CuO/ZnO Nanocomposite Gas Sensors Developed by a Plasma-Assisted Route

DOI: 10.1002/cphc.201101062

CuO/ZnO Nanocomposite Gas Sensors Developed by a Plasma-Assisted Route Quentin Simon,[a] Davide Barreca,*[b] Alberto Gasparotto,[a] Chiara Maccato,[a] Eugenio Tondello,[a] Cinzia Sada,[c] Elisabetta Comini,[d] Giorgio Sberveglieri,[d] Manish Banerjee,[e] Ke Xu,[e] Anjana Devi,[e] and Roland A. Fischer[e] CuO/ZnO nanocomposites were synthesized on Al2O3 substrates by a hybrid plasma-assisted approach, combining the initial growth of ZnO columnar arrays by plasma-enhanced chemical vapor deposition (PE-CVD) and subsequent radio frequency (RF) sputtering of copper, followed by final annealing in air. Chemical, morphological, and structural analyses revealed the formation of high-purity nanosystems, characterized

by a controllable dispersion of CuO particles into ZnO matrices. The high surface-to-volume ratio of the obtained materials, along with intimate CuO/ZnO intermixing, resulted in the efficient detection of various oxidizing and reducing gases (such as O3, CH3CH2OH, and H2). The obtained data are critically discussed and interrelated with the chemical and physical properties of the nanocomposites.

1. Introduction Among the technological challenges related to the development of innovative gas sensors, the reliable detection of flammable, toxic and explosive species represents a key issue for safety control in both domestic and industrial environments.[1, 2] In this regard, quasi-1D nanostructures based on oxide semiconductors (SCs) have become the object of intensive investigation as potential building blocks for nanoscale sensors, due to the enhanced modulation of their electrical properties upon interactions with adsorbed species.[2–4] To obtain improved sensor response and selectivity, various efforts have been focused on the modification of quasi-1D structures by the introduction of functional activators, such as noble metals (e.g. Au, Pd, Pt) or oxide nanoparticles (NPs).[5–7] This interest is motivated by the unique features derived from low dimensionality and the synergistic interplay between each component.[8–11] In addition, an attractive option is offered by the controlled formation of oxide–oxide p–n junctions, which, due to interfacial electron-transfer processes, can extend the space charge region, thus improving the functional properties.[8, 9] In addition, the different sensing mechanisms of p- and n-type SCs can be advantageously exploited to develop unique material characteristics by tuning the amount and distribution of the constituents.[10–13] Among the potential candidates, cost-effective and nontoxic CuO/ZnO composites represent attractive systems. ZnO, which is an n-type SC with a wide band gap (Eg = 3.4 eV), displays interesting sensing performances towards a broad category of gases.[3, 14–16] In a different way, CuO, which is a p-type SC with a narrow band gap (Eg = 1.2 eV), is a less explored alternative, although it has demonstrated appreciable responses to reducing analytes at moderate temperatures.[17–20] Their combination, yielding CuO/ZnO composites, has led to promising results in H2 and CO detection,[21–26] with performances strongly dependent on the system nano-/micro-organization.[21, 23, 25, 26]

2342

To date, CuO/ZnO composites for gas-sensing applications have been synthesized by thermal treatment of powder samples,[22, 24] liquid-phase methods,[21, 23] hydrothermal routes,[25] and, ultimately, hybrid vapor/liquid processes.[26] Nevertheless, synthetic approaches to CuO/ZnO nanosystems, which combine a high surface-to-volume ratio and an intimate interfacial contact of the components, are still in demand. In particular, major requirements are the use of soft conditions to prevent the formation of Cu–Zn–O ternary phases, which are detrimental to functional performances,[27] and the synthesis of the desired nanostructures supported on suitable substrates, which enables their direct integration in sensor fabrication. [a] Dr. Q. Simon, Dr. A. Gasparotto, Dr. C. Maccato, Prof. E. Tondello Department of Chemistry Padova University and INSTM 35131 Padova (Italy) [b] Dr. D. Barreca CNR-ISTM and INSTM Department of Chemistry Padova University 35131 Padova (Italy) E-mail: [email protected] [c] Dr. C. Sada Department of Physics and CNISM Padova University 35131 Padova (Italy) [d] Dr. E. Comini, Prof. G. Sberveglieri SENSOR Lab Department of Chemistry and Physics Brescia University and CNR-IDASC 25133 Brescia (Italy) [e] Dr. M. Banerjee, Dr. K. Xu, Prof. A. Devi, Prof. R. A. Fischer Lehrstuhl fr Anorganische Chemie II Ruhr-University Bochum 44780 Bochum (Germany) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201101062.

