- Email: [email protected]
Preparation of nanocrystalline nickel oxide thin films by sol–gel process for hydrogen sensor applications
Materials Science and Engineering C 32 (2012) 2230–2234
Contents lists available at SciVerse ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Preparation of nanocrystalline nickel oxide thin films by sol–gel process for hydrogen sensor applications A.M. Soleimanpour, Ahalapitiya H. Jayatissa ⁎ MEMS and Nanotechnology Laboratory, Mechanical, Industrial and Manufacturing Engineering (MIME) Department, 2801 West Bancroft St. MS 312, The University of Toledo, Toledo, OH 43606, United States
a r t i c l e
i n f o
Article history: Received 4 January 2012 Received in revised form 4 June 2012 Accepted 11 June 2012 Available online 24 June 2012 Keywords: Nickel oxide Sol–gel coating Crystallinity Electrical properties Hydrogen sensor
a b s t r a c t Preparation of nanocrystalline NiO thin films by sol–gel method and their hydrogen (H2) sensing properties were investigated. The thin films of NiO were successfully deposited on the glass and SiO2/Si substrate by a sol–gel coating method. The films were characterized for crystallinity, electrical properties, surface topography and optical properties as a function of calcination temperature and substrate material. It was found that the films produced by this method were polycrystalline and phase pure NiO. The H2 gas sensitivity of these films was studied as a function of H2 concentration and calcination temperature. The results indicated that the sol–gel derived NiO films could be used for the fabrication of H2 gas sensors to monitor low concentration of H2 in air quantitatively at low temperature range (b200 °C). © 2012 Elsevier B.V. All rights reserved.
1. Introduction Recently, metal oxide thin films have been investigated for gas sensor applications at low temperature. In particular, monitoring of hydrogen (H2) in air is very important because of rapid increase in H2 usage in industries, laboratories, and energy technologies. Also, sensors that operate with low power consumption and at low temperature are important for continuous monitoring of H2 concentrations in ambient conditions in a safe manner. However, the current commercial gas sensors suffer from severe limitations due to their high power consumption, high measurement temperature and low sensitivity [1–4]. Therefore, there is an urgent need for new sensor materials that can be used to design sensors with low power consumption. In order to further improve the gas sensor performance, nanocrystalline metal–oxide semiconductor materials have been widely studied for the fabrication of gas sensors [4–6]. Nickel oxide (NiO) thin films are p-type and transparent semiconductor with a band gap energy in the range of 3.6–4.0 eV [7]. NiO thin films have been investigated for electrochromic devices, gas sensors, organic light emitting diodes (OLED), and display devices [8–11]. Thin films of NiO have been fabricated by different methods such as reactive sputtering [12], chemical vapor deposition [13], reactive pulsed laser deposition [14] and sol–gel method [15]. Nickel is a transition metal which can have multiple oxidation states such as Ni 2+ and Ni 3+. Also, NiO exhibits a p-type semiconducting behavior due non-stoichiometry of metal oxides [2]. In recent years, metal oxide ⁎ Corresponding author. Tel.: +1 419 530 8245; fax: +1 419 530 8206. E-mail address: [email protected] (A.H. Jayatissa). 0928-4931/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.msec.2012.06.007
thin films have been fabricated by sol–gel process because of its cost effectiveness, scalability and reproducibility. The sol–gel process produces high quality thin films on different substrates. The sol–gel derived metal oxide films have several important properties such as high porosity, uniformity, and nanocrystallinity, which are key factors related to gas sensing properties such as sensitivity and response and recovery times [16]. In the present study, p-type and transparent NiO films were prepared by a sol–gel process on the glass and SiO2/Si substrates. The film properties were investigated as a function of calcination temperature and substrate material. The effectiveness of substrate and calcination temperature was investigated by measuring the gas sensing properties of NiO. It was found that the H2 sensors fabricated using these NiO can be operated at a temperature as low as 175 °C with high sensitivity, fast response and recovery times. The results were highly reproducible and repeatable with high material stability in a wide temperature range. In this paper, these experimental results are presented. 2. Experimental procedure NiO thin films were synthesized by a sol–gel coating method on the alkali free glass substrate and oxidized silicon wafer (SiO2/Si), where the thickness of SiO2 was around 1 μm. The nickel nitrate hexahydrate was used as the starting material and it was dissolved in equal amount of isopropanol alcohol and polyethylene glycol 200 to get a 0.1 M solution. The prepared solution was stirred at a temperature of 40 °C for 60 min to yield a transparent and homogenous solution. A dilute ammonium hydroxide solution was introduced to the
AM. Soleimanpour, A.H. Jayatissa / Materials Science and Engineering C 32 (2012) 2230–2234
solution to produce nano nickel hydroxide particles. In order to eliminate particle agglomeration, Triton X-100 (Alfa Aesar) was added to the solution prior to the addition of ammonium hydroxide solution. The prepared solution was used to coat on alkali-free glass and SiO2 (1 μm)/Si substrates by a spin coater. The spin coater was set to a rotation speed of 3000 min −1 for 30 s. In order to remove extra solvent and the organic residuals, the coated samples were heated up to 350 °C for 5 min after each coating step. This step was repeated for ten times for each sample to produce a uniform multilayer thin film. Finally, the coated samples were calcinated at different temperatures for 3 h in a tube furnace in air. Table 1 lists sample numbering and their calcination temperatures and substrate materials. The crystal structure of the samples was studied with X-ray diffraction (XRD) measurements using a Cu Kα radiation (45 kV, 40mA, Panalytical X-ray). The surface topography and particle size were studied using scanning electron microscope (SEM, JEOL 6100). The optical transmittance measurements of NiO thin films were extracted using a double beam UV/Visible spectrometer (UV-1650 PC Shimadzu) as a function of wavelength in the range of 250–1050 nm. The electrical resistivity of the samples was measured in the temperature range of 25–300 °C under vacuum condition (30 mTorr). The Fresnel approach was used to theoretically estimate the optical properties and the thickness of the films [17]. A program developed using Matlab software to theoretically estimate the refractive