- Email: [email protected]
Vacuum deposited WO3 thin films based sub-ppm H2S sensor
Materials Chemistry and Physics 134 (2012) 851e857
Contents lists available at SciVerse ScienceDirect
Materials Chemistry and Physics journal homepage: www.elsevier.com/locate/matchemphys
Vacuum deposited WO3 thin films based sub-ppm H2S sensor Niyanta Datta a, Niranjan Ramgir a, *, Manmeet Kaur a, Mainak Roy b, Ranu Bhatt a, S. Kailasaganapathi a, A.K. Debnath a, D.K. Aswal a, S.K. Gupta a a b
Thin Films Devices Section, Technical Physics Division, Bhabha Atomic Research Center, Trombay, Mumbai 400 085, India Solid State Chemistry Section, Chemistry Division, Bhabha Atomic Research Center, Trombay, Mumbai 400 085, India
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 July 2011 Received in revised form 13 February 2012 Accepted 8 March 2012
A simple method of vacuum deposition using W foils has been utilized to fabricate Au-incorporated WO3 thin film sensors. Incorporation of Au has been demonstrated to improve both the sensitivity and the selectivity of the sensor films towards H2S. The effect of operating temperature, Au loading and gas concentrations have been investigated and correlated with the observed sensitivity values to determine the optimum conditions for realizing H2S sensor with better sensing properties. The sensor film containing 2.32 at.% Au detected H2S selectively with an enhanced sensitivity of S ¼ 16 (1 ppm) at an operating temperature of 250 C. The enhanced response kinetics has been confirmed further using Raman and work function measurements. Ó 2012 Elsevier B.V. All rights reserved.
Keywords: Thin films Vacuum deposition Defects Surface properties
1. Introduction WO3, an n-type wide band gap material, has demonstrated its potential for realizing highly sensitive and selective sensors towards different gases [1,2]. Similar to other semiconducting oxides like SnO2 and ZnO, its surface is characterized by the presence of oxygen vacancies. These oxygen vacancies act as donor levels and contribute in the governing sensing mechanism [3]. Various sensors including C2H5OH, O3, NH3, NOx, H2S, have been attempted and realized using WO3 in conventional thin films as well as nanoforms [4e7]. Thin film based sensors have been commonly deposited using techniques like micromachining, sputtering, thermal evaporation and solegel [8]. To improve sensitivity and selectivity, the host matrix is generally modified with various sensitizers including Pd, Pt, Ag, Au, and often a bimetallic sensitizer is also considered [9e12]. In particular, Au incorporation resulted in improved sensing characteristics like faster response time and increased sensitivity values and accordingly, has been found to be a promising sensitizer [13]. Modulation of sensing properties based on Au has been mainly ascribed to the electronic sensitization mechanism. Use of Au has been demonstrated to improve the sensing performance towards different gases like NO2, NOx H2S, H2, and NH3 [14e18]. Of these hydrogen sulfide (H2S) is one of the
* Corresponding author. Tel.: þ91 22 25595839; fax: þ91 22 25505296. E-mail address: [email protected] (N. Ramgir). 0254-0584/$ e see front matter Ó 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.matchemphys.2012.03.080
highly toxic and flammable gases that is being employed extensively in various industrial applications. In particular, its exposure affects human’s nervous system and could cause loss of consciousness at a very low concentration. Its threshold limit value has been set to 10 ppm and hence detection at low or sub-ppm concentrations has become the focus of research. Besides, as the catalytic decomposition of H2S is known to occur between 300 and 400 C [19], the H2S sensors must be operated below 300 C. Ionescu et al. have reported the sub-ppm detection of H2S upto 20 ppb with a sensitivity value of w1.1 (S ¼ 1.25 towards 50 ppb) at an operating temperature of 250 C. Although the sensors responses were good, nanoparticles formation via advanced gas deposition unit is tedious [2]. Tao et al. have investigated the effect of different doping element on H2S sensing properties of WO3 films deposited by RF plasma sputtering [10]. There report lacks the systematic investigation of doping element and its concentration in the host matrix to realize a better sensing material. Hence to know and understand under what conditions sensitizers imparts sensitivity towards a particular gas demands a detailed investigation. In the present work we report a systematic investigation of the gas sensing properties of pure and Au-incorporated WO3 sensor thin films towards H2S. The sensor films have been realized using a simple method of vacuum deposition. We have investigated the effect of sensor operating temperature, Au loading, and gas concentration on sensing properties to determine the optimum conditions for the detection of H2S gas. Additionally, Raman and work function measurements both with and without the test gases
852
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857
The steps involved in the fabrication of sensor films are illustrated schematically in Fig. 1. In brief, Al2O3 substrates were first cleaned by ultrasonicating in solvents namely trilene, acetone and methanol. Contact electrodes were then predefined on these substrates using Pt-wire and a high temperature drying Au paste followed by curing at 850 C for 10 min. The predefined electrodes were protected by masking with molybdenum foils. W films (w600 nm thick) were then vacuum deposited using W-foil (purity: 99.99%, dimensions 1 4 cm) under application of a very high current (w80 A) in a base vacuum of 2 104 mbar. In order to prepare Au-incorporated WO3 films a sandwich layer of W(300 nm)/Au (t nm)/W (300 nm) (t taking values 1.86, 3.3, 7.4 and 13.3 nm corresponding to 0.55, 1.23, 2.32 and 3.95 at.%, as measured using energy dispersive X-ray analysis (EDAX)) were deposited onto alumina substrates. The resulting films were then subjected to post deposition annealing at 600 C for 1 h in quartz tubular furnace under constant oxygen (O2) flow of 100 sccm. In order to make electrical contacts 120 nm thick Au layer was deposited using a mask having 400 mm spacing.
