Nitrogen dioxide sensing properties of sprayed tungsten oxide thin film sensor: Effect of film thickness

Journal of Colloid and Interface Science 451 (2015) 245–254

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science www.elsevier.com/locate/jcis

Nitrogen dioxide sensing properties of sprayed tungsten oxide thin film sensor: Effect of film thickness V.V. Ganbavle a, S.V. Mohite a, G.L. Agawane b, J.H. Kim b, K.Y. Rajpure a,⇑ a b

Electrochemical Materials Laboratory, Department of Physics, Shivaji University, Kolhapur 416004, M.S., India Department of Materials Science and Engineering, Chonnam National University, 300 Yongbong-Dong, Buk-Gu, Gwangju 500-757, South Korea

g r a p h i c a l a b s t r a c t NO2 concentration100 ppm o Operating temperature=200 C

100 100 ml

Gas response (%)

80

75 ml

60

125 ml

40

50 ml

20

0 0

100

200

300

400

500

600

700

800

Time (s)

a r t i c l e

i n f o

Article history: Received 4 March 2015 Accepted 1 April 2015 Available online 7 April 2015 Keywords: NO2 sensor WO3 Spray pyrolysis XPS Photoluminescence Selectivity

a b s t r a c t We report a study on effect of film thickness on NO2 sensing properties of sprayed WO3 thin films. WO3 thin films varying in thicknesses are deposited onto the glass substrates by simple spray pyrolysis technique by varying the volume of spray solution. Thin film gas sensors are characterized by using X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), atomic force microscopy (AFM) and photoluminescence (PL) techniques to study their physical properties. Film having thickness 745 nm has shown highest gas response of 97% with 12 and 412 s response and recovery times, respectively towards 100 ppm NO2 concentration. Gas response of 20% is observed towards 10 ppm NO2 at 200 °C operating temperature. Sensitivity of the optimal sensor is 0.83%/ppm when operating at 200 °C with 10 ppm lower detection limit. The response of the sensor is reproducible and WO3 films are highly selective towards NO2 in presence of mist of various interfering gases viz. H2S, NH3, LPG, CO and SO2. Ó 2015 Elsevier Inc. All rights reserved.

1. Introduction The perception of the surroundings based on mere senses is related closely to the progress of human/living being life. The smell ⇑ Corresponding author. Fax: +91 231 2691533. E-mail address: [email protected] (K.Y. Rajpure). http://dx.doi.org/10.1016/j.jcis.2015.04.001 0021-9797/Ó 2015 Elsevier Inc. All rights reserved.

or more specifically, the gas detection is a quite complicated sensorial practice which has an effect on human’s decisions and actions. Humans are unable to detect number of gases such as CO, CO2, CH4, H2, NO, NO2 by smell above certain concentrations which can have fatal consequences if remained undetected and uncontrolled [1]. The massive usage of vehicles produce polluting gases such as CO, NOx, SO2, particles, and hydrocarbons [2] thus,

V.V. Ganbavle et al. / Journal of Colloid and Interface Science 451 (2015) 245–254

it is obligatory to develop sensitive gas sensors to detect even subppm of specific analyte gas in a complex gas mixture [3]. According to various studies done so far and available reports by the various environment monitoring/protection agencies significant air pollutants are O3, NO2, SO2, CO, and particulate matter (lead) [2,4]. Among these pollutants risks to human lives due to NO2 is high. When inhaled, NO2 can be easily detected by its smell but at the same time, it is toxic. NO2 detection by its smell is easiest way provided its concentration is below 4 ppm; otherwise it anaesthetizes the nose [5]. Thus, many times when NO2 is leaked it remains unnoticed without use of some external sensing agency which may result in potential health risks. As far as NOx sensing is concerned, WO3 is the most popularly used material. Currently some of the WO3 sensors are capable of measuring accurately in the 10 ppb range, which is well below the emergency exposure limit for human [6]. WO3 nanowires were prepared by a vapour transport method using WO3 powder as a raw material. The sensor made of the nanowires as thin as 50 nm showed the highest response towards 3 ppm NO2 at a low operating temperature of 100 °C [7]. Irregular nanosheets of WO3 have been prepared and high response to NO2 even at sub-ppm level was observed by You et al. [8]. Using TeO2 templates An et al. [9] synthesized WO3 nanotubes and used for the detection of low NO2 concentrations ranging from 1 to 50 ppm. A report discussing the effect of thickness variation of the WO3 thin film on the NO response is available [10]. It is reported that thickness of the film affects the response and recovery kinetics of the sensor [11]. Dependence of the gas sensing performance on the thickness of thin films for various gases and materials is studied [12–17]. Desired thickness of the films is tuneable by controlling the preparative parameters. In spray pyrolysis method, by controlling the quantity of precursor solution to be sprayed, desired thickness of the thin films can be achieved. We have made an attempt to develop a gas sensor based on WO3 thin films deposited by simple spray pyrolysis technique. Variation in the thickness of the films has effect on various properties such as morphology, porosity, defects, composition, etc. and hence on the gas sensing behaviour. The effect of thickness variation on the physicochemical and NO2 sensing properties of WO3 thin films deposited by varying the film thickness is discussed in the present work.