indices and the thickness of NiO film based on the experimental values of the reflectance data from bi-layer of NiO/glass or NiO/SiO2/Si substrates. The value of the thickness was estimated around 94±5 nm. The gas sensing devices were fabricated by vacuum deposition of a thin layer of gold (~75 nm) on the surface of NiO films followed by photolithography and etching in a KI+I2 solution with a comb like structure with 200 μm length and 20 μm width and spacing. The sensors were installed in a custom-build closed chamber (200 cm 3) with Au connecting electrodes. The sensors were placed on a resistive heater coupled with a DC regulated power supply while the temperature was fixed at 175 °C using a Ni–Cr thermocouple in contact with NiO films. Total air flow rate through the chamber was kept constant at 100 standard cubic centimeters per minute (sccm) for all measurements. The samples were installed inside the chamber and maintained in air for 60 min at 175 °C before introducing H2 gas for the first time. This step was necessary for the chamber to reach to stable temperature before starting each test. High purity H2 was introduced to the air flow through the closed chamber and the air flow rates were manipulated by needle valves and mass flow controllers. After introducing the gas into the chamber, the resistance of the sensor was measured using a high mega ohm multimeter (Keithly, Model-1200) and LabVIEW software. 3. Results and discussion 3.1. Characterization of microstructure Fig. 1 shows the XRD spectra of NiO thin films deposited on alkali free glass and SiO2/Si substrates. The calcinations were carried out at different temperatures to investigate the effect of temperature on the crystallinity. In the XRD patterns, peak appeared corresponding to (200) and (111) crystal planes of NiO. The XRD patterns indicate that the films have a polycrystalline structure and that the degree of Table 1 Description of nickel oxide films investigated in this study. Sample
Substrate
Calcination temperature (°C)
G-500 G-550 S-550 S-700
Glass Glass SiO2/Si SiO2/Si
500 550 550 700
2231
Fig. 1. XRD patterns of NiO films described in Table 1 (measured intensity data were shifted 200 for clarity).
crystallinity depends on the substrate materials as well as the calcination temperatures. When the temperature increased, the full width at half maximum (FWHM) of XRD peaks decreased indicating that the crystallinity was enhanced and the grain size was increased by increase of the calcination temperature. The average grain size of NiO films was calculated using Scherer formula as 45, 62 and 68 nm for G-550, S-550 and S-700, respectively. These results indicate that the average crystal size depends on the calcination temperature and substrate materials. Fig. 2 shows the SEM images of sol–gel coated NiO thin films on glass and SiO2/Si substrates. The surface morphology was studied with SEM as a function of calcination temperature and substrate material. It can be seen that the G-500 film had a uniform distribution of small size particles on the surface. The grain size was increased as the increase of the calcination temperature. For the film fabricated on SiO2/Si substrate, the grain size increased by increasing the calcination temperature to 700 °C (see sample S-700). Although calcination temperature was same, the grain size NiO film changed when different substrates were used (G-550 and S-550). In this case, the same sol-mixture was coated on glass and SiO2/Si substrates under identical conditions and calcination was done by keeping both samples together in the same temperature zone of the furnace. The effect clearly indicated that the substrate has a significant influence in the grain size. The result can be interpreted as due to the fast thermal expansion of SiO2/Si compared with the glass substrate. A cubic-like structure was observed on the surface of NiO films deposited SiO2/Si substrates. The shape of grains was changed and the size was increased rapidly at high temperature and the film calcinated at 700 °C exhibited a largest grain size among this set of samples. Here, both XRD patterns and SEM images of NiO films indicated that the crystal size increased and the film porosity decreased when calcination temperature was increased. The crystal size, crystallinity and porosity are three most important factors that affect the gas sensitivity of metal oxide thin film devices. The distribution of the grain size was inserted for each film. 3.2. Electrical and optical properties Fig. 3 shows the change of resistance as a function of temperature for NiO films in the range of 25–300 °C. The results indicate that the change of resistance with temperature has a typical Arrhenius behavior. When the calcination temperature was increased the electrical resistivity of NiO films increased. Generally, the resistivity of the film can be affected by isotropic background scattering due to external surface and grain boundaries [18]. The fitted line using the Arrhenius equation can be used to estimate the activation energy [10]. It is clear that the slope of the fitted line increases with increasing the calcination temperature. The activation energy of S-550 and S-700 were calculated as 0.23 eV and 0.30 eV, respectively. The activation energies of G-500 and G-550 were close to 0.22 eV. Phase pure stoichiometric NiO
2232
AM. Soleimanpour, A.H. Jayatissa / Materials Science and Engineering C 32 (2012) 2230–2234
Fig. 2. SEM images of NiO samples (a) G-500, (b) G-550, (c) S-550 and (d) S-700.
crystal is a perfect insulator; however, considerably high electrical conductivity can be achieved by increasing the amount of defects in the crystal [10,19,20]. As mentioned earlier, the polycrystalline NiO exhibits p-type conductivity due to metal cation deficiency [2]. When the films are heated at high temperature, the resistance of films is increased. It is noted that, the electrical property of NiO thin film strongly depends on the defects in the crystals such as interstitial defects and vacancies [21]. Fig. 4 shows the optical transmittance spectra of NiO films coated on the alkali free glass substrates. In this case, we could not calcinate the samples more than 550 °C on glass substrate because strain point
of this glass was around 572 °C [22]. Both samples exhibited transmittance larger than 65% in visible and near IR range and the transmittance was increased with the increase of calcination temperature. Samples have an absorption edge in the 300–380 nm range. When calcination temperature was increased from 500 °C to 550 °C, the optical gap of the films increased from 3.575 to 3.650 eV, respectively. The intrinsic bandgap of NiO has been reported to be 3.7 eV, which is very close to the measured values [18,23,24]. Moreover, these results indicated that the calcination temperature did not affect the bandgap or transmittance of films significantly.
