Journal of Solid State Chemistry 226 (2015) 192–200
Contents lists available at ScienceDirect
Journal of Solid State Chemistry journal homepage: www.elsevier.com/locate/jssc
Ammonia sensing properties of V-doped ZnO:Ca nanopowders prepared by sol–gel synthesis E. Fazio a, M. Hjiri b,c, R. Dhahri b,c, L. El Mir b,d, G. Sabatino a, F. Barreca a, F. Neri a, S.G. Leonardi c, A. Pistone c, G. Neri c,n a
Dipartimento di Fisica e di Scienze della Terra, Università di Messina, Viale Ferdinando Stagno d’Alcontres 31, I-98166 Messina, Italy Laboratory of Physics of Materials and Nanomaterials Applied at Environment, Faculty of Sciences in Gabes, 6072 Gabes, Tunisia c Dipartimento di Ingegneria Elettronica, Chimica e Ingegneria Industriale, Università di Messina, Viale Ferdinando Stagno d’Alcontres 31, I-98166 Messina, Italy d Al Imam Mohammad Ibn Saud Islamic University (IMSIU), College of Sciences, Department of Physics, Riyadh 11623, Saudi Arabia b
art ic l e i nf o
a b s t r a c t
Article history: Received 18 November 2014 Received in revised form 13 February 2015 Accepted 25 February 2015 Available online 5 March 2015
V-doped ZnO:Ca nanopowders with different V loading were prepared by sol–gel synthesis and successive drying in ethanol under supercritical conditions. Characterization data of nanopowders annealed at 700 1C in air, revealed that they have the wurtzite structure. Raman features of V-doped ZnO:Ca samples were found to be substantially modified with respect to pure ZnO or binary ZnO:Ca samples, which indicate the substitution of vanadium ions in the ZnO lattice. The ammonia sensing properties of V-doped ZnO:Ca thick films were also investigated. The results obtained demonstrate the possibility of a fine tuning of the sensing characteristics of ZnO-based sensors by Ca and V doping. In particular, their combined effect has brought to an enhanced response towards NH3 compared to bare ZnO and binary V-ZnO and Ca-ZnO samples. Raman investigation suggested that the presence of Ca play a key role in enhancing the sensor response in these ternary composite nanomaterials. & 2015 Elsevier Inc. All rights reserved.
Keywords: Sol–gel V-doped ZnO:Ca NH3 sensor
1. Introduction In recent years much effort has been devoted to the research on semiconducting metal-oxides based materials, which has led to many promising applications ranging from photovoltaics, photocatalysis and gas sensing [1–4]. For this latter application, metal oxides such as TiO2, SnO2 and ZnO were found to be very useful in developing sensors for combustible and toxic gases, mainly for the control of environmental pollution, home and industrial safety, etc. [5–10]. In particular, ZnO-based sensing devices have been proposed for real-time monitoring of a variety of gases such as hydrocarbons, oxygen, CO, H2 and NO2 species [11–20]. Although ZnO display good response to these gases, enhancing the sensing characteristics of ZnO has drawn much attention from researchers. The most popular strategies employed to enhance sensor performance are: (a) the control of material’s stoichiometry and morphology in order to increase the effective area of gas adsorption [21] and, (b) the use of additives which can promote the gas target reactions on the surface of ZnO semiconductor [13]. ZnO can be doped with different additives. For example, Huang
n
Corresponding author. E-mail address:
[email protected] (G. Neri).
http://dx.doi.org/10.1016/j.jssc.2015.02.021 0022-4596/& 2015 Elsevier Inc. All rights reserved.
et al. studied the doping process of ZnO with many elements [22]. Group III and IV elements such as Al, Ga, In and Sn are the most largely used. When these dopant atoms incorporate into ZnO materials, they replace Zn host atoms, which can release free electrons and contribute to higher conductivity or improved carrier mobility in ZnO materials and enhanced gas sensing properties [23,24]. In this work, we report a study about V- and Ca-doped ZnO nanohybrids as sensing materials for resistive gas sensors. Vdoping has been previously reported to enhance the sensitivity of metal oxides such as WO3 and TiO2 to NH3 and SO2, respectively [25,26]. As regards Ca-doping, literature data evidenced the enhanced stability of some Ca-doped metal oxides [27]. Furthermore, incorporating Ca2þ ions into ZnO lattice induces a reduction in near band edge emission and hence an enhancement in the oxygen vacancies which could be advantageous for gas sensing [28]. No data are instead available in literature about the ternary Vdoped ZnO:Ca system, in which both V and Ca acts as co-dopant. Previously, we prepared Ca doped ZnO nanoparticles by a sol–gel method [29]. Vanadium-doped ZnO samples were also prepared by a solid state reaction [30]. Taking into account these previous studies, we tried to prepare the ternary V-doped ZnO:Ca system. The synthesis by sol–gel route, the characterization of the V-doped ZnO:Ca nanoparticles obtained and their sensing properties in the detection of ammonia are here reported.
