Journal ofMaterials Chemistry C

PAPER

Publ

ishe

d on

26

Apr

il 20

13. D

ownl

oade

d by

BE

IJIN

G I

NST

ITU

TE

OF

TE

CH

NO

LO

GY

on

06/1

1/20

13 1

3:35

:17.

View Article OnlineView Journal | View Issue

aKey Laboratory of Synthetic and Natural Fu

of Chemistry (Ministry of Education), Nor

China. E-mail: xglin707@163.com; Tel: +86bCollege of Chemistry & Chemical Engin

Laboratory of Chemical Reaction EngineerincCollege of Mechanical & Material Engin

Institute, China Three Gorges University,

E-mail: lidongsheng1@126.com; Tel: +86 71

† Electronic supplementary informa10.1039/c3tc30189c

Cite this: J. Mater. Chem. C, 2013, 1,4153

Received 30th January 2013Accepted 26th April 2013

DOI: 10.1039/c3tc30189c

www.rsc.org/MaterialsC

This journal is ª The Royal Society of

Synthesis of mesoporous Bi2WO6 architectures and theirgas sensitivity to ethanol†

Danjun Wang,ab Yanzhong Zhen,ab Ganglin Xue,*a Feng Fu,b Xuemei Liua

and Dongsheng Li*c

Uniform hierarchical multilayered Bi2WO6 architectures were synthesized by a facile template-free

hydrothermal process, and their synthesis conditions and formation mechanism were carefully

investigated. XRD, XPS, FE-SEM, and HR-TEM techniques were employed to determine their phase

composition, morphology and microstructure. Nitrogen adsorption and desorption isotherms were

conducted to examine the specific surface area and the pore nature of the as-prepared sample. The

results show that the as-prepared Bi2WO6 architectures consist of secondary nanoplates and have a

mesoporous nest-like morphology with a diameter of 3–4 mm, and have very large specific surface areas.

The largest surface area of 47.72 m2 g�1 is achieved when synthesized at 190 �C for 2.0 h. Furthermore,

the Bi2WO6 samples were fabricated into a gas sensor, and the experimental results showed that the

samples exhibited high sensitivity and a fast response–recovery to ethanol gas at lower temperatures

(300 �C). For 100 ppm ethanol, the sensitivity of the best sensor (GS2) was 34.6, which is about 3-fold

higher than the reported mesoporous ZnO based gas sensor because its mesoporous structure provided

a high surface-to-volume ratio and surface accessibility for the ethanol. A plausible enhancement gas

responding mechanism of the nest-like Bi2WO6 sensors was also proposed based on the structure and

response–recovery properties.

1 Introduction

In recent years, environmental pollution has become more andmore serious. There has been an increasing demand for gassensors for better environmental control of the discharge ofpollutants. Different kinds of gas sensing materials based onsemiconductor metal oxides,1,2 conducting polymers,3 andcarbon black-polymer composites,4–6 have attracted muchattention. Perovskite-type oxides, such as La0.68Pb0.32FeO3,La0.7Sr0.3FeO3 and LnFeO3 (Ln¼ La, Sm, and Eu) nanoparticles,have been widely investigated in the elds of gas sensors.7–9

Especially, mesoporous oxide nanostructures with well-alignedpore structures have been reported to show very high gasresponses and rapid gas responding kinetics10–12 which areattributed to their high surface area and well-dened porousnanostructures, respectively. The literature to date reveals that

nctional Molecule Chemistry, Department

thwest University, Xi'an, 710069, P. R.

29-88302604

eering, Yan'an University, Shaanxi Key

g, Yan'an, 716000, P. R. China

eering, Functional Materials Research

Yichang, Hubei, 443002, P. R. China.

7-6392538

tion (ESI) available. See DOI:

Chemistry 2013

mesoporous nanostructures can increase both gas response andresponse speed simultaneously and substantially because theycan provide an effective gas diffusion path via well-alignedporous structure without sacricing a high surface area.13

Ethanol is one of the most common and volatile organiccompounds, its fast and effectively detection is very importantin some elds, such as testing alcohol levels of drivers, moni-toring chemical synthesis, etc. To date, many metal oxides, suchas ZnO,14–16 Fe2O3,17 In2O3,18 and WO3,19 have been employed tofabricate ethanol gas sensors because of their sensitivity toethanol. However, much work needs to be done to improve thesensitivity of those materials and further to explore newethanol-sensitive materials. Bismuth tungstate (Bi2WO6) is oneof the Aurivillius oxide families of layered perovskites. As awidely investigated versatile material, Bi2WO6 has manyattractive and important physical properties such as ferroelec-tric piezoelectricity, pyroelectricity, catalytic property, oxideanion conduction, nonlinear dielectric susceptibility, andluminescent properties.20,21 Many studies have shown that theperformance of Bi2WO6 can be greatly inuenced by the size,morphology and structural features.22–30 To this end, manydifferent synthetic processes have been developed for thecontrollable synthesis of Bi2WO6 archietectures.31–36 Li and co-workers synthesized 3D Bi2WO6 architectures by a PVP-assistedhydrothermal method.34 Zhang's group synthesized 3DBi2WO6 architectures via a triblock pluronic P123-assisted

J. Mater. Chem. C, 2013, 1, 4153–4162 | 4153

Journal of Materials Chemistry C Paper

Publ

ishe

d on

26

Apr

il 20

13. D

ownl

oade

d by

BE

IJIN

G I

NST

ITU

TE

OF

TE

CH

NO

LO

GY

on

06/1

1/20

13 1

3:35

:17.

