highly sensitive thin film nh3 gas sensor operating at room

A vailable online at www.sciencedirect.com Sensors and Actuators B 129 (2008) 888–895 Highly sensitive thin film NH 3 gas sensor operating at room temperature based on SnO 2 /MWCNTs composite Nguyen Van Hieu a,b,∗ , Luong Thi Bich Thuy a , Nguyen Duc Chien a,b,c a International Training Institute for Materials Science (ITIMS), Hanoi University of Technology (HUT), Viet Nam b Hanoi Advanced School of Science and Technology (HAST), Hanoi University of Technology (HUT), Viet Nam c Institute of Engineering Physics (IEP), Hanoi University of Technology (HUT), Viet Nam Received 21 February 2007; received in revised form 26 September 2007; accepted 27 September 2007 Available online 13 October 2007 Abstract A SnO 2 /MWCNTs composite-based NH 3 sensor working at room temperature was fabricated by thin film microelectronic technique. The gas- sensitive composite thin film was prepared by using both commercially available multi-walled carbon nanotubes (MWCNTs) and nanosized SnO 2 dispersion. Microstructure and surface morphology of the composite were investigated and they revealed that the MWCNTs were still present and well embedded by SnO 2 particles in the composite powder as well as in the composite thin film at calcination temperatures up to 550 ◦ C. The effect of the preparation process of the sensitive composite thin film on gas-sensing properties was examined, and the preparation process parameters such as MWCNTs content, MWCNTs diameter, calcination temperature, and film thickness were optimized. At room temperature, the optimal composite sensor exhibited much higher response and faster response-recovery (less than 5 min) to NH 3 gas of concentrations ranging from 60 to 800 ppm, in comparison with the carbon nanotubes-based NH 3 sensor. Based on the experimental observations, a model of potential barrier to electronic conduction at the grain boundary for the CNTs/SnO 2 composite sensors was also discussed. © 2007 Elsevier B.V. All rights reserved. Keywords: Nanocomposites; Carbon nanotubes; Gas sensors 1. Introduction SnO 2 -based sensors have been extensively investigated since they can detect a wide variety of gases with high sensitivity and good stability at low production cost [1–3]. However, like other semiconductor type gas sensors, SnO 2 sensors should be operated above room temperature, which brings about much inconvenience for practical applications and sometimes it is even unsafe for detecting combustion gases [4–6]. Currently, SnO 2 and noble metal doped SnO 2 -based sensors are commercially available [7,8]. Still, much effort has been made to improve gas-sensitivity as well as to reduce operating temperature by introducing dopants or decreasing SnO 2 particle size to the nanoscale (<10 nm) [2,4,5,9]. ∗ Corresponding author at: International Training Institute for Materials Sci- ence (ITIMS), Hanoi University of Technology (HUT), No. 1 Dai Co Viet Road, Hanoi, Viet Nam. Tel.: +84 4 8680787; fax: +84 4 8692963. E-mail address: hieunv-itims@mail.hut.edu.vn (N. Van Hieu). Carbon nanotubes (CNTs) special geometry and their amaz- ing feature of being all surface reacting materials offer great potential applications as gas sensor devices working at room temperature. It has been reported that the CNTs are very sensi- tive to surrounding environment. The presence of O 2 ,NH 3 ,NO 2 gases and many other molecules can either donate or accept electrons, resulting in an alteration of the overall conductiv- ity [10,11]. Such properties make CNTs ideal for nanoscale gas-sensing materials, and CNTs field effect transistors and conductive-based devices have already been demonstrated as gas sensors [12–15]. However, the CNTs still have certain lim- itations for gas sensor application such as long recovery time, detection of limited gases, and strong influence of humidity and other gases. Recently, the combinations of metal oxides such as SiO 2 , TiO 2 , SnO 2 , and the CNTs have been paid much attention for various applications such as photocatalytic, anode materials for lithium-ion batteries as well as gas sensors [16–20]. The combi- nation can be conducted by different ways such as SiO 2 /CNTs, TiO 2 /CNTs, and SnO 2 /CNTs composite [16–18], SnO 2 -coated CNTs [20], SnO 2 -filled CNTs [21], and SnO 2 -doped with CNTs 0925-4005/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2007.09.088 N. Van Hieu et al. / Sensors and Actuators B 129 (2008) 888–895 889 [22]. Gas sensors based on SnO 2 -coated CNTs and SnO 2 -doped with CNTs have been