E. Fazio et al. / Journal of Solid State Chemistry 226 (2015) 192–200
The detection of ammonia is highly desirable in environmental gas analysis, the automotive industry, the chemical industry and for medical applications [31]. Very low concentrations of ammonia, around 1 ppm, are need to be detect for biomedical applications and in environmental pollution control (for example short term and long term exposure limits of NH3 are been set at 35 and 25 ppm, respectively). However, accurate, reliable, and continuous ammonia monitors are required also to detect higher concentration of this gas as for example in personal protective equipment for employees who work regularly with anhydrous ammonia and are subject to overexposure either to the liquid or the vapor [32]. Among the other sensor typologies, resistive sensors based on metal oxide semiconductors are the most simple and versatile. They operate on the principle of conductance change due to chemisorption/reaction of ammonia gas molecules on the semiconducting sensing layer [33]. It is well known that the behavior of resistive sensors is strongly dependent on the size and shape of the semiconducting material, the presence of dopant, etc. [34]. Then, the gas sensing study undertaken was aimed to establish the key factors influencing the ammonia sensitivity of the novel Vdoped ZnO:Ca samples synthesized.
2. Experimental 2.1. Synthesis Vanadium doped zinc oxide-Ca nanopowders with constant loading of calcium and varying amount of vanadium (viz. 1, 2, 3 and 4.0 at %) were prepared by a sol–gel method, following the preparation procedure already adopted by El Mir et al. [35], using 16 g of zinc acetate dihydrate [Zn(CH3COO)2 2H2O] as a precursor in 112 ml of methanol. After 10 min of magnetic stirring at room temperature, an adequate quantity of ammonium metavanadate corresponding to V/Zn atomic ratios of 0.02, 0.03, 0.04 and 0.05 and an adequate quantity of calcium chloride–hexahydrate [CaCl2 6H2O] corresponding to Ca/Zn atomic ratios of 0.04 were added. After 15 min under magnetic stirring, the solution was placed in an autoclave and dried in supercritical conditions with ethyl alcohol (EtOH). The obtained powder was then heated in a furnace for 2 h at 700 1C in air. Reference samples of ZnO doped with calcium or vanadium were also prepared. In the following, the nanopowders are labelled as ZnO:Ca, ZnO:V and ZnO: CaV1, ZnO:CaV2, ZnO:CaV3 and ZnO:CaV4.
193
XPS spectra were acquired using a K-Alpha system of Thermo Scientific, equipped with a monochromatic AlKα source (1486.6 eV) and operating in constant analyser energy (CAE) mode with a pass energy of 200 eV and 50 eV, respectively, for survey and high resolution spectra, and a spot size of 400 mm. The binding energy shifts were calibrated keeping the C1s position fixed at 285 eV. Surface charging effects were avoided using an electron flood gun. The analysis procedure was performed with the Avantage software of K-Alpha system and every core level photoemission peak was deconvoluted with Gauss–Lorentzian shape functions with the same FWHM (1.4–1.7) for all the considered subbands. Further details about instrumentation utilized and fitting procedure of XPS spectra can be found elsewhere [36]. Raman spectra, excited by a 532 nm YAG laser, were recorded using a Jobin–Yvon–Horiba Raman spectrometer equipped with an Olympus BX40 confocal microscope. The scattered light was collected, in backscattering geometry, by the microscope and dispersed by a Horiba XploRA monocromator, equipped with a 600 line mm grating, and using a Peltier charge-coupled device array as the sensor. The spectra were acquired in the 200– 2000 cm 1 spectral region. 2.3. Sensing tests Sensor devices were fabricated mixing V-doped ZnO:Ca samples with water to obtain a paste and then printing it on alumina substrates (3 mm 6 mm) supplied with interdigited Pt electrodes and heating element. Heating element controls very precisely the operating temperature of the sensing devices, ensuring high stability of the sensor signal. Before sensing tests, the sensor was conditioned in air for 2 h at 300 1C. Electrical measurements were carried out in the working temperature range from 50 to 300 1C. Sensing tests were performed in a lab apparatus which allows to operate at controlled temperature and to perform resistance measurements while varying the ammonia concentration in the carrier stream. Measurements were performed under a dry synthetic air total stream of 100 sccm, collecting the sensors resistance data in the four points mode by means of an Agilent 34970 A multimeter. The gas response is defined as the ratio S ¼R0/R, where R0 represents the electrical resistance of the sensor in dry air and R the electrical resistance at different ammonia concentration.
3. Results and discussion 2.2. Characterization 3.1. Samples morphology and composition SEM images of samples surface were acquired with a FEI Inspect S instrument, coupled with an Oxford INCA PentaFETx3 EDX spectrometer, with a resolution of 137 eV at 5.9 keV (Mn Kα1) and equipped with a nitrogen cooled Si(Li) detector. The spectral data were acquired at a working distance of 10 mm with an acceleration voltage of 20 kV, counting times of 60 s, with approximately 3000 counts/s (cp). The results were processed by the INCA Energy software. Morphology and spatial homogeneity of the nanocomposites were further investigated by transmission electron microscopy (TEM) using a JEOL JEM 2010 microscope operating at 200 kV microscope operating at 200 kV. XRD patterns were recorded in the 2θ range from 201 to 801 using a Bruker a Philips X-Pert diffractometer with a Ni β-filtered CuKα radiation (1.54178 Angstrom). The average crystallite size, d, has been estimated by means of the Scherrer equation: d¼
0:9λ B cos θB
where λ is the X-ray wavelength, θB is the maximum of the Bragg diffraction peak (in radians) and B is the full width at half maximum (FWHM) of the (1 0 1) XRD peak.