View Article Online

hydrothermal process.37 Also, Ma et al. studied the self-assem-bled 3D Bi2WO6 architectures by using citrate anions as acomplexing agent in the hydrothermal process,38 and supposedthat, in the formation process of the Bi2WO6 nanostructure,citrate could selectively absorb at the crystal face of Bi2WO6

nanoplates and direct the orientation and self-assembly of theseprimary nano-building blocks, resulting in the formation ofBi2WO6 nanostructure. Tian et al. obtained nest-like Bi2WO6

nanostructure using CTAB as template.39 These approaches arepromising, however, the processes involved additives as atemplate or structure-directing reagent. Therefore, it remains asignicant challenge to develop simple, effective, economical,and template-free approaches for the synthesis of Bi2WO6

architectures. Recently, our group has successfully synthesized3D nest-like mesoporous Bi2WO6 architectures by a hydro-thermal method without any additives, and Ag nanoparticleswere deposited on the surface of Bi2WO6 via a following facilephotoreduction process. The photocatalysis experimentalresults revealed that Ag-loaded Bi2WO6 exhibited high photo-catalytic activity compared with pure Bi2WO6.40 Very recently,Chen and co-workers have reported that the Aurivillius typeBi2MO6 (M ¼ W, and Mo) exhibited signicant impedancechanges of 4–5 orders of magnitude with increase of the relativehumidity from 11% to 95%,41 which inspired us to investigatethe gas sensitivity of Bi2MO6-based materials. In this work, wereport the synthesis and the growingmechanism of the uniformhierarchical multilayered Bi2WO6 architectures with diameterof 3–4 mm prepared via a facile template-free hydrothermalprocess, and furthermore, the as-prepared Bi2WO6 sampleswere fabricated into a gas sensor, and the experimental resultsshowed that the samples exhibited a remarkable improvementof sensing properties for ethanol compared to the reportedmesoporous ZnO or WO3 sensors. The effects of morphology,size of crystallites, surface areas of the mesoporous Bi2WO6 ongas sensing properties were also discussed.

2 Experimental section2.1 Synthesis of Bi2WO6 architectures

All of the chemicals were analytical grade and used withoutfurther treatment. Bi2WO6 architectures were synthesized byhydrothermal reaction under autogenerated pressure. In atypical procedure, 0.4 g (1.0 mmol) Bi(NO3)3$5H2O was rstdissolved in 20 mL of nitric solution (0.4 mol L�1), and then10 mL of (NH4)10W12O41 solution with concentration of0.0082 mol L�1 was added drop-wise to the above solutionunder vigorous magnetic stirring at room temperature. Understirring, the diluted NaOH and HNO3 (4 mol L�1) were used toadjust the pH value to 1. Aer being magnetically stirred atroom temperature for 2 h, the resulting precursor suspensionwas transferred into a 50 mL Teon-lined autoclave. The auto-clave was then sealed and maintained at 190 �C for differenttimes. Subsequently, the autoclave was cooled to roomtemperature naturally. Aer ltration, the yellowish whiteprecipitate was collected and washed with distilled water andabsolute ethanol several times, then dried under vacuum at80 �C for 4 h. The prepared samples were denoted as S0.5, S1,

4154 | J. Mater. Chem. C, 2013, 1, 4153–4162

S2, S6 and S12, respectively, with the number referring to thereaction time.

2.2 Characterization

The purity and crystalline structure of the samples were carriedout with a Shimadzu XRD-7000 X-ray diffractometer (XRD) withCu Ka radiation (l ¼ 0.15418 nm). The accelerating voltage andthe applied current were 40 kV and 30 mA, respectively. X-Rayphotoelectron spectroscopy (XPS) was recorded on a PHI-5400X-ray photoelectron spectrometer. The morphology of thesample was observed by eld emission scanning electronmicroscopy (FE-SEM, JSM-6700F). The sample was furtherinvestigated by high-resolution transmission electron micros-copy (HR-TEM). HR-TEM studies were carried out on a JEM-2100 electron microscope at an accelerating voltage of 200 kV.The selected area electron diffraction (SAED) pattern wasobtained from HR-TEM to determine if the sample is singlecrystalline. The pore diameter distribution and surface area ofBi2WO6 architectures were tested by nitrogen adsorption–desorption (Automated Physisorption and ChemisorptionAnalyzer, Micromeritics ASAP 2010).

2.3 Fabrication and measurement of gas sensor

Gas sensing of Bi2WO6 samples was performed in a WS-30Astatic gas-sensing system (Zhengzhou Wei-Sheng ElectronicsTechnology Co., Ltd., Henan, P. R. China). The gas sensor is aside-heated type made as follows: Bi2WO6 samples were mixedwith ethanol in an agate mortar to form the paste. The paste wascoated on a ceramic tube (provided by Zhengzhou Wei-ShengElectronics Technology Co., Ltd. The internal diameter of theceramic tube is 0.8 mm. The external diameter of ceramic tubeis about 1.4 mm, and the length of tube is about 4 mm)equipped with a pair of Pt wire on Au electrode. A spring-like Ni–Cr wire was inserted into the ceramic tube to provide theoperating temperature (Fig. 1). The coating thickness is about0.25 mm. All the sensors were aged at 400 �C overnight toimprove their stability. The sensors are labeled as GS1, GS2,GS6, and GS12, respectively, by the same naming scheme asdescribed above. The sensor signal voltage (Vout) was collectedby a computer at a constant test voltage of 5 V. The sensitivity, S,is determined as the ratio, Ra/Rg, where Ra is the resistance in airand Rg is the resistance in the tested gas atmosphere. Theresponse time is dened as the time taken for the sensor toreach 90% of the saturation value aer the sensor is exposed togases, and the recovery time is dened as the time taken for thesensor to decrease to 10% of the saturation value aerthe removal of gases. When testing, the calculated amount ofthe target liquid was injected into the chamber by a micro-injector on the evaporator in the back of the chamber andgasied quickly. The used formula can be written as:

Vl ¼ VC4M

22:4rp� 10�6 � 273:15þ TR

273:15þ TB

(1)

where Vl is the liquid volume (mL), VC is the volume of testchamber (mL), 4 is the required volume fraction, M is themolecular weight (g mol�1), p is the purity of the selected

This journal is ª The Royal Society of Chemistry 2013

Fig. 1 Experimental setup for gas sensing.

Paper Journal of Materials Chemistry C

Publ

ishe

d on

26

Apr

il 20

13. D

ownl

oade

d by

BE

IJIN

G I

NST

ITU

TE

OF

TE

CH

NO

LO

GY

on

06/1

1/20

13 1

3:35

:17.