reported by Liu et al. [20] and by Wei et al. [22], respectively. Recently, the composites of metal oxides (SiO 2 ,WO 3 , and SnO 2 )/CNTs for gas-sensing application have also been reported [19,23,24]. However, their gas-sensing per- formance has not yet been much improved in comparison to SnO 2 as well as CNTs-based sensors. Furthermore, currently reported composite sensors still operate at elevated tempera- tures. To the best of our knowledge, it seems that experimental data on SnO 2 /CNTs composite-based gas sensors operated at room temperature are still lacking. The NH 3 gas sensors based on CNTs [25,26] and SnO 2 [27–31] have been extensively investigated. The SnO 2 -based gas sensors can detect NH 3 gas with good sensitivity and response- recovery time, but it only operates at elevated temperatures. In contrast, the CNTs-based sensors can detect NH 3 gas at room temperature, but their sensitivity is still low and response- recovery time is still very long. In this paper, we present our current research on gas-sensing properties of SnO 2 /MWCNT composites, in which we aim to take advantage of both SnO 2 and CNTs to develop room temper- ature gas sensors to detect NH 3 gas with much better response and shorter response-recovery time, compared to those of the sensors based on the SnO 2 or CNT material alone. 2. Experimental The SnO 2 dispersion (15% nanoparticles with particle size of 10–15 nm dispersed in water) purchased from Chemat Tech- nology Inc. (US) [32] was used for the preparation of the gas-sensing material. Two kinds of MWCNTs with different diameters (d < 10 and d = 60–100 nm), their lengths of 1–2 ␮m and their purity of 95% were used in this study (they were pur- chased from Shenzhen Nanotech Port Ltd. Co., China [33]). The gas-sensing element based on a SnO 2 -MWCNTs compos- ite was fabricated in the following manners. At the beginning, MWCNTs bundles and cetyltrimethyl ammonium bromide (C 16 TMAB A.R., Merck) were added and dispersed in the SnO 2 dispersion by ultrasonic vibration for about 1 h to obtain a well- mixed suspension. The immersion-probe ultrasonic with a high power up to 500 W(Model VC-505, Sonics, US) was used. Then, the suspension of CNTs and SnO 2 nanoparticles was deposited on the Pt interdigitated electrode by means of spin-coating. The MWCNTs/SnO 2 composites with different MWCNTs con- tents were prepared for the sensitive thin film fabrication. The thickness of the sensitive thin films was controlled by varying spin-coating speed. The coating layer was dried in air for 24 h and subsequently calcinated for 1 h at different temperatures and conditions (vacuum or air atmosphere). The interdigitated elec- trode was fabricated using the conventional photolithographic method with a finger width of 100 ␮m and a gap size of 70 ␮m. The fingers of interdigitated electrode were fabricated by sput- tering 10 nm Ti and 200 nm Pt on a layer of silicon dioxide (SiO 2 ) with the thickness of about 100 nm thermally grown on top of a silicon wafer. The microstructure of the composite thin film was charac- terized by X-ray diffraction (XRD, Cu Ka radiation), using Bruker-AXS D5005. The morphology of the sensing layers was verified by field-emission scanning electron microscope (FE- SEM, 4800 Hitachi, Japan). Fig. 1. Apparatus for gas sensor testing. 890 N. Van Hieu et al. / Sensors and Actuators B 129 (2008) 888–895 The gas-sensing measurements were carried out as follows. The sensor was first placed on a hot plate and electrically con- nected by tungsten needles, and then all were loaded in a glass chamber (see Fig. 1). The desired NH 3 gas concentrations, obtained by mixing NH 3 gas with air using a computerized mass flow control system (AALBORG model GFC17S-VALD2- A0200), were injected into the chamber subsequently. The injection of a certain amount of the mixed gas was accurately controlled by a computer. After a duration of time, the chamber was purged with air and the experiment was repeated for another cycles. The electrical resistance response during testing was mon- itored by a precision semiconductor parameter analyzer (HP4156A). The sensor response (S) for a given measurement was calculated as follows: S = R gas /R air , where R gas and R air are electrical resistances of the sensor in a tested gas and in air, respectively. 