- Email: [email protected]
Chlorine gas sensors using one-dimensional tellurium nanostructures
Talanta 77 (2009) 1567–1572
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Chlorine gas sensors using one-dimensional tellurium nanostructures Shashwati Sen a , Madhvi Sharma a , Vivek Kumar c , K.P. Muthe a , P.V. Satyam b , Umananda M. Bhatta b , M. Roy d , N.K. Gaur c , S.K. Gupta a,∗ , J.V. Yakhmi a a
Technical Physics & Prototype Engineering Division, Bhabha Atomic Research Centre, Mumbai 400085, India Institute of Physics, Sachivalaya Marg, Bhubaneswar, Orissa 751005, India c Department of Physics, Barkatullah University, Bhopal, MP 462026, India d Chemistry Division, Bhabha Atomic Research Centre, Mumbai 400085, India b
a r t i c l e
i n f o
Article history: Received 7 August 2008 Received in revised form 23 September 2008 Accepted 24 September 2008 Available online 17 October 2008 Keywords: Nanotubes Gas sensor Band gap Impedance spectroscopy Grain boundary
a b s t r a c t Tellurium nanotubes have been grown by physical vapor deposition under inert environment at atmospheric pressure as well as under vacuum conditions. Different techniques such as scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD) and optical absorption have been utilized for characterization of grown structures. Films prepared using both types of tellurium nanotubes were characterized for sensitivity to oxidizing and reducing gases and it was found that the relative response to gases depends on the microstructure. Nanotubes prepared at atmospheric pressure (of argon) showed high sensitivity and better selectivity to chlorine gas. Impedance spectroscopy studies showed that the response to chlorine is mainly contributed by grain boundaries and is therefore enhanced for nanotubes prepared under argon atmosphere. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Investigations on the growth and characteristics of quasi onedimensional structures (Q1D) are important as they provide inherently enhanced integration density of devices, have novel physical effects due to reduced dimensionality and have high surface to volume ratio desirable for some applications [1]. A variety of inorganic materials including elemental and compound semiconductors have been synthesized in various nano-forms [2]. Tellurium is a low band gap semiconductor and is useful for various applications such as optical recording, thin film transistors, strain sensitive devices, infrared detectors, gas sensors and thermoelectric devices [3,4]. Due to anisotropic crystal structure, Te is amenable to the growth of one-dimensional nanostructures as nanowires and nanotubes. In earlier studies, we have investigated pure Te thin films as room temperature gas sensors for H2 S, NH3 and NO [5–7]. However, there are no reports on application of Te nanostructures as gas sensors or Te thin films for detection of chlorine gas. Chlorine is a widely used chemical in many industrial processes and is very harmful when emitted into environment. It may be detected by sophisti-
∗ Corresponding author. Tel.: +91 22 25593863; fax: +91 22 25505296. E-mail address: [email protected] (S.K. Gupta). 0039-9140/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2008.09.055
cated techniques as gas chromatography, chemical detecting tubes and electrochemical sensors [8–10]. Sensors based on semiconductor oxides such as In2 O3 and WO3 have also been employed for the detection of this gas but most of them require elevated temperatures for operation [11,12]. Patil and Patil have reported room temperature detection of Cl2 gas at high concentrations (300 ppm) using CuO modified ZnO [13]. In a recent report, Sb-doped, SnO2 films with nanoporous structure have been employed for room temperature chlorine sensing [14]. As tellurium thin films have been shown to detect many reducing and oxidizing gases at room temperature [5–7], it is desirable to explore Te nanostructures for sensitivity to chlorine. In most of the studies reported in the literature, nanostructures of Te have been grown by solution routes such as reduction of TeO2 or orthotelluric acid by hydrazine and oxidation of NaHTe in the presence of surfactants [15–18]. Nanostructures have also been prepared by hydrothermal and solvothermal processes. Few groups have reported the growth of Te nanostructures by vapor phase techniques. Geng et al. [19] have synthesized Te nanobelts from Al2 Te3 in a furnace under Ar gas flow. Mohanty et al. [20] have synthesized Te nanotubes with triangular cross-section and some with hexagonal cross-section by evaporation of Te in a tubular furnace in the presence of Ar gas at intermediate pressure of 1–1.5 Torr and deposition temperature of 150–200 ◦ C. They reported the growth of nanowires and nanorods in case of Si (1 1 1) and sapphire (0 0 1)
1568
S. Sen et al. / Talanta 77 (2009) 1567–1572
Fig. 1. SEM images of typical (a) type-I sample—Te nanotubes grown in furnace, (b) type-II sample—Te micro-rods and (c) type-III sample—Te nanotubes on silicon. (d) TEM and HRTEM images of type-III sample. Inset of (a) shows TEM of a type-I nanotube.