View Article Online

solutions, TR and TB are the temperature at ambient and testchamber (�C), respectively. In order to eliminate the effect ofwater gas, the synthetic dry air (mixture gas of N2 and O2, v/v,4 : 1, relative humidity, 0%) was introduced to the chamber for10 min to replace the moist air at the beginning of the testing,then the target liquid was injected into the chamber. Further-more, we also investigate the inuence of ambient air humidityon the gas response of as-fabricated sensors. The controlledambient environments for all measurements were achievedaccording to the literature.41 WO3 gas sensor, known for its goodsensitivity, was prepared as a reference. According to the liter-ature,42 mesoporous WO3 based gas sensor was also fabricatedto compare the gas sensing properties with our mesoporousBi2WO6 based gas sensor.

3 Results and discussion3.1 Crystallinity and morphology

The phase structure, crystallinity and purity of as-obtainedproducts were examined by XRD and XPS measurements. Atypical XRD pattern is shown in Fig. S1.† All of the diffractionpeaks can be readily indexed to pure orthorhombic phaseBi2WO6 (JCPD card no.73-1126). No characteristic peaks ofother impurities were observed. Aer renement, the crystalparameters of Bi2WO6 were obtained: a ¼ 5.458(2) A, b ¼5.432(8) A and c ¼ 16.403(4) A. According to Scherrer equation,the crystallite size is ca. 16.8 nm. Moreover, the productspresent high crystallinity when reaction time reaches 2 h. In theinset of Fig. S1,† a standard XRD pattern of Bi2WO6 is given. Theintensity ratio of (200) or (020) peak to (113) in our result is 0.75,obviously larger than the standard value of 0.185, which indi-cates that the crystal has special anisotropic growth along the(200)/(020) plane, in good agreement with the plate-likemorphology of Bi2WO6. The X-ray photoelectron spectroscopy(XPS) was carried out to determine the chemical compositionand the valence states of the as-obtained 3D Bi2WO6 architec-ture. The XPS results illustrate that the sample contains Bi, W, Oelements and a trace amount of carbon (Fig. 2A). The XPS peakfor C element is due to the adventitious hydrocarbon from theXPS instrument itself. Fig. 2B–D shows the high-resolutionspectra of the Bi 4f, W 4f and O 1s regions, respectively. Thespin-orbit components of Bi 4f consists of two peaks withbinding energies around 159.2 eV and 164 eV (Fig. 2B), corre-sponding to Bi3+ in the crystal structure. Peaks at 37.5and 35.2 eV, as shown in Fig. 2C, corresponding to W 4f5/2 and

This journal is ª The Royal Society of Chemistry 2013

W 4f7/2, respectively, can be assigned to a W6+ oxidation state ofBi2WO6.43 As shown in Fig. 2D, the asymmetric XPS of O 1sindicated that oxygen species were present in the form of alattice with hydroxyl oxygen on the surface.44,45 The O elementmay be tted into three kinds of chemical states: peaks at529.66 and 531.04 eV correspond to the crystal lattice, which areconsistent to different chemical environments of oxygenelement in [WO4]

2� and [Bi2O2]2+ layers,46 and the weak peak at

531.77 eV to surface absorbed oxygen species. Furthermore,peak areas were determined for the quantitative elementalanalysis of Bi, W, and O element in the as-obtained 3D Bi2WO6

nanostructure. The atomic ratio of Bi to W to O is about2 : 1.02 : 6, which further proves that the samples are pureBi2WO6 phase without any impurities.

Fig. 3 displays the typical FE-SEM images of the sample byhydrothermal reaction at 190 �C for 2 h. It can be seen thatas-obtained Bi2WO6 exhibits a uniform nest-like three-dimen-sional (3D) Bi2WO6 architecture with scales of 3–4 mm (Fig. 3a–d). Interestingly, the nest-like Bi2WO6 3D architecture consistsof many secondary nanoplates. Further information about theBi2WO6 architecture were obtained from FE-SEM images asgiven in Fig. 3d and TEM images in Fig. 4e and f. From thesegures, it can be seen that the nest-like 3D Bi2WO6 architectureis built up by nanoplates. In addition, a selected area electrondiffraction (SEAD) pattern of the whole superstructure was alsorecorded (inset of Fig. 3e). It is clear that nest-like 3D Bi2WO6

architectures are polycrystalline in nature, and the architectureis organized by the nanocrystalline subunits. However, thepattern has arc-like spots with obvious symmetry, which directlydemonstrates that the superstructure should consist of single-crystal building blocks.47,48 The single crystalline nature andparameters of nanoplates were also conrmed by HR-TEM(Fig. 3f). Fig. 3f is the enlarged TEM image of the areamarked bya red circle in Fig. 3e. Obviously, the nanoplates can be seen,which assemble into one layer with their sides. The SAEDpattern (inset picture in Fig. 3f) exhibits a regular and clearsquare diffraction spot array and can be indexed to singlecrystalline Bi2WO6. From the high-resolution TEM (HR-TEM)shown in Fig. 3f (inset picture in right corner), the latticeinterplanar spacing is measured to be 0.195 nm, correspondingto the (220) planes of orthorhombic Bi2WO6 and is wellconsistent with the XRD pattern in Fig. S1.† This result alsoconrms the single structure of the nanoplates.

In order to investigate the porous structure of the as-synthesized Bi2WO6 architectures, N2 adsorption–desorption

J. Mater. Chem. C, 2013, 1, 4153–4162 | 4155

Fig. 3 (a) Low-magnification and (b–d) high-magnification FE-SEM images of the as-obtained 3D Bi2WO6 architectures. (d) SEM-image of an individual nest-likearchitecture. (e) TEM image and its corresponding SAED of an individual nest-like Bi2WO6 architecture. (f) Enlarged TEM and HRTEM image of a simple nanoplateshowing its single-crystal nature.

Fig. 2 XPS spectra of as-obtained Bi2WO6 architectures. (A) The survey spectra, and the high resolution XPS spectra of the sample, (B) Bi4f, (C) W4f and (D) O1s.

Journal of Materials Chemistry C Paper

Publ

ishe

d on

26

Apr

il 20

13. D

ownl

oade

d by

BE

IJIN

G I

NST

ITU

TE

OF

TE

CH

NO

LO

GY

on

06/1

1/20

13 1

3:35

:17.

