+Kheel-2017-Hydrogen Gas Detecti

Met. Mater. Int., Vol. 23, No. 1 (2017), pp. 214~219doi: 10.1007/s12540-017-6343-3Hydrogen Gas Detection of Nb2O5 Nanoparticle-Decorated CuONanorod SensorsHyejoon Kheel1, Gun-Joo Sun1, Jae Kyung Lee1, Ali Mirzaei2, Seungbok Choi3, and Chongmu Lee1,*1Department of Materials Science and Engineering, Inha University, Incheon 22212, Republic of Korea 2Department of Materials Science and Engineering, Shiraz University, Shiraz, Iran 3Department of Mechanical Engineering, Inha University, Incheon 22212, Republic of Korea(received date: 16 May 2016/accepted date: 29 June 2016)Pristine and Nb2O5 nanoparticles-decorated CuO nanorods were prepared successfully by a two step process: thethermal evaporation of a Cu foil and the spin coating of NbCl5 solution on CuO nanorods followed by thermalannealing. X-ray diffraction was performed to examine the structure and purity of the synthesized nanoatuc-tures. Scanning electron microscopy was used to examine the morphology and shape of the nanostuctures. TheNb2O5 nanoparticles-decorated CuO nanorod sensor showed responses of ~217.05-862.54%, response timesof ~161-199s and recovery times of ~163-171 s toward H2 gas with concentrations in a range of 0.5 - 5%at the optimal working temperature of 300°C. The Nb2O5 nanoparticle-decorated CuO nanorod sensor showedsuperior sensing performance to the pristine CuO nanorod sensor for the same H2 concentration range. Theunderlying mechanism for the enhanced hydrogen sensing performance of the CuO nanorods decoratedwith Nb2O5 nanoparticles is discussed.Keywords: semiconductors, nanostructured materials, oxidation, electrical properties, scanning electron microscopy(SEM)1. INTRODUCTIONAs an attractive energy source, hydrogen (H2) is expected to substitute carbon-based fuels and become a promising clean, renewable, and common fuel for the future that can overcome some of the critical problems such as shortages of fossil fuel resources, global warming and air pollution [1]. In spite of its advantages, unfortunately H2 is tasteless, col-orless, odorless, and very explosive and inflammable [2]. Its low flammable limit (4 vol%) and ultra-small molecular size, which make difficult confinement and containment of it, demands very cautious handling [3]. These properties of hydro-gen demands an accurate detection technique and continu-ous monitoring for safe industrial use [4]. In this regards, gas sensors might be very promising detectors for hydrogen gas because they are cheap, simple, reliable.Recently, semiconducting metal oxide (SMO) one-dimen-sional (1D) nanostructures such as nanorods (NRs) have been widely studied as sensor materials because of their high specific area and excellent carrier transporting characteristics [2]. So far, n-type SMOs have been intensively studied as gas sensing materials [2-4]. In contrast, few studies have been made on p-type SMO gas sensors. Nevertheless, the poten-tial of p-type SMO as gas sensing materials should not be underestimated. In addition to several shortcomings, p-type SMOs have advantages that n-type SMOs do not possess as follows: (1) good catalysts [5-8] to promote selective oxida-tion of various reducing gases; (2) formation of p-n hetero-structures by combining with a p-type SMO can also be used to enhance the gas sensing performance by altering the gas-sensing characteristics of gas sensors near the heterointer-faces [9]; (3) the higher oxygen absorption of p-type SMOs, i.e., higher oxygen solubility in p-type SMOs leading to low humidity dependence and rapid recovery kinetics [10,11]. Therefore, a combination of p-type SMOs with n-type SMOs, i.e., formation of p-n heterostructures would exhibit enhanced sensing performance compared