3. Results and discussion 3.1. Microstructure characterizations The XRD patterns of blank SnO 2 , MWCNTs and MWCNTs/ SnO 2 composite are compared in Fig. 2. The most intense two peaks of MWCNTs correspond to the (0 0 2) and (1 0 0) reflections, respectively. Only SnO 2 in crystalline phase can be indexed from the patterns for SnO 2 and the composite. It is note- worthy that the characteristic peaks of MWCNTs can hardly be identified from the patterns of the composite. Although the most intense peak of MWCNTs corresponding to (0 0 2) reflec- tion overlaps the peak of crystalline tin oxide (1 1 0) reflection, the composite presents a symmetric peak of the crystalline tin oxide corresponding to (1 1 0) reflection in its diffraction pat- terns. Additionally, the other intense peak of MWCNTs due to (1 0 0) reflection between 40 ◦ and 50 ◦ , where no peak can be attributed to SnO 2 , is also absent for the composite. This Fig. 2. XRD patterns of (a) SnO 2 wt%, (b) MWCNTs wt%, and (c) SnO 2 - 10 wt% MWCNTs composites (Cu Ka radiation). Fig. 3. SEM images of (a) SnO 2 wt% nanoparticles and (b) SnO 2 -10 wt% MWCNTs nanocomposites annealed at 550 ◦ C in the vacuum of 10 −2 Torr. observation can be hypothesized that the MWCNTs are well embedded in the SnO 2 matrix. The FE-SEM images of blank tin oxide and 10 wt% MWCNTs/SnO 2 composite powder samples after heat treat- ment at 550 ◦ C in vacuum (10 −2 Torr) are shown in Fig. 3a and b, respectively. Spherical fine particles (around 10 nm) were observed in the blank tin oxide sample. This is just a rough esti- mation of the size of particles because of the limitation of the FE-SEM method. One notes that the particle size of the tin oxide in suspension solution was indicated by the producer to be less than 15 nm [32]. As in the composite, it was found out that the CNTs disperse well and separate from each other clearly (see, Fig. 3b) and CNTs are well embedded by spherical tin oxide nanoparticles. Our sensing element is of a thin film type. There- fore, the morphology of the composite thin film after the heat treatment at 550 ◦ C in vacuum of 10 −2 Torr was also verified by the FE-SEM, and the result is shown in Fig. 4. It is observed that there are many fiber-like protrusions emerged from the SnO 2 matrix, which may indicate that the CNTs are most embedded in the SnO 2 . The CNTs on the surface of the composite thin film are also coated by SnO 2 nanoparticles as indicated in the inset of Fig. 4. The diameter of the coated MWCNTs fibers is around 40 nm, which is larger than that of the pure MWCNTs (d < 10 nm). It has been reported that there is a good attachment of SnO 2 nanoparticles on CNTs due to the electrostatic inter- action between the tin oxide nanoparticles and the MWCNTs, which is quite strong so that the inner SnO 2 nanoparticles immo- N. Van Hieu et al. / Sensors and Actuators B 129 (2008) 888–895 891 Fig. 4. SEM image of SnO 2 -10 wt% MWCNTs composites thin films annealed at 550 ◦ C in the vacuum of 10 −2 Torr. bilized on the MWCNTs are stable[17,20,34,35]and agrees with our experimental observations, where CNTs are well embedded in the SnO 2 /MWCNTs composite of both powder and thin film formations. 3.2. Gas sensor characteristics 3.2.1. Sensor response at room temperature and sensing mechanism Fig. 5 plots a typical response curve to NH 3 gas at room tem- perature of a 10 wt% MWCNTs/SnO 2 composite sensor. This composite was calcined at 500 ◦ C in vacuum of 10 −2 Torr. The response curve shows that the resistance of the sensor varies over time with various cyclic tests. It can be seen that the resistance increases upon exposure to NH 3 gas and it returns to the original value upon exposure to the air. Since NH 3 is an electron donating gas, the increase ofthe sensor resistance can be hypothesized that the composite sensing layer behaves as a p-type semiconductor. We should note that the SnO 2 thin film cannot have resistance as Fig. 5. Sensor response of 10 wt%-MWCNTs/SnO 2 composite sensor calci- nated at 500 ◦ C in the vacuum of 10 −2 Torr to different concentrations of NH 3 gas. low as that of the composite at room temperature. Additionally, it cannot respond to NH 3 gas at room temperature. This implies that the response of the composite sensor should be mainly contributed by the MWCNTs, which have been well known to behave as a p-type semiconductor [12–15,25–28]. Comparing with the CNTs-based NH 3 sensor [27,29] and the SnO 2 -based NH 3 sensor [29–31] reported previously, as-synthesized com- posites SnO 2 /MWCNTs-based sensors have a higher response to NH 3 gas at room temperature. The exact mechanism of the high response of the SnO 2 /MWCNTs composites as a sensing material is still not clear. However, we speculate that the enhancement of the response to NH 3 gas of the composite