The deposited films were characterized by various techniques. To identify the crystal structure, the thin films were characterized by X-ray diffraction (XRD) technique and the diffraction patterns were recorded using Bruker D2 phaser X-ray diffractometer using Cu Ka radiation of wavelength 1.5406 Å. Monochromatic X-ray beam of energy 1253.6 eV was used in X-ray photoelectron spectroscopy (XPS) study (XPS, Physical Electronics PHI 5400, USA); which reveals information about chemical composition and valence state of the elements. Surface morphology and topography of the films were studied using scanning electron microscopy (SEM, Model JEOL JSM-6701F, Japan) and atomic force microscopy (AFM, Digital Instrument, Nanoscope III). Absorption spectra obtained from UV– Visible spectrophotometer (UV–vis 1800 Spectrophotometer, Shimadzu) was used to determine optical band gap. Room temperature photoluminescence (PL) spectra were recorded using Perkin–Elmer luminescence spectrometer model: LS55 to study the defects. Thickness of the films was measured by means of surface profiler (XP-1 Stylus Profiler, Ambios Technology Inc.). Gas sensing measurements were carried out in locally fabricated gas sensing unit equipped with Keithley electrometer (6514). Sensor of size 1 cm  1 cm was fabricated and silver contacts were drawn for good electrical contacts. Thin film sensor was mounted in 250 ml airtight container where it is preheated at required temperature using temperature controller. Sensor was heated until its resistance stabilized and the time required for this was around three hours. Thin film sensors were then exposed to the analyte gas of desired concentration in the gas sensor unit and change in the resistance was monitored using Keithley (6514) electrometer. After each successive measurement, fresh air was passed into the test box and then required amount of analyte gas was injected into the box to obtain a desired concentration. Selectivity studies were carried out by monitoring change in resistance of the film by purging various gases of desired concentration. Various canisters of CO, NH3, SO2, LPG, H2S and NO2 gases having 2000 ppm gas concentration were used as analyte gases procured from Shreya Enterprises Pvt. Ltd. Mumbai, Maharashtra, India. 3. Results and discussion 3.1. X-ray diffraction studies

2. Experimental

XRD patterns of WO3 thin films with different thickness along with the stick pattern of JCPDS card are presented in Fig. 1. XRD

PTA thus formed was then diluted using double distilled water to 15 mM and used for the spray deposition. Thickness of the films was varied by varying solution quantity of the solution in four subsequent steps as 50, 75, 100, and 125 ml and the films were labelled as W345, W675, W745 and W801, respectively according to their thickness (Table 1). PTA was sprayed on to the preheated glass substrates at optimized substrate temperature of 425 °C [19]. The nozzle to substrate distance was kept 28 cm and the spray rate was maintained at optimized value of 5 ml/cc.

(200)

(142)

(-223) (041) (-232)

(022) (-202) (202)

745 nm (222)

ð1Þ

801 nm

(120) (-112)

The substrates were cleaned by the procedure discussed elsewhere [18]. The peroxotungstic acid (PTA) of 0.5 mM concentration was prepared by dissolving (2.757 g) tungsten metal powder (AR grade, 99%; loba chimie Pvt. Ltd., Mumbai) in 30 ml of 30% hydrogen peroxide (H2O2) (AR grade; Thomas Bakers, Pvt. Ltd. Mumbai). The solution was initially kept in an ice bath for an hour to control the exothermic reaction. It is then rigorously stirred for 48 hours at room temperature until all the powder was completely dissolved according to reaction (1).

(002) (020)

2.1. Materials synthesis

W þ H2 O2 þ H2 O ! WO3  nH2 O2  mH2 O þ D

2.2. Materials characterization

Intensity (A.U.)

246

675 nm JCPDS 43-1035 20

25

30

35

40

45

50

345 nm 55

60

2θ (degree) Fig. 1. X-ray diffraction patterns of WO3 thin with different thickness.

247

V.V. Ganbavle et al. / Journal of Colloid and Interface Science 451 (2015) 245–254

Solution quantity (ml)

Thickness (nm)

Crystallite size ‘D’ (nm)

50 75 100 125

345 675 745 801

28 57 69 67

400

W 4f

C 1s W 4d3/2 W 4d5/2

600

200

0

Fig. 2. Survey scan XP spectra of WO3 thin films with 675 and 745 nm thickness.

1.6

Table 1 Variation of crystallite size and thickness of thin films deposited with various solution quantities.

800

Binding energy (eV)

3.2. X-ray photoelectron spectroscopy

675 nm

W 4f and 5p

4f7/2 35.91 eV

4f5/2 38.04 eV

1.4

6+

1.0

W

5

Intensity, 10 (cps)

1.2

0.8 0.6

4f5/2 37.21 eV 4f7/2 34.38 eV

5p3/2 41.28 eV

0.4 0.2

5+

W

0.0 46

44

42

40

38

36

34

32

30

Binding energy (eV)

1.6

745 nm

W 4f and 5p

1.4

4f7/2 35.55 eV

4f5/2 37.68 eV

1.2

6+

1.0

W

5

Intensity, 10 (cps)