The morphology of the V-doped ZnO:Ca samples was investigated by SEM. In Fig. 1 are reported some typical images taken from these samples, showing round particles with dimension between 2 and 5 μm, having a rough porous fine-grained microstructure. EDX pattern is also reported, evidencing the presence of Zn, O, Ca and V. Fig. 2a–d report a series of TEM pictures acquired from these samples. On all samples, single grains having size in the 50– 200 nm range were observed. HRTEM image shows well defined fringes indicating a high crystallinity of the formed grains by this method. Occasionally, the presence of nanofibers, has been also noticed (see Fig. 2a). As these one-dimensional structures were not observed on the pure ZnO and ZnO:V samples, we hypothesize that they are likely due to presence of separated phases rich in Ca. 3.2. Samples microstructure XRD analysis has been carried out to investigate the microstructure of ZnO:CaV powders (Fig. 3). All samples are highly
194
E. Fazio et al. / Journal of Solid State Chemistry 226 (2015) 192–200
Fig. 1. SEM images and EDX patterns of some V-doped ZnO:Ca samples. (a) ZnO:Ca; (b) ZnO:CaV4.
crystalline. The main XRD reflection peaks (indicated by the red square symbols in the figure) are located in the 32–401 range and are assigned to the (1 0 0), (0 0 2) and (1 0 1) reflections of ZnO hexagonal wurtzite crystal structure. On the other hand, some other weak features can be observed in the XRD spectra (see the green line symbol in the figure). According to JCPDS database [37], they were identified and attributed to the basic zinc chloride (ZnCl2 4Zn[OH]2 and to basic salts of Zn such as Zn5(OH)8Cl2–H2O. The average crystallite size, d, has been estimated by means of the Scherrer equation from the full width at half maximum (FWHM) of the (1 0 1) diffraction peak. The d values obtained were around 90 nm for all the samples, in fairly agreement with TEM results. XPS measurements were carried out in order to obtain information about the surface composition of the samples and the charge state of vanadium in the ZnO:CaV nanopowders. The XPS wide scans of some representative samples are shown in Fig. 4. The peaks due to Cl, C, Ca, V, O and Zn were identified. The atomic percentage of all the species was estimated taking into account the relative atomic Scofield’s sensitivity factors and the results are shown in Table 1. The calculated O/Zn ratio, considering only the zinc bonded to oxygen, is in the 0.92–1.15 range while V/Ca ratio increases from 0.047 up to 0.115 going from the ZnO:CaV1 sample to the ZnO:CaV4 one. Interestingly, the V/Ca ratio determined by XPS is lower than based on the nominal composition, indicating an enrichment of calcium on the surface with respect to vanadium. High resolution XPS spectra for Zn2p, O1s, V2p and Ca2p regions are shown in Fig. 5. All the peaks appear to have an evident asymmetry, suggesting the presence of many components. The core line of Zn2p3/2 (Fig. 5a) was deconvoluted using three Gauss–Lorentzian subband and a Shirley-type background
function. The component at 1021.8 70.1 eV corresponds to the ZnO bulk [38,39], that at 1022.870.1 eV can be ascribed both to zinc hydroxide and to zinc carbonate [38] while the high binding energy component at 1023 7 0.1 eV can be ascribed to zinc chloride due to the presence of residual chlorine [40,41]. Finally, the XPS line-shapes of the ZnO:CaV3 and ZnO:CaV4 samples are well reproduced adding a new component, assigned to the zinc acetate dehydrate [42]; also this contribution can be considered a residue of the preparation method. The O1s spectrum (Fig. 5b), has been decomposed into three contributions at about 530.5, 531.9 and 533.1 eV. The component with low binding energy (530.5 70.1 eV) corresponds to O2 on normal wurtzite structure of hexagonal Zn2 þ ion array, surrounded by Zn (or the substitution of V) atoms with their full complement of nearest-neighbor O2 ions. Its intensity is a measure of the amount of oxygen atoms in a fully oxidized stoichiometric surrounding [43]. The second peak centered at 531.970.1 eV can be attributed to the O2 ions in the oxygendeficient regions within the ZnO matrix, whose intensity partly would represent the variation in terms of oxygen vacancies concentration [43,44]. The peak at 533.1 70.3 eV can be ascribed to the chemisorbed oxygen of the surface, such as absorbed H2O or absorbed O2, as well as carboxylate- or carbonate-type species [38,40,43,45]. As in the Zn2p lineshape, another contribution (located at about 535.5 71 eV) and due to the zinc acetate dehydrate [42] is added to fit the data of the ZnO:CaV3 and ZnO: CaV4 samples. The analysis of V2p (Fig. 5c) line-shape is complicated by the presence of O satellite peak in the same binding energy range [46]. To correctly determine the vanadium atomic concentration in the co-doped samples and define the vanadium charge state, we have correlated the O1s and O satellite peaks
E. Fazio et al. / Journal of Solid State Chemistry 226 (2015) 192–200
195
Fig. 2. TEM pictures of the samples. (a) ZnO:Ca; (b) ZnO:V; (c) ZnO:CaV4; (d) HRTEM images from one of the hybrid nanoparticles.