3. 3. Characterization of gas sensor properties H2 sensitivity of NiO films was investigated as a function of H2 concentration in the air. A series of systematic investigations was carried out to identify the best performance, durability and repeatability of NiO-based sensors. It was observed that the NiO/Au interface changed in a rapid manner and electrical contact between NiO film and Au layer degraded above 200 °C for several hours of heating. Thus, we have selected 175 °C as the testing temperature of these films. When Au coated electrodes were tested at 175 °C, the electronic properties did not change even for three days of continuous heating
Fig. 3. Dependence of resistance on temperature for NiO films as a function of temperature, (a) G-500 and G-550, and (b) S-550 and S-700.
Fig. 4. Optical transmission spectra of G-500 and G-550 deposited on glass with respective to the air.
AM. Soleimanpour, A.H. Jayatissa / Materials Science and Engineering C 32 (2012) 2230–2234
2233
in the air. The sensor characteristics were reproducible for many heating and cooling cycles in the air and H2 at 175 °C. Since NiO is a p-type semiconductor, the electric resistance of NiO sensor increases when H2 is introduced to the sensor testing chamber. The sensor response described in this paper was estimated with the following formula [25]
S¼
Rg −Ra 100 Ra
:
ð1Þ
where, Rg and Ra are the resistance of sensor in the tested gas and the air, respectively. The response value depends on the flow rate of H2 at a given temperature. Fig. 5 shows the dynamic response of NiO-based sensors fabricated on glass (G-550) and SiO2/Si (S-550) substrates. Here, sensors were tested at 175 °C for three different H2 concentrations of 0.1%, 0.2% and 0.3%. The S-550 sensor has a higher response than G-550 sensor. According to the XRD patterns and SEM images of this sample, the G-550 film consists of smaller grains than S-550. Although these two samples were fabricated at the same conditions, the main difference was observed in crystal size but not in electrical properties. Therefore, large porosity and crystal defects in G-550 film can be attributed to faster H2 sensitivity. Fig. 6 shows the dynamic response of G-500 (a) and S-700 (b) sensors operated at 175 °C. It can be see that the G-500 has faster response than S-700 supporting the previous interpretation regarding the results in Fig. 5. When the calcination temperature was increased the low energy crystal defects were eliminated from the NiO films and thus, longer response time could be observed. Fig. 7 displays one complete cycle of response and recovery of the different sensors versus elapsed time for H2 gas concentration of 3000 ppm. By comparing response time (defined as the time required to reach 90% of the final equilibrium value) of different samples, the response time was calculated as 153 s, 248 s, 393 s and 1019 s for sensors fabricated with G-500, G-550, S-550 and S-700, respectively. It can be noted that the G-500 and S-700 have shortest and longest response times, respectively. This fact indicates that the increase of
Fig. 5. Dynamic response characteristics of NiO thin films exposed to 1000, 2000 and 3000 ppm of H2 at 175 °C: (a) G-550 and (b) S-550.
Fig. 6. Dynamic response characteristics of NiO thin films exposed to 1000, 2000 and 3000 ppm of H2 at 175 °C: (a) G-500 and (b) S-700.
grain size can decrease the change of carrier concentration in NiO due to low porosity and low surface to volume ratio of the crystallites. Fig. 8 compares the response of sensors at different concentrations for four sensors which prepared in this study. When the calcination temperature was increased from 500 to 700 °C, a slight increase in gas sensor response was observed. These facts indicate that crystallinity plays a significant role in metal oxide gas sensors [26]. It is also known that the NiO has a catalytic effect for producing H2 ions at the surface of NiO at elevated temperatures. Also, adsorption/desorption phenomena are enhanced at elevated temperatures [27]. These results clearly support the current understanding of interaction between H2 and metal oxide surfaces. Fig. 9a shows the dependence of sensitivity of G-550 sample on wide range of H2 concentrations in the air at 175 °C. It clearly indicates that the NiO thin films can be used to detect H2 quantitatively. The dependence of sensor response is linear for low H2 concentrations; however, it changes at a higher concentration due to the limitation of exposed sites on the NiO grains. Repeatability is an important factor related to practical applications [28,29]. Fig. 9b shows the repeatability of the G-550 sample under 3000 ppm of H2 gas at 175 °C. It indicates that though sensor runs for few subsequent cycles, no significant change can be observed among response and recovery and response
Fig. 7. The response for all NiO thin films to 3000 ppm of H2 at 175 °C.
2234
AM. Soleimanpour, A.H. Jayatissa / Materials Science and Engineering C 32 (2012) 2230–2234
4. Conclusion
Fig. 8. The response of NiO thin films to 1000, 2000 and 3000 ppm of H2 at 175 °C.
times. Small deviation can be attributed to small error in inserting target gas with mass flow meters etc. The p-type behavior of NiO is due to non-stoichiometry of the prepared samples. Vacancies in cation sites play a key role in the change of carrier concentration in NiO thin films. For each metal vacancy, two electron holes are formed. Oxygen molecules adsorbed on NiO surface at room temperature [30]. At elevated temperatures (100–300 °C), the oxygen molecules adsorb on the surface of the grains and form O − and O2−, which can increase the hole density and decrease the resistance [31,32]. With introducing the H2 molecules to the chamber, H2 molecules replaced the adsorbed oxygen molecules and produce water vapor. This effect decreases hole concentration and therefore, the resistance of NiO film increases [33]. There are only a few metal oxides exhibiting p-type semiconducting behavior and gas sensing capabilities [9,31]. When we use n-type metal oxide films for gas sensor fabrication, the resistance reduces in the presence of reducing gases such as H2. Thus maximum achievable response is less than 100%. However, when a p-type metal oxide semiconductor is used, the increase of resistance can be larger than 100% and thus, signal to noise ratio of gas sensors can be maintained at a higher level. Therefore, use of p-type metal oxide to monitor reducing gases is favorable over n-type metal oxide semiconductors.