The X-ray photoelectron (XP) spectra of two representative samples viz. W675 and W745 are shown in Fig. 2. Stoichiometry and study of valance states of WO3 thin films was carried out by using XPS technique. Moreover, this technique also gives information about band structure of material and valence band maximum. It is well studied that tungsten (W) can exist in various chemical states from 4+ to 6+ by forming covalent bond with oxygen and a series of oxides having well-characterized structural features are formed [23]. XPS data indicates that the WO3 films are sub-stoichiometric. The survey scan spectra consists of several peaks belonging to various core levels of W and O which indicates that W and O are the main constituent elements of the films. Along with these, presence of C is observed which is originated from surface adsorbed carbon species. Various peaks due to Auger electron emission and satellite shake ups near to the photoelectron peaks are also observed. The doublet separation between various core levels of W (p, d and f) decreases with increase in l which is in accordance with theory. The peak at 927 eV is due to the O KLL Auger electron transition [24]. The peak at around 36 eV is due to the electrons from 4f core level and peaks at 248 and 260 eV correspond to 4d5/2 and 4d3/2 core levels of tungsten (W), respectively. Peaks at 428 and 496 eV are assigned to 4p3/2 and 4p5/2 core levels of W [25]. The stoichiometry decreases for higher solution quantities and thus the oxygen content has been reduced with increase in film thickness. Fig. 3 shows the detailed narrow scan W 4f core level spectra of W675 and W745 thin films. Both the spectra exhibit well resolved doublets which were deconvoluted using Gaussian fitting and Shirley background was subtracted. Deconvolution gives four peaks belonging to two doublets with different intensities indicating presence of W in two chemical states viz. W5+ (blue) and W6+ (red), respectively. Fig. 3(a) shows intense 4f doublet belonging to W6+ oxidization state of W with the components W 4f7/2 peak

1000

O 1s W 4p1/2 W 4p3/2

1200

W 4s

675 nm

O KLL

745 nm

Intensity (A.U.)

patterns match well with the JCPDS card No. 43-1035 confirming monoclinic crystal structure. A characteristic triplet of peaks viz. (0 0 2), (0 2 0), (2 0 0) of monoclinic phase of WO3 [20] is observed for all the films with preferred orientation along (2 0 0). The peak position is found to be independent of the film thickness however the peak intensity depends on the thickness. Peak intensity increases with increase in the film thickness, indicating increase in the crystalline component of the films with increase in solution quantity. The intensity of the (2 0 0) peak increases till W745 and decreases with further increase in film thickness W801. The decrease in peak intensity for W801 is due to the formation of powdery material which can be wiped easily thereby making surface of the films less adherent [21]. The crystallite size (D) estimated using the Scherrer formula corresponding to most intense (2 0 0) plane is presented in Table 1. It increases up to W745; however for W801 with more spray quantity (125 ml), the crystallite size seems to be saturated. Similar behaviour of the crystallite size as a function of film thickness is reported by Bouderbala et al. [22]. Gas response of the films increased with increase in the crystallite size and gas response is maximum for W745 film.

0.8 0.6

5p3/2 40.95 eV

4f5/2 36.85 eV

4f7/2 34.04 eV

0.4 0.2

5+

W

0.0 46

44

42

40

38

36

34

32

30

Binding energy (eV) Fig. 3. Narrow scan XP spectra of the 4f core level of W of WO3 thin film with 675 and 745 nm thickness.

at 35.91 eV and W 4f5/2 peak at 38.04 eV. Peak with lower intensity for 4f doublet having components 4f7/2 at 34.38 eV and 4f5/2 at 37.21 eV associated with W5+ oxidation state is seen Fig. 3(a) [26]. Fig. 3(b) shows intense 4f doublet belonging to W6+ oxidization state (red color) of W with components W 4f7/2 peak at

V.V. Ganbavle et al. / Journal of Colloid and Interface Science 451 (2015) 245–254

3.3. Scanning electron microscopy

3.0

675 nm

O 1s

531.05 eV

2.8

2.4 2.2

5

Intensity, 10 (cps)

2.6

2.0 1.8 1.6

532.59 eV

1.4 1.2 1.0 538

536

534

532

530

528

526

Binding Energy (eV) 3.0 2.8

745 nm

O 1s

530.95 eV

2.6 2.4

5

35.55 eV and the W 4f5/2 peak at 37.68 eV. The lower intensity 4f doublet having components 4f7/2 at 34.04 eV and 4f5/2 at 36.85 eV associated with W5+ oxidation state (blue colour) are also visible [26]. The BE of all the 4f doublets of W in this work are in good agreement with the literature value characterizing the W6+ and W5+states in WO3 (Table 2) [27,28]. The ratio of area under the curves of the hexavalent W for both films is 0.75, which is supported by the spin-orbit splitting theory of 4f levels [26]. The second doublet associated with pentavalent W (W5+) with poor intensity has a lower binding energy than that of the doublet associated with W6+. With increase in oxidation state, the number of electrons of the atom taking part in screening of core electrons decreases. This causes the increase in BE at higher oxidation state of an element. Due to the occurrence of oxygen vacancy, the electronic density of corresponding W atom increases which raise the screening of its nucleus and thus, the 4f level energy is expected to be at a lower binding energy [29]. The intensity ratio of W5+/W6+ is 0.23 and 0.25 for W675 and W745 films, respectively. This implies the decrease in stoichiometry of the films with increase in film thickness. This is also observed in the photoluminescence (PL) studies where density of oxygen deficiency increases with increase in film thickness. Gas response increases with increase in the film thickness and it is observed to be maximum for the W745 film. The narrow scan XP spectra of O 1s core level is shown in Fig. 4(a) and (b) for W675 and W745, respectively. Both spectra show shoulders due to the different kinds of oxygen species. With increase in film thickness, slight shift in the BE of lattice oxygen towards lower energy side is observed confirming decrease in the stoichiometry of the film with higher thickness [30].