View Article Online

isotherms and Barrett–Joyner–Halenda (BJH) pore size distri-bution analysis were performed. As shown in Fig. 4, thenitrogen adsorption and desorption isotherms can be catego-rized as type IV isotherm with a hysteresis loop observed in therelative (P/P0) range of 0.4–1.0, which implies the presence ofmesopores in the size of 2–50 nm.34 In additional, the hysteresisloop shis to a higher pressure on approaching P/P0 z 1, whichsuggests that macropores are also present. This is alsoconrmed by the corresponding pore-size distribution thatshows two peaks for mesopores along with macropores up to

4156 | J. Mater. Chem. C, 2013, 1, 4153–4162

100 nm in size (inset in Fig. 4). These pores presumably arisefrom the spaces among the nanoplates within the nest-like 3DBi2WO6 architecture. The 3D nest-like Bi2WO6 architecturespossess small mesopores of ca. 8.5 nm and large mesoporeswith a peak at about 45 nm. As revealed by FE-SEM observation,the smaller pores may be generated during the crystal growthprocess, whereas the larger pores can be attributed to the spacebetween the intercrossed Bi2WO6 nanoplates.48 The BET surfacearea of nest-like 3D Bi2WO6 architecture was calculated from N2

isotherms and is 47.72 m2 g�1. The single-point total volume of

This journal is ª The Royal Society of Chemistry 2013

Fig. 4 Nitrogen adsorption–desorption isotherms and Barrett–Joyner–Halenda(BJH) pore size distribution plot (inset picture) of as-prepared nest-like 3D Bi2WO6

architecture.

Paper Journal of Materials Chemistry C

Publ

ishe

d on

26

Apr

il 20

13. D

ownl

oade

d by

BE

IJIN

G I

NST

ITU

TE

OF

TE

CH

NO

LO

GY

on

06/1

1/20

13 1

3:35

:17.

View Article Online

pores at P/P0 ¼ 0.991 is 0.452 cm3 g�1. The porous structure andthe large BET surface area may endow the as-prepared 3D nest-like Bi2WO6 architecture with novel applications.

3.2 Formation mechanism of the mesoporous 3D Bi2WO6

architecture

Time-depended experiments were performed with Bi2WO6

crystal growth time of 0.5, 1.0, 2.0, 6, and 12 h by using(NH4)10W12O41$5H2O and Bi(NO3)3$5H2O as starting materials

Fig. 5 FE-SEM images of the samples prepared by hydrothermal process at 190 �C fof samples (i).

This journal is ª The Royal Society of Chemistry 2013

under the hydrothermal temperature at 190 �C. FE-SEM imagesof products formed at different evolution stages are presentedin Fig. 5. In the initial stage (hydrothermal treatment for 30min), a number of spherical nanoparticles with a diameter of20 nm (Fig. 5a) can be seen, indicating that Bi2WO6 nano-particles are precipitated in the solution as a result of homo-geneous nucleation. As the reaction time prolonged to 60 min,some nanoplates with the thickness of 20–30 nm can bedetected, coexisting with a mass of spherical nanoparticles(Fig. 5b). With the elongation of reaction time, the nanoplatesgrow gradually with oriented self-assembly and the sphericalnanoparticles disappear. Aer 2 h of reaction, the as-obtainedS2 sample exhibits a nest-like 3D architecture with scales of 3–4 mm (Fig. 5c and d). Interestingly, the nest-like 3D Bi2WO6

architectures have many small secondary nanoplates (Fig. 5eand f). Increasing the hydrothermal treatment time to 6 h and12 h, the morphology (Fig. 5g and h), BET surface area of theproducts have no obvious change (Fig. S2†). The specic surfaceareas of the samples obtained for the reaction time of 0.5 h and1.0 h are 25.6 and 32.6 m2 g�1, respectively (Fig. S2a and b†).The largest surface area of 47.72 m2 g�1 is achieved when thesample was synthesized at 190 �C for 2.0 h (Fig. S2c†). Extendingthe hydrothermal treatment time to 6 h, BET surface area of theproducts have no obvious change, the value is 46.5 m2 g�1. TheXRD pattern of samples in Fig. 5i is well consistent with theFE-SEM results. It can be seen that some weak peaks of ortho-rhombic Bi2WO6 have already emerged aer 0.5 h of hydro-thermal reaction. With increasing the reaction time, thediffraction peaks have become stronger gradually. As shown in

or different times; (a) 0.5 h, (b)1 h, (c–f) 2 h, (g) 6 h and (h) 12 h; and XRD patterns

J. Mater. Chem. C, 2013, 1, 4153–4162 | 4157

Journal of Materials Chemistry C Paper

Publ

ishe

d on

26

Apr

il 20

13. D

ownl

oade

d by

BE

IJIN

G I

NST

ITU

TE

OF

TE

CH

NO

LO

GY

on

06/1

1/20

13 1

3:35

:17.

View Article Online

Fig. 5i, well crystalline orthorhombic Bi2WO6 are obtained withsharp diffraction peak aer hydrothermal treatment for 2 h.

The effect of hydrothermal condition, such as, pH value,reaction temperature and species of tungsten source on thephase structure and morphology and of Bi2WO6 was alsoinvestigated (Fig. S3–S5†). The result indicated that pH value,hydrothermal temperature and tungsten source is vital to theformation of mesoporous nest-like Bi2WO6 architecture(detailed in the ESI†). When the pH value increased from 1.0 to3.0 while xing the other hydrothermal conditions to beconstant, the morphology of Bi2WO6 is shown in Fig. S4,† fromwhich the ower-like Bi2WO6 nanostructure can be observedwhile nest-like nanostructure disappeared. Increasing the pHvalue to 5.0 and 7.0, the morphology of the hydrothermalproduct is distinctly different. It can be seen that the multidisc-like Bi2WO6 nanostructure are the main products. As the pHvalue continues to increase to 9.0, the three-dimensionalstructures fall apart, and the nanoplates of Bi2WO6 arerandomly piled up with no typical aggregation. These resultswere partly consistent with the literature result reported byWang's group.37 However, our process has the advantages ofsimplicity, speed and lower energy consumption.