room temperature. Prior to measurements sensor samples were exposed to H2S at an elevated temperature of w100 C. The change in signal was better observed for higher dose of exposure (100 ppm) and hence was chosen for the experiments. Kelvin probe contact potential difference (CPD) measurements were carried out using SKP Kelvin Probe 4.5 from KP Technology Ltd. UK. All the measurements were performed at room temperature and ambient conditions using Au electrode having tip diameter of 2 mm. For better average value of the work function the electrode tip was scanned across the sample surface (Raster scan) and the relative variation in the CPD was measured. Sensor samples were exposed to 100 ppm of H2S at an elevated temperature of w100 C before measuring the work function [20]. Sensing measurements were performed in a static system as described elsewhere [21]. In brief, sensor films were mounted in a stainless steel test-chamber (volume: 250 cm3) equipped with a temperature control unit. The desired temperature was achieved using a Pt-wire based heater attached to the backside of the sensor film. The desired concentration of the test gas was achieved by injecting the measured quantity of commercial gas inside the chamber. Commercial grade gases having 1000 ppm concentration of desired gas in N2, obtained in half liter cylinders were used to control the final concentration. The resistance of the film was monitored and acquired as a function of time using a personal computer equipped with Labview software. Recovery of the sensors was achieved by opening the housing to the atmosphere. Sensor response or sensitivity (S) was calculated using the relation:
2.2. Instruments and analysis
S ¼
Surface morphology of the as-grown films was investigated using scanning electron microscopy, SEM, (TESCAN, model: TS 5130MM) equipped with EDAX unit. X-ray diffraction measurements were performed using Rigaku expert machine with CuKa radiation having 1.5406 Å wavelength. Raman spectra were recorded on a Labaram-I spectrometer in a backscattering geometry at
for reducing gases namely H2S, CH4 and NH3, and
have been performed to elucidate the nature of interaction with the sensor film. Our studies indicates that Au-incorporated WO3 sensor films could detect sub-ppm (50 ppb, S ¼ 2.7) H2S selectively at an operating temperature of 250 C. 2. Experimental section 2.1. Sensor fabrication
S ¼
Ra Rg
Rg Ra
(1)
(2)
for oxidizing gases namely Cl2, NO2 and NO, where Ra and Rg are resistances in air and test gases, respectively. Response and
Fig. 1. Steps involved in the fabrication of (a) pure WO3 and (b) Au-incorporated WO3 sensor devices.
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857
853
recovery times were defined as the time needed for 90% of total change on exposure to test gas and fresh air, respectively. 3. Results and discussion 3.1. Structural and morphological characteristics The resultant W films after vacuum depositions are black in color, which when oxidized at an elevated temperature of 600 C in O2 environment gets transformed into greenish yellow color. Fig. 2(a)e(e) shows the SEM images of pure and Au-incorporated WO3 thin films. It is seen that pure film consists of large number of microparticles having an average grain size of w5.4 mm containing microcracks and even cavities. Additionally, these particles are seemed to consist of large number of nanosized grains with size <100 nm. The sensor films color changes from greenish yellow to violet purple upon incorporation of Au. Besides it also causes elimination of microcracks. Additionally, grain size decreases (Fig. 2(f)) and porosity increases with Au incorporation. Fig. 3 shows the XRD profile of Au-incorporated (2.32 at.%) WO3 thin film. The pattern matches with the monoclinic crystal structure of WO3 [22] having lattice parameters of a ¼ 7.309, b ¼ 7.545 and c ¼ 7.689 Å, respectively. 3.2. Gas sensing properties of pure and Au-incorporated WO3 sensor films Fig. 4 shows the sensing response curves recorded for pure WO3 towards 1 ppm of H2S at an operating temperature of 250 and 225 C, respectively. The sensitivity of sensor film towards H2S is not affected much with the operating temperature however a strong variation in sensor films recovery time is seen. Recovery time is observed to decrease with increase in operating temperature. From Fig. 4(c) it is clear that the sensor responded with fairly good sensitivity and a reasonable recovery time at an operating temperature of 250 C and hence was chosen for all the further experiments. To determine the optimum Au concentration in the host matrix a systematic investigation of Au loading and its effect on sensitivity
Fig. 3. XRD profile of Au (2.32 at.%) incorporated WO3 sensor thin film.
towards H2S have been carried out. A typical bell shaped curve (S vs. Au concentration) as shown in Fig. 5(a) with maxima for WO3 thin films containing 2.32 at.% Au was observed. The low Au concentration has a discontinuous distribution causing insufficient catalytic activity leading to smaller sensitivity, while higher concentration gives continuous distribution leading to lower initial resistance and thence lower sensitivity value [23]. An overall trend of decrease in the air resistance (measured at fixed temperature of 250 C) with Au incorporation is seen (Fig. 5(a)) that further corroborates the findings. Fig. 5(b) shows the selectivity histogram of pure and Au-incorporated (2.32 at.%) sensor film at an operating temperature of 250 C towards 1 ppm of different gases. Interestingly, both pure and Au-incorporated samples exhibited a maximum sensitivity towards H2S. However, upon Au incorporation the sensor became highly sensitive and thereby selective towards H2S. A negligible response towards all the other interfering gases namely NH3, Cl2, CO, CH4, CO2 and NO2 was observed. Our initial findings suggested that pure WO3 responded with maximum sensitivity towards H2S in comparison to that of Au modified films
Fig. 2. SEM images of (a) pure and Au-incorporated WO3 thin films containing (b) 0.55, (c) 1.23, (d) 2.32 and (e) 3.95 at.% Au, and (f) a plot of average grain size vs. Au content in WO3 thin film.
854
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857
Fig. 4. Sensors response curves at (a) 250, (b) 225 C and (c) the variation in sensitivity and recovery times as a function of an operating temperature for pure WO3 thin films measured upon exposure to 1 ppm of H2S.