Intensity, 10 (cps)

248

2.2 2.0 1.8 1.6

532.57 eV

1.4 1.2

Fig. 5 shows SEM images of the WO3 thin films with different thickness. For W345, wire network of the film starts to grow and with increase in film thickness for the film W745 dense network of the wires is seen. For W675 the internal network of wire-like morphology of the films is just started to grow from single scale to dual scale network where network of two different size wires is observed. The wire network becomes denser with increase in film thickness and gives rise to dual scale wire network (Fig. 5(e)) for W745 film. The network of thick wires of 0.5– 0.8 lm diameters and another internal network with diameters around 0.1–0.3 lm inside the outer network is clearly visible in Fig. 5(d). This internal network of wires is due to the collision and decomposition of fine droplets onto the substrate where repetition of the decomposition and overlapping of droplets takes place as the solution quantity to be sprayed is increased. This may give rise to dual scale wire network (Fig. 5(e)). This network of thin wires inside the thicker wires, collapse with increase in solution quantity to 100 ml (W745). Spherical grains having diameter ranging from 0.1 to 0.3 lm are observed inside the network of wires making films more porous having protrusion of size 0.3–0.8 lm size. This increases the effective surface area for the gas adsorption and hence the gas response. This is an indication

1.0 538

536

534

532

530

528

526

Binding Energy (eV) Fig. 4. Narrow scan XP spectra of O1s oxygen core level of WO3 thin film with 675 ml and 745 nm thickness.

that the pores formation induces diffusion process allowing the gas molecules to diffuse within the surface of film and improves the gas response. Porous morphology is observed for W745 which exhibits highest gas response [31]. However, it can be achieved either by reducing the grain size or by obtaining porous structures. Smaller grains could be transferred into the larger grains by continuous heating films and which causes drift in the response of the films over the time [32]. Thus, porous microstructure of the films is preferred in gas sensor application. Porous morphology is observed for W745 which exhibits highest gas response. This is due to its porous dual scale surface morphology that enhances diffusion and adsorption of analyte gas. For much higher film thickness i.e. for W801, dual scale wire network is seen to be collapsed and morphology of film changes to more compact network. 3.4. Atomic force microscopy

Table 2 Comparison of observed and standard doublet separation for various core levels of tungsten for films having 675 and 745 nm film thickness. Core level of WO3

4f (+6) 4f (+5) 4d 4p

Observed doublet separation (eV) W675

W745

2.13 2.83 12.49 68.42

2.13 2.81 12.35 67.64

Standard doublet separation (eV)

2.10 – 12.50 67.10

Fig. 6 shows AFM images of WO3 thin films with different thickness. AFM images are recorded in tapping mode over 3  3 lm2 scan area. Table 3 presents route mean square (RMS) roughness of the films. When solution quantity increased, agglomeration of the grains is observed and spherical grains have developed between the wire structures of the WO3 films. Agglomeration of the grains is observed with increase in film thickness due to the large amount of available material. Small and spherical grains are observed for W675 and number of grains increases for W745 film. With further

V.V. Ganbavle et al. / Journal of Colloid and Interface Science 451 (2015) 245–254

249

Fig. 5. SEM images of WO3 thin films with (a) 345, (b) 675, (c) 745 and (d) 801 nm thickness, (e) lower magnification SEM image of WO3 thin film with 745 nm thickness.

increase in solution quantity, for W801, agglomeration of the grains resulting in larger grains. The size of these spherical grains increases with increasing film thickness as observed in Fig. 6(d). This results in decrease in surface to volume ratio which in turn decreases gas response at higher film thickness. Excessively smaller grain size is not suitable since drift in gas response is observed due to coalescence of smaller grains into larger grains [32]. Statistics based on the AFM images have been used to investigate RMS roughness of the WO3 films. Its evaluation is strongly influenced by solution quantity sprayed and thus film thickness. It is observed that with increase in film thickness, roughness of the films increases and thus effective surface area available for the gas adsorption increases. This in turn helps increasing the gas response of the films till W745. Overgrown grains (whitish region) of larger size are observed for W801 which affects the gas sensing property. 3.5. Optical absorption and photoluminescence studies Optical absorption is measured over the range of 300–1100 nm (not shown). Absorption increases with increase in solution

quantity due to increased film thickness. Graph of (ahm)1/2 versus hm was plotted (not shown) and indirect band gap of the WO3 was found to be 2.67 eV which is in well agreement with reported band gap of the WO3 [33]. Room temperature photoluminescence (PL) spectra of WO3 thin films with different thickness recorded over 350–700 nm range by using 325 nm excitation wavelength is shown Fig. 7. In order to recognize contributions from all components, PL spectrum is deconvoluted into multipeaks using Gaussian fitting (inset of Fig. 7). After deconvolution, various peaks aroused due to different types of defects are observed and are mentioned in Table 4. Various peaks due to the near band edge emission (NBE), due to the O and W vacancies, band to band transition, and transitions due to interstitial atoms are observed. NUV peaks (389 and 409 nm) at higher energy side are attributed to an electron hole radiative recombination in deep level defects; the lower-energy peak was assigned to localized states in the band gap [34]. NUV emission band at around 389–409 nm is due to the deep level oxygen vacancies [35]. The stronger NUV emission is indicative of the existence of high degree of oxygen vacancies in the WO3 samples. It is observed that NUV

250

V.V. Ganbavle et al. / Journal of Colloid and Interface Science 451 (2015) 245–254

Fig. 6. AFM images of WO3 thin films with (a) 345, (b) 675, (c) 745 and (d) 801 nm thickness.

Table 3 RMS roughness of WO3 thin films with different thickness. Thickness (nm)

RMS roughness (nm)

345 675 745 801

47 75 83 106

peak intensity is minimum for W745 film; it decreases first (for 75 ml) and then increases for higher film thickness. Blue emissions near band edge at the wavelength 465 and 487 nm are due to the interstitial oxygen [36], and green emission at the wavelength 2.4

1.8

Intensity, 10 (cps)

2.2

6

Intensity, 10 (cps)

6

2.0 1.8 1.6 1.4 1.2

1.6 1.4 1.2 1.0 0.8 0.6 0.4

3.6. Electrical resistivity study

0.2 0.0 350

400

50 ml 125 ml 100 ml 75 ml

1.0 0.8 0.6

450

500

550

600

650

700

Variation of resistance with respect to inverse of temperature (1000/T) is shown in Fig. 8 for the WO3 thin films with different

Wavelength (nm)

Table 4 Peak values of fitted PL spectra of WO3 thin films with different thickness.