sensors may result from the p–n hetero-junction formed by CNTs and SnO 2 nanoparti- cles, which has been indicated by Wei et al. [22]. The model is similar to the p–n junctions of sensing materials, which have been investigated by several authors [36–39]. The p–n semiconductor/SnO 2 gas sensor has been demonstrated to work at room temperature. They have proposed that the change in barrier height or in the conductivity of the SnO 2 sensitive layer may modulate the depletion layer at the p–n junction of the Si substrate. This change of the depletion layer in the p–n junction, induced by the sensitive SnO 2 layer, may cause the improve- ment in the performance of the gas sensor at low operating temperature. However, the SWCNTs-doped SnO 2 sensor behaves as an n-type semiconductor, while our MWCNTs/SnO 2 composite sensor behaves as a p-type semiconductor. A plausible expla- nation is that the composite has a much higher CNTs content; as a result, the major conducting carriers are the holes, which are mainly contributed by CNTs. When the MWCNTs/SnO 2 composites are exposed to NH 3 gas, NH 3 molecules may inter- act with the MWCNTs by replacing the pre-adsorbed oxygen [19,26,40], while NH 3 adsorbs and mutually interacts with oxy- gen on the surface of SnO 2 , resulting in oxidation of NH 3 gas at the surface and removing the oxygen accordingly. Therefore both of these effects can modulate the potential barrier of the hetero-junction formed by MWCNTs and SnO 2 and can change the conductivity of the compositematerial during the exposure to NH 3 gas. Apparently, this possible mechanism requires further experimental and theoretical investigations. Fig. 6 is to show estimations of the response and recovery times of our best sensor, in which optimized parameters such as MWCNTs content, thermal treatment condition and thick- ness were selected (will be shown later). In this figure, the time interval between measured points is 2 s. It can be seen that the response-recovery time is less than 5 min. Fig. 6 also shows that the response occurred immediately after few seconds of gas- injection in the chamber. The response time from A to B (Fig. 6) is the time needed for the gas in the testing chamber to become homogenous (see Fig. 1). Previous reports have shown that the CNTs-based sensor can detect various gases at room temper- ature, but the response and recovery times are quite long, of the order of 1 h [14,25,26]. This is hardly acceptable in prac- tice with CNTs-based sensors. So, the use of SnO 2 /MWCNTs composites for the gas sensor can somehow overcome the problem. 892 N. Van Hieu et al. / Sensors and Actuators B 129 (2008) 888–895 Fig. 6. Response to NH 3 of 15 wt%-MWCNTs/SnO 2 composite sensor calci- nated at 530 ◦ C. 3.2.2. Effect of the MWCNTs content It has been realized that when operated at room tempera- ture SnO 2 -based sensors do not respond to NH 3 gas unlike the MWCNTs-based sensors. Therefore, it was predicted that the content of MWCNTs in the composites strongly affected the response of the composite sensors. So various MWCNTs/SnO 2 composites-based sensors, in which the MWCNTs content (weight ratio of MWCNT to SnO 2 ) was varied such as 5%, 10%, and 15%, were characterized. Fig. 7 shows the sensor response versus NH 3 gas concentration of the composite sensors with different MWCNTs contents. It shows that the response of the sensors depends strongly on the NH 3 gas concentration and the slope (R/C) of the curve for linear fit is large enough (0.03–0.05) for the gas sensor applications. These values are comparable with that previously reported for CNTs-based sensor [14]. Fig. 7b plots the dependence of sensor response to 200 ppm NH 3 gas on the MWCNTs content. These sensors were annealed at 530 ◦ C in vacuum of 10 −2 Torr. It can be seen in Fig. 7b that the response of the sensor to NH 3 gas increases with increasing MWCNTs content. However, the relation between the response and the NH 3 gas concentration is less linear in the case of high MWCNTs content (see Fig.7a). So far we cannot increase the MWCNTs content in the composite because the dispersion of a higher MWCNT content in SnO 2 sol is not good enough and we cannot get repeatable results. To overcome the problem, further studies are needed. 