 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemPhysChem 2012, 13, 2342 – 2348

CuO/ZnO Nanocomposite Gas Sensors In this context, the present contribution reports a bottomup approach to CuO/ZnO nanocomposites on polycrystalline alumina. After the initial growth of quasi-1D ZnO structures by PE-CVD, the dispersion of CuO particles was performed by RF sputtering of copper and ex situ annealing in air. The unique activation of cold plasmas in both PE-CVD and RF sputtering processing steps[28] resulted in the formation of tailored material features. A detailed characterization of the chemical, structural, and morphological properties of the nanosystems as a function of the amount of CuO, carried out by means of a multi-technique approach, is presented. In addition, a preliminary investigation of CuO/ZnO sensing properties towards various flammable/toxic gases is reported for the first time, focusing on the most attractive results in view of technological applications.

2. Results and Discussion

sectional FE-SEM observations and EDS line scans along the deposit thickness were carried out (Figure 1, right). As observed, a conformal coverage of Al2O3 particles by columnar ZnO aggregates, growing perpendicularly to the substrate surface, was present for all samples. As revealed by glancing incidence X-ray diffraction (GIXRD, see below), these quasi-1D arrays grew along the < 001 > direction. The mean column diameter (  20 nm) was of the same order of magnitude of ZnO Debye length LD,[14–16] which is a favorable feature to promote complete depletion upon interaction with the analytes, thus yielding improved sensing responses. The column length changed from 260 to 300 nm upon going from sample S-5 to S-15; a result attributed to the increase in CuO content with RF power. Indeed, EDS line scans indicated that CuO was mainly concentrated at the tip of ZnO columns, although indepth penetration also took place, in line with secondary ion mass spectrometry (SIMS) results (see above and Figure 5 below). Quantitative EDS analyses showed a linear increase of

To tailor the overall copper content in the composite systems, three different settings of RF power (5, 10, and 15 W) were used. In the following, samples are indicated as S-P, in which P is the RF power (5, 10, or 15 W) adopted during Cu sputtering. 2.1. Material Characterization The morphological organization of CuO/ZnO composites was investigated by field-emission scanning electron microscopy (FE-SEM). Plane-view images (Figure 1, left) showed that the large globular grains of polycrystalline alumina were uniformly decorated by nanoaggregates, with features directly dependent on the adopted synthetic conditions. For specimen S-5, small and well-defined particles could be discerned, whereas upon increasing the RF power used for Cu deposition, a more compact and smoother surface was obtained (sample S-15). Nonetheless, irrespective of the synthetic conditions, the obtained nanomaterials were characterized by an appreciable porosity; an important issue in view of their gas-sensing utilization. To attain a deeper insight into the mutual in-depth distribution of CuO and ZnO, as well as on the system growth mode, crossChemPhysChem 2012, 13, 2342 – 2348

Figure 1. Plane-view (left) and cross-sectional (right) FE-SEM micrographs for the different CuO/ZnO samples. Energy-dispersive X-ray spectroscopy (EDS) line scans recorded along the lines marked in cross-sectional images. Arrows indicate the direction of abscissa increase.