Fig. 9. (a) The response of G-550 to 1000–40,000 ppm of H2 at 175 °C and (b) repeatability data for G-550 to 3000 ppm of H2 at 175 °C.
Thin films of NiO were fabricated by a sol–gel coating process on glass and SiO2/Si substrates. The microstructure of these films indicates that the size of crystallites, surface morphology and the film composition depend on the substrate material and calcination temperature. The electrical and gas sensing properties indicate that the NiO films have a p-type conductivity with low thermal activation energy of electrical conductivity. Also, the band gap of the materials is within the values reported in the literature. The sensor response depends on the concentration, testing temperature and calcination temperature and the substrate material. Therefore, all these parameters can be optimized to manufacture the gas sensors to monitor H2 in the air at a low temperature, where flammable H2 gas can be detected in a wide range of concentrations. The gas sensing properties of NiO sensors indicated that the gas sensors could be operated below 200 °C with reasonably fast response time and recovery time. The concentration of H2 below 0.1% in the ambient conditions can be easily detected with the NiO thin film-based sensors. Acknowledgment This work was supported by the National Science Foundation (Grant number: CMMI 0933069) of USA. References [1] A. Qureshi, A. Mergen, A. Altindal, Sens. Actuators, B 135 (2009) 537–540. [2] I. Hotovy, J. Huran, L. Spiess, R. Capkovic, S. Hascik, Vacuum (2000) 300–307. [3] M. Kitao, K. Izawa, K. Urabe, T. Komatsu, S. Kuwano, S. Yamada, Jpn. J. Appl. Phys. 33 (1994) 6656–6662. [4] A.M. Soleimanpour, Y. Hou, A.H. Jayatissa, Appl. Surf. Sci. 257 (2011) 5398–5402. [5] K. Yoshimura, T. Miki, S. Tanemura, Jpn. J. Appl. Phys. 34 (1995) 2440–2446. [6] H. Kumagai, M. Matsumoto, K. Toyoda, M. Obara, J. Mater. Sci. Lett. 15 (1996) 1081–1083. [7] D. Adler, J. Feinleib, Phys. Rev. B 2 (1970) 3112–3134. [8] I. Hotovy, J. Huran, P. Siciliano, S. Capone, L. Spiess, V. Rehacek, Sens. Actuators, B 103 (2004) 300–311. [9] C. Cantalini, M. Post, D. Buso, M. Guglielmi, A. Martucci, Sens. Actuators, B 108 (2005) 184–192. [10] P.S. Patil, L.D. Kadam, Appl. Surf. Sci. 199 (2002) 211–221. [11] W.-L. Jang, Y.-M. Lu, W.-S. Hwang, W.-C. Chen, J. Eur. Ceram. Soc. 30 (2010) 503–508. [12] I. Hotovy, J. Huran, J. Janı, A.P. Kobzev, Vacuum 51 (1998) 157–160. [13] E. Lindahl, M. Ottosson, J.O. Carlsson, Surf. Coat. Technol. 205 (2010) 710–716. [14] B. Sasi, K.G. Gopchandran, Sol. Energy Mater. Sol. Cells 91 (2007) 1505–1509. [15] A.A. Al-Ghamdi, W.E. Mahmoud, S.J. Yaghmour, F.M. Al-Marzouki, J. Alloys Compd. 486 (2009) 9–13. [16] G. Korotcenkov, Mater. Sci. Eng., R 61 (2008) 1–39. [17] O.S. Heavens, Optical Properties of Thin Solid Films, Butterworths Scientific Publications, 1955. [18] A.F. Mayadas, M. Shatzkes, Phys. Rev. B 1 (1970) 1382–1389. [19] A. Bosman, C. Crevecoeur, Phys. Rev. B 144 (1966) 763–770. [20] M.N. Islam, M.O. Hakim, H. Rahman, J. Mater. Sci. 22 (1987) 1379–1384. [21] A. Mallikarjuna Reddy, A. Sivasankar Reddy, P. Sreedhara Reddy, Vacuum 85 (2011) 949–954. [22] K. Kondoh, M. Mizuhashi, J. Non-Cryst. Solids 178 (1994) 189–198. [23] R. Romero, F. Martin, J.R. Ramos-Barrado, D. Leinen, Thin Solid Films 518 (2010) 4499–4502. [24] A.C. Sonavane, A.I. Inamdar, P.S. Shinde, H.P. Deshmukh, R.S. Patil, P.S. Patil, J. Alloys Compd. 489 (2010) 667–673. [25] M. Gautam, A.H. Jayatissa, Mater. Sci. Eng., C 31 (2011) 1405–1411. [26] Y.-D. Wang, L.-F. Yang, Z.-L. Zhou, Y.-F. Li, X.-H. Wu, Mater. Lett. 49 (2001) 277–281. [27] W. Wei, X. Jiang, L. Lu, X. Yang, X. Wang, J. Hazard. Mater. 168 (2009) 838–842. [28] H. Meixner, U. Lampe, Sens. Actuators, B 33 (1996) 198–202. [29] H. Steinebach, S. Kannan, L. Rieth, F. Solzbacher, Sens. Actuators, B 151 (2010) 162–168. [30] S.T. Shishiyanu, T.S. Shishiyanu, O.I. Lupan, Sens. Actuators, B 107 (2005) 379–386. [31] K. Wetchakun, T. Samerjai, N. Tamaekong, C. Liewhiran, C. Siriwong, V. Kruefu, A. Wisitsoraat, A. Tuantranont, S. Phanichphant, Sens. Actuators, B 160 (2011) 580–591. [32] M. Stamataki, D. Tsamakis, N. Brilis, I. Fasaki, A. Giannoudakos, M. Kompitsas, Phys. Status Solidi A 205 (2008) 2064–2068. [33] A.H. Jayatissa, P. Samarasekara, G. Kun, Phys. Status Solidi A 206 (2009) 332–337.