When hydrothermal temperature was changed while xingthe other hydrothermal conditions to be constant, themorphology of Bi2WO6 is shown in Fig. S5(a–e).† Fig. S5† showsthe FE-SEM images of samples obtained under differenthydrothermal temperature for 2 h. As shown in Fig. S5,† thesamples exhibit sphere-like nanoparticles with the size of about20 to 50 nm under 140 �C and 160 �C, respectively. Furtherincreasing the hydrothermal temperatures to 170 �C, theobtained samples begin to appear in a plate-like morphology,which indicates that the Bi2WO6 nanoparticles begin to growalong the certain direction. With an increase of hydrothermaltemperature to 180 �C, more and more plate-like morphologyBi2WO6 was formed coexisting with a mass of spherical nano-particles. Fig. S5(e)† shows the FE-SEM images of Bi2WO6

obtained under 190 �C hydrothermal reactions for 2 h. It can beseen that as-obtained Bi2WO6 exhibits an uniform nest-like 3DBi2WO6 architecture with scales of 3–4 mm. Moreover, whenxing the other conditions, only using Na2WO4$5H2O instead of(NH4)10W12O41$5H2O, FE-SEM image of the resulting sample isshown in Fig. S5(f).† Interestingly, the as-obtained sampleexhibit multilayer-disk morphology aer hydrothermal reac-tion. So, we can draw a conclusion that pH value, hydrothermal

Fig. 6 Schematic description of the growth process of nest-like 3D Bi2WO6 archite

4158 | J. Mater. Chem. C, 2013, 1, 4153–4162

temperature, and species of tungsten source is vital to theformation of the 3D nest-like Bi2WO6 architecture. On the basisof our experimental results and literature reported,20,26,37,38

relevant chemical reactions can be proposed as follows:

Bi(NO3)3 + H2O / 2HNO3 + BiONO3 (2)

BiONO3 + H2O / Bi2O2(OH)NO3 + HNO3 (3)

(NH4)10W12O41$5H2O + 10HNO3 / 10NH4NO3 + 12H2WO4(4)

H2WO4 / 2H+ + WO42� (5)

Bi2O2(OH)NO3 + H+ + WO42� / Bi2WO6Y + HNO3 + H2O(6)

Based on the above experimental results, it can be concludedthat the intricate crystal growth process of primary nano-particles proceeds simultaneously with oriented assembly of theas-formed nanoplates into a 3D architecture. The main processis illustrated in Fig. 6. In the rst step, spherical amorphousBi2WO6 nanoparticles form aer (NH4)10W12O41 aqueous solu-tion is added to Bi(NO3)3 solutions. With prolonging the reac-tion time, Bi2WO6 nanoparticles begin to crystallize in asupersaturated solution during the initial dissolution andrenucleation process. Our experimental result is clear that pHvalue of the precursor suspensions greatly affect the shape ofthe nal products Bi2WO6. When reaction occurs at low pH, theconcentration of H+ ions is much higher than that of OH� ions,which restrains the hydrolysis of Bi3+. Thus, the nucleation rateof Bi2WO6 is much faster than the rate of crystal growth due tothe existence of large quantities of Bi3+. Large quantities ofBi2WO6 nuclei tend to aggregate together to form larger andmore thermodynamically stable spherical particles. Subse-quently, Ostwald-ripening, as a common phenomenon incrystallization processes, takes place. According to the Gibbs–Thomson law,20,34 larger Bi2WO6 crystals grow at the expense ofsmaller crystalline nuclei and nally Bi2WO6 nanoplates areattained. Meanwhile, the above anisotropic growth processoccurs simultaneously with oriented assembly process of theBi2WO6 nanoplates.49 The individual Bi2WO6 nanoplates inti-mately contact each other in a side-by-side manner and tend torotate adequately to meet the low-energy conguration inter-face. As the orthorhombic Bi2WO6 is constructed by a numberof alternating (Bi2O2)n

2+ layers and perovskite-like (WO4)n2�

layers, stacking along the c axis, the oppositely charged layers

ctures.

This journal is ª The Royal Society of Chemistry 2013

Paper Journal of Materials Chemistry C

Publ

ishe

d on

26

Apr

il 20

13. D

ownl

oade

d by

BE

IJIN

G I

NST

ITU

TE

OF

TE

CH

NO

LO

GY

on

06/1

1/20

13 1

3:35

:17.

View Article Online

may result in the formation of nanoplates with polar charges ontheir top and bottom surfaces. Thus, Bi2WO6 generally exhibitsa plate-like anisotropic structure.20 Due to electrostatic effects,these nanoplates can connect to each other. Subsequently, nest-like architectures can be formed under the certain hydro-thermal temperature without any template or surfactant (Fig. 6).

Fig. 8 Gas responses of sensors fabricated by as-prepared Bi2WO6 samplesunder different hydrothermal times (the operating temperature is 300 �C, and theconcentration of all gases is 20 ppm, relative humidity (RH), 0%).

3.3 Gas sensing properties

It is well known that morphology, size and surface structure ofnanostructures can dramatically change its optical, electrical,and physical properties.48,50–55 So, it could be expected that sucha uniform hierarchical multilayered Bi2WO6 architecturesmight be a potential candidate for gas sensing applications. Todetermine the optimum working temperature for our Bi2WO6

based sensors, four sensors were fabricated from as-preparedBi2WO6 samples under different hydrothermal time, and theirgas sensing properties were investigated. The responses ofsensors to 20 ppm ethanol gas in dry air were examined vs.temperature (Fig. 7). It is obviously seen that gas responses of allthe four sensors increase with increasing temperature, and thegas response of GS2 sensor is higher than those of other sensorsat all temperatures. The response shows a maximum at 300 �C.Aerwards, the responses decrease again. So, the optimumoperating temperature of 300 �C has therefore been performedin further experiments.

The most desirable sensor would be one which is sensitive toonly a single gas and is not affected by others at all. However,most semiconductor sensors suffer from a lack of gas selectivity.The responses of the Bi2WO6-based sensors to 20 ppm ofethanol, n-isopropanol, cyclohexane, and petroleum ether at300 �C were investigated and the results are shown in Fig. 8. Theresults show that all the sensors exhibit higher responsesensitivity to ethanol than other gases, and GS2 sensor exhibitsthe best ethanol gas response because its mesoporous structureprovides a high surface-to-volume ratio and large surfaceaccessibility for the ethanol molecules entering into andcoming out (Fig. 5).