 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemphyschem.org

2343

D. Barreca et al.

Figure 2. Evolution of the Cu wt % as a function of RF power. The quantification was made by EDS by considering only Cu and Zn signals to avoid contributions of oxygen from the Al2O3 substrate.

the overall Cu amount as a function of the adopted RF power (Figure 2). As a whole, the above results indicated the possibility of controlling CuO content and in-depth dispersion through the correct choice of synthetic conditions. X-ray photoelectron spectroscopy (XPS) surface analyses were performed to attain a detailed characterization of the Zn and Cu chemical environments. Survey spectra (Figure 3)

of the surface Cu/Zn atomic percentage ratio yielded values ranging from 3.0 to 6.0 upon increasing the RF power from 5 to 15 W. This feature, in line with EDS data (compare Figure 2), suggested a direct dependence of the deposited copper amount on the adopted sputtering conditions. The Cu2p photoelectron peak (Figure 3, inset) displayed intense “shake-up” satellites; a clear fingerprint for the presence of Cu(II) species.[29] This feature, confirmed by the copper Auger parameter (see the Experimental Section; a  1851.6 eV), provided evidence for the presence of CuO,[29, 30] demonstrating full oxidation of the deposited copper species after thermal treatment. As a whole, XPS results suggested the formation of CuO/ZnO composites. Finally, it is worth noting that the carbon signal fell to noise level after 5 min of Ar + erosion, indicating that contamination was due to atmospheric exposure and was limited to the sample surface. The in-depth chemical composition of CuO/ZnO materials was investigated by SIMS analysis, with particular attention to the copper distribution along the ZnO thickness. In agreement with XPS results, the presence of carbon (signal not reported) was limited to the sample surface, confirming the high-purity of the obtained specimens. Depth profiles of samples S-5 and S-15 are displayed in Figure 4. As observed, the outermost sample region was Curich, a phenomenon particularly evident at higher RF power, that is, higher copper loading (sample S-15). Upon increasing the sputtering time, a decrease in the Cu ionic yield and a concomitant increase in the Zn signal up to a stable value took

Figure 3. Surface XPS survey spectrum of the CuO/ZnO S-5 specimen. The Cu2p photoelectron peak is shown in the inset. BE = binding energy.

showed the sole presence of Zn, Cu, O, and C. Irrespective of the adopted RF power, the presence of zinc signals was systematically detected, indicating only a partial coverage of ZnO matrices by copper species. All specimens possessed Zn Auger parameters (see the Experimental Section; a  2010.0 eV) in agreement with the value reported for pure ZnO.[14] Calculation

2344

www.chemphyschem.org

Figure 4. SIMS depth profiles for specimens S-5 (top) and S-15 (bottom).

 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemPhysChem 2012, 13, 2342 – 2348

CuO/ZnO Nanocomposite Gas Sensors place. This phenomenon indicated an efficient dispersion of CuO NPs throughout the ZnO matrix. Finally, both copper and zinc signals underwent a drop-off at the interface with the alumina substrate. Irrespective of the processing conditions, the O ionic yield remained almost constant throughout the deposit thicknesses, in line with the co-presence of CuO and ZnO observed by XPS. The system structure and crystallinity were characterized by GIXRD. In addition to alumina signals,[31] the patterns of the obtained samples (Figure 5) revealed the sole presence of the

Figure 6. Responses as a function of working temperature for sample S-5 towards CH3CH2OH (500 ppm), CO (500 ppm), H2 (1000 ppm), CH3COCH3 (100 ppm), CH4 (50 ppm), and O3 (70 ppb).

Figure 5. GIXRD patterns of CuO/ZnO nanosystems. Al2O3 substrate reflections are marked by #.