3.2.3. Effect of the MWCNTs diameter It was shown that the diameter of CNTs strongly affected the electronic properties as well as gas-adsorption/desorption behavior [41–43]. Therefore, in this work, we also stud- ied the effect of MWCNTs diameter on the response of the MWCNTs/SnO 2 composites-based sensor. Fig. 8 shows the response of two composite sensors, which were fabricated by using MWCNTs with diameters of lower than 10 nm and in the range of 60–100 nm. We observe that the composites using MWCNTs with the larger diameter has higher response. This effect can be explained by the fact that the MWCNTs embed- ded in SnO 2 behave as nanochannels for the gas diffusion in Fig. 7. Effect of MWCNTs content on the response of the composite sensors: (a) the response of the different composites vs. NH 3 gas concentration and (b) the response to 200 ppm NH 3 gas vs. MWCNTs content. Fig. 8. Effect of the MWCNTs diameter on the response of the thin film com- posite sensors. N. Van Hieu et al. / Sensors and Actuators B 129 (2008) 888–895 893 the composite materials. So, larger diameter MWCNTs would increase the number of gas molecules adsorbed on the material. The effect should be especially strong with the gas having larger molecules like NH 3 . Specific surface area (SSA) of these CNTs was examined by the BET method (data not shown here). The SSA of MWCNTs with diameter of <10 and 60–100 nm were 242.2 and 45.2 m 2 /g, respectively. In principle, the material with a higher SSA would have a better gas response. However, we have observed an opposite effect for our case. So, the SSA fac- tor cannot be a piece of evidence for the difference in the sensor response. 3.2.4. Effect of thermal treatment conditions This step was dedicated to investigate if any improvement could be obtained in the detection of NH 3 gas by chang- ing the thermal treatment conditions. In this experiment, the composite sensors were calcinated at various temperatures of 400, 450, 500, and 550 ◦ C. The calcination at 500 and 550 ◦ C was carried out at vacuum of 10 −2 Torr to avoid the burn- ing of CNTs because a thermal gravimetric analysis (TGA) characterization (not shown here) pointed out that the MWC- NTs in the composites started to burn out at temperature of 548 ◦ C in the air. Fig. 9a shows the response of the sen- sors calcinated at different temperatures. It clearly indicates that the sensors calcinated at higher temperatures have much higher response. The sensor calcinated at 550 ◦ C has the high- est response in this experiment. To be sure, we carried out another experiment, in which the sensor was calcinated at a temperature of 530 ◦ C in vacuum of 10 −2 Torr. As indicated in Fig. 9b, the response of this sensor is better than other cases. So, this can be considered as an optimized calcination temperature. It has been reported in literatures [25–31] that SnO 2 /CNTs composite sensors have better performance compared to the SnO 2 and CNTs-based NH 3 gas sensors. So, we believe that the contacts between SnO 2 nanoparticles and CNTs contribute to the improvement of sensing performance of the composite sensors. Increasing the annealing temperature may result in the improvement of the contact between SnO 2 nanoparticles and CNTs, and therefore, the sensing performance of the device. However, the higher calcinated temperature may also result in burning of CNTs by residual oxygen or damaging of CNTs structure, and thus the response decreases. 3.2.5. Effect of the film thickness It is well known that the thickness of the sensitive layer has a great influence on the gas-sensing performance of thin film sen- sors, which has provided a much better platform to produce high performance gas sensors [44–49]. In this work, we also explored the effect of the thickness of the composite sensing layer on the response, to find optimized thickness for the composite gas sen- sor. The sensing layer was fabricated by mean of spin-coating. The thickness of the film was therefore controlled by the spinner speed as well as the deposition time. Fig. 10 shows the effect of the composite film thickness on the sensor response. It can be seen that the sensor response to NH 3 gas of the SnO 2 /MWCNTs composite gas sensor first increases as the thickness increases up Fig. 9. Effect of calcinationtemperature on the response ofthe composite sensor: (a) calcination temperatures from 400 to 550 ◦ C with step of 50 ◦ C and (b) calcination temperatures of 500, 530, and 550 ◦ C. to 400 nm but it decreases when the thickness further increases to 600 nm. The result of previous studies by experiment and simula- tion on semiconductor oxide thin film showed that generally the response dropped as the thickness of the sensitive film increased Fig. 10. Effect of the film thickness on the response of the composite sensors. 