to pristine n-type SMOs [12]. Cupric oxide (CuO) is a p-type metal oxide [13-16] drawing significant attention owing to its highly catalytic activity [17] and many other applications in fields of lithium-copper oxide electrochemical cells [18], field emission devices [19], high temperature superconductors [20], infrared photodetectors [21], etc. CuO nanostructured gas sensors have attracted grow-ing attention due to its high surface area-to-volume ratio, adsorption capability of oxygen and catalytic activity [22]. In this study, pristine CuO NRs and Nb2O5 nanoparticles-decorated CuO NRs were fabricated by the direct thermal oxidation of Cu foils [23] and spin coating of NbCl5 solution followed by annealing at high temperature. Finally, their H2*Corresponding author: cmlee@inha.ac.kr KIM and SpringerHydrogen Gas Detection of Nb2O5 Nanoparticle-Decorated CuO Nanorod Sensors215 gas sensing properties were examined.2. EXPERIMENTAL PROCEDURENb2O5-decorated CuO NRs were by the direct thermaloxidation of Cu foils [23] and spin coating of NbCl5 solutionfollowed by annealing at high temperature. First, CuO NRswere prepared by oxidizing a Cu foil (99.8%, thickness:0.25mm, Alfa Aesar) thermally in a quartz tube furnace.Cu foil pieces were placed at the center of a 20 mm-dia. quartztube in the heating zone of a vacuum tube furnace. Oxygenand nitrogen gases were introduced to the system after 10 minevacuation. The furnace was then heated to 500°C. At the temperature of 500°C, O2/N2 mixture gas was introduced into the quartz tube to initiate CuO oxidation. After heating for 1 h, the furnace was cooled to room temperature and the products (CuO NRs) were taken out.Subsequently, the CuO NR samples were decorated with Nb2O5 nanoparticles using a solvothermal method. An eth-anolic niobium chloride solution (C2H5OH:NbCl5=1000:1) was prepared in a vial. The CuO NR samples were immersed in the ethanolic niobium chloride solution (C2H5OH:NbCl5 = 1000:1) in a vial in a home-made ultraviolet (UV) box. The samples were then transferred to an annealing quartz tube furnace and annealed at 500°C for 1 h in O2 ambient. The O2 gas flow rate and process pressure were 200 cm3/min and 0.8 Torr, respectively.Scanning electron microscopy (SEM, Hitachi S-4200) was performed to examine the morphology of the synthesized samples. X-ray diffraction (XRD, Philips X`pert MRD dif-fractometer) measurements were performed to examine the structures of the samples by glancing angle using Cu-Kαradiation (0.1541 nm) at a glancing angle of 0.5°. Interdigital electrode (IDE) patterns to be used as the elec-trodes of the sensor were fabricated by the sequential sput-ter-deposition of ~200 nm-thick Ni and ~50 nm-thick Au thin films followed by photolithography. On the other hand, the pristine and Nb2O5 nanoparticles-decorated CuO NRs were separately dispersed by ultrasonicating in a mixture of deion-ized water (5 ml) and isopropyl alcohol (5 ml), followed by drying at 90°C. A solution containing the NRs (10μl) was dropped onto the SiO2(200 nm)/Si substrates equipped with a pair of the IDE s with a gap of 20μm (Fig. 1).The gas sensing properties of the multi-networked NR sensors fabricated using the pristine and Nb2O5-decorated CuO NRs were measured at 300 °C in a quartz tube placed in a sealed chamber. H2 gas was introduced into the testing tube to obtain 0.5-5% H2 concentrations while monitoring the electrical resistance of sensors. One sensing cycle con-sisted of introducing synthetic air at a flow rate of 200 cm3/min for 5 min to achieve a baseline, switching to H2 gas/air mixture at a flow rate of 200 cm3/min for 5 min, followed by switch-ing back to synthetic air at a flow rate of 200 cm3/min for a 10 min for sensor recovery. The