zinc oxide wurtzite structure, with characteristic reflections at 2q = 34.5, 36.2, and 47.68.[32] The strong intensity of the (002) ZnO peaks was attributed to anisotropic crystallite growth along the < 001 > direction, in line with previous literature reports.[14, 15, 33, 34] The absence of appreciable reflections for CuO could be related to its high dispersion and/or low crystallite size. As a general rule, no appreciable variation in the ZnO peak positions occurred upon Cu sputtering/annealing, suggesting the absence of significant structural modification of the pristine matrix and enabling to discard the formation of Zn–Cu–O ternary phases. 2.2. Gas-Sensing Performances Preliminary gas-sensing tests were performed as a function of CuO loading in the CuO/ZnO nanocomposites (see Figure S1 in the Supporting Information). In the following, the results obtained for sample S-5, exhibiting the best performances, are presented and discussed in detail. Figure 6 displays a general overview of sample S-5 response values to selected concentrations of various analytes. As observed, a monotonic response increase with working temperature was detected for all of the tested reducing gases, that is, CH3CH2OH, CH3COCH3, H2, CO, and CH4, due to enhanced reactions between analytes and adsorbed oxygen species at higher temperatures [see below, Eq. (1)–(3)]. Interestingly, the responses provided by the present CuO/ZnO sensors to ethanol, ChemPhysChem 2012, 13, 2342 – 2348

acetone, and hydrogen were appreciably higher than those reported for ZnO and ZnO–TiO2 materials,[10, 14, 15] and comparable to those yielded by Ag/ZnO systems.[35] These results suggested that CuO/ZnO nanocomposites present an improved detection efficiency to the target gases with respect to bare ZnO, an important issue for eventual technological applications.[36] To this aim, the sensor selectivity is a further main concern, since many analytes can simultaneously interact with the sensing element under real-world conditions.[15] In the present case, the use of working temperatures between 100 and 300 8C favored the discrimination of ozone with respect to reducing gases, since the response to O3 exceeded the corresponding values towards the other analytes by more than one order of magnitude (Figure 6). As regards reducing species, whose detection efficiency was enhanced at 4008C, responses to ethanol were always higher than those to CO and H2. As a whole, the above features indicated a certain selectivity of the present CuO/ZnO nanosystems, depending on the correct choice of the operational conditions. Ozone detection indicated a change in the sensitivity and selectivity pattern, since the response exhibited maximum-like behavior as a function of the working temperature, the best operating conditions being at 300 8C (Figure 6). This trend is qualitatively similar to that reported by us for Ag/ZnO systems,[35] despite the response values achieved by CuO/ZnO nanocomposites exceeded the previous ones by one order of magnitude. To our knowledge, these values are the best ever obtained for ozone detection by ZnO-based sensors, highlighting the potential of the present systems as highly efficient sensors for practical utilization. To attain a deeper insight into the functional behavior of the system, attention was subsequently focused on the study of dynamic responses (see Figure 7 for ozone and ethanol). The sensor displayed n-type behavior, with a conductance decrease (increase) upon contact with oxidizing (reducing) gases.[4, 15] Based on the above characterization data, such features can be related to the high dispersion of CuO particles over ZnO SC

 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemphyschem.org

2345

D. Barreca et al.

Figure 8. Response of sample S-5 as a function of analyte concentrations: a) O3 (working temperature = 200 8C); b) CH3CH2OH and H2 (working temperature = 400 8C). Figure 7. Isothermal dynamic responses of specimen S-5 to square concentration pulses of ozone (top) and ethanol (bottom). Working temperatures = 200 and 400 8C, respectively.

matrices, constituting the percolation network.[37] In this case, the mechanism for gas detection involves the initial chemisorption of O2 from air on the system surface, generating active oxygen species, such as O [Eq. (1)]:[14, 15, 38] O2 ðgÞ þ 2e Ð 2 O ðadsÞ

ð1Þ

Subsequently, the interaction with oxidizing [Eq. (2)] or reducing [Eq. (3)] species leads to the reactions shown:[14, 15, 38] O3 ðgÞ þ e Ð O ðadsÞ þ O2 ðgÞ