Fig. 7 Temperature dependence of gas response of sensors (the concentrationof ethanol gas was 20 ppm, relative humidity (RH), 0%).

This journal is ª The Royal Society of Chemistry 2013

Sensitivity is an important factor of chemical sensors, ahigher sensitivity can usually allow for a low detection limit.56,57

The sensitivity of Bi2WO6 gas sensor to different ethanolconcentrations are compared in Fig. 9, showing the sensitivityof Bi2WO6 gas sensor is improved with increasing gas concen-tration. The GS2 sensor reaches up to 34.6 on exposure to100 ppm of ethanol gas compared with 12.4, 9.6 and 7.4 for GS6,GS12, and GS1 gas sensors, respectively. In other words, the gassensitivity of GS2 sensor on exposure to 100 ppm of ethanol gasis 2.8, 3.6 and 4.8 times that of GS6, GS12, and GS1 gas sensors,respectively. Response and recovery times are also importantparameters in a gas sensor. Fig. 10 shows the response–recoverycurves of GS2 sensor to ethanol with concentration between 1and 100 ppm at 300 �C. The response–recovery time increaseswith increasing ethanol concentration. When the concentrationof ethanol increases to 5, 10, 20, 50 and 100 ppm, correspondingrecovery time is 13, 16, 19, 21, 24 and 27 s, respectively. The gasresponse of GS2 sensor is better than that of other sensors,which can be ascribed to the porous structure and its large BETsurface area of the S2 sample (Fig. S2†). And the porous

Fig. 9 Responses vs. ethanol concentration for the gas sensors based on theBi2WO6 samples prepared under different hydrothermal times (relative humidity(RH), 0%).

J. Mater. Chem. C, 2013, 1, 4153–4162 | 4159

Fig. 10 Response–recovery curve of GS2 sensor when operated at 300 �C(RH 0%).

Journal of Materials Chemistry C Paper

Publ

ishe

d on

26

Apr

il 20

13. D

ownl

oade

d by

BE

IJIN

G I

NST

ITU

TE

OF

TE

CH

NO

LO

GY

on

06/1

1/20

13 1

3:35

:17.

View Article Online

structure composed of 3D Bi2WO6 architectures have largercontact areas with tested gas to improve the gas sensitivity. Inaddition, the sensor exhibited satisfactory sensing properties atlow temperature (300 �C). For 100 ppm ethanol, the sensitivitywas 34.6, which is about 3-fold higher than reported meso-porous ZnO based gas sensor.48 For comparison, mesoporousWO3 was also synthesized according to the literature,42 andmesoporous WO3 based sensor was fabricated. The results areshown in Fig. S6.† Ethanol gas sensing tests showed that theresponse and recovery time of 3D mesoporous Bi2WO6 archi-tectures was more rapid than that of WO3 based gas sensor.Moreover, the inuence of relative humidity (RH) on the gassensor was also investigated. The result indicated that humiditydependence is low at 300 �C under low relative humidity. Whenthe humidity is more than 30%, the sensitivity (Rg/Ra) exhibits aslight increase (Fig. S7a†). However, the relative response (SRH/S0) is nearly invariable (Fig. S7b†). Where SRH is the sensitivityvalue at certain RH value, S0 is the sensitivity value at RH ¼ 0%.

3.4 The sensing mechanism of mesoporous Bi2WO6

It is well known that oxygen ionosorption removes conductionelectrons and thus lowers the conductance of semi-conductors.56–58 For Bi2WO6-based sensor, the change of resis-tance is mainly caused by the adsorption and desorption of gasmolecules on the surface of the sensing structure. The experi-mental observations described above can be explained by thesurface-depletion model,57–59 as shown in Fig. 11. When the as-

Fig. 11 Schematic diagram of the proposed reaction mechanism of Bi2WO6-based

4160 | J. Mater. Chem. C, 2013, 1, 4153–4162

fabricated sensor is exposed to air, molecular oxygen can adsorbon the surface of Bi2WO6 nanostructure and form O2

�, O2� andO� by capturing electrons from the conduction band (CB).These lead to the formation of a chick space-charge layer whichincreases the potential carrier, and thus results in a higherresistance (Fig. 11a). When the sensor is exposed to ethanol,which is a reductive gas, the gas would react with adsorbedoxygen species on the Bi2WO6 surface to form CO2 and H2O(Fig. 11b). This leads to an increasing carrier concentration ofthe sample and decreasing resistance of the sensor.

According to the model, ethanol molecules can be oxidizedby the surface absorbed oxygen species and consequently, thedepleted electrons would be released to Bi2WO6. The electronicstates of the adsorbed oxygen species on the surface of Bi2WO6

material change as follows:14,16,59

O2(gas) / O2(ads) (7)

O2(ads) + e� / 2O2�(ads) (8)

2O2�(ads) + e� / 2O�

(ads) (9)

where the subscript ads represents the state of adsorption.Reactive oxygen species adsorbed on the surface of Bi2WO6,such as O2

�, O2�, and O� depend strongly on temperature.58 Atlow temperatures, O2

� is commonly chemisorbed. At hightemperature, O� and O2� are commonly chemisorbed, however,while O2

� disappears rapidly.60 Since our sensor was operated at300 �C, the O� species were more important than other oxygenadsorbates.61–63 When the sensor was exposed to ethanol, theoxygen adsorbates react with ethanol as follows:

(C2H5OH)(ads) / (C2H5OH)(ads) (10)

(C2H5OH)(ads) + 7O�(ads) / 2CO2(gas) + 3H2O(gas) + 7e� (11)

Above reactions release electrons, and the released electronsgo back into the conduction band of Bi2WO6, which leads to thedecreasing of resistance of the sensors.

4 Conclusion

In summary, uniform hierarchical multilayered Bi2WO6 archi-tectures with diameter of 3–4 mm were successfully synthesizedvia a facile template-free hydrothermal process. The as-prepared samples have very large specic surface areas, and the

gas sensor. (a) In air and (b) in ethanol.

This journal is ª The Royal Society of Chemistry 2013

Paper Journal of Materials Chemistry C

Publ

ishe

d on

26

Apr

il 20

13. D

ownl

oade

d by

BE

IJIN

G I

NST

ITU

TE

OF

TE

CH

NO

LO

GY

on

06/1

1/20

13 1

3:35

:17.