through their binding energy difference and the areas ratio for an undoped sample (not reported here). Then, we have deconvoluted the core line of the doped samples with two components: a doublet peak and a single peak for the V2p and O satellite contributions, respectively. This latter was built starting from the O1s peak, keeping fixed the ratio of the areas to 0.125 while the binding energy difference was fixed to the value of 13.2 70.1 eV. Moreover, the ratio of the area of the two doublet peaks A (V2p3/2)/A (V2p1/2) was kept fixed at 0.5, as well as the binding energy difference Eb ¼ Eb(V2p1/2) Eb(V2p3/2) at 7.770.2 eV. From this analysis, we verified that the peak position of the individual 2p3/2 component is located at 517.3 70.2 eV and, thus, it can be attributed to the V2O5 phase. No other oxidation phase is evident for the vanadium. Moreover, the Ca2p shape (see Fig. 5d) was deconvoluted with three doublet, located at 346.1, 347.1 and 348.3 70.1 eV, to take into account of the presence of three species, namely CaO, CaCO3 and CaCl2 [40]. Finally, the C1s band is decomposed into three components (not shown). The C–C contribution located at 284.970.1 eV, the middle peak at 287.070.1 eV due to the C–O bonds, while the peak at high binding energy at about 289.7 70.1 eV can be associated to carboxylates and/or zinc carbonates (ZnCO3). Also in this case, the latter contributions are likely due to the residue of
the zinc acetate dehydrate and/or to contamination for sample’s air exposure [45,47]. As suggested by Ballerini et al. [38], having taken into account of the ZnCO3 contribution in the C1s deconvolution helps to ascribe the third component of the O1s peak to carboneous species. The Raman spectra are shown in Figs. 6 and 7. In Fig. 6 are reported the spectra of binary ZnO:Ca and ZnO:V samples. We outline that the wurzite ZnO structure belongs to the space group C46v. Thus, the Raman active zone center optical phonons predicted by the group theory are: A1 þ 2B1 þE1 þ2E2. The phonons of A1 and E1 symmetry are polar phonons and they exhibit different wavenumbers for the transverse optical (TO) and longitudinal optical (LO) phonons. Non-polar phonon modes with symmetry E2 have two wavenumbers: E2 (high) mode is associated with oxygen atoms and E2 (low) mode is associated with Zn sub-lattice. Then, the B1 modes are infrared and Raman inactive modes (namely silent modes) [48]. Raman spectrum of ZnO:Ca sample is characterized by all the main ZnO vibrational modes: a peak at 325 cm 1 assigned to the transverse optical (TO) phonons of the A1 mode and related to ZnO structural defects like oxygen vacancies, Zn interstitials and also their complexes; a peak at 430 cm 1 assigned to the E2 (high) Raman vibrational mode typical of the ZnO bulk in the wurtzite
196
E. Fazio et al. / Journal of Solid State Chemistry 226 (2015) 192–200
Fig. 3. XRD spectra of (a) ZnO:Ca, (b) ZnO:CaV2, (c) ZnO:CaV3 and (d) ZnO:CaV4 samples. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 1 Surface elemental composition (from XPS) of the investigated samples. XPS (at%)
ZnO ZnOCa ZnOCaV1 ZnOCaV2 ZnOCaV3 ZnOCaV4
Fig. 4. Wide scans of some representative samples.