ð2Þ

CH3 CH2 OHðgÞ þ 6 O ðadsÞ Ð 2 CO2 ðgÞ þ 3 H2 OðgÞ þ 6e

ð3Þ

On this basis, contact of the sensor with a strong oxidant (ozone in the present case) results in electron withdrawal [Eq. (2)], which is responsible for the ultimate conductance decrease. In a different way, the interaction with a reducing gas, such as ethanol [Eq. (3)], releases electrons into the system conduction band, increasing the charge carrier concentration and yielding the observed conductance enhancement (Figure 7). Careful inspection of the dynamical responses shows that the conductance displayed a rapid variation at the beginning of each gas pulse, followed by a slower evolution up to the end of the pulse. This behavior, which was more pronounced in the case of ozone, suggested that the rate-limiting step of

2346

www.chemphyschem.org

the overall sensing process was analyte adsorption on the sensor surface.[14] Further efforts were devoted to analyzing the sensor response as a function of gas concentration. Representative data for ozone, ethanol, and hydrogen are plotted in Figure 8. Interestingly, over the tested range, the sensors presented a response, S, proportional to the analyte concentration, C, following the power law [Eq. (4)]: S ¼ A  ½CB

ð4Þ

in which A is a constant typical of the sensing element and B, with a value usually equal to 1 or 0.5, depends on the charge of the surface species and the stoichiometry of the reactions involved.[14, 15, 35, 39] The observed linear trends, indicating the absence of appreciable saturation phenomena, are of interest in view of possible quantitative applications of the present sensing devices. The responses of the nanosystems obtained suggest a superior potential for technological applications. This result is correlated to the peculiar quasi-1D system morphology, with a high surface-to-volume ratio and ZnO column diameters comparable to 2LD. As already mentioned, the latter feature leads to electrical properties highly sensitive to the surrounding medium, since a significant portion of aggregates can be completely depleted (Figure 9). Additionally, the high dispersion of CuO nanoparticles, producing intimate CuO/ZnO intermixing, is likely to promote an enhanced detection efficiency, thanks to cooperative effects in the reactivity of the two oxides. The

 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemPhysChem 2012, 13, 2342 – 2348

CuO/ZnO Nanocomposite Gas Sensors

Figure 9. a) Schematic representation of the columnar CuO/ZnO arrays on polycrystalline Al2O3, showing the percolation path through the ZnO network. b) Sketch of a single CuO/ZnO column, illustrating depletion due to interactions with adsorbed species, such as O , and the formation of p-CuO/nZnO interfaces.

electron-transfer processes occurring at p-CuO/n-ZnO heterojunctions can in fact improve charge carrier separation and lifetime, thus promoting reactions involved in the sensing process (see above) and resulting in enhanced conductance/resistance modulation upon interaction with the analytes.

3. Conclusions An innovative approach to nanostructured CuO/ZnO composites supported on polycrystalline Al2O3 has been developed. After the growth of ZnO columnar arrays by PE-CVD, CuO particles were deposited by RF sputtering of copper, followed by ex situ annealing in air. The process enabled the synthesis of high-purity CuO/ZnO systems, characterized by tailored morphological features. Interestingly, control of the RF power during copper sputtering allowed simultaneous modulation of the CuO/ZnO ratio and CuO dispersion, resulting in an intimate intermixing of the two components. Such features, along with the high surface-to-volume ratio of ZnO matrices, enabled the development of highly efficient sensors, in particular, for ozone and ethanol. Interestingly, the use of working temperatures  300 8C favored the discrimination of ozone with respect to the above-reported reducing gases; an important point for the development of selective gas sensors. Further experiments will be devoted to a more detailed investigation of the system selectivity as a function of the operating conditions. Basing on the results presented herein, appealing perspectives for the development of this work also concern further studies of the system behavior as a function of CuO loading in ZnO matrices. In addition, the high versatility of the proposed approach can be extended to the synthesis of other kinds of nanosystems with tailored characteristics, in view of further improvements in the sensor efficiency and selectivity.