894 N. Van Hieu et al. / Sensors and Actuators B 129 (2008) 888–895 [46,47]. It was also concluded that this relation was prone to dras- tic alterations with respect to the microstructural defects present in the sensitive film employed. Our experimental result shows that if the MWCNTs/SnO 2 composite is too thin, the response will be decreased. Such gas-response dependent on the thickness of the composite MWCNTs/SnO 2 thin film sensor has not been clear so far. It seems that the thin film composite cannot well embed CNTs in the film due to the fact that CNTs have a rela- tively large diameter, ranging from lower than 10 to 60–100 nm. Like semiconductor oxide gas sensors, the response of the com- posite gas sensors could relate with the reactivity and diffusion of gas molecules inside the gas-sensing layers [49]. Therefore, a increase in thickness of the thin film composite sensors results in a decrease in the response due to the increase of the diffusion length of gas [49]. 4. Conclusion A new composite MWCNTs/SnO 2 thin film gas sensor has been successfully developed with high response and good response and recovery in detection of NH 3 gas at room tem- perature. The composite sensor can solve the problems of SnO 2 -based and carbon nanotubes-based sensors; the former cannot detect NH 3 gas at room temperature and the latter has very long recovery and response times in detection of NH 3 gas at room temperature. The preparation of the MWCNTs/SnO 2 composite thin film sensor was simple which both commercial SnO 2 nanoparticles dispersion and MWCNTs were used; the fabrication process involved the dispersion of MWCNTs in the SnO 2 dispersion using an ultrasonic high power immersion-probe and subsequent spin-coating and thermal treatment. The response of the MWCNTs/SnO 2 composite thin film gas sensor strongly depends on the preparation process of the sen- sitive film. The composite thin film with the MWCNTs content of 15 wt%, the MWCNTs diameter of 60–100 nm, the calcina- tion temperature of 530 ◦ C under vacuum of 10 −2 Torr, and the film thickness of 400 nm are optimal conditions. This result also implies that these conditions need to be optimized for practical applications of the composites of semiconductor oxides/carbon nanotubes as the gas sensors in general. The observations of the film morphology revealed that the MWCNT bundles were embedded in the SnO 2 nanoparticles materials. According to this result, a model of a potential barrier to electronic conduction at the grain boundaryfor the composites of CNTs/semiconductor oxide sensorsisa plausible explanation. Acknowledgements This work is financially supported by VLIR-HUT project, Code AP05/Prj3/Nr03. The authors also acknowledge Grant No. 405006 (2006) from the Basic Research Program of the Ministry of Science and Technology (MOST) and for partly financial support from Third Italian-Vietnamese Executive Pro- gramme of Co-operation in S&T for 2006–2008 under project title, “Synthesis and Processing of Nanomaterials for Sensing, Optoelectronics, and Photonic Applications”. References [1] D. Kotsikau, M. Ivanovskaya, D. Orlik, M. Falasconi, Gas-sensitive prop- erties of thin and thick film sensors based on Fe 2 O 3 –SnO 2 nanocomposites, Sens. Actuator B 101 (2004) 199–206. [2] P. Ivanoc, E. Llobet, X. Vilanova, J. Brezmes, J. Hubalek, X. Corrig, Devel- opment of high sensitivity ethanol gas sensors based on Pt-doped SnO 2 surfaces, Sens. Actuator B 99 (2004) 201–206. [3] A. Teeramongkonrasmee, M. Sriyudthsak, Methanol and ammonia sensing characteristics of sol–gel derived thin film gas sensor, Sens. Actuator B 66 (2000) 256–259. [4] R.S. Niranjan, V.A. Chaudhary, I.S. Mulla, A novel hydrogen sulfide room temperature sensor based on copper nanocluster functionalized tin oxide thin films, Sens. Actuator B 85 (2002) 26–32. [5] E. Comini, G. Faglia, G. Sberveglieri, UV light activation of tin oxide thin films for NO 2 sensing at low temperatures, Sens. Actuator B 78 (2001) 73–77. [6] K. Anothainart, Light enhanced NO 2 gas sensing with tin oxide at room temperature: conductanceand work function measurements, Sens.Actuator B 93 (2003) 580–584. [7] Figaro Eng. Inc., http://www.figarosensor.com. [8] Microsens Inc., http://www.microsens.ch. [9] R.S.Niranjan, Y.K. Hwang, D K. Kim, S.H. Jhung, J S. Chang, I.S. Mulla, Nanostructured tin oxide: synthesis and gas-sensing properties, Mater. Chem. Phys. 92 (2005) 384–388. [10] H. Ulbricht, G. Moos, T. Hertel, Interaction of molecular oxygen with single-wall carbon nanotube bundles and graphite, Surf. Sci. 532–535 (2003) 852–856. [11] S. Santucci, S. Picozzi, F. Di