structure and two contributions at about 585 cm 1 and 656 cm 1, assigned to the longitudinal optical (LO) phonons of the A1 mode and to the multiple-phonon Raman processes, respectively. Further, we outline the absence of the ZnO Raman feature centred at about 379 cm 1 and assigned to the E1(TO) mode. On the other hand, the Raman-active υ1 and υ4 symmetrical stretching vibration of carbonate group (i.e. CaCO3 and Ca(OH)2 species) centred at around 385, 708, and 1090 cm 1 and the contribution at 765 cm 1 due to α(COO) vibrations of Zn(CH3COO)2 are evident together with the ZnO ones [49,50]. The sample ZnO:V shows, along with
Ca
V
O
Zn
V/Ca
O/Zn
0 7.84 4.67 7.18 4.89 3.74
0 0 0.22 0.58 0.38 0.43
45.78 36.75 44.15 40.15 44.17 44.46
49.54 29.73 37.27 33.17 38.70 43.73
– 0 0.047 0.081 0.078 0.115
0.92 1.24 1.18 1.21 1.14 1.02
the ZnO peaks, also the characteristic vanadium oxide features in the 750–1000 cm 1 region [51,52]. In Fig. 7 are shown the Raman spectra of the ternary vanadiumdoped ZnO:Ca samples. They are characterized by some bands located in the 280–375 cm 1 and in the 750–1000 cm 1 region [51,52], together the ZnO ones (see Fig. 7a–d). According to literature data [49], the features (weak here) in the 280– 375 cm 1 are assigned to the bending vibration of the bridging V–O–V (doubly coordinated oxygen). Going from the ZnO:CaV1 to the ZnO:CaV4 samples, the peak at around 325 cm 1 which represents the A1 (TO) mode of ZnO, shifts at lower wavenumber (by indicating the presence of vanadium ions in the ZnO:Ca matrix) while a new contribution (shoulder) at around at 313.6 cm 1 can be observed. This contribution indicates the mode which might be due to the bending vibration of the triply coordinated oxygen (V3–O) bonds [51]. However, the V3–O oxidation state of vanadium should be very weak since the more intense expected peak at around 520 cm 1, assigned to the V3–O
E. Fazio et al. / Journal of Solid State Chemistry 226 (2015) 192–200
197
Fig. 5. Results of sub-bands model fitting procedure for the Zn2p (a), O1s (b), V2p (c) and Ca2p (d) core level photoemission peaks.
Fig. 6. Raman spectrum of (a) ZnO:Ca and (b) ZnO:V samples.
stretching mode resulting from edge-shared oxygens in common to three pyramids, is not clearly evident. Further, all the investigated samples show always the E2 (high) mode, indicating that the
wurtzite structure of ZnO is still maintained after the vanadium oxide doping. The modes towards higher wavenumber side (750– 1000 cm 1) arise after the vanadium doping, i.e. they were absent
198
E. Fazio et al. / Journal of Solid State Chemistry 226 (2015) 192–200
Fig. 7. Raman spectra of the ZnO:CaV nanopowders.
in the ZnO:Ca undoped sample. Particularly, the barely peak between 770 and 790 cm 1 is generally assigned to the doubly coordinated oxygen (V2–O) stretching mode which results from corner-shared oxygens in common to two pyramids. The presence of these vibrations indicates the layer-like structure of V2O5 phase. The prominent peaks at 850 cm 1 and 865 cm 1 are related to a particular configuration of the vanadium pentoxide, generally associated to host matrix disorder. The presence of polarizability disorder on the basic V2O5 units, induces an enhancement of the Raman intensity from the vibrations of such units [52]. A useful comparative system is given by the surface structures of V2O5 mixed with other oxides supported on alumina catalysts [53], where a strong Raman band in the 800–900 cm 1 range can be observed. Moreover, a sharp peak can be envisaged nearly at 930 cm 1 attributed to V4 þ ¼O bonds. The V4 þ ¼O bonds are due to a direct conversion from V5 þ ¼ O bonds and/or breaking of the single oxygen bonds involving V4 þ ions [54]. Another evidence of the occurrence of this conversion is given considering that no Raman peak is evident at 1027 cm 1, indicating the absence of the V5 þ ¼ O stretching mode of terminal oxygen atoms, generally present on the surfaces of cluster in amorphous vanadium-based films [55]. In comparison to the sample without vanadium, it can be observed a change of the broad band in the 1085–1155 cm 1 region, previously assigned to some vibrational modes of both ZnO and CaCO3 composite. In particular, for the ZnO:CaV4 sample, it is evident that the first of these two contributions, centred at around 1089 cm 1, prevails on the other one. At the same time, the two features in the 800–900 cm 1 are well defined and they are characterized by narrower FWHM and increased intensity. Further, a change in the ratio between the intensity of the two main peaks at 800–900 cm 1 with the V/Ca content has been clearly observed. Finally, looking to the modes towards lower wavenumber side
(200–500 cm 1), both the Raman features related to ZnO or vanadium oxide are substantially modified, which should indicate the substitution of vanadium ions in the ZnO lattice. 3.3. Ammonia sensing tests The sensing properties of the synthesized ternary V-doped ZnO:Ca, binary ZnO:Ca and ZnO:V and pure ZnO samples, were investigated. Fig. 8 shows the typical transient response observed with these sensors, at the operating temperature of 250 1C, towards pulses of NH3 at concentration ranging from 250 ppm to 4000 ppm. A reversible behavior has been noted, with a fast decrease of the baseline resistance when ammonia was added to carrier gas passing over the sensors. The long recovery time, suggests a strong adsorption of ammonia on the surface of the sensing layer, so it can be hypothesized that the desorption process of ammonia or its intermediate species from the surface is slower than ammonia absorption one. The observed decrease of resistance can be interpreted on the basis of classical theory of resistance change after interaction with a reducing species on the surface of n-type semiconductors [56,57]. As is well known, this behavior is derived from the reaction of the adsorbed oxygen species on the surface of semiconductor oxides with the test gas molecules. On exposure to reducing gases, the surface reaction between the adsorbed oxygen species and gas molecules will occur. The nature and concentration of chemisorbed oxygen species strongly depend on temperature. At the operating temperature of our sensors, O and O2 adsorbed species are usually chemisorbed. Consequently, surface reactions, as that below described, can occur on the surface of the sensing layer: 2NH3 þ10O(ad) -2NO þ3H2Oþ 10e
E. Fazio et al. / Journal of Solid State Chemistry 226 (2015) 192–200
250 ppm 500 ppm 1000 ppm 2000 ppm 4000 ppm
10
T = 250°C
0
1000
2000
3000
4000
Time (s) Fig. 8. Variation of the resistance of the sensor ZnO:CaV1 to pulses of NH3 at different concentration. Inset shows the calibration plot.