Experimental Section Synthesis CuO/ZnO nanocomposites were prepared on suitably precleaned[14] polycrystalline Al2O3 substrates (3  3 mm2 ; thickness = 250 mm; average surface roughness = 70 nm) through a two-step RF sputtering/PE-CVD process. First, columnar ZnO arrays were grown by the PE-CVD technique using a bis(ketoiminato) zinc(II) precursor, Zn[CH3O(CH2)2NC(CH3)=CHC(CH3)=O]2.[40] The precursor, vaporized ChemPhysChem 2012, 13, 2342 – 2348

at 150 8C, was transported into the deposition chamber by means of an Ar flow (rate = 60 sccm) through gas lines maintained at 170 8C to avoid condensation phenomena. Separate Ar and O2 flows (rates = 15 and 20 sccm, respectively) were directly introduced into the reactor. On the basis of previous results,[14, 15, 41] growth experiments were performed under optimized conditions (total pressure = 1.0 mbar; substrate temperature = 300 8C; RF power = 20 W; duration = 60 min). Subsequently, RF sputtering processes were carried out from a Cu target (Cu 99.95 %, Alfa Aesar) in Ar plasmas (Ar flow rate = 10 sccm; total pressure = 0.30 mbar; experiment duration = 150 min), at a substrate temperature as low as 60 8C, to prevent undesired variations of the ZnO matrix characteristics during depositions.[35] Ex situ annealing in air at 400 8C for 60 min was finally performed to ensure complete copper oxidation to CuO and stabilize the obtained CuO/ZnO materials before gassensing tests.

Characterization FE-SEM and EDS measurements were performed by using a Zeiss SUPRA 40VP instrument (acceleration voltages = 10.0–20.0 kV), equipped with an Oxford INCA x-sight X-ray detector. ZnKa1 and CuKa1 X-ray signals were acquired along cross-section line scans. XPS analyses were carried out by using a PerkinElmer F 5600ci spectrometer at pressures lower than 108 mbar, using a standard MgKa excitation source (1253.6 eV). BE values (standard deviation =  0.2 eV) were corrected for charging by assigning a value of 284.8 eV to the C1s line of adventitious carbon. Peak fitting was performed by a least-squares procedure, adopting Gaussian–Lorentzian peak shapes. Zn and Cu Auger parameters were determined as previously reported.[14, 29] The atomic composition was calculated by using FV5.4A sensitivity factors. Ar + erosion was carried out at 3.0 kV (Ar partial pressure = 5  108 mbar). SIMS analysis were carried out by using a Cameca IMS 4f mass spectrometer, using a Cs + primary beam (14.5 keV, 15 nA) and negative secondary ion detection. Rastering over a 150  150 mm2 area was performed, whereas the secondary ions were collected from a subregion close to 8  8 mm2 to avoid crater effects. Charging phenomena were compensated for by means of an electron gun. Beam blanking mode and high mass resolution configuration were adopted to improve in-depth resolution and avoid mass interference artifacts, respectively. GIXRD patterns were collected at a fixed incidence angle of 18 by using a Bruker D8 Advance diffractometer equipped with a Gçbel mirror and a CuKa source (powered at 40 kV, 40 mA). Gas-sensing tests were performed in a temperature-stabilized sealed chamber (20 8C, relative humidity level = 40 %) in the 100– 400 8C interval. Higher working temperatures were intentionally avoided to prevent undesired alterations of the obtained nanocomposites. A constant synthetic air flow (0.3 L  min1) at atmospheric pressure was used as a carrier gas for the dispersion of the analytes in the desired concentrations.[42] Conductometric sensing devices, fabricated by sputtering of Pt through a shadow mask, consisted of interdigitated electrodes (gap = 200 mm) over CuO/ ZnO surfaces and heaters on the backsides of Al2O3 substrates.[42] The sensor responses, defined for an n-type SC towards oxidizing (S) and reducing (S’) analytes, were obtained by using Equations (5) and (6), respectively:[10, 14, 35, 42] S ¼ ðRf R0 Þ=R0 ¼ DR=R0

ð5Þ

S0 ¼ ðGf G0 Þ=G0 ¼ DG=G0

ð6Þ

 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemphyschem.org

2347

D. Barreca et al. in which R0 (G0) corresponded to the baseline resistance (conductance) in air, whereas Rf (Gf) were the corresponding steady-state values upon analyte exposure. Resistance (conductance) values were obtained by the measured current values and the constant applied voltage of 1 V. Before sensing measurements, all samples were prestabilized at the working temperature for 8 h. Repeated experiments under the same operational conditions yielded stable and reproducible sensor responses for one month (estimated uncertainty =  5 %).