Gregorio, L. Lozzi, NO 2 and CO gas adsorption on CNTs: experiment and theory, J. Chem. Phys. 119 (2003) 10904–10910. [12] A. Modi, N. Koratkar, E. Lass, B. Wei, Miniaturized gas ionization sensors using carbon nanotubes, Nature 424 (2003) 171–173. [13] T. Someya, J. Small, P. Kim, C. Nuckolls, J.T. Yardley, Alcohol vapor sensors based on single-walled carbon nanotube field effect transistors, Nano Lett. 3 (2003) 877–881. [14] J. Li, Y. Lu, Q. Ye, M. Cinke, J. Han, M. Meyyappan, Carbon nanotube sensors for gas and organic vapor detection, Nano Lett. 3 (2003) 929–933. [15] R. Ionescu, E.H. Espinosa, E. Sotter, E. Llobet, X. Vilanova, X. Correig, A. Felten, C. Bittencourt, G. Van Lier, J.C. Charlier, J.J. Pireaux, Oxygen functionalisation of MWNT and their use as gas sensitive thick-film layers, Sens. Actuator B 113 (2006) 36–46. [16] J. Xie, V.K. Varadan, Synthesis and characterization of high surface area tin oxide/functionalized carbon nanotubes composite as anode materials, Mater. Chem. Phys. 91 (2005) 274–280. [17] M.H. Chen, Z.C. Huang, G.T. Wu, G.M. Zhu, J.K. You, Z.G. Lin, Synthesis and characterization of SnO–carbon nanotube composite as anode material for lithium-ion batteries, Mater. Res. Bull. 38 (2003) 831–836. [18] W. Wanga, P. Serp, P. Kalck, J.L. Faria, Photocatalytic degradation of phe- nol on MWNT and titania composite catalysts prepared by a modified sol–gel method, Appl. Catal. B: Environ. 56 (2004) 301–308. [19] O.K. Varghese, P.D. Kichambre, Gas sensing characteristics of multi-wall carbon nanotubes, Sens. Actuator B 81 (2001) 32–41. [20] Y L. Liu, H F. Yang, Y. Yang, Z M. Liu, G L. Shen, R Q. Yu, Gas sensing properties of tin dioxide coated carbon nanotubes, Thin Solid Films 497 (2006) 355–360. [21] L. Zhao, L. Gao, Filling of multi-walled carbon nanotubes with tin(IV) oxide, Carbon 42 (2004) 3251–3272. [22] B Y. Wei, M C. Hsu, P G. Su, H M. Lin, R J. Wu, H J. Lai, A novel SnO 2 gas sensor doped with carbon nanotubes operating at room temper- ature, Sens. Actuator B 101 (2004) 81–89. [23] C. Bittencourt, A. Felten, E.H. Espinosa, R. Ionescu, E. Llobet, X. Correig, J J. Pireaux, WO 3 films modified with functionalized multi-wall carbon nanotubes: morphological, composition and gas response studies, Sens. Actuator B 115 (2006) 33–41. [24] A. Wisitsoraat, A. Tuantranont, C. Thanachayanont, V. Patthanasettakul, P. Singjai, Electron beam evaporated carbon nanotubes dispersed SnO 2 thin films gas sensor, J. Electroceram. 17 (2006) 45–47. N. Van Hieu et al. / Sensors and Actuators B 129 (2008) 888–895 895 [25] S.G. Wanga, Q. Zhang, D.J. Yang, P.J. Sellin, G.F. Zhong, Multi-walled carbon nanotube-based gas sensors for NH 3 detection, Diamond Relat. Mater. 13 (2004) 1327–1332. [26] J. Kong, N.R. Franklin, C. Zhou, M.G. Chapline, S. Peng, K. Cho, H. Dai, Nanotube molecular wires as chemical sensors, Science 287 (2000) 622–625. [27] M. Arab, F. Berger, F. Picaud, C. Ramseyer, J. Glory, M.M. -L’Hermite, Direct growth of the multi-walled carbon nanotubes as tool to detect ammo- nia at room tempertuare, Chem. Phys. Lett. 433 (2006) 175–181. [28] E. Bekyarova, M. Davis, T. Burch, M.E. Itkis, B. Zhao, S. Sunshine, R.C. Haddon, Chemically functionalized single-walled carbon nanotubes as ammonia sensors, J. Phys. Chem. B 108 (2004) 19717–19720. [29] V.V. Kovalenko, A.A. Zhukova, M.N. Rumyantseva, A.M. Gaskov, V.V. Yushchenko, I.I. Ivanova, T. Pagnier, Surface chemistry of nanocrystalline SnO 2 : effect of thermal treatment and additives, Sens. Actuator B 126 (2006) 52–55. [30] J. Kaur, S.C. Roy, M.C. Bhatnagar, Highly sensitive SnO 2 thin film NO 2 gas sensor operating at low temperature, Sens. Actuator B 96 (2006) 1090–1095. [31] Y D. Wang, X H. Wu, Q. Su, Y F. Li, Z L. Zhou, Ammonia-sensing characteristics of Pt and SiO 2 doped SnO 2 materials, Solid State Electron. 45 (2001) 347–350. [32] Chemat Technology Inc., 9036 Winaetka Avenue, Northridge, CA, http://www.chemat.com. [33] Shenzhen NanoTech Port. Co. Ltd. (China), http://www.nanotubes.com.cn. [34] W Q. Han,A. Zettl, Coatingsingle-walled carbonnanotubes with tin oxide, Nano Lett. 3 (2003) 681–683. [35] L. Zhao, L. Gao, Coating of multi-walled carbon nanotubes with thick layers of tin(IV) oxide, Carbon 42 (2004) 1858–1861. [36] W. Zhang, E.A. de Vasconcelos, H. Uchida, T. Katsube, T. Nakatsubo, Y. Nishioka, A study of silicon Schottky diode structures for NO x gas detection, Sens. Actuator B 65 (2000) 154–156. [37] W. Zhang, H. Uchida, T. Katsube, T. Nakatsubo, Y. Nishioka, A study of silicon Schottky diode structures for