1
ZnOCaV4
ZnOCaV3
ZnOCaV2
ZnOCaV1
ZnOV
2
ZnOCa
3
ZnO
Sensor response (R0/R)
1000 ppm NH3
Sensor
Fig. 9. Sensor response of the tested sensors at 1000 ppm of NH3.
NH3 behaves as a reducing gas donating electrons back to ZnO conduction band. Therefore, the potential barrier in depletion region is decreased, leading to the increase (decrease) in the conductance (resistance) of the sensor. All the sensors showed this behavior, indicating that the presence of Ca and V dopant did not altered the typical n-type semiconductor behavior of pure ZnO. In the inset of Fig. 8 is reported the calibration curve for this sensor in the ammonia concentration range from 250 ppm to 4000 ppm. It can be noted a linear trend when data are plotted in a log–log graph. In Fig. 9 is reported the response to 1000 ppm of ammonia in air of all sensor investigated and operating all in the same experimental conditions. With respect to pure ZnO taken as a reference, the binary ZnO:Ca shows a poorer response to ammonia. The samples containing vanadium shows, on the average, a higher response. It seems evident then that vanadium is responsible of this enhanced response, in agreement with a previous report [25]. However, it also appears from these data that only vanadium doping is nearly ineffective to enhance the response of the pure ZnO (see the ZnOV sample). The response of ZnO:CaVx sensors appears instead clearly dependent on the V content, and is higher for the ZnO:CaV1 sample. The sensitivity to ammonia for this specific sensor has been evaluated to be around 2.85 10 3 ppm 1. The sensor, although not very sensitive in absolute, display better performance than ones, for example, based on carbon nanotubes which are also more expensive [58].
From the morphological characterization we can exclude that the enhancement in the sensor response is linked to a change in the particle shape. Besides, XRD indicated that the particle size are almost unchanged after the doping process. Then, can be excluded also a correlation with the surface area of the ZnO:CaVx sensing layers. Due to the different ionic radii of Ca2 þ ion (1.14 Ǻ) and of vanadium in its V4 þ or V5 þ valency states (0.72 and 0.68 Ǻ, respectively) [9] compared to that of the Zn2 þ ions (0.88 Ǻ), this presumably should causes a distortion in the ZnO crystal structure. However, XRD data indicated instead that Ca and V likely accommodate in the crystal lattice of ZnO causing only a minimal distortion. This could be due to a sort of compensation effect, considering the opposite variation of ionic radius of V and Ca with respect to Zn. Doping divalent cation Zn sites by V4 þ or V5 þ ions creates a mixed valency in the original ZnO hexagonal structure. The mixed valence creates charge polarity between Zn–O and V–O bonds and this could be responsible of the sensing behavior observed. In order to better clarify the action mechanism of Ca and V in activating the sensor response, we focused our attention on the Raman peaks at 850 cm 1 and 865 cm 1. Depending on Ca and V content, the relative intensity between these two bands changes markedly. At low V content the band at 850 cm 1 has a higher intensity. However, the band at 865 cm 1 increases of intensity increasing the V content. It is well known that these bands are sensitive to local arrangements of V on the surface of host matrix [52,53]. Because the sensing mechanism rely on reactions occurring on specific surface sites, this change of arrangements could give place to different sensing characteristics. In the specific, the band at 850 cm 1, appears to be associated with the most sensitive V sites, as revealed by plotting the sensor response as a function of the I850/I865 ratio, representing the ratio between the intensity of the band at 850 cm 1 and the band at 865 cm 1 (Fig. 10). As the intensity of the above specified Raman bands depends on the relative V/Ca ratio, it can be suggested that the presence of Ca has a key effect in enhancing the sensor response in these ternary composite nanomaterials. A deeper investigation has been undertaken in order to clarify the origin of these Raman bands and the role of Ca in induce the intensity modifications observed. From the practical point of view, we noted the good stability and selectivity of these sensors (see Table 2). Sensor devices have been tested for prolonged times (around one–two months) and no remarkable degradation of the performances has been noted. Further, interferences coming from other reducing gases like CO
3.5
Sensor response (R0/R)
Resistance (MΩ)
100
199
ZnO:CaV1
3.0 2.5 ZnO:CaV3
ZnO:CaV2
2.0 ZnO:CaV4
1.5 ZnO:Ca
1.0
0.0
0.5
1.0
1.5
I850/I865 ratio Fig. 10. Response of the sensor investigated as a function of the I850/I865 ratio.