Acknowledgements The research leading to these results has received funding from the European Community’s Seventh Framework Program (FP7/ 2007-2013) under grant agreement no. ENHANCE-238409, as well as from COFIN-PRIN 2008 and Padova University PRAT 2010 (no. CPDA102579) projects. A. Ravazzolo (CNR-ISTM, Padova, Italy) is gratefully acknowledged for technical assistance. Keywords: copper sputtering · CuO · gas sensors · chemical vapor deposition · ZnO [1] V. Aroutiounian, Int. J. Hydrogen Energy 2007, 32, 1145. [2] E. Comini, C. Baratto, G. Faglia, M. Ferroni, A. Vomiero, G. Sberveglieri, Prog. Mater. Sci. 2009, 54, 1. [3] C. C. Li, Z. F. Du, L. M. Li, H. C. Yu, Q. Wan, T. H. Wang, Appl. Phys. Lett. 2007, 91, 032101. [4] N. Singh, A. Ponzoni, R. K. Gupta, P. S. Lee, E. Comini, Sens. Actuators B 2011, 160, 1346. [5] T.-J. Hsueh, S.-J. Chang, C.-L. Hsu, Y.-R. Lin, I. C. Chen, Appl. Phys. Lett. 2007, 91, 053111. [6] T. Jinkawa, G. Sakai, J. Tamaki, N. Miura, N. Yamazoe, J. Mol. Catal. A: Chem. 2000, 155, 193. [7] K.-W. Kim, P.-S. Cho, S.-J. Kim, J.-H. Lee, C.-Y. Kang, J.-S. Kim, S.-J. Yoon, Sens. Actuators B 2007, 123, 318. [8] Z. Wang, Z. Li, J. Sun, H. Zhang, W. Wang, W. Zheng, C. Wang, J. Phys. Chem. C 2010, 114, 6100. [9] C. W. Na, H.-S. Woo, I.-D. Kim, J.-H. Lee, Chem. Commun. 2011, 47, 5148. [10] D. Barreca, E. Comini, A. P. Ferrucci, A. Gasparotto, C. Maccato, C. Maragno, G. Sberveglieri, E. Tondello, Chem. Mater. 2007, 19, 5642. [11] Y. Zhu, C. H. Sow, T. Yu, Q. Zhao, P. Li, Z. Shen, D. Yu, J. T. L. Thong, Adv. Funct. Mater. 2006, 16, 2415. [12] A. Wisitsoraat, A. Tuantranont, E. Comini, G. Sberveglieri, W. Wlodarski, Thin Solid Films 2009, 517, 2775. [13] W. J. Moon, J. H. Yu, G. M. Choi, Sens. Actuators B 2002, 87, 464. [14] D. Barreca, D. Bekermann, E. Comini, A. Devi, R. A. Fischer, A. Gasparotto, C. Maccato, G. Sberveglieri, E. Tondello, Sens. Actuators B 2010, 149, 1.