[24]. One of the reasons for the difference in the sensitivity values in the two cases could be the gas sensing measurement units that were used in both the cases. In particular, earlier measurements were performed in a static system 250 cm3 (cylindrical) with one end enclosed/rested on a heater (w5 cm diameter). In the present work the sensing measurements were preformed with Pt-wire (R ¼ 2e3 U) heater attached at the backside of the sensor thereby providing a localized heating. In former case at high temperatures the whole set-up was heating up and is probably responsible for the difference in measurements. A systematic investigation of the effect of gas concentration was then performed at an operating temperature of 250 C. Fig. 6 shows the response curves of pure and Au-incorporated WO3 sensor films recorded with the increasing concentration of test gas. As is evident from the figure sensitivity is found to increase with gas concentration. Au-incorporated WO3 films exhibited a better sensing performance in comparison with that of pure WO3 sensor films. For example, towards 0.7 ppm of H2S, pure WO3 films showed a sensitivity of 2.9 with a response and recovery time of 106 s and 32 min, while Au-incorporated sample (2.32 at.%) exhibited an increase in the sensitivity to 12 with a response and recovery times of 88 s and 18 min, respectively. Moreover, Au-incorporated thin films were able to detect as low as 50 ppb of H2S with a sensitivity value of 2.7. For practical applications, the reproducibility and the stability of sensing-device are also important. In the present case, the sensors developed under identical conditions (same and different batches) exhibited similar response towards H2S. Moreover, the stability measurements performed over a period of 15 days indicate them to be quite stable. For example, the sensor sample with 2.32 at.% Au exhibited a variation of about 8% in its initial resistance value for this period. A detailed investigation of the long term stability of the sensor films is underway.
3.3. Raman and work function measurements Raman investigations of the sensor film before and after exposure to H2S have also been performed as it is known to give the “fingerprint” of the material. Fig. 7(a) shows the Raman spectra (recorded at room temperature) of a bare WO3 films before and after exposure to H2S. Both the spectra invariably exhibit a series of sharp peaks at wavenumbers lower than 250 cm1 attributed to vibrational modes of the WO3 lattice i.e. the translational and rotational motion of WO6 octahedra in the same unit cell. They also exhibit peaks at intermediate frequency region (200e400 cm1) assigned to the bending modes of OeWeO bridging bonds and at high frequency region (600e900 cm1) attributed to the stretching modes of WeO bonds [25,26]. The Raman peaks observed at approximately 60, 73, 136, 274, 328, 718 and 809 cm1 and a shoulder at w96 cm1 are characteristic of crystalline monoclinic g-phase of WO3. The monoclinic g-phase is the most stable room temperature phase of WO3 and has been reported for both WO3 powder and nanoforms [27,28]. The fingerprint Raman peak at approximately 33 cm1 could not be detected for the samples due to experimental limitations [notch cut off]. Additionally, the absence of the 950 cm1 Raman peak in both the samples is indicative of their high crystalline nature [29]. Apart from the Raman peaks of the g-phase, an additional peak at w40 cm1 attributed to the orthorhombic b-phase is observed. The low frequency Raman modes involve tilted motion of the oxygen cages around the W atoms and/translational motion of WO6 octahedra in the same unit cell. Strong polarizability modulation of these lattice vibrational modes during the evolution of a new phase makes Raman spectroscopy a more sensitive technique than XRD for studying phase transition in WO3. Reversible transition from g to b phase has been reported at w300 C for microcrystalline WO3
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857
855
Fig. 5. Response curves of (a) pure and (c) Au-incorporated (2.32 wt.%) WO3 thin films towards sub-ppm concentration of H2S at an operating temperature of 250 C and corresponding variation in sensitivity with gas concentration (c) and (d).
powder [25] and for nanowires above 227 C [26]. However, room temperature observation of the b-phase for the vapor deposited thin film samples is an interesting and notable exception. Upon exposure to H2S, intensity of the Raman peaks due to the g-phase relative to that of the peak at 40 cm1 (considered to be the fingerprint of the b-phase) reduced significantly, but their peak positions remained almost unaltered. The two peaks at w60 and 73 cm1 almost merged to form a single broad band. Since, interaction of WO3 with H2S is likely to occur at the film surface, the observation indicates a further stabilization of the b-phase on the film surface at the expense of the g-phase. Additionally a decrease
in the intensity of the peaks corresponding to the stretching modes was also observed. Raman spectra, before and after exposure to H2S, of Au-incorporated WO3 are depicted in Fig. 7(b). Although the spectrum of the unexposed sample is qualitatively similar to that of bare WO3, intensity of the peak at w40 cm1 is much higher than that at 60 cm1. Moreover, the broad hump at w96 cm1 in the spectrum of bare WO3 appears as a distinct peak for AueWO3. These clearly indicate the relative enhancement of the orthorhombic b-phase. Upon exposure to H2S, intensity of the lattice and stretching modes reduced marginally, whereas no significant change in the intensity
Fig. 6. (a) Effect of Au loading on sensitivity towards 1 ppm H2S, and on the air resistance at an operating temperature of 250 C. (b) Selectivity histogram pure WO3 and Auincorporated WO3 thin film sensors towards different gases at an operating temperature of 250 C.
856
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857
Fig. 7. Raman spectra recorded for (a) pure WO3 and (b) Au-incorporated WO3 sensor films before and after exposure to H2S gas at room temperature.
of the bending modes of WO3 was observed. Interestingly, in contrast to the bare WO3, intensity of the peak at w40 cm1 relative to that at 60 cm1 got reduced after exposure to H2S. Augmentation of Raman signals due to the surface enhanced resonance Raman effects of the embedded Au may contribute towards the anomalous intensity variation for AueWO3 sensor films. Fig. 8 shows the work function area scan recorded for 3.3 mm2 of sensor samples before and after exposure to 100 ppm of H2S. Upon exposure to H2S, the CPD over the measured area is observed to vary. The color variation indicates the localized variation in the CPD. Table 1 illustrates the final work function values of the sensor samples calculated taking the average of the relative variation in the CPD. As is clearly evident the incorporation of Au in the WO3 resulted in work function decrease by 0.25 eV. Besides exposure to H2S causes changes in the work function of pure and Au-incorporated WO3 samples by 30 and 11 meV, respectively. The small
variation in the work function of Au-incorporated sample is attributed to the faster reaction kinetics (recovery time). This result is in accordance with and corroborates further the findings of Raman investigations. 3.4. Gas sensing mechanism It is established that interaction of H2S with WO3 leads to the generation of large number of oxygen vacancies [10]. The interaction leads to the introduction of new energy levels in the mid gap states of WO3. Besides, incorporation of doping elements in particular Au has been known to impart the sensitivity in the host material mainly via electronic sensitization mechanism [26]. In this the additives enhance the rate of reaction by providing the additional active sites. The interaction with the test gas mainly occurs on the Au surface and the corresponding changes
Fig. 8. 2-D work function map of pure WO3 sensor films (a) before and (b) after exposure to H2S and Au-incorporated WO3 (c) before and (d) after exposure to H2S.