0.4 0.2 0.0 350

531 nm occurs because of oxygen at interstitial sites forming (WO3 + Oi) [37]. The blue emission at around 486 nm is attributed to the band-to-band transition of wire-like WO3 structures as already pointed out by researchers [35,38]. This is the strongest emission observed compared with all other emissions. It has to be noticed that emission through quantum confinement effects can be ruled out because the diameter of the WO3 wires is larger than that of the critical radius [39]. Band to band transition around 487 nm is observed for all the films and its peak position does not vary with film thickness and excitation wavelength. All PL emission spectra feature an additional weak red emission peaks at 616 and 647 nm. The intensity of these peaks increases first till W745 and then decreases as the film thickness increases. Red emission is attributed to the monovalent ions (H+) at the interstitial sites which is also reported by Shi et al. [40]. These H+ ions are either the ions remained unreacted or occur due to the adsorbed water in the film surface which is also observed in XPS studies.

Thickness (nm)

400

450

500

550

600

650

700

Wavelength (nm) Fig. 7. Photoluminescence spectra of WO3 thin films with different thickness. Inset: Gaussian fitting of PL spectrum of film with 745 nm thickness.

345 675 745 801

Peak position (from left) 1 (nm)

2 (nm)

3 (nm)

4 (nm)

5 (nm)

6 (nm)

7 (nm)

8 (nm)

390 390 389 389

407 407 409 410

433 432 439 439

466 466 465 465

486 487 487 488

530 530 531 531

615 616 616 617

646 647 647 647

251

V.V. Ganbavle et al. / Journal of Colloid and Interface Science 451 (2015) 245–254

thickness. Relatively high resistivity is observed for the W345 film and it decreases with increase in film thickness. As the crystallinity increases, grain size increases thereby minimizing the grain boundary scattering losses, which results in decrease in resistance with increased film thickness. So the number of electron trap states reduces and hence the carrier concentration increases [41]. Also, increase in the thickness of films results in decreasing the grain boundary effect due to the broadening neck due to the agglomeration. This leads to elongated mean free path which in turn decreases the scattering and thus, the resistivity of the film [42]. Lower electrical conductivity of WO3 films is due to the oxygen deficiency and interstitial W atoms. Various earlier reports, [43,44] reveals that a very small decrease in oxygen content of WO3 can affect its electronic structure and increase in the conductivity by significant amount is observed. Electrons induced by these defects contribute to conduction, resulting in n-type WO3 films. The resistivity decreases with increasing temperature, because with the addition of thermal energy, electron could be set free from the lattice for conduction. Initially at room temperature resistance of the film is within the order of a few giga ohms but with increase in temperature, the resistance of the films fall to a value of tens of mega ohm. Maximum gas response is observed at 200 °C which implies that optimum value of resistance is necessary for application of film as gas sensor. Activation energy for all the films is calculated using Arrhenius equation and tabularized in Table 5. Activation energy increases with increase in film thickness till W745 may be due to the variation in the stoichiometry which increases the collision of the charge carrier concentration. 3.7. Gas sensing properties Gas response of the sensor is measured over a range of operating temperatures and for varied gas concentrations. Selectivity of the sensor is examined by measuring gas response towards various oxidizing and reducing gases. Gas response is calculated by using Eq. (2).

S¼

  jRg  Ra j  100 Ra

ð2Þ

Response time is measured as time taken by the sensor to reach 90% of its maximum gas response upon purging analyte gas in the test chamber and recovery time is the time taken by the sensor to reach 10% of the maximum gas response value upon removal of gas.

Table 5 Activation energy values of WO3 thin films with different thickness. Thickness (nm)

Activation energy (eV)

345 675 745 801

0.38 0.39 0.46 0.40

3.7.1. Determination of operating temperature It is well studied that sensor can be operated over a wide range of temperatures but highest response towards analyte gas is observed at a certain temperature [30]. Optimization of operating temperature was carried out by measuring gas response over the range of 50–250 °C. The gas response towards NO2 has been studied for W745 film at various temperatures and variation of gas response with respect to operating temperature towards 100 ppm NO2 is shown in Fig. 9. Overall, maximum response of 98% has been recognized at 200 °C. In addition, the influence of the operating temperature on kinetics of the sensor is remarkably significant. Thermal energy assists the adsorption of NO2 in the  form of NO 2 or 2NO by charge transfer from electrons or through the pre-adsorbed oxygen species, respectively. Gas response of the films does not exhibit significant difference at lower operating temperatures such as 50 °C, however, at 100 °C the gas response is increased to 18% and it is irreversible. Resistance of the sensor does not regain its original value after removal of the NO2 gas even after 3 h. A slight increase in gas response is observed at 150 °C while at 200 °C highest gas response towards NO2 is noticed, therefore further measurements were carried out at 200 °C. The improved sensor performance due to increase in operating temperature can be ascribed to the enhanced adsorption reactions and dissociation of adsorbed analyte species which increase the diffusion through the wire structure of WO3 in the surface. This causes maximum coverage to target gas thereby increasing gas response [12]. The dynamic NO2 sensing transients of WO3 thin films with different thickness are presented in Fig. 10. Gas response increases with increase in film thickness till W745 and it decreases with further increase in film thickness. Increased crystallinity, improved surface morphology and roughness are the things which enhance the gas response. Response and recovery times for all the films towards 100 ppm NO2 operated at 200 °C are summarised in Table 6. The response of the sensor is quick but it takes more time

17 16 15

90

745 nm 675 nm 345 nm

13

80

Gas response (%)

14

ln (ρ )

100 NO2 concentration = 100 ppm

801 nm

12 11 10 9 8

70 60 50 40 30 20

7

10

6

0

5 2.0

2.2

2.4

2.6

2.8

3.0

3.2

1000/T (K-1) Fig. 8. Variation of resistivity ln (q) with respect to inverse of temperature for the WO3 thin films with different thickness.