2348

www.chemphyschem.org

[15] D. Barreca, D. Bekermann, E. Comini, A. Devi, R. A. Fischer, A. Gasparotto, C. Maccato, C. Sada, G. Sberveglieri, E. Tondello, CrystEngComm 2010, 12, 3419. [16] Y. Chen, C. L. Zhu, G. Xiao, Nanotechnology 2006, 17, 4537. [17] D. Barreca, E. Comini, A. Gasparotto, C. Maccato, C. Sada, G. Sberveglieri, E. Tondello, Sens. Actuators B 2009, 141, 270. [18] Y.-S. Kim, I.-S. Hwang, S.-J. Kim, C.-Y. Lee, J.-H. Lee, Sens. Actuators B 2008, 135, 298. [19] M. L. Zhong, D. C. Zeng, Z. W. Liu, H. Y. Yu, X. C. Zhong, W. Q. Qiu, Acta Mater. 2010, 58, 5926. [20] C. Yang, X. Su, F. Xiao, J. Jian, J. Wang, Sens. Actuators B 2011, 158, 299. [21] H. Bae, G. Choi, Sens. Actuators B 1999, 55, 47. [22] J. D. Choi, G. M. Choi, Sens. Actuators B 2000, 69, 120. [23] C. S. Dandeneau, Y.-H. Jeon, C. T. Shelton, T. K. Plant, D. P. Cann, B. J. Gibbons, Thin Solid Films 2009, 517, 4448. [24] S.-J. Jung, H. Yanagida, Sens. Actuators B 1996, 37, 55. [25] J. X. Wang, X. W. Sun, Y. Yang, K. K. A. Kyaw, X. Y. Huang, J. Z. Yin, J. Wei, H. V. Demir, Nanotechnology 2011, 22, 325704. [26] N. Wu, M. Zhao, J. Zheng, C. Jiang, B. Myers, S. Li, M. Chyu, S. X. Mao, Nanotechnology 2005, 16, 2878. [27] M. B. Rahmani, S. H. Keshmiri, M. Shafiei, K. Latham, W. Wlodarski, J. du Plessis, K. Kalantar-Zadeh, Sens. Lett. 2009, 7, 621. [28] A. Gasparotto, D. Barreca, D. Bekermann, A. Devi, R. A. Fischer, C. Maccato, E. Tondello, J. Nanosci. Nanotechnol. 2011, 11, 8206. [29] D. Barreca, A. Gasparotto, E. Tondello, Surf. Sci. Spectra 2007, 14, 41. [30] D. Briggs, M. P. Seah, Practical Surface Analysis, 2nd ed., Wiley, New York, 1990. [31] JCPDS, card no. 36-1451, 2000. [32] JCPDS, card no. 42-1468, 2000. [33] X. Liu, X. Wu, H. Cao, R. P. H. Chang, J. Appl. Phys. 2004, 95, 3141. [34] P. Romero-Gmez, J. Toudert, J. R. Snchez-Valencia, A. Borrs, A. Barranco, A. R. Gonzalez-Elipe, J. Phys. Chem. C 2010, 114, 20932. [35] Q. Simon, D. Barreca, A. Gasparotto, C. Maccato, E. Tondello, C. Sada, E. Comini, A. Devi, R. A. Fischer, Nanotechnology 2012, 23, 025502. [36] N. Barsan, D. Kojziej, U. Weimar, Sens. Actuators B 2007, 121, 18. [37] E. Comini, G. Faglia, G. Sberveglieri, Solid State Gas Sensing, Springer, New York, 2009. [38] D. Barreca, D. Bekermann, E. Comini, A. Devi, R. A. Fischer, A. Gasparotto, M. Gavagnin, C. Maccato, C. Sada, G. Sberveglieri, E. Tondello, Sens. Actuators B 2011, 160, 79. [39] M. R. Mohammadi, D. J. Fray, Acta Mater. 2007, 55, 4455. [40] D. Bekermann, D. Rogalla, H.-W. Becker, M. Winter, R. A. Fischer, A. Devi, Eur. J. Inorg. Chem. 2010, 2010, 1366. [41] Q. Simon, D. Barreca, D. Bekermann, A. Gasparotto, C. Maccato, E. Comini, V. Gombac, P. Fornasiero, O. I. Lebedev, S. Turner, A. Devi, R. A. Fischer, G. Van Tendeloo, Int. J. Hydrogen Energy 2011, 36, 15527. [42] D. Barreca, A. Gasparotto, C. Maccato, C. Maragno, E. Tondello, E. Comini, G. Sberveglieri, Nanotechnology 2007, 18, 125502.

Received: December 28, 2011 Published online on April 24, 2012

 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemPhysChem 2012, 13, 2342 – 2348