200
E. Fazio et al. / Journal of Solid State Chemistry 226 (2015) 192–200
Table 2 Sensitivity of the ZnOCaV1 sensor to NH3 at various times and to different interfering gases. Gas/day
Sensitivity (ppm 1) 103
NH3/0 NH3/7 NH3/23 NH3/50 CO/0 CH4/0
2.85 2.58 2.77 2.69 0.17 0.01
and CH4, have been also evaluated. The results have shown the negligible effect of these gases on the sensor response, indicating the good selectivity of the sensors. Furthermore, they appear to be sensitive in a wide range of concentration, compared to other sensors described in literature which have the tendency to saturate almost completely at high concentration. 4. Conclusions In summary, ternary V-doped ZnO samples have been synthesized by a simple sol–gel method. These hybrid nanostructures present the typical the wurtzite structure. Raman spectra revealed clearly the interaction of ZnO with the dopant Raman spectra showed a modification in the ZnO raman modes, which clearly indicate the presence of interactions between the ZnO structure and the dopant. Resistive sensors based on the V-doped ZnO:Ca samples showed also enhanced responses to NH3 compared to binary ones, likely due to a sinergic action between V and Ca. However, further studies are need for a deeper understanding of the dopinginduced sensing properties in these hybrid materials. Acknowledgments The authors are grateful to A.B.A.L. ONLUS Messina (http:// www.abalmessina.it) for the support to Raman characterization. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]
H. Bai, X. Li, C. Hu, X. Zhang, J. Li, Y. Yan, G. Xi, Sci. Rep. 3 (2013) 2204. N. Barsan, D. Koziej, U. Weimar, Sens. Actuators, B 121 (2007) 18–35. S.T. Mckinstry, P. Muralt, J. Electroceram. 12 (2004) 7. K. Nomura, H. Ohta, K. Ueda, T. Kamiya, M. Hirano, H. Hosono, Science 300 (2003) 1269. K.L. Chopra, S. Major, D.K. Pandya, Thin Solid Films 102 (1983) 1. T. Seiyama, A. Kato, K. Fjiishi, M. Nagatani, Anal. Chem. 34 (1962) 1502. A. Jones, T.A. Jones, B. Mann, J.G. Firth, Sens. Actuators 5 (1984) 75. S. Basu, A. Dutta, Sens. Actuators, B 22 (1994) 83. G. Sberveglieri, P. Nelli, S. Groppelli, F. Quaranta, A. Valentini, L. Vasanelli, Mater. Sci. Eng., B 7 (1990) 63. T. Seiyama, A. Kato, K. Fjiishi, M. Nagatani, Anal. Chem. 34 (1962) 1502. A. Jones, T.A. Jones, B. Mann, J.G. Firth, Sens. Actuators 5 (1984) 75. S. Basu, A. Dutta, Sens. Actuators, B 22 (1994) 83. U. Lampe, J. Müller, Sens. Actuators 18 (1989) 269. E. Traversa, A. Bearzotti, Sens. Actuators, B 23 (1995) 181.
[15] J.D. Choi, G.M. Choi, Sens. Actuators, B 69 (2000) 120. [16] C. Baratto, G. Sberveglieri, A. Onischuk, B. Caruso, S. di Stasio, Sens. Actuators, B 100 (2004) 261. [17] E. Fazio, A.M. Mezzasalma, G. Mondio, T. Serafino, F. Barreca, F. Caridi, Appl. Surf. Sci. 257 (2011) 2298. [18] E. Fazio, A.M. Mezzasalma, G. Mondio, F. Neri, R. Saija, Appl. Surf. Sci. 272 (2013) 30. [19] H. Chen, Y. Liu, C. Xie, J. Wu, D. Zeng, Y. Liao, Ceram. Int. 38 (2012) 503. [20] J. Xu, Y. Shun, Q. Pan, J. Qin, Sens. Actuators, B 66 (2000) 161. [21] E. Fazio, A. Cacciola, A.M. Mezzasalma, G. Mondio, F. Neri, R. Saija, J. Quant. Spectrosc. Radiat. Transfer 124 (2013) 86. [22] G.Y. Huang, C.Y. Wang, J.T. Wang, Physica B 405 (2010) 158. [23] J.-H. Lee, B.