50

100

150

200

250

o

Operating temperature ( C) Fig. 9. Variation of gas response of W745 film towards 100 ppm NO2 at different operating temperatures.

V.V. Ganbavle et al. / Journal of Colloid and Interface Science 451 (2015) 245–254

Recovery time (s)

9 24 12 15

193 294 412 679

and 100 ppm. This is because of saturation in adsorption at higher concentrations of NO2. With further increase in NO2 concentration (>100 ppm), the gas response remains nearly unchanged. Response and recovery times of the W745 thin film sensor are shown in Table 7. It is observed that, response time increases with increase in gas concentration till 80 ppm and remains nearly constant for 100 ppm NO2 concentration. Recovery time of the sensor increases linearly with increase in gas concentration. Recovery time is relatively much higher than the response time. This is due to the porous films structure which takes more time to desorb all the gas molecules from the sensor surface. Further higher gas response is observed with increased NO2 concentration; having large number of gas molecules adsorbed on the film surface which have to be desorbed. Sensitivity of the sensor is calculated by plotting graph of gas concentration versus gas response. Slope of the graph gives the sensitivity of the sensor and it is observed to be 0.83%/ppm for W745 film operating at 200 °C. Higher response is observed for W745 film without modifying the sensor surface with any catalyst such as Pt or Pd. This fact motivated to test sensor response at lower concentrations of 10 ppm which is half of the immediately dangerous limit to health concentration (IDLH). Gas response for W745 film towards 10 ppm NO2 at 200 °C is measured and shown in Fig. 12. Also, the reproducibility of the sensor is assessed for 10 ppm NO2 in six successive cycles of measurements and exhibited 20% gas response. Variation in the sensor response is less than 2% for various measurements at same gas concentration of NO2. Also, the recovery time for lower gas concentrations is low which supports use of the sensor for lower concentration detection. To determine the selectivity of sensors, the response towards H2S, NH3, LPG, SO2, and CO is calculated and plotted at the fixed gas concentration of 100 ppm and operating temperature of 200 °C as shown in Fig. 13. It is observed that sensor W745, exhibits extreme response towards NO2 in comparison with various interfering gases and little response towards H2S, NH3, LPG, SO2, and

NO2 concentration=100 ppm o Operating temperature=200 C

745 nm

o

Gas response (%)

Gas response (%)

80

675 nm

60

801 nm

40

345 nm

20

Operating temperature=200 C

100

100 ppm

100

Response time (s)

345 675 745 801

80 ppm

3.7.3. Effect of NO2 concentration on gas response and selectivity studies The NO2 sensing transient response for W745 sensor is shown in Fig. 11. Five consecutive cycles were recorded for 40 min of measurements towards various concentrations of NO2 ranging from 20 to 100 ppm. The results indicate that the sensor shows less response towards lower concentrations and it increases with increase in gas concentration which is consistent with theory. It can be seen that W745 shows the highest gas response with fast response time (12 s) at 200 °C operating temperature. Gas response increases linearly with increase in NO2 concentration and it saturates at higher gas concentration (not shown). Initially, the response increases linearly when the NO2 concentrations are increased from 20 to 60 ppm and it increases gradually for 80

Thickness (nm)

80

60

40

60 ppm

3.7.2. NO2 sensing mechanism Upon exposure of film to NO2, absorption of NO2 followed by chemisorption occurs in two possible ways. NO2 has an unpaired electron. After forming covalent bond between nitrogen and oxygen, one of the atoms remains with a single unpaired electron. This single unpaired electron is the reason why chemisorption of NO2 is more likely than that of the other gases. This unpaired electron in N of NO2 forms bond with the surface oxygen of WO3 and subsequently promote the chemisorptions [46]. Chemisorption of NO2 increases the width of potential barrier more than other gases hence giving higher change in resistance as compared with other gases. Interaction of NO2 with the surface of the sensor is of two types. The monomolecular adsorption of NO2, which is observed at oxidized surface, whereas; the dissociation of NO2 takes place at the oxygen deficient centres [47]. Thus, partially reduced cations i.e. W+5 play the role of NO2 chemisorption centres at comparatively low temperature 200 °C as discussed elsewhere [30].

Table 6 Response and recovery times of WO3 thin films towards 100 ppm NO2 operated at 200 °C.

40 ppm

for recovery of the sensor. For porous microstructure of thin film, the gas diffuses and chemisorbed at grain boundaries well inside the surface and it takes longer time to completely desorb the gas. This causes the increase in recovery time of the sensor with increase in film thickness. It is important to note that the sensor resistance is not completely recovered to initial value for all the films in the given time interval. Maximum gas response is observed for the W745 film and it decreases for higher film thickness, as for W801. Response kinetics obeys Arrhenius law which implies that increase in working temperature leads to decrease in response and recovery times [45].