substrate, respectively [21]. Metraux and Grobéty [22] have grown Te nanostructures on Si (1 1 1) substrate and aluminium foil by heating Te in an induction furnace (in the presence of magnetic field). They observed formation of nanotubes on both substrates under argon ambient at 1 mbar pressure. However, at 10−7 mbar, the formation of nanowires on aluminium foil and platelets on Si (1 1 1) was observed. Interestingly, when the depositions were carried out in a high vacuum coating unit (thus without the assistance of magnetic field) the formation of dendritic Te filaments was observed which led to the inference that the presence of magnetic field is primarily responsible for the observed growth behavior. Recently Wang et al. [23] reported the growth of ultrawide Te nanobelts by vapor deposition in a horizontal quartz tube at evaporation temperature of 350 ◦ C under vacuum conditions. Similarly, Chen et al. [24] have fabricated Te nanowires of different orientations, on NaNO2 floccules by vacuum deposition at 10−5 mbar. Thus we find that diverse Te nanostructures have been obtained by tuning the growth conditions. As the microstructure of materials affects their functionality [25], we expect suitable Te nanostructures to show selective response for detection of desired gases. Here, we report the growth of Te nanotubes at atmospheric pressure in a horizontal quartz tube and under high vacuum conditions. The gas sensing characteristics of these nanostructures were investigated and Te nanotubes prepared under atmospheric pressure were found to detect chlorine gas at room temperature at very low concentration of 0.5 ppm. Impedance spectroscopy studies showed that the response to chlorine is mainly due to resistance changes at grain boundaries. 2. Experimental Growth of Te nanostructures was carried out at atmospheric pressure as well as under vacuum conditions. At atmospheric pres-
sure, the growth was carried out under argon atmosphere in a tubular furnace as described earlier [26]. Briefly, Te powder was loaded in an alumina boat and placed in a 1-m long quartz tube. The furnace temperature was raised to 550 ◦ C in presence of Ar gas flow (150 cm3 /min) and maintained at this temperature for 2 h. Te nanotubes were found to deposit on quartz tube in the direction of gas flow at temperature of 30–50 ◦ C. Micro-rods (whiskers) of Te were found to grow in the high temperature (350–400 ◦ C) zone of the furnace. Similar deposits were also obtained on different substrates placed in the tube. In the case of synthesis carried out by thermal evaporation under vacuum, high purity Te powder was loaded in a molybdenum boat. The depositions on Si (1 1 1) substrates (maintained at 100 ◦ C) were carried out under vacuum of 2 × 10−5 mbar at a rate of 10 Å/s maintained using a quartz crystal thickness monitor. Microstructure of the grown Te was studied by scanning electron microscopy (SEM) using TESCAN make VEGA MV2300T/40 system. The chemical composition was obtained by energy dispersive Xray (EDX) analysis, while structural information was obtained by X-ray diffraction (XRD) spectra obtained by employing Cu K␣ radiation. TEM analysis was carried out using JEOL 2010 UHR system and the images were recorded using a 20× Gatan camera. For TEM, powder of Te nanostructures grown in furnace and scraped powder from silicon substrates in case of Te grown in high vacuum were utilized. Powders were dispersed in propanol by ultrasonication and a drop of this solution was placed on carbon-coated Cu grid. The impedance spectroscopy measurements were carried out using AUTOLAB make potentiostat–galvanostat (model PGSTAT 302). For study of gas sensitivity, Te grown in horizontal furnace was dispersed in ethanol to make a suspension, which was painted on a glass plate (called type-I sample) with two pre-deposited gold electrodes and thick film gold contacts were made on individual whiskers of tellurium (called type-II samples). In case of Te deposited on Si (1 1 1) substrates under vacuum (called type-III
S. Sen et al. / Talanta 77 (2009) 1567–1572
Fig. 2. XRD spectra of: (a) type-I, (b) type-II and (c) type-III samples (peak marked # corresponds to SiO2 ).
samples), Au electrodes were thermally evaporated on nanostructured Te film. Electrical contacts were made by soldering silver wires to the gold pads, using indium solder. The sensitivity towards different gases was determined by measuring the resistance in atmosphere and on exposure to different gases using Keithley multimeter. The measurement setup has been described elsewhere [5]. All the sensitivity measurements were carried out at room temperature. Sensitivity to some of the gases was also studied using Te thin films deposited on alumina substrates (called type-IV samples). On exposure to oxidizing gases such as NO and Cl2 the resistance of all the samples was found to decrease while on exposure to reducing gases such as H2 S and NH3 , the resistance increased, as expected for p-type nature of Te. The response (S) to different gases is defined by
(Ra − Rg ) Ra
S =
where Ra and Rg are the resistances of the sensors in air and gas, respectively. It may be noted a factor of 4 change in resistance corresponds to S = 3 for a reducing gas (where Rg > Ra ) and S = 0.75 for an oxidizing gas (where Ra > Rg ). The response time is defined as the time taken to attain 90% of the change in resistance and the recovery time is the time taken by the sensors to return back to 10% of its original resistance. 3. Results and discussion SEM and TEM micrographs of various samples are shown in Fig. 1 and corresponding X-ray diffraction spectra are shown in Fig. 2. SEM micrograph of low temperature deposit at atmospheric pressure (type-I samples) showed presence of nanotubes (Fig. 1a). Inset of this figure shows the TEM image of a single nanotube. Detailed characterization and growth mechanism of these nanotubes has been
1569
reported earlier [26]. The tubes had outer diameter of 100–500 nm, wall thickness of ∼10 nm and maximum length up to 3 m. All the tubes were found to be hexagonal in shape and their size was independent of the nature of the substrate. Deposits in quartz tube at temperatures of 350–400 ◦ C were found to consist of Te micro-rods as shown in Fig. 1b (type–II samples). The whiskers had a hexagonal cross-section with a diameter of 2–4 m and length ranging from 1 to 10 mm. Depositions under vacuum were carried out at a substrate temperature of 100 ◦ C, as films at this temperature were found to be crystalline in earlier studies [6]. Si (1 1 1) substrates were used due to their good lattice matching with Te [22]. SEM of these samples (type-III) showed growth of nanotubes with diameter of 70–150 nm (Fig. 1c). The size distribution of the nanotubes was found to be quite narrow in this case as compared to type-I samples. TEM image (Fig. 1d) showed that the nanotubes had prong-like structure with tubes showing split at the end as reported in earlier study [26]. The HRTEM image at the tip of these prongs showed the lattice spacing of ∼0.2 nm corresponding to the (0 0 3) direction of Te, confirming the growth direction to be along c-axis. Our result of nanotubes deposition under vacuum conditions is in contrast with that of Metraux and Grobéty [22] who have not observed the formation of Te nanostructures when depositions were performed in a high vacuum coating unit. X-ray