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857 Table 1 Work function values recorded for sensor samples before and after exposure to H2S. Sample name
F, before exposure (eV)
F, after exposure (eV)
Pure WO3 Au-incorporated WO3
5.221 4.973
5.191 4.962
are transferred immediately to the host matrix reflected as a fast drop in the resistance of the sample. This is expected to contribute for improving both the response and the recovery times of the sensor films. Further, the oxygen vacancies can easily be reoxidized upon exposure to fresh air thereby assuring complete recovery. 4. Conclusions In summary, we have deposited pure and Au-incorporated WO3 sensor thin films using a simple technique of vacuum deposition. Effect of operating temperature, gas concentration, and Au loading has been explored and correlated with the observed sensitivity values to determine the optimum condition for improved sensing performance towards H2S. The sensor film containing 2.32 at.% of Au detected H2S selectively and sensitively (S ¼ w16) at an operating temperature of 250 C. The enhanced reaction kinetics is mainly attributed to Au and has been confirmed using both the Raman and the work function measurements. Our results clearly indicate that the WO3 thin films can be made selective and sensitive towards sub-ppm H2S by a proper control over the process parameters namely operating temperature, doping element (Au) and its concentration in the host matrix. Acknowledgments This work is partly supported by “DAE-SRC Outstanding Research Investigator Award” (2008/21/05-BRNS) and “Prospective Research Funds” (2008/38/02-BRNS) granted to D.K.A. Authors also thank Dr. A.K. Chauhan and Mr. Arvind Kumar for their help with Xray diffraction and work function measurements, respectively.
857
References [1] V. Srivastava, K. Jain, Sens. Actuators B 133 (2008) 46. [2] R. Ionescu, A. Hoel, C.G. Granqvist, E. Llobet, P. Heszler, Sens. Actuators B 104 (2005) 132. [3] M. Bendahan, J. Guérin, R. Boulmani, K. Aguir, Sens. Actuators B 124 (2007) 24. [4] M. Akiyama, J. Tamaki, N. Miura, N. Yamazoe, Chem. Lett. (1991) 1611. [5] A. Labidi, E. Gillet, R. Delamare, M. Maaref, K. Aguir, Sens. Actuators B 120 (2006) 338. [6] C.N. Xu, N. Miura, Y. Ishida, K. Matsuda, N. Yamazoe, Sens. Actuators B 65 (2000) 163. [7] J.-H. Yoon, H.-J. Lee, Sens. Lett. 9 (2011) 46. [8] L. Jia, W. Cai, H. Wang, Appl. Phys. Lett. 96 (2010) 103115. [9] T. Maekawa, J. Tamaki, N. Miura, N. Yamazoe, Chem. Lett. (1992) 639. [10] W.H. Tao, C.-H. Tsai, Sens. Actuators B 81 (2002) 237. [11] B.T. Marquis, J.F. Vetelino, Sens. Actuators B 77 (2001) 100. [12] P. Montmeat, J.-C. Marchand, R. Lalauze, J.-P. Viricelle, G. Tournier, C. Pijolat, Sens. Actuators B 95 (2003) 83. [13] Z. Xu, J.F. Veletino, R. Lec, D.C. Parker, Electrical properties of tungsten trioxide films, J. Vac. Sci. Technol. A 8 (1990) 3634. [14] H. Xia, Y. Wang, F. Kong, S. Wang, B. Zhu, X. Guo, J. Zhang, Y. Wang, S. Wu, Sens. Actuators B 134 (2008) 133. [15] S.J. Ippolito, S. Kandasamy, K. Kalantar-zadeh, W. Wlodarski, Sens. Actuators B 108 (2005) 154. [16] X. Wang, N. Miura, N. Yamazoe, Sens. Actuators B 66 (2000) 74. [17] C.V. Ramana, S. Utsunomiya, R.C. Ewing, C.M. Julien, U. Becker, J. Phys. Chem. B 110 (2006) 10430. [18] M. Penza, G. Cassano, F. Tortorella, Sens. Actuators B 81 (2001) 115. [19] R.B. Slimane, F.S. Lau, R.J. Dihu, M. Khinkis, Production of hydrogen by superadiabatic decomposition of hydrogen sulfide, in: Proceedings of the 2002 U.S. DOE Hydrogen Program Review, NREL/CP-610-32405. [20] N.S. Ramgir, M. Ghosh, P. Veerender, N. Datta, M. Kaur, D.K. Aswal, S.K. Gupta, Sens. Actuators B 156 (2011) 875. [21] N.S. Ramgir, S. Kailasa Ganapati, M. Kaur, N. Datta, K.P. Muthe, D.K. Aswal, S.K. Gupta, J.V. Yakhmi, Sens. Actuators B 151 (2010) 90. [22] Powder Diffract, File, JCPDS Int. Centre Diffract. Data, PA 19073-3273, U.S.A., 2001. [23] R.S. Niranjan, V.A. Chaudhary, S.R. Sainker, K.R. Patil, K. Vijayamohanan, I.S. Mulla, Sens. Actuators B 79 (2001) 132. [24] N.S. Ramgir, S. Kailasaganapati, M. Kaur, S. Mishra, N. Datta, D.K. Aswal, S.K. Gupta, J.V. Yakhmi, AIP Conf. Proc. 1349 (2011) 679e680. [25] T. Siciliano, A. Tepore, G. Micocci, A. Serra, D. Manno, E. Filippo, Sens. Actuators B 133 (2008) 321. [26] J. Gabrusenoks, A. Veispals, A. von Czarnowski, K.-H. Meiwes-Broer, Electrochim. Acta 46 (2001) 2229. [27] E. Cazzanelli, C. Vinegoni, G. Mariotto, A. Kuzmin, J. Purans, Solid State Ionics 123 (1999) 67e74. [28] D.Y. Lu, J. Chen, H.J. Chen, L. Gong, S.Z. Deng, Y.L. Liu, N.S. Xu, Appl. Phys. Lett. 90 (2007) 041919. [29] K. Nonaka, A. Takase, K. Miyakawa, J. Mater. Sci. Lett. 12 (1993) 274.