Contents lists available at SciVerse ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Preparation of nanocrystalline nickel oxide thin films by sol–gel process for hydrogen sensor applications A.M. Soleimanpour, Ahalapitiya H. Jayatissa ⁎ MEMS and Nanotechnology Laboratory, Mechanical, Industrial and Manufacturing Engineering (MIME) Department, 2801 West Bancroft St. MS 312, The University of Toledo, Toledo, OH 43606, United States
a r t i c l e
i n f o
Article history: Received 4 January 2012 Received in revised form 4 June 2012 Accepted 11 June 2012 Available online 24 June 2012 Keywords: Nickel oxide Sol–gel coating Crystallinity Electrical properties Hydrogen sensor
a b s t r a c t Preparation of nanocrystalline NiO thin films by sol–gel method and their hydrogen (H2) sensing properties were investigated. The thin films of NiO were successfully deposited on the glass and SiO2/Si substrate by a sol–gel coating method. The films were characterized for crystallinity, electrical properties, surface topography and optical properties as a function of calcination temperature and substrate material. It was found that the films produced by this method were polycrystalline and phase pure NiO. The H2 gas sensitivity of these films was studied as a function of H2 concentration and calcination temperature. The results indicated that the sol–gel derived NiO films could be used for the fabrication of H2 gas sensors to monitor low concentration of H2 in air quantitatively at low temperature range (b200 °C). © 2012 Elsevier B.V. All rights reserved.
1. Introduction Recently, metal oxide thin films have been investigated for gas sensor applications at low temperature. In particular, monitoring of hydrogen (H2) in air is very important because of rapid increase in H2 usage in industries, laboratories, and energy technologies. Also, sensors that operate with low power consumption and at low temperature are important for continuous monitoring of H2 concentrations in ambient conditions in a safe manner. However, the current commercial gas sensors suffer from severe limitations due to their high power consumption, high measurement temperature and low sensitivity [1–4]. Therefore, there is an urgent need for new sensor materials that can be used to design sensors with low power consumption. In order to further improve the gas sensor performance, nanocrystalline metal–oxide semiconductor materials have been widely studied for the fabrication of gas sensors [4–6]. Nickel oxide (NiO) thin films are p-type and transparent semiconductor with a band gap energy in the range of 3.6–4.0 eV [7]. NiO thin films have been investigated for electrochromic devices, gas sensors, organic light emitting diodes (OLED), and display devices [8–11]. Thin films of NiO have been fabricated by different methods such as reactive sputtering [12], chemical vapor deposition [13], reactive pulsed laser deposition [14] and sol–gel method [15]. Nickel is a transition metal which can have multiple oxidation states such as Ni 2+ and Ni 3+. Also, NiO exhibits a p-type semiconducting behavior due non-stoichiometry of metal oxides [2]. In recent years, metal oxide ⁎ Corresponding author. Tel.: +1 419 530 8245; fax: +1 419 530 8206. E-mail address: [email protected] (A.H. Jayatissa). 0928-4931/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.msec.2012.06.007
thin films have been fabricated by sol–gel process because of its cost effectiveness, scalability and reproducibility. The sol–gel process produces high quality thin films on different substrates. The sol–gel derived metal oxide films have several important properties such as high porosity, uniformity, and nanocrystallinity, which are key factors related to gas sensing properties such as sensitivity and response and recovery times [16]. In the present study, p-type and transparent NiO films were prepared by a sol–gel process on the glass and SiO2/Si substrates. The film properties were investigated as a function of calcination temperature and substrate material. The effectiveness of substrate and calcination temperature was investigated by measuring the gas sensing properties of NiO. It was found that the H2 sensors fabricated using these NiO can be operated at a temperature as low as 175 °C with high sensitivity, fast response and recovery times. The results were highly reproducible and repeatable with high material stability in a wide temperature range. In this paper, these experimental results are presented. 2. Experimental procedure NiO thin films were synthesized by a sol–gel coating method on the alkali free glass substrate and oxidized silicon wafer (SiO2/Si), where the thickness of SiO2 was around 1 μm. The nickel nitrate hexahydrate was used as the starting material and it was dissolved in equal amount of isopropanol alcohol and polyethylene glycol 200 to get a 0.1 M solution. The prepared solution was stirred at a temperature of 40 °C for 60 min to yield a transparent and homogenous solution. A dilute ammonium hydroxide solution was introduced to the
AM. Soleimanpour, A.H. Jayatissa / Materials Science and Engineering C 32 (2012) 2230–2234
solution to produce nano nickel hydroxide particles. In order to eliminate particle agglomeration, Triton X-100 (Alfa Aesar) was added to the solution prior to the addition of ammonium hydroxide solution. The prepared solution was used to coat on alkali-free glass and SiO2 (1 μm)/Si substrates by a spin coater. The spin coater was set to a rotation speed of 3000 min −1 for 30 s. In order to remove extra solvent and the organic residuals, the coated samples were heated up to 350 °C for 5 min after each coating step. This step was repeated for ten times for each sample to produce a uniform multilayer thin film. Finally, the coated samples were calcinated at different temperatures for 3 h in a tube furnace in air. Table 1 lists sample numbering and their calcination temperatures and substrate materials. The crystal structure of the samples was studied with X-ray diffraction (XRD) measurements using a Cu Kα radiation (45 kV, 40mA, Panalytical X-ray). The surface topography and particle size were studied using scanning electron microscope (SEM, JEOL 6100). The optical transmittance measurements of NiO thin films were extracted using a double beam UV/Visible spectrometer (UV-1650 PC Shimadzu) as a function of wavelength in the range of 250–1050 nm. The electrical resistivity of the samples was measured in the temperature range of 25–300 °C under vacuum condition (30 mTorr). The Fresnel approach was used to theoretically estimate the optical properties and the thickness of the films [17]. A program developed using Matlab software to theoretically estimate the refractive indices and the thickness of NiO film based on the experimental values of the reflectance