electrical resistance of the gas sensor was determined by measuring the electric current at 300°C flowing at an applied voltage of 1V between the two Ni/Au IDEs. A Keithley sourcemeter- 2612 was used for this measurement. The details of the sensing test procedures are described elsewhere [24]. The electrical resistances of the NR sensors were monitored while H2 gas was introduced into the testing tube. The response of the sensors to H2 was defined as R g/R a×100 (%), where R g and R a are the electrical resistances in the sensors in the H2 gas and air, respectively.3. RESULTS AND DISCUSSION3.1. Structures of the pristine and Nb2O5 nanoparticle-decorated CuO NRsFigure 2 shows the XRD patterns of both the pristine and Nb2O5 nanoparticles-decorated CuO NRs. The pristine CuO NR sample exhibited the reflections from simple cubic Cu2O (JCPDS no. 05-0667) and monoclinic CuO (JCPDS no. 45-0937), as well as those from pure Cu, which might be due to the Cu foil substrate. This result suggests that both Cu2O and CuO phases coexist in the NRs. Copper is thermal oxi-dized according to the following reactions [25]:4Cu + O2 2Cu2O (1) Fig. 1.Schematic of the gas sensor.Fig. 2. XRD pattern of the pristine and Nb2O5-decorated CuO NRs.216Hyejoon Kheel et al.2Cu 2O + O 2 4CuO (2)During thermal oxidation process a dense Cu 2O film covers the Cu foil surface first, then CuO NRs grows on the Cu 2O film {18}. Of the two cupric oxide phases the CuO phase is dominant at 500°C. The Cu 2O peaks in the XRD patterns of the two samples might be caused by the Cu 2O layer under-neath the NRs, not by the nanorods. On the other hand, the NRs consists mostly of the CuO phase. In addition to cupric oxide and cupper peaks, the (111), (002) and (134) peaks from Nb 2O 5 were observed in the diffraction pattern of the Nb 2O 5nanoparticles-decorated CuO NRs, which are due to the pres-ence of Nb 2O 5 nanoparticles on the surface of the CuO NRs.Figure 3 shows a SEM micrograph of the synthesized CuO NRs decorated with Nb 2O 5 nanoparticles. The inset in Fig. 3shows that the NRs have diameters less than 50 nm and lengths of up to 5 μm and that the diameters of the Nb 2O 5 nanopar-ticles on the CuO NRs range from 20-70 nm. Also the NRs tended to grow perpendicular to the substrate and there was a small gap between neighboring NRs, which along with the high surface area of the NRs, makes them favorable for gas sensing applications.3.2. Comparison of the H 2 sensing performance between the pristine and Nb 2O 5 nanoparticles-decorated CuO NR sensorsFigure 4 shows the responses of both sensors to 0.5% H 2gas, where it clearly shows the response of Nb 2O 5 nanopar-ticles-decorated CuO NRs is higher than the pristine CuO NRs over the entire test temperature range. At the optimal working temperature of Nb 2O 5 nanoparticle-decorated CuO NR sensor the response of the decorated CuO NR sensor is more than three times higher than the pristine one. Accord-ing to this experiment all remainder tests were conducted atthe optimal temperature i.e. 300°C. According to [26,27], hydro-gen gas sensors that can operate at temperatures from 30 to 1000°C are needed. Our sensor can be used for relatively high temperature applications that need to detect H 2.Figures 5(a) and (b) present the dynamic response of the pristine and Nb 2O 5 nanoparticles-decorated CuO NR sen-sors to H 2gases with different concentrations, respectively.Fig. 3. SEM micrograph of the synthesized CuO NRs decorated with Nb 2O 5 nanoparticles. Inset, enlarged SEM image of Nb 2O 5-decoratedCuO NRs.Fig. 4. Response of the pristine and Nb 2O 5-decorated CuO NR sen-sors versus temperature.Fig. 5. Dynamic response of (a) the pristine CuO NR sensor and (b)Nb 2O 5-decorated CuO NR