View Article Online

largest surface area of 47.72 m2 g�1 is achieved when synthe-sized at 190 �C for 2.0 h. The samples were used to fabricate gassensors, and the responses to ethanol, n-isopropanol, cyclo-hexane, and petroleum ether were investigated. The resultsshow that the sensors exhibit higher response sensitivity toethanol than other gases, faster response–recovery and betterselectivity than the reported mesoporous ZnO or WO3 sensors.For 100 ppm ethanol, the sensitivity of the best sensor (GS2) was34.6, which is about 3-fold higher than reported mesoporousZnO based gas sensor. The 3D hierarchical multilayered mes-oporous Bi2WO6 based gas sensor exhibits higher sensitivitythan the other tested sensors because of the following reasons:(a) the mesoporous structure has a high surface-to-volume ratioand large surface accessibility for the ethanol to adsorb andentering into and come out of the mesoporous structure freely.(b) The large surface–volume ratio of the 3D mesoporousstructure of Bi2WO6 structure might give rise to more potentialadsorption sites to adsorb ethanol molecules, therebyimproving the sensitivity. (c) The 3D mesoporous structure ofBi2WO6 may help ethanol gas molecules penetrate the surfaceof Bi2WO6 architectures to contact with inner oxygen species ofthe surface of the materials, thus, increase the gas response. So,the ethanol molecules could not only be absorbed onto thegrains of the outer surface, but also fully interact with those onthe inner surface of the mesoporous structure, resulting in thehigh gas response and fast recovery.

Acknowledgements

This work is supported by the National Natural Science Foun-dation of China (nos 20973133, 21073106). It is also nanciallysupported by the key industry plan of Yan'an City (Grant no.2011kg-13).

References

1 C. H. Wang, X. F. Chu andM. M. Wu, Sens. Actuators, B, 2006,113, 320–323.

2 Q. Y. Ouyang, L. Li, Q. S. Wang, Y. Zhang, T. S. Wang,F. N. Meng, Y. J. Chen and P. Gao, Sens. Actuators, B, 2012,169, 17–25.

3 H. Bai and G. Q. Shi, Sensors, 2007, 7, 267–307.4 M. C. Lonergan, E. J. Severin, B. J. Doleman, S. A. Beaber,R. H. Grubbs and N. N. Lewis, Chem. Mater., 1996, 8, 2298–2312.

5 S. G. Chen, J. W. Hu, M. Q. Zhang, M. W. Li and M. Z. Rong,Carbon, 2004, 42, 645–651.

6 T. Hyido, T. Ohoka, Y. Shimizu and M. Egashira, Sens.Actuators, B, 2006, 117, 359–366.

7 L. Zhang, J. F. Hu, P. Song, H. W. Qin, K. An, X. D. Wang andM. H. Jiang, Sens. Actuators, B, 2006, 119, 315–318.

8 H. Suo, F. Q.Wu, Q. Y. Wang, G. F. Liu, F. B. Qiu, B. K. Xu andM. Y. Zhao, Sens. Actuators, B, 1997, 45, 245–249.

9 T. Chen, Z. L. Zhou and Y. D. Wang, Sens. Actuators, B, 2009,143, 124–131.

10 G. S. Devi, T. Hyodo, Y. Shimizu and M. Egashira, Sens.Actuators, B, 2002, 87, 122–129.

This journal is ª The Royal Society of Chemistry 2013

11 T. Wagner, T. Waitz, J. Roggenbuck, M. Froba, C. D. Kohland M. Tiemann, Thin Solid Films, 2007, 515, 8360–8363.

12 T. Waitz, T. Wagner, T. Sauerwald, C. D. Kohl andM. Tiemann, Adv. Funct. Mater., 2009, 19, 653–661.

13 J. H. Lee, Sens. Actuators, B, 2009, 140, 319–336.14 T. J. Hsueh, C. L. Hsu, S. J. Chang and I. C. Chen, Sens.

Actuators, B, 2007, 126, 473–477.15 W. Y. Wu, J. M. Ting and P. J. Huang, Nanoscale Res. Lett.,

2009, 4, 513–517.16 O. Lupan, V. V. Ursaki, G. Chai, L. Chow, G. A. Emelchenko,

I. M. Tiginyanu, A. N. Gruzintsev and A. N. Redkin, Sens.Actuators, B, 2010, 144, 56–66.

17 O. K. Tan, W. Cao, W. Zhu, J. W. Chai and J. S. Pan, Sens.Actuators, B, 2003, 93, 396–401.

18 H. Y. Lai, T. H. Chen and C. H. Chen, CrystEngComm, 2012,14, 5589–5595.

19 X. L. Li, T. J. Lou, X. M. Sun and Y. D. Li, Inorg. Chem., 2004,43, 5442–5449.

20 C. Zhang and Y. F. Zhu, Chem. Mater., 2005, 17, 3537–3545.21 N. A. McDowell, K. S. Knight and P. Lightfoot, Chem.–Eur. J.,

2006, 12, 1493–1499.22 J. W. Tang, Z. G. Zou and J. H. Ye, Catal. Lett., 2004, 92, 53–

56.23 C. Y. Wang, H. Zhang, F. Li and L. Y. Zhu, Environ. Sci.

Technol., 2010, 44, 6843–6848.24 H. B. Fu, C. S. Pan, W. Q. Yao and Y. F. Zhu, J. Phys. Chem. B,

2005, 109, 22432–22439.25 F. Amano, K. Nogami and B. Ohtani, J. Phys. Chem. C, 2009,

113, 1536–1542.26 L. S. Zhang, W. Z. Wang, Z. G. Chen, L. Zhou, H. L. Xu and

W. Zhu, J. Mater. Chem., 2007, 17, 2526–2532.27 J. Wu, F. Duan, Y. Zheng and Y. Xie, J. Phys. Chem. C, 2007,

111, 12866–12871.28 G. K. Zhang, F. Lu, M. Li, J. L. Yang, X. Y. Zhang and

B. B. Huang, J. Phys. Chem. Solids, 2010, 71, 579–582.29 Z. J. Zhang, W. Z. Wang, M. Shang and W. Z. Yin, J. Hazard.

Mater., 2010, 177, 1013–1018.30 L. Zhou, W. Z. Wang and L. S. Zhang, J. Mol. Catal. A: Chem.,