data from bi-layer of NiO/glass or NiO/SiO2/Si substrates. The value of the thickness was estimated around 94±5 nm. The gas sensing devices were fabricated by vacuum deposition of a thin layer of gold (~75 nm) on the surface of NiO films followed by photolithography and etching in a KI+I2 solution with a comb like structure with 200 μm length and 20 μm width and spacing. The sensors were installed in a custom-build closed chamber (200 cm 3) with Au connecting electrodes. The sensors were placed on a resistive heater coupled with a DC regulated power supply while the temperature was fixed at 175 °C using a Ni–Cr thermocouple in contact with NiO films. Total air flow rate through the chamber was kept constant at 100 standard cubic centimeters per minute (sccm) for all measurements. The samples were installed inside the chamber and maintained in air for 60 min at 175 °C before introducing H2 gas for the first time. This step was necessary for the chamber to reach to stable temperature before starting each test. High purity H2 was introduced to the air flow through the closed chamber and the air flow rates were manipulated by needle valves and mass flow controllers. After introducing the gas into the chamber, the resistance of the sensor was measured using a high mega ohm multimeter (Keithly, Model-1200) and LabVIEW software. 3. Results and discussion 3.1. Characterization of microstructure Fig. 1 shows the XRD spectra of NiO thin films deposited on alkali free glass and SiO2/Si substrates. The calcinations were carried out at different temperatures to investigate the effect of temperature on the crystallinity. In the XRD patterns, peak appeared corresponding to (200) and (111) crystal planes of NiO. The XRD patterns indicate that the films have a polycrystalline structure and that the degree of Table 1 Description of nickel oxide films investigated in this study. Sample
Substrate
Calcination temperature (°C)
G-500 G-550 S-550 S-700
Glass Glass SiO2/Si SiO2/Si
500 550 550 700
2231
Fig. 1. XRD patterns of NiO films described in Table 1 (measured intensity data were shifted 200 for clarity).
crystallinity depends on the substrate materials as well as the calcination temperatures. When the temperature increased, the full width at half maximum (FWHM) of XRD peaks decreased indicating that the crystallinity was enhanced and the grain size was increased by increase of the calcination temperature. The average grain size of NiO films was calculated using Scherer formula as 45, 62 and 68 nm for G-550, S-550 and S-700, respectively. These results indicate that the average crystal size depends on the calcination temperature and substrate materials. Fig. 2 shows the SEM images of sol–gel coated NiO thin films on glass and SiO2/Si substrates. The surface morphology was studied with SEM as a function of calcination temperature and substrate material. It can be seen that the G-500 film had a uniform distribution of small size particles on the surface. The grain size was increased as the increase of the calcination temperature. For the film fabricated on SiO2/Si substrate, the grain size increased by increasing the calcination temperature to 700 °C (see sample S-700). Although calcination temperature was same, the grain size NiO film changed when different substrates were used (G-550 and S-550). In this case, the same sol-mixture was coated on glass and SiO2/Si substrates under identical conditions and calcination was done by keeping both samples together in the same temperature zone of the furnace. The effect clearly indicated that the substrate has a significant influence in the grain size. The result can be interpreted as due to the fast thermal expansion of SiO2/Si compared with the glass substrate. A cubic-like structure was observed on the surface of NiO films deposited SiO2/Si substrates. The shape of grains was changed and the size was increased rapidly at high temperature and the film calcinated at 700 °C exhibited a largest grain size among this set of samples. Here, both XRD patterns and SEM images of NiO films indicated that the crystal size increased and the film porosity decreased when calcination temperature was increased. The crystal size, crystallinity and porosity are three most important factors that affect the gas sensitivity of metal oxide thin film devices. The distribution of the grain size was inserted for each film. 3.2. Electrical and optical properties Fig. 3 shows the change of resistance as a function of temperature for NiO films in the range of 25–300 °C. The results indicate that the change of resistance with temperature has a typical Arrhenius behavior. When the calcination temperature was increased the electrical resistivity of NiO films increased. Generally, the resistivity of the film can be affected by isotropic background scattering due to external surface and grain boundaries [18]. The fitted line using the Arrhenius equation can be used to estimate the activation energy [10]. It is clear that the slope of the fitted line increases with increasing the calcination temperature. The activation energy of S-550 and S-700 were calculated as 0.23 eV and 0.30 eV, respectively. The activation energies of G-500 and G-550 were close to 0.22 eV. Phase pure stoichiometric NiO
2232
AM. Soleimanpour, A.H. Jayatissa / Materials Science and Engineering C 32 (2012) 2230–2234
Fig. 2. SEM images of NiO samples (a) G-500, (b) G-550, (c) S-550 and (d) S-700.
crystal is a perfect insulator; however, considerably high electrical conductivity can be achieved by increasing the amount of defects in the crystal [10,19,20]. As mentioned earlier, the polycrystalline NiO exhibits p-type conductivity due to metal cation deficiency [2]. When the films are heated at high temperature, the resistance of films is increased. It is noted that, the electrical property of NiO thin film strongly depends on the defects in the crystals such as interstitial defects and vacancies [21]. Fig. 4 shows the optical transmittance spectra of NiO films coated on the alkali free glass substrates. In this case, we could not calcinate the samples more than 550 °C on glass substrate because strain point
of this glass was around 572 °C [22]. Both samples exhibited transmittance larger than 65% in visible and near IR range and the transmittance was increased with the increase of calcination temperature. Samples have an absorption edge in the 300–380 nm range. When calcination temperature was increased from 500 °C to 550 °C, the optical gap of the films increased from 3.575 to 3.650 eV, respectively. The intrinsic bandgap of NiO has been reported to be 3.7 eV, which is very close to the measured values [18,23,24]. Moreover, these results indicated that the calcination temperature did not affect the bandgap or transmittance of films significantly.