NO x gas detection, Sens. Actuator B 49 (1998) 58–62. [38] A. Kunimoto, N. Abe, H. Uchida, T. Katsube, Highly sensitive semicon- ductor NO x gas sensor operating at room temperature, Sens. Actuator B 65 (2000) 122–124. [39] N. Savage, B. Chwieroth, A. Ginwalla, B.R. Patton, S.A. Akbar, P.K. Dutta, Composite n–p semiconducting titanium oxides as gas sensors, Sens. Actuator B 79 (2001) 17–27. [40] A. Cheng, W.A. Steele, Computer simulation of ammonia on graphite: low temperature structure of monolayer and bilayer films, J. Chem. Phys. 92 (1990) 3858–3864. [41] A. Javey, M. Shim, H. Dai, Electrical properties and devices of large- diameter single-walled carbon nanotubes, Appl. Phys. Lett. 80 (2002) 1064–1066. [42] P X. Hou, S T. Xu, Z. Ying, Q H. Yang, C. Liu, H M. Cheng, Hydro- gen adsorption/desorption behavior of multi-walled carbon nanotubes with different diameters, Carbon 41 (2003) 2471–2476. [43] K. Seo, K.A. Park, C. Kim, S. Han, B. Kim, Y.H. Lee, Chirality- and diameter-dependent reactivity of NO 2 on carbon nanotube walls, J. Am. Chem. Soc. 127 (2005) 15724–15729. [44] A. Galdikas, S. Kaiulis, G. Mattogno, A. Mironas, A. Napoli, D. Senulien, A. Etkus, Thickness effect of constituent layers on gas sen- sitivity in SnO 2 /[metal]/metal multi-layers, Sens. Actuator B 58 (1999) 478–485. [45] P. Montmeat, R. Lalauze, J P. Viricelle, G. Tournier, C. Pijolat, Model of the thickness effect of SnO 2 thick film on the detection properties, Sens. Actuator B 103 (2004) 84–90. [46] F.H. - Babaei, M. Orvatinia, Analysis of thickness dependence of the sensitivity in thin film resistive gas sensors, Sens. Actuator B 89 (2003) 256–261. [47] G. Sakai, N.S. Baik, N. Miura, N. Yamazoe, Gas sensing properties of tin oxide thin films fabricated from hydrothermally treated nanoparticles- dependence of CO and H 2 response on film thickness, Sens. Actuator B 77 (2001) 116–121. [48] A. Zahab, L. Spina, P. Poncharal, Water-vapor effect on the electrical con- ductivity of a single-walled carbon nanotube mat, Phys. Rev. B 63 (2000) 1000–1002. [49] G.Sakai, N. Matsunaga, K.Shimanoe, N. Yamazoe, Theory of gas diffusion controlled sensitivity for thin filmsemiconductor gas sensor, Sens. Actuator B 80 (2001) 125–131. Biographies Nguyen Van Hieu received his MSc degree from the International Training Institute for Material Science (ITIMS), Hanoi University of Technology (HUT) in 1997 and PhD degree from the Department of Electrical Engineering, Univer- sity of Twente, Netherlands in 2004. Since 2004, he has been a research lecturer at the ITIMS. In 2007, he worked as a post-doctoral fellow, Korea University. His current research interests include the nano-architectures of carbon nanotubes, oxide semiconductors and oxide semiconductor nanowires for chemical sensors. Luong Thi Bich Thuy received the BS degree in physics at Hanoi University of Education in 2004, and MSc degree in materials science from the International Training Institute of Material Science (ITIMS), Hanoi University of Technology (HUT), in 2006. Her research interest is the development of semiconductor oxide/carbon nanotubes composites gas sensors. Nguyen Duc Chien received the engineering degree in electronic engineering at Leningrad Electrotechnical University, Russian, in 1976, and the MSc and PhD in microelectronics at Grenoble Polytechnique University, France, in 1985 and 1988, respectively. He has worked as associated professor at the Institute of Engineering Physics (IEP), Hanoi University of Technology (HUT). From 1989 to 1990 he worked as a visiting professor at the Grenoble University, France. From 1992 to 2006 he was a vice director of the International Training Institute for Materials Science (ITIMS), HUT, where he established a Laboratory of Microelectronics and Sensors. Since 2003 he has been the Director of the IEP, HUT. His research interests include: characterizations and modeling of MOS devices, nanomaterials for chemical sensor, biosensor, optoelectronic materials and devices, and MEMS devices. He has been a leader of many national research projects related to microelectronic devices and functional nanomaterials. Dr Nguyen Duc Chien is a member of Physics Society of Vietnam and Vietnamese Materials Research Society. . A vailable online at www.sciencedirect.com Sensors and Actuators B 129 (2008) 888–895 Highly sensitive thin film NH 3 gas sensor operating at room temperature based. concentrations of NH 3 gas. low as that of the composite at room temperature. Additionally, it cannot respond to NH 3 gas at room temperature. This implies that