20 ppm

252

20

0

0 0

100

200

300

400

500

600

700

800

Time (s) Fig. 10. Transient gas response of WO3 thin films with different thickness towards 100 ppm NO2 at different operating temperatures.

0

400

800

1200

1600

2000

2400

Time (s) Fig. 11. Transient gas response of W745 film towards various NO2 concentrations at 200 °C operating temperature.

253

V.V. Ganbavle et al. / Journal of Colloid and Interface Science 451 (2015) 245–254 Table 7 Response and recovery times of WO3 film with thickness 745 nm towards various NO2 concentrations at 200 °C operating temperature. Gas concentration (ppm)

Response time (s)

Recovery time (s)

20 40 60 80 100

3 8 11 12 12

151 204 280 328 412

Table 8 Selectivity coefficient of WO3 film with thickness 745 nm towards various gases. Gas

Selectivity coefficient

H2S NH3 LPG CO SO2

14 11 12 10 8

4. Summary and conclusions

10 ppm

Gas response (%)

20

15

10

5

0 0

500

1000

1500

Time (s) Fig. 12. Reproducible transient gas response of W745 film towards 10 ppm NO2 concentration at 200 °C operating temperature.

References

100 Gas concentration=100 ppm o Operating temperature=200 C

Gas response (%)

80

60

40

20

0

H2S

NH3

LPG

CO

SO2

NO2

Gas Fig. 13. Selectivity of W745 film for various gases at 200 °C towards 100 ppm gases concentration.

CO, gases. In order to quantify the selectivity, the selectivity coefficient of the sensor is calculated using Eq. (3) and the values are tabularized in Table 8.

St K ¼ Si

The effect of film thickness on the physicochemical and gas sensing properties of WO3 films deposited on glass substrates is studied. Sensor response towards a range of concentrations of NO2 and operating temperatures was measured. XPS and PL analysis revealed that the films are oxygen deficient and this deficiency increases with increase in film thickness. SEM and AFM studies showed porous structure of the films with high surface roughness which enhanced gas response. Higher gas response (97%) was observed for the film with 745 nm thickness. The sensor was found to respond rapidly, reproducibly, and selectively towards NO2 at 200 °C operating temperature with 10 ppm lower detection limit. The gas response towards 10 ppm NO2 is 20% and W745 exhibits maximum sensitivity around 0.83%/ppm. The response time of the sensor was around 12 s, and recovery time was 412 s for optimal sensor towards 100 ppm NO2 and operating at 200 °C. Collectively, porous nanostructured WO3 thin film of thickness 745 nm deposited by spraying 100 ml of precursor solution is suitable for the detection of less concentration of NO2 at moderately low operating temperature.

ð3Þ

Selectivity coefficient varies from 8 to 14 for W745 film operating at 200 °C towards 100 ppm gas concentration. Pronounced selectivity towards NO2 in presence of SO2 is observed, which is another major representative coexisting air pollutant. Films are more selective towards NO2 due to highly oxidizing and reactive nature of NO2 [48] and the operating temperature 200 °C is also optimum for selective NO2 sensing.

[1] A. Tricoli, M. Righettoni, A. Teleki, Angew. Chem. Int. Ed. 49 (2010) 7632–7659. [2] http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=194645, U.S. Environmental Protection Agency. Integrated science assessment for oxides of nitrogen-Health criteria, EPA/600/R-08/071, July 2008. [3] L. Wang, A. Teleki, S.E. Pratsinis, P.I. Gouma, Chem. Mater. 20 (2008) 4794– 4796. [4] B. Brunekreef, S.T. Holgate, The Lancet 360 (2002) 1233–1242. [5] G.F. Fine, L.M. Cavanagh, A. Afonja, R. Binions, Sensors 10 (2010) 5469–5502. [6] M.M. Arafat, B. Dinan, S.A. Akbar, A.S.M.A. Haseeb, Sensors 12 (2012) 7207– 7258. [7] D. Meng, N.M. Shaalan, T. Yamazaki, T. Kikuta, Sens. Actuators B: Chem. 169 (2012) 113–120. [8] L. You, Y.F. Sun, J. Ma, Y. Guan, J.M. Sun, Y. Du, G.Y. Lu, Sens. Actuators B: Chem. 157 (2011) 401–407. [9] S. An, S. Park, H. Ko, C. Lee, Ceram. Int. 40 (2014) 1423–1429. [10] Z. Meng, A. Fujii, T. Hashishin, N. Wada, T. Sanada, J. Tamaki, K. Kojima, H. Haneokad, T. Suzukid, J. Mater. Chem. C 3 (2015) 1134–1141. [11] Y. Liu, J. Parisi, X. Sun, Y. Lei, J. Mater. Chem. A 2 (2014) 9919–9943. [12] P.S. Shewale, G.L. Agawane, S.W. Shin, A.V. Moholkar, J.Y. Lee, J.H. Kim, M.D. Uplane, Sens. Actuators B: Chem. 177 (2013) 695–702. [13] R.R. Salunkhe, C.D. Lokhande, Sens. Actuators B: Chem. 129 (2008) 345–351. [14] X. Du, S.M. George, Sens. Actuators B: Chem. 135 (2008) 152–160. [15] H.J. Nam, T. Sasaki, N. Koshizaki, J. Phys. Chem. C 111 (2007) 9105–9109. [16] C. Mercado, Z. Seeley, A. Bandyopadhyay, S. Bose, J.L. McHale, ACS Appl. Mater. Interfaces 3 (2011) 2281–2288. [17] Y. Liu, J. Parisi, X. Sun, Y. Lei, J. Mater. Chem. A 2 (2014) 9919–9943. [18] V.V. Ganbavle, G.L. Agawane, A.V. Moholkar, J.H. Kim, K.Y. Rajpure, J. Mater. Eng. Perform. 23 (2014) 12041213. [19] V.V. Ganbavle, J.H. Kim, K.Y. Rajpure, J. Electron. Mater. 44 (2015) 874–885. [20] D. Meng, T. Yamazaki, Y. Shen, Z. Liu, T. Kikuta, Appl. Surf. Sci. 256 (2009) 1050–1053. [21] A.V. Moholkar, S.M. Pawar, K.Y. Rajpure, P.S. Patil, C.H. Bhosal, J. Phys. Chem. Solids 68 (2007) 1981–1988. [22] M. Bouderbala, S. Hamzaoui, B. Amrani, Ali H. Reshak, M. Adnane, T. Sahraoui, M. Zerdali, Physica B 403 (2008) 3326–3330. [23] Y. Baek, K. Yong, J. Phys. Chem. C 111 (2007) 1213–1218. [24] S. Fardindoost, A.I. zad, F. Rahimi, R. Ghasempour, Int. J. Hydrogen Energy 35 (2010) 854–860. [25] M. Ranjbar, S. Fardindoost, S.M. Mahdavi, A.I. zad, N. Tahmasebi, Sol. Energy Mater. Sol. Cells 95 (2011) 2335–2340. [26] S. Ma, F.G. Amar, B.G. Frederick, J. Phys. Chem. A 107 (2003) 1413–1423.