3. 3. Characterization of gas sensor properties H2 sensitivity of NiO films was investigated as a function of H2 concentration in the air. A series of systematic investigations was carried out to identify the best performance, durability and repeatability of NiO-based sensors. It was observed that the NiO/Au interface changed in a rapid manner and electrical contact between NiO film and Au layer degraded above 200 °C for several hours of heating. Thus, we have selected 175 °C as the testing temperature of these films. When Au coated electrodes were tested at 175 °C, the electronic properties did not change even for three days of continuous heating
Fig. 3. Dependence of resistance on temperature for NiO films as a function of temperature, (a) G-500 and G-550, and (b) S-550 and S-700.
Fig. 4. Optical transmission spectra of G-500 and G-550 deposited on glass with respective to the air.
AM. Soleimanpour, A.H. Jayatissa / Materials Science and Engineering C 32 (2012) 2230–2234
2233
in the air. The sensor characteristics were reproducible for many heating and cooling cycles in the air and H2 at 175 °C. Since NiO is a p-type semiconductor, the electric resistance of NiO sensor increases when H2 is introduced to the sensor testing chamber. The sensor response described in this paper was estimated with the following formula [25]
S¼
Rg −Ra 100 Ra
:
ð1Þ
where, Rg and Ra are the resistance of sensor in the tested gas and the air, respectively. The response value depends on the flow rate of H2 at a given temperature. Fig. 5 shows the dynamic response of NiO-based sensors fabricated on glass (G-550) and SiO2/Si (S-550) substrates. Here, sensors were tested at 175 °C for three different H2 concentrations of 0.1%, 0.2% and 0.3%. The S-550 sensor has a higher response than G-550 sensor. According to the XRD patterns and SEM images of this sample, the G-550 film consists of smaller grains than S-550. Although these two samples were fabricated at the same conditions, the main difference was observed in crystal size but not in electrical properties. Therefore, large porosity and crystal defects in G-550 film can be attributed to faster H2 sensitivity. Fig. 6 shows the dynamic response of G-500 (a) and S-700 (b) sensors operated at 175 °C. It can be see that the G-500 has faster response than S-700 supporting the previous interpretation regarding the results in Fig. 5. When the calcination temperature was increased the low energy crystal defects were eliminated from the NiO films and thus, longer response time could be observed. Fig. 7 displays one complete cycle of response and recovery of the different sensors versus elapsed time for H2 gas concentration of 3000 ppm. By comparing response time (defined as the time required to reach 90% of the final equilibrium value) of different samples, the response time was calculated as 153 s, 248 s, 393 s and 1019 s for sensors fabricated with G-500, G-550, S-550 and S-700, respectively. It can be noted that the G-500 and S-700 have shortest and longest response times, respectively. This fact indicates that the increase of
Fig. 5. Dynamic response characteristics of NiO thin films exposed to 1000, 2000 and 3000 ppm of H2 at 175 °C: (a) G-550 and (b) S-550.
Fig. 6. Dynamic response characteristics of NiO thin films exposed to 1000, 2000 and 3000 ppm of H2 at 175 °C: (a) G-500 and (b) S-700.
grain size can decrease the change of carrier concentration in NiO due to low porosity and low surface to volume ratio of the crystallites. Fig. 8 compares the response of sensors at different concentrations for four sensors which prepared in this study. When the calcination temperature was increased from 500 to 700 °C, a slight increase in gas sensor response was observed. These facts indicate that crystallinity plays a significant role in metal oxide gas sensors [26]. It is also known that the NiO has a catalytic effect for producing H2 ions at the surface of NiO at elevated temperatures. Also, adsorption/desorption phenomena are enhanced at elevated temperatures [27]. These results clearly support the current understanding of interaction between H2 and metal oxide surfaces. Fig. 9a shows the dependence of sensitivity of G-550 sample on wide range of H2 concentrations in the air at 175 °C. It clearly indicates that the NiO thin films can be used to detect H2 quantitatively. The dependence of sensor response is linear for low H2 concentrations; however, it changes at a higher concentration due to the limitation of exposed sites on the NiO grains. Repeatability is an important factor related to practical applications [28,29]. Fig. 9b shows the repeatability of the G-550 sample under 3000 ppm of H2 gas at 175 °C. It indicates that though sensor runs for few subsequent cycles, no significant change can be observed among response and recovery and response
Fig. 7. The response for all NiO thin films to 3000 ppm of H2 at 175 °C.
2234
AM. Soleimanpour, A.H. Jayatissa / Materials Science and Engineering C 32 (2012) 2230–2234
4. Conclusion
Fig. 8. The response of NiO thin films to 1000, 2000 and 3000 ppm of H2 at 175 °C.
times. Small deviation can be attributed to small error in inserting target gas with mass flow meters etc. The p-type behavior of NiO is due to non-stoichiometry of the prepared samples. Vacancies in cation sites play a key role in the change of carrier concentration in NiO thin films. For each metal vacancy, two electron holes are formed. Oxygen molecules adsorbed on NiO surface at room temperature [30]. At elevated temperatures (100–300 °C), the oxygen molecules adsorb on the surface of the grains and form O − and O2−, which can increase the hole density and decrease the resistance [31,32]. With introducing the H2 molecules to the chamber, H2 molecules replaced the adsorbed oxygen molecules and produce water vapor. This effect decreases hole concentration and therefore, the resistance of NiO film increases [33]. There are only a few metal oxides exhibiting p-type semiconducting behavior and gas sensing capabilities [9,31]. When we use n-type metal oxide films for gas sensor fabrication, the resistance reduces in the presence of reducing gases such as H2. Thus maximum achievable response is less than 100%. However, when a p-type metal oxide semiconductor is used, the increase of resistance can be larger than 100% and thus, signal to noise ratio of gas sensors can be maintained at a higher level. Therefore, use of p-type metal oxide to monitor reducing gases is favorable over n-type metal oxide semiconductors.