diffraction spectra of different structures are shown in Fig. 2. Powder diffraction spectra were measured for type-I and type-II samples and grazing angle XRD was obtained for typeIII samples due to their thin film nature. All of the samples are found to be single crystalline Te. All the peaks could be indexed for the hexagonal structure of tellurium with, space group P31 21 and lattice parameters of a = b = 4.49 and c = 5.9 Å. This shows that samples are high quality Te crystals without any impurity phase. It is seen that relative intensity of (1 0 0) peak in whiskers (spectrum b) is quite large compared to sample of type-I (spectrum a). This is because these are c-axis oriented whiskers of large size and lie parallel to the substrate during XRD. In case of spectra (c) we also observe a peak corresponding to SiO2 as these Te nanotubes (typeIII sample) have been grown on Si substrate with thin oxide layer. The variation in relative intensities of peaks in three spectra arises due to different orientations of crystallites. Because of the anisotropic nature of Te lattice, it shows tendency to from 1D structures with Te adatoms having greater tendency of sticking along c-axis. In case of growth in furnace at atmospheric pressure, formation of nanowires and nanotubes by vapor–solid (VS) mechanisms has been described earlier [26]. In this case we observe nanotubes at temperature of 30–50 ◦ C irrespective of the nature of the substrate. However this temperature is not sufficient for epitaxial growth of Te on a given substrate. In the case of vacuum deposition at 100 ◦ C, the growth of crystalline Te is governed by surface kinetics as Te adatoms gain sufficient energy to occupy favorable lattice sites on the substrate. For Si (1 1 1) substrate, 3asurface Si(1 1 1) = 2abulk Te (where asurface and abulk are lattice parameters on surface of silicon and bulk crystalline Te) with a mismatch of only ∼0.9%. This leads to preferential nucleation of Te (0 0 1) parallel to Si (1 1 1) [22]. The Te adatoms continue to find and attach themselves to the energetically favorable sites leading to the formation of oriented tubes. Optical absorption spectra and photoluminescence studies provide information about defects in semiconductors. UV–vis absorption spectra were recorded for different samples and the results are shown in Fig. 3. The value of optical band gap was estimated using Mott and Davis model for direct allowed transitions in semiconductors using equation: ˛()h = B(h − Egap )
m/2
1570
S. Sen et al. / Talanta 77 (2009) 1567–1572
Fig. 3. Optical absorption spectra of: (a) type-I (b) type-II and (c) type-III samples.
where ˛ is the absorption coefficient, Egap is the optical band gap, h is the energy of incident photon, B is a constant and m = 1 for direct transitions. To determine the band gap, we have plotted (˛h)2 as function of (h). The absorption in the 3–6 eV region for Te has been attributed to transitions from bonding to antibonding states [27,28]. Band gap for type-I nanotubes is seen to be less than that of whiskers (type-II) and type-III nanotubes indicating larger number of defects or states in the gap for type-I samples. The defects in the structure are important as they may contribute to interaction with gases but have disadvantage of reducing mobility of carriers. Thus it is expected that type-I samples may act as better sensor material due to increased defects. Response of different samples was studied for many gases and Table 1 gives typical response to H2 S, NH3 , NO and Cl2 gases. It is found that (a) response of whiskers to all gases is quite small compared to other samples, (b) both whiskers and thin films have better response to H2 S in comparison to Cl2 and NO and (c) type-I samples have maximum sensitivity for detection of Cl2 . Small response of whiskers is understandable as these have much smaller surface area compared to other samples. To understand the better sensitivity of nanotubes to chlorine, we have performed impedance spectroscopy measurements on type-I and type-III samples on exposure to 4 ppm of chlorine and H2 S gases and the results are shown in Fig. 4. The impedance spectra have been analyzed using equivalent circuit as shown in Fig. 4(c). It consists of two RC networks consisting of RBulk and CBulk for intragrain region and RGrain and CGrain for grain boundary region. R0 is additional frequency independent resistance. Values of various parameters obtained from impedance spectra (before and after exposure to gases) are given in Table 2 for two types of samples. Type-I samples show only grain boundary contribution. This is understandable as these have been prepared by dispersion of nanotubes in solution and coating on substrates that yield very poor grain to grain connectivity. Type-III samples Table 1 Typical response of different Te samples to H2 S, NH3 , NO and Cl2 gases at concentration of 8 ppm. Gas
H2 S NH3 Cl2 NO
Fig. 4. Impedance spectra before and after exposure to 4 ppm of Cl2 and H2 S gases at room temperature for typical (a) type-I and (b) type-III samples. The scattered points show experimental data and the solid lines represent fitting obtained with the values given in Table 2. (c) Shows the equivalent circuit used for analysis.
Response to gases for samples of type I (nanotubes)
II (whiskers)
III (nanotubes)
IV (thin films)
0.33 0.0204 0.75 0.47
0.087 0.009 0.043 0.009
0.18 0.026 0.43 0.27
1.85 0.1 0.57 0.30
show both intragrain and intergrain (grain boundary) contributions and it is found that grain boundary region has much higher contribution in response to Cl2 in comparison to intragrain region. On the other hand, contributions of intragrain and grain boundary regions towards response to H2 S are similar. A similar and in fact,
S. Sen et al. / Talanta 77 (2009) 1567–1572
1571
Table 2 Parameters obtained by fitting of impedance spectroscopy data to equivalent circuit for type-I and type-III samples. Parameters have been measured before and during exposure to H2 S and Cl2 gases at 4 ppm concentration. Parameter
R0 () RBulk () CBulk (F) RGrain () CGrain (F)
Type-I sample (atm. pressure)
Type-III sample (vacuum deposited)
Unexposed
Cl2 exposed
H2 S exposed
Unexposed
Cl2 exposed
H2 S exposed
7770 – – 8200 1.29 × 10−7
5300
8600
3150 1.01 × 10−6
11,400 1.16 × 10−7
10 400 2.21 × 10−8 370 8.6 × 10−8
10 390 2.147 × 10−8 176 1.45 × 10−7
10 440 2 × 10−8 500 6.63 × 10−8
often higher contribution of intragrain region to H2 S is in agreement with earlier studies on thin films [5]. A higher contribution of grain boundary region towards response to chlorine explains better sensitivity of nanotube samples to this gas in comparison to thin film and whisker samples. This also explains best response of type-I samples to chlorine. As type-I samples have good response to chlorine, these were further studied. Response of a typical sensor to chlorine at 2 ppm concentration is shown in Fig. 5(a). A response time of 30 s and recovery time of 2 h is observed for a typical sample. Response of another film to chlorine at different concentrations is shown in Fig. 5(b). The recovery time is found to increase with concentration of Cl2 however, the resistance of the film is found to recover to its base resistance even when exposed to 8 ppm of gas indicat-
Fig. 6. Response of a typical type-I sample to H2 S, NO, Cl2, acetone, NH3 , CO and CH4 gases at 4 ppm and H2 at 2% concentration.