Contents lists available at SciVerse ScienceDirect
Materials Chemistry and Physics journal homepage: www.elsevier.com/locate/matchemphys
Vacuum deposited WO3 thin films based sub-ppm H2S sensor Niyanta Datta a, Niranjan Ramgir a, *, Manmeet Kaur a, Mainak Roy b, Ranu Bhatt a, S. Kailasaganapathi a, A.K. Debnath a, D.K. Aswal a, S.K. Gupta a a b
Thin Films Devices Section, Technical Physics Division, Bhabha Atomic Research Center, Trombay, Mumbai 400 085, India Solid State Chemistry Section, Chemistry Division, Bhabha Atomic Research Center, Trombay, Mumbai 400 085, India
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 July 2011 Received in revised form 13 February 2012 Accepted 8 March 2012
A simple method of vacuum deposition using W foils has been utilized to fabricate Au-incorporated WO3 thin film sensors. Incorporation of Au has been demonstrated to improve both the sensitivity and the selectivity of the sensor films towards H2S. The effect of operating temperature, Au loading and gas concentrations have been investigated and correlated with the observed sensitivity values to determine the optimum conditions for realizing H2S sensor with better sensing properties. The sensor film containing 2.32 at.% Au detected H2S selectively with an enhanced sensitivity of S ¼ 16 (1 ppm) at an operating temperature of 250 C. The enhanced response kinetics has been confirmed further using Raman and work function measurements. Ó 2012 Elsevier B.V. All rights reserved.
Keywords: Thin films Vacuum deposition Defects Surface properties
1. Introduction WO3, an n-type wide band gap material, has demonstrated its potential for realizing highly sensitive and selective sensors towards different gases [1,2]. Similar to other semiconducting oxides like SnO2 and ZnO, its surface is characterized by the presence of oxygen vacancies. These oxygen vacancies act as donor levels and contribute in the governing sensing mechanism [3]. Various sensors including C2H5OH, O3, NH3, NOx, H2S, have been attempted and realized using WO3 in conventional thin films as well as nanoforms [4e7]. Thin film based sensors have been commonly deposited using techniques like micromachining, sputtering, thermal evaporation and solegel [8]. To improve sensitivity and selectivity, the host matrix is generally modified with various sensitizers including Pd, Pt, Ag, Au, and often a bimetallic sensitizer is also considered [9e12]. In particular, Au incorporation resulted in improved sensing characteristics like faster response time and increased sensitivity values and accordingly, has been found to be a promising sensitizer [13]. Modulation of sensing properties based on Au has been mainly ascribed to the electronic sensitization mechanism. Use of Au has been demonstrated to improve the sensing performance towards different gases like NO2, NOx H2S, H2, and NH3 [14e18]. Of these hydrogen sulfide (H2S) is one of the
* Corresponding author. Tel.: þ91 22 25595839; fax: þ91 22 25505296. E-mail address: [email protected] (N. Ramgir). 0254-0584/$ e see front matter Ó 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.matchemphys.2012.03.080
highly toxic and flammable gases that is being employed extensively in various industrial applications. In particular, its exposure affects human’s nervous system and could cause loss of consciousness at a very low concentration. Its threshold limit value has been set to 10 ppm and hence detection at low or sub-ppm concentrations has become the focus of research. Besides, as the catalytic decomposition of H2S is known to occur between 300 and 400 C [19], the H2S sensors must be operated below 300 C. Ionescu et al. have reported the sub-ppm detection of H2S upto 20 ppb with a sensitivity value of w1.1 (S ¼ 1.25 towards 50 ppb) at an operating temperature of 250 C. Although the sensors responses were good, nanoparticles formation via advanced gas deposition unit is tedious [2]. Tao et al. have investigated the effect of different doping element on H2S sensing properties of WO3 films deposited by RF plasma sputtering [10]. There report lacks the systematic investigation of doping element and its concentration in the host matrix to realize a better sensing material. Hence to know and understand under what conditions sensitizers imparts sensitivity towards a particular gas demands a detailed investigation. In the present work we report a systematic investigation of the gas sensing properties of pure and Au-incorporated WO3 sensor thin films towards H2S. The sensor films have been realized using a simple method of vacuum deposition. We have investigated the effect of sensor operating temperature, Au loading, and gas concentration on sensing properties to determine the optimum conditions for the detection of H2S gas. Additionally, Raman and work function measurements both with and without the test gases
852
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857
The steps involved in the fabrication of sensor films are illustrated schematically in Fig. 1. In brief, Al2O3 substrates were first cleaned by ultrasonicating in solvents namely trilene, acetone and methanol. Contact electrodes were then predefined on these substrates using Pt-wire and a high temperature drying Au paste followed by curing at 850 C for 10 min. The predefined electrodes were protected by masking with molybdenum foils. W films (w600 nm thick) were then vacuum deposited using W-foil (purity: 99.99%, dimensions 1 4 cm) under application of a very high current (w80 A) in a base vacuum of 2 104 mbar. In order to prepare Au-incorporated WO3 films a sandwich layer of W(300 nm)/Au (t nm)/W (300 nm) (t taking values 1.86, 3.3, 7.4 and 13.3 nm corresponding to 0.55, 1.23, 2.32 and 3.95 at.%, as measured using energy dispersive X-ray analysis (EDAX)) were deposited onto alumina substrates. The resulting films were then subjected to post deposition annealing at 600 C for 1 h in quartz tubular furnace under constant oxygen (O2) flow of 100 sccm. In order to make electrical contacts 120 nm thick Au layer was deposited using a mask having 400 mm spacing.
room temperature. Prior to measurements sensor samples were exposed to H2S at an elevated temperature of w100 C. The change in signal was better observed for higher dose of exposure (100 ppm) and hence was chosen for the experiments. Kelvin probe contact potential difference (CPD) measurements were carried out using SKP Kelvin Probe 4.5 from KP Technology Ltd. UK. All the measurements were performed at room temperature and ambient conditions using Au electrode having tip diameter of 2 mm. For better average value of the work function the electrode tip was scanned across the sample surface (Raster scan) and the relative variation in the CPD was measured. Sensor samples were exposed to 100 ppm of H2S at an elevated temperature of w100 C before measuring the work function [20]. Sensing measurements were performed in a static system as described elsewhere [21]. In brief, sensor films were mounted in a stainless steel test-chamber (volume: 250 cm3) equipped with a temperature control unit. The desired temperature was achieved using a Pt-wire based heater attached to the backside of the sensor film. The desired concentration of the test gas was achieved by injecting the measured quantity of commercial gas inside the chamber. Commercial grade gases having 1000 ppm concentration of desired gas in N2, obtained in half liter cylinders were used to control the final concentration. The resistance of the film was monitored and acquired as a function of time using a personal computer equipped with Labview software. Recovery of the sensors was achieved by opening the housing to the atmosphere. Sensor response or sensitivity (S) was calculated using the relation:
2.2. Instruments and analysis
S ¼
Surface morphology of the as-grown films was investigated using scanning electron microscopy, SEM, (TESCAN, model: TS 5130MM) equipped with EDAX unit. X-ray diffraction measurements were performed using Rigaku expert machine with CuKa radiation having 1.5406 Å wavelength. Raman spectra were recorded on a Labaram-I spectrometer in a backscattering geometry at
for reducing gases namely H2S, CH4 and NH3, and
have been performed to elucidate the nature of interaction with the sensor film. Our studies indicates that Au-incorporated WO3 sensor films could detect sub-ppm (50 ppb, S ¼ 2.7) H2S selectively at an operating temperature of 250 C. 2. Experimental section 2.1. Sensor fabrication
S ¼
Ra Rg
Rg Ra
(1)
(2)
for oxidizing gases namely Cl2, NO2 and NO, where Ra and Rg are resistances in air and test gases, respectively. Response and
Fig. 1. Steps involved in the fabrication of (a) pure WO3 and (b) Au-incorporated WO3 sensor devices.
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857
853
recovery times were defined as the time needed for 90% of total change on exposure to test gas and fresh air, respectively. 3. Results and discussion 3.1. Structural and morphological characteristics The resultant W films after vacuum depositions are black in color, which when oxidized at an elevated temperature of 600 C in O2 environment gets transformed into greenish yellow color. Fig. 2(a)e(e) shows the SEM images of pure and Au-incorporated WO3 thin films. It is seen that pure film consists of large number of microparticles having an average grain size of w5.4 mm containing microcracks and even cavities. Additionally, these particles are seemed to consist of large number of nanosized grains with size <100 nm. The sensor films color changes from greenish yellow to violet purple upon incorporation of Au. Besides it also causes elimination of microcracks. Additionally, grain size decreases (Fig. 2(f)) and porosity increases with Au incorporation. Fig. 3 shows the XRD profile of Au-incorporated (2.32 at.%) WO3 thin film. The pattern matches with the monoclinic crystal structure of WO3 [22] having lattice parameters of a ¼ 7.309, b ¼ 7.545 and c ¼ 7.689 Å, respectively. 3.2. Gas sensing properties of pure and Au-incorporated WO3 sensor films Fig. 4 shows the sensing response curves recorded for pure WO3 towards 1 ppm of H2S at an operating temperature of 250 and 225 C, respectively. The sensitivity of sensor film towards H2S is not affected much with the operating temperature however a strong variation in sensor films recovery time is seen. Recovery time is observed to decrease with increase in operating temperature. From Fig. 4(c) it is clear that the sensor responded with fairly good sensitivity and a reasonable recovery time at an operating temperature of 250 C and hence was chosen for all the further experiments. To determine the optimum Au concentration in the host matrix a systematic investigation of Au loading and its effect on sensitivity
Fig. 3. XRD profile of Au (2.32 at.%) incorporated WO3 sensor thin film.
towards H2S have been carried out. A typical bell shaped curve (S vs. Au concentration) as shown in Fig. 5(a) with maxima for WO3 thin films containing 2.32 at.% Au was observed. The low Au concentration has a discontinuous distribution causing insufficient catalytic activity leading to smaller sensitivity, while higher concentration gives continuous distribution leading to lower initial resistance and thence lower sensitivity value [23]. An overall trend of decrease in the air resistance (measured at fixed temperature of 250 C) with Au incorporation is seen (Fig. 5(a)) that further corroborates the findings. Fig. 5(b) shows the selectivity histogram of pure and Au-incorporated (2.32 at.%) sensor film at an operating temperature of 250 C towards 1 ppm of different gases. Interestingly, both pure and Au-incorporated samples exhibited a maximum sensitivity towards H2S. However, upon Au incorporation the sensor became highly sensitive and thereby selective towards H2S. A negligible response towards all the other interfering gases namely NH3, Cl2, CO, CH4, CO2 and NO2 was observed. Our initial findings suggested that pure WO3 responded with maximum sensitivity towards H2S in comparison to that of Au modified films
Fig. 2. SEM images of (a) pure and Au-incorporated WO3 thin films containing (b) 0.55, (c) 1.23, (d) 2.32 and (e) 3.95 at.% Au, and (f) a plot of average grain size vs. Au content in WO3 thin film.
854
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857
Fig. 4. Sensors response curves at (a) 250, (b) 225 C and (c) the variation in sensitivity and recovery times as a function of an operating temperature for pure WO3 thin films measured upon exposure to 1 ppm of H2S.
[24]. One of the reasons for the difference in the sensitivity values in the two cases could be the gas sensing measurement units that were used in both the cases. In particular, earlier measurements were performed in a static system 250 cm3 (cylindrical) with one end enclosed/rested on a heater (w5 cm diameter). In the present work the sensing measurements were preformed with Pt-wire (R ¼ 2e3 U) heater attached at the backside of the sensor thereby providing a localized heating. In former case at high temperatures the whole set-up was heating up and is probably responsible for the difference in measurements. A systematic investigation of the effect of gas concentration was then performed at an operating temperature of 250 C. Fig. 6 shows the response curves of pure and Au-incorporated WO3 sensor films recorded with the increasing concentration of test gas. As is evident from the figure sensitivity is found to increase with gas concentration. Au-incorporated WO3 films exhibited a better sensing performance in comparison with that of pure WO3 sensor films. For example, towards 0.7 ppm of H2S, pure WO3 films showed a sensitivity of 2.9 with a response and recovery time of 106 s and 32 min, while Au-incorporated sample (2.32 at.%) exhibited an increase in the sensitivity to 12 with a response and recovery times of 88 s and 18 min, respectively. Moreover, Au-incorporated thin films were able to detect as low as 50 ppb of H2S with a sensitivity value of 2.7. For practical applications, the reproducibility and the stability of sensing-device are also important. In the present case, the sensors developed under identical conditions (same and different batches) exhibited similar response towards H2S. Moreover, the stability measurements performed over a period of 15 days indicate them to be quite stable. For example, the sensor sample with 2.32 at.% Au exhibited a variation of about 8% in its initial resistance value for this period. A detailed investigation of the long term stability of the sensor films is underway.
3.3. Raman and work function measurements Raman investigations of the sensor film before and after exposure to H2S have also been performed as it is known to give the “fingerprint” of the material. Fig. 7(a) shows the Raman spectra (recorded at room temperature) of a bare WO3 films before and after exposure to H2S. Both the spectra invariably exhibit a series of sharp peaks at wavenumbers lower than 250 cm1 attributed to vibrational modes of the WO3 lattice i.e. the translational and rotational motion of WO6 octahedra in the same unit cell. They also exhibit peaks at intermediate frequency region (200e400 cm1) assigned to the bending modes of OeWeO bridging bonds and at high frequency region (600e900 cm1) attributed to the stretching modes of WeO bonds [25,26]. The Raman peaks observed at approximately 60, 73, 136, 274, 328, 718 and 809 cm1 and a shoulder at w96 cm1 are characteristic of crystalline monoclinic g-phase of WO3. The monoclinic g-phase is the most stable room temperature phase of WO3 and has been reported for both WO3 powder and nanoforms [27,28]. The fingerprint Raman peak at approximately 33 cm1 could not be detected for the samples due to experimental limitations [notch cut off]. Additionally, the absence of the 950 cm1 Raman peak in both the samples is indicative of their high crystalline nature [29]. Apart from the Raman peaks of the g-phase, an additional peak at w40 cm1 attributed to the orthorhombic b-phase is observed. The low frequency Raman modes involve tilted motion of the oxygen cages around the W atoms and/translational motion of WO6 octahedra in the same unit cell. Strong polarizability modulation of these lattice vibrational modes during the evolution of a new phase makes Raman spectroscopy a more sensitive technique than XRD for studying phase transition in WO3. Reversible transition from g to b phase has been reported at w300 C for microcrystalline WO3
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857
855
Fig. 5. Response curves of (a) pure and (c) Au-incorporated (2.32 wt.%) WO3 thin films towards sub-ppm concentration of H2S at an operating temperature of 250 C and corresponding variation in sensitivity with gas concentration (c) and (d).
powder [25] and for nanowires above 227 C [26]. However, room temperature observation of the b-phase for the vapor deposited thin film samples is an interesting and notable exception. Upon exposure to H2S, intensity of the Raman peaks due to the g-phase relative to that of the peak at 40 cm1 (considered to be the fingerprint of the b-phase) reduced significantly, but their peak positions remained almost unaltered. The two peaks at w60 and 73 cm1 almost merged to form a single broad band. Since, interaction of WO3 with H2S is likely to occur at the film surface, the observation indicates a further stabilization of the b-phase on the film surface at the expense of the g-phase. Additionally a decrease
in the intensity of the peaks corresponding to the stretching modes was also observed. Raman spectra, before and after exposure to H2S, of Au-incorporated WO3 are depicted in Fig. 7(b). Although the spectrum of the unexposed sample is qualitatively similar to that of bare WO3, intensity of the peak at w40 cm1 is much higher than that at 60 cm1. Moreover, the broad hump at w96 cm1 in the spectrum of bare WO3 appears as a distinct peak for AueWO3. These clearly indicate the relative enhancement of the orthorhombic b-phase. Upon exposure to H2S, intensity of the lattice and stretching modes reduced marginally, whereas no significant change in the intensity
Fig. 6. (a) Effect of Au loading on sensitivity towards 1 ppm H2S, and on the air resistance at an operating temperature of 250 C. (b) Selectivity histogram pure WO3 and Auincorporated WO3 thin film sensors towards different gases at an operating temperature of 250 C.
856
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857
Fig. 7. Raman spectra recorded for (a) pure WO3 and (b) Au-incorporated WO3 sensor films before and after exposure to H2S gas at room temperature.
of the bending modes of WO3 was observed. Interestingly, in contrast to the bare WO3, intensity of the peak at w40 cm1 relative to that at 60 cm1 got reduced after exposure to H2S. Augmentation of Raman signals due to the surface enhanced resonance Raman effects of the embedded Au may contribute towards the anomalous intensity variation for AueWO3 sensor films. Fig. 8 shows the work function area scan recorded for 3.3 mm2 of sensor samples before and after exposure to 100 ppm of H2S. Upon exposure to H2S, the CPD over the measured area is observed to vary. The color variation indicates the localized variation in the CPD. Table 1 illustrates the final work function values of the sensor samples calculated taking the average of the relative variation in the CPD. As is clearly evident the incorporation of Au in the WO3 resulted in work function decrease by 0.25 eV. Besides exposure to H2S causes changes in the work function of pure and Au-incorporated WO3 samples by 30 and 11 meV, respectively. The small
variation in the work function of Au-incorporated sample is attributed to the faster reaction kinetics (recovery time). This result is in accordance with and corroborates further the findings of Raman investigations. 3.4. Gas sensing mechanism It is established that interaction of H2S with WO3 leads to the generation of large number of oxygen vacancies [10]. The interaction leads to the introduction of new energy levels in the mid gap states of WO3. Besides, incorporation of doping elements in particular Au has been known to impart the sensitivity in the host material mainly via electronic sensitization mechanism [26]. In this the additives enhance the rate of reaction by providing the additional active sites. The interaction with the test gas mainly occurs on the Au surface and the corresponding changes
Fig. 8. 2-D work function map of pure WO3 sensor films (a) before and (b) after exposure to H2S and Au-incorporated WO3 (c) before and (d) after exposure to H2S.
N. Datta et al. / Materials Chemistry and Physics 134 (2012) 851e857 Table 1 Work function values recorded for sensor samples before and after exposure to H2S. Sample name
F, before exposure (eV)
F, after exposure (eV)
Pure WO3 Au-incorporated WO3
5.221 4.973
5.191 4.962
are transferred immediately to the host matrix reflected as a fast drop in the resistance of the sample. This is expected to contribute for improving both the response and the recovery times of the sensor films. Further, the oxygen vacancies can easily be reoxidized upon exposure to fresh air thereby assuring complete recovery. 4. Conclusions In summary, we have deposited pure and Au-incorporated WO3 sensor thin films using a simple technique of vacuum deposition. Effect of operating temperature, gas concentration, and Au loading has been explored and correlated with the observed sensitivity values to determine the optimum condition for improved sensing performance towards H2S. The sensor film containing 2.32 at.% of Au detected H2S selectively and sensitively (S ¼ w16) at an operating temperature of 250 C. The enhanced reaction kinetics is mainly attributed to Au and has been confirmed using both the Raman and the work function measurements. Our results clearly indicate that the WO3 thin films can be made selective and sensitive towards sub-ppm H2S by a proper control over the process parameters namely operating temperature, doping element (Au) and its concentration in the host matrix. Acknowledgments This work is partly supported by “DAE-SRC Outstanding Research Investigator Award” (2008/21/05-BRNS) and “Prospective Research Funds” (2008/38/02-BRNS) granted to D.K.A. Authors also thank Dr. A.K. Chauhan and Mr. Arvind Kumar for their help with Xray diffraction and work function measurements, respectively.
857
References [1] V. Srivastava, K. Jain, Sens. Actuators B 133 (2008) 46. [2] R. Ionescu, A. Hoel, C.G. Granqvist, E. Llobet, P. Heszler, Sens. Actuators B 104 (2005) 132. [3] M. Bendahan, J. Guérin, R. Boulmani, K. Aguir, Sens. Actuators B 124 (2007) 24. [4] M. Akiyama, J. Tamaki, N. Miura, N. Yamazoe, Chem. Lett. (1991) 1611. [5] A. Labidi, E. Gillet, R. Delamare, M. Maaref, K. Aguir, Sens. Actuators B 120 (2006) 338. [6] C.N. Xu, N. Miura, Y. Ishida, K. Matsuda, N. Yamazoe, Sens. Actuators B 65 (2000) 163. [7] J.-H. Yoon, H.-J. Lee, Sens. Lett. 9 (2011) 46. [8] L. Jia, W. Cai, H. Wang, Appl. Phys. Lett. 96 (2010) 103115. [9] T. Maekawa, J. Tamaki, N. Miura, N. Yamazoe, Chem. Lett. (1992) 639. [10] W.H. Tao, C.-H. Tsai, Sens. Actuators B 81 (2002) 237. [11] B.T. Marquis, J.F. Vetelino, Sens. Actuators B 77 (2001) 100. [12] P. Montmeat, J.-C. Marchand, R. Lalauze, J.-P. Viricelle, G. Tournier, C. Pijolat, Sens. Actuators B 95 (2003) 83. [13] Z. Xu, J.F. Veletino, R. Lec, D.C. Parker, Electrical properties of tungsten trioxide films, J. Vac. Sci. Technol. A 8 (1990) 3634. [14] H. Xia, Y. Wang, F. Kong, S. Wang, B. Zhu, X. Guo, J. Zhang, Y. Wang, S. Wu, Sens. Actuators B 134 (2008) 133. [15] S.J. Ippolito, S. Kandasamy, K. Kalantar-zadeh, W. Wlodarski, Sens. Actuators B 108 (2005) 154. [16] X. Wang, N. Miura, N. Yamazoe, Sens. Actuators B 66 (2000) 74. [17] C.V. Ramana, S. Utsunomiya, R.C. Ewing, C.M. Julien, U. Becker, J. Phys. Chem. B 110 (2006) 10430. [18] M. Penza, G. Cassano, F. Tortorella, Sens. Actuators B 81 (2001) 115. [19] R.B. Slimane, F.S. Lau, R.J. Dihu, M. Khinkis, Production of hydrogen by superadiabatic decomposition of hydrogen sulfide, in: Proceedings of the 2002 U.S. DOE Hydrogen Program Review, NREL/CP-610-32405. [20] N.S. Ramgir, M. Ghosh, P. Veerender, N. Datta, M. Kaur, D.K. Aswal, S.K. Gupta, Sens. Actuators B 156 (2011) 875. [21] N.S. Ramgir, S. Kailasa Ganapati, M. Kaur, N. Datta, K.P. Muthe, D.K. Aswal, S.K. Gupta, J.V. Yakhmi, Sens. Actuators B 151 (2010) 90. [22] Powder Diffract, File, JCPDS Int. Centre Diffract. Data, PA 19073-3273, U.S.A., 2001. [23] R.S. Niranjan, V.A. Chaudhary, S.R. Sainker, K.R. Patil, K. Vijayamohanan, I.S. Mulla, Sens. Actuators B 79 (2001) 132. [24] N.S. Ramgir, S. Kailasaganapati, M. Kaur, S. Mishra, N. Datta, D.K. Aswal, S.K. Gupta, J.V. Yakhmi, AIP Conf. Proc. 1349 (2011) 679e680. [25] T. Siciliano, A. Tepore, G. Micocci, A. Serra, D. Manno, E. Filippo, Sens. Actuators B 133 (2008) 321. [26] J. Gabrusenoks, A. Veispals, A. von Czarnowski, K.-H. Meiwes-Broer, Electrochim. Acta 46 (2001) 2229. [27] E. Cazzanelli, C. Vinegoni, G. Mariotto, A. Kuzmin, J. Purans, Solid State Ionics 123 (1999) 67e74. [28] D.Y. Lu, J. Chen, H.J. Chen, L. Gong, S.Z. Deng, Y.L. Liu, N.S. Xu, Appl. Phys. Lett. 90 (2007) 041919. [29] K. Nonaka, A. Takase, K. Miyakawa, J. Mater. Sci. Lett. 12 (1993) 274.