Fig. 9. (a) The response of G-550 to 1000–40,000 ppm of H2 at 175 °C and (b) repeatability data for G-550 to 3000 ppm of H2 at 175 °C.
Thin films of NiO were fabricated by a sol–gel coating process on glass and SiO2/Si substrates. The microstructure of these films indicates that the size of crystallites, surface morphology and the film composition depend on the substrate material and calcination temperature. The electrical and gas sensing properties indicate that the NiO films have a p-type conductivity with low thermal activation energy of electrical conductivity. Also, the band gap of the materials is within the values reported in the literature. The sensor response depends on the concentration, testing temperature and calcination temperature and the substrate material. Therefore, all these parameters can be optimized to manufacture the gas sensors to monitor H2 in the air at a low temperature, where flammable H2 gas can be detected in a wide range of concentrations. The gas sensing properties of NiO sensors indicated that the gas sensors could be operated below 200 °C with reasonably fast response time and recovery time. The concentration of H2 below 0.1% in the ambient conditions can be easily detected with the NiO thin film-based sensors. Acknowledgment This work was supported by the National Science Foundation (Grant number: CMMI 0933069) of USA. References [1] A. Qureshi, A. Mergen, A. Altindal, Sens. Actuators, B 135 (2009) 537–540. [2] I. Hotovy, J. Huran, L. Spiess, R. Capkovic, S. Hascik, Vacuum (2000) 300–307. [3] M. Kitao, K. Izawa, K. Urabe, T. Komatsu, S. Kuwano, S. Yamada, Jpn. J. Appl. Phys. 33 (1994) 6656–6662. [4] A.M. Soleimanpour, Y. Hou, A.H. Jayatissa, Appl. Surf. Sci. 257 (2011) 5398–5402. [5] K. Yoshimura, T. Miki, S. Tanemura, Jpn. J. Appl. Phys. 34 (1995) 2440–2446. [6] H. Kumagai, M. Matsumoto, K. Toyoda, M. Obara, J. Mater. Sci. Lett. 15 (1996) 1081–1083. [7] D. Adler, J. Feinleib, Phys. Rev. B 2 (1970) 3112–3134. [8] I. Hotovy, J. Huran, P. Siciliano, S. Capone, L. Spiess, V. Rehacek, Sens. Actuators, B 103 (2004) 300–311. [9] C. Cantalini, M. Post, D. Buso, M. Guglielmi, A. Martucci, Sens. Actuators, B 108 (2005) 184–192. [10] P.S. Patil, L.D. Kadam, Appl. Surf. Sci. 199 (2002) 211–221. [11] W.-L. Jang, Y.-M. Lu, W.-S. Hwang, W.-C. Chen, J. Eur. Ceram. Soc. 30 (2010) 503–508. [12] I. Hotovy, J. Huran, J. Janı, A.P. Kobzev, Vacuum 51 (1998) 157–160. [13] E. Lindahl, M. Ottosson, J.O. Carlsson, Surf. Coat. Technol. 205 (2010) 710–716. [14] B. Sasi, K.G. Gopchandran, Sol. Energy Mater. Sol. Cells 91 (2007) 1505–1509. [15] A.A. Al-Ghamdi, W.E. Mahmoud, S.J. Yaghmour, F.M. Al-Marzouki, J. Alloys Compd. 486 (2009) 9–13. [16] G. Korotcenkov, Mater. Sci. Eng., R 61 (2008) 1–39. [17] O.S. Heavens, Optical Properties of Thin Solid Films, Butterworths Scientific Publications, 1955. [18] A.F. Mayadas, M. Shatzkes, Phys. Rev. B 1 (1970) 1382–1389. [19] A. Bosman, C. Crevecoeur, Phys. Rev. B 144 (1966) 763–770. [20] M.N. Islam, M.O. Hakim, H. Rahman, J. Mater. Sci. 22 (1987) 1379–1384. [21] A. Mallikarjuna Reddy, A. Sivasankar Reddy, P. Sreedhara Reddy, Vacuum 85 (2011) 949–954. [22] K. Kondoh, M. Mizuhashi, J. Non-Cryst. Solids 178 (1994) 189–198. [23] R. Romero, F. Martin, J.R. Ramos-Barrado, D. Leinen, Thin Solid Films 518 (2010) 4499–4502. [24] A.C. Sonavane, A.I. Inamdar, P.S. Shinde, H.P. Deshmukh, R.S. Patil, P.S. Patil, J. Alloys Compd. 489 (2010) 667–673. [25] M. Gautam, A.H. Jayatissa, Mater. Sci. Eng., C 31 (2011) 1405–1411. [26] Y.-D. Wang, L.-F. Yang, Z.-L. Zhou, Y.-F. Li, X.-H. Wu, Mater. Lett. 49 (2001) 277–281. [27] W. Wei, X. Jiang, L. Lu, X. Yang, X. Wang, J. Hazard. Mater. 168 (2009) 838–842. [28] H. Meixner, U. Lampe, Sens. Actuators, B 33 (1996) 198–202. [29] H. Steinebach, S. Kannan, L. Rieth, F. Solzbacher, Sens. Actuators, B 151 (2010) 162–168. [30] S.T. Shishiyanu, T.S. Shishiyanu, O.I. Lupan, Sens. Actuators, B 107 (2005) 379–386. [31] K. Wetchakun, T. Samerjai, N. Tamaekong, C. Liewhiran, C. Siriwong, V. Kruefu, A. Wisitsoraat, A. Tuantranont, S. Phanichphant, Sens. Actuators, B 160 (2011) 580–591. [32] M. Stamataki, D. Tsamakis, N. Brilis, I. Fasaki, A. Giannoudakos, M. Kompitsas, Phys. Status Solidi A 205 (2008) 2064–2068. [33] A.H. Jayatissa, P. Samarasekara, G. Kun, Phys. Status Solidi A 206 (2009) 332–337.