ing a reversible change on exposure to chlorine gas. Investigations to reduce recovery time are being carried out. Response of a typical type-I sample to many oxidizing and reducing gases is shown in Fig. 6. It is observed that these samples are quite selective for response to chlorine. The mechanism of response of Te to oxidizing or reducing gas has been reported in earlier studies [4]. Te is a p-type semiconductor and adsorption of oxygen (in normal air) leads to trapping of electrons and therefore introduction of holes in the lattice. This reduces the resistance. The resistance in further reduced on adsorption to oxidizing gases that trap electrons (such as Cl2 ), and increased on exposure to reducing gases that remove the adsorbed oxygen. 4. Conclusion The growth of Te nanotubes has been carried out via a onestep physical vapor deposition process at atmospheric pressure as well as under high vacuum conditions. Te nanotubes grown under atmospheric conditions in a horizontal furnace showed high sensitivity towards Cl2 gas at room temperature. It is found that chlorine has larger influence on grain boundary resistance compared to intragrain resistance leading to better response of nanostructure materials. References
Fig. 5. (a) Response of a type-I sensor to 2 ppm chlorine. Inset gives expanded view to show saturation at 2 ppm. (b) Response and recovery characteristics for different concentrations of Cl2 gas for another type-I sample.
[1] J.G. Lu, P. Chang, Z. Fan, Mater. Sci. Eng. R 52 (2006) 49. [2] C.N.R. Rao, F.L. Deepak, G. Gundiah, A. Govindraj, Prog. Solid State Chem. 31 (2003) 5. [3] T. Ikari, H. Berger, F. Levy, Mater. Res. Bull. 21 (1986) 99. [4] S. Sen, K.P. Muthe, N. Joshi, S.C. Gadkari, S.K. Gupta, Jagannath, M. Roy, S.K. Deshpande, J.V. Yakhmi, Sens. Actuator B 98 (2004) 154. [5] S. Sen, V. Bhandarkar, K.P. Muthe, M. Roy, S.K. Deshpande, R.C. Aiyer, S.K. Gupta, J.V. Yakhmi, V.C. Sahni, Sens. Actuator B 115 (2006) 270.
1572
S. Sen et al. / Talanta 77 (2009) 1567–1572
[6] V. Bhandarkar, S. Sen, K.P. Muthe, M. Kaur, M. Senthil Kumar, S.K. Deshpande, S.K. Gupta, J.V. Yakhmi, V.C. Sahni, Mater. Sci. Eng. B 131 (2006) 156. [7] V. Bhandarkar, S. Sen, K.P. Muthe, S.K. Deshpande, S. Bhattacharya, S.K. Gupta, Asian J. Phys. 14 (2005) 97. [8] A. Galdikas, Z. Martunas, A. Setkus, Sens. Actuator B 7 (1992) 633. [9] D.H. Dawson, D.E. Williams, J. Mater. Chem. 6 (3) (1996) 409. [10] T. Miyata, T. Hikosaka, T. Minami, Sens. Actuator B 69 (2000) 16. [11] J. Tamaki, Sens. Actuator B 83 (2002) 190. [12] T.V. Belysheva, L.P. Bogovtseva, J. Anal. Chem. 61 (2006) 494. [13] D.R. Patil, L.A. Patil, Sens. Actuator B 123 (2007) 546. [14] A. Chaparadza, S.B. Rananavare, Nanotechnology 19 (2008) 245501. [15] B. Mayer, Y.N. Xia, J. Mater. Chem. 12 (2002 1875). [16] R. Zheng, W. Cheng, E. Wang, S. Dong, Chem. Phys. Lett. 395 (2004) 302. [17] Z. Liu, Z. Hu, J. Liang, S. Li, Y. Yang, S. Peng, Y. Qian, Langmuir 20 (2004) 214.
[18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]
Z. Liu, Z. Hu, Q. Xie, B. Yang, J. Wu, Y. Qian, J. Mater. Chem. 13 (2003) 159. B. Geng, Y. Lin, X. Peng, G. Meng, L. Zhang, Nanotechnology 14 (2003) 983. P. Mohanty, T. Kang, B. Kim, J. Phys. Chem. 110 (2006) 791. P. Mohanty, J. Park, B. Kim, J. Nanosci. Nanotechnol. 6 (2006) 3380. C. Métraux, B. Grobéty, J. Mater. Res. 19 (2004) 2159. Q. Wang, G.-D. Li, Y.-L. Liu, S. Xu, K.-J. Wang, J.-S. Chen, J. Phys. Chem. C 111 (2007) 12926. H. Chen, H. Lu, Y. Nie, J. Zhang, M. Zhang, Q. Dai, S. Gao, S. Kan, D. Li, G. Zou, Phys. Lett. A 362 (2007) 61. B. Lui, H. Zeng, J. Am. Chem. Soc. 126 (2004) 8124. S. Sen, U.M. Bhatta, V. Kumar, K.P. Muthe, S. Bhattacharya, S.K. Gupta, J.V. Yakhmi, Cryst. Growth Des. 8 (1) (2008) 238. R. Swan, A.K. Ray, C.A. Hogarth, Phys. Stat. Sol. A 127 (1991) 555. U.K. Gautam, C.N. Rao, J. Mater. Chem. 14 (2004) 2530.