sensor.Hydrogen Gas Detection of Nb 2O 5 Nanoparticle-Decorated CuO Nanorod Sensors 217The resistance of the sensors increased rapidly upon expo-sure to H 2 gas and the resistance recovered completely to the initial value upon stopping the H 2 gas supply and intro-ducing air. The responses of the NR sensors were quite sta-ble and reproducible for repeated testing cycles. In both cases,the sensors exhibited a higher response when they were exposed to higher concentrations of hydrogen. As an increase in H 2gas concentration activates the reactions at the surfaces of sensors due to the larger coverage of surface, the response of the sensors would increase accordingly. Also both sensors exhibited a p-type electrical behavior to H 2 gas, which might be because a p-type SMO, CuO is a major constituent of both sensors. The baseline resistance of the Nb 2O 5 nanoparticle-decorated CuO NR sensor is, interestingly, higher than that of the pristine CuO NR sensor, due to the higher intrinsic resistance of Nb 2O 5 than that of CuO.Figure 6 shows the responses of both sensors at different H 2 concentrations. The pristine CuO NR sensor shows an almost linear trend with very a low slope with increasing the H 2 concentration. This means that the sensor almost reached a saturation response owning to its smaller surface area. In contrast, the response of the Nb 2O 5 nanoparticles-decorated CuO NR sensor increases progressively with increasing the H 2 concentration. Even at a maximum con-centration of H 2 (5%), it did not reach the saturation value.It seems that decorating CuO NRs with Nb 2O 5 resulted in an increased surface area. The main reasons for the higher response of the Nb 2O 5 nanoparticles-decorated CuO NR sensor will be discussed in the next section.Figures 7(a) and (b) show the response/recovery times of both sensors. The difference in response time might be due to that in the saturation time and mean residence period of the H 2 molecules on the surface of the sensor material. For a low H 2 concentration, the H 2 molecules spend a relatively long time reacting with the adsorbed oxygen species. For ahigh H 2 concentration, a larger amount of H 2 molecules are available for reaction with the adsorbed oxygen species, so that the response time is shorter. Change of the recovery time with H 2 concentration might be due to the followings:When air is introduced into the testing tube, oxygen mole-cules are adsorbed by the sensor material surface and react with the H 2 molecules. The complete desorption at the CuO-Nb 2O 5 interface takes more time than at the outer Nb 2O 5surface, resulting in a longer recovery at higher H 2 concen-trations, where H 2 molecules exist at the CuO-Nb 2O 5 interface.In both cases, Nb 2O 5-decorated CuO NRs have shorter response and recovery times. The difference in response time between the pristine and Nb 2O 5-decorated CuO NRs at higher concentrations (4 and 5%) is more significant.This is probably due to the larger surface area of the Nb 2O 5-decorated CuO NRs. Also the recovery times are shorter than the response times, which is probably due to the lower desorption energy of H 2relative to the adsorption energy.Fig. 6. Response of pristine and Nb 2O 5-decorated CuO NRs versushydrogen concentration.Fig. 7. (a) Response times and (b) recovery times of the pristine and Nb 2O 5-decorated CuO NR sensors for different H 2 concentrations.218Hyejoon Kheel et al.3.3. Sensing MechanismIn the pristine CuO NR gas sensor, in air, the oxygen mole-cules are adsorbed by the CuO NR surfaces and the adsorbedoxygen molecules are ionized to O2-, O- or O2- by releasingholes to the valence band of CuO. The stable oxygen speciesare O2-, O- and O2- below 150 °C, between 150 and 400 °C,and