-O. Park, Thin Solid Films 426 (2003) 94–99. [24] M. Hjiri, L. El Mir, S.G. Leonardi, A. Pistone, L. Mavilia, G. Neri, Sens. Actuators, B 196 (2014) 413. [25] I. Jimenez, A. Vilà, A. Cornet, J.R. Morante, IEEE Sens. J. 1 (2002) 409. [26] D. Morris, R.G. Egdell, J. Mater. Chem. 11 (2001) 3207. [27] B.-Ki Min, S.-Don Choi, Sens. Actuators, B 108 (2005) 119. [28] B. Karthikeyan, T. Pandiyarajan, K. Mangaiyarkarasi, Spectrochim. Acta, A: Mol. Biomol. Spectrosc. 82 (2011) 101. [29] L. El Mir, F. Ghribi, M. Hajiri, Z. Ben Ayadi, K. Djessas, M. Cubukcu, H.J. von Bardeleben, Thin Solid Films 519 (2011) 5787. [30] B. Karthikeyan, T. Pandiyarajan, K. Mangaiyarkarasi, Spectrochim. Acta, Part A 82 (2011) 97. [31] B. Timmer, W. Olthuis, A. Van den Berg, Sens. Actuators, B 107 (2005) 666. [32] Pocket Guide to Chemical Hazards, NIOSH Publications, Pocket Guide to Chemical Hazards, Cincinnati, OH 1994. [33] P.T. Moseley, B.C. Tofield (Eds.), Solid State Gas Sensors, Adam Hilger, Bristol, 1987. [34] G. Korotcenkov, Mater. Sci. Eng., B 139 (2007) 1. [35] L. El Mir, J. El Ghoul, S. Alaya, M. Ben Salem, C. Barthou, H.J. von Bardeleben, Physica B: Condens. Matter 403 (2008) 1770. [36] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray Photoelectron Spectroscopy, Published by Perkin-Elmer Corp., Eden Prairie, MN, USA, 1992. [37] M.C. Morris, H.F. McMurdie, E.H. Evans, B. Paretzkin, H.S. Parker, N. C. Panagiotopoulos, JCPDS Database, Malcolm Baldrige, 1981. [38] L.L. Yang, Q.X. Zhao, M. Willander, X.J. Liu, M. Fahlman, J.H. Yang, Appl. Surf. Sci. 256 (2010) 3592. [39] E. Fazio, S. Patanè, S. Scibilia, A.M. Mezzasalma, G. Mondio, F. Neri, S. Trusso, Curr. Appl. Phys. 13 (2013) 710. [40] 〈http://srdata.nist.gov/xps/Default.aspx〉. [41] M.F. Guimon, G. Pfister-Guillouzo, M. Bremont, W. Brockmann, C. Quet, J. Y. Chenard, Appl. Surf. Sci. 108 (1997) 149. [42] M. Chen, X. Wang, Y.H. Yu, Z.L. Pei, X.D. Bai, C. Sun, R.F. Huang, L.S. Wen, Appl. Surf. Sci. 158 (2000) 134. [43] R.N. Gayen, S. Majumder, S. Varma, A.K. Pal, Mater. Chem. Phys. 123 (2010) 138. [44] F. Hai-Bo, Y.S. Yan, Z.P. Feng, W.H. Yuan, L.X. Lin, J.C. Mei, Z.Q. Sheng, C.Y. Hai, W.Z. Guo, Chin. Phys. Lett. 24 (2007) 2108. [45] K. Ogle Ballerini, M.-G. Barthés-Labrousse, Appl. Surf. Sci. 253 (2007) 6860. [46] Q. Guo, S. Lee, D.W. Goodman, Surf. Sci. 437 (1999) 38. [47] X. Ji, Y. Chen, G. Zhao, X. Wang, W. Liu, Tribol. Lett. 41 (2011) 113. [48] S. Karamat, R.S. Rawat, P. Lee, T.L. Tan, R.V. Ramanujan, W. Zhou, Appl. Surf. Sci. 256 (2010) 2309. [49] B. Leng, F. Jiang, K. Lu, W. Ming, Z. Shao, CrystEngComm 12 (2010) 730. [50] M. Galvan-Ruiz, L. Baños, M.E. R.-Garcia, Sens. Instrum. Food Qual. Saf. 1 (4) (2007) 169. [51] W. Chen, L. Mai, J. Peng, Q. Xu, Q. Zhu, J. Sol. State Chem. 177 (2004) 377. [52] E. Cazzanelli, G. Mariotto, S. Passerini, W.H. Smyrl, A. Gorenstein, Sol. Energy Mater. Sol. Cells 56 (1999) 249. [53] M. Vuurman, I.E. Wachs, J. Mol. Catal. 77 (1992) 29. [54] R.J.H. Clark, The Chemistry of Titanium and Vanadium, Elsevier, New York, NY, 1968. [55] S.H. Lee, H.M. Cheong, M.J. Seong, P. Liu, C.E. Tracy, A. Mascarenhas, J.R. Pitts, S. K. Deb, Solid State Ionics 165 (2003) 111. [56] S. Santangelo, G. Faggio, G. Messina, E. Fazio, F. Neri, G. Neri, Sens. Actuators, B 178 (2013) 473. [57] S. Santangelo, E. Fazio, F. Neri, G. Faggio, G. Messina, G. Neri, J. Phys. D: Appl. Phys. 46 (2013) 125303. [58] N.D. Hoa, N. Van Quy, Y. Cho, D. Kim, Sens. Actuators, B 127 (2007) 447.