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Chlorine gas sensors using one-dimensional tellurium nanostructures Shashwati Sen a , Madhvi Sharma a , Vivek Kumar c , K.P. Muthe a , P.V. Satyam b , Umananda M. Bhatta b , M. Roy d , N.K. Gaur c , S.K. Gupta a,∗ , J.V. Yakhmi a a
Technical Physics & Prototype Engineering Division, Bhabha Atomic Research Centre, Mumbai 400085, India Institute of Physics, Sachivalaya Marg, Bhubaneswar, Orissa 751005, India c Department of Physics, Barkatullah University, Bhopal, MP 462026, India d Chemistry Division, Bhabha Atomic Research Centre, Mumbai 400085, India b
a r t i c l e
i n f o
Article history: Received 7 August 2008 Received in revised form 23 September 2008 Accepted 24 September 2008 Available online 17 October 2008 Keywords: Nanotubes Gas sensor Band gap Impedance spectroscopy Grain boundary
a b s t r a c t Tellurium nanotubes have been grown by physical vapor deposition under inert environment at atmospheric pressure as well as under vacuum conditions. Different techniques such as scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD) and optical absorption have been utilized for characterization of grown structures. Films prepared using both types of tellurium nanotubes were characterized for sensitivity to oxidizing and reducing gases and it was found that the relative response to gases depends on the microstructure. Nanotubes prepared at atmospheric pressure (of argon) showed high sensitivity and better selectivity to chlorine gas. Impedance spectroscopy studies showed that the response to chlorine is mainly contributed by grain boundaries and is therefore enhanced for nanotubes prepared under argon atmosphere. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Investigations on the growth and characteristics of quasi onedimensional structures (Q1D) are important as they provide inherently enhanced integration density of devices, have novel physical effects due to reduced dimensionality and have high surface to volume ratio desirable for some applications [1]. A variety of inorganic materials including elemental and compound semiconductors have been synthesized in various nano-forms [2]. Tellurium is a low band gap semiconductor and is useful for various applications such as optical recording, thin film transistors, strain sensitive devices, infrared detectors, gas sensors and thermoelectric devices [3,4]. Due to anisotropic crystal structure, Te is amenable to the growth of one-dimensional nanostructures as nanowires and nanotubes. In earlier studies, we have investigated pure Te thin films as room temperature gas sensors for H2 S, NH3 and NO [5–7]. However, there are no reports on application of Te nanostructures as gas sensors or Te thin films for detection of chlorine gas. Chlorine is a widely used chemical in many industrial processes and is very harmful when emitted into environment. It may be detected by sophisti-
∗ Corresponding author. Tel.: +91 22 25593863; fax: +91 22 25505296. E-mail address: [email protected] (S.K. Gupta). 0039-9140/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2008.09.055
cated techniques as gas chromatography, chemical detecting tubes and electrochemical sensors [8–10]. Sensors based on semiconductor oxides such as In2 O3 and WO3 have also been employed for the detection of this gas but most of them require elevated temperatures for operation [11,12]. Patil and Patil have reported room temperature detection of Cl2 gas at high concentrations (300 ppm) using CuO modified ZnO [13]. In a recent report, Sb-doped, SnO2 films with nanoporous structure have been employed for room temperature chlorine sensing [14]. As tellurium thin films have been shown to detect many reducing and oxidizing gases at room temperature [5–7], it is desirable to explore Te nanostructures for sensitivity to chlorine. In most of the studies reported in the literature, nanostructures of Te have been grown by solution routes such as reduction of TeO2 or orthotelluric acid by hydrazine and oxidation of NaHTe in the presence of surfactants [15–18]. Nanostructures have also been prepared by hydrothermal and solvothermal processes. Few groups have reported the growth of Te nanostructures by vapor phase techniques. Geng et al. [19] have synthesized Te nanobelts from Al2 Te3 in a furnace under Ar gas flow. Mohanty et al. [20] have synthesized Te nanotubes with triangular cross-section and some with hexagonal cross-section by evaporation of Te in a tubular furnace in the presence of Ar gas at intermediate pressure of 1–1.5 Torr and deposition temperature of 150–200 ◦ C. They reported the growth of nanowires and nanorods in case of Si (1 1 1) and sapphire (0 0 1)
1568
S. Sen et al. / Talanta 77 (2009) 1567–1572
Fig. 1. SEM images of typical (a) type-I sample—Te nanotubes grown in furnace, (b) type-II sample—Te micro-rods and (c) type-III sample—Te nanotubes on silicon. (d) TEM and HRTEM images of type-III sample. Inset of (a) shows TEM of a type-I nanotube.
