. Kong, J. Wang, T. Luo, F. L. Meng, X. Chen, M. Q. Li and
J. H. Liu, Analyst, 2010, 135, 368.
10 J. Polleux, A. Gurlo, N. Barsan, U. Weimar, M. Antonietti and
M. Niederberger, Angew. Chem., Int. Ed., 2006, 45, 261.
11 J. J. Chen, K. Wang, L. Hartman and W. L. Zhou, J. Phys. Chem. C,
2008, 112, 16017.
This journal is ª The Royal Society of Chemistry 2012
View Article Online
12 L. Francioso, A. M. Taurino, A. Forleo and P. Siciliano, Sens.
Actuators, B, 2008, 130, 70.
13 C. Nayral, T. Ould-Ely, A. Maisonnat, B. Chaudret, P. Fau,
L. Lescouz
eres and A. Peyre-Lavigne, Adv. Mater., 1999, 11, 61.
14 J. Zhang, S. R. Wang, M. J. Xu, Y. Wang, B. L. Zhu, S. M. Zhang,
W. P. Huang and S. H. Wu, Cryst. Growth Des., 2009, 9, 3552.
15 Y. D. Zhang, Z. Zheng and F. L. Yang, Ind. Eng. Chem. Res., 2010,
49, 3539.
16 Z. P. Wei, M. Arredondo, H. Y. Peng, Z. Zhang, D. L. Guo,
G. Z. Xing, Y. F. Li, L. M. Wong, S. J. Wang, N. Valanoor and
T. Wu, ACS Nano, 2010, 4, 4785.
17 X. L. Zhang, F. Z. Huang, A. Nattestad, K. Wang, D. C. Fu,
A. Mishra, P. B€auerle, U. Bach and Y.-B. Cheng, Chem. Commun.,
2011, 47, 4808.
18 Y. G. Li, B. Tan and Y. Y. Wu, Chem. Mater., 2008, 20, 567.
19 Z. Zheng, L. Huang, Y. Zhou, X. W. Hu and X. M. Ni, Solid State
Sci., 2009, 11, 1439.
20 Y. Ren, W. K. Chim, S. Y. Chiam, J. Q. Huang, C. Pi and J. S. Pan,
Adv. Funct. Mater., 2010, 20, 3336.
21 B. Liu, H. Q. Yang, H. Zhao, L. J. An, L. H. Zhang, R. Y. Shi,
L. Wang, L. Bao and Y. Chen, Sens. Actuators, B, 2011, 156, 251.
22 W. Zhou, M. Yao, L. Guo, Y. M. Li, J. H. Li and S. H. Yang, J. Am.
Chem. Soc., 2009, 131, 2959.
23 K. L. Zhang, C. Rossi1, P. Alphonse and C. Tenailleau,
Nanotechnology, 2008, 19, 155605.
24 J.-H. Lee, Sens. Actuators, B, 2009, 140, 319.
25 H.-R. Kim, A. Haensch, I.-D. Kim, N. Barsan, U. Weimar and
J.-H. Lee, Adv. Funct. Mater., 2011, 21, 4456.
26 G. X. Zhu, H. Xu, Y. J. Liu, X. Xu, Z. Y. Ji, X. P. Shen and Z. Xu,
Sens. Actuators, B, 2012, 166–167, 36.
27 N. G. Cho, H.-S. Woo, J.-H. Lee and I.-D. Kim, Chem. Commun.,
2011, 47, 11300.
28 M. Ahsan, T. Tesfamichael, M. Ionescu, J. Bell and N. Motta, Sens.
Actuators, B, 2012, 162, 14.
29 X. Chen, Z. Guo, W.-H. Xu, H.-B. Yao, M.-Q. Li, J.-H. Liu,
X.-J. Huang and S.-H. Yu, Adv. Funct. Mater., 2011, 21, 2049.
30 P. Sun, L. You, Y. F. Sun, N. K. Chen, X. B. Li, H. B. Sun, J. Ma and
G. Y. Lu, CrystEngComm, 2012, 14, 1701.
31 Z. Wang, Z. Li, J. Sun, H. Zhang, W. Wang, W. Zheng and C. Wang,
J. Phys. Chem. C, 2010, 114, 6100.
32 J. Wang, L. M. Wei, L. Y. Zhang, Y. F. Zhang, C. H. Jiang and
Y. F. Zhang, CrystEngComm, 2012, 14, 1629.
33 J. Wang, L. Y. Zhang, L. M. Wei, C. H. Jiang and Y. F. Zhang, Nano-
Micro Lett., 2011, 3, 264.
34 Y. Y. Wang, N. T. Hu, Z. H. Zhou, D. Xu, Z. Wang, Z. Yang,
H. Wei, E. S.-W. Kong and Y. F. Zhang, J. Mater. Chem., 2011,
21, 3779.
35 A. Hagfeldt and M. Gratzel, Chem. Rev., 1995, 95, 49.
36 Y. Liu, J. Dong, P. J. Hesketh and M. Liu, J. Mater. Chem., 2005, 15,
2316.
37 R. K. Bedi and I. Singh, ACS Appl. Mater. Interfaces, 2010, 2, 1361.
38 L. N. Han, D. J. Wang, Y. C. Lu, T. F. Jiang, B. K. Liu and Y. H. Lin,
J. Phys. Chem. C, 2011, 115, 22939.
J. Mater. Chem., 2012, 22, 20038–20047 | 20047
Electronic Supplementary Material (ESI) for Journal of Materials Chemistry
This journal is © The Royal Society of Chemistry 2012

Supporting Information

Zinc-doped nickel oxide dendritic crystals with fast response and

self-recovery for ammonia detection at room temperature

Jian Wang, Liangming Wei, Liying Zhang, Jing Zhang, Hao Wei, Chuanhai Jiang and Yafei Zhang*

Key Laboratory for Thin Film and Microfabrication of the Ministry of Education, Research Institute of Micro

/ Nano Science and Technology, School of Materials Science and Engineering, Shanghai Jiao Tong

University, Shanghai 200240, China.

Fig. S1 Thermogravimetric (TG) curve of Zn-doped NiO dendritic crystals (increasing temperature
from 30 oC to 900 oC in air).

1
Electronic Supplementary Material (ESI) for Journal of Materials Chemistry
This journal is © The Royal Society of Chemistry 2012

Fig. S2 Thermogravimetric (TG) curve of Zn-doped NiO dendritic crystals (fixed the temperature at
450 oC, increasing time from 0 to 8 h in air).