2007, 268, 195–200.31 S. O. Alfaro and A. Martınez-de la Cruz, Appl. Catal., A, 2010,

383, 128–132.32 S. H. Chen, S. X. Sun, H. G. Sun, W. L. Fan, X. Zhao and

X. Sun, J. Phys. Chem. C, 2010, 114, 7680–7688.33 H. B. Fu, S. C. Zhang, T. G. Xu, Y. F. Zhu and J. M. Chen,

Environ. Sci. Technol., 2008, 42, 2085–2091.34 Y. Y. Li, J. P. Liu, X. T. Huang and G. Y. Li, Cryst. Growth Des.,

2007, 7, 1350–1355.35 J. G. Yu, H. G. Yu, B. Cheng and C. Traplis, J. Mol. Catal. A:

Chem., 2006, 249, 135–142.36 D. V. Bavykin, V. N. Parmon, A. A. Lapkin and F. C. Walshi, J.

Mater. Chem., 2004, 14, 3370–3377.37 L. S. Zhang, W. Z. Wang, L. Zhou and H. L. Xu, Small, 2007, 3,

1618–1625.38 D. K. Ma, S. M. Huang, W. X. Chen, S. W. Hu, F. F. Shi and

K. L. Fan, J. Phys. Chem. C, 2009, 113, 4369–4374.39 Y. Tian, G. M. Hua, N. Li, M. Fang and L. D. Zhang, J. Alloys

Compd., 2011, 509, 724–730.

J. Mater. Chem. C, 2013, 1, 4153–4162 | 4161

Journal of Materials Chemistry C Paper

Publ

ishe

d on

26

Apr

il 20

13. D

ownl

oade

d by

BE

IJIN

G I

NST

ITU

TE

OF

TE

CH

NO

LO

GY

on

06/1

1/20

13 1

3:35

:17.

View Article Online

40 D. J. Wang, G. L. Xue, Y. Z. Zhen, F. Fu and D. S. Li, J. Mater.Chem., 2012, 22, 4751–4758.

41 K. B. Zheng, Y. Zhou, L. L. Gu, X. L. Mo, G. R. Patzke andG. R. Chen, Sens. Actuators, B, 2010, 148, 240–246.

42 Z. P. Zhnag, H. P. Sun, X. Q. Shao, D. F. Li, H. D. Yu andM. Y. Han, Adv. Mater., 2005, 17, 42–47.

43 J. H. Ryu, S. Y. Bang, W. S. Kim, G. S. Park, K. M. Kim,J. W. Yoon, K. B. Shim and N. Koshizaki, J. Alloys Compd.,2007, 441, 146–151.

44 O. Lupan, L. Chow, L. K. Ono, B. R. Cuenya, G. Y. Chai,H. Khallaf, S. Park and A. Schulte, J. Phys. Chem. C, 2010,114, 12401–12408.

45 L. Armelao, M. Fabrizio, S. Gialanella and F. Zordan, ThinSolid Films, 2001, 394, 89–95.

46 C. Hu, Y. Q. Lan, J. H. Qu, X. X. Hu and A. M. Wang, J. Phys.Chem. B, 2006, 110, 4066–4072.

47 Z. C. Wu, C. Pan, Z. Y. Yao, Q. R. Zhao and Y. Xie, Cryst.Growth Des., 2006, 6, 1717–1719.

48 J. Li, H. Q. Fan and X. H. Jia, J. Phys. Chem. C, 2010, 114,14684–14691.

49 J. G. Yu, H. Guo, S. A. Davis and S. Mann, Adv. Funct. Mater.,2006, 16, 2035–2041.

50 D. J. Wang, D. S. Li, L. Guo, F. Fu, Z. P. Zhang and Q. T. Wei,J. Phys. Chem. C, 2009, 113, 5984–5990.

51 K. M. Kim, H. R. Kim, K. I. Choi, H. J. Kim and J. H. Lee, Sens.Actuators, B, 2011, 155, 745–751.

4162 | J. Mater. Chem. C, 2013, 1, 4153–4162

52 D. Calestani, M. Zha, R. Mosca, A. Zappettini, M. C. Carotta,V. Di Natale and L. Zanotti, Sens. Actuators, B, 2010, 144, 472–478.

53 Y. Zheng, T. Zhang and L. Qiao, Mater. Lett., 2009, 63, 843–846.

54 R. C. Singh, O. Singh, M. P. Singh and P. S. Chandi, Sens.Actuators, B, 2008, 135, 352–357.

55 N. V. Hieu and N. D. Chien, Phys. B, 2008, 403, 50–56.56 J. Zhang, X. H. Liu, X. Z. Guo, S. H. Wu and S. R. Wang,

Chem.–Eur. J., 2010, 16, 8108–8116.57 J. Z. Ou, M. H. Yaacob, M. Breedon, H. D. Zheng,

J. L. Campbell, K. Latham, J. D. Plessis, W. Wlodarski andK. Kalantar-Zadeh, Phys. Chem. Chem. Phys., 2011, 13,7330–7339.

58 Z. Gergintschew, H. Forster, J. Kositza and D. Schipanski,Sens. Actuators, B, 1995, 26, 170–173.

59 J. Cerda Belmonte, J. Manzano, J. Arbiol, A. Gracia,J. Puigcorbe, A. Vila, N. Satate, I. Gracia, C. Cane andJ. R. Morante, Sens. Actuators, B, 2006, 114, 881–892.

60 P. P. Sahay, J. Mater. Sci., 2005, 40, 4383–4385.61 Z. X. Yang, Y. Huang, G. N. Chen and Z. P. Guo, Sens.

Actuators, B, 2009, 140, 549–556.62 S. W. Choi, J. Y. Par and S. S. Kim, Nanotechnology, 2009, 20,

465603.63 Y. Zhang, J. Q. Xu, Q. Xiang, H. Li, Q. Y. Pan and P. C. Xu, J.

Phys. Chem. C, 2009, 113, 3430–3435.

This journal is ª The Royal Society of Chemistry 2013