above 400°C, respectively [28]. In this study, predom-inant species might be O- because the operating temperaturewas 300°C. The ionization of oxygen molecules causes an increase in carrier concentration and formation of an accu-mulation layer at the surface of the CuO NRs, leading to a decrease in resistance:1/2O2(g) = O-(ad) + h+ (1)In this case, the accumulation plays a role of a conduction channel.Upon exposure to H2 gas, the ethanol molecules adsorbedby the CuO NR surfaces are oxidized by reacting with the preexisting oxygen species at the surfaces as follows:2H2 + 2 O-(ad) + 2 h+ 2H2O (2) Reaction (2) capture holes from the valence band of CuO,forming a depletion layer, leading to an increase in the resis-tance of the CuO NRs. Consequently, the adsorption of H2gas and air accompanies the modulation of the conductionchannel width, leading to the change in resistance.The enhanced sensing properties of the decorated CuO NR sensor might be caused mainly by the two electronic effects: (1) modulation of conduction channel width and (2) modulation of the potential barrier height at the CuO-Nb2O5 interface accompanying the absorption and desorption of H2 gas.The modulation of the conduction channel width of the Nb2O5-decorated CuO NR sensor is more significant than that of the pristine CuO NR sensor (W c4-W c3 > W c2-W c1 in Fig. 8), because the depletion layer forming on the Nb2O5 side in the Nb2O5-decorated CuO NR sensor in air is thicker than that in the pristine CuO NR sensor owing to the exis-tence of the accumulation layer on the CuO side [28], causing the difference in conduction channel width between the two sensors. A potential barrier also develops at the p-CuO/n-Nb2O5 interface due to the built-in potential (Fig. 8) [29,30]. Modulation of the potential barrier height occurs during the adsorption and desorption of H2 gas. (NOTE: Desorption of H2 gas corresponds to adsorption of air.) The potential barrier height at the p-n junction changes from V1 (in H2 gas) to V2 (in air) as shown in Fig. 8. Of these two effects, the modulation of the conduction channel width might be predominant over the modulation of the potential barrier height because the Nb2O5-decorated CuO NR sensor as well as the pristine CuO NR sensor exhibited a p-type electrical behavior [Figs. 5(a) and (b)]. If the modulation of the potential barrier height had been dominant, the decorated CuO NR sensor would have exhibited an n-type electrical behavior, differently from the pristine CuO NR sensor.Besides the two electrical effects, the enhanced sensing performance of the decorated CuO NR sensor are partially attributed to physical effects such as the increase in the surface area of the NRs [31,32] and the formation of more active sites for gas adsorption by decoration of the CuO NRs with Nb2O5 particles [33]. Brunauer-Emmett-Teller (BET) anal-ysis was performed to verify the specific surface area of the Nb2O5-decorated CuO nanorods was larger than that of the pristine CuO nanorods. The measured specific surface areas of the pristine and Nb2O5-decorated CuO nanorods were 3.17 and 10.26 m2/g, respectively. The specific surface area of the pristine CuO nanorods was more than two fold larger than that of the pristine CuO nanorods.4. CONCLUSIONSNb2O5-decorated CuO NRs were synthesized using a two-step process: the synthesis of CuO NRs using a thermal evaporation method followed by the deposition of Nb2O5 nanoparticles using a solvothermal method. The optimal working temperature of the decorated NR sensors was deter-mined to be 300°C. The Nb2O5-decorated CuO NRs exhib-ited stronger and faster response to H2 gas than the pristine CuO NRs. The underlying mechanism of the improved sens-ing properties of the