254

V.V. Ganbavle et al. / Journal of Colloid and Interface Science 451 (2015) 245–254

[27] R. Nyholm, A. Berndtsson, N. Martensson, J. Phys. C: Solid State Phys. 13 (1980) L1091–L1096. [28] W. Grunert, E.S. Spiro, R. Feldhaus, K. Anders, G.V. Antoshin, K.M. Minachev, J. Catal. 107 (1987) 522–534. [29] P.J. Gellings, H.J.M. Bouwmesster, The CRC Handbook of Solid State Electrochemistry, CRC Press Inc., Cleveland Ohio, 1997. pp. 656. [30] V.V. Ganbavle, S.V. Mohite, J.H. Kim, K.Y. Rajpure, Curr. Appl. Phys. 15 (2015) 84–93. [31] F. Sun, S.B. Liu, F.L. Meng, J.Y. Liu, Z. Jin, L.T. Kong, J.H. Liu, Sensors 12 (2012) 2610–2631. [32] Y.F. Sun, S.B. Liu, F.L. Meng, J.Y. Liu, Z. Jin, L.T. Kong, J.H. Liu, Sensors 12 (2012) 2610–2631. [33] P.P. Gonzalez-Borrero, F. Sato, A.N. Medina, M.L. Baesso, A.C. Bento, G. Baldissera, C. Persson, G.A. Niklasson, C.G. Granqvist, A.F. Silva, Appl. Phys. Lett. 96 (2010) 061909. [34] K. Lee, W.S. Seo, J.T. Park, J. Am. Chem. Soc. 125 (2003) 3408–3409. [35] J.Y. Luo, F.L. Zhao, L. Gong, H.J. Chen, J. Zhou, Z.L. Li, S.Z. Deng, N.S. Xu, Appl. Phys. Lett. 91 (2007) 093124. [36] H.L. Hartnagel, A.L. Dawar, A.K. Jain, C. Jagadish, Semiconducting Transparent Thin Films, Institute of Physics Publishing, Bristol and Philadelphia, 1995. [37] J. Diaz-Reyes, J.E. Lores-Mena, J.M. Gutierrez-Arias, M.M. Morincastillo, H. Azucena-Courtecatl, M. Galvan, P. Rodriguwz-Fragoso, A. Mendez-Lopez,

[38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48]

Advances in Sensors, Signals and Materials, Optical and structural properties of WO3 as a function of the annealing temperature, WSEAS Press, 2010. pp. 99– 104. J. Wang, P.S. Lee, J. Ma, J. Cryst. Growth 311 (2009) 316–319. M.T. Chang, L.J. Chou, Y.L. Chueh, Y.C. Lee, C.H. Hsieh, C.D. Chen, Y.W. Lan, L.J. Chen, Small 3 (2007) 658–664. C. Shi, Y. Wei, X. Yang, D. Zhou, C. Guo, J. Liao, H. Tang, Chem. Phys. Lett. 328 (2000) 1–4. S.S. Shinde, P.S. Shinde, C.H. Bhosale, K.Y. Rajpure, J. Phys. D Appl. Phys. 41 (2008). 105109-6. K.Y. Rajpure, C.D. Lokhande, C.H. Bhosale, Thin Solid Films 311 (1997) 114– 118. S. Bai, K. Zhang, L. Wang, J. Sun, R. Luo, D. Li, A. Chen, J. Mater. Chem. A 2 (2014) 7927–7934. R. Chatten, A.V. Chadwick, A. Rougier, P.J.D. Lindan, J. Phys. Chem. B 109 (2005) 3146–3156. A.A. Vasiliev, V.V. Malyshev, Sens. Actuators B: Chem. 189 (2013) 260–267. W. Gao, Y. Ling, X. Liu, J. Sun, Int. J. Min., Metall. Mater. 19 (2012) 1142–1148. A. Stanoiu, C.E. Simion, S. Somacescu, Sens. Actuators B: Chem. 186 (2013) 687–694. A.A. Tomchenko, G.P. Harmer, B.T. Marquis, J.W. Allen, Sens. Actuators B: Chem. 93 (2003) 126–134.