substrate, respectively [21]. Metraux and Grobéty [22] have grown Te nanostructures on Si (1 1 1) substrate and aluminium foil by heating Te in an induction furnace (in the presence of magnetic field). They observed formation of nanotubes on both substrates under argon ambient at 1 mbar pressure. However, at 10−7 mbar, the formation of nanowires on aluminium foil and platelets on Si (1 1 1) was observed. Interestingly, when the depositions were carried out in a high vacuum coating unit (thus without the assistance of magnetic field) the formation of dendritic Te filaments was observed which led to the inference that the presence of magnetic field is primarily responsible for the observed growth behavior. Recently Wang et al. [23] reported the growth of ultrawide Te nanobelts by vapor deposition in a horizontal quartz tube at evaporation temperature of 350 ◦ C under vacuum conditions. Similarly, Chen et al. [24] have fabricated Te nanowires of different orientations, on NaNO2 floccules by vacuum deposition at 10−5 mbar. Thus we find that diverse Te nanostructures have been obtained by tuning the growth conditions. As the microstructure of materials affects their functionality [25], we expect suitable Te nanostructures to show selective response for detection of desired gases. Here, we report the growth of Te nanotubes at atmospheric pressure in a horizontal quartz tube and under high vacuum conditions. The gas sensing characteristics of these nanostructures were investigated and Te nanotubes prepared under atmospheric pressure were found to detect chlorine gas at room temperature at very low concentration of 0.5 ppm. Impedance spectroscopy studies showed that the response to chlorine is mainly due to resistance changes at grain boundaries. 2. Experimental Growth of Te nanostructures was carried out at atmospheric pressure as well as under vacuum conditions. At atmospheric pres-
sure, the growth was carried out under argon atmosphere in a tubular furnace as described earlier [26]. Briefly, Te powder was loaded in an alumina boat and placed in a 1-m long quartz tube. The furnace temperature was raised to 550 ◦ C in presence of Ar gas flow (150 cm3 /min) and maintained at this temperature for 2 h. Te nanotubes were found to deposit on quartz tube in the direction of gas flow at temperature of 30–50 ◦ C. Micro-rods (whiskers) of Te were found to grow in the high temperature (350–400 ◦ C) zone of the furnace. Similar deposits were also obtained on different substrates placed in the tube. In the case of synthesis carried out by thermal evaporation under vacuum, high purity Te powder was loaded in a molybdenum boat. The depositions on Si (1 1 1) substrates (maintained at 100 ◦ C) were carried out under vacuum of 2 × 10−5 mbar at a rate of 10 Å/s maintained using a quartz crystal thickness monitor. Microstructure of the grown Te was studied by scanning electron microscopy (SEM) using TESCAN make VEGA MV2300T/40 system. The chemical composition was obtained by energy dispersive Xray (EDX) analysis, while structural information was obtained by X-ray diffraction (XRD) spectra obtained by employing Cu K␣ radiation. TEM analysis was carried out using JEOL 2010 UHR system and the images were recorded using a 20× Gatan camera. For TEM, powder of Te nanostructures grown in furnace and scraped powder from silicon substrates in case of Te grown in high vacuum were utilized. Powders were dispersed in propanol by ultrasonication and a drop of this solution was placed on carbon-coated Cu grid. The impedance spectroscopy measurements were carried out using AUTOLAB make potentiostat–galvanostat (model PGSTAT 302). For study of gas sensitivity, Te grown in horizontal furnace was dispersed in ethanol to make a suspension, which was painted on a glass plate (called type-I sample) with two pre-deposited gold electrodes and thick film gold contacts were made on individual whiskers of tellurium (called type-II samples). In case of Te deposited on Si (1 1 1) substrates under vacuum (called type-III
S. Sen et al. / Talanta 77 (2009) 1567–1572
Fig. 2. XRD spectra of: (a) type-I, (b) type-II and (c) type-III samples (peak marked # corresponds to SiO2 ).
samples), Au electrodes were thermally evaporated on nanostructured Te film. Electrical contacts were made by soldering silver wires to the gold pads, using indium solder. The sensitivity towards different gases was determined by measuring the resistance in atmosphere and on exposure to different gases using Keithley multimeter. The measurement setup has been described elsewhere [5]. All the sensitivity measurements were carried out at room temperature. Sensitivity to some of the gases was also studied using Te thin films deposited on alumina substrates (called type-IV samples). On exposure to oxidizing gases such as NO and Cl2 the resistance of all the samples was found to decrease while on exposure to reducing gases such as H2 S and NH3 , the resistance increased, as expected for p-type nature of Te. The response (S) to different gases is defined by
(Ra − Rg ) Ra
S =
where Ra and Rg are the resistances of the sensors in air and gas, respectively. It may be noted a factor of 4 change in resistance corresponds to S = 3 for a reducing gas (where Rg > Ra ) and S = 0.75 for an oxidizing gas (where Ra > Rg ). The response time is defined as the time taken to attain 90% of the change in resistance and the recovery time is the time taken by the sensors to return back to 10% of its original resistance. 3. Results and discussion SEM and TEM micrographs of various samples are shown in Fig. 1 and corresponding X-ray diffraction spectra are shown in Fig. 2. SEM micrograph of low temperature deposit at atmospheric pressure (type-I samples) showed presence of nanotubes (Fig. 1a). Inset of this figure shows the TEM image of a single nanotube. Detailed characterization and growth mechanism of these nanotubes has been
1569
reported earlier [26]. The tubes had outer diameter of 100–500 nm, wall thickness of ∼10 nm and maximum length up to 3 m. All the tubes were found to be hexagonal in shape and their size was independent of the nature of the substrate. Deposits in quartz tube at temperatures of 350–400 ◦ C were found to consist of Te micro-rods as shown in Fig. 1b (type–II samples). The whiskers had a hexagonal cross-section with a diameter of 2–4 m and length ranging from 1 to 10 mm. Depositions under vacuum were carried out at a substrate temperature of 100 ◦ C, as films at this temperature were found to be crystalline in earlier studies [6]. Si (1 1 1) substrates were used due to their good lattice matching with Te [22]. SEM of these samples (type-III) showed growth of nanotubes with diameter of 70–150 nm (Fig. 1c). The size distribution of the nanotubes was found to be quite narrow in this case as compared to type-I samples. TEM image (Fig. 1d) showed that the nanotubes had prong-like structure with tubes showing split at the end as reported in earlier study [26]. The HRTEM image at the tip of these prongs showed the lattice spacing of ∼0.2 nm corresponding to the (0 0 3) direction of Te, confirming the growth direction to be along c-axis. Our result of nanotubes deposition under vacuum conditions is in contrast with that of Metraux and Grobéty [22] who have not observed the formation of Te nanostructures when depositions were performed in a high vacuum coating unit. X-ray diffraction spectra of different structures are shown in Fig. 2. Powder diffraction spectra were measured for type-I and type-II samples and grazing angle XRD was obtained for typeIII samples due to their thin film nature. All of the samples are found to be single crystalline Te. All the peaks could be indexed for the hexagonal structure of tellurium with, space group P31 21 and lattice parameters of a = b = 4.49 and c = 5.9 Å. This shows that samples are high quality Te crystals without any impurity phase. It is seen that relative intensity of (1 0 0) peak in whiskers (spectrum b) is quite large compared to sample of type-I (spectrum a). This is because these are c-axis oriented whiskers of large size and lie parallel to the substrate during XRD. In case of spectra (c) we also observe a peak corresponding to SiO2 as these Te nanotubes (typeIII sample) have been grown on Si substrate with thin oxide layer. The variation in relative intensities of peaks in three spectra arises due to different orientations of crystallites. Because of the anisotropic nature of Te lattice, it shows tendency to from 1D structures with Te adatoms having greater tendency of sticking along c-axis. In case of growth in furnace at atmospheric pressure, formation of nanowires and nanotubes by vapor–solid (VS) mechanisms has been described earlier [26]. In this case we observe nanotubes at temperature of 30–50 ◦ C irrespective of the nature of the substrate. However this temperature is not sufficient for epitaxial growth of Te on a given substrate. In the case of vacuum deposition at 100 ◦ C, the growth of crystalline Te is governed by surface kinetics as Te adatoms gain sufficient energy to occupy favorable lattice sites on the substrate. For Si (1 1 1) substrate, 3asurface Si(1 1 1) = 2abulk Te (where asurface and abulk are lattice parameters on surface of silicon and bulk crystalline Te) with a mismatch of only ∼0.9%. This leads to preferential nucleation of Te (0 0 1) parallel to Si (1 1 1) [22]. The Te adatoms continue to find and attach themselves to the energetically favorable sites leading to the formation of oriented tubes. Optical absorption spectra and photoluminescence studies provide information about defects in semiconductors. UV–vis absorption spectra were recorded for different samples and the results are shown in Fig. 3. The value of optical band gap was estimated using Mott and Davis model for direct allowed transitions in semiconductors using equation: ˛()h = B(h − Egap )
m/2
1570
S. Sen et al. / Talanta 77 (2009) 1567–1572
Fig. 3. Optical absorption spectra of: (a) type-I (b) type-II and (c) type-III samples.
where ˛ is the absorption coefficient, Egap is the optical band gap, h is the energy of incident photon, B is a constant and m = 1 for direct transitions. To determine the band gap, we have plotted (˛h)2 as function of (h). The absorption in the 3–6 eV region for Te has been attributed to transitions from bonding to antibonding states [27,28]. Band gap for type-I nanotubes is seen to be less than that of whiskers (type-II) and type-III nanotubes indicating larger number of defects or states in the gap for type-I samples. The defects in the structure are important as they may contribute to interaction with gases but have disadvantage of reducing mobility of carriers. Thus it is expected that type-I samples may act as better sensor material due to increased defects. Response of different samples was studied for many gases and Table 1 gives typical response to H2 S, NH3 , NO and Cl2 gases. It is found that (a) response of whiskers to all gases is quite small compared to other samples, (b) both whiskers and thin films have better response to H2 S in comparison to Cl2 and NO and (c) type-I samples have maximum sensitivity for detection of Cl2 . Small response of whiskers is understandable as these have much smaller surface area compared to other samples. To understand the better sensitivity of nanotubes to chlorine, we have performed impedance spectroscopy measurements on type-I and type-III samples on exposure to 4 ppm of chlorine and H2 S gases and the results are shown in Fig. 4. The impedance spectra have been analyzed using equivalent circuit as shown in Fig. 4(c). It consists of two RC networks consisting of RBulk and CBulk for intragrain region and RGrain and CGrain for grain boundary region. R0 is additional frequency independent resistance. Values of various parameters obtained from impedance spectra (before and after exposure to gases) are given in Table 2 for two types of samples. Type-I samples show only grain boundary contribution. This is understandable as these have been prepared by dispersion of nanotubes in solution and coating on substrates that yield very poor grain to grain connectivity. Type-III samples Table 1 Typical response of different Te samples to H2 S, NH3 , NO and Cl2 gases at concentration of 8 ppm. Gas
H2 S NH3 Cl2 NO
Fig. 4. Impedance spectra before and after exposure to 4 ppm of Cl2 and H2 S gases at room temperature for typical (a) type-I and (b) type-III samples. The scattered points show experimental data and the solid lines represent fitting obtained with the values given in Table 2. (c) Shows the equivalent circuit used for analysis.
Response to gases for samples of type I (nanotubes)
II (whiskers)
III (nanotubes)
IV (thin films)
0.33 0.0204 0.75 0.47
0.087 0.009 0.043 0.009
0.18 0.026 0.43 0.27
1.85 0.1 0.57 0.30
show both intragrain and intergrain (grain boundary) contributions and it is found that grain boundary region has much higher contribution in response to Cl2 in comparison to intragrain region. On the other hand, contributions of intragrain and grain boundary regions towards response to H2 S are similar. A similar and in fact,
S. Sen et al. / Talanta 77 (2009) 1567–1572
1571
Table 2 Parameters obtained by fitting of impedance spectroscopy data to equivalent circuit for type-I and type-III samples. Parameters have been measured before and during exposure to H2 S and Cl2 gases at 4 ppm concentration. Parameter
R0 () RBulk () CBulk (F) RGrain () CGrain (F)
Type-I sample (atm. pressure)
Type-III sample (vacuum deposited)
Unexposed
Cl2 exposed
H2 S exposed
Unexposed
Cl2 exposed
H2 S exposed
7770 – – 8200 1.29 × 10−7
5300
8600
3150 1.01 × 10−6
11,400 1.16 × 10−7
10 400 2.21 × 10−8 370 8.6 × 10−8
10 390 2.147 × 10−8 176 1.45 × 10−7
10 440 2 × 10−8 500 6.63 × 10−8
often higher contribution of intragrain region to H2 S is in agreement with earlier studies on thin films [5]. A higher contribution of grain boundary region towards response to chlorine explains better sensitivity of nanotube samples to this gas in comparison to thin film and whisker samples. This also explains best response of type-I samples to chlorine. As type-I samples have good response to chlorine, these were further studied. Response of a typical sensor to chlorine at 2 ppm concentration is shown in Fig. 5(a). A response time of 30 s and recovery time of 2 h is observed for a typical sample. Response of another film to chlorine at different concentrations is shown in Fig. 5(b). The recovery time is found to increase with concentration of Cl2 however, the resistance of the film is found to recover to its base resistance even when exposed to 8 ppm of gas indicat-
Fig. 6. Response of a typical type-I sample to H2 S, NO, Cl2, acetone, NH3 , CO and CH4 gases at 4 ppm and H2 at 2% concentration.
ing a reversible change on exposure to chlorine gas. Investigations to reduce recovery time are being carried out. Response of a typical type-I sample to many oxidizing and reducing gases is shown in Fig. 6. It is observed that these samples are quite selective for response to chlorine. The mechanism of response of Te to oxidizing or reducing gas has been reported in earlier studies [4]. Te is a p-type semiconductor and adsorption of oxygen (in normal air) leads to trapping of electrons and therefore introduction of holes in the lattice. This reduces the resistance. The resistance in further reduced on adsorption to oxidizing gases that trap electrons (such as Cl2 ), and increased on exposure to reducing gases that remove the adsorbed oxygen. 4. Conclusion The growth of Te nanotubes has been carried out via a onestep physical vapor deposition process at atmospheric pressure as well as under high vacuum conditions. Te nanotubes grown under atmospheric conditions in a horizontal furnace showed high sensitivity towards Cl2 gas at room temperature. It is found that chlorine has larger influence on grain boundary resistance compared to intragrain resistance leading to better response of nanostructure materials. References
Fig. 5. (a) Response of a type-I sensor to 2 ppm chlorine. Inset gives expanded view to show saturation at 2 ppm. (b) Response and recovery characteristics for different concentrations of Cl2 gas for another type-I sample.
[1] J.G. Lu, P. Chang, Z. Fan, Mater. Sci. Eng. R 52 (2006) 49. [2] C.N.R. Rao, F.L. Deepak, G. Gundiah, A. Govindraj, Prog. Solid State Chem. 31 (2003) 5. [3] T. Ikari, H. Berger, F. Levy, Mater. Res. Bull. 21 (1986) 99. [4] S. Sen, K.P. Muthe, N. Joshi, S.C. Gadkari, S.K. Gupta, Jagannath, M. Roy, S.K. Deshpande, J.V. Yakhmi, Sens. Actuator B 98 (2004) 154. [5] S. Sen, V. Bhandarkar, K.P. Muthe, M. Roy, S.K. Deshpande, R.C. Aiyer, S.K. Gupta, J.V. Yakhmi, V.C. Sahni, Sens. Actuator B 115 (2006) 270.
1572
S. Sen et al. / Talanta 77 (2009) 1567–1572
[6] V. Bhandarkar, S. Sen, K.P. Muthe, M. Kaur, M. Senthil Kumar, S.K. Deshpande, S.K. Gupta, J.V. Yakhmi, V.C. Sahni, Mater. Sci. Eng. B 131 (2006) 156. [7] V. Bhandarkar, S. Sen, K.P. Muthe, S.K. Deshpande, S. Bhattacharya, S.K. Gupta, Asian J. Phys. 14 (2005) 97. [8] A. Galdikas, Z. Martunas, A. Setkus, Sens. Actuator B 7 (1992) 633. [9] D.H. Dawson, D.E. Williams, J. Mater. Chem. 6 (3) (1996) 409. [10] T. Miyata, T. Hikosaka, T. Minami, Sens. Actuator B 69 (2000) 16. [11] J. Tamaki, Sens. Actuator B 83 (2002) 190. [12] T.V. Belysheva, L.P. Bogovtseva, J. Anal. Chem. 61 (2006) 494. [13] D.R. Patil, L.A. Patil, Sens. Actuator B 123 (2007) 546. [14] A. Chaparadza, S.B. Rananavare, Nanotechnology 19 (2008) 245501. [15] B. Mayer, Y.N. Xia, J. Mater. Chem. 12 (2002 1875). [16] R. Zheng, W. Cheng, E. Wang, S. Dong, Chem. Phys. Lett. 395 (2004) 302. [17] Z. Liu, Z. Hu, J. Liang, S. Li, Y. Yang, S. Peng, Y. Qian, Langmuir 20 (2004) 214.
[18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]
Z. Liu, Z. Hu, Q. Xie, B. Yang, J. Wu, Y. Qian, J. Mater. Chem. 13 (2003) 159. B. Geng, Y. Lin, X. Peng, G. Meng, L. Zhang, Nanotechnology 14 (2003) 983. P. Mohanty, T. Kang, B. Kim, J. Phys. Chem. 110 (2006) 791. P. Mohanty, J. Park, B. Kim, J. Nanosci. Nanotechnol. 6 (2006) 3380. C. Métraux, B. Grobéty, J. Mater. Res. 19 (2004) 2159. Q. Wang, G.-D. Li, Y.-L. Liu, S. Xu, K.-J. Wang, J.-S. Chen, J. Phys. Chem. C 111 (2007) 12926. H. Chen, H. Lu, Y. Nie, J. Zhang, M. Zhang, Q. Dai, S. Gao, S. Kan, D. Li, G. Zou, Phys. Lett. A 362 (2007) 61. B. Lui, H. Zeng, J. Am. Chem. Soc. 126 (2004) 8124. S. Sen, U.M. Bhatta, V. Kumar, K.P. Muthe, S. Bhattacharya, S.K. Gupta, J.V. Yakhmi, Cryst. Growth Des. 8 (1) (2008) 238. R. Swan, A.K. Ray, C.A. Hogarth, Phys. Stat. Sol. A 127 (1991) 555. U.K. Gautam, C.N. Rao, J. Mater. Chem. 14 (2004) 2530.