multi-networked Nb2O5-decorated CuO Fig. 8. Energy band diagrams of the p-CuO-n-Nb2O5 system and sche-matic diagrams of the pristine and Nb2O5-decorated CuO NR sensors showing the conduction channel widths and potential barrier heights in air and H2 gas.Hydrogen Gas Detection of Nb2O5 Nanoparticle-Decorated CuO Nanorod Sensors219NR sensor to H2 gas was explained by the modulation of the conduction channel width, the modulation of the poten-tial barrier height at the Nb2O5-CuO interface, the increased surface area of the NRs, and more active sites for gas adsorp-tion created by decoration of the CuO NRs with Nb2O5. We believe that p-type SMOs could also be efficiently used for fabricating gas sensors, if only they were combined with n-type SMOs. This paper showed the improved sensing prop-erties of the p-type SMO nanostructures decorated with n-type SMO nanoparticles and elucidated the underlying mech-anism for the improved gas sensing properties.ACKNOWLEDGEMENTThis work was supported by Basic Science Research Pro-gram through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2010-0020163).REFERENCES1.W. J. Buttner, M. B. Post, R. Burgess, and C. Rivkin, Int. J.Hydrogen Energ.36, 2462 (2011).2.S. Park, S. Park, S. Lee, H. W. Kim, and C. Lee, Sens. Actua-tors B202, 840 (2014).3.T. Hübert, L. Boon-Brett, G. Black, and U. Banach, Sens.Actuators B157, 329 (2011).4.H. G. Z. Wang and Y. Hu, Sensors12, 5517 (2012).5.K. Pirkanniemi and M. Sillanpää, Chemosphere48, 1047(2002).6.M. M. Bettahar, G. Costentin, L. Savary, and J. C. Lavalley,Appl. Catal. A-Gen.145, 1 (1996).7.Y. Motooka and A. Ozaki, J. Catal.5, 116 (1966).8.S. S. Kaye and J. R. Long, J. Am. Chem. Soc.127, 6506 (2005).9.Z. Zhang, C. Shao, X. Li, C. Wang, M. Zhang, and Y. Liu,ACS Appl. Mater. Interfaces2, 2915 (2010).10.H.-R. Kim, A. Haensch, I.-D. Kim, N. Barsan, U. Weimar,and J.-H. Lee, Adv. Funct. Mater.21, 4456 (2011).11.H.-R. Kim, K.-I. Choi, K.-M. Kim, I.-D. Kim, G. Cao, andJ.-H. Lee, Chem. Commun.46, 5061 (2010).12.S. Park, H. Kheel, G.-J. Sun, H. W. Kim, T. Ko, and C. Lee,Met. Mater. Int.22, 730 (2016).13.S. Park, H. Ko, S. An, W. Lee, S. Lee, and C. Lee, Ceram.Int. 39, 5255 (2013).14.H. Meng, W. Y ang, K. Ding, L. Feng, and Y. Guan, J. Mater.Chem. A3, 1174 (2015).15.H. J. Lee, T. G. Kim, H. Oh, W. Lee, and H. Ryu, Korean J.Met. Mater. 53, 445 (2015).16.U. H. Lee, H. J. Lee, and C. M. Park, Korean J. Met. Mater.54, 469 (2015).17.K. Zhou, R. Wang, B. Xu, and Y. Li, Nanotechnology17,3939 (2016).18.L. B. Chen, N. Lu, C. M. Xu, H. C. Y u, and T. H. Wang,Electrochim. Acta54, 4198 (2009).19.Y. W. Zhu, T. Y u, F. C. Cheong, X. J. Xu, C. T. Lim, V. B.C. Tan, C. H. Sow, et al.Nanotechnology16, 88 (2005).20.A. H. Macdonald, Nature414, 409 (2001).21.S. B. Wang, C. H. Hsiao, S. J. Chang, K. T. Lam, K. H. Wen,S. C. Hung, et al.Sens. Actuators A171, 207 (2011). 22.C. Y ang, F. Xiao, J. Wang, and X. Su, Sens. Actuators B207,177 (2015).23.C. Yang, X. Cao, S. Wang, L. Zhang, F. Xiao, X. Su, and J.Wang, Ceram. Int.41, 1749 (2015).24.S. Park, S. An, Y. Mun, and C. Lee, ACS Appl. Mater. Inter-faces5, 4285 (2013).25.H. Kim, C. Jin, S. Park, S. Kim, and C. Lee, Sens. Actua-tors B161, 594 (2012).26.G. Korotcenkov, S. D. Han, and J. R. Stetter, Chemical109,1402 (2009).27.J. M. Lim, K. C. Shin, and H. W. Kim, Thin Solid Films475, 256 (2005).28.N. Barsan and U. Weimer, Electroceram.7, 143 (2001).29.Z. Zhu, C.-T. Kao, and R.-J. Wu, Appl. Sur. Sci.320, 348(2014).30.D. R. Patil and L. A. Patil, Talanta 77, 1409 (2009).31.H.-J. Kim and J.-H. Lee, Sens. Actuators B192, 607 (2014).32.F. Qu, J. Liu, Y. Wang, S. Wen, Y. Chen, S. Ruan, et al.Sens.Actuators B199, 346 (2014).33.H. Zhang, J. Feng, T. Fei, S. Liu, and T. Zhang